Advances in Inorganic Chemistry and Radiochemistry 2 (1960)

A d v a n c e s in
INORGANIC CHEMISTRY
AND
RADIOCHEMISTRY
Volume
This Page Intentionally Left Blank
Advances in
INORGANIC CHEMISTRY
AND
RADIOCHEMISTRY
EDITORS
H. J. EMELEUS
A. G. SHARPE
University Chemical laboratory
Cambridge, England
VOLUME 2
7960
ACADEMIC PRESS INC.
NEW YORK
CopyrightO, 1960, by Academic Press Inc.

ALL RIQHTS RESERVED
NO PAllT O F THIS BOOK MAY RE REPRODUCED IN ANY FORM,

BY PHOTOSTAT, MICROFILM, OR ANY OTHER MEANS,
WITHOUT WRITTEN PERMISSION FROM T H E PUBLISHERS.
ACADEMIC PRESS INC.

111 FIFTHAVENUE
NEWYORK3, N. Y.
United Kingdom Edition

Published by
ACADEMIC PRESS INC. (LONDON)LTD.
17 OLI) QUEENSTREET, LONDON 8.w. 1
Librury of Congress Cutulog Card Number: 59-7682
PBIXTED IN THE UNITED STATES OF AMERIC.4
LIST OF CONTRIBUTORS
MARGOTBECKE-GOEHRING,
Department of Chemistry, University of
Heidelberg, Heidelberg, Germany
GEORGEH. CADY,Department of Chemistry, University of Washington,
Seattle, Washing ton
J . D. DUNITZ,Organic Chemistry Laboratory, Swiss Federal Institute
of Technology, Zurich, Switzerland
JOHNEISCH,Department of Chemistry, St. Louis University, St. Louis,
Missouri ':'
OLAVFOSS,Chemical Institute, University of Bergen, Bergen, Norway
K. FRITZE,Department of Chemistry, McMaster University, Hamilton,
Ontario, Canada
HENRYGILMAN,Department of Chemistry, Iowa State College, Ames,
Iowa
.JOSEPH .J. KATZ,Argonne National Laboratory, Argonne, Illinois
C. C. MCMULLEN,
Department of Physics, McMaster University, Iiamilton, Ontario, Canada
L. E. ORGEL,University Chemical Laboratory, Cambridge, England
IRVING
SHEFT,Argonne Nationtrl Laboratory, Argonne, Illinois
F. G. A. STONE,
Department of Chemistry, Harvard University, Cambridge, Massachusetts
1. G. THODE,Department of Cheniistry, McMaster University, Harnilton, Ontario, Canada
* Present
Address: University of Michigan, Ann Arbor, Michigan

V
CONTENTS
CO~lHIHUTUHS
Stereochemistry of Ionic Solids

J . D . DUNITZ
AND L . E. ORCEL
I . Introduc-tion
. . . . . . . . . . .
I1. Ionic Coordinittion Polyhedra . . . . . . .
I11. Simple Ionic Compounds
. . . . . . . .
.
.
.
.
.
.
.
.
.
.
1
2
5
12
34
45
52
54
56
IV . The Influence of Crystal Fields on Transition-Metal Stereochemistry

. . . . . . .
V . Stercochemistry of B-Subgroup Ions
VI . Small Ion Distortions and Rattling
. . . . . . .
VII . Conclusions
. . . . . . . . . . . . .
VIII . Appcndix: Ionic Radii . . . . . . . . . . .
References . . . . . . . . . . . . . .
Organometallic Compounds
JOHN EISCHAND HENRYGILMAN

I . Introduction
. . . . . . . . . .
I1. Nrw Prep:v:itivc Trdiniqucs and Procedures . . .
I11. Structure and Bonding in Organometallic Comliountls .
.
.
.
.
.
.
IV . Reactivity and Itraction Pathways of Organometallic Conipounds . .

V . Applications of Organometallic Reagents in Research and Technology .
VI . Present Trcnds in Organometallic Rescarrh . . . . . . .
List of Symbols . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . .
61
73
83
85
98
99
99
99
Fluorine-Containing Compounds of Sulfur
GEORGEH . CADY
. . . . . .
I . Introduction
. .
. . . . . .
I1. Fluoridcs of Sulfur . . . . . . . . . . . . .
I11. Sulfur Oxyhnlides Containing Fluorine . . . . . . . .
IV . Pluorooxyacids of Sulfur
. . . . . . . . . . .
V . Other Inorganic Compounds Containing Sulfur and Fluorine . . .
VI . Compounds Containing Carbon, Sulfur, Fluorine, and Elrmcnts other
than Oxygen . . . . . . . . . . . . . .
VII . Perfluoroalkanesulfonic A d s . Their Anhydrides. Salts. and Esters
VIII . Certrtin Compounds Containing the -(SO2)Bridge . . . . .
References .
. . . . . . . . . . . . . .
105
107
115
123
128
133
140
142
148
Amides and lmides of the Oxyacids of Sulfur
MARGOT
BECKE-GOEHRING
I . Introduction
. . . . . . . . . . . . .
I1. Sulfur Imides Derived from Oxides of Sulfur in Oxidation States
between (0) and (111) . . . . .
I11. Sulfur Imides Derived from Sulfur Dioxide
vii
.
.
.
.
.
.
.
.
.
.
.
.
159
.
.
161
174
...
C!ONTENTS
Vlll
IV . Aniidcs of Oxyncids of Sulfur (IV) . .

V . Sulfur Imidcs Derived from Sulfur Trioxide
VI . Amides of Oxyacids of Sulfur (VI) . .
Refcrenccs . . . . . . . .
.
.
.
.
.
.
.
.
. . . . . . .
179
182
185
190
Halides of the Actinide Elements

Jossprr J . KATZAND IRVING
SHEFT
I . IntroductioIi
. . . . . . .
I1. Survey of tlie Actinide Element Halides .
111. Preparative . . . . . . . .
. . . . .
IV . Propcrtics of Halides
Rrferencrs . . . . . . . .
.
.
.
.
.
.
.
.
.
.
. . . . .
. . . . .
. . . . . .
195
195
202
224
233
Structures of Compounds Containing Chains of Sulfur Atoms
OLAV Foss
1. Introduc+ion
. . . . . . . .
. . . . . . .
I1 . Elemental Sulfur
I11. The Polysulfide Ions
. . . . . .
IV . The Sulfur Hydridcs and Halides . . .
V . Organic Disulfidcs :tnd Polysnlfitfes . . .
V1 . Sulfur Chains Terminated by Cyano Groups .
VII . Sulfur ChiLins Terminated by Sulfonyl Groups .
VIIT . Sulfur Chains Terminated by Sulfonate Groups:
IX . Unsymmetrical Compounds . . . . .
X . Comments on Structural Evidcnce . . .
XI . Stereochemistry
. . . . . . .
Referencrs . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
The Polytliionltlcs
. . . . . .
. . . . . .
. . . . . .
. . . . . .
237
238
240
241
243
248
250
254
265
266
267
273
Chemical Reactivity of the Boron Hydrides and

Related Compounds
F G. A . STONE
I . Introduction
. . . . . . . . . . .
IT . Reactions of the Boranes and Their Derivatives . . .
I11 . Recent Advanccs in the Chemistry of the Borohydrides .
IV . Conclusion . . . . . . . . . . . .
References . . . . . . . . . . . .
.
.
. .
. .
.
279
283
305
308
309
Mass Spectrometry in Nuclear Chemistry
H . G. THODE.
C. C. MCMULLEN.AND K . FHITZE
I . Introduction
. . .
TI. Mass Assignmcnt
. .
111. Half-life Dc.t.erminttt.ions
IV . Neutron Capture Cross Sections .

V . Fission Yields . . . . .
References . . . . . .
AUTHORINDEX . . . . .
SUR.JECT
INDEX
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
315
323
324
327
332
360
366
387
STEREOCHEMISTRY OF IONIC SOLIDS

D. Dunitz and L. E. Orgel
J.
Organic Chemistry Laboratory, Swiss Federal Institute of Technology, Zurich, Switzerland

a n d University Chemical laboratory, Cambridge, England
I. Introduction
IV. The Influence of Crystal Fields on Transition-Metal Stercoclieniistry ,

A. Octahedral Coordination . . . . . . . . . . . .
B. The Jahn-Teller Effect
. . . . . . . . . . . . .
12
12
15
23
25
28
29
33
11. Ionic Coordination Polyhedra

111. Simple Ionic Compounds .
C.
D.
E.
F.
G.
Square Coplanar d* Complexes . . . . . . . .

Tetrahedral Coordination . . . . . . . . . .
Octahedral vervusTetrahcdra1 Coordination . . . .
The Structure of Compounds with the Spinel Arrangement
Concluding Remarks on Jahn-Teller Distortions . . .
V. Stercochemistry of B-subgroup Ions. . .

A. Ions with do Closed Shell Configurations
B. The 1nc:rt Pair . . . . . . .
VI. Small Ion Distortions and Ratt,ling .
VII. Conclusions .
VTII. Appendix . .
A. Ionic Radii .
References .
.
.
.
.
.
.
.
.
.
.
.
34
34
40
45
.
,
52
.
. . . . . . . . . . . . . . . . .
54
54
5lj
I . Introduction
Once upon a time it was believed that the structures of most binary
solid compounds of the metals, other than alloys and related substances,
could be described adequately in terms of an ionic model in which metal
ions and anions are treated as charged, more or less compressible, spheres.
This model leads to a classical electrostatic theory of bonding and is relatively well understood. Naturally, modern developments in vslcncy theory
have tended t o emphasize those situations in which a simple ionic theory is
inadequate and so have occasionally led to the multiplicatioii of hypotheses
to explain observations which could just as well he interpreted in terms of
an electrostatic theory. This has tended to obscure the fact that the ionic
1
,I. D. 1)IJNITX . 4 N D 1,. E. ORGBL
model still forms il satisfactory background to t i great part of the theory

of the stereochemistry of metal compounds.
I n this review wc shall emphasize the ionic viewpoint, but not because
we are undware of the importance of covalent bonding. We believe that only
after the irrelevance of much that has been presented as qualitative stereochemical evidence for covalency has been appreciated, does the importance
of quantitative studies, for example of unpaired electron distributions by
paramagnetic resonance experiments, become clear. Our program, then, is
to interpret as Inany features as possible of the stereochemistry of metal
compounds from the point of view of an ionic theory; and only incidentally
to draw attention to the modifications that are required if the covalent
effects (which, to a greater or lesser extent, influence the properties of all
compounds) are to b~ taken into :wcount.
We slid1 for the most part illustrate our general arguments by examples
of typically ionic solids; oxides, halides, hydrates, etc. However, it must
he appreciated that fwtors important in determining the structures of
ionic solids also decide the stereochemistry of discrete molecules and complex ions. WP shall not, therefore, hesitate to discuss the stereochemistry
of isolated molecules or complex ions in solution insofar ils they are relevant
to tlie generid problem of the stereochcmistry of met a 1 ions.
II. Ionic Coordination Polyhedra (76, 147)
Wc first examine the iniplicatioiis of a very much oversiniplified ionic

theory. Wc suppose that:
(1) Ions w e charged, incompressible, lionpolarizable spheres.
(2) An arrangement of ions of one charge about tin ion of the opposite
charge is stable only if the central ion is in contact with each of its neighbors. This places a lower limit to the ratio of the radius of the cwitral ion
to that of the surrounding ions for each type of coordination polyhedron
(Table I). Since positive ions are almost always smaller than the negative
ions which accompany them in crystals this restriction means effectively
that the coordination number of small metal ions is often limited by the
radius ratio rules, while the coordination numbers of anions are rarely
limited in this way.
(3) The coordination number is as large ah possible, subject to condition ( 2 ) .Of course, the coordination number may also be restricted by the
composition of the substance if it is necessarily present as discrete molecules, for example Tic&
in the gas phase.
(4) The arrangement of the coordinated groups minimizes the electrostatic repulsion energy between them.
It follows from (4)that an ideal MX2 complex should be linear, for this
STI3HEOCHEMISTRY OF I O S I C SOLIDS
is the arrangement which maximizes the X-X distance and so minimizes

the electrostatic repulsion between X ions. In a similar way the repulsion
between X ions in an MX3 complex is minimized if the molecule is planar
and the X ions are a t the vertices of an equilateral triangle. There are two
other cases in which the preferied configuration depends only on symmetry,
namely MX4 and MXs complexes which should adopt regular tetrahedral
and regular octahedral structures respectively. These symmetry determined
arrangements are illustrated in Fig. 1.
(b)
(c)
FIG. 1. Symmetry determined coordination polyhedra: (a) planar equilateral triangle, (b) tetrahedron, (c) octahedron.
I n all other important cases the optimum configuration is not determined by symmetry alone. Although it is possible to arrange eight equivalent X anions at the vertices of a cube about a central M ion as in Fig. 2a,
this is not the configuration which minimizes the electrostatic repulsion
energy, for clearly the square antiprism shown in Fig. 2b is more favorable.
(a)
(b)
FIG. 2. Possible arrangements for 8-coordination: (a) cube (the CsCl structurc:),
(b) quart: Archimidean antiprism.
The only configurations of MX5, MX?, and MXs groups which maintain
the equivalence of all M X bonds are the energetically unfavorable planar
ones. Some of the configurations which lead to relatively favorable electrostatic energies are illustrated in Fig. 3.
So far we have made use of condition (4) t o determine the ideal configuration for each type of MX, complex. Now we must determine for each
coordination number the values of the ratios of the radii TM and rx of the
M and X ions which are consistent with (2) and (3). Clearly a linear MX2
arrangement (but not a nonlinear one) satisfies (3) for any value of the
radius ratio. On the other hand in a planar MX3complex the M ion would
rattle inside the triangle of X ions if the ratio r M / r x of the two ions were
less than 0.155. Thus according to (2) a very small A ion cannot form an
MX, complex with n large X ion, but must instead form an MX2 complex.
(a)
0))
(c)
(di
FIG.3. Some favorable arrangernerits for MXs, MX7, and MXBgroups: (a) square
pyramid (note that thc square of ligands is not necewnrily coplanar with the metal ion),
(h) trigonnl bipyramid, (c) octahedron distorted by expanding one face and inserting
an vxtm ligand on thc line through its center, (d) trigonal prism with additional ligands
on the linw through the centers of the thrcc rectangular fnccs.
In the s:tnie way it follows from (3) that an MX3 complex should only be
stable if the radius ratio lies in the range 0.155 6 rM/rx 6 0.225. Yrocceding in this way me can derive thc ranges of the radius ratio consistent with
ench type of coordination as shown in Table I.
LIMITINQ
TABLE I
RADIIJE
RATIOSFOR VARIOUS
Polyhcdrori
Equi1:ttcral triangle
Tetrnhcdron
Trigon a1 bi prism
Sqiiarc pyramid
Octahedron
Square hrc1iiniidc:m nntipriarri
Cube
COOR1)lNATION POLYHELlRA
Coordination
number
Minimum
radius ratio
0.155
0.225
0.414
0.414
0.414
0.645
0.732
The MX6 complex may now be considered in the light of conditions (2)
and (3). I n a trigorid bipyrsmid the angle between axial and radial bonds is
just go, that is the X-X distance is no smaller than in the octahedron.
Thus (3) leads us to expect that the octahedron will bo preferred to the
trigonal pyramid. An exactly similar argument shows that the octahedron
will also be preferred to a square prism. Thus, except in isolated molecules
or ions, we should not expect t o find MXI,arrangements.
This discussion shows that the widespread occurrence of regular tetrahedrally and octahedrally coordinated metal ions calls for no special interpretation in terms of bond hybridization, and so forth. It is rather the other
four or six coordinated structures which require espl:i,iiation; for cs:tniplr,

we shall need a special explaiitltioti of the widesprcttd occurrence of planar
coordination among the conipounds containirig the Ni++, Pd++ and Pt++
ions.
The arguments from electrostatic theory apply formally only to substances made up froin monoatornic ions. However, there is reason to believe
that they are qualitatively correct in a much less restricted context. Thus,
although it cannot be proved generally arid rigorously, it is very plausible
that regular octahedral and tetrahedral coordination are the preferred
arrangements for the disposition of most dipolar molecules or polyatomic
anions around a central positive ion. They are simply the arrangements
which place the ligands as f~ apart as possible, and since ligands usu:illy
repel one mother they are to be expected if the metal ion behaves as a
spherically symmetric charge distribution. Thus the octahedral character
of [Co(NH,)@+ no more implies covalent bindirig than does the octahedral
environment of Naf in NaCI. Conversely, the planarity of [Pt(NH,)d]++
requires comment in just the same way as does the square coplanar enviroiiment of Pt++in PtO.
I l l . Simple Ionic Compounds
The compounds which according to almost any theory should be the

most nearly ionic in character are the alkali halides. It is, therefore, particularly interesting to compare their structures with those predicted by
our simple theory. We see in Table I1 that only the NaCl rocksalt (Fig. 4 )
TABLE I1
O F ALKALIHALIDES(I6U).
RADIUSRATIOSAND CRYSTALSTRUCTURES
WHERETwo STRCJCTURES
ARE LISTED,THE FIRSTREPRESENTS
THE S T A R L E M o I ) r m x r r I O N
I,i
iuiL
1;
IZb
cs
c1
Br
I
a
Rocksalt structure (Fig. 4).

CsCl structure (Fig. 2a).
arid the CsCl structures (Fig. 2a) occur. In the NaCl structure each cation
is surrounded by a regular octahedron of anions and each anion by a regular
octahedron of cations, and the structure conforms to our deductions about
stable coordination polyhedra. On the other halid, the coordination about
J. D. DUNITZ A N D L. E. ORGEL
e:wh ioii iii the CsCI structure is cubit: and so docs riot correspond to the
ideal arrangemerit for isolated A& groups.
A study of a iiurriber of other compounds which arc generally considered
to be among the most completely ionic shows that cubic coordination is not
confined to the alkali halides. It occurs, for example, in the fluorite structure in which niany ionic MX2 compounds crystallize. On the other hand
the isolatcd [laFsI3- group has the square antiprism arrangement (64)predicted by the simple theory. This underlines a point of great importance to
the understariding of ionic crystal structures, namely, that the requirement
that a structure can be extended indefinitely in space imposes severe restrictions on the types of coordination which are possible. Cubic coordiriation
can be extended indefinitely, but it is not possible to form an extended
FIG.4. The rocksalt structure.
structure in which every ion is surrounded by a square antiprism of ions

of the opposite charge. Thus a cubic structure is stable for a crystal as a
whole, although it does not provide individual ions with their optimum
environments; the stability of individual coordination polyhedra is sacrificed in the interest of long-range order.
To proceed further with our comparison of theory with experiment we
require values of the radii of the ions concerned. The values recommended
by Pauling (114) are given in Table I11 and a discussion of their derivation
in the Appendix.
A comparison of thc limiting rsdius-ratio values given in Table I with
the radius ratios of alkali halides in Table I1 shows that the observed
structures are not always as predicted. According to the simple theory, the
rocksalt structure should be stable only within the range 0.414 6 T M / T X
6 0.732. Thus LiCI (0.33), LiBr (0.31), and LiI (0.28) should have tetrahedrally coordinated structures, while KB (0.98), RbB (1.09), and CsF
(1.24) should have eight or twelve coordinated structures. These disagreements are perhaps not surprising in view of the crudeness of the approxirnatioris involved, but more realistic models do not lead immediately to an
explanation of the persistence of the rocksalt structure not only in the alkali
TABLE I11
CRYSTALRADIIA N I ) UNIVALENT
CRYSTALRADII(IN BRACKETS) OF CERTAIN
IONS.AFTER PAULINC
(114). VALUESIN ANGSTROM
UNITS
Li+ 0.60
Na+ 0.95
K+ 1.33
Rb+ 1.48
Cs+ 1.69
cu+ 0.96
Ag+ 1.26
Au+ 1.37
Ti4+ 0 . 6 8 [O ,961
Zr4+ 0.80 11. 091
Ce4+ 1.01 [ I . 271
Ge4+ 0.53 [O. 761
Sn4+ 0.71 [0.96]
Pb4+ 0.84 [I.Oti]
TI+ 1.44
Bo++ 0.31 [0.44]

Mg++ 0.65[O. 821
Cat+ 0.99[l.18]
Sr++ 1.13[1.32]
Ba++ 1.35 11.531
Zn++ 0.74[0.88]
Cd++ 0.97[l.141
Hg++ l.lO[1.25]
F- 1.36
C1- 1.81
Br- 1.95
I- 2.16
Empirical ionic radii

Mn++ 0.80
Fe++ 0.75
Co++ 0.72
Ni+' 0 . 7 0
A13+ 0.50[O.721
sc3+ 0.81 [1.06]
Y3+ 0.93 /l.20]
La3+ 1.15 [l ,391
Ga3+0.62(0.811
In3+ 0.81 [1.04]
T13+ 0.95 [l.151
O-- 1.40[1.76]
S-- 1.84 I2.191
Ti3+ 0.69
V3+ 0.66
Cr3+ 0.64
Mn3+ 0.62
FcJS 0.60
halides but also in silver halides, including AgF (0.93), and in oxides and
sulfides of the alkaline earths, and of many other divalent metal ions, often
with radius ratios far outside the predicted limits (Table IV). On the other
hand, the CsCl structure is confined under ordinary conditions to CsCI,
CsBr, CsI, and the ammonium halides (where it is stabilized by a favorable
hydrogen bonding arrangement). Since the Madelung constant for this
structure is somewhat greater than for the 6-coordinated (1.763 compared
with 1.746), one would certainly expect its more widespread occurrence in
compounds with suitable radius ratios than is observed.
The most important NIX structures involving tetrahedral coordination
are the cubic ZnS sphalerite (Fig. 521) and the hexagonal ZnS wurtzite
(Fig. 5b) arrangements. It is striking that halides and sulfides of metal
ions with d6 and dl0 shells have a tendency to crystallize in the sphalerite
structure; for example, the cuprous halides, AgI, HgS, MnS, CdS, and
ZnS. (The last three also occur in the wurtzite modification, as do the
oxides of Zn and Be). (See Table V.) Here again, the simple ionic theory
fails to account for the facts for (1) the radius ratios of some of these compounds are compatible with a 6-coordinated structure, and (2) interatomic
distances calculated from the usual ionic radii (decreased by 5y0 to com-
J. I). DUNITZ A N D L. E. ORGEL
TABLE IV
THE ROCKSALT
STRUCTURE (160)
S O M E I O N I C CRYSTALS WITH
Ionic
radius sum m/rx
-
KJ3r
KI
RbF
RbCl
IibI3r
RbT
CSF
CsCl
AgF
AgCl
AgBr
Ionic
radius sum
(A)
TY/TX
(A)
(A)
~-
LiF
IdCI
I,il%r
LiI
NaI'
NaCl
NaRr
NaT
1<F
KCI
M-X
Compound
01
67
75
00
31
2 81
2 98
3 23
2.67
3 14
3 29
3 52
2 82
3 27
3 43
3 66
3 00
3 51
2 46
2 77
2 89
2
2
2
3
2
1.96
2.41
2.55
2.76
2.31
2.76
2.90
3.11
2.69
3.14
3.28
3.49
2.84
3.29
3.43
3.64
3.05
3.50
2.62
3.07
3.21
0.44
0.33
0.31
0.28
0.70
0.52
0.49
0.44
0.98
0.73
0.68
0.G2
1.09
0.82
0.76
0.69
1.24
0.93
0.93
0.70
0.65
2.05
2.39
2.53
2.75
2.20
2.15
2.12
2.10
2.37
0.46
0.71
0.81
0.96
-
MnO
FeO
COO
NiO
CdO
TaO
2.10
2.41
2.57
2.76
2.12
2.04
2.22
2.16
2.13
2.10
2.35
2.21
MnS
PbS
2.61
2.96
2.64
0.43
-
MgO
CaO
SrO
BaO
Ti0
vo
__
0.57
0.53
0.51
0.50
0.69
-
pensate for the change from 6- to 4-coordination) do not agree with the
observed distances in a number of cases. The largest discrepancies are
observed for compounds of Cu+, Agf, and Hg++, and, as we shall show in a
later section, it is just for these cations that deviations from the predictions
of the simple ionic picture should be most marked. However a satisfactory
and detailed interpretation of these structures is still wanting.
(a)
(b)
FIG.5. Tetrahedrally coordinated MX structures; (a) sphalorite, (b) wurtzite.
TABLE V
SOME CRYSTALS WITH TETRAHEDRALLY COORDINATED STRUCTURDS
Compound
Be0
BcS
ZnO
ZnS
CdS
HgS
MnS
CuF
CUCl
CuRr
cur
Ad
Sphalerite
M-X
Wurtzite
M-X
(A)
-
2.10
-
2.36
2.52
2 53
2.43
1.85
2.35
2.46
2.62
2.80
(160)a
Ionic radius
sumb
(4
d - x
(A)
1.64, 1.65
1.95, 1 . 9 8
2.33, 2.33
2.51, 2 . 5 3
2.41, 2 . 4 4
-
1.62
2.04
2.03
2.45
2.67
2.79
2.50
2.20
2.63
2.76
2.96
3.24
0.22
0.17
0.53
0.40
0.53
0.60
0.43
0.71
0.53
0.49
0.44
0.58
-_
-
The interatomic distanccs in these compounds can be closely reproduced by a special

set of covalent radii (114, p. 179).
The ionic radius sum has been multiplied by 0.95 to allow for the change from 6- to
4-coordination (see Appendix).
When we turn t o ionic solids of the composition MX2 we fi.id that predictions based on the simple ionic model are fulfilled surprisingly well. The
two principal structural types are illustrated in Fig. 6. In the fluorite structure, the cation M is surrounded by 8 anions in a cubic arrangement, and
the anion X is surrouiided by 4 cations a t the corners of a tetrahedron. We
expect this structure to be stable only for values of the radius ratio greater
than 0.73 and in general this is the case. Moreover, the observed M-X
n
(4
(b)
FIQ.6. Ionic MXI structures (a) fluorite (b) rutile. I n (\) the shaded and unshaded
atoms occur in alternate layers separated by about 1.5-1.7 A.
10
J. D. DUNITZ AND L. E. ORGEL
distances agree well with those calculated from the ionic radii (Table VI).
For only one crystal ZrOz (0.62), is the radius ratio smaller than the required minimum, and it is noteworthy that it adopts the fluorite structure
only in an unstable high-temperature modification. The alkali metal oxides,
M20 (except CspO),crystallize in the so-called antifluorite arrangement,
in which the positions of anions and cations are reversed relative to those in
the fluorite structure.
SOME CRYSTALS
WITH
Compound
CaF2
SrF2
BaD2
Cd R
Ibkz
HgFz
SrClz
nscI2
ThOz
UOz
CeOz
Pro2
HfOz
Zr02
a
THE
TABLE VI
8-COORDINATED
FLIJORITE
STRUCTURE
(160)a
M-X
Ionic radius
(A)
morn
2.36
2.51
2.68
2.33
2.56
2.40
3.02
3.18
2.42
2.37
2.34
2.32
2.21
2.19
2.35
2.49
2.71
2.33
TMITX
(A)
__
2.46
2.94
3.16
-
2.41
-
2.20
0.87
0.97
1.13
0.84
0.92
0.73
0.84
0.72
-
0.62
The radius ratios are calculated from the univalent radii (see Appendix and ref.
114, p. 365).
Fluorides and oxides of composition MX2 adopt the 6-coordinated

rutile structure when the radius ratio is less than the critical value, 0.73.
Since this structure is tetragonal, the crystal symmetry does not impose a
perfectly rcgular octahedral environment on the metal ions. Nevertheless,
the observed deviations are usually quite small (Table VII). The largest
deviation from octahedral symmetry occurs in FeF2, where, according t o a
recent analysis (9),four bonds in a plane are significantly longer (0.13 8)
than the remaining two axial bonds. We shall return to this point in a later
section. The rutile structure should be unstable if the radius ratio is less
than 0.414 and accordingly we find that GeO2 with rM/rx on the limit
crystallizes also in the tetrahedrally coordinated quartz structure (150).
Quite different octahedrally coordinated structures are observed in the
bromides, iodides (and some chlorides) of divalent metal ions with suitable
radius ratios. In these, the MX6 octahedra are not linked in a three-dimen-
11
sionul network but in two-dimensional sheets, separated from each other

only by anion-anion contacts. The halide ions are thus shielded from each
other by cations on one side only. This type of structure is incompatible
with the simple ionic model but is readily understood in terms of Van der
Waals attraction between the large, highly polarizable, anions if it is supposed that the charge on the anions has been greatly reduced by polarization
or covalent bonding.
TABLE VII
SOME CRYSTALS WITH THE 6-COORDINATED
RUTILESTRUCTUREa
Ionic radius
Compound
SUJrrl
(A)
2.13,2.10
2.12,1.99
2.05,2.03
2.02,1.99
2.04,2.02
2.00,l.YS
2.16, 2.17
1.94,1.99
2.05,2.06
1.87,1.91
0
Bracketed values of
TM/TX
2.16
2.11
2.ox
2.0B
2.10
2.01
2.08
2.11
1.93
0.59
0.55
0.53
0.52
0.54
0.48[O.CiO]
-
0.49 I0.551
0.51[0.551
0.38[O.431
are calculated from the univalent radii.
Our review of simple ionic compounds shows that the radius-ratio rules
are a useful guide to the stereochemistry of halides and oxides, but that the
quantitative predictions are not to be taken too seriously. This is no doubt
in part attributable t o the crudeness of the model from which the rules are
derived, and in part to uncertainties in the values of the ionic radii. There
are many examples in which the coordination number increases with increasing radius of the metal ion in closely related compounds; for example,
Mg++ in MgFz is 6-coordinated but CaFz has a n 8-coordinated structure.
It is in accounting, at least semiquantitatively, for such relations that the
radius-ratio rules are most important.
I n the simple structures which we have discussed so far the local
environment of the central metal ion has been tetrahedral, octahedral, or
cubic. We can describe all of these as being of cubic symmetry since they
include some or all of the characteristic symmetry elements of a cube.
They possess the maximum symmetry compatible with the coordination
number and, with the reservations already noted for eightfold coordination,
are just those environments which would follow from the properties of the
simple ionic model described in the previous section. There are many
12
J. D.
DUNITZ AND L. E. ORGEL
structures where the symmetry of the local eiivironment of the central metal
ion is lower than cubic. We shall discuss some of these in the following sections, and shall usually adopt the view that they correspond to distortions
from the cubic symmetry which is to be expected from the simple ionic
model.
IV. The Influence of Crystal Fields on Transition-Metal Stereochemistry
In this section we shall develop the electrostatic crystal-field theory to

a sufficient extent to give an account of its stereochemical implications
particularly with reference to the distortions from cubic symmetry which
are characteristic of certain groups of transition-metal compounds. The
treatment is not meant to be complete and the reader is referred to the
many reviews (51, 72, $7, 102, 106) for detailed references, derivations,
qualifications, disputations, and applications.
According to this theory, a transition-metal ion in a compound or complex is supposed to be subjected to an electrostatic field produced by the
molecules and ions in its environment, particularly those in its first coordination sphere. Since the ligands which we consider are invariably either
negative ions or polar molecules so oriented that their unshared pairs point
directly a t the metal ion, the field is always roughly equivalent to that due
to a set of negative point charges placed about the metal. The electrostatic
theory is concerned with the effect of this field on the orbital energies of the
metal d electrons.
A. OCTAHEDRAL
COORDINATION
We consider first an octahedral arrangement of negative charges about
the metal. The five d orbitals have the forms indicated in Fig. 7, in which
the coordinate axes are taken to lie along the bond directions. It is clear
that the d,, and dZt+ orbitals have substantial amplitudes in the bond
directions, but that the a,,
and d,. orbitals tend to avoid these directions. The energy of an electron in the d,, or dZl+ orbitals will therefore be
raised substantially by the repulsive field of the ligands while the energy
of an electron in the d, dug,or d,, orbitals will be less affected. It is obvious
that d, dyi, and cl,, orbitals have an identical environment and hence are
degenerate, that is, they have the same energy. And it can be shown by
group theory or by direct calculation that the d,, and dz1+ orbitals are also
degenerate. Thus the five d orbitals in an octahedral field split into two
groups, a lower group of three and an upper group of two, the two groups
usually being designated by the labels t z p and ep respectively (or sometimes
by y 6 and y3 or by d, and dy respectively). The energy separation between
13
4'
FIG.7. Cross sections of the five d orbitals chosen in real form (61,Fig. 1).
the two groups of orbitals is usually designated as A, and the energies of

the individual orbitals are
E'(t2,)
and
E(e,)
+ 35 A
-
referred t o the weighted mean energy in the electrostatic field as zero.

We now consider the distribution of electrons on metal ions in octahedral environments. If only one d electron is present it must clearly occupy
one of the more stable set of tzs orbitals. If more than one d electron is present, however, complications may arise for in such cases there are two
competing tendencies a t work.
(1). The electrons tend, as far as possible, t o occupy the more stable
t2* orbitals in preference to the less stable e, orbitals.
(2). The electrons tend to go into different orbitals with their spins
parallel, since this results in a lower intra-atomic repulsion energy between
them and in a more favorable exchange energy.
14
J. I). DUNITZ AND L. E. ORGEL
With two, or three d electrons present, both of these tendencies can be

satisfied simultaneously by placing the electrons in the different tzg orbitals
with their spins parallel. For four to seven d electrons there are two favorable arrangements : one which places the maximum number of electrons
compatible with the Pauli Principle in the more stable tzO orbitals, and the
other which maintains a maximum number of unpaired spins. With more
than seven electrons there is again a unique arrangement which gives at
the same time the best distribution among the orbitals and the maximum
number of unpaired spins. This is illustrated in Table VIII.
In the case that the metal has four to seven d electrons it is clear that a
large value of A favors the low-spin configurations for if A is large enough
it must finally force the electrons into the tzQ orbitals, notwithstanding the
opposed tendency to maintain a maximum number of unpaired spins. For
this reason the low-spin configurations are often said to correspond to the
strong field situation. In a similar way, the exchange forces must dominate
and maintain a maximum number of unpaired electrons if A is sufficiently
small. High-spin complexes therefore correspond to the weak field limit.
TABLE VIII
d ELECTRON
ARRANGEMENTS
IN OCTAHEDRAL
COMPLEXES
High spin
(weak field)
Number of
d electrons
Low spin
(strong field)
t20
T T
t t t
T T T
t
ti
~
r r
T1 Tl t
Tl T1 ti
Ti TI tl
T
T
T
T
T
T T
t T
tl
e,
12P
PU
T
tl
tl
Gain in
orbital energy
in strong field
0
0
T t
M t
M tl tl
Tl tl Tl
2A
2A
A
0
0
The complex [ C O ( N H ~ ) ~
provides
] ~ + a good example of the strong field
case. The ground state is (tzs)'j with all d electrons paired and the complex
is accordingly diamagnetic. In the ease of [Fe(H20)$+, which is formally
isoelectronic with [Co(NH3)6I3+A is not large enough to break down the
tendency of the d electrons to maintain their spins parallel, so that we find
a paramagnetism corresponding to four unpaired electrons, that is, the configuration is (t2g)4(eg)2,as shown in column 2 of Table VIII.
In the purely electrostatic theory which has been presented, covalent
bonding between the ligands and the metal ions has not been taken into
STEREOCHEMISTRY O F IONIC SOLIDS
15
consideration. I t can easily be shown however, that molecular-orbital theory

leads t o exactly the same kind of splitting of the d orbitals, into a lower,
nonbonding t2, set of three degenerate orbitals, and an upper, antibonding
pair of e, orbitals. The inclusion of covalent bonding in a more complete
theory (ligand field) would only affect the magnitude of the energy separation which is, in any case, usually determined empirically, from analysis of
molecular spectral data. It is found from such analyses that A varies somewhat from metal to metal, and also depends on the nature of the ligands.
For hydrates and oxides of divalent ions of the first transition-metal series
it is in the range 7,500-12,500 cm-', and for those of the corresponding
trivalent ions, in the range 14,000-21,000 cm-l. For compounds of the
second and third transition-metal series, A is substantially larger than in
the first series. We may also note that the magnitude of pairing energies
can also be estimated empirically from atomic spectral data.
B. THE JAHN-TELLEH
EFFECT
So far, it has beeu assumed that the ligands around the metal ion are in
a regular octahedral arrangement. We now go on to show that this is
not possible for certain configurations of the d electrons. The theory is essentially that of Jahn and Teller (68),but we shall develop it in a qualitative
way rather than in the elegant form in which they first presented it. Historically, their ideas were originally applied to the interpretation of the
magnetic properties of transition-metal compounds (13'7); the stereochemical applications are much more recent (35, 104).
The physical basis of the Jahn-Teller effect is best illustrated by an
example. As we have seen, the ground state of the cupric ion (d9) surrounded
by a regular octahedron of ligands would be (t2s)6(e,)3.The vacancy in the
d shell can occur either in the d,, orbital or in the dZLy2orbital, so that the
ground state is degenerate. Let us first suppose that the hole is in the
dZL,, orbital. I n this case the nuclear charge of the ion is more completely
shielded in the z-direction than in the x- and y-directions; that is, the atom
develops a greater electronegativity in the xy-plane than perpendicular
to this plane (Fig. 8). It follows from this anisotropic electronegativity that
the force acting on the ligands in the xy-plane is more strongly attractive
than that acting on the ligands in the zdirection. Thus if the ligands in
the z-direction are in equilibrium there will still be an extra attractive
force acting on the ligands in the xy-plane, which will tend to pull them in
to a shorter distance. We see that in equilibrium the distance t o the ligands
in the zy-plane must be shorter than that to those along the z-axis; that is,
four short, and two long bonds are formed. If the hole in the e, orbitals is
in the d,, orbital then exactly the opposite result follows, namely that two
16
J. D. DUNITZ AND L. E. OltGEL
short bonds to ligands in the z-direction and four long bonds in the sy-plane
are formed.
We have considered the two simplest cases, but it should be remembered
that the vacancy in the d shell could also occur in an orbital which was an
arbitrary mixture of dZLIIand d,, orbitals, in which case the distortion
would be correspondingly more complicated. It turns out that while it is
easy to show that some distortion must occur, it is very difficult to be sure
which distortion produces the most stable equilibrium configuration. Opik
and Pryce (105) have concluded that the configuration with four near
b
FIO.8. Anisotropic electronegativity arising in an octahedral complex for a
with a hole in the d z 1 4 orbital.
d9
ion
neighbors and two distant ones should always be most stable, but Liehr
and Ballhausen (80) have questioned this conclusion. Experimentally, all
but one of the examples studied crystallographically agree with the conclusion of Opik and Pryce, and it is not yet clear whether the single exception, K2CuF4 (18),is due to an inadequacy of their theory for regular
octahedral complexes or to the noncubic environment of the metal ion
enforced by the nature of the basic crystal structure, which even in its
undistorted form imposes a strongly tetragonal rather than cubic symmetry
on the metal.
There is one important characteristic of the Jahn-Teller mechanism
that is independent of detailed models. If the unstable symmetric configuration of an isolated complex has a center of symmetry, this will be preserved
in the distorted configuration. This is not necessarily the case with the
near degeneracy instability to be discussed in later sections.
There is another way of looking at the Jahn-Teller effect in de complexes
which, while equivalent to the foregoing discussion, throws further light on
Recent work has shown that the structures of KCuFs (Sh, 66a), RbCuFa (65a),
and KCrFB(SSc;., 6%) are tetragonal distortions of the cubic perovskite structure with
c/a < 1. This new evidence seems to support the conclusions of Liehr and Ballhausen,
but a final decision on the matter must be postponed until more detailed rcsults become
available.
STEHEOCHXMISTHY OF I O N l C SOLIDS
17
the principles involved. In Fig. 9 we show the d orbital energy level diagram
for a regular octahedral arrangement of ligands, and for a n octahedral
arrangement slightly distorted in such a way as t o produce four shorter
bonds in the xy-plane and two longer bonds along the z-axis. If we have to
accommodate niiic d electrons in the levels of the undistorted structure the
FIG.9. Splitting of d orbital energy levels in a regular octahedrd fidd and in an octahedral fidd distorted to produce four shorter bonds in thc zy-plane and two longer bonds
along t h e z-axis.
orbital energy is just -8A. If, however, we use the levels of the distorted
structure, the orbital energy may be lowered to -+A - *& by leaving the
hole in the d shell in the dSy2 orbital.l"Thusthe restoring forceswhich would
maintain a regular octahedral environment if the central ion had a filled
shell are opposed by forces which tend to distort the octahedron so as to
lower the d electron orbital energy. The detailed theory shows th a t the distorting forces must always prevail.
At this point we can conveniently generalize our arguments to apply
to other ions with incomplete d electron shells (36, 87). First, we observe
that filled, or half-filled shells of tzsor ep electrons do not lead t o instability.
Consider, for example, two electrons with parallel spins in the e, pair of
orbitals. If one electron occupies the dz2orbital, the other must occupy the
dZ2-p orbital. Hence the ion is equally shielded in all bond directions and
exerts the same attraction on each ligand. In terms of the energy level
diagram, one electron would be stabilized by distortion by a n amount
+& but the energy of the other would be increased by exactly the same
lo A complementary argument applies if the distortion is reversed t o give four long and
two short bonds.
18
J. D. DUNITZ A N D L. E. OltGEL
amount, so thtit there is no net tendency to distort. We see, therefore, that

for the high-spin configurations of d3, d6, d8, and dLoand for the low-spin
configuration of d6 (Table VIII), a regular octahedral arrangement of
ligands is stable against distortion. On the other hand, the high-spin
configuration of d4: (t2,)3(e,>1 and the low-spin configuration of d7: (t2u)6(e,)1
are similar to the d9 configuration in that they have a degeneracy involving
the d,, and dZLy2orbitals. Consequently, similar distortions of the octahedral environment are to be anticipated.
This leaves for consideration the high-spin configurations d1: (t2g)1,
d 2 : ( t 2 , ) 2 , d 6 : (t2u)4(e,)2, d': ( t ~ , ) ~ ( eand
, ) ~the low-spin configurations d4: (t2g)4
and d5: (LJ6 which have degeneracies involving the three tzsorbitals. Arguments similar to those already given show that distortions from the regular
octahedral environment must again lead to a lowering of the orbital energy.
It is evidmit, however, that for tz, orbital degeneracies, the sense of the distortion corresponding to minimal orbital energy should depend on the number of electrons in the t 2 , level. Thus, the arrangement with four short and
two long bonds leads to the energy level diagram of Fig. 9. For one (or
four) tzu electrons the amount of extra stabilization arising from this distortion is 66,. It is easy to see that for the opposite distortion the corrcsponding stabilization would be twice as great and hence that ions with
(tz,)' or (t2u)4configurations should have two near and four more distant
neighbors. For two (or five) tar electrons the situation is reversed. The
preferred environment corresponds to that indicated in Fig. 9 with a stabilization of #al while the opposite distortion produces a lowering of the
orbital energy of only
These conclusions cannot, however, be taken very
seriously on account of an important distinction which exists between the
e, arid t 2 , orbitals. As we have already noted, the pair of e, orbitals point
directly a t the ligands while the three tzo orbitals point in directions which
lie between the ligands. This means that a change from regular to distorted
octahedral environment has a much smaller effect on the energies of the
tp, orbitals than on those of the e, orbitals. Thus, although the qualitative
considerations outlined above are still valid, the magnitude of a1 is quite
small compared with that of 6,. This implies that not only should JahnTeller distortions due to tzodegeneracy be much smaller than those due to e,
degeneracy, but also that various factors, particularly spin-orbit coupling,
which we have not taken into consideration in this elementary account
will have a much greater relative importance and may affect even our qualitative conclusions. Furthermore, there is no good basis for oux assumption
that Jahn-Teller distortions necessarily lead to tetragonally distorted
environments; they might be trigonal. The position, then, is that the JahnTeller distortions anticipated for tzo degeneracies are difficult to calculate
and are, in any case, small.
19
There is, perhaps, oiily one example where an appreciable distortioii

from octahedral symmetry might be interpreted as a result of t~ degeneracy.
As we have seen from Table VII the largest deviation from octahedral symmetry in the transition-metal fluorides with the rutile structure occurs in
FeF2, where the Fe++ ion is surrounded by 4 F- at 2.12 8 and by 2 F- at
1.99 A (9). This is just the kind of distortion which was predicted by the
simple theory for the high-spin d6: (t2u)4(eu)2 configuration of the Fe++ ion.
On the other hand, the opposite distortion which would be expected in
CoF2 does not seem to occur. It has also been suggested that the very small
distortions from the rocksalt structure which occur a t low temperatures
in FeO (132) and COO (49) may be attributable to a Jahn-Teller mechanism
C
(1)
(2)
(3)
FIG.10. Comparison of structures of CoC12 . 2H20 and CuC12 2IIzO: (1) structures
projected down respective c-axes, (2) projection of CoClz . 2Hz0 chain down b-axis,
(3) projection of CuCl2 . 2H20 chain down b-axis (35).
(74). But complex magnetic ordering processes are associated with these
transformations (79), and the distortions from cubic symmetry must to
some extent be due to the anisotropic nature of the interactions between
metal ions. Moreover, if the Jahn-Teller mechanism is to be invoked as an
explanation of the distortions in FeO and COO, to what are we to ascribe
the very similar distortions which occur also in MnO (132) and NiO ( l 2 1 ) ,
where the Jahn-Teller effect should produce virtually no disturbance of
the regular octahedral environment? It would appear that neither theoretical considerations nor the available experimental evidence indicate that the
Jahn-Teller effect arising from tzU degeneracies is an important factor in
the gross stereochemistry of transition-metal ions, at least of those of the
first series.
The distortions from octahedral symmetry arising from ep degeneracy
are much more important. In the case of cupric compounds, such distortions have been recognized for many years, particularly in their more
extreme manifestations, as, for example, in CuO (tenorite) where the Cu++
ion is surrounded by 4O-- anions in a square planar arrangement(133).
20
J. I). UUNITZ AND L. E. ORGEL
Square coordination of this kind has 1,ceii iiiterpretcd as a consequence of

dsp2 covalent bonding and has been regarded as more or less characteristic
of cupric compounds. A review of the environment of the Cu++ ion in a
number of structures reveals, however, that square coordination is quite
rare, that the usual coordination is distorted octahedral, with four short
and two long bonds, and that all degrees of distortion can be recognized
(56).Thus, in compounds containing copper-chlorine bonds, while the short
Cu-C1 distances do not vary much from compound to compound (2.262.32 A) the long distances vary over a much wider range (2.65-3.05 A),
TABLE IX
INTERATOMIC
DISTANCES
I N SOMECu++X, COORDINATION
POLYHEDRA
Compound
Distances (A)
CUO (1%)
CuCli (146)
CuBrs (60)
CSCuCls (246)
CuF2 (11)
(CrF2) (67)
KzCuFr (78)
4 0 at 1.95 (square coplanar)

4 C1 at 2.30, 2 C1 at, 2.95
4 Br at 2.40, 2 Br at 3.18
4 C1 at 2.30, 2 C1 at 2.65
4 F at 1.93, 2 F at 2.27
4 P at 1.98-2.01, 2 F at 2.43
2 F a t 1.95, 4 F at 2.08
as is only to be expected from the general manner in which potential energy
would vary with internuclear distance. In this context, the square coplanar
complexes should be regarded not as representative of a special type of
coordination involving a special kind of covalent bonding but simply as
the limiting case of the usual distortion from octahedral symmetry with the
two distant ligands removed completely. Interatomic distances in some
Cu++X, coordination polyhedra are shown in Table IX.
Cupric fluoride provides an especially interesting case. The cubic
fiuorite structure which was proposed many years ago has now been shown
to be incorrect. The revised structure (11)is rutile-like but the environment
of the Cu++ ions is so strongly deformed from regular octahedral that the
crystal symmctry is reduced from tetragonal to only monoclinic. As we
have already mentioned, the difluorides of the other transition-metal ions,
Mn++, Fe++, Co++, Ni++, and Zn++, occur in the tetragonal structure with
regular or only slightly distorted octahedral environments. The missing
member, CrF2, has now been studied (67) and found t o deviate from the
ordinary rutile structure in just the same way as CuFa. In the chromous
compound, the coordination consists of 2F- a t 1.98, 2F- at 2.01, and 2Fa t 2.43 A.
Copper oxide is the only transition-metal monoxide which does not have
the rocksalt structure and the only one for which large deviations from cubic
21
symmetry are predicted, apart from chromous oxide, the structure of which
is not established with certainty. Although this compound has been
described as crystallizing with the hexagonal wurtzite structure (86), the
very limited X-ray data which have been obtained are not in good agreement with this. They are, indeed, just as compatible with a tenorite-like
structure but are not sufficiently detailed to permit an unequivocal interpretation. It is of interest that in chromous sulfide, the coordination around
Cr++ is of the distorted octahedral kind with 4 S at 2.43 8 and 2 S a t 2.88 8
(69)*
Very few simple binary compounds of Mn3+ are known. The trifluoridc
is a distorted form of the VFa transition-metal trifluoride structure with
2 Mn-F a t 1.79, 2 Mn-F at 1.91, and 2 Mn-F a t 2.09 8 in a distorted
octahedral arrangement (62). The tendency of MnS+ions to distort from a
regular octahedral environment is best seen in the deviations from cubic
symmetry which occur in M++Mnz3+04compounds with the spinel arrangement, but we may note also the distorted octahedral environment which is
found in the mineral manganite, HMn02, where the Mn3+ ion has four
nearest 0-- neighbors in a plane at distances of 1.85 and 1.92 8 and two
more distant oxygens a t 2.30 8 (23).
One might well ask at this stage: granted that distortions from regular
octahedral coordination do occur in compounds of Cu*, Crft, and Mn3+,
could such distortions not equally well be ascribed to covalent bonding
involving dsp2 hybridization on the central ion? To such a question our
answer must be a qualified affirmative, for it should be clear that the alternative descriptions of the distribution of electrons in orbitals on the central
ion and on the ligands are, in general, complementary rather than in conflict with one another, a t least insofar as occupancy of the more stabIe tzll
orbitals is concerned. Both theories agree, moreover, that in a d9 complex
with the usual type of distortion, two of the remaining three d electrons
occupy the d,, orbital. A difference arises, however, concerning the placing
of the last electron. I n the crystal field theory the last electron is placed
in the least stable of the eo orbitals, the d,t+ orbital. In the valence bond
theory this electron has to be promoted to a 4 p orbital, since the remaining
d orbital is required for the formation of the set of dsp2 hybrid orbitals. The
presence of an unpaired 4p electron is, however, incompatible with recent
evidence from paramagnetic resonance (14, IS) and opticaI studies ( I S , 105)
and is, in any case, difficult to reconcile with the marked resistance of Cu++
to further oxidation. Thus the valence bond model, in its traditional form,
while serving as a basis for the discussion of the stereochemical evidence, is
incompatible with many other experimental facts.
Similar deviations from regular octahedral coordination are to be
anticipated for d' ions with the low-spin (t2,)6(eo)1 configuration. A good
22
J . D. DUNITZ AND L. E. ORGEL
example occurs in the low-temperature modification of NaNiO2, where the

Ni3+ion has 4 oxide neighbors at 1.95 A and two more distant at 2.17 8 to
complete the octahedron (38). The low-spin cobaltous complexes have
been divided into two groups, with magnetic moments of 1.7-2.0 B.M. and
2.1-2.9 B.M. respectively. The first group are believed to be octahedral
complexes, the second planar (dd), but detailed X-ray structural evidence is
la eking.
So far we have discussed distortions from octahedral symmetry, in
which all the ligands could, in principle, be equivalent. While this is the
case of interest from the limited point of view of the Jahn-Teller effect,
the general arguments apply equally to other systems, for example, CuC12.2
pyridine or CuC12.2H20. There should, quite generally, be four more
than averagely strong bonds in a plane and two weaker bonds whenever
the d,, orbital is doubly occupied and the &+ orbital singly occupied,
whatever the nature of the ligands. Although we cannot follow the arguments here, it should be mentioned that ligand field theory gives some
indication about which of the ligands are likely to be attached by short
bonds and which by long in a complex of known composition (35).
There are, for example, several cases where substitution of Cuf+ for
another divalent ion results in the formation of a related though significantly different structure in which a less symmetrical environment of the
cupric ion is attained. The structures of CoC12.2H20 (134) and CuC12.2H20
TABLE X
INTERATOMIC DISTANCES
Compoiind
IN
SOME MIXEDCUPRIC COORDINATION
Refs.
Distances
POLYHEDRA
(A)
4 0 (OH) at 1.98, 2 0 at 2.71

4 0 (OH) at 2.01, 2 OH at 2.41
(64, 88) 2H20 at 2.01, 2 C1 at 2.31, 2 C1 at 2.98
(30) 2Hz0 at 2.01, 2 C1 at 2.29, 2 C1 at 2.93
(138) 4 0 at 1.97, H20 at 2.20, Cu at 2.64
(77) 4 0 at 2.00, 2 Ha0 at 2.36
(90) 2 N at 1.99, 2 0 at 2.03, 2H20 at 2.53
(34) 2 N at 2.02, 2 C1 at 2.28, 2 C1 at 3.05
(96)
4 N at 2.05, H20 at 2.59, HzO at 3.37
(43) 4 N at 1.94, 0 at 2.43
(5s) 2 N at 1.95, 4 C1 at 2.76
(63) 2 N at 2.03, 4 Br at 2.88
(148)
ChClp 2H20
KzCuClr 2HzO
CU( ~ c c ~ L L
. H20
~c)~
4Hz0
C~(forrnate)~
Cu proline 2H20
CuClz . 2 pyridine
CU(NHI)&
SO, . HgO
Cn-dimcthylglyoximc
Cu(NHa)z Clz
Cu (NH& Brt
(54, 88) are shown diagrammatically in Fig. 10. The two structures are very
similar and indeed are virtually identical in projection down their c-axes.
Both consist of chains running parallel to e but in these chains the square
coplanar arrangement of the four chlorine atoms about.each cobalt atom is
23
replaced by a rectangular arrangement in the copper structure. In both

cases the water molecules are situated on lines normal to the planes containing the chlorine atoms. A similar relationship holds in the comparison
of the structures of cobalt dipyridine dichloride and copper dipyridine
dichloride (54), which are again virtually identical in one projection but
significantly different when the third coordinate is taken into account. Once
again the environment of the copper atoms is a distorted version of the more
regular environment of the cobalt atoms, with four strong bonds in a
plane and with two weaker ones normal to this plane. In Table X we give
some interatomic distances in these ((mixed coordination compounds of
Cu++. Compounds of Cr++ have not been so extensively studied, but it is
noteworthy that the only other substance known to have the curious
dimeric structure of cupric acetate monohydrate is the corresponding
chromous compound (155).
C. SQUARE
COPLANAR
dx COMPLEXES
We turn now to a slightly different example of a distortion from octahedral symmetry, namely that of ds ions. We have seen that in an octahedral environment the most stable electronic configuration is (tzo)6(e,)z
with the spins of the two electrons in the R , pair of orbitals parallel (triplet
state) and that for this configuration the octahedral environment is stable.
There is another possible configuration in which the two e, electrons go into
a single orbital with their spins antiparallel. This corresponds to the lowest
singlet state which, in a regular octahedral environment, is, of course, less
stable than the triplet because of increased Coulombic repulsion between
the electrons and a less favorable exchange energy. If the regular octahedron
is now distorted so as t o form four short and two long bonds, it can be shown
by calculation (92) that while the energy of the triplet state is unchanged
to first approximation, the singlet state splits into two components, one
more stable, the other less stable, as shown in Fig. 11. If the distortion is
sufficiently large, the lower component of the singlet may become more
stable than the triplet, and hence become the ground state. This is the
explanation of the occurrence of planar complexes of dx ions which are
invariably diamagnetic.
One can see, in a qualitative way, what has happened by reference
to Fig. 9. As increases, there is a tendency for both electrons to go into
the dp4orbital, but this can only be done by pairing the spins and is therefore opposed by the unfavorable exchange energy. Thus small distortions
do not occur. If, however, the distortion becomes large enough for the
orbital stabilization to compensate for the loss of exchange energy the
singlet state becomes more stable. The situation is quite different from that
24
J. D. DIJNITZ ANI) L. E. OKGEL
which occurs in dY complexes where thc ground state is stabilized hy distortions, for then all magnitudes of distortion are possible.
The square coplanar configuration for diamagnetic d8 ions can, of course,
be simply interpreted in terms of the formation of four dsp2 hybridized
orbitals directed towards the corners of a square. The eight d electrons
must accordingly be accommodated in the remaining four d orbitals with
paired spins, thus leading to diamagnetism. It is of interest that this argument was first used by Pauling (113) in 1931 to predict square-coplanar
coordination in diamagnetic Ni++ complexes before it had been verified
experimentally. Since then, conclusive support for the hypothesis has been
forthcoming.
/
I
Singlet
*
P
C
Triplet
Increasing dlstortlon
--+
FIQ. 11. Energies of (t2p)6(ey)2 singlet and triplet states rn a function of distortion
from regular octahcdral symmetry.
In Ni++ compounds, diamagnetic complexes are not formed with

typically (ionic ligands but only with those which give rise to rather
large crystal field splittings (we can refer to such ligands as low-spin or
covalent), for example, cyanide ion in the complex ion [Ni(CN)$ - and the
chelating agents dimethylglyoxime, dithiooxalate, etc. In such compounds
it is quite common to find isomorphism between the corresponding compounds of Ni++, Pd++, and Pt++.
With ions of the second and third transition series, however, spin-paired
complexes are formed with ligands, which according to almost any criteria,
would be included among those forming the most ionic compounds. This
difference arises because the 4d and 5d orbitals extend to greater radial
distances from the metal ion than the 3d orbitals and hence the effect of
the ligand field on the splitting of the orbital energies will be greater, and
the interelectronic repulsions will be dimiriished. Both of these effects
25
must increase the tendency of the electrons to go into the lowest available orbitals, and hence must ellcourage spin pairing. Thus, Pd++ is
paramagnetic only in the fluoride PdF2 with the octahedrally coordinated
rutile structure ( 7 ) , and no paramagnetic Pt++ compounds are known.
Square coplanar coordination has been found in all the other examples
which have been studied; for example, in the oxides PdO and PtO, the
sulfides PdS and PtS, the chloride PdC12,and in a large number of complex
ions [PdCL]--, [PtCl,]--, [Pt(NH&l++, [Pt(CN)41--, [Pt(N02)4]--,
etc. (147). I n one or two cases, for example Pt(NH3)2Br2, there is weak
residual bonding to complete a distorted octahedral environment @ I ) ,
but this is the exception rather than the rule. As far as is known, all AuS+
compounds are diamagnetic, square coplanar complexes. A recent example
is represented in AuCh which has been shown to consist in the solid state
of planar dimeric molecules (31). Vapor-density measurements indicate that
dimeric molecules must also exist in the gas phase of gold(II1) chloride
and bromide.
D. TETRAHEDRAL
COORDINATION
(36)
In the field due to a tetrahedron of negative ions about a positive transition-metal ion, the d orbitals are split into a lower doubly degenerate set of
e orbitals, and an upper triply degenerate set of tz orbitals. If we take the
coordinate axes parallel to the sides of the enclosing cube (Fig. 12), then
FIG.12. Showing the symmetrical relationship of the dzy, d,,, and d,, orbitals which
comprise the triply degenerate t2 orbitals in a tetrahedral complex (2%).
these two sets comprise the dS-2 and d,, orbitals, and the d,,, d,,, and d,,
orbitals, respectively. Thus the splitting is just as for the octahedral case,
but with the orders of the levels inverted. (A precisely similar splitting
occurs with eight ligands at the vertices of a cube.) In the case of a tetrahedral environment, it is the t 2 orbitals which interact most strongly with
the ligaiids, so we shall consider only distortions which depend on the
degeneracy of these orbitals.
26
J . D. DUNITZ .4KD L. 8:. OHGEL
Large distortions from regular tetrahedral symmetry are to be anti&

pated for the high-spin configurations (e)2((tz)*, (e)2(t2)a, (e)4(t2)4, and
(e)4(t2)6. The type of distortiou which should occur is, as we shall see, different for the da and d8 configurations on the one hand, and the d4 and d9
configurations on the other.
To destroy the threefold orbital degeneracy it is quite sufficient to introduce any distortion which removes the cubic symmetry of the complex.
For the present, however, we shall consider only tetragonal distortions
since rhombohedra1 distortions, etc., do not seem to have been discovered
as yet. In discussing these tetragonal distortions we shall retain the usual
crystallographic nomenclature and utilize an orthogonal coordinate system
with a = b # c. If the enclosing cube (Fig. 12) is tetragonally distorted,
we must distinguish between two possible cases. (1) c > a involves an elongation of the tetrahedron along one of its fourfold alternating axes; (2)
c < a involves a corresponding flattening of the tetrahedron. In both
cases, the triply degenerate t z orbital splits into a singlet and a doublet but
which of these lies lower depends on the ratio c/a. In the limiting case of
(2), c/a = 0 the tetrahedron becomes a square in the zy-plane. The d,
orbital concentrated in this plane is therefore destabilized: the d,, and dun
orbitals constitute the degenerate doublet which is stabilized. This type of
splitting is shown in Fig. 13 from which it is clear that the orbital degen-
FIG. 13. Splitting of d orbital energy levels in a regular tetrahedral field and in tctragonally distorted fields with c/a < 1 (left,)and c/a > 1 (right). These distortions may be
described as flattening cr elongation, respectively, of the tetrahedron along an IS axis
(36).
eracy is removed for the d 4 and d9 configurations but not for d3 and d8.
The opposite distortion, elongation of the tetrahedron ( c / a > 1) produces
just the reverse splitting, that is, the singlet d, orbital is stabilized and
the degenerate pair of orbitals, d,, and dualare destabilized. This distortion
leaves the d 4 and d9 systems degenerate but removes the degeneracy of
d3 and ds. Thus we expect d4 and d9 systems to flatten, and d8 and d8 systems
to elongate.
Such distortions can be understood very easily by considering the distortions produced by one extra electron (8and d8) or one hole (d4and d8)
in a half-filled or filled set of t l orbitals. The two situations are shown
S T E R E O C H E M I S T R Y OF I O N I C S O L I D S
27
diagrammatically in Fig. 14. The extra tz electron repels the ligands and so
causes elongation of the tetrahedron, while the positive hole attracts the
ligands and so causes flattening.
It cannot yet be claimed that the experimental evidence for such distortions from tetrahedral coordination is so well established as for the
octahedral case, but this is largely attributable to the comparative rarity
of structures in which transition-metal ions occur in a tetrahedral environment. There appear t o be no examples of tetrahedral coordination about
d3 and d4 ions, and only fragmentary information is available for d8 and
ds ions. It may be recalled that tetrahedral coordination was anticipated by
Pauling (114) for paramagnetic 4-covalent complexes of nickel(I1) on the
grounds that, since all the d orbitals are occupied by electrons from the
metal, the covalent bonding must occur through sp3 hybridized orbitals.
It is now known that many formally 4-covalent complexes are, in fact,
@
----- -
d 3 , d8
tzfl
----- -
d,
d9
(4
(b)
FIG.14. Effect of (a) a single electron in the tz orbital (Tz state), (b) a single hole in
state) on the stereochemistry of a tetrahedral complex. Only tetragonal
the t~ orbital (TI
distortions aro considcrcd (36).
octahedrally coordinated, with two or more of the ligands shared between

pairs of metal atoms, and it is only recently that definite evidence for tetrahedral coordination in such compounds has been obtained. The paramagnetic compound (Ph3P)zNiClzhas been shown by X-ray analysis to be tetrahedral (139),in contrast to the diamagnetic (Et3P)zNiBrzwhich is coplanar
(122). But since in this compound the ligands are not all equivalent, it is
not really relevant to a discussion of distortions from regular tetrahedral
symmetry. The CuC14- - ion however in the compound CszCuC14is known
to be flattened (61) in just the way that the theory would predict while
the CoC14-- ion in the structurally similar CszCoC14 (115) is much more
nearly regular. Some other examples of distortions from tetrahedral symmetry occur in NiCrz04and CuCrz04,and we shall mention these in our
discussion of the spinel structure.
Although it is formally possible to have low-spin complexes of d3, d4,
d6, and d6 ions, none are known. The reason for this is that the splitting
A between the e and t z levels is much smaller in a tetrahedral than in an
octahedral complex. The purely electrostatic theory predicts a ratio of
4/9 between the two cases under equivalent conditions, and the limited
28
spectroscopic evidence for tetrahedrally coordinated complexes in solution

and in glasses suggests that this figure is fairly close to the mark. Such a
low A would be too small to cause spin pairing for 3d electrons, but for 4d
and a fortiori for 5d electrons, it seems possible that spin pairing might
occur even in tetrahedrally coordinated compounds. This might explain
the diamagnetism of the ReCI4- (see 45) ion since the Re3+ions could have
the configuration (e)4(tz)0.
E. OCTAHEDRAL
VERSUS TETRAHEDRAL
COORDINATION
(36,37)
In addition to the factors which we outlined on p. 2 for assessing the
relative stabilities of the various coordination polyhedra, the number of d
electrons present in the central ion must also play an important role. Indeed, since transition-metal ions of a given charge have approximately the
TABLE XI
CRYSTAL
FIELDSTABILIZATIONS
IN TERMS
OF A/lO FOR TRANSITION-METAL
IONS
I N OCTAHEDRAL
A N D TETRAHEDRAL
ENVIRONMENTS
Ion
NOof delectrons
Ti'+ V4+
Ti++,VJt, Cr4C
V++, CrS+, Mn4+
Cr++,MnJ+, Fe4+
Mn++,Fe3+, C04+
E'c++,C O ~ +Ni4+
,
Co++,Ni3+
Ni++, CiiJ+
CU++
c u t , Zn++
0
Octahedral
stabilization
Tetrahedral
stabilization
2
3
4
5
6
7
8
12
6
12
9
10
ti0
8a
40
0
12
12
6 O
8"
4.
These ions have substantial additional Jahn-Teller stabilizations.
same ionic radius and differ principally only in the number of d electrons
present, we would anticipate that this additional factor would be the
determining one in deciding their relative preferences for octahedral and
tetrahedral coordination.
We have shown that for an ion in octahedral coordination, the energies
of the tzo and e, orbitals referred to the mean d orbital energy are -+Aa
and +gAal respectively. This means that for the electronic configuration
(tze)"(e,)" the crystal field stabilization energy is Aa(44m - 6n)/lO. Similarly, in tetrahedral coordination, the corresponding stabilization for the
configuration ( e ) p ( t . J P is A'(6p - 4q)/10, where we remember that At is
only about 40% of A" under equivalent conditions. These stabilizations are
29
shown in Table XI for different transition metals. It is clear from the results
that Cr3+ and Ni++ should show the strongest preference for octahedral
coordination and that Mn++ and Fe3+ will have no particular preference
insofar as crystal field effects are concerned. We have already mentioned the
comparative rareness of high-spin tetrahedral complexes of Ni++, which
contrasts with the fact that several tetrahedral complexes of Co++ are
known; for example, the CoC14-- ion in CseCoC14(115) and Cs3CoCls (116)
and the blue complexes of CoClz with organic bases (15, 101) (the violet
complexes have Co++ in octahedral coordination). From Table X I we
should expect Co++ to show a greater preference for tetrahedral coordination than Ni++.
In order to make a more critical test of the theory, we must look for a
structure in which the transition-metal ion has the possibilities of both
octahedral and tetrahedral coordination, and preferably one in which a
given transition-metal ion may be exchanged for another without substantial alteration in the atomic arrangement. Such a structure is the spinel
arrangement in which a large number of XY 204
compounds crystallize,
where X is a divalent metal ion and Y is AP+, Cr3+, Mn3+, Fe3+, or Go3+.
This structure is also of importance in the discussion of distortions from
cubic symmetry and we shall now consider it in some detail.
F. THESTRUCTURE
OF COMPOUNDS
WITH
THE
SPINELARRANGEMENT
The mineral spinel MgA1204 is the prototype of this structure. The

basic framework consists of a cubic close packed arrangement of oxide ions.
Such a framework contains both octahedral and tetrahedral interstices.
Starting off from any oxygen, we find an octahedral site by moving in a
direction parallel to any cube edge to a point halfway to another oxygen.
Similarly, a tetrahedral site is found by moving in a direction parallel to
any cube diagonal. If all the octahedral sites are occupied by metal ions,
we have the rocksalt structure (Fig. 4).If all the tetrahedral sites are occupied we have the fluorite structure (Fig. 6a), and if only every alternate
tetrahedral site is occupied we have the sphalerite structure (Fig. 5a).
There is a special way of occupying half of the octahedral sites and oneeighth of the tetrahedral sites which gives the spinel structure (Fig. 15).
In a normal spinel XY204, the trivalent Y ions occupy the octahedral
sites, and the divalent X ions the tetrahedral sites. There is also the possibility, however, that half of the Y ions occupy the tetrahedral sites, the
remaining half together with the X ions being distributed over the octahedral sites-the so-called inverted arrangement. Intermediate arrangements are also possible. If the atomic proportion of the divalent ion occurring in the octahedral sites is expressed as 6, then 6 = 0 for a normal spinel
30
J. D. DUNITZ AND 1,. l3. ORGEL
Octahedral ion o t 1/8
Tetrahedral ion
Oxygen at 3/8
FIQ.15. Arrangement of atoms in the spinel structure. The lower half of the cubic
unit cell is shown projected down a cube axis, the heights of the a t o m being indicated
either by numbers (0, Ji,
or $4) or by vertical (36) and horizontal (36) shading. One
octahedron and one tetrahedron are outlined for clarity.
and 6 = 1 for an inverse spinel. For completely random distribution of the

two kinds of ion between the two kinds of site, 6 = 0.67.
As we have mentioned, it is possible to assess the relative tendencies of
transition-metal ions to adopt octahedral or tetrahedral coordination from
the respective crystal field stabilizations. The results of Table XI could
TABLE XI1
CRYSTALFIELD
STABILIZATIONS
(KCAL/MOLE)
ESTIMATED
FOR TRANSITION-METAL
OXIDES
Ion
Octahedral
stabilization
Tetrahedral
Stabilization
Excess octahedral
stltbilizlttion
0
7.9
14.8
8. G
6.4
14.0
25.5
16.0
9.6
0
0
4.0
7.4
20.6
15.2
6.9
12.8
37.7
22.8
0
Mn++
Fe++
Cot"
Ni++
Cu++
TiS+
V3+
Cra+
Mns+
Ft!'+
0
11.9
22.2
29.2
21.6
20.9
38.3
53.7
32.4
0
31
STEIiEOCHEMISTRY OF IONIC SOLIDS
serve as a rough basis for such a theory but it is better to express the crystal
field stabilizations in a semiquantitative form by estimating A for the individual transition-metal ions in an environment of oxide ions. This can be
done from a consideration of the absorption spectra of transition-metal
oxides and hydrates (37) and the results are given in Table XII. The experimental results on cation distribution in spinels derived from X-ray and
neutron diffraction measurements are given in Table XI11 and are seen
to be in excellent agreement with expectations.
TABLE XlII
EXPERIMENTAL
CATIONDISTRIBUTIONS
AMONG TETRAHEDRAL
AND
OCTAHEDRAL
SITESIN SPINELS
(FOR DETAILED REFERENCES SEE 37).
normal spinel, M++ in tetrahedral sites
8 = 0,
8 = 1,
inverse spinel, M++ in octahedral sites
d = 0.67, random distribution of M+ and M3+over tetrahedral and octahedral sites
Mg++
0
Aluminates
Chromites
Ferrites
Manganites
Cobaltites
Mn++
0.9
0
0
0
0 2
0
.-
Fe++
Co++
0
0
1
?
-
0
0
1
y
0
Ni++
Cu++
Zn++
0.1
0
0
0
0
0
0.76
0
Thus the ions with the greatest excess octahedral stabilization are
Cr3+ and Mn3+,and in agreement with this we see that all the chromites
and manganites (except NiMn20s) are normal spinels. For Fe3+, on the
other hand, there should be no preference, as far as crystal field effects are
concerned, for either type of coordination, so that this ion should be displaced from the octahedral sites fairly easily. In agreement with this, all
the ferrites except for ZnFe2Ol are a t least partially inverted. Of the
divalent ions, the one with the greatest excess octahedral stabilization is
Ni++, and we see that this is the only divalent ion capable of displacing
APf or Mn3+ from the octahedral sites. The compounds CoMnzOd and
FeMnzOl have been described as having a random distribution of cations
over the available sites (195) but since electron-transfer reactions (for
example, Co++ Mn3+ = Co3+ Mn++) may occur here it seems doubtful if they are really to be classed as manganites. In contrast to Mn3+ions,
Mn++ ions should be easily displaced from the octahedral sites.
The ideal spinel structure has cubic symmetry but there are a number
of cases where one cube axis becomes longer or shorter than the other two
to give a structure with only tetragonal symmetry. Two of these cases
32
J. D. DUNITZ AND L. E. OIZGEL
have been studied in detail. I n CuFezOr(c/a = l.OS), an inverse spinel with

Cu++ exclusively in the octahedral sites, the tetragonal distortion has been
shown (117) to be associated with a distortion of these sites from regular
octahedral symmetry, in such a way as to surround the metal ion by four
nearest neighbors in the ab-plane and two more distant neighbors along the
c-axis. The tetrahedral sites, in which only Fe3+ ions occur, retain their
regular tetrahedral symmetry. CuCra04 is a normal spinel with c/a = 0.91
and here the distortion is associated with a flattening of the oxygen tetrahedron about each Cu++ ion so as to make it approach to a square in
the ab-plane (118). We have seen other examples of both of these types
of distortion in our account of the stereochemistry of Cu++ and have
attributed their occurrence to the Jahn-Teller effect.
In the other cases where tetragonal distortions have been observed for
spinels, detailcd structural evidence is not available. In the absence of such
evidence, it is reasonable to suppose that they arise in a similar fashion
since, in almost all cases, they occur in compounds for which a large deviation from cubic symmetry would be anticipated as a result of the JahnTeller effect. NiCrz04, a normal spinel, becomes tetragonal below 310K.
A t room temperature, 15' below the transition point, the axial ratio c / a
is 1.025 ( 8 5 ) and it presumably increases on further lowering of the temperature. We believe that this distortion results from the occurrence of Ni++
in tetrahedral coordination and that it consists of an elongation of the oxygen tetrahedra in the direction of a cube axis. We recall that opposite distortions should occur for tetrahedrally coordinated ions with d8(Ni++)and
d9(Cu++)and hence the change in c/a from greater to less than unity on
passing from NiCrZO4to CuCrz04was to be expected.
For the normal spinels CdMn204( l 2 5 ) ,Mn,O4([Mn++][Mn2"+]O4)
(120)
MgMnz04and ZnMnn04(120, 126), all with c/a ratios greater than unity,
the deviations from cubic symmetry are probably associated with the tendency of the Mn3+ ion to distort from regular octahedral coordination so
as to have four nearest neighbors in a plane and two more distant neighbors
in the direction normal to the plane. The random spinels CoMnz04and
FeMrizO4 presumably have sufficient Mn*+ ions in octahedral sites to give
the same effect. It is more difficult to account for the c/a values of unity
reported for normal CuMntOd and inverse Mn[NiMn]Od (125). In both of
these cases, the tetrahedral sites should flatten towards a square (c/a < 1)
while the octahedral sites should distort t o give four long and two short
bonds (c/a > 1). There is a way of distorting the spinel structure to satisfy
both of these requirements simultaneously. It consists of displacing every
0- ion from its ideal position away from its nearest tetrahedrally coordinated ion in a direction parallel to a diagonal of a given cube face. Such an
arrangement has c/a = 1, but it is nevertheless only of tetragonal sym-
STEREOCHEMISTHY OF IONIC SOLIDS
33
metry. It should be possible to detect such a deviation from cubic symmetry

from single crystal studies but not, however, from powder samples, and a
definite conclusion on the matter must await such evidence.
G. CONCLUDING
REMARKS
ON JAHN-TELLER
DISTORTIONS
This is a convenient point to recall a number of limitations and also a
number of further applications of the foregoing discussion :
1. We have neglected spin-orbit coupling, which may well prove to be
important particularly in tetrahedrally coordinated compounds.
2. We have been able to predict only that distortions must exist in
certain cases and that they maintain the center of symmetry if the metal
ion is octahedrally coordinated. We have not been able to specify the details
of the distortion which minimizes the energy.
3. We have not considered the role played by cooperative forces in the
crystal in trapping the distorted configurations. In an isolated complex
the effects which we have discussed lead to the existence of a set of equivalent, distorted configurations. In a d9 complex, for example, a tetragonal
distortion could occur about any one of a set of three mutually perpendicular axes. The internal motion of the complex at room temperature would
then lead to frequent exchanges between these distorted configurations
unless the activation energy for the process were unexpectedly high. In
any case, in an assembly of noninteracting molecules, each configuration
would be represented an equal number of times. In a crystalline compound
containing distorting ions regularly arranged on a lattice it is the interaction between the distortions which leads, a t sufficiently low temperatures,
to the trapping of each distortion in fixed orientation to all the others. Only
if this happens can X-ray methods reveal Jahn-Teller distortions, for example, in spinels.
4. We have restricted our discussion to ions in their ground states as
this is all that is necessary in stereochemistry. In treating the optical properties of solids, particularly their fluorescence and phosphorescence, it may
be anticipated that Jahn-Teller distortions of excited states will be of great
importance.
I n cubic crystals containing a small amount of a foreign ion such as
Cu++ or Ni+ at low temperatures it is found (58) that each metal ion has
a distorted environment, but that the tetragonal axis of this environment
may have a number of equivalent orientations relative to the crystal as a
whole. At higher temperatures the direction of the tetragonal axes changes
so very rapidly that each ion experiences on the average a cubic crystal
field. Thus impurity ions in a cubic crystal exhibit just the properties
expected for isolated complexes, as indeed they should.
34
J. D. DUSITZ AND L. E. ORGEL
V. Stereochemistry of B-Subgroup Ions
A. IONS
WITH dl0 CLOSEDSHELLCONFIGURATIONS
(73,107)
In this section we discuss the stereochemistry of B-subgroup ions with
filled d shells. The most striking feature is that in addition t o forming tetrahedrally and octahedrally coordinated compounds, these ions exhibit a
more or less pronounced tendency t o occur in linear environments, for
example, Ag+ in the linear cation [Ag(NH,)s]+, Au+ in the linearly coordinated infinite chains of AuI, and Hg++ in the discrete molecules of HgClz
which exist in both the gas and crystalline phases. We believe that two influences are a t work in determining these structures. One is related t o the
Jahn-Teller effect, and the other to covalent bonding. We shall discuss them
in this order and then consider their relative importance.
We have already shown in a previous section that if a transition-metal
ion has an odd number of electrons in the e, orbitals, then it must necessarily distort a regular octahedral environment in which it is placed. The
orbital degeneracy which would occur in the highly symmetrical environment is broken down by a lowering of the symmetry, thus conferring an
additional degree of stability to the molecule or complex. We now wish t o
illustrate an example of a more general but less powerful principle, namely
that the existence of low lying excited states often leads to a tendency to
distort from a highly symmetric environment. If the excited states are
sufficiently close in energy t o the ground state, then a distortion must occur
in a number of important situations.
Consider an ion with the configuration d'O. The ground state is nondegenerate, ' S , but there are also fairly low lying excited states 3D and 'D
arising from the configuration dgs.2In a cubic environment these states
cannot be mixed with the ground state, since neither the tzoorbitals nor the
e, orbitals have the same transformation properties as the s orbital. I n a
tetragonally distorted environment, however, this is no longer true, and the
most favorable electronic configuration is then described by d10-bs6. In
other words, some of the ID excited state would be mixed with the ground
state. This is most easily seen by considering the separate orbitals. The
wave functions for the &, s, and (l/fi)(& f s) orbitals are illustrated in
Fig. 16.
It is clear that the (l/Z/2)(& - s) orbital is concentrated-in-'the
s) orbital
zy-plane. If this orbital were occupied, and the (l/fi)(ds2
left empty, then the ligands in the zy-plane would be repelled more than
those along the z-axis. A regular octahedron would therefore distort by
We neglect spin-orbit coupling here, although it would have to be taken into account
in quantitative calculations, especially if very heavy ions are concerned.
35
lengthening the bonds in the xy-plane and shortening those in the z-direction. Conversely, a distortion of this nature would stabilize the (1/<2)
(dz2- s) orbital with respect to both the ( l / d % ) ( d z 2 s) and the pure dzz
orbitals. The opposite distortion, that is, one leading to four short bonds and
s) orbital more
two long ones, would of course make the (1/<2)(dz2
stable than either the ( l / d 2 ) ( d z 2- s) or the dzzorbital.
Let us first suppose that in a regular octahedral environment the e,
pair of d orbitals and the s orbital are degenerate, and consider what would
happen to the lowest singlet state.3 Since the d(e,) and s orbitals are degen-
(a)
FIG.16. (a) The d,z and s orbitals, (b) hybrid orbitals concentrated along the z-axis
and in the sy-plane (107, Fig. 2).
erate, it requires no intra-atomic reorganization energy to form an arbitrary

mixed orbital from them. If the pure dz2orbital, (together with the other d
orbitals), is used there is no tendency to distort, but if the occupied orbital
is a linear combination of s and d then the arguments used in the case sf the
d9 configuration show that distortion must occur with a resulting stabilization of the system. (A theorem related to the Jahn-Teller theorem would
apply quite rigorously here, provided that a vibration of appropriate symmetry to mix the two accidentally degenerate orbitals exists.) Thus, either
the ( l / d ) ( d z 2 s) orbital would be occupied and we should get four
short bonds, or the ( l / f l ) ( d z 2 - s) orbital would be occupied and we
should get two short bonds. There is also the possibility that some other
combination of the dz2,dz2+, and s orbitals would be occupied. I n this case
we should get a distortion t o a lower than tetragonal symmetry but always
so as to maintain the center of symmetry.
Now suppose that the eo orbitals are more stable than the s orbital by
an amount E,. Mixing of the dz2 and s orbitals to give an orbital a& (1 - a2)%swould still cause a distortion and resultant stabilization of the
octahedron, but this would have to be offset by the promotion energy
(1 - a2)E,required to achieve the mixing. If E, is sufficiently small then
3 We neglect here the fact that in such a system the ground state would in fact be a
triplet. This is no longer so, however, if the d-s separation is sufficiently large, as it is in
all practical cases.
38
distortion must occur, for we have seen that this is so if E , = 0; that is,
if the d,, and s orbitals are degenerate. Similarly, if E , is sufficiently large
it is clear that little advantage is obtained by mixing the d,, and s orbitals.
Thus it becomes a quantitative question to decide whether any given
compound will be stable against distortion or not. Detailed calculations
have not been made, but it is an experimental fact that it is just those dt0
ions which have small d-s separations, Au+, Hg*, Ag+, and Cu+, which
have a strong tendency to give linear complexes (see Table XIV).
TABLE XIV
ENERGIES
(EV) OF LOWEST
dss
AND
d9p
STATESABOVE
THE
d'0
GROUND
STATE(723,
(98)
dQs
@P
Ca+
Zn++
Ag+
Cd++
Au+
Hg++
TI$+
2.7
8.2
9.7
17.1
4.8
10.0
9.9
17.3
1.9
7.8
5.3
14.7
9.3
18.3
The essential difference between the Jahn-TelIer effect and the near
degeneracy effect which we have just discussed arises because the orbital
stabilization is linear in the extent of distortion for the first case (for
sufficiently small distortions, at any rate) but quadratic in the second. The
restoring forces are always quadratic in the extent of distortion. Thus the
Jahn-Teller effect produces a distortion no matter how large the normal
restoring forces, while the near degeneracy effect only produces distortion
if the quadratic dependence of the stabilization energy on distortion is
greater than that of the restoring forces (see Fig. 17).
We turn now to the argument from covalent bonding considerations.4
(a)
(b)
(c)
FIG.17. Potential energy curvea for (a) Jahn-Teller distortion, (b) near-degeneracy,
small distorting force, and (c) near-degeneracy, large distorting force. The restoring
potentid is shown as a thin h e , the distorting potentid &8%. thick h e , the resulting
potential as a dashed line.
This argument seems to have no precise equivalent in an ionic theory so we consider it here.
37
It has been argued (73) that the boiids formed by Hg++ are more covalent
than those formed by Cd++and Zn++ because the s and p orbitals are more
stable in the first case. This argument is not good enough as it stands,
because an increase in the stability of these orbitals on the metal ion should
lead to a uniform increase in bond strength, but not to a reduction in the
coordination number. We believe, theref ore, that the s-p separation, as
well as the absolute energies of these orbitals, is important in producing
a lowering in the coordination number. If the s-p separation were very
large we should, from the point of view of a covalent bonding theory,
e cpect discrete MX compounds, insofar as this is allowed by the composition of the substance; for example, in monovalent halides. If the s-p separation were small, we should expect high coordination numbers using sp3 or
sp3d2hybridized orbitals. For intermediate separations, we should expect
linear coordination using s p orbitals or possibly trigonal coordination using
sp2 hybridized orbitals. Thus the greater tendency of Hg++, compared with
Cd++ and Zn++, to form linear bonds may be partially attributed to the
larger s-p separation (by about 2 ev) in Hg++. Similar arguments might
apply to the other dL0ions which form linear complexes.
It seems likely that both the covalent and ionic mechanisms play some
part in determining the unusual stereochemistry of these dlo ions. The
electrostatic arguments do not show whether two or four short bonds are
to be anticipated, but analogy with the case of the Cut+ion would suggest
the latter. In practice, the other configuration is invariably observed and
we are inclined to attribute this to the importance of covalent factors
neglected in the ionic theory. On the other hand, there is a drastic change
in the d-s separation on going from Zn++ and Cd++ to Hg++, which seems
more in keeping with the drastic change in typical coordination number
than does the comparatively small change in the s-p separation. It is not a t
present possible to decide which effect, if either, is the dominant one.
One must remember too that even if the tendency to distort may be
largely electrostatic in origin, its consequences may include ones that can
be understood only in terms of covalent bonding. Once distortion has
occurred, for whatever reason, the shorter bonds will certainly be stronger
and, in the conventional sense, more covalent. There is, for example, no
question that the bonds in the HgCI2are more covalent than those in CdCL:
the point at issue is rather to establish the reasons why Hg++prefers to form
these two more covalent short bonds, and Cd++ to form six equivalent,
longer, more ionic bonds.
Mixing of the d and s orbitals in tetrahedrally coordinated dl0 ions (108)
can occur only if the symmetry is lowered by an off-center movement of
the ion (T2vibration). In this respect it is interesting that certain unusual
38
J. D. bUh'ITE A N D I,. E. ORQEL
features of the X-ray scattering from Cu+ and Ag+ halides can be understood only on the hypothesis that the metal ions have extremely large
amplitudes of vibration or, in some cases, if there exist subsidiary potential
energy minima displaced from the centers of the coordination tetrahedra
(66)'
We now proceed to discuss the stereochemistry of dl0 ions in the light
of the preceding discussion, beginning with the monovalent ions. The
linear coordination about the metal atom in AuI, a structure composed
of infinite chains (144)
Au
/\
/ \
Au
Au
'I/
Au
I'
is in contrast to the tetrahedral coordination in CuI and AgI, and establishes

the position of Au+ as the ion with the greatest tendency to the formstion of linear bonds. At the other end of the scale, the position of Cu+ is
established from a comparison of the structures of the complex cyanides
KM'(CN)2. The silver and gold compounds contain discrete linear complexes N-C-M-C-N
(63), but KCu(CN)z has an entirely different
structure in which the complex anion consists of infinite chains of alternating copper atoms and cyanide groups with the extra cyanide group attached
through carbon to euch copper atom (32). The angle N-Cu-C within a
chain is close to tetrahedral and the Cu atom lies about 0.27 from the
plane of its three nearest neighbors, thus achieving a trigorla1 pyramidal-like
~oordination.~"
In this respect it is interesting that Cu+ readily forms higher
complex cyanides [Cu(CN)& - and [Cu(CN)$- which are presumably
both tetrahedrally coordinated, while Ag+ forms [Ag(CN)3]-- only with
difficulty, and Au+ forms only the linear double cyanide. Compounds of
hg+ are known with octahedral (AgF, AgCI, AgBr), tetrahedral (AgI),
and linear coordination (AgCN), the coordination number generally
decreasing with increase in the electron-donor properties of the attached
ligands although with many exceptions. Thus AgzO, by the usual criteria
an ionic substance, has a linear arrangement of oxygen about the silver
ions (149).
Among the cuprous compounds which have been studied by X-ray
methods, linear coordination is quite rare but it does occur in cuprite, CuzO
(loo),which is isostructural with argentous oxide, and in cuprous chromite
(33) and ferrite (127). In these three linearly coordinated oxide structures,
4a The coordination about Cu+ in KCu(CN)* may be described a8 trigonal (39) with
2 C a t 1.92 and 1 N a t 2.05 4. There is, however, an extra C, belonging to a different
chain, at 2.68 A near the apex of the tetrahedron defined by the three nearest neighbors.
STEREOCHEMISTRY OF IONIC SOLIIX3
39
the Cu-0 distances are appreciably shorter than the sum of the ionic radii.
The only other example of linear coordination in a cuprous compound
occurs in the cuprous diazoaminobenzene derivative. This molecule has
been shown (22) to be dimerized in the solid state, with formation of the
eight-membered ring shown in Fig. 18.
FIG.18. Drawing of the diazoaminobenzene: copper complex.
Of the divalent dl0 ions, Zn++ and Cd++ are invariably found in highly
symmetrical environments with coordination number 4 or 6, and it is only
Hg++ which shows a tendency towards linear coordination. The most
ionic compound, HgF2, has the 8-coordinated fluorite structure (Table VI),
compatihle with the rather large ionic radius (1.10 b) which has been
assigned to the Hg++ ion. In crystals of HgCl2 (18, 47), HgBrz (140), and
the yellow modification of HgI2 (467, the Hg++ ions are surrounded by six
halide groups in a much distorted octahedral arrangement with two
coaxial bonds much shorter than the other four. This type of coordination
seems to be quite common among mercuric compounds: it occurs also in
the complex halides NH4HgC13 (55) and K2HgCld.H20 (89) and in HgCL.2
pyridine (48) (the last is almost isostructural with C0CI2.2 pyridine and
CuC12.2 pyridine (34) but the distortion of the octahedral environment is
even more extreme). In all of these compounds the distortion from octahedral coordination has gone so far that discrete HgX2 molecules, similar
to those which occur in the gas phase (50),may be said to exist in the crystals. Such discrete molecules also occur in the complex of HgClz with 1:4
dioxane (57) but here there are only two (rather than four) more distant
oxygen neighbors. In the red, low-temperature form of the iodide (10) the
Hg++ ions are, however, in tetrahedral coordination and no discrete molecules can be recognized.
The recently determined structure of mercuric oxide has two short
collinear Hg-0 bonds and two much weaker bonds at right angles (6).
The sulfide is dimorphic. One form has the tetrahedrally coordinated
sphalerite structure (Table V) : the other, cinnabar (6),contains infinite
chains like those of the oxide. Here, however, the packing of the chains is
different and is such as to complete a very distorted octahedral environment
40
J. D. DIJNITZ AND L. E. ORGEL
about the metal. An unusual structure has been reported for the compound
2HgC12.Hg0 (I.%), containing trichlormercury oxonium cations with Cs
symmetry: the O-Hg-Cl angle is 175, close to linear.
It is seen that dl0 ions exhibit a considerable variety in their coordination potentialities, from octahedral and tetrahedral with little or no distortion from regularity, to highly distorted linear coordination. The ions
with large d-s separations seem to be immune from distortion, but the
tendency to distortion does not always run parallel with the d-s separations
of the different ions. If this were the sole determining factor we should
expcct Ag+ to show a smaller tendency to linear coordination than Cu+,
whereas the contrary is sometimes the case. The importance of covalent
bonding is indicated by the increased tendency to linear coordination with
increase in the electrondonor properties of the ligands, for a given metal
ion. Here again, however, the correlation is only very rough, and there are
many exceptions. Finally, the frequent occurrence of polymorphic forms
with different coordinations would suggest that the balance between the
various factors that are involved may be very sensitive to a number of
environmental features which we have neglected in ou<discussion.
B. THE INERTPAIR
It is well known that the presence of pairs of unshared electrons in the

valency orbitals of an atom affects the stereochemistry of the molecule in
which it occurs. This is equally evident from the nonplanarity of molecules
such as NH, and PHI and from the irregular coordination often found about
Sn+f or Pb* ions. As in the case of many stereochemical facts, interpretations of the influence of the unshared pair can be given either from the
electrostatic or from the covalent point of view. Here we restrict our discussion to the most ionic compounds of the metals and hence use the
electrostatic approach which is then more appropriate. We leave for other
publications detailed discussions of the quantitative importance of covalency.
Let us first consider the structure of tetragonal (red) lead monoxide
from the conventional viewpoint. The Pb++ion has the environment iilustrated in Fig. 19 with four nearest neighbors (2.30 A) forming a square
FIG.19. Environment of Pb++ion in red PbO.
STEitE:OCHlCMISTItY OF IONIC SOLIUS
41
on one side of the metal ion and four more distant neighbors (4.3 A) completing a much distorted cube (24, 99). It is usually supposed that the
inert pair occupies an s p hybridized orbital pointing away from the four
near neighbors and that it is electrostatic repulsion between the unsymmetrically distributed lone pair electrons and the ligands which accounts
for the characteristic deviation from cubic symmetry.
Almost all theories agree that the deviation from cubic symmetry of
the environment is associated with a nonspherical charge distribution of
the unshared pair. We take this for granted and enquire in more detail
about the relation of the electronic properties of the metal ion and the
nature of the anion to the extent of distortion (110). Our arguments follow
closely those already outlined for the problem of the distortions associated
with dlo ions.
Ions with an outer, doubly occupied s orbital, such as TI+ and Pb++,
differ from all other filled shell ions in that they have a low lying excited
configuration . . . (ns)(np), obtained by promoting an electron from the
occupied s to the unoccupied p orbital (6s-6p) in the case of TI+ and Pb*.
The important P state of this excited configuration is 7.4 ev above the
ground state in the case of TI+,for example. Since the s orbital is spherically
symmetrical while the p orbital changes sign on reflection in the origin,
only antisymmetric vibrations can mix them together. When the octahedron distorts in one of the ways shown in Fig. 20 the stabilization accompanying sp mixing tends to counteract the normal forces which hold the
octahedron in its regular configuration. An exactly analogous argument
holds for distortions from regular cubic coordination. If the s-p separation
is sufficiently small then the regular environment must be unstable. However, it may require low temperatures to reveal the nature of the distortions,
for a t high temperatures the distortions a t different sites may not be ordered
in which case the crystal, as a whole, would behave as if it were cubic.6
It should be noted that these distortions are sharply distinguished from
those predicted for dlo ions in that they lead t o environments without a
6
Its your move said the Queen.

Oh dear answered Alice. But I dont know whore to move. One move seems just
as good as another and 1 really cannot make up my mind which is best.
When I say its your move, youve got to move whether you like it or not. Alice
looked a t the board and tried to concentrate on the game but no matter how hard
she tried all the moves seemed exactly equivalent. Take her away and give her
a cold compress cried the Queen angrily--she wont play properly and shes
spoiIing the game for me.
She turned to Alice and explained, in a kinder tone, Its for your own good, my
dear. Youre looking a little feverish but once youve cooled down youll be much
more cooperative and make just the same move as everyone &P. Then 111 be the
winner! (See 1%).
42
J.
I).
D U N I T Z AND L. E. ORGEL
center o j syrrumetry. It is the nature of the low lying excited states which
dictates the nature of the characteristic distortions, and these are quite
different for dlo and d10s2 ions.
In our discussion of the stereochemistry of (inert pair ions we shall
restrict ourselves to compounds which have structures that may reasonably
be regarded as distortions of simple ionic ones. It must be emphasized that
a large number of otherwise relevant compounds do not fall into this category although the environment of the metal nearly always lacks a center
FIG.20. Distortions from octahedral coordination preserving (a) a fourfold axix,

(b) a twofold axis, (c) a thrcefold axis.
of symmetry. For example, the high-temperature modifications of the oxides
Asz03and Sbz03,like the vapor phases of these compounds, contain discrete M,Os molecules (1) with the characteristic cage structure of PdOS
(52, 94). On the other hand, as we have seen, no discrete PbO molecules
can be recognized in the structure of red PbO, which can be regarded as a
distortion of the typically ionic CsCl structure. Exactly the same structure
is found for the (inert pair oxide, SnO (99).
FIG.21. Environment of the metal ion in GeS and SnS. The interatomic distances are,
for GeS and SnS respcctively (a) 2.47, 2.68, (b) two: 2.64, 2.62, (c) two: 2.91, 3.27, and
(d) 3.00, 3.39 b.
In the same way, the structures of GeS (151) and SnS (65) can be regarded as distortions of the rocksalt structure. The deviations from octahedral coordination are not quite the same but are very similar in that in
both compounds the metal ion is much closer to three of its neighbors than
to the other three (Fig. 21). Another kind of deviation from octahedral
coordination is found in TIF which also possesses a distorted rocksalt structure. The early structure determination in which the TI+ ion was at the
43
center of a distorted octahedron has now beeii showii to be incorrect. In

the revised structure (56)each metal ion has five nearest neighbors and one
more distant. In T11 (59), InBr (199), and In1 (71) this kind of distortion
has become more extreme: there are only five neighbors, one corner of the
octahedron being unoccupied (Fig. 22). TlCl and TlBr, on the other hand,
(&.
"...-3.87..--.
-=-
Fro. 22. Environment of the TI+ ion in TII. Regarded as eithcr 5- or 7-coordination
the distortion from an octahedron can be clearly rccogniacd (distances in Angstrom
units).
appear to have the undistorted CsCl structure a t ordinary temperatures

(150). A 5-coordinated structure, based on a distorted octahedron but with
one corner missing also occurs in complex salts of the type M+SbF4and
MZ+SbFS (39).
Pronounced off-center distortions of a rather different kind occur in
tetragonal PbTiOa (194) and orthorhombic PbZr03 (70). Both crystals
have deformed perovskite structures. In the ideal cubic structure (Fig.
23) each Pbff ion would be in cubic closest packing with 12 oxygen neigh-
FIG. 23. The cubic perovskite (AB03) structure. The small B ion is octahedrdly
surrounded by oxygens, the large A ion is in cubic closest packing with 12 oxygen neighbors. Removal of the large A ion gives the Rc03 structure.
bors, but in the titanate the Pb++ ion poves by 0.5 8 from its ideal position
to give four nearest neighbors (2.53 A) arranged in a similar fashion as in
red PbO, four more distant ones (2.80 A), and four very distant ones (3.20
A). In the zirconate there is also a large displacement of the Pb++ ion, but
this tiFe along a threefold axis of the cube, to give three nearest neighbors
(2.55 A) at the corners of a rather flat trigonal pyramid, with six more dis-
44
J. D . DUNITZ AND L. E. ORGEL
tarit arid three very distant rdghbors. It must be emphasized that ill these
compounds, the Ti4+and Zr4+ ions also undergo considerable displacements
from the centers of their surrounding oxygen octahedra. The nature of these
is discussed in the next section.
Extreme forms of the trigonal distortion shown in PbZrOs occur in
orthorhombic (yellow) PbO and in Pb304(26). In both of these compounds,
the environment of the Pb++ ions is best described as 3-coordinated, based
on tetrahedral coordination but with one corner missing, and with a rather
indefinite number of next nearest neighbors. The short Pb-0 distances are
about 2.24 8. A very similar 3-coordinated arrangement of oxygens about
the Sb3+ion occurs in compounds of the type M++Sb204(isostructural with
Pb304) and also in the low-temperature modification of Sb203,and about
the Bi3+ ion in Biz03(66). These structures, however, can scarcely be regarded as distortions of simple ionic ones and are perhaps better considered in the usual way, as involving covalent bonding, with the inert
pair pointing away from the other ligands. But although the approach from
distortion of regular ionic packing may be misleading if taken too seriously.
it does enable us to recognize the intermediate stages in the transition from
high-coordinated, regular, ionic environments to low-coordinated, unsymmetrical, covalent ones.
The low-coordination structures, so characteristic of the oxygen compounds of Pb++, do not seem to occur in the halides. At least two modifications of PbFz are known to occur, one with the 8-coordinated cubic fluorite
structure, the other, held in common with PbC12, PbBrz, BaC12, BaBrn,
etc., with the metal ions in 9-coordination as in Fig. 3d. Interatomic distances are not known with precision for any of these compounds and it
may also be mentioned that the evidence for the plane of symmetry in this
structure cannot be regarded as convincing by modern criteria. The interatomic distances reported for the fluoride (%) make five Pb-F bonds
(2.50 A) shorter than the other four (2.85 A) : for the chloride (17),one dis
tance is 2.67 8, two are about 2.90 8, and the other six are greater than
3 8. The iodide is also dimorphic (150),occurring with the CdL and CdCL
structures, in both of which the metal ions are presumably octahedrally
coordinated. Thus the (inert pair does not seem to play such an important
role in determining the structures of the halides, as it does for the oxides.
I t is possible, making use of known values of ionic polarizabilities, to
calculate semiquantitatively the extent to which polarization stabilizes
the distorted structures (110).The stabilization energy, like the restoring
potential, increases quadratically with the extent of the distortion and is
of the same order of magnitude. Particularly interesting is the R+ dependence of the stabilization on the interatomic distance R, for this suggests
that the small anions F- and 0- - should lead to particularly pronounced
45
stereochemical effects. This is in contrast to the predictions of a purely

covalent model of the distortion mechanism, which would lead to larger
distortions in, say, sulfides than in oxides. It is, therefore, particularly
interesting that metallic PbS has the rocksalt structure but that both
forms of PbO are strongly distorted. Similarly, TlCl has a cubic structure
but TlF is distorted. On thc other hand, one would certainly expect a much
more pronounced inert pair distortion in PbFz than probably occurs,
particularly in the low-temperature PbClz-like modification.
VI. Small Ion Distortions and Rattling (109)
Up till now, the distortions from cubic symmetry that we have discussed have all been attributed to various features of the electronic structure of the metal ion which lack spherical symmetry and which hence lead
to more or less strongly directional electrostatic interactions with the
environment. In other words, these deviations from ideal ionic packing
result from the breakdown, in a number of cases, of the assumption made
on p. 2 that ions can be regarded as spheres.
We now consider a quite different type of distortion, a good example
of which occurs in the structure of vanadium pentoxide, VZOS (27). The
environment of the V6+ ion is illustrated in Fig. 24. Each V6+ ion has five
0
FIG.24. Environment of the V6+ion in vanadium pentoxide (distances in angstrom
units).
oxide neighbors in a trigonal bipyramid-like arrangement at distances

ranging from 1.54 to 2.02 8. At the very great distance of 2.81 A from the
V6+ ion occurs a sixth oxide neighbor which can be regarded as completing
a very distorted octahedron. Clearly, none of the electronic factors previously introduced can play any important role in this very marked offcenter displacement which, although rather extreme, is typical of a whole
class of distortions to be found in the higher oxides of transition metals.
46
J. D. DUNITZ AND L. E. QRGEL
We believe that this class of distortions has its origin in the breakdown
of other assumptions of the elementary ionic theory, namely, that ions are
incompressible and that an arrangement of ions of one charge about an
ion of opposite charge is stable only if the central ion makes simultaneous
contact with each of its neighbors. We shall see that if we relax these restrictions we do not necessarily predict a sharp transition from octahedral to
tetrahedral coordination a t the limiting ratio, 0.414. On the contrary, a
more realistic consideration of the potential energy of aggregates of ions
shows that a slowly developing instability of the octahedral structure with
respect to the tetrahedral may occur as the size of the central cation decreases. This instability leads successively to a loosening of the central
cation at the center of its surrounding octahedron, to off-center displacements characteristic of ferroelectric and antiferroelectric substances, and
finally to such large distortions that the coordination of the central ion
becomes fivefold or tetrahedral. According to this view, 5-coordinated
structures do not necessarily indicate a curious bonding tendency of the
metal ion but merely represent a stage in the gradual transition from
6-coordinated to 4-coordinated structures.
We suppose that a cation M is surrounded by a regular octahedral
arrangement of oxide ions which, for the moment, may be regarded as fixed,
as they would be if they formed part of a close-packed array. The cations
can then be imagined as occupying the octahedral interstices of such a
close-packed oxide structure. If the cation is sufficiently large (Fig. 25a),
it will be subjected to compression by the fixed environment of oxide ions,
and the potential energy diagram describing the motion of M away from
its central position is of the type shown, for any direction of displacement.
On the other hand, if the cation is sufficiently small, it is no longer compressed. For small displacements, its potential energy depends solely on
Coulombic forces, and only for large displacements do the repulsion forces
come into play. The potential energy diagram is of the form shown in Fig.
25b, and the symmetrical configuration is now one of maximum potential
energy. (Obviously, if the dimensions of the oxide octahedron become very
large, an MO molecule and five oxide ions is more stable than the set of
widely separated ions.)
Somewhere between these two extremes there must be a radius of the
metal ion for which the potential energy curve is $ut for small displacements from the center of the octahedron (Fig. 25c). We call the corresponding interionic distance the maximum contuct distance LO,
for insofar as analogies with classical particles are permissible, the flatness of the potential
energy curve may be said t o correspond to contact between compressible
spheres. If the interionic distance increases the metal ion becomes free to
rattle: if it is decreased the metal ion is subjected to compression. Note
47
too that the maximum contact distance does not correspond to the sum of
the ionic radii as usually defined (see Appendix).
Our assumption that the anions are fixed and form part of a rigid
extended framework within whose octahedral interstices the cations are
free t o move is not, of course, very realistic except, perhaps, when the anions
form a close-packed arrangement. For some other structures it would seem
equally plausible (or implausible) to assume that the cations were rigid
and that the anion environment could be distorted, though only in such
a way as to make an extended framework possible. Neither of these models
is in any sense complete but they are distinct in that they may lead to rather
3 0
V
(a)
(b)
(c)
FIG.25. Variation of potential energy with displacement of a metal ion from the
ccnter of a fixed oxide octahedron (for clarit>yshown in two dimensions): (a) large ion,
(b) small ion, (c) ion of intermediate size.
different kinds of distortion. If the anions are fixed, then the distorted
structure must lack a center of symmetry. If they are free to move to any
extent, the distorted structure may or may not be centrosymmetric. In
fact, most of the higher oxides of transition-metal ions do have structures
based on close-packed, or nearly close-packed anion arrangements so that
off-center displacements of the metal ion are the rule, while symmetric
collapsing of the coordination polyhedron is found much less frequently,
if a t all.
Some comments on distortions from cubic eightfold coordinated oxide
structures are of interest in this connection. Hafnium dioxide, HfOz, has
the 8-coordinated fluorite structure (see Table VI), while ZrOz only adopts
this structure in an unstable high-temperature modification. In the stable,
room-temperature modification, baddeleyite, the coordination is only sevenfold but can still be regarded as derived from a cube with four oxygens a t
the base and one a t one of the upper corners, the remaining two being a t
the midpoints of the cube edges between the unoccupied corners (87a).
Alternatively the Zr atom may be regarded as occupying the common
apex of a tetragonal and of a trigonal pyramid, so oriented that one atom
48
J . D. DUNITZ AND L. E. ORGEL
of the oxygeii square is directly opposite another of the oxygen triangle.

The Zr-0 distances range from 2.04 to 2.26 A. A quite different kind of
distortion from a cube occurs in zircon, ZrSi04. Here (78a),the symmetry
of the Zr environment has been reduced from cubic to only 42m (DM),
and
the cube collapses into two tetrahedra rotated 90" with respect to each
other, one flattened (Zr-0, 2.15 A), the other elongated (Zr-0, 2.29 8).We
se3 that when the size of the central ion is reduced beyond the point a t
which cubic coordination becomes unstable, we do not necessarily get
octahedral coordination but rather distorted cubic environments which,
a t least in the two examples mentioned, lack a center of symmetry.
(a)
(b)
(4
FIQ.26. Potential energy diagrams showing possible types of transition from octahedral (11) to tetrahedral (I)Coordination; -, stable; -,unstable. (a) Transformation
--
from stable octahedral to stable tetrahedral coordination, (b) development of unstable

octahedral coordination before transition to stable tetrahedral coordination, (c) unstable
octahedral coordination changing continuously to tetrahedral coordination (109).
We now consider the consequences of continuously reducing the size

of the central cation in an octahedrally coordinated structure. First, there
may be another structure, perhaps quite unrelated and involvhg tetrahedral or other coordination, which becomes stable relative to the octahedral structure before the latter becomes unstable with respect to small
displacements, that is, before the maximum contact distance is reached.
In this case there is a sudden change from regular octahedral to regular
tetrahedral coordination, as shown in Fig. 26a. The oxides of the divaleiit
transition-metal ions have structures in accord with this model, for, in the
absence of Jahn-Teller effects, they are either octahedrally or tetrahedrally
coordinated.
Another possibility is that instability with respect to small distortions
occurs before any tetrahedrally coordinated structure becomes stable
relative to the octahedral one. In this case, Fig. 26b, we have a range of
metal radii for which the structure is neither regular octahedral nor regular
tetrahedral, but, rather, distorted octahedral. At high enough temperatures,
the metal ions will generally be distributed at random over offcenter sites
if the distortion is small, but at lower temperatures cooperative interactions
betwecn the ions will lead to a structure involving a definite arrangement
49
of displaced metal ions in almost regular octahedral coordination. The

ferroelectric and antiferroelectric oxides with distorted perovskite or ReOs
structures belong to this class. We may carry this argument further by supposing that extensive distortion is possible before a tetrahedral structure
becomes stable. Then, just as for ions with a pair of s-electrons, the distorted
environments will correspond to a coordination number of less than six,
possibly with a sufficient number of more distant neighbors to complete a
very distorted environment, as for example, in V Z O ~ .
Finally, there is the possibility that a sufficiently distorted octahedral
structure may be able to go over continuously into a tetrahedral structure,
as shown in Fig. 26c. An example of this is seen in the MOI oxides. WOa has
a structure (2) which is only a slight distortion of the cubic Re03 type (see
Fig. 23) in which the metal ions occur in the octahedral interstices of a
nearly close-packed array of oxide ions. In MOO,, the distortion from the
ideal structure has increased. Although the coordination can still be described as octahedral, one oxygen is much more distant than the other
five, giving a square pyramidal arrangement of nearest neighbors (4). In
Cr03 (28) the distortion has become extreme. The structure contains chains
-CrOz-O-CrOz-O-CrO~in which each Cra+ ion is tetrahedrally
coordinated, but the chains are arranged in such a way that there are two
very distant neighbors (3.10 A) completing a very distorted octahedral
environment.
WOs, Moo3, and Croattherefore, illustrate three stages in the transition
from octahedral to tetrahedral stereochemistry : slightly distorted, strongly
distorted, and, finally, tetrahedral with two scarcely recognizable additional
ligands.
Since instability to distortions increases as the size) of the metal ion
becomes smaller, it is possible to arrange the various ions of high valency
in approximate order of size from a comparison of the crystal structures
of their oxides. We shall not carry out this analysis in detail but shall draw
attention to a number of important features.
The very largest ions will be 8-coordinated and indeed ThOz, CeOz,
Proz, UOz, HfOz, NpOz, PuOz, and Am02 occur with the regular fluorite
structure (150).For ZrOz,as already mentioned, the stable crystal modification is only 7-coordinated (87u) but in the mineral zircon, ZrSiOl, the Zr4+
ions are again in distorted 8-coordination (1.2, 78a, 138). We may suppose that Zr4+ is the smallest of the above mentioned tetravalent ions.
Next come those ions whose dioxides have the 6-coordinated rutile structure, PbOn, NbOz, OsOz, RuOz, SnOz, IrOz, TaOz, TiOz, and GeOz (see
Table VII and ref. 150). In the last two members, the oxygen-oxygen
distances are such that large anion-anion repulsions must come into play,
and the octahedron cannot contract any further. This suggests that Ti4+
50
J. 1). DUhITZ AND L. E. OHGEL
(and Ge4+) have L close to Lo for the 6-coordinated rutile structure. Thc
metal-oxygen distances have been determined in both TiO, and Ge02 with
good accuracy (9). I n both compounds the octahedr? are not far from regular; Ti-0, 1.94 8 (four), 1.99 8 (twice); Ge-0, 1.87 A (four), 1.91 8 (twice).
The shortest 0-0 distances are 2.52 and 2.40 A respectively.
That Ti4+is close to the critical size is much more convincingly shown
by the occurrence of this ion in a large number of oxide ferroelectrics of the
general formula A++TiO3. At high enough temperatures, these compounds
are cubic with the perovskite structure (regular octahedral coordination),
but, on lowering the temperature, transitions to ferroelectric distorted
structures, characterized by offcenter displacements of the Ti4+ion, OCCUT.
In tetragonal BaTi03, the displacement is along a fourfold axis of the octahedron, the Ti-0 distances being 1.87, 2.00 (four), and 2.17 8 (44). At
lower temperatures, a rhombohedrally distorted structure occurs, the displacement of the Ti4+ion being now presumably along the threefold axis
(96). In tetragonal PbTiO3, the displacement of the Ti4+ ion is even greater
than in the barium compound, the distances being 1.78, 1.98 (four), and
2.38 8 (124). Here a large complementary displacement of the Pb++ ion
also takes place to give the metal ion four nearest neighbors, the arrangement being square pyramidal as in PbO (see Section V.B).
The structure of V 0 2 is closely related t o that of Ti02. It is a monoclinic
distorted rutile-like structure in which the V4+ ion has moved away from the
center of the surrounding octahedron, approximately along a threefold
axis, t o give three short bonds, 1.86, 1.87, and 1.76 A, and three long ones,
2.01, 2.03, and 2.05 (3). The displacement is thus similar to that which
probably occurs in the rhombohedrally distorted perovskites. A distortion
something like that in tetragonally distorted perovskites occurs in hydrated
vanadium dioxide, V2O4.2H20in which there is one V-0 bond of 1.65 8,
much shorter than any other (41). The structures of Mooz, WOz, TcOz,
and Re02 are closely related to that of VOZ (91). These facts suggest that
V4+, Mo4+, W4+, Tc4+, and Re4+ have radii considerably smaller than the
critical contact radius for an oxide octahedron. The possibility that the distortions are attributable to a Jahn-Teller effect can be discounted since this
should always produce an arrangement distorted in such a way as t o
preserve a center of symmetry, contrary t o the observed displacements of
the metal ions from the center of their surrounding octahedra. Nevertheless, some sort of bonding interaction between metal ions, involving the
use of the d electrons, cannot be excluded, and in order to contrast the
crystal chemistry of Ti4+with that of smaller metal ions without introducing complications due to the presence of d electrons, one is obliged to go to
V6+ and Cr6+.
Pentavalent vanadium occurs in a variety of different oxide environ-
STEREOCHEMISTRY O F I O N I C SO LI D S
51
rnents, and may be 4-, 5, or &coordinated. Isolated V04 tetrahedra occur

in the orthovanadates, for example, in HiVO, (119) arid in CrV04 (19).
In the metavanadates, NaV03 (128) and KVO3 (40),chains of tetrahedra
joined by corners are found. In KVO,.H20, the V5+ ion is in an almost
exactly trigonal bipyramidal environment (29). In V206, as we have seen,
the vanadium ion is again 5-coordinated but with a further distant oxide
neighbor to complete a much distorted octahedron. The compound Nav6015
contains V6+ ions in two different types of distorted octahedral environments (143). I n one type the V-0 distances vary from 1.57 t o 2.33 8, in
the second type the distortion is more extreme, one V-0 bond being 2.68 A
so that the ion can be considered to be 5-coordinated. We have already
drawn attention to the gradually increasing departure from ideal ionic
packing which takes place as we go from WOS, through Moo3, to Cr03.
A rather similar relationship exists between the, structures of NaTa03,
NaNb03, and NaV03. While the compounds of the heavier elements are
ferroelectrics based on the perovskite structure, the vanadate is a tetrahedrally coordinated structure. Pentavalent niobium and tantalum appear
to have octahedral coordination, often distorted, in all of their oxide compounds. Those of the composition, A+Nb03, A+Ta03, have perovskite-like
structures (Ids),those of the composition A++Nb20~,A++Ta2o6have the
niobite structure (20). In niobite itself, (Fe,Mn)Nb206,the Nb6+ ion is in
octahedral coordination but is displaced from the center towards one of
its neighbors to form one rather short bond, 1.86 8, the other five contacts
being from 2.04 to 2.12 8 (131). Compounds of the formula A3+Nb04,
A3+Ta04have a rutile-like structure, with the two kinds of metal ion randomly distributed over the octahedral sites (20).
It would seem that W6+ and Mo6+ions are able to occur in both tetrahedral and octahedral (usually deformed) coordination. Although most of
the simple tungstates and rnolybdates have structures based on tetrahedral
Coordination the series Ca3WOs,SrsWO6, Ba3W06, and a number of related
molybdates possess octahedrally coordinated (possibly deformed) structures (130). The more complex para-, meta-, and poly-acid salts (81,82, 83)
and the heteropoly-acids (75) are also octahedrally coordinated. Sodium
dimolybdate Na2M0207(and the isostructural ditungstate) contains highly
distorted Moo6 octahedra and MOO, tetrahedra (84).
These selected examples, and a number of others, suggest that for the
transition-metal ions with empty d shells, the contact radius decreases (and
instability t o distortion increases) in the approximate order Hf 4+, Zr4+,
Ti4+, Ta6+, Nb5+, W6+, Mo6+, V5+, Cr6+,Mn'+ (Table XV). Ferroelectrics
are largely restricted to compounds containing the group of ions of intermediate size, Ti4+,Ta5+,Nb5+,W6+. We can see, in a qualitative way, why
this should be so. Larger ions are too stable in regular octahedral environ-
52
TABLE
xv
CUURLUNI\TION ~ I L Y H E D H A IN OXII)EB OW 'rHANCjITION-METAL
WITH
Ion
Pauling
ionic radius
Hf'+
Zr4+
0.80
Ti4+
0.68
Tab+
Nb6+
0.70
WE+
M 06'
0.62
V'+
0.59
CrE+
0.52
Mn7+
0.46
(A)
IONS
INERTGAS CONFIGURATIONS
Coordination
Cubic
Cubic
Distorted cubic
Octahedral
Octahedral
Distorted octahedral
Octahedral
Octahedral
Tetrahedral
(Square pyramidal +1 distant)
Tetrahedral
Very distorted octahedral
(3 near
3 distant)
Trigonal bipyramid (+1
distant)
Trigonal pyramid
Tetrahedral
Tetrahedral (+2 distant)
Tetrahedral
Tetrahedral
-+
Examples
HfOI
Z r O l ( > 1600C)
ZrOa, ZrSiO,
BaZrOa
TiOt, BaTiOa (>120")
BaTiO3
NaTaOJ (>475")
NaTaOa
NaNbOs (>640j
FeNbzOs, NaNbOs
WO,
Caw08
Moos
CaMoOd
vos
VaOa
KVOa . Ht0
NaV03
CrOt
BaCrOc
KMnO,
The simple ionic theor predicts transition from octahedral to tetrahedral oxide
environment at r y 0.58 %,aasuming re-- = 1.40 k
ments, while smaller ions form structures which are so strongly distorted
that no phase transition occurs below the melting point, and, if the structure
is polar, the field required to reverse the polarization cannot be attained.
We must emphasize, too, that the idea of a critical ionic radius at which
"rattling" of metal ions within their oxide environments begins is part of
the Mason and Mathias theory of ferroelectricity (93).
VII. Conclusions
We have seen that many characteristic features of metal ion stereochemistry which involve the occurrence of ions in sites of unexpectedly low
symmetry can be explained without recourse to detailed arguments about
53
the nature of the chemical bonding. We have emphasized particularly that

ions with incomplete d electron shells may intrinsically have less than
spherical symmetry. In such cases the electronic structure may be incompatible with the existence of a cubic environment about the metal ion.
I n a rather similar way ions which have all their electrons in closed shells
may nevertheless show a strong tendency to occur in distorted environments if there exist sufficiently low lying excited states. Essentially the electronic distribution is then so easily distorted that the ion becomes less than
spherically symmetrical once it is brought into contact with ions of the
opposite charge. The distortion of the electronic cloud corresponds to a
simple polarization in the case of (ns)2ions, and hence the environment
must lack a center of symmetry if it is to produce an appreciable stabilization energy. With dl0 ions the nature of the polarization is different (quadrupole polarization) and a quite different environment is required.
In addition to these effects of the electronic properties of the free ions
on stereochemistry we have also discussed the effect of ionic size somewhat
more fully than is usual. Again we find that certain distortions from cubic
symmetry can be understood qualitatively in terms of the properties of
ions without detailed consideration of covalency.
In every case we have emphasized that the distortions are not without
effect on the covalent character of the bonds involved. We have tried to
show that the tendency to distortion could often have an electrostatic
origin, and hence that other than stereochemical arguments must be used
in determining the degree of covalency both in distorted and in undistorted
complexes. Such methods include the study of antiferromagnetic interaction, nuclear magnetic resonance spectra, and paramagnetic resonance
spectra.
For convenience we have usually supposed environments of less than
cubic symmetry to arise by distortion from fully cubic ones. This is particularly useful when the distortions are small, but inadequate if the deviations
from cubic symmetry are very large. Even in the latter case our general
principles should apply, for example, centrosymmetrical environments
should be common for d9 and dl0 ions but not for sa ions. The justification
for our approach is that it provides a useful bridge between cubic and other
structures which enables the cause of the unusual stereochemistry to be
more easily discerned.
Finally we note that once the various types of characteristic distortions
have been recognized, predictions of new properties can be made. For example, oxides of s2 ions should form mixed crystals with oxides of transition
metals in sufficiently high valencies, since, for different reasons, they are
both susceptible to antisymmetric distortion. This seems to be the case.
54
J. D. DUNITZ AND L. E. OltGBL
Again, cubically coordinated T1+ salts should be particularly susceptible t o

polarization by applied fields; that is, they should have anomalously large
dielectric constants.
VIII. Appendix: Ionic Radii
Of the several systems of 5onic radii which have been proposed to

account for the approximately additive relationships which exist between
the observed interatomic distances in ionic crystals, the one which has come
into most general use is that of Pauling (111,114).This is not, as is variously
supposed, either a set of empirical radii derived purely from the experimental data which it is, in turn, supposed to reproduce, or one which has
been derived purely from theoretical considerations. It is a semiempirical
system in the sense that from a very limited set of experimental data, certain relationships are derived using approximate theories of atomic and
crystal structure, which adequately account for a much wider set of data.
Since no precise physical significance can be attached to the concept of
atomic or ionic radius (the electronic wave functions approach zero asymptotically), the radii to be assigned are those which reproduce the equilibrium
interatomic distances in ionic crystals. These distances depend on the balance between the attractive and repulsive forces, and thus not only on the
electron distributions of the ions but also on the crystal structure and the
radius ratios. Pauling assumes that the relative sizes of a pair of isoelectronic ions are inversely proportional to the effective nuclear charges operating on the outmost electron shell; that is
7-1
C,/(Z - S ) ,
and divides the experimental interatomic separation in this ratio. The

quantity C , is a constant for a given isoelectronic sequence, depending
only on the total quantum number of the outmost electrons; Z is the atomic
number; and S is a screening constant. For neon-like ions, S is 4.52 and
hence the effective nuclear charges for Na+ and F- are 6.48e and 4.48e
respectively. We now have
rl(Na+)
rl(F-)
rl(Na+)
4.48,
6.48
+ rl(F-) = 2.31 A,
and hence rl(Na+) = 0.95

and zl(F-) = 1.36 A. From the observed
interatomic distances-Na-F, 2.31 A; K-Cl, 3.14 A; Rb-Br, 3.43 8;Cs-16,
3.85 A-the effective radii of the singly charged ions are calculated and
8 This value is that observed in CsI reduced by 2.7% to give the value for 6-coordination, assuming n = 12 (see p. 56).
STEREOCHEMISTHY OF I O S I C SOLIDS
55
the coristaiits C, are then used to derive a set of univalent radii for
multiply charged ions with inert gas configurations. These LLunivalent
radii
are the radii which multivalent ions would possess if they were to retain
their electron distribution but enter into Coulomb attraction a s if they were
univalent. They do not, as they stand, reproduce observed interatomic
distances. I n order to convert them into crystal radii we have to take into
account the effect of the ionic charge on the equilibrium distance. From
the Born equation we have
BeZ
Ti=--- AeV
&
+%
where V is the potential energy; A , the Madelung constant; and B, the

Born coefficient.
Hence,
dV - Ae2z2 nBeZ
R2
RR+
dR
At equilibrium dV/dR = 0 and R, = (nB/Az2)1/(n-1).If the ions were to
enter into Coulomb attraction as if they were monovalent, with the repulsion coefficient unchanged, the equilibrium interatomic distance would be
R1 = (nB/A)(n-)
corresponding to addition of univalent radii, and hence
R,
R1/z2/(-),
Approximate values of n are known for various types of ions and are used
to calculate the crystal radii of Table III.7
Since the ionic radii of Table I11 have been obtained with reference to
the NaCl type of structure as standard, it is not t o be expected that they
should apply t o other types without corrections to take into account possible variation of A and B. For two different structures with Madelung
constants A 1 and Az, and repulsion coefficients B , and B:! we have
If i t is now assumed that B , and Bz are proportional to the respective coordination numbers for the two structures, we can virtually ignore the ratio
A 1 / A 2which is always close to unity and write
7 The ionic radius assigned to Lit (0.60 A) is obtained from a further cxperimental
quantity, namely the observed Li-0 distance in Liz0 (2.00 A) from which thc cnlcnlated
ionic radius of 0-- (1.40 d) is subtracted.
56
Thus, for a change in coordiiiation number from 6 to 8 and for n = 9, the

standard ionic radius must be multiplied by (8/6)1M = 1.036. For a change
from 6 to 4, the corresponding factor is 0.950.
In the derivation of these ionic radii, it has been assumed that the repulsion coefficient B depends only on the coordination number; that is, on the
number of anion-cation contacts, but if the radius ratio is close to or less
than the lower limit, anion-anion contact occurs and the additional Born
repulsion will lead to equilibrium with the attractive Coulomb forces at
a larger distance than that given by the sum of the ionic radii. This phenomenon of double repulsion is shown (see tabulation) by the lithium
halides especially. In a more detailed treatment, Pauling (112, 114) has
~
Sum of ionic radii (A)

Observed distance (A)
Radius ratio rM/rx
~~
~~~
LiF
LiCl
LiBr
LiI
1.96
2.01
0.44
2.41
2.57
0.33
2.55
2.75
0.31
2.76
3.02
0.28
used a modified form of the Born expression for the lattice energy COIIsidered now not as a function of the cation-anion distance alone, but taking
into account also the dependence of the Born repulsion energy on anionanion and cation-cation distances. The equilibrium value of R now depends
not only on the magnitudes of the ionic radii but also on the radius ratio
Bo
(Tmt
+ rr)F
(c),
where F is a rather complicated function of the radius ratio. Pauling has

evaluated this function for the NaCl structure, and has shown that it compensates exactly for the deviations from strict additivity in the alkali
halide series.
For crystals in which the charge of the anion and cation are unequal it is
more appropriate to use the univalent radii for estimating the radius ratio.
This is because the stabilities of the various coordination polyhedra depend
on the relative magnitudes of cation-anion and anion-anion repulsions.
For calculating the absolute distances between multiply-charged ions in
such crystals, it is sufficient t o use the direct sum of the crystal radii without correction (see ref. 114, p. 365).
REFERENCES
1. Almin, K. E., and Westgren, A., Arkiv Kemi, Mineral. Geol. 168, No. 22 (1942).
8. Andersson, G., Acta Chem. Scand. 7 , 154 (1953).
3. Andersson, G., Acta Chem. Scand. 10, 623 (1956).
4. Andersson, G., and Maw&, A., Acta Chem. Scand. 4, 793 (1950).
6. Aurivilliua, K. L., Acta Chern. Scand. 4, 1413 (1950).
57
6'. Aiirivillius, I<. L., Acta Chem. Sca7~d.10,852 (1956).
7. Bartlett,, N., and Maitland, R., Acta Cryst. 11, 747 (1958).
8. Baur, W. H., Ada Cryst. 9, 515 (1950).
9. Baur, W. H., Acta Cryst. 11, 488 (1958).
10. Bijvoet, J. M., Claascn, A., and Karssen, A., Proc. Acad. Sci. Amsterdam 29, 529
(1926).
11. Billy, C., and Haendler, H., J. Am. Chem. SOC.79, 1049 (1957).
12. Binks, W., Mineral. Mag. 21, 176 (1926).
13, Bjerrum, J., Ballhausen, C. J., and Jsrgensen, C. K., Acta Chem. Scand. 8, 1275
(1954).
14. Bleaney, B., and Stevens, K. W. H., Repts. Progr. in Phys. 16, 108 (1953).
16. Bokai, G. B., Malinowski, T. I., and Ablov, A. V., Kristallograjiya 1, 49 (1956).
16. Bowers, K. D., and Owen, J., Repts. Progr. i n Phys. 18, 304 (1955).
ir. Braekken, H., 2. Krist. 83, 222 (1932).
18. Braekken, H., arid Scholten, W., 2. Krist. 89, 448 (1934).
19. Brandt, K., Arkiv Kemi, Mineral. Geol. 178, No. 6 (1943).
20. Brandt, K., Arkiv Kemi, Mineral. Geol. 178, No. 15 (1943).
21. Brosset, C., Arkiv Kemi, Mineral. Geol. 268, No. 19 (1948).
22, Brown, I. D., and Dunitz, J. D., Abstr. Papers Intern. Conf. on Co-ordination
Chem. London Paper 87 (1959).
d3. Buerger, M. J., 2. Krist. 896, 163 (1936).
24. Bystrom, A., Arkiv Kemi, Mineral. Geol. 208, No. 11 (1945).
26. Bystrom, A., Arkiv Kemi, Mineral. Geol. 248, No. 33 (1947).
26. Bystrom, A., Arkiv Kern,( Mineral. Geol. 268, No. 13 (1948).
27. Bystrom, A., Wilhelmi, K. A., and Broteen, O., Acta Chem. Scand. 4, 1119 (1950).
28. Bystrom, A., and Wilhclmi, K. A., Acta Chem. Scand. 4, 1131 (1950).
29. Christ, C. L., Clark, J. R., and Evans, H. T., Acta Cryst. 7, 801 (1954).
30. Chrobak, L., 2. Krist. 88, 35 (1934).
Sf. Clark, E. S., Templeton, D. H., and MacGillavry, C. H., Acta Cryst. 11, 284 (1958)
82. Cromer, D. T., J . Phys. Chem. 61, 1388 (1957).
33. Dannhauser, W., and Vaughan, P. A., J . Am. Chem. SOC.
77, 896 (1955).
34, Dunitz, J. D., Acta Cryst. 10, 307 (1957).
36. Dunitz, J. D., and Orgel, L. E., Nature 179, 462 (1957).
S6. Dunitz, J. D., and Orgcl, L. E., J . Phys. Chem. Sc2ids 3, 20 (1957).
87. Dunite, J. D., and Orgel, L. E., J . Phys. Chem. Solids 3, 318 (1957).
38. Dyer, L. D., Borie, B. S., and Smith, G. P., J . Am. Chem. SOC.76, 1499 (19.54).
39. Edstrand, M., Inge, M., and Ingri, N., Acta Chem. Scand. 9, 122 (1955).
89a. Edwards, A. J. and Peacock, R. D., J . Chem. Soc. p. 4126 (1959).
40. Evans, H. T., and Block, S., Am. Mineralogist 39, 327 (1954).
41. Evans, H. T., and Mrose, M. E., Acta Cryst. 11, 56 (1958).
42. Figgis, B. N., and Nyholm, R. S., J. Chem. SOC.p. 12 (1954).
43. Frasson, E., Rardi, R., and Bczzi, S., Acta Cryst. 12, 201 (1959).
44. Frazer, B. C., Danncr, R. R., and Pepinsky, R., Phys. Rev. 100,745 (1955).
46. Gillespie, R. J., and Nyholm, R. S., Quart. Revs. 11, 339 (1957).
46. Gorskii, V. S., Physik. 2. Smujetunion 6, 367 (1934).
47. Grdenic, D., Arkiv Kemi 22, 14 (1950).
48. Grdenic, D., and Krstanovic, I., Arkiv Kemi 27, 143 (1955).
49. Greenwald, S., Acta Cry&. 6, 396 (1953).
60. Gregg, A. H., Hampson, G. C., Jenkins, G. I., Jones, P. L. F., and Sutton, L. E.,
Trans. Faraday SOC.33, 852 (1937).
58
J. D. DUNITZ -4ND L. E. ORGEL
E., Qua~t.Reus. 11, 381 (1957).

6%. Hampson, G. C., and Stosick, A. J., J . Am. Chem. SOC.60, 1814 (1938).
6.9. Hanic, F., and Cakajdova, I. A., Actu Cryst. 11, 610 (1958).
64. Harker, D., 2. Krist. 93, 136 (19%).
55. Harmscn, E. J., 2.Krist. 100, 208 (1939).
56. Harvey, K. B., private communiration (1959).
67. Hassel, O., and Hvoslef, J., Acta Chem. Scand. 8, 1953 (1954).
58. Hayes, W., Discussions Faraduy Sac. 26, 58 (1958).
6Q. Helmholz, L., Z. Krist. 96, 129 (1936).
60. Hclmholz, L., J. Am. Chem. Sac. 69, 886 (1947).
61. Helmholz, L., and Kruh, R. F., J. Am. Chem. Sac. 74, 1176 (1952).
62. Hepworth, M. A., Jack, K. H., and Nyholm, R. S., Nature 179, 211 (1957).
63. Hoard, J. L., 2. Krist. 84, 231 (1933).
64. Hoard, J. L., Martin, W. J., Smith, M. E., and Whitney, J. F., J. Am. Chem. Soe.
76, 3820 (1954).
66. Hofmann, W., 2. Krist. 92, 161 (1935).
66a. Hoppe, R., Angew. Chem. 71, 457 (1959).
66, Hoshino, S., J. Phys. Soc. Japan 12, 315 (1957).
66a. Ijdo, D. J. W., private communication (1959).
67,Jack, K. H., and Maitland, R., Proc. Chem. SOC.p. 232 (1957).
08. Jahn, TI. A., and Teller, E., Proc. Ray. Sac. Al61, 220 (1987).
69. Jellinek, F., Actu Cryst. 10,820 (1957).
70. Jona, F., Shirane, G., Marsi, F., and Pepinsky, R., Phys. Rev. 106, 849 (1957).
71. Jones, R. E., and Templeton, D. H., Acta Cryst. 8, 847 (1955).
72. Jsrgensen, C. K., Proc. 10th Solvay Congr. in Chem., Brussels p. 355 (1956).
73. Jsrgrnsen, C. K., Thesis, Copcnhagen (1957).
74. Kanamori, J., Pragr. Theoret. Phys. (Kyoto) 17, 177, 197 (1957).
76. Kcggin, J. F., Proc. Roy. Sac. 8144, 75 (1934).
76. Ketelaar, ,J. A. A., Chemical Constitution, 2nd rev. ed. Elsevicr, Amhcrdam
1958.
77. Kiriyama, H., Tbamoto, IT., and Matsuo, K., Acla Cryst. 7, 482 (1954).
78. Knox, K., J . Chem. Phys. 30, 991 (1959).
78a. Krstanovic, I. R., Actu Cryst. 11, 896 (1958).
79. Li, Y . Y.,Phys. Rev. 100, 627 (1955).
80. Liehr, A. D., and Ballhausen, C. J., Ann. Phys. ( N . Y . )3, 304 (1958).
81. Lindqvkt, I., Arkiv Kemi 2, 325 (1950).
82. Lindqvist, I., Arkiv. Kemi 2, 249 (1950).
83. Lindqvkt, I., Acta Chem. Scund. 4, 551 (1950).
84. Lindqvist, I., Acta Chem. Scund. 4, 1066 (1950).
86. Lotgering, F. K., Philips Research Rept. 11, 190 (1956).
86. Lux, H., and Proeschal, E., 2.Anorg. Chem. 267, 73 (1948).
87. McClure, D. S., 1.Phys. CAem. Solids 3, 311 (1957).
87a. McCullough, J. D., and Trueblood, K. N.,A d a Cryat. 12, 507 (1959).
88. MacGillavry, C. H., and Bijvoet, J. M., 2. Krist. 94, 231 (1936).
89. MacGillavry, C. H., Wilde, J. H., and Bijvoet, J. M., 2. Krist. 100, 212 (1939).
90. Mathieson, A. McL., and Welsh, H. K., Acta Cryst. 6, 599 (1952).
91. Magneli, A., and Andersson, G., Actu Chem. Scand. 9, 1328 (1955).
92. Maki, G., J. Chem. Phys. 28, 651 (1958).
93. Mason, W. P., and Mathias, B. T., Phys. Rev. 74, 1622 (1948).
94. Maxwell, L. R., Hendricks, 8. B., and Deming, L. S., J . Chem. Phys. 6, 626 (1937).
61. Griffith, J. S., and Orgcl, L.
59
96. Mazai, F., Acta Cryst. 8, 137 (1955).

96. Megaw, H. D., Acta Cryst. 6, 739 (1952).
97. Moffitt, W., and Ballhausen, C. J., Ann. Rev. Phys. C h m . 7, 107 (1956).
98. Moorc, C.A., Atomic Energy Levels, Natl. BUT.Standards (U.S.) C ~ T C467,
. Vol.
I1 (1952); Vol. I11 (1958).

99. Moorc, W. J., and Pauling, L., J . Am. Chem. SOC.63, 1392 (1941).
100. Niggli, P., 2. Krist. 67, 253 (1922).
101. Nyholm, R. S., Quart. Revs, 7, 377 (1953).
102. Nyholm, R. S., PTOC.
10th Solvay Congr. in Chem., Brussels p. 235 (1956).
103. Opik, U., and Pryce, M. H. L., PTOC.
Roy. SOC.A238, 425 (1957).
104. Orgel, L. E., J . Chem. SOC.p . 4756 (1952).
106. Orgel, L. E., J . Chem. Phys. 23, 1004 (1955).
106. Orgel, L. E., PTOC.
10th Solvay Congr. in Chem., Brussels p. 289 (1956).
107. Orgel, L. E., J . Chem. Soc. p . 4186 (1958).
108. Orgel, L. E., J . Phys. Chem. Solids 7, 276 (1958).
109. Orgel, L. E., Discussions Faraduy SOC.26, 138 (1958).
110. Orgel, L. E., J . Chem. SOC.p. 3815 (1959).
111. Pauling, L., PTOC.Roy, SOC.8114, 181 (1927).
112. Pauling, L., J . Am. Chem. SOC.60, 1036 (1928).
118. Pauling, L., J . Am. Chem. Soc. 63, 1367 (1931).
114. Pauling, L., The Nature of the Chemical Bond, 2nd ed. Cornell University
Press, Ithaca, N. Y . , 1940.
116. Porai-Koshits, M. A., Trudy Znst. Krist. Akad. Nauk. S.S.S.R. 10, 288 (1954).
116. Powell, H. M., and Wells, A. F., J . Chem. SOC.p. 359 (1935).
117. Prince, E., and Trcuting, R. G., Acta Cryst. 9, 1025 (1956).
118. Prince, E., Acta Cryst. 10,554 (1957).
119. Qurashi, M. M., and Barnes, W. H., Am. MineraZogist 38, 489 (1953).
120. Romeijn, F. C., Philips Research Rept. 8, 304 (1953).
121. Rooksby, H. P., Acta Cryst. 1, 226 (1948).
2.22. Scatturin, V., and Turco, A., J . Znorg. & Nuclear Chem. 8, 447 (1958).
1.25. Scavnicar, S., and Grdenic, D., Acta CTySt. 8, 275 (1955).
124. Shirane, G., Pepinsky, R., and Fraser, B. C., Acta Cryst. 9, 131 (1956).
126. Sinha, A. P. B., Sanjana, N. R., and Biswas, A. B., Acta Cryst. 10,439 (1957).
1.26. Siwel, I., Through the Mirror-plane. Hcirus (1959).
1.97. Sollcr, W., and Thompson, A. J., Phys. Rev. 47, 664 (1935).
i28. Ssrum, H., Kgl. Norske Videnskab. Selskabs Forh. 16, 39 (1943).
129. Stephenson, N. C., and Mellor, D. p., Australian J . Sn. Research AS, 581 (1950).
150. Stcward, E. G., and Rooksby, H. P., Acta Cryst. 4, 503 (1951).
131. Sturdivant, J. H., 2.Krist. 76, 88 (1930).
23.2. Tombs, N. C., and Rooksby, H. P., Nature 166, 442 (1950).
ISS. Tunell, G., Posnjak, E., and Ksanda, C. J., 2.Krist. 90, 120 (1935).
134. Vainstein, B. I., Doklady Akad. Nauk S.S.S.R. 68, 301 (1949).
156. van Niekerk, J. N., and de Wet, J. F., Acta Cryst. 6, 501 (1953).
136. van Niekerk, J. N., and Schoening, F. R. L., A d a Cryst. 6, 227 (1953).
137. van Vleck, J. H., J . Chem. Phys. 7, 472 (1939).
158. Vcgard, L., Phzl. Mag. [7] 1, 1151 (1926).
139. Venanzi, L. M., PTOC.Intern. Symposium on Coordznation Compounds, Rome, 1967
p. 137 (1958).
140. Verweel, H. J., and Bijvoet, J. M., 2. Krist. 77, 122 (1931).
142. Vousden, P., Acta Cryst. 4, 373 (1951).
60
J.
n.
DUNITZ AND L. E. ORGEL
143. Wadsley, A. D., Acla Cryst. 8, 695 (1955).

144. Weiss, Alarich, and Weiss, Armin, Z.Nalurforsch. 116, 604 (1956).
146. Wells, A. F., J . Chem. SOC.p. 1662 (1947).
146. Wells, A. F., J . Chem. SOC.p. 1670 (1947).
14r. Wells, A. F., Structural Inorganic Chemistry, 2nd ed. Oxford University Press,
London and New York, 1950.

148. Wells, A. F., Ada Crgst. 4, 200 (1951).
149. Wyckoff, R. W. G., Am. J . Sci. 3, 184 (1922).
160. Wyckoff, R. W. G., Crystal Structurw. Intersciencc, New York.
161. Zachariasen, W. H., Phy8. Rev. 40, 917 (1932).
ORGANOMETALLIC COMPOUNDS
John Eisch* a n d Henry Gilman
Department of Chemistry. St Louis University. St Louis. Missouri. and Department of Chemistry.

Iowa State College. Ames. Iowa
I. Introduction . . . . . . . . . . . . . .
A . Concept of the Carbon-Metal Bond . . . . . .
B. General Characteristics of Organometallic Compounds .
C. Development of Organometallic Chemistry . . . .
D . Renaissance in Organometallic Chemistry and its Sources
. . . . .
. . . . .
.
.
.
. . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
I1. New Preparative Techniques and Procedures

.
A . Metal Hydrides and Olefins . . . . . . . . . . . . .
B. Aluminum Alkyls as Alkylating Agents . . . . . . . . . .
C . Methylenation via Diaaomethane . . . . . . . . . . .

D. Organometallic rz-Complexes . . . . . . . . . . . .
E. Organometallic Compounds of Transition Metals and Metalloids . .
F. Preparation of Vinyl Organometallic Compounds in Tetrahydrofuran
G . Alkali Metal Adducts with Unsaturated Hydrocarbons
. . . .
H . Electrolytic Preparation of Organometallic Compounds . . . .
111. Structure and Bonding in Organometallic Compounds . . . . . .
A. Autocomplexation . . . . . . . . . . . . . . .
B. Carbon-Metal Bonding . . . . . . . . . . . . . .
C. Metal-Metal Bonding . . . . . . . . . . . . . .
IV. Reactivity and Reaction Pathways of Organometallic Compounds . .
A . Structure and Reactivity . . . . . . . . . . . . .
B. Cleavage of the Carbon-Metal Bond . . . . . . . . . .
C. Reaction Pathways . . . . . . . . . . . . . . .
V . Applications of Organometallic Reagents in Research and Technology .
A . Synthesis . . . . . . . . . . . . . . . . . .
3 Technology . . . . . . . . . . . . . . . . .
VI . Present Trends in 0rganomet.alIit: Research . . . . . . . . .
List of Symbols .
References
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
61
61
63
70
72
73
73
77
78
78
80
81
82
82
83
83
84
85
.
.
85
85
88
89
.
.
.
98
98
98
99
. 99
. 99
I . introduction
A . CONCEPTOF THE CARBON-METAL

BOND
1. Metallic Character
Classically organometallic chemistry is considered to deal with substances containing a direct carbon-metal bond . The scope of this field. how* Present address: University of Michigan. .4nn Arbor. Michigan .
(i 1
62
JOHN EISCH A N D HENRY GILMAN
ever, expaiids as the definition of a metallic elemelit is liberalized. If twth

physkal properties such as high thermal and electrical conductivity, rcmarkable cohesiveness and luster, and chemical characteristics understood
under the term base-former, are considered to define metallic character,
probably three-fourths of the elements have acceptable metallic credentials.
The demarcation between organometallic compounds and certain excluded
types, therefore, is bound to be hazy if based upon such a general notion of
metallic character, and the distinction can only become sharp by arbitrary
selection of typical metals. Certain organometalloid compounds, such
as of boron, silicon, arsenic, selenium, tellurium, and even iodine, exhibit
some behavior reminiscent of organometallic types, and their consideration
in organometallic chemistry can be justified by the more general treatment
of the elements thus achieved. However, divergence from typical organometallic character is to be expected of such metalloid types for two reasons.
First, since the carbon-metal bond owes its reactivity in part to the difference in the electron-attracting tendencies of the atoms involved, the
diminished polarity of the carbon-metalloid bond will be reflected in its
lessened reactivity. Second, reactions typical of organometallic compounds
such as rapid oxidation, facile cleavage by protic agents, and complex
formation seem to require unfilled electronic orbitals in the metallic atom
of the carbon-metal bond. That certain organic derivatives of the metalloids
and nonmetals (Groups IVA-VIIA) have no low-lying unfilled orbit,als
rationalizes their inertness toward Lewis bases and oxidation of their
carbon-element bond. For the purposes of organometallic chemistry it
may be useful to define an organometallic compound in electronic terms;
that is, as one containing carbon cr-bonded to a more electropositive element
which has available low-lying p-orbitals in the combined state. By way of
illustration this definition would include boron alkyls, but exclude silicori
and arsenic alkyls, as the former has an electronic sextet about horoii,
whereas i n the latter two the metalloid atonis possess octets.
2. Evolving Concept of the Chemical Bond
A sustaining influence on the growth of orgaiiometallic chemistry has

been the recognition of the variegated nature of the carbon-metal bond.
Indeed, from its inception wi+ the study of zinc alkyls by Frankland (39)
in 1849 organometallic chemistry has contributed consistently and significantly to valence theory. Franklands notion of valence as a fixed number,
Although silicon has available 3d-orbitaL3, and arsenic available 4d-and 4f-orbitals,
these are higher in energy. The electronic definition would limit tho term organometallic
compound to those alkyls of Groups I to I11 and to any transition metal alkyl containing a u-carbon-metal bond. The common feature of these alkyls would be the availability
of np-orbitals on the metal atom.
OHGANOMETALLIC COMPOUNDS
63
Gombergs and Paneths (4s) studies of free organic radicals, the bridge
structure of metal alkyls (lOS), the organic sandwich complexes of the
transition metals (71, SS), all have in their turn spurred the theory of the
chemical bond on to fresh viewpoints.
Almost every recognized bond type finds its counterpart in organometallic chemistry. To the traditional ionic and covalent carbon-metal
bonds involving outer s- and p-orbitals of the metal have been added the
aromatic-transition metal r-complexes, where, for example, *-electrons of
the cyclopentadienyl anion, the benzene system, or the cycloheptatrienyl
cation help bind the organic moiety to the transition metal by utilization
of the latters &orbitals. These new organometallic r-complexes in general
do not exhibit the properties of ordinary organometallic compounds. In
addition, the ?r-bonding involves a t least several carbon atoms in relation
to one metal atom, so that the concept of an isolated carbon-metal bond
becomes less meaningful. Finally, the vigorous research devoted to such
r-complexes in recent years combines with the above-mentioned factors to
make their consideration in a separate chapter desirable (see Vol. 1, of this
series, Chapter 2 by Prof. E. 0. Fischer and Dr. H. P. Fritz).
B. GENERALCHARACTERISTICS
OF OHGANOMETALLIC
COMPOUNDS
To place the recent advances in organometallic chemistry in proper
perspective and to serve as an introduction to the chemist unfamiliar with
this branch of chemistry, it is thought desirable t o indicate here the salient
features of traditional organometallic chemistry. The reader desiring
further details may consult any of several excellent references (24, 7.2, 76,
101).
1. Organometallic Bonding
a. Ionic Type. By an ionic organometallic compound is meant that in

which the carbon atom in the organic moiety is anionic in character and is
attracted t o the metal cation in a nondirectional, electrostatic fashion. A
close analogy to ionic metal alkyls is the alkali halide. The phenomena
associated with electrostatic bonding in the latter case, such a s formation
of electrolyte solutions, high melting point, and insolubility in nonsolvating
solvents are encountered with the alkyls of sodium, potassium, rubidium,
and cesium. The sodium alkyls, ReNa , thus are nonvolatile, generally
colorless solids which are insoluble in hydrocarbons and which will decompose before melting. Their violent reaction with protic solvents prevents
observation of electrical conductivity in these solutions, but solutions in
diethylzinc are electrolytes.
Factors favoying the ionization of the carhon-metal bond are a high
64
JOHN EISCH AND HENRY GILMAN
electropositivity of the meta1 as in alkali and alkaline earth alkyls and the
formation of LL carbanion of some stability, for example,
R---C=CB
CHZ=CN-CH~e.
and
The former, an aoet,ylenic anion, is stabilized hy the increased electronegativity of sp-hybridized carbon and the latter, the ally1 anion, by the r-delocalization of the negative charge.
b. Covalent Type. The majority of organometallic compounds possesses
a directed, though polar, covalent bond between the carbon and metal
atoms. This implies that the electron cloud representing the position probability of the electron pair is denser about the carbon atom. The degree of
unequal sharing of the electron pair in the u-bond,
6-
d-k
C-M,
is assased by reference to empirically established electronegativity scales

whose values seek to assign relative electron-seeking tendencies to the
elements. Although there is some disagreement among the various scales
proposed, the values in general increase with electron-attracting or nonmetallic character. Typical values from the widely accepted Pauling scale
rltrige from 0 to 4:Cs, 0.7; Ba, 0.8; Mg, 1.2; AI, 1.5; B, 2.0; C, 2.5; and F,
4.0. Suggested relationships exist for converting electronegativity differences between the carbon and metal atom into partial ionic character
percentages (101). Such measures of carbon-metal bond polarity present a
useful semiyuantitative tool for correlating chemical reactivity with structure. Other factors, however, such as available metal orbitals, steric demands, and electron cloud polarizability often must be given explicit eonsideration (Section 1V.A).
With the directional covalent bonding occurring in this class of organometallic compounds the molecular unit and hence molecular structure
become distinct and meaningful. Since such molecular units have polar
bonds, however, intermolecular forces often cause extensive association.
Trimethylaluminum, for example, exists as a stable dimer, [A1(CH3)&and
others exist as polymers (Section 1II.A).
c. Transition Metal &Orbital Complexes. The enhanced stability and
unique properties of this class of compounds point to a distinct variety of
covalent bonding. Transition metals, that is, elements with unfilled (n-1)
d-orbitals, tend to form stable complexes with various unsaturated organic
systems, although they yield only unstable organometallic compounds of
the conventional type. These ?r-complexes seem to involve sharing between
the unsaturated systems r-orbitals and the unfilled d-orbitals of the transitioti metal. Besides the aromatic types already mentioned in Section I.A.2,
complexes with olefins, acetylenes, and dienes have been characterized.
ORGANOMETALLIC C O M P O U N D S
65
d. Borderline arid ~ ~ x c l ?Clusses.

~ ~ d The admittance of the al)ove transition metal complexes as a branch of organometallic chemistry reopens the
question of whether certain metallic derivatives of organic systems are not
also qualified to be included. If the delocalized cyclopentadienyl anion is a
fitting moiety for an organometallic compound, then one would ask whether
an enolate-anion, containing some of the negative charge on carbon, is riot
also acceptable. Perhaps the recognition of two main branches of organometallic compounds would resolve the difficulty : on the one hand, classical
organometallic compounds, as delineated above; on the other hand, organometallic a-complexes involving organic a-electron donors (without regard
for carbon-metal bonds) and metal atom acceptors (compare Section I.A.2).
Enolate salts could be included in this latter group, as could any organic
type possessing unshared electron pairs which could assume 7r-donor bond
character:
R-A-M
e e
t-)
R-A=M.
This would embrace alkoxides (ROM), carboxylate salts (RCOOM),

amides (H,NM) and other similar metallic salts of acidic organic
compounds. Classical organometallic chemistry would then limit itself to
those compounds containing a direct a-bond between carbon and a metal.
2. Chemical Properties
a. Reactivity of the Carbon-Metal Bond. The most characteristic feature

of the carbon-metal bond is the ease with which it gives way t o C-C, C-H,
C-0, C-X bonds on the one hand, and M-X, M-0 and M-N bonds on the
other hand. Consequently, organometallic compounds generally are reactive toward various inorganic and organic reagents. This might be summarized by pointing out that many organometallic compounds are spontaneously flammable in air, decompose upon relatively mild heating,
hydrolyze readily and tend to add to unsaturated organic compounds. Their
great reactivity has long made them unexcelled synthetic tools in the
laboratory. Although any one chemical reaction is probably not sufficiently
unique to characterize a carbon-metal bond, the presence of a carbon-metal
bond in a given substance can usually be ascertained if several of the following reactions are observed.
b. Behavior toward Inorganic Reagents. (1) Oxygen. The most arresting
property of many organometallic compounds is their strongly exothermic
oxidation in air. Alkyls of alkali, alkaline earth, and boron families usually
generate sufficient heat in oxidizing to catch fire. As illustrated for the dimer
of triethylaluminum controlled oxidation in the absence of moisture; however, leads to the formation of an alkoxide:
[iu(C~H&12 302 -+ [A~(OCZHS)& heat.
(1)
66
JOHN EISCH A N D IIENItY GILMAN
(2) Hydrogoti. Hydrogen, on the other hand, can cleave many carbonmetal bonds smoothly with the formatiori of the corresponding hydrocarbon and mctnl hydride. I n a remarkable example of such hydrogenolysis
phenylpotassium has been found to be cleavcd by hydrogen a t ordinary
temperatures in the ttbscnce of pressure or catalysts (45):
CaH&
+ Hz
C& f KB.
--*
(2)
The synthetic utility of hydrogenolysis in organometallic synthesis recently

has been re-emphasized in the work of Ziegler and Koster with aluminum
and boron alkyls (Sections 1I.A and 1V.C).
(3) Water. The sensitivity of orgmometallic compounds to hydrolytic
cleavage varies from the spectacular to the mild: alkyls of Groups IA, IIA,
IIIA react with water explosively to yield the corresponding hydrocarbon
and metallic hydroxide :
Alkyls of mercury, silicon, tin, among others, show a slow cleavage in

aqueous medium and usually require an acidic or basic medium to effect
complet,ecleavage. 111any event, such hydrolytic cleavages furnish a general
method of analysis of organometallic compounds. If the group R in R,M
gives rise to ti volatile hydrocarbon, RH, this may be quantittttively cstimated as a check on either the purity of a known organometallic sample or
the riurnber of carbon-metal bonds per formula in a new substance. Moreover, the resulting aqueous solution can be analyzed to determine the
amount of metal cation formed.
(4) Mineral Acids. Similar to water, almost any desired inorganic acid
can effect full or partial cleavage of several carbon-metal bonds in an
orgtinornetallic compound. The most useful synthetically are, however, the
gaseous hydrogen halides dissolved in inert solvents such as benzene and
chloroform. Again, with very reactive organometallic compounds the
reaction is completed rapidly:
RzMg
+ ZTIX
2RH f MgX2.
(6)
The rextion may be viewed as an analog of the reaction of a stronger acid

with the salt of a weaker acid. Here the stronger acid, HX, liberates the
weaker acid, RH. If the original organometallic system is moderately reactive and has several carbon-metal bonds, conditions can usually be found
which lead predominantly to a stepwise cleavage of these bonds. An
illustration is the preparation of either triphenyltin halide or diphenyltin
67
dihalide by treating tetraphenyltin in chloroform solution with varying

amounts of hydrogen halide gas:
(CsHd4Sn HX
(C6HdaSnX -k CsH6,
(CsHa)&
2HX ----t (CsHa)zSn& -k 2Cs&
-+
(7)
(8)
(5) Halogens. The foregoing comments concerning the selective cleavage of certain carbon-metal bonds in a molecule are also valid for the
behavior toward molecular halogen. A favorite reagent for converting fully
alkylated metal derivatives into alkylmetal halides is iodine. For a metal
bearing different organic radicals there is often a preference as t o which
group is cleaved. For example, a dialkyldiaryllead compound will lose
preferentially aromatic groups upon treatment with iodine :
RzArzPb
+ 212
RzPbIz 2ArI.
(9)
(6) Metals. The behavior of metals towards organometallic compounds

depends on whether the metal alkyl is simple or mixed. By a simple organometallic compound is meant one bearing only R groups on the metal
(R,AI) ; by a mixed type is understood one having both R and acid-radical
or X groups on the metal (R2AICI). The effect of an active metal on mixed
organometallic compounds is usually the extraction of the X group and the
formation of the simple organometallic compound :
+
+
+
+
3R2AICl 3Na -+2R3A1 3NaCl

rRGeCl zK + (RGe), sKCI.
+ Al,
(10)
(11)
The simple organometallic compounds usually react with metals by displacement, when the more reactive can displace the less reactive metal
from the compound. The most important example of this is the utilization
of organomercury compounds to obtain the more reactive metal alkyls of
such metals as sodium, lithium, and aluminum:
RzHg
+ sNa
--f
2RNa
+ Na-amalgam.
(12)
Although the displacement power of a metal approximately parallels its

oxidation potential in aqueous solution, there are indications that this is
not always the case.
(7) Metal Halides and Other Salts. The more reactive organometallic
types will exchange their alkyl groups for the halogen or acid anion of the
salt of some less reactive metal or metalloid. Thus the reactive lithium and
magnesium (Grignard) alkyls can form metal alkyls from many other metal
salts. For example,
SnC14
+ 4RMgX
--f
R4Sn(RaSnCl,R2SnC12, RSnC13)
+ 4MgXC1.
(13)
(8) Lewis Bases. The unfilled electronic orbitals in the metal atoms of
many organometallic compounds favor complex formation by acceptance
68
JOHN EISCH A N D IWNRY QILMAN
of electronic pairs from suitable donors or Lewis bases. A wide variety of

dvnors including ammonia, amines, ethers, hydride ions, halide ions, arid
carbanioris form strong complexes with the alkyl derivatives of Groups
1-111. The driving force toward complex formation is so pronounced in
Group IIIA alkyls that they complex with themselves to form bridge dimers
(except boron).
c. Behavior toward Organic Reagents. Although not of immediate concern to the inorganic chemist the reactivity of organometallic compounds
toward organic systems further delineates the chemical character of this
bond and may suggest andogous applications in inorganic chemistry.
(1) Addition to Multiple bonds. Organic compounds containing carbon
bonded doubly or triply either to another carbon atom or a nonmetal such
as 0 or N generally add organometallic reagents. It seems logical to include
the inorganic reagent, carbon dioxide, under this heading.
(C&)&=CHz
+ It-Li
+ (CeHs)nCLi-CHzR.
(14)
R
H
It -C=O
+ R-Mg-X
O=C=O
+ R-Nrt
C--0-MgX.
(15)
+ R-C-0-Na.
(16)
8,
The facility with which organometallic reagents add to unsaturated systerns varies so noticeably, that a useful classification of organometallic
reactivity can be based on the ease of addition to certain unsaturated systems. Highly reactive types (It-Na) will react with both olefinic and carbony1 linkages, moderately reactive types (R-Mg-X)
will react only
with the carbonyl linkage, and unreactive types (R2-Hg) will react with
neither. It should be borne in mind, however, that solvent and other experimental conditions can influence the reactivity of a given organometallic
type (compare Section IV.A.1).
A sensitive, qualitative color test (50a) for R-M types of a t least
moderate reactivity has been devised based upon the addition of R-M t o
the carbonyl group of Michlers ketone. This test is very useful in monitoring the formation or consumption of organometallic reagents.
(2) Metalation. The exchange occurring between R-M and organic
compounds containing an acidic hydrogen bonded to carbon is termed
metalation.
R-I
+ R-M
-+
R-H
+ R-M.
(17)
Due to the electronegativity difference between hydrogen and carbon the

hydrogen atoms in all organic compounds may be considered acidic to some
69
extent. Metalation of an organic compouiid hy RM, however, is more facile

when the carbon atom to be metalated is adjoined by more electronegative
groups, or when the carbon anion formed is stabilized by charge delocalization. The metalation of anisole in the ortho position by RLi (Formula I)
0-CH,
I
illustrates the former case, whereas the metalation of toluene (Formula
11) by phenylsodium shows the favorable influence of charge delocalization
(Formula 111).
CHzNaO
I
I1
111
The synthetic utility of metalating organic systems with organolithium

compounds has been reviewed recently (4913).
(3) Halogen-Metal Interconversion. Quite analogous to metalation is
the observed exchange between R-M and organic halides :
R-X
+ R-M
---t
R-M
+ R-X.
(19)
Just as aromatically bonded hydrogen is more acidic than aliphatically

bonded hydrogen, it is observed that aromatically bonded halogen is more
cationic than the halogen in alkyl halidrs. Thus aryl bromides react with
alkyllithium compounds to form aryllithium derivatives and alkyl bromides :
Ar-Br
+ R-Li
---f
Ar-Li
+ It--Rr.
(20)
The interested reader is referred to a recaent survey of this reaction for its
scope and limitations (69a).
3. Physical Properties
As previously mentioned almost every recognized type of bonding is

represented in the diversified types of organometallic compounds known.
Consequently, no extensive generalization of the physical properties of all
organometallic compounds obtains. There seem t o be three broad classes of
R,M as t o physical properties: ionic or saline, associated covalent, and
unassociated covalent. The existence of ion pairs in ionic R,M types and the
strong bonding existing between molecular units in associated covalent
types (as in dirnerized aluminum alkyls and polymeric beryllium dkyls)
70
render corlclation of physical properties difficult,.Only i r i Ihe third class,

uiitissociatcd covalent, RM types, are there readily understandable trends
in the phyHical properties within the family. This is true of the alkyls of
the carbon family ( C , Si, Ge, Sn, Pb). Mendeleeffs enviable success in
predicting accurately the boiling point and density of tetraethylgermane
some fifteen years before Winkler first prepared it was possible because
such alkyls are unassociated.
4. General Preparative Methods

The classical methods which have proved most serviceable for the
preparatiori of organornetallic compounds already have received some
indirect attention in Section I.B.2. However, for convenience the equatioris
are given here.
(a) lntersctiori of metal arid alkyl halide:
Mm
+ nR-X
+ R,MX,-,.
(21)
(b) Metal exchange:
M R-M -+ M R-M.
M ~nort)reactive than M
(c) Alkyl-halogen exchange:
+ MX, + nMX MR,(or MRnXm).

R-M more reactive than RnM
&-M
(23)
(d) Metal-hydrogen exchange (metalation)

R-M
+ R-I
-+ R-H
+ R-M.
R-Br
+ R-M.
(24)
(e) Htilogen-met:il exchange:

R-M
+ R-Br
---f
(25)
An extensive review of these and other specialized methods of preparing

organometallic compounds has appeared (70).
c. 1)EVELOPMXNT
O F ORGANOMXTALLTC CHEMISTRY
1. Signijicant Periods
In retrospect the history of organometallic chemistry appears to have

been molded largely by key discoveries in synthesis. New preparative techniques have given prominence successively to different metal alkyls. Each
class of metal alkyls in its turn has had its own impact upon various aspects
of pure and applied chemistry. Some of the more fruitful discoveries which
71
made new organometallic types readily accessible are: (a) Franklands (39)
discovery of zinc alkyls, 1849; (b) Dimroths (34) studies on aromatic
mercuration, 1898; (c) Grignards (55) preparation of orgarlomagnesium
halides, 1900; (d) Schlenks (106) researches on organoalkali compounds,
1914; (e) Zieglers (141) preparation of lithium alkyls, 1930; (f) the preparation of ferrocene (71, 86) as the first sandwich organometallic s-complex,
1951 ; arid (g) Zieglers (140) preparation of aluminum alkyls via olefin and
hydride, 1955.
2. Contributions to the General Advance of Chemistry
Interestingly enough, organometallic chemistry was born of the

researches of Bunsen (18a) and Frankland (1849) concerning the existence
of uncombined radicals or element imitators. In isolating cacodyl which was
thought to be (CH&As, Bunsen thought that he had obtained the first
lLfree radical. As mentioned in Section I.A.2, however, organometallic
chemistry has amply repaid valence theory this debt of parentage by
furnishing the structural chemist with new challenges and insights into the
nature of the chemical bond.
Some other contributions of organometallic compounds to fundamental
research are: (a) the detection of free alkyl radicals by the pyrolysis of lead
alkyls; (b) the classification of hydrocarbon acidity via organoalkali compounds; (c) the study of Lewis acid-base interactions with Group I11
alkyls; (d) the development of the concept of electron-deficient compounds
by the study of metal alkyls; (e) the discovery of stereospecific olefin polymerization; and (f) the investigation of nucleophilic additions to unsaturated organic compounds via reactive metal alkyls.
The importance of organometallic compounds in synthesis to both the
organic and inorganic chemist should be apparent from the discussion in
Section I.B.2. However, the sensitivity of many organometallic compounds
to solvolysis and oxidation, their poisonous and flammable character, the
necessity of utilizing large quantities of solvent and relatively expensive
starting materials, all have tended to suppress until recently their widespread industrial use. The production of tetraethyllead as a n antiknock
additive, of methylchlorosilanes as silicone intermediates, and of organomercury antiseptics may be cited as exceptions.
3. Bridge between Inorganic and Organic Chemistry
The study of the carbon-metal bond offers excellent opportunity for a

rapprochement of the inorganic and the organic chemist. The fertility of
an interdisciplinary field lies in the meeting of different viewpoints. Here
the extensive knowledge of the electronic and steric characteristics of
organic groups can he pooled with the known chemistry of the metals,
72
JOHN EISCH A N D HENRY GILMAN
thus leading to a morc intimate and broader understanding of the chemical

bond.
D. RENAISSANCE:
IN ORGANOMETALLIC
CHEMISTRY
AND ITS SOURCES
In recent years organornetallic compounds have ceased to be exclusively
the object of academic research. Almost every major industrial chemical
research laboratory in the world has initiated extensive research in organometallic chemistry. In addition to the general increase in chemical research
following World War I1 the prominence of this field can be owed to certain
pacesetting discoveries. A principal factor in awakening industrial interest
has been tho researches of Ziegler and co-workers with aluminum alkyls.
The realization that these alkyls with various co-catalysts can convert
olefins, dienes, and acetylenes into superior polymers and ring compounds
has been of first-rate importance to organic synthesis. The utility of these
same alkyls in the preparation of boron alkyls and hydrides, in the electroplating of aluminum, and the synthesis of tetraethyllead has stimulated
the imagination of the inorganic chemist.
A similar role has been played by Wittig and co-workers in extending
the horizons of aromatic organometallic chemistry. Investigation of the
complexes formed between aryl alkali compounds as phonyllithium and
Lcwis bases of type R,M has resulted in the synthesis of ylides of nitrogen
and phosphorus, metalloids in rare valence states, and organometallic
a-complexes of unique structure.
A thoroughly unexpected postwar breakthrough was the synthesis of the
first aromatic type of transition metal sandwich complex known as dicyclopentadienyliron (11) or ferrocene. Realization of this intriguing structure
has stimulated the synthesis of many similar types.
Two major fields of inorganic chemistry which have had a considerable
impact on recent organometallic chemistry are those of metal hydride and
metal carbonyl research. The former owes much of its present vigor to the
studies of Schlesingers group (69),whereas metal carbonyls have received
much attention from Hieber and co-workers (66).Schlesingers aluminum
hydride complexes proved very helpful in Zieglers study of aluminum
alkyls, and metal carbonyl chemistry has guided the development of the
field of transition metal *-complexes.
The amount of recent research in organometallic chemistry has gone up
exponentially and many researchers have contributed significantly. Space
does not permit consideration of each metal alkyl type, however. The following discussion will aim rather at emphasizing new preparative methods
of some generality, AS well as recent viewpoints on the structure and reactivity of carbon-metal bonds.
OIlGANOMETALLIC COMPOUNDS
73
11. New Preparative Techniques and Procedures
A. METALHYDRIDES
A N D OLEPINS
1. Lithium and Aluminum Hydrides
The addition of metal and metalloid hydrides to carbon-carbon double

bonds is not a new reaction, having been observed from time t o time with
silanes of the type R3SiH under free-radical conditions (42, 85) and with
boron hydrides (68). The versatility of such hydride-olefin interactions,
nevertheless, first became evident with the recent researches of Ziegler
with lithium and aluminum alkyls (139). The observation that attempted
distillation of ethyllithium led to decomposition into lithium hydride,
ethylene, and higher olefins prompted the following formulation of the
reaction course (see 18):
CH3-CHz-Li
F? CH2=CH2
LiH,
CH,CHELi+ CH3CH&H2CH2Li,
CH2=CH2
CHBCHZCH2CHZLiF? CH3CH2CH=CH2 LiH, etc.
(26)
As this implied that LiH can add reversibly to ethylene and thus effect its
catalytic conversion t o higher olefins, Ziegler was prompted to study the
behavior of alkali hydrides towards ethylene in hopes of achieving polymerization. However, tarry reaction products hindered its realization with
lithium hydride. Utilization of lithium aluminum hydride instead led to a
rapid conversion of ethylene into higher a-olefins. This in turn suggested
that perhaps aluminum alkyls and hydride were more active than lithium.
Consequently, the fundamental chemistry of aluminum alkyls and hydrides
was extensively studied. The fruits of this research to the field of organic
synthesis have been very abundant: the synthesis of higher olefins, fatty
alcohols, and desirable polymers such as high-density polyethylene and
stereospecific polymers of dienes.
Because of the widespread utility of aluminum alkyls preparative
methods have been developed which place these compounds among the
potentially cheapest organometallic reagents. Since aluminum hydride
adds to ethylene a t moderate temperatures,
(AlHa)"
+ 3nCHz=CH*
+ nAl(CH2CHJ3,
(27)
a simple, elegant synthesis of aluminum alkyls was envisioned: the formation of aluminum hydride in situ by the reaction of activated aluminum
and hydrogen gas and then subsequent introduction of the olefin:
A1
+ #Hz + 3 C H 2 4 H R -+ Al(CHZCH2R)a.
(28)
Two limitations on the synthesis are, first, that the olefins utilized must
have a terminal double bond (a-olefins) and, second, that the use of the
74
JOHN EISCII AND HENRY OILMAN
simplest olefin, ethylene, and other straight chain olefins in such a one-step
synthesis is not feasible. The reason is that a t the required temperature
(12O-150)triethylduminum reacts with ethylene to form higher aluminum
alkyls (see Sectlion IV.C.6). The latter so-called growth or Aufbau
reaction [Eq. (84)]is useful for producing higher olefins, but it must be
avoided in synthesizing triethylaluminum. Thus a useful two-step procedure is employed : first, triet,hylaluminum (prepared via conventional methods), aluminurn, arid hydrogen are interacted a t higher temperatures to
form diethylaluminum hydride:
2(CHaCH2)3AI
+ A1 + 3H2
3(CII&Hz)zAIH,
(29
then the ethylene is introduced a t a lower temperature (60-70), where the

undesired growth reaction is slow, but where diethylnluminum hydride
can still add smoothly to et,hylene (140):
(CHaCHZ)gAIH
+ CHz=CH2
-+
Al(CH&T,),.
(30)
Other interesting aspects of this new synthesis of aluminum alkyls stem

from thc fact that hydride additions to olefiris are reversible. In the first
place this means that favorably constituted aluminum alkyls should lose
olefin readily upon being heated. Such branched-chain alkyls as trii~obut~ylaluminum will cvolve isobutylene upon heating a t 150-180 :
IT
AI(CHy--C-CHa)3
CITS
A
-+ HAl(CHzCHCHs)z
CHa
+ CHFCCH~
CH,
The resulting diisobutylaluminum hydride is an excellent liquid reducing

agent toward unsaturated systems. Likewise, branched beryllium alkyls
such as diisopropyl and di-t-butyl derivatives decompose to give the
corresponding olefins and polymeric beryllium hydride (65, 59). Analogously, diethylmagnesium yields upon pyrolysis magnesium hydride (40).
The existence of an equilibrium between hydride and olefin on the one hand,
and aluminum alkyl on the other makes possible a transalkylation process
known as the displacement or Verdrtingung process. If one heats a n aluminum alkyl with a different olefin, one can expect the following:
Al(CH2-CH&)3 F! HAl(CHz-CHzR)z
CH-CHR
HAI(C1f~- CH2R)Z CHz=CHR + AI(CH2CH2R)f(CH,-CH2R)
1T
IIAl(CH2CHzR)(CHz-CH2R)
+ CH,=CHR,
(32)
etc.
The equilibrium point will depend on the temperature arid the nature of the
R and R groups. However, by using the added olefin (CH2=CHR) in
generous excess and/or by allowing the generated olefin (CH,=CHR) to
escape, one can effectively prepare new aluminum alkyls. Thus the heating
of triisobutylaluminum with decene-1 a t 125 leads to a practically qilaii-
75
titittive formation of tri-n-decylaluminurn, since the volatile isobutylene

escapes :
AI(CH&HCH,),
+ ~CHZ=CH--C~HI,
-+
Al(C1oHzt)a
+ 3CH2=CCHa t .
I
CH,
(33)
CH3
The generality of this transalkylation reaction is reflected in the successful

synthesis of higher organoboron (73) and organogallium alkyls (142) from
olefins with triisobutylborane and triisobutylgallium, respectively.
2. Boron Hydrides
The synthesis of trialkylboranes by addition of diborane (68) or aluminum borohydride (9) to olefins and other unsaturated systems has found
only limited application because of the hazards of working with boron
hydrides. The aforementioned discussion of aluminum alkyls, nevertheless, points up the potentialities of analogous reactions with boron. The
current intense interest in (exotic or high-energy fuels involving boron has
caused many researchers to refocus their attention on the addition of boron
hydrides to olefins. In place of utilizing free diborane, new procedures have
either used masked boron hydride or generated it in situ. Working in
Zieglers laboratories Koster has found that the transalkylation reaction
utilizing triisobutylborane will effect the addition of borane t o both terminal and internal double bonds as in a-methylstyrene and cyclohexene
(73):
B(CHZCHCH3)a
CHa
+ 3CsH,-C=CHz
J
--t
B(-CHz-CHCeH6)3
+ 3CHz=CCH,.
I
CH3
CHs
(34)
CH3
The tri-2-phenylpropylborane formed can be oxidized and hydrolyzed

to 2-phenylpropanol-1, thus representing [as with aluminum alkyls (IdO)]
an overall anti-Markownikoff hydration of a-olefins. Besides employing
trialkylboranes as a source of boron hydrides, similar additions have been
accomplished by heating olefins with boron hydride-amine complexes. As
the latter are not as oxygen-sensitive, they are easier to handle (73):
3CH*=CHR
+ BH3: NR,
B(CHz-CH2R)a
+ NR3.
(35)
A procedure for bringing about the interaction of olefins with boron hydrides generated in situ recently has been reported by Brown and Subba
Rao (lQ,15).This reaction which has been termed hydroboratioii is effected
by adding the olefin to a mixture of sodium borohydride and aluminum
chloride in diethylene glycol dimethyl ether. Other ethers (tetrahydrofuran) and other haIides (gallium and boron halides) are also effectual when
used with sodium borohydride. Added to the great convenience and high
yield (90%) accompanying this synthesis of trialkylboranes, the reaction
76
seems able to isomerizc olcfiiis with iiitcriial double bonds into tri-n-alkylboraim.2 Illustrations of this are the hydroborations of 3-hexene and a
mixture of isomeric decencs. In each case the principal product obtained by
oxidation and hydrolysis of the trialkylborane is the primary alcohol.
This implies that an equilibrium between boron hydride and olefin versus
the trialkylborane is set up. Stability considerations seem to lead to predominance of the tri-n-alkylborane:
CHnCH2CH=CHCH2CHs
+ R2BH F? RaBF? b B H + 2-hcxcne, ete.
(3G)
3. Hydrides of Group IV
As an extrapolation of the peroxide-catalyzed addition of trichlorosilane

t o olefins (38, 116) new aryl silanes and germanes of the type (CeH&MR
have been prepared by the addition of triphenylsilane and triphenylger2~
mane to olefins in the presence of benzoyl peroxide ( ~ $86).
+ CHFCH-R
(CsHs)rSiH
-+ (C6H6)sSiCH2CHZ-R.
(37)
There is indication that triphenylgermane is more reactive than triphenylsilane in this reaction. The germane hydride is able to add to the double
bond in triphenylallylgermane in high yield, whereas triphenylsilane will
not add under the same conditions (47).
Since triphenyltin hydride tends to interact with benzoyl peroxide to
yield triphenyltin benzoate and other products (44), a peroxide-catalyzed
addition of this hydride to olefins is overshadowed by decomposition products. However, polar unsaturated compounds such as acrylonitrile add
triphenyltin hydride readily via a supposed ionic mechanism (122,123). The
reaction offers a new pathway to functional organotin derivatives of high
biocidal activity:
(COHs)oSnH
+ CIIFCHCN
--$
(CsHs)8SnCH2CH2CN.
(88)
The possibility of synthesizing alkyl and alkenyl silanes from olefins

and acetylenes with silane, SiHd, has been investigated by White and
Rochow (128). With ethylene various ethyl derivatives of silane and disilane
were obtained under conditions suggestive of a free-radical mechanism :
4. Stereospeci5city of Hydride Addition

There is reason to believe that alkylaluminum and alkylboron hydrides
approach and add to the unsaturated linkage in a strictly cis-fashion. Al8 Attempts to prepare tri-t-butylborane and tri-seobutylborane from boron trichloride and the corrasponding Grignard reagent demonstrate the ease with which these
types rearrange: for example, tri-isebutylborane is formed instead of tri-t-butylborane
(6%.
77
though this is difficult to demonstrate with ordinary olefinic systems, the

cis-character of the addition has been proved with acetylenes. Thus diethylaluminum hydride adds smoothly to diphenylacetylene and diethylacetylene at room temperature to give pure cis-olefins upon hydrolysis
(133).If one accepts the unlikelihood of rearrangement during hydrolysis,
the addition must be formulated :
R
+ R-CrC--R
RZAlH
--+
R-A1
c=c
R
By the use of deuterated dialkylaluminum hydride for the addition and/or
heavy water for hydrolysis, specifically deuterated olefins may be prepared. Likewise, hydroboration of certain olefins has now been shown to
occur in a cis-fashion (16).
B. ALUMINUMALKYLSAS ALKYLATINGAGENTS
The availability of the aluminum alkyls in pure state makes their use
as alkylating agents for the synthesis of other organometallic compounds
quite feasible. One of the general approaches is the interaction of the metal
or metalloid halide with the trialkylaluminum (105):
3MXn
+ nR3AI
3MR,
+ nAlX3.
(41)
This method has the advantage over the use of ethereal solutions of magnesium and lithium alkyls in that the metal alkyls can be prepared etherfree. In certain cases, however, not all the alkyl groups of Rs& are available.
Thus in the preparation of gallium and indium alkyls the reaction proceeds
according to the following equation (142):
GaXa
+ 3AIR3 + GaR3 + 3AIRzX.
(421
In the synthesis of boron alkyls, on the other hand, the employment of

BFa allows the following stoichiometry (73) :
BF3
+ AlRI + BR3 + AlF,.
(43)
Another important point in this procedure is that occasionally the HA1Cl

formed can interfere with high yields. Although gallium bromide gives 85%
yields according to Eg. (42), gallium chloride yielded only 4 0 4 5 % of
triethylgallium. The decrease in yield is due to the binding of one-half of
the gallium chloride by the diethylaluminum chloride :
GaCls 3Al(CzH& -+ Ga(CzH&

3Al(CzH&C1
GaCla 3Al(CzH&Cl --+ Ga[Al(CzH&Cl&.
(44)
78
JOHN EISCH AND HENItY GILMAN
This hypothesis is made quite probable by the isolation of the complex

between gallium chloride and diethylaluminum chloride. Moreover, inclusion of potassium chloride in the reaction mixture to complex with the
dialkylaluminum halide and free the gallium chloride raises the yield to
90% of triethylgallium (142).
3KC1
+ Ga[Al(CzH&Cl&
c. METHYLENATION
VIA
GaCla
+ 3K[A1(C2H&C121.
(45)
D3AZOMETHANE
Recently, interest has been revived in the preparation of chloromethyl

derivatives of the more electronegative metals by the interaction of the
metal halide with diazomethane. Thus germanium tetrachloride upon
treatment with an ethereal diaeomethane solution and copper powder at
-GO" yields 94% of ehloromethylgermanium trichloride (110):
GcC14
+ CHzNz
cu
ClCH2GeCb
-60'
+ NZ1' .
Vigorous investigation, especially by Yakubovich and co-workers, indicates

that the scope of this reaction includes the metal halides of mercury, thallium, germanium, tin, arsenic, antimony, and bismuth. Metalloid halides of
silicon and phosphorus also react (107). The promise of the reaction lies
in the formation of organometallic compounds containing the potentially
reactive Cl-CH2grouping (109). Since the net effect is the insertion of
the methylene group, --CH2-,
into MX,, the reaction has been termed
methylenation. Another related reaction which allows the transformation
of certain hydrides into organometallic compounds has been discovered by
Lesbre (78), namely the reaction of diaeoalkanes with stannanes. Tri-nbutylsttlrinane yields methyltri-n-butyltin with diazomethane whereas
tri-n-butyltin chloride does not react (see 111) :
(C4H&SnH
+ CH2Nz
-+
(C,Hp)&KX18
+ Nz T .
(47)
Use of substituted diazoalkanes, such M B ethyl diaxowetate, leads to functional tin alkyls. The potential applications in preparing interesting alkylboranes from boron hydrides should be apparent.
U. ORGAX'OMETALLIC
U-COMPLEXES
As has been mentioned in Section I.A., most organometallic compounds have unfilled electronic orbitals and hence can display Lewis acid
behavior. In recent years Wittig (134) has devoted intensive study to the
preparative possibilities of complexes between different aryl organometallic
compounds. In most cases the bond between the organic moiety and metal
79
has v-character, being symmetrical to the bond axis. This type is exemplified
by complexes between phenyllithium arid organometallic acceptors such
as triphenylborane and diphenylberyllium :
C&Li
Li[B(C&.),].
B(C6H6)3 -+
(48)
Such complexes possess sharply reduced chemical reactivity and consequently they often tend to stabilize the valence state of the acceptor metallic atom. Lithium tetraphenylboroiiate requires heating in acid solution
in order to effect cleavage of the boron-carbon bonds and is quite stable in
air toward oxidation. The acceptance of the phenyl anion has satisfied the
electronic demands of boron. A direct preparation of analogous alkyl complexes has been realized by heating lithium aluminum hydride with ethylene
under pressure (139):
LiAlH,
+ ~CHFCIJZ
--$
Li[AI(CzH&].
(49)
The attainment of outermost filled s- and p-orbitals (complete octet) is

the limit of first period elements. With metalloids of higher principal
quantum number d-orbitals can also be filled. Pentaphenylantimoriy, tetraphenyltellurium, and tripheiiyliodine form complexes of decreasing stability
with phenyllithium:
LiCeH5
+ Sb(CaH&
+ Li[sb(Cd&)~].
(50)
Another system forming a-complexes involving d-orbitals seems to be

dicyclopentadienyldiphenyltitanium which forms complexes with one or
more equivalents of phenyllithium (118, 119) :
(C5HS)2Ti(CeHS)zLiC6H6+ Li[(c~H~)zTi(CsH5)3].
(51)
This also appears to be the case with the previously elusive triphenylchromium. Hein and Weiss (61) have isolated from the interaction of
chromium (111) chloride and an excess of phenyllithium a solid red substance which is the etherate of such a complex:
CrC&
+ 6LiCeH5
-+
Lia[Cr(CeH6)6]f 3LiCl.
(52)
No sandwich or r-complex seems to be involved. One might observe in

general that the isolation of organometallic compounds of low stability
may often succeed if complex formation with solvent, metallic salt, or an
RM type is provided for (See Section IV.A.3). For example, disubstituted
platinum (11) and palladium (11) derivatives have been prepared successfully as the bis(triethy1phosphine) complexes in the following reaction
(19, 20) :
2RLi
+ M[(C2H6)3P]&
+ M[(C~H~),PIZR,2LiX.
(53)
An intriguing new organometalloid complex type has resulted from

renewed attempts to prepare pentacovalent nitrogen derivatives. These
80
are the phosphorus and nitrogen ylides (134). If onc treats a tetramethylammonium salt with phenyllithium, one obtains a salt complex of lithium
halide with trimethylammonium methylide :
(N(CH&]@I8
+ LiCSHs
+ (CH&N-CH*.
LiI
+ CBHa.
(54)
Although this nitrogen ylide allows of only a betaine structure, the phosphorus analog may involve some Ir-double bonding:
This has led to interesting phosphorus-boron complexes. Pentaphenyl

phosphorus will donate a phenyl anion to triphenylborane:
P(C&)s
+ B(CaI%a)a+ [P(csHs)r]@[B(CeHs)rl@.
(56)
Analogously, interaction with triphenylphosphorus methylide leads to a

betaine complex :
E. ORGANOMETALLIC
COMPOUNDS
OF TRANSITION
METALSAND
METALLOIDS
As an extension of the previous discussion, it can be observed that difficultly preparable organometallic compounds can often be obtained by
careful choice of structural features. Besides utilizing the stabilizing influence of complexing agents the location of electronegative substituents on
the metal seems to stabilize the carbon-metal bond. The notion has been
advanced that the strong negative field of an oxygen or halogen atom decreases the promotion energy necessary for a given hybridization of tho
metals orbitals and consequently stabilizes that state. Herman and Nelson
(63)found, for example, that in attempting to synthesize ordinary titaniumcarbon bonds they could obtain an isolahle phenyltitariium (IV) triisopropoxide by interaction of phenyllithium with the titanium orthoester:
Ti(i-OCaHp)r LiCeHs
--j
( C ~ H K ) T ~ ( O C ~LiOC3H7.
H~)
(58)
Replacement of more than one isopropoxyl group led to very unstable

products; hence, the presence of these groups promotes the stability of the
phenyl-titanium bond. However, another factor not to be overlooked is
that the corresponding alkyltitanium trialkoxides were too unstable t o be
isolated. This may indicate the stabilizing influence of both the alkoxide
(Formula IV) and the phenyl group (Formula V) is not only due t o the
strong negative field of these groups, but also to *-electron donation
into titanium atoms unfilled orbitals.
81
R-0
0-R
IV
R-0
0-R
V
Whatever the explanation, the enhanced stability of organometallic

cornpourids bearing phenyl, chloro, or similar groups is often observed.
The long sought-after methyl derivatives of titanium have been prepared
recently by the interaction of trimethylaluminum with titanium tetrachloride. The CH3TiC13and (CH3),TiC12types thus isolated (37) are much
more stable than might be suspected. Again, dicyclopentadienyldiphenyltitanium results from treating dicyclopentadienyltitanium dihalide with
phenyllithium (118). The phenyl groups seem to be bonded in a u-fashion.
Other organometallic compounds of transition metals containing a n ordinary sigma carbon-metal bond seem to be alkylmanganese pentacarbonyls
(27), phenylvanadium bromide ( 8 ) , methylcobalttetracarbonyl (66), and
bcnzylchromium salts (2).
The organic moiety most suitable to establish a stable sigma carbonmetal bond in difficult cases seems t o be an electronegative radical such as
the phenyl or acetylenic group. Examples of new and interesting phenyl
organometallic compounds are triphenylchromium (64), pentaphenylantimony, tetraphenyltellurium, and triphenyliodine (134). Nickel and
other metals can be transformed into acetylides (89).
A class of organornetalloid compounds awakening a considerable amount
of current interest is the perfluoroalkyls. The high electronegativity of
fluorine confers interesting properties on these derivatives. Trisperfluorornethylstibine, for example, has sharply reduced donor properties and can
even behave as an acceptor towards pyridine. The general method of
synthesis of antimony or arsenic derivatives entails the interaction of the
metal with trifluoromethyl iodide t o give (CF&M, (CF&MI, and CF3MI2.
(29, 35,
Reaction of (CF3),MI with metals leads to (CF&M-M(CF&
58).
F. PREPARATION
OF VINYLORGANOMETALLIC
COMPOUNDS
IN
TETRAHYDROFURAN
The excellent donor or complexing characteristics of tetrahydrofuran
have popularized its use in organometallic chemistry. The finding of
Normant (91) that difficultly accessible vinyl and aryl Grignard reagents
can easily be prepared from the organic halide and magnesium in tetrahydrofuran, has made vinyl-type organometallic compounds just as available as other organornetallic types. I n typical recent utilizations, vinyl
82
JOHN ElSC11 A N D HENLLY GILMAN
c1eriv:ttives of mercury (6), germanium (108), tin ( l l d ) , and Group VB

elements (84)h i v e been prepztred by treating the respective rnctdlic halide
with vinylmagnesium halide in tetrahydrofuran. These vinyl derivatives
promise to be valuable in both theoretical and industrial application. The
situation of ,z ?r-bond in the vinyl group adjacent to the metal atommay
allow px -dr electronic interactions, whereas the industrial chemist envisions
such viiiyl derivatives as potentially polymerizable organic intermediates.
G. ALKALIMETALADDUCTS
WITH UNSATU~LATED
HYDHOCAHBONS
Schlenk's (106) classical work on the addition of alkali metals t o unsaturated carbon-carbon bonds has been the subject of recent re-evaluation. The highly colored solid obtained by Schlenk and co-workers by
shaking biphenyl with lithium, for example, was formulated as (Formula
VI). A 2 : 1 adduct of this kind is not compatible with the physicochemical
VI
VII
studies of Weissman, Lipkin, and Paul. By working in tetr,zhydrofuran

they obtained a green solution containing alkali metal and biphenyl in 1 :1
proportions. Coriductimetric and especially electron spin resonance results
are compatible with the free radical-anion formulation (Formula VII)
(7, 80, 127). Similar 1:1 or 2: 1 adducts of alkali metal and many aromatic
hydrocarbons arc now readily obtainable by simply stirring the hydrocarbon and metal in tetrahydrofuran. The pronounced paramagnetic
character of such solutions makes them electronically :malogous to solutions of alkali metals in liquid ammonia (67').Whether these adducts should
be considered as organometallic compounds is u moot question.
H.
ELECTROLYTIC PREPARATION OF
ORQANOMETALLIC COMPOUNDS
A recent procedure for the preparation of lead and other alkyls of potential industrial application is an electrolytic process utilizing an aluminum
cathode, a lead anode, and an electrolyte of N a F 2Al(CZH,)s. Passage of
current leads to a quantitative formation of tetraethyllead a t the anode
and the deposition of high-purity aluminum on the cathode. It has been
suggested thatj the purified aluminum obtained as a by-product may help
meet the rost of electrical current and raw materials (146):
A1(99%)
31%
+ 3 1 1 ~+ 3CzH4
--*
+ 2[AI(C,ITs)sl?NnF'
Al(C2Hs)a.
electrolysis
+ 41\1(99.95%)
(59)
+ 3Pb(C213~)~+ 2NnF.
83
Ill. Structure and Bonding in Organometallic Compounds
A. AUTOCOMPLEXATION
Characteristic of most reactive RM compounds is their existence as two
or more R,M units closely associated with each other. The nature of these
metal alkyl complexes has been extensively studied via X-ray crystal
analysis by Rundle and co-workers, who showed that the a-carbon atom
of the alkyl groups forms a bridge between metal atoms in different RM
units. Thus, the dimer of trimethylaluminum may be viewed as
/
.,
CR,
(CHdzAI
Al(CH3)z.
CH,
Steric interactions explain why the degree of association of R3A1 decreases

in the order: methyl > ethyl > isopropyl (103). To bring such unique
bonding in consonance with modern electronic theory the electron pair of
the bridge bond is considered to be delocalized in a three-center (AI-C-Al)
molecular orbital. Ilundle notes that such electron-deficient compounds
(that is, metal having more low-energy orbitals than valence electrons)
will assume a configuration in which all low-energy orbitals may in some
way be involved in the bonding (103). This principle rationalizes the
behavior of polymeric dimethylberyllium (114), tetrameric tetramethylplatinum (103), and tetrameric trimethylindium (1).
Lithium alkyls in ether or benzene show a mean degree of association of
from three t o seven, whereas phenyl- and benzyllithium are dimeric in
ether (17, 1.95). The lower degree of association in ether may stem from
etherate formation. The structure of these auto-complexes may be analogous to that of beryllium and aluminum alkyls, or perhaps a lithium atom
acts as a Lewis acid; that is, Li@[Li(CsHa)2]. Wittig (135)favors this formulation over a phenyl bridging scheme.
The structure of the most commonly encountered organometallic,
namely the Grignard reagent, has been under dispute since its discovery.
Recent investigations confirm that RMgX is dimerized and solvated in
ether (113). Bridging by halogen atoms seems to be responsible for the
observed association, although alkyl bridges cannot be excluded. Some
attempt has been made to correlate the degree of association with the nature
of different alkyl groups and halogen in RMgX, but no clear pattern
emerges.
Recently, the equilibrium postulated by Schlenk to exist between
various species in the Grignard reagent :
A
RzMg. MgBrz F? l&Mg
+ MgBrz @B 2RMgBr,
(W
84
JOIIN EISCH A N D HENRY GILMAN
h:ts t m i i subjcoted to :t r:tdio:totive exchuiige tcst. Since 811 equimo1:v mixture of dietliylmagncsiurn and magnesium bromide dissolved in ether displttys the same kinetics toward 1-hexyne as the ordinarily prepared Grignard solutioii (Section IV.A.3), it seems reasonable to consider that both
involve the same or siinilar complexes. Now the mixing of radioactive
MgZsBr2with R2Mg and subsequent precipitation of MgBr2 with dioxane
gives almost complete (95%) recovery of MgZ8.If equilibrium B were involved, one would expect 507' recovery. A better formulation of the Grignard reagent thus appears to be RzMg MgBrz arid equilibrium A (31).
B. CARBON-METAL
BONDING
Current work promises to lead to a better understanding of the carbonmetal bond. The determiriation of the dissociation energies of the carbonmetal bonds of mercury (22, 36, 99),cadmium (77, 99),zinc (98), and the
alkyls of Group I11 (83, 120) and Group V elements (81, 82) promises to
offer an approach to estiinating intrinsic bond energies of the a-covalent
bonds. To do this, however, one will have to have some way of calculating
the valence state energies of the combined metal atoms. Thus, though the
boron-carbon bond has a greater mean dissociation energy than that of the
gallium-carbon bond, the valence state energy of sp2-hybridized gallium
would appear to be greater than that of boron. Consequently, the difference
in intrinsic borid energies may be still smaller or reversed in sign (83).
Another aspect of the carbon-metal bond which is receiving increased
attention is d o n o r interactions between adjacent unsaturated systems
and unfilled nietnl orbitals. Chatt and Williams (21) have sought to assess
the importance of dr-pr bonding in Group IVA elements by comparing the
dissociation constants of benzoic acid substituted in the para position by
R3M, where M = C , Si, Ge, Sn. Enhanced acidity of the carboxylic acid is
iriterpretcd 11s resulting from the withdrawal of r-electrons from the
the phenylene ring, thus facilitating proton loss (Formula VIII). Aryl-
VIII
IX
boraiies also evidence such ?r-interactions. Electric moment measurements

of aryldichloroboranes in solution (28) give values compatible with resonance structures such as Formula IX.
Another manner in which intramolecular electron donation can help to
fill the available orbitals of the metal atom is via a cyclic u-bond between
85
the metal atom and some function having an unshared pair. Such derivatives have an enhanced chemical stability. P-Aminoethyl diphenylborinate
(Formula X) (79) and 1-(4-ethoxybutyl)diethylaluminum (Formula XI)
(3) exemplify this interesting type of bond.
XI
C. METAL-METAL
BONDING
A word should be inserted about those organometallic compounds
thought to contain metal-metal bonding. This field may well receive considerable attention in future research. Various compounds of the type
R3-M-M-R3
have been prepared involving elements of Group I V (10,
12,13,46,95). Although susceptible to cleavage by sodium-potassium alloy,
they do not seem to dissociate into free radicals except when M is carbon
and R is an aryl radical. In Group I11 alkyls there is some evidence that
the colorless dianion exists,
[(CH3)3Ga-G~(CHa)31ee,
as does the sodium triphenylborarie adduct,

[(CsHs),--B--B-(CsHs)rl@
( 8 4*
The existence of hexamethyldiplatinum, (CH3)sPt-Pt(CHa)a (49) and the

possible existence of R-Hg-Hg-R
(64) are further instances. (Compare
ref. 103 for possible metal-metal bonding in associated alkyls.)
IV. Reactivity and Reaction Pathways of Organometallic Compounds
A. STRUCTURE
AND REACTIVITY
1. Nature of the Metal
It seems fair to state that the nature of the metal is more important than
the R group in determining the gross behavior of R,M types. Relative
reactivities of organometallic compounds within a given family or horizontal period of the periodic chart have been estimated by their rate of addition
to unsaturated compounds (for example, benzonitrile). The resulting empirical correlations show that R,M types of the A family metals in Groups
86
JOHN EISCIi A N D HENRY GILMAN
IA and IIA increase in reactivity with increasing atomic weight. Conversely, R,M compounds of the B family metals of the same group decrease
in reactivity. Thus in Groups IA and IIA, RCs and RnBa compounds are
the most reactive whereas in Groups IR and IIB, RCu and RzZn have the
greatest reactivity. In a given period, however, the reactivity seems to
decrease with atomic number increase; hence, the order, RNa > R,Mg
> RIAL These rules covering the most typical organometallic types can
be supplemented by generalizations for Groups 111-VII (43). Since the
reactivity of R,M types increases with increasing metallic character of M,
broadly speaking, the reactivity of R,M varies somewhat inversely to the
electronegativity index of M (101). Indeed, there is a tendency to readjust
electronegativity values of metals by reason of the chemical behavior of
their organometallic compound^.^ It should be remembered, however, that
these reactivity correlations are drawn largely from the behavior of R,M
in ether solution containing metal halides. Since Group I11 alkyls form very
stable complexes with both ether and metal halides, possibly their observed
lower order of reactivity compared to alkyls of Group I and I1 cannot be
related to the intrinsic reactivity of their carbon-metal bond. It may simply
exprcss the relative stability and inertness of R3M complexes. The great
reactivity of ether-free RAI discussed in Sections II.A.1 and IV.C.6 points
out the need for further reactivity evaluations.
An additional hindrance to relating the reactivity of R,M types to the
electronic configuration of the metal involved is the tendency for extensive
intermolecular associations of organometallic compounds. The extent to
which alkyls of the boron family dimerize is a function of both the size and
electric field of the metal atom. The reactivity of monomeric triethylborane
versus that of dimeric triethylaluminum is therefore less readily analyzed
into electyonic effects.
2 . Nature of the Organic Moiety
With a given metal atom the reactivity of R,M is profoundly influenced

by the nature of the R group. A recent example is the addition of Grignard
reagents to the >C = N- linkage in benzophenone anil. The ally1 Grignard
reagent adds almost quantitatively, whereas the n-propyl analog gives no
discernible reaction in the same period of time (45). Empirical reactivity
series of various organomagnesium and other organometallic compounds
have been established (43).Such experimentally determined electronegativities of organic radicals do not distinguish the various structural factors
3The interaction of RLi with (CsH&MH (M = Si, Ge, Sn) is illustrative. While
the silicon and tin hydrides undergo some hydrogen-alkyl exchange to give (CeH&M-R,
the germanium analog undergoes metalation t o yield (GH&Ge-Li (&a). This supports
a higher electronegativity for Ge versus Si and Sn.
ORGANOMETAL1,IC COMPOIJNDS
87
involved. ~latioiwliz:Ltioi1of reactivity based upon the electroiiic :md steric

factors involved in the H group would be preferable. This cannot be done a t
present without somc ambiguity for the reasons discussed in the previous
section. However, some beginning can be made. In the first place, one can
envisage that steric factors could increase or decrease the reactivity of the
carbon-metal bond. Thus if R is bulky, the compound R-M may be less
associated as (R-M), and hence the monomeric R-M may react more
readily. Moreover, the repulsive forces existing between the metal atom aiid
a bulky organic group may weaken the carbon-metal bond and thereby
aid reaction. The greater reactivity of the 2,4,6-trimethylphenyl Grignard
reagent compared to that of the phenyl reagent versus benzonitrile is a
case in point:
R-Mg-Br
+ CeH5CSN -+ CeH&=N-MgBr.
(61)
Conversely, a decreased reactivity of R-M types involving bulky R groups

can be related to steric interference either with solvation by ether or with
the substrate molecule. Such complexation seems to be important in aiding
heterolysis of the carbon-metal bond. Witness the failure of t-butylmagnesium halide to add to di-t-butyl ketone, although methylmagnesium halide
adds with no special difficulty.
In the second place, the electronic nature of the R group can modify
the polarity of the carbon-metal bond. The importance of ?r-bonding in the
structure of metal alkyls is discussed in Section 1II.B. As far as chemical
reactivity is concerned, however, a greater electron-attracting power of
R should enhance the polarity of the carbon-metal bond and thus increase
its chemical reactivity. In addition, if the R group is unsaturated, the
reactivity of the organometalh compound changes markedly. Situation of
the unsaturation on the carbon atom of C-M bond as in phenyl, vinyl, or
acetyleiiic 1%-M types usually results in some stabilization of the bond if
n-metal orbits1 overlap can occur:
-C=C-M
ts
-C-C=M.
@
That such bonding is operative in many carbon-metal bonds is indicated

tiy the superior stability of phenyl and acetylenic derivatives of the transition metals (Section 1I.E) and the reduced reactivity of vinylboranes
(92, 93). Trivinylborane, for example, has a reduced acceptor tendency towards ammonia and does not react with oxygen. Investigations have aimed
at looking for similar effects with vinylstannanes (112) and vinylgermanes
(108). Yet in alkynylgermanes and alkynylstannanes it appears that the
adjacent triple bond labilizes the carbon-metal bond. Both types undergo
88
facile hydrolysis, arid t)is(triethylgermyl)~cetyIenc!is rather sensitive to

oxygen (56, ,579.
If instead the H group contains unsaturation a, fl to the carbon-metal
bond, the lability of the bond is increased. With the possible exception of
ionic alkyls (already ionized) the ally1 and benzyl derivatives of a given
nietul show the greater reactivity. The serisitivity of tetrabenzyllead and
of tribcnzylbismuth (4) to oxidation exemplifies this increased reactivity.
One can rationalize this enhanced reactivity by the ease of cleaving the
bond :
- -C-CHn-M
-B
- =C-CH,@
+Ma.
(62)
Whether the cleavage is heterolytic, as above, or homolytic, the allylie-type

anion or radical would be stabilized by Ir-electron delocalization.
3. Role of the Reaction Medium
The importarice of solvent in the reactions of the polar carbon-metal

bond has aIready been adverted to. Evidence indicates that reactions of
lithium and magnesium alkyls go more rapidly in ether than in benzene.
The acceleration seems to be due to more extensive complexation of ether
molecules with the metal moiety which thereby facilitates heterolysis of
the carbon-metal bond. On the other hand, solvation seems to aid the formation of many difficult-to-obtain organometallic compounds. It would
appear that such solvent complexation stabilizes certain valence states of
metals and hinders homolytic decomposition. The recent widespread introduction of solvents more basic than diethyl ether, namely, tetrahydrofuran
(THF) and glycol ethers (10, 91, 100, l o g ) , has played a significant part in
new synthetic schemes. Herwig and Zeiss (64) have obtained the longsought triphenylchromiurn (111) from chromium (111) salts and phenyl
Giignard in THF, whereas the saue reaction in dicthyl ether leads to aromatic ?r-complexes of chromium, Heins polyphenylchromiums (60, 138).
Besides facilitating such displacement reactions THF and other solvents
aid the reaction of metals with organic compounds. The preparation of
unsaturated hydrocarbon-metal adducts is immeasurably more rapid in
such ethers. Thus, although lithium and biphenyl react very slowly in
diethyl ether to form a 1:1adduct (several days), the same reaction in THF
proceeds momentarily (compare Section 1I.G).
B. CLEAVAGE
OF
THE
CARBON-METAL
BOND
Highly ionic carbon-metal bonds such as presumably exist in organosodium compounds consist of carbanion-sodium cation pairs and the reac-
89
tioii with the substrate niolec.ulc inay be viewed a s a nucleophilic attack
-Ce
+ s+A--B6-
--f
-C-A
\
+ BB or -C-A-BB.
/
(63)
A similar behavior is to be expected of the highly polar, covalent carbonmetal bond as occurs in lithium, magnesium, and aluminum alkyls. Here,
however, initial complexation with the substrate seems to precede the
heterolysis of the carbon-metal bond:
C-0
+ R-M
+ Ce-O-Me
Where the nature of the metal or experimental conditions render such

auxiliary complexation with either the substrate or the solvent less favorable, there is an increased tendency towards homolysis of the carbon-metal
bond. Treatment of orgaiiolithium and organomagnesium alkyls with transition metal halides leads to metal halide reduction and hydrocarbon, suggestive of the transitory formation of metal alkyls (48), decomposing
thereupon into a lower salt and free alkyl radicals :
MX,
+ RLi
-+
LiX
+ RMX,-l
+ MX,-I
+ R..
(65)
The detection of free radicals from such Grignard reactions using anthracene as a scavenger has been studied recently (90). The extremely unstable
silver alkyls resulting from the treatment of lead alkyls with silver nitrate
decompose rather easily to yield alkyl radicals (compare behavior of isohut1-enylsilver) (61, 5.2).
C. REACTION
PATHWAYS
I. Oxidation
Depending upon the metal concerned, controlled oxidation of metal
alkyls may cease before all carbon-metal bonds are converted to C-0--M
groupings. It is interesting to compare the behavior of triethyl derivatives
of Group 111elements toward air oxidation: the aluminum analog oxidizes
completely to (C2H60)& the boron derivative may form either (C2B6)B(OCzH6)z or (C2H&B(OC2H6),and triethylgallium and triethylindium stop
after having formed (CZH5)~MOC2H6.
Presumably the presence of the
ethoxy group leads to stabilization towards further oxidation by inter-
90
JOHN EISCH AND HENRY GlLMAN
molecular complexation as dimer in the diethylgallium cthoxide (XI1)

@4,14.2):
XI1
and as polymer in the diethylindiuIri ethoxide (infusible solid) (104, 142) ;

or by intramolecular complexation in
0 @
RIB-OR
ct
hB-0
-R.
Although air oxidation may not oxidize all carbonmetal bonds in an

orgariometallic compound, reagents such as alkaline hydrogen peroxide
cleave such compounds completely to the alcohol and metal base :
R,,--M
11202
NaOH
nR-OH
+ M"+ + n0I-F.
(66)
Such reactions have great preparative significance for alcohol production

from olefiris (Section II.A.2).
The mechanism of the air oxidation is not completely clear. It is significompounds react
cant that excepting organornetalloids only those It,-M
whose metal has unfilled p-orbitals of the highest principal yuaritum number. Since the oxygen molecule is a diradical in the ground state, interaction
between R,,-M and oxygen may involve initially a one-electron bond between the metal's p-orbital and the oxygen with a subsequent 1,3-shift
(via a three-electron bridge bond'?) of the R group to yield the hydroperoxide salt: (compare Section IV.C.4).
R
Such a formulation seems more appealing than initial complexation, as

alkyls of boron and zinc oxidize readily despite their seeming reluctance to
form complexes with oxygen-type donors. That hydroperoxides are formed
arid that these give eventually the alkoxidcs has been demonstrated for
several metal alkyls. Recently the corresponding alkyl hydroperoxides
have been prepared in good yield by treating the Grignard reagents with
oxygen-saturated solvents a t - 70" (124, 225). Similarly peroxides of boron
91
(96) arid cadmium alkyls (30) have been isolated by controlled oxidation.
The rapidity of such oxidations makes a radical chain process attractive,
but neither aldehydes nor aromatic amines appear to interfere with the
reaction (125). Thus the formulation given in Eq. 67 or an analogous electron-transfer process would seem more acceptable.
a
2. Hydride Formatiioiz
The formation of metal hydrides from organonietallic compounds is an

interesting transformation attracting more and more attention. There are
several ways in which this conversion can be brought about. The first
method, that of hydrogenolysis via molecular hydrogen, has been mentioned in Section I.B.2. Within the last few years this method has gained
new prominence as a feasible preparation of boron hydrides. It has been
found that trialkylboranes, prepared readily from boron trifluoride and
aluminum alkyls, can be cleaved by hydrogen under pressure a t 200" in
the presence of amines, to give BH,-amine complexes. Such complexes
can be converted into alkali borohpdrides by the addition of the desired
alkali hydride (74):
H3B:NR3
100"
+ NaH -+ether
NaBHd
+ NRI.
Similarly, the useful Iiquid reducing agent, diethylaluminum hydride, can

be readily made by heating triethylaluminum with hydrogen under
pressure :
Al(CzHs)a
+ Hz
-+
HAl(C2Ha)Z
+ CzHe..
(69)
Almost nothing is known about the effect of catalysts on these hydrogenolyses, but the known behavior of such metal alkyls suggests that at higher
temperatures an equilibrium obtains between the olefin and metal hydride
(Section II.A.l):
R
H H
R-C-C-M+
H H
\
H/
C=C
+MH.
\H
The olefin may be reduced subsequently to the alkane, leaving the metal
hydrideS4That such an equilibrium is likely is strongly supported by the
products of the thermal decomposition of triisobutylgallium. At 160"
this compound decomposes to give gallium metal, hydrogen, and isobutylene
in nearly a 1:1.5:3.0 ratio. Very little isobutane is found. This suggests
Alternatively, onc can envisage a direct hydrogenolysis of the carbon-metal bond
to form the alkane and the metal hydride. The driving force would be the liberation of
the weaker acid (RH) by the stronger acid (H2).
92
JOHN EISCH AND HENItY GILMAN
an equilibrium involving unstable gallium hydride, which at this temperature reverts to the elements (142).
Another method of preparing metal hydrides from organometallic
compounds is by means of the following exchange reaction:
R-M
+ M'-H
-+ R-M'
+ M-H.
(71)
A recent example is the preparation of a magnesium hydride, HMgX, by

treating the Grignard reagent with diborane (129):
+ BzI&
GR-Mg-X
-+
GHMgX
+ 2BRs.
(72)
The method seems to offer promise of alkylating certain boron hydrides

(compare 1.26).In addition, hydrides of tin and silicon seem to undergo an
analogous reaction. Thus it has been observed that triphenyltin hydride
reacts with methyllithium to yield mainly lithium hydride and triphenylmethyltin (60).
A third procedure follows from the above and subsequent discussions
and that is the controlled thermal elimination of olefin from It--M to yield
metal hydrides where the latter are sufficiently stable:
H I3
R-
&I - 4 -'t M
-+
MH
+ R-CH=CHz.
(73)
H H
Diisobutylaluminum hydride can be prepared conveniently by heating the

triisobutylalurninum between 150" and 180" to remove one equivalent of
isobutylene (compare Section II.A.1).
3. Cleuvuge by Active Hydrogen Compounds
The negatively polarized carbon atom in organometallia cornpounds

formally can be considered as a carbanion, Re, either actually (RNa)
or potentially (ILAl). Since the corresponding hydrocarbon, RH, is usually
a very weak acid, the carbanion, Re,is a very strong base. Consequently,
most H-A compounds, where A is an electronegative element or moiety,
will give up a proton to Re and thus cleave the carbon-metal bond:
R-M
+ H-A
+ R-H
+ M-A.
(74)
Cleavage reactions of R-M by means of water and acids cited in Section I.B.2 are of this general type. Other suitable H-A species are NH1,
RNH2, R2NH, ROH and acidic hydrocarbons such as terminal alkynes
(It-C-C-H)
and cyclopentndiene types. Although such reactions have
as their driving force the generation of the weaker acid, the mechanism of
cleavage of It-M by active hydrogen compounds is not fully elucidated
at the present. There are certain recognizable similarities between this
93
cleavage reaction and orgaaliomet:tllic oxidation. Again it is seeii that H-M

compounds susceptible to facile cleavage possess unfilled low-lying p-orbitals on the metal atom. I t would seem that initial attack on R-M is
complexation by H-A if the A moiety possesses an unshared electron pair.
Separation of R-H may then occur by a 1,3 shift. The hydrolysis of RJI
can be represented schematically:
R
[RI-AI]~
+ 2HOH
\ /
2Al:O-H
H
2R-H
+ 2RzA10H.
(75)
The plausibility of this formulation is enhanced by the known resistance of

trialkylboranes to hydrolysis (121). This stability would seem understandable in terms of steric shielding of the relatively small boron atom by the
alkyl groups to approach of water molecules. Alcoholysis of Group I11
alkyls is of varying facility. Contrasted with inertness of RIB, the aluminum
alkyls undergo rupture of all three bonds in alcohol:
RaAl
+ 3R-OH
-+ 3RH
+ 3(RO)sAl,
(76)
but gallium (142) and indium alkyls (104) allow cleavage of only one carbon-metal bond, leading to RzMOR. As both hydrolysis and alcoholysis
involve the rupture of oxygen-hydrogen bonds, some researchers have
sought to learn whether this bond is broken in the rate-determining step
by employing deuterated water and alcohol. The deuterated compound
should react more slowly if the hydrogen-oxygen bond is ruptured in the
rate-determining step. One study of the hydrolysis of methylmagnesium
iodide showed no apparent preference for breaking 0-H over 0-D bonds in
the reaction; the methane obtained contained deuterium in the same pro.
of certain
portion as deuterium was present in the water ( 9 1 ~ )Alcoholysis
organolithium and Grignard reagents in deuterated methanol showed a
slight acceleration over ordinary methanol. This unusual isotope effect
might suggest a stronger oxygen-hydrogen bond in the transition state over
that in the reactants, but the cause is not yet certain (130). Since both the
RLi and RMgX compounds prepared in ether from the metal and RX
contiin metal salts, and are themselves extensively solvated and associated,
the system is perhaps too complex to permit unambiguous conclusions
to be drawn.
The interaction of 1-alkynes and alkyl Grignard reagents has been
the subject of much recent attention (32, 33, 136). By measurement of
the rate of evolution of RH the reactivity of RMgX towards 1-hexyne was
evaluated :
R-Mg-X
+ c 4 H g c ~ C - H + CdHgCEC-Mg-X
+ RH.
(77)
94
JOHN ElSCH A N D HENRY GILMAN
The Grigiard reagents showed the following order of incrc:tsing reactivity :

methyl < n-propyl < ethyl < iso-propyl < allyl. Although the electrolysis decomposition pot,entials of these Grigriard reagents increase in the inverse order, the complex character of the organomagnesium halides in
ether prevent,s any simple mechanistic interpretation of this correlation
(32). It has been proposed that the reactivity of RMgX is proportional to
the number of /3-hydrogens in R. However, the reactivity differences might
be due to varying degrees of autocomplexation of the Grignard reagents
(compare Sectiori IV.A.1). Experiments with l-hexyne-dl showed that the
rate of cleavage was decreased by a factor of 4.3 compared to ordinary
1-hexyne. 111 the light of previous discussion this would suggest that the
C-H bond is ruptured in the rate-determining ptep (33).
4. Lewis Acid Character

Much recent investigation has been devoted to evaluating the tendency
of organometallic compounds to act as electron pair acceptors (116). The
possibilities of oarbanions acting as donors is considered in Section I1 .D
dealing with org:wometallic o-complexes. Three main types of donors have
been considered in these studies: first, amines and their Group V analogs;
second, ethers and their Group VI analogs; and third, halide ions. Although
Groups IA, IIA, and IIIA mctnl dkyls form Lewis salts with the aforementioned biiscs, the stability of such salts with metal alkyls of the same
period increases with the group number: thus, the stability of R,M C OR2
seems to increase in the order, RNa < R2Mg < R,Al. Such a gradation
is reasonable in terms of tin increasing nuclear charge in M (for elements of
the same principal quantum number) causing an enhanccd elcctronegativity, arid consequently greater acceptor tendencies in R-M. The variation
of acceptor propertties for metal alkyls of a given periodic family has been
cxurnined most thoroughly for Group IIIA elements, both by comparing
the heats of dissociation of R,M t NR, complexes, as Coates has done
(26), and by correlating dipole moments of solutions of R3M in donor solveiits, as carried out by Strohmeier (117). Both methods give the following
sequence of decreasing acceptor character: B < A1 > Ga > I n > T1. The
exceptional behavior of boron alkyls versus trialkylamine, that is, the formation of no measurably stable complex, is attributed to steric hindrance
by the neighboring alkyl groups in establishing a stable boron-nitrogen
bond (F-strain). The observed variation from aluminum alkyls through
thallium dkyls can be owed to the increasing atomic radius of M in R,M
causing a decreasc i n the bond strength of M-N via the shielding effect of
uridcrlyirig filled electron levels. As the effective nuclear charge of M is
weakened, the bond streiigth of the Lewis salt decreases. A similar variation in acceptor strength has been noted with the metal alkyls of Group
11: Be < Mg > Zn > Cd > Hg.
95
To a first approximation these trends seem to carry over t o ether and

halide complexes also. For example, the lower aluminum alkyls form very
stable monoetherates, R,,AI 4- OR2 (5), whereas gallium and especially
indium alkyls can be prepared in ether solution and isolated free of ether
(104). A striking illustration of this difference in complexation tendencies
among family members is the behavior of Group I11 alkyls towards alkali
fluorides : triethylaluminum complexes with either sodium or potassium
fluoride, the corresponding gallium alkyl complexes only with potassium
fluoride, and t,he indium alkyl complexes with ricither fluoride. This sharp
distinction has interesting applications in separating these metal alkyls
from each other (142).
The formation of such Lewis complexes may find increasing application
in stereochemical studies. Two cases in point might be cited. First, etherate
formation allows the isolation of the optically active tri-2-methylbutylaluminum etherate (97). This compound is stable to racemization up to
100". Second, diphenylpalladium (11) has recently been prepared as its
bis-triethylphosphine complex in both a cis- and a trans-planar complex
form (20). The latter seems to represent the first example of geometrical
isomerism among organometallic compounds (Section 1I.D).
A final point concerning Lewis complex possibilities should be mentioned. The adducts formed between Group I11 organometallics and
alkali metals are somewhat remindful of Lewis salt formation except only
one electron is involved (compare oxidation of R-M, Section IV.C.1) :
R,B
+
+ Na' + [RaB-. Na].
(78)
However, recent re-examination of (C6H6)3BNa has shown it to be diamagnetic, suggesting a better formulation to involve a dimer [R.3B-BH.3] "
(23).
5. Thermal Dissociation
Itupture of the carbon-metal bond in R,M upon heating seems to

involve either homolytic cleavage leading to free-radical fragments, R' and
R,-l-M, or heterolysis giving rise t o metal hydride and olefin. The latter
reaction, representing the reversal of metal alkyl formation from olefins
and metal hydrides, takes place with the more reactive alkyls (compare
Sections II.A.1 and IV.C.2). Since it seems that metal hydrides add to
unsaturates in a cis fashion, one might expect elimination in the same sense,
rather than trans elimination [Eq. (73)], the less substituted olefin being
formed preferentially (compare behavior of sodium alkyls) (87, 88). This
suggests that methyl derivatives should be more stable, since the analogous
alkene would be the reactive methylene; methyllithium gives CH2Li2 a t
200-240" and LiC =CLi a t 420, whereas ethyllithium decomposes over
100" (139).This mode of deromposition is apparent in certain of the reac-
96
JOHN EIBCH AND HENRY GILMAN
tions of orgaiiornetallic compounds : first, one can sometimes eliminate

M-H from R-M by addition of a suitable uomplcxing agent [that is,
(CeHJaB].Wittig has developed the following olefin synthesis (13.4):
Second, certain ll-M compounds interact with carbonyl compounds to

give reduction products explained, formally a t least, by the carbonyl
compound removing M-H from H.-M (via XIII) :
XI11
The homolytiu thermal dissociation of R-M, previously used hy

Pttneth in generating and studying free alkyl radicals, occurs with the more
electronegative alkyls, such as those of mercury, lead, and the metalloids.
Currently there is much interest in evaluating carbon-metal bond dissociation energies from such pyrolyses (Section 1II.B).
In cases involving the pyrolyses of organometalloid compounds where
the carbon-metalloid bond has a high thermal stability, it has been found
that heterocyclic organic compounds of novel structure are formed. Presumably the pyrolysis of trimethylhorane leads to, among other products,
an eight-membered ring (Formula XIV) (55), and the heating of tetraInethylsilane yields besides a six-membered ring the interesting bicyclic
type (Formula XV) (41):
97
6. A d d i t i o n to U r m t u r u t e d Oryuriic Compounds
The truly significant advance in the interaction of R-M with unsaturated organic compounds in the last several years is admission of the
-C=Cand -C = C- linkages as ordinary functional groups in organometallic synthesis. The instances of the addition of organomagnesium
compounds to carbonyl, nitrile, or other heteroatom unsaturates of the
type M = N are countless, but similar reactions with the olefinic and
acetylenic groupings have been of an ad hoc nature (see 72). What seems
to have delayed recognition of the additive possibilities of R-M on olefins
is the traditional utilization of R-M in ether solution where etherate
formation depresses the reactivity of the carbon-metal bond. Postwar research by Ziegler and co-workers (140) with lithium and subsequently
aluminum alkyls has brought to light two main types of interactions, the
addition reaction [Eq. (Sl)] and the displacement reaction [Eq. (82)]:
H
C=C
H
H H
R-C=CI-I
+ R-M
H H
R-C-C-M
H H
H H
-+ R-C-C-M,
H H
H H
+ R-CzCH
H H
+ R-C-C-M.
H H
(82)
The utility of the displacement or (Verdrangung reaction in the synthesis

of other organometallic compounds is considered in Section 1I.A.1. However, in organic synthesis two courses of the addition reaction are possible:
either the olefin or acetylene reacts with a given carbon-metal linkage in a
ratio of 1:1, this adduct being worked up to give product; or the unsaturated compound reacts with the carbon-metal bond successively to give
long carbon chains. Illustrative is the behavior of (C2Hb)aI with an
a-olefin, ethylene. One reaction observed is simple addition in ratio of
C-A1 :olefin of 1 : 1:
(CzH,),AI
+ 3CHzzCHz
(~C4Hg)aAl.
(83)
However, the tri-n-butylalurnirium can react further with ethylene to

give higher aluminum alkyls to yield the net equation:
(CzH,),Al
(CZH4)a-CZH6
+ zCHZ=CH~+ AI--(CZH~)~-CZHL
\
z =u
+b +c
(84)
(CzHJc-CzHs
This (growth reaction is hindered by concurrently occurring olefin displacement reactions [Ey. (32)]. In essence, however, it presents a method
of building long carbon chains from ethylene by using the initial aluminum
alkyl as a framework. The exciting aspect is that introduction of certain
98
JOHN EISCII AND HENRY GILMAN
traiisitiori rnetd h:clides in such a mixture accelersttcs growth of the carboii

chain to an astounding rate, so that polyethylene polymers having average
molecular weights up to three million are obtained (140).
Extension of these observations to acetylenes and coiijugatcd dienes
has led to interesting carbon ring compounds such as hexasubstituted
benzenes from KC =CIt and cyclododecatriciic-1,5,9 from butadierie
(132).
V. Applications of Organornetallic Reagents in Research and Technology
A. SYNTHESIS
In the discussions of the latest advances in the preparation and reactions
of orgaiiometdlic corripouiids the import to industrial synthesis has been
alluded to. Development of feasible methods for the large-scale prepara t1011
of aluminum alkyls has placed in the hands of the industrial chemist a
tool for convcrting lower olefins into higher aluminum ulkyls (Scction
IV.C.G), or to higher fatty alcohols, RCH&H20H (Section II.A.2). Resides thus obtaining useful detergent and fatty acid intermediates from
olefins, cntalytic qiiantit,ies of aluminum or other metal nlkyls with R tmnsition metal salt co-catalyst convert olefins to polyalkylenes of superior
properties. High-density polyethylene, isotactic polypropylene, 1,2- and
1,4-polyalkadienes, arid synthetic natural rubber have resulted from such
metal alkyl-catalyzed polymerizations (compare 131).
The pr:ict,icality of preparing rrietal and metalloid hydrides in certain
cases from the readily ttccessikle metal alkyls has been receiving everiilcreasing attention. Eithei by elimiriating one or more moles of olefiri
from (RCH&H2),M thermally, or by cleaving the organometallic compourid with hydrogen, one can obtain pure hydrides or partially alkylated
hydrides. Diborane, obtained in the latter manner, is the keystone of syntheses leading to high-energy rocket fuels, whereas dialkylaluminum
hydrides have been suggested 2s liquid reducing agents for functions susceptible to LiNH4.
B. TECHNOLOGY
The applications of orgariometallic agents in technology show every
promise of increasing. The traditioiial use of tetraethyllead in motor fuels
may eventurlly he replaced by some other type. Tin alkyls have gained
some recent prominence both for the biocidal prorerties of certain tristlkyltin derivatives and for the ultravioletdisco1oratio:i protection offered vinyl
plastics by dibutyltin dilaurate. Also attention should be drawn to the possi-
99
bilities of metal plating and lead alkyl production by means of electrolytic

reactions employing alkali fluoride-aluminum alkyl complexes. Finally,
there has beeii some investigation (143) of' the possibility of metal purification by means of preparing the R-M compound with the crude metal and
then pyrolyzing the purified R-M compound to free the metal.
VI. Present Trends in Organometallic Research
Recent orgaiiomettlllic investigation hits witnessed a resurgence of one

of the most primary aspects of chemical research, namely the preparation
and isolation in pure state of many new and intriguing types of compounds.
This synthetic aspect seems destined to continue flourishing in the years
ahead. Albeit preparative techniques have been refined to permit the formation of some type of carbon-metal bond with almost any known metallic
element, our understanding of the factors governing thc structure and reactivity of organometallic compounds is largely of a qualitative, empirical
nature. It would seem therefore that future research in this field will aim
at elucidating the structure of associated metal alkyls not only in the pure
state but also in solution. Once a clearer picture is obtained of which species
predominate, say in a solution of R,M in ether, then the next step of interpreting the kinetic data of organometallic reactions in terms of known
R-M species becomes more feasible. Correlations of reactivity with variations of the alkyl and metal moieties in It-M then can be more directly
related t o electronic and steric factors. In the case of unassociated metal
alkyls, it would seem that there will be increasing effort expended in evaluating the electron-transmitting properties of carbon-metal bonds in terms
of inductive and niesomeric effects.
Orgaiiometallic chemistry, therefore, will he a topic of mutual interest
to the theoretical and to the synthetic chemist for many years to come.
LIST OF SYMBOLS
a pi interaction between a d- and a p- orbital.
d,-p,
6 - , S+ a partial separation of charges.
8,@ a full separation of charges.
t
a donor bond.
REFERENCES
1. Amma, E. L., and Rundle, R. E., J . Am. Chem. SOC.80, 4141 (1958).
2. Anet, F. A. L., and Leblanc, E., J . Am. Chem. SOC.79, 2650 (1957).
3. Bahr, G., and Mullcr, G. E., Chem. Ber. 88, 251 (1955).
4. B&hr,G., and Zoche, G., Chem. Bur. 88, 542 (1955); 90, 1176 (1957).
6. Baker, E. B., and Sisler, H. II., J . A7tt. Cherrt. Soc. 76, 4825, 5193 (1953).
6. Uartocha, B., and Stone, F. G. A., Z. Natwforsch. 13b, 347 (1958).
7 . de Boer, E., and Weissman, 6. I., J . Am. Chem. SOC.
80, 4549 (1958).
100
JOHN NISCH AND HXNRY GILMAN
8. Bonncr, E. P., Carrick, W. L., Kluibcr, R. W., Rcichle, W. T., Rugg, I?. M., Smith,
J. J. and Wartnian, L. H., Proc. 1SSrd Meeting Am. Chem. SOC.p. 19R (1958).
9. Brokaw, R. S., and Pease, R. N., J. Am. Chem. SOC.72, 3237 (1950).
10. Brook, A. G., J . Am. Chem. SOC.76, 77 (1954).
11. Brook, A. G., Cohen, H. L., and Wright, G. F., J. Org. Chem. 18, 447 (1953).
12. Brook, A. G., and Gilman, H., J . Am. Chem. SOC.76, 278 (1954).
13. Brook, A. G., Gilman, H., and Miller, L. S., J . Am. Chem. SOC.76, 4759 (1953).
14. Brown, H. C., and Subba Rao, n. C., J . Am. Chem. Sac. 78, 5694 (1956).
15. B r o w , H. C., and Subba Rao, B. C., J . Org. Chem. 22, 1135 (1856).
16. Brown, 1. C., and Zwcifel, G., J . Am. Chem. SOC.81, 247 (1959).
17. Brown, T. L., and Rogers, M. T., J . Am. Chem. SOC.79, 1859 (1957).
18. Bryce-Smith, D., J . Chem. SOC.p. 1712 (1955).
18a. Bunsen, R., Ann. 42, 14 (1842).
19. Calvin, G., and Coates, G. E., Chem. CPI Znd. p. 160 (1958).
20. Chatt, J., and Shaw, B. L., communication to Calvin and Coates (19).
81. Chatt, J., and Williams, A. A., J. Chem. Soc. p. 4403 (1954).
2% Chernick, C. L., Skinner, H. A., and Wadso, I., Trans. Furaduy SOC.62, 1088
(1956).
2s. Chu, T.L., J . Am. Chem. Sor. 76, 1730 (1953).
24. Coates, G. E., Organo-metallic Compounds, Methuen, London, 1956.
26. Coatcs, G. E., and Glockling, F., J . Chem. Soc. p. 2526 (1954).
26. Coates, G. E., and Whitcombe, R. A., J . Chem. SOC.p. 3351 (1956).
27. Cotton, F. A., Chem. Revs. 66, 551 (1955).
88. Curran, C., McCusker, P. A., and Mskowski, H. S., J . Am. Chem. Soc. 79, 5188
(1957).
2.9. Dale, J. W., Emeleus, H. J., Hasseldine, R. N., and Moss, J. H., J . Chern. Soc.
p. 3708 (1957).
SO. Davies, A. G., and Packer, J. E., Ghem. CPI Znd. p. 1177 (1958).
31. Dessy, R. E., Handler, G. S., Wotia, J. H., and Hollinpworth, C. A., J . a4m.
Chem. SOC.79, 3476 (1957).
39. Des~y,R. E., Hollingsworth, C. A., and Wotia, J. H., J . Am. Chem. SOC.77, 4410
(1955).
33. Dessy, R. E., Wotiz, J. H., and Hollingsworth, C. A., J . Am. Chem. SOC.70, 358
(1957).
34. Dimroth, O., Ber. 31, 2154 (1898).
36. Emcleus, H. J . , and Hasscldine, R. N., Science 117, 311 (1953).
36. Fairbrother, D. M., and Skinner, H. A., Trans. farm la^ SOC.62, 956 (1956).
S7. Farhwerkc, Horhst, Belgian Patent 553,477 (1956).
38. Fischcr, A. K., West, R. C., andRochow, E. G., J . Am. Chem. SOC.76,5878 (1954).
39. Frankland, E., Ann. 71, 171 (1849); 86, 347 (1853).
40. Freundlich, W., and Claudel, B., Bull. soe. chim. Prance p. 967 (1956).
41. Fritz, G., and Raabe, B., 2.anorg. u. allgem. Chem. 286, 149 (1950).
42. Gadsby, G. N., Research 3, 338 (1950).
43. Gilman, H., in Organic CherniBtry (H. Gilman, ed.), 2nd ed., Vol. 1,pp. 489-580.
Wiley, New York, 1943.
44. Gilman, I., and Eisch, J., J . Org. Chem. 20, 763 (1955).
45. Gilman, H., and Eisch, ,T., J . Am. Chem. SOC.79, 2150 (1957).
46. Gilnian, TI., and Gcrow, C. W., J . A7n. Chem. SOC.
77, 5509 (1955).
46a. Gilman, H., and Gerow, C. W., J . Am, Chem. Soc. 78, 5435 (1956).
47. Gilman, I., and Gwow, C. W., J. Am. Chem. SOC.
79, 342 (1957).
48. Gilman, H., Jones, R. G., and Woods, L. A., J . Am. Cham. SOC.76, 3615 (1954).
ORGANOMETALLIC COMPOUN 1)s
101
49. Gilman, H., Lichtenwaltw, M., and Bcnkcser, R. A., J . A m . Chem. SOC.76, 2063
(1953).
49a. Gilman, H., and Morton, J. W., i n Organic Reactions (R. Adam, ed.), Vol. 8,
Chapter 6. Wiley, New York, 1954.
50. Gilman, H., and Rosenberg, S. D., J . A m . Chem. SOC.76, 3592 (1953).
60a. Gilman, H., and Schulze, F., J . A m . Chem. SOC.47, 2002 (1925).
61. Glockling, F., J . Chem. SOC.p. 716 (1955).
62. Glockling, F., J . Chem. SOC.p. 3640 (1956).
63. Goubeau, J., and Epple, R., Chem. Ber. 90, 171 (1957).
64. Gowenlock, B. G., and Trotman, J., J . Chem. SOC.p. 2114 (1957).
66. Grignard, V., Compt. rend. acud. sci. 130, 1322 (1900).
66. Hartmann, H., and Ahrens, J. U., Angew. Chem. 70, 75 (1958).
67. Hartmann, H., and IIonig, H., A n g m . Chem. 70, 75 (1958).
68. Haszeldine, R. N., and West, B. O., J . Cheni. SOC.p. 3880 (1957).
69. Head, E. L., Holley, C. E., Jr., and Rabideau, S. W., J . A m . Chem. SOC.79, 3687
(1957).
60. Hein, F., and Eisfeld, K., 2.anorg. u. allgem. Chem. 292, 162 (1957).
61. Hein, F., and Weiss, R., 2. anorg. u. allgem. Chem. 296, 145 (1958).
62. Hennion, G. F., McCusker, P. A., Ashby, E. C., and Rutkowski, A. J., J . Am.
Chem. SOC.79, 5190 (1957).
63. Herman, D. F., and Nelson, W. K., J . A m . Chem. SOC.76, 3877 (1953).
64. Herwig, W., and Zeiss, H. H., J . A m . Chem. SOC.79, 6561 (1957).
65. Hieber, W., FIAT Reviews of German Science 1939-1940: Inorganic Chemistry,
Pt. 11, pp. 108-145. Dieterich, Wiesbaden, 1948.
66. Hieber, W., Vohler, O., and Braun, G., 2. Naturforsch. 13b, 192 (1958).
67. Holmes-Walker, W. A., and Ubbelohde, A. R., J . Chem. SOC.p. 720 (1954).
68. Hurd, D. T., J . A m . Chem. SOC.70, 2053 (1948).
69. Hurd, D. T., Introduction to the Chemistry of the Hydrides. Wiley, New York,
1952.
69a. Jones, R. G., and Gilman, H., i n Organic Reactions (R. Adams, ed.), Vol. 6,
Chapter 7. Wiley, New York, 1951.
70. Jones, R. G., and Gilman, H., Chem. Revs. 64, 835 (1954).
71. Kedy, T. J., and Pauson, P. L., Nature 168, 1039 (1951).
72. Kharasch, M. S., and Reinmuth, O., Grignard Reactions of Nonmetallic Substances, pp. 1384. Prentice-Hall, New York, 1954.
73. Koster, R., Angew. Chem. 68, 383 (1956).
74. Koster, R., Angew. Cliem. 69, 94 (1957).
75. Koster, R., and Ziegler, K., Angew. Chem. 69, 94 (1957).
76. Krause, E., and von Grosse, A., Die Chemie der metallorganischen Verbindungen. Horntracger, Berlin, 1937.
77. Laurie, C . M., and Long, L. H., Trans. Faraday SOC.63, 1431 (1957).
78. Lesbre, M., Bull, sor. chim. France p. 1204 (1957).
79. Letsinger, R. L., and Skoog, I., J . A m . Chem. SOC.77, 2491 (1955).
80. Lipkin, D., Paul, D. E., Townsend, J., and Weissman, S. I., Science 117, 534 (1953).
81. Long, L. I., J. Chern. Sac. p. 3410 (1956).
82. Long, L. H., Trans. Faraday SOC.63, 1606 (1957).
85. Long, L. H., and Sackman, J. F., Trans. Faraday SOC.64, 1797 (1958).
84. Maier, L., Seyferth, D., Stone, P. G. A., and Rochow, E. G., J . Am. Chem. Soc.
79, 5885 (1957).
86. Merten, I., and Gilman, H., J. Am. Chem. SOC.76, 5798 (1954).
86. Miller, S. A., Tehhath, J. A., and Tremaine, J. F., J . Chem. SOC. p. 632 (1952).
102
JOHN ETSCII A N D HENItY GILMAN
Morton, A. A., I,:inphcv-, E. O., J . Ojg. Chew. 20, 839 (1955).

Morton, A. A., Lanpher, E. O., J . Org. Clicm. 21, 93 (1956).
Nast, R., Vcstcr, K., and Gricssharnmer, II., Chem. Rer. 90, 2678 (1957).
Norman, R. 0. C., and Waters, W. A., J . Cherr~.Sor. p, 950 (1957).
31. Normant, H., Hull. SOC. chzm. France p. 728 (1957).
91a. Orchin, M., Wendcr, I., and Frirdcl, R. A., Anal. Chem. 21, 1072 (1949).
92. Parsons, T. D., and Ritter, D. M., J. Am. Chent. Soc. 76, 1710 (1954).
95. Parsons, T. D,, Silverman, M. 13., and Ritter, D. M., J . Am. Chem. SOC.79, 5091
(1957).
94. Paul, D. E., Lipkin, D., and Wcissman, S. I., J . Ant. Chem. Soc. 78, 116 (1956).
06. Pedlcy, J. B., Skinner, H. A., and Chernick, C. L., Trans. Faraday Soe. 63, 1612
(1953).
96. Petry, R.C., and Verhoek, F. K., J . Am. Chem. Sue. 78, 6416 (1950).
97. Pino, P., Lardicci, L., and Lorrnzi, G. P., Angew. Chem. 70, 599 (1958).
98. Prier, 8. J. W., and Trotman-Dickenson, A. F., Trans. P a r d a y Soc. 63, 1208
(1953).
98. Priec, S. J. W., and Trotman-Dickenson, A. F.,Trans. Faradav Soc. 63,939 (1957).
100. Ramsdcn, IT. E., Balint, A. E., Whitford, W. R., Walburri, J. J., and Cscrr, R.,
J . org. Chem. 22, 1202 (1967).
101. Rochow, E. G., Hurd, D. T., and Lewis, R. N., The Chemistry of Orgnnometallic
Compounds. Wiley, New York, 1957.
102. Rosenberg, S. D., Gibbons, A. J., and Ramsden, H. E., J . A ~ MChem.
.
Soc. 79, 2137
(1957).
103. Rundlc, R. E., J . Pkys. Chem. 61, 45 (1957).
104. Runge, F., Zimmermann, W., Ifeiffw, H., and Pfciffcr, I., 2. anorg. u. allgem.
Chem. 267, 39 (1951).
105. Swharkin, L. I., and Ochlohysstin, 0 . J., Repts. Arad. Sri. f i k r . S.S.R. 116, 236
(1957); Chem. Zelatr. 129, 11463 (1958).
106. Schlenk, W., Appenrodt, J., Michrtel, A., and Thal, A., Her. 47, 473 (1914).
107. Sryfcrth, D., C/icm. Revs. 66, 1155 (1955).
108. Seyferth, D., J . A m . Chem. SOC.79, 2738 (1957).
109. Scyfcrth, D., J . Am. Chem. Svc. 79, 5881 (1957).
110. Seyferth, D., and Rorhow, E. G., J . Am. Chem. Sor. 77, 907 (1955).
111. Seyfcrth, D., and Rochow, E. G., J . A m . Chem. SOC.77, 1302 (1955).
112. Seyferth, D., and Stone, 17. G. A., J . Am. Chem. Sor. 79, 515 (1957).
11.9. Slough, W., and Ubbelohde, A. R . , J . Chem. SOC.p. 108 (1955).
114. Snow, A. I., and Rundlc, It. E., A d a Cryst. 4, 348 (1951).
116. Sprirr, J. L., Jr., Zimmerman, R. E., and Webster, J., J . Am. Chem. Svc. 78, 2278
(1956).
116. Stone, F. C. A., Chem. Revs. 68, 101 (1958).
117. Strohmcicr, W., and Humpfner, K., 2. Elektrorhem. 61, 1010 (1957).
118. Sumrneia, L., and Uloth, R. H., J . Am. Ciiem. SOC.76, 2278 (1954).
111). Summers, L., Uloth, R. H., and Holmcs, A., d . Am. Che7n. SOC.77, 3604 (1955).
120. Tannenbaum, S., and Schaeffcr, P. Y.,J . Am. Chem. Sor. 77, 1385 (1955).
121. Ulmschncider, D., and Goubeau, J., Chem. Rer. 90, 2733 (1957).
122. van der Kerk, G . J. M., Chem. di Ind. p. 609 (1958).
12.3. van dcr Kcv-k, G..J. M., Luijten, J. G. A., m d Noltes, J. G., Chem. & I r ~ dp. 352
(1956).
124. Walling, C., and Buckler, S. A., J. Am. Ctiem. SOC.76, 4372 (1953).
226. Walling, C., and Ihxklrr, 8. A., J . Am. Chem. SOC.77, 6032 (1955).
87.
88.
89.
80.
103
Wartik, T., and Schlesinger, H. I., J . Am. Chem. SOC.76, 835 (1953).
Weissman, 6. I., Tuttle, T. R., Jr., and de Boer, E., J . Phys. Chem. 61, 28 (1957).
White, D. G., and Rochow, E. G., J . Am. Chem. Soc. 76, 3897 (1954).
Wiberg, E., and Htrebel, P., Ann. 607, 9 (1957).
Wibcrg, K. 13., J . Am. Chem. Soc. 77, 5987 (1955).
Wilke, G., Angew. Chem. 68, 306 (1956).
Wilkc, G., Angew. Chem. 69, 397 (1957).
133. Wilke, G., and Mdller, H., Chem. Her. 89, 444 (1956).
lt?4. Wittig, G., Angew. Chpm. 70, 65 (1958).
135. Wittig, G., Mcyer, F. J., and Langc, G., Ann. 671, 184 (1951).
136. Wotiz, J. IT., Hollingsworth, C. A., and Dcssy, R. E., J . Am. Chem. Soc. 77, 103
(1955).
137. Zeiss, H. IT., and IIcrwig, W., J . Am. Chem. SOP.78, 5959 (1956).
138. Zeiss, H. H., and Tsutsui, M., J . Am. Che7n. Soc. 79, 3062 (1957).
139. Ziegler, K., Urennst0.f-Chem. 36, 321 (1954).
140. Ziegler, K., Angew. Chem. 68, 721 (1956).
141. Ziegler, K., and Colonius, H., Ann. 479, 135 (1930).
142. Ziegler, K., and Eisch, J., unpublished studies presented in part before the Orgariometallic Symposium, sponsorcd by the Cinrinnati Section, A.C.S., October 29,
1958.
143. Ziegler, K., and Gcllt,rt, H-G., Angew. Chem. 67, 218 (1955).
144. Ziegler, K., Gellert, H-G., Martin, H., Nagel, K., and Schneider, J., Ann. 689, 91
(1954).
145. Ziegler, K., and Idimkuhl, H., 2. anorg. u. allgem. Chem. 283, 414 (1956).
126'.
127.
128.
129.
130.
131.
132.
FLUORINE-CONTAINING COMPOUNDS OF SULFUR
George H Cady
Deportment of Chemislry. University of Washington. Seattle. Washington
I . Introduction
105
I1. Fluorides of Sulfur

. .
. . . . . . . . . . . . . 107
A. Sulfur Hexafluoride . . . . . . . . . . . . . . . 107
B. Disulfur Decafluoride . . . . . . . . . . . . . . . 113
C. Sulfur Tetrafluoride . . . . . . . . . . . . . . . 113
D . Disulfur Difluoride and Sulfur Difluoridc . . . . . . . . . 114
111. Sulfur Oxyhalides Containing Fluorine . . . . . . . . . . .
. . . .
A . Oxyhalides Containing One Atom of Sulfur per Molecule
13 Compounds Containing Two or More Atoms of Sulfur per Molecule .
115
115
120
IV . Fluorooxyacids of Sulfur . . . . . . . . . . . . . . .
A . Fluorosulfonic Acid, Its Salts and Its Esters . . . . . . . .
B. Salts of Fluorodisulfuric Acid
. . . . . . . . . . . .
C . Fluorosulfinic Acid and Its Salts
. . . . . . . . . . .
123
123
127
127
V. Other Inorganic Compounds Containing Sulfur and Fluorine

A Compounds Containing Phosphorus, Sulfur, and Fluorine
B Compounds of Sulfur, Nitrogen, and Fluorine . . . .
C Compounds Containing Sulfur, Antimony, and Fluorine .
D Sulfur-Containing Compounds Formed by the Addition of
E. The Compound SC14 '2AsFs . . . . . . . . .
.
.
.
.
. . . 128
. . . 128
. . . 130
. . . . 131
B R . . 133
. . . . 132
.
.
VI . Compounds Containing Carbon, Sulfur, Fluorine, and Elements other than

Oxygen . . . . . . . . . . . . . . . . . . . .
133
A . Compounds wit.h Fluorine Bound to Carbon but Not to Sulfur . . . 133
B . Compounds with Fluorine Bound to Sulfur and Carbon . . . . . 138
VIL . Pcrfluoroalkanesulfonic Acids, Their Anhydrides, Salts, and Esters

. .
A . Trifluoromethanesulfonic Acid, CFaSOsH . . . . . . . . .
B. Esters of Trifluoromethanesulfonic Acid
. . . . . . . . .
C . Salts of Trifluoromethanesulfonic Acid . . . . . . . . . .
D. Other Perfluoroalkanesulfonic Acids and Derivatives . . . . . .
E. Trifluoromethyl TrifluoromethanethiolsuIfonate, CFsS02SCF3 . . .
V I J I . Certain Compounds Containing the -(SOz)A . Sulfonyl Fluorides . . . . . . .

References .
140
140
140
141
141
141
Bridge . . . . . . 14"
. . . . . . . . . 142
. 148
I. Introduction
I n the years since World War I1 a remarkable growth has occurred in

the knowledge of compounds of sulfur which contain fluorine The number
105
106
GEORGE
n.
CADY
of known compounds of this type has increased several fold arid thc information about some substances known for many years, particularly sulfur
hcxafluoride, has cxparided greatly. The reason for this growth is t o be
found largely in the awakened interest in fluorine chcmistry initiated by
the atomic energy program. A part of the effort devoted t o fluorine chemistry in the universities, industrial laboratories, and government laboratories
in the United States, England, Germany, and, to a lesser degree, Russia,
hits been sperit on compounds of sulfur. The laboratory which has bccn
the most, productive of published results is that of Carnbridge University.
A sccorid factor which is responsible for the discovery of many new compnunds is the electrochemical technique of Simons (282, 283) for the synthesis of fluorides by the electrolysis of hydrogen fluoride containing various
solutes. By this method many sulfonyl fluorides and perfluoroalkane sulfonic acids have been prepared for the first time.
Still a third stimulus is thc potential usefulness in industry of some of
the compounds. This factor has been particularly important in research on
sulfur hexufluoride, fluorosulfonic acid, arid the perfluoroalkanesulfonic
acids.
The large number of fluorinc-containing compounds of sulfur results
from the ability of sulfur to form strong covalent bonds with itself or with
fluorine, oxygen, carbon, chlorine, phosphorus, and nitrogeii. I t is also
irnportarit that sulfur may have coordination numbers of 2, 3, 4,5, and 6.
Formulas of a few compounds which illustrate the versatility of sulfur arc:
0 0
SFe, FbSSFb, F3CSF6, SF4, OSk'2, 0SF4, FSOSF, FSOOSF,

0 0
0 0
F
H-N-S
0
0
0
HOSF, FSF, HOSCFa, FaCSSSCFX, FJCSCI, SPFB, C102S03F,
0
0
0
I
II
S-N=S
F
II
I
Most of the recent synthescs of new compounds have to do with: (1)

substances containing - CP, or other fluorocarbon radicals, (2) substances
containing -SF6radicals, (3) oxyfluorides or salts containing oxyfluoride
anions, (4)compounds of S, N, and F.
It may be expected that new compounds will continue to come to light.
Kesearch in applied chemistry should find use both for the new and the older
compounds. Some of the substances should, for example, prove to be valuable reagents. This applies particularly to very reactive compounds such as
0
0
IIOSCF,, FxCSF, FbSOF, SF4, S0Fq
0
0
~iridothers.
(FaCS)z, SNF3,
FLUOILINE-CONTAINI~G COMPOTTNDS OF SULFUR
107
As chemists become increasingly aware of these compounds more and

more studies shall probably be made. The opportunities for studies by
physical chemists are particularly great. For example, almost no studies
of kinetics of reactions involving these fluorides have been made.
This review covers nearly the complete literature on fluorine-containing
compounds of sulfur u p through 1956. That is the last year for which a subject index for Chemical Abstracts or Chemisches Zentralblatt was available. Articles appearing through 1958 are also included when known to the
reviewer. In the case of carbon compounds the review deals with the first
member of an homologous series, for example
0
HOSCFj
0
but does not go into detail about other members of the series. No attempt
has been made to cover double salts such as Al&304)3. 4A1F3 . 12H20,
K3HS207F2. H20, etc. These have been reviewed previously (211).
Other reviews of a part of the material covered in this article are to be
found in the literature (2, 37, i78a, 211, 263, 27'2, 306).
The earliest reference used in this review is dated 1888. In it and in the
larger paper which followed in 1889 Thorpe and Rodgers (299) reported
the preparation and properties of SPF,. With the exception of its length,
a paper of this quaIity would be readily accepted for publication today.
This was a good beginning for fluorine-containing compounds of sulfur.
II. Fluorides of Sulfur
A. SULFURHEXAFLUORIDE
Sulfur hexafluoride was first produced by Moissan (215) in 1891 and
prepared and studied by Moissan and Lebeau (216) in 1900 using the combustion of sulfur in fluorine as the preparative method. Although sulfur
hexafluoride may be obtained in other ways, including the combustion in
fluorine of compounds of sulfur (225, 227, 228) and the electrolysis of solutions of sulfides, for example hydrogen sulfide in liquid hydrogen fluoride
(244, 280), the original method of preparation is still preferred (100, IOi,
190, 211, $41, 269, 330) and is now used industrially (100, 101). To obtain
pure sulfur hexafluoride the crude gas may be scrubbed with a solution
of KOH or NaOH to remove H F and lower fluorides of sulfur. The gas is
then heated to about 300"-400" to convert SzFlo to SFs and SF,. The latter
may be removed by a second treatment with base (100, 190, 269, 270). In
a typical case a sample of the combustion products of sulfur was found to
contain 94.3% (SF, S2FIo),1.9% (SF4 S2Fz),2.4% SOz, 0.4% HF ,
TABLE I
PROPERTIES
OF SULFURHEXAFLUORIDE
Reference
Vapor pressure
Solid (-72" to -51"): loglo P,, = 8.7648 - 1231.3/T
Liquid (-50" to 40"): log,, P,, = 7.308 - 907.98/T
331
21%
Transition temperature (OK)
92
94.3
Sublimation temperature 760 mm, ("C) -63.8
168
Melting point ("C)
168
-50.8"
Critical temperature ("C)
45.642, 45.547
318, $04
Critical pressure (atm)
37.193,37.113,36.8
318,204,219
Heat of transition (kcal/mole)
0.3835
92
Heat of sublimation (kcal) at -63.8'
5.64,5.57
SSl,l43, 168
Heat of fusion (kcal)
1.201
92
Free energy of formation,

1 atm, 25", (cal/mole)
235,000
531
Heat of formation, gas,

1 atm, 25", (cal/mole)
262,000
531
Heat capacity
Entropy, gas, 1 atm, 25"
See references:
69.43, 69.6
Other references
145, 168, 226, 228, 271
271
4, 212,242
m
5:
168, 331
89,90,9%,208,268
92, 331
a
d
329
Density
Gas at 20C (gm/liter)

Solid (gm/ml)
Temperature ("C)
Liquid (gm/ml)
Temperature ("C)
Critical density (gm/ml)
Solubility in water, cm3 SFs per cma H20
Temperature ("C)
Solubility in nitromethane, cma gas per
cm3 liquid
Temperature ("C)
6.093 at 753.5 mm
2.683
2.51
-195
-50
1.878
1.819
1.787
-50
-45
-39
0 . 7 4 , 0 . 7 3 ,0.7517
0.055
25
271
168
1.722
-20
1.37
20
2,212, ZSY, 168,242 204

318, ?W,.G
0.076
14.9
9Y
0.377
0.363
25 19.35
97
Specific magnetic susceptibility
(-0.300
_+
Surface tension (dyne/cm)

Temperature ("C)
11.63
-50
8.02
-20
Dielectric constant, 27.5'

Pressure (mm)
1.00191
708
loniaation potential (ev), 1st electron

From spectrum
Calculated
19.3
16.15
20.1
Viscosity of gas, cgs units

Temperature ("C)
1.537 X 10'
22.5
Collision diameter of molecule,

determined from viscosity
4.77;i
0.003) X 1 0 3
1.00123
457
1.00034
131
144
237
284
99
188
11.9
296,231
80
1.871 X lo-'
100
81
81
d
0
E0
110
GEORGE H. C A D I
arid 1.0% inerts while the refined product contaiiicd 99.870SFs,SF, S2F2
8 ppm arid 0.2% inerts (190).
The gas is colorless, odorless, nontoxic, arid inert. It is not changed by
electrical strcbs just below the corona point (82),but it is decomposed slowly
by spark-over or corona electrical discharge (S2, 276) giving lower fluorides
of aulfur aiid fluorides of the metals used as the electrodes. It does not
react with water or with a basic solution but, it does react vigorously with
a hot alkali metal. The gas is riot toxic; however, it has some depressant
action upon the ccritrul nervous system (45)and has inild anesthctic properties (312).In spite of this, rats may live in an atmosphere of 80% SF6arid
20% 0 2 for periods up to one day with no sigris of poisoning (187).
Sulfur hexafluoride is hold in cylinders containing 100 Ib of the material
at about $3 per lb (1959 price). I t is used as the electrical insulator in coaxial
cables, high voltage X-ray transformers, and high voltage generators (35,
a l l ) . Its availability is a stimulus to research dealing with the substance.
Many studies use the substance because it is inert or because it is made
up of nearly spherical molecules. There is also much theoretical interest
in its structure and in the nature of the chemical bond involved. Many
studies are related to its usefulness as an electrical insulator.
The first work on sulfur hexafluoride as a n electrical insulator was
reported by Watson and Rsmaswamy in 1934 (317). Many other studies
have followed (1, 15, 17, 31, 35, 38, 42, 43, 53, 55, 69, 82, 105, 107, il3,
l@-lt54, 203, 240, 241, 272, 293, 315, 316, 323, 335). The high dielectric
strength of the gas results from the tendency of the molecules to capture
electrons easily, thereby preventing a cascade breakdown (1, 15, 272).
There is a resonance capture of electrons of about 2 ev energy to form SF6and SF6- in about equal amounts ( 1 ) . F- is also formed easily (316).
Bombarding electrons of much higher voltage may form positive ions but
not SF6f. Appearance potentials for positive ions as observed by a mass
spectrometer itre SF6f 15.9 volts, SF4+ 18.9 volts, SF,+ 20.1 volts, SF2+
26.8 Volts, SF+ 31.3 Volts, F+ 35.8 volts, Sf 37.3 Volts, SF4++40.6 Volts,
and SFz++46.5 volts (88). The lack of a parent peak is commonly found
for fluorides. Apparently a mass spectrometer operating with negative ions
would be a desirable tool for studies of fluorides.
Physical properties of sulfur hexafluoride are listed in Table I. Thesc
include various values of the critical constants. Several excellent studies
of the critical phenomena have been made to learn whether the predictions
of Harrison and Mayer (125) are correct. They suggested in 1938 that there
could be :t range of temperature above the observed critical point (disttppcaranae of meniscus) i i i which the slope of pressure versus volume
isotherms is zero. Their arguments have been criticized by Zimm (334) in
111
1951 and Mayer (200) has agreed that Zimm's arguments are a s plausible
as those of Harrison and Mayer. Sulfur hexafluoride has been chosen by
Schneider and co-workers (4,204)and by Wentorf (318)as a test substance.
The latter has found that a liquid meriiscus can be observed a t 45.64"
but not a t 45.66". The pressure versus volume isotherms have regions of
zero slope up to 45.68' but not higher. The observed ranges of zero slope
are shown in Table 11.
RANGEO F SLOPE= 0 IN P
Temperature ("C)
Range,
Units = ml/gm.
TABLE I1
VOLUMEISOTHERMS
OF SF, (318)
VERSUS
45.52
From
To
Pressure (atm)
45.62
45.64
45.66
45.68
1.188
1.591
1.212
1.563
1.225
1.550
1.24
1.53
1.26
1.49
37.084
37,176
37.192
37.21
37.23
These observations are in agreement with the idea that the critical
point is a single temperature and not a range of temperatures. It does
appear, however, that the coexistence curve for gas and liquid has a very
flat top. This is responsible for the rather large differences in critical densities which have been reported. Orthobaric densities for liquid and vapor
(Table 111) have been determined by Miller et al. (212) from 9" to 43'. In
this work liquid sulfur hexafluoride was found t o be a good solvent for
nitrogen.
TABLE I11
ORTHOBARIC
DENSITIES
O F SFe (212)
Temperature ("C)
Density (gm/ml)
Liquid
Gas
1.47
0.14
20
1.37
0.19
30
1.26
0.27
40
1.10
0.35
42
1.07
0.39
43
1.03
0.43
Schneider (266, 267) has studied the conduction of sound of 600-kc

frequency in the neighborhood of the critical temperature. As the temperature rises the velocity of sound in both liquid and gas decreases up to the
critical temperature. At the critical temperature and pressure the velocity
is 121.5 meter/sec. Above this temperature the velocity increases. There
is a very sharp maximum in the absorption of sound over a range of about
1" with the peak a t the critical temperature. From these data the heat
capacity a t constant volume, near the critical temperature, has been calculated (268).
112
GEORGE H. CADY
In spite of the fact that thc parnchor (143.0 tit -50") was a t one time
corisidcred evidence for a nonsymmetrical structure (237) this evidence was
questioned (284, Z43) and it now appears certain that the fluorine atoms
in SFs are a t the corners of a regular octahedron, the S-F distance being
close to 1.56 A (5,28,SS,89,102,176,234). It is considered that 3d orbitals
are used in bond formation and several theoretical papers deal with the
nature of the S-F bond (58, 59, 60, 80, 103, 142, 274). The force constant,
0.344 megadyne/cm, obtained from vibration data corresponds to an S-F
distance of 1.56 (5). Studies of the Raman and infrared spectrum (83,
89, 102, 176, 177, 260, 329, $35') indicate that SFabelongs to the Oh point
group. The spectra are consistent with the regular octahedral structure.
The nmr spectrum for fluorine in SF6 consists of a single peak (118) such as
would result from a symmetrical molecule.
Iritcrmolecular forces involving sulfur hexafluoride molecules have been
diecussed in several papers (91, 121, 122, 194, 250, 296). Other studies include: (a) molecular volume (254), (b) stopping of alpha particles (16, II?'),
(c) transfer of energy by collision (205), (d) mutual diffusion of Hz and SFe
(Lgf), (e) mutual solubilities of gases, including SFe, in water (197), (f)
salting out of dissolved gases (219), (g) compressibility (193) (h) Faraday
effect (161), (i) adsorption on dry lyophilieed proteins (f4), (j) effect of
pressure on electronic transitions (231), (k) thermal relaxation of vibrational states (232),(1) ultraviolet spectrum (295), (m) solubility in a liquid
fluorocarbon (230).
The chemically inert character of sulfur hexafluoride is responsible for
the almost complete lack of exchange of fluorine atoms between SFa and
H F (249). It docs react with hot alkali metals, however, and a study has
been made of the rate of reaction of Na atoms with SF6gas using the sodium
diffusion flame technique. The rate constzmt at 247' is 2.23 X
cm3
mole-' sec-* and the energy of activatioii for the reaction SFt, Na + Sk'6
NaF, is about 37 kcal. A film of sodium 011 a glass wall does not react
with SFe at room temperature. The reaction sets in a t about 200" (67).
The fluorides, SVB, SFr, and S2F2, have no effect upon the viscosity of
liquid sulfur in the range 180-195" (93). Sulfur hexafluoride forms a solid
hydrate which has a crystal constant of 17.21 A. It decomposes just above
0" (285).
Sulfur hexafiuoride accelerates the pyrolysis of paraffin hydrocarbons
(164), lowers the octane number of gasoline containing lead tetraethyl
(189), removes silicon from a platinum catalyst when heated to 800 to 1000O
(206) and catalyzes the reaction of ammonia with a kctone and aldehyde
to give a substituted pyridine (196). It may be used at high pressure to fill a
fuse. When the fuse "blows" an arc is prevented (210).
YLUOHlrU'l+CONTAINII\iG COMPOUNDS OF SULFUlt
113
By the sublimation of sulfur hexafluoride in 1933 Denbigh and WhytlawGray (66, 67) found a small liquid residue which they identified as S2F10.
From 20 liters of crude SF6 gas they recovered only about 20 ml of SZFIO
vapor. This method of preparation has been confirmed by others (34, 100,
101, 269, 270, 307). The substance is a colorless volatile liquid which has
a surface tension of 13.9 dyne/cm a t 0C. From its vapor pressure, log,o P,,
= 7.95 - 1530/T, its heat of vaporization is calculated to be 7000 cal/mole
(67).Liquid SzFlohas a specific electrical conductivity somewhere between
and 10-14 ohm-' cm-1, a dielectric constant of 2.030 at 10"and a density
of 2.081, 2.054, and 2.028 gm/ml a t 4",12", and 20, respectively. Its dipole
moment is 0 (155). Each suljur atom is linked octahedrally to five fluorine
atoms a t a distance of 1.56 A and to the other sulfur atom a t a distance of
2.21 8 (7, 129).
Disulfur decnfluoride has to a limited dcgrce the inert character of sulfur
hexafluoride. It does not react a t 20" with water, mercury, or copper hut
it does react slowly with a solution of sodium hydroxide. It reacts with hot
mercury or copper (67).When the gas is passed through a hot tube a t about
400" it decomposes according to the equation SZFM
-+SFF, SF4 (100,190,
269,270). Within the range 433455K the reaction is homogeneous and of
first order with a rate constant described by the equation, K = 2.99
x lozoe-49*200/RT. The very high frequency factor suggests a chain reaction
(307). The compound is more toxic than phosgene. Lung injury and death
within 18 hr occur to rats placed in an atmosphere containing one part
per million of S2FI0(187).The 50% lethal dose, by intravenous injection in
dogs as a lecithin emulsion, is 5.79 mg/kg of body weight (259).
C. SULFURTETRAFLUORIDE
Sulfur tetrafluoride was identified in 1929 by Fischer and Jaenekner
(96)who produced the gas by heating a mixture of sulfur, cobalt trifluoride,
and calcium fluoride (this inert material was used to reduce the vigor of
the reaction and to prevent explosions). The compound has also been obtained by: (a) burning sulfur in fluorine (34,100,101,190,270),(b) heating
SzBrzwith IF5 (265) (c) combustion of CSz in fluorine highly diluted by
nitrogen (280), (d) decomposition of CFISF5 into CF4 SF, by sparkover electrical discharge (279), and, perhaps, (e) spark-over electrical discharge in sulfur hexafluoride (272). (This process gives one or more lower
fluorides of sulfur but it is not certain that it gives SF,).
The substance is colorless. It fumes in air, has an irritating bad odor

(96), a surface tensioii a t 200K of 25.70 dynes/cm, and a heat of vaporiaation a t the boiling point of 6320 cal/mole (34).A high Trouton constant
suggests that thc liquid is somewhat associated. Equations showing the
effect of temperature upon various physical properties tire given below (34).
Vapor pressure, 160-224K:
Siirfacc tcnkm, 190-230K :
Density, 170-2001i :
Coefficicnt of Cii expansion, 170-200'K:
loglo PI,, = 8.8126 - 1381/T

61.36 - 0.17832'
d (gm/ml) = 2.5471 - 0.003142'
= 0.00170
y (dynelcm) =
The irifrarcd and Raman spectra indicate that the molecule has no sym(70).This structure is confirmed by the nrnr specniet,ry (point, group CZ,~)
trum for liquid SF4a t - 100". There are two equally strong sets of triplet
peaks which suggest two pairs of identical fluorine atoms. As the SF, is
warmed, the triplets become broad single peaks above -94" and at higher
tcmperaturcs the t w o broad peaks merge into one. This effect indicates
a rapid exchange of fluorine atoms (66). The structure suggests that hybrid
sp3d-orbitals are used for bond formation giving a trigonal bipyramid with
an unshared electron pair occupying one of the "belt" positions (56).
Before the above studies were made, the structure was thought to be tetrahedral (175, 298).
Sulfur tetrafluoride reacts with a little water to give SOFz and HF (70).
I t is readily absorbed by a solution of sodium hydroxide (100, 270) and a t
first it was thought, to react with mercury (96). Ilater work has shown that
the SF, docs riot rcact with mercury but that the crude material contains
reactive impurities whic,h can be removed by shaking with mercury (191).
Sulfur tetrafluoride combines with boron trifluoride to form the compound SF, . BF3 which sublinics a t itbout 80" under a pressure of 1 atm.
When this compound is heated wit,h sodium fluoride the boron trifluoride
reacts to give NaBF4. Sulfur tetrafluoride combines with arsenic pentafluoride to form SE'4 . AsF6 (s,ublimes at 190") arid with SbF6 to form
SF, . ShF6 (mclts at 245"). These substances may be donor-acceptor addition compounds (6) but it is thought that SFa . BFI is composed of the ions
SF3+and BF4- (56).
K.eviews of the chemistry of sulfur tetrafluoridc may be found in some
of the references (34, 37, 211, 263, 306).
D.
DISU1,FUR I)[FLTJORIDE
AND
("SULFUR
MONOFLUORIDE")
SULFUIZ
DIFLUORIDIC
I t is well known that the fluorination of sulfur produces one or more

lower fluorides which are unstable at, room temperature and deposit, sulfur
FLUORINE-CONTAISING COMPOTJNDS OF STSLFUIZ
115
011 the walls of the container. In 1923 Cciitticrzwcr and Strenk (46) obtained
such a product by heating a mixture of sulfur and silver fiuoride. From the
density and analysis of the gas (not distilled to isolate the various substances present) they chose S2Fzas the formula. They later (47) reported
a melting point of -105.5" and a boiling point of -99' but found that a
liquid remained even a t room temperature when the product boiled away.
Strenks (292) found that SiF4was present in his crude product and concluded that the earlier physical constaiits should not be trusted. Ruff
(253), in a preliminary report on incomplete work of Jaenkner, said in
1933 that SzF2 boils a t -38.4' and freezes a t -120.5', and that SFz (the
first report of this compound) boils at about -35'. He also said that because of the proximity of boiling points, the compounds SF4, SF2, SzFz,
and SOFZare difficult t o separate from each other (253). In this work
neither SF2 nor S2F2 was obtained pure. Trauta and Ehrmann (306) found
the product obtained from sulfur and silver fluoride to be a mixture, but
they were unable to isolate a pure sample of S2Fz.They obtained a mixture which they thought to be SF2 and S2F2.Dubnikov and Zorin (75)
described in 1947 various methods of producing crude mixtures of SF2
and S2F2but did not isolate either compound. Still others (82, 166, 199,
270, 272) have worked with these lower fluorides of sulfur but no one has
clearly reported isolating either SFz or S2F2.
An infrared spectrum has been reported for SzF2 (199) and infrared
spectra have been used in the Bell Laboratories t o identify both SFZand
SzF2 as products of the decomposition of SF6 by electrical discharge (82).
It is not said how the control samples of SFz and S2Fz were prepared nor
are the spectra given. The products, SF2and SzF2 are described as being
fairly stable when pure but readily hydrolyzed by a trace of water to give
H F and SO2 (sulfur not mentioned) (82).These lower fluorides are absorbed
by NaOH solution (46, 306), and hydrolysis by water is said to give HF,
H2S03,and S (292) The gas is said to attack mercury (306) and, when
pure, not to attack mercury (4'7). It is poisonous (292, 82).
If the reader is now confused about S2F2and SF2, he is in the same position as the writer.
Ill. Sulfur Oxyhalides Containing Fluorine
A. OXYHALIDES
CONTAINING
ONE ATOMOF SULFURPER MOLECULE
1. Thionyl Fluoride, F2SO
Thionyl fluoride was first prepared by Meslans (207) in 1896 and four
years later it was described in more detail by Moissan and Lebeau (216,
217) who produced it by the reaction of thionyl chloride with fluorine or
116
GEORGE H. CADY
(preferal~ly)itrseiiic trifluoride. Steinkopf :itid Herold also used arsenic

trinuoride (288) Other reactions which have been used to produce thionyl
fluoride are : (a) refluxing thionyl chloride over antimony trifluoride containing some antimony pentachloride (21) (b) heating together thionyl
chloride and hydrogen fluoride in the presence or in the absence of a
catalytic amount of aritimony pentafluoride (314, 320) (c) contacting
vanadium pentafluoride with sulfur dioxide a t room temperature (98%
yield of S0F.J (50) (d) the action of fluorine upon cold sodium thiosulfate
(229, 239) (low yield of SOFz), (e) heating together S4NdrCuO, and HF in
a sealed bomb a t 100' (655), (f) heating together iodine pentafluoride and
thionyl chloride (165) (g) passing thionyl chloride vapor over potassium
fluorosulfinate, KSOZF,a t 150" (273).
Thionyl fluoride is a colorless gas having an irritating odor. The substance does not attack glass (at 20") or mercury, but it reacts rather slowly
with water to give hydrofluoric and sulfurous acids. With ammonia it gives
OS(NH,), (207).It reacts with hot glass but is inert toward iron a t red heat
and toward several other metals a t temperatures up t o 125' or more (21).
The vapor pressure of the liquid is given by the equation loglo P,,
= 30.333 - 1908.4/T - 8.1053 loglo T. The critical temperature and pressure are 89.0"and 55.3 atm, respectively (21). Infrared and Raman spectra
have been interpreted (12, 233, 329) and the microwave spectrum giv:s
moments of inertia consistent with the structure: TSO = 1.412 f 0.001 A,
rSF = 1.585 f 0.001 A, L FSF = 92"49' f 5', L OSF = 106" 49' f 5'. The
Stark effect has also been studied (96).By use of method of linear comhination of atomic orbitals (LCAO), Moffitt has concluded that the SO hind
in thionyl fluoride and i n sulfuryl fluoride is a double bond (21.4).
2. Thionyl ChZorojZuoride, SOClF
When Booth and Mericola (1940) allowed thionyl chloride to reactr

with antimony trifluoride containing some antimony pentachloride thcy
obtained as products both thionyl fluoride and thionyl chlorofluoride,
SOClF (21). The compound was also produced in the laborutory of Otto
Ruff (1937) but the work was not published until 1951. For this work iodine
pentafluoride was heated with thionyl chloride (165).
Thionyl chlorofluoride is somewhat more reactive than thionyl fluoride
but less so than thionyl chloride. It reacts with mercury and is readily
hydrolyzed. Upon standing at room temperature for several months it
disproportionates to a considerable extent into thionyl chloride and thionyl
fluoride. The reactivity of SOClF a t elevated temperatures may be due
largely to SOC1, formed by disproportionation (21).When thionyl chlorofiuoride is chilled quickly it congeals to a glass (165) but it can be made to
117
form crystals which melt a t -139.5". The influence of temperature upon

density and vapor pressure are shown by the equations:
d(gm/ml) = 1.576 - 0.00224 t
log,o P,, = 7.0466 - 12782.2/T
loglo P,, = 7.83 - 1409/T
+ 0.1268 log10 T
Ref. 165
Ref. 21
Ref. 165.
where t is in degrees centigrade and T is in degrees Kelvin.

3. Thionyl Tetrafluoride, SOF,
Moissan and Lebeau (1902) mixed fluorine and thionyl fluoride a t room
temperature and obtained a new product. From the change in pressure
during the reaction they considered the product to have the formula SOF,.
They did not definitely identify the substance (217). The substance was
prepared and studied in the laboratory of 0. Ruff in 1937 but the work
was not published until after World War 11. The reaction of Fzwith SOF2
took place in the presence of platinum a t about 150" (165, 174). Thionyl
tetrafluoride has also been produced from SOFn and Fz in a copper tube
reactor packed with a fluorinated copper, AgFt catalyst (78).
Thionyl tetrafluoride is a colorless gas which reacts with water rapidly
forming SOZFz, HF, H+, and SOIF-. With NaOH solution it gives F- and
SOsF-. It reacts slowly with mercury a t room temperature to give mercurous fluoride and thionyl fluoride (78, 165). The density and vapor pressure of SOFl are given by the equations:
d(gm/ml)
d(gm/ml)
loglo P,,
log10 P,,,
P
= 1.653 - 0.00360 t(OC)
Ref. 165
Ref. 78
Ref. 165
= 7.76 - 1092/T
= 7.2349 - 8.5958 X 102/T - 2.6275 X 104/T2
Ref. 78
= vapor pressure of liquid; heat of vaporization = 5090 cal/mole
= 2.6963
- 4.1121 X
1O-I
T("K)
The mass spectrum is known (78) and there is only one nmr "line" for
fluorine in the nmr spectrum of thionyl tetrafluoride a t room temperature
(79). This probably means that there is a rapid exchange of fluorine atoms
within the molecule.
4. Pentafluorosulfur H y p oJzuorite , F OF
I n 1955 Dudley, working with Cady and Eggers, obtained pentafluorosulfur hypofluorite by the reaction of fluorine, in excess, with thionyl
fluoride, or sulfur dioxide, a t 200" in the presence of copper coated with
fluorides of silver (78).It is probable that thionyl tetrafluoride is a n intermediate in the process and that it combines with fluorine to give FsSOF. !
118
GEORGE H. C.4I)Y
Pentafluorosulfur hypofluorite is a very reactive colorless gas having an

odor resembling that of oxygen fluoride. It reacts with aqueous KI and
aqueous KOH according to the equations:
+ + H20 = SOZFz + 2HF + 1 2 + 2FFhSOF + BOH- = 5F- +

+ 31120 + SOalFsS01~ 21-
$ 0 2
Equations for vapor pressure and density of the liquid are:

log,,
Y,,
d(gm/ml)
= 6.03633 - 4.2035 X 102/T

=
- 7.836 X 104/!P
2.788 - 3.722 X 10-3T
There are two widely scpmitcd %nes7for fluorine in the nmr spectrum of
F,SOF. As would he expected for this molecule, the line a t the lower magnetic field is only one fifth as strong as the other. Since the five fluorine
atoms attached to the sulfur cause only one line at room temperature
it is probable that a rapid exchange occurs among these five atoms. At a
lower tcrnperature this lir~cshould break into a complex structure (78,
79)
Electron diffraction by pentafluorosulfur hypofluorite is consistent with

the following structural features for the molecule: TSF = 1.53 b, rop =
1.43 A, rso = 1.64 A, LFSF = about 90) L S O F = about tetrahedral,
somewhat grcater than 103,L FSO (opposed F and 0 atoms) slightly different from 180. The structure about the sulfur atom is essentially octahedral (61).
5. Sulfuryl Fluoride, S02F2
Moissan and Tlebeau (1901) produced sulfuryl fluoride by the combination of sulfur dioxide with fluorine (217).Other p~-occsseswhich have been
used to produce the gas are: (ti) the thermal decomposition of barium
fluorosulfonate or certain other fluorosulfonates (133, 221, 303), (b) the
reaction of sulfur dioxide with chlorine and hydrogen fluoride in the
presence of activated charcoal at 400 (II), (c) the reaction of sulfur dioxide
and chlorine with potassium or sodium fluoride a t 400 (328), (d) the disproportionation of sulfuryl chlorofluoride a t 300-400 (328), (e) the reaction of sulfuryl chloride with a mixture of antimony trifluoride and antimony pcntuchloride a t about 250 (%), (f) the reaction of sulfur dioxide
with silver difluoride (867, (g) the reaction of thionyl fluoride with oxygen
in an electrical discharge (314))(h) electrolysis of a solution of fluorosulfonic
acid in hydrogen fluoride (264)) (i) the reaction of fluorine with sodium
sulfate, sodium sulfitc or sodium t1iiosulf:itc (229, %.9), (j) the reactiori
of hydrogen fluoride with sulfuryl chloride (320).
I n many ways sulfuryl fluoride is much like sulfur hexafluoride. Its
FLUORINE-CONTAINING COMPOUNDS O F SULFUR
119
rate of hydrolysis even in a basic solution is slow; its boiling point is low;
it is stable (217, 306), but decomposes slowly at about 1000" to give SOFz
(221);it is a very good gaseous electrical insulator (17, 31). Although the
hydrolysis in water is very slow, the reaction with highly concentrated
hydrofluoric acid is almost instantaneous (3ZU). The hydrolysis in alcoholic
potassium hydroxide is more rapid than in aqueous potassium hydroxide
(217 ) .
The physical properties of sulfuryl fluoride are not well known. As one
may see from Table VI there is considerable doubt about its melting and
boiling points. Many other physical constants have not been measured a t
all. Over the temperature range -155" to -83" the vapor pressure is
represented by the equation loglo P,,,, = 7.593 - 1023/T and the boiling
point obtained from this equation is -56.2" (229). Good studies of the
Raman (12) and infrared spectra (12, 238) have been made. An electron
diffraction stvdy gives scattering c2nsistent with the structure: TSF =
1.56 zt 0.02 A, TSO = 1.43 f 0.02 A, L F S O = 105 f 2", LFSF = 100
f lo", LOSO = 130 f 10" (290). The microwave spectrum gives a
dipole moment of 0.228 f 0.004 Debye units :indois consistent with a
Czu symmetry corresponding to TSF = 1.570 f 0.01 A, T S O = 1.370 f 0.01
A, L FSF = 92" 47' f 30' and L OSO = 129" 38' f 30' (98). Molecular
orbital (LCAO) calculations assuming the use of 3d orbitals of the sulfur
atom indicate that the SO bond is essentially a double bond (214).
6. Sulfuryl Chlorujuoride, SOzClF
I n 1936 Booth and Hermann (19) reported the preparation of sulfuryl

chlorofluoride by the reaction at 300" of sulfuryl chloride with antimony
trifluoride containing some antimony pentachloride. Others have used a
similar procedure (8G,113). The compound has also been produced by
heating together a t 80" sulfuryl chloride and ammonium fluoride (528),
by the reaction of sulfuryl chloride with CoF3, MnF3, or AgFz (LOI), and
by the reaction of pyrosulfuryl chloride, Sz05C12, with iodine pentafluoride
(265). Sulfuryl chlorofluoride is colorless and intermediate in reactivity
between SOJ.32 and SOZF?. It reacts with water but not with glass, mercury, or brass at room temperature (19). It decomposes at 300-400" in
metal apparatus to give SOz, Clz, and SOzFz (3%'). At 0" its surface tension
is 17.2 dyne/cm. (19).
7 . Sulfuryl BromoJEuoride, SOzBrF
In 1951 Jonas reported the preparation of sulfuryl bromofluoride in the

laboratory of 0 . Ruff in 1937. The substance was produced by the reaction
of: (a) BrF3 with CC13SO&1, (b) sulfur dioxide with bromine and bromine
trifluoride. It is very stable but reacts vigorously with water. The density
120
GEORGE H. CADY
of the solid a t liquid air tcmperaturc is 3.16 gin/ml and for the liquid the
density may be obtained from thc equL t1011
d = 2.175
- 0.00289 t(C).
Vapor pressures correspond to the equation

loglo P = 8.03 - lGlO/T (166).
8. Fluorine Fluorosulfonate, S03F2
Dudley, working with Cady and Eggers, in 1955 obtained fluorine

fluorosulfonate by combining sulfur trioxide with fluorine at about 200,
or by the reaction of fluorine with sulfamic acid. The compound is colorless
and it has an odor resembling that of oxygen fluoride. It reacts with a solution of sodium hydroxide and with potassium iodide solution according to
the equations:
S03J3.2
SOaF,
+ 20H-
+ 21-
+ J3- + HnO f 40%

+ F- + Ia
= SO3F-
SOaF-
The influence of temperature upon vapor pressure nnd density of the liquid
are represented by the equations (76):
loglo P,,
= 6.56476
2687 X 10* G.3906 X

- 6r7
II
104
d = 2.4314 - 0.00325 T
There are two doublet peaks of equal intensity in the nmr spectrum (79).
This fact together with the chemical reactions given above is in accord
with the structure
0
080F
F
B. COMPOUNDS
CONTAINING
Two
PER MOLECULE
011
M o m ATOMSOF SULFUIZ
1. Bis(pentu~uoros~~1jur)
Peroxide, F&OOSF6
This compound was first prepared by J. W. Dale and D. A. MacLeod

of the Defense Research Chemical Laboratories, Ottawa, Canada. The only
publication is that of Harvey and Bauer who determined its structure by
electron diffraction and who used the name (disulfur decafluorodioxide.
The substance was obtained as a minor product of the combustion of sulfur
in fluorine. It is diamagnetic and hns the stzueture of a peroxide with the
structural parameters: TSF = 1.56 f 0.02 A, roo = 1.47 f 0.03 8, rso
= 1.66 f 0.05 A, LSOO = 105 f 3) dihedral angle SOOS (looking along
121
axis through the two 0 atoms) = 107 f 5". The SFS groups are octahedral (128). The value of rso ill this compound is much larger than that
in SOFz or SOZFz.
811
2. Pyrosulfuryl Fluoride (Disulfuryl Fluoride), S206F2
In 1951 Hayek and Koller produced pyrosulfuryl fluoride by refluxing

sulfur trioxide with antimony pentafluoride (136). The compound has
also been obtained by: (a) the thermal decomposition of a product made by
the combination of sodium fluoride with sulfur trioxide (132), (b) the reaction of iodine pentafluoride with sulfur trioxide (265), (c) the thermal
decomposition of certain fluorosulfonates (133), (d) heating together
HSOsF and Asz06 (135), (e) the reaction a t room temperature of sulfur
trioxide with vanadium pentafluoride (50), (f) mixing 60-98% sulfuric
acid a t 25" with solid [Ca(S20eF)z?]formed by adding an excess of sulfur
trioxide t o calcium fluoride at 100-300" (221, 222).
Pyrosulfuryl fluoride is a colorless liquid which hydrolyzes rather sIowly
to give HSOsF. It is not very solubIe in cold concentrated sulfuric acid
but is soluble in carbon tetrachloride or benzene. It is poisonous. When
heated to about 400" it decomposes into sulfuryl fluoride and sulfur trioxide
(134, 135, 136, 221). The equation for its vapor pressure is
l0g1.loP,, = 8.015 - 1662/T.
This gives a heat of vaporization of 7600 cal/mole (50).The infrared spectrum and the nmr spectrum are consistent (291) with the structural formula
0 0
FSOSF.
00
3. Pyrosui&ryl ChloroJluoride, SzO&lF
Engelbrecht (1953) produced pyrosulfuryl chlorofluoride by the reaction of pyrosulfuryl chloride with silver fluoride a t about 80". The reaction
also gave some sulfuryl chlorofluoride.
Pyrosulfuryl chlorofluoride is a colorless liquid which hydrolyzes slowly
and does not attack dry glass or mercury. From the vapor pressures the
heat of vaporization a t the boiling point is 8070 cal/mole (87).
Vapor pressure (mm of Hg)
49.2
68.0
133.0
319.2
567.7
799.4
Temperature ("C)
32.7
40.0
53 . 1
73.9
90.7
101.7
4. S206F4,M o n o p e r o x y t e t r u ~ ~ o r o ~ i ~ uVl fIu)roxide

(
In 1952 Wannagat and Mennicken (314) described the reaction of oxyUpon distilling the
gen with thionyl fluoride in an ozonizer a t -50--60".
122
GEORGE H . CADY
product, they found iri addition t,o iiiiused reactniits the substances sulfuryl
fluoride, Sd )5F4 arid an irripurc material having appr~xirnat~ely
the composition SzOsFz.The yields were small and the work was difficult because
of the small arnounts of materials involvcd. S406F4 decomposes above - 20
into SOzF2and 0 2 . It oxidizes I- to 12. It may have the structure
F-S(
0-0
F.
0 q -
5. Peroxydisulfuryl DiJEuoride, S20BF4
The product of Warinagat arid Mennicken (314) (see Section III.R.4)

contairied a srnnll amount of material which was not obtained pure, but
had a cornposition close to Sz06F2.The properties of this matcrial differed
eonsiderahly from those of a pure form of SsOsPz,peroxydisulfuryl difluoride, reported in 1957 by Dudley and Cady (77).It is, therefore, not certain
that the two substances are the same. The compound is easily produced
by the reaction of fluorine with an excess of sulfur trioxide a t about 250
or by the reaction of fluorine fluorosulfonate with sulfur trioxide.
Peroxydisulfuryl difliioride is a oolorlcss liquid which reacts with water
according to the equation :
SsOsk,
+ HzO
2803F-
+ 2H+ + 40s.
Wheii reachting with potassium iodide solution, one molc of S20sF2liberates

one mnlc of iodine. Equations for the vapor pressure and density are:
loglo PllIIrl
= 5.49916 - 1.2925 X 102/T - 2.5921 X 106/T2
d(gm/ml) = 2.3959 - 2.434 x 10-3 T
The iiifrarod
:tiid
nmr spectra are in accord with the structural formula

0
I
I
l?-S--(-
0- &q-1<3. (77)
I
6. T r i s u l f u q l Fluoride, S3OsF2
A compound of the above empirical formula and prcsumal~lyof the

structure
0
was described by Lehmann and Kolditx in 1953 (186).T o prepare the substance they saturated liquid siilfur trioxide with boron trifluoride and to
FLUORINE-CONTAINING COMPOUKDS OF SULFUR
123
the resulting solution then added 70% sulfuric acid while keeping the
materials cooled by ice. This released S308F2 as a separate liquid phase
which could be removed by distillation. They also formed the substance
by the destructive distillation of KBF4 . 4S03 at about 65".
Trisulfuryl fluoride is a colorless liquid which hydrolyzes slowly in
KOH solution according to the equation
S30&'2
+ 40H-
2SOaF-
+ S04- + 2Hz0.
This reaction indicates that the two fluorine atoms are not attached to the
same sulfur atom in the molecule (186).
IV. Fluorooxyacids of Sulfur
A. FLUOHOSULFONIC
ACID,ITS SALTSAND ITS ESTERS
1. F E Z L O ~ O S U ZAcid,
~ O ~ ~HS03F
C
I n 1892 Thorpe and Kirman reported the preparation and properties

of fluorosulfonic acid. They condensed together sulfur trioxide and hydrogen
fluoride and removed the excess of the latter by sweeping with a stream of
carbon dioxide (298). The acid is now made commercially by combining
sulfur trioxide with hydrogen fluoride (157, 158, 159, 211, 289). A convenient laboratory method is the distillation of a mixture of potassium acid
fluoride and oleum in glass apparatus (211). Other procedures which have
been used are: (a) distilling a mixture of a fluoride and oleum (163, 213,
256, 302, 304), (b) distilling a mixture of HS03Cl and NHaF or a fluoride
of an alkali or alkaline earth metal (160), (c) adding sulfur trioxide to a
mixture containing sulfuric and hydrofluoric acids ( S l l ) , (d) adding hydrogen fluoride t o chlorosulfonic acid (320).
The anhydrous acid is a colorless liquid which fumes in air and dissolvcs
readily in water. In solution the acid is highly ionized as shown by electrical
conductivity (326). The SOSF- which is formed a t first, reacts eventually
to give SO4' and HF. When a little water is mixed with fluorosulfonic acid,
an equilibrium of the type H2S04 H F $ HS03F HzO is established.
This reaction has been studied by Traube and Reubke (305) and by Lange
(181). Thorpe and Kirmann (298) found a boiling point of the acid of
162.6' (the value still given in the literature) and they considered that slow
decomposition occurred a t the boiling point to give, presumably, sulfuryl
fluoride and sulfuric acid. In contrast to this Ruff (252) found the acid to
be stable up to 900C. The anhydrous acid is a poor electrical conductor
but it acts as an ionizing solvent,. Dissolved KS03F, according to studies
by Woolf (327), is highly ionized in HSOBFand the high molecular conductivity indicates a charge transfer mechanism of the Grotthus chain
124
GEORGE H. CADY
type. The self-ionization of the anhydrous acid gives mostly S03F- and a
solvated proton. Such ions as HzF+ and SzOaF- may also bc present. In
anhydrous HS03F antimony pentafluoride acts as an acid, probably forming SbFe- as it does in liquid HF. Other acidic substances are AuF3,TaFs.
and PtFd. Among the bases are fluorosulfonates, such as KS03F, and the
fluorides AsF,, SbF3, BrF3, and IF&.Even HClO4 is a proton acceptor and,
therefore, a base (327).
The abnormally high heat of neutralization, 14.46 kcal/mole, of fluorosulfonic acid (526) has not been fully explained.
Glass is not appreciably attacked by the liquid acid but the part of a
glass container in contact with the vapor over the liquid is attacked slowly,
The cold anhydrous acid does not attack S, C, Se, Te, Pb, Ag, Cu, Zn,
Fe, Cr, or Mn but it slowly dissolves Hg and Sn. At higher temperatures
it dissolves S, Pb, Hg, and Sn rather rapidly. Rubber, cork, wood, and
scaling wax are attacked (209).
Fluorosulforiic acid is a useful catalyst and reagent for the production
of important organic compounds. Uses given in the literature are: (a) a
catalyst for the alkylation of hydrocarbons (71, 127, 162, 286,297), (b) a
catalyst for the isomerization of hydrocarbons (39, 192, 547, 248), (c) a
catalyst for condensation of various mixtures of organic materials (163,
166,583),(d) a reagent to cause ring closure (62,146), (e) removal of metals
from petrolcum (54), (f) desulfurization of petroleum products (48), (g)
improvement of viscosity of lubricating oil (44), (h) production of esters
and salts of fluorosulfonic acid (42,180,184,246,262), (i) an agent to polymerize rosin (26), (j) an agent to attack cellulose (all?),(k) a selective solvent for separating hydrocarbons (%+), (1) a solvent to absorb and recover
hydrogen fluoride used as a catalyst for the alkylation of paraffin hydrocarbons (I.%), (m) a reagent for the production of DDT (257), (n) a component of a solution used for the anodic polishing of stainless steel (333).
The chemistry of fluorosulfonic acid, its salts and its esters has been
reviewed by Lange ( 1 7 8 ~ ) .
2. Pluorosulfonate ion, SO$-
The fluorosulfonate ion is prescnt in aqueous solutions of fluorosulfonic

acid or its salts. It hydrolyzes slowly according to the equation S03F+ HzO-+ SO4= F- 2H+. In a slightly alkaline solution the reaction
is first order with respect to the fluorosulfonate ion with a heat of activation of 17.9 kcal and an entropy of activation of -31 cal/degree. In acidic
solutions the rate is more rapid than in neutral or basic solutions (258).
Studies of the infrared (276, 277) and Raman spectra (278) of fluorosulfonates have been made.
Since the fluorosulfonate ion has the same sort of structure and charge
+ +
125
as the perchlorate ioii it is not surprisiitg that certain fluorosulfonates are

isomorphous with the corresponding perchlorates ( 178,,276,321). Isomorphous salts have been obtained for the anions BF4-, s@$-, C104-, Mn04(321).
In general the salts of fluorosulfonate ion are very soluble in water, the
only highly insoluble one being that formed by the organic base, nitron
(211).
3. Salts Containing Fluorosulfonate Ion

A procedure which has been used commonly t o prepare fluorosulfonates
is the addition of sulfur trioxide t o a fluoride. This method was used by
Traube in 1913 (301, 302j and has since then been employed by others
(133, 221). Even chloryl fluoride, C102F, adds SO3. The product, chloryl
fluorosulfonate, C102S03F, is a red liquid at room temperature. It was discovered independently in 1954 by Woolf (325) and by Schmeisser and
Ebenhoch (263).Another convenient method for preparing fluorosulfonates
is t o add a fluoride of a metal to fluorosulfonic acid and then t o distil off
hydrogen fluoride (303). Fuming sulfuric acid and a fluoride may also be
used (301, 303). Sodium chloride and fluorosulfonic acid react when heated
to give HC1 and NaS03F (252).In some cases the reaction of sulfur trioxide
with a fluoride gives a double salt. For example, tung3ten hexafluoride
gives a product which probably is WF2(S03F),(49).The discoverers, Clark
and Emelbus (1957) say that it, like NbF3(S03F)2and TaF3(S03Fj2,is a
liquid of low volatility a t room temperature and that it hydrolyzes readily.
Another method used to produce double salts, such as AICI(S03F)2,
TiCl(S03F)2,and ZrF3S03F, is the reaction of a chloride of a metal with
fluorosulf onic acid ( 137 j .
Arsenic trifluoride has been found by Engelbrecht et al. (88) t o combine
with sulfur trioxide t o form a volatile compound 2AsF3 3so3. This has
been confirmed by Muetterties and Coffman (221) who found that SOs and
SbF3 appear t o give Sb(S03F)3.The nmr spectrum of 2AsF3 3S03 consists
of three FI9 peaks of relative intensity 3 :2 : 1, the peak of intensity 3 corresponding t o fluorine bound to sulfur. This suggests the structure:
0
F-S-0
\ /
\*/
As
/\0
\/
/ \
I
I
0-S-F
0
The substance behaves as a fluorinating arid sulfoiiatiiig agent (221).
126
GEORGE H . CADY
Melting points for ti few Auorosu1foii:ites :ire given ill Table V I (211,
303).
The pyrolysis of barium strontium and zinc: fluorosulfonates a t about
500" produces a gas which is largely sulfuryl fluoride (221,303). By a similar
process calcium fluorosulforiate gives sulfur trioxide and only a little S02F2
(221).
Infrared spectra are known for the fluorosulfonates of ammonium,

sodium potassium, rubidium, and cesium (276, 27'7, 278).
TABLE I V
MELTINQPOINTS
OF FLUOROSULFONATES
Salt
Tcmpcrature ("C)
NH&W$F
LiSOoF
245
360
311
KSOaF
Salt
Temperature ("C)
RbSO3E'
304
CsSOsF
2Y2
Nitrosyl fluorosulfonate, NOSO3F, was produced by Lange (179)

(1927) by the reaction of Nz03with HS03F. It may also be formed by the
reaction of BrF3 with (N0)8@7 (324) or as a by-product of the oxidation
of a perfluoroalkane sulfonyl fluoride by an excess of nitrogen dioxide a t
and
~ FNO, give C,F&OF and NOSO3F
about 550". For example, C ~ F I ~ S O
(275). Nitrosyl fluorosulfonate is a white solid a t room temperature. It is
very hygroscopic. An impure form of the compound, as obtained by Lange
( I T $ ) , begins to soften a t about 118"and is completely molten a t about 140".
Nitronium fluorosulfonate, N02S03F,is formed by the reaction of a n
excess of nitrogen dioxide with a mixture of sulfur trioxide and bromine
trifluoride (Woolf, 1950, 324) or by the reaction of dinitrogen pentoxide
.
compound is a
with fluorosulfonic acid (Goddard et al., 1950, 1 1 1 ~ )The
white solid composed of the ions NOz+ and SOP-, as shown by its Rumari
. melts a t about 200' and reacts with water to give H+,
spectrum ( 2 0 9 ~ )It,
NO3- and S03F- ions ( 1 1 1 ~ ) .
4. Esters of Fluorosulfonic Acid

Several aryl esters of fluorosulfonic acid were prepared by Lange (1930)
by decomposing an aryldiazonium fluorosulforiatc, such as C6H6NZS03F
a t about 100' (180, 184). A method for preparing a diazonium fluorosulfonate is to add HSO,F to an aqueous solution of a diazonium salt and to
then remove the precipitiited diazonium fluorosulforiate (145). Alkyl
esters were prepared by Meyer and Schramm (1932) by the reaction of
fluorosulfonic acid with an ether or an alcohol. For example, C2H60Hand
HSOIF gave CZH~SOJF
(209). Calfee and Florio produced esters by adding
CH,=CF, or C H C k C F 2 to fluorosulfonic acid a t a temperature below
30" (41). The resulting esters are useful as fumigants for stored grain (40).
FLUORINE-CONTAINING COMPOUNDS O F S U L F U R
127
A few physical constants of some of these esters are given in the literature (41,180, 184, 209, 246).
B. SALTSOF FLUOHOI)ISULFUIZIC
ACID
Lehmann and Kolditz (1953) condensed an excess of sulfur trioxide on
potassium fluoride and later distilled off the excess a t about 35". White
crystals of potassium fluorodisulfate, KS2OsF, remained (185). Muetterties
and Coffman thought that they obtained Czt(S206F)z by the action of a
large excess of SO3upon CaF2 but they did not analyze the material (221).
C.
FLUOROSTJLFINIC
ACIDAND
ITS s . 4L T S
1. Fluorosuljinic Acid, HSO2F?
Seel and Riehl (1955) have studied the system HF-SO, mid havc
found a single compound, HF SOz, melting a t -84"C, which may he
fluorosulfinic acid, HOSO. It is not at all certain that the substance has this
F
structure (273).
2. Salts of Fluorosuljinic Acid
Fluorosulfinatm of the alkali metals, other than lithium, and of the

tetramethyl ammonium radical are made, according to Seel and Riehl
(1955), by allowing sulfur dioxide (liquid or gaseous) to combine with
fluorides of the alkali metals or with (CH3),NF (273).The larger the cation,
the more rapidly the reaction occurs. The salts contain the S02F- ion. As
one might expect, KSOzF is structurally similar to KC103.
Potassium fluorosulfiriate reacts with the chlorides S02C12, SOClz,
CsHsS02Cl,PC13, P0Cl3, AsC13, CICHzCOCI, and CsH6COC1 to give the
corresponding fluorides. The potassium salt is a more convenient fluorinating agent than SbF3, Fz,or HF. Thc preparation of thionyl fluoride may be
accomplished, for example, by passing a stream of thionyl chloride vapor
in nitrogen over a bed of potassium fluorosulfinate in a glass tube at 150".
Other reactions of fluorosulfinates are (273):
+ HzO (g) 2MHFz + MzSzOs + 2502

+ 2H2O (1) = MHSOa + MHFz + HzS03
liq. Sot
(CH,)rNS02F + MI = MS02F (s) + (CHa)dNI
175"
2MS02F + SOz
2MSOxF + S
MSOiF + F2 SO$z + blF
MSO2F + Clz = SOzClF + MCl
4MS02F
2MS02F
128
CEUHGE H . CAUY
V. Other Inorganic Compounds Containing Sulfur and Fluorine
A. COMPOUNDS
CONTAINING
PHOSPHOHUS,
SULFUR,
AND FLUORINE
1. Phouphorua Thiojtuoride, PSF,
Thorpe and Rodgers (299) (1888) produced phosphorus thiofluoride

by heating a mixture of PbF2and P2Saor a mixture of PbF2, S, and P. They
also uscd the reaction
Others have used modifications of the first of the above procedures (183,
313). The Swarts reaction involving PSCl, or PSBr3 and SbF, has also been
used (18, 29).
PSF3 is a colorless gas which burns in air. The products are said t o be
PF,, SO2,and, probably, P0F3(299). Another author gives PF,, P20,, and
SO2as the products (211).Phosphorus thiofluoride reacts slowly with water
and more rapidly with a solution of sodium hydroxide (183, 299). The
hydrolysis goes by steps giving HF and HzS (or F- and S= or HS-) and
phosphoric acids (or their anions) such as those shown in the hydrolysis
scheme proposed by Lange arid Askitopoulos (183),
OH
SPFa 4 SP--F
-+
' g
OH
OH
SP-OH
OH
-+
\F
S1'-OH
-+
OP-OH
O
'H
\OH
By the reaction with cold NaOH Solution Lange has obtained thc salt
0
NaS-P-F
(182). Ammonia reacts with phosphorus thiofluoride to give
F
(NH&PSF, a white solid (183, 299) that hydrolyzes in moist air to form
[PS(NH2)2]0H(211). PSFBattacks some metals (313) and hot glass (299).
The vapor pressure of PSF3is given by the equation
loglo P,,
= 7.5882
- 1038.8/T.
This gives a heat of vaporization of 4684 kcal/mole. The critical temperature is 72.8" and the critical pressure is 37.7 atm (18).
Electron diffraction (290) gives a structure having the parameters:
r P F = 1.51 f 0.02 A, r& = 1.85 f 0.02 A, L F P F = 99.5 f 2", LFPS
= 118 f 2'. From the microwave spectrum (131, 322) the parameters
are: r p S = 1.87 f 0.03 A, rPF= 1.53 f 0.02 A, LFPF = 100.3 f 2".
129
Some thermodynamic properties have been calculated from the infrared

spectrum (336).
6. Phosphorus ThiochloroJEuorides
Booth and Cassidy (1940) produced PSCIFz and PSClzF by the reaction
of PSCl3 with a mixture of antimony trifluoride and antimony pentachloride
at about 75" (18). These compounds hydrolyze in moist air arid react
rather slowly with potassium hydroxide solution. They do not ignite spontaneously in air as readily as does PSF,; however, a mixture with air is
spontaneously explosive (18).
Equations for the vapor pressures are:
log,, P,,
log,, P,,,
7.5100 - 1292.7/T (for PSCIF,)

= 7.6596 - 1613.9/T (for PSC12F)
The heats of vaporization per mole are 5703 cal and 6863 cal for PSClF2
and PSC12F, respectively. Phosphorus thiochlorodifluoride has a critical
temperature of 166.0' and a critical pressure of 40.9 atm.
3. Phosphorus ThiobromoJEuoricles
Booth and Seabright (1943, 26) produced PSBrFz and PSBr2F by the
reaction of PSBr3 with SbF3 at about 65". By the same sort of reaction
Delwaulle and Francois (66) have prepared PSClBrF from PSCIBr2. These
compounds hydrolyze slowly in basic solutions. They are not highly reactive in general but few chemical reactions are known (22).
Structures for two of these compounds are known to resemble that of
PSF3. The parameters as determined by electronodiffraction (36) are: for
PSBr2F,rps = 1.87 f 0.05 A, TPF = 1.50 f 0.10 A, TpBr = 2.18 =I=0.13 A,
LBrPBr = 100 f 3"; for PSBrF,, r p s = 1.87 f 0.04 TPF = 1.45 f 0.08
A, TPBr = 2.14 f 0.05 A, LFPBr = 106 f 3". Equations for the vapor
pressures of the same two compounds are:
logloP,,
log,, P,,
7.4674 - 1827.3/T (for PSBr2F up to 760 mm)

= 7.6970 - 1484.8/T (for PSBrF2 up to 760 mm)
=
The heats of vaporization per mole arc 8351 and 6775 cal for PSBrZF and
PSBrFz, respectively (22).
4. Phosphorus DiJEuoroisothiocyanate, PFzNCS

In 1947 Anderson ( 3 ) reported the preparation of PFzNCS by the reaction of P(NCS)3 with SbF3 at about 175". An equation for the vapor pressure is log,, P m m = 7.7045 - 1752.9/T and the heat of vaporization is
8080 cal/mole. The compound is completely miscible with carbon disulfide
and its stability is "fairly high" (3).
130
GEORGE H . CADY
8
6. DiethyZ Fluorothiophosphate, (C2H60)J'F
S
White (1952, 319) produced the ester, ((
s s
60)Zp
', by
warn .ng
together hydrogen fluoride and (CZH,0)2POP(OC2H6),. The product is

soluble in ethanol, acetone, and ether. It is slightly soluble in water and it
hydrolyzes but does not attack glass. The odor is nauseating but the toxicity
is not particularly high.
B. COMPOUNDS
OF SULFUR,
NITROGEN,
A N D FLUOKINE
S4N4F4is prepared (Glemscr el al., 1955, 109, 110) by adding AgF2 to
a solution of S4N4in carbon tetrachloride arid then warming the mixture
t o a temperature near its boiling point for about 15 minutes. This process
gives a red solution from which white crystals of S4N4F4may be separated
as the solvent is rcmoved. Upon being warmed, the compound starts to
decompose a t about 128' and it becomes completely molten a t 153'. Its
solubility in carbon tetrachloride a t 20' is 3.44 gm/liter. An X-ray study of
the solid indicates a tetragotal structure and a space group of Pz2lc-D$d
with a = 9.2 and c = 4.3 A. The unit cell contains two formula weights
of S4N4F4.
A molecule has a dipole moment of 0.
The compound hydrolyzes slowly in moist air. With warm NaOH
solution it gives NH,OH, F- and SO,". This iridicates an oxidation riurriber
of $4 for sulfur. All of the above properties are in accord with the proposed
structure
F-&=N-,C-F
11
SNzFzis a colorless gas formed by refluxing a solution of S4N4in carbon

tetrachloride with AgFz for a period of about two hours (109, 110). The gas
reacts rapidly with warm sodium hydroxide solution forming NHIOH,
F-, and SO,-. It also reacts with mercury and stopcock grease. Vapor
pressures a t a few temperatures are:
Vapor pressure (mm of Hg)
Temperature
(OK)
16.0
191
141.6
281.9
445.4
668.9
226
239
249
259
131
FLUORINE-CONTAIKISG COMPOUNDS O F S U L F U R
+ -
The proposed structure is F-N'S-N-F

(109, 110). At its boiling point
SN2F2 decomposes into NP, NSF, SNFI, and another colorless liquid compound of S, N, and F melting a t -20' and boiling a t -60" (108).
SNF in an impure form presumably containing SN2F2 has been obtained
by the fractional distillation of crude SN2F2(109, 110). The composition
of the molecule has been learned by analyzing (for S and F) and measuring
the densities of several samples of the impure material. Hydrolysis in
NaOH solution gives NH40H, F-, SO3', and S=. As the impure gas stands
for a long time, green-gold crystals of solid S,N,F, collect on the walls of
the glass container. The proposed structure of S N F is S=N-F.
SNF3 is obtained by passing a gaseous mixture of SNF and SNZF, over
AgF2a t about 20". (Glemser and Schroder 1956, 111). The thermal decomposition of SNzF2 a t its boiling point proceeds slowly and forms Nz, SNF,,
-20", bp -60") which contains the
NSF, and a colorless liquid (mp
elements N, S, and F (108).The compound SNF3 is a colorless gas which is
hydrolyzed by water. With NaOH solution it gives N H 4 0 H , F-, and
SO3=. A heat of vaporization for the liquid of 5210 cal/mole is found from
vapor pressure measurements. The proposed structure is FzS=NF.
NSF is formed along with SiF, and N2 when SNzFPis decomposed in a
quartz glass vessel a t 250" (108). This compound is considered to be an
It is isoelectronic with
isomer of SNF and t o have a structure N=S-F.
SOz t o which it bears a physical resemblance. The gas is colorless and it
reacts with NaOH solution to give NH,OH, F-, and SO3=. By hydrolysis
the isomer, SNF, gives NH4 0 H,F-, SOa' and S=. A few vapor pressures are:
Vapor pressure (mm of ~ g ) 21
Temperature (OK)
1213.2
59
220.9
104
158
258
645
230.7
239.3
250.5
273.2
The heat of vaporization is 5300 cal/mole.
C. COMPOUNDS
CONTAINING
SULFCJR,
ANTIMONY,AND FLUORINE
(SbF&S is obtained (Aynsley et al., 4a) by the reaction of sulfur with
antimony pentafluoride. I t is a white solid and is stable up to about 200".
The related compounds (SbF5)Se and (SbFK)5Teare also known.
is formed by combining antimony pentafluoride with sulfur
S b F K SO2
.
dioxide (4a). It is a white solid which melts a t 57" to a mobile colorless
liquid. It reacts rapidly with water forming sulfur dioxide. Arsenic pentafluoride does not combine with sulfur dioxide.
132
GEORGE H. CADY
D. SULFUR CONTAINING COMPOUNDS FORMED

BY
OF BF3
THE:
ADDITION
1. Compounds in Which BFs I s Loosely Bound
Studies by H. S. Booth and co-workers of the freezing points for systems

involving boron trifluoride have shown the existence a t low temperatures
of solid compounds of low stability. They decompose easily to give BFB
and the other components. The compounds which contain sulfur are given
in the tabulation.
Formula
BFa 7H2S
BFs * HeS
BFa SOFr
BFa * SO?
2. SF4 * BF3 (6,56)
This compound is discussed in Section 1I.C.

3. Boron TriJEuoride and Sulfates
Baumgarten arid Miiller (1936, l o ) , also Baumgarten and Hennig

(1939, 9), found that boron trifluoride combines with certain sulfates at
about 250-300' to form the compounds Na2S04 BF3, K2S04 BF3,
Cs2SO4 2BF8, and TlzSOI * BF3. The potassium salt becomes completely
molten when heated to 240" and a t about 260' it decomposes. The salts are
white crystalline solids a t room temperature.
A related compound, KBF4 4SOa (Baumgarten, 1940) is formed by
adding sulfur trioxide in excess to potassium fluoroborate a t 45-70'. After
removal of the excess SO3 white crystalline KBF4. 4803 remains (8).
When this solid is heated, it begins to decompose at about 65" forming SO3
and S308F2 which distil away (186). Proposed structures are
oso
O F 0
K OSFBFSO
O F 0
O O F O O
(8) and K(FSOSOBOSOSF1
oso
(186).
O O F O O
E. THE COMPOUND
SC1, 2AsF3
Sulfur tetrachloride and arsenic trifluoride combine readily to form
SC14 2AsF3 (Ruff and Thiel, 1904, 261). The substance is a pale yellow
solid which is very hygroscopic. It hydrolyzes rapidly in water or NaOH
solution and attacks organic matter. Glass is attacked only slowly.
FLUORINE-CONTAINING COMPOUNDS O F SULFUR
133
VI. Compounds Containing Carbon, Sulfur, Fluorine, and Elements

other than Oxygen
A. COMPOUNDS WITH FLUORINE

BOUNDTO CARBON
T O SULFUR
BUT
NOT
1. Thiocarbonyl Fluoride, CSFz
The literature regarding thiocarbonyl fluoride, CSF2, is confusing. It

is doubtful whether the pure compound has been isolated; consequently,
its existence is not certain. There is a report (30) that the compound was
prepared by 0. Ruff using the reaction of IF6 with CSZand that it melted
a t - 136" and boiled a t -40". When the same reaction was tried by Haszeldine and Kidd (1953,138) no CSFz was found as a product, but the use of
partially hydrolyzed IFs caused the formation of SOFZ, a substance having
properties like those reported for CSFz. Bennett et al. (1950, 13) and
Brandt et al. (27') reported CSFz to be formed by heating sulfur with CF31,
but further work by Haszeldine and Kidd (138) with the same reaction
was not reported to give thiocarbonyl fluoride. Still later (1955) Haszeldine
and Kidd (141) said that they obtained a product thought t o be impure
CSFz by the reaction of CF3SH with NaF a t room temperature. The product was not formed by the same reactants a t 55", but it was assumed t o
be an intermediate product of the hydrolysis of CFZSH (141). In 1953
Tyczkowski and Bigelow (310) described a n impure sample of a material
considered to be thiocarbonyl fluoride. It was a sinall part of the product
resulting from the action of fluorine upon carbon disulfide and it had a
boiling point of -46" and melting point of - 134". The authors expressed
uncertainty about the formula for the compound.
2. TriJuoromethanethioE, CF3SH
The compound CF30H has riot been prepared. Instead, one obtains
COFz and HF.
Hydrogen chloride reacts with (CF3S)zHg to form mercuric chloride
and CFBSH (138). The compound reacts with NaOH solution to give COO=
F-, and S= ions. Under prolonged irradiation by ultraviolet light it decomposes into Hz, CHF3, (CF&S, (CF3S)z,and S (138). As CF3SH reacts with
S
an excess of water, a yellow liquid (CF3SCF) is formed a t first. Carbonyl sulfide is also produced. With a little water in a glass vessel the products
include COS, SiF4, and, perhaps, CSF,. At room temperature sodium
S
fluoride and CFZSHreact in glass to form CF3SCF, COS, SiFI, and, per-
134
GEORGE H. CADY
haps, CSF2. The latter is
iiot
ot)taiued a t 55". Equal numbers of moles of
S
CFaSHand anhydrous NH3react in a glass vessel to give CF&I?, COS, SiF,,
(CF,S)tCS, NH4F, CFBCNHt?, and, perhaps, CSF2 (138).
S
3. Other RfSH Compounds
Known compounds of this type are C3F7SH (141) and H(CF,CFt),C H 8 H (94).
4. TrifEuoroniethanesulfenylChloride, CF3SCI
When (CF3S),Hg reacts with chlorine in a glass tube at about -20"
the products obtained are: HgClt, CF3SHgCl, (CF3S)2,and CF3SCI (Haszeldine and Kidd, 1953, 138). When the temperature of the reaction is increased to 20" the yield of CF3SCl is reduced. Under the influence of ultraviolet light (CF3S)2reacts with Clt to form CFBCl. Trifluoromethanesulfenyl chloride reacts with NaOH solution. With mercury it gives HgClt
and (CFfi)2. With CF3SH it slowly gives HCl and (CF,S)t. There is a
slow reaction with H2S which forms CF3SSSCF3.The compound CF3SCl
has a golden yellow color and vapor pressures as shown:
Vapor pressurc (mm of Hg)
Tcmperaturc ("C)
262.2
-25.1
316.0
-21.3
446.6
-13.7
740.3
760.0
-1.3
-0.7
6 . Other H.fSCI Compouiids
Knuriyaiits and Fokin (1955,171) have prepared ClCF2CF8C1by combining SzC12 or SClz with C2Fl at about 110" and by the reaction of Clz
with (ClCFtCF2S)t at 100". Several chemical and physical properties are
known. The reaction of SC12 with C2F4 produces both CICFtCF2SCI and
CICF2CF2SSCIfor which physical constants are known (171).
The compound CICFzCHzSCl is formed when chloriiie reacts with
(CICF2CH2S)2(169). Physicoal constants of CICFzCH2SClare known as are
several of its chemical reactions and the nature of the compounds formed
by the reactions (169). For example, CICF2CH2SClreacts with (CtH6)tNH
in (CtH&O to form CICFtCHSN(C2H6)t.
6. Bis(trifluorornethyl) sulfide, (CF3)tS
The ultraviolet irradiation of (CF3S)2in a silica tube produces sulfur

arid bis(trifluoromethy1) sulfide (Brandt et al., 1952,26). I t is also obtained
by the irradiation of (CF3S)2S(158) and by t,lie thermal decomposition of
FLUOHINE-CONTAINING
COMPOUNDS OF SULFUR
135
(CF3S)2 or (CF3S)ZHg (141). The compound strongly resists hydrolysis by

NaOH solution being scarcely, if a t all, attacked even a t 15OOC. Its vapor
pressure follows the equation
loglo P,,
= 7.816
- 1239.1/T
This corresponds to a heat of vaporization of 5650 cal/mole (26).Structural

garameters obtained by $ectron diffraction are (24): TCF = 1.328 f 0.001
A, rcs = 1.828 A 0.015 A,LCSC = 105.6 f 3".
7 . Other I ~ z Compounds
S
Containing F Atoms in the Radicals
Among the products of reaction of C3F71with sulfur is a small yield of

a material, bp about go", thought probably to be (CYF7)2S(Hauptschein
and Grosse, 1951, 130). A similar reaction produces [H(CF2)6CH2]2S(94).
An olefin such as CzF4,C2CIF3, C2BrF3, C2F31reacts with SClz or S d X
under pressure a t about 100-150" to give sulfides, disulfides, and trisulfides (Raasch, 1948, 245; Knunyants and co-workers, 1955, 169, 171)
including (CF&ICF2)&. Compounds of the type R2S are also prepared
by the reaction of an olefin such as CH2=CF2 with a sulfenyl chloride such
as CICF2CH2SCl (169, 171). By this means the compounds (ClCFzCH2)2S,
CICFZCH~SC~H~OCI,
CICFZCH~SCZHF,,
CICHzCH2SCHzCF2CI,and others
have been produced. Some of their chemical and physical properties are
known (169). The compound CICFzCHzSClreacts with C6H6MgBrto give
ClCFzCH2SCsH6, and C1CF2CH2Brreacts with Na and CzH6SH in methanol to form CzH6SCF2CH2Br(169). Still another method for making
R2S-type compounds is the reaction of RSH with an olefin in the presence
of an alkali (170, 172). Compounds of the type RISRPare known in which
RI is CHB,C ~ H Si-C3H7,
,
or C6H6, and Rz is -CF2CFC1H, -CFCICFCIH,
or -CFzCFzH. The compounds CH3SCF=CFCI and C2H6SCF=CFCI
and CF~CHFCFZSCH~
are also known (170, 172).
Another method for preparing RISRz compounds is by the reaction of
SbF3 and SbcI6 with a sulfide containing chlorine (308). For example,
CHYSCC13is converted to CHYSCF3.Other compounds produced in this
way are CF3SCH2Cl, CF2CISCH2Cl, CF3SCFzH, and CH3SCF2C1.Some
of their physical constants are known.
8. Bis(trifEuoromethyZ)polysulfides, CF:{SSCF,,CF3SSSCF3, CF3SSSSCF3
The reaction of CFJ with sulfur a t about 260" produces CF3SSCF3,

bis(trifluoromethy1) disulfide (Bennett e l al., 13,26,85).This same reaction
a t about 300" also produces small yields of CF3SSSCF3,bis(trifluoromethy1)
trisulfide, and CFBSSSSCF3,bis(trifluoromethy1) tetrasulfide (Haszeldine
and Kidd, 1953, 1%). Rot,h the disulfide and the txisulfide are also formed
136
(;EOIIGP H . C A U Y
by the reaction of IF6 with CSz at 195" (138). Other reactions which produce (CF3S)2are: (a) ultraviolet irradiation of CFSSH, (b) reaction of C12
with (CF,S),Ig, (c) reaction of I-Ig with CFaSCl, (d) reaction of CF3SC1
with CF3SH. (e) ultraviolet irradiation of CF3SC1,(f) hydrolysis of CF3SCI
(138, l4O), (9)reaction of mercuric fluoride with carbon disulfide a t 250"
or more (220).Bis(trifluoromethy1) trisulfide is f ormcd as hydrogen sulfide
reacts slowly with CF3SC1(138).
These polymlfides are stable to air, water, and mercury at 20" but they
react with warm NaOH solution (138). Thermal decomposition of (CF3S)Z
at 320" gives CSz, CF.,, and (CF3)?;S(141). Ultraviolet irradiation of
CF3SSSCF3gives S and CF3SCFs (1%). When irradiated with ultraviolet
light, (CF3S),reacts with mercury to givc (CF3S)tHg (26) and with chlorine
to give CFsSCl (1%). At 360" chlorine and (CF3S)2 react in the dark to
form CF3C1and S2C12.Cobalt trifluoride and (CF3S)Z produce CF3SFs (ZG).
The reaction of (CF3S)2with concentrated sulfuric acid in glass a t 120"
gives S, S02, COP,and SiF, (139).
The vapor pressure of (CFaS)2follows the equation:
loglo P,,
= 7.765
- 1506.1/T
From this the heat of vaporization is 6880 cal/mole (26).

Structural parameters found by electron diffraction are given in the
tabulation.
(CFaSh
(CFaS)zS
1.334 kO.013 1.829 f0.017 2.053 f0.019 105.4f3"

1.340k .011 1.848f .015 2.068It .016
S---S
103.8 *3"
CFa
FsC'
9. Other PoEysuQ5des Containing C--F
Groups
CsFJ reacts with sulfur at about 250' forming C3F,SSC3F,, bis(pcrfluoro-n-propyl) disulfide, bp 122.2", and C3F7SSSC3F,, his(perfluoro-npropyl) trisulfide, bp 153" (Hauptschein and Grosse, 1951, 130). (For
similar work see references 94 and 141 .) Disulfidcs are also produced by
the reaction of SClz or S2CI2at about 100-150" with olefins such as C2F4,
C2CIFs,etc. (Raasch, 1948, 245; Knunyants and co-workers, 1955, 169,
171). The known polysulfides which have been prepared by the latter
method include (CICF&F2S)2S, (ClCF2CF&3)2, (CICF&HZS)z, (CICFX-
FLUORINE-CONTAINING
137
COMPOUNDS OF SULFUR
CHzS)2, and ClCF&FzSSC6H,oCl. Properties of some of these are known

(169, 171).
10. TriJluoromethylthiometal Compounds, (CF3S)zHg,etc.
When trifluoromethyl disulfide (CF3S)2is in contact with mercury and

the mixture is irradiated by ultraviolet light, the two substances combine
to form bis(trifluoromethylthio)mercury, (CF3S)zHg (Brandt et al., 26).
Bis(heptafluoropropyIthio)mercury, (C3F,S),Hg has also been produced
(141). The compound, (CF3S)zHg, is also formed by the reaction of mercuric fluoride with carbon disulfide a t about 250". (CF3S)2Hg melts a t
37-38" (220).
When (CF3S),Hg and AgN03 react in aqueous solution trifluoromethylthiosilver, CF3SAg is formed. A similar copper compound is made a t 100150" by heating a mixture of powdered copper with (CF3S)2Hg.Extraction
with (CZH&O followed by evaporation of the solvent gives CF3SCu. When
this compound is recrystallized from CH3CN the solid has the formula
CF3Cu . CH3CN (220).At 50" CSClz reacts slowly with (CF3S)zHgto form
CF3SHgC1, trifluoromethylthiomercuric chloride and (CF3)zCS3. Pyrolysis
in glass of (CF3S)zHg a t 220" gives COS, CF3SCF3, COZ, SiF4,and perhaps,
CSFz (141).The reaction of (CF3S)2Hgwith HCI a t room temperature gives
HgClz and CF3SH, and the reaction with Clz a t about -22" gives (CF3S)2,
CF3SCI, and HgClz (138).
11. BistriJluoromethyl Trithiocarbonate, (CF3S)ZCS,
S
and Trijluoromethyl Fluorodithioformate, CFsSCF
When equal numbers of moles of NHs and CF3SH are mixed in a glass
vessel a t -80" and the mixture is then allowed t o warm to room temperature, the reaction which occurs produces CF3SCF, (CF3S),CS, COS, SiF,,
S
NH4F, and, perhaps, CF3SCNH2 and CSFz (Haszeldine and Kidd, 1955).
S
S
The compound (CF3S)&S is also produced by the reaction of CF3SCF with
(CF3S)zHga t 70", and by the reaction of (CF3S)2Hgwith CSClz a t 50".
(CF,S)zCS is a red liquid which docs not react with water a t room temperature but does react a t 80" giving COS as a product. It reacts with 15y0
NaOH at room temperature forming F-,S=,and other products.
S
CF3SCF is a yellow liquid which reacts with water a t 75" forming
CF3SH, COZ, and COS. With NaOH solution it gives F-, S=,and other
products (141).
138
GEOHGE E I . CADY
R. COMPOUNDS
WITH FLTJORINE
ROUNDTO SULFURAND C A R ~ O N
I . Derivatives o j Sulfur Tetrafluoride
Tyczkowski and Bigelow, 1953, obtained CF3SF3 (trifluoromethylsulfur
trifluoridc) arid F8SCF2SFs together with CF4, SFe, SF,, CF,SFa, S2FI0,
SF&F2SF6, and, perhaps, CSF2 by the reaction at about 48" of carbon
disulfide with fluorine highly diluted by nitrogen (310). CF3SF3has also
been formed by the reaction of (CF3)2SF4 with CF,N=CF2 a t about 520"
(73).
CF,SF, resembles SF, in its reactivity. It attacks glass slowly a t room
temperature and reacts with aqueous base (73).
The electrolysis of hydrogen fluoride containing carbon disulfide and a
little sodium fluoride gives a gaseous product containing CF3SF6,CF2(SF&,
and CF,(SF&, difluoromethyleriebis(su1fur trifluoride (51).
2. Derivatives o j Sulfur Hexajluoride
( I . TriJluorovtiethylsulfur PentriJluoride, CF38Fs. The cornpourid CF3SFs
(Silvey and Cady, 1950,279) is prepared by: (a) the electrolysis of hydrogen
fluoride coritaiiiing carbon disulfide (51, 280) or dimethyl sulfide (73), (b)
the reaction of cobalt trifluoride with carbon disulfide, methyl mercaptan
(279), or his(trifluoromethy1) disulfide (as), (c) the reaction of fluorine
with methyl mcrc:tptan in the presenre of a porous mass of copper coated
with fluorides of silver (W79), (d) the reaction of fluorine with CS, (310).
In the third process a compound of the empirical formula CSHF? is formed
(279).
Trifluoromethylsulfur peritafluoridc rcscmblcs sulfur hexafluoride by
being an excellent gaseous electrical insulator (104, 105, 123, 124) and by
being highly resistant toward hydrolysis. Like sulfur hexafluoride it reacts
with a hot alkali metal. I t is more easily decomposed by a. spark than
SF6 and the products are CF, and SF, (279). When the gas ;s passed through
n nickel tube a t 450" there is substantially no decomposition; a t 500" there
is a slow decomposition, or reaction with the nickel tube, to give CZFB and
SF4. As a mixture of CFBSFb and perfluoropropylenc, C3F6,passes through
a nickel tube a t about 500-520O the reaction produces SF4, CPe, C6FI2
(three isomers), C6F14, and C7Fl, (72).
The compound CSHF, reacts readily with NaOH solution to give Farid other products. It decomposes in a spark-over electrical discharge to
give CF4, CHF3, and SF4 (279).
For CFsSF6, the vapor pressure is given by the equation log,, P,,,
= 6.71988 - 757.795/T - 53,771.9/T2; while the corresponding equation
for CSHF, is log,, PIlllr,
= 6.64570 - 759.863/T - 80,107.2/T2. The aorre-
139
sponding heats of vaporization per molc are 5,380 and 6,100 cal/mole (279).
The moment of inertia (167) found from the microwave :pectrum of
CF3SF6corresEonds to the structural parameters: f S F = 1.57 A, TCF = 1.35
A, rcs = 1.86 A, LFCF = 107" 30',LFSF = 90'.
b. Other R,SF6 Compounds (73, 147). The electrochemical process is
used to produce CzF6SF6 (bp, 135"), C3F7SF6 (bp, 42"), C,F~SFE,(bp,
68.7"), and (C2F6)2NCFzCF2SF6
(bp, 123'). Some properties of these compounds are known.
c. DiJluoromethyZenebis(sulfurpentajluoride), F&CF2SF6.A small yield
of F6SCF2SF6results from the action of diluted fluorine upon carbon
disulfide (Tyczkowski and Bigelow, 1953, 310). The same substance is
formed in small yield by the electrolysis of CS2 in hydrogen fluoride containing sodium fluoride (Clifford et aZ., 5 1 ) .
d. SF5CFzSF3(see Section VI.B.l).
e. Bis(triJEuoromethyZ)suZfurTetraJEuoride, (CF3)2SF4and Other R.f2SF4
Compounds. A compound of the formula (CF3)2SF4is one of the products
obtained in small yield by the electrolysis of (CH&S in hydrogen fluoride
TABLE V
SYNTHESIS
AND PROPERTIES
OF R#F4 COMPOUNDS
Organic solute
in H F
R,,SF4 compound of Boiling point

related structure
("C)
70
116
154
70
n~z6
d25
Reference
1.2753 1.836 73, 147

1.2856 1.865
147
1.2943 1.903
147
1.2766
147
C H ~ C H ~ S C H ~ ~CFzCFzSF4CFzCFz
H~
'
80
CHzCHzOCHzCHzS CFzCFzOCFzCFzSF~
(C~F~)ZNCF~CFZSF;123
[ (CZHK)~NCH~CH~SIZ
1.3015 1.859
1.2921
147
147
(51, 73). When a mixture of the compound with CF3N=CF2 is heated to

520" the reaction products include CF4, CzFs, SF4, and CF3SF3 ( 7 3 ) .
(CF3)&3F4
reacts with CF,CF=CF, a t 518" to form SF,, C3F8, and Cd?,z
(mostly neo-CsFd ('74).
The electrochemical process has also been used to obtain the derivatives
of sulfur hexafluoride shown in Table V.
The nuclear magnetic resonance spectra of R,,SF4 and of RfSF6 compounds indicate: (1) the fluorine nuclei of the Rf groups are more shielded
by electrons than those of the -sF6 and =SF4 groups; (2) the -sF6
groups have four identical fluorine atoms and one (presumably at the apex
of a square pyramid) which is different; (3) in Hf,SF4 there is only one
resonance peak for the =SF4 fluorine atoms; therefore, the two Rf groups
140
GEORGE H . CADY
are attached a t 180" from each other; (4)in CFzCFZ0C~F2CF2SF4

there are
two sets of triplet peaks for the =SIP4 group; therefore, the angle L CSC is
about 90" rather than 180" (225).
VII. Perfluoroalkanesulfonic Acids, Their Anhydrides, Salts, and Esters
A. TRIFLUOHOMETHANESULFONIC
ACID,CF3S03H
The compound CF3S03Hwas first reported by Haszeldine and Kidd
in 1954 (139). I t is produced by adding concentrated sulfuric acid to
Ba(CFaS03)2,or some other salt of trifluoromethanesulfonic acid, and then
distilling off CF3S0,H (29, 36, 114, 159). The procedures differ in the
method for first obtaining the salt. The known methods are: (1)electrolysis
of hydrogen fluoride containing CH3S02F or CH3S02C1 (29, 36, 114) to
give CF3SOZF which is later caused to react with a base t o give a salt such
as Ba(CF3SO& or KCFsS03, (2) reaction of hydrogeri peroxide with
(CF3S)zHg to give an acidic solution from which Ba(CF3S0& is obtained
throughneutralization by BaC03(139), ( 3 ) reaction of a base with CF3S02Cl
(140).
Some properties of trifluoromethanesulfonic acid are like those of
fluorosulfonic acid. For example, the boiling points are almost identical,
166' and 163", rcspectively, and both are very strong acids when dissolved
in water (29, 116, 139). The acid is stable up to at least 350" but at 650" it
decomposes into S02F2,COF2, CHF3, and COZ.It is soluble in water, alcohols, ethers, and ketones (116). With water it gives the compound CF3S03H
. H2O (36, 116) which can be distilled unchanged, its vapor pressure being
15 mm at 120". It reacts with PCla to form CF8S02Cl (29, 116).
The anhydride of trifluoromethanesulfonic acid is prepared by the reaction of the acid with phosphorus pentoxide (36, 116) or phosphorus pentachloride (29). The aiihydride is a good esterifyirig agent (116). It reacts
(69).
with aniline in (C2H&0 to give CF3S03H . NH2$ and CF~SOZNHC#J
Trifluoromethanesulfonamidc, CF3S02NH2 (69,Ii'd), melts a t - 119".
B. ESTERSOF TRIFLUOROMPTHANESULFONIC
ACID
CFSSOaH and ethylene react a t room temperature to form the ester
CF3S03C2Hs and a low polymer of C2H4 (116). The ethyl ester is also obtained by the reaction of the acid with ethanol. CFaS03Hcombines with
(C*H&O to give [ (C2H&0H+CF3S03-] which decomposes when heated
forming CF3SO3CzH6(116). A general method for preparing esters of perfluoroalkane sulfonic acids is to allow the silver salt to react with an alkyl
iodide (29, 11.6).
FLUORINE-CONTAINING COMPOUNDS OF SULFUH
141
Esters of perfluoroalkane sulfonic acids are alkylating agents. For

example, CF3So3CZH6reacts with benzene to give CeH5C2H5and CaH4(C2H6)2 (116).
C. SALTSOF TRIFLUOROMETHANESULFONIC
ACID
Methods for preparing salts have been given in Section V I L A . Salts
which are known include NaCF3S03(mp, 248') (114), KCF3S03 (mp, 230")
(29), AgCF3S03,and B~(CFJSOJZ
(mp above 370") (114). The silver salt
is soluble in benzene (139) and the barium salt is soluble in water, acetone,
and alcohol. It does not act as a strong electrolyte in water (139).
D .
OTHERPERFLUOROALKANESULFONIC
ACIDS AND DERIVATIVES
Numerous perfluoroalkane sulfonyl fluorides, RfSOzF, are prepared by

the electrolysis of sulf onyl chlorides in liquid hydrogen fluoride. From these,
the corresponding perfluoroalkanesulfonic acids are obtained by : (1) reaction of the sulfonyl fluoride with a base to give a salt, (2) distillation of the
acid from a mixture of the salt with concentrated sulfuric acid (29,36,115).
These acids are much stronger and more stable than hydrocarbon sulfonic acids and perfluorocarboxylic acids. The higher acids are useful as surface active agents. The fluorocarbon chains of the higher acids are both
hydrophobic and oleophobic. The potassium salts may be used as fusible
fluxes, bonding agents and heat exchange liquids, or as high temperature
lubricants (29).A few properties of known acids are: CzF,S03H, bp, 175 or
178", surface tension a t 25' = 21 dyne/cm, pyrolysis gives SO2, CF&OF,
COF2, n-C4Flo (29,115);C6Fl1SO3H. HzO, bp, 212O, mp, 129'; C6F13S03H,
vapor pressure = 3.5 mrn at 95"; n-CsF1$03H, bp, 249'; perfluoro-p-trifluoromethylcyclohexane sulfonic acid, bp, 240, stable in water up to 250"
(29);CBFi7S03H, bp, 259' (115).
Anhydrides and salts of a few of the above acids are known (29,36,116).
Alcohols of the type RfCHzOH react with p-toluenesulfonic acid to form
0
esters of the type RfCH20SC7H,. Several esters of this kind are known (94,
0
195,300)
0
E. TRIFLUOROMETHYL
TRIFLUOROMETHANETHIOLSULFONATE,
CF3SSCF3
0
CF,SCI reacts slowly with Zn(CF3S02)2at room temperature to form
CF3S02SCF3.The slow hydrolysis of C F K I by the addition of successive
142
GEORGE H. CADY
small amounts of water produces both (CF3S)2and CE'&30BCF3.Trifluoromethyl trifluoromethanethiolsulfonate is a colorless liquid which hydrolyzes
in a basic solution to give CHF3, F-, COB, and S= or polysulfide ion (140).
0
VIII. Certain Compounds Containing the -S-
Bridge
A. SIJLPONYL
FI~TJOI~IIIES
I . P~rJEuorualkanesulfonyIFluorides
Trifluoromethanesulfonyl fluoride, CFsS02F,is prepared by the electrolysis of CH,S02CI in hydrogen fluoride. The higher perfluoroalkancsulfonyl fluor des are prepared similarly. This application of the electrolytic
process of Simons (281,282) was made independently by three parties: (a)
Brice and Trott, 1956 (29); (b) Gramstad and Haszeldine, 1956 (114);
(c) Burdon et al., 1957 (36).
In these compounds the S-F bond is not very reactive. For example,
CF,SO,F reacts only slowly with water and it does not react with CzIIaOH
at 100" (114, 116) They react with bases to give salts of perfluoroalkanesiilfonic acids (L.9,36, 114). They are oxidized by NO2 a t high temperatures.
For example, CaF17802F
and NO:! react at about 550' to give C~F15COF
and
NOS0,F (275). A few reactions of CF3S02Fare (1114, 216):
(CzHshNH
CF3SOaF - CFaSOzN(CzHsh
NHa
CFaSOZF
CF~~OZNH~
C2H6ONU
CFaSOzF
CFSS03Na
+ (Cz1ls)zO
CHaMgI
CF&30,11'
CFaSOiCHa
(CxHs)aO
+ CFaSOzCH&30zCFa
Sodiiim borohydride reacts with C8F17SOzF in ether and alcohol a t 16" to

give pcrfluoro-n-octanesulfinate ion, CsFIvS02-. The S-benzylthiuronium
salt, of this ion melts a t 130" (56).
Known perfluoroalkiinesulfonyl fluorides include (29, 11,5) :
CzF6SOzP; bp, 7.5"
CaF,SOzF; bp, 36"

CdFSS02F; bp, 64"
CaFIiS09; bp, 90,
T L D * ~1.2881
,
C~FI~SOPF;
bp, 115"; n ~ " ,1.2918
C~FI~SOZF;
bp, 135"; %D*', 1.297
FLTJORINE-CONTAINIRG COMPOUNDS OF SULFUR
143
2. Other Suljonyl Fluorides, RSO,F
I n 1927 Steinkopf reported the preparation of several aromatic sulfonyl

fluorides (287). For example, he obtained benzene sulfonyl fluoride,
C6H5SOzF,by the slow addition of fluorosulfonic acid to benzene a t about
20". The crude product was poured over ice and was then extracted in
ether. After evaporating the ether, the benzene sulfonyl fluoride was refined
by steam distillation. Another method for preparing aromatic sulfonyl
fluorides is by the reaction of an aromatic sulfonic acid, such a s C6H&OaH,
with fluorosulfonic acid (202). By boiling a mixture of a n aromatic sulfonyl
chloride with a concentrated solution of potassium fluoride in water, the
chloride is converted to a fluoride (62).
Methanesulfonyl fluoride, CHBSOZF, and other alkanesulfonyl fluorides
were prepared by Davies and Dick, 1932, by refluxing CH3S02Cl over
solid ZnFa or by heating an alkanesulfonyl chloride with 70% K F solution
in water (63).
Methanesulfonyl fluoride boils at 124" (63,11.4-).It is hydrolyzed slowly
by boiling water and more rapidly by NaOH soIution giving F- ar-d the
sulfonate (63). Benzenesulfonyl fluoride boils at 203" and reacts with: (a)
NH,OH t o give +S02NH2, (b) aniline to give +SOzNH+, (c) ROH to give
+SOaR (287).
F
b!A
TABLE VI
PHYSICAL
PROPERTIES
OF VOLATILE
FLUORINE-CONTAINING
COMPOUNDS
OF SULFUR
Formula
Name
bp ("C)
Density
Trouton
constant mp ("C) at "C
(@;m/d)
no
at "C
Reference
rp
References to
ir = infrared
uv = ultraviolet
Ra = Raman
Fluorides of sulfur
StFz.?
SK?
SF4
SFs
@I0
Disulfur difluoride
Sulfur difluoride
Sulfur tetrafluoride
Sulfur hexafluoride
Disulfur decafluoride
?
-351
-38 to -40
-63.6 sub.
29.25
27.1
23.0
-121
-50.8
-54.2
1.9191 -73
1.912 (1)-50
2.515 ( ~ ) - 5 0
2.08
47, 263, SO6

ir 82, 199
8665
ir 82
34,74,96,138,310 ir 70; Ra 70
168, 2@, 870, 871 ir 83, 89, 102, 176,
177; Ra 89,329;
uv 896
ir 83,84
67,165
soaz
FOmSOzF
FO2SOOSO2F
SIOjFc
FsSOOSFs
Thionyl fluoride
Sulfuryl fluoride
Thionyl tetrafluoride
Pentafluorosulfur hypofluorite
Fluorine fluorosulfonate
Pyrosulfurvl fluoride
Peroxydisulfuryl
difluoride
Monoperoxytetrafluorodisulfur(V1) oxide
Bis (pentafluorosulfur)
peroxide
-43.8
-52?
-49
-35.1
22.6
-31.3
51.0
67.1
22.1
23.5
22.4
-129.5
- 1201
-99.6 1.808
-86.0 1.947
-158.5 1.784
-48
1.75
-55.4 1.645
-35 dec.
-95
-19
-95
0
M
?
P
U
-57
-47
-74
20
35.5
g l , 807, 117,863
11, 78, 817,306
78, 166, 174
78
ir
76
60,132, 136, 136
77,2314
ir 76
ir 50
ir 77
314
1.82
Oxyfluorides of sulfur
SOFz
SOeF2
SOF,
SFsOF
20
128
8SS;
Ra 12, 329
ir 238; Ra 12
ir 78
ir 78
SaOaFa
SOClF
SOzClF
SOzBrF
FOzSOSO&l
Trisulfuryl fluoride
Thionyl chlorofluoride
Sulfuryl chlorofluoride
Sulfuryl bromofluoride
Pyrosulfuryl chlorofluoride
120.0
12.2
7.1
40.0
100.1
20.7
22.5
21.6
-139.5
-124.7
-86.0
-65
1.86 25
1.576 0
1.623 0
2.12 21
1.797 20
186
21, 165
19
165
87
Fluorosulfonic acid and its esters

HSOSF
2AsF8 '3SOa
CHzSOaF
CBsS03F
Fluorosulfonic acid
Methyl fluorosulfonate
Phenyl fluorosulfonate
Other esters of HSOIF
162.6
141.9
92
180
-87.3 1.740 18
1.427 16
298, 211, 209

88, 221
209
180, 184
41, 184, 209, 946
Phosphorus thiofluoride. etc.

PSFs
PSClFz
PSClZF
PSBrFz
PSBr2F
PSBrClF
Phosphorus thiofluoride
Phosphorus thiochlorodifluoride
Phosphorus thiodichlorofluoride
Phosphorus thiobromodifluoride
Phosphorus thiodibromofluoride
Phosphorus thiobromochlorofluoride
Phosphorus difluoroisothiocyanate
Diethyl thiofluorophosphate
-52.3
6.3
21.2
20.4
-148.8
-155.2 1.484 0
64.7
20.3
-96.0 1.590 0
35.5
22.0
125.3
21.0
-136.9
22
-75.2 2.390 0
2f?
1.96
166
22.2
-95
ir SS6
18
1.940 0
98
90.3
18,18S, 313
18
65
1.452 20 1.4978 20 3
1.4188 25 319
R a 65
TABLE 1'1 (Continued)

Formula
Name
bp ("C)
Density
Trouton
constant mp ("C) at "C
(gm/ml)
References to
nu at "C
Reference
+
i
c,
ir = infrared
uv = ultraviolet
Ra = Raman
Compounds of sulfur, nitrogen, and fluorine

SNF
NSF
SNFI
SNzF,
4 8
-23
-11
19.1
20.8
20.7
SL'TIF,
-79
1.38 -60
-81
1 9 2 -80
-108
1.57 -80
153(d) 2 326 20
110
108
111
110
109
Compounds with fluorine bound to carbon but not to sulfur; oxygen absent
CSFZ?
Thiocarbonyl fluoride
Trifluoromethanethiol
Trifluor omet hanesul fenyl
chloride
Bis (trifluoromethyl)
sulfide
Bis(trifluoromethy1)disulfide
Bis(trifluoromethy1) trisulfide
Bis (trifluoromethy1)tetrasulfide
Bis (trifluoromethyl)
thiocarbonate
Trifluoromethyl
fluorodithioformate
- 134?
? -46, -40?
-36.7
-0.7
-22.2
34.6
4:
138
yellow 138
22.5
36
22.4
13,16,138
Trifluoromethylsulfur
trifluoride
Y
c
uv 138
uv 138
ir 87; uv 138
86.4
ir, uv, 87, 138
1.4023 20 138
uv 138
135
1.4608 20 138
uv 138
110
red 141
43.9
-i
4
F:
F
0
5.
3
-4
yellow 141
Compounds with fluorine bound to carbon and sulfur; oxygen absent

CFaSFj
ts
15, 141, 175,310
-110
73, 510
ir,
UV,
141
ir, uv, 141
CF,(SF3),
FsSCRSFs
CFaSFs
CF,(SFdz
(CFdzSF,
Difluoromethylenebis(sulfur trifluoride)
Trifluoromethylsulfur
pentafluoride
Difluoromethylenebis(sulfur pentafluoride)
Bis (trifluoromethy1)sulfur tetrafluoride
35
51
26
-- 2 0 . 4
21.3
-51
-86.9
-70
60.5
20.5
310
51, Y2, 73
2Y9
51, 310
ir, uv 51
51
ir 51
29, 116, 139
ir 139
36, 116
ir llfi
29, 36, 116
ir 116
ir 51; uv 51
Trifluoromethanesulfonic acid and derivatives

CF~SOBH
Trifluoromethanesulfonic
acid
CF3S03H. H20 Hydronium trifluoromethanesulfonate
(CF&30~)20
Trifluoromethanesulfonic
anhydride
CF3S08CH3
Methyl trifluoromethanesulfonate
CF3S02SCF,
Trifluoromethyl trifluoromethanethiolsulfonate
CF3SOzF
Trifluoromethanesulfonyl
fluoride
CFaSOzCI
Trifluoromethanesulfonyl
chloride
166, 162
120 at
15 mm
80 5, 84
45
34
99
1.3238 25 29, 114
ir 114. 116
70
1.3480 17 140
ir 140
-21.7
-23
31.6
33
22.2
29,114
ir 114
29,140
ir 140
Trifluoromethyl sulfones, etc.

CF3SOzCH3
Methyl trifluoromethyl
sul fone
CF$OzCH&l
Chloromethyl trifluoromethyl sulfone
CF2C1SO&H3 Methyl chlorodifluoromethyl sulfone
(CF&~O&CHZ Bis(trifluoromethanesulfonvl)methane
130
140
165
191
1 4 . 0 1.5141 20 1.3486 20 29,116,308

-90
21
ir 114
1.6533 20 1.3859 20 508

1.5685 20 1.4050 20 308
29
ir 116
148
GEORGE H. CADY
The hydroc~trl~oiisulfo~iyl
fluorides are poisonous (224) but not extremely toxic (236). In some dyes the presence of an -SO9 group cause8 a
lightening and brightening of the shade (236).
Further information regarding the many sulf onyl fluorides not mentioned
above may be found in the literature (62, 63, 64, 114, 120, 202, 261, 262,
246, 287, 309).
3. Per~uoroalkunesuZ~on~Z
Chlorides, RfS02C1
Trifluoromethanesulfonyl chloride is produced by : (a) the reaction of

trifluoromethariesulfonic acid with phosphorus pentachloride a t loo", (b)
the reaction of trifiuoromethanesulfenyl chloride, CF3SC1, with an excess
of chlorine in the presence of water (Haszeldine and Kidd, 1955, 140; see
also 29 and 116).
CF3S02C1 is hydrolyzed only slowly at room temperature. With hot
water it is converted to CF3SOaH within a few hours. The reaction with
NaOH solution is fast a t room temperature. It reacts slowly with zinc dust
and water to give trifluoromethanesulfinyl ion, CF~SOZ-.From the solution
the salts CF3S02Na- HzO and (CF8SO&Zn 3Hz0 have been prepared
(140). Trifluoromethanesulfonyl chloride reacts with liquid ammonia t o
form the amide, CFsS02NH2, and with aniline to give the analide (29).
4. Sulfones
Methyl trifluoromethyl sulfone, CHsS02CF3,was prepared by Truce
et al. (1952, 308) by oxidizing CHaSCF3with CrOa in glacial acetic acid at
95'. The compounds CH2C1S02CFs and CHaS02CC1F2 were prepared
similarly (308). Another method for preparing CH3S02CF3 uses the reaction of CFaS02Fwith CHaMgI in ether (116).This reaction also produces
bis(trifluoromethanesulfonyl)methane, CF3S02CI-12SO&F3.
CH3S02CF3is n colorless odorless liquid which resists hydrolysis (308).
Infrared spectra are known for CH3SOZCF3 and CFaSOnCII&302CF3(116).
REFERENCES
A. J., and Hannay, N. U., J . Chem. Phys. 21, 119 (1953).
2. Allied Chemical and Dyo Corp., General Chemical Division, Product Data Sheet
NO. SX-ti (1947).
3. Anderson, 1. H., J . Am. Chem. SOC.69, 2495 (1947).
4 . Atack, D., and Schncider, W. G., J . Phys. & Colloid Chem. 66, 532 (1951).
412. Aynsley, E. E., Peacock, R. D., and Robinson, P. L., Chem. & Ind. p . 1117
(1951).
6. Badger, R. M., J . Chem. Phys. 3, 710 (1935).
6. Bartlctt, N., and Robinson, P. L., C h m . & Znd. p. 1351 (1956).
7. Rauer, S. H., J . Phys. Chem. 58, 343 (1952).
8. Baumgarten, P., Ber. 73, 1397 (1940).
9. Htiumgartw, P., and Hennig, H., Ber. 72, 1743 (1939).
1. Ahcrn,
FLUORINE-CONTAINING COMPOUNDS OF SULFIJK.
149
10. Baumgarten, P., and Miiller, E., Rer. 69B,2688 (1936).

11. Belf, L. J., British Patent 727,062 (1955); Chem. Abstr. 49, 11970 (1955).
f2, Bender, P., and Wood, J. M., J . Chem. Phys. 23, 1316 (1955).
1.3. Bennett, F. W., Brandt, G. R. A., Emelkus, H. J., and Haszeldine, R. N., Nature
166, 225 (1950).
14. Benson, S. W., and Ellis, D. A., J . Am. Chem. Soc. 73, 2095 (1950).
16. Berg, D., and Dakin, T. W., J . Chem. Phys. 36, 179 (1956).
16. Boitner, T. E., and Hurst, G. S., Phys. Rev. 93, 1236 (1954).
17. Bonch-Bruevich, A. M., Glikina, M. V., and Hokhberg, B. M., J . Exptl. Theorel.
Phys. (U.S.S.R.) 10, 171 (1940); Chem. Abstr. 34, 7723 (1940).
18. Booth, H. S., and Cassidy, Mary C., J . Am. Chem. SOC.62,2369 (1940).
19. Booth, H. S., and Herrmann, C. V., J . Am. Chem. SOC.68,63 (1936).
20. Booth, H. S., and Martin, D. R., J . Am. Chem. SOC.64,2198 (1942).
21. Booth, H. S., and Mericola, F. C., J . Am. Chem. SOC.62,640 (1940).
22. Booth, H. S., and Seabright, C. A., J . Am. Chem. SOC.
66, 1834 (1943).
23. Booth, H. S., and Walkup, J. H., J . Am. Chem. SOC.66, 2334 (1943).
24. Bowen, H. J. M., Trans. Faraday SOC.60,452 (1954).
26. Braidwood, C. A., and Hovey, A. G., U. S. Patent 2,419,185 (1947); Chem. Abstr.
41, 4658 (1947).
26. Brandt, R. A., EmelBus, H. J., and Hmzeldine, R. N., J . Chem. Soc. p. 2198
(1952).
%7. Brandt, R. A., Emelbus, H. J., and Haszeldine, R. N., J . Chem. SOC.p. 2549
(1952).
28. Braune, H., and Knoke, S., 2. physik. Chem. (Lapzig) B21, 297 (1933).
89. Brice, T. J., and Trott, P. W., U. S. Patent 2,732,398 (1956; Chem. Abstr. 60,
13982 (1956).
SO. British Intelligence Objectives Sub-committee, BIOS, Final Report 1595, Item
No. 22; from ref. 211.
S1. The British Thomson-Houston Co., Ltd., British Patent 532,670 (1941); Chem.
Abstr. 36, 855 (1942).
S2. Brockway, L. O., J . Am. Chem. SOC.66, 1941 (1944).
S3. Brockway, L. O., and Pauling, L., Proc. Natl. Acad. Sci. U . S.19, 68 (1933).
34. Brown, F., and Robinson, P. L., J . Chem. SOC.p. 3147 (1955).
36. Buechner, W. W., Van de Graff, R. J., Sperduto, A., Burrill, E. A., McIntosh,
L. R., and Urquhart, R. C., Phys. Rev. 69, 692 (1946).
S6. Burdon, J., Farazmand, I., Stacey, M., and Tatlow, J. C., J . Chem. SOC.p. 2574
(1957).
S7. Burg, A. B., in Fluorine Chemistry (J. H. Simons, ed.), Vol. 1, pp. 90-95,
Academic Press, New York, 1950.
58. Burhop, E. H. S., Massey, H. S. W., and Page, G., Phys. Rev. 80, 107 (1950).
39. Burwell, R. L., Maury, L. G., and Scott, R. B., J . Am. Chem. SOC.76, 5828
(1954).
40. Calfee, J. D., U. S. Patent 2,570,917 (1951); Chem. Abstr. 46, 1666 (1952).
41. Calfee, J. D., and Florio, P. A., U. S. Patent 2,628,972 (1953); Chem. Abstr. 48,
1413 (1954).
4%. Camilli, G., and Chapman, J. J., Gen. Elec. Rev. 61, No.2, 35 (1948).
4s. Camilli, G., and Chapman, J. J., Trans. Am. Znst. Elec. Engrs. 66,1463 (1947).
44. Carnell, P. H., U. S. Patent 2,538,293 (1951); Chem. Abstr. 46, 2290 (1952).
46. Carpenter, F. G., Am. J . Physiol. 173, 471 (1953).
46. Centnerszwer, M., and Strenk, C., Ber. B66, 2249 (1923).
150
GEORGE; H. CADY
Centnerszwer, M., and Strenk, C., Ber. B68, 914 (1925).

Chenicek, J. A., U. S. Patent 2,486,519 (1949); Chem. Abstr. 46, 1249 (1952).
Clark, H. C., and EmnelBus, H. J., J . Chem. SOC.p. 4778 (1957).
Clark, H. C., and EmclEus, H. J., J . Chem. Soc. p. 190 (1958).
61. Clifford, A. F., El-Shamy, I<. H., Emelbus, H. J., and Haazeldine, R. N., J. Chem.
Soc. p. 2372 (1953).
62. Coates, G. E., and Glockling, F., J . Chem. Soc. p. 1376 (1951).
53. Cobinc, J. D., h'lec. Eng. 74, 520 (1955); Chem. Abstr. 49, 14403 (1955).
64. Coons, E. A,, U. S. Patent 2,611,735 (1952); Chem. Abstr. 47, 2969 (1953).
56. Cooper, F. S., lJ. S. Patcnt 2,221,671 (1940); Chem. Abstr. 36, 1328 (1941).
66. Cotton, F. A., George, J. W., and Waugh, J. S., J . Chem. Phys. 28, 994 (1958).
57. Cowen, H. C., Riding, F., and Warhurst, E., J. Chem. Soc. p. 4168 (1953).
68. Craig, D. P., and Magnusson, E. A., J . Chem. Phys. 26, 383 (1956).
59. Craig, D. P., and Magniisson, E. A., J. Chem. SOC.p. 4895 (1956).
60. Craig, D. P., Maccoll, A., Nyholm, R. S., Orgel, L. E., and Sutton, L. E., J. Chem.
SOC.
p. 332 (1954).
6f. Crawford, R. A., Dudlcy, F. B., and Hedberg, K., private communication (1959).
62. Davies, W., and Dick, J. H., J . Chem. Soc. p. 2104 (1931).
63. Davics, W., and Dick, J. H., J . Chem. SOC.
p. 483 (1932).
64. Davies, W., and Dick, J. H., J . Chem. Soc. p. 2042 (1932).
66. Delwaiille, M. Id., and Francois, F., Compt. rend. 226, 1308 (1947).
66. Dcnbigh, K. G., and Whytlaw-Gray, R., Nalure 181, 783 (11133).
67. Dcnbigh, K. G., and Whytlaw-Gray, R., J . Chem. SOP.p. 1346 (1934).
68. Diehler, V. TI.,and Mohler, F. L., J . Research Natl. Bur. Standards ( U . 8.) 40, 25
(1948).
69. Dohretsov, L. N., Ltokludy Akad. Nauk. (S.S.S.R.) US, 1547 (1948); Chem. Abstr.
43, 454 (1949).
70. Dodd, R. E,, Woodward, L. A., and Roberts, H. L., Trans. Faraday SOC.62, 1052
(1956).
7 1 . Dominion Tar and Chemical Co., Ltd., I3ritish Patent 668,283 (1952); Chem.
Abstr. 47, 2212 (1953).
7% Drcsdner, R. D., J . Am. Chem. SOC.77, 6633 (1955).
73. Dresdncr, R. D., J . Am. Chrm. Soc. 79, 69 (1957).
74. Drcsdner, R. D., Mao, T. J., and Young, J. A., J . Am. Chem. SOC.
80, 3007 (1958).
75. Dubnikov, L. M., and Zorin, N., J . Ben. Chem. (U.S.S.R.) 17, 185 (1947); Chem.
Abstr. 42, 51 (1948).
76. Dudley, F. B., Cady, G. H., and Eggers, D. F., J . Am. Chem. Soc. 78, 290 (1956).
77. Diidlcy, F. n., and Cady, G. H., J . A m . Chem. Soc. 79, 513 (1957).
78. Dudley, F. B., Cady, G. H., and Eggcrs, D. F., J . Am. Chem. Soc. 78, 1553 (1956).
79. Dudley, F. B., Shoolcry, J. N., and Cady, G. H., J . Am. Chem. Soc. 78,568 (1956).
80. Duncan, A. B . F., J. Chem. k'hya. 20, 951 (1952).
81. Earwickrr, G. A., and Fear, E. J. P., Chem. & Ind. p. 903 (1954).
82. Edrlson, D., Bieling, C. A,, and Kohman, G. T., Znd. Eng. Chem. 46, 2094 (1953).
83. Edelson, D., and McAfee, K. B., J . Chem. Phys. 19, 1311 (1951).
84. Edelson, D., J . Am. Chem. SOC.74, 262 (1952).
86. EmelBus, H. J., Bull. SOC. chim. France p. 909 (1953); also Fortschr. chem. Forsch.
2, 609.
86. EmelBm, H. J., and Wood, J. I?., J . C h m . SOC.
p. 2183 (1948).
87. Engelbrecht, A., Z . anorg. u. allgem. Chem. 273, 269 (1953).
88. Engelbrecht, A., Aignesberger, A., and Hayek, E., Monatsh. 86, 470 (1955).
47.
48.
49.
ii0.
FLUORINE-CONTAINIKG COMPOUNDS OF SULFTJR

89. Eucken, A., and Ahrms, H., 2.physik. Chem. (Lcipzig)
151
B26, 297 (1034).

30. Euckcn, A., and Bert>rani,A., Z. physik. Cheni. (Leipzig)B31, 361 (1936).
91. Euckcn, A., and Sauter, F., 2.physik. Chern. (Leipzig) B26, 463 (1934).
92. Eucken, A., and Schroder, E., 2.physik. Chem. (Leipzig) B41, 307 (1938).
93. Fanelli, R., J . Am. Chem. Sac. 70, 1956 (1948).
94. Faurote, P. D., and O'Rear, J. G., J . A m . Chem. Sac. 78, 4999 (1956).
96. Ferguson, R. C., J . Am. Chem. Sac. 76, 850 (1954).
96. Fischer, J., and Jaenckner, W., 2. angew. Chem. 42, 810 (1929).
97. Friedman, H. L., J . Am. Chem. Sac. 76, 3294 (1954).
98. Fristrom, R. M., J . Chem. Phys. 20, 1 (1952).
99. Fuoss, R. M., J . Am. Chem. Sac. 60, 1633 (1938).
100. Gall, J. F., U. S. Patent 2,482,379 (1949); Chem. Abstr. 43, 3575 (1949).
101. Gall, J. F., U. S.Patent 2,555,739 (1951); British Patent 651,668 (1951); Chem.
Abstr. 46, 7757 (1951).
102. Gaunt, J., Trans. Faraday Sac. 49, 1122 (1953).
103. Gaunt, J., Trans. Faraday Sac. 60, 546 (1954).
104. Geballe, R., and Linn, F. S., J . Appl. Phys. 21, 592 (1950).
105. Geballe, R., and Reeves, M. L.,Phys. Rev. 92,867 (1953).
106. Germann, A. F. O., and Booth, H. S., J . Phys. Chem. 30, 369 (1926).
107. Gibson, J. W., and Miller, C. F., J . Electrochem. Sac. 100, 265 (1953).
108. Glemser, O., and Haesler, H., 2.anorg. u. allyem. Chem. 287, 54 (1956).
109. Glemser, O., Schroder, H., and Haesler, H., Naturwiss. 42, 44 (1955).
110. Glemser, O., Schroder, H., and Haesler, H., 2. anorg. u. allgem. Chem. 279, 28
(1955).
111. Glexser, O., and Schroder, H., 2.anorg. u. allgem. Ciiem. 284, 97 (1956).
111a. Goddard, D. R., Hughes, E. D., and Ingold, C. K., J . Chem. Sac. p. 2559 (1950).
112. Goldstein, L., A n n , Physik. 151 9, 723 (1938); Chem. Abstr. 32, 7811 (1938).
113. Gowland, T. B., British Patent 527,713 (1940).
114. Gramstad, T., and Hasaeldine, R. N., J . Chem. Sac. p. 173 (1956).
115. Gramstad, T., and H~zeldine,R. N., J . Chem. Sac. p. 2640 (1957).
116. Gramshd, T., and Hasaeldine, R. N., J . Chem. Sac. p. 4069 (1957).
117. Gray, L. H., PTOC.
Cambridge Phil. Sac. 40,72 (1944); Chem. Abstr. 38,3902 (1944).
118. Gutowsky, H. S., and Hoffman, C. J., J . Chem. Phys. 20, 200 (1952); 19, 1259
(1951).
119. Halls, E. E., Ind. Chemist 26, 222 (1949).
120. Ham, N. S., and Hambly, A . N., Australian J . Chem. 6, 33 (1953).
121. Hammann, S. D., and Lambert, J. A., Australian J . Chem. 7, 1 (1954).
12.2. Hammann, S. D., Lambert, J. A., and Thomas, It. D., Australian J . Chem. 8,
149 (1955).
123. Harrison, M. A., and Geballc, R., Phys. Rev. 83,884 (1951).
124. Harrison, M. A., and Geballe, R., Phys. Rev. 91, 1 (1953).
126. Harrison, S. F., and Mayer, J. E., J . Chem. Phys. 6, 101 (1938).
126. Hartman, B. F., U. S. Patent 2,434,040 (1948); Chem. Abstr. 42, 2063 (1948).
127. Hartough, H. D., and Kosak, A. I., J . Am. Chem. Soe. 69, 3093 (1947).
128. Harvey, R. B., and Bauer, S. H., J . A m . Chem. Sac. 76, 859 (1954).
129. Harvey, R. B., and Bauer, S. H., J . A m . Chem. SOC.76, 2840 (1953).
130. Hauptschein, M., and Grosse, A. V., J . A m . Chem. SOC.73, 5461 (1951).
131. Hawkins, N. J., Cohen, V. W., and Koski, W. S., J . Chem. Phys. 20, 528 (1952).
132. Hayek, E., Austrian Patent 173,679 (1953); Chem. Abstr. 47, 2442 (1953).
133. Hayek, E., Csaloun, A., and Krismer, B., Monatsh. 87, 741 (1956).
152
GEORGE H. CADY
Ilayrk, E., and Czxloun, A,, Monakh. 87, 790 (19%).

Hayek, E., Aigncsbergcr, A., and Engelbrecht, A.,Monnlsh. 86, 735 (1!355).
Hayek, E., and Koller, W.,Monatsh. 82, 942 (19.51).
Hayek, E., I'uschmann, J., and Czaloun, MonalsR. 86, 359 (1954).
Haszeldine, R. N., and Kidd, J. M., J. Chem. Soc. p. 3219 (1953).
Haszeldine, R. N., and Kidd, J. M., J. Chem. Soc. p. 4228 (1954).
Hameldine, R. N., and Kidd, J. M., J . Chem. Sac. p. 2901 (1955).
Haszeldine, R. N., arid Kidd, J. M., J. Chem. Soc. p. 3871 (1955).
Heath, D. F., and Linnett, J. W., Trans. Faraday SOC.46, 264 (1949).
143. Henkel, P., and Klernm, W., 2. anorg. u. allgem. Chem. 222, 65 (1935).
144. Henkel, P., and Klemm, W., 2. anorg. u. allgem. Chem. 222, 70 (1935).
146. Hentrich, W., Hardtmann, M., and Ossenbeck, A., U. S. Patrnt 1,847,513, (1932);
Chem. Abslr. 26, 2469 (1932).
146. Hoffa, E., German Patent 464,087 (1928); Chem. Abstr. 22, 4130 (1928).
147. Hoffman, F. W., and Simmos, T. C . ; with rkck, R. B., Holler, H. V., IZats, T.,
Koshar, R. J., Larsen, E. R., Mulvaney, J. E., Rogers, F. E., Singleton, R., and
Spark, R. S., J . Am. Chem. Soc. 79, 3424 (1 957).
148. Hokhberg, 3. M., E'lsktrichestuo No. 3, 15-19 (1947); Chem. Abstr. 41, 7282
(1947).
149. Hokhberg, B. M., and Oksman, Ya. A., J . Tech. Phys. (U.S.S.R.) 11, 1058
(1941); Chem. Abstr. 56, 4382 (1942).
150. Hokhberg, B. M., and Oksman, Ya. A., J . Phys. (tJ.S.KE.) 6, 39 (1941); Chem.
Abstr. S8, 2561 (1944).
151. Hokhberg, B. M., Reinov, N. M., and Glikina, M. V., J . Tech. Phys. (U.S.S.R.)
12, 3, 8 (1942); Chem. Abstr. 57, 1528, 1303 (1943).
162. Rokhbcrg, B. M., and Reinov, N. M., Doklady Akad. Nauk. (S.S.S.R.) 70, 837
(1950); Chem. Abstr. 48, 8565 (1950).
163. Hokhberg, B. M., and Zandberg, E. Y a . ,J. Tech. Phys. (U.S.S.R.) 12,65 (1942);
Chem. Abstr. 37, 1326 (1943).
164. Hokhberg, B. M., and Zandberg, E. Ya., Compt. rend. acad. sn'. U.R.S.S. 65, 511
(1940); Chem. Abstr. 41, 2635 (1947).
156. Hollies, N. R. S., and MeIntosh, R. L., Can. J . Chem. 29, 494 (1951).
156. Huckel, W., Nachr. Akad. Wiss. Gottingen, Math.-physik. KZ. p. 36 (1940).
167. I. G. Farbenindustrie, A. G., British Patent 496, 658 (1938); Chem. Abslr. 33,
3540 (1939).
168. I. G. Farbenind., A. G., French Patent 808,991 (1937); C h m . Abstr. 31, 6833
(1937).
169. I. G. Farbenind., A. G., British Patent 479,450 (1938); Chem. Abstr. 82, 5165
(1938).
160. Iler, R. K., U. S. Patent 2,312,413 (1943); Chem. Abstr. 57, 4864 (1943).
161. Ingersoll, L. R., and Liebenberg, 1).H., J. Opt. SOC.Am. 46, 538 (1956).
161. Ipatieff, V. N., and Linn, C. B., U. S. Patent 2,428,279 (1947); Chem. Abstr. 42,
353 (1948).
16s. Ipatieff, V. N., and Linn, C. R., U. S. Patent 2,412,946 (1947); Chem. Abstr. 41,
5296 (1947).
164. Jach, J., and Hinshelwood, C. N., Proc. Roy. Soc. 8229, 143 (1955); A231, 145
(1955).
166. Jonas, H., 2. anorg. u. dlgem. Chem. 266, 273 (1951).
166. Kemp, W. E., Canadian Patent 445,094 (1947); Chem. Abstr. 42, 2823 (1948).
167. Kisliuk, P., and Silvey, G. A,, .I. Chem. Phys. 20, 517 (1952).
134.
235.
136.
137.
138.
139.
140.
141.
14s.
153
168. Klcmm, W., and ITtbnkrl, l., 2. anorg. u. allgem. Chem. 207, 73 (1932).
169. Knunyants, I. L., and Bykhovskaya, E. G., Izvest. Akad. Nauk SSSR., Oldel.
Khim. Nauk. 852 (1955); Chem. Abstr. 60, 9322 (1956).
170. Knunyants, I. L., and Fokin, A. V., Izvest. Akad. Nauk SSSR., Otdel. Khim.
Nauk. 261 (1952); Chem. Abstr. 47, 3221 (1953).
171. Knunyants, I. L., and Fokin, A, V . , Izvest. Akad. Nauk SSSR., Otdel. Khim.
Nauk. 705 (1955); Chem. Abstr. 60, 7069 (1956).
172. Knunyants, I. L., Shehekotikin, A. I.,and Fokin, A. V., Izvest. Akad. Nauk SSSR.,
Otdel. Khim. Nauk. 282 (1953); Chem. Abstr. 48, 5787 (1954).
173. Kwasnik, W., and Scherer, P., in Recent Research Work on Fluorine and
Fluorine Compounds, pp. 10, 26. U. 8, Dept. of Commerce, Washington, D. C.,
1947, quoted in ref. 310.
174. Kwasnik, W., FIAT Rev. Ger. Sci. 1939-1946, Inorg. Chern. 1, (1948).
175. Lachman, F., J . Chem. Phys. 22, 1459 (1954).
176. Lagemann, R. T., and Jones, E. A., J . Chem. Phys. 19, 534 (1951).
177. delattre, A., J . Chem. Phys. 20, 520 (1952).
178. Lange, W., Ber. B60, 962 (1927).
178a. Lange, W., in Fluorine Chemistry ( J . H. Simons, ed.), Vol. 1, pp. 167-182.
Academic Press, New York, 1950.
17B. Lange, W., Ber. B60, 967 (1927).
180. Lange, W., German Patent 532,394 (1930); Chem. Abstr. 26, 153 (1932).
181. Langc, W., 2. anorg. u. allgem. Chem. 216, 321 (1933).
182. Lange, W., German Patent 669,384 (1938); Chem. Abstr. 33, 2660 (1939).
183. Lange, W., and Askitopoulos, K., Ber. 71, 801 (1938).
184. Lange, W., and Muller, E., Ber. B63, 2653 (1930).
185. Lehmann, H. A., and Kolditz, L., 2. anorg. u. allgem. Chem. 272, (1953).
186. Lehmann, H. A , , and Kolditz, L., 2. anorg. u. allgem. Chem. 272, 73 (1953).
187. Lester, D., and Greenberg, L. A., Arch. Ind. Hyg. Occupational Med. 2, 348 (1950);
Chem. Abstr. 44, 10143 (1950).
188. Linke, R., 2. physik. Chem. (Leipzig) B48, 193 (1941).
189. Livingston, H. K., Hyde, J. L., and Campbell, M . H., Ind. Eng. Chem. 41, 2722
(1949).
190. Loudermilk, E. R., Danehower, R. G., and Miller, H. C., J . Chem. Educ. 28, 246
(1951).
191. Luchsinger, W., Dissertation, Breslau (1936), as quoted in ref. % l l .
192. Lynch, C . S., and Wood, J. E., 111, U. S. Patent 2,223,180 (1940); Chem. Abstr. 36,
1798 (1941).
193. MacCormack, K. E., and Schneider, W. G., J . Chem. Phys. 19, 845 (1951).
194. MacCormack, K. E., and Schneidcr, W . G., J . Chem. Phys. 19, 849 (1951).
196. McBee, E. T., Campbell, D. H., and Roberts, C. W., J. Am. Chem. SOC.77, 3149
(1955).
196. Mahan, J. E., U. S. Patent 2,742,474 (1956); Chem. Abstr. 60, 16876 (1956).
197. Mannheimer, M., Chemist-Analyst 46, 8 (1956); Chem. Abstr. 60, 5350 (1956).
198. March, N. H., Proc. Cambridge Phil. Soc. 48, 665 (1952).
199. Matutano, J. R. B., and Otero, C., Anales real soc. espak fis. y. quim. (Madrid)
61B, 223 (1955); Chem. Abstr. 49, 15486 (1955).
200. Mayer, J. E., J . Chem. Phys. 19, 1024 (1951).
201. McCann, H. G., and Trout, H. Q., U. S. Patent 2,562,432 (1951); Chem. Abstr. 46,
696 (1952).
154
GEORGE H. CADY
202. McCoy, G., I n n m i , C. E., :tnd Haims, P. C., U. S. 1:h:nt 2,AXA,202 (1954);
Chem. AOstr. 49, 12536 (1955).
203. McCormick, N. R., and Craggy, J. D., Brit. J . A p p l . Phys. 6, 171 (1954); Chem.
Abstr. 46, 12488 (1954).
204. McCormack, K . E., and Schneider, W. G., Can. J. Chem. 29, 699 (1951).
205. McCrath, W. D., and Ubbclohde, A. R., Proc. Roy. Soc. A227, 1 (1954).
206. McHnrd, J. A , , and Winger, R. A,, U. S. Patent 2,699,411, (1955); G e m . Abstr.
49,5257 (1955).
207. Mcslnns, M., Bull. soc. chim. France 16, 391 (1896).
2138. Meyer, E. G., and Bucll, C. E., J . Chem. Pliys. 16, 744 (1948).
209. Meyer, J., and Schramm, G., 2. anorg. u. allgem. Chern. 206, 24 (1932).
209a. Millen, D. J., J . Chem. Sac. p. 2606 (1950).
215. Miller, H. C., U. S.Patent 2,539,261 (19.51); Chem. Abslr. 46, 2800 (1951).
211. Miller, I T . C., and Gall, J. F., I T L Eng.
~ . Chenz. 42, 2223 (1950).
212. Miller, H. C., Verdelli, L. S., and Gall, J. I?., Ind. Eng. Chem. 43, 1126 (1951).
213. Mitchell, W. F., and Grant-MacKay, J. A., U. S. Patcnt 2,702,233 (1955); Chem.
Abstr. 49,8571 (1955).
214. Moffitt, W., Proc. Roy. Soc. A200, 409 (1950).
216. Moissaii, H., Ann. chim. et phys. [6] 24, 239 (1891).
216. Moinsan, IT., and Iobcau, P., Compt. rend. 130, 865, 984 1436 (1900).
217. Moissm, H., and Lebenu, P.,Com,pt. rend. 132, 374 (1901); Ann. ch,im. et. p l q s .
[7] 26, 145 (1002).
218. Mollering, C. H., J . prakt. Chem. 134, 209 (1932).
$19. Morrison, T. J., and Johnstone, N. B. B., J . Cham. Soc. p. 31555 (1955).
220. Muetterties, E. L., U. S. Patent 2,729,663 (1956); Chem. Abstr. 60, 11362 (1956).
221. Muct)tcrtics, E. L., and Coffman, D. D., J . Am. Chem. Soc. 80,5914 (1958).
222. Muetterties, E. L., I T . S. Patcnt 2,801,904 (1957); Chem. Abstr. 61, 15911 (1957).
223. Muller, N., Lauterbur, P. C., and Svat<os,G. I?., J . AVL.Chem. Soc. 79,1043 (1957).
224. Myers, D. K., and Kenip, A, Jr., Nature 173, 33 (1954).
226. Nendorffcr, J., Compt. rend. 231, 1070 (1950).
226, Neudorffer, J., Compt. rend. 232, 2102 (1951).
227. Neudorffcr, J., Compt. rend. 234, 1983 (1952).
228. Neudorffcr, J., Ann. chim. (Paris)8, 501 (1953); Ch.em. ADstr. 49, 4621 (1955).
229. Neudorffer, J., Compt. rend. 236, 706 (1953).
230. Ncwcomc, M. M., and Cudy, C. H., J . Am. Chem. Soc. 78, 5216 (1956).
2.91. Nostrand, E. D., and Duncan, A. 11. F., J . AVL.C h n . Soc. 76, 3377 (1954).
232. OConnor, C. L., J . Acoust. Soc. AWL.26, 3Gl (1954); C. A. 49,20 (1955).
233. OLoane, J. K., and Wilson, M. K., J . C/LSWL.
Phys. 23, 1313 (1955).
234. Palmer, W. G . , Bndeavour 12, 124 (1953); Chem. Abstr. 47, 9686 (1953).
256. Parker, R. P., and Hoffmann, C. M., U. S. Patent 2,427,995 (1947); Chem. Abstr.
42, 375 (1948).
236. Pattison, F. L.M., Nature 174, 737 (1954).
257. Pearson, l
G.,.and Robinson, P. L., J . Chem. Soc. p . 1427 (1933).
238. Perkins, W. D., and Wibon, M. K., J . Chem. Phys. 20, 1791 (1952).
239. Picon, M., and Domange, L., Compt. rend. 236, 704 (1953).
240. Pollock, IT. C., and Cooper, F. S., Phys. Rev. 66, 170 (1939).
241. Porter, R. W., Chem. Eng. 66, No. 4, 102 (1948).
2.42. Prideaux, E. n. R., J . Ctiein. Soc. 89, 316 (1906).
243. Prideaux, E. B. R., Chem. & Znd. 63, 351 (1934).
244. Prober, M., U. S. Patent 2,717,235 (1955); Chem. Abstr 60, 94 (1956).
FLUORINE-CONTAINING
COMPOUNDS O F SULFUR
155
24<5. Raasch, M. S., U. S. Patent 2,451,411 (1948); Chem. Abstr. 43, 6646 (1949).
246. Renoll, pul. W., J . Am. Ckeni. Sor. 64, 1489 (1942).
247. Roebuck, A. K., and Evering, B. L., U. S. Patent 2,564,080 (1951); Chern. Abstr.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
46, 1756 (1952).

Roebuck, A. K., and Evering, B. L., J . Am. Chem. Soc. 76, 1631 (1953).
Rogers, M. T., and Katz, J. J., J . Am. Chem. Sor. 74, 1375 (1952).
Rowlinson, J. S., J . Chem. Phys. 20, 337 (1952).
Ruff, O., and Thiel, K., Ber. B37, 4520 (1904).
Ruff, O., Ber. 47, 656 (1914).
Ruff, O., 2. angew. Chem. 46, 739 (1933).
Ruff, O., Ebert, F., and Menzel, W., 2. anorg. u. allgem. Chem. 207, 46 (1932).
Ruff, O., and Thiel, K., Ber. 38, 549 (1905).
Ruff, O., and Braun, H. J., Ber. 47, 646 (1914).
Ruggeberg, W. H. C., and Torrans, D. J., Ind. Eng. Chem. 38, 211 (1946).
Ryss, I. G., and Gibanova, T. A., J . Phys. Chem. (U.S.S.R.) 29, 1822 (1955);
Chem. Abstr. 60, 9121 (1956).
R., and Broun, E. C., Arch. Znd. Hyg.

Occupational itfed. 8, 436 (1953).
260. Schatz, P. N., and Hornig, D. F., J . Chem. Phys. 21, 1516 (1953).
261. Scherer, O., and Petri, H., German Patent 907,775 (1954); Chem. Abstr. 49, 2480
(1955).
262. Scherer, O., and Hahn, H., German Patent 936,090 (1955); Chem. Abstr. 60, 9441
(1956).
265. Schmcisser, M., and Ebenhoch, F. L.,Angew. C h e m 66,230 (19.54).
264. Schmidt, H., and Schmidt, H. D., 2. anorg. u. allgem. Chem. 279, 289 (1955).
266. Schmidt, W., Monatsh. 86, 452 (1954).
266. Schneider, W. G., J . Chem. Phys. 18,1300 (1950).
267. Schneidcr, W. G., Can. J . Chem. 29, 243 (1951).
268. Schnoidcr, W. G., and Chynowcth, A., J . Chem. Phys. 19, 1607 (1951).
269. Schumb, W. C., Znorg. Syntheses 3, 119 (1950).
270. Schumb, W. C., Znd. Eng. Chem. 39,421 (1947).
271. Schumb, W. C., and Gamble, E. L., J . Am. Chern. SOC.
62, 4302 (1930).
272. Schumb, W. C., Trump, J. G., and Pricst, G. L., Ind. Eng. Chem. 41, 1348 (1949).
273. Sccl, F., and Richl, L., Z. anorg. u. allgem. Chem. 282, 293 (1955).
274. Senent, F., Anales real soc. espa6. fis. y quim. (Madrid) 47B, 665 (1951).
27ii. Scvtmon, W. A., and Brice, T. J., J . Am. Chem. SOC.80, 2313 (1958).
276. Sharp, D. W. A., J . Chem. SOC.p. 3761 (1957).
277. Sharp, D. W. A., and Sheppard, N., J . Chem. SOC.
p. 674 (1957).
178. Siehert, H., Z. anorg. u. allgem. Chem. 289, 15 (1957).
279. Silvcy, G. A., and Cady, G. H., J . Am. Chern. SOC.72, 3624 (1950).
,280. Silvey, G. A., and Cady, G. H., J . Am. Chem. SOC.
74, 5792 (1952).
281. Simons, J. H., and co-workcrs, Trans. Electrochem. Soc. 96,47 (1949).
282. Simons, J. H., ed., Fluorine Chemistry, Vol. I, pp. 414-420. Academic Press,
New York, 1950.
285. Simons, J. H., Passino, H. J., and Archer, S., J . Avn. Chem. SOC.63, 608 (1941).
284. Spong, A. H., Chem. & Znd. p. 312 (1934).
285. Stackelberg, M. V., and Jahns, W., 2. Elektrochem. 68, 162 (1954).
286. Standard Oil Development Co., British Patent 537,589 (1941); Chem. Abstr. 36,
1328 (1942).
287. Stcinkopf, W., J . prakt. Chem. 117, 1 (1927); 128, 63 (1930).
259. Saunders, J. P., Shoskcs, M., DeCarlo, M.
156
GEORGE H. CADY
$88. Steinkopf, W., and Herold, J., J . prakt. Chem. 101, 79 (1920).
28.9. Stephenson, R., and Watson, W. E., U. S. Patent 2,430,963 (1947); Chem. Abstr.
42, 2734 (1947).
290. Stevenson, D. P., and Russell, H., Jr., J . Am. Chem. SOC.61, 3264 (1939).
29f. Strehlow, R. A., J. Chem. Phys. 21, 2101 (1953).
299. Strenks, K., Acta. Univ. Latuimsis, Kim. Fakultat. Ser. 1, 233 (1930); Chem.
Abstr. 26, 2067 (1931).
$93. Sudzuki, M., J . Znst. Elec. Engrs. Japan 61, 636 (1941).
894. Sweep, E., Dutch Patent 78,909 (1955); Chem. Abstr. 60, 6776 (1956).
296. Ta-Kong Liu, Moe, G., and Duncan, A. B. F., J . Chem. Phys. 19, 71 (1951).
296. Thomaus, G., J . chim. phys. 49, 323 (1952); Chem. Abstr. 46, 8912 (1952).
297. Thomas, C. L., U. S. Patent 2,313,103 (1943); Chem. Abstr. 37, 5076 (1943).
298. Thorpe, T. E., and Kirman, W., J . Chem. Soc. 61, 921 (1892).
299. Thorpe, T. E., and Rodgers, J. W., J. Chem. SOC.63, 766 (1888); 66, 306 (1889).
$00. Tiers, G. V. D., Brown, H. A., and Reid, T. S., J . Am. Chem. SOC.76, 5978 (1953).
$01. Traube, W., Ber. 46, 2513 (1913).
302. Traube, W., Ber. 46, 2525 (1913).
503. 'lhubc, W., Hoerenz, J., and Wunderliah, F., Ber. B62, 1272 (1919).
304. Traube, W., and Lange, W., Ber. B67, 1038 (1924).
306. Traube, W., and Reubke, E., Rer. BS4, 1618 (1921).
308. Trautz, M., and Ehrmann, K., J . prakt. Chem. 142, 79 (1925).
307. Trost, W. It., and McIntosh, R. L., Can. J . Chem. 29, 508 (1951).
308. Truce, W. E., Birum, B. H., and McBee, E. T., J . Am. Chem. SOC.74,3594 (1952).
309. Truce, W. E., and Hoerger, F. D., J . Am. Chem. Soe. 76, 3230 (1954).
310. Tyczkowski, E. A., and Bigelow, L. A., J. Am. Chem. SOC.76, 3523 (1953).
811. U. 6. Rubher Co., British Patent 643,824 (1950); Chem. Abstr. 46, 3874 (1951).
312. Virtue, R. W., and Weaver, R. H., Anesthesiology 13, 605 (1952); Chem. Abstr. 47,
4484 (1953).
313. Vovnesenskii, S. A., and Duhnikov, L. M., J . Gm. Chem. (U.S.S.R.) 11, 507
(1941); Chem. Abstr. 36, 7309 (1941).
314. Wannagat, U., and Mennicken, G., Z . anorg. u. allgem. Chem. 278, 310 (1955).
316. Warren, J. W., Hopwood, W., and Craggs, J. D., Proc. Phys. SOC.B63, 180 (1950).
318. Warren, J. W., Marriott, J., and Craggs, J. D., Nature 171, 514 (1953).
3i7. Watson, H. E., Rao, G. G., and Ramaawamy, K. L., Proc. Roy. SOC.8143, 558
(1934); 8166, 144 (1936); Chem. Abstr. 28, 2583 (1934); 30, 7406 (1936).
918. Wentorf, R. H., Jr., J. Chem. Phys. 24, 607 (1956).
319. White, W. E., and Hood, A., J . Am. Chem. SOC.74, 853 (1952).
320. Wiechert, K., Z . anorg. u. allgem. Chem. 261, 310 (1950).
391. Wilke-Dtifurt, E., Balz, G., and Weinhardt, A,, Z . anorg. u. allgem. Chem. 186,
417 (1930).
322. Williams, Q., Sheridan, J., and Gordy, W., J. Chem. Phys. 20, 164 (1952).
323. Wilson, W. A., Simons, J. H., and Brice, T. J., J . Appl. Phys. 21, 203 (1950).
SZ4. Woolf, A. A., J . Chem. SOC.p. 1053 (1950).
326. Woolf, A. A., J. Chem. Soc. p. 4113 (1954).
32%'. Woolf, A. A., J . Chem. Soc. p. 2840 (1954).
927. Woolf, A. A., J. Chem. Soc. p. 433 (1955).
328. Wyoski, M. M., J. Am. Chem. SOC.72, 919 (1950).
3.29. Yost, D. M., Proc. Indian Acad. Sci. 88, 333 (1938); Chem. Abstr. 33, 4873
(1939).
YSO. Yost, D. M., Inorg. Syntheses 1, 121 (1939).
157
33f. Yost., D. M., and Clarasen, W. H., J. A m . Cham. Soc. 66, 885 (1933).
332. Yost, D. M., Htrffens, C. C., and Gross, S.T.,J. Chem. Phys. 2, 311 (1934).
333. Young, C. B. F., and Hesse, K. R., Metal Finishing 46, No. 2, 6384; No. 3, 64
(1947).
334. Zimm, B. H., J . Chem. Phys. 19, 1019 (1951).
336. Zandberg, E. Ya., J . Tech. Phys. (U.S.S.R.) 17, 299 (1947); Chem. Abstr. 42, 814
(1948).
336. Ziomek, J. S., Piotrowski, E. A., and Walsh, E. N., Phys. Rev. 98, 243 (1955).
AMIDES AND IMIDES OF THE OXYACIDS OF SULFUR

Margot Becke-Goehring
Department of Chemistry. University of Heidelberg. Heidelberg. Germany
I . Introduction
. . . . . . . . . . . . . . . . . . .
159
11. Sulfur Imidcs Derivcd from Oxides of Sulfur in Oxidation States between (0)
and (111) . . . . . . . . . . . . . . . . . . . .
161
A . Heptasulfur Diimide
. .
. .
. . . . . . 161
B . Hexasulfur lmidc
. . . . . .
C. Tetrasulfur Tetraimidc . . . . .
D . Other Sulfur Imidcs . . . . . .
E. The Compound HaN& . . . . .
F. Amidcs of the Lowcst Sulfur Oxyacids
. 165
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
164
168
169
174
111. Sulfur Imidcs Dcrivcd from Sulfur Dioxide . . . . . . . . . . 174

A. Sulfur Diimide . . . . . . . . . . . . . . . . . 174
13. Thionyl lmidc and Its Isomcrs
. . . . . . . . . . . . 176
IV. Amidcs of Oxyacids of Sulfur(1V) . . . . . . . . . . . . . 179

A . Thionyl Amide . . . . . . . . . . . . . . . . . 179
. . . . . . . . . . . . . . . . 180
B . lmidodisulfamide
C. Amidosulfuric Acid and Imidosulfuric Acid . . . . . . . . . 181
V. Sulfiir lmides Derived from Sulfur Trioxidc . . . . . . . . . . 182
A.SuIfimide. . . . . . . . . . . . . . . . . . . 182
B. Sulfanuric Acid . . . . . . . . . . . . . . . . . 184
VI . Amides of Oxyacids of Sulfur(V1) .
A Sulfuryl Amidc . . . . .
B. lmidodisulfamide
. . . .
C. Amidosulfonic Acid . . . .
D . Imidodisulfonic Acid Monoamid(.
Reference8 .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . . . . . . . . 185
. . . . . . . . . 185
. . . . . . . . . 188
. . . . . . . . . 188
. . . . . . . . . . 190
190
I. introduction
Acid amides are derived from thc oxyacids of nonmetallic elements by

replacement of the OH group of the acid by NH2. the amide group . Similarly. nonmetallic oxides yield their amino analogs when the oxygen of the
oxide is replaced wholly or in part by the NH group. which is isosteric with
oxygen. In general. this substitution is not effected by direct reaction of
the oxide or acid with ammonia . Other methods for preparing the amides
and imides have frequently to be devised . The products are often more
stable than the corresponding oxides and oxyacids .
159
160
MARGOT BECKE-GOEHRING
TABLE I
OXIDES
AND
OXYACIIM
NITROGEN
ANALOGS
OF SULFUR WITH THE
Sulfur oxides
Formula
Nitrogen analogs
Irormula
Stability
LOW
STNH
SiNR
&NMe
Very low
Very low
S4(NR)z
S,(NH)'
S4NR)r
S4(NMeh
LOW
HiNaSi
NasNA
S(NMe)o
OSNH
(OSNH),
(0SNH)r
HOSN
Na (OSN)
3-
+44- Only RSOH and RzSO
Only esters and organo

derivatives
+
+
+
+
+
+
Low
LOOW
LOW
LOW
(O~SNMC)~
(0zSNR)a
(OzSNMC) 4
(0zSNRh
[S(NHs)&
[S(NHa)slz
&E+NH
S(NRa)a
&S(O)NH
Only esters
Only salts
Only salts
Only salts
++
OS(NRdz
SO1 . NHI (?)
O B (NHoh
Salts and orgaiio derivatives'
02s(OH)NHI
Salts and organo derivatives
Sa(NRa)a
HN [(S(0)NHz)11
Ag-salt
-
HN(SOaNHd2
SOzNTTz
NH, . N
+
+
+
+
+
+
+
JAOW
- Oniy esters and other

NHI(0SN)
organo derivatives, salts
Stability
/
\
SO, . ONHI
H O s S (NH-SOz) .OH
H~NSOI-(NH-SOZ)~NHZ
Low
LOW
Low
+
+
Low
+
+
+
+
+
+
Low
+
+
+
+
AMIDES AND IMIDES O F THE OXYACIDS OF SULFUR
161
The numerous oxides and oxyacids of sulfur which are known are listed
in the accompanying table I together with the nitrogen analogs which have
been found up to the present: R denotes an organic radical and Me a metal
atom.
II. Sulfur lmides Derived from Oxides of Sulfur in Oxidation States
between (0)and ( I l l )
A. HEPTASULFUR
IMIDE
I n the normal stable forms of elementary sulfur, that is, in the rhombic
and monoclinic modifications as well as in the melt up to 160" (%), it is
known that molecules are present which contain eight atoms. These molecules form puckered eight-membered rings which may be represented by
Formula (I).It has been found that either one or four sulfur atoms in these
eight-membered rings may be replaced by the N H group. When one sulfur
. ,
s' s'
I
I
I
I
NH
S
'/
(11)
atom is replaced one obtains heptasulfur imide [Formula (II)].

This substance was first discovered by Macbeth and Graham (85)
among the products of the reaction between disulfur dichloride and ammonia. In addition to the main reaction products, namely, tetrasulfur tetranitride, S4N4,sulfur and ammonium chloride, these authors found white
laminae, melting a t 105", which resembled sulfur and gave a transitory blue color with alcoholic alkali. Macbeth and Graham thought that
the compound had the formula SeNH2. Later Arnold et al. (11) thought it
to be SI8N3H6,and finally Arnold and his co-workers discovered the correct
f0rmu.a) SINH. The compound was then studied further by Goehring et al.
(49,and the constitut on was elucidated as that shown by Formula (11).
Heptasulfur imide is formed from elementary sulfur or from the product
of the reaction of sulfur with sulfur trioxide by treatment with liquid
ammonia a t room temperature. A disproportionation reaction takes place
in which, in addition t o HZS and S4N4, S7NH is produced from the elementary sulfur or S203( I , 39).
In the preparation of S7NH it is better to use the reaction of sulfur
chlorides with ammonia. This can occur both in organic solvents (9, 15, 45)
162
a d in water (%3). Dimethylformitmide is a very coiivciiieiit sulveiit fur the

reaction. Becke-Goehring et al. (15) assume in this case that the reaction
occurs in the manner described in the following paragraph.
In the polar solvent there is first a dissociation of the SzClZ. Ammonia,
as a nuclcophilic reagent, attaches itself to the cation SsCi+. The unstable
intermediate (111) will then have a considerable tendency to split off
positive groups itrid is able t o react with further SZC12 and so lengthen the
sulfur chain. The stepwise reaction which lengthens the chain leads firially
to compound (IV), which can stabilize itself by splitting off HCl and closing
the ring to form S7NH.
+ CI- + NH,
C1- S-Sf
Cl-S-S-CI
Cl- (H3N-S-S-Cl)
+ Cl--S-S-Cl-+
+
+ CI-(H.yN-&-CI)
---t
(Cl-S-SNH3)
C1-
(111)
Cl-(H,N-S-S-S-CI)
+ HCI
+ H*N-S7--CI
+ SCI,
--*
HCI
+ 87NH
(IV)
Heptasulfur imide melts a t 113.5" without decomposition, but the melting point is strongly depressed by small admixtures with sulfur. A eutectic
between STNIXand Ss is observed at 91.5" with 46% of sulfur (25). Othcrwise the physical behavior of STNHstrongly resembles t,hat of elementary
sulfur. Thus, for example, it is not wetted by water but dissolves in many
organic solvents-indeed more readily than sulfur itself.
On rapidly heating STNH in vacuum above the melting point the following reaction occurs:
787NH ---t 2112s
+ NHS + 47s + 3Nz - 450.1 k d .
From the heat of this reaction the enthalpy of formation of 87NH from
its elements may be calculated as -67.4 kcallmole. It is seen that S7NH
is unexpectedly stable-more so than would be anticipated from a knowledge of the electronegativities of the partners forming the bond. This
stability can be traced back to resonance energy if the limiting forms
H
N
s/
I
s
I
s
I
\sI
s
I
s
\N'
I
I1
AMIDES AND IMIDES O F THE OXYACIDS OF SULFUR
163
(Va-d) are taken into consideration. If this view is correct, there should be
a considerable decrease in the stability of the eight-membered ring system
in the transition from the hydrogen compound to the ion (VI).
(VI)
By decomposition of ethereal solutions of heptasulfur imide with triphenylmethyl sodium, an olive green sodium salt Na(NS7) may be prepared (17)
which probably contains the ion (VI) and which, as expected, is much more
reactive than S7NH itself. For example, NaNS7 undergoes hydrolysis even
in contact with moist air.
The hydrogen atom which, as the infrared spectrum shows (%), is
linked with the nitrogen atom of the S7NH molecule, can be replaced
readily by mercury, as would be expected. By decomposing S7NH with
mercury(I1) acetate yellow-white Hg(NS7)z is obtained (93) and bright
yellow Hgz(NS7)2
may be precipitated from a solution of S7NH in dimethylformamide by mercury(1) nitrate (56).The mercury compounds are rapidly
decolorized a t room temperature with decomposition.
Compounds in which the hydrogen is replaced by a n organic group are
more stable than the more or less salt-like metallic derivatives of S7NH. As
examples, mention may be made of the reaction product with formaldehyde
(49,93) Formula (VII), in which both the whole OH group or its hydrogen
atom may be replaced by other organic groups; the reaction product with
acetyl chloride (VIII) ; and the corresponding product with benzoyl chloride (IX) (49).
s-s-s
/
S
\
8-s-s
s-s-s
H
N-C-OH
/
H
s-s-s
\
/
N-C-CH,
(IX)
Reaction of SsNH with other inorganic acid halides is also vigorous. Thus,
interaction with SCl2, with the addition of pyridine as a n acceptor for HCI,
164
gives good yields of crystalline yellow S16N2 (Xa-c), which melts without
decomposition a t 137"; whereas with SzClzthe compound S M N(XI)(16)
~
is
obtained (mp, 95-100"). Particularly S16N2 appears to be stabilized by
resonance, as is indicated by Formulas (Xa-c).
s-s-s
s-s-s
\ - - -/
N4S-N
s-s-6
/
S
\
5-8-
s--8-s
\ - I @
\
N=s-
s-s-s
e *\ s-s--s
/@
(Xa!
s/
\
/N-s-s--N
s--s--s
\S
s- s- s/
(XU
The reaction of S7NH with formaldehyde indicates that S7NH is able

to function as a Lewis base, that is, as a nucleophilic reagent. This mode of
/
S
\
s-s-- s
S-S-S
\
IN
/ I
+ CH&
s-s-s
S-S-S
TI
\*N-(3-010
I -
/I
H H
/s-s-s
S
\
S-S-S/
13
N-C-011
reaction is even more apparent in the interaction of S7NH with the strong
Lewis acid, sulfur trioxide. A carmine red SOaadduct, S7NHS0, is formed,
which very probably has the formula (XII)(@).
/
S
s--8-s
\
IN-H
/
s- s---s
s-s-s
+ SO3 -+ S/
\
s-s-s
\I
/
N-SO@ + S
/*
\
s-s-s
s-s-s
\
N-SOaH
/
(XII)
B. HEXASULFUHDIIMIDE
It has been shown recently ( 1 1 5 ~that

)
the reaction between S2C12and
ammonia results riot only in ammonium chloride, sulfur, heptasulfurimide,
and S4N4,but also in the compound SO(NH)Z.
The compounds S7NH and
S,(NH), can be separated by adsorption chromatography on aluminum
oxide. The hydrogen atoms in the compound are acidic like those in S7NH
AMIDES A N D IMIDES O F THE OXYACIDS O F SULFUR
165
and can he estirnatcd quaiititativcly hy the method of Zerewitiiioff. The

colorless, crystallized compound is soluble ill many organic solvents and
insoluble in water. The crystals change color to yellow by heating up to
120" and melt a t 140" with decomposition, forming a red liquid. The compound has the formula (Ve).
C. TETRASULFUR
TETRAIMIDE
As has already been mentioned, it is possible formally to replace not
only one or two sulfur atoms in its eight-membered ring [Formula (I)] by
NH, but also four such atoms. In this way one obtains tetrasulfur tetraimide (XIII).
HN-S-NH
SI
sI
I
HN-SNH
(XIII)
This compound was discovered by Wolbling (118), who reduced tetrasulfur tetranitride, S4N4,with tin(I1) chloride. Formula (XIII) was first
proposed by Arnold ( l o ) ,although Meuwseri (89) had already suggested
a ring structure on the basis of a molecular weight determination, though
he placed the hydrogen atoms on sulfur. After evidence for N H groups in
the molecule and a sulfur oxidation number of +2 had been adduced from
the chemical properties (10) and from the Kor X-ray emission spectrum of
the sulfur (32), the presence of N-H bonds was established from the
infrared spectrum (83).
FIG.1. Molecular structure of Sd(NH), (84, 105).
The structure of S,(NH), is completely elucidated (84, 105). The colorless substance crystallizes in the orthorhombohedral system. There are
four molecules in the unit cell and the molecule has the configuration shown
in Fig. 1. Both the sulfur and the nitrogen atoms have a square planar
arrangement. According to Sass and Donohue the S-F-S angle is 122"
and the N-S-N angle 103.4". The N-S distance is 1.67 A. The molecule of
H4N4S4is seen to be extraordinarily similar to that of elementary sulfur.
I66
MARGOT BECKR-GOEBRIXG
As in the case of S,NEI, hydrogen it1 84(NH)4may be replaced by metals.

Attempts t o prepare an alkali salt of tetrasulfur tetraimide go back to
Meuwsen (go), who allowed KNHz to react with S4(NH)4and showed
KNS . KNH2 was probably formed. Becke-Goehring and Schwara (17)
studied the reaction of S4(NH)4with triphenylmethyl sodium and obtained
orange-red solid Na4(N4S4).It is likely that citron yellow Na2(H2N4S4)is
also formed. These sodium salts are extremely unstable. They react with
water with inflammation, but even in moist air a rcaction may occur
which is accompanied by detonation.
These experiments already indicate that the ion (XIV) is of low stability, and this conclusion is supported by attempts to prepare complex salts
with this ion as a ligand. It is found that S4(NH)4reacts rapidly and quantitatively with lithium aluminum hydride, Li(AIH4). Hydrogen and a
solid white compound Li(Al(N4Sa)) [Formula (XV)] result. This cornpound is very unstable:
(XIV)
(XV)
When well dried it detonates on touch. On careful hydrolysis (XV) may be

reconverted to the hydrogen compound (XIII) (M).
Cornpounds of heavy metals with the (N5)4 group are more stable than
those substances which contain the ion (XIV), in which the bonding must
be supposed to be largely heteropolar. Prom copper(1) chloride and S4(NH)I
in pyridine solution one obtains brown-black solid (CuNS).t (54), and redbrown (AgNS)4 miiy be prepared similarly. Copper(I1) chloride likewise
yields the solid yellow compound (XVI), while copper(I1) nitrate reacts
CI--Cu-N--S-NH
IIL-~-CU-C1
(XVI)
only partially to yield ( C U N O ~ ) ~ H ~which

N~S~
is,the analog (XVI). Mainly
Cu(NO& tends to oxidize S,(NH), and to convert it to &N4. The mercury
compounds are quite stable, yellow Hg4(N4S4)resulting from the reaction
of Hg2(N0& with S4(NH)4 in dimethylformamide (66), while yellow
Hg2(NS), and yellow Hg,,(NS),, is obtained by the reaction of S4(NH)4
with mercury(I1) acetate (92).
For all of these heavy metal compounds the molecular weight is not yet
167
established and it is only a matter of conjecture that the ring system

remains intact, as in formulas (XVI) and (XVIT). It is highly probable,
Me-N-S-N-Me
s
I
s
I
M~-N-s-N-M~
(XVII)
however, that the metal in the compound is linked to nitrogen. To check

this point the silver compound was reacted with ethyl iodide (54),and the
product was then reduced and hydrolyzed. Formation of ethylamine
showed that the ethyl group replacing the silver atom was bonded to
nitrogen.
Just as in the case of STNH, organic derivatives of S4(NH), may be
prepared. The methylol compound (XVIII) (10, 89) has been known for
a long time, and from it, further derivatives may be prepared (10). Arnold
H
C6Hh-N
C-N-S-N-C
HO-CHz-N-S-N-CHz-OH
0)
l'i
HO--CH~-N-S-N-CHrOH
GHs-N
\o
\
(XVIII)
N-COHI
N-CsH6
H
(XIX)
(10) prepared (XIX) from S4(NH), and phenylisocyanate. Lengfeld and

Stieglitz (79) obtained S4(NH>,derivatives by a n entirely different route
involving the interaction of ethylamine with SClz. In this way a yellow oil
is obtained t o which the author ascribed formula (XX). Whether this oil
is really the pure product (XX) or a mixture of it with other polymeric
compounds of the formula (CzH,NS), remains uncertain. It seems the more
questionable since Stone and Nielsen (109) obtained from methylamine
and SClz a mixture of polymers (CH3NS),, in which higher polymeric subCzH5-N-S-N-CzH5
I
I
S
S
CaIIa-CHz-N-S-N-CHzCsHs
CzHs-N--S-N-CzH6
(XX)
CsH6-CHz-N-S-N-CHz-CsHa
(XXI)
sI
I
sI
I
stances (with a molecular weight of 600 and more) were present to a co;isiderable extent. Jenne (67) obtained a cryst,alline derivative (mp, 163"),
163
which probably has the Formula (XXI) from benzylamine and SClz a t
high dilution. This product is almost certainly a monomer, but the molecular weight has not yet been definitely established. Stone and Nielsen prepnred recently the analog methyl compound ( 1 0 9 ~ ) .
D. OTHEHSULFURIMIDES
In addition t o SSmolecules which constitute the stable modifications of
sulfur, Ss molecules may he obtained, the structure of which has been elucidated (28) and which form a puckered six-membered ring, Formula ( X X I I ) .
The question then arises as t o whether one or more sulfur atoms in this
ring system can he replaced by the NH group, that is, whether there are
compounds such as (XXIII) or (XXIV).
s-s
s-s
s s
\s
S/
\S-d
s--s
/
If-N
\
N-H
(XXIII)
(XXII)
\
N-II
/
s-s
(XXIV)
I t has not yet proved possible to prepare these sulfur imides, but
derivatives of at leust (XXIV) have been found in which the hydrogen of
the iinide group is replaced by an organic radical (XXV). The first repre5-s
/
R-N
\
\
N-R
/
s--s
(XXV)
sentative of these N,N'dialkyl tetrasulfur dinitrides was found by Levi

(81) in very small yield when he allowed SzClzto react with ethylaminc and
distilled the product with steam. This group of compounds becomes readily
accessible if S2C12is allowed to react with alkylamines in ether solution
using the Ruggli-Ziegler dilution principle (14).
H
R-N
II
\"
c1
c1
N-13.
+ 4HoNR
-+
s-s
4[HsNR]Cl
+ R-N
\%S/
CzHs, CIzHzs, CsHn, CsHsCHZ, CbHaCH,.CH2,
s-s
N-R
A M I D E S AND 1MII)E:S OF THE: OXYACIUS O F S U L F U R
169
Sulfur imides in which the N H groups are incorporated in sulfur chains

and chain-like sulfur imides in general, have not yet been prepared. There
are also no organic derivatives of these compounds unless the polymeric
substances of Stone and Nielsen (XXVI) (109), which are derived from the
hypothetical imide (XXVII) are counted as such.
H
(XXVI)
-N-S-N-S-N-S(XXVII)
E. THE COMPOUND
H3N3S2
If the nitrides of sulfur, S4N4and S2Nz (M), are dissolved in liquid
ammonia, then a reaction occurs a t about -70" for S2Nzand a t about
-40" for S4N4. A dark red solution is produced as the reaction proceeds,
and after driving off ammonia, a red mass of the composition S4N4,2NH3
(104) or S2Nz,NH3(39)is obtained. The solution spectra of these ammoniates in the visible and ultraviolet regions are identical, as are the X-ray
powder photographs of the two solids. Accordingly the two substances are
almost certainly the same. Since volatile SzN2 is readily formed from the
ammoniate, it is probable that the S4N4molecule is decomposed during the
reaction with ammonia into S2Nzfragments. The process can then be formulated as follows:
If this view is correct, the ammoniate should behave as an acid in liquid

ammonia (22),that is, the solution should conduct and the hydrogen should
be replaced by metal in a neutralization reaction. Actually the conductivity
of liquid ammonia is increased considerably when S4N4dissolves in it and
the solution is warmed to -47" (I?'). Thus with a S2Nz. NH3 molarity of
1.67 X
mole/liter, the specific conductivity is 0.104 X
at
21.01 X
mole/liter it is 0.62 x
and a t 13.24 x
moIe/Iiter,
ohm-' cm-1.
2.84 X
The metal amides are the outstanding bases in the ammonia system with
which acid (XXVIII) can be neutralized. In order to convert this acid
(XXVIII) into a salt, Berg and Goehring (19) reacted its solutions in ammonia with KNH,. A yellow solid with the composition K3N3S2was pro-
170
MAHGOT BECKE-GOEHRING
duced. This potassium compound proved to be unusually sensitive to dry,

and, especially, to moist air, and reacted with inflammation.
The yellow solid KjNSS2 may be a salt of the ammoniatc (XXIX), or,
as was a t first assumed, an equimoleaular mixture of the salts (XXX) and
(XXXI). Since no difference was observed in the reactivity of the H atoms
of the nrrirnoniute during the reaction with potassium amide, the theory
favored init,ially was that there was first a dissociation of H3N3S2 t o HNS
Ka(-fl- -%-N=S--N-]
-
K [-%4]
- - -
&[ IN=%N-] -
(XXX)
(SSIX)
(XXXI)
and S(NH&, followed by salt formation. Later, however, it proved possible,

by using a much milder neutralization reaction, to establish a difference
between reactivit8yof thc separate hydrogen atoms of (XXXII), as would
be expected for an imide-amide. The SzNz NH3 was titrnted in ethereal
solution with triphenylmethyl sodium solution ( l y ) , and two color changes
were observed. At the first the red solution became citron yellow. When
twice the volume of the triphenylmethyl solution required to produce this
color is then added there is u further color change to red. After the first color
change, there was a deposit which was brown when dry and had the composition Na(-N=S=N-S-NH2).
After the second color change, a yellow
solid phase is obtained, which has the formula Na3(-N=S=N-S-N=).
The latter is very unstable and detonates on touch.
It is seen from all these experiments that H3N3S2is a substance with
acid properties. The limiting Rtmctiires (XXXIIa-d) must probably be
considered. The imide hydrogen should be replilccd readily by metal atoms,
but the amide hydrogen is also replacable in this way.
H-~=S==N-%~-II
- _ -
e e
II--~-~%N-&-~--H
_ _ _ _
I
H
H
(b)
(8)
@_e
H-N-S-N-3-N_ _
_ _
11
H--N--S- -
-N=S-N-H
I
In the rcwtion with triphenylmethyl sodium the molecule of H,N&

certainly remains intact, since the isolation of the moriosodium salt would
otherwise be difficult to understand. It is still uncertain whether, in the
reaction with potassium amide, the molecule is decomposed or remains
intact; that is, whether the potassium salt has the formula (XXIX) or
(XXX) and (XXXI).
Cleavage of the molecule (XXXII) into two fragments with sulfur in
AMIDES A N D IMIDES OF THE OXYACIDS OF SULFUR
171
diff ererit oxidation states, which cannot be established with certainty in

the reaction with KNHz, is however observed if the compound is decomposed with PbIz or Pb(N03)z or with Hg12 in liquid ammonia as solvent
(104). I n the reaction with lead salts, green Pb(NS)2. NH, which is
sparingly soluble in ammonia, is precipitated. After filtering this substance, yellow-green Hg(N2S)NH3may be precipitated by adding HgI2.
The metallic salts may also be added in the reverse order.
The structure of Pb(NS)2NH3has now been elucidated (117). It crystallizes in the monoclinic system and there are four molecules in the unit
cell. The molecule, which is not planar, has the formula (XXXIII), with
the N1-Pb-S' bond angle 78"; N-S-N, 114";Pb-N1-S, 115", and S-N2-S, 103".
NI-SZ
,I,
\-
S'--NZ -
(XXXIII)
The following interatomic distances hav: been found: Pb-S, 2.75 A;

Pb-N1, 2.27 8; S1-N2, 1.72 A; S2-N1, 1.70 A; S2-N2,1.60 A. Since the distance between N2 $nd the lead atom of the neighboring molecule is relatively small (2.83 A), one can perhaps assume the existence of a secondary
valency from this nitrogen atom to the next molecule. The position of ammonia has not yet been determined with certainty.
The structure of Pb(NS)2 shows that H3N3S2undergoes intermolecular
disproportionation under the influence of heavy metal salts such as PbIp.
2H3NsSz + HzN&
+ 2HZNsS.
The products of the disproportionation reaction react with heavy metal

salts forming compounds of low solubility. A similar disproportionation of
sulfur nitride ammoniate was proposed earlier by Meuwsen (go), but he
assumed the formation of H4N4S4,whereas, in fact, as follows from Formula
(XXXIII), the acid from which the lead salt is derived has only half the
molecular weight.
Thallous nitrate reacts with (XXXII) in the same way as lead iodide
and nitrate. Red-brown 2Tl(NS), * NHs may be precipitated from ammonia
solution and may be converted to red-brown Tl(NS)3 (54). Ochre colored
T13(NS)*is precipitated from alcoholic solution. With copper(1) chloride,
brown C U ( N S )is~ precipitated, while with silver nitrate red-brown Ag(NS)2
is obtained (54). The structures of these metallic derivatives are not yet
established.
A series of inner complex salts of metals of the eighth subgroup is derived
from the hydrogen compound HzN2S2, on which the lead compound
(XXXIII) is based. The possibility of forming such complexes was first
172
r~(:opiii~ed
by (;oeliriiig and Debo (42) from the iiitensc! red color which
results from tho interaction of Bcllucis salt, K4[Ni(CN),],, arid S4N4in
alcoholic solution. Wciss (If5 ) later succeeded in isolating the cornpourid
K(Ni(CN)2(HN2S2))tis one of the products of this reaction. Since this
cornpourid is diamagnetic and the presence of hydrogen hss been established unambiguously by thc Zerewitinoff-Tschugaeff method (116),
formula (XXXIV) is very probably correct, with nickel in the bivalent
state. It is found that other complex compounds with the same sulfur
(XXXIV)
(XXXVa)
(Me = Ni, Co, Pd, Pt and probably Fe aho)
imide as a ligarid are also very readily made. The nickel and cobalt carhonyl
compounds, (Ni(C0)4) arid CoZ(C0)8 or HCo(C0)4 are allowed to react
with S4N4 in an indifferent organic solvent. Alternatively, NiCI2, CoC12,
PdCI2, or H2(PtCle) may be used in alcoholic or dimethylformamide solution (26,36).Heavy metal compounds of the formula (XXXVa-e) may be
isolated from the reaction products, in some cases in very good yield. On
prolonged boiling of NiClz with S4N4 in alcohol, compounds other than
(XXXVa-e) are formed in small amounts (102). Compounds of the type
(XXXVa-e) are well crystallized and deeply colored. The nickel and cobalt
compounds are deep violet, the palladium compound is red-brown and the
platinum compound blue-black. These substances are freely soluble in
many organic solvents. They were originally formulated as Me(NS)4
without hydrogen, largely on account of their formation from carbonyls
and S4N4,since it was not upparent how the formation of (HS2N2)- could
arise. Piper has, however, proved that the compounds actually contain
hydrogen (101), and subsequently Weiss and Becke-Goehring have established the presence of hydrogen by the Zcrcwitinoff-Tschugaeff method
(116). An iron coinpourid Fe(NS)*, which is formed in the reaction of
Fe(C0)6with S4N4,also belongs to this series (40). Whether it has the formula (XXXVa-e) or is hydrogen free is still unknown. It is striking that the
iron compound does not dissolve in nonpolar organic solvents.
The constitution of the platinum complex has been elucidated by
crystal structure investigation (82),
and the molecule (XXXVa-e) has been
found to be completely planar. The interatomic distances and bond angles
are given in Fig. 2. As would be expected, the nickel compound is diamagnetic, and the cobalt and iron compounds show paramagnetism, corresponding with one and two unpaired d-electrons, respectively (41). These
AMIDES AND IMIDES OF THE OXYACIDS O F SULFUR
173
FIG.2. Interatomic distances and bond angles in Pt(HN2SZ)T (Lindqvist and Weiss, 89).
properties are quite unexpected, a t least as far as cobalt and iron compounds are concerned, and they indicate that an abnormal electronic state
of the metal is stabilized by an unusual ligand field. The planar structure,
the high stability, and the special physical properties, lead one to suppose
that the compounds possess special possibilities of resonance. In fact,
numerous mesomeric resonance structures may be formulated, some of
which are given by Formulas (XXXVa-e).
(XXXVb)
(XXXVC)
IN=a
R=N
i -X I
N-SI
IS--N
- 1H
I -
H
(XXXVe)
(XXXVd)
174
MARGOT BECKE-GOEHRINC
The copper compounds which are obtained by thereuction of CuCI2.2H2O

or CuBrz with S4N4in dimethylformarnide are clearly similarly constituted
(34). If Formula (XXXVIa-c) is assumed in accordance with the analytical
results, which fit the formula CuXzNzS2,copper will have an oxidation
number of +3. As the deep violet solid chloro compound is diamagnetic,
the physical properties of the compound appear to support this theory.
Resoriarice stabilization of the complex is also possible here.
6 - NI
X
X/
/ \-
\N--, I
X \ C " . TN-S[
I -
IT
(XXXVla)
H
(XXXVIb)
\c"JJ-f
/ \
N-S(
I
H
(XXXVIC)
(X = C1, Ur)
F. AMIDESOF
THE
LOWESTSULFUROXYACIDS
The compound HsN3S2is an irnide-amide of sulfur. Amides of sulfur with

a simpler structure, such as the compounds (XXXVII-XL) have not yet
H2N-S-NHZ
(XXXVII)
I,N-S-S-NHI
(xxxvlll)
H2N-S-S-S-NH2
(XXXIX)
TT~N---S-S-S-S-NH~
(XL)
been prepared, though organic derivatives of these compounds, in which

hydrogen is replaced by alkyl radicals, are well known and quite stable.
Derivatives of (XXXVII) and (XXXVIII) were prepared quite early
(78, 94, 95). Levi (80) has made derivatives of (XXXIX) by the reaction
of diulkylamines with sulfur arid lead oxide. Derivatives of (XL) are available by several routes; for example, by reaction of sulfur with secondary
arniries and HgO or PbO, by reaction of S7NH with amines, by reaction of
S2Cl2 with NHa and dimethylformamide, and, finally, from secondary
amines with S2C12, a method which also gives compounds with 1, 2, and 3
sulfur atoms in the chain (68).
I l l . Sulfur lmides Derived from Sulfur Dioxide
A. SULFUR
DIIMIDE
If both oxygen atoms in sulfur dioxide (XLI) are replaced by isosteric
imide groups, sulfur diimide (XLII) is obtained. As in the dioxide, sulfur
in this compound should have an oxidation number of +4. Derivatives of
this diimide were obtained in the reaction of S4N4with NH, and HgIz
(104). The compound HgNZS, which results in this reaction through the
175
A M I D E S AND IMIDES O F THE OXYACIDS OF SULFUR
qe
01
q/-
SY
Ib
\-
916
xo,
(XLU
N-H
N@--H
4
IS
r"S/-
\
N-H
\N
-~ - H
(XLII)
intermediate H3N3S2, should be regarded as a mercury derivative of

(XLII).
Sulfur diimide should be obtained most simply by ammonolysis of a
halide of tetravalent sulfur. Sulfur tetrachloride or trithiazylchloride
(XLIV) are available as acid halides for such a reaction. The former is very
readiIy decomposed by ammonia, but S(NH)2 cannot be identified among
the reaction products which consist chiefly of S4N4,NH4C1, and chlorine
(43).Only an organic derivative of (XLII) could be prepared in a solvolytic
reaction of this sort with SC14, in which butylamine gave the n-butylimide
(XLIII) (53).
CH~-CH~-CH~-CH~-N=S=N-CH,-CH~-CH~-CH~
(XLIII)
Thiazyl halides, such as thiazyl chloride (XLIV) or thiazyl bromide,

the moIecuIar weight of which is unknown, are rapidly solvolyzed by ammonia, in which they dissolve to give a red color. Analysis shows the prod-
c1
I
H
(XLVI)
-NH,
SAg+
IN
S
I'N1
HN=S ,-/S=
N
Ag
(XLVII)
NH
176
w t to be :t niixturc of ammonium hdicle and [S(NH&],. Compounds in

which sull'ur has tin oxidatioii iiumber of +4, however, disproportioriate
readily to sulfur nitrides (tervalent S) and compounds of sexvalent sulfur.
The derived hydrogen compound may, however, be isolated as a metallic
derivative when the structure is stabilized (20). Thus the product for
which the Formula (XLVI) is proposed interacts with potassium amide
in liquid ammonia to give solid yellow Kz(NzS), while with Hgh, one
obtains from liquid ammonia yellow-green solid Hg(N2S)NH3, which is
transformed to yellow Hg(N2S) in high vacuum at 90".
When the reaction product of (XLIV) is taken up with ammonia and
precipitated with silver nitrate, red-brown silver salts of variable composition are obtained. When part of the ammonia is removed carefully from the
filtrate one obtains :Lyellow crystalline silver salt which analysis shows to
have the composition AgN& and which can very probably be represented
by formula (XLVII). This silver compound is light sensitive but otherwise
stable at room temperature. It detonates with extreme violence on impact
(86)*
B. THIONYL
IMIDEAND ITSISOMERS
Reaction between thionyl chloride and ammonia should in principle
lead to a further sulfur imide (XLVIII), derived from sulfur dioxide by
replacing one oxygen atom by the isoteric N H group.
(XLVIII)
(XLIX)
P. W. Scherik (106) studied the reaction between OK12 and NH3 in the
gas phase at low temperature. He obtained a product which was colorless
and solid below -85' and melted a t about -85' to a colorless liquid with
an appreciable vapor pressure a t - 70'. Analysis and vapor density determinations indicated the foririula OSNH, and one can assume with some
certainty that this volatile substance is a true analog of sulfur dioxide with
the structure (XLVIII).
Monomeric thioriyl imide polymerizes very rapidly. Even a t above
-60" a yellow solid separates from the liquid which, with further warming
becomes glass-clear yellow-brown, red, and firially brown. The brown
polymer, which does not dissolve in organic solvents, is found from the
infrared spectrum to contain NH groups (I@, SO that Formula (XIAX)
proposed by Schenk appears to be established.
Reaction between thionyl chloride and ammonia takes a different course
177
a t room temperature in an organic solvent such as chloroform. The solution

becomes deep red, and, particularly if lime is added to fix the HCI, the
following decomposition occurs almost quantitatively :
OSCIZ NH,
+ CaO
--t
HOSN
+ CaClx + HzO.
The red solid compound HOSN may be isolated after evaporating the
solvent. It is not, however, very stable and readily goes over into the brown
polymer (XLIX). The red compound, which is monomeric in organic
solvents, is an isomer of Schenks thionyl imide. It polymerizes so readily
that only chemical methods have hitherto been applied successfully in
establishing the structure as (La) (18).
This is transformed readily by excess of thionyl chloride into the acid chloride, which goes over t o the stable trimer (XLIV).
-
H-C-S_=N
+ OSCIz
--t
Cl-S==%j
-
+ SO2 = + HCI
C1
(XLlV)
In solvents which are not proton acceptors (L) may be shown by cryoscopic measurements to be present in the undissociated state. If the solvent
is itself a proton acceptor it follows that the ions (LIa-c) must result.
In a n aqueous medium the ion undergoes rapid hydrolysis, the results
of which are especially apparent if the bisulfite ion HSO, is added to the
reaction medium. The changes taking place may then be represented by
the following scheme:
t2HS0,-
\+
HSOz-
178
MARGOT HECKE-GOEHHING
When ammonia is used as solvent a deep red solution is formed which c*ontairis the ammonium salt (N&) (OSN). At elevated temperatures the
ammonia also exerts a solvolytic action.
The ion (NS0)- is more st:tblc than the hydrogen cornpourid. This is
a t once upparcut if one considers the lirnit,ing formulas (Lla-c) which
describe the ion. Thc incidence of mesomerism becomes iriuch smaller in
passing from the ion to the hydrogen compound. In fact, the latter car1 also
be represented by several limiting formulas (La-c), but (Lb) is improbable
l ~ c a u s eneighboring atoms have formal charges of the same sign, while
(Lc) d s o has only ti small effcct as the charge distribution does not correspond with the electronegativities of the elements. The resonumce-stabil i e d ion is present in the lattice of the sodium salt which may readily be
prepared by treating itii ethereal solution of HOSN with triphenylmethyl
sodium (18). The solid brown-violet salt Na(0SN) is, in contrast to the
hydrogen compound HOSN, stable for some days when air and moisture
are excluded, ttnd thc tendency to polymerizihori shown hy (La) is no
lotiger observed. A red-silver salt Ag(N80) can also be prepared (51). As
might be expccted, this is much less stable than thc sodium salt, since the
ability of the ion (LI) to show resonance is milch restricted by polarization.
Stjartirig from imidosulfintimide (LII) , the ummonolysis product of red
HOSN, i t is possible to obtain a further isomer of thionyl imide. Thus,
when hydrogen chloride is condensed on to (LII) a yellow solution results
H2N-SO-NH-SO-NH~
(LW
from which, on distilling off the HCI, yellow crystals separate. The yellow
substancc passes over into t)he brown polyrricr (XLlX) with great ease.
The absorption spectrum of the yellow substance is very similar to that of
CI?H~NSO
(LIII) (69). Probably, therefore, it has the same arrangement
of atoms as in (LIII) or (XLVIII). The isomer must differ from monomeric
thioriyl imide in its degree of polymerization, but the great invtdility of
the substance does not allow this to be determilied. It is striking that the
yellow isomer is able not, only to go over rapidly to brown polymeric thionyl
179
iniidc a t low temperatures (above -60") hut can also yield thc red isomer
when it is rapidly warmed to about $60" (51),
e e
C2HsN-60
C2Hb-N-S-0
e e
(LIII)
Finally, yet a fifth isomer of thionyl imide may be obtained in quite

another way. Fluck arid Becke-Goehring (33) started with tetrasulfur tetraimide (XIII) and oxidized it by heating with air at 110-120". Among the
products was the solid red tetramer of thionyl imide (LIV) which was soluH
OS-N-SO
HA
I
I
NH
US-N-SO
H
(LW
ble in organic solvents such as methanol. The tetramer could not be transformed into the other isomers. The followiiig scheme shows the transformations of these various isomers.
OSCI,
+ NH3
H-0-S=N
H
c1
I
,!,-,!,7,!!\
-,
s
s
II
lot
II
101
101
_I
HN-S-NH
CI-S\
HN- S-NH
- -.
!I'
FIG.3. The interrelationship of the isomers of thionyl imide.
IV. Amides o f Oxyacids o f Sulfur( I V )
A. THIONYL
AMIDE
Schiff first suggested that the diamide of sulfurous acid was formed in
the reaction of thionyl chloride with ammonia (107).Later Meslans (87,88)
180
MARGOT BECKR-GOEHRINB
thought it likely that the same substaiiee wtit~sformed from thionyl fluoride
and nmnonia ;
OSX2
+ 4NHa
-+
OS(NIS2)Z
+ 2NH4X.
These speculations have not so far been confirmed. In the reaction between
thionyl chloride and ammonia, using liquid ammonia and cooling carefully,
a compound with sulfur and nitrogen in the atomic ratio of 1 :2 does indeed
result (3'7); but it shows the same absorption spectrum in solution as salts
of HOSN, so that its structure is probably not that of an amide of sulfurous
acid, but of the isomeric ammonium salt NH4(OSN).This salt goes over
very readily to (LII). Meslans' experiments could not be confirmed either
by Moissan and Lebeau (97, 98, 99) or in later work, although it was
thought possible that thionyl amide is formed from OSFz and NH, (96).
I t is not surprising that thionyl amide (LV) has hitherto not been
obtained. Its analog, sulfurous acid, has also not been made in the pure
state and readily decomposes to SOn, so that one can expect OS(NH2)z to
react in an analogous way.
OH
(04/\
NHz
+s'-
-1120
(0-s
/
\
Nl@
4
0lO-k
- ?/
. NH,$
NIT2
OH
(LV)
In coiitrast with the hydrogen compound (LV), its organic derivatives, in

which H is replaced by alkyl, are well known.
B. IMIDOSULFINAMIDE
The ammonium salt of thionyl imide, NH4(OSN), is transformed in a
few hours a t room temperature into imidosulfinamide (LII), with evolution of ammonia;
BNR(0SN) + NHs
+ H2N-SO-NH--SO-NH,.
(LII)
Imidosulfiriarnide can also be obtained directly from OSCl2 and NH,,
without first making NH,(OSN) (48). Findly, it can be made from SO2
arid NH, by allowing one molecular proportion of NHs to react with two
of SO2a t - 10' to -20" to form orange-red SO2 2NH3. The latter decomposes with loss of NH, and SOz and (LII) can be isolated from the decomposition products (48).
The diamide of imidosulfinic acid is a yellow solid which is relatively
stable. A silver salt AgN(SONHAg)23H20may be prepared from it. When
one attempts to recrystallize the substance, which comes down in the
amorphous state, from water, HOSN, which can be recognized by the
181
AMIDES AND IMIDES OF THE OXYACIDS O F SULFUR
transitory red color, and also a well-crystallized compound with the composition S3N402H4--6
(110) are formed. Molecular weight determination in
water shows this to be an ammonium salt for which the Formula (LVI) is
proposed.
GVI)
The imide from which the salt (LVI) is derived is the oxygen analog of the
imide corresponding to the silver salt (XLVII). Its probable mode of formation entails initially the loss of ammonia and HOSN from two molecules of HN(SONH&. The resulting trimeric thionyl imide then tends to
form its anhydride
0
II
2HN(SONH2)2-+ NH3
+ I-IOSN +
HN/
o=s
\NH
s=o
-HzO
--+
\d
H
(LVII)
o=s
)Nle
s=o
"'
H
(LVIII)
Loss of water from (LVII) corresponds to the loss of ammonia from the
corresponding imide (XLVI) (cf. p. 175).
C. AMIUOSULFURIC ACID AND IMIDOSULFURIC

ACID
When sulfur dioxide reacts with excess of ammonia a t low temperature
in an indifferent solvent, one obtains, according to Divers and Ogawa
(27), the colorless ammonium salt of imidosulfinic acid (LIX). Schumann
(108) thought that he had isolated the corresponding free acid (LX) from
the product of the reaction of ammonia with excess of sulfur dioxide. The
constitution of the yellow reaction product SOzNHI has, however, not yet
been elucidated, especially since its molecular weight cannot be determined.
0
[II,N-S-O]NHI
H2N-S-OH
el2
[0=8 (NH,)z][SO31
\-
GI
/s-NH3
10
(LIX)
(LX)
(Lm
(LXII)
182
The compound SOzNH3 could be (LX) or a thionyl ammonium sulfite

(65), but it may also be considered as a simple adduct of NH, and SO2
(LXII). It shows a strong tendency to disproportionation to compounds

with a sulfur oxidation number of +6 and others with sulfur in a lower
oxidation state (47).
Organic derivatives of compound (LX) are again well known and are
formed by addition of SO, to primary or secondary amines. A pure inorganic
derivative of (LX) is found in compound (LXIII), which is formed from
OSF, and animoriia as a gas and which is soluble in ether (52). The tautorncric forms (LXIV) and (LXV) are possible, though (LXIII) polymerizes
very readily to a yellow solid (LXVI).
(LXIII)
(LXIV)
(LXV1)
(LXV)
Organic derivatives of the type R*NS(O)F have also been prepared

from (1,XIII). The existence of an imidosulfuric acid (LXVII) is more
probttble than that of an nmidosulfuric acid. According to Ephraim and
Piotrowski (31),the ammonium salt of this acid (LXVIII) is formed when
SOa is added to excess of NHa. It seems that there are two compounds with
the composition SOn2NH3, one of which is formed a t very low temperatures
in the form of glistening plates (48),and a compound which is obtained a t
room temperature. The latter is probably the triammonium salt of imidodisulfinic acid.
0
II
HN
S-OH
S-OH
11
0
(LXVII)
S--0.
II
NHd
0
(LXVIII)
V. Sulfur lmides Derived from Sulfur Trioxide
A. SULFIMIDE
Vurious sulfur iniides based on sulfur trioxide are possible. Thus one,
two, or three oxygen atoms in the monomeric trioxidc could be replaced
by the isosteric NH group, and numerous imides can be formulated which
183
A M I D E S AND IMIDES OF THE OXYACIDS OF SULFUR
are derived from the various polymers of sulfur trioxide by a similar

replacemelit of 0 by NEE. IIantzsch aiid Hull ( G I ) were the first to prepare
such a polymeric sulfirnide. They obtaiiied derivatives of the sulfimide
(LXIX) both from the products of the decomposition of sulfuryl chloride
with ammonia and also from the product of a rearrangement which sulfuryl
amide undergoes when heated to 180-200.
The exothermic reaction which occurs when sulfuryl amide is heated
was discovered by Traube ( I l l ) .
OzS(NHd2 + N H I ( N S O ~
HN
0 2
0 2
/ \Ni
\NH
so2
02s
0 2
@02S
CH3-N
li
b028
N
H
(LXX)
H
(LXIX)
N-CHa
02s
\-/
N/
/ \
so2
N/
CH3
(LXXI)
The ammonium salt which results from the transformation of sulfuryl

amide can be converted to a silver salt, AgNS02H20, by silver nitrate (64)
and this readily gives AgNS02. The silver salt with methyl iodide yields
the N-methyl derivative (LXXI) (61). Since molecular weight determinations show the methyl derivative to be a trimer, it may also be assumed
that the salts, and particularly that of silver, are also trimeric. Attempts to
isolate the hydrogen compound (LXIX) have so far failed, but Heinze
and Meuwsen (64) have prepared many salts of (LXIX). They also were
able t o show that when the silver salt was decomposed with an equivalent
quantity of hydrochloric acid a tribasic acid resulted which was to some
extent stable in water. It readily lost two protons, hut one proton was
split off only with difficulty. The ion (LXX) clearly has a high stability.
The dipyridinium salt of the trimeric sulfimide, which is derived from the
ion (LXX), can be obtained ( l a ) in a yield of 70% by reaction of amidosulfuric acid chloride with pyridine at 20:
3HZNSOzCI
+ 5CsHsN
(0zSNH)a 2CsHsN
+ 3[CsHsNH][Cl]
Since sulfimide is a derivative of sulfur trioxide, it was natural to

attempt its synthesis by direct ammonolysis of sulfur trioxide. Appel and
Goehring (4) reacted NH, with excess so3 in nitromethane solution. The
product was (NH4)2S3010,
formation of which shows that a process yielding
water occurred during the ammonolysis.
+
+
SO3 NH3 --+ HNSOz H20

3SOa H20 ---t H2S3010
HaSiOla 2NH3 -+ (NH&SaOio
184
Actu:tlly a suhatanoe W ~ Hprcseiit in thc nitronirtharie solution which

was acid in water arid for which the coniposition (IINSO,), * HzS04was
adduced. Various silver salts could be precipitated from the aqueous solution with silver nitrate and could be separated by their solubility. In addition to the silver salt of the trimeric sulfimide, (AgNSO&, a further silver
salt of the formula AgNSO, was obtained which was converted by methyl
iodide into the tetramer of CH3NS02 (LXXII), thus proving the existence
of derivatives of tetrameric sulfimide (LXXIII), as well as of the trimer. A
salt of LXXIII is also obtained from H2NS02CI and pyridine ( l a ) .
0%
CH3-N-S-N-CH3
I
I
0 2
I
I
802
CHa-N-S-
H-N-S-N-H
N-CTIS
0 2
H 0
2
HO,H -N-SI?H
'
Ip'
H-N-S-
-H
O*
(LXXIII)
0 2
(LXXII)
(LXXIV)
Thc quantity of trimeric and tetrameric sulfimide rccovercd from the

reaction products of NH3 with excess SO1 in nitroinethane did not in any
way correspond with that expected from the equations given above and
only 10% of the quantity of sulfimide expected was isolated. The residue
which had the composition (HNS02),H2SOr proved to he polysulfimide
sulfonic acid (LXXIV) (5). This may be regarded as an imido derivative
of the polysulfuric acids. Amides of these acids have already been obtained
by Ephraim and Michel (SO) by decomposing sulfuryl chloride with
tmmonia.
Definite sulfimide sulfonic acids are formed from SO3 adducts and sulfamide in a melt the pyrid nium salt of the sulfamide disulfonic acid,
HO(02SNH)2-S03H was obtained. The arialogous reaction with imidosulfamide results in the dipyridinium salt of trisulfimide sulfonic acid
HO(O&3NH)3-S03H. Besides the pyridinium salts, several potassium,
ammonium and sodium salts were prepared (92a).
B. SULFANURIC
ACID
As Hantzsch and Stuer (62) suggested, trimeric sulfimide (LXXVa)

should have a n isomer (LXXVb), to which the name sulfanurie acid has
been given. However, this acid, like sulfimide, cannot be isolated as the
hydrogen compound itself, though the acid chloride (LXXVI) is readily
accessible. It can be made by the reaction of amidosulfonic acid with
PCI, and thermal decomposition of the product (73, 7 4 ) .
HO-SOZNH2
3Cl-SO2
+ 2PCIs -+
3HCl+ OPCIa CI-SO2-NPCIs

. NPCla + 3POCls [Cl-S(O)N]s
AMIDES AND IMIDES O F THE OXYACIDS O F SULFUR
185
A second method is to decompose a mixture of sulfuryl chloride and

thionyl chloride with ammonia (44). It is also possible to obtain sulfanuryl
chloride by oxidizing thiazyl chloride (LXXVII) with sulfur trioxide (50).
0 2
H-N
S
/ \
02s
N-H
SO2
\ /
N
H
(LXXVa)
(LXXVb)
c1
//
SO8
(LXXVI)
\N
c---
I1
cl-s\N/s-cl
(LXXVIl)
Sulfanuric acid (LXXVb) should be the first product of the hydrolysis

of sulfanuryl chloride (LXXVI), but this substance is clearly very unstable.
Even on careful acid hydrolysis, imidosulfamide (LXXVIII) and sulfuric
acid are formed immediately.
H2N . SO2 . NH . SO2 * NH2
(LXXVIII)
It is striking that compound (LXXVa) is clearly much more stable than

the isomer (LXXVb) and that hydrolysis leads to different products.
Whereas sulfimide in an acid medium gives sulfuryl amide, amidosulfonic
acid and sulfuric acid, the hydrolysis of sulfanuric acid leads to imidosulfamide (86).
VI. Amides of Oxyacids of Sulfur ( V I )
A. SULFURYL
AMIDE
Sulfuryl amide (LXXIX), the diamide of sulfuric acid, has been known
for a very long time. It was discovered by Traube (111, 112) and isolated
as a colorless solid; mp, 92-93". It is formed in a quite normal way both in
the ammonolysis of sulfuryl halides (see, for example, refs. 22,26,29,7Y,IOS,
119) and also in the ammonolysis of sulfur trioxide (6). It is common to
both modes of preparation that other reactions occur besides the formation
186
of sulfuryl amide. Thus sulfuryl fluoride reacts fairly smoothly to form

sulfuryl amide (IIS),but sulfuryl chloride is able, according to the conditions and also by using excess of ammonia, t o give salts of sulfimide (LXIX),
imidodisulfaniide (LXXVIII), or long-chain sulfuryl imido amides (LXXX)
(21, SO), in addition to sulfuryl amide. The situation is similar in the
reaction between sulfur trioxide and ammonia, when only by working
with excess of liquid ammonia is an appreciable yield (tipproximately
10%) of sulfuryl amide formed (6).
H2N-?[-N-S-]nk%-NHz
1 0 2
Oa
HzN-S-NHz
(LXXlX)
(LXXX)
According to current views on the forniation of sulfuryl amide (6, 44), a

condensation reaction probably leads in each case to the formation of
(LXXIX).
0zSCI2e OZSCl+ c1-+ (0zSNHz)'

HCI
2NHs & NHr+ NHBOzSNHz+ NIIt- + HzN - SO2 - NH2
OzSCI'
+ HNH2
+
OzSO
+ HNHz
OzSNHz
-t 09SNHz+
OH2NH3 NHi+ NHzNH2- 3 HzN - SOz - NIIz
Since, however, both SO, and the ion OnSC1+,which is analogous to sulfur
trioxide, are not only capable of undergoing condensation reactions but, as
Lewis acids, also have a strong tendency to undergo addition reactions, the
formation of sulfuryl amide cannot be the only reaction of these substances
with ammonia.
An interesting and almost quantitative formation of sulfuryl amide is
based on the use of sulfuryl diisocyanate (LXXXII), which can be made
by the action of cyanogen bromide on sulfur trioxide (GO),or by the
reaction of chlorosulfonyl isocyanate (LXXXI), with silver cyanate (3).
(LXXXI)
FI
H O
N-CbO
\
N -C=O
I3 0
H
Scheme 1
N=C=O
(LXXXII)
AMIDES AND IMIDES O F THE OXYACIDS O F SULFUR
187
Hydrolysis of this substance leads by intermediate steps t o sulfamide

(Scheme 1).Ammonolysis, on the other hand, yields a cyclic mixed derivative of sulfuric acid and carbonic acid (LXXXIII). Sulfuryl amide gives
salts and also numerous organic derivatives. A silver salt (AgNH)ZS02
(LXXXIII)
is readily obtained in an aqueous medium. Sulfuryl amide is a strong acid

in liquid ammonia and both a dipotassium salt, (KNH)zS02, and also a
tetrapotassium salt, (KzN)$02, are formed by reaction with KNHZ.
The number of organic derivatives is very large, for all four H atoms of
sulfuryl amide can be replaced by alkyl, aryl, or acyl groups, either singly
or together. The tendency of sulfuryl amide to stabilize itself by forming a
six-membered ring, which is seen in the formation of the ring system of
sulfimide on heating sulfuryl amide (compare p. 183), is also apparent in
many reactions with products containing carbon; for example, in the formation of (LXXXV) from dicyanamide and amidosulfonic acid (114). A
sulfuryl amide with an adamantine structure (LXXXVI) which is produced
0 2
0 2
0 2
N/
CI-c
I1
N'
I/
c-c1
'o/
(LXXXIV)
(LXXXV)
(LXXXVI)
from sulfuryl amide, ethylene diamine, and formaldehyde as a white solid

(mp, 195-196"), can be used in the analytical detection of sulfuryl amide
(100).
188
B. IMIDODISULFAMIDE
Imidodisulfamide (LXXVIII) is formed extraordinarily readily from

sulfuryl amide when it is heated for some time with dilute sodium hydroxide
solution (76). The resulting sodium salt is decomposed with sulfuric acid
and the iniidodisulfamide dissolved out with acetone. Like sulfuryl amide,
imidodisulfamide is also a colorless solid (mp, 169') which dissolves in water
and in some organic solvents, such as acetone or ethyl acetate.
0 2
HzN-S-NHz
12
+ H2N-S-NHt
(
H
+ H2N-S-N-S-NHt
0 2
-+ NHa
0 2
(LXXVIII)
Imidodisulfamide is a strong acid, which when titrated with sodium hydroxide solution behaves as if it were monobasic. Particularly the hydrogen on
the imide nitrogen atom is readily replaced by metals, though with silver
ions also a tri-silver salt, AgNH-SO2-NAg-SO2-NHAg,
is obtained
(73, 7 6 ) .
C. AMIDOSULFONIC
ACID
When sulfuryl amide or imidodisulfamide is heated with strong sodium
hydroxide solution, the sodium salt of amidosulfonic acid results. The free
acid itself may be prepared in various ways; for example, froni urea and
sulfuric acid, from many adducts of sulfur trioxide and ammonia, from
hydroxylammonium salts and sulfur dioxide, or from SO? and acetoxime.
Arnidosulfonic acid, which is colorless and melts at 205' wit,h decompouition, has, as Baumgarten supposed ( I S ) , the structure (LXXXVII).
I1 0
1 i[
1- @N-b--OB
I/
H O
(LXXXVII)
1
H-N@=S=O
I
08
H O
I I1
t)
H-N-S-OB
- Ii
(LXXXVIII)
According to crystal structure investigations of Kanda and King ( 7 l ) ,

the S-N distance is 1.73 A and the S-0 distance is 1.48 A. The amidosulfonate ion, in which the S-N distance is smaller (66,72),has the formula
(LXXXVIII).
Amidosulfonic acid is a strong monobasic acid in water but is dibasic
in ammonia (24) and reacts, for example, with sodium, forming the salt
NaNH SOaNa. The two hydrogen atoms on nitrogen are readily substituted if one allows it t o react with mercury(I1) salts: one can, for example,
prepare a salt HgN-SOINa or HgN-S03K.
189
The derivatives of amidosulfonic acid are very numerous. The acid

halides will be considered particularly here. Amidosulfonic acid chloride
may be prepared in the following way:
Chlorosulfonyl isocyanate (LXXXI) is formed in the reaction of
cyanogeri chloride with sulfur trioxide in a 1:1 molar ratio (57, 58), and
this is saponified by water vapor (59) or formic acid ( 2 ) to amidosulfonic
acid chloride (LXXXIX), H2NS02C1. It is colorless and crystalline and
melts at 40-41'. The reaction with water is explosive, amidosulfonic acid
and HCl being formed.
H,O
0 2
O=C=N-s-Cl-
0 H 0,
[HOC--N - S - C j
-H2N--S--C1
(LXXXI)
02
+ C02
(LXXXIX)
From amidosulfonic acid chloride the corresponding acid fluoride (XC)

may be made by reaction with potassium fluoride in boiling acetonitrile
(7), or by the reaction of (LXXXVI) with NaF followed by careful hydrolysis (70). Amidosulfonic acid fluoride is colorless, melts a t 8", and is soluble
0 2
H2N-S-F
(XC)
in organic solvents such as ether or chloroform. Towards water, in which it

dissolves, it is astonishingly stable. The aqueous solution hydrolyzes only
slowly t o amidosulfonic acid and hydrogen fluoride. Amidosulfonic acid
fluoride can also be obtained (2a) in a yield of 40% by ammonolysis of
disulfurylfluoride, SzObFz:
FSO2-O-SOzF
+ 2NH3 + FS03NH4 + NH~SOZF
Purely inorganic derivatives of amidosulfonic acid are accessible

through the reaction of HaN-SO, with PCls (73, 74) in which, as already
mentioned (compare p. 184) the compound (XCI) results.
0 2
Cl,P=N--S-CI
(XCI)
OSN-S03Me
(Me = Na, K)
(XCII)
The compound (XCI) is naturally, as in an acid chloride, sensitive to

hydrolysis. With water it yields amidosulfonic acid and phosphoric acid
quantitatively; the high stability of the nitrogen-sulfur bond, which is
manifestly greater than that of the phosphorus-nitrogen bond, is clear
from these results.
190
MARGOT RECKE-GOEHRING
Thionyl chloride reacts in an analogous manner to PC16 with mercury

amidosulfonate (QI), though not with free amidosulfonic acid. Thionyl
amidosulforiates (XCII) are quite stable and are clearly polymers.
D. IMIDOSULFONIC
ACID MONOAMIDE
Just as a diamide and a monoamide are formed from sulfuric acid,
it should be possible to prepare amides from disulfuric acid.
OH
NHz
02s
NHz
0 2
0 2
HO-S-0-S-OH
02 H 0 2
I~N--S-N-S-NHz
(LXXVIII)
SO
NH
/
02s
\OH
02 H 0 2
H2N--S-N-S-OH
(XCIII)
It would be expected that in the ammonolysis reaction the oxygen bridge

would disappear and, in its place, an NH bridge would be formed. From
this point, of view, imidodisulfamide (LXXVIII), which has already been
described, is a diamide derived from disulfuric acid. The question then
arises as to whether the corresponding monoamide (XCIII) can exist.
AppeI et al. (8) were able to prepare a diammonium salt of (XCIII) by
the reaction of pyrosulfuryl chloride, S20d212or trisulfuryl fluoride, SKhF2,
with ammonia.
(XCIV)
The compound (XCIV) is probably identical with the compound described earlier by Hayek et al. (68)as a diammoniuni salt of a pyiosulfamide. The pyridiriium salt of (XCIII) can be obtained too, when c6p 6N603
is allowed to react with an excess of siilfamide ( 9 2 ~ ) .
REFERENCES
I. Appel, R., Dissertation, Heidelberg (1951).

l a . Appel, R., and Rerger, G . , 2. anorg. Chem. in press.
2. Appel, H., and Berger, G., Ber. 91, 1339 (1958).
fa. Appel, R., and Eisenhauer, G., Angew. Chem. 70, 742 (1956).
3. Appcl, R., and Gerber, H.,Ber. 91, 1200 (1958).
4. Appel, R., and Goehring, M., 2.anorg. Chem. 271, 171 (1953).
5. Appel, R., and Huber, W., 2. nnorg. Chem. 276, 338 (1954).
6. Appel, R., and Huber, W., Ber. 89, 386 (1956).
7 . Appel, R., and Senkpiel, W., Angew. Chem. 70, 572 (1958).
8. Appel, R., Voigt, G., and Sadek, E. H.,
Natunoissenschaflen 21, 496 (1956).
9. Arnold, M. H. M., British Patent 544,577 (1942); U. S. Patents 2,364,414 (1944);
2,372,046 (1945); 2,382,845 (1945).
191
10. Arnold, M. H. M., J. Chem. Soc. p. 1596 (1938).

11. Arnold, M. H. M., Hugill, A. J. C., and Hutson, J. M., J . Chem. Sac. p. 1645
(1936).
12. Audrieth, L. F., and Kleinberg, J., Non-Aqueous Solvents. Wiley, New York,
1953.
Baumgarten, P., Ber. 62, 820 (1929).
Becke-Goehring, M., and Jenne, H., Angew. Chem. 70, 399 (1958).
Becke-Goehring, M., Jenne, H., and Fluck, E., Ber. 91, 1947 (1958).
Becke-Goehring, M., Jenne, H., and Fkkali6, V., Ber. 92, 855 (1959).
I?. Becke-Goehring, M., and Schwarz, R., Z. anorg. Chem. 296, 3 (1958).
18. Becke-Goehring, M., Schwarz, R., and Spisss, W., Z. anorg. Chem. 293, 294
(1957).
19. Berg, W., and Goehring, M., 2. anorg. Chem. 276, 273 (1954).
$0. Berg, W., Goehring, M., and Malz, H., Z. anorg. Chem. 283, 13 (1956).
21. Broja, G., German Patent 814,143 (1951).
82. Broja, G., German Patent 815,642 (1951).
23. Brustier, V., and Garcia-Fernandez, H., Bull. soc. chim. France p. 1299 (1955).
24. Butler, M. J., and Audrieth, L. F., J . Am. Chem. Soc. 61, 914 (1939).
26. Daum, K. W., Goehring, M., and Weiss, J., 2. anorg. Chem. 278, 260 (1955).
26. Degering, E. F., and Gross, G. C., U. S. Patent 2,407,481 (1946).
$7. Divers, E., and Ogawa, M., J . Chem. Soc. 77, 327 (1946).
28. Donohue, J., Caron, A., and Goldish, E., N d u r e 182, 518 (1958).
29. Ephraim, F., and Gurewitsch, M., Ber. 43, 138 (1910).
30. Ephraim, F., and Michel, F., Ber. 42, 3833 (1909).
31. Ephraim, F., and Piotrowski, H., Ber, 44, 379 (1911).
32. Faessler, A., and Goehring, M., Naturwissenschaften 39, 169 (1952).
33. Fluck, E., and Becke-Goehring, M., Z. anorg. Chem. 292, 229 (1957).
34. Fluck, E., and Goehring, M., Z. Naturforsch. l l b , 756 (1956).
36. Fluck, E., Goehring, M., and Weiss, J., Z. anorg. Chem. 287, 51 (1956).
36. Gee, G., Chem. Soc. (London) Spec. Publ. No. 12, 247 (1958).
37. Gehrig, O., Dissertation, Heidelberg (1953).
38. Goehring, M., Quart. Revs. 10, 437 (1956).
$9. Goehring, M., Ergebnisse und Probleme der Chemie der Schwefelstickstoffverbindungen. Akademie-Verlag, Berlin, 1957.
40. Goehring, M., and Daum, K., Z. anorg. Chem. 282, 83 (1955).
41. Goehring, M., Daum, K., and Weiss, J., 2. Naturfursch. 106, 298 (1955).
42. Goehring, M., and Debo, A., 2. anorg. Chem. 273, 319 (1953).
43. Goehring, M., and Ebert, J., Z. Ndurforsch. lob, 241, (1955).
44. Goehring, M., Heinke, J., Malz, H., and ROOS,G., Z. anorg. C h m . 273, 200
(1953).
45. Goehring, M., Herb, H., and Koch, W., Z. anorg. Chem. 264, 137 (1951).
46. Goehring, M., and Hohenschutz, H., Naturwissenschaften 40, 291 (1953).
47. Goehring, M., and Kaloumenos, H.-W., Z.anorg. Chem. 263, 137 (1950).
48. Goehring, M., Kaloumenos, H.-W., and Messner, J., 2. anorg. Chem. 264, 48
(1951).
49. Goehring, M., and Koch, W., Z. Naturforsch. 7b, 634 (1952).
60. Goehring, M., and Malz, H., 2. Naturforsch. 9b, 567 (1954).
51. Goehring, M., and Messner, J., 2. anorg. Chem. 268, 47 (1952).
52. Goehring, M., and Voigt, G., Ber. 89, 1050 (1956).
53. Goehring, M., and Weis, G., Angm. Chem. 68, 678 (1956).
13.
14.
15.
16.
192
MARGOT RECKE-GOEHRING
64. Goehring, M., Weiss, J., and Zirker, G., Z . anorg. Chem. 278, 1 (1955).
66. Goehring, M., and Zirker, G., 2. Naturforsch. lob, 58 (1955).
66. Goehring, M., and Zirker, G., 2.anorg. Chem. 286, 70 (1956).
67. Graf, R., Ber. 89, 1071 (1956).
68. Graf, R., German Patent 944,009 (1956).
69. GrtLf,R., German Patent 937,645 (1956).
60. Graf, R., German Patent 940,351 (1956).
61. Hantzsch, A., and Holl, A., Ber. 34, 3430 (1901).
68. Hantzsch, A., and Stuer, B. C.,Ber. 38, 1022 (1905).
63. Hayek, E., Engelbrecht, A., rind Wagner, F., Monatsh. 88, 939 (1952).
64. Heinee, G., and Meuwsen, A., Z. anorg. Chem. 276, 49 (1954).
65. Jander, G., Knoll, H., and Ipmig, H., 2. anorg. Chem. 232, 229 (1937).
66. Jeffrey, G. A., and Stadler, H. P., J. Chem. SOP.p. 1467 (1951).
67. Jenne, H., Dissertation, Heidelberg (1958).
68. Jmne, H., and Becke-Goehring, M., Ber. 91, 1950 (1958).
69. Jensen, K. A., and Hofmann-Bang, N., Ann. 648, 95 (1941).
70. Jonas, H., and Voigt, D., Angm. Chem. 70, 572 (1958).
71. Kanda, F. A., and King, A. J., J. Am. Chem. SOC.73, 2315 (1951).
72. Iietelaar, J. A. A., and Heilmann, E. L., 2. Krist. 103, 41 (1940).
73. Kirsanov, A. V., Communs. Acad. Sci. U.S.S.R. 426 (1950), see Chem. Zentr.
122, 3168 (1951).
74. Kirsanov, A. V., J. Gen. Chem. U.S.S.R.22, 88 (1952).
76. Kirsanov, A. V., and Zolotov, Y . M., J. Gen. Chem. U.S.S.R. 20, 1790 (1950).
76. Kirsanov, A. V., and Zolotov, Y. M., J. Gen. Chem. U.S.S.R. 22, 2038 (1952).
77. Kirsanov, A. V., Zolotov, Y. M., and Jegorova, N. L., J. Gen. Chem. U.S.S.R. 20,
2251 (1950).
78. Lengfeld, F., and Stieglitz, J., Ber. 28, 575 (1895).
79. Lcngfeld, F.,and Stieglitz, J., Ber. 28, 2742 (1895).
80. Levi, T. G., Guzzetta 61, 286 (1931).
81. Levi, T. G., Gazeetta 61, 294 (1931).
88. Lindqvist, I., and Weiss, J., J. Inorg. & Nuclear Chem. 6 , 184 (1958).
83. Lippincott, E. R., and Tobin, M. C., J. Am. Chem. Soc. 73, 4990 (1951).
84. Lund, E. W., and Svendsen, S. R., Acta Chem. Scand. 11, 940 (1957).
86. Macbeth, A. K., and Graham, H., Proc. Roy. Irish A d . 86, 31 (1923).
86. Male, II., Dissertation, Heidelberg (1964).
87. Medam, M., Bull. SOC. chim. France 16, 391 (1896).
88. Meslans, M., Chem. Ztg. I, 308 (1896).
89. Meuwuen, A., Ber. 62, 1959 (1929).
90. Mcuwsen, A,, Ber. 64, 2301 (1931).
91. Meuwsen, A,, Colloq. Intern. Union Pure and A p p l . Chem., Mcnster, 1954 p. 130
(1955).
92. Meuwsen, A,, and Losel, M., 2.anorg. Chem. 271, 217 (1953).
9ba. Meuwsen, A., and Reiohelt, H., Angew. Chem. 71, 162 (1959).
93. Meuwsen, A., and Schlossnagel, F., 2. anorg. Chem. 271, 226 (1953).
94. Michaelis, A., Ber. 28, 165 (1895).
96. Michaclis, A+,Rer. 28, 1013 (1895).
96. Miller, H. C., and Gall, J. F., Znd. Eng. Chem. 42, 2224 (1950).
97. Moissan, H., and Lebeau, P., Ann. chim. phys. 26, 145 (1902).
98. Moissan, H., and Lebeau, P., Bull. SOC. chim. France 27, 240 (1902).
99. Moissan, H., and Lebeau, P., Compt. rend. mad. sci. 130, 1436 (1900).
100. Paquin, A. M., Angew. Chem. 860, 319 (1948).
101. Piper, T. S., Chem. & Ind. 37, 1101 (1957).
10.2. Piper, T. S., J . Am. Chem. Soc. 80, 30 (1958).
103. Rommel, O., German Patent 838,743 (1952).
10.4. Ruff, O., and Geisel, E., Ber. 37, 1573 (1904).
1G6. Sass, R. L., and Donohue, J., Acta Cryst. 11, 497 (1958).
106. Schenk, P. W., Ber. 66, 94 (1942).
107. Schiff, H., Liebigs Ann. Chem. u. Pharm. 102, 113 (1857).
108. Schumann, H., 2. anorg. Chem. 23, 43 (1900).
109. Stone, B. D., and Nielsen, M. L., J . Am. Chem. Soc. 79, 1264 (1957).
109a. Stone, B. D., and Nielsen, M. L., J . Am. Chem. SOC.81, 3580 (1958).
110. Spiess, W., Dissertation, Heiddberg (1957).
1 1 1 . Traube, W., Ber. 26, 2472 (1892).
112. Traube, W., Ber. 26, 610 (1893).
11s. Traube, W., and Reubke, E., Bsr. 66, 1656 (1923).
114. Walter, H. A,, U. S. Patent 2,449,520 (1942).
116. Weiss, J., 2.Naturforsch. 12b, 481 (1957).
116~.Weiss, J., Angezo. Chem. 71, 246 (1959).
116. Weiss, J., and Becke-Goehring, M., 2.Naturforsch. 13b, 198 (1958).
117. Weiss, J., and Neubauer, D., 2.Nuturforsch. 13b, 459 (1958).
118. Wolbling, H., 2. anorg. Chem. 67, 281 (1908).
119. Wood, F. C., and Battye, A. E., Chem. & Znd. 62, 347 (1933).
193
HALIDES OF THE ACTINIDE ELEMENTS

Joseph J. Katz and Irving Sheft
Argonne National laboratory, Argonne, Illinois
I. Introduction .
. 195
11. Survey of Actinide Element Halides .
111. Preparative . . . .
A. General Considerations
B. Lower Fluorides . .
C. Higher Fluorides . .
D. Chlorides . . . .
E. Bromides and Iodides
195
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
202
202
. . . . . . . . . . . 205
. . . . . . . . . . . 211
. . . . . . . . . . . 219
. . . . . . . . . . . 221
. .
IV. Properties of Halides
A.Structure. . . . .
B. Volatility . . . . .
C. Change of State . . .
D. Thermochemistry
. .
E. Electronic Configuration
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . . . . . . . 224
. . . . . . . . 224
. . . . . . . . 226
. . . . . . . . . . . . . .
228
. 229
232
I. Introduction
The binary halides of the actinide elements constitute a most important

class of compounds. The actinide elements themselves are, of course, vital
to all presently practical means for releasing nuclear energy, and the halides
of these elements are leading protagonists in nuclear technology. And, as
is the case for other groups of elements in the Periodic Table, the halides
are particularly convenient for many kinds of scientific investigations and
correlations. Much of what is now known of the electron configuration of
the actinide elements, for example, has been learned by experiments with
halides. A recapitulation and reassessment of the halides of the actinide
elements in the light of recent developments thus becomes appropriate.
The present review will be confined to the solid, binary halides, and principal attention will be focused on developments which have occurred since
the subject was last reviewed (51). For an interesting and authoritative
review of the entire subject of the actinide elements, the recent account of
G. T. Seaborg (74) can be warmly recommended.
I I . Survey of the Actinide Element Halides
The actinide elements now constitute a fairly numerous group of compounds (Table I). An examination of Table I reveals a number of interesting
195
196
JOSEPH J. KATZ A N D IRVING SHEFT
regularities as well as a number of puzzling and unresolved features. It is

well known that the actinide elements exhibit a much greater tendency to
exist in multiple oxidation states than do the related lanthariide elements.
Table I1 summarizes the known oxidation states of the actinide elements
and demonstrates their greater ease of oxidation. It also permits a comparison between the oxidation states of a particular element and its known
halides. It is immediately obvious that uranium, neptunium, and plutonium, which possess the greatest tendency to multiplicity in oxidation
TABLE I
HALIDES
OF
Ac
Th
Pa
THE
ACTINIDEELEMENTS"
NP
Pu
PuFs
PuF~
PUd?,?
Br
AcBra
Am
Cm
AmF3
AmF4
CmF3
CmF4
[PUF',]
PuFB
PuC13
AmCh
PuBra
AmBr
PUIS
Am4
ThBrc
a Parentheses indicate an incompletely characterized compound. Square brackets

indicate compounds presumed capable of existence but not as yet prepared.
state, form the largest number of halides. The ability t o form higher oxidation states diminishes rapidly in proceeding to elements of higher atomic
number; this is reflected in the difficulty in producing higher halides with
higher members of the actinide series. In addition to the regularities that
are evident, there are also a number of important instances where u11expected discrepancies occur. I t may therefore be instructive to examine
the situation in detail.
Little need be said about the halides of actinium. All of the binary
halides of this element have been prepared by conventional procedures,
and the straightforward, uncomplicated chemical behavior that actinium
exhibits generally is also evident here. Thorium, however, presents quite
a different problem. Thorium gives no indication of existence in aqueous
solution in an oxidation state lower than f4. From considerations of the
197
HAL I DE S OF T H E ACTINIDE E L E M EN TS
oxidation potentials it appears quite certain that Th(II1) would reduce

water strongly, and lower oxidation states of this element appear to be
incapable of more than a transitory existence in aqueous media. The situation with respect t o possible lower valent solid compounds is more confused, and despite the activity of a number of workers, it is still not possible
to give a n unequivocal answer to the question of the existence of lower
halides of thorium. It appeared from the early work of Warf (85) that
thorium tetrafluoride is not reduced to ThFB on heating with thorium
TABLE I1
OXIDATION
STATESOF THE ACTINIDE ELEMENTS"
Atomic number
Element
Oxidation states
89
Ac
3
90
Th
(3)
4
93
Np
3
(3)
3
4
4
4
4
6
5
6
6
6
91
Pa
92
94
Pu
3
(
5
6
95
96
Am Cm
3
3
4
)
4
5
6
97
Bk
3
98
Cf
3
0 The characteristic, stable oxidation state of the element in solution is indicated by

boldface type. A doubtful, or unstable, state is enclosed in parentheses.
metal at 1175' and 1600'. On the other hand, Anderson and D'Eye (4),
Hayek et al. (423,44) and Jantsch et al. (48) all reported the preparation of
lower chlorides, bromides, and particularly iodides by thermal decomposition of the appropriate tetrahalide, or by reduction of the tetrahalide with
metallic thorium or aluminum at elevated temperatures. Iodides of the
composition ThI, and ThIz were described as dark metallic grey in color
with a violet or brownish tinge, and were considered to be analogous to
the lower iodides of zirconium that are also described in the literature
(69, 90).
Watt et al. (84) have carried out a very careful investigation of the
reduction of thorium tetraiodide by metallic thorium. I n these experiments,
known weights of iodine and thorium, corresponding to different reaction
ratios, were mixed and heated a t 550' for 3 days. Any tetraiodide present
was removed by sublimation, and the nonvolatile residue was separated
into a series of fractions by treatment with liquid ammonia, aqueous perchloric acid, and dimethyl formamide. Compositions with a I/Th ratio of
3 dissolved in liquid ammonia with evolution of hydrogen; thorium(1V)
amidotriiodide 4-ammoniate, Th(NH2)13 .4NH3, could be recovered from
the liquid ammonia solution. However, in experiments where I/Th ratios
varied from 2 to 1, the experiments were interpreted as indicating the
presence in the nonvolatile reaction product of ThIz and ThI, the first of
these insoluble in liquid ammonia but soluble in anhydrous dimethyl
198
JOSEPH S. KATZ AND IRVING SHEFT
formamide, the latter being insoluble in both solvents. Aqueous perchloric

acid was employed to separate the lower halides from unreacted thorium
metal, because the halides were reported to dissolve in this reagent whereas
the metal does not. Despite the care with which the experiments were
carried out, several questions remain. The original reaction product is
clearly not a pure phase, and inferences based on extraction into dimethyl
formamide or perchloric acid do not rigorously establish the formulas of the
phases present in the reaction product. Even dilute aqueous perchloric
acid has considerable oxidizing power (46),and it is not immediately apparent that an aqueous extract with an I/Th ratio of 1 necessarily implies
that the original material treated had the composition ThI. Further, Watt
and co-workers remark that T h L reduces liquid ammonia readily, but the
lower iodides do not; it might be anticipated that ThIz arid ThI should be
even more powerful reducing agents than ThI,. Young (90) found that
ZrIodissolved in water to give a solution with transient but strong reducing
properties. Similar experiments do not appear to be recorded for ThI,.
Whereas ThI, was described by Anderson arid D'Eye (4) as a volatile substance, and by Hayek et al. (44) as undergoing disproportionation over the
temperature range 500-600", Watt and co-workers find it to be essentially
nonvolatile and stable to disproportionation. Particulttrly puzzling is the
absence of definitive X-ray diffraction data, despite the fact that the
preparntions are made under conditions which should foster crystal growth.
Anderson tind D'Eye obtained an X-ray diffraction pattern for ThI, and
two different patterns for the phase designated ThI,; the diiodide structure
was interpreted in terms of a n hexagonal layer-lattice structure similar to
other metal diiodides. Hayek et al. (44) made the somewhat surprising
statement that the lower thorium iodides are so sensitive to X-rays as to
precludc crystallographic study, while Jantsch et al. (48) were unable to
obtain crystal structure data either by X-ray methods or by electron diffraction. The study of the lower halides of thorium poses a serious experimental problem, necessitating as it does the separation and characterization of solid, nonvolatile mixtures of compounds. An answer might be more
readily forthcoming if alternative synthetic procedures that would yield
single products could be devised. It is to be hoped that the investigators
active in this field will persist in their difficult task in the interest of clarifying the situation.
The gaps in the list of protactinium halides are due not so much to the
instability of the missing compounds but rather to difficulties in characterization. Until very recently, protactinium was the least studied of the
actinide elements, primarily because of the great difficulties hitherto experienced in obtaining milligram amounts of pure protactinium-231. The lack
of pure protactinium and the difficult chemistry of the element have made
199
for slow progress. Nevertheless, sufficient work has been done to show conclusively that halides of both protactinium(1V) and (V) exist as stable,
and in the case of the (V) compounds, volatile compounds. The absence of
protactinium bromides from Table I therefore can be expected t o be soon
remedied. The parentheses enclosing PaC16 in Table I reflect analytical
difficulties; the existence of a volatile higher chloride of protactinium
appears unequivocal, but as it is not isostructural with the corresponding
niobium and tantalum compounds, and in the absence of conventional
analytical data, the formula of the volatile higher protactinium chloride
is not regarded as fully established.
Uranium possesses the largest and best studied array of halides of any
of the actinide elements. Both uranium tetrafluoride and uranium hexafluoride have considerable technological significance. Uranium tetrafluoride
is produced on a large scale for the preparation of uranium metal which in
turn is used as fuel in nuclear reactors. The tetrafluoride also serves as a
starting material for the preparation of uranium hexafluoride. Uranium
hexafluoride, by virtue of its high volatility, is used in the thermal diffusion
separation of the isotopes U235and U238.Consequently, uranium tetra- and
hexafluorides have been studied very intensively. While the situation
with respect t o physical data is quite satisfactory, the situation regarding
chemical behavior is in a much less satisfactory position, and only in the
very recent past have chemical studies with UFa been taken up. The
uranium halides are also unique in the existence of a group of compounds
that have been designated intermediate fluorides. Since discovery of the
intermediate fluorides some 15 years ago, Iittle has appeared in the literature. Intermediate in composition between UF, and UFs, the compounds
U4Fl,(UF4.26)and U Z F ~ ( U F ~are
. ~ Oobtained
)
by reaction of uranium tetrafluoride with uranium hexafluoride ( 2 ) . The existence of these compounds
raises the question as to whether similar compounds exist intermediate
between UF6and UF6, and whether other actinide elements will form intermediate halides. Mandleberg et al. (58) have reported the existence of an
intermediate fluoride of plutonium which they believe is Pu4F17. Although
not obtained pure, and although X-ray identification is incomplete, it is
likely that the compound is indeed an intermediate fluoride. This interesting observation would seem to merit further study.
The neptunium halides illustrate the dictum that higher oxidation states
become more difficult to attain for the higher actinide elements. Although
on the whole neptunium resembles uranium rather than plutonium, the
lower halides of neptunium are distinctly more stable than the corresponding uranium compounds, and distinctly more severe oxidizing conditions
are required to prepare NpF,. Although uranium tetraiodide can be prepared, it is a relatively unstable substance; the corresponding neptunium
200
JOSEPH J. KATZ AND IRVINQ SHEFT
compound appears to be unstable to the extent that it does not exist under
ordinary conditions. Perhaps the most curious feature of the neptunium
halides is the absence of a neptunium pentafluoride. Despite a not unconsiderable amount of research, no compound NpF6 has been characterized,
although the (V) oxidation state is ordinarily the most stable neptunium
oxidation state in solution. The stability of the neptunium (V) state
would thus appear to arise from hydration and oxygenation.
In the case of plutonium, attainment of higher oxidation states is much
more difficult than for either uranium or neptunium, It is only in the case
of the fluorides that it is possible to prepare other than a tripositive halide.
Repeated attempts have been made to prepare plutonium tetrachloride,
as for example by subjecting plutonium trichloride to the action of chlorine
under high pressure, and it has been concluded that the dissociation pressure of PuC14 to PuCla and chlorine at room temperature must be of the
order of several thousand atmospheres. It thus appears unlikely that
tetrachlorides or tetrabromides of plutonium will be capable of synthesis.
The vacancies in the list of plutonium halides thus occur, not by default,
but because of the instability of the higher halides. As in the case of neptunium, no plutonium pentafluoride has been identified. Even though the
(V) oxidation state of plutonium is much less stable than is the case for
neptunium, it would appear probable that PuF6 should exist as a stable
entity.
The (111) state is by far the most stable oxidation state for americium,
and it is possible only by the use of fluorine to prepare a higher halide. Americium tetrafluoride has been prepared by direct fluorination, but attempts
to prepare AmF6 or AmFB have so far been fruitless. It is a reasonable
presumption that AmFBdoes not exist as a stable compound under ordinary
conditions.
Curium resembles americium in the difficulties attendant on preparing
higher halides. Here also it would appear that curium tetrafluoride is the
highest stable halide, and although no halides other than the fluorides have
as yet been prepared, curium trichloride, tribromide, and triiodide should
exist. It must be pointed out, however, that curium is an intensely radioactive element. All of the earlier chemical studies were carried out with
Cm242,
a 162.5day alpha emitter. Efforts to prepare curium tetrafluoride
using this isotope failed, but it was possible to prepare the higher fluoride
using one of the longer lived curium isotopes, Cm244,
a 19-yr alpha emitter.
Presumably Cm244, or an even longer lived isotope such as curium possesses,
may be required for the preparation of the other halides of curium.
Table I11 contains a list of the isotopes of the actinide elements with
half-lives for radioactive decay sufficiently long to permit the manipulation
20 1
of macroscopic quantities. At the present, it does riot appear likely that

actinide elements above einsteinium will have isotopes sufficiently stable
to permit of work on the macro scale.
The isotopes A+', Th232,Pa231,and U238occur in nature. The remaining
isotopes listed in Table I11 are purely synthetic. Most of the chemical
investigation carried out to date on neptunium, plutonium, americium, and
Am241,and Cm242.Plutonium,
curium have utilized the isotopes NpZ3',Pu239,
americium, and curium are known, however, to have long-lived isotopes.
TABLE I11
SUITABLE
FOR INVESTIGATIONS
ACTINIDE
ELEMENT
ISOTOPES
WITH MACROSCOPIC
QUANTITIES
Element
Ac tini um
Thorium
Protact,inium
Uranium
Neptunium
Plutonium
Americium
Curium
Berkelium
Californium
Einsteinium
Isotope
Ac227
Th232
Pa231
U238
NpZ3'
P1PY
IJP
Pu244
Am2"
Am243
Cm242
Cm244
Cme4'
Cm248
13k247
Bk249
Cf249
Cf25'
E264
Half-life
22.0 yr
1 . 3 9 X 10loyr
34,300 yr
4 . 5 0 X lo8 yr
2 . 2 0 X lo6 yr
24,360 yr
3.79 X lo6 yr
7 . 6 X lo7 yr
458 yr
7,600 yr
162.5 day
19 yr
2 4 X 107 yr
4 . 7 X lo5 yr
7 X lo3 yr
314 day
360 yr
660 yr
280 day
Alpha disintegrstt,ion
per minute per milligram
1.59 X
2.47 X
1.00 x
7.42 x
1.52 X
1.36 X
8.66 X
4.29 x
7.19 x
4.29 X
7.38 x
1.72 X
<8.04 X
6.79 X
4.60 X
3.71 X
8.86 X
4.80 x
4.07 X
1011
102
108
102
lo6
lo8
loc
104
109
lo8
10'2
10"
lo4
loc
lo8
lot2
lo9
109
10l2
When these isotopes become available, chemical studies will be greatly

simplified, and the complications introduced by the radioactivity of the
actinide elements will be substantially minimized. The longest-lived isotopes of berkelium, californium, and einsteinium are still fairly short-lived
substances, and macroscopic amounts have a tremendous associated radioactivity. Nevertheless, it should eventually be possible to prepare and study
the solid halides of the actinide elements through the element einsteinium
using weighable amounts of reactants. This remains for the future, however.
The special experimental problems associated with highly radioactive substances are considered below.
202
JOSEPH J . KATZ AND IRVING SHEFT
I l l . Preparative
A.
GwlcriAL
CONSIDEIIATIONS
While the reactions employed in the synthesis of the actinide element

htdides are for the most part quite conventional, the actual experimental
procedures are unusual and distinctive. This circurnstarlce arises from the
radioactivit,y of the actinide elements, and the chemical and physiological
consequences which ensue. Referring to Table 111, it can be seen that the
emission of alpha particles by the shorter-lived actinide isotopes is prodigious. Alpha particles do not usually pose an external radiation threat,
but ingestion is quite another matter. For example, Pu239inhaled into the
luiigs is to a considerable extent transferred to and deposited in the skeleton. Long continued alpha emission in the bone in a localized region may
result in the development of bone cancer, and for the human about one
microgram of Pu239fixed in the skeleton is considered a potentially lethal
dose (55). While all of the actinide elements are not metaholized in the
smie way, nevertheless, all except uranium are easily fixed in the skeleton,
and all present an ingestion hazard. Means have therefore been devised to
miriirriize rontact between the investigator and his experimental material.
Figure 1 shows an arrangement frequently used for the manipulation of
intense alpha-emitting isotopes. Essentially a ventilated hood (fume cupboard) for radiochemical research, i t is converted for use with alphaemitters by the insertion of u glove panel. This seals the hood, and provides
tt set of gloves through which experimentatiori is carried out. A plastic
port for introducing and removing itpparatiis and samples from the hood is
is also provided. When the hood must be opened, air is drawn from the face
of the hond through the set of air filters shown in the back of the hood to
keep contnminants from leaving the hood. The particular hood illustrated
is equipped for the manipulation of plutonium and fluorine. Another piece
of apparatus that finds extensive use is shown in Fig. 2. This is a ventilated
vacuum frame. Many of the preparations of actinide halides are curried
out by vacuum line manipulation, and often the vacuum line is too large
to fit into a gloved-box. The entire vacuum line can be placed in the enclosure shown and very large amounts of air can be moved past the apparatus.
The door slides and the entire container may be closed, or any particular
panel may be moved to allow access to a portion of the appuratus. In rase
of accident, the ventilation provisions are adequate to prevent material
from entering the room. Most of the experimental work with the extremely
hazardous plutonium hexafluoride has heen carried out in such an enclosed
vacuum line. l3agnall (8) gives a very useful description of laboratory
techniques applicable in marly cases to work with actinide compounds.
203
Micro techniques have been very useful in chemical studies of the

actinide elements, and particularly so for the study of the halides. Historically, micro methods were required because of the very small amounts
of the synthetic actinides originally available for experiment. However,
even when large amounts of neptunium, plutonium, and other actinide
elements became available, micro methods were found to offer many
advantages, and they have therefore continued to be of great service to
this day. W. H. Zachariasen was able to obtain and interpret X-ray diffraction patterns on samples of the order of a few micrograms and the
FIG.1. Ventilated hood for radiochemical manipulation.
results obtained in this way have been of the greatest utility in the identification of new compounds. The hazards attendant on the preparation of
compounds of such short-lived isotopes as AcZz7or Cm242are greatly decreased by such techniques because such very small amounts are used. In
the case of' AcZz7,which generates a host of beta- and gamma-emitting
daughters by radioactive decay, it is a necessity and not a convenience to
employ microgram amounts. When large amounts of AcZz7are used, conventional X-ray diffraction methods are no longer applicable, as the beta
and gamma background produced by the daughter activities fogs the X-ray
film. While special methods can be employed to obtain X-ray diffraction
204
JOSEPH J. KATZ ANI) IRVING SIIEPT
data in the presence of a considerable gamma background, nevertheless,

it is much more convenient to resort to micro methods.
Capillary techniques can be successfully used in the range of a few
micrograms to perhaps a milligram of sample. The essence of the capillary
technique is to use the X-ray capillary as the reaction vessel. The preparative and purification procedures are carried out in the capillary tube used
FIa. 2. Ventilated vacuum frame hood.
to secure the X-ray diffraction pattern, thus avoiding losses in transfer

and possible exposure of the sample to air or moisture. The application of
capillary techniques to the synthesis of actinide element compounds has
been described by Fried and Davidson (34) and Sheft and Fried (76). For
many purposes, experimentation on about the milligram scale appears to
be optimum in scale. This is particularly true for work with neptunium and
205
plutonium, and much of the cliemistry of the solid compounds of these

elements has been learned from work on the milligram scale. More convcntional chemical procedures may be employed 011 the milligram scale but
here also X-ray methods are basic.
As might be expected, the bulk of the research activities relating to
halides of the actinide elements has concerned itself with compounds
important in nuclear technology. The preparation and properties of UF4,
UFs, and PuFBhave consequently received much attention. Another reason
for the interest in fluorides is the ability of fluorine to evoke highest oxidation states. The actinide elements themselves have not all attracted the
same amount of study; thus, nothing has appeared in the literature on
actinium since the pioneering study of Fried et al. (36). The same is true
for the protactinium halides where nothing appears in the literature subsequent to the original communication of Sellers et al. (76). Presumably
this situation is a consequence of the great difficulty in obtaining pure
protactinium, and it is reasonable to assume that the paucity of information
on the protactinium halides will soon be rectified because of the success of
Maddock and co-workers in isolating large amounts of pure protactinium
(40)*
B. LOWERFLUORIDES
The lower fluorides of the actinide elements are prepared by reaction
with hydrogen fluoride, either in aqueous solution or at elevated temperatures in the dry way. The latter is the more usual procedure. The methods
used for the preparation of UF, are typical of those used for the preparation
of the lower fluorides of any of the actinide elements. The usual starting
materia1 for the preparation of uranium tetrafluoride is the trioxide, UOa,
obtained by thermal decomposition of urariyl nitrate hexahydrate,
U02(N0& 6H20. The trioxide is reduced by hydrogen to uranium dioxide,
which is in turn hydrofluorinated to UF4:
IJOs
UOz
+ Hz
+ UOz
120,
+ 4HF + UP, + 2Hz0.
For the other actinide elements the usual starting material will be the
dioxide. Although these reactions appear to be simple, they are in fact
quite complex. Uranium trioxide as prepared by thermal decomposition
of uranyl nitrate hexahydrate, uranium peroxide, or ammonium diuranate
may exist in any one of at least four crystal modifications and in an important amorphous form. The reduction of UOS appears to be related both to
the crystal form and to the surface area, and depending on these and perhaps other still unspecified variables, the hydrofluorination reaction on the
206
JOSEPH J.
IiKrz
AND IRVING SHEFT
uraiiiuin dioxide so olhtiiied may procccd smoothly to essciitially oomplete reactiori, or only very slowly and incompletely. Iiiihlman and Swinehart (62)have reported a detailed study of the factors involved in obtaining
quantitative hydroflnorinntiori of UOt. The rate of hydrofluorination of a
given sample of uranium dioxide depends upon the temperature a t which it
has been prepared (Fig. 3). If UO, is reduced a t a temperature above
6501 apprecitible sintering occurs, and the dioxide obtained is of low surface area and low reactivity. A number of competing processes also are
I00
80
D
W
c
o
a
60
0
TIME (minutes)
FIQ.3. Rnle of hydrofluorination of uraniuiii dioxido snmplcs at 575" (56).Surriplc A ,

UOa rtduccd at, 4-50";B , at 540"; C , atJ780".
involved in the hydrofluorinntiori reaction. Although thc rate of the reaction increases with temperature, should siriteririg occur, the rate of hydrofluorination may nppenr to decrease with temperature. The initial hydrofliiorinatiori rciictiori is very rapid even at 40Oo1and local temperatures
may be attained coiisidcr:hly in excess of the riorninitl temperature because
of the heat of reaction. The more reactive the original starting material,
the greater is the likelihood of incomplete reaction because of sintering.
Another complicating factor is the reversibility of the hydrofluorination
reaction; the retictmts and products will come to an equilibrium determined
by the flow characteristics of the reaction system. The value of the equilibrium coristtrit of the hydrofluorination reaction is in doubt, and values
( ~ H~ ,I oI Fare equilibrium pressures in
for K,, where K , = ~ H ~ o ~ / ~ H F ~ and
atmospheres) ranging from 30 to 400 have been reported (79). A rapid flow
of hydrogen fluoride is thus desirable in order to displace the equilibrium
in the desired direction. The problem of siriteririg is merely troublesome on
HALIDES O F THE ACTINIDE ELISMENTS
207
the laboratory scale, but it can make the difference between success and
failure on the large scale. Consequently, fluidized bed processes have been
developed that considerably minimize these difficulties (49).
Hydrogen fluoride is not the only reagent that can be used for the hightemperature dry-way preparation of uranium tetrafluoride. Cacciari et al.
(14) have described the preparation of uranium and t,horium tetrafluorides
by the reaction a t 400" of uranium trioxide or thorium dioxide with fluorinated hydrocarbons:
+ 2CFzCIz
Tho2 + 2CFZC12
UOI
+ COz + COClz + C12,

+ ThFr + 2COC12.
--t
UF,
The reagent CFzCIz (Freon-12) was first introduced for this purpose by
Booth et al. (11) and has tbe considerable advantage that, unlike hydrogen
fluoride, it can be used in quartz apparatus.
Because hydrates of uranium tetrafluoride, as well as the hydrates of
lower fluorides of the actinide elements, generally undergo extensive
hydrolysis when attempts are made to dehydrate them, preparative methods involving aqueous solutions have fallen into disuse. If fluoride ion is
added t o an aqueous solution of U4+,hydrated UF4 ' 2 . 5 H20 is obtained.
This hydrate can readily be dehydrated to the monohydrate UF4 HzO,
but the removal of the last molecule of water is a matter of some difficulty.
Nevertheless, aqueous methods appear to bc having a new vogue, presumably because of the advantages inherent in carrying out the preparation in
aqueous solution at moderate temperatures. Allen et al. (3) have described
the preparation of uranium tetrafluoride from aqueous solutions of uranyl
ion, UOz++. Tin(II)chloride, copper(I)chloride, chromium(II)chloride, and
titanium(II1)chloride are all effective agents for the reduction of UOz++
to U4+.However, a catalytic reduction process in which sulphur dioxide is
used as the reducing agent was found to be especially useful. While SO,
does not ordinarily reduce UOz++, reduction is rapid in the presence of
catalytic quantities of copper ion. Sulphur dioxide presumably reduces Cuff
to CU+, and this species is known t o reduce UOz++. Hydrofluoric acid is
added t o the uranyl solution, and the reduction and precipitation are carried
out with SO, in the presence of a small amount of copper ion. The reaction
is carried out a t 85-95' and the product obtained is UF4 '0.75 HZO. This
can he dehydrated to anhydrous UF, by heating in a hydrogen fluoride
gas stream. An interesting variant is the use of electrolytic reduction for
the wet-way preparations of uranium tetrafluoride (7.2). The dehydration
of thorium tetrafluoride has also received study. D'Eye and Booth (Z0,21)
have studied the precipitation of thorium tetrafluoride from aqueous solution. Three hydrates have been described: ThF, . 2.5-3.0 H,O (ThF, . zH20,
pseudo-tetragonal), ThF, .0.5-2.0 H20, (ThF, yH,O, orthorhombic), and
208
JOSEPH J. KATZ AND llLVING SHEFT
a hydrate of probable composition ThF4 .0.25 HzO (ThF4 . zHzO). The

first two hydrates were prepared by adding a slight excess of hydrofluoric
acid to a Th4+ solution, and the phase ThF4. zHzO was prepared by the
thermal degradation of the higher hydrates. ThF4 * zHzO is obtained when
the precipitate is separated from the mother liquor immediately after precipitation, whereas ThF, . yHZO is formed when the hydrated precipitate
is allowed to remain in contact with the mother liquor for 48 hr. Rerrioval
of water at 300" in air or vacuum gave a product of composition ThF4 .0.1fh4+
Seporated from mother liquor

a f t e r standing for 48 hours
4 0 % aqueous H F
Seporoted from mother liquor
immediately a f t e r precipitation
.ThF4
Woter o t
r\
looo for
2000 vocuo
l8hrs
--
ThF4
. i.35
. phost
Woter ot 100' for 18 hrs
H
O
,
phase 3
0.2 HzO (phase 2) different from that of monoclinic ThF, (phase 1). Phase
1 could be obtained from phase 2 by heating in vacuum a t 400'. The complex relations existing in this superficially simple system are shown in Fig. 4.
It is on occasion necessary to prepare uranium tetrafluoride from uranium hexafluoride, particularly when dealing with material highly enriched
in UZas.The conversion of uranium hexafluoride to the tetrafluoride has
has thercf ore attracted considerable study, and the present situation is
summarized by Sniiley and Brater (80). Uranium hexafluoride can be reduced with hydrogen. The thermodynamic equilibrium, even a t room
temperature, is in favor of the formation of uranium tetrafluoride, but the
reaction requires a high energy of activation to inaugurate reduction. The
HALIDES O F THE ACTINIDE E L E M E N T S
209
heat required to initiate the reaction can be supplied either by the reactor
walls or by the highly exothermic reaction of hydrogen with a small amount
of elemental fluorine which is added to the uranium hexafluoride. Trichloroethylene can also be utilized as a reducing agent, but the product uranium
tetrafluoride may be contaminated with organic by-products.
The transuranium element lower fluorides can be prepared by quite
analogous procedures. Westrum and Eyring (88) have described a platinum
hydrofluorination apparatus (Fig. 5 ) that is very useful on the microgram
FIG.5. Platinum hydrofluorination apparatus (88). A , platinum spheres with filter

disks of sintered platinum; B, brass needle valves; C,stainless steel block clamps; D,
platinum reactor; E, platinum vessel; F, t o manifold; G, furnace; H , plastic bubbler
with fluorocarbon oil.
or milligram scale, and has been used for the preparation of americium
trifluoride by the hydrofluorination of Amz03 with anhydrous hydrogen
fluoride a t 650". The trifluorides of the actinide elements (where they are
stable) can be prepared in a variety of ways. Uranium trifluoride is best
obtained by reduction of the tetrafluoride, and aluminum is the most convenient reagent for this purpose(70) :
UFI
+ Al-UFI
900C
+ AlF 7'.
A procedure of this kind is practically obligatory for the preparation of UF3

because the uranium(II1) oxidation state is only barely stable in aqueous
solution; attempts to precipitate solid compounds from a Ua+ solution
invariably yield compounds of uranium(1V). For the transuranium elements, however, either aqueous or high-temperature dry-way procedures
210
riiay be employd. Plut3onium trifluoride is insoluble in water arid may

therefore be prepared by precipitation from ityueous solution by addition
of hydrofluoric acid to a Pu3+ solution. Anhydrous PuF3 can be obtained
from the hydrated trifluoride by heating in a stream of hydrogen fluoride
at 200-300". Addition of F- ion to a solution of plutonium(1V) forms
PuF4 . 2.5 HzO, analogous t o UF4 . 2.5 H2O obtained under comparable
conditions. Attempts to dehydrate PuF4 * 2.5 HzO yield Pup3 and not the
expected PuF4. Ditwson et al. (19) account for this state of affairs by the
reaction scheme :
+ 2H20 a P u O ~+ 4HF
3PuFI + P u O ~ 4PuF3 +
4PuR + 2HzO 4PuFa + 4HF +
I'uE'I
0 2
0 2
This behavior is clearly consistent with the increasing stability of the lower
oxidation states as one proceeds to actinide elements of higher atomic
number.
Both plutonium trifluoride arid tetrtifluoride can be prepared by hydrofluorination of plutonium dioxide. The oxidation-reduction relatioils of
plutonium(II1) and (IV) are such that it is readily pussible to specify which
of the fluorides shall be the product. If the reaction is carried out under
reducing conditions, then the product is PuF3, arid under oxidizing conditions, PuF4:
GOO'C
+ iH2
P u F ~+ 2Hz0,
PUOZ+ 4HF + Oz - P U F +
~ 2Hz0 +
+
PUOZ 3HF
55O'C
02.
Cylinders of hydrogen fluoride stored for any length of time contain hydrogen gas formed by reaction with the iron container. Hydrogen fluoride
drtiwn from such a cylinder will yield only PuF3 because of the reducing
atmosphere provided by the hydrogen. To obtain PuF4, oxygen may be
added to assure the absence of reducing agents; under these conditions
PuF4 is the sole product. Similar considerations apply to NpFs arid NpF4.
Of the trarisplutonium elements, one of the most interesting results
obtained in recent years relates to the preparation of curium tetrafluoride.
The (111) oxidation state is the characteristic one for curium, and curium
trifluoride is readily prepared by precipitation from aqueous solution (27).
Attempts to prepare curium tetrafluoride by fluorination of the trifluoride,
analogous to the successful preparation of americium tetrafluoride from the
trifluoride, were uniformly unsuccessful as long as the short-lived isotope
Cm242 (162.5-day alpha-emitter) was used. Asprey et al. (7 ) were able t o
convert CmFl to CmF4 with elemental fluorine a t 400", provided Cm'44,
211
an alpha-emitter of 18.4-yr half-life was used. Presumably the longer-lived

isotope causes less radiation damage t o the compound. Curium tetrafluoride
is monoclinic, and is isostructural with uranium tetrafluoride and the other
actinide element tetrafluorides.
C. HIGHERFLUOEIDES
The higher fluorides of the actinide elements are certainly among the
most interesting of the actinide halides. The volatility of the MFs compounds makes them of great practical interest, and the unique opportunity
to study a rare-earthlike electron configuration in the gaseous state makes
the triad UF6, NpF6,and PuF6of equally great scientific interest. Considerable information on these compounds has been obtained in recent years,
mainly by Weinstock and co-workers, and this area of research has seen
perhaps the most interesting and significant work in the entire area of
actinide halide research in recent years.
Uranium hexafluoride is prepared industrially on a very large scale for
uranium isotope separations. All compounds of uranium, including of course
the metal, react with fluorine to yield the hexafluoride if the reaction is
allowed t o proceed under sufficiently vigorous conditions. In general, UF,
is prepared by the fluorination of uranium tetrafluoride, this being the
procedure that requires the least amount of the expensive reagent, fluorine:
+ 3F2 + UFs + 3 0 2 ,
3UFe + 402,
UOz + 3F2 UFs +
UF4 + Fz UFs.
UOa
U&
+ 9Fz
02,
The fluorine content of UF, is derived from the relatively inexpensive reagent, hydrogen fluoride, so that only two of the six fluorine atoms of UFBare
obtained from elemental fluorine. Brater and Smiley (1.2) have given an
excellent description of the laboratory scale preparation of UFB by the
UFd-F, reaction.
Labaton and Johnson (53) have made a kinetic study of the fluorination of uranium tetrafluoride by fluorine. The reaction was studied in the
temperature range 265-348" by following the change in weight-of the solid
phase with a copper-beryllium spring balance. No detectable reaction is
noted a t 220", but the reaction rate is measurable at 230". The rate a t which
UF6 is produced is in agreement with the kinetics expected for the reaction
between a gas a t a spherical interface that is continuously being reduced
due to reaction with the gas. The rate law can be expressed by the equation
of Anderson (5) for the case of a gas-solid reaction governed by a diminish-
212
JOSKPH J. KATZ AND IRVING SHEFT
ing interface. From the temperature variation of the rate, an activation

energy of 19.1 kcal/mole, for UFd prepared by hydrofluorination, is deduced. The mechanism of the reaction can be visualized in the following
terms : Formation of uranium hexafluoride proceeds by adsorption of fluorine on the surface of the uranium tetrafluoride particles. The fluorine
migrates into the crystal lattice with the formation of a nonstoichiometric
intermediate compound of variable composition richest in fluorine near the
surface. Uranium hexafluoride is continuously evolved from the surface
as long as the partial pressure of uranium hexafluoride in the system is less
than the dissociation pressure of the intermediate to form UFe. The particles of uranium tetrafluoride diminish in size as the reaction proceeds, The
migration of fluoride ions into the uranium tetrafluoride lattice is conccived
as an essential feature of the mechanism. Apart from the possible impropriety of designating the intermediate fluorides participating in the reactor
as nonstoichiomctric compounds, the suggested mechanism appears to be
a reasonable one.
Elemental fluorine is not the only reagent that can be used for the
fluorination of uranium tetrafluoride. The halogen fluorides were first used
for the preparation of actinide fluorides by Emelbus et al. (26, 26). Particularly useful are chlorine trifluoride and bromine trifluoride. The former is
usually used as a gas because of its low boiling point of 12, whereas bromine trifluoride boiling at 126 can be used as a liquid phase reagent. Both
of these reagents have been employed in the processing of uranium fuels
from nuclear reactors; the irradiated uranium, fission products, and plutonium arc converted to fluorides, and are separated by taking advantage
of differences in volatility. The uranium is recovered as uranium hexafluoride and is obtained in a very high degree of both chemical and radiochemical purity by fractional distillation (9,47).Labaton (54) has reported
u kinetic study of the fluorination of uranium tetrafluoride by gaseous
chlorine trifluoride. The mechanism of the reaction, according to Labaton,
is essentially the same for chlorine trifluoride as for fluorine. The reaction
rate passes through a maximum at 105, falls to a minimum at 148 and
then rises again. The lower reaction rate constants in the temperature interval are attributed to the lower rate a t which chlorine trifluoride reacts with
the surface intermediate fluoride as compared to uranium tetrafluoride;
above 150, intermediate fluorides do not form and the rate of hexafluoride
formation rises again. The activation energy is much lower and the reaction
proceeds a t consideraly lower temperatures than is the case when elemental
fluorine is used. The halogen fluorides thus appear to have definite advantages in the preparation of uranium hexafluoride. The phase diagrams for
the uranium hexafluoride-chloride trifluoride and uranium hexafluoridebromine pentafluoride have been studied (23). The uranium hexafluoride-
213
chloriiie trifiuoride system shows complete miscibility in the liquid phase

a t 67" but partial miscibility at 75" over a limited composition range; no
azeotropes exist in this system. The uranium hexafluoride-bromine pentafluoride system exhibits complete miscibility over the range 68-80', and
has a single maximum vapor pressure nzeotrope at a composition of approximately 0.1 mole fraction of uranium hexafluoride. The uranium hexafluoride-bromine trifluoride system has:also
been investigated (SO) ; the
system is of the simple eutectic type and exhibits positive deviations from
ideality.
Whether elemental fluorine or one of the halogen fluorides is employed,
preparation of uranium hexafluoride by fluorination requires the preparation of elemental fluorine, because the preparation of the halogen fluorides
require elemental fluorine. Fried and Davidson (33, 37') discovered, however, that it is possible to obtain uranium hexafluoride by the reaction of
uranium tetrafluoride with dry oxygen a t 800' according to the reaction
2UF4
+ Oz
-+
UFs
+ UOsFz.
The reaction can be made the basis for a cyclic process for the preparation
of uranium hexafluoride without the use of elemental fluorine:
+
+
+
2UF4 0 2 ---t UFs UOzFz

UOZFZ H2 3 UOZ 2HF
UOz 4HF 2H2O UF4
+
+
--j
The fluorine content thus is obtained entirely from low-cost hydrogen

fluoride. This reaction, because of its potential large-scale utility, has
therefore been the subject of a number of investigations. Ferris (28) has
examined the side reactions and thermodynamics of the oxidation in the
temperature range 600-900". In addition to the expected products UFB and
UOzF2,Ferris observed the formation of uranium pentafluoride. Its formation was postulated to be the result of a side reaction between UFG and
UF,. Detailed experiments indicated also that the reaction
3UF4
0 2
-+
21JF6
+ UO2F2
makes a significant contribution to the observed stoichiometry at elevated

temperatures. In the temperature range studied, only slight decomposition
of uranyl fluoride was noted, in accordance with thermodynamic predictions. In a subsequent publication, Ferris (29) has examined the kinetics
of the uranium tetrafluoride-oxygen reaction by a thermogravimetric
method in the temperature range 550-815'. The reaction shows a firstorder dependence on the surface area of the uranium tetrafluoride; the rate
is also dependent on the oxygen partial pressure. The rate for the reaction
214
JOSEPH J. KATZ AND IRVING SHEFT
with dry oxygeu caii be oxpresbecl :is k = 8.31 X lo6 exp (-45,YOO/RT).
where k has the units of moles per mioiite per syuttre meter.
As usually prepared, uranium hexafluoride may be contaminated with
other volatile fluorides, and hydrogen fluoride, molybdenum hexafluoride,
vanadium oxyfluoride are among the common and objectionable impurities
that may be present. Purification is usually achieved by distillation procedures and a considerable amount of useful and practical information on
purification proredures has recently become available (24, 62, 81). A procedure for the purification of UF, that has been recently introduced by
Cathers et al. (15) involves the formation of a complex compound betweeii
uranium hexnfluoride and an alkali metal fluoride. Ruff and Heinzelmann
(68) first observed complex formation between uranium hexafluoride and
sodium or potassium fluoride, and this was confirmed by Martin el al. (69).
These reports were received with a certain reserve, because efforts made during World War 11 to prepare addition compounds of urnnium hexafluoride
were singularly unsuccessful (41). Nevertheless, complex compounds are
indeed formed, according to the reaction
UF6
+ 3NaP 4 Na3LTF,.
The renction is reversible, atid uranium hexafluoride is evolved on heating;

it is this factor that makes the addition compound useful for the recovery
or purificatiori of uranium hexafluoride. The dissociation pressure of the
complex compound in the temperature range 80-320" is given by the
equation log p = 10.88 - (5.09 X l03/T), where p is the partial pressure
of UF, in millimeters of mercury, nnd T is the nbsolute temperature. A
decomposition pressure of 1 atm is calculated to occur at 363", and an enthdpy of formation of -23.2 kcallmole of UFais deduced for the formation
of the complex. The kinetics of the reaction between uranium hexafluoride
and sodium fluoride have been studied by Massoth and Hensel (60, 61).
The reaction appears to proceed initially at a rate determined by the surface
area of the sodium fluoride, and then to become diffusion limited. The rnte
of the reaction is given by the expression k = 5.0 X lo8exp (- 13,00O/RT)
per hour; the activation energy is thus 13.1 f 0.2 kctd/mole of uranium
hextrtluoride.
A study of the uranium hexafiuoride-metal fluoride complexes was
undertaken by Sheft et al. (7'7) by the use of radioactive fluorine-18 exchange methods. The exchangeability of fluorine betWC?eKl uranium hexafluoride and metal fluorides caii be used as a measure of the extent to which
reaction occurs. If a complex compound actually forms between uranium
hexafluoride and a metal fluoride, then the fluorine atoms become equivnlent and randomization of the radioactivity occurs. Complete exchange
of the fluorine atoms is found to occur between sodium fluoride labeled
215
HALIDES OF TIIE ACTINIDE ELEMENTS
with F1*and uranium hexafluoride. The data are consistent with the
formation of UF6 2NaF rather than UF6 * 3NaF; the exchange results
may be interpreted on this basis or that the complex UF6 3NaF does
form, but that only three of the six fluorine atoms originally in the UFSare
capable of exchange with sodium fluoride. The former alternative seems
the more likely. At a temperature of 350, where the uranium hexafluoridesodium fluoride addition compound is essentially completely dissociated,
no exchange of fluorine is observed to occur. The evidence for the formation
of a true complex between sodium fluoride and uranium hexafluoride thus
appears quite good. No exchange was observed to occur between uranium
hexafluoride and lithium fluoride, potassium fluoride, silver fluoride, and
zirconium tetrafluoride labeled with F1*. The ability to form a n additive
compound with uranium hexafluoride thus appears to be limited to sodium
fluoride of the metal fluorides tested. Since the X-ray diffraction pattern
of the sodium fluoride complex has not as yet been interpreted, a geometrical reason for the observed specificity, although probable, remains to
be established.
Neptunium hexafluoride was first prepared by A. E. Florin and has
recently been studied in detail by Weinstock and his collaborators. Neptunium hexafluoride can be prepared by the fluorination of NpF3 or NpF4 a t
500". Since neptunium hexafluoride requires a rather higher temperature
of formation and is less stable than uranium hexafluoride, a n ingenious
apparatus tthat enables the reaction to proceed a t elevated temperatures
and still allows the product to be rapidly cooled is employed (Fig. 6) (57).
Liquid fluorine drips from the liquid-nitrogen cooled condenser onto the
heated starting material, NpF3 or NpF4. A high local concentration of
fluorine is thus maintained in the reaction zone; convection currents carry
the product NpFB to the condenser where it i s trapped. Neptunium hexafluoride is bright orange in color, unlike colorless uranium hexafluoride, and
appears t o be susceptible to photodecomposition.
Although the existence of a volatile higher fluoride of plutonium had
been surmised from tracer experiments, positive evidence of the existence
of PuF6 was obtained hy Florin (31) who first prepared the compound. An
investigation of the preparation and properties of PuFBwas also conducted
by Mandleberg el al. (58): but in recent years the compound has been
most intensively studied by Weinstock and his collaborators. Plutonium
hexafluoride can be prepared by a variety of procedures similar to those
used for the preparation of uranium hexafluoride. The most widely used
method consists in the fluorination of plutonium tetrafluoride with elemental fluorine. Whereas uranium hexafluoride can be prepared by the
analogous reaction a t 300", and neptunium hexafluoride a t 500", the
preparation of plutonium hexafluoride by this reaction appears t o require a
216
JOSEPH J . KATE AND IIIVINU SHEYT
temperature of 750' in order to achieve useful rates of reaction. Since

plutonium hexufluoride is alleged to be thermally unstable, it is essential
that the freshly formed plutonium hexafluoride be rapidly chilled to a
temperature where the rate of decomposition is slow. Weinstock and Mulm
(85) have described a reactor for simultaneously carrying out the preparaTo Vocuum
i'
I
To Vocuum
S y s t e m ond
Fluorine Tonk
Evocuoted Oswor
Groove;
for Closure
FIG.6. Reactor for neptunium hexafluoride (57).
tion at a high temperature and at the same time providing a low-temperature condenser (Fig. 7). The starting charge of plutonium tetrafluoride is
placed in a nickel crucible which is inserted into the coil of an induction
heater. Fluorine is then introduced into the reaction vessel. The coil, which
is of copper tubing, is cooled by the circulation of liquid nitrogen and thus
serves the dual purpose of heating the plutonium tetrafluoride and simul-
HALIDES O F THE ACTINIDE ELEMENTS
217
taneously providing a condenser. Gram amounts of plutonium hexafluoride

have been prepared in this reactor. The fluorination of plutonium tetrafluoride by elemental fluorine has been studied in a flow system in the
temperature range between 100" and 600". Activation energies between
10 and 12 kcallmole were obtained. In contrast to earlier workers, Steindler
et al. (86)and Adams et al. (I) found that plutonium tetrafluoride can be
P P
TO FLUORINE TANK,
STORAGE VESSELS
CIRCULATION
TO INDUCTION
HEATER
FIG.7. Reactor for preparation of plutonium hexafluoride (86).A , nickel dish filled
with Puff; B, brass reactor can; C, removable cover for loading reactor; D, tongue and
groove; Teflon gasket for closure; E, nickel supporting rod for nickel dish; F , coil of
3/16-in. 0.d. copper tubing; G, Fluorothene window; H , Teflon seal and insulator; I ,
Micalex insulator.
fluorinated to the hexafluoride at modest temperatures of the order of 200".

These observations are confirmed by Robb et al. (66) who found that
decomposition products of plutonium hexafluoride can be refluorinated
at 200". The reason for the discrepancy in the effect of temperature on the
reaction rate of the fluorination of plutonium tetrafluoride is not clear,
but it may be related to differences in the surface area and characteristics
of the plutonium tetrafluoride, and on the complex dependence of the
reaction on the fluorine pressure. In any event it appears quite certain that
218
JOSEPH J. KATZ A N D IRVING SHEFT
under certain conditions conversion of plutonium tetrafluoride to the hexitfluoride can be carried out at a fair rate even a t temperatures as low as
170'. These observations have obvious significance for volatility procedures
for the recovery of plutonium from neutron-irradiated uranium.
Although thc fluorination of PuF4is the most useful preparative method,
plutonium hexafluoride can be prepared by a variety of alternative reactions (58):
+ l2TTF + -+ 2PuF8 + 6H20,

2PuF4 + 0
2 + PuFa + PUOZFZ.
2P1102
0 2
Although the oxidation of PuF4 by oxygen very likely proceeds by a

mechanism similur to that which obtains in the UF4-02 reaction, dctuils
are a t the present lacking on the PuF4-02reaction. The behavior of plutonium trifluoride with oxygen deserves t o be noted. When plutonium
trifluoride is heated in oxygen a t 600", oxidation to a mixture of plutonium
tetrafluoride and plutminm dioxide occurs (35, 19): 4PuF;, 0 2
3YuF4 PuOa. Thc reaction is reversible and oxygen has been observed
as a reaction product of plutonium dioxide and plutonium tetrafluoride.
Plutonium hexafluoride is a reddish-brown solid, and like neptunium
hexafluoride, is light sensitive. Again, as in the case of iiranium and neptunium hexafluoridcs, plutonium hexafluoride may be handled and stored in
yuartz or Pyrex equipment with no sign of decomposition provided the
glass has been thoroughly outgassed and the hexafluoride is completely
free of h y d r o p i fluoride. Any of the actinide hexafluorides may be heated
to 200" in quartz apparatus with little if any nttack on the container.
The alpha activity of the actinide elements int,roduces a complication
in assessing the stability of their compounds. Not only does the alpha activity present tixi extraordinary health hazard in the case of plutonium hexaffuoride, and to n. lesser extent nepturiium hexafluoride, but the continuous
passage of alpha particles, t,hrough the compounds results in radiation
decomposition. Shiflett et a1 (78) have studied the alpha-particle decomposition of uranium hexaflnoride. Compared to plutonium hexafluoride
the effects of the self-pruduced alpha particles are very small. Uiiless
uranium is enriched in 1JZ34 or U236,radiation effects are very difficultj to
or U236emits about 50
detect, but uranium enriched a hundredfold in UZa4
tirncs the number of alpha par ticles as does normal uranium, and here the
effects m e yuite appreciable. In a hundredfold-enriched uranium, about
0.2y0of the molecules present would become ionized in the course of a year
by self-alpha irradiation and the extent of decomposition might become
significant. For normal uranium, the effects of alpha particles 011 uranium
hexafluoride were examined in the gas phase by irradiation with alpha particles from MnZz2.The decomposition products of the uranium hexafluoride
are fluorine and intermediate uranium fluorides; the products react with
219
each other to reform uranium hexafluoride by a radiation-induced process.

The rate of alpha decomposition is essentially independent of the te,mperature. For GI the number of molecules of uranium hexafluoride decomposed per 100 ev of energy absorbed, a value of 0.90 was found; for M l n ,
the number of molecules of uranium hexafluoride decomposed per ion pair,
a value of 0.27 was calculated. The latter number is based on the assumption of a loss of 30 ev by an alpha particle when it produces one ion pair.
Compounds of plutonium provide a much more striking example of the
effects of alpha-particle irradiation than in the case of uranium compounds.
This circumstance arises from the very much shorter half-life for alpha
let alone P8.
The rate
decay of Pu239in comparison to that of U234or U2361
of radiation decomposition in solid plutonium hexafluoride is about 1.5%
per day. From a decomposition rate of 1.5% per day and the assumption
that all the energy of the alpha particles is absorbed in the solid, the average
energy absorbed in the decomposition of one plutonium hexafluoride molecule is 30 ev. This value is very similar to the average energy loss by alpha
particles per ion-pair formation for a wide variety of substances. The
decomposition product of plutonium hexafluoride is a green substance that
turns white upon exposure to air; both of these products are compounds
of plutonium(IV), and no evidence for the formation of plutonium(II1) or
(V) compounds has been found. The solid radiation decomposition product
apparently fails to reach a steady state equilibrium. In order to minimize
the loss of plutonium hexafluoride on storage because of radiation damage,
it is desirable to store the compound as the vapor. Under these conditions,
a considerable fraction of the alpha particles escapes capture and is absorbed
in the walls of the container. The pressure of the gas and the geometry of
the container thus play important parts in the conservation of the hexafluoride. The lower the pressure of the gas and the smaller the diameter
of the containing vessel, the greater will be the probability that a given
alpha particle will collide with the wall rather than a plutonium hexafluoride molecule. By appropriate storage conditions the average decomposition rate can be reduced to a value as low as 0.06% per day. It would
appear, however, that all of the circumstances relating to the radiation
decomposition are not understood, since stabilities considerably greater
than would be predicted are on occasion encountered. Clearly, further
information regarding the achievement of a steady state, and the conditions
under which a back reaction might occur are to be desired.
D. CHLORIDES
The actinide chlorides can be prepared by a variety of more or less
standard chemical procedures. As these halides are considerably less volatile than the hexafluorides, the experimental complications and health
220
hazards are considerably reduced. All of the actinide chlorides citn he made
by direct combination of the elements, but this procedure is seldom resorted
to, partly because d the great difficulties encountered in obtaining many
of the actinide element metals, and partly because of the difficulty in controlling the reaction. The reaction of halogen with actinide carbides,
nitrides, and hydrides very often offers the advantage that the reaction
can be more readily controlled. However, such procedures do not obviate
the necessity of producing metal, since carbides, nitrides, and hydrides
can normally be obtained in a reasonable way only from the metal. If the
element is readily available in metallic form, as is the case for thorium,
uranium, and plutonium, then the reaction of the actinide hydride with
hydrogen chloride is a very convenient procedure :
250C
u + %ItUHI
+ 3HC1-
UBI
2M)"30O0C
UClr
+ 3H2.
Far more important are procedures involving the chlorination of the

oxides. Since the actinide element oxides are very stable in a thermochemical sense, powerful chlorinating agents are required. Thus neither chlorine
nor hydrogen chloride will react to any extent with actinide element oxides,
and reagents such as carbon tetrachloride, thionyl chloride, phosphorus
pentachloride, or phosgene are required. Of these, carbon tetrachloride
has perhaps been most widely used. This reagent can be used either in a
vapor phase reaction, or in the liquid phase under elevated pressure, and,
depending on the exact conditions, UCL, UCls, or UCle can be obtained
from uranium dioxide as the starting material. The chlorination of urnnium
dioxide and plutonium dioxide by carbon tetrachloride has been the subject
of a very thorough investigation by Budaev and Vol'skii (13). The complexity of the vapor phase reaction can be judged by the fact that Budaev
and Vol'skii find it necessary to postulate some 20 reactions that are supposed to occur during the chlorination of uranium dioxide by carbon tetrachloride. The exhaust gases are found to contain carbon dioxide, phosgene,
carbon monoxide, and chlorine. Under conditions of incomplete reaction
at 400" to 500, two intermediate oxychlorides, UOCla and UOCl, are reported to occur; these may be intermediates in the reaction. Uranium(V)
oxychloride, UOC13, consists of reddish-brown crystals that are readily
soluble in water, acetone, or methanol, but are insoluble in benzene or carbon tetrachloride. A study was also made of the composition of the solid
products and gases formed in the chlorination of plutonium dioxide by
carbon tetrachloride at 700'. The principal reactions are :
+ CCI,
PuO2 + ZCCI,
PUOl
+ c02 + +a,,
PUClS + 2co + $Cle;
+ PUClS
--t
221
and to a small extent

PUOl
+ 2CCL
-+
PUCl3
+ 2coc12 + +ClZ.
Another procedure extensively used for the preparation of plutonium trichloride is the reaction of plutonium oxalate with hydrogen chloride.
R.elated to carbon tetrachloride is the reagent hexachloropropene,
C13CCC1= CClz; bp, 210". Uranium trioxide reacts with hexachloropropene at reflux temperature according to the equation (45)
UOa
+ 3ClaCCC1 = cclz
UC14
+ 3ClzC = c c l c o c l + clz.
This is perhaps the most convenient method for the preparation of uranium
tetrachloride. If the reaction is carried out a t lower temperatures, UCI6and
UCls are the primary products; on heating they lose chlorine to form uranium tetrachloride. Panzer and Suttle (64) have reported the isolation of
a dark red condensation compound with trichloroacrylyl chloride of the
composition 5UCla CClz = CClCOCl that is readily decomposed to uranium tetrachloride. The formation of such a coordination compound by
uranium pentachloride is reasonable, as this uranium compound always
occurs in a coordinated structure. Pure uranium pentachloride appears
to exist as the dimer (UCI&, possibly as (UC14+) (UCls), and forms coordination compounds with phosphorus pentachloride and thionyl chloride.
Hexachloropropene reacts only with uranium oxides in a higher oxidation
state, that is, UOs or U308. Uranium dioxide and other dioxides do not
react. When Np30s is treated with hexachloropropene, no reaction is
observed (39). Plutonium dioxide does not react with hexachloropropene,
but plutonium trichloride can be obtained in good yield and purity when
plutonium(II1) oxalate is heated with hexachloropropene at 180-190' for
18 hr (16).
Freeman and Smith (3%')have prepared the anhydrous chlorides of a
number of lanthanides and of thorium by dehydrating the hydrated
chlorides with thionyl chloride. Although efforts to obtain anhydrous
plutonium trichloride in this way were unsuccessful, it is believed that this
may be a useful procedure for actinide elements such as actinium, americium, and curium that have a particularly stable (111) oxidation stage. In
general, aqueous methods for preparing tetrachlorides are of little value; but
anhydrous trichlorides, particularly of the transuranium elements, can be
obtained readily from the hydrated trichlorides by dehydration in an atmosphere of hydrogen chloride.
E. BROMIDES
AND IODIDES
Little has appeared since the preparation of such compounds was last
reviewed (61). Iodides and bromides are best prepared from the elements.
TABLE I V
CRYSTAL STRUCTURE DATAFOR THE
A%CTINIDE HALIDES"
Crystal structure
Color
Compound
SPme group
or structure type
hF3
w-hite
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Hexagonal
Or thorhombic
Ort horhombic
Orthorhombic
Orthorhombic
Orthorhombic
Orthorhombic
Orthorhombic
Monoclinic
Reddish brown
MonocIinic
UF,
White
Black
Purple or black
Purple
Pink
White
Red
White
Emerald green
pink
White
Red
Green
Green
Green
m'hite
Black
Brown
Bright green
Yellow
Lattice parameter
smmetry
Cs/mmc
LaF,
LaFr
LaF3
LaF3
UCls
C6dm
C63/m
C6Jm
G6dm
Ucls
C6dm
CWm
Ccmm
Ccmm
Ccmm
Cnm
Ccmm
Ccmm
Ccmm
c62h
bo
ao
(b)b
cn
Density
(sm/cm3)
4.17
4.146
4.108
4.087
4.067
7.53
7.348
7.273
7.240
7.225
7.88
8.95
9.12
9.32
9.53
7.62
7.442
7.405
7.380
7.38
8.06
7.942
7.917
12.65
12.62
12.6
14.01
14.03
14.03
14.0
13.1
4.55
4.320
4.273
4.238
4.25
4.68
4.440
4.382
9.15
9.13
9.12
10.01
9.95
9.92
9.9
8.6
4.81
5.51
5.58
5.70
5.78
5.85
6.53
6.62
6.62
6.69
4.11
4.09
4.11
4.32
4.30
4.30
4.31
11.01
a2 = 126 xk 1'
6.76
6.82
6.92
5.71
Green
Monoclinic
c2/c
12.82
10.74
126"lO'
10.64
a2 = 126"lO'
10.57
az = 126"lO'
10.47
a2 = 126'10'
10.45
a2 = 126"
30'
8.41
6.70
8.41
6.8
8.28
7.0
az =
Green
Monoclinic
ZrF4
12.70
Brown
Monoclinic
c2/c
12.62
Tan
Monoclinic
UF4
12.49
Greenish tan
Monoclinic
UF4
12.45
White
Greenish yellow
Green
Red brown
White
Brown
Tetragonal
Tetragonal
Tetragonal
Tetragonal
Tetragonal
Monoclinic
I4/amd
UClr
I4/amd
I4/amd
8.473
8.377
8.296
8.29
8.963
10.92
2/c-/-
8.69
8.20
8.16
7.468
7.482
7.487
7.46
7.946
7.05
4.60
4.87
4.92
5.69
5.55
a* = 93O9'
Reddish brown
White
Black
Black
Black
-~
White
White
Red brown
White
Brown
Reddish brown
Black
Monoclinic
UBr4
Cubic
143m
Distorted
UF4
14/m
I32d
Tetragonal
Tetragonal
Monoclinic
Orthorhombic
Orthorhornbic
Orthorhombic
Hexagonal
Pnma
Pnma
C3m
7.06
8.471
9.900
9.91
10.97
5.207
5.21
6.94 (exp)
5.81
6.45
3.81 (exp)
5.060
5.00
6.04
3.56
4.472
5.209
6.525
11.473
8.966
8.97
For detailed references to the original literature see Kata and Seaborg ( 5 1 ) ,Seaborg and Katz (?'S), and Kate and Rabinowitch (50).
b Lattice parameters, when published as k X units, have been converted to Angstrom units.
c Douglass and Staritzky (22).
224
Aluminum bromide and iodide, however, have special utility in the eonversion of oxides to the corresponding halides. Prigent (6'5) has reported
that UOa heated with carbon tetrabromide a t 165" forms UBr4. Uranium
tetrabromide has also been prepared by reaction of uranium dioxide with
a tenfold excess of carbon tetrabromide a t 175" and subliming away the
UBr4formed (22).
IV. Properties of Halides
By virtue of the fact that the actinide elements constitute a family of

elements whose electronic configurations are related to each other in a
quite unique way, a comparison of the physical and chemical properties of
a series of related compounds has more than the usual interest. In this section some of the more important properties of the actinide element halides
are summarized in tabular form, and a number of topics of current interest
are described.
A. STRUCTURE
Crystal structure data obtained by X-ray diffraction methods for the
actinide element halides are collected in Table IV. Crystal structure determinations have been most important in identifying new compounds of the
actinide elements; the data are sufficiently extensive now for use in drawing
conclusions regarding systematic trends and relations among the actinide
elements. The tetrafluorides, for instance, supply one of the best illustrations of an actinide contraction that is entirely similar to the well-known
lanthanide contraction (Table V).
TABLE V
IONICRADIIOF M4+ ACTINIDEIONS'
~
Com pound
Ion
4+
Np4+
Pu4+
Am'+
Cm4+
Ionic radius
(A)
0.93
.92
,90
.89
.88
~~
Asprey et al. (6).
The vibrational spectra of all of the known actinide element hexafluorides can be interpreted on the basis of the structure of a regular octahedron. Malm et al. (66) have measured the infrared spectra of NpFa and
PuFBand have deduced the fundamental vibration frequencies (Table VI).
225
HALIDES OF THlC ACTINIDE ELEMENTS
Only v3 is directly observed in the infrared. All of the frequencies for

NpF6 lie intermediate between UF6 and PuF6. The infrared data make it
possible to deduce the value of the Raman active frequencies; in the case
of uranium hexafluoride, a discrepancy was revealed between the value for
the Raman spectrum as deduced by the method of Gaunt (38)from infrared measurements and the experimentally determined Raman spectrum
of BigeIeisen et al. (10). Consequently, Claassen et al. (17) have measured
TABLE VI
FUNDAMENTAL
VIBRATION
FREQUENCIES
OF UFa, NpFs,
Designation
Symmetry
species
VI
alg
YL
EL7
v3
flu
v4
f , II
v5
fi"
V6
f2u
AND
PuF,"
Frequencies (cm-1)
activity
Raman, p
Raman, d p
Infrared
Infrared
Raman, d p
Inactive
UFG
667
535
623
181
202
140
NPFG
648
528
624
200
206
164
PUFG
628
523
615
203
211
171
Malm et al. (66).
the Raman spectrum of gaseous uranium hexafluoride, and found v1 to bc

666.6 f 0.3 cm-l and v2, 535 f 5 em-]. These new Raman values are in
excellent agreement with values deduced by combination from the infrared
data, and are the values that appear in Table VI. The infrared and Raman
data then lead to the octahedral structure of symmetry Oh for the actinide
element hexafluorides.
Schomaker et al. (71) have made an electrondiffraction study of the
structure of NpF6and PuF6. An F-F bond angle of exactly 90" is obtained,
in agreement with the structure of a regular octahedron, and the derived
metal-to-fluorine distances are given in Table VII.
TABLE VII
METAL-FLUORINE
BONDLENGTHS
OF ACTINIDEELEMENT
HEXAFLUORIDES
FROM ELECTRON
DIFFRACTION
STUDIES"
M-F
U--F
Np--F
Pu-I?
Distance
(A)
1.994
1.981
1.969
Schomaker et al. (71).
Here also the metal-fluorine distance decreases with increasing atomic

number in a manner analogous to the lanthariide contraction in crystals.
226
JOSEPH J. KATZ A N D 1 H V l N U SIiEFT
13. VOLATILITY
The actinide element halides run the entire gamut of volatility from
very high melting and refractory substances such as the trifluorides to the
exceedingly volatile hexafluorides. Table VIII lists the actinide element
TABLE VIII
VAPOR IREBSUREDATAFOR ACTINIDEELEMENT
HALIDES~
+B
log pmm= - A / T
Compound mP
Solid
IhCL
IJClr
ThBn
1JHrd
ThI,
770
590
679
519
21120
24828
12000
15910
15000
152x1
15173
16860
14370
10886
12910
10427
9628
10900
1J2Eod
Dec.
DCC.
Dec.
Drr.
7315
7143
2942
4166
IIIFI
AmFn
IJCl3
PuCln
1JBr3
Pu13rr
PI113
ThFrb
111Yd
AmF,c
U,Fi7d
n-1Jbbd
&UFbd
1425
835
760
730
681
1110
1037
Liquid
B
12.468
36.888 - 7.048 log T
10.0
12.726
12.5
13.386
29.18 - 5.035 log 7
11.986
36.07 - 7.554 log 1
7.727
14.30
13.2995
11.73
14.56
19400
11.273
12590
9.509
12360
13008
1.5270
18124
10.321
30.16 - 6.042 log 2
10. tcz1
37.97 - 7.554 log 2
7987
7205
7440
7060
31500
9.57
9.65
9.56
9.71
9.09
13.68
12.75
7.634
10.71
For dctailrd references to the original literature Are Kahz and YeaLorg (61), Seaborg
and Katz (7S),and Katz and Rabinowitch (50).
* Darnell and Keneshea (18).
Yakovlev and Kasynkov (89).
d Equilibrium vapor pressure of UFs for thr disproportionation reaction.
a Dccomposcs.
halides for which quantitative vapor pressure data exist ; qualitative observations arc collected in Table IX, and data relating to the volatile hexufluorides are given in Table X.
Weinstock et al. (87) have made an exceedingly thorough study of the
vapor pressure of neptunium and plutonium hexafluorides (Table X). The
vapor pressures of NpF6and PuFaare compared to the UF6 vapor pressures
227
HALIDES OF T H E ACTINIDE ELEMENTS
dclermined hy Oliver et al. (63) in Fig. 8. The relative volatilities of the

liquids are in the expected order uF6 > NpF6 > P U F 6 , but the relative
volatilities of the solids exhibit a more complicated behavior with NpFB
having the greatest volatility over the measured solid range. For all molecuTABLE IX
QUALITATIVE
OBSERVATIONS
O N VOLATILITY
OF ACTINIDE ELEMENT
HALIDES
Compound
Sublimation
temperature ("C)
Compound
Sublimation
temperature ("C)
UF3
AcClr
NPCh
AmCI3
AcBrr
NpBrr
ArnBr~
<1000
NpI3
Am&
UFI
PaC14
NpCla
NpBr4
800
900
1000
400
500
500
950
750-800
850
800
800
850-900
lar parameters such as the vibration frequencies, interatomic distances,

and the size of the unit cell in the solids, neptunium hexafluoride is intermediate between the other two actinide hexafluorides. It is, therefore, surprising t o encounter such an anomaly in the vapor pressures of the solids.
Weinstock et al. (87) have considered a number of possible explanations to
TABLE X
VAPORPRESSURE
DATAFOR ACTINIDEELEMENT
HEXAFLUORIDEP
Compound Temperature
range ("C)
0-64
64-1 16
~ ~ ~ ( 1 1 116-230
NpFdc)
0-55.10
NpFs(1)
55.10-76.82
PuFa(c)
0-51.59
PUFs(1)
51.95-77.17
UFa(C)*
UFa (1)
Vapor pressure equation

log p,,
= 6.38363
0.00753771 - 942.76/(t
183.416)
= 6.99464 - 1126.288/(t
221.963)
= 7.69069 - 1683.165/(t
302.148)
= -2892.0/T
2.6990 log T
18.48130
= -1191.1/T
2.5825 log T 0.01023
= -2095.0/T
3.4990 log T 0.39024
+
+
+
+
+
- 1.5340 log T + 12,14545
-
+
+
-1807.5/T
Taken from Oliver et al. (63),and Weinstock el al. (87).

(c) crystalline.
(1) = liquid.
E=
account for the anomalous properties of neptunium hexafluoride. It would

appear that explanations based on presumed differences in the Van der
Wads forces in the solid or on changes in internal energy do not account
for the situation. The suggestion is considered by Weinstock, Weaver, and
228
JOSEPH J. KATZ AND IHVING SHEFT
Malm that the volatility anorrialy of solid NpFs is related to the fact th:tt
of the three actinide hexafluorides, neptunium hexnfluoride is the only one
found to be paramagnetic. If the paramagnetism of neptunium hexafluoride
is reflected in a higher magnetic entropy for the solid than either its vapor
or liquid possesses, an increase in magnetic entropy would also tend to
decrease the vapor pressure of the solid. This is in the opposite direction
to that actually observed. It may, therefore, be necessary to account for
3c
---..___
Pu F6
2,:
2.c
15
I .c
L
5 22.99
5
3.0
3.1
3.2
33
3.3
34
3.4
3
3.5
5
3.6
x 103
FIO.8. Vapor pressure of UF6, NpFs, and PUFB(67).
the observations in terms of nn electronic effect, in which the ground electronic state of the solid is perturbed relative to the ground state of the vapor
to an extent and in the direction which results in a reduced value of the heat
of vaporization. A more definitive assessment of the situation must await
further experimental data.
C. CHANGE
OF STATE
Thermochemical data for heats of fusion, vaporization, and sublimation
are collected in Table XI. Data relating to change in state for the actinide
hexafluorides are given in Table XII.
The triple points are seen to decrease in the order UFB, NpFa, and
PuFe. It can also be seen that, unlike uranium hexafluoride, neptunium
hexafluoride and plutonium hexafluoride possess a liquid range a t atmospheric pressure. Heats of sublimation and vaporization are shown in Fig. 9,
also taken from the paper of Weinstock et al. (87).
229
HALIDES O F T H E ACTINIDE ELEMENTS
TABLE X I
THERMOCHEMICAL
DATAFOR CHANGEOF STATEOF
HALIDESOF ACTINIDEELEMENTS~
Formula
ThFd
ThCla
ThBr4
ThI,
UR
UCI,
UCI,
U13r:l
UBr4
NpF4
PUF>
I'd1'4
PuC13
PuBrr
PuI3
AmF3
AmF4
Process
Temperature
("C)
Sublimation
Vaporization
Sublimation
Vaporization
Vaporization
Sublimation
Vaporization
Vaporization
Vaporization
Sublimation
Vaporization
Vaporization
Sublimat ion
Vaporization
Sublimation
Vaporization
Sublimation
Fusion
Vaporization
Sublimation
Vaporization
Sublimation
Sublimation
Fusion
Vaporization
Snblimation
Sublimation
Fusion
Vaporization
Sublimation
Fusion
Vaporization
Fusion
Sublimation
Sublimation
903
1511
770
770
921
697
697
857
1036
1030
1417
835
835
590
590
752
752
752
519
510
6 13-818
1056
1289
500-851
760
760
681
681
385-640
AH
AS
(kcal/molc)
(cal/mole/deg)
77.1 kO.9
69.9fl.5
59
36.5
35.0
44
34.5
30.4
31.5
68.9
63.2
57.5
63
54
46.3
36.0
68.6
11.0
57.6
41.9
43.8
71.8
96.6
7.9
88.7
68.0
72.8
15.2
57.6
69.9
13.4
56.5
(12)
112.67 f0.12
49.8
41.7
55.6
36.3
29.3
47.2
35.5
26.9
52.6
48.3
34.1
57
49
53.6
41.7
66.9
10.7
56.2
52.9
43.8
46.2
72.9
5.5
51.8
46.2
63
14.7
57.8
74.0
14.0
60.0
22.2
a For detailed referenccs to the original literature see Katz and Seaborg (61),
Rossini
et al. (67), Darnel1 and Keneshea (18), and Yakovlev and Kosyakov (89).
D. THERMOCHEMISTRY
Heats and entropies of formation are assembled in Table XIII. It is
somewhat disconcerting to realize that many of the values given in the
table are estimates, and that experimental values are still lacking.
230
TABLE XI1
DATARELATING
TO PHASE
TRANSITIONS
FOR HEXAFLUORIUES
OF THE ACTINIDE ELEMENT6"
Boiling or Vapor pressure

Triple
Heat of Entropy of
Compound sublimation
at 0C
(oc) mm of Hg fusion
fusion
(cal/mole) (cal/mole/dcg)
point ("C)
(mm of Hg)
~
17.65
20.8
17.9
56.54
55.18
62.16
IJF6
NpFa
l+Fs
~~
4588
4198
4456
1139.6
758.0
533.0
64.05
55.10
51.59
13.61
12.79
13.72
Weinstock ct al. (87).
One interesting point that requires mention is the question of the thermodynamic stability of plutonium hexafluoride. Weinstock and Malm
(85) have concluded that plutonium hexafluoride is unstable toward dissociation by the reaction
PuFBS PuFs
+ Ft.
Thermal decomposition appears to be quite small a t 200, but decomposition of the hexafluoride is practically complete a t 280' in 1 hr. The thermodynamic instability of plutonium hexafluoride toward dissociation has the
12.500
=0-
'
Solid
Liquid
7500
12.000
=\
2;
t"
11,500
I1.000
280
300
320
540
360
]6000
380
TEMPERATURE (.K )
FIG.9. IIents of sublimation and vaporization of U F , NpFe,, and PUPS(87).
consequence that the hexafluoride becomes more stable as the temperature

is raised. In spite of the evidence that indicates that plutonium hexsfluoride is a thermodynamically unstable substance, it can be stored for
long periods of time a t room temperature with rather small losses. This
suggests the possibility that in the course of thermal decomposition the
reaction does not come to equilibrium; it is known that the reverse reaction,
23 1
TABLE XI11
HEATSAND ENTROPIES
OF FORMATION
OF ACTINIDE
ELEMENT
HALLDW~
Formulac
AH,"
(kcal/mole)
(-410 f10)
(-260 f10)
(-220*10)
( - 169 _+ 10)
( -478)
-284.5 f 0 . 5
-229
- 132
( -477)
( -357)
-443 f 3
-182?0.5
-933.8
-483.7
-485.2f1.5
-505
-517
-213 . O f1 . 5
-251.2f0.5
-262.1f0.5
-272.4f0.6
- 170.1 f0.7
-196.6f0.5
-114.7k0.7
-127.0f0.7
( -360 f 2)
(-428 f 3 )
(-463 f 3)
(-216 f 2 )
(-237f1)
-246d
(-174 f1)
(-183 f 1)
(-120 f1)
-374.6f0.5
(-424 f 4 )
( -392)
-230.1 f 0 . 3
-187.8
(-133 k 1)
( - 394)
( - 400)
-251.3
AFf"
(kcal/mole)
So
(cal/mole/deg)
33.95 f 0 . 4
(48.3)
( -339)
-421
- 1727.5
-884.0
-458.2
-458.7 1 . 5
-485
-486
-196.9
-230.0
( -237.4)
-241.5
( - 164.7)
( - 188.5)
( - 115.3)
(-126.1)
(-356)
( -400)
(-214)
( - 182.4)
(- 133.6)
Values in this table are from Kata and Seaborg (61).
* Parentheses are estimated values.
For the solid state unless otherwise indicated.

Rosaini et al. (67).
(26)
36.93f .2
149.0
75.3
48.0
43.3f . 5
90.76 f1
54.45
37.99
47.4
(62)
68.3
(49)
(58)
(56)
(65)
(37.9 f0 . 5 )
88.31
(38)
(49)
(56)
232
the rcfluorinatioa of plutonium tetrafluoride, is a slow reaction a t low temperatures. An experimental determination of the heat of formation of plutonium hexafluoride will be required before the point can be settled. This
is a matter of considerable difficulty because the reaction of plutonium
hexafluoride with water is a complex one. It produces plutonium in d l of
its oxidation states in variable amounts, along with variable amounts of
oxygen, ozone, and perhaps oxygen fluoride. Consequently a simple measurement of the heat of solution does not seem to provide the means whereby
the heat of formation can be deduced.
E. ELECTRONIC
CONFIGURATION
Roth uranium hexafluoride arid plutonium hexafluoride (4.2) possess a
small temperaturedependent paramagnetism. The slight paramagnetism of
uranium hexafluoride is expected because all of its valence electrons are
paired in the six fluoride bonds of UFe. For PuFa, the situation is surprising.
Plutonium hexafluoride contains two nonbonding 5f electrons and might,
therefore, be expected to be strongly paramagnetic. The absence of paramagnetism in plutonium hexafluorjde has been explained as a result of the
octahedral electric field produced by the six fluoride ions. In such a field,
an f-term is split into three levels, fs, fa, and fc, with fs, corresponding to
the onedimensional representation rz,lying lowest. If the two nonbonding
electrons in plutonium hexafluoride are regarded as occupying the fa level
with paired spins, then the small susceptibility observed for plutonium
hexafluoride can be accounted for. The magnetic susceptibility of neptunium hexafluoride has been measured by Weinstock and Malm (86).
Neptunium hexafluoride, like other actinide compounds containing nonbonding electrons, shows a fairly large temperature-dependent magnetic
susceptibility. The octahedral field provided by the fluoride ions results
in an uriusrial behavior for the paramagnetism of NpF8. The susceptibility values are markedly smaller at corresponding temperatures than
those observed for the isoelectronic compound sodium neptunyl acetate,
NaNp02(CH3C00), or than those calculated for the spin-only and unquenched orbital angular momentum cases.
The oxidation number $4 is not known in aqueous solutions of americium and curium. The measurement of the absorption spectra of americium
and curium tetrafluorides by Asprey and Keenan (7) is thus a valuable
contribution to the study of the electronic spectra of the actinide elements.
Special techniques were devised to measure these spectra with microgram
amounts of CmF4and AmF4over the region 3500-20000 b. Table XIV lists
the positions and the relative intensities of the observed maxima. The
agreement of the trivalent fluoride spectra with those of the corresponding
233
TABLE XIV
ABSORPTIONSIWCIXAMAXIMA(A) OP YM- A N L , TETRAFLUORIDES

OF
CURIUM A N D AMERICIUM~
(.kSl>RICY A N D NEENAN,
7)
AmF8
CmF3
AmFl
CmFl
3600 w
3652 w
3760 m
4003 w
4150 w
4378 w
4530 w
5006 s
5085 s
5148 m
7900 s
8250 w
9015 w
9918 w
10460 m
10660 w
11120 w
3740 m
3780 m
3950 s
4490 w
3760 s
4530 s
4860 w
5360 w
5680 s
6390 m
7030 m
7440 m
8620 m
9180 m
3865 m
4010 m
4118 m
4504 s
4607 ni
6730 w
6960 w
7650 w
7915 w
8560 s
9100 w
10975 w
16120 w
s, sharp peak; m, moderately sharp peak;
w, weak or very broad peak.
trivalent ions is very close. Absorption in 4$ gadolinium(II1) is confined to

regions below 3200 A. The electronic transitions of 5$ curium(II1) lie at
lower energies because the absorption bands are moved towards the visible.
This shift, as well as the very existence of curium(IV), emphasizes that the
5.f electrons in the actinide elements are less tightly bound than the corresponding 4f electrons in the lanthanide elements.
REFERENCES
1. Adams, M. D., Steunenberg, R. K., and Vogel, R. C., Argonne National Laboratory
Report ANL-5796 (1957).
2. Agron, P., Grenall, A., Kunin, R., and Weller, S., Paper 62 in Chemistry of Uranium: Collected Papers, Document TID-5290, p. 652. U. S. Atomic Energy Commission., Technical Information Service Extension, Oak Ridge, Tennessee, 1958.
3. Allen, R. J., Petrov, H. G., and Magno, P. J., Znd. Eng. Chem. 60, 1748 (1958).
4. Anderson, J. S., and DEye, R. W. M., J. Chem. SOC.Suppl. 2,244 (1949).
5. Anderson, J. S., Bull. SOC.Chim. France 20, 781 (1953).
6. Asprey, L. B., Ellinger, F. H., Fried, S., and Zachariasen, W. H., J . Am. Chem. Soe.
79, 5825 (1957).

7 . Asprey, L. B., and Keenan, T. K., J . Znorg. & Nuclear Chem. 7, 27 (1958).
8. Bagnall, K. W., Chemistry of the Rare Radioelements, p. 32 el seq. Butterworths,
London, 1957.
9. Bernhardt, H. A., Barber, E. J., and Gustison, R. A., Ind. Eng. Chem. 61, 179
(1 959).
234
JOSEPII J. KATZ AND IRVING SHEFT
10. Bigclciscn, J., Mayt:r, &I. G., Stevonsim, P. C., and Turkwich, J., ,I. Cliem. Ihys.
16, 442 (1948).
11. Booth, 11. S., Krmny-Ergen, W., and IIeath, B. E., J . Am.. Chern. SOC.68, 1969
(1948).
12. Bratctr, D. C., and Smiley, S. H., Progress in Nuclear Energy, Series I11 Process
Chcmistry, Vol. 2, p. 136. Pergarnon, London, 1958.
13. Budaev, I. V., and Volskii, A. N., Proc. 2nd Intern. Conf. on Peaceful IJses of Atomic
Energy A/CONF. 16/P/2195 (1958).
2 4 . Cacciari, A., Fizzotti, C., Gsbaglio, G. M., and Dc Leone, R.,Energie nuc16uire 1,
11 (1957).
25. Cathers, G. I,, Bennett, M. R., and Jollcy, R. L., ZrLd. Eng. Chem,. 60, 1709 (1958).
16. Christensen, E. L., and Mullins, L. J., LOSAlnmos Scientific Laboratory Report
LA-1431 (1952), decl. Sept,. 1955.
17. Clanssen, H. H., Weinstock, B., and Malm, J. G., J. Chem. Phys. 26, 420 (195.6).
28. Darnell, A. J., and Keneshea, F. J., Jr., J . P h p . Cham. 62, 1143 (1958).
19. Dawson, J. K., Elliot,, R. M., Hurst, R., and Trnswell, A. E., J. Chem. SOC.p. 558
(1954).
20. DEye, R. W. M., and Booth, W., J . Inorg. c!& Nuclear Cham. 1, 326 (1955).
22. DEye, It. W. M., and Booth, W., J . Inorg. & Nuclear Chem. 4, 13 (1957).
22. Douglnss, R. M., and Stttritzky, E., Anal. Chem. 29, 459 (1957).
23. Ellis, J. F., and .Johnson, K. D. B., J. Inorg. LY: Nuclear Ch.em. 6 , 194 (1958).
$4. Ellis, J. F., Brooks, L. H., and Johnson, K. D. B., J . Inorg. & Nuclear Ch.em. 6 , 199
(1958).
M. Emelbus, I. J., Maddock, A. G., Miles, G. L., arid Sharpc, A. G.,J. C h e m SOC.
1). 1991 (1948).
bfi. Emelbus, H. J., and Woolf, A. A., J. Chem. Soc. p. 1G4 (1950).
by. Feay, D. C., Thesis, University of California, Ucrkeley (1954), issued as Universily
of California Radiation Labomtory Report, UCRL-2547 (1954).
%8. Ferris, 1,. M,, J . A m . Cheni. Soc. 79, 5419 (1957).
2-9. Ferris, L. M., Ind. Eng. Chent. 61, 200 (1959).
30. Fischor, J., and Vogol, R. C., J . A m . Chem. Sot. 76, 4829 (1964).
32. Florin, A. E., Tannenbaurn, I. R., and Lemons, J. F., J. I m r g . & Nuclenr Cham. 2,
368 (1958).
32. Frccman, J. H., iind Smith, M. C., J. Zfiorg. ck Nuclear Cheni. 7, 224 (1958).
33. Fried, S.,nnd Davidson, N. R., U. H. Atomic Energy Commission Report AECD2981 (1945).
34. Fried, S., nnd Davidson, N. R., J . Am. Chem. SOC.70, 3539 (1948).
35. Fried, S., and Dnvidson, N. It., Paper 6.11 i t i The Traimurtmium Elements,
National Nuclear Enitrgy L%ries, Division IV, Volume 148. McGraw-Hill, New
York, 1949.
36. Fried, S,, Hagemann, F., and Znchariascn, W. H., J. Am. Cliem. Soc. 72, 771 (1950).
37. Frictd, S., and Davidson, N. R., Paper ti7 in Chemistry of Uranium: Collected
Papers, Document, TID-Fj290,p. 888. IJ. 8. Atoinic Energy Commission, Technical
Informat,ion Service Extension, Oak Ridge, Tennessee, 1958.
38. Gaunt, J., Trans. Parnday SOC.60, 546 (1954).
39. Gibson, G., Gruen, D. M., and Katz, J . J., J. Am. Chem. Soc. 74, 2103 (1952).
40. Goblc, A. G., Goldcn, J., Maddwk, A. G., and T u r n , D. J., Progress in Nuclear
Energy, Swies IT1 Process Chemistry, Vol. 2, 1). 86. Pergarnon, London, 1958.
42. Grossc, A. V., Manhat,t,an Project Report A-83 (1941).
4%. Gruen, D. M., Malm, J. G., and Weinstock, B., J. Chem. Phys. 24, 905 (195G).
43. Hayek, E., and b h n e r , T., Ezperienta 6, 114 (1949).
235
44. Hayek, E., Rehner, T., and Frank, A., Mmatsh. 82, 575 (1951).
46. Hermann, J. A., and Suttle, J. F., Inorg, Syntheses 6, 143 (1957).
46. Hindmann, J. C., and Wehner, P., J . Am. Chem. Soc. 76, 2869 (1953).
47. Hyman, H. H., Vogel, R. C., and Katz, J. J., Progress in Nuclear Energy, Series
111, Process Chemistry, Vol. 1, p. 261. Pergamon, London, 1956.
48. Jantsch, G., Homayr, J., and Zemek, F., Monatsh. 86, 526 (1954).
49. Jonke, A. A., and Levitz, N., Progress in Nuclear Energy, Series 111, Process
Chemistry, Vol. 2, p. 122. Pergamon, London, 1958.
60. Katz, J. J., and Rabinowitch, E., The Chemistry of Uranium. McGraw-Hill,
Ncw York, 1951.
61. Kata, J. J., and Seaborg, G. T., LThe Chemistry of the Actinide Elements.
Methuen, London, 1957.
62. Kuhlman, C. W., Jr., and Swinehart, B. A., Znd. Eng. Chem. 60, 1774 (1958).
63. Labaton, V. Y., and Johnson, K. D. B., J . Znorg. & Nuclear Chem. 10, 74 (1959).
64. Labaton, V. Y., J . Inmg. I4 Nuclear Chem. 10, 86 (1959).
65. Langham, W. H., Proc. 7th Hot Labs. and Equipment Conf., Cleveland, Ohio p. 256
(1959).
66. Malm, J. G., Weinstock, 13., and Claassen, H. H., J . Chem. Phys. 23, 2192 (1955).
57. Malm, J. G., Weinstock, B., and Weaver, E. E., J . Phys. Chem. 62, 1506 (1958).
58. Mandleberg, C. J., Rae, H. K., Hurst, R., Long, G., Davies, D., and Francis, K. E.,
J . Znorg. I4 Nuclear Chem. 2, 358 (1956).
59. Martin, H., Albers, A., and Dust, H. P., 2. anorg. u. allgem. Chem. 266, 128 (1951).
60. Massoth, F. E., and Hensel, W. E., Jr., J . Phys. Chem. 62, 479 (1958).
61. Massoth, F. E., and Hensel, W. E., Jr., J . Phys. Chem. 63, 697 (1959).
62. Mcars, W. H., Townend, R. V., Broadly, R. D., Turissini, A. D., and Stahl, R. F.,
Ind. Eng. Chem. 60, 1771 (1958).
63. Oliver, G. O., Milton, H. T., and Grissard, J. W., J . Am. Chem. SOC.76, 2827 (1953).
64. Panzcr, R. E., and Suttle, J. F., J . Inorg. R- Nuclear Chem. to be published.
65. Prigent, J., Compt. rend. acad. sci. 238, 102 (1954).
66. Robb, W. L., Brandon, R. J., Meyers, R. L., and Galpern, H. N., Knolls Atomic
Power Laboratory Report KAPL-1728 (1957).
67. Rossini, F. D., Wagman, D. D., Evans, W. H., k v i n c , S., and Jaffe, I., Null. Bur.
Standards ([I.8.) Circ. 600, (1952).
68. Ruff, O., and Heinzelmann, H., Z . anorg. u. allgem. Chsm. 72, 63 (1911).
69. Ruff, O., and Wallstein, R., 2. anorg. Chem. 128, 96 (1923).
70. Runnals, 0. J. C., Can. J . Chem. 31, 694 (1953).
71. Schomaker, V., Kimura, M., Smith, D., and Weinstock, B., t o be published.
72. Schulz, W. W., Neuvar, E. W., Carroll, J. L., and Burns, R. E., Znd. Eng. Chem. 60,
1768 (1958).
73. Sfaborg, G. T., and Katz, J. J., eds., The Actinide Elements. McGraw-Hill,
New York, 1954.
74. Seaborg, G. T., The Transuranium Elements. Yale University Press, New Haven,
Conn., 1958.
75. Sellers, P. A., Fried, S., Elson, R. E., and Zachariasen, W. H., J. Am. Chem. Soc.
76, 5935 (1954).
76. Sheft, I., and Fried, S., J . Am. Chem. Soe. 76, 1236 (1953).
77. Sheft, I., Hyman, H. H., Adams, R. M., and Kata, J. J., Abstr. 186th Meeting Am.
Chem. Soc., Boston. p. 28R (1959).
78. Shiflett, C. H., Steidlitz, M. E., Rosen, F. D., and Davis, W., Jr., J . Znorg. & Nuclear
Chem. 7, 210 (1958).
236
JOSEPH J. KATZ ASD IRVING SHEFT
79. Smiley, S. H., and Brater, D. C., Progress in Nuclear Energy, Series I11 Process
80. Smiley, S. H., and Brater, D. C., Progress in Nuclear Encrgy, Series I11 Process
81. Stahl, R. F., Znd. Eng. Chem. 60, 1771 (1958).
86. Steindler, M. J., Steidl, D. V., and Steunenberg, R. K., Argonne National Laboratory Report ANL-5875 (1958).
83. Warf, J. C., J. Am. Chem. Snc. 74, 1864 (1952).
84. Watt, G. W., Sowards, D. M., and Malhotra, S. C., J . Am. Chem. Soc. 79, 4908
(1957).
86. Weinstock, B., and Malm, J. G., J. Znorg. k Nuclear Chem. 2, 380 (1956).
86. Weinstack, B., and Malm, J. G., J. Chem. Phgs. 27, 594 (1957).
87. Weinstock, B., Weaver, E. E., and ,Malm, J. G., J. Inorg. & Nucleur Chern. 11,
104 (1959).
88. Westrum, E. F., Jr., and Eyring, L., J . Am. Chcm. Soc. 73, 3896 (1951).
89. Yakovlev, G. N., and Kosyakov, V. N., Proc. 2nd Intern. Conf. on Peaceful Uses of
Atomic Energy A/CONF. l6/P/2127 (1958).
313.Young, It. C., J. Am. Chem. Sor. 63, 2148 (1931).
STRUCTURES OF COMPOUNDS CONTAINING CHAINS

OF SULFUR ATOMS
Olav Foss
Chemical Institute. University of Bergen. Bergen. Norway
I . Introduction
. . . . . . . . . . . . . . . . . . 237
I1. Elemental Sulfur
I11. The Polysulfide Ions .
. 240
. 241
IV The Sulfur Hydrides and Halides
V. Organic Disulfides and Polysulfides . . .

A . Various Physicochemical Evidence . .
B. Structure Determinations by Electron and
C. Di- and Triselenides, Ditellurides . .
. . . . . . . . . 243
. . . . . . . . . 244
X-ray Diffraction Methods . 247
. . . . . . . . . 248
VI . Sulfur Chains Terminated by Cyano Groups .
VII . Sulfur Chains Terminated by Sulfonyl Groups

A . Types of Compounds . . . . . . .
B. X-ray Studies . . . . . . . . .
. 248
. 250
. 250
. .
.
.
VIII . Sulfur Chains Terminated by Sulfonate Groups: The Polythionates

A . Introductory Remarks
. . . . . . . . . . . .
B. Structure Determinations by X-ray Methods . . . . . .
C. Bond Lengths in the Sulfur Chains . . . . . . . . .
D. Limit of the Polythionate Series . . . . . . . . . .
.
.
.
.
.
IX . Unsymmetrical Compounds .
X . Comments on Structural Evidence .
. 238
. 254
254
. 256
. 263
. 264
265
266
XI . Stereochemistry . . . . . . . . . . . . . . . . .
A . Nonplanarity of Sulfur Chains . . . . .
. . . . . .
B. Cyclic Disulfides . . . . . . . . . . . . . . . .
C . Selenium and Tellurium Analogs
. . . . . . . . . . .
D . Peroxy Compounds . . . . . . . . . . . . . . .
XI1. Refcrenccs .
251
267
267
271
271
272
. 273
I Introduction
The ability of sulfur to form chains is displayed in the allotropes of the

element. and in a variety of compounds built up of sulfur chains terminated
by other atoms or groups. The maximum number of sulfur atoms in the
chain of compounds so far isolated and characterized. varies with the
237
238
OLAV FOSS
nature of the eiid groups; iii some it is four or six, and in others reaches
eight.
The major structural problem with regard to sulfur chain compounds
originates in their tendency t o give off sulfur when acted upon by bases or
by sulfur acceptors like sulfite or cyanide. Are the sulfur atoms all divalent
and part of unbranched chains, or are the labile oms attached in a different
way, through coordinate sulfur-sulfur bonds'? Closely corinected are problems concerning reactivity, and the mechanisms by which desulfurations
and shortenings of chains take place. There is a t present convirlcing evidence, chemical as well as physical, that in compounds containing chains
of sulfur atoms the chains are unbranched. Indeed, branching has not been
found in any instance. This review is concerned with the physical evidence,
particularly from structure determinations by X-ray methods; the chemical
side is mentioned but only leading references are given.
II. Elemental Sulfur
The element in its various allotropes presents the prototype structures

of the chains in conipouiids. An eight-membered ring occurs in orthorhombic, y-monoclinic, and probably also p-rnonoclinic sulfur; a six-mcrnbered ring in the rhombohedral form; and long helical chains in fibrous and
probably also in viscous sulfur.
The structure of orthorhombic or a-sulfur was redetermined a. few
years itgo by Abrahams (2), The SSring has a puckered iicrown" form, with
a sulfur-sulfur bond length of 2.04 A, a sulfur valency angle of 107.5",
and a dihedral angle of 99" between successive planes through three and
three sulfur atoms. Ile Haan (46) has recently reported a crystal structure
andysis of y-sulfur. the monotropic monoclinic allotrope discovered by
(:erne& in 1883 and by him called nacreous sulfur; the Ss rings also occur
there. No structure analysis appears to have been made of ordinary monoclinic sulfur, p-sulfur; Das (42) was not able to obtain X-ray photographs
of it owing to too rapid transformation into a-sulfur. The preliminary results
of a crystal structure ttnalysis of rhombohedral or p-sulfur have been reported by Donohue et al. ( 4 7 ) ; the primitive rhombohedral cell contains
six sulfur atoms in a puckered ring of sixfold molecular symmetry.
Liquid sulfur is almost entirely SSup to 160"; a t higher temperatures
equilibrium mixtures of Se and polymeric forms predominate. Work by
Gee (128) and paramagnetic resonance measurements by Gardner and
Fraerikel (127) show that the polymers are probably dirndical chains with
n maximum :Lversge length of about lo6 atoms a t approximately 170".
Schenk (196) has recently reviewed the evidence for long chain molecules
STRUCTUBES OF S U L F U R CHAIN COMPOUNDS
239
in liquid sulfur; the alternative theory propounded by Krebs (154) is that

large rings occur instead of open chains.
When viscous sulfur is chilled in water and strongly extended, so-called
fibrous sulfur results. It gives an X-ray diagram and its structure has
recently been elucidated by Prins et al. (190). The fibers contain two constituents, one consisting of very long, roughly close-packed helixes, and the
other of small crystals of y-sulfur (S8 rings) located in needle-shaped holes
of the helix structure. Each helix has periods of 10 atoms in three turns
on a cylinder of 0.92 A radius, the length of a period being 13.7 A. This
gives a S-S bond length of 2.04 A and a sulfur valency angle of 107", as
in the s8 ring, and a dihcdral angle of 87" against 99" in the ring.
The same helixcs occur, according to Prins, Schenk and Wachters, in
the so-called co-sidfur, which is the insoluble residue remaining when chilled
viscous sulfur or flowers of sulfur are extracted with carbon disulfide, and
which can be produced also in other ways, for instance by hydrolysis of
disulfur diohloride. This form is hexagonal, as was discovered in 1938 by
Das, who has rccently (42) reviewed his X-ray powder studies on allotropes
of sulfur.
The structure of the Ss molecule in the gas phase has been determined by
Lu and Donohue (168) by electron diffraction; the same puckered ring
occurs as in the solid state. Assignment of infrared and Raman frequencies
has been made by Guthrie et al. (139),who also discussed the relative stabilities of possible forms of the ring in the gas phase.
The different conformations of molecular sulfur in the eight- and sixmembered rings and in the extended helixes are examples of rotational
isomerism, encountered also in derivatives. The all-cis configuration of S6
and s8 transforms into the all-trans of the helixes on breaking of a bond and
rotation of approximately 180" about the symmetrically located bond.
An eight-membered ring of seven sulfur atoms and a n imide group,
heptasulfur imide SyNH, may be briefly mentioned here. According to
Goehring et at. (134) the crystals are orthorhombic, with four molecules
per unit cell, and space group probably Dzh16-Pnma. If this is so, the molecule must have mirror plane symmetry in the crystals, and since it is certainly nonplanar, it lies across the mirror plane with the imide group and
one sulfur atom in the plane. On this basis, a structure quoted by Goehring
(132) can hardly be correct.
The largest molecule of oxygen is triatomic ozone, but selenium and tellurium form longer chains in the elemental state, like sulfur. The crystal
modifications of selenium
structures of the two red, monoclinic (aand
have been determined by Burbank (38)and Marsh et al. (170). I n these,
the molecules are Se8 rings of the same "crown" form as s8 sulfur, while
a)
240
OLAV FOSS
in hcxagorial selenium and tellurium infinite helixes occur. The hexagonal

space group being enantiomorphous, each single, untwinned crystal contains only right-handed or only left-handed helixes.
Ill. The Polysulfide ions
The highest alkali polysulfides are hexasulfides, while in salts with

organic bases, up to nonasulfides appear to exist. From the systematic
studies by Feh6r and co-workers of the systems sodium-sulfur ( G I ) , potassium-sulfur (6@, and rubidium- and cesium-sulfur ( 7 l ) , the only polysulfides in the di- through hexasulfide series which do not exist as individuals
in the solid state, are sodium trisulfide and hexasulfide, rubidium tetrasulfide and hcxasulfidc, and cesium tetrasulfide. The others, among them
the whole series of potassium salts from di- through hexasulfide, were isolated in a pure state and gave characteristic X-ray powder patterns. The
salts were in most cases prepared from the elements in liquid ammonia;
some of them such as sodium tetrasulfide and potassium tri- and pcntasulfide crystallize well from ethanol and were obtained from sulfur and the
monosulfide in this solvent.
The color of the alkali polysulfides changes with increasing sulfur content from the light yellow of disulfides through orange to red or red-brown.
Hexasulfides liberate sulfur when dissolved in water, which is understandable from the measurements of Arntson et al. (8) of the solubility of orthorhombic sulfur in aqueous sodium monosulfide. The equilibrium value of
n in Sn-- was found to be approximately 4.8 a t 2 5 O , and varied at 50 from
approximately 4.6 in dilute solutions to approximately 5.4 at higher
concentrations.
Salts of hydrogen polysulfides with organic bases have been known for
a long time; a strychnine hexasulfide was prepared by Hoffmann in 1868.
ltccently Krebs (166) obtained cyclohexylammonium hexasulfide, several
heptasulfides, and triethyl- and di-n-propylammonium nonasulfide, from
the amine, hydrogen sulfide, and sulfur in benzene. In the same way but
using aqueous media, Krebs and Muller (165)prepared ethylenediammonium trisulfide, triethylenediammonium penta-, hexa-, and heptasulfides,
and tetramethylenediammonium hexasulfide, and determined unit cells
and space groups from single-crystal X-ray photographs. Three of the salts
crystallize in the enantiomorphous space group D2a-P21212with four formula units per unit cell. The salts are relatively unstable.
The di- and trisulfide ions offer no structural problem apart from bond
lengths, and the bond angle in the trisulfide. For the tetrasulfide and higher
polysulfides branched structures can be imagined, such as a tetrathiosulfate atruetura for the pentasulfide. The recent crystal structure deter-
Sl'ItUCTURES OF SULFUR CHAIN COMPOUNDS
241
minations by Abrahanw and Grison (5) of cesium hexasulfide, and by

Abrahams (1) of barium tetrasulfide monohydrate, provide proof of
unbranched sulfur chain structures in these salts and, by inference, in other
polysulfides.
The crystals of barium tetrasulfide monohydrate, Basd . HzO,-aEe
orthorhombic, with four formula units in a cell of dimensions a = 9.67 A,
b = 7.99 8, c = 7.81 8. The space group is D23-P21212.The unbranched,
nonplanar tetrasulfide ions lie on twofold axes of symmetry, with only righthanded or only left-handed forms in each unit cell.
Cesium hexasulfide, CszSe, cryst$Iises in the triclinic space group
C+Pi with a = 11.53 A, b = 9.18 A, c = 4.67 A, a = 89.2", p = 95.2",
y = 95.1", and two formula units per unit cell. The hexasulfide ions are
unbranched and nonplanar and have the shapes of extended helixes.
No crystal structure analysis of a pentasulfide has been reported. As
for di- and trisulfides, Miller and King (174) worked out a structure for
barium trisulfide on the basis of X-ray powder data, but, as pointed out
by Wyckoff (231), the structure cannot be considered as well established.
Sodium disulfide (61) occurs in a low-temperature (a)and a high-temperature (p) form which, according to Foppl (81), are isomorphous with sodium
peroxide and lithium peroxide, respectively. Foppl obtained single crystals
of /3-Na2S2and has tentatively reported (81) a length of 2.25 8 for the disulfide bond in this salt; the value appears unreasonably large in view of
the lengths 2.02-2.11 A found for bonds between divalent sulfur atoms in
other compounds. For instance, in hydrogen disulfide (211) the S-S bond
is 2.05 A. The 0-0 distance in the peroxide ion, 1.49 8, which is known
accurately from Foppl's work (80) on alkali peroxides, is the same as in
covalent peroxy compounds.
IV. The Sulfur Hydrides and Halides
The crude yellow oil resulting from the interaction of sodium polysulfide
solutions with an excess of hydrochloric acid was discussed by Scheele in
1777, but one and a half centuries later the only hydrogen polysulfide isolated in a pure state was the trisulfide, obtained by Bloch and Hohn (29)
in 1908 by fractional distillation of the crude ,oil under reduced pressure.
Mills and Robinson (175) in 1928 prepared hydrogen pentasulfide from
ammonium pentasulfide and formic acid, but according to later work (64)
the product may not have been quite pure. Only during the last 15 years
have Feh& and co-workers isolated and characterized the complete hydrogen polysulfide series up to and including the octasulfide.
The preparation and properties of hydrogen trisulfide (58),tetrasulfide
(59),and penta- and hexasulfide (60) were described by Feh6r and Baudler
242
OLAV FOSS
in 1947-1940. The tetra-, peiita-, and hexasulfides were ohtaincd by fractional distillation in high vacuum of crude hydrogen polysulfide oil, the
hexasulfide only with difficulty. The crude oil normally does not contain
any hydrogen tri- and disulfide, but yields these on cracking. PehBr, Lsue,
and Wirikhaus (68) have later improved the procedure and the purity of
the products. Feh& and Winkhaus (78) in 1956 reported the preparation
of hydrogen penta- and hexasulfides, and also hepta- and octasulfides, by
a new method consisting in the condensation of two moles of hydrogen disulfide with one mole of mono-, di-, tri-, or tetrasulfur dichloride. An excess
of hydrogen disulfide must be used, but this compound is more volatile
than the products and hence easily removed afterwards. The hepta- and
octasulfide could not be difitilled but were practically pure. By varying
the sulfur content and the relative amounts of the starting hydride and
chloride, the method can be used to prepare mixtures of hydrogen polysulfides with average chain lengths of from 6 to 30 sulfur atoms.
The hydrogen polysulfides tue yellow liquids, which from the tctrasulfide upwards have no definite frwzing points. They are labile with respect
to change into sulfur and hydrogen sulfide, and decompose on contact with
powdered qutrtz and glass, wood, or paper, or traces of basic substances :
the same applies to hydrogen disulfide.
The situation with regard to polysulfur dichlorides arid dibroniides
resembles that for the hydrogen polysulfides. Although the presence of
trisulfur dichloride in heated rnixtures of disulfur dichloride and sulfur
had been assumed by Spong (209)and others, it was not, until a few years
tigo that FehBr, Naused, and Weber (72) prepared polysulfur dichlorides
S,C12 with n = 3 through 6. The compounds were made by an adaption
of the method used for the preparation of hydrogen polysulfides, by now
condensing hydrogen mono-, di-, txi-, and tetrasulfides with an excess of
monosulfur dichloride. By the use of disulfur dichloride instead of monosulfur dichloride, FehBr and RistiB (73) similarly prepared penta-, hexa-,
hepta-, and octasulfur dichlorides.
The polysulfur dichlorides are orange-red, oily liquids, of deepening
color with increasing chain length. S3C12 can be distilled in high vacuum,
S4C12only under partial decomposition. At -20' they can be kept for some
weeks without appreciable change; a t room temperature half of a sample of
SdC12decomposed in about three weeks.
Hydrogen bromide reacts with di- and polysulfur dichlorides with elimination of hydrogen chloride, without rearrangement of the sulfur chain,
to give di- and polysulfur dibromides. In this way, Feh6r and 12istii: (74)
prepared practically pure di- through octasulfur dibromides. These are
brown-red liquids, the color becoming less intense with increasing sulfur
content.
STRUCTURES OF SULFUR CHAIN COMPOUNDS
243
Feh& has in two recent reviews (55, 56) discussed preparative methods
and properties of sulfur chain compounds studied by his group. The success
of the work appears to be due to, beside a remarkable preparative skill, the
use of Raman spectra for the identification and characterization of the different members of the series.
The Raman spectra are consistent with unbranched structures of the
sulfur chains; they display characteristic S-S stretching and S-S-S
bending frequencies but no frequencies corresponding to coordinate sulfur-sulfur bonds.
Further support for unbranched structures in hydrogen polysulfides and
polysulfur dichlorides comes from molar volumes, refractions, and viscosities; these are discussed in Section V on organic disulfides and polysulfides.
In the case of hydrogen disulfide, the S-S bond length as determined
hy Stevenson and Beach (211) from electron diffraction indicates that
branching does not occur. Electron diffraction studies by Palmer (184)
and Guthrie (138) on disulfur dichloride show a n unbranched, nonplanar
structure. Smyth (206) has discussed the dipole moments of hydrogen disulfide, disulfur dichloride, and diselenium dichloride on the same basis;
and Hooge and Ketelaar (145),the vibrational spectra of hydrogen disulfide
and disulfur difluoride, dichloride, and dibromide.
V. Organic Disulfides and Polysulfides
These make up the largest class and display a wide variety of terminal
groups, from simple and substituted alkyls and aryls to acyls and amino
acid residues. Disulfides are usually quite stable; polysulfides are less
stable but can, like the disulfides, be obtained in a variety of ways. Schoberl
and Wagner (198) have recently reviewed preparative methods for tri-,
tetra-, and pentasulfides. Only a few hexasulfides are known; for example,
dimethyl, diethyl, and dibeneyl hexasulfides which were prepared by Bohme
and Zinner (34) from the alkyl hydrogen trisulfides by oxidation with iodine; they are yellow oils a t room temperature. The bis(thiocarbamy1)hexasulfides studied by Levi (164)and Blake (28) are crystalline and apparently stable.
Cyanogen polysulfides, although organic in the sense that the bonding
atom of the terminal group is carbon, constitute a unique series and are considered separately, in Section VI. Another, not strictly organic, series which
may be briefly mentioned here, is the sulfur amides, R2N-Sn--NR2
with
n = 1 through 4.The tri- and tetrasulfur derivatives are formed (147, 163)
when a secondary amine, like piperidine or dimethylamine, reacts with
sulfur in benzene in the presence of lead oxide or mercury oxide as a sulfide
acceptor. Goehring (130) has studied the Raman spectrum of disulfur
244
OLAV FOSS
bis(dimethylamide), and Jenne and Becke-Goehring (147) reported that

the infrared spectra of di-, tri-, and tetrasulfur diamides in the range 7005000 em-' are practically identical. Two oxygen disulfur analogs are known,
namely, the alkoxides S2(0CH&and Sz(OCzHa)2,
but no higher members
of this series. The structure of the alkoxides has been discussed by Goehring
(130) on the basis of Raman spectra and measurements by other authors
of dipole moments and diamagnetic susceptibilities.
The sulfur-sulfur bond in di- and polysulfides is easily cleaved, particularly by nucleophilic reagents. The reactions lead to interconversions and
desulfurations and have often been taken to indicate branching of the sulfur
chains, but are actually consistent with unbranched structures (87, 97).
A. VARIOUSPHYSICOCHEMICAL
EVIDENCE
The literature on structure of organic di- and polysulfides is quite extensive. In the last decade convincing evidence has accumulated in favor of
unbranched sulfur chain structures, as contrasted with branched, coordinated ones.
1. Molar Volume and Refraction, Viscosity
Accurate measurements have been carried out by Feh& and co-workers

on dimethyl di- through tetrasulfides and diethyl di- through pentasulfides
(65), hydrogen di- through octasulfide (67, 69), and monosulfur through
hexasulfur dichlorides (72). Within a homologous series S,X2, the molar
volume and Lorentz-Lorenz refraction and the logarithm of the dynamic
viscosity are linear functions of chain length. The relation (69):
Molar volume
= 2u
Molar refractioii
+ (n - 2)b
is followed within experimental error, a being the cuntribution of the end

groups SX and b the increment for each inner sulfur atom. For molar refractions, RD~O,b = 8.6, 8.9, 8.8, and 8.9 em* in the hydrogen polysulfidc,
polysulfur dichloride, and dimethyl and diethyl polysulfide series, respectively, while for molar volumes, V*olb = 16.4, 16.2, 16.3, and 16.3 emS.
As pointed out by Fehbr (56)the constancy of the increments b supports
the contention that the sulfur atoms are similarly linked together to form
unbranched chains. It should be noted, though, that the regular increments
are consistent with but do not prove unbranched structures; a homologous
series having both hetero atoms or groups attached a t the same end of the
sulfur chain might well display similar properties.
Measurements by Minoura (178) of molar refraction of dibenzyl and
245
di-p-tolyl di-, tri-, and tetrasulfide give increments between 9.1 and 12 em3
per sulfur atom; in the di-n-hexadecyl di- to tetrasulfide series (230) the
increments are 8.2 and 11 cm3.
2. Dipole Moments
Dipole moments in benzene of di-n-hexadecyl mono-, di-, tri-, and tetrasulfide have been reported by Woodrow, Carmack, and Miller (230), of
dimethyl and diethyl di- and trisulfide and di-n-propyl disulfide by Smyth
and co-workers (160,225),of the three isomeric dibutyl disulfides by Rogers
and Campbell (193), and of diphenyl disulfide and p-substituted diphenyl
disulfides by Gur'yanova (137).
In dialkyl di- and polysulfides, sulfur is the negative end of the dipole.
The moment of a di-n-alkyl disulfide is approximately 2.00 D and of a trisulfide approximately 1.65 D, irrespective of the length of the carbon
chain. In a tetrasulfide the moment is again higher; thus, the alternating
values 1.47,12.00, 1.63, and 2.16 D, respectively, were obtained by Woodrow et al. (230) for the di-n-hexadecyl mono- through tetrasulfide series.
In di-t-butyl disulfide (193), the moment 1.86 D is lower than in normal
disulfides owing to steric effects. In diphenyl disulfides (137) the moments
vary with the nature of the substituent, and in di-p-nitrophenyl disulfide
(4.31 D in dioxane at 20') sulfur is the positive end of the dipole.
The dipole moments support an unbranched chain structure for di- and
polysulfides, and as referred to later, provide information about the stereochemistry of sulfur chains.
3. Diamagnetic Susceptibility
Fava and Iliceto (54) measured the susceptibilities of n-octane, di-n-butyl

mono-, di-, tri-, and tetrasulfide and impure penta- and hexasulfide, and
concluded on the basis of the found sulfur increments that the sulfur chains
are unbranched. The fact that the increments decrease slightly with
increasing chain length was interpreted as indicating partial double-bond
character of the sulfur-sulfur bonds.
4. Ultraviolet Spectra
Measurements have been made by Koch (152, 153) on disulfides and
diethyl, diphenyl, and di-2-benzthiazyl tetrasulfide and dicyclohexyl
hexasulfide, and by the following authors on complete series R2S, with
n = 1 through 4: Baer and Carmack (12) for R = n-hexadecyl, Minoura
for R = benzyl (176) and p-tolyl ( l 7 7 ) , and Haszeldine and Kidd (142)
for R = trifluoromethyl. Harris (140) has recorded the spectra of the di-onitrophenyl mono- through pentasulfide series, Gorin and Dougherty (135)
those of the di-n-alkyl disulfides from methyl through butyl, and dimethyl
246
OLAV FOSS
trisulfide, and Schotte (199-203) has recently carried out extensive spectrochemical studies of noncyclic and cyclic disulfides as well as of some
polysu lfides.
The ultraviolet spectra of polysulfides are difficult t o interpret in terms
of structure. However, as first pointed out by Koch (152),the fact that the
various polysulfides all display characteristic absorption of the same spectral type, resembling the diiilkyl disulfide spectrum displuccd towards
greater wave lengths and intensities, provides (in view of the known structure of disulfides) strong evidence in favor of unbranched structures also
in polysulfidcs.
Noricyalic d$ulfides normally absorb in the near ultraviolet with a peak
at about 2500 A and more strongly in the vacuum ultraviolet with maximum a t about 2000 A. Studies of shifts in the position of the first peak have
led to interesting results, to be discussed later, concernirig the barrier to
rotation about sulfur-sulfur bonds.
5 . Infrured and Barnan Spectra
The S-S stretching frequency gives rise t o only rather weak absorption
in the infrared, probably because the symmetrical vibration of the bond
involves little change of dipole moment. It is strong and easily detected in
the Itaman spectrum, and usually occurs between 440 and 510 cmn-'. The
S-S--S
bending frequency normally lies a t about 150-250 cmn-', and the
S-S torsional frequency a t about 100-120 cm-l. Recent careful assignments of vibr:ttional frequencies have been made by Scott and co-workers
for dimethyl disulfide (204,diethyl disulfide (205), arid sulfur Sa (139),
and by Hooge arid Ketelaar (14.') for hydrogen disulfide, dimethyl disulfide,
and the disulfiir dihalidcs.
The vibrational frequencies are consistent with unbrttnchcd, noriplunar
structures. The fact that the S-S stretching frequency, although weak,
occurs in the infrared, excludes a planar, centrosymmetric trans form for
disulfides. In the Raman spectra of di- and polywilfides, no lines are found
which can be assigned to a coordinate sulfur-sulfur bond; such a bond should
be characterized by a rather strong line and its absence implies that branching docs not orrur.
Recent studies of infrared spectra include those by Brandt, EmelBus,
and Haszeldine (36) on bis(trifluoromethy1) disulfide and by Haszeldine and
Kidd (14%')on the tri- and tetrasulfides; the former authors have summarized available infrared and ltaman data on disulfides. The infrared and
ltaman spectra of the corresponding trichlorornethyl series have been
recorded by Feh6r arid Uerthold (63).Schotte measured and exhaustively
discussed the infrared spectra of various types of disulfides (199, $00, 201,
STRUCTURES O F SULFUR CHAIN COMPOUNDS
247
203) as well as of diethyl tri- and tetrasulfide, tetrathiodiglycolic acid, and

three l12,3-trithianes (202). Feh6r and Winkhaus (79)recently reported the
Raman spectra of a mono- through tetrasulfide series derived from chloral.
The work of FehBr, Krause, and Vogelbruch (65) on dimethyl and diethyl
polysulfides, in which, beside viscosity and molar volume and refraction,
the Raman spectra also were studied, provides particularly convincing
evidence for unbranched structures in these compounds.
BY ELECTRON
AND X-RAY
B. STRUCTURE
DETERMINATIONS
DIFFRACTION
METHODS
1. DisulJides
The structures of dimethyl disulfide and bis(trifluoromethy1) disulfide

have been determined by electron diffraction, by Stevenson and Beach
(211) and by Bowen (35))respectively. Crystal structure determinations
of the following disulfides have been carried out: di-p-bromophenyl disulfide
by Toussaint (216); N,N'-diglycyl-L-cystine dihydrate by Yakel and
Hughes (232) ; hexagonal L-cystine by Oughton and Harrison (183);
L-cystine hydrochloride by Steinrauf et al. (210); and formamidinium disulfide diiodide and dibromide monohydrates (154).
Nonplanar, unbranched disulfide groups occur in all these compounds.
are, with the exception
The values reported for the length of the S-S
of the rather inaccurate value of 2.14 f 0.08 A in di-p-bIomopheny1
disulfide, remarkably invariant, namely 2.042 A with an average deviation of
0.005 A. The sulfur valency angle lies in the range 99-107", and the dihedral angle between the planes of the valencies of the two sulfur atoms,
between 79" and 105'.
The crystal structures of some cyclic disulfides have also been determined and are commented on in Section XI.
bent
2. Trisulfides
Donohue and Schomaker (48) have determined the structure of dimethyl trisulfide by electron diffraction, and Bowen (35) the structure of
bis(trifluoromethy1) trisulfide. The crystal structure of one representative,
di-2-iodoethyl trisulfide, was first reported by Dawson and Robertson (44)
and the carbon positions later revised by Donohue (46).
The trisulfide groups are unbranched and nonelanar, with reported
S-S bond lengths of 2.04 f 0.02 8, 2.065 f 0.016 A, and 2.05 f 0.04 A,
and S-S-S
angles of 104 f 5', 103.8 f 3", and 113 f 2 O , respectively.
The last angle, in di-2-iodoethyl trisulfide, is exceptionally large.
248
C. Dr-
OLAV FOBS
AND
TRISELENIDES,
DITELLURIDES
The largest number of selenium atoms in covalent selenium chain

compounds isolated so far appears to be three, and even so, the representative compounds known are relatively few in number. In the case of tellurium, the maximum number is two. Iiheinboldt (191) has recently
reviewed preparative methods for di- and triselenides, ditellurides, and
compounds containing mixed sulfur-selenium and sulfur-tellurium chains.
The unbranched structure of diselenides, triselenides, and ditellurides
is firmly established, although less structural work has been done on them
than on disulfides and polysulfides. Rogers and Campbell (192) have studied
the dipole moments of diphenyl, di-p-tolyl, di-p-bromophenyl, and dibenzyl
diselenide, and later (193) those of diethyl and di-t-butyl diselenide. The
moments of diselenides are normally a little lower (by about 0.1 D) than
of the corresponding disulfides. Bergson (25) recently discussed ultraviolet
and infrared spectra of noncyclic and cyclic diselenides; the diselenides
were found to be more suitable than the disulfides for a study in the ultraviolet region, whereas in the infrared region the reverse is true. The same
author (23) has given a preliminary report on the ultraviolet spectra of
two aliphatic ditellurides, and Farrar (61) one on the spectrum of diphenyl
ditelluride. Farrar supposed that ditellurides readily dissociate into free
radicals; however, according to magnetic measurements on diary1 ditellurides at temperatures up to 80" this is not the case (63).
A crystal structure analysis of diphenyl diselenide has been carried out
by Marsh (169) and of the isomorphous di-p-chlorophenyl diselenide and
ditelluride by Kruse, Marsh, and McCullough (157). Bowen (36) in his
electron diffraction work on trifluoromethyl derivatives also studied the
diselenide. The results show nonplanar, unbranched diselenide and ditelluride groups.
The crystal structures of two triselenides, namely cyanogen triselenide
and di-p-toluenesulfonyl triselenide, are known; these are referred to in
other sections.
VI. Sulfur Chains Terminated by Cyano Groups
The first member of the series, cyanogen sulfide S(CN)a, was prepared
by Lassaigne in 1828; the second, thiocyanogen (SCN),, is known from the
classic work of Sijderback published in 1919. The next two homologs,
cyanogen tri- and tetrasulfide, were described in 1922 by Lecher and
Wittwer (162) and Lecher and Goebel (161), respectively; and Feh6r and
Weber (77) in 1958 prepared the penta-, hexa-, hepta-, and octasulfide
and thus brought the series into line with the hydrogen polysulfide and
polysulfur dichloride and dibromide series.
swiucrums
249
OF S U L F U I ~ CHAIN COMPOUNDS
The method for the preparation of cyanogen tri- through octasulfides

is to react mono- through hexasulfur dichlorides with an excess of mercury
thiocyanate suspended in chloroform or carbon disulfide. The compounds
are colorless or faintly yellow crystals, or greenish yellow liquids. The
trisulfide melts and decomposes at 92-94", the tetrasulfide melts a t -3",
and the pentasulfide also below 0", the hexasulfide at 38-39', while the
hepta- and octasulfide form glasses at low temperatures. The trisulfide is
sparingly soluble in most solvents a t room temperature; the others are
more soluble. At temperatures below 0' they can be kept unchanged for
weeks, but at higher temperatures a fairly rapid change to yellow, polymeric
products takes place.
The thiocyanate group has pseudohalogen properties, and the cyanogen
polysulfides may alternatively be termed sulfur thiocyanates. The behavior
of the tri- and tetrasulfides towards nucleophilic reagents (82) is consistent
with the thiocyanate nomenclature.
TABLE I
AXIAL LENGTHS
AND
Compound
S(SCNh
Se(SCN)2
Se(SeCN)2
(1)OF
CYANOGEN 'rRISULFIDE
ISOMORPHOUS
SELENIUM
ANALOGS
Refcrence
10.12
9.87
10.07
12.83
13.03
13.35
4.34
4.44
4.48
181
6
.Q4
The three lowest members of the corresponding selenium series are

known. Cyanogen triselenide, or selenium diselenocyanate Se(SeCN)z, was
prepared by Verneuil ($19) in 1886, by treatment of potassium selenocyanate with oxidizing agents; the yellow crystals melt at 132" and are quite
stable in the dark. Three analogs containing mixed sulfur-selenium chains
have been described, namely diselenium dithiocyanate Se2(SCN)2and the
compound SeS(SCN)2by Baroni (16) in 1936, and selenium dithiocyanate
Se(SCN)*by Ohlberg and vander Meulen (180) in 1953.
The presence of unbranched sulfur chains in the cyanogen polysulfides
is evident from the Raman spectra recorded by Feh6r and Weber (7'7).
The characteristic S-S stretching and S-S-S
bending frequencies occur,
and the C r N frequency varies with increasing chain length in a way
analogous to unbranched methylene a,odinitriles.
The crystal structures of selenium dithiocyanate (181) and selenium
diselenocyanate (6) have been determined by X-ray diffraction. The crystals of these compounds and of sulfur dithiocyanate (94) are isomorphous,
and the three structures are accordingly analogous. The space group is
DzhI6-PrLma with four molecules per unit cell, of dimensions as listed in
Table I. A mirror plane of molecular symmetry is crystallographically
250
OL.4V
FOBS
required; the rnolccules are unbranched and rionplanar and lie across the
mirror plane with only the middle sulfur or selenium atom in the plane.
The dipole moment of sclenium diselenocyanute in benzene a t 25" has
been measured by Rogers and Gross (194). The observed value, 3.98 D,
can he accounted for (194) by assuming free rotation about the Se-Se
bonds, or by postulating a mixture of rotational isomers.
VII. Sulfur Chains Terminated by Sulfonyl Groups
Thcse compounds, disulfonyl mono-, di-, and polysulfidcs, are intermediates between organic mono-, di-, and polysulfides nnd the polythionates, in the sense that they are formally derived from the polythionio acids
by substitution of alkyl or aryl groups for the hydroxyl groups of thc acids.
The monosulfides, like the trithioiiate ion, contain three sulfur titoms in ti
chain, but only the middle one is divalent. There is thus 110 real structural
problem with regard to branching or nonbranching in disulfonyl monosulfides and trithiortates.
A. TYPESOF COMPOUNDS
The first rcpresoritative, di-p-toluenesulfonyl trisulfide, was discovered

by Blomstrand (31) in 1870. On treatment of sodium ~-tolucnethiosulfonntr.
with iodinc in ethanol, he obtained the trisulfide instend of the expected
disulfide. Otto arid Troegcr (182) in 1891 showed that the disulfidc is
initially formed hut rearranges easily into mono- and trisulfide. By the
adion of iodine or chlorine 011 potassium benzene- and p-toluenethiosulfonate they prepared the complete series of mono-, di-, and trisulfides. Troeger
and Hornurig (217 ) extended the henzene- and p-toluenesulfonyl series
to include the tctrasulfides; all members were obtained from mono- or
disulfur dichloride by reaction with an excess of the appropriate sodium
or potassium sulfinate or thiosulfonate in an inert solvent.
The compounds are solids a t room temperature, the melting points being
133', 76-77', 103", and 84-85' for dibenzenesulfonyl mono-, di-, tri-, and
tetrasulfides, respectively, and 138', 114*, 183', and YO" for the p-toluene
compounds. Christiansen (39) observed an unstable form, mp 160", of'
di-p-toluenesulfonyl trisulfide.
The only aliphatic derivatives reported are dimethaiiesulfonyl disulfide
(86) and trisulfide (88), and diethariesulfonyl tri- and tetr:isulfides (82).
The first two are solids with mp 61" and 70', respectively, the last two were
obtained as oils.
Disulfonyl disulfides possess pseudohalogen properties (86), arid the triand tetrasulfides in reactions with nucleophilic reagents behave as sulfur
25 1
thiosulfonates (82).The thiosulfonnte nomenclature is particularly suitable

for some trisulfide analogs, namely selenium and tellurium thiosulfonates,
Se(S202R)2and Te(SzOzR)2.Of these; the methane (88) and benzene and
p-toluene (90) derivatives have been prepared; with one exception the
crystals are isomorphous with those of the corresponding trisulfides.
The selenium analogs, dibenzenesulfonyl and di-p-toluenesulfonyl
mono-, di-, and triselenide, are obtainable (89) from diselenium dichloride
and the sodium sulfinates.
Blomstrand (31) in 1870 formulated di-p-toluenesulfonyl trisulfide with
an unbranched sulfur chain structure, in analogy with the polythionates.
The dipole moments of di-p-toluenesulfonyl mono-, di-, tri-, and tetrasulfide were reported by Christiansen (39) in 1928, but the observed values,
0.75, 1.24, 1.22. and 1.25 D, respectively, gave little information about the
structure. Decisive evidence of unbranched chains in the compounds has
come from X-ray crystallographic work.
B. X-RAYSTUDIES
Unit cells and space groups of disulfonyl mono- and disulfides are given
in Table 11, and of trisulfides in Table 111, together with those of selenium
and tellurium analogs. Data for only one tetrasulfide, the dibenzenesulfonyl
compound (95), are available. This has a large unit cell not favorable for
structure analysis.
1. Monosuljides and Monoselenides
The four derivatives listed (Table 11) are isomorphous. A complete

structure analysis of dibenzenesulfonyl monosulfide has been carried out
by Mathieson and Robertson (172) and of the monoselenide by Furberg
and a y u m (126). The molecules possess twofold symmetry in t,he crystals,
a twofold axis passing through the divalent sulfur or selenium atom. The
bond angles a t these atoms are 106.5' and 105", respectively, and the S-S
and S-Se bond lengths, 2.07 and 2.20 A.
2. Disulfides and Diselenides
The structure of the only aliphatic representative, dimethanesulfonyl

disulfide, was reported by Sorum (207) in 1953. The sulfur chain has an
unbranche!,
nonplanar structure, with S-S bond lengths of 2.10, 2.06,
and 2.10 A ( f 0 . 0 3 A) and S-S-S
bond angles of 104". The SSS/SSS
dihedral angle was found to be approximately 90".
No structure analysis of any aromatic derivative has been made, but the
unbranched structure can hardly be in doubt. The crystals of di-p-toluene-
252
OLAV FOSS
TABLE I1
UNIT CBLLSAND SPACEGHOUIW
O F UISULYONYL
MONO-AND DISULFIDES
AND SELENIUM
ANALOGS*
Compound
Axial lengths
a
b
'
(A)
c
~~
15.90
16.16
5. 53 15.90 113"
5.61 15.86 112"
16.52
16.83
5. 86 18.90 120'
5.77 19.09 119"
5.52 15.78 10.05 97.6"

12.32 6.10 20.75 109"
5.55 14.27 23.22 108"
* Ms
methnnesulfonyl, ns
43
9.98 15.09 11.26

10.17 15.50 11.28
=
Molecular
symmetry Reference
Spacc
group
95"
94O
P21/c
Twofold
''
axis
43
89
None
207
P2Jc 4
P2i/e 4
None
None
93
93
P&/e
None
89
benaencsulfonyl, Ts = p-toluencsulfonyl.
sulfonyl disulfide and diselenide are isomorphous and the structures are
therefore analogous.
No molecular symmetry is crystallographically required in any of the
disiilfonyl disulfides or diselenides. The only possible symmetry element
for a nonplunar disulfide or diselenide group is a twofold axis; a center
would lead to a planar trans structure.
S. TrisulJidesand Triselenides,and Compounds with Mixed Three-Membered
Sulfur-Selenium and Sulfur-Tellurium Chains

These are pentathionic compounds, and have been the subject of more
extensive studies.
The three methanesulfonyl derivatives, sulfur, selenium, and tellurium
dimethnnethiosulfonates, are isomorphous. Structure analyses have been
carried out, starting with tellurium dimethanethiosulfonate and using
tellurium as a heavy atom in the first stages. So far, the structure of the
tellurium compound has been published in detail (l$S),and an electron
density projection along the short crystal axis of the trisulfide (98) has
shown the unbranched sulfur chain structure.
Of the eight aromatic representatives, six crystallize in the same tetragonal trapezohedral space group and are isomorphous. These include all
four p-toluenesulfonyl compounds, namely, the trisulfide, triselenide, and
the two with mixed chains. The only ones which do not belong to the isomorphous series are dibeneenesulfonyl triselenide, which is triclinic, and
tellurium dibenzenethiosulfonate which has one orthorhombic and one
253
STRUC'IUILES OF SULFUR CHAIN COMPOUNDS
TABLE 111
GROUPSOF DISULFONYL
TRISULFIDES
SELENIUM
AND TELLURIUM
ANALOGS*
U N I T CELLS AND SPACE

AND
Axial lengths
a
b
Compound
-~
(A)
c
Space
group
~~
Molecular
2 symmetry Reference
~~
S(SMs)z
SC(SMS)~
Te(SMs)i
S(SBsh
Se(SBs)z
11.35
11.40
11.45
5 . 2 2 16.17 91"
5 24 16.26 91"
5.30 16.35 91"
7.81
7.81
- 26.32 - 26.58 -
14.48 11.20 10.50

'
13.93
SesBss
8.25
S(STS)~
Se(STs)a
Te(STs)2
SeaT~z
7.71
7.71
7.74
7.80
7.71 15.72 06"

8 . 5 5 12.82
-
* Ms = methanesulfonyl, Bs
t Orthorhombic.
29.54 29.65 29.93 29.80 -
P21/n
~~
None
88
p41212
axis
Pbcn
P21/c
{Tr:iinic 2
90
90
axis
None
95
None
90
4s
P41212
Twofold
axis
90
90
90
benaenesulfonyl, Ts = p-toluenesulfonyl.
$ Monoclinic dimorph.
5 LY = 105", B = 95", y = 101'.
monoclinic dimorph. A twofold axis of molecular symmetry is required by

the tetragonal space group and twofold symmetry also by the orthorhombic
one.
The crystal structure of orthorhombic telluriuni dibenzenethiosulfonate
was solved (179) by fairly straightforward use of the heavy atom technique.
The molecular symmetry element is a twofold axis; the other possibility
permitted by the space group, a center, would have led to a tellurium
valency angle of 180. The structure, although orthorhombic, is related to
the tetragonal one and was of substantial aid in a subsequent structure
analysis of tetragonal tellurium di-p-toluenethiosulfonate (109). The
divalent S-Te-S
chains are unbranched, with the sulfonyl groups in
trans positions as demanded by the twofold axis.
An electron density projection (110) of another representative of the
tetragonal series, di-p-toluenesulfonyl triselenide, shows the triselenide
in the trans form, as contrasted with the cis form in cyanogen triselenide.
On the basis of the isomorphism, the same unbranched, trans structures can
be inferred for di-p-toluenesulfonyl trisulfide and selenium di-p-toluenethiosulfonate, as well as for the two tetragonal benzenesulfonyl compounds.
254
OLAV FOSS
Vlll. Sulfur Chains Terminated by Sulfonate Groups: The Polythionates
A. INTHODUCTORY
REMARKS
The history of the polythionates is a long one and fraught with problems
as to structure and reactivity. Dalton in 1808 in his A New System of
Chemical Philosophy commented on the chemical nature of the constituents of the liquid which later came to bear the name of Wackenrodcr. The
three lowest acids were discovered in 1840 to 1846, and Debus in 1888 from
Wackenroders liquid isolated a salt that analyzed for potassium hcxathionate, hut the existence of hexathioriic acid was later doubted, and was
not definitely established until the work of Weitz and Achterberg (220)
in 1928. The literature on polythionates up to 1927 has been reviewed by
Kurtenacker (f58);
a later review is by Goehring (131) in 1952.
Thc polythioiiutes havc no selenium counterparts, containing seleriium
and oxygcri only, in the ions. However, compounds with divalent selenium
chains and terminal sulfonate groups exist, although in accord with the
lower chain-forming capacity of selenium relative to sulfur, the stability
decreases r:qidly with increasing number of selenium atoms in the chtiin.
Sclenotrithionic acid was discovered and isolated as the potassium salt,
K2Se(S03)2,by Ilathke in 1865. Not until quite recently did YanitskiI
and Zelionkaite (256-238) isolate the first salt of diselenotetrathionic acid,
the nionohydrste KzSez(SO& * H20, and of trisclenopentathioriic acid, the
nitron (Nt) salt Ntz HzSen(SO&. The series has thus so far three members,
like the disulfonyl mono-, di-, and triselenide series. The disulf oriyl derivatives appear to be more Fit,able.
Two acids with mixed three-mcnibered chains are known, namely
selenopentathionic and telluropentathioiiic acid. They are derived from
pentuthionic acid by substitution of selenium or tellurium for the middle
sulfur atom of the latter, and behave in reactions with nueleophilic reagents
as thiosulfates of divalent selenium and tellurium. The first salts of sclenopentathionic (84) and telluropentathionic (85) acid were isolated in 1949,
although the formation reactions in aqueous solutions had been known
earlier. Wood (229)has by paper electrophoresis shown that the preparative
methods Icad to pure products.
The polythioriic acids are strong, diprotic acids. The ions decompose in
alkaline solutions and also, more or less readily, in strongly acid solutions,
and free, unsolvated acids had not been prepared until quite recently:
FehBr, Schotten, and Thomas (76) in 1958 announced the solvent-free
synthesis of polythionic acids. Nothing was said about their stability a t
normal temperatures.
Blomstrand (50) in 1869 and one year later Mendelejeff (173) were the
255
first to formulate the polythionates with unbranched sulfur chains terminated by sulfoiiate groups. As remarked in 1927 by Bassett and Durrant
(IY), these formulas were more or less accepted ever since, although it has
been recognized that the evidence in support of them was slender. Bassett
and Durant themselves preferred formulas with branched sulfur chains,
and so, indeed, did many workers, mostly on the ground that the sulfur
atoms of tetra-, penta-, and hexathionate which are so readily given off by
the action of basic reagents like hydroxide, sulfite, and cyanide ions, must
be bonded differently from the others and therefore cannot be part of unbranched chains.
Hertlein (143) in 1896 measured refraction, viscosity, arid electrical
conductivity of aqueous potassium tri-, tetra-, and pentathionates and
decided in favor of unbranched structures, and so did Martin and Metz
(171) in 1924 on the basis of thermochemical measurements. Spa& and
Popper (208) in 1939 reported the refraction of sodium tetrathionate;
Grinberg (136)has later commented on the ionic refraction of tetrathionate
as indicating negative charge on the divalent sulfur atoms. However, from
Hertleins measurements the increments per sulfur atom, from tri- to tetrathionate and tetra- to pentathionate, are about 9 cm3. This is the same as
in other series, for example, in the polysulfur dichlorides (72) where the
sulfur atoms are hardly negative.
The ultraviolet absorption spectra of the tri-, tetra-, and pentathionate
ions were measured by Lorenz and Samuel (167) in 1931 and their results
have been confirmed (9, 150). Andersen and Asmussen (7) studied the
Faraday effect of aqueous potassium tri-, tetra-, and pentathionates, and
concluded that the polythionates behave as a homologous series. Particularly indicative of unbranched structures are the measurements of the
Raman spectra of potassium tri-, tetra-, and pentathionate by Fhcken and
Wagner (49) in 1948, of the K a X-ray fluorescence of sulfur in potassium
tri- through hexathionates by Paessler and Goehring (50) in 1952, and of
the diamagnetic susceptibility of the same salts by Croatto et al. (41) in
1952; a study of the Itaman spectrum of aqueous sodium tetrathionate in
1950, however, did not enable Gerding and Eriks (129) to draw any safe
conclusion with reference to the real configuration of the tetrathionate
ion. Palmer (185) in 1954 in reviewing Paulings General Chemistry
(187)remarked that In presenting structures for the tetra- and hexathionic
ions on p. 368 the author gives no hint that they are controversial, and that
inorganic chemists still eagerly await a settlement by diffraction methods.
Paulings structures, the Blomstrand-Mendelejeff unbranched ones, are
a t present substantiated through X-ray work. Also, the chemistry of the
polythioiiates is fairly well understood (37, 4U, 52, 82, 87, 96) on the basis
of unbranched sulfur chain structures.
256
OT,AV FOSS
U. S T ~ ~ J C
I ~ TTE RU
M I~
N A~
T I OEN S BY X-RAYMETHODS
IJntil 1950 the only X-ray crystallographic work on polythionates was

Zachariasen's structure determination of potassium trithionate (239)
published in 1934, a partial structure analysis of thallium trithionate by
Ketelaar and Sanders (151) in 1936, and unit cell and space group determinations of rubidium trithionate (151) and potassium tetrathionate
(218). Since then, surveys of unit cells and space groups of tetra-, penta-,
arid hexathionates and selenotrithionates, selenopentathionates, and
telluropentathionates have been made, and structure determinations of
two tetrathionates, three pentathionates, one selenopentathionate, two
telluropentathionates, and two hexathionates have been carried out.
The sulfur chains, or in the case of seleno- and telluropentathionates
the SS-Se-S-S
and S-S-Te-S-S
chains, are in all salts unbranched and nonplanar. The tetra-, penta-, and hexathionate ions contain
two, three, and four divalent sulfur atoms, respectively, in a chain between
sulfonate groups; the latter, togcther with the nearest divalent sulfur atom,
constitute thiosulfate groups which have the shapes of distorted tetrahedrons. The tetrathionate ion can alternatively be regarded as built up of
two thiosulfate groups, linked together through the divalent sulfur atoms,
and the penta- and hexathionate ions of one and two divalent sulfur atoms,
respectively, to which are attached two thiosulfate groups.
1. Trithionates and Selenotm'thionates
Potassium trithionate is orthorhornbic bipyramidal, space group

L)zh16-Pnma with four formula units per unit cell. From Zachariasen's
(239) structure analysis, carried out in the early thirties by trial and error
methods, the trithionate ion has mirror plane symrnctry in the crystals,
lying with all three sulfur atoms and two of the oxygen atoms in the mirror
plane. The rubidium salt (151) is isomorphous with the potassium salt. The
third trithionate for which data are available, the thallium salt, is monoclinic prismatic, space group Czl,'-C2/c with four formuIa units per unit
cell (151).Here, the trithionate ion has a twofold axis as symmetry element.
X-ray crystallographic data on potassium, rubidium, cesium, and
ammonium selenotrithionates (108) reveal no isomorphism between any
of these and the above trithionates, although the structures of the triand selenotrithionate ions are undoubtedly analogous. Barium selenotrithioiiate dihydrate is orthorhombic, space group probably D2t,l7-Cmcmwith
eight formula units per unit cell (108). According to Patterson and preliminary Fourier projections the selenium atoms appear to lie in the intersection line of two mirror planes, with the sulfur atoms in one of the planes.
STltUCTUIZES OF SULFUR CHAIN COMPOUNDS
257
The selenotrithionate ion would thus have the symmetry Czn-rnrn in this
salt, but the structural data are incomplete.
2. Tetrathionates
The first structure determinatioii of a tetrathionate, the barium salt,

was published in 1954, 20 years after that of potassium trithionate.
Barium tetrathionate dihydrate is monoclinic prismatic, space group
Czh6-P21/c, with four formula units in a cell of dimensions a = 5.18 8,
b = 9.47 8, c = 19.09 8, p = 96'. The structure analysis (99) led to an
unbranched, nonplanar structure for the sulfur chain of the tetrathionate
ion, and the following bond lengths and angles (=t0.03 A and 5'2'):S1-S2
= 2.10 A, s 2 - s 3 = 2.02 8, &-s4= 2.13 8, L S ~ - S Z - S=~ 104', LS2-S3-S4
= 102'. A value of 90' was derived for the SlSzS3/S2S3S4
dihedral angle.
FIG. 1. The tetrathionate ion as seen along its twofold axis in the sodium salt (102).
These results have later been confirmed through a more accurate structure analysis of sodium tetrathionate dih drate. This salt is monoclinic
c = 5.47 A, p = 105.6', and
sphenoidal, with a = 14.50 A, b = 6.37
two formula units per unit cell. The space group is C2W2, which requires
that the tetrathionate ion lies on a twofold axis of symmetry. The b- and
c-axis projections were solved through Patterson and Fourier syntheses
and refined through difference syntheses. The atomic coordinates (101)
give a length of 2.019 A for the middle S-S bond across the twofold axis,
and 2.116 d for the end S-S bonds, with estimated standard deviations
of 0.01 A (see Fig. 1). There is, thus, no doubt that the middle bond, between the two divalent sulfur atoms of the tetrathionate ion, is shorter
than the two other S-S bonds. The S 4 - S bond angles are 103.8 f 0.5'
and the SSS/SSS dihedral angle is 90.4 f 1".
The sulfonate groups have trigonal symmetry within the accuracy of
the analysis. The axis of the trigonal pyramid does not quite coincide with
the direction of the OsS-S bond, the 0-S-S
angles being 108', 108',
and 99".
The crystals of sodium tetrathionate dihydrate, although monoclinic,
are submicroscopically twinned (91, 93) and appear orthorhombic ; they
258
OLAV E'OSS
have, indeed, earlier been described as orthorhombic. The space group,

C 2 W 2 , is enantiomorphous, and a unit cell contains only one of the enantiomorphous forms of the tetrathionate ion. The twins thus appear to consist of alternating submicroscopic lamellae of right- and left-handed forms.
The isomorphous potassium (218) and rubidium (91) salts have large
cells with two tetrathionate ions in t,he asymmetric unit.
3. Pentuthionates, Selenopentathionates, a n d Telluropentathionates
Unit cells arid space groups of alkali salts and barium salts are listed in
Tables I V and V.
Among the alkali salts, seven hemitrihydrates, namely, three pentathionates, two seleiiopeiitnthionrtes, and two telluropciituthionutcs, arc
isomorphous. The unit cells are large, and the crystal structure is not
known, but the isomorphism shows that the structures of the three anions
TABLE IV
IJNIV CEI,I,SA N I ) S P A C E GROUPSOF ALKALISALTSOF PENTATHIONIC,
SELENOPENTATHIONIC,
A N D TELLUROI'ENTATHIONIC
ACID
8alt,
Axial lerigthti
a
8 . 9 4 22.86
(A)
C
Space
group
9.18
9.27
9.33
9.24
9.25
9.36
9.56
20.21
9 . 2 2 12.26
Z Rcfomrico
5.53
20.12
20.84
20.50
20.87
20.61
21.17
21.79
9 . 6 9 18.03
9.76 18.19
12.32
12.52
12.61
12.50
12.59
12.48
12.93
111
10.3
io3
8
111
103
103
105
103
I
Ort,horhombic 8
105
to3
4
4
Its
106
6.31
6.41
11.22 5 . 5 5 16.67 91"

5 . 0 3 18.90 11.61 100"
P2Jn
P2,lC
91
are analogous. The only salt in Table IV for which a structure analysis has
been carried out is ammonium telluropentathionate (106). The telluropentathionate ion has a trans form in this salt, the sulfonate groups being
located on opposite sides of the S-Te-S
plane (Fig. 5).
The crystal chemistry of the barium salts is interesting. There are five
dihydrates, which crystallize in three different space groups, a triclinic, an
orthorhombic, and a monoclinic one. Two dimorphs of barium pentathionate dihydrate exist, one triclinic and one orthorhombic, and two of
259
STRUCTTJRES O F SULFUR CHAIN COMPOUNDS
barium telluropent,athioiiate dihydrate, one triclinic and w e monoclinic.

The acetone and tetrahydrofuran solvatex represent a fourth, closely related
structure type. Structure determinations of five salts have been made;
namely, of triclinic (115) and orthorhombic (124) barium pentathionate
dihydrate, barium pentathionate hydrate acetonate (119), orthorhombic
barium selenopentathionate dihydrate (114), and monoclinic barium telluropentathionate dihydrate (120).These include two representatives of
the orthorhombic structure type and one of each of the three others.
TABLE V
UNIT CELLSAND SPACEGROUPSOF BARIUM
SALTSOF PENTATHIONIC,
AND TELLUROPENTATHIONIC
ACID*
SELENOPENTATHIONIC,
Salt
Axial lengths
a
b
(A)
c
BaS(Sz0& . 2HZO
BaTe(SzOa)2 '2HzO
5.01 10.37 11.54

4.99 10.57 12.38
BaS(SzO& . 2Hz0
BaSe(SZO&* 2Hz0
5 . 0 1 10.31 21.81
4.99 10.37 22.23
BaS(S20& . H 2 0 . A
BaSe(S20& * HzO . A
BaS(S203)Z . HzO . T
BaSe(S2Oa)z . H 2 0 . T
BaTe(S20r)z . HzO T
5.04
5.02
5.03
5.03
5.00
10.47
10.56
10.56
10.69
10.82
13.61
13.78
13.81
13.97
14.21
Space
Molecular
2 symmetry Reference
group
fl
Pi
None
116
116
Mirror
12.4
11.4
104"
105'
104"
105"
106"
Mirror
plane
118
Mirror
plane
116
180
BaTe(&O& 2H2O
4.99 10.59 23.61 98"
BaSe(S203)z . H20 . +D 5.02 10.59 22.72 102"
BaS(Sz03)2 ' HzO . E

BaSe(SsO& H2O . E
} Pcab
None
117
P2Jc
None
116
BaTo(&03)z . 3H20
9.18 13.27 21.25

9.19 13.30 21.37
11.16
5.24 21.30 107"
A2/m
* In solvates: A
=
#a=
(Y
= acetone, T = tetrahydrofuran, D = dioxane, E = ethanol.

1090, fl = 970, y = 90".
1060, fl = 1010, y = 90".
The penta-, selenopenta-, and telluropentathionate ions have mirror

plane symmetry in the crystals. This is crystallographically required in all
four structure types except the triclinic one, but is realized even there.
The ions thus have a cis form in these mlts, with the terminal sulfonate
groups on the same side of the plane through the three middle atoms. The
occurrence of the cis form in the barium salts is perhaps due to the oxygencoordinating power of the barium ion, and favorable lattice conditions for
the mirror-plane cis arrangement.
The dimensions of the pentathionate ion are within the experimental
260
OLAV FOSS
error thgs:rme in the threc studied stilts (see Fig. 2). The average value for
thc length of the terminal S-S bonds is 2.12 A and of the middle S--8
bonds 2.04 A, with estimated standard deviations of 0.02 8. The sulfur
valency angles are 105' and 106'. In the seleno- and telluropentathionate
ions (Figs. 3-5) the observed values for the terminal S--S bonds range from
2.10 to 2.13 A, with the same average as in the pentathionate ion, 2.12 A.
The crystals of barium penta-, selenopenta-, and telluropentathionate
dihydrates and acetone and tetrahydrofuran solvates are built up of layers
parallel to the c crystal plane, and show a corresponding perfect cleavage
FIQ.2. The pentathionate ion as it occurs in the barium salts (116,119,184). A mirror
plane of symmetry, normal to the plane of the paper, passes through the middle sulfur
atom, and through the selenium and tellurium atom of Figs. 3 and 4 respectively.
along this plane. The thickness of the layers is approximately 11 A in the

dihydrates and approximately 13 &
. iiz the solvates, and is equal to the e
spacing when there are two formula units pcr unit cell, and half the c
spacing whcn there are four. The atomic arrangement within a layer is the
same in all four structure types, and the structures differ only in the way
of packing of the layers. In the triclinic dihydrates, the oblique 8- and
wangles lead to relative displacements of successive layers in the directions
of the a- and b-axes, while in the acctone and tetrahydrofuran solvates,
displacerrierits occur in the a-axis direction only. In the orthorhombic
dihydrates, adjacent layers are related through glide planes in the interface, and in the monoclinic dihydrate, through twofold rotation axes in
the interface.
In the dihydrate structures, one of the water molecules is firmly held
in the lattice, being coordinated to two barium ions and forming hydrogen
bonds to two sulfonate oxygen atoms. The other water molecule is loosely
bound, to one barium ion only, and has a relatively spacious environment;
it is this water molecule which in the solvates is replaced by acetone or
tetrahydrofuran. The oxygen atom of the organic molecule is coordinated
to a barium ion, like the oxygen atom of the replaced water molecule, and
is located between the barium ion and the nearest layer interface, with the
rest of the molecule directed towards the interface. In the dioxane solvate
STHUCTUHES OF S U L F U R CHAIN COMPOUNDS
261
FIG.3. The selenopcntathionatc ion in orthorhombic barium selenopentathionate

dihydratc (114).
FIG.4.Tho cis form of the telluropentathionntc ion as it occurs in monoclinic barium

telluropentathionate dihydratc (120).
FIG.5. The trans form of the telluropentathionate ion as found in the ammonium
salt (106).
262
OLAV FOSS
of Table V, isornorphous with monoclinic barium telluropcntathioritItc

dihydrate, the dioxane moleculc presumably extends across the layer interface and replaces two water molecules, one in each of two adjacent layers,
The layers are thereby held more firmly together, arid the characteristic
cleavage is lacking.
Two ethanol solvates and a trihydrate, conclude Table V. Strontium
sclenopentathionate gives a dihydrutc (95)isomorphous with the two orthorhombic barium dihydrates.
4. Hexathionates
A preparative and X-ray crystallographic survey (111) of salts of hexathionic acid led to unit cell and space group data for two salts, the crystals
of which are, in view of the rather unstable nature and poor crystallization
power of hexathionates, surprisingly well developed and stable. One is a
potassium barium double salt, and the other a cobalt complex salt.
The crystals of potassium barjum hexathionate, KzBa(SsOs)a,are monoclinic prismatic, with Q = 11.58 A, b = 10.81 A, c = 9.14 A, p = 112', and
two formula units per unit cell. The space group is C~h"P2/c, which requires
;B
FIG.6 . Thc cis-cia form of the hexathionate ion as seen along tho c crystal axis of thc
potassium barium salt (108,106).
that the barium ions lie in special, twofold positions. The structure was
solved (102) through projections along the b- and c-axes, by direct Fourier
synthesis based on the strongest reflections with positive signs. The projections have later been refined (106) through difference syntheses.
The sulfur chain of the hexathionate ion is unbranched and nonplanar,
and has in this salt the cis-cis configuration of the Ss ring. The terminal
S-9 bonds arc 2.10 8 arid the three middle S-S bonds 2.04 A, with estimated standard deviations of 0.02 8 (see Fig. 6). The bond ttrigles a t the
divalent sulfur atoms are lolo, 113", log", and loo", and the SSS/SSS
dihedral angles, 108",91", and 105",from one end of the chain to the other.
STRUCTURES OF S U L F U R CHAIN COMPOUXDS
263
The differences hetwceii values for correspondiiig angles in the two halves
of the ion are within the experinieritul errors.
The second salt, trans-dichloro-dien-cobalt(II1) hexathionate monohydrate, is orthorhombic pyramidal, with two f2rmula units in a cell of
dimensions a = 12.12 A, b = 19.13 8, c = 6.43 A. The space group C'22y8Pba2 requires that the hexathionate ions lie on twofold axes of symmetry.
FIQ. 7. Electron density projection along the twofold axis of trans-dichloro-diencobalt (111) hexathionate monohydrate, showing the trans-trans form of the hexathionate
ion. Reproduced, with permission, from Acta Chem. Scand. 13, 201 (1959).
An electron density projection (Fig. 7) along the c-axis (107)shows an unbranched structure for the sulfur chain of the hexathionate ion also in this
salt, but here, the chain has the extended trans-trans form of the helixes of
fibrous sulfur.
c. BONDLENGTHSIN
THE SULFUR C H A I N S
Two types of sulfur-sulfur bonds occur in the polythionates, namely,

between divalent sulfur atoms in the middle of the chains and between one
divalent and one sulfonate sulfur atom a t the ends. The weighted average
of the observed values for the length of the terminal bonds, in nine different
264
OL.4V FOSS
salts including one sebiopentathionate arid two telluropetitathioiiat~s, is

2.11 A with an average deviation of 0.01 A. The middle bonds have within
the errors the same length as the S-S bonds in orthorhombic sulfur, 2.04 A,
which is also the value found for S-S bonds in organic di- and trisulfides.
The difference in length between the two types of bonds indicates, apart
from a possible effect of different hybridization of a-bond orbitals at divalent and sulfonate sulfur, that bonds between divalent sulfur atoms
possess some pd r-bond character, or, what is less probable, that the terminal bonds are longer than single bonds.
D. LIMITOF
THE
POLYTHIONATE
SERIES
Weitz and Achterberg (220) in 1928 remarked that the mother liquor
from the preparation of potassium pentathionate from acidified thiosulfate
with arsenite as a catalyst-the Raschig method-contains appreciable
amounts of polythionates with sulfur content higher than that of hexathionate. Kurtenacker and Matejka (169) in 1936 confirmed that higher
polythionic acids are formed in thiosulfate-arsenite solutions at high concentrations of hydrochloric acid. The highest acid detected was decathionic
acid; a product K2Sz06with 2 about 10 was characterized by X-ray powder
lines.
The extensive studies of Weitz and co-workers on higher polythionates
were published in full in 1956. Weitz and Spohn (224) found that in the
Raschig method for pentathionate, from 30 to 40% of the total yield may
bc obtained as higher potassium polythionates &Sz06 with z from 6 to 12.
They used fractional crystallization of benzidinium, tolidinium, or transdichloro-dien-cobalt (111) salts for analysis of crude products, but did not
succeed in isolating pure salts higher than hexathionate; products with
higher sulfur content were all mixtures. Comparing different methods for
the preparation of potassium hexathionate, Weitz, Becker, and Gieles
(221) stated that only the nitrite-thiosulfate method of Weitz and Achterberg give a pure product and not a mixture of higher and lower homologs.
Thus, from a product prepared according to Goehring and Feldmann (133)
from disulfur dichloride and thiosulfate, fractions with values of z from 4 to
15 were obtained. Feh6r (YO) has later remarked, though, that the WeitzAchterberg and Goehring-Feldrnann methods for hexathionate give identical products.
In 1953-1954 Yanitskii arid Valanchunas (233,$34) reported the isolation of benzidinium octathionate from the reaction between sodium thiosulfate, sodium hydrogen sulfide, sodium hydrogen sulfite, and hydrochloric acid. This method yields, depending on the relative amounts of the
reaction partners, still higher polythionates; thus, Yanitskii and co-workers
265
(235) recently reproduord microphotogr:Lphs of crystalline nitrou (Nt)

salts Ntz H2X,06with z = 13and 18 and dimethylphenylbenzylammonium
salts (ClgH18N)&&,with z = 6, 9, 12, and 13. In view of the experience
of Weitz and co-workers with regard to fractional crystallization of higher
polythionates, one cannot feel sure that the products represent pure individuals and not mixtures of homologs. The ordinary analytical methods
give the average sulfur content only, and the only safe way t o ascertain
the purity of a higher polythionate appears to be through chromatography
or electrophoresis. Pollard et al. (189) and Bighi and Trabanelli (27) have
worked out methods for paper chromatography, and Wood (228) a technique for paper electrophoresis.
Schmidt (197) in 1956 reported the preparation of a n ether solvate of
heptathionic acid from hydrogen pentasulfide and sulfur trioxide, and of
solutions and potassium salts of octa-, deca-, and dodecathionic acids. Feh6r
et al. (75) in a preliminary report have described the use of condensation
reactions between chlorosulfonic acid and higher hydrogen polysulfides,
and between thiosulfuric acid and polysulfur dichlorides. They report the
synthesis of free octa-, nona-, deca-, undeca-, and dodecathionic acids, and
reproduce paper chromatograms of, among other samples, one of pure
potassium octathionate.
The results of Weite and co-workers (222, 223) indicate that the polythionate series extends up to and including the hydrophilic OdBns sulfur
sols. These appear to be sodium polythionates Na2S,0a with x from 50 to
100. The stability of the polythionates, with respect to liberation of sulfur,
decreases monotonously up to x about 20 and then increases as the properties approach those of the sulfur sols.
IX. Unsymmetrical Compounds
In the compounds considered so far the sulfur chains are terminated

by the same kind of atom or group a t both ends. However, mixed organic
disulfides and also other types of unsymmetrical di- and polysulfides exist.
with It = o-nitrophenyl and
Organic disulfur chlorides, RS-S-C1
9-anthryl, were described as early as 1922. Harris (140) has recently studied
properties and reactions of the o-nitrophenyl compound, and Himel and
Edmonds (144) in patents described reactions of the t-butyl analog; other
aliphatic disulfur chlorides were reported by Bohme and van Ham (33)
and Feh6r and Kruse (66) in 1958. With thiols, thioacetate, xanthates, and
dithiocarbamates these compounds give unsymmetrical trisulfides. Bohme
and Clement ($2) in 1952 prepared acyl-disulfur and acyl-polysulfur chlorides, arid derivatives such as an unsymmetrical pentasulfide.
were reported in 1954 by
Alkyl-hydrogen di- and trisulfides R-S,-H
266
OLAV FOSS
Bohme and Zinner (Y4), who prepared the methyl, ethyl, and henzyl
derivativcs. With iodine, the trisulfides give symmetrical hexttsulfides.
Feh& and Kruse (66) prepared phcnyl-hydrogen tri- and tetrasulfide and
n-butyl-hydrogen tetrasulfide and studied their Raman spectra. The compounds are sensitive to bases, and salts are apparently unstable, although
White (226) in 1918 reported a stable sodium salt of 2-naphthyl-hydrogen
disulfide.
Further examples include some in which the sulfur chains carry an alkyl
or aryl group, or a hydrogen atom, at one end and a sulfonate group a t
have been studied in soluthc other. Sulfenyl thiosulfates R-S-S-SO3was recently identified by
tions (83),and a trisulfide R-S--S-S-SO,Szczepkowski (212) as a minor product from the reaction of tetrathionate
with cysteine. The trisulfide is evidently formed through a nucleophilic
displacement of a sulfite group of tetrathionate, a less predomimnt reaction
were first
path (96). So-called sulfane-monosulfonic acids, H--S,--SOsH,
prepared by Schmidt (196) from hydrogen di- and polysulfides and sulfur
trioxide, and later by Feh6r and co-workers (75) by use of chlorosulfonic
with n = 8
acid instead of sulfur trioxide. The ionized species, -&--SOathrough 2, have been postulated (87) as intermediates in the reaction of
sulfur Se with sulfite ion to give thiosulfete, and analogous unsymmetrical
intermediates, 3,-CN arid -Sn--P(C6H6)3, occur in the reactions of sulfur

S, with vyanide ion (17, 87) and with triphcriylphosphine (18).
X. Comments on Structural Evidence
Structure determinations by X-ray and electron diffraction methods

have demonstrated the presence of unbranched chains of up to six divalent,
sulfur atoms in inorganic polysulfides, four divalent ones in salts of polythionic acids, and three in disulfonyl, dialkyl, and cyanogen derivatives.
The synthetic arid Raman-spectroscopic work of Feh6r has established the
presence of unbranched structures with up to five sulfur atoms in organic
polysulfides, and eight in the cyanogen polysulfide, polysulfur dichloride,
and hydrogen polysulfide series. Tho results are supported by extensive
physicochemical evidence of less conclusive nature.
There exist in the literature claims concerning the occurrence of branching in particular compounds; the more definite of these have been refuted
in recent years. Backer and ao-workers (10, 11) formulated some cyclic
di- and trisulfides with branched sulfur groupings; from a comparison of
their ultraviolet and infrared spectra with those of other di- and polysulfides, Schotte (202) concluded that branching does not occur. Baroni
(15), on the basis of parachor measurements, assigned isomeric branched
STRUCTURES OF SIJLFUR CHAIN COMPOUNDS
267
structures to two diethyl pentasulfides; this claim has been refuted by

Feh6r et al. (66),who were able to prepare only one diethyl pentasulfide the
Raman spectrum of which agreed with an unbranched structure. The supposed low-melting isomer of di-2-naphthyl diselenide, to which a coordinated structure was ascribed by Loevenich el al. (166) was, according to
Bergson (99),
probably a mixture of monoselenide and normal, unbranched
diselenide.
In fact, no substantiated example of branching is known. Feh6r (p. 313
in ref. 56) states that "We have tried for many years to prepare branchedchain polysulfides, but without success."
The only instances in which sulfur forms only one bond to sulfur and
none to another atom, are apparently in the polysulfide ions as terminal
atoms, and in the thiosulfate and thiosulfonate ions, S-SO,-and
S-S0,R-,
in which the S--S bond appears to be partly double. Thus,
crystal structure determinations of sodium thiosulfate pentahydrate (213)
and sodium methanethiosulfonate monohydrate (100) have shown the S-S
bond to be 1.97-1.98 A, as compared with 2.04 1 in orthorhombic sulfur.
The reason (87) why coordinated structures are unstable relative to
unbranched ones may be that a sulfur atom as an electroneutral, nonpolar
acceptor can add only to donors which are less electronegative, or to donors
which can expand the valency shell for *-bonding.
XI. Stereochemistry
Fairly accurate values for bond lengths and angles in sulfur chains are
known from X-ray and electron diffraction work; some have been commented on in earlier sections. A recent review is by Abrahams (3) in 1956.
A. NONPLANARITY
OF SULFURCHAINS
In the noncyclic compounds for which structural data are available,
a dihedral angle of about 90" occurs between the planes of the valencies of
adjacent sulfur atoms. Reported values lie between 74" in cesium hexasulfide (3) and 108" in barium pentathionates (119). In six- and particularly in five-membered cyclic disulfides, the angle is smaller, and may for
steric reasons also become larger, but planar, trans groups with 180"
dihedral angle have not been observed.
1. Origin and Magnitude of the Barrier to Internal Rotation
The barrier to rotation about a bond between two divalent sulfur atoms
is thought to arise principally from the mutual interaction of the unshared
pairs of 3pr-electrons on the sulfur atoms. This explanation was first put
268
OLAV FOSS
forward by Penney and Sutherland (188) in 1934 to account for the nonplanarity of hydrogen peroxide and hydrazine, and was discussed with
reference to elemental sulfur and sulfur compounds by Pauling (186) in
1949. The cr-bonding orbitals of sulfur are here assumed to be nearly pure
p in character, and the second lone pair on each sulfur atom to occupy
mainly the %-orbital. The interaction between the two unshared pairs of
3p~electronson adjacent sulfur atoms has a maximum when the orbitals
are parallel (dihedral angle cp = 0" or 180") and a minimum when they are
orthogonal (cp = 90"). This leads to a simple twofold rotation barrier.
Important information concerning the 3pAnteraction in disulfidcs has
recently come from studies of ultraviolet spectra. As pointed out by Barltrop et al. (14) and Schotte (601, 203) the absorption peak near 2500 8
displayed by normal noncyclic disulfides is in five-membered cyclic disulfides shifted towards longer wave lengths, to about 3300 A. According to
molecular orbital calculations by Bergson (2420, the red shift is due to a
decrease in excitation energy of 3p~electronswith decreasing dihedral
angle. The 3pr-interaction satisfactorily accounts for a dihedral angle of
90" as the most stable conformation of a disulfide in the ground state, but
is not a simple Coulombic repulsion between charge clouds; in order to
calculate the torsion energy from the red shift it is necessary to know both
the overlap integral of the 3p-electrons and the dihedral angle. For the
barrier height, that is, the torsion energy when cp = 0", in a five-membered
cyclic disulfide, 1,2-dithiolane-4-carboxylic acid, Bergson (65) arrived at a
value of 14 kcal/mole.
Two other approaches beside Bergson's have been used to derive torsion
barriers for S-S bonds; namely, from the torsional frequency of the bond
as observed in Raman spectra, and from measurcd thermodynamic data
as compared with statistically calculated ones. According to Scott and coworkers the torsional frequencies in dimethyl disulfide (bod), diethyl
disulfide (205), and disulfur dichloride (139) correspond to barrier heights
of 9.5, 13.2, and 14.2 kcal/mole, respectively. On the other hand, agreement between calculated and observed entropy arid heat capacity of
dimethyl disulfide was obtained (146) by use of 6.8 kcal/mole for the effective barrier height. From observed and calculated heat capacities, Fehdr
and Schulze-Rettmer (7%) arrived at a value of 2.7 kcal/mole for the barrier
height in hydrogen disulfide.
2. Rotational Isomerism
The nonplanarity of sulfur chains gives rise to rotational isomerism,
namely, to two enantiomorphs of a disulfide, two forms of a trisulfide of
which one is a pair of enantiomorphs, and to three pairs of enantiomorphs
of a tetrasulfide. In gases and solutions, equilibrium mixtures probably
269
oc(:ur, whereas in crystals, with fixed atomic positions, only oric of the forms
will normally be present. Depending on the symmetry of the crystal and
the number of molecules per unit cell, the cell contains both or only one of
a pair of enantiomorphs.
a. Disuljdes, tetrathiomtes. The nonplanarity of the valencies of two
bonded sulfur atoms was first suggested by Feh& and Baudler (67) for
hydrogen disulfide from analogy with hydrogen peroxide. The exact value
of the dihedral angle in hydrogen disulfide is uncertain (227). It is 101" in
N,N'-diglycyl-L-cystine dihydrate (%%?),79" in L-cystine hydrochloride
(210), 105' in formamidinium disulfide diiodide monohydrate and 89" in
the corresponding dibromide (1O4), and 76" in barium tetrasulfide monohydrate (1). The average of these values is exactly 90, and this is the angle
found in barium tetrathionate dihydrate (99) and sodium tetrathionate
dihydrate (101).
Rogers and Campbell (193) have concluded from dipole moment measurements that the dihedral angle in di-2-butyl disulfide is larger than normal
owing to spatial interference of the bulky t-butyl groups. This was first
discussed by Koch (162) on the basis of ultraviolet absorption; the characteristic peak of normal disulfides a t 2500 is for di-t-butyl disulfide shifted
towards so low: wave length that it is overlapped by the strong absorption
at about 2000 A. However, as pointed out by Bergson (21, 25), the blue
shift is not a simple steric effect but principally an electronic (hyperconjugative and inductive) effect. Schotte (199)has demonstrated the same spectral effect for disulfides with carboxylic group substituents in a-positions
to the disulfide bond.
The enantiomorphs of a nonplansr disulfide sometimes become separated in crystals. Thus, diphenyl disulfide (43)and probably also di-p-tolyl
disulfide (43) crystallize in enantiomorphous space groups with only one
enantiomorph per unit cell, and so do barium tetrasulfide monohydrate (1)
and sodium tetrathionate dihydrate (101). The same applies, of course, to
L-cystine and other optically active disulfides.
b. Trisulfides, pentathionates. Two rotational-isomeric forms of trioccur. The atoms or groups X are rotated about
sulfides X-S-S-S-X
90" out of the plane of the three sulfur atoms, either to the same side of the
plane-&$-or
to opposite sides-trans. There are two enantiomorphous
trans forms. This holds also if X is sulfur; the cis form of the then fivemembered sulfur chain is a segment of the Ss ring of orthorhombic sulfur,
while the trans forms are present in the helixes of fibrous sulfur. A cis form
possesses a mirror plane of symmetry, and a trans form a twofold axis, provided that the two atoms or groups X are like, and that bond lengths and
angles are the same in both halves of the molecule.
In the crystals of di-2-iodoethyl trisulfide (46) the configuration about
270
OLAV FOSS
the trisulfide group is trarbs, a twofold axis of molecular synmetry being

crystallographically required. The trans form occurs also in the crystals
of dibenzenesulfonyl and di-p-toluenesulfonyl trisulfide ; here also the
molecules are required to lie on twofold axes (43). These three trisulfides
crystallize in the enantiomorphous space group, D4pP4,2,2 (or the enantiomorph) with four molecules per unit cell. A unit cell, and whole single crystals provided twinning does not occur, thus contains only one of the trans
enantiomorphs. The configuration of dimnethanesulfonyl trisulfide is also
trans (98); here, the space group is centrosymmetric and both enantiomorphs are present in each unit cell.
The cis form occurs in cyanogen trisulfide, where the molecules lie
across mirror planes of symmetry (94). The trisulfide group of the pentathionate ion occurs in the cis form in the three barium salts (115, 119, 124)
for which structure determinations have been carried out, a mirror plane of
symmetry being demanded by the space group in two of the salts. In the
six- and seven-membered rings of 1,2,3-trithianes (202) and the one known
l12,3,5,6-pentathiepane (125) the configuration of the trisulfide group is
necessarily cis.
The existence of the two forms, cis and trans, of a trisulfide was apparently first considered by Baudler (20) with reference to the Raman spectrum
of hydrogen trisulfide, and later by Kushner et al. (160) and Woodrow
el al. (230)on the basis of dipole moments of dialkyl trisulfides in solutions.
The author (92) has discussed the occurrence of the forms in crystals.
c. ~ e t r a ~ l ~ d hexathionates.
es,
In compounds X-S-S-S-S-X,
three different rotational-isomeric forms, each consisting of a pair of
enantiomorphs, are possible, as was pointed out by Woodrow et al. (230)
in a discussion of the dipole moment of di-n-hexadecyl tetrasulfide. In the
six-membered chain of the four sulfur atoms and their terminal bond partners, the fifth atom may be located cis or trans relative to the first, and the
sixth similarly relative to the second. The three forms which thus arise
are, one cis-cis in which the fifth and sixth atom are in cis positions relative
to the first and second, respectively; one cis-trans, and one trans-trans. All
three forms may be right- or left-handed; a cis-trans form is the mirror
image of a trans-cis. A cis-cis form passes into a trans-trans, and vice versa,
through a rotation of approximately 180" about the middle sulfur-sulfur
bond. The cis-cis form is present in the Ss ring, and the trans-trans in the
helixes of fibrous sulfur; these two forms possess a twofold axis of symmetry
provided that bond lengths and angles are equal in the two halves.
The trans-trans forms occur in the crystals of cesium hcxasulfide (5) and
of transdiehlorodien-cobalt (111) hexathionate monohydrate (lor), and
the cis-cis forms in the crystals of potassium barium hexathionate (102),
these being the only compounds containing tetrasulfide groups for which
structurtil data are available.
271
HTltUCTUHES OF SULFUR CHAIN COMPOUNDS
B. CYCLICDISULFIDES
When a disulfide group is part of a ring of small size, the group cannot,
owing t o the geometrical requirements of the ring, maintain the normal
dihedral angle of about 90". Thus, according to crystal structure determinations, the dihedral angle is 60" in the six-membered ring of racem-1,2dithiane-3,6dicarboxylic acid (112), and 27" in the five-membered ring of
l12-dithiolane-4-carboxylicacid (122). With a simple cosine-type barrier
to rotation about the S-S bond, the latter angle corresponds to a S-S
torsion energy equal to 80% of the barrier height. Beside torsion, there is
also stretching and bending strain in the rings; Bergson and Schotte (26)
have, on the basis of structural data, carried out a conformational analysis
of the rings of racem-l12dithiane-3,6-dicarboxylic
acid and 172-dithiolane4-carboxylic acid and concluded that the total strain is a t least 5 kcal/mole
in the l12-dithiane ring and 16 kcal/mole in the dithiolane ring. The large
strain in the latter is in accordance with the highly unstable nature (14) of
l12-dithiolane itself; in derivatives, the substituents stabilize the ring to
some extent.
There exist some unsaturated five-membered cyclic disulfides which
are remarkably stable, among them the 1,2dithia-4-cyclopentene-3-thiones
and 3,5-diimino-l,2,4dithiazolidines,
the so-called trithiones and thiurets,
respectively. Here, the S-S Spa-interaction is either absent, or compensated by conjugative stabilization of the rings. Kehl and Jeffrey (149) have
I
s=c /
\
S--S
\C-H
/
(1)
+ / \
HzN=C
C-NHI
84
(11)
made a crystal structure analysis of 4-methyl-trithione1 (I); the molecule

was found to be planar within the experimental accuracy. I n the crystals of
thiuret hydroiodide, the cation of which (11) is a diamino and not a diimino
derivative, the disulfide group lies across a mirror plane and is thus exactly
planar (121);the cation as a whole is planar or nearly so.
C. SELENIUM
AND TELLURIUM
ANALOGS
The same barrier t o rotation about selenium-selenium and telluriumtellurium bonds exists as in the case of sulfur. Bergson (25) has recently
extended his discussion of p7r-interaction in disulfides (24) t o diselenides.
There have not been so many determinations of dihedral angles in
selenium and tellurium compounds, but the available data show angles of
272
OLAV FOYY
about 90". Thus, the reported value for the dihedral angle in diphenyl
diselenide (169)is 82", in di-p-chlorophenyl diselenide arid djtelluride (167)
74" and 72", respectivcly, and in cyanogen triselenide (6) 94". In a-monouliriic (38) and 0-monoclinic (170) selenium the angles are 102" and in
hexagonal selenium and tellurium (S), 101". Dihedral angles dependent on
rotation about sulfur-selenium and sulfur-tellurium bonds have becri
reported for selenium dithiocyanate (181) and barium selenopentathionate
dihydrate (114) and for several telluropentathionic compounds. In the
first two the angles are 79' and 109', respectively, and in the latter (106,
109, 120, 123, 179) the values range from 79" to 103".
Planar trans disclcriide or ditelluride groups, with 180' dihedral angles,
have not been encountered. The smallest S e - S e dihedral angle determined
is 56" in the six-membered ring of racem-l,2diselenane-3,6dicarboxylic
acid (106, 113). In saturated five-membered cyclic diselenides the dihedral
angle is, for gcorrietrical reasons, necessarily smaller, and in unsaturated
oms probably zero, but no structure determinations have been carried out.
l'riselenides, and compounds with mixed three-membcred chains like
seleno- and telluropeiitiitliioncLtcfi, display the same rotational isomerism
as trisulfides. Examples already referred to include the cis triselenide
group of cyanogen triselenide arid the trans one of di-p-toluenesulf onyl
triselenide, and likewise, the cis telluropentathioriate ion in the barium
salts as contrasted with the lrans form occurring in the ammonium salt.
D. PEROXY
COMPO~JNDS
These are, with one notable exception, nonplanttr like disulfides,

disclenides, and ditellurides; indccd, the nonplanarity of hydrogen peroxide
was the first example discussed, by Penney and Sutherland (188) in 1934.
Incidentally, here pd *-bonding can make no contribution to the nonplanarity, as has been suggested by Krebs (154) for sulfur and sulfur compounds.
Rain and Giguhrc ( I S ) have recently discussed dihedral angle and barrier
height in hydrogen peroxide on the basis of infrared and other data. From
the crystal structure analysis of hydrogen peroxide by Abrahams et aZ. (4),
the dihedral angle is 94 =t 1.5" at -20". Harvey and Bauer (141) in an
electron-diffraction study of bis(pentafluorosulfur) dioxide found a dihedral
angle of 107 f 5". Dialkyl and diacyl peroxides are also nonplnnar, according to dipole moment measurements by Lobunez, Rittenhouse, and Miller
(165). The dipole moment of di-2-butyl peroxide indicates a larger than
normal dihedral angle, namely about 125', owing to the spatial requirements of the 2-butyl groups, as earlier deduced on the same basis by Rogers
arid Campbell (19s).
It is remarkable, then, that in the peroxydisulfate ion the peroxy group
273
is planar truns. The crystal structures of ammonium and cesiuni peroxydisulfate were determined as early as 1934 by Zachariaseri and Mooney
(240) and that of the potassium salt one year later by Keen (148). The
latter is triclinic, while the ammonium and cesium salts are monoclinic and
isomorphous; in both space groups a center of symmetry is required for
the peroxydisulfate ion. As stated earlier, no such structure has been found
for disulfide, diselenide, or ditelluride groups.
A planar trans structure has recently been reported for a hydrazine
derivative. Whereas hydrazirie itself is noriplanar owing to interaction between pa-electron pairs as in compounds of sixth-group elements, Tomiie,
Koo, and Nitta (215) found that diformylhydrazine is centrosymmetric in
the crystalline state and its hydrazine group thus planar. Tomiie (214) has
subsequently performed a molecular orbital calculation of the electronic
structure of diformylhydrazine, and concluded that Coulombic repulsion
between charged atoms in different halves of the molccule is the primary
cause of the planar structure. A similar explanation may apply in the case
of the peroxydisulfate ion. Also, different hydridization at the peroxy
oxygen atoms, or r-bonding between these and sulfur, may lower the
plr-interaction and contribute to the planarity.
REFERENCES
1 . Abrahams, S. C., Acta Cryst. 7, 423 (1954).
2. Abrahams, S. C., Acta Cryst. 8, 661 (1955).
3. Abrahams, S. C., Quart. Revs. 10, 407 (195ti).
4. Abrahams, S. C., Collin, R. L., and Lipscomb, W. N., Acta Cryst. 4, 15 (1951).
5. Abrahams, S. C., and Grison, E., Acta Cryst. 6,206 (1953).
6. Aksnes, O., and FOSS,
O., Acta Chem. Scand. 8, 1787 (1954).
Y. Andersen, E. B., and Asmussen, R. W., J. Phys. Chem. 36, 2827 (1932).
8. Amtson, R.H., Dickson, F. W., and Tuncll, G., Science 128, 716 (1958).
9. Awtrey, A. D., and Connick, R. E., J. Am. Chem. SOC.73, 1842 (1951).
10. Backer, H. J., and Evenhuis, N., Rec. trav. ci~im.66, 129, 174 (1937).
11. Backer, H. J., and Tamsma, A. F., Rec. trav. chim. 67, 1183 (1938).
12. Baer, J. E., and Carmack, M., J . Am. Chem. SOC.71, 1215 (1949).
13. Bain, O., and Gigubre, P. A., Can. J . Chem. 33, 527 (1955).
14. Barltrop, J. A., Hayes, P. M., and Calvin, M., J . Am. Chem. SOC.76,4348 (1954).
15. Baroni, A., Atti reale accad. naz. Lincei [6] 14, 28 (1931).
16. Baroni, A., Atti reale accad. naz. Lincei [S] 23, 139 (1936).
17. Bartlett, P. D., and Davis, R. E., J. Am. Chem. SOC.80, 2513 (1958).
18. Bartlett, P. D., and Meguerian, G., J. Am. Chem. SOC.78, 3710 (1956).
19. Bassett, B., and Durrant, R. G., J . Chem. SOC.p. 1401 (1927).
$0. Baudler, M., Thesis, University of Gottingcn, p. 89 (1946), cited in Gmelins
Handb. unorg. Chem. 9/Bl, 150 (1953).
21. Bergson, G., Arkiv Kemi 9, 121 (1956).
23. Bergson, G., Acta Chem. Scand. 11, 571 (1957).
274
OLAV FOSH
25. Ibrgson, G.,Arhiu Kemi lS, 11 (1!%8).

26. Bergsoti, G., and Schotte, L.,Arkiv Kemi 13, 44 (1958).
27. Bighi, C., and Trabanelli, G., Ann. cliim. (Rome) 46, 1186 (1955).
28. Blake, E. 8., J . A m . Chem. Soc. 66, 1267 (1943).
29. Rloch, I., and Hohn, F., Ber. 41, 1971 (1908).
30. 13loinstrand, C. W., Cliemie der Jftztzeit p. 157 (1869).
31. Blomstrand, C. W., Ber. 3, 957 (1870).
<?2.Biihmne, H., and Clement, M., Ann. 676, 61 (1952).
33. Bohmc, H., and van Ham, G., Ann. 617, V2 (1958).
34. Hiihme, H., and Zinner, G., Ann. 686, 142 (1954).
35. Rowen, H. J. M., Trans. Faruduy SOC.60, 1 (1954).
J6. Brandt, G. R. A,, Emelkus, I. J., and Hmzeldine, R. N., J . Chem. Soc. p. 2549
(1952).
37. Brodskir, A. I . , i n "Radioisotopes in Scientific Research (R. C. Extcrmunn, d.),
Vol. 2, p. 41. 1t:rganion Press, London, 1958.
.?8. Burbttnk, R. D., Acta Cryst. 4, I40 (1951).
3.9. Christinnscn, .J. A.,2.Eleklrochern. 34, 638 (19%).
40. Christisiuon, J. A., and Drost-Hansen, W., Nature 164, 759 (1949).
41. Croatt,o, [J., knva, A., aud Scatturin, V., Gazz. chirn. ital. 82, 552 (1952).
42. DM, S. R., Colloq. Intern. CJnion Pure and A p p l . Chem., Mdnsler, 1954, 11. 103
( I 95.5).
43. Dawson, I. M., Mathicson, A. McL., aid Robertson, J. M., J . Chem. SOC.p. 322
( 1948).
44. Dawson, 1. M., and Robertson, J. M., J . ClwttL. Soc. p, 1256 (1948).
4n. De H a m , Y. M., Physica 24, 855 (1958).
40. Donohue, J., J . Am. Chem. SOC.72, 276 (1950).
47. Ihnohuc, J., Carson, A., and Goldish, E., Nature 182, 518 (1958).
48. Donohue, J., and Schornakcr, V., J . Chem. Phys. 16, 92 (194X).
4!!. Eucken, M., and Wagner, J., Acta Plry~.AusLriaca 1, 339 (1948).
60. Faessler, A,, and Goehring, M., Naturwissenschaftm 3@,169 (1952).
61. Iaarrar, W. V., ltesearch 4, 177 (1951).
66. Fava, A., (Jazz. chim. ital. 83, 87 (1953).
63. Iqava, A., private communication (1954).
64. Fava, A., and Iliceto, A., Ann. clritn. (Rowe) 43, ,509 (1053).
66. Fch6r, Y., Anyew. Chern. 67, 337 (195.5).
66. FehBr, F., Chemical Society Symposia, Brist)ol, 1958, p. 305. Cliern. Soc. (London)
S p e c . lubl. NO.12 (1958).
67. FehBr, F., and Baudler, M., 2. Elektrochem. 47, 844 (1941).
158.Fchgr, F., and Raudlcr, M. , 2. anarg. u. atlqem. Clram. 264, 2.51 (1947).
69. FehBr, F., and Baudler, M., 2. unorg. u. allgem. Cltum. 264, 289 (1947).
60. Fchbr, F., and Baudler, M.,
2. unorg. u. allgem. Chem. 268, 132 (1949).
61. Fehkr, F., and Berthold, H. J., 2. anorg. u. allgem. Chem. 273, 144 (1953).
6%. Fehbr, F., and Berthold, H. ,J., 2.anorg. u. allgem. Chem. 274, 223 (1953).
63. Fchkr, F., and Berthold, H. J., Chem. Ber. 88, 1634 (1965).
64. Fehik, F., and Berthold, R., Z . anorg. u.al2gem. C l ~ s n 290,
.
251 (1957).
66. Fehdr, F., Kraise, G., and Vogelbruch, K., Chem. Ber. 90, 1570 (1957).
66. Eeher, F., and Kruse, W.,
Chem. Ber. 91, 2528 (1958).
67. FchBr, F., Kruse, W., and Lauc, W., 2. anorg. u.allgem,. Chem. 292, 203 (1957).
68. FehtthBr, F., Laue, W., and Winkhaus, G., 2. anorg. u. allqem. Chem. 288, 113 (19.56).
60. EchEr, F., Laue, W., and Winkhaus, G., 2.anorg. u. allyem. Cham. 290, 52 (1957).
STHUCTURES O F SULFUIt CHAIN COMPOUNDS
275
70. FehBr, F., Meyer, L., Nauscd, K., Ristid, S., and Weber, H., Angew. Chem. 69,392
(1957).
71. FehBr, F., and Naused, K., 2. anorg. u. allgem. Chem. 283, 79 (1956).
72. FehBr, F., Naused, K., and Weber, H., 2. anorg. u. allgem. Chem. 290, 303 (1957).
73. FehBr, F., and RistiE, S., 2.anorg. u. allgem. Chem. 293, 307 (1958).
74. FehBr, F., and Risti6, S., Z. anorg. u. allgem. Chem. 293, 311 (1958).
75. FehBr, F., Schotten, J., and Thomas, B., Z. Naturforsch. 13b, 624 (1958).
76. FehBr, F., and Schulee-Rettmer, R., Z. anorg. u. allgem. Chem. 296, 262 (1958).
77. FehBr, F., and Weber, H., Chem. Ber. 91, 642 (1958).
78. FehBr, F., and Winkhaus, G., Z. anorg. u. allgem. Chem. 288, 123 (1956).
79. FehBr, F., and Winkhaus, G., Chem. Ber. 91,217 (1958).
80. Foppl, H., 2. anorg. u. allgem. Chem. 291, 12 (1957).
81. Foppl, H., Angew. Chem. 70, 401 (1958).
82. Foss, O., Kgl. Norske Videnskab. Selskabs Skrifter 1946, No. 2 (1947).
83. FOSS,
O., Acta Chem. Scand. 1, 307 (1947).
84. Foss, O., Acta Chem. Scand. 3, 435 (1949).
8Y. FOSS,
O., Acta Chem. Scand. 4, 404 (1950).
88. Foss, O.,Acta Chem. Scand. 6, 115 (1951).
89. FOSS,O., Acta Chem. Scand. 6, 508 (1952).
90. FOSS,
O., Acta Chem. Scand. 6, 521 (1952).
91. FOSS,
O., Acta Chem. Scand. 6, 802 (1952).
9.2. Foss, O., Acta Chem. Scand. 7, 1221 (1953).
BS. Foss, O., Acta Chem. Scand. 8, 469 (1954).
94. Foss, O., Acta Chem. Scand. 10,136 (1956).
97. Poss, O., in Organic Sulfur Compounds (N. Kharasch, ed.), Vol. 1. Pergamon
Press, London, 1960.
,98. Foss, O., Furberg, S., and Hadler, E., Acta Chem. S f a n d . 6, 1417 (1951).
99. Foss, O., Furberg, S., and Zaehariasen, H., A c f a Chem. Scand. 8, 459 (1954).
100. Foss, O., and Hordvik, A., Acta Chem. Scand. 11, 1443 (1957).
101. Foss, O., and Hordvik, -4.,
Acta Chem. Scand. 12, 1700 (1958).
10.2. Foss, O., Hordvik, A., and Palmork, K. H., A d a Chem. Scand. 12, 1339 (1958).
103. FOSS,
O., and Jahr, J., A c h Chem. Seand. 4, 1560 (1950).
104. Foss, O., Johnsen, J., and Tvedten, O., Actn Chern. Scand. 12, 1782 (1958).
105. Foss, O., and Johnsen, J., unpublished work (1958).
106. Foss, 0 , and Larssen, P. .4.,Acta Chem. Scund. 8, 1042 (1954).
107. FOSS,
O., and Maroy, K., A c f a Chem. Scand. 13, 201 (1959).
108. Foss, O., and Morch, 0. H., Acta Chem. Scand. 8, 1169 (1954).
109. Foss, O., and i)yum, P., Acta Chem. Scand. 9, 1014 (1955).
110. Foss, O., and i)yum, P., unpublished work (1955).
111. Foss, O., and Palmork, I<. H., Actn Chem. Scund. 12, 1337 (1958).
11.2. FOSS,
O., and Reistad, T., Acta Chem. Scand. 11, 1427 (1957); and unpublished
work (1957).
113. FOSS,O., and Schotte, L., Acta C h a . Scund. 11, 1424 (1957).
114. Foss, O., and Tjomsland, O., Acta Chem. Scand. 8, 1701 (1954).
276
OLAV POSS
117. Fws, O., and Tjomsland, O., Actu Chem. Scund. 10, 421 (1956).
118. Foss, O., m d Tjomsland, O., Acta Chem. Scand. 10, 424 (1956).
119. Fws, O., and Tjomsland, O., Acta Chem. Scand. 12, 44 (1958).
121. Foss, O., and Tjomsland, O., Actu Chem. Scand. 12, 1799 (1958).
i22. F o q O., and Tjomsland, O., Acta Chem. Scand. 12, 1810 (1958).
123. FOSS,
O., and Vihovdr, E. H., Acta Chem. Scand. 8, 1032 (1954).
i24. Foss, O., and Zachariasen, H., Acta Chem. Scand. 8, 473 (1954).
126. Fredgs, A., Acta Chem. Stand. 12, 891 (1958).
126. Furberg, S., and Oyum, P., Acta Chem. Xcand. 8, 42 (1954).
197. Gardncr, D. M . , and Fracnkcl, G. K., J . Am. Chem. SOC.78, 3279 (1956).
i28. Gee, C., Xci. Progr. 43, 193 (1955).
129, Gerding, I., and Eriks, K., Rec. trav. chim. 89, 724 (1950).
i30. Gochring, M., Chem. Ber. 80, 219 (1947).
131. Goehring, M., Fmtschr. chrm. Forsch. 2, 444 (1952).
i32. Goehring, M., Ergebnisse und Probleme der Chcmic der SchwefelstickntoffVerbindungen, p. 57. Akadcmie-Vrrlag, Berlin, 1957.
133. Goehring, M., and Feldmann, U., 2. anmg. u. allgem. Chem. 267, 223 (1948).
lS4. Goehring, M., Herb, H., and Koch, W., 2. anorg. u. allgem. Chem. 264, 137 (1951).
1.56. Gorin, C., and Dougherty, G., J . Org. Chem. 21, 241 (1956).
156. Grinberg, A. A., Zhur. Prilclad. Khim. 21, 425 (1948); Chem. Abstr. 43, 8567
(1948).
1S7. Guryanovn, E . N . , Zhur. Fiz. Khim. 24,479 (1950); Chem. Abstr. 44, 8181 (1950).
138. Guthrie, G., Thesis, California Inst. Technol. (1949), quoted by Pauling (186).
1%. Guthrie, G. B., Scott, D. W., and Waddington, G., J . Am. Chem. Soc. 76, 1488
(1954).
140. Harris, J . F., Thesis, University of Pennsylvania (1953), Univ. Microfilms, Ann
Arbor, Michigan, Publ. No. 4927.
141. Harvey, R. B., and Baucr, S. H., J . Am. Chem. Soc. 78, 859 (1954).
14% Haszeldine, R. N . , and Kidd, J. M., J . Chem. SOC.p. 3219 (1953).
143. Hertlein, H., 2. physik. Chem. 19, 287 (1896).
144. Himel, C. M . , and Edmonds, L. O., U. S. Patents 2,520,401 (1950); 2,572,567,
2,572,845, 2,574,829 (1951); Chom. Abstr. 44, 10735 (1950); 46, 6149, 9589, 4565
(1952).
i46. Hoogc, F. N., and Ketelaar, J. A. A., Rec. trau. chim. 77, 902 (1958).
146. Hubbard, W . N . , Douslin, D. R., McCullough, J. P., Scott, D. W., Todd, S. S.,
Mresrrly, J. F., Hossenlopp, I. A., George, A., and Waddington, G., J . Am. Chem.
Soc. 80, 3547 (1958).
1.47. Jmnc, H . , and Becke-Goehring, M., Chem. Ber. 91, 1950 (1958).
148. Keen, R. C., 2. Krist. 91, 129 (1935).
149. Kehl, W. L., and Jeffrey, G. A., Acta Cyst. 11, 813 (1958).
160. Kerker, M., J . Chem. Phys. 19, 1324 (1951).
161. Ketelaar, J. A. A,, and Sanders, J. K., J . Chem. Phys. 4, 621 (1936).
162. Koch, H. P., J . Chem. Soc. p. 394 (1949).
169. Koch, H. P., J . Chem. SOC.p. 401 (1949).
164. Krebs, H., Z . Naturforsch. lab, 795 (1957).
166. Krebs, H., and Muller, K. II., 2. unmg. u. allgem. Chem. 281, 187 (1955).
1.56. Krrbs, H., Weber, E. F., and Balters, H., 2.anurg. u. allgem. Chem. 276, 147 (1954).
167. Kruse, F. H . , Marsh, R. E., and McCullough, J. D., Acta Cryst. 10, 201 (1957).
i68. Kurtenacker, A,, Abeggs Handb. anorg. Chem. IV (1) I, 541 (1927).
277
169. Kurtenacker, A., and Matcjka, I<.,Z. unorg. u. allgem. Chem. 229, 19 (1936).
160. Kushner, L. M., Gorin, G., and Smyth, C. P., J . Am. Chem. SOC.72, 477 (1950).
161. Lecher, H., and Goebel, A.,Ber. 66, 1483 (1922).
162. Lecher, H., and Wittwer, M., Ber. 66, 1481 (1922).
163. Levi, T. G., Gazz. chim. ital. 61,286 (1931).
164. Levi, T. G., Guzz. chim. itut. 61, 373 (1931).
166. Lobunez, W., Rittenhouse, J. R., and Miller, J. G., J . Am. Chem. SOC.80, 3505
(1958).
166. Loevenich, J., Fremdhg, H., and Pohr, M., Ber. 62,2856 (1929).
167. Lorenz, L., and Samuel, R., 2.physik. Chem. B14, 219 (1931).
168. Lu, C.-S., and Donohue, J., J . Am. Chem. SOC.66, 818 (1944).
169. Marsh, R. E., Ada Cryst. 6,458 (1952).
170. Marsh, R. E., Pauling, L., and McCullough, J. D., Actu Cryst. 6, 71 (1953).
171. Martin, F., and Metz, L., 2. anorg. u. allgem. Chem. 127, 82 (1924).
172. Mathieson, A. McL., and Robertson, J. M., J . Chem. SOC.p. 724 (1949).
173. Mendelejeff, D. I., Ber. 3, 870 (1870).
174. Miller, W. S., and King, A. J., 2. Krist. 94,439 (1936).
176. Mills, H., and Robinson, P. L., J . Chem. SOC.
p. 2326 (1928).
176. Minoura, Y., J . Chem. SOC.
Japan, Pure Chem. Sect. 73, 131 (1952); Chem. Abstr.
47, 3118 (1953).

Y., J . Chem. SOC.Japan, Pure Chem. Sect. 73, 224 (1952); Chem. Abstr.
46, 6933 (1952).
178. Minoura, Y., J . Chem. SOC.Japan, Pure Chem. Sect. 76, 869 (1954); Chem. Abstr.
48, 12494 (1954).
179. Oyum, P., and FOSS,
O., Actu Chem. Scand. 10, 279 (1956).
180. Ohlberg, S. M., and vander Meulen, P. A., J . Am. Ckem. SOC.76, 997 (1953).
181. Ohlberg, S. M., and Vaughan, P. A., J . Am. Chem. SOC.76, 2649 (1954).
182. Otto, R., and Troegcr, J., Ber. 24, 1125 (1891).
183. Oughton, B. M., and Harrison, P. M., Acta Cryst. 10, 479 (1957).
184. Palmer, K. J., J . Am. Chem. Soc. 60,2360 (1938).
186. Palmer, W. G., Actu Cryst. 7, 381 (1954).
186. Pauling, L., PTOC.
Natl. Acad. Sci. U.S.36, 495 (1949).
187. Pauling, L., General Chemistry, 2nd ed., p. 368. Freeman, San Francisco, 1953.
188. Penney, W. G., and Sutherland, G. B. B. M., J . Chem. Phys. 2, 492 (1934).
189. Pollard, F. H., McOmie, J. F. W., and Jones, D. J., J . Chem. SOC.p. 4337 (1955).
190. Prins, J. A., Schenk, J., and Wachters, L. H. J., Physicu 23, 746 (1957).
191. Rheinboldt, I-I., in Houben-Weyl Methoden der organischen Chemie, Vol. 9,
177. Minoura,
pp. 1086-1118. Thieme, Stuttgart, 1955.

292. Rogers, M. T., and Campbell, T. W., J . Am. Chem. SOC.69,2039 (1947).
193. Rogers, M. T., and Campbell, T. W., J . Am. Chem. Soc. 74, 4742 (1952).
194. Rogers, M. T., and Gross, K. J., J . Am. Chem. SOC.74, 5294 (1952).
195. Schenk, J., Physica 23, 325 (1957).
196. Schmidt, M., Z. anorg. u. allgem. Chem. 289, 158 (1957).
197. Schmidt, M., Z. anorg. u. allgem. Chem. 289, 175 (1957).
198. Schoberl, A,, and Wagner, A., in Houben-Weyl Methoden der organischen
Chemie, Vol. 9, pp. 83-92. Thieme, Stuttgart, 1955.
199. Schotte, L., Arkiv Kemi 8, 579 (1955).
200. Schotte, L., Arkiv Kemi 9,299 (1956).
201. Schotte, L., Arlciv Kemi 9,309 (1956).
202. Schotte, L., Arkiv Kemi 9, 361 (1956).
278
OLAV FOSS
M03. Scholte, L., Athw Ketnz 9, 441 (1956).

204 Scott, D. W., Finkr, I. T,, Gross, M. E,Guthric, G. 13., and Tluffman, IT. M ,
J. Atti. Chetti. SOC.72, 2424 (1950).
206. Scott, D. W., lfinke, H. L., McCullough, J. P., Gross, M. E., Pennington, R. E ,
and Waddington, G., J. Am. Chem. SOC.74, 2478 (1952).
206. Smyth, C. l., Dielcctric Behavior and Struetiire, pp. 3813388. McGraw-Hill,
New York, 1955.

Sorum, H., A d a Chem. Scand. 7 , 1 (1953).
Spac;, C.S., and Popper, E., 2.physik. Chem. 8184, 367 (1939).
Spong, A. H., J . Chem. Sor. p. 1547 (1933).
Steinrauf, L. K., Peterson, J., and Jensen, L. H., J. Arn Chern. SOC.80, 3835
(1958).
811. Stevenson, D. P., and Beach, J. Y., J. Am. Chern. Soc. 60, 2872 (1938).
212. Szczepkowski, T. W., Nature 182,934 (1958).
213. Taylor, P. G . , and Bcevers, C . A., Acta Cryst. 6, 341 (1952).
214. Tomiie, Y., Acta Cryst. 11, 875 (1958).
216. Tomiie, Y., Koo, C. H., and Nitta, I., A d a Cryst. 11, 774 (1958).
816. Toimsaint, J., Bull. SOC. chim. Belges 64, 319 (1945).
217. Troeger, J., and Hornung, V., J . prakl. Chem. [a] 60, 113 (1899).
218. Tuncll, G., Mcrwin, H. E., and Ksanda, C. J., Am. b. Sci. 1.51 36A,361 (1938).
21.9. Verneuil, A,, Ann. chzm. el phys. [6] 9, 289 (1886).
620. Weitz, E., and Achtcrberg, F., Ber. 81, 399 (1928).
221. Weitz, E., Becker, F., and Gieles, K., Chem. Ber. 89, 2345 (1956).
222. Weitz, E., Becker, F., Gieles, K., and Alt, B., Chew. Ber. 89, 2353 (1956).
229. Weitz, E., Giclcs, K., Singer, J., and Alt, B., Chem. Ber. 89, 2365 (1956).
224. Weitz, E., and Spohn, K., Chem. Ber. 89, 2332 (1956).
22.6. Wcstlakc, H. E., Laqucr, H. L., and Smyth, C. I., .I. Am. Chem. S0c.72~436(1950)
226. White, G. N., J. Chem. Soe 113, 60s (1918).
227. Wilson, M. X., and Badgcr, R. M., J . Chem. Zhys. 17, 1232 (1949).
228. Wood, H. W., J. Phol. Sci. 2, 154 (1954).
260. Wood, H. W., Chem. d? Ind. p. 468 (1956).
230. Woodrow, C. C., Carmack, M., and Miller, J. G., J. Chem. Phys. 19, 951 (1951).
23t. Wyrkoff, It. W. G., Crystal Structures, Vol. 1, Chapter V, p. 24. Interscience,
New York, 1948.
232. Yakcl, H. L., Jr., nnd Hughes, E. W., Arta Cryst. 7, 291 (1954).
233. YanitskiI, I. V., and Valanchuoas, I. N., Sbornik Statu Obshchei Khtm., Akad.
Nauk 5. S.S.R. 1, 732 (1953); Chem. Abstr. 49, 8023 (1955).
234. YanitskiI, I. V., and Valanchunas, I. N., Zhur. Obshchei Kham. 24, 790 (1954);
Chem. Abstr. 48, 13.510 (19.54); .I. Gen. Chem. U.S.S.R. (hngl. trunsl.) 24, 793 (1954).
235, Yanitskir, 1. V., Valanchunns, I. N., and Tuchaitc, 0. Y., Zhur. Neorg. Khim. 3,
2087 (1958).
236. YanitskiI, I. V., and Zelionkaite, V. I,, Zhur. ObshcheZ Khim. 26, 841 (1955); Chem.
Abstr. 49, 13003 (1955); J. Gen. Chem. U . S. S. R. (Engl. trans!.) 26, 805 (1955).
$37. Yanitskir, I. V., and Zclionkaite, V. I., Zhur. Neorg. Khim. 2, 1349 (1957); Chew.
Abstr. 62, 3581 (1958).
238. YanitskiI, I. V., and Zelionkaite, V. I., Zhur. Neorg. Khim. S, 1755 (1958).
239. Zachariaseo, W. H., 2.Krzs!. 89, 529 (1934).
240. Zacharimcn, W. H., and Moorwy, R. C. L., Z.Krist. 88, 63 (1934).
207.
208.
209
210
CHEMICAL REACTIVITY OF THE BORON HYDRIDES

AND RELATED COMPOUNDS
F. G. A.
Stone
Deparfment of Chemistry, Harvard University, Cambridge, Massachusetts
I. Introduct,ion .
. 279
11. Reactions of the Boranes and Their Derivatives

.
A. General Principles . . . . . . . . .
. 283
. . . . . . . 283
B. Reactions Leading t o Formation of Boron-Nitrogen Bonds . . . .

C. Reactions Leading t o Formation of Boron-Phosphorus, Boron-Arsenic, or
Boron-Antimony Bonds
. . . . . . . . . . . . . .
D. Reactions Leading t o Formation of Boron-Oxygen Bonds . . . . .
E. Reactions Leading t o Formation of Boron-Sulfur or Boron-Schium
. . . . . . . . . . . . . . . . . . .
Bonds
F. Reactions Leading to Formation of Boron-Carbon Bonds . . . . .
G. Reactions Leading to Formatlion of Boron-Halogen Bonds. . . . .
111. Recent Advances in the Chemistry of t.he Borohydrides
IV. Conclusion
References
.
.
.
.
.
.
285
294
297
298
299
304
305
. 308
. 309
I . Introduction
Alfred Stock (131) and his collaborators were the first to characterize
a series of boranes. Properties and many reactions reported by them (19121936) have required but little revision to this day. For one hundred years
prior to the investigations of Stock and his co-woi kers, the boron hydrides
had been produced intermittently, but they were not identified or analyzed
correctly. Stocks work stimulated others, notably Burg and Schlesinger,
to enter this field so that from 1930 until 1942, the date of an important
review ( I l l ) , structural studies were carried out, and the abandon with
which the boranes enter into chemical combination with themselves and
with other molecules became evident. Development of improved, or even
new, experimental methods has necessitated reinterpretation of most of
the early structural results. As would be expected, preparative studies
carried out prior t o about 1942 have required less revision, providing a firm
foundation on which a great extension of synthetic work could be built.
Until fairly recently interest in boron hydride chemistry was largely
academic but was sustained by the failure of the boranes, and certain
compounds related to them, to conform with the usual rules relating chemi279
2x0
F. G . A. STONE
cal composition with classical valence theory. This has now changed, and
in little more than a decade the boron hydrides and their derivatives have
assumed an importance far greater than could possibly have been envisaged
hy pioneer workers in this field. This is a conscquerice of the production of
certain boron compounds on an industrial scale for use as fuels and fuel
additives (I&') and as reducing agents in organic and inorganic chemistry
(56)*
Compounds rich in both boron tirid hydrogcil are especially suitable as

chemical fuels because they combine such properties as high energy content
per unit weight with high density and with low melting point. Of all thc
elements hydrogen has the highest heat of combustion, 52,000 BTU/lb a t
25C. Carbon has a heat of combustion of 14,500 BTU/lb. Hydrocarbons,
therefore, have intermediate hc:h of combustion. Thus kerosene burned in
jet engines yields about 18,600 BTlJ/lb. Boron itself releases 25,000
BTU/lb when burncd, arid when it is combined with hydrogcri in the form
of diborane (13zH6)cornbustion yields 32,000 Brl'lJ/lb. The hydrides R5Hs
id RloHIParc' only a little less efficient in this rcspect. The boranes are thus
theoretically far superior fuels to herosene.' However, actual fueIs currently
being produccd are believed t o be alkylated boranes. Introduction of carbon
decreases rfficicwy, but the resulting compound would be easier to manipulate than a simple boron hydride.
New principles of propulsion may in time make chemical fuels obsolete.
Furthermore, boron is a somewhat rare element. From the point of view of
industrial iritcrcst in this field, it is fortunate that compounds containing
boron and hydrogen have other applications, for example, as reducing
agcnt 8.
Boron hydride structures and structures of moIecules related to them, as
well as other properties associated with bonding problems in these compounds, have been reviewed in the first volume of this series (83).Here it
is intended t o survey chemical reactivity of the boranes and of their
derivatives.
Rapid expansion of boron research with isolation of many new substances
has led t o a proposal that there should be a new scheme of nomenclature
for boron cornpounds (90, 100). Most workers now reporting their result,s in
the English language literatme do so in accordance with the new scheme.
Although the subjcct is still being debated with vigor and there has been
no international agreement on the subject, the new system of naming will
be used here, except where long-established compounds are involved and
trivial designations have received widespread usage. The new system may
A recent useful article on this subject is by D. R. Martin (83~).
However, it now
appears that performance of the borane fuels has not lived up to expectation, and costs
arc much higher than anticipated. See Chem. Eng. News 37, No. 36, 38 (1959).
CHEMICAL HEACTIVITY OF THE BORON HYDRIDES
281
be illustrated by considering the BH, group, long referred to as borine

but now better described as borane, thereby emphasizing its relation to the
commonest boron hydride, diborane (B2Hs). A natural extension is to consider most compounds as being derived from borane, thus calling alkyl
compounds RaB, RZBX, and RBX2 trialkylboranes, dialkylhaloboranm,
and alkyldihaloboranes, respectively; for example, (CH3)3B, trimethylborane; (C2H6)2BBr,diethylbromoborane; and C2HsBF2,ethyldifluoroborane. Other new names put forward have received far less support; for
*
example, tetrahydridoborate(II1) for the BH4- ion.
Through the years the boron hydrides and their derivatives have been
reviewed several times, and the reader would do well to refer to these articles (6, 10, 27, 57, 111, 132, 146, 147) for a more detailed survey of borane
chemistry than is possible in the space available here.
Well-authenticated boron hydrides are listed in Table I. Characterization of these hydrides does not preclude the isolation of others (83).Properties known at present to be common to all boranes include thermal instability; susceptibility to hydrolysis, to alcoholysis, and to oxidation; toxicity;
ability to react with ammonia and with many electron-pair donor molecules
[for example, (CHa)sN]; and a tendency to undergo substitution of their
hydrogen atoms by halogen or alkyl groups.
As might be expected, the hydrides possess these properties to varying
degrees. At ambient temperatures decomposition of pure diborane is less
than about 10% in a year but at 100" it rapidly forms hydrogen and other
hydrides. On the other hand, tetraborane-10 and pentaborane-1 I decompose into other boranes fairly rapidly at room temperature. Similarly,
whether or not a borane will inflame on contact with air depends not only
on the borane but also on the conditions of temperature and pressure. Most
of these general properties have been known for a long time, but in the last
ten years much work has been done to pIace knowledge of them on a more
quantitative basis. The following studies illustrate this point. Pressure and
composition explosion limits have been determined for diborane-oxygen
(91, 97) and pentaborane-8oxygen (2, 92) mixtures. The results have been
interpreted in terms of branched-chain reactions, with oxygen atoms probably taking part in chain propagation. Experiments have shown (119, 120,
144) that the hydrolysis of diborane and of pentaborane-9 very likely
involves the intermediate
+-
1320. BHI,
and Stock's (131) observation that pentaborane-9 reacts only slowly with
water at room temperature has been shown (121) to be due to poor miscibility of the hydride with water. In the presence of a mutually miscible solvent
(dioxane) the rate of hydrolysis of pentaborane-9 is fast. A kinetic study
TAkBLEI
PHYSICAL
CONSTANTS
O F THE VOLATILE
BORON
HYDRIDES
Name
Melting
point ("C)
Boiling
point ("C)
Diborane-6
Tetraborane-10
Pentaborane-9
Pentaborane-li
Hexaborane-10 Enneaborane-15
-165.5
- 120
-46.6
- 123
-63.2
-92.5
18
48
63
82.2
-20
-
Decaborane-14
99.7
213
u-
CHEMICAL REACTIVITY OJ? T H E BORON HY'DRIDES
283
( 7 , 8) has been made of the reactions between decaborane-14 and various

alcohols, and deuterium exchange of decaborane-14 and deuterium oxide
has been examined (68). Partial hydrolysis of pentaborane-11 below room
temperature yields tetraborane-10 (IS). This reaction illustrates a new
principle-interconversion of boranes by appropriately chosen chemical
reagents rather than interconversion via thermal decomposition type
reactions. A further example of this procedure is the synthesis of the
hitherto rare B6HI0,several physical properties of which have recently
been reported (56u), by treating BsHll with dimethyl ether or certain other
basic reagents (1%).
11. Reactions of t h e Boranes and Their Derivatives
A. GENERALPRINCIPLES
On the basis of known behavior most boron hydride reactions may be
assigned to two main classes. The first class involves removal of a borane
group from a boron hydride structure, while the second class involves
nonsymmetrical cleavage of the hydrogen bridge bonds in a borane.2
Removal of a borane group often occurs when a borane is treated with a
nucleophilic reagent. The initial product of this type of reaction is an adduct
of the borane group, but the nature of the donor part of the adduct controls
the subsequent fate of the whole. For all such compounds there is a tendency, varying from complex to complex, to dissociate reversibly into donor
and acceptor parts. Thus whereas (CH&S * BH3 (49) is appreciably dissociated in the gas phase at 50", the compound (CH3)QP. BH3 (43) may be
heated to 200' without appearance of significant amounts of borane decomposition products.
Some borane adducts decompose irreversibly. Thus if the ligand atom
has hydrogen bonded to it, elimination of molecular hydrogen from the
complex occurs either as soon as it is formed [ ] or on heating; for example,
BzHo
+ ROH + [R(H)O. BHa]
+ ROBHz
+ FL
+ (RO)aB + BZH6, (38)'"

(CH3)zNBHz + Hz. (36, S7,148,149)3
(CHa)?NH
+ BzHa
disproportionation
heat
(CH3)zNH. BH3 ---+
(R0)zBH
*a In this chapter, parenthesized numbers appearing t o the right of chemical equations are reference numbers throughout.
* Since this article was submitted to the publisher, two papers concerned with classification of the reactions of the boranes in relation to their structures have appeared
(83a, 85a).
3 This compound, N,N-dimethylaminoborane (mp, 75"), like most BH2 derivatives
polymerizes. When prepared by this method, the solid is dimeric but a monomerdimer equilibrium exists in the gas phase, complicated by some disproportionation t o
284
F. G. A. STONE
Reactions related to these are the iriariy which occur in the gas phase
between boranes and molecules with insufficient dative bonding power to
permit formation of a definite complex even at low temperature; for
example,
IbTI, HpAs + Hz (HzAsBHz)~,

BzHs (CF3)2PII I32 f [(CF&PBHz]s.
(1%)
(311
An alternative route for irreversible decomposition of a borane adduct involves transfer of hydride from boron to the ligand, usually occurring with
a rapidity sufficient to prevent isolation of an initial complex. Indeed, existence of an adduct at the first stage of reaction can often only be inferred
from the nature of the final product; for example,
<[I: ]
l 3 z T I ~ CsH6 + CH,
BpHr,
+ CHaCTO
- BH,
+
+ [CHJCHZCHZBH~]
* (CsH7)aB,
disproportionation
[CHsCHO . BHal+ [CHsCHzOBHzI
(59)
>
disproportionation
(CzHpO)zBH
+ (CzHsO)&.
(16)
The boron hydrides function as Lewis acids in these reactions involving

borane. In this connection important energy factors are dissociation of the
boron hydride in the gas phase to give a borane fragment, vaporization of
the base if it is a liquid or solid, adjustment of borane and of the base to
configurations prcscnt in the final product, energy released in dative-bond
foimation and in condensation of product to a liquid or solid (13.2, 233).
Thus each reaction must be examined individually, and although data for
component parts of such an eriergy cycle are lacking, the cycle can be used
qualitatively for discussion of real or possible compounds. Physical properties of several borane addition compounds are presented in Table 11. Heats
of dissociation and heats of formation of a number of borane coordination
compounds have been measured. Borane exists in its standard state as
B2H6, so that the measurements nrcessai ily irivolve this dimer. However,
the enthalpy change for the reaction B2Ha(B,= 2BH3(,) has been estimated
as 28.4 kcal/mole (5, 84). Hence strength of the dative bond in a borane
itddition compound can be determined from data involving diborane if
[(CH&N]zBH and (CH&NBzHs. As will be described in Section II,B, a trimeric form of
(CH&NBH2 has a ho been reported. In most cases occurrence of borane derivatives like
(CH&NBHa or (CaHs0)2BHas monomers can be accounted for in terms of the existence
of double bonding involving ovcrlap of a filled p,-orbital on the donor atom (namely
N or 0) with the empty boron p,-orbital. See references (10) and (155) for more detailed review of thiis ideti.
CHEMICAL 1tEACTIVITY OF THE BORON HYDRIDES
285
strength of the bond is defined (153) as the eiithalpy change accompanying

the reaction,
Ligand . BHa(gj+ Ligand(,)
+ BHa(,).
More recently another mode of reaction of some of the boranes has been
recognized. Certain reagents under rather carefully controlled conditions
induce nonsymmetrical cleavage of the hydrogen bridge bonds. As will be
discussed in more detail later (Section II,B), formation of the diammoniate of diborane probably involves this mechanism.
The sodium-diborane reaction (72) provides another example of unsymmetrical cleavage of a double bridge bond.
2Na
+ 2BaHe + NaBH, f NaBP8
Besides removal of borane groups and unsymmetrical bridge-bond cleavage, other types of boron hydride reaction are known but at present appear
to be much Iess common;2for example, loss of protons, as in the ionization
of Bl0HI4in dioxane-water or ethanol-water solutions (61, 69),
B I o H I ~ (G
~ QBloH-la(aq)
)
and in salt formation by
BlOH14
(CeH6)3P+CHz-
+ H+(aq),
(65),
+ BioHu
[(CsH~hPCHslBioHla,
or internal exchange of hydrogen atoms as in the reaction betweenfBloH,,

and acetonitrile to give BloH12(CNCH3)2.The latter substance is best
regarded as a substitution derivative of BloH14- (83),not a derivative of
B10H14,with the boron-hydrogen arrangement having either the 2632 or
0814 topology (139).
B. REACTIONS
LEADING
TO FORMATION
OF BORON-NITROGEN
BONDS
Discovery of the djammoniate of diborane (BZHa, 2NH3) and borazene (B3N3H6),better known by its older name of borazole, by Stock (131)
indicated that the boron hydrides could be used to prepare interesting
boron-nitrogen compounds. This approach to boron chemistry was rapidly
extended when the principles discussed in the preceding section were
clearly recognized. Thus the compound (CH3)gN * BH3, the first borane
adduct to be characterized (S9), pointed the way to the synthesis of a
number of similar substances directly from diborane and amines; for exam-
f3
TABLE I1
COORDINATION
COMPOCNDS
OF BORANE
Compound
Melting
point
BoiIing
pointa
("C)
("C)
h 3 0
P ( m )
=
-A/T
+B
a
AH
(vapor)
(kcal/mole)
Slowly decomposes, liberating H? at room temperature

54
-49.3
Solid at -78.5
( p h = 10.5 mm)
36
- 22.6
-22.4 t o -21.5
94
-4
-33
103
73.5-74.5
-33
Solid a t -78
2-4"
55
-23 to -21
- 16.5
-116.1
- 137
150
Unstable
2329
8.400
I
10.7
-
174
2337
8.100
10.7
85.5
2263
9.191
10.3
171
2202
7.846
10.1
Vapor pressure 12 mm at 100"
Vapor pressure 14 mm at 125"
Vapor pressure
mm a t 87"
2933
9.531
154
2420
8.553
11.1
Unstable
Loses HZ on warming to room temperature
h e s HZ on warming to room temperature
Vapor pressure 6 mm a t 25"
Slightly volatile at. 40"
Slightly volatile a t 25"
Vapor pressure 3.8 mm at 26"
-61.8
1038.9
7.8061
4.760
-64
1040
7.850
4.750
Trouton
constant"
(eu)
References
183, 124
1 , 12, 96, 148
149
25.3
-
43
134
23.9
28.9
22.7
1 , 96,148,149
43
134
39
80
80
80
43
26.0
134
94,71
45
45
45
11
11
11
22.5
23
86
39
?
rn
c3
( C H h O . BH,
Solid a t -778.5
Unstable
110
62
97
132.5
119.5
63.2
1244
2346
2321
2394
1732
6.592
9.220
8.602
8.978
8.030
10.7
10.6
11.0
7.9
17.0
29.0
26.2
27.9
23.5
49,95,154
44, 49, 58
49
49
58
hiss
= 18 mm)
(CH2140. BH,
(CH3)eS. BH,
(CHs)aS. BH,
(CPHS)SS
* BH3
(CH&Se. BH3
-34
-40 to -38
-34 to -32
In order t o calculate the boiling point, a large extrapolation of the vapor pressure equation is involved, so these values must be considcred as approximate. Furthermore, whenever vaporization of an addition compound is accompanied by an increase in dissociation,
extrapolation of the vapor pressure equation cannot give the true boiling point. This is exemplilied by tetrahydrofuran-borane, the boiling
point of which is less than that of the ligand [tetrahydrofuran; bp 64"]. The large entropies of vaporization of molecular addition cornpounds are a consequence of further dissociation on vaporization, although dipole association in the condensed phase probably enhances
the effect to some extent. A few addition compounds do not have high Trouton constants even though they are known to be highly dissociated in the gas phase. Such adducts are no doubt partially dissociated even as liquids.
(1
5>
F
&
4
0
LIJ
288
F. G . A. STONE
ple, CbHsN RH3 (17)' CHSNH, * BlIs (112, 148, 149), (CH&NH H H 3
(36, 148, 146), and (CH&NH BH3 (33).Moreover, many tidducts of this
type provide a I oute to substituted derivatives of borane, because in those
instances where the donor atom carries a proton the complex will elimirizltc
hydrogen, either incipiently or on heating; for example,
2(C11[2)rNI BFT, --Ha

+
bz
/c"z\
CH2
\
hWII21
(33)
/
2
Azetidine-borane
Azetidinoborane
loo0
3CHaNHz. BHa A (CHsNHBH2)a
N-methylamincborane
+ 3H2
(18)
N-methylaminobornnc
( C H & ~ ~ St 3Ht
N-trimethylbornsole
(112,151)
Aminoboranes undergo an interesting reaction when heated with diborane; a borarie group is absorbed and amiriodiboranes are formed (35,56).
The structure of N,Ndimethylaminodiborane has been determined by
electron diffraction (70);it (I) is related to the structure of diborane and
to the probable structure of dimeric N,N-dimethylaminoborane (11).3 It
is likely that all aminodiboranes have similar structures. The first member
of the group, HzNH2Hs, may be prepared by passing diborzlnc over its
CHa
H
HaC
CHI
\N/
\ / \B/
B
I
'
\H'
'11
CJT,
\N'
\y
H3C
(1)
I1
\ /+\
/
B - -13
\H
II/
/ \
CH3
(11)
dizlmmoniate a t 88", and st n pressure of about 1 atrn (113). It decomposes

slowly into a polymer, (H2B. NH2),, and diborane. Alkylaminodiboraries
are niore stable. N,N-Dimethylaminodiborancmay be stored indefinitely
a t room temperature.
The effect of ammonia on diborsne is very complex. A variety of products are produced depending on conditions. At - 120' using very specific
techniques, diborane and excess of animonia forni the diamnioaiate, B2Hs,
2NH3 (89, 110, 131). The latter was a t first formulated as (N134+)2(B2H4)-2
(131,146), but this structure was later shown to be incorrect (26, 110), con4 Borane adducts of various dkylpyridines havc been preparcd. They were studied
in order t o correlate their stability with certain steric effects in organic chcmistry. See
Brown and Domash (1%).
Formillti
-66.5
-54,s to -54.4
-39.0
-69.4 to -6X.S
-96.1
-146 to -143
-45.4
-63.5
~~~
~~~
_____
From the vapor pressure equation.
76.2
66.8
50.3
51
54
86.6
121
101
121.8
148
2097
0.00643
2158.56
0.00SOG
1727.64
0.004661
loglop(mm) = S.0S1
loglop(mm) = 7.974
2090.3
0.00533
log,, p(rnm) = 8.537
224s
0.00616
2205
0.004539
2410
0.00510
6.677
7.518S3
5.3370
- 16S6T-1
- 16697'-I
6.15245
- 21177'-I
6.6928
5.7217
6.1606
21.0
19.6
20.6
23.S
23.3
21.0
23.9
20.2
20.6
19.6
113
36
96
.95
3.5
33
53
33
SS
33
vitiriiig cvidcticx: :ig:iiiist it, cntiiitig from :iii isot,opic: t>r:iwr st.udy using
A IIs (26). lhis sho~vedtli:it 1-1) csc.h:inge I)ct\vccn solute atid sol\.ctit
occ-urretl o d y hetwcti K-hoiitlcd hydrogens. L:it,er t,hr strrictitrc! (NI14+1( I l n R * IT112 HH3)- was suggcstetl (110) maitily bcc~nuscwith sodiiirri i t 1
lic luid :iI I i I 110 t ii:t t h r diam inn I i i:it c give oi ic cqit iv:t lei I t of hytl rogcii . This
was takon :is evidence for thc: presetice of one immoniiiiii ioii per pair of
hornti n toins in t he st,riict,iire.The cwnsi it.ution of tlic idit)orntiediminioni:it(! Iias rccent ly tm:n p1:iccd on n niuc!h firmer I):isis hy st,iitlicsof Parry tirid
his cw-workcrs (S8,88, 117, 122, 124, 125). Thore is tiow no doul)t that the
di:~iiitiioni:~tecont:iins the horohytlridc ioti, BH4-. This wis first, itidic:itcd
thc \\.ark of (;. W. Schacffcr~t ( I / . (101) \rho fo~ititlt h t \\.hell BJI6, 2NH3
\v:is trcitetl \\itll sodiiitii, sodium 1)oIohydriclewis produced. Later it was
showti ( 1 1 7 ) t.h:it [lIg(NH3)~](B114)2
cwuld be prcc*ipit:ittxl from :L liquid
:itiiinotii:i solution of the di:trnmorii:it,c.. I h t h c r m o r e , t,hc Ihmnti spcetrum
of RpHe 2Kl13 is in :~c:cordwith the prcscti(:e of :iBH,- ion (1,W).
It should
l x tiotctl t1i:it the liorohydride ioti had not, bccti disc.ovc.rcc1 at, t,lic time
Svhlcsiiiger :tiid Burg (110) proposed the structure (NH4+)(H3BNIz1~113)-.
Oticc t hc prrsence of borohydridc in the st,ruct,urc wis cst~nl)lishcd,the
itlcvi of hnvitig : ~ i i:miinniiiiirn ion siniultaneonsly present liccwiic unlikely
l)eeniisc :itnriioniiitii 1)orohydricle is unst,nlilc (87). hlorcovcr, forni:it,ioii of
oiic e(1iiiv:ilrtit of hydrogcii I)y tre:it.iiig t,he dinnitnoni:tt,e of dil)or:me ivitli
sodiut~iiii liqiiid :itniiioni:i ( 1 10) is not :I spceifia t,cst for miitnotiiiitii ion,
Init merely iriclic+:ites thc prcsenc*c of ;I 131iinst,cd-I,o\vry :wid (117). h
idcn ( l ( 1 1 ) that nilinioniuin ntid borohydrjdc ions cnti csist together in the
c1i:immoni:itc hecause they :ire st,:ildized by n H,13NI12 group, namely,
XH4+(BII,NH,)BH,-, is difficult to reconcile with ccrt,:iin espwiment:tl
ol)servat,iotis (11 7 ) . A rc:isonahle expl:in:it,ion of the results is poswil)lc if
the presetice of tlic c:it,ioii [H,B(NII&]+ iti the dintiitiiotii:it,e is :issumed.
This accords with the isolation of chloride and t m m d e s:ilts [II?B(NH3)?]5
from re:ictions 1)etween the dianimoiii:itc :ind the :ippropriat,e :iiiiniotiiiiiii
hnlitlc, with the itiertticss of the c1i:imnioni:itc towirds sodiurii horohytlridc,
: i d wit,h the behavior of tlic di:mitnoiii:itc on hydrolysis, ns n.cll :is with
certain otlicr propcrt.ics. Furtherinore, n n S-ray diff r:ic:t,ion study of
(NH&BH2CI dcmonstr:ttcs th:it the c:itioi) [H2R(NH3)2]+is present iii
this coinpourid (S/to). Thus the tli:iriinionint,c of diborane is now best rcg ~ r d c das hsvitrg thc struc:t,urc [I1213(NH3)2]+
[13114]-. Quite dist.iiic-tfrom
this compou~idis niiitiioiii:i-borniic (123, Is/t),prepared by tlic re:ic.t ions,
(cdI6)zo
IAi1%I14 NH4<:! ___ + LiCl
+ II,N .
HI13
+ 112
I ~ t i i t i i o i i i : t - t ) o r ~ i t i cis
' thc?rni:illy less s t d h thaii tlic "di:tiiitiiotii:itc" with
r(qm:t to loss of hytlrogoti. It, is iiioiiotiicric, :tiid uiilike the "di:itiit~ioiii:tte~~
is soluth in ethar.
When diborwe and excess of ninmonia :ire heated together, boron imitle
:iiid ultitnntety 1)oroti iiitrideb arc produced. If dibortiiie i i t i d nmnioniu i i i
:i ratio of 1 :2 :ire he:ited, or if the diammotiiate of diboratie is heated above
2W0, thc conipouiid bornzole(II1) (bp, 53") is obtained, but in yields o d y
:is high as 45%. Ror:tzole decotiiposes slowly iii the liquid phase atid even
more slo\vly in the gas ph:ise. Its boroii ntonis cat1 act as weak electroti
:tcceptors aiid its iiitrogen atoilis :is weak donors; thus hydrogeri chloride
adds to lxmzole, aiid oil heating B,B',B"-trichlorol)or:ixole is produced.
111 :I sitiiilar rnaiiiic'r water re:tcts with t)ornzolc. This rcnctivity err1phasixc.s
how t)oraxolc differs from hctizene :md is rc1:ittxl to loctilization of electrotis
I1
II
S
HI3/ N h lI
III{ '."l$lI
TIN:
c t
:NII
,'
1%
II
TIN
N II
11
(111)
on nitrogen, for 1vhic.h therc is ndditiotial evidence from spectroscopic

studies (93). A very large tiuint)er of substituted bor:izoles are knowii.
Reartioils by wliicli horazolc : i t d sotile of its derivatives hnvc twcri prepnrerl
iiicludc:
:$LiUII, :3NlI,(;l
(niixtiirc- of solids)
((T?
m0
-+
I I s )?(
I%rNJTti 9IL
(:IO-Y5 ";)
yicld)6
+ 3LiC1
(104)
(75)
6 In this rcnc-tioil tliv hexngonnl or " w l i t c grnphitc" forin of t l i q snbstnncc: is obtained

It is composed of rond~wsed(BN), rings. A cubic forin, bomzon, analogoiLs to diarnond
11:~s rcccntly bccm prcp:ircd. It, is ofteii s t n t d t,hrit. t h : struetiires of graphite and hexagoiiul IJ(J~OII nitritlc art' sinrilnr. 'I'hcy arc rc!l:tttd but differ in an important respect.
1,octiIiztit ion of clcctrons on the nitrogen :itnnLs in boron nitrido lea& to a different
iiiodc of Inyrr pwkiiiK, with horon utoins in one lnycr diroctly under nitrogen atonw in
tinot hcr. I n ;irrorthnc!o wit 11 this, hcxtigonnl tioron nitride is n poor electrical condlIvt,or.
6 Rcccmtly Eiiic.16us and Vitlclii (5Ua) have shown that if hydrazinc hydroc-hloritlc is
iiscd in this rcnrt ion iristcd of aniinoniiiiii chloride, thc yield 3.i n w r 50(;;,.
hent
3RRC1 ONHI
[RBNH13 (iNII,CI
(R = n-C,IIy, CsIIa, CICII:CII, etc.)
AS \\o~tldt)c rspc.ctc4, hytlrolytic: :itit1 thcrrn:il st,nbility of the suhtituted

I)(mtzoks is clcpciitl(vit on t,he substituent groups. Thus U-trimethyl-Nt~ril)hetiylbor:izolcis hut little :iffec:tcd by \v:iter :it loo, \vhcrc;is the
~\-tri:ilk~lbor:tzol~~s
arc rc:idily hydrolyzed.
A I I cstmsi\x field of I,oroii cheinistry is :issoc:i:itcul with ro:ic:t,ions het\\.ccii t,ha Iiighcr I)tirtiiics :incl :iititiiorii:t or atnitics. Stock (Z:?1) wis rcspoiisitdc lor disc+o\vryof n scrics of :iitiiiioiiiiit,es of the Iiiglic!r i)oron hydrides
:i11c1 for their t,heriit:il clec:otiipositioii it) to Imxzole ntid tioiivo1:itilc boroiiiiiLrogcti (:ont:iiriiiig ni:itcri:ils. The properties of all these :inimotii:ttes are
\\orthy ol further itivcstig:itioii siricc recently it has I)eeti showti (77, 78)
t,h:it tc*ti.at)ornticb-IUforms ti diaiiitiiotiiLitc atid tint :i ~~tctr~itiriiriioiii~it(!
:ISrcport,ccl by Stock (131). The Ldiiiriir~iotii:ite))
of tctr:it)orniie h:is propertiw coiisisteiit with the structure [HzB(NH3)2]+(H3Hg]-(77, 78, 7Du). This
ittisymtiictrical clenvtigc by amniotiiu of the hriclge bonds of n highcr borntie
is i i i contmst to the behavior of terti:iry uriiiiics towards tetrtrborane, pentatwr:uie-3, or pental)oratie-ll. The ultimate products of the renctiori betweeti
7
111 tlw prcsciice of cwI)nlt c:itlrlyst yicltls
(GOU).
iii
this: rciict,ion nro iiinrkedly iIicr(w(x1
tetr:il)oranc :ind c x c e s of trimethylaminc :ire thrw moles of (CH 3)3N BH,

and (BHz)n polymer (42). With a 1 : 1 niolar ratio of rc:ictniits, howcver,
tetr:iborane and trimethylaiirinc yield a mixture of (CH3),N HH, and
(CH3)3N H3H7.In a similar rniiiiiier the adduct C&N * R3H7has been
obtained from pyridine arid tetraborane (73). With excess of the nitrogen
bases the RaII.~-:idducts decompose giviiig an amine-boranc and dibornne.
The B&I7 group has n strong affinity for pyridine sirice the reaction,
+ ChIIaR . BII3
141110
-+
Ct.IIbN . BaH,
+ I32Hs
occurs. Adducts arc also forriicd 1)ctween B3H7 and diinethyl sulfide,
ethylene glycol ethers (73),:tnd aminonia (77, 86),
H4Mio.
LNIT,
+ HCl
(Celr1)zo
-78
(NH3)zBITk:I
+ (C:Ilfi)?O. BaII7 + Hz
111 coiitrast to tctral)oraiie, no reactions I\ ere observed between pentnlmranc-9 :~nd we:A I ,ewis lxises such as teti nhydrofuran or diriiethyl
sulfide (74). With t rimethyl:imine, however, :it -78 (28) or t i t room ternpcraturc (74), pe1itJ~or:ine-9:iffords the white sublim:iIh histrimethylamirie pcrit:ilmwie-9, [ (CH3)3K;l2B6H9.If the latter bul>st:iricc is heated
(H),
or left in c*ont:ict uith ariiinc (74), c1eav:igc occurs uith ptrtial conw r h n to ((II8),N 731T3 :itid (I<Hz)n.
In the :thscnce of exwss of :tminc,
1~i~trimcthyl:iniitic
pcbtitahor:uie-9 slowly forms (CH3)3N BI, : i d trinicthylamiiir tetr:iIioran(-6, ((11,)JV
I 3 J I ~ (74). Dirnethy1:tmiiic or the
Ar-n~ethylnn~iiiol)ornrics
:h
reniove boranc from B ~ Hon
Q heati ng, but form
as well :I cwiipouiid first fonnulatcd :is I (CH3)2NJ3B& (28).6
8 Very rrrcwtly thr rcuction brtnccsn I3JID and nxwss of dimcthyla~~~inohor~tllc.

hw
bren rcinvestigntrd ( 4 6 ~ )On
. the brwis of new acid-mcthanolysis dnta and an N M R
study the conipound rcyorted tm [ ( C H I ) ~ N ] ~ B ~hns
I T ~now
bccn fornidatrd us [(CH3),NBH& a ncw trimeric form of dimcthylaminoborane. This has bccn confirmed by an
X-ray diffrnction Rtiidy by Trrfonas nnd Lipscomb ( 2 % ~ ) . The compound in question
is inrrt to wntrr, acid, rind mcthnnol at nmbirnt temprratures (in this reupect it is quite
compnrriblc (If?) to the N-mrt Iiylnminoboranc triniw), and is stablc in vuczio to a littlt1
over 300. With mcthnriol and hydrochloric acid, howrvrr, it rcnrts coxnpletrly a t 80-90
over a pcriod of 16 hr or lcss. Compared with other bornne derivativcs the stability of
294
Y. G . A . S T O N E
'Hie int,crtic+ion of t,riiiicthyl:tiiiiiie or certaiii ot,licr Lewis bases ivit h

p(!iit,nhorarie-11also results in symmetrical cleavage of RH, groups. Nevcrtlielcss, t,he ultimate reactions are more compliciited than t,he re:wt,ions wit.h
other boranes becaiise of t.he instat)ilit,y of pciitahorniic-11 (55). rl'lic
dcc~:~boranc-dimethylaniiiicsystein is complic:itcd. When solid JjlOH14 is
t r e : i t d with dimethyl:tmiiie, :t series of adducts containing 1, 2, or 3 niolcs
of :tinine per mole of decaborane tend to form (52). 11ec~ihor:tne:tiid pyridine :dso yield a niiniber of prodiwts ( 5 1 ) . It, appears tlittt most of these
product,s coiisist, of two pyridine molecules atttiched to u c1ec:iborane rnolccwlc which has lost oiic or two hydrogen atoms.
As a conclusion to this section it should bc meiit,ioncd that rcact,ioiis
l)et,wccndihor:tne arid scver:rI other iiitrogeri Lewis bases bcsidcs those niciit ioiied al)ovc have heen investigated. 'I'ht~se nitrogen donors include hytlrnzinc :tiid its riiethyl deriwtives ( I S O ) , hydroxylaminc :tiid its ,V-methyl
tloriv:itives (&), : ~ i i dO-i~iethylhydraxyl:tniiiicarid its N-methyl dcriv:it,ivrs
(17 ) . In gencwl, hortine adducts arc formed ('l'ablc 11) n4iic.h arc thernicilly
iiiist:ildc with rcy>c!ct, t,o rcleasc of hydrogen.
HOItOS-XIISESIC', O1t H O I t O S - r \ S T I h l O S Y
130Sl)S
I3y usiiig t)oroii hydridcs as tlic l)oron sourc.~,n iiumlwr of c.onipouiitls

niuy be madc in which horoii is Imiderl to phosphorus or its coiigencrs.
'I'hcrmal st'nhilit.ics of :inahgous hortiric tlerivat.ivcs of t,Iicse elements
dcc:rc:ise from phosphorus tanantimony in accord with t.hc genenil ohserv:it ioii that heavier clenieiits forin we:iker c o d e n t Imids th:tn do lighter
chi~ciits.Hiwever, st.rength of R h i i d between ti pnrt,icular (iroup \' clcinciit a,ntl the hoi*:tnegroup is dependent on :t numbcr of factors iricluding
the ii:iture of suhstituents on the Group V atom. 'I'lius t.hc compouiid
( U I n ) n P. 1311, is very st:rble (43) since strong donor groups are on phosphorus, nhere:rs t,lie complex FJ' HHs (86) wit,h elcctroiwgntive groups
on t,hc ligantl :itom ratdily tlissoci:itc,s inlo its cornponeiits. Xniminolysi~
of pliosphorus t,rifluoride-l,ortinc a t low txmpemtures niay he represented by
The conipouiid triniiiido-pliosphorus-borniic,(H&) J

' BI13, is a sublimd h white solid soluble in ether and liquid ammoniii. An ethane-like struc.
295
CHEMICAL REACTIVITY O F T H E BORON HYDRIDES
ture has been suggested for phosphorus trifluoride-borane with a a-bond

between boron and phosphorus (85).However, a s will be discussed in Section II,F, it is possible that the bonding may not be as simple as this.
Diborane and phosphine a t - 110" form a solid B2Hs,2PH3,dissociation pressure 200 mm a t 0" (66).Its constitution is unknown but the properties are more like those of ammonia-borane than those of the diammoniate
of diborane, so it may be a simple borane adduct H3P BH3. Diborane
reacts with biphosphine well below room temperature t o produce PzH4
B ~ B Ga ,compound thermally more stable than P2Ha 2BF3 (9). As might
be expected, diborane and phosphine react slowly at room temperature,
or more rapidly on heating, forming a polymer [BPH,], and hydrogen.
Similarly, CH3PH2* BH3 ( 4 S ) , with more than one proton bonded to the
ligand atom, releases hydrogen on heating, producing nonvolatile polymeric
materials. When adducts such as (CHJ2PH BH3 or (CH3)ZAsH BH3
are heated on the other hand, loss of hydrogen ceases after evolution of one
mole and formation of borane derivatives, [(CH3)2MVBH2],(Mv = P or
As) (4sl134). These dimethylphosphino- and dimethylarsinoboranes are
obtained principally as trimers, although some tetlameric and higher
polymeric forms are simultaneously produced. The trimers and tetramers
are volatile crystalline solids with exceedingly high thermal and hydrolytic
stability for molecules containing B-H bonds. Thus dimethylphosphinoborane trimer and tetramer are hydrolyzed only slowly by concentrated
hydrochloric acid near 300", and [(CH3)2PBH2]3is stable in vucuo to 400",
although it ignites in air a t about 250". The arsinoboranes are somewhat
less stable, as would be expected from the weaker bonding powers of arsenic.
I t is likely that these trimeric and tetrameric phosphino- and arsinoboranes
have ring structures (IV,V). Indeed, in the case of [(CH&PBH,], this has
HZ
As(CHB)P
(CH&AS
B'
HL!
been confirmed by an X-ray diffraction study (63).It has been suggested

(43) that the extraordinary stability of these compounds is due t o a novel
type of bonding in which the regular dative a-bonds are strengthened by
*-bonds. These ?r-bonds could arise through interaction of vacant d-orbitals
on phosphorus or arsenic with electron density from the B-H linkages. In
296
I!'. G. A. STONE
this nimmer the virtual 1~l)scnccof hydridic character in the B-H bands
can be accounted for. This hypothesis gains some support from the observation that in [(CH&PBH2I3the HBH angle is near 120', considerably
greater than tetrahedral, and as if B-H electrons are attracted to phosphorus above and below the puckered ring (63). Any B-Y ?r-bonding in
the cyclic phosphinoborunes should irlcrease in irnportaiice as electronegativity of the groups on phosphorus is increased. The P-B dative u-bonds,
however, would be weakened. It is interesting, therefore, that the volatile
solids [ (CF,),PBH2], and [ (CF3)2PBH&have been prepared by reaction
betweeri diborane and (CI?3)2PFor (Cl?&PH (31). These fluorocarbon
phosphinoboranes are hydrolytically much less stable than the methyl
compounds, but they are thermally stable up to 200'. Their existeiicc is
difficult to account for in terms of polymerization of (CF&PBH2 groups
merely by Y-B udative bonds because the (CF&P group is a very weak
electron pair donor. The compouod (CF&PH does not form a recognizable
adduct with borane.
A number of compounds [RJ'BH,], have now been prepared (41, I@),
some by pyrolysis of borane adducts obtained from diborarie and sceondary
phosphines, for example,
(CHI)(i-C3Hr)PH. BHa
cgrZo(CgH&-'H
-+
-31 [CHj(i-C3H7)PBHzls + Hn
. BH, + 51 lcycZo(CsH11)2PBHz13 + H2
others starting from organohalophosphines, for example,

It2PCI
+ NnI<H, -+ RtPH.
BH3 + N d J
1pyrolysiv
W'BHzh
(R = Et, fh, etc.)
The discovery (28a) that when the adduct (CHJ*P2(BH&is heated to

170-200" the phosphinoborane [(CH~)ZPBHZI,
(n = 3, 75% yield; n = 4,
15% yield) is produced suggests that, diphosphines may also provide routes
to phosphorus-boron polymers. In the above reactions the phosphinoboranes are obtained mostly as trimers with small amounts of tetrainer and
only traces of higher polymers. In the search for truly inorganic polymer
systems it would be useful if the phosphinoboranes could be obtained with
a high degree of polymerization. Recently a novel technique has been
partially successful in increasing the yield from pyrolysis of a phosphineborane of highly polymeric [R2PBH2], ( 1 4 0 ~ )When
.
a phosphine-borane
is pyrolyecd i n the presence of about ten mole per cent of a strong Lewis
297
base [for example, (C2H6):{N,(CHJsP, or (CH,),N], a significant proportion (approximately 50%) of the phosphinoboraiie produced is in the form
of high polymer. Under these conditions (CH&PH BH3, for example,
affords a brittle white polymer with a degree of polymerization of about
80 (molecular weight approximately 6000), while when CHs(C2H6)PH BH3
is heated in the presence of Lewis base, a truly plastic polymer (mp, 118126") is produced. It is probable that in these reactions the Lewis base
functions to stabilize the monomer, namely, RzPBHz N(C2H6)3, the
phosphorus atom in one such unit then bonding to the boron atom in another. It is likely that treatment of higher boranes with organophosphines
will afford interesting substances with phosphorus-boron bonds. In this
connection the compound BloHl,[ (CsH6)3P]2should be noted (69). This
substance, stable without melting t o above 300, may be prepared by
refluxing decaborane and triphenylphosphine in diethyl ether. One mole of
hydrogen is evolved per mole of compound produced. Treatment of
BloHl2(CH3CN)2with triphenylphosphine also yields BlJ12[(CsH,) 3PlZ.
The BloH12unit can be transferred from one ligand t o another, behaving
as a discrete electrondeficient species (69).
Recently some stibinoboranes have been reported (34), and as would be
expected antimony-boron bonds are considerably weaker than those between arsenic and boron. Thus (CH3)BSb * BHs exists only well below room
temperature, whereas (CH3)aAs . BH3 shows instability only above 80".
Indeed, if Lewis acid-base displacement reactions are used as a criterion
of stability, then relative stabilities of the BH3 adducts of trimethyl compounds of Group V elements are in the order: (CH3)aP BH3 > (CH3),N
BH3 > (CH3),As BHs > (CH3)sSb BH3 (135). Reaction between diborane and (CH&SbH, or better (CH3)dSbZ a t looo, yields (CH&SbBH,
(bp, est. 70") (34). This compound is monomeric. Failure of (CH3)2SbBH2
to polymerize via antimony-boron bonds is easily understood in terms of
the weak dative bonding power of antimony. However, the BH, group in
(CH3),SbBH2should then pass to a dimer form as in diborane, because
the monomer would not be stabilized by Sb-B p,-p, bonding, as is monomeric (CHJ2NBH2. Antimony is too large to form double bonds of the
p,-p, variety. I n view of this it has been suggested (34) that monomeric
(CH3),SbBH2 exists through the ability of antimony to hybridize its
filled 5p-orbital with the appropriate 5d to give a pd hybrid long enough for
effective overlap with the vacant p,-orbital of a planar boron atom.
9
D. REACTIONS
LEADINGTO FORMATION
OF BORON-OXYGEN
BONDS
There are oxygen counterparts t o many of the reactions which occur
between boranes and nitrogen compounds, but with important differences
298
P.
A. STONE
C;.
because the donor power of oxygen is less than that of nitrogen, and the
0-H bond has a greater protonic activity towards B-H than does the
N-H bond. Thus borane forms a series of etherates analogous to the
amine-boranes but with much lower stability. In the gas phase at ambient
temperatures the borane-etherates (Table 11) are completely dissociated,
but their relative stabilities as determined from volatility measurements
(49), phase studies (154), and from Raman spectra studies (96), are in the
order (CHJ4O * BH3 > (CH&O BHa > (C2H&0 BH3. Furthermore,
some of the higher horanes react with ethers in a manner analogous to their
reactions with tertiary amines. Tetraborane-10 on treatment with either
diethyl ether or tetrahydrofuran forms diborane, and the unstable solids
(C2H6)20* BsH7 and (CH2)40 B3H7, respectively (73).
It was learned long ago that boranes and methanol release hydrogen
rapidly when mixed. Moreover, if the borane is in excess an unstable
polymeric species CHIOBH~appears which subsequently disproportionates
to di- and trimethoxyboranes and diborane (38). Dimethoxyborane,
(CHSO)zBH, the oxygen arialog of [(CH&NI2BH, is the first member of a
series of monomeric3 dialkoxyboranes. These are formed when diborane
or some of the higher boranes are treated with aldehydes, ketones (16,
l 5 6 ) , or alkene oxides (156). The dialkoxyboranes disproportionate more
readily than do the aminoboranes.
E. REACTIONS
LEADING
TO FORMATION
OF BORON-SULFUR
OR BORONSELENIUM
BONDS
The effect of diborane on several sulfur compounds and one selenium
compound has been studied. No compounds from higher boranes and sulfur
compounds have yet been reported apart from an adduct (CH3),S B3H7
derived from B4H10(73, 74). The complex CHsSH BH3 may be prepared
from its components a t -78O, but even a t this temperature it begins to
lose hydrogen to make (CH3SBH2),polymer (44). In contrast to compounds
like (R2PBH2)3and [ (CH3)1A~HH2]3,
polymer bonding in (CH3SBH2), is
weak. This is shown by the readiness with which (CH3SBH2),yields a series
of low polymers when heated over its melting range of 65-80", and more
especially by removal of CH3SBH2groups, with formation of a complex
(CH3)aN * BH2SCH3, when (CH3SBH2), is treated with trimethylamine.
The adduct (CH3)3N BH2SCH3reacts with a stream of diborane at elevated temperatures to form (CH3)3N BH3 and the interesting derivative
CH3SB2H6.The latter very probably has a sulfur-bridge structure analogous
to the nitrogen-bridged aminodiboranes, but it is much less stable than
these compounds, easily reverting to diborane and (CH3SBH2),.
With dimethyl sulfide and dimethyl selenide, diborane forms well-de-
299
fined BHa complexes (Table 11). The compound (CH3),S BH3 is unusual
in being more stable than either (CH&S BF3or (CH3)20 BH3. Similarly
(CHz),S BH3 is more stable than (CHz)4S BF3 or (CH2),0 . BH3, and
(C2H&S BH3 is more stable than (C2H&S * BF3 or (C2H&0 BH3
(49, 58). In forming more stable complexes with oxygen donors than with
sulfur donors, boron trifluoride parallels the behavior of trimethylaluminum
(50) and trimethylgallium (48). By exhibiting the reverse behavior, borane
differs from most of those Group I11 acceptor molecules which have been
studied. Furthermore, on the basis of electronegativity effects alone it
would be expected that boron trifluoride would always form more stable
complexes than borane. However, there is considerable evidence suggesting that the acceptor power of boron trifluoride is reduced by F-B dative
a-bonding. Thus heats of formation of the trimethylamine adducts of boron
trifluoride and borane, and of the pyridine adducts of these two acids are
very similar (58). Nevertheless, with heavier donor atoms (P, As, Sb, S,
and Se) borane forms much stabler complexes than does boron trifluoride,
even though it is necessary to supply considerable energy to diborane to
break its bridge bonds to supply borane groups. These observations have
been interpreted in terms of there being something unusual about the
borane group. Perhaps the atomic orbitals of the three hydrogen atoms of
borane can combine, thereby forming a p,-like orbital which can then overlap with an empty orbital of a ligand. There would thus be a a-bond
strengthening the a-bond in certain borane complexes, but not in those
involving oxygen or nitrogen since these elements possess no low-lying
vacant orbitals (58). The effect with BH3 would be something like the
hyperconjugation of organic chemistry except that with borane the
effect would be greater due to the charge separation in the dative a-bond.
Some support for this idea comes from a study of the infrared spectrum
of (CH3)2S. BH3 (94). The boron-hydrogen stretching frequencies in
(CH&S BH3are higher than the corresponding frequencies in the boraneetherates, indicating greater charge transfer to boron in the latter, yet
these are much weaker complexes than their sulfur analogs. However,
supplementary a-bonding is in the reverse direction to the donor a-bond,
so the total effect in dimethylsulfide-borane could be less negative charge
on boron resulting in a higher B-H stretching frequency.
F. REACTIONS
LEADING
TO FORMATION
OF BORON-CARBON
BONDS
A number of reactions of the boranes lead to synthesis of boron-carbon
bonds. With boron alkyls the boron hydrides undergo important alkylation
reactions, but only in the case of diborane have these reactions been
described in detail. When diborane is mixed with trimethylborane at room
300
F. G . A. STONE
temperature, a mixture of methyldihoranes is obtained (116).These methyldiboranes readily exchange methyl groups with each other, and the interconversions have recently been studied kinetically using gas partition
friictornctry (138). Unlike its isomer 1,l-dimethyldiborane, the compound
1,2-dimethyldiborane is hardly present a t all in an equilibrium mixture of
methyldihoranes. It may be prepared by removal of a borane group from
CH3BzH,with tetrahydrofuran (SO),
CHjB2Hs
+ (CH2)rO
(CH2)rO. UH,
+ 31 CH,HB(HZ)BHCIT~.
1,2-Dimcthyldiborane has been used to prepare a number of compounds

wherein boron is bonded to one methyl group; for example, (CH&NBHCH,
(30) and (CHaO)zBCH3 (25).
Besides the methyl compounds a number of other alkyldiboranes have
been prepared from diborane and trialkylboranes (111, 128). The alkylated
diboranes can also be obtained by treating at elevated temperatures
NaRH4, LiRH4, or LiAlH4 with R3B, arid cithcr LL hydrogen halide or a
boron trihdide (8%).
Even when trimethylborane is taken in great excess for reaction with
diborane, only four hydrogen atoms of the hydride are replaced by methyl
groups. Tetramethyldiborane is fairly easily removed by fractional condensation from the reaction mixture. Since it behaves chemically as a source
of (CH&BH fragments] it has been used to develop a considerable chemistry of the (CH&B-group, for example,
+ 31 [(CHJ)~BH]Z C H ~ J H. BH(CIT& + CHsSB(CH& + Hz,

(CHj)zNH + [(CH3)zBH]z+ (CHahNH. BH(CH&+ (CH&NB(CHs)z + Hz,
CHjHH
(44)
(29)
With sodium in liquid ammonia the two halves of a molecule of tetramethyldiborane behave differently. One half forms H,N BH(CH&,
which is converted to (CH&BNHY and Hz, and the other half forms a
salt Nat2HB(CH& (52). The latter when trcated with trimethylborane in
liquid ammonia forms Na2HH2(CH3)s
which very likely contains a boronboron bond. When pure, both NazHB(CHaj2and Na2HB2(CH3)6
are stablc
in oucuo to about 100. A calcium compound CaHB(CH3j2 NH, has also
becn prepared (4S), but in contrast to its sodium analog it does not act as
a Lewis base towards trimethylborane.
Few boron-carbon bonded derivatives of the higher boranes have been
reported, but well-authenticated alkyl- and bcnzyl-decaboranes have been
prepared from the remarkable Decaborane Grignard (54, 126).
+
ether
I l i o H ~ CHxMgI -+
BioHisMgI 4-2CeHaCHzCI --+
BloHlaMgI
+ CH4
CsHbCHzBloHls
+ MgC12 + GH5CHzI.
30 1
CHEMICAL REACTIVITY O F THE BORON HYDRIDES
Best yields of alkyl-substituted decaboranes are obtained from reactions

between the decaborane Grignard and alkyl fluorides rather than from
decaborarie Grignard and other alkyl halides.
The compound borane-carbonyl occupies a special position in any discussion of molecules derived from boron hydrides with boron-carbon bonds.
Carbon monoxide reacts with diborane (20 atm/room temperature) to
give an adduct of the borane group which does not rearrange or lose hydrogen (39).The adduct OC * BH, does, however, dissociate into carbon monoxide and diborane a t ordinary temperatures. No carbonyls of other
Group I11 acceptor molecules have been prepared so far (133), and it has
been suggested (2'7, 58) that borane-carbonyl owes its existence t o the ability of the hydrogen atoms of borane to transfer electrons to a vacant
pT-orbital in carbon monoxide.
The behavior of carbon monoxide toward borane would, therefore, be
similar t o its behavior towards metals in the metal carbonyls in which
multiple bonding between the metal and carbon monoxide groups has long
been postulated. The carbon monoxide groups of metal carbonyls may
sometimes be replaced by phosphorus trifluoride groups. It has been suggested (4'7) that this is possible because phosphorus trifluoride functions
as a ligand like carbon monoxide, that is, PF, can donate electrons in a
dative a-bond, as well as accept them in a r-bond, since phosphorus has
vacant 3d-orbitals. It is therefore, interesting that the phosphorus trifluoride complex of borane has been prepared (85) (see above). Its existence
could be due t o phosphorus-boron multiple bondinglgalthough it should be
recognized that there is no general agreement on this idea. Nevertheless,
the concept of the H,B group being able to release electron density t o a
ligand with vacant orbitals as well as to accept it in a classical a-dative
bond accords well with many experimental facts (58).
Diborane is not unique among the boron hydrides in its ability to form
a carbonyl. High pressure reaction of either pentaborane-11 or tetraborane-10 with carbon monoxide forms a substance which behaves like
borane-carbonyl in its manner of decomposition. The formula of this
polyborane-carbonyl has recently been established as OC B4H, (mp,
- 114.5"; bp, 59.6") ( 4 1 ~ )It. reacts with trimethylamine without release
of carbon monoxide.
A different approach to the synthesis of boron-carbon bonds from the
boranes involves addition of B-H bonds to carbon-carbon double bonds.
9 Boron trifluoride does not form a PF3 adduct. However, even though PF3is a very
weak electron pair donor, by using a strong enough acceptor atom (boron is rather weak
in this respect compared with aluminum or gallium) it might well be possible to make a
PF, adduct of a Group IT1 acceptor molecule in which the dative bonding could be
described in terms of a classical u-bond.
302
F. G . A . STONE
The reaction
proceeds slowly in the gas phase at ordinary temperatures, but more rapidly
a t 80" (76).
Recently it has been found that the speed of addition of dibontne t o
olefiris is remarkably increased if an ether is present (21). In spite of its
great speed this hydroboration reaction in ether is fairly selective. Thus
treatment of an equimolar mixture of 1- and 2-hexene with a deficiency
of diborane, followed by refluxing, yielded tri-n-hexylborane. Under the
influence of heat the organoborane from 2-hexene isomerized into tri-nhexylborarie (22). Similarly a mixture of 2-, 3-, 4-, and &decerles treated
with diborane in ether, heated and then subsequently oxidized gave an
80% yield of ldecanol. Diborane may thus be used to transform olefins
into alcohols.
A logical extension of the ether-catalyzed diborane-olefin reaction was
the discovery that amineboranes, for example, (CH3)aN BH3 or C6HBN *
B H , and olefins react to yield triorganoboranes ( l a , 66, 81, 118).
Reactions between diborane and molecules containing carbon-carbon
pr-pr bonds do not always involve simple addition of B-H to the double
bond. Thus the principal volatile products of the gas-phase diboranevinylsilane reaction are silane and hydrogen. In the presence of an ether,
however, the silyl group is not cleaved from vinylsilane by borane and
addition of B-H to the carbon-carbon bond does take place. Similarly
in Pyrex bulbs at SO" the volatile products of the reaction between diboiane
arid perfluoropropene are CF4 and SiF4, the latter formed presumably by
hydrogen fluoride attack on the reaction vessel (129). How varied the effect
of diborane on an organic molecule can be is illustrated by the action of
the hydride on fluoroethylenes such as vinyl fluoride. See Fig. 1. In the
iibsence of ether, all fluoroethylenes with diborane give the same mixture
of products (4). The main reaction products are boron trifluoride, ethyldifluoroborane, diethylfluoroborane, and triethylborane. Composition of
the mixture, however, depends markedly on the fluoroethylene taken for
reaction. With tetrafluoroethylene distribution of boron in the volatile
products is in the order: BF3 > C&6BF2 >> (C2H6),BF> (C2H6)aB,while
with vinyl fluoride it is: CzHsBFz> (CzHs)2BF>> BFI > (CzH6)3B. The
products may be accounted for by the reaction scheme outlined (Fig. 1) (4).
In the presence of ether, however, preliminary results indicate that borane
adds to the double bond of vinyl fluoride t o some extent without rupture
of the carbon-fluorine bond (129). In this manner organoboranes with
fluorine atoms in the side chain may be obtained in small quantity. In the
absence of ether, in a manner similar to the reaction between diborane
CHEMICBL IZEACTIVITY OF THE BORON HYDRIDES
303
and vinyl fluoride, treatment of vinyl chloride with diborane yields ethyldichloroborane, diethylchloroborane, triethylborane, and boron trichloride (129). In this reaction, however, significant quantities of chloroB H 3 / r
CFz:CFH + BHzF
'1
CF2CFH+BHFz
/ BH3
CFHCFHt BH$
BHzF
blymer
CF2CFH t BF3
BHz\
(CFHCFHt BHFz
CFH:CFH+BF3
or
or
(C2H5I2BFt CzF3H etc.
(CzH5)3B
FIG.1. The effect of diborane on fluoroethylenes.
ethylboranes are also formed, demonstrating that addition of borane to the

double bond of the vinyl halide is almost as important a process as reduction
of the vinyl halide to ethylene.
When diborane is led into vinyl chloride in ethereal solvents like diglyme, tetrahydrofuran, and diethyl ether, thermally unstable mixtures
are obtained which defy separation (6'7). In an attempt to simplify this
behavior, vinyl chloride in dimethyl ether was treated with diborane a t
304
B'.
L;. A . S'lWNX
- 80". Again, however, in working up the products exothermic decomposition often occurred, but it was possible to isolate in low yield 8-chloroethylboron dichloride as its crystalline dimethyl etherate. I n a similar manner,
diborane and ally1 chloride in dicthyl ether afford a mixture of tri-y(chloropropy1)-borane and di-7-(chloropropy1)-boron chloride. Hydrolysis
of these chloropropyl horzlncs yields cyclopropane.
It is probable that a whole new area of boron chemistry will result from
compounds obtained from reac tioris betwceii the higher borancs arid compounds containing carbon-carbon double bonds. In this respect discovery
of the new compound dimcthylenetetrahorane, C2H4B4H~,
is particularly
intriguing (64). This compound, which may be formed in 70% yield from
tctraborarie and ethylene, appears to have a cyclic bridged structure in
which one hydrogen atom of each of the two BH2 groups in the tctruborarie
structure has been lost, the two boron atoms then being bonded to each
other by a -CH2CH2-- group.
G. RXACTIONS
LEADINGTO
~ ~ O I i M I 4 T I O OF
N
BORON-HALOGEN
BONDS
This section of this article would not he complete wit,hoiit some mention
of reactions of the boranes and their derivatives which lead to boronhalogen bond synthesis. It was noted above that reactions between diborane
and halogen-substituted ethylenes led to boron-halogen as well as boroncarbon bond synthesis. In addition to these reactions, and preceding them
chronologically, it was observed that several of the more stable boranes
(hH6, &Hs, RloHla) under cmefully controlled conditions undergo substitution reactions with halogens, hydrogen halides, or boron halides.
Usually only one or two of the hydrogen atoms of the borane can be replaced by halogen without breakdown of the hydride structure. Compounds
prepared directly from the boranes and having both boron-hydrogen and
boron-halogen bonds include H2H6RP, R2H&I, B6H8Cl, BsHsBr, B~HsT,
BIOHIJ (two isoincrs), B10H13Br,and BIOH& (two isomers). The structures of several of these compounds have been reviewed by Lipscomb (85).
The compound B,H,CI can only be isolated with difficulty since it is
transformed into RC13 and B2H6 rapidly at room temperature, disproportionatimg much more readily than does BBHKBr.
Indeed, mixtures of boron
trichloride and diborane at ambient temperatures exhibit pressures which
are additive for the two components. In the presence of ethers, however,
diborane and boron trichloride react readily a t room temperature to form
chloroborane etherates (24);
+
+
BzHs BClj
BaTTs 4BCls
+ 3Rz0 + 3R20.BHzCl.
+ GRzO tjRzO BHCl2.
-+
305
These chloroborane etherates can also be prepared from alkali metal borohydrides (Section 111).
For the preparation of unusual boron compounds the hydrogen atoms
of certain derivatives of borane can also be replaced by halogen; for
example,
[ ( C H ~ ) Z P B H ~ I ~ GHF
anhydrous
[(CHs)zPBH&
(CHs),N.BHa
+ GCHaX
HC1
anhydrous
[(CHs)zPBF&
AlXi
+ 8Hz,
(140)
+ 6CH4,
(140)
+ Hz.
(24
[(CHo)zPBXzla
(X = C1, Br, I)
-+
--t
(CFI3)oN * BHzCl
Ill. Recent Advances in the Chemistry of the Borohydrides
Preparation of the first borohydrides, those of aluminum, beryllium,

and lithium, by Schlesinger, H. C. Brown, Burg, and Sanderson (40, 106,
114, 116), suggested the existence of similar compounds of other elements.
This field was rapidly extended by studies related t o the search for solid
sources of hydrogen gas, volatile uranium compounds, and new methods
for preparing diborane (107).
Further stimulus came on the discovery that the borohydrides were
very useful reducing agents in inorganic and organic chemistry. Borohydrides of some 35 metals have now been reported in the literature,
including those of elements of Groups I, 11, and 111, those of several subgroup elements (for example, Zn, Cu, Ti, Cr, Mn, etc.), those of several
lanthanide elements, and those of a few of the transuranic elements. A still
more recent development has been a report of triple hydrides (for example,
NaA1(BH4)4,Ca[A1(BH4)4]2)by Wiberg and co-workers (147).
Borohydrides vary greatly in volatility, arid in such properties as thermal stability, ease of oxidation or hydrolysis. Decreasing thermal stability
and ease of oxidation correlate roughly with increasing electronegativity
of the metal. Thus sodium borohydride is stable to 300" in dry air, and to
400" in vucuo, whereas titanium(II1) borohydride decomposes completely
within a few days a t ordinary temperatures. Some borohydrides, such as
those of copper and silver, have been observed only at low temperatures.
Potassium and sodium borohydrides can be crystallized from cold water,
whereas aluminum borohydride is hydrolyzed explosively. Sodium borohydride, having a face-centered cubic structure (l27), is essentially ionic,
but volatile borohydrides like AI(BHJ3, (CH3)&aBH4, and U(BH4),
represent fairly nonpolar situations, with varying degrees of covalent
bonding of metal ion to BHd- groups via three center bonds like those in
diborane.
306
F. G. A. STONE
Sodium borohydridc, now curnmcrcittlly available, may be prepared

without recourse t o diborane by the reaction (108):
4NaH
+ B(OCFI&-
225-275'
+ NaBHd
+ 3NaOCHa.
Although alternative synthetic routes are available for some of thc borobydrides, most can be prepared from sodium or lithium borohydrides by
metathesis; for example,
FeCh
+ 3LiBHr
AlCls
-45"
3LiCl
ether
+ 31 Hz + 1 BZHB+ Fe(BH&,
decomposes
at 0"
+ 3NeBH4 ____ + 3NnCI + AI(BIId)S,
LiCl
+ NaRHd
---+ CHaONa + CsBH,,
(7 0.9)
bp 44.9
NaCl
+ Idi13H4,
( I 09)
CHaOH
CH~OCS N ~ L B H ~
[Cr(NH&]F,
(1as)
heat
+ 3NaBH4
liq. NH3
3NaF
-45'
+ [Cr(NH&I(BH&. 21 NHs.
(3)
(87)
Under suitable conditions boron trifluoride will liberate diborane from

lithium or sodium borohydrides according to the equation (18):
3MBHa
+ 4BF1
--j
3MBP4
+ 2BzZIn.
Although several other methods for preparing the simplest borane are
known (132),this method represents a convenient laboratory route to the
hydride. However, when sodium borohydride is used, the above equation
does not completely represent the course of reaction unless sodium borohydride is added in increments to boron trifluoride, both reagents being
dissolved in djglyme.l0 If a deficiency of boron trifluoride is added t o sodium
borohydride dissolved in diglyme, the reaction,
NaBHd
+ .j1 BzHs dklyme

-+
NaBHI . BHs
occurs (as), yielding the interesting new substance NaBzH7. Only on addition of excess of boron trifluoride are considerable amounts of diborane
liberated. Similarly if boron trichloride is used to form diborane from sodium borohydride, only stoichiometric quantities of reagents (BCl, :NaRH4,
1 :3) give quantitative yields. As in the analogous boron trifluoride reaction
n deficiency of the Lewis acid gives low yields of dibornne. Furthermore,
10 Sodium borohydride is insoluble in the ordinary et,hers but is soluble in diglyme,
the dimethyl ether of diethylene glycol.
C H E M I C A L R E A C T I V I T Y O F THE BORON HYDRIDES
307
unlike the boron trifluoride situation, use of an excess of boron trichloride

also decreases the yield. This is because ether solutions of boron trichloride
absorb diborane to form a new class of compound, the mono- and dichloroborane etherates (24) (See Section I1,G). The boron trichloride-sodium
borohydride system can be represented by the equations,
+
+
+
7NaBH4 BC13 -+ 3NaCl+ NaBzH7,

3NaBHd BCla 43NaCl 2BzH6,
NaBH,, BCl, RzO -+ NaCl 2Rz0. BA2CI.
+
+
Sodium borohydride is useful in the preparation of hydrides of several

elements, for example,
GeOz
NaBHd
n+
GeH4.
Lithium borohydride has also found some use in this respect, for example,
(159)
CeHsMClz 2LiBHd 4 C~ILMH2 B&

(M = P, As, Sh)
+ 2LiCl
In aqueous solution many metal ions are reduced by sodium borohydride

to borides [for example, Ni(I1) and Co(II)], to the metal [for example,
Ag(I)], or t o a lower oxidation state [for example, Ce(1V) converted to
Ce(III)]. Some of these reactions have found analytical application; for
example, analysis of iron, wherein NaBH4 replaces SnC12 or the Jones
Reductor.
The borohydrides have found their greatest use as reducing agents in
organic chemistry (56). Reduction products are obtained in nearly quantitative yield and the reactions are highly selective. Reducing power of the
borohydride group depends on the nature of the associated cation, on the
solvent, and on the presence or absence of other substances (19). Thus
sodium borohydride does not normally reduce olefinic double bonds, but
in diglyme solution in the presence of aluminum chloride reduction readily
takes place, thereby providing a route to triorganoboranes, or even alcohols
by subsequent oxidation of the borane, followed b y hydrolysis of the boron
ester (20). Aluminum borohydride may be an intermediate in this reaction,
but if so it is not present in the sodium borohydride-aluminum chloride
reagent in more than trace amounts. Table IV illustrates the selectivity
of sodium borohydride and related substances as reducing agents in organic
chemistry. In the construction of Table IV no attempt has been made to
take solvent effects into account. Choice of solvent sometimes determines
308
F. G . A. STONE
whether reduction takes place or not. Suitable solvents for sodium horohydride are diglyme, isopropyl alcohol, or even water or methanol. Lithium
borohydride and lithium aluminum hydride may be used in a variety of
etherfi, but not in water or alcohols since they react vigorously with these
substances.
Finally mention should be made of the existence of a variety of substituted borohydrides, as well as of the existence of aluminohydrides.
Many of these suhstances, especially lithium aluminum hydride, have,
TABLE IV
SODIUMBOROHYURIDE
A N D RELATED
COMPOUNDS
A S SELECTIVE
REDUCING
AGENTS
IN ORU.4NlC CHEMISTRY
++ Iteacts rapidly, + Reacts a t
H(:HO
RR'CO
RCOCl
RCOOR'
RCOOII
RCOOM
HCHzCl
RCN
RN0z
RCH=CI12
++
++
++
at 25-100"
ti
++
++
++
+
+
modcrate rate,
++
++
++
++
++
- Reacts hlowly or ncgligibly.
++
++
f+
+-
++
++
+
++
-
-ti-
++
++
++
++
++
++
++
++
+
like the parent borohydrides, revolutionized the preparat,ive techniques

of organic and inorganic chemistry. Examples of substitutcd borohydrides
include trialkoxyborohydridcs, NaBH(0R) 3 (14, 15), sodium triphenylborohydride, NaRH(CsH6)3(156),lithium phenylborohydride, LiB(CsHs)H3
(150),and sodium triformatoborohydridc, NaBH(02CH)3(141, 143).
IV. Conclusion
This review, concerned with reactivity of the boranes with no attempt

to make the references exhaustive, has demonstrated the current and intense activity in this area of chemistry. The reader may already have
recognized that nearly 60% of the references quoted refer to work published
in this field during the last five years. With the application of new methods
of preparation and separation, it is evident that boron compounds will be
described in increasing numbers. Meanwhile, it is hoped that the subject
matter reviewed here will be of help to research workers and teachers alike.
CHEMICAL EEACTIVITY OF THE BOKON HYDRIDES
309
REFERENCES
R. D., Carter, J. C., and Taylor, R. C., J . Am. Chem. SOC.

81, 3550 (1959).
l a . Ashby, E. C., J . Am. Chem. SOC.81, 4791 11959).
2. Badcn, IT. C., Bauer, W. H., and Wiberly, S. E., J . Phys. Chem. 62, 331 (1958).
3. Banus, M. D., Bragdon, R. W., and Hinckley, A. A., J . Am. C h m . Soc. 76, 3848
(1954).
4. Bartocha, B., Graham, W. A. G., and Stone, F. G. A., J . Inorg. & Nuclear Chem.
6, 119 (1958).
5. Bauer, S. H., J . Am. Chem. Soc. 78, 5775 (1956).
6. Bauer, S. H., Encyclopedia Chem. Tech. Suppl. 1, 103 (1957).
7. Beachell, H. C., and Meeker, T. R., J . Am. Chem. Soc. 78, 1790 (1956).
8. Beachell, H. C., and Schar, W. C., J . Am. Chem. SOC.80, 2943 (1958).
9. Beichl, G. J., and Evers, E. C., J . Am. Chem. Soc. 80, 5344 (1958).
10. Bell, R. P., and EmelBus, H. J., Quart. Revs. 2, 132 (1948).
11. Bissot, T. C., Campbell, D. H., and Parry, R.W., J . Am. Chem. Soc. 80, 1868
(1958).
12. Bissot, T. C., and Parry, R. W., J . Am. Chem. Soc. 77, 3481 (1955).
13. Boone, J. L., and Burg, A. B., J . Am. Chem. Soc. 80, 1519 (1958).
13a. Boonc, J. L., and Burg, A. B., J . Am. Chem. SOC.81, 1766 (1959).
13b. Brown, H. C., and Domash, L., J . Am. Chem. Soc. 78, 5384 (1956).
14. Brown, H. C., and Mead, E. J., J . Am. Chem. Soc. 76, 6263 (1953).
15. Brown, H. C., Mead, E. J., and Shoaf, C. J., J . Am. Chem. SOC.78, 3616 (1956).
16. Brown, H. C., Schlesinger, H. I., and Burg, A. B., J . Am. Chem. SOC.
61, 673 (1939).
17. Brown, H. C., Schlesingcr, H. I., and Cardon, S. Z., J . Am. Cliem. SOC.64, 328
(1942).
18. Brown, H. C., Schlesinger, H. I., Gilbreath, J. R., Kats, J. J., J . Am. Chem. SOC.
76, 195 (1953).
19. Brown, H. C., and Subba Rao, B. C., J . Am. Chem. Soc. 78, 2582 (1956).
20. Brown, H. C., and Subba Rao, B. C., J . Am. Chem. Soc. 78, 5694 (1958).
21. Brown, H. C., and Subba Rao, 3. C., J . Org. Chem. 22, 1136 (1958).
22. Brown, H. C., and Subba Rao, B. C., J . Org. Chem. 22, 1136 (1958).
23. Brown, H. C., and Tierney, P. A., J . Am. Chem. Soe. 80, 1552 (1958).
24. Brown, H. C., and Ticrney, P. A,, J . Znorg. & Nuclear Chem. 9, 51 (1959).
25. Burg, A. B., unpublished observationn.
26. Burg, A. B., J . Am. Chem. SOC.69, 747 (1947).
27. Burg, A. B., Record Chem. Progr. (Kresge-Hooker Sci. Lib.) 16, 159 (1954).
28. Burg, A. B., J. Am. Chem. SOC.79, 2129 (1957).
%a. Burg, A. B., J . Inorg. & Nuclear Chem. 11, 258 (1959).
29. Burg, A. B., and Banns, J., J . Am. Chem. Soc. 76, 3903 (1954).
30. Burg, A. R., and Boone, J. L., J . Am. Chem. SOC.78, 1521 (1956).
31. Burg, A. B., and Brendcl, G., J . Am. Chem. Soc. 80, 3198 (1958).
32. Burg, A. B., and Campbell, G. W., J . Am. Chem. Soc. 74, 3744 (1952).
33. Burg, A. B., and Good, C. D., J . Inorg. & Nuclear Chem. 2, 237 (1956).
34. Burg, A. B., and Grant, L. R., J . Am. Chem. Soc. 81, 1 (1959).
36. Burg, A. B., and Kuljian, E. S., J. Am. Chem. Soc. 72, 3103 (1950).
36. Burg, A. B., and Randolph, C. L., J . Am. Chem. SOC.71, 3451 (1949).
37. Burg, A. B., and Randolph, C. L., J . Am. Chem. SOC.73, 953 (1951).
38. Burg, A. B., and Schlesingcr, H. I., J . Am. Chern. SOC.66, 4020 (1933).
1. Alton, E. R., Brown,
310
F. G. A . STONE:
A. B., and H(:hl(:singcr, 11. I., J. Am . Chctn. Soc. 69, 780 (1937).
40. Burg, A. 13., and Ychlcsinger, 1. I., J . Am. Chem. SOC.62, 3425 (1940).
41. Burg, A. B., Slota, P. J., and Mahlcr, W., Abstr. Papers 129th Metting Am. Chern.
Soc., Dallas p. 4Q (1956).
41a. Burg, A. B., and Spielman, J. R., J . Am. Chem. SOC.81, 3479 (1959).
42. Burg, A. B., and Stone, F. G. A., J . Am. Chem. SOC.76, 228 (1953).
43. Burg, A. B., and Wagner, R. I., J . Am. Chem. Soc. 76, 3872 (1953).
44. Burg, A. B., and Wagncr, R. I., J . A m . Chem. Suc. 76, 3307 (1954).
45. Campbell, D. I-I., Birsot, T. C., and Parry, R. W., J . Am. Chem ROC. 80, 1549
(1958).
46. Campbell, G. W., J . Am. Chern. Soc. 79, 4023 (1957).
46a. Campbell, G. W., and Johnson, L., J . Am. Chem. SOC.81, 3800 (1959).
47. Chatt, J., and Williams, A. A., J . Chem. SOC.p. 3061 (1951).
48. Coates, G. E., J . Chem. ASOC.
p. 2003 (1951).
49. Coyle, 1.D., Kacsz, H. D., and Stone, F. G. A., J . Am. Chem. SOC.81,2989 (1959).
60. Davi&ofi, N., and Brown, H. C., J . Am. Chem. SOC.64, 316 (1942).
60a. EmelBus, 11. J., and Videla, G . J., J. Chem. Soc. p. 1306 (1959).
61. Fetter, N. R., and Burkardt, L. A., Abstr. Papers 136th Meettng Am. Chem. Soc.,
Boston p. 45M (1959).
56. Fitch, S. J., and Laubcngaycr, A. W., J . Am. Chem. Sac. 80, 5911 (1958).
63. Forsyth, M. W., Hough, W. V., Ford, M. D., Hefferan, G. T., andEdwards, L. J.,
Abstr. Papers lS6th Meeting Am. Chem. Suc., Boston p. 40M (1959).
64. Gallaghan, J., and Siegel, B., J . Am. Chem. SOC.81, 504 (1959).
5t. Gamble, E. L., and Gilniont, P., J . Am. Chem. SOC.62, 717 (1940).
66. Gaylord, N. G., Reduction with Complex Metal Hydridw. Interscience, New
York, 1956.
66a. Gibbins, S. G., and Shapiro, I., J . Chem. Phys. 30, 1483 (1959).
67. Bmelins Hnndb. Arwrg. ChenL. Bor, 13 (1954).
68. Graham, W. A . G., and Stone, F. G. A., J . Inorg. & N,ucleai Chern. 3, 104 (lC56).
69. Graham, W. A. G., andStone, F. G. A., Chem. & Ind. p. 1096 (1957).
60. Groszos, S. J., and Stafiej, S. F., J . Am. Chem. Sac. 80, 1357 (1958).
61. Gutcr, G. A., and Schaeffer, G. W., J . Am. Chem. SOC.78, 3546 (1956).
62. Guter, C. A., and Schaeffcr, G. W., Abstr. Papers l3lst Meeting Am. Chem. Soc.,
Miami p. 3R (1957).
83. Hamilton, W. C., A d a Cryst. 8, 199 (1955).
64. Harrison, n. C., Solomon, I. J., Hites, R. D., and Klein, M. J., Abstr. Papers
135th Meeting Am. Chem. SOC.,Boston p. 38M (1959).
65. Hawthorne, M. F.,J . Am. Chem. Sac. 80, 3480 (1958).
86. Hawthorne, M. F., J . Org. Chem. 23, 1788 (19.58).
67. Hawthorne, M. F., and Dupont, J. A,, J . Am. Chem. Sac. 80, 5830 (1958).
58. Hawthorne, M. F., and Miller, J. J., J. Am. Chem. SOC.80, 754 (1958).
69. Hawthorne, M. F., and Pitochelli, A. R., J . Am. Chem. Sac. 80, ti685 (1958).
70. Hedberg, K., and Stosick, A. J., J . Am. Chem. Soc. 74, 954 (1952).
7 i . Hewitt, F., and Holliday, A. K., J . Chem. SOC.
p. 530 (1953).
72. Hough, W. V., Edwards, L. J., and McElroy, A. D., J . Am. Chern. Soc. 80, 1828
(1958).
73. Hough, W.
Ford, M. D., and Edwards, L. J., A b s t ~Pupers
.
132nd Meettag Am.
Chem. Soc., New Yorlc p. 15N (1957).
74. Hough, W. V., Ford, M. D., and Edwards, 1,. J., Abstr. Papers 135th Meeting Am.
Chem. SOC., BOStWtk p. 38M (1959).
$9. Burg,
v.,
CHEMICAL HEACTIVITY ON THE BORON HYDRIUES
311
76. Hough, W . V., Schaeffer, G. W., Dzurus, M., and Stewart, A. C., J. Am. Chem.
Soc. 77, 864 (1955).
76. Hurd, D. T., J. Am. Chem. Soc. 70, 2053 (1948).
77. Kodama, G., Dissertation Abstr. 19, 435 (1958).
78. Kodama, G.,and Parry, R. W., J. Am. Chem. SOC.79, 1007 (1957).
79. Kodama, G., and Parry, R. W., Abstr. Papers 135th Meeting Am. Chem. Soc.,
Boston p. 3 3 M (1959).
7Ba. Kodama, G., Parry, R. W., and Carter, J. C., J . Am. Chem. SOC.81, 3534 (1959).
80. Koster, R., Angm. Chem. 69, 94 (1957).
81. Koster, R., Angew. Chem. 69, 684 (1957).
82. Laubengayer, A. W., and Brown, C. A., J. Am. Chem. Soc. 77, 3699 (1955).
83. Lipscomb, W. N., Advances in Inm-g. Chem. Radiochem. 1, 146 (1959).
83a. Lipscomb, W. N., J . Inorg. & Nuclear Chem. 11, 1 (1959).
83b. Long, L. H., and Wallbridge, M. G. H., Chem. d? Ind. p. 295 (1959).
8%. Martin, D. It., J. Chem. Educ. 36, 208 (1959).
84. MrCoy, R.E., and Baiicr, S. H., J . Am. ChPm. SOC.78,2061 (1956).
84a. Nordman, C.E., and Peters, C. R., J . Am. Chem. SOC.81,3551 (1959).
86. Parry, R.W., and Bissot, T. C., J . Am. Chem. Soc. 78, 1524 (1956).
81ia. Parry, R. W., and Edwards, L. J., J . Am. Chem. Soc. 81, 3554 (1959).
86. Parry R. W., and Kodama, G., 16th Congr. Intern. Union Pure and Appl. Chern.,
Paris p. 483 (1957).
87. Parry, R. W., Schultz, D. R., and Girardot, P. R., J . Am. Chem. Sac. 80, 1 (1958).
88. Parry, R. W., Schultz, D. R., Shore, S. G., and Kodama, G., Abstr. Papers 133rd
Meeting Am. Chem. SOC.,
Sun Francisco p . 23L (1958).
89. Parry, R. W., and Shore, S. G., J . Am. Chem. Soc. 80, 15 (1958).
90. Patterson, A. M . , Chem. Eng. News 34, 560 (1956).
91. Price, F. P., J . Am. Chem. Soc. 72, 5361 (1950).
96. Price, F. P., J . Am. Chem. SOC.73, 2141 (1951).
93. Rector, C. W., Schaeffer, G. W., and Platt, J. R., J . Chem. Phys. 17, 460 (1949).
94. Rice, B.,Galiano, R. J., and Lehmann, W. J., J . Phys. Chem. 61, 1222 (1957).
96. Rice, B., Livasy, J., and Schaeffer, G. W., J . Am. Chem. SOC.77, 2750 (1955).
96. Rice, B., and Uchida, H. S., J . Phys. Chem. 69, 650 (1955).
97. Roth, W., and Rauer, W. H., J. Phys. Chem. 60, 639 (1956).
98. Ruigh, W. L., et al., Research on Boron Polymers, WADC Tech. Rept. 55-26,
Parts I-IV. Wright Air Devclopment Center, Wright-Patterson Air Force Base,
Ohio.
99. Ryschkewitsch, G. E., Harris, J. J., and Sislcr, H. H., J . Am. Chem. Soc. 80, 4515
(1958).
100. Schaeffer, G. W., A M - . Papers 133rd Meeting Am. Chem. SOC.,Sun Francisco
p. 2L (1958).
101. Schaeffer, G.W., Adams, M. D., and Koenig, F. J., J . Am. Chem. SOC.78, 725
(1956).
102. Schaeffer, G.W., and Emilius, M., J . Am. Chem. Soc. 76, 1203 (1954).
103. Schaeffer, G. W., Roscoe, J. S., and Stewart, A. C., J . Am. Chm. SOC.78, 729
(1956).
104. Schacffer, G. W., Schacffcr, R., and Schlesinger, H. I., J . Am. Chem. Soc. 73, 1612
(1951).
106. Schaeffer, R., Steindler, M., Hohnstedt, L., Smith, H. S., Eddy, L. R., and Schlesinger, H. I., J. Am. Chem. SOC.76, 3303 (1954).
106. Schlcsinger, H. I., and Brown, H. C., .I. Am. Chem. Soc. 62, 3429 (1940).
312
b. 0. A . EITONE
107. Bohltrsingcr, H. I., Brown, H. C., et al., J . Am. Chem. SOC.78, 186 (1953).
108. Schlcsinger, H. I., Brown, H. C., and Finholt, A. E., J . Am. Chem. Soc. 76, 205
(1953).
I., Brown, H. C., and Hyde, E. K., J . Am. Chem. Sac. 76, 209
(1953).
110. Schlesingcr, H. I., and nurg, A. B., J . Am. Chenc. Soc. 60, 290 (1938).
111. Schlcsinger, H. I., and Burg, A. D., Chem. Revs. 31, 1 (1942).
11.2. Schlesingcr, II. I., Rittcr, D. M., and Burg, A. B., J . Am. Cbem. Soe. 60, 1296
(1938).
115. Schlcsinger, IT. I., Ritter, D. M., and Burg, A. B., J . Am. Chem. Soc. 60, 2297
(1938).
114. Srhlcsingcr, H. I., Sanderson, R. T., and Burg,A. B., J . Am. Chem. Soc. 61, 536
(1939).
116. Schlcsingcr, H. I., Sanderson, R. T., and Burg, A. B., J . Am. Chem. Soc. 62, 3421
(1940).
116. Svhlcsingor, H. I., and Walker, A. O., J . Am. Chem. Soc. 67, 621 (1935).
117. Schults, D. R., and Parry, R. W., J . Am. Cbem. SOC.80, 4 (1958).
118. Seyfcrth, I)., J . Am. Chem. Soc. 81, 1844 (1959).
119. Shapiro, I., and Weks, 1%.G., J . Pliys. Chem. 67, 219 (1953).
f.20.Shapiro, I., and Weiss, H. G., J . Am. Chem. Soc. 76, 1205 (1954).
1.21. Shnpiro, T., and Wciss, 11. G., J . Am. Chem. Soc. 76, 6020 (1954).
1.22. Short,, S. G., Girardot, P. R., and Parry, It. W., J . Am. Chem. SOC.80, 20 (1958).
1M. Shore, S. G., and Parry, R. W., J . Am. Chem. Soc. 77, 6084 (1955).
f.24. Shore, S. G., and Parry, R. W., J . Am. Chem. Soc. 80, 8 (1958).
1.25. Shore, S. G., and Parry, R. W., J . Am. Chem. Soc. 80, 12 (1958).
f.26. Siegel, B., Mack, J. L., Lowe, J. V., and Gallaghan, J., J . Am. Chem. Soc. 80, 4523
(1958).
1.27. Soldate, A. M., J . Am. Chem. Soc. 69, 987 (1947).
128. Solomon, I. J . , Klcin, M. J . , and Hattori, K., J. Am. Chem. Sac. 80, 4520 (1958).
129. Stafford, S. L., and Stone, F. G. A., unpublished observations (1959).
130. Stt.indlt.r, M. .J., and Brhltbsingcr, H. I., J . Am. Chem. Sac. 75, 756 (1953).
131. Stock, A., Hydrides of Boron and Silicon, Cornell University Press, Ithaea, N.Y.,
1933.
13.2. Stonc, P. G. A., Quart. Revs. 9, 174 (1955).
f33. Stono, F. G. A,, Chem. Kms. 68, 101 (1958).
1.94. Stone, F. G. A,, and Burg, A. B., J . Am. Cbem. SOC.76, 386 (1954).
155. Stone, F. G. A., and Emt.lBus, H. J., J . Chem. Sor. p. 2755 (1950).
136. Taylor, R. C., Schults, D. R., and Emery, A. R., J . Am. Chem. SOC.80, 27 (1958).
236u. Trefonas, L. M., and Lipscomb, W. N., J . Am. Chem. Soc. 81, 4435 (1959).
137. Turner, H. S., and Warnc, Et. J., Chem. k Znd. p. 526 (1958).
138. van Alten, L., Seeley, G. R., Oliver, J., and Rittrr, D. M., Abslr. Papers i 3 d r d
Meeting A m . Chrm. Soc., Sun Francisco p. 38L (1958).
139. van der Maas Reddy, J., and Lipscomb, W. N., J . Am. Chem. Soc. 81, 754 (1959).
140. Wagner, R. I., et al. Phosphinoborine Polymcrs, WADC Technical Report
57-126, Parts I, 11, and 111, 1957, 1958 and 1959, Wright Air Development
Center, Wright-Patterson Air Force Base, Ohio.
14Oa. Wagner, R. I., and Caserio, F. F., J . Inorg. & Nuclear Chem. 11, 259 (1959).
141. Wartik, T., and Pearson, R. K., J . Am. Chem. Soc. 77, 1075 (1955).
14.2. Wartik, T., and Pearson, R. K., J . Znorg. & Nuclear Chem. 7, 404 (1958).
143. Weilmuenster, E. A , Ind. Eng. Chem. 49, 1340 (1957).
109. Schlcsinger, I.
C H E M I C A L REACTIVITY O F THE UOHON HYDIUDES
313
Weiss, H. G., and Shapiro, I., J . Am. Chent. Soc. 76, 1221 (1953).
Wiberg, E., Ber. 69, 2816 (1936).
Wiberg, E., Naturwissenmhaften 36, 212 (1948).
Wiberg, E., 16th Congr. Intern. Union Pure and A p p l . Chem., Paris p. 183 (1957).
Wiberg, E., and Bola, A., Rer. 73, 209 (1940).
Wibcrg, E., nolz, A., and Buchheit, P., 2. anorg. Chem. 266, 285 (1948).
Wiberg, E., Evans, J. E. F., and Noth, H., 2. Naturforsch. 13b, 265 (1958).
Wibcrg, E., Hertwig, K., and Bola, A., 2. anorg. Chem. 266, 141 (1947); 266, 177
(1948).
166. Wibcrg, E., and Modritzer, K.,2. Naturforsch. 12b, 123 (1957).
163. Wilson, M. K., and Piper, T. S., J . Inorg. & Nuclear Chem. 4, 22 (1957).
154. Wirth, H. E., Massoth, F. E., and Gilbert, D. X., J . Phys. Chem. 62, 870 (1958).
166. Wittig, G., Keicher, G., Ruckert, A., and Raff, P., Ann. 663, 110 (1949).
166. Zhigach, A. F., Kazakova, Y. B., and Kigel, R. A., Doklady Akad. Nauk S.S.S.R.
106, 69 (1956).
144.
146.
146.
147.
148.
149.
160.
161.
MASS SPECTROMETRY IN NUCLEAR CHEMISTRY

H. G . Thode.
C. C. McMuIlen.
and
K . Fritze
Departments of Chemistry a n d Physics. McMoster University. Hamilton. Ontario. C a n a d a
I . Introduction . . . . . . . . . . . . . . . . . . . 315
A. Instrumental Sensitivity
. . . . . . . . . . . . . . 317
B. Absolute Abundance Measurements . . . . . . . . . . . 318
C. Contamination Problems . . . . . . . . . . . . . . 320
D. Isotope Dilution . . . . . . . . . . . . . . . . . 321
IT. Mass Assignment
IIT. Half-life Determinations .

A. Decay Method . . .
13. Daughter-Growth Method
C. Specific Activity Method
. 323
. 324
325
. 325
. 326
. . . . . . . 327
. . . . . . . 327
. . . . . . . 328
. . . . . . . . . . .
329
. . . . . . . . . . . 231
IV. Neutron Capture Cross Sections

A. Quantitative Determinations
B. Relative Cross Sections . .
C. Absolute Cross Sections
.
D . Effective Cross Sections
.
.
.
.
.
.
.
.
.
.
.
.
.
V. Fission Yields . . . . . . . . . . . . .
A. Introduction . . . . . . . . . . . .
13. Relative Mass Chain Yields . . . . . . .
C. Absolute Mass Chain Yields . . . . . . .
D. Fine Structure in Mass Yield Curve . . . . .
E. Neutron Induced and Spontaneous Fission in Nature
F. Indepcnderit. Fission Yields . . . . . . . .
References
.
.
.
.
.
.
.
. . . . . 332
. . . . . 332
. . . . . 334
. . . . . 340
. . . . . 343
. . . . . 347
. . . . . 352
. . . . . . . . . . . . . . . . . . . 360
I . Introduction
Mass spectrometers have been used extensively in the field of nuclear

chemistry. and the development of instruments of high sensitivity in recent
years has made them indispensable tools in this field. In particular. the use
of isotope dilution techniques has extended the usefulness of mass spectrometers in the solution of nuclear problems.
The important developments in nuclear chemistry in which mass spectrometers have played a role will be reviewed in this chapter . These include
the identification and determination of radioactive isotopes and of stable
daughter products in nature. the accurate determination of half-lives for
315
316
H. G. THODE, C. C. MCMULLEN, AND K. FRITZE
aertaiu radioactive isotopes, iicutroti cross sections of both stable and

r;tdionrtivc isotopes, and, finally, the accurate determination of mass chain
yields and independent yields of fission products from the various elements
which undergo fission.
In all these applications, sensitive instrumoiits for the determiiintion
of isotope abundances are required. In this connection, the conventional
ionic-type mass spectrometer is used almost exclusively for the determination of absolute and relative isotope abundances with the exception of the
wide use of infrared methods for the determination of hydrogen to deuterium ratios.
The early mass spectrographs or mass spectrometers, were used mainly
to identify the isotopes of the elements and t o measure their abundances.
Using the parabola method of positive ion analysis, J. J. Thomson (107)
showed conclusively that the element neon consisted of at least two types
of atoms, one of mass 20 and the other of mass 22. A third stable isotope of
mass 21 was later shown to exist. In 1919, Aston (6) designed the first
mass spectrograph which could be used to measure isotope ubuiidanccs
arid succeeded in making abundance determinations of a number of elements. In 1918, Ilempster (24) developed a new type of instrument which
employed a method used earlier by Classen (90) to determine the electron
charge-to-mass ratio. This instrument which uses the geometric focusing
properties of a homogeneous magnetic field on a stream of charged particles,
forms the basis of most ionic-type mass spectroscopes used today. These
instruments are comprised of four basic components, u sample system, a n
ion source, tin analyzer, niid an ion-detector system. In general, ions which
are produced at the source, either by elcctron impact or by thermionic
emission from a hot filament, are accelerated by a voltage V through a system of slits. The beam of ions produced is then deflected by a uniform magnetic field II in the analyzer, according to the relation:
where e is the ionic charge, m is the mass, arid R the radius of curvature of
the ion path in the magnetic field. By a suitable adjustment of the parameters V and H , ions of any desired mass may be collected by a Faraday cup
and measured. The measured current of positive ions is the basis of relative
abundance calculations.
By 1930, Aston and Dempster, between them, had investigated the isotopes of most of the then known elements. However, their instruments
were of low sensitivity and many isotopes were missed. For example, Aston
in his work on the isotopes of ruthenium (7) was unable to detect n u Q8
which was later found to have a natural abundance of 1.9%. Also, the iso-
MASS SPECTROMETRY I N NUCLEAR CHEMISTItY
317
topes C13,N16,01, and 0lRwent undiscovered until the late 20s when thcir
existence was established by optical spectroscopy (45).
A. INSTRUMENTAL
SEXSITIVITY
Following the pioneer work of Aston and Dempster, there were rather
limited advances made in the field of isotopic abundance measurements until
1935 when A. 0. Nier (85) introduced a number of improvements to mass
spectrometer design. Since then, many more refinements have been made,
particularly with regard to more efficient ion sources and to more sensitive
ion detectors, so much so, that the mass spectrometer has become an
extremely sensitive instrument for the detection of practically all of the
elements, and its sensitivity remains second only to that of radioactive
counting techniques.
With the development of the electron multiplier by Bay (10) and Allen
(3) and its subsequent incorporation into a mass spectrometer as a n ion
detector @ I ) , it has been possible to increase the sensitivity of a standard
mass spectrometer by as much as a factor of lo4.
Also new and more efficient ion sources have been produced both for
gas and solid samples (8). For example, Inghram (56) modified a surface
ionization source to incorporate three filaments and succeeded in analyzing
samples of the heavy elements as small as 10-l2 gm, when the source is
used in conjunction with an electron multiplier detector. The over-all
efficiency of a surface ionization source depends upon the number of
neutral molecules impinging on the ionization surface a t the operating
temperature of the filament. In the case of many elements which have high
ionization potentials, it is found that the vapor pressure of the compound
containing the element reaches a high value before a suitable ionization
temperature can be achieved. The multiple filament source circumvents
this difficulty when the sample is dried on one of two sample filaments which
are in close proximity to a high temperature ionizing filament. By controlling the current in the sample filament, it is possible to set the sample
evaporation rate a t any given level. The neutral atoms which then impinge
on the hot ionization filament are ionized and accelerated into the magnetic
analyzer.
F. A. White et al. (120) have employed a filament which is folded in the
shape of a V into which the sample is placed. Neutral atoms which are
evaporated from the hot tungsten are given an additional opportunity to
become ionized on the walls of the tungsten V before they reach the first slit
of the ion source. Both the V-shaped filament and the multifilament ion
sources have proved successful in improving the efficiency of ion production
for elements having high ionization potentials.
318
H. G. THODE, C. C. MCMTJLLEN, A N D K . FIlITZE
R. ABSOLUTEABUNDANCE
MEASUREMENTS
Sources of error in isotope abundance measureincuts have been discussed

in detail elsewhere (106). The reproducibility of results with a modern
iristrumcrit is usually better than 0.1%. Furthermore, it is possible to detect
differences in isotope ratios of better than 0.02%using isotope ratio instruments which employ simultaneous ion collection systems. These high-precision isotope ratio instruments are used extensively to study equilibrium
and kinetic isotopc cffccts in chemistry and to measure the variations in
isotope abundances that occur in nature.
Although the isotope ratios of samples may be compared with high precision, the absolute errors may be quitc large, usually of the order of 1.0%.
These errors may be attributed to the method of introducing the sample
into the mass spectrometer or to discriminating effects within the ion source,
analyzer, or detector regions of the instrument (26).
For the mass spectrometric analysis of gaseous elements, the sample is
introduced into the ionization chamber through what is known as a leak.
Molccular flow will occur through the leak when the mean free path of the
molecules is at, least 10 times greater than the diameter of the leak. The
velocity of flow will then be inversely proportional to the square root of
the isotopic mass. At the same timc, molecular flow will occur when the
gases are pumped out of the ionization chamber. Provided the volume of
the gas reservoir behind the leak is large so that depletion of a light isotope
is small during analysis, cancellation of the discrimination in the gas flow
system occurs. Thus the gas in the ionization chamber will have the same
composition as that of the reservoir.
On the other hand, viscous flow will dominate in the leak when conditions are such that the mean free path is small compared t o the leak dimensions. I n this instance, the flow rate is independent of the isotopic mass and
a correction which is proportional to 1 0is necessary t o compensate for
molecular flow from the ionization chamber.
When analyzing elements which are ionized by evaporation from an
oven or heated surface, isotopic fractionation is once again a serious problem. This fractionation which again depends upon 1 1 / M is large when the
isotopic masses are small. It is therefore essential that each ionization
process be investigated to determine the amount of discrimination which
occurs during the analysis of a particular element.
When electron impact sources are employed, some mass discrimination
effects can be attributed to space charges which build up in the ionization
chamber. The magnitude of these effects may be determined by measuring
the isotope ratio as a function of the ionizing current arid then extrapolating
to zero current. In addition, isotope effects may occur in the dissociation
319
of molecules by electron impact. Since there is no satisfactory explanation

for this phenomenon, care must be exercised when mass spectrometric
studies of molecular species are undertaken (106).
Other effects which are of a minor nature but still require attention when
making absolute abundance measurements are those which occur in the
analyzer tube and at the collector. Residual gases in the instrument can
provide a source of error when weak ion currents are being measured. These
can usually be minimized by baking the spectrometer a t temperatures
between 300 and 400C in order to outgas the entire source and analyzer
system before the sample is introduced. Small angle ion scattering can, in
general, be overcome to a large extent by proper design of the instrument.
The tandem analyzer design of Inghram and Hess (58) and White and
Collins (118) practically eliminates this effect and it is then possible to
measure the ion currents of adjacent masses which have an abundance
ratio of approximately lo8.This type of tandem or double analyzer mass
spectrometer is being used in a number of laboratories to determine either
a rare isotope in nature (119, 120) or the formation of an isotope in some
nuclear process, in each case in the presence of a very high concentration
of the adjacent or target isotope. Such instruments are, of course, ideal
for the study of certain (n, 2n) reactions and are perhaps the ultimate in
low-level detection work. It should be pointed out, however, that the
abundance sensitivity of mass spectrometers using a single magnetic
analyzer may be considerably improved by using improved pumping systems and operating at lower pressures.
At the collector, secondary electron emission, which occurs when the
energetic ions are stopped, is prevented by incorporation of a suppressor
grid in the detector assembly. A less elegant method is to collect the ions
in a long V-shaped electrode from which very few secondary electrons will
escape. When an electron multiplier is employed to detect the ions, a mass
discrimination will occur when the integrated electron current from the
multiplier is used as a measure of the ion current. When pulse-counting
techniques are used to detect the ions, this discrimination disappears provided each ion produces an electron pulse which is of sufficient magnitude
to be counted by some counting rate-measuring device. It is obvious that
accurate absolute abundance measurements require very careful studies of
a number of discrimination effects which occur in the mass spectrometer
if an absolute accuracy of the order of 0.1% is to be achieved.
In 1950, Nier (86) reported the redetermination of the relative abundances of the isotopes of a number of elements with an accuracy of 0.1%.
The mass spectrometer used in these studies was calibrated for mass discrimination effects by means of carefully prepared synthetic mixtures of
A36 and A40. Samples of elements for which accurate isotope ratios are
320
H . G . TfIODE, C . C . MCMULLEN, A N D K. PHITZE
known may he used as secondary standards for instrument calibration

purposes.
C. CONTAMINATION
PROBLEMS
Ever sinec the time of Aston, contamination problems have plagued
the mass spectroscopists. Ions of hydrocarbon fragments which fall a t
almost every mass present a serious problem. These contamination problems are even greater when instruments of high sensitivity are used in
low-level detection work. Small residual ion currents due to hydrocarbons
persist even after long baking and inert gas flushing of the mass spectrom-
I
I28
I
IM
I32
MASS NUMBER
I
124
136
FIG 1. Mass spectrogram of the xenon isotopcs for the thcrmnl ncutron fission of
P1lP (62).
eter. However, the hydrocarbon problem may be overcome by increasing

the resolution of the instrument until the isotopic masses in question are
resolved from the hydrocarbons corresponding to the same masses. This
may be done with conventional instruments simply by using narrower
entrance arid exit slits in the mass spectrometer, providing the instrument
sensitivity is sufficient to measure the decreased ion currents which result.
Figure 1 shows a mass spectrogram of the xenon isotopes in B sample of
321
fission gas from the neutron-induced fission of Pu239(6%).The isotopes

Xe131,Xe132,Xe134,and Xe136are all produced in high yield and hydrocarbon
contamination peaks did not show up a t the low sensitivities used to
measure them. However, in the case of the isotopes of XelZ8,XelZ9,and
Xe130, measured a t sensitivities 1000 to 5000 times higher, small hydrocarbon peaks did appear on the high mass side of each isotopic peak. When
these doublets are completely resolved it is then possible to identify and
measure rare isotopes such as XelZxeven though the hydrocarbon contamination produces 4 times the ion current. In the particular example shown
in Fig. 1, the sample contained approximately 10-I2gm of Xe12*.The limit
of detection in this case was of the order of 10-13 gm.
D. ISOTOPE
DILUTION
The mass spectrometer method of chemical analysis which employs
isotope dilution techniques has wide application in nuclear chemistry and
physics, particularly in low-level detection work. There are many problems
which require a quantitative measure of the amount of a particular element
or isotope present in a sample in amounts less than one part per million.
Sensitive mass spectrometers and the availability of tracer isotopes make
the solution of these problems possible.
The isotope dilution method first proposed by Hevesy (49) has been
used extensively by Rittenberg (94) with stable isotope tracers for the determination of the elements and compounds of hydrogen, carbon, and nitrogen.
Now that separated isotopes of many elements are more readily available,
the method has been greatly extended and the isotope dilution technique
has become a very effective tool in quantitative trace analysis. By adding
to a sample a known amount of a separated or enriched isotope of the element under study and then measuring by means of a mass spectrometer,
the resulting change in the isotopic abundances of the element, it is possible
to calculate the absolute number of atoms which were present in the original
samples. For example, Xelz8which is produced by neutron irradiation of
112 (P7
5 Xe128)may be used as a tracer isotope in the absolute
determination of xenon. In this case, a small quantity of XelZ8tracer is
measured out with a system of gas pipettes and added to the gaseous Sample. The xenon may then be extracted from the sample employing the usual
techniques of rare gas purification. Finally, the change in isotopic composition of the xenon is determined mass spectrometrically. From this procedure, the number of atoms of the various isotopes of xenon in the original
sample may be determined. In cases where it is known that the isotopic
composition of the element under investigation is different from the com-
322
H. G . THODE, C. C. MCMlJLLEN, AND K. FHITZE
position found in nature (for example, fission products), measured yu:tntitics of the normal element can be used as diluent. I n the isotope dilution
method, the tracer is measured and added to the sample under study prior
to any chemical separation of the element in question. A quantitative
recovery of the tracer and unknown is then unnecessary since the analysis
depends only on the ratio of the isotopes in the tracer-unknown mixture
which is not significantly affected by separation procedures. This is a n important feature of the method since it is very difficult and tedious t o obtain
100% yields in trace analysis. A hypothetical isotope dilution measurement
of lanthanum in a solid sample (59)is illustrated in Fig. 2.
I PPM OF
LAf3' ADDED AS TRACER
THEREFORE,FROM ISOTOPIC RATIO,

LANTHANUM IN SAMPLE IS 0 5 PFM
LA138
IFROM TRACER1
J
I38
I
I39
I
140
MASS NUMBER
14'1~.2. Hypothetical isotope dilution dcterminstion of lanthanum (69).
One serious difficulty which is encountered in the determination of

minute samples or of trace impurities in a sample is that of contamination
from reagents. By analyzing samples of different size it is possible to evaluate the extent of this interference. In general, extreme care must be taken
in the preparation of reagents to ensure that the contamination by trace
quantities of an element is reduced to a level which can be tolerated.
Since stabIc isotope dilution depends on the mass spectrometric determination of isotope ratios, the method is generally restricted to elements
which have a t least two stable isotopes. However, in the case of mono
isotopic cesium, the long-lived fission product Cs13sor Cs137 can be used as
a tracer for thn analysis of this element. I n all, some 68 elements can be
MASS SPECTROMETRY I N NUCLEAR CHEMISTRY
323
analyzed hy this method (57). Modern mass spectrometers which employ

gm of
electron multipliers for ion detection may detect as little as
most elements. In the case of gas analysis it has been possible, using highvacuum techniques, to measure 1 part in lo8 of krypton in xenon (117).
By eliminating the difficulty with gas scattering it should be possible to
obtain sensitivities of 1 part in lolo.
Application of trace element techniques has brought fruitful results in
the area of nuclear science. Some of the important results obtained to date
are discussed in the following sections.
11. Mass Assignment
During the last 25 years, a large number of radioactive species has been
discovered. One of the prime objectives of nuclear chemists has been to
assign the atomic number and the mass number to the nuclides which are
responsible for observed activities. In general, the atomic number is found
by chemical analysis, whereas the mass assignment is most unambiguously
obtained by methods involving mass spectrometric techniques.
As early as 1937, Smythe and Hemmendinger (97) separated the isotopes of potassium in a high intensity mass spectrometer and showed conclusively that the activity which had been discovered 30 years earlier (19) is
associated with K40. In the same year, the known @-activity of natural
rubidium was assigned to RbS7using two different methods. Hemmendinger
and Smythe (46) used exactly the same experiment as for potassium,
whereas Mattauch (78) showed that strontium isolated from an old
rubidium-rich mineral was essentially pure Sr87 (see also Section 111).
Since 1945, mass assignments have been made to nearly 200 artificially produced radioactive species. This has been made possible on the one hand
by the enormous development in instrumentation during the war years
and, on the other hand, by the ability to produce a large number of radioactive isotopes in sufficient quantity to allow this type of study. One technique uses a mass spectrograph to separate the isotopes which are deposited
either on photographic plates or on metal collectors. In the photographic
method, the known spectrum of the stable element is used for mass calibration. Any additional lines can be attributed to artificially produced isotopes
and activities can be assigned to their respective masses using the transfer
plate technique (43).
In this method a second photographic plate is placed in contact with the
original collector plate. The radiations emitted by any radioactive nuclide
cause, after a suitable exposure, a blackening on the second plate and thus
makes possible the identification of the nuclide or nuclides responsible for
the radioactivity. An alternative method is to scan the first plate with a
324
H . G . THODE, C. C. MCMULLEN, AND K . FRITZE
rtidintioti dctcator having a fine slit. These tecfiiiiyues have t w n widely

used by workcrs in this field (41, 71, 72, 93).
Approximate measurements of the half-life have been employed, whenever possible, t o establish the identity of the radioactive species involved.
If a line corresponding to one isotope is found on the transfer plate, mass
assignment of the activity is certain. However, if the transfer plate shows
more than one line, a rough half-life measurement must be made in order to
associate the proper mass with the accurate half-life obtained from other
experiments. For example, Hess et al. (4'7) analyzed neutron-irradiated
rhenium in a mass spectrograph, collecting the ions on a photographic
plate. Two successive radioautographs of this plate were made, the first
one being exposed for 36 hr, the second one for 180 hr. All three plates were
then developed tirid cornparisoils of the lines made, using the primary plate
as a mass marker. The 36-hour plate showed two weak lines a t masses 186
tind 188, whereas, the second plate revealed only one line tit mass 186.
On the other hand, the decay curve for neutron-irradiated rhenium can be
resolved into two components having half-lives of 17 hr (73) and 91 hr (22).
Since no line appeared a t mass 188 on the second transfer plate the shorter
of the two half-lives must be associated with this rhenium isotope.
Iristetid of the conventional mass spectrograph, other types of mass
spectroscopes have been used in an effort to obtain mass assignments. A
time-of-flight mass separator, constructed a t the University of Californitk,
has been used extensively to study high-energy fission and spallation products (85). This instrument features an absolute mass calibration which
renders unnecessary the observation of ion currents from isotopes of known
mass. The investigation of iodine isotopes obtained as spallation products
of cesium ( 1 ) is a11 example of the use of electromagnetic isotope separators
not, only for mass assignment purposes but also for obtaining correct values
for the half-lives of several isotopes of the same element. Normally, the
isotopes are collected on thin metal foils which are cut into strips for counting, each strip containing the atoms which have been collected a t one mass
position.
Ill. Half-life Determinations
The obvious method of determining the half-life of a radioactive sample

is to measure the change in activity of the sample in successive intervals
of time. Another satisfactory procedure is to measure the disappearance
of a radioactive parent nuclide or the growth of a daughter nuclide using
mass spectrometric techniques. In the former method, emitted particles
are counted, whereas, the latter method measures directly the decrease in
the amount of the parent isotope, or the increase in the amount of the
325
daughter isotope, both relative t o a stable isotope of their respective element. I n addition, a combination of counting and mass spectrometric
measurements can be used effectively to measure half-lives which cannot
be readily determined by other methods.
A. DECAYMETHOD
The half-life T+ of any radioactive species is given by the following
equation:
T,
-0.6932/(1n NINo),
(2)
where N / N o is the fraction of the atoms remaining after a decay time t .

Thode and his collaborators (104, 112) who discovered the long-lived
fission product Kra5, have assigned its half-life as 10.27 f 0.18 yr by
measuring several times during a period of 7 yr the change in Kr8s/Kr84and
Kr85/Kr86 ratio. Since both Krs4and Krss are stable isotopes, the measured
ratios can be used directly to obtain N / N o . In principle, either of the two
ratios is sufficient to determine T,. However, the Kra6/Krs4ratio provides
a means of correcting for any normal krypton which could be introduced
accidentally because of frequent handling of the sample over a long period
of time. Essentially the same procedure has been used to determine the
5.271 f 0.002-day half-life of Xe133(74) and a 33 f 2-yr half-life of Cs13
(121).Half-lives, ranging from about 50 yr to a few days, can be determined
using this method. The lower. half-life limit is governed by the difficulty in
obtaining a sufficient quantity of the isotope in question for mass spectrometric analysis. Since the sensitivity of mass spectrometers changes
from element t o element, this lower limit cannot be well defined. As in the
case of counting techniques, the upper half-life limit is determined by the
length of time during which observations are made relative to the half-life
of the isotope under study.
B. DAUGHTER-GROWTH
METHOD
Since each radioactive decay forms one atom of a different element, the
accumulation of the daughter product can be followed as a function of time.
This build-up can be expressed by the following equation :
(3)
where N d is the number of daughter atoms; N Othe number of parent atoms
at t = 0; t the elapsed time; and T,, the half-life of the parent isotope.
Since the fractional change in Nd is particularly large for values of t which
326
I*. G . THODE, C. C. MCMULLBN, AND K. FIITZE
are small in comparison to the half-life, the daughter-growth method permits a more accurate determination of long half-lives in a relatively short
period of time. For example, work is in progress in this laboratory to determine, mass spectrometrically, the amount of SrS7which has been formed
from the @decayof the Rbg7in 20-gm samples of strontium-free rubidium
perchlorate. Using the currently accepted half-life of 5 X 1Olo yr (2) for
HbR7approximately 7 X lo-" gm of SrS7would be formed during a 2-yr
period. It is felt that this quantity of SrE7can be accurately determined
employing a known amount of SrgOfor isotope dilution. In principle, the
same approach was used in 1938 when Strassmann and Walling (101)
analyzed an old mica from Manitoba, Canada, for its rubidium and strontium content. They calculated the half-life (6.3 X 10*Oyr) from the geological age of the mineral after Mattauch (78) had proven, mass spectrographicnlly, that the chemically isolated strontium consisted of 99.7% Srs7.By a
careful analysis of several minerals from a number of pegmatites, using
isotope dilution methods, Aldrich and his co-workers (2) have shown the
half-life of Rbg7to be (5.0 f 0.2) X 1O'O yr. However, in experiments of
this kind, the accuracy of the half-life is limited by the value obtained for
the geological age, a difficulty which is circumvented in the experiment
discussed above.
C. SPECIFIC
ACTIVITY
METIIOD
Another accurate, but experimentally more difficult, method of determining the half-life is to evaluate the components of the differential decay
equation :
Ti
0.693
dN/dt
-~
N'
(4)
The disintegration rate, d N / d t can be measured by absolute counting and

the number of atoms N determined by means of mass spectrometric techniques. Bartholomew et al. (9) isolated the chlorine from irradiated KCl
and measured the disintegration rate of a known amount of chlorine using
4r beta counting methods. Another portion was analyzed by means of a
mass spectrometer for the ClS6content. Combining these results, they obtained a half-life of 3.08 X lo6 yr for C P . In cases where the amount of
the radioactive isotope is small, aliquots of a stock solution of the element
are used for 47r counting and to determine the concentration by means of
isotope dilution. This procedure was followed by Wiles and Tomlinson
(122) and by Brown et al. (16) to determine the half-life for Cs13'. The difference in the results of the two groups (26.6 yr and 30.0 yr, respectively)
is difficult to explain since the corrections applied to the 4~counting meas-
327
urenieiits were cssciitially the same in both ~ascs.This method has been
used to determine the half-life of Be"' (SO), SrgO(123), U232(96),and other
isotopes.
IV. Neutron Capture Cross Sections
Although a number of different reactions can be initiated wheii an element is irradiated with neutrons, this discussion will be restricted essentially
to slow neutron reactions of the (n,y) type. These reactions produce a mass
increase of unity in any isotope involved and, consequently, cause a change
in the isotopic composition. The magnitude of these changes will be determined for the various isotopes by the neutron capture cross section, the
neutron flux, and the irradiation time.
)
can be described by the following
Quantitatively the ( n , ~reaction
equation:
N = N o e-"'@,
(5)
where N o is the number of target nuclei a t the beginning of the irradiation;

N , the number of target nuclei remaining a t the end of the irradiation; uc,
the capture cross section; 4, the neutron flux; and t, the irradiation time.
The form of this equation immediately suggests that the approach to
measure cross sections is similar to the one used for the determination of
half-lives. However, cross section measurements are complicated by the
fact that uC is a function of the neutron energy and by the difficulty jn
obtaining a reliable value of the neutron flux. The product isotope of a n
(n,7)reaction can be either radioactive or stable. In the former case, either
counting methods or mass spectrometric techniques can be utilized in order
to obtain information concerning the capture cross section. When the product isotope is stable, only mass spectrometria methods are applicable.
A. QUANTITATIVE
DETERMINATIONS
Dempster (25) for the first time showed that the isotopic composition
of an element could be modified by a long neutron irradiation and proved
conclusively that Cd"3 was responsible for the large cross section which
had been observed previously in normal cadmium. His spectrograms of
irradiated cadmium revealed a marked decrease in the intensity of the
Cd113 line and a corresponding enhancement of the Cd114line. Lapp, Van
Horn, and Dempster (68) also showed that the large cross section found in
samarium and gadolinium could be attributed to the isotopes Sm14Q,Gd166,
and CrdIS7.Figure 3 shows a mass spectrogram of the gadolinium isotopes
before and after irradiation with thermal neutrons. It can be seen that
328
H. G . THODE,
C. C. MCMIJLLEN, A N D K. FHITZE
AF TER
IRRADIATION
0EFORE
IRRADIATION
I-
z
W
I
I
:
I
I
:
3
0
I
154
IS6
I
158
160
154
156
I58
160
MASS NUMBER
FIG.3. Mass spcctrogrrtm of gadolinium before and after thcrmnl ncritron irradiBtion
(110).
(hi*65
and Gd16' have been completely converted to Gdlb6and Gd168,respectively, as a result of neutron absorption.
H. EELATWE
Caoss SECTIONS
With a mass spectrometer it is possible to measure relative changes in
isotopic composition very accurately. One can, therefore, obtain some yuantitative information coriccrr~ingcapture cross sections. In the case of a
hypothetical element consisting of three stable isotopes, A , B, and C,having
successive mnsses where A and C have measurable cross sections, the number of atoms of the different isotopes, after an irradiation of time duration
1, can bc expressed as follows:
N A t = N Ae-"A6t,
~
N s t = N B O N A O (-~ ecUAACt),
Nct = N C ~ ~ - " " ~ ~ ,
(6)
(7)
(8)
where N A t , N B t , Nct are the number of atoms of isotopes A , B, and C after

the irradiation; N A ~NBO,
, Nco are the number of atoms of isotopes A , B,
and C a t the beginning of the irradiation (1 = 0) ; U A , U C , the neutron capture cross section of isotopes A and C , respectively; 4, the neutron flux;
and t, the irradiation time.
aA
329
Ry combining Eqs. (6), (7), and (S), the following expression relating
and (TCcan be derived :
The values for NA~,co and N B O / N Aare

O obtained from the known
abundances of the element before irradiation, whereas, N c t / N ~ tand
N B ~ , A ~can be calculated from the results of the mass spectrometric
analysis of the irradiated material. The accuracy of spectrometric measurements are usually good to f 0 . 5 % or better.
It should be realized that this hypothetical example is a rather simple
one and in practice, the problem is frequently complicated by one or more
of the following factors: (1) the number of stable isotopes is larger; (2)
some of the irradiation products are radioactive but have a considerable
cross section which might lead to another stable isotope; (3) the capture of a
neutron by a particular isotope can produce isomers, which, in turn, are
likely to have different cross sections (in this case also the branching ratio
for the production of the two isomers would have to be taken into account);
(4) if the irradiation position in the reactor is such that the contribution
of fast neutrons is significant, allowance may have to be made for other
types of nuclear reactions. It is obvious that these complicating factors
will make the mathematical treatment more complex. In this connection,
the reader is referred t o a paper by Hayden, Reynolds, and Inghram (44)
which deals with the reactions induced in europium by slow neutrons.
Studies of the isotopic capture cross sections of a number of other clement^
have been performed by other workers (48, 54J 111).
C. ABSOLUTECROSSSECTIONS
It will be noted in most of the publications th a t absolute capture cross

sections have been tabulated. In order to convert from relative t o absolute
isotopic cross sections, a knowledge of the absolute elemental cross section
is required. Since the elemental cross sections are rarely known to better
than &lo%, the error in the isotopic cross section will be determined by
the accuracy of the assigned elemental cross sections.
If Eq. (5) is rewritten in the following form
(T
(In N I N d / &
(10)
it can be seen that a mass spectrometric determination of N/NO and a

knowledge of the integrated neutron flux 4tJ will permit a n evaluation of
the absolute capture cross section. The integrated flux can be determined
by the simultaneous irradiation of a flux monitor with the sample. Boron
330
H . L;. TIIODE, C . C. MCMULLEN, AND K. FHITZE
arid cobalt have been used for this purpose (35,Ol).In the former case, thc
change in the Bl0/B11 ratio resulting from the B1"(np)Li7 reaction is measured and $t calculated, using the known Bl0cross section
111 the latter case, the disintegration rate of Coeo produced during the
irradiation is measured and $t calculated using this information
The accuracy of the value of the integrated flux will be largely determined
by the accuracy of the cross section assigned to the monitor material.
Up to this point, discussion has been restricted to multi-isotope elements. Since one obtains only isotopic ratios from mass spectrometric
analyses, the disappearance of a single isotope cannot be measured directly
if the radioactive product is short-lived. However, in principle, a method
which is very similar t o the daughter-growth method for determining halflives (Section 111) should be applicublc. For example, the neutron irradiation of Au197 yields Au198 which decays with a 2.70-day half-life to HglgX.
After a suitable irradiation and decay time, the gold is dissolved, a known
quantity of mercury added, and the isolated mercury analyzed in a mass
spectrometer. From the increase in the Hg1Q8/Hgly9ratio the number of
AuIgRatoms produced can be calculated. As a result, the cross section can
be obtairied providing the neutron flux 4 is sufficiently small that the
( n , ~Aulyy
)
reaction is of no consequence.
The largest neutron capture cross section ever found is the oiie for
Xe136. This isotope which has a 9.2-hr half-life (61) is one of the high yield
fission products and, therefore, an exact knowledgc of its cross section is
extremely important from the standpoint of nuclear technology. The fission
product chains involving XelXbnnd Xe136can he represented as follows (61):
/115.3m Xe136m
Tcls6 -+
G.7h 1'36
'9.2h
I
1
Te136+ 8Gs
--+
1
xo135
/"
1Wy C P -+ stable nn"5
(11)
11
stable
If the irradiation of the fissile material is carried out a t u low neutron

flux the X ~ 1 3( n~ , ~Xe1s6
)
reaction will not significantly alter the cumulative
yields of the Xe136and Xe136mass chains. As the flux increases, the number
of Xe136atoms present a t any time will become larger and a significant
increase in the yield of the 136 mass chain, at the expense of the 135 mass
33 1
chain, will result. Equation (12) gives the fractional increase of Xe136owing
to neutron capture of Xe136during an irradiation.
136
xeq
xc136
~-
j;=o
t=T
(NO.
atoms ~e135present a t any time t )
136
Y,la6
x fT
xF x
u x p dt
(12)
Xe!l;/6
where F is the thermal iicutron flux (neutrons cm-2 hr-l); axelas,

the thermal
neutron capture cross section for Xe136;Y p , the fission yield of
the
total number of fissions per hour; and T, the irradiation time (hr).
This increase can be determined by measuring either the cesium or
xenon isotopic ratios. These results combined with the low flux fission yields
can then be used to calculate the cross section of Xe136.The ideal experiment would be one in which the flux remains constant during the entire
irradiation so that the amount of Xe135present at any time is not dependent
on flux variations. This approach has been used by Fickel and Tomlinson
(28), who arrived at a value of 3.2 X lo6 barns (1 barn =
cm2) for
the neutron cross section of Xe136.Cross sections of other fission products
have been measured (12, 65, 76) although the conditions in these experiments were such that the effect of flux variations could not be evaluated.
D. EFFECTIVE
CROSSSECTIONS
In order to produce changes in isotopic composition of sufficient magnitude to be measured by a mass spectrometer, it is necessary to utilize the
high neutron flux obtainable in a nuclear reactor. This flux will consist of
neutrons having widely varying energy; and in the case of a well-moderated
reactor, the neutron spectrum can be thought to consist of a Maxwellian
distribution in energy with an added epithermal component which varies
inversely as the neutron energy. The neutron temperature which corresponds to the classical temperature of the Maxwellian component can be
expected to vary from one point to another in the reactor. In other words,
the neutron energy spectrum will depend on the reactor geometry. It is
obvious that any value for the capture cross section computed from mass
spectrometric data will onIy represent that so-called effective cross section
for the particular irradiation position. This effect will be most pronounced
in cases where the cross section departs markedly from a l/u energy dependence (neutron resonances). Although such cross-section results are of considerable value to reactor technologists, it would be more satisfactory to
express these cross sections as the value for 2200-meters/sec neutrons. Westcott (115) describes the effective cross section by the following equation :
= ao(g
+ 4,
(13)
332
H. G . THODE, C. C. MCMULLEN, AND K. FRITLE
where ug is the 2200-meters/sec cross section, g and s are a measure of the

departure of the cross section from a l / u form (for a l / u law: g = 1 and
s = 0), and r measures the relative number of epithermal neutrons in the
iieutron spectrum. For further details concerning the effective cross section
for a number of isotopes the reader is referred to the publication of Westcott
(116).
V. Fission Yields
A. INTRODUCTION
Since the discovery of fission, a great effort has been made to unravel
the maze of fission products which are produced in the various fission
processes. In this work, mass spectrometric methods have been used extensively. These methods have been particularly fruitful in the determination
of relative and absolute fission yields.
In the fission of a heavy nucleus, many products are formed in detectable amounts throughout the mass region 72 to 161. In general, thermal
neut,ron fission results in two unequal mass fragments which fly apart with
a total kinetic energy of about 170 MeV. The most probable heavy mass is
around 139 and the most probable light mass is around 95.
Since the primary fragments formed have excitation energies of about
11 and 9 Mev for the light and heavy fragments, respectively, they disintegrate by neutron emission (114). It is evident from the asymmetry of fission
neutron emission (126)that these neutrons, referred to as prompt neutrons,
are emitted after the fragments have acquired 90% of their kinetic energy,
the decay constant of the highly excited fragments being less than 2 X 10's
sec-'. The average number of neutrons emitted per fission v is 2.47,2.51,2.90
for the thermal neutron induced fission of U2s6,Was,and Puzse(114).
The mechanisms and data of the fission process have been reviewed
recently by Leachman (7'0). Several different approaches have been used
in an effort to explain the asymmetry of the fission process as well as other
fission parameters. These approaches include : developments of the liquid
drop model (60,51,IOZ), calculations based on dependence of fission barrier
penetration on asymmetry (54), the effect of nuclear shells (62>79, S l ) ,
the determinations of the fission mode by level population of the fragments
(18, 33, 84), and finally the consideration of quantum states of the fission
nucleus at the saddle point (16, 108). All these approaches require a mass
formula whereby the masses of the fission fragments far removed from
stability may be detcrmined. The lack of an adequate mass formula has
hindered the development of a satisfactory theory of fission. The fission
process is highly complex and it is not surprising that the present theories
fall short of a full explanation.
333
For the thermal ueutron fission of U2335,

the process can be represented
by the following equation:
U235
+n+
U236* 4
ZIAml + ZzBq
+ v(n) + 173 MeV,
(14)
+ +
where 21 2 2 = 92, and ml mz v(n) = 236.

The primary fission products A and B are formed with a proton-toneutron ratio which is less than that for stable nuclei of the same mass.
The fragments are therefore proton deficient and undergo a series of P-decays in a chain until they attain stability. Each mode of fission, therefore,
leads to two such chains. For example, mass chains 133 and 101 may be
complementary masses in the above fission process where v the average
number of neutrons emitted per fission is 2. In this case, the mass chains
are as follows
2.3d X P
6.3m Te1331
0.72 7
-+
---f
4.lm SbIJ3
0.28 \
2m
Te133
0.024 7
20.8h 1133
0.970\
stable Cs133
1 7
5.27d XeIa3
(15)
14.6m Mol01-j 14.3m Tc1'J' -+stable Rule*
For the fission of each heavy nucleus, it is important that we have as

much information as possible about the fission products, for both practical
arid theoretical reasons. In general, we wish to identify all the products,
to determine their genetic relations one to the other, to determine their
nuclear properties, half-lives, neutron cross sections, etc., and finally, to
determine the fission yields of the various nuclides (independent yields) and
of the various mass chains (cumulative yields or total of the independent
yields along a chain). The yield is defined as the percentage probability per
fission that a given nuclide or mass chain will be formed.
Whereas the yields of the individual nuclides give the charge distribution for fission fragments of the same mass, the cumulative yields or chain
yields give the probability of the various modes of fission or mass distribution for a given fission process. It has been shown (38,87) in studies of the
distribution of nuclear charge in fission that in most cases the last two
members of a fission product chain are formed almost exclusively from the
decay of precursors, and only a small fraction of the yields of the last two
members result from direct formation from primary fission fragments
(prompt neutron emission only). Therefore, a fission yield measured for
any of the later members of a mass chain, particularly for the stable or
long-lived end products, usually represents the total yield of that chain.
Independent yields found for shielded nuclides (nuclides immediately f ollowing a stable isotope in a mass chain) although of considerable interest
334
H. G . THODE, C. C. MCMULLEN, AND K. FRITZE
from the poirit of view of charge distribution (see Section V.1") itre, in gciieral, very small and add little to the total chain yield.
B. RELATIVEMASS CHAINYIELDS
1. Relative Yields; the M a s s Spectrometric Method
It was clear from the time of the first fission product studies that the
mass spectrometer would eventually be used in separation problems, mass
identification, and isotope abundance measurements. However, the early
work on fission products involved very small samples of material and only
radiochemical methods were considered sensitive enough to identify and
follow the radioactive isotopes. In 1945, Thode and Graham (104)succeeded
in obtaining muss spectrograms of the xenon and krypton isotopes formed in
the thermal neutron fission of U236.
mc
a
0
C
II
--
FIQ4. Mass spcctrogram of fission product xenon from neutron-irradiated Uza6(104).
These early spectrograms shown in Figs. 4 arid 5 were obtained with a

180" gas-type mass spectrometer, using fission gas samples of 0.1 to 1 mm3
in volume at N.T.P. (105).The importance of this method of investigating
fission products was immediately recognized. Stable and long-lived isotopes
of xenon and krypton which are end products of fission chains, were identified for the first time. Since these products were extracted from the uranium
several years after the neutron irradiation and sirice the precursor nuclides
were knowri to he all relatively short-lived, it could be assumed that the
MASS S P E C T R O M E T H Y IN NUCLEAR C H E M I S T R Y
335
total yield of each mass chain would be in the form of these stable isotopes.
The relative abundances of these isotopes, therefore, gave directly the relative yields of the respective mass chains. Many of these cumulative yields
have been obtained by measuring, with radiochemical methods, the relative
number of atoms of iodine in the various radioactive chains. However, to
obtain the yields in this manner, the data must be corrected for the halflives of the isotopes measured and for the relative counter'efficiencies for
radiations of different energies. Because corrections of this kind are difficult and uncertain, the early radiochemical yield da t were uncertain by a t
c
C
E
a
0
0
C
78
80
82 84 I
Mass Number
FIG.5. Mass spectrogram of fission product krypton from neutron-irradiated U2"

(104).
least ~ 2 0 7 Evcri
~ . today, with modern counting techniques, yields determined by radiochemical methods are rarely better than f5%. The mass
spectrometer, on the other hand, provides a method of determining directly
the relative yields of a large number of mass chains for each element investigated with an accuracy of 1.0% or better. In the case of the xenon and
krypton isotopes, relative yield data were provided in the first instance for
8 mass chains (104).
2. Fission Yields and Problems of Chain Branching
a. Independent isomers ( K P ) . The identification of K P which had a

low yield relative to masses 84 and 86 in fission gas samples 1-2 yr old
(Fig. 5 ) was of particular interest (104). Prior to 1945, the only KrS6
isomer known had a 4.5-hr half-life, an activity first noted by Saell (98),
336
li. 0. TIIODE, C. C. MCMULLEN, AND K. FRITZE
while studying the (d,p) reaction on krypton. In 1945, Hoagland and Sugarman (53)found an activity with a minimum half-life of 10 yr when investigating the gaseous activities in fission products. They were able to show
that this was due to an isotope of krypton and assigned it either to KrS6
or Kre7. The long-lived isotope of krypton revealed in the mass sppect,rogr:tms of gaseous fission products showed the definite existence of a longlived isotope of krypton a t mass 85 and hence an isomer of the known 4.5-hr
krypton of that mass. By following the decrease of this isotope over a period
of a few months, Thode and Graham were able to show that it decayed
with a half-life of about 9.4 yr (104). A more recent mass spectrometric
rncusurement made with samples 1-8 yr in age, an elapsed time of the same
order as that of the half-life, gives the more accurate value of 10.27 f 0.2
yr (11%).The krypton isotope a t mass 85 has therefore two isomeric states,
n long-lived one with a half-life of 10.27 yr and a short-lived one with a
4.4-hr half-life. Since the shorter-lived isomer will have completely decayed
before the fission gases were extracted from the uranium samples, the mass
spectrometer result8 of Fig. 5 will indicate only the yield of the 10.27-yr
isomer. The low yield found for this isomer indicated a chain branching
with n large portion of the short-lived isomer decaying directly to RbS5
t i i d only a small part, 25%, decaying first t o the 10.27-yr isomer or ground
state of Kr. By making the assumption that mass yields for chains 83,
84, 85, and 86 fall on a smooth curve (Fig. 5 ) , the total yield of the 85 muss
chain may be determined. This procedure, used by Wanless and Thode
(113) gave a branching ratio for the decay of KrS6of 0.29, the 4.4-hr Kr86m
decaying 77.5% directly to stable R b s and 22.5% through the 10.27-yr
KP6 isomer as follows (61):
7
40s
So86
+ 3rn
Ur*s
4.4k K P m
10.775
o . m stable Rbs6
I
7
(16)
10.3~
Krs6
The value of this branching ratio measured directly from P--decay data
is 0.30 (11) whereas a direct measure of the fission yields of Kr86and RbS
using mass spectrometry and isotope dilution gave a value of 0.28 (13).
The assumption that the Krsa yield falls on a smooth mass yield curve in
the case of the thermal neutron fission of U235therefore appears t o be
justified.
b. Delayed neutron emitters. All fission yield determinations must take
into account chain branching due to delayed neutron activities and neutron
capture processes. Both Kr87m and Xe137 each with one neutron more than
ti closed shell (51 and 83, respectively) are instantaneous neutron emitters
(99) and Lead to chain branching as follows (61):
337
+n
1<P
fI
/ I i r 8 7 m
17s So87 -+ 55s B P

Xe.130
p 5 X 10LayRbs7
17.8~~1K P
+n
Because of the delayed neutron activities of 55 sec and 24 sec which

are derived from p-decay half-lives of Brs7 and 113, respectively (the
neutron emission is instantaneous and follows the p--emission), the two
mass chains 86 and 136 gain in yield at the expense of the 87 and 137
chains, respectively. However, in general, these effects are small, amounting to a few percent of the yield of a chain at most.
c. Neutron capture processes. Neutron capture processes, however, can
lead to a large increase in the yield of one chain at the expense of the chain
one mass unit lower, and, of course, the extent of this chain branching will
depend upon neutron capture cross sections of the nuclide in question and
the conditions of irradiation.
In general, only nuclides with high capture cross sections lead to appreciable chain branching a t moderate neutron fluxes. An important example
of a pile poison or a high neutron absorber is the 9.2-hr XeIa6which has
been discussed in Section IV. In this case, the burn up is appreciable
even at fluxes of the order of 5 X 10l2 neutrons/cm2/sec. The neutron
capture process in the 135 mass chain is indicated below (61):
Because of this capture process, the mass spectrometer abundance

data for the xenon isotopes from fission will show a high yield value of the
136 mass chain, and similar data for the cesium isotopes will show a correspondingly low value a t mass 135. The percentage increase in the Xe136yield
due to this neutron capture process can, of course, be calculated, provided
we have accurate information concerning the number of fissions per second,
the half-lives of 1135and Xe136,the neutron capture cross section u of Xe136
at the neutron temperature of the irradiation, and, finally, the length of
the irradiation period (Section IV).
d. Relative cumulative yields of isotopes of xenon and krypton from
U236 nJission. To obtain accurate relative yields, however, it is far better
to avoid any corrections for neutron capture processes by irradiating sam-
338
H . G. TIIODE, C. C. MCMULLEN, A N D K. PltlTZE
plw for investigation at low fluxes, where such processes are negligiblc.
The early mass spoctrorneter yield determinations of the isotopes of xenon
and krypton for the thermal neutron fission of U236were repeated with this
in mind (113). Fission gases for study were extracted from uranium
irradiated in the thermal column of a ieactor at fluxes of the order of
10lo neu trons/cm2/sec.
The relntive abundances of the xenon and krypton isotopes obtained
with these samples provide the best relative yields of mass chains ending
in these isotopes for the thermal neutron fisaiori of U236.Table I gives this
TABLE r
KRYPTONISOTOPE
ABUNDANCES
IRRADIATED IN NRX THERMAL COLUMN"
FISSION 1'RODUCT X E N O N AND
FOR U R A N I U M
Atom
Isotope
abundnnco
Absolute fksion
yield*
(%)
14.10
25.93
7.59
52.38
13.42
20.08
36.91
29.59
Xc'J*/Xe'aa(0.443)
i%)
0.557
1.02
0.300
2.07
2.93
4.3s
8.06
6.46
6.59
P'ission product isotope abundanccs (14, 113).

Xc131/Xc13J rntios (14).
c Abfiolute yield C S ' ~
(6.59%) (go), also XeIS3yield (6.62% )(60).
relative yield data normalized to 6.59% a t Xe133. This latter absolute yield
was determitied hot,h a t Xelsv(6'0) and Cs':is (90) by methods described in
Section V.C, and in addition the partial yield curves for krypton and xenon
were tied together by an absolute deterinination of the number of xenon
arid krypton atoms by means of the isotope dilution technique (Section I).
3. The Cumulative Yield of 5.27-Day Xe133
Mass spectrometer yield determinations nre not limited to stnble and

long-lived end products of fission chains. I n 1950, Macnamara, Collins, and
Thode (7'7) found it possible to extract fission gases from irradiated uranium
disks without a long "cooling" period and were able to determine the yield
and half-life of Xe'33. The mass spectrograms obtained were similar to those
given in Figs. 4 and 5, except for the appearance of XeIa3,and, of course,
precursor had not fully
the Xe131abundance wns low since the 8-dtiy
decayed. The fission yield of the 133 mass chain for U236
fission is calculated
339
from the mass spectrometer abundance ratio of Xe133to Xe13* obtained

after an irradiation time t , a "cooling time" t l , and at a time tz after extraction of the fission gases.
The two mass chains being comparedare as follows:
0.8m Sb134+44m Te134L52.5m1134+stableXeIa4
p63m Te133m
4 . l m Sb133
'2m
(19)
2.3d Xe133m
20.8h
Te1337
1133
Y
'
/*stable
Cs133
5.27d Xe133
As the period of "cooling" is long compared to the half-lives of the

Xe134 precursors, the total mass 134 chain will decay to Xe134.Therefore,
the total number of Xe134atoms at time of analysis is given by:
(20)
Nxe1u = fY134t)
where Y134 is the yield of the 134 mass chain and f the neutron flux during irradiation time t. The ratio of the total number of Xe133atoms to
the total number of XelS4atoms a t the time of analysis is therefore given
by the equation:
[at tz1 - exp (-Adz)
NXp
Yxe134t
Nxelas
.[I-
hz exp ( - A d ) - h1 exp (-hi)

hz - A 1
and Xe133,respectively.
where h1 and hz are the decay constants of
The measured mass spectrometer ratio of Xe133/Xe1aat tz (time of analysis) which may be equated to the left side of the equation, therefore makes
possible the calculation of the yield Y133of the mass chain.
Since the independent yield of Csla3will be negligible the total yield
obtained for the 133 mass chain should be the same whether measured at
Xe133or Cs133.The Xe133cumulative yield therefore provides a link between
the relative yields of 8 mass chains ending in stable and long-lived isotopes
of xenon and krypton and of the 3 mass chains ending in stable and longlived isotopes of cesium.
4. Relative Yields Over Entire Mass Range

The mass spectrometric method of fission product yield measurements
can be applied to the entire mass range of fission products. Inghram et al.
(55) investigated fission products in the rare earth region and showed how
the partial yield curves obtained from isotope abundance measurements
for the various elements can be fitted together by the method of stable
340
H . 0. THOUE, C. C. MCMULLEN, AND K. FlZITZE
isotope dilution. This method has been used extcnsively in the analysis of
fission products (13,14,44,59,60),and as described in Section I, makespossible the determination of the absolute number of atoms of each nuclide
in a mixture of fission products and extends the mass spectrometric method
of yield determinations to the whole mass range.
One important feature of the mass spectrometric method of fission yield
determinations is the ease with which a large number of mass chains may
be investigated a t the same time for a mixture of fission products from a
single irradiation, Since ions for the various elements are emitted from a
sample on a hot filament at different temperatures, isotope abundance
measurements of a number of elements may be made on a single sample
without chemical separation. It is, therefore, possible to obtain a large
body of yield data for a particular fission process with high relative nccuracy (*I%).
Fission yields of the stable and long-lived isotopes of xenon, ccsium,
barium, cerium, neodymium, samarium, krypton, rubidium, strontium,
zirconium, molybdenum, ruthenium, have been measured mass spectrometrically for the thermal neutron fission of U236(40, 61, 90),UZaa(4, 31,
61, 82, loo), and Pu23B(32, 61, 66, 124). I n addition, the relative yields of
the xenon and krypton isotopes produced in the fast neutron fission (fission
spectrum neutrons) of U238and Th232(64, 113) have been determined.
C. ABSOLIJTEMASS CHAINYIELDS
1 . Total Yields Normalized l o 100%
The relative yields of the isotopes of the aforementioned (Section

V.B.4) twelve elements that can be measured mass spectrometrically give
reliable yields of about 43 Inass chains. These, taken together with a number
of yields dctermined radiochemically , give essentially the whole mass yield
curve, which may be normalized to 100% in order to obtain absolute yields
for the various mass chains. This procedure is used when mass chain yields
for most of the chains of appreciable yield for either the light or heavy
fragments are known. AbsoIute yields good to &37$ have been obtained in
this way.
2. Direct Measurement of Absolute Yield
To obtain absolute yields directly, it is necessary to determine, not only
the absolute number of atoms of a nuclide produced in fission, but, also,
the total number of fissions. The determination of the number of fissions in
a sample of a fissile element irradiated with neutrons is a most complex
problem and cannot be discussed in detail here. In general, the number of
341
fissions can be determined by measuring the change in the B1O/BII ratio in

a BF, flux monitor irradiated simultaneously in the same neutron flux as
the fission sample. Conversion to the number of fissions is then effected by
the use of the known ratio of Bl0absorption cross section to the fission cross
sections (91). Cobalt and samarium have also been used as flux monitors.
In the case of the cobalt, the amount of Cowformed from a known amount
of cobalt can be determined by ionization chamber measurements, using
standard cobalt sources for comparison. In the case of the samarium, the
change in Sm160/Sm149
ratio is determined by mass spectrometry.
In the determination of the number of fissions in an irradiated sample by
the use of flux monitors, account must be taken of the flux depression in the
sample due to self-shielding to obtain an effective flux. Also, the capture
cross sections of the monitors and the fission cross sections of the sample
are neutron energy dependent. It is, therefore, necessary to know the energy
distribution of the neutrons or the neutron temperature and to determine
effective cross sections (Section IV). This can be done by using two monitors
such as cobalt and samarium, the one monitor being used to determine the
neutron temperature corresponding to the neutron distribution as described
by Fritze et al. (36).
3. Summary of Absolute Yield Determinations
Katcoff (1958) has prepared tables of fission product yields from lowenergy neutron-induced fission of U236,Th2a2,and Pu239(61). In the compilation of these tables he has given first priority to mass spectrometric
data since these are more accurate (f3%) than most radiochemical data
(f10%). Since these tables were prepared, the absolute yields of some 43
mass chains from the thermal neutron fission of Pu239have been redetermined using mass spectrometric methods (29, 36). These yields are considered accurate to better than 1% 011 a relative basis and better than 3%
on an absolute basis. The earlier determinations have been shown to be in
considerable error because of incomplete extraction of fission products
from PuOz.
4. Interpretation of Mass Yield Curvesfor U236+ n and Pu239+ n Fission

The mass yield data for the thermal neutron induced fission of UZa6
and Puzasare plotted in Figs. 6 and 7, respectively. In each case the mass
yield curve is folded over so that complementary masses fall toge:her
assuming that 2.5 neutrons (U2a6 n) and 3 neutrons (Puz3g n) are
emitted per fission. This means that the sum of the corresponding masses
totals 233.5 and 237, respectively.
A complete mass yield curve makes it possible to calculate the total
342
H. 0. THODE, C . C. MCMULLEN, AND K . FRITZE
niimber of neutrons, v, released in a fission process. For examplc, in the neutron-induced fission of Pu239
v = 240
- 2('
mass X yield
1yield
From the data of Fickel and Tomlinson (29)a value of 2.80 f 0.08 is
obtairied assuming an absolute accuracy of f 3 y Ofor the measured fission
yields. This is in agreement with the world-accepted value of 2.90 (114).
MASS
FIQ.6. Comparison of the light and heavy mass yields in thc t,hcrmul neutron fission
of UZJ6(91). KEY:-, heavy; ---,light.
A more detailed consideration of Fig. 7 shows that the corresponding

mass chains in the light and heavy groups have equal yields in the mass
ranges 138-154,83-99, indicating the emission of three neutrons per fission
on the average for modes of fission in this region. Although there is a dearth
of measured fission yields in the 96-118 and 118-131 mass region, the displacement of the curves in Fig. 7 of approxiniately 1 Inass unit suggests
that on the average only 2 neutrons are emitted per fission in modes of
MASS SPECTROMETRY IN NUCLEAH. CHEMISTRY
343
fission leading to these masses. Finally, in the mass region 132-138, 105-99,
the yields of the corresponding masses in the light and heavy regions do not
correspond a t all. This i s the region of the fine structure in the neighborhood of the 82 neutron shell.
In the case of neutron-induced fission of U236there is again good correspondence between fission yields for the light and heavy fragments in the
138-154,83-99 mass regions (Fig. 6). The correspondence occurs, however,
for the average neutron emission of 2.5 per fission. As in the case of the PuzS9
12
of
FIG.7. Comparison of the light and heavy m a s yields in thc thermal neutron fission
(27, 29). KEY:0-0, heavy masses; 0--0,
light; @, light and hoavy.
Pu239
fission, the curves appear to be displaced in the mass region 117-131 and
103-117 indicating correspondence of yields for light and heavy fragments
for the emission of 2 neutrons per fission on the average rather than 2.5
neutrons. Again, there i s poor correspondence between the light and heavy
mass yields in the mass region where fine structure occurs. These yield
curves have suggested another possible explanation of the fine structure
discussed in the next section.
D. FINESTRUCTURE
IN MASSYIELDCURVE
1. Fine Structure in 82 Neutron Shell Region
The early fission yield data obtained by radiochemical methods, although only good to & l o % a t best, did give the familiar double hump mass
344
H. G . THODE, C. C. MCMULLEN, AND K. FRITZE
yield curve and showed that the fission process was asymmetric with the
most probable modes of fission grouped around masses 95 and 140. Within
the limits of the radiochemical determinations, a smooth curve could be
drawn through the yield values for the various masses. However, the first
mass spectrometric investigations of fission products showed up irregularities and established the existence of fine structure in the mass yield curve
(104). In particular, the yield of XelS4is high relative to the two adjacent
masses and there is a sharp peak at masses 133 and 134 in the mass yield
curve for UZaa n fission (Fig. 4).
It is interesting to note that this fine structure falls in the neighborhood
of the 82 neutron shell. Four or five mass chains in this range each have
one nuclide with 82 neutrons with a significant independent yield, and in
the case of the 134 mass chain, this nuclide Te134is the most probable member of the chain according to the Coryell, Glendenin postulate of charge
distribution for a given mass in fission (see Section V.F) .
Glendenin (37)attempted to explain the high yields of the 133 arid 134
mass chains by assuming that nuclei with 83 neutrons (which has already
emitted the usual number of prompt neutrons) would have a high probability of boiling off a prompt neutron to form the most stable 82 neutron
configuration rather than emit /3- or y rays as in the ordinary case. According to this scheme, the 134 mass chain would be increased by the reaction
Telaa (83 neutrons) -+ Te1M (82 neutrons)
+ n;
and decreased by the reaction

Sb4 (83 neutrons) -+ Sb***(82 neutrons)
(23)
+ n.
If TelS6has a higher independent yield than Sb184,then the 134 chain

will gain more than it loses by this mechanism and the XeIs4yield will be
high.
Although reaction (1) should be greater than reaction (2) on the busis of
the simple theory of charge distribution in fission, this mechanism cannot explain the observed fine structure. Wiles et al. (Ibl), therefore, postulated that in addition to the Glendenin mechanism high yields must result
from a structural preference for isotopes with 82 neutrons in the primary
fission act. Glendenin et al. (39)reached a similar conclusion after observing
a high yield at mass 100, the complementary fragment to the 134 mass
chain (Fig. 6).
Investigations of the xenon isotope yields in the neutron-induced fission
of UZaa(SI),U2@( I I S ) , Pu238(X?),and PuZ41(36) have also been carried out.
In all cases, there is evidence of some increased yield in the xenon mass
range. The sharpness of the peak at mass 134 varies considerably, however,
and in the case of UZa3 n fission, there is little evidence of fine structure
MASS SPECTROMETHY I N NUCLEAR CHEMISTRY
345
from the xenon isotope yields alone. The entire fission yield curve for
U233 n fission, however, does indicate an increased yield in this mass
range spread out over a larger number of mass chains.
The complete fission yield curve obtained for the neutron-induced fission
of Pu239has suggested another explanation of the fine structure in the mass
yield curve in the 132-136 mass range (27,29).Since the folded over fission
yield curve of Fig. 7 discussed in Section Y.C.4 indicates correspondence of
yields of the light and hcavy fragments for the emission of 3 neutrons on the
average in one region of mass division and 2 neutrons in another, there
must be a mass region in which a transition occurs.
Fickel and Tomlinson ($7) have suggested that an increased stability
of primary fragments in the heavy mass region, perhaps those having either
82 neutrons or 50 protons, may cause a decrease in neutron emission a t
this point. Such a decrease in neutron emission would cause a bunching
up of yields such as those found in the 131-136 mass range. The total effect
would be to increase the yields in the 131-136 mass range by one half of the
yield of a single mass chain in this region (3% of fission). A corresponding
decrease in the yields could occur in the 100-106 range by approximately
the same amount. It is seen from Fig. 7 that the total yield discrepancies
in these two complementary mass regions do differ by about 6%. The
complete mass yield curves obtained for the neutron-induced fission of
U236 suggest a similar explanation of the fine structure in the mass yield
curve at xenon (masses 131-136).
This latter explanation of fine structure is a plausible one. However,
complete fission yield curves for the neutron-induced fission of other fissile
nuclides are needed to add further light on this problem.
2. Fine Structure in the 50 Neutron Shell Region
It is perhaps significant that irregularities in the mass yield curve occur

also in the region of the 50 neutron shell. These irregularities appear in the
yields of mass chains ending in stable isotopes of krypton. Figure 8 shows
the krypton yields for Th232(64),U233($I), U236(14), and UZ3*
(113) obtained
mass spectrometrically. It is apparent that the mass yield curve in this
region moves to higher yields as the mass of the fissile nuclide decreases.
This indicates that the mass yield curve for the light fragments is actually
displaced. The mass displacement of these curves, for a fixed fission yield, is
approximately equal to the mass difference between the fissile nuclides.
This implies that the heavy fragment portion of the mass yield curve remains a t a relatively constant position. A closer examination shows a fine
structure a t masses 84 and 85. It has been suggested that since the relative
yields of the krypton isotopes may be determined mass spectrometrically
with an accuracy of 1% arid a precision of O.l%, there is little question
346
H . G. THODE, C. C. MCMULLEN, A N D K. FRITZE
about the existence of fine structure effects which amount to a 40% change
in the isotope ratio.
The high 84 yield in all four eases probably arises from a preference for
this chain in the original fission process or cross chain branching from the
85 mass chain. The behavior of the 85 yield depends on the fissile nuclide
and low for Tha32and UZa8.
The deviaand is high for U233,normal for P6,
tion of the 85 yield from the smooth mass yield curve can be correlated
with the Z / A value of the fissile nuclide. This suggests that a branching
83
84
85
86
MASS NUMBER
FIG.8. lhe krypton yields for nciitron-indilccd fission of Th2, U233, U*JS,and
Uz38
(641.
(neutron emission) may occur in thc 85 chain since the indepcrident yield
of such a prompt neutron emitter would be related to the Z / A value of
the fissile nuclide. The nuclide in question must have a fairly high independent yield nnd possess a low threshold for neutron emission. Kennett
and Thode (64) point out that these conditions w e fulfilled by &eS6 which
has 51 neutrons, one more than a closed shell. As pointed out by Glendenin
(37) the loss of the loosely held 51 neutron would probably occur only where
the nuclide was formed as a fission fragment (after the usual prompt neutron emission) in a highly excited state. Siiice the fractional yield of the
mass chain occurring a t SeB6(primary yield) will depend on Z / A , any
chain branching at this nuclide should show about the same percentage
effect for U238and Th232.
3. Fission Products from Several Fission Processes
Because of the fine structure in the mass yield curve in the 131-136
mass range, ratios of the xenon isotopes in this region vary markedly from
347
one fission process to another. This makes possible a convenient method of

unraveling a mixture of fission products produced from several different
fission processes.
The early studies of fission products from neutron-induced fission of
U236were carried out with samples of natural uranium irradiated with
pile neutrons which were not completely thermalized. A comparison of
the mass spectrograms of the xenon isotopes extracted from one of these
samples with those extracted from uranium depleted in U236indicated a
14% contribution from U238fission in the natural uranium samples (113).
Also, a sample of natural uranium irradiated in a fast reactor showed a
60% contribution from U238fission. The mass spectrometric method can
and UZs6
be conveniently used to measure the fission contribution from U238
in fuel elements of reactors where natural or enriched U236fuels are used.
E. NEUTRON-INDUCED
AND SPONTANEOUS
FISSION
IN NATURE
Soon after the discovery of fission, Petrzhak and Flerov (92) demonstrated with ionization chambers that uranium undergoes fission spontaneously and emits neutrons in the process. Later, Segre (95)used separated isotopes of uranium and showed that U238has a spontaneous fission rate of
about 23 times that of U235,the half-lives for spontaneous fission of U238and
U235being 8.04 X 1015 yr and 1.87 X 1017 yr, respectively. These results
suggested the presence of fission products in uranium minerals. In 1950,
xenon and krypton fission products were identified in old uranium minerals
using mass spectrometric methods (75).The patterns of xenon and krypton
isotopes found indicated an asymmetric mass yield curve similar to that
obtained in the neutron fission of
and U238.
and Neutron Fission of U235

There were, however, certain marked differences. First, the mass yield
curves in the xenon and krypton mass ranges were steeper for natural
fission. Since there is much less energy available for spontaneous fission
than for neutron fission, the process is expected to be more selective (steeper
curves). This was considered strong evidence that spontaneous fission of
U238is involved in natural fission. Second, the abnormally high yields found
at masses 133 and 134 (fine structure in the mass yield curve) in the neutron
fission of U235are shifted toward the lower masses for natural fission which
and neuis the same direction of shift found between neutron fission of U236
Third, the xenon-to-krypton ratio was found to be much
tron fission of P8.
higher in natural fission gases. From this and other evidence, Macnamara
and Thode (75) concluded that a t least a large part of the fission gases in
uranium minerals resulted from the spontaneous fission of U238.
Later, it was shown that both neutron fission of UZ35and spontaneous
1. Spontaneous Fission of
348
Ha
a.
TIIODE, C. C. MCMULLEN, AND
K. FRITZE
fission of Uysxoccur in nature (YO, 117') and that the xenon and krypton
isotope pattern found is a composite for the two processes. The fission
isotope patterns of xenon and krypton were found to vary from one sample
to another, depending on the proportion of the two fission processes which
have taken place. Since the mass 129 chain yield is about 1.0% for the neutron fission of U236and was found to be very nearly zero for the spontaneous
fission of UZa8,the yield of this isotope was taken as a n index of the prowhich has
portion of neutron fission of U236and spontaneous fission of U288
occurred in a given mineral. Several minerals investigated were found to
contain less than 0.02% Xe12g,indicating that at least 98% of the fission
Other samples showed up
gas was due t o the spontaneous fission of U2s8.
to 25% neutron fission of UZsb.There are, of course, two sources of neutrons
for this process, spontaneous fission and (a,n)reactions on light elements.
2. Aholute Yields
a. Spontaneous fission of Uzs8. Recently, absolute cumulative yields of
the xenon and krypton fission products from six uranium mirierals were
determined mass spectrometrically, using isotope dilution techniques (126).
The uranium ores were crushed arid sieved to about 100 mesh and finally
coriceritratcd to a density greater than 3 by gravitational separation with
diiodomethane. A homogeneous aggregate of uranium mineral concentrate
was produced by this procedure suitable for the extraction and analysis of
the fission gases. Greater than 99.5% of the fission gases was released by
heating the mineral concentrate in a molybdenum crucible in a vacuum
induction furnace over a period of 12 hr at 1300C. The absolute yield calculation for the spontaneous fission of 1JZax
depends on the independent
determination of the number of spontaneous fissions which have taken place
and the number of atoms of a given fission product which has been foImed
from this process.
The total number of U238atoms which have undergone spontaneous
fission in a mineral sample since deposition are calculated from the U238
content, the age, and the rates a t which the U238nucleus decays by alpha
emission and by sporittmeous fission according to the relation :
N = WAr/A,(eXaT - 1);
(24)
where W is the present number of U2s8

atoms in the mineral sample; Xf,
the UZs8spontaneous fission decay constant [8.62 f 0.30 X 10-17/yr (96)];
A,, the U238decay constant [ = 1.54 X lO-'O/yr (SS)];T, the period during
which fission products accumulated, assumed to be given by the lead isotope
age measurement.
The total number of atoms of each fission product xenon and krypton
isotope in the niineral sample was determined by means of mass spectrom-
MASS SPECTHOMETHY IN NUCLEAH CHEMISTRY
349
etry and the isotope dilution technique discussed in Section I. In this connection, known quantities of the tracer isotopes Xe128,Kr80, and K P , prepared by the neutron irradiation of iodine and bromine in the N R X
reactor a t Chalk River, were added to the impure fission gases liberated
from the uranium mineral concentrate. Finally, the fraction of the fission
gas in each mineral sample which resulted from the spontaneous fission of
U238was obtained from a breakdown of each natural fission gas sample
into its various fission components. This analysis makes use of the known
relative abundances of the xenon and krypton isotopes for the various
possible fission processes separately and a comparison of these with the
relative abundances obtained for the fission gases extracted from the various
minerals. As mentioned in Section V.D.3, the ratio of the xenon and krypton
isotopes in fission product gases varies markedly, depending on the nature
of the fission processes because of the fine structure which occurs in this
region.
b. NeutronJission 0fU238. A preliminary analysis of the data showed that
it was not possible to separate satisfactorily the six mineral isotope patterns
into a U236neutron-induced component and a UZ3*spontaneous fission
component. However, a satisfactory analysis was obtained when a third
component, U238neutron-induced fission, was considered along with the
two processes proposed earlier (SO). Table I1 shows the breakdown obtained
for the six samples investigated.
c. Proportion of neutron-induced Jission. These results indicate that there
is a correlation between the percentage of uranium in a mineral concentrate and the percentage of neutron fission. This relationship is shown in
Fig. 9. The percentage of uranium in the mineral concentrate should be
approximately equal to the percentage of uranium in the mineral itself,
since most of the rock matrix was removed by gravitational separation with
diiodomethane. It had been assumed earlier that the presence of the rare
earth elements which have large thermal neutron absorption cross sections,
is largely responsible for the low proportion of neutron fission in uraninites.
The occurrence of U238 neutron fission in uranium minerals suggests much
larger neutron energies than had been suspected and indicates that the
rare earths have little effect on the amount of neutron fission in a uranium
mineral. The fact that the Jahola Lake uraninite (high rare earths) and the
Cinch Lake pitchblende, with equal uranium concentrations in the concentrates, have the same proportion of neutron fission, supports this conclusion.
d. Losses of fission gas from uranium minerals. Included in Table I1
are the ratios of the number of atoms of Xela6determined to the number
of fissions for each sample (spontaneous fission component only). These
ratios expressed as percentages, can be interpreted as absolute yields only
if no fission products have been lost from the mineral. The value obtained
CI
c
?
e
0
3
TSBLE I1
FISSION-PRODUCT
COMPONENTS
IN NATURAL
FISSIOX
Mineral
Urn8 Spontaneous fission (70)

USs Neutron fission (yo)
U2%Neutron fission (yo)
Xela (Spontaneous fission)
x 100
C*38(Soontaneous fission)
Cinch
Lake
95.86
4.14
0.00
6.30k0.38
0.0 f 6 . 0
Jahala
Lake
Beaverlodge
Lake
96.15
3.85
0.00
91.82
8.18
0.00
1.59k0.26
4.38f0.18
30.5 f 6 . 6
74.8 f 7 . 3
Great Bear
Lake
-m
Eagle
&line
Belgian
Congo
81.65
15.19
3.16
80.42
17.22
2.36
68.65
25.23
6.12
4.78f0.23
3.21f0.13
24.1 f 7 . 1
49.0 f 6 . 3
3.51f0.17
44.3f 6 . 5
+z
e
MASS SPECTROMETHY I N NUCLEAR CHEMISTIlY
35 1
for the Cinch Luke sample is in agreement with absolute yield estimates
and therefore suggests that inert gas losses from this mineral have been
small. If the lead ages are assumed to be the period during which spontaneous fission products accumulated, then the low yields of Xe136for the
other five minerals must be due to a loss of inert gas during geological time
and these losses are given in Table 11. If it is assumed that there was no
35
25
30
35
40
45
U 238 CONCENTRATION
50
55
(%I
FIG.9. Relation between the amount of neutron fission and uranium concentration
in a uranium mineral or ore concentrate (126).
loss of fission gases from the Cinch Lake mineral since it was deposited,
then the spontaneous fission yields obtained for this sample are absolute.
The fact that there is little evidence of fractionation of the two rare gases,
xenon and krypton, for this sample supports this assumption. I n any case,
the absolute yields of mass chains ending in stable isotopes of xenon and
krypton from the spontaneous fission of U238obtained for the Cinch Lake
sample are minimum values.
3. Absolute Yield Data for the Spontaneous Fission o j U238
Recently, absolute yields of Mog9and several iodine isotopes from spontaneous fission of
determined by radiochemical methods were reported
by Parker and Kuroda (89) and Ashizawa and Kuroda (5). These yields,
along with the absolute yields of the xenon and krypton isotopes obtained
352
11. G. THODE, C. C . MCMULLEN, AND K. FRITZE
for the Cinch Lake sample are given in Table 111. The good agreement between the xenon and iodine yields a t masses 131, 132, and 134 is in accord
with the assumption that no fission product xenon and krypton had been
lost from the Cinch Lake mineral. The large losses of rare gas fission prodiwts indicated by most of the uranium minerals prevents the general use
TABLE I11
ABSOLUTE
YIELDSIN U*38SPONTANEOUS
FISSION
MU8
Element
Yield4 (%)
83
84
85
86
99
131
132
133
134
135
136
Kr
Kr
0.0327 -I 0,0028
0.122 f 0.012
Kr
0.951 f 0.057
Xe
Xe
0.524 f 0.031
3. 83 f 0 . 2 2
X(?
5. 14
fO.31
Xe
6.30
f 0.38
Element
Mo
I
I
I
I
I
Yield (%)
*
*
6 . 4 .5b
0.4 f .lc
3.6
.4c
1.5 f . 3 O
5 . 2 f .5c
5.1 f . 5 O
All data in this column were obtained by the authors.

Parker and Kuroda
c Ashizawa and Kuroda
a
of uccumulnted rare gas fission products for age determinations. Nevertheless, the fractionation which accompanies loss should be a sensitive test of
product age, that is, the period during which
the vttlidity of x L4fissi~n
fission products havc accumulated without loss.
F. INDEPENDENT
FISSION
YIELDS
For n complete understuridirig of the fission process, it is necessary to
know both the charge and mass distribution of the fission fragments. Fission yield studies to date have been mainly concerned with mass distribution or the total or cumulative yields of a given mass. The question of charge
distribution for n given mass involves the primitry or independent yields
of the fission fragments. Thc measurement of independent yields along a
mass chain is very difficult with present techniques because of the very
short half-lives of most members and the correction iiecesssry for the decay
rate and the production rate.
Attempts have been made to determine cumulative yields for some of
the short-lived krypton and xenon isotopes by gas sweeping a solution of
uranium undergoing fission. However, quantitative removal of inert gases
MASS SPECTROMETRY IN NUCLEAR CHXMISTHY
353
from solution was not achieved. Wahl (109)found that inert gases are quantitatively removed from barium or uranyl stearate and by utilizing welldesigned experiments obtained values for the fractional cumulative yields
of these short-lived inert gases.
1. Shielded Nuclides
For the most part, however, independent yield data are confined to
about a dozen nuclides that are near stability and are shielded by stable or
relatively long-lived isobars.
For example, in the mass chain
the independent yields of tin, antimony, and tellerium accumulate at stable
TelSoand thus any yield of the shielded I'30 is, therefore, the independent
fission yield of that isotope. Since the primary fission fragments emit prompt
of 1130 will be given by the equation:
neutrons, the independent yield YIIM
Y~iao= P~iiofo
+ P~infz+
P~eifi
P1wf3
Ppf4
+. . .,
(26)
where PI^^, PIIJl,PIlJr,PIIS,arid PIluare the primary yields of 1130,

a t the instant of fission (before prompt neutrons are
1
1 and
emitted), andfo, f1, fi, f3, and f4 are probabilities that 0, 1 , 2 , 3, or 4 neutrons
will be emitted per fragment. The yields of some of the shielded isotopes,
for example
have been determined radiochemically (37). However,
the detection of a shielded nuclide in the presence of isotopic activities of
much greater yield is possible only when its radiation characteristics and
half-life are sufficiently different from those of the higher yield isotopes to
permit identification by decay and absorption measurements. The measurement of the yield of a n isotope like XeIB0by means of a high-sensit vity
mass spectrometer avoids these difficulties and makes possible a direct and
accurate determination of the independent yields of such shielded isotopes.
2. Mass Spectrometric Determination of Shielded Nuclide Yields
Kennett and Thode (62) examined rare gas fission products with a conventional 90" direction-focusing mass spectrometer using a n electron multiplier for a detector. The sensitivity of this instrument was several orders
of magnitude greater than that of an ordinary single-ion collection spectrometer and was sufficient to permit a complete analysis of the xenon isotopes including the very low yield shielded nuclides Xelz8 and XeI3O, in
samples containing as little as 10loatoms of these isotopes. Figure 1 shows a
typical mass spectrogram obtained for a sample of fission product xenon
extracted from irradiated plutonium. These spectrograms showed, for the
first time, the presence of XelZ8and Xe130 as fission products. Whereas the
354
11. G . THODE, C. C . MCMULLEN, AND K . FKITZE
more abundant isotopes Xel3I, Xe132, Xc134, and Xe136represent the cumulative yields of their respective mass chains, the yields of Xe128arid Xe130
represent the independent yields of the shielded isotopes
and 113",respectively, along with a small contribution of the independent yields of the two
xenon isotopes themselves
of peak heights).
The determined X C ' ~ * / Xand
~ ~ Xe130/Xe131
~*
ratios are used to calaulatc
the fraction of the 128 and 130 mass chains appearing as primary fission
fragments. Corisideration of the neutron capture contribution must also he
made since tlhe reactions 11*' ( n , ~ ) and IlZ9( n , ~I1go
) are both possible.
Prom these considerations, the following relation may be derived :
(measured yicld of XcI3O) = (indcpcndcnt yield of 1130)
(cont,rihution from capture in the 129 chain)
where a is the fraction of the 130 chain formed as

4 is the flux; t the
irradiation time; u the neutron capture cross section of
and Y Areprescrits the total chain yield for mass number A . Actually, the neutron capturc process is small a t moderate neutron fluxes and in such cases the second
term in Eq. (27) can be neglcctcd. An analogous calculation gives the independent yield of
A similar investigation, in the krypton region rcvcaled the presence of
KrE0and Krs2indicating that Brsoand B P are also product,s of fission. Since
the two selenium nuclides preceding the BrX0and BrS2in their respectivc
mass chains are stable, these isotopes can only be formed directly in fission
and their yields are independent yields. The independent yields of Bra0
and Brsz arc, therefore, determined from the yields of their daughters
Kr*O and KrS2measured mass spectrometrically. As in the analogous xenon
case, the independent yields of the KrS0and Krs2 will be less than 1% of
that of the corresponding isotopes of bromine, and the total yields of these
isotopes may therefore be ascribed to the parent isotopes Br8*and BrS2.
It is thcreforc possible, by means of a simple mass spect,rometric analysis
of a single fission gas sample, to determine directly and with considerable
accuracy the independent yiclds of 4 shielded nuclides. To date, the primary
yields of I'2Rand
have been determined for the thermal neutron induced
fission of U233,U236, and Pu239, and preliminary yields of Brsoand Bra*have
been ohtained for the thermal neutron fission of Uz3 and WS6(68).
S . Charge Distribution Postulates
The accurate independent yields of

arid 1130 obtained for U233,U236.,
and Pu239fission provide a severe test for any hypothesis of charge distribu-
355
tion. Since both isotopes are formed as the same element and therefore the
value of Z is constant, &lid since the most probable charge 2, for mass
chains 128 and 130 is in the region of the 50-proton shell, a discrepancy
between theory and experiment might be expected here.
a. Equal chain length postulate. The early independent yield data
obtained by Glendenin, Coryell, and Edwards (58),Glendenin (37), and
Pappas (88) were sufficient to determine a semiempirical charge distribution which is assumed independent of the mass chain, the fissioning nucleus,
FIG.10. The empirical charge-distribution curve as determined by Glendenin (37).
and the neutron energy. The hypothesis of equal charge displacement of

Glendenin assumes that the most probable charges for both fragments in
fission are equally displaced from the center of the valley of beta stability.
The probability curve is of the Gaussian type centered around Z,, the most
probable charge for the mass chain considered, and is shown in Fig. 10.
The value of Z , is given by:
Z p
=ZA
- >i[ZA
Z(AI--A--131)]
2';
(28)
where Z A is the most stable charge for mass A which is found from the BohrWheeler stability line; A l and 2' are the mass number and charge of the
fissioning nucleus, and v is the average number of neutrons emitted per
fission. Therefore, the most stable charge for each fragment satisfies the
equation :
( Z p
- ZA)light
( Z p
- ZA)heavy.
(29)
356
H. G. THODE, C. C. MCMULLEN, A N D K. FRITZE
Pappas found better agreement with experimental values when he

calculated the 2,-values using the lines of maximum stability given by
Coryell, Brightsen, and Pappas (23). These lines have discontinuities at
nuclear shell boundaries. He defined 2, by:
2, = Z A - $ 5 [ z A
- Z(AI--A+I)
- z'],
(30)
so that 2, corresponds to the mass chain A - 1 since prompt neutron

emission is thought to occur from the fragments themselves. I n most cases
2, is nonintegral.
The most probable charge 2, for the mass chains 128 and 130 for the
various fissioning nuclei may be calculated using either the Glendenin or
Pappas equation.
Figure 11 shows the determined fractional chain yields relative to the
smooth curves obtained semiempirically by Glendenin and Pappas. It is
CHAIN POSITION (Z-Zp)

FIQ.11. Comparison of the independent yield data of 1x28 and Ilm (62)with thc chargedistribution curve of (A) Glendenin (37) and (B) Pappns (88).
clear that the yields do not agree with these curves and that a smooth
curve cannot be drawn through all the points. In both cases, the calculated
Z is too large and there is a displacement according to mass.
b. Postulate of maximum energy release. Since all the 1130 yields lie on the
curve, the assumption made by Glendenin (37) that the curve shown in
Fig. 10 is independent of the mass and charge of the fissioning nuclide
appears valid within the accuracy of the yield measurements. Also, the
357
assumption that the chargedistribution curve is independent of the mass

of the fission fragments seems to be confirmed by radiochemical determinations. Kennett and Thode (62),therefore, concluded that the displacement of
the 112*yields is due mainly to the method used to calculate 2,. Since
it might be expected that 2, would remain close to 50 over several masses
in the neighborhood of the 50-proton shell, they proposed that shell effects
be taken into account more directly in the determination of 2, for a given
mass. This was done by postulating a 2, which would yield the greatest
energy release in the fission act. For example, in the fission of Pu239,the
value of 2, is defined in such a way that
M(240, 94) - 3M(1,0) - M ( A , 2,)
- M(237 - A , 94 - 2,)
(31)
is maximum.
For Eq. (31) to have its maximum value, accurate values for masses
far removed from the stability line are required and it is necessary to make
extrapolations by means of a mass formula which will include shell effects.
The mass formula of Kumar and Preston (67)in which shell effects and spin
terms have been included, was chosen as the most applicable for the evaluation of Eq. (31). Preliminary calculations of 2, made over the mass range
A = 126 to A = 136 for U233,Was, and Pu239,are plotted in Fig. 12. The
effect of the 50-proton shell and 82-neutron shell are clearly indicated. It
is seen that the value of 2, remains very nearly 50 from A = 128 to A = 132
which is quite different from the behavior of 2, determined from the equal
chain length hypothesis shown as Z', in Fig. 12. It turns out that the value
of 2, calculated from energetic considerations is 0.4 charge units higher a t
mass 128 than that calculated from the equal chain length equation. This
indicates that the charge-distribution curve is displaced by 0.4 charge units
a t mass 128 due to the influence of the 50-proton shell. Actually, it is seen
from Fig. 11 that the measured 128 yields are displaced by 0.5 charge units
in the same direction. Therefore, the independent yield results a t IlZsand
I13O seem to support the postulate of maximum energy release for the division of nuclear charge in fission. Also, the values of 2, calculated by the
method of maximum energy release predict that the independent yields
of 1131 and
will be low and will lie to the left of the curve in Fig. 11 just
as the 1128 yields are high and lie to the right. This has been confirmed by
the results of Wahl who found that the independent yields of Te131 and
1132 were low. In general, independent yields plotted on the basis of a 2,
calculated from the maximum energy release postulate give a chargedistribution curve which although different from that obtained from the
equal chain length postulate does account for the present independent
yield data equally as well. In addition, the maximum energy release postulate does indicate shell effects when the Kumar and Preston mass formula
358
H . C. THODE, C. C. MCMULLEN, AND K. FRITZE
is used and can account for the accurately determined yields of

and
1130 which otherwise present an anomaly.
In the evaluation of 2, from the postulate of maximum energy release,
other mass formulas may be used for extrapolating to masses far removed
from the stability line. In this connection, the mass formulas of Hay and
53
52-
51
u*=
+------
----49
N
'
126
128
126
128
I30
**
53
2 '30-
130
:82
I
132
134
132
136
134
136
Newton (42) and of Cameron (I?') have been tried and the values of 2,
calculated do not show sharp discontinuities at shell crossings. The mass
formula of Cameron takes shell effects and spin terms into account much as
does the mass formula of Kumar and Preston, however, the mass formula
of the Iatter is specifically used for isobaric sequences which show discontinuities in P-decay energies versus charge curves when shells are crossed.
M.48S SPECTROMETRY IN NUCLEAIl CHEMISTRY
359
c. Multineutron evaporation from primary $ssion jrayments. Recently,

Cameron (18) used the statistical theory of fission with a new mass formula
as the basis for a semiempirical procedure for correlating fission yields and
kinetic-energy distributions. This procedure enabled him to calculate
primary yields of individual nuclides. Since his calculations indicate
extremely small primary yields for nuclides near the valley of beta stability,
he concluded that the small independent yields found experimentally for
shielded nuclides have resulted from the evaporation of 3 or 4 neutrons from
nuclei of approximately maximum primary yield on a mass chain.
He is, therefore, suggesting that in Eq. 26 for the independent yield of
I13O, the fourth and fifth terms are dominant; whereas, in the case of other
nuclides in the 130 mass chain of moderate yield, the second and third terms
are dominant since single and double neutron emissions are more probable.
I t is clear that 2 - Z,, the displacement from the most probable charge,
decreases from 113' to
and the primary yields therefore increase markedly. On the other hand, the probability of multineutron evaporation from
the primary fragments will decrease as the number of neutrons emitted
per fragment increases beyond one. The question arises as to the relative
importance of these two effects.
If multineutron evaporation is responsible for the independent yields of
shielded nuclides, then, of course, these yields would not be predicted by
any simple hypothesis for the division of charge in fission which does not
take into account the distribution of fission neutron numbers.
Leachman (69)and, more recently, Terrell(lO3) have made calculations
of the probability P of emission of any given integral number of prompt
neutrons from fission making certain assumptions as to the distribution of
excitation energy among the fission fragments. Although the calculations
are very complex and accurate experimental data are required to fix the
many parameters of the theory, some estimate can be made of the probability that a fission fragment will emit a given number of neutrons. In the
most favorable case for multineutron emission, where it is assumed that all
the excitation energy is always on one fragment, the probabilities of 3
and 4 neutron emission are such that terms 4 and 5 of Eq. (26) could much
more than account for the experimentally determined independent yield
of I*S0providing the primary yields of 1133 and 1134 predicted by the simple
chargedistribution theories are correct. However, the primary yields predicted for
are too low to account for the independent yield
and
found for
on the same basis.
The question of whether the independent yields of shielded nuclides
result from multineutron evaporation from primary fragments as has been
suggested is therefore still unsettled. It is perhaps significant that the
for the neutron-induced fission of Pu239
independent yields of IlZsand
360
II. G. THODE, C . C. MCMULLEN, AND K. FRITZE
did iiot change with un increase it1 thc neutron ciicrgy froin thc thermal
range to the range of energies in the fission spectrum.
In any case, accurate independent yields of shielded nuclides can be
determined mass spectrometrically. Interpretation of these yields will
require further measurements, and, in particular, a study of the independent yields of shielded nuclides ns a function of neutron energy should
be of interest. I n this connection, mass spectrometric methods will undoubtedly be used to determine independent yields both of shielded nuclides and
short-lived isotopes for which satisfactory sweep experiments can be
designed.
REFERENCES
1. Angnrd, P., Anderson, G., Burgman, J. O., and Pappas, A. C., J. Inorg. & Nuclear
Chem. 6, 105 (1957).
2. Aldrich, L. T., Wctherill, G. W., Tilton, G. R., arid Davis, G. L., Ihus. RPU.103,
1045 (1956).
3. Allen, J. S., Phys. Rev, 66,966 (1939).
4. Anikina, M. P., Aron, P. M., Gorshkov, V. K., Ivanov, R. N., Krizhansky, L. M.,
Kukavadee, G. M., Murin, A. N., Reformatsky, I. A., and Ershler, B. V., Proc.
h d Irilarn. Conf. on Peaceful Uses of Atomic Energy, Geneva 16, 446 (1958).
6. Ashizawa, F. T., and Kuroda, P. K., J. Inorg. (e- Nuclear Chem. 6, 12 (1957).
6. Aston, F. W., Phil. Mag. [6] 38, 707 (1919).
7. Aston, F. W., Proc. Roy. SOC.AlS2, 487 (1931).

8. Barnard, G. P., Modern Mass Spectrometry. Institute of Physics, London, 1953.
9. Bartholomew, R. M., Boyd, A. W., Brown, F., Hawkins, R. C., Lournbury, M.,
and Merritt, W. F., Can. J . Phys. 93, 43 (1955).
20. Bay, Z., Nature 141,284, 1011 (1938); Rev. En. Instr. 12, 127 (1941).
11. Bergstrom, I. M., Siegbahn Commem. Volume p. 355 (1951).
f2. Bidinmti, D. R., Fickel, H. R., and Tomlimon, R. H., Proc. 2nd Intern. Conf. on
Peaceful Uses of Atomic Energy, Geneva 16, 450 (1958).
13. Bladw, A. T., and Thodc, H. G., 2.Naturforsch. 10a, 838 (1955).
14. Blades, A. T., Fleming, W. H., and Thode, H. G., Can. J . Chem. 34, 233 (1956).
16. Bohr, A., Proc. 1st Intern. Conf. on Peaceful Uses of Atomic Energy, Ueneva 2 , 151
(1 9.5.5).
16. Brown, F., Ilnll, G. R., and Walter, A. J., J. Znurg. & Nuclear Chem. 1, 241 (1955).
17. Cameron, A. G. W., Can. J. Phys. 36, 1021 (1957).
18. Cameron, A. G. W., Proc. 2nd. Intern. Cmf. on Peaceful Uses of Atomic Energy,
Geneva 16, 425 (1958).
19. Campbell, N. R., and Wood, A., Proc. Cambridge Phil. Sac. 14, 15 (1907).
$0. Classm, J., Jahrb. Hamburg, Wiss. Anst. Seiheft (1907).
21. Cohen, A. A., Phys. Rev. 63, 219 (1943).
22. Cork, J. M., Shriffer, R. G., and Fowler, C. M., Phys. Rev. 74, 1657 (1948).
23. Coryell, C. D., Brightsen, R., and Pappas, A. C., Phys. Rev. 86, 732 (1952).
24. Dempster, A. J., Phys. Rev. 11, 316 (1918).
26. Dcmpster, A. J., Phys. Rev. 71, 829 (1947).
86. Duckworth, H. E., Mass Spcctroscopy. Cambridge University Press, London
and New York, 1958.
361
27. Fic:kt:l, H. R., and lcmlinson, It. II., unpublishcd work, McMmter University,
Hamilton, Canada (1958).
28. Fickel, H. R., and Tomlinson, R. [I., Can. J . Pliys. 37, 531 (1959).
29. Fickel, H. R., and Tomlinson, R. H., Can. J . Phys. 37, 916, 926 (1959).
SO. Fleming, W. H., and Thode, H. G., Phys. Rev. 92, 378 (1953).
31. Fleming, W. H., Tomlinson, R. H., and Thode, H. G., Can. J . Phys. 32, 522 (1954).
S2. Fleming, W. H., and Thode, H. G., Can. J . Chem. 34, 193 (1956).
33. Fong, P., Phys. Rev. 89, 434 (1953).
S4. Frenkel, J., J . Phys. (U.S.S.R.) 10, 533 (1946).
56. Fritzc, K., McMullen, C. C., and Thode, H. G., Proc. 2nd Intern. Conf. on Peareful
Uses of Atomic Energy, Geneva 16, 436 (1958).
36. Fritze, K., McMullen, C. C., and Thode, H. G., unpublished work, McMaster
University, Hamilton, Canada (1958).
S7. Glendenin, L. E., Tech. Rept. No. 35, Laboratory for Nuclear Sci. and Eng.,
M. 1. T. (1953); also Ph.D. thesis M. I. T. (1949).
58. Glendenin, L. E., Coryell, C. D., and Edwards, R. R., in National Nuclear Energy
Series, Div. IV. Vol. 9, p. 489. Paper 52. McGraw-Hill, New York, 1951.
39. Glendenin, L. E., Steinberg, E. P., Tnghram, M. G., and Hess, D. C., Phys. Rev.
84, 860 (1951).
40. Glendenin, L. E., Steinberg, E. P., Flynn, K. F., Hayden, R. J., and Inghram, M.
G., Mass spectrometric data quoted by E. P. Steinberg and L. E. Glendenin, Proc.
1st Intern Conj. on Peaceful Uses of Atomic Energy, Geneva 7, 3 (1955).
41. Grandsen, M. M., and Boyle, W. S., Phys. Rev. 82, 447 (1951).
42. Hay, I. W., and Newton, T. D., Can. J . Phys. 36, 195 (1957).
4s. Hayden, R. J., Phys. Rev. 74, 650 (1948).
44. Hayden, R. J., Reynolds, J. H., and Inghram, M. G., Phys. Rev. 76, 1500 (1949).
46. Hertzberg, G., Spectra of Diatomic Molecules. Van Nostrand, New York, 1950.
46. Hemmendinger, A., and Smythe, W. R., Phys. Rev. 61, 1052 (1937).
47. Hess, D. C., Hayden, R. J., and Inghram, M. G., Phys. Rev. 72, 730 (1947).
48. Hess, D. C., and Inghram, M. G., Phys. Rev. 76, 300 (1949).
49. Hevesy, G., and Paneth, F. A., A Manual of Radioactivity, 2nd ed. Oxford
University Press, London and New York, 1938.
60. Hill, D. J., and Wheeler, J. A., Phys. Rev. 89, 1102 (1953).
61. Hill, D. J., Proc. 2nd Intern. Conf. on Peaceful lises of Atomic Energy, Geneva 16,
244 (1958).
G2. Hill, R. D., P l ~ y s .Rev. 98, 1272 (1955).
63. Hoagland, E. J., and Sugarman, N., in National Nuclear Energy Serics, Div. IV,
Vol. 9, p. 635. Paper 69. McGraw-Hill, New York, 1951.

64. Inghram, M. G., Hess, D. C., and Hayden, R. J., Phys. Rev. 71, 561 (1947).
66. Inghram, M. G., Hayden, R. J., and Hess, D. C., Phys. Rev. 79, 271 (1950).
66. Inghram, M. G., and Chupka, W. A., Rev. Sci. Instr. 24, 518 (1953).
67. Inghram, M. G., J . Phys. Chem. 67,809 (1953).
68. Inghram, M. G., and Hayden, R. J., A Handbook of Mass Spectroscopy, Nuclear
Sci. Series, Report No. 14. National Academy of Sciences, National Research
Council, Washington, D. C., 1954.
69. Inghram, M. G., Ann. Rev. Nuclear Sci. 4, 81 (1954).
60. Katcoff, S., and Rubinson, W., Phys. Rev. 91, 1458 (1953).
61. Katcoff, S., Nucleonics 16, No. 4, 78 (1958).
62. Kennett, T. J., and Thode, H. G., Phys. Rev. 103, 323 (1956).
362
H . G . THODE, C . C. MCMIJLLEN, AND K, FRITZE
63. I<c.nnc4tJ,T. J., and Thocle, 11. G., J. Inorg. ck Nuclear Chetn. 6, 253 (1958).
64. Bennett, 1. J., and Thodr, 11. G., Can. J . thys. 56, 969 (19.57).
66. Kicnbcqer, C. A.,Phys. Ken. 76, 1561 (1949).

66. Krizhansky, L. M., Maly, Ya., Murin, A. N., and Preobrazhensky, B. K. Soviet J.
Atomic Energy 2, 334 (1957).
67. Kumar, I<., and Ircston, M. A., Can. J. Phys. 55, 298 (1955).
68. Lapp, R. E., Van Horn, J. R., and Dempster, A. J., Phys. Rev. 71, 745 (1947).
69. Leachman, R. B., Phys. Rev. 101, 1005 (1956).
70. Leachman, R. B., Proc. 2nd Intern. Conf. on Peaceful Uses of Atomic Energy, Geneva
16, 229 (19.58).
71. Lewis, L. G., and Hayden, R. J., Z2ev. Sci. Znstr. 19, 599 (1948).
72, Lewis, L. G., and Hayden, R. J., Rev. Sci. Instr. 19, 922 (1948).
73. Lindner, M., Phys. Rev. 84, 240 (1951).
74. Macnamara, J., Collii, C. B., and Thode, H. G., Phys. Rtw. 76, 532 (1949).
76. Macnamara, J., and Thode, 11. G., Phys. Rev. 80, 471 (1950).
76. Mwnamara, J., and Thode, H. G., Phys. Rev. 80, 296 (1950).
77. Macnamarrt, J., Collins, C. B., and Thode, H. G., Phys. Rev. 78, 129 (1950)
78. Mattauch, J., Naturwissenschaften 26, 189 (1937).
79. Mayer, M. G., Phys. Rev. 74, 235 (1948).
80. McMillan, E. M., Phys. Rev. 72, 591 (1947).
81. Meitner, I,., Arkiv Pysik 4, 383 (1951).
88. Melaika, E. A., Parker, M. J., Petruska, J. A., and Tomlinson, R. H., Can. J . Chem.
33, 830 (1955).
grS. Michel, M. C., and Templcton, D. H., Phys. Rev. 93, 1422 (1954).
84. Newton, T. D., Atomic Energy Can. Ltd., Chalk River Rept. C-842-A1p. 307 (1956).
86. Nier, A. O., Rev. Scz. Instr. 6, 254 (1935).
86. Nier, A. O., Phys. Rev. 77, 789 (1950); 79, 450 (1950).
87. Pappas, A. C., Proc. 1st Intern. Conf. on Irawful Uses of Atomic Energy, Geneva 7,
19 (1955).
88. Pappm, A. C., Tech. HRpt. No. 63, Laboratory for Nuclear Sci. M. I. T. (1953).
89. Parker, P. L., and Kuroda, P. K., J. Inorg. & Nuclear Chem. 6, 153 (1958).
90. Petruska, J. A., Thode, H. G., and Tomlinson, R. H., Can. J . Phys. 33,693 (1955).
g f . Ptbtruska, J. A., Melaika, E. A., and Tomlinson, R. H., Can. J. Phys. 35, 640 (1955).
$2. Petrshak, I<. A., and Plwov, G. N., J . Exptl. Theoret. Phys. I1.S.R.R. 10, 1013
(1949).
$3. Rall, W., Phys. Rev. 70, 112 (1946).
94. Rittenberg, D., J. Appl. Phys. 13, 561 (1942).
96. Segc, E., Phys. Rm. 86, 21 (1952).
96. St.llers, P. A., Stevens, C . M., and Studicr, M. H., Phys. Rev. 94, 952 (1954).
97. Smythe, W. R., and Hemmendinger, A.,Phys. Rev. 61, 178 (1937).
98. Snell, A. H., Phys. Rev. 62, 1007 (1937).
99. Snell, A. H., Levinger, J. S.,Meiners, E. P., Sampson, M. B., and Wilkinson, R.,
Phys. Rev. 72, 545 (1947).
100. Steinberg, E. P., Glendenin, L. E., I n g h r m , M. G., and Hayden, R. J., Pllys. Rev.
95, 867 (1954).
101. Strassmam, F., and Walling, E., Ber. deut. chem. Ges. 71B, 1 (1938).
102. Swiatecki, W. J., Phys. Rev. 85, 178 (1951).
10s. Terrell, J., Phys. Rev. 108, 783 (1957).
204. Thode, H. G., and Graham, R. L., Can. J . Research A26, 1 (1947); Montreal
Report MX 129 Natl. Research Council of Canada (1945).
363
105. Thode, H. G., NucZeonics 3, 14 (1948).
H. G., Ann. Rev. Phys. Chem. 4, 95 (1953).

107. Thomson, J. J., Proc. Roy. SOC.889, 1 (1914).
108. Vladimirsky, J., J . Exptl. Theoret. Phys. U.S.S.R. 32, 822 (1957).
109. Wahl, A. C., J . Inorg. & Nuclear Chem. 6, 263 (1958).
110. Walker, W. H., Ph.D. thesis, McMaster University, Hamilton, Canada (1954).
111. Walker, W. H., and Thode, H. G., Phys. Rev. 90, 447 (1953).
112. Wanless, R. K., and Thode, H. G., Can. J . Phys. 31, 517 (1953).
113. Wanless, R. K., and Thode, H. G., Can. J . Phys. 33, 541 (1955).
114. Weinberg, A. M., and Wigner, E. P., Physical Theory of Neutron Chain Reactors, p. 106. University of Chicago Press, Chicago, Ill., 1958.
115. Westcott, C. H., Walker, W. H., and Alexander, T. K., Proc. 2nd Intern. Conf. on
Peacefd Uses of Atomic Energy, ffeneva16, 202 (1958).
116. Westcott, C. H., Chalk River Puhl. No. 407, Atomic Energy of Canada Ltd. (1957).
117. Wetherill, G. W., Phys. Rev. 92, 907 (1953).
118. White, F. A., and Collins, T. L., Appl. Spectroscopy 8, No. 4, 169 (1954).
119. White, F. A., Collins, T. L., and Rourke, F. M., Phys. Rev. 97, 566 (1955).
120. White, F. A., Collins, T. L., and Rourke, F. M., Phys. Rev. 101, 1786 (1956).
121. Wilus, D. R., Smith, B. W., Horsley, R., and Thode, H. G., Can. J . Phys. 31, 419
(1953).
122. Wiles, D. M., and Tomlinson, R. H., Phys. Rev. 99, 188 (1955).
123. Wiles, D. M., and Tomlinson, R. H., Can. J . Phys. 33, 133 (1955).
124. Wiles, D. M., Petruska, J. A,, and Tomlinson, R. H., Can. J . Chem. 34, 227 (1956).
125. Wilson, R. R., Phys. Rev. 72, 189 (1947).
126. Young, B. G., and Thode, H. G., Can. J . Phys. 38, 1 (1960).
106. Thode,
AUTHOR INDEX
Numbers in parentheses are reference numbers and are included to assist in
locating references in which the aut,hors names are not mentioned in the text.
Numbers in italics indicate the page on which the reference is listed.
A
Aagard, P., 324 ( I ) , 360
Ablov, A. V., 29 (15), 67
Abrahams, S. C., 238, 241, 267 (31, 269
( l ) , 270 (5), 272 (31, 273
Achterberg, F., 254, 264, 278
Adams, M. D., 217, 233, 290 (1011, 311
Adams, R. M., 214 (771, 236
Agron, P., 199 (21, 233
Ahern, A. J., 110 (11, 148
Ahrens, H., 108 (89), 112 (89), 144 (89),
Ashby, E. C., 76 (62), 101, 302 ( l a ) , 309

Ashizawa, F. T., 351, 360
Askitopoulos, K., 128 (1831, 145 (183),
165
Asmussen, R. W., 255,273

Asprey, L. B., 210, 224, 232, 233
Aston, F. W., 316, 3G0
Atack, D., 108 (4), 109 (41, 111 (4), f4S
Audrieth, L. F., 169 (121, 188 (24), 191
Aurivillius, K. L., 39 (5, 6), 66, 67
Awtrey, A. D., 255 (9), 273
Aynsley, E. E., 131, l 4 S
161
Ahrens, J. V., 88 (561, 101

Aignesberger, A., 121 (135), 125 (88),
144 (135), 145 (881, 160, 162
Aksnes, O., 249 (61,272 (6), 273

Albers, A,, 214 (591, 236
Aldrich, L. T., 326 (2), 300
Alexander, T. K., 331 (115), 363
Allen, J. S., 317, 360
Allen, R. J., 207, 233
Almin, K . E., 42 ( l ) , 56
Alt, B., 265 (222, 2231, 278
Alton, E. R., 286 ( l ) , 309
Amma, E. I,., 83 ( l ) , 99
Andersen, E. B., 255, 273
Anderson, G., 324 ( l ) , 360
Anderson, H. H., 129, 145 (31, 14s
Anderson, J. S., 197, 198, 211 ( 5 ) , 235
Andersson, G., 49 (2, 4), 50 (3, 91),50,5S
Anet, F. H. L., 81 (2), 9<9
Anikina, M. P., 340 (4), 360
Appel, R., 161 ( I ) , 183 (In, 4), 184 (la,
5 ) , 185 (6), 186 (3, 6), 189 (2, 2a, 7),
190 ( S ) , 190
Appenrodt, J., 71 (106), 82 (106), 102
Archer, S., 124 (283), 155
Arnold, M. IT. M., 161 (9, l l ) , 165, 167
B
Backer, H. J., 266, 273
Baden, H. C., 281 (2), 309
Badger, R. M., 112 (5), lQS, 269 (2271,
27s
Bahr, G., 85 (3), 88 (4), 99

Baer, J. E., 245, 273
Bagnall, K. W., 202, 233
Bain, O., 272, 273
Baker, E. B., 95 (5), 99
B a h t , A. E., 88 (loo), 102
Ballhausen, C. J., 12 (97), 16 (go), 21
(13), 67, 58, 59
Bakers, H., 240 (1561, 276

Balz, G., 125 (321), 156
Banus, J., 300 (291, 309
Banus, M. D., 300 (29), 306 (3), 309
Barber, E. J., 212 (91, 233
Bardi, R., 22 (43), 67
Barltrop, J. A,, 268, 271 (141, 273
Barnard, G. P., 317 (8), 3GO
Barnes, W. H., 51 (119), 69
Baroni, A,, 249, 266, 273
Bartholomcw, R. M., 326, 360
Bartlett, N., 11 (7), 25 (7), 67, 114 (61,
(101, 190, 191
Arntson, R. H., 240 (8), 273

Aron, P. M., 340 (4), 360
132 (6), 14s
Bartlett, P. D., 266 (17, 18), 273

Bnrtochn, B., 82 (61, 9.0, 302 (4), 309
36.5
366
AUTHOR INDEX
Ihc;r;eCt, H., 255, 273

Battye, A. k;., 185 ( l l 9 ) , 193
Baudler, M., 241 (58, 59, 601, 269, 270,
f l 3 , 274
Bauer, S. H., 113 (7, 1291, 121 (128),
144 (12S), 148, 161, 272, 276, 281 (6),
284 (51, 309, 311
Baucr, W. II., 281 (2, 97), 509, J i l
Baumgarten, P., 132 (81, 149, 149, 188,
191
Baur, W. H., 10 (9), 11 (8, 91, 19 (91,

50 (91, 67
Bay, Z., 317, 360
Beach, J. Y., 241 (2111, 243, 247 (211),
278
Beachrll, H. C., 283 (7, 8), 309
Bcck, R. B., 139 (1471, 162
Bccke-Gorhring, M. (Goeliring M.), 161
(15, 391, 162, 163 (17, 39, 49, 561, 164
(16, 461, 165 (321, 166 (54, 55, 56),
167 (54), 168 (141, 169 (17, 19, 38,
350, 170 (17), 171 154), 172 (25, 35,
40, 41, 42, 1161, 174 (34, 681, 175 (43,
531, 176 (18, 201, 177 (181, 178 (18,
51), 179 (511, 180 (481, 182 (47, 48,
521, 183 (4), 185 (44, 50), 186 (44),
190, 191, 192, 193, 239, 243 (147),
244, 254, 255, 264, fig, 23'6
Becker, F., 264, 265 (2221, 678
Beevers, C. A., 267 (2131, 27' 8
Beichl, G. J., 295 (9), 309
Belf, L. J., 118 (111, 144 ( l l ) , 14s
Bell, R. P., 281 (101, 284 (lo), 309
Bender, P., 116 (121, 119 (121, 144 (12),
149
Benkeser, R. A., 85 (491, 101
Bennett, F. W., 133, 136, 146 (131, 149
Bennett, M. R., 214 (15), 234
Benson, S. W., 110 (141, 149
Berg, D., 110 (15), 14.9
Berg, W., 160 (191, 176 (201, 1.91
Berger, G., 183 ( l a ) , 184 ( l a ) , 189 (21,
190
Rergson, G., 248 (23, 25), 267, 268, 269,
271 (24, 25, 261, 273, m4
Bergstrorn, I. M., 336 ( I l ) , 360
Bernhsrdt, H. A,, 212 (91, 233
Berthold, H., 240 (61, 621, 241 (61, 64),
246, 274
Bertram, A,, 108 (go), 161
Beaai, S., 22 (43), 67

Bidinosti, D. It., 331 (121, 360
Bieling, C. A., 110 (821, 115 (821, 144
(82), 160
Bigeleisen, J., 225 (lo), 234
Bigelow, L. A,, 133, 138 (310), 139, 144
(3101, 146 (310), 147 (310), 166
Bighi, C., 265, $74
Bijvoet, J. M., 22 (881, 39 (10, 89, 1401,
67, 68, 69
Billy, C., 20 ( l l ) , 67
Binks, W., 49 (121, 67
Birum, B. H., 135 (3081, 147 (308), 148
(3081, 166
Bissot, T. C., 286 (11, 12, 45, 85), 288
(121, 293 (121, 294 (11, 45, 851, 295
(85), 301 (851, 309, 510, 311
Biswas, A. B., 31 (1251, 32 (125), 69
Bjerrum, J., 21 (131, 67
Blades, A. T., 336 (131, 338 (14), 340
(13, 14), 345 (141, 3GO
Blake, E. S., 243, ,974
Bleaney, B., 21 (14), 67
Bloch, I., 241, 274
Block, S., 51 (40), 67
Blomstrand, C. W., 250, 251, 254, 874
Bohme, H., 243 (34), 265, 266, fj74
Bohr, A., 332 (15), 360
Boitner, T. E., 112 (16), 149
Bokai, G. B., 29 (15), 67
Bolz, A., 283 (148, 1491, 286 (148, 149),
288 (148, 149, 1511, 292 (151), 313
Bonch-Bruevich, A. M., 110 (17), 119
f17), 149
Bonner, E. I?., 81 (81, 100
Boone, J. L., 383 (13, 1 3 4 , 300 (30), 909
Booth, H. S., 116 (211, 117 (21), 119,
128 (18, 22), 129 (18, 221, 132 (20, 23,
106), 44 (21), 145 (18, 19, 21, 22),
149, 161, 207, 234
Booth, W., 207, 208 (21), ,934
Borie, B. S., 22 (38), 57
Bowen, H. J. M., 135 (241, 136 (24), 149,
247, 248, ,974
Bowers, K. D., 21 (16), 67
Boyd, A. W., 326 (9),360
Boyle, W. S., 324 (411, 361
Braekken, H., 16 (18), 39 (181, 44 (17).
67
Bragdon, R. W., 306 (31, 309
AUTHOR INDEX
367
Burdon, J., 140 (36), 141 (36), 142 (36),

Braidwood, C. A., 124 (25), 140
147 (36), 149
Brandon, R. J., 217 (66), 235
Brandt, G. R. A., 133 (131, 136 (13), 146 Burg, A. B., 107 (37), 114 (371, 149, 279,
281 (27, 1111, 283 (13, 13a, 36, 37, 38,
(131, 140, 24G, 274
Brandt,, K., 51 (19, 201, 57
431, 284 (16, 31, 1341, 285 (391, 286
(34, 36, 39, 43, 134), 287 (441, 288
Brandt, R. A,, 133, 134, 135 (26), 136
(26, 33, 35, 36, 110, 112, 113), 289
(26), 137, 138 (26), 146 (26, 27), 149
(33, 35, 36, 113), 290 (26, I l O ) , 293
Brater, D. C., 206 (79), 208, 211, 234, 236
(28, 42), 294 (431, 295 (43, 1341, 296
Braun, G., 81 (66), 101
(28a, 31, 41), 297 (34), 298 (16, 38,
Braun, H . J., 123 (256), 155
44), 300 (25, 29, 30, 32, 44, I l l ) , 301
Braune, H., 112 (281, 149
(27, 39, 41a1, 305 (251, 309, 310, 312
Brendel, G., 284 (31), 296 (31), 309
Brice, T. J., 110 (323), 126 (275), 140 Burgman, J. O., 324 ( I ) , 360
(29), 141 (29), 142 (29, 2751, 147 Burhop, E. H. S., 110 (381, 149
Burkardt, I,. A., 294 (511, 310
(29), 148 (29), 149, 155, 156
Burns, R. E., 207 (721, 235
Brightsen, R., 356, 360
Burrill, E. A,, 110 (35), 149
Broadly, R. D., 214 (62), 235
Brockway, L. O., 112 (33), 129 (321, 149 Burwell, R. L., 124 (39), 149
Butler, M. J., 188 (24), 191
Brodskii, A. I., 255 (37), 274
Rykhovskaya, E. G., 135 (1691, 136
Broja, G., 185 (221, 186 (211, 191
(169), 137 (1691, 153
Brokaw, R. S., 75 (9), 100
Brook, A. G., 85 (10, 12, 131, 88 (lo), 100 Bystrom, A., 41 (241, 44 (25, 261, 45
(27), 49 (28), 57
Brooks, 1,. H., 214 (241, 234
Brosset, C., 25 (21), 57
Broteen, O., 45 (27), 67
C
Brown, C. A,, 292 (821, 911
Cacciari, A,, 207 (141, 234
Brown, E. C., 113 (2591, 156
Brown, F., 113 (341, 114 (341, 144 (341, Cady, G. H., 107 (2801, 112 (230), 113
(279, 280), 117 (78, 791, 118 (78, 79),
149, 326 (91, 360
120 (76, 79), 122, 138 (279, 2801, 139
Brown, H. A,, 141 (3001,166
(2791, 144 (76, 77,781, 147 (279), 150,
Brown, H. C., 75 (14, 15), 77 (16), 100,
i54, 156
284 (16), 288 (13b, 171, 298 (16), 299
(50), 302 (21, 22) 304 (241, 305, 306 Cakajdova, I. A., 22 (531, 68
(18, 23, 108, log), 307 (19, 20, 241, Calfee, J. D., 124 (411, 126 (40, 411, 127
308 (14, 15), 309, 310, 311, 312
(411, 145 (41), 149
Calvin, G., 79 (lY), 100
Brown, I. D., 39 (22), 67
Brown, R.. D., 286 (11, 309
Calvin, M., 268 (14), 270 (141, 273
Cameron, A. G. W., 332 (181, 358, 359,
Brown, T. L., 83 (171, 100
Brustier, V., 162 (23), 191
360
Camilli, G., 110 (42, 431, 149
Bryce-Smith, D., 73 US), 100
Buchheit, P., 283 (149), 286 (1491, 288 Campbell, D. H., 141 (195), 163, 286 (11,
45), 294 (11, 45), 309, 310
(1491, 313
Campbell,
G. W., 293 (46a), 300 (32, 461,
Buckler, S. A,, 90 (124, 1251, 91 (1251,
309, 310
108
Campbell, M. H., 112 (1891, 163
Budaev, I. V., 220, $34
Campbell, N. R., 323 (19), 360
Buechner, W. W., 110 (351,149
Campbell, T. W., 245, 248 (192, 1931, 269
Buell, C. E., 108 (2081, 154
(193), 272, 277
Bnerger, M. J. Z., 21 (23), 57
Cardon, S.Z., 288 (17), 309
Bunsen, R., 71 (18:%),700
Carmack, M., 245, 270 (230), 273, 278
Burbank, R. I).,239, 272 (381, ,274
368
AUTHOR INDEX
Camell, P. H., 124 (441, 140

Cmon, A., 168 (28), 101
Carl.renter, F.G., 110 (451, 140
Carrick, W. L.,81 (8), 100
Carroll, J. L., 207 (721, 2S5
Carson, A., 238 (471, 874
Cartor, J. C., 286 ( l ) , 292 (79a), 293
(79n), 30.1, 311
Cirsorio, F. F., 296 (140a), 312
Cassidy, Mary C., 128 (18), 129 (18), 145
(la,149
Calliers, C. I., 214, 2.34
Centnersziver, M., 115 (46, 47), 144 (47),
1.69, 150
Chapman, J. J., 110 (42, 43), l @
Chatt, J., 79 (20), 84 (21), 95 (201, 100,
301 (471, 310
Chenirck, J. A,, 124 (48), 160
Cherniek, C. L., 84 ("21, 85 ($51, 100,lU2
Christ, C. I,., 51 (291, 57
Christensen, E. I,.,221 (161, $34
Christianscn, J. A., 250, 251, 255 (40),
-v4
CIirobak, L. X., 22 (301, 57

Chu, T. L., 95 (23), 100
Chupka, W. A., 317 (561, S61
Chynoweth, A,, 108 (2681, 166
Claasen, A ., 39 (101, 57
Claassen, H. H., 224 (56), 225 (56), 234,
236
Clark, E. S., 25 (31), 67
Clark, H. C., 116 (50), 121 (501, 124 (491,
144 (50), 160
Clark, J. R., 51 (291, 57
Clawen, J., 316, 360
Claudel, B., 74 (40), 100
C ~ X I S SW.
~ RH.,
, 108 (331), 167
Clement, M., 265, 274
Clifford, A. F., 131 (51), 139 (51), 147
(51), 160
Coates, G. E., 63 (24), 74 (25), 79 (19),
90 (241, 94 (26), 100, 124 (52), 160,
299 (48), 310
Cobine, J. D., 110 (53), 150
Coffman, n. D.,
118 (221), 119 (221), 121
(2211, 125 (2211, 126 (2211, 127
(321), 145 (221), 154
Cohen, A. A,, 317 (211, 360
Cohcn, H. L., 100
Cohen, V. W., 128 (131), 151
Collin, It. L., 272 (41, 273

Collins, C. B., 325 (74), 338, 363
Collins, T. L., 317 (120), 319, 363
COlORlUS, H., 71 (1411, 103
Connick, R. E., 255 (9), 273
Coons, E. A., 124 (54), 150
Cooper, P. S., 110 (55, 240), 160, 164
Cork, J. M., 321 (22), 360
Corycll, C. D., 333 (38), 355, 356, 360,
3G1
Cotton, F. A,, 81 (27), 100, 114 (561, 132
(b6), 160
Cowen, H. C., 112 (571,150
Coyle, T. D., 283 (491, 287 (491, 298
(49), 299 (491, 310
Craggs, J. D., 110 (203, 315, 316), 166
Craig, D. P., 112 (58, 59, 60), 160
Crawford, R. A., 118 (61), 160
Croatto, V., 255, 274
Cromcr, D. T., 38 (32), 67
Cscrr, R., 88 (1001, 108
Curmn, C., 84 (281, 100
Czaloun, A., 118 (1331, 121 (133, 1341,
125 (133, 1371, 161, lG2
D
Dakin, T. W., 110 (15), 149
Dale, J. W., 81 (291, 100
Danehower, R. G., 107 (190), 110 (1901,
113 (190), 163
Danner, H. R., 50 (44), 67
Dannhauser, W., 38 (33), 57
Darnell, A. T., 226, 229, 234
Das, S. R., 238, 239 ( 4 3 , ,274
Daum, K. W., 172 (25, 40, 411, 191
Dsvidson, N., 299 (50), 510
Davidson, N. R., 204, 213, 218 (351, 254
Davies, A. C., 91 (301, 100
Davies, D., 199 ( 5 8 ) , 215 (581, 218 (581,
$36
Davies, W., 143 (62, 63), 148 (62, 63, 641,
150
Davis, G. L., 326 (2), 3G0
Davis, R. E., 266 (171, 273
Davis, W., Jr., 218 (781, 635
Dawson, I. M., 247, 252 (43), 253 (431,
269 (431, 270 (43), 2'74
Dawson, J . K., 210, 218 (191, 234
Debo, A., 172, 191
369
AUTHOR INDEX
de Boer, E., 82 (7, 1271, 99, 103

De Carlo, M. R., 113 (2591, 155
Degering, E. F., 185 (261, 191
De Haan, Y. M., 238, 274
de Lattre, A,, 112 (1771, 165
De Leone, R., 207 (141, 254
Delwaulle, M. I,., 129, 145 (6.51, 150
Deniing, I,. S., 42 (941, 58
Demyster, A. J., 316, 327, 3G0, 362
Denbigh, K. G., 113 (66, 671, 144 (671,
150
Dessy, R . E., 84 (311, 93 (32, 33, 1361,

94 (32, 331, 100, 103
de Wet, J. F., 23 (1351, 59
D'Eye, R. W. M., 197, 198, 207, 208 (21),
2393, 234
Dick, J. H., 143 (62, 631, 148 (62, 63, 64)
160
Dickson, F. W., 240 (81, 273

Diebler, V. H., 110 (681, 150
Dimrotii, O., 71 (341, 100
Divers, E., 181, 191
Dobretsov, L. N., 110 (69), 160
Docid, R. E., 114 (70), 144 (70), 160
Domnnge, L., 116 (2391, 118 (2391, 154
Domash, L., 288 (13b1, 309
Donohue, J., 165 (1051, 168 (281, 191,
193, 238, 239, 247, 269 (461, 274, 277
Dougherty, G., 245, 276
Douglass, R. M., 223, 224 (221, 234
Douslin, D. R., 268 (146), 276
Dresdner, R. D., 138 (72, 731, 139 (73,
741, 144 (741, 146 (731, 147 (72, 73),
150
host-Hansen, W., 255 (40), 274

Dubnikov, 1,. M., 115 (751, 128 (3131,
145 (313), 160, 156
Duckworth, H. E., 318 (26), 360
Dudley, F. B., 117 (78, 791, 118 (61, 78,
791, 120 (76, 791, 122, 144 (76, 77,
781, 150
Duncan, A. B. F., 109 (80, 231, 2951, 112
(80, 231, 295), 144 (2951, 150, 164,
156
Dunitz, J. D., 15 (351, 17 (361, 19 (351,

20 (351, 22 (34, 351, 23 (341, 25 (36),
26 (361, 27 (361, 28 (36, 371, 31 (371,
39 (22, 341, 57
Dupont, J. A,, 303 (671, 3f0
Durrant, R. G., 255, R3
Dust, H. P., 214 (591, 235

Dyer, L. D., 22 (381, 57
Dewius, M., 291 (751, 311
E
Earwicker, G. A., 109 (811, 160
Ebenhoch, F. L., 125, 156
Ebert, F., 112 (2541, 175 (43), 166
Ebert, J., 175 (43), 191
Eddy, L. B., 292 (1051, 311
Edelson, D., 110 (82), 112 (831, 115 (821,
144 (82, 83, 841, 160
Edmonds, L. O., 265, 276
Edstrand, M., 43 (391, 67
Edwards, A. J., 16 (39a1, 67
Edwards, L. J., 283 (%a), 285 (721, 293
(73, 741, 294 (531, 298 (73, 741, 310,
511
Edwards, R. R., 333 (381, 355, 361

Eggers, D. F., 117 (781, 118 (781, 120
(761, 144 (76, 781, 160
Egorova, N. L., 185 (771, 192
Ehrmann, K., 107 (306), 114 (3061, 115,
119 (3061, 144 (3061, 166
Eisch, J., 75 (1421, 76 (441, 77 (1421, 78
(1421, 86 (45), 90 (1421, 92 (1421, 93
(1421, 95 (142), 100
Eisenhauer, G., 189 (2a1, 190
Eisfeld, K., 88 (601, 101
Ellinger, F. H., 224 (61, 233
Elliot, R. M., 210 (191, 218 (19), ,934
Ellis, D. A,, 112 (141, 149
Ellis, J. F., 212 (231, 214 (241, 2.34
El-Shamy, K . H., 138 (511, 139 (511, 147
(511, 160
Elson, R. E., 205 (751, 236
EmelBus, H. J., 81 (29, 351, 100, 116 (50),
118 (861, 119 (861, 121 (501, 124 (491,
133 (13, 271, 134 (261, 135 (13, 26,
851, 136 (261, 137 (261, 138 (26, 511,
139 (511, 144 (501, 146 (13, 26, 271,
147 (511, 14.9, 160, 212, 234, 246, W4,
281 (101, 284 (101, 291, 298 (1351,
309, 310, 312
Emery, A. R., 290 (136), 312

Emilius, M., 307 (1021, 311
Engelbrecht, A., 121 (87, 1351, 125 (881,
144 (1351, 145 (87, 881, 150, 162, 190
(631, 192
370
AUTHOR INDEX
Ephr:tiin, P., 182 (31), 184, 185 (291, 186

(30), 191
Epple, R., 96 (53), 101
Eriks, K., 255, 276
Ershler, B. V., 340 (41, 360
Eucken, A., 108 (89, 90, 92), 112 (89,
91), 144 ("9, I d 1
Eucken, M., 255, f 7 4
Evans, H. T., 50 (41), 51 (29, 401, 67
Evans, J. E. F., 308 (150), 913
Evans, W. H., 229 (671, 231 (67), 235
Evenhuis, N., 266 (lo), 273
Evering, B. L., 124 (247, 2481, 166
Evers, E. C., 295 (9), 309
Eyring, L., 209 (881, 2%
F
Faessler, A,, 165 (32), 191, 255, 974
Fairbrother, D. M., 84 (361, 100
Fanclli, R., 112 (931, 161
Farazmand, I., 140 (361, 141 (36), 142
(36), 147 (36), 149
Fnrrar, W-.V., 248, 274
Faurote, 1'. D., 134 (941, 135 (941, 136
(94), 141 (941, 161
Fava, A., 245, 248 (53), 255 (41, 521, QQ
Fear, E. J. P., 109 (811, 160
Feay, n. C., 210 (271, 234
FehCr, F., 240 (61, 62, 71), 241 (58, 59,
60, 61, 641, 242 (68, 72, 73, 74, 78),
243, 244 (65, 67, 69, 721, 246 (631,
247, 248, 249, 255 (721, 264, 265, 266,
267, 368, 269, f l 4 , 276
Feldmann, U., 64 (1331, 276
Fcrguson, R. C., 116 (95), 151
Ferris, L. M., 213 (28, 29), 234
Fetter, N. R., 294 (51), 510
Fickel, H. It., 331 (121, 341 (291, 343
(27, 291, 345 (27, 291, 560, 861
Figgis, B. N., 22 (42), G7
Finholt, A. E., 306 (1081, 312
Finke, H. I., 246 (204, 205), 268 (204,
2051, 278
Fischer, A. K., 76 (38), 100
Fischer, J., 113 (96), 114 (961, 144 (96),
161, 213 (30)' 234
Fitch, S. J., 294 (52), 310
Fizzotti, C., 207 (14), 254
Fleming, W. H., 338 (14), 340 (14, 31,
321, 344 (31, 321, 345 (14, 311, 348

(30), 349 (30), 860, 361
Flcrov, G. N., 347, 362
Florin, A. E., 215, 254
Florio, P. A., 124 (41), 126 (41), 127
(411, 145 (411, 1.49
Pluck, E., 161 (15), 162 (151, 172 (351,
174 (341, 179, 191
Flynn, K. F., 340 (401, 361
Fohr, M., 267 (16G), H7
Piippl, H., 241, 276
Fokin, A. V., 134 (171), 135 (170, 171,
172), 136 (171), 137 (1711, 155
Fong, P., 332 (33), 361
Ford, M. D., 293 (73, 74), 294 (531, 298
(73, 741, 510
ForsyLh, M. W., 294 (53), 310
Foss, O., 244 (87, 971, 247 (104), 249 (6,
82, 941, 250 (82, 86, 881, 251 (82, 88,
89, 90, 95), 252 (89, 93, 98, 1231, 253
(88, 90, 95, 109, 110, 1791, 254 (84,
851, 255 (82, 87, 96), 256 (108), 257
(91, 93, 99, 101), 258 (91, 103, 106,
ill), 259 (114, 115, 116, 117, 118,
119, 120, 124), 260 (115, 119, 1241,
261 (106, 114, 120), 262 (95, 102, 105,
l l l ) , 263 (107), 266 (83, 87, 961, 267
(87, LOO, 119), 269 (99, 101, 1041, 270
(92, 94, 98, 102, 107, 115, 1241, 271
(112, 121, 1221, 272 (6, 105, 106, 109,
113, 114, 120, 123, 179), 273, 275, 276,
277
Fowler, C. M., 324 (22), 360
Fraenkcl, G. K., 238, 276
Francis, K. E., 199 ( 5 8 ) , 215 (581, 218
(58), 236
Fmncois, F., 129, 145 (65), 150
Fwnk, A,, 197 (44), 198 (44), 236
Frankland, E., 62 (391, 71 (391, I00
Frasson, E., 22 (43), 67
Frazer, B. C., 43 (124), 50 (44, 124), 67,
69
Fredga, A., 270 (1251, 276
Freeman, J. H., 221, 234
Fremdling, H., 267 (166), 277
Frenkcl, J., 332 (341, 561
Freundlich, W., 74 (40), 100
Fried, S., 204, 205 (751, 213, 218 (35),
224 (61, 233, 284, 236
Friedel, R. A., 93 (91a), 10.2
AUTHOR INDEX
Friedman, H. L., 109 (97), 161

Fristrorn, R. M., 119 (981, 161
Fritz, G., 96 (411, 100
Fritze, K., 330 (351, 341 (351, 344 (361,
361
Fuoss, R. M., 109 (991, 151

Furbcrg, S., 251 (1261, 252 (981, 257 (991,
269 (991, 270 (981, ,275, 276
G
Gabaglio, G. M., 207 (141, 234
Gadsby, G. N., 73 (421, 76 (42), 100
Galiano, R. J., 299 (941, 311
Gall, J. F., 107 (100, 101, 2111, 108 (212),
109 (2121, 110 (2111, 111 (2121, 113
(100, 101), 114 (100, 2111, 123 (2111,
125 (2111, 126 (211), 128 (2111, 145
(2111, 161, 154, 180 (961, 192
Gallaghan, J., 300 (54, 1261, 310, 312
Galpern, H. N., 217 (661, 235
Gamble, E. L., 108 (2711, 109 (271), 144
(2711, 156, 295 (551, 310
Garcia-Fernandez, H ., 162 (23), 191
Gardner, D. M., 238, 276
Gaunt, J., 112 (102, 1031, 144 (1021,
161, 225, 234
Gaylord, N. G., 280 (561, 307 (561,310
Geballe, R., 110 (1051, 138 (104, 105, 123,
1241, 161
Gee, G., 161 (361, 191, 238, 27G
Gehrig, O., 180 (371, 191
Geisel, E., 169 (1041, 171 (1041, 174
(1041, 1.93
Gellert, H-G., 99 (1431, 103
George, A,, 268 (1461, 276
George, J. W., 114 (561, 132 (561, 150
Gcrber, H., 186 (31, 190
Gerding, H., 255, 276
Germann, A. F. O., 132 (1061, 161
Gerow, C. W., 76 (47), 85 (461, 86 (46a),
100
Gibanova, T. A., 124 (2581, 166

Gihbins, S. G., 283 (56a), 310
Gibbons, A. J., 88 (1021, 102
Gibson, G., 221 (391, 234
Gibson, J. W., 110 (1071, 151
Gielcs, K., 264, 265 (222, 2231, ,2778
Gigukre, P. A., 272, 273
Gilbert, D. X., 287 (154), 298 (1541, 313
371
Gilbreath, J. R., 306 (181, 309

Gillespie, R. J., 28 (451, 67
Gilman, H., 63 (431, 66 (431, 68 (50a1,
69 (49a, 69a), 70 (70),73 (851, 76 (44,
47, 851,85 (12, 13, 46, 49), 86 (43, 45,
46a), 89 (481, 92 (501, 100, 101
Gilmont, P., 295 (551, 310
Girardot, P. R., 290 (87, 1221, 811,312
Glemser, O., 130, 131 (108, 109, 110, 1111,
146 (108, 109, 110, 1111, 161
Glendenin, L. E., 333 (381, 340 (40, 1001,
344, 346, 353 (371, 355, 356, 358, 361,
962
Glikina, M. V., 110 (17, 1511, 118 (171,
149
Glockling, F., 74 (251, 89 (51, 521, 100,

101, 124 (521, 160
Goble, A. G., 205 (40), 234
Goddard, D. R., 126, 161
Goebel, A., 248, $77
Goehriny, M., see Becke-Goehring, M.
Golden, J., 205 (401, 234
Goldish, E., 168 (281, 191, 238 (471, 274
Goldstcin, L., 109 (112), 161
Good, C. D., 288 (331,289 (331, SO9
Gordy, W., 128 (322), 166
Gorin, G., 248 (1601, 270 (1601, 276, 277
Gorshkov, V. K., 340 (4), 360
Gorskii, V. S., 39 (461, 57
Goubeau, J., 93 (1211, 96 (531, 101, 102
Gowenlock, B. G., 85 (541, 101
Gowland, T. B., 119 (1131, 151
Graf, R., 186 (601, 189 (57, 58, 591, 1.92
Graham, H., 161, 196
Graham, R. L., 325 (1041, 334, 335 (1041,
336, 344 (1041, 362
Graham, W. A. G., 284 (591, 287 (581,
299 (581, 301 (581, 302 (4), 309, 310
Gramstad, T., 140 (114, 1161, 141 (114,
115, 116), 142 (114, 115, 1161, 143
(114), 147 (114, 116), 148 (114, 1161,
161
Grandsen, M. M., 324 (411, 361

Grant, L. R., 286 (341,297 (341, SO9
Grant-Mae Kay, J. A., 123 (213), 154
Gray, L. H., 112 (117), 151
Grdenic, D., 39 (47, 481, 40 (1231, 67, 68
Greenberg, L. A,, 110 (187), 113 (1871,
163
Greenwald, S., 19 (491, 67
372
AUTHOR INDEX
Gregg, A. H., 39 (50), 57

Grenall, A., 199 (2), 233
Griesshammer, H., 81 (89), 106
Griffith, J. S., 12 (51), 13 (51), 65
Grignard, V., 71 (55), 101
Grinberg, A. A,, 255, 2Y6
Grison, E., 241, 270 (5), 873
Grissard, J. W., 227 (63), 235
Gross, G. C . , 185 (261, 191
Grow, K. J., 250, 277
Gross, M. E., 246 (204, 2051, 268 (204,
205), 975
Gross, S. T., 112 (332), 157
Crosse, A. V., 135, 136, 161, 214 (41), 334
Groszos, S. J., 292 (601, 310
Cruen, D. M., 221 (39), 232 (42), 234
Gurewitsch, M., 185 (291, 191
Gurynnova, E. N., 245 (137), 976
Gustison, R. A,, 212 (9), 233
Guter, G. A., 285 (61, 621, 310
Guthrie, G., 243, 676
Guthrie, G. B., 239, 243, 246 (139, 204),
268 (139, 204), 276, 278
Cutowsky, H. S*,
112 (118), 151
H
Hndlw, I$., 252 (98), 270 (98), 176
Hrtendler, I%.,
20 (111, 67
Haesler, H., 130 (109, 1101, 131 (108, 109,
1101, 146 (108, 109, 1101, 161
Hagemann, F., 205 (36), 634
Hahn, H., 148 (262), 166
Haines, P. G., 143 (202), 148 (2021, 154
Hall, C. R., 326 (16), 3GV
Halls, E. E., 110 (119), 151
Ham, N. S., 148 (120), 161
Hambiy, -4.
N., 148 (120), 151
Hamilton, W. C., 295 (63), 296 (631, 810
Hammann, S. D., 112 (121, 122), 151
Hampson, C. C., 39 (50), 42 (521, 67, 68
Handler, G. S., 84 (31), 100
Hanic, F., 22 (531, 65
Hannay, N. B., 110 ( l ) , 145
Hantxsch, A., 183, 184, 196
Hardtmann, M., 126 (145), 16Z
Harker, D. Z.,22 (541, 65
Harmsen, E. J., 39 (55), 68
Harris, J. F., 245, 265, 976
Harris, J. J., 292 (991, 311
Harrison, B. C., 304 (64), 310

Harrison, M. A., 110 (1251, 138 (123,
124), 161
Harrison, P. M., 247, 277
Hartman, B. F., 124 (126), 161
Hartmann, H., 88 (56, 57). 101
Hartough, H . D., 124 (127), 161
Harvey, K. B., 39 (47), 43 (56), 68
Harvey, R. B., 113 (129), 121 (1281, 144
(128), 161, 272, 676
Hassei, O., 39 (57), 68
Haserldine, R. N., 81 (29, 35, 581, 100,
101, 133 (13, 271, 134 (26, 138, 141),
135 (13, 26, 1411, 136 (26, 138, 139,
140, 141), 137 (26, 138, 141), 138 (26,
511, 139 (511, 140 (114, 116, 139,140),
141 (114, 115, 116, 1391, 142 (114,
115, 116, 140), 143 (1141, 144 (138),
146 (13, 26, 27, 138, 1411, 147 (51,
114, 116, 139, 1401, 148 (114, 1161,
149, 160, 161, 156, 245, 246, 274, 676
Hattori, K., 300 (128), 316
Hauptschein, M., 135, 136, 161
Hawkins, N. J., 128 (1311, 151
Hawkins, R. C., 326 (91, S60
Hawthorne, M. F., 283 (68), 285 (65),
297 (69), 302 (66), 303 (67), 310
Hay, I. W., 358, 362
Hayden, R. J., 323 (431, 324 (47, 71, 72),
329 (54), 339 (55), 340 (40, 44, loo),
361, 366
Hayek, E., 118 (133), 121 (132, 133, 134,
135, 136), 125 (88,133, 1371, 144 (132,
135, 1361, 145 (B),160, 161, 152, 190,
192, 197, 198, 234, 636
Haycs, P. M., 268 (14), 271 (141, 973
Hayes, W., 33 (58), 55
Head, E. L., 74 (591, 101
Heath, D. F., 112 ( l a ) , 162
Heath, R. E., 207 ( l l ) ,234
Hedherg, K., 118 (61), 160, 288 (701, 310
Hefferan, G. T., 294 (53), 310
Heilmann, E. L., 188 (72), 192
Hein, F., 79 (611, 88 (60), 101
Heinke, J., 185 (441, 186 (441, 191
Heinze, G., 183, 192
Heinxelmann, H., 214 (68), 236
Helmholz, L., 20 (601, 27 (61), 43 (59),
58
Hemmendinger, .4.,323, 362
373
AUTHOR INDEX
Hcndricks, S. U., 43 (94), 68

Henkel, P., 108 (143, 1681, 109 (144, 1681,
144 (168), 152, 163
Hennig, H., 132, l @
Hennion, G. F., 76 (62), 101
Hensel, W. E., Jr., 214, ,236
Hentrich, W., 126 (1451, 152
Hepworth, hl. A,, 21 (62), 58
Herb, H., 161 (451, 191, 239 (1341, 276
Herman, D. F., 80 (631, 101
Herniann, J. A,, 221 (45), 235
Hcrold, J., 116, 166
Herrmann, C. V., 119, 145 (19), 149
Hertlein, H., 255, 276
Hertwig, K., 292 (1511, 313
Hertzberg, G., 317 (45), 361
Herwig, W., 81 (641, 88 (64), 101, 103
Hess, D. C., 319, 324, 329 (48, 54), 339
(551, 344 (39), 561
Hesse, K . R., 124 (3331, 157
Hevesy, G., 321, SG1
Hewitt, F., 286 (71), 310
Hieber, W., 72 (65), 81 (66),101
Hill, D. J., 332 (50, 5 1 ) , 361
Hill, R. D., 332 (521, 361
Himel, C. M., 265, 276
Hinckley, A. A., 306 (3), 309
Hindmann, J. C., 198 (461, 235
Hinshelwood, C. N., 112 (1641, 152
Hites, R. D., 304 (641, 310
Hongland, E. J., 336, 36'1
Hoard, J. L., 6 (64), 38 (63), 68
H6hn, F., 241, 274
Hoerenz, J., 118 (3031, 125 (3031, 126
(3031, 166'
Hocrger, F. D., 148 (309), 166
Hoffs, E., 124 (146), 162
Hoffman, C. J., 112 (118), 152
Hoffman, F. W., 139 (147), 162
Hoffmann, C. M., 148 (2351, 164
Hofmann, W., 42 (651, 68
Hofmann-Bang, N., 178 (691, 192
Hohenschutz, H., 164 (461, 191
Hohnstedt, L., 292 (105), 311
Hokhberg, B. M., 110 (17, 148, 149, 150,
151, 152, 153, 1541, 118 (17), 149,162
Hall, A., 183, 192
HolIer, H . V., 139 (147), 162
Holler, C . E., Jr., 74 (59), 101
Hollidny, A. I<., 286 (71), 310
Hollies, N. R. S., 113 (159, 144 (1551,

162
Hollingsworth, C. A., 84 (31), 93 (32, 33,
136), 94 (32, 331, 100, 103
Holmes, A., 79 (1191, 108
Holmes-Walker, W. A., 82 (67), 101
Homayr, J., 197 (481, 198 (481, 236
Honig, H., 88 (57), 101
Hood, A,, 130 (319), 145 (3191, 166
Hooge, F. N., 243, 246, 276
Hoppe, R., 16 (65a), 58
Hopwood, W., 110 (3151, 156
Hordvik, A,, 257 (101), 262 (1021, 267
(loo), 269 (IOI), 270 (102), 276
Hornig, D. F., 112 (260), 155
Hornung, V., 250, 278
Horsley, R., 325 (121), 344 (121), 363
Hoshino, S., 38 (66), 58
Hossenlopp, I. A,, 268 (146), 276
Hough, W. V., 285 (721, 291 (75), 293
(73, 74), 294 (531, 298 (73, 74), 310,
311
Hovey, A. G., 124 (25), 149

Huhbard, W. N., 268 (146), 276
Huher, W., 184 (5), 185 (6), 186 (6), 190
Hiickel, W., 115 (1561, 162
Hiimpfner, K., 94 (117), 102
Huffman, H. M., 246 (2041, 268 (204),
BS
Hughes, E. D., 126 ( l l l a ) , 161
Hughes, E. W., 247 (232), 269 (232), 23'8
Hugill, A. J. C., 161 ( I l ) , 191
Hurd, D. T., 63 (101), 64 (1011, 72 (69),
73 (68), 75 (681, 86 (101), 101, 102,
302 (76), 311
Hurst, G. S., 112 (16), 149
Hurst, R., 199 (581, 210 (19), 215 (58),
218 (19, 581, 234, 235
Hutson, J. M., 161 ( l l ) ,191
Hvoslef, J., 39 (571, 58
Hyde, E. K., 306 (1091, 318
Hyde, J. L., 112 (189), 153
Hyman, H. H., 212 (471, 214 (7'0, 235
I
Ibamoto, H., 22 (77), 58
Ijdo, D. J. W., 16 (66a), 55
Her, R. K., 123 (160), 152
Ilireto, A,, 245, 274
374
AUTHOR INDEX
Ininiig, H., 182 (651, 192
Inge, M., 43 (39), 67

Ingersoll, L. R., 112 (1611, 152
Inghram, M. G., 317, 319, 322 (591, 323
(571, 324 (47), 329 (48, 541, 339, 340
(40, 44, 59, IOO), 344 (391, 361, 36.2
Ingold, C. K., 126 ( I l l s ) , 161
Ingri, N., 43 (39)' 67
Inman, C. E., 143 (2021, 148 (202), 15,i
Ipatieff, V. N., 123 (163), 124 (162, 163),
,Jonkc, '1.ti.,207 (491, W G

Jflrgensen, C. li., 12 (72), 21 (131, 34
(731, 36 (73),37 (731, 67
K
Ktiesz, H. D., 283 (491, 287 (491, 298
(491, 299 (491, 310
Kaloumenos, H.-W., 180 (481, 182 (47,
48), 191
Kanamori, J., 19 (74), 86

Kanda, F. A., 188 (71), 192
Ivanov, R. N., 340 (41, 3G0
Ksrsscn, A,, 39 (101, 67
Katcoff, S., 330 (el), 336 (611, 337 (611,
J
338 (601, 340 (60, 611, 341 (35), 361
Kats, J. J., 112 (2491, 166, 195 (511, 212
Jach, J., 112 (164), 162
(47), 214 (77), 221 (39, 51), 223, 22G,
Jack, K. H., 20 (671, 21 (621, 68
229, 231, 234, 236, 306 (18), 309
Jaenckncr, W., 113 (961, 114 (961,144
Katz, T., 139 (147), 162
(961, 161
Kazakova, Y. B., 298 (156), 313
Jaffe, I., 229 (67), 231 (671, 235
Kealy, T. J., 63 (71), 71 (71), 101
Jahn, H. A., 15 (68),68
Kcen, R. C., 273, 276
JLthns, W., 112 (2851, 165
Keenan, T.K., 210 (7), 232, 2S3
Jahr, J., 258 (103), 276
Keggin, J. F., 51 (751, 68
Jander, G., 182 (651, 192
Kchl, W. L., 271, 276
Jantsch, G., 197, 198, 2%
Keicher, G., 308 (155), 313
Jeffrey, G. A., 188 (66), l D Z , 271, 276
Kemp, A., Jr., 148 (224), 16.4
Jellinck, P., 21 (69), 68
Kemp, W. E., 124 (1661,162
Jenkins, G. I., 39 (501, 67
Jcnne, H., 161 (151, 162 (15), 164 (16), Keneshea, F. J., Jr., 226 (IS), 229, 2.34
167, 168 (14), 174 (681, 192, 192, 243 Kennett, T. J., 320 (62), 321 (62), 331
(631, 340 (641, 345 (64), 346, 353,
(147), 244, $76
354 (621, 356, 357, 358 (62), 361, 302
Jensen, K. A., 178 (691, 192
Jensen, L. H., 247 (210), 269 (2101, 278 Kerker, M., 255 (150), 276
Johnsen, J., 247 (104), 262 (1051, 269 Ketelaar, J. A. A., 2 (76), 68, 188 (72),
192, 243, 246, 256 (1511, $76
(1041, 272 (1051, f l 6
Johnson, K. D. B., 211, 212 (23), 214 Kharasch, M. S., 63 (72), 97 (72), 101
Kidd, J. M., 133, 134 (138, 141), 135
(24), 234, 236
(141), 136 (138, 139, 140, 141), 137
Johnson, L., 293 (46a), 310
(138, 141), 140 (139, 1401, 141 (139),
Johnstone, N. B. B., 112 (2191, 164
142 (1401, 144 (1381, 146 (138, 141),
Jollep, R. L., 214 (151, 234
147 (139, 1401, 162, 245, 246, 276
Jona, F., 43 (701, 68
Jonas, H., 116 (165), 117 (165), 120 (165), Kienberger, C. A., 348 (651, 362
144 (165), 145 (165), 162, 189 (701, Kigel, R. A,, 298 (166), 313
Kimura, M., 225 (71), 236
192
King, K. J., 188 (711, 192, 241, 277
Jones, D. P., 265 (189), 277
Kiriyama, R., 22 (77), 68
Jones, E. A., 112 (176), 144 (176), 163
Kirman, W., 123 (2981, 145 (298), 166
Jones, P. L. F., 39 (50), 67
Jones, R. E., 43 (711, 68
Kirsanov, A. V., 184 (73, 741, 185 (77),
188 (75, 76), 189 (73, 741, 192
Jones, R. G., 69 (69a), 70 (70), 89 (48),
100, 101
Kisliuk, P., 139 (l67), 162
162
375
AUTHOR INDEX
Klcin, M. J., 300 (1281, 304 (641, 310,

312
Kleinberg, J., 169 (121, 191
Klemm, W., 108 (143, 1681, 109 (144,
1681, 144 (168), 152, 169
Kluiber, R. W., 81 (81, 100
Knoll, H., 182 (651, lC92
Knoke, S., 112 (281, 149
Knox, K., 20 (78),55
Knunyants, I. I,., 134 (1711, 135, 136, 137
(169, 171), 163
Koch, H. P., 245, 246, 269, 2YG
Koch, W , 161 (451, 163 (491, 191, 239
(1341, 27G
Kodama, G., 290 (881, 292 (77, 78, 79~ 1,
293 (77, 79a, 861, 294 (791, 311
Koenig, F. J., 290 (1011, 311
Koster, R., 75 (731, 77 (73), 91 (74), 101,
286 (801, 302 (811, 311
Kohman, G. T., 110 (821, 115 (821, 144
(821, 150
Kolditz, I,., 122, 123 (186), 127 (185), 132
(186), 145 (186), 155
Koller, W., 121 (1361, 144 (1361, 262
Koo, C. H., 273 (2151, 2YS
Kosak, A. I., 124 (127), 162
Koshsr, R. J., 139 (1471, 162
Koski, W. S., 128 (1311, 151
Kosyakov, V. N., 226, 229, 2356
Krasny-Ergen, W., 207 (111, 234
Krause, E., 63 (761, 101
Krause, G., 244 (65), 247, 267 (65),dY4
Krebs, H., 239, 240, 272, 2YG
Krismer, B., 118 (133), 121 (1331, 125
(1331, 161
Krizhansky, L. M., 340 (4, 661, 360, 962
Krstanovic, I . R., 39 (481, 49 (78a1, 67,
58
Kruh, R. F., 27 (611, 68
Kruse, F. H., 248, 272 (157), dY6
Kruse, W., 244 (671, 265, 266, R'4
Ksanda, C. J., 19 (1331, 20 (1331, 69, 256
(218), 258 (218), 278
Kuhlmnn, C. W., Jr., 206 (521, 236
Kukavadze, G. M., 340 (4), 960
Kuljisn, F,. S., 288 (351, 289 (35), SO9
Kumar, I<., 357, 302
Kunin, R., 199 (2), 233
Kuroda, P. K.,
351, 3e0, 362
Knrtcnacker, A,, 254, 264, 276, $77
L
Labaton, V. Y., 211, 212, 236
Larhman, F., 114 (175), 153
Lagemann, R. T., 112 (176), 144 (1761,
163
Lambert, J. A,, 112 (121, 1221, 161
Lange, G., 83 (135), 103
Lange, W., 107 (178a1, 123 (3041, 124
(180, 1841, 125 (1781, 126 (179, 180,
1841, 127 (180, 184), 127 (182, 1831,
145 (180, 183, 184), 163, 266
Langham, W. H., 202 (551, 235
Lanpher, E. O., 95 (87, 881, 102
Lapp, R. E., 327, 362
Laquer, H. L., 245 (2251, 278
Lardicci, T,., 95 (971, 10.9
Larsen, E. R., 139 (1471, 26.9
Larssen, P. A,, 258 (1061, 261 (1061, 272
(106), $76
Laubengayer, A. W., 292 (821, 294 (52),
810, 911
Laue, W., 242 (68), 244 (67, 691, 274
Laurie, C. M., 84 (771, 101
Lauterbur, P. C., 140 (2231, 264
Leachman, R. B., 332, 359, 362
Lebeau, P., 107 (2161, 115, 117 (217), 118
(217), 119 (2171, 144 (2171, 154, 180
(97, 98, 991, 192, 193
Leblanc, E., 81 (21, 99
Lecher, H., 248,flY
Lehmann, H. A,, 122, 123 (186), 127
(1851, 132 (186), 145 (1861, 163
Lehmann, W. J., 299 (94), 311
Lehmkuhl, H., 82 (1451, 103
Lemons, J. F., 21.5 (31), 234
Lengfeld, F., 167, 174 (781, 192
Lesbre, M., 78 (78), 101
Lestcr, D., 110 (187), 113 (187), 155
Letsinger, R. I,., 85 (79), 101
Levi, T. G., 168 (811, 174, 292, 243 (163),
277
Levine, S., 229 (671, 231 (67), 236
Levinger, J. S., 336 (99), 362
Levitz, N., 207 (49), 236
Lewis, L. G., 324 (71, 72), 362
376
AUTIIOB INDEX
McCann, H. G., 119 (2011, 163

McClure, 1). S., 17 (871, 68
Maccoll, A,, 112 (601, 160
MacCormack, K. E., 108 (2041, 109
(204), 111 (204), 112 (193, 194), 153,
164
McCormick, N. R., 110 (2031, 164
McCoy, C., 143 (2021, 148 (202), 164
McCoy, R. E., 309, 311
McCullough, J. D., 47 (87a1, 48 (87a1,
49 (87a), 68, 239 (170), 248, 272,
(157, 1701, 877
McCullough, J. P., 246 (2051, 268 (146,
205), 976, 278
McCusker, P. A,, 76 (621, 84 (281, 100,
101
McElroy, A. D., 285 (721, 310
MacGillavry, C. H., 22 (881, 25 (311, 39
(89), 67, 68
McGrsth, W. D., 112 (205), 164
McHard, J. A,, 112 (2061, 164
McIntosh, I,. R., 110 (351, 142
McIntosh, R. L., 113 (155, 307), 144
(1551, 162, 156
Mack, J. L., 300 (126), 31%
McMillan, E. M., 327 (go), 362
McMullen, C. C., 330 (351,341 (351, 344
(361, 381
Macnamara, J., 325 (74), 331 (761, 338,
347 (75), 362
McOmie, J. F. W., 265 (1891, ,277
Maddock, A. G., 205 (401, 212 (251, 23j
MagnEli, A,, 49 (4), 50 (911, 66, 68
Mngno, P. J., 207 (3), 233
Magnumon, E. A., 112 (58, 591,150
Mahan, J. E., 112 (1961, 163
Mahlcr, W., 296 (411, 310
Maier, L., 82 (84), 101
Maitland, R., 11 (7), 20 (671, 25 (71, 67,
68
Maki, C.,23 (92), 58
Makowski, H. S., 84 (28), 100
Malhotra, S. C., 197 (841, 236
Malinoaski, T. I., 29 (151, 67
Malm, J. G., 215 (571, 216, 217 (85), 224,
M
226 (171, 226 (871, 227 (871, 228
(571, 230, 232 (421, 234, 236, 236
McAfee, K. B., 112 (831, 144 .(83), 160
McBec, E. T., 135 (3081, 141 (1951, 147 Maly, Ya., 340 (66), 362
Malz, H., 176 (20, 861, 185 (44, 50, 861,
(308), 148 (3081, 163, 166
186 (44), 191, 192
Macbeth, A, K., 161, 192
Lewis, R. N., 63 (1011, 64 (1011, 86

(101), 102
Li, Y. Y., 19 (79), 6s
Ikhtenwaltcr, M., 85 (491, 101
Liebenberg, D. H., 112 (1611, 162
Liehr, A. D., 16 (801, 68
Lindner, M., 324 (73), 362
Lindqvist, I., 51 (81, 82, 83, 84), 68, 172
(821, 173, 1%
Linke, R., 109 (188), 163
Linn, C. B., 123 (1631, 124 (162, 1631,
162
Linn, F. S.,138 (1041, 161
Linnctt, J. W., 112 (1421, 162
Lipkin, D., 82 (801, 101, 102
Lippincott, E. R., 165 (831, 192
Lipscomb, W. N., 272 (41, 873, 280 (831,
281 (831, 283 (83a1, 285 (83, 1391,
293, 304 (831, 311,312
Livasy, J., 287 (951, 311
Livingston, H. K., 112 (189), 163
Lobunez, W., 272, $77
Losel, M., 166 (921, 102
Loevenich, I'., 267, 277
Long, G., 199 (581, 215 (581, 218 (581,
23'6
Long, L. H., 84 (77, 81, 82, 83),101, 300
(83b1, 311
Lorenz, L., 255, $77
Lorenzi, G. P., 95 (971, 108
Lotgering, F. K., 32 (85), 68
Loudermilk, E. R., 107 (1901, 110 (190),
113 (1901, 153
Lounsbury, M., 326 (9), 300
Lowe, J. V., 300 (1261, 312
Lu, C.-S., 239, 277
Imhsinger, W., 114 (191), 163
LIU, T.-K., 109 (2951, 112 (2951, 144
(295), 166
Luijten, J. G. A., 76 (1231, 102
Lund, E. W., 165 (841, ID2
Lux, H., 21 (861, 68
Lynch, C. S., 124 (1921, 163
AUTHOR INDEX
377
Mandlcberg, C. J., 199, 215, 218 ( 5 8 ) , Merritt, W. F., 326 (91, 360
Mcrten, K., 73 (851, 76 ( 8 5 ) , 101
256
Merwin, H. E., 256 (218), 258 (2181, 278
Mannhcimcr, M., 112 (197), 153
Meslans, M., 115, 116 (207), 144 (207),
Mao, T. J., 139 (74), 144 (74), 150
164, 179, 192
March, N. H., 114 (198), 153
Messerly, J. F., 268 (146), 276
Maroy, K., 263 (107), 270 (107), 275
Messner, J., 178 (51), 179 (51), 180 (481,
Marriott, J., 110 (316), 156
182 (48), 191
Marsh, R. E., 239, 248, 272 (157, 1701,
Meta, L , 255, 2Y7
276, 277
Martin, D. R., 132 (20), 149, 280 (83c), Meuwsen, A,, 163 (93), 165, 166 (92), 167
(89), 171, 183, 184 (92a), 190 (91,
311
92a), 192
Martin, F., 255, 277
Meyer, E. G., 108 (2081, 154
Martin, H., 214, 235
Meyer, F. J., 83 (135), 103
Martin, W. J , 6 (64), 58
Meyer, J., 124 (209), 126 (209), 127
Mason, W. P., 52 (93), 68
(209), 145 (209), 154
Massey, H. S. W., 110 (38), 149
Massoth, F. E., 214, 235, 287 (154), 298 Meyer, L., 264 (701, 275
Meyers, R.L., 217 (66), 236
(154), 313
Michael, A,, 71 (1061, 82 (106), I02
Matejka, K., 264, 277
Michaelis, A,, 174 (94, 95), 192
Mathias, B. T., 52 (93), 58
Mathieson, A. McL., 22 (go), 85, 251, Mirhel, F., 184, 186 (30), 1.91
252 (431, 253 (431, 269 (43), 270 Michel, M. C., 324 (831, 362
Miles, G. L., 212 (251, 234
(43), 274, 9 7
Millen, D. J., 126 (209a), 154
Mrttsuo, K., 22 (771, 58
Miller, C. F., 110 (107), 151
Mattj~uch,J., 323, 326, 3&?
Matutano, J. R. B., 115 (1991, 144 (199), Miller, H. C., 107 (190, 211), 108 (212),
109 (212), 110 (190, 2111, 111 (212),
153
112 (210), 113 (190), 114 (211), 123
Maury, L. G., 124 (39), 14s
(2111, 125 (2111, 126 (211), 128
Maxwell, L. It., 42 (94), 58
(211), 145 (211), 153, 154, 180 (961,
Mayer, M. G., 225 (lo), 234, 332 (79),
192
362
Maycr, J. E., 110 (125), 111 (200), 151, Miller, J. G., 245, 270 (230), 272, 277,
153
278
Miller, J. J., 283 (68), 310
Mazzi, F., 22 (951, 43 (70), 68, 59
Miller, L. S , 85 (13), 100
Mead, E. J., 308 (14, 15), 309
Miller, S. A., 63 (86), 71 (861, 101
Mears, VC'. H., 214 (621, 255
Miller, W. S., 241, 277
Meeker, T. R., 283 (7), 309
Mills, H., 241, 277
Megaw, H. D., 50 (961, 59
Meguerian, G., 266 (181, 273
Milton, H. T., 227 (631, 235
Meiners, E. P., 336 (99), 362
Minoura, Y., 244, 245, 277
Meitner, L., 332 (Sl), 362
Mitchell, W. F., 123 (2131, 164
Melaika, E. A , 330 (911, 340 (82), 341 Moe, G., 109 (2951, 112 (295), 144 (295),
(911, 342 (91), 362
156
Mellor, D. P., 43 (129), 59
Modritzer, K., 307 1152), 313
Mendelejeff, D. I., 254, 277
Mollering, C. H., 124 (218), 154
Mennicken, G., 116 (314), 118 (314), 121, Morrh, 0. H., 256 (1081, 275
122, 144 (314), 156
Moffitt, W., 12 (971, 59, 116 (214), 119
Menzel, W., 112 (254), 155
(214), 164
Mericola, F. C , 116 (211, 117 (21), 144 Mohlcr, F. I,., 110 (68), 150
(21), 14.5 (211, 149
Moissan, H., 107 (215, 216), 115, 117
378
AUTHOR INDEX
(217), 118 (217), 119 (2171, 144

(2171, 164, 180 (97, 98, 99), 192, 193
Moonay, R. C. L., 273, 278
Moorc, C. A,, 36 (981, 69
Moore, W. ,J., 41 (99), 42 (991, 69
Morrison, T. J., 112 (2191, 164
Morton, A. A., 95 (87, 88), 102
Molton, .I. W., 69 (49n), 101
Moss, J. H., 81 (291, 100
Mrose, M. TS., 50 (411, 67
Miillcr, E., 124 (184), 126 (184), 127
(184), 132, 145 (1841, 149, 155
Miiller, G. E., 8.5 (3), 99
Miillcr, H., 77 (133), 105
Miiller, K. H., 240 (155), 270
Mucttmties, E. L., 118 (2211, 119 (221),
121 (221, 222), 125 (221), 126 (221),
127 (221), 136 (220), 137 (2201, 145
(221), 154
Muller, N., 140 (2231, 164
Mullins, J,. J., 221 (16), 2%$
Miilvancy, J. E., 139 (147), 162
Murin, A. N., 340 (4, 66), 300, 362
Myers, D. K., 148 (224), 164
N
Nagcl, K., 103
Nast, R., 81 (891, 102
Naused, K., 240 (711, 242 (72), 244 (72),
255 (72), 264 (70), ,276
Nelson, W. K., 80 (631,101
Neubaucr, D., 171 (117), 193
Nciidorfier, J., 107 (225, 227, 228), 108
(226, 228), 116 (2291, 119 (229), 164
Nenviir, B. W., 207 (72), 2S5
Ncwcome, M. M., 112 (2301, 164
Newton, T. D., 332 (84), 358, S61, 362
Nieluen, M. I,., 167, 168 (109a1, 193
Nier, A. O., 317, 319, 362
Niggli, P., 38 (1001, 69
Nitta, I., 273 (2151, 8 8
Noth, H., 308 (1501, d l 5
Noltes, J. G., 76 (123), 102
Nordman, C. E., 290 (84a), 311
Norman, R. 0. C., 89 (901, 102
Normant, H., 81 (91), 88 (911, 10.8
Nostrand, E. D., 109 (231), 112 (231),
164
Nyholm, R,. S., 12 (1021, 21 (62), 22 (42),
0
O(*hlobysstin, O., 77 (105), 102
OConnor, C. L., 112 (2321, 164
Oyurn, P., 251, 253 (109, 110, 1791, 272
(109, 1791, 876, 876, Z77
Ogawa, M., 181, 191
Ohlberg, S. M., 249 (180, 1811, 272 (1811,
877
Oksman, Ya. A,, 110 (149, 150), 169
Oliver, G . O., 227, 236
Oliver, J., 300 (138), $19
OT,oane, J. K., 116 (233), 164
Opik, U., 16 (1031, 69
Orchin, M., 93 (Sla), 102
ORear, J. G., 134 (94), 135 (94), 136
(941, 141 (94), 151
Orgcl, 1,. E., 12 (51, 106), 13 (511, 15
(35, 104), 17 (361, 19 (351, 20 (3%,
21 (1051, 22 (35), 25 (361, 26 (36),
27 (361, 28 (36, 371, 31 (371, 34
(1071, 35 (1071, 37 (1081, 41 (1101,
44 (1101, 45 (1091, 48 (log), 67, 112
(60), 160
Ossenberk, A., 126 (1451, 152
Otero, C.,115 (1991, 144 (199), 153
Otto, R., 250, 377
Oughton, B. M., 247, 277
Owen, J., 21 (161, 67
P
Pwker, J. E., 91 (30), 100
Page, G., 110 (381, 148
Palmer, I<. J., 243, 277
Palmer, W. G., 112 (234), 164, 255, ,977
Palmork, K. H., 258 (ill), 262 (102,
1111, 270 (1021, 876
Paneth, F. A,, 321 (491, 361
Panser, R. E., 221, 235
Pnppas, A. C., 334 ( l ) , 333 (871, 355, 356,
360, 3G2
Paquin, A. M., 187 (loo), 193
Parker, M. J., 340 (82), 362
Parker, P. L., 351,368
Parker, R. P., 148 (2351, 164
Parry, R . W., 283 (85a), 286 (11, 12, 45,
85, 123, 124), 288 (12, 891, 290 (87,
379
AUTHOR INDEX
88, 89, 117, 122, 123, 124, 1251, 202

(78, 79a), 2'33 (12, 79a, 86), 294 (11,
45, 79, 851, 295 (8.9, 301 (8.51, 30$),
310, 311, 312
Parsons, T. D., 87 (92, 93), 102
Passino, H. J., 124 (2831, 155
Pntt,erson, A. M., 280 (go), 311
Pattison, F. L. M., 148 (236), 154
Paul, D. E., 82 (80), lU1, 1W
Pauling, L., 9 ( l l 4 ) , 10 (114), 41 (99),
42 (99), 59, 112 (33), 149, 239 (170),
255, 268, 272 (1701, 277
Pauson, P. L., 63 (711, 71 (711, 101
Peacock, It. D., 16 (39a), 57, 131 ( 4 ~ 1 ,
148
Pearson, It. K., 308 (141, 142), 312

Petmson, T. G., 109 (237), 112 (237),
154
Pease, R. N., 75 (91, 100
Pcdley, J. B., 85 (95), 102
Penney, W. G., 268, 272, 27Y
Pennington, R . E., 246 (2051, 268 (205),
278
Pepinsky, R., 43 (70, 124), 50 (44), 124,
57, 59
Perkins, W. D., 119 (2381, 144 (238), 154
Petcrs, C. R., 290 (84a), 311
Peterson, J., 247 (2101, 269 (210), 278
Petri, H., 148 (261), 155
Petrov, H. G., 207 (3), 233
Petruska, J. A,, 330 (91), 338 (go), 340
(82, 90, 124), 341 (91), 342 (91), 362,
363
Petry, R. C., 91 (96), 102

Petrzhak, K. A., 347, 362
Pfeiffer, H., 90 (104), 93 (104), 95 (104),
102
Pfeiffcr, I., 90 (104), 93 (104), 95 (104),
102
Picon, M., 116 (239), 118 (239), 154

Pino, P., 95 (97), 102
Piotrowski, E. A,, 129 (336), 145 (336),
157
Piotrowski, H., 182 (31), 191
Piper, T. S., 172 (101, 102), 193, 307
(1531, 313
Pitochelli, A. R., 297 (691, 310
Platt, J. R., 291 (93), 311
Pollard, F. H., 265, 277
Pollock, H. C., 110 (240), 154
Poppcr, E., 255, 278

Porai-Koshit.s, M. A., 27 (115), 29 (115),
59
Porter, R. W., 107 (2411, 110 (240, 154

Posnjak, E., 19 (1331, 20 (133), 59
Powell, H. M., 29 (116), 69
Preobrazhensky, B. K., 340 (66), 362
Preston, M. A., 357, 362
Price, F. P., 281 (91, 921, 311
Price, S. J. W., 84 (98, 991, 102
Prideaux, E. B. R., 108 (242), 109 (2421,
112 (243), 144 (2421, 164
Priest, G. L., 107 (2721, 110 (272), 113
(272), 155
Prigent, J., 224, 236
Prince, E., 32 (117, 1181, 59
Prins, J. A., 239, 277
Prober, M., 107 (2441, 154
Proeschal, E., 21 (861, 58
Pryce, M. H. L., 16 (103), 59
Puschrnann, J., 125 (137), 152
Q
Qurashi, M. M., 51 (119), 59
R
Raabe, B., 96 (41), 100
Raasch, M. S., 135, 136, 155
Rabideau, S. W., 74 (59), 101
Rabinowitch, E., 223, 226, 235
Rae, H. K., 199 (581, 215 (581, 218 (58),
235
Raff, P., 308 (1551, 313
Rall, W., 324 (931, 362
Ramaswamy, K. L., 110 (317), 156
Ramsden, H. E., 88 (100, 102), 102
Randolph, C. L., 283 (36, 371, 286 (36),
288 (361, 289 (36), 309
Rao, G. G., 110 (317), 166
Rector, C. W., 291 (93), 311
Reeves, M. L., 110 (1051, 138 (105), 151
Reformatsky, I. A,, 340 (4), 360
Rehner, T., 197 (43, 441, 198 (441, 234,
235
ReicheIt, H., 184 (92al, I90 (92a), 192
Rcichle, W. T., 81 (81,100
Reid, T. S., 141 (300), 156
Rheinboldt, H., 277
380
AUTHOR INDEX
Rcirtmu~,li,O., 63 (72), 97 (72), 101

Rrinov, N. M., 110 (151, 152), 168
Reistad, T., 271 (1121, 176
Rekalic, V., 164 (16), I91
Rcnoll, M. W., 124 (2461, 127 (2461, 145
(246), 148 (2461, 166
Reubke, E., 123 (305), 166, 186 (1131,
ID3
Reynolds, J. H., 329, 340 (44), 361
Rheinboldt, H., 248, g 7
Hicc, R., 287 (951, 298 (961, 299 (941, 511
Riding, F., 112 (571, 160
Riehl, L., 116 (273), 127 (273), 166
Ristic', S., 242 (73, 741, 264 (70), f16
Rittenberg, D., 321, 362
Rittcnhouse, J. R., 272, xT7
Ritter, D. M., 87 (92, 93), 102, 288 (112,
113), 289 (113), 300 (138), 512
Robb, W. L., 217, 236
Roberts, C. W., 141 (195), 165
Roberts, H. L., 114 (70), 144 (701, 160
Robertson, J. M.,247, 251, 252 (43), 253
(43), 269 (43), 270 (43), 274, 277
Robinson, P. L., 109 (237), 112 (2371, 113
(34), 114 (6, 34), 131 (4a), 132 (6),
144 (34), 148, 149, 164, 241, 277
Roohow, E. G., G3 (1011, 64 (101), 76
(38, 128), 78 (110, I l l ) , 82 (841, 86
(1011, fU0,lU1, IU2,103
Rodgers, J. W., I07 (2991, 128 (299), 166
Roebuck, A. K., 124 (247, 248), 166
Rogers, F. E., 139 (147), 158
Rogers, M. J., 83 (17), 100, 112 (249),
166, 245, 248 (192, 193), 250, 269,
272, fly
R.omeijn, F. C., 32 (1201, 69
Rommel, O., 185 (103), 193
Rooksby, H. I)., 19 (121, 132), 51 (1301,
69
Roos, G., 185 (441, 186 (441, 191
Roscoe, J. S.,306 (103), 311
Rosen, 14'. D., 218 (78), 235
Rosenherg, S. D., 88 (102), 92 (50), 101,
102
Rossini, F. D., 229, 231, 236
Roth, W., 281 (97), 311
Rourke, F. M.,317 (120), 319 (119, 120),
363
Rowlinson, J . S., 112 (2501, 166
Ruhinson, W., 338 (601, 340 (GO), 361
c
Rwkcrt, A,, 308 (155), 313

Iluff, O., 107 (253), 112 (254), 114 (2531,
116 (2551, 123 (2561, 124 (2521, 125
(252), 132, 144 (253), 166, 169 (1041,
171 (104), 174 (104), 193, 197 (69),
214 (681,256
Kugg, F. M., 81 (8), 100
Ruggeberg, W. H. C., 124 (257), fG5
Ruigh, W. L., 292 (981,911
Rundle, R. E., 63 (103), 83 (I, 103, 1141,
85 (1031, 99, 102
Runge, F., 90 (104), 93 (1041, 95 (1041,
102
Runnals, 0. J. C., 209 (701, 236

Russell, H., Jr., 128 (290), 166
Rutkowski, A. J., 76 (62), 101
Ityschkewitmh, G. E., 292 (991, 511
Ryss, I. G.,124 (258), 166
s
Sacharkin, L. I., 77 (105), 108
Sackman, J. F., 84 (83), 1Of
Sadek, E. H., 190 (8), 190
Sampson, M. B., 336 (99), 562
Samuel, R., 255, D7
Sanders, J. K., 256 (151), d7G
Sanderson, R. T., 305, 312
Srtnjana, N. It., 31 (125), 32 (1251, 69
Sam, R. L., 165 (105), 195
Snunders, J. P., 113 (Zag), 156
Sauter, F.,112 (91), 161
Scatturin, V., 27 (122), 69, 255 (41), 27'4
Scavnicnr, S.,40 (123), 69
Schaeffer, G. W., 280 (loo), 285 (61, 62),
287 (951, 290 (101), 291 (75, 93,
104), 307 (1021, 310, 311
Schaeffer, 1'. F., 84 (120), 102
Schneffer, R., 291 (104), 292 (105), 306
(103), Slf
Schnr, W. C., 283 (81, SO9
Schats, P. N., 112 (260), 166
Schenk, J., 238, 239 (190), 8 7
Sehenk, P. W., 176, 193
Scherer, O., 148 (261, 262), 166
Scherer, P., 146 (173), 163
Schiff, H., 179 (107), 193
Schlenk, W., 71 (106), 82 (lost, 102
Schlesingcr, H. I., 92 (126), 102, 279, 281
(1111, 283 (38), 284 (16), 285 (39),
AUTHOlZ INDEX
286 (39), 288 (17, 110, 112, 113), 289

(1131, 290 (llO), 291 (109, 292
(1051, 294 (130), 298 (16, 38), 300
(111, 116), 301 (39), 305, 306 (18, 108,
109), 309, $10, 311, 312
Sclilossnagel, F., 163 (93), 192
Schmeisser, M., 125, 155
Schmidt, H., 118 (264), 165
Schmidt, H. D., 118 (264), 155
Schmidt, M., 265, 266, 2YY
Schmidt, W., 113 (265), 119 (265), 121
(265), 155
Schneider, J., 103
Schneidcr, W. G., 108 (4, 204, 268), 109
(4, 204), 111 (4, 204, 266, 267, 268),
112 (193, 194), 148, 153, 15.4, 155
Schoberl, A,, 243 (1981, 277
Schoening, F. R. I,., 22 (1361, 59
Scholten, W. Z., 16 (181, 39 (18),57
Schomaker, V., 225, 235, 247, 274
Schotte, L., 246 (199, 200, 201, 202, 2031,
247 (202, 203), 266, 268, 269, 270
(2021, 271 (261, 272 (113), 974, 275,
BYY, 2Y8
Schotten, J., 254, 265 (751, 266 (751, 275
Schramm, G., 124 (2091, 126 (2091, 127
(209), 145 (2091, 154
Schroder, E., 108 (921, 161
Schroder, H., 130 (109, 110), 131 (109,
110, I l l ) , 146 (109, 110, I l l ) , 161
Schultz, D. R., 290 (87, 88, 117, 136), 311,
312
Schula, W. W., 207 (72), 235
Schulse, F.,68 (50a), 101
Schulzc-Itettmer, R., 268, 275
Schumann, H., 181 (108), 193
Schumb, W. C., 107 (269, 270, 2721, 108
(271), 109 (271), 110 (2721, 113
(269, 270, 2721, 114 (270), 115 (270,
272), 140 (270, 271), 155
Schwarz, R., 163 (17), 166, 169 (17), 170
(17), 176 (IS), 177 (18), 178 (18), 191
Scott, D. W., 239 (139), 246 (139), 204,
205), 268 (139, 146, 204, 205), Q6,
2Y8
Scott, R. B., 124 (39), 1.49
Seaborg, G. T., 195 (51, 74), 221 (51),
223, 226, 229, 231, 235
Senbright, C. A., 128 (221, 129 (22), 145
(22), 149
381
Seel, F., 116 (2731, 127 (2731, 166

Seeley, G. R., 300 (1381, 312
Segre, E., 347, 348 (95), 362
Sellers, P. A,, 205, 235, 327 (96), 368
Senent, I?., 112 (274), 155
Senkpiel, W., 189 (71, 190
Severson, W. A,, 126 (275), 142 (2751, 166
Seyferth, D., 78 (107, 109, 110, I l l ) , 82
(84, 108, 112), 87 (108, 112), 101, 102,
302 ( I l S ) , 312
Shapiro, I., 281 (119, 120, 121, 144), 283
(56a), 310, 312, 313
Sharp, D. W. A,, 124 (276, 2771, 125
(276), 126 (276, 277), 155
Sliarpe, A. G., 212 (25), 234
Shaw, B. L., 79 (201, 95 (201, 100
Sheft, I., 204, 214, 235
Shehekotikin, A. I., 135 (172), 163
Sheppard, N., 124 (277), 126 (277), I56
Sheridan, J., 128 (322), 15F
Shiflett, C. H., 218, 235
Shirane, G., 43 (70), 43 (124), 50 (1241,
58, 69
Shoaf, C. J., 308 (151, 309
Shoolery, J. N., 117 (79), 118 (79), 120
(79), 150
Shore, S. G., 286 (123, 1241, 288 (891, 290
(88, 89, 122, 123, 124, 125), 3f1, 312
Shoskes, M., 113 (259), 155
Shriffer, R. G., 324 (22), 360
Siebert, H., 124 (2781, 126 (2781,165
Sicgel, B., 300 (54, 126), 310, 3f.9
Silverman, M. B., 87 (931, 102
Silvey, G. A,, 107 (280), 113 (279, 2801,
138 (279, 280), 139 (167, 279), 147
(279), 152, 156
Simons, J. H., 106 (281, 2821, 110 (3231,
124 (283), 142, 155, 166
Simmos, T. C., 139 (1471, 152
Singer, J., 265 (223), 2Y8
Singleton, B., 139 (1471, 152
Sinha, A. P. B., 31 (125), 32 (12.51, 50
Sisler, H. H., 95 (51, 99, 292 (99), 311
Skinner, H. A., 84 (22, 361, 85 (951, 100,
102
Skoog, I., 85 (791, 101

Slota, P. J., 296 (411, 310
Slough, JV., 83 (113), 102
382
AUTHOR INDEX
S ~ ~i i l cy,
6. H., 206 (791, 208, 211, $34,
236
Smith, R. W., 325 (1211, 344 (121), S6S

Smith, D., 225 (711, 235
Smith, G. P., 22 (381, 67
Smith, H. S., 292 (105), 311
Smith, J. J., 81 (8), 100
Smith, M. C., 221, 234
Smith, M. E., 6 (641, 68
Smyth, C. P., 243, 245, 270 (1601, 277,
278
Smythe, W. R.,323, 302
Snell, A. H., 335 (98), 336 (991, 362
Snow, A. I., 83 (114), 102
Soldate, A. M., 305 (127), 313
Sollcr, W., 38 (1271, 59
Solornon, I. J., 300 (128), 304 (64), 310,
312
Rerum, H., 51 (US), 50, 251, 252 (207),

278
Sowards, D. M., 197 (84), 236
Spa&, G. S., 255, 278
Sparks, R.S., 139 (1471, 162
Speier, J. L., Jr., 76 (115), 102
Sperduto, A., 110 (351, 149
Spielman, J. R., 301 (41n), 820
Spiess, W., 176 (181, 177 (181, 178 (181,
181 (llO), 101, 103
Spohn, K., 264, 278
Spong, A. H., 109 (284), 112 (2841, 156,
242, 278
Stacey, M., 140 (361, 141 (36), 142 (361,
147 (36), Id!)
Stackelberg, M. V., 112 (285), 155
Studler, 13. P., 188 (661, 192
Stafford, S.L., 302 (1291, 303 (1291, 312
Stafiei, S. P., 292 (601, 310
Stuhl, R. P., 214 (62, 81), 296, 230
Staritxky, E., 223, 224 (221, 234
Steffens, C. C., 112 (3321, 157
Steidl, D. V., 217 (821, 286
Steidlitr, M. E., 218 (781, $36
Steinberg, E. P., 340 (40, 1001, 344 (39),
361, 362
Steindler, M. J., 217, 236, 292 (1051, 294
(1301, 311, 312
Steinkopf, W., 116, 143 (287), 148 (2871,
165, 168
Steinrauf, 1,. K., 247, 269 (210), 278

Stephenson, N. C., 43 (129), 69
Stephenson, R., 123 (2891, 16G

Steunenberg, R. K., 217 (1, 821, $33, 286
Stevens, C. M., 327 (961, 362
Stevens, K. W. H., 21 (141, 57
Stevenson, D. P., 128 (290), 156, 241
(2111, 243, 247 (2111, 278
Stevenson, P. C., 225 (lo), 234
Steward, E. G., 51 (1301, 69
Stewart, A. C., 291 (751, 306 (1031, 311
Stieglite, J., 167, 174 (78), 192
Stock, A., 279, 281, 285, 288 (131), 292,
812
Stone, B. D., 167, 168 (109a), 169, 193
Stone, F. G. A., 82 (6, 84, 1121, 87 (112),
94 (1161, 99, 101, 102, 281 (1321, 283
(49), 284 (59, 132, 133, 134), 285
(133), 286 (134), 287 (49, 581, 293
(421, 295 (134), 297 (1331, 298 (49,
135), 299 (49, 581, 301 (58, 133), 302
(4, 1291, 303 (1291, 306 (1321, 309,
810, 8lb
Stosick, A. J., 42 (521, 55, 288 (70), 510
Straasmann, F., 326, 362
Strebel, P., 92 (129), 103
Strehlow, R. A., 112 (2911,166
Strenk, C., 115 (46, 471, 144 (471, 149,
160
Strenks, K., 115, 156
Strohmeier, W., 94 (1171, 102
Studier, M. H., 327 (96), S6,9
Stuer, B. C., 184, 192
Sturdivant, J. H., 51 (1311, 59
Subba Rao, B. C., 75 (14, 151, 100, 302
(21, 22), 307 (19, 201, 309
Sudzuki, M., 110 (293), 156
Sugarman, N., 336, 361
Summers, I,., 79 (118,119), 81 (1181, 102
Suthcrland, G . B. B. M., 268, 272, 277
Suttle, J. F., 221 (451, 236
Sutton, L. E., 39 (501, 67, 112 (601, 150
Svatos, G. F., 140 (223), 154
Svendsen, S. R., 165 (841, 1.02
Sweep, E., 124 (2941, 166
Swiatecki, W. J., 332 (102), 362
Swinehart, B. A., 206 (52), 235
Szczepkowski, T. W., 266, 278
T
Tamsma, A. F., 266 (111, 273
383
AUTHOR INDEX
Tarinenbautn, 1. R., 215 (31), $34

Tannmbaum, S., 84 (120), 102
Tatlow, J. C., 140 (36), 141 (361, 142
(36), 147 (361, 149
Taylor, P. G., 267 (213), ,278
Taylor, R. C., 286 ( I ) , 290 (136), 50.9,
512
Tebboth, J. A,, 63 (86), 71 (861, 101

Teller, E., 15 (681, 68
Templeton, D. H., 25 (311, 43 (71), 67,
324 (83), 56.2
Terrell, J., 359, 562
Thal, A., 71 (106), 82 (106), 102
Thiel, K., 116 (255), 132, 156
Thode, H. G., 318 (1061, 319 (1061, 320
(62), 321 (62), 325 (74, 121), 329
(ill), 330 (35), 331 (63, 76), 334
(105), 335 (104), 336 (13, 112), 338
(14, 90, 113), 340 (13, 14, 31, 32, 64,
90, 113), 341 (35), 344 (31,32, 36, 104,
113, 121), 345 (14, 31, 64, 113), 346,
347 (75, 113), 348 (30, 126), 349 (30),
351 (126), 353, 354 (62), 356, 357,
358 (621, seo, 361, 362, 365
Thomaes, G., 112 (296), 166
Thomas, B., 254, 265 (75), 266 (75), 276
Thomas, C. L., 124 (2971, 166
Thomas, R. D., 112 (122), 161
Thompson, A. J., 38 (1271, 59
Thomson, J. J., 316, 563
Thorpc, T. E., 107 (299), 123 (2981, 128
145 ~ w )166
,
Tierney, P. A., 304 (24), 306 (23), 307
(24), SOD
Tiers, G. V. D., 141 (300), 166
Tilton, G. R., 326 (2), 360
Tjomsland, O., 259 (114, 115, 116, 117,
118, 119, D O ) , 260 (115, 119), 261
(114, 1201, 267 (119), 270 (115, 119),
271 (121, 122), 272 (114, 120), $76,
276
Tobin, M. C., 165 (83), 192
Todd, S. S., 268 (1461, R G
Tombs, N. C., 19 (1321, 69
Tomiie, Y., 273, 978
Tomlinson, R. H., 326, 327 (1231, 330
(911, 331 (12), 338 (901, 340 (31, 82,
90, 124), 341 (29, 91), 342 (91), 343
(27, 29), 344 (311, 345 (27, 29, 31),
360, 361, S G 303
~
m~),
Toms, D. J., 305 (40), 234

Torrans, D. J., 124 (257), 155
Toussaint, J., 247, 278
Townend, R. V., 214 (62), 256
Townsend, J., 82 (go), 101
Trabanelli, G., 265, ,974
Traube, W., 118 (303), 123 (302, 304,
3051, 125 (301, 302, 3031, 126 (303,
166, 183, 185, 186 (1131, 193
Trauts, M., 107 (306), 114 (306), 115, 119
(306), 144 (306), 166
Trefonas, L. M., 293, 512
Tremaine, J . F., 63 (86), 71 (861, 101
Treuting, R. G., 32 (1171, 69
Troegcr, J., 250 (2171, 677, 278
Trost, W. R., 113 (3071, 166
Trotmnn, J., 85 (541,101
Trotman-Dickenson, A. F., 84 (98, 99),
102
Trott, P. W., 140 (as),141 (291, 142 (291,

147 (29), 148 (29), 149
Trout, H. Q., 119 (201), 153
Truce, W. E., 135 (3081, 147 (3081, 148
(308,309), 166
Trueblood, K. N., 47 (87a), 48 (87a), 49
87a), 68
Trump, J. G., 107 (272), 110 (272), 113
(2721, 166
Truswell, A. E., 210 (19), 218 (19), 234
Tsutsui, M., 88 (138), 105
Tuchaite, 0. Y., 265 (235), 278
Tunell, G., 19 (133), 20 (133), 69, 240 (8),
256 (218), 258 (218), ,973, 2775
Turco, A,, 27 (122), 69
Turissini, A. D., 214 (62), 25.5
Turkevich, J., 225 (lo), 254
Turner, H. S., 292 (137), 512
Tuttle, T . R., Jr., 82 (127), 103
Tvedten, O., 247 (1041, 276
Tyczkowski, E. A,, 133, 138 (310), 139,
144 (310), 146 (310), 147 (310), 166
U
Ubbelohde, A. R., 82 (671, 83 (113), 101,
102, 112 (205), 164
Uchida, H. S., 298 (96), 311
Ulmschneider, D., 93 (121), 102
Uloth, R. H., 79 (118, 119), 81 (Ilk?), 102
Urquhart, R. C., 110 (35), 149
384
AUTHOR INDEX
Wagner, A., 243, 277

Wagner, F.,190 (63), 192
Wagner, J., 255, 274
Vainstein, B. I., 22 (134), 60
Wagner, R. I., 283 (43), 286 (43), 287
Valanchunas, I. N., 264, 265 (2351, 278
(441, 294 (431, 295 (431, 296 (140,
van Alten, L,, 300 (138), S l 2
140a), 298 (44), 300 (44), 305 (140),
Van de Gmff, R. J., 110 (351, 140
310, 312
van der Kerk, G. J. M., 76 (122, 1231,
Wahl, A. C., 353, 363
102
van der Maas Retidy, J., 285 (1391, 512 Walburn, J. J., 88 (1001, 102
Walker, A. O., 300 (1161, 312
vander Meulen, 1'. A,, 249, ,077
Walker, W. H., 328 (110), 329 (1111, 331
van Ham, G., 265, 274
(115), S63
Van Horn, J. R., 327, 362
van Niekerk, J. N., 22 (136), 23 (1351, Walkup, J. H., 132 (231, 148
Wallbridge, M. G. H., 300 (83b), 511
69
Walling, C., 90 (124, 125), 91 (125), 102
van Vlcck, J. T., 15 (137), 69
Vaughan, I?. A., 38 (33), 67, 249 (181), Walling, E., 326, 362
Wallstein, R., 197 (69), 236
272 (181), 277
Walsh, E. N., 129 (336), 145 (3361,167
Vegard, I,., 49 (1381, 69
Walter, A. J., 326 (16), 360
Venanzi, L. M., 27 (139), 69
Verdelli, L. S.,108 (212), 109 (212), 111 Walter, H. A., 187 (114), 193
Wanless, R. K., 325 (112), 336 (112), 338
(212), 164
(113), 340 (113), 344 (1131,345 (113),
Verhoek, F. H., 91 (961, 102
347 (113), 565
Verneuil, A., 249, 278
Wannagat, V., 116 (314), 118 (314), 121,
Vcrwcel, H. J., 39 (1401, 60
122, 144 (314), 156'
Vester, K., 81 (891, 10%
Warf, J. C., 197, l S 6
Videla, G. J., 291, 310
Vihovde, E. R., 252 (1231, 272 (123), 27G Warhurst, E., 112 (57), 150
Warne, R. J., 292 (137), 312
Virtue, R. W., 110 (312), 166
Warren, J. W., 110 (315, 3161, 166
Vladimirsky, J., 332 (1081, 983
Vogel, R. C., 212 (471, 213 (30), 233, g54, Wart.ik, T., 92 (126), 102, 308 (141, 142),
318
236
Vogclbruch, K., 244 (65), 247, 267 (651, Wartman, L. H., 81 (81, 100
Waters, W. A . , 89 (go), 102
g4
Watson, H. E., 110 (3171, 166
Vohler, O., 81 (66), 101
Watson, W. E., 123 (289), 1GF
Voigt, D., 189 (70), 102
Watt, G. W., 197,230
Voigt, G., 182 (52), 190 (81, f90, 101
Waugh, J. S., 114 (561, 132 (56), 150
Vol'skii, A. N., 220, 234
Weaver, E. E., 215 (571, 216 (571, 226
von Groase, A,, 63 (76), 101
(871, 227, 228, 230 (87), 236, 236
Vousden, P., 51 (1421, 69
Vovnesenskii, S. A,, 128 (3131, 145 (313), Weaver, R. H., 110 (3121, 166
Weber, E. F., 240 (156), 276
166
Weber, H., 244 (72), 248, 249, 255 (72),
264 (701, 276
W
Webster, J., 76 (115), 10.2
Wehner, P., 198 (461, $36
Wachters, L. H. J., 239 (190), Z77
Waddington, G., 239 (1391, 246 (139, Weilmuenster, E. A., 280 (1431,512
Weinberg, A. M., 332 (114), 342 (114),
205), 268 (139, 146, 205), %76, 278
363
Wadsley, A. D., 51 (143), GO
Weinhardt, A,, 125 (321), 156
Wadso, I., 84 (22), 100
Wagman, D. D., 229 (671, 231 (67), 836 Weinstock, B., 215 (57), 216, 217 (85),
AUTHOR INDEX
224 (56), 225 (17, 56, 711, 226, 227,

228 (571, 230, 232 (42), 234, 236, 236
Weis, G., 175 (531, 191
Weiss, Alaninso, 38 (144), 60
Weiss, Arniin, 38 (1441, 60
Weiss, H. G., 281 (119, 120, 121, 144),
312, 313
Weiss, J., 164 (115a), 166 (541, 167 (54),
171 (54, 117), 172 (25, 35, 41, 82, 115,
116), 173, 101, 192, 193
Weiss, R., 79 (61), 101
Weissman, S. I., 82 (7, 80, 127), 90, 101,
102, I05
Weitz, E., 254, 264, 265, 278
Wellcr, S.,199 (21, 233
Wells, A. F., 2 (147), 20 (145, 1461, 22
(148), 25 (1471, 29 (1161, 69, GO
Welsh, H. K., 22 (go), 58
Wender, I., 93 (91a), 102
Wentorf, R. H., Jr., 108 (3181, 109 (318),
111 (318), 156
West, B. O., 81 (581, 101
West, R. C., 76 (38), 100
Westcott, C. H., 331, 332, 36'3
Westgren, A., 42 (11, 66
Westlake, H. E., 245 (225), $78
Westrum, E. F., Jr., 209 (88), 236
Wetherill, G. W., 323 (117), 326 (2),
348 (117), 360, 363
Wheeler, J. A., 332 (50), 361
Whitcombe, R. A., 94 (261,100
White, D. G., 76 (128), 103
White, F. A,, 317, 319, 363
White, G. N., 266, 278
White, W. E., 130 (3191, 145 (319), 156
Whit>ford, W. R., 88 (loo), 102
Whitney, J. F., 6 (64), 68
Whytlaw-Gray, R., 113 (66, 67), 144 (67),
160
Wiberg, E., 92 (1291, 103, 281 (146, 147),

283 (148, 149), 286 (148, 149), 288
(145, 148, 149, 151), 292 (151), 305,
307 (1521, 308 (150), 313
Wiberg, K. B., 92 (129), 93 (130), 103
Wiherly, S. E., 281 (2), 309
Wiechert, K., 116 (3201, 118 (320), 119
(3201, 123 (3201, 166
Wigner, E. P., 332 (114), 342 (114), 303
Wild?, J. H., 39 (891, 58
385
Wiles, D. M., 326, 327 (1231, 340 (124),

363
Wiles, D. R., 325 (1211, 344, 363

Wilhelmi, K. A., 45 (27), 49 (281, 67
Wilke, G., 77 (1331, 98 (131, 132), 103
Wilke-Dofurt, E., 125 (321), 166
Wilkinson, R., 336 (999),362
Williams, A. A., 84 (21), 100, 301 (471,
310
Williams, Q., 128 (322), 166
Wilson, M. K., 116 (233), 119 (238), 144
(238), 154, 269 (227), 278, 307 (153),
913
Wilson, R. R., 332 (1251, 363
Wilson, W. A., 110 (3231, 156
Winger, R. A., 112 (206), 154
Winkhaus, G., 242 (781, 244 (69), 247,
274, 275
Wirth, H. E., 287 (1541, 298 (154), 913

Wittig, G., 78 (1341, 80 (134), 81 (1341,
83 (135), 96 (1341, 103, 308 (155),
313
Wittwer, M., 248, 277
Wdbling, H., 165, 193
Wood, A,, 323 (19), 360
Wood, F. C., 185 (119), 195
Wood, H. W., 254, 265, 278
Wood, J. E., 111, 124 (192), 163
Wood, J. F., 118 (86), 119 (861, 150
Wood, J. M., 116 (12), 119 (12), 144 (12),
149
Woodrow, C. C., 245, 270, 278

Woods, L. A,, 89 (48), 100
Woodward, I,. A., 114 (701, 144 (701,
150
Woolf, A. A,, 123 (326), 124 (326, 3271,

125, 126, 166, 212 (261, 234
Wotiz, J. H., 84 (31), 93 (32, 33, 136),
94 (32, 331, 100, 103
Wright, G. F., 100
Wunderlich, F., 118 (303), 125 (303), 126
(3031, 166
Wyckoff, R. W. G., 5 (150), 8 (150), 9
(150), 10 (150), 38 (149), 43 (150), 44
(150), 49 (150), GO, 241, 278
Wyoski, M. M., 118 (328), 119 (3281,
166
Y
Yakel, H. L., Jr., 247, 269 (232), 278
386
AUTIIOR INDEX
Xelionkaite, V. I., 254, $78

Zemek, F., 197 (48), 198 (48), 236
Zhigsch, A. F., 298 (156), 313
Zicgler, K., 71 (140, 1411, 73 (1391, 74
(1401, 78 (140, 142), 77 (142), 78
(142), 79 (1391, 82 (145), 90 (1421,
92 (142), 93 (1421, 95 (1421, 97 (1401,
98 (1401, 99 (143), 101, 103
Zimm, B. H., 110 (334), 167
Zi~nmerman,R. E., 76 (115), 102
Zimmermann, W., 90 (1041, 03 (104), 95
(1041, 102
2
Zinner, C., 243, 266, 374
X:Lcliuriasen, W, H., 42 (1511, 60, 205 Xiomck, J. S., 129 (3361, 145 (3361, 167
(36, 75), 224 (6), 233, 234, 236, 256, Zirker, G., 163 (561, 166 (54, 55, 561, 167
(54), 171 (54), 198
257 (991, 259 (124), 260 (1241, 269
Zoche, G., 88 (4), 90
(991, 270 (124), 273, 276, 276, 5"s
Zandhcrg, E. Ya., 110 (153, 154, 3351, Zolotov, Y. M., 185 (771, 185 (75, 76),
102
162, 167
Zeiss, H. H., 81 (641, 88 (64, 1381, 101, Zorin, N., 115 (751, 160
Zweifel, C., 77 (16), 100
103
Yakavlev, C. N., 226, 229, 2%

Yanitski;, I. V., 254 (236, 237, 2381, 264,
265 (2351, 878
Yost, D. M., 107 (330), 108 (329, 3311,
112 (329, 332), 116 (3291, 144 (3291,
166, 167
Young, 8 . G., 348 (126), 351 (1261, 363
Young, C. B. F., 124 (333), 167
Young, J. A,, 139 (74), 144 (74), 160
Young, R. C., 197 (go), 198 (go), 286
SUBJECT INDEX
A
Actinide element halides, 195-233
bromides, 221-224
change of state, 228-229
chlorides, 219-221
crystal structure, 222-223
electronic configuration, 232-233
fluorides, 205-211
iodides, 221-224
properties, 224-233
structure, 224-225
thermochemistry, 229
volatility, 226-228
Actinium ,
halides, 196, 222
isotopes, 201
oxidation state, 197
Alkali halides, radius ratios of, 5, 6-7
Aluminates, spinel strut-ture, 31
Aluminum alkyls, 71, 72, 89, 91, 93
as alkylating agents, 77-78
dimers of, 69
reactions of, 97
technological applications, 98
Aluminum hydride, 73-75, 92
Aluminum, triethyl, 65, 86
Aluminum, trimethyl, 64, 83
Americum,
halides of, 196, 200, 222, 223-224, 233
isotopes, 201
oxidation state of, 197
Amidosulfonic acid, 188-190
Amidosulf uric acid, 181- 182
Amino diboranes, 289
Anisole, metalation of, 69
Anisotropic electronegativity, 16
Antimony, organometallic compounds of,
79, 81
Aryl silanes, 76
Aufbau reariion, 74
B
Berkelium,
isotopes, 201
387

Beryllium,
alkyls, 69, 74
isotope, half-life, 327
Bismuth alkyls, 88
Bis(pentafluorosu1fur) peroxide, 120-121,
144
Bis(trifluoromethy1)
polysulfides,
135-
136, 146
Bis(trifluoromethy1) sulfide, 134, 146

Bis(trifluoromethy1) sulfur tetrafluoride,
139, 147
Bistrifluoromethyl trithiocarbonate, 137,

146
Borane(s),
coordination compounds of, 286-287
fuels, 280
reactivity, 86, 87
Boron,
alkyls, 89, 90, 94, 96
-antimony compounds, 294-297
-arsenic compounds, 294-297
-carbon bonds, 299-304
-halogen compounds, 304-305
hydrides, 75-76
chemical reactivity of, 279-308
properties, 3 0 M 0 8
volatile, physical constants of, 282
-nitrogen compounds, 285-294
organometallic compounds, 84
-oxygen compounds, 297-298
-phosphorus compounds, 294-297
-selenium compounds, 298-299
-sulfur compounds, 298-299
Boron fluoride, 132
Boron trifluoride, 132
Bromides, of actinide elements, 221-224
C
Cadmium,
ion, stereochemistry, 37, 39
organometallic rompounds, 84, 91
Californium,
isotopes, 201
Oxidation stat,e, 197
388
SURJECT INDEX
Carlmi-iiictd bond, rcitclivil.y, 05

Ccsiuni iuotopc, Idr-lilc detcrininriliori,
325-326
Cesium chloride, 5-6
Chemical bond concept, 62-63
Chlorides, of actinide elements, 219-221
Chlorine isotope, half-life det,erminalion,
326
Chromites, as spinel structures, 30
Chromium-organornctallic compounds,
79, 81, 88
Chromium oxide, 52
Cinnnbnr, 39
Cobnlt,
ammonium ion, 5, 14
organomctallic compounds of, 81
Cobalt chloride, stereochemistry, 19
Cobaltites, ns spinel structures, 31
Coordination,
octahedral, 12-15, 28-29
polyhcdral, 2-5
tetrahedral, 25-29
Copper,
compounds, coordination, 20, 22
ion, stereochemistry of, 32, 37, 3839
Copper chloride, stereochemistry, 19
Copper oxide, stereochemistry, 20-21
Crystal fields, influence on transitionmetal st,ereocheniistry, 12-33
Crystal radii, of various ions, 7
Cupric fluoride, stereochemistry, 20
Curium,
halides, 196, 200, 21Ck211, 222-224, 233
isotopes, 201
Cyanogen polysulfides, 243, 248-250
D
Daughter-growth method, of isotope
half-life determination, 325-326
Decay method, of isotope half-life dctrrmination, 325
Diaaoaminobenzenc-copper complex, 39
Diazomethnnc, methylntion using, 78
Diethyl fluorothiophospliate, 130
Difluoromethylrnr bis(su1fur pentafluoride), 139, 147
Diphenyl tin hnlitlr, 66
Disulfidrs, 247
Disulfonyl polydlitles, 250-253

Disulfnr clecnfluoridc, 113, 144
Disulfur difluoridc, 114-115, 144
E
Electrons, inert, 40
F
Ferrites, as spinel structures, 31
Ferrocene, 71-72
Fission,
neutron-induced, 347-352
spontaneoue, 347-352
yields, 352-360
Fluorides, of actinide elements, 205-219
Fluorine fluorosulfonate, 120, 144
Fluorine-sulfur compounds, 105-148
with antimony, 131
with nitrogen, 130-131
properties of volatile, 144-147
Fluorite, %lo, 29
Fluorodisulfuric acid, salts of, 127
Fluorooxy acids of sulfur, 123-127, 145
Fluorosulfinic acid, 127
Fluorosulfonste ion, 124-125
Fluorosulfonic acid, 123-124, 145
rsters, 126127, 145
G
Gadolinium, mass spectrogram, 328
Gallium alkyls, 77, 84, 89-91, 93, 95
Germane hydride, 76
Germanes, reactivity of, 87
Germanium organorrietallic compounds,
82
Glycol ethers, in orgunorrietallic compound syml,liesis, 88
Gold,
compounds, stmeochemistry, 25
ion, stereocliemistry, 34, 38
Grignard reagent, 84, 86, 93-94
H
Hafnium oxides, st,ereochemistry of, 47,
52
Halitit,s, of :tclinide clemenls, 195-233
389
SUBJECT INDEX
Hept,asulfur imide, 161-164

Hexasulfurdiimide, 164-165
Hexathionates, 262-263
Hydrides, addition, stereospecificity of,
7e-77
Hydrogen diaulfide, 243
I
Iniidosulfainitle, 188
Imidosulfinamide, 180-181
Imidosulfonic acid monoamide, 190
Imidosulfuric acid, 181-182
Indium alkyls, 77, 89, 90, 93, 95
Iodides, of actinide elements, 221-224
Iodine organometallic compounds, 79-81
Ionic roordination polyhedra, 2-5
Ionic radii, 54-56
Ionic solids, stercochemistry, 1-56
Ionic theory, 2
Isotope(s),
abundance measurements of, 318-320
dilution technique , 321-323
half-life determination of, 324-327
J
Jahn-Teller effect, 15-23, 33, 36
K
Krypton,
determin>ttion in xenon, 323
isotopes, 325, 336, 345-346
L
Lanthanum, isotope dilution deterniinntion, 322
Lead,
ion, stereochemistry, 40-41, 4445
organometallic compounds, 82, 88, 89
Lithium,
alkyls, 71, 89
association of, 83
reactivity of, 88, 97
thermal dissociation, 95-96
ion, ionic radius of, 55-56
organometallic compounds of, 79, 82
phenyl, 72
Lithium hydride, 73-75, 92
M
Magnesium alkyls, 84, 87, 88
Magnesium hydride, 74
Manganese,
compounds, stereochemistry, 21
Manganese oxide, stereochemistry, 52
Manganites, as spinel structures, 31
Mason and Mathias theory of ferroelectricity, 52
MN-SS,
alignment of radioactive species, 323324
uliain yields, of isotopes, 340-343
separator, time-of-flight, 324
yield curve, fine structure in, 343-347
Mass spectrometer(s),
isotope dilution changes in, 321-323
sensitivity, 317
tandem analyzer, 319
Mass spectrometry, in nuclear chemistry,
315-360
Mercuration, aromatic, 71
Mercury,
ion, stereochemistry, 34, 37, 3 9 4 0
organometallic compounds, 82-84
Metal,
-carbon bond, 6 1 6 3
carhonyls, 72
hydrides, 73-77
olefins, 73-77
oxides, stereochemistry, 48-52
plating, organometallic compounds use
in, 99
Metalation, of organometallic compounds, 68-69
Metalloids, organometallic compounds
of, 8W1
Methylenation using diazomethane, 78
Molecular-orbital theory, 15
Molybdenum,
ion, stereochemistry, 51, 52
oxide, stereochemistry, 49
Monoperoxytetrafluorodisulfur oxide,
121-122, 144
N
Neptunium,
halides, 196, 199-200, 215, 222, 223, 224
isotopes, 201
390
SUBJECT INDEX
oxidation statc, 197

Neutron,
capture cross sections, 327-332
capture processes, 337
emitters, delayed, 336-337
Nickel,
coordination compounds, 5, 24, 27
ion, stereochemistry of, 32
Niobite, stereochemistry, 51
Niobium,
oxide, stereochemistiy, 52
Nuclides, shielded, 353-354
0
Orgnnoboron alkyls, 75
Organogallium alkyls, 75
Organolithium compounds, 69
Organomagnesium halides, 71
Organometallic compounds, 61-99
addition to multiple bonds, 68, 97-98
applications in technology, 98-99
autocomplexation, 83-84
bonding, 63-64
carbon-mctal bonding in, 84-85, 88-89
characteristics, 63-70
chemical propertics, 65-69
cleavage of, 92-94
elwtrolytic prepitrtllion, 82
halogen-metal interconversion, 69
history, 70-72
hydride formation, 91-92
Lewis acid character, 94
metal-metal bonding in, 85
metalation of, 68-69
organic moiety influencc on, 86-88
oxidation of, 8%91
physical properties, 69-70
preparation, 70
reactions of, 66, 68, 85-98
structure, 83-85
thermal dissoriation, 95-96
uses, 98-99
vinyl-type, 81-82
Organometalloid compounds, 62
P
Palladium compounds, 5, 25
Pentafluorosulfur hypofluorite, 117-118,

144
Perfluoroalkanesulfonic acids, 140-142
Perfluoroalkanesulfonyl chlorides, 148
Perovskite structure, 43, 49
Peroxydisulfuryl difluoride, 122, 144
Peroxysulfur compounds, 272-273
Phosphorus difluoroisothiocyanate, 129,
145
Phosphorus thiobroniochlorofluoride, 145
Phosphorus thiobromofluorides, 129, 145
Phosphorus thiochlorodifluoride, 145
Phosphorus thiochlorofluorides, 129, 145
Phosphorus thiodibromofluoride, 145
Phosphorus thiofluoride, 128-129, 145
Platin um ,
compounds, strreochemistry of, 25
organometallio compounds, 79, 85
Plutonium,
halides, 196, 200, 210, 215-219, 220, 222224
isot,opes, 201
oxidation states, 197
Polysulfide ions, 240, 241
Polysulfides, organic, 243-248
Polythionates, 254-266, 269-270
Potmaium isotopes, 323
Protactinium,
halides, 196, 222, 223
isotopes, 201
Pyrosulfur chlorofluoridc, 121, 145
Pyrosulfur fluoride, 121
Pyrosulfuryl chloride, 144
R
Rocksalt, crystal structure of, 6, 29
Rubidium isotopcs, 323, 326
Rutile, 9, 11
5
Selenium chain compounds, 248, 249, 251252, 254, 271-272
Selcnotrithionates, 256
Silicon alkyls, pyrolysis of, 96
Silver ion, stereochemistry, 34, 3s
Sodium alkyls, 95
Sodium borohydride, 308
SUBJECT INDEX
Sodium chloride, c;rystal structure, 5

Specific activity method, of isotope lialflife determination, 326-327
Sphalerite, 7-9, 39
Spinel arrangement, 29-33
Stannanes, 78, 87
Stereochemistry of ionic solids, 1-50
Strontium isotope half-life, 327
Sulfanuric acid, 184-185
Sulfimide, 182-184
Sulfones, 148
Sulfonyl fluorides, 142-143
Sulfur,
amides, 243
chain compounds, 237-273
containing organo groups, 248-250
containing sulfonate groups, 254-265
containing sulfonyl groups, 250-253
rotational isomerism, 268-270
selenium analogs, 271-272
stereochemistry, 267-273
structure studies, 266-267
tellurium analogs, 271-272
unsymmetrical 265-266
elemental, 238-240
fluorine-containing compounds of, 105148
halides, 241-243
hydrides, 241-243
imides, 161-176
oxides, compared with nitrogen analogs, 160
oxyacids, 159-190
Sulfur chlorofluoride, 119
Sulfur difluoride, 114-115, 144
Sulfur diimide, 174-176
Sulfur hexafluoride, 107-112, 138-139, 144
Sulfur monofluoride, 114, 115, 144
Sulfur oxyfluorides, 115-123, 144-145
Sulfur tetrafluoride, 113-114, 138, 144
Sulfur trioxide, 182-184
Sulfuryl amide, 185-188
Sulfuryl bromofluoride, 119-120, 145
Sulfuryl chlorofluoride, 145
Sulfuryl fluoride, 118-1 19, 144
T
Tantalum,
39 I
oxide, stereochemistry of, 52

Tellurium,
chain compounds, 248, 252-253, 254,
271-272
ion, stereochemistry of, 41, 43, 44-45,
54
organornetallic compounds, 79, 81
Tenorite, stereochemistry, 19
Telraethylgermane, 70
Tetrahydrofuran, in organometallie compounds, 88
Tetrasulfur tetraimide, 165168
Tetrathionates, 257-258
Tliiocarbonyl fluoride, 133, 146
Thionyl amide, 179-180
Thionyl chlorofluoride, 116-117, 145
Thionyl fluoride, 114-116, 144
Thionyl imide, 176-179
Thionyl tetrafluoride, 117, 144
Thorium,
halides, 196-198, 220-221, 223
isotopes, 201, 207
oxidation states, 197
Tin,
alkyls, biocidal properties of, 98
ion, stereochemistry, 40
Titanium,
ion, stereochemistry, 50
organometallic compounds of, 79, 80
Toluene, metalation, 69
Transition-metal stereochemistry, 12-33,
52
Trifluoroniethane sulfenyl chloride, 134,
146
Trifluoromethanesulfonic acid, 140-141,
147
Trifluoromethanethiol, 133-134, 146
Trifluoromethyl fluorodithioformate, 137,
146
Trifluoromethyl sulfones, 147
Trifluoromethylsulfur trifluoride, 146
Trifluorometliyl thiometal compounds,
137
Trifluoromethyl trifluoromethanethiolsulfonate, 141-142, 147
Triphenyltin halides, 66
Trisulfides, 247
Trisulfuryl fluoride, 122123, 145
392
SUBJECT INDEX
rrurlgstell,
ion, st.ci,ooclieinistry, 51
oxide, stereochemktry, 49, 52
oxides, stereochemistry of, 50, 52

Verdriingung process, 74, 97
W
U
Wurzite, 7-9
Uranium,
fission products from, 334-335
halides, 196, 198-199, 205-206, 208-209,
211, 215, 220-220, 222-224
isotopes, 201
half-life ticterminrttion of, 327
thermal neutron fimion of, 333
xenon yield from, 337-338
minerals, 347-351
oxidation strrtcs, 197
spontmcous fission of, 347-348
V
Vanadium,
ion, stereochemistry of, 45, 50
organomrtallic compounds, 81
X
Xenon,
isotope half-life determination of, 325,
330, 338339
mass spectrogram of, 320
Z
Zinc,
ion, stereochemistry of, 37, 39
organomet,allic compounds, 71, 84, 90
Zircon, stereorhemistry, 49
Zirconium,
ion, stcreoc1iemist)rryof, 44, 47

Advances in Inorganic Chemistry and Radiochemistry 2 (1960)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advances in Inorganic Chemistry and Radiochemistry 2 (1960)

Uploaded by

Copyright:

Available Formats

A d v a n c e s in

This Page Intentionally Left Blank

ACADEMIC PRESS INC.

CopyrightO, 1960, by Academic Press Inc.

NO PAllT O F THIS BOOK MAY RE REPRODUCED IN ANY FORM,

ACADEMIC PRESS INC.

United Kingdom Edition

Librury of Congress Cutulog Card Number: 59-7682

PBIXTED IN THE UNITED STATES OF AMERIC.4

Address: University of Michigan, Ann Arbor, Michigan

This Page Intentionally Left Blank

Stereochemistry of Ionic Solids

IV . The Influence of Crystal Fields on Transition-Metal Stereochemistry

JOHN EISCHAND HENRYGILMAN

IV . Reactivity and Itraction Pathways of Organometallic Conipounds . .

Fluorine-Containing Compounds of Sulfur

Amides and lmides of the Oxyacids of Sulfur

IV . Aniidcs of Oxyncids of Sulfur (IV) . .

Halides of the Actinide Elements

Structures of Compounds Containing Chains of Sulfur Atoms

Chemical Reactivity of the Boron Hydrides and

Mass Spectrometry in Nuclear Chemistry

IV . Neutron Capture Cross Sections .

STEREOCHEMISTRY OF IONIC SOLIDS

Organic Chemistry Laboratory, Swiss Federal Institute of Technology, Zurich, Switzerland

IV. The Influence of Crystal Fields on Transition-Metal Stercoclieniistry ,

11. Ionic Coordination Polyhedra

Square Coplanar d* Complexes . . . . . . . .

V. Stercochemistry of B-subgroup Ions. . .

,I. D. 1)IJNITX . 4 N D 1,. E. ORGBL

model still forms il satisfactory background to t i great part of the theory

II. Ionic Coordination Polyhedra (76, 147)

Wc first examine the iniplicatioiis of a very much oversiniplified ionic

is the arrangement which maximizes the X-X distance and so minimizes

four or six coordinated structures which require espl:i,iiation; for cs:tniplr,

The compounds which according to almost any theory should be the

Rocksalt structure (Fig. 4).

FIG.4. The rocksalt structure.

structure in which every ion is surrounded by a square antiprism of ions

STEREOCHEMISTRY OF IONIC SOLIDS

Bo++ 0.31 [0.44]

Empirical ionic radii

J. I). DUNITZ A N D L. E. ORGEL

STEREOCHEMISTRY OF IONIC SOLIDS

The interatomic distanccs in these compounds can be closely reproduced by a special

J. D. DUNITZ AND L. E. ORGEL

Fluorides and oxides of composition MX2 adopt the 6-coordinated

STEREOCHEMISTRY OF IONIC SOLIDS

sionul network but in two-dimensional sheets, separated from each other

are calculated from the univalent radii.

DUNITZ AND L. E. ORGEL

In this section we shall develop the electrostatic crystal-field theory to

STEREOCHEMISTRY OF IONIC SOLIDS

the two groups of orbitals is usually designated as A, and the energies of

referred t o the weighted mean energy in the electrostatic field as zero.

J. I). DUNITZ AND L. E. ORGEL

With two, or three d electrons present, both of these tendencies can be

STEREOCHEMISTRY O F IONIC SOLIDS

consideration. I t can easily be shown however, that molecular-orbital theory

J. D. DUNITZ AND L. E. OltGEL

amount, so thtit there is no net tendency to distort. We see, therefore, that

STEREOCHEMISTRY OF IONIC SOLIDS

There is, perhaps, oiily one example where an appreciable distortioii