Bowens Reaction Series

Railsback's Some Fundamentals of Mineralogy and Geochemistry
Bowen's Reaction Series II: A silicon-centered explanation

potential than Si4+, to less substitution, and
thus from less cation-cation repulsion to
more (see the green labels below). In the
middle branch, the spinels have no Si4+ at all
and thus obviously minimize Si4+-Si4+
repulsion. Zircon, inserted in the series
below, has some Si4+, but only half as much
as quartz (see the brown labels below).
To summarize, as we pass down each
branch of Bowen's Reaction Series, we go
from less Si4+-Si4+ repulsion to more. That
means we go from more readily-formed
minerals to less readily-formed ones. As a
result, we go from those that can form at
higher temperatures to those that will only
form at lower temperatures.
That's a good explanation
of why the minerals of Bowen's
Reaction Series are in the
An Si4+-centered explanation of the order of Bowen's Reaction Series
order they are. However, it is
a silicon-centered explanation
Minerals
Less
Spinels
Olivine
Anorthite
- it pays little attention to the
forming
at
Si4+ - Si4+
(Chromite,
(calcic
highest
other cations in the minerals
Magnetite,
repulsion temperaplagioclase)
involved. With a little more
Pyroxenes
etc.)*
tures
(e.g. augite)
insight, we can reach a more
Plagioclase
complete explanation. That's
Zircon
Feldspars
what we'll find in Parts III and
Amphiboles
(e.g.,
IV of this series of pages on
Albite
hornblende)
Bowen's Reaction Series.
(sodic
formation of a Si-rich mineral becomes a
disfavored process, relative to formation of
minerals with cations of lesser ionic
potential.
With that in mind, we can see why each
branch of Bowen's Reaction Series goes in
the order that it does. In the discontinuous
series from olivine to biotite, we progress
from minerals with no linkage of silica
tetrahedra to much linkage of silica
tetrahedra (and thus much "potential" for
repulsion) (see the blue labels below). In
the continuous series, where all tetrahedra
share all four oxygens and so linkage is
maximized, we progress from anorthite to
albite. We thus progress from more
substitiution of Al3+, a cation of lower ionic
Le
s
tet s Si 4+
rah
i
sit edra n
es
l
Some
Si4+
of
ge
n k a dra
s l i ah e
L e s a te tr
c
si li
No Si4+
One might ask, why are the minerals in

Bowen's Reaction Series arranged as they
are - why can olivine, anorthite, and the
spinels form first, and quartz form last? If
anything, we might expect a melt rich in
silica to produce silica as its first mineral,
rather than its last. Why instead is the silica
mineral usually last?
The answer usually given (and it's a good
answer) is that the order of crystallization
from a melt is dictated in part by the
repulsions between cations. Formation of a
mineral requires cations, bundles of positive
charge, to come into close proximity in a
mineral structure. Si4+ is a cation of high
ionic potential (i.e., high density of charge,
defined as charge divided by radius). Thus
plagioclase)
Biotite
All
cations
are Si4+
n
Potassium Feldspar
Mo
4+ i l
i
r
s il e l
S ra
Muscovite
ic a i n k
re hed
o
tet a g e
M tra e s
rah o Quartz
te sit
ed f
ra
Minerals
More
forming at
4+
Si - Si4+ lowest
repulsion temperatures
*Bowen's (1922) original formuation of the "reaction principal" included this middle
branch for the spinels. Zircon is added here for the sake of completeness.
LBR BowensRxnSeriesSketch 07 rev 1/2007
Bowen's Reaction Series III: Melting temperatures of oxides

This page is a step on the road to
understanding Bowen's Reaction Series, the
subject of a series of pages in this book. Pages
IV and V will finish the job.
The diagram below is a bit of the periodic
table. Superposed on it are contours showing
the melting temperatures (Tm) of the oxides of
the cations in this part of the table. Tm is
greatest for the swath of cations from Mg2+ to
Th4+.
On this part of the periodic table, ionic
potential (charge radius) increases from lower

left (z/r = 0.67 for Rb+) to upper right (z/r = 45
for N5+). The highest values of Tm for the
oxides are thus those of the oxides of cations of
intermediate ionic potential, rather than cations
of low or high ionic potential. The reasons for
that relationship are shown below: low ionic
potential only allows weak electrostatic bonding
to O2-, and high ionic potential causes
repulsions between cations, but intermediate
ionic potential hits the happy medium of strong
bonds without cation-cation repulsions.

If we make a profile across the periodic
table, we can construct the diagram at lower
right, where melting temperature of the oxides is
plotted against ionic potential. With this
diagram, we can proceed to Bowen's Reaction
Series IV, where we will plot the minerals of
Bowen's Reaction Series on this diagram.
Conceptual model of the behavior of

oxides of hard (and intermediate) cations
Li
Li
2681
1700
Na
Be
2+
1193
Mg2+
3125
3+
723
4+
216
Al 3+
2345
Si
4+
1996
Li + Be2+ B 3+
5+
ross
of c elow
h
t
Pa ion b
t
sec
500
5+
Na
855
Ca
3200
Sc 3+
Ti 4+
Ca2+ Sc3+ Ti 4+
Rb+ Sr 2+ Y 3+
290
2103
V 5+
943
charge
radius
Rb
673
2938
Ba
2+
2286
3+
La
Zr
4+
3123
3+
Hf
4+
3173
2580
3493
1785
Ta
5+
2058
Mo
6+
1074
6+
1745
1500
Th 4+
Nb
5+
2000
2500
0
300
3000
2500
200
1500
Cs
Sr
2+
V 5+ Cr 6+
Zr 4+ Nb5+ Mo6+
Cs + Ba2+ La 3+ Hf 4+ Ta 5+ W6+
Cr 6+
10
00
S 6+
6+
Mleting temperature (C) of

oxides of hard cations
2+
32
Mg2+ Al 3+ Si 4+ P5+
N 5+
16
K+
C 4+
H+
Contours of ionic potential:
Melting T(K) of oxides of hard cations
Cations
High z/r
Strong
O2
Rb
bonds, but
Intermediate cation-cation
Low z/r
repulsion
z/r
Weak cation- Strong cationoxygen bonds oxygen bonds Thus less
stable
Thus less
Thus stable
solids
stable solids
solids
Melting temperatures of
3000
2500
2000
Dry
1500
1000
500
Wet
Fe3+ Ti4+
Fe2+
K+ Na+ Ca2+
Mg2+ Cr3+ Zr4+
Al3+
0
1
Low ionic potential
(weak cation-O2bonds)
3
4
5
6
7
Intermediate ionic potential
(Stable cation-O2- frameworks)
Si4+
8
9
10
High ionic potential
(cation-cation repulsion)
Ionic potential (charge radius) of cations
LBR BRSIII05 1/2007
Bowen's Reaction Series IV:

Toward a broader explanation
Cation comprising more than half

of cation atoms in mineral's formula
L Left side of Bowen's Reaction Series:
DiscontinuousSeries
Middle track of Bowen's Reaction Series:
Spinels etc.
Right side of Bowen's Reaction Series:
Continuous Series
M
Average ionic
potential of mineral
Cation comprising half of the cation

atoms in mineral's formula
Melting
temperatures
of oxides
of hard
cations
3000
Forsterite (Mg-olivine)
Chromite
Mleting temperature (C) of oxides of hard cations
Magnetite
L
Anorthite (Ca-plagioclase)
2500
M
Ilmenite
Titanite (sphene)
Augite
2000
Zircon
Plagioclase: Ab50An50
1500
Hornblende
Dry
Biotite
R
Albite (Na-plagioclase)
1000
Muscovite
Wet
K-feldspar
500
Quartz
K+ Na+ Ca2+ Fe2+ Mg2+
0
Cr3+ Fe3+Zr4+ Ti4+ Al3+

4
Si4+
7
10
Intermediate ionic potential
(weak cation-O2- bonds)
(Stable cation-O2- frameworks)
(cation-cation repulsion)

LBR BowensRxnSeriesSFMG04 rev. 1/2007
Low ionic potential
Very generalized order of igneous crystallization
Part II of this series of pages on Bowen's Reaction

Series provided a Si4+-centered explanation of
Bowen's Reaction Series. This page tries to explain
Bowen's Reaction Series in terms of all the cations
involved. It draws heavily on Part III of this series,
which looked at the melting temperatures of oxides to
assess the importance of ionic potential in determining
the temperature of formation of minerals.
The diagram below shows the compositions of
minerals that form in the crystallization of silicate melts.
The different shades of blue correspond to the different
branches of Bowen's Reaction Series. The sizes of the
squares correspond to the relative abundance of the
cations in each mineral.
Graphically speaking, there are two things to note.
First, the large squares are in the middle of the
diagram at its upper (high-temperature) end, but they
drift to the right as one moves down through the
diagram to lower temperatures. Secondly, the leftmost
squares drift to the left as one passes downward
through the diagram from higher temperatures to lower
temperatures.
So what does that mean? The minerals high in
Bowen's Reaction Series - the minerals that form at
higher temperatures - are minerals that preferentially
incorporate cations of intermediate ionic potential like
Mg2+. On the other hand, the minerals lower in
Bowen's Reaction Series - those that form at lower
temperatures - are minerals that increasingly
incorporate either cations of lower ionic potential like
Na+ and K+ and/or the cation of exceptionally high
ionic potential, Si4+. The cations of low ionic potential
form weak bonds to O2-, explaining their entry into
minerals only at low temperatures, and cations of high
ionic potential have such focused positive charge that
they repel each other when brought in close proximity,
as they must be in a Si4+-rich mineral.
Minerals in igneous rocks:

Cation comprising less than half of
cation atoms in mineral's formula
Pyroxenes
(e.g. augite)
Amphiboles
(e.g.,
hornblende)
Spinels
(Chromite,
Magnetite,
etc.)*
Zircon
Plagioclase
Feldspars
More cations of low

ionic potential (e.g.,
Minerals
K+) forming weak
More
forming at
2bonds with O
Si4+ - Si4+ and/or more cations lowest
temperarepulsion
of high ionic
tures
potential (e.g., Si4+)
*Bowen's (1922) original formuation of the "reaction principal" included this middle
generating cationbranch for the spinels. Zircon is added here for the sake of completeness.
cation repulsions
H+
Cations
N
1
16
Mleting
temperature (C)
of oxides of hard
cations
All
cations
are Si4+
Li
S6+
Melting temperatures of
3000
2500
2000
Dry
1500
1000
500
Fe3+ 4+
Fe2+
Na+Ca2+ Mg2+ 3+ 4+Ti 3+
+
K
Al
Cr Zr
High z/r
charge
= ionic potential
radius
Si4+
Strong
O2
bonds, but
Intermediate
cation-cation
Low z/r
z/r
repulsion
Strong cationWeak cationThus less
oxygen bonds
oxygen bonds
stable solids
Thus stable
Thus less stable
solids
solids

Li
Rb
Wet
0
1
2
3
4
5
6
7
8 9 10
Low ionic potential Intermediate ionic potential
2(weak cation-O bonds) (Stable cation-O2- frameworks) (cation-cation repulsion)
K+ Ca2+ Sc3+ Ti 4+ V5+ Cr 6+
Cs+ Ba2+ La3+ Hf4+ Ta5+ W 6+
Minerals
forming at
highest
temperatures
Biotite
32
Rb+ Sr 2+ Y 3+ Zr 4+ Nb5+ Mo6+
More cations of
intermediate
ionic potential
(e.g., Mg2+)
forming stable
bonds with O2-
Albite
(sodic
plagioclase)
Conceptual model of the

behavior of oxides of hard
(and intermediate) cations
Na+ Mg2+ Al 3+ Si4+ P5+
Anorthite
(calcic
plagioclase)
Less
Si4+ - Si4+
repulsion
Mo
+ in l
Potassium Feldspar
4
i
s i l re
S ra
i c a lin
Muscovite
e hed
r
k
te t a g
s
o
ra e
M etra site
he of Quartz
t
dr
a
Contours of ionic potential:

Li + Be2+ B3+ C4+ N5+
Le s
tetrah s Si 4+ in
edral
sites
of
Less linkage
ra
silica tetrahed
This is the fifth of five pages

explaining the order of the minerals
in Bowen's Reaction Series. The
usual first minerals to form are those
rich in cations of intermediate ionic
potential, which bond strongly to O2but don't have such highly focused
charge that they set up cation-cation
repulsions. The usual later, or
lower-temperature, minerals either
have cations of low ionic potential,
which only bond weakly to O2-,
and/or they have the cation of high
ionic potential, Si4+, which generates
cation-cation repulsions that
destabilize their mineral structures at
higher temperatures.
Olivine
Some
Si4+
Bowen's Reaction Series V:

A summary explanation
No Si4+
Bowen's
Reaction
Series
K
LBR BRSV06 rev. 1/2007
Generalized Trends in Silicate Minerals in Igneous Rocks:

Gray arrows indicate paths of causality
Discontinuous
reaction series
Magnesium- & iron-rich
Olivine
Pyroxenes
ore d
ens
Continuous
reaction series
Si4+ closer toge the

and thus lo r cau
wer
s
me es m
ltin o
g re
te
m
Ca-Plagioclase
Increasing
linkage of
Increasing
Increasing
silicate
density
melting
tetrahedra, Increasing
of
in crystals viscosity minerals temperature
of magma
of minerals
or in
in the presence
magma
of H O
(e.g. augite)
Amphiboles
(e.g. hornblende)
Biotite (black mica)

K(Fe,Mg)3AlSi3O10(OH)2
K-Feldspar
KAlSi3O8
mica)
Sialic/Felsic
Cl u
d
ke
s
tetr ters of lin 't
n
a he
a
d ra c
fl ow a di ly.
re
CaAl2Si2O8
Na-Plagioclase
(albite)
NaAlSi3O8
SiO2
Silicon- and
aluminum-rich
(anorthite)
Muscovite (white
Quartz
on = m
i on
ls re.
pu tu
re era
p
Mafic
(Fe,Mg)2SiO4
KAl3Si3O10(OH)2
More ir
Plagioclase is a
kind of feldspar
Felspar-like
Macroscopic
results:
Crystal shapes:
elongate chain silicates,
planar layer silicates,
equant framework
silicates.
Patterns of volcanism:
Explosive eruptions and steepsided volcanoes of sialic
magmas, vs. flowing volcanism
and gently-sloping volcanoes
or flood basalts of mafic
magmas.
Global topography:
Buoyant sialic
continents with tops
above sea level, and
dense mafic ocean
floor below sea level.
Global geography:
Differentiation
of sialic continental
crust by partial
melting.
LBR 1121SilicateMineralsTable14 1/2002; rev 8/2007
Rate of cooling of magmas and the texture of igneous rocks

Slow cooling:
Fast cooling:
Maximum
possible
travel for
atom at left
Maximum
possible
travel
Multiple
possible
random
walks
Longer cooling time
allows longer random
walks of atoms to sites on crystals
Size of crystals:
Descriptive terms on the
basis of whether crystals
are visible to naked eye:
Shorter cooling time

only allows shorter random
walks of atoms to sites on crystals.*
coarse-grained
phaneritic
fine-grained
aphanitic
From the Greek word

for visible
Genetic terms on the basis

of whether the rock forms
inside the Earth or at its surface:
Genetic terms derived from
the names of Roman gods:
intrusive
extrusive
plutonic
volcanic
From Pluto, the god

of the underworld
From Vulcan, the god

of the forge
Another way to look at this is the much

greater surface area of smaller crystals
provides more surface area and thus
more sites for atoms to join
in a shorter time.*
(at least most of the crystals are

tiny and invisible; some earlyforming crystals may be large,
giving a porphyritic rock)
*The explanations above assume that many small
crystals will exist with fast cooling. Nucleation of
many small crystals happens as fast cooling induces a high saturation state that allows crystal
nuclei to grow despite their large ratio of surface
area to volume, which otherwise thermodynamically
disfavors their survival .
LBR
6/2012
SFMGIntrusiveExtrusiveOrigin01.odg
20
Wet
melting
of basalt
I S
1500
Pressure (kilobars)
Depth (km)
50
50
Along the dry-melting boundary, increased vibrational

energy of atoms with increasing temperature breaks
minerals apart to cause a transition to liquid (melting).
Because these liquids are less dense than the corresponding solid minerals, decreasing pressure also favors
melting, so that the boundary slopes as it does.
Compared to the dry melting boundary discussed above,

the wet-melting boundary falls at a lower temperature
because H 2 O molecules bond to the cations of minerals
and help liberate those cations, breaking down minerals to
generate melt.
Dry
melting
of basalt
1500
50
The curves at left are only schematic and would move with
changing rock composition and/or changing water content.
20
10
500
1000
Temperature (C)
Pressure (kilobars)
Pressure (kilobars)
Pressure (kilobars)
Heat moving outward from Earth's core or lower

mantle drives melting. This is the most intuitively
obvious potential cause of melting (add heat to
make something melt), but it is the least
volumetrically significant of the three origins
shown here.
For sources of the concepts shown here, and for
more about partial melting, see Part II of this pair.
Heat
Melt
Thermal-plume-driven partial melting
Depth (km)
Depth (km)
1500
The position of the

rock in P-T space
does not change;
instead, the relevant
boundary changes.
500
1000
Temperature (C)
Wet
melting
of basalt
Dry
melting
of basalt
0 Earth- 500
1000
surface
conditions Temperature (C)
Depth (km)
20
10
10
Origins of melts and magmas I
Dry
melting
of basalt
50
Melts (molten mineral material) and magmas

(melt plus newly-formed crystals) cool to form
igneous rocks. Places where igneous rocks
form are not evenly distributed on Earth;
instead, they are restricted to three general
locations. This page explains those three
general locations in term of how melts
originate. We begin with the diagram at right
and move to the three diagrams below.
20
10
0
1500
H2O-induced partial melting
500
1000
Temperature (C)
Pressure-release partial melting
Melt
Water released from the subducting oceanic plate at a

convergent plate boundary induces melting that produces island arcs and continental arcs. The former are
largely basaltic; the latter include both grano-dioritic
intrusive rocks and basaltic to rhyolitic volcanics.
H2O
lt
Solid
Partia
l melt
at convergent plate boundaries
Melt
Asthenosphere
As plates of oceanic
lithosphere move
apart at divergent plate
boundaries (at midocean ridges), underlying asthenospheric rock flows
upward. Its upward movement requires a decrease in
pressure that allows melting and thus produces melt.
These melts form the Mid-Ocean Ridge Basalts
(MORB) and underlying gabbros of the oceanic crust.
Note the role of moving oceanic plates in both of these origins of

melts, which are volumetrically the two dominant sources of melts.
Solid
Part
ial m
e
Solid
Partia
l melt
Solid
Partia
l melt
at divergent plate boundaries
Solid
Partia
l melt
I S
Dry
melting
of basalt
1000
1500
50
Depth (km)
1622
Tm(PH2O = 5 kb)
Along the dry-melting boundary,

increased vibrational energy of atoms
with increasing temperature breaks
minerals apart to cause a transition to
liquid (melting). Because these liquids
are less dense than the corresponding
solid minerals, decreasing pressure also
favors melting, so that the boundary
slopes as it does.
Compared to the dry melting

boundary discussed above, the wetmelting boundary falls at a lower
temperature because H 2 O molecules
bond to the cations of minerals and
help liberate those cations, breaking
down minerals to generate melt.
1722
1553
1391
1890
0.25
1.0
(Al+Si)
2.0
Mafic
Sialic
Pressure (kilobars)
LBR OriginofMagamas. . .
9/2010
The curves at left are only schematic

and would move with changing rock
composition and/or changing water
content.
1530
(Mg+Ca)
1118
Tm(1 atm)Dry
Melting temperatures (C)
Temperature (C)
500
Wet
melting
of basalt
Origins of melts and magmas II

20
10
0
0 Earthsurface
conditions
Minerals
Forsterite (Mg2SiO4)
1234
1295
Anorthite (CaAl2Si2O8)
1077
Diopside (CaMgSi2O6)
Solid
Partia
l melt
Part I of this series drew on concepts presented in webpages by Prof.

Jonathon Miller of San Jose State University (geosun.sjsu.edu/~jmiller/
Geo1_Lec5_IgRocks.html) and by Dr. John W. Merck of the Universiy of
Maryland (www.geol.umd.edu/~jmerck/geol100/lectures/10.html). The
table above is adapted from Table 4 of Railsback 2007 (American
Mineralogist, v. 92, p. 356-369). These pages also draw on the comments
of Drs. Robert Hawman and James Wright of the University of Georgia.
748
876
SiO2 minerals (Quartz etc.)

K-feldspar (KAlSi3O8)
Albite (NaAlSi3O8)
Water released from the subducting oceanic plate at a convergent plate boundary induces
melting that produces island arcs
and continental arcs. In the former
case, the result is largely basalt,
whereas the latter produces both
grano-dioritic intrusive rocks and
basaltic to rhyolitic volcanics.
The diagram and text at right were used in Part I

of this pair of pages to explain why most melts
originate where they do. The aspect of the
diagram least considered in Part I was the label
partial melt. This page tries to redress that
deficiency.
The table at right shows that the melting
temperatures of sialic silicate minerals are
generally less than those of mafic silicate
minerals, especially when water is present.
Thus, when a rock (an assemblage of minerals)
is heated, not all of its minerals melt at the same
time. Instead, those with lower melting temperatures melt preferentially to give a partial melt that
is typically more sialic than the original rock.
That thought about the lower melting
temperatures of sialic silicate minerals is
especially true in wet melting conditions where
H 2 O is present (see the blue column at right).
Thus, when dehydration of a subducting oceanic
plate releases H 2 O that induces melting in the
scenario shown both in Part I and below, the
result is a significantly more sialic magma than
the original melting rock. It was this preferential
production of sialic melt that led the University of
Georgia's Professor James A. Whitney, a famous
igneous petrologist, to give one of his UGA
colloquium lectures the succinct title Why are
there granites? Because there are oceans.
H2O-induced partial melting

at convergent plate boundaries
Melt
HO
2
Solid
Partia
l mel

Bowens Reaction Series

Uploaded by

Document Information

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

Bowens Reaction Series

Uploaded by

Copyright:

Available Formats

Railsback's Some Fundamentals of Mineralogy and Geochemistry