The Science of the Total Environment 223 Ž1998.

87]97

Oxidation treatment of a sulphide-bearing scrubber dust

from the Mansfeld Region, Germany: Organic and
inorganic phase changes and multi-element partition
coefficients between liquid and solid phases

Maurice Morency a,b , Holger Weiss c , Klaus Freyer d , James Bourne b,U ,
Denise Fontaine b , Raymond Mineau b , Monika Moder
¨ d , Peter Morgenstern d ,
d b
Peter Popp , Michel Preda , Hanns-Christian Treutler d , Rainer Wennrich d
a
Institute des Sciences de l’En¨ ironnement, Uni¨ ersite´ du Quebec
´ a` Montreal,
´ Quebec,
´ Canada H3C-P8
b
´
Departement des Sciences de la Terre, Uni¨ ersite´ du Quebec
´ a` Montreal,
´ Quebec,
´ Canada H3C-3P8
c
Industrial and Mining Landscapes Department, Umweltforschungszentrum, Leipzig-Halle, Germany
d
Department of Analytical Chemistry, Umweltforschungszentrum, Leipzig-Halle, Germany

Received 20 April 1998; accepted 12 July 1998

Abstract

A low temperature oxidation process has been developed to separate Pb from Zn in a complexly contaminated
scrubber dust. The material consists of approx. 50% galena ŽPbS. and sphaleriterwurtzite ŽZnS., and also an
amorphous component, a variety of different silicate and carbonate phases as well as naturally-occurring radionu-
clides and oil and grease. The two-stage process consists of a grinding stage followed by placing this material into a
hydrogen peroxide solution. The resultant oxidation reaction is violently exothermic } the temperature rises
spontaneously to approx. 968C. The amount of solid phases remaining after completion of the reaction was reduced
by approx. 40% and consisted largely of insoluble lead sulphate which retained the bulk of the radionuclides. Almost
all of the zinc was placed into solution by this process as was the Cd, Re and Cu. The total PAH content of the
original Theisenschlamm Ž468 ppm. was reduced to 11.25 ppm in the residual sediment and the PCDDrPCDF
concentrations were reduced by approx. 40%. The radionuclides are almost totally concentrated in the solid phase.
The method shows considerable promise as a separation technique for very fine-grained sulphide-bearing residues.
Q 1998 Elsevier Science B.V. All rights reserved.

Keywords: Oxidation; Scrubber dust; Sulphide-bearing; Organic and inorganic phase; Liquid and solid phase

U
Corresponding author.

0048-9697r98r$ - see front matter Q 1998 Elsevier Science B.V. All rights reserved.
PII S0048-9697Ž98.00271-X
88 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97
1. Introduction
The Permian Kupferschiefer black shale forma-
tion has been mined for copper in Sachsen-Anhalt,
Germany for more than 800 years. The mineral-
ized shale was processed with coal in a blast
furnace at a temperature of approx. 13008C. A
scrubber dust by-product, known as the Theisen-
schlamm, contains a variety of heavy metals, car-
bon dust, hazardous hydrocarbons and a substan-
tial amount of radioactive elements, ŽFlachowsky
et al., 1992; Freyer and Treutler, 1992; Lorenz et
¨
al., 1992; Glaser et al., 1993; Lorenz, 1994..
Roughly 220 000 tons of Theisenschlamm slurry
was generated which has been stored in a large
impoundment. This material still represents a ma-
jor contaminant risk in the area as it is not sealed
from the surface or ground water.
A description of the physical and chemical
characteristics of the Theisenschlamm is available
in Weiss et al. Ž1997.. The slurry recovered from
the scrubber contains substantial amounts of
heavy metals Ž18% zinc, 14% lead dry wt..,
organic carbon compounds including poly-
cyclic aromatic hydrocarbons ŽPAH., polychlori-
nated dibenzo-p-dioxin and dibenzofuran
ŽPCDDrPCDF. and radioactivity Ž5000]40 000
Bqrkg 210 Pb andror 210 Po. ŽLorenz et al., 1992;
Lorenz, 1994.. Noteworthy are the concentrations
of both hazardous ŽAs, Cd, Tl. and valuable ŽAg,
Re, Ge. elements. The Theisenschlamm consists
of particles with a median diameter of 1.25 m m
but scanning electron microscope studies show
that these particles are themselves comprised of
still smaller particles of sub-micron size. The pre-
dominant mineral phases present are ZnS
Žsphaleriterwurtzite ., galena ŽPbS. and secondary
anglesite ŽPbSO4 .. The sulphides appear to have
a distorted atomic structure and thus do not
strictly correspond to naturally occurring minerals
with normal ASTM diffraction patterns, most
likely due to ionic substitution of Zn, Pb, Cu and
Fe.
Several attempts have been made to develop a
hydrometallurgical or pyrometallurgical process
to extract the valuable contents of the material,
however, none of them were found to be ap-
propriate ŽLorenz et al., 1992; Lorenz, 1994; Wei-
landt, 1994.. The treatment of this heterogeneous
material requires a flexible technique that can
deal with the above mentioned characteristics and
at the same time minimize the impact on the
environment and convert the waste into valuable
products. Conventional separation techniques
cannot be successfully applied due to the ag-
glomerated state and small grain size of the mate-
rial.
A series of experiments were carried out with a
view to developing a low-temperature oxidation
process in which the sulphide phases would be
converted to sulphates, thereby making it possible
to separate the soluble zinc sulphate from the
more stable lead sulphate. This possibility only
exists if an oxidizing agent can act efficiently on
the variety of complex phases present in the
Theisenschlamm. The advantage of such an ap-
proach is that the zinc in solution could be recov-
ered by standard metallurgical methods whereas
the lead sulphate, including the radionuclides,
could be processed by lead smelting ŽLeipner,
1994; Weilandt, 1994.. Results show that most
elements strongly prefer either the liquid or the
solid phases. The partition coefficients presented
below should prove applicable to any fine-grained
sulphide-bearing material.
2. Experimental section
2.1. Analytical methods
X-ray powder diffractograms, used for the phase
identification of the whole powder sample, were
made using a Siemens diffractometer D 5000,
step scanner with a cobalt tube of 1.2 kW coupled
to a silicon detector. The interpretation was made
with the assistance of search match software.
The SEM ŽHitachi S 2300 with a Kevex energy
dispersive system. was used to determine the mi-
cro-chemical composition and element distribu-
tion as well as the texture and size of the differ-
ent chemical phases. Samples of dried slurry
powder were embedded in a polymer from which
a polished section was obtained and coated with
either gold or carbon.
Grain size measurements were performed on a
Galai CIS-100 which uses a laser beam of 1-m m
 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97
diameter. The lower resolution of the instrument
is approx. 0.2 m m. The machine is coupled to an
optical system which allows visual observation of
the morphology andror the degree of agglomera-
tion of the particles. The measurement was made
in an aqueous medium which maintained the
particles in suspension using ultrasound as well as
a mechanical stirring device.
DTArTGA measurements, using a Linseis
thermobalance with a furnace capable of heating
up to 16008C were made to determine the ther-
mal stability of this polyphase system. Approxi-
mately 250 mg of material were used for the
measurements.
The major element concentrations as well as
several trace element concentrations were de-
termined using X-ray fluorescence analyses ŽXRF.
and ICP-MSrICP-AES in order to obtain repre-
sentative values for the different elements.
For XRF, two sample preparation techniques
were employed: Ž1. a fusion technique with lithium
tetraborate during which the samples were heated
to a maximum of 9508C to assure complete disso-
lution and homogenization of the melt. The resul-
tant glass bead was analyzed using a Philips PW
1600 simultaneous X-ray fluorescence spectrome-
ter; and Ž2. samples were dried in air at a temper-
ature of 1058 C, ground in an agate mill, mixed
with a wax binder Ž20%. homogenized in agate
and pelletised at a pressure of 140 MPa. The
measurements were done using wave-length-dis-
persive X-ray fluorescence spectrometry ŽSiemens
SRS 3000..
For ICP-MS and ICP-AES, analyses the sam-
ples were prepared using an aqua regia digestion
according to DIN 38414r7. For ICP-MS analysis
ŽELAN 5000, Perkin Elmer., the samples were
diluted to 1:100 and 1:1000 in doubly distilled
water. After applying pneumatic nebulization and
multi-element standard solutions, the ICP-MS re-
sults were verified using the standard addition
technique. For ICP-AES analysis ŽSpectroflame,
Spectro A.I.., the digestion solution was diluted
to 1:10 and 1:100 with doubly distilled water.
Prior to determination of the organic com-
pounds, the samples were treated by Soxhlet ex-
traction Ž12 h with a toluenermethanol ratio of
89
10:1.. For PAH determination by HPLC, the sam-
ples were then subjected to a solvent exchange
Žcyclohexane., ultrasonic treatment Ž5 min., a sil-
ica gel clean-up and a final solvent exchange.
A GCrMS analysis was used to obtain additio-
nal information concerning the categories of or-
ganic substances present in the samples. Starting
with the sample preparation methods described
above, each sample was concentrated to a total
amount of 1 ml. The complex mixtures were sepa-
rated by gas chromatography using a Hewlett
Packard HP 5890 series II instrument. A directly
coupled mass spectrometer ŽMSD Hewlett
Packard. was used to analyze the components
separated by the gas chromatograph using elec-
tron impact ionization of 70 eV.
For the determination of PCDDrPCDF a Fin-
nigan MAT double focussing mass spectrometer
including a Hewlett Packard HP 5890 series II
gas chromatograph was used.
Two techniques of measurement are required
in order to determine the radioactivity concentra-
tions of 210 Pb and 210 Po } alpha spectrometry
for 210 Po and high resolution gamma spectrome-
try for 210 Pb.
For alpha spectrometry, the samples were first
put in solution with concentrated hydrofluoric
acid. The solutions were then spiked with an
aliquot of a tracer using 209 Po. The fractions were
evaporated to dryness and electrodeposited on
polished steel disks for subsequent counting.
The gamma spectrometry data were collected
using a high-purity germanium ŽHPGe. coaxial
low-energy detector, n-type, covered with a beryl-
lium window of 0.5 mm in thickness.
3. Oxidation experiments
Samples of the Theisenschlamm were ground
in an attritor and dried prior to treatment with
hydrogen peroxide. The grinding was done to
create fresh surfaces on the constituent particles
which would be more reactive. Also, the reduc-
tion of the grain size would help to make the
reaction more complete and more exothermic.
Our standardized procedure consisted of two sep-
arate steps } a mechanical pre-treatment phase
90 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97

Žgrinding. and the oxidizing reaction per se. The Table 1

whole process is referred to as the Theisen- Chemical evaluation of the grinding process
schlamm-Oxidation-Process ŽTSOP.. Element Ground W and - 0.45 Proportion
Theisenschlamm Žmgrkg. in solution
Žmgrkg. Ž%.
3.1. Grinding procedure and associated chemical
changes Al 23 600 0.7 0.09
B 111 2 17.21
Ca 8800 540 41.45
Cd 360 0.04 0.12
For the grinding procedure, 37 g of the Co 80 0.51 7.56
Theisenschlamm were placed in an attritor along Cr 890 0.51 0.69
with 300 ml of distilled water. Mechanical erosion Cu 14 100 0.12 0.01
Fe 30 850 0.84 0.04
of the steel balls used in the attritor causes Fe
K 6700 290 33.31
contamination of the ground Theisenschlamm. Mg 2500 98 31.14
The standardized grinding time used was 1 h, Mn 670 12 22.39
after which the sample was dried. This procedure Ni 510 2.2 6.42
was used to produce the standard starting mate- Pb 122 000 3.6 0.03
Zn 163 000 830 5.65
rial for subsequent experiments, which has been
called 33A. There was considerable weight loss
Stot 113 000 1500 13.60
during the grinding } the 37 g of wet Theisen-
schlamm initially present Žequivalent to roughly Oil and grease 8700 n.m. ]
30 g when dried due to a humidity content of
Notes. The ground Theisenschlamm column shows the chemi-
approx. 18%. was reduced to 26 g after attrition
cal composition of the Theisenschlamm after having been
and drying. Table 1 presents the chemical evalua- ground in water and then dried. The column W and - 0.45
tion of the grinding process together with the refers to the analysis of the water phase, which contains
concentration of a limited number of elements in species dissolved in the water as well as solid particles in
the original, dried Theisenschlamm. Also shown suspension which were small enough to pass through the
0.45-m m filter. The last column shows the relative distribution
is the composition of the Theisenschlamm after between the Theisenschlamm and the water phases.
grinding Ž33A } weight 26 g. as well as the Abbre¨ iation. n.m., Not measured.
composition of the 300-ml of initially distilled
water. The Theisenschlamm]water mixture was
filtered after grinding using a 0.45-m m mem- for Ca to almost 0% for Pb and Cu. It is clear
brane. Thus the analysis of this liquid phase ŽW that the composition of the Theisenschlamm be-
and - 0.45. includes not only ionic species but fore and after grinding differ considerably, how-
also suspended particles smaller than the mem- ever, only the concentrations of the most easily
brane. water-soluble elements are affected } the con-
The weight loss during attrition indicates that a centrations of the elements of economic interest
certain quantity of the elements in the wet Že.g. Pb, Zn. or of environmental interest Že.g. Cd,
Theisenschlamm were dissolved during the grind- Tl. are not significantly modified.
ing process. The amount in solution was roughly
estimated by multiplying the W and - 0.45
column by a factor of 300 Žas this column repre- 3.2. The peroxide reaction
sents the concentration in 300 g of water. and the
33A column by a factor of 26 Žthere being approx.
26 g of powder remaining after grinding.. The A mass of 5 g of sample 33A was then treated
proportion of each element that is in the water with 350 ml of 30% hydrogen peroxide solution.
phase is then presented in the final column of The slurry was continuously stirred magnetically
Table 1. This proportion varies from roughly 40% and the temperature continuously monitored dur-
 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97
ing the process. The resultant reaction was vio-
lent and strongly exothermic. At the peak of the
reaction a temperature of 968C was measured and
the slurry began to fume and froth } vigorous
stirring was required to prevent the slurry from
boiling over. After leaving the solution overnight,
the slurry was filtered and the solid residue ex-
amined by diffraction and SEM. The chemical
composition of the liquid was determined by ICP.
The radioactive elements, as detected by alpha
counting, and the Pb are preferentially concen-
trated in the sediment at the bottom of the beaker
Ž43A., whereas the zinc is displaced into solution
Ž43L..
To check for the possible influence of aging,
three groups of samples Ž10 samples total. of 33A
material were prepared as described above, and
the dried powders were stored in an oven at 658C
for periods of 1, 2 and 3 weeks, respectively, prior
to reaction with the peroxide. The behaviour of
all three groups of samples was the same. Using
results from the four 1-week samples for sake of
discussion only, it was observed that:
v the temperature rises spontaneously from 208C
to approx. 30]378C in approx. 10 min;
v it then rises more slowly to approx. 458C after
30 min;
v this is followed by a sharp rise to more than
908C after 40 min; and
v the temperature then drops rapidly to room
temperature after approx. 2 h.
Results for the other two groups were similar,
the major difference being that more time is
required to attain the maximum temperature. This
is probably due to surface oxidation of the powder
by contact with air during the waiting period.
Five grams of 33A powder was used in each of
these 10 experiments. After reaction, the solid
residue remaining in the container was filtered,
dried and weighed. The residue typically weighs
approx. 3 g } thus there is a 40% weight loss
during the reaction process. There is also a volume
loss of liquid due to frothing and an evolution of
a gas phase, one component of which was identi-
fied as hydrogen.
91
4. Results
4.1. Mineralogical changes resulting from the
peroxide reaction
Fig. 1 is a composite diffractogram showing the
profile for the starting material prior to reaction,
the residual sediment after reaction, as well as a
diffraction profile for the sediment obtained by
total evaporation of the liquid phase Žafter the
reaction had completely terminated.. The starting
material Ž33A. is shown to be a mixture of phases,
predominantly galena, sphaleriterwurtzite and
anglesite, whereas the diffractogram of the resid-
ual sediment after reaction Ž43A. indicates the
presence of anglesite only. In particular, two peaks
for ZnS, Žat 2 u s 33.28 and 35.78., are visible on
the ground Theisenschlamm curve whereas they
are absent from the middle curve. The diffraction
pattern for the liquid phase Ž43L. is dominated by
boyleite ŽZnSO4 ? 6H 2 O.. As the liquid phase was
almost totally clear after the reaction had termi-
nated, it is certain that the boyleite formed during
the evaporation process.
These results indicate that the zinc and lead
sulphides are oxidized and the zinc has been
dissolved. It was determined that the lead sul-
phide is converted to lead sulphate which is a
stable phase under surface conditions. It is
probable that the negligible residual zinc in the
sediment is situated within a more stable phase
and is not readily available for reaction with the
hydrogen peroxide. It has been shown by Weiss et
al. Ž1997. that some grains in the residue are
complete mixtures of Pb and Zn. It is therefore
possible that an anglesite coating forms around
the larger grains during reaction, thereby shield-
ing the Zn from further dissolution.
The ground Theisenschlamm Ž33A. has a me-
dian diameter of 1.20 m m for the component
particles whereas the same value for the sediment
after TSOP treatment is 0.82 m m.
4.2. Impact of the peroxide treatment on the
inorganic chemistry
Table 2 presents results of the analyses of the
92 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97

Fig. 1. Composite diffractogram showing the profile for the starting material prior to reaction, Ž33A., the residual sediment after
reaction, Ž43A. as well as a diffraction profile for the sediment obtained by total evaporation of the peroxide solution Žafter the
reaction had completely terminated Ž43L..

sediment after treatment Ž43A. as well as the
composition of ground Theisenschlamm Ž33A. and
the composition liquid phase after cessation of
the reaction Ž43L.. The most important feature of
Table 2 is seen by comparing the PbrZn ratio for
the ground Theisenschlamm prior to treatment
with that of the sediment remaining in the con-
tainer after treatment } in the former it is
roughly 0.75:1 whereas in the latter it is approx.
12:1. The same ratio in the liquid phase is 1:700.
The percentage of the original mass that was put
into solution for each element was calculated by
assuming 3 g of sediment remaining after treat-
ment and 280 cm3 of liquid phase. These values
are shown graphically in Fig. 2 in relationship to
their concentrations in 33A. Results show that
93.7% of the zinc originally present was placed
into solution whereas the comparable figure for
Pb is approx. 0.2%. In other words, the separation
of Pb from Zn is almost complete. Lead, S, Si and
Fe are the most important of the elements in the
residual sediment, whereas in the liquid phase,
the total of Zn q S is 92.6% of the total amount
present, with Cu being an important accompany-
ing element. Distribution coefficients, Žcalculated
as: K SrL s mass of element in sedimentrmass of
element in liquid. are also shown in Table 2 and
provide information on the partition of the ele-
ments between these two respective phases.
4.3. Impact of the peroxide treatment on the organic
chemistry
Results of the determination of the PAH con-
centrations using the EPA method 610 are pre-
sented in Table 3. The data show that PAH
values in 33A are much higher than in the resid-
ual sediment 43A. The total PAH concentration
in the untreated material is 468.9 mgrkg, whereas
in the residual sediment it is 11.25 mgrkg. This
represents a reduction of almost 98%.
The results of the identification of several cate-
gories of compounds were classified into groups
and are presented in Table 4. Sample 33A con-
tains only a few low volatile components, such as
the methyl and dimethylbenzene. The main part
of the hydrocarbons is a mixture of polyaromatic
compounds with a higher boiling point, however,
 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97 93

Table 2
Composition of a typical sample of ground Theisenschlamm and the composition of the products after hydrogen peroxide treatment

Element Ground Sediment Liquid % In sediment KSrL

Theisenschlamm 43A Žmgrkg. 43L Žgrl.
33A Žmgrkg.

Ag 520 810
Al 23 600 31 500 42.15 89 8.0
As 4080 7600 0.19 100 429.0
B 111 135 1.07 57 1.4
Ba 1670 1930
Bi 165 280
Ca 8800 2200 8.8 73 2.7
Cd 360 30 5.1 6 0.06
Co 80 35 1.12 25 0.33
Cr 890 1160 1.56 89 8.0
Cu 14 100 1215 183.5 7 0.07
Fe 30 850 43 700 1.45 100 323
Ga 78 140
Ge 31 32 0.23 60 1.5
Hg 110 130
In 23 45
K 6700 8350 27 77 3.3
Mg 2500 5200 12 82 4.6
Mn 670 580 5.82 52 1.1
Mo 600 880 0.02 100 471
Na 670 1400 4.3 78 3.5
Ni 510 360 4.57 46 0.84
Pb 122 000 182 000 3.75 100 520
Re 63 9 1 9 0.10
Stotal 113 000 39 900 1680 20 0.25
Sb 1470 2400 0.06 100 429
Se 720 320 8.4 29 0.41
Si 96 700 20 98 52
Sn 17 200 26 500
Th 2 3
Ti 5000 5580
Tl 320 290 2.18 59 1.4
U 30 32 0.12 74 2.9
Zn 163 000 15 800 2500 6 0.07
Zr 970 1135

Element Ground Theisen Sediment Liquid % In sediment KSrL

33A ŽkBqrkg. 43A ŽkBqrkg. 43L ŽkBqrl.
210
Pb 22.1 27.1 0.004 ) 99 ) 99
210
Po 16.6 19.9 0.2 ) 99 ) 99

additional alkylated homologues and partially hy-
drogenated aromatics were also detected, as were
some hydroxyaromatics and O-heterocyclic com-
pounds, quinoline andror isoquinoline as well as
dibenzothiophene derivatives. The data are pre-
sented as semi-quantitative countsrg of sample.
The GCrMS analysis of the residual sediment
is shown in column 2 of Table 4. The high boiling
temperature of the mixture has been reduced.
The values in Table 4 clearly show the decrease
in the concentration of aromatic compounds.
Specifically, the quinolinerisoquinoline deriva-
tives are absent and their alteration products, the
aromatic amino compounds, appear.
94 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97

Fig. 2. Element distribution pattern between the solid Žleft, 43A. and liquid Žright, 43L. phases after the TSOP treatment.

The analysis of the extract of the liquid phase
indicates a low concentration of hydrocarbons.
Water soluble substances, such as ben-
zenemethanol and benzenedicarbonyl acids are
predominant, however, volatile compounds, such
as the xylenes are also present. It is also apparent
that the concentration of the toxicologically rele-
vant N- and S-heterocyclic compounds has been
significantly reduced. Sample 43L contains insig-
nificantly small quantities of hydrocarbons, pre-
dominantly alkanes, some low boiling aromatics
and water soluble compounds, such as phthalates.
The results of the PCDDrPCDF analyses of
the two solid samples 33A and 43A are listed in

Table 3
Results of PAH measurements on ground Theisenschlamm and TSOP reaction products

Compound Ground Resid. sediment Liquid Proportion KSrL

Theisenschlamm after treatment phase in sediment
33A Žmgrg. 43A Žmgrg. 43L Žmgrg. Ž%.

Naphthalene 1.9 0.20 0.0140 13.3 0.15

Acenaphthene 0.3 0.03 0.0012 21.1 0.27
Fluorene 2.9 0.12 0.0033 28.0 0.39
Phenanthrene 111.7 0.04 0.0100 4.1 0.04
Anthracene 10.7 bd 0.0009
Fluoranthrene 71.0 1.90 0.0070 74.4 2.91
Pyrene 74.5 2.31 0.0014 94.6 17.68
Benzow axanthrancene 21.2 1.16 bd
Chrysene 62.3 2.20 bd
Benzow b xfluoranthrene 45.6 1.64 bd
Benzow k xfluoranthrene 4.7 0.27 bd
Benzow axpyrene 18.0 0.57 bd
Dibenzow a,h xanthracene 1.2 bd bd
Benzow g,h,i xperylene 31.7 0.63 bd
IndenoŽ1,2,3.pyrene 11.2 0.18 bd

Total 468.9 11.25 0.0354

 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97 95

Table 4
Results of GCrMS analyses of ground Theisenschlamm and TSOP reaction products

Compounds Ground Solid residue Liquid Ž43E.

Theisenschlamm Ž33A. after treatment Ž43A. Žarear148 ml water
Žarearg sample Žarearg sample =1 000 000.
=1 000 000. =1 000 000.

Alkanes 4.93 1.69 0.052

Polycyclic aromates 21.69 0.62 0.017
Benzene derivatives 0.41 0.49 0.411
Naphthaline derivatives 1.86 0.40 0.0018
Biphenyls andror 20.67 0.004 tr.
Dibenzofuranes
Diphenylethanes tr. 0.009 0.058
Quinoline andror 0.2 bd.
Isoquinoline derivatives
Aromatic amines 0.012
Dibenzothiophen derivatives 5.67 0.07 bd.
Fluorenon-9 0.43 0.005 tr.
Benzenemethanol 0.04 0.04 0.039
Styrene 0.03 0.035 0.014

Abbre¨ iations. b.d., Below detection limit; tr., trace amounts present.

Table 5 smaller chlorosubstituted isomer range. It is pos-

Results of PCDDrPCDF analyses on ground Theisenschlamm
sible that some of the more volatile Žtetrachloro-
and the sediment remaining after TSOP treatment
substituted . isomers might have been overlooked
Compound Ground Solid residue during the analysis } this is because an open
Theisenschlamm after treatment beaker was used during the exothermic oxidation
Ž33A. Žppb. Ž43A. Žppb. reaction.
Sum TCDF 86.0 47.1 A chemical change, involving the degradation
Sum PeCDF 40.3 30.1 or transformation of the dioxins and furans would
Sum HxCDF 10.5 14.7 produce changes in the isomeric pattern. How-
Sum HpCDF 8.4 6.2 ever, when compared with 33A, the relative pro-
OCDF 1.4 2.5
portion of the PCDDrPCDF compounds in the
Sum PCDF 146.6 100.6
43A sample was the same } only the total amount
Sum TCDD 5.9 0.7 varies.
Sum PeCDD 0.8 1.1
Sum HxCDD 1.8 0.3 5. Radioactivity
Sum HpCDD 1.9 bd
OCDD 5.5 bd
Sum PCDD 15.9 2.1 Alpha and gamma spectroscopy was carried out
on the starting material 33A as well as on the
PCDDq PCDF 162.5 102.7 TSOP reaction products 43A and 43L to de-
Abbre¨ iation. bd, Below detection limit of approx 0.05 ppb. termine the activity of 210 Po and 210 Pb, respec-
Also shown are the International Toxicity Equivalents ŽITE. tively. The radioactivity value associated with
210
expressed in ngrkg. Po in sample 33A prior to reaction was de-
termined to be 16.6" 0.4 kBqrkg. After the per-
Table 5. The sum total concentration of each oxide treatment, the value in the residual sedi-
group is shown in m grkg Žppb. of dried material. ment Ž43A. increased to 19.9 " 0.4 kBqrkg
A decrease in PCDDrPCDF concentration after whereas the value for the liquid phase Ž43L. was
peroxide treatment is observed, particularly in the less than the detection limit of 0.2 Bqrl, Žsee
96 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97
Table 2.. The radioactivity value associated with
210
Pb in sample 33A prior to reaction and in the
sediment Ž43A. were determined to be 22.1" 0.6
kBqrkg and 27.1 " 0.9 kBqrkg, respectively
whereas the value for the liquid phase was less
than the detection limit of 4 Bqrl. Mass balance
calculations indicate that virtually all of the 210 Pb
and 210 Po is concentrated in the sediment Ži.e. in
the lead sulphate phase., after treatment. The
total radioactivity value for the liquid phase is
- 4.2 Bqrl.
6. Discussion
The partition coefficients shown in Tables 2
and 3 clearly indicate that certain elements or
organic compounds are preferentially concen-
trated in either the liquid or solid phases after the
oxidation described above. In particular, the Zn,
Cd, Cu and Re partition strongly into the liquid
phase whereas the solid phase contains almost all
the As, Sb, Fe, Mo, Pb, Žincluding 210 Pb and
210
Po. and a high proportion of the Al, Cr, and Si
Ž) 89%.. Of the 24 elements evaluated, 11 of
them were concentrated Žmore than 75% of the
original mass present. in the sediment whereas
another four elements were more than 75% con-
centrated in the liquid phase. Those elements
normally present in silicates, such as Al, Mg, Na,
K and Fe Žand Si. are concentrated in the resid-
ual sediment since silicates are normally not able
to be attacked by hydrogen peroxide. The chemi-
cal separation between Pb and Zn is related to
the different solubility constants for ZnSO4 and
PbSO4 . It is therefore clear that this simple
method can be used as a first order procedure for
the separation of complex mixtures of residual
sulphides and silicates into two fractions, each of
very different chemical composition.
Such a treatment may have considerable poten-
tial application in remediation projects, particu-
larly for very fine-grained material. In the case of
the Theisenschlamm, lead, zinc and sulphur are
the most important components. Depending on
the technical specifications required by different
industries, the separation of the Zn from the Pb
described here may by itself generate an
economically useful product which could be used
as input feedstock. Alternatively, the process may
provide an initial treatment of such material which
would require further processing prior to obtain-
ing a suitable product.
It is important to emphasize the flexibility of
this technique. As it is a chemical treatment, it is
most suitable for very fine-grained material. It is
not influenced by the chemical composition of the
mixture, however, it is influenced by the possible
presence of external oxide coatings on the surface
of the particles. Conventional hydrometallurgical
and pyrometallurgical methods require significant
capital investment which would not be economi-
cally feasible considering the relatively small
amount of material to be treated ŽLeipner et al.,
1991; Lorenz et al., 1992..
7. Conclusions
1. Hydrogen peroxide is an excellent oxidising
agent for the Theisenschlamm. It caused the
near complete separation of the Zn from the
Pb, yielding a hydrated zinc sulphate that
could potentially be used in an electrolytic
process for zinc products and a lead sulphate
which could be used in lead smelting.
2. The elements considered can be divided into
three groups: those with a solidrliquid parti-
tion coefficient - 1:3 ŽZn, Cu, Cd, Re.; and
those with a solidrliquid partition coefficient
) 3:1 ŽPb, As, Sb, Mo, Fe, Si, Na, Cr, Mg, K,
Al. with the remainder ŽU, Ca, Ge, B, Tl, Mn,
Ni, Co and Se. being present in subequal
amounts in both liquid and residue.
3. The small amount of Zn remaining in the
solid residue after treatment does not appear
to be accessible at the grain size at which the
experiments were conducted.
4. The 210 Pb is strongly associated with the solid
phase Ži.e. lead sulphate. as is the 210 Po.
5. The peroxide treatment reduces the concen-
tration of PAH in the solid residue by almost
98% and the PCDDrPCDF concentrations
by approx. 40%.
6. After peroxide treatment, the liquid phase
contains insignificantly small quantities of hy-
drocarbons, predominantly alkanes, some low
 M. Morency et al. r The Science of the Total En¨ ironment 223 (1998) 87]97 97

boiling aromatics and water soluble com- Leipner K, Korb J, Hein K. Technisch-wissenschaftliche be-
pounds, such as phthalates. trachtungen zur verarbeitung von mansfelder Theisen-
schlamm Žflugstaub.. Erzmetall 1991;44:560]565.
Leipner K. Verhalten der Hauptbestandteile der Mansfelder
Theisenschlamms bei dessen pyrometallurgis cher Ver ar-
References beitung. In: Mansfeld eV, editor. Sanierungs-verbund,
Theisenschlamm-tagungsband zum kolloquium vom 7.12.93,
Flachowsky J, Mattusch P, Morgenstern P, Wennrich R. Un- Mansfeld, 1994.
tersuchungen zur anorganischen schadstoffverteilung in Lorenz R. Erfahrungen, versuche und projekte zur Theisen-
¨
Theisenschlammen, ¨
boden ¨ ¨
und oberflachenwassern im schlammverwertung. In: Mansfeld eV, editor. Sanier-
bereich der ehemaligen krughutte,¨ ¨
eisleben. In: Glaser ungsverbund ŽHrsg... Theisenschlamm-tagungsband zum
H-R, Jacob E, Schulz G, Stocker¨ G, Weiss H, editors. kolloquium vom 7.12.93. Mansfeld, 1994.
Untersuchungen zum rekultivierungsvorhaben der
Lorenz R, Tacke M, Hartmann K, Weilandt E. Theisen-
¨
gemeinnutzigen sanierungsgesellschaft mansfelder land
¨
schlamme im mansfelder revier, verwertung oder deponie?
¨
mbH auf dem gelande ¨
der ehemaligen krughutte, eisleben.
} ein konzept zur verwertung. Metall 1992;46:955]957.
UFZ-Bericht, 1992.
Weilandt E. Verfahrenskonzepte zur aufarbeitung des mans-
Freyer K, Treutler H-C. Radiometrische untersuchungen an
Theisenschlamm- und bodenproben aus dem bereich der ¨ zink- und
felder Theisenschlamms zu einem einsatzstoff fur
¨ eisleben. In: Glaser
ehem krughutte, ¨ H-R, Jacob E, Schultz ¨
bleihutten. In: Mansfeld eV, editor. Sanierungsverbund.
¨
G, Stocker G, Weiss H. Untersuchungen zum rekul- Theisenschlamm-tagungsband zum kolloquium vom 7.12.93;
¨
tivierungsvorhaben der gemeinnutzigen sanierungsge- Mansfeld, 1994.
sellschaft mansfelder land mbH auf dem gelande¨ der ehe- Weiss H, Morency M, Freyer K et al. Physical and chemical
¨
maligen krughutte, eisleben. UFZ-Bericht, 1992. characterization of a complexly contaminated scrubber dust
¨ H-R, Glasser
Glaser ¨ W, Jacob G, Weiss H. Abfallbeseitigung } a by-product from copper smelting in Sachsen-Anhalt,
in bergbaugebieten: altstandorte und altlasten } sicherung Germany. Sci Total Environ 1997;203:65]78.
oder sanierung? Ein fallbeispiel aus dem mansfelder land.
In abfallbeseitigung und deponien. Umweltgeologie Heute
1993;1:49]58.