f/EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-049 Sept. 1981
Project Summary
Investigation of
New Techniques for
Control of Smelter Arsenic
Bearing Wastes
Anil K. Mehta
Arsenic is the twentieth most
abundant element in the earth's crust
and is known to appear in 245 mineral
species. It is found in appreciable
concentrations in association with
sulfide deposits with arsenopyrite
(FeAsS) the most common form.
Because of its relative abundance and
modes of occurrence, it is a contam-
inant of coals, iron ore, phosphate
rock, and nonferrous metals. Domes-
tically, arsenic is produced as arsenic
trioxide recovered as a by-product
from copper smelting and is used to
produce arsenic metal and some 45
other compounds of commercial
significance. Uses include pesticides,
animal hide and wood preservatives,
feed additives, metal alloys, glass and
pigment manufacture, solar cells, and
catalysts. Most compounds of arsenic
are extremely toxic and the element is
also a suspected carcinogen.
Because of the potential hazards
associated with arsenic disposal, the
Industrial Environmental Research
Laboratory at Cincinnati, Ohio, spon-
sored a research project at the Mineral
Research Center, Montana Tech
Alumni Foundation, to investigate a
variety of approaches to the fixation of
arsenic-bearing wastes, particularly
smelter flue dusts, to render them
harmless to the environment. The
program was comprised of three
principal elements: (1) separation of
the arsenic from metal values by
leaching; (2) incorporation of the
arsenic into a fixation matrix; and (3)
leach testing of samples of the fixed
product to determine suitability for
environmental protection. Stabiliza-
tion of the arsenic by incorporation in
slag matrices and in cements, mortars,
clays and concrete was investigated.
The most promising technique found
was stabilization by dissolution in a
slag matrix to form a solid solution
upon cooling. The slag fixation prod-
ucts yielded leachate concentrations
of from 0.02 to 4.5 ppm for arsenics in
slag loadings ranging from 7 to 24
percent arsenic.
This Project Summary was devel-
oped by EPA's Industrial Environ-
mental Research Lab., Cinn., OH. to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
The flue dusts generated in the
process of copper smelting contain
appreciable amounts of copper and
other metal values including precious
metals. These flue dusts may also
contain appreciable amounts of arsenic
trioxide where arsenic (sulfide) occurs
in the ore body. In the past, it had been
common practice at some smelters to
-------
collect metal values in the flue dust by
means of high temperature electrostatic
precipitation (ESP), with much of the
arsenic trioxide passing the ESP in the
vapor state and being emitted to the
atmosphere. Now, smelters processing
high arsenic concentrates cool the flue
gas to desublimate the arsenic trioxide
for control by a particulate arrestor. Two
types of control are practiced: (1) a high
temperature ESP, cooling, and cold ESP;
and (2) cooling and cold fabric filter.
Therefore, smelters processing concen-
trates having a medium to high arsenic
content generate residuals containing
arsenic trioxide. Present practice is to
store these dusts. Where the residuals
contain metal values, they are stored for
retrieval and processing for metal
recovery at some future time when
economic recovery becomes possible.
There is now only one smelter in the
United States that treats flue dust to
produce arsenic for sale. The decline m
demand and the competition from
overseas sources to supply arsenic
compounds have created an unfavor-
able economic condition that has
resulted in rather large inventories of
accumulated flue dust in storage.
Because of the toxic nature of the flue
dust, this storage must be in weather
proof structures. At this time, it does not
appear likely that sufficient market
capacity for arsenic or its compounds
will be found to consume the stockpiles
being generated. Disposal of arsenic in
some environmentally acceptable form
may, therefore, become necessary.
Eventually the main receptorfor arsenic
will be the land via landfill disposal or
solution ponding. If land disposal of
arsenic is to be practiced, the volume of
dust should be reduced. This can be
accomplished by concentrating the
arsenic content. The objective is to
remove metallic values from the dust,
thereby leaving a residue ready for
subsequent metallurgical treatment for
fixation of the arsenic.
Montana Tech Minerals Research
Foundation (MRC) began an EPA sup-
ported project in November 1976 (EPA
Grant R-804-1-595) to investigate
means of treating arsenic containing
copper smelter flue dusts A compre-
hensive literature survey was initiated
to gather information about arsenic that
could possibly lead to a technique or
techniques for the safe disposal of this
material. Information was sought on the
general chemistry and geochemistry of
arsenic, the extractive metallurgy of
arsenic and its removal as an impurity
from metallurgical systems, wastewater
purification, and existing control
equipment and techniques. The survey
resulted in a collection of over seven
hundred references. The MRC arranged
for the participation of interested
industry representatives to provide
insight into the potential usefulness of
the research and performance of certain
aspects of the experimental work.
An experimental research program
was devised and initiated in March
1977. This program was structured to
investigate methods to separate the
arsenic from the flue dust so that the
metal values could be recovered, and
the arsenic fixed in a form which would
be stable in the environment, this
permitting disposal to the land without
extraordinary safety precautions. A
number of approaches were investi-
gated, and extensive leach tests of the
fixed products were conducted over
periods ranging from 6000 to 9000
hours.
The experimental program was initially
structured around the hydrometallurgi-
cal treatment of smelter flue dust to
separate metal values from the arsenic
trioxide and subsequent recovery of the
arsenic through precipitation as an
arsenate or by sorption. In an indepen-
dent approach, a number of schemes for
arsenic fixation in the form of flue dust,
arsenic trioxide, and calcium and iron
arsenates were investigated. This work
lead to the discovery that calcium
arsenate was an excellent form for
fixation in selected media. Because of
this, methods to convert the arsenic in
flue dust to calcium arsenate in a dry,
elevated temperature process was
pursued. A technique for using calcium
compounds such as lime was developed
for possible use for converting the
arsenic trioxide in flue dust to form
suitable for fixation.
Extraction of Arsenic
from Flue Dusts
A program of leaching experiments
was organized to investigate the solu-
bility characteristics of a variety of
arsenic-bearing flue dust materials
collected from domestic primary pro-
ducers of copper, zinc, and lead (Tables
1 and 2). Leaching experiments were
conducted using six conventional
lixiviants (water, sulfuric acid, ferric
sulfate, ferric chloride, aqueous am-
monia, and sodium hydroxide) in various
combinations. Leach extraction was
determined as a function of time,
temperature, and reagent concentration.
Table 1. Copper Smelter Flue Dust Composition Ranges
Element Percent in dust
Cu
Pb
Zn
As
Bi
Ag
Au
5-15
8-15
5-15
3-30
0-2
3-13 02/ton
0-0.2oz/ton
Table 2. Example Copper Smelter Flue Dust Assays
Assay, %
Element
As
Bi
Cd
Cu
Fe
Pb
Sb
Sn
Te
Zn
S
Mo
Ag
Dust 1
20.0
0.65
0.94
4.76
6.1
5.8
0.63
0.75
0.27
14.3
10.0
—
—
II
29.4
0.94
0.65
3.91
1.43
5.3
1.26
1.60
0.69
12.3
7.9
—
—
Ill
25.6
0.88
0.89
4.19
1.43
5.43
0.74
0.5
—
14.2
9.1
0.09
5. 4 02: /ton
IV
3.54
0.33
0.41
19.2
14.6
5.01
0.16
—
.046
7.12
12.4
—
10.31 oz /ton
-------
Results of the leaching tests indicated
that up to 99 percent of the contained
arsenic in several of the dusts examined
could besolubilizedby at least one of the
hxiviants tested under appropriate
conditions Due to the complex and
variable mineralogy of the arsenic in the
dusts examined, no universally effective
method of leaching capable of complete
dissolution of arsenic was found The
complexity of both the chemical and
mmeralogical compositions of smelter
flue dusts also resulted in widely
dissimilar solubility characteristics of
other heavy metal constituents giving
rise to a wide range of solution treatment
problems following leaching. All leach
tests performed resulted in the dissolu-
tion of significant amounts of copper,
iron, zinc, cadmium, antimony, and
tellurium in addition to arsenic.
Although a solvent that would be
applicable to all flue dusts was not
identified, each flue dust responded
effectively to at least one of the solvents.
Specific parameters of reagent concen-
tration, time, temperature, and liquid-
solid ratio must be determined experi-
mentally for each individual flue dust in
order to maximize arsenic extraction,
achieve selectivity with respect to other
metal constituents, and to conserve
chemical reagents No single combina-
tion of lixiviant composition and leaching
conditions is considered optimum for a
diversity of flue dust waste materials.
Most flue dusts responded well to a
combination of sulfatization followed by
either water or sulfuric acid leaching.
The added advantage of nearly complete
extraction of copper, zinc, and cadmium
favors the use of sulfuric acid as
opposed to the other reagents examined
by providing a degree of segregation of
other dust constituents in addition to
nearly complete arsenrc extraction In
this way, a sizable fraction of the
original dust containing a variety of
impurities deleterious to most nonfer-
rous smelting operations can be isolated
from the arsenic as well as several
important metals of value. Such a solid
residue constitutes an important means
of impurity rejection in most smelting
operations.
Fixation of Arsenic
The recovery and fixation of arsenic
forms from the leach liquors and other
arsenic compounds readily available in
the lab, such as arsenic trioxide,
calcium, and iron arsenates, was the
subject of broad based investigation. A
number of arsenic "getter" compounds
were considered in this screening study.
The fixed products were subjected to
leach tests to determine the stability of
these products. Additionally, a program
of arsenic fixation experiments was
organized to investigate the stabili-
zation of arsenic by dissolution in slag
matrices, and by encapsulation in
cements, mortars, clays, and concrete
The success or failure of the stabilization
was determined by reasonably long-
term leach tests. These investigations
were based upon the use of a suitable
dry form of arsenic as a starting material
for the fixation experiment Such forms
included unprocessed flue dust, calcium
arsenate, iron arsenate, and arsenic
trioxide. Both ambient and elevated
temperature approaches were studied.
Sorption/Fixation from Leach
Solutions
Laboratory tests to effect the recovery
of arsenic from pregnant leach solutions
included precipitation of arsenic by
chemisorption' on hydrated ferric
oxides generated by hydrolysis of ferric
sulfate present in the leach liquor, and
on particulate surfaces of phosphate
rock, bentonite clay, and cement plant
cottrell dust. The arsenic content of
sulfate-based leach liquors can be
effectively recovered as a solid residue
by coprecipitation with hydrated ferric
oxide at mildly acidic solution pH
Residual arsenic concentration in
solution can be reduced to well below
one part per million when such precipi-
tation is carried out in the presence of
naturally occurring calcium phosphate
(collophanite), although a reversion of
arsenic solubility to approximately one
part to seven parts per million occurs
when arsenic-bearing hydrated ferric
•oxide precipitates are equilibrated with
pure water over periods of several
months. The use of particulate material
to chemisorb arsenic from solutions
does not appear to be a viable technique.
Excessively large quantities of solid
particulate would be required for
successful removal of arsenic to
desirable concentration levels There-
fore, large quantities of sludge material
would have to be handled and stored
Clays
The concept employed here was to
cause physical entrapment of the
arsenic with a clay matrix. Bentonite
and Kaolin were used as the clay
sources. Bentonite contains a high
percentage of montmonllonite clay
mineral. Arsenic sources used were
arsenates of iron and calcium and
arsenic trioxide
The mixtures in specified weight
ratios were mechanically pelletized,
then roasted to give dense and physically
strong structures The pellets were
formed with specific arsenic concentra-
tions and then heated at various
roasting temperatures and times As a
variation, some pellets were coated
with ceramist's glazing compound and
glazed These pellets were subjected to
the immersion type leach test.
The pH of the solution was allowed to
take its own course in most of the cases
unless specified. For arsenic analysis,
two ml of the leach solution were taken
out at specified time intervals and
replaced with two ml of the distilled
water The pellets of iron or calcium
arsenates resulted in very low leaching
(i e., less than 1 ppm, with most samples
less than 0 5 ppm) at arsenic load ings as
high as 75 percent. However, the
arsenic trioxide pellets proved to be
unsuitable, yielding high leachate con-
centrations of from 40 to 100 ppm, as
for pellet loadings as low as 0 5 to 1.0
percent
Cement
Arsenic compounds were mixed in
different proportions with the cements
Three treatments were investigated
(a) The materials were mixed with
water, cast, and cured
(b) Materials were mixed and roasted
at different specified time and
temperatures Water was added
next, the resulting mixtures were
cast and cured.
(c) Materials were mixed with water,
cast, and cured The cubes were
then roasted at specified time and
temperatures.
A major reason for setting of cements
into hard and dense structures is the
crystallization of tricalcium silicate (3
CaO-Si02>. Arsenic trioxides and calcium
arsenates have been used in portland
cement mixtures for special uses (i e ,
retarded early-setting time, early high
strength, and resistance to chemical
action of water). Addition of arsenic
compounds is limited to certain quanti-
ties as they retard the setting of
cements Failure to achieve a good
sample which is solid, unfractured, and
hard to break is the result of retardation
in crystallization of tricalcium silicate
compound in the cements Retardation
by admixtures like arsenic compounds
-------
is so enhanced at higher concentrations
that cement setting does not take place
and the final product is soft and
crumbly.
The cast samples of arsenic cement
mixtures were leach-tested for arsenic
release using a shaker test. Leaching
characterisitics of roasted arsenate
containing cements are good.
Slags
Arsenates were used as starting
materials mainly for their arsenic
carrying ability (i.e., they have a low
vapor pressure and will not boil out of
the system before the slag becomes
molten) Flue dust or arsenic oxide
cannot be used directly because it will
vaporize away from the system before it
has a chance to dissolve
The solvents chosen for study are
based on two primary considerations
1 The availability and low cost of the
materials, and
2. The potential for high arsenic
dissolution.
The first of these considerations is
fulfilled by both by-product slags from
smelting operations and impure clays.
The second consideration (i.e , the
concentration of arsenic that can be
dissolved), can only be postulated
because only a few phase diagrams are
available for arsenic containing systems.
These considerations are no more
important than the requirement for the
solvents to retain arsenic in such a way
as to restrict its release to the environ-
ment. This requirement may be termed
as chemical fixation of arsenic Chemical
fixation is defined for the purpose of this
report as the formation of a chemically
stable compound containing arsenic
that upon weathering will not release
arsenic to the environment in detri-
mental quantity. Arsenic containing
glass (slag) is one of the chemical
fixation possibilities
Glasses are defined by the American
Society for Testing Materials as "an in-
organic product of fusion which has
been cooled to a rigid condition without
crystallization." For the purposes of this
summary, a slag will be considered a
glass. Slag is a multicomponent oxide
mixture generated as a by-product from
many smelting operations It is normally
treated for storage eithei by water
quenching it or slow cooling it The
water quenched product is mostly non-
crystalline while the slow cooled
product is a matrix of crystalline and
non-crystalline phases.
The experimental procedure used for
preparation of the arsenic containing
slag was as follows: The solid arsenate
was weighed in the desired proportion
and added to a fireclay crucible. The
material or a mixture of the materials to
be used as the solvent was added to the
crucible in solid form. The crucible was
then placed in an electric muffle furnace
at the desired temperature. Furnace
atmosphere was not controlled and was
assumed to be of slightly oxidizing
nature. The samples were melted,
removed, and poured into sample
recovery pans They were allowed to
cool in air.
Lime Roasting for Arsenic
Fixation
Because of the success with the high
temperature techniques for arsenic
fixation, a pyrometallurgical route to
conversion of arsenic trioxide to a
refractory compound was sought.
Based upon the literature review and
the results obtained from the pellet
roasting experiments, a number of
arsenic gettering compounds were
screened for their ability to retain
arsenic trioxide under oxidative roast
conditions as an arsenate form.
Many getter compounds were con-
sidered in this screening study (e.g , iron
bearing compounds such as FeaOs,
FeS2, reverberatory slag); aluminum
bearing compounds (e.g., AI203, kaolin);
and calcium bearing compounds (e.g.,
CaO, CaC03, power plant fly ash,
phosphate ore and phosphate slag). The
experimental procedure was to mix
proportioned amounts of the getter
material with either AsaO3 or flue dust,
pelletize (1/2-mch diameter spheres),
and dry at 50°C for 24 hours. The pellets
were then stored in vials until roasted
Roast tests were performed in an air
atmosphere in a muffle furnace. The
temperature range investigated was
200°to 400°C After roasting, the
pellets were weighed, ground, and
analyzed for arsenic content by standard
analytical procedures. Identification of
the arsenic compound present was by x-
ray diffraction
The most effective calcium containing
getters were CaO and Ca(OH}2. Those
found to be ineffective were waste
products that contained calcium, such
as cement plant dust (75% CaCOs), a
power plant fly ash (calcium dissolved in
slag), phosphate ore [calcium asCa^POa}?
CaF2] and a phosphate slag (calcium ES
CaSi03>. This study shows that the
arsenic trioxide can very effectively be
converted to calcium arsenate by low
temperature, short time air' roasting.
This approach permits conversion ol
arsenic-bearing flue dusts to a form
suitable for incorporation into a slag for
disposal. Alternatively, the process may
potentially be adapted to permit recovery
of metal values either through: (1]
separation and recovery of the arsenic
trioxide from the flue dust by volatili-
zation and condensation; (2) separation
of the arsenic trioxide from the flue dust
by volatilization and recovery by reaction
with a high temperature lime bed; or (3!
return of the roasted pellets to the
smelting furnace where the calcium
arsenate is incorporated in the slag.
Conclusions
A summary of the most significant
results from the leach tests is given in
Table 3. These results show that arsenic
in the form of calcium or iron arsenate
can be incorporated into a variety of
matrices, yielding stabilized products
which, when subjected to extensive
leach testing, permit only minimal
releases to the environment. Limited
testing with arsenic oxides showed thai
this form is not suitable for direct
fixation in any of the matrices investi-
gated in this study because sufficiently
low leachability could not be obtained at
reasonable arsenic-to-matrix ratios
Hydrometa/lurgy
The results of the experimental
leaching studies on the dissolution of
arsenic from smelter flue dust wastes
indicate that leaching can be an
effective impurity rejection method
dependent upon the specific chemical
and mineralogical compositions of such
materials. Specific parameters 01
lixiviant composition, time, temperature,
and liquid-to-solids ratio must be
determined empirically for each indi-
vidual flue dust in order to maximize
arsenic extraction, achieve selectivity
with respect to other heavy metal
constituents, conserve chemical rea-
gents, and mollify reaction conditions
No single combination of lixiviant
composition and leaching conditions is
considered optimum for a diversity of
flue dust waste materials.
The arsenic content of sulfate-based
leach liquors can be effectively recovered
as a solid residue by coprecipitation
with hydrated ferric oxide at mildly
acidic solution pH. Residual arsenic
concentration in solution can be reduced
-------
Table 3. Summary of Fixation Test Results
Matrix
Clay
Cement
None
Cement
Cement
Clay
BF slag
BF slag
RF slag
RF slag
Cement
Cement
Cement
Cement
Concrete
Clay slag
Arsenic form
As203
As2O3
Iron arsenate
CA or FA
FA
CA or FA
FA
CA
FA
CA
FA
CA
CA
CA
CA or FA
FA
Arsenic
loading
1%
<25%
JOO%
25%
25%
to 75%
<20
<25
<10
<20
<10%
<25%
<25%
<10%
4%
<15
Leach
time (hrs)
168
9,OOO
1,700
9,OOO
9.OOO
9,OOO
8,OOO
8,000
8.OOO
8.0OO
9,000
9,000
9,000
9,000
6,OOO
7,800
Arsenic
concentration
40-100 ppm
<25
6-7
<2.0
.7
-------
Table 4. Arsenic Extraction from Arsenic Doped Copper Reverberatory Slag
A rsenic extraction'1'
Arsenic in
slag (%)
0.54™
0.77™
2.1
3.3
5.2
8.8
17.4
19.2
23.5
Exposed time
(hrs.)
7320
8304
1536
1536
1536
7944
7992
8309
7152
As concentration
{mg/liter)
0.070
<0.038
0.24
0.34
0.14
0.48
2.19
1.21
2.21
wLeachant: deionized water; pH - 6, so/id/liquid ratio - 1/100.
™As received commercial copper reverberatory slag.
Table 5. EPA Toxicity Test Results for Doped Slag Systems
Analysis of leach solution from extraction test (mg/l)
Arsenic in
slag (%)
0.54
0.77
2.1
3.3
5.2
9.1
19.4
23.5
As
0.016
0.047
0.448
0.421
0.901
0.415
0.802
1.791
Cd
0.093
0.000
0.000
0.000
0.000
0.001
0.002
0.001
Cr
O.016
0.007
0.006
0.004
0.007
0.007
0.007
0.008
Pb
0.226
0.149
0.169
0.500
0.150
0.148
0.149
1.142
EPA designated concentration of contaminants for characteristic toxicity {mg/l)
As 5.0
Cd 1.0
Cr 5.0
PbS.O
small. But the loss in metal values at
smelters that treat high arsenic concen-
trates would be significant.
It is recommended that further study
be directed toward determining the
potential for recovering the metal
values in the flue dust while disposing
of the arsenic via recycling roasted lime-
flue dust mixtures back into the rever-
beratory smelting furnace The arsenic
(now in a non-volatile form) should
preferentially distribute to the slag
phase and the copper and silver should
distribute to the matte phase from
which they can be subsequently re-
covered. Experimental evidence has
been generated by Luigi at the Anaconda
smelter and shows that, as the lime
content of the slag phase is increased,
the arsenic content of the slag phase
also increases. Therefore, the slag
phase should function as an arsenic
bleed from the system. If the arsenic is
forced into the slag phase by increasing
the lime content of that phase, the
question then becomes what happens
to the distribution of the other constitu-
ents? Presently available distribution
data show that the copper and silver
should distribute to the matte phase.
Anil K. Mehta is with the Mineral Research Center, Montana Tech Alumni
Foundation, Butte, MT
John O. Burckle is the EPA Project Officer (see below)
The complete report is in two volumes, and is entitled "Investigation of New
Techniques for Control of Smelter Arsenic Bearing Wastes."
Volume I. Experimental Program (Order No PB 81-231 581; Cost- $21 50)
Volume II Literature Review (Order No PB 81-231 599, Cost: $12 50)
These reports will be available only from: (prices are subject to change)
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone- 703-487-4650
The EPA Project Officer can be contacted at.
Industrial Environmental Research Laboratory
U.S Environmental Protection Agency
Cincinnati, OH 45268
•fr U S GOVERNMENT PRINTING OFFICE, 1981 - 757-012 "7323
-------
Postage and
United States Center for Environmental Research Fees Paid
Environmental Protection Information Environmental
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED Third-Class
Bulk Rate
I F K L 0 1 2 0 H ft U
0 A V I i' K F t
t P A K / 11 V
A X, nv! njv ^30
L P| A7A
------- |