United States
Environmental Protection
Agency
Office of Research
and Development
Washington, DC20460
EPA/600/R-92/105
August 1992
&EPA Arsenic & Mercury
Workshop on Removal,
Recovery, Treatment, and
Disposal
Abstract Proceedings
Alexandria, Virginia
August 17-20, 1992
-------
-------
EPA/600/R-92/105
August 1992
ARSENIC AND MERCURY
WORKSHOP ON REMOVAL, RECOVERY, TREATMENT,
AND DISPOSAL
Abstract Proceedings
Coordinated by:
Science Applications International Corporation
Ft. Washington, PA 19034
Work Assignment Managers:
Ronald J. Turner
Risk Reduction Engineering Laboratory
Office of Research and Development
Jose Labiosa
Waste Management Division
Office of Solid Waste and Emergency Response
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
i WASTE MANAGEMENT DIVISION
OFFICE OF SOLID WASTE AND EMERGENCY RESPONSE
US. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
Printed on Recycled Paper
-------
DISCLAIMER
The abstracts contained in this Abstract Proceedings do not necessarily reflect the views of the Agency
and no official endorsement should be inferred. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
11
-------
FOREWORD
Today's rapidly developing technologies and industrial practices frequently carry with them the
increased generation of materials that, if improperly dealt with, can threaten both public health and the
environment. The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's
land, air, and water resources. Under a mandate of national environmental laws, the Agency strives to formulate
and implement actions leading to a compatible balance between human activities and the ability of natural
systems to support and nurture life. These laws direct the EPA to perform research to define our environmental
problems, measure the impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning, implementing, and managing
research, development, and demonstration programs to provide an authoritative, defensible engineering basis
in support of the policies, programs, and regulations of EPA with respect to drinking water, wastewater,
pesticides, toxic substances, solid and hazardous waste, and Superfund-related activities. This publication is
one of the products of that research and provides a vital communications link between researchers and users.
The workshop on removal, recovery, treatment, and disposal of arsenic and mercury wastes was
convened to provide a forum for the exchange of state-of-the-art information among industrial, commercial, and
academic experts and the EPA's research and regulatory program offices. This Abstract Proceedings highlights
the technical presentations of the workshop.
Additional information may be obtained by contacting the authors or the EPA offices.
E. Timothy Opp,elt, Director
Risk Reduction Engineering Laboratory
in
-------
ABSTRACT
A workshop on removal, recovery, treatment, and disposal of arsenic and mercury wastes was held in
Alexandria, Virginia on August 17 through 20,1992. The goals of the workshop were: 1) to examine the chemical
fundamentals and analytical issues related to As and Hg compounds, 2) to disseminate information of the state
of practice in source reduction, in technologies that recover or remove As and Hg from industrial wastes and
recycling or reuse processes, and 3) to present a discussion of existing and emerging technologies that treat
Industrial wastes or contaminated soil and water, and the storage and disposal of treated wastes.
This Abstract Proceedings is organized into two sections: Session A - Arsenic and Session B - Mercury
and contains extended abstracts of the paper presentations. The abstracts cover papers on fundamentals-
analytical techniques/characterization; removal, recovery, and reuse; and treatment, storage and disposal '
IV
-------
TABLE OF CONTENTS
SESSION A - ARSENIC
Introduction to Arsenic Chemistry and'Analysis ., , 1
Arsenic Chemistry in Relation to the Disposal and Stability ,
of Metallurgical Extraction Wastes 4
Arsenic: Supply, Demand, and the Environment ,..;... 8
Recovery of Arsenic as a Raw Material for Reuse 13
Alternative Preservative Systems: Pros & Cons ....;,.........20
Potential for Recovery of CCA from F035
Wood Preserving Operations;... , -.;...; :..... ..24
Reuse of Wood Preservative That Contains Arsenic 27
Osmose Water Purification System To Remove CCA
Contaminants from Water 30
The Cashman and Other Hydrometallurgical Process Treatments
of Polymetallic Arsenical Dusts, Sludges, and Wastes 33
Arsenic Extraction from Silt and Clays 39
The Behavior of Arsenic in a Rotary Kiln Incinerator 43
Removal of Arsenic from Wastewaters and Stabilization of
Arsenic Bearing Waste Solids 46
Treatment of Landban-Varianced Arsenic Wastes at TSDFs 51
Characterization of Arsenic-Containing Mining/Smelting Wastes in the
Clark Fork Basin, MT and Some Potential Remedial Technologies 56
Solidification/Stabilization of Arsenic Compounds 60
Vitrification of Waste Streams Containing RCRA
Metal Compounds 64
SESSION B - MERCURY
Elemental Mercury in Soil and the Subsurface:
Transformations and Environmental Transport
Research Program for Dealing with Mercury in Soil at
Natural Gas Industry Sites
Mercury Containing Hazardous Wastes:
Generation and Potential Reduction
Recent Advances in the Analytical Techniques for the Quantification
of Mercury and Mercury Compounds in Different Media
Mercury in Sediments - How Clean Is Clean?
Effect of Chemical Form of Mercury on the Performance of Dosed
Soils in Standard Leaching Tests: EP and TCLP
Inter-Laboratory Testing for Mercury by TCLP and Source
Reduction in the Lamp Manufacturing Industry
Management of Medical Mercury Battery Wastes
Through Source Substitution
Recovery of Mercury D-009 and U-151 Waste from Soil
Using Proven Physical and Gravimetric Methods
Treatment and Mercury Recovery from Electrical
Manufacturing Wastes
.69
.70
.73
.77
.78
.81
.86
.92
.96
.99
-------
TABLE OF CONTENTS (Continued)
Development of BOAT for the Thermal Treatment of
K-106 and Certain D-009 Wastes 100
Mercury Removal with lonac Ion Exchange Resins 103
Development of Bacterial Strains for the Remediation
of Mecurial Wastes 106
The Recovery of Mercury from Mineral Extraction Residues
Using Hydrometallurgical Techniques 110
High Vacuum Mercury Retort Recovery Still for Processing
EPA D-009 Hazardous Waste 113
Non-Thermal Processing of K-106 Mercury Mud 117
Biological and Physio-Chemical Remediation of
Mercury-Contaminated Waste 121
VI
-------
INTRODUCTION TO ARSENIC CHEMISTRY AND ANALYSIS
Edwin A. Woolson
EPL Bio-Analytical Services, Inc.
395 N. Memorial Parkway
Harristown, IL 62537
USA
Tel: (217) 963-2143
Arsenic at waste sites may be present from many natural or man-made sources. In all cases, the
arsenic present may be subject to metabolism and/or movement (1) through or from the site.
Arsenic occurs naturally in about 245 mineral species (2). These range from the native element or
alloys to arsenides, sulfides, sulfosalts, and oxidation products (oxides, arsenites, and arsenates). In
addition, a variety of complex organoarsenic compounds may be present at sites containing waste from coal
utilization or oil production.
These compounds can oxidize in aerated soil to arsenates and be subsequently reduced to arsenites,
various alkylarsines, and trimethylarsine oxide (TM AO). The arsines are volatile and can become dispersed
in air to return ultimately as oceanic sediments (3). Organoarsines can be oxidized to methanarsonate or
cacodylate. These are a part of the natural cycling of arsenic in the environment. These reactions area result
of oxidation, reduction, and microbial activity.
In sampling for arsenic residues, the nature of the residues expected based on past contamination or
environmental concerns must be considered in choosing sample-handling techniques. For instance, if the
only concern is how much arsenic is present and not which actual species is present, then no special
handling techniques need be employed. However, if differences between arsenate (+5) and arsenite (+3)
need to be determined, then considerable care must be taken in handling the samples. Organic arsenicals,
in general, are not affected by handling during the sampling or analysis process.
To a large degree, the soil Eh determines the ratio of +3 to +5. Soil Eh is not a function of a single
compound or component but of a combination of factors: e.g., iron content, pH, microbial population, and
moisture content.
In general, as iron increases, arsenite/arsenate decreases. Arsenate is the predominant arsenic form
in aerobic soils. Arsenite, however, is formed at Eh <300 millivolts (mV) over a pH range of 4 to 8, typical
soil values. As the Eh decreases due to flooding or a variety of conditions, the arsenite content increases
while arsenate decreases. With active microbial populations, some reduction of cacodylic acid or
methanearsonic acid to volatile alkylarsines is frequently observed.
What does this imply for waste site arsenic determinations? If the concern is whether the arsenic
present is arsenite, the Eh and pH should be measured at the time of sampling. If Eh is >300 mV, arsenate
will be predominant and no precautions to prevent oxidation should be necessary (2). If however the Eh is
<300 mV, care must be taken to prevent oxidation after sampling. The container may be purged with
nitrogen. The sample must be analyzed as quickly as possible since oxidation/reduction can occur even
when the sample is frozen. Lowering the temperature slows the reaction.
If the sample is aqueous instead of solid, the addition of ascorbic acid will slowthe oxidation of arsenite
to arsenate (4, 5). Analysis should be performed as soon after sampling as possible.
Afterobtaining and preparing the sample, the effect of the extractant on changes in the arsenic species
must be considered. Oxidizing reagents such as sulfuric or nitric acids will oxidize arsenite to arsenate.
Dilute aqueous salts or non-oxidizing acids (e.g., hydrochloric, phosphoric acids) will not significantly affect
the +3 to +5 ratio, although as with sampling, samples should be processed quickly. Arsenite will convert
to arsenate with time. The extractant will not change the organic arsenicals to inorganic unless acid
digestion is employed, in which case all arsenicals are converted to arsenate.
1
-------
In aqueous 8 normal (N) hydrochloric acid (HCI) extracts, benzene (and perhaps toluene) will extract
arsenite but not arsenate. This extract can then be back-extracted with water to allow separation of arsenate
and arsenite.
Following extraction, a cleanup separation is frequently necessary depending on the analysis
technique to be used. Metals may be extracted with organic complexation reagents (e.g., dithiazone,
quinolinol sulfate) depending on which metals are causing the difficulty. For example, high iron (Fe) levels
in extracts will interfere in liquid chromatographic separations of arsenite/arsenate. Iron complexed with
quinolinol sulfate can be adsorbed by charcoal and removed as an interferant (6). Aluminum on the other
hand is not removed.
The necessity for removal of interfering cations will depend on the metals present at specific sites.
Metals which mayinterfere include but are not limitedto: lead, Fe, copper, cobalt, manganese, and calcium.
The degree of interference is probably related to the amount of arsenic in solution due to solubility product
considerations and will vary at every site.
Detection methods for environmental analysis generally require nanogram to microgram detection
ranges (3). One of the detection methods frequently used involves hydride generation (7,8). The hydrides
after generation may then be trapped on glass beads using liquid nitrogen (5,9). Detection can be by DC
discharge (7), microwave emission (10), electron capture, orflame ionization (11) on a gas chromatograph
following differential heating.
Arsenite is converted to arsine by sodium borohydride at any pH, whereas arsenate is reduced at pH
<4.0. In practice, arsenite is determined at pH >5 and arsenate, methanearsonate, cacodylate, and TMAO
are converted to their respective arsines at pH 1.5 (3).
Another means of separating arsenicals involves ion exchange column chromatography. An AG 50W-
X8 cation exchange resin (12) can be used to separate arsenicals. At pH >1.5, arsenite and arsenate elute
before methanearsonate, followed by cacodylate. Using a mixed resin, AG 1-X8 (100 to 200 mesh), the
order of elution is arsenite < methanearsonate < arsenate < cacodylate (13).
Ion chromatography may be interfaced with a quartz furnace detecting a continuous hydride flow by
atomic absorption spectroscopy. In this instance, elution order proceeds: cacodylate < arsenite <
methanearsonate < arsanillate < arsenate (14).
A refinement of the technique involves High-Performance Liquid Chromatography (HPLC) in combi-
nation with graphite furnace atomic absorption (15-17). An intermediate strength anion exchange column
is eluted with a waterto 0.2 molarammonium carbonate using a gradient (6). The effluent is passed through
a sample cup which holds 50 microliters before going to waste. The flowing sample stream is sampled at
about 50-second intervals and the graphite furnace fired. The peaks resulting from furnace firings are
summed, the background subtracted, and response factors calculated. The elution order forthis system is:
TMAO < arsenite < cacodylate < methanearsonate < arsenate. This technique has been used to analyze
soil and water samples.
REFERENCES
1. Woolson, E.A. Fate of arsenicals in different environmental subtrates. IQ: Environ. Health
Perspectives. Ch. 19.1977. pp. 73-81.
2. Subcommittee on Arsenic, National Research Council. Chemistry or Arsenic and Distribution of
Arsenic in the Environment. In: ARSENIC, Medical and Biologic Effects of Environmental
Pollutants, National Academy of Sciences, Washington, DC, 1977. pp. 4-79.
3. Braman, R.S. Environmental reaction and analysis methods, in: B. A. Fowler (ed.), Biological and
environmental effects of arsenic. Vol. 6. Ch 3. Topics in environmental health. Elsevier Science
Publishers B.V., Amsterdam, Netherlands, 1983. pp. 141-154.
-------
4. Irgolic, K.J., Stockton, R.A., and D.C. Chakraborti. Determination of Arsenic and arsenic
compounds in water supplies. In: Lederer, W.H. and Fenterheim, R.J. (eds.), Arsenic: Industrial,
Biomedical, Environmental Perspectives. Van Nostrand Reinhold Co., New York. 1983. pp. 282-
306.
5. Feldman, C. Improvements in the Arsine Accumulation-Helium Glow Detector Procedure for
Determining Traces of Arsenic. Anal. Chem. 51: 664-669,1979.
6. Woolson, E.A.and Aharonson, N. Separation and Detection of Arsenical Pesticide Residuesand
Some of Their Metabolites by High Pressure Liquid Chromatograms (etc.). J. Ass. Offic. Anal.
Chem. 63(3): 523-528,1980.
7. Braman, R.S. and Foreback, C.C. Methylated Forms of Arsenic in the Environment. Science 182:
1247,1973.
8. Braman, R.S., Johnson, D.L, Foreback, C.C., Ammons, J.M. and Bricker, J.L. Separation and
Determination of Nanogram Amounts of Inorganic Arsenic and Methylarsenic Compounds. Anal.
CjTem.49:621,1977.
9. Johnson, D.L. and Braman, R.S. Alkyl and Inorganic Arsenic in Air Samples. Chromosphere 6:
333-338,1975.
10. Talmi, Y. and Bostick, D.T. Determination of Alkylarsenic Acids in Pesticide and Environmental
Samples by Gas Chromatography with a Microwave (etc.). Anal. Chem. 47:2145- 2150,1975.
11. Andrea, M.O. Determination of Trace Elements From Sewage Sludge Fertilizer in Soils and
Plants. Anal. Chem. 49: 820,1977.
12. Yamamoto, M. Determination of Arsenate, Methanearsonate, and Dimethylarsinate in Water and
Sediment Extracts. Soil Sci. Soc. Am. Proc. 39: 859-861,1975.
13. Grabinski, A.A. Determination of Arsenic (III), Arsenic (V), Monomethylarsonate, and
Dimethylarsinate by Ion-Exchange Chromatography with Flam (etc.) Anal. Chem. 53: 966-968,
1981.
14. Ricci, G.R., Shepard, L.S., Colovos, G. and Hester, N.E. Ion Chromatography with Atomic
Absorption Spectrometric Detection of Organic and Inorangic Arsenic Species. Anal. Chem. 53:
610-613,1981.
15. Brinckman, F.E., Blair, W.R., Hewett, K.L.and Iverson, W.P. Graphite Furnace Atomic Absorption
Spectrophotometers as Automated Element-Specific Detectors for High Pressure Liquid (etc.). >L
Chromatogr. Sci. 15: 493-403,1977.
16. Stockton, R.A. and Irgolic, K.J. The Hitachi Graphite Furnace- Zeeman Atomic Absorption
Spectrometer as an Automated Element- Specific Detector for High (etc.). Int. J. Environ. Anal.
Chem. 6:313-319. 1979.
17. Irgolic, K.J., Woolson, E.A., Stockton, R.A., Newman, R.D., Bottino, N.R., Zingaro, R.A., Kearney,
P.C.,tPyles, R.A., Maeda, S., McShane, W.J. and Cox, E.R. Characterization of arsenic
compounds formed by DaphniamagnaandTetraselmischuifrom inorganic arsenate.Jn: Environ.
Health Perspectives. Ch. 19.1977. pp. 61 -66. and references cited there.
-------
ARSENIC CHEMISTRY IN RELATION TO THE DISPOSAL AND STABILITY OF
METALLURGICAL EXTRACTION WASTES
Robert G. Robins
HydroMet Technologies Limited
31-45 Smith Street
Marrickville NSW 2204
Australia
Tel: 61 25171188
Fax: 61 25199468
The work which is summarised in this presentation was conducted mostly at the University of New
South Wales while theauthorwas Head of the Department of Mineral Processing and Extractive Metallurgy.
He is now Visiting Professor at the Centre for Minerals Engineering at the same university where the work
on arsenic stabilisation is being continued as one of the projects in an Australian Government sponsored
Co-operative Research Centre for Waste Management and Pollution Control. The project on Stabilisation
of Arsenic Wastes is costed at $1.85M AUD over the 3-year period commencing in 1992.
Most of the material used in this presentation is described in detail in the publications (1 -21) cited in
the references.
ARSENIC HYDROMETALLURGY (9-14)
The removal of arsenicfrom process solutions and effluents has been practised by the mineral process
industries for many years. Removal by existing hydrometallurgical techniques is adequate for present-day
product specifications but the stability of waste materials for long-term disposal will not meet the regulatory
requirements of the future. This presentation reviews briefly the aqueous inorganic chemistry of arsenic as
it relates to the hydrometailurgical methods that have been applied commercially for arsenic removal,
recovery, and disposal, as well as those techniques which have been used in the laboratory or otherwise
suggested as a means of eliminating or recovering arsenic from solution.
The various separation methods include: oxidation- reduction, precipitation and thermal precipitation,
coprecipitation, adsorption, electrolysis, solvent extraction, ion exchange, membrane separation, precipi-
tate flotation, ion flotation, and biological processes. The removal and disposal of arsenicfrom metallurgical
process streams will become a greater problem as minerals with much higher arsenic content are being
processed in the future.
PRECIPITATION OF ARSENIC (1-6,10,12,18)
The insolubility of certain inorganic arsenic(III) and arsenic(V) compounds is the basis of most
hydrometallurgical arsenic-removal processes and the insoluble product is often the disposed material.
Recently it has been shown that these disposal residues are unstable and will produce a leachate that
contains arsenic. "
The most common method of removing arsenic from aqueous process streams is to precipitate the
sulfide As?S3, or calcium arsenate or ferric arsenate. The sulfide has its lowest solubility below pH = 4 but
that solubility is significantly higher than has been generally accepted. The sulfide is not usually the form
disposed in residues as it is easily oxidised, however there have been attempts to use As2S3 in landfill in
which anaerobic conditions are maintained and also in cement cast admixes. Recent work on biological
formation of arsenic sulfides may have an application in treating process residues.
There are a number of calcium arsenates that can be precipitated from arsenic(V) solutions. At high
pH, obtained by lime addition in excess, arsenic concentrations can be reduced to <0.01 mg/L. Those
calcium arsenates which are precipitated at pH >8 are not stable with respect to the carbon dioxide in the
atmosphere converting them into calcium carbonate with release of arsenic to solution.
-------
Arsenic(V) can be precipitated from process solutions below about pH = 2 with iron(lll) to form ferric
arsenate, FeAsO4.2H2O (scorodite). At lowprecipitationtemperaturesthecompound is very small in particle
size (~ 10nm) and is X-ray amorphous. At temperatures above about 90 degrees C the precipitated
compound is crystalline (>100nm) and has a solubility about two orders of magnitude lower than the
amorphous material (particle size effect). Ferric arsenate exhibits incongruent solubility at pH > about 2 and
will convert very slowly to ferrihydrite, which initially has extremely small particle size (~5nm) and a
stoichiometry near Fe2O3.2H2O.
The ferrihydrite is in fact "arsenic-bearing ferrihydrite" since it strongly adsorbs the released arsenic,
and at least in the short term prevents high arsenic solution levels. Ferric arsenate is therefore not stable
in the neutral to high pH region and so is not suited for direct disposal. Ferric arsenate is also not stable in
alkaline cement cast admixes.
There are other metal arsenates, such as those of Fe(ll), Zn(ll), Cu(ll), and Pb(ll), which are less
soluble (and more stable) in the neutral pH region than the calcium arsenates or ferric arsenate, but these
have not been seriously considered as disposal forms. Iron(ll) arsenate should be investigated further in
this respect.
ADSORPTION OF ARSENIC ON FERRIHYDRITE (15, 21-23)
Much attention has been directed to the removal of arsenic from hydrometallurgical process solutions
and wastewaters by precipitation and coprecipitation with iron(lll). At high concentrations of iron(lll) and
arsenic(V) and at low pH, precipitation results in the formation of ferric arsenate, FeAsO .2H O. At low
concentrations of arsenic(V) and high iron(lll) concentrations the coprecipitation of arsenic with ferric
hydroxide occurs and is probably the most effective method of removal of arsenic from aqueous solutions
and leads to a solid phase that is stable at least for a year or so. The solid coprecipitate has been referred
to as "basic ferric arsenate" and in 1985 a controversy began as to whether the coprecipitated material was
in fact a compound of iron(lll) and arsenic(V) or simply an adsorptive binding of arsenic with ferric
oxyhydroxide. There was at that time sufficient evidence to support the latter contention, but the use of the
term "basic ferric arsenate" still exists and formulae such as FeAsO4.xFe(OH)3 are used.
Recently, EXAFS studies have shown that in these materials arsenic(V) is adsorbed to ferrihydrite as
a strongly bonded inner-sphere complex with either monodentate or bidentate attachment. Monodentate
attachment apparently occurs near the optimal pH 4 to 5 for adsorption.
The adsorption of arsenic(lll) on ferrihydrite has also been investigated and although it seems an
efficient process there is no evidence that the adsorbed species is arsenic(lll). It may be that during the
process, air oxidation of arsenic(lll) will occur with some ease, as has been shown in preliminary
experiments.
TESTING FOR LONG-TERM STABILITY
Testing methods for evaluating the stability of hazardous waste residues have been defined by the US
Environmental Protection Agency (EPA) in several "Background Document for Toxicity Characteristic
Leaching Procedure" publications. The test methods do not adequately assess the long-term stability of
arsenical residues. Improved test methods must be designed which also include a characterisation of
physical properties and chemical components (mineralogy) so that predictions of behaviour can be made.
REFERENCES
1. Robins, R.G. The Solubility of Metal Arsenates. Paper D-1- 3.1m Proceedings of the MMIJ-AIME
Joint Meeting. Tokyo, 1981. pp. 25-43.
2. Robins, R.G. The Solubility of Metal Arsenates. Metallurgical Transactions B12: pp. 103-109,
1981.
-------
3. Robins, R.G. The Solubility of Metal Arsenates II. Paper presented at TMS/AIME Annual
Conference, Dallas, TX. Feb. 1982.
4. Robins, R.G. and Tozawa, K. Arsenic Removal from Gold Processing Waste Waters: The
Potential Ineffectiveness of Lime. Can. Min. & Met. Bull. April: 171-174,1982.
5. Robins, R.G. The Stabilities of Arsenic(V) and Arsenic(lll) Compounds in Aqueous Metal
Extraction Systems. 1m K. Osseo-Assare and J.D. Miller (eds.), Hydrometallurgy, Research
Development and Plant Practice. TMS. Warrendale, PA, 1982. pp. 291-310.
6. Nishimura, T. Tozawa, K. and Robins, R.G. The Calcium- Arsenic-Water System. Paper JD-2-1.
1m Proceedings of MMIJ/Aus. IMM Joint Symp., Sendai, 1983. pp. 105-120.
7. Robins, R.G. The Stability of Arsenic in Gold Mine Processing Wastes, im V. Kudryk et al (eds.),
Precious Metals: Mining Extraction and Processing. AIME. Warrendale, PA, 1984. pp. 241-249.
8. Robins, R.G. The Stability of Barium Arsenate: Sheritt's Barium Arsenate Process. Metallurgical
Transactions B16:404-406,1985.
9. Robins, R.G. The Aqueous Chemistry of Arsenic in Relation to Hydrometallurgical Processes.
Paper No. 1. 1m Proceedings, Impurity Control and Disposal. CIM Annual Meeting, Vancouver,
Canada, August 1985. pp. 1-1 to 1-26.
10. Nishimura, T. Tozawa, K. Ito, C.T. and Robins, R.G. The Calcium-Arsenic-Water-Air System. Ibid.
Paper No. 2, pp. 2-1 to 2-19.
11. Lefebre, M.S. and Robins, R.G. Membrane Separations of Arsenic from Hydrometallurgical
Process Streams. Ibid. Paper No. 10 pp. 10-1 to 10-9.
12. Robins, R.G. The Solubility and Stability of Scorodite, FeAsO..2H2O; Discussion. Amer. Mineral.
72: 842-844, 1987.
13. Robins, R.G. The Solubility and Stability of Ferric Arsenate. Paper presented at TMS/AIME
Annual Meeting, Denver, CO. Feb. 1987.
14. Robins, R.G. Arsenic Hydrometallurgy. In: Reddy, Hendrix and Queneau (Eds.) Arsenic
Metallurgy Fundamentals and Applications. TMS. Warrendale, PA, 1987. pp. 215-247.
15. Robins, R.G. Huang, J.C.Y., Nishimura, T. and Khoe, G.H. The Adsorption of Arsenate Ion by
Ferric Hydroxide. Ibid. pp. 99-112.
16. Robins, R.G. and Glastras, M.V. The Precipitation of Arsenic from Aqueous Solution in Relation
to disposal from Hydrometallurgical Processes, in: Proceedings, Research and Development in
Extractive Metallurgy, Aus.l.M.M., May 1987.
17. Khoe, G.H. and Robins, R.G. The Complexation of Iron(lll) with Sulphate, Phosphate or Arsenate,
J. Chem. Soc., Dalton Trans., 1988: 2015-2021.
18. Robins, R.G. The Stability and Solubility of Ferric Arsenate: An Update. In: D.R. Gaskell (ed.),
EPD Congress '90. TMS. Warrendale, PA, 1990. pp. 93-104.
19. Khoe, G.H., Huang, J.C.Y. and Robins R.G. Precipitation Chemistry of the Aqueous Ferrous
Arsenate System. In: D.R. Gaskell (ed), EPD Congress '91. TMS. Warrendale, PA, 1991. pp. 103-
105.
20. Robins, R.G. and Javaweera L.D. Arsenic in Gold Processing. Mineral Processing and Extractive
Metallurgy Review 9: 255- 271.1992.
-------
21. Robins, R.G. Wong, P.L.M.Nishirnura,T.,Khoe,G.H.and Huang, J.C.Y.Basic Ferric Arsenat&s-Non
Existent, in: J.P. Hager(ed.), EPD Congress 1992. TMS. Warrendale, PA, 1991. pp. 31-39.
22. Waychunas, G.A. Rea, B.A. Fuller, C.C. and Davis, J.A. Surface Chemistry of Ferrihydrite: I EXAFS
Studies of the Geometry of Coprecipitated and Adsorbed Arsenate. Geochimica et Cosmochimica
Acta, to be published.
23. Fuller, C.C. Davis, J.A. Waychunas, G.A. and Rea, B.A. Surface Chemistry of Ferrihydrite: II Kinetics
of Arsenate Adsorption and Coprecipitation Geochimica et Cosmochimica Acta, to be published.
-------
ARSENIC: SUPPLY, DEMAND, AND THE ENVIRONMENT
J. Roger Loebenstein
US Bureau of Mines
8107th St. NW
Mail Stop 5208
Washington, DC 20241-0002
USA
Tel: (202) 501-9416
SUPPLY
Arsenic is recovered as a byproduct of processing certain complex ores that are mined mainly for
copper, lead, zinc, gold, and silver, and its supply is dependent to a large extent on the demand for these
metals. In addition, the supply of arsenic is affected by environmental constraints and the economics of
producing a relatively low-value byproduct from low-grade source materials. In general arsenic is regarded
as atroublesome impurity in smelting and refining and thus high-arsenic source materials may be penalized
at the smelter and refinery or avoided at the mine.
Arsenic trioxide (As2O3 or white arsenic), the most important commercial compound, is currently
produced principally in Belgium, Chile, China, France, Mexico, the Philippines, and the former USSR, as
shown in Table 1. Table 1 shows a country as a producer if that country produced at least 3,000 metric tons
in any of the years shown. Currently, the two largest producers are Chile and China. Both countries are
expected to increase production capacity progressively over the next few years.
TABLE 1. MAJOR PRODUCERS OF ARSENIC TRIOXIDE (Metric Tons)
Country 1971 1981 1991
Belgium6
Chile
China0
France
Mexico
Philippines
Sweden
U.S.S.R."
United States
—
8,023
11,483
™
17,450
7,150
9,000
3,000
-—
___
e5,200
6,517
6,900
7,750
7,800
3,000
7,000
10,000
3,000
4,960
5,000
2,500
7,000
Many changes have occurred in producing countries overthe last year. Historically Sweden has been
an important producer, but production was permanently shut down in 1991, reportedly for environmental
reasons. In addition, a Finnish producer ceased operations in 1991 because of low prices for arsenic
trioxide. In France one of the two major producers of arsenic trioxide from gold ores, Salsigne, declared
bankruptcy. The producers in Mexico and Belgium are under increased pressure to reduce production by
governments concerned with environmental problems associated with arsenic and its potential release to
the environment.
The result of many developed countries cutting back production has been to spur lesser developed
countries to increase production to fill the supply shortfall. For example, US imports of arsenic trioxide from
China have grown from nothing in 1989, to about 1,100 metric tons in 1990, and finally to about 7,100 metric
tons in 1991, representing over one-fourth of total US arsenic trioxide imports. Consumers report that the
Chinese arsenic is generally poor in quality, 97 to 98% pure, and contains a high degree of moisture and
-------
foreign material, such as rocks and pebbles. However, in future years, China will probably improve itsquality
control rather than risk losing its customers. China is believed to be the only country in the world that
produces arsenic trioxide directly from arsenic ores.
Chile was the second largest source of arsenic trioxide imports in 1991, providing slightly less than
one-quarter of total US imports. Cia. Minera San Jos6 Ltd. produced arsenic trioxide as a byproduct of gold-
copper ores roasted at its El Indio smelter. The company is expected to increase production when a new
roaster is added to its smelter, sometime in the next year.
For many years, until 1985, arsenic trioxide was produced in the US at ASARCO Incorporated's
smelter in Tacoma, WA. The area surrounding the smelter is currently a Superfund site being cleaned up
under Environmental Protection Agency (EPA) supervision.
Nearly 80% of ASARCO'S arsenic trioxide production in 1981 was derived from high-arsenic copper
concentrates imported from the Philippines. The rest of the arsenic was produced from domestic copper
concentrates. When ASARCO shut down in 1985, production was shifted from the US to the Philippines,
as shown in Table 1.
DEMAND
All arsenic consumed in the US in 1991 was derived from imported sources. The most important
arsenic chemical imported was arsenic trioxide, with minor amounts of arsenic acid and arsenic sulfide
imported as well. Demand for arsenic is shown in Table 2. As can be seen, the decline in agricultural
chemicals has been made up by the increase in wood preservative use.
TABLE 2. ESTIMATED US DEMAND FOR ARSENIC (Metric Tons, Arsenic Content1)
1971
1981
1991
Agricultural chemicals
(herbicides, desiccants) 15,600 8,900 5,000
Glass 2,000 1,000 900
Industrial chemicals
(wood preservatives) 970 9,100 14,300
Nonferrous alloys and electronics 570 600 1,000
Other 500 400 400
Total
19,640
20,000
21,600
Arsenic trioxide contains 75.7% arsenic by weight. ~
Source of data: U.S. Bureau of Mines, Arsenic chapter from Mineral Facts and Problems, 1980 and 1985
editions, and author's estimates for 1991.
Arsenic trioxide was imported and then converted to arsenic acid for use in the production of arsenical
wood preservatives by three major companies: Hickson Corp., CSI, and Osmose Corp.
Hickson International PLC, Castleford, UK, the parent company of Hickson Corp., has wood
preservative plants in Australia, New Zealand, the Republic of South Africa, and the UK, as well as in the
US.
LaPorte PLC, UK, is the parent company of CSI, Mineral Research and Development, and Rentokill.
CSI manufactures wood preservatives in the US, and Mineral Research and Development formulates
arsenical Pharmaceuticals for veterinary purposes in the US. Rentokill produces wood preservatives in
Liverpool, UK.
Only a few agricultural uses for arsenic remain. ISK Biotech (formerly Fermenta Corp.), Mentor, OH,
produced the arsenical herbicide monosodium methanearsonate (MSMA) at a plant in Houston, TX.
Atochem Corp., Bryan, TX, was a major producer of arsenic acid for use by cotton growers and wood
preservative companies. A company in Israel, Luxembourg Chemicals & Agriculture Ltd. .produced sodium
cacodylate, an arsenical used as a bait preparation against ants.
Minor amounts of arsenic acid are used by glass companies such as Corning Glass, Corning, NY.
Arsenic in the form of arsenic acid is used as a fining agent to disperse bubbles that tend to form when certain
types of glass are produced. As a result of EPA regulations that became effective in 1986, arsenic emissions
from glass furnaces have been reduced. As shown in Table 2, arsenic usage by the glass industry has
declined since 1971.
-------
Chromated copper arsenate (CCA) is the most common arsenic- based wood preservative. Other
arsenic-based wood preservatives include ammoniacal copper arsenate (ACA) and f luor chrome arsenate
phenol (FCAP). Arsenic trioxide is also used as a reagent in ore flotation.
During the 1930s and 1940s, an estimated 45,000 metric tons of arsenic-based insecticides were used
annuallyinthecompoundParisgreenCcopperacetoarseniteJandthearsenatesofcalcium.chromium.lead,
and sodium (1). For many years, until the late 1970s, agricultural chemicals were the most important end
use for arsenic, but in the future, becauseof environmental problems, agricultural chemical use will probably
decline to a negligible amount. The only two remaining major agricultural uses for arsenic are as herbicides
(MSMA and DSMA) for control of weeds and arsenic acid to aid in the mechanical harvesting of cotton in
Oklahoma and Texas. Depending on weather conditions during the cotton harvest, more or less arsenic acid
is used. An early frost tends to help dry cotton bolls and reduce arsenic acid consumption.
ENVIRONMENTAL REGULATION
Arsenic has long posed environmental problems because of its well known toxicity, and more recently
as a result of the Government's finding that inorganic arsenic is a human carcinogen. Most of the published
literature on arsenic emissions dates back to the mid-1970's when the EPA studied the US nonferrous
smelting industry and the arsenic industry in depth. In 1978, the Occupational Safety and Health
Administration (OSHA) promulgated the final standard on the occupational exposure to inorganic arsenic.
OSHA had concluded that inorganic arsenic was a carcinogen and that worker exposure had to be limited.
In 1980, EPA listed inorganic arsenic as a hazardous air pollutant, based on its findings that inorganic
arsenic was carcinogenic to humans, and thatthere was significant public exposure tothepoliutant.ini 983,
EPA estimated that over 85% of the 1,200 tons per year of atmospheric arsenic emissions came from copper
smelters and glass manufacturing plants. EPA identified other source categories for which standards were
not proposed: primary lead and zinc smelters, zinc oxide plants, arsenic chemical manufacturing plants,
cotton gins, and secondary lead smelters (2).
As a result of the necessity to comply with Federal and local regulations on atmospheric emissions of
sulfur dioxide and arsenic, ASARCO closed its copper smelter and associated arsenic recovery plant at
Tacoma, WA, in 1985. At that time, ASARCO was the only producer of arsenic in the US. Currently, there
is no arsenic produced in the US.
EPA regulates the uses of inorganic arsenic under provisions of the Federal Insecticide, Fungicide,
and Rodenticide Act (FIFRA) (3). A pesticide product may be sold or distributed in the US only if it is
registered or exempt from registration under FIFRA. Before a product can be registered as a pesticide, it
must be shown that it can be used without "unreasonable adverse effects on the environment", and without
causing "any unreasonable risk to man and the environment taking into account the economic, social and
environmental costs and benefits of the use of the pesticide". It is then the responsibility of the proponent
of initial or continued registration to prove that the pesticide meets the risk and benefit standard of FIFRA
above. If at anytime EPA determines that a pesticide does not meet this standard for registration, EPA may
cancel the registration under Section 6 of FIFRA.
In 1986, EPA issued its final rules on arsenic emissions from copper smelters and glass manufacturing
plants. In the following year, EPA issued its preliminary position to cancel most of the nonwood pesticide
uses for inorganic arsenicals. Included in the list were lead arsenate, calcium arsenate, sodium arsenite,
sodium arsenate, and arsenic trioxide.
Registrations for lead arsenate, calcium arsenate, and sodium arsenite have been voluntarily
canceled. The sole registrant of the growth regulator use of lead arsenate requested voluntary cancellation
in 1987; the registrant of products containing calcium arsenate requested voluntary cancellation in 1989;
and the registrant of products containing sodium arsenite requested voluntary cancellation in 1990 (4).
In 1991, EPA announced its preliminary decision to cancel the registration of products containing
arsenic acid used as a desiccant on cotton. EPA provided for a period of hearings before making its final
10
-------
decision (4). EPA estimated that current usage of arsenic acid is 2 to 3 million pounds per year, representing
5 to 10% of total US acreage (3). Using 2.5 million pounds as an estimate, this quantity represents a usage
of about 1,130 metrictonsof arsenicacid peryear (containing about 600 metrictons of arsenic). Subtracting
600 from the 5,000 metric tons for agricultural chemicals shown in Table 2 indicates that about 4,400 metric
tons of arsenic were used in herbicides in 1991.
ENVIRONMENTAL PROBLEMS
In January 1992, a Panamanian ship traveling from New York to Baltimore via the Delaware Bay
encountered a storm off the coast of Cape May, NJ, and lost four containers of arsenic trioxide weighing
a total of 75 metric tons. Subsequently, the Justice Department filed suit against the owner of the
Panamanian freighterto help pay forthe cleanup. Some $2.7 million in Superfund cleanup money has been
approved by EPA to help cover the costs of the extensive search and recovery of the arsenic- containing
drums. Total cost of recovery is estimated at about $5 million. To date, only some of the drums have been
recovered. The Commerce Department has prohibited fishing in the vicinity of the spill.
Joseph Brown, a toxicologist with California's Environmental Protection Agency in Berkeley, has
recommended that California's Department of Health Services lower its 50 parts per billion (ppb) drinking
water standard and establish an even lower regulatory goal of 2 parts per trillion arsenic (5). The 50 ppb is
the current Federal standard. A new riskassessment by California's EPA indicates that lifetime consumption
of drinking water with levels of arsenic at 50 ppb presents a one-in-100 risk of cancer.
In California, hazardous water has been found primarily in 200 deep wells in Kings, Kern, San Joaquin,
and Sonoma counties. The arsenic is believed to enterwater supplies from underground deposits in highly
porous rock where wells are often drilled, particularly in desert areas and near mountains.
ENVIRONMENTAL STUDIES
The Bureau of Mines is in the process of conducting a series of material balance studies which examine
all stages of the flow of materials: from extraction through processing, manufacture of products, use of
products, and finally disposal. In these studies, estimates are made of the amounts of material lost at several
points along the flow.
Figure 1 is a preliminary drawing of the US materials balance for arsenic in 1989. The figure is
separated between processing and materials in use in order to show that no arsenic is consumed from
domestic sources—only from imported sources.
Atmospheric emissions
800
Leach dumps, tailings, slag
9,200
Loss to enYironment
(Mostly from herbicides * desiccanu)
X ml
Fabrication losses
(Mostly from glass industry)
20
Figure 1. US materials balance for arsenic in 1989 (metric ton units).
11
-------
On the processing side, arsenic is discarded in leach dumps, tailings, slag, or emitted to the
atmosphere. As shown, most arsenic losses, about 10,000 metric tons per year, are emitted during the
processing of copper, zinc, and lead. On the use side, estimates are still being developed to determine the
quantity of arsenic dissipated to the environment and the quantity being disposed of in landfills and
municipal waste dumps, as indicated by theXand Yshown in Figure 1. In 1989, based on demand estimates
by the Bureau of Mines, about 4,900 metric tons of arsenic were used in herbicides and desiccants and lost
to the environment in the same year in which they were used.
The problem of estimating Y shown in Figure 1 is more difficult because the quantity of arsenic
disposed of in landfills depends on the useful life of the arsenical product. The useful life of treated lumber,
automobile batteries, and glass products that contain arsenic can vary from 10 to 30 years. Using 20 years
as an average, US demand forarsenic in 1969 (1989 less 20 years) was 16,300 metric tons for agricultural
uses and 4,300 metric tons for all other uses. Therefore, Y could be about 4,300 metric tons.
SUMMARY COMMENTS
More research is needed to find ways to remove arsenic from the environment or combine it in such
a way that it becomes less harmful to the environment. Arsenic is present throughout the Earth's crust and
can never be completely eliminated. The cost of reducing arsenic contamination in groundwater must be
weighed against its detrimental health effects.
REFERENCES
1. Will, R. ArsenicTrioxide and Arsenic Metal, im Chemical Economics Handbook, SRI International
Menlo Park, CA. April 1991,12 pp.
2. Edelstein, D. US Bureau of Mines. Other Metals, Ch. from the 1983 Minerals Yearbook.
3. U.S. Environmental Protection Agency, Office of Pesticides and Toxic Substances, Office of
Pesticide Programs. Inorganic Arsenicals, Technical Support Document, Oct. 1991, 53 pp.
4. Federal Register, US Environmental Protection Agency. Inorganic Arsenicals; Preliminary Deter-
mination to Cancel Registration of Pesticides Containing Inorganic Arsenicals Registered for Non-
wood Preservative Use: Availability of Technical Support Document; Notice of Intent to Cancel
Oct. 7,1991, pp. 50576-50585.
5. Arsenic in Water: Bigger Cancer Threat. Science News. April 18,1992. p. 253.
12
-------
RECOVERY OF ARSENIC AS A RAW MATERIAL FOR REUSE
(CASE STUDIES - MINING WASTES, FLUE DUSTS)
Robert D. Arsenault
Aminex Company
SE 80 Arkada Court
Shelton, WA 98584
USA
Tel: (206)427-8331
INTRODUCTION
Arsenic (As) may be recovered from a variety of byproduct sources and wastes to be used as a raw material
to manufacture the arsenical wood preservatives, chromated copper arsenate (CCA) and ammoniacal copper-
zinc arsenate (ACZA). The byproduct source may be crude, impure arsenic trioxide (As2O3) from smelter flue
dust piles or it may be slimes, sludges, and filter cakes from metal refineries, speisses from smelters, arsenate
solutions from manufacturing or arsenopyrite wastes and jig or float concentrates from mines. Currently,
though, most CCA is made from arsenic acid, which is made from various arsenic trioxides.
In some of these byproducts the value of the copper (Cu) content may be the economic incentive to
consider recovery of the Cu and As together for CCA and ACZA. In others the zinc (Zn), cobalt (Co), gold (Au),
silver (Ag), or other metal content may be the economic incentive. In each case the recovery method is one
designed for the specific raw material source and the contaminants to be recovered.
Crude As2O3 can be purified of contaminants and oxidized to arsenic acid, which is used directly to make
CCA or ACZA. In this case the contaminants may be antimony (Sb), selenium (Se), bismuth (Bi), Indium (In),
lead (Pb), Zn, iron (Fe), Au, Ag, sulfur (S) and copper (Cu). It may be desirable to recover some of these
contaminants for their value depending on their concentration in the raw material.
Sodium arsenite or arsenate concentrates and solutions can be used directly to make copper arsenate.
The arsenite is oxidized to arsenate with oxygen or hydrogen peroxide and copper sulfate is reacted in the
arsenate solution to make copper arsenate [Cu2AsO4»OH and Cu3(AsO4)2], which is solubilized with ammonia
for ACZA or solubilized with chromic acid for CCA. Care must be taken that undesirable contaminants are not
present in the alkaline arsenic solution that will cause CCA or ACZA sludges or corrosion of treating plant
equipment.
For example, a sodium arsenate solution byproduct from the manufacture of arsanilic acid and 3-nitro-4-
hydroxy phenyl arsonic acid, which was previously wasted by reacting with lime to produce calcium arsenate,
was converted to copper arsenate for the solubilization with acqueous ammonia for ACZA. However, the
sodium arsenate solution had excessive chloride in it from hydrochloric acid (HCI) used in the precipitation of
the 3-nitro feed additive product. The resulting copper arsenate paste for use in ammoniacal solution, even after
washing, had 0.48% Cl -, which would be a corrosion hazard for treating plant equipment and hardware on
treated wood. Tests of the chloride-contaminated copper arsenate, when used in formulation to make CCA 2%
treating solution, showed that 40 ppm Cl caused 0.2 mil/yr corrosion rate on carbon steel. This is unacceptable
corrosion. This is equivalent to 0.2% Cl in the CCA on a 100% basis.
Some of the available crude As2O3 has considerable Sb.
Other sources have considerable Fe. Both of
these elements cause CCA sludges and/orsurface residues on CCA-treated wood, and Fe causes sludges and/
or surface residue on ACZA treated wood. The maximum amount of Fe that can be tolerated in CCA solution
without a rapid increase in precipitation (1) is 75 ppm. The total of Fe and Pb should be limited to 100 ppm (1).
Therefore, the extraction and separation processes used for crude sources of arsenic and copper for making
CCA and ACZA must be tailored to avoid excessive amounts of contaminants in the arsenic acid or copper
arsenate final products.
The case studies which follow are examples of how As was recovered for industrial use and other examples
are given which show potential sources and purification routes.
13
-------
CASE STUDY 1 - EQUITY SILVER MINE, HOUSTON, BC, CANADA
This mine has chalcopyrite (CuFeS2) copper ore and tetrahedrite [(Cu, Fe)12Sb4S13] silver ore. Associated
with the tetrahedrite is tennantite [(Cu, Fe)12As4S13] The Cu ore and Ag ore were ground to 300 mesh and
leached with a sodium hydrosulfite (NaHS) sodium hydroxide (NaOH) solution (2). The leach solution contained
160 g/L NaOH and 120 g/L NaHS with 200 ppm of Cu catalyst as copper sulfate (CuSO4«5H2O). The pH was
adjusted to 8 or 9 with sulfuric acid (H2SO4). The pregnant leach solution contained 50 to 60 gl_l Sb and 10 to
15 g/L As.
After leaching, separating from the gangue, and recovering the Cu and Ag-laden filtrate, a trace of copper
quinone catalyst was added (3). Sulfur was added as sodium sulfide (Na2S), and pressure oxidation in an
autoclave with oxygen produced sodium pyroantimonate [NaSb(OH)6] and sodium pyroarsenate [NaAs(OH)e
(3)]. This autoclave with NaHS, NaOH, O2, Na2S, Na3SbS3 and Na3AsS3yielded a filtrate containing NaOH,
Na2SO2, Na2SO4, NaAs(OH)6, and Na2S2O3 The sodium pyroantimonate was insoluble and was easily filtered
off with good separation. This first product was sold.
The filtrate can then be acidified with sulfuric acid to a pH of 8 and copper sulfate added to react with the
NaH,AsO4»5H2Othat resulted from the NaAs(OH)6 in solution. The copper arsenate formed [Cu3(AsO4)2»4H2O
and Cu2AsO/OH] can be recovered as filter cake, washed and sold for making CCA and ACZA. Recovery of
As is usually greater than 95 to 98% by this method (4).
The filtrate can be recycled for subsequent autoclave charges until the sodium sulfate maximum solubility
is reached at approximately 32° C (transition point).
CASE STUDY 2 - TEXASGULF CANADA, TIMMONS, ONT, CANADA
The Zn hydrometallurgical process utilized a zinc sulfate solution for electrolysis called electowinning. Zn
electrolyte typically contained impurities of Cu, Co, Ni, and Cd. These are detrimental to the plating of Zn and
must be removed priorto electrolysis. These elements were removed by a hot copper sulfate/arsenic trioxide
/ zinc dust purification procedure. Apparently, zinc dust displaces Cu and As from solution which are thought
to precipitate as a metallic couple. Zinc dust ordinarily does not displace Co and Ni from solution, but in the
presence of the Cu/As couple such metals are quantitatively precipitated. The byproduct is a cement Cu cake
containing various amounts of Zn, Cd, Co, Ni, and As.
This cement Cu cake had greatly reduced market value because of its high As content. However, the As
and Cu could be recycled so that neither CuSO4«5H O nor As2O3 needed to be purchased. Instead, the Cu
and As were recovered as copperarsenate and recycled, and the Cu cake was marketable at higher prices. The
recovery was accomplished in four basic operations: (1) acid leaching, (2) cobalt removal, (3) caustic leach ,(4)
arsenic removal.
The zinc sulfate ore was conventionally roasted to form zinc oxide, and then leached with sulfuric acid to
form zinc sulfate solution. This feed solution typically contained 0.5 to 1.0 g/L Cu, 20 to 30 ppm Co, 1 to 2 ppm
Ni, and also some Cd. The feed solution was treated with the Zn dust and copper arsenate in order to precipitate
the Co, Cu, Ni, and Cd. The final Co levels must be less than 0.1 ppm in order to ensure sufficient purity of the
electrolytic Zn.
The precipitated cement Cu filter cake consisting of 20% solids was subjected to a 2% sulfuric acid leach
at 95 degrees C which solubilized the Zn, Co, and Cd but left the As in the residue. After about 2 hrs the acid
leach slurry was neutralized with NaOH to a pH of 3.5 to 4.0. This precipitated any Cu in solution. The filtrate
went to the Co removal circuit where 100% recovery was achieved using potassium permanganate and NaOH
for oxidizing the Co and other metals.
The filter cake containing As and Cu was leached with caustic to dissolve the As, using a temperature of
95 degrees and additional NaOH solution and vigorous addition of air to the slurry at a rate of 500 cu. ft. / min
/dry ton of acid leach residue. After about 6 to 8 hrs. approximately 95% As recovery was achieved as NaH2AsO4
•5H?O. The filter cake with high Cu content (approx. 65% Cu, 3% Zn, 1 % Cd, and less than 0.5% As) was then
sold to a smelter as an upgraded cement Cu cake. The filtrate, containing the As, was reduced in pH to about
8 with addition of H2SO4. Copper sulfate was added, the amount varying with the As content, and the slurry
maintained for about 2 hrs. A copper arsenate filter cake (approx. 25% Cu, 27% As) was recovered for use
in replacing the purchased arsenic trioxide which was formerly used in purifying electrolytic zinc solution.
14
-------
Arsenic recovery (75 tons recycled) was about 98%. The filtrate containing 20 to 30 gpl Na was recycled to the
leaching plant as reagent for jarosite precipitation.
Later it was discovered that zinc arsenate was co-precipitating with copper arsenate, so that while the As
in the filter cake was a constant 27%, the Cu percentage varied widely. It was also discovered that zinc arsenate
could be used as a purification agent, thus lowering the need for zinc dust and for operating at low pH levels.
CASE STUDY 3 - PURIFICATION OF ARSENIC TRIOXIDE - MT. ISA MINES LTD (MIM), AUSTRALIA
Crude arsenic trioxide from various smelter sources may be contaminated with Pb, Sb, Si, Se, Au, Ag, Co,
and Fe. Some typical examples are shown in Table 1.
Crude arsenic trioxide (As2O3) may be leached by hot water (5) or by ammoniacal solution, and the As
dissolves into solution and is recovered by drying while some metals that cause sludge in CCA and ACZA, such
as Fe, remain in the residue. Au can be recovered in the residue for further processing.
Hot water extraction has the advantage that Fe, Pb, Ag, and Si would not be dissolved from the crude As2O3
and would not cause sludges in CCA and ACZA solutions or surface residues on treated wood. The Ag and
Au can be recoveredfortheirvalue in the residues. Arsenic trioxide of greaterthan 99% purity can be produced.
In the hot water extraction process As2O3 is soluble to the extent of about 7.5% at 210 degrees F, but at 90
degrees F its solubility is 3%. Therefore, 4.5 Ibs of As2O3 would be purified for every 110 Ibs of water heated
to boiling and cooled.
Energy requirements are much less when aqueous ammonia is used to leach As2O3 from crude material.
However, certain metals, including Pb, Sb, Zn, Co and oxides or sulfates of Ag are soluble in ammoniacal
solution. Silver salts value cannot be recovered unless reduced with hydrogen gas (6) or other reducing agent
in ammoniacal solution. Pb will cause precipitates and sludges in CCA and ACZA. However, when these
contaminants are not present, ammoniacal solution is the most economical purification method. Fe and Au are
left behind in the residue and the Au can be recovered for its value.
The solubility of As O in aqueous ammonia at 70 degrees F increases from 3% at 0% NH3 to about 20%
at 6% NH3 and then decreases to 7% at 8% NH3concentration and 2% at 10% NH3. At 160 degrees F the solubility
of As2O3 is 55% at 9 to 10% N H3 and decreases on either side of this maximum. Therefore, hot ammonia solution
is a more energy efficient extraction method than hot water if the dissolved impurities can be tolerated.
CASE STUDY 4 - SHERRITT GORDAM MINES, LTD., FORT SASKATCHEWAN, ALBERTA , CANADA
Sherritt Gordan used a hydrometallurgical process for recovery of non-ferrous metals such as Ni, Co and
Ag Sulfuric acid hot solutions were the medium in which the metal ions were reduced to metallic form by a partial
pressure of hydrogen gas. Other reducing agents that can be used include sodium borohydride.
The hydrogen reduction process began with leaching the finely ground ore or filter cake with hot
concentrated sulfuric acid. The recommended leach temperature for dissolution of silver is 200 degrees C,
which is 130 degrees C below the boiling point of H2SO4. Ag precipitates as silver sulfate upon dilution with water
or by crystallization when leach liquor is cooled below 170 degrees C. During dissolution the evolved gas
consists of 98% SO and 2% H2. Though silver sulfate precipitates on cooling the sulfuric acid, the silver salt
contains entrained H2SO4 and is difficult to handle. Therefore, it is desirable for Ag recovery to dilute the leach
solution with water at a minimum of 66% dilution. This will avoid the necessity of cooling the system. Also,
dilution with water precludes the formation of a double ammoniacal -sulfate silver salt precipitate and a slower
rate of silver reduction.
Silver sulfate was dissolved in ammonia solution, forming soluble ammines. The maximum solubility
occurs at NHs/Ag molar ratio range of 0.5/1.0 to1.5/1.0. Increasing the temperature to 50 degrees C increases
silver salt solubility. This solubility behavior is similar to that of arsenic trioxide in ammoniacal solution.
Increasing the solute concentration above the maximum causes a dramatic lowering of the solution concentra-
tion of Ag. Hydrogen gas partial pressure of 5 to10 kg/cm2 and a temperature of 110 to 125 degrees C are the
recommended reduction conditions.
The leaching phase with concentrated H2SO4 causes As, Ag, Se, Cd, and Cu to dissolve. Other metals,
15
-------
N
CO
in p cq
o o c\i
CQ
CD CO
p T-;
T^ CD
O
in
O5
p cq
h-' T-'
to
O
CM
CO C\J C\J
CO T- O
OJ
C)
I
o
UJ
CO
cc
<
UJ
Q
DC
O
CO
r>
o
cc
z
I
UJ
UJ
UJ
UJ
$
o
O
OJ
CD
O
O
p
U3
UJ
UI
UJ
I-
01
o
cc
LU
Q.
UJ
X
o
£C
Q.
Q.
O)
<
o
o
CM
N ON
o t- o
c
o
o o
oJ
CD O
C
o 2
d o
^•^gCM £
CD O CD O
CO
cq
CD
cq
oi
CO
ci
CO
CO
£
CM T-_
CD O"
pop
co co' to'
S
m co
O CD -
0>
CO
2
o
-o
15
_o
o
o
§
3
£
CO
8
o
2
32
£3
g|
§ m a)
^2-|
1^-
OB O O
•= § «
= 1
111
CO CO'
c w S
° ol
8~ S
o "
o in 5
CM^-I
CO i— c-
c<»
0>
CO
LU
O
CC
1
CO
I
CO
cos
<
in
CQ O Q UJ Ll_
LO
-H-*
ax-
16
-------
Including St>, Pb, Ni, Sn, and Te become insoluble salts in H2SO4. Dilution and cooling with water to optimum
acid concentrations causes the dissolved salts to precipitate. Arsenic can be separated from the sulfuric acid
leach solution by treating it with hydrogen sulfide, forming arsenic trisuifide, As2S3, which is insoluble in water.
By filtering, it can be recovered and then leached with nitric acid to form a mixture of arsenic acid, H3AsO4 and
sulfuric acid, with NO2 as an off-gas. Alternatively, the dilute acid leach solution containing the soluble As and
any Se or Cd can be neutralized with caustic and oxidized with oxygen, forming sodium arsenate. This can then
be reacted with copper sulfate to form copper arsenate for use with CCA and ACZA.
Other metals are recovered by hydrogen reduction in ammoniacal solution. These metals include Cu in
ammonium carbonate solution (7), Cu in ammonium sulfate solution (8), nickel in ammonium carbonate system
and molybdenum in ammonia - ammonium sulfate system. Copper in sulfuric acid is also reduced by hydrogen
to the metal (9). Reduction of these metals in ammoniacal solution occurs at 150 to 200 degrees C and 20 to
30 kg/cm2of hydrogen partial pressure.
CASE STUDY 5 - EXTRACTION OF ARSENIC FROM SPEISSES AND OTHER SOLID SOLUTIONS BY
OXIDATIVE LEACHING
Various sludges, slimes, filter cakes, arsenopyrites and speisses from refineries, mines and smelters can
be leached with either acid or alkaline solutions to recover metals and arsenic. The metals may be extracted
more rapidly in an autoclave under oxygen pressure for oxidation of the metals. Examples of these methods
are given below.
1. Speiss containing approximately 60% Cu, 21 % As, 15% Pb, 1.8% Sb, 1.7% Fe, 1 % S, 0.65% Sn, 0.56%
Ag, 0.4% Ni, 0.2% Zn, 0.06% Se and 0.38 oz/ton Au was ground to -200 mesh and leached with concentrated
sulfuric acid at 100 degrees C under an oxygen partial pressure of 100 psi for 1 hour. Extraction of Cu and As
were 99.7% and 91.5% respectively. Both Pb and Ag were only slightly solubilized under these oxidizing
conditions, each averaging about 10 ppm in solution. The Cu and As were precipitated from the undiluted leach
solutions by neutralizing with sodium carbonate to pH 6. The product contained approximately 45% Cu, 16%
As, 0.5% Sb,1.9% Fe, 0.2 %Zn and smaller amounts of other elements. This material has approximately three
times too much Fe for CCA and ACZA raw materials. Additional processing by dissolving the product in
ammoniacal or caustic solution would precipitate out the Fe.
2. The speiss in Example 1 was leached with ammonia/ammonium carbonate solution with an exothermic
reaction at approximately 55° C under an oxygen partial pressure of 20 psi for 1 hour. Cu and As were extracted
at 99.3% and 90% efficiency. However, a considerable amount of Pb was solubilized , ranging from 3.8 to 6
gpl in the leach solution. This resulted in the Pb coprecipitating with As when the solution was boiled to expel
NH3 and CO2. The product had 44% Cu, 16% As, and 0.45% Pb. The Pb caused a precipitate in CCA solution
made with the product.
3. Sherritt Gordan Mines, Ltd., Fort Saskatchewan, Alberta uses ammoniacal solution leaching of Co, Cu
and Ni after pressure leaching of arsenopyrite with oxygen to convert sulfur to SO4 - As is released and oxidized.
Co and Cu are soluble in ammoniacal solutions. The ammonium sulfate byproduct is sold as a fertilizer. The
ammoniacal oxygen leach process was also used by International Nickel Co. of Canada Ltd. in the Port
Colborne, Ontario, and Sudbury Districts.
4. Kennecott Minerals Co. used a patented oxidizing sulfuric acid leach in a pressure autoclave to extract
Cu from flue dust. The dust contained 13.8% Cu, 11% As, and 17% Pb. After leaching, the dry residue contained
1.3% Cu, 12.5% As, 2.2% Bi, 0.4% Sb, 20% Pb as insoluble sulfate, 12.5% Fe, and 0.8% Mo. The Pb could
have been leached with NaCI brine, forming soluble lead chloride. The residue could have been leached with
caustic to extract the As from FeAsO4«2H2O/Fe3(AsO4)2»6H2O.
5. Sumitomo Metal Mining, Ltd. usesaprocess (21 patents)that incorporates pressure oxidation leaching
of nickel cobalt sulfide residue from a hydrometallurgical treatment process of laterite ore. The process utilized
continuous oxygen pressure with sulfuric acid leaching to produce cobalt sulfate and nickel sulfate in solution.
Versatic acid was used to solvent extract the metals by controlling pH with NH3 The metals are stripped from
the versatic acid with hydrochloric acid to give a solution of nickel and cobalt chlorides. Co extraction from a
nickel-cobalt chloride solution is performed in a mixer-settler using tri-n-octylamine. Co is completely extracted
and separated from the nickel and then stripped with water or dilute HCI to yield cobalt chloride.
17
-------
The NH3 used for pH control is recovered as (NH4)2SO , which is subjected to a double decomposition
reaction with Ca(OH)2 at high temperature to recover NH3 and gypsum.
(NH4)2S04 + Ca(OH)2 -> 2NH3 + CaSO4 + 2H2O
The NH3 is recovered and recycled to the extraction process.
This process, or variations of it, could be utilized for high As/Co ores in North America where gravity
concentrates contain 40 to 50% As, 10 to 15% Co and 1500 oz/ton Ag. Ores from mines near Cobalt, Ontario
such as Pan Continental, Agnico Eagle, Silverfields, Teldyne and Canadaka produce high As/Co/Ag concen-
trates. Combinations of oxygen/ammoniacal solution leaching would leach the As, Co and Ag from the
concentrates. Dilution with sulfuric acid and optimized pH control would cause formation of double ammine
sulfate salts, allowing selective As and Ag precipitation and Co filtration.
6. The U. S. Bureau of Mines utilized ferrous chloride-oxygen pressure leaching of a Co/As/Ag ore such
as the Canadaka Mine, Cobalt, Ontario concentrates. The concentrate contained 9.7% Co, 0.36% Cu, 2.8%
Ni, 45.5% As, 10.6% Fe, and 720 oz/ton Ag. The major mineral was cobaltite (CoAsS). During the ferrous
chloride-oxygen leach, Co, Cu, and Ni are solubilized as chlorides and As is converted to arsenate ion. Silver
was rejected to the leach residue as the insoluble chloride. It was subsequently extracted by cyanidation, though
it could have been extracted with acqueous NH3. Extractions of Co, Ni, and Cu were in the 97 to 98% range.
Silver extractions were 99% with a cyanide consumption of 4 Ib/ton of concentrate. During ferrous chloride-
oxygen treatment 23.5% of the As was solubilized. However, during this extraction the objective was to leave
the As insoluble. As was removed from process solutions by raising the pH to 2.1 with NaaCO3 and adding ferric
chloride, required to precipitate insoluble FeAsO4»2H2O (10).
The FeAsO4«2H2O precipitate contains Fe+3and As*5. It will pass the Environmental Protection Agency
(EPA) Toxicity Characteristic Leaching Procedure (TCLP) test for As leachability. However, this could be
further processed to produce copper arsenate for reuse in making CCA and ACZA. Hydrochloric acid will
solubilize FeAsO4»2H 2O , but removal of Fe and Cl from AsO4+3 would be difficult. FeAsO4»2 H2O could be
reduced to Fe (AsO4) 2«6H?O, with Fe+2 and As*5 with a reducing agent such as SO2 (11) and then solubilized
in ammoniacafcopper solution to precipitate out the Fe. TheCu3(AsO4)2»4H2O and Cu2AsO4«OH would remain
in the filtrate for subsequent recovery on evaporation and recovery of the ammonia.
A concentrate containing 45.6% As, 12% Bi, 9% Co, 4% Fe, 6.3% Ni, 2.8% S, and 11.9 oz/ton Ag was
oxidized with chloride-oxygen at 100 degrees C and 50 psi O2 for 24 hrs. About 25 to 33% of the As was
solubilized; and 99%of the Co and 97 to 98% of the Ni and 99% of the Ag was solubilized. However, the process
was designed to precipitate As as FeAsO4»2H2O by adding FeCI2»4H2O; and PbS was added to extract Bi from
the ore when extraction was 60% complete. When FeCI2»4H 2O concentration was doubled Bi extraction was
85% and As extraction was 92%. CuCI2could replace FeCI2 to produce copper arsenate for use in CCA and
ACZA manufacture.
CASE STUDY 6 - CONVERSION OF CALCIUM ARSENATE AND PYRITE WASTES TO COPPER
ARSENATE FOR CCA AND ACZA
Calcium arsenate is a residual waste material that is stored at several locations where there were Au and
Ag mining sites and Cu and Ag smelters years ago. At one site there is about 1500 to 2000 tons of calcium
arsenite/calcium arsenate mixture in a holding pond with a double layer liner. This material is 11 % As, 7% Sb,
15% Fe, 3% Pb and 32% Ca. Another BC mine has a calcium arsenate sludge contaminated with Sb from a
silver mining operation.
Calcium arsenate wastes can be leached with sulfuric acid to produce sodium arsenate and gypsum. This
has not been commercially done for the production of arsenicals for wood treating. However, if a copper
arsenate can be produced free from objectionable impurities this is a source of arsenic. Calcium will cause
sludges in both CCA and ACZA solutions. Sulfate will interfere in the chemistry of ACZA ammoniacal fixation
reactions with Cu, Zn, and As.
Since some of the As is likely in the form of arsenite, the calcium arsenate wastes should be slurried and
batch leached with sulfuric acid and overpressure of oxygen or air in order to oxidize the arsenite. A hot brine
leach would be necessary to remove the Pb. This would be followed by a recovery of the iron in a hot caustic
solution, causing the As to become available as sodium arsenate, from whence it can be reacted with copper
sulfate to make copper arsenates.
18
-------
The Getchell Mining property near Winnemuca, Humboldt County, NV has ponded arsenic sulfides that
were stored in the 1960s from an Au mining operation of arsenical ore containing the minerals orpiment and
realgar. The ore floatation concentrate contained 30 to 40% As. The sulfide material in the holding ponds
contains 3 to 3.5% As. The ponding of the wastes was done in the 1962 to 1963 period. This property at one
time had a roaster and arsenic trioxide was produced Another Au mining operation in South Dakota produces
hugequantitiesof arsenopyrite waste. Arsenopyrite (FeAsS) mineral wastesare present at Leadville, CO; Lead,
SD; Thompson Falls, MO; Polaris, ID; Eureka, NV; Big Creek, ID; Balmertown, Ontario, Canada; Roxbury, CN
and other mining sites. Glaucodot mineral [(Co,Fe)AsS], proustite (Ag3AsS3), cobaitite (CoAsS), smaltite
(CoAs2),skutterudite(CoAs3),erythrite[Co3(As04)2»8H20], annabergite [Ni3(AsO4)2-8H2O] all occuratCobalt,
Ontario, Canada. There are 29 known arsenical minerals. All of these wastes and minerals present the potential
of recovery of precious metals and As. However, the high sulfur content or the high calcium content of
arsenopyrites and calcium arsenates, respectively, make recovery impractical at this time. The possibility exists
for sulfur oxidizing bacteria such as Thiobacillus or sulfate production by Chromatia bacteria from the sulfide
ores. If this is feasible it would be an inexpensive recovery method. Otherwise, oxidative leaching would be
required.
SUMMARY
Each type of arsenic recovery system must be tailored to recover all of the valuable metals while producing
an arsenic oxide or copper arsenate that is free of objectionable levels of Fe, Pb, Sb, Cl, and Mn that might cause
corrosion or sludges in treating solutions and surface residues on the CCA and ACZA treated wood. Process
flow charts, major equipment characteristics, and operating parameters are available from the author and from
the cited references.
REFERENCES
1. Hartford, W. H. The practical chemistry of CCA in service, hr Proceedings of the Eighty-Second
Annual Meeting of the American-Wood Preservers' Association. 1986.82:28.
2 Sill, Metallurgical Resources, Inc. Improvements in process for treating complex ores. British Pat-
ent 834,253. May 4, 1960.
3. Vogt, J. W. Extraction of antimony from antimony sulfides bearing solids. U.S. Patent 4,096,232June
20,1978.
4. Freeman, G. M. and Dulson, J. E. Zinc hydrometallurgical process. U. S. Patent 4,049,514. Sept 20,
1977.
5. British Patent 368,316.
6. Kunda, W. Hydrometallurgical process for recovery of silver from silver bearing materials. Paper
presented at 1979 Third International Precious Metals Conference. Chicago, IL. May 8-10,1979.
7. Kunda, W. and Evans, D. J. I. Hydrogen reduction of copper to powder from ammine carbonate
system. im.Hausner, N. H. (ed.). Modern Development in Powder Metallurgy. Vol 4., Plenum Press.
1971. P. 49.
8. Evans, D.J.I., Romanchuk, S. and Mackiw, V. N. Copper powder by hydrogen reduction techniques.
Canadian Mining and Metallurgical Bulletin. 1961. 54(591): 523.
9. Genik-Sas, Berezowsky, R. M. Production of copper by gaseous reduction. U. S. Patent 4,018,595.
April 19, 1977.
10 Scheiner, B. J. and Lindstrom, R. E. Leaching Complex sulfide concentrates using aqueous chloride
oxidation systems. Colorado Mining Association Mining Yearbook. 1978. P.133.
11. Madsen, B. W., Dolezal, H. and Bloom, P. A. Processing arsenical flue dusts with sulfur dioxide and
sulfuric acid to produce arsenic trioxide. Paper presented at 1981 TMS/SME Hydrometallurgy
Chemical Processing Committee. AIME Annual meeting. Chicago, IL, Feb. 22-26,1981.
19
-------
ALTERNATIVE PRESERVATIVE SYSTEMS: PROS & CONS
H. M. Barnes and D. D. Nicholas
Mississippi Forest Products Utilization Laboratory
Department of Forest Products
School of Forest Resources
Mississippi State University
PO Drawer FP
Mississippi State, MS 39762-5724
USA
Tel: (601) 325-2116
INTRODUCTION
This paper provides an overview of the advantages and disadvantages of new generation wood
preservative systems and treating processes. Historically, the wood-preserving industry has used three major
preservative systems: creosote, pentachlorophenol, and waterborne inorganics. The oldest wood preservative
system, creosote, was initially used for the domestic treating industry in the 1870s when crossties were treated
for the L & N Railroad in Pascagoula, MS using the full-cell process developed by Bethell in 1838. Commercial
treating processes remained unchanged until the early 1900s with patents being issued for the Rueping and
Lowry empty-cell processes for use with oilborne systems.
The first patented waterborne system, acid copper chromate, was brought to the marketplace in 1929 and
was followed shortly by a new oilborne system, pentachlorophenol ("penta"), in 1931. Before the decade was
out, the remaining two major commercial waterborne systems, chromated copper arsenate (CCA) and
ammoniacal copper arsenate (ACA) were developed. Historical summaries can be found in the literature (1,2).
In recent years, changes in treatment technology and preservative systems in the US have arisen from
two principal factors: (1) environmental concerns, including promulgated air and water quality standards, and
the effect of treated wood on man and non-target organisms; and (2) the energy crisis, especially in regard to
oil and oil-based preservative systems. Of these two, environmental concerns predominate. Fourcriteria should
be used to judge candidate wood-preservative chemicals.
The first of these is safety. Wood preservatives should be safe to handle and use with respect to both
processing operations and final product consumption.
Additionally, these systems must be effective in order to provide the consumer with acceptable product
performance. In this regard, both the amount of biocide (retention) and the penetration of the bipcide
are critical.
Preservative permanence is an additional requirement for acceptable performance. All preservatives
will deplete from treated wood overtime, but acceptable systems do so at a slow rate. Both processing
and exposure parameters affect the permanence of preservative systems. p
Lastly, preservative systems must be cost-effective. Not only the raw material cost of the components
of a given system need to be considered, but also the energy and processing costs along with
environmental costs must be taken into account.
Competitive materials should be judged using the same criteria. Often, alternatives to treated wood
products appearenvironmentally benign, but when examined closelyyield high environmental and energy costs
for their production. The advantages for using a renewable natural resource should not be overlooked. It has
been estimated that the failure to control wood-destroying insects and fungi would require the additional cutting
of over 300 million acres of forests, a significant consideration in these days of concern about global warming.
Currently, over 336 million cubic feet of wood are treated with wood preservatives and fire retardants. As shown
in Figure 1, waterborne preservatives represent the largest component of this production.
20
-------
Figure 1. Volume of wood treated in 1990 by preservative type.
Pentachlorophenol
9%
Creosote Solutions
23%
Fire Retardanta
1%
Waterborne Salts
67%
ALTERNATIVE PRESERVATIVE SYSTEMS
In the past 15 years, a considerable amount of research activity has focused on the development of new
biocides for wood preservative systems. These new candidates include waterborne, oilborne, and aqueous
emulsion systems which generally exhibit lower mammalian toxicity and should pose less danger to the
environmentthanthetraditional systems. Two waterborne preservativesthatshowpromiseare dualcomponent
systems: ammoniacal copper/didecyldimethylammoniumchloride and cupramine/dimethyldithiocarbamate.
Borates also show promise, but their use is limited because they are highly susceptible to leaching.
Several new biocides have potential as oilborne preservative systems. These include an isothiazolone,
chlorothalonil, iodopropynylbutyl carbarnate, and propiconazole. In addition to these biocides, copper naphthenate
and oxine copper are reemerging as commercial wood preservatives. The environmental impact of these new
biocides varies but in general they pdse fewer problems. All of these biocides are effective wood preservatives
but have limitations. For example, iodopropynylbutyl carbarnate and propiconazole are effective only in above-
ground applications. Furthermore, although the isothiazolone and chlorothalonil are excellent broad spectrum
wood preservatives, they are less cost- effective than cu rrently used systems. The liability of new copper- based
systems to copper-tolerant fungi is also a concern.
Emulsion systems hold promise for the future. The advent of micro-emulsions may allovv for the
combination of otherwise dissimilar biocides into effective formulations which would require less preservative
for effectiveness. At the same time these systems eliminate the need for excessive amounts of petroleum based
carriers.
Research to date suggests that it is desirable to incorporate durable water repellents into preservative
formulations in order to improve the weathering and leaching performance of treated wood products. These
systems are essentially wax emulsion systems, and several commercial products are available for use with CCA
treatments. Other approaches, such as the incorporation of oil in CCA systems, also hold promise for better
product performance and reduced leaching rates. Experimental polymer systems have been researched as co-
treatments for CCA treatment with some success. It is anticipated that further progress in this area will be
forthcoming in the future.
ALTERNATIVE TREATING PROCESSES
The Bethell, orfull-cell, process was developed in 1838 by John Bethell to treat timbers for the Royal Navy
with creosote. The process utilizes an initial vacuum (22 in Hg or greater) to evacuate the air from the wood.
After evacuation, the treating cylinder is filled with preservative fluid and the pressure is increased toa maximum
dependant upon the species of wood being treated. Pressure is maintained until the wood cell lumens are
completely filled with preservative solution. This process is used primarily with waterborne preservatives such
as CCA and results in the maximum absorption of solution into the wood.
In orderto reduce the uptake of solution and minimize the dripping of preservative solution and subsequent
holding time after treatment, commercial treaters have turned to a modified full-cell process. In this process, a
short term (10 min), low intensity (nearly equal to 10 in Hg) vacuum is used in conjunction with a final vacuum
period after the pressure period. Wood treated using this modification typically exits the treating cylinder with
significantly lower moisture content.
21
-------
A modified Lowry process, designated the alternating pressure method (APM), has been used success-
fully in Australia for many years to treat partially seasoned pine with CCA preservatives. After introduction of
the preservative, pressure is quickly cycled between maximum pressure and atmospheric pressure. This
process has the advantage of being less energy intensive since the need for initial kiln-drying is eliminated.
In situ fixation of CCA has been achieved with the patented MSU Process developed by W. C. Kelso, Jr..
The process is a modification of empty-cell processing whereby the CCA components are fixed in the wood prior
to removal of the wood from the cylinder. Typically, 95 to 99% of the components are fixed yielding a product
with half of the water found with wood treated with a full-cell process. The process eliminates dripping and the
wood is safe to handle immediately upon exiting the cylinder.
Considerable research has been undertaken to accelerate the fixation of chromium and arsenic containing
preservatives. Rapid fixation schemes generally use one of four methods: hot air heating, hot water fixation,
steam fixation, or hot oil heating. Each method has its advantages and disadvantages. The key factors affecting
fixation are wood moisture content, temperature, concentration, and time. In general, moist wood is essential
to proper rapid fixation since the fixation reactions are essentially ionic. Failure to maintain a moist environment
can lead to excessive leaching of components.
Compared to fixation at ambient temperature, accelerated fixation generally proceeds at a rate one to three
times faster for each 10 degree C rise in temperature. Leaching rates for CCA components also depend upon
the concentration. There is evidence that the leaching of arsenic from CCA-treated wood in laboratory studies
is greater forconcentrations below commercial production norms. Presumably this is because the reaction sites
on the wood are filled by adsorbed copper leaving the arsenic in a less fixed state. At normal commercial
retentions, this behavior disappears. Care must be taken in interpreting small-scale laboratory results,
Implications from which may not be applicable to commodities treated in full production sizes. An excellent
discussion of accelerated fixation can be found in the literature (3).
DISPOSAL OF TREATED WOOD
Disposal of wood products treated with CCA, penta, and creosote is becoming a major environmental
issue. Restrictions on landfill disposal are increasing and will soon preclude this option. Therefore, other
methodsof disposal are being studied. In this regard, creosote-treated wood isthe easiest to dispose of because
it can be incinerated in commercial boilers without creating environmental problems. Penta-treated wood can
be incinerated, but scrubbers are needed and licensing is difficult. CCA-treated wood is also more troublesome,
and additional research will be required to develop cost-effective systems.
One approach to the disposal of organics has been to remove the metal hardware and shred the wood.
The wood is then placed in a reactor to remove organics and the residual wood is composted. With CCA,
research has focused on extraction/destruction of the wood followed by recovery and reuse of the preservative
components. Reuse of treated products in large sizes (e.g., poles) is being accomplished to a limited extent by
peeling and retreatment. This disposal problem will undoubtedly accelerate the acceptance of new preservative
systems. Data currently being generated should help develop total life cycle management costs fortreated wood
products.
FUTURETRENDS
Environmental concerns are likely to continue to drive the treated-wood industry. Advances in plant design,
process control, and sanitation continue to improve in order to meet existing and future regulations. The
development of new and/or modification of existing preservative systems will likely provide the industry with new
alternatives that will make it easier to comply with the ever changing environmental and safety regulations.
However, it will also require that the industry become more sophisticated because the new preservative systems
willprobably require much morestringentquality control measures in ordertoattain equivalent performance and
thus enable the industry to compete with alternate products.
The move on the part of the industry to a "use category" system for setting standards should allowfor more
specific preservatives for specif ic uses. The use of combination biocides has the potential for reducing the total
amount of preservative injected into wood, hence lowering potential environmental impact. Advancing treatment
technology, such as supercritical fluid treatment, may allow treaters of the future to put biocides into the wood
structure in such a way that they are more effective and less leachable and this will permit the use of lower
concentrations.
22
-------
REFERENCES
1. Barnes, H. M. Trends in the wood-treating industry: state-of- the-art report. Forest Prod. J. 35(1 ):13-
22.1968.
2. Barnes, H. M. Novel pressure processes for treating wood with preservatives.^ Proceedings, Tenth
Annual Meeting of the Canadian Wood Preserving Assoc. Canadian Wood Preserving Association,
Mississauga, Ontario, 1989. pp. 127-143.
3. Anderson, D. G. The accelerated fixation of chromated copper preservative treated wood. Proceed-
ings, American Wood- Preservers' Assoc. 86:129-151.1990.
23
-------
POTENTIAL FOR RECOVERY OF CCA FROM F035 WOOD PRESERVING
OPERATIONS
Tom Lewis III
Lewis Environmental Services, Inc
RJ Casey Industrial Park
Pittsburgh, PA 15233
USA
Tel: (412) 322-8100
Fax: (412) 322-8109
INTRODUCTION
Lewis Environmental Services, Inc., has developed the ENVIRO-CLEAN PROCESS: a patent
pending process forthe reclamation of metals such as chromium, copper, zinc, and cadmium from process
waste streams. This two-step process utilizes granular activated carbon and electrolytic recovery to
produce a saleable metallic byproduct. The process generates no sludge or hazardous waste and the
effluent meets Environmental Protection Agency (EPA) discharge limits based on metal levels. The process
has effectively demonstrated that it can treat a matrix of multiple metals in a single stream with positive
results. This process has treated waste sources from the wood preserving, metal finishing, and painting
industries.
Undera Small Business Pollution Prevention Grant, Lewis Environmental Services Inc., modified its
ENVIRO-CLEAN PROCESS as an overall approach to remediating hazardous soil contaminated with
Chromated Copper Arsenate (CCA) salts for a major producer of wood- preserving chemicals. Currently,
the producer offers a CCA- contaminated soil disposal service to its licensees. Shortly, the cost of disposing
of this material will increase dramatically, impacting the producer's ability to continue offering this service.
The new disposal technique will require stabilization of the contaminated soil with cement. This technique
is expensive and will rely on the availability of landfill space. Wood-preserving producers are investigating
new recovery approaches which will eliminate the disposal of hazardous CCA waste. It was positively
demonstrated that CCA-contaminated soil can be acid leached and upgraded to a nonhazardous level. The
acid leaching technique is simple and does not require elaborate equipment design or process conditions.
CCA PROCESS DESCRIPTION
The process consists of a counter-current acid leaching technique to contaminated CCA soil where
it would pass the EPA Leach Toxicity Test and be reclassif ied as nonhazardous waste. The volume of leach
acid with the recovered chromium and copper metals will be recycled back into the operation. Any rinse
waters required to clean the processed soil will be processed by the ENVIRO-CLEAN PROCESS. The
ENVIRO-CLEAN PROCESS operates in a very simple manner similarto a filter; the system consist of a tank,
pump, and two carbon filters. This system design and sizing makes it applicable to small and medium size
waste generators. Also the granular carbon is reusable and capable of recovering heavy metals over
numerous treatment cycles. The implementation of this combined process scheme would eliminate the
need to dispose of CCA-contaminated soil, which is a hazardous waste, as well as recover valuable metals
such as copper and chromium.
SOIL CHARACTERISTICS
A 5-galIon sample of CCA-contaminated soil was received from the producer for testing. The soil was
green/gray in color with pieces of wood and stones. The acid leached soil was free flowing and light gray
in color. Table 1 presents the analysis of the contaminated and the acid leached soils for chromium and
copper.
24
-------
TABLE 1. SOIL ANALYSIS VIA ACID LEACHING
tern
Plant Acid Leached
Soi1 Soi1
Removal
Chromium
(mg/KG)
Copper
(mg/KG)
6070.0
4370.0
58.6
221 .0
99.03
94.94
PROCESSED SOIL CHARACTERISTICS AND MANAGEMENT
All the customer soil samples (five samples) passed the Toxicity Characteristic Leaching Procedure
(TCLP) test. The two main leach constituents chromium and arsenic averaged 0.8 mg/kg and 0.9 mg/kg
respectively over the various TCLP tests. The leach acid analysis from a second stage sample contained
3330,13300, and 22990 mg/L of chromium, copper, and iron. Table 2 presents the TCLP test results from
a typical soil sample.
TABLE 2. TCLP TEST RESULTS OF ACID LEACHED SOIL1
Tests
Resu1ts (mg/1)
Arseni c
Chromi um
Copper
0.71
0. 22
0.77
'The leachate Information - TCLP: Sol id Material. 100% Solids
charged to the extractor, Initial pH 3... 5
ENVIRO-CLEAN PROCESSING RESULTS
The acid leach samples were water washed with one volume of water equal to the acid leach volumes.
These solutions contained moderate levels of chromium, copper, and arsenic; the resulting concentrations
of the combined wash waters were 41.4,94.8, and 3.0 mg/L respectively. There was consistent repeatability
of leaching based on TCLP test results. All samples passed based on the standardized leach procedure.
Wash water generated from the final cleaning of processed soil was treated through an activated
carbon system. The activated carbon system consisted of two filters each containing 180 grams of activated
carbon. About 2000 ml of wash water was treated at a f lowrate of 15 ml/min. The effluent was crystal clear.
The effluent contained on average 0.01, 0.01, and 0.22 mg/L respectively for chromium, copper, and
arsenic. Table 3 list the results of the carbon treatment test.
25
-------
TABLE 3. CARBON TREATMENT DATA
Sampl
feed
time
time
time
time
time
le
30
60
90
120
150
chromium(mg/ 1 )
mi
mi
mi
mi
mi
n
n
n
n
n
41
0
0
0
0
0
.40
.0
1
.02
.0
.0
.0
1
1
1
copper (mg/ 1 )
94.90
0.01
0
0
0
0
arsen
3
0
0
0
0
0
i c
.0
.2
.3
.2
. 1
. 1
(mg/)
2
2
7
4
ENVIRONMENTAL ADVANTAGES
Wood preserving operations gain the following environmental benefits by utilizing the ENVIRO-
CLEAN PROCESS:
1. It allows a wood treating site or chemical manufacturer to perform closed-loop recovery of
valuable chemicals and water.
2. The production and handling of CCA-treated wood products generates a hazardous waste, in the
form of CCA-contaminated soil. This soil is "treated" and reclaimed, eliminating the need to
dispose of this material.
3. Chromium, copper, and arsenic are recycled back into the process. The water used to wash the
soil is processed through the ENVIRO-CLEAN PROCESS to remove trace metals and the rinse
water is reused as a source of clean water.
FACILITY DESCRIPTION
Lewis Environmental Services, Inc, is located in Pittsburgh, PA and is a permitted waste recycler under
Pennsylvania's Beneficial Recycling and Reuse classification. We are in operation and able to receive liquid
waste or activated carbon utilized in our process for heavy metal recovery.
The wood preserving chemical producer is located in Georgia, and plans to install a soil cleaning
system based on our design. We are planning to conduct a pilot scale test in 1992 to verify process
parameters and scale-up factors.
26
-------
REUSE OF WOOD PRESERVATIVE THAT CONTAINS ARSENIC
William Joseph Baldwin
Hickson Corporation
3941 Bonsai Road
Conley, Georgia 30027
USA
Tel: (404) 362-3970
INTRODUCTION
Inorganic pentavalent arsenical compounds of varying formulations have been used as wood
preservatives in substantial quantities for over 50 years (1). Currently, there are five arsenical wood
preservative formulations listed in The American Wood-Preservers' Association Standards (2). These
include three chromated copper arsenate (CCA) formulations, i.e., Types A, B, and C, ammoniacal copper
arsenate and ammoniacal copper zinc arsenate. Of these, CCA Type C, which consists of 34.0% arsenic
(as As2Os), 18.5% copper (as CuO), and 47.5% hexavalent chromium (as CrO3), is the predominant
arsenical wood preservative used in the US today. CC A-treated wood is used in applications such as decks,
docks, foundation and marine piling, fences, and utility poles.
The manufacture and use of arsenical wood preservatives is heavily regulated. The formulation of
CCA is regulated by the Occupational Safety and Health Administration (OSHA); however, the Environmen-
tal Protection Agency (EPA) governs the use of arsenical preservatives and regulates stormwater runoff.
Moreover, EPA's effluent guidelines for arsenical plants prohibit the discharge of process wastewaters into
navigable waters.
Although hazardous waste generated at arsenical plants has been regulated since the initiation of the
program in 1980, EPA amended its regulations under the Resource Conservation and Recovery Act
(RCRA) in late 1990 by listing wastewaters, process residuals, preservative dripping, and spent preserva-
tives as hazardous. Furthermore these regulations, referred to as Subpart W, establish standards for
treating plant design, operation, inspections, and closure (3).
PROCESS DESCRIPTION AND FEED STREAM CHARACTERISTICS
CCA is shipped to the treating plant by tank truck as a 50% concentrate with the composition noted
previously and 50% water. Upon arrival at the plant, the preservative is transferred to a concentrate storage
tank. When needed, the arsenical preservative is diluted with waterto a 1 to 2% working ortreating solution
through a closed system metered mixing device and transferred to a work tank.
Arsenical preservatives are applied by pressure processes. Two methods, referred to as full cell and
empty cell, are available; however, in practice, full cell and a hybrid of the two referred to as modified full
cell are used commercially. The difference is that in a full cell treatment (as the name implies) the wood cells
are full of treating solution; in a modified full cell treatment there is less liquid in the wood.
In general, during arsenical treatment, untreated wood is placed on rail ortram cars which are winched
or pushed into the treating cylinder. The door is closed, a vacuum is applied to the cylinder to remove air
from the wood cells, and the treating solution is transferred through piping from the work tank to the treating
cylinder. Pressure is applied to force the treating solution into the wood after which the solution is returned
to the work (storage) tank for reuse. A final vacuum is pulled with all excess preservative returned to the
work tank, pressure is returned to atmospheric, the door is opened, and the treated wood is pulled from the
cylinder and placed on the drip pad.
In accordance with the regulations, existing pads must be assessed by an independent registered
professional engineer. This annual assessment must document items such as materials of construction,
slope, curb or berm, structural strength, and crack repair.
All solution dripping onto the pad, as well as washdown water and rainwater, flows to a collection sump
27
-------
from which it is transferred to the dilution water tank. It should be noted that at some plants this transfer
involves filtration, e.g., bag or sockfilters, to remove wood chips, dirt, labels, etc. fromthe liquid. The dilution
water is in turn blended with additional CCA concentrate to make fresh treating solution.
As noted previously, treatment with the CCA preservative requires the use of water as a diluent. As
such, CCA plants collect and use, in process, all dripping, rainwater, run on, and water used to clean the
drip pad. Contaminated process water is not discharged.
RESIDUAL CHARACTERISTICS AND MANAGEMENT
The amount of dripping, now classified as hazardous waste, collected from an arsenical plant varies
by the treating process previously described. Dripping is also a function of species, e.g., southern pine vs.
hem-fir; dimension, e.g., 2 x4 in. vs. 2 x 10 in. vs. 5/4 x6 in. etc.; temperature and moisture content. "Typical"
dripping can rangefromlessthanaguart (from tram dripping only) to 50to100gajlonsperchargefora rapid
cycle plant treating southern pine.
As mentioned previously, EPA's listing also applies to wastewatersthat come in contact with process
contaminants, e.g., rainwater that mixes with wood dripping on the pad. So, regardless of concentration,
rainwater collected on the drip pad is considered hazardous waste until it is reclaimed. Table 1 illustrates
the volume of rainwater that will be classified as hazardous at arsenical plants of varying size throughout
the US.
TABLE 1. HAZARDOUS WASTE GENERATION AT TREATING PLANTS DUE
TO SUBPART W
Plant Location
Southern Washington
Houston, TX
Mobile, AL
Atlanta, GA
Nashville, TN
Pittsburgh, PA
Richmond, VA
Chicago, IL
Average
Annual
Rainfall
36.80"
47.07"
64.64"
48.61"
48.49"
36.30"
44.07"
33.34"
Pad Size
75'x100' 100'x150 100'x250'
Annual Tons of Hazardous Waste
718
918
1260
948
946
708
859
650
1435
1836
2521
1896
1891
1416
1719
1300
2392
3060
4202
3160
3152
2360
2865
2167
It should be noted that prior to the Subpart W listing, the majority of treating plants were small quantity
generators, i.e., generated less than 1 ton of waste per month.
ENVIRONMENTAL ADVANTAGES
The arsenical wood preserving industry collected and reused process water, i.e., dripping and
rainwater, prior to the listing of such materials as hazardous in 1990. Nevertheless, this material is now
classified as hazardousand it is conservatively estimated that 400, OOOtons of "rainwater" are collected, i.e.,
generated, annually by the 429 arsenical plants in the US. Fortunately, this water can be used in process
and thus avoids disposal. Moreover, wood properly treated with arsenical preservatives has a service life
seven to ten times greater than untreated wood, thus conserving our forest resources.
28
-------
REFERENCES
1.
2.
3.
Hartford, W. Chemical and Physical Properties, in: D.D. Nicholas (ed.) Wood Deterioration and
Its Prevention by Preservative Treatments. Vol. 2. Preservatives and Preservative Systems
Syracuse University Press, Syracuse, New York, 1983. p. 37.
Book of Standards, American Wood-Preservers' Association, Woodstock, Maryland, 1991.
55 Federal Register 50450, December 6,1990.
29
-------
OSMOSE WATER PURIFICATION SYSTEM TO REMOVE CCA CONTAMINANTS
FROM WATER
A. A. (Gus) Staats
Osmose Wood Preserving, Inc.
PO Drawer O
Griffin, GA 30224
USA
Tel: (404) 228-8434
INTRODUCTION
The Osmose Water Purification System is manufactured by Zenon Environmental Systems exclu-
sively for Osmose Wood Preserving, Inc. and the wood treating industry. Osmose Wood Preserving, Inc.
is an American-based company which manufactures chromated copper arsenate (CCA) used in wood
preservation. Zenon Environmental Systems is a Canadian-based firm which specializes in watertreatment
systems. With the advent of Stormwater and other regulations which affect the wood treating industry,
Osmose went in search of a watertreatment system that could consistently produce clean water from water
contaminated with CCA.
Osmose located Zenon Environmental Systems, which developed the present system to remove CCA
from contaminated water. CCA is a waterborne wood treating preservative that is used in many markets
across the country. Watercontaminated by CCA can be cleaned to a state where the arsenic and chromium
values are at or below the Drinking Water Standard. In addition, it will produce a concentrate that can be
reused in the treating process. This state-of-the-art water treatment system can be used to purify surface
water such as ponds, streams, stormwater runoff or groundwater. The Osmose Water Purification System
has been successfully employed to clean up both surface water and groundwater.
PROCESS DESCRIPTION
The Osmose Water Purification System can be modified to meet the specific needs of each particular
waste stream of CCA contaminated water. A typical system would be composed of three components. The
first would be a prefilter; next would be the primary method of membrane separation; and the last phase of
the system would be an ion exchange polishing unit. The prefiltration stage will remove suspended solids
from the wastewater stream. The size of the filtration unit will depend on the suspended solid particle size.
Removal of the suspended solids is necessary to protect the reverse osmosis membrane system. The
reverse osmosis membrane system is the main portion of the treatment train which is utilized to remove
approximately 95% majority of the dissolved CCA from the wastewater. A high pressure process pump is
utilized by the RO unit to force wastewater across the surface of a selective reverse osmosis membrane.
The membrane selectively allows the passage of water molecules through and rejects dissolved contami-
nants.
Two exit streams result from this separation process. The first stream is called the RO concentrate
which is composed of membrane-rejected material and has the highest concentration of undesirable
contaminants. The second stream is called the RO permeate which is membrane-accepted material which
has a very low concentration of the contaminants. The final stage of the system that can be employed, if
necessary, is the ion exchange system which will remove most of the remaining metals from the RO
permeate. The ion exchange system is composed of anion and cation resins which remove negatively and
positively charged ions, respectively.
FEED STREAM CHARACTERISTICS
Table 1 is a summary of chemical analysis which was achieved during the actual use of the Osmose
Water Purification System during an accidental spill of CCA wood treating solution in Canada. As can be
seen in the data, the RO unit removed between 99.1 and 99.8 percent of the target contaminants. Table 2
gives the comparison of effluent quality between the waste influent, treated effluent, and Vancouver tap
water.
30
-------
TABLE 1. SUMMARY OF CHEMICAL ANALYSIS
Sample
Stream
RO "Feed
RO Concentrate
RO Permeate
Ion Exchange
Discharge Tank
RO % Removal
IX % Removal
Chromium
mg/L
40.1
90.8
0.3658
0.0030
0.0333
99.1%
100.0%
Copper
mg/L
5.15
22.4
0.0085
0.0043
0.0060
99.8%
99.9%
Arsenic
mg/L
24.4
57.7
0.0394
0.0251
0.0229
99.8%
99.9%
TOC
mg/L
33*
N/A
N/A
17
N/A
N/A
48%
*Estimated
RESIDUAL CHARACTERISTICS AND MANAGEMENT
The waste streams generated through this system are the permeate and the concentrate streams. The
concentrate stream can be recycled back into the RO, or if the appropriate percentage removal has been
achieved, it can be removed and utilized in a current wood treating plant as makeup water ordisposed. The
permeate stream, in most cases, can be discharged to a local POTW or used as makeup water in a wood
treating facility. The advantage of this system is that it produces very little hazardous waste (filterwash resin,
etc.).
ENVIRONMENTAL ADVANTAGES
As stated previously, there is very little hazardous waste produced when utilizing this system because
the separated waste water streams can be recycled back into a treating plant for use as makeup water.
Alternative systems require regeneration of resins, disposal of filters, or disposal of some type of chemical
binding agent. Depending on the feed stream, the RO membrane will usually last 2 to 3 years before a
thorough cleaning is necessary. The ion exchange resin beds are periodically flushed and regeneration is
not needed on a regular basis because of the reduced concentrations of metals which enter the ion
exchange system.
FACILITY DESCRIPTION
The Osmose Water Purification System can be permanently placed onsite or used as a completely
mobile emergency unit. The system is produced in 10,20,50 and 100 gallon per minute (gpm) units which
can be combined or upgraded as the need arises. The system is very versatile and has been proven in field
operations throughout the US and Canada.
31
-------
TABLE 2. EFFLUENT QUALITY COMPARISON
Element
Analyzed
Phosphorus
Arsenic
Lead
Cobalt
Barium
Chromium
Copper
Aluminum
Magnesium
Strontium
Molybdenum
Cadmium
Nickel
Manganese
Vanadium
Zinc
Iron
Calcium
Waste
Influent
mg/L
0.7
23.9
<0.005
<0.004
0.04
34.7
11.5
0.6
3.6
0.1
0.01
<0.001
<0.02
0.57
<0.001
1.2
0.79
29.3
Treated
Effluent
mg/L
<0.01
<0.04
<0.005
<0.004
<0.004
<0.001
<0.002
<0.01
<0.01
<0.002
<0.001
<0.001
<0.002
<0.001
<0.001
<0.002
<0.002
<0.01
Vancouver
Tap Water
mg/L
<0.01
<0.04
<0.005
<0.004
<0.004
<0.001
1.2
0.09
0.15
0.07
<0.001
<0.001
<0.002
0.01
<0.001
0.05
0.07
1.4
REFERENCES
1. Connell, P.J., and Marr, T.A. Emergency Response Spill Cleanup of Wood Treating Waste, July
6,1990.
2. Donison, H., Whittaker, H., Tremblay, J., and Dutton, R., Development and Demonstration of a
Mobile Reverse Osmosis Absorption Treatment System for Environmental Cleanups.
32
-------
THE CASHMAN AND OTHER HYDROMETALLURGICAL PROCESS TREATMENTS
OF POLYMETALLIC ARSENICAL DUSTS, SLUDGES, AND WASTES
Richard S. Kunter
Richard S. Kunter & Associates
25958 Genesee Trail Road, No. K324
Golden, Colorado 80401-5742
USA
Tel: (303) 526-1868
Fax:(303)526-1718
INTRODUCTION
Artech Systems, Inc. has developed a low pressure and temperature chloride leach process called the
"Cashman Process" to extract metals from arsenical flue dusts and residues and fix the arsenic in an
environmentally stable form as ferric arsenate. The process was pilot tested at Hazen Research in an
integrated plant including continuous recycle from August 1989 to October 1989 during which several tons
of flue dust were processed. Based on this pilot program, the process was deemed technically feasible and
produced commercially salable products. Residues from this pilot program were subjected to a long-term
stability test jointly designed by PTI environmental services and the US Environmental Protection Agency
(EPA).
Pressure oxidation particularly of sulfide ores containing arsenic such as practiced by Homestake
Mining Company at the Mclaughlin Mine in California also precipitates arsenic as ferric arsenate from
aqueous slurries acidified with sulfuric acid; however the pressure and temperatures requirements of the
process are higher than that required by strong chloride brine leaches such as the "Cashman Process".
Less base metals are soluble in sulfuric acid compared to hydrochloric and they therefore remain in the
residue.
The economic recovery of metals from ores, concentrates and byproducts having a significant arsenic
content represents a considerable challenge to the metallurgical industry.
A wholly pyrometallurgical process approach is becoming less acceptable because of ever more
stringent regulatory and environmental constraints. In a copper smelter, for example, attempts to keep flue
dust in closed circuit result in the buildup of impurities which cannot be tolerated because of the deleterious
effects of arsenic, bismuth, antimony, etc. on copper quality. In these smelters such problems can be
overcome by utilizing a hydrometallurgical process to treat a portion or all of the resulting flue dust. This
would give the pyrometallurgical facility the advantage of lowering or eliminating the flue dust recirculating
load, thereby increasing the facility capacity, and at the same time allowing the facility to process feeds with
a greater arsenic content. However, in instances of treating arsenic rich feeds, a wholly hydrometallurgical
approach may offer the cleanest solution.
In the selection of a hydrometallurgical process, one of the most important concerns should be the
long-term stability of the arsenic-containing residue. Ever-changing environmental standards should
convince the smelting industry that the superior environmental performance of a process has great
advantages, both politically and operationally. The arsenical residue disposal problem will grow since many
of the future sources of the common base metals (copper, lead, zinc), as well as gold, have arsenic contents
that effectively preclude their processing by conventional pyrometallurgical methods alone.
The fume/flue dust or electrostatic precipitator dust that is recovered from the off-gases of many
smelter and roasting operations typically contains arsenic and metals of value, and is often considered
hazardous. In some instances, these hazardous flue dusts have been disposed of without regard to their
hazardous nature. Because the arsenic component of the flue dust is in a chemically unstable/soluble form,
weathering of such flue dust dumps can lead to the contamination of local groundwater systems.
33
-------
The metallurgical industry is faced with two problems in the treatment of arsenic containing feeds. In
addition to the development of economic extraction processes for metals, the industry must also develop
processes which address ever-increasing environmental concerns. From a combined economic and
engineering point of view, there is a great incentive to solve both problems simultaneously. This means
developing flow sheets that allow for the economic recovery of the contained metals as well as the
stabilization of the arsenic in a residue that will meet current and future environmental regulations.
Hydrometallurgical processes such as the Cashman Process and pressure oxidation provide the necessary
solutions in many cases.
THE CASHMAN PROCESS
The Cashman Process was initially developed to treat arsenic-containing base metal sulf ide concen-
trates. Subsequent work showed that by the choice of appropriate operating parameters, the basic
principles of the process could be applied to effectively treat arsenical flue dusts.
The Cashman Process is a hydrometallurgical process utilizing a hydrochloric acid leach under
somewhat elevated temperature and pressures which solubilizes metals of economic interest such as
copper, zinc, lead, silver, gold, nickel, cobalt, and bismuth and simultaneously leaves toxic elements such
as arsenic in an environmentally stable leach residue. The Cashman Process uses a patented hydrochloric ,
acid leach process (Patent No. 4,655,829) developed by CSS Management Corporation in Skykomish,
Washington. Artech is the licensee of the Cashman Process and is currently conducting efforts to identify
and develop commercial applications for the Cashman Process.
To bring about the oxidation of arsenic(III) to arsenic(V) necessary for the precipitation reactions as
well as the oxidation of the base metal sulfides and arsenopyrite, oxygen (sometimes as air) is used. The
effectiveness of the Cashman Process depends largely on the leach chemistry as well as efficiency of
oxygen dispersion throughout the reaction slurry, which is achieved by use of a low temperature, low
pressure autoclave with an efficient mixing and gas injection system.
Materials of construction are always a matter of concern when developing design engineering and
operating criteria. When calcium chloride is used as one of the reactants in the Cashman Process, the
pregnant leach slurry contains solubilized base metal chlorides which are corrosive toward standard
materialsof construction. Forthese reactions, acid-resistant brick-lined autoclaves are required. Ancillaries
such as piping and pumps must also be corrosion resistant. Although materials of construction are a
concern, reliable components are readily available.
Reagent additions are determined by the chemistry and mineralogy of the feed. Operating conditions
for each feed type are typically of lowtemperature and pressure. The reactor is heated indirectly with steam.
Reaction time depends on grind size, pulp density, and mineralogy and is typically 30 to 60 minutes for a
flue dust that contains only small quantities of sulfides.
STABILITY OF ARSENIC-CONTAINING SOLIDS
The chemistry of arsenic is complex and much of the early work on the composition and stability of
arsenic-containing solids precipitated from solution is difficult to interpret. In most hydrometallurgical
processes, the arsenic is precipitated as ferric arsenate (FeAsO4). Recent studies have indicated that in
orderto minimize the solubility of the arsenic in the precipitate, the Fe:As cation ratio in the precipitate should
be greater than the inferred stoichiometry of 1:1. This means that excess ferric ions and other cations are
co-precipitated with the ferric arsenate. In addition, the studies indicate the temperature and pH (acidity)
of precipitation should be controlled to produce a crystalline ferric arsenate product.
These conditions are difficult to maintain in typical commercial hydrometallurgical operations. Many
processes designed to recover valuable base metals from arsenic rich sources are also designed to
minimize iron dissolution because the recovery of base metals from iron rich solutions is often difficult. This
apparent dissolution selectivity is achieved by either actually preventing the iron minerals from dissolving,
or causing them to dissolve simultaneously and reprecipitate in a form other than ferric arsenate. This
results in an arsenic compound that is unstable and has relatively high solubility.
34
-------
In summary, most of the proposed hydrometallurgical processes for treating arsenic-containing feeds
result in residues that are not chemically stable and not of low solubility because (a) the acidity (pH) regime
in which the arsenic compound was formed is unsatisfactory and/or (b) there is insufficient cation
concentration in solution to precipitate the arsenate anion "completely".
In order to meet environmental disposal requirements, other processes may require additional
processing stages, either to re- treat the leach residue, and/or to precipitate the soluble arsenic from the
leach slurry before disposal or treatment for recovery of valuable metals. The Cashman Process obviates
the need for these additional stages because it permits the precipitation of all the arsenic in the feed in a
stable non-soluble form in a single unit operation. Furthermore, Artech is successful at producing residues
that better the toxicity tests using stoichiometric iron to arsenic ratios of 1.2 :1. Selected samples of the
leach residues have been chemically and mineralogically characterized and shown to be mixtures of
predominantly crystalline ferric arsenate and gypsum.
CASHMAN PROCESS DESCRIPTION
In base metal smelters, dusts emanating from roasting, smelting and converting operations typically
contain arsenic, bismuth, cadmium and antimony as process impurities as well as valuable base and noble
metals. The base metals are present in the fume/flue dust as complex mixtures of oxides, ferrites, etc. while
the arsenic is typically present as arsenic(l 11) oxide. In order to dissolve the base metals from the flue dust,
during the Cashman Process a reaction slurry of about 20% solids is made acidic. If the flue dust has a
relatively low iron content, an additional iron source is required. Calcium chloride is used so that the base
metals are dissolved as chloro complexes. The arsenic is co-precipitated primarily as ferric arsenate but
also as a mixture of copper/iron and other complex metal arsenates with gypsum (hydrated calcium sulfate)
which are very stable.
Implementing the Cashman Process at an existing smelter site is generalized as follows: After solid/
liquid separation of the leached slurry, the pregnant leach liquor can either be processed by a combination
of selective precipitation, solvent extraction/ion exchange and electrowinning techniques to recover the
dissolved base metals separately, or in bulk form as a mixed oxide / hydroxide / carbonate base metal
precipitate which can be recycled to the smelting furnace. This proprietary conversion reaction (called the
metathesis reaction) has been developed by Artech to convert chlorides to oxide products, and to recycle
the chloride in the Cashman Process, thereby substantially eliminating residual chloride content in the
products which might cause corrosion damage in subsequent unit operations.
Generally a plant utilizing the Cashman Process would include the following:
Feed preparation
Autoclave leaching, filtration, and disposal of process residue
Metals recovery
Barren stream recycle
FEED PREPARATION
Mineralogical and metallurgical test work are used to identify the optimum reagent additions required
to provide maximum metals recovery and maximum environmental performance. Most feed constituents of
less than 85 mesh generally do not require grinding, although screening and wet grinding can be utilized,
where necessary. The material to be treated is slurried with recycled calcium chloride brine, sulfuric acid,
and a source of iron (if sufficient iron is not contained in the waste) and is then introduced into the autoclave.
AUTOCLAVE LEACHING, FILTRATION, AND DISPOSAL OF RESIDUE
Materialsare leached in an autoclave undersomewhat elevated temperature and pressure conditions
for 30 to 120 minutes. The hot leached slurry is then pressure filtered to produce a solid residue and a leach
liquor. The solid residue generally contains sulfates, iron oxides, silica, sulfides, and ferric arsenate. The
leach liquor may contain dissolved metals such as copper, lead, zinc, silver, mercury, bismuth, cadmium,
35
-------
and gold. The solid residue is washed with water priorto disposal. Metals recovery operations are applied
to the filtered leach liquor and the barren liquor containing the chloride is then recycled.
METAL EXTRACTIONS
The metal extractions for the same eight feedstocks are shown in Table 1.
TABLE 1. CASHMAN PROCESS METAL EXTRACTIONS FOR EIGHT FEEDSTOCKS (1)
Material
Feed #
Feed #
Feed #
Feed #
Feed #
Feed #
Feed #
Feed #
1
2
3
4
5
6
7
8
Gold
93.75%
N/D
95.78%
69.53%
N/D
90.37%
49.63%
64.38%
Silver
N/D
93.10%
N/D
74.03%
N/D
72.05%
88.59%
96.77%
Copper
80.
40%
N/D
89.
90.
94.
90.
96.
94.
11%
71%
45%
56%
44%
41%
Lead
47
61
46%
57%
N/D
80
42
45
86
61
91%
35%
40%
27%
74%
Zinc
97
98
74
81
83
56
90
97
84%
11%
89%
98%
34%
78%
59%
12%
Bismuth
N/D
N/D
N/D
N/D
29.95%
98.30%
94.71%
92.63%
Cadmium
98.78%
N/D
95.37%
99.09%
98.85%
91.44%
98.47%
99.37%
N/D = Not Determined
FEED STREAM ANALYSES
The analyses of eight feedstocks and residues processed by the Cashman Process are presented in
Table 2.
TABLE 2. ANALYTICAL RESULTS FOR CASHMAN PROCESS RESIDUES AND UNPROC-
ESSED FEEDS (1)
Material
Feed # 1
Residue # 1
Feed # 2
Residue # 2
Feed # 3
Residue # 3
Feed # 4
Residue # 4
Feed # 5
Residue # 5
Feed # 6
Residue # 6
Feed # 7
Residue # 7
Feed # 8
Residue # 8*
11
5
11
7
5
4
2
3
1
1
7
8
2
2
22
IS
As
.6200
.4200
.2500
.9400
.3400
.9200
.9650
.1300
.4150
.0100
.7700
.5750
.0600
.0800
.9000
.9000
Ag
0. 0110
0.0010
0.0160
0.0010
0.0190
0.0010
0.0090
0.0030
0.0061
0.0017
0.0163
0.0022
0.0160
n/a
0.0603
0.0015
Analytical Results
Ba Cd
n/a
n/a
n/a
n/a
n/a
n/a
0.0540
n/a
n/a
n/a
0.1792
n/a
n/a
n/a
n/a
n/a
0
0
0
0
1
0
0
0
0
0
0
0
2
0
.5340
.0065
n/a
n/a
.1656
.0070
.0050
.0100
.1860
.0160
.1660
.0030
.1560
.0017
.9500
.0140
Cr
0.0160
0.0106
so. ooi
n/a
0.0115
0.0133
<0.001
n/a
: 0.0069
i n/a
0.0110
n/a
0.0125
^ 0.0118
: n/a
n/a
Hg
0.0040
0.0001
<0.001
<0.001
<0.001
0.0002
0. 0001
0. 0001
0.0001
n/a
0.0003
n/a
0.0012
0. 0030
<0. 0001
n/a
2.
1.
44.
14.
1.
1.
23.
4.
0.
0.
" 9.
1.
1.
1.
25.
7.
Pb
6800
4080
5000
9500
8980
8960
7000
9500
4920
2700
5600
5500
9700
0740
9000
4900
Se
n/a
n/a
n/a
n/a
n/a
n/a
0.0004
n/a
0.0001
n/a
0. 0002
n/a
n/a
n/a
n/a
n/a
n/a = Not Analyzed
36
-------
MET ALS RECOVERY
The best method of recovering metals from the leach liquor will vary, depending upon the quantity and
the type of metals present. Thef ollowing discussion first focuses on the most likely metal recovery methods,
but also identifies alternatives developed for Anaconda flue dust which could be employed, if necessary.
Not all flue dusts or residues would necessarily have present all of the metals cited in the tables in this
abstract or alternatively may have other metals not mentioned present.
The Cashman Process can be utilized at existing smelting facilities to eliminate hazardous flue dust
from the smelting process. Treating the flue dust would generate two products: a non-hazardous residue
primarily consisting of CaSO4 (calcium sulfate), SiO2 (silica) and FeAsO4 • 2H20 (ferric arsenate or
scorodite); and a mixed metal oxide/carbonate precipitate that could be recycled to the smelter feed to
combine with and, in part, replace the fluxing components.
Artech has piloted a proprietary elevated temperature and pressure precipitation procedure (metathe-
sis) which allows the precipitation of metal oxides and hydroxides from a chloride solution. Using this
procedure to precipitate all the metals solubilized during the leaching stage produces a precipitate
consisting of metal oxides, metal hydroxides, metal carbonates, and CaO. This precipitate could then be
either recycled to the smelting furnaces or directed to other metal recovery operations at the facility. For
instance, the precipitate could be digested with sulfuric acid and directed to existing solvent extraction
circuits.
Continuous pilot-scale tests on the treatment of the flue dusts have confirmed flow sheets for
producing products such as copper oxide or electrolytic copper, zinc oxide, bismuth oxychloride, lead
sulfate, silver cement, cadmium sponge, and mercury. Both batch and continuous production have been
piloted for Artech's proprietary metal oxychloride to metal oxide conversion reaction (metathesis).
Bechtel Corporation has completed an engineering design of a Cashman Process plant for a
Superfund flue dust project incorporating many of the metal recovery steps discussed in this abstract. The
Bechtel report states that the "process plant is designed to be zero discharge and has no water effluent".
RESIDUE CHARACTERISTICS
The toxicity results for the eight feedstocks illustrated in this abstract are shown in Table 3.
TABLE 3. TOXICITY RESULTS FOR CASHMAN PROCESS RESIDUES AND FEEDS (1)
Material
Peed # 1
Residue #
Peed # 2
Residue #
Peed # 3
Residue #
Feed # 4
Residue #
Peed # 5
Residue #
Peed # 6
Residue #
Peed # 7
Residue #
Feed # 8
Residue #
Type of
Test
EP TOX
1 EP TOX
EP TOX
2 EP TOX
EP TOX
3 EP TOX
EP TOX
4 EP TOX
EP TOX
S EP TOX
TCLP
6 TCLP
TCLP
7 TCLP
TCLP
8 * TCLP
Extract
Number As
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
2
1
1
1
1
1
52
0
23
0
28
0
0
0
0
3
0
0
0000
1300
0000
3500
0000
0030
6000
0400
3500
6000
0099
0500
1,300. 00
1
200
2,100.00
1
.100
0
0
0
0
<0
0
0
0
0
0
0
0
0
0
0
0
Ag
0090
0500
0050
0200
001
0350
0100
0400
0010
0001
0300
0100
0190
0100
.0090
.0100
Toxicity Results (ppra)
Ba Cd Cr
<0
<0
0
0
<0
<0
0
0
<0
<0
0
0
0
0
0
0
001
001
0100
0050
001
001
4000
0100
0001
0001
0990
0900
0190
0100
.0600
.0100
11.4000
0.0700
26.0000
0.8000
33.6900
0.4870
10.000
0.0400
33.2200
0.0660
65.000
0.160
36.600
0. 030
41.400
0.169
0.0070
0.0660
0.0050
0.0050
0.0040
0.0210
<0.001
0.0200
0.0280
0.0230
0.0400
0.0400
0.2000
0.0100
0.0090
0.0100
Hg
0.0010
0.0030
0.0004
0.0006
<0.001
0.0001
0.0004
0.0005
<0. 0001
<0.0001
0.0001
0.0001
0.0020
0.0006
0.0004
0.0009
Pb
0.
2
1
2
2
1
7
2
53
8000<0
7000<0
28000
94000.
2000<0
5000<0
00000.
38000.
40000.
1.3000<0.
0
0
1
1
9
15
67000.
14000.
69000.
56000.
.28000.
.90000.
3s
. 001
.001
0005
0340
.001
.001
0009
0001
0001
0001
3500
0350
2000
0560
0580
1000
* This sample is in the process testing stage as of 2/25/91 (only two scoping tests performed to
date).
NOTE: The EPA has identified feed # 5 as a K064 waste
N/A = Not Applicable
37
-------
LONG-TERM STABILITY TESTING
Independent test work conducted in 1989 and 1990 by PTI Environmental Services (2) of Boulder,
Colorado to simulate the long term stability of leach residues derived from treating copper smelter flue dusts
by the Cashman Process substantiate the longterm stability claimsof the Cashman Process leach residues.
The EPA approved the procedures used by the independent test firm for this test work. In the test work,
arsenic and cadmium in aqueous solution were extracted from leach columns and their concentrations were
monitored over several months. The report on these tests concluded that it would take 4 to 8 million years
to gradually leach all of the arsenic from a 3-meterthickpile of Cashman Process residues openly exposed
to the environment. Additionally, the suitability of several substrates to mitigate any solubilized arsenic and
cadmium were investigated. Some of these tests demonstrated that the leachate which passed through the
substrate passed drinking water quality standards for cadmium and arsenic. The long term stability test
information from the above report provides a basis for making a technically appropriate choice of disposal
sites.
ADVANTAGES OF THE CASHMAN PROCESS LEACH OVER OTHER PRESSURE LEACH
SYSTEMS
• Oxygen partial pressure is not the rate determining factor in the oxidation of iron and arsenic, but
oxygen dispersion and pulp agitation are critical.
• Temperatures and pressures are significantly lower than for other autoclave systems.
• During the oxidation leach, the arsenic is precipitated in a highly stable form in a residue that
significantly betters the TCLP test requirements. Cashman Process residues, if left exposed will
take millions of years to gradually release their potentially hazardous constituents back into the
environment.
• The process leaches base and precious metals and precipitates the arsenic in a single stage
leach (other processes require two or more stages).
ADVANTAGES OF THE CASHMAN PROCESS LEACH OVER CONVENTIONAL FLUE DUST
MANAGEMENT OR TREATMENT PROCESSES
• The Cashman Process could replace the usual procedure of recycling flue dusts within the
smelter, thereby allowing for increased smelterthroughput, increased metal purities and less dust
handling problems.
• Smelters could accept ores and concentrates higher in impurity element content, and still achieve
product specifications.
• Readily leachable arsenic trioxide or other arsenic complexes are converted to stable ferric
arsenate. Arsenic is estimated to leach from these materials over millions of years, which would
virtually eliminate loading of groundwater with arsenic.
• Long-term management and potential liabilities associated with discarding waste products can
be minimized.
• Waste volumes can be reduced by the removal of metals.
• Natural resources are conserved because otherwise discarded metals are returned to industrial
use.
REFERENCES
1. Kunter, R.S. and Bedal, W.E. The Cashman Process Treatment of Smelter Flue Dusts and
Residues. In: R.G. Reddy, W.P. Imrie, and P.B. Queneau (eds.), Residues and Effluents
Processing and Environmental Considerations. The Minerals, Metals, Materials Society,
Warrendale, Pennsylvania, 1992. p. 269.
2. Davis, A. Anaconda Smelter Flue Dust Cashman Process Residue Long-Term .Stability Study.
Report by PTI Environmental Services, Boulder, Colorado, 1990.
38
-------
ARSENIC EXTRACTION FROM SILT AND CLAYS
Adolfo R. Zambrano and Dennis D. Chilcote
BioTrol, Inc.
11 PeaveyRoad
Chaska,MN55318
USA
Tel: (612) 448-2515
INTRODUCTION
The application of BioTrol Soil Treatment System (BSTS) to the remediation of arsenic-contaminated
soils results in the separation of a residual fine fraction of the soil which essentially carries the bulk of the
arsenic and which is constituted primarily by a matrix of silt and clay(s). In general, this fraction does not
warrant a high remediation cost by extraction techniques due to its small volume and as a result, alternative
technologies such as solidification and stabilization become very attractive.
The opportunity to study an entire site with the characteristics of the fine fraction previously described
was a welcome challenge.
SITE CHARACTERISTICS
The Whitmoyer Laboratories Superfund Site, located in Lebanon County, Myerstown, PA, was
dedicated to the production of veterinary pharmaceutical products and for some 30 years produced organo-
arsenical compounds, primarily arsanilic acid (C6H8-AsNO3) and carbarsone (C7H9-AsN2O4). On-site
treatment of contaminated groundwater made extensive use of arsenic precipitation as calcium and ferric
arsenates which were stored in lagoons.
Two soil samples from this site were used in this study: an "average" soil sample having an arsenic
content of 1035 ppm, and the "worst" soil sample having an arsenic level of 3650 ppm. Both samples had
average values for silt and clay of about 85% and Cation Exchange Capacities of 14 mEq/100 g.
EXPERIMENTAL DESIGN
In order to compensate for the lack of an arsenic species distribution in the soils, the performance of
mineral acids and bases alone and in combination with complexing agents and polar organic solvents were
evaluated to select the two most effective extractive agents.
DESIGN IMPLEMENTATION
The search for the best two leaching agents was carried out as follows: 1) A series of 12 single-stage
leaching extraction tests was performed with a single extractant or a combination of extractants plus a
complexing agent or polar solvent using the "average" soil sample; 2) The best four performing extractants
or extractant-complexing agent combination from the previous series were retested using the "worst" soil
sample; 3) The best two performing extractants from the four-test series were re-evaluated in multistage
cross-current leaching extraction tests to simulate a continous operation.
All leaching extraction stages were conducted using a 35% solids slurry and were extracted for 4 hours
at room temperature (22 degrees C ± 2).
RESULTS AND DISCUSSION
Single Stage Extractions
"Average" Soil: Acidic Leaching at pH = 1.0. These results are summarized in Table 1. Although the
39
-------
three acids tested gave poor performances, sulfuric acid appeared to outperform both the nitric and
hydrochloric acids. The performance of hydrochloric acid was most improved in the presence of sodium
thiosulfate.
TABLE 1. ACIDIC ARSENIC EXTRACTION FROM "AVERAGE" SOIL AT pH = 1
REAGENTS
NAME
LEVEL,
Ib/STon
PROCESSED STREAMS
Name
keorl
Total TCLP- Arsenic
Arsenic, Arsenic, Extracted,
ppm ppm %
H2SO4
HNO3
HC1
HC1
DP-3
HC1
Na2SO3
HC1
Na2S2O3
165
190
140
120
1.3
135
20
175
20
Soil
Leachate
Soil
Leachate
Soil
Leachate
Soil
Leachate
Soil
Leachate
Soil
Leachate
0.30
0.98
0.28
1.16
0.27
1.05
0.28
1.05
0.26
1.63
0.28
1.15
974.0
17.9
1119.0
0.5
1051.0
0.4
836.0
2.0
1109.0
2.2
896.0
31.6
1.4
5.7
0.2
0.2
0.9
1.3
12.7
"Average" Soil: Basic Leaching at pH = 11.0. These results are given in Table 2. These data show that
the performance of the hydroxyl ions is improved by the presence of the carbonate ions and the performance
of these two ions is in turn improved by the presence of Cyquest DP-6 antiprecipitant. Also the performance
of hydroxyl ions is substantially improved in the presence of ethanol.
TABLE 2. BASIC ARSENIC EXTRACTION FROM "AVERAGE" SOIL AT pH = 11
REAGENTS
NAME
LEVEL,
Ib/STon
PROCESSED STREAMS
Name
kgorl
Total
Arsenic,
ppm
TCLP- Arsenic
Arsenic, Extracted,
ppm %
NaOH
NaOH
Na2CO3
NaOH
DP-6
Na2CO3
NaOH
Ethanol
10
20
380
14
1.3
370
56
200
Soil
Leachate
Soil
Leachate
Soil
Leachate
Soil
Leachate
0.27
1.35
0.29
1.22
0.30
1.01
0.30
0.97
1070.0
26.0
1148.0
69.2
691
124.0
691.0
140.0
1.4
0.1
10.8
20.2
37.7
62.0
40
-------
'Worst" Soil: Basic Leaching at pH = 11.0. The comparable performance of four lixiviant solutions are
summarized in Table 3. Any one of these lixiviants could be used to optimize a conceptual process
remediation scheme.
TABLE 3. BASIC ARSENIC EXTRACTION FROM "WORST" SOIL AT pH = 11
REAGENTS
NAME
Total TCLP- Arsenic
LEVEL, PROCESSED STREAMS Arsenic, Arsenic, Extracted,
Ib/STon Name kgorl ppm ppm %
NaOH
Na2S2O3
Na2CO3
NaOH
NaOCl
Na2CO3
NaOH
DP-6
Na2CO3
NaOH
Ethanol
Na2CO3
95
20
375
55
20
335
115
1.3
375
100
200
347
Soil
Leachate
Soil
Leachate
Soil
Leachate
Soil
Leachate
0.30
1.15
0.30
1.17
0.30
1.27
0.30
1.19
1440.0
658.0
1450.0
536.0
1206.0
568.0
1538.0
489.0
0.2
0.2
0.2
0.3
63.7
59.0
66.6
55.8
Multistage Simulations
Typical results for the two types of soil used in this study are depicted in Figure 1 for the evaluation
of Na2CO3 and Cyquest DP-6 but without NaOH additions.
55-
50-
.- 45-
c
g
| 40-
LU
O
0)
35-
30
25
20
0
2 3
Stages of Extraction
"Average" Soil
"Worst" Soil
Figure 1. Arsenic extraction using sodium carbonate and Cyquest DP- 6.
41
-------
CONCLUSIONS
The experimental data and the general observations during the course of the treatability study can
be summarized as follows:
Leaching with sulfuric, nitric, or hydrochloric acid at pH = 1 does not accomplish any significant
arsenic removal.
Under basic conditions, sodium carbonate in combination with sodium hydroxide and a chelating
agent can extract about 65% of the total arsenic. In the absence of sodium hydroxide the extraction
efficiency decreases. (Compare results in Table 3 and Figure 1.)
For leaching extraction times beyond 4 hours, solid-liquid separation by filtration becomes very
difficult.
Because ethanol can also extract more than 60% of the total arsenic, its utilization as the major
solvent will also drastically reduce the filtration difficulties associated with soil matrices made up
primarily of siltand clay(s). Solvent recovery and recycling can be accomplished through distillation.
A conceptual process flowsheet for the removal of arsenic- bearing contaminants other than the
ferric arsenates is shown in Figure 2 for the case of extraction with sodium carbonate and Cyquest
DP-6. An order-of-magnitude installed cost of $5 million was estimated within a -30% +50%
accuracy for a 20 tons per hour remediation plant. The operating cost to remediate 100,000 cubic
yards of soil was estimated at $160 per cubic yard, exclusive of excavation, debris removal, and
equipment capital recovery.
/CONTHMINRTED
SOIL ,
STOCKPILE /
1
SCREEN
CIRCUIT
i
CONCURRENT
LERCHING
CIRCUIT
SOLIDS
DEWRTERING
RND RINSING
LERCHED SOIL
^
7
/^ 10 MESH
\. OVERSIZE
\- -.
/H2S04
RNO I
H2S /
LERCHHTE
PRECIPITRTION
1
) PROCESS
IRTER
BIOTRERTMENT
1
RRSENIC SULFIDE
BYPRODUCT
Figure 2. Conceptual site remediation flowsheet.
42
-------
THE BEHAVIOR OF ARSENIC IN A ROTARY KILN INCINERATOR
Robert C. Thurnau
US Environmental Protectipn Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
USA
INTRODUCTION
Arsenic has always been an element of environmental interest. Its toxicity to the human body has been
well documented, and its regulation from an environmental standpoint has appeared in several statutes. With
the toxicity problems of hazardous waste and its treatment/disposal being better understood, releases of trace
elements (especially arsenic) through incineration were thought to be a potentialdanger. The data base to
understand and predict the impact from incinerating arsenic-laden wastes was very sparse (1 -5) and thus the
Environmental Protection Agency/Office of Research and Development (EPA/ORD)undertookthetasktostudy
arsenic's fate under a variety of combustion conditions. This paper will discuss the partitioning of arsenic as
observed during a controlled incineration study on an arsenic-containing synthetic waste and on a Superfund
soil containing arsenic as a primary contaminant.
APPROACH
EPA/ORD operates a research facility dedicated to the study of hazardous waste incineration and the
conditions that influence the different environmental release pathways. To address the question of arsenic
partitioning to the different streams, ORD through its Incineration Research Facility (IRF) systematically
evaluated the following variables:
Kiln and afterburner temperature ,
Waste feed chlorine content
Initial arsenic concentration
Effects on Toxicity Characteristic Leaching Procedure (TCLP)
Forthe parametric study, a synthetic mixture containing arsenic trioxide and other metals was mixed into
a homogenous solution and metered onto a clay/organic matrix as the material was fed to the incinerator. For
the study of the Superfund soil, the soil matrix was fed to the incinerator in the same condition as received, but
the total chemical identify of the arsenic was unknown. The test conditions for the Superfund soil were similar
to the parametric test conditions except the kiln oxygen was a new variable.
Samples of kiln ash, scrubber liquor, and scrubber exit flue gas were collected with EPA-validated
sampling methods and analyzed for arsenic to determine its partitioning.
RESULTS
The expectation at the beginning of the experiments was that arsenic, with its relatively high vapor
pressure, would be volatile and be increasingly displaced from the waste as the combustion temperature
increased. The data collected however, showed that the arsenic, although influenced by the kiln temperature,
was tending to be found in the ash. This relationship was not only clear, but reproducible with a linear regression
correlation coefficient of 0.97 for the parametric test data. The same type of trend was also observed for the
arsenic in the Superfund soil with a regression coefficient of 0.84.
A series of experiments were carried out to determine the effect of the afterburnertemperature on arsenic
partitioning. The data collected for afterburner temperatures of 1800, 2000, and 2220 degrees F indicate that
the partitioning of arsenic isindependent of afterburnertemperature. This experiment was not conducted on the
Superfund soil.
The interaction of chlorine with trace metals during incineration to form new compounds with different
boiling points has been suggested as a possible mechanism for explaining the observed partitioning data for
several of the trace metals studied. To determine if changes in waste feed chlorine concentration had an effect
on the partitioning of arsenic, tests were conducted in which feed chlorine concentration was varied from 0 to
43 -
-------
8 percent, whilethe other experimental parameters were held constant (kiln temperature=871 degreesC (1600
degrees F), afterburner temperature = 1093 degrees C (2000 degrees F)).
The results for arsenic partitioning to the kiln ash, flue gas, and scrubber liquor were collected and under
the conditions studied during these tests, arsenic partitioning did not vary with changes in feed chlorine
concentration. This observation is consistent with equilibrium calculations which do not predict a relationship
between arsenic vapor pressure and chlorine content.
The feed rate of arsenic during the parametric tests was designed to be constant throughout the test series,
but in reality varied from 1173to2112 mg/hr. Data collected illustrate the effect of varying the feed rateof arsenic
on partitioning to the incinerator effluents for all parametric tests conducted at constant kiln temperature. For
the range of feed rates experienced during this test series, no statistical correlation was found between the
arsenic feed rate and arsenic partitioning to the different effluents.
TCLP values for arsenic in the Superfund soil and in the kiln ash were also collected. In all cases, the act
of incinerating the soil increased the mobility of the arsenic in the ash as evidenced by the higher arsenic
concentrations in the leachate compared to the untreated soil. These data do not show a statistical relationship
between variations in kiln temperature and kiln ash arsenic TCLP values. However, the data do suggest a
relationship between kiln oxygen levels and the ash TCLP arsenic values. Specifically, at the higher kiln exit
oxygen concentrations the kiln ash TCLP values for arsenic were lower. On average, kiln exit oxygen
concentrations around 11 % resulted in arsenic TCLP values that were about 4.6 times lower than the 7% kiln
exit oxygen concentrations.
DISCUSSION AND CONCLUSIONS
Previous data gathered on the partitioning of arsenic during the incineration of hazardous wastes were
Inconclusive with regard to its distribution, and thus the potential for environmental release was not fully
understood. With the implementation of the Land Ban Rules, there was additional concern on the part of EPA
that the chemical composition of hazardous waste incinerated in the future would change. Metals such as
arsenic might be found at higher concentrations, thus increasing the potential for additional arsenic releases
to the environment.
Incineration data generated during tests at the IRF showed that the distribution of arsenic to the different
residual streams associated with hazardous waste incineration (in both a synthetic matrix and a "real world"
matrix) generally favored the solid phase, or ash, rather than the gas or vapor phase. Both sets of data showed
thatthe temperature of the kiln directly affected how much arsenic was volatilized. Overthe temperature range
studied, 816to 927 degreesC (1500 to 1700 degrees F)forthe parametric study and 816to982degreesC(1500
to 1800 degrees F) for the Superfund soil, the amount of arsenic volatilized increased with increased
temperature. The increase in volatility was much less pronounced in the parametric study with the net change
being only 9% over the 200 degrees F temperature range. The change in arsenic partitioning to the ash was
more pronounced in the Superfund soil, with an observed difference of 36% over a temperature range of 300
degrees F.
An increase in arsenic volatility with increased kiln temperature is predicted based on the thermodynamic
data, but the extent to which arsenic is retained by the ash in these experiments was unexpected. The chemical
form that the arsenic assumed during incineration may be responsible for some of the unexpected partitioning
to the ash. Also, thearsenic may have become chemically or physically bound with the solid matrix and not easily
released to the flue gas.
The TCLP results associated with the Superfund soil residuals were very interesting. Incineration altered
thesoil matrixto the extentthat arsenic was much more mobile in the kiln ash than in the native soil. Even though
the concentration of arsenic in the native soil was high, it was retained by the soil, which was not a hazardous
waste as defined by TCLP. The leachate concentrations were well below the regulatory TCLP limit of 5 mg/L.
As the amount of oxygen in the kiln decreased, the arsenic TCLP values for the kiln ash increased, although
the kiln ash leachate concentrations did remain below the regulatory limit. Again, the chemical form of the
arsenic and the components of the matrix are unknown, but probably are responsible forthis phenomenon. The
importance of this observation will be reflected in the operating conditions of an incinerator burning this soil,
along with the added costs for subsequent ash treatment, if required.
The incineration of soils contaminated by organics and arsenic may be a feasible remediation option for
44
-------
destroying the organic compounds. The results of these studies suggest that by selectively optimizing the
incineration conditions, arsenic can be retained in the ash where it can be effectively handled.
REFERENCES
1. Oppelt.E.T. Incineration of Hazardous Waste: A Critical Review. JAPCA 37(5). May 1987. pp. 558-
86.
2. Fournier, D. F., Whitworth W. E., Lee J.W., Waterland L.R.,Thurnau, R.C. and Carroll, G. J. Pilot Scale
Evaluation of the Fate of Trace Metals in a Rotary Kiln Incinerator with a Venturi Scrubber/Packed
Column Scrubber. EPA/600/2-90/043a, U.S. Environmental Protection Agency, Cincinnati, OH,
September 1990.
3. Wall, H. O. and Richards, M.K. The Incineration of Arsenic- Contaminated Soils Related to the
Comprehensive Environmental Response, CompensationandLiability(CERCLA).JniProceedingsof
the Sixteenth Annual Hazardous Waste Research Symposium, Cincinnati, OH, April 1990. EPA/600/
9- 90/037.
4. , Waterland, L. R., King, C., Vocque, R. C., Richards, M.K., and Wall, H. Pilot Scale Incinerability
Evaluation of Arsenic and Lead Contaminated Soils From Two Superfund Site. Presented at 1991
Incineration Conference, Knoxville, TN, May 13-17,1991.
5. Carroll, G. J., Thurnau, R. C., Mournighan, R. E., Waterland, L. R., Lee, J. W. and Fournier Jr., D. J.
The Partitioning of Metals in Rotary Kiln Incineration. inj_Proceedings of the Third International
Conference on New Frontiers for Hazardous Waste Management, U.S. Environmental Protection
Agency, Washington, DC. EPA/600/9-89/072.
45
-------
REMOVAL OF ARSENIC FROM WASTEWATERS AND STABILIZATION OF
ARSENIC BEARING WASTE SOLIDS
L.G. Twidwell, K.O. Plessas, and T.P. Bowler
Metallurgy-Mineral Processing Engineering Department
Montana College of Mineral Science and Technology
Butte, MT 59701
USA
Tel: (406) 496-4208
INTRODUCTION
An extensive compilation of references concerned with arsenic removal from wastewaters, stabiliza-
tion of waste materials, and treatment of arsenic bearing wastes has been assembled by the authors. This
compilation will be available at the conference. The emphasis of the presentation will be on removing
arsenic from wastewaters and stabilizing arsenic-bearing solids. A review of the literature on arsenic
removal and stabilization will be presented and experimental results from studies conducted at Montana
Tech will be discussed. Studies to be discussed include: arsenic removal from solution; stabilization of
arsenic-bearing waste materials; and treatment of arsenic bearing metallurgical waste and byproducts.
STUDIES AT MONTANA COLLEGE OF MINERAL SCIENCE AND TECHNOLOGY
Arsenic Removal from Solution
Comba (1) investigated the removal of arsenic from solution by the formation of mimetite, a lead
chloroarsenate (Pb,.(AsO4)3CI). His results demonstrated that arsenic could be effectively removed from
solution as a crystalline precipitate to concentrations below the detection limit (by AA analysis), i.e., the
arsenic concentration was below 0.2 ug/L (ppb). The free energy of formation for mimetite was determined
to be -625 ± 2 kilocalories/mole. The equilibrium stability diagram for the lead-arsenate-chloride system
(using the determined free energy value) ispresentedin Figure 1. Note the low solubility of arsenic at natural
waterpH levels. The filterability of the mimetite was excellent because the morphology of the precipitate was
small crystalline spherites. The lead left in solution could be stripped from solution as lead hydroxide, lead
carbonate, or lead phosphate.
Pb(D.D1moles/I]-As(0.005)-CI(0.05]
O)
o
in
<
-6
-g
-10
-12
c HjAsO,,-
-------
presented in Table 1. These solution/solid mixtures have been aged in closed containers for approximately
4 years and are currently being reanalyzed for solution arsenic and for phase identification.
TABLE 1. ARSENIC REMOVAL FROM SOLUTION
System
Starting Cone.
Ca-As
Ca-As-Cl
Ca-As-P
Ca-As-P-Cl
Arsenic (V), ug/lit
106
91.6
0.7
<0.5
<0.5
PH
12.7
12.6 (8 samples)
12.6
12.6 (5 samples)
Bowler (3) investigated the removal of arsenic from solution by a variety of techniques. He
accomplished essentially complete removal by precipitation/adsorption using calcium hydroxide additions.
His results are in agreement with the results reported previously by Nishimura and Tozawa (4) and Robins
and Huang (5). The solid products pass the Toxicity Characteristic Leaching Procedure (TCLP) test for
arsenic but, as has been reported by Robins (6), the solid product is not appropriate for long-term chemical
pond storage.
Bowler (3) and Plessas (7) investigated arsenic removal by precipitation at elevated temperatures.
They demonstrated that arsenic may be effectively removed from elevated pH and temperature solutions
by precipitation of maghemite/magnetite. The test work was conducted in the pH range 7 to 11 (3), 8 to 10
(7) at temperatures of 70 to 80 degrees C (3), 90 degrees C (7). Successful arsenic removal was achieved
to below the drinking water standard (7). The precipitated products passed the TCLP test.
Honores (8) noted in a study (concerned with the recovery of metal values from copper smelter slags)
that during the removal of copper by iron cementation that greater than 90% of the arsenic was also
removed. Plessas (7) followed up on this noted effect and investigated arsenic removal by cementation
using iron. She investigated the removal of arsenic from synthetic wastewaters (at various pH levels) by
flowing solutions over iron scrap (in a column setup). Arsenic concentrations in the range 1 to 5 mg/L were
achieved at pH levels of 5 to 6. It is presently not clear whether the removal was by reduction of arsenate
to arsenic metal or was due to adsorption by oxidized iron on the surface of the iron scrap.
Stabilization of Arsenic-Bearing Waste Materials
Four experimental studies have been conducted at Montana Tech on stabilization of solid waste
materials. The technologies utilized were vitrification (9,10) (one study) and cement/lime stabilization (11-
13) (three studies). Twidwell and Mehta (10) proposed that a way to dispose of coppersmelterflue dust was
via conversion of the arsenic to calcium arsenate which could then be dissolved in copper smelter slag.
Copper slags were doped with arsenic by dissolution of calcium arsenate in molten slag. Slags were then
subjected to the EP toxicity test (refer to Table 2) and also leach tested in a water environment for up to 7
years. Arsenic release was minimal from the glassy slag test materials.
Three studies have been conducted that used cement or cement/lime mixtures for stabilizing various
arsenic-bearing materials, e.g., copper smelterflue dust (11), calcium and iron arsenate contaminated soils
(12), and ferric hydroxide precipitated solids (13). Tang (11) investigated the stabilization of copper smelter
flue dust with cement/lime and the influence of stripping copper from flue dust (by a pyrometallurgical
process) on the stability of the final arsenic bearing residue. He demonstrated as a part of his study that flue
dust was stabilized by the addition of cement (25%)/lime (10%), i.e., the mixtures passed the TCLP test for
arsenic, lead, and cadmium. This result was in agreement with test results generated by ARCO (14) (who
have signed a ROD for disposal of 360,000 tons of flue dust by cement/lime stabilization). The findings of
47
-------
both studies agreed that passing the TCLP test required that sufficient cement/lime must be present to
ensure that the TCLP solution pH be in the range 10 to 11.
Atreatability study (on NPL site material) conducted by Twidwell and Chatwin (12) demonstrated that
cement stabilization of calcium and iron arsenate/arsenite contaminated soils were not stabilized by cement
alone. The presence of cement and a high pH in the TCLP test solution were not sufficient for the mixtures
to pass the TCLP test. An additional stabilization roast was required. A brief summary is presented in Table
3. The roast/cement stabilization technology was chosen as the alternative for cleanup at the Whitmoyer
NPL site.
TABLE 2. EP TEST RESULTS FOR DOPED COPPER REVERBERATORY SLAG SYSTEMS
As in Slag, %
0.5'
0.81
2.1
3.3
5.2
9.1
19.4
Extraction Solution Analysis, mg/L
As"
0.016
0.047
0.448
0.421
0.901
0.415
0.802
Cdb
0.093
0.000
0.000
0.000
0.000
0.001
0.002
Cr"
0.016
0.007
0.006
0.004
0.007
0.007
0.007
Pbb
0.226
0.149
0.169
0.150
0.150
0.148
0.149
Zn
0.30
0.010
0.082
0.084
0.151
0.060
0.036
Cu
' 0.239
0.474
0.526
0.270
0.294
0.050
0.008
* As received (undoped) copper reverb slag
b EPA designated characteristic concentration for As, Cr, Pb is 5 rng/L; for Cd is 1 mg/L
TABLE 3. STABILIZATION OF WHITMOYER NPL SITE MATERIALS
Arsenic Content, %
1.3
17.8
Cement/Waste
3
3R
1
1R
0.5
0.5R
3
3R
1
1R
TCLP, ing/liter (pH)
5.02 (11.3)
0.11 (11.4)
15.80 (11.0)
0.43 (11.5)
39.9 (11.6)
0.45 (11.5)
5.0 (11.8)
0.99(11.7)
72.2 (11.4)
10.8 (10.3)
R designates roasting at 700°C for 1 hour.
A third study was conducted by Twidwell and McGrath (13) to evaluate whether organic arsenic
(monosodium methylarsonate, MSMA) could be stripped from a salt brine solution (containing approxi-
mately 2 g/L arsenic) and the product stabilized by cement. Five solution treatment techniques were
investigated (four precipitation and one solvent extraction). Thefinal recommended procedure, i.e., the use
of ferric precipitation (with an iron/arsenic mole ratio of ten) resulted in excellent results, e.g., 2.3 gpl arsenic
was reduced to 8 to 30 ug/L. TCLP results on the solid products ranged from 0.296 to 0.715 mg/L.
The products from the ferric stripping studies were subjected to roast stabilization (chosen because
of the success achieved in the Twidwell, Chatwin study and the success achieved by Tozawa, Nishimura
and Umetsu (15)). Roast stabilization was unsuccessful. The roasted solids actually leached more arsenic
in the TCLP tests than the unroasted precipitated products. Roasting in the presence of cement (1 part
cement/1 part ferric product) was successful and resulted in TCLP values that were three to five times less
than the results on the untreated ferric product.
48
-------
Treatment of Arsenic-Bearing Metallurgical Wastes and Byproducts
One of the major arsenic-bearing metallurgical waste byproducts from smelting is flue dust. Vast
quantities of flue dust currently exist; some containing up to 20% arsenic (9). Flue dust continues to be
experimentally investigated at Montana Tech. Anderson (16) experimentally surveyed the use of a number
of elevated temperature roasting techniques for recovering arsenic from copper smelter flue dusts. He
specified optimum experimental conditionsfor removing arsenicasarsenicoxidefromthedusts.Mehta (17)
developed means for producing calcium arsenate (by roasting) in the flue dust and Mehta and Twidwell (10)
investigated disposal of flue dust by dissolution in smelter slag.
Newhouse (18) and Arratia (19) applied the Copper Segregation Process (20) to the dust and to the
dust mixed with lime, respectively. They demonstrated that the presence of lime mixed with the dust retained
arsenicand allowed forapproximately 60% copper recovery. Tang (11) studied the recovery of copperfrom
cement/lime stabilized flue dust (as stated previously). Neira (21) developed reductive roast processes for
stripping arsenic from flue dust as relatively high purity metal (99.5%).
Several investigations have been performed to develop a process for producing copper without the
formation of arsenic- bearing flue dust. Beuerman and Twidwell (22,23) investigated the recovery of copper
from high arsenic concentrates (and the retention of arsenic in the concentrate residue) by a non-smelting
segregation treatment. Excellent copper recovery was achieved but arsenic was volatilized^ Gregory (2)
continued the research to demonstrate complete retention of arsenic in the concentrate residue (without the
emission of arsenic or sulfur dioxide gas) by a lime roasting/segregation process. He demonstrated over
90% copper recovery without any arsenic emission.
Other byproduct treatment processes include studies by Flynn (25) to remove arsenic from lead blast
furnace speiss as arsenic metal and by Peterson (26,27) who developed techniques to recoverarsenicf ram
speiss as arsenic metal or as arsenic sulfide.
REFERENCES
1. Comba, P. Removal of Arsenic from Process and Wastewater Solutions, M.S. Thesis, Montana
College of Mineral Science and Technology, Butte, MT, April 1987. 137 p.
2. Twidwell, L.G. The Removal of Toxic Species from Waste Solutions by the Formation of Stable
and Filterable Precipitates, Quarterly Report, USBM Mineral Industry Waste Treatment and
Recovery Generic Center, Univ. of Nev., Reno, NV, Sept. 1988. 4 p.
3. Bowler, T. Removal of Arsenic from Wastewater Solution by Precipitation Techniques, M.S.
Thesis, Montana College of Mineral Science and Technology, Butte, MT, May 1992. In Progress.
4. Nishimura, T., K. Tozawa. Removal of Arsenic from Waste Water by Addition of Calcium
Hydroxide and Stabilization of Arsenic- Bearing Precipitates by Calcination, Impurity Control and
Disposal, CIM 15th Annual Hydromet. Conf., Vancouver, BC Canada, August 1985. P3, pp. 1 -20.
, 5. Robins, R.G., J.C.Y. Huang, T. Nishimura, G.H. Khoe. The Adsorption of Arsenate Ion by Ferric
Hydroxide, im Arsenic Metallurgy Fundamentals and Applications, (eds): R.G. Reddy, J.L.
Hendrix, P.B. Queneau, AIME-TMS, Phoenix, AZ, January 1988. pp. 99-114.
6. Robins, R.G. Reference to be supplied.
7. Plessas, K. Arsenic Removal from Process and Wastewater Solutions, M.S. Thesis, Montana
College of Mineral Science and Technology, Butte, MT, August 1992. In Progress.
8. Honores, C. Recovery of Copper and Zinc from Copper Reverberatory Slag, M.S. Thesis,
Montana College of Mineral Science and Technology, Butte, MT, May 1992.100 p.
9. Mehta, A.K, Investigations of New Techniques for Control of Smelter Arsenic Bearing Wastes,
EPA-600/S2-81-049, PB81-23581, Cincinnati, OH, Sept.'1981. : ,
49
-------
10. Twidwell, L.G., A.K. Mehta. Disposal of Arsenic Bearing Copper Smelter Flue Dust, Nuc. Chem.
Waste Management, Vol. 5,1985. pp. 297-303.
11. Tang, X. Recovery of Copper from Stabilized Copper Smelter Flue Dust, M.S. Thesis, Montana
College of Mineral Science and Technology, Butte, MT, May 1992. 85 p.
12. Twidwell, L.G., T.D. Chatwin. Treatability Study for the Whitmoyer Laboratories Site Arsenic
Stabilization: Cement Casting, Final Report to NUS.WA No. 200-3LC9,Treatability Studies 4938-
89S-1041, July 1989.106 p. .
13. Twidwell, L.G., S. McGrath. Survey of Arsenic Removal and/or Stabilization Techniques Applied
to Fermenta Waste Solutions and Solids, Final Report to Fermenta, ASC, December 1989.84 p.
14. Canonie Environmental Services. Anaconda Smelter NPL Site Flue Dust Operation Unit Reme-
dial Design Work Plan, Draft Report to ARCO, Anaconda, MT, March 1992.
15. Tozawa, K., T. Nishimura, Y. Umetsu. Removal of Arsenic from Aqueous Solutions, 16th Annual
CIM Conf. of Met., Vancouver, Canada, August 1977.10 p.
16. Anderson, C.G. A Survey of Roasting Techniques to Volatilize Arsenic and Antimony from Copper
Smelter Flue Dust, M.S. Thesis, Montana College of Mineral Science and Technology, Butte, MT,
February 1984.126 p.
17. Mehta, A.K. Fixation of Arsenic in Smelter Flue Dust, M.S. Thesis, Montana College of Mineral
Science and Technology, Butte, MT, April 1978.
18. Newhouse, J.P. Segregation Process Applied to Copper Smelter Flue Dust, M.S. Thesis,
Montana College of Mineral Science and Tephnology, Butte, MT. 108 p.
19. Arratia, J. Optimization of Lime Roasting-Segreqation Treatment of Flue Dust, M.S. Thesis,
Montana College of Mineral Science and Technology, Butte, MT, Ma'y 1985.
20. Marcuson, S.W. Application of the Segregation Process to Roasted Copperconcentrates, Mineral
Science and Engineering, Vol. 12, No. 1, January, 1980.
21. Neira, M. Recovery of Elemental Arsenic from Copper Smelter Flue Dust by Volatilization, M.S.
Thesis, Montana College of Mineral Science and Technology, Butte, MT, May 1990. 90 p.
22. Beuerman, K. A Study of the Copper Segregation Process Applied to Copper Concentrate from
the Weed Concentrator at Butte, Montana, M.S.Thesis, Montana College of Mineral Science and
Technology, Butte, MT, April 1982.178 p.
23. Twidwell, L.G., and K. Beuerman. The Segregation Process-Applied to Dead Roasted Copper
Concentrates: The Distribution of Impurities, Proc. Australas. Inst. Min. Metall., No. 289,
November 1984. pp. 295-302.
24. Gregory, P. The Segregation Process Applied to Lime Roasted Copper Concentrates: The
Distribution of Impurities, M.S. Thesis, Montana College of Mineral Science and Technology,
Butte, MT, May 1992. In Progress.
25. Flynn, H. A Survey of Pyrometallurgical Methods for Lead Speiss Treatment, M.S. Thesis,
Montana College of Mineral Science and Technology, Butte, MT, May 1980. 80 p.
26. Peterson, M. Removal of Arsenic and Antimony from Lead Smelter Speiss, M.S. Thesis, Montana
College of Mineral Science and Technology, Butte, MT, May, 1985. 79 p.
27. Peterson, M..L.G. Twidwell. Removal of Arsenic from Lead SmelterSpeiss, J.Haz. Materials, Vol.
12, Elsevier Science Publishers, Amsterdam, 1985. pp. 225-229.
50
-------
TREATMENT OF LANDBAN-VARIANCED ARSENIC WASTES ATTSDFs
Jesse R. Conner and Paul R. Lear
Chemical Waste Management, Inc.
Geneva Engineering and Technology Center
1950 South Batavia Ave.
Geneva, Illinois 60134
USA
Tel: (708) 513-4808
INTRODUCTION
Normally, traditional chemical stabilization processes are able to immobilize arsenic in contaminated
soils, incinerator ashes, and other wastes to well below US Environmental Protection Agency (EPA)
requirements using the Toxicity Characteristic Leaching Procedure (TCLP) for remedial actions where the
waste is to be landfilled on-site after treatment. A sampling of results from treatability studies and
commercial stabilization operations at Chemical Waste Management Inc. (CWM) is given in Table 1.
TABLE 1. TYPICAL STABILIZATION RESULTS ON ARSENIC-CONTAINING WASTES
Waste Description
Contaminated soil
Hazardous waste (H.W.) landfill leachate
H.W. Incinerator ash, fluid bed
H.W. Incinerator pond sludge
ti
Pesticide sludge
II H
Pesticide soil
. il it
Phosphoric acid filter cake
Test
Type*
TCLP
TCLP
TCLP
TCLP
EPT
EPT
Cal WET
EPT
Cal WET
TCLP
Arsenic Content (mg/I or mg/kg)
K^^^^^^«Mm^^^^B
Total
(mg/kg)
56.9
,44.0
47.6
24.7
24.7
1250.0
1250.0
400.0
400.0
7950.0
Untreated
Leach
0.10
4.20
0.07
0.30
0.30
19.0
52.0
0.60
28.0
70.0
Treated
Leach
0.04
0.016
0.019
<0.01
<0.01
0.14
5.20
0.27
.6.50
1.580
* Test Methods: TCLP = Toxicity Characteristic Leaching Procedure; EPT = Extraction
Procedure Toxicity Test; Cal WET = California Waste Extraction Test
However, in California a different test, the WETor CAM, is used to establish acceptable immobilization
for such on:site disposal. This test is much more severe, especially for arsenic (most especially where the
arsenic is present as an organic species) and lead, and often results in WET failures where the material
passes the TCLP. CWM has developed a simple, cost-effective, one-step stabilization formulation that
allows the treatment of most of these wastes to pass both the TCLP and WET protocols. For more difficult
cases, more complex, multi-step processes were required; they are discussed in following paragraphs.
Certain arsenical species, arsenic trisulfide and organic arsenicals primarily, are very difficult to
stabilize by conventional chemical means. Because of this, EPA established a Best Demonstrated Available
51
-------
Technology (BOAT) leaching standard based on vitrification (1) for D004 (arsenic characteristic) and
transferred this standard to K031, K084, P011, P012, P036, P038 and U136 non-wastewater codes in the
"third-third" Land Disposal Restrictions (LDR). Due to lack of treatment capacity, EPA granted a 2 year
national capacity variance for these waste codes, until May 8, 1992 (2). Since 1988 CWM has been
conducting extensive research and development on arsenic stabilization at its Geneva Research Center.
This work has resulted in an extensive database on immobilization of arsenic in a wide variety of waste
streams and to different regulatory standards and test protocols.
This database was shared with the EPA before the third-third LDRswere finalized in 1990. Atthattime,
the chemical stabilization of arsenic trisulfide sludge from phosphoric acid manufacture and organo-
arsenicals in a variety of wastes had still not been successfully accomplished on a commercial scale. More
recently however, CWM research has resulted in practical formulations and processes for nearly all of these
especially difficult-to-treat wastes. This paper discusses these two difficult types of arsenic-bearing waste
streams. Waste characteristics, process description, and treated residue characteristics are discussed for
each waste type.
ARSENIC TRISULFIDE WASTE FROM PHOSPHORIC ACID PRODUCTION
Arsenous trisulfide (As2S3), especially in high concentrations, is very difficult to immobilize chemically
because of its solubility under the high pH conditions of conventional stabilization processes. This waste
is produced from the "wet process" of food-grade H3PO4 manufacture from phosphate rock. Arsenic is an
impurity in the rock, and must be remoyed from the raw acid by precipitation with H2S gas. The precipitate
is removed by settling and filtration
Waste Characteristics
These wastes typically are highly acidic from the residual phosphoric acid, with arsenic concentrations
in the 0.5 to 4% range, although some samples may contain even higher concentrations. Other solids in the
waste stream consist of other metals precipitated along with the arsenic, as well as large amounts of filter
aids such as diatomaceous earth and other solid impurities. In some cases, the waste also contains cement
used to eliminate free liquid in the filter cake or sludge. It is bright yellow in color and has a relatively mild
sulfide odor. This waste is unusual in that while most metal sulfides decompose in acids to yield soluble
metal salts and H2S, arsenous trisulfide is stable and relatively insoluble under acid conditions, but fairly
soluble in alkalies.
Process Description
A large number of conventional stabilization techniques were tested on this waste, with either negative
or highly variable results. Research into the literature of arsenic chemistry indicated that the best approach
would be to oxidize the As*3 to As*5 under alkaline conditions, forming the arsenate anion which can then
be precipitated by cations such as Ca*2 or iron to form low- solubility species. This approach, in fact, worked
well using lime, calcium, or sodium hypochlorite, and a stabilization reagent. A similar treatment scheme
was developed by a waste generator (3). We have now demonstrated at full production scale a reliable two-
step process that is being implemented at several CWM sites. Full-scale testing has been completed and
CWM fixed sites were in production as of May 8, 1992 when the existing treatment variance terminates.
Characteristics of the Treated Residue
Both laboratory and full-scale tests yielded arsenic TCLP leaching levels below 0.5 mg/L in all cases,
and below 0.05 mg/L (the drinking water standard) in many instances. Table 2 gives a summary of bench-
scale testing results and Table 3 the full-scale pilot test results at two different CWM sites. These results
show that the system chosen is effective forthe stabilization of arsenous trisulfide wastes. Furthermore, full-
scale pilot tests yield results comparable to the bench-scale results. The process is not sensitive to the type
of oxidizing agent used: all three tested yielded good results. Similarly, a variety of conventional stabilization
reagents can be used. This flexibility allows a treatment site to tailor its procedure to cost-effectiveness of
reagents available and to other facility restraints.
52
-------
TABLE 2. SELECTED RESULTS OF BENCH-SCALE STABILIZATION TREATABILITY
TESTING FOR ARSENOUS TRISULFIDE WASTES
Sample
1
2
2
3
4
5
Alkaline
Reagent
Hydrated Lime
Hydrated Lime
Hydrated Lime
Cement* *
Hydrated Lime
50% NaOH
Oxidation
Reagent
Ca(OCI)2
Ca(OCI}2
H202
Ca(OCI)2
Ca(OCI)2
NaOCI
Stabilization
Reagent
Bed Ash
CKD*
CKD
Hydrated
Lime
CKD
Bed Ash
Arsenic Content {mg/l or mg/kg)
' Total
(mg/kg)
13000.
37200.
37200.
10300.
6390.
1750.
Untreated
Leachate
> 100.0
> 100.0
> 100.0
>30.0
>30.0
>20.0
Treated
Leachate
0.25
0.22
0.04
0.03
0.04
2.0
Cement Kiln Dust
e Pretreated at generator's facility
TABLE 3. SELECTED RESULTS OF FULL-SCALE STABILIZATION TREATABILITY TEST-
ING FOR ARSENOUS TRISULFIDE WASTES
Sample
*
t
1
3
1
CWM Site
Lake Charles
Emelle
Lake Charles
Alkalinity
Reagent
50% NaOH
Cement**
Hydrated Lime
Oxidation
Reagent
NaOCI
Ca(OCI)2
NaOCI
Stabilization
Reagent
Bed Ash
Lime/Cement
Bed Ash
Arsenic
Content in
Treated
Waste
Leachate
(mg/l)
0.43
<0.10
<0.14
* See Table 2 for waste description and bench scale results
** Pretreated at generator's facility
The exact cost of our method for arsenous trisulfide wastes depends heavily on reagent usage,
especially that of the oxidizing agent, and therefore on the arsenic content of the waste as well as on other
factors. Reagent cost is expected to range in the area of $250 to $550 per ton of waste treated. This is high
compared to most stabilization systems, but is still less expensive than other options that are commercially
available.
53
-------
VETERINARY PHARMACEUTICAL PROCESS WASTES AND HIGH-LEVEL ORGANIC PESTICIDE
CONTAMINATED SOILS AND SLUDGES
Waste Characteristics
Species such as cacodylic acid (dimethylarsinic acid) present in wastes containing weed killers and
other organo-arsenic compounds are difficult to immobilize by conventional methods. Additional organic
compounds encountered in veterinary Pharmaceuticals include 4-hydroxy-3-nitrobenzene arsonic acid, 2-
hydroxy-3-nitrobenzene arsonic acid, 4-nitro-phenylarsonic acid, and their metal salts. These organics may
be encountered in highly concentrated form and in organic matrixes, or as contaminants in soils, sludges,
and other wastes. In the former case, incineration is usually indicated, and it is very expensive because of
the necessity to dilute the organic with other incineration feed stock because of arsenic concentration
problems. In the latter case however stabilization is preferable from both cost and operational bases, if the
LDR requirement of 5.0 mg/L by the TCLP test can be met.
Five samples of organo-arsenical wastes from a remediation site were tested in this project at
laboratory scale. Three were concentrated in arsenic, at 14,300 to 31,700 mg/kg, and two were slightly
contaminated soils at about 140 mg/kg. The analyses and leachabilities of the untreated wastes are shown
in Table 4.
TABLE 4. SELECTED RESULTS OF BENCH-SCALE STABILIZATION TREATABILITY
TESTING FOR ORGANO-ARSENICAL WASTES
Sample
1
1
1
2
2
2
3
3
3
4
4
Waste Type
Concentrated Waste #1
n n
ti it
Concentrated Waste #2
it n
n n
Concentrated Waste #3
n it
n ii
Contaminated soil #1
Contaminated soil #2
Stabilization Reagent
Portland cement
Quicklime
Proprietary system #1
Portland cement
Quicklime
Proprietary system #2
Portland cement
Quicklime
Proprietary system #2
Portland cement
Portland cement
Arsenic Content (mg/l or mg/kg)
Total
(mg/kg)
1 4300.
1 4300.
1 4300.
1 9800.
1 9800.
I 9800.
31700.
31700.
31700.
139.
134.
Untreate
d
Leachate
518.
518.
518.
427.
427.
427.
1060.
1060.
1060.
1.20
1.36
Treated
Leachate
60.5
16.7
16.4
9.61
5.00
0.42
36.4
21.1
1.48
<0.25
<0.25
Process Description
Again, a large number of conventional stabilization techniques were tested on this waste, with either
negative or highly variable results. It was evident from the literature that the destruction of the organic
species, at least to the point of releasing the arsenic, would be necessary so that arsenic could be
immobilized as a metal arsenate. A major research project was initiated in cooperation with a generator to
deal with certain wastes in this category, and preliminary results to date are being presented here for the
first time. After considerable testing, a method was found that accomplished this goal in most cases. This
process is considered proprietary by CWM and may be the object of patent applications. It is now being
tested at pilot scale at one of CWM's Treatment, Storage, and Disposal Facilities (TSDFs).
54
-------
Characteristics of the Treated Residue
Table 4 gives some of the results of this program. In the case of the soils, it was found the arsenic
teachability was not a problem, since the untreated soils leached at levels below the LDR requirement.
Nevertheless, some testing was done to determine whether this level could be further reduced by
conventional stabilization methods. A simple Portland cement system gave good results, reducing
leachability to below limits of quantitation (LOQ) for this matrix, <0.25 mg/L.
The untreated concentrated wastes exhibited high total levels of arsenic and very high leaching levels.
In general, the leaching level increased with the total concentration of arsenic in the waste. With wastes #2
and #3, conventional stabilization with cement or lime reduced the leachability substantially, but not to the
required LDR levels. The proprietary system #2 was effective in reducing leachability to well belowthe LDR
level of 5.0 mg/L. In the case of waste #1, however, this proprietary system, and another one (#1) achieved
only the same leaching level as quicklime - more than 16 mg/L - and well above LDR requirements. However,
leachability was substantially reduced by a factor of about 10, taking into account the effect of dilution from
the stabilization reagents and process water.
Additional work is underway to improve immobilization of waste #1. In addition, continuous treatability
testing is being done on other organo-arsenical wastes to improve and optimize the process. The arsenic
work has already had extended application to the immobilization of other organo-metal problem wastes.
i
FACILITIES
Chemical Waste Management Inc. operates central Resource Conservation and Recovery Act
(RCRA) TSDFs at a number of locations throughout the US. Seven of these TSDFs perform stabilization
treatment and disposal on a wide variety of hazardous wastes. It is currently expected that two of these
facilities will process arsenous trisulfide wastes, and likely organic arsenicals as well. These facilities are
at Lake Charles, Louisiana and Emelle, Alabama. All of the CWM facilities can stabilize and dispose of the
more conventional arsenic wastes.
REFERENCES
1. U.S. EPA. Federal Register 55(106): 22558, June 1,1990.
2. U.S. EPA. Federal Register 55(106): 22636, June 1,1990.
3. Fisher, D.O. and Lannert, K.P., U.S. Patent 4,948,516, August 14,1990.
55
-------
CHARACTERIZATION OF ARSENIC-CONTAINING MINING/SMELTING WASTES IN
THE CLARK FORK BASIN, MT AND SOME POTENTIAL REMEDIAL TECHNOLOGIES
Mike Bishop,
U.S. EPA
Federal Building
301 S. Park
Drawer 10096
Helena, MT 59626
USA
Tel: (406) 449-5414
INTRODUCTION
The headwaters of the Clark Fork River are located in southwestern Montana. Within the area known as
the upper Clark Fork River Basin, world-class mining and smelting operations existed for a period exceeding
100 years. During the course of these operations a variety of arsenic-containing extraction, benef iciation, and
processing waste streams were released. The historic releases of mining and smelting wastes have resulted
in situations where threats exist to human health and the environment.
Four Superfund sites are located within the upper Clark Fork River Basin that collectively involve one of
the largest Superfund investigations in the nation (1). The National Priorities List (NPL) sites located within the
Basin include the Silver Bow Creek/Butte Area, Montana Pole, Anaconda Smelter, and Milltown Reservoir sites
(Figure 1). Of these sites, all but the Montana Pole site involve a variety of mining and smelting wastes that
contain characteristic arsenic concentrations.
Figure 1. Location of Superfund sites in the Clark Fork Basin.
'ILLTOWN RESERVOljfsiTE
i-Milltown ""~-""
SILVER BOW CREEK/
BUTTE AREA SITE
ONTANA POLE SITE
Clark Fork Drainage Basin
City
Town
State Highway
Interstate Highway
- River or Stream
Superfund Sites
56
-------
PURPOSE
The purpose of this presentation is to discuss analytical methods used to characterize these NPL sites,
review concentrations of arsenic that are typical of the waste types associated with these NPL sites, and to
summarize technologies that have been used within the Clark Fork Basin to remediate potential exposure to
these wastes. Current efforts to develop innovative technologies for the treatment of mining wastes will also be
discussed.
ANALYTICAL METHODS
Three analytical approaches are generally used in association with the Clark Fork Superfund sites to
characterize arsenic concentrations in the solid wastes described previously including: the Environmental
Protection Agency (EPA) Toxicity Characteristic Leaching Procedure (TCLP) method 1311, the EPA contract
lab program (CLP) acid extractable metals (CLP SOW 788) and X-ray fluorescence (XRF).
The circumstances for applying each of these procedures are explained in a regional Clark Fork Lab
Analytical Protocol (2), and Laboratory Analytical Protocol for X-ray Fluorescence Analysis of Solid Media (3).
Differences in analytical values (4) and problems associated with each analytical approach will be discussed.
WASTETYPES
Waste types identified at the NPL sites previously identified include waste rock, mill tailings, flue dust,
arsenical wood treating waste, slag, and contaminated stream sediments.
Waste rock is the rejected material resulting from mining operations which is usually disposed of near the
source. Over 125 years of mining in Butte, approximately 100 major waste rock dumps were developed
containing millions of yards of waste covering hundreds of acres. Waste rock was also extensively distributed
along railroad lines as ballast. The chemical composition of waste rock is highly dependent on the source
materials being mined (5).
Tailing deposits resulting from numerous smelting operations in the upper Clark Fork Basin are extensive.
In Butte three major accumulations of tailings exist including the Yankee Poodle, Clark, and Colorado tailings.
The Clark and Colorado tailings alone include over 1,250,000 cubic yards of material (5). Additional large
volumes of tailings occurat the Ramsey Flats directly adjacent to Silver BowCreek. Tailings from the Anaconda
Smelter are contained in the Red Sands area and the Anaconda and Opportunity tailings ponds. These tailings
deposits, which have resulted from over 100 years of smelting operations, extend overan area exceeding 4,000
acres and are up to 90 feet in depth (6).
The texture of the tailings deposits decreased as more aggressive methods of ore processing developed,
allowing lower concentration ores to be economically processed. The chemical composition of tailings deposits
are as variable as the source of ore being mined at the time and the geochemical processes that have
secondarily altered the original matrix (7). Generally, higher concentrations of arsenic and other metals are
associated with older tailings from higher grade ores which were processed using less efficient extraction
technologies.
Flue dust is a smelting waste that results from collection of solid particulates before they are released to
the atmosphere via smokestacks. Flue dust represents one the highest concentration waste streams associated
with the Clark Fork Basin (8). Arsenic concentrations in flue dust generally exceed the 1 percent level (10,000
ppm) and may range as high as 15 percent. The Anaconda Smelter N PL site has more than 315,000 cubic yards
of flue dust in a variety of storage sites. , -
Treatment of mine timbers with arsenic became popular after it was discovered that timbers treated with
creosote posed a potential fire hazard in the underground mines. The Rockertimberf raming and treatment plant
(Figure 1) used arsenic produced at the Anaconda Smelter for treatment of mine timbers until 1957. During a
cleanup action in 1989 approximately 1,000 cubic yards of wood treating waste was removed that contained
arsenic concentrations in the range of 1 to 30 percent (9). The site is undergoing additional investigation to
determine the arsenic concentrations of waste remaining that must be remediated.
57
-------
Slag is the vitrified waste remaining following roasting and removal of the precious metals. On the
Anaconda Smelter NPL site 1.7 million cubic yards of heap roasting slag deposits remain (6). As with the mining/
smelting wastes, the heap roasting represents a low technology process applied to high concentration ores
resulting in high concentration wastes. In addition to the heap roasting slag, massive quantities of granulated
slag also remain on the Anaconda Smelter site (10).
During the 150 years of mining which occurred in the Clark Fork Basin, Silver Bow Creek, a stream which
forms the headwaters of the Clark Fork River, was historically used as an industrial sewer. During this long
history, it is estimated that 3to 10 million cubic yards of mining/smelting wastes were released to the river system
where they have mixed with stream sediments. This waste type has characteristic concentrations different than
the waste sources themselves.
REMEDIAL TECHNOLOGIES
Remediation efforts within the Clark Fork Basin have included removal/disposal of waste materials, partial
removal/addition of amendments/capping, removal, partial removal and capping, lime amendments, calcium
oxide and calcium hydroxide/lime amendments, deep plowing, injection and a variety of tilling procedures. An
overview of these technologies will be presented.
INNOVATIVE TECHNOLOGIES
Remedial project managers are required to use innovative technologies when there is a potential for
comparable or superior treatment performance or implementability, fewer (or lesser) impacts than other
approaches, or lower costs for similar levels of performance than demonstrated technologies. The availability
of innovative technologies formining wastes (containing arsenic) are limited. Congress recognized the need for
the further development of promising new treatment technologies for mining wastes when they appropriated
funds for the establishment of a pilot program for the treatment of mining waste (11).
This mine waste pilot program, operating under the direction of EPA's Risk Reduction Engineering Lab,
through the Department of Energy's Component Development and Integration Facility and the Montana College
of Mineral Science and Technology, has conducted a literature search of technologies that may have application
to mining waste. Mine waste types have also been researched.
Three mine waste types have been selected for demonstrating new technologies. These include: acid mine
drainage at a remote alpine site, in-situ treatment of near stream mining waste deposits, and treatment of acidic
waters such as those residing in the abandoned Berkeley open pit mine. An update of the status of the mine
waste pilot project will be provided.
REFERENCES
1. U.S. Environmental Protection Agency and Montana Department of Health and Environmental
Sciences. Clark Fork Superfund Sites Master Plan. 1990. 42 pp.
2. ARCO. Clark Fork River Superfund Site Investigations Laboratory Analytical Protocols.
3. Ashe Analytical and Morrison Knudsen Environmental Sciences. Clark Fork River Superfund Site
Investigation's Laboratory Analytical Protocol for X-ray Fluorescence Analysis of Solid Media. I.
Laboratory Grade Instrumentation Methods. 1992.11 pp.
4. Franklin, J.C. and Riley, J.A. Evaluation of Copper Mine Tailings Using EPA Leaching Procedures.
Montana State University Reclamation Research Unit Publication No. 9003. Fifth Billings Symposium
on Disturbed Land Rehabiliation Volume II. 1990. 389 pp.
5. Camp Dresser & McKee Inc. Draft Final Report Butte Soils Screening Study. 1988.
6. CH2M Hill. Final Work Plan for Anaconda Smelter RI/FS, Anaconda, Montana. 1984.194 pp.
7. TetraTech. Geochemistry Report. 1986. 445pp.
58
-------
8. Dames and Moore. Flue Dust Remedial Investigation Summary Report. 1989.140 pp.
/
9. Keystone Environmental Resources Inc. Work Plan Remedial Investigation Feasibility Study (RI/FS)
Rocker Timber Framing and Treating Plant Operable Unit. Rocker, Montana. 1991.
10. TetraTech. Granulated Slag Pile Draft Stage I Remedial Investigation Report. 1985. 84pp.
11. U.S. Environmental Protection Agency. Interagency Agreement between the U.S. Environmental
Protection Agency and the U.S. Department of Energy for the Pilot Program for the Treatment of
Mining Waste, Butte, Montana. 1992. 8 pp.
59
-------
SOLIDIFICATION/STABILIZATION OF ARSENIC COMPOUNDS
Frank K. Cartledge
Department of Chemistry
Louisiana State University
Baton Rouge, LA 70803-1804
USA
Tel: (504) 388-3459
INTRODUCTION
Solidification/stabilization (S/S), also known as chemical fixation or encapsulation, is widely applied
to waste streams and contaminated soils. The most common form of the technology uses a cement or
pozzolanic binder to convert the waste to a solid (if necessary) and depending on the constituents of the
waste stream and the binder, may reduce toxicity and/or water solubility of hazardous materials and may
create a monolithic waste form that limits contaminant mobility due to its low permeability and small surface
area.
The process is most commonly applied in cases where the contaminants of concern are heavy metals
in cationlc forms, e.g., Cd2+, Cr3*, Pb2+. However, applicability to a wide variety of waste materials including
As wastes has been proposed. The author is not aware of a commercial scale treatment where As
concentrations inthe wastestream were high; howeverthere have been remediations of contaminated soils
containing low concentrations of As and many bench-scale studies have been carried out.
PROCESS DESCRIPTION
Several detailed descriptions of S/S have appeared that include extensive discussions of the range
of applicability of the technology to a variety of waste streams and also descriptions of the equipment
available to carry out the process (1,2). Cementitious S/S is essentially similar to concrete production and
handling. The waste is combined with cement, water, sometimes aggregate material, and sometimes
admixtures of a number of types, and mixed under low shearing force. The product may resemble a concrete
slurry which is pourable and moldablefora period of several hours, and which can have appreciable strength
after curing. Alternatively, a soil-like solid material with little strength may be produced.
In cases for which S/S has been classified as a Best Demonstrated Available Technology (BOAT),
some reasonably clear guidelines exist from the US Environmental Protection Agency (EPA) specifying
performance criteria for the product. These include: a) the demonstration of no free liquids using the paint
filter and liquids release tests; b) development of a minimum strength (usually specified as 50 psi) as
evidence that cementing reactions have taken place; and c) meeting waste stream-specific leachability
requirements (3).
Depending upon the ultimate disposal of the treated wastes, additional performance criteria or more
stringent leachability requirements may be imposed by state or Federal regulatory agencies. For instance,
an extensive hydrogeologic characterization of the disposal site maybe required and may justify standards
of waste treatment that are closer to drinking water standards.
Binding agents that are commonly used are Portland cement, cement and fly ash, lime and fly ash,
and dust from cement kilns or lime kilns. The matrix that is formed in these cases is essentially the same
and consists of calcium silicate hydrates and smaller amounts of calcium aluminate hydrates. In cases
where lime is the main constituent, the curing process is somewhat different and involves absorption of
carbon dioxide from the air to form calcium carbonate as the principal matrix constituent. A variety of
additives is used in the process, including:
Standard cement additives - set accelerators or plasticizers
Clays - adsorb both metal ions and organics
60
-------
Silicates - enhance gel formation and precipitate metal ions
Silica fume - reduce permeability and increase strength
Emulsifiers - reduce interferences from organics
Iron salts - reducing agents for Cr(Vl) and form insoluble complexes
Sulfide or phosphate salts - form less soluble mixed salts
The cement- or pozzolan-based S/S process generally does not require sophisticated equipment or
extensive worker training. In contrast, blending the waste with liquid asphalt or bitumen is more energy
intensive, more complex from the point of view of volatilization of components of the waste, and somewhat
more expensive. This technology is not widely practiced.
APPLICATIONS TO ARSENIC COMPOUNDS
Arsenic wastes are grouped together for discussion of treatment standards in the land disposal
regulations promulgated by the EPA (4). Considerable discussion is allocated to S/S, since it was
considered by EPA to be a "potentially applicable technology." However, S/S is not currently considered
BOAT for any arsenic waste or wastewater. The following quote indicates the reasons:
"EPA has relatively inconclusive performance data for stabilization of arsenic in three different
wastes using nine different binders. Analysis of these data indicates that the effectiveness of any
particular stabilization binder appears to be highly dependent upon the waste types. This result
is what might be expected giving (sic) the chemical nature of arsenic and the relative sensitivity
of the effectiveness of stabilization processes with respect to the presence of organics and
organo-metallics."
Treatment standards include thefollowing:fornonwastewaters, 5.6 mg/Lor5.0mg/L in EP (or Toxicity
Characteristic Leaching Procedure—TCLP) leachates; for wastewaters, 5.0 or 0.79 mg/L in total compo-
sition. Clearly, EPA would like to establish the 0.79 mg/L standard forall As wastewaters, but it has received
many comments that such a stringent standard is not always attainable.
In specific language, EPA does not preclude the use of S/S fortreatment of As (particularly inorganic
As) wastes, but recommends that its use be determined on a case-by-case basis. Given the wide range of
chemical characteristics of As wastes, such a position is quite reasonable. Nevertheless, as a result of these
misgivings, there has not been a large-scale demonstration or remediation in which As concentrations in
the wastes have been high or for which As was the major toxic element of concern. On the other hand, there
are many results available from treatability or leaching studies on a laboratory scale, and these are
summarized in the following paragraphs.
Inorganic arsenic wastes have been most closely studied. Arsenic does not form an insoluble
hydroxide and hence the mechanism that operates during S/S of many heavy metal cations does not apply
for As. As(lll) is commonly observed in solution as AsO2-, and As(V) as AsO43-, and these oxo-anions are
typically water-soluble. Arsenic sulfides are-also commonly encountered in wastes, and these may also
have significant solubilities under the basic conditions typical of S/S. The leachate results from a variety of
As-containing wastes subjected to S/S with a number of binders have been summarized by Conner (1).
When the As concentrations in the waste are high (>635 mg/kg), leachate concentrations often can
be brought below 5 mg/L, but usually not below 0.79 mg/L. Treatment sometimes actually increases the
leachability, presumably because of different speciation under basic conditions. Similar results have been
reported for other TCLP leaching studies, and column and long-term equilibrium batch leaching experi-
ments indicate that As leaches at an elevated rate compared to most heavy metal cations.
In our work (refer to Table 1), both arsenate and arsenite have been solidified with a wide range of
binders using a weight percentage of 10% with respect to As, and the results range from entirely satisfactory
(arsenate using Portland cement and soluble sodium silicate or silica fume additives) to very unsatisfactory
(arsenate or arsenite using 1:1 Portland cement and type F fly ash).
61
-------
TABLE 1. TCLP LEACHATE CONCENTRATIONS FROM SOLIDIFIED WASTES:
1 YEAR (OR 28 DAYS) OF CURE
Leachate Concentration, mq/L
Binder'
As
As(V
Cr
OPC
OPC, no gypsum
1:1 OPC:Type F FA
20:1 OPC:SiOj
20:1 OPC:Na2Si03
10:1 OPC:Bentonite
10:1 OPC:Organoolay
20:1 Type IA:Si02
20:1 Type IA:NajSiO3
White
20:1 White:Na2SiO3
Lumnite
Refcon
Pyrament
35 (48)
43 (2.2)
0.9 (0.3)
25 (24)
28 (1.8)
46 (1.4)
40 (16)
41 (5.2)
35 (44)
54(1.0)
38 (6.0)
4 (400)
190 (680)
1.7
1.7(2.1)
3.1 (2.7)
430 (540)
2.3 (2.4)
2.0 (3.5)
2.4 (30)
1.1 (3.0)
2.5 (3.1)
1.6 (1.2)
0.9 (1.7)
1.0 (2.5)
240 (140)
1,90 (150)
47
1.4(1.7)
94 (7.9)
0.4 (0.1)
0.4 (0.1)
10 (3.6)
15 (8.0)
4.4
0.4 (0.4)
0.7 (0.2)
0.5 (0.2)
0.2 (0.2)
0.4 (0.2)
0.8 (0.2)
1.4
2500 (1400)
3200 (2400)
2900 (1600)
2500 (1600)
2500 (1800)
2400 (1700)
2800 (2400)
2800 (1900)
• 2700 (3300)
2900 (2000)
2800 (2000)
3700 (2300)
3600 (1900)
2800
- uro = Type I portland, FA = fly ash, SiO2 = silica fume, Na2Si03 = Type N soluble sodium silicate, White = portland
with low iron content. Lumnite and Refcon are specialty cements that are for refractory applications and high in alumina
content. The water to binder weight ratio is 0.5, and the metal to binder ratio is 0.1.
b The salts used are Pb(NO3)2, NaAsO2, Na2HAsO4'7H2O, Cr(NO3)3-9H20, and Na2Cr04'4H20.
Arsenate has been classified as a moderate retarder of cement and tricalcium silicate hydration
reactions based on calorimetry (5). We have investigated both arsenite and arsenate effects using solid-
state nuclear magnetic resonance spectroscopy and find that % hydration of the cement silicate phase is
significantly retarded by arsenate, and more so by arsenite. After 1 year of cure however, both kinds of
arsenic salts show only mild overall retardation in terms of % hydration, but both salts are very great
retarders of silicate polymerization. Thus, the silicate matrix is quite different in the presence of As(lll) and
We have periodically monitored our samples by powder X-ray diffraction in order to attempt to identify
crystalline phases that are forming as a result of waste addition. In most cases any new crystalline salts,
if formed, are present in such low concentrations that they are not evident in the already complex diffraction
pattern of the cement. This is not true with arsenates. A crystalline material is formed in major amounts,
NaCaAsC), »7.5H2O, and this species is present in both cement and cement/fly ash binders. Even when
As(lll) is solidified with cement, minor amounts of the same crystalline product are formed. The formation
of this salt depends upon the availability of significant amounts of Na as well as Ca, and in all cases we are
adding the arsenic compounds as Na salts.
While following the hydration reactions of As(lll) samples solidified in cement/fly ash, we noted an
initial normal, albeit slow, conversion of the aluminate phase from tetrahedral to octahedral. After 28 days
of cure, however, that conversion had begun to reverse itself, and after 14 months almost all the octahedral
Al had reverted to tetrahedral Al. Furthermore, there is some indication that the silicate phase is undergoing
depolymerization. The spectral changes leading to the latter conclusion are still within the range of
experimental error(±4%),butthey clearly indicate the necessity to monitorthese changes over even longer
periods of time.
After seeing the effects noted above in the As(III)/OPC/FA mixture, we quickly ran some OPC/FA
control samples containing no waste. We see similar effects in both the aluminum and silicon spectra,
although the extent of reversion is lower at 1 year in the absence of As. Both As(lll) and As(V) appear to
catalyze the aluminate phase changes. Such long-term effects that alter the matrix have serious
consequences for the application of leach modelling to long-term predictions of release rates for contami-
nants. The As(V) samples, for instance, show clearly enhanced leachability after 1 year of cure compared
62
-------
to 28 days (Table 1). We are continuing to monitoirthese samples, and will have additional data after2 years
of cure to report in August.
S/S of As-containing waste streams clearly shows promise as an effective treatment technology.
However, significant questions about the range of applicability remain. Furthermore As salts induce
changes in the S/S binding matrixthat first become evident after long cure times, suggesting that teachability
testing will have to be carried out after much longer cure times than is currently practiced for S/S.
REFERENCES
1. Conner, J.R. Chemical Fixation and Solidification of Hazardous Wastes. Van Nostrand Reinhold,
New York, 1990.
2. Barth, EIF. et al. Stabilization and Solidification of Hazardous Wastes. Noyes Data Corp., Park
Ridge, NJ, 1990.
3. 40 CFR268.10-268.12.
4. Federal Register, 55, 22556-22561; June 1,1990.
5. Thomas, N.L. Corrosion problems in reinforced concrete: why accelerators of cement hydration
usually promote corrosion of steel. J. Mater. Sci. 22: 3328,1987.
63
-------
VITRIFICATION OF WASTE STREAMS CONTAINING RCRA METAL COMPOUNDS
J. G. Hnat, John Patten, and Christopher Jian
Vortec Corporation
3770 Ridge Pike
Collegeville, PA 19426-3158
USA
Tel: (215) 489-2255
INTRODUCTION
Abstract
Vortec Corporation has been developing an Advanced Combustion and Melting System (CMS) which
oxidizes the organic constituents of a hazardous waste and chemically combines the heavy metal
components into a glass matrix. This paper summarizes Vortec's ongoing research work on the vitrification
of Resource Conservation and Recovery Act (RCRA) metal containing wastes. Both laboratory analyses
and pilot scale testing are being conducted. The work has been supported by Department of Energy (DOE)
and Environmental Protection Agency (EPA) as well as by internal Vortec research funds. A significant
portion of Vortec's industrial process research is devoted to the development of innovative approaches to
waste recycling and the vitrification of organic and heavy metal contaminated waste materials.
Previous Research
The development of the Vortec Combustion and Melting System (CMS) initially concentrated on the
melting of glass batch materials with the objective of improving the efficiency of glass manufacturing
operations. However, the application of the technology to vitrifying waste materials was soon suggested.
Vortec has recently processed several industrial waste materials containing high concentrations of heavy
metals including arsenic. The preliminary data suggest that a portion of the heavy metals are being retained
in the glass matrix. Further work is being conducted by Vortec (the SITE program) to identify the partitioning
of the metals as they are processed accurately.
Glass manufacturing operations in the past have added approximately 3000 ppm of As203 (arsenic
trioxide), as an agent to assist in the removal of bubbles formed during the manufacture of glass materials.
In general, the removal of thebubbles (called fining) produqesasuperiorquality product that is aesthetically
pleasing to the customer. The mechanism by which the bubbles are removed involves the evolution of
oxygen from the arsenic compound, at elevated temperatures, with the subsequent migration of theoxygen
into the already existing bubbles. As the bubble diameter increases, the buoyant force driving the bubble
to the surface of the glass melt is increased and the time at temperature required to remove the bubbles
is reduced.
The use of arsenic as a fining agent in glass is obsolete. Less expensive combinations of compounds
are now being used. However there is a body of research on the behavior of arsenic trioxide in glass melts.
These data would suggest that as much as 90% of the arsenic compound could be retained in the glass,
and usual practice would expect anywhere from 50 to 90% to be retained depending on the temperature
at which the melter is being operated and the type of glass being manufactured.
Brief descriptions of previous research programs that have direct application to the remediation of
materials containing RCRA metals are presented in following paragraphs. Characterization of the waste
streams for a selected number of the materials are presented along with typical Toxic Characteristic
Leaching Procedure (TCLP) results. Each waste stream has significant variations in composition and
successful vitrification will require the addition of different amount and composition of glassmaking
components.
Vortec is developing a multi-fuel capable advanced high temperature process heater for melting and
other high temperature industrial applications, based on patents previously developed. Vortec designed,
fabricated, and installed a nominal 15 to 20 ton/day advanced process heater test loop in Harmarville, PA.
64
-------
The system has been used to test the melting of waste glass, waste fiberglass contaminated with organics,
utility flyash, and waste dust. This program is a DOE-supported effort underthe Advanced Coal-Fired Glass
Melting Process Heater Program, DOE Contract DE-AC22-91PC91161.
Vortec Corporation is the prime contractor on a DOE SBIR program to demonstrate the application of
Vortec's CMS, using coal as a fuel, to waste fiberglass recycling. Phase I of the program demonstrated the
feasibility of oxidizing the organic contaminants in waste insulation fiberglass material, and produced a
cullet suitable for recycling to a manufacturer's furnace.
Vortec has completed Phase I of an EPA SBIR contract to demonstrate the feasibility of using the CMS
with natural gas as the primary fuel, to oxidize the organic contaminants in waste insulation fiberglass
material, and produce a cullet suitable for recycling to a manufacturer's furnace. The program is currently
in Phase II, the objective of which is to characterize effluents from the system to assure environmental
compliance. The Hazardous Flyash and Industrial Process Dust Vitrification project, also for EPA, has as
its objective the assessment and feasibility demonstration of vitrifying municipal solid waste incinerator
flyash in the CMS to produce products that can pass TCLP, thus helping to recycle this material into value
added products.
Vortec is currently conducting a SITE program for EPA. The objective of this program is to verify the
effectiveness of the CMS for the oxidation and vitrification of contaminated soils in dry or slurry form that
are representative of materials found at EPA Superfund sites.
Materials Previously Processed
The CMS has successfully demonstrated the vitrification of several different classes of materials.
Postconsumer waste glass has been processed.in the form of commercially obtained glass cullet. During
the startup phase of the development program, glass cullet was the feed material that was the easiest and,
most uniform in physical properties, simplifying its melting and control. Subsequently, waste insulation
fiberglass, utility flyash, municipal solid waste incinerator flyash, aluminum processing dust, and sewage
sludge incinerator ash have all been successfully processed to form a vitrified product capable of passing
the TCLP test. Vortec is preparing to conduct further testing to demonstrate the CMS's capability to vitrify
additional contaminated soils and sludges, auto shredderfluff, and filtercakes of various compositions from
the chemical processing industry.
PROCESS DESCRIPTION
System Description
t
Vitrification tests are being conducted using Vortec's CMS located at the University of Pittsburgh
Applied Research Center in Harmarville, PA. An isometric sketch of the pilot plant test system arrangement
is shown in Figure 1. The major components in the test system are a counter-rotating vortex (CRV)
combustor/suspension preheater, a cyclone melter, a glass separator/reservoir, material storage facilities,
a pneumatic feed assembly, a cullet quench assembly, and a flue gas scrubber assembly. Vortec's test
facility also contains supporting subsystems and assemblies forcoolingwatersupply, combustion airsupply
and preheating, instrument and service air supply, natural gas and coal fuel supply, flue gas cleanup, and
system control. Descriptions of the major components in the pilot scale test system are presented in
following paragraphs.
f
MAJOR COMPONENT DESCRIPTIONS
Counter-Rotating Vortex Combustor/Preheater
The CRV combustor is designed to provide rapid suspension preheating of the waste materials to their
melting temperature. This is accomplished via convective and radiative heating of the feedstock materials
in the combustion zone of the CRV combustor.
\ . .
65
-------
SEPARATION CHAMBER
GAS SEAL
GULLET CART
Figure 1. Illustration of pilot scale cyclone melting system.
A variety of fossil fuels can be utilized in the CRV combustor, including pulverized coal, oil, and gas.
Inasmuch as the feedstocks are preheated in suspension and turbulently mixed with the combustion air
introduced into the combustor, the process also provides an effective means of incinerating any organic or
other combustibles in the waste materials. In most applications, fuel and waste along with other appropriate
ingredients can be introduced axially along the centerline of the CRV combustor.
Combustion air is introduced tangentially through two inlet arms such that the direction of rotation of
the air streams are opposed to each other. The CRV preheater assembly provides a region of intense
turbulence in the upper section of the CRV combustor for good flame stability, followed by a region of plug
flow at the exit of the CRV combustor.
The CRV combustor/preheater is refractory lined and water cooled with the water temperatures and
refractory temperatures continuously monitored. Typical reactor temperatures in the CRV combustor vary
between 1200K (1700 degrees F) and 1900K (2960 degrees F) depending on the type of feedstock being
utilized. The nominal air preheat temperature for the experiments is 870K (1100 degrees F). In spite of the
relatively high air preheat temperature utilized, NOx levels are low, typically less than 200 ppm, due to the
rapid temperature quenching by the feedstock materials.
Cyclone Melterand Separation/Reservoir Assembly
Combustion products and preheated batch particulate material exiting the CRV combustor/preheater
are introduced tangentially into the cyclone melter where melting occurs and the molten glass is separated
centrifugally from the gases. The cyclone melter is refractory lined and water cooled with water tempera-
tures continuously monitored. The molten glass and combustion gases exit tangentially from the cyclone
melter and enter a rectangular, refractory lined separator/reservoir where the vitrified product (glass) is
removed from the system. The combustion gases are directed to a flue gas quench/scrubber assembly, and
the molten vitrified product exits a tap hole on the bottom of the separator/reservoir.
Theseparator/reservoirassembly also contains an overflowpipe that interfaces with acullet collection
cart in the event that the glass flow entering the separation chamber exceeds the flow capability of the tap
66
-------
hole. The molten glass exiting the tap hole in the glass separator/reservoir assembly falls into a chute where
water is introduced to quench the glass and transport the quenched glass (cullet) into a cullet collection cart.
Flue Gas Quench/Scrubber Assembly
The gas quench assembly consists of a channel connected to the discharge of the glass separator/
reservoir assembly with waterspray nozzles forquenching the exhaust gases from nominally 1700 degrees
K (2600 degrees F) to 365 degrees K (197 degrees F). The flow passes to a scrubber and is then discharged
to the atmosphere by an induced draft fan.
Pulverized Material Storage and Pneumatic Feed Assembly
Pulverized batch materials are stored in a4.25 cubic meter (150cubicfoot) pneumatictransport vessel
which is supported on weigh cells to determine feed rate during operation. A fugitive dust control system
creates a negative pressure at the vessel interface and draws d usty air throug h a fabric filter located outside
the test building, exhausting clean air to the environment.
Mixing is accomplished in the vessel via a pneumatic system designed for this particular application.
A calibrated rotary feeder meters material from the storage vessel to the process. The material is
pneumatically transported to the process at solids-to-air weight ratios up to 60-to-1.
STREAM CHARACTERISTICS
Typical Feed Stream Compositions
Vortec has been conducting demonstration testing using its CMS facility at U-PARC. These tests
typically process 1000 to 1500 Ibs per hourof batch composed of a combination of waste and glass making
additives. Table 1 presents the RCRA metals content of several waste streams that are being investigated
using the Vortec CMS. These wastes, in some cases, contain heavy metals at concentrations that make the
passing of the TCLP uncertain.
Vortec is conducting a SITE demonstration test in which a clean surrogate soil will be spiked with
oxides and salts of the four RCRA heavy metals as indicated in Table 1. This soil will be mixed with additives
to produce a batch composition satisfactory for the formation of a vitreous product. Extensive testing will
be conducted during the test to establish the partitioning of the metals arsenic, cadmium, chromium, and
lead.
As required by the SITE Q/A plan, the batch material entering the system, the effluent water leaving
the air pollution control device, and the flue gases will all be sampled and tested for these heavy metals.
In addition, the glass stream will be tested for total metals and also to determine if the vitrified product will
pass TCLP.
Product Characteristics
As indicated in Table 1, six waste streams are being investigated to determine if the waste can be
vitrified and to establish that the vitrified product will pass the TCLP test. As indicated in the table, four of
the six vitrified products tested have passed TCLP even though the untreated waste has concentration of
heavy metals several orders above the limits.
1992.
The other two wastes, spent potliners and EPA surrogate soil, are scheduled for CMS testing later in
67
-------
FACILITY DESCRIPTION
Vortec CMS Pilot Plant Facilities
Vortec's advanced combustion/melting test system is installed in the physical plant facilities at the
University of Pittsburgh Applied Research Center (U-PARC),Harmarville, PA. Vortec has the exclusiveuse
of the High Bay Area within Building B11, which has plan dimensions of 40 x 100 ft, and a height of 64 ft
This area includes a towerfor support of test equipment, and a 5-ton bridge crane. The tower is a structural
steel frame with four elevations above grade, extending to a height of 43-1/4 ft and plan dimensions of 15
x 40 ft. The crane support is arranged so that crane service is provided to the full plan area of the high bay.
Truck access to the test area is provided by a 15 ft high x 15-1/3 ft wide roll-up door. Except for the
slurry preparation and feeding subsystem, the facility has all the equipment and utilities necessary to
conduct this research and development program.
TABLE 1. WASTE STREAM CHARACTERIZATION
Elements
Arsenic
Barium
Cadmium
Chromium
Load
Mercury
Selenium
Silver
Contaminated
PPM
540
NR
30
52
380
6
50
606
Soil
Spent Potliners
Average
PPM
13
100
2.5
16
22
25
0.5
1
EPA Surrogate
Soil
Vortec SITE Test
PPM
200
Nl
200
1000
2500
Nl
Nl
Nl
Waste Water
Treatment
Incinerator Ash
PPM
27
390
10
610
860
<.01
17
4.46
MSW Flyash
PPM
981
984
57
601
3433
<200
2
<30
Hazardous
Baghouse Dust
PPM
N=!
NR
523
957
13641
NR
NR
NR
Typical TCLP Results from Vitrified Product
Arsenic
Barium
Cadmium
Chromium
Load
Mercury
Slenium
Silver
0.9
0.14
0.001
0.02
<0.003
ND
NR
NR
ND-Not Detected
NI-Not Included
NR-Not Reportred
Test Scheduled
For Late 1992
Test To Be
Conducted
June 17,1992
0.05
<1.2
<0.07
<0.10
0.5
<0.0003
<0.05
<0.01
<0.05
<1.0
0.05
0.1
3.3
0.0093
<0.050
<0.02
<0.5
<0.05
0.2
1.5
ND
<0.025
ND
68
-------
ELEMENTAL MERCURY IN SOIL AND THE SUBSURFACE: TRANSFORMATIONS
AND ENVIRONMENTAL TRANSPORT
Ralph R. Turner -
Environmental Sciences Division
Oak Ridge National Laboratory1 , «
Oak Ridge, Tennessee 37831-6036
Disposals and spills of both small and large quantities of elemental mercury in terrestrial environments
have been a common occurrence. The subsequent behavior and environmental fate of such mercury is
controlled by a host of factors, not all of which are fully understood at this time. The physical properties and
solubilities of elemental mercury and its oxidation products can be used to predict how disposals and spills
will behave, but the range of the environmental settings is so large that relying on theoretical predictions
alone can be risky. Important physical properties of elemental mercury to consider are its liquid state, high
density, high surface tension, and volatility.
When spilled onto porous media the liquid state and high density will tend to drive the mercury ever
deeper. This tendency is counteracted somewhat by the high surface tension of the metal which favors
"beading" and trapping in pore space. Although much less volatile than water, elemental mercury does
evaporate sufficiently to create a significant vapor pressure in surrounding airand may migrate significantly
through porous media via this mechanism. While the solubility of elemental mercury is relatively low, the
metal is subject to oxidation to far more soluble compounds. However, solubility of pure compounds is not
a very good indicator of mobility in the environment.
Soluble species of mercury exhibit a very high affinity for solids which often bind the mercury so tightly
that only destruction of the solid will release the mercury. In addition, soluble mercury can react with other
compounds to form secondary compounds, e.g., sulfides, under certain conditions which are far more
insoluble than the original metal. Under some conditions a small fraction of inorganic mercury in soils may
be converted to alkylated species (e.g., monomethyl and dimethyl). The latter process is only poorly
understood at this time but is likely to be important in the evaluation of human and ecological risk.
Experience in characterizing and cleaning up several spills of elemental mercury at a nuclear weapons
plant in Oak Ridge, Tennessee is useful to review. Early attempts to recover quantitatively large ( 50 tons)
spills of the metal onto soil or unsealed floors usually met with failure due both to the rapid percolation into
every available pore and fissure. Attempts at excavation created new fissures which simply drove the
mercury deeper. Smaller (fewpounds)spillstendedto remain in theuppersoilprofileasfinelydividedbeads
which tended to coalesce if the profile was disturbed by sampling or excavation.
Soij borings at former spill sites have not generally detected very much mercury perhaps reflecting
operation of the "nugget effect" in these explorations for spilled mercury. With a few exceptions,
groundwater monitoring has likewise failed to show any consistent relationship between spill sites and the
groundwater concentrations. Inclusion of suspended matter in groundwater analyses has seemed to
account for most of the elevated mercury concentrations. Results of US Environmentaf Protection Agency
(EPA) standardized leaching protocols (EP, TCLP) applied to soils from the site have shown no relationship
between total mercury concentration in the soil (ug/g) and performance on the leaching tests (mg/L).
Variability in the forms of mercury present in the soils has been shown to account for this observation.
These experiences have suggested that better methods of characterizing spill sites are needed and
that availability of physical and chemical speciation information is essential (1) to fully characterize sites,
(2) to conduct a realistic risk assessment, and (3) to select the most appropriate corrective action.
1Oak Ridge National Laboratory is managed for the US Department of Energy by Martin Marietta
Energy Systems, Inc. under contract No. DE-AC05-84OR21400.
69
-------
RESEARCH PROGRAM FOR DEALING WITH MERCURY IN SOIL AT NATURAL GAS
INDUSTRY SITES
David S. Charlton, Craig R. Schmit, Daniel J. Stepan, and Frank W. Beaver
Energy and Environmental Research Center
University of North Dakota
Grand Forks, ND 58202
USA
Tel: (701) 777-5000
James M. Evans
Environmental and Health Research
Gas Research Institute
8600 West Bryn Mawr Avenue
Chicago, IL 60631
USA
Tel: (312) 399-8329
INTRODUCTION
The past use of elemental mercury in monitoring and control instrumentation at natural gas industry sites
has resulted in soils contaminated by mercury. The most common types of mercury-filled instrumentation in the
gasindustry are flowmeters (manometers) that are used to measure the volumeof gas flowing through metering
stations.
The Energy and Environmental Research Center (EERC) has been contracted by the Gas Research
Institute (GRI), with support from the US Department of Energy (DOE), to investigate potential issues related
toandalternativeremediationtechnologiesforelemental mercury spills. The investigation of elemental mercury
spills is an outgrowth of a broader research program focused on issues related to a variety of potential
contaminants in gas industry wastes and products.
SOURCES OF MERCURY CONTAMINATION
Mercury-filled flowmeters have been the traditional design used in the gas industryforthe past half century
Even though mercury flowmeters are being replaced with newer designs that use no mercury, a number of
meters with elemental mercury still exist. Flowmeters may contain 8 to 10 pounds of mercury. Soils in the
immediate area of some mercury flowmeters have become contaminated with mercury due to leakage, spills,
equipment failure, vandalism, and operator error. A better understanding of these sources of mercury
contamination has resulted in more effective containment systems and improved management practices,
resulting in a drastic reduction in mercury reaching the soil surface.
However gas metering sites remain that were contaminated in the past with elemental mercury. Because
of thesmall amounts of elemental mercury involved, most of these sites probably have relatively small volumes
of mercury-contaminated soil, perhaps one to two cubic yards per site. Nevertheless, the gas industry is
interested in the most efficient, cost-effective, and feasible technologies for remediating these sites.
ENVIRONMENTAL REGULATIONS
Mercury became a concern within the gas industry first as a worker safety issue. Some gas companies
voluntarily began cleaning up their sites because of this issue. Site cleanups resulted in mercury-contaminated
soilsthat were sentto hazardous waste landfillsfordisposal. Of particular concern, therefore, to thesgas industry
has been the May 8,1992 deadline underthe Land Disposal Restrictions (LDR), also called "Land Ban After
that deadline, certain types of mercury-contaminated wastes can no longer be sent to hazardous waste landfills
without meeting specific treatment requirements.
70
-------
ForD009 wastes with total mercury concentrations greater than 260 mg/kg, which are typical of many gas
metering sites, the required Best Demonstrated Available Technology (BOAT) is roasting or retorting, with
condensation and recovery of the volatilized mercury. Within the gas industry, there are concerns that roasting
or retorting may not be the most economical and efficient technology for dealing with mercury-contaminated
soils and that the national capacity does not exist to deal with the potential volume of contaminated soils that
could require treatment.
The regulations do provide mechanisms for addressing these two concerns. One available mechanism is
to apply fora treatability variance when the BOAT is inappropriate for the waste. Another available mechanism
is to apply for a national capacity variance when it can be demonstrated that there is insufficient capacity
nationwide to handle the volume requiring treatment. However, both of these mechanisms can cause
substantial delays in initiating a remediation effort.
An additional regulatory consideration is that individual states can impose more restrictive corrective
action levels than those required at the Federal level. As an example, a number of states currently require
cleanup to background levels, determined on a case-by-case basis.
The regulatory issues discussed above are primarily those affected by the Resource Conservation and
Recovery Act (RCRA) and its amendments. However, there are other Federal programs and regulations that
also affect the cleanup of mercury-contaminated soils, including those related to the Occupational Safety and
Health Administration (OSH A); the Comprehensive Environmental Response, Compensation, and Liability Act
(CERCLA, "Superfund"); and the Clean Air Act Amendments.
MERCURY RESEARCH PROGRAM
To addressjhe potential environmental impacts of elemental mercury at gas industry sites in the context
of a changing regulatory framework, an interdisciplinary research program was designed and implemented.
Specifically, within the context of that program:
A mercury workshop was organized and conducted forthe gas industry at GRI headquarters in Chicago,
Illinois, on February 10-11, 1992. A total of 62 people, representing the gas industry, the research
community, and remediation companies, participated in the workshop. A proceedings volume will be
available in the near future (1).
A literature review was conducted and a bibliographic database was constructed to identify relevant
mercury- related literature. A summary report is being written based upon key references (2).
Computer models are being assessed for their applicability to mercury migration in the subsurface.
A risk assessment model is being developed for mercury- contaminated sites. Risk models can be used
as mechanisms for setting site-specific cleanup levels or they can be used in ranking sites for cleanup.
Sampling, preservation, and analytical protocolsfor mercury-contaminated solidsand liquids are being
evaluated and developed as necessary. .j
Laboratory experiments are being conducted evaluating: 1) leaching techniques applied to mercury-
contaminated materials and 2) mercury speciation in subsurface environments.
Mercury-contaminated gas industry field sites are being characterized and instrumented in several
; parts of the country. Sites have been selected to represent a range of site-specific variables.
Existing and developing remedial technologies are being reviewed to deal with mercury in soils. A
summary report is being written based upon this review (3). In addition, one remedial technology has
actually been tested, and others will be tested in response to a recently circulated Request for Proposal
(RFP).
REMEDIATION OPTIONS
A major goal of this research program is to develop an understanding of mercury contamination at gas
industry sites so that the impacted materials can be remediated in.an efficient and cost-effective manner. To
evaluate the range of remediation options available to the gas industry, information was solicited from
companies and researchers that have available or are developing remedial technologies for mercury-
71
-------
contaminated soils. The remediation options were grouped into six categories: 1) chemical treatment, 2)
physical treatment, 3) thermal treatment, 4) biological treatment, 5) immobilization, and 6) electrolytic treatment.
A technology based on one of these six remediation options, physical separation, was tested using native
soil fromthe southwestern USspiked with elemental mercury. This particular remedial technology was selected
fortesting because a unit was already underdevelopment as a mobile prototype unit and preliminary information
suggested thatthe unit, based upon established mining technologies, might be effective in separating elemental
mercury from soil.
Other technologies will be selected for development and testing based upon an evaluation of proposals
solicited through the recent circulation of an RFP. That RFP was sent to companies and researchers'known to
have experience in dealing with elemental mercury or related contamination.
SUMMARY
This ongoing research program has been designed to address a range of key issues regarding elemental
mercury contamination at gas industry sites. An integrated, multidisciplinary research approach is being used
to develop a better understanding of the complex interactions that can occur between mercury and a range of
variables at the sites. A better understanding of these complex interactions is crucial to the design and
evaluation of remedial technologies that are suitable for application by the gas industry where mercury
contamination has occurred.
REFERENCES . .
1. Charlton, D.S. and Harju, J.A. (eds.). la: Proceedings of the Workshop on Mercury Contamination at
Natural Gas Industry Sites. Gas Research Institute, Chicago, Illinois, 1992. p. 217.
2. Henke, K.R.andSchmit, C.R. Review of mercury-related topics with relevance to the issue of mercury
contamination at gas industry metering sites. Gas Research Institute, Chicago, Illinois. (In prepara-
tion.)
3. Stepan, D.J. and Gust, H. Evaluation of remediation technologies applicable to mercury-contamina-
tion issues of the natural gas industry. Gas Research Institute, Chicago, Illinois. (In preparation.)
72
-------
MERCURY CONTAINING HAZARDOUS WASTES: GENERATION AND POTENTIAL
REDUCTION
Wayne Brooke-Devlin
Versar Inc.
6850 Versar Center, PO Box 1549
Springfield, VA 22151
USA
Tel: (703) 750-3000 x312
In 1983 (the last year for which figures were available), US industries consumed the following
percentages of that year's mercury supply (1):
Alkaline batteries 48%
Chlorine and caustic soda ...16%
' Electrical and instruments 13%
Paints and other 23%
Recognizing that these industries produce mercury-containing waste, the US Environmental Protec-
tion Agency (EPA) has characterized the following industrial wastes and promulgated Best Demonstrated
Available Technology (BOAT) standards for their treatment prior to land disposal:
D009 - Any waste that is characteristically hazardous based on the concentration of mercury in the
leachate as determined by the Toxicity Characteristic Leaching Procedure (TCLP)
P065 - Mercury fulminate
P092 - Phenylmercuric acetate
U151 -Mercury
K071 - Brine purification muds from the mercury cell process in chlorine production, where
separately prepurified brine is not used
K106 - Wastewater treatment sludge from the mercury cell process in chlorine production
MERCURY CELL BATTERIES
The largest single use of mercury is in the production of batteries. This industry utilizes mercury and
several mercury compounds in the manufacture of mercury cell batteries.
During battery manufacturing, mercury-containing wastewaters can result from the cleanup of metallic
mercury and mercury compound spills, or from washes of process equipment. Mercury- containing
nonwastewaters can result from off-specification product, from spills of mercury metal or mercury
compounds, and from wastewater sludges generated in the treatment of mercury containing wastewaters.
These wastes will most likely be categorized as D009 and/or U151 code wastes and will be treated to the
required mercury levels (0.030 mg/L (total) for U151 wastewaters, 0.20 mg/L (total) for D009 wastewaters,
0.20 mg/L (TCLP) for D009 nonwastewaters and U151 retorting residues, and 0.025 mg/L (TCLP) for U151
nonwastewaters that are not retorting residues) through the suggested BOAT methods (sulfate precipitation
for wastewaters; roasting or retorting for nonwastewaters).
In 1975the EPA estimated that battery production, which accounted for 39% of domestic mercury use
that year, contributed less than 0.9% of that year's quantity of mercury iost to the environment through
industrial waste (2). Now, with the adoption of BOAT treatment levels, it is probable that the percentage of
industrial mercury lost to the environment through battery production waste streams is even smaller. In
industries where small on-site spills, periodic equipment washes, non-standard product, and wastewater
73
-------
sludges are the major sources of waste, it is not surprising that only small amounts of mercury-containing
wastes are produced. When treated to BOAT levels the amount of waste due to battery manufacturing is
not expected to be significant.
The adoption of nickel-cadmium or lithium battery cells would eliminate mercury cell wastes. However,
though both types of cell are in production, they have not yet replaced mercury-containing batteries.
In 1983 something on the order of 3,700,000 pounds of mercury entered the environment via end-user
battery disposal (3). EPA figures show that in 1973 about 5% of battery mercury was being recycled. The
significance of waste from the mercury using industries pales in comparison to the waste from disposable
batteries. Increased battery recycling has the potential to reduce overall mercury wastes greatly.
CHLORINE/CAUSTIC SODA PRODUCTION VIA MERCURY CELLS
The second largest use of mercury is in the production of caustic soda and chlorine. This industry
employs a flowing stream of mercury as the cathode terminal in an electrolytic decomposition cell.
In caustic soda/chlorine manufacturing, mercury-contaminated wastewaters can be produced during
the washing of cell room floors and end boxes and from direct contact cooling of the hydrogen stream.
Mercury-containing nonwastewaters can be produced from wastewater treatment and the precipitation of
brine purification solids and from retort ash. The EPA has identified K071 and K106 as two types of wastes
which result from caustic soda production.
As wastewaters and nonwastewaters the K106 waste code covers many of the mercury-contaminated
wastes from mercury cell caustic soda plants. The K071 code applies to plants which do not use prepurified
salt and must therefore include a solids precipitation step before allowing brine solution to enterthe mercury
cell. Since some of this brine is made up from a recycle stream of spent brine from the mercury cell, some
mercury may also precipitate in this step, thus contaminating the precipitation solids.
BOAT calls for the treatment of K071 wastes by acid leaching since their low concentrations (25 to 120
mg/kg) of mercury make retorting difficult. Under BOAT, K106 wastewaters are to be treated by precipitation
using sulfate or hydrazine followed by filtration to reduce their total mercury concentrations to 0.030 mg/L.
The K106 nonwastewaters are to be retorted until their mercury concentrations are below 0.20 mg/L TCLP
for wastes which initially contained concentrations above 260 mg/kg, and below 0.025 mg/L TCLP for
wastes which initially contained a lesser concentration of mercury.
In 1975 the EPA estimated that of the 449,608 kg of mercury used in the caustic soda industry that
year about 229,760 kg (51.1 %) was lost to the environment as wastes. In 1980 the EPA estimated that of
the 325,768 kg of mercury used by the caustic soda industry that year approximately 154,000 kg (47.3%)
was lost as waste. Though the 3.8% decrease in the ratio of mercury used to mercury waste during this
period is not great, the 28% decrease in total mercury use and the resulting 33% decrease in overall mercury
loss is very significant.
During this time period the percentage of total caustic soda production capacity provided by mercury
cell plants fell from 25% to 19% (4). This trend continued through the 80s. In 1980 there were 27 mercury
cell plants in operation. In 1984 there were 24, and in 1990 the EPA estimated that only 20 mercury cell
plants remained in operation. It is expected that with the passage of time the remaining facilities will also
be phased out. However, even without this gradual attrition, immediate reductions in mercury waste are
expected to occur due to the application of BOAT treatment standards.
Two alternatives to mercury cell technology are currently available: diaphragm cell and membrane
cell. Neither of these technologies produces a caustic with as high a degree of purity as that from mercury
cells. Nevertheless, all new US installations during the last 20 years have been of one of these two types,
which indicates the industry's desire to abandon the potentially hazardous mercury technology.
ELECTRICAL AND INSTRUMENTATION USES
Mercury's high density, volatility, low electrical resistivity, and other physical properties make it an
74
-------
ideal material for a number of electrical and instrumental uses. Mercury has found use as the contact
material in some types of electrical switches, as the conductor in a number of electrical devices, and as the
vapor in some gaseous discharge lamps. It has also been widely employed in the manufacture of
thermometers, barometers, and other pressure/temperature sensing devices (5).
Due to the wide variety of manufacturing processes involved, it is difficult to determinethecontribution
of the electrical and instrumental uses of mercuiy tothe problem of mercury-contaminated waste. However,
it is probable that in these manufacturing industries (as in battery manufacturing) such wastes occur
primarily as a result of spills and process equipment washes. Therefore, electrical and instrument
production wastes are likely to be of the same types (D009 and U151) and on the same order as battery
wastes, and when treated to the previously-described BOAT standards forthese wastes will constitute only
a small fraction of annual mercury wastes.
Though still used, mercury switches and devices have largely been replaced by electronic equivalents.
Electronic devices have also to some degree replaced both mercury thermometers and barometers.
Nevertheless, the increasing demand for outdoor security lighting and the greater accuracy of mercury-
based pressure and temperature devices ensure that mercury-containing electrical equipment and devices
will remain in production for the foreseeable future. Fortunately, end use recycling of mercury from these
products is not difficult. In 1975 the EPA estimated that of the 191,000 kg of mercury in electrical and
instrumental end products approximately 82,700 kg (43%) was recycled. Again, improved recycling could
greatly reduce the total waste from these sources.
PAJNTS AND OTHER SOURCES
Phenylmercuric acetate (P092) and phenylmercuric oleate when used as a bactericide in stored latex
paint or as a fungicide in applied latex paint accounted for slightly more than 12% of the mercury used in
1983. In 1990 the EPA found that these compounds were still being manufactured and used as
preservatives in latex paints. In paint formulation operations wastes are likely to be mainly due to spills and
equipment washes. Therefore, these wastes are likely to be on the same order as those from battery
production.
BOAT treatment of P092 wastewaters calls for oxidation of the organic mercury compound followed
by precipitation and filtering to yield a wastewater with a total concentration of less than 0.030 mg/L The
P092 nonwastewaters are to be treated to mercury concentrations of either 0.20 or 0.025 mg/L TCLP
(depending upon whethertheir initial concentration exceeded 260 mg/kg) through incineration or retorting.
The small expected amount of waste and the applicable BOAT standards are likely to make the contribution
of paint production to mercury wastes minimal. Recycling of used paint is not considered a viable waste
reduction method.
Dental amalgam represents a further 3% of the mercury used in this category. Wastes from this source
are difficult to assess but are expected to decline as composite materials gain greater acceptance in the
dental profession. In the past, recycling of recovered amalgam has been excellent due to the value of the
material. The application of D009 and U151 are expected to improve this recovery in the future.The
remaining 8% of annual mercury consumption has in the past been used for the production of explosives,
pigments, Pharmaceuticals, and chemical catalysts.
In recent years mercury fulminate (P065) has been eliminated from explosives manufacturing due to
its instability. (Though BOAT levels have been established for this compound, they find no current
application.) Where possible mercury compounds have been replaced by other compounds in pigments,
and by less toxic compounds in medicine. Their use in the catalytic manufacturing of plastics has largely
been supplanted by alternate processes. The contribution of these sources to mercury wastes is likely to
continue to decrease. The wastes from these sources are expected to be minimized through application of
D009andU151 BDATs.
In conclusion, recycling (or replacement) of mercury batteries, and further reductions in the use of
mercury cells in the production of caustic soda hold the greatest promise for significantly reducing mercury
wastes. Based on current trends the number of mercury cell chlor-alkali plants are expected to decrease
75
-------
with time. The wastes from these and other mercury-using industries are expected to be diminished by
adoption of BOAT treatment limits. The future of battery recycling or replacement efforts has yet to be
determined.
REFERENCES
1. U.S. Bureau of Mines. Mercury. Im Mineral Facts and Problems. Preprint from Bulletin 675. U.S.
Department of the Interior, Washington, DC, 1985.
2. DeCarlo, V. J. Multimedia Levels Mercury. EPA-560/6-77-031. U.S. Environmental Protection
Agency, Office of Toxic Substances, Washington, DC, 1977.
3. Rosengrant, L. Final Best Demonstrated Available Technology (BOAT) Background Document
for Mercury Containing Wastes. U.S. Environmental Protection Agency, Washington, DC, 1990.
4. U.S. Environmental Protection Agency. Review of National Emission Standards for Mercury.
EPA-450/3-84-014. U.S. EPA Office of Air Quality Standards and Planning, Triangle Park, NC,
1984.
5. Drake, H. J. Mercury. Jni Kirk-Othmer Encyclopedia of Chemical Technology. 3rd ed. John Wiley
and Sons, NY, NY. 1978.
76
-------
RECENT ADVANCES IN THE ANALYTICAL TECHNIQUES FOR THE
QUANTIFICATION OF MERCURY AND MERCURY COMPOUNDS IN DIFFERENT
MEDIA
Eric Prestbo and Nicolas S Bloom " /
Brooks Rand, Ltd
3950 6th Avenue, Northwest
Seattle, Washington 98107
USA
Tel:(206) 632-6206
Research and development at Brooks Rand is focused on improved methods for the collection and
analysis of individual mercury species in the entire range of environmental samples (air, water, soil, and
tissue). The long-term goal of this work is to understand the sources, sinks, and chemical transformations
of mercury in the environment. By doing so, it will be possible to isolate the most efficient methods to
remediate existing mercury problems and also monitor the result of a remediation strategy.
The method used is a cold vapor atomic fluorescence technique, based upon the emission of 253.7
nm radiation by excited Hg° atoms in an inert gas stream. The challenge has been to extract quantitatively
and separate individual mercury species in complex environmental sample matrices. Recently, new
advances have been made in the laboratory which have improved upon the existing techniques. Also,
mercury sampling methods have historically beenf lawed, resulting in concentrations that are systematically
high. Furthermore, accurate results depend on proper storage and sample handling in the laboratory.
The most recent findings regarding improved analytical techniques, sample handling, and sample
storage of mercury in environmental samples will be presented. A comparison will be made to other current
methods, exploring both the advantages and disadvantages inherent in the analytical methods for mercury.
Finally, future research questions and needs will be addressed.
77
-------
MERCURY IN SEDIMENTS - HOW CLEAN IS CLEAN?
Gary N. Bigham and Elizabeth A. Henry
PTI Environmental Services
1601 Trapeio Road, Waltham, MA 02154
USA
Tel: (617) 466-6681
INTRODUCTION
Sediments are often the repository for both organic and inorganic contaminants in aquatic environ-
ments. Numerous techniques have been developed to characterize the distribution of contaminants and to
estimate or predict effects on water quality and aquatic biota. This paper will describe several approaches
to development of sediment quality criteria and their applicability to mercury. The behavior of mercury in
aquatic systems will be discussed and the requirements for effective sediment criteria for mercury will be
described.
SEDIMENTCRITERIA
In the 1970s, regulatory consideration of sediments was related to permitting dredging and dredged
material disposal. Techniques were also developed to estimate the potential water quality impact of aquatic
disposal of dredged sediments (1). Since then, research on the behavior of sediment contaminants in situ
and their effects on benthic macroinvertebrates and other organisms has led to attempts to develop
sediment quality criteria to protect aquatic organisms. -
Several methods have been developed for deriving sediment criteria (2). In general, the methods
evaluate the toxicity of either interstitial water or bulk sediment to aquatic macroinvertebrates. Interstitial
water methods include comparing measured contaminant concentrations to water quality criteria, perform-
ing bioassays, and predicting the concentration of selected contaminants based on the total sediment
concentration and organic carbon concentrations. Sediment criteria methods based on analysis of bulk
sediments include bioassays of spiked orunspiked sediment, characterization of benthic macroinvertebrate
' communities, or both.
Methods such as the Sediment Quality Triad (3) and the Apparent Effects Threshold (4) link site-
specific observations of contaminant concentration, macroinvertebrate community structure, and bioassay
results. All of these methods evaluate effects on or potential toxicity to aquatic organisms residing in contact
with the sediments.
MERCURY IN THE AQUATIC ENVIRONMENT
Although some forms of mercury can be directly toxic to aquatic macroinvertebrates, the primary
hazard of mercury in sediments is an indirect one. Mercury undergoes methylation in sediments and it is
the methylmercury species which are highly bioaccumulated in the aquatic food chain. Methylmercury has
been observed to be concentrated by 106 in fish tissue compared to the surrounding water (5). The hazard
to humans is related to consumption of methylmercury-contaminated fish. Sediment quality criteria related
to toxicity to macroinvertebrates are, therefore, not appropriate for mercury.
Mercury methylation in sediments is a biologically-mediated reaction (6) and is favored in anoxic
sediments over oxic sediments (7). Although many bacteria and fungi are capable of methylation, the
principal methylators of mercury in both saline and freshwater anoxic sediments are sulfate-reducing
bacteria (8). Measuring the activity of anaerobic sulfate-reducing bacteria may be important for predicting
the rate of methylmercury production.
Once methylmercury is created, its behavior in an aquatic ecosystem will be influenced by a number
of factors. Methylmercury is strongly adsorbed to organic and inorganic particles and can undergo bacterial
78
-------
demethylation. The pH of the system will influence the chemical form of the methylmercury (i.e., chloride,
hydroxide, or ionic). Each of these factors can be significantly different between, for example acidified
seepage lakes, eutrophic lakes, rivers, or estuaries.
The mechanism by which methylmercury enters the food chain and is eventually transferred to fish
tissue is not well known. Methylmercury produced in sediment may move into the water column by one or
more of the processes of diffusion, advection, resuspension, or bioturbation. Measurement of in situ
methylmercury flux rates may be possible now with the development of new analytical techniques for very
low(< 1 part per trillion) methylmercury concentrations (9). Once in the water column, methylmercury,can
become incorporated into phytoplankton by diffusion through the cell wall or by adsorption to the cell
surface. Methylmercury will then accumulate in zooplankton which graze on phytoplankton and similarly
in planktivorous and piscivorous fish. '
Despite numerous studies, there is little evidence that a direct correlation exists between mercury
concentration in sediments and the methylmercury concentration in fish tissue. Other parameters such as
lakewater pH, lake volume and maximum depth, and lake water alkalinity appear to influence methylmer-
cury accumulation in fish (10, 11). Sediment quality criteria for mercury based on sediment mercury
concentration toxicity are, therefore, not directly related to human health or ecological risk.
REQUIREMENT FOR SEDIMENT MERCURY CRITERIA
Sediment mercury criteria should be based on achieving acceptable concentrations of mercury in fish
tissue. Based on new analytical techniques, measurement of in s/Yuratesof methylmercury production may
be possible. Indirect measurements may also be useful. For example, the sulfate-reduction rate in
sediments may be an indicator of methylmercury production rate. The actual mechanism by which
methylmercury becomes associated with phytoplankton is critical. Methylmercury-cell association may be
influenced by ambient water composition. Much research remains to be done to identify the factors which
influence the rates of mercury methylation.
REFERENCES ( '
1.
2.
3.
4.
5.
6.
7.
8.
Engler, R.M. and Mathis, D.B. Dredged-material disposal strategies. In: M.E. Champ and P.K.
Park (ed.), Oceanic Processes in Marine Pollution. Vol. 3. Marine Waste Management-
Science and Policy, Malabar, Florida, 1989. p. 53. ,
Giesy, J.P. and Hoke, R.A. Freshwater sediment quality criteria:Toxicity Bioassessment. in: R.
Baudo, J. Giesy, and H. Muntau(ed.), Sediments: Chemistry and Toxicity of In-Place Pollutants '
Chelsea, Michigan, 1990. p. 265.
Chapman, P.M. Sediment quality criteria from the sediment quality triad; an example Environ
ToXicol. Chem. 5:957-964,1986 ~
EPA Report of the sediment criteria subcommittee of the ecological processes and effects
committee. EPA-SAB-EPAC-90-018, U.S.Environmental Protection Agency, Washington, DC,
i yyo. 2o pp.
Gilmour, CC and Henry, E. A. Mercury methylation in aquatic systems affected by acid deposition
Environmental Pollution 71:131-169,1991
Jensen, S. and Jernelov, A. Biological methylation of mercury in aquatic organisms Nature
(London) 223: 753-743, 1969.
Compeau, G. and Bartha, R. Methylation and demethylation of mercury under controlled redox,
pH and salinity conditions. Applied Environmental Microbiology 48:1023-1027,1984.
Compeau, G. and Bartha, R. Sulfate-reducing bacteria: Principal methylators of mercury in
anoxic estuarine sediment. Applied Environmental Microbiology 50: 498-502,1985.
79
-------
9. Bloom, N. Determination of picogram levels of methylmercury by aqueous phase ethylation
followed by cryogenic gas chromatography with cold vapour atomic florescence detection. San,
J. Aquat. Sci. 46:1131 -1140,1989.
10. Hakanson, L, Nillsson, A., and Andersson, T. Mercury in fish in Swedish lakes. Environmental
Pollution 49:145-162,1988.
11 Wren, C.D., and MacCrimmon, H.R. Mercury levels in sunfish, Lepomis gibbosus, relative to pH
and other environmental variables of Precambrian Shield lakes. Can, J. Fish, and Aquat. Sci. 40:
1737-1744,1983.
80
-------
EFFECT OF CHEMICAL FORM OF MERCURY ON THE PERFORMANCE OF DOSED
SOILS IN STANDARD LEACHING TESTS: EP AND TCLP (1)
Ralph R. Turner
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, TN 37831
USA
Tel: (615) 574-7856
INTRODUCTION
There are few regulatory and/or advisory criteria for mercury in contaminated soils, sediment, and
waste. Typically, action or cleanup guidelines are developed on a case-by-case basis. The most widely
applied criterion for mercury in solid material is the performance of solid samples analyzed with leaching
procedures developed by the U. S. Environmental Protection Agency (EPA). These protocols, along with
others, have been used to classify wastes underthe Resource Conservation and Recovery Act (RCRA) (1).
Leaching results, expressed as the concentration of mercury in the laboratory-generated leachate, are
compared to 200 M9/L- Results above this criterion are classified as having failed the RCRA "toxicity
characteristic." The 200 ug/L limit represents 100 times the Drinking Water Standard for mercury and
requires the sample to contain at least 4 ug/g. The latter value is derived by multiplying the established limit,
200 ug/L, by the solution-to-solid ratio in the test, 0.02 L/g.
Theoretically, any sample containing total mercury less than 4 ug/g could not fail the protocol and be
accordingly classified as hazardous under RCRA. Rarely, if ever, will all the mercury in a soil or waste be
completely leachable in a protocol. Therefore, soil containing considerably higher concentrations of
mercury than 4 ug/g could theoretically pass the protocol. On September 25, 1990, an earlier leaching
protocol, called the Extraction Procedure (EP), was replaced with the Toxicity Characteristic Leaching
Procedure (TCLP) (40 CFR 261). While there are some minor differences between the TCLP and the EP
protocols, most notably the composition of the extraction fluid, the established limit remained 200 ug/L.
Application of the EP to a variety of mercury-contaminated soils and solid wastes from US Department
of Energy (DOE) facilities in Oak Ridge, Tennessee has not shown any consistent relationship between test
results and total mercury concentrations. Low yields of leachable mercury (less than the RCRA limit of 200
ug/L) have not been uncommon for samples containing several thousand ppm of total mercury and even
for samples exhibiting visible beads of mercury. Conversely, some soils with relatively low concentrations
of total mercury have exhibited high leachability of mercury (>200 ug/L) and been classified as hazardous
under RCRA.
In addition to the previously-described anomaly, highly mercury- contaminated sediments removed
from storm sewers initially exhibited low leachability of mercury (17 out of 21 samples yielded leachate with
<200 ug/L after initial dewatering). Subsequently, these sediments exhibited high leachability of mercury
(31 out of 31 samples yielded leachate with >200 ug/L following re-analysis after dry storage for up to 3
years). The increase in solubility was hypothesized to have been caused by surface oxidation of elemental
mercury contained in the dried sediments.While still in the storm sewer any oxidation products on the
surface of the elemental mercury would be washed away continuously. Immediately after drying, the
solubility of the mercury would be controlled mainly by the solubility of clean elemental mercury (reported
to be about 60 ug/L), and thus less than the 200 ug/L RCRA limit. The oxidation product would initially be
insoluble mercurous oxide (HgO) which, in turn, may partially decompose to more soluble mercuric oxide.
The reported solubility of mercuric oxide (about 53000 ug/L) is sufficient to yield a leachate exceeding 200
M9/L
As previously discussed, pure mercury compounds have different solubilities in aqueous solutions.
Generally, organic and inorganic mercury compounds (except mercuric sulf ide) are sparingly soluble while
elemental mercury is relatively insoluble in aqueous solutions (2). For this reason, mercuric sulfide and
elemental mercury, if submitted in pure form to either of these protocols, would be expected to pass at much
81
-------
higher concentrations than compounds such as mercuric oxide. The adsorption capacity of soils and solid
wastes is also expected to reduce the solubility of any pure compound added to the soil or waste.
The objectives of this study were to determine which forms of mercury in soil exhibit the highest
leachability as analyzed by the EP and TCLP protocols and to determine whether headspace mercury vapor
readings from soil samples are indicative of leaching performance. The four mercury compounds studied
were elemental mercury (Hg°), mercuricsulfide(HgS), mercuric oxide (HgO), and mercurousoxide (Hg2O).
Whether the two leaching procedures yield equivalent results was also addressed.
MATERIALS AND METHODS
Uncontaminated soil was collected from the Bear Creek Valley on the Oak Ridge Reservation. The
soil was a shaley silt loam from the Ap horizon. These soils are acidic (pHwater= 4.7) and exhibit cation
exchange capacities ranging from 5 to 10 mEq/100 g. Dried soil was crushed using a mortar and pestle and
sieved through a2 mm screen. Additional analyses to characterize the dry sieved soil included total mercury
(0.72 ug/g), organic carbon (2.08%), and particle size distribution (18.27% sand, 71.26% silt, and 10.46%
clay). Glass bottles (4 oz) were filled with 100 g of dry, sieved soil, and fitted with teflon-lined closures. Four
different mercury compounds (Hg°, HgS, HgO, and Hg2O) at four different concentrations (0,100,1000,
10000 ug/g) were studied. Each treatment was done in triplicate. Elemental mercury was first washed with
distilled waterto remove oxide, and only shiny beads of Hg° were used to dose the soil. After dosing, samples
were shaken vigorously to blend the mercury compounds with the soil.
Shaken samples were allowed to equilibrate for at least 1 h before headspace vapor readings were
taken. A Jerome Model 411 Gold Film Mercury Vapor Analyzer was used to measure the mercury vapor at
room temperature (22to24degreesC). To take a reading, each sample bottle was opened enough to insert
the Jerome sampling tube and a 10 s scan was performed. Sample headspace volume (about 50 ml_) was
considerably less than the volume (125 ml) of airdrawn by the Jerome and thus uncontrolled dilution of the
headspace sample occurred. Each sample was read three times with the exception of some of the soils
dosed with the highest amounts of Hg° which saturated the Jerome.
Leaching protocols were carried out using the entire contents of the sample bottle in orderto eliminate
any lack of homogeneity that could arise from taking sub-samples. While the two protocols are very similar,
some minor differences exist (40 CFR 261). Because the soil collected from the Bear Creek Valley site was
already acidic (pH <5.0), no acetic acid was added in the EP extraction. The extraction fluid for the TCLP
contained acetic acid and thus contrasts with the EP fluid in this potentially important aspect.
RESULTS AND DISCUSSION
The results from the trials wherein the Jerome analyzer was used to measure headspace mercury
vapor were not always highly reproducible based on the replicate trials. General trends among the
compounds were observed. Elemental mercury yielded some of the most variable vapor results, especially
within the same concentration. These highly variable measurements may be due to the way each sample
was shaken and the corresponding position of the mercury bead(s) at the time of measurement or to the
uncontrolled dilution with room air during headspace sampling. Generally, the higher the dosed soil
concentration the higher the headspace vapor reading. For the Hg° dosed soil very few readings could be
obtained from samples dosed to 10,000 ug/g because the gold film on the Jerome monitor saturated (>1.99
mg/m3).
Because the Jerome analyzer is engineered to detect elemental mercury and the other compounds
tested have very low or no vapor pressures, it was not initially expected that the other three test compounds
(HgS, HgO, and Hg2O) would yield detectable vapor. Mercuric sulfide showed the most consistent
headspace mercury vapor readings. At all concentrations, HgS yielded 0.000 to 0.001 mg/m3 of Hg.
Mercuric oxide samples at the 100 and 1000 ug/g dosing levels also resulted in very low mercury vapor
readings (0.000 to 0.014 mg/m3). At 10,000 ug/g, the vapor readings in soil dosed with HgO ranged from
0.006 to 0.081 mg/m3. Mercurous oxide caused increased vapor concentrations with increased soil
concentration. Detection of significant elemental mercury vapor over Hg2O may be explained by thefact that
this compound is subject to disproportionation and may, therefore, contain some elemental mercury.
82
-------
Results from the two leaching protocols are summarized in Table 1. As expected, different test results
were obtained for the different mercury compounds, dosed concentrations, and leaching protocols. None
of the dosed concentrations forHg0 and HgSyielded leachate concentrations in EPorTCLP exceeding 200
ug/L, the RCRA limit. In fact, the EP leachate concentrations in the HgS dosed soils are similarto the blank
concentration of non-dosed soil (Table 1). Using the EP protocol, only HgO at doses of 1000 and 10,000
ug/g and Hg2O at 10,000 ug/g resulted in leachate concentrations exceeding 200 ug/L.
Leachate concentrations exceeding the RCRA limit were also found at the 1000 and 10,000 ug/g
doses for both HgO and Hg2O using the TCLP protocol. With one exception (HgO at 100 ug/g) the TCLP
yielded similar or higher leachate concentrations than the EP (see subsequent discussion). For HgO and
Hg2O at the 10,000 ug/g dose level, the TCLP leachate concentrations were 10 times higher than the EP
results. The TCLP leachate concentrations for soil dosed with Hg2O at the 10,000 ug/g dose are similarto
the pure compound solubility value (53 mg/L) given in Kaiser and Tolg (2). A possible explanation for the
much higherthan expected solubility of Hg2O could be a result of disproportionation of mercurous mercury
to elemental and mercuric mercury. This could also explain why the Hg2O dosed soils yielded high vapor
readings. The purity of the various mercury compounds was not verified independently and thus, sufficient
HgO could have been present in the Hg2O to increase the apparent solubility of Hg2O.
The sample headspace vapor results clearly indicated that presence or absence of headspace vapor
could not be used to predict performance of a soil or waste on the leaching protocols. Nonetheless,
measurement of headspace vapor may provide other useful information about potential human and
ecological risks associated with the tested sample. For example, our study has shown that a soil sample
containing one percent by weight (10,000 ug/g) elemental mercury could pass the TCLP leaching test but
TABLE 1. AVERAGE MERCURY CONCENTRATIONS OBTAINED FROM TWO LEACH-
ING TESTS*
Dose
100
1000
10000
0
Compound
Hg°
HgS
HgO
Hg20
Hg°
HgS
HgO
Hg20
Hg°
HgS
HgO
Hg2o
BLANK
EP Results
1.86" 0.21
0.38° 0.17
65.7" 8.77
42.8" 0.56
17.2b 3.72
0.22° 0.05
220." 46.0
114." 20.16
47.6b 5.24
2.48C 1.19
6270." 861.
3200." 694.
0.42 0.31
19.4
76.3
24.2
2.3
37.6
36.4
36.2
30.4
19.1
83.1
23.8
37.5
129.
TCLP Results
G/g/L) S.E. C.V.(%)
8.89b
. 1.63°
38.5"
53.8"
25.5"
1.67°
258."
491."
51.6°
2.15d
57200."
36600."
0.30
0.29
0.61
0.99
9.68
4.39
0.29
10.5
195.
0.46
0.39
7030.
1590.
0.03
5.74
65.0
4.47
31.2
29.8
29.9
7.07
68.9
1.55
31.6
21.30
7.52
16.7
*For each dosing levi*and test, averages followed by the same letter ("•"•c'd) are not significantly
different at the 5% level (Duncan's Multiple Range Test applied to log-transformed.data).
83
-------
have a sample headspace vapor concentration (e.g., >2 mg/m3) that exceeds both the "derived" ambient
air (0.001 mg/m3;28 CFR Pt I! 38) and the industrial hygiene (0.05 mg/m3;29 CFR 1910.1) standards by
orders of magnitude. This fact alone is a compelling reason to treat such a sample as a hazardous waste
regardless of its leaching performance.
The leaching results suggest that reliance on the EP or TCLP alone to classify waste orsoil could lead
to a significant underestimation of the risks posed by a poorly characterized solid and inappropriate
disposal. Wastes or soils containing mercury predominantly in the elemental or sulfide forms would be
particularly subject to misclassification if only limited process or site knowledge is available. We suggest
that a tiered approach, incorporating headspace vapor analysis, analysis for total mercury, and application
of TCLP, is necessary to fully evaluate the degree of hazard associated with mercury- contaminated soils
and wastes.
Ideally, information on the speciation of mercury in the solid phase would also be desirable, but such
measurements are expensive and well beyond the analytical capabilities of most laboratories. Obviously,
headspace vapor analysis conveys some information about speciation of the mercury and this analysis can
be performed routinely at low cost. No standardized protocol for performing headspace mercury vapor
measurements currently exists but developing such protocols is straightforward and can be patterned after
headspace analysis forvolatile organic compounds. Analytical capability to perform total mercury analyses
is almost universal and entails less equipment and produces less chemical laboratory waste than
performance of a leaching protocol such as the TCLP. Thus, it seems logical always to screen samples for
headspace mercury vapor and to perform total mercury analysis prior to making a decision about the need
to perform the more expensive leaching protocol also.
Currently there exists no federally regulated, uniform guideline for total allowable mercury levels in
soil. After May 8,1992 debris and soil failing TCLP and containing mercury at a concentration >260 ug/g
was banned from disposal by landfilling (40 CFR 268) and such material now requires pretreatment
(thermal) before landfilling. Soil containing less than 260 ug/g but that fails the TCLP will be classified as
hazardous. The US DOE's Oak Ridge Y-12 Plant in Tennessee has adopted a 12 ug/g interim maximum
concentration for mercury in soil as proposed by Bashor and Turri (3). This value incorporates potential
mercury risks from inhalation and direct ingestion, and is similar to the value (20 ug/g) suggested by EPA
(40 CFR 55:145, July 27,1990) as an "action level". More recently, Revis et al. (4) recommended that a 722
ug/g limit of mercury in soil would suffice for the Y-12 Plant site because a large percent of the mercury
contamination had been bound in the HgS form, at least in contaminated sediment and floodplain soil.
This study has confirmed Revis1 suggestion that the form of mercury contamination must be assessed
in orderto establish an appropriate limit. Soil can'contain much hig her concentrations of relatively insoluble,
non-volatile HgS and probably constitute negligible human health risk where smaller concentrations of other
mercury forms should trigger immediate attention. Transformations among forms of mercury, especially
those that increase the mobility and bioavailability of mercury obviously complicate this issue and bear
further consideration.
There are currently no regulatory criteria or guidelines which apply to waste or soil samples which do
exhibit detectable headspace mercury vapor. We suggest that headspace mercury vapor concentrations
that exceed 10% of the industrial hygiene workroom standard of 0.050 mg/m3 could constitute an inhalation
hazard undercertain circumstances (e.g., confined or poorly- ventilated space). Solid waste or soils which
exhibit such headspace vapor concentrations should be classified as hazardous. We also suggest that
materials exhibiting detectable headspace mercury vapor (e.g., 0.001 mg/m3 using the Jerome Model 411
analyzer or similar instruments), but less than 0.005 mg/m3, should be classified as hazardous if total
mercury exceeds 20 ug/g. These proposed headspace vapor guideline values are preliminary and may
need further support based on risk analysis.
84
-------
REFERENCES
1. U.S. Environmental Protection Agency. Test Methods for Evaluating Solid Waste; Volume 1C:
Laboratory Manual Physical/Chemical Methods. SW-846.1986.
2. Kaiser, G., and Tolg, G. Mercury. lm Hutzinger, O. (ed.), Handbookof Environmental Chemistry:
Anthropogenic Compounds. Springer Verlag, Berlin, 1980. pp. 1-58.
3. Bashor, B. S., and Turri, P. A. Arch. Environ. Contam. Toxicol. 15:435.1986.
4. Revis, N. W., Osborne, T. R., Holdsworth, G., and Hadden, C. Arch. Environ. Contam. Toxicol.
19:221.1990.
85
-------
INTER-LABORATORY TESTING FOR MERCURY BY TCLP AND SOURCE
REDUCTION IN THE LAMP MANUFACTURING INDUSTRY
Vincent D. Meyer
National Electrical Manufacturers Association
GTE Electrical Products
100 Endicott Street
Danvers, MA01923
USA
Tel: (508) 750-2322
INTRODUCTION
This paper presents the results of extensive inter-laboratory testing of fluorescent lamps by the Lamp
Technical Committee of the National Electrical Manufacturers Association (NEMA) in cooperation with the
Environmental Protection Agency (EPA) using the Toxicity Characteristic Leaching Procedure (TCLP).
Included in this paper will be a report on the significant progress the lamp industry is making to reduce the
amount of mercury used in fluorescent lamps.
Disposal of spent fluorescent lamps containing small amounts (about 0.01 %) of mercury is becoming
a problem. Disposers of spent fluorescent lamps encountered a serious problem in determining if these
lamps exceed the regulatory threshold of 0.2 mg/L when applying the standard EPA test for to.xicity
characteristics under the Resource Conservation and Recovery Act (RCRA) and reported their concerns
to US lamp manufacturers. In dealing with this situation, US lamp manufacturers through NEMA reported
to the EPA Officeof Solid Waste the wide variation in results obtained by thedisposers of fluorescent lamps
and initiated a series of inter-laboratory tests.
The inter-laboratory testing had three objectives:
To confirm the variation among various laboratories in performing TCLP on fluorescent lamps
To isolate within the TCLP the step or steps that give rise to the variations
To attempt to determine the testing conditions or causes of the variation in results and to suggest
methods of obtaining consistent test results
The first two objectives have been successfully met by this series of inter-laboratory tests; however,
it was not possible to find testing conditions which would eliminate the large variation in test results.
METHODOLOGY
The testing was designed to isolate the effect of each step in the TCLP on variability. Beginning with
the final step in TCLP of analysis for mercury, the testing progressed through four phases:
(I) Analysisof standard solutions of mercury to quantitatively determine theability of each laboratory
to analyze for soluble mercury
(II) Analysis of aliquots of solutions obtained after the filtering step in TCLP, prepared at one
[laboratory, to determine quantitatively the variability when analyzing for soluble mercury in the
extraction fluid
(III) Extraction of crushed lamps prepared at one laboratory to determine the agreement among
laboratories when extracting, filtering, and analyzing crushed lamps
(IV) TCLP tests for mercury in whole fluorescent lamps to determine quantitatively the agreement
among laboratories when the complete TCLP is performed on fluorescent lamps
Each phase included a sample containing no mercury, a blank. In each phase the samples were
86
-------
identified only as A, B.C.Dand laboratories were requested to perform the appropriate procedure for TCLP
and/or analysis for mercury.
Standard 4-ft. fluorescent lamps were used. All lamps were selected from lamps made consecutively
on a high-speed lamp manufacturing machine with a history of consistently producing duplicate lamps.
Blanklamps were made by turning off the mercury dispensing apparatus during the manufacturing process
immediately after the test lamps were produced.
IMPLEMENTATION
' Initially seven different laboratories were chosen to participate in these inter-laboratory tests. The
laboratories included both the laboratories of US lamp manufacturers and independent laboratories
experienced in performing TCLP. During the final phase of the testing a laboratory chosen by EPA Was
included. The laboratories were numbered 1 to 8 and otherwise not identified.
Phase I - Samples of standard solutions of mercury were prepared by dissolving metallic mercury in
nitric acid followed by appropriate dilution. The solutions were shipped to the individual laboratories with the
following instructions:
"This is the first in a series of tests designed to improve the inter-laboratory correlation of the TCLP
for mercury. The enclosed four solutions contain mercury in
-------
This is the fourth in a series of tests designedto improve the inter-laboratory correlation of the TCLP
for mercury. The enclosed four samples are fluorescent lamps manufactured to our specifications
and aged for specific time periods. Using the entire sample in each case, perform TCLP extraction
and filtration as per 40 CFR Parts 261,264,265,268, 271 and 302 dated June 29,1990 Analyze
for mercury concentration using method 7470. Report results in mg/L. Please keep records of how
the procedure was interpreted. See attached notes on recording procedure."
Each laboratory was also provided with a questionnaire to record details of the procedure used in
performing the TCLP. i
RESULTS
TABLE 1. PHASE I, STANDARD SOLUTIONS OF MERCURY
Sample Identification
As Made
Laboratory 1*
Laboratory 2
Laboratory 3
Laboratory 4
Laboratory 5
Laboratory 6
Laboratory 7
Concentration of Mercury in mg/L
IBzA. IR-B IR-C
0.000 0.113 0.565
0.000
0.000
0.001
0.000
0.002
<0.01
0.0550
Mean (excluding Lab 1) 0.011
Std Dev 0.022
Coefficient of Variation- SD/M 1.92
Recovered (Mean/As Made) n.a.
0.010
0.090
**
0.080
0.108
0.10
0.1310
0.102
0.019
0.191
90%
0.050
0.490
0.410
0.585
0.529
0.49
0.5650
0.512
0.063
0.123
91%
IR-D
1.017
0.100
0.990
0.780
1.000
0.944
0.94
0.8600
0.919
0.084
0.092
90%
* After the fact Laboratory 1 reported a calculation error of a factor of 10
* 'Solution B lost at Lab 3
TABLE 2. PHASE II, SOLUTIONS AFTER THE FILTERING STEP IN TCLP
Sample Identification
Description
Laboratory 1
Laboratory 2
Laboratory 3
Laboratory 4
Laboratory 5
Laboratory 6
Laboratory 7
Mean
Std Dev
Coefficient of Variation- SD/M
Concentration of Mercury in mg/L
3 lamps
0.518
0.57
0.43
0.51
0.52
0.35
0.601
0.500
0.085
0.170
0.2mg/L HgO
0.204
0.23
0.19
0.250
0.242
0.20
0.348
0.238
0.054
0.225
2R-C
blank
<0.001
0.008
0.004
n.d.
0.005
<0.01
n.d.
<0.004
0.004
0.626
2R-D
1 lamp
0.699
0.71
0.60
0.69
0.782
0.55
0.827
0.694
0.096
0.138
88
-------
TABLE 3. PHASE III, CRUSHED LAMPS
Sample Identification
Type of Lamp
Laboratory 1
Laboratory 2
Laboratory 3
Laboratory 4
Laboratory 5
Laboratory 6
Laboratory 7
Mean
Std Dev
Coefficient of Variation- SD/M
Concentration of Mercury in mg/L
CL-1 CL-3 CL-4 Mean, CL-2
Normal Normal Normal - of Three No Hg
0.010±. 007 <0.0005
0.40 ± .08 n.d.
0.49 ± .06 n.d.
0.51 ± ,06 n.d.
0.94 ± .35 <0.01
0.105±.03 <0.001
3.09± 2.11 n.d.
0.793
1.21
1.53
0.0067
0.31
0.43
0.56
0.55
0.082
1.57
0.0056
0.46
0.48
0.45
1.23
0.14
2.20
0.0176
0.44
0.55
0.53
1.04
0.093
5.50
0.501
0.518
1.03
0.709
0.763
1.08
1.167
1.94
1.66
Sample Identification
Type of Lamp
Laboratory 1
Laboratory 2
Laboratory 3
Laboratory 4
Laboratory 5
Laboratory 6
Laboratory 7
Laboratory 8
Mean
Std Dev
Coefficient of Variation- SD/M
TA13LE 4. PHASE IV, WHOLE LAMPS
Concentration of Mercury in mg/L
WL-1 WL-2 WL-4 Mean
Normal Normal Normal of Three
0.021
0.56
0.39
0.31
1.3
0.01
0.08
0.320
0.013
0.52
0.39
0.52
1.52
0.02
0.10
0.650
0.0006
0.65
0.53
0.57
2.8
0.02
0.07
0.760
0.374
0.421
1.13
0.467
0.492
1.053
0.675
0.911
1.35
0.505
0.630
1.25
WL-3
NoHg
0.012± 0.01
0.58± 0.07
0.44± 0.08
0.47± 0.14
1 .87± 0.81
0.02± 0.006
0.08± 0.02
0.577± 0.229
0.021
n.d.
n.d.
n.d.
0.02
<.01
n.d.
<0.001
DATA ANALYSIS
Phase I demonstrated that the individual laboratories are capable of performing quantitative analysis
for soluble mercury over the range of concentrations of interest with a coefficient of variation of about 0.14
(mean of the Coefficients of Variation excluding the blank) and recoverabout 90% of theformulated amount
of mercury.
Phase II demonstrated that all of the laboratories are capable of quantitatively recovering soluble
mercury from the leaching solution. The soluble mercury solution was prepared by dissolving enough HgO
in the extraction solution to produce a mercury concentration of 0.20 mg/L. The mean for all of the
laboratories was 0.238 mg/L or 19% higher than the formulated amount. The coefficient of variation was
0.225. It is difficult to explain the results for samples 2R-A (3 lamps) and 2R-D (1 lamp). The filtered extract
for the sample 2R-A (3 lamps), which presumably has three times the amount of leachable mercury as
sample 2R-D (1 lamp) yielded less detected mercury. The amount of mercury extracted did not correlate
in a simple way with the amount of extractable mercury in the samples.
Phase III results clearly show that there is a dramatic increase in the variation in results when the
leaching and filtering steps by individual laboratories are included in the testing on very similar lamps. The
89
-------
coefficient of variation has increased to about 1.5. For the previous phases the coefficient of variation
averaged about 0.15 (excluding blanks).
Phase IV results clearly show that there remains a wide variation in results obtained by different
laboratories when the laboratories use similar lamps to perform TCLP for mercury. The coefficient of
variation was about 1.2.
During Phase IV of the testing each laboratory completed a questionnaire describing details of the
procedure they used. This was an attempt to find differences in interpretation and implementation of the
TCLP for mercury in fluorescent lamps. While difference among the laboratories were identified, no
correlation could be found between the results of TCLP and these procedural differences.
Specifically, no correlation could be found between the number of tumbling containers used for each
lamp, the pH of the solution after extraction or the number of pieces that the aluminum end cap was cut into
before extraction.
Insummary, these inter-laboratorytests clearly showthatthe laboratories are capableof quantitatively
analyzing for mercury and can quantitatively recover soluble mercury in the TCLP solution, but wide
variation in test results among different laboratories originates in the extraction and filtration steps of the
TCLP when performing TCLP for mercury in fluorescent lamps.
SOURCE REDUCTION
All fluorescent lamps contain elemental mercury. Standard 4- and 8-ft. lamps contain between 35 and
75 mg. Very small fluorescent lamps including the compact fluorescent contain about 10 mg or less.
Lamp manufacturers have for many years been attempting to reduce the amount of mercury in
fluorescent lamps. Significant progress has been made in this source reduction effort. A recent survey by
NEMA of current manufacturing practices revealed a 14% reduction in the average mercury content of a
standard 4-ft. 1-1/2 inch diameter cool-white fluorescent lamp from 1985 to 1990. These manufacturers also
predicted a further decrease of about 35% by 1995. (Refer to Table 5.)
TABLE 5. PREDICTED DECREASE IN MERCURY IN FLUORESCENT LAMPS
YEAR
1985
1990
1995
MERCURY FILL
48 mg
42 mg
27 mg
These mercury reductions have been achieved by the use of more efficient dosing techniques during
the manufacturing process incorporated into newer high speed production equipment. The projected figure
for 1995 of 27 mg of mercury used in a 4-ft fluorescent lamp is close to the present practical limit needed
for efficient lighting performance so that reductions below this amount will probably be very small.
Nevertheless, 27 mg of mercury represents only about 0.01% of the total weight of a 4-ft. lamp.
It has been shown that the disposal of lamps containing mercury accounts for approximately 0.2% of
the total discards of mercury (1). These lamps are efficient sources of white light, typically being three to
four times more efficient than an incandescent lamp. Consequently, their use results in significantly less
power use along with associated reductions in a variety of power plant stack emissions, including mercury.
Studies show that the substitution of incandescent lamps by lamps containing mercury results in a net
reduction of environmental mercury releases (1). Typically, for the average mix of fossil fuels, nuclear
power, and hydro used in the US, this reduction is approximately three times the mercury contained in the
lamp, depending on lamp type.
90
-------
A recent Dutch government study concerning the environmental aspects of lighting concluded:
"It is remarkable how the total quantity of mercury that is released when the incandescent lamp is
used is largerthan when, for example, the fluorescent lamp is used, when the fluorescent lamp has
a bad name on account of the mercury it contains."
REFERENCE
1. Lovins, A. and Sardinsky, R. State of the Art: Lighting. Competitek, Rocky Mountain Institute,
Snowmass, Colorado. March 1988 Edition. 231 pp.
91
-------
MANAGEMENT OF MEDICAL MERCURY BATTERY WASTES THROUGH SOURCE
SUBSTITUTION
John L Price
Broward County Office of Integrated Waste Management
200 Park Central Boulevard South, Suite 3
Pompano Beach, FL 33064
USA
Tel: (305) 968-3861
In the early 1980s, solid waste managers in Broward County, Florida, began to integrate mercury
management into Broward's overall waste management system. Led by T. M. Henderson, Director of the
Resource Recovery Office, early efforts focused on proper management of mercury after entry into the
system. Late in 1990, Henderson shifted gears into diversion of mercury-containing wastes before they
reached facilities and landfills. Utilizing a source substitution strategy, the county devised a program which
Iscurrently diverting nearly one ton of mercury yearly from just one point source: mercury medical batteries.
Both solid waste and medical officials were surprised this substitution proved to provide a cost savings to
area hospitals.
BACKGROUND
Predominantly urban Broward County, Florida, contains about 1.2 million people in the Fort Lauder-
dale area of south Florida. Among the population characteristics impacting municipal solid waste (MSW)
management decisions are a large senior segment and a sizable seasonal winter influx.
The cornerstones of Broward's solid waste management system are two mass burn resource recovery
facilities with a combined 4500 ton per day capacity. Best available emission control equipment installed
after careful selection and at considerable expense ensures that mercury (and other) emissions are well
below existing and forecasted Federal, state, and local limits. Ash residues are placed in secure monofills
adjacent to plant sites. Residential and commercial recycling activities augment these facilities. Two
landfills with multiple liners, leachate collection systems, and monitoring wells round out the system. Four
Broward cities utilize a parallel MSW composting facility with unprocessable residues entering resource
recovery facilities or landfills.
IDENTIFYING THE SOURCE AND MANAGEMENT OPTIONS
Dry cell batteries were identified as a major source of mercury in the Broward County MSW stream
in a 1987 waste composition study by Cal Recovery Systems. "Button" and medical batteries were targeted
as major mercury sources to be diverted from Broward's MSW before the Resource Recovery Facilities
began operations in mid 1991.
In early February 1991, concurrent with implementation of its button battery recycling program, the
county began research and planning on the medical battery phase of its expanded mercury management
program. Preliminary investigation of the costs of a county-funded recycling effort showed that it was
prohibitive. Further, the fact that these battery wastes were commercially generated as opposed to the
household-generated button batteries raised another issue: Should county dollars be used to assist only
certain businesses (hospitals) in defraying costs of doing business (waste management)?
These considerations prompted the decision that county staff would function as an information
resource presenting the problem and options to area hospitals regarding proper battery and mercury-
containing waste management. The hospitals, having legal responsibility for management of waste
generated, would be asked to select the option appropriate to their individual operations. The selected
option had to comply with applicable waste management regulations. County staff would control and
monitor the hospitals'response.
92
-------
During the research phase of the project, one hospital professional was extremely valuable to county
staff. In addition to explaining medical battery uses, E. Kurnia, a biomedical engineer at Memorial Hospital
in Hollywood, also suggested what was to become the preferred course of action for area hospitals
regarding mercury medical batteries: substitution of an alternate type of battery. Memorial Hospital was the
first area hospital to implement changes and to realize a cost savings as a result. Other hospital, battery
industry, and governmental professionals at the local, state, and national level also provided helpful
information.
s
Discussions with appropriate hospital personnel in Broward County hospitals over the next three
months surveyed the types, uses, and quantities of medical mercury batteries. Mercury-containing wastes
from broken medical instruments, spill cleanup, and used thermostats and switches were also discussed.
Understanding of applicable Federal, state, and local waste management requirements was confirmed. It
was emphasized that mercury batteries and all mercury-containing wastes were not acceptable at either
Broward County resource recovery facilities or landfills.
Current waste management procedures were surveyed/Alternative waste management options were
presented and discussed with particular attention to costs. The following options were presented in order
of desirability from an environmental impact point of view (most through least). This order of desirability in
many cases held true from a cost point of view as well (least through most).
1. Substitute a less hazardous alternative
2. Return used batteries to supplier
3. Manage via hazardous waste management vendor
Discussions with hospital personnel confirmed that mercury 8.4 volt medical batteries were widely
used to power telemetry cardiac monitors in Intermediate Care or "step down" units and portable Holler
cardiac monitors. One hospital had switched to the more benign zinc air battery in its telemetry cardiac
monitors about five years ago because it performed better. Battery usage per year per hospital ranged from
100 to nearly 16,000 depending on the number of beds and occupancy rates (Table 1). Almost without
exception, area hospitals were disposing mercury medical batteries in regular or biohazardous waste
containers.
ONE TON OF MERCURY DIVERTED
When advised of the composition of these batteries and viable options, most hospital administrators
were quick to change use and'waste management procedures. The most popular option with respect to
mercury 8.4 volt medical batteries was to substitute a zinc air battery. The zinc air battery is usually more
expensive but in many cases lasts longer than its mercury counterpart. One drawback is that the zinc air
battery will continue to discharge even when not in use.while the mercury will not. However, this is not a
problem when the monitors are in constant use as in most hospitals.
Four hospitals switched to a combination of zinc air and low mercury alkaline batteries due to
equipment compatibility and cost considerations. Stocks of mercury batteries (usually less than one month's
supply) were depleted or returned to suppliers. I n many cases, mercury batteries from stocks were collected
when discharged and returned to suppliers for proper management or managed as hazardous wastes
through established vendors.
By the summer of 1991 most hospitals had eliminated their use of mercury batteries or initiated proper
management policies. The amount of mercury diverted from Broward's waste stream is staggering. Each
8.4 volt mercury battery weighs about 1.8 ounce (1) of which nearly 1/2-ounce is mercury (2). Based on
surveys more than 63,000 of these batteries were entering Broward's waste stream from hospitals alone
each year. That is in excess of 1,750 pounds or nearly one ton of mercury per year (Table 1).
Area hospitals managed to divert this ton of mercury and save money in the process. Taking into,
account the extra expense of alternate batteries and the avoided costs of managing mercury battery wastes,
93
-------
TABLE 1. DIVERSION OF 8.4 VOLT MERCURY MEDICAL BATTERY WASTES FROM
BROWARD COUNTY, FLORIDA, MSW STREAM 1991 BY HOSPITAL
Number of Beds
Using Monitors
Mercury Batteries
Number (Ibs mercury)
Management
Opt i on
1 .
2.
3.
4.
5.
6.
7.
8.
9.
10.
11 .
12.
13.
14.
15.
16.
70+
40
45
52
32
n/a
16
16
18
48
n/a
8
88
35 +
12
0
15,
12,
5,
5,
4.
4,
3,
3,
3,
3,
2,
708
893
844
328
176
000
600
216
000
000
300
96
0
0
0
(442)
(363)
(164)
(150)
(117)
(113)
(101 )
(90)
(84)
(84)
(65)
(3)
(0)
(0)
(0)
(0)
Substitute Zinc
Substitute Zinc
Substi tute Zi nc
Substitute Zinc
Substitute Zinc
*Return to Vendor
*Return to Vendor
Substitute Zinc
Substitute Zinc
Substitute Zinc,
Al kal i ne
**Substitute Zinc
Substi tute
Al kal ine
Use Ni -Cadmi urn
Use Al kal i ne
Use Zinc
***Use Low Mercury
Al kal i ne
Total
63,161 (1776)
Note: "Zinc":Zinc Air type batteries; "Ni-Cadmiurn":Nickel
Cadmium rechargeable batteries; "Alkaline":9 Volt Alkaline
batteries with standard mercury content; "Low Mercury Alkaline":9
Volt Alkaline batteries with less than .025% mercury content.
* Corporate investigation of nationwide Zinc Air substitution
** Existing use of Low Mercury Alkaline in some applications v
normally powered by Mercury batteries
*** No telemetry monitors: Holter monitors only
94
-------
there is a net savings in most cases. At least one hospital was able to purchase alternate batteries at the
same cost. With.a longer alternate battery life, this hospital was able to cut battery costs by more than 25%
(based on the first 2 month's experience) before taking into account the additional avoided costs of waste
management (3). This was the key in program adoption and continuation.
MONITORING
Although cost savings go far toward ensuring hospital adherence to these new waste management
policies, county staff structured three procedures to reinforce and monitorthe situation. Both the public and
private sectors were involved.
First, private waste haulers who servicearea hospitals agreedto specifically notifytheircustomersthat
mercury batteries were not acceptable in their regular trash containers. Customers were further informed
that discovery of such disposal would result in substantial cost increases and/or termination of service.
Second, stepped-up load inspections of hospital containers were initiated at the Broward resource recovery
facilities soon after they began operation. This roughly coincided with the time when most hospitals had
adopted new waste management practices. Ongoing inspections targeting hospital loads among others are
part of operational plans at both resource recovery facilities and landfills.
. Lastly, the Broward County Office of Natural Resource Protection agreed to perform battery use and
waste management inspections at all hospitals as part of their annual hazardous materials licensing
inspection. The effectiveness of these annual inspections was shown within a few weeks. An imminent
scheduled inspection prompted a call from a hospital bipmedical engineer to confirm that his battery use
and management procedures would be acceptable. To date no deviation from acceptable waste manage-
ment practices has been detected. However, monitoring will be ongoing.
CONCLUSIONS
When informed of battery waste management implications and options, Broward area hospitals were
cooperative (and in some cases enthusiastically so) with county staff in improving the situation. Functioning
as an information exchange resource with input from many hospital professionals, one county project
coordinator was able to facilitate the diversion of nearly one ton of mercury per year from Broward's waste
stream in just a few months. Costs to Broward County and the hospitals were minimal and, indeed, in most
cases, cost savings were realized by the hospitals. This cooperative approach based upon information
exchange will be helpful as Broward County expands its management efforts to other types of commercially
generated wastes.
NOTES
1.
Weight of one discharged Duracell TR146X 8.4 volt mercury battery as weighed on postal scale
at Broward County Governmental Center Mailroom 2/20/91.
Duracell TR146X mercury battery contains 27% HgO according to 3/8/91 Telefax from George
Wallis, Director of Environmental Affairs, Duracell, Needham, MA. HgO contains 92.6% mercury
by weight according to George Riley, Director, Environmental Monitoring Division, Broward
County Office of Natural Resources Protection, 4/9/91 telephone conversation.
E. Kurnia, Biomedical Engineer, Memorial Hospital of Hollywood, FL, 5/29/91 telephone conver-
sation.
95
-------
RECOVERY OF MERCURY D-009 AND U-151 WASTE FROM SOIL USING PROVEN
PHYSICAL AND GRAVIMETRIC METHODS
Michael W. Chintis
GZA GeoEnvironmental, Inc.,/Hunter Mining Co.,
141 East Palm Lane
Phoenix, AZ 85004
USA
Tel: (602) 495-1833
INTRODUCTION/BACKGROUND
GZA GeoEnvironmental, Inc., in conjunction with Hunter Mining Co., (GZA/HM), has developed a
viable, cost-effective, and practical process to recover elemental mercury from soil. Many techniques were
investigated with most proving inefficient or impractical for field use. GZA/HM came to the conclusion that
modifications and refinements of mining technology would produce results most compatible to US
Environmental Protection Agency (EPA) standards and its philosophy of protecting human health and the
environment. The technology developed removes/recovers in excess of 99.8% of the elemental mercury
present in the soil.
In focusing on the removal of elemental mercury as the primary objective three important tasks have
been accomplished: 1) the primary source of mercury mobility throughout the environment described in the
EPA's Toxicological Profile for Mercury published in December 1989 (i.e., transformation from elemental
to ionic and volatilization) can be eliminated; 2) any remaining residuals of mercury salts in the effluent
process streams can then be cost- effectively treated or immobilized, if required; and 3) a valuable and
reusable resource can be recovered for use as was originally intended.
The GZA/HM process uses only physical and gravimetric techniques to separate mercury from the spill
matrixes and does not incorporate any thermal or chemical methods that could increase the possibility of
creating new waste streams or vapor releases to the atmosphere.
PROCESS DESCRIPTION
Material Sizing
The first step, and arguably the most important, is to size the matrix containing the elemental mercury
uniformly. For gravimetric systems to work properly with a high degree of efficiency, material entering the
separation equipment must be of a uniform size in order to take full advantage of varying specific gravities.
The material sizing process uses pressure washing and screening to produce the uniform matrix required
to enter the separation section of the unit.
The proper mesh size screen was selected after tests demonstrated that washed material rejected was
free of elemental mercury. Additionally, most of the mercury oxidation products passed through the screen
as well and were not present in the discharge.
In summary, when soil containing elemental mercury is adequately washed and scrubbed, mercury
in all its forms will pass through a properly-sized screen, thus greatly reducing the amount of material
requiring further processing.
Separation of Mercury from Matrix
Numerous pieces of equipment were tested in determining which apparatus would most efficiently
recover/remove elemental mercury from a soil matrix. The best device proved to be a dual rotating bowl
mechanism followed in series by a cleanup concentrate process for the specific matrix to complete mercury
separation from heavy sands. A description of the separation process follows.
96
-------
The properly sized material containing the elemental mercury enters an apparatus consisting of two
vessels operated in series. The vessels consist of dual rotating bowls inside a housing. The inner walls of
the bowls are rubber or plastic spirals designed specifically to capture elemental mercury. Slurried mesh
material is fed to the center of the bottom of the bowl. The heavier elemental mercury and insoluble mercury
salts are caught in the lower riffles of the bowl and the lighter tails overflow the top of the bowl. The heavier
material containing the mercury, mercury salts, and black sands is discharged from the bottom of the bowl.
The elemental mercury is siphoned off and the remaining material is passed to the second stage for further
separation of heavy ends.
This staged process consistently produces reduction in mercury concentrations exceeding 99.8% and
can operate at ranges of 1/2 to 6 tons per hour. Atypical demonstration test for this staged process cleaned
93.0% of the material volume to levels below 11 ppm total mercury concentration in a single pass (from an
initial concentration of approximately 5000 ppm).
In summary, this demonstration succeeded in reducing the mercury concentrations in soil from 5000
ppm to 11 ppm in 93% of the soil volume.
EVALUATION OF EFFLUENT STREAMS
The GZA/HM process, by not incorporating any form of thermal vaporization or chemical reaction to
clean soil, produces only four possible effluent streams. These streams are accounted for as follows:
Cleaned soil produced from two areas of operation: 1) the screen reject material which has been
washed at the entry to the recovery process; and 2) tailings from the rotating bowls. Tailings from the
concentrator stage can be reintroduced back into the rotating bowls. The cleaned soil exiting the process
from screening and separation typically represents 93% of the total soil volume and consistently contains
less than 15 ppm mercury.
Elemental mercury recovered from the concentration stage is returned to stock for use as was
originally intended.
Heavy sands containing minute particles of elemental mercury plus mercuric oxides in a magnetite
matrix are stockpiled. When operations at a site are complete the small volume of heavy sands/matrix
containing residual mercury will require some further processing. Disposal and/ortreatment options forthis
low-volume residual will depend on client and oversight agency agreement parameters.
Water is recycled during operations and since no chemicals are introduced to the system to solubilize
mercury, a buildup of ionic concentration of mercury has not been observed and is not anticipated. This has
been substantiated by sampling of water during demonstration testing.
QUALITY ASSURANCE/QUALITY CONTROL
Demonstration tests and treatability studies have been conducted with EPA QA/QC protocols and
analytical results and mass balance reports are available upon request. Samples were taken to ensure
representativeness of matrix and delivered to a certified environmental laboratory for analysis. Field and
internal lab QA/QC protocols plus sample analyses were conducted with procedures consistent with EPA's
procedures for verification of Best Demonstrated Available Technology (BDAT)l
The tests/demonstrations were conducted at a reputable mining research and development company.
Some QA/QC procedures used in gathering this data were modified from the guidelines presented in EPA
Guidebook for Submission of Data for the Land Disposal Restriction (LDR) program for the following
reasons: a
Elemental mercury does not disperse evenly throughout a soil matrix; therefore, GZA/HM increased
the quantity of the sample for total digestion analysis from 0.2 grams to as much as 10 grams. The
increase in sample size will assure representativeness and negate the possibility of an uneven
distribution of mercury residuals generating false positive data.
97
-------
When tests were conducted using spiked matrix, accurate weighing of mercury inoculant was
substituted fortotal mercury analysis in preparing the pre-test matrix and evaluating test efficiency.
SUMMARY
The GZA/HM process utilizes the high specific gravity of mercury and adapts basic mining techniques
into a process capable of recovering 99.8% of elemental mercury from soil matrixes. Once a valuable
resource has been recovered, any trace mercury residuals remaining in the soil discharge can be
immobilized or be subjected to further reduction techniques (if the trace concentration levels exceed any
oversight parameters).
This process is adaptive to many site variables such as concentration levels of mercury, soil types,
presence of utilities, remoteness, and quantity of waste. A small mobile unit has been built and mounted
on a 25-foot trailer pulled by a 3/4-ton truck. This unit is designed specifically for the Natural Gas Industry
and is capable of cleaning 2 tons per hour.
This mercury recovery process was designed to comply with state/local health and environmental
regulations. By using gravity separation in a water medium the chance of mercury vaporization has been
virtually eliminated. Workers will only be subjected to handling high concentrations when siphoning the
elemental mercury discharged from the spiral concentrator to storage flasks.
An added advantage to the gravity separation approach is that most of the equipment the process unit
comprises is readily available in the mining community. Additionally, this equipment has been proven
reliable through years of operation and can be assembled to handle large or small volumes of contaminated
soil.
Using a process that does not employ thermal or chemical methods, GZA/HM has developed a
patented process compatible with EPA goals established by the Resource Conservation and Recovery Act
(RCRA) which consistently produces recovery rates exceeding 99.8%. This technology is capable of
processing the soil volumes necessary to meet estimated LDR national capacities becoming effective in
May 1992. Additionally, waste once destined for land disposal can now be processed on-site, eliminating
the continuing liability and responsibility a generator faces when hazardous waste is placed in a landfill.
98
-------
TREATMENT AND MERCURY RECOVERY FROM ELECTRICAL
MANUFACTURING WASTES
H. H. Dewing and R. B. Schluter
US Bureau of Mines
Rolla Research Center r
1300 N. Bishop Avenue, PO Box 280
Rolla, MO 65401
USA • . -
Tel: (314) 364-3169
The US Bureau of Mines Rolla Research Center characterized mercury-containing waste and used
a thermal desorption process to remove and recover the contained mercury. The wastes were generated
by an electrical parts plant engaged in the assembly of mercury- containing switching devices and consisted
of discarded Bakelite phenolic resins and paper insulating materials mixed with soil and other trash. The
average mercury content of the materials was about 400 ppm. Numerous characterizations of the waste
showed that the mercury was tightly adsorbed and could not be readily removed or concentrated by leaching
or gravity separation techniques.
Both laboratory and rotary kiln thermal desorption tests were conducted and showed that mercury
content of the wastes was reduced to less than 15 ppm in 2 h at 400 degrees C. Air flow through the retort
improved thermal desorption efficiency while vacuum operation was of little benefit in lowering temperature
requirements. Off-gases from decomposition of organic materials were burned in a high temperature
afterburner. The off-gases were subsequently cooled in a heat exchangerto condense the water vapor, and
finally, the mercury vapor was condensed and adsorbed from the off-gas by activated charcoal. Mercury
recovery was over 99.99% of the desorbed mercury.
- ' \
Thermal desorption processes have had wide application to many mercury-containing wastes, and
historical experience in mercury mining has demonstrated the potential cost effectiveness.
99
-------
DEVELOPMENT OF BOAT FOR THE THERMAL TREATMENT OF K106 AND CERTAIN
D009 WASTES
Arthur E. Dungan
The Chlorine Institute, Inc.
2001 L Street, NW, Suite 506
Washington, DC 20036
USA
Tel: (202) 775-2790
INTRODUCTION
The Land Disposal Restrictions (LDR) forthe Third Third Scheduled Wastes (55 FR 22520; June 1,1990)
prohibit the land disposal of certain wastes including K106 and D009 nonwastewaters in the high mercury
subcategory beginning May 8,1992. The final rule requires that such wastes be retorted/roasted prior to land
disposal.
K106 wastes are listed wastes resulting from treatments of the effluent from the mercury cell chloralkali
process. D009 are characteristic wastes exhibiting toxicity due to mercury. These wastes are produced in
chloralkali manufacturing facilities using the mercury cell process. Seventeen facilities accounting for about
15% of the chlorine produced in the US use this technology.
Prior to the promulgation of the final rule pertaining to the Third Third Scheduled Wastes, the Chlorine
Institute (the "Institute") commented to the Environmental Protection Agency (EPA) that the proposed Best
Demonstrated Available Technology (BOAT) forthe mercury cell process was, in fact, not demonstrated. The
Institute and the various mercury cell chloralkali producers argued that the K106 and D009 wastes generated
in the chloralkali process were different from the cinnabar ore processing that was the basis of the BOAT.
Nevertheless, EPA finalized the rule based on this technology.
After the regulations were promulgated, the Institute and seven mercury cell chloralkali producers
determined that the best course of action to follow was to attempt to develop the technology that could be used
to treat the wastes.
In the summer of 1990, Hazen Research, Inc. was contacted to develop and demonstrate technology for
thermal treatment of nonwastewaters containing mercury. The project consisted of two phases. Phase 1
included the characterization of representative wastes, establishment of the affected variables and completion
of preliminary design and costs. Phase 2 of the project included the construction and operation of a pilot plant
to demonstrate the thermal process. This work was begun in late 1990 and was completed in 1991.
The process objectives were as follows:
(1) Develop the technology to treat K106 and D009 process wastes that meet the LDR requirements:
ash that is less than 260 ppm total mercury and less than 200 ppb Toxicity Characteristic Leaching
Procedure (TCLP) extract mercury.
(2) Recover the metallic mercury.
«/•
(3) Obtain operating and design data for the engineering, construction/and operation of commercial
plants for each participant.
(4) Meet the requirements of project's research, development and demonstration (RD&D) permit for
emissions to the atmosphere and effluents to the local publicly owned treatment works (POTW).
100
-------
PROCESS DESCRIPTION
The K106 nonwastewaters tested contained the following:
24 - 57% Water
1 - 12% Mercury, dry basis
0.4- 15% Sulfur, dry basis
3 - 15% Inorganic salts, primarily sodium chloride, dry basis
The bulk of the remainder was filteraid which, depending on the type used, could be carbon or
diatomaceous earth.
The D009 nonwastewaters that were the primary focus of the testing were process wastes. These were
either graphite or activated carbon used to make and purify the sodjum hydroxide or sludges from the cell house
sump. Mercury content of these wastes varied from less than 1 % to more than 20%.
The wastes were first fed into a furnace (thermal processor) equipped wJth a two stage afterburner. The
gas leaving the furnace was cooled through two stages of scrubbing and cooling using a combination of water
and sodium hydroxide. A final stage of scrubbing and cooling in a venturi and separator vessel was employed.
Metallic mercury was recovered from each stage of the scrubbing/cooling process. The trail gas exiting the final
stage of the cooling was passed through sulfur impregnated carbon to remove residual mercury before being
discharged to the atmosphere. The ash produced in the furnace was periodically removed. Analyses of the ash
were taken to determine total mercury and TCLP mercury extract. The water collected from the various
scrubbers/separators was collected in holding tanks. Prior to discharge to the POTW, it was treated in a system
similar to that used in the mercury cell process to generate the K106 wastes. Ambient air checks of the area
surrounding the process equipment were conducted to determine fugitive mercury emissions.
PILOT PLANT TEST RESULTS
1. Ash Levels - Levels were consistently below the EPA requirements ranging between 0.03 and 64 ppm
total mercury. TCLP levels varied from less than 1 to 23 ppb.
2. Mercury Balance - Table 1 summarizes the overall mercury balance for the entire project. Individual
balances on each of the test runs varied widely. The reason for the large variation is believed to be differences
in the mercury content of the wastes fed to the furnace and relatively short running times. The mercury collected
in the carbon columns and the effluent treatment system can be fed back into the furnace with the bulk of it
recovered.
TABLE 1. OVERALL MERCURY DISTRIBUTION AND
BREAKDOWN OF OUTPUT STREAMS
Overall Mercury Distribution
Quantity, Pounds
Input Stream, Total 712
Output Stream, Total 705
Breakdown of Output Streams
Recovered Mercury 467
Mercury in Carbon Columns 210
Mercury in Effluent Treatment Solids 28
Mercury in Ash 0.04
Total 705
3. Mercury emissions to atmosphere through the stack were very low. Actual emissions were less than
0.003 Ib/hr (less than 0.006% of mercury fed to the furnace.)
101
-------
Ambient mercury levels were monitored throughout the project. In some instances mercury levels
exceeded the OSHA limits mandating the need for respirators. Early in the process testing, the unit was shut
down and modified to reduce fugitive mercury emissions, primarily from the furnace. Additional modifications
will be made in the commercial units to reduce fugitive emissions.
It is expected that the commercial facilities will be abletoachieve existing NESHAPregulationsfor mercury
cell chloralkali facilities.
4. AH tests were conducted in accordance with appropriate testing and sampling procedures and
applicable quality assurance/quality control protocols.
ALTERNATETECHNOLOGY
Thermal processing of certain mercury containing wastes has been practiced for many years through
batch thermal processors (retorts). This technology has been limited to the processing of relatively high levels
of mercury in the waste (typically in excess of 30%) and with small throughputs. Based on the historical data
available, the Institute and the participating members believed that this batch technology could not be applied
to high volume, low mercury content wastes. The newly developed technology can be scaled up or down
depending on the throughput required.
SUMMARY
The project allowed for the successful development of the technology to treat K106 and certain D009
wastes in accord with the LDR requirements. Four of the participating companies have initiated construction of
such facilities. These companies have been granted case-by-case extensions to the LDR to allow for continued
land disposal of these wastes until May, 1993. While the startup date for the facilities varies for the different
companies, all are working diligently to achieve a startup date by May, 1993.
Capital costforsuchafacility can vary depending on many factors, such as size, location, required utilities,
etc. It is believed that a facility to handle from 2 to 20 tons per day of waste would cost about $5 million. Raw
materials consist primarily of fuel for heating the wastes to the required temperature along with sodium
hydroxide for scrubbing. Staffing for long-term operations and maintenance have not been determined.
It has been reported in the press that one member company which is building a single facility to handle
the wastes form two chloralkali plants expects to invest $6.5 million and employ 10 people in its operations.
Based on past discharges of mercury to hazardous wastes landfills, mercury recovery should be about 25 tons
per year.
FURTHER INFORMATION
This abstract is developed from documents submitted to EPA on a confidential business information basis
in December, 1991 by the Institute in support of case-by-case extension applications of several member
companies. Although the mechanism has not been finalized, it is the intention of the participants to share this
technology to non-participants on a cost recovery basis.
The cost of the project was approximately $3.5 million. The technology is applicable to wastes generated
from mercury cell chloralkali processes and may have application elsewhere. The Institute is prepared to let
anyone who signs a secrecy agreement read the detailed report at the Institute's offices. Contact is Art Dungan
(Tel: 202 775-2790).
ACKNOWLEDGMENT
The technology described was developed through a task group of the Chlorine Institute, Inc. The
technology isownedbythe seven participating companies. These include ASHTA Chemicals, Inc., BFGoodrich
Chemical Corporation, Georgia Pacific Corporation, Olin Corporation, PioneerChior Alkali Company, Inc., PPG
Industries, Inc. and Vulcan Materials Company. These companies are applying for a patent for this technology.
The project manager for the members was J. W. Hutchins of PPG. The technology was developed at Hazen
Research, Inc., Golden, Colorado. Project manager for Hazen was C.W. Kenney.
102
-------
MERCURY REMOVAL WITH IONAC ION EXCHANGE RESINS
F. X. McGarvey
Sybron Chemicals, Inc.,
Birmingham, NJ 08011
USA
Tel: (609) 893-1100
Ion exchange resins have proven to be useful in the removal of mercury from aqueous streams,
particularly at low concentration levels (1 to10 ppb). The application of ion exchange resins requires an
understanding of the chemistry of mercury as it occurs in the natural environment. Mercury has been used
in the electrochemical industry as conductive liquor in the preparation of caustic. Mercury is also used in
gold mining and may be associated with cyanide complexes. Mercury will also occur in water supplies
contaminated with mercury as an industrial waste.
Ion exchange applications usually treat mercuric salts. For example, water supplies which contain
chlorides will frequently find the mercuric complex HgCI3- or HgCI4=. These compounds are quite stable and
can be picked up on anion exchange resins.
M[HgCI3-] + Resin Cl -> Resjn HgCI3 + MCI
The anion exchange resin can be regenerated with strong acid solutions, but this is done with difficulty
since the mercury salts are not highly ionized and have low conductivities in solution. Also mercury forms
complexes with many organic compounds including carbon itself. One of the most difficult salts from the
standpoint of the food chain is Hg(CH3)2, mercury dimethyl. Compounds of this type enter the food chain
resulting in mercury accumulation in fish. Since these compounds are not ionized, they are difficult to
remove by conventional ion exchange (1).
While conventional ion exchange resin will concentrate mercury complexes, these resins are not
particularly selective for mercury ions. Resin containing the iminodiacetic acid group will pick up mercury
selectively from calcium and magnesium. This type of resin can be helpful with cationic mercury. In
conjunction with anion exchange resins, mercury can be reduced significantly in potable waters.
Another type of resin contains a S-H+ functionality on the aromatic ring of a styrene-divinylbenzene
copolymer. This resin functions on the insolubility of sulfide compounds of mercury and forms almost
irreversible linkage with mercury so long as oxidation of the sulfur group is prevented. Regeneration is
possible with 20% sulfuric acid in cases where resin recycle is desired.
Other resins selective for mercury (2) are also under study. One promising resin is the macroporous
glycidylmethacrylate- ethylenedimethacrylate type copolymer which has a selective group formed when the
copolymer is contacted with ammonia
R-COOH2CH-CH2NH
I
OH
This resin has been shown to remove mercury from waste streams containing copper, cadmium, zinc,
and other divalent cations. Regeneration can be achieved with nitric acid. While there several groups
working on selective resin for mercury removal, none have shown the absolute selectivity of the -SH group.
The sensitivity of this functional group for oxidation indicates that caremust be taken if this resin canfunction
in the field.
103
-------
IONAC ION EXCHANGE RESIN STUDIES
Several studies are underway with lonac ion exchange resins. These are summarized in following
paragraphs.
Anion Exchange Resins
lonac AFP-329 and lonac ASB-1P have been used in southern New Jersey to treat a potable supply
which contains mercury. Initial tests showed that mercury was present as the chloride complex and a field
study has been underway for two years conducted by the State of New Jersey. This is an example of a
potable water study using resins which are included in the FDA list of acceptable resins shown in CFR
173.25.
Thio Resin
Several studies have been underway with SR-4 to remove mercury from waste streams. One
installation used this resin to remove mercury from battery wastes from an industrial plant. The mercury level
was reduced from 30 ppb to less than 0.5 ppb. No regeneration was attempted.
SR-4 was used to remove mercury from a cyanide solution containing gold and silver. This was
achieved without removal of gold and silver cyanide complexes which would have been absorbed with the
mercury complex if conventional resin had been used (3).
lonac SR-4 has been used to remove mercury from waste streams associated with mercury cell
caustic wastes and this effort continues at various plants.
Iminodiacetate Resins
Where mercury occurs as a divalent cation, SR-5 can be used to selectively remove mercury in some
cases. Copper and cobalt tend to compete effectively against mercury, but sufficient selectivity can be
achieved to make additional tests of interest.
ECONOMICS OF MERCURY REMOVAL
Most toxic streams contain mercury in low concentration so that a selective resin can be used in an
irreversible fashion with final disposal of the loaded resin in a hazardous waste site or in some type of
oxidative chemical destruction where the mercury is held or recovered. The cost savings can be substantial
as illustrated in the following example.
If mercury is present at 5 parts per billion (5 ug/L) level and must be reduced by 90% to 0.5 part per
billion, lonac SR-4 has a capacity of 0.7 equivalent per liter of resin:
Volume waste (5 ug/liter) = 0.7(201/2) grams per liter
0.7(100.5)
Volume = •
• = 15.6 x 109 liters/liter resin
4.5x10-9
It is extremely unlikely that an ion exchange bed would be able to treat 15x109 bed volumes without
serious packing and general failure of the column. A value of 10,000 bed volumes may be achievable in
cases where the waste is quite free of suspended solids. For this case a cubic foot of SR-4 values at $3007
ft3 could treat almost 40,000 gallons of waste and concentrate the waste into a cubic foot volume (7.5
gallons). The degree of concentration is more than 5000. Based on $100 to dispose of a liquor per 100
gallons ($1/gallon), the disposal of resin would be about $750 as compared to $40,000. Under these
conditions resins can easily justify a single cycle.
104
-------
REFERENCES
1. McGarvey.F.andHauser, E. Removal of Mercury from Industrial Wastes as Well as from Potable
Supplies. Chemistry in the Environment, Turino, Italy. September 1987.
2. Jehlickova, A., Svec, F. and Kalal, T. Reactive Polymers. XXVI. Sorption Properties of the
Glycidylmethacrylate- Ethylenedimethacrylate Copolymer Modified with Ammonia. Die Angewandte
Makromolekulare Chemie, 81, 87(1979).
3. McGarvey, F. Selective Ion Exchange Resins in Gold Processes: Mercury Removal and Gold
Recovery. 95th Annual Convention, Northwest Mining Association, Spokane, Washington.
December 1989.
105
-------
DEVELOPMENT OF BACTERIAL STRAINS FOR THE REMEDIATION OF MERCURIAL
WASTES
Joanne M. Horn
Center for Environmental Diagnostics and Bioremediation
University of West Florida
11000 University Pkwy.
Pensacola, FL32514
USA
Maren Brunke, W.-D. Deckwer, and K.N. Timmis
National Research Center for Biotechnology
Mascheroder Weg 1
3330 Braunschweig
Germany
INTRODUCTION
The biotoxicity of mercury (Hg) resides in the ability of charged forms of the metal to bind and inactivate
enzymes. Organomercurial compounds also accumulate in higher organisms and cause systemic disease.
While Hg(ll) and Hg-containing compounds are highly toxic, they are still being anthropogenically mobilized into
the biosphere at sufficient rates to create environmental hazards.
Since the total amount of extant Hg is finite and unchangeable, remediation of Hg pollution can only be
aimed at altering its ionic form to a less toxic species and/or sequestering it, ideally in forms which can be
recycled for further use.
An alternative remediation strategy currently being examined involves converting Hg(l I) to the more inert,
volatile elemental form [Hg(0)] by employing the mechanism of bacterial Hg resistance which relies on the
reduction of Hg(ll) to Hg(0) through the activity of a cytosolic mercuric reductase. Such bacterial reduction
systems might be utilized for either on-site remedial situations, since volatilized Hg(0) is both less toxic and less
bioavailable, or within contained bioreactors from which the reduced Hg(0) could potentially be recovered.
Specific bacterial mercury resistance (Hgr) involves both the mercuric reductase and, since Hg(ll) also
inactivates membrane proteins, a Hg-specific transport system which protects the membrane and other cellular
components by shuttling Hg from the cell surface to the intracellular mercuric reductase.
Narrow spectrum Hgr confers resistance only to mercuric ions. Broad spectrum mercury resistance
includes, in addition to the transport and reductase enzymes, an organomercurial lyase which protonolytically
cleaves carbon-Hg bonds. The broad spectrum Hgr systems, through the successive activities of the lyase and
the reductase, additionally provide resistance to organomercurial compounds.
Both narrow and broad spectrum Hgr mechanisms have been found in a wide range of bacteria. Hgr
structural genes are encoded in a single operon that is primarily regulated by the tnerR gene product, which is
itself transcribed contiguous to, but in the opposite direction from, the merTPABDoperon. The merTand merP
gene products are involved in Hg(ll) uptake and transport, merA specifies the mercuric reductase, and merB
encodes the organomercurial lyase. merD, the most promoter-distal gene identified, has been associated with
a transcriptional coregulatory function.
We report here the generation and selection of Pseudomonas putida strains with heightened mercurial
detoxification properties by constitutive hyperexpression of the merTPAB genes. We have also genetically
combined the ability to reduce Hg(ll) at increased rates under conditions of high concentration with a benzene
degradative pathway. Derivative strains are thus able to dissociate the chemically dissimilar components of an
organomercurial compound, phenylmercuric acetate (PMA), into its metal and aromatic elements, and
separately detoxify each. While these strains were originally conceived for intended use in contained
bioreactors for treatment of industrial waste streams, they have attributes which may contribute to on-site
treatment of Hg-containing pollutants in contaminated soils and waters as well.
106
-------
This project was conducted underthe auspicesof the Biochemical Engineering and Microbiology Divisions
of the National Centerfor Biotechnology, Braunschweig, Germany, where process development utilizing these
strains is currently in progress.
PROCESS DESCRIPTION
Constitutive overproduction of proteins conferring detoxification of and resistance to organomercurials
was accomplished by random mutagenesis of two strains of P. putida with a mini-transposon containing the
merTPAB genes. The mini-transposon encodes the organomercurial resistance determinants flanked by the
inverted repeats of Tn5, which have been abbreviated to 19 basepairs. The transposase gene (tnp) is located
externally with respect to the TnStransposing element, and merRand merD, specifying mer regulatory proteins
were eliminated from the construct (1).
Since mini Jn5-merTPAB insertions into recipient chromosomes were assumed to be generally random
with respect to target locii, it was reasoned that insertions downstream of proximal host promoters could result
in increased mer expression, with a coincident rise in Hg(ll) and PMA resistance (PMAr). Thus, subsequent
selection for mini-Tn5me/TPAB-mutagenized P. putida derivatives which demonstrated increased resistance
to organomercurials was undertaken. Isolates were selected which showed up to 86% increase in Hgr over a
control P. putidastram harboring a singlecopy of a wildtype narrow range mercurial resistance operon (encoded
by Tn507) under control of its own promoter. Resistance to organomercurials was also increased: some
derivatives demonstrated growth in up to 80ug/ml PMA, while 10ug/ml is normally used as the minimal inhibitory
concentration of this compound.
Since the enzymatic activity of the merA gene product, mercuric reductase, is assayable in crude cell
extracts, it was used as a measure of mergene expression. Up to a four fold increase in mercuric reductase
activity (above the Tn507-containing control strain) was found in given isolates. Furthermore, the expression
of merA, in contrast to the activity of wildtype mer operons, was found to be constitutive; the addition of Hg(ll)
was not required for mini-Tn5-encoded merTPAB expression.
Elevated merTPAB expression corresponded with increases in Hgr and PMAr. A statistically significant
correlation between mer expression, as indicated by mercuric reductase activities, and resistance to
(organo)mercurials was demonstrated. These findings were in sharp contrast to previous studies which have
attempted to produce the same results in Escherichia coll. It has been shown that as mer expression increases
as a result of heightened gene dosage in E. coll, but these increases are not manifested in a concomitant rise
in Hgr of whole cells; rather, increased expression is cryptic (2-4).
It has been shown that Hgr in E coll is limited by Hg(ll) transport, and it was postulated that a limitation
in the availability of membrane insertion sites for the MerT transport protein may account for this observed lack
of correlation between mer expression and resistance (4). The contrasting correlative relationship observed
in P. putida not only allowed selection of overproclucers, butcontributestowardstheirpotential utility in mercurial
treatment schemes (see following).
Growth rates of P: putida merTP/4B-mutagenized derivatives were determined to assess the effect of
merTPAB hyperexpression on culture growth rates. While a P. putida strain containing a wildtype inducible
narrow spectrum mer operon was unable to grow when initially exposed to toxic concentrations of HgCI2, the
P. putida me/TP/lB-hyperexpressing isolates showed no effects of Hg(ll) on culture growth, even when
inoculated into media containing Hg(ll) at normally toxic concentrations.
One P. putida host strain mutagenized with mini-Tn5mer7Py4B, P. putida F1, contains a chromsomally-
encoded pathway forthe complete catabolism of benzene and toluene (5). P.putidaFI is capable of using these
compounds as a sole source of carbon and energy, resulting in their total mineralization. Since the mini-Tn5
merrPAB-containing F1 derivatives encode the organomercurial lyase they are presumably capable of cleaving
the Hg(ll) moiety from PMA, and using the produced benzene as an energy source. When tested for this ability
We found these derivatives were indeed able to grow on minimal salts media with PMA as a sole carbon source.
By coupling Hg(ll) cleavage and reduction with benzene degradation, a novel catabolic pathway was generated.
107
-------
FEED STREAM CHARACTERISTICS
The described P. putida merTPAB hyperproducing strains are highly resistant to (organo)mercurials,
demonstrate increased rates of Hg(ll) reduction, and constitutiveiy express these capabilities. Thus they can
survive in and detoxify highly contaminated mercurial-containing wastes. This may be a virtue in dewatered
effluents for example, or for potential on-site remediation efforts.
Cell death due to toxic effects of high Hg(l I) concentrations before expression can be fully induced, which
was observed in P. putida strains containing the wildtype Hgr system, is prevented. Likewise, in mercurial
wastes which are too low in concentration to induce the reducing capability of the wildtypeoperon, these isolates
would also prove useful, as their reductive capability is completely independent of Hg concentration.
Combining the overexpression of Hg-C cleavage and reduction with benzene/toluene catabolism allows
the resulting strain to detoxify Hg and benzene components, both chemically dissimilar moieties of PMA,
separately by different pathways present in the same recombinant organism. Mixtures of organomercuric
compounds could potentially be treated, as the lyase has wide substrate specificity and can cleave Hg from a
variety of organomercurials; the reductase would then act on the produced Hg(ll). Since this process is
enzymatically based it specifically targets mercuric salts and organomercurial compounds.
When benzene or toluene are present, either as copollutants or as the organic component of an
organomercurial, they would be mineralized completely. Since we have shown that these activities are
sufficiently high to support growth on PMA, it may be possible to drive this system without requiring the addition
of exogenous nutrients. Furthermore, since benzene is totally catabolized, there is no possibility of producing
toxic aromatic endproducts which might hinder further cellular metabolism.
PRODUCT CHARACTERISTICS AND USE
Initial process development of bioreactors utilizing Hg(ll)-reducing bacteria hasshown that reduced Hg(0)
can be retained within a fixed bed bioreactor. This offers the possibility of reclaiming Hg(ll)removedfrom reactor
input in a concentrated, significantly less toxic and potentially reutilizable form.
ENVIRONMENTALADVANTAGES
Currently available technologies for Hg removal involve either ion exchange oradsorptive matrixes. These
methods have the disadvantages of being non-selective, they can generate a secondary solid waste, are costly,
and since Hg is non-selectively and/or permanently bound, it cannot be recovered for further recycling. Since
the biotransformative system described here however is enzymatically driven, it is selective. Furthermore it
does not utilize expensive resins, and the reduced Hg(0) can potentially be recovered from closed bioreactor
systems. ,
Since these bacterial strains are derivatives of naturally-occurring soil organisms, the risks of deliberately
releasing them for on-site bioremediation may be significantly reduced. They also have the ability to survive
under normal environmental conditions. Their complement of Hg reducing genes are stably inserted into the
chromosome and by themselves, have no capability to transfer to other locii within the same organism or to the
native biota.
The product of this system, Hg(0)l, is relatively inert in comparison to its substrate, and thus less toxic to
biological organisms. Hg(0) is also less bioavailable, since under a wide variety of conditions it is volatile, or
at least insoluble. The bioreduction of mercury therefore produces a form of the metal which has significantly
less impact on the biosphere.
REFERENCES
1. Herrero, M., de Lorenzo, V., and Timmis, K.N. Transposon vectors containing non-antibiotic
resistance selection markers for cloning and stable chromosomal insertion of foreign genes in
gram-negative bacteria. J. Bacteriol. 172:6557,1990.
2. Nakahara, H., Kinscherf, G., Silver, S., Miki, T., Easton, A.M., and Rownd, R.H. Gene copy number
effects in the meroperon plasmid NR1. J. Bacteriol. 138:284,1979.
108
-------
3. NiBhriain, N., Silver, S., and Foster, T.J. Tn5 insertion mutants in the mercuric ion resistance genes
derived from plasmid R100. J. Bacteriol. 155:690,1983.
4. Phillippidis, G.P., Malmberg, L-H., Hu, W-S., and Schottel, J.L. Efffect of gene amplification on the
mercuric reduction activity of Escherichia coll. Appl. Environ. Microbiol. 57:3558,1991.
5. Gibson, D.T., Koch, J.R., and Kallio, R.E. Oxidative degradation of aromatic hydrocarbons by m i -
croorganisms. I.enzymatic formation of catechol from benzene. Biochemistry 7:2653,1968.
109
-------
THE RECOVERY OF MERCURY FROM MINERAL EXTRACTION RESIDUES USING
HYDROMETALLURGICAL TECHNIQUES
Robert G. Robins and Lakshman Jayaweera
HydroMet Corporation Limited
31-45 Smith Street
Marrickville NSW 2240
Australia
Tel: 61 2 517 1188
Fax: 61 2 519 9468
HydroMet Corporation Limited commenced its business in Sydney, Australia in 1987. It has since
become a public company listed on the Sydney Stock Exchange. In April 1992 a new share issue to raise
capital for plant development was oversubscribed and the company will proceed to the construction of a
treatment plant at Newcastle and additional plant at Port Kembla (both in New South Wales, Australia).
The company's principal activities are in the industrial waste management and resource recovery
sectors of the mineral processing industry. Its activities include the processing of industrial smelter residues
and the manufacture of a range of metal and chemical products from those residues. Presently the company
is the largest producer of high grade selenium in the southern hemisphere. HydroMet undertakes general
hydrometallurgical research and development specialising in contaminated waste treatment.
PROCESS DEVELOPMENT
The basis of process development for mercury removal and recovery from mineral concentrates and
residues is thermodynamic modelling of the system, taking into account the speciation of mercury in
aqueous solution and the solubilities of both mercury(l) and mercury(ll) compounds. Computer modelling
techniques involve the use of a number of well-known codes such as MINTEQA2 but also an in-house
program, DIASTAB, for plotting stability diagrams. Some of the diagrams that have been generated with
DIASTAB will be used to illustrate the stabilities of mercury species in relation to recovery processes.
The selective leaching of mercury from some concentrates and solid residues can be accomplished
by the formation of a range of soluble mercury (I I) species that can be obtained by complexation with certain
reagent Hgands (such as chloride, bromide, chlorobromide, cyanide, bromocyanide, ammonium, acetate,
citrate, EDTA, etc.).Thepotential effectiveness of these reagents is illustrated bytheir effect onthe solubility
of mercuric oxide.
The removal of mercury from hydrometallurgical solutions has been accomplished by the unit
processesof precipitation, ion exchange, solvent extraction, adsorption, membrane diffusion, cementation,
electrolysis, precipitate and ion flotation, and using biological methods. The most common process is the
precipitation of an insoluble mercury(l) or mercury(ll) compound (such as oxide, carbonate, sulfide, sulfite,
sulfate, phosphate, basicsulfate, iodate, and several organic mercury(l) compounds). The relative solubility
of some of these compounds will be illustrated with stability diagrams.
HydroMet has been involved in developing a number of commercial (or tentatively commercial)
processes for the removal and recovery of mercury from mineral concentrates and solid residues. The
mercury end-product in most cases has been a high-grade mercury concentrate that is sold on the
international market. Some of the treatment processes that have been developed and some of the
processes investigated are described briefly in following paragraphs.
SELECTIVE LEACHING OF SULFIDE CONCENTRATES
The presence of mercury in sulfide concentrates causes problems in most cases and results in
commercial penalties. In some concentrates the mercury exists as a separate simple mineral such as
110
-------
cinnabar and can often be selectively dissolved. HydroMet has found that an acidic (HCI) chiorobromide
leach with a hypochlorite-bromine oxidant was extremely effective in the case of one particular Australian
complex sulfide concentrate. Initial mercury concentrations as high as 6 g/kg, in one particularsample, could
be reduced to less than 5 mg/kg in the first stage. A commercial process has been developed for this
concentrate in which the mercury is recovered from the leach solution as a mercury(l) sulfite precipitate.
Many variations of the mercury sulfide/halide leach process have been investigated by HydroMet
since most processes have been concentrate-specific. In the case of more complex mercury sulfide impurity
minerals, which often contain arsenic, antimony, bismuth, thallium, selenium etc., more complex leach
procedures have been adopted in laboratory investigations, but mercury recovery in those cases was more
difficult.
A more conventional leach process for cinnabar is hypochlorite leaching (sometimes electrolytically
assisted) followed by either carbon adsorption or zinc cementation to yield a Zn-Hg amalgam.
CEMENTATION OF MERCURY
The recovery of elemental mercury from complex leach solutions with iron, zinc, and aluminium has
been investigated with a view to several different specific applications. Some of these will be mentioned.
RECOVERY OF MERCURY FROM GOLD-CYANIDE SOLUTIONS
Mercury often occurs in precious metal bearing minerals and forms soluble cyanide complexes during
a cyanide leach process. The mercury is then finally deposited on the adsorption carbon, on the electrolysis
cell cathodes and in the tailings pond. In all cases there are further problems.
A number of ion exchange resins (including Duolite A-7, lonac A-305, and Schering TN02327) have
been investigated for removing both the mercury and gold as cyanide complexes from the leach solutions.
Gold and mercury can be selectively eluted but not with a complete separation. The ion-exchange process
appears to have an application in some Australian situations.
SELECTIVE PRECIPITATION OF MERCURY SULFIDE
Mercury has been.selectively precipitated with thioacetamide to yield sulfide from a copper-mercury
solution obtained by sulfuric acid leaching of a hydrometallurgical process residue. Mercury sulfide is
insoluble over a broad range of pH but forms sulfide complexes with sulfide and bisulfide ions which result
in a small degree of solubility in neutral-to-alkaline solutions. Thioacetamide will quickly precipitate mercury
sulfide from a sulfate solution at a pH of about 2 so that it can be removed by filtration before significant
copper is also precipitated.
Thioacetamide and thiourea are useful sulfide precipitating reagents at low pH since the generation
of hydrogen sulfide, as would be obtained with sodium sulfide or bisulfide solutions, is minimal.
TREATMENT OF MERCURY-SULFUR RESIDUES
Sulfuricacid is often produced from SO2gas generated by roasting sulfide concentrates which contain
trace levels of mercury. The roaster gas is contaminated with mercury and this is usually partly removed
by weak acid scrubbers and mist Cottrells. The wash liquor from these units can then be treated to remove
the dissolved mercury by cementation with aluminium metal pellets to produce a solids residue which is
predominantly elemental mercury and sulfur in strong physical association (about 15% Hg and 70% S, with
Pb and SiO2 among other constituents).
This residue is treated in a HydroMet-designed plant to upgrade the mercury by leaching the sulfur with
sulfite solution and then recovering a high-grade elemental sulfur. The mercury residue can then be retorted
to produce elemental mercury that is converted to a range of chemicals.
111
-------
RECOVERY OF MERCURY AND SELENIUM FROM ROASTER GAS
Another process for removing mercury from roaster gases is a sulfatization procedure similar to that
adopted at the Outokumpu zinc plant at Kokkola in Finland in 1970. After dust removal from the mercury
bearing off-gases, they are contacted with a recirculating 85 to 90% sulf uric acid in a sulfatizing tower and
then in a weak acid scrubber to remove chloride which is in the form of HCI gas and HgCI2 gas. When
selenium-bearing concentrates are treated in this way the product from the sulfatizing unit contains
selenium. This product is washed to remove the soluble salts leaving a complex mercury-selenium
precipitate which can be represented by the formula:
HgSO4.xHg(0,S,Se,Te).
The HydroMet process for recovering both mercury and selenium from this residue involves a
controlled potential sulfite-chloride leach procedure in which mercury can be precipitated as Hg2SO3 and
selenium recovered as the element.
ETHYLENE LEACHING
HydroMet is currently developing a process in which ethylene gas is the reagent used to form a strong
complex with mercury(ll). The process is operated at a gas pressure of 4 atmospheres and after the leach
procedure HgO can be precipitated by releasing pressure.
112
-------
HIGH VACUUM MERCURY RETORT RECOVERY STILL FOR PROCESSING EPA
D-009 HAZARDOUS WASTE
Bruce Lawrence
Bethlehem Apparatus Company, Inc.
890 Front Street
Hellertown, PA 18055
USA
Tel: (215) 838-7034
INTRODUCTION
Bethlehem Apparatus Company developed a continuous-feed vacuum triple distillation process for
mercury in 1955. Since then, high purity instrument grade triple distilled mercury has been sold to industrial,
institutional, and educational accounts principally in the US and more recently to several foreign accounts.
In the early 1970s Bethlehem developed a mercury retort process for the recycling of fever and industrial
thermometers. This process has gone through a variety of changes over the past 15 years and now is a
relatively high-vacuum system that can handle a wide variety of mercury-bearing waste products.
PROCESS DESCRIPTION
Mercury-bearing materials are placed into the retort either in a 55-gallon steel drum or on trays. Larger
materials can be placed directly into the retort without the support of a container. The retort is a bell type
unit that is hydraulically raised and lowered. Once the materials are loaded onto the stationary base, the
top bell unit is lowered and fastened. The joining members of the retort are sealed with a water cooled
gasket. Once sealed, a vacuum is drawn down to approximately 25 inches mercury column. Condensers,
water traps, and particle traps are located between the retort and the vacuum pump.
Inside the retort are electrical radiant heaters that raise the inside temperatures to 1250 degrees F.
Mercury is vaporized at these temperatures and directed to the condenser where the saturated air stream
of mercury vapor is condensed because of the water cooled jackets. Condensed mercury is allowed to spill
into a reservoir for transfer to our triple distillation process.
Water vapor or dirt particles that are carried with the mercury vapor pass through the mercury
condenser and are subsequently trapped in chambers between the vacuum pump and the condenser. If
there are large amounts of water in the material processed, then water will spill into a trap located directly
on the vacuum pump.
Exhaust air from the vacuum pumps is drawn through a series of treated activated charcoal absorption
units to remove mercury vapor.jchlorine and acid fumes, and odors. The final exhaust air is vented directly
into the condenser room where it can be readily measured from mercury vapor content. Since the exhaust
air is vented into a work area, the mercury vapor levels are kept to below .05 milligrams per cubic meter
(current OSHA standard). (See Figure 1.)
Major process streams include mercury batteries, thermometers, mercury switches (glass and metal),
quartz lamps, thermocouples, ig nitron tubes, fluorescent lamps, and absorption-activated charcoal.
Currently aqueous solutions cannot be handled since a large proportion of the mercury salts are carried over
in the distillation process. Materials like mercury chloride also cannot be handled since the system is
predominantly stainless steel.
Major equipment characteristics include:
A stainless steel chamber capable of holding a~29-inch mercury column vacuum at 1,350 degrees
temperature
113
-------
Refrigeration
Compressors
VACUUM
RETORT
HEATING
CHAMBER
Water
Chilled
Condenser
Solids from charcoal
air and water treatment
recirculated
cooling
water
mercury vapor
absorption
Single
Distillation
condensed mercury
Particle
Filter/Water
Trap
Triple
Distillation
chlorine and acid
vapor absorption
! I *
Triple
Distilled
Mercury
Customers
Charcoal absorption
>of mercury from
waste water
4 t
ToPOTW
exhaust
to room
odor absorption
Exhaust
Vacuum
Pump
Figure 1. Mercury vacuum retort system.
114
-------
A stainless steel connection from the heated vacuum chamber to the water cooled stainless steel
condenser
A stainless steel condenser with independent waste connections for water cooling flow
Water pump to transfer water through condenser and return to chilled water source
Chilled water source including independent pumps, compressor, and temperature controllers
Filter media to trap dirt particles and retain condensed water
Vacuum pump capable of pulling 28-inch mercury column
Water cooling attachment to vacuum pump connected to independent chilled water supply
Activated charcoal media to absorb mercury vapor, chlorine or acid fumes, and odors
FEED STREAM CHARACTERISTICS
Typical feed stream characteristics include standard metal and glass materials. Most plastics can be
processed; however, polyvinylchloride is kept to a minimum because of the chlorine content. Other halogen
materials are also kept to a minimum. The only metals rejected are lithium, arsenic, and thallium. Quartz
glass is accepted; however, it must be crushed if the mercury is contained inside a hermetically sealed unit.
Dirt, soils, and sludge-like material can be processed as long as there is not a high degree (over 40%)
of water content, if mercury is contained in an aqueous solution then it must be filtered or precipitated into
a fairly solid form before the material can be retort processed. If feed streams have a high proportion of water
then the water carries a large proportion of the solids during distillation. The solutions could be run at lower
temperatures to distill off the water but this takes a long time and causes problems in the distillate collection
systems.
Radioactive materials are not accepted, nor are explosives, acids, and alkaline materials. Due to
regulatory restraints we can not process any materials from wastewater treatment processes or any listed
wastes with K, U, or P codes. We also do not handle any organomercurials.
- Our system is a batch-feed process that handles 55-gallon drum volumes of material for each batch.
If a feed stream is a large quantity of low mercury concentration (less than 5%) we would preferthe generator
concentrate the mercury intosomething like charcoal filter media where the mercury levels are much higher.
PRODUCT USE
All mercury recovered from the retort process is pumped to our scrap mercury processing system.
Here the mercury is pumped into a single distillation unit that subsequently feeds into our continuous-feed
vacuum triple distillation units. The distilled mercury is then packaged and sold to industrial accounts for use
in items like fluorescent lighting, dental amalgams, and mercury switches.
Scrap metals that are recovered from the mercury retorts are sold to scrap metal recycling
organizations. We have recycled steel, aluminum, brass, copper, nickel/cadmium, lead, and zinc in this
manner.
Oil is sometimes generated from the condensation of certain types of plastics that are run through the
retort. In these cases the oil is mixed with our vacuum pump waste oil, which is filtered and run through
mercury absorption activated charcoal. The cleaned oil is sent off-site for fuel blending.
RESIDUAL CHARACTERISTICS
There are seven types of residual materials from the retort process:
Mercury - triple distilled and sold, approximately 25 tons a year.
115
-------
Glass, dirt, and other residuals (referto Table 1)-sentto industrial landfill for non-hazardous waste
approximately 20 tons a year.
Metals sent to scrap metal recycling centers, approximately 20 tons per year.
Water, which is pH balanced, filtered, and treated in activated charcoal for mercury removal.
Processed water is than sent to local sewage treatment plant. Current amounts average 20 to 50
gallons per day.
Oil from the retort and vacuum pumps. Filtered, treated, and fuel blended. 50 to 100 gallons per
year. . _
Solids from the charcoal absorption and waste water treatment. Returned to retort for mercury
removal.
Residues from still bottoms that contain low-level mercury with high levels of lead, cadmium, and
silver. This is sent off-site for treatment and disposal in hazardous waste landfills. Annual quantity
is approximately 1/4 ton.
TABLE 1. TOTAL MERCURY AND TCLP TEST RESULTS ON RESIDUAL MATERIALS THAT
WERE PROCESSED IN THE BETHLEHEM VACUUM RETORT SYSTEM (1)
SAMPLE
TOTAL MERCURY
MGKG
TCLP EXTRACTION
Mercury Capillary Tubes
Switches, Metal
Charcoal, activated
Condenser Muds
Plastic Relays
Tyvek &, Rubber Gloves
Switches.glass
Zinc air cells
4.47
30.2
1.13
9.2
65
.57
2
1.72
.068
.0058
<0.01
.001
.0025
<0.001
.0002
.0003
ENVIRONMENTAL ADVANTAGES
Bethlehem's high-vacuum retort system offers the greatest environmental advantage over other
systems because of the very low air emissions. Since our exhaust air is low in volume (approximately 1 cubic
foot per minute per retort) we can readily scrub all the mercury from the air prior to emission. While this can
theoretically be done with larger-volume air systems, the amount of scrubbing material required is much
greater. In addition there is virtually no fugitive air emission from the outside of the retort. Low-vacuum
systems will give off high levels of mercury vapor which causes a work environment hazard as well as
emission problems.
FACILITY DESCRIPTION
Bethlehem Apparatus Company isa recycling facilityfor characteristic D009 mercury. We have seven
retort recovery furnaces operational, withfourmore updated units under construction. Bethlehem will need
to use 1 of the 11 retorts to process its own material. The remaining 10 units are available for processing
materials from outside generators. Each retort unit can process 35 tons of material per year. Since the
process does not require permitting, we can increase our capacity to 22 retort units in our existing facility.
REFERENCE
1. All analytical data performed by Benchmark Analytics, Hellertown, PA 18055
116
-------
NON-THERMAL PROCESSING OF K106 MERCURY MUD
Michael Rockandel
Innochem Engineering Ltd.
#850 -1441 Creekside Dr.
Vancouver, BC
Canada V6J4V3
Tel: (604) 736-3381
Fax:(604)736-3110
INTRODUCTION (1,2)
Electrolytic mercury cell chlorine producers use a sulfidation technique in wastewater treatment to
achieve Department of Energy (DOE) standards for water disposal. The sulfide treatment process
generates an Environmental Protection Agency (EPA) designated K106 mud which is classified as a
hazardous high mercury subcategory waste. The K106 mud was banned from landfilling on May 8,1991.
Industry has continued to dispose of the material to landfill under a 2-year national capacity variance.
In the "Regulations on Land Disposal Restrictions" (40CFR 268.41) the EPA has designated retorting
as the Best Demonstrated Available Technology (BOAT) and the standard treatment technology for K106
mud. The EPA recognized the potential deleterious impact upon retorting of chloride salts in the K106 mud
and identified the possibility of pretreating prior to retorting. On April 16,1992 the EPA granted permission
to Georgia-Pacific to utilize the Innochem non-thermal treatment process to treat their K106 mud generating
a low mercury subcategory disposable waste and a high mercury subcategory concentrate requiring
thermal processing.
Innochem Engineering has developed a hydrometallurgical process for the treatment of K106 mercury
contaminated wastes. The process can achieve a measure of performance equivalent to that offered by
retorting. The process offers advantages over the retorting alternative including: no atmospheric emission
potential, operator familiarity, continuous operation, and low capital cost.
The process has been tested in the laboratory and in a continuous mini-plant. The test work was
conducted at the Montana College of Metallurgy, Science and Technology in Butte, Montana under the
guidance of L. Twidwell. Chemical analysis of the treated muds and extracts was performed at certified state
laboratories. Muds from three different chloralkali plants, a decommissioned chloralkali plant, and
elemental mercury contaminated soils have been tested.
A commercial plant is being designed for Georgia-Pacific in Bellingham, Washington. The plant will
have a capacity of about 200 tons/year of K106 mud and is scheduled for startup in the fall of this year.
TESTPROGRAM
The mercury content of the K106 solids can vary from under 1 % to near 20%. The typical mercury
content however is 1 to 5%. The mercury is in various forms including: mercurous and mercuric chlorides,
sulphide, elemental mercury, and species adsorbed on clay and activated carbon components. Total
mercury extraction is favoured by a combination of both acidification and oxidation. The on-site abundance
of inexpensivesodium hypochlorite and drying acid (sulfuric acid) ma'dethesethe logical reagents. Areview
of the available literature and simulation of the chemistry with Eh-pH predominance diagrams streamlined
the hunt for the optimum process parameters.
Several alternatives were considered for recovery of mercury from the leach solution. Low cost,
simplicity of application, lowtoxicity, and ease of removal from solution made iron cementation the preferred
route.
117
-------
An extensive test program was undertaken comprising both laboratory investigation and continuous
pilot confirmation. Several leaching variables were examined with the key ones being: pH, Oxidation
Reduction Potential (ORP), temperature, type of acid, residue washing, and contact time. Cementation
variables explored were: pH, retention time, type of iron, and contactor design.
The program successfully developed a treatment technology for K106 mud. The residue analysis was
typically 60 to 100 ppm total mercury with Toxicity Characteristic Leaching Procedure (TCLP) <0.025 mg/
L. The key test findings are summarized below:
A two-stage leach is preferred.
Leaching kinetics are very fast.
Sodium hypochlorite must be added well in excess of stoichiometric levels to be effective.
Elevated chloride tenor (10%) is essential.
Moderate temperature elevation (50 degrees C) is beneficial.
There is no performance advantage of hydrochloric versus sulfuric acid.
Residue washing is critical. A brine wash is required at high mercury levels.
Toachieve the very low mandated TCLP levels (<.025mg/L), stabilization of the residue with sulfide
is preferred.
Cementation is best performed in a rotating contactor.
Inexpensive ground steel scrap is an effective cementation media.
Optimum cementation pH is in the range of 2 to 4.
PROCESS DESCRIPTION
The Georgia-Pacific plant in Bellingham, Washington will employ the Innochem leach technology.
Significant variation has been observed in the composition of the muds tested which could result in
modification of some of the leach parameters. The following description which is primarily directed at the
Georgia- Pacific mud should be considered a process generalization.
Currently at the Georgia-Pacific facility, wastewater consisting of runoff, spills, and brine purges is
sulfidized at ambient temperature at a pH near 5. The low percent solids sulfidization product is clarified in
Adams Tube Filters yielding a water effluent which achieves DOE disposal guidelines. The backflushed
solids are thickened and filtered on a rotary drum filter. The cake is disposed of in restricted landfill (K106
mud).
The wastewater treatment process will be modified to include a clarifying thickener between the
sulfidation reactor and the Adams Filter. The reduced solids loading feeding the Adams filters will
substantially increase the operating cycle and life of the filtertubes while reducing the usage of filteraid. The
impact upon waste generation could be significant with current K106 mud containing 60 to 80% by weight
filteraid. The thickened solids are delivered to the treatment process at 10 to 20% by weight solids.
The first stage leach conducted at a pH of 5 to 6 extracts 98.5% of the mercury from the K106 mud
(30,000 ppm Hg dry basis) producing a residue of about 300 ppm mercury. Sodium hypochlorite is
proportioned to the incoming mud flow and is typically 70 to 100 #/DT of K106 mud. The leach product is
then transferred to the first of two vertical wash towers (thickeners). The overflow solution is transferred to
cementation while the washed and thickened residue is pumped to the second leach.
The second leach is aggressive with pH = 2 to 3. Sodium hypochlorite is added at 20 #/ DT of initial
K106 mud. Thesecond leach reduces the mercury to under 100 ppm, increasing the recovery to 99.8%. The
second leach product is washed in the second vertical washing tower using either plant waste brine or
118
-------
cementation product solution. The tower overflow solution is used as a washing fluid in the first wash tower.
The underflow will be dewatered on an existing conventional rotary vacuum filter. At Georgia-Pacific the filter
cake will be repulped and refiltered with the K071 brine mud prior to disposal. The residue designated as
a K106 and K071 low mercury subcategory waste wilt be disposed of in a restricted landfill.
The first wash tower overflow containing 1,000 to 2,000 mg/L mercury flows to the cementation step.
The pH is reduced to the range of 2 to 4 prior to flowing into the rotary contactor. Mercury cements onto the
surface of the iron and reduces to <1 mg/L in under 30 minutes of contact time. Iron dissolves to produce
about 2,000 mg/L ferrous ions. The cementation solution can be oxidized with subsequent iron removal and
be recycled to the second washing thickener or alternatively can be wasted to the sulphide treatment step.
The soft mercury amalgam formed on the surface of the iron is attrited by the mill's tumbling action and
overflows the contactor.
The cement at about 40 to 50% mercury is dewatered and washed on a filter press prior to packaging
in polyethylene drums. The cementation product containing 99.8 % of the starting mercury represents about
7.5% of the starting material weight. The cement impurities are essentially iron and its alloying agents. The
cement is an ideal feedstock for thermal refining and will be custom refined.
FEED STREAM CHARACTERISTICS
The composition of K106 wastewater treatment mud can vary substantially depending upon
the flexibility built into the chloralkali plant, the quality of the feed sodium chloride, and the plant's operating
status. The K106 mud can be contaminated with soil, carbon filter backflush material, brine, mud, and
f loorspills. Several K106 muds have been examined by SEM-EDAX. The range of analysis is shown in Table
1.
TABLE 1. ANALYSIS OF K106 MUD (DRY BASIS)
ELEMENT
RANGE OF ANALYSIS (%).
TYPICAL ANALYSIS (%)
s
Fe
Si
Mg
Cl
Ca
Na
Al
K
0.7 -
0.5 -
16
0.0 -
1.0 -
0.1 -
1.0 -
1.7 ~
• 0.2-
2.8
5.7
40
24.4
14.0
11.0
7.6
4.2
0.5
2.7
1 2'
37
2
' 1
4
2
2
0.3
The mercury can be present in various forms but mercuric sulphide is predominant. Other forms of
mercury which are believed to be present include: mercuric and mercurous chloride, elemental mercury,
and species adsorbed on clay and activated carbon particles. The analysis of mercury can again vary
dramatically. Waste, high in activated carbon can contain up to 20% mercury. The analysis of the materials
examined in this study is shown in Table 2.
The sulfide precipitate is very fine and is typically under 5 microns in size. The K106 will generally have
40% of its mass contained in the minus 5 micron fraction. This largely accounts for the high percentage of
filteraid added to the wastewater clarification step.
PROCESSADVANTAGES
The Innochem mercury treatment process offers many advantages over the alternative retorting
(BOAT), the most important being:
Produces a mercury concentrate suitable for thermal refining
119
-------
Technology familiar to operators of chloralkali plants
Technology which uses common on-site chemicals
Process without atmospheric emission potential
Technology which operates continuously
Technology applicable to D009 wastes
Low energy requirements
Low capital cost
TABLE 2. MERCURY CONTENT OF K106 MUDS (DRY BASIS)
SAMPLE DESCRIPTION
MERCURY ANALYSIS (%)
Georgia-Pacific #1
Georgia-Pacific #2
Georgia-Pacific #2
Company X
Company Y
3.0
3.3
3.5
3.9
0.8
REFERENCES
1. Twidwell, L.G. Treatability Study - Recovery of Mercury From K106 Waste - Phase II Final Report,
March 1992.
2. Scheiner, B.J., Lindstrom, R.E. and Shanks, D.E. Recovery of Mercury From Cinnabar Ores by
Electrooxidation, Report of Investigation 7750, United States Department of the Interior, Bureau
of Mines.
120
-------
BIOLOGICAL AND PHYSIO-CHEMICAL REMEDIATION OF MERCURY-
CONTAMINATED HAZARDOUS WASTE
Conly L Hansen
Nutrition and Food Science & Biological and Irrigation Engineering
Utah State University
Logan, UT 84322-8700
USA
Tel: (801) 750-2188
David K. Stevens
Civil and Environmental Engineering
Utah State University
Logan, UT,84322-4110
USA
Tel: (801) 750-3229
INTRODUCTION
Mercury pollution is a serious problem for humans and other animals. Mercury can be removed from
water by chemical precipitation; for example, a method for removing ionic mercury from water is addition
of sodium sulfide. The insoluble mercuric sulfide (HgS) that forms is removed as a sludge that is then
landfilled(l).
Chemical methods of mercury detoxification are not adequate. It has become evident that mercury can
be solubilized from HgS under conditions that could be present in a landfill; Thiobacillus ferrooxidan&can
facilitate solubilization and volatilization of Hg° from HgS (2,3):
Biological detoxification, using mercury-resistant bacteria in a completely mixed, aerobic biological
treatment process has been shown to have a capability for long-term removal of mercury from polluted water
or soil slurry (4, 5). Detoxification is an enzyme- catalyzed process in which ionic mercury is reduced to
volatile metallic mercury by mercuric reductase, a soluble flavoprotein located in the cytoplasm (6-8).
Mercuric reductase can convert both Hg* and Hg+* to Hg° (9). The elemental form is easily removed from
the growth medium.'thus facilitating the removal of mercury from polluted water (10). A schematic of the
process for removal of Hg2+ from aqueous solutions is shown in Figure 1.
TRAP 2 TRAP 1 SPLASH TRAP
EFFLUENT
Figure 1. Schematic of biological process for removing Hg2* from aqueous solutions
121
-------
METHODS
Experiments to better characterize the process were carried out at Utah State University between
September 1991 and May 1992. The growth substrate was a defined buffered minimal salts media
supplemented with sucrose. It consisted of: 3.0 g potassium phosphate monobasic (KH2PO4); 7.0 g
potassium phosphate dibasic (K,,HPO4-3H2O); 1.0 g ammonium sulfate ([NHJ2SQ4); 2.0 g sucrose
(C^HjjO,.) and 1.0 L distilled water. Aeration rates were held constant at 1.5 LPM to simplify analysis of
the fate of the reduced mercury.
Three influent mercury concentrations (nominal levels of 2,20, and 40 mg/L), 3 mean cell residence
times (12,20, and 28 hr), and 3 temperatures (15,22.5, and 30 degrees C) were employed in a 23 factorial
design with a replicated center point, at a nominal influent chemical oxygen demand (COD) of 2,500 mg/
L. The experimental design is given in Table 1.
To start the continuous runs; a batch culture was inoculated with 5% (v/v) stored culture and the
defined growth medium and mercury stock solution were added to the reactor on a 1:1 volumetric basis to
obtain the desired influent concentration. The reactor was allowed to operate and was monitored
periodically for influent and effluent Hg and COD, effluent Volatile Suspended Solids (VSS), bacterial
Colony Forming Units (CPU), Dissolved Oxygen (DO), pH, redox potential, and Hg volatilization for a time
period equivalent to about four Hydraulic Residence Times (HRT) (mean cell residence times) to achieve
steady-state conditions.
After about four residence times, the reactor reached steady state, as indicated by constant
concentrations of reactor COD, VSS, [Hg], CPU, DO, pH, and redox potential (data not shown). After the
reactor reached steady state, sampling was performed at least once each residence time for these
parameters to characterize the reactor under the run conditions. At least two samples were used for each
determination. At the end of a run, the culture was stored at 4 degrees C, the fermenter was cleaned, feed
flow rate was modified, and the culture was allowed to reach steady state at the new levels. Parameters were
averaged over the steady-state period.
Samples were preserved in a .04% sulfuric acid solution and stored at 4°C until analysis. COD was
determined using the Closed Reflux, Colorimetric Method (11). VSS was determined using the APHA
method, "Fixed and Volatile Solids Ignited at 550°C" (11). The 50 ml samples were collected from the reactor
using a grab technique.
Influent, effluent, and off-gas total mercury was determined by the Cold-Vapor Atomic Absorption
Spectrometric Method (11). A BuckScientific Mercury Analyzer System, Model 400A was used foranalysis.
Five ml samples were placed in 5ml of 50% nitric acid and stored at 4 degrees C until analysis. Off-gases
were trapped in a solution consisting of 3% potassium dichromate and 14% sulfuric acid in distilled,
deionized water.
Additional measurements included reactor DO, pH, and redox potential to ensure the reactor was
aerobic and well buffered. DO was determined using an Orion Research Model 97-08 Oxygen Electrode
connected to an Orion Research Model 720 pH/ISE Digital Meter. The pH was determined using a Corning
Scientific Instruments Model 10 pH Meter. Redox potential was measured with a Corning Model PS-19 ORP
Meter.
RESULTS AND DISCUSSION
Averaged steady-state concentrations for selected measured parameters are shown in Table 1. A
number of observations from the data are discussed in following paragraphs. A factorial analysis of these
experiments was carried out to determine which of the variables studied had a significant impact on the
mercury removal performance of the system. The results of the analysis for influent mercury concentration,
temperature, and hydraulic retention time are given in Table 2. Results for the other parameters measured
are not presented here but are available in a full report from the authors.
122
-------
TABLE 1. EXPERIMENTAL DESIGN AND SELECTED RESULTS
Temperature
-1
-1
1
1
-1
-1
1
1
0
0
HRT
-1
1
-1
1
-1
1
-1
1
0
0
Hgln
-1
-1
-1
-1
1
1
1
1
-0.31579
-0.31579
[Hg]e
mg/L
0.028
0.010
0.053
0.230
18.632
17.876
15.358
4.213
1.087
1.970
Removal CPU Redox
Efficiency % #/mL potential
98j
99.
97.
88.
6.1
8.'
15.
88.
92.
86.
6% 195,000
5% 10,080,000
4% 710,000
5% 4,520,000
5% 440
r% 314
0% 290,000
1% 3,600,000
8% 22,460,000
9% 44,330,000
189.2
175.5
171.4
179.8
268.6
249.0
256.4
146.0
140.9
136.3
-1 : low level of indicated variable
0, -0.3 16 : medium level of indicated variable
1 : high level of indicated variable
TABLE
2. SIGNIFICANCE OF
Variable
ESTIMATED
MAIN EFFECTS AND INTERACTIONS
Significance of
Effect on
[Hg]e % Removal
Main effects
Temperature (T)
Hydraulic Residence Time (HRT)
Influent Hg ([Hg],)
1 - means effect of variable is not significant at the 95% confidence level
2 + means effect of variable is significant at the 95 % confidence level
The growth behavior of the microorganisms responsible for the reduction of Hg2+ to Hg° and the effect
of mercury concentration on that behavior were examined. During all runs of the experiment, the system
was operating very close to washout conditions; in all cases, conventional measures of process perfor-
mance showed that the system had failed. The COD removal was essentially zero and VSS, a commonly
used measure of active biomass in wastewater treatment, was near zero. Thus, conventional approaches
to evaluation of growth kinetics of the bacteria responsible for mercury removal were not used here.
Instead of using VSS as the measure of biological activity, CPU was much more indicative of mercury
removal performance. Figure 2 shows the relationship between mercury removal efficiency and CFUs.
From this plot, there is a clear threshold CFU value of about 106 /ml, below which there is little or no Hg
removal and above which Hg removal is nearly complete. This behavior has been observed previously in
our laboratory and suggests that these mercury-resistant organisms are able to detoxify their environment
123
-------
120%
100%-
8
o 80% H
w
I
60%-
40%-
S 20%-
S
Threshold Region
Below Threshold
Above Threshold
0%T 1—i i i i mi 1—i I i mil 1—i i i i nil 1— 1—I i i mil 1—i I I i
100 10,000 1,000,000 100,000,000
1.000 100,000 10,000,000
Colony Forming Units (#/mL)
Figure 2. Mercury removal efficiency vs. colony forming units
by removing mercury at very low biomass densities, and thus if-their numbers are above the approximate
threshold of 108 units/ml,, good mercury removal efficiency should be possible.
At the higher concentrations, some of the influent mercury precipitated, probably as mercuric
hydroxide, mercuric oxide, or mercuric chloride (12) to a level'of soluble reactor [Hg], of 15 to 18 mg/L. The
solubility of these minerals is very low - in pure water at pH 7, the solubilities of Hg(CI)2, Hg(OH)2, and HgO
are approximately 10'10 -1O'12 molar Hg2+ at 25°C. The high ionic strength (>0.12 M) in the reactor medium
increases the solubility, as does the formation of Hg complexes with chloride, hydroxide, sulfate, and
ammonium to the levels seen here. In equilibrium with HgO precipitate, the concentration of the complex
Hg(OH)2 will be about 10"1 M or 20 mg/L Hg and may dominate the mercury in solution in the reactor (12).
This value is consistent with the concentrations observed in the reactor during runs 5, 6, and 7 as shown
in Table 1. Thus, in wastewaters it is unlikely to expect reactor soluble mercury concentrations in excess
of 20 to 30 mg/L
The runs for which CPU are below the threshold correspond to those runs for which the nominal influent
[Hg2*] was 40 mg/L. This suggests the possibility of inhibition or toxicity of mercury to these organisms at
elevated concentrations.
ACKNOWLEDGMENTS
The research described in this article has been funded in part by the U.S. Environmental Protection
Agency under cooperative agreement CR-814831-01-0 and by the Utah Agricultural Experiment Station,
Utah State University. It has not been subjected to the EPA's peer and administrative review and therefore,
may not necessarily reflect the views of the Agency, and no official endorsement should be inferred.
REFERENCES
1. Bhattacharyya, D., Jumawan A.B. Jr, and Grieves, R.B. Separation of toxic heavy rnetals by
sulfide precipitation. Separation Sci. and Tech. 14(5):441-45, 1979.
2. Silver, M. and Torma, A.E. Oxidation of metal sulfides by Thiobacillus ferroxidans grown on
different substrates. Can. J. of Microbiol. 20:141 -147,1974.
3. Baldi, F. and Olson, G.J. Effect of cinnabar pyrite oxidation by Thiobacillus ferrooxidans and
cinnabar mobilization by a mercury-resistant strain. Appl. Environ. Microbiol. 53:772- 776, 1987.
124
-------
4. Hansen, C.L., Stevens, O.K., Zhang, S. and Warner, D.W. Biologically Enhanced Removal of
Mercury From Contaminated Soil. Accepted for publication in Proceedings of the 85th Annual
Meeting of Air & Waste Management Association, Pittsburgh, PA 15230,1992.
5. Hansen, C.L., Zwolinski, G, Martin, D. and Williams, J.W. Bacterial Removal of Mercury from
Sewage. Biotechnology and Bioengineerinq Journal 26:1330-1334,1984.
6. Summers, A.O. and Sugarman, L.I. Cell-free mercury(ll)- reducing activity of a plasmid-bearing
strain of Escherichia coll J. Bacteriol. 119:242-249,1974.
7. Summers, A.O., Schottel, J., Clark, D. and Silver, S. Plasmid- borne Hg(ll) and organomercurial
resistance. Jri: Microbiology, D. Schlessinger, Ed. American Society of Microbiology. Washington,
DC, 1974. pp. 219-226.
8. Silver, S., and Kinscherf, T.G. Genetic and biochemical basis for microbial transformations and
detoxification of mercury and mercurial compounds, in: A.M. Chakrabarty (ed.), Biodegradation
and detoxification of environmental pollutants. CRC Press, Inc., Boca Raton, Fla., 1982. p. 85.
9. Izaki, K. Enzymatic reduction of mercurous and mercuric ions in Bacillus cereus. Can.J.Microbiol.
27:192-197,1981.
10. Summers, A.O., and Silver, S. Mercury resistance in plasmid- bearing strain of Escherichia coll.
J. Bacteriol. 112:1228-1236. 1972.
11. APHA-AWWA-WPCFStandardMethodsfortheExaminationofWaterandWastewater. 15thed.
American Public Health Association, 1015 Fifteenth Street NW, Washington, DC, 1989.
12. Lindsay, W.L. Chemical Equilibria in Soils. John Wiley and Sons. New York, New York, 1979.449
PP-
*U.S. GOVERNMENT PRINTING OFFICE:! 992-750-002/60102
125
-------
------- |