United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-95/012
April1995
&EPA
21st Annual RREL
Research Symposium
Abstract Proceedings
-------
ill
iiJiiiiiiiniiH m»\ (iii iiiiiiiniiiiiuiiiy
i ii in 111 in iiiiiiii mi in mi in in in i iiiiiiiiiiiiiiiiH
Ii illlll! iliiii.ii IK'Ii' :;"!K :i
!,!!!'! IN j Hi: L.' I 4 li'l.l liliHIUiill InlliluiliKII ly
,{• ." IilliKI i iiSiilWiftMl ( ..... «!; 'i •'.'••, ii'tli" 'SI , l ( fir '.I ..... • H "" -1: INI • ili'jSIn 'SiWi^HJ ]• I . f y 'Wr
it1 liiil 1 1 ...... l^fllWi.'jt!1'.1 l! tilJu.'tl! 'Ti;i ...... .' .'|t.'Vltttu.»ty !,.»i..«!
'" liliil "I'll Pl'illlt I" II ui'1!, II I1 "' 1 "ill » i'i ' illili'l ' '! II! ' ' ' I ' ;l|i|' A1 1/ I ' 'I/ ' W Jill1 1 11 III II 'I ' II ' Iii" IJ ' f 'II I fll :' II I ' il| ' III, ' '! ' I |i f!
'Ill- !•. ...... iiiiihdilrtiJ!" .",',. iifc ..... iikiiJiii.i' ,;li. ii ,- is ..... i.1.; ..... *!!"UI>,i!-H.i:iir'l:!! [[[
' illlllK I ilSij111:1;!! rill, ivii ::',,'',, li'lli ir 'IlilVil.,'!:,! ::il! '!; !' • ' ':•>! „ ..... ',!!!!» Ipi "if If; I (I'll; |il. Ill ' il"i,( i I'HittW^^ ill ....... I1W ill '',• i ISTJtM'i'lT''i1'!, " Bli'Si
I illlli1: 'iSd It ..... S ;!,is ', • • ' . ii!!!!!1' I1 tl", i ti.P 'It ', It- ' '.ill < ' liillM •('• ** ...... t:l < ,i'!f:"M
'iiiiiiTi i' „ i riii i " 11, if' i;11::' ;sii' 'i i,:!si i s i'i,: 1111 iM i11 "i i11; I i I'lrii,;"'' i1 r i''''' ,,ir! i" 11: iiiiHia:!11"!!11! ii™ i! y IIM^ SLiisaiiiiiw^^^^^^^^^^^ t iio fHiri .nil list! 'in
> •£ i 'ii'liiiil'i'i ,vi* '>• '.' ,!': i * • HA f-A v il, 11";'' ii t: ii'iini ii 'iiiiiiii.!!!!: ii'i' v' < ;>" i;iiwWfMfSiBifiif'iBiii liliii^^^^^^^^^^^^^^
1 f f , ,111 Illlll , l.l,!:;!,'1 311" I', n1' .IK I1'1' ',, II1! ill""'!!1'" ,',111 'i'"1 , w"1 III'I I'll]'111' I11' I"! |'I I i1
IKK ! ' ill!1!1,!1"!*!!!1 V
II II II I I I I II
IIIIIIII I I
III 111
111!,
11" it
IJlii!
lililf'i
ill II
I I II IIIIIIII I
I II II
111 111 III
-------
Margins for all trext pages--
same as originals.
EPA/600/R-95/012
April 1995
21st ANNUAL RREL RESEARCH SYMPOSIUM
ABSTRACT PROCEEDINGS
Coordinated by:
Science Applications International Corporation
Lower Gwynedd, PA 19002
Contract No. 68-C2-0148
Work Assignment No. 3-7
Work Assignment Manager:
Emma Lou George
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
Printed on Recycled Paper
-------
NbTICE
"" ,, ii jji , . ., :," • ' ;;; : . ij- ,;, i Hi , I , j% : ,; ;:' j] • : \ , / .
The abstracts presented in these Proceedings (with the exception of those from the Hazardous
Substance Research Centers and National Institute of Environmental Health Sciences) have been reviewed
in accordance with the U.S. Environmental Protection Agency's peer and administrative review policies and
approved for presentation and publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii:!* '«! ilifii!
,'! !l
"ffv . ': i 11I11I|I»I| !'t II ! liiiilll I
-------
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, pollution prevention, solid and hazardous waste, and Superfund-related activities. This
publication is one of the products of that research and provides a vital communication link between
researchers and users.
These Abstract Proceedings from the 1995 Symposium provide the results of projects recently
completed by RREL and current information on projects presently underway. Those wishing additional
information on these projects are urged to contact the author or the EPA Project Officer.
RREL sponsors a symposium each year in order to assure that the results of its research efforts
are rapidly transmitted to the user community.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-------
ABSTRACT
The 21st Annual Risk Reduction Engineering Laboratory (RREL) Research Symposium was held in
Cincinnati, Ohio, April 4-6,1995. The purpose of this Symposium was to present the latest significant research
findings from ongoing and recently completed projects funded by the U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory.
These Proceedings are organized into two sections. Part One contains extended abstracts of the paper
presentations. Part Two contains abstracts of the poster displays. Subjects include pollution prevention
demonstrations and life cycle analysis; remediation technologies from the SITE Program, RREL technologies,
and oil spills remediation technologies; drinking water and wastewater technologies; municipal solid waste
technologies; and hazardous waste technologies.
IV
-------
CONTENTS
PART ONE - PAPER PRESENTATIONS
A Practical Approach to Choosing a Substitute Solvent
Garry Howell, U.S. EPA, RREL 1
Onsite Recycling of Electric Arc Furnace Dust: The Jorgensen Steel Facility
Ivars Licis, U.S. EPA, RREL 5
A Replacement Solvent Cleaner/Degreaser Study at Duffy Electric and Machine Company
Johnny Springer, U.S. EPA, RREL .. 9
Evaluation of Supercritical Carbon Dioxide Spray Technology to Reduce Solvents in a Wood
Finishing Process
Paul Randall, U.S. EPA, RREL ... 14
Demonstration of Alternatives for Vapor Degreasers
Dean Menke, University of Tennessee 19
Research Support to the Pollution Prevention Program of the U. S. Postal Service
Terri Hoagland, U.S. EPA, RREL 24
Life Cycle Design Framework and Demonstration Projects
Gregory Keoleian, University of Michigan 27
Industrial Assessment Database for Energy Efficiency and P2
Peter Polomski, Rutgers University 29
Applications of Supercritical Fluid Technology to Pollution Prevention and Waste Minimization
Ronald Turner, U.S. EPA, RREL 35
Development of High-Rate Trickle Bed Biofilter
Richard Brenner, U.S. EPA, RREL 38
Integrated System for Treating Soil Contaminated with Wood Treating Wastes
Carolyn Acheson, U.S. EPA, RREL 45
-------
Development of a Multi-Level Protocol for Determination of Biodegradation Kinetics and Bioavailability
of Organic Compounds to Enhance In-Situ Bioremediation
Henry Tabak, U.S. EPA, RREL ; 49
Design and Testing of an Experimental In-Vessel Composting System
Carl Potter, U.S. EPA, RREL 56
Evaluation of Abiotic Fate Mechanisms in Soil Slurry Bioreactor Treatment
John Glaser, U.S. EPA, RREL 61
Biotransfomnation of Sulfur and Nitrogen Oxides in Stack Gases
Rakesh Govind, University of Cincinnati 64
Anaerobic Biodegradation of Pesticides and TNT: SITE Emerging Technology Demonstration
Wendy Davis-Hoover, U.S. EPA, RREL 71
Iron Enhanced Abiotic Degradation of Chlorinated Hydrocarbons
Chien Chen, U.S. EPA, RREL .74
Testing the Sonotech Pulse Combustor in a Pilot-Scale Rotary Kiln Incinerator
Marta Richards, U.S. EPA, RREL 79
The Relative Effectiveness of Mineral-Based Sorbents for Metal Capture in a Bench-Scale Incinerator
Gregory Carroll, U.S. EPA, RREL .84
Research on the Low-Temperature Decontamination of Soil
John Rindt, University of North Dakota ..'..- 89
A Summary of Two Recycling Technology Evaluations Completed Under the MITE Program
Lynnann Hitchens, U.S. EPA, RREL .-. , 94
Full-Scale Leachate-Recirculating MSW Landfill Bioreactor Assessments
David Carson, U.S. EPA, RREL .98
Removal of Mercury from Stack Gases by Activated Carbon
Radisav Vidic, University of Pittsburgh 103
Application of Ambersorb 563 Adsorbent Technology for Treatment of Chlorinated Organics in
Groundwater
Russell Frye, Roy F. Weston, Inc 108
vi
-------
Geosafe In Situ Vitrification SITE Demonstration
James Hansen, Geosafe Corporation 113
Unterdruck-Verdampfer Brunnen (UVB): An In Situ System for Remediation of Contaminated Aquifers
Michelle Simon, U.S. EPA, RREL . . . 120
SITE Demonstration of Terra-Kleen Response Group's Mobile Soyent Extraction Process
Mark Meckes, U.S. EPA, RREL .125
Electrokinetic Remediation: Technology Status
Yalcin Acar, Louisiana State University 129
SITE Demonstration of Bioremediation of Cyanide at the Summitville Colorado Site
Leslie Thompson, Pintail Systems, Inc. . -.".'.- 134
Reductive Photo-Dechlorination (RPD) Process for Safe Conversion of Hazardous Chlorocarbon Waste
Streams
Moshe Lavid, ENERGIA, Inc 140
Treatment of Hazardous Landfill Leachate by the Rochem DTM Membrane Separation Technology
Douglas Grosse, U.S. EPA, RREL 145
SITE Demonstration of the SVVS Technology
Paul dePercin, U.S. EPA, RREL . 149
Use of Sulfate-Reducing Bacteria in Acid Mine Drainage Treatment
Thomas Powers, U.S. EPA, RREL .153
Buffing, Burnishing, and Stripping of Vinyl Asbestos Floor Tile
Bruce Hollett, U.S. EPA, RREL 159
Contaminants and Remedial Options at Pesticides Sites - A Technical Resource Document
Richard Koustas, U.S. EPA, RREL 162
Contaminants and Remedial Options at Selected Metals Contaminated Sites - A Technical Resource
Document
Michael Royer, U.S. EPA, RREL 166
VII
-------
Overview of Technical Resource Document for Solvent Contaminants and Remedial Options at
Superfund Sites
Uwe Frank, U.S. EPA, RREL 172
SITE Demonstration of the Dynaphore/Forager Sponge Technology to Remove Dissolved Metals from
Contaminated Groundwater
Carolyn Esposito, U.S. EPA, RREL 177
Case Studies of the High Voltage Electron Beam Technology for Wastewater Treatment
Mary Stinson, U.S. EPA, RREL
181
SITE Demonstration of the ZenoGem™ Technology to Treat High Strength Wastewaters
Daniel Sullivan, U.S. EPA, RREL
186
Recovery of Monomers from Recycled Plastics
Laura Sharp, University of North Dakota
189
Two-Phase Ozonation of Chlorinated Organics
Dibakar Bhattacharyya, University of Kentucky
194
Evaluation of Base-Catalyzed Decomposition (BCD) Process for New York/New Jersey Harbor
Sediment Decontamination
Dennis Timberlake, U.S. EPA, RREL 199
Electrokinetic Soil Remediation: Impact of Aqueous Phase Properties on Soil Surface Charge and
Electroosmotic Efficiency
Leland Vane, U.S. EPA, RREL 203
Base Catalyzed Decomposition - New Hydrogen Donors and Reaction Activators/Catalysts
Fred Kawahara, U.S. EPA, RREL
207
Swirl Technology - Proper Design, Application, and Evaluation
Richard Field, U.S. EPA, RREL
211
A Mufti-Chambered Stormwater Treatment Train for the Treatment of Stormwater
Robert Pitt, University of Alabama at Birmingham
217
Groundwater Contamination from Stormwater Infiltration
Tom O'Connor, U.S. EPA, RREL (Storm & Combined Sewer)
222
Vill
-------
Colloidal Contaminants in Urban Runoff
Mark Wiesner, Rice University . • • • • 227
Modelling Air Emissions from Contaminated Dredged Materials
Louis Thibodeaux, Louisiana State University . 232
Bioremediation of Contaminated Sediments
Joseph Hughes, Rice University • • • • 236
Pollutant Fluxes to Aquatic Systems via Bed-Sediment Processes
K.J. Valsaraj, Louisiana State University . • 243
Accelerating Research and Development of Innovative Technologies
Norma Lewis and Randy Parker, U.S. EPA, RREL ... 247
Supercritical Extraction of Polychlorinated Biphenyls from Soils and Sediments: Remediation and
Possible Risks
Lawrence Tavlarides, Syracuse University , 248
By-Products from Supercritical Water Oxidation: Pathways, Kinetics and Mechanisms
Phillip Savage, University of Michigan 254
Oxidation and Hydrolysis of Acetic Acid and Methylene Chloride in Supercritical Water as a Means of
Remediation
Philip Marrone, Massachusetts Institute of Technology 258
Soil Lead Remediation: Is Removal the Only Option?
James Ryan, U.S. EPA, RREL , - - 260
Review of Mathematical Modeling for Evaluation of SVE Applications
James Mercer, GeoTrans, Inc • 265
Representative Sampling and Analysis of Heterogeneous Soils
Ann Leitzinger, U.S. EPA, RREL -272
Bioremediation of Trinitrotoluene by a Ruminal Microorganism
Kenneth Williamson, Oregon State University 278
Full-Scale In-Situ Bioremediation of Trichjoroethylene in Groundwater: Preliminary Modeling Studies
Mark Goltz, Stanford University 283
ix
-------
A Microbiological Surrogate for Evaluating Treatment Efficiency
Eugene Rice, U.S. EPA, RREL 288
Evaluation of Membrane Performance and Fouling by Pyrolysis- GC/MS
Thomas Speth, U.S. EPA, RREL 293
Biofiltration for Control of Volatile Organic Compounds (VOCS)
Dolloff Bishop, U.S. EPA, RREL 298
Potential Cryptosporidium Surrogates and Evaluation of Compressible Oocysts
Sylvana Li, U.S. EPA, RREL 305
Remote Monitoring and Control of Small Drinking Water Package Plant Systems
John Teuschler, U.S. EPA, RREL 309
Control of Copper Corrosion of Household Plumbing Materials
Michael Schock, U.S. EPA, RREL 312
Hydraulic Fractures As Subsurface Electrodes: Early Work on the Lasagna Process
Larry Murdoch, University of Cincinnati 318
Hydraulic Fractures as Anaerobic Biological Treatment Zones
Wendy Davis-Hoover, U.S. EPA, RREL 323
Advances in TCE-Degrading Thermophilic, Mesophilic, and Alkaline-Resistant Organisms
Steven Vesper, University of Cincinnati 325
Integrated Pneumatic Fracturing and Bioremediation for the In-Situ Treatment of Contaminated Soil
David Kosson, Rutgers University 329
Fundamental Criteria for the Better Design of Secondary Combustion Chambers
, Guido Sacchi, MIT 333
Kinetics of Metal Leaching from Heavy Metal Contaminated Soils
John Van Benschoten, SUNY - Buffalo 338
Intrinsic Bioremediation of Chlorinated Solvents at the St. Joseph, Ml Aquifer-Lake Michigan Interface
Peter Adriaens, University of Michigan 344
-------
Monitoring and Control of Pseudomonas sP. Strain KC for In-Situ Bioaugmentation of Carbon
Tetrachloride Contaminated Sites
Craig Criddle, Michigan State University 349
Remediating Munitions Contaminated Soils
Pat Shea, University of Nebraska 354
Risk Reduction Through Waste Minimizing Process Synthesis
L.T. Fan, Kansas State University 359
Prepared Bed Reactor for Full Scale Remediation of Soil Contaminated with Wood Preserving Wastes:
Field Bioremediation Evaluation
Ronald Sims, Utah State University . . ., . • .-. • ,, 364
PART TWO - POSTER DISPLAYS
The EPANET Water Quality Model
Lewis Rossman, U.S. EPA, RREL ,.. -,- 368
Removal of Alkanes from Drinking Water Using Membrane Technologies
Carol Fronk, U.S. EPA, RREL . . : , • • •
An Evaluation of Drinking Water Samples Treated with Alternative Disinfectants
Kathleen Patterson, U.S. EPA, RREL . . . ...;... 370
A Measurement Method for Pollution Prevention Progress
Dave Stephan, U.S. EPA, RREL •,. ,371
Pollution Prevention Possibilities for Small and Medium-Sized Industries: The Waste Reduction
Innovative Technologies Evaluation (WRITE) Project Summary
Ivars Licis, U.S. EPA, RREL . . 372
Pollution Prevention Research and Outreach
Ruth Corn, U.S. EPA, RREL , 373
Computer Supported Information System for Measuring Pollution Prevention Progress
Rada Olbina, U.S. EPA, RREL ...... 374
XI
-------
QA of Fortune
Kim McClellan, U.S. EPA, RREL 375
Treatability Data Base, Version 5.0
Glenn Shaul, U.S. EPA, RREL 376
Mine Waste Technology Program
Thomas Powers, U.S. EPA, RREL 377
Comparison of Reactor Designs for Biological Ex-Situ Soil Treatment
Carl Potter, U.S. EPA, RREL 378
Integrated Pneumatic Fracturing/Bioremediation for the In-Situ Treatment of Contaminated Soil
Uwe Frank, U.S. EPA, RREL 379
Results of a SITE Demonstration of the COGNIS-TERRAMET® Lead Extraction Process to Remove
Heavy Metals from Soil
Michael Royer, U.S. EPA, RREL 380
RREL Site Remediation Technical Support Program
Benjamin Blaney, U.S. EPA, RREL 381
Location of Leaks in Pipelines Using Acoustical Principles
Anthony Tafuri and Robert Hillger, U.S. EPA, RREL 382
Alternative Treatment Technology Information Center Computer Database System
Daniel Sullivan, U.S. EPA, RREL 383
Clean Technology Demonstrations for the 33/50 Chemicals
Diana Kirk, U.S. EPA, RREL 384
Pollutant Element Foms within Sludges Generated by Treatment of Two Acid Mine Waters with Lime,
Inorganic Sulfide and Sulfate Reducing Bacteria
Bill Chatham, Montana Tech 385
XII
-------
ACKNOWLEDGMENTS
The 21st Annual RREL Research Symposium was planned in partial fulfillment of Contract No. 68-
C2-0148, Work Assignment No. 3-7 by Science Applications International Corporation (SAIC) under
sponsorship of the U.S. Environmental Protection Agency. Emma Lou George of the Risk Reduction
Engineering Laboratory (RREL) was the Work Assignment Manager responsible for coordinating this
project. The conference program and activities were planned by a committee consisting of the following
individuals: Emma Lou George, Daniel Sullivan, Lou Garcia, Franklin Alvarez, Randy Parker, and Walter
Feige of RREL and Lisa Kulujian of SAIC.
xiii
-------
-------
A PRACTICAL APPROACH TO CHOOSING A SUBSTITUTE SOLVENT
S. Garry Howell and Johnny Springer
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
E. Thomas Marquis
Austin Laboratories Huntsman Corporation
7114N. LamarBIvd.
Austin, TX 78752
INTRODUCTION
Organic solvents have a multitude of uses in industrial processes, and, although there is some
progress being made in replacing them with aqueous systems, many applications still require them. Many
cleaning operations which require merely removal of light soils and easily emulsified oil or grease have
been changed to aqueous systems with excellent results. Other applications, such as paint removal,
cleaning of easily corrodible metal surfaces, removing polymer coatings, or baked on carbon, still may
require application of an aggressive solvent. This paper will focus on coating removal, or paint stripping,
and consider several instances where substitutes were chosen, and attempt to explain why (or why not)
these substitutes performed well.
A major use of organic solvents is in the formulation of solvent coatings such as paints, enamels,
varnishes, and lacquers. This is an extremely broad and specialized field, and will not be covered in this
brief discussion. Instead, we will discuss the other end of the spectrum, that is, removing coatings which
have reached the end of their useful life. The techniques for choosing solvents for coating removers can
follow almost the same rules used to formulate them, except that while the resins used to formulate a
coating are soluble in the proper solvent mixes, many coatings cure or crosslink after application and
become insoluble, and merely swell or soften. Indeed, complete dissolution of the coating to be removed
is undesirable, since a softened coating may be removed in large pieces, instead of as a viscous solution.
A brief outline of progress in solvent selection indicates that the earliest discovery that solvent
mixtures were more effective than pure solvents was made by paint and varnish formulators. As knowledge
advanced, the Edisonian approach of trial and error was replaced by reasoning; if one had a clue about
the chemical structure of both the resins being dissolved, and of a group of available solvents, then trial
and error could be reduced. Others postulated that physico-chemical parameters of both the solvents and
resins or other solutes could be measured and described mathematically using solution theory. Modern
solution theory relies heavily on work done 40 or more years ago by Hildebrand1 and expanded upon by
Hansen2. Hansen proposed that there were three contributors to the solubility parameter as developed by
Hildebrand which he designated Sd, Sp, and Sh, each contributing in its own way to the overall solubility of
a compound, and that a close numerical match (about 1 or 2 Hildebrand(H)) is required for mutual solubility
of two compounds. Sd quantifies the dispersion or van der Waals intermolecular forces; these are the
primary forces at work in hydrocarbons. Sp is a polar component common to acids, alcohols, etc. Sh is
the hydrogen bonding parameter which designates the bond between hydrogen and two unshared electrons
of another molecule. There are several tables of these values available, notably the Solvent Properties
chart listing over 200 solvents prepared by Dr. Tom Marquis and published by Texaco Chemical Company,
(now Huntsman Corporation). Huntsman Corporation has also developed a computer program which will
calculate the solubility parameter of mixtures of solvents. Others have both tables and computer solvent
selection programs available today3, although some programs concentrate more on such factors as
evaporation rate, smog formation, and flash points.
-------
RREL and the U.S. Air Force recently funded an extensive assessment of pollution prevention
opportunities at Tinker Air Force Base (TAFB), Oklahoma. Among the processes evaluated by the
contractor, Battelle, were a number of depainting and degreasing operations where either toxic, flammable,
or ozone depleting solvents were used. One of these was a depainting operation where aircraft radomes
were depainted by showering them with methyl ethyl ketone (MEK).
The radomes to be depainted at TAFB had three coats of paint; an epoxy primer, a urethane rain
erosion layer, then a low visibility topcoat. When time came to overhaul an aircraft, the radomes were
removed, the coatings were scored, and the radome placed under a shower of MEK until the coating
bubbled and could be squegeed off; the operation typically takes 1 1/2 to 3 hours. Because of the
flammability, volatility, and toxicity of MEK, the shower was enclosed in a large closed booth with forced
ventilation. During the process, MEK vapors were discharged to the atmosphere, and about 50,000 pounds
were lost in 1992. One of the coauthors, Mr. Johnny Springer, was Project Officer on a task to determine
if mixture of solvents would perform almost as well as MEK, be less volatile, have a higher flash point, and
be less toxic. After looking through a table of solubility parameters, it appeared that a low volatility, non
toxic solvent such as propylene carbonate (PC) blended with others to match the S0, might make a
candidate replacement solvent. Dr. Marquis agreed to try to develop a substitute mixture having the
requisite characteristics, as he thought a single solvent would not be effective. Development of these
formulations and some guidance on developing new ones are the subject of this paper.
METHODOLOGY
Using the Hildebrand solubility parameter of MEK as a rough starting point, Dr. Marquis made up
several blends of solvents for laboratory testing. All of the blends contained propylene carbonate and
N-methyl pyrrolidone (NMP) and at least one other solvent. The combinations were not chosen on their
removal efficiency alone; vapor pressure (a large contributor to evaporation rate), flash point, and toxicity
were also considered. Small sections of a condemned B-52 radome were immersed in the solvent blends
in open beakers and compared to MEK, however it was noted that the volatile MEK evaporated rapidly,
which was overcome by enclosing those beakers in polyethylene bags. Table 1 lists the time required for
the paint to completely Bubble so that it could be easily wiped off. MEK took up to 12 times as long to
bubble the coatings in this test as the mixtures did. This did not compare to the experience of the
operators of the depainting operation at TAFB; their experience with the shower head arrangement
indicated that complete removal of the paint usually occurred in less than three hours.
TABLE 1
Solvent
MEK
Dry
Blend 1
Blend 2
Blend 3
sd
7.8
8.9
8.9
8.5
SP
4.4
5.6
5.7
3.8
sh
2.5
3.5
3.5
4.2
Strip Time
Hours
24
4
2
4
Composition by Weight
(%)
PC NMP DBE
33.33% each
25 50 25
15 15 40
DPM 15 MIAK 15
MEK - methyl ethyl ketone; PC - propylene carbonate; NMP - N-methyl-2-pyrrolidone;DBE - Dibasic esters
(Du Pont) (mixture of dimethyl esters of succinic, glutaric, and adipic acids);DPM - Dipropylene glycol
monomethyl ether;MIAK - Methyl isoamyl ketone
-------
Looking at the solubility parameters of the blends indicates that they were all fairly similar to those
of MEK, and therefore should not differ too greatly in solvent activity. Questioning of the engineer in charge
of the TAFB operation revealed that the removal time with MEK decreased after the first run or so, and that
he suspected water absorbtion from the air could be the reason. From this information, the contract
laboratory reasoned that enclosing the beakers was a mistake, and found that open beakers of MEK
performed better. They then found that adding 12.5% of water to the MEK yielded bubbling times of 30
minutes. This contradiction of conventional wisdom was not apparent at first, since one would think that
water would detract from, rather than augment, removal efficiency. An explanation of this was sought by
reevaluation of the MEK solubility parameters, especially Sh. Looking again at Table 1, Sh of MEK is 2.5,
only 1 Hildebrand (H) lower than Blend 2, the best of the three tested. In general a match within 1 H is
considered adequate for solvent matching, but knowing the effect of water, the strongest hydrogen bonder,
on the process, the effect on this parameter was calculated.
Solvent blend 2 in Table 1 is clearly the best paint stripper of the three blended solvents. Its
hydrogen bonding parameter is 3.5 compared to 2.5 for MEK, and Sp at 5.7 is clearly higher than MEKs
4.4 H. It is then quite likely that an important part of this blends effectiveness is the hydrogen bonding
ability of one or more of its components. NMP is hygroscopic, that is, it readily absorbs water out of the
air. Blend 2, having 50% NMP, quite likely absorbed water during the time it was in the beaker with the
radome section. The calculated Sh of both MEK and NMP are increasing as water is added, and at 12.5%
water, MEKs Sh is about 4.3H; Blend 2 only requires 4% water reach 4.3H. This may partially explain why
MEK/water is an effective stripper for this particular coating.
Solvent blends 1 and 3 are apparently affected by Sp and a third factor, dielectric constant, closely
related to it. An attempt to incorporate this factor into an overall equation is underway, and will be the
subject of a subsequent paper.
What's happening? Why does Blend 2 work almost as well as MEK/water? Why doesn't plain water remove
the coating? While there is no trite answer to the questions, an effective remover/stripper must have the
following characteristics:
A. Ability to diffuse into the film. Only one of the components of a blend may do this alone,
but it may carry the others with it, as diffusion implies solubility, which means that polymer
chains are separated4 by solvent.
B. The proper solvent power to move coating molecules apart so that the film is softened
and/or swelled. In this, and in (A.) above, the solvent must have at least one parameter
which matches the coating film. For example, a styrene/butadiene rubber will hardly be
affected by MEK but will swell considerably in hydrocarbons such as toluene. This is easily
determined by experimentation, but also by looking at the solvent properties chart, and a
list of polymer solubility parameters5'6. In this case, should the solvent be too effective and
make the surface sticky, adding a poor or moderately hydrogen bonding solvent could
produce a good swelling or removal agent.
C. Adhesion of a coating to a substrate can occur in several ways. In the case of the radome
example above, it was apparently largely through hydrogen bonding, and increasing the
hydrogen bonding parameter of the solvent with the addition of water allowed water to
occupy sites at the polymer/substrate interface, weakening the polymer/substrate bond.
Conversely, using toluene to remove a hydrocarbon coating may work, but if not
crosslinked, the coating might dissolve completely, making a viscous solution almost as
difficult to remove as the coating. This might be overcome by adding another component
which is a moderate hydrogen bonder which would keep the resin from going completely
into solution.
D. When the coating polymer is polar, as polyurethanes are known to be, a solvent of high
3
-------
dielectric constant may aid in separating charged areas, thus making space for other
solvents to soften the film.
If a table of polymer solubility parameters includes the one you wish to strip, so much the better,
but if not, a practical technique for developing a stripper/ remover would be to find one that does the job,
however volatile, toxic, or odoriferous it may be, and try to find its solubility parameters in the literature.
From this point, one may develop a solvent blend which closely matches these parameters.
We are continuing work on finding solvent substitutes; our next task, which is to the reporting stage,
is finding a blend which eliminates NMP, as it has recently come under suspicion of being a mutagen.
Findings from the MEK/NMP/PC study above will help in this task. Our next planned project is to find a
substitute for the phenol/methylene chloride stripper used on complete aircraft. This might prove to be
much more difficult than stripping a relatively small, smooth surfaced radome. There are problems of
corrosion, volume of solvent, and overall cost, since active solvents are more expensive.
REFERENCES
1. Hildebrand, J.H., Scott R.L The Solubility of Non Electrolytes Reinhold, New York, 1950.
2. Hansen. H.J. J. Paint Technol. (1967) 40.197.
3. In addition to Texaco-Huntsman, Dow Chemical Co., Shell, Arco, BP, Eastman, and EXXON have
computer programs for choosing solvents.
4. Hildebrand reasoned that the energy required to separate two molecules could be measured by
the heat of vaporization. While this measurement is feasible for small molecules, polymers usually
decompose before they vaporize, so their solubility parameters are measured indirectly, by finding
solvents which will dissolve or soften them. The solubility parameters of the solvents are then used
to estimate that of the polymer.
5. Brandrup, J., Immergut E.H. Polymer Handbook Wiley Interscience, New York 1989 p. VIII 519.
6. Barton, A.F.M. CRC Handbook of Solubility and Other Cohesion Parameters CRC Press Boca
Raton, FL 1983.
ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Ms. Ann Hooper and colleagues of F.W. Envirosponse
and to Carlos Nazario of Tinker Air Force Base for their observations which led to this paper.
Thanks also Dr. William B. Jensen of the University of Cincinnati for much valuable insight on the aspects
of hydrogen bonding/acid-base interactions.
-------
ONSITE RECYCLING OF ELECTRIC ARC FURNACE DUST:
THE JORGENSEN STEEL FACILITY
Ivars J. Licis
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513)569-7718
Robert C, Bermark
Washington State Department of Ecology
PO Box 47600
Olympia, WA 98504-7600
(360) 407-6812
INTRODUCTION
The steel-making industry produces a large amount of Electric Arc Furnace (EAF) dust as part of
normal production. This waste is listed as K061, defined as "emission control dust/sludge from the primary
production of steel in electric arc furnaces" under 40 CFR 261.32. A glass making technology called Ek
Classification™ (hereafter called "the Process") has been developed by Roger B. Ek and Associates, Inc.
(hereafter called "the Developer") to recycle EAF dust and convert it, along with other byproducts of the
steel-making industry, into marketable commodities.
This Process was evaluated under the Waste Reduction Innovative Technology Evaluation (WRITE)
Program. The project was designed and conducted in .cooperation with the Washington State Department
of Environmental Quality, the Process Developer and the host test site, the Earle M. Jorgensen (EMJ) Steel
Company of Seattle, Washington. Test personnel for EPA were supplied by SAIC Inc., on contract to EPA.
The overall objectives of the project were to conduct a pilot scale evaluation of the Process,
investigate if toxic metals are leached from the products (such as colored glass and glass-ceramics;
ceramic glazes, colorants, and fillers; roofing granules and sand-blasting grit; and materials for Portland
cement production). Three glass recipes (Glass I, II, and III) were designed by the developer for potential
use at EMJ. The EPA portion was focused on determining the toxic metals concentrations of the Glass
II recipe, evaluating the P2 impact of using this Process in comparison to traditional methods of waste
treatment and disposal, and assessing the economics of both.
METHODOLOGY
For this project, a portable pilot-scale process furnace was utilized. The furnace was located in the
steel-melting area of the EMJ plant so that fugitive emissions could be collected with the steel plant's dust
collection system and routed to the baghouse. Natural gas burners were used for the initial melting of the
glass and to heat the furnace to its operating temperature of approximately 2,500° F. The furnace was also
equipped with molybdenum metal electrodes, which could be used for partial or complete electric heating
of the furnace. For this test, the electrodes were also tested to provide criteria for designing a full scale
system, including requirements for the electric power supply.
EMJ supplied samples of feed materials for the test. Chemical analyses were performed on the
feed materials to serve as a basis for mixing the three recipes. The feed materials used for the production
of the glass included EAF dust, spent steel slags, spent refractories, mill scale, and grinding swarf.
Approximate batch size was 300 Ibs.
The Developer's test program consisted of the sampling and analysis of all three glass recipes, and
process monitoring throughout the test.
-------
The Glass II recipe was sampled and analyzed as part of the EPA testing activities. This recipe was
used to prepare glaze, iron silicate for Portland cement production and sandblasting grit.
Glass was dispensed either into molds to form castable products or into a granulator (water quench)
to form granular products. Once ladling or pouring was complete, the furnace was refilled with glass batch
and the process was repeated.
Samples obtained were split for analysis by both EPA's laboratory (NET Pacific, Inc.) and the
Developer's laboratory (Sound Analytical Services, Inc.) and placed in 1-L glass jars. The primary granular,
and a composite of the castable samples were subjected to the TCLP in accordance with SW-846 Method
1311 and subsequently analyzed for eight RCRA medals plus zinc. The TCLP analyses performed on Glasses
I and III by the Developer's laboratory were reported to be within TCLP limits.
EXHAUST
- FWWACZ lttn»ACTO«V LINING
y NATURAL OAS
» tURMCM NOZZLE
CLASS IATCN
MOLYBDENUM
ELECTRODE
Figure 1. Ek Classification™ Test Furnace for EAF Dust
RESULTS
Results from analyzing leachability characteristics for are shown in Table 1. The leachable metals
concentrations, in both the castable and granular samples were within the TCLP limits for all compounds
analyzed. Barium, chromium, lead, and zinc were the only compounds detected in either of the EPA
samples. Comparison of these data to those obtained by.the Developer produced similar results (for the
granular product only).
-------
TABLE 1. TCLP RESULTS AND COMPARISON TO REGULATORY LIMITS
FOR SAMPLES FROM EMJ
EPA HW No,1
D004
D005
D006
D007
D008
D009
D010
D011
—
Contaminant
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Zinc
EPA
Castable
Sample2
(mg/L)
< 0.0025
0.043
< 0.0035
0.050
0.067
< 0.00008
6
<0.001
<0.0092
0.95
EPA
Granular
Sample2
(mg/L)
<0.0025
0.025
< 0.0035
0.13
0.120
< 0.00008
6
<0.001
< 0.0092
0.60
Developer
Granular
Sample
(mg/L)
<0.2
<0.1
<0.1
0.1
<0.1
< 0.002
<0.3
<0.1
0.6
Regulate
ry Level3
(mg/L)
5.0
100.0
1.0
5.0
, 5.0
0.2
1.0
5.0
NR
Hazardous Waste Number
2 Average of duplicate samples
3 Regulatory levels taken from 40 CFR ch. 1
NR Not Regulated
(7-1-90 Edition), Section 261.24, Table 1
The major P2 impact of implementing this Process is the double effect of precluding the treatment
and burial of the EAF dust in a hazardous landfill while negating the requirement of mining and processing
virgin material for the purpose of making identical products. For this limited evaluation, the pollution
involved in making glass products the traditional way, vs. recycling, the relative energy consumption,
resource depletion, etc. were not quantified.
The recycling of other, non-hazardous wastes into the Process, is an added benefit.
The economic estimates indicate that a profitable operation is possible. Industrial product prices
ranged from $2/ton (Portland cement materials) to $650/ton (glass ceramics/ architectural tiles feedstocks)
based on 1991 information. Assuming that the average price of products sold is in the middle of the price
range, and based on the Developer's projection of operating costs, in a ten-year period, the Process could
produce a gross profit of $63,195,000. Avoiding disposal costs (at around $200/ton) could save another
$43,040,000, for a total of $106 million. The capital costs for this facility would be $10,500,000 Additional
benefits could be reduced liability, and avoidance of administrative costs for permits and managing of
-------
hazardous waste under the old system. The actual savings realized will depend heavily on the types and
amounts of products sold.
Air emissions and process wastewater were not analyzed for this test. For full scale applications
these may need to be investigated under actual operating conditions.
Stack gas sampling data were previously gathered during earlier, pilot-scale tests at the Oregon
Steel Mill (OSM). Although these data suggest acceptable air emissions, the data do not satisfy EPA stack
testing protocols and standards. Prior to this project, on the basis of historical and glassmaking data from
other sources, the Developer was allowed to proceed in building a full-scale facility at OSM by the Oregon
Department of Environmental Quality.
The full report and project summary, titled "Recyclin g of Electric Arc Furnace Dust: Jorgensen Steel
Facility", by Trevor W. Jackson, et al., are being published as EPA reports.
CONCLUSIONS:
The glass product types which were prepared by the Process and tested as part of this study
resulted in relatively non-leachable glasses. The concentrations measured were lower than those
allowed under RCRA regulations for TCLP.
• The Process utilizes both listed and non-hazardous foundry wastes to replace constituents that
would otherwise be mined and processed from virgin sources as additives for glass-making. Ideally
this results in both a conservation of resources and reduction of hazardous and non-hazardous
wastes for disposal.
• This project did not focus on investigating compliance issues in terms of air emissions and waste
water generated during batch charging, melting, quenching and drying of the three glass products.
Although there is historical evidence for glassmaking, for a rigorous evaluation of species and
concentrations, this type of testing would need to be evaluated at full scale under actual operating
procedu res/conditions.
8
-------
A REPLACEMENT SOLVENT CLEANER/PEGREASER STUDY
AT DUFFY ELECTRIC AND MACHINE COMPANY
Johnny Springer, Jr.
Waste Minimization, Destruction
and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
(513) 569-7542
Bruce Sass
Battelle
505 King Avenue
Columbus, Ohio 43201-2693
(614)424-6315
INTRODUCTION
Duffy Electric & Machine Company repairs and rebuilds electric motors. The company
overhauls large electric motors (AC and DC with greater than 15 hp output). The company also
overhauls small electric motors. The process involves gross cleaning of electromechanical devices to
achieve a level of cleanliness that facilitates inspection, repair, and testing.
The cleaning system used in this study is comprised of a cleaning unit and a rinse unit. The
cleaning unit is the IBR Series 400 parts washer, made by Inter Basic Resources, Inc., of Grass Lake,
Michigan. The unit features an 11-inch-diameter impeller blade with 100-rpm rotation mounted at the
inside bottom of an immersion tank. Cleaning fluid is circulated through a 50-u.m filter at 5 gallons per
minute, at 1 foot of head pressure to create turbulence in the cleaning fluid. Parts are cleaned as the
fluid impinges on them. A 100-VAC electric motor powers the unit. The motor turns the impeller blade
and runs the pump that circulates the fluid. The immersion well is 16 in x 19 in x 11.5 in and it holds
approximately 14.5 gallons. Safety switches shut off the air motors if the loading door is open. The
cleaning tank contains an insertion heater. The ester bath is heated to approximately 130°F. Light
scrubbing is used to dislodge heavy grease and other difficult-to-remove contaminants.
The IBR alcohol rinser consists of a chamber that encloses a manifold sprayer. The tube-shape
manifold is fitted with an array of nozzles to spray isopropyl alcohol (IPA) onto the parts from a variety
of angles; the nozzles project a flat spray. Pneumatic pumps power the unit. The power supply is a
high-pressure air line (110 to 120 psig) connected to the shop's air compressor. Alcohol drawn from a
5-gallon drum by an internal pump is sprayed inward toward the center of the manifold, which moves
up and down at about 4 cycles per minute while spraying to better reach all of the part surfaces. While
IPA is being delivered, runoff is caught by a pan at the bottom of the unit, flows to a drain, and is
pumped by a smaller air motor back into the 5-gallon drum. When the IPA spray is deactivated after 5
minutes, the drain pump operates for another minute. Cabinet air is then purged by forcing air into the
chamber through a venturi and exhausting through a hose to the outside of the building. The IPA
rinser operates at ambient temperature.
The cleaner used in this study, Petroferm BIOACT™ 285, was selected because it is
representative of its class of material. BIOACT™ is a mixture of high molecular-weight aliphatic esters
and can be categorized as a semi-aqueous fluid. The cleaner is meant to be used without dilution and
must be rinsed with alcohol, such as IPA, rather than with water. The alcohol is a technical grade that
is at least 98% IPA, with the remainder being water. The IPA rinse is completely miscible with the
ester solvent. The IPA evaporates rapidly due to its high vapor pressure ( 3 mm Hg @ 68° F).
-------
METHODOLOGY
The product quality evaluation involved analytical testing to ensure that the new technology
provides an acceptably cleaned product. Cleaning effectiveness was evaluated to show whether the
new process cleans electric motor parts to an acceptable level of cleanliness (judged visually by an
experienced technician at Duffy).
A more quantitative approach also was used to monitor cleaning effectiveness. Two similar
motors were selected from the shop's inventory, each was disassembled and a rotor, housing with
stator, and end cover were saved. The two sets of motor parts were contaminated with an oil/carbon
mixture, representative of the soils found on actual motor parts. The soiled parts were heated in a
105°C oven for 16 to 24 hours (MIL-C-85570B) and then cleaned at regular intervals using the cleaning
system. Cleaning performance was determined by measuring residual soluble surface material on the
parts being cleaned in the cleaning system.
The parts were cleaned using the IBR cleaning system, then visually inspected by a Battelle
technician and by Duffy staff. Then the parts were cleaned by agitation in a 1-L bath of hexane. Parts
were classified as Set A or Set B. Afterward, the hexane was evaporated onto platinum weigh dishes
to determine nonvolatile matter according to ASTM D 1353, i.e., the amount of residue remaining after
cleaning.
Certain parameters for monitoring the condition of the cleaner and alcohol also were checked,
including appearance, color, nonvolatile matter (alcohol only), specific gravity, and pH of a water extract
of the ester cleaner.
One of the product quality parameters was designed to show possible adverse effects of the
cleaner and IPA on wire insulation materials. Tests were conducted to evaluate whether the
elastomers are compatible with BIOACT™ 285. Swell ratio (by weight and thickness change, ASTM
method D 2765) was measured in small (approximately 2-inch-square) coupons of the elastomers
Buna-N, Hypalon M, silicone, and neoprene. These materials are used for electrical insulation on the
wire leads of older electric motors.
Waste volume was determined by measuring the volumes of spent cleaner and IPA after
completion of testing. After cleaning about 108 parts over a twelve week period, the testing was
terminated and the cleaner and alcohol were disposed of.
The economic evaluation compared the costs of the new cleaning system with what is currently
used in the shop. A return on investment and payback period for installation of a new ester cleaning
system were not calculated because at this shop the ester-based cleaning system cost more to use on
an annual basis.
RESULTS
The data in Table 1 show that the residual soil levels on both sets of motor parts varied
consistently over the course of the study. Except for the first measurement (day 8), residue
measurements from day 37 to day 86 are within 1 to 2g per set of motor parts. The higher values on
day 8 are believed to be due to removal of debris from the motors or to incomplete cleaning, because
these results are not consistent with the remainder of the cleaning runs. Higher residue measurements
at 71 and 79 days are believed to be caused by soil-loading of the ester cleaner.
The check of appearance and color was used to assess how soiled the cleaner and alcohol
had become. By the 7th week of the study, or after washing 30 parts, the cleaner and rinse were very
darkly colored. After 100 parts were cleaned, they appeared highly contaminated. Specific gravity is
useful to track how much soil loading the cleaner and alcohol experience. Both solutions showed a
10
-------
small, monotonic increase in specific gravity over time. The main contaminants expected to be present
in the cleaner are oils and suspended solids; in the alcohol, the main contaminant is dragout cleaner.
The pH of the cleaning solution was measured for solution acidity or alkalinity changes over
time. This test was done to determine whether exposure to moisture over time caused any acid
increase in the ester cleaner. Acidity was determined by extracting samples of the cleaner with water
(ASTM D 2110). The pH dropped rapidly, reaching a steady state of about 5.04 in the seventh week.
In all cases, the materials in the compatibility tests experienced a net decrease in weight and thickness,
which is probably caused by removal of process oils, colorants, stabilizers, and other additives. The
loss of these constituents could have a negative effect on the performance of the elastomers.
In evaluating waste reduction/pollution prevention potential, the new cleaner achieves waste
reduction through the elimination of solvent air emissions and potential discharge of wastewater to the
environment because it is a nonaqueous process. However, the spent liquid must be disposed of.
Contaminants in the ester cleaner primarily are oil, grease, and shop dirt. Therefore, the cleaner itself is
assumed to present little environmental or health hazard during use.
Annual solvent usages were calculated to be 51.4 gal of cleaner and 55.2 gal of IPA. These
values represent a worst-case estimate for the cleaner because it was not fully spent at the time the
study was concluded. However, the cleaning performance was determined to be degrading. The
annual volumes of waste liquids were calculated to be 39.0 gal of cleaner and 39.1 gal of IPA.
The petroleum solvent formerly used by Duffy Electric Company had been supplied at a rate of
360 gal per year. Industry estimates indicate that about half of the amount of petroleum solvents
supplied is recovered. The remainder is lost due to dragout, evaporation, and spillage during transfers.
About 180 gal of spent solvent can be collected for distillation and later reuse, and another 180 gal is
unrecoverable. In contrast, the alternative cleaning system generates 106.6 gal of solvent (total) per
year, of which 16.1 gal of IPA is unrecoverable due to evaporation, and 78.1 gal of liquid waste is
produced.
Table 2 gives the annual operating costs of both the existing petroleum solvent cleaner and the
alternative ester-based cleaning system. The major operating cost of the new cleaning system is the
cost of the ester cleaner and IPA rinse. If a heating element is used, energy usage also needs to be
considered. Disposal costs may vary depending on the system currently in use. At this shop a
contractor retrieves the used petroleum solvent for recycling and supplies the shop with a clean
recycled product for use.
CONCLUSIONS
The ester-based cleaner used at Duffy Electric performed adequately for cleaning parts when
used in conjunction with a heater or with manual scrubbing for final touch-up cleaning. Soiled parts
were cleaned in the ester system to determine the amount of residue remaining after cleaning to
determine the life of the cleaner and to observe the rate of degradation. In general, parts removed from
the alcohol rinse were visibly dry and residue-free upon removal,
The economics of the ester cleaner are not advantageous when compared to the recycled
petroleum solvent used at Duffy Electric. Because the same contractor handles the supply and recovery
of the petroleum solvent, its cost is relatively low. Currently, the ester cleaner is not recycled by a
contractor, although the potential for recycling does exist. When the economics of the ester cleaning
system are compared to those of other existing systems such as the use of chlorinated solvents, the
results will differ. The ester cleaning system is likely to be less expensive than chlorinated solvent use.
The pollution prevention aspects of the ester-based cleaning system must be looked at on a
case-by-case basis. In the shop that participated in this study, Safety-KIeen is contracted to provide a
11
-------
recycled petroleum solvent and to perform recycling on the spent product, so the solid waste generated
is minimal. Currently, the ester cleaning waste and the alcohol rinse are sent to hazardous waste
disposal sites, and therefore the pollution prevention benefits for this shop in using an ester cleaner are
limited to reduced waste toxicity and decreased VOC emissions. With further commercialization of the
product, recycling may become more common and the benefits more apparent. In shops that use
chlorinated solvents, which are usually more regulated, the pollution prevention benefits of an ester
cleaner become more obvious.
Air emissions also need further study when evaluating the benefits of an ester- and alcohol
based cleaning system. Industrial suppliers of petroleum solvents estimate that approximately 50% is
unrecoverable. The petroleum distillate solvent contains an unknown amount of toluene, ethylbenzene,
and other hazardous air pollutants. Certain chlorinated solvents result in evaporation of high
percentages of these compounds. Many such compounds are included in the 33/50 Program list and
are regulated as hazardous air pollutants (HAPs) under the Clean Air Act. The ester-based cleaner
itself does not result in any significant evaporation. Although the alcohol rinse has a high evaporation
rate, its constituents are not as hazardous as those found in petroleum solvents or chlorinated solvents.
Health hazards, although not specifically studied in this testing, are important to weigh when
considering the ester-based cleaning system. The ester cleaner does not have the defatting properties
common in other cleaners that lead to skin irritations and dermatitis.
TABLE 1. TOTAL SOLIDS RESIDUE ON PARTS
Weight of Solids (g/part set)
Sample
ID
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
H13
Motor
Part
A
B
A
B
- A
B
A
B
A
B
A
B
Sample
Date
6/17/93
6/17/93
7/16/93
7/16/93
7/27/93
7/27/93
8/6/93
8/6/93
8/19/93
8/19/93
8/27/93
8/27/93
Days of
Study
8
8
37
37
48
48
58
58
71
71
79
79
Primary
3.22
3.05
1.99
1.44
1.83
1.33
1.24
1.62
2.60
1.50
2.79
2.73
Duplicate
3.27
3.29
1.76
1.24
2.06
1.55
1.01
1.66
1.98
1.30
3.42
2.81
Average
3.25
3.17
1.87
1.34
1.94
1.44
1.13
1.64
2.29
1.40
3.11
2.77
12
-------
TABLE 2. ANNUAL OPERATING COSTS
New Cleaning System
Ester: 51.4 gal @ $20.00 per gal
Isopropyl Alcohol: 55.2 gal @ $3.00 per gal
Disposal: 78.1 gal @ $2.50 per gal
Total8
$1,028
$166
$195
$1,389
Existing Petroleum Solvent System
Solvent purchase and disposal
Total"
$1,070
$1,070
a Total does not include labor, energy, and small maintenance costs.
b Total does not include cost of drying the parts, should faster drying be necessary.
13
-------
EVALUATION OF SUPERCRITICAL CARBON DIOXIDE SPRAY TECHNOLOGY
TO REDUCE SOLVENTS IN A WOOD FINISHING PROCESS
Paul M. Randall
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
(513) 569-7673
INTRODUCTION
The purpose of this evaluation was to provide an objective evaluation of the use of supercritical
carbon dioxide (CO2) as an alternative technology for spray applied in wood finishing processes using
reduced solvent formulations. Union Carbide has pioneered this technology under the UNICARB™
trademark. In the UNICARB™ process, the solvent-like properties of supercritical CO2 are employed to
replace a portion of the organic solvent in the conventional coating formulation. The supercritical CO2
acts as a diluent solvent to thin the viscous coating just prior to application so that the coating can be
atomized and applied with a modified spray gun. According to Union Carbide, 30 to 80 percent of the
organic solvent in a coating formulation can be replaced with the supercritical fluid.
This evaluation addressed the issues of product quality, pollution prevention potential, and
process economics. The testing was conducted at the Pennsylvania House Furniture Company in
White Deer, PA. The White Deer facility produces cherry and oak chairs, stools, dining room tables,
and four poster beds. At the time of the evaluation, the White Deer plant had been using the
UNICARB™ process to apply nitrocellulose lacquer finish on their chair line for over a year with good
results. Testing was done to quantify and qualify these results.
METHODOLOGY
In this technology evaluation, three issues were examined:
• Product Quality: to show that a coating applied by this technology meets company standards
for a quality finish.
• Pollution Prevention Potential: to demonstrate that use of spray application technology for
solvent replacement in coatings reduces volatile organic compounds (VOCs) released during
finishing operations.
Economics: to document the cost to install and operate this pollution prevention technology on
an existing spray coating finish line.
In the product quality evaluation, the objective was to determine whether nitrocellulose lacquer
applied by the UNICARB™ process provided a wood finish of equal or better quality than does the
conventional nitrocellulose formulation and spray technique previously used by Pennsylvania House. At
Pennsylvania House, the appearance and quality of the final finish are judged through visual
examination by inspectors on the coating line. Special attention is given to gloss, smoothness, and lack
of surface defects such as blisters or pinholes. In this project, product quality was evaluated through
independent evaluations performed by Pennsylvania House staff and coatings experts at Battelle. Nine
chair-back splats(three sample sets) were prepared by Pennsylvania House staff on the chair finishing
line. All panels were finished by the same production methods that typically are used on the chair line.
14
-------
The pollution prevention potential of the UNICARB™ process is based on reducing emissions
of organic solvents without adding to other wastestreams. In this project, VOC emissions from the
coating formulation and spray booth wastes(including solvent laden filters and nitrocellulose dust) were
evaluated. In the UNICARB™ process, most of the fast- and medium-drying solvents are removed from
the formulation, and the slow-drying solvents are adjusted slightly for better film formation. Supercritical
CO2 is used to replace the fast- and medium-drying solvents. Thus, the supercritical CO2 spray
process was expected to reduce VOC emissions from the chair-finishing line at the Pennsylvania House
plant.
The objective of the economic analysis was to determine the payback period for the switch to
the supercritical spray process from the conventional system. This was accomplished by comparing the
operating costs of the UNICARB™ process to the conventional finishing process. The initial investment
by Pennsylvania House in capital equipment and installation costs was considered, as were operating
costs, which include materials, waste disposal, labor, and utilities. A return on investment (ROI) and
payback period for the conversion was calculated.
For further details, a Quality Assurance Project Plan (QAPP) was prepared (USEPA & Battelle,
1993) which described the approach and scientific rationale.
RESULTS
The results of the product quality assessment indicated that the nitrocellulose lacquer finish
applied by the supercritical CO2 process is of equal or better quality than the finish obtained by
conventional methods. Product quality verification was supported by coating experts, who typically will
be more critical of the more subtle aspects of the coating finish, and by non-expert examinations. Both
groups indicate that the UNIGARB™ process will provide a finish with acceptable consumer appeal.
Thus, it is concluded that the use of this process for the application of nitrocellulose lacquer finishes on
the chair line at Pennsylvania House in no way compromises the quality of the finished product. These
results are supported by the fact that Pennsylvania House has not seen an increase in returns since
the inception of the UNICARB™ process. Further, the number of chairs that have to be reworked in
the plant to remove finish defects before shipping has decreased, indicating that the production
efficiency has remained the same or increased slightly since the UNICARB™ process was imple-
mented.
Pennsylvania House records indicate that it takes approximately 16 oz of the conventional
formulation to apply the two coats needed to achieve the desired quality in the finished product. The
UNICARB™ process required about 7 oz of the reduced solvent formulation per furniture unit to
achieve the same quality. There are two reasons that a smaller volume of coating is, required when the
UNICARB process is used: the higher solids content of the UNICARB™ formulation means more resin
is transferred to the substrate per volume of formulation sprayed; and the increased viscosity of the film
deposited by the UNICARB™ process inhibits film buildup by soaking into the wood substrate.
Data on the volatile content of the conventional nitrocellulose coating and the UNICARB™
coating were obtained from the coatings formulator (Lilly) and from direct determination of percent
volatile/percent solids content of samples collected. VOC contents were determined by methods
outlined in ASTM D2369. Liquid samples of both the conventional nitrocellulose coating and the
nitrocellulose coating modified for the UNICARB™ process were collected for laboratory analysis.
Coating formulation samples were taken from drums of coating material received by Pennsylvania
House from Lilly. The supplier, lot number, and date were recorded in the laboratory record book for
each field sample. At the White Deer facility, drums of coating are temperature equilibrated in the paint
room for at least 24 hours before mixing and dispensing to the spray guns. Three samples were
pumped into glass jars with Teflon® lid liners from a drum of conventional solvent-based nitrocellulose
and three from a drum of the nitrocellulose coating reformulated for the UNICARB™ process. Sample
containers were sealed to prevent evaporation during shipping.
15
-------
The Material Safety Data Sheets (MSDS) for each of the formulations
indicate the UNICARB™ coating is formulated using 16.7 percent less solvents (on an absolute basis)
than the conventional formulation. Only 9.67 percent of the UNICARB™ formulation is comprised of
Hazardous Air Pollutants(HAP) materials compared to 30.46 percent for the conventional formulation.
On a per-gallon-of-coating-sprayed basis, this would result in a relative decrease in VOC emissions of
22.81 percent, with a 68.25% decrease in HAPs using the UNICARB™ formulation and 5.9 Ib/gal for
the conventional system.
In order to determine the reduction in VOC emissions on an annual basis, the number of gal of
each formulation sprayed per year (Q) was estimated using the following equation:
Q = PVxDOPxOZ/16x1/D
where PV is the daily production volume, DOP is the days of operation per year, OZ is the amount of
coating sprayed per unit, and D isjthe density of the coating formulation. An average production rate of
250 chairs a day was assumed, with 200 production days a year. Pennsylvania House operates one
shift a day. Density is reported on the MSDSs as 8.0 Ib/gal for the UNICARB™ formulation and 7.7
Ib/gal for the conventional formulation. The total volume of each coating sprayed per unit has been
established at 7 oz for the one-coat UNICARB™ process and 16 oz for the two coats of conventional
formulation. Under these assumptions, 2734.4 gal of UNICARB™ formulation are needed compared to
the 6495.5 gal of the conventional formulation needed to finish the same number of units.
Using the VOC content reported on the MSDSs, this corresponds to an annual VOC emission
reduction of 67.5 percent when the newer process is used. Even if the VOC content of the formulations
were the same, an annual reduction in VOC emissions of 57.9 percent would be achieved because of
the decreased amount of formulation needed per unit using the UNICARB™ process to achieve the
same quality of finish.
Supercritical CO2 is used in the UNICARB™ process to decrease VOC emissions. The
reduced solvent formulation used by Pennsylvania House requires approximately 2.43 Ib of CO2 for
every gal of coating concentrate sprayed with the UNICARB™ process. The amount of CO2 released
annually into the atmosphere can be determined based on annual usage of the UNICARB™
formulation. This has been determined previously to be 2734.4 gal for an annual production of 50,000
units. Using this as a basis, the annual emission of CO2 from the finishing process is expected to be
6645 Ibs (2.43 Ib/gal x 2734.4 gal).
Carbon dioxide is not being produced through use of the UNICARB™ process. It is simply
being used as a substitute solvent to thin and aid in the spray atomization process. The CO2 used in
this technology is supplied by various distributors of CO2 which obtain and sometimes purify CO2
generated as a by-product of other chemical processes. Thus, CO2 used in processes such as
supercritical CO2 spray application of coatings does not actually contribute to the emission of CO2 into
the atmosphere.
Coatigg overspray at Pennsylvania House is collected on dry filters that are compressed and
stored in 55-gal drums for disposal by landfill. One drum, at a disposal cost of $150/55 gal drum, can
hold about 200 compacted filters and solid debris. Waste products generated will include dry and
solvent-laden filters and nitrocellulose dust, both loose and trapped in the filters. No liquid waste was
generated. Because Pennsylvania House does not separate waste by production lines, no physical
data were available for the solid wastestream analysis. However, discussions with Pennsylvania House
management and staff consistently indicated that the solid and liquid wastestreams were unaffected by
the conversion to the supercritical CO2 technology.
The chair-finishing process is the same except for the application of the nitrocellulose lacquer
finish. Using the old process, two booths were in operation which required cleaning and maintenance.
16
-------
With the UNICARB™ process, only one booth is needed. The transfer efficiency of the modified
UNICARB™ spray gun and the air-assisted spray gun are both approximately 50% based on Union
Carbide records. However, due to the increased solids content of the UNICARB™ formulation, more
solid waste is generated from the overspray by the UNICARB™ process. The 28 dry filters in each of
the two spray booths needed for the conventional two-coat finishing process were changed once per
week for a total disposal rate of 56 filters/week. The 28 dry filters in the one spray booth required for
UNICARB™ finishing are changed twice per week for a total of 56 filters/week. Dry paint and dust from
the booths is packed in the disposal drums with the filters, but no increase or decrease in the total
volume of these products was noted. In conclusion, no change was observed by Pennsylvania House
in the volume of solid waste generated by converting to the supercritical CO2 spray process on the
chair line.
The results of the pollution prevention analysis clearly indicate that a reduction in VOC emis-
sions occurred with use of the UNICARB™ process. The only new by-product of the process intro-
duced into the wastestream is CO2, but market information clearly indicates that the CO2 sold com-
mercially is itself a by-product of other production processes. Thus, the emission of CO2 from the
UNICARB™ process is not considered a detriment to the environment. Although the higher solids
UNICARB™ formulation made it necessary to clean the spray booth more often, the difference in waste
generation was offset by the fact that only one spray booth is needed with the UNICARB™ process.
The initial investment for the UNICARB™ process was $ 58,000, of which $ 46,000 was for
equipment purchase and $ 12,000 for installation of the equipment. The operating costs were based
on production of 50,000 chairs per year. The UNICARB™ process costs of $ 45,546 included $ 35,848
for the coatings formulation and $ 9,698 for the CO2 equipment rental and concentrate. The
conventional formulation cost was $ 46,883. By converting from a two-coat process to the one-coat
UNICARB™ process, Pennsylvania House was able to decrease its utility costs by $ 11,000 because
there was one less booth to operate and labor costs by $ 46,000 because one less finisher and one
less sander were needed. No change was assumed for line waste handling and disposal costs or for
finishing line maintenance.
The economic evaluation demonstrated a positive return on investment after the first year with
a total payback period within 3 years if gas utility savings are included, and 5 years if gas utilities are
not included.
This analysis reflects the economics of the actual operation is use on the chair line at the time
of this evaluation. Implementing the UNICARB finishing process on the chair line at Pennsylvania
House resulted in substantial annual savings in both utilities and labor. Cost savings could be realized
from a decrease in raw materials costs, but these savings are offset by the leasing fees for the CO2
tank and pump at Pennsylvania House. Additional savings could be realized by decreasing the size of
the existing ovens to reflect the change to a one-coat system.
CONCLUSIONS
This evaluation of supercritical CO2 spray technology for application of solvent-borne coating fo-
cused on three aspects: product quality, pollution prevention potential, and process economics.
The quality of the one-coat nitrocellulose lacquer finish applied at Pennsylvania House Furniture
Company by supercritical CO2 spray technology was demonstrated to be equal to or better than the
quality of the two-coat finish applied by conventional air-assisted airless spray in this evaluation. In
production, the furniture finish passes or fails on the basis of subjective evaluation of the total appear-
ance by Pennsylvania House experts and ultimately by the customers. Quality of the supercritical CO2
finish was supported by subjective evaluations by Pennsylvania House staff, coatings experts in the
Battelle Coatings Group, and a group of non-experts, as well as by Pennsylvania House's records on
customer acceptance and rates of in-plant defect corrections spanning more than one years's produc-
17
-------
tion line use of the supercritical CO2 spray technology support.
Release of volatile organic compounds during the finish process was reduced at Pennsylvania
House by the supercritical CO2 spray technology. The CO2 used in this process is recovered from the
wastestream of other industrial processes so it is not an additional contributor to global warming.
Overall CO2 may be decreased because many organic solvents that can add CO2 to the wastestream
are eliminated from the coating formulation. An annual reduction in VOC emissions in the range of 57%
to 67% was demonstrated. Much of this reduction occurred because supercritical CO2 is used at
Pennsylvania House to apply a one-coat finish. The conventional finish process required two coats of
nitrocellulose lacquer. Solid waste remained the same.
Capital investment costs incurred by Pennsylvania House will be recovered in the first 5 years
of operation. Most of the economic benefit gained from conversion to the supercritical CO2 process can
be attributed to the reduction in labor and operating costs on the chair line at the White Deer plant.
This technology is one approach to reducing VOC emissions in the application of solvent-borne
coatings. Product quality can be maintained and operating costs can be decreased. Capital costs will
vary with each implementation but a favorable payback period can be anticipated, in light of the findings
of this evaluation.
This technology evaluation focused on a single product type and coating formulation wood
furniture industry. However, this specific supercritical CO2 spray technology seems adaptable to a
number of solvent-borne coating formulations and products.
REFERENCES
Battelle Memorial Institute, Quality Assurance Project Plan for Evaluation of Supercritical Carbon
Dioxide Technology to Reduce Solvent in Spray Coating Applications, May 13, 1993.
Heater, KJ. and Parsons, A.B., Olfenbuttel, R.F., Evaluation of Supercritical Carbon Dioxide
Technology to reduce solvent in spray coating applications," EPA/600/R-94/063, U.S. Environmental
Protection Agency, Cincinnati, Ohio, April 1994.
FOR MORE INFORMATION
Paul M. Randall
U.S. Environmental Protection Agency
Pollution Prevention Research Branch
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
Telephone: 513-569-7673
Fax: 513-569-7111
Email: Randall.Paul@epamail.epa.gov
18
-------
DEMONSTRATION OF ALTERNATIVES FOR VAPOR DEGREASERS
Dean Menke - Center for Clean Products
Rupy Sawhney - Department of Industrial Engineering
University of Tennessee
327 South Stadium Hall
Knoxville, Tennessee 37996-0710
(615)974-8979
INTRODUCTION
The "Cleaner Technology Demonstrations for the 33/50 Chemicals" is a cooperative agreement
project between the Center for Clean Products and Clean Technologies and the U.S. EPA. Though
originally designed to support the 33/50 Program, the results of this RREL-funded research will have a
broad range of applications within industry and offer pollution prevention benefits beyond the 33/50 goals.
The overall objective of this project is to evaluate substitutes of the 33/50 chemicals in order to encourage
reductions in their use and release within specified priority use clusters. Priority use clusters, identified in
the "Product Side of Pollution Prevention: Evaluating Safe Substitutes for the 33/50 Chemicals" report,
are products and/or processes that consume a significant fraction of the 33/50 chemicals (1). The first
evaluation, presented here, focused on the metal and parts degreasing priority use cluster and specifically
substitutes for solvent degreasing processes that eliminate the use of the chlorinated degreasing solvent
dichloromethane, tetrachloroethylene, 1,1,1 -trichloroethane, and trichloroethylene.
In this study the Center for Clean Products worked directly with an industry partner to
demonstrate substitute feasibility and to gain actual industrial information. Calsonic Manufacturing
Corporation (CMC) is aggressively pursuing less polluting alternatives to solvent degreasing and agreed
to participate as the Center's industrial partner to demonstrate solvent degreasing substitutes. CMC
manufactures automotive parts included heaters, blowers, cooling units, motor fans, radiators, auxiliary oil
coolers, and exhaust systems. Over the past four years, CMC had evaluated and implemented a number
of environmental improvements to completely eliminate 1,1,1-trichloroethane (TCA) from their degreasing
processes. This research focused on two of these improvements; an aqueous wash system which
replaced five vapor degreasers of the radiator manufacturing line, and a no-clean processing alternative
(i.e. application of an evaporative lubricant which does not require cleaning for subsequent processing)
which eliminated two vapor degreasers of the condenser manufacturing line.
METHODOLOGY
The technical, environmental, economic, and national impact evaluations performed for the
.aqueous wash system and no-clean alternatives employed at the CMC facility had the following specific
objectives:
1. technical evaluation
n evaluated the substitutes' effects on process and product performance as
compared to the solvent degreasing processes
2. environmental evaluation
a evaluated the releases and off-site transfers of the 33/50 chemicals in the
production process compared to the substitutes' chemical releases and transfers
3. economic evaluation
n evaluated the costs, traditional and nontraditional, of the substitutes as compared
to the 33/50 chemicals
4. national impact evaluation
n evaluated and compared the overall life-cycle national environmental impacts of
replacing the 33/50 chemicals with the substitutes
19
-------
Data required to perform the technical, environmental, and economic evaluations were collected
from CMC through data request tables, site visits, and interviews with CMC employees. Data request
tables, completed by CMC employees and during site visits, allowed for the collection of process
information including capital costs, operating and maintenance costs, utilities consumption, and production
data. Questions concerning generation rates and disposal costs of waste (hazardous and non-hazardous)
and wastewater accompanied the data request tables, as well as questions concerning permitting
requirements. Tables and questions were directed at operations both before and after the process
changes.
Site visits and interviews allowed Center staff to become familiar with the day-to-day operations of
each CMC manufacturing line of interest. This information was used to extend the traditional economic
evaluation by using activity-based cost accounting. Activity-based cost accounting specifically identifying
the frequencies, durations, costs, and possible chemical emissions for every activity required to operate
and maintain the solvent degreasers and alternative systems. Direct manufacturing activities, as well as
indirect support activities (e.g. paper work, waste management, supervision) were identified and included
in the evaluation.
These evaluations of CMC, supplemented by on-line databases and literature sources, were used
to estimate the national environmental impacts that could occur if entire industrial sectors replaced solvent
degreasing systems with the alternatives.
RESULTS
For this study, process and product performance were used as the two parameters to evaluate
the technical feasibility of the alternative cleaning systems. As part of a continuous manufacturing line, the
cleaning process (or no-clean alternative) has the potential to influence both of these parameters.
Process performance was defined as the rate of production. Product performance was based on the part
reject rate per unit of production which was determined from the leak test records of every unit
manufactured. The production and part reject rates when the solvent degreasing processes were on-line
were used as the baseline for comparisons with the alternative processes.
Production rates and part reject rates were both established through historical records and
employee interviews. Evaluation of this data revealed that the production rate of either process line
(radiator or condenser) was not affected by the change to the alternative system. Neither was the part
reject rate of the condenser line, both before and after the process change to the no-clean alternative.
The part reject rate for the radiator line, however, did significantly decrease after the aqueous wash
system was installed. By implementing the aqueous wash system, and through the efforts of a Radiator
Task Force established by CMC, the leak detection rate of the radiator line was decreased nearly 77
percent.
Though the alternative processes eliminated TCA releases and transfers from the radiator and
condenser process lines, other chemical releases and transfers resulted from their implementation.
Therefore, it was necessary to evaluate multiple medias (land, air, and water), as well as hazardous and
nonhazardous wastestreams, to capture the full impact of the changes to the alternative processes.
Air releases and off-site transfers, reported to the 1992 Toxic Release Inventory (TRI), were the
predominant releases and transfers of TCA from CMC's manufacturing facility. Table 1, below,
summarizes these releases and transfers, and shows how they decreased over the past four years. TRI
only requires facilities to report total releases and transfers of a chemical, not process-by-process
releases or transfers. Therefore, specifically identifying the contribution to the overall reductions from
either the radiator or condenser process lines was not possible. However, chemical use records for these
process line, and employee interviews establish the following estimates:
1. the radiator process line, consuming 250,400 Ib. of TCA for solvent degreasing in 1990,
released 115,000 lb./yr. in 1990, 86,800 lb./yr. in 1991, and 0 lb./yr. in 1992; and
2. the condenser process line, consuming 88,550 Ib. of TCA for solvent degreasing in 1992,
released 75,500 lb./yr. in 1992, and 0 lb./yr. in 1994.
The implementation of these alternatives eliminated this consumption of TCA and the releases and
20
-------
transfers associated with its use.
The implementation of the aqueous wash system for the radiator line, however, generated an
8,400 gallon/day water wastestream. Treated at an on-site pretreatment facility, this wastewater
represents a significant waste management change. A nonhazardous, oily wastestream, skimmed from
the surface of the aqueous wash reservoirs, was also a newly generated wastestream of the aqueous
wash system. The no-clean alternative, by applying an evaporative lubricant to eliminate the need for
parts cleaning, generated a new source of volatile organic compound (VOC) emissions to air. Based on
lubricant consumption records, and assuming 100 percent evaporation, approximately 4,000 pounds/year
(1.7 pounds/day) of volatile organics are emitted to the air from this alternative process.
TABLE 1. CMC TRI-REPORTED RELEASES AND TRANSFERS OF TCA
Year
1990
1991
1992
1993
1994*
TCA Air Emissions
(lb./yr.)
425,756
194,622
176,239
89,446
66,800
Percent Change
-
-54.3
-9.4
-49.8
-25.3
TCA Off-Site
Transfers (lb./yr.)
233,530
338,525
206,345
194,975
109,000
Percent Change
;
45.0
-39.0
-5.5
-44.1
Values estimated from eleven months of TCA purchase records and trends of previous years
The traditional economic evaluation, results of which are presented in Table 2, indicated return on
investments in as little 0.3 years (CMC-determined Rl for the condenser line). The activity-based costs
accounting economic evaluation had not been complete at the time of this abstract publication. However,
initial review of the activities recorded during site visits to CMC identified significant differences in the
required activities between the solvent degreasing processes and those of the alternative systems. These
differences centered around two operations; one being the activities required to manage toxic chemicals
and toxic waste; the other was the costs associated with the treatment of the aqueous system's
wastewater. These results will be available by the time of the presentation, and copies of the methodology
and results will be available.
TABLE 2 - COMPARISON OF SPECIFIC TRADITIONAL COSTS
Costs
Capital Investment
Chemical Costs
Waste Disposal *
Radiator
Degreasers
not avail.
$182,490
$20,000
Aqueous System
$463,595
$21,400
$12,430
Condenser
Degreasers
not avail.
$67,040
$13,735
Evap. Lube.
$44,000
$4,720
$0
Chemical releases and transfers occur through out their life cycles; from their production, use, and
disposal. Significant changes in these emissions can occur if entire industrial sectors were to implement
alternatives to solvent degreasing similar to those of CMC. Therefore, a life-cycle, multi-media approach
21
-------
to the national environmental impact evaluation was used to capture the overall environmental impacts of
the alternatives.
Production facility releases and transfers of the chlorinated degreasing chemicals, in TRI reporting
year 1992, totaled 1,286,823 Ib. An estimated 34 percent of the chlorinated solvents produced in the U.S.
were used in solvent degreasing applications in 1992 (2). Using a life-cycle approach, some fraction of
the production emissions may be attributed to solvent degreasing; 34 percent to the production releases,
establishing the potential upper boundary, equalled 440,000 Ib. The EPA estimates that 24,500 solvent
degreasers were operational in 1992 within the US (3). These solvent degreasers consumed
approximately 440 million pounds of chlorinated solvents. Based on this information, the EPA also
established a 1992 air emission baseline from these 24,500 solvent degreasers at 283.5 million pounds
(4). Eliminating the use of chlorinated chemicals in solvent degreasing processes would greatly reduce or
eliminate these emissions, both associated production releases and transfers, as well as the use and
disposal releases and transfers. Phase-out regulations for TCA will reduce the use and releases/transfers
of TCA regardless of the degree of which these alternatives are implemented.
The alternatives to solvent degreasing also have their life cycle environmental releases and
transfers associated with them. Aqueous detergents may include in their formulations surfactants,
saponifiers, chelators, corrosion inhibitors, and stabilizers. Specific examples from each of these additive
classes were analyzed. Disposal of the water wastestreams may have significant effects on publicly
owned treatment works (POTW). The POTW infrastructure of the nation was evaluated, and the potential
impact the aqueous wash systems have on the infrastructure was established. A similar life-cycle
approach was used to evaluate the mineral-spirits-based evaporative lubricants.
CONCLUSIONS
A significant number of studies are being conducted, or have been completed, which evaluate the
effectiveness of cleaning alternatives. These studies primarily focus on one of the four evaluations
performed in this study; little integration of all potential issues is attempted. This cooperative agreement
with EPA expands the existing knowledge of alternatives to solvent degreasing by integrating technical,
environmental, and economic issues, as well as addressing the life-cycle attributes of the alternatives on a
nation scale.
The technical feasibility of CMC's process changes has proven to be positive. Significant
reductions in toxic chemical releases and transfers were a result of the process changes, while other
wastestreams were generated which required different management schemes. The traditional economic
evaluation of this study did not reveal any unique conclusions. However, the activity-based cost
accounting method did identify the costs associated with managing toxic chemicals and wastes, costs
normally absorbed by the company as overhead. Finally, the national impact evaluation identified the
importance of a life-cycle approach to evaluate pollution prevention projects. Though the alternatives
evaluated in this research eliminate chlorinated chemical emissions, there are new wastestreams and
wastestream constituents that must be addressed.
REFERENCES
1. Product Side of Pollution Prevention: Evaluating Safe Substitutes of the 33/50 Chemicals,
EPA/600/R-94/178, U.S. Environmental Protection Agency, Office of Research and Development,
September 1994
2. Product Side of Pollution Prevention: Evaluating Safe Substitutes of the 33/50 Chemicals,
EPA/600/R-94/178, U.S. Environmental Protection Agency, Office of Research and Development,
September 1994
3. National Emission Standards for Hazardous Air Pollutants: Halogenated Solvent Cleaning -
Background Information Document, EPA-453/R-93-054, U.S Environmental Protection Agency,
22
-------
Office of Air Quality, November 1993.
4. National Emission Standards for Hazardous Air Pollutants: Halogenated Solvent Cleaning -
Background Information Document, EPA-453/R-93-054, U.S. Environmental Protection Agency,
Office of Air Quality, November 1993.
FOR MORE INFORMATION:
Diana Kirk
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Drive, MS-466
Cincinnati, Ohio 45268
(513)569-7674
23
-------
RESEARCH SUPPORT TO THE POLLUTION PREVENTION PROGRAM
OF THE US POSTAL SERVICE
N. Theresa T. Hoagland
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Pollution Prevention Research Branch
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7783
INTRODUCTION
The USPEA Risk Reduction Engineering Laboratory's (RREL) Pollution Prevention Research
Program focuses on the question "How should consumer, government and industrial products and
processes be designed, manufactured, used, and/or performed so that their manufacture, use, disposal,
or performance will have a minimal effect on the environment." Research projects addressing this issue
are divided into five areas: (1) Clean Technology Projects; (2) Clean Products Projects; (3) Longer Term
P2 Research; (4) P2 Assessments; and (5) Cooperative P2 Projects with Other Federal Agencies.
Under the general heading of Cooperative P2 Projects with Other Federal Agencies, a specific
research program exists at RREL, which is referred to as the WREAFS Program (Waste Reduction
Evaluations At Federal Sites). Many of the projects with the Departments of Defense (DOD) and Energy
(DOE), are funded under the Strategic Environmental Research and Development Program (SERDP).
The primary components of the WREAFS program are: (1) opportunity assessments; (2) research,
development and demonstration; and (3) technology transfer.
Most of the WREAFS projects involve Pollution Prevention Opportunity Assessments (PPOA) and
have been conducted in conjunction with DOD facilities; however, several "civilian" agencies are now
beginning to participate in the program, as well. One major player which is emerging as a leader in
addressing P2 on an agency-wide basis is the US Postal Service (USPS). Like most agencies, the
USPS' operations entail many industrial-type processes that are common to the private sector industrial
complex.
METHODOLOGY ,
The PPOAs were conducted by an assessment team that was composed of personnel from EPA,
USPS, and SAIC, under contract with EPA. The assessments followed the procedures described In the
EPA Report, Facility Pollution Prevention Guide (EPA/600/R-92/088), which identifies the four major
phases of an assessment: (1) Planning and Organization, which includes organization and goal setting;
(2) Assessment, which includes a careful review of a facility's operations and waste streams and the
identification and screening of potential options to reduce waste; (3) Feasibility Analysis, including an
evaluation of the technical and economic feasibility of the options selected and subsequent ranking of
options; and (4) Implementation, which involves procurement, installation, implementation, and
evaluation.
Many of the pollution prevention opportunities identified during WREAFS projects involve low-
cost changes to equipment and procedures that can often be implemented by the facility without
extensive engineering evaluations. Other pollution prevention opportunities identified require further
study before full implementation can be realized. Typically, opportunities requiring further evaluation are
those that have the potential to affect the process and/or require the use of new procedures or
equipment. In such cases, it may be necessary to conduct demonstration projects.
24
-------
Depending on the nature and state of development of the pollution prevention option selected
for demonstration and evaluation, these projects may include: (1) process design, (2) detailed design
and specification, (3) system procurement, (4) installation and start-up, (5) monitoring, and (6) reporting.
Some projects may require bench-scale and/or pilot testing prior to, or as a part of, the demonstration
project. Other projects may utilize full-scale equipment directly on the production line.
RESULTS
An example of the PPOA process as applied by the US Postal Service is the assessment
conducted at a USPS facility complex in Buffalo, NY. The complex encompasses 25 acres with a
vehicle maintenance facility (VMF) and general mail facility (GMF). At the VMF, a staff of automotive
technicians maintain a fleet of 1,200 vehicles, which range from light delivery vehicles to 18-wheel tractor
trailers. In a year, approximately 2,500 to 3,000 maintenance and repair jobs are performed and 500
vehicles completely painted with USPS colors at this facility alone, which is only one of 350 such
facilities nationwide.
The GMF consists of a three-story building and a one-story 276,000 square foot mail processing
floor (another building on the site houses the Computerized Forwarding System and the Undeliverable
Bulk Business Mail (UBBM) operation). Dock positions for more than 50 trucks are located here and the
building is occupied 24 hours every day by up to 1,515 full-time employees. The facility receives nearly
two million pieces of mail each day for processing; nearly three million pieces are sent out daily for
delivery. The Buffalo GMF processes all mail from the eastern seaboard destined for Canada, and as
such serves as a concentration center for mail transportation equipment.
These industrial-type facilities generate industrial-type pollution. The GMF generates
approximately 537 tons of waste per year, at an annual solid waste disposal cost of approximately
$42,000. The wastes are comprised of 253 tons of old corrugated cardboard, two tons of computer
paper, 46 tons of mixed office paper, 30 tons of metals, 13 tons of plastic film, 192 tons of
"undetermined wastes," and an undetermined amount of machinery maintenance wastes (oil, grease and
parts-cleaning solvent). Typical wastes that were produced by the VMF include about 4,100 gallons of
waste petroleum naphtha generated annually during brake and engine parts cleaning operations and
about 300 antifreeze were disposed of as hazardous waste.
As an example of the PPOA findings and recommendations:
Paint wastes comprise approximately 16 percent of all hazardous wastes generated by the VMF
and are responsible for half of the hazardous waste management costs. The three primary PPOA
recommendations for reduction of paint related wastes are described below.
• Investigate water-based or high-solids paints - Investigations at the VMF indicate that water-
based primers may be substituted for conventional ones without affecting performance. Water-
based top coatings, however, have been shown to exhibit insufficient durability. One solution
currently under evaluation at the VMF involves the use of a waterborne primer with an acrylic
enamel top coat.
« Conversion to HVLP paint application systems - HVLP spray guns operate at a much lower
pressure than conventional spray guns, resulting in considerably lower paint waste due to
bounce-back. Two HVLP spray guns were purchased for the VMF and have resulted in
improved-quality paint jobs and savings in paint usage. Because the equipment was installed
only recently, the VMF has not been able to provide accurate data on the incurred savings. The
VMF's conservative estimate is a twenty percent reduction in paint usage per vehicle, with an
associated reduction in VOC emissions. These figures are based on a four-month test period.
25
-------
• Install a paint-gun washer station - At present, paint guns are washed in an open solvent tank.
The PPOA recommended using specially designed enclosed paint-gun washing stations which
reduce VOC emissions by 75 to 90 percent. The VMF has plans to purchase a paint-gun washer
station during FY94.
CONCLUSION
Many of the recommendations made in the PPOA have been successfully implemented at the
Buffalo facility. Approximately 624 tons of solid waste were disposed of by the GMF in 1992, incurring
disposal costs of $42,120. In fiscal year 1993, waste generation was reduced to approximately 244 tons.
Additional cost reductions are expected as further plans are implemented.
Aerosol chemical use has been eliminated in all postal facilities in the western New York District.
Portable sprayers are used to dispense maintenance products, which are purchased in bulk five gallon
containers. As a result, significant reductions have been achieved in waste aerosol cans and packaging.
In 1992, over 2,400 aerosol paint cans were used at the VMF.
In the past, approximately 4,100 gallons of waste petroleum naphtha were generated annually
during brake and engine parts cleaning operations at the VMF. This represented about one third of the
total hazardous waste costs. The VMF has recently converted to an aqueous cleaner for cleaning
brakes, which will eliminate over 2,000 pounds of hazardous waste annually. In addition, a longer-lasting
solvent for cleaning automotive parts is in use which should also result in a substantial reduction in
hazardous waste generation.
The Buffalo PPOA was just the beginning. PPOAs are, or have been conducted at other
representative USPS facilities throughout the United States. These facilities include: a forensic
laboratory; an engineering and research development laboratory; a stamp distribution center; a bulk mail
center; an area supply center; and customer service centers. Results of these assessments and P2
plans will be shared with other Federal agencies and the private sector as they are received.
26
-------
LIFE CYCLE DESIGN FRAMEWORK AND DEMONSTRATION PROJECTS
Gregory A. Keoleian
National Pollution Prevention Center
University of Michigan
Dana Building 430 E. University
Ann Arbor, Michigan 48109-1115
(313)764-1412
Mary Ann Curran
Project Officer
Pollution Prevention Research Branch
Risk Reduction Engineering Laboratory
Office of Research and Development
US Environmental Protection Agency
Cincinnati, OH 45268
The life cycle design project was sponsored by the US Environmental Protection Agency's
Pollution Prevention Research Program and was organized into two phases: Phase I - development of
the life cycle design framework (published in Life Cycle Design Guidance Manual (EPA/600/R-92/226))
and Phase II - two demonstration projects and refinement of the framework (published in Life Cycle
Design Demonstration Projects: Profiles of AT&T and AlliedSional (in press))
Life cycle design is a proactive approach for integrating pollution prevention and resource
conservation strategies into the development of more ecologically and economically sustainable
products. The specific goal of life cycle design is to minimize the aggregate risks and impacts
associated with a product life cycle, which encompasses raw materials acquisition through materials
processing, manufacture and assembly, use and service, retirement, disposal, and the ultimate fate of
residuals. Environmental impacts include resource and energy depletion and ecological and human
health effects resulting from environmental releases and wastes. The following key elements of the life
cycle design framework are outlined: a firm's environmental management system, needs analysis and
project initiation, specification of design requirements, selection and synthesis of design strategies for
minimizing environmental burden, and evaluation of design alternatives using environmental analysis
tools.
Environmental issues can not be addressed in isolation of other design criteria. Multicriteria
requirements matrices were developed as a key tool for systematically identifying and evaluating
performance, cost, legal, and cultural requirements in addition to environmental criteria. Balancing
these criteria in successful designs requires the participation of members of a cross-functional design
teams. Industrial designers, process engineers, corporate executives, product development managers,
production workers, environmental health and safety staff, purchasers, accountants, marketers, and
legal staff each play a critical role in developing cleaner products.
Life cycle design is influenced by a complex set of factors. Design teams are challenged by
external factors that affect the design process such as government regulations, market demand, public
and scientific understanding of environmental risk, and existing infrastructure. Within a corporation
successful application of life cycle design depends on a firm's environmental management system,
which includes environmental policies and goals, environmental performance measures, organizational
design, education programs and other factors.
27
-------
In the Life Cycle Design Demonstration Projects, both AT&T Bell Labs and AlliedSignal, Filters
and Spark Plugs applied life cycle design to the development of cleaner products. AT&T focused on
achieving greater material and energy efficiency, a higher degree of recyclability, and fewer toxic
constituents and releases in their design of a business telephone terminal. AlliedSignal developed
design criteria to guide the improvement of future engine oil filters, the AlliedSignal team considered a
cartridge filter with a reusable housing and a single-use, spin-on design. Both AT&T and AlliedSignal
concluded that multicr'rteria requirements matrices are a useful tool for organizing, identifying, and
evaluating the complex set of life cycle issues affecting the design of a product system. Major
accomplishments and difficulties in implementing life cycle design are highlighted for each project.
The University of Michigan Research Team, in cooperation with the US EPA, is currently
investigating life cycle design of automobile powertrains, automotive coatings, beverage packaging, and
photovoltaics.
28
-------
Industrial Assessment Database for Energy Efficiency and P2
Michael Muller and Peter Polomski
Office of Industrial Productivity and Energy Assessment
Introduction
The Energy Analysis and Diagnostic Center/ Industrial Assessment Center (EADC/IAC)
Program, in existence since 1976, is a federally funded service provided to small to medium sized
manufacturing firms. The program is funded out of the Office of Industrial Technology of the U.S.
Department of Energy with support from the Pollution Prevention Research Branch of the U.S.
Environmental Protection Agency. Over 5000 energy assessments have been performed by teams
made up of faculty and students from engineering schools at universities throughout the United States.
Normally the teams perform a one day site visit at a manufacturing plant which follows an extensive
pre-assessment data gathering function. Following the site visit, the assessment team prepares a
written report for the industrial client which includes information about the plant's resource use,
processes, waste handling, and other operations. In addition, several assessment recommendations
(ARs) are written up with sufficient engineering design to provide for anticipated savings,
implementation costs and simple payback for each AR. Nearly 50% of the more than 35,000
recommendations have been reported implemented by the industrial clients.
The program has been highly successful since its inception with four schools. With 30
universities now active Centers, combined assessments encompassing both energy conservation and
waste minimization began on a limited basis during the 1994 program year. Expansion of this effort to
all participating schools is planned for 1996.
Program benefits are not limited to the manufacturers served, students educated and faculty
enriched but also extend to the community at large through data generated by the program. Since
1980, the data has been compiled from the assessments performed under this program and since FY
93 has been made available to the public via internet links to the Office of Industrial Productivity and
Energy Assessment (OIPEA) at Rutgers University. For FY 95, a major revision to the database and
the accompanying Assessment Recommendation Coding (ARC) scheme was undertaken to bring the
structure in line with the combined nature of the program.
The EADC/IAC Program Database
The EADC/IAC Program Database has evolved from a flat compendium of data issued annually
to a fully relational dataset updated daily. Data provided from the EADC/IAC reports is uploaded in
spreadsheet "boilerplates" via the internet to OIPEA where it is processed through various translation
algorithms and ultimately appended to the database. Responsibility for the accuracy of the data rests
with the uploading Center but consistency checks are included within the translation schemes.
The database consists of two separate datasets: the Assessment Database [ASSESSxxx.DBF]
contains information pertaining to each individual assessment and the Recommendation Database
[RECCxxx.DBF] where information pertinent to each recommendation arising from the site visit is
logged (Table 1). The database relation exists via a derived common identifier field consisting of the
EADC/IAC name and report number which together comprise the ID field in the assessment and a
portion of the SUPERID field in the recommendation database. To speed data transfer, most derived
fields such as the ID field in the recommendation data set have been left to the user.
29
-------
TABLE 1. FIELDS IN THE EADC/IAC DATABASES
Field Name
id
eadc/iac
repnum
visitdata
sfc
sales
employees
plant_area
products
resources
produnits
prodlevel
prodhours
numars
ASSESSMENT 1 1
Description
Unique identifying number given to all
records based on EADC/IAC Name &
Report # .
The Identifier assigned to each EADC
/IAC.
The number assigned by the
EADC/IAC to the visit & report.
r
The date of the assessment. |
The Std. Industrial Code for the
plant's principle product.
The annual sales in dollars for the
site reported by the client.
The total number of employees on
the site.
The total amount in square feet of
area used for production and office
support purposes.
Principle products of the plant (in
words).
The total number of resources
tracked at the plant
The units of production for the
principle product (coded).
The client reported total number of
units produced annually.
Client reported annual production
hours.
Field Name
superid
ar_number
appcode
arctype
arc
impdate
fmpstatus
impcost
impconser
impsaved
psourccode
pconserved
psaved
The total number of ARs
recommended in this report. |
"Rie annual usage and cost of energy
taken from actual bills.
nrgoosttot
Total energy cost for this client ||
ssourccode
^conserved
ssaved
RECOMMENDATION
Description
The unique identifying number given to all
records based on EADC/IAC Name, Report #
&Recc#.
The recommendation number sequentially as
it appears in the report
Application for recommendation (see ARC
Manual).
Recommendation type (see ARC Manual).
The code representing the specific
recommendation made (see ARC Manual).
Client reported date of implementation of this
Assessment Recommendation.
Client reported implementation status of this
Assessment Recommendation.
Client reported implementation cost. This cost
may be estimated.
Client reported amount of energy conserved
upon implementation of the Assessment
Recommendation.
Client reported amount of money saved upon
implementation of the Assessment
Recommendation.
The Primary Resource [coded]. This resource
may not necessarily be the most important
resource involved in the Assessment
Recommendation, but is usually chosen
based on greatest usage before conservation
measures are suggested (logic follows for
secondary & tertiary).
The amount of primary resource conserved.
The primary resource's dollar savings for this
Assessment Recommendation.
The Secondary Resource involved in this
Assessment Recommendation.
The amount of secondary resource conserved
The secondary resource's dollar savings.
30
-------
The annual production and cost of
waste.
wstcosttot Total waste cost for the client.
comments General comments about the
assessment.
fy The fiscal year in which the
assessment was performed.
tsourccode
tconserved
tsayed
rebate
incremntal
descript
The Tertiary Resource involved in this
Assessment Recommendation.
The amount of Tertiary resource conserved.
The Tertiary resource's dollar savings.
Indicative of whether the Assessment
Recommendation included a rebate.
Indicates if the AR is to be implemented on an
incremental basis- included in the database
for the first two years only:
Description in words of the AR.
Resource Stream Tracking in the Database
Program expansion into the waste reduction / pollution prevention arena necessitated a
revamping in the way resources are thought of in the database. Spirited debate centered around which
database modifications would expand the presented information without sacrificing understanding of the
issues. Ultimately it became clear that any suggestions or considerations for change must represent a
clear and unmistakable improvement over past practice. The judgments made often became a
compromise between two or more equally valid points of view with the simpler argument holding sway.
Perhaps the easiest decision centered on the number of resource streams to track ("track or
tracking" refer to information inclusion in the Assessment Database in terms of cost and usage and in
the Recommendation Database where savings and reductions appear). Waste reduction
recommendations tend to involve energy, resources, and waste. An increase from two to three streams
allows recognition of tertiary detail which in the past would have been buried in the recommendation
calculations. Recommendations have been calculated that involve more than three streams (even six
or more!) but good database management precludes carrying so many as this burdens the database
with a large number of empty fields acting merely as place holders.
The stream definition question focused debate on tracking issues important enough for
inclusion without becoming bogged down in minutiae. Previously, the only items tracked involved
energy sources, additional income, material costs and operating costs. Stream types now define four
primary areas of interest:
0 Energy
0 Waste Reduction
0 Resource Costs
<> Production
The energy streams remain unaffected. Waste minimization and pollution prevention plays
require the obvious consideration of liquid and solid waste (hazardous and non-hazardous) and
gaseous waste in the waste reduction category. Water disposal is differentiated from other liquids
conceding the importance of this shrinking resource. Resource costs break out detail from the old
material and operating cost fields. Material is broken into primary raw material and ancillary material
cost. Personnel changes and administrative costs cover the operating cost category. Water
consumption costs become germane to many waste reduction recommendations as to warrant inclusion
as a distinct heading. The production streams track one primary product and one byproduct only.
31
-------
Waste Reduction Assessment Recommendation Coding
Concurrent with the inclusion of waste reduction efforts in the industrial assessment is the
addition of a new coding scheme for the assessment recommendations. Older coding consisted
primarily of Energy Conservation Opportunities (ECOs) and the need arose for a coding system for
recommendations involving enhancements in energy efficiency, waste minimization and manufacturing
productivity. Most recommendations can be collected into groups that focus either on the same system
or on the same general strategy for enhancement. Attempts were made to develop a coding scheme
which would be consistent along either one of those lines, but neither approach proved satisfactory.
The resulting organization of recommendations has been done in an "expert system" fashion.
Therefore, the code has been assembled to best collect recommendations which would be considered
together by an experienced professional. For example, recommendations for energy savings for air
compressors (a system) are grouped. In a similar fashion, waste heat recovery (a strategy)
recommendations are collected together.
A coding system like this will change frequently as new technologies and strategies reach the
manufacturing floor. The ARC consists of a code as follows:
X.YYYY.Z
The first number, "X" is the recommendation type. Examples are 2 for energy savings, 3 for
waste reduction etc. The second four numbers, "Y1Y2Y3Y4", detail the strategy being employed. The
final number, "Z" is the application of the strategy, indicating whether the recommendation impacts the
process, the building and grounds, or other application (Table 2).
Database Strengths and Weaknesses
The EADC/IAC Program Database is a unique representation of over 15 years of manufacturing
site visits by engineering faculty. Updated daily, the database reflects the latest in industrial
assessment techniques, energy and waste costs for small to medium size industrial plants throughout
the U.S., recommended / implementation costs and savings for a large number of recommendations
involving a variety of resource streams, and other useful data. The database has been used
successfully around the world for pre-assessment preparation, utility Demand Side Management
planning, government policy recommendations, in conjunction with case studies to gain insights into the
investment criteria required for successful recommendations, and many other applications.
The database must still be viewed for what it is: a collection of information generated by human
beings. There can be errors made in almost any step of the data collection process. There are biases
in the data and disagreements over what constitutes an acceptable technology. Earlier datasets do not
provide the level of detail found after FY 94. Electricity data does not break out demand, etc. This
should be remembered during any and all analysis of the data.
Database Location
The database access location for downloading purposes is the OIPEA Home Page on the
World Wide Web (W3) at: "http://OIPEA-WWW.rutgers.edu".
32
-------
FOR MORE INFORMATION
Michael R. Muller
Director, Office of Industrial Productivity and Energy Assessment
Rutgers, The State University of New Jersey
Piscataway, NJ 08855-1179
(908)445-3655
TABLE 2. WASTE MINIMIZATION / POLLUTION PREVENTION ARC GROUPS
Major Group (Y1)
Sub Group (Y2)
Heading (Y3)
Operations
Procedures
WasteStream Contamination
CAD/CAM
Process Specific
Material Application
Stripping
Scheduling
Desulfurization/Slag Mgmnt
Reduction / Elimination
Product Specifications
Byproduct Use
Dragout Reduction
Rinsing Strategies
Miscellaneous
General
Equipment
General
Fault Tolerance
Painting Operations
Process Specific Upgrades
Tank Design
Automation
System Monitoring
Treatment / Minimization
General
Neutralization
Removal of Contaminants
Material Concentration
Water Use
General
Close Cycle Water Use
Reduction
Water Quality
Chlorination
33
-------
Recycling
Waste Disposal
Liquid Waste
General
Sludge Maintenance
Miscellaneous
Maintenance
Raw Materials
Cleaning / Degreasing
Spillage
Other
Solvents
Other Solutions
Solids
Mechanical Cleaning
Reduction of Cleaning
Rag Use
Preventive Maintenance
Operations
Hardware
Leak Reduction
Miscellaneous
Use Reduction
Emission Reduction
Material Replacement
Solvent Recovery
Water-Based Substitutes
Other Substitutes
General
34
-------
APPLICATIONS OF SUPERCRITICAL FLUID TECHNOLOGY TO POLLUTION PREVENTION AND
WASTE MINIMIZATION
Ronald J Turner
U.S. Environmental Protection Agency
Office of Research and Development
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
INTRODUCTION
Many commonly used solvents and extractants are hazardous materials. In the past, careless
disposal of these materials resulted in widespread contamination of groundwaters as proven by the
large number of Superfund sites having this kind of contamination. Lack of adequate worker protection
also resulted in a significant health risk in the workplace from contact with the liquid and vapor forms of
the solvents.
One approach to the prevention of pollution and the reduction of worker exposure from
conventional solvents is the use of less hazardous substitutes. One of the more novel substitutes is
carbon dioxide. Although carbon dioxide is a gas at ordinary conditions, it can be liquified by
application of pressure. In its supercritical state, it exhibits good solvent properties. Separation of the
carbon dioxide from the extracted materials can be accomplished by pressure reduction, adsorption
onto activated carbon, or with a membrane process, e.g., decaffeination of coffee.
Supercritical fluid extraction has been studied extensively on a laboratory scale to remove
organic compounds from hazardous mixtures and contaminated waters. Few such systems have been
employed commercially. One used liquified propane/butane to demonstrate extraction of organics from
sediment at New Bedford Harbor, Massachusetts, under the EPA's Superfund Innovative Technology
Evaluation program. However, supercritical carbon dioxide (SCCO2) may be the preferred choice for
industrial extractions where relatively low temperatures are required.
Carbon dioxide is a non-polar solvent which is relatively non-toxic, non-flammable, with a critical
temperature of only 31 degrees C and a critical pressure of 73 atmospheres: Extraction with SCCO2
offers several advantages over conventional solvent extraction, including minimization of organic liquid
waste generation. Its solvent strength and low viscosity dependence on temperature and pressure
makes SCCO2 attractive for replacement or recovery of organic compounds from liquid and solid
materials. The use of SCCO2 in a remediation process allows rapid extraction of toxics and relatively
easy solute concentration and recovery. Brady et.al.(1) successfully extracted PCBs, DDT and
toxaphene from two types of contaminated soil using SCCO2. They observed that over 90 percent of
the PCBs could be extracted in less than one minute. The SCCO2 extraction of naphthalene,
anthracene, hexachlorobenzene and phenol from soil has also been reported.
The EPA is supporting a competitive university-based supercritical carbon dioxide research
program administered by the Risk Reduction Engineering Laboratory, Cincinnati. The goal of this
program is to develop and demonstrate current and new uses of SCCO2 in industrial processes as
pollution prevention and waste minimization alternatives to chlorinated and other hazardous solvents.
The degree to which this goal is accomplished will be determined by the acceptance of the SCCO2
processes for commercialization. A major factor limiting the commercial applications of supercritical
fluid technology is the lack of reliable data for the design of the extraction or synthesis units.
This paper presents a brief overview of the university and federal laboratory collaborative efforts
focusing on processes using supercritical carbon dioxide as a substitute for hazardous solvents. The
SCCO2 technology applications under this program include extraction of natural pharmaceutical
materials, phase-transfer catalysis, solvent replacement in chemical synthesis, temperature-solubility
35
-------
relationships, and separation of organic materials from soils and slurries. A paper on the extraction of
heavy metals with SCCO2 was presented by Ataai et.al. at the 87th National Meeting, A&WMA (2).
This work is also supported by the EPA.
OVERVIEW OF THE SUPERCRITICAL CARBON DIOXIDE RESEARCH PROGRAM
Selective Oxidation in Supercritical Fluids (Department of Energy, Los Alamos National Laboratory)
This research partnership will develop new, environmentally-friendly, selective catalytic oxidation
routes to the production of oxygenated hydrocarbons, using supercritical carbon dioxide as the reaction
medium in place of hazardous process solvents. This should reduce by-product wastes and improve
downstream separation of products/catalysts. For example, a new class of heterogeneous oxidation
catalysts, titanium silicates, catalyze a number of useful reactions including epoxidation, alcohol
oxidation, aromatic hydroxylation, and ammoxidaition. Titanium silicates and analogs will be examined
in both supercritical carbon dioxide and water/carbon dioxide mixtures using a number of substrates
(alkenes, arenes, alcohols). The reactivity of olefins with several palladium catalysts in supercritical
carbon dioxide/water mixtures will also be investigated, as there is considerable industry interest in new
routes to polymer intermediates or precursors.
Supercritical Fluids for Phase-Transfer Catalysis (Georgia Institute of Technology)
This research, under C.A. Eckert, will study the replacement of toxic and hazardous fluids with
supercritical carbon dioxide to design "cleaner" manufacturing processes. This includes catalyst
solubility studies and determinations of reaction kinetics. Phase-transfer catalysis in supercritical fluids
will be demonstrated, providing opportunities for faster heterogeneous reaction rates at milder
conditions, tailored to achieve a variety of process goals. In correspondence with the EPA, Dr. Eckert
cited an example of a successful phase-transfer catalysis reaction in a supercritical fluid which
produced benzyl bromide from a quaternary ammonium bromide salt and benzyl chloride. This reaction
is representative of an industrially important class of processes. A major chemical company is
collaborating with Dr. Eckert to identify specific reactions where phase-transfer catalysis can be used
for pollution prevention. This partnering will improve commercialization opportunities.
Environmentally Benign Pharmaceutical Processing (Georgia Institute of Technology)
This study will employ supercritical carbon dioxide in a separation process, replacing alcohols
and chlorinated solvents to extract compounds of therapeutic value from natural organic materials. This
activity is being conducted under Dr. A.S. Teja, and includes study of solubility, extraction, and
analytical characteristics of taxol and taxol-related compounds. The present separation method uses
exhaustive ethanol extraction, partitioning between water and chloroform, followed by successive
chromatographic separations and crystallization from aqueous methanol. Dr. Teja proposes to combine
supercritical extraction with high pressure adsorption and other techniques to yield a novel hybrid
separation process. His initial studies indicate that the hybrid process will eliminate the use of
chlorinated solvents and reduce the use of alcoholic solvents by 85%. This process could potentially
be used for the production of a large class of taxol-like Pharmaceuticals.
Supercritical Carbon Dioxide Extraction at Elevated Temperatures (University of North Dakota)
The goal of this study is to provide fundamental solubility data for a variety of semivolatile
organic compounds, including polycyclic aromatics, and to determine the effect of temperature on CO2
extraction efficiencies of both polar and non-polar organic compounds. Solubility is expected to
increase dramatically with elevated temperatures (i.e., above typical 50 to 100 degrees C) for many
components thought to be insoluble, extending the usefulness of SCCO2 for industrial processing and
extractive clean-up. The results of the investigations by Dr. Steven Hawthorne should apply to the
supercritical carbon dioxide extraction of organic contaminants from soils, sediments, spent sorbents,
36
-------
contaminated catalysts, and industrial waste sludges. The data generated in this study will also be
used to predict solubility of organics in SCCO2, and will support the development of processes to
replace hazardous organic solvents with SCCO2.
The Use of Supercritical Fluids for Waste Minimization (Louisiana State University)
This research will evaluate the practicality of SCCO2 to reclaim industrial organic materials by a
non-destructive extraction process. Drs. K.M. Dooley and F.C. Knopf will utilize LSU's bench-and pilot-
scale units to find the best conditions for specific extractions, and to implement new operating
strategies for supercritical fluid extractions applicable to process streams. A countercurrent bubble
column mode which can be directly scaled up to industrial practice will be operated on difficult-to-extract
materials. The types of compounds to be tested include aqueous streams of alcohols, nitriles, and
nitroaromiatics; soils contaminated with polynuclear aromatics; and slurries of soils containing the above
classes of compounds.
COMMERCIALIZATION OF SUPERCRITICAL CARBON DIOXIDE
Industrial chemists and engineers should begin to consider the potential uses of supercritical
fluids as less hazardous solvents in chemical separation/synthesis applications. The successful
commercial application of supercritical carbon dioxide technology for pollution prevention and waste
minimization in organic solvent-using operations will help U.S. companies become more competitive by
improving their manufacturing process efficiencies, reducing waste and associated disposal costs, and
enhancing environmental compliance.
The EPA will assist its partners in this program with technical transfer of reports, presentations,
and publications. Formation of partnerships between the technology developers and industry will also
aid commercialization. Another possibility under consideration is a future EPA workshop to highlight the
results of the supercritical fluid research under the EPA program with the inclusion of similar efforts
conducted under private or other public sponsorship.
REFERENCES
1. Brady, B.O., Dooley, K.M., Knopf, E.C., and Gambrell, R.P., Ind. Eng. Chem. Res.,
26,26(1987).
2. Beckman, E.J., Ataai, M.M., Extraction of Mercury, Arsenic, and Lead from Soils Using
Supercritical Carbon Dioxide and Novel, CO2-Philic Chelating Agents, 87th Annual
Meeting, A&WMA, Cincinnati, Ohio (1994).
37
-------
DEVELOPMENT OF HIGH-RATE TRICKLE BED BIOFILTER
Francis L. Smith, George A. Sorial, Makram T. Suidan, and Pratim Biswas
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, OH 45221-0071
Richard C. Brenner
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
INTRODUCTION
Since enactment of the 1990 amendments to the Clean Air Act, the control and removal of
volatile organic compounds (VOCs) from contaminated air streams have become a major public
concern (1). Consequently, considerable inte'rest has evolved in developing more economical
technologies for cleaning contaminated air streams, especially dilute air streams. Biofiltration has
emerged as a practical air pollution control (APC) technology for VOC removal. In fact, biofiltration
can be a cost-effective alternative to the more traditional technologies, such as carbon adsorption
and incineration, for removal of low levels of VOCs in large air streams (2). Such cost effectiveness
is the consequence of a combination of low energy requirements and microbial oxidation of the VOCs
at ambient conditions.
Preliminary investigations (3) were performed on three media: a proprietary compost mixture;
a synthetic, monolithic, straight-channeled (channelized) medium; and a synthetic, randomly packed,
pelletized medium. These media were selected to offer a wide range of microbial environments and
attachment surfaces and different air/water contacting geometries. The results of this preliminary
work demonstrated that 95+% VOC removal efficiency could be sustained by all three media at a
toluene loading of 0.725 kg COD/m3-d, but at different empty bed contact times (EBCTs). For the
pelletized medium, this performance could be achieved at an EBCT of 1 min, for the channelized
medium at 4 min, and for the compost medium at 8 min. Both synthetic media developed headless
over time, with the pelletized medium showing a pressure drop in excess of several feet of water
after sustained, continuous operation. These results left open the question of which medium could
provide the optimum combination of high VOC elimination efficiency at high loading with minimum
pressure drop.
This paper discusses the continuing research being performed for development of biofiltration
as an efficient, reliable, and cost-effective VOC APC technology. The objectives of the recent
research were to conclude the evaluation of the three media and to develop workable strategies for
the removal and control of excess biomass from the (ultimately) selected pelletized attachment
medium.
METHODOLOGY
The biofilter apparatus used in this study consisted of three, independent, parallel biofilter
trains, each containing 4 ft of attachment medium: biofilters "A", "B", and "C". A detailed schematic
and equipment description is given elsewhere (4). Biofilter "A" was filled with a proprietary compost
mixture, "B" with a Corning Celcor® channelized medium, and "C" with a Manville Celite® pelletized
medium. Biofilters "A" and "B" were square and had an inner side length of 5.75 in., and biofilter "C"
was round with an inside diameter of 5.75 in. The air supplied to each biofilter was highly purified for
complete removal of oil, water, CO2, VOCs, and particulates. After purification, the air flow for each
38
-------
biofilter was split off, the VOCs were injected into it, and then it was humidified and fed to the
biofilters. The air feed was mass flow controlled, and the VOCs were metered by syringe pumps.
The flow direction of the air and nutrients inside each biofilter was downward. Each biofilter was
insulated and independently temperature controlled.
Buffered nutrient solutions were fed to biofilters "B" and "C". A detailed description of the
nutrient composition is given elsewhere (4). Each of these biofilters independently received a
nutrient solution containing all the necessary macro- and micro-nutrients with a sodium bicarbonate
buffer. The nutrients required in biofilter "A" were included as part of the original compost mixture.
RESULTS
Biofilter "A": This biofilter run on the compost medium was made to evaluate the effect of
temperature, and then loading, on toluene removal efficiency. Figures 1a and 1b summarize the
biofilter performance. The biofilter was started up and, after some operational difficulties, stabilized
by day 10 at 52°F, 50 ppmv toluene, 2 min EBCT, and a removal efficiency of about 58%. On day
17, the temperature was raised to 60°F, resulting in a rise in efficiency to about 75%, which
decreased after day 24 into the 60's, and after day 32 into the 50's. On day 41, the temperature was
increased to 70°F, resulting in a gradual increase in efficiency to about 75% by day 47. On day 53,
the temperature was increased to 80°F, resulting in an increase in efficiency into the low 80's. On
day 61, the temperature was increased to 90°F, resulting in a further increase in efficiency to the mid
90's (Figure 1a). After day 77, the feed was increased slowly to about 95 ppmv toluene, resulting in
a drop in efficiency to about 88%. Further increases in the feed concentration to a maximum of 180
ppmv toluene on day 139 resulted in a further decline in efficiency to about 58% (Figure 1b). The
run was terminated on day 215.
Biofilter "B": This biofilter run was made on the synthetic channelized medium to evaluate
the effect of temperature, and then nutrient feed rate, on removal efficiency. The biomass in the
channels of the medium remaining from the previous run was removed by hydroblastjng the eight 6-
in. high medium blocks from top and bottom. The corners of these square blocks were filled with
grout to provide a "round" active block. This last step was taken to match a round block cross
section with the round pattern of the nutrient delivery spray nozzle. Figure 2 shows the biofilter
performance as a function of time. The biofilter was started up at 52°F, 50 ppmv toluene, .and 2 min
EBCT. By day 36, the removal efficiency had drifted over a range from about 62% to 80%. On day
36, the nutrient feed rate was increased from 30 to 60 L/day while keeping the mass loading of the
nutrients constant. The increased nutrient flow rate effectively doubled the wetting cycle from 20
sec/min to 40 sec/min. An immediate increase in efficiency to 99% was observed, which then quickly
dropped and ranged by day 50 between about 30% and 70%. On day 50, the nutrient feed rate was
increased to 90 L/day (increasing the wetting cycle to 60 sec/min), but the efficiency dropped from
69% and ranged by day 67 from about 22% to 65%. On day 67, the temperature was raised from
52°F to 60°F and the efficiency increased to 66%. By day 75, the efficiency was 87% and this level
was maintained to day 83. After day 83, the temperature was raised further, in 10°F steps, to 90°F
but the efficiency did not improve. In fact, for the rest of the run, at 90°F and 60 L/day, the efficiency
ranged between about 58% and 83%. The run was terminated on day 152.
Biofilter "C": The first biofilter run on the synthetic pelletized medium was made to evaluate
the effect of pressure drop, and then temperature, on toluene removal efficiency. The biofilter was
charged with pellets used in the previous run. These pellets were washed by hand in hot water
(150°F) until the accumulated surface biomass had been removed and the pellets were free flowing.
Figure 3 presents the biofilter performance as a function of time. The biofilter was started up at 52°F,
50 ppmv toluene, and 2 minutes EBCT. By day 21, the removal efficiency was 99% and, by day 27,
it had reached 100% and remained at this level until day 50. From day 51 to day 57, the EBCT was
gradually reduced to 1 minute, causing the efficiency to drop, to 84%. Subsequently, the toluene
removal efficiency rapidly increased to the low 90's and remained in that range until day 81. On day
39
-------
o
a
l>
K
a
V
100
90
80
70
Toluene Loading
0.45 leg COD/mSday
. EBRT = 2 minutes
50 -
50
60
70
80
90
100
Temperature, F
Figure la. Effect of Temperature on the Performance of the Compost Biofilter
100
90
"S 80
r>
O
a
V
v
a
V
70
60
50
Temperature — 90 F
EBRT - 2 minutes
0.0
x
0.4 '
• ' * • ' ._f ' ' - ' - -«
0.6
0.6
1.0
1.2
1.4 1.6
Toluene Loading, kg COD/day m
1.8
2.0
Figure Ib. Effect of Toluene Loading on the Performance of the Compost Biofilter
40
-------
100
100
80
ft
04
*
a
o
O
a
o
o
'oluene Loading
3
.44 kg COD/m day
80 80 100 120 140 160
Sequential Date, days
Figure 2. Performance of Channelized Biofilter with
Respect to Toluene Removal at an EBRT of
2 Minutes
a
«•*
o
o
a
o
Pi
41
-------
82, the temperature was raised to 60°F and the efficiency steadily rose until complete biodegradation
of the toluene was reached on day 89. This essentially 100% efficiency in toluene removal was
maintained through day 97. During the period between day 54 and day 97, pressure drop across the
system increased from 0.2 to 5.5 in. water. From day 97 to day 111, the efficiency dropped steadily
from 100% to 86% while the pressure drop increased from 5.5 to 6.0 in. water. On day 112, the
temperature was increased to 70°F and the efficiency rebounded by day 113 to a peak value of 97%,
after which it dropped to 85% by day 188. On day 119, the temperature was raised to 80°F and the
efficiency rose to about 89% by day 120. During the period from day 112 to 120, the pressure drop
increased from 6 to 18 in. water. By day 128, the efficiency had steadily dropped from 89% to 77%
as the pressure drop increased from 18 to 27 in. water. This pattern of a steady loss of efficiency
with a coincident increase in pressure drop suggested the development of short circuiting within the
biofilter medium due to biomass accumulation, which resulted in a significant reduction in actual
contact time.The first run was terminated on day 128.
The second biofilter run on this medium was conducted to evaluate routine biomass control
by backwashing. The biofilter was charged with a 50:50 mixture of fresh pellets and pellets from the
previous run. The used pellets were thoroughly washed by hand in tepid water (90°F) until the
accumulated surface biomass had been removed and the pellets were free flowing. Figure 4 shows
the biofilter performance as a function of time. The filter was started up at 90°F, 50 ppmv toluene,
and 2 min EBCT. By day 4, the removal efficiency was 100%. (Note: this second run, started up
with pellets washed in tepid water, contrasts with the slower start-up in the first run, where the pellets
were washed with hot water.) On day 8, the feed was increased to 250 ppmv toluene and the
efficiency dropped to 97%, ranged between 92% and 98% until day 25, when it again reached 99%.
Subsequently, the efficiency dropped as low as 86% before regaining 99% on day 81, after which the
efficiency was nearly always 99+%. Initially, backwashing was performed once per week by using
100 L of fresh water at a rate of 6 gpm. After day 28, the frequency was increased to twice per
week and, after day 38, the volume was increased to 200 L. These changes were made because
measurable pressure drop was observed between backwashings. On day 73, the backwash rate
was increased to 15 gpm in order to induce full fluidization. Although the pressure drop increase was
minimal, the efficiency did not improve, suggesting some form of channelizing within the bed.
Therefore, on day 80, the length of the backwash period was increased to 1 hour by recirculating the
backwash water. After this final adjustment, the toluene removal efficiency, as mentioned above,
achieved and sustained 99+%. During this latter period, the total volume of water used per
backwash was optimized to 120 L. Of this volume, 70 L were used for the 1-hour backwash recycle,
while the remaining 50 L were used to flush the released solids from the reactor.
CONCLUSIONS
A marked improvement in toluene removal efficiency with increasing temperature was
demonstrated in this study for the compost mixture, the channelized medium, and the pelletized
medium. The direct consequence of this finding is that much less medium would be needed for a
biofilter operating at 90°F than at 52°F, resulting in a proportional reduction in capital cost. The
economic tradeoff with the cost of heating the incoming air should usually favor operation at these
warmer conditions.
The modest performance of the compost mixture with respect to increased loading
complimented our earlier findings with respect to decreasing EBRT (3). Unfortunately, implicit
limitations of the experimental apparatus may have resulted in reduced performance. Specifically,
the manufacturer recommended using a width-to-depth ratio of 1/1, rather than 1/8. They also stated
that from their experience the only effective means of controlling bed moisture content was to weigh
the entire biofilter. This was impossible with the heavy stainless steel unit used here, which was
bolted to a support frame. Several moisture measurement and control strategies were attempted, but
it was never possible to be certain that the bed moisture content was consistently at the reported
optimum range, i.e., between about 50% and 60% (5,6). The sometimes erratic performance may
42
-------
>»
o
d
v
w
"a
o
V
CA
Toluene Loading kg COD/m day
0.92
40 60 80 100
Sequential Date, days
120
140
Figure 3. Performance of Pelletized Biofilter with Respect to Toluene
Removal at 1 and 2 Minutes EBRT without Backwashing
ca
o
a
o
K
100
90
80
70
60
50
40
30
20
10
Efficiency
Toluene Loading kg COD/m day
2.27
20
15
a
10
SH
Q
a>
(-
3
w
to
-------
have been influenced by variations in bed moisture content. However, it can be seen that the best
removal efficiencies achieved by the compost mixture were better than exhibited by the channelized
media but worse than exhibited by the pelletized media.
The performance of the channelized medium also confirmed our earlier findings that this
medium is distinctly inferior to the pelletized medium (3). The best performance was achieved during
the use of new medium blocks. After biomass accumulation within the channels and subsequent
removal by hosing, the performance never regained the previous, still modest, levels. Attempts to
adjust nutrient flow as a means of testing the effect of the duration of the wetting in the nutrient
application cycle did not overcome the previously demonstrated efficiency limitations. The more
erratic performance of this medium after removal of the biomass suggests that this medium may be
unsuitable for sustained efficiency after periodic cycles of biomass removal. This erratic
performance, due to suspected random uneven plugging of channels by biomass, coupled with its
relatively low overall removal efficiency, difficulty in biomass removal, and intrinsicly high medium
cost suggests that this medium may not be a viable option for this application.
The pelletized medium exhibited the best and most consistent performance of the three
media tested. It rapidly achieved high removal efficiencies at high toluene loadings. As the first run
demonstrated, however, an excessive accumulation of biomass, shown by a rise in the pressure drop
across the medium, results in a substantial loss in efficiency, followed by a very rapid rise in pressure
drop. This suggested that efficient, sustained performance might be achieved through early and
periodic control of biomass accumulation by backwashing. In the second run, the implementation of
a suitable backwashing strategy for biomass control was achieved by using full medium fluidization.
This strategy permitted sustained operation of the biofilter at high loadings with efficiencies
consistently at 99+%.
REFERENCES
1. Lee, B. Highlights of the Clean Air Act Amendments of 1990. J.Air Waste Manage. Assoc.
41(1):16, 1991.
2. Ottengraf, S.P.P. Exhaust Gas Purification, in: H.J. Rehn and G. Reed (eds.), Biotechnology.
Vol. 8. VCH Verlagsgesellschaft. Weinham, 1986.
3. Serial, G.A., Smith, F.L., Smith, P.J., Suidan, M.T., Biswas, P., and Brenner, R.C. Evaluation
of Biofilter Media for Treatment of Air Streams Containing VOCs. ]n: Proceedings of the
Water Environment Federation 66th Annual Conference and Exposition, Paper No. AC93-
070-002. Anaheim, California, 1993. pp. 429-439.
4. Sorial, G.A., Smith, F.L, Smith, P.J., Suidan, M.T., Biswas, P., and Brenner, R.C.
Development of Aerobic Biofilter Design Criteria for Treating VOCs, Paper No. 93-TP-52A.04.
Paper presented at 86th Annual Meeting and Exhibition of Air and Waste Management
Association, Denver, Colorado, June 13-18, 1993.
5. Bohn, H.L. Biofiltration: Design Principles and Pitfalls, Paper No. 93-TP-52A.01. Paper
presented at the 86th Annual Meeting and Exhibition of Air and Waste Management
Association, Denver, Colorado, June 13-18, 1993.
6. Van Lith, C., David, S.L., and March, R. Design Criteria for Biofilters. Trans. Inst. Chem. Enq,
686:127-132, 1990.
FOR MORE INFORMATION
Richard C. Brenner
Risk Reduction Engineering Section
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
513/569-7657
44
-------
INTEGRATED SYSTEM FOR TREATING SOIL
CONTAMINATED WITH WOOD TREATING WASTES
Carolyn M. Acheson and Richard C. Brenner
U. S. EPA Risk Reduction Engineering Laboratory, Cincinnati, OH 45268
Amid P. Khodadoust, Gregory W. Wilson, Karen M. Miller,
and Makram T. Suidan • . , .
University of Cincinnati, Department of Civil and Environmental Engineering
•'..-.,• , Cincinnati, OH 45221-0071
INTRODUCTION
Approximately 20% of the hazardous waste sites undergoing
bioremediation are contaminated with wood treating wastes, primarily
compounds such as pentachlorophenol (PCP), creosote, polycyclic aromatic
hydrocarbons (PAHs), and other hydrocarbons (1). A process that combines
soil washing with sequential anaerobic and aerobic biotreatment is being
integrated to remediate soil contaminated with these wood treating wastes.
By extracting the target compound from the soil, soil washing facilitates
degradation by mobilizing the target compound and expanding the range of
feasible remediation technologies. Additional flexibility is possible
since soil washing can be conducted in an in-situ or ex-situ format. In
this process, the wash solution is initially bioremediated in an anaerobic
environment. Mineralization of the target compound is completed
aerobically. Based on preliminary, results, the integrated process could
meet the target cleanup level for PCP in approximately 45% of the
bioremediation sites (1).
Process development began by independently evaluating soil washing
and target compound degradation. PCP contaminated soils were the initial
focus, but this work is currently being extended to include soils
contaminated with both PCP and PAHs. In addition, based on promising
results from the soil washing and degradation evaluations, these individual
unit operations are being integrated to form a complete process to
remediate soils contaminated with wood treating wastes. This complete
process incorporates soil washing, soil wash solution recycling, and
biodegradation of the target compounds and is outlined in Figure 1.
RESULTS
Soil Washing ........
An ethanol and water mixture (mass fractions with soil: 0.02 soil,
0.49 ethanol, 0.49 water) has been selected as the soil washing solution.
Compared to other ethanol-water mixtures, this mixture'•achieves the
highest removal efficiency, 65% to 100% depending on soil size^ > •
contaminated age, and soil washing format. Comparing ethanol-water soil
washing to the analytical procedures of soil sonication and soxhlet
extraction, soil washing removes between 93% and 114% of the PCP removed
by these analytical methods (2). With the eguimass mixture of ethanol and
water, essentially complete extraction is achieved in a 30 minute contact
time. The ability of the ethanol-water mixture to remove PAHs from soil
is being evaluated using naphthalene, acenaphthalene, pyrene, and
benzo(b)fluoranthene as model compounds.
45
-------
Wash Solution Recycle
E, P,W
Wash Feed
E, W
4 Soil Feed
P, S(F)
.
Prefilter
Clarifier
Overflow
E, F.P.W
E, P, W
R O Filter
RO Retentate
E, P,W
Nutrient
Feed
Clarifier
Bottoms Water
Feed
W
E, P, S, W
Prefilter
Retentate
E, F, P, W
Filtrate
Outlet
E, P, W
1
i
Soil
Washer
2
E, P, S, W
Dewaterer
Clean Soil
E, P, S(F), W
Caas
Anaerobic
Expanded
Bed
GAG
Reactor
Nutrient
Feed ^
fc
B, MCP, DCP,
TCP, P
Aerobic
Expanded
Bed
GAG
Reactor
Outlet Gas_
Outlet Liquid _
B,W
Figure 1. Integrated Process for Remediating Soil Contaminated with Wood
Treating Wastes. The diagram details the flow of the
following components: Biomass (B), Ethanol (E), Soil Fines
(F), PCP (P), Soil (S), and Water (W).
46
-------
Recycling Wash Solution
To reduce costs and the volume of PCP bearing liquid, the soil wash
liquid will be concentrated and the removed ethanol and water will be
recycled to the first soil washing unit. Reverse osmosis (RO) has been
selected as the wash solution recycle operation to avoid the high energy
costs of distillation. The feasibility of reverse osmosis will be
determined in large part by the amount of wash solution which can be
recycled and the economics of this process step.
Biodegradation
The retentate from the wash solution recycle process, the more
concentrated wash solution, will be fed to an anaerobic expanded-bed
granular activated carbon (GAG) reactor. The expanded-bed GAC reactor
consists of a 96.5 cm Plexiglas tube with an internal diameter of 10.2 cm
and an influent and effluent header. The reactor volume is 10 L including
the recycle loop. Two of these reactors have been operated for over 40
months to evaluate variables such as PCP loading and retention time. When
operated at retention times of 1.2 and 2.3 hours, greater than 99.97% of
the influent PCP (0.6 g/day) is dechlorinated to monochlorophenol (MCP).
The outlet stream from the reactor operating with the 2.3 hour retention
time contains: 0.55 mmol MCP/L, 0.02 mmol dichlorophenol (DCP)/L, 0.005
mmol phenol/L, 0.0008 mmol PCP/L, and 0.00035 mmol trichlorophenol
(TCP)/L. Inhibition of the microbial cultures in these reactors,
presumably by elevated levels of MCP, was alleviated by decreasing the PCP
feed rate to the current level of 0.6 g/day. Ethanol is also included in
the reactor feed stream at a level of 33.3 g/day; however, the ethanol
feed in the integrated process will be determined by the wash solution
recycle process unit.
The outlet stream from the anaerobic expanded-bed GAC reactor will
be fed to an aerobic expanded-bed GAC reactor. Under aerobic conditions,
the remaining chlorinated compounds should be mineralized (3). Since
analysis of the anaerobic reactor is nearly complete, characterization and
optimization of the aerobic reactor will begin shortly.
In addition, experiments to characterize an anaerobic expanded-bed
GAC reactor fed PCP and the four model PAHs have begun. An 11 L slurry-
type reactor is also being evaluated. These reactors are also feed
ethanol as a primary substrate.
CONCLUSIONS
A process integrating soil washing and sequential anaerobic-aerobic
biodegradation may be an effective method of remediating wood treating
wastes. The process meets the target cleanup level in 45% of the PCP
bioremediation sites, or less than 35 mg PCP/kg soil. Soil washing with a
mixture of ethanol and water removes PCP at levels comparable to those
achieved through the analytical techniques of soil sonication and soxhlet
extraction. Following soil washing, some of the soil washing solution
will be recycled. The concentrated PCP bearing liquid is fed to an
anaerobic expanded-bed GAC reactor where 99.97% of the PCP is reduced to
MCP. The outlet from the anaerobic reactor feeds an aerobic expanded-bed
reactor that should mineralize the remaining chlorinated phenols.
Current efforts are directed towards adapting the process to soils
contaminated with both PCP and PAHs, evaluating reverse osmosis as a means
of recycling the soil washing solution, and integrating the process.
47
-------
REFERENCES
1. Bioremediation in the Field No. 11. EPA/540/N-94/501, U. S.
Environmental Protection Agency, 1994. 63 pp.
2. Khodadoust, A. P.; Wagner, J. A.; Suidan, M. T.; Safferman, S. I.
Solvent Washing of Aged PGP-Contaminated Soils. Paper presented in
1993 at the Water Environment Federation, Anaheim, California.
October, 1993.
3. Rochkind, M. L. and Blackburn, J. W. Microbial Decomposition of
Chlorinated Aromatic Compounds. EPA/600/2-86/090, U. S.
Environmental Protection Agency, Cincinnati, Ohio, 1986. 269 pp.
FOR MORE INFORMATION
Carolyn M. Acheson
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
(513) 569-7190
48
-------
DEVELOPMENT OF A MULTI-LEVEL PROTOCOL FOR DETERMINATION OF BIODEGRADATION
KINETICS AND BIOAVAILABILITY OF ORGANIC COMPOUNDS TO ENHANCE IN-SITU
BIOREMEDIATION
Henry H. Tabak
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
Rakesh Govind, C. Fu, X. Yan, S. Pfanstiel, and C. Gao
Department of Chemical Engineering
University of Cincinnati
Cincinnati, Ohio 45221-0171
INTRODUCTION
Knowledge of biodegradation kinetics in soil is needed to understand the efficacy of in-situ and ex-situ
bioremediation technologies. Laboratory studies to determine biodegradation rates can be used as screening tests
to determine the rate and extent of bioremediation that might be attained during remediation, and to provide design
criteria. (Graves et al, 1991; (1) Tabak et al, 1992; (2) Govind et al, 1993; (3) Tabak et al, 1994, 1995) (4,5)
Traditionally, in-situ biodegradation kinetics have been determined using soil microcosms (6,7) which are difficult
to model mathematically. Laboratory studies involving soil slurry reactors have been reported by Bachmann et al.
(8) Kaplan and Kaplan (9), Milhelcic and Luthy (10) and Brunner et al. (11). Currently there is no systematic
methodology to quantitatively determine biodegradation kinetics of contaminants in compacted soil systems.
Biodegradation in soil is a fairly complex process which involves diffusion of contaminants in the porous
soil matrix, adsorption to the soil surface, biodegradation in the biofilms existing on the soil particle surface and
in the large pores as well as in the bound and free water phase after desorption from the soil surface. In soil slurry
reactors, biodegradation of contaminant occurs both in the liquid phase by soil microorganisms desorbed from the
soil matrix and by the biofilms immobilized on the soil particle surface. In compacted soil systems, biodegradation
occurs in the free and bound water phase primarily by the soil immobilized microorganisms and the contribution
of water suspended microbiota is small due to low water content. In this paper a systematic protocol based on three
types of bioreactors and one representative contaminant, phenol, has been developed to determine the biokinetic
parameters of the suspended and immobilized microbiota and the transport parameters of contaminant and oxygen
in the soil matrix.
EXPERIMENTAL PROCEDURES AND MATERIALS
In this study, three types of bioreactors, shown in Figure 1, to determine the biokinetic parameters of the
suspended and immobilized microbiota were: (1) slurry bioreactor, where soil at 5% slurry concentration is
vigorously mixed with the contaminant, dissolved in water with nutrients; (2) wafer reactor, where a thin wafer of
soil is spiked with contaminant and nutrients dissolved in water, to obtain a 50% total soil moisture content; and
(3) porous glass tube reactor, where sieved soil with contaminant is packed in a porous glass tube with moisture
content identical to the wafer reactor.
In the soil slurry reactor there are no limitations of oxygen, which freely diffuses into the well-stirred
slurry, nutrients, which are initially dissolved in the aqueous phase and water. Hence, the biodegradation rate in
soil slurry reactors depends on the intrinsic biokinetic rate, microorganism concentration in the soil matrix and
inherent diffusivity of the contaminant. In the soil wafer reactor, oxygen diffuses freely through the thin soil matrix,
and hence biodegradation rate is controlled by the water content in addition to the other intrinsic parameters, as in
49
-------
the case of the soil slurry reactor. In the porous tube reactor, biodegradation rate is controlled by the water content
and oxygen diffusivity and other intrinsic biokinetic parameters. The porous tube reactor provides a better estimate
of biodegradation rates for in-situ bioremediation than the soil wafer and soil slurry reactors.
Biodegradation rates of phenol in the three soil reactor systems were determined by electrolytic
respirometry (Voith Respirometer) used to continuously measure the cumulative oxygen uptake and to simultaneously
monitor the carbon dioxide evolution. Appropriate concentrations of phenol stock solution in distilled deionized
(DD) water were added to the reactor system containing Organization for Economic Co-operation and Development
(OECD) nutrient solution composed of inorganic salts, trace salts, vitamin solution and/or yeast extract solution and
different soil concentrations. Stock solution containing OECD nutrients with no phenol were used for control
reactor flasks. In addition, radiolabelled contaminant was used to confirm the oxygen uptake from the contaminant
rather than due to normal soil respiration.
The soil was characterized to determine soil moisture content, clay, sand and silt content, percent organic
matter, soil pH, cation exchange capacity, bulk density, and nutrients in soil. The soil selected for the study was
uncontaminated silt loam soil with the following characteristics: soil moisture 17%, organic matter 2.9%,
classification silt loam, cation exchange capacity 6.5, soil pH 6.1, bulk density 1,06, nutrients in soil (ppm)
phosphorus 17, potassium 90, magnesium 80, calcium 1100 and sodium 17. Table 1 summarizes the soil
characterization data. The Brunauer, Emmett and Teller (BET) specific surface area was 20.27 m2/g, BET void
volume was 01029 cm3/g and the BET average pore diameter was 58A. The soil porosity was 6.53 %. Soil particle
size distribution was determined using outline by Day (1965) (12). Soil porosity, pore size distribution, pore volume
and surface area was determined by nitrogen adsorption using Micrometrics ASAP 200.
Abiotic adsorption and desorption kinetics of the contaminant into the soil matrix and oxygen uptake data
obtained for the soil slurry, soil wafer and porous tube reactors are used in conjunction with detailed mathematical
models to derive the intrinsic biokinetic and transport parameters. To determine the adsorption and desorption
characteristics, the soil was air dried and passed through 2 mm sieve. Abiotic adsorption and desorption equilibria
were quantified using the Freundlich isotherm equation. A systematic protocol for conducting the abiotic adsorption
and desorption kinetics was initially developed using soil slurry systems.
Soil Slurry Experiments
To simulate biodegradation of substrate in soil slurry bioreactors (1) 20 g of soil was mixed with 250 mL
of DD water containing OECD nutrients and phenol at different concentration levels; (2) 10 mL of KOH solution
(pH 9.5) was injected in carbon dioxide trap; (3) Reactor vessel was closed and cumulative continuous oxygen
uptake measurements were made using respirometer. Reactor vessels were stirred continuously with teflon coated
stirring bars. (4) KOH solution was withdrawn from CO2 trap and the amount of CO2 evolved, was quantified by
measuring the change in pH of KOH solutions. (5) In radiorespirometric studies both gaseous and aqueous 14 CO2
and radiolabelled intermediates were analyzed by Packard scintillation counting apparatus.
Soil Wafer Experiments
To more closely simulate the actual biodegradation of phenol in intact soils, soil wafer studies were
developed for respirometer experimentation. The wafer system consisted of a thin layer (or wafer) of soil with a
known moisture content. Procedures for the wafer experiments were as follows: (1) 20.0 g of soil and measured
amount of water, approximately 20 to 30 mL in volume were placed in reaction flask, well mixed to give unifonn
initial biomass concentration, and the entire flask was weighed; (2) the flask(s) were placed in fume hood overnight
and re-weighed. The weighing procedure was repeated until weight of flask indicated that the soil contained the
desired moisture content; (3) the soil wafer was contaminated with 2.5 to 10.0 mL of experimental stock solution,
depending on the desired concentration, using a syringe. Steps 3 thru 5 for the slurry experiments were followed
without stirring the soil wafer.
50
-------
Microporous Tube Experiments
Microporous tube reactors were developed for use in electrolytic respirometer to study the effects of oxygen
diffusivity. The porous glass tubes used were purchased from Corning (Corning, NY) and were made from vycor
glass with pores averaging 40 Angstroms in diameter. This pore size was selected because it was found to be
optimum for containing the soil and water within the tube and allow free flow of oxygen into the soil. The porous
glass tube pore sizes exceeding 40 A° were found not to effect the oxygen uptake results. The experimental
procedure was as follows: (1) the microporous glass tubes were placed in the laboratory oven set at 200°C and left
overnight to evaporate any water or contaminant in the pores or on the surface of the tubes; (2) the tubes were
removed from oven and allowed to cool for several hours to room temperature. The tubes were placed in a beaker
containing DD water to initially saturate the pores in the glass; (3) the tubes were removed from the water and filled
with 20.0 g of soil, plugging both ends with glass wool; (4) the soil was contaminated evenly in the tube with 1.25
to 5.0 mL of experimental stock solution by inserting a syringe through the plugs at either end and mixing the soil
around with the syringe as much as possible while injecting the solution; (5) two tubes were placed into each flask
and procedures according to steps 3 through 5 for the slurry experiments were followed.
Adsorption and Desorption Experiments
For adsorption experiments, 50 g soil sample were placed in a 250 mL glass bottle. Various concentrations
of each compound, namely 10, 25, 50, 100 and 150 mg/L were used. The volumes of the total solution added to
each bottle were 250 mL in order to maintain a minimum head space. The head space was important, as it not only
affected the degree of mixing during stirring but also controlled loss of phenol due to volatilization. Blanks
containing only the phenol solution allowed the measurement of volatilization losses. In order to prevent phenol
degradation in solution, 1 mL HgCl2 saturated solution was added to each bottle. All adsorption experiments were
conducted in a fume hood, so that the temperature of adsorption could be controlled at 24°C.
The mixture in the bottle was stirred using a magnetic stirrer for 96 hours. The samples at 2, 4, 6, 8, 12,
24, 36, 48, 72, and 96 hours from the beginning of the adsorption experiment were filtered using 0.45 /xm silver
membrane and placed in a 50 mL sample vial for extraction. The extraction method was based on U.S. EPA
Method 604 & 610.
The GC used was a HP-5890A model equipped with a flame ionization detector. The following conditions
were used throughout this study; initial oven temperature 60°C, initial hold time 5 minutes oven temperature rate
8°C/min., final oven temperature 280°C, final hold time 5 minutes, injector temperature 225°C, detector
temperature 310°C, makeup gas (nitrogen) flowrate, 35 mL/min, detector gas flowrate 32 mL/min. hydrogen and
435 mL/min. air, carrier gas (helium) 2 mL/min., column HP-5 methyl silicon gum and 5m x 1.53mm x 2,54mm
film thickness, and software: HP Chemstation.
After adsorption equilibrium was attained, the solution in a 250 mL glass bottle was mixed with 250 mL
distilled deionized water in a large 500 mL glass bottle to start the desorption experiment. AH the conditions for
the desorption experiment were the same as for adsorption experiment. Samples for quantifying desorption were
treated the same way as for adsorption studies, although the time taken to achieve desorption equilibrium was
significantly longer. ;
RESULTS AND DISCUSSION
The Freundlich isotherm adsorption parameters (Ka and 1/n) and desorption parameters (Kd and 1/h) as
well as the values for porosity of adsorbent (e) and density of adsorbent (p) are listed in Table 2. Phenol adsorption
equilibrium isotherm was highly non-linear compared to the desorption equilibrium isotherm and the extent of
adsorption was significantly higher than the desorption extent. This suggested that significant amounts of phenol
remained irreversibly bound to the soil matrix.
51
-------
Figure 2 shows the oxygen uptake data for the slurry, wafer and porous tube reactors for lOOmg/L phenol
concentration. Clearly, the oxygen uptake curve for the slurry reactor reaches a higher plateau than the curves for
the soil wafer and porous tube reactors, indicating that more phenol is being degraded in the slurry reactor. The
soil wafer reactor degrades more phenol than the porous tube reactor. Furthermore, the slurry reactor data attain.
a plateau value faster than the wafer and porous tube oxygen uptake data. This shows that biodegradation rates in
soil slurry reactors are the highest since there are no limitations of oxygen, nutrients and water and biodegradation
occurs both in the liquid phase by the suspended microorganisms and by the soil immobilized biofilms. In the soil
wafer reactor, there is no oxygen limitation and the biodegradation rate is governed by the contaminant desorption
and subsequent degradation in the free and bound water phase by the soil immobilized biofilms. Since the water
content of the soil in the wafer reactor is significantly less than in the slurry reactor, the biodegradation rate is also
lower than in the slurry reactor. In the porous tube reactor, in addition to the limited water content in the soil, as
in the case of the soil wafer reactor, oxygen diffusion in the soil matrix is also limited. This limits phenol
degradation only in the outer region of the tube and phenol present in the interior of the tube does not biodegrade
due to unavailability of oxygen.
This demonstrates, that in-situ bioremediation rates are significantly lower than biodegradation rates
achievable in soil slurry reactors due to limited water content and oxygen diffusivity. Furthermore, nutrient
limitations may further limit bioremediation rates in contaminated soil sites. While bioventing approaches may
maximize availability of oxygen, delivery of water and nutrients are still major limitations for maximizing in-situ
bioremediation rates.
Detailed mathematical models were developed for analyzing the oxygen uptake data from the soil slurry,
wafer and porous tube reactors. It was found that in the soil slurry reactor, significant degradation of contaminant
occurs in the aqueous phase by the suspended soil microorganisms rather than by the soil immobilized biofilms.
Biodegradation rates in soil wafer and porous tube reactors increase linearly with contaminant concentration and
active microbiota concentration. 81 % of the phenol added initially was biodegraded in the soil wafer reactor and
64% degraded in the porous tube reactor. As shown in Figure 2, the mathematical model fitted the experimental
data quite well. The soil slurry reactor data were used to derive the biokinetic parameters for the suspended and
immobilized microorganisms. These parameters when used with the appropriate amount of free water, were used
to fit the wafer reactor data. The wafer reactor data were used to obtain additional information with no oxygen
limitations and the porous tube reactor data provided quantitative estimation of oxygen diffusivity in the soil matrix.
The porous tube reactor data was used to derive the oxygen diffusivity in the soil phase.
Determining the oxygen profile in the porous tube soil using the model showed that the radial oxygen
concentration decreases rapidly attaining a zero value at a radial distance of 0.25R from the tube center, where R
is the radius of the porous tube. This again confirmed that there were oxygen limitations in the porous tube reactor
due to limited oxygen diffusion in the soil matrix.
In the soil slurry reactor there are no limitations of oxygen, which freely diffuses into the well-stirred
slurry. Hence, the biodegradation rate in soil slurry reactors depends on the intrinsic biokinetic rate, microorganism
concentration in the soil matrix and inherent diffusivity of the contaminant. In the soil wafer reactor, oxygen
diffuses freely through the thin soil matrix, and hence biodegradation rate is controlled by the water content in
addition to other intrinsic parameters, as in the case of the soil slurry reactor. In the porous tube reactor,
biodegradation rate is controlled by the water content in. addition to other intrinsic parameters, as in the case of the
soil slurry reactor. In the porous tube reactor, biodegradation rate is controlled by the water content and oxygen
diffusivity and other intrinsic biokinetic parameters. The porous tube reactor provides a better estimate of
biodegradation rates for in-situ bioremediation than the soil wafer and soil slurry reactors.
Experiments with uniformly labeled C1* phenol and measurement of carbon dioxide evolution by absorbing
the KOH solution and scintillation counting showed that the net oxygen uptake (actual uptake minus the oxygen
uptake in the control flask) was solely due to phenol degradation. This verified our initial assumption that the net
52
-------
cumulative oxygen uptake in each type of soil reactor could be used to derive the biokinetic and transport
parameters.
DESCRIPTION OF PROTOCOL
The overall protocol for evaluating soil biodegradation kinetics involves the following steps: (1) Measure
cumulative oxygen uptake in soil slurry reactors. The slurry reactor model is used to obtain the aqueous and soil
phase Monod biokinetic parameters. (2) Measure cumulative oxygen uptake in soil wafer reactor. Use the
biokinetic parameters determined from the slurry model and detailed wafer model to determine diffusivity of
contaminant in soil matrix. (3) Measure cumulative oxygen uptake in porous tube reactor. Use the parameters
determined from soil slurry and wafer models to calculate oxygen diffusivity in the slurry reactor which was also
the highest when compared to the other reactors. The soil slurry reactor oxygen consumption attains a plateau value
in about 1 day, while the wafer and tube reactor's cumulative oxygen uptake attains a plateau value in about 3.5
days.
The oxygen uptake data for the soil slurry, wafer and tube reactors were analyzed using a detailed
mathematical model to determine the best-fit values for the model parameters. The slurry reactor was used to
determine the biokinetic parameters for the suspended and soil immobilized microbiota in soil from porous tube
reactor model. (4) Measure radiolabelled carbon dioxide evolution using uniformly labeled compound hi soil slurry,
wafer and porous tube reactors to verify that oxygen uptake was due to compound mineralization. The biokinetic
parameters, compound and oxygen diffusivity in soil determined using the protocol can be used to model kinetics
of in-situ bioremediation.
REFERENCES
1. Graves, D.A., C.A. Lang and M.E. Leavitt. 1991. "Respirometric Analysis of the Biodegradation of
Organic Contaminants in Soil and Water." Applied Biochem. Biotech. 28/29: 813-826.
2. Tabak, H.H., C. Gao, L. Lai, X. Yan, S. Pfanstiel, I.S. Kim, and R. Govind. 1992. "Methodology for
Testing Biodegradability and Biodegradation Kinetics of Toxic Organics in Soil." Gas. Oil. Coal and
Environmental Biotechnology. V. IGT. Chicago. Illinois. (Eds.) C. Akin, R. Markushewski and J. Smith.
3. Govind, R., X. Yan, L. Lai, S. Pfanstiel, C. Gao, and H.H. Tabak. 1993. "Development of Methodology
to Determine the Bioavailability and Biodegradation Kinetics of Toxic Organic Pollutant Compounds in
Soil." In R.E. Hinchee, D.B. Anderson, F.B. Metting, Jr. and G.D. Sayles (Eds.) Applied Biotechnology
for Site Remediation, pp. 229-239, Lewis Publishers, Boca Raton, FL.
4. Tabak, H.H., C. Gao, L. Lai, X. Yan, S. Pfanstiel, I.S. Kim, and R. Govind. 1994. "Determination of
Bioavailability and Biodegradation Kinetics of Phenol and Alkyl Phenols in Soil." ACS Symposium Series
554 "Emerging Technologies in Hazardous Waste Management IV. (Eds.) D. William Tedder and F.G.
Pohland.
5. Tabak, H.H., C. Gao, L. Lai, X. Yan, S. Pfanstiel, C. Fu, and R. Govind. 1995. "Determination of
Bioavailability and Biodegradation Kinetics of Polycyclic Aromatic Hydrocarbons in Soils." ACS
Symposium Series. "Emerging Technologies in Hazardous Waste Management V. (Eds)
D. William Tedder and F.G. Pohland.
6. Chang, F. 1985. "Microbial Biodegradation of Crude Petroleum and Model Hydrocarbons in Simulated
Deep Soil Environment." ASM Abstracts. 85th ASM Annual Meeting. Las Vegas, NV., March 3-7,
1985. Q-50, 266,
53
-------
7. Michaels, G.B., J.P. Laplante and DJ. Schneck, 1988. "Biodegradation of Hydrocarbons by Indigenous
Microbial Populations in an Arid Soil Ecosystem." ASM Abstracts. 88th ASM Annual Meeting. Miami
Beach, FL. May 8-13, 1988 Q-136, 305.
8. Bachman, A., P. Walet, P. Wunen, W. de Brain, J.L.M. Huntiens, W. Roelofsen, and A.J.B. Zehnder.
1988. "Biodegradation of Alpha- and Beta-Hexachlorocyclohexane in a Soil Slurry under Different.Redox
Conditions." Appl. Environ. Microbiol. - 54. No. 1, 143-149.
9. Kaplan, D.L. and A.M. Kaplan. 1985. "Biodegradation of N-Nitrosodimethylamine in Aqueous and Soil
Systems." Appl. Environ. Microbiol. 50, No. 4, 1077-1086.
10. Mihelcic, J.R. and R.G. Luthy. 1988. "Microbial Degradation of Acenaphthene and Naphthalene under
Denitrification Conditions in Soil-Water Systems." Appl. Environ. Microbiol. 54. No. 5. 1188-1198.
11. Brunner, W., F.H. Sutherland and D.D. Focht. 1985. "Enhanced Biodegradation of Polychlorinated
Biphenyls in Soil by Analog Enrichment and Bacterial Inoculum." J. Environ. Oual. 14, No. 3, 324-328.
12. Day, P.R. 1965. "Particle Fractionation and Particle Size Analysis." In: C.A. Black (Ed) Methods of
Soil Analysis. 43. American Society of Agronomy, Madison, WI.
FOR MORE INFORMATION CONTACT:
Henry H. Tabak
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268
Phone: 513/569-7681
porous glass tube
packed soil
(a) Soil slurry reactor
(b) Soil wafer reactor
(c) Porous tube soil reactor
FIGURE 1 Schematic of Soil Slurry, Soil Wafer and Soil Porous
Tube Respirometric Reactors
54
-------
TABLE 1
Soil Characterization Data
Particle size
Oay(<2nm) 24%
Silt (7-44 urn) 58%
Sand(>44aai) 18%
Classification SUtLoun
Organic Muter 2.9%
Cation Exchange Capacity 6.5
SoilpH 6.1
Bulk density 1.06
1/3 bar moisture 30.7%
15 bar moisture 14.8%
Nutrients in soil Cppm)
Phosphorus 17 (h)
: Pota&am 90 (m)
Magnesium 800)
Calcium 1100(m)
Sodium 17 (v)
Note: h « high; m - medium; 1" tow, v - voy low
TABLE 2 Adsorption/Desorption Soil Equilibrium and Soil Porosity
and Density Parameters
K,
0.0105
1/n
0.84
K,
0.01259
1/n1
0.77*
e
0.06525
P,(g/L)
2250
R(mm)
0.334
Only for the slurry model
70
" Experimental
Experimental
* Experimental
— Model
40 50 60
Time (hr.)
FIGURE 2 Oxygen Uptake Curves for Soil Slurry, Wafer and Porous Tube
Reactors for Phenol. The line indicates the Fit by the
Mathematical Model Developed for each Reactor Type
55
-------
DESIGN AND TESTING OF AN EXPERIMENTAL IN-VESSEL COMPOSTING SYSTEM
Carl L. Potter and John A-. Glaser
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
Majid A. Dosani, Srinivas Krishnan, Timothy A. Deets, and E. Radha Krishnan
IT Corporation
11499 Chester Road
Cincinnati, OH 45246
INTRODUCTION
Composting has received much attention as a potential technology for treating solid
waste. Most of that attention has been focused on treatment of municipal solid waste,
sewage sludge, yard trimmings, and agricultural wastes. More recently, composting has been
investigated as a remediation technology for hazardous wastes. Laboratory and field-scale
work has been conducted to determine the fate of pesticides, hydrocarbons, and explosives in
the composting environment (Ziegenfuss et al., 1991).
In the composting of non-hazardous materials, the objectives are to stabilize and
oxidize organic materials, reduce the volume of waste, reduce the moisture content of waste,
and destroy pathogens. Composting of hazardous waste includes the same objectives plus
detoxification of hazardous substances into innocuous end-products.
Optimal application of biotechnology to large-scale compost systems is based on a
working understanding of processes and mechanisms involved in composting of organic
material. Currently, commercial compost operations are operated as black-box systems where
optimization is largely achieved through trial and error. Large-scale treatment of
hazardous waste will require optimal controls to meet the specified end points.
Some proponents of compost treatment have claimed significant success in destruction
of hazardous wastes without strong data to support their claims. Disappearance of parent .
compounds has been used to claim that microorganisms successfully degraded waste chemicals.
However, some toxic chemicals could potentially adsorb to, or react with, humic substances
in the compost and become undetectable. by chemical analysis. Such toxicants might later •
desorb from humus and migrate to the biosphere. This emphasizes the need for well
controlled studies to rigorously document degradation rates and identify metabolic products
of hazardous chemicals, metabolically active microbial species, and mechanisms of hazardous
chemical transformation in compost systems.
We have designed and tested closed bench-scale compost reactors to evaluate
composting processes using contaminated soils. Identification of suitable co-compost and
bulking agents, appropriate ratios of soil to organic components and effective aeration
strategies and rates have been selected as major factors requiring investigation.
This research program is designed to develop a thorough engineering analysis and
optimization of composting as a process to treat soil contaminated with hazardous waste.
Bench-scale composters serve as diagnostic tools to predict treatment effectiveness of
larger systems. Fully enclosed, insulated reactors permit reliable data collection on
microbial population dynamics and fate of toxic chemicals during soil composting.
56
-------
The goal of the compost study is to evaluate the potential use of compost systems in
remediation of soils contaminated with hazardous chemicals. In pursuit of this goal, we
have developed bench-scale composters to model large-scale systems. We are currently
studying the ability of compost microorganisms to biodegrade polynuclear aromatic
hydrocarbons (PAHs) in in-vessel reactors located at the U.S. EPA Test & Evaluation (T&E)
Facility in Cincinnati, OH. Soils contaminated with PAHs have been obtained from the Reilly
Tar Pit Superfund site in St. Louis Park, MN for use in these studies.
BACKGROUND
Optimum conditions for composting may vary depending on a number of factors, but
generally aerobic conditions with 45° - 55°C temperature, 40% - 60% moisture, and a carbon to
nitrogen ratio of 20:1 to 30:1 have been considered best. Mesophilic composting (35° to
55°C) might prove to be the most effective for destruction of wastes. However, it may not
always be practical to maintain temperature below 55°C from an economic standpoint if it
requires too much energy to maintain lower temperature.
Composting differs from other ex-situ soil treatment systems in that bulking agents
are added to the compost mixture to increase porosity. This allows air to flow through the
contaminated soil and maintain aerobic conditions. Microbial metabolism of substrate
generates heat raising the temperature. Temperature changes in the pile bring about changes
in the microbial ecology. Some microbes will fare better at one temperature, while others
will take over as the temperature rises or falls.
The conventional aerobic compost process passes through four major microbiological
phases identified by temperature: mesophilic (30° - 45°C); thermophilic (45° - 75°C);
cooling; and maturation. The greatest microbial diversity has been observed in the
mesophilic stage.
The early stage of composting, where the temperature rises rapidly, is called the
active stage of composting. In an active compost pile, temperature can easily exceed 55°C,
and temperatures as high as 70°C have been reported. Compost temperature above 55°C becomes
lethal to most bacteria. Microbes found in the thermophilic stage have been spore forming
bacteria (Bacillus spp.) (Nakasaki, et al., 1985) and thermophilic fungi (Strom, 1985;
Fogarty and Tuovinen, 1991).
As aerobic microbes are killed during the thermophilic stage, aerobic activity (heat
production) slows and temperature declines back into the mesophilic range below 45°,
Microbial recolonization during the cooling phase brings the appearance of mesophilic
bacteria and fungi whose spores withstood the high temperatures of the thermophilic stage.
A maturation, or curing, phase is the final compost stage. By this stage, most .
digestible organic matter has been consumed by the microbial population, and the composted
material is considered stable. Fungi and Actinomycetes may dominate the eco-system during
the curing stage.
REACTOR DESIGN
Prototype composter evaluation at the U.S. EPA T&E facility has proceeded through
several different designs. The performance of each design was evaluated by conducting a
treatability experiment using PAH-contaminated St. Louis Park soil. For our design
criteria, one particular prototype offered considerable versatility. This design was used
to build stainless steel reactor units.
57
-------
Ten 55-gal. insulated, stainless steel composters have been fabricated to perform
closely monitored treatability studies. These fully enclosed, computer monitored, bench-
scale reactors hold about 1/4 cu. yd total compost mixture.
The reactor units stand upright with air flowing vertically through the compost
mixture. Enclosed units permit online analysis of oxygen, carbon dioxide, and methane at
inlet and exit locations. A data capture system accumulates these data and transmits them
to the central computer that monitors and controls each reactor. XAO traps in the exit line
of each composter permit trapping of volatile organic compounds (VOCs) for analysis.
The bottom of each reactor consists of 2 in. gravel contained above a conical
collection system for periodic sampling of any leachate leaving the reaction mixture. Mass
balance studies on soil contaminants are possible by direct sampling of the reaction mixture
at different depths through bung holes in the lid, together with capture of VOCs and
leachate leaving the reactor.
Periodic determination of compost moisture content in each reactor unit permits
adjustment of total moisture content in the compost matrix to 40X to 50%. Moisture
condensers inside compost units promote recycling of moisture. Otherwise, with typical air
flows, each unit could lose significant amounts of water daily. If moisture falls below
40X, a water distribution system inside the reactor may be used to facilitate water addition
to the reaction mixture without opening the reactor.
Cylindrical reactor design permits mixing of reactor contents by rolling each unit on
a drum roller at desired intervals. Mixing can be used to break up anaerobic pockets and to
avoid packing of the compost mixture. All reactors are mixed simultaneously by placing them
on rollers over a modified conveyor belt that forces the reactors to turn in unison.
Baffles inside reactors promote mixing of the compost during rolling.
An insulated space between the reactor core and the outer shell reduces heat loss
from the reactor during aerobic activity. Heating coils in this space provides the option
of warming the reactor to accelerate composting during startup. Each composter houses five
thermocouples connected to a central computer for online temperature measurements.
Thermocouples reside at four equally spaced locations within the compost mixture and a fifth
thermocouple tracks ambient temperature outside the reaction vessel.
One operational scheme permits temperature control by introduction of ambient air
through a computer controlled valving system. If the mean temperature of the middle two
reactor thermocouples exceeds a predetermined high value, the computer switches that unit to
high air flow (20 L/min) to cool the reaction mixture. After the high-temperature unit
cools to a specified low temperature, the computer switches the unit back to low air flow (5
L/min) to reduce further heat loss from the reaction mixture.
CURRENT RESEARCH
Current studies focus on defining acceptable operating conditions and process
characteristics in order to establish suitable parameters for treatment effectiveness.
Parameters of interest include aeration, moisture dynamics, heat production, and physical
and chemical properties of the compost mixture.
Aeration studies evaluate porosity (air flow) in the compost system, and attempt to
discover relationships between free air space, forced air flow, and composting rate.
Aeration studies also investigate roles of anaerobic and aerobic metabolism in chemical
degradation. Anaerobic pockets may benefit the process by initiating degradation of
58
-------
recalcitrant compounds, especially highly chlorinated compounds, via reductive metabolism.
After an initial reductive step, aerobic biodegradation of toxicants may proceed more
readily. The research program will attempt to.identify optimal aeration rates and pile
mixing frequency for the most effective combination of anaerobic/aerobic conditions for
biodegradation of recalcitrant substrates. These studies will investigate whether forced
anaerobiosis and inoculation with a facultative anaerobe prior to development of aerobic
compost conditions enhances biodegradation of toxic wastes.
Moisture dynamics studies evaluate rates of change in.moisture content for different
regions of the compost reactor. Moisture may be lost through evaporation and convection.
Changes in moisture content are attributable to factors such as aeration rate, temperature,
and compost composition (e.g., soil type and co-compost material).
Heat production may be highly variable throughout the compost reactor. We have
devised a method to continually monitor temperature changes (heat production) at various
reactor locations. Bench-top composters are insulated to control heat loss, thereby
mimicking a large scale compost pile where heat is lost by ventilation and water evaporation
more than by conduction.
A preliminary study was performed on PAH-contaminated soil mixed with 1% by weight
fresh cow manure and 50% by volume ground corn cobs. Reactor core temperature ranged from
20°C to 60°C when moisture content was below 52%. However, when moisture content exceeded
57%, temperature remained below 30°C, suggesting that aerobic metabolic activity was
inhibited at moisture content 57% under these conditions.
Critical properties of the compost mixture include density (g/cc), pH changes in
various reactor.locations, pressure drop across the pile if it is actively aerated, solids
fraction, moisture content, and concentrations of organic compounds. These investigations
focus on potential to enhance biodegradation by manipulation of physical and biological
parameters that influence the process. These studies will also investigate whether recycling
mature compost material into fresh compost enhances biodegradation of contaminants.
Early microbiological studies have focused on enumerating total microorganisms and
determining the presence of PAH degraders. Current studies focus on characterizing changes
in biological activity during the four stages of composting and identifying microbial
species responsible for significant biodegradation of PAHs during each compost stage.
Reappearance of fungi and other mesophiles (e.g. Actinomycetes) during the cooling stage is
also of interest.
FUTURE RESEARCH
Future investigations will, include.treatability studies on pentachlorophenol and
other soil contaminants. Research on technical developments necessary to improve composting
applications for degradation of hazardous waste will remain as a primary objective. To
judge the abilities of microorganisms to degrade hazardous wastes in soil under various
compost conditions, emphasis will be placed on determining mass balance of contaminants and
degradation products in compost, and identifying microbial species responsible for
biodegradation of contaminants.
REFERENCES
Fogarty, A.M and Tuovinen, O.H. 1991. Microbiological degradation of pesticides in yard
waste composting. Microbiol. Rev. June 1991:225-233.
59
-------
Nakasaki, K., Sasaki, M, Shoda. M., and Kubota, H. 1985. Change in microbial numbers during
thermophilic composting of sewage sludge with reference to C02 evolution rate. Appl.
Environ. Microbiol. 49(1):37-41.
Strom, P.P. 1985. Identification of thermophilic bacteria in solid-waste composting. Appl.
Environ. Microbiol. 50(4):906-913.
Ziegenfuss. P. Scott and Williams Richard, T. 1991. Hazardous materials composting. J.
Haz. Mater. 2S. 91-99.
FOR MORE INFORMATION
Carl L. Potter, U.S. Environmental Protection Agency, Risk Reduction Engineering Laboratory,
Cincinnati, Telephone (513) 569-7231, FAX (513) 569-7105.
-------
EVALUATION OF ABIOTIC FATE MECHANISMS
IN SOIL SLURRY BIOREACTOR TREATMENT
John A. Glaser and Paul T. McCauley
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
Majid A. Dosani, Jennifer S. Platt, and E. Radha Krishnan
IT Corporation
11499 Chester Rd.
Cincinnati, Ohio 45246
INTRODUCTION
Biological treatment of contaminated soil slurries may offer a viable technology for soil
bioremediation. Slurry bioreactor treatment of soils, however, has not sufficiently progressed to be a
durable, reliable, and cost-effective treatment option.1
Critical to the evaluation of slurry bioreactors is a better description of pollutant mass transfer
during the treatment phase. Losses attributable to abiotic means are generally overlooked in field
application of the technology. Discussions with EPA regional personnel and inspection of active soil
slurry bioreactor operations have identified operational problems such as foaming which could result in
possible abiotic loss.2 Field bioslurry operations have adopted various approaches to reduce foaming:
1.) the addition of defoaming agents, 2.) the reduction of rotational speed of the agitator, and 3.) the
reduction of gas flow through the bioreactor system. The foaming phenomena is viewed as a nuisance
rather than a potential beneficial removal mechanism. In the case where the pollutants have a specific
gravity less than water, once desorbed from the slurried soil, they would rise to the surface as in a
flotation process. Our hypothesis was that foam formation could be related to this pollutant release
process. If this analysis has merit, it is possible that the operational strategy used in the field is
counterproductive since a separated contaminant phase is re-entrained with a partially cleaned soil
material.
We have conducted two bench-scale slurry bioreactor treatability studies, at the U.S. EPA Testing &
Evaluation Facility in Cincinnati, Ohio, which were designed to investigate some of the operating factors
leading to foam formation and identify the most advantageous means to deal with foaming. The initial
study has been previously presented as a general treatability study for treatment of creosote
contamination in a soil.3 During this study, foaming became a major problem for operation. The foaming
conditions were mitigated by use of defoamer and, in the more extreme cases, through reduction of the
mixer rotational speed and gas flow. A subsequent study which was devoted specifically to investigating
the causes and conditions of foaming using a different batch of soil from the same site as the earlier
study showed little foaming at the very beginning of the study.
METHODOLOGY
Treatability Study3
A soil from St. Louis Park, Minnesota was contaminated with polynuclear aromatic hydrocarbons
(PAHs) and used to evaluate the performance of bench scale slurry reactors. The bench-scale bioslurry
reactor was fabricated from a 8-L glass conventional resin kettle with a four-port cover fitted with
standard taper joints. The reactor vessel was fabricated to have three sample ports located at 5, 10,
and 15 cm vertically from dead center of the reactor bottom. The ports in the reactor cover permitted
introduction of the stirring shaft, influent and effluent gas lines, and a thermocouple temperature probe
61
-------
into the reactor. Operational slurry volume was 6 L, representing (75%) of the total reactor volume.
Ten bench-scale reactors were used to assess; the effect of engineering variables on the
degradation of PAH constituents over a 10-week treatment period.
Experimental Design
Experimental variables selected for this study were solids loading, rotational speed of the mixing
impeller, and dispersant. Soil solids concentrations of 10% and 30% (dry weight basis) were tested.
Two mixing speeds were evaluated. A high mixing rate was selected for complete off-bottom
suspension. A low mixing rate was arbitrarily set at 200 rpm lower than the high mixing rate. Effective
high mixing rates were found to be 650 rpm and 900 rpm for 10% and 30% soil solids, respectively.
The dispersant (Westvaco, Reax 100M) was added to test its ability to minimize foam production.
Two separate reactors were operated under abiotic conditions to serve as bio-inactive control
reactors. Formaldehyde was used as a biocide in these reactors and maintained at 2% residual
concentration.
The following monitoring and operating conditions were held constant for the reactors:
Dissolved Oxygen (>2 mg/L)
pH (range of 6-9)
Ambient Temperature (recorded daily)
Treatment Duration (10 weeks)
Nutrients (C:N:P ratio = 100:10:1)
Antifoam (as needed to control foam)
Foam Study
To evaluate the conditions and causes of foam formation, a subsequent study was designed. This
investigation used the monitoring conditions specified for the treatability study and was conducted using
six bench-scale slurry reactors. Each reactor was loaded at 30% solids with an initial volume of 6 L.
The study design employed two reactors which were permitted to develop foam, two reactors in which
the foam formation was suppressed through the use of defoamer; and two reactors where
formaldehyde was used to suppress biological activity.
RESULTS
The high solids control reactor showed the greatest amount of foaming in the treatability study. This
is a surprising event since our expectation was that foaming was related to the formation of
biosurfactants by the microbiota. The addition of formaldehyde to the control reactor was the only other
explanation for the foam formation observed. The other reactors had foaming problems but this reactor
was very noticeable in contrast to the others. Higher solids loading was also observed to contribute to
foam formation.
In contrast with the initial treatability study, the designed foaming study showed little foam formation.
There was a small amount of foaming on the first day of operation. There were a few differences in the
physical characteristics of the soil and the operation of the bioslurry reactor between the two studies
which may have contributed to decreased foam formation in the designed foam study. Although soil
from the same site was used for both the treatability study and foam study, the batch of soil used for
the foam study was coarser (<1/4") which may have resulted in lower PAH concentrations in the
foaming study and lead to decreased foam formation. Furthermore, the air flow rate used for all six
reactors in the foam study (1 ft3/min) was approximately 20 percent of that used in the treatability study
(5 ft3/min) which could have also contributed to decreased foam formation.
62
-------
CONCLUSIONS
The lack of foam formation in the one study devoted to its analysis was a set back to our initial goal.
We will continue to investigate foam formation and appropriate controls for soil slurry bioreactors. The
unpredictable nature of foaming events must be better understood to provide dependable operating
conditions required for field scale bioslurry applications.
REFERENCES
1. U.S. EPA, Engineering Bulletin: Slurry Biodegradation, EPA/540/2-90/016 Cincinnati, Ohio.
2. Jerger, D.E., Erickson, S.A., and Rigger, R.D. "Full-Scale Slurry Phase Biological Treatment of
Wood Preserving Wastes at a Superfund Site." 1994 Draft Manuscript.
3. Glaser, J.A., Dosani, M.A., McCauley, P.T., Platt, J.S., Opatken, E.J., and Krishnan, E.R. Soil
Slurry Bioreactors: Bench Scale Studies. ]rv Twentieth Annual RREL Research Symposium:
Abstract Proceedings. EPA 600/R-94-011. United States Environmental Protection Agency,
Cincinnati, Ohio, 1994. p. 127.
FOR MORE INFORMATION:
John A. Glaser
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
(513) 569-7568 (Voice)
(513) 569-7105 (FAX)
63
-------
JSFORMATION OF SULFUR AND NITROGEN OXIDES IN STACK GASES
Rakesh Govind, Rama Puligadda
Department of Chemical Engineering
University of Cincinnati
Cincinnati, OH 45221
and
Dolloff F. Bishop
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26W. Martin Luther King Drive
Cincinnati, OH 45268
INTRODUCTION
The Nation's large supply of high sulfur coal and increasingly stringent emission regulation led to priority
development of advanced innovative processes for treating pollutants in flue gases from coal combustion. The
principal pollutants in flue gases, sulfur oxides (802,803) and nitrogen oxides (NOX) cause acid rain. Thus, the
Department of Energy's Clean Coal Program is funding projects to commercialize technologies that minimize
emission of sulfur and nitrogen oxides at power plants. The emerging technologies should be applicable to coal-
fired combustors in the electric utility, industrial and commercial sectors. The technologies should also achieve
high degrees of reliability and have low operating and capital costs. The status of flue gas cleanup technology is
summarized in the folowing sections.
Flue gas cleanup technology provides emission control of sulfur dioxide, nitrogen oxides, and particulate
matter (PM) in combustion of coal using two generic categories of processes: One category, dry sorbent injection,
uses dry sorbents such as limestone injected directly into the combustion zone or downstream of a coal- fired boiler,
to capture sulfur dioxide insitu. In this strategy, a low-NOx combustion environment is usually maintained to aid
capture. The second category, post combustion gas cleanup, uses sorbents in a slurry, aqueous liquor, or dry
powder injected into the combustion gas leaving coal-fired boilers to effect post combustion capture of SO2-
Processes in both categories employ various approaches to control emission of particulate matter (PM).
In the last 10 to 20 years, a wide variety of dry sorbent injection and post combustion gas cleanup systems
(1-12) have evolved. A representative system for dry sorbent injection is the Limestone Injection Multistage Burner
(LIMB) used at utility power- generating stations in existing pulverized coal-fired (PC) boilers firing run-of-mine
(ROM) coal. The boilers are equipped with conventional particulate matter (PM) controls (typically an electrostatic
precipitator). The LIMB system injects dry alkali metal limestone or hydrated lime into the furnace operating with
low-NOx combustion environment achieved through modifications of the burner assembly or combustion operating
conditions. LIMB is an emerging technology undergoing research and development at the bench, pilot, prototype,
and demonstration plant levels. The thrust of ongoing research is to identify factors that govern performance to
optimize removal efficiency.
Numerous activities by both private and public sectors to improve emission control systems are
summarized as follows:
1. Research for control of particulate matter (PM) is focussed on performance improvement and
optimization. Due to concern for health risks from trace element and inhalable particulate matter,
emphasis is placed on the removal of submicron-sized particles.
64
-------
2. Research in existing flue gas desulfurization (FGD) systems is focused on the use of organic acids or
magnesium salts to enhance SC>2 removal efficiency and reagent utilization. Results indicate that a
removal efficiency of 95 % can be achieved at reduced operating costs.
3. Research in existing and new FGD systems is to reduce fresh water consumption and clean-up of waste
water discharges since these requirements may improve siting and operational problems at coal power
plants.
4. A promising low-cost FGD option for 862 control is dry injection of sorbent in flue gas before the fabric
filter. This process has been demonstrated by the Electric Power Research Institute (EPRI) at full-scale
and is applicable to both new and existing low-sulfur coal power plants. Additional research is proceeding
on high-sulfur coal applications for use with electrostatic precipitators, for improved waste fixation and
disposal, for system optimization and for use of lower-cost alternate reagents.
5. Advanced limestone/gypsum FGD processes are being developed for application which produce
marketable gypsum by forced oxidation of the spent slurry.
6. For post-combustion control of NOX, selective catalytic and selective non-catalytic reduction systems are
the most advanced. Pilot- scale systems of these two technologies have been successfully tested on coal-
fired power plants. However, these processes are more expensive than modification of burners for low
NOX emission. Major improvements are needed in the process control subsystems, extension of catalyst
life, elimination of ammonia leakage, and reduction in overall cost.
Long term research efforts are focused on advanced SC>2 control, combined SC>2/ NOX control and PM
removal. The primary emphasis in advanced SC>2 control technologies is placed on reagent regeneration and
saleable product processes to eliminate or minimize solid waste disposal problems.
Biological desulfurization of coal is a pretreatment method that attempts to remove both pyritic and
organic sulfur in coal before combustion. In the past, several laboratory studies have been reported to remoye sulfur
from coal using microorganisms (13-18). Two main microorganisms that have been studied are Thiobacillus
ferroxidans, a mesophilic (10°C-37°C) bacterium isolated from acid mine drainage, and Sulfolobus
acidocaldarius, a thermophillic (55°C-80°C) bacterium. Both organisms aerobically oxidize pyritic sulfur,
primarily FeS.2, and organic sulfur to soluble sulfates. Atlantic Research Corporation, Virginia (19) developed a
unique microorganism (CBI) capable of removing a portion of the organic sulfur from coal. CBI primarily attacked
thiophenic sulfur in coal, removing the sulfur as a water soluble sulfate. In laboratory batch and bench-scale
continuous systems, CBI removed up to 47% of the organic sulfur. The sulfur removal depended on coal type,
particle size and surface properties. Attia and Elzeky (20) used thiobacillus feirooxidans bacteria to selectively
modify the pyrite surface chemical behavior to improve pyrite sedimentation in the flotation of coal. They studied
the effects of variables (bacterial conditioning time, concentration of bacterial cells, pH, temperature and nutrient
solution) on the flotation behavior of pyrite and coal suspensions. Their results indicated that about 40% reduction
in the pyrite content of coal was achieved with bacterial pre-conditioning of less than 15 minutes, while coal
flotation was unaffected.
METHODOLOGY,
Current sulfur and nitrogen oxide control approaches for contaminated gas streams such as stack
gases have limitations and are relatively expensive. Practical methods for removal of carbon dioxide are not
available. In this paper, the controlled use of biotransformation to remove these oxide pollutants from
contaminated gases involves complex biooxidation and bioreduction of the oxides in appropriate; bioreactors,
ultimately to elemental sulfur, nitrogen gas and biomass. Two separate bioreactor experiments were conducted to
investigate the application of biological reactors for removal of sulfur dioxide, nitrogen oxides and carbon dioxide
from stack gases.
Biofilter Studies ,
An experimental biofilter was operated for 5 months to study biological removal of sulfur dioxide,
nitrogen oxides and carbon dioxide from stack gases. A 2-inch diameter stainless-steel tube was used to construct
the biofilter and ceramic straight-passage monolith with 100 cells/square inch, obtained from Corning Glass, Inc.,
65
-------
NY, was used as the support media. The height of the biofilter was 3 feet. Stack gas, with fiowrate that was varied
in the range 1-2 liters/minute, was introduced at the top of the biofilter. The recycled OECD nutrients (21) were
also trickled down the biofilter cocurrent with the flow of the stack gas. 30 mg of Thiobacillus (obtained from
ATCC culture bank and grown in specialized culture reactors) was suspended in 0.5 liters of OECD nutrients and
recirculated through the biofilter to seed the ceramic media with the desired organisms. About 16 gms of the
culture was attached after 2 hours of recirculation of the suspended culture through the biofilter.
Stack gas, with the following composition: Carbon dioxide (41.56%), Nitrogen (50%), Oxygen (8%),
Sulfur dioxide (0.4%), nitrogen oxides (0.04%), was introduced at the top of the biofilter at an initial fiowrate of 1
liter/minute, which resulted in a 2-minute gas phase empty-bed residence time. OECD nutrients with the following
composition: (21): KH2PO4 (85 mg/L)), K2HPO4 (217.5 mg/L)), Na2HPO4.2H2O (334 mg/L), NH4C1 (25 mg/L),
MgSO4.7H2O (22.5 mg/L), CaCl2 (27.5 mg/L) and FeCl3.6H2O (0.25 mg/L), MnSO4.H2O (0.0399 mg/L),
H3BO3 (0.0572 mg/L), ZnSO4.7H2O (0.0428 mg/L), (NH4)6Mo7O24 (0.0347 mg/L), FeCl3.EDTA (0.1 mg/L),
and yeast extract (0.15 mg/L) were also introduced at the top of the biofilter at a fiowrate of 4 liters/day. The
biofilter was operated at ambient temperature of 25°C.
Suspended Culture Bioreactor
A bioreactor experiment was conducted to evaluate SO2/NOX removal from stack gases by suspended
culture bacteria. The design and operating conditions for the bioreactor have been summarized in Table 1. Stack
gas, consisting of carbon dioxide, nitrogen, sulfur dioxide, nitrogen oxides and oxygen was bubbled at a rate of 1.5
liters per minute through the bioreactor. The bioreactor, a 4-inch diameter glass column, 3 feet tall, was filled with
2 feet of deionized distilled water and 20 gms of activated sludge as seed. There was no further addition of
activated sludge to the reactor. The reactor was mixed by the incoming stack gas flow. The reactor detention time
was 4 minutes for the gas stream. The reactor detention time has not been optimized.
The incoming gas stream and the exiting gas stream compositions (refer to Table 1) were analyzed by a
sulfur dioxide and nitrogen oxides Beckman Analyzer. OECD nutrient media was batch fed daily to the reactor
(0.2 liters) and an equivalent amount of liquid was withdrawn (Table 1) from the reactor. The exit gas was also
analyzed using a gas chromatograph.
RESULTS
Biofilter Studies
Complete removal of sulfur and nitrogen oxides was observed in this experiment. No oxygen or hydrogen
sulfide was present in the exit gas stream. The carbon dioxide composition was reduced to 32.4%, from an initial
feed gas composition of 41.56%. The composition of nitrate in the exit liquid stream was 12 ppm and sulfate
composition was 10 ppm. Free sulfur was obtained from the liquid stream. The amount of free sulfur, quantitated
by filtering and weighing, closely agreed with the complete conversion of sulfur dioxide to sulfur.
The biofilter consisted of two distinct zones: (1) oxidation zone, in which the sulfur dioxide and nitrogen
oxides were converted abiotically and biologically to sulfate and nitrate in the liquid phase; and (2) reduction zone,
in which the sulfate and nitrate were converted to hydrogen sulfide, sulfur and nitrogen gas. The hydrogen sulfide
formed also reacted with sulfur dioxide in the gas phase to form sulfur, which exited with the nutrient liquid flow.
The following are the possible reactions occuring in the oxidation and reduction zones:
Oxidation zone: Biological reactions
Bacteria
1/2 O2 + H2O + SO2 -> SO4= + 2H+
Bacteria
1/2 O2 + H2O + NO > NO3- + 2H+
66
-------
Bacteria
1/2 O2 + H2O + 2NO2
2NO3- + 2H+
Sulfur dioxide is also oxidized to sulfate abiotically.
Reduction zone: Biological reactions
Bacteria
SC>42~ + organic source ----------- > H2S + alkalinity
Bacteria
NC>3~ + organic source ----- •: ----- > 1/2 N2 + alkalinity
Bacteria .
+ organic source ------------- > S^ + alkalinity
With nutrient recycle, the alkalinity (HCO^") produced in the bioreduction zone neutralizes the acidity
produced due to the formation of sulfate and nitrate in the oxidation zone. Waste activated sludge was
continuously added to the biofilter and primarily served as the organic source for the bioconversion of sulfate to
hydrogen sulfide. The steady-state pH of the exit nutrient stream was 9.2. The majority of the carbon dioxide in
the gas stream was absorbed by the bicarbonate to form carbonate. The H2S produced in the reduction zone
combined with sulfur dioxide (SO2) in the gas phase to form free elemental sulfur.
2H2S + SO2 — > 3S° + 2H2O
The major problem encountered in this experiment was the senstivity of the exit gas composition to inlet
stack gas flowrate. If the stack gas flowrate was abruptly changed even by a small amount, the conversion of sulfur
dioxide would immediately drop and then again achieve 100% conversion after a period of 2-3 hours. For
example, at a stack gas flowrate of 1.2 liters/minute, immediately after being changed from a steady-state flow of 1
liter/minute, the conversion of sulfur dioxide immediately dropped from a steady-state value of 100% to 85%; i.e.,
15% of the incoming sulfur dioxide was observed in the exit gas stream. Nitrogen oxides were also observed in the
exit gas stream, although their composition was very low due to low inlet composition. However, if the biofilter
was operated for about 3 hours at a steady-state stack gas flowrate of 1.2 liters/minute, then the sulfur dioxide and
nitrogen oxide conversion again reached a steady-state value of 100%. * .
The instability of the biofilter process to changes in stack gas flowrate was mainly due to disruption of the
oxidation-reduction zones, due to penetration of oxygen, present in stack gas, into the reduction zone. However,
after the reactor had been operated for some time, usually a few hours, bacterial growth responded to restore the
balance of the oxidation-reduction zones, thereby again attaining complete conversion of the sulfur and nitrogen
oxides to sulfur and nitrogen gas.
Suspended Culture Bioreactor
After acclimation, complete removal of sulfur dioxide, nitrogen oxides and oxygen was observed in the
bioreactor for a period of two months. 30% of incoming carbon dioxide was removed in the bioreactor. The reactor
pH remained at near neutral values. The chemoautotrophic sulfate reducing bacteria also use carbon dioxide from
the stack gas to support biomass development. Thus CO2 conversion to biomass and alkalinity (HCC^') also
Occurs in the process. Chemoautotrophic bacteria function over a wide range of temperature from normal ambient
to high temperatures nearing the boiling point of water. Temperature optimization of the processes have not been
attempted. About 25% of the sulfur (SO2) entering the reactor was released as H2S in the exit gas. The carbon
dioxide in the exit liquid was determined by acidification and measurement of CO2 evolution. The bulk of the
carbon dioxide removed from the stack gas was converted to carbonate in the liquid phase. The biomass produced
by chemoautotrophic sulfate reducing bacteria supplied some of the organic material needed for sulfate reduction.
Waste activated sludge was added to the reactor and primarily served as the organic source for sulfate reduction.
67
-------
Analysis of the 0.2 liter aqueous stream from the bioreactor revealed that the nitrate/nitrite composition
during the two month study varied between 50-120 ppm, the sulfate concentration was steady at 180 ppm.
Qualitatively, particulate sulfur was present in the bioraass.
The naturally adapted organisms and the water in the bioreactor support the following complex reactions:
Dissolution and Oxidation in Water:
SO2 + H2O ==> H2SO3
2SO2 + 2H20 + 1/2O2 -> H2SO4 + H2SO3
H2SO3 + 1/2O2 —-> H2SO4
2NO + O2—->2NO2
2NO2 + H2O —-> HNO2 + HNO3
CO2 + H2O => H2C03
Bioreduction:
CH3COO- + SO42' ==
3CH3COO~ + 4SO32' =
2H+ + S2'
SO42' + 2ET1- + 3H2S
Bacteria
Bacteria
=> H2O + CO2 + HCO3- + S2'
=> 3H2O + 3CO2 + 3HCO3-+ 4S2'
H2S
4S
4H2O
Nitrite and nitrate ions are also biologically reduced to nitrogen gas. The alkalinity (HCO3~) produced by
the reduction neutralizes the acids produced in the oxidation reactions and by the dissolution of the gases in the
water of the bioreactor. The steady-state pH in the bioreactor was 9.0.
Sulfur, biomass, sulfate and nitrate ions are removed from the reactor in the effluent aqueous stream. The
small residuals of soluble sulfate and nitrate ions may be separated from the particulate sulfur and biomass in the
stream. With or without separation, the soluble sulfate and nitrate ions may be reduced in a second reactor using
waste organic sources, such as sewage sludge.
CONCLUSIONS
Biological conversion of sulfur dioxide and nitrogen oxides in stack gases is possible using either
immobilized biofilms or suspended culture reactors. Complete conversion of sulfur and. nitrogen oxides are
achieved in the reactors with the final products being sulfur and nitrogen gas, respectively. Detailed reactions
explaining the experimental data have been presented in this paper. Biotransformation of stack gas contaminants
offers an alternative approach to manage combustion of high sulfur coal, without requiring landfilling or disposal
of sludge.
REFERENCES
1. McCrea, D.H., Forney, A.H., and Myers, J.G., Recovery of Sulfur from Flue gas Using a Copper Oxide
Absorbent. J.Of Air Pollution Control Association, Vol. 20. No. 12. 819, Dec. 1970.
2. Strakey, J.P., Bauer, E.R., and Haynes, W.P., Removal of SO2 and NOX from flue gad in Fluidized bed
copper oxide process. Presented at the Sixth National Conference on Energy and Environment, Pittsburgh,
Pa., May 21-24, 1979.
68
-------
3. Demski, R.J., Gasoir, S.J., Bauer, E.R., Jr., Yeh, J.T., Strakey, J.P., and Joubert, J.I., Simultaneous
removal of SC>2 and NOX from flue gas employing a fluidized bed copper oxide process. Presented at the
1092 Summer National Meeting of the AIChE, Cleveland, Ohio, August 29- September 1, 1982.
4. Yeh, J.T., Demski, R.J., Strakey, J.P., and Joubert, J.I., PETC Fluidized bed copper oxide process for
combined SC>2 /NOX removal from flue gas. Presented at the 1984 Winter National meeting of the
AIChE, Atlanta, Ga., March 11-14, 1984.
5. Drummond, C.J., Yeh, J.T.,Joubert, J.I., and Ratafia- Brown, J.A., The design of a dry, regenerative
fluidized bed copper oxide process for the removal of sulfur dioxide and nitrogen oxides from coal fired
boilers. Presented at the 78th Annual Meeting and Exhibition of the Air Pollution Association, Detroit,
Michigan, June 16 -21, 1985,
6. McArdle, J.C., Leshick, D.G., and Williamson, R.R., Sorbent Life Cycle Testing of the Fluidized bed
copper oxide process, Presented at the 1987 Spring National Meeting of the AIChE, Houston, Texas,
March 29-April 2, 1987.
7. Tseng, H.P., Haslbeck, J.L., and Neal, L.G., Evaluation of the NOXSO combined NOX/SO2 flue gas
treatment process. Report to the U.S. Department of Energy , Contract No. DE-AC22-FE6-0148,
September 30, 1983.
8. Haslbeck, J.L., and Neal, L.G., Technical evaluation of the NOXSO combined NOX/SO2 flue gas
treatment process. Presented at the 1985 Spring National Meeting of the AIChE, Houston, Texas, March
24-28, 1985.
9. Yeh, J.T., Drummond, C.J., Haslbeck, J.L., and Neal, L.G., The NOXSO process : Simultaneous removal
of SO2 and NOX from flue gas. Presented at the 1987 Spring National Meeting of the AIChE , Houston,
Texas, March 29- April 2, 1987. ,
10. Markussen, J.M., Yeh, J.T., and Drummond, C.J., Enhanced removal of nitrogen oxides in a spray dryer
using a lime slurry containing sodium hydroxide. Presented at the 1986 Spring National Meeting of the
AIChE, New Orleans, April 6-10, 1986.
11. Walker, R.J., Drummond, C.J., and Ekmann, J.M., Evaluation of Advanced separation techniques for
application to flue gas cleanup process for the simultaneous removal of sulfur dioxide and nitrogen oxides,
DE/PETC/TR-85/7 (DE85013006), June 1985.
12. Joubert, J.I., Drummond, C.J., Development of advanced flue gas emissions control technologies at the
Pittsburgh Energy Technology Center, in "Coal Combustion", ed by J. Feng, Hemisphere Pub. Corp. N.Y.
P. 853, 1988. '
13. Kargi, F., and Robinson, J.M., Microbial Desulfurization of coal by Thermophilic Microorganism
Sulfolobus Acidocaldarius, Biotechnology and Bioengineering, 24, p. 2115. (1982).
14. Kargi, F., Microbiological coal desulfurization, Enzyme Microb. Technpl. 4. Jan. p. 13. (1982).
15. Kos, C.A., Poorter, R.P.E., Bos, P., and Kuenen, J.G., Geochemistry of sulfides in coal and microbial
leaching experiments, Int. Conf. on coal science , Dusseldorf, p.842. (1981).
16. Dugan, P.R., Apel, W.A., Microbial desulfurization of coal, in : Metallurgical Application of Bacterial
Leaching and Related Phenomena, L.E. Muer, A.E. Torma, and J.A. Brierly (BOS). Academic Press.
N.Y., p. 223. (1978).
17. Detz, C.M., and Barvinchak, G., Microbial desulfurization of coal, Min. Congr. J. 65. p.75. (1979).
69
-------
18. Isbister, ID., and Kbylinski, E.A., microbial desulfurization of coal, in : Processing and Utilization of
High Sulfur Coal, ed. by Y. A. Attia, Elsevier, N.Y., p. 627, (1985).
19. Murphy, J., Riestenbery, E, Mohler, R, Marek, D., Beck, B., and Skidmore, D., Coal desulfurization by
microbial processing, in : Processing and Utilization of High Sulfur Coal, ed. by Y.A. Attia Elsevier,
N.Y., p. 643. (1985).
20. Attia, Y.A., Elzeky, M.A., Biosurface modification in the separation of pyrite from coal by froth flotation.
in : Processing and Utilization of High Sulfur Coals, ed. by Yosry. A. Attia, Elsevier, N.Y. p. 673. (1985).
21. OECD Guidelines for Testing of Chemicals, EEC Directive 79/831, Annex V, Part C: Methods for
Determination of Ecotoxicity, 5.2 Degradation. Biotic Degradation. Manometric Respirometry. Method
DGXI, Revision 5, 1983.
FOR MORE INFORMATION:
Rakesh Govind
Department of Chemical Engineering
University of Cincinnati
Cincinnati, OH 45221
Phone: 513/556-2666
Table 1. Design and Operating Conditions of the Suspended Culture Bioreactor
Diameter of reactor
Height of liquid in reactor
Amount of activated sludge
added at start of experiment
Gas phase retention time
Inlet Gas flowrate
Outlet Gas flowrate
Inlet Gas Composition:
4 inches
24 inches
20gms
4 minutes
1.5 liters per minute
1.2 liter per day
Exit gas composition:
Reactor temperature and pH
Amount of nutrient media
fed daily to reactor
Amount of reactor liquid
withdrawn daily
Exit liquid composition
(average) nitrate/nitrite
sulfate
free sulfur
Carbon dioxide
Nitrogen
Oxygen
Sulfur dioxide
Nitrogen oxides
Carbon dioxide
Nitrogen
Moisture
Hydrogen sulfide
Sulfur oxides
Nitrogen oxides
27°C; 9.0
41.56%
50.00%
8.00%
0.40%
0.04%
36.4%
62.6%
0.88%
0.12%
0.00%
0.00%
0.2 liters (OECD composition)
0.2 liters
100 ppm
180 ppm
present
70
-------
ANAEROBIC BIODEGRADATION OF PESTICIDES AND TNT:
SITE EMERGING TECHNOLOGY DEMONSTRATION
Wendy Jo Davis-Hoover
Risk Reduction Engineering Laboratory (RREL)
5595 Center Hill Ave.
Cincinnati, Ohio 45224
513-569-7206; fax 513-569-7879
Trevor Jackson
SAIC
Idaho Falls, Idaho
208-528-2131; fax 208-528-2197
INTRODUCTION
J. R. Simplot, Inc. is a multiconglomerate whose initial interest was in the production of
potatoes in the Pacific Northwest. This has lead them into the agricultural chemical business including
pesticides. Dinoseb (a carcinogenic herbicide used to kill the foliage of potato plants allowing for easer
harvesting) was used by them and after its ban in the U.S., J. R. Simplot realized the only acceptable
way to dispose of dinoseb contaminated soils was to incinerate them at considerable cost. Thus
J.R.Simplot became interested in developing a disposal solution that would be less cost prohibitive and
as effective as incineration.
J. R. Simplot submitted a proposal and was accepted into the SITE Emerging Technology
Program to optimize the anaerobic biodegradation of nitroaeromatics (dinoseb) contaminated soils with
the addition of water, pH buffers, and potato waste as a source of starch. This work ultimately resulted
in a good understanding of the biological and chemical process of destruction (optimum conditions of
35 and 37 degrees C, and pH of approximately 7, leads to destruction in 14 days) of both dinoseb and
TNT and in several peer reviewed publications (1-7).
METHODOLOGY
The process was then tested by the SITE Demonstration Program at a small field scale at
much less than optimum conditions at Ellensburg, Washington in the cold summer of 1993 on soil
contaminated with an average of 27.3 milligrams of dinoseb per kilogram (mg/kg) of soil. The process
was also evaluated at the previously functioning Weldon Spring Ordnance Works near St. Louis in the
even colder fall/winter of 1993/94 on soil contaminated with an average of 1510 mg of TNT/kg of soil.
Sixty one samples were randomly collected from the excavated one half inch screened
contaminated soil and analyzed for dinoseb by HPLC using a method developed for this evaluation.
The 39 cubic yards of contaminated soil was moved into a truck-trailer sized reactor while 2-5 percent
of the weight was added as potato waste. Dinoseb-free water was added at a 1 U kg of contaminated
soil ratio. The pH was buffered to about 7.1 and an inoculum of previously treated soil (5 gallon
bucket) was added. Two mixers were used with the dinoseb bioreactor to keep the soils and liquids
mixed but not to add oxygen. The process is initiated by allowing the aerobic organisms to consume
the potato waste and thus utilize the oxygen. The anaerobic conditions that result allow the anaerobic
organisms to degrade the dinoseb and any byproducts. A negative control of excavated, screened, and
dinoseb-contained soil was kept at the site and analyzed before and after the testing for dinoseb. After
treatment, 41 randomly generated areas of the bioreactor were sampled, and the sediments, and eluent
from these samples were analyzed for dinoseb and any known byproducts along with a toxicity screen
which showed the presence of other pesticides.
71
-------
The TNT contaminated soils were treated in the same manner except the soils had a higher
clay count and thus the mixers were ineffective. To obtain the soil/water interface, the bioreactor was
therefore lanced every 7-14 days. As the ambient temperature dropped significantly at this site, heaters
were added to keep the bioreactor from freezing. TNT levels were determined by EPA method 8330
at time zero, 5 months and 9 months. Toxicity tests were analyzed on time zero and at 5 months.
RESULTS
No dinoseb was detected in any samples after only 23 days of remediation even though the
temperature was as low as 18 degrees C. The concentrations of nitroaniline, parathion, malathion, and
4,4'-DDT were also found to be decreased in this process in just 23 days.
The winter conditions in St. Louis threatened to freeze the bioreactor so heaters were added to
bring the temperature up from 4 degrees C. Sampling at roughly 5 months showed a 95 percent
reduction in the concentration of TNT, and significant reductions in toxicity were measured by root
elongation, early seedlings, and earthworm reproduction tests. After sitting for another 4 months the
TNT contamination in the soil was reduced by 99.4 (95 % C.I. is 98.3-99.9) percent.
CONCLUSIONS
This process has shown substantial removal of dinoseb, TNT, their known byproducts, and
toxicity in general (as measured by several methods) at costs that are significantly less than soil
incineration even at non-optimized conditions. Testing has also shown that nitroaniline, parathion,
Malathion, and 4,4'-DDT levels can also be reduced using this technology. Although the soil is
excavated and thus disrupted it is left in better condition with this process than with incineration. Thus
it appears to be a viable alternative to soil incineration for these contaminants. More complete
information on these demonstrations can be found in their Innovative Technology Evaluation Reports
(8, 9) and Technology Evaluation Reports (10, 11).
REFERENCES
1. Kaake, R,, Roberts, D., Stevens, T., Crawford, R., and Crawford D. Bioremediation of Soils
Contaminated with the Herbicide 2-sec-Butyl-4,6-Dinitrophenol (Dinoseb). Applied and
Environmental Microbiology, May 1992, p. 1683-1689.
2. Spiker, J., Crawford, D., and Crawford, R. Influence of 2,4,6-Trinitrotoluene (TNT)
Concentration on the Degradation of TNT in Explosive-Contaminated Soils by the White Rot
Fungus Phanerochaete Chrysosporium. Applied and Environmental Microbiology, Sept. 1992,
p. 3199-3202.
3. Funk, S., Roberts, D., Crawford, D., and Crawford, R. Initial-Phase Optimization for
Bioremediation of Munition Compound-Contaminated Soils. Applied and Environmental
Microbiology, July 1993, p. 2171-2177.
4. Kaake, R., Crawford, D., and Crawford, R. Optimization of Anaerobic Bioremediation Process
for Soil Contaminated With the Nitroaromatic Herbicide Dinoseb (2-Sec-Butyl-4,6-Dinitrophenol).
Jn: Applied Biotechnology for Site Remediation. Lewis Publishers, Boca Raton, Ann Arbor,
London, Tokyo.
5. Roberts, D., Kaake, R., Funk, S., Crawford D., and Crawford, R. Anaerobic Remediation of
Dinoseb from Contaminated Soil - An On-site Demonstration, jn: Applied Biochemistry and
Biotechnology, Vol. 39/40, 1993.
72
-------
6. Daniel, B., Korus, R., and Crawford, D. Anaerobic Bioremediation of Munitions-Contaminated
Soil. In Press, October 1993.
7. Funk, S., Crawford, D., Roberts, D., and Crawford, R. Two-Stage Bioremediation of TNT
Contaminated Soil. In Press, October 1993.
8. U.S. Environmental Protection Agency. Ex-Situ Bioremediation Technology for Treatment of
Dinoseb-Contaminated Soils. Innovative Technology Evaluation Report. U.S. EPA, Cincinnati,
Ohio, 1995.
9. U.S. Environmental Protection Agency. Ex-Situ Bioremediation Technology for Treatment of
TNT-Contaminated Soils. Innovative Technology Evaluation Report. U.S. EPA, Cincinnati,
Ohio, 1995.
10. U.S. Environmental Protection Agency. Ex-Situ Bioremediation Technology for Treatment of
Dinoseb-Contaminated Soils. Technology Evaluation Report. U.S. EPA, Cincinnati, Ohio, 1995.
11. U.S. Environmental Protection Agency. Ex-Situ Bioremediation Technology for Treatment of
TNT-Contaminated Soils. Technology Evaluation Report. U.S. EPA, Cincinnati, Ohio, 1995.
FOR MORE INFORMATION
Wendy Jo Davis-Hoover
RREL
5595 Center Hill Ave.
Cincinnati, Ohio 45224
513-569-7206; fax 513-569-7879
For report requests call 513-569-7794
73
-------
IRON ENHANCED ABIOTIC DEGRADATION OF
CHLORINATED HYDROCARBONS
Chien T. Chen
U.S.EPA, Releases Control Branch
2890 Woodbridge Avenue, Edison, NJ 08837-3679
(908) 906-6985
INTRODUCTION
Since the 1970s, several researchers have investigated the ability of certain zero-valent
metals or alloys to enhance the degradation of halogenated organic compounds in contaminated
water. Iron, zinc, aluminum, brass, copper, and stainless steel have been studied at various times
with varying degrees of success. Gillham and O'Hannesin have recently made a literature review
and conducted tests on 14 halogenated aliphatic compounds using zero-valent iron as an enhancing
agent (1). The results showed that rapid dehalogenation occurred on all of the compounds tested
except dichloromethane. Based on these test results, EnviroMetal Technologies, Inc. proposed to
remediate groundwater contaminated with chlorinated organic compounds using this technology.
The EPA Superfund Innovative Technology Evaluation (SITE) program has accepted this
technology for demonstration. This demonstration project will include two processes, above
ground reactor and in situ permeable wall. The demonstration on the above ground reactor is being
conducted at a site in Wayne, New Jersey. The main contaminants at this site are
tetrachloroethene (PCE) and trichloroethene (TCE). The in situ permeable wall process will be
conducted at a site in upstate New York. This site is a shallow sand aquifer containing TCE,
dichloroethenes, and 1,1,1-trichloroethane.
METHODOLOGY
Before the site demonstration, laboratory experiments were conducted on the contaminated
groundwater to evaluate the efficiency of the process , under simulate site conditions. The
procedures of this and the two demonstration processes are described below.
Column tests were undertaken in the laboratory to simulate the dynamic conditions of
groundwater. The columns were constructed of plexiglass with a length of 50 cm and an internal
diameter of 6.5 cm. Seven sampling ports were positioned at distances of 2.5, 5, 10, 15, 20, 30,
40 and 50 cm from the inlet end. The columns also allowed for collection of samples from the
influent and effluent solutions. A total of three column experiments were set up, one control and
two reactive columns. The control column was packed with 100% silica sand. One reactive
column was packed with a mixture of 50% (by weight) granular iron and 50% (by weight) silica
sand, while the other reactive column was packed with 100% granular iron. The mass of iron to
volume of solution ratio for the 50% iron column was 3.4 g : 1 ml and for the 100% iron column
was 8 g : 1 ml.
The above ground reactor was constructed of fiberglass with a height of 9 feet and an
internal diameter of 8 feet. The reactor was packed from the bottom to the top with: 6" of pea
gravel, 66" of granular iron, 24" of groundwater and 12" of air. The total weight of iron was
18003 kg with a porosity of about 0.4. The pore volume was calculated to be about 3400 L. Five
sampling ports were positioned at distances of 21, 36, 48, 66 and 84 inches from the inlet end
(top of the water level). The influent and effluent solutions were also collected. During the
demonstration, groundwater was extracted from the site by passive recovery techniques, using
tiles placed in the bottom of collection trenches (about 14 ft below the ground surface).
74
-------
Groundwater flowed through the collection trenches to a sump, where it was pumped to the above
ground reactor at a rate of 1.9 L per minute. This would let the water to have the residence time
in the reacting medium of about 26 hours during which the TCE and PCE in the contaminated
water would be completely dechlorinated according to the half lives of those two compounds
obtained in the column test.
The in situ permeable reactive wall will be constructed by filling a trench with iron and inert
materials. Groundwater will move naturally through the wall. The contaminants are expected to
be dechlorinated after passing the reactive wall.
RESULTS
Column Tests
Tables 1-3 show the concentrations of PCE and TCE at different sampling positions with
various pore volumes of the groundwater passing through the three columns. The results show
that (1) for the control column containing only silica sand, the concentrations of PCE and TCE
remained relatively constant over the entire length of the column, with the influent concentration
very similar to the effluent concentration; (2) for the 50% iron/sand column, the observed decline
in concentration with distance reflects that both PCE and TCE were degraded gradually along the
column, at a distance between 15 and 20 cm from the influent, both compounds had degraded to
below the detection limit (2.5 ug/L for PCE and 3.0 ug/L for TCE); (3) for the column with 100%
granular iron, rapid declines in concentrations of PCE and TCE were observed, at a distance
between 2 and 10 cm from the influent, both compounds had degraded to below the detection
limit. The production of small amounts of 1,1-dichloroethene (1,1-DCE), cis-1,2-dichloroethene
(cDCE), and vinyl chloride (VC) were observed in both reacting columns. However, 1,1-DCE
disappeared rapidly in both columns, cDCE and VC were eventually dechlorinated in the 100% iron
column. They were not completely removed in the 50% iron/sand column.
TABLE 1
CONTROL COLUMN (100* SILICA SAND)
Distance Along Column (cm)
Influent
0 2.5
PCE
TCE
Pore
Volume
1.03
2.15
5.00
1.03
2.15
5.00
Organic
12394
11031
8305
508
432
298
5
10
15
20
30
Effluent
40 50
Concentration ,(ug/L)
12272
9978
8867
465
442
273
11863
9798
9640
490
334
305
2914
8709
9241
319
317
291
9997
8050
9213
444
384
293
10718
9932
9349
404
415
304
8896
10947
9012
400
423
287
2858
11504
10196
242
437
330
7709
10061
8841
234
387
273
Flow Velocity = 44 cm/day (1.44 ft/day)
Porocity = 0.39
75
-------
TABLE 2
REACTIVE COLUMN I (50/50 Iron/Silica Sand)
Distance Along Column (cm)
Influent
0
2.5
10
15
20
30
40
Effluent
50
PCE
TCE
Pore
Volume
1.25
6.73
11.28
15.73
21.67
24.95
31.00
33.86
38.81
40.92
1.25
6.73
11.28
15.73
21.67
24.95
31.00
33.86
38.81
40.92
Organic Concentration (ug/L)
11882
9327
9438
9434
8934
4593
4360
5062
4846
4316
462
312
379
390
361
1160
1042
986
1136
938
73
5317
7037
6646
4188
3073
1630
3114
2842
2406
133
704
589
564
495
730
374
621
534
400
0.7
2790
3781
3712
2480
1851
2000
1681
1805
1314
0
748
636
656
397
500
516
365
324
213
0
0
0
47
249
285
388
302
486
379
0
0
90
127
143
132
157
108
141
105
0
0
0
0
0
0
0
0
1
3
0
0
0
0
0
0
4
4
12
12
Flow Velocity = 55 cm/day (1.8 ft/day)
Porocity = 0.34
The test results show that the degradations of both PCE and TCE follow first order kinetics.
Therefore the half lives of these two compounds in the two reactive columns were calculated
according to the equation:
t1/2 = 0.693/k (1)
where t1/2 is the half life (the time at which the initial concentration, C0, declines by one-half), k is the
first order rate constant which is calculated according to the following equation:
In (C/C0) = -kt
(2)
Where C is the concentration in solution at time t. From the experimental results, the half lives of PCE and
TCE in the reaction columns are calculated to be as follow:
(1) For 50* iron/sand column - PCE:0.67 hrs, TCE:1.10 hrs
(2) For 100% iron column - PCE:0.42 hrs, TCE:0.52 hrs
Demonstration Using Above Ground Reactor
The sampling was started after about 3 pore volumes of water had passed through the reactor.
The water samples were collected about once per week. So far, four sets of samples have been
collected and analyzed. The results are shown in Table 4. In the first three sets of samples, all PCE
and TCE were completely dechlorinated after passing through 12" of iron. No chlorinated
hydrocarbons were observed (detection limit for all the probable products was 1.0 ug/L). In the fourth
set of samples, the concentration of PCE was very high (13000 ug/L), so it did not completely
disappear until the solution passed through about 24" of the reactive iron. The quantities of TCE
increased first and then disappeared after the solution had passed through 24" of iron.
76
-------
TABLE 3
REACTIVE COLUMN II (100% Grandular Iron)
Distance Along Column (cm)
•• i<
Influent
0 2.5
Pore
Volume
PCE 1.01
4.47
8.73
13.88
19.25
21.79
27.19
29.62
33.02
34.85
TCE 1.01
4.47
8.73
13.88
19.25
21.79
27.19
29.62
33.02
34.85
5
10
15
20
30
Effluent
40 50
Organic Concentration (ug/L)
1 1963
8227
9602
9847
7610
4535
4476
5190
4241
4407
452
282
336
384
321
1222
1069
1005
1020
998
0
1
462
1532
1136
923
1109
865
682
926
0
6
391.7
541
349
356
380
298
169
262
0
0
0
0
6
14
59
57
67
96
0
0
0
10
29
26
54
36
35
46
0
0
0
0
0 •
0
0
0
0
0
0
0
0
0
0
0 "
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0
o
0
o
o
o
0
o
0
o
o
o
0
Flow Velocity = 47 cm/day (1.54 ft/day)
Porocity = 0.40
TABLE 4
ABOVE GROUND REACTOR
From the Influent (inch)
21s 36
48
# The reaction with iron started at 24" from influent port
* This is the effluent solution
** Not determined
66
84
96
PCE
TCE
••
Pore
Volume
2.4
8.8
14.4
25.6
2.4
8.8
14.4
25.6
Concentration (ug/L)
6200
9800
9700
13000
110
69
130
110
6100
Nd**
Nd
8700
0
Nd
Nd
93
0
Nd "
Nd
4200
0
Nd
Nd
330
0
Nd
Nd
0
0
Nd
Nd
0
0
Nd
Nd
0
0
Nd
Nd
0
0
Nd
Nd
0
0
Nd
Nd
0
0
0
0
0
0
0
0
0
77
-------
The production of cDCE was observed between 0 and 12" below the iron surface. It was 92, 18 and
0 ug/L respectively at 12", 24" and 42" below the iron surface respectively. Small quantity of vinyl
chloride (4.9 ug/L) was observed at 24" which disappeared at 42". At this sampling point, 1.5 ug/L
of methylene chloride was observed which degraded to below 1.0 ug/L below the iron surface. The
reaction process appeared to follow sequential dechlorination from PCE to TCE to cDCE and then to
vinyl chloride and methylene chloride. However, all the chlorinated compounds were completely
dechlorinated at the end if there was enough reactive iron.
CONCLUSIONS
The results from the column tests show that the 100% iron column can dechlorinate PCE and
TCE much faster than the 50% iron/sand column. Also, the former column can completely
dechlorinate the by-products but not the later column. Therefore the 100% iron column was used in
the pilot scale experiment. The results obtained thus far showed that PCE, TCE and their chlorinated
products can be completely dechlorinated in the above ground reactor.
The potential limiting factors of this technology such as: the possible formation of precipitates
that might clog the treatment reactor's pore space and/or coat the iron surface and thus limiting water
flow or reducing the reactivity of iron; the consumption of iron by its reaction with the chlorinated
compounds to produce soluble iron compounds; and the increasing pH over time and thus impeding
the reaction rate, are being studied. The approximate capital as well as the short and long term
operating and maintenance costs are being evaluated.
The above ground reactor may replace the air stripper and activated carbon as methods of
remediating halogenated organic compounds. The in situ permeable wall may permit passive treatment
for a long period of time and thus eliminating the disposal or treatment of the pumped groundwater.
Acknowledgments
The author wishes to acknowledge S. Krishnamurthy of the United States Environmental
Protection Agency (EPA) for his helpful discussion and critical review. He also wishes to thank Kim
Lisa Kreiton of PRC Environmental Management, Inc (PRO for her efforts in collecting and calculating
the data in the above ground reactor demonstration and William Matulewicz of Rhode Engineering for
his providing of the column test results. This research work was funded by EPA under Contract No.
68-CO-0047 to PRC.
REFERENCES
1. Gillham, R.W., and O'Hannesin, S.F. Enhanced Degradation of Halogenated Aliphatic by Zero-Valent
Iron, Ground Water 32, 958-967, 1994. ".
2. Superfund Innovative Technology Evaluation Program, Technology Profiles, Seventh Edition,
EPA/540/R-94/526, Office of Research and Development, Washington, DC, 1994, 174-175.
78
-------
TESTING THE SONOTECH PULSE COMBUSTOR IN A PILOT-SCALE ROTARY KILN INCINERATOR
Marta K. Richards
united states Environmental Protection Agency
Office of Research and Development, Risk Reduction Engineering Laboratory
26 w. Martin Luther King Drive
Cincinnati, Ohio 45268
513/569-7692
Larry R. waterland
Acurex Environmental corporation
incineration Research Facility
Jefferson, Arkansas 72079
501/541-0004
Kenneth c. Partymiller
PRC Environmental Management, inc.
5326 Paris Pike
Georgetown, Kentucky 40324
502/867-1397
INTRODUCTION
This is a report on an incinerator performance test of an innovative soundwave-based
combustion burner system that was patented, designed, and fabricated by its developer -
sonotech, inc., of Atlanta, Georgia. The superfund innovative Technology Evaluation (SITE)
program Demonstration of the sonotech, inc., tunable-pulse combustion burner technology
was conducted in the fall of 1994 at USEPA'S incineration Research Facility (IRF) in Jefferson,
Arkansas. Sonotech claims the technology will provide benefits when applied In a variety of
combustion processes. The burner system Incorporates a natural-gas-ftred burner, the pulse
frequency of which can be tuned to induce large amplitude sonic pulsations inside a
combustion process unit, such as an incinerator, sonotech claims these pulsations will
increase the efficiency of a combustion process by promoting better mixing of the
combustion gases, in the SITE Demonstration of its technology, sonotech's technicians
retrofitted a pulse combustion burner in the government's pilot-scale rotary kiln
incineration system (RKS) housed at the IRF.
Sonotech claims their burner significantly improves the performance of the incineration
process by increasing the rates of mixing (i.e.,momentum) and mass and heat transfers
within an incinerator. These, in turn, increase the incineration rate, reduce the amount of
air required for incineration, reduce the severity of puffs, and reduce pollutant emissions.
The developer claims these improvements reduce capital investment and operating costs in
a wide variety of incineration systems and improve their performance.
TEST OBJECTIVES
-------
increased incinerator capacity or productivity
increased principal organic hazardous constituent (POHO destruction and
removal efficiency (ORE)
Decreased flue gas carbon monoxide (GO) emissions
Decreased flue gas nitrogen oxides (iMOx)-emissions
Decreased flue gas soot emissions ..
Decreased combustion air requirements
Decreased auxiliary fuel requirements
Test data were also developed for the secondary Test program Objectives, which were to
allow evaluating whether the application of the sonotech technology, when compared to
conventional, non-pulsating combustion:
• Reduced the magnitude of transient puffs of co and total unburned
hydrocarbon (TUHO
• Allowed reduced incineration costs
• caused significant changes in:
— Distribution of hazardous constituent trace metals among the
Incineration system discharge streams (kiln bottom ash,
scrubber liquor, baghouse flyash, and baghouse exit flue gas)
— The mobility of the toxicity characteristic leaching procedure (TCLP) trace
metals from kiln bottom ash, scrubber liquor, and baghouse flyash
The last secondary objective does not relate to any sonotech claim, however, fates of
metals and their teachabilities are of general interest to the EPA IRF research program.
TESTING PROTOCOL
The waste feed for all tests was a mixture of contaminated materials from two
manufactured gas plant (MOP) superfund sites, one site was the Peoples Natural Gas
Company (coal gasification) site In Dubuque, lowa, and the other was an oil gasification site
In the southeast U.S. The final waste feed was composed of coal and sludge from the
Peoples site mixed with the contaminated soil borings and tar from the oil gasification site.
The final composite feed contained polynuclear aromatic hydrocarbons (PAHS) and other
semlvolatile organic compounds (svocs), benzene and other volatile organic compounds
(VOCs), metals, and other contaminants, it was determined that, to guarantee meaningful
calculation of DREs of specified POHCS,, spiking of the waste feed with benzene and
naphthalene was necessary. The spiked composite feed contained 13,500 mg
naphthalene/kg feed and 9000mg benzene/kg feed. ,
Twelve field incineration test runs were conducted. Four incinerator system operating
conditions - each In triplicate - addressed the Demonstration objectives. The four
conditions were:
• conditions conventional combustion at baseline/typical operation
• condition 2: conventional combustion at its.maximum feedrate
• conditions: sonotech combustion at the same feedrate as condition 2
• Condition 4: sonotech combustion at its maximum feedrate
This testing program would allow comparison of sonotech performance with conventional
combustion and permit evaluation of the developer performance claims.
80
-------
Table 1 shows the samples that were taken and the analyses that were performed.
TABLE 1. SAMPLING AND ANALYTICAL TEST PROGRAM
samples \ Analyses
Feed
Kiln Ash
Scrubber Liquor
Baghouse Ash
AB Exit Partial late
BH Exit Flue Gas
AB Exit Flue Gas
SVOCS VG
cs Metals PCDDS/PCDI
X X X
X XX
X X X
X XX
X
:S TOC TCLP
X
X
X
X
x
XXX
XX X X
Note: svocs = semivolatile organic Compounds
vocs = volatile organic compounds
PCDDS/PCDFS = Polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans
TOC = Total organic carbon
TCLP = Toxicity Characteristic Leaching Procedure of metals
AB = Afterburner
BH = Baghouse ;
TO determine if performance satisfied the test objectives, continuous emission monitors
(CEMs) constantly measured the oxygen (02), CO, carbon dioxide (C02), Mox, and TUHC levels in
the flue gas streams.
The test matrix was designed to give data that allowed evaluation of the vendor claims as
follows. Test condition 1 provided emissions, POHC ORE, metals partitioning, and metals
leachability data corresponding to baseline, typical, well-controlled incinerator operation.
Test condition 2 provided the same kinds of data for operation corresponding to maximum
feedrate possible under conventional Incinerator operation, which is borderline acceptable
operation approaching noncompliance with typical permit limits. Test condition 3 provided
similar data corresponding to test condition 2, except with the sonotech unit running,
while test condition 4 corresponds to the maximum feedrate possible with the sonotech
pulse combustor operating.
RESULTS
preliminary results of the test program related to the primary test objectives are
reported here. Test condition 3 data are used to evaluate the last six primary objectives by
comparing results with the those obtained under test conditions 1 and 2 data, comparing
test condition 4 data to those from test conditions 1,2, and 3 allowed evaluation of the first
primary objective. The numerical-average results are shown in Table 2.
81
-------
TABLE 2. OPERATING DATA AND RESULTS BY TEST CONDITION
Parameter
waste feedrate,
Ib/hr
Kiln Gas temperature, °F
Afterburner gas temperature,1^
Afterburner exit 02, %
Test Conditions*
Conventional Combustion Pulse
1: 2:
61.0 72.8
1,720 1,730
2,000 2,000
9.33 9.27
3:
73.6
1,700
2,000
8.73
Combustion
4:
82.4
1,700
2,000
8.51
Heat input, kBtu/hr
waste
Auxiliary fuel
Total
combustion air,
dscf/hr
Afterburner exit co, ppm @ 7% 02
Afterburner exit NOX, ppm @ 7% 02
Afterburner exit soot emission
rate rroc in particulate, %)
517 617
1,670 1,540
2,190 2,160
37,300 36,800
15.2 20.3
90.0 81.7
<1.0
-------
5. Average soot emissions, measured by TOC analyses of tne afterburner exit flue gas
particulate and comparing soot emission rates, were lower with pulse combustion than
conventional combustion at equivalent feedrates.
6. combustion air requirements, determined from stoichiometric calculations, showed
small decreases with pulse combustion.
7. Fuel requirements for all four test conditions were statistically similar.
For the secondary test program objectives:
1. The frequency of transient puffs of CO decreased with the pulse combustor in
operation.
2. The average normalized distribution of the trace metals did not vary significantly from
test to test.
3. The effect of the sonotech combustor on incinerator residue toxicity characteristics, as
measured by the TCLP, was inconclusive due to very low levels of metals in the kiln ash and,
therefore, the leachate.
CONCLUSIONS
The most important difference measured with the sonotech combustor in operation
was the ability to increase the capacity of the EPA RKS. This increased capacity resulted in
the same or slightly reduced CO, NOX, and soot emissions. Additionally, the benzene and
naphthalene ORES were equal or marginally improved with the sonotech burner in
operation. The increase in incinerator capacity was also realized without increasing fuel or
air consumption to the system.
FOR MORE INFORMATION:
Marta K. Richards
united states Environmental Protection Agency
Office of Research and Development, Risk Reduction Engineering Laboratory
26 w. Martin Luther King Drive
Cincinnati, Ohio 45268
513/569-7692
83
-------
THE RELATIVE EFFECTIVENESS OF MINERAL-BASED SORBENTS FOR
METAL CAPTURE IN A BENCH-SCALE INCINERATOR
Gregory J. Carroll
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
513/569-7948
Shyam Venkatesh
Acurex Environmental Corporation
c/o NCTR, Bldg. 45
Jefferson, AR 72079
501/541-0004
INTRODUCTION
Given the concern over the emission of hazardous constituent trace metals from incinerators,
there is currently considerable interest in the potential use of mineral-based sorbents for capturing and
retaining those metals in the incinerator "ash" discharges (fly ash and bottom ash).
Most of the research completed to date has focussed on quantifying the effectiveness of
various proposed sorbents for capturing vaporized metals from the flue gas. In such applications, it is
theorized that vaporized metals will react with the sorbent particles at elevated incinerator temperatures
or heterogeneously condense onto the sorbents as the flue gas cools. In the absence of available
condensation sites, vaporized metals will primarily undergo homogeneous condensation, forming a fine
fume. Thus, the goal with flue-gas sorbents is to make particles available with which the metals can
react or upon which the metals can condense. Metals bound to larger sorbent particles will be more
effectively collected by air pollution control systems (APCSs) than metals presented as a fine fume.
Studies completed to date suggest that chemical reaction between the metal and the sorbent dominates
over physical adsorption, offering the additional advantage of reduced potential for metal leaching from
collected particulate.
Other researchers have studied the incorporation of sorbents into the solid feed. This approach
seeks to capture and bind the metals in the incinerator bottom ash, preventing them from exiting with
the combustion gases. Research completed to date suggests that for this approach to be effective, the
metal should become volatile in the incinerator environment and chemically react with the sorbent
material.
The subject test program was designed to further investigate this second approach by
screening several minerals for their suitability as sorbent materials for capturing metals in the solid bed
and preventing their release to the flue gas. In addition to capturing the metals, an ideal sorbent would
retain them in the ash when disposed, so that extraction of the ash by the toxicity characteristic
leaching procedure (TCLP) would yield a leachate with metals concentrations below respective
regulatory levels. Accordingly, the objective of this screening program was to evaluate several
candidate sorbents with respect to: [1] the degree to which they facilitate retention of trace metals in the
ash/solid bed discharged from an incinerator; and [2] the degree to which they retain trace metals in the
solid bed when subjected to TCLP extraction.(l)
METHODOLOGY
Test Equipment
The screening tests comprising this program were conducted in the bench-scale thermal
84
-------
treatability unit (TTU) at the U.S. EPA Incineration Research Facility (IRF). The TTU is a small
commercial pathological incinerator that has been modified to allow: continuous test material feed and
treated material (e.g., ash) removal; variable treatment temperature control; and expanded process
operation monitoring. .
The combustor portion of the TTU consists of three chambers. The charge chamber is
designed to accept the solid material feed stream and corresponds to the primary combustion chamber
(or kiln portion) of a waste incinerator. The retention chamber, which directly follows the charge
chamber, is designed to effect further organic constituent destruction and corresponds to the secondary
combustion chamber (or afterburner) of a waste incinerator. The breaching chamber serves as a
second-stage afterburner. Volumes of the charge chamber, retention chamber, and breaching chamber
are 0.82 m3 (29 ft3), 0.67 m3 (23.5 ft3), and 0.10 m3 (3.5 ft3), respectively. Each of the three chambers
is fired with natural-gas-fueled burners.
Test Program
The test program consisted of 50 tests, including two duplicates of one test condition. Test
variables were sorbent material type, solid bed temperature, feed chlorine content, and metal form in
the feed.
Six sorbents were evaluated. The first five (silica, diatomaceous earth, kaolinite, bauxite, and
alumina) were selected based on the most promising results from other researchers and make up a
sorbent-material spectrum ranging from pure silica to pure alumina. The attapulgite clay used in past
IRF trace metals studies was tested as the sixth sorbent.
The approximate mineral content of diatomaceous earth, kaolinite, and bauxite are given in
Table 1. Silica is presumed to be pure SiO2; alumina is presumed to be pure AI2O3; and the attapulgite
clay is a hydrated magnesium aluminum silicate [(Mg,AI)5Si8O22(OH)4-4H2O] containing some dolomite
[Ca,Mg(CO3)2], calcite [CaCO3], and silica.
Three solid bed temperatures were tested: 540°C, 700°C, and 870°C (1000°F, 1300°F, and
1600°F). Two feed chlorine contents were tested: 0% and 4% by weight. Polyvinyl chloride powder
was added to the mixtures as the chlorine source.
TABLE 1. APPROXIMATE SORBENT MINERAL COMPOSITION ••-,-;,
Si02
AI203
Fe203
TiO2
CaO
MgO
Other oxides
. Diatomaceous earth
90.4
6.5
2.3
'
0.2 »
0,3
0.3
Kaolinite
52
45
1
-. - • 2
.....
'
— '
Bauxite
11
84
, ; '.-•> 5 '.''.,
ri
' ~ - < mm
'
Sorbent impact on the retention of the following trace metals was evaluated in the program:
arsenic, cadmium, chromium, lead, and nickel. Two forms of incorporating the metals into the feed
mixtures were evaluated. Past trace metals tests at the IRF have used aqueous metal spike solutions
containing soluble nitrate salts of the metals (with the exception of arsenic which has been added as
85
-------
For continuity, this form was one of the two used in the tests.
The second form of metal spiking was a metal compound "dispersion". The dispersion
consisted of metal compound powders suspended in a liquid carrier analogous to pigments dispersed in
paint or ink. Chromium, cadmium and lead were present as metal oxides in the dispersion, while
arsenic and nickel were introduced as sulfides and carbonates, respectively. Because of a difficulty
achieving homogenous feed mixtures with the metal compound dispersion, the discussion of results in
this abstract is limited to the 36 tests in which the aqueous metal spike solution was used.
For each test, weighed amounts of the appropriate mixtures of sorbent, PVC (for tests with
chlorine-containing feed), and metal spiking formulations were added to a TTU quartz tray. Charge
depth was held constant at nominally 2 cm, corresponding to a charge volume in the tray of
approximately 300 cm3. Target feed metal concentrations were also constant as follows: arsenic (200
mg/kg); cadmium (50 mg/kg); chromium (150 mg/kg); lead (250 mg/kg); and nickel (150 mg/kg).
Charge mass ranged from 50 g to 320 g, depending on sorbent bulk density.
Tray residence time in the TTU was approximately 20 minutes at the target solid bed
temperature. Based on the results of other research, no further reaction between the sorbent and
metals was expected after the first 10 minutes at target temperature.
Sampling and Analysis
For each test, TTU gas temperatures and solid bed temperatures were measured. Samples of
unspiked sorbent, metal spiking solution, TTU feed, and TTU discharge were collected for metal
analyses. Additionally, one TTU feed sample from each of the sorbent/metal formulation combinations
and TTU discharge samples from every test were subjected to TCLP extraction, and the resulting
leachates analyzed for trace metals.(2) Quality assurance samples were prepared as well.
Aqueous liquid samples were digested using EPA Method 3010.(3) Solid samples were
digested using a microwave-assisted HNOg/HF procedure,(4) Analyses of each of the latter digestates
for arsenic were by Method 7060 [GFAA].(3) All other digestates analyses were by Method 6010
[ICAP].(3)
RESULTS
Feed Samples
Despite the use of a rigorous digestion method, analyses of feed samples often yielded
concentrations below those expected (based on prepared metal spike solutions). Of the five metals,
cadmium and lead exhibited the largest difference between target and measured concentrations. On
average, metal concentrations in the attapulgite clay and kaolinite were below target more so than the
other matrices.
Since the same preparation and analysis methods were applied to both the feed and treated
material samples, it is assumed that any sampling or analysis bias should manifest itself in both sample
types (i.e., the feed and treated material associated with a particular sorbent). Therefore, comparisons
between the feed samples and treated material were based on measured (as opposed to prepared)
feed concentrations.
Retention of Metals in Ash
Metal volatility was judged by examining treated material (ash) concentrations relative to feed
concentrations (adjusted for loss of mass during treatment); the greater the reduction in concentration,
the greater the volatility. On average, there was little difference among the sorbents for limiting the
86
-------
volatility for chromium and nickel. With no chlorine in the feed, there was also little difference between
the sorbents for cadmium. With 4% chlorine present in the feed, cadmium appeared to be less volatile
from kaolinite, relative to the other sorbents. Finally, for the tests both with and without chlorine in the
feed, arsenic appeared to be less volatile from the attapulgite clay and lead less volatile from kaolinite,
compared to the other sorbents.
In addition to the overall differences noted above, changes in temperature affected the retention
of metals in several of the sorbents when chlorine was present in the feed. Arsenic volatility in the
silica, diatomaceous earth, and alumina matrices increased with increasing temperature, while the
reverse occurred in the attapulgite clay. Cadmium volatility in diatomaceous earth, alumina, and
bauxite increased with increasing temperature. Chromium volatility was relatively constant across the
temperature range. Lead volatility increased with increasing temperature for the alumina and bauxite
matrices. Nickel volatility in the attapulgite clay decreased with temperature.
In cases where chromium appeared to impact metal behavior (independent of temperature), it
consistently increased the volatility of the metal. Such increases were seen for arsenic in bauxite;
cadmium in silica and attapulgite clay; and lead in silica. Chromium and nickel volatility did not appear
to be impacted by chlorine in any of the matrices.
Resistance to Leaching by TCLP
"Fractional teachability" represents the ratio of the TCLP-extracted metal to the total metal
measured in the sample. In nearly every case, teachabilities from treated samples were less than the
corresponding feed samples. Attapulgite clay, bauxite, kaolinite, and alumina were better for limiting
arsenic teachability during tests with and without chlorine in the feed. With no chlorine in the feed, all
of the sorbents had similar teachabilities for cadmium and lead. With feed chlorine, attapulgite clay,
kaolinite, and diatomaceous earth were better for limiting cadmium and lead teachability, as was bauxite
for cadmium teachability. With one exception, all of the sorbents had similar chromium teachability;
chromium was very easily leached from the treated attapulgite clay samples. Nickel teachability did not
differ significantly among the sorbents for the tests with chlorine in the feed. With no chlorine in the
feed, alumina, attapulgite clay, bauxite, and kaolinite were better at limiting nickel teachability.
Increases in temperature consistently led to decreased teachability of cadmium from each of
the sorbents and of arsenic, chromium, lead, and nickel from diatomaceous earth. Arsenic teachability
from the other sorbents was not affected by temperature. When chlorine was not present in the feed,
chromium teachability from silica and alumina decreased with increasing temperature, as did lead
teachability from kaolinite and nickel teachability from silica. With chlorine in the feed, lead teachability
from silica, and nickel teachability from silica, alumina and bauxite decreased as temperature rose.(5)
CONCLUSIONS
Given the screening nature of these tests, and considering the analytical problems encountered
with many of the feed samples, the results of the tests should be used with caution. Nevertheless, a
number of preliminary conclusions may be made. Combining the two sorbent criteria of limited metal
volatilization and reduced teaching by TCLP, kaolinite and attapulgite clay appear to be the most
promising sorbents for arsenic; kaolinite for cadmium; kaolinite, diatomaceous earth, and attapulgite
clay for lead. With few exceptions, the sorbents showed comparable performance for the other
matrix/metal combinations; chromium teachability appeared to increase in the attapulgite clay matrix
with increased temperature.(5)
REFERENCES
1. Acurex Environmental Corp. Quality Assurance Project Plan for Evaluating the
Effectiveness of Additives as Sorbents for Metal Capture Using the Thermal Treatability
87
-------
Unit. Contract No. 68-C9-0038, U.S. Environmental Protection Agency, Cincinnati, Ohio,
1993.
2. Method 1311 Toxicity Characteristic Leaching Procedure (TCLP). 40 CFR Part 261, Appendix
II.
3. Test Methods for Evaluating Solid Waste: Physical/Chemical Methods.
SW-846, 3rd Edition, Revision 1, U.S. Environmental Protection Agency, 1992.
4. Methodology for the Determination of Metals Emissions in Exhaust Gases from Hazardous
Waste Incineration and Similar Combustion Processes.
40 CFR Part 266, Appendix IX.
5. Acurex Environmental Corp. Evaluation of the Effectiveness of Additives as Sorbents for Metal
Capture Using a Thermal Treatabilrty Unit - Draft Report. Contract No. 68-C9-0038, U.S.
Environmental Protection Agency, Cincinnati, Ohio, 1994.
FOR MORE INFORMATION
Gregory J. Carroll
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory (ML 481)
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
513/569-7948
88
-------
RESEARCH ON THE LOW-TEMPERATURE DECONTAMINATION OF SOIL
Melanie D. Hetland and John R. Rindt
University of North Dakota Energy & Environmental Research Center
PO Box 9018
Grand Forks, ND 58202-9018
(701) 777-5000
INTRODUCTION
Decontamination of organically contaminated soil requires at least two processes:
1) separation of the contaminants from the soil and 2) destruction or stabilization and recovery of
the organic contaminants. Techniques used to convert the organic contaminants must produce
environmentally acceptable products. A process is seen as more favorable if the inorganic portion
remaining after destruction of the organic contaminants retains a "natural" state, since ideally the
residual should be returned to the original ecological system.
Low-temperature plasma (LTP) processing has the potential to offer decontamination
equivalent to incineration with decreased production of harmful by-products and without changing
the morphology of the inorganic matrix. Low-temperature plasmas are formed by electric discharge
(rather than thermal means) at conditions that do not result in thermal equilibrium. It is possible
with an LTP system for typical plasma reactions to occur at nearly ambient bulk temperatures.
The purpose of this project is to perform the appropriate engineering evaluations needed to
scale up an LTP soil decontamination process developed at the Energy & Environmental Research
Center (EERC). The LTP process under development uses oxygen plasma, resulting in oxidation of
the contaminant species. Previous research at the EERC proved that the technique successfully
separates organic contaminants from soil. The success of this previous research indicated that the
use of LTP as a decontamination technique should be investigated further, with possible scaleup to
real-world application.
METHODOLOGY
A program plan consisting of four phases was developed to enable the design and
construction of a real-world-scale prototype unit. The Phase 1 work included operation of the
existing EERC LTP unit with integrated gas analysis to verify that the fate of all organic species
placed in the reactor could be determined. Phase 2 consists of gathering the engineering
information needed for prototype design, while Phase 3 consists of the design, fabrication, and
shakedown testing of the prototype. The design, fabrication, and integrated shakedown testing of
any add-on equipment necessary for decontamination using the prototype constitute the Phase 4
work effort. This paper summarizes the testing performed during the Phase 1 carbon fate study.
The Phase 1 testing was performed on the EERC bench-scale LTP system. The system
consists of a radio frequency (rf) generator, a reactor center, a vacuum system, a feed gas system,
and an elaborate product gas collection system. The rf generator and the reactor center were
manufactured by the International Plasma Corporation. The quartz reaction chamber is 8 inches
long and 4 inches in diameter and has been modified to contain the hardware necessary for
agitation during operation. A pair of rotating aluminum trays was used to agitate the sample during
operation. The trays each have a lip on one longitudinal edge and are aligned so that, when turning
on the rotary coupling, the sample falls from one tray to the other through the plasma. A vacuum
pulls the plasma-producing gas through the chamber, while the rf energy is transmitted across the
electrodes surrounding the sides of the reaction chamber.
89
-------
The product gas collection system was designed and fabricated for use in all phases of the
project. The system consists of a bypass loop, two product gas collection bags or bladders
contained in two vessels, and three syringes that are used to transfer the gas from the collection
bags to the analytical system. An on-line gas chromatograph (GO was available for analysis of the
product gas stream during the Phase 1 testing.
A statistically designed experimental matrix was used to gather data concerning the fate of
the carbon species. The matrix was designed to vary five independent variables: oxygen flow
rate, reactor pressure, rf power, reaction time, and percent carbon in sample. The statistical
analysis of the data will indicate which operating parameters affect contaminant removal and
destruction efficiencies and how these efficiencies are affected. The matrix was designed so that
the test points would be performed in random order to minimize skewing of data that can occur
when one variable is held constant for several tests in a row. Tests were performed in blocks to
determine whether significant changes occurred in the system over the course of the testing.
All tests were performed using decolorizing carbon as the contaminant source and washed,
60- to 100-mesh sand as the inert matrix. Pure oxygen was used as the plasma-producing gas.
The use of carbon and pure oxygen limited the various products of incomplete combustion (PICs)
that could be produced, thereby greatly simplifying the product gas analysis. Both the feed and
product solid samples were analyzed for carbon, hydrogen, nitrogen, sulfur, and ash content. Feed
and product gas samples were analyzed using gas chromatography.
RESULTS
The purpose of this study was not to maximize contaminant removal and destruction
efficiencies, but rather to lay the foundation for the testing that will take place during the remaining
phases. The goals of the Phase 1 testing included verification that:
• Mass balance closures of ±5% are possible.
• The analytical scheme that is in place is capable of accurate results.
• The fate of the contaminant species can be accurately tracked.
• PICs can be identified and quantified with the existing product gas collection system and
the on-line GC.
A mass balance closure criterion of ±5% was chosen as a project quality assurance
objective because data gathered from tests with worse mass balances would not accurately
represent the tests. Therefore, as a first step, a mass balance was calculated for each test, and
those tests not meeting the criterion were repeated. A total of seven tests were repeated based
upon initial mass balance calculations. A change in the mass balance calculation was made to
account for minor contamination by air, resulting in only two tests that did not meet the mass
balance closure criterion. The results of the Phase 1 testing are summarized in Table 1. Table 1
clearly shows that mass balances meeting the criteria were easily obtained. The average mass
balance was 100.46%.
Quality assurance (QA) check samples were randomly submitted to verify that the
analytical results accurately described the samples. Analyses of both the solid and gaseous
products agreed with the calculated values well within the error range of the analytical
instrumentation. Average differences between the calculated values and the analyzed values for
the solid samples were +0.05 wt% carbon, -0.004 wt% hydrogen, -0.002 wt% nitrogen,
+0.084 wt% sulfur, and -0.20 wt% ash. Instrument error is ±1.0 wt% for ash and ±0.5 wt%
for the other components. The gas analysis checks were equally accurate, with average
90
-------
CO
LU
O
CO
CC
OOOOOOC33OOOO)OCJlOOOO>OcnOO05O
J_ CO O
o j-5
• O C-
a.
i_ 03
O b: —
4-1 13 -
«o w 2
co CD ii
.1 CD-P
4.1 ^ L_
o E-?
CC
c
c co CD
LOLOLOOLOLOLOLOOtOLOOLOOO06OOOOOggLOLC5LOLOLOOOLO
cNir^oi^i^^cMOCMc^orvooooLoooiflOOor^eM CM i^. CN o o CM
LOIOLO lOlOLOLO LOLOLO
LotnintOLO
LOLOLOLOlLO
LOLO uo in LO in LO in in
co cdLOT-cdcdiricd^-cdcDr-LO1
COlriLOLOLOr-Olri
co co co co co co co co r>r>«i^«rs«oooocooocoooo5O5O>cno)O)coco'*^!-<5i''*^r
LOLOininir)iOLomif>LomLnLOLOLOiOLOir5LnmLoir>LOLOcococococococp
0-CLa.Q.a-aLQ-Q.a.CLQ.a.a.Q-a.a.D-CLQ.a.Q-DLQ-Q-Q-Q.a.a.D-Q-D.Q-
x
.«
CO
E
CD
U3
CO .
d1.!
to to
T3 3
c w
CO "co
15
E.tt
+3 £
CD O
<-• 03
m "5.
C |
O CO
ft
»«. CO
CO CO
o ra
C3) CO
'fj "co
'>^ P
o £
"to ID
CD a)
E •-
S c"
•2 E
Q. Q)
CD £2
= °
< CO
91
-------
differences less than 1 mol% for ail components. Instrument error for the GC is ±2 mol%. The
GC detection of water in the calibration gas is an analytical anomaly, the cause of which has yet to
be defined but should diminish as the project progresses to larger-scale equipment and samples.
The results summarized in Table 1 show that it is possible to accurately track the fate of
the carbon species in the LTP system. Carbon balances were calculated for each of the tests and
range from 92.0% to 149.5%. The mean carbon balance of 126.4% is close to the median value
of 128.3%, indicating that the values are tightly grouped and therefore representative. Although
these balances appear to be somewhat high when considering that a perfect balance would be
100%, all are within analytical instrumentation error of 100%.
The product gas composition was calculated for each test and is summarized in Table 2.
By far the most prevalent species found was oxygen. This is logical since excess oxygen was
provided for reaction. Carbon dioxide (the primary product of the reaction of oxygen and carbon),
carbon monoxide (the expected PIC), water vapor, and hydrogen were also found to be present.
The GC was calibrated to detect hydrogen sulfide and various hydrocarbon gases, but as none
were detected these results were not included in the table. A minute amount of methane was
detected in the product of the test having the largest contaminant loading in the feed sample.
The reproducibility of LTP processing can be seen by comparing the product gas
composition of Tests P572, P584, and P594. These tests were performed at nominally the same
operating conditions and contaminant loadings. As Table 2 shows, the product gas compositions
of these tests are nearly identical, indicating that the results are reproducible.
Contaminant removal and destruction efficiencies were calculated for each of the tests,
even though it was not the goal of the Phase 1 testing to maximize these efficiencies. These
values are presented in Table 1. Removal efficiencies ranged from about 4% to about 76%, while
destruction efficiencies for the removed contaminants ranged from about 76% to about 97%.
These encouraging results imply that, when operating conditions are optimized, it should be
possible to meet the project goals of removal of at least 99 wt% of the contaminants and
destruction of 99.99 wt% of the removed contaminants.
The statistical analysis has not yet been completed. It is expected that the statistical
results will indicate which operating conditions should produce the highest contaminant removal
and destruction efficiencies. Such information will be used in upcoming project phases.
CONCLUSIONS
Several conclusions can be drawn from the results:
• It is possible to obtain mass balance closures of ±5 wt% with the existing LTP system.
• The analytical scheme currently in place accurately analyzes both solid and gaseous
products.
• The fate of the contaminant species can be tracked in detail.
• The gaseous products of all tests consisted primarily of excess oxygen, carbon dioxide,
and carbon monoxide.
• The tests are very reproducible.
In general, it can be said that both the mechanical and analytical aspects of the LTP system met or
exceeded expected performance standards based upon previous research at the EERC.
92
-------
TABLE 2. PRODUCT GAS SLATES FOR PHASE 1 TESTS
Test No.
P560
P562
P564
P565
P566
P567
P568
P569
P572
P575
P576
P577
P580
P583
P584
P585
P588
P589
P590
P591
P593
P594
P595
P596
P638
P639
P640
P641
P642
P644
P645
P647
O
(mol%)
71.2
77.2
89.9
85.2
87.9
91.5
88.1
70.3
75.8
81.4
88.2
91.5
86.3
80.0
79.3
99.6
84.3
92.9
85.3
80.0
70.3
78.5
71.9
68.5
83.2
70.4
79.7
93.0
74.6
83.6
83.6
73.9
CO,
19.1
16.4
7.6
11.2
9.5
6.3
8.2
20.4
15.3
14.7 '
9.4 .
7.5
9.7
13.5
16.3
0.0
12.6
5.6
12.8
16.2
20.9
16.4
19.6
22.0
10.9
18.9
13.2
3.9
14.9
10.5
10.5
17.6
H
1.0
0.8
0.8
0.7
O.3
0.5
0.6
1.4
0.9
0.7
0.5
0.6
0.7
0.5
0.7
0.0
0.7
0.4
0.4
0.4
0.9
0.5
0.5
1.2
0.7
0.6
0.5
0.5
0.6
0.4
0.4
0.7
H2O
2.4
2.1
1.3
1-7
1.7
1.4
2.4
1.6
2.6
1.9
1.4
0.0
1.1
0.9
0.6
0.4
0.8
0.9
0.8
2.1
0.9
1.4
1.0
0.8
2.7
3.2
4.7
2.3
2.9
4.8
4.8
2,9
CO
6.3
3.5
0.4
1.3
0.6
0.3
0.8
6.3
5.4
1.4
0.5
0.4
2.2
5.1
3.1
0.0
1.7
0.2
0.6
1.3
7.0
3.2
7.1
7.5
2.4
7.0
2.0
0.2
7.0
0.7
0.7
4.9 ,
GH£
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
FOR MORE INFORMATION
For more information, please contact:
Dr. C.C. Lee, Technical Project Manager
U.S. EPA - Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513) 769-7520
93
-------
A SUMMARY OF TWO RECYCLING TECHNOLOGY EVALUATIONS
COMPLETED UNDER THE MITE PROGRAM
Lynnann Hitchens
Risk Reduction Engineering Laboratory
5995 Center Hill Road
Cincinnati, Ohio 45224
(513) 569-7672
INTRODUCTION
Through the Municipal Solid Waste Innovative Technology Evaluation (MITE) Program, EPA has
evaluated a number of promising innovative technologies for municipal solid waste (MSW) recycling.
These technologies have varied from collection and processing methods to evaluations of products made
from post-consumer material. The MITE program was conceived with the purpose of providing objective,
third-party evaluations on environmental performance and costs of innovative municipal solid waste
management technologies and this program is administered for EPA by the Solid Waste Association of
North America (SWANA), representing the public sector solid waste management needs. The
technologies are selected via an annual solicitation and review by an Advisory Group made up of recycling
coordinators and solid waste managers.
Once the technologies are selected and the evaluation is performed, the results are published in a
report that is distributed to the information users or the general public. The benefit for technology
developers in participating is that they receive objective, published information to assist them in the
commercialization and marketing of their technology.
One unique point separating this innovative environmental technology evaluation program from
others is that it is often the public sector (the municipality, county, or solid waste district) advancing MSW
technology through research and development, and partnerships with private companies and trade
organizations. A significant number of applicants to the MITE program have been cities or non-profit
groups. This is the case with the two evaluations profiled below.
This summary will address two evaluations focusing on the collection of municipal solid waste and
designs that foster cost-effective recycling and source reduction.
METHODOLOGY
The first project profiled is a weight-based municipal solid waste collection system. The second
project is an evaluation of three co-collection systems for solid waste and recyclables.
Weight-Based Collection Pilot in Farmington, Minnesota
The City of Farmington, Minnesota applied to the MITE program for an evaluation of their
proposed pilot for a weight-based MSW collection system. The purpose of this collection system is to
charge waste generators according to the weight of the solid waste placed at the curb for disposal. The
more waste generated, the higher the disposal costs would be. There is an economic incentive in
generating less waste for disposal and Farmington theorized that this type of system would increase the
amount of material set out for recycling, a service for which there is no charge, and also foster source
94
-------
reduction at the residential level.
In order to implement this type of collection system with a minimum additional collection cost, the
City attempted to retrofit a side loading collection vehicle to provide weight information for billing purposes.
This vehicle was already in service, providing fully automated solid waste collection. The retrofit had three
main components: the first consisted of load ceils mounted to the vehicle's gripper mechanism. These
load cells would weigh each refuse container during pick-up. The second component of the retrofit was
the use of a Trovan radio-frequency identification (RFID) transponder attached to each refuse container.
The transponder released coded information containing the address of the container when placed in
proximity of the antenna, mounted on the refuse collection vehicle. These two components were tied into
the third item, an on-board computer, which translates the information into an address identification and
container weight. Thus, the weight of the refuse is correlated to an address for billing purposes.
It was agreed that Farmington would obtain the vehicle retrofit and pilot test it prior to the MITE
evaluation. Unfortunately, there were many difficulties encountered in the development of this system. It
was discovered that the load cells were not durable enough for Farmington's collection needs. There was
also a problem with the accuracy of the weight measurements. The National Institute of Standards and
Technology (NIST) has established national weights and measure requirements for on-board weighing
systems. These systems have to meet Class III requirements, or an allowed tolerance level of accuracy
of one pound for the first 500 pounds, two pounds for 501 to 2,000 pounds, and three pounds for 2,001 to
4,00 pounds.
Through pilot testing it was determined that approximately 90 percent of all weights recorded on
the truck were within two pounds of accuracy. However, since full residential containers weighed less
than 500 pounds, the project was not consistently conforming to Class III standards. Accuracy testing
was performed by weighing the refuse containers first with a certified digital scale, then weighing the
containers with the on-board system.1
There were also other factors affecting the accuracy of the load cells. .Considerations such as the
slope of the road surface, excessive wind speed, and shifting weight due to liquids in the container caused
unpredictable errors in measurements. Because of noncompliance with the Class III standard, billing was
never permitted by the Minnesota State Department of Weights and Measures.
Based on EPA's assessment of this pilot, there are a number of recommendations that can be
made to municipalities and others attempting to conduct a weight-based collection pilot. Most important is
to establish performance specifications for system development. These allow more control that design
specifications. A second recommendation is to involve crew and staff in the development and
implementation of the system. This will allow a smoother transition during implementation. It is also
important to provide adequate training to staff and education to the waste generators.
The pilot, though not fully successful and not able to be fully tested, was a valuable learning
experience for Farmington and other municipalities and companies. In 1994, a number of private firms,
both here and in Canada made significant progress in the development of truck mounted systems for
weight-based collection for solid waste. In fact, a number of systems have applied to, and are expected to
receive NIST certification for their equipment, and allow public marketing of the systems. There is no
doubt that the research and development done by the City of Farmington contributed to this rapid
development.
Evaluation of the Weight-Based Collection Project in'Farmington, Minnesota: A MITE
Program Evaluation. EPA/600/R-94-164.
95
-------
Co-Collection of Solid Waste and Recyclables
At the present time, the majority of community recycling programs that use of curbside collection
utilize a separate vehicle in which to collect recyclables. These vehicles vary in design, depending on the
overall recycling system. If the recycling program makes use of a Materials Recovery Facility, or MRF, to
separate recyclables for market, the recyclables are often collected commingled in a single compartment
collection vehicle. If the program makes use of a curb sorting system, the collection vehicle can take on a
number of multi-compartment designs, depending on the amount and type of recyclables included in the
recycling program.
It has been shown in numerous cost studies that collection costs are a significant portion of the
total solid waste management cost. It has also been concluded in a number of instances that the "add-on"
of the curbside recycling collection - separate facilities, crews, and collection vehicles - may not be the
most efficient solid waste management system design. Simply speaking, all solid waste services should
be considered in an integrated manner, looking at long term program needs. In an attempt to minimize
collection costs, several communities across the United States are changing to a "co-collection" system,
where the solid waste and recyclables are collected simultaneously in the same vehicle.
There are several co-collection system designs. The first and simplest system uses a collection
vehicle (perhaps already in use for MSW only) with a single compartment. Recyclables and waste are
separately bagged by the waste generators, picked by concurrently, and placed in the same vehicle. The
bags of recyclables are separated from the bagged waste at a centralized facility. The second system is
the use of a vehicle with separate compartments, the number depending on the number of items that are
collected and the amount of curb-side sorting that occurs. The third unique system uses a dual
compartment refuse container designed and manufactured for use specifically with a dual compartment
automated collection vehicle.
Three separate systems will be evaluated in this project. They include a co-collection vehicle
manufactured by Oshkosh Truck Corporation as utilized by Durham, North Carolina; the Western
Curbside Collector co-collection truck manufactured by May Manufacturing Company, as utilized by
Loveland, Colorado; and Ruckstell California Sales, Inc. co-collection refuse container system as utilized
by Visalia, California.
One component of the evaluation consists of on-route data collection by the vehicle crew. Crew
training and all data management will be done by RW Beck and Associates. RW Beck will also perform a
sort of the recyclable stream for both the pre-existing two truck system and the co-collection system to
evaluate any changes in quality of recyclables. Information to be collected includes times on and off
route, time between stops, unloading and break time, the number of pick-ups, estimated distance between
stops, pounds collected, driver distance, and crew number. A number of different routes will be evaluated
over the two week data collection effort.
The information collected will be used for an economic analysis and a comparison of the quality of
the recyclables. Costs per household will be calculated for each of the systems and they will be compared
to the more traditional two vehicle systems for refuse and recyclable collection. There will also be an
attempt made to identify key factors that determine system cost effectiveness and efficiency.2
The co-collection system evaluations are presently underway and it is expected that draft results
will be available by May 1995.
.Co-Collection System Evaluation, Draft Scope of Services. RW Beck and Assoc. 10/94.
96
-------
CONCLUSIONS
US EPA's MITE Program, in cooperation with its partner, the Solid Waste Association of North
America (SWANA), seeks to perform objective assessments of innovative technologies for solid waste
management systems. Through these efforts, information on environmental effectiveness and costs on
these technologies is spread to the user community, in the hopes that it will foster environmentally and
economically sound municipal solid waste management.
REFERENCES
1. RW Beck and Associates. Draft Scope of Work: Co-Collection System Evaluation,
October 24, 1994.
2. SCS, Engineers. Evaluation of the Weigh-Based Collection Project in Farmington, Minnesota: A
MITE Program Evaluation. EPA/600-R-94/164, U.S. Environmental Protection Agency,
Cincinnati, Ohio 1994. 17pp.
3. Steuteville, R. Early Results with Co-Collection. Biocycle. February 1993.
For more information contact: Lynnann Kitchens, Risk Reduction Engineering Laboratory,
5995 Center Hill Road, Cincinnati, Ohio 45224, (513) 569-7672.
97
-------
FULL-SCALE LEACHATE-RECIRCULATING MSW LANDFILL BIOREACTOR ASSESSMENTS
David A. Carson
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory (ML-CHL)
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268-3001 USA
INTRODUCTION
The integrated waste management hierarchy philosophy continues to develop as a useful tool to
solve solid waste issues in an environmentally responsible manner. Recent statistics indicate that
approximately two thirds of municipal solid waste in the United States is disposed in landfills. EPA research
has continued to develop and refine the variety of technologies, materials, and operational techniques to
further reduce risk from this most often used form of waste disposal.
Current landfill operational technique involves the preparation of a waste containment facility, the
filling of the waste unit, installation of the final cover, and the maintenance of the unit. The goals being to
isolate the waste from people, and to minimize infiltration of water, thus minimizing releases of moisture
and decompositional gases into the environment. This method of operation has proven to be reasonably
effective in waste disposal, effectively minimizing risk by collecting the liquid that percolates through the
waste, called leachates, at the bottom of the landfill, and controlling landfill gas with collection systems.
Effective gas collection often results in utilization of the gas for other purposes.
Concerns over the longevity of containment systems components present questions that cannot be
answered without substantial performance data. Landfills, as currently operated, serve to entomb dry
waste. Therefore, the facility must be maintained in perpetuity, consuming funds and ultimately driving up
waste collection costs. Further, there is a concern about rare but possible environmental impacts from
leakage through lining or cover components and possible catastrophic failures of the landfill system.
This presentation will describe a new form of solid waste landfill operation, it is a technique that
involves controlled natural processes to break down landfilled waste, and further minimize risk to human
health and the environment.
A landfill operated in an active manner will encourage and control natural decomposition of
landfilled waste. This can be accomplished by collecting leachate, and reinjecting it into the landfilled
waste mass. Keeping the waste mass moist will lead to a largely anaerobic system with the capacity to
rapidly stabilize the landfilled waste mass via physical, chemical and biological methods. The system has
proven the ability to breakdown portions of the waste mass, and to degrade toxic materials at the laboratory
scale.
METHODOLOGY
Experiments are designed to compare similar landfill cells as each is operated in a different
manner, either wet or dry. The experiments are time consuming and require long-term commitment to
research. The projects are also very costly because of their magnitude. To gather data and analyze
results, U.S. EPA embarked on a research program to study this operational technique for MSW landfills in
the early 1980s. Bench-scale laboratory research on the anaerobic bioreactor in MSW landfills was
conducted with the assistance of The Georgia Institute of Technology (1). Researchers found that by
simply reinjecting common MSW landfill leachate under controlled sequences that the following benefits
could potentially be realized at full scale:
• expansion of landfill capacity through volume reductions induced by biological decomposition of
MSW in the landfill, reducing the actual number of landfills that must be sited
98
-------
• improved quality and quantity of recoverable landfill gases (methane and carbon dioxide) through
controlled biological reactions
• toxicity reduction of the MSW mass through biological decomposition and immobilization in the
waste mass resulting in lower pollutant concentration in leachate
» reduced post-closure monitoring time due to toxicity reduction
Building upon this fundamental research, EPA sponsored further research to take this operational
technique to the field in the form of pilot-scale landfill test cells (2). Anticipating the opportunity to build on
this research, full-scale landfills were sought and selected to prove the technology at full-scale. One landfill
was selected near Gainesville, Florida, and another landfill project is in the construction phase near
Rochester, New York.
Table 1. U.S. EPA Landfill Bioreactor Projects in 1994
Location
Georgia
Delaware
Florida
New York
Ohio
Scale
Lab
Pilot
Full
Full
Full
Description
Model Landfill Lysimeters
Test Cells 2 each, 1 acre each, wet vs. dry
Active Fill - 6 acres
Active Fill - 10 acres
Remediation of Existing Landfill
Status
Completed '93
Covered
Filling
Filling
Planning
Ref
1
2
3
4
5
In addition to these activities, EPA is tracking related projects in the United States, and around the
world. EPA is also pursuing the application of this technology in a remedial manner at an unlined,
uncontrolled landfill.
Leachate-Recirculation Process
Prototype methods which have been utilized to date include surface spraying, surface ponds,
vertical injection wells with and without wicks, and horizontal surface infiltration devices. Generally,
additional costs for leachate recirculation materials are relatively low. For proper execution of a leachate
recirculation system at a modern MSW landfill, the following components are generally considered to be
necessary:
• • : • ' ' i ' *s
• a composite lining system comprised of a single or double composite compacted soil and a
geomembrane, with leachate handling system.
• MSW placed at a density determined to accomrpodate leachate recirculation .
• daily cover that does not significantly affect the continuous passage of moisture through the landfill
from top to bottom (this rules out many traditional MSW landfill daily cover practices)
• a leachate reproduction system that is concealed within the landfill enclosure, that only uses
native landfill leachate
• an active gas collection system as part of a comprehensive landfill cap design
• a landfill cover capable of providing modern landfill cap functionality, that can maintain integrity as
the leachate recirculation process causes landfill volumes to decrease,
99
-------
• a trained landfill operator who understands the daily operational requirements of a leachate-
recirculating landfill
Many of these issues remain as engineering challenges that need to be resolved in full-scale
implementation.
The leachate-recirculation process is fundamentally simple, requiring relatively minor changes over
the current designs. However, the number of variables involved to demonstrate control over the reactions in
these studies are numerous, and every attempt is made during experimental design to control the number
of variables. Operator training is critical, and outputs from these projects are aimed at operator control. As
a matter of practicality, experiments to evaluate materials that may be added to the landfill during filling to
enhance degradation have been relegated to future experimentation. Possible additions include waste pre-
processing, microbiological additives, waste sludge, gases, or the control of temperature of the waste
mass.
Concurrent with EPA research, other research projects are underway in the USA and around the
world. There are new projects being initiated in California, Florida and New Hampshire, and underway
around the world in Canada, Germany, Denmark, Italy, Sweden, Japan, and United Kingdom. The
collective database formed in the study of these landfills will supply enough information to assess the
performance of this operational technique in the field.
Landfilling in the Future
As landfills continue to evolve as sophisticated waste disposal and decomposition facilities, there
are new goals anticipated for the future. Foreseen is a waste management park that will involve centrally
located management of a variety of waste streams, of which a landfill will provide a necessary service.
Waste water, sludges, composted waste and green waste, MSW materials recovery facilities, incinerators,
and recycling facilities require inputs and outputs that can be integrated resulting in more efficient
processing of MSW.
The landfill could produce an output of low to medium grade compost that could be used for soil
amendments in roadways and earth works. Gas from the landfill could be used to generate electricity or
power collection vehicles. The landfills could be arranged in a turntable-type arrangement so that cell #1 of
the landfill can be constructed, filled, covered, and placed under leachate recirculation, then on to cell #2,
where the process is repeated, and so on. When all cells are complete, the operator returns to cell #1,
where the contents could be mined (7), with recyclables and compost removed, the bottom lining inspected
and replaced when necessary, and the entire process again. While this technique will likely not eliminate
the need to site new landfills, it will significantly increase the useful life of landfills when constructed to
operate as bioreactors.
RESULTS
The status of active EPA sponsored projects are described here. The effort has benefitted from
the active participation of major waste management companies in the USA, and from the collaborative
research efforts of EPA's colleagues, and their participation is gratefully acknowledged. Results to date are
derived from the following projects: :
Laboratory Lysimeter Studies
Completed in 1993, the research performed at the Georgia Institute of Technology and University
of Pittsburgh showed leachate recirculation offers rapid and complete stabilization of the landfilled waste
masses. The system proved resiliency to toxic loadings that caused retardation but not defeat of the
stabilization process. The study showed that landfills are capable of biological and physicochemical
100
-------
reactions to attenuate waste constituents via reduction, precipitation, and matrix capture for heavy metals,
and biotic and abiotic transformation, and sorption for organic materials. The study also showed that the
performance of the bioreactor landfill can be monitored with leachate and gas parameters during significant
phases of the reaction. The study concluded that the bioreactor landfill design is viable, and offers a
significant improvement over traditional landfill operation (1).
Pilot-Scale Landfill Test Cells - Delaware
-. :Two 1 -acre pilot-scale landfill test cells are currently operated by the Delaware Solid Waste
Authority in Sandtown, Delaware. Two identical cells were constructed, one operated traditionally (dry) and
the other employing leachate-recirculation. The cells have been operating for approximately 1.5 years, and
although the data is being analyzed, preliminary results have shown that the trends described in the early
laboratory research are repeated in the field (2).
Full-Scale Landfill Project - Alachua County, Florida
-. The first of two full-scale landfill projects was selected near Gainesville, Florida. The landfill is
operated by the City with assistance from The University of Florida, engineering firms, and others. The
landfill is approximately half full, and leachate recirculation systems are being installed as filling progresses.
A nearby landfill at that site has performed this technique in a less sophisticated manner, utilizing surface
infiltration ponds in previous years with success (3).
Full-Scale Landfill Project - Monroe County, New York
The second full-scale landfill project is in the construction phase near Rochester, New York. EPA
is assisting the New York State Energy Research and Development Authority (NYSERDA) in their efforts to
study how this landfill operational technique may achieve complementary goals of environmental
protection, and the potential for enhanced energy production through gas recovery (4).
Remediation Project - Ohio
An unlined landfill is under study in Cincinnati, Ohio for application of leachate recirculation as an
alternative to pumping and treating groundwater at an uncontrolled pre-regulatory landfill. The project is
significantly more complicated due to the absence of a bottom liner, but researchers are convinced that
groundwater can be controlled and barriers can be constructed to accommodate leachate recirculation at
the site (5). A feasibility study is in final draft status, and the project will soon enter its second phase to
complete pre-construction details.
Comprehensive Study - University of Central Florida
A comprehensive evaluation of active landfill bioreactor projects is underway at the University of
Central Florida (6). Researchers there are compiling data from these projects, other American projects,
and data from projects from around the world by site visitation and interactions with specialty groups like the
International Energy Agency's Landfill Gas Expert Working Group. Specific projects aim to assess moisture
distribution of the leachate into the waste mass through computer models. This will assist designers in the
configuration of the injection system, and may lead to guidance on placed waste densities that will
accommodate leachate recirculation. A conference is planned for the autumn of 1994.
CONCLUSIONS
Landfills are currently designed and operated to serve as containment systems. While originally
designed to entomb waste, the modern landfill has evolved into a more technically advanced containment
system, with sophisticated controls and operational techniques. The landfill of the future will be protective
of human health and the environment, will degrade the waste mass in the landfill, and will be reusable,
101
-------
allowing the waste to be excavated for the cell to be inspected and refilled as part of an integrated waste
management park.
REFERENCES
1. United States Environmental Protection Agency. 1993. Behavior and Assimilation of Organic and
Inorganic Priority Pollutants Codisposed with Municipal Refuse, Volumes 1 and 2. EPA/600/R-
93/137a and b (NTIS PB93-222198 and 93-222206). Risk Reduction Engineering Laboratory,
Cincinnati, Ohio.
2. Vasuki, N.C. 1993. "Practical Experiences with Landfill Leachate Recirculation in Pilot and Field
Scale Units, Proceedings: 31st Annual International Solid Waste Exposition, Solid Waste
Association of North America, San Jose, California, August 3.
3. Townsend, T., and W. L. Miller. 1993. Alachua County Southwest Landfill Bioreactor Project,
Proceedings: Madison Waste Conference, Madison, Wisconsin, September.
4. Babcock, J., and Reis, J. 1993. "Monroe County's Leachate Recirculation Project," Modern
Double Lined Landfill Management Seminar, New York State Department of Environmental
Conservation, Saratoga Springs, NY, March 2-4.
5. United States Environmental Protection Agency. 1993. Leachate Recirculation for Remedial
Action at MSW Landfills, DRAFT FINAL REPORT. Risk Reduction Engineering Laboratory,
Cincinnati, Ohio.
6. Reinhart, D. R., and Carson, D. A. 1993. "Full-Scale Application of Landfill Bioreactor
Technology," Proceedings: 31st Annual Solid Waste Exposition of the Solid Waste Association of
North America, San Jose, CA (August 2-5,1993).
7. United States Environmental Protection Agency. 1993. Evaluation of the Collier County, Florida
Landfill Mining Demonstration. EPA/600/R-93/163. Risk Reduction Engineering Laboratory,
Cincinnati, Ohio.
FOR MORE INFORMATION
Please contact David Carson, U.S. Environmental Protection Agency, Risk Reduction Engineering
Laboratory (ML-CHL), 26 W. Martin Luther King Drive, Cincinnati, Ohio, 45268-3001, USA, at (513) 569-
7527.
102
-------
REMOVAL OF MERCURY FROM STACK GASES BY ACTIVATED CARBON
Radisav D. Vidic
University of Pittsburgh
943 Benedum Hall
Pittsburgh, PA 15261
412-624-1307
INTRODUCTION
On combustion, the trace elements in the incinerator feed stream are partitioned between the
bottom ash (slag) stream, and a flue gas stream containing suspended fly ash and vapors of volatile
elements or compounds. A further partitioning of the flue gas stream takes place in the particulate
emission control devices that efficiently remove larger fly ash particles but are less efficient for vapors
and finer particles. Ash from the bottom of the combustion chamber and ash removed by the particulate
control devices are flushed with water to ash ponds, where elements may be leached from the ash and
enter the aquatic environment in runoff. Small particles and vapors are discharged to atmosphere
through exhaust stacks and enter the terrestrial and aquatic environments by wet or dry deposition.
Environmental control agencies, researchers, and general public have become increasingly
concerned with the mobilization of trace elements to the environment from solid and hazardous waste
incinerators. Mercury is the trace element of particular concern since, during combustion, most of the
mercury present in the influent stream is transferred into the vapor phase due to its high volatility. There
is a considerable evidence in the literature that currently used pollution abatement technologies (flue gas
clean-up and particulate control devices) are not capable of controlling gas phase mercury emissions.
Once discharged to the atmosphere, mercury persists in the environment and creates long term
contamination problem. Furthermore, well documented food chain transport and bioaccumulation of
mercury, together with high toxicity to mammals and severe health problems caused by the ingestion of
mercury even at low levels, require strict control of mercury emissions from solid and hazardous waste
incinerators.
Activated carbon adsorption is a unit process that offers great promise for achieving high quality
air emissions with respect to mercury and other trace elements that might be present in gases emitted
from solid and hazardous waste incinerators. This study is designed to evaluate the rate of vapor-phase
mercury removal by virgin and sulfur impregnated activated carbons under various process conditions.
The specific process conditions that will be evaluated for their effect on the rate and mechanism of
mercury uptake include temperature, moisture content, oxygen partial pressure, and presence of other
compounds and trace elements in the vapor-phase. Accurate description of the kinetics of mercury
removal by activated carbon is an essential component in establishing design procedures that would
ensure successful application of this efficient technology for mercury control.
METHODOLOGY
Experimental set-up that is used in this study is shown in Figure 1. Carrier gas used in the
experiments is first passed through a drier and a 0.22 ^m nylon filter to remove impurities and is then
split into two streams. One stream goes through a regulating valve, mass flow controller, gas moisturizer
103
-------
(if needed), and into a mixing chamber. The other gas stream goes through a regulating valve, mass
flow controller and a mercury generator. Hydrogen chloride can also be added to this gas stream.
r=n Mass Flow Controler
— Glass Tube
Teflon Tube
txi Valve
t£i Regulating Valve
* 3-Way Valve
Mercury Trap
...Pressure
•® Gauge
Mercury
Permeation Cell
and Heater
HCI Mixing
Reservoir chamber
Water Bath
Holding
Reservoir
Figure 1. Schematic Representation of Experimental System
Mercury is added to the gas stream by passing the gas around a permeation cell that contains
mercury and is submerged into a constant-temperature water bath. Desired mercury concentrations in
the vapor phase is accomplished by adjusting the temperature of the water bath. HCI is added to the gas
stream by injecting a known flowrate of a gas which contains a known concentration of hydrogen chloride
in nitrogen. After the mixing chamber, the mercury laden gas stream passes through a holding reservoir
that is placed into the system to increase the total system volume and improve the accuracy of
experimental results. During the preparation phase, which precedes each measurement of mercury
uptake by activated carbon, the gas stream is directed around the activated carbon bed until a steady-
state conditions with respect to gas composition are achieved.
Gas stream then passes through a quartz cell placed in a light path of an atomic absorption
spectrophotometer (AAS) which enables direct and continuous measurement of elemental mercury in the
vapor phase. Finally, the gas stream is passed through appropriate impinger solutions to facilitate
measurements of other forms of mercury in the gas phase as well as other trace compounds that might
be present in the gas stream. Impinger solutions are placed in 250 ml_ gas washing bottles. The
collection of a particular compound from the gas phase into an impinger solution is performed for a
period of 60 minutes after which the fresh solution will be placed on line. The mercury trap at the exit
from the system is a GAG column charged with 100 g of sulfur impregnated activated carbon (HGR).
Once the composition of the gas stream is stabilized at predetermined levels, the regulating
valve is closed to prevent any infusion of mercury into the holding reservoir and the gas stream exiting
the AAS is directed back into the holding reservoir thereby, forming a completely closed system. The
compressor in the system is operated at a very high flow rate (up to 3 cfm) to ensure complete mixing in
104
-------
the system throughout the experiment. The flow of the gas is then directed to a differential (short)
activated carbon bed positioned inside the furnace and the uptake of mercury by activated carbon is
monitored with time.
Elemental mercury in the gas stream is measured continuously at a wavelength of 253.7 nm
using a quartz gas cell placed in the AAS system which is first calibrated according to the procedure
described by Shendrikar et a/. (1) while the total mercury is determined by directing a gas through a 200
mL impinger solution containing 1.5% KMnO^ and 10% H2SO4that collects all mercury species from the
gas phase. The total mercury concentration in the impinger solution is then analyzed using a cold vapor
atomic absorption spectrophotometric method (2).
The proposed research objectives will be accomplished through the following activities:
1) Evaluate the effect of temperature and mercury concentration in the gas phase on the rate of
mercury adsorption by virgin and sulfur impregnated activated carbons for a wide range of mercury
concentrations (100 -1000 |j.g/m3) and temperatures (80-200 °C) that are representative of gaseous
emissions from incineration plants. Activated carbons selected for this part of the study are bituminous-
coal based virgin activated carbon designed for vapor-phase applications (BPL) and sulfur impregnated
activated carbon (HGR) (Calgon Carbon Corporation, Pittsburgh, PA).
2) Evaluate the effect of activated carbon particle size on the rate of mercury uptake from the
gas phase. Smaller particle sizes would provide easier access of mercury to the surface active sites
where adsorption occurs, thereby improving the kinetics of adsorption. However, injecting smaller GAC
particle sizes would facilitate faster build-up of pressure drop at the fabric filter which is used to control
particulate emissions. Consequently, the frequency of fabric filter cleaning would increase leading to
lower run times. Therefore, it is necessary to determine an optimal particle size that would satisfy these
opposing requirements. The intention of this task of the study is to determine the extent of the
improvement in adsorption kinetics with the reduction in activated carbon particle size. Activated carbon
particle sizes that will be used in such experiments include 30x40, 60x80, and 170x230 U.S. Standard
Mesh sizes. These kinetic experiments will be performed at 140 °C using an influent mercury
concentration of 500 j^g/m3 as representative of incinerator flue gas conditions. Once the optimum
activated carbon particle size is established during this phase of the study, it-will be used in the rest of
the experiments outlined below.
3) Determine the effect of environmental conditions on the rate mercury uptake by activated
carbon. The key environmental conditions that need to be evaluated include moisture content and
oxygen partial pressure in the gas stream. Effects of moisture on the adsorption kinetics will be
evaluated using dry and moisture saturated nitrogen as a carrier gas. Dry nitrogen and dry air will be
used as carrier gases in the experiments designed to evaluate the effect of oxygen partial pressure on
the adsorption kinetics. All the kinetic experiments included in this task of the study will be conducted at
three different temperatures (80,140, and 200 °C) using an influent mercury concentration of 500 ^g/m3.
All possible combinations of moisture content and oxygen partial pressure in the gas stream will be
employed to study the effect of temperature on chemical interactions between mercury, oxygen, and
moisture, and the related impact on the rate of mercury uptake by activated carbon.
4) Evaluate the effect of the presence of halogen compounds (most notably chlorine, Cl) in the
gas stream on the kinetics of mercury uptake by activated carbons. Cl can react with elemental mercury
and change its oxidation state and therefore, influence the affinity of mercury towards the carbon
surface. This task of the study includes kinetic experiments at three different temperatures (80, 140 and
200 °C) using different ratios of mercury and chlorine concentrations in the influent stream. Typical
chlorine in-stack vapor-phase concentrations for incineration plants range from 400 to 1000 (J.L/L as
hydrogen chloride, HCI (3). Two different HCI concentrations will be tested in this study, namely 400 and
1000 (J.L/L while the concentration of mercury in both cases will be 500
105
-------
RESULTS
An experimental system capable of producing constant concentrations of vapor-phase mercury
was successfully constructed and tested at the outset of this study. Furthermore, an AAS was calibrated
by trapping a carrier gas containing different concentrations of vapor-phase mercury in the impinger
solutions. It was discovered that mercury adsorbs to glass walls of the impinger bottles and the problem
was solved by rinsing the glassware with concentrated nitrohydrochloric acid. The calibration curve
generated from these experiments is shown in Figure 2.
50
60 70
Water Bath Temperature [°C]
Figure 2. Vapor-Phase Mercury Concentration as a Function of Water Bath Temperature
Following the calibration of the instrument, the influence of mercury concentration and ambient
temperature on the adsorptive capacity of virgin GAG was evaluated using the fixed-bed breakthrough
experiments conducted with 100 mg of GAG and a carrier gas flow rate of 1 L/min. The results of these
experiments are summarized in Figure 3.
1.6+1
« T = 25 °C
O T = 90 °C
T=140°C
1.6-2
20
40
60 80
Mercury Concentration [ng/m3]
100
Figure 3. Effect of Temperature and Mercury Concentration on Virgin GAC Adsorptive Capacity
106
-------
CONCLUSIONS
As is apparent from Figure 3, the increase in mercury concentration leads to an increase in virgin
GAG adsorptive capacity for vapor-phase mercury. However, higher temperatures significantly reduce
the ability of virgin activated carbon to remove mercury from the stack gases. Both of these trends were
expected since the adsorption process is an exothermic reaction that is driven by the concentration of
adsorbate in the fluid phase.
What is also apparent from Figure 3 is that the virgin GAC has very low capacity for the removal
of vapor-phase mercury under the conditions that are prevalent in the incineration plant stacks (high
temperatures). Furthermore, the presence of other trace inorganic and organic compounds in the stack
gas would further diminish its capacity for the retention of mercury due to competitive adsorption.
Therefore, it appears that sulfur impregnated carbon would be better suited for the control of mercury
emissions from solid and hazardous waste incinerators for the following reasons: a) the capacity of sulfur
impregnated GAC is several orders of magnitude above the capacity of virgin GAC; and b) the entire
adsorptive capacity of sulfur impregnated GAC can be used exclusively for vapor-phase mercury since
this carbon can not adsorb any other trace organic or inorganic compounds in the stack gas.
Next goal in this study is to evaluate the effects of temperature and mercury concentration on
the ability of sulfur impregnated activated carbon to remove mercury from the gas phase (similar to
those presented in Figure 3 for virgin GAC) followed by the evaluation of the rates of vapor-phase
mercury uptake by both carbons. Based on the results of these experimental studies, a mathematical
model that accounts for both diffusion rate and the rate of reaction between sulfur and vapor-phase
mercury will be develop to assist in predicting the performance of these activated carbons in the control
of mercury emissions from solid and hazardous waste incinerators.
REFERENCES
1. Shendrikar, A.D., Damle, A., and Gutknecht, F. Collection Efficiency Evaluation of Mercury^
Trapping Media for the SASS Train Impinger Solution. EPA-600 S7-84-089, U.S. EPA,
Washington, DC., 1984.
2. APHA-AWWA-WPCF Standard Methods for the Examination of Water and Wastewater, 17th
ed., Port City Press, Baltimore, MD, 1989.
3. Hall, B, Schager, P., and Lindqvist, O. (1991) "Chemical Reactions of Mercury in Combustion
Flue Gases." Water, Air, and Soil Pollut, 56, 3.
For More Information:
Luis Garcia
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
513-569-7932
107
-------
APPLICATION OF AHBERSORB 563 ADSORBENT TECHNOLOGY
FOR TREATMENT OF CHLORIHATED ORGANICS IH GROUNDWATER
Russell W. Frye, Joseph F. Hartino, and Russell E. Turner
Roy F. Ueston, Inc.
1 Weston Way
West Chester, PA 19380
610-701-6173
Eric G. Isacoff and Deborah A. Plantz
Rohm and Haas Company
Research Laboratories
Spring House, PA 19477
215-641-7255
INTRODUCTION
Roy F. Weston, Inc. (WESTON.), in conjunction with Rohm and Haas Company (Rohm and Haas), conducted a
field pilot study to demonstrate the technical feasibility and cost-effectiveness of Ambersorb* 563
(A-563) carbonaceous adsorbent for the remediation of groundwater contaminated with volatile organic
compounds (VOCs). The project was conducted under the Emerging Technology Program of the EPA Superfund
Innovative Technology Evaluation (SITE) program.
Ambersorb adsorbents are a family of patented, synthetic, tailorable carbonaceous adsorbents that were
developed by Rohm and Haas in the 1970's for the treatment of contaminated water. In specific applications,
Ambersorb adsorbent technology may offer a cost-effective alternative to air stripping or granular activated
carbon (GAC), which are typically used in pump and treat systems for remediating groundwater contaminated
with organic compounds.
Ambersorb adsorbents have been found to be effective in the removal of low levels of VOCs and other
synthetic organic compounds from contaminated water (1). Previous applications using Ambersorb adsorbents
have demonstrated several key performance benefits over GAC (2,3,4,5). Ambersorb 563 adsorbent can be
regenerated onsite using steam, solvents, or other techniques, permitting the recovery of a concentrated
organic stream which can be disposed of or reclaimed. Ambersorb 563 adsorbent has a significantly greater
adsorption capacity than GAC for chlorinated hydrocarbons when the contaminants are present at low
concentrations. Ambersorb 563 adsorbent systems can operate at higher flow rates than GAC systems, while
maintaining effluent water quality below drinking water standards.
METHODOLOGY
The Ambersorb technology demonstration was conducted at Pease Air Force Base (AFB) in Newington, New
Hampshire. The base has been included on the National Priorities List (NPL) and WESTON has been conducting
an Installation Restoration Program (IRP) Stage 3 Remedial Investigation (RI) at Pease AFB over the past
several years. Based on a review of groundwater data for various sites at Pease AFB, Site 32/36 was
selected for the Ambersorb 536 adsorbent field trial. The groundwater in this area is contaminated with a
number of chlorinated organics including vinyl chloride (VC), 1,1-dichloroethene (1,1-DCE), cis-1,2-
dichloroethene (cis-1,2-DCE), trans-1,2-dichloroethene (trans-1,2-DCE), and trichloroethene (TCE).
The Ambersorb technology demonstration project under the SITE program used a 1-gallon-per-minute (gpm)
continuous pilot system, consisting of two adsorbent columns, to evaluate the treatment of groundwater from
Site 32/36 at Pease AFB. The field study was performed over a 12 week period during the spring/summer of
1994. The testing program included four service cycles and three steam regenerations. A summary of the
conditions for the service cycles and steam regenerations is provided in Tables 1 and 2, respectively.
108
-------
TABLE 1. SUMMARY OF CONDITIONS FOR SERVICE CYCLES
Parameter
Column configuration
Column Diameter (Inches)
Column Height (feat)
Flow Rate (gpm)
Hydraulic- Loading (gpnVK2)
Flow Rate Loading (gpnVff)
Empty Bed Contnct Time (EBCT) (minutes)
Days to Saturation Breakthrough for Each
Column*
Estimated Number of Bed Volumes (BV)
Treated*
Cycle 1
Cot A
A-563 Virgin
4
2
0.44/0.87"
5/10
2.SJS
3.0/1.5
12/6
6,000
CoLB
F-tOO Virgin
6
3
0.2»0.58'
1.5/3.0
0.5/1.0
15.0/7.5
12/6
1,000
Cyck»2
Cot A
A-563
Regen
Lag
4
2
0.87
10
2.5*
3.0'
12
6,000
Col B
A-563
Virgin
Lead
4
2
0.87
10
5
1.6
6
6,000
Cyde3
COLA
A-563
Regen
Lead
4
2
0.87
10
5
, 1.5
6
6,000
CoLB
A-563
Regen
Lag
4
2
0.87
10
2.ff
3.0*
12
6,000
Cycle 4
Cot A
A-563
Regen
Lag
4
2
0.87
10
2.5-
3.0f
12
6,000
CotB
A-563
Regen
Lead
4
2
0.87
10
5
1.5
6
6,000
'Flowrate was increased after 7 days.
'''Total system loading with columns operating in series.
Estimated value predicted from model using influent vinyl chloride concentration of 5 ug/L.
TABLE 2. SUMMARY OF CONDITIONS FOR STEAM REGENERATION
Parameter
Column Regenerated
Temperature (°C)
Time (hours)
Total Bed Volumes
Regeneration 1
Cycle 1, Column A
155
16
7.6
Regeneration 2
Cycle 2, Column B
145
16.5
7.0
Regeneration 3
Cycle 3, Column A
140
17.5
8 9
The first cycle consisted of a direct comparison of the performance of Ambersorb 563 adsorbent and
Filtrasorb 400 (F-400) GAC. The remaining cycles evaluated Ambersorb 563 adsorbent under varying process
conditions. .Concentrations of VOCs in the influent contaminated groundwater and the treated column effluent
were monitored during each cycle to establish breakthrough curves. Process parameters including groundwater
influent flowrate, temperature, and pressure were also monitored at periodic intervals throughout the field
trial.
Following contaminant breakthrough, the service cycles were terminated. Steam regeneration of the lead
Ambersorb adsorbent column was then performed onsite. The regeneration process yielded a condensate
consisting of a distinct separable organic layer,and an aqueous phase. Both the aqueous and organic phases
were measured and analyzed to assess regeneration efficiency.
A test to demonstrate the use of a "superloading" adsorbent column to treat the aqueous condensate from
a typical steam regeneration process was also conducted during the field trial. The superloading column
takes advantage of Ambersorb adsorbent's higher adsorption capacity at higher concentrations. The aqueous
phase from the third steam regeneration was passed through an Ambersorb adsorbent superloading column with a
diameter of 2 inches and a bed height of 2 feet. Effluent samples from the superloading column were
collected and analyzed for VOCs.
109
-------
RESULTS
The average VOC levels measured in the influent groundwater during each of the service cycles is
summarized in Table 3. Note that due to analytical limitations, the vinyl chloride and 1,1-DCE
concentrations for certain cycles were estimated based on the amount of the contaminant subsequently
recovered during regeneration.
TABLE 3. INFLUENT GROUNDWATER QUALITY
Compound
VC
1,1-DCE
cis-1,2-DCE
trans-1,2-DCE
TCE
Maximum Contaminant Level
(ug/L)
2
7
70
100
5
Average Influent Concentration (ug/L)
Cycle 1
A-563
3.4+
0.31+
312
101
4,330
F-400
3.9+
0.31+
329
101
4,120
Cycle 2
A-563
4.8
5.8+
351
121
4,490
Cycle 3
A-563
5.7
6.6+
373
116
3,600
Cycle 4
A-563
11
<10
357
89
3,883
National Revised Primary Drinking Water Regulations (40 CFR 141.61).
+Estimated concentration based on recovery during regeneration of A-563 column.
Vinyl chloride, cis-1,2-DCE, trans-1,2-DCE, and TCE were present in the influent groundwater at
concentrations exceeding the maximum contaminant levels (MCL) established in the National Revised Primary
Drinking Water Regulations. TCE was the contaminant measured at the highest average concentration, ranging
between 3,750 ug/L and 4,330 ug/L.
The results of the Ambersorb 563 adsorbent demonstration study are summarized for each cycle in
Table 4.
TABLE 4. DEMONSTRATION STUDY PERFORMANCE RESULTS
Compound
VC
1,1-DCE
cis-1,2-DCE
trans- 1,2-DCE
TCE
Bed Volumes to Drinking Water Standard
Cycle 1
A-563
7,900
>12,600
9,000
>12.600
7,700
F-400
1,740
4,270
3,700
4,890
4,820
Cycle 2*
A-563
8,500
>1 1,800
10,500
>1 1,800
9,300
Cycle 3*
A-563
6,500
>7,500
8,500
>11,450
5,300
Cycle 4*
A-563
5,000
>15,690
9,900
>15,600
7,500
Results presented for the lead column.
The monitoring results for Cycle 1 column effluents showed that both the Ambersorb 563 adsorbent and
GAC achieved water quality below the drinking water standards for each compound.' However, when comparing
the bed volumes treated to the drinking water standard breakthrough for each contaminant, Ambersorb 563
adsorbent, while operating at 5 times the flowrate loading as Filtrasorb 400, treated approximately two to
five times the volume of water as GAC. The TCE breakthrough curves for Cycle 1 are shown in Figure 1.
Comparison of the performance results for the Ambersorb adsorbent for the four cycles indicated minimal
loss in adsorption capacity during the course of the field trial. The reduction in bed volumes treated to
the drinking water standard breakthrough observed in Cycles 3 and 4 resulted from the increase in vinyl
chloride concentration in the influent groundwater. There was no significant reduction in capacity for any
of the other contaminants.
110
-------
10
9 --
8 -•
a
•2 6 i
•§ 5
Filtrasorb 400
Ambersorb 563
MCL
0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000
Bed Volumes
Figure 1. Cycle 1 Ambersorb 563 and Filtrasorb 400 Breakthrough Curves for Trichloroethene.
The results of the steam regenerations are provided in Table 5.
TABLE 5. STEAM REGENERATION RESULTS
Parameter
Total Bed Volumes
VOC Mass Recovery in 3 Bed Volumes (%)
Total VOC Mass Recovery (%)
VOC Mass Recovery Associated with Organic Phase (%)
Regeneration 1
7.6
70
79
86
Regeneration 2
7.0
70
74
90
Regeneration 3
8.9
90
>95
81
The steam regeneration results indicated that a significant recovery of the VOC mass loaded onto the
Ambersorb adsorbent during the service cycles, ranging from 74% to >95%, was achieved during the
regeneration process. The results showed that the bulk of the VOC mass recovery occurred within the first 3
bed volumes of steam as condensate. Furthermore, the results indicated that approximately 80% to 90% of the
VOC mass recovered was associated with the easily separable organic phase.
The differences observed in the VOC mass recovery for the three steam regenerations may be related to
the dehydrohalogehation of chlorinated organics under elevated temperatures. The pH profiles for the steam
condensate for each of the regenerations suggest the possibility of a dehydrohalogenation mechanism.
The results of the superloading test indicate that the condensate was effectively treated to levels
below the drinking water standards. A total of 15 bed volumes of condensate, which averaged 700,000 ug/L
VOCs (predominately TCE) was passed through the superloading column at an EBCT of 7.5 minutes (8.0 BV/hr).
The effluent samples from the superloading column showed no detectable leakage of any VOCs for 11 bed
volumes. TCE at 2 ug/L was the only compound detected in the final superloading column effluent sample
collected after 15 bed volumes.
111
-------
CONCLUSIONS
Ambersorb 563 adsorbent is an effective technology for the treatment of groundwater contaminated with
chlorinated organics. The effluent groundwater from the Ambersorb 563 adsorbent system consistently met
drinking water standards. The adsorption capacity of the Ambersorb system remained essentially unchanged
following onsite regeneration of the adsorbent and multiple service cycles.
Direct comparison of the performance of Ambersorb 563 with Filtrasorb 400 using the bed volumes treated
to the drinking water standard breakthrough indicated that Ambersorb 563 adsorbent was able to treat
approximately two to five times the volume of water as GAC while operating at 5 times the flowrate loading.
Onsite steam regeneration was successfully demonstrated. The steam regenerations yielded a separate
organic phase which contained approximately 80% to 90% of the total VOC mass loaded onto the adsorbent. The
majority of VOC recovery was shown to occur within 3 bed volumes of steam as condensate.
The principle of super-loading was demonstrated as an effective treatment method for the aqueous
condensate layer resulting from the steam regeneration of the Ambersorb adsorbent. A condensate stream
containing 700,000 ug/L VOCs was treated to below the drinking water standards using a super-loading column.
REFERENCES
1. Neely, J.W. and Isacoff, E.G. Carbonaceous Adsorbents for the Treatment of Ground and Surface
Waters. Marcel Dekker, Inc., New York, New York, 1982.
2. Vandiver, H. and Isacoff, E.G. THM Reductions with Ambersorb 563 Adsorbent. Paper presented at 41st
Annual Conference, Society of Soft Drink Technologists, Albuquerque, New Mexico. April 19, 1994.
3. Parker, Jr., G.R. Comparison of Ambersorb 563 Carbonaceous Adsorbent and Granular Activated Carbon
for the Removal of TCE from Water at Short Empty Bed Contact Times. Presented at American Institute
of Chemical Engineers Annual Conference, Miami Beach, Florida. November 4, 1992.
4. Isacoff, E.G., Bortko, S.M., and Parker, Jr., G.R. The Removal of Regulated Compounds from
Groundwater and Wastewater using Ambersorb 563 Carbonaceous Adsorbent. Paper presented at American
Institute of Chemical Engineers Conference, Miami Beach, Florida. November 3, 1992.
5. Parker, Jr., G.R. and Bortko, S.M. Groundwater Remediation using Ambersorb Adsorbents. Paper
presented at the Florida Environmental Chemistry Conference. October 30-November 1, 1991. '
For More Information:
Ronald J. Turner
Project Officer
U.S. EPA
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
513-569-7775
Russell E. Turner
Project Manager
Roy F. Weston, Inc.
1 Weston Way
West Chester, PA 19380
610-701-3097
112
-------
GEOSAFE IN SITU VITRIFICATION SITE DEMONSTRATION
James E. Hansen, Director
Brett E. Campbell, Site Manager
Crag L Timmerman, Projeqt Manager
Geosafe Corporation
2950 George Washington Way, Richland, WA 99352
(509)375-0710 ,
INTRODUCTION
This extended abstract presents the results of the EPA SITE Program Demonstration that was
conducted on Geosafe Corporation's In Situ Vitrification (ISV) technology at the Parson Chemical/ETM
Enterprises Superfund Site in Grand Ledge, Michigan. The significance of the demonstration results
are also related to other Geosafe project experience and the current state of the ISV technology.
The ISV technology is a joule-heated electric melting technology that treats contaminated soil and
other earthen materials (e.g., sediment, sludge, flyash, mill tailings) for the primary purposes of destroy-
ing, removing, or immobilizing hazardous, radioactive, and mixed contaminants. ISV may be applied to
materials in their original location within the ground, or to materials placed in a specific location or con-
tainer, below grade or above grade, for purposes of treatment. A melt is typically initiated at the sur-
face of the material to be treated. Joule heating occurs as electric current flows through the molten
material, thus causing the melt to increase in temperature and adjacent material to melt. Typical melt
temperatures range from 1,600 to 1,800°C. Single Melts as large as 1,400 tons and exceeding 20-ft in
depth have been achieved. Adjacent melts fuse together to form a single contiguous monolith.
Contaminants may be destroyed, immobilized, and/or removed during ISV. The high temperature
typically destroys organics by pyrolysis. The predominant disposition of heavy metals is chemical or
physical incorporation within the resulting vitreous monolith, which produces a permanent immobilization
result. Some vaporizable contaminants may be removed by the process heat without undergoing
destruction or immobilization. The specific disposition that may be expected for contaminants at a
given site depends on many waste- and site-specific variables. Off-gas treatment is employed to treat
and/or remove vaporized contaminants and to ensure gases evolved from the process are safe for re-
lease.
The ISV process and equipment system is illustrated in Figure 1. ISV is a truly mobile technology
with the majority of process equipment being permanently mounted on trailers.
The Parsons Chemical site was previously owned and operated from 1945 through 1979 by the
former Parsons Chemical Works, Inc. which was involved in the mixing, manufacturing, and packaging
of agricultural chemicals including pesticides, herbicides, solvents, and mercury-based compounds.
These activities resulted in the contamination of soil around the manufacturing facility and in ditch
locations of a drainage system that flowed approximately 1/4 mile to a nearby creek. The typical depth
of contamination around the site was 5-ft or less. A total of about 3,000 cu-yd of soil was found to be
contaminated with a broad range of organics and metals associated with the Parsons Chemical Works'
activities at the site. The four contaminants of primary regulatory concern, and their maximum concen-
trations measured in the site soil, included: chlordane (89,000 ppb), 4,4'-DDT (340,000 ppb), dieldrin
(87,000 ppb), and mercury (34,000 ppb). Dioxins were also found on the site at very low levels.
The site soil is a silty clay with some sand present. A sandy layer exists approximately 8 to 10-ft
below grade. Water flows through this layer, in a north easterly direction, on a variable basis related to
recent rain and snowfall. The site soil is relatively homogeneous with nearly zero rock content. The
113
-------
Backup Off-Gas
System
Thermal
Oxidizer
(Option)
Secondary Waste Recycling
to Future Melts
Figure 1. ISV Processing and Equipment Schematic
soil wet density is 1.8 ton/cu-yd; the dry density is 1.5 ton/cu-yd. The soil was found to be very low in
load bearing capacity when wet, and very hard and strong when dry. These conditions made it difficult
to work with regardless of the season.
METHODOLOGY
EPA and the Michigan State Department of Natural Resources (MDNR) established ARARs for
the site including the following cleanup standards for the treated (vitrified) soil: chlordane (1,000 ppb),
4,4'-DDT (4,000 ppb), dieldrin (80 ppb), and mercury (12,000 ppb).
The contaminated soils from around the site were consolidated into a 16-ft deep trench for ISV
treatment purposes. The treatment trench was laid out in a manner that would accommodate nine
individual melts involving about 400 cu-yd of contaminated soil each. The treatment trench construction
included a feature to intercept and divert groundwater that might flow into the site through the sandy
layer at the 8 to 10-ft level. This was accomplished by placing a layer of cobble rock under the com-
plete treatment area, deeper cobble-filled trenches along the outside (north-south) edges of the treat-
ment area, and vertical walls of cobble around the complete treatment volume. In addition, two cobble-
filled sumps were placed at the northernmost corners of the treatment area to serve as accumulation
points for intermittent pumping of intercepted water to a nearby drainage ditch. Figure 2 illustrates the
configuration of the treatment trench.
At the time of the project design, it was also believed that the cobble walls would be beneficial to
the project as thermal barriers, which would limit melt width and thus minimize the extent of overmelting
into clean adjacent soil. In order to construct the cobble walls, concrete walls were placed as vertical
forms on the interior of the cobble (side toward the contaminated soil), and pressed wood sheeting was
used as forming on the exterior (native soil) side. During the vitrification portion of the project, concern
114
-------
^ Pumping sumps
il
jjl
;i,
\
•j
%
|
it
•J
i-i
•$
V*i
f\
•**
Ij
a
^f
p
&
VV? f3?5ft»e"SB I
•
1
i'
r.
28X28 I
melt cell |
(typ) — '
;!
Concrete ^^
walls "^
*
&
^i
^- tf
If &
^^ 1
1
2
~*3
4
""""^
6
8
0
^
!• I
?• \
i^
{••
i<
i
5
7
g
How direction of \
i
V
%
j
i
i
^
j
•{
perched water CobWe wall with
drain underneath
Intercept trench
\
. (installed mid-project)
Figure 2. Treatment Trench Configuration
developed over having the intercept water and cobble
materials located immediately adjacent and beneath
the contaminated soil. Therefore, about midway
through the project, another intercept trench was con-
structed away from the treatment trench for purposes
of intercepting the groundwater before it could reach
the vicinity of the treatment trench. This new intercept
trench was successful in minimizing the amount of
water reaching the original intercept and diversion
system.
The SITE Demonstration Program identified one
critical objective and seven secondary objectives for
their evaluation of the demonstration. They deter-
mined that the demonstration would be performed on
the sixth melt performed at the site. Their objectives
were:
• To determine if the final soil cleanup levels
were achieved (critical objective)
• To evaluate the teachability characteristics
of the vitrified product using the TCLP
• To determine the approximate levels of
residual contaminants in the vitrified soil
o To characterize the pesticide and mercury
content of the off-gas scrubber water
« To evaluate emissions from the process
• To identify operational parameters of the
technology
To develop operating cost estimates and projections, and to assess equipment reliability
• To examine potential technical, institutional, operational, and safety impediments related to
the use of the ISV technology.
Extensive sampling and analysis of the demonstration soil volume was performed to establish a
pre-vitrification contamination level. Composite sample results for the contaminants of concern were:
chlordane (2,000 ppb), 4,4'-DDT (72,000 ppb), dieldrin (12,000 ppb), and mercury (12,000 ppb). These
values were reported as estimates since they were less than the reporting detection limit for the meth-
ods used, but greater than the method detection limits.
Geosafe utilized conventional ISV processing methods for performance of the individual melts.
The "feeding electrode" concept was employed, wherein the depth of the electrodes were controlled by
feeders. The depth of the melt could be determined at each electrode by lowering the electrodes to the
bottom of the melt. Two notable changes from the initial operating mode were made midway through
the project. First, because of the tendency for the soil conditions at the site to produce wider than usual
melts, and because of the limited performance of the cobble walls to limit melt width, it was determined
115
-------
to employ refractory concrete barrier panels in place of the cobble rock for the last five of the melts.
Second, due to the presence of a nonhazardous but sometimes offensive odor present within the off-
gas emissions, a thermal oxidizer was added as an off-gas polishing step midway through the project to
eliminate the odor.
RESULTS
SITE Demonstration Program data indicated that cleanup objectives for the site were achieved.
Table 1 presents pre-test, post-test, and regulatory limit values for the contaminants of concern. It
should be noted that the pre-test value for chlordane was found to be below regulatory limits, although
prior samples from the site were well above limits. The fact that the materials were excavated and
staged for treatment could be expected to provide a mixing of soils and an "averaging down" of con-
tamination levels. It should also be noted that the ISV process is a destruction and removal process for
organics, and is primarily a removal process for mercury. Mercury is notable compared to other heavy
metals in that its low solubility in silicate melts, and its high vapor pressure at ISV melt temperatures,
result in nearly total removal by vaporization to the off-gases. This behavior is in contrast to other
heavy metals of environmental concern, which predominantly are retained in and immobilized by the
vitrified product.
Vitrified product TCLP data was obtained for the target contaminants as well as other priority
pollutant metals (see Table 2). It should be noted that the indication of organics being present at below
the detection limit for the analytical method used is an improbable result due to the fact that organics
cannot exist at the temperatures experienced in the ISV melt. The TCLP results were all very good,
and were in agreement with many prior tests of ISV vitrified product that show the benefit of vitrification
for immobilization of heavy metals and destruction of organics.
Results of stack gas emissions analyses are presented in Table 3. The less than values for the
target pesticides are based on detection limit values for the analytical methods used. No pesticides
were detected in the off-gas. Because of the presence of other metals within the soil in addition to the
target metal contaminant (mercury), the metals arsenic, chromium and lead were designated as critical
analytes of the off-gas emissions. As indicated in Table 3, all metals emissions results were in compli-
ance with established regulatory guidelines. Arsenic was below reporting detection limits in all samples.
The purpose of the wet scrubber in the off-gas treatment system is to remove paniculate and
condensible vapors that may escape the treatment zone and enter the off-gases. The concentration of
such materials in the scrub solution increases over time. SITE Demonstration Program analyses of the
scrubber water confirmed that it contained volatile organics, partially oxidized semivolatile organics,
mercury, and other metals. Geosafe sent the scrubber solution offsite for treatment and disposal at the
TABLE 1. COMPARISON OF PRE-TEST AND POST-TEST SOIL
CONTAMINANT CONCENTRATIONS WITH CLEANUP CRITERIA*
Contaminant
Pre-Test
Post-Test
Regulatory Limit
Chlordane
4,4'-DDT
Dieldrin
Mercury
<80**
13,000
4,600
3,800
<80
<16
<16
33
1,000
4,000
80
12,000
* All results by SITE Demonstration Program; all units are ng/kg
** < values indicate not detected at or above presented value (detection limit)
116
-------
TABLE 2. VITRIFIED PRODUCT TCLP DATA FOR CONTAMINANTS OF INTEREST*
Contaminant
Chlordane
4,4'-DDT
Dieldrin
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
* All results by SITE
** *• wall ice inrtir>£rto n
Post-Test
Concentration (ng/kg)
<80**
<16
<16
3,100
3,600
NA
13,000
8,600
33
NA
NA
Demonstration Program
n* riotortari ait rvr otvn/o nracoi
TCLP
Result (ng/L)
<0.50
NA
NA
13
440
<5.0
<10
1,100
0.18
<300
<10
nioH i /oil u» /rteto/tfirhn limrf\
Regulatory
Limit (|ig/L)
30
NA
NA
5,000
100,000
1,000
5,000
5,000
200
1,000
5,000
TABLE 3. STACK EMISSIONS PERFORMANCE DATA COMPARED TO
REGULATORY GUIDELINES*
Contaminant
Chlordane
4,4'-DDT
Dieldrin
Arsenic
Chromium
Lead
Mercury
Emissions
Value (Ib/hr)
<1.0X10**.
<2.0X1O6
<2.0X1ff8
<1.9X-10*
2.1 X10T5
2.8X10^
1.1 XIO^4
Regulatory
Limit (Ib/hr)
2.5 X101
1.0 X10'2
2.8 X ICJ4
***
"Jdcft
***
5.9X10^
* All results by SITE Demonstration Program
** < values indicate not detected at or above presented value (detection limit)
*** Emissions levels were deemed acceptable by MDNR
117
-------
end of the project. Alternatively, it is believed that several means may be employed to remove the
contaminants from the water, such that the contaminants could be replaced in the soil for additional de-
struction/immobilization treatment by a subsequent ISV melt. Such alternatives were not attempted
during the demonstration.
The SITE Demonstration Program also evaluated the reliability and operating efficiency of the ISV
process and equipment. The demonstration melt was found to take 10 days to complete, melting
approximately 600 tons of soil at an average specific power consumption of 0.72 MWh/ton. The ISV
process equipment operated very well during the demonstration with only minimal downtime for addition
of electrode segments and minor adjustments.
Lastly, the SITE Demonstration Program performed order of magnitude (+50%, -30%) estimates of
cost for ISV processing based on the costs incurred during the demonstration melt. They estimated the
cost for three cases involving different quantities of material treated and depth of treatment, as follows:
Case 1: 1,700 tons at 5-ft depth - $740/ton; Case 2: 5,700 tons at 15-ft depth - $430/ton; and Case 3:
7,900 tons at 20-ft depth - $370/ton. It is noted that ISV costs should be computed on a per ton basis
since the throughput capability is related to melting mass as opposed to volume. Processing rates vary
directly with soil wet density. It is appropriate to use wet density in such determinations since the
process consumes time and energy for removal of water as well as for melting soil.
CONCLUSIONS
The SITE Demonstration Program reported that the ISV technology performed well relative to all
the critical demonstration objectives, and that the ISV technology should be applicable to other sites
with similar contaminants and soil conditions.
i
The demonstration reporting noted that the contamination conditions at the Parsons Chemical site
were not severe enough to enable measurements and calculation of destruction and removal efficien-
cies (DREs) for the pesticides treated at the site. This is a typical problem where contamination levels
are too low relative to the capabilities of analytical methods to detect contaminants at low enough levels
to determine DREs in the range of 4-9s (99.99%) or greater. Relative to this limitation of the demon-
stration, Geosafe notes that a large number of ISV tests have been performed that have demonstrated
repeatedly the capabilities of the process to attain DRE's in the range of 99.99% to >99.9999% for
volatile, sem'rvolatile, and nonvolatile organics. One such demonstration was performed at large-scale
immediately after the Parson Chemical Project. This demonstration was Geosafe's National TSCA
Demonstration Project which was performed at a private site in Region 10. The site contained PCBs to
a maximum level of 17,000 ppm. The project was performed under the auspices of EPA's TSCA orga-
nization, as a demonstration in support of Geosafe's application for a National TSCA Operating Permit
for PCB treatment. The project clearly demonstrated the capability of the process to exceed 99.9999%
ORE for PCBs.
Geosafe concurs with the SITE Demonstration Program's order of magnitude cost estimates for
ISV based on the conditions experienced at the Parsons Chemical site. However, there are several
factors that should be noted relative to considering the cost of ISV at other sites. It should first be
recognized that specific conditions at the Parsons Chemical site made it a more costly endeavor than
may be typical for other sites. The primary determinants of cost are four: 1) the price of electricity at
the site, 2) the depth of processing, 3) the amount of water to be removed by the process, and 4) the
amount of clean soil that is melted to ensure that the full target volume has been treated. The Parsons
Chemical project was at the higher end of all these variables except processing depth.
Because of the large amount of electricity purchased, it is typically possible to obtain a negotiated
prq'ect rate of approximately 1/2 the local residential rate. The price of electricity for ISV within the
118
-------
U.S. typically falls in the range of $0.02/kwh (for large government applications) to $p.06/kwh. The
price of power at the demonstration site totalled about $0.07/kwh. The 15-ft processing depth at the
site is considered to be a good economical depth for ISV. Deeper processing depths are more eco-
nomic than shallow depths because overall equipment utilization efficiency (time spent melting versus
time spent moving equipment between melts) is greater for deeper depths. The amount of water that
had to be removed at the demonstration site was considered to be extreme compared to typical sites.
Whereas the process will operate on fully saturated soils, the energy and time-related costs associated
with removal of water from the soil favors drier soil conditions. At the Parsons Chemical site, the satu-
rated soil conditions, the ability of the high clay content soil to absorb water, and the probability that
additional water entered the treatment zone from the sandy layer during the project, resulted in unusu-
ally high water removal requirements at this site. The soil moisture content and other properties also
had the effect of producing melts with a higher width:depth ratio than was expected. This resulted in
the necessity to overmelt, widthwise, into clean soil in order to allow additional time to reach the desired
melt depth. Such overmelting has a significant cost impact. Geosafe believes that these factors con-
tributed to the cost of this project being 10 to 20% higher than might be expected at more favorable
sites. It should be noted that the Parsons Chemical site was Geosafe's first large-scale remediation
project. A large number of improvements to the equipment and operating methods were made during
and since the project with the effect of improving overall operating efficiencies. Geosafe's current cost
estimate for ISV treatment of typical nonradioactive sites within the U.S. is in the range of $350-450/ton.
Lastly, Geosafe notes that the ISV technology has significant adaptability to varying site condi-
tions. For example, modifications can be made to the off-gas treatment system to accommodate un-
usual contaminant conditions or specific State emissions standards as required. In addition, there is a
broad range of application configurations that may be employed to most economically treat contaminat-
ed materials. The ISV technology may be a preferred technology for challenging sites due to its unique
capabilities, including: 1) the ability to simultaneously treat mixtures of contaminant types, 2) high
treatment efficiencies (contaminant destruction, removal, immobilization), 3) high volume reduction (25-
50% for soils), and 4) onsite and in situ safety benefits.
REFERENCES
The reader is referred to the full suite of SITE Demonstration Program documents and the video
for details regarding this demonstration. The reader is referred to Geosafe Corporation literature for
details regarding other test, demonstration, and remediation projects that have been performed using
the ISV technology.
FOR MORE INFORMATION:
Ms. Teri Richardson, Technical Project Manager
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7949
119
-------
UNTERDRUCK-VERDAMPFER-BRUNNEN (UVB):
AN IN SITU SYSTEM FOR REMEDIATION OF CONTAMINATED AQUIFERS
Michelle A. Simon
US EPA
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
(513) 569-7469 (phone) (513) 569-7676 (fax)
Roger R. Argus and Benjamin L. Hough
PRC Environmental Management, Inc.
4065 Hancock Street, Suite 200
San Diego, CA 92110
(619) 225-1883 (phone) (619) 225-9985 (fax)
INTRODUCTION
Traditionally, contaminated groundwater is pumped to a surface facility
for treatment, often by air stripping. An innovative technology, the
Unterdruck-Verdampfer-Brunnen (UVB), German for Vacuum Vaporizing Well, is an
in situ groundwater remediation technology that combines air-lift pumping and
air stripping to clean aquifers contaminated with volatile compounds.
Additionally, the developer claims that in some cases the technology is
capable of simultaneous recovery of soil gas from the vadose zone. An
evaluation of this process is discussed in this abstract.
A UVB system consists-of a single well with two hydraulically separated
screened intervals installed within a single permeable zone. The air-lift
pumping occurs in response to reduced pressure introduced at the wellhead by a
blower. This blower creates a vacuum that draws water into the well through
the lower screen and the water rises toward the top of the well. A
submersible puma ensures a flow rate of approximately 20 gallons per minute
(1.26x10" meter per second). Air stripping occurs as ambient air, also
flowing in response to the vacuum, is introduced through a sieve plate located
within the upper screened section of the well.
Air bubbles form in the water column causing volatile compounds to
transfer from the aqueous to the gas phase. The rising air transports
volatile compounds to the top of the well casing, where they are removed by
the vacuum blower. The blower effluent is passed through granular activated
carbon before being released to the atmosphere.
The transfer of volatile compounds is further enhanced by a stripping
reactor located immediately above the sieve plate. The stripping reactor
consists of a fluted and channelized column that facilitates the transfer of
volatile compounds to the gas phase by increasing the contact time between the
two phases and by minimizing the coalescence of air bubbles.
Once the upward stream of water leaves the stripping reactor, the water
falls back through the well casing and returns to the aquifer through the
120
-------
upper well screen. This return flow to the aquifer, coupled with the inflow
at the bottom of the well, circulates groundwater around the UVB well. The
extent of the circulation pattern is known as the radius of influence, which
determines the volume of water affected by the UVB.
The UVB technology is a process patented by IE6 mbH in, Reutlingen,
Germany. IE6 Technologies, Inc., located in Charlotte, NC, markets the
technology in North America. I EG teamed with Roy F. Weston, Inc. to.
demonstrate the UVB technology at March Air Force Base (AFB), CA. March AFB
allowed the US EPA Superfund Innovative Technology Evaluation (SITE) program
to evaluate the technology. The SITE program retained PRC Environmental, Inc.
to evaluate the performance of the UVB system "at March.
METHODOLOGY
The objective of the demonstration was to determine the efficiency of
UVB system, the reduction in concentration of trichloroethene (TCE) in the
aquifer, and the radius of influence of the UVB system within the aquifer.
Two series of wells were placed downgradient of the UVB system, as
depicted in Figure 1. The first series of wells were placed 40 feet from the
system; the second 85 feet away. Each of the two series had three wells: a
shallow well screened at a depth of 37 tb 57 feet below ground surface (bgs),
an intermediate well screened at a depth of 65 to 75 ft bgs, and a deep well
screened from 90 to 105 ft bgs. The two shallow wells were screened at the
same depth as the UVB system outlet; the intermediate wells were screened at
the same depths as the UVB inlet. The deep wells were screened below the
anticipated UVB circulation cell. The wells were sampled monthly and samples
were analyzed for volatile organic compounds.
A dye trace study was performed on the system. No dyes were detected in
the aquifer prior to start of this dye study. A two-dye approach was
implemented. Sodium fluorescein, a bright green dye, was injected in the
system outlet, in order to determine whether water discharged from the UVB
system could be detected in any of the surrounding wells. A bright yellow-
orange dye, rhodamine WT (red-purple xanthene) was injected into the
intermediate well placed 40 feet downgradient of the UVB system, to determine
whether this well was within the system's radius of influence.
RESULTS
Figure 2 is a graph of the change of the TCE concentration for the inlet
versus outlet water. When the system operated properly, the inlet TCE
concentrations averaged 33.75 micrograms per liter (M9/L) and the outlet
concentrations averaged less than 1 fig/L (97% removal efficiency). Twice
during the twelve month period, the system required adjustments.
Figure 3 is a graph of concentration of TCE versus time for the wells
located 40 feet downgradient of the UVB system. Prior to UVB system startup,
samples from the shallow well had TCE concentrations of 530 ng/L; the
intermediate well - 750 /jg/L; and the deep well - 100 /ig/L. Initially, the
TCE concentration in the water from all six wells remained constant or
121
-------
&
5
I
I
122
-------
increased. The TCE concentration in samples from the shallow well and the
intermediate well increased to peak values of 620 and 2000 /jg/L, respectively,
after three months of system operation. Then, the concentration of TCE
declined. TCE concentration in the deep well remained constant at 110 jig/L.
Figure 4 is a graph of the TCE concentration for the wells located 85
feet downgradient of the UVB system. The TCE concentrations in the samples
collected from the outer well cluster initially increased after system
startup, also. However, the concentrations peaked after seven months of UVB
operation. The shallow well initially had a TCE concentration of 650 [ig/l,
which increased to 980 after seven months and then reduced to 290 after twelve
months. Samples from the intermediate well measured 120, 640, 210 /ig/L TCE,
for the same respective time periods.
As for the dye study, the fluorescein dye that was injected into the
outlet of the UVB was observed at the inner shallow well after 48 days. It
was also observed in the deep inner cluster well after 59 days. It was never
observed in the inner intermediate well which is located between the these
two. The fluorescein dye was not observed in any other wells for the four
month duration of the test, including those located 85 feet downgradient of
the UVB.
The rhodamine dye that was injected into the inner intermediate well was
observed upgradient at the UVB system after 85 days. The rhodamine dye was
also observed in the inner deep well as soon as 17 days. Rhodamine dye was
not detected in any other wells.
CONCLUSIONS
The UVB system can successfully remove TCE from inlet contaminated
groundwater to a level less than 1 ng/L in the outlet under normal operating
conditions. The dye study confirmed that the radius of influence for this
application is at least 40 feet. A modeling study performed by the vendor
predicted a radius of influence of approximately 80 feet. The TCE
concentrations in samples from wells within the radius of influence initially
increased but eventually declined to one-half their original levels after 12
months of treatment. March AFB continued to operate the UVB system for an
additional 6 months and monitored its performance. These data show a
continual decline in TCE concentration for the wells affected by the UVB.
FOR MORE INFORMATION
Contact Michelle A. Simon at the address above.
123
-------
260
0 1 234 6 6 7 8 9 li> 11 12
MONTH AFTER STARTUP
FIGURE 2
UVB SYSTEM INLET AND OUTLET
012346S78S10 11 12
MONTH AFTER STARTUP
FIGURE 3
INNER CLUSTER OF WELLS
2600
ui
o
o
u
Ul
o
600
MONTH AFTER STARTUP
FIGURE 4
OUTER CLUSTER OF WELLS
8 9 10 11 12
124
-------
SITE DEMONSTRATION OF TERRA-KLEEN RESPONSE GROUP'S MOBILE SOLVENT
EXTRACTION PROCESS
Mark C. Meckes
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 46268
(513)569-7348
Scott W. Engle and Bill Kosco
PRC Environmental Management, Inc.
644 Linn Street, Suite 719
Cincinnati, Ohio 45203
(513)241-0149
INTRODUCTION
In 1986 the Environmental Protection Agency (EPA) Office of Solid Waste and Emergency
Response (OSWER) and the Office of Research and Development (ORD) established the SITE
Program to Promote the development and use of innovative technologies to clean up uncontrolled
hazardous waste sites across the country. The SITE .Program is composed of four major elements:
the Demonstration Program, the Emerging Technologies Program, the Measurement and Monitoring
Technologies Program and the Technology Transfer Program. The Demonstration Program is
designed to provide engineering and cost data for selected technologies, by evaluating their ability to
treat wastes from Superfund sites.
The Terra-Kleen Response Group (TKRG) requested that EPA's SITE Program evaluate their
mobile solvent extraction technology, and was selected for evaluation under the Demonstration
Program. This technology is a batch process system which uses proprietary solvents to separate
organic contaminants from soils, sediments, and/or sludges. Organic contaminants are concentrated
during processing, thus reducing the volume of hazardous wastes for final disposal. Therefore, this
technology is non-destructive.
TKRG's solvent extraction process is transportable and can be configured to treat both small or
large quantities of solids. System components are often available from local suppliers throughout the
United States. This reduces setup time and can reduce the amount of down time associated with
equipment replacement.
Processing begins following excavation of contaminated solids and loading them into extraction
vessels. The vessels are covered and clean solvent at ambient temperature and pressure is pumped
into each one. Organic contaminants in the solids leach into the solvent without the aid of a mixing
device. Contaminated solvent then flows into a clar'rfier where heavy solids are separated by gravity
from the solvent. Clarified solvent is pumped through a mfcrofilter which removes fines, and then
through a proprietary regeneration unit which concentrates the organic contaminants. Clean solvent,
discharged from the regeneration unit, is stored in a holding tank for reuse. This sequence of treatment
steps is repeated until contaminant concentrations of the solids within the extraction vessels are
reduced to a desired level. Some residual solvent still remains with the solids. Therefore, at this point
in the process all solvent carrying lines are drained and vacuum extraction is used to reduce the
concentration of solvent in the solids.
Treated solids are typically removed from the vessels by a front end loader and returned to the
site. Concentrated contaminants are removed and disposed of off-site in accordance with applicable
regulations. Regenerated solvent may be used for treatment of solids at other waste sites.
125
-------
The SITE Demonstration of this technology was conducted during May 1994 in cooperation with
the Naval Environmental Leadership Program (NELP) at Navel Air Station North Island (NASNI) which
is located near San Diego, California. TKRG was contracted by NASNI to treat five tons of
polychlorinated biphenyl (PCS) contaminated soil. This project was considered to be a pilot-scale
demonstration of the capabilities of the TKRG's solvent extraction process. The primary objective of
this demonstration was to determine if the process could achieve a soil clean up level of <2.0 mg/kg
total PCB.
METHODOLOGY
A backhoe was used to excavate five tons of PCB contaminated soil from NASNI Site 4. The
excavated soil was then homogenized by using the front end loader of the backhoe for mixing. Five
extraction vessels were tared and each were filled with approximately one ton of the homogenized soil
and weighed. A sampling grid was laid out across the top of each vessel and a core sampler was used
to collect seven samples from each vessel. The seven samples were composited by vessel such that a
composite sample represented the contents of an individual vessel. These samples were analyzed for
PCBs, volatile organic compounds (VOCs), and semivolatile organic compounds (SVOCs) in
accordance with Test Methods for Evaluation of Solid Wastes Physical/Chemical, 1992 (SW-846).
Following sample collection clean solvent was added to each extraction vessel and permitted to
drain into a clarifier. Samples of the extraction solvent were periodically collected from the drain lines
leading to the clarifier. The clarified solvent was pumped through a 5 micron bag filter and TKRG's
proprietary solvent regeneration system. Samples of the regenerated solvent were collected
periodically and analyzed for PCBs. Regenerated solvent was pumped into a storage tank and was
held there until it was reutilized for subsequent extraction cycles.
When the solvent in all of the five extraction vessels had finished draining to the clarifier, the
drain lines were closed and a second extraction cycle was initiated. The extraction cycles were
continued until the measured concentration of PCBs in the drained solvent was <1.0 nng/L. At this
point, the suction side of a centrifugal blower was connected to the drain lines of each extraction vessel
and was operated continuously for three days. This was done to remove residual solvent from the
solids. Following this treatment a mixture of nutrients and microorganisms were added to the extraction
vessels to biologically degrade any remaining solvent. After two weeks of this biological treatment a
sampling grid was reestablished for each extraction vessel and seven core samples of the solids were
collected, composited, and analyzed as described above.
RESULTS
The characteristics of the soils obtained from each extraction vessel prior to treatment are
shown in Table 1. Untreated soil was predominantly a dry sand with an average moisture content of
0.83 percent and 93.6 percent of the solids retained on a 0.075 mm screen. Polynuctear aromatic
hydrocarbons (PAHs), hexachlorodibenzo furans (HxCDFs), and pentachlorodfcenzo furans (PeCDFs)
were identified, but only at very tow concentrations. Bis(2-Ethylhexyl)phthalate was also identified at
<0.7 mg/kg in the untreated soils. Other analyses showed the average Oil and Grease concentration at
759 mg/kg and the average total PCB concentration at 144 mg/kg.
126
-------
TABLE 1
PHYSICAL AND CHEMICAL CHARACTERISTICS OF UNTREATED SOILS
Extraction Vessel
Particle Size
(% >0.075 mm)
Moisture Content (%)
Total PAHs (mg/kg")
Total HxCDF (ug/kg")
Total PeCDF (ug/kg')
Total PCB (mg/kg")
Oil & Grease (mg/kg")
A
92.7
0.79
2.10
0.658
<.409
130
753
B
93.4
0.79
2.23
0.629
0.144
140
726
C
934
0.79
3.38
0.647
<.343
134
713
D
93.5
0.80
1.55
0.848
0.162
146
737
E
95.0
0.99
1.94
0.704
0.218
170
868
Avg.
93.6
0.83
2.24
0.697
<.255
144
759
S.D.
• 0.85
0.09
0.69
0.089
-
15,7
62.5
* dry weight
A total of 11 extraction cycles were completed over 7 days for each of the 5 one ton batches of
contaminated soil. Three days of vacuum extraction were then completed to reduce the concentration
of residual solvent in the treated solids. Upon completion of vacuum extraction, a mixture of
microorganisms and nutrients was added to each extraction vessel. After two weeks soil samples were
collected from each vessel and analyzed for contaminants.
Low initial contaminant concentrations at or pear method detection limits precludes the use of
PAH, HxCDF, and PeCDF data for evaluation of the process. The results of PCB and Oil and Grease
analyses and removal efficiencies for each are shown in Table 2.
Solvent concentrations in treated solids ranged from 4.7 to 3.6 percent prior to vacuum
extraction. Following three days of vacuum treatment residual solvent in the soBds was found to be
between 2.7 and 1.8 percent. Fourteen days of biological treatment further reduced solvent
concentrations to <0.5 percent.
TABLE 2
CONTAMINANT CONCENTRATIONS AND REMOVAL EFFICIENCIES OF TREATED SOILS
Extraction Vessel
Oil and Grease (mg/kg")
Removal Efficiency (%)
Total PCB (mg/kg)
Removal Efficiency (%)
A
327
56.6
1.79
98.6
B
300
58.7
1.63
98.8
C
252
64.6
1.79
98.7
D
167
77.3
1.88
98.7
E
318
63.4
1.96
98.8
Avg.
273
64.0
1.81
98.7
S.D.
65.8
8.1
0.12
0.1
dry weight
127
-------
CONCLUSIONS
TKRG's solvent extraction process was effective in removing PCBs and Oil and Grease from
dry sandy soils. The extraction process was operated for 11 cycles to achieve a predetermined
remediation goal of <2.0 mg/kg PCS in soil. Oil and Grease concentrations in the contaminated soils
were also reduced to <330 mg/kg. All five batches treated met the remediation goal demonstrating that
treatment of solids was consistent. Vacuum extraction combined with biological treatment was able to
reduced the concentration of residual solvent in treated solids to <0.5 percent. Further reduction of
residual solvent concentrations in solids may have been possible if treatment would have been allowed
to continue.
FOR MORE INFORMATION:
Contact: Mark C. Meckes
U.S. EPA
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
128
-------
ELECTROKINETIC REMEDIATION: TECHNOLOGY STATUS
YalcinB. Acar
Civil and Environmental Engineering Department
Louisiana State University
Baton Rouge, LA 70803
Phone:(504)388-8638
Akram Alshawabkeh, Robert J. Gale, Robert E. Marks and Susheel Puppala
ELECTROKINETICS INC.
The Louisiana Business and Technology Center
Suite 105, South Stadium Drive
Baton Rouge, LA 70803
(504)388-3992 ,
Randy Parker
Risk Reduction Engineering Laboratory, USEPA
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7271
and
MarkBricka
US Army Engineer Waterways Experiment Station
Hal's Ferry Road
Vicksburg, Ml 39180
(601) 634-3700
INTRODUCTION
The demand to develop innovative and cost-effective in-situ remediation technologies in waste
management stimulated the effort to employ conduction phenomena in soils |using an electric field to
remove chemical species from soils (1-6). This technique variably named as elecfrokinetic remediation,
electro-reclamation, electrokinetic soil processing, electro-chemical decontamination , electrorestoration
or electrochemical soil processing uses low-level DC in the order of mA/cm2 of cross sectional area
between the electrodes or an electric potential difference in the order of a few volts per cm across
electrodes placed in the ground in an open flow arrangement. The groundwater in the boreholes or an
externally supplied fluid (processing fluid) is used as the conductive medium. Open flow arrangement at
the electrodes allows ingress and egress of the processing fluid or the pore fluid into or out of the porous
medium. The low-level DC results in physico-chemical and hydrological changes in the soil mass leading to
species transport by coupled and uncoupled conduction phenomena in the porous media. Electrolysis
reactions prevail at the electrodes. The species input into the system at the electrodes (either by the
electrolysis reactions, or through the cycling processing fluid) and the species in the pore fluid will be
transported across the porous media by conduction phenomena in soils under electric fields. This
transport coupled with sorption, precipitation and dissolution reactions comprise the fundamental
mechanisms affecting the electrokinetic remediation process. Extraction and removal are accomplished by
electrodeposition, precipitation or ion exchange either at the electrodes or in an external extraction
system placed in a unit cycling the processing fluid (1,2,4).
129
-------
Electrokinetic remediation technology has recently taken significant strides. Electrokinetics Inc. of
Baton Rouge has completed large-scale pilot studies using spiked and naturally contaminated soil
deposits under the USEAPA SITE program. In collaboration with the US Army Waterways Experiment
Station, Electrokinetics Inc. is currently carrying out a field study of extracting lead from soils at a Firing
Range that belongs to the US Army. This demonstration study will be independently evaluated by the
USEPA under the SITE program. The purpose of this paper is to present some of the recent
developments in this technique and to outline the ongoing activities.
BACKGROUND
Three pilot-scale studies were conducted by Electrokinetic Inc. under the SITE program; two pilot-scale
tests using kaolinite spiked with lead at initial concentrations of 850 mg/kg, 1,500 mg/kg and another
using fine sand and kaolinite mixture spiked with lead at 5,322 mg/kg. The kaolinite used had lead
adsorption capacity of about 1,100 mg/kg. Lead nitrate salt is used as the source of lead. Tap water is
used both as the catholyte and the anolyte. Other details of testing are presented by Alshawabkeh and
Acar (3) and EK (5). Figure 1 shows the lead concentration profile after 123 days of processing at a
current density of 133 uA/cm2. More than 90 % of the lead in the soil is transported across to the cathode
compartment. Lead prematurely precipitates close to the cathode compartment at its hydroxide solubility
value if the chemistry of the electrolyte at the electrodes is not altered or controlled (unenhanced
electrokinetic remediation). One objective of these pilot-scale tests was to formalize and validate the
principles of multi-species transport under electric fields. An appreciation of the relation between the
mechanics and chemistry is only possible when precipitation close to the cathode compartment is allowed.
Therefore, pilot-scale tests did not employ any enhancement technique.
The development of the theoretical formalisms pertaining to multi-species transport under electric fields,
preparation of the associated numerical model and design/analysis packages are supported by the
USEPA under the Gulf Coast Hazardous Substance Research Center program at Lamar University. The
total lead profile predictions of the model pertinent to the the specific initial and boundary conditions in
2000
1800
1600
1400
1200
1000
800
600
400
200
Figure 1. Lead concentration profile after 123 days of processing kaolinite/fine sand mixture spiked at
5,322 mg/kg (3).
130
-------
the pilot-scale test presented in Fig. 1 are given in Figure 2. The acid generated by the electrolysis
reactions at the anode and the lead released in to the pbre fluid either by the dissolution or by the
aqueous phase reactions travel towards the cathode compartment under the multi-species transport
phenomena in soils under electric fields. Lead precipitates close to the cathode compartment at its
hydroxide solubility value with the increase in the hydroxide concentration due to the prevailing
cathodic reaction. The 50 day predictions shown in Fig. 2 are compared with the pilot-scale test results in
Fig. 3. The agreement between the theoretical model and the pilot-scale test results demonstrate that the
principles of the process are quite well rationalized and understood. The design analysis package which
will include the generalized model is developed through joint collaborative effort between Electrokinetics
Inc. and US Army Waterways Experiment Station.
The need to overcome the shortcoming of precipitation close to the cathode compartment prompted
Electrokinetics Inc. and the US Army Waterways Experiment Station to collaborate and evaluate the
feasibility of employing different techniques in enhancement of the process. The objective of the study
was to promote transport of the positively charged species in to the catholyte where they could be
removed either by electro deposition, by membrane separation or through ion exchange schemes. Acar
and Alshawabkeh (1) discuss further the heed for enhancement and propose different enhancement
techniques which prevent the encountered precipitation. Figure 4 shows the extraction of different
species in time from a 'real world' soil using the acetic acid cathode depolarization technique. The solubility
of the species by the anodic acid and their migration under electrical fields are important considerations in
eiectrokinetic remediation. The most soluble species will be the first ones to come out in solution and
hence be transported towards the respective electrodes.
^ 140DD
? 12000
f 10000
Q SOOD
600D
40 OP
20DP
Figure 2. Prediction of total lead concentration across the electrodes using the Finite Element Model for
Multi-Species Transport in Soils under Electrical Fields (FEM-MTSE) [Boundary and Initial Conditions
pertinent to the Pilot-Scale Study are employed; 100 elements are used] (3,5].
131
-------
—•—Model
—O -PST3 Top Layer
-El -PST3 Middle Layer
--EB--PST3 Bottom Layer
CB
o
H
0 10 20 30 40 50 60 70
Distance From Anode (cm)
Figure 3. Comparison of lead profile predicted by the FEM-MSTEF with those measured in the pilot-scale
study (3,5).
Although acetic acid depolarization technique is successfully used in remediating lead, chromium, zinc
and other heavy metals from soils retrieved from sites across the nation and also from Europe, the
technique generates significant amounts of liquid that needs to go through a secondary processing.
Electrokinetics Inc. has developed an electrode system (CADEX™) that promotes electrodeposition of
the species and minimizes the need for secondaiy processing.
In collaboration with the US Army Waterways Experiment Station, Electrokinetics Inc. is currently
conducting a field-scale demonstration study at Fort Polk, Louisiana. The site is located in a creek. The
surface deposits within the first two feet are contaminated with lead at concentrations of 3,500 mg/ kg (±
500 mg/kg). The lead in the deposits leached from the bullets left in the firing range over years of
exposure to the environment. Preliminary chemical speciation and corrosion studies indicate that minute
quantities of lead leach from the individual particulates and the contamination is mainly due to the quantity
of the bullets left at the site. An area of about 2,000 ft2 at this site is being processed. CADEX™
electrode system is used at this site. Onsite remediation is expected to continue for about six months to a
year and the goal of the study will be to continue the process until levels of 100 mg/kg or less are reached.
USEPA will have an independent evaluation of the demonstration study through the SITE program.
SUMMARY AND CONCLUSIONS
Electrokinetic extraction technique has gone through the phases of bench-scale testing and pilot-scale
testing. Precipitation of species close to the cathode compartment has been a bottleneck for the
process. Acetic acid depolarization technique and other depolarization schemes have been developed
(4). Extraction of heavy metals by bench and pilot-scale tests from 'real world' soils retrieved from sites
across the nation and from Europe demonstrate that the technique may be efficiently and cost-effectively
used at selected sites. Currently, a field demonstration study is ongoing in the USA.
A theoretical model has been developed and its numerical implementation has been completed. The
predictions of this model compared with the results of the pilot-scale studies demonstrate that the
132
-------
500 1000 1500 2000 2500 3000
TIME (h)
Figure 4. Breakthrough of Different Species in Enhanced Electrokinetic Remediation of a Specimen from
a US Army site (EK1994)
principles of the technique have been well rationalized. Currently, the model is being generalized and
conditioning schemes are being incorporated. Design/analysis packages are being developed.
Construction guidelines will be written upon completion of the field study. When cost-effectiveness and
technical feasibility of other remediation optionsprohibit their use, electrokinetic remediation is expected
to offer an alternative at sites contaminated with inorganic species. The technique also is in the process Of
being scrutinized and developed for injection, specifically to engineer insitu bioremediation of organic
species through injection of process additives, nutrients and microorganisms.
REFERENCES
1. Acar, Y.B., Alshawabkeh, A. Principles of Electrokinetic Remediation . Environmental Science and
Technology, vol. 27. n. 13. pp. 2638-December 1993.
2..Acar, Y. B., Gale R, J., Alshawabkeh, A., Marks, R. E., Puppala, S., Bricka, M., Parker, R. "Electrokinetic
Remediation: Basics and Technology Status," Journal of Hazardous Materials. Elsevier Science B.V.,
Amsterdam, Netherlands, 40(3), February 1995, pp. 117-137.
3. Alshawabkeh, A., Acar, Y. B., "Electrokinetic Remediation: I. Pilot-Scale Tests; II. Theory" Journal of
Geotechnical and Gedenvironmental Engineering. ASCE. (in review)
4. EK(1994). An Investigation of Selected Enhancement Techniques in Electrokinetic Remediation,
Report submitted to US Army Waterways Experiment Station, Electrokinetics Inc., Baton Rouge,
Louisiana, 1993, 160 p.
5. EK(1995). Theoretical and Experimetal Modeling of Removing Contaminants from Soils by an Electrical
Field, Report submitted to USEPA by Electrokinetics Inc., Baton Rouge, Louisiana, 375 p. (in press).
6. Lageman, R.; Wieberen, P.; Seffinga, G. Electro-ReclamationfTheory and Practice. Chemistry and
Industry, Society of Chemical Industry, London, pp. 585-590, 1989.
7. Shapiro, A. P.; Probstein, R.F. (1993)" Removal of Contaminants from Saturated Clay by
Electroosmosis," Environmental Science and Technology. 1993, 27 (2), pp. 283-291.
133
-------
SITE Demonstration of Bioremediation of Cyanide
at the Sumrnitville Colorado Site
Leslie C. Thompson
Pintail Systems, Inc.
11801 E. 33rd Ave. Suite C
Aurora, CO 80010
(303) 367-8443
Randy Fischer
Dames & Moore
633 17th Street, Suite 2500
Denver, CO 80202
(303) 294-9100
Scott W. Beckman
Science Applications International Corp
411 Hackensack Ave, 3rd Floor
Hackensack, NJ 07601
(201) 489-5200
Introduction
The Sumrnitville Mine in southern Colorado is located in the San Juan mountains at an
average site altitude of 11,500 feet. Sumrnitville was the site of mining operations that began in 1873
with the discovery and development of gold placer and lode deposits. The site was actively mined for
gold, silver and copper between 1873 and 1947. From 1947 to 1986 the mine area was inactive until
the Sumrnitville Consolidated Mining Corporation, Inc. (SCMCI) a wholly-owned subsidiary of Galactic
Resources, Ltd. started an open pit mine and heap leach operation at the site.
SCMCI ran a large tonnage open-pit and cyanide heap leach operation from 1986 to 1992.
Gold ore .(approximately 10 million tons) was mined, crushed and stacked on a lined, bowl-shaped
leach pad. The mine experienced problems with water balance and unplanned solution discharges
from the start of the mine life. Solution containment complications and ineffective water treatment
contributed to environmental problems. Despite the production of 249,000 troy ounces of gold during
the mine operation SCMCI was unable to meet remedial requirements arid notified the state of
Colorado of its intention to file a Chapter VII bankruptcy in December 1992. The EPA Region VIII
Emergency Response Branch took over site operations on December 16, 1992 to prevent a cata-
strophic release of hazardous substances to the environment. The Sumrnitville Mine site was added
to the National Priority List in June 1994.
There are multiple sources of contamination, at the site due to historic and SCMCI mining
operations. Emergency response operations at the site have prevented releases of severely
contaminated solution and studies are underway to define a permanent solution to detoxification or
neutralization of the various mine waste units. This report addresses demonstration of an innovative
bioremediation technology for treatment of cyanide and soluble teachable metals in the heap and
heap solutions.
The heap leach pad consists of approximately 10 million tons of cyanide-leached ore and 90
to 150 million gallons of process solution. EPA Region VIII commissioned a Focused Feasibility Study
(FFS) and Report of Investigation (Rl) to evaluate remedial options for the Heap Leach Pad (HLP).
The RI/FFS was completed by Morrison Knudsen Corporation and submitted to EPA Region VIII on
August 19, 1994.
A Request for Proposal (RFP) was issued by Environmental Chemical Corporation (ECC) in
October 1993 at the request of USEPA Region VIII, Department of the Interior and Bureau of
Reclamation. The RFP requested interested companies to provide information on their ability to
134
-------
implement innovative treatment technologies to improve treatment efficiency and reduce cost of
treatment of the heap leach pad spent ore and leachate solutions. Dames & Moore and Pintail
Systems, Inc. (PSI) jointly submitted a proposal suggesting application of biotreatment processes for
treatment of the spent ore and process solutions in the HLP. The proposal was accepted for
feasibility demonstration under the EPA Superfund Innovative Technology Demonstration Program
(SITE) with additional funding from EPA Region VIII.
The primary objectives of the Dames & Moore/Pintail Systems proposal were to:
1. Demonstrate the feasibility of spent ore and process solution cyanide bio-detox.
2. Develop site-specific biotreatment processes for spent ore and process solution cyanide
detox.
3. Provide treatment data for use in the RI/FS and Record of Decision (ROD) for the Spent Ore
and Entrained Solutions operable units at the Summitville Mine.
4. Immobilize potentially teachable metals including zinc, copper, manganese, iron and arsenic
within the heap to improve water quality.
Tests and demonstrations outlined in the proposal were conducted in PSI's Aurora, Colorado lab
and pilot plant and at the mine site. Spent ore treatability testing included waste characterization,
bacteria isolation and bioaugmentation, parallel column treatment tests, data evaluation and reporting.
Biotreatment processes for heap, tailings and process solution detox have been proven at other
mine sites in a variety of environments. Biological processes are both site-specific and waste-specific
and must be individually engineered and tested for each "mine waste. Successfully adapting treatment
bacteria to the spent ore environment is a key to developing successful bioremediation potential.
Working with a biotreatment population that has been specifically adapted to the ore and augmented
to improve cyanide metabolism insures that biotreatment will be effective.
Cyanide metabolism is known to occur in several species of bacteria. Bacteria that have the
capacity for enzymatic hydrolysis of ionic cyanide or metallo-cyanide compounds use the carbon
and/or nitrogen of the cyanide to meet their nutritional needs. The end-products of cyanide metabo-
lism are natural and non-toxic.
LABORATORY WORKPLAN
Bacteria Source - Isolation and Development
Cyanide decomposition bacteria were isolated from Summitville spent ore samples and were
augmented for remediation testwork with mine ores, tailings and waste rock. Preparation for the
column tests included:
1. Isolating native bacteria from the spent ore;
2. Adapting the treatment bacteria to the .Summitville Mine spent ore in a series of stress and
waste infusion media;
3. Characterizing growth and enhancing the new "cyanooxidans" population.
4. Demonstrating bacteria growth and cyanide decomposition in an ore leach flask test.
135
-------
Column Test Design
The following workplan was used for the column tests:
1. 75-85 kg of spent ore collected by SAIC and the SITE program was loaded into each of six
6"x10' PVC columns. The columns were-fitted with a perforated screen and a tapped end-cap
to-allow treatment solutions to percolate through the ore and be collected for analysis. The
columns were set up as follows:.
Column #1: sulfide zone ore, percolation leach biotreatment;
Column #2: oxidized ore, 25-90 ft depth, percolation leach biotreatment;
Column #3: oxidized ore, 90-130 ft depth (saturated zone), saturated with HLP solution,
percolation biotreat;
Column #4: oxidized ore, 0-90 ft depth, rinsed zone (1993 peroxide rinse program),
percolation leach biotreatment;
Column #5: oxidized ore, 0-25 ft depth, percolation leach biotreat;
Column #6: Control Column, oxidized ore, 90-130 ft depth (saturated zone), saturated with
HLP solution, percolation leach with peroxide-treated HLP solution.
2. The Detox population of bacteria was grown to working strength and transferred to a dilute
nutrient solution for application in a percolation leach.
3. Thirty gallons" of barren solution supplied by SAIC and ECC were used to saturate the ore in
Columns 3 and 6 to simulate treatment in a saturated ore zone.
4. The treatment solutions were applied to each column at a nominal rate of 0.004 gpm/ft2.
5. Bio-leach or detox barren solutions from the columns were collected and analyzed for total
cyanide. Total cyanide, weak acid dissociable (WAD) cyanide, gold and select metals were
analyzed in column leachate solutions during the course of the test.
6. Data was collected to allow calculation of contaminant reduction related to tons of treatment
solution applied per ton of ore.
7. Leachate solutions were analyzed for copper to determine metal mobilization due to bacteria
processing. Metal analyses were run on an Inductively Coupled Spectrophotometer and ah
Atomic Absorption Spectrophotometer.
8. Split samples of spent ore and column leachate solutions were collected by SAIC and were
submitted to a contract laboratory for confirmation analysis,. „ ; '- •
Pilot Test Data
The data collected from the pilot column ore treatment program is presented in Figures 1 and 2.
Treatment .compliance for successful cyanide detoxification was 0.2 mg/L WAD cyanide measured in
column leachate solutions. The control peroxide rinse column of saturated zone ore did not achieve
compliance with a WAD cyanide standard. All other column treatments reached compliance levels.
Total and WAD cyanide were plotted against the tons of solution applied per ton of ore. The amount
of biotreatment solution required for complete cyanide detoxification is projected to be 25-30% of the
amount of solution required by conventional chemical rinse detox treatments.
136
-------
SUMM11VILLE FFS COLUMN TESTS
LEACHATE SOLUTION WAD CYANIDE
100 -3
0.2
0.4 0.6 0.8 1 1.2
TONS SOLUTION PER TON ORE
1.4
1.6
— •— C1 SULFIDE BIO
-<&— C2 OXIDE BIO
— *- C3 OXIDE SAT BIO
-fr— 04 OXIDE BIO
-&- C5 0-25 FT BIO
-Q- C6 CONT DETOX BARREN
Figure 1. Column Leachate Solution WAD Cyanide
100 -3
SUMMITVILLE FFS COLUMN TESTS
COLUMN LEACHATE TOTAL CYANIDE
0.2
0.4 0.6 0.8 1 1.2
TONS SOLUTION PER TON ORE
C1:SULF!DE
CZ- OXIDE
C3: OX, SATBIOTREAT
1.4
C4: OXIDE BIO
C5: OXIDE BIO
C6:SAT DETOX BARREN
Figure 2. Column Leachate Total Cyanide
137
1.6
-------
A secondary treatment goal in the Focused Feasibility Study was to reduce the amount of
teachable or soluble metals in the spent ore and entrained heap leachate solution. Pintail Systems
has observed a reduction of soluble metals such as copper in other pilot and field treatment
applications. This test was designed to quantify reduction of metals in column leachate solution, and
to identify any remineralized products in the column tests. Copper in column leachate solutions is
shown in Figure 3 for Column 3 and Column 6. Column 3 was the biotreatment test column run as a
saturation zone sample. Column 6 was a control rinse using a barren solution detoxified with
hydrogen peroxide in a saturation rinse.
Numerous species of bacteria, fungi and yeasts are capable of accumulating many times their
weight In soluble metals. Both living and dead biomass are effective in removing soluble metals from
waste streams containing gold, silver, chromium, cadmium, copper, lead, zinc, cobalt and others.
Bacteria found in natural and extreme environments have developed a wide variety of metabolic
functions to adapt to these environments. These natural microbial functions contribute to global
mineral cycling that continuously forms, transforms and degrades minerals and metals in the
environment. Biomineralization is described as a surface process associated with microorganism cell
walls where the remineralization occurs. The biogeochemical activities initiated by microorganisms in
ores, soils, surface and groundwater environments can dominate the formation and transformation of
those mineral environments.
The metal remineralization process is catalyzed by biological processes alone and by biological
processes initiating physical and chemical processes causing an alteration of the micro-environment.
During the course of the column tests a series of observations were made on the changing surfaces
in the test column. These observations are the basis of the following hypothesis for formation of bio-
minerals.
1. Bacteria added to the ore columns attach to the ore surfaces forming a "bio-slime" layer,
2. Soluble metals bind to cell walls and extra-cellular products excreted by the microorganisms
(exopolymers, pigments, waste organics, etc.),
3. Metal hydroxides, oxides and carbonates are formed in the primary "bio-slime" layer as
amorphous mineral pre-cursors. Curing or maturation of the amorphous slimes suggests that
a molecular rearrangement of the hydroxy-metals to more stable forms occurs,
4. Stabilization of the amorphous precipitates forms a remineralization nucleation crystal template
for further mineralization to occur. The micro-environment alteration and bacteria metabolism
continue to catalyze the remineralization by on-going formation of organometallic compounds,
precipitates and transformation of metal oxidation states. The biornineralization appears to
follow a sequential and "layered" development on many of the surfaces. Some of the possible
minerals formed include calcite, gypsum, bornite, pyrite and covellite.
CONCLUSIONS
Cyanide detox in spent ore is a function of solution application efficiency and bacterial use of
cyanide. The ore biotreatment in this test gave similar results and comparable detox time to prior PSI
experiences. The treatment bacteria adapted well to the spent ore environment and effected a rapid
detoxification of cyanide in spent ore and ore solutions. The Summitviile ore.is a suitable candidate
for a field biotreatment.
The biological treatment column achieved a >99% removal of weak acid dissociable cyanide with
138
-------
SUMMSTVSLLE FFS COLUMN TESTS
LEACHATE SOLUTION COPPER
too
(XL
LU
CL
C3_
o
o
1
0.01
C6 BARREN/H202 RINSE
0.1-
0.4 0.6 0.8 1
TONS SOLUTION PER TON ORE
Figure 3. Column Leachate Solution Copper
application less than an average of 0.5 tons of solution per ton of ore. Total cyanide in columrr
leachate solutions at the end of the test was <0.5 mg/L indicating that bacterial action in the
treatment solution will metabolize strong metal-cyanide compounds. A field treatment of a leach pad
cell or other division of spent ore could be planned for a treatment program using less than 0 5 tons
of solution per ton of ore.
Biotreatment in this study achieved a greater reduction in total cyanide in a shorter application
than chemical treatments can achieve. In Situ biotreatment is the most efficient heap cyanide detox
as compared to peroxide rinse treatments. The data generated in this study indicate that the
biotreatment processes have the potential to operate as an effective field treatment. Biological
treatments are projected to be cost and time competitive with chemical rinse treatments.
The objectives of the pilot column tests were met in this biotreatment demonstration.
1. Existing strains of cyanide-oxidizing bacteria were adapted to grow- in the ore environment and
to use cyanide as a carbon and/or nitrogen source,
2. Flask and column tests of the adapted, augmented treatment population verified that bacteria
would grow and metabolize soluble cyanide in the Summitville spent ore.
3, Cyanide was detoxified in biotreatment tests in spent ore and column leachate solutions with
application of less than an average of 0.5 tons of solutions per ton of ore. Cyanide levels did
not reach a 0.2 mg/L discharge criteria with the peroxide kill, saturated, barren rinse test
column with application of more than 1.5 tons of treatment solution per ton of ore.
139
-------
REDUCTIVE PHOTO-DECHLORINATION (RPD) PROCESS FOR
SAFE CONVERSION OF HAZARDOUS CHLOROCARBON WASTE STREAMS
Moshe Lavid, Suresh K. Gulati, Moisey Teytelboym
ENERGIA, Inc.
Research and Consulting
P.O. Box 470, Princeton, NJ 08542
609-799-7970
INTRODUCTION
A novel technology designated "Reductive Photo-Dechlorination" (RPD) has
been developed and successfully tested for environmentally safe treatment of
waste streams containing hazardous chlorinated hydrocarbons. This RPD process
employs ultraviolet (UV) light in a reducing atmosphere and at moderate
temperatures' to efficiently convert chlorocarbon contaminants into valuable
hydrocarbons such as methane, ethane, ethylene, acetylene and hydrogen
chloride. The UV light promotes carbon-chlorine bond cleavage and long-chain
radical reactions with the hydrogenous bath gas leading to the
thermodynamically and kinetically favored hydrocarbon products at a conversion
of +99%.
The RPD process is schematically shown in Figure 1. The pilot-scale
prototype consists of five main units: (1) Input/Mixer; (2) Photo-thermal
Reactor; (3) Scrubber; (4) Separator/Storage,- and (5) Recycling. Chlorinated
waste streams can be introduced in one of three ways: liquid, vapor or
adsorbates (to activated carbon). Chlorocarbon solvents are fed into a
vaporizer, mixed with a reducing gas and passed into the Photo-thermal
Reactor. Air laden with chlorocarbon vapors is first passed through a
separator (condenser) which removes chlorinated materials as liquid.
Chlorinated contaminants adsorbed onto activated carbon are removed as vapors
by purging with a mildly heated reducing gas. Then, the vapors are passed
into the Photo-thermal Reactor.
Reducing Gas
Recycle
Chlorocarbons
Waste Stream
Photo-thermal
Reactor
Hydrocarbons
Reducing Gas
Make-Up
Figure 1. ENERGIA's Reductive Photo-Dechlorination (RPD) Process.
140
-------
The Photo-thermal Reactor is the heart of .the RPD technology. Here the
mixture is irradiated and heated. The UV light breaks the C-C1 bond and the ,
temperature sustains long-chain radical Reactions. After a suitable residence
time, conversion and dechlorination are fully completed. Hydrogen chloride is
scrubbed from the mixture which proceeds to the separator. After separation,
excess reducing gas is recycled back to the Input/Mixer. Valuable hydrocarbon
products are collected and sold. There is also an option for recycling a
portion of the hydrocarbon products as an auxiliary fuel to heat up the Photo-
thermal Reactor. :
METHODOLOGY
The RPD treatment was successfully .tested with a representative
chlorocarbon 1,1,1-trichloroethane (TCA). The tests were conducted in a
Photothermal Annular Flow Reactor, schematically illustrated in Figure 2. The
photothermal Annular Flow chamber is the .volume bounded by two concentric
tubes: an outer pyrex tube (1 = 51 cm, 0 = 4.5 cm) wrapped, with a heating
tape, and an inner suprasil quartz tube (1 = 71 cm, $ =2.5 cm) containing a
low-pressure Hg lamp. The heated section of the reactor is insulated with a
ceramic blanket. The reactor can be heated up to 750 °C.
The low-pressure Hg lamp .is specially designed to EHERGIA's
specifications. Inert gas (N2 or Ar) , transparent to UV irradiation, is
continuously flown over the lamp, to displace air surrounding•the lamp so that
DV light is not attenuated by oxygen (present in air). This also reduces the
thermal load due to external heating. To; avoid lamp damage, the lamp
electrodes are located in the cool end-zones of the reactor. And, if
necessary, they can be cooled by auxiliary fans.
Mixtures of Chlorocarbon/H2/Ar flow1through the electrically heated
reactor and are exposed to UV irradiation along the constant temperature
thermal zone. Samples for GC and GC/MS analyses are drawn at the entrance and
exit of the photothermal chamber.
Key Features of the. annular flow reactor are: - i. Operating Conditions:'
T = ambient - 650 ,°C, - P = 1 atmosphere, ii.. Long -residence time, r = 10 - 260
sec, iii.lLarge reactor volume = 451 cc, iv. Uniform exposure to UV
irradiation, v. Optical path length = 1 dm, vi. Accessible wavelengths: 185,
254, 313 nm (Hg), 222 nm (Cd), 215 nm (Zri).
Low Pressure Hg-Lamp
= 12.6mm
Pyrex Tube
To Trap/Vent
G.C. Sampling Port
UV Transparent
Inert N2
Power Cotd
(0.94Amps, IZOvolls)
Venl
Electrode
O Ring Seal
Heating Tape ^- Suprasil Quartz Tube
Reactants
Figure 2. Schematic of the Annular Flow Reactor.
141
-------
Experiments were performed without and with UV light, under otherwise
identical conditions. Experiments with the UV light off are designated
Reductive Thermal-only (RT), and experiments with the light on are designated
Reductive Photo-Dechlorination (RPD).
The conditions selected for both sets of experiments (RT and RPD) are:
Flow Mixtures:
Pressure:
Temperature:
r Range:
Light Source:
1.0% TCA/45.0% H2/Balance Ar
1 atmosphere
430 °C, 520 °C, 590 °C
15 - 260 sec
Low-Pressure Hg Lamp
These conditions were chosen to optimize the process for achieving +99%
(conversion and dechlorination).
RESULTS
Experimental results are presented in terms of Percent Conversion,
Percent Dechlorination and Selectivity. Percent Conversion is the percent
consumption (disappearance) of the parent molecule (i.e. TCA), without regard
to the identification of the products. Percent Dechlorination is the percent
conversion of TCA to chlorine free products. Selectivity is the normalized
molar fraction of each product relative to all products.
Percent Conversion and Percent Dechlorination were measured as a
function of residence time (T) at three temperatures, 430, 520 and 590 °C.
At the lower temperature (430 °C) , percent conversion was very low for
RT (less than 30%) and significantly higher (80%) for RPD. At higher
temperatures (520 °C and 590 °C) , complete conversion (100%) was obtained for
both RT and RPD processes. Consequently, the difference between the two
processes was only in their respective selectivities. As will be pointed out
below, RPD demonstrated almost complete dechlorination, whereas RT did not.
Percent dechlorination obtained during RT treatment at 430 °C and at
15 - 60 sec was quite low -5% and at 260 sec it was still low, about 9.0%.-
However, during RPD treatment, it increased from 9 to' 47% with the same
increase in T. At slightly higher temperature, i.e., at 520 °C, during RT
treatment, it increased very slowly from 30 to 40%, with the increase'in
residence time from 15 to 260 sec. In contrast, during RPD treatment, it
increased rapidly from 41 to 92% with modest increase in residence time (to
130 sec). At 590 °C (see Figure 3), percent dechlorination obtained during
RPD treatment was remarkable. Even at the shortest studied r (15 sec),
reasonable amount of dechlorination (-65%) was obtained. It increased
dramatically to +98% (at T = 100 sec) and at r > 100 sec, +59% dechlorination
was obtained. There was also improvement during RT treatment. Percent
dechlorination increased monotonically from 37 to 77%, with increasing in
residence time up to 260 sec. But it fell short of complete dechlorination
obtained by RPD.
These experimental results have demonstrated greater than 99% conversion
and complete dechlorination at 590 °C and r a 60 sec. Figure 4 depicts
products selectivity. It clearly shows the advantage of RPD (black bars) over
Reductive Thermal-only (RT) (open bars) treatment under otherwise identical
conditions. In both cases conversion is +99%. However, while the RT is
limited to 51% dechlorination, the RPD exhibits +99%. It is apparent that the
RPD process is capable of safe and efficient conversion of all chlorinated
hydrocarbon contaminants to valuable hydrocarbon products (mainly ethane and
methane).
142
-------
c
_o
"<*3
-co
J>
.c
o
100
80 -•
60 -•
40 -
20 -
00
O
-f-
50 100 150 : 200
Total Residence Time (sec)
250
300
Figure 3. Percent Dechlorination as a Function of Residence Time (r)
during RT/RPD Treatment at 590 °C.
100
80
.t±
•I 60
o
JD
0>
(/) 40
20
LLl
O
O
O
Q
O
n
i
CM
O
RT,
Conversion = 99% +
Dechlorination = 51 %
RPD,
Conversion = 99 % +
Dechlorination = 99% +
O
in
x
CM
O
O
CM
Z
CM
O
CM
O
Figure 4. Product Selectivity during RT/RPD Treatment of'TCA
at 590 °C and 60 sec.
In order to compliment the experimental effort, a detailed kinetic
modeling study was also performed. It provided insight into the reaction
mechanism and identified the key elementary steps determining the reaction
pathways for rapid conversion of chlorocarbons to environmentally acceptable
products. Using this kinetic model, a comparison was made between
experimental data and computer modeling results for RT and RPD treatments of
TCA. A good agreement was evident. ;
143
-------
WASTE APPLICABILITY
The RPD process can be effectively applied to liquid or gaseous waste
streams containing saturated and unsaturated chlorocarbons. It was tested for
TCA, TCE, DCE, DCA, vinyl chloride, ethyl chloride, DCM, and chloroform. It
may also be applicable to PCE, carbon tetrachloride, and chlorinated aromatics
to be tested.
The RPD process is specifically cost-effective for the following on-site
remedial operations:
(1) In-situ treatment of chlorinated wastes discharged from Soil Vapor
Extraction (SVE).
(2) Direct treatment of Off-Gas streams containing chlorocarbons.
(3) On-site regeneration of Activated Carbon saturated with chlorocarbons
removed by adsorption from waste streams.
(4) Pretreatment of waste streams entering Catalytic Oxidation systems,
reducing the chlorine content and thereby promoting oxidation and
longevity.
(5) Small-scale, on-site remediations in R&D and testing laboratories,
chemical hoods, clean rooms, etc.
CONCLUSIONS
Bench-top experimental results have successfully demonstrated the
feasibility of the RPD process. They have established that the optimal
operating conditions for achieving +99% dechlorination of TCA are ~ 590 °C
and r a 60 sec. Preliminary cost analysis showed that the RPD is extremely
competitive with other remedial processes. Its estimated cost is less than
$l/lb. of treated chlorocarbons. The RPD process has successfully completed
the bench-top developmental stage arid is the basis on which a pilot-scale
prototype unit is being constructed. A demonstration SITE program will follow
after which the RPD technology will be available for commercialization.
For more information:
EPA Project Manager: Michelle Simon
U.S. EPA Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, OH 45268
Tel: 513-569-7469
Fax: 513-569-7676
144
-------
TREATMENT OF HAZARDOUS LANDFILL
LEACHATE BY THE ROCHEM DIM
MEMBRANE SEPARATION TECHNOLOGY
Douglas W. Grosse
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
(513)569-7844
Kyle Cook
Science Applications International Corporation
10260 Campus Point Drive
San Diego, CA 92126
(619)546-6117
Introduction
The Disc Tube™ Module (DTM) technology was developed by Rochem Separations Systems,
Inc. and designed to remove a variety of organic and inorganic contaminants from liquid hazardous
waste streams. The DTM technology is a membrane based technology utilizing an innovative process
configuration which allows for treatment of aqueous waste streams such as landfill leachate.
Historically, membrane based technologies have been used as a secondary polishing step in treating
effluents to meet pretreatment discharge standards. The DTM technology was designed to treat liquids
containing higher dissolved solids, turbidity and contaminant levels than previously possible with
conventional membrane processes.
The DTM process was evaluated under the Superfund Innovative Technology Evaluation (SITE)
program at the Central Landfill in Johnston, Rhode Island during the months of August and September
1994. Approximately 33,000 gallons of hazardous landfill leachate were treated by the DTM technology
using reverse osmosis membranes. The leachate contained moderate to high levels of volatile organic
contaminants, low levels of heavy metals, and high total dissolved solids. The Developer (Rochem)
claimed that the technology was capable of (1) achieving a high percent rejection of the contaminants
of concern; (2) recovering ;>75% permeate (treated water); and (3) allowing for a greater resistance to
scaling and fouling of membranes than conventional membrane processes.
Technology
The DTM technology is a modular system which utilizes reverse osmosis (RO), ultrafiltration, or
microfiltration membrane materials in a modular configuration which offers large feed flow channels for
high feed flow velocity, generating a crossflow of feed water over stacked membranes. Purportedly,
these system design features allow the DTM greater tolerance for dissolved solids and turbidity; hence,
a greater resistance to scaling and fouling of the membranes. Suspended particulates are easily
flushed away from the membranes and out of the module during operation.
Figure 1 shows a cutaway of a disc tube module. The modular disc tube is comprised of
conventional RO membranes which are formed into octagonal cushions by ultrasonically welding two
membranes together at the edges and separated by a layer of porous spacer material on the inside.
These cushions are supported by a plastic disc which holds the membrane cushion in place and forms
the flow channels for the feed. These membrane cushions are alternately stacked with hydraulic discs
on a tension rod. The tension rod houses a conduit for permeate collection from each membrane
cushion to facilitate discharge from the module into a product recovery tank. O-rings seal the permeate
145
-------
water channels from the feed water channels. A stack of cushions and discs is housed in a pressure
vessel; whereby, flanges seal the ends of the module and provide the feed water input, permeate and
reject outlet connections. The number of discs per module, number of modules and the membrane
materials can be custom designed to suit the application.
Permeate
(Product)
Concentrate
(Reject)
Pressure Vessel
End Flange Tension Rod
The RO membranes used for this demonstration test are more permeable to water than to
organic and/or inorganic impurities. When pressure is applied to a solution (leachate), the aqueous
phase (permeate) passes through a semipermeable membrane which, in turn, rejects the contaminants
carried by the solution, causing them to be retained in a brine solution (concentrate). These impurities
are selectively rejected by the RO membranes. The percentage of water that passes through the
membrane is a function of operating pressure, membrane type and concentration of the contaminants.
Methodology
The DTM technology was evaluated for its ability to meet percent rejections established for the
target contaminants (VOCs >90%, TOG >92%, TDS >99% and metals 99%) while operating at a
treated water recovery rate of 75 percent or greater. In addition, the DTM system was evaluated for
treating .a hazardous leachate without experiencing an unacceptable level of membrane fouling or
scaling, resulting in a loss of membrane flux (flow rate/membrane area).
The Central Landfill site comprises two areas: a 121 acre area and a 33 acre expansion area.
The 121 acre area of the site is where disposal of hazardous and non-hazardous wastes took place in
the past. Several wells on the site were being used to intercept leachate from a "hot spot" area (0.5
acre) of the site to prevent off-site migration of contaminated liquids. Leachate from one of these wells
(MW91ML7) was characterized during a pump test previously conducted. This well was located
downgradient from the hot spot area of the landfill; thereby, containing higher concentrations of
contaminants than other wells on the site. Leachate produced from this well was deemed to be the
146
-------
most suitable for the DTM field demonstration. Table 1 gives the target constituent profile of the
leachate generated from well MW91ML7.
A three-stage DTM system was used for the field demonstration. Two stages were used in
series to treat the landfill leachate to produce a final permeate. Both stages utilized standard thin film
composite (TFC) RO membranes. The permeate from the first-stage DTM was used as a feed for the
second-stage DTM for further removal of contaminants. A third-stage high pressure unit (HPU) was
used to further squeeze the concentrate rejected by the first-stage to increase permeate recovery. The
permeate from the HPU was directed to the second-stage DTM unit. Each unit was protected by media
(sand) and cartridge (10 micron) filters to remove suspended particulates. Acid was also added at the
first-stage and HPU for pH control. The system processed approximately 33,000 gallons of hazardous
leachate at an average feed rate of 5 gpm. The system operated long enough to allow for several
cycles of membrane cleaning (up to 8 hours a day, for 19 days).
Baseline testing was performed prior to and immediately following the Rochem Demonstration
test to determine if there had been a loss of membrane flux due to fouling and scaling. Baseline testing
procedures, as specified in the QAPP, were modified to account for varying conditions (pressure, flow
rates, temperatures and brine concentrations) during the field test in accordance with the Standard
Practice for Standardizing Reverse Osmosis Performance Data ASTM D 4516-85. This method allows
for collected field data to be normalized against a chosen set of standard operating conditions
(average). Using this method, the change in membrane flux was determined after the demonstration
test was completed. A saline solution of known conductivity was used for baseline testing.
Table 1 - Target Contaminant Profile (MW91ML7)
Constituent
VOC
Average Concentration
(ug/D
Chlorobenzene
Ethylbenzene
Toluene
Xylenes
1,2 - Dichlorobenzene
1,4 - Dichlorobenzene
Metals
Calcium
Magnesium
Sodium
Barium
Iron
Manganese
Potassium
Strontium
Other AnaSytes
Total Dissolved Solids
, TOC
pH
OO.UUU
1,600
2,800
2,300
29,000
1,600
(mg/l)
140
200
560
1.4
49
22
110
0.82
(mg/I)
4,100
360
6.8 Units
147
-------
Results
During the field demonstration, both laboratory and field samples were collected for analyses.
Reid measurements included: pressure, temperature, flow, pH, turbidity, and conductivity of the process
liquids. A portable organic vapor detector monitored for fugitive emissions during field operation.
Laboratory samples were collected from: (1) the raw feed; (2) first-stage feed, reject and permeate; (3)
second-stage reject and final permeate; and HPU feed, reject, and permeate. The target contaminants
were analyzed along with TOG, IDS, total solids, MBAs and ammonia (see Table 1).
Based upon results obtained from preliminary data reduction, the Developers claim for percent
rejection of the target contaminants appears to be met; especially, when considering established
standard deviations (SD). Rejection of heavy metals and total dissolved solids was excellent by
significantly surpassing the claim for >99% rejection. Similarly, percent rejection of TOG (>96.3) far
exceeded the claim of >92%. For target VOCs the Developer's claim of >90% rejection was overall
exceeded with 95% rejection of 1, 2-Dichlorobenzene (one of the primary contaminants of concern),
and >95% for toluene, xylenes and ethylbenzene. Although, chlorobenzene only achieved 87% and 1,
4-Dichlorobenzene >87.6% rejection, 95% confidence intervals were calculated to be ±3.7% and
±4.1%, respectively.
The second of the three critical claims established for this demonstration pertained to the
achievement and maintenance of a treated water recovery rate of 75% or greater. Recovery rates were
calculated for each day of operation. There were a total of three (3) days of downtime attributed to
system maintenance, resulting in 18 days of actual field data collection. The system water recovery
rates ranged from, approximately, 65% to as high as 86%. Overall, the average recovery rate of 73.3%
was achieved for the demonstration with a 95% confidence interval of ±2.6%.
The third critical objective of the demonstration test was to determine the extent, if any, of
membrane scaling and fouling as measured by the difference in membrane flux before and after system
field operation. It was considered an indication of membrane scaling and fouling if a loss of greater
than 10% flux was obtained. Baseline procedures were employed for both the first-stage unit and the
HPU, separately. One set of actual operating data (average operating conditions) was selected as the
standard for each unit. At this time, not all of the data have been completely reviewed. In discussions
with the Developer, preliminary results are in question (>10%A) due to the premise that the membranes
were not adequately broken in. Manufacturers typically recommended 50-100 hours of continual
running before the baseline can be determined for a given membrane on sea water systems. For
leachate, Rochem has recently discovered that the same "break-in" period can be 200 hours and more
before a baseline determination can be adequately made.
Conclusions
*
Preliminary results of the SITE demonstration of the DTM system shows that significant percent
removal of target contaminants was achieved, exceeding the Developer's claim in most cases. The
DTM technology can effectively treat hazardous landfill leachate to the extent that the treated water
(permeate) can meet effluent discharge standards (the Rhode Island DEM discharge limits).
The DTM process was also capable of approximating (on the average) the claim of 75% or
greater for water recovery efficiency. During peak days of operation as high as 86% was achieved.
Key findings of this SITE demonstration including analytical results will be discussed in detail
with the submittal of the Technology Capsule, Innovative Technology Evaluation Report and a
videotape.
148
-------
SITE Demonstration of the SVVS Technology
Paul R. de Percin
Risk Reduction Engineering Laboratory
US Environmental Protection Agency
16 West Martin Luther King Dr.
Cincinnati, Ohio 45268
INTRODUCTION
The Subsurface Volatilization and Ventilation System is an integrated technology used for attacking all
phases of volatile organic compound (VOC) contamination in soil and ground water. The SWS technology
promotes in-situ remediation of soil and groundwater contaminated with organic compounds through the
injection of air into the saturated and unsaturated zones, and extraction of vapors from the vadose zone.
Through this process, volatile and semivolatile organic compounds are stripped from the soil and
groundwater. The subsurface circulation of air also increases dissolved oxygen concentrations in the
saturated zone, capillary fringe, and vadose zone, thereby promoting aerobic microbiological processes.
The contaminated air extracted from the wells can be treated at the surface before being discharged to the
environment.
The SVVS process was evaluated under the SITE program at the Electro-Voice, Inc. (EV) facility in
Buchanan, Michigan. The soils were contaminated with aromatic hydrocarbons, and halogenated and non-
halogenated volatile and semivolatile organic compounds (SVOCs) through discharge into a dry well.
Baseline data indicated that approximately 1,000 kg of VOC and SVOC contamination was present in the
dry well area soils, principally in a subsurface sludge layer. The developer claimed that their technology
would reduce the sum of seven target VOCs by 30% over a one year period.
TECHNOLOGY
The SVVS process utilizes soil vapor extraction in conjunction with in-situ bioremediation to clean soil,
sludge, and groundwater. A typical SVVS installation is comprised of a series of air injection and
vacuum/extraction wells designed to circulate air below ground to 1) increase the flow of oxygen in the soil
to enhance the rate of organics destruction by indigenous soil microbes and 2) volatilize and remove
volatile organic contaminants from the soil. This system consisted of three individually plumbed rows of
alternating vacuum extraction and air injection wells referred to as reactor lines. Each reactor line is
plumbed to a single central vapor control unit (VCU) used to house air injection and vacuum pumps and
gauging, as well as emissions control equipment.
The injection wells are installed below the groundwater table and are used to inject air into the
groundwater. The Developer claims that the air strips volatile contaminants from the soil and water as it
percolates through this saturated zone. Extraction wells installed in the vadose zone pull the percolated
air through the soil under vacuum, further stripping contaminants. In addition, the increase in air circulation
in the soil, specifically oxygen, increases the rate of biodegradation by soil microbes, according to the
developer, and transforms contaminants into harmless end products such as carbon dioxide and water.
To aid in the circulation process, sand chimneys can be installed. These are sand-packed borings which
provide passive airflow between the subsurface layers, increasing both the soil vapor extraction and the
biodegradation rates.
The SWS process generates one major wastestream - vapors from the vacuum extraction wells.
Depending upon regulatory requirements, the extracted air may be treated above ground or released
directly to the atmosphere. In the early stages of SWS implementation, the overall rate of mass transfer
of contamination to the vapor phase may exceed biodegradation rates. It is during this period, which lasts
anywhere from two weeks to a few months, that extracted vapors may need to be treated above ground
before release to the atmosphere. However, the magnitude of treatment will decrease steadily over this
149
-------
period until biodegradation rates surpass the net rate of transfer of contaminant mass into the circulating
air. When this point is reached, the vapor extraction off-gas will consist predominantly of carbon dioxide,
which is the major gaseous by-product resulting from the biodegradation process. To reduce these costs
further and promote additional VOC destruction, the SWS design employs the use of proprietary biofilters
for treatment of the extracted vapors.
If required by permits, off-gas extracted from the vacuum extraction wells can be routed through a
configuration of Biological Emissions Control™ (BEG™) units (a patent pending system which, according
to the developer, through biodegradation, achieves up to 80% reductions in concentrations of VOCs in
stack emissions at approximately 20% of traditional emission control costs). The off-gas is then expelled
to the atmosphere through a vent pipe affixed to the extraction pump. Vacuum extraction emissions may
also be favorably controlled within regulatory limits by adjusting the air injection and vacuum extraction
rates. However, if the levels of VOCs in the off-gas are in excess of acceptable levels, the off-gas exiting
the BEC™ units can then be routed through an activated carbon adsorption unit as a final polishing step
prior to discharge.
METHODOLOGY
The SWS process was evaluated for its ability to reduce volatile organic contaminants in the vadose
zone soil of the "dry well" area at the Electro-Voice, Inc. site in Buchanan, Michigan. The primary objective
of the demonstration was to evaluate the developer's claim of a 30% reduction in the sum of seven specific
volatile organic compounds (i.e., benzene, toluene, ethylbenzene, xylene, tetrachloroethene, trichloroethene,
and 1,1-dich!oroethene) in vadose zone soils of the treatment plot over a 12 month period of operation.
A one year time frame was chosen for testing purposes only, and the reduction claim does not reflect the
limits of the technology. Under an actual remedial clean-up, the system may require a longer time than
was possible during the present study.
Reductions in the volatile organics were proposed to occur through the combined effects of in-situ
biodegradation and soil vapor extraction. These reductions were evaluated by comparing the sum of the
concentrations of the select volatile organic contaminants in the matrix prior to system startup and after 12
months of system operation. Secondary objectives were established to determine the relative contributions
of in-situ biodegradation and vapor extraction to the removal and degradation of volatile organics from the
subsurface.
Soil samples were collected from borings within the physical boundaries of the SWS system and
sampled in a manner such that the entire vertical section of the vadose zone was represented. Five distinct
subsurface zones were identified based on lithology and contaminant occurrence. These included the upper
horizon (above the contaminant source), sludge layer (predominant source of contamination), and lower
horizons A1, A2, and B (below the contaminant source).
Since the developer's claims were to reduce seven volatile organic contaminants by 30%, benzene,
toluene, ethylbenzene, and xylenes (BTEX), tetrachloroethene (PCE), trichloroethene (TCE), and 1,1-
dichloroethene (1,1 -DCE) were considered the critical analytes for this demonstration. Analyses were also
performed on select samples for the following non-critical parameters: total carbon (TC), total inorganic
carbon (TIC), nutrients (nitrate, phosphate), total metals plus mercury, cyanide, pH, and particle size
distribution (PSD). An additional objective of this demonstration was to develop data on operating costs
for the SWS technology.
The extracted vapor streams were analyzed by continuous emission monitoring (CEM) for O2, CO2, and
total hydrocarbons (THC). Grab samples of the extracted vapor stream were collected for determining the
concentration and distribution of individual volatile organic compounds.
Shut-down tests were periodically performed to assess the presence and magnitude of biological
processes in the destruction of organic constituents in the subsurface. During a shut-down test, the
150
-------
injected air stream is temporarily turned off resulting in the cessation of oxygen delivery to the subsurface.
If there is a robust aerobic microbial population in the subsurface, the available oxygen will be quickly
depleted. The shut-down test tracks the magnitude arid rate of oxygen drop-off over a twenty-four hour
period.
RESULTS
At the Electro-Voice site, the SWS process achieved an overall 80.6% reduction of the sum of the
seven critical VOCs over a one year period from vadose zone soils. This level of reduction greatly
exceeded the developer's claim of a 30% reduction over a one year time frame. The average concentration
of the sum of the seven analytes from the hot zone in the study area, prior to installation of the SWS, was
341.5 mg/kg. The average concentration of the sum of the seven analytes after one year of operation was
66.20 mg/kg.
The data reveals that the most contaminated zone, the sludge layer, had an average reduction of
81.5%. The other less contaminated horizons exhibited reductions ranging from 97.8% to 99.8%. The
reductions over the area! extent of the site, as determined from the individual boreholes, ranged from 71%
to over 99%. This indicates the system operated relatively uniformly over the entire vadose zone of the
treatment plot, and no significant untreated areas were encountered, regardless of VOC concentration or
lithology.
The shut-down testing indicates that microbiological activity was stimulated at the site. Due to the
inherently high organic content of the soil, it was not clear how much of this stimulation was due to
contamination. The microbiological activity, as determined from the first shut-down test, was greatest in
portions of the site where the VOCs were greatest, and least active in areas of the site where the
contamination was small or absent. Seasonal variations as evidenced in the background wells, where
presumably no contamination existed, introduced uncertainty in data interpretation. A comparison of three
shut-down tests indicates that biological activity was greatest during the beginning of remediation and
progressively decreased throughout the remainder of the demonstration, but at a rate that was less than
the VOC mass removal rates attributed to vapor extraction alone. This would indicate that biological
processes play an increasingly important but not a dominant role, relative to vapor extraction, as the
remediation proceeds.
An analysis of the volatiles from the vapor extraction outlet reveals that the highest mass of volatiles
was removed during the early phase of the project. Furthermore, the mass of volatiles in the off-gas
stream gradually decreases to a low and constant level after approximately 230 days of operation.
The SWS experienced no major operational problems over the twelve month study period. Once
implemented, the system was easy to monitor and required minimal maintenance and/or operator attention.
The Biological Emissions Control™ (BEG™) unit, installed to biologically degrade VOCs from the off-gas
stream, was removed from the system after a few months of operation and was not evaluated. Dispersive
air modelling results showed that contaminant concentrations were below established air quality standards
and discharge criteria for the site were met without any additional treatment.
The SWS was installed at the site based on contaminant distribution information derived from remedial
investigation data. During the baseline sampling event under the SITE Demonstration, it became evident
that a portion of the system was installed within a clean area of the site. Operation of the system was
easily adjusted while maintaining the existing hardware to concentrate remedial action in more
contaminated areas. However, installation of the system in the non-contaminated area impacted costs
since materials and labor were expended. The excess installation did not in any way impact the
performance of the system. This situation stresses the importance of accurately defining the extent and
magnitude of contamination prior to the implementation of in-situ technologies. In-situ technologies may
require site characterization in greater detail than is commonly available from remedial investigations.
151
-------
CONCLUSIONS
The results from the demonstration indicate the SVVS technology greatly exceeded their reduction
claims by providing a site average 80.6% reduction of volatile organics in the vadose zone. Furthermore,
aerial and vertical reductions across the site did not indicate the presence of any zones that were not
treated by the system. The SWS process proved to be reliable and required minimal operator oversight.
The technology did not experience significant operational difficulties during the evaluation period.
The SWS process is applicable to sites contaminated with gasoline, diesel fuels, and other
hydrocarbons, Including halogenated compounds. The system is very effective on benzene, toluene,
ethylbenzene, and xylene (BTEX) contamination. The process can also be used to contain contaminant
plumes through its unique vacuum and air injection techniques. The technology should be effective in
treating soils contaminated with virtually any material that has some volatility or biodegradabilHy. The
technology can be applied to contaminated soil, sludge, free-phase hydrocarbon product, and groundwater.
By changing the injected gases, anaerobic conditions can be developed, and a microbial population can
be used to remove nitrate from groundwater. The aerobic SWS can also be used to treat heavy metals
in groundwater by raising the redox potential of the groundwater and precipitating the heavy metals.
The cost to remediate 21,300 yd3 of vadose zone soils during a full-scale cleanup over a 3-year period
at the Electro-Voice Superfund site in Buchanan, Ml was estimated to be $192,237 or $9/yd3, not including
effluent treatment and disposal. The majority of this was incurred in the first year, primarily due to well
drilling and associated site preparation. If effluent treatment and disposal, using vapor phase granular
activated carbon, had been included, this would have added $164,500 to the first year of remediation and
brought the total cleanup figure to $356,737 ($16.75/yd3). This would have accounted for over 45% of the
total cleanup costs.
152
-------
USE OF SULFATE REDUCING BACTERIA IN ACID MINE DRAINAGE TREATMENT
Thomas J. Powers
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
(513) 569-7550
INTRODUCTION
The environmental impacts caused by Acid Mine Drainage (AMD) were first recorded in .
1556 by Georgius Agricola. In the United States 10,000 miles of streams and 29,000 surface
acres of impoundments are estimated to be seriously affected by AMD. Abandoned surface
mines are estimated to contribute about 15% of the drainage, while active mines (40*) and
shaft and drift mines (45*) contribute the remainder.
AMD results when metal sulfide minerals, particularly pyrite (FeS2), come in contact
with oxygen and water. Acid generation occurs when metal sulfide minerals are oxidized
according to the Initiator Reaction:
FeS2(pyrite)
H20 .---
Fe2+ + 2S042' + 2H+
Figure 1 illustrates the pyrite oxidation cycle. This reaction is one of many that results
in increased metal mobility and increased acidity (lowered pH) of the mine water. Figure 2
puts forth the oxidation of Iron Sul fides. The oxidation of ferrous sulfate is accelerated
by bacterial action of Thiobacillus ferrooxidans, a naturally occurring bacterium that at pH
3.5 or less, can rapidly accelerate the conversion of dissolved Fe (ferrous iron) to FeJ+
(ferric iron), and can act as an oxidant for the oxidation of pyrite. Ferric ions, as well
as other metal ions, and the sulfuric acid have. a deleterious influence on the biota of
streams receiving AMD.
The Lilly/Orphan Boy Mine, located in the Elliston Mining District of Powell County,
Montana, was selected as the Sulfate Reducing Bacteria (SRB) technology demonstration site.
The mine is situated on a patented claim on Deerlodge National Forest Land about 11 miles
south of Elliston, Montana. This abandoned mining operation consists of a 250 -foot shaft,
four horizontal workings, and some stoping. The shaft is flooded with AMD to the 74-foot
level and is discharging about 3 gallons per minute (gpm) at a pH of 3.0 from the adit
associated with this level. The total aluminum, arsenic cadmium, copper, manganese, iron,
and zinc concentrations of the Lilly/Orphan Boy Mine Water are shown in Table 1. The pH
range observed in the mine water is also presented.
METHODOLOGY
The main objective of conducting laboratory- scale tests was to determine the
effectiveness of bacterial sulfate reduction in the reduction of metal concentrations at the
ambient shaft water pH of 3.0 and temperature of 8°C. The temperature of the shaft water
ranged from 6°C, at the portal, to 11°C, in the shaft, over a one-year period (5/93 to 7/94)
during which time samples were collected from the site. This objective was facilitated by
monitoring the laboratory- scale system over a period of 60 days through the collection and
analysis of multiple samples, principally metals concentration in the influent and effluent
waters. The sulfur cycle was observed throughout Phase 1 and Phase 2 of the pilot plant SRB
experiment (Figure 3).
The experimental design of the laboratory- scale testing allowed for the comparison of
two different bacterial preparation methods and three different substrate layering methods.
The testing was performed at a typical shaft water temperature (8°C) and with influent water
obtained from the shaft of Lilly Orphan Boy Mine one week prior to start up.
153
-------
FeS2(pyrite)
_ + H2O —* Fe2+ + 2SO42' + 2H+
Initiator reaction
2+
Slow spontaneous,
bacteria catalyze f o<
Fast spontaneous
(bacteria may also
catalyze)
FeS2(Pyrite)
*Fe3+
Figure 1. The Pyrite Oxidation Cycle
FERROUS SULFATE BY NON-BIOTIC OXIDATION :
2 FeS2 + 2 H2O + 7 O2-4-2 FeSO4 + 2 1^804
BACTERIAL OXIDATION OF FERROUS SULFATE BY
Thiobacillus ferrooxidans:
4 FeSO4 + O2 + 2H2SO4—¥ 2 Fe2(SO4)3 + 2 H2O
FERRIC SULFATE IS REDUCED AND PYRITE OXIDIZED BY
CHEMICAL REACTION :
Fe2(SO4)3 + FeS2 —* 3 FeSO4 + 2 S
2S + 6Fe2(SO4)3 + 8 H2O^^12 FeSO4 + 8 H2SO4
ELEMENTAL SULFUR OXIDATION BY Thiobacillus thiooxidans:
2S + 3O +
Figure 2. The Oxidation of Iron Sulfides
1154
-------
Pilot scale testing was conducted in eight packed-bed reactors (4 foot, vertical,.
plexiglass columns) operated in up-flow configurations. This design was selected to closely
model the actual conditions of the hydrogeologic;flow of water entering the shaft via
groundwater up-flowing to the adit level, and leaving the shaft through the mine adit.
Eight tests were conducted in parallel. Four reactors had diameters of 14 inches while the
remaining four reactors were 18 inches in diameter. The reactors were housed in a wooden,
insulated, air-conditioned enclosure during the entire testing period to help maintain a
temperature representative of the shaft water, nominally 8°C.
Mud from the Lilly/Orphan Boy Mine was used as a source of Sulfate Reducing Bacteria
(SRB) for the pilot-scale testing. Approximately three cubic feet of mud was obtained from
the portal area of the mine. In order to maintain the SRB in an anaerobic environment, the
mud was placed in plastic bags within five gallon buckets until it was mixed with cow manure
and decomposed woodchips to make the organic substrate that was packed into the reactors.
RESULTS
Seven of the eight pilot SRB reactors were very effective in reducing all metal
concentrations from the Lilly/Orphan Boy Mine shaft water. The one reactor (using a fishnet
containment system) allowed the untreated water to by-pass the substrate reactor. Water
sampled in the shaft of the Lilly/Orphan Boy Mine was characterized and typical results are
shown in Table 1. The Project Goals are also included in Table 1.
Table 2 shows no data for arsenic. This is' because Sulfate Reducing Bacteria (SRB)
were collected from mud outside the adit of the Lilly/Orphan Boy Mine. This mud contained
high levels of arsenic that, when mixed with the manure, solubilized. Therefore, the data
collected from the effluent were considered to be invalid measurements of the capabilities
of the process in arsenic removal. ;
TABLE 1. LILLY/ORPHAN BOY.SHAFT WATER CHEMISTRY (M6/L)
Fe
29.0
1.0
Mn
6.28
2.0
LABORATORY -SCALE
Cu
0.24
0.1
TESTING
Zn
25.6
GOALS:
4.0
Al
8.09
LILLY ORPHAN
1.0
Cd
0.24
BOY
0.1
As
1.02
PILOT MINE
0.03
Soluble SO/2 pH
Sulfide
0.75 277 2.8-3.0
PROJECT
--- -'--•'. 6-8
The main objective of conducting laboratory-scale tests was to determine the
effectiveness of bacterial sulfate reduction in:the mitigation of metal concentrations at
the ambient shaft water pH (Table 1) and temperature (Nominally 8°C). The temperature of
the shaft water ranged on average from 6°C, at the portal, to 11°C, in the shaft, over a
one-year period (5/93 to 7/94) during which samples were collected from the site. This
objective was facilitated by monitoring the laboratory-scale system over a period of 60 days
through the collection and analysis of multiple samples, principally metals concentration in
the influent and-effluent waters. '
The experimental design of the laboratory-scale testing allowed for the comparison of
two different bacterial preparation methods and three different substrate layering methods.
The testing was performed at a typical shaft water temperature (8°C) and with influent water
having a representative pH (3.0) of the ambient shaft water.
155
-------
TABLE 2. SRB TECHNOLOGY RESULTS
Measurement
Iron (Fe)1
Manganese (Mn)2
Copper (Cu)3
Zinc (Zn)3
Aluminum (AT)1
Cadmium (Cd)3
Soluble Sulfide
Sulfate (S04'z)
PH
Effluent (mg/L)
0.17
1.84
0.0069U
0.172
0.03U
0.0047U
3.12
257
7.6
Change
99# reduction
71% reduction
97% reduction
99% reduction
99% reduction
98% reduction
416% increase
8% reduction
Approximately 4.6 pH
unit increase
'Reduction probably due to hydroxide precipitation resulting from pH increase.
Deduction probably due to either carbonate precipitation or absorption.
Deduction due to sulfide precipitation.
U - non-detect (the number before the "U" is the instrument detection limit).
FIELD TESTING
The field testing of the 50 ton in situ bioreactor is in progress and will continue
until at least December 1995. The final reporting is scheduled for April 1996. The field
demonstration of sulfate-reducing-bacteria technology required the placement of an organic
Substrate containing the SRB at positions in the two vertical shafts and horizontal tunnel
(ADIT) of the mine where the natural flow of shaft water must pass through the substrate
(Figure 4). The biological action of the SRB and subsequent chemical reactions 1) reduce
the amount of sulfate in the water, 2) form soluble hydrogen sulfide, 3) form metal sulfides
from a reaction of the metallic ions with the hydrogen sulfide, 4) precipitate the metal
sumdes within the substrate and 5) increase the pH of the effluent discharge to 7.6+.
CONCLUSIONS
Sulfate reduction was accomplished by sulfate reducing anaerobic bacteria. Sulfate-
reducing bacteria decompose simple organic compounds (substrate) using sulfate as a terminal
electron acceptor, thus producing soluble hydrogen sulfide and bicarbonate as metabolic
products. The hydrogen sulfide reacts with dissolved heavy metals in the water to form
insoluble metal sulfides. The bicarbonate has the effect of increasing pH. These two
processes have the potential to greatly reduce or eliminate the problems associated with
some rural remote abandoned acid generating mine sites.
156
-------
INORGANIC LOOP
ORGANIC LOOP
SULFATASES
ESTERIFICATION
DISSIMILATORY.
SULFATE I |
R OS03H
SULFATE
ESTERS
SULFATE
ASSIMILATION
s°
ELEMENTAL
SULFUR
_
w
H2S
HYDROGEN
SULFIDE
• ^^
SULFIDE ASSIMILATION
1 DESULFURIZATION
I
» * .
RSH
THIOLS
Figure 3. The Sulfur Cycle
HEADFRAME
8 CABLES SECURED AT HEADFRAME
INJECTION WELLS
EXPECTED WATER
LEVEL AT 74ft,
LAND
APPLICATION
SRB SUBSTRATE MIX
(MINE
! SHAFT
•PLATFORM
UPDWELLING MINE WATER
Figure 4. Lilly Orphan Boy Mine Profile
157
-------
The Phase 1 (laboratory pilot plant) was successful in demonstrating the Sulfate-
Reducing Bacteria (SRB) process. Metals were reduced by 90+# and the effluent discharge
levels in milligrams per liter were within the project's goals in seven of the eight column
reactors used in the laboratory experiment. The process worked at 8°C with a retention time
of approximately 4-1/2 days.
The Phase 2 (Field demonstration of the SRB technology) required the placement of an
organic substrate containing the Sulfate Reducing Bacteria (SRB) at positions in the
vertical shafts and the horizontal adit of the Lilly Orphan Boy Mine where the natural flow
of water was forced through the substrate. The effluent discharge criteria appears to have
been met using the current data available.
REFERENCES
1. MSE, Inc. Laboratory Report for the Sulfate-Reducing Bacteria Laboratory
Demonstration. Mine Waste Technology Pilot Program, Activity III, Project 3. MSE,
Inc.. Butte, Montana, October 1994.
2. Odom. J.M. and Singleton, R. The Sulfate-Reducing Bacteria: Contemporary
Perspectives. ISBN 0-387-97865-8; ISBN 3-540-97865-8; QR92, S8585, Pages 211-
249). 1992.
For more information, contact Thomas J. Powers, U.S. Environmental Protection Agency,
Risk Reduction Engineering Laboratory, 26 West Martin Luther King Drive, Cincinnati, Ohio
45268, (513) 569-7550.
158
-------
BUFFING. BURNISHING. AND STRIPPING OF VINYL ASBESTOS FLOOR TILE
Bruce A. Hollett, Alva Edwards, and Patrick J. Clark
U.S. Environmental protection Agency
Risk Reduction Engineering Laboratory
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
(513) 569-7654
John R. Kominsky, Ronald W. Freyberg, and James M. Boiano
Environmental Quality Management, Inc.
1310 Kemper Meadow Drive
Cincinnati, Ohio 45240
INTRODUCTION
!
Studies were conducted to evaluate airborne asbestos concentrations during the three
principal types of preventative maintenance (low-speed spray-buffing, ultra high-speed
burnishing, and wet-stripping) used on asbestos-'containing floor tiles. These were done
under pre-existing and prepared levels of floor care maintenance. Airborne asbestos
concentrations were measured before and during each floor care procedure to determine the
magnitude of the increase in airborne asbestos levels during each procedure. Airborne total
fiber concentrations were also measured for comparison with the Occupational Safety and
Health Administration's (OSHA) Permissible Exposure Limit (PEL) of 0.1 f/cm3. Low-speed
spray-buffing and wet-stripping were evaluated on pre-existing floor conditions and three
levels of prepared floor care conditions (poor, medium, and good). Ultra high-speed
burnishing and wet-stripping were evaluated on two levels of prepared floor care conditions
(poor and good). Floor care conditions were defined in consultation with the Chemical
Specialty Manufacturers Association and other representatives of floor-care chemical
manufacturers. Controlled studies were conducted in an unoccupied building at the
decommissioned Chanute Air Force Base in Rantoul, Illinois, with the cooperation of the U.S.
Air Force. The building offered approximately 8600 ft2 of open floor space tiled with 9-inch
by 9-inch resilient floor tile containing approximately 5% chrysotile asbestos.
METHODOLOGY
Configuration for Experiments '. '
Approximately 6500 ft2 of floor space was isolated as the experimental test area. A
containment shell was constructed to provide five equally-dimensioned test rooms, each with
approximately 1300 ft2 of floor space and 7-foot ceiling height. The ceiling and walls were
covered with polyethylene. Four high-efficiency particulate air (HEPA) filtration units were
placed in the hallway outside of the five test rooms to ventilate the test rooms and reduce
the airborne asbestos concentrations to background levels after each experiment.
Upon completion of the low-speed spray-buffing and wet-stripping experiments, the test
area was reconfigured to accommodate the ultra high-speed burnishing and wet-stripping
6500 ft2
experiments. The test area was reconfigured to 'provide a single test room of approximately
*• of floor space and 7-foot ceiling height.
Experimental Design
Low-speed spray-buffing was first evaluated on the pre-existing floor-care condition.
Low-speed spray-buffing of the pre-existing floor-care condition was evaluated five times,
once in each of the five test rooms. Wet-stripping (including polish and sealant removal)
was also evaluated on the pre-existing floor-care condition. Wet-stripping of the pre-
existing floor-care condition was evaluated five times, once in each of the five test rooms.
159
-------
Low-speed spray-buffing was evaluated on three levels of prepared floor-care conditions:
(1) poor floor-care condition defined as a floor with one coat of sealant and one coat of
polish; (2) medium floor-care condition defined as a floor with one coat of sealant and two
coats of polish; and (3) good floor-care condition defined as a floor with two coats of
sealant and three coats of polish. Each floor-care condition was evaluated five times, once
in each of the five test rooms, to yield a total of 15 experiments. Wet-stripping after low-
speed spray-buffing was evaluated on two levels of floor-care conditions (medium and good).
This comparison addresses the effectiveness of two coats of sealant versus one coat of
sealant to limit the extent of airborne asbestos concentrations during polish removal. Wet-
stripping of each of the two floor care conditions was evaluated five times, once in each of
the five test rooms, to yield a total of ten experiments.
Ultra high-speed burnishing was evaluated on two levels of prepared floor-care
conditions: (1) poor floor-care condition (defined as a floor with two coats of sealant and
one coat of polish) and (2) good floor-care condition (defined as a floor with two coats of
sealant and four coats of polish). Each floor-care condition was evaluated four times to
yield a total of eight experiments.
Wet-stripping after ultra high-speed burnishing was also evaluated on two levels of
floor-care conditions (poor and good). Each of the two floor care conditions were evaluated
four times to yield a total of eight experiments.
Analytical Methods
The mixed cellulose ester filters were prepared and analyzed in accordance with the non-
mandatory transmission electron microscopy (TEM) method specified in the Asbestos Hazard and
Emergency Response Act (AHERA) Final Rule (October 30, 1987; 52 CFR 4826) to achieve a
sensitivity of 0.005 s/cc, plus the specific length and width of each structure were measured
and recorded. The phase contrast microscopy (PCM) samples were prepared and analyzed
according to the National Institute of Occupational Safety and Health (NIOSH) Method 7400,
with an analytical sensitivity of about 0.01 f/cc. Specific quality assurance procedures
outlined in the AHERA rule were used to ensure the precision of the collection and analysis
of air samples, including filter lot blanks, open and closed field blanks, and repeated
sample analyses.
Statistical Methods
The relative change in airborne asbestos concentration was measured by the ratio of the
average concentration during the specific maintenance procedure to the average concentration
before the maintenance procedure. These ratios were then compared by taking the natural
logarithm and comparing the averages by standard analysis of variance (ANOVA) techniques.
RESULTS
Low-speed spray-buffing and wet-stripping was first evaluated on the pre-existing floor-
care condition. Larger (and statistically significant) increases in the TEM airborne
asbestos concentrations were observed during wet-stripping than during spray-buffing. None
of the individual PCM concentrations exceeded the OSHA PEL of 0.1 f/cm3- Consequently,
8-hour time-weighted average (TWA) concentrations based on these measured levels would not
exceed the OSHA PEL. The highest individual PCM concentration (0.023 f/cm3) was measured
during wet-stripping.
For the prepared floor studies, the mean relative increase in TEM airborne asbestos
concentrations during low-speed spray-buffing tended to decrease as the floor care condition
improved (i.e., poor condition resulted in a larger relative increase than medium, and medium
condition showed a larger relative increase than good); however, the differences between the
three levels of floor care were not statistically significant (p = 0.1149).
Larger (and statistically significant) increases in TEM airborne asbestos concentrations
were observed during wet-stripping of floors in medium condition than on floors in good
condition. The relative increase in airborne asbestos concentrations (i.e., compared to
baseline measurements) was approximately 14 times greater, on average, during wet-stripping
160
-------
of floors in medium condition than during wet-stripping of floors in good condition. The
stripping solution used on these floors was designed to remove only the polish from the
floor, leaving the layer(s) of sealant on the floor. Therefore, although significant
increases in airborne asbestos concentrations were observed during wet-stripping of floors in
both medium and good condition, the extra layer of sealant on floors in good condition
appears to significantly decrease the airborne asbestos levels that were generated by the
activity. Overall, significantly larger increases (p = 0.0001) in airborne asbestos
concentrations were observed during wet-stripping than during low-speed spray-buffing (this
comparison was restricted to floors in medium and good condition since wet-stripping was not
evaluated on floors in poor condition). The relative increase in airborne asbestos
concentrations was approximately 18 times greater, on average, during wet-stripping than
during low-speed spray-buffing.
PCM Concentrations ;
None of the individual PCM concentrations exceeded the OSHA PEL of 0.1 f/cm3.
Consequently, 8-hour TWA concentrations based on these measured levels would not exceed the
OSHA PEL. The highest individual PCM concentration (0.032 f/cm3) was measured during low-
speed spray-buffing.
Ultra High-Speed Burnishing and Uet-Stripping Experiments
Similar increases in airborne TEM asbestos concentrations were seen during ultra high-
speed burnishing and wet-stripping of floors in both poor and good condition. No floor
condition or maintenance procedure resulted in significantly higher or lower increases in
mean airborne asbestos concentration.
Overall, ultra high-speed burnishing and wet-stripping resulted in an 11-fold
statistically significant increase, on average, in airborne asbestos concentration.
The ultra high-speed burnishing operation produced a fine, pale yellow, powdery dust
from the wax and/or sealant. PCM concentrations measured during ultra high-speed burnishing
were significantly higher than those measured during stripping. The elevated concentrations
measured during ultra high-speed burnishing were'due primarily to the white dust generated
during the process. The fine dust particles (pulverized wax/sealant) that measured greater
than 5 /JIB in length and had a length-to-width aspect ratio of 3:1 were counted as fibers
(NIOSH Method 7400, A Counting Rules). The corresponding TEM concentrations show that the
PCM concentrations do not reflect an accurate indication of the airborne asbestos
concentrations.
The 8-hour TWA concentrations were calculated by assuming zero exposure beyond that
which was measured during the experiment. None of the 8-hour TWA concentrations measured
during wet-stripping (after ultra high-speed burnishing) exceeded the OSHA PEL of 0.1 f/cm3
for total fibers. '
CONCLUSIONS
This study shows that low-speed spray-buffing, ultra high-speed burnishing, and wet-
stripping of asbestos-containing resilient floor tile can be sources of airborne asbestos in
building air. Greater releases of airborne asbestos were observed during wet-stripping than
during low-speed spray-buffing when performed on floors in both pre-existing condition and
prepared conditions. The results of this study further suggest that multiple layers of
sealant applied to the floor prior to the application of-the floor finish can reduce the
release of asbestos fibers during polish removal,
For more information, the EPA Work Assignment Manager, Alva Edwards, can be contacted at
the U.S. Environmental Protection Agency, Risk Reduction Engineering Laboratory, 26 West
Martin Luther King Drive, Cincinnati, Ohio 45268, (513) 569-7693.
161
-------
CONTAMINANTS AND REMEDIAL OPTIONS AT PESTICIDES SITES-
A TECHNICAL RESOURCE DOCUMENT
Richard N. Koustas
USEPA Edison Facility
2890 Woodbridge Avenue
Edison, NJ 08837
Phone: 908-906-6898
INTRODUCTION:
Pesticide contamination includes a wide variety of compounds resulting from manufacturing,
improper storage, handling, disposal, and/or agricultural processes. Remediation of pesticide-
contaminated soils can be a complicated process, as most pesticides are mixtures of different
compounds rather than pure pesticide. The remedial manager is faced with the task of selecting
remedial options that will meet established cleanup levels. There are three principal options for
dealing with pesticide contamination: containment/immobilization, destruction, and
separation/concentration. This paper is condensed from the technical resource document (TRD)
"Contaminants and Remedial Options at Pesticide Sites" and provides a brief summary on treatment
technologies that are available or those being developed for pesticide contamination. Technologies
that have not produced performance data are not included nor are water treatment technologies.
This paper focuses on potential remediation techniques of soils.
CHARACTERISTICS OF PESTICIDES
Pesticides, as defined by the U.S. Federal Environmental Pesticide Control Act, are "...any
substance or mixture of substances intended for preventing, destroying, repelling, or mitigating any
insect, rodent, nematode, fungus, weed or any other form of terrestrial or aquatic plant, animal life,
or virus, bacteria, or other microorganism which the Administrator declares a pest". Pesticides
include insecticides, fungicides, herbicides, acaricides, nematocides, and rodenticides as well as any
substance or mixture of substances intended for use a plant regulator, defoliant, or desiccant.
Pesticides do not include such substances as fertilizers or veterinary medicines. Pesticide wastes
are generally complex chemical mixtures and not pure pesticides. These mixtures can include
solvents, carriers and other components that will have a direct effect on toxicity, mobility, transport,
and treatment.
Several classification criteria are utilized when grouping pesticides. Conventional
classification methods are based on the applicability of a substance or product to the type of pest
control desired. In addition, the EPA has its own classifications under the Resource Conservation
and Recovery Act (RCRA) and Superfund. For the purpose of treatment, pesticides may be
classified based on three characteristics: water solubility, contains metals or contains halogen.
Therefore, TRD categorizes pesticides into four waste groups based on data needs for available
treatment technologies:
WG01 - Inorganic pesticides
WG02 - Halogenated water insoluble organics
WG03 - Halogenated sparingly water soluble organics and organo-linked compounds
WG04 - Non-halogenated organics and organo-linked compounds.
The TRD provides details of the four pesticide waste groups and gives examples of commonly found
pesticides. These groups are subdivided further to show the chemical class or family each pesticide
belongs to according to their molecular structure or key functional group. Applicable treatment
technologies for each waste group are also provided. References to pesticides and pesticide wastes
in this document use the above waste group categories.
162
-------
Most pesticides are adsorbed easily on soils because of their molecular weight. In fact,
adsorption of pesticides on the soil surface is a dominant factor that affects the extent of the site
contamination. As a rule, when applied properly, pesticides migrate slowly. Concentrated pesticide
from a spill or leak can move more quickly into the subsurface, especially if the pesticide is an
aqueous phase or under the influence of percolating water. Mechanisms of pesticide fate and
transport that affect the extent of site contamination include: adsorption, biodegradation,
volatilization, downward and lateral migration, and photolysis.
Selecting a remedial strategy includes considering the individual contaminant's toxicity,
persistence, migration pathways, and rate of transport from a site. The wide range of physical and
chemical properties of pesticides also influences the selection of an appropriate remedial technology
or combination of technologies (known as a treatment train). It is important to gain information
specific to the pesticide(s) present in order to effectively identify the treatment technology(ies) that
is most applicable and cost effective.
The following sections present an overview of treatment technologies that have been used
to treat, destroy, or remove pesticides from soil contaminated with pesticides. The technologies
discussed are at various stages of development and application and are separated into
immobilization, destruction and separation/concentration technologies.
IMMOBILIZATION OPTIONS
Immobilization technologies minimize or prevent the migration of the contaminant by utilizing
physical barriers to inhibit the flow of groundwater through contaminated soil. The use of chemical
reactions and/or physical interactions may also be employed to retain or stabilize the contaminant to
prevent its migration into the surrounding environment. Immobilization techniques limit the mobility
of the pesticides arid provide no detoxification or volume reduction. These technologies include
containment, stabilization/solidification (S/S) and in-situ vitrification (even though in-situ vitrification
is considered a thermal process, its primary goal is to immobilize the contaminant).
Containment of contaminant plumes via capping or vertical barriers is often a component of
an overall remediation plan. These technologies are used with groundwater extraction and
treatment. In S/S technologies, a chemical reagent is mixed with the contaminant to physically bind
or chemically react to reduce its mobility or restrict contact with a mobile phase. Cement based
processes are more suitable for inorganic pesticides (WG01) and are often effective in reducing the
mobility or leachability of metals.
In-situ vitrification (ISV) is used to melt contaminated soils and convert them into a glass-like
material of very high stability and chemical inertness. Typically, an electric current is used to
generate temperatures of up to approximately 2,000°C in order to fuse the soil. Gases containing
vaporized water and organic substances are collected at the surface for treatment. ISV is
commercially available and can be used to destroy or capture organic pesticides in waste groups
WG02, WG03, and WG04 as well as to immobilize; inorganic pesticides,(WG01) within the melt.
DESTRUCTION TECHNOLOGIES
Destruction technologies are used to reduce or eliminate toxicity and often decrease volume
and/or mobility. Contaminated soils that are treated using destruction technologies regularly require
materials handling steps such as excavation, dewatering, conveying, and screening prior to
treatment. Additionally, large volumes of soil may require separation or concentration of the
contaminants in order to reduce the total volume to be treated. Destruction technologies include
thermal, chemical, and biological treatments.
163
-------
Thermal destruction technologies include incineration and ultra-high temperature processes
(e.g., plasma-arc). Incineration is commercially available that utilizes thermal energy (heat) in the
presence of oxygen to combust the pesticides and other organic contaminants in soils. Incineration
can be used to remediate soils contaminated with organic pesticides (WG02, WG03, and WG04).
Incineration of inorganic pesticides from WG01 may result in volatilization of metals into off gases or
concentration of metals in the ash.
Chemical destruction technologies that are potentially applicable for remediating pesticide-
contaminated soils include chemical oxidation, dehalogenation, hydro processing/heteroatom
removal, and hydrolysis/neutralization. These technologies use chemical reactions to reduce or
eliminate the contamination in soil containing pesticides.
Chemical oxidation is the most well known of the chemical destruction technologies.
Although its use in environmental remediaton has been very limited, it as been used successfully by
the chemical processing industry. This treatment may be used to treat organic pesticides (WG02,
WG03, and WG04). The possible formation of partially oxidized compounds due to incomplete
oxidation should be considered. Variants of oxidation technology may employ air, oxygen or other
oxidizing agents.
Dehalogenation removes the halogen atoms from halogenated organic compounds, resulting
in significantly reduced toxicity. This technology is applicable to halogenated organic pesticides in
waste groups WG02 and WG03. This technology has been demonstrated at the bench and pilot-
scale levels as effective on pesticide-contaminated soils.
Hydrolysis/neutralization is a chemical treatment process that results in the destruction of
the contaminant under acidic or alkaline conditions. The products of this reaction generally consist
of smaller and less toxic molecules: however, an incomplete chemical reaction may produce toxic
by-products. This technology can be used to detoxify waste groups WG03 and WG04 and has been
demonstrated at the bench and pilot-scale levels.
Bioremediation is typically applicable fpr treatment of media contaminated with low
concentrations of organic pesticides (WG01, WG02, WG03). Oxygen and nutrients are often
supplied to the soil to allow microorganisms to metabolize the organic compounds present. The
presence of inorganic pesticides (WG01) may inhibit the microbiological degradation process. This
technology is available at full-scale.
SEPARATION/CONCENTRATION OPTIONS
Separation/concentration options utilize physical or chemical processes to concentrate
contaminants by removing them from the media in which they are contained. These technologies do
not reduce the toxicity of the contaminants or inhibit mobility. Rather, they reduce the volume of
material requiring treatment by collecting the contaminents in a concentrated form or transferring
them to another medium that is easier to handle or treat. These technologies include radio
frequency (RF) heating, soil washing, thermal desorption, and solvent extraction.
Radio frequency heating utilizes electromagnetic energy to heat soil to temperatures as high
as 300°C. Large volumes of contaminated soil can be heated, with resultant organic contaminant
removal via steam stripping, boiling, and vaporization. This technique can be used to remediate soils
contaminated with pesticides that typically volatilize between 80°C and 300°C, such as halogenated
volatile alphatics in waste group WG03. Other pesticides that have been removed include WG02
pesticides such as aldrin, dieldrin, endrin, and isodrin. This technology is currently in the pilot and
field-scale stage and has been demonstrated at the Rocky Mountain Arsenal.
164
-------
Soil washing is used to remove contaminants from soils by dissolving or suspending them in
a wash liquid. Fine particles such as clay and silt containing adsorbed insoluble contaminants also
may be separated from the coarser soil fractions using soil washing. The wash liquid is then
separated using conventional water treatment methods, while the cleaned coarse fraction that
composes the bulk of the soil volume is backfilled or reclaimed. Soil washing can be used to treat
soils contaminated with pesticides from all four waste groups (WG01, WG02, WG03, WG04). This
technology is available for full-scale use and has been used to remove dieldrin in a bench-scale study
at the FMC Fresno Superfund site.
Thermal desorption is used to remove volatile and semivolatile organic chemicals from
contaminated soils by increasing the soil temperature. Depending on the vapor pressure of the
contaminants in the soil, temperatures are raised to 200°-1000°F to desorb organic compounds.
The compounds are then transferred to the vapor phase and are either destroyed, condensed and
reclaimed, or treated with carbon. This technology is commercially available and can be used to
remove pesticides in waste groups WG02, WG03, and WG04 from contaminated soil.
Solvent extraction uses organic solvents to remove organic compounds from contaminated
soils and reduces the contaminant volume by transferring the contaminant from the soil and
concentrating it in an extract phase. Solvent extraction treatment is applicable to soils
contaminated with organic pesticides (waste groups WG02, WG03, WG04). Solvent extraction is
commercially available. ,
REFERENCE:
USEPA Contaminants and Remedial Options for Pesticides-Contaminated Sites.
EPA Number - Not yet available.
FOR MORE INFORMATION:
Richard N. Koustas, U.S. EPA, Risk Reduction Engineering Laboratory,
2890 Woodbridge Avenue, (MS 106), Edison, New Jersey 08837
Phone: (908) 906-6898, Fax, (908) 906-6990
165
-------
CONTAMINANTS AND REMEDIAL OPTIONS AT SELECTED METALS CONTAMINATED SITES
- A TECHNICAL RESOURCE DOCUMENT
Michael D. Royer
Technical Support Branch/RREL
U.S. EPA
Bldg.# 10 (MS-104)
2890 Woodbridge Avenue
Edison, NJ 08837-3679
Phone: 908-321-6633
Lawrence A. Smith
Battelle
505 King Avenue
Columbus, OH 43201-2693
Phone: 614-424-3169
INTRODUCTION
A technical resource document, Contaminants and Remedial Options at Selected Metals-
Contaminated Sites, has been produced to assist site remediation managers to select treatment
technologies for contaminated soils, sludges, sediments, and waste deposits at sites where
inorganic arsenic (As), cadmium (Cd), chromium (Cr), mercury (Hg), or lead (Pb) are the primary
contaminants of concern. These five metals have been addressed because of their toxicity,
industrial use, and frequency of occurrence at Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA) sites and in Resource Conservation and Recovery Act
(RCRA) hazardous wastes. This document should prove useful to all remediation managers,
whether their efforts fall under federal, state, or private authority, and whether they are applying
standards from RCRA, CERCLA, and/or state programs.
METHODOLOGY
A diligent effort was made (subject to the key limitations noted below), to identify, collect,
analyze, and organize information, data, and pertinent references that a remediation manager
would find useful for identifying and selecting remedial alternatives for soils, sediments, sludges,
and waste deposits in which the principal contaminants are As, Cd, Cr, Hg, or Pb and selected
inorganic compounds of these metals. The types of information collected to support preparation of
this document include the following.
• Background information on As, Cd, Cr, Hg, Pb, and associated inorganic compounds
regarding mineral origins, processing, uses, common matrices, chemical forms, behavior,
transport, fate, and effects.
• Existing remediation performance data, listed below, in rough order of desirability: (a)
full-scale remediation of As, Cd, Cr, Hg, and Pb contaminated sites; (b) technology
demonstrations on As, Cd, Cr, Hg, or Pb contaminated sites under the EPA Superfund
Innovative Technology Evaluation Program; (c) RCRA As, Cd, Cr, Hg, and Pb bearing
hazardous wastes for which Best Demonstrated Available Technologies have been
established; (d) waste applicability/capacity information for treatment technologies as
described in technology guides and the EPA Vendors' Inventory of Superfund Innovative
166
-------
Treatment Technologies (V1SITT) database; (e) feedstock specification information for
primary or secondary smelting or recycle/re-use markets; (f) Records of Decision (RODs)
and corresponding summaries for As, Cd, Cr, Hg, and Pb contaminated sites; (g)
Treatability test data on As, Cd, Cr, Hg, and Pb contaminated matrices where the results
are well-documented and in an accessible form (e.g., Alternative Treatment Technology
Information Center [ATTIC] and the Risk Reduction Engineering Laboratory treatability
database) (h) Superfund National Priority List sites where As, Cd, Cr, Hg, or Pb
contaminated media is a primary concern and remedial options are or will be under
evaluation.
It is assumed that the remediation manager is familiar with appropriate policy issues (RCRA,
CERCLA, and state), site characterization, sampling methods, analytical methods, risk assessment,
determination of cleanup levels, and health and safety plans. It is also assumed that the
manager or available support staff is familiar with widely available references from which physical
and chemical data for the 5 metals of interest and their compounds can be obtained.
Containment and water treatment technologies are primarily addressed by reference, since
they are well described and evaluated in recent, available documents, which are referenced in
Section 4 of the technical resource document. ,
To avoid overlap with existing or forthcoming documents, information collection and
coverage of four specific types of metals sites [lead battery recycling, wood preserving (As, Cr),
pesticides (As, Hg), and mining] was intentionally limited to selected cases where innovative
technologies have been chosen or applied.
In the interests of simplicity, brevity, and due to constraints imposed by limited project
resources, the reference document does not attempt to systematically address remediation of
organometallic compounds, organic-metal mixtures, and multi-metal mixtures. For example, while
incineration is noted as a potential pretreatment for an organic-metal-soil mixture, the effects of As,
Cd, Cr, Hg, or Pb on the technical and economic feasibility of incineration are not discussed.
Another example is that several RCRA Best Demonstrated Available Technologies are cited for
multi-metal wastes, but there is no discussion on how, in general, one should select a remedial
technology for a multi-metal waste.
RESULTS
An approximately 200-page technical resource document has been produced. Section 4
(Remedial Options) and the appendices cited therein form the heart of the document. This section
begins with a brief general discussion of the key applicable or relevant and appropriate regulations
that influence cleanup goals. Soil and groundwater action levels and risk goals are tabulated for 24
metal-contaminated sites. TCLP limits for metals ir) selected metal-bearing RCRA characteristic
hazardous wastes are also tabulated.
The bulk of Section 4 addresses the immobilization, and separation/ concentration
technologies that are potentially applicable for remediating metal-contaminated solids, with the
main emphasis on soils. Each technology is addressed in a similar manner.
• A technology description is provided, followed by a discussion of typical treatment trains, and
a discussion of the applicability of the technology to various wastes. Specific reference is
made to the 5 metals of interest, when applicable information is available.
• The status (e.g., bench, pilot, full-scale, applications to Superfund remediation) and
performance of the technologies are also discussed and, where sufficient examples exist,
tabulated.
167
-------
• Cost factors and costs are also discussed with cost estimates often being drawn from
applicable Superfund Innovative Technology Evaluation (SITE) program Applications Analysis
Reports.
• Finally, data needs for assessing the applicability of each type of technology are tabulated.
The sub-section on immobilization addresses solidification/ stabilization (cement-based and
polymer microencapsulation) and vitrification (in situ and ex situ) technologies. Containment
technologies (capping and vertical and horizontal barriers) are noted, but only addressed by
reference since: (1) the type of metal contaminant is not crucial to containment system selection,
and (2) there is a recent, readily available EPA document (EPA 625/6-91/026) that already
addresses the topic at the desired level.
Separation/concentration technologies are subdivided into two categories:
• Technologies applicable for excavated solids:
— physical separation technologies [i.e., screening, classification, gravity separation, magnetic
separation, and flotation];
— soil washing technologies [i.e., extraction via water, solvents, or solutions containing
surfactants, chelating agents, acids, or bases], and
— pyrometallurgical separation technologies [i.e., Waelz kiln, flame reactor, molten metal bath,
secondary lead smelting via reverberatory and blast furnaces, submerged arc furnace, and
mercury roasting and retorting], and
• Technologies applied in situ (i.e., soil flushing, and electrokinetics).
Water treatment options are very briefly discussed, and a summary table is provided. As with
containment options, limited coverage is provided due to the availability of other recent, available
EPA documents that address the topic in an adequate manner.
Section 4 on Remedial Options is complemented by a number of key appendices.
• Appendix B summarizes 68 technologies applicable to metals-contaminated media that are
undergoing evaluation in the SITE program.
• Appendix C summarizes 67 innovative metals-contaminated technologies from 16 technology
categories. This information was excerpted from EPA's Vendor Inventory of Superfund
Innovative Treatment Technologies (VISITT) database version 2.0.
* Appendix D lists and briefly describes 44 selected metals-contaminated National Priority List
Sites.
• Appendix E summarizes Best Demonstrated Available Technologies (BOAT) for 60 RCRA
hazardous wastes that contain As, Cd, Cr, Hg, and Pb.
• Appendix F supplements the separation/concentration technology portions of Section 4 by
providing a review of metal recycling options for metal-contaminated wastes from CERCLA
Sites.
168
-------
Section 2 briefly identifies typical mineral origins, industrial uses, and Superfund matrices of
inorganic As, Cd, Cr, Hg, and Pb. Section 3 addresses possible chemical forms for the 5 metals
under various conditions. Also described in Section 3 are typical environmental transport,
partitioning, and transformation phenomena for the 5 metals in air, soil and sediment, and surface
water and ground water. Section 3 also includes a brief overview of the human and environmental
toxicity of the five metals and some of their compounds.
CONCLUSIONS
1. The technical resource document consolidates and organizes a substantial body of information
pertinent to the remediation of As, Cd, Cr, Hg, and Pb contaminated soils, sediments, and
sludges. During the development of remedial investigation and feasibility study (RI/FS) reports,
users should be able to considerably reduce time and effort required to identify, describe, and
make preliminary assessments of remedial technologies applicable to metals-contaminated
sites.
2. The chemistry of these metals, particularly arsenic, mercury, and chromium, is quite complex.
Significant differences in solubility, volatility, or toxicity are observed for various species of all
five metals. These property differences may have a substantial impact (positive or negative)
on the effectiveness, implementability, and cost of remedial alternatives. The technical
resource document (Sections 2, 3, and portions of 4) clearly indicates the potential difficulties
and thus alerts the RPM to the critical need for early and continuous consideration of the
chemistry of these metals by a knowledgable person during site characterization, and remedy
evaluation, selection, design, and implementation.
3. In addition to no action and excavation/offsite disposal, thirteen technologies were identified as
be- ing potentially applicable to the remediation of metals-contaminated soils, sediments, and
sludges.
• Although by no means appropriate for all metal-matrix combinations, the mostly broadly
applicable technologies for metals-contaminated soils are capping (not addressed in TRD),
vertical barriers (not addressed in TRD), cement-based solidification/ stabilization, screening,
gravity separation, and soil washing/acid extraction.
• A second tier of technologies is applicable to a much narrower range of situations due to
either effectiveness, implementability, or cost limitations. This second tier of technologies
includes horizontal barriers, vitrification, polymer microencapsulation, flotation,
pyrometallurgical separation, soil flushing, and electrokinetics.
• The beneficial role of chemical treatment (e.g., oxidation, reduction, neutralization) is
recognized, but not addressed as a separate technology, since it is always closely coupled
with another technology (e.g., solidification/ stabilization, vitrification) when treating metal
contaminated soils.
• Biotreatment (e.g., extraction of metals from soils using bacteria) was considered for
inclusion in the technical resource document, but rejected due to its early stage of
development.
• Only a very limited number of facilities recover the five metals in forms and concentrations
likely to be arising from Superfund site remediation. Appendix F lists pertinent facilities,
enabling the reader to easily identify and contact recyclers closest to the site to determine
interest, acceptance criteria, and costs. Sixteen potential recyclers were identified for lead
wastes and 7 for mercury wastes. For 18 other RCRA and specialized metal-bearing
169
-------
wastes containing one or more of the 5 rnetals of interest, only 18 additional potential
recyclers were identified.
4. Inorganic arsenic is difficult to treat successfully. It has multiple valences and can interconvert
between species depending on pH and oxidation/reduction potential and species present. Care
must be taken during treatment processes to ensure that volatile arsenic compounds are not
formed. Arsenic forms anionic compounds in water, and thus does not form insoluble
hydroxides during cement-based stabilization/ solidification. Solidification/stabilization may be
applied in instances where arsenic is present in low concentrations. Polymer
microencapsulation is an option for arsenic, but no instances of its application were identified.
Best Demonstrated Available Technology (BOAT) for arsenic-bearing RCRA hazardous wastes
is vitrification, but no Superfund applications of this technology have occurred to date.
Potentially viable separation/concentration options include screening, gravity separation, and
soil washing/acid/base extraction. These technologies have been selected for remediation of
several wood preserving sites. Recycling of recovered arsenic is not very promising ~ there is
little demand and therefore little or no capacity for recycling arsenic.
5. Inorganic mercury also tends to be difficult to treat. It may be converted by microorganisms
under some conditions to volatile organomercury compounds. It is amenable to
stabilization/solidification, but only at very low concentrations (e.g., for RCRA hazardous
wastes, S/S is a BOAT for Hg < 260 mg/Kg). Mercury does not form insoluble hydroxides,
and is not amenable to cement-based S/S at higher concentrations. Polymer
microencapsulation is a potential option for mercury immobilization. Physical separation
techniques {e.g., screening or gravity separation) and soil washing should be applicable to
removal of mercury from soils. These technologies are being evaluated by the Gas Research
Institute and U.S. Department of Energy. Given its low solubility in glass and low boiling
point, mercury is not a good candidate for vitrification. However, mercury's high vapor
pressure, low boiling point, and ready decomposition of its oxides enables it to be separated
from soils via thermal desorption and roasting. Only seven facilities appear to be potentially
applicable to processing mercury recovered from Superfund remediation.
6. Chromium (III) forms insoluble hydroxides and is amenable to cement-based
stabilization/solidification. Chromium (VI) does not form insoluble hydroxides, but it can be
subjected to S/S after reduction to Chromium (III). Vitrification may be technically feasible,
but cost is likely to pose a problem. Physical separation and soil washing are potentially
applicable and have been selected for several wood preserving sites. Soil flushing has been
applied at two chromium-contaminated sites. Electrokinetics shows some promise, particularly
if success is demonstrated for in situ treatment of clayey soils.
7. Inorganic lead forms insoluble hydroxides, and cement-based solidification/stabilization has
been applied full-scale to numerous lead-contaminated soils. Lead is amphoteric, so pH must
be carefully controlled during S/S processing to ensure that lead solubility remains at a
minimum. Although vitrification and polymer microencapsulation would appear to be
technically feasible for lead-contaminated soils, solidification/stabilization would be expected
to cost less to implement. Screening, gravity separation, and soil washing/acid extraction
have been implemented at two lead-contaminated sites. Whether these
separation/concentration technologies can economically attain treatment goals must be
determined on a site specific basis. Recycling of lead has been accomplished as part of some
lead site remeditions, but the value of the recovered lead typically does not offset processing
costs. Soil flushing may be applicable in some circumstances. Electrokinetics may prove to be
useful for clayey soils.
170
-------
8. Inorganic cadmium occurs mostly in the + 2 valence state and does not exhibit amphoteric
behavior. It is amenable to stabilization/solidification, although pH must be maintained in the
alkaline range to ensure that leaching does not occur. Although vitrification and polymer
microencapsulation would appear to be technically feasible for cadmium-contaminated soils,
solidification/ stabilization would be expected to cost less to implement. Screening, gravity
separation, and soil washing/acid leaching are commercially available and would appear to be
applicable, but examples of implementation where cadmium was a key contaminant were not
found. Facilities that will take cadmium-concentrate from a Superfund remediation are scarce.
Soil flushing and elecrokinetics may be applicable for special circumstances.
9. While this technical resource document consolidates information from the past in an attempt
to accelerate and improve decisions in the future, it is recognized that site-specific factors
ultimately drive the selection of the remedial alternative for any particular site. The remedial
action objectives should be clearly established and cleanup levels designated. It is of particular
importance to develop reasonable estimates of the volume, distribution, and physical and
chemical composition of each significant contaminant/co-contaminant/medium combination at
the site that will require remediation. It is similarly important to clearly define the parameters
(e.g., total metal(s) concentration, leachable metals, filtered/ unfiltered aqueous metal
concentrations), test methods (e.g., TCLP, EP Toxicity Test, other leaching test, total waste
analysis), and numerical goals that will be employed to measure treatment effectiveness. A "
risk assessment should consider transport and fate of contaminants using the best methods
available including Eh-pH, equilibrium and/or transport models where applicable.
REFERENCES
Battelle - Columbus Division. Contaminants and Remedial Options at Selected Metals-
Contaminated Sites. EPA # Not Assigned Yet,
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1995. 200 pp.
FOR MORE INFORMATION:
Michael D. Royer, Technical Support Branch/RREL,
U.S. EPA, Bldg. #10, (MS-104), 2890 Woodbridge Avenue,
Edison, NJ 08837-3679. Phone (908) 321-6633, FAX (908) 321-6640.
or
Lawrence A. Smith, Battelle, 505 King Avenue, Columbus, OH 43201-2693
Phone:614-424-3169
171
-------
OVERVIEW OF TECHNICAL RESOURCE DOCUMENT FOR SOLVENT CONTAMINANTS AND
REMEDIAL OPTIONS AT SUPERFUND SITES
Uwe Frank, Daniel Sullivan, Kenneth Wilkowski
U.S. EPA
2890 Woodbridge Avenue
Edison, New Jersey 08837
(908) 321-6626
Fax (908) 906-6990
INTRODUCTION
Solvent contamination is a persistent problem found at numerous Superfund sites. Solvents
at Superfund sites are usually halogenated or nonhalogenated organic liquids whose source can be
traced to many manufacturing, industrial and commercial processes and uses. Almost half of all
Superfund sites listed in the record of decision (ROD) summary database contain solvents as a
contaminant.
The contamination usually emanates from the improper disposal of solvents used in a wide
variety of industrial and commercial applications such as manufacturing of chemicals, preparation
of products and use as cleaning agents. Also the inadequate storage, mishandling and improper
applications of solvents have significantly contributed to the problem. Consequently, this has
resulted in affecting every medium including soil, sediment, sludge, sub-surface strata, and
secondary contamination into air and water, both surface and groundwater. This persistence
throughout the environment and the mobility of solvents has created a potentially serious health
hazard from toxic exposures to humans. To mitigate this threat, a remedial project manager (RPM)
is faced with the challenge of selecting a remedial technology which will achieve established
cleanup goals.
Consequently, EPA's Office of Research and Development (ORD) developed a technical
resource document (TRD) to assist RPMs and other remedial personnel in making this decision.
This TRD was published November 1994, and is available through EPA's Center for Environmental
Research Information (EPA/600/R-94/203). The publication discusses solvents, their properties as
environmental contaminants and most importantly, the selection of treatment technologies available
to reach established cleanup levels in soil at solvent contaminated Superfund sites.
The TRD classifies the appropriate remedial technologies for solvent contaminated soil into
3 main categories; separation, destruction and immobilization. "Separation" includes a group of
technologies that extract or separate the contaminant from the soil matrix and form a concentrate
or medium that can be more effectively treated. Therefore, this group of technologies prepare or
treat solvent contaminated soil matrices for further treatment by either destruction or recovery.
The "destruction" technologies aim to permanently remove the contamination problem and
includes those processes that destroy the contaminant by the use of thermal, chemical or
biological means. "Immobilization" is the category that includes those methods that stop the
spread or minimize the migration of solvent contamination either through the construction of
physical barriers, through chemical reactions, or by a combination of both. The remediation of
most Superfund solvent sites usually requires a combination of treatment technologies and
contaminant control methods. Even if only a single compound type or chemical class is present,
generally no single technology is capable of remediating an entire site. As a result, a treatment
train is developed which may include immobilization, separation and destruction technologies to
achieve site specific objectives and prerequisite cleanup levels.
The technologies discussed in this TRD are in different development stages; proven,
innovative, and emerging. For example, some such as incineration and capping, have been proven
172
-------
on a commercial scale. Others, such as rnicrobial degradation and soil flushing, are less proven, or
innovative, and will require site-specific treatability tests to ensure they can meet the established
cleanup levels. Emerging technologies, such as horizontal barriers, are expected to be appropriate
but have yet to be shown effective in site remediation. To facilitate the remedy selection process
and accelerate future cleanups some of these technologies are also being designated as
presumptive remedies. This approach is one tool of acceleration within the Superfund Accelerated
Cleanup Model (SACM) as is described in an EPA fact sheet (EPA 540/F/93/048).
IN-SITU SEPARATION TECHNOLOGIES
The TRD describes separation technologies as those which remove contaminants from soil to
form a stream for either contaminant collection or destruction. They include in-situ technologies such
as soil vapor extraction (SVE), steam extraction, radio frequency (RF) heating, and soil flushing; for ex-
situ treatment they include thermal desorption (TD), soil washing and solvent extraction.
SVE is an in-situ process that applies a vacuum to sub-surface soil formations through a
series of wells and removes volatile solvent contaminants. The technology is being widely applied
and is well documented in this TRD and referenced SVE publications. SVE is applicable to vadose
zones having permeabilities of 10"10 cm2 or greater and contaminants with vapor pressures greater
than 0.5 Torr. Innovative technologies such as hydraulic and pneumatic fracturing have been
demonstrated through EPA's SITE program and found effective for extending SVE to less
permeable formations.
SVE is usually combined with carbon adsorption or thermal destruction to treat the vapor
and liquid waste streams. Other combinations include condensation, biological degradation, and
ultraviolet oxidation. Contaminated groundwater can be treated and discharged. Highly
contaminated soil tailings must be collected for treatment by another technology, such as
incineration. Typical costs for SVE range from $10 - $150 per ton.
I
Steam extraction or steam injection physically separates volatile and semivolatile organics
from soil, sediment, and sludge. The process uses a combination of thermal and mechanical
energies generated by steam, hot air, infrared elements, and electrical systems to volatilize
contaminants from soil into the vapor phase. Steam extraction is an emerging technology that
appears promising, particularly if used in conjunction with SVE. Due to the heating of soil, steam
extraction can remove more of the less volatile compounds than SVE. Limited field experience has
demonstrated 90% removal of volatile and semivolatile compounds, with better recovery in higher
permeability zones and lower recoveries of high aqueous-phase solubility compounds in lower
permeability regions. The technology can be used in combination with carbon adsorption to treat
both condensed water and vapor streams. Estimates place costs for a stationary system at $50-
$300 per cubic yard.
Radio frequency heating is an innovative treatment technology for rapid and uniform in-situ
heating of large volumes of soil. The process uses electromagnetic wave energy to heat the soil
evenly to the point where volatile and semivolatile contaminants are vaporized in the soil matrix,
and vented electrodes recover the formed gases. The concentrated extracted gas and particulate
streams can be incinerated or subjected to other treatment methods such as carbon adsorption.
The technology is applicable to material that typically volatilize in the temperature range of 80° to
300°C such as aliphatic and aromatic fractions of jet fuels and gasoline, chlorobenzene, trichloro-
ethylene, dichloroethane, and tetrachloroethylene, and is predominantly for sandy soils.
Costs vary between $50 - $90 per ton.
173
-------
Soil flushing is another innovative technology that extracts contaminants from soil with
water or other suitable aqueous solutions. Soil flushing introduces extraction fluids into soil using
an in-situ injection or infiltration process. This method may apply to all types of soil contaminants,
including halogenated aliphatics, aromatics, polar organic compounds and metals. It is used in
series with other treatments that destroy contaminants or remove them from the recirculating
extraction fluid and groundwater, such as carbon adsorption, metal precipitation and air stripping.
Costs range from $50 - $120 per cubic yard..
EX-SITU SEPARATION TECHNOLOGIES
Thermal desorption is also a separation technology but is used for excavated soil. It uses
indirect or direct heat exchange to volatilize contaminants and water from soil into a carrier gas
stream for further treatment. The carrier gas stream may be either air or an inert gas. Depending
on the process selected, this technology heats contaminated media to temperatures between 200°
and 1,000° F. Off-gases may be burned in an afterburner, catalytically oxidized, condensed for
disposal, or captured by carbon adsorption beds. Thermal desorption can successfully treat most
of the contaminants found at solvent sites. It cannot effectively separate metals (arsenic,
cadmium, lead, zinc, chromium) or PAHs with boiling points above 1,000° F. Mercury, a volatile
metal, can be treated with some thermal desorption units. Bench, pilot, and full scale studies have
demonstrated that thermal desorption achieves treatment efficiencies of 99 percent or greater for
VOCs and semivolatile organic compounds (SVOCs). Some higher temperature units can treat
PCBs, pesticides, and dioxins/furans. Costs range from $80 - $350 per ton.
Soil washing, also for excavated soil, is a water-based process for mechanically scrubbing
excavated soil to remove contaminants by dissolving or suspending them in the wash solution, or
by concentrating them into a smaller volume of soil through particle size separation techniques.
Soil washing systems that incorporate both techniques yield the greatest success for soils
contaminated with heavy-metal and organic contaminants. The soil-washing process uses various
additives {surfactant, acids, chelating agents) to increase separation efficiencies. After
successful testing, the washed soil can be returned to the site or reclaimed. The aqueous phase
and the clay/silt/sludge fraction contain high concentrations of contaminants. These two streams
become waste feed for other on-or off-site treatment by incineration, thermal desorption, or
bioremediation. Soil washing is only now becoming more popular in the United States, but has
been more widely used in Europe. Costs are from $50 - $205 per ton.
Solvent extraction is similar to soil washing but uses organic solvents to remove
contaminants from soil instead of water solutions. Solvent extraction is more appropriate for
organic contaminants than inorganics and metals; it reduces contaminant volume by concentrating
them in the extract phase. The three broad categories of the solvent extraction process are
conventional solvent extraction, critical solution temperature fluid, and supercritical fluid extraction.
Solvent extraction is not generally used for extracting inorganics/metals, but would be specified for
a solvent contaminated site if there were other, more difficult to remediate compounds, such as ,
PCBs, which could not be treated by SVE or bioremediation. Costs are high, ranging from $100-
$500 per ton.
DESTRUCTION TECHNOLOGIES
The destruction technologies described in this TRD are for remediation of contaminated soil,
sludge, and sediment at solvent sites and can be divided into three categories: thermal, chemical,
and biological. Destruction technologies either destroy or detoxify hazardous wastes by altering
the chemical structure of the constituents or breaking down the chemical structure into its basic
components. They include incineration, pyrolysis, biodegradation and chemical dehalogenation.
174
-------
Incineration treats organic contaminants in solids, liquids, and gases by subjecting them to
temperatures greater than 1,000° F in the presence of oxygen. This causes the volatilization and
combustion of the organic contaminants and converts them to carbon dioxide, water, hydrogen
chloride, nitrogen oxides, and sulfur oxides. Three common types of incineration systems can treat
contaminated solids; the rotary kiln, the infrared incinerator, and the circulating fluidized bed (CFB)
units. The rotary kiln and the infrared units contain a primary chamber that usually operates in the
temperature range of 1,000° F - 1,800° F. The rotary kiln is a refractory-lined, slightly-inclined,
rotating cylinder that serves as a combustion chamber. The infrared unit uses electric resistance
heating elements or indirect-fired radiant U-tubes tb heat material passing through the chamber on
a conveyor belt. The CFB uses air to circulate and suspend waste in a combustion loop and
operates from 1,500° F - 1,800° F. Pyrolysis differs from incineration in that it uses heat in the
absence of oxygen to volatilize and decompose organics including PAHs, PCBs, and dioxins. Clean
solids are discharged and the gases are condensed or incinerated. It operates between 1,000° F -
2,200° F. Pyrolysis is best for a mixture of organic wastes which have PCBs or dioxin as a
component because it is energy intensive and expensive and may not be appropriate for solvents
only. Costs for incineration and pyrolysis are similar ranging from $200-$ 1,500 per ton excluding
added costs for waste stream treatment.
Chemical dehalogenation is a remedy applicable to contaminated soil, sludge and sediment
at solvent sites. The dehalogenation process is effective potentially in detoxifying chlorinated
organic contaminants such as dioxins, PCBs and chlprobenzenes. This converts the more toxic
compounds into less toxic, sometimes more water-soluble products and leaves compounds that are
more readily separated from the soil and treated, in the dehalogenation of chlorinated aromatic
compounds, a nucleophilic substitution reaction replaces a chlorine atom with an ether or hydroxyl
group. Dehalogenation of chlorinated aliphatic compounds occurs through an elimination reaction
arid the formation of a double or triple carbon-carbon bond.
Bioremediation uses microorganisms to biochemically degrade or transform organic
contaminants. It attempts to foster and optimize the natural bioremediation and biotransformation
processes which occur in soils. Complete degradation of organic contaminants to carbon dioxide,
water, and inorganic products may be achievable in some cases. In terms of degree of
contaminant removal and final residual levels, the extent of treatment achievable in bioremediation
depends upon various factors including the types of contaminants present, and processes used,
and site-specific environmental conditions. In general, bioremediation does not achieve
contaminant destruction efficiencies comparable to incineration. Performance comparisons with
other contaminant removal or destruction processes should be made on a case-by-case basis. Even
when lower contaminant removal or destruction is achieved, as long as remedial action goals are
met, bioremediation may be a favored alternative based upon factors such as cost,
implementability, and public acceptability.
Three principal bioremediation processes generally apply to soils at solvent sites: solid-
phase, slurry-phase, and in-situ bioremediation. The solid-phase method places contaminated soil
in a thin layer in a lined treatment bed. It is relatively simple and inexpensive and may be effective
for a wide range of contaminants, but is land intensive due to the thin soil layer required for aerobic
microbial activity.
Slurry-phase bioremediation involves excavated soil or sludge that is mixed with water in a
tank or lagoon to create a slurry, which is then mechanically agitated. The procedure adds
appropriate nutrients and controls the level of oxygen, pH, and temperature. Potential advantages
of slurry-phase treatment as compared to solid-phase bioremediation includes the possibility for
more effective treatment due to the high degree of mixing, and the effective contact between
contaminated soils and nutrients.
175
-------
In-situ bioremediation promotes and accelerates natural processes in undisturbed soil.
Under appropriate conditions, this technology can destroy organic contaminants in place without
the high costs of excavation and materials handling. It can involve recirculation of extracted
groundwater that is supplemented above ground with nutrients and oxygen. Vacuum or injection
methods can supply oxygen to the subsurface soil. Bioventing combines in-situ bioremediation
with SVE to destroy semivolatile and some nonvolatile compounds that cannot be treated by SVE
alone. It is an emerging technology that takes advantage of aerobic biodegradation of organics by
using forced air to carry an adequate supply of oxygen to subsurface soils. Recent studies also
have attempted to bioremediate chlorinated solvents by cometabolism using low-molecular weight
alkanes such as methane.
Costs for bioremediation range from $80-$ 150 per cubic yard for slurry and solid phase
treatments, and $8-$15 per pound of contaminant for in-situ treatment.
IMMOBILIZATION TECHNOLOGIES
Technologies described in this group by the TRD include containment and
solidification/stabilization (S/S) processes designed to minimize contaminant migration.
Containment is a common component in the overall remediation of a solvent site and involves
capping systems and vertical/horizontal barriers. It also requires that long-term monitoring be
performed.
Capping can range from a native soil cover or plastic sheets to a RCRA subtitle C composite
cover. Common vertical barriers include slurry walls in excavated trenches, group curtains formed
by injecting grout into soil borings, cement-bentonite filled borings or holes formed by withdrawing
beams driven into the ground, and sheet-pile walls formed of driven steel. Horizontal barriers
underlie a sector of contaminated soil without removing any material. S/S technologies either
physically reduce the mobility of a contaminant (solidification) or chemically alters or binds the
contaminant (stabilization). Stabilization can be performed without solidification, while
solidification usually includes stabilization. Solidification also includes the use of binders for waste
bulking to facilitate the handling of liquid wastes.
Costs for the containment technologies are affected by many variables and are outlined in
the TRD. Costs for some common S/S systems range from $75-$400 per ton for landfilling on site.
WATER TREATMENT TECHNOLOGIES
Although the focus of this TRD is soil remediation, groundwater treatment is also generally
addressed. Groundwater treatment can be performed by both in-situ and ex-situ technologies. The
effectiveness of an ex-situ technology is dependent upon removing contaminants along with the
groundwater. This process usually is referred to as pump and treat. If it proves too difficult, other
technologies such as steam extraction or surfactant flushing can be used to improve contaminant
removal with the groundwater. Groundwater technologies are discussed in more detail in this TRD
and referenced publications. The following summarizes the water treatment technologies
applicable to solvent sites:
Separation/Concentration Technologies
Adsorption
Filtration
Ion exchange
Air stripping
Membrane separation
Precipitation
Oil/water separation
Air sparging
Destruction Technologies
Chemical oxidation
Biological treatment
Reverse osmosis
176
-------
SITE DEMONSTRATION OF THE DYNAPHORE/FORAGER SPONGE TECHNOLOGY
TO REMOVE DISSOLVED METALS FROM CONTAMINATED GROUNDWATER
Carolyn R. Esposito
U.S. EPA Risk Reduction Engineering Laboratory,
2890 Woodbridge Avenue
Edison, New Jersey 08837
(908) 906-6895
Gary Vaccaro
Science Applications International, Corp
411 Hackensack Avenue, Third Floor
Hackensack, New Jersey 07601
(201) 489-5200
INTRODUCTION
A Superfund Innovative Technology Evaluation (SITE) demonstration was conducted ;of the
Dynaphore/Forager Sponge technology during the week of April 3, 1994 at the N.L.
Industries Superfund Site in Pedricktown, New Jersey. The Forager Sponge is an open-celled
cellulose sponge incorporating an amine-containing chelating polymer that selectively absorbs
dissolved heavy metals in both cationic and anionjc states. This technology is a volume reduction
technology in which heavy metal contaminants from an aqueous medium are concentrated into a
smaller volume for facilitated disposal.
The developer states that the technology can be used to remove heavy metals from a wide
variety of aqueous media, such as groundwater, surface waters and process waters. The sponge
matrix can be directly disposed, or regenerated with chemical solutions. For this demonstration the
sponge was set up as a mobile pump-and-treat system which treated groundwater contaminated
with heavy metals. The demonstration focused on the system's ability to remove lead, cadmium,
chromium and copper from the contaminated groundwater over a continuous 72-hour test. The
removal of heavy metals proceeded in the presence of significantly higher concentrations of
innocuous cations such as calcium, magnesium, sodium, potassium and aluminum.
METHODOLOGY
The Forager Sponge is an open-celled cellulose sponge which contains a water-insoluble
polyamide chelating polymer for the selective removal of heavy metals. The polymer is intimately
bonded to the cellulose so as to minimize physical separation from the supporting matrix. The
functional groups in the polymer (i.e., amine groups in the polymer backbone and pendent carboxyl
groups) provide selective affinity for heavy metals in both cationic and anionic states, preferentially
forming coordination complexes with transition-group heavy metals (groups IB through V.IIIB of the
Periodic Table). The order of affinity of the polymer for metals is influenced by solution parameters
such as Ph, temperature, and total ionic content. The following affinity sequence is generally
expected by Dynaphore:
Cd++ >Cu++ > Fe + + +
Pb+ + > Au(CN)2- > Se04'2 :
> Ca + + > Mg + + » Na -
> Au+ + + > Mn-
> AsO4-3 > Hg + +
>Zn++
Cr04'2
UO/2
Ni+ +
> Ag +
> Co + + >
> AI+ + + >
The high selectivity for heavy metals and the low selectivity for alkali and alkaline earth metals (Na + ,
K + , Mg+ +, and Ca+ +) is especially useful for the treatment of contaminated natural waters which
may contain high concentrations of these innocuous chemical species. These monovalent and divalent
cations do not interfere with or compete with absorption of heavy metals from contaminated waters.
177
-------
The sponge is highly porous and thus promotes high rates of absorption of ions. Absorbed ions
can be eluted from the sponge by techniques typically employed for regeneration of ion exchange
resins. Following elution, the sponge is ready for the next absorption cycle. The useful life of the
media depends on the operating environment and the elution techniques used. Where regeneration is
not desirable or economical, the sponge can be compacted to an extremely small volume to facilitate
disposal. The metal-saturated sponge can also be incinerated with careful attention given to the
handling of resultant vapors.
The sponge can be used in columns, fishnet-type enclosures, or rotating drums. For this
demonstration, the sponge was utilized in a series of four 1.7 cubic foot columns situated on a mobile,
open trailer-mounted unit measuring approximately 50 square feet. Each column was 5 feet in height
with an 8 inch inside diameter. The columns were connected for upward flow series and each column
contained a removable fishnet bag filled with about 24,000 half inch sponge cubes. The trailer was
equipped with a.wastewater pump, water heater, and both a rotameter and positive displacement type
flow totalizer.
This technology was evaluated over a continuous 72-hour operational period, resulting in a total
treatment volume of approximately 4,300 gallons. Groundwater was pumped from an existing
monitoring well into a 6500 gallon influent storage tank. Although concentrations of some of the
critical metals exceeded cleanup goals for the site, the groundwater in the influent storage tank was
then spiked with nitrate solutions of lead, copper, and cadmium to assure effective evaluation
(quantification) of the developer's claim. The spiked solutions were added to the influent storage tank
approximately 24 hours prior to the start of the demonstration and the tank was kept well mixed via
recirculation throughout the demonstration.
The groundwater was pumped from the influent storage tank through the four-column system at
a treatment flow rate of 1 gpm or 0.08 bed volumes per minute. The influent temperature was raised
approximately 15°C to increase reaction rates in order to improve absorption of the critical metals. The
treated effluent was initially discharged to a 250 gallon portable tank from which it was subsequently
pumped to a 20,000 gallon effluent storage tank for transport off-site for treatment at a local POTW.
Grab samples for analysis of critical parameters were collected from the raw influent, final
effluent and intermediate column effluent points. In addition, equal volume 24-hour composite samples
were collected for total metals, chemical oxygen demand, total suspended solids, total dissolved solids,
sulfate and gross alpha and beta radioactivity. Since the developer reported that replacement or
regeneration of the columns should not be necessary, side tests on laboratory scale columns treating
standard metal salt solutions were performed to aid in evaluating the absorption capacity and
regenerative capabilities of the sponge.
In addition to the lead, cadmium and chromium which were the contaminants of concern at the
NL site and therefore the critical parameters for this study, copper was also considered a critical
parameter because of the high removal efficiency observed in predemonstration treatability tests.
The developer claimed that the technology would achieve at least a 90% reduction of lead and
copper, an 80% reduction in cadmium and a 50% reduction of chromium (as trivalent chromium) in the
groundwater.
178
-------
RESULTS
Analytical results of the critical parameters for the raw influent and final effuent are presented in
Table 1. These data show that the treatment claims for cadmium, copper, and lead were achieved.
However the developer did not achieve treatment claims for chromium. The treatment claim was based
on comparing the mean concentration of the raw influent to the mean concentration of the final
effluent.
Table 1: Treatment Performance for Critical Metals
Parameter
Cadmium
Chromium
Copper
Lead
90% Confidence
Interval for Avg.
Influent Cone.
(ug/L)
537 + /- 11
426 + /-31
917 +/- 14
578 +/- 12
90% Confidence
Interval for Avg.
Effluent Cone.
(ug/L)
56 +/- 13
290+/-30
25 '+/- 0
18 +/-3
90% Confidence
Interval for
Percent Removal
(%)
90 +/- 2.7
32 +/- 5.8
97 +/- 0.04
97 +/-0.59
Developer's
Treatment Claim
(%)
80
50
90
90
Effective removal of cadmium, copper and lead was achieved in the presence of a groundwater
pH ranging from 3.1-3.8; a sulfate concentration of approximately 20,000 mg/L; a TDS concentration
of approximately 23,000 mg/L and disproportionately higher concentrations of other cations such as
magnesium (72 mg/L), potassium (82 mg/L), aluminum (149 mg/L), calcium, (224 mg/L) and sodium
(6000 mg/L). The technology's low affinity for these cations was supported by the near zero removal
rates of these ions. ;
Effective removal of chromium (based on the 50% claim) was achieved within the first 10 hours
of operation until performance markedly decreased. The decrease in the removal efficiency could be the
result of the sponge's higher affinity for the other critical metals. Based on the developer's regeneration
tests on sponge cubes taken from the demonstration columns, regeneration is feasible for lead, copper
and cadmium. '
Following completion of the demonstration, the sponges were easily removed from the columns
via an overhead pulley system installed on the trailer. They were then placed in plastic sleeve bags and
hand compacted into 55-gallon drums. A waste compacting firm was able to compact 16 sleeves of
sponges into one drum, using a standard industrial waste compactor.
The cost to treat heavy metal contaminated groundwater over a one year period with the
Forager Sponge technology is estimated at $340/1000 gallons, assuming the sponges are not
regenerated and are replaced upon saturation; or $238/1000 gallons, assuming the sponges are
regenerated twice providing for three useful treatment cycles. This cost estimate assumes groundwater
characteristics are similar to the demonstration groundwater and the cadmium, lead and copper are
treated to demonstration performance claims utilizing a four-column, pump-and-treat unit similar to the
demonstration unit. The system would operate 24 hours a day, 7 days a week at a flow rate of 1 gpm
resulting in a total treatment volume of approximately 525,000 gallons.
179
-------
CONCLUSIONS
The Dynaphore Inc. Forager Sponge represents a new approach for the selective removal of
heavy metals from water. The developer contends that the cost can be greatly reduced by performing
treatability tests on the particular water to be treated and thereby modifying the chelating polymer
structure for that particular contaminated water. In addition, the sponge can be used for in-situ
applications, e.g., placing a single container bag unit of sponge vertically at the neck region of
convergent groundwater collecting zone downstream of the plume. The sponge can also be placed in a
trench installation for groundwater remediation employing a multitude of containers of sponge
horizontally placed in a sand bag type of arrangement. Containers of sponge can also be utilized tea-
bag style or placed within a conduit through which water flows by gravity effect. Loose sponge is
being evaluated in tumble operations in soil washing treatments and in stirred tanks for sludge
remediation.
FOR MORE INFORMATION
Carolyn R. Esposito, EPA Office of Research and Development, Risk Reduction Engineering Laboratory,
2890 Woodbridge Avenue (MS-106), Edison, New Jersey 08837 (908) 906-6895
180
-------
CASE STUDIES OF THE HIGH VOLTAGE ELECTRON BEAM
TECHNOLOGY FOR WASTEWATER TREATMENT
William J. Cooper, Thomas D. Waite, Charles N. Kurucz, Michael G. Nickelsen, David C. Kajdi
High Voltage Environmental Applications, Inc.
9562 Doral Boulevard
Miami, Florida 33178
Friedemann Gensel
High Voltage Environmental Applications, Inc. - Deutschland GmbH
Magdeburger Str 66, 39167 Eichenbarleben
Mary K. Stinson
USEPA Technical Support Branch/RREL (MS-104)
2890 Woodbridge Avenue
Edison, New Jersey 08837-3679
INTRODUCTION
The development of innovative technologies for the remediation of contaminated sites is
continually being considered as a treatment option for several reasons. From an economic
standpoint, the treatment costs of conventional technologies continues to increase and from an
environmental impact view, treatment technologies are sought that destroy contaminants without
creating additional disposal problems.
The high energy electron beam process has been shown to be effective in destroying many
organic compounds associated with contaminated sites1'3. The experiments to date have been
conducted at both bench and full scale. Bench scale studies are performed using a 5000 Curie 60Co-
y source whereas full scale studies have been completed at a pilot facility that has a daily capacity
of 170,000 gallons (120 gallons per minute, gpm). jhe application of this technology to problems
that face EPA could result in cost effective treatment of contaminated groundwater, soils, sediments
and sludges, as well as demilitarization (chemical weapons), and as an industrial treatment unit
process.
To provide "on-site" treatability/feasibility studies ,as well as small scale clean-ups we have now
developed a mobile system. This system is capable of treating up to 50 gpm, employing a 500 kV,
40 mA accelerator. The system is self contained and can be used in remote locations if necessary.
We have received an US Environmental Protection Agency, SITE Emerging Technology project
which has as its focus the application of this process to complex mixtures of contaminants in
groundwater, wastewater, soils, sediments, and sludges. This paper will discuss the details of
several case studies conducted at full scale using the mobile treatment process arid review data
obtained from on-going studies that have direct benefit to EPA and may assist groups within the
environmental clean-up field to more effectively accomplish their mission.
METHODOLOGY
Aqueous Chemistry of High-Energy Electrons. The scientific foundation for the use of high
energy electron beam in water treatment is based in the field of radiation chemistry4. The purpose
of this section is to provide a brief overview of aqueous-based radiation chemistry which is intended
to familiarize the reader with the chemistry and show its applicability to achieving the goals and
objectives of this project.
181
-------
Irradiation of pure water results in the formation of electronically excited states and/or free
radicals along the path of the electron. 10~7 sec after the electron has passed through a solution the
products that are present are shown in Equation 1:4iB
H20 -\AA-> [2.7] OH- + [2.6] e'aq + [0.6] H- + [0.7] H202 +[2.6] H30+ + [0.45] H2 [1]
Unlike photochemical reactions where one photon of light initiates one (molecular) reaction, a
high energy electron is capable of initiating several thousand reactions as it dissipates its energy.
The efficiency of conversion of high energy electron radiation to a chemical process is defined as G
(shown in brackets in Equation 1). G is the number of radicals, excited states or other products,
formed or lost in a system absorbing 100 eV of energy. Of the products formed in Equation 1, the
most reactive species are the oxidizing, hydroxyl radical (OH-), and the reducing, aqueous electron
(e'^,) and hydrogen atom (H-). Thus, the chemistry of primary interest in the high energy electron
irradiation process is that of these three species. The concentrations of these reactive species, at
several radiation doses, are summarized in Table 1.
TABLE 1. ESTIMATED CONCENTRATION OF TRANSIENT REACTIVE SPECIES AT
SEVERAL DOSES USING HIGH ENERGY ELECTRON IRRADIATION.
H-
OH-
H,0,
DOSE (Mrad)
mM
1
2
4
2.7
5.4
10.8
0.6
1.3
2.6
2.8
5.6
11.2
0.7
1.4
2.8
High energy electron beam irradiation is the only process that is capable of forming both highly
oxidizing and highly reducing reactive species in aqueous solutions at the same time and in relatively
the same concentration. Furthermore, no other advanced oxidation process has the capability of
generating as high an overall free radical yield per unit energy input as high energy electron beam
treatment.
Aqueous Electron. The e"aq is a powerful reducing reagent with an E° (e"aq + H — > 1/2H2) of
2.77. The reactions of the e"aq are single electron transfer, the general form of which is:
e'aq + SN -—> SN'1 ' [2]
The e"«, reacts with numerous organic chemicals and of particular interest to the field of toxic and
hazardous wastes are the reactions with halogenated compounds. A generalized reaction is shown
below:
RCI -—> R- + cr
[3]
Thus, reactions involving the e"aq may result in the dechlorination of organohalogen compounds.
Further reaction of the organic radical formed could result in the complete destruction of the
compound. The e".q also reacts with other organic compounds and would contribute to the removal
of these compounds from aqueous solutions.
182
-------
Hydrogen Atom. The hydrogen atom accounts for approximately 10% of the total free radical
concentration in irradiated water. The H- undergoes two general types of reactions with organic
compounds, hydrogen addition and hydrogen abstraction.
An example of a typical addition reaction with an organic solute of interest in contaminate source
water is that of benzene:
H- + C6H6 -T—> C6H7»
The second general reaction involving the H- is hydrogen abstraction:
H- + CH3OH > Ha- + • CH2OH
[4]
[5]
Because of the relatively small second order reaction rate constant of H- with the common radical
scavengers found in natural waters, it is possible that this transient radical may be important in
removing some of the compounds of interest in drinking water treatment. This treatment process is
the only one in which this particular radical is formed.
Hydroxyl Radical. The types of reactions that involve the OH- are addition, hydrogen abstraction,
electron transfer, and radical-radical recombination.
Addition reactions occur readily with aromatic and unsaturated aliphatic compounds. The
resulting compounds are hydroxylated radicals:
OH-
CH2 = CH2
HOCH2-CH2»
[6]
Hydrogen abstraction occurs with saturated and many unsaturated molecules, e.g., aldehydes
and ketones:
OH-
CH3-CO-CH3
CH2COCH3
H2O
[7]
Case Studies. All case studies presented in this paper were conducted in Halle-Dieskau,
Germany, utilizing a mobile electron beam treatment unit. The electron beam and ancillary
equipment are mounted in a 48 foot tractor trailer. The trailer is completely self contained, self-
shielded, and has the capability of treating waste flows of up to 50 gpm.
In all cases aim3 sample of the waste to be treated was transported by tank truck to the mobile
unit and transferred into an internal 2 m3 nalgene vessel. The waste stream was then recirculated
through the electron beam treatment process, at a flow rate of 25 gpm, for a total of anywhere
between 50 and 199 minutes (a total of 3 to 14 recirculation passes based on time), depending on
the wastewater characteristics. The electron beam was maintained at full power for the entire
study. Influent and effluent samples were collected by German scientists every one or two
recirculation passes; again depending on the waste stream treated.
183
-------
RESULTS
Table 1 summarizes the results obtained from one of the case studies. The results obtained in
this study were typical of those observed for all studies conducted.
TABLE 1. SUMMARY OF ELECTRON BEAM REMOVAL OF VARIOUS ORGANIC
COMPOUNDS FROM CASE STUDY #1.
Analyte
Z Phenol //g
Benzene JJQ
Toluene //g
Ethylbenzene fjg
m/p-Xylene //g
o-Xylene fjg
Z Aromatics fjg
Naphthalene ng
Fluorene ng
Phenanthrene ng
Anthracene ng
Pyrene ng
Benzo-(a)anthraceneng
Fluroanthene ng
Z PAH ng
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
L1
Zero
Dose
2230
6200
498
118
201
108
7125
15207
728
48
138
56
12
86
16277
0.7 1.4 2.8
Mrad Mrad Mrad
220 60 20
150 10 3.5
9 <1 BMDL
2 BMDL BMDL
4 <1 BMDL
2 BMDL BMDL
167 10 3.5
<20
<20
<10
<10
BMDL
BMDL
<10
BMDL
Overall
Removal
99.10
99.95
100
100
100
100
99.95
>99.87
>97.25
>79.17
>92.75
100
100
>88.37
100
BMDL = below method detection limit.
CONCLUSIONS
Work to date has demonstrated that high energy electrons generated by electron beam
accelerators can effectively and efficiently treat complex mixtures of hazardous compounds from
aqueous solutions. Additionally, the electron beam equipment used for this study can easily be
scaled up to handle increased waste stream flows, thereby making this innovative technology
economically competitive with existing treatment processes.
184
-------
REFERENCES
1. Cooper, W.J., M.G. Nickelsen, D.E. Meacham, T.D. Waite and C.N. Kurucz. (1992): High
Energy Electron Beam Irradiation: An Advanced Oxidation Process for the Treatment of Aqueous
Based Organic Hazardous Wastes. Wat. Poll. Res. J. Canada. 27:69-95.
2. Kurucz, Charles N; Waite, Thomas D.; Cooper, William J.; Nickelsen; Michael G. (1991): High
Energy Electron Beam Irradiation of Water, Wastewater and Sludge. Chap.1 In: Advances in
Nuclear Science and Technology. Vol.22. (Eds: Lewins, J; Becker, M) Plenum Press, New York,
1-43. .
3. Nickelsen, Michael G; Cooper, William J; Kurucz, Charles N; Waite, Thomas D. (1992): Removal
of Benzene and Selected Alkyl-Substituted Benzenes from Aqueous Solution Utilizing Continuous
High-Energy Electron Irradiation. Environ. Sci. Technol. 26(1). 144-152.
4. Spinks, J.W.T. and R.J. Woods, An Introduction to Radiation Chemistry. John Wiley & Sons,
Inc., New York, 1990, pp 574.
5. Buxton, G.V., C.L. Greenstock, W.P. Helman and A.B. Ross. Critical review of rate constants for
reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (-OH/-0") in aqueous
solution. J. Phvs. Chem. Ref. Data, 17: 513-886, 1988.
FOR MORE INFORMATION:
Mary K. Stinson
USEPA Technical Support Branch/RREL (MS-104)
2890 Woodbridge Avenue
Edison, New Jersey 08837-3679
(908)321-6683
185
-------
SITE DEMONSTRATION OF THE ZENOGEM™ TECHNOLOGY
TO TREAT HIGH STRENGTH WASTEWATERS
Daniel Sullivan
US EPA
2890 Woodbridge Ave
Edison, NJ 08837
(908) 321-6677
Michael Merdinger
Foster Wheeler Environmental Corporation
8 Peach Tree Hill Road
Livingston, NJ 07039
(201) 535-2379
William Kosco
PRC Environmental Management
644 Linn Street - Suite 719
Cincinnati, OH 45203
(513)241-0149
F. Anthony Tonelli
Zenon Environmental Systems
845 Harrington Court
Burlington, Ontario
Canada L7N 3P3
(416) 639-6320
INTRODUCTION
High strength organic wastewaters are encountered at hazardous waste sites in the form of
leachate and in some cases groundwater. The ZenoGem™ Process is designed to remove
biodegradable materials, including most organic contaminants, from wastewater to produce a high
quality effluent. This technology was accepted into EPA's Superfund Innovative Technology Evaluation
(SITE) program in summer 1992; this paper summarizes the technology demonstration performed at a
Superfund site in 1994.
DESCRIPTION OF THE TECHNOLOGY
The ZenoGem™ Process consists of an integrated bioreactor and ultrafiltration membrane
system, or ultrafilter. After equalization, wastewater enters the bioreaclor, where contaminants are
biologically degraded. In this tank, a biomass develops which contains bacterial cultures that break
down organic contaminants. Ideal conditions for biomass growth are maintained, including introduction
of air to assure sufficient aerobic conditions and optimal process temperatures. The contents are
constantly mixed by the introduction of air bubbles through a series of manifolds from the tank bottom.
The tank is totally enclosed; air is recycled and a purge is emitted through a carbon adsorption unit
before being discharged into the atmosphere. A mixture of sludge solids and un- filtered wastewater
from the ultrafilter is recycled back to the bioreactor and remains in the treatment system for periods of
several weeks. The bioreactor's size is significantly reduced because of this long sludge retention time.
Conversely, the hydraulic residence time in the bioreactor is relatively short.
The ultrafilter receives feed flow from the bioreactor. It separates treated wastewater from
biological solids and soluble materials with higher molecular weights. Ultrafittration (UF) is a pressure-
driven (typically at 60 to 70 pounds per square inch) cross flow filtration process in which the water to
186
-------
be processed flows tangentially over the surface of a membrane filter capable of separating both
insoluble materials (bacteria, colloids, suspended solids) and higher molecular weight soluble materials
from the treated water. The membrane system consists of a series of tubes, in ten-foot modules and
approximately three inches in diameter, into which the cylindrical membrane filters are inserted. Feed
flow to the UF system is continuously pulled off the bioreactor and fed into the UF system: the treated
filtrate (or permeate) flows through the membrane while the remaining feed is concentrated and
returned to the bioreactor.
SITE DEMONSTRATION OF THE TECHNOLOGY
The SITE demonstration of the technology was conducted at the Nascolite Superfund site in
New Jersey during the period from September through November 1994. The groundwater at this 17.5
acre site is contaminated with volatile organic compounds (VOCs) which result from past operations at
the facility, which included manufacturing of polymethyl methacrylate plastic sheets, commonly known
as acrylic or plexiglass.
The results of the remedial investigation/feasibility study for this site confirmed extensive
contamination of the groundwater with VOCs. In March 1988, a record of decision (ROD) was signed
requiring remedial actions of ground water extraction with on-site treatment and re-injection of the
treated effluent. Methyl methacrylate (MMA) is the major contaminant, with groundwater levels
approximating 12,000 mg/l. Other contaminants at the site include a number of VOCs, including
toluene, ethylbenzene, carbon disulfide, styrene, 2,4-dimethylphenol, benzene, trichloroethyiene, vinyl
chloride.
For demonstrations, the ZenoGem™ process equipment has been mounted inside a trailer.
Prior to the EPA SITE demonstration, it has performed major demonstrations on two types of residues:
(a) a combination of firefighting training residues and contaminated groundwater containing burned and
unburned fuel, and (b) wastewater containing aircraft fire fighting foam compounds, oil, greases,
benzene, toluene, ethylbenzene, xylenes and suspended solids.
RESULTS
The results of the 89-day test run conclusively demonstrated the efficient removal of MMA and COD
from the highly contaminated groundwater. Due to the timing of the end of the demonstration and the
deadline for this extended abstract, the data presented is preliminary and has not yet been validated.
The results of the analytical data are shown in Table 1. The concentration of MMA and COD in the
groundwater feedstock, permeate, and product are shown in Table 1 as 10-day averages up to the
shock loading on November 8. During start-up, process feed rate was approximately 300 gallons per
day (gpd) and was gradually increased to 563 gpd by October 11, 1994. After the initial shock loading,
the process feed rate started at 50 gpd and lined out at 150 gpd. The resulting reductions in MMA and
COD are also shown. The MMA removal was essentially 100% for the entire operation. The COD
removal for the permeate averaged 84% up to the shock loading and 95% after the shock loading. The
product COD removal averaged 97% for the periods in which data are available.
The shock loading test demonstrated the flexibility of the process in handling a sudden increase of
concentration of contaminants. The process was able to withstand the increased concentration and
with the reduction of feedrate after the 4-hour shock was able to achieve the 95% reductions in MMA
and COD very quickly, this indicated the dynamic nature of this integrated process.
187
-------
TABLE 1. DEMONSTRATION TEST RESULTS
Feedstock
MMA COD
Permeate
Product
MMA
COD
MMA
COD
Date
9/2-9/11*
9-13-9/25
9/26-10/6
10/7-10/19
10/20-10/31
11/1-11/7
11/8
11/8
11/9-11/18
11/20-11/23
Cone.
(mg/1)
1618
2307
2209
1720
2142
2134
2020
7140
6984
8342
Cone.
(mg/1)
3467
5564
5847
4694
4973
9317
5910
19600
17585
19921
Cone.
(mg/1)
2.968
0.743
0.027
0.018
0.020
0.020
0.020
0.020
0.017
0.025
%
Red.
99.80
99.95
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
Cone.
(mg/1)
467
452
577
485
1007
3214
1170
1180
927
936
%
Red.
84.28
81.22
90.54
89.09
78.21
80.83
80.20
93.98
94.67
95.27
Cone.
(mg/1)
N/Af
N/A
N/A
N/A
0.020
0.020
0.020
0.020
0.020
0.020
%
Red.
N/A
N/A
N/A
N/A
100.0
100.0
100.0
100.0
100.0
100.0
Cone.
(mg/1)
N/A
N/A
N/A
N/A
54
198
100
121
165
868
%
Red.
N/A
N/A
N/A
N/A
99.01
96.73
98.31
99.38
99.16
95.99
*Average values for days noted.
tData not available
Overall, the process ran very smoothly. The system was computer controlled with an alarm system
that activated a beeper retained by the operator. The run demonstrated that the process operation was
so smooth and flexible that unattended operation is extremely viable for extended periods.
The groundwater used for the demonstration was very odoriferous and contained free product. The
resulting product from the process was odorless, absent of suspended solids, and was accepted by the
local sewage treatment plant for disposal at $22.50 per 6,000-gallon tanker. The sampling and
analyses of all of the tankers showed a clear and odorless product.
CONCLUSIONS
The process was very effective in reducing highly concentrated organic contamination to POTW
disposable levels. The process effectively demonstrated the complete removal of MMA and over 95%
removal of COD.
The process demonstrated unattended operation, flexibility during sustained operations, and an
ability to easily handled a four-fold increase in contaminant concentration during shock loading. The
process could operate smoothly over a wide range of conditions and would recover quickly from upsets
encountered in Superfund operations such as loss of electricity, quadrupling of feed concentration, free
product in feedstock, and adverse weather conditions.
188
-------
RECOVERY OF MONOMERS FROM RECYCLED PLASTICS
Laura L. Sharp (913) 651-4736, Robert O. Ness Jr. (701) 777-5209
University of North Dakota Energy & Environmental Research Center
15 North 23rd Street
Grand Forks, ND 58203
and
Jose M. Sosa (713) 884-0507 •
Fina Oil and Chemical Co.
PO Box 1200
Deer Park, TX 77536
INTRODUCTION
Plastics make up approximately 20% by volume of the material disposed of in landfills in the
United States. The increased interest in recycling has focused attention on ways to expand our current
recycling efforts. Such efforts to recycle more of our plastic waste must include versatile processes that
can address the heterogeneous nature of postconsumer plastics streams in the most cost-effective
manner. Types of commodity plastics typically found in a postconsumer stream include high-density
polyethylene (HOPE), low-density polyethylene (LDPE), linear low-density polyethylene (LLDPE),
polypropylene (PP), polyethylene terephthalate (PET), polyvinyl chloride (PVC), and polystyrene (PS). In
addition to plastics such as these, a number of organic and inorganic constituents will be present,
including paper, paint, food, and various metals. These constituents are present as a result of
introduction into the plastics during manufacturing (to give a plastic product selective properties) or as
residual matter from use by the consumer. The Energy & Environmental Research Center (EERC) is one
of several groups in the United States and Europe that, over the last several years, has worked toward
developing a process to thermally break down postconsumer plastics to hydrocarbon liquids and gases.
Such a process, sometimes referred to as thermal depolymerization, thermal recycling, or feedstock
recycling, produces hydrocarbon liquids and gases that could be used for the manufacture of new
plastics or other petroleum products. The specific slate of products depends on processing conditions.
The EERC has completed studies on various aspects of thermal depolymerization using fluidized-bed
technology, including a fundamental examination of the products as a function of temperature, bed
material, and feed mix (1). This and subsequent studies have identified several relatively high-value
products possible from the process, including ethylene (C2~), propylene (C3~), and butylenes. An U.S.
Environmental Protection Agency (EPA)-sponsored program at the EERC, just beginning, proposes to
better define optimal process conditions for making these olefins (propylene, ethylene, and butylenes).
Past work at the EERC has also indicated that optimal processing conditions exist for these olefin yields.
The proposed the EPA work is based on information, presented here, that was obtained in studies
completed at the EERC under the sponsorship of the American Plastics Council (APC) and the U.S.
Department of Energy (DOE).
METHODOLOGY
Two sets of screening tests were completed on the EERC's 1-4-1 b/hr continuous fluid-bed reactor
(CFBR) bench-scale test unit (Figure 1). The intention of these tests was to determine the effect of
temperature and gas residence time on the olefin yield from a postconsumer plastics thermal
decomposition process. Two different bed materials, sand and CaO, were used. From earlier work, it
was apparent that product yields in general are dependent on feed material, decomposition temperature,
gas residence time, and bed material. Twenty-five tests were run in the CFBR, the parameters and yield
results of which are given in Table 1. The base blend material, used for the first set of tests, consisted of
60% HOPE, 20% PS, and 20% PP virgin resin (percentages are on a weight basis). The postconsumer
189
-------
Condensation Train
Feed
Hopper
Gas
To Mass Spectrometer
Back-Pressure
Valve
Glycol
Coolers
Solids
Feeder
Preheaters
Liquid
EERCLS07220.CDR
Figure 1. 1-4-lb/hr CFBR unit.
blend, used for the second set of tests, consisted of the following approximate composition 59.2%
HOPE, 20.1% PET, 0.6% PVC, 10.7% PS, 10.7% PP, and 4.7% LDPE. The base blend was chosen for
the first set of tests in order to separate the coking effects from PET (if any) from the coking that may
result from severe temperatures. While coke formation was not quantified, it is desirable to know if it is a
result of the particular composition of feed material or of the temperature chosen.
RESULTS
Figure 2 shows the C2~ + C3~ yields for the three test series performed with the base blend at
residence times of 6, 16, and 30 seconds. Yields are calculated as weight of a particular product divided
by the moisture-ash free feed material. The C2~ + C3" yield will be referred to as "the olefin yield." Yield
percent Is defined as mass of a particular product divided by mass of moisture-, ash-free plastic fed.
Even though other olefins (i.e., butylenes) are present in the gas stream, these two components are the
most prevalent and will be used as an indication of conversion to olefins. Other components in the
product gas stream include H2, CH4, CO, CO2, C2H6, C3H8 and other hydrocarbon gases up to C5's. The
majority of the product gas stream is N2 since it is used as the reactor fluidization gas. A small amount
of uncondensed liquid are also present. The three test series using base blend were all completed in a
sand bed. All three residence time series show an initial increase in olefin yield with temperature, a peak
at about 700°C, then decreased olefin yield. This is consistent with earlier work with HOPE, which
showed that at the higher temperatures (750°-850°C), liquids became more aromatic, liquid yield
Increased, and gas yield decreased (1). For the 6-second residence time tests, the C2~ maximum is at
about 750°C, and for C3~ it is at about 700°C. CH4 yield increases continually with temperature; H2
increases until 825°C, where yield drops slightly. For the 16-second residence time test, C3" yield
peaks at about 650°C, and C2" yield at about 700°C. CH4 and H2 yields increase with temperature over
the entire temperature range. For the 30-second residence time series, C3~ yield is highest for the
190
-------
TABLE 1. CFBR TEST RESULTS
Run
Number
Temp. °
Liquid Yield,
C wt%
Gas Yield,
wt%
. 6-secosd Residence Time,
M422
M423
M424
M425
M426
M427
M431
M430
M429
M428
600
650
700
750
800
825
600
650
700
750
72
45
30
35
32
50
1 6-second
47
32
32
35
28
55
70
65
68
49
C2- Yield,
wt%
fiase| Blend,
2.1
8.0
17.9
21.0
17.4
14.7
ftesidence Time, Base Blend;,
53
68
68
65
30-second Residence Time
M432
M433
M434
M435
600
650
700
750
48
51
30
62
50
49
70
36
11.0
15.3
22.6
21.0
, Base Blend,
11.9
14.6
24.6
10.3
C3v Yield, H2 Yield,
wt% wt%
ttpsjg,
2.6
9.7
15.8
0.2
0.9
0.0
10 psig,
14.0
17.2
15.2
5.6
3Qpsig,
15.2
13.8
7.4
0.6
6-second Residence Tirne, Pasteonsumer Plastics Blend,
M437
M438
M439
M440
M441
600
650
700
750
800
30
13
8
13
8
69
82
83
82
90
3.6
9.3
16.5
18.2
17.0
5.1
9.9
14.2
6.9
1.4
6-secan.cJ Residence Time* Posfconsurner Plastics Blend,
M442
M443
M444
M445
600
650
700
750
51
30
29
28
47
60
69
69
7.9
16.8
19.4
32.1
" 15-secanrf Residence Time, Pasteansumer
M447
M446
700
750
28
30
59
53
21.7
22.0
10.3
19.3
13.9
0.0
Plastics,
6.9
0.0
Sand Sect
0.0
0.1
0.3
0.4
1.2
1.0
Sand Bed
0.2
0.3
0.7
0.9
Sand Bed
0.2
0.4
1.1
1.2
10 psig, Sand Bed
0.0
0.3
0.3
0.5
1.1
10 psig, Gap Bed
0.4
0.7
0.9
1.5
10 pstg, CaO
1.3
1.2
CH4 Yield,
wt%
0.7
2.4
6.0
9.6
14.0
13.4
4.6
6.9
14.0
16.7
7.8
10.2
27.2
20.0
1.2
3.9
6.0
8.7
10.4
2.7
6.0
8.4
17.7
18.4
18.3
Material
Balance
107
108
98
95
95
86
104
103
104
103
101
96
105
88
104
108
107
110
104
108
96
122
117
106
117
600°C test. C2~ and CH4 yields peak at around 700°C, and H2 yield increases over the entire
temperature range. It should be noted that no tests were completed above 750°C for the 16- and 30-
second residence time series because of prohibitive amounts of coking.
Figure 3 shows the C2~ + C3" yields for the three test series performed with postconsumer
plastics at residence times of 6 and 15 seconds. Test series were .completed using a sand bed at
6 seconds and a CaO bed at 6 and at 15 seconds. For the series in a sand bed, maximum olefin yield
occurred at 700°C, similar to the series using the base blend. Ethylene yield peaked at 750°C, and C3~
yield at 750°C. The series in a CaO bed, however, had its maximum olefin yield at 650°C. Here, C2~
yield again peaked at 750°C, while C3~ reached a maximum at 650°C. CaO has been observed to act
as a cracking agent, lowering the temperature required to produce a specified liquid yield, relative to
sand (1). Only two tests were completed at the 15-second residence time because of a lack of time.
These two points, though, are seen to follow the same general trend of lower olefin yield with increased
temperature after the maximum.
191
-------
40
30
]o
.2
> 30
20
EERCBN11071.CDR
600
650 700 750
Temperature, ° C
800
825
6-Second Residence Time
30-Second Residence Time
16-Second Residence Time
Figure 2. C2~ + C3~ Yield from base blend in a sand bed.
40
EERCBN11070.CDR
30
TJ
20
600
650
700 750
Temperature ° C
800
825
6-Second Residence Time, Sand Bed
15-Second Residence Time, CaO Bed
6-Second Residence Time, CaO Bed
Figure 3. C2~ + C3~ yield from postconsumer plactics.
192
-------
Referring to Table 1, the base blend tests had maximum combined Cg~ and C2~ yields at 700°C
for all three residence times examined. The highest combined C3~ and C2~ yield occurred at a
residence time of 16 seconds (38%), followed by the 6-second residence time (34%). For the
postconsumer tests, the highest combined yield was at a temperature of 700°C for the 6-second sand
bed test (31%), and at a temperature of 650°C for the 6-second CaO bed test (36%). The highest
overall C3~ and C2~ yield, 38%, occurred at a 16-second residence time and a temperature of 700°C,
using the base blend.
Conditions for producing olefins will also produce significant quantities of liquids. The character
of the liquids from the tests presented here was determined using a gas chromatography flame
ionization detector (GC-FID) to obtain simulated boiling point distributions and also by comparing this
information with gas chromatography mass spectroscopy (GC-MS) data from known plastics
decomposition products to identify specific components. These liquids are fairly aromatic and may be
useful as a chemical feedstock.
CONCLUSIONS
The two most important conclusions from the olefin work performed to date at the EERC are
1) olefin yield from a thermal depolymerization process depends on decomposition temperature, gas
residence time, feed material mix, and bed material and 2) a temperature can be identified for maximum
olefin yields (at a fixed gas residence time and with a specific bed material). For an inert (sand) bed, the
temperature for maximum olefin yield is about 700°C, and for a CaO (catalytic) the temperature is about
650°C. The existence of a maxima for olefin yield is of particular importance; if plastics are
decomposed at a temperature higher than the optimal temperature for olefin yield, overall gas yield
decreases, and the relative percentages of C2~, C3~, and C4~ in the gas stream decrease, while those of
H2 and CH4 increase. CaO bed material acts as a catalyst, enabling processing at lower temperature
with equivalent yields, relative to a sand bed. This effect was also observed in an earlier
depolymerization studies at the EERC (1). In the upcoming work for EPA, temperature and residence
time will be focused to a smaller range of conditions to more narrowly define the optimal process
conditions. The effect of pressure is also being contemplated* as the olefins produced will be introduced
into an (pressurized) olefin plant.
The effect of PET in thermal depolymerization processes has been examined in the APC/DOE-
sponsored projects, as well as in a project for 3M. PET will be present in postconsumer streams in
concentrations of 10% to 30% and has been seen to have a significant impact on fluid-bed operation
and on liquid products. PET will continue to be considered in all of the research efforts concerning
thermal depolymerization at the EERC.
A decision, based on an economic evaluation of the process will have to be made as to the fate
of the liquids coproduced from a plastics-to-olefins process. These liquids will either be considered a
coproduct or recycled to the process. Future efforts will include characterization of these liquids.
REFERENCES
1. Sharp, L.L.; Ness, R.O.; Aulich, T.A.; Randall, J.C. Thermal Recycling of Plastics," EERC
publication, March 1994.
FOR MORE INFORMATION
Don Brown, Office of Research and Development Risk Reduction Engineering Laboratory, USEPA
(MLK-437), Cincinnati, OH 45268, (513) 569-7630.
193
-------
TWO-PHASE OZONATION OF CHLORINATED ORGANICS
Dibakar Bhattacharyya, Amy Freshour, and Dan West
Department of Chemical Engineering
University of Kentucky
Lexington, KY 40506-0046
(606)-257-2794
INTRODUCTION
In the last few years the amount of research being conducted in the field of single-phase ozonation has
grown extensively. However, traditional aqueous-phase ozonation systems are limited by a lack of
selective oxidation potential, low ozone solubility in water, and slow intermediate decomposition rates.
Furthermore, ozone may decompose before it can be utilized for pollutant destruction since ozone can be
highly unstable in aqueous solutions. Naturally occurring compounds such as NaHCOs also affect ozone
reactions by inhibiting the formation of OH- free radicals. To compensate for these factors, excess ozone
is typically supplied to a reactor. Since ozone generation requires considerable electric power
consumption (16-24 kWh/kg of 03), attempts to enhance the ozone utilization rate and stability should
lead to more efficient application of this process to hazardous waste treatment.
To improve the process, ozonation may be more efficiently carried out in a two-phase system
consisting of an inert solvent (saturated with 03) contacted with an aqueous phase containing pollutants.
From practical considerations, the non-aqueous phase must meet the following criteria: 1) non-toxic, 2)
very low vapor pressure, 3) high density (for ease of separation), 4) complete insolubility in water, 5)
reusability, 6) selective pollutant extractability, 7) high oxidant solubility, and 8) extended 03 stability.
Previously published studies (1) have indicated that a number of fluorinated hydrocarbon compounds fit
these criteria. For this project, FC40 (a product of 3M Co.) was chosen due to its low vapor pressure (3
mm Hg) and high specific gravity (1.9). The primary advantages of the FC40 solvent are that it is non-
toxic, reusable, has an ozone solubility 10 times that of water, and that 85 % of the ozone remains in the
solvent even after 2 hours. This novel two-phase process has been utilized to study the rapid destruction
of pentachlorophenol (PCP), 1,3 dichlorobenzene (DCB), trichloroethylene (TCE), and organic mixtures.
METHODS
Ozonation studies were conducted using pentachlorophenol (PCP), trichloroethylene (TCE), and 1,3-
dichlorobenzene (DCB) (Aldrich Chemical) as parent compounds. For all experiments, the non-aqueous
solvent was FC40, obtained from 3M Co. The initial concentration ranges were varied for each
compound: PCP at 10 -100 mg/L for pH 3.5 -11.7, TCE at 1442 mg/L for pH 4.4 -10.0, and DCB at 26
mg/il and 78 mg/L for pH 2.1 -11.2. The first stage of the study was to determine the partitioning
coefficient for each compound studied. The next stage consisted of determining the effect of various
variables on parent compound degradation (such as the ratio of 03 to pollutant, reaction time, and initial
pH).
Analysis of the parent compounds in the aqueous phase were conducted following the guidelines
outlined EPA Methods 624 and 625 using a Hewlett-Packard 5890 Series II Gas Chrornatograph with an
attached 5971A Quadrupole Mass Selective Detector (GC/MS) equipped with NIST/EPA/MSDC 49K Mass
Spectral Database. For analysis of the solvent phase, the parent compounds were extracted with
methanol and analyzed by GCMS. Besides quantifying the parent compounds, selected intermediates
and reaction products were also analyzed. Organics acids from PCP ozonation were monitored on HPLC
194
-------
was 3 mg/L with ± 3% reproducibility. Chloride concentration in the aqueous phase was measured by an
Orion combination electrode (Model 96-17B) with a reproducibility of ± 4%. In order to measure ozone
concentration in the solvent phase, the modified indigo dye method was utilized (2). For quick analysis of
ozone, the concentration could be determined by measuring the absorbance at X = 290 nm with a Bausch
and Lomb Spectronic 1001 spectrophotometer.
RESULTS
The major focus of this research was to quantitatively establish the feasibility of degrading chlorinated
organics in a short reaction time. The major variables studied were O3 dosage, feed pH, and reaction
time. The experiments were characterized in terms of parent compound degradation, pH drop, free
chloride formation, and selected intermediate identification. The only source of ozone is the solvent
(FC40) phase. Feed ozone dosage is reported in terms of molar ratio of ozone to parent compound (M).
Percent chloride formation is the amount of chloride measured in the aqueous phase after an ozonation
run compared to the maximum amount of chloride that can be released by the parent compound. For
100% chloride formation, the compound has achieved complete dehalogenation.
A distribution coefficient, KD, describes the extent that a compound partitions between the FC40 phase
and the aqueous phase. The experimentally determined KD at pH < pKa is 10.2 for PGP, 5 for DCB, 40
for TCE, and 0.003 for chlorendic acid. For ionizable compounds such as phenols, KD is dependent on
pH. The KD for PCP at pH 10.3 is 0.03.
To demonstrate ozone stability in the fluorinated hydrocarbon, the solvent (in this case, FC40) was
saturated with ozone and removed from the ozone source. No significant change in O3 concentration was
observed after the first five minutes, and after two hours in a closed container with no head space, the 03
concentration remained steady at 85% ± 2% of the initial concentration. The effect of water phase on
ozone decomposition was also established. At least 77% of the initial O3 concentration remained even
after 2 hours (92 mg/L O3). It should be noted that the solvent used in this experiment had been used
repeatedly in ozonation experiments (washed between experiments) for over a year, proving its reusability
and capacity for long term ozone stability. Free radical scavengers (3), for example bicarbonate, are
compounds which compete with other hazardous organics for the OH« during single phase ozonation. An
experiment following the previous conditions was performed with the addition of 20 mM NaHCO3 to the
aqueous phase. The addition of sodium bicarbonate (20 mM) increased the ozone stability, with an ozone
decomposition of only 10% after two hours.
To study the effect of initial pH on the degradation of PCP, experiments were conducted at pH 3-11 at
an ozonation contact time of 5 minutes. Using a 10 mg/L solution of PCP, it was found that there was ~
95% PCP degradation at initial pH 10.5 for M = 4.3, compared to ~ 92% PCP degradation at pH around 4.
Since the partitioning coefficient is significantly lower at a higher pH, this result indicates that mass
transfer is not a limiting factor in the degradation of PCP in the two-phase system, and that a higher pH is
preferred for better PCP destruction. An enhanced reaction rate at a high pH is consistent with single
aqueous phase ozonation results reported in the literature (4). PCP ozonation kinetics were studied at an
initial pH of 3.5-4.5 and 10.3. The experiments indicate that a total of 79% degradation of PCP occurs
within 30 seconds and nearly 92% total degradation occurs after 1 minute (pH=4, M=4.3). Faster
destruction were found for PCP at a basic pH (M=6.66). 95% PCP degradation occurred in the first 30
seconds. Since the parent compound degradation was rapid, the order of the reaction at high pH was
determined by varying the feed PCP concentration. The feed was varied from 10 -100 mg/L while the
initial pH, ozone dosage, and reaction time were held constant. First order kinetic rate constants were
determined for the high and low pH reactions. These rate constants are respectively 200 min"1 and 6.7
min" . Calculations prove the reactions to be first order.
195
-------
Within 1 minute, the majority of the bound chloride for PCP was converted to free chloride. All of the
free chloride (Cf) remained in the aqueous phase, since Cl" has no affinity for the FC40 phase.
Approximately 70% of the bound chloride in the system was converted to free after 1 minute. The
remaining 20 - 30 % of the chloride may have been bound within intermediate compounds instead of being
converted to free chloride. 90% dechlorination of PCP occurs after 15 minutes in the two-phase batch
reactor (M=6.66, pH 10.3). Bicarbonate is a hydroxyl radical scavenger and it has been shown that it can
alter pollutant decomposition in a single-phase aqueous system (5). The addition of 20 mM NaHCOa
decreased PCP degradation in the two-phase system by only 3% after 5 minutes reaction time (M=4.3, pH
10.3).
In an attempt to identify some intermediate organic acids formed during PCP ozonation, HPLC analysis
was used on the ozonated products. One intermediate was shown to be oxalic acid. Studies by Stowell
showed that oxalic acid is a product formed from single phase ozonation of 2-chlorophenol (6). From a
calibration of oxalic acid on the HPLC, the maximum concentration of acid produced by ozonation of a 100
mg/L solution of PCP (M = 6.66, 60 minutes, pH 10.3) was determined to be 30 mg/L. Assuming that all
the PCP is degraded to oxalic acid only (no CO2), it would theoretically yield 101 mg/L acid. In order to
simulate a continuous reaction system, one study was conducted by continuously bubbling O3 into the
reactor. The ozonation of 100 mg/L PCP (pH 11.7) yielded complete dehalogenation for a 5 minute
ozonation (100% chloride formation) and 70 mg/L oxalic acid formation.
For comparison purposes, a single aqueous phase ozonation system was operated with 100 mg/L
PCP initially at pH 10.3. Ozone was bubbled directly into the aqueous phase (400 ml) at a rate of 56.5
mg/min with no. solvent present. PCP degradation and chloride formation were shown to occur at a slower
rate than in the two-phase system. A value of 0.154 min"1 was determined as the first order rate
constant (R2 = 0.96) for PCP degradation. This value is 3 orders of magnitude lower than the rate
constant in the two-phase system at the same pH (10.3).
PCP not only degrades faster in the two-phase system, but also the O3 dosage requirement is lower.
For 400 mL of solution of 100 mg/L PCP it would require at least 14 minutes reaction time (1200 mg O3)
to achieve 90% PCP degradation, whereas over 95% PCP degradation can be acquired for the same
volume in less than 5 minutes (48 mg 03 if equal volumes solvent and water) with the two-phase system.
These calculations lead to 25 times more ozone requirement in the traditional single-phase system. This
value is significant because ozone generation requires substantial electrical power which increases
operating costs.
The effect of molar ratio was found to be negligible for the degradation of both 1,3-dichlorobenzene
(DCB) and trichloroethylene (TCE) so long as O3 was in excess. To determine the effect of time on DCB
degradation, individual runs were conducted with a feed solution of 77.3 mg/L and the reactions were
quenched at 1, 5, and 10 minutes. Results showed that degradation occurred mainly in the first minute,
and then appeared to cease after the first minute. Approximately 35 % of the DCB had been degraded
and 30 % of the chloride was released, and these values remained relatively constant for the rest of the
reaction, even after 120 minutes. To determine why the reaction slowed after the first minute, the
experiments were repeated and the ozone concentration in the FC40 was quantified using the modified
indigo dye method. The ozone concentration dropped significantly within the first five minutes. With a low
ozone concentration, the ozone might not be in excess, and may have become a limiting reactant.
To study the effect of ozonation time on TCE, several semi-batch experiments were conducted in the 1
L reactor vessel. For the experiments, 500 mL of FC40 was saturated with ozone. Next the ozone was
turned off and 300 mL of water buffered to pH 10.3 with 0.05 M KH2PO4added to the vessel. Finally
0.732 mg of TCE was pipetted into the vessel and the ozone was turned on again. A steady-state ozone
concentration of 30 mg/L was measured in the solvent phase by the modified indigo dye method. If the
solution were in equilibrium, these conditions would result in an aqueous TCE concentration of 36 mg/L
196
-------
and a solvent phase concentration of 1442 mg/L. For a reaction time of 110 minutes, complete
mineralization of TCE was achieved. The plot of the natural log of the percent chloride released versus
time proved that the mineralization reaction was first order with an apparent rate constant of 0.036 min"1
In order to rapidly dehalogenate TCE, other researchers have utilized UV light (7,8). As a comparison to
literature, Sundstrom reported a first order rate constant of 0.0903 min"1 for the degradation of TCE in a
H2O2/UV system (8). This indicates that TCE degradation by two-phase ozonation is 3.5 times faster
than H2O2/UV.
Previously, studies had been conducted with only one parent compound present initially. Therefore,
experiments were conducted with a mixture of compounds to study the effect that the presence of other
organics might have on the degradation of each compound. The mixture examined contained 21.0 mg/L
pentachlorophenol, 38.6 mg/L 1,3-dichlorobenzene, and 29.3 mg/L trichloroethylene. Pollutant
degradation was determined by analyzing free chloride released, PCP concentration in the aqueous
phase, and DCB concentration in the solvent phase. The chloride results from a batch reaction with pHi
10.3 (unbuffered) and equal volumes of solvent and water show that ~ 65% total dehalogenation occurs
within 10 minutes. Within one minute, the pH dropped to 6.2, thus indicating significant organic
degradation. The KQ values were taken into account when calculating compound concentrations. PCP
degradation occurred similarly as it did in a system containing only PCP with ~ 95% degradation in 5
minutes, but DCB degraded faster than in a DCB only system with 70% degradation within 1 minute. This
behavior might be explained by an increase in the production of OH» that are formed as a result of PCP
degradation, however, increased PCP concentration did not increase DCB or TCE degradation.
A two-phase system with a continuous supply of O3 was operated with the same mixture solution. It
was found that the total chloride released is greater than it is for a batch reaction with as much as 78%
dechlorination after 30 minutes. Although at 5 minutes ;there is a comparable amount of CI" released
(approx. 65%) for both the batch and semibatch systems, at 10 minutes, where there is no increase in
dechlorination for the batch system (probably due to O3 requirements), there is an increase with the
continuous ozone supply.
CONCLUSIONS
Overall, the two-phase ozonation system showed superior performance over single-phase systems in
several areas: higher ozone solubility and stability, lower ozone generation requirements, faster pollutant
destruction rates, and the potential for selective oxidation. Also, the solvent was reused repeatedly for
over a year, proving its reusability. The solubility of ozone in FC40 is ten times higher than in water.
Where O3 decomposes completely within 40 minutes in water (pH 6), ozone in FC40 has decomposed
only 10 % within the same amount of time. A comparison between single and two-phase ozonation
systems with PCP showed that ~ 25 times more ozone must be used for a single phase system to achieve
at minimum 90% PCP degradation.
In terms of pollutant degradation, the first order reaction rate constants for PCP and TCE were orders
of magnitude larger in the two-phase system than in single phase systems. The presence of a free radical
scavenger (sodium bicarbonate) in the aqueous phase did not inhibit the degradation of PCP. Actually, 20
mM NaHCO3 decreased the rate of ozone decomposition by a factor of ~ 10. Complete dechlorination' of
PCP was still achieved after only 5 minutes when bicarbonate was present.
DCB and TCE were shown to degraded faster when reacted in a solution of PCP, DCB and TCE than
in a single-pollutant system. It was also shown that using a continuous supply of ozone to the two-phase
system increased the amount of free chloride released for the mixture solution. .
197
-------
REFERENCES
1. Stitch, F. A. and D. Bhattacharyya. Ozonolysis of Organic Compounds in a Two-Phase Flurocarbon-
Water System. Env. Progr. 6: 224-229,1987.
2. Bhattacharyya, D., T. F. vanDierdonck, S. D. West, and A. R. Freshour. Two-Phase Ozonation of
Chlorinated Organics. J. Haz. Mat. In Press, 1994.
3. Singer, P. C. Assessing Ozonation Research Needs in Water Treatment. Journal AWWA. 82(10): 78-
88,1990.
4. Hoigne1, J. and H. Bader. Rate Constants of Reactions of Ozone with Organic and Inorganic
Compounds in Water-II, Dissociating Organic Compounds. Water Res. 17: 185-194,1983.
5. Reckhow, D. A., B. Legube and P. C. Singer. The Ozonation of Hazardous Halide Precursors: Effect
' of Bicarbonate. Water Res. 20(8): 987-998,1986.
6. Stowell, J. P., J. N. Jensen and A. S. Weber. Sequential Chemical/Biological Oxidation of 2-
" Chlorophenol. Env. Sci. Tech. 26(9-11): 2085-2087,1992.
7. Glaze, W. H. The Chemistry of Water Treatment Processes Involving Ozone, Hydrogen Peroxide and
UV Radiation. Ozone Sci. & Eng. 9: 335-352,1987.
8. Sundstrom, D.W., H.E. Klei, T.A. Nalette, D.J. Reidy, and B.A. Weir. Destruction of Halogenated
Aliphatics by Ultraviolet Catalyzed Oxidation with Hydrogen Peroxide", Hazardous Waste &
Hazardous Materials, 3:101-110, 1986.
FOR MORE INFORMATION: Contact Richard P. Lauch, U.S. EPA, RREL, Cincinnati, OH 45268,
Phone No. 513-569-7237
198
-------
EVALUATION OF BASE-CATALYZED DECOMPOSITION (BCD) PROCESS
FOR NEW YORK/NEW JERSEY HARBOR SEDIMENT DECONTAMINATION
Dennis L. Timberlake
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
INTRODUCTION ,
Dredging and disposal of sediments from the New York/New Jersey Harbor (Harbor) are
conducted on a regular basis to ensure that shipping channels are maintained for safe navigation.
Sediments that accumulate in jthese areas may contain a variety of contaminants at concentrations that
pose a range of potential risks to ecological and human health. Heavy metals (Hg, Cd, Pb, Ni, Cu, Zn,
and As), chlorinated pesticides, polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls
(PCBs), and dioxins are the major contaminants of concern in the Harbor.
Most sediments dredged from the Harbor have been disposed at the New York Dredged
Material Disposal Site which is located six miles off the New Jersey coast. Such material must pass
recently revised tests for ocean disposal of dredged material, however. The updated testing manual
contains more stringent chemical and biological testing guidelines for determining the suitability of
dredged sediments for ocean disposal. As a result, the volume of contaminated dredged material
prohibited from unrestricted ocean disposal may increase dramatically. Although alternatives to ocean
disposal have been investigated on a preliminary basis, presently there are no long term non-ocean
disposal alternatives in operation in the Harbor region.
The U.S. Army Corps of Engineers, in consultation with the U.S. Environmental Protection
Agency, identified existing technologies for the treatment of contaminated dredged material. Screening
level bench-scale tests were performed with the most promising technologies. The Base-Catalyzed
Decomposition (BCD) process was one of the technologies tested at the bench-scale level which
demonstrated high removal efficiencies for dioxin. The fact that the BCD process actually destroyed
dioxins rather than merely removing them was a significant advantage.
Based on the promising screening level bench-scale tests, the decision was made to conduct a
field demonstration of the BCD process. Prior to conducting the field demonstration, it was necessary
to conduct detailed bench-scale tests, the results of which would be used to size equipment and
determine operating conditions for the pilot unit. This paper details the results of these bench-scale
tests.
METHODOLOGY
BCD is a two-stage chemical process which operates at moderate temperatures to remove
organic contaminants and decompose chlorinated hydrocarbons from soils and sediments. In the first
stage, sodium bicarbonate is added to the sediment, which is heated to about 340 C. As a result, (a)
water is evaporated and separated from the sediment into one waste stream, and (b) organic
contaminants are partially decomposed and removed in another waste stream. In the second stage,
sodium hydroxide, a hydrogen-donor oil, and a catalyst are added to the organic waste stream at 340
C.
Sediment was collected from three Harbor sites: Newark Bay, Arthur Kill, and Newtown Creek.
The plan was to use one of the sediments for initial tests to determine optimum conditions and then to
199
-------
treat all three of the sediments at the optimum conditions. The Newtown Creek sediment was chosen
as the sediment for the optimization tests because it had the highest concentration of the higher
chlorinated dioxin, furan, and PCB congeners. The untreated sediment had a total dioxin concentration
of 18 ppfa (10 ppt 2,3,7,8-TCDD), a total furan concentration of 4 ppb, and a total PCB concentration of
1.5 ppm.
RESULTS
Three process variables were evaluated during the Stage 1 optimization tests: contact time,
sodium bicarbonate dosage, and water content of the feed. The tests revealed that a contact time of
one hour, a sodium bicarbonate dosage of 10%, and the use of a predried feed material resulted in the
greatest destruction of dioxins. In addition to the treated sediment, the process also produced an
aqueous condensate and an oil condensate.
While some destruction of chlorinated organics occurred in the Stage 1 reaction, the bulk of the
chlorinated organics were transferred to the oil condensate. The oil condensate was to be treated by
addition of the BCD reagents, but the aqueous condensate required only polishing. Oil extraction,
flocculation, and carbon treatment were considered as treatment options. Testing revealed that a
combination of flocculation and carbon treatment provided the highest removal efficiencies of
chlorinated organics from the aqueous condensate.
Funds did not exist to treat all three sediments as planned, so only the Newtown Creek
sediment was run at the optimum conditions. Four runs were performed to collect sufficient sidestream
material to allow for characterization. The oil condensate underwent the Stage 2 BCD reaction with the
addition of sodium hydroxide, catalyst, and a hydrogen donor. The aqueous condensate was treated
with flocculants followed by carbon.
Preliminary results from the bench-scale testing indicate that the percent removals of dioxins,
furans, and PCBs from the sediment were >99% with total dioxin and furan concentrations in the
treated sediment at non-detect levels and total PCBs at 22 ppb. The aqueous condensate was
relatively free of contaminants prior to flocculation and carbon treatment and these contaminants were
at non-detect levels after polishing. The bulk of the contaminants ended up in the oil condensate where
Stage 2 BCD treatment resulted in destruction efficiencies >99%.
There was no evidence of dioxin/furan destruction in the Stage 1 reactor, but there was
evidence of partial dechlorination and congener shifting. Essentially all of the dioxins and furans ended
up in the oil condensate where Stage 2 destruction efficiencies exceeded 99%. About 50% of the
PCBs were completely dechlorinated in the Stage 1 reactor. The remaining 50% showed up in the oil
condensate and were completely destroyed during the Stage 2 reaction.
Using the bench-scale data, a mass balance of the BCD process revealed that for every ton
(2000 Ibs) of wet sediment (67% water content) treated, 0.33 tons (660 Ibs) of treated solids, 163
gallons of aqueous condensate, and 1.3 gallons of oil condensate would be generated. Contaminants
from one ton of contaminated sediment would be concentrated into the 1.3 gallons of oil condensate
where Stage 2 treatment would completely dechlorinate the remaining dioxins, furans, and PCBs.
CONCLUSIONS
The BCD process was successful in destroying chlorinated organic material present in Newtown
Creek sediments collected from the New York/New Jersey Harbor. Destruction efficiencies for dioxins,
furans, and PCBs were >99% and concentrations in the treated sediment were at or below detection
limits.
200
-------
Having completed the bench-scale testing portion of this study, effort is now being directed
towards the pilot-scale field demonstration of the BCD process. The demonstration is being planned for
the summer of 1995.
FOR MORE INFORMATION CONTACT:
Dennis L. Timberlake :
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513)569-7547
201
-------
ELECTROKINETIC SOIL REMEDIATION: IMPACT OF AQUEOUS PHASE PROPERTIES ON
SOIL SURFACE CHARGE AND ELECTROOSMOTIC EFFICIENCY
Leland M. Vane and Gwen M. Zang
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Dr. (M/S 443)
Cincinnati OH 45268
Phone:(513)569-7799
INTRODUCTION
The in situ remediation of a contaminated soil is an exercise in mass transfer limitations. The
challenge is to mobilize the contaminant and transport it to a treatment/collection zone or to deliver
nutrients, microorganisms, or destruction chemicals to degrade the contaminant where it resides. For soils
with high hydraulic permeabilities, mobilization or treatment solutions can be hydraulically delivered to the
contaminated zones. Mobilized contaminants and degradation products can be removed in the same
manner. However, pressure driven hydraulic delivery/removal in low permeability soils, such as clays, is
Impractical.
One method of transporting solutions and compounds in low permeability soils is the application of an
electric current to the soil in a process called Electrokinetic Soil Remediation (ESR). This form of
remediation utilizes the response of charged molecules and particles to an applied voltage gradient to
effect the movement of pollutants. Driving the remediation are the electrokinetic phenomena of
electroosmosis, ion migration (electromigration), and electrophoresis. As depicted in Figure 1, most soil
particles, including clays, carry a negative surface charge. When the soil is immersed in an electrolyte, the
particles attract cations, creating a positively charged boundary layer (referred to as the charged double
layer) next to the surface of the soil particles. Application of a voltage difference across a section of soil
causes movement of the ions and associated water within the double layer toward the cathode (electron
source). The remainder of the pore fluid moves in the same direction as the double-layer fluid due to
viscous drag interactions. This net flow of pore fluid due to an applied voltage gradient is termed
"electroosmosis" (EO). EO can be utilized to remediate contaminated soils in situ by flushing out the pore
fluid and contaminants (or to deliver nonionic nutrients, surfactants, etc.).
The ions in the bulk pore fluid also respond to the applied voltage gradient with the anions being
driven to the anode and cations driven to the cathode. This movement of aqueous ions and ion-complexes
in response to the voltage gradient is referred to as ion migration or electromigration. Electromigration can
be used to recover ionic contaminants from soil even in unsaturated soils (1). Larger charged molecules
and particles also move due to an applied voltage gradient (electrophoresis). Substances which fall into
this latter category include cationic or anionic surfactant molecules and micelles, clay particles, and
polyelectrolytes. The degree to which each electrokinetic phenomenon occurs depends on the properties
of the soil/pore fluid matrix including the degree of saturation, ionic strength of pore fluid, types of
ions/charged particles present, pH of pore fluid, temperature, porosity of soil, soil composition (% clay, type
of clay, etc.), and the surface charge of the soil particles. In this paper, the effect of pore fluid properties on
the surface charge of clays and the resultant effect on electroosmosis in saturated clays will be examined.
When a charged particle is suspended in an electrolyte, ions with a charge opposite to that of the
particle will concentrate in the charged double layer. The velocity of the particle (Vp) when placed in an
electric field is dependent on the viscosity of the fluid (n), the applied voltage gradient (£), and the zeta
potential (0, or surface charge, of the particle as described by Smoluchowski's classic equation (2):
v =
p
n
0)
202
-------
Similarly, when a packed bed of clay particles is saturated with an electrolyte and exposed to a
voltage gradient, the electroosmotic volumetric flow rate (geo) resulting from the movement of solvated
ions concentrated outside the stationary layer is described by the Helmholtz-Smoluchowski equation (2):
'-An =
n
An
(2)
where /is the bulk current density in the sample and A0, A, and n are the bulk conductivity, total cross-
sectional area, and porosity of the sample, respectively. The ?, of most charged particles is dependent on
solution pH, ionic strength, types of ionic species, temperature, and type of mineral. According to
equation (2), the electroosmotic flow rate observed in packed beds of charged particles should also be a
function of these same parameters. Investigators of electrokinetic soil remediation processes have
observed the development of often dramatic pH, conductivity, temperature, and species concentration
gradients. The pH gradients arise from electrolysis reactions which occur at the powered electrodes
(shown below for unreactive electrodes):
2 H2O - 4e~ ** O2t +• 4 H+ (Anode)
2 H2O
2e- - H,t +2 OH' (Cathode)
(3)
The anode region becomes acidic (pH as low as 2) while the cathode region is basic (pH as high as 12).
Unfortunately, models for electrokinetic soil remediation processes have only recently included the Z, as a
variable (3,4). Probstein et al. incorporated t, as a fitting parameter, constant over the entire sample, while
Eykholt introduced ?, as a function of pH. Recently, Jacobs et al. mentioned the inclusion of pH and ionic
species/concentration effects on ^ in future versions of their model (5).
When electroosmosis is relied upon to transport contaminants (as with uncharged contaminants with
or without nonionic surfactants), the time required to remediate a site is proportional to the electroosmotic
flow rate - assuming that the contaminant in the pore fluid is in equilibrium with the sorbed contaminants.
Therefore, based on equation (2), variations in I, and £ directly impact the remediation time. Even when
electromigration is the desired transport process, the amount of electroosmosis must be factored into the
remediation scheme. While £ can be independently controlled at the electrodes, £ is determined solely by
the subsurface conditions which may be dramatically affected by the remediation process. For this reason,
it is critical that the ^ of the soil be evaluated based on the expected conditions during the remediation as
well as based on depth and position at the site.
METHODOLOGY ,
The zeta potentials of small particles in dilute suspensions have been routinely measured using
instruments which range from relatively simplistic manually operated optical electrophoresis devices to
more involved automated light scattering devices. These instruments allow for the rapid determination of
zeta potential as a function of solution properties with a high degree of control over these properties.
Therefore,, a wide range of variables can be investigated in a reasonable amount of time. Conversely,
bench-scale electroosmosis experiments in compressed clay samples suffer from little control over
variables such as pH, conductivity, and types of ionic species both as a function of position in the sample
and of time. As a result, it is difficult to distill £j information from bench-scale data, although the larger scale
data is required to more fully evaluate the process under field conditions.
The effect of pH, ionic strength, clay type, and ionic species on the zeta potential was evaluated using
an automated rnicroelectrophoresis instrument (ZetaSizer 4, Malvern Instruments). Dilute suspension
(0.1 g/L) samples of kaolinite, bentonite, and a local silty clay soil were analyzed. Bench-scale
electroosmosis experiments were performed with 3" diameter kaolinite clay samples of 1" or 4" length. The
samples were placed between electrolyte reservoirs in which platinum electrodes were immersed. The pH
and conductivity of the electrolyte was kept constant as was the electric current through the sample. A
constant fluid level in the reservoirs was maintained by gravity overflow to a receiving container. The
electroosmotic flow rate was determined from the weights of the inflow and outflow bottles for each
203
-------
reservoir. The voltage gradient across the sample was measured using platinum mesh electrodes pressed
into the ends of the sample cell.
RESULTS
The variation of zeta potential with pH for the kaolinite, bentonite, and soil samples in 0.01 M KCI is
shown in Figure 2. It is apparent that the zeta potentials of bentonite and the soil sample were relatively
insensitive to pH. However, the zeta potential of kaolinite was found to be a strong function of pH, ranging
from -50 mV under basic conditions to approximately 0 mV at pH = 2.0. Equation (2) predicts that the
electroosmotic flow rate-voltage gradient ratio (qeJE) is proportional to the zeta potential. As a result, the
flow/voltage ratio should be pH dependent for packed beds of kaolinite. In Figure 3, the flow/voltage ratio
for a 4" long packed kaolinite bench-scale sample is plotted alongside the dilute suspension kaolinite zeta
potential data (absolute value), both as functions of pH. As predicted, the flow/voltage ratio displays the
same pH dependence as the zeta potential. For the same voltage gradient, the electroosmotic flow rate
was about 3 times greater at pH=5.0 than at pH = 2.3. The zeta potential results suggest even more
dramatic increases in flow at higher pHs.
The effect of various cations, anions, and ionic strength on the zeta potentials of bentonite and
kaolinite was also investigated. Changes in ionic strength (KCI) did not greatly alter the zeta potential of
either kaolinite or bentonite. More significant changes were observed when the concentrations of the +2
cations Pb2*, Ca2+, and Cu2+ were increased. For example, the magnitude of the zeta potential of
bentonite was reduced by 1/3 when the Pb2+ concentration was increased to 100 ppm in 0.01 M KCI. This
same concentration of Pb2+, Cu2"1", or Ca2+ (in 0.01 M KCI @ pH=4) was sufficient to reverse the sign of the
charge on kaolinite, indicating that the direction of electroosmotic flow would be reversed. Reduction of the
background electrolyte concentration from 0.01 to 0.0005M KCI resulted in kaolinite charge reversal at
10ppmPb2+.
CONCLUSIONS
Zeta potential results indicate that the electroosmotic efficiency (flow/voltage ratio) in bentonite should
be relatively insensitive to pH and ionic strength variations. The zeta potential of kaolinite, however, was
found to be quite sensitive to pH. The electroosmotic efficiency for kaolinite was found to be equally
sensitive to pH. Zeta potential results further indicate that the electroosmotic efficiency as well as the
direction of electroosmosis in kaolinite will be impacted dramatically by the presence of metal cations.
These results suggest that zeta potential measurements could be used to study the impact on
eiectrosmotic efficiency of initial site conditions as well as conditions expected during an electrokinetic
remediation process.
REFERENCES
(1) Lindgren, E.R.; Mattson, E.D.; Kozak, M.W.; Electrokinetic Remediation of Contaminated Soils, NTIS
Document DE91 018683, SAND-91-0726C.
(2) Hunter, R.J.; Zeta Potential in Colloid Science: Principles and Applications, Academic Press: London,
1981.
(3) Shapiro, A.P.; Probstein, R.F.; "Removal of Contaminants from Saturated Clay by Electroosmosis,"
Environ. Sci. Technol.. 27,283-291,1993.
(4) Eykholt, G.R.; Driving and Complicating Features of the Electrokinetic Treatment of Contaminated
Soils, Ph.D. Dissertation, The University of Texas at Austin, 1992.
(5) Jacobs, R.A.; Sengun, M.Z.; Hicks, R.E.; Probstein, R.F.; "Model and Experiments on Soil
Remediation by Electric Fields," J. Environ. Sci. Health. A29,1933-1955,1994.
204
-------
' Electrical Qirrent
Figure 1. Illustration of Electrokinetic Processes in Soil Pore
10
0 -
-10 -
I
. .$ -20 -
. I
°~ ™_i
ra -OU -
o>
- N
"'-40 -
-50 -
-60
D Kaolinite
•^ Bentonite
O Soil (fine fraction)
n i,. •'. i '•', i , i. •, i ri ' i ' i '
1 2 3 4 5 67 8 ',.9 , 10, 11 '
PH
Figure 2. Variation of Zeta Potential with pH for Samples of Kaolinite, Bentonite, and a Silty Clay Soil.
(25° C, 0.01 M KCI, KCI-treated samples)
205
-------
ou
SS-
.55
"c
.2
n on
"S3
NJ,
T3
co -in -
0
'to
a:
1 -
Li.
• Fl,
O Ab
^
iw Ratio (m
solute value
\
Jday*voft)
ofzetapot
^
ential
»
2 3 4 56 7
PH
Figure 3. Variation of Flow-Voltage Ratio (mLyday*volt) and Absolute Value of Zeta Potential (mV) with pH
for 3" dia. Kaolinite Packed Column (4" length) and 0.1 g/L Kaolinite Suspension, respectively.
FOR MORE INFORMATION
Leland M. Vane
U.S. EPA - Risk Reduction Engineering Laboratory
26 W. Martin Luther King Dr. (M/S 443)
Cincinnati OH 45268
Phone: (513) 569-7799 FAX: (513) 569-7677
e-mail: VANE.LELAND@EPAMAIL.EPA.GOV
206
-------
BASE CATALYZED DECOMPOSITION - NEW HYDROGEN DONORS
AND REACTION ACTIVATORS/CATALYSTS
Fred K. Kawahara and Peter M. Michalakos
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
INTRODUCTION
A significant challenge today is the need to develop a fast and complete detoxification process
suitable for the destruction of polychloro biphenyl (PCBs) and all those chlorinated organic compounds
lying as waste in soils, sediments, water, air and concentrates in waste dumps. This destruction is
necessary because PCBs are toxic toward living species: trout have been killed by water containing 8
parts per billion (ppb) of PCBs while shrimp become lifeless when the water is contaminated with only 1
ppb of the toxic substance. Commercial mixtures of PCBs contain more than 50 compounds and are
widely used in plasticizers, lubricants, hydraulic fluids, paints, inks, and coatings, as well as heat
conductors in transformer fluids. These very stable compounds have persisted since the 1960s, have
been recycled through the food chains, and have been found in humans where they cause nausea,
dermatitis, bronchitis, and cancer. Since 1976 their manufacture has been banned by the U.S.
Environmental Protection Agency owing to the cumulative storage of PCBs in the human body and their
extremely high toxicity. Thus, methods for the removal and/or destruction of halogenated organic
compounds such as PCBs are required.
The Risk Reduction Engineering Laboratory (RREL) at the EPA in Cincinnati, Ohio, confirmed
in January 1989, that a chemical reaction conducted under basic conditions completely removed
covalently bound chlorine from a variety of organic compounds. This reaction was referred to as Base
Catalyzed Decomposition (BCD). 12 Laboratory tests since 1989 have been conducted to confirm the
ability of the BCD process to dechiorinate 2,4-D and 2,4,5-T, completely converting the organically-
bound chlorine to NaC1.
BCD chemistry has repeatedly demonstrated that it can destroy up to 100,000 ppm of PCBs in
dielectric fluids and can also destroy these materials in the lower concentrations found with fluids in a
soil matrix. The treatment is similar to the above with variation.1'2 (An account of substitution
processes at low and high temperatures, can be found in the monograph on innovative site remediation
technology).3
The present Base Catalyzed Decomposition (BCD) process under prescribed conditions of two
hours of heating and stirring at 340°C will hydrodechlorinate Aroclor 1242 to the extent of 97 to 99%.1i!!
An investigation was undertaken of products and the yield of biphenyl from the dechlorination of
PCBs resulting from the BCD reaction. One objective of the investigation was to determine the
conditions under which complete hydrodechlorination of PCBs can occur at 340° C or lower at two
hours or less. A second objective was to isolate and identify products in the "BCD" reaction mixture
and thus cast light upon the hydrogen donor-transfer mechanism. Moreover, as the hydrogen radical
has been observed to evolve from the donors in a copius manner, it was thought that the surplus,
unused gas may be temporarily preserved in the reaction mixture by use of another reagent.
(Amorphous carbon has been shown to be somewhat ineffective). An additional aim was to seek an
effective, inexpensive catalyst that may accelerate the rate of reaction and shorten the reaction time
required for complete hydrodechlorination of the PCBs.
207
-------
EXPERIMENT
Reagents
Aroclor 1242 (Monsanto) contained 1 wt % monochlorobiphenyl, C^NgCI,; 13%
dichlorobiphenyl, C12H8C12; 45% trichlorobiphenyl, C12H7C3; 31% tetrachlorobiphenyl C1?H6C14 and 10%
pentachloroblphenyl, C12H6C15. The following blends were used as solvents: n-paraffinic bright stock
(PBS, Vulcan Oil Company, Cincinnati); naphthenic bright stock (NBS Vulcan Oil Company, Cincinnati),
LW-110 (Sun Oil distillate) and SOW oil (Valvoline lubricant oil, Ashland Oil).
The following were tested as hydrogen donors: No. 5 Fuel Oil, (N5 Boswell Oil Company,
Cincinnati); a derivative of No. 5 Fuel Oil; polyisobutene, (L14 Amoco Oil Company, Naperville, IL);
polyalphaolefins, Rextac 2715 (R15) and Rextac 2780 (R80 Rexene Products, Odessa, TX); and
trisopropyl benzene (Aldrich).
Materials tested as hydrogen acceptors or catalysts include proprietary catalyst; Parasol AN-3S-
-20% substituted naphthalene-- (Amoco Oil Company, Naperville, IL); Indopol L-14, polybutene polymer,
(Amoco Oil); anthracene (An); 97+%, (Aldrich); phenanthrene (Ph), 99%, (Aldrich); and iron powder, Fe,
325 mesh size, (Aldrich).
Reactions
Reactions were conducted batchwise in a 250 ml three-neck flask fitted with two water-cooled
reflux condensers, placed in series, a mechanical stirrer, and a thermocouple. The effect of a stainless
steel reactor was briefly investigated as indicated in reaction K-182. The flask was heated by a heating
mantle connected to a variable autotransformer and insulated with glass wool. Flow of evolved gas
was indicated by a rotameter attached to the second serial top of the reflux condenser. The reaction
mixture of 45g of solvent oil, 5g hydrogen donors/acceptors, 2.4g sodium hydroxide, 1-3g catalyst and
1g Aroclor 1242 (the nominal PCB content was 20,000 mg/kg) was heated to 280°-340° C for 2-4 h.
Samples were obtained from the flask before and after reaction at 25-60°C and diluted 1:18 w:w with n-
hexane (gas chromatographic grade, Fisher Scientific).
Analytical Methods
Reagents and reaction samples were analyzed by cool, on-column gas chromatography using a
mass selective detector (GC-MSD). A Hewlett-Packard 5890 GC was equipped with Hewlett-Packard
5972 MSD and a J&W deactivated fused silica guard column (5 m X 0.25 mm ID) attached in series
with a J&W DB-5MS fused silica capillary column (30 X 0.25 ms ID, with 0.25 um stationary phase
coated with DB-5) The following isobaric (12 psig) program was used for the GC oven temperature:
40°C for 5 min, 25°C/min until 180°C, hold for 2.5 min, 5°C/min until 280°C, hold at 280°C for 10 min
(or until the signal returns to baseline). Peak areas were summed for the extracted ion chromatogram
(EIC) of each class of homologs and compared to calibration solutions of Aroclor 1242 in oil/hexane.
Internal standards, spikes, and other quality control measures were used as related in EPA Method
8270B.4
RESULTS
Reagent Characterization
The multicomponent hydrocarbon reagents were characterized by GC-MS without a guard
column. Various alkanes were detected in PBS including alkyl cyclohexanes, alkyl heptanes, and do-,
hexa-, hepla-decanes. The major aromatic and alkane constituents observed in the total ion
chromatogram of N5 included alkyl benzenes, alkyl naphthalenes, alkyl anthracene or phenanthrenes,
and decanes. Manufacturer's information for the mixture of No. 6 and No. 2 is known to be 60%
208
-------
alkanes 20% alkyl benzenes for the No. 5 fuel oil of Boswell Oil.
Reaction Solvents
Several high boiling solvents were studied: Paraffinic Bright Stock (PBC), Naphthenic Bright
Stock (NBS), 50 W lubricating oil (50 W) (W110 oil (LW110). These were found, from the results of the
BCD reaction, to possess varying degrees of hydrogen donation capability which is a function of the
geometric constitution of the molecule.
For example, allylic hydrogen attached to naphthalene has a much lower bond strength than
when attached to a nonaromatic moiety.
From our test results the hydrogen donor capability appears to be in the order of
50W>NSB>PBS>LW110. Differences in bond strengths, due to the effect of double bond and its
location and due to the aromatic structural differences have significant influences in donor-acceptor
qualities.
Acceptors-Donors
Test results, from the reaction in which naphthalene or dihydronaphthalene was employed
separately, provided additional information that the aromatics are indeed acceptor-donor type of the
hydrogen free radicals. However, they can be donor-acceptor-donor type if alkyl groups on the
aromatic nucleus are substituted.
Polymers
Polymers act as donor type, releasing only hydrogen free radicals, when they are linear and
prepared from a non-aromatic polymer or copolymer. An example is polypropylene.
Polymer may function as donor-acceptor type if an aromatic monomer is used. An example is
a polymer made with styrene as a monomer in a copolymer.
Heavy Fuel Oil and Related Products
A limited number of heavy fuel oil or related products were studied as donor and/or acceptor of
hydrogen free radicals. These are the following: No. 5 fuel oil (Boswell), No. 5 fuel oil (Ashland),
refined oil N464, refined oil K158, refined oil K160.
From the research data No. 5 fuel oil and the N464 appear to possess better efficacy in the
conversion of the polychlorobiophenyl to biphenyl as indicated by the percent of PBCs remaining in the
test results column using N464 itself or with R15 (a polypropylene polymer).
A New Catalyst
To the modified "BCD" type reaction which utilizes a base, an oligomer (L14), petroleum solvent
3S and paraffinic bright stock, we introduced metallic iron particles of mesh size 325 as a possible aid
for increasing the reaction rate or decreasing the reaction time. In this reaction 2.6g of L-14, 2.7g of 3S
in 34g of PBS with 2.7g of powdered iron were added in a glass flask and heated to 340° C. After one
hour, only 3% of PCBs remained. In a similar run when no iron was used 45% and 25% of PCBs
remained after two and four hours. ,
209
-------
CONCLUSION
We have met the objectives given in the introduction. A number of new hydrogen donors were
used in the BCD reaction. One reaction mechanism pathway was discovered; its usage has led to
hydrogen radical donor-acceptor transfer-donor reaction. In so doing we have been able to preserve
some of the hydrogen radicals generated via use of the dicyclic aromatic which was converted to
dihydronaphthalene which then as donor became naphthalene. A new catalyst which is inexpensive
has been found effective in this hydrodechlorination of PCBs.
REFERENCES
1. Rogers, C. J., and Kernel, A. Hazpac (1991) 233.
2. Rogers, C. J, Kornel, A. and Sparks, H. L. U.S. Patent 5,064,526, Nov 12, 1991.
3. Weftzman, L. V., Gray, K., Kawahara, F. K., Peters, R. W. and Verbicky, J., "Chemical
Technology," Anderson, W. C., ed (1994) American Academy of Environmental
Engineers, Annapolis, MD. Kawahara, F. K. "Substitutions Reactions for the
Detoxification of Hazardous Chemicals. Report RZ008 Dec. 1994, 35 pages.
4. "Test Methods for Evaluating Solid Wastes," Office of Solid Wastes, U.S. EPA Nov.
1992, SW 846, Third Edition. (Revision 2), Method 8270B.
FOR MORE INFORMATION
Carl A. Brunner
513/569-7655
210
-------
SWIRL TECHNOLOGY: PROPER DESIGN. APPLICATION. AND EVALUATION
Richard Field and Thomas P. O'Connor
Storm and Combined Sewer Control Research Program
U. S. Environmental Protection Agency (MS-106)
2890 Woodbridge Avenue
Edison, NJ 08837-3679
INTRODUCTION (908)321-6674
Swirl and vortex technologies have been with us for over thirty years now, ever since Bernard
Smisson incorporated a cylindrical vortex-type combined sewer overflow (CSO) regulator/settleable-
solids concentrator into the Bristol, England sewerage system back in the early 1960's. In the early
1970's the U.S. Environmental Protection Agency (EPA) conducted a series of projects to develop and
demonstrate swirl flow regulator/settleable-solids concentrator (swirl) technology. These projects
resulted in the EPA swirl and helical-bend flow regulators/settleable-solids concentrators and the swirl
degritter. New generations of this technology emerged after the EPA versions were developed including
the Fluidsep™ and the Storm King™ vortex-hydrodynamic separators. However, despite different
designs and applications, the main intent of the technologies are the same, i.e., to use the forces that
arise from a change in flow direction to enhance settleable-solids separation from the storm flow. A
variety of opinions have developed regarding the application of these technologies varying from
overwhelming support to detractions that question their effectiveness. This abstract will show that
proper design and placement in the sewerage system results in effective use of swirl technology.
Reliable swirl pollution control efficiency determination is principally dependent on proper sampling
and suspended- and settleable-solids analysis techniques of the influent and effluent. Simultaneous
flowrate measurement is also important. Without the complete capture ,of heavy and stratified
suspended solids (SS) across the influent flow channel or water column, the apparent performance of
the swirl will be less than the actual. Particle-settleability tests which are presented, must be conducted
before and after installation, but especially before in order to decide if the inertial characteristics of SS in
the storm flow warrants the use of a swirl.
General Description. Swirl SS separation performance depends upon its design flowrate, Qd and
the fraction of separable solids included in the total storm-flow influent SS. The swirl was developed
through hydraulic modeling studies which used representative settleable particles (based on the Froude
Number and Stokes Law) to simulate grit and organics with specific gravities (SG) = 2.65 (fine sand)
and 1.2, respectively and effective diameters (de) from 0.2 to 5 mm. It is important to appreciate this
aspect of the swirl's development and not expect significant removals of fine-grained and/or low-
specific gravity particles. Floatables were simulated with a SG range between 0.9 and 0.96 and de's
between 5 and 500 mm.
Figure 1 is an isometric of the swirl. Flow enters the swirl tangentially and follows the peripheral
wall of the cylindrical shell (vortex chamber) creating a swirling-flow pattern. The swirling action causes
separable solids to be concentrated at the bottom of the unit. The underflow (6 to 10% of the inflow at
Qd) containing the concentrated separable solids discharges through a foul-sewer outlet in the bottom,
while the clarified supernatant exits the top through an annular overflow weir and weir-plate arrangement
into a central downshaft for storage, treatment, or discharge. A baffle configuration captures floatables
in the supernatant and directs them for containment under the weir plate and weir skirt. The floatables
are carried out in the underflow as the storm flow subsides and the water level in the swirl falls. During
low- and dry-flow conditions all flow discharges from the bottom foul-sewer outlet. Design information
for the swirl is contained in the EPA publication "Design Manual: Swirl and Helical Bend Pollution Control
Devices" (Sullivan et al. 1982). The design hydraulic loadings (HL) of the swirl typically ranges from 30
to 60 gpm/ft2 and may vary as a function of the particle-settling-velocity distribution and associated SS
removal objectives.
211
-------
A Inlet Ramp
B Flow Detector
C Scum Ring
0 Overflow Weir and Weir Plate
E Spoitors
F FtoatabJes Trap
G Foul Sawar Outlet
H Floor Gutters
I Downshaft
J Secondary Overflow Wair
K Secondary Gutter
Figure 1. Isometric Drawing of Swirl (Sullivan etal. 1982)
Advantages of the swirl include the three simultaneous functions of flow regulation, settleable-solids
concentration, and floatables capture, static operation (no moving parts to induce vortex-flow patterns).
and high-rate operation (smaller size and lower cost than sedimentation). Comparatively, Storm King™
and Fluldsep™ designs are deeper and flow regulation is not a primary function. Disadvantages include
potential underflow pumping requirements (especially for the Storm King™ and Fluidsep™ due to their
relative depth) and a relatively dilute underflow which requires further treatment. Since the foul-orifice
diameter is large enough to avoid blockages (> 1 ft diameter), pretreatment by coarse screening is
unnecessary (unless the foul underflow is pumped).
Figure 2 is an elevation view of the swirl degritter. The swirl degritter (Shelley et al. 1981 and
Sullivan etal. 1974a, 1977 and 1982) is a variation of swirl technology which does not have a continuous
underflow to the wastewater treatment plant (WWTP). Instead a relatively dry mass of settleable solids
(grit/detritus) collects in a 60° conical-bottom hopper for intermittent removal.
METHODOLOGY
To properly incorporate swirl and vortex technologies into a combined-sewerage or stormwater-
drainage system, their limitations, control functions, applicability, and design idiosyncrasies must be
clearly understood. Two major problems that continue to occur, which hinder the effectiveness and
reputation of swirl and vortex technologies are: (1) inadequate or faulty sampling techniques for
assessment of their applicability, design, and control performance evaluation; and, (2) their inappropriate
application or placement in the sewerage system.
212
-------
OUTLET t
MOTOR
GRIT WASHER
AND ELEVATOR
Figure 2. Elevation View of Swirl Degritter (Sullivan ef a/. 1982)
Proper sampling, flow measuring, and analyzing is a must for swirl application assessment, design,
and treatability evaluation. The monitoring and analyses needed to properly assess potential application,
design, and evaluate the removals of and by a swirl are expensive and complex; however, it may save
additional construction costs. Money spent on monitoring and analyses that effectively assesses
settleable solids concentration and expected swirl performance may result in lower long-term cost by
eliminating unnecessary swirl facilities and/or additional controls.
Design. Prior to selecting the swirl for a combined-sewerage or separately-sewered-stormwater
system, representative samples, collected by use of appropriate sampling techniques, should be
analyzed for particle-settling-velocity distribution versus SS and associated pollutant content. This
analysis will afford the proper assessment of the swirl's applicability. If the storm flow does not contain
enough paniculate matter with the associated settling velocities of grit-like particles (SG ;> 2.65 and de ;>
0.2 mm) and swirl separable-organic particles (SG ;> 1.2 and de * 1.0 mm), then application of the swirl
will prove futile and alternative technologies should be used, e.g., standard, lower-cost flow regulators for
situations where flow regulation is the basic objective.
If particle-settling velocities indicate that swirl technology is applicable then, hydrological and
hydraulic studies must be conducted to determine the Qd. This analysis of flow must be done on a
long-term basis to achieve the best Qd and settleable-solids removal prediction by directly measuring
flowrates and mathematically modeling for calibration and verification. The use of the design storm
concept for flowrate determination and the "polluted segment" for settleable-solids analysis will not give a
proper design. To best determine whether the swirl is suited for a particular site, pilot-scale testing
should also be conducted. While this increases preliminary design costs, it is only a fraction of the
overall construction cost. A pilot test can be set up by diverting a pilot stream from the storm-flow
mainstream. A pilot test is the only way of providing direct analyses of swirl treatability.
213
-------
Under-designing the swirl will lead to poor treatability, while over-designing will lead to a costly and
partially useless facility. Qd selection should also consider the use and advantage of the secondary-
overflow weir (emergency spillway) located on the periphery of the swirl. The secondary- overflow weir
provides flow relief to maintain a level of settleabie-soiids separation by decreasing internal vortex flow.
Design of secondary-weir elevation should limit the f lowrate to the central downshaft to about twice Qd
and allow overflows a few (two to six) times a year while also alleviating upstream flooding.
If site constraints are encountered, the swirl affords some flexibility in its sizing. Depending on
depth (e.g., difference in elevation between the trunk combined sewer and the interceptor) and area
constraints of the site, the height to diameter configuration of the swirl can be adjusted, i.e., by using a
greater diameter for a lesser height (Sullivan et al. 1974b and 1982). Design flexibility is also given for
separable-solids removal as a function of swirl size, i.e., the greater the removal the larger the required
size. Accessibility must also be included for the purposes of maintenance and monitoring. A careful
assessment of additional cost requirements must be made before applying the swirl when underflow
pumping is required.
Treatabilitv Evaluation. To determine the swirl's settleable-solids treatment effectiveness, sampling
techniques have to provide the representative fraction of relatively heavy SS in the influent and effluent
channels. Some of the equipment and strategies recommended to achieve this goal are:
1) taking grab samples or 0.2 mm aperture screening (at intermittent, short intervals) of the treated
overflow effluent. These samples should not contain significant amounts of grit-like or relatively
large organic particles. Capture of theoretically-separable solids indicates poor treatability.
2) sampling for influent SS. To do this properly, samples must be taken across the flow's entire
cross-section. This entails having a sampling system with intake velocities greater than the main
stream velocity to draw up the heavier particles (particles to be separated out by the swirl) and
multi-leveled ports to capture stratified heavy particles. This is essential for determining the SS
separation performance of the swirl. Alternatively a "slice-sampler" capable of automatically taking
a cross-sectional slice or segment of the whole influent pipe cross-section can be used.
3) performing SS analyses. Four recommended methods of SS analyses are: one for settleable
solids (gravimetrically) (Standard Methods 1985) and three for settling-velocity distribution
(classical, Brombach or German, and NIVA).
The first item (above) seems obvious but is never practiced. The second is complex and costly;
however, without proper and comparative sampling techniques, reliable data and a base of results and
performance cannot be established.
Samples taken of the influent and effluent have to be properly analyzed for SS and associated
pollutants. A gravimetric (mg/l) settleable-solids analysis (Standard Methods 1985) should be
conducted. Three methods for determining the particulate-settling-velocity distributions are the:
classical, Norwegian Institute for Water Research (NIVA), and Brombach or German. The Brombach and
NIVA methods specifically designed for the relatively high concentration of heavier particles in storm-
generated flows, require less analyses, smaller testing volumes (column heights of 60 and 70 cm,
respectively compared to 5 ft or greater for the classical method), and are amenable to field use. They
also provide better representation of high-settling velocity SS at the test starting time, tQ since the
samples are released into a settling column containing tap water from an elevated reservoir. It is almost.
impossible to attain a homogeneously mixed sample at t0 using a classical settling column.
Floatables Capture Analysis. Besides settleable solids, the swirl also removes floatable matter. A
reliable analysis for the removal of floatables has not been done. These analyses are recommended for
the assessment of swirl floatables capture ability: eyeball the floatable capture effectiveness; capture of
the influent floatables with a coarse screen; and capture of the downstream clarified effluent with a
coarse screen.
214
-------
Placement and Application. Where the swirl is placed in the control system affects its pollutant
removal effectiveness. Swirls have been improperly placed in certain projects. Swirls should not be
placed downstream of storage/sedimentation basins or after a grit chamber since swirls were developed
to remove the relatively heavy paniculate matter that would be removed by upstream sedimentation.
Effective placement or application of the swirl includes:
• upstream of a CSO storage/sedimentation facility. The swirl reduces the operation and
maintenance necessary to remove grit and sludge from the storage facility as the heavier settleable
solids are diverted directly to the WWTP.
• downstream of in-line or in-sewer CSO storage. An automated system will be able to maintain an
optimum flowrate (<; twice CL) to the swirl which will increase its design effectiveness.
• as a CSO flow regulator/settleable-solids concentrator alone with the relatively clear effluent
discharging directly to the receiving water. This applies only to cases where the receiving water
can tolerate a partially treated CSO and still meet its water quality objectives.
• upstream of CSO tunnel storage to eliminate the grit removal burden.
• upstream of created wetlands to alleviate settleable and floatable solids problems.
• using the swirl degritter when coarse or preliminary treatment of CSO is the objective and flow
regulation or splitting is not required.
• using the swirl degritter on the swirl foul-sewer underflow pipe to remove detritus, especially in
cases requiring pumping (Pisano etal. 1984) and where the interceptor has a relatively flat slope to
decrease wear and sedimentation, respectively.
• as separate stormwater discharge control.
RESULTS
The following case studies demonstrate some of the principles mentioned above. One of the most
common occurrences is not making use of settling-velocity data before the swirl is selected for an
application. This was inadvertently demonstrated at San Francisco, CA, where a small pilot-scale swirl
was tested on dry-weather sanitary sewage. The swirl, as predicted, was not suitable since the settling
velocities of the SS in the sanitary sewage were too low and not in the swirl's range of concentrating
effectiveness.
In another case, the Northeast Boundary swirl facility in Washington D.C. (O'Brien and Gere 1992)
(comprised of three 57-ft diameter swirls each having a Qd = 133 MGD, with a nominal HL = 35
gpm/ft2) averaged 38% SS-mass loading Removal for all the storm events; however, most of this
removal, i.e., 26% was accomplished by direct foul-underflow diversion to the WWTP or Reduction and
not by the swirl concentrating effect. Efficiency (= Removal - Reduction) was less than expected,
averaging 12%. The primary cause of low Efficiency was the relatively low SS-settling velocities of
particles in the influent CSO (based on settling-velocity-distribution analyses).
The previously mentioned "slice-sampler" was specially constructed during a 1980-81 demonstration
in Lancaster, PA (Pisano et al. 1984). Prior to using the slice-sampler, the Lancaster swirl had shown
"...negligible to negative solids treatment efficiency..." and further that "...samples taken manually for
settleability analyses typically contained SS concentrations much lower than concentrations of
samples..." taken by the slice-sampler. As compared to a Manning model 6000 sequential sampler, the
slice-sampler collected samples 1.5 to 7 times more concentrated with SS. The non-flow weighted
average SS Removal and Efficiency were 55% and 37%, respectively.
A demonstration in Syracuse, NY (Drehwing etal. 1979) using a 12 ft diameter swirl having a Qd =
6.8 MGD with a nominal HL = 40 gpm/ft2 was evaluated by samplers with an adequate intake sampling
technique. SS Mass loading Removal and Efficiency averaged 52 and 18%, respectively.
215
-------
CONCLUSIONS
Certain procedures must be followed to maximize swirl application including:
• predetermination of the particulate-settling-velocity distribution of representative samples and/or a
pilot study to assess treatment suitability;
• proper placement of the swirl as part of the storage/treatment system (e.g., not downstream of a
storage/sedimentation basin, grit chamber, or flow regulator);
• proper selection of Qd based on a long-term hydrological/hydraulic study;
• use of the secondary- (emergency-) overflow weir on the swirl chamber's outside cylindrical wall to
enhance separable-solids concentration at high influent flowrates;
• capture of heavy and sometimes stratified particles for treatability and suitability assessment; and
• applying the right technology, e.g., using the swirl degritter when treatment alone is the objective.
The swirl is a high-rate, relatively small and low-cost device which serves three simultaneous
functions: flow regulator, settleable-solids concentrator, and floatables-material collector. The
performance of swirl devices is dependant on the settling characteristics of the SS and fraction of
dissolved solids In the storm flow. When correctly combined with other controls of the combined-
sewerage or separately-sewered-stormwater system, swirl devices will serve a beneficial purpose.
REFERENCES
Drehwlng, FJ. et al., Disinfection/Treatment of Combined Sewer Overflows, USEPA, Edison NJ.
EPA-600/2-79-134, 1979.
Field, R. and H. Masters, Swirl Device for Regulating and Treating Combined Sewer Overflows, USEPA,
Edison, NJ. Report EPA-625/2-77-012, 1977.
O'Brien and Gere Engineers, Inc., CSO Abatement Program: Segment I - Performance Evaluation, Water
and Sewer Utility Administration, Washington, DC, April 1992.
PIsano, W.C. et al., Swirl and Helical Bend Regulator/Concentrator for Storm and Combined Sewer
Overflow Control, Environmental Design & Planning, Inc., Hanover, MA. EPA-600/2-84/151, 1984.
Shelly, G J. et al., Field Evaluation of a Swirl Degritter at Tamworth, New South Wales, Australia, G.J.
Shelly Consulting Engrs., Tamworth, N.S.W., Australia. EPA-600/2-81-063, 1981.
Sullivan, R. H. et al., The Swirl Concentrator as a Combined Sewer Overflow Regulator Facility, American
Public Works Association (APWA), Chicago, IL. EPA-R2-72-008, 1972.
Sullivan, R.H. et al.., The Swirl Concentrator as a Grit Separator Device, APWA, Chicago, IL.
EPA-670/2-74-026, 1974a. .
Sullivan, R.H. et al., Relationship between Diameter and Height for Design of a Swirl Concentrator as a
Combined Sewer Over-flow Regulator, APWA, Chicago, IL. EPA-670/2-74-039, 1974b.
Sullivan, R.H. et al., Field Prototype Demonstration of the Swirl Degritter, APWA, Chicago, IL. Report
EPA-600/2-77-185, 1977.
Sullivan, R.H. et al., Design Manual - Swirl and Helical Bend Pollution Control Devices, APWA, Chicago,
IL. EPA-600/8-82/013, 1982.
216
-------
A MULTI-CHAMBERED STORMWATER TREATMENT TRAIN FOR
THE TREATMENT OF STORM WATER
Robert Pitt, Brian Robertson,
Department of Civil and Environmental Engineering,
University of Alabama at Birmingham,
Birmingham, Alabama 35294
Richard Field
U.S. Environmental Protection Agency
Storm and Combined Sewer Pollution Control Program
2890 Woodbridge Avenue
Edison, New Jersey 08837
INTRODUCTION
Runoff from paved parking and storage areas, and especially gas station areas, has been
observed to be heavily contaminated with concentrations of many pollutants being 3 to 600 times
greater than typical receiving water criteria. These paved areas are usually found to contribute most
of the toxicant pollutant loadings to the stormwater outfalls in residential and commercial areas.
PAHs, the most commonly detected toxic organic compounds found in urban runoff, are mostly
from fossil fuel combustion. These compounds are very hard to control by eliminating their use,
and, unfortunately, their control by typical stormwater management practices is not well
understood. The major benefit of this research project will be to better understand how these
toxicants can be controlled at critical source areas, especially automobile service facilities, with the
use of a special treatment device or Multi-Chambered Treatment Train (MCTT).
METHODOLOGY
Earlier bench scale treatability studies of this U. S. Environmental Protection Agency (EPA)
sponsored research found that the most beneficial treatment for the removal of stormwater
toxicants (as measured using the Microtox™ test) included column settling for at least 24 hours
(generally 40% to 90% reductions), screening through at least 40 micron (\un) screens (20% to
70% reductions), and aeration and/or photo-degradation for at least 24 hours (up to 80%
reductions) (1). Based on these tests and a computer model based upon hydrologic conditions in
Birmingham, Alabama, a pilot-scale MCTT has been constructed and is currently being tested on the
University of Alabama at Birmingham campus. Two full-scale MCTT units are also scheduled to be
constructed in Wisconsin as part of their EPA 319 grant.
The MCTT includes a special catchbasin followed by a two chambered tank that is intended to
reduce a broad range of toxicants (volatile, paniculate, and dissolved). The runoff enters the
catchbasin chamber by passing over a flash aerator (small column packing balls with counter-current
air flow) to remove highly volatile components. This catchbasin also serves as a grit chamber to
remove the largest (fastest settling) particles. The second chamber serves as an enhanced settling
chamber'to remove smaller particles and has inclined tube or plate settlers to enhance
sedimentation. This chamber also contains fine bubble diffusers and sorbent pads to further
enhance the removal of floatable hydrocarbons and additional volatile compounds. The water then
enters the final chamber at a slow rate to maximize pollutant reductions. The final chamber contains
a mixed media (sand and peat) slow filter, with a filter fabric layer. The MCTT would typically be
sized to totally contain all of the runoff from a 12.7 millimeter (mm) (0.5 inch) rain from a typical
0.2 hector (ha) (0.5 acre) gas station. If the area is larger, then multiple, or larger, units will be
needed. Figure 1 is a diagram of this device.
217
-------
Colchbosln Main Settling Chamber
Packed Column ~ sorbent pillows
aerators ~" "n* bubble aerators
— tube settlers
Filtering Chamber
— sorbenl filter fabric,
- mixed media filter layer
(sand and peat)
- filter fabric
— gravel packed
underdrain
Rgure 1. Multi-Chambered Treatment Train (MCTT) schematic,,
Catchbasins have been found to be effective in removing pollutants associated with coarser
runoff solids (2,3). High reductions in total and suspended solids (SS) (up to 44% reduction,
depending on the inflowing water rate) were indicated by a number of prior studies. While relatively
few pollutants are associated with these coarser solids, their removal will decrease maintenance of
the other chambers of the MCTT. The size of the MCTT catchbasin sump is controlled by three
factors: the runoff flow rate, the suspended solids (SS) concentration in the runoff, and the desired
frequency at which the catchbasin will be cleaned so as not to sacrifice efficiency. Figure 2 is a
plot of the accumulation of SS versus accumulative rain for approximately sizing the catchbasin.
•"'SOaigfL ^BOmgfL ~*~ 100 mgfL •*• 200 mglL
"Z-JOOntg/L +-SOOmglL •* 800 mg/L & 1000 ntg/L
100 "
Accumulative SS Retained (ft^3/ncre)
0.01-^-r
10 20 30 40
Accumulative Rain (In)
Figure 2. Accumulation of suspended solids in catchbasin sump.
218
-------
The main settling chamber mimics completely mixed settling column bench-scale tests and uses
a treatment ratio of depth to time for removal estimates. In addition to housing plate or tube
settlers, the main settling chamber contains floating sorbent "pillows" to trap floating grease and oil
and contains a fine bubble diffuser. The settling time in the main settling chamber usually ranges
from 20 to 70 hours.
For the pilot-scale MCTT set up to capture runoff from a parking and vehicle service area on
the campus of the University of Alabama at Birmingham, the catchbasin/grit chamber is a 25
centimeter (cm) vertical PVC pipe containing about 6 liters (L) of 3 cm diameter packing column
spheres. The main settling chamber is about 1.3 square meters (m2) in area and 1 meter (m) deep
which with a 48 to 72 hour settling time should result in a median toxicity reduction of about 90 to
95%. The filter chamber is about 1.5 m2 in area and contains 0.5 m of sand and peat directly on
0.15 m of sand over a fine plastic screen and coarse gravel that covers the underdrain. A
Gunderboom filter fabric also covers the top of the filter media to distribute the water over the
filter surface by reducing the water infiltration rate through the filter and to provide additional
pollutant capture.
During a storm event, runoff from the parking lot is pumped into the catchbasin/grit chamber
automatically. During filling, an air pump supplies air to aeration stones located in the main settling
chamber. When the settling chamber is full, all pumps and samplers cease. After a quiescent
settling period of up to 72 hours water is pumped through the filter media and discharged.
In evaluating the pilot scale MCTT located in Birmingham, samples are collected before and
after each chamber of the device. To better estimate fate and treatability of toxicants, samples are
partitioned into filterable ("dissolved") and non-filterable ("paniculate") components before being
analyzed for a wide range of toxicants using detection limits from about 1 to 10 micrograms per
liter (jag/L) and conventional pollutants. Constituents being analyzed include heavy metals (copper,
cadmium, lead, and zinc) and organics (phenols, PAHs, phthalate esters, herbicides, and pesticides).
Particle size distributions, using a Coulter Multi-Sizer He M, are also being made, in addition to
conventional analyses for COD, major ions, nutrients, suspended and dissolved solids, turbidity,
color, pH, and conductivity. Samples are also screened using the Microtox™ toxicity test to
measure relative reductions in toxicity.
RESULTS
For runoff of 13 storm events treated by the MCTT over a period of four months, 10 were
found to be above a threshold value of 12% as measured using the Microtox™ test. The reduction
in toxicity ranged from 70 to 100%, with a median reduction of 90% which was the value
predicted by initial design. The toxicity of the filtered samples were also substantially reduced (50 to
100%, median of 88%). As expected, almost all of the total sample toxicity reduction occurred in
the sand and peat filter and in the main settling chamber. This can be seen in Figure 3 which is a
plot of relative toxicity through the MCTT for runoff treated from each storm event. The results
from the specific metal and organic analyses are not yet available, but these toxicity data indicate
excellent toxicant removals, as expected based on earlier bench-scale tests. "•;;
' ''T> •
Overall COD reductions varied substantially ranging from 0 to 100% (median of 53%) with
influent concentrations ranging from 0 to 197 mg/L (median of 56 mg/L). Most of these reductions
occurred in the filtering chamber. Dissolved COD reductions also showed substantial variations with
concentration increases occurring in the filtering unit for some events and most of the dissolved
COD reductions occurring in the main settling chamber.
219
-------
c g
£ 70
Inlet
Catch Basin Settling Chamber
Figure 3. Relative toxicity through MCTT.
Peat-sand
Inlet Catch Basin Settling Chamber Peat-sand
Figure 4. Suspended solids concentration through MCTT.
Inlet Catch Basin Settling Chamber
Figure 5. Turbidity changes through MCTT.
Peat-sand
Suspended solids reductions ranged from 25 to 100% overall for samples with concentrations
ranging from 7 to 137 mg/L. This includes increases that occurred in the sand and peat filter due to
limited media washout during filtration. The first filtration run resulted in the largest washout, as
would be expected. The suspended solids reductions before the filter was much greater, being
220
-------
about 70 to 100%. Figure 4 shows suspended solids concentrations through the MCTT for each
storm. Turbidity, shown in Figure 5, and color also experienced substantial increases during the
filtration process (turbidity going from 2 to 16 NTU in the influent to the complete treatment unit to
effluent values of 2.4 to 7.7 NTU, still relatively low, and color going from 20 to 58 HACH units in
the influent to 30 to 100 HACH units in the effluent). The peat also apparently caused a reduction
in pH, from 6.29 to 7.27 in the filter influent to 5.93 to 6.78 in the effluent.
CONCLUSIONS
This research examined the design of a multi-chambered tank to collect and treat stormwater
runoff from critical urban source areas, including gas stations, oil change facilities, transmission
repair shops, and other auto repair facilities. The collected runoff will first be treated in a catchbasin
chamber where larger particles will be removed by settling. The water would then flow into a main
settling chamber containing oil and grease sorbent material where it will undergo a much longer
treatment period (20 to 70 hours) to remove finer particles and to remove oil residues. In practice,
the MCTT would be utilized as a subterranean unit for space limited sites.
These preliminary results show that the treatment unit is providing substantial reductions in
stormwater toxicants (both in paniculate and filtered phases), organics, and suspended solids.
Slight increases in turbidity and color and about a unit in pH reduction also occurred during the
filtration step. The filter unit appears to be responsible for most of the toxicity reductions. However,
the main settling chamber also resulted in substantial reductions in the dissolved toxicity fraction,
total and dissolved COD, suspended solids, turbidity, and color. The catchbasin/ grit chamber also
showed suspended solids reductions. The use of the MCTT is seen to be capable of reducing a
broad range of stormwater pollutants that have been shown to cause substantial receiving water
problems (4).
REFERENCES
1. Pitt, R., Field, R., Lalor, M., and Brown, M. Urban Stormwater Toxic Pollutants:
Assessment, Sources, and Treatability. Water Environment Research. February 1995.
2. Lager, J. and W. Smith. Catchbasin Technology Overview and Assessment. U.S.
Environmental Protection Agency. 1976.
Pitt, R. and P. Bissonnette. Characterizing and Controlling Urban Runoff through Street and
Sewerage Cleaning. U.S. Environmental Protection Agency. Storm and Combined Sewer
Program, Risk Reduction Engineering Laboratory. EPA/600/S2-85/038. PB 85-186500.
Cincinnati, Ohio, June 1985.
Pitt, R. "Biological Effects of Urban runoff Discharges." In: Urban runoff and Receiving
Water Systems: An Interdisciplinary Analysis of Impact, Monitoring, and Management.
Engineering Foundation and ASCE. Lewis Publishers, Chelsea, Michigan, to be published in
1995.
For More Information: Richard Field
U.S. Environmental Protection Agency
Storm and Combined Sewer Pollution Control Program
2890 Woodbridge Avenue
Edison, New Jersey 08837
Phone: (908) 321-6674
221
-------
GROUNDWATER CONTAMINATION FROM STORMWATER INFILTRATION
Robert Pitt, Shirley Clark and Keith Farmer
Department of Civil and Environmental Engineering, The University of Alabama at Birmingham
University Station, Birmingham, AL 35294-4461
(205) 934-8430
Richard Field and Thomas P. O'Connor
Storm and Sewer Pollution Control Program, U.S. Environmental Protection Agency
2890 Woodbridge Avenue, Edison, NJ 08837
(908) 321-6674
INTRODUCTION
The research summarized here was conducted during the first year of a 3-yr cooperative agreement
(CR819573) to identify and control stormwater toxicants, especially those adversely affecting
groundwater. The purpose of this research effort was to review the groundwater contamination literature
as it relates to stormwater.
Prior to urbanization groundwater is recharged by rainfall-runoff and snowmelt infiltrating through
pervious surfaces including grasslands and woods. This infiltrating water is relatively uncontaminated.
Urbanization, however, reduces the permeable soil surface area through which recharge by infiltration
occurs. This results in much less groundwater recharge and greatly increased surface runoff. In
addition the waters available for recharge carry increased quantities of pollutants. With urbanization,
waters having elevated contaminant concentrations also recharge groundwater including effluent from
domestic septic tanks, wastewater from percolation basins and industrial waste injection wells, infiltrating
stormwater,- and infiltrating water from agricultural irrigation. The areas of main concern that are covered
by this paper are: the source of the pollutants, stormwater constituents having a high potential to
contaminate groundwater, and the treatment necessary for stormwater.
METHODOLOGY
An extensive literature review of stormwater pollutants that have the potential to contaminate
groundwater was collected by searching prominent databases. This paper, a condensation of a larger,
more detailed report (Pitt et al. 1994), addresses the potential groundwater problems associated with
stormwater toxicants and describes how conventional stormwater control practices can reduce these
problems. Potential problem pollutants were identified, based on their mobility through the unsaturated
soil zone above groundwater, their abundance in stormwater, and their treatability before discharge.
This information was used with earlier EPA research results of toxicants in urban runoff sheet flows (Pitt
and Field 1990) to identify the possible sources of these potential problem pollutants.
Recommendations were also made for stormwater infiltration guidelines in different areas and monitoring
that should be conducted to evaluate a specific stormwater for its potential to contaminate groundwater.
RESULTS
Sources of Pollutants. Tables 1 and 2 summarize toxicant concentrations and likely sources or
locations having some of the highest concentrations found during an earlier phase of this EPA-funded
research (Pitt and Field 1990). The detection frequencies for the heavy metals are close to 100% for all
source areas, and the detection frequencies for the organics ranged from about 10% to 25%. Vehicle
service areas had the greatest frequencies and/or quantities of observed organics.
222
-------
TABLE 1. CONCENTRATIONS OF HEAVY METALS IN OBSERVED AREAS
Toxicant Highest Median
Cadmium Vehicle service area runoff
Chromium Landscaped area runoff
Copper Urban receiving water
Lead CSO
Nickel Parking area runoff
Zinc Roof runoff
(ng/i) Highest Observed
8 Street runoff
100 Roof runoff
160 Street runoff
75 Storage area runoff
40 Landscaped area runoff
100 Roofrunoff
220
510
1250
330
130
1580
TABLE 2. MAXIMUM CONCENTRATIONS OF TOXIC ORGANICS FROM OBSERVED SOURCES
Toxicant
Benzo(a)anthracene
Benzo(b)fluorantnene
Benzo(k)fluoranthene
Benzo(a)pyrene
Fluorantnene
Naphthalene
Phenanthrene
Pyrene
Chlordane
Butyl benzyl phthalate
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
1,3-Dichlorobenzene
Concentration Detection Significant Sources
(jig/L) Frequency(%)
60 12 Gasoline, wood preservative
226 17 Gasoline, motor oils
221 17 Gasoline, bitumen, oils
300 17 Asphalt, gasoline, oils
128 23 Oils, gasoline, wood preservative
296 13 Coal tar, gasoline, insecticides
69 10 Oils, gasoline, coal tar
102 19 Oils, gasoline, bitumen, coal tar,
wood preservatives
2 13 Insecticide
128 12 Plasticizer
204 14 Fumigant, solvents, insecticides,
paints, lacquers, varnishes
217 14 Pesticide manufacturing
120 23 Pesticide manufacturing
Potential Contaminates to Grouhdwater. NUTRIENTS. Nitrates are one of the most frequently
encountered contaminants in groundwater (AWWA 1990). Phosphorus contamination has not been as
widespread or as severe as that of nitrogen compounds. Nitrate is highly soluble (> 1 kg/L) and will
stay in solution in the percolation water.
PESTICIDES. Urban pesticide contamination of groundwater can result from municipal and
homeowner use for pest control and the subsequent collection of the pesticide in stormwater runoff.
The greatest pesticide mobility occurs in areas with coarse-grained or sandy soils without a hardpan
layer, and with soils that have low clay and organic matter content and high permeability (Domagalski
and Dubrovsky 1992). Pesticides decompose in soil and water, but the total decomposition time can
range from days to years. In general, pesticides with low water solubilities, high octanol-water
partitioning coefficients, and high carbon partitioning coefficients are less mobile. The slower moving
pesticides that may better sorb to soils, have been recommended for use in areas of groundwater
contamination concern.
OTHER ORGANICS. The most commonly occurring organic compounds found in urban
groundwaters include phthalate esters and phenolic compounds. Polycyclic aromatic hydrocarbons
(PAHs) have also been found in groundwaters near industrial sites. Groundwater contamination from
organics occurs more readily in areas with sandy soils and where the water table is near the land
surface (Troutman etal. 1984).
223
-------
METALS. Studies of recharge basins receiving large metal loads found that most of the heavy
metals are removed either in the basin by sedimentation or in the vadose zone. The order of attenuation
in the vadose zone from infiltrating stormwater is: zinc (most mobile) > lead > cadmium > manganese
> copper > iron > chromium > nickel > aluminum (least mobile) (Harper 1988).
SALTS. Sodium and chloride used for deicing collects in the snowmelt and travels down through
the vadose zone to the groundwater with little attenuation. Salts that are still in the percolation water
after it travels through the vadose zone will contaminate the groundwater (Sabol et al. 1987; and Bouwer
1987). Studies of depth of pollutant penetration in soil have shown that sulfate and potassium
concentrations decrease with depth, whereas sodium, calcium, bicarbonate, and chloride concentrations
increase with depth (Close 1987; Ku and Simmons 1986).
MICROORGANISMS. Viruses have been detected in groundwater where stormwater recharge basins
were located short distances above the aquifer (Vaughn et al. 1978). The factors that affect the survival
of enteric bacteria and viruses in the soil include pH, antagonism from soil microflora, moisture content,
temperature, sunlight, and organic matter (Jansons et al. 1989; and Tim and Mostaghim 1991). The
major bacterial removal mechanisms in soil are straining at the soil surface and at intergrain contacts,
sedimentation, sorption by soil particles, and inactivation.
Treatment of Stormwater. Table 3 summarizes the filterable (dissolved solids) fraction of toxicants
found in storm runoff sheetflows from many urban areas found during an earlier phase of this EPA-
funded research (Pitt and Field 1990). Pollutants that are mostly in filterable forms have a greater
potential of affecting groundwater and are more difficult to control with the use of conventional
stormwater control practices which mostly rely on sedimentation and filtration principles. Fortunately,
most of the storm-flow toxic organics and metals are associated with the nonfilterable (suspended
solids) fraction. Possible exceptions include zinc, fluoranthene, pyrene, and 1,3-dichlorobenzene.
Pollutants in dry-weather storm drainage flows, however, tend to be much more associated with filtered
sample fractions and would not be as readily controlled with the use of sedimentation (Pitt et al. 1994).
TABLE 3. FILTERABLE FRACTIONS OF STORMWATER TOXICANTS FROM SOURCE AREAS
Metals
Cadmium
Chromium
Copper
Iron
Lead
Nickel
Zinc
Filterable (%)
Fraction
20 to 50
<20
Small, amount
<20
Small amount
>50
Organics
Benzo(a)anthracene
Fluoranthene
Naphthalene
Phenanthrene
Pyrene
Chlordane
Butyl benzyl phthalate
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
1,3-Dichlorobenzene
Filterable (%)
Fraction
None found in filtered fraction
65
25
None found in filtered fraction
95
None found in filtered fraction
Irregular
Irregular
None found in filtered fraction
75
Sedimentation is the most significant removal mechanism for particulate-reiated (nonfilterable)
pollutants. Volatilization and photolysis are other important pollutant removal mechanisms in wet-
detention ponds. Biodegradation, biotransformation, and bioaccumulation (into plants and animals) may
also occur in larger and open ponds. Infiltration devices can safely deliver large fractions of the surface
flows to groundwater, if carefully designed and located (EPA 1983). Grass-filter strips may be quite
effective in removing paniculate pollutants from overland flows. The filtering effects of grasses, along
224
-------
with increased infiltration/recharge, reduce the paniculate sediment load from urban landscaped areas.
Grass swales are another type of infiltration device.
CONCLUSIONS AND RECOMMENDATIONS
With a reasonable degree of site-specific design considerations to compensate for soil
characteristics, infiltration may be very effective in controlling both urban runoff quality and quantity
problems (EPA 1983). This strategy encourages infiltration of urban runoff to replace the natural
infiltration capacity lost through urbanization and to use the natural filtering and sorption capacity of soils
to remove pollutants; however, the potential for some types of urban runoff to contaminate groundwater
through infiltration requires some restrictions. Infiltration of urban runoff having potentially high
concentrations of pollutants that may pollute groundwater requires adequate pretreatment or the
diversion of these waters away from infiltration devices. The following general guidelines for the
infiltration of stormwater and other storm drainage effluent are recommended in the absence of
comprehensive site-specific evaluations:
• Divert away from infiltration devices - dry-weather storm drainage effluent (probable high
concentrations of soluble heavy metals, pesticides, and pathogenic microorganisms); combined
sewage overflows (poor water quality with high pathogenic microorganism concentrations and
clogging potential); snowmelt runoff (potential for having high concentrations of soluble salts);
runoff from manufacturing industrial areas (potential for having high concentrations of soluble'
toxicants); and construction site runoff (high suspended solids (sediment) concentrations, which
would quickly clog infiltration devices).
•»; Runoff from other critical source areas (e.g., vehicle service facilities and large parking areas)
should receive adequate pretreatment to eliminate the groundwater contamination potential before
infiltration.
• Runoff from residential areas (the largest component of urban runoff in most cities) is generally the
least polluted urban runoff flow and should be considered for infiltration.
Most past stormwater quality monitoring efforts have not adequately evaluated stormwater's
potential for contaminating groundwater. These are the urban runoff contaminates with the potential to
adversely affect groundwater (with the most prominent and/or analyses recommendations in
parentheses):
nutrients (nitrates); salts (chloride); VOCs (if expected in the runoff [e.g., runoff from manufacturing
industrial or vehicle service areas] could screen for VOCs with purgeable organic carbon analyses)-
pathogens (especially enteroviruses, if possible, along with other pathogens [e.g., Pseudomonas
aeruginosa, Shigella, and pathogenic protozoa]); bromide and total organic carbon (to estimate .
disinfection by-product generation potential, if disinfection by either chlorination or ozone is being
considered); pesticides, in both filterable and total sample components (lindane and ehlordane);
other organics, as filterable and total sample components (1,3 dichlorobenzene, pyrene,
fluoranthene, benzo(a)anthracene, bis(2-ethylhexyl)phthaiate, pentachlorophenol, and
phenanthrene); and heavy metals, as filterable and total sample components (chromium, lead,
nickel, and zinc).
The following urban runoff components can adversely affect infiltration and injection operations:
sodium, calcium, and magnesium (calculate sodium adsorption ratio to predict clogging of clay soils);
and suspended solids (determine the need for sedimentation pretreatment to prevent clogging).
REFERENCES •. .
AWWA (American Water Works Association). Fertilizer Contaminates Nebraska Groundwater. AWWA
Mainstream. 34 (4): 6, 1990.
225
-------
Bouwer, Herman. Effect of Irrigated Agriculture on Groundwater. Jour, of Irrigation and Drainage Eng.
ASCE. 113 (1): 516-535, 1987.
Close, M.E. Effects of Irrigation on Water Quality of a Shallow Unconfined Aquifer. Water Resources
Bulletin. 23 (5): 793-802, 1987.
Domagalski, J. L and Dubrovsky, N. M. Pesticide Residues in Groundwater of the San Joaquin Valley,
California. Journal of Hydrology. 130 (1-4): 299-338,1992.
EPA. Results of the Nationwide Urban Runoff Program. NTIS No. PB 84-185552, U.S. Environmental
Protection Agency, Water Planning Division, Washington, D.C., December 1983.,
Harper, Harvey H. Effects of Stormwater Management Systems on Groundwater Quality. Final Report for
DER Project WM190. Florida Department of Environmental Regulation, 1988.
Jansons, J., Edmonds, L. W., Speight, B. and Bucens, M. R. Survival of Viruses in Groundwater. Water
Research. 23 (3): 301-306,1989.
Ku, H. F. H. and Simmons, D. L Effect of Urban Stormwater Runoff on Groundwater beneath Recharge
Basins on Long Island, New York. U.S. Geological Survey (USGS) Water Resources Investigations
Report 85-4088. USGS, Denver, Colorado, 1986.
Pitt, R., and Field, R. Hazardous and Toxic Wastes Associated with Urban Stormwater Runoff. 'In:
Proceedings of the 16th annual RREL Hazardous Waste Research Symposium: Remedial Action,
Treatment and Disposal of Hazardous Waste. U.S EPA, Cincinnati, OH, EPA/600/9-90-37,1990.
Pitt, R., Clark, S. and Farmer, K. Potential Groundwater Contamination From Intentional and Non-
Intentional Stormwater Infiltration. EPA/600/SR-94/129. U.S. EPA, Cincinnati, Ohio, 1994.
Sabol, G. V,, Bouwer, H. and Wierenga, P. J. Irrigation Effects in Arizona and New Mexico. Journal of
Irrigation and Drainage Engineering, ASCE. 113 (1): 30-57,1987.
Tim, U. S. arid Mostaghim, S. Model for Predicting Virus Movement through Soil. Groundwater. 29 (2):
251-259, 1991.
Troutman, D.E., Godsy, E. M., Goerlitz, D. F. and Ehrlich, G. G. Phenolic Contamination in the Sand-and-
Gravel Aquifer from a Surface Impoundment of Wood Treatment Wastewaters, Pensacola, Florida. USGS
Water-Resources Investigations Report 84-4230. USGS, Denver, Colorado, 1984.
Vaughn, J.M., Landry, E.F., Baranosky, L.J., Beckwith, C. A., Dahl, M. C. and Delihas, N.C. Survey of
Human Virus Occurrence in Wastewater Recharged Groundwater on Long Island. Applied and
Environmental Microbiology. 36 (1): 47-51,1978.
FOR MORE INFORMATION: Richard Field, EPA Project Officer
Chief, Storm and Combined Sewer Pollution Control Program
U.S. Environmental Protection Agency (MS-106)
Edison, NJ 08837-3679
(908) 321 - 6674
226
-------
COLLOIDAL CONTAMINANTS IN URBAN RUNOFF
M.R. Wiesner, G. Characklis, and D. Brejchovd
Department of Environmental Science and Engineering
Rice University
P.O. Box 1892
Houston,Texas 77251
Phone:713-285-5129
INTRODUCTION
Rivers typically account for the largest single sediment input to lacustrine and estuarine systems.
Pesticides, metals, chlorinated hydrocarbons and other contaminants sorbed to particles carried by rivers
may be transported to the sediments of lakes and estuaries. Urban runoff is often an important non-point
source of particles to surface waters. For example, under average conditions, over 40% of the suspended
solids entering Galveston Bay are estimated to originate from high density urban and residential non-point
sources and, in some cases, the fractions of specific contaminants entering Galveston Bay associated
with these particles (such as oil and grease) approach 100%1.
Information on the size distribution of particles in urban runoff has been surprisingly limited despite
its importance in providing a strategy for pollution control. While many studies have been done on urban
runoff, in only a few instances has the role of specific particle fractions as transport vectors for pollutants
in urban runoff been considered.
In this paper,,we summarize field data on particle size distributions and contaminant concentrations
in urban runoff and streams obtained over a 2-year period. Much of the previous research done in this
area has concentrated on particles, larger thanIO p,m in diameter. While larger particles may represent a
significant fraction of suspended materials, we show that the majority of the particles and particle surface
area in runoff from one urban center resides within the smaller size fractions.
Inextricably linked to the problem of pollutant transport is the question of whether or not there is
significant aggregation of particles in the runoff stream. The degree of aggregation that occurs in the
runoff stream will directly affect the fate and transport of the particles, and thus the pollutants sorbed to
them. For example, if sub-micron particles aggregate significantly in the runoff stream, particle settling
rates will be accelerated in lakes and estuaries, as will the transport of particles and associated
contaminants to the sediments.
METHODOLOGY
An urban waterway (Brays Bayou) and several of its tributaries were sampled over a two year
period. Brays Bayou drains areas of Fort Bend and Harris counties in Texas within the Houston
metropolitan area. The watershed selected as a study area for this research begins at the start of Brays
Bayou in west Houston and continues down through southwest Houston until the Bayou reaches South
Main St. This portion of the watershed equates to a drainage area of approximately 92 sq. mi. The
majority of the watershed is under stress from a wide variety of urban uses consisting mostly of residential
and commercial properties. Research was concentrated primarily on four sampling sites. The sites are,
from west to east; 1) Keegan's Bayou at Roark Rd., 2) Brays Bayou at Gessner, 3) Poor Farm Ditch,
and 4) Brays Bayou at Main Street.
Samples were obtained by dropping a 12-quart, detergent cleaned, polyethylene bucket into the
center of the runoff stream from overpasses spanning the Bayou. The bucket containing the sample was
retrieved by means of a rope attached to the handle. Approximately one liter of sample was retained in
two 500 ml polyethylene bottles. Each bottle had been washed with Alconox® cleaner, rinsed three times
with distilled water and then given three final rinses with Milli-Q®, deionized water. Samples were
transported back to the lab for analysis. Surface water samples were obtained under both storm and
ambient conditions. Upon returning to the lab, all samples were analyzed for turbidity, pH, particle size
distribution, total suspended solids (TSS), total organic carbon (TOG), and a suite of metals. Particle size
analyses were performed using an electronic particle counter (Coulter Electronics Ltd., Luton, England).
The Coulter Multisizer operates on the electronic sensing zone (ESZ) principle. The effective
measurement range for such devices usually varies from 2-60% of the orifice diameter. A 19u.m orifice
tube was used for these experiments with an effective measurement range of approximately 0.5 to 15 |j.m.
Samples were also filtered through various size membranes in order to analyze the composition of the
227
-------
filtrates in each size fraction. The filtration system consisted of four 250 ml vacuum filter holders
(Nalgene, Prod. 315-0047) mounted on 1000 ml Erlenmeyer flasks. Each flask was connected to a
vacuum pump (Cast, No. 1HAB25BM100X) by means of Tygon® (R-3603) tubing. The cutoffs for
membranes used in the fractionation procedure were reported by the manufacturer to be 0.45 urn, 0.1 urn
and 105 daltons. All of these membranes are cellulose nitrate filters (Sartorious AG, Goettingen,
Germany, 47 mm circular).
Membrane-fractionated samples were analyzed for suspended solids, TOG, and metals. All
determinations of suspended solids were made according to procedure 2540D in Standard Methods2. In
the best cases, solids measurements yielded data on the mass of particles in the size ranges >20 urn,
0.45 to 20 urn, 0.1 to 0.45 u.m and 105 daltons to 0.1 |im. Unfortunately, it was often difficult to obtain
data for the smallest size range, as the 105 dalton filter would frequently foul to the extent that filtration
was not possible.
In preparation for metals analysis, each filtrate was concentrated through a Nitric Acid-Hydrochloric
Acid Digestion, method 3030F in Standard Methods2. Digestion resulted in concentrating the samples 5-
10 times. Samples were stored in detergent washed, acid rinsed glass vials and refrigerated at 4"C until
analysis. Metal concentrations (Fe, Sr, Ba, Pb, Zn, Ca, Cu and Mn) were measured by Inductively
Coupled Plasma (ICP) -Atomic Emission Spectrometry (Perkin-Elmer).
Samples were also obtained for the purpose of evaluating particle aggregation under controlled
conditions. In aggregation experiments, "background" or storm water was transported back to the
laboratory, and the largest particles were allowed to settle from suspension. The supernatant was then
transferred to a 2-liter beaker and mixed at a prescribed speed. Mixing conditions as a function of paddle
rotational speed were known for this system. An rpm was selected toprovide the mean velocity gradient,
G, of approximately 70 per second estimated to occur in the bayou under stormwater flow conditions.
Particle size distributions were measured at regular time intervals to evaluate the kinetics of particle
aggregation in the system and to quantify the collision efficiency factor, a.
Aggregation was further investigated by following a plug of water in Brays Bayou from the west side
of Houston through the Houston Ship Channel. During this phase of the project, three more sampling
sites were added to the four used for stormwater sampling ; Brays Bayou at Old Spanish Trail, Brays
Bayou at the Houston Ship Channel, and the Houston Ship Channel at the San Jacinto Monument. The
Houston Ship Channel represented the last leg of stormwater travel from Houston to the Galveston Bay.
It is an unconcreted, man-made channel approximately a half mile wide by the time it reaches the San
Jacinto Monument. Particles from Brays Bayou have been significantly obscured by dilution and
subsequent industrial additions by the time they reach this point. There is also a significant increase in
salinity at this sampling location due to its proximity to the Bay. These samples were analyzed for pH,
turbidity, particle size distribution, fractional organic carbon and fractional metals in the same manner as
those measured during the stormwater phase of this work.
RESULTS
Particle size data show that more than 90% of the particle number exists between 0.45(im and 2urn.
Photon Correlation Spectroscopy measurements also show a peak in the distribution varying from 0.4|J.m
to 0.7nm. Particle size distributions in runoff were taken under both storm and ambient conditions.
Although the profiles under both sets of conditions were similar, the concentration of particles in the storm
samples was much higher, both in particle number and calculated particle surface area ( Figure 1).
These higher concentrations, coupled with much higher flow rates in the runoff stream during a storm
event, produce high particle loadings during storm events. ':
TOC analysis indicated that the paniculate phase was composed of 1 to 1,0% organic carbon.
Functional groups on particulate organic carbon may complex metals and enhance the transport of metals
with particles. Under storm conditions TOC concentrations increased as the bayou water progressed
downstream. TOC tended to be bimodally distributed between the largest and smallest (dissolved) size
fractions (Figure 2). Thus, metals readily complexed by TOC would also be expected to exhibit a bimodal
distribution with respect to size fraction.
Metals analysis of the samples yielded detectable levels of Fe, Ba, Cu, Mn, Pb and Zn. The metals
concentrations in the runoff, expressed per mass of suspended solids, were comparable with
concentrations reported in previous studies evaluating the metals content of Galveston Bay sediments.
With regards to all metals analyzed, except Zn, there was good correlation between the profiles of the
runoff samples and the Bay sediments. While elevated concentrations of zinc have been reported in the
228
-------
1.4x108
-0.4 -0.2
0.2 0.4 0.6 0.8 1 1.2
log dp
Avg Backgrd (n=3)
Avg Storm (n=4)
—e— Background S.A.
—ft— Storm S.A.
Figure 1. Particle size distributions at Brays Bayou at Gessner averaged over 3 sampling dates for
background conditions and 4 separate storm events.
Storm Avg (n=4)
Backgrd Avg (n=3)
-------
Figure 3. Variations in average zinc concentrations during storm and background conditions at the
four sampling sites.
Intensive sampling of individual storm events produced a similar picture of the materials loadings to
this urban waterway. Different fractions of materials were "eluted" to the bayou at different times during
storm events. For example, larger particles tended to peak in concentration earlier during a storm than
did smaller particles. This is reflected in the data summarizing settleable (within 15 minutes) and
unsettleable suspended solids as compared with particle number concentrations (Figure 4). Similarly,
metals tended to peak in concentrations at different times. Iron concentrations peaked at approximately
the same time as the unsettleable solids (4 to 8 hours) and remained above background concentrations
for at least 24 hours after the onset of the storm. In contrast, zinc concentrations peaked at 4 hours and
then decreased to, or below, background concentration. Although zinc and TOC concentrations
correlated significantly (r2 = 0.8) TOC concentrations were observed to remain above background levels
for at least 24 hours after the onset of the storm events monitored.
Laboratory simulations of mixing and particle aggregation in Brays Bayou indicated that significant
aggregation is likely to occur in runoff water within 1-3 hrs. The number concentration of particles
measured by the particle analyzer with a 19 jim orifice increased consistently to a maximum in the first
few hours of mixing. Particle number then displayed a drop back down to approximately the original
concentration, eventually settling into a slow, steady decline. After 24 hours of continuous mixing in the
lab, large aggregates of particles were readily visible. The original increase in particle number is
attributed to the aggregation of particles which are initially below the 0.45p.m threshold of the particle
analyzer. As these particles reach the threshold there is a rise in the number of measurable particles.
However, a monotonically decreasing number concentration was observed in aggregation experiments
when particle concentrations were measured using a 50 |im orifice. Because this orifice provides a
window on a slightly larger fraction of particles, the effect of aggregation into the smaller size classes
observed with the 19 jim orifice is masked. A least-squares fit to a semi-log plot of particle concentration
yields values of the collision efficiency factor on the order of 10'2 (Figure 5).
Although it is evident from laboratory mixing studies that significant aggregation of particles entering
the bayou is probable, aggregation was not evidenced in samples obtained following a a single plug of
water down the bayou; particle size distributions remained largely unchanged with distance. Aggregation
of particles present in the upper reaches of the bayou may be masked by the introduction of particles from
the numerous small tributaries that feed into Brays Bayou as it makes its way through the city.
CONCLUSIONS
The number and surface area concentrations of particles in the urban runoff studied were dominated
by particles 2 to 5 urn in diameter. Zinc and barium were observed to be distributed bimodally with
respect to the size fractions where they predominate, while lead and iron tended to associate almost
exclusively with the largest size fraction. Most metals are associated with size fractions either larger than
0.45 Jim or less than 105 daltons. This bimodal distribution of metals is similar to that observed for total
organic carbon. Materials in different size classes were observed to reach a maximum in concentration at
230
-------
V
different times during a storm, larger particles reaching their maximum concentration
particles.
300-1 -&• — r 3.0x108
before smaller
Backgrd.. Settleable
Backgrd _ .Unsettjeabje
-B— SettSS
—&— UnsettSS
Figure 4. Comparison of suspended solids and particle number concentrations in Brays Bayou at
Main Street on 27 August 1993.
0.0 f
-0.1
-0.2
-0.3
y = - 2.1211e-2 - 1.1691e-3x RA2 = 0.835
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Time (min)
Figure 5. Log particle number concentration (normalized to initial concentration) in Brays Bayou at
Main Street during storm event on 12 May 1994.
LITERATURE CITED
1. Newell, C. J., Rifai, H. S. and Bedient, P. B. Characterization of Non-Point Sources and Loadings to
Galveston Bay. Galveston Bay National Estuary Program, (1991).
2. Clesceri, L. S., Greenberg, A. E. and Trussell, R. R. Standard Methods for the Examination of Water
and Wastewater. (1989).
231
-------
MODELLING AIR EMISSIONS FROM CONTAMINATED DREDGED MATERIALS
L J Thibodeaux1, K T Valsaraj1, D D Reible1 and J M Brannon2
Department of chemical engineering,
Louisiana State University, Baton Rouge, LA 70803
2U S Army Waterways Experiment Station,
Vicksburg, MS
INTRODUCTION:
A confined disposal facility (CDF) is a diked area for gravity separation and dredged
material solids. When contaminated dredged material is placed in a CDF, the potential exists for
volatile organic chemicals (VOCs) associated with the sediment to be released to the air.
Sediments from the New Bedford Harbor (NBH) Superfund Site, MA. contain significant amounts of
polych^rinated biphenyls (PCBs), some of which may be released to the air during evaporative
drying in a CDF.
Models for evaluation of volatile emissions to air during dredged material disposal have
been developed (Thibodeaux, 1989). These theoretical models may be applied to calculate
potential PCB emissions from CDFs proposed for containment of NBH sediment. Four locales
associated with a CDF operation were identified as separate volatile sources. These locales
were: (1) the sediment (dredged material) relocation locale, (2) the exposed sediment locale,
(3) the ponded sediment locale and, (4) the vegetation-covered sediment locale. The exposed
sediment locale- was ranked the highest. Field or laboratory emission data suitable for
comparison to model predictions were not available for any of the locales. Brannon (1989)
reported some preliminary data for locale 2 for the emission of Aroclor-1242 from a drying
sediment exposed to air under laboratory conditions. This paper compares the experimental values
against theoretical predictions.
METHODOLOGY:
A series of four experiments were performed with an emission isolation flux chamber. This
apparatus was placed atop the sediment to collect the PCB vapors as air was passed over the
sediment in the chamber for 1 hour. The vapors were trapped in fluorisil tubes and analyzed by
gas chromatography. Two adsorption tubes, and an air rate of 900 cnr/min were used. In general
the experiments can be placed into two classes, wet and dry. The first three experiments were
performed with sediment saturated with water; these are called wet sediment experiments. The
last experiment showed significant cracking and drying of the surface sediment. The soil
porosity decreased from 0.774 on day 1 to 0.103 on day 10, while the solid fraction increased
from 0.33 to 0.88 during this period. The laboratory was maintained at 20±0.3 °C and low relative
humidity throughout the experiment. The experiment number 4 shall be termed the dry sediment
experiment. . •
The following is a brief presentation of the model that applies for the exposed sediment
locale. Contaminated sediment that is wet and exposed directly to air results in the highest
VOC emission rates. VOCs sorbed on the particles at the soil surface have a relatively short
pathway to the air. The top layers eventually become depleted of the chemical. Continuing
losses will come from within the soil pores. The following equation applies for this transient
volatilization rate
r 7i t 1
D f 1 + Pb }
a J?"*
1/2 -]_
^A
(1)
is the air-filled porosity (m3.nT3), ka is the air-side mass transfer coefficient (rn.lr1), and Ca
is the chemical concentration in the background air (/jg.nr). The above equation is for an
232
-------
unsaturated (with respect to water) or dry sediment. As the sediment looses its water content
by evaporation and drainage, the air-soil equilibrium partition constant changes. Three water
content regimes exist and require as many partition constants. These are wet, damp and dry. Each
can be expressed by the simple equilibrium expression, C& =K*WA. The partition constant
for the wet case is the ratio of Henry's constant to the soil-water partition constant,
* , ,
PA A where Kd is the
H
K* = —-, whereas for the dry case at soil loadings it is
K* =
W
soil-water equilibrium constant (n^.g'1), p
adsorption
constant
on the sediment for a monolayer coverage
er equilibrium constant (nr.g'1), pA is the chemical vapor pressure (atm), B, is the B
ion constant (dimensionless), M, is the chemical molecular weight (jug/mol"1), R is the g
t (= 82 x 1CT6 atm.m3.moT1.K-1), T is the temperature (K) and, W.* is the chemical loadi
BET
gas
loading
RESULTS:
In the section in which the experimental data was reviewed it. was noted that the
experiments fall into two classes. These were the initial period with the wet sediment and the
long-term run with the dry sediment. Equation (1) applies for both cases with the appropriate
wet and dry equilibrium' expressions given earlier. '
During the initial phase of the experiment evaporation time is short (i.e, t -.0) and the
sediment surface is wet. If the incoming air is also PCB free Equation (1) simplifies to
. *A=kAWAK* • (2)
where K* is given by the equation for the wet case. At 25°C it is 0.132 g.m"3. The air side mass
transfer coefficient can be estimated using four different methods (Thibodeaux and Scott 1985).
The estimates for Arpclor 1242 at 25°C range from 30 to 96 cm.h"1. Based on the range of
coefficients the predicted emission rates for A-1242 with a sediment loading of 887 /ug.g"1 ranges
from 36 to 113 /vg.nr.h . The experimental values ranged from 12 to 62 ^/g.m^.h"1. These values
are shown in Figure 1. The fact that the predicted and experimental ranges substantially overlap
each other suggests that the wet soil portion of the model is correct.
By the third day of the experiment the sediment was essentially dry. In this case, time
is large and Equation (10 applies as written. The air-sediment partition constant should reflect
the dry state ,df the sediment.
A procedure based on the theory of gas mixtures adsorbing competitively on a solid surface
that extends the classical Brauner-Emmett-Teller (BET) model to account for water vapor was used
to estimate K* for Aroclor 1242 on dry sediment (Valsaraj and Thibodeaux, 1988). A key factor
in the procedure is the surface area of the soil. In the case this was determined from the
fractional organic matter and clay of the sediment. Together with the sediment surface area S
(rrr'.g"1) so determined the monolayer coverage of Aroclbr 1242 on the sediment ;WA* was computed.
Other relevant parameters used were: B\ = 20, 18.4% of S active for adsorption and Aroclor 1242
density was 1.5 g.cm"3. All other Aroclor 1242 properties were from Thibodeaux (1989) / Two
estimates of K for the dry case were obtained; one for the estuarine composite sediment 0.056
kg.I"1 and the other for the sediment from the hot spot 0.016 kg.I"1.
The necessary information is available in Equation (1) to arrive at predicted values of
the flux. The working form of Equation (1) is
887 K*
0.097(tJT*)1/2 + 3.29
(3)
with t in days, K* in g.m"3 and NAin //g.m^.h"1. The value of kA was 0.304 m.h"1, DAa was 0.035 6.
in cm .s , the soil bulk density was 0.69 g.cm" and porosity ea was 0.774'.
4/3
The predicted values for Aroclor 1242 flux assuming dry soil conditions appear in Figure
1. The experimental values are also given alongside. Comparison of predicted and measured flux
values should be done for t * 3 days, since the sediment was wet or damp before this time. The
average measured fluxes for days 3 through 10 is 0.885 ± 0.432 //g.m~2.h , whereas the predicted
233
-------
FIGURE 1. AROCLOR 1242 VAPORIZATION RATE FROM
DREDGED SEDIMENT
100
< 50
'0
15
UJ
cc
z
o
H
N
QC 10
O
Q.
5
o
8-
K = O.I32g/m3
Initial Period Experiments
o data
• model
K=.056g/m3
Long-Term Experiments
o data
model
K = .OI5g/m3
.04
K =.0074 g/m3
468
TIME , t (days)
10
12
Figure 1. Aroclor 1242 vaporization rate from dredged sediment.
fluxes were 15 and 4 jug.m^.h"1 for the two calculated values of K*. At best the predicted values
are 4.7 times larger than the measured values. The model predicts a very weak time dependence
on the rate. At day 10 the soil-side mass tranfer resistance accounts for only 1.1% of the
transport resistance for the organic. The measured values also seem to display a time
dependence. This behavior may be due to several factors including cracking and other soil
porosity changes with time. However, it appears that K is the primary factpr that controls the
flux in this set of measurements. A value of 0.0074 g.m"3 in the model equation yields predicted
fluxes in line with the experimental values. The model predicted and measured values appear in
Figure 1. An overall mass balance indicates that 2.32 g of Aroclor 1242 were present initially
in the isolation flux chamber. Using 1.7 /^g.m^.h"1 as the average evaporation rate yields 59 ug
lost from the sediment in 10 days. This is 0.0026%, so the bulk of the original Aroclor 1242
remained in the soil and did not evaporate.
CONCLUSIONS:
The theoretical model does a satisfactory job in predicting the Aroclor 1242 emission rate
from exposed sediment. The measured values for wet sediment averaged 26 ± 13 /vg.m .h . The
model yielded values of 36 to 113 A/g.m .h depending on the kA values used. The measured values
for the dry sediment averaged 0.855 ± 0.432 jug.m'2.h. The most sensitive parameter in the model
was K*. the air-sediment partition constant. In general, the theoretical model overpredicts the
measured values by a factor of 1.4 to 18.
Additional experimental measurements are needed to reduce this range of uncertainty and
explore the effects of cracking. These experiments should involve re-wetting the soil in order
to observe any increase in flux. Field measurements at a CDF site should be performed with the
234
-------
isolation flux chamber or a similar method. Laboratory experiments should be performed and the
air-sediment partition constant K* measured for damp and dry conditions. This critical parameter
has a .very weak data base (Valsaraj and Thibodeaux, 1988).
REFERENCES:
Thibodeaux L J (1988): Theoretical Models for Evaluation of Volatile Emissions to Air During
Dredged Material Disposal with Applications to New Bedford Harbor, Massachussetts,
Miscellaneous Paper EL-89-3, US Army Engineer Waterways Experiment; Vicksburg, MS.
Brannon, J M (1989): Laboratory Assessment of Volatilization from New Bedford Harbor Sediment,
Memorandum for Record, WESES-A, US Army Engineer Waterways Experiment Station. Vicksburg,
MS.
Thibodeaux L J and H D Scott (1986), Chapter 4 in Environmental Exposure from Chemicals. Volume
1. W B Neely and G E Blau (editors), CRC Press, Boca Raton, FL.
Valsaraj K T and L J Thibodeaux (1988), J. Haz. Materials, 19: 1979 - 1989.
ACKNOWLEDGEMENT: ' *
This work was supported by a grant from the US Environmental Protection Agency (Grant No:
R 819165-01) to the Hazardous Substance Research Center (South and Southwest). Dr. Dale Manty
is the Project Officer.
235
-------
BIOREMEDIATION OF CONTAMINATED SEDIMENTS
J B. Hughes, V. Jee, and C. H. Ward
Department of Environmental Science and Engineering
P.O. Box 1892
Rice University
Houston, TX 77251
INTRODUCTION
Contaminants in bottom sediments have historically been considered to have minimal
environmental impact because they are buried, sorbed or electrostatically bound to clay particles, or
incorporated into humus. Physical and chemical conditions such as alkalinity, pH, and redox of the
sediments also play a part in sequestering contaminants. (2,3) As long as the sediments are
undisturbed, the contaminants are considered stabilized and not an immediate environmental problem.
Resuspension of bottom sediments makes contaminants more available for dispersal into the marine
environment. Events that can cause resuspension include storm surges, construction activity, and
dredging.
A conceptual model of the effect of a resuspension event on contaminated sediments is shown
in Rgure 1. During resuspension, sediment particles move from an anaerobic to aerobic environment,
changing their redox characteristics, and allowing the indigenous aerobic bacteria to grow and utilize
certain classes of contaminants as energy sources. The contaminants are also more available for use
because the mixing energy imparted to the particles during resuspension enhances mass transfer,
allowing contaminants to enter the aqueous phase more rapidly.
The contaminants targeted in this research are polynuclear aromatic hydrocarbons (PAHs), a
class of contaminant commonly found in bottom sediments near highly industrialized areas. A major
source of PAH contamination is the combustion of fossil fuels. (8,5) Other sources are industrial
wastewater effluents, petroleum spills and oilfield produced brine disposal into waterways. PAHs
consist of multiple benzene rings fused in linear, angular, and cluster arrangements. PAHs are non-
ionic and hydrophobic and as such, are lipid soluble and can be bioaccumulated in the food chain.
Exposure to polycyclic compounds has also been associated with higher risks for cancer.
Previous research has shown that PAHs can be biodegraded. Size and structure, i.e., number
and configuration of condensed rings, can affect compound disappearance. (4,7) PAHs of up to three
condensed rings have been shown to serve as growth substrates. Compounds with more than three
rings may be subject to cometabolic degradation.(l) Probably the most important parameter affecting
blodegradation is the redox character of the sediment or sediment/water system. Studies of sediment
slurries under various redox conditions have shown that rates of compound disappearance and
mineralization increase with increasing redox potential. (3,6)
The focus of this research was to examine the relationship between resuspension and
blodegradation of PAHs in lab scale slurry reactors. The rate and extent of contaminant release from
the sediments into an uncontaminated water column was determined. Oxygen demand of initially
anaerobic sediments were investigated. Then rate and extent of phenanthrene biodegradation was
examined. Rnal partitioning of the phenanthrene, after the degradation test, was determined based on
mass balance calculations made on the radiolabelecl carbon in the tracer. Several factors which may
influence the design or operation of bioreactors used for remediation of contaminated sediments were
also evaluated.
METHODOLOGY
Sediments used in these experiments were collected from Dickinson Bayou, Galveston County,
Texas. The natural environment is esutarine with total dissolved solids (TDS) of the water measuring
236
-------
5,000 parts per million (ppm). An artificial seawater (ASW) medium based on Difco marine broth
without the organics and diluted to the proper IDS was used in all tests. The sediments were
artificially contaminated with phenanthrene (a 3-ring PAH). Special care was taken to keep the
sediments anaerobic, with all handling done in a nitrogen atmosphere glove box.
Concentration in the various system compartments - liquid, solid, and vapor was monitored
using a Hewlett Packard HP 5890 gas chromatograph with photo-ionization detector (GC-PID).
Sediments were extracted using a modification of EPA Method 3550 (Sonication) with dichloromethane
(DCM) as the solvent. Liquids were extracted by shaking with DCM. The sorbent of the ORBO
hydrocarbon vapor traps were also extracted by shaking in DCM. 14C labeled, phenanthrene was
added to trace mineralization in the biodegradation tests. Radiation levels were measured using a
Beckman Model IS 3801 Scintillation counter.
The slurry reactor used in the experiments is shown in Figure 2. CO2 traps were used in the
biodegradation tests. The traps were removed for the abiotic contaminant release experiments.
Sediments and artificial seawater (ASW) medium were loaded into the reactors inside the glove box.
Three solids loadings were tested - 5, 10, and 15% dry weight sediment. ASW used was purged for 30
minutes with N2 gas before being brought into the anaerobic environment of the glove box.
The first experimental series determined the rate and extent of phenanthrene release from the
sediments to uncontaminated liquid under various conditions of mixing and aeration, these were short
term tests (48 hours) with an emphasis on early time measurements. Each solids loading was tested
under three mixing conditions - (1) unmixed, (2) continuous mixing, and (3) initial mixing only, both with
and without aeration.
Mineralization to CO2 was tracked in the biodegradation tests using 14C tracer and dual CO2
traps containing 1N KOH. All reactors in this test series were continuously aerated. At the end of the 7
day biodegradation tests, the reactors were poisoned with mercuric chloride and samples of each
system compartment (solid, liquid, and vapor trap) was extracted. A mass balance approach was
utilized to determine final partitioning of carbon from degraded phenanthrene. Non-solvent extractable
carbon was measured in the sediment fraction using a modified chemical oxygen demand (COD) test.
Output gases from the oxidation were passed through dual CO2 traps and the residual radiation
measured by scintillation counting.
Additional tests were performed to evaluate the effect of various parameters on rate and extent
of phenanthrene mineralization. Factors examined included intermittent aeration, nutrient amendment,
and bioaugmentation. The objectives of these tests were to reduce volatilization losses and shorten
degradation time.
RESULTS
Contaminant release into the aqueous phase of sediment/water slurries was rapid in mixed
reactors. Solids loading has no effect on rate of contaminant release and little effect on extent with
mixing. All reactors reached 80 to 90% of theoretical aqueous concentration within 30 minutes.
Mixing, itself, had the greatest effect on rate and extent. Rate of contaminant release was 2 to 4 times
slower for unmixed systems. Maximum extent of release was also less than half that seen in mixed
reactors. This is shown by Figure 3 which shows contaminant release in reactors loaded with 10%
solids.
Aeration effects appear to be masked by mixing effects. No differences in release rates were
seen between mixed reactors which were aerated and those that were not. In unmixed reactors,
aerated systems showed faster rates of contaminant release than unaerated systems, indicating that
the aeration process itself provides some degree of mixing.
237
-------
Input of energy resuspends
contaminated sediments
Aeration of particle
PAH
desorbs
changes redox
^
* k.
/
* aerobic
/
Increases potential
for aerobic degradation
>»«
« •
PAH degraders present,
Oo limited, PAH sorbed
/
anaerobic
Figure 1. Conceptural model of the resuspension of process and the effect it has on the ability
of indigenous bacteria to biodegrade polynuclear aromatic hydrocarbons (PAHs).
Air
Inlet
Sediments
CO2 traps
Air
Volatile HC
trap
300ml
slurry reactor
Magnetic stirrer
Figure 2. Slurry reactor used in experiments. CO2 traps are included only in biodegradation
tests where radiolabeled CO2 is a byproduct of phenanthrene mineralization.
238
-------
Several observations can be made about the biodegradation experiments. Volatilization over
the 7 day test period was considerable. A two (2) day lag period was seen before significant
mineralization occurred in all reactors. Maximum extent of mineralization appeared to be reached
within five (5) days in mixed systems. The effect of mixing on mineralization is clearly evident. Figure
4 shows mineralization in reactors loaded with 10% solids. Significantly higher mineralization was seen
in reactors which were mixed.
Mass balance calculations were performed based on tracking of the radiolabeled carbon. Table
1 shows the mass balance for mixed reactors at the end of the biodegradation test. Analysis of the
mass balance partitioning in relationship to expected partitioning based on stoichiometry indicates that
the phenanthrene was being used as a growth substrate by the indigenous bacteria.
The amount of phenanthrene lost to volatilization is much higher than might be expected from
low vapor pressure (6.8 X10"4 torr.). Therefore the effect of intermittent aeration on mineralization and
stripping was tested. The test results show that not only does mineralization still occur without
continuous aeration, but also that the final extent of mineralization is higher for the intermittently aerated
reactor.
Reactors were amended with various amounts of fertilizer to determine if the low nutrient
(nitrogen and phosphorus) content of the ASW could be causing the lag period seen before
mineralization. Nutrient amendment had no effect on rate and extent of mineralization. Augmenting the
slurry reactor with an aged slurry (2 - 3. days), however, did eliminate the lag time.
CONCLUSIONS
It can be concluded from this series of experiments that resuspension of anaerobic sediments
can affect the degradation of phenanthrene sorbed to them. Sediments with contamination levels of 50
ppm phenanthrene were remediated in lab scale slurry reactors to the point that only trace
phenanthrene was found after 5 days. Mixing and aeration, natural byproducts of the resuspension
process, were the only treatment used. Due to the time required to see complete degradation, it is
unlikely that the mixing and aeration provided solely by dredging can be considered a remedial
treatment.
The lab scale slurry reactors used in this project show that this type of reactor holds promise as
a potential remediation methodology. A slurry reactor can be defined as an enclosed system where
sediments and water are maintained in a homogeneous slurry over a fixed period of time. It is
envisioned that the reactors will be a batch treatment process. As such its size will be limited.
Ex-situ reactors for sediment remediation could be built on barges or on shore. Both the
bottom sediments and a liquid phase would have to be moved to the reactor to form the slurry. If
hydraulic dredges are used, the sediments are removed in slurry form and could be pumped directly to
the reactors. In-situ reactors would isolate small areas of the contaminated bottom and utilize the
overlying water to make the slurry. Potential methods of isolation include caissons, sheet piling, or
other types of physical barriers.
Design factors that should be considered include mixing intensity, aeration, and the use of
sequential treatment cells. The lab scale reactors used had minimal to insufficient mixing capacity.
Any field or full scale test should be designed to provide adequate mixing for the desired slurry. Mixing
not only enhances mass transfer of the contaminant from the sediment to the aqueous phase, but also
helps to maintain oxygenation of the slurry when open to the atmosphere. Optimum design will utilize
intermittent aeration or some form of chemical oxidation, for example addition of peroxide. The purpose
will be to minimize contaminant loss due to volatilization or stripping.
Since it was shown that augmenting the reactor with aged slurry eliminated the lag period, an
239
-------
o
c
o
o
w
3
O
0)
3
CP
140
120-
100-
Cont Mix Aerated
Cont Mix Unaerated
Initial Mix Aerated
D ~ Unmix Aerated
Unmix unaerated
— 80-
£
o
Q)
0)
Q.
3 4
Time, hrs
Figure 3. The effect of mixing and aeration on contaminant release in slurry reactors loaded
with 10% solids. Error bars represent one standard deviation.
C
O
*£*
(0
-- *
^
345
Time, days
^£^
6
8
Figure 4. Phenanthrene mineralization in slurry reactors loaded with 10% solids. Error bars
represent one standard deviation. Curves with symbols are for the abiotic controls.
240
-------
overall treatment scheme utilizing some form of sequential treatment seems prudent. Treatment times
were reduced by 40% when the reactors were seeded, going from 5 to 3 days to reach maximum
extent of mineralization. Ex-situ treatment would require two reactors. Slurry from the initial reactor
would be used to seed the second reactor which then would be used to seed the first reactor, etc. jrv
situ treatment cells could be built so that each successive cell would be seeded by the previously built
cell. Removable barriers for only two cells would be required. Cell construction could be timed so that
the materials from the first cell would be moved to build the third cell, etc.
From the research conducted, ft appears that engineering intervention would be most effective
in treating small areas of highly contaminated sediments before dredging. This assumes that in-situ
reactors would be more cost effective than ex-situ reactors. Since batch biological treatment requires,
at minimum, 3 days, an in-situ approach also eliminates problems associated with sediment storage
that would arise if the sediment were treated after dredging.
REFERENCES
1. Cerniglia, C. E. and Heitkamp, M. A. 1989. Microbial Degradation of Polycyelic Aromatic
Hydrocarbon (PAH) in the Aquatic Environment. Metabolism of Polycylic Aromatic Hydrocarbons
in the Aquatic Environment. Boca Raton, FL, CRC Press.
2. Cullinane, M. J., Averett, D. E., Shafer, R. A., Male, J. W., Truitt, C. L, Bradbury, M.R., Requegnat,
W. E. 1990. Contaminated Dredged Material, Control, Treatment and Disposal Practice. Park
Ridge, New Jersey, Noyes Data Corporation.
3. Hambrick, G. A., DeLaune, R.D., et al. 1980: "Effect of Estuarine Sediment pH and Oxidation-
Reduction Potential on Microbial Hydrocarbon Degradation." Applied and Environmental
Microbiology 40(2): 365-369.
4. Heitkamp, M. A. and Cerniglia, C. E. 1987. "Effects of Chemical Structure and Exposure on the
Microbial Degradation of Polycyelic Aromatic Hydrocarbons in Freshwater and Estuarine
Ecosystems." Environmental Toxicology and Chemistry 6: 535-546.
5. Hites, R. A., LaFlamme, R. E., et al. 1980. Polycyelic Aromatic Hydrocarbons in Marine/Aquatic
Sediments: Their Ubiquity. Petroleum in the Marine Environment. Washington, D. C., American
Chemical Society. 289-311.
6. Leduc, R., Samson, R. et al. 1992. "Biotic and Abiotic Disappearance of Four PAH Compounds
from Flooded Soil Under Various Redox Conditions." Wat. Sci. Tech 26(1-2): 51-60.
7. Lee, R. F., Gardner, W. S. et al. 1978. "Fate of Polycyelic Aromatic Hydrocarbons in Controlled
Ecosystem Enclosures." Environmental Science & Technology 12(7): 832-838.
8. Wakeham, S. G., and Farrington, J. W. 1980. Hydrocarbons in Contemporary Aquatic Sediments.
Contaminants and Sediments. Ann Arbor, Michigan, Ann Arbor Science. 3-32.
241
-------
Table 1 Mass balance and partitioning of radiolabeled carbon for mixed reactors after 7 day
biodegradation tests.
5% Solids
% Initial % STDV
Mineralized
NS Extract
Volatilized
Sediment
Liquid
26.8
26.9
44.1
2.3
8.3
20.9
36.4
25.2
95.7
24.1
10%
% Initial
41.1
33.7
15.2
0.8
6.5
Solids
% STDV
7.5
19.9
59.9
175.0
61.5
15% Solids
% Initial % STDV
30.5
59.4
8.5
5.2
2.6
12.8
7.4
17.6
61.5
73.1
Total
108.4
6.1
97.4
3.5
108.1
4.6
242
-------
POLLUTANT FLUXES TO AQUATIC SYSTEMS VIA BED-SEDIMENT PROCESSES
V Popov, L Maranto, K T Valsara.i1. D D Reible, L J Thibodeaux, M A Todaro and J Fleeger
Departments of Chemical Engineering and of Zoology
Louisiana State University
Baton Rouge. LA 70803
Phone: 504 388 1426
INTRODUCTION
Hydrophobia organic contaminants (HOCs) sorb strongly to sediments and partition weakly
into the-porewater and overlying water. This leads to the detection of HOCs in sediments long
after their original introduction to the environment. Water bodies with active sediment
processes have larger fluxes of HOCs to overlying water. In the absence of sediment resuspension
by erosive processes, the normal life cycle activities of benthic organisms will predominate in
the transport of particles from within the sediment bed to the sediment-water interface. As a
result the HOCs associated with the particles are released to the water column. This process
is called bioturbatwn and is the focus of this paper. There are a number of species that act
as bioturbators. The most prevalent ones, especially in contaminated sediments across many sites
in the U.S. are the Tubificidae species that burrow within the sediment and defecate at the
surface.
Capping with clean sediment is a possible remediation measure for isolating the aquatic
system from contaminated sediments (Thoma et al, 1993). The placement of the cap will
effectively increase the pathlength for contaminant transport by diffusion and adyection and,
will also decrease pollutant release by direct bioturbation of contaminated particles. This
paper describes our experiments on contrasting the flux of HOCs from bioturbated capped and
uncapped sediments in small laboratory microcosms with those of control microcosms (capped and
uncapped) without bioturbators.
METHODOLOGY
The experiments were conducted in laboratory microcosms which have been described in
detail elsewhere (Reible et al, 1994). They were designed to simulate the flow of water over
a sediment surface. Each microcosm was 15 cm x 5 cm x 4.5 cm in dimension and was constructed
out of 0.64 cm thick plexiglass. Two flat overflow slots were located on both ends of the
microcosms to make water flow more uniformly over the sediment surface. Fifteen such microcosms
were used. The total sediment thickness in each was 3.5 cm that included the clean sediment used
as cap in some of the microcosms. Seven of them were control microcosms that contained no
organisms. Of the seven, three were uncapped and four were capped with clean sediment to a
thickness of 0. 5 cm. The remaining eignt microcosms contained 200 organisms each. -Of-the
eight, four were uncapped and four were capped with cleam sediment to a thickness of 0.5 cm.
The sediment used was obtained from a local area (Bayou Manchac, Baton Rouge, LA). It had an
organic carbon content of 2.8% and a porosity of 0.67. The sediment was processed and inoculated
with a mixture of three polyaromatic hydrocarbons (pyrene. phenanthrene and dibenzofuran). In
this paper we report our results only for pyrene. The pyrene concentration in the sediment was
31.9+1.1 mg/kg. The sediment inoculation procedure was described in an earlier publication
(Reible et al, 1994). The bioturbators used were Limnodrillus hoffmeisteri , head-down feeders
often found in many freshwater estuarine environments (Stimpson et al, 1985).
In the experiments reported here, the average flow rate of water over the sediment was
500 ml/h. This high flow rate ensured that water-side mass transfer resistance was negligible
in the transport of pyrene from sediment to water. The effluent water was collected periodically
to estimate the contaminant concentration on an HPLC by an EPA standard method (Reible et al,
1994). The flux of contaminant was determined from N = C AW At, where AV is the volume collected
in time At. The experiments were run for sixty days and then the sediment Was cored into 4 mm
thin sections and analyzed for pyrene. •
^•Author for correspondence
243
-------
RESULTS , :
The measured fluxes of pyrene from the experimental microcosms are shown in Figure 1.
The open symbols represent the chemical flux from the bioturbated systems while the solid symbols
represent the flux from the non-bioturbated control microcosms. All fluxes represent the average
of four replicate microcosms (except the controls which represent three microcosms); the error-
bars indicate the standard deviation in the measured fluxes. The standard deviation among the
replicate microc9sms was typically about 202, indicating the variability associated with the worm
biological activity, liquid extraction and concentration procedures for HPLC analysis. In figure
1 are shown the pyrene flux from both the capped and uncapped Bayou Manchac sediments populated
with 200 worms. For comparison the pyrene flux from the control channels are also presented.
Pyrene flux in uncapped control channels-decreased gradually from ~ 250 ng/cm2/d at the start of
the experiment to about 10 ng/cmz/d after 25 days. The flux remained stable at the lower value
for approximately 30 days. In the capped control channels the pyrene flux observed in the outlet
water was zero indicating the effectiveness of the cap in retarding the movement of a
hydrophobic compound from the underlying sediment. The data is in agreement with the observed
trend in earlier experiments on capping (Thoma et al, 1993). The organisms were added after six
days of the experiment, the influence of which is obvious in the flux of pyrene. Whereas the
pyrene flux in the control channels was about 60 to 70 ng/cnr7d at the time of worm incorporation
into the sediment, the bioturbated microcosms showed a flux of 140 ng/cm2/d. Even towards the
end of the experiment the flux in the bioturbated microcosms remained about 60 'ng/cnrYd. In the
capped microcosms populated by the worms, .the flux was quite stable during the course of the
experiment and varied between 30 to 40 ng/cnfVd. Towards the end of the experiment, the fluxes
from both the capped and uncapped sediments with worms were almost identical.
Figure 2 represents the sediment concentration profile for pyrene in uncapped control
without worms, capped control without worms, uncapped sediment with worms and capped sediment
with worms. The profiles in the uncapped controls without worms showed the expected exponential
profile pver the first 4 mm of the contaminated layer where the sediment concentration decreased
from an initial loading of 32 mg/kg to 7 to 12 mg/kg. In the uncapped microcosms with worms the
depletion of pyrene was larger at the surface. In the capped control without worms, the expected
effect of decreased movement of pyrene at the surface was apparent; in fact, there was only
minimal amout of pyrene that migrated into the clean cap. In the microcosm with cap and worms,
there was a clear indication of the worm transporting pyrene through the cap. It was observed
that the worm fecal pellets collected at the surface contained about 5 mg/kg of pyrene, while
the underlying 3 mm of the cap contained no pyrene. The depletion layer in this case was - 10
mm. • . . i
The above results clearly indicate" that bioturbation significantly increased the
contaminant flux from the sediment over control capped and uncapped microcosms. The total
contaminant flux from the sediment can be regarded as the sum of the diffusive flux from the
porewater and the bioturbation flux. The latter is a combination of the sediment particle
transport and water pumping during feeding activities by worms. Initially the diffusion from
the sediment surface will be large; however, with time the effective depletion of the surface
sediment will lead to a decrease in the rate of molecular diffusion. As the diffusive flux
decreases with time relative significance of bioturbation will increase. The bioturbation flux
from the sediment can be determined by subtracting the flux in the controls without worms from
those of the microcosms with worms. For pyrene the average bioturbation related fluxes estimated
thus were: 55±14 ng/cnr/d in the uncapped microcosms, and 33+6 ng/cnrYd in the capped microcosms.
Therefore the decrease in pyrene flux as a result of the cap was only 402 in the presence of the
worms. The flux due to bioturbation as obtained above was observed to remain relatively constant
for pyrene during the course of the experiment.
An effective bioturbation mass transfer coefficient can be defined as the ratio of the
flux due to bioturbation to the contaminant concentration in the sediment:
~ N
(1)
The numerator represents the bioturbation flux, viz., the difference in flux between the'
microcosms with the species (N obs) and the microcosms without the species (Ncont). pb is the
244
-------
E
o
X
300
250 -
200 -
150 -
g 100
50 -
0 -
Control - uncapped
Control - capped
•-A-- Bioturbed - uncapped
Bioturbed - capped
10
—, p—
20 30
time (days)
40
50
60
Figure 1. Sediment-to-water flux of pyrene in the various microcosms as a function of
time.
sediment bulk density and Wsfi4 is the sediment concentration of pyrene. The effective
bioturbation mass transfer coefficient for pyren'e was 0.7 cm/y. This is comparable to measured
values in the field for other compounds which range from 0.1 to 10 cm/y (Reible et al, 1994).
Assuming that the effective bioturbation depth is limited to - 3 cm, an effective bioturbation
diffusion constant can be estimated from Db10 = Kb • (3 cm), which for pyrene would be 2.1 cmVy.
CONCLUSIONS • :
The major focus of these experiments was to evaluate the effects of a conveyor-belt worm
Limnodrilus hoffmeisteri on capping as a .control technology for contaminated sediment
remediation. The overall 'approach was to inoculate the sediment with contaminants and then
monitor its effluent concentration in small microcosms inhabited by the species. The pyrene flux
in uncapped control microcosms decreased gradually from 250 ng/cnr/d at the start of the
experiment to about 10 ng/cm2/d after 30 days. The flux then remained constant.for the next 30
days. The addition of the worms after six days of the experiment markedly increased the flux
of pyrene. The flux due to bioturbation from an uncapped sediment was 55+14 ng/cm/d and from
a capped sediment was 33±6 ng/cmz/d. Towards the end of the experiment the flux from both capped
and uncapped sediments were similar. The effective bioturbation mass transfer coefficient for
pyrene was 0.7 cm/y. Capping effectively reduced the contaminant release to the water column
only in the absence of the worms. In the presence of the worms, a 5 mm layer of cap was found
to be less effective in retarding the movement of pyrene from the underlying contaminated
sediment to the overlying water. ' .
245
-------
35-
30 -
1?
==£
g 25-
c
_o
'!* 20 -
•4— •
c
0)
o
I 15-
<1>
c
0)
& 1°-
"c
(U
1 5-
u
CO
0 -
A _
/"^^ A A •''•'"*
/ A ''• ^^
/ A-' • 'A ^o
18
Figure 2. Sediment concentration profile for pyrene in the different microcosms as a
function of depth into the sediment.
REFERENCES
Reible D D, V Popov. K T Valsaraj, L J Thibodeaux, M Dikshit, F Lin. M A Todaro and J F Fleeger
(1994), Contaminant fluxes from sediment due to tubificid oligochaete bioturbation, Water
Research ( under revi ew).
Stimpson K S. DO Klemm and J K Hiltunen (1985), Freshwater tubificidae: Annelida Oligochaeta.
In A Guide to the Freshwater Annelida (Polychaeta, Naidid and Tubificid Oligochaeta and
Hirudenia) of North America, Klemm D J (ed), Kendall/Hunt, Dubuque, IA, p.44.
Thoma. G J, D D Reible, K T Valsaraj and L J Thibodeaux (1993), Efficiency of capping
contaminated sediments in situ 2. Mathematics of diffusion-adsorption in the capping
layer. Environ. Sci. Technol., 27, 2412-2419.
ACKNOWLEDGEMENT
This work was funded by a grant (R 819165-01) from the US EPA to the LSU Hazardous Substance
Research Center (South and Southwest). ,
246
-------
ACCELERATING RESEARCH AND DEVELOPMENT OF INNOVATIVE TECHNOLOGIES
Norma M. Lewis and Randy A. Parker
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Office of Research and Development
26 West Martin Luther King Drive
Cincinnati, OH 45268
(513)569-7665
The U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation
(SITE) Program encourages the development of innovative technologies for faster, more effective, and
less costly treatment of hazardous waste. The SITE Program demonstrates full scale technologies
through the Demonstration Program, and provides financial and technical assistance to developers of
bench- and pilot-scale technologies through the Emerging Technology Program. Other SITE programs
support assessment of innovative technologies for site characterization (Monitoring and Measurement
Technologies Program), and for dissemination of information on innovative technologies (Technology
Transfer Program).
The primary focus of this panel discussion is the SITE Emerging Technology Program. This
Program brings government and the private sector together to foster and accelerate research and
development of innovative technologies for commercialization. The Emerging Technology Program
seeks to accelerate the development of innovative an alternative technologies by:
• Entering into Cooperative Agreements with innovative technology developers for
technical and financial assistance;
• Providing access to national and international scientific and engineering information
resources;
• Promotion of technologies to a level of confidence where they can be demonstrated in
the field; and
• Encouraging technologies with potential to move into the commercial marketplace.
The panel discussion will bring together representatives of government agencies, private sector
and university developers of innovative technologies, and representatives of environmental policy
makers in an open forum to discuss the opportunities and impediments associated with the
development of innovative technologies. Panelists and audience members will discuss the various
aspects of innovative technology development from both the government and private sector
perspectives.
247
-------
SUPERCRITICAL EXTRACTION OF POLYCHLORINATED BIPHENYLS FROM SOILS AND
SEDIMENTS: REMEDIATION AND POSSIBLE RISKS
Lawrence L. Tavlarides*, Pairu Chen, Wu Zhou, Wayne S. Amato
Department of Chemical Engineering and Materials Science, Syracuse University
320 Hinds Hall, Syracuse University
Syracuse, NY 13244-1190
Tel: (315) 443-1883
Fax:(315)443-2559
INTRODUCTION
It is estimated that there exists about 350 million pounds of polychlorinated biphenyls (PCBs) in
landfill and other storage, and another 24 million pounds in sediments, soil, vegetation and animals
(1). The potential environmental threat of the large amount of PCBs has called for the development
of effective PCB cleanup techniques. Although some PCB remediation technologies such as
incineration and in-situ vitrification have been commercial available, many efforts are still being
conducted in order to develop more economic and social acceptable methods (1-3). Among these
efforts, the application of supercritical fluid extraction (SFE) in the removal of toxic organics from
environmental samples is receiving much attention due to the unique properties of supercritical fluids
(SCFs) such as low viscosity, high diffusivity, and easily tunable solvent power (2-5). A process
concept advanced at Syracuse University will be presented here.
The process to treat PCB contaminated soils with supercritical technique may involve some risks.
These risks could include additional contamination of underground waters due to fracture and
permeation through underground structure when contaminated soils are excavated; removal of
organic matter from soil and rendering it inert; hazardous operation of large scale high pressure
mobile apparatus; and hazards in surface transportation of PCB soil extracts to a central site for
supercritical water oxidation (SCWO) destruction. These issues will be discussed.
METHODOLOGY
We are developing a new generation of soil remediation technology at Syracuse University to clean
soils and sediments contaminated with chlorinated hydrocarbons such as PCBs as well as polycyclic
aromatic hydrocarbons (PAHs). The two-stage concept advanced is shown in Figure 1. First (Figure
1a), the toxic PCBs or PAHs are extracted from the sediments or soils using SFE with high-pressure
carbon dioxide fluids. The clean soils/sediments are returned to the site, the concentrated toxic
organics are separated from the supercritical fluids for further processing, and the supercritical fluids
are recycled. In the second process step (Figure 1b), SCWO, the concentrated toxic organics are
destroyed by wet oxidation to form harmless carbon dioxide, water and hydrochloric acid.
A laboratory scale extraction unit is employed to obtain desorption data for PCB removal from
laboratory-spiked and native contaminated soils with supercritical carbon dioxide and cosolvents.
Various conditions of extraction for a variety of soils/sediments have been studied. A flowsheet of
the unit is shown in Figure 2. The fixed-bed unit can be operated at pressures and temperatures as
high as 680 atm and 100 °C.
248
-------
Contaminated
Soil/Sediment
Pressure
Reducer
PCBs, PAHs, PCDDs etc.
(to SCWO Process)
COz/Cosolvent Recycle
(a) Supercritical Fluid Extraction(SFE)
(From SFE Process)
PCBs, PAHs, PCDDs, etc.
r-&4c02, HCI,
etc.
| Compressor
(b) Supercritical Water Oxidation(SCWO)
Figure 1. Supercritical Methods for Soil Clean-up
249
-------
tf
fc.,3 ByPass ^ Mloometertng Diy Test Meter
—fVft—I r-C^)-».Vent
cv
Cold Trap
Figure 2. Schematic of Laboratory SFE System
RESULTS
Experiments have been executed with the laboratory scale unit in a wide range of experimental
conditions such as temperature, pressure, cosolvent type and concentrations, soil type and initial
PCS concentration. A summary of results are shown in Table 1 which shows the contact times
necessary to achieve sub 10 ppm removal of PCB congeners (Aroclor 1248) from various soil types
at specific initial concentration.
TABLE 1. SUMMARY OF EXPERIMENTS WITH SUB 10 PPM REMOVAL OF AROCLOR 1248
Soil Initial Pressure Temp. Cosolvent Moisture Contact Residual
Cone. (atm) (°C) (mol%) (wt.%) Time Cone.
(PPm) (min.) (ppm)
Till 1150
Sediment
(Native, St. 370
Lawrence
River)
Sediment
(Native, 71
Hudson River)
Surficial 4200
(4" - 12")
P1
P2
P2
P2
P1
P1
P1
P1
P1
P1
P2
P2
P2
T1
T2
T3
T1
T2
T2
T2
T2
T3
T3
T2
T3
T2
5
5
5
0
5
5
5
5
5
5
5
5
5
10
10
10
10
0
10
0
20
0
20
10
10
0
60
45
30
120
60
60
45
45
60
60
60
30
30
2.6
3.2
3.0
3.4
18.9
15.1
6.5
5.7
4.3
3.2
4.4
6.7
8.0
Sand
950
P1
T2
10
15
1.2
These results show that sub 10 ppm residual concentrations can be achieved at fixed-bed contact
times of 15 to 60 minutes.
250
-------
500
450
400
-' 350
c
•s 300
is
| 250
3 200
I 15°
E 100
50
0
0
•
•A
•
O
„ A
•XX
I
AMoisture=10 wt.%. T2
OMo1s«ure=10 wt,%. T3(>T
XMolsture=0 wt.%. T3(>T2
A
O
*
4
k
^ .
¥
2)
— i— I
250
•g- 200
•I 150
g
o
m
•a
100
50
I
0
, o
A
A
O
A
AMoisture=0 wt.%. P1
OMoisture=10 wt.%. P2(>P1)
<
I
A
<
10 20 30 40 50
Extraction Time (min)
60
10 20 30 40 50
Extraction Time (min)
60
Figure 3. Temperature and Moisture Effects on
SFE of Arolor 1248 from Lab-spiked Till
Conditions: P2, MOD1=5 moi%,
Initial Concentration=1150 ppm.
Figure 4. Moisture Effects on SFE of PCBs from
Contaminated Sediment of St. Lawrence River
. Conditions: T2, MOD1=:5 mol%,
Initial Concentration=370 ppm.
Figures 3 and 4 are the experimental results whereby the residual Aroclor 1248 concentration is
plotted versus time of extraction during which the supercritical carbon dioxide with cosolvent is
passed through the bed of contaminated soil. In Figure 3, the results for till are illustrated. Higher
temperature of T3 is preferred to remove Aroclor 1248 from wet till to sub 5 ppm. At T2, it takes 45
minutes, but 30 minutes at T3. It is also seen that a rapid drop occurs without moisture present, to
about 25 ppm after 3.5 minutes of contact, however, an equilibrium value is reached and the final
concentration is 12 ppm.
In Figure 4, it is seen that the residual concentration drops to 19 ppm between 30 to 45 minutes for
extraction of Aroclor 1248 from contaminated, St. Lawrence River sediment of initial concentration of
370 ppm for either dry or 10 wt.% moisture content sediment. These results occur although the dry
samples were extracted at a lower pressure. The residual concentrations remain at the same level
even after 60 minutes of extraction which indicates the achievement of equilibrium value. Desprption
rate models are being applied to correlate these data. . ,
Economic estimates have been made comparing treatment costs for supercritical extraction of
PCBs from soil with other competing technologies. This comparison is based on an up-date of work
by Carpenter (15) and has built into it the soil handing costs and shown in Table 2.
The comparison shows the favorable economics for SFE/SCWO of PCBs from soils.
CONCLUSIONS
The results of laboratory scale SFE clearly indicate that supercritical carbon dioxide/cosolvent
fluids can remove PCBs in various contaminated soil matrices to sub 10 ppm level under proper
conditions of contact.
An economic analysis based on laboratory scale unit results shows that the SFE/SCWO of PCBs
from soils to be a favorable process. ,
251
-------
Despite the advantages of the supercritical technology for remediation of contaminated soils, risks
may still occur during the processing procedures. The risks involved would be disturbing
underground structure and causing dispersion of contaminants beyond the hazardous waste site;
removal of organic matter from soil and producing an inert soil which may be unsuitable for reuse;
hazards in transporting PCB laden extracted hydrocarbons to a central site for SCWO destruction;
and hazardous operation of large scale high pressure mobile apparatus. These risks do not appear
insurmountable, however.
SCWO studies are in progress to determine conditions for acceptable PCB destruction efficiencies,
kinetics of the reaction, and economic estimates of this process step.
A bench scale unit has been fabricated and installed for experiments to confirm the laboratory
conditions and to establish suitable extractor geometry and contacting schemes for a viable process.
TABLE 2* COMPARISON OF TREATMENT PROCESSES
Process (order of rank)
Cost Range ($/m3 processed)
O.H.M. Methanol Extraction
Advanced Electric Reactor (thermal treatment)
Acurex Solvent Wash (hexane/trichlorotrifluoro-
ethane extraction
Bio-Clean
Vitrification
LARC (isopropanol extraction and radiation
treatment)
Modar Supercritical Water
Soilex Solvent Extraction (kerosene extraction)
SFE/SCWO (estimate in this work)*
$513 to $658
$1063 to $1206
$251 to $728
$245 to $474
$326 to $702
$286 to $430
$320 to $938
$1096 to $1169
$220 to $270
* First eight process costs taken from Carpenter (15); 1985 dollars updated to 1994 dollars using 1.35
M&S factor.
1 Assumptions for this estimate are given in the text.
REFERENCES
1. Amend, L. J. and Lederman, P. B. Critical Evaluation of PCB Remediation Technologies.
Environmental Progress 11 (3):173-177.1992.
2. Chen, P.; Zhou, W.; Zhang, J. and Tavlarides, L. LDesorption of Polychlorinated Biphenyls from
Soils with Supercritical Fluids. Environmental Science and Technology in preparation.
3. Zhou, W.; Chen, P. and Tavlarides, L. L. Supercritical Fluid Extraction of Polychlorinated Biphenyls
from Sediments. Environmental Science and Technology in preparation.
4. Richter, M.; Chen, P.; Zhou, W. and Tavlarides, L. L. Modeling of Desorption of Polychlorinated
Biphenyls from Soils under Supercritical Conditions. Ind. Eng. Chem. Res, in preparation.
5. McHugh, M. and Krukonis, V. In: M.A. Stoneham (ed.),Supercritical Fluid Extraction, Butterworth
Publishers, 1986, p. 10.
6. Chester, T. L.; Pinkton, J.D. and Raynie, D.E. Supercritical Fluid Chromatography and Extraction.
Anal. Chem. 64:153R-170R, 1992.
7. Hawthorne, S.B.; Miller, D.J.; Burford, M.D.; Langenfeld, J.J.; Eckert-Tilotta, S. and Louie, P.K.
Factors Controlling Quantitative Supercritical Fluid Extraction of Environmental Samples. J.
Shrank 642:301-317,1993.
8. Burford, M.D.; Hawthorne, S.B. and Miller, DJ. Extraction Rates of Spiked versus Native PAHs
from Heterogeneous Environmental Samples Using Supercritical Fluid Eixtraction and Sonication
in Methylene Chloride. Anal. Chem. 65:1497-1505,1993.
252
-------
9. Hawthorne, S.B.; Krieger, M.S. and Miller, D.J. Supercritical Carbon Dioxide Extraction of
Polychlorinated Biphenyls, Polycyclic Aromatic Hydrocarbons, Heteroatom-Containing Polycyclic
Aromatic Hydrocarbons, and n-Alkanes from Polyurethane Foam Sorbents. Anal. Chem. 61:736-
740,1989.
10. Langenfeld, J.J.; Hawthorne, S.B.; Miller, D.J. and Pawliszyn, J. Effects of Temperature and
Pressure on Supercritical Fluid Extraction Efficiencies of Polycyclic Aromatic Hydrocarbons and
Polychlorinated Biphenyls. Anal. Chem. 65:338-344.1993.
11. Hawthorne, S.B.; Langenfeld, J.J.; Miller, D.J. and Burford, M.D. Comparison of Supercritical
CHCIF2, N2O, and CO2 for the Extraction of Polychlorinated Biphenyls and Polycyclic Aromatic
Hydrocarbons. Anal. Chem. 64:1614-1622.1992.
12. Fahmy, T.M.; Paulaitis, M.E.; Johnson, D.M. and McNally, M.E.P. Modifier Effects in the
Supercritical Fluid Extraction of Solutes fro Clay, Soil, and Plant Materials. Anal. Chem. 65:1462-
1469,1993.
13. Bwadt, S. and Johansson, B. Analysis of PCBs in Sulfur-Containing Sediments by Off-line
Supercritical Fluid Extraction and HRGC-ECD. Anal. Chem. 66:667-673,1994.
14. Alexandrou, N. and Pawliszyn, J. Supercritical Fluid Extraction for the Rapid Determination of
Polychlorinated Dibenzo-p-dioxins and Dibenzofurans in Municipal Incinerator Fly Ash. Anal.
Chem. 61:2770-2776.1989.
15. Carpenter, B.H. PCB Sediment Decontamination-Technical Economic Assessment of Alternative
Treatments. EPA 600/2-86/112. U.S. Environmental Protection Agency, 1986.
253
-------
BY-PRODUCTS FROM SUPERCRITICAL WATER OXIDATION:
PATHWAYS. KINETICS. AND MECHANISMS
Phillip E. Savage* and Sudhama Gopalan
University of Michigan, Chemical Engineering Department
Ann Arbor, Ml 48109-2136 phone (313) 764-3386
INTRODUCTION
Supercritical water oxidation (SCWO) is an emerging technology for the ultimate destruction of
organic wastes. Organic compounds and oxygen can be intimately mixed in a single homogeneous
aqueous phase at supercritical conditions (Tc = 374°C, Pc = 218 atm). Thus, the rapid oxidation
reactions are unhindered by inter-phase transport limitations that could occur at subcritical conditions
where multiple phases exist. Savage et al. (1) describe current research into SCWO reactions in their
review of reactions at supercritical conditions.
The rational design, optimization, control, and analysis of SCWO processes requires a knowledge
of SCWO kinetics and potential byproducts formed from the oxidation of real pollutants. Our research
group has focused on the oxidation of phenolic compounds (2-10), and this presentation will provide an
overview of our most recent work.
METHODOLOGY
All oxidation experiments were performed in a reactor that nominally operated isothermally,
isobarically, and in plug flow. Aqueous solutions of 02 and the phenol were prepared separately and
used as the reactor feed streams. The feed streams were pressurized and pumped through the reactor
using two liquid chromatography pumps. The phenol and oxygen streams were preheated in separate
coils of Hastelloy C-276 tubing. These preheater lines meet in a mixing union, where the temperature
was measured by a thermocouple and the mixed feed streams enter the 4 m. long by 0.125 in. O.D.
Hastelloy reactor. The preheater lines, mixing tee, and reactor coil are housed in an isothermal fluidized
aluminum-oxide bath equipped with a temperature controller.
The reactor effluent was cooled in two consecutive tube-in-tufae heat exchangers and
decompressed in a back pressure regulator. The exiting stream was separated into gas and liquid
phases (at ambient conditions) in a liquid trap. The gas flow rate was measured with a bubble meter at
the outlet of the system, and the gas stream was then sent to an on-line gas chromatograph (GC)
equipped with a thermal conductivity detector (TCD). The liquid flow rate was measured, and samples of
the liquid phase were retained for analysis.
A reverse-phase high-performance liquid chromatograph (HPLC) was used to determine the
concentration of phenolics in the liquid effluent samples. Analyses were performed isocratically with a
mobile phase of water/acetonitrile in a 5:2 v/v ratio flowing at 1 ml/min. The UV absorbance at a
wavelength of 210 nm was monitored, and detector response factors were determined experimentally.
Before additional products in the liquid phase were analyzed, the samples were concentrated
because many products were present in low concentrations. 20 ml_ of the reactor effluent was extracted
with three successive 10 ml_ aliquots of dichloromethane. The 30 mL volume of this organic phase was
subsequently reduced to 1 mL using a Kuderna-Danish concentrator in a water bath at 50°C. 10 mLpf a
dichloromethane solution containing a standard was added to each sample prior to concentration.
Reaction products in these concentrated samples were identified by GC-MS. The reaction
products were quantified by GC with a flame ionization detector (FID). When a suspected reaction
product was available commercially, we positively identified that product by matching both the mass
spectrum and retention time with those of the authentic sample. The FID response factor was then
determined experimentally for these compounds. Other suspected products, for which the authentic
compound was not available commercially, were tentatively identified by inspecting the mass spectra and
254
-------
matching them to spectra stored in the GC-MS computer database. The response factors for these
products were assumed to be equal to those determined experimentally for chemically similar
compounds. The GCs were equipped with12 m x 0.2 mm O.D. HP-1 capillary columns. Helium served
as the carrier gas, and 1 ml_ of the sample was injected in the splitless mode.
RESULTS
Thornton et at. (5), Li et al. (3), Martino et al. (10), and Gopalan and Savage (8) identified several
multi-ring products (e.g. dibenzofuran, 2 and 4-phenoxyphenol, dibenzo-p-dioxin and 2,2'-biphenol) from
the oxidation of phenols in supercritical water at temperatures up to 480°C. The formation of dimers is
important because some of these are potentially more toxic than the reactant. By determining the
concentrations, molar yields, and selectivities of many of these products, we have been able to develop
reaction networks that describe the formation of these multi-ring species.
The data of Thornton and Savage (7) for phenol SCWO, and Li et al. (2,3) for SCWO of 2-
chlorophenol show some general trends for product selectivities. The dimer selectivity tends to decrease
with increasing phenol conversion while the selectivity to gases (CO, COa) tends to increase with
increasing phenol conversion. We developed a reaction network that was consistent with these qualitative
trends in the experimental product selectivities. This network considered phenol, dimers, carbon oxides,
and "others" as the distinct chemical species. In this network, phenol is consumed by parallel paths that
form dimers and "other" products. The dimers decompose to the ring-opening and other products. Finally,
the "other" products are oxidized to CO and
The mathematical modeling of each network involved differential equations for each of the
product categories. To determine the rate parameters for each of the four reactions in the model, we
employed SimuSolv (11). The parameter estimation process incorporated an integration routine in
addition to the optimization. A linear multistep predictor-corrector integration routine, LSODE, was
employed in concert with the generalized reduced-gradient method. The entities for which experimental
data were obtained; viz. phenol, dimers, and gases, were the response variables. The criterion for
parameter estimation was the maximization of the log likelihood function for each of the response
variables. The reaction orders (a/.b,-) for the organic compound and oxygen for each pathway were
determined from data at a single temperature, 460°C.
The reaction orders obtained from the regression of data at 460°C were then fixed in the
subsequent parameter estimation. Also, the water order for each step was fixed at 0.42, the value
obtained for phenol disappearance kinetics. Data at four temperatures; 420, 440, 460 and 480°C were
then used to determine the Arrhenius parameters for the four individual reactions. Table 1 displays the
results of the parameter estimation process for the network model. The units are kcaj, moles, liters, and
seconds.
Table 1. PARAMETERS IN REACTION NETWORK MODEL
Reaction 1 Reaction 2
Reaction 3
Reaction 4
a
b
E. (kcal/mol)
d
A(M0.58-a-bs-1)
k @ 460°C,250 atm
0.89 ± 0.26
0.49 ±0.23
10.0 ±4.7
101.8 + 1.5 .
0.026± 0.009
0.86 + 1.19
0.48 ±1.15
25.5 ±15.0
105.6±5.5
.0034 ± 0.0001
0.86 ± 0.30
0.60 ±0.32
28.8 + 4.1
108.5 + 2.6
0.398 ± 0.023
0.50 + 0.15
1.49 ±0.31
44.8 ± 9.0
1Q13.5 + 2.8
0.293 ± 0.054
(M = moles of carbon per liter)
We used the model with the parameters in Table 1 to predict the results of phenol SCWO reported by
Thornton and Savage (6,7), and Li et al. (2). Figure 1 shows a comparison of the effect of residence time
on the phenol conversion and the dimer and gas selectivities at 380°C and 278 atm. The curves, which
give the model predictions, capture the trends in the experimental data well.
255
-------
0.8
0.6
o
•a 0.4
0.2
Of->
10 20 30 40 50 60
Time (s)
70 80
Figure 1. Product Yield vs. Residence Time for Phenol SCWO at 380°C and 278 atm
The estimated activation energies in Table 1 enable us to compare the rate constants for phenol
consumption along the competing parallel paths as a function of temperature. The initial dimer
selectivities calculated from the model (calculated as ki/ki-d<2) range from 89% at 420°C to 78% at
480"C. These rate constants when extrapolated to 600°C predict an initial selectivity to dimers of ~45%,
which shows that even at higher temperatures dimerization and ring-opening reactions are likely to be
equally important. Examining lower temperatures we note that at 380°C, nearly 95% of the initial
selectivity of phenol is towards dimers as shown in Figure 1. These results show that dimerization is the
main primary pathway for phenol. In fact, at 380°C, the path from phenol to ring-opening and other
products is negligible, and the reaction network could be reduced to a set of consecutive reactions.
The model also shows that although the formation of dimers cannot be prevented, the selectivity
to dimers decreases with increasing phenol conversion (Figure 1). At complete conversion of phenol the
selectivity towards dimers is essentially zero. Also, the destruction of dimers is aided by higher
temperatures as is evident from the activation energy for the destruction of dimers (29 kcal/mol)
exceeding that for dimer formation (10 kcal/mol). This difference in activation energies translates to an
Increase in the ratio of the rate constant for destruction of dimers to that for the formation of dimers
(kg/k1) by a factor of 7 when the temperature increases from 380°C to 480°C.
The parameters in Table 1 also show that higher temperatures have the desirable effect of favoring the
formation of gases. This is evident from the activation energy for the formation of gases (45 kcal/mol)
exceeding the activation energies for the steps leading to the formation of the ring-opening and other
Intermediates (25 and 29 kcal/mol).
The reaction orders for phenol in Table, 1 for the two primary paths are nearly equal, so varying
the phenol concentration will have little effect on the product selectivities. The reaction orders for oxygen
are also nearly equal for all the steps except for step 4 which has an oxygen order of 1.5. This high
oxygen order indicates that high oxygen concentrations will result in higher rates of formation of gases.
256
-------
In addition to modeling the reactions using empirical reaction networks, we have developed
detailed chemical kinetics models based on the reaction mechanisms. This work (9) focuses on adapting
gas-phase combustion kinetics to SCWO conditions as the basis for a quantitative model.
CONCLUSIONS
The products of phenol oxidation in SCW can be categorized as dimers, gases, and ring-opening and
other products. The variation of the selectivity of phenol to these products with residence time,
temperature and species concentrations can be quantitatively described using a reaction network and the
parameters in Table 1. The network includes parallel pathways for phenol to dimers and to ring-opening
and other products, secondary decomposition of dimers to ring-opening and other intermediates, and
oxidation of these intermediates to carbon oxides.
Dimerization of phenol is the main primary pathway for phenol consumption between 380 and 480°C.
A successful strategy for treatment of phenolic wastes by SCWO must ensure the destruction of these
dimers and the formation of CO2 in high selectivities. Our quantitative reaction model showed that long
residence times and high reaction temperatures favor the destruction of dimers and the formation of CO
and CO2- Moreover, the influence of intermediate reaction products (dimers and other products) on the
rate of formation of the carbon oxides shows that phenol disappearance kinetics alone constitutes
insufficient information for process design, identifying and quantifying reaction intermediates and
developing accurate quantitative reaction models is of vital importance for any rational SCWO process
design.
REFERENCES
1. Savage, P. E., Gopalan, S., Mizan, T. I., Brock, E. E., Martino, C. J. Reactions at Supercritical
Conditions: Fundamentals and Applications. AlChEJ. accepted. 1995.
2. Li, R., Thornton, T. D., and Savage, P. E. Kinetics of CO2 Formation from the Oxidation of Phenols in
Supercritical Water. Environ. Sci. Technol.. 26. 2388.1992.
3. Li, R., Savage, P. E., and Szmukler, D. 2-Chlorophenol Oxidation in Supercritical Water: Global
Kinetics and Reaction Products. AlChE J.. 39.178.1993. ' ,
4. Thornton, T. D., and Savage, P. E. Phenol Oxidation in Supercritical Water. J. Supercritical Fluids. 3,
240,1990.
5. Thornton, T. D., LaDue, p. E., and Savage, P. E. Phenol Oxidation in Supercritical Water: Formation
of Dibenzofuran, Dibenzp-p-dioxin, and Related Compounds. Environ. Sci. Technol.. 25,1507,1991.
6. Thornton, T. D., and Savage, P. E. Kinetics of Phenol Oxidation in Supercritical Water. AlChEJ.. 38,
321,1992. •:.->•
7. Thornton, T. D., and Savage, P. E. Phenol Oxidation Pathways In Supercritical Water. Ind. Eng.
Chem. Res.. 31. 2451.1992. ,
8. Gopalan, S., and Savage, P. E. Reaction Network for Phenol Oxidation in Supercritical Water: A
Comprehensive Quantitative Model. AlChE J. accepted. 1995.
9. Gopalan, S., and Savage, P. E. Reaction Mechanism for Phenol Oxidation in Supercritical Water 4
Phys. Chem. 98. 12646. 1994.
10. Martino, C. J., Savage, P. E., and Kasiborski, J. Kinetics and Products from o-Cresol Oxidation in
Supercritical Water. Ind. Eng. Chem. Res, submitted. 1995.
11. Steiner, E. C., Rey, T. D., and McCroskey, P. S. SimuSolv Modeling and Simulation Software
Reference Guide. The Dow Chemical Company, 1990. '
FOR MORE INFORMATION
Prof. Phillip E. Savage, University of Michigan, Chemical Engineering Department, Ann Arbor, Ml 48109-
2136, Phone (313) 764-3386, e-mail: psavage@engin.umich.edu
257
-------
OXIDATION AND HYDROLYSIS OF ACETIC ACID AND METHYLENE CHLORIDE IN
SUPERCRITICAL WATER AS A MEANS OF REMEDIATION
Philip A. Marrone, Russell P. Lachance, Joanna L. DiNaro,
Brian D. Phenix, Jerry C. Meyer, Jefferson W. Tester,
William A. Peters
Chemical Engineering Department and Energy Laboratory
Massachusetts Institute of Technology
Room E40-455
77 Massachusetts Avenue
Cambridge, MA 02139
(617) 253-3401
K.C. Swallow
Merrimack College
North Andover, MA 01845
INTRODUCTION
Supercritical water oxidation (SCWO) is a promising technology proposed for the destruction of
hazardous organic wastes. Unlike its well known behavior under ambient conditions, water above its
critical point (374° C, 221 bar) has properties similar to that of a nonpolar solvent, primarily due to the
effect of a decrease in hydrogen bonding and density that occurs near and above the critical point.
The result is that nonpolar organics and oxygen exhibit complete solubility in supercritical water, while
polar species such as inorganic salts are insoluble and precipitate out. In the single homogeneous
phase formed, oxidation of organics with oxygen in supercritical water is rapid and complete to CO2 and
H2O. Organic heteroatoms such as halogens, sulfur, or phosphorus are converted to inorganic acids
(HCI, H2SO4, HgPO^ which precipitate as salts when neutralized with added base, while nitrogen is
converted to N2 and N2O. No NOX compounds are formed due to the relatively low temperatures that
exist in the SCWO process (400 - 650°C) relative to that of air incineration processes (typically 900 -
1300°C). Oxidation in supercritical water is thus an appealing means of destroying toxic organic
compounds while simultaneously separating out undesired inorganics by precipitation. Applications to
decontaminating soils and dilute aqueous wastes are of special interest. Earlier work has demonstrated
high destruction efficiencies for various organics in SCWO (1).
METHODOLOGY
Our methodology includes laboratory-scale experiments and interpretive mathematical modeling to
deduce global kinetic parameters and a fundamental mechanistic understanding of the destruction
pathways of model organic compounds. These model organics are compounds that are either typical
and representative wastes and soil contaminants themselves, refractory intermediate compounds
formed in the oxidation of a more complex compound, or surrogates for compounds too dangerous to
use in the lab (such as chemical weapon agents) but that are chemically similar in some respect to the
more toxic compound they represent. Waste destruction efficiencies, yields, and identities of resulting
products are measured over ranges of key operating variables (e.g. temperature, pressure, reaction
time, organic/oxidant feed ratio, possible presence of solids) pertinent to normal and upset conditions.
This paper describes the experimental results of oxidation and hydrolysis (no added oxygen) in
supercritical water of two model compounds that have been investigated in our lab - acetic acid and
methylene chloride.
258
-------
RESULTS AND CONCLUSIONS
Acetic acid (CH3COOH) and methylene chloride (CH-jClj) oxidation and hydrolysis in supercritical
water were examined in an isothermal plug flow tubular reactor. Acetic acid experiments were
conducted over the range of 425-600° C, 160-263 bar, and 4.4-10 sec reactor residence times, with
stoichiometric, substoichiometric, and superstoichio metric oxygen to organic feed ratios. Acetic acid
oxidation was globally observed to be 0.72 ± 0.15 order in acetic acid and 0.27 ± 0.15 order in oxygen
(differing from the usual trend of first-order in organic and zero-order in oxygen), with an activation
energy of 168±21 kJ/mol and pre-exponential factor of 1099±17. Complete conversion was achieved
above 550° C and 8 sec reactor residence time at 246 bar. Major products formed included CO, CO2,
H2, and CH4. For hydrolysis, where some conversion was observed but not as much as under similar
oxidation conditions, an assumed first order global rate expression had an activation energy of 94±17
kJ/mol and pre-exponential factor of 1044±1-1. Pressure (water density) variations were observed to
have a small effect on acetic acid oxidation.
Methylene chloride experiments were carried out at 450-600° C, 246 bar, and 4-9 seconds reactor
residence time, with O^CH2C\2 feed ratios ranging from substoichiometric to superstoichio metric. Major
products of both hydrolysis and oxidation of CH2CI2 were carbon monoxide (CO), carbon dioxide (CO2),
formaldehyde (HCHO), methanol, (CH3OH), hydrogen (H2), and hydrochloric acid (HCI). Small amounts
of methane (CH4) were detected above 575° C, and trace amounts (< 1% of total gas volume) of
chloromethane (CH3CI), 1,1 -dichloroethylene (C2H2CI2), cis-1,2-dichloroethylene,
trans-1,2-dichloroethylene, chloroform (CHCI3), and trichloroethylene (C2HCI3) wero found in the vapor
phase effluent.
Unlike that seen with earlier model compounds such as acetic acid, conversion of methylene
chloride under hydrolysis conditions was very similar to that under oxidation conditions. In fact, the rate
of destruction of CH2CI2 was globally determined to be zero order in O2 concentration. The presence of
O2 did affect the product distribution, however, with more CO and CO2 and less HCHO, CH3OH, and H2
produced under oxidation conditions than under hydrolysis conditions. Complete conversion of CHZCI2
was achieved above 563°C at a reactor residence time of 6 seconds under oxidation conditions.
Significant corrosion was observed in the coil used to preheat the organic feed. This corrosion
occurred in a hot but subcritical region within one to two feet of the coil entrance, indicating
considerable CH2CI2 breakdown before reaching the reactor and oxygen feed. All of this experimental
evidence suggests a global oxidation mechanism of initial CH2CI2 reaction with water to form HCI,
HCHO and/or CH3OH, and then reaction of these intermediates with O2 to form CO and CO2.
REFERENCE
1.
Tester, J.W., H.R. Holgate, F.J. Armellini, P.A. Webley, W.R. Killilea, G.T.
Hong, and H.E. Earner, Oxidation of Hazardous Organic Wastes in Supercritical
Water: A Review of Process Development and Fundamental Research, ACS
Symposium Series 518: Emerging Technologies in Hazardous Waste
Management III. D.W. Tedder and F.G. Pohland, Eds. Washington D.C. (1993).
259
-------
SOIL LEAD REMEDIATION: IS REMOVAL THE ONLY OPTION?
James A. Ryan and Pengchu Zhang
US EPA Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
(513) 569-7653
INTRODUCTION
Lead, a naturally occurring metal, has always been present in soils, surface waters and ground
waters. Lead content of agricultural soils ranges from > 1 mg/kg to 135 mg/kg with a median value of 11
mg/kg (1). Inner-city neighborhoods in most of our major cities have mean or median soil Pb
concentrations in excess of 1000 mg/kg (2-6) with values as high as 50,000 mg/kg being reported (7).
Most of these elevated lead concentrations observed in the urban soils are assumed to come from various
anthropogenic sources: industrial emissions, vehicular emissions and exterior lead paint (8). Additionally,
lead has been added to soil as the insecticide lead arsenate, impurity in fertilizers as well as from mining
and smelting activities (9). Further, lead is a contaminant of concern in about one third of the National
Priority List (NPL) sites and over 400 Superfund sites have excessive soil Pb concentrations (10). Thus,
its use by society; paints, chemical additives, tools and weapons, as well as other consumer and industrial
products, coupled with inadequate disposal or recycling by society have caused environmental systems
(soils) to become repositories for the metal. It is also apparent that not only are soils the repository for
environmentally released Pb, but it is retained in the zone of addition (9).
According to the Center for Disease Control (CDC) lead poisoning is the most common and most
devastating environmental disease affecting young children. Over the past decade the blood Pb level
associated with impairment has decreased from 25 >g/dL to 10 ng/dL, as we have learned that levels
above 10-15 u.g/dL can significantly reduce IQ and learning ability in children (11). Because of the
reduction of Pb in automotive emissions, and reduction of Pb in food due to changes in canning technology,
both food and automotive emission Pb levels have decreased nearly 10 fold in the past 15 years (12).
During the same time frame median blood Pb levels in suburban children have fallen from about 20-25
u.g/dL in 1970 to 3-4 u.g/dL in 1990. With the normal variance (and varied amounts from Pb in plumbing
systems) some suburban children exceed 15 ixg/dL. But over 50% of children in the center city exceed
15 u.g/dL limit (13). Thus, lead risk to young children is now recognized as the most sensitive limit for Pb
in the environment (11). CDC estimates that Pb poisoning in children costs billions of dollars in medical
and special education expenses and decreased future earnings. Lead paint, Pb in drinking water and Pb
in soil are the major sources of exposure. Children exposed to high levels of soil and dust Pb have been
found to have high blood Pb in numerous cases (7). Lead in soils contaminated by smelter emissions,
automotive emissions, or paint residue have been found to cause increased blood Pb in children when soil
Pb exceeds 500-1000 mg/kg (14-17). In other cases, social factors and/or soil chemical factors alter
exposure and/or bioavailability of the soil Pb and little or no increases in blood Pb are observed even with
soils containing 5000 mg Pb/kg (18). Further, Pb in mining soils appears to have lower bioavailability than
Pb in urban dusts (19-21). Cotter-Howells and Thornton (18) reported low blood Pb levels in children living
in an area with soils (about 5000 mg/kg) derived from PbS mining wastes. Studies have found the
relationship (slope of blood Pb/soil Pb) for children in smelter and urban areas to range from 1.1 to 7.6
ng/dL/1000 mg/kg, while for children in mining areas the relationship ranges from 0 to 4.8 (j.g/dL/1000
mg/kg (22). Suggesting that Pb in soils contaminated from mining activities is less bioavailable than Pb
in soil derived from urban and smelting sources. Three possible explanations have been offered for the
observations: the size of the Pb containing particle, the species of Pb in soil, and the geochemical matrix
incorporating the Pb species. These results are interpreted as indicating that because of specific
adsorption, soil Pb bioavailability increases with increasing soil Pb concentration and that the form of soil
Pb alters its bioavailability.
Programs to reduce exposure from Pb paint and Pb in drinking water are moving forward. No
program exists on contaminated soil Pb because, according to CDC and EPA, there is insufficient
260
-------
information available on which to base such a program. They state that far less is known about the
hazards of soil Pb- and how to address those hazards- than about paint or water. Thus, information is
needed to better characterize the problem, determine pathways of exposure, and determine effective
remediation methods.
Remediation treatments for soils attempt to capitalize on the differences in physical and chemical
properties between a contaminant and soil constituents. For example, remediation efforts for metal contami-
nated sites use properties such as solubility, density, particle size distribution, surface chemistry, boiling
point or magnetic susceptibility to allow separation and recovery. Metals found as relatively soluble
species or weakly sorbed to soil clays might be solubilized by the application of mild organic acids. If the
metals are present as separate mineral particulates, then their typically higher density might permit the
physical separation of these species from the less dense aluminosilicate and organic constituents of soils.
Otherwise, these forms could be bound in a solid cement or vitreous glass matrix. If the metal species are
volatile, then a soil heating methocJ might allow recovery. Separation methods relying on the magnetic
susceptibility of ferromagnetic or strongly paramagnetic metal species have also been attempted. More
detailed discussions of remediation technologies can be found in the literature (23-28).
Many biological, chemical and physical process have been proposed for soil remediation. Some
of the processes can be either applied to excavated soil or used in situ. However, reduction in exposure
to soil Pb has typically been accomplished by soil removal for off site disposal, covering, or diluting by
mixing with uncontaminated soil. Cost, logistical concerns, and regulatory requirements associated with
excavation, ex situ treatment and disposal can make in situ treatment an attractive option. Our current
understanding of Pb exposure and factors which effect its bioavailability as well as its environmental
chemistry may allow development of less costly and environmentally less disruptive methods of
remediation.
METHODOLOGY
In response to the need for cost effective technology to immobilize Pb, we have collaborated with
Ohio State University to examine the feasibility of Pb immobilization by phosphate rocks. This approach
is based on the hypothesis that Pb phosphates are the most insoluble Pb minerals, these materials are
resistant to acid weathering, and these materials are less bioavailable than other Pb forms. The
experimental approaches utilized in this research have been laboratory scale solution studies, resin
studies, dialysis studies, soil studies, and feeding studies.
RESULTS
We have shown that Pb is rapidly and effectively precipitated from solution by orthophosphate
(aqueous P, hydroxyapatite, or phosphate rock) to form a series of Pb phosphates (29-31). We have used
hydroxyapatite and phosphate rock as the primary P source and have shown that they are effective in
attenuating Pb in aqueous solution, exchangeable form and contaminated soil material, to below the U S
EPA action level of 15 /ig/L dissolved Pb. Phosphate rocks from Florida, North Carolina and Idaho are also
shown to be effective in removing Pb from aqueous solution (29). The final product of Pb immobilization
is primarily hydroxypyromorphite (Pb10(PO4)6(OH)2), which is stable even at pH as low as 3. Results of
chemical and x-ray diffraction (XRD) analysis, scanning electronic microscopy (SEM), and scanning
transmission electronic microscopy (STEM) strongly support the mechanism of dissolution of hydroxyapatite
and precipitation of hydroxypyromorphite. Aqueous P concentration is a key factor in determining the
effectiveness of Pb immobilization and the formation of hydroxypyromorphite. Thus pH is important since
it determines solubility of hydroxyapatite or phosphate rock. Hydroxyapatite or phosphate rock not only
supplies P to immobilize Pb, but also provides Ca to replace Pb from exchange sites. This is especially
important in contaminated soils and solid wastes where the bulk of the Pb is in labile rather than soluble
form.
We have also shown that hydroxyapatite can effectively immobilize aqueous Pb in the presence
261 ..
-------
of common soil solution anions: NCV, CI', F, SO/', and CO/' (30). Lead concentrations were reduced
from initial levels of 5 -100 mg/Lto below the EPA dissolved Pb action level of 15 //g/L except at very high
CO32" concentrations, whose levels are unlikely to be found in contaminated soils and wastes.
Hydroxyapatite was transformed to hydroxypyromorphite after reaction with Pb(NO3)2 in the presence of
NO3", SO42-, and CO32'; to chloropyromorphite (Pb10(PO4)6CI2) after reaction with PbCI2; and to
fluoropyromorphite (Pb10(PO4)6F2) atter reaction with PbF2, respectively. Hydroxyapatite dissolution followed
by precipitation of hydroxypyromorphite, chloropyromorphite, or fluoropyromorphite was the main process
during the reaction, but Pb adsorption by hydroxyapatite and cation substitution of Pb for Ca on
hydroxyapatite may also have occurred.
We have also investigated the effects of metals such as Zn, Cd, Ni, Cu, Fe2*, and Al on Pb
Immobilization by hydroxyapatite as well as the effectiveness of hydroxyapatite in attenuating these metals
(31). We have shown that not only do these metals have no significant effect on Pb immobilization by
hydroxyapatite at low concentrations (<20 mg/L), but also these metals themselves are removed by
hydroxyapatite. The amount of metals removed depends on the concentrations of Pb and metals, as well
as the types of metals. At higher concentrations (>20 mg/L), Cu is most effective in inhibiting Pb
immobilization by hydroxyapatite, followed by Fe2*, Cd, Zn, Al, and Ni. Hydroxypyromorphite was the only
mineral detected by XRD besides hydroxyapatite after Pb reaction with hydroxyapatite in the presence of
these metals. The amounts of hydroxypyromorphite formed decreased with an increase in metal
concentrations, according to XRD. The order of inhibition of hydroxypyromorphite formation is positively
correlated with the solubility of the other metal phosphates and supports a mechanism of competitive
precipitation.
Hydroxyapatite also removed Pb2* from Pb-EDTA solution in the presence of excess EDTA.
Indicating that basks Ca-phosphates can sequester Pb even in the presence of strongly complexing organic
ligands. In longer term experiment, hydroxyapatite was effective in immobilizing Pb2+ for up to 16 weeks,
confirming the stability of the reaction product. In another experiment, a mixture of hydroxyapatite and
hydroxypyromorphite were reacted with anion exchange resin. Lead concentrations in the suspension of
hydroxyapatite and hydroxypyromorphite were low (< 168 nmol/L) in spite of the fact that the anion
exchange resin extracted PO43" from solution, which again demonstrated the stability of
hydroxypyromorphite over hydroxyapatite. Additionally, the mixture of hydroxyapatite and
hydroxypyromorphite was also reacted with aqueous Ca2* to study the possibility of Ca2* substitution for
Pb2* on hydroxypyromorphite. Higher Ca2* concentrations resulted in slightly higher Pb2* concentrations
(< 158 nmol/L), possibly as a result of hydroxyapatite precipitation and hydroxypyromorphite dissolution.
We have illustrated that exchangeable Pb can be precipitated by reacting hydroxyapatite with a cation
exchange resin saturated with Pb. No hydroxypyromorphite was detected in the hydroxyapatite residue,
indicating the reaction did not occur in bulk solution. The resin was coated with hydroxypyromorphite
indicating that hydroxyapatite not only immobilized Pb, but supplied cations (Ca) to displace Pb from the
resin which, was immediately precipitated as hydroxypyromorphite. In further evaluations of the reaction
we mixed hydroxyapatite with several Pb contaminated soils and allowed them to react for several days.
Sequential extraction of the samples illustrated that the hydroxyapatite addition reduced the labile fractions
of soil Pb (soluble and exchangeable) and increased the residual fraction of soil Pb.
In animal feeding experiments we have illustrated that Pb bioavailability followed the order: Pb-
acetate » contaminated soil> pyromorphite = control and that the addition of apatite or rock phosphate
to the contaminated soil reduced the bioavailability of the contaminated soil Pb. Thus, illustrating that the
formation of phyromorphite in soils not only reduce the solubility of the soil Pb, but reduce its bioavailability.
In fact even without allowing time for reaction, the addition of the phosphate (apatite or rock) to the
contaminated soil was effective at reducing soil Pb bioavailability.
CONCLUSIONS
Our results strongly demonstrate that both hydroxyapatite and phosphate rocks were effective in
reducing Pb solubility and bioavailability through dissolution of hydroxyapatite or phosphate rocks and
262
-------
precipitation of pyromorphite. The effective and rapid Pb2* immobilization from solution and contaminated
soils by hydroxyapatite or phosphate rock, the limited effects from other minerals, anions, and cations, the
apparent environmental stability of the reaction products/along with the ready availability and low-cost of
hydroxyapatite or phosphate rock suggest that this approach might have great merit for cost-effective in
situ immobilization of Pb contaminated water, soils, and wastes.
REFERENCES
1. Holmgren, G.G.S., M.W. Meyer, R.L. Chaney, and R.B. Daniels. 1993. Cadmium, lead, zinc, copper,
and nickel in agricultural soils of the United States of America. Jour. Environ. Quality. 22:335-348
2. Angle, C.R., M.S. Mclntire, and A.V. Colucci. 1974. Lead in air, dustfall, soil, housedust, milk, and
water: correlation of blood lead in urban and suburban school children/Trace Substah. Environ.
Health, 8:23-29.
3. Johnson, D.E., J.B. Tillery, and R.J. Prevost 1975. Levels of platinum, palladium, and lead in
populations of Southern California. Environ. Health Presspect. 12:27-33.
4. Bornschein, R. 1986. Lead in soil in relation to blood lead levels in children. Trace Substan. Environ.
Health. 20:322-332.
5. Mielke, H.W., J.L. Adams, P.L. Reagan, and P.W. Mielke. 1989. Soil-dust lead and childhood lead
exposure as a function of city size and community traffic flow: the case for lead abatement in
Minnesota. Environ. Geochem. Health.
6. Madhaven, S., K. Rosenman, and T. Shehata. 1989. Lead in soil: recommended maximum
permissible levels. Environ. Research. 49:136-142.
7. Chaney, R.L., H.W. Mielke and S.B. Sterrett. 1989. Speciation, mobility and bioavailability of soil
lead. [Proc. Intern. Conf. Lead in Soils: Issues and Guidelines. B.E. Davies and B.G. Wixson (eds.)].
Environ. Geochem. Health 11 (Supplement):!05-129.
8. EDF, 1990. Environmental Defense Fund. "Legacy of Lead: America's Continuing Epidemic of
Childhood Poisoning". (Washington DC: EOF, March, 1990)
9. Davies, B.E. 1990. Lead. In Heavy Metals in Soils. B.J. Alloway (Ed) Blackie and Son Ltd., Glasgow.
10. Komianos, W.L. 1992. "Managing the risk of lead exposure". Environmental Protection (July/August
1992).
11. U.S. Department of Health and Human Services. 1991. Preventing Lead Poisoning in Young
Children. A statement by the centers for disease control October, 1991.
12. Bolger, P.M., C.D. Carrington, S.G. Caper and M.A. Adams. 1991. Reductions in dietary lead
exposure in the United States. Chem. Spec. Bioavail. 3:31-36.
13. Agency for Toxic Substances and Disease Registry (ATSDR). 1988. The Nature and Extent of Lead
Poisoning in Children in the United States: A Report to Congress. DHHS Doc. No. 99-2966. US Dept.
Health Human Service, Public Health Service. Atlanta, GA.
14. US CDC (Center for Disease Control). 1985. Preventing lead poisoning in young children: A
statement by the Centers for Disease Control. Jan. Atlanta, GA., No.99-2230.
263
-------
15. US-EPA (Environmental Protection Agency) 1989. OSWER Directive #9355.4-02 "Interim Guidance
on Establishing Soil Lead Cleanup Levels at Superfund Sites. Sept. 1,1989.
16. US-EPA. (Environmental Protection Agency)1986. Air Quality Criteria for Lead., EPA-600/8-83/02cF.
17. Duggan, M.J. and M.J. Inskip. 1985 Childhood exposure to lead in surface dust and soil: A
community health problem. Public Health Rev. 13:1-54.
18. Cotter-Howells, J. and I. Thornton. 1991. Sources and pathways of environmental lead to children
in a Derbyshire mining village. Environ. Geochem. Health 13:127-135.
19. Freeman, G.B., J.D. Johnson, J.M.Killinger, S.C.Liao, P.I.Feder, A.O.Davis, M.V.Rudy, R.LChaney,
S.C.Lovre, and P.D.Bergstrom. 1992. Relative bioavailabiiity of lead from mining waste soil in rats.
Fundamental and Applied Toxicology 19:388-398.
20. Davis, A., M.V. Rudy, and P.D.Bergstrom. 1992. Mineralogic controls on arsenic and lead
bfoavailability in soils from the Butte mining district, Montana.U.SA Environ. Sci. Techol 26:461-468
21. Rudy, M.V., A.Davis, J.H Kempton, J.W. Drexler, and P.O. Bergstrom. 1992. Lead bioavailabiiity:
Dissolution kinetics under simulated gastric conditions. Environ. Sci. Techol. 26:1242-1248.
22. Steele, M.J., B.D. Beck, B.L. Murphy, and H.S. Strauss. 1990. Assessing the contribution from lead
in mining wastes to blood lead. Reg. fox. Pharm., 11:158-190.
23. US-EPA. (Environmental Protection Agency)1994. Engineering Forum Issue. Considerations in
deciding to treat contaminated soils in situ. EPA/540/S-94/500
24. US-EPA. (Environmental Protection Agency)1994. A literature review summary of metal extraction
processes used to remove lead from soils. EPA/600/SR-94/006
25. Wommel, S. etal, 1993," Cleaning of contaminated soils - a treatment concept," in Contaminated Soil
'93, edited by F. Arendt, Kluwer, The Netherlands, pp. 1287-1294.
26. Pearl, M. and P. Wood, 1993, "Separation Processes for the Treatment of Contaminated Soil," in
Contaminated Soil '93, edited by F. Arendt, Kluwer, The Netherlands, pp. 1295-1304.
27. Swartzbaugh, J. etal, 1992, "Remediating Sites Contaminated with Heavy Metals, Parts I, II and Hi,"
Hazardous Materials Contra], Nov/Dec 1992.
28. Sims, R., 1990, "Soil Remediation Techniques at Uncontrolled Hazardous Waste Sites, A Critical
Review," Journal of the Air and Waste Management Association, Vol. 40, No. 5, May 1990, pp. 704-
732.
29.
30.
31.
Ma, Q.Y., S.J. Traina, T.J.Logan, and JARyan. 1993. In situ lead immobilization by apatite. Environ.
Sci. Techno!. 27:1803-1810.
Ma, Q.Y., TJ.Logan, S.J. Traina, and JARyan. 1994. Effects of NO3-, cr, F, SO/, and CO/on
Pb2* immobilization by hydroxyapatite. Environ. Sci. Technol. 28:408-418.
Ma, Q.Y., S.J. Traina, T.J.Logan, and JARyan. 1994. Effect of aqueous Al. Cd, Cu, Fe(ll), Ni, and
Zn on Pb immobilization by hydroxyapatite. Submitted to Environ. Sci. Technol.28:1219-1228
264
-------
REVIEW OF MATHEMATICAL MODELING FOR EVALUATION OF SVE APPLICATIONS
David L. Jordan1
James W. Mercer
GeoTrans, Inc.
46050 Manekin Plaza, Suite 100
Sterling, Virginia 20166
703-444-7000
INTRODUCTION
Soil vapor extraction (SVE), a demonstrated technology, enhances the removal of volatile,
chemicals from the subsurface through application of a vacuum at an extraction well to induce air flow
through the subsurface toward the well. As of 1991, SVE comprised 13% of selected remedies at
Superfund sites, and approximately 7% of leaking underground storage tanks. The flow of air
enhances volatilization of compounds from the residual NAPL phase in soil pores and from the
dissolved phase in soil pore water. The technology is particularly applicable to relatively volatile
organic compounds (Henry's law constant > 10^ atm-m3/mole) residing in the vadose zone. The
technology may also be applicable for removal of volatile light non-aqueous phase liquids (LNAPLs)
floating on the water table or entrained in the capillary fringe, if the chemicals of concern have high
vapor pressures (e.g., benzene). During SVE, contaminant removal is expected to be enhanced by
decreasing soil moisture. As percent moisture decreases, air permeability increases. Increased soil
organic carbon content will increase sorption to the soil matrix, decreasing SVE efficiency.
Heterogeneous flow conditions also affect the efficiency of contaminant removal, with higher flow
zones (preferential flow zones) cleaning up faster than low flow zones (less permeable zones).
Air sparging, another SVE-related technology, generally involves use of injection wells to inject
gas (typically air) into the saturated zone below areas of contamination. Ideally, dissolved, separate-
phase and sorbed contaminants will partition into the injected air, effectively creating an in-situ air-
stripping system. This can take place within a single-well system or the stripped contaminants can be
transported in the gas phase to the vadose zone and collected by SVE wells. The advantage of such
a system is that the treatment of groundwater and soil takes place in-situ, reducing the need for
disposal of treated material. Although air sparging is a physical/chemical treatment process, the
addition of air has the potential to promote biodegradation.
The SVE process involves installation of vacuum extraction wells or trenches at strategic
locations and depths. Air extraction can also be combined with air injection. The spacing of wells or
trenches depends on soil properties such as permeability and porosity. Where the objective is to
remove both air and water, dual vacuum extraction wells may be used. The injection wells for air
sparging can be vertical or inclined, ranging to horizontal. Effective design and prediction of system
performance can be difficult, depending on site conditions.
Tools are now available, in the form of numerical models, that allow one to both screen for the
potential feasibility of SVE, and design and estimate performance of the system. While modeling
should not be considered an end in itself, it provides a means by which to quantify some of the
important SVE operating processes. Modeling can provide estimated answers for numerous questions
concerning the feasibility and usage of SVE. Screening models can be used in conjunction with site
characterization data and best professional judgment to determine the potential feasibility of SVE at a
contaminated site. Flow and transport models can then be used to enhance the system design
1Now at Daniel B. Stephens & Associates
265
-------
process and estimate performance. The work performed as part of this effort included a review of
models that can be applied to SVE applications. This review includes a summary of critical information
required in a SVE application. It also includes a model selection process, model usage guidelines, and
case studies.
METHODOLOGY
At an "Integrated In Situ Treatment System Design Workshop" that took place on August 10 and
11, 1993 in Edison, New Jersey, a need was identified to provide environmental managers with
guidance on how models may be used to (1) determine the viability of using SVE, (2) if viable, help
design the SVE system, and (3) estimate system performance. The methodology used to provide this
guidance was a literature review and analysis of the various codes that may be applied to SVE. The
literature review, and basic information on SVE system design, are provided. This includes
introductory material, model selection tips, and example applications. In addition, information is
provided on flow and transport theory.
Applicable codes were divided into the categories of screening, air flow, and compositional flow
and transport. For each of these categories, currently available models were compiled and reviewed.
Several example applications utilizing a number of the codes are presented, along with three case
studies.
RESULTS
The results of this review is a guidance document that highlights the following topics and guide
the user through the processes of selecting and applying models to SVE sites. Technical information
Is provided in order to: (1) Determine the types of problems that can be addressed by modeling; (2)
Highlight the methods that are commonly used to solve such problems; (3) Assist potential users in
determining the presence or absence of a need for modeling at their site and, if a need is shown to
exist, selecting a model for their site; (4) Identify and illustrate the major processes governing air flow
and contaminant transport in the subsurface; (5) Present a discussion of model data needs; (6) Review
available commercial and public domain codes; and (7) Present a suite of example model applications
and case studies.
CONCLUSIONS
Modeling can provide estimated answers for numerous questions concerning the feasibility and
usage of SVE. Screening models can be used in conjunction with site characterization data and best
professional judgment to determine if SVE at a contaminated site is feasible. Flow and transport
models can be used to enhance the system design process and estimate performance. In some
cases, no complex model is necessary, and decisions can be made based on simple analytical
solutions and/or best professional judgment. Geographical information systems (GIS) can provide
valuable assistance in organizing and presenting site data graphically in order to enhance the remedial
alternative selection process.
Table 1 presents a summary of the screening, air flow, and compositional flow and transport
codes that were evaluated. For screening, these models include the Hyperventilate and VENTING
codes, as well as other analytical solutions. Air flow models available at this time include AIRFLOW,
CSUGAS, and AIR3D. For compositional flow and transport, the VENT2D/VENT3D model is available
and capable of simulating contaminant transport and removal via SVE.
The selection and application of any model will ultimately lie with the model user. This
document attempts to guide the potential model user through a decision-making process that is
266
-------
intended to help decide how and when to select a model, to make users aware of the processes
governing flow and transport in the vadose zone, and highlight the limitations of model results.
FOR MORE INFORMATION
Chi-Yuan Fan
Superfund Technology Demonstration Division
Risk Reduction Engineering Laboratory
Edison, New Jersey 08837
908-906-6924
267
-------
15
M
1
p
Hardware/Softwar
Requirements
CO
Q
113
CO
CD
1 Advantag
15
O
CD
CL
"55
CO
CO
£
LU
Available from
O
o
CL
2
eg
o
t*=
L.
'co
I
?**
CO
C
Provides rapid
Calculates ail
O)
_c
CD
CD
0
CO
iL
§.
=
CO
CM
O
EPA/600/R-93/
(EPA ORD
Publications,
513-569-7562)
Compatible:
80386/80387
c
Q
1*3
_3
s
CD
£
it-
CD
estimates for
determination
co"
permeability,
well flow rate
*s
CO ^^
"H o
coprocessor or
80486, 4 MB RAM,
CO
3
CD LU
Lf} ^^
Q
— 'co
3 CD
o -o
co 2
•s
^,
•— •
15
Q.CO
1
8 w
k* E
CO -
E ^
^-xO
o^^:
_O
L<^
LU jo
DOS 3.1 or higher,
Microsoft Windows
CO
E
CD
**-*
1
Provides rapid
removal from
several
i
°> c
S-'5
Q- co
Ss
JD
]§
'ra
^
CO
a.
151
O
.E IcT
estimates of
contaminant
idealized
diffusion-
CD
o>
OT|
^^ C
S w
§?
O -
E S-
o o
£ 0
CO
Q.
5
0
c 'S
.— CO
concentrations
extracted gas,
limited
scenarios
O
1 £
-.53
JJ
CL
s
CO
concentration
0)
co
CO
>,
over time for
1
u.
p
multiple
constituents
o
o
<
IS
J^ T-
•-s«
II
t£.
o c
Q. O
I
o
°. o
(N O
ri cp
II
A5
tn o
w ^
D w
o'Q
D. £
s^
CO c
I
268
-------
Availability
£
Hardware/Softwe
Requirements
CO
0
3
E
_i
8?
1
3
Capabilities
o
to
15
^^
6
^
<
CD
CO
.1
I
Calculates
Hydrogeologic
Software
m
compatible,
80386/80486, 4 U
1
.0
CO
0)
%
2£
-§.5
M
US
^.E
pressure
distribution in
Q
|
CO
1
O)
o
o
esi
8
Q
1
O
Q.
UJ
1
CO
CO
f—
o
HJ
simplifies mod
CO
'•o
2
CO
(RR#2)
Bolton, Ontario^
i_
higher, mouse an
math coprocessor
D.
1
•o
CO
•3'
QL
domain,
calculates air
CO
a:
CO c^
CD ^^
i C^ ^^
' CO
|§^
c!l£
80386-based
machines
recommended
i_
o
D)
•D :§ c
ro CD
u_
o
•'
269
-------
-i
'
.TO .2
tn
ra
oo
>£
to o 10
CO
ro .E
2
o> e o c 5 o
QQLUODLLcoQ.
0)
8
in
O
CD
CO
O
(U
0)
-2- ™ o) c
Q -^ ro ™
O)
CN
CO CO
'co E
>» tu
ro "o
ro
ra
£
£.2 -a
CD
.
'•S Q-
0) p 4r! o
UJ .&•.£ 'to
cp
ra
Q.
8
co
o>
m O
•g _ -43
ro E 3
"
ro
o
ro .S!i
O > =5
to CD
CO
0)
_
- -
o ro
8 -,1
•tj 2 s
tn ^— —^
8
ro
O
8.
CD
CD
CO
o
,
o:
270
-------
z <
CO
£
Q
CO
1
1
CO
^,-
15
J9
'co
>
<
1
Q)
E
.2
1
a:
co
.0
•4—*
a
E
CO
0)
CD
(0
1
15
a
CO
O
0)
Q.
0)
"O
o
c
§
0)
m
^
_-g
>
D
6
0
DL
S
CO
2
13
1
CD
.N
•c
(D
0}
.Q
CO
1
I
6
Calculates
1
0=
-^
1
LU
•55
0)
CO ^
*; m
5 °°
g O)
CO Z
0 o
•* S
s£
I
oo
^^
*= "co
(0 p
Q. c
II
CO
o '^
^"x.
CO
CO
CM o
x .E
in CD
CM ^2
> CD
is
"TO t:
c: o
O Q.
'4= CO
II
E-o
8 ro
c
pressure
distribution ii
t
1
I 1
Q
1
LLJ
s
5x1
cp
CM
0
8
0
-
o
CO
8
2
Q.
c
o
2
5
•s
2
£
73
.0
CO
1
0)
1
o
0
10
O5
f>
CD
.0
CL
m
CN
m
i_r
0)
CD
Ic
o
o
CO
0)
£
o
fc
15
CO
1
1
constituent
.22
*-•
B-
|
a.
CO
concentratiol
CD
£
Q.
CO
CD
jQ
i
CD
1
0
CO
=3
8
the subsurfa
271
-------
REPRESENTATIVE SAMPLING AND ANALYSIS OF HETEROGENEOUS SOILS
Ann Leitzinger
U. S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7635
Joseph Evans
Science Applications International Corporation
545 Shoup Avenue
Idaho Falls, ID 83402
(208) 528-2168
Sam Hayes
Science Applications International Corporation
Suite 403
635 W. Seventh Street
Cincinnati, OH 45203
(513) 723-2600
Ken Partymiller
PRC Environmental Management, Inc.
5326 Paris Pike
Georgetown, KY 40324
(502) 867-1397
Gregory Swanson
S-Cubed
3398 Carmel Mountain Road
San Diego, CA 92121
(619) 587-7221
INTRODUCTION
Sampling and analysis presents specific problems with heterogeneous
soils. These problems have been carefully evaluated in several projects for
the Risk Reduction Engineering Laboratory (RREL), specifically Superfund
Innovative Technology Evaluation (SITE) demonstrations. Of particular concern
is the inability of standard analytical methods to account for heterogeneous
samples. While a typical analytical method may normally accommodate only 2
grams (g) of sample (and therefore requires that "oversized material" be
removed from a sample before analysis, potentially biasing analytical
results), a representative sample aliquot from a heterogeneous site may
require a sample size that is orders of magnitude greater than this. The
inability to analyze representative samples may be the result of various
factors, including the presence of oversized material (such as rocks or other
debris); highly concentrated contaminants (such as lead chunks or tar balls),
which may preclude taking large laboratory samples that would overwhelm
instrument capabilities; or volatile contaminants which cannot be mixed to
272
-------
create a more homogeneous aliquot because of the volatility of the contaminant
being investigated. This paper contains several case studies that present the
problems associated with sampling and analyzing heterogeneous soils;
approaches to specific problems encountered in each of the projects discussed;
and the relative success of these approaches.
Approaches previously recommended for selected RREL projects include the
following: (1) screening pre-treated soil and then treating and analyzing the
screened fraction; (2) grinding large amounts of material, followed by
homogenization and analysis of a representative aliquot; (3) screening,
sieving, and then analyzing separate portions of a sample; (4) removing
oversized material from a sample, analyzing the sample portion that remains,
and correcting for the percentage of oversized material in the sample; and (5)
analyzing larger sample sizes. While one approach may be valid for a
particular site or type of contamination, the case studies presented below
emphasize how particular approaches were tailored to the perceived problems.
For example, heterogeneous soils may include large rocks, (which are usually
relatively free of contamination) or tar balls, which are highly concentrated
forms of organic contamination. Excluding either of these materials may bias
sample results. Typical laboratory glassware or the unavailability of
concentrated 1aboratory spikes can also present problems particularly when
analyzing larger sample sizes. Grinding and sieving samples may lead to loss
of contaminants through volatilization or may release contamination from the
inside of the oversized material, thereby producing biased results. In
addition, there may be physical limitations to sampling based on the sample
type. For example, oily samples or samples containing metallic lead may not
mix, sieve, or grind well.
Data comparability is also an important consideration when sampling.
Therefore, it may be more important to use a consistent approach from one
event to the next, rather than to use a more accurate technique at a
subsequent event. Another consideration, when sampling, is project cost
limitations which may prevent the analysis of numerous samples necessary to
fully characterize the contamination at a specific site. The examples
presented below illustrate approaches used by RREL investigators to alleviate
specific sampling and analysis problems associated with heterogeneous soils.
METHODOLOGY
A variety of sources were used during the preparation of this paper. A
literature search was performed and previously published documents were
reviewed. In addition, individuals considered knowledgeable in the area of
sampling were contacted to provide further information. Quality Assurance
Project Plans for RREL projects were also surveyed from the last five years to
identify projects for the accompanying case studies.
The case studies below are structured as follows. First, a brief
project description is presented with a discussion of site characteristics,
the technology, the nature of the problem, the options considered, and the
option chosen. The results of the chosen option are then presented, with the
last section of each case study presenting the lessons learned for that
particular site.
273
-------
CASE STUDY 1 - RADIO FREQUENCY (RF) HEATING
The RF Heating technology uses electromagnetic energy in the RF band to
heat contaminated soil. The soil is heated by providing power to an array of
antennae located in drilled bore holes within the soil. Petroleum
hydrocarbon, volatile, and semivolatile contaminants present are volatilized,
collected, and treated. Soil moisture is also volatilized to provide a steam
sweep inside the treatment zone. The organic vapors and moisture are
recovered by applying a vacuum to vapor extraction wells near the antennae.
Particle size distribution (PSD) results indicated that the top 4.9
meters (m) of soil consisted of 25 to 48 percent gravel by weight, and that
between 5.5 m and 9.1 m below the ground surface, the soil consisted primarily
of gravel (64%-86%).
Due to the large amount of gravel, considerations were given to the
manner in which a representative sample might be analyzed. The options
considered included: (1) grinding and homogenizing; (2) screening and
analyzing the different portions; and (3) removing the rocks and analyzing the
portion that remained, with subsequent correction of the results based on PSD
results for each sample. The third option was chosen. The analytical results
were evaluated both with correction and without. This comparison indicated
that the chosen method did not change the site evaluation.
CASE STUDY 2 - BIOGENESIS SOIL WASHING TECHNOLOGY
The BioGenesis soil washing technology uses a proprietary solution to
transfer organic contaminants from a soil matrix to a liquid phase. The
technology uses a specially designed wash unit that can treat contaminated
soil at rates of up to 10 metric tons per hour. After washing, soil is stored
in containers, and the liquid waste stream is either treated or recycled. The
combination of the surfactant that remains on the soil and the water left over
from washing enhances microbial activity, which biodegrades any residual
contamination.
The critical parameter measured in soil samples was petroleum
hydrocarbons. To determine the removal efficiency of petroleum hydrocarbons,
multiple waste streams were sampled and analyzed, including feed soils and
treated soils. The presence of large rocks and tar balls prompted questions
about the homogeneity of the soils and the representativeness of the sampling
process. As a result, the sampling staff began subjectively excluding both
rocks and tar balls from pre- and post-treatment samples. In addition, the
means for excluding tar balls differed between pre- and post-treatment
samples, which introduced further sampling bias.
To address the above concerns, the first samples were recollected by
sampling large amounts of soil (346 kilograms). This entire amount of soil
was screened through a 12.7 millimeter (mm) diameter screen; separated by hand
into soil, tar ball, and rock fractions; and weighed to determine the percent
composition of each. An aliquot from each fraction was analyzed. For rocks,
100 g to 200 g aliquots were obtained before rinsing the rocks with solvent.
A 1 g aliquot of the tar ball material was also obtained and dissolved in
274
-------
freon. The solvent extract from both matrices was analyzed separately for
petroleum hydrocarbons. The analytical data was then mathematically
recombined to yield sample results. Results showed that the presence of rocks
and tar balls were offsetting factors and did not affect the petroleum
hydrocarbon data during the demonstration.
CASE STUDY 3 - TORONTO HARBOUR SOIL RECYCLING DEMONSTRATION
The Toronto Harbour demonstration is a three-phase operation consisting
of a soil washer (used to separate the soil into different sized sample
streams), a chelating process (for removal of metals), and a bioslurry process
(to treat organics remaining in the fine material). SVOCs, metals, and oil
and grease were the critical parameters for this project.
Two concerns, discovered in the field, were associated with
representative sampling. First, the feed material consisted of a significant
amount of rock and contained high concentrations of organics. Secondly, the
oversize sample stream from the soil washer consisted of rocks ranging from
about 6 mm to 25 mm in diameter.
To address these concerns, the feed material sample size was increased
to 100 g to improve representativeness. Samples were then split by the
laboratory into soil and rock fractions using a 9.5 mm sieve, and analyzed
separately. This allowed the soil fraction to be homogenized and analyzed by
SW-846 Method 3540 (soxhlet extraction) using 2 g aliquots. The rock fraction
obtained from the 100 g aliquot was also analyzed by Method 3540 to ensure
soil clinging to the outside of the rocks would be fully included.
For the oversize sample stream, the original plan was to grind and
homogenize the samples. This was rejected as the grinding of these samples
would expose surfaces of the rock that would not otherwise be exposed for
extraction thus creating a non-representative sample. It was decided that a
200 g aliquot would be obtained for the laboratory and the entire sample would
be extracted and analyzed for organics. Any organic contaminants on the
surface of the rocks could then easily be extracted by using a solvent wash.
This allowed large quantities to be extracted without the usual limitations
associated with laboratory glassware. For both the oversized fraction and the
feed material, several sample aliquots had similar analytical results
indicating samples were representative.
CASE STUDY 4 - BESCORP SOIL WASHING PROCESS
The Bescorp process uses a combination of soil washing, size
segregation, and density separation to remove heavy metals from soils. This
demonstration was carried out at a previous lead-acid battery reclamation
site. Native soil consisted of approximately 50% gravel (>6.4 mm), 40% sand
(6.4 mm to 0.1 mm), and 10% fines (<0.1 mm). Lead tended to be most
concentrated in the fine fraction, but the coarser fractions also contained
significant fractions of lead as battery posts, lead plates, and lead
materials attached to battery casings.
275
-------
This project represented a serious sampling and analytical challenge in
that it required sampling a heterogeneous soil for lead that was present both
as lead chunks and as fines. The primary challenge was measuring total lead
in the coarse fractions. The standard laboratory digestion procedure digests
only a 1 g aliquot, which is unsuitable for samples containing pieces of rock
and lead that can weigh several grams each.
For this project, a 1 kilogram (kg) sample was first dried and then
screened into four weighed fractions: 64 mm to 6.4 mm, 6.4 mm to 2 mm, 2 mm
to 0.1 mm, and < 0.1 mm. Each fraction was examined visually for pieces of
lead, which were removed and weighed. The two larger fractions were then
placed in a polyethylene bottle and digested with 6% HN03 overnight on a
rotating apparatus, while 2 g aliquots of the two smaller fractions were
placed in beakers and digested with SW-846 Method 3550 and analyzed. The
masses measured were then combined to arrive at an average concentration for
the entire sample.
The approach described above was satisfactory for the sand and fine
fractions, but marginal for the larger fractions due to the intermittent
occurrence of lead chunks. A larger initial sample size (e.g., 50 to 100 kg)
could have alleviated this problem.
CASE STUDY 5 - THE COGNIS INC. (COGNIS) TERRAMET™ LEAD EXTRACTION PROCESS
The COGNIS Inc. TerraMet™ lead extraction process uses chemical
leaching to remove metals bound to sands and fines (including silts and
clays). Sands and fines are separated before leaching, and particulate lead
is collected for recycling by density separation. Lead present in the liquid
phase is precipitated for recovery, and the leaching solution is recycled.
Samples were collected from the feed soil and several sand and fine
streams before and after leaching. Obtaining representative samples of some
of the coarse-grained soil streams became an issue due to the presence of lead
shards in the samples. Grinding soil samples prior to analysis was determined
to be the most appropriate method of obtaining representative samples.
Originally, soils were to be ground finely enough to pass through a 0.11 mm
sieve. However, based on time and cost evaluations, it was decided that this
approach would not be applicable in this situation. Samples from the Cognis
demonstration are presently being analyzed to determine the most appropriate
methods of sample preparation and analyses.
CASE STUDY 6 - BIOVENTING FOR ENHANCED BIODEGRADATION
The Bioventing process is an in-situ biological process that increases
aerobic degradation of organic materials in soil. Specifically, several wells
are drilled into the contaminated soil plot and air is pumped into the wells
for a period of several years to increase aerobic activity and thereby
stimulate biodegradation. In-situ soil samples are taken at pre- and post-
demonstration periods to determine removal efficiency for the process.
While at the site, it was observed that soil cores contained what
appeared to be a significant number of rocks and creosol balls, which were
276
-------
later referred to as cinder chunks. The rocks and cinder chunks were
separated from the soil and saved, the soil samples that remained were then
homogenized in the field and sent to the laboratory for analysis as per
standard procedures. The total volume of the rocks and cinder chunks was
determined for the entire site. A ratio was then calculated for the volume of
rocks and cinders in comparison to the total soil volume at the site. This
was based upon estimations of soil core volumes and the number of soil cores
collected. A representative sample of the rocks and cinder chunks was then
analyzed by the laboratory. It was determined that no significant bias
existed by excluding these materials in the analyses. Thus, the analysis of a
rock/cinder sample was necessary, but the number of analyses was minimized,
with the results obtained from the soil analyses verified as being
representative of the entire site.
CONCLUSIONS
RREL investigators perform a variety of research projects and frequently
encounter unusual sampling situations. The following conclusions have been
drawn from heterogeneous soil sampling efforts associated with RREL projects:
• No single approach for sampling and analyzing heterogeneous soils will
be applicable for all sites or technologies.
• Multiple approaches for sampling and analyzing heterogeneous soils may
need to be used at different sampling locations for a specific
technology.
• New approaches for sampling and analyzing heterogeneous material must
continue to be developed.
• The risk of collecting unusable data is reduced.by developing an
adequate sampling approach before an actual sampling episode. Before
.selecting an approach, investigators should review the site
characteristics, the technology, the nature of the problem, and the
approaches available. Once an approach is selected, investigators
should review its success and document the outcome.
In light of the above conclusions, RREL will continue to handle the
sampling of heterogeneous soils on a case-by-case basis, while drawing on the
knowledge base it has developed. RREL anticipates maintaining a database of
heterogeneous sampling approaches, and their results, so that RREL
investigators can improve their future planning decisions.
FOR MORE INFORMATION CONTACT: Ann Leitzinger
Technical Project Manager
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7635
277
-------
BIOREMEDIATION OF TRINTTROTOLULENE BY A RUMINAL MICROORGANISM
Taejin Lee, Graduate Research Assistant
Department of Civil Engineering
Dr. Kenneth J. Williamson, Professor . '
Department of Civil Engineering
(503)737-6836
Dr. A. Morrie Craig, Professor
College of Veterinary Medicine
(503)737-4282
Oregon State University
Corvallis , Oregon 97331-2303
INTRODUCTION
2,4,6-trinitrotoluene (TNT) has been widely used for the production of explosives because
of its low boiling point, high stability, low impact sensitivity, and safe manufacture. TNT is known to
produce adverse health effects from occupational exposure including increased incidences of
aplastic anemia, liver damage, dermatitis, ocular disorders, and gastrointestinal distress. In
addition, TNT is of ecological concern based on its toxicity to certain aquatic organisms. More
than 1,100 military facilities, each potentially contaminated with munitions waste, are expected to
require treatment of more than one million cubic yards of contaminated soils. The cost associated
with remediation of these sites has been estimated to be in excess of $1.5 billion.
Recently, researchers have studied ruminal microorganisms in relation to their ability to
degrade xenobiotic compounds. Many of these organisms are strict anaerobes with optimal
redox potentials as low as -420 mV . Ruminal organisms have been shown capable of destroying
some pesticides, such as parathion, p-nitrophenol, and biphenyl-type compounds; thiono isomers,
(8, 33); and nitrogen-containing heterocyclic plant toxins such as the pyrrolizidine alkaloids (5).
Many of these compounds have structures similar to TNT.
A TNT-degrading ruminal microorganism has been isolated from goat rumen fluid with
successive enrichments on triaminotoluene (TAT) and TNT. The isolate, designated G.8, utilizes
nitrate and lactate as the primary energy source. G.8 was able to tolerate and metabolite levels of
TNT up to the saturation point of 125 mg/1.
Objectives
Based upon the results of previous studies, the present research was directed at
understanding the'process of TNT degradation by the denitrifying ruminal microorganism G.8 (5).
The objectives of this study were:. 1) to identify specific metabolites and define TNT destruction
pathways and 2) to understand TNT and the metabolites transformation mechanisms on the
different primary electron acceptors.
METHODOLOGY
Growth Medium
The media used to incubate the isolate in serum bottles consisted of (mg/1) CI^CHOHCOONa
(500), MnSO4.H2O (8.5) FeSO4.7H2O (10), KNO3 (2000), ZnSO4.7H2O (5), CaCl2.2H20 (24),
Na2HPO4.H2O (2550), H3BO3 (1.5), NaMo4.2H2O (1.5), NaH2PO4.H2O (975), CoCl2,.6H2O (0.6),
278
-------
Na2EDTA (21.5), yeast extract (100), CuCl2.2H2O (0.05), MgSO4.7H2O (30), and NiCl2.6H2O (1).
The nutrients were mixed and well-stirred. pH was adjusted to 7.0 with 0.1 N NaOH, then boiled for
3 minutes under argon gas flow and dispensed into serum bottles, which were then stoppered and
autoclaved for 25 minutes. To prevent the precipitation of salts from the medium, 0.3 ml of 800
ppm CaCl2'2H2O was injected following autoclaving. An 0.2 % inoculum was used for all
experiments. Serum bottles were cultivated anaerobically at 37°C for given periods of time using a
mechanical shaker. Ethanol was used to prevent contamination at the contact points for the
inoculant and samples, and sample aliquots were indexed and frozen at -14°C.
Analytical Methods
TNT, 4-amino,2,6-dinitrotoluene (4A26DNT), 2,6-dinitrotoluene (26DNT), 2-amino,6-
nitrotoluene (2A6NT), and 2-nitrotoluene (2NT) were measured by high performance liquid
chromatography (HPLC),. Gas chromatography/mass spectrometry (GC/MS) analysis was
performed using a Hewlett-Packard Model 5988A connected to a Model 5890 gas chromatography
(GC) with a XTI-5 fused-silica capillary column. A Dionex 4000i ion chromatography (1C) with a
HPIC-CS3 anion column was used to measure anions such as nitrate and nitrite. Culture turbidity
in the serum tubes was measured using a spectrophotometer at 660 r\m absorbance (Bausch and
Lomb, Spectronic 20D) for cell-growth analysis.
Microorganism
A TNT degrader was isolated from rumen fluid with successive enrichment of TNT and
TAT.. The isolated microorganisms from a fistulated goat was named G.8. The gram negative G.8
was identified based fatty acid analysis as a Escherichia Coli with similarity index (SI) of 91.4
percents.
RESULTS
f-
G.8 growth was monitored by optical density using different energy sources including
nitrates, nitrites, or TNT to investigate the relation of TNT to G.8 metabolism. G.8 growth occurred
in the presence of only nitrate, but not on nitrite, or TNT, indicating that nitrate serves as a viable
terminal electron acceptor. Fermentation was not observed with the presence of lactate only in the
medium. G.8 growth was detected in the absence of lactate (TNT and nitrate present), suggesting
that TNT could have been potentially used as the carbon source and electron donor. G.8 growth
was observed in the medium without lactate (only nitrate present) as much as in the medium
containing TNT and lactate, probably from the use of the yeast extract as the electron donor and
carbon source.
Additional experiments were conducted to investigate the influence of TNT on the
utilization of primary energy source with G.8 growth. Nitrate was converted to nitrite completely
within 10 hours, followed by further reduction of the nitrite. Nitrate was converted to nitrite and then
less transformation of the nitrite in the medium was observed when TNT is present.
In the preliminary experiments,G.8 growth was not observed with nitrite or TNT as the
primary electron acceptor, indicating G.8 does not initially utilize the nitrite or TNT. When TNT was
absent from the medium, further transformation of .the nitrite from nitrate was observed and less
transformation of the nitrite with presence of TNT in the medium. Thiese results demonstrate that
nitrite or TNT conversions takes place in the presence of nitrate. It suppose that the nitrite and TNT
appear to compete as cometabolic electron acceptors, with the TNT being preferrentially used.
G.8 Experiments to Determine Metabolic Pathways
Based on evidence of co-metabolic TNT transformation, we decided to attempt to identify
the TNT biotransformation pathways and limiting factors. TNT and each of its transformation
279
-------
products were incubated individually with a G.8 medium containing lactate as the electron donor
and nitrate or nitrite as the primary electron acceptor. A medium containing lactate and TNT
metabolites was also prepared.
TNT Transformation
G.8 was incubated for three days in serum bottles containing 20 mg/I of TNT. From Bottle
A (containing lactate and nitrate as primary energy sources), the metabolites were identified as
4A26DNT. The complete TNT transformation in Bottle A and the non-transformations of TNT hi
Bottles B (containing lactate and nitrite as primary energy sources) and C (containing only lactate)
showed that TNT is co-metabolically degraded in the presence of nitrate, and not degraded hi the
presence of nitrite.
4A26DNT Transformation
G.8 was incubated for eight days in serum bottles containing 20 mg/I of 4A26DNT. In
Bottle B, 4A26DNT was reoxidized to TNT and deaminated to 26DNT. In Bottle A, only negligible
amounts of TNT and 26DNT were detected. No 4A26DNT transformation occurred in Bottle C.
Nitrites hi the medium appeared to stimulate the G.8 reduction and oxidation systems.
26DNT Transformation
In the same manner described for the 4A26DNT incubations, 26DNT was incubated for 20
days with G.8 in Bottles A, B, and C. The transformation of 26DNT to 2A6NT was not detected or
was negligible in Bottles A and B. The reduction of the 26DNT ortho-position nitro group was
detected hi bottle C, suggesting that G.8 utilized 26DNT as a primary electron acceptor.
2A6NT Transformation
During 18 days of incubation, 2A6NT in Bottle B was transformed to 2NT and 26DNT.
Oxidation and reduction of 2A6NT trends were identical to those for the 4A26DNT incubation. The
absence of a primary electron acceptor in Bottle C did not stimulate 2A6NT transformation, and the
oxidation of 2A6NT was faster than the reduction processes.
2NT and 2AT Transformations
During 18 days of incubation, 2NT metabolites were not detected from either HPLC or
GC/MS analysis, but it was noteworthy that 2-aminotoluene (2AT) completely disappeared from
Bottles A and B, suggesting that 2AT transformed more rapidly than 2NT. It is reasonable to
suppose that the 2AT transformation products converted rapidly to other metabolites, but that the
transformation of 2NT to 2AT was too slow. The 2AT transformation products were identified as o-
cresol, toluene, and small amounts of 2NT. These trends for amino group transformation were
identical to those for 4A26DNT and 2A6NT.
TNT transformation pathways were established with series connections of each transfor-
mation products presented above. Proposed pathways for the TNT degradation are shown hi
Figure 1. The TNT biotransformation steps shown are the reduction process, wherein the nitro
group is reduced to an amino group, followed by deaminization and oxidation when the amino
group was present. As observed previously, para-nitro group reduction was the most susceptible
transformation .
280
-------
NO2
OH
Figure 1. Proposed pathways for TNT biotransformation with G.8 incubation. The pathways are
established with series connection of the identified transformation products by GC/MS and HPLC,
281
-------
CONCLUSIONS
The G.8 isolate was capable of reducing the TNT nitro group to an amino group and was also involved in
deaminization as a co-metabolite, resulting in nitrogen free compounds such as toluene or o-cresol.
The reduction (deaminization) and oxidation (hydroxylation) reactions took place simultaneously when the
amino group was present From previous studies, it had been reported that oxidation and reduction processes
occurred jointly subject to anaerobic conditions. Vogel et al. (4) demonstrated that Cytochrome P450 could mediate
both the oxidation and reduction reactions for degradation of halogenated aliphatic compounds. For Pseudomonas
sp. strain PN-1, it was also found that oxidation steps proceeded under conditions of denitrification (1). CO2 and
chloroform formations from carbon tetrachloride in the culture of Escherichia Coli K-12 was found by Griddle et
al. (1). These mechanisms are coincident with the present mechanisms of G.8.
Degradation of compounds is dependent upon the characteristics of the parent compounds, the microbial
consortium, and environment factors (3). In the experiments described, the patterns of TNT metabolite
transformation were dependent on the type of electron acceptors. The presence of nitrates in the medium stimulated
the reduction of para-positioned nitro groups, and nitrites stimulated the deaminization and hydroxylation processes.
The absence of such primary energy sources as nitrates or nitrites stimulated the reduction of die ortho-positioned
nitro groups.
It is assumed that the incomplete TNT transformations are related to inadequate combinations of the terminal
electron acceptors. Nitrate consumption in the medium was complete while TNT was transformed to 4A26DNT and
26DNT. For certain periods of time, nitrite accumulated in the medium does not support G.8 metabolism and subject
to termination of 26DNT reduction process, as was indicated at 26DNT transformations with presence of nitrite in
the medium.
From these findings, it is concluded that an appropriately designed sequencing reactor system, or other
alternatives to control such energy source as nitrates, or nitrites in the TNT degradation system, could result in
obtaining the full transformation of TNT to o-cresol or toluene.
REFERENCES
1. Berry, D. F., A. J. Francis, and J. M Bollag. 1987. Microbial Metabolism of Homocydic and Heterocyclic
Aromatic Compounds under Anaerobic Conditions. Microbiol. Rev., 51:43-59.
2. Griddle, C. S., J. T. Dewitt, and P. L. McCarty. 1990. Reductive Dechlorination of Carbon Tetrachloride
by Eschfricfaa coU K-12. Appl. Environ. Microbiol., 56:3247-3254
3. Nicholson, D. K., S. L. Woods, J. D. Istok, and D. C. Peek 1992. Reductive Dechlorination of
Chlorophenols by a Pentachlorophenol-Acclimated Metanogenic Consortium. Appl. Environ. Microbiol. 58:
2280-2286.
4. Vogel, T. M., S. C. Griddle, and P. L. McCarty. 1987. Transformation of Halogenated Aliphatic
Compounds. Environ. Sci. Techonol. 21:722-736
5. Wachenheim, D. E., L.L. Blythe, and A. M. Craig. 1992. Characterization of Rumen Bacterial Pyrrolizidine
Alkaloid Biotransformation in Ruminants of Various Species. Vet Hum. Toxicol. 34:513-517.
282
-------
FULL-SCALE IN SITU BIOREMEDIATION OFTRICHLOROETHYLENE IN GROUNDWATER:
PRELIMINARY MODELING STUDIES
Mark N. Goltz, Brett T. Kawakami, and Perry L. McCarty
Western Region Hazardous Substance Research Center
Department of Civil Engineering
Stanford University
Stanford, CA 94305-4020
Phone:415-723-8574
Fax:415-725-3164
INTRODUCTION
A full-scale study of in-situ bioremediation is being planned for implementation at Edwards Air Force
Base. The bioremediation system that is being proposed has been developed over 8 years of research
and testing in the laboratory and at a pilot field site located at Moffett Naval Air Station in Mountain View,
California. Studies conducted at the Moffett field site have demonstrated that trichloroethylene (TCE),
the contaminant found at Edwards, can be effectively biodegraded cometabolically through the
introduction into the subsurface of a primary substrate (such as phenol or toluene) and an oxygen source
(such as hydrogen peroxide) to support the growth and energy requirements of a native population of
microorganisms (1,2,3).
One of the main questions that needs to be answered, prior to full-scale demonstration of this
technology on the Edwards TCE plume, is how best to mix a primary substrate, an oxygen source, and
TCE, and subsequently get the mixture to the microorganisms. At Moffett Reid, mixing of these three
components was accomplished above ground, with the mixture then introduced into the subsurface
through an injection well. In the full-scale demonstration, the TCE will, of course, already be in the
groundwater. A major objective of the demonstration will be to investigate how a primary substrate and
an oxygen source can be efficiently mixed and transported to indigenous microorganisms, in order to
promote cometabolic degradation of TCE. For the demonstration, it is anticipated that a subsurface
recirculation system, similar to that described by Herrling (4) and McCarty and Semprini (5) will be used.
The proposed remediation system is shown in Figure 1. The system consists of a pair of recirculation
wells, each screened at two depths. In operation, a submersible pump installed between the two screens
of each well would draw TCE contaminated water into the well at one of the screened intervals. The
primary substrate (toluene) and oxygen source (hydrogen peroxide) will then be introduced into the well
through feed lines, and the water, which contains TCE, toluene, and oxygen, will be discharged into the
aquifer from the second screened interval. An in-situ treatment zone will be created in the aquifer, around
the discharge screen. The second treatment well will extract water from the injection zone of the first
well, and inject water at the level where the first well is extracting water. Thus, water will circulate through
the aquifer, from one treatment well to the other. The injection/extraction levels of each well can be
periodically reversed.
METHODOLOGY
To help plan and design the demonstration, a computer model which incorporates all the significant
flow and transport processes was used. A two-dimensional flow code was used, to simulate pumping
through the two treatment wells. The streamtubes which were output by the flow code were used as input
for the transport code. The one-dimensional transport code (6,7) was applied along each streamtube.
The transport code includes the microbial processes of bacterial growth, toluene and oxygen utilization,
and the cometabolic transformation of TCE coupled with the transport processes of advection, dispersion,
and rate-limited sorption onto aquifer solids. Model simulations could be run for different parameter
values, with TCE concentrations output as a function of space and time. In addition, TCE concentrations
may be numerically integrated over space, to obtain plots of TCE mass versus time.
283
-------
i
85
fcl
L
•c
Q.
if
I
•u
pi
-------
RESULTS
Based on preliminary data obtained from the Edwards demonstration site and the Moffett Reid
experimental site, and using "best-guess" design values for the treatment system, a base line simulation
was run. Base line parameter values are listed in Table 1.
TABLE 1. PARAMETER VALUES USED FOR BASELINE SIMULATION OF BIOREMEDIATION MODEL
Parameter
Distance between treatment wells
Flow rate through each treatment well
Treatment well screen length
Concentration of oxygen source (hydrogen
peroxide) added
Time-averaged concentration of primary
substrate (toluene) added
Toluene pulse length
Initial TCE concentration in aquifer
Sorption distribution coefficient ,
Sorption/desorption rate coefficient
TCE retardation factor
Cometabolism rate constant
TCE half-saturation coefficient '
Value
10.0 meters
15.0 gallons per minute
4.0 meters
100 milligrams per liter (mg/L)
9 mg/L
20 minutes every 8 hours
0.2 mg/L
0.2 milliliter per gram
0.010 per daV:
2.0 •-'!, -"/
0.01 5 per day
0.5 mg/L
The simulation shows how TCE mass declines in both the dissolved and sorbed phases over the four
months planned for the demonstration (figure 2). Notice that initially, the mass of TCE in the dissolved
and sorbed phases is equal. This is due to the fact that the retardation factor used for the simulation is
2.0, and TCE concentrations in the dissolved and aqueous phases are initially assumed to be at
equilibrium. Over approximately the first 50 days of the simulation, dissolved mass drops relatively
rapidly, due to biodegradation of TCE in the aqueous phase. After about 70 days, the mass of TCE in the
aqueous phase becomes relatively constant, while the mass in the sorbed phase shows a slow but steady
decline. What has happened is that after about 70 days, slow desorption has begun to control the rate of
mass removal, A pseudo-steady state has been attained where the rate of biodegradation of dissolved
TCE equals the rate at which TCE desorbs from the aquifer solids into the aqueous phase. Thus, based
on the assumed parameter values, we see that the decline in total mass after the initial 70 days of
bioremediation is controlled by the rate of TCE desorption from aquifer solids. A critical value in the
simulations is thus the sorption/desorptjon rate coefficient, a value which was estimated. Because of its
importance, this value needs to be determined experimentally in the laboratory for the Edwards site. The
simulation also demonstrates that using preliminary design values, a significant drop in TCE mass may be
expected to be seen in the demonstration area over the course of the four month demonstration. Figure 3
shows aqueous and sorbed TCE concentrations between the two treatment wells along one of the
streamtubes after 120 days of bioremediation. Note how the aqueous concentration drops rapidly near
the injection Well, and then slowly rises. The rapid drop results since most of the microbial mass is
predicted by the model to reside close to the injection well. This prediction is consistent with the results
from the Moffett Reid studies. The slow rise in TCE concentrations is due to the relatively clean water
moving through the aquifer towards the extraction vyell. The clean water causes sorbed TCE to desorb,
thus leading to the concentration rise in the aqueous phase. Also note that the TCE concentrations after
120 days are well below the initial values of 0.2 mg/L in the dissolved phase and 0.04 mg/kg in the sorbed
phase. ; - • • •'/ ;"- •--' - •••;- - - - • •'-•'•'•
CONCLUSIONS ••'•;.
Based on reasonable parameter values and results of prior work at an experimental field site,
preliminary modeling studies have been used to show that the planned operation of an in-situ aerobic
cometabolic bioremediation system can be expected to result in observable decreases in TCE mass and
concentration at the proposed technology demonstration site.
285
-------
Initial Total Mass
0 10 20 30 40 50 60 70 80 90 100110120
Time (days)
Figure 2. TCE mass vs. time
0.040
0.000
0 5 10 15 20 25 30 35
Distance from Injection Well (m)
Figure 3. Dissolved and sorbed TCE concentration vs. distance
between wells
286
-------
REFERENCES
1. Hopkins, G. D., Semprini, L, and McCarty, P. L. Microcosm and In Situ Field Studies of Enhanced
Biotransformation of Trichloroethylene by Phenol-Utilizing Microorganisms. Applied and Environmental
Microbiology. 59(7): 2277-2285.1993.
2. Hopkins, G. D., Munakata, J., Semprini, L., and McCarty, P. L. Trichloroethylene Concentration
Effects on Pilot Reid-Scale In-Situ Groundwater Bioremediation by Phenol-Oxidizing Microorganisms.
Environmental Science and Technology. 27(12): 2542-2547.1993.
3. Hopkins, G. D. and McCarty, P. L. Field Evaluation of In Situ Aerobic Cometabolism of
Trichloroethylene and Three Dichloroethylene Isomers Using Phenol and Toluene as the Primary
Substrates. Submitted Environmental Science and Technology. 1994.
4. Herrling, B. Hydraulic circulation system for in situ bioreclamation and/or in situ.remediation of
strippable contamination. In: R.E. Hinchee and R.F. Olfenbuttel (eds.), On-Site Bioreclamation.
Butterworth-Heinemann, Boston, Massachusetts, 1991. pp. 173-175.
5. McCarty, P.L. and Semprini, L. Ground-water Treatment for Chlorinated Solvents. ]n: R.D. Norriset
al. (eds.), Handbook of Bioremediation. Lewis Publishers, Boca Raton, Florida, 1993. pp. 87-116.
6. Lang, M. M. Design and Optimization of In Situ Bioremediation Relying on Cometabolic Degradation.
Ph.D. Dissertation, Stanford University, Stanford, CA, 1995. 185pp.
7. Semprini, L. and McCarty, P. L. Comparison Between Model Simulations and Reid Results for In Situ
Biorestoration of Chlorinated Aliphatics: Part 2. Cometabolic Transformations. Ground Water. 30(1): 37-
44,1992.
FOR MORE INFORMATION
Mark N. Goltz
Western Region Hazardous Substance Research Center
Department of Civil Engineering
Stanford University
Stanford, CA 94305-4020
Phone:415-723-8574
Fax:415-725-3164 , ' •
Email: goltz@cive.stanford.edu
287
-------
A HICROBIOLOGAL SURROGATE FOR EVALUATING TREATMENT EFFICIENCY
Eugene W. Rice, Kim W. Fox, Richard J. Miltner,
Darren A. Lytle and Clifford H. Johnson
Drinking Water Research Division
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
INTRODUCTION
In this study we report on the use of a microbial surrogate system which
can be used to evaluate the efficiency of various unit processes used in
drinking water treatment for the removal of microbial contaminants. The
proposed procedure uses Gram-positive, mesophilic, aerobic spore-forming
bacteria as the surrogate organisms. These bacteria do not pose a public
health threat and are naturally occurring in most surface water supplies. The
aerobic spore-formers are easy to culture and are present throughout the
treatment train. This group of organisms consists primarily of species of the
genus Bacillus. These organisms form endospores which are ellipsoidal to
spherical in shape and measure on average approximately 0.5 X 1.0 X 1.5
micrometers, and are environmentally resistant. Like pathogenic Giardia cysts
and Cryptospon'dium oocysts endospores of aerobic bacilli may be found far
into the treatment train.
MATERIALS & METHODS
Both naturally occurring aerobic spore-forming bacteria and pure
cultures of Bacillus subtilis spores were used in this study. B. subtilis
spores were purchased from a commercial laboratory (Raven Biological
Laboratories, Omaha, Nebraska 68106). The protocol for the enumeration of the
aerobic spores was a modification of a procedure used for the detection of
Bacillus spores in raw milk (1). The following equipment is used in the
protocol:
• Sterile Erlenmeyer flasks with screw-caps or
sterile bottles with closures.
• Thermostatically controlled water bath
with shaker for constant agitation
Thermometer
Sterile pipettes
60 x 15 mm Petri dishes with loose lids ;
47 mm, 0.45 fan porosity membranes
Membrane filtration apparatus
Incubator, 35°C
The medium used for growing the bacteria consisted of nutrient agar plus
the dye trypan blue. The dye was added to aid in the visualization of the
colonies.
288
-------
The formulation for the medium is:
• Nutrient Agar
Peptone 5 g
Beef extract 3 g
Agar 15 g
• Trypan blue dye 0.015 g
• Distilled water 1,000ml
Heat to boiling dissolve
completely.
Sterilize in the autoclave
for 15 minutes at 15 psi (121°C)
pH 6.8 ± 0.2 at 25°C
The spore-forming bacteria were enumerated using the membrane filter , .,
method. The following steps outline the experimental procedure:
• Place sample in sterile flask and cover
with cap
• Prepare pilot flask with same volume
as sample and thermometer
• Place sample flask(s) and pilot flask in
water bath at 82°C and agitate throughout
heat treatment
• Lower temperature of bath to 80°C when
flask(s) reaches 79°C
• When contents of Flask(s) reach 80°C, keep
samples in bath for an additional 12 min.
• Immediately cool sample flask(s) in ice bath
• Membrane filter (0.45 p m pore size) appropriate
volumes and place membrane on medium
• Incubate plates 20-22 hr at 35°C
• Count colonies
Heating.the sample to 80°C inactivates vegetative bacterial cells but
does not inactivate spores. During aerobic incubation at 35°C the spores
germinate and are counted as bacterial colonies.
Total coliform bacteria were enumerated by the membrane filter procedure
using m-Endo LES agar, and heterotrophic plate count (HPC) analyses were
conducted by the spread plate procedure using R2A agar with incubation
of plates at 28°C for 5-7 days. Cultures of B. subtilis spores were used in
the bench-scale coagulation studies (jar tests). Standard procedures were
used to conduct the jar test experiments (3). An electronic particle counter,
HIAC/ROYCO, model 9064 (Pacific Scientific Co., Montclair, Calif.) was used
for particle sizing and counting analysis.
RESULTS & DISCUSSION
Levels of indigenous aerobic spores found in the Ohio River over a ten
month period are listed in Table 1. These results are typical of the numbers
of aerobic spore-formers which would normally be encountered in a surface
water supply.
289
-------
Table 1.
Levels of Indigenous Spores
in
Ohio River Water
1994 '
Month CFU/100 ml
January
February
March
April
May
June
July
August
September
October
7,500
31,000
11,000
27,000
5,800
3,700
3,900
310,000
2,600
1,800
290
-------
Table 2 lists the mean log removals of several microbial parameters and
total particles (>1 /jm) and particles in the size range of 3.1-7 tm in a pilot
conventional drinking water treatment plant sampled daily over seven days.
This table lists the log removals achieved through the settling and filtration
processes. The spores were the only microbial parameter which could be found
throughout the treatment train and their removals closely paralleled both
total particle counts and counts in the 3.;l-7 /wn size range. Table 2 also
indicates that the spores, like Giardia and Cryptospon'dium, are much more
resistant than coliforms or HPC to chlorination.
Log removals of spores and particles in a full-scale water treatment
plant which utilizes Ohio River water are listed in Table 3. These data
collected over three seasons again show the parallel between particle counts
and levels of indigenous spores.
CONCLUSIONS
Indigenous aerobic bacillus spores can serve as surrogates for particle
removal and potentially for Giardia and Cryptospon'dium removals. Generally,
optimal removals were achieved with spores when optimal removal efficiencies
were achieved for turbidity and particle removals. The aerobic spore forming
bacteria are ubiquitous in nature and thus can be used to evaluate treatment
without adding potential surrogates to the source water. Since these bacteria
and their spores are primarily soil organisms their numbers increase with
increasing amounts of run-off. Unlike the general heterotrophic plate count
parameter, these organisms would not propagate in the treatment plant (e.g. as
the result of increased AOC after ozonation) and are more resistant to
disinfection. Also, unlike inanimate particles, which are counted by
electronic particle counters, the spores would not be created or broken up
while passing through various unit processes. While these organisms have no
public health significance and are not indicators of fecal contamination, they
may serve as useful indicators of treatment plant efficiency.
REFERENCES
1. American Public Health Assoc. 1985. Standard Methods for the Examination
Dairy Products. 15th ed. Am. Pub. Health Assoc. Washington, D.C.
2. American Public Health Assoc. 1992. Standard Methods for the Examination
of Water and Wastewater. 18th ed. Am. Pub. Health Assoc. Washington,
D.C.
291
-------
Table 2.
Removals During Pilot Study
Parameter
Total col i forms
HPC
Spores
3.1-7 /im particles
Total particles
Log10
Settled
0.47
0.38
0.85
1.11
0.91
Removal
Filtered
Chlorinated
>1.79*
>4.58*
2.11
2.10
1.70
* Below detection limit:
Total coliforms 1 cfu/100 ml
HPC 1 cfu/ml
Table 3.
Seasonal Removal - Ohio WTP
Log10 Removal from Raw Water
Spores 3-6 /zm Total
, Particles Particles
Winter
settled
finished
Spring
settled
finished
Summer
settled
finished
1.35
2.86
1.51
2.80
2.11
2.08
1.62
3.52
1.08
2.22
0.61
2.09
1.37
3.26
1.47
2.50
0.73
1.97
292
-------
EVALUATION OF MEMBRANE PERFORMANCE AND FOULING BY
PYROLYSIS-GC/MS
Thomas F. Speth
USEPA ' ' ;'
26 W Martin Luther King Dr.'
Cincinnati, OH 45268
(513)569-7208
William R. Fromme '
and R. Scott Summers
University of Cincinnati
Cincinnati, OH 45221
INTRODUCTION
The loss of membrane efficiency due to fouling is one of the main impediments to the
development of membrane processes for use in drinking water treatment. Membrane fouling is
dependent on the water quality and the membrane properties and construction. In general, fouling is
defined as the accumulation of material on the surface, or in the pores, of a membrane that decreases
the water flux through the membrane. The consequences of fouling can be severe, fouling can reduce
the water flux through a membrane up to 90 percent (1). There are five broad fouling categories:
sparingly soluble inorganics, colloidal or particulate matter, dissolved organics, chemical reactants, and
microorganisms.
Dissolved organics and colloidal matter are considered to be the most serious of the foulants
due to the difficulty in removing them with pretreatment processes. Membranes can be fouled either by
cake-layer formation or adsorption. Cake-layer formation refers to a layer of material on top of the
surface of the membrane. Cake-layer fouling can be removed hydrodynamically. Adsorption refers to
materials attached within the pores of the membrane. This type of fouling is generally considered
irreversible although chemical cleaning can be effective in some cases. The nature of the organic
molecules that foul nanofiltration membranes is not known, knowledge of the characteristics of the
natural organic matter responsible for membrane fouling may lead to optimal membrane and
pretreatment selection. Improved nanofiltration performance will lead to an increased use of
nanofiltration as a drinking water treatment method.
Pyrolysis-gas chromatography/mass spectroscopy (Py-GC/MS) has the ability to elucidate
natural organic matter characteristics. Pyrolysis is used to thermally cleave organic molecules into
volatile fragments that are then separated by gas chromatography and identified by mass spectroscopy.
The Py-GC/MS method used herein was developed by Bruchet et_aJL_(2) at the Lyonnaise des Eaux in
Paris, France. Bruchet's method categorizes natural organic matter in terms of polysaccharides,
proteins, polyhydroxyaromatics, and aminosugars. The fractions are calculated relative to a known
standard mixture of dextran, bovine serum albumin, cellulose acetate, and Fluka humic acids. A simple
empirical algorithm was developed for computing relative peak areas among the various classes of
biopolymers (2).
The objective of this research is to relate the productivity of membrane systems that treat
different waters to natural organic characteristics as determined by Py-GC/MS. Membrane studies will
be completed with a cellulose acetate membrane that has a nominal molecular-weight cutoff of 500
Daltons (Amicon YC05), and a thin-film polyamide membrane (Fluid Systems) that has a nominal
molecular-weight cutoff of 200 Daltons. Waters from the Ohio River (southern Ohio), East Fork Lake
(southern Ohio), and Manatee County (Florida) were used. The three waters have very different
293
-------
characteristics. The Manatee County water (MCW) has a very high dissolved organic carbon (DOC)
concentration (near 20 mg/L). It also has very low alkalinity (12 mg/L as CaCO3) and hardness (29
mg/L as CaCO3). Py-GC/MS results indicate that the Manatee County water is very aromatic in nature.
The Ohio River water (ORW) is low in DOC (approximately 2.5 mg/L). It is moderately hard (110 mg/L
as CaCO3), and has a moderate alkalinity (56 mg/L as CaCO3). Preliminary Py-GC/MS results suggest
that it has aromatic and proteinaceous character. The East Fork Lake water (EFLW) has a moderate
DOC (approximately 5 mg/L), and has the highest hardness (126 mg/L as CaCO3) and alkalinity (100
mg/L as CaCO3) of the three waters. The EFLW Py-GC/MS results indicate a polysaccharide nature.
METHODOLOGY
The membrane experiments were conducted in a 450 ml dead-end pressure vessel that was
stirred. The membrane experiments were run until the flux decreased to approximately 50 percent of
the initial flux, or until a steady flux was obtained. At the end of each experiment, the concentrate was
collected and rotoevaporated to a concentration that allowed for adequate organics to be spiked into
50u,l pyrolysis tubes. Likewise, the feed and the permeate streams were rotoevaporated to a similar
concentration. The membrane was then turned over and approximately 100 ml of organic-free water
was used to backflush the membrane. Finally, the membrane was turned back over and a short run
with organic-free water was completed to check the flux recovered after backflushing. The backflush
sample was also rotoevaporated to the same concentration as the other samples. The Manatee water
experiments had a layer of material on top of the membrane that did not come off with backflushing.
This material was physically scraped off and analyzed.
The DOC, total dissolved solids (TDS), and ultraviolet @254nm (UV) of the feed, concentrate,
and numerous permeate samples were assessed. These results allowed for a mass balance to be
completed over the membrane experiment. Also, small aliquots of the rotoevaporated samples were
taken and diluted into 100 ml of organic-free water. DOC, TDS, and UV of the diluted samples were
assessed. This was completed to determine the volume of the samples to be spiked into the pyrolysis
tubes to obtain the same total DOC in all of the tubes (130 ng). It also allowed for determination of the
amount of organics that adhered to inorganic precipitation during rotoevaporation. Rotoevaporation
losses of organics may be important with regard to the pyrolysis results if the organics removed were
not similar in make-up to the organics that remained in solution.
Each Py-GC/MS sample was run in triplicate. The pyrolysis tubes were weighed before and
after pyrolysis to confirm that similar sample sizes were being used, and to show that similar amounts
of the samples were being pyrolyzed. The three waters being used were each concentrated on two
separate occasions to confirm that each water's characteristics were not changing over time.
After the fouling of the membranes, a small (O.Smg) sample of each of the fouled membranes
was pyrolyzed. The membrane sample was measured on a balance sensitive to 0.01 mg. As with the
previously discussed samples, each Py-GC/MS sample was run in triplicate. Clean membranes were
also pyrolyzed in triplicate to determine the background pyrograms.
The analytical procedure for Py-GC/MS entailed ramping the pyrolysis sample from 35°C to
750°C at 20°C/ms in a heated-filament pyrolysis probe. The final temperature was held constant for 20
seconds. Splitless injection to the GC was used. The GC was ramped from 35°C to 260°C at a rate of
2°C/min. A 60-meter Supelco WAX-10 column was used.
RESULTS
Table 1 contains the flux decline results for the matrix of experiments performed in this study.
The Manatee County water showed the greatest flux reduction for the Amicon membrane experiments.
294
-------
The East Fork Lake water showed the greatest flux reduction for the Fluid System membrane
experiments. The reason for this behavior is made clear by electron microscope analysis. Electron
microscopy of the fouled Amicon membranes showed no inorganic deposits. This indicates that the
fouling agents were organic in nature. Electron microscopy of the Fluid System membranes for the
EFLW and ORW experiments, however, showed large deposits of calcium (most probably calcium
carbonate). The Manatee County water showed no inorganic fouling with the Fluid System membrane.
TABLE 1. FLUX DECLINE RESULTS FOR MATRIX OF EXPERIMENTS
Membrane/Water
Liters of Water Passed
Percentage Flux Decline
Amicon/Manatee County Water
Amicon/Ohio River Water
Amicon/East Fork Lake Water
Fluid System/Manatee County Water
Fluid System/Ohio River Water
Fluid System/East Fork Lake Water .
7.8
21.5
18.6
6.8
6.8
3.9
35
30
15
55
60
68
As mentioned above, the Amicon membranes were fouled by organics. Therefore, it is not
surprising that the Manatee County water, which had the highest DOC concentration, showed the
greatest amount of fouling. The organic nature of the foulant was also confirmed by the UV, DOC, and
TDS balances of the membrane run. The Amicon membrane had a large enough molecular-weight
cutoff to pass inorganic ions while rejecting most of the organic molecules.
The Fluid System membrane rejected both organics and a great deal of the inorganic ions.
This was shown by UV, DOC, and TDS mass balances over the membrane experiments. This was
expected because of the Fluid System membrane's designation as a "softening" membrane. The East
Fork Lake water had the greatest concentration of hardness and alkalinity, and hence the greatest
amount of inorganic fouling. The Fluid-System/MCW experiment showed a lack of inorganic fouling as
mentioned above. This was due to the water's extremely low hardness and alkalinity. This was visually
confirmed by noticing that the final concentrate solutions of the Fluid-System/EFLW and the Fluid-
System/ORW experiments were salty in appearance, whereas the Fluid-System/MCW concentrate was
clear.
Py-GC/MS of the Manatee County water showed that the water is primarily polyhydroxyaromatic
in nature. Polyhydroxyaromatics accounted for over half of the organics. Proteins, polysaccharides,
and aminosugars accounted for a much lower percentage of the total. The pyrolysis results for the feed
water, concentrate, permeate, and membrane scrapings showed that the concentrate had a marked
decrease in polysaccharides as compared to the feed water. The pyrolysis results of the other fractions
were roughly similar to the feed-water results.
Figure 1 contains the mass balance for the Amicon/MCW experiment. The mass balance
shows that the biopolymer groups were primarily found in the concentrate stream as expected due to
the high organic rejection. The permeate, backflush, and membrane scrapings contained small
amounts of all the biopolymers. The mass balance closed very well for the protein,
polyhydroxyaromatic, and aminosugar groups. However, a large amount of the polysaccharides fed to
the system was not accounted for. When the membrane was pyrolyzed, the unaccounted for
polysaccharides were found. Figure 1 also shows the amount of polysaccharides found in the
membrane. This additional amount of polysaccharides closes the mass balance reasonably well.
Therefore, these results suggest that polysaccharide material was irreversibly adsorbed (operationally
defined) to the membrane, and was responsible for the membrane fouling.
295
-------
100
•o
o>
3
0
80 -
60 -
20 -
a>
i "I
O O
s o
CO
a
O CB o
a. m to
<3
SJ.
g>
-5.
S
1
«
I I
SJ.
o>
POLYSACCHARIDE
PROTEIN
I 1 g
a. m to
AROMATIC
O
O 0.
£ &
AMINOSUGAR
Rgure 1. Pyrolysis-GC/MS mass-balance results for Amicon-YC05 membrane run with
Manatee County water.
The Amicon/MCW results agree with the work of Mallevialle et al. (3) who showed that
polysaccharides were one of the fouling agents for ultrafiltration membranes. Mallevialle et al. (3)
indicated that inorganics also were present in the fouling mechanism. In this work, however, the
electron microscopy results do not show an inorganic presence.
The Fluid-System/MCW results show that all of the Bruchet et al. (2) groups were rejected, and
thereby remained in the concentrate. However, the mass balance indicates that significant portions of
all four groups remained unaccounted for. Co-elution problems with membrane pyrolysis fragments
obscured the membrane pyrolysis results for polysaccharides and polyhydroxyaromatics. Therefore, it
is unknown as to whether they were irreversibly adsorbed to the membrane.
The EFLW results are unique. For the Amicon experiment, all four Bruchet et al. (2) groups
were rejected by the membrane and thereby remained in the concentrate solutions. The mass
balances suggest that no significant arnounts of organics were adsorbed to the membrane. Direct
pyrolysis of the membrane showed a similar result. The very low flux drop over the course of the
Arnteon/EFLW experiment also suggests that the membrane was not significantly fouled. The
membrane run was stopped because TDS, UV, and DOC broke through the membrane at elevated
levels. The lack of fouling was surprising because of the high polysaccharide concentration in the East
Fork Lake water.
The Fluid System/EFLW pyrolysis results suggest that proteinaceous material was adsorbed
onto the membrane. The percentage of flux reduction that was caused by the proteinaceous material in
relation to that caused by the above mentioned inorganic foulant is unknown. The combined mixture of
inorganic and organic foulants corresponds to the results of Mallevialle et al. (3).
296
-------
The ORW results are also unique. The concentrates for both the Amicon-YC05 experiment and
the Fluid System experiment show high concentrations of polysaccharides. The pyrolysis mass
balances, however, give inaccurate results. The inaccurate results suggest that the high ratio of salts
to organics affected the pyrolysis results. The ORW concentrate solutions had the greatest amount of
organics removed by salt precipitation during rotoevaporation.
The varied fouling results for the different waters and different membranes suggest that it may
be possible to choose a pretreatment process that delays flux reduction. For inorganic contaminants,
acid addition or antiscalent agents can be used, as they are currently. For organic foulants, Py-GC/MS
can identify the major foulant and a pretreatment process could be selected to remove the foulant.
Also, if membrane properties can be related to fouling agents, the choice of the optimal membrane can
be made from analysis of the water.
CONCLUSIONS
The cellulose-acetate-based Amicon-YC05 membrane was fouled to the greatest extent by the
Manatee County water which had high organic levels. The polysaccharide fraction of the Manatee
County water was determined to be responsible based on a mass balance over the membrane
experiment and by direct evaluation of the fouled membranes. The East Fork Lake water did not foul
the Amicon-YC05 membrane severely. The pyrolysis-GC/MS results for the East Fork Lake experiment
confirmed that organics were not adsorbing onto the membrane. The Ohio River water results were
inconclusive, probably due to the presence of salts. Electron microscopy results showed that there was
no inorganic fouling of the Amicon membranes by any of the waters.
The polyamide-based Fluid-System membrane was fouled to the greatest extent by the East
Fork Lake water which had the highest hardness and alkalinity. Calcium carbonate precipitation was
deemed to be the causative fouling agent by electron microscopy for both the East Fork Lake water
and Ohio River water. The East Fork Lake water pyrolysis results suggest that proteins also adsorbed
to the membrane with the inorganic foulants. The Ohio River water pyrolysis results were inconclusive
probably due to the presence of salts. The Manatee County membrane experiment showed no
inorganic fouling, The pyrolysis-GC/MS results for the Manatee County water experiment suggest that
all four biopolymer classes may be involved in the observed fouling.
REFERENCES
1. van den Berg, G.B. and Smolders, C.A. Flux Decline in Membrane Processes,
Filtration and Separation. 2. 115-121. 1988.
2. Bruchet, A., Rousseau, C., and Mallevialle, J. Pyrolysis-GC-MS for Investigating High-
Molecular Weight THM Precursors and Other Refractory Organics, Jour. Amer. Water Works
Assoc.. 82(9): 66-74. 1990.
3. Mallevialle, J., Anselme, C., and Marsigny, O. Aquatic Humic Substances: Influence on Fate
and Treatment of Pollutants, Suffet, IH and P MacCarthy (eds), ACS Press, 1987, pp. 749-767.
FOR MORE INFORMATION
Thomas F. Speth
US Environmental Protection Agency
26W Martin Luther King Drive
Cincinnati, OH 45268
(513) 569-7208
297
-------
BIOFILTRATTON FOR CONTROL OF VOLATILE ORGANIC COMPOUNDS (VOCS1
Dolloff F. Bishop
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
Rakesh Govind
Department of Chemical Engineering
University of Cincinnati
Cincinnati, Ohio 45221
INTRODUCTION
Air biofiltration is a promising technology for control of air emissions of biodegradable volatile organic
compounds (VOCs). In conjunction with vacuum extraction of soils or air stripping of ground water, it can be
used to mineralize VOCs removed from contaminated soil or groundwater. The literature (1) describes three
major biological systems for treating contaminated air bioscrubbers, biotrickling filters and biofilters.
Bioscrubbers, which are not evaluated here, use counter current gas-liquid spray columns with microorganisms
freely suspended in the aqueous phase. Biofilters and biotrickling filters use microbial populations in biofilms
immobilized on support media to degrade or transform contaminants in air. Filter media can be classified as:
bioactive fine or irregular particulates, such as soil, peat, compost or mixtures of these materials; pelletized,
which are randomly packed in a bed; and structured, such as monoliths with defined or variable passage size
and geometry. The media can be made of sorbing and nonadsorbing materials. Nonbioactive pelletized and
structured media require recycled solutions of nutrients and buffer for efficient microbial activity and are thus
called biotrickling filters. Filters with bioactive fine or irregular particulates as media, referred to as biofilters,
usually do not recycle solutions of nutrients and buffers to prevent media compaction and gas channeling. All
filters humidify the contaminated air before biotreatment.
Soil biofilters are relatively large because soil pores are smaller and compounds have low permeability
in soil. They also have limited bed depths, required for maintaining humidity in soil and minimizing pressure
drop. Peat/compost biofilters, used commercially, are suitable for treating large volumes of air containing
biodegradable VOCs at low concentrations (<200 ppmv). However, both soil and peat/compost biofilters are
susceptible to channeling and maldistribution of air and require periodic media replacement.
Extensive work has been conducted to improve biofiltration by EPA's Risk Reduction Engineering
Laboratory and the University of Cincinnati in biofilters using pelletized and structured media and improved
operational approaches. Representative VOCs in these studies included compounds with a range of aqueous
solubilities and octanol-water partition coefficients. The compounds include iso-pentane, toluene, methylene
chloride, trichloroethylene (TCE), ethyl benzene, chlorobenzene and perchloroethylene (PCE) and alpha (a-)
pinene. Comparative studies were conducted with peat/compost biofilters using isopentane and a-pinene.
Control studies were also conducted to investigate adsorption/desorption of contaminants on various media using
mercuric chloride solution to insure the absence of bioactivity.
MATERIALS AND METHODS
The typical experimental set-up of a biofilter bed (Figure 1) packed with support media was operated at
steady-state conditions to characterize process performance. Contaminated air, synthesized by injecting VOC
stock solutions into a controlled air stream, was passed through the biofilter bed. The inlet air was humidified
298
-------
-Simple Point
Blofflter
Sample point;
Flow
Controller!
Syringe
Pump
(for
Substrate)
Rotameter
— Sample Point
Air
Sample Point
Nutrient
Solution
pump
Figure 1. Typical experimental set-up of biofilter system.
299
-------
before contamination. In the case of pelletized and structured media, solutions of nutrients and buffers were
introduced at the top of the bed and allowed to trickle countercurrent to the air flow through the media. The
solutions were collected at the bottom of the bed and recycled with appropriate additions for nitrogen
consumption and pH control. The water in the solutions insured effective wetting of the media.
Initially the pelletized and structured geometry (straight-passages) media filters were seeded with
acclimated biomass from an activated sludge process. Attachment or encapsulation of the seed on the various
media was successful. Garden variety peat humus (2204.6 gms) was thoroughly mixed with foam fluff (144.8
gms) providing a peat mixture with high void fraction. Similarly, compost (1358.5 gms) and foam fluff (145.5
gms) were mixed thoroughly for the compost column studies. Foam fluff was made by shredding polystyrene
foam into pieces with an average size of 2 mm. Control studies conducted with foam fluff showed that no
VOC adsorption occurred on the foam material. Contaminated air for all studies was prepared by injecting
appropriate VOC stock solutions into the humidified air flow controlled by a mass flow controller.
Gas samples to characterize biofilter performance were collected from the inlet and outlet ports of the
biofilter using gas tight syringes. Samples were collected in 250 mL glass sampling tubes by diverting the gas
flow through the tubes for 1 hour. The 250 microliter samples were injected into a gas chromatograph (Tracer
585 GC, with 703 PID 10.0 ev lamp, 1000 Hall BCD, Tekmar LSC 2000 purge and trap concentrator). Liquid
samples (5 mL) were collected in syringes with luer locks for purge and trap analysis.
In evaluating mineralization of the chlorinated VOCs, 100 mL samples of the effluent nutrient solution
were taken and mixed with 2 mL of 5 M sodium nitrate to adjust the ionic strength, and analyzed for chloride
ion with an Orion solid state combination electrode (9617 BN) in an Accumet 1003 pH/mV/ISE meter. The pH
of the effluent nutrient solutions was measured by the Accumet 1003 pH/mV/ISE meter. Carbon dioxide
concentrations in the inlet and outlet gas streams were determined in a Fisher 1200 gas partitioner.
RESULTS
The studies evaluated peat and compost biofilters, pelletized (activated carbon, ceramic and
encapsulated biomass) biofilters and structured geometry (ceramic and carbon coated, ceramic straight-passages)
biofilters. Biofilter performances, expressed in terms of removal efficiencies of the compounds were calculated
from the amount of compound removed per unit time in the biofilter, expressed as a percentage of the amount
of that compound entering the biofilter per unit time.
Peat and compost biofilters:
Studies of peat and compost biofilters treating slightly soluble iso-pentane at high (360-960 ppmv)
concentrations in humidified air were completed using the indigenous microorganisms and nutrients of each
media to establish bioactivity. Initial abiotic control tests on peat and compost revealed poor adsorption of iso-
pentane. The effluent air content of iso-pentane became equal to the influent concentration (350 ppmv) in about
30 minutes.
Pseudo-equilibrium removal efficiencies were established for iso-pentane by operating at selected air
retention times (2 minutes to 13 minutes) until the increasing removal efficiencies reached steady removal
values. Equilibrium removal values, achieved in one to three weeks of operation at each retention time,
increased with decreasing influent iso-pentane concentration and increasing gas residence time. The compost
biofilter exhibited substantially higher removal efficiencies at the 360 ppmv iso-pentane concentration than the
peat biofilter. At higher influent concentrations, the two materials exhibited roughly similar performances.
Water content in peat and compost effected removal efficiencies. Performance studies, with varying
water content in the media, revealed critical water contents of 0.48 gm of water/gm of peat and 0.58 gm of
water/gm of compost, below which biofilter performance deteriorated substantially. The loss in removal
300
-------
efficiencies below the critical water content were caused by irreversible shrinkage of the media permitting gas
by-passing. Subsequent addition of water did not expand the media nor eliminate the cracks and voids.
Removal of iso-pentane exhibited optimal removal efficiencies at 0.56 gm of water/gm of peat and 0.67 gm of
water/gm of compost. As the water content increased above the optimal value, removal efficiencies gradually
decreased for both media. These decreasing efficiencies probably resulted from partial filling of bed void
fraction with free water and increased mass transfer difficulties for the slightly soluble iso-pentane. The losses
of efficiency, above and below the optimal water content, illustrated the need for maintaining appropriate media
water content through effective humidification of influent air entering peat and compost biofilters.
Studies of temperature effects on removal efficiency in fully humidified air revealed, for both media,
increasing efficiencies with increasing temperatures and maximum removal at 35°C. Below 25°C, the removal
efficiencies decreased nearly linearly with temperature. Thus for influent air temperatures below 25 °C,
substantial improvement in biofilter performance could be achieved by heating the air to increase bed
temperature.
A study using a-pinene was also conducted in a compost biofilter to assess the effects of pH on media
with limited buffering capacity. Monounsaturated a-pinene, a major contaminant in press vent and dryer gas, is
a slowly degradable complex bridged ring compound (2) and biodegrades through a variety of pathways,
producing organic acids and neutral compounds that accumulate as intermediate products. The study revealed
that a compost biofilter with an air retention time of 3.5 minutes supported nearly complete removals and
mineralization of the a-pinene at an inlet concentration of 25 ppmv. At inlet concentrations of 50 ppmv and
above, however, the removal efficiencies and mineralization sharply decreased. Tests on mixtures of the
compost and water revealed that organic acids were accumulating in the compost bed, limiting degradation and
indicating a need for additional buffer capacity in the compost for inlet a-pinene concentrations above 25 ppmv.
Biofilters with pelletized media
Two biofilters, one packed with 3 mm activated carbon pellets (void fraction .0.35), the other with 6
mm porous ceramic (celite) pellets (void fraction 0.40) were studied to characterize performance of biofilters
randomly packed with pellets and using nutrient recycle. The research also revealed performance differences
and similarities between adsorptive and non-adsorptive media. The activated carbon biofilter, operated at 2
minutes of gas retention time and a nutrient solution loading of 1000 L/m2day, initially revealed high adsorptive
removals of the inlet contaminants; toluene (520 ppmv), methylene chloride (180 ppmv), and TCE (25 ppmv)
followed by a rapid decrease in contaminant removal as the wetted carbon became saturated with contaminants.
In this first biofilter study, the initial buffered nutrient solution had insufficient ammonia which limited biomass
growth and substantially prolonged start-up time by delaying biomass growth. Indeed, the carbon media
fortuitously became saturated with adsorbed VOC contaminants, especially with toluene. Subsequent increase of
ammonia in the nutrient solution accelerated biofilm development. , Removal efficiencies increased to nearly
100% for all three contaminants. The buffered nutrient also maintained the liquid phase pH above 6.2. The
activated carbon biofilter continued to effectively remove (>99%) of the contaminants for an additional 3
months before excessive pressure losses and flooding occurred.
The start-up of the ceramic celite biofilter with appropriate nutrients at a nutrient solution loading of
250L/m2day and a 2 minute air retention time revealed gradual increases in contaminant removals as the seeded
biomass grew on the celite pellets. After about 40 days, sufficient biofilm growth occurred to substantially
degrade (>95%) four of the five inlet contaminants; toluene (450 ppmv), methylene chloride (150 ppmv), ethyl
benzene (25 ppmv) and chlorobenzene (40 ppmv). In contrast to the performance of the activated carbon
biofilter, the fifth inlet contaminant, TCE (25 ppmv) was only partially removed by the bioactivity in the celite
biofilter.
Folsom et at. (3) showed that TCE is usually degraded only as a secondary substrate in the presence of
a primary metabolite, such as toluene or phenol. Examination of contaminant concentration profiles in the celite
301
-------
biofilter revealed that toluene was rapidly removed in the bottom third of the biofilter. As a result toluene was
not available as a primary metabolite in the upper two-thirds of the filter. Thus the slowly degrading TCE was
degraded only in the bottom third of the column.
In the activated carbon biofilter, toluene should also rapidly degrade, probably in the bottom third of
the biofilter. Apparently the adsorbed toluene from the prolonged start-up of the activated carbon biofilter acted
as a toluene reservoir desorbing toluene to the attached biofilm on the carbon over the full height of the biofilter
bed. The desorbing toluene produced complete degradation of the TCE at the 2 minute retention time in
activated carbon bed but was unavailable over full height of the biofilm in the nonadsorbing celite biofilter, thus
limiting TCE degradation.
The celite biofilter with larger void space in the 6mm bed continued to substantially remove all of the
contaminants except TCE for about 210 days of operation after full acclimation. The biofilter then exhibited
increasing pressure losses and flooding similar to that which occurred in the activated carbon biofilter. The
biofilm plugging the biofilter, however, adhered less strongly to the celite pellets than to the carbon pellets and
was easier to remove. Studies on the degree of mineralization of the contaminants due to biodegradation were
also conducted on both pelletized media. Abiotic control columns, identical to the biofilter beds but without
biomass seeding, revealed breakthrough of the contaminants typical for adsorptive and non-adsorptive beds.
The amounts of contaminant absorbed in the small volume of nutrient solution were negligible compared to that
in the inlet air. There were no increases in carbon dioxide concentration in the air stream.
In the carbon and celite biofilters, carbon dioxide increases in the air stream during treatment of
contaminated air ultimately revealed nearly complete mineralization of the removed contaminants. Initially, a
portion of the contaminants removed by the biofilters were converted to biomass. With increasing biofilm
buildup and negligible releases of biomass, carbon dioxide increases in the effluent air approached increases
expected from full mineralization of the removed contaminants. This indicated that the biomass decay rates
began to approach the biomass growth rates in both biofilters. Chloride ion increases in the effluent nutrient
solution also revealed complete conversion of organic chlorine in the contaminants to chloride ion. These
results along with the absence of partially degraded VOC products in the effluent air, verified by gas
chromatography measurements, confirmed efficient mineralization of the removed VOCs.
A comparative study of a-pinene biodegradation was also conducted in a celite biofilter with trickling
nutrient and buffer solution. In contrast to the compost biofilter, the results revealed that nearly complete
removal and mineralization occurred at biofilter air retention times of 4 minutes for inlet concentrations of a-
pinene of 25, 50, and 75 ppmv. Reduction of the air retention time to 1.5 minutes at an inlet a-pinene
concentration of 75 ppmv did not significantly reduce celite biofilter performance.
PCE, a VOC recalcitrant to aerobic biodegradation, and TCE are easily dechlorinated under anaerobic
conditions. The partially dechlorinated products are also easily degraded under aerobic conditions. Thus a
nonbiodegradable hydrogel pellet (3 by 6mm cylinder), encapsulating biomass from an activated sludge process,
was developed to retard oxygen transport, producing an anaerobic core and an aerobic outer zone for
synchronous anaerobic/aerobic treatment of contaminated air. The pellet included an outer stainless steel mesh
for structural stability. Biofilter tests of the pellets with trickling nutrient and buffer solution revealed complete
mineralization of inlet TCE (21 ppmv) and PCE (20 ppmv) at air retention times of 1.5 minutes and 4 minute,
respectively. The hydrogel pellets exhibited long term stability.
Structured Geometry (Straight-Passages) Biofilters
Biofilter studies on straight-passage media with trickling nutrient and buffer solutions at typical loadings
of 150 L/m2/day included a celite plate biofilter and extruded cordierite and carbon coated cordierite biofilters.
The celite plate biofilter constructed of thick-walled plates with low surface area per unit volume (10 m2/m3),
required relatively high air retention times (15 minutes) for nearly complete biodegradation of four of the five
302
-------
inlet contaminants; toluene (450 ppmv), methylene chloride (150 ppmv), ethylbenzene (25 ppmv), and
chlorobenzene (40 ppmv). As in the biofilter with celite pellets, the fifth contaminant, TCE (25 ppmv), was
only partially removed (35%) because of lack of a cometabolite in the upper region of the biofilter. Increasing
the inlet toluene concentration in air to 850 ppmv and subsequently increasing the air flow rate in an attempt to
increase the availability of cometabolite produced little impact on TCE removal. Addition of phenol to the
nutrient solution, as an alternative TCE cometabolite available over the entire bed height, however, increased
the TCE removal, ultimately to nearly 100% removal.
Biomass release from straight-passages celite media was observed in the effluent nutrient solution. The
amount of released biomass depended upon the nutrient flow rate and biomass loading. In addition, during 240
days of operation, the straight-passage media in the bench-scale filter never exhibited pressure drop buildup.
These results suggested possible self-cleaning or easier cleaning for straight-passages media compared to the
extensive cleaning required for pelletized media.
Biofilter studies on isopentane using extruded (straight-passages) cordierite and carbon and activated
carbon coated (0.1 mm films) cordierite media, with high surface area per unit volume (80m2/m3) and high void
fraction (80%), were evaluated at isopentane concentrations varying from 360 to 960 ppmv in the inlet air to
compare their removal efficiencies with those of peat and compost biofilters. The removal efficiencies of
slightly soluble isopentane in the cordierite biofilter, as a function of contaminant concentration and air retention
time, were similar to those observed in peat and compost biofilters. The removal efficiencies in the cordierite
biofilter were limited by poor adhesion of the biofilm on the vertical walls. The removal efficiencies in the
carbon coated and activated carbon coated cordierite biofilter, however, were substantially better than observed
in peat, compost or uncoated cordierite media. The average biodegradation rates of isopentane in activated
carbon coated cordierite biofilters were also significantly greater than those in the other biofilters, substantially
reducing biofilter size.
Media surface effects on start-up and dynamic response to sudden contaminant concentration changes
were examined for straight-passages cordierite media without coatings and with coatings of resin, carbon (0.1
mm) and activated carbon (0.1 mm) as a function of time. Biofilter start-up with carbon coated cordierite was
faster than uncoated and resin coated cordierite media. Start-up for activated carbon coated media was fastest.
High adsorption capacity and strong adhesion of the activated carbon coating accelerated the start-up process.
Similarly, dynamic responses to sudden contaminant concentration changes were also fastest on the cordierite
media coated with activated carbon.
Biodegradation Kinetics
Kinetic experiments on aerobic degradation were conducted for various VOC on a range of media
types. Monod kinetics satisfactorily described the biodegradation of the tested VOCs and led to development of
design models for each type of biofilter media. The kinetic results on isopentane revealed that activated carbon
coated or activated carbon media provided best overall performance for slightly soluble VOCs. For soluble and
moderately soluble VOCs, non adsorbing pelletized and structured media with good biofilm adhesion provided
removals and kinetics equivalent to those obtained in biofilters with activated carbon or carbon coated media.
CONCLUSIONS
Biofilters using porous pellet or structured (straight-passages) media and recycled nutrient and buffer
solutions have potential advantages over conventional soil or peat/compost biofilters for removal of
biodegradable VOCs from contaminated air. These include:
• improved distribution of air flow and moisture control in the media with lower operating pressure
losses and improved process performance;
303
-------
• improved neutralization of acid degradation products (pH control);
• increased capacity for efficiently treating higher VOC loading (up to 800 ppmv); and
• capability for removal of excess biomass from the biofilter, thus preventing clogging and eventual
media replacement.
Biofilters with porous pellet media, however, require development of effective cleaning methods for
field-scale applications. Biofilters with structured media (straight-passages) offer easier cleaning options,
including release of biomass (self-cleaning) into the recycling nutrient and buffer solution. Biofiltration with
activated-carbon-coated or activated carbon pelletized or structured media is an effective technology for control
of slightly soluble VOCs. Novel media supporting anaerobic/aerobic treatment extend the range of application
of biofiltration to VOCs recalcitrant to aerobic biodegradation.
REFERENCES
1. Ottengraf, S.P.P. and R. Diks, "Biological Purification of Waste Gases," Chim Oggi 8 (5), 41 (1990).
2. Gibbon, G.H., Millis, N.F., and Pirt, SJ. "Degradation of a-pinene by Bacteria," Procs IV IPS:
Forment. Technol. Today. 609-612, (1972).
3. Folsom, B.R., Chapman, P.J., and Pritchard, P.H. "Phenol and Trichloroethylene Degradation by
Pseudomonas Cepacia G4: Kinetics and Interactions Between Substrates." Appl. Environ. Microbiol..
56; 1279, (1990).
FOR MORE INFORMATION:
Dolloff F. Bishop
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, OH 45268
Phone: 513/569-7629
304
-------
POTENTIAL CRYPTOSPORIDIUM SURROGATES AND EVALUATION
OF COMPRESSIBLE OOCYSTS
Sylvana Y. Li, James A. Goodrich, James H. Owens, and Robert M. Clark
Drinking Water Research Division, US EPA
26 W. Martin Luther King Dr., Cincinnati, OH 45268
Gene E. Wflleke
Institute of Environmental Sciences
Miami University, Oxford, OH 45056,
Frank W. Schaefer III
EMSL, US EPA
26 W. Martin Luther King Dr., Cincinnati, OH 452683
INTRODUCTION
Cryptosporidium has been recognized as an important waterborne agent of gastroenteritis and a biological
contaminant in drinking water. The widespread presence of Cryptosporidium in surface source water and
either untreated or insufficiently treated drinking water has led to Cryptosporidium outbreaks in the United
States and worldwide.
Cryptosporidium is highly resistant to commonly used drinking water disinfectants. Typical dosages of
chlorine or monochloramine currently used in water plants appear to have no significant inactivation effect on
Cryptosporidium.1 High-dose ozonation may constitute an effective treatment technology to inactivate
Cryptosporidium in drinking water.2 Among the conventional control practices, filtration3 and high
temperature distillation4 appear to be the potentially viable technologies for protection against
Cryptosporidium in drinking water.
As employed in many water plants, filtration is likely to be the most practical treatment technology
utilized for Cryptosporidium removal in the near future. Consequently, accurate and reliable methods for
evaluation of Cryptosporidium removal rates for filtration-based systems are necessary to assist States in
determining drinking water quality and complying with the up-coming national standard for Cryptosporidium
in drinking water. Furthermore, searching for reliable and non-hazardous surrogates for evaluation of
treatment plant efficiency has been intensified because of the potential health risk associated with
Cryptosporidium. Additionally, during the filtration procedure Cryptosporidium may squeeze and fold through
pores size of the filtration systems that are smaller than the diameter of the organism; a fraction of these
Cryptosporidium oocysts may still remain a certain degree of viability. These uncertainties are critical for the
evaluation and optimization of filtration-based physical treatment systems.
The in-house research studies described below consist of two parts. One is a potential surrogate study
using bag filtration systems at the US EPA Test & Evaluation Facility in Cincinnati, Ohio. The second is
Cryptosporidium compressibility and viability investigation.
METHODOLOGY
Parti.
Potential Surrogate Studies
The potential surrogates for the non-preserved Cryptosporidium tested include 4.5 /un polystyrene beads, 1-
25 yxm particle counts, 4-6 /nm particle counts, turbidity, and preserved Cryptosporidium. A given volume of
influent solution spiked with either polystyrene beads or Cryptosporidium was injected into the raw water
ahead of the inlet pump of the bag filter system. An influent sample was collected from the spiked solution
before each test. An effluent sampling consisted of collecting approximately 5% of the total effluent volume
using a manifold and a 1.0 /xm polycarbonate membrane filter. The effluent sample was concentrated to
305
-------
approximately 8 ml using centrifugation (1200 rg) to a given small volume. A flow meter was used to
determine the total effluent volume of each test. The duration of each test was also recorded. Both
collected influent and effluent samples were scanned and examined under a UV microscope to determine the
number of beads or Cryptosporidium oocysts using a hemacytometer. The preserved Cryptosporidium oocysts
were from neonate male holstein calves and stored in dichromate. Indirect fluorescent antibody was used to
label the effluent Ciyptosporidium oocysts. Particle count analyses were conducted using a Met One" particle
counter equipped with a light scattering liquid 211 sensor. The turbidity measurements were performed with
a Hach* 2100P portable turbidimeter.
Three bag filter systems were used in the studies. Bag filter #1 was supplied by Strainrite Inc. Lewiston,
Maine, and Bag filter #2 was manufactured by Filtration Systems, a division of Mechanical Mfg. Co. at
Sunrise, Florida. Bag Filter #3 was obtained from the 3M Company. The studies were performed using 50
psi inlet pressure as recommended by the manufacturers, at varying pressure drops across the'bag filter and
influent flow rates of 125, 25 and 40 gpm.
Part 2. Compressibility and Viability Tests
A unique compressibility test device was developed which made it possible to examine Cryptosporidium
passage through a membrane of designated pore size using a manifold under various inlet pressures. The non
preserved Cryptosporidium oocysts that we used for this study came from neonate male holstein calves. A
given number of Cryptosporidium oocysts were suspended in 0.01% Tween 20 (v/v) contained in a 5 L
pressure reservoir. The influent Cryptosporidium oocysts passed through the system equipped with a 25 mm
polycarbonate membrane filter. The collected effluent was centrifuged to approximately 200 jul and then was
scanned and counted under the microscope. The membrane filter used in this study was made of
polycarbonate with a pore size of 3 um, which is smaller than the Cryptosporidium average size 4-6 /xm.
Polystyrene beads were used to demonstrate membrane integrity; no beads were detected. This advantageous
filter characteristic allows precise determination of removal rates for both the beads and Cryptosporidium at
various inlet pressures.
Three blank tests without a filter were performed under 50 psi inlet pressures in order to evaluate
Cryptosporidium loss in the system. Four tests were conducted under 50 psi inlet pressure and one test under
25 psi with the 3 pm filter. Non-preserved Cryptosporidium was used in the blank tests. In order to compare
the system loss difference between non-preserved and preserved Cryptosporidium, one blank test using
preserved Cryptosporidium was performed under 50 psi inlet pressure.
For the compression test using the 50 psi inlet pressure conditions, a viability test was performed on the
non-preserved Cryptosporidium recovered from the effluent. Cryptosporidium concentrated from the effluent
was fed to neonatal mice at two dosages, 340 and 3400 Cryptosporidium oocysts per animal. A control sample
consisted of 18 mice fed with 63 unffltered non-preserved Cryptosporidium oocysts per animal.
RESULTS
The initial bag filter testing results indicated that the removal rates for 4.5 /mi polystyrene beads, turbidity,
and particle counts in the 1-25 fan. and 4-6 pirn size ranges, vary significantly among the three bag filters. The
log reduction ranges from 0.14 to 3.42 for the 4.5 /*m beads, from 0.04 to 1.89 for turbidity, from 0.09 to 2.84
for 1-25 fim. particle counts, and 0.06 to 2.99 for 4-6 um particle counts. Initial bag filter testing for preserved
Cryptosporidium reveals a log reduction of 1.8.
Inspection of the bag filter testing results reveals varying relationships between the 4.5 /un beads, 4-6 pirn
particle counts, 1-25 ^tm particle counts, turbidity and non-preserved Cryptosporidium. The correlations
between log reductions for beads and turbidity or particle counts for three bag filters at varying pressure
drops and inflow rates, are consistent with the linear model:
Log (beads) = a Log (turbidity or particle count) +fi
306
-------
where a is the slope constant of the linear correlation, and /? is the regression constant. The 4.5 pun beads
and turbidity have a linear correlation with a and /) of 0.52±0.15 and 0.04+0.24, respectively. The squared
coefficient of determination (R2) is 0.79. The beads and 4-6 pun particle count have a 1:1 correlation in
removal rates as shown by the linear regression results: a and /3 are 0.97+0.35 and -0.29±0.54 respectively,
and a squared coefficient of determination (R2) is 0.79. On the other hand, 1-25 pun and particle counts
display the correlation departured from 1:1 relationship and closer to the correlation exhibited between beads
and turbidity. The linear correlation between 4.5 pun beads and 1-25 pun particle counts has the slope
constant (a) of 0.73 and coefficient of determination (R2) of 0.55.
Detailed size analyses show that the size distribution of non-preserved Cryptosporidium, 4.5 pun beads, and
particle counts, primarily determines their similarities and dissimilarities in the removal rates for bag filters.
The similar removal rates between 4.5 pun beads and non-preserved Cryptosporidium are consistent with their
similar size distributions. Cryptosporidium has the size ranging 3.50-6.50 pun with an abundance peak at 4.01-
4.50 pun, whereas the 4.5 pun beads range in size from 3.50-7.00 pun with the peak at 4.51-5.00 pun.
Furthermore, the 4-6 pun particle counts have a similar size distribution with 4.5 pun beads. In contrast, the
1-25 pun particle counts have two abundance peaks at 5.01-6.00 pun and <2.00 pun. This size distribution is
significantly different from those of the Cryptosporidium and 4.5 pun beads. The difference appears to be the
explanation for the distinct removal rates of 1-25 pun particle counts from those of 4.5 pun beads and non-
preserved Cryptosporidium.
The determined correlations in removal rates among Cryptosporidium and its potential surrogates do not
appear to be the artifacts of influences from operational parameters including pressure drop, flow rate, or
spiked influent level of surrogates. Testing results for the beads at varying pressure drops and flow rates for
bag filter #1 show that the variation in log reduction is approximately 0.5, within the range of experimental
variations determined with duplicate testing. Similarly, the number of influent beads does not appear to have
a statistically significant influence on the removal rates.
Compression study results lead to two important observations regarding non-preserved Cryptosporidium
removal efficiency of the filtration process and the reliability of surrogates. First, about 0.7% of the non-
preserved Cryptosporidium at 3.50-5.00 pun size passed through the 3 pun polycarbonate filter at 50 psi inlet
pressure. In contrast, the polystyrene beads of the same size range were completely removed. Different
removal rates of the non-preserved Cryptosporidium and the beads are believed to reflect the ability of
Cryptosporidium to squeeze and fold through the pressurized membrane filter. Furthermore, compressibility
tests at 25 psi inlet pressures show an increased removal rate of Cryptosporidium. Only about 0.0021% non-
preserved Cryptosporidium passed through the filter. It is possible that the Cryptosporidium squeezing and
folding effect on the removal rate of the filter may diminish at lower inlet pressures.
Secondly, 4 blank tests on non-preserved Cryptosporidium at 50 psi inlet pressure lead to the observation
that 53.61 ±5.30% of Cryptosporidium in oocysts in the influent was lost in the system. However, the result
from one blank test on preserved Cryptosporidium at the same experimental conditions show the system loss
less than 1%. Such a large difference in system loss between preserved and non-preserved Cryptosporidium
indicate that the non-preserved Cryptosporidium not only can squeeze and fold through membrane filter, but
also adhere to physical surfaces at applied inlet pressure. Both characteristics of the Cryptosporidium have
substantial influence on apparent removal rates.
Initial viability test results show that the non-preserved Cryptosporidium collected from effluent of
compression study are still viable. Five of seven mice were infected with a dose of 340 Cryptosporidium
oocysts per animal. Ten of ten mice were infected with a dose of 3400 Cryptosporidium oocysts per animal.
Control samples of 63 Cryptosporidium oocysts per dose infected 13 out of 18 animals.
CONCLUSIONS
The Cryptosporidium surrogate investigation, compressibility study and viability test, lead to the following
conclusions:
Accuracy and reliability of the Cryptosporidium surrogates are determined by similarities of size
307
-------
distribution between Ctyptosporidium and its surrogates. The 4.5 pun polystyrene beads have a similar size
distribution as the Cryptosporidium with a peak distribution in the 4-5 pun range. Consequently, the 4.5 pun
beads and Cryptosporidium show the closest removal rates. This similarity will be further characterized in
future studies. Studies on surrogate interchangeabflity show that the 4-6 pun particle counts and 4.5 pun beads
have a 1:1 correlation in removal rate. On the other hand, 1-25 pun particle counts and turbidity are not
directly interchangeable surrogates with 4.5 pun beads. As the latter most closely resemble the
Cryptosporidium, both 1-25 pun particle counts and turbidity are less reliable surrogates for Cryptosporidium
removal evaluation.
Compression test results demonstrate that unlike the 4.5 pun beads, a small fraction of 3.50-5.00 pun
Cryptosporidium oocysts squeeze and fold through smaller pore sizes of a membrane filter. This unique
characteristic of Cryptosporidium leads to apparently lower removal rates than that of the 4.5 pun beads.
However, the influence of Cryptosporidium size change at the membrane filter appears diminished at low inlet
pressures. A large fraction of Cryptosporidium oocysts that passed through membrane filter remain viable.
Therefore, a high removal rate might be required for filtration-based treatment technology to ensure safety of
the treated drinking water.
It is noted that both surrogate and compressibility studies are presently on-going. Results from future
studies will be supplemental to the conclusions presented here.
REFERENCES
1. Pontius, F. W. Protecting the Public Against Cryptosporidium. J. AWWA. p. 18-22 and cont'd on p. 122-
123. August, 1993
2. Finch, G. R., Black, E. K, Gyurek, L., and Belosevic, M. Ozone Inactivation of Cryptosporidium Parvum
in Demand-free Phosphate Buffer Determined by in Vitro Excystation and Animal Infectrvity. Applied
and Environmental Microbiology, v. 59, p. 4203-4210.1993
3. Chapman, P. A. and Rush, B. A. Efficiency of Sand Filtration for Removing Cryptosporidium Oocysts
from Water J. Med. Microbiol. v. 32, p. 243-245. 1990
4. Payer, R. Effect of High Temperature on Infectivity of Cryptosporidium Parvum Oocysts in Water.
Applied and Environmental Microbiology, v.60, p. 2732-2735. 1994
For more information:
Sylvana Y. Li, Systems & Field Evaluation Branch
Drinking Water Research Division, 26 W Martin Luther King Dr.
Cincinnati, OH 45268, (513)-569-7074, (513)-569-7185(Fax)
308
-------
REMOTE MONITORING AND CONTROL OF SMALL DRINKING WATER PACKAGE PLANT
SYSTEMS
John M. Teuschler, James A. Goodrich, Benjamin W. Lykins, Jr.
, and Robert M. Clark
Drinking Water Research Division
US EPA
26 W. Martin Luther King Dr.
Cincinnati, OH 45268
INTRODUCTION
Remote telemetry, metaphorically referred to as the "electronic circuit rider", offers an attractive
alternative to an "on the road" operator. Substantial cost savings in addition to the improvement of drinking
water quality can be achieved through the use of the remote telemetry. Using remote telemetry, an
operator has the ability to call a remote data logger unit and review the operating parameters in "real time"
of a plant within minutes using the central site personal computer. Likewise, several sites may be reviewed
in this manner. An operator normally requires hours to physically visit the remote drinking water plant sites.
Therefore, remote telemetry can significantly reduce operator time required to visit remote plants.
With remote telemetry, the operator has the ability to display the current operating parameters of a
remote drinking water (DW) plant on the central site computer. The operator then, can prioritize the plant
site visits accordingly. Also, immediate attention can be given to a drinking water plant with critical
problems. Preventive maintenance can be scheduled for plants with operating parameters showing the
need for maintenance. The overall effect from remote telemetry is an improvement in operator efficiency
and hence a net improvement in water quality.
In addition to displaying parameters when the operator calls a remote site, the remote data logger unit
has the ability to call the operator via a phone line to alert the operator of an immediate problem.
Commercial remote sensing units have the ability to auto dial a pre-programmed phone number to alert
an operator of site problems. This is typically referred to as "exception reporting." The operator defines
the conditions which cause the remote unit to dial the pre-programmed phone numbers. The exception
limits can be set for turbidity, pressure, flow, etc.. The operator of a specific plant defines the limits for the
plant. Through exception reporting, water quality is improved by quickly alerting the operator of potentially
harmful site conditions. Again, the net effect of remote telemetry is the improvement in operator efficiency
thereby improving the overall, drinking water quality at the consumers tap.
METHODOLOGY
This project demonstrated remote telemetry using commercially available hardware and telemetry
software. Specifically, a CUNO ultrafiltration drinking water package plant was remotely monitored and
controlled. Monitoring consisted of permeate flow, pH, finished turbidity, hi and low trans-membrane
pressure, chlorine residual, raw pressure, permeate pressure, reject flow, raw flow, and raw turbidity.
Control of the plant via electrically actuated proportional valves included: raw water flow, hi trans-membrane
flow, lo trans-membrane flow, and reject flow. The telemetry software allowed the "on line" data display,
data analysis, trend reporting, and history reporting. Proportional control of the valves was accomplished
via the remote telemetry software.
Remote telemetry hardware requirements include an Acromag Inc. remote data logger, a central site
generic personal computer, a modem for communications, and selected monitoring instrumentation. The
central site control and acquisition software is commercially available from Total Systems Resources (TSR).
The process control and data acquisition software from TSR contain the hardware drivers necessary to
communicate with the specific remote data logger used in this project.
309
-------
Data Logger:
A commercially available remote data logger was purchased from Acromag Inc.. The data logger is
a microprocessor unit with programmable memory, random access memory, a serial port, analog
input/output and digital input/output. The remote logger is customized to a specific site. Information
concerning the analog to digital conversions and the digital to analog control can be downloaded to the
data logger's Electrically Erasable Programmable Read Only Memory (EEPROM) from the central site
personal computer. Power failures will not cause the unit to loose specific site control information.
The data logger does not use electromechanical digital storage devices like disks or tapes for data or
program storage. Read only memory (ROM) is used for CPU control and is hard coded at the factory. The
ROM software controls the operation of the CPU. It is similar to a mini disk operating system (DOS).
Commands are received in predefined format via the serial link, and appropriate actions are taken.
Approximately seventy commands are available in the Acromag ROM system. The commands range
widely in capability. The commands are as simple as "IS" which is identify station. And the commands
are as complex as "LL" or log list all the data remotely stored.
The versatility of the microprocessor unit is a result of its field programmability. User field program
control is stored in EEPROM. The field program control defines data collection rates, data channels to be
collected, data channels to be used as analog outputs, communicates with the central site computer, and
in general, controls the customization of the data acquisition and control of the remote data logger.
The microprocessor random access memory (RAM) stores data from the sensors. Approximately 9000
data points can be stored with the data logger used in this project. The data points are collected at user
programmed data rates. The remote data logging is activated by the operator. The RAM is the internal
buffer for the collected data. At operator selected intervals, remote data is downloaded from the data
logger to the central site computer for display,and time series plotting. After dumping the data to the
central site, the RAM is free to start storing a new series of data points.
The microprocessor unit contains a serial communication link. It is through this link that the
microprocessor receives control information and sends the collected information to the central site
computer. The communication port is user programmable via the standard serial port capability.
Programmability includes baud rate, parity checking, etc.. The microprocessor unit's serial link may be
hardwired to a modem, radio transmission link, or other serial ASCII communication systems.
Data loggers typically have the ability to communicate with a variety of data acquisition modules. These
include analog to digital (A/D) modules, digital to analog (D/A) modules, or digital input/output modules.
As discussed earlier, the data logger is able to be customized to a particular site. System setup personnel
customize the data logger to the site requirements. The customization is transparent to an operator. Data
from the sensors will be numerically displayed as flow, pressure, turbidity, etc. on the computer display.
All raw data transformations will be computed internally in the central site software. The transformations
will be programmed by the setup personnel.
Most monitoring instrumentation can be specified with a 4 to 20 ma current loop interface. The 4 to
20 ma interface was chosen as the standard for this project. All instrumentation for the project was ordered
with the 4 to 20 ma current loop interface. This included the "data in" as well as the "data out." The
remote monitored parameters included chlorine, turbidity, pH, pressure, temperature, and flow.
Control of the package plant is accomplished using 115 volt AC electrically actuated control valves.
For proportional control, the actuator movement is initiated with the standard 4 to 20 ma current loop. Note
that 4 ma. is the full closed control signal. And 20 ma. is the full open signal. Feedback potentiometers
on the actuator control valves indicate the current position of the control valve. The feedback
potentiometers will allow the current position of the valve to be displayed, upon system bootup. Control
of DW site valves is accomplished via the remote control system. The operator may initiate a change in
310
-------
the process control, and instantly evaluate the effect of the process control change through the real time
display of the monitored parameters.
Data acquisition and control software:
The central site software package, can communicate via modem, radio, or direct link with the remote
data logger. The central site software provides a high level (mouse driven), sophisticated interface to the
remote data logger. The software package used in this demonstration has extensive graphical capability
which can provide complete monitoring and control information at a glance. The software is user friendly
and can be easily applied to a variety of small drinking water package plant systems, nationwide.
The central site software communicates with the remote data logger by channels (one to sixteen). In
this project, for standardization purposes, all Acromag units were programmed in a standard configuration.
Channels 1 to 12 were defined as analog to digital modules. Channels 13 to 16 were defined as digital
to analog modules. Channels 1 to 12 monitored remote parameters. Channels 13 to 16 controlled remote
servo valves for process control.
All sites will require some customization of the software channels. Software channel 1 may be turbidity
at site 1. Whereas at site 2 turbidity may be on channel 4. The system software setup personnel require
some computer expertise. The computer expertise required is equal to a Harvard Graphics user or a
Freelance Graphics user. For sites lacking "in house" computer expertise, the central site software vendor
will gladly customize the software for a nominal fee.
In this demonstration, the software was partially customized by Drinking Water Research Division
personnel. Customization by DWRD personnel included: adding the remote servo valves for variable flow
rate control, adding a number of additional sensors not in the original system setup file. Modifying the
graphics display to accommodate the additional sensors and control valves. Modification of the system was
relatively painless. Sufficient technical support was obtained from both the software and hardware vendbrs.
CONCLUSIONS
As more small DW system operations utilize personal computers for billing, personnel tracking, etc.,
remote telemetry adds an additional cost benefit to the acquisition of a personal computer. The computer
availability at the central office, in addition to attractive pricing of remote data loggers and data logger
software makes remote "telemetry and control" an affordable option to a small systems operation. Cost
of the remote data logger is under $5,000. Central site software is approximately $3,000 without
customization. Sensors at the remote DW site usually have an option for an external computer interface
(4 to 20 ma output). The external 4 to 20 ma interface is typically under $200. The fax can also function
as a 9600 baud modem for data acquisition and control.
Computer expertise is probably the main impediment to remote telemetry. However, with each new
generation, computers and computer control are becoming more commonplace. Remote telemetry and
control is quickly coming into the reach of small system operations both from the standpoint of cost and
personnel capabilities.
For more information:
John M. Teuschler
Systems & Field Evaluation Branch
Drinking Water Research Division
26 W. Martin Luther King Drive
Cincinnati, OH 45268
311
-------
CONTROL OF COPPER CORROSION OF HOUSEHOLD PLUMBING MATERIALS
Michael R. Schock
and
Darren A. Lytle
Drinking Water Research Division
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
26 West Martin Luther King Drive
Cincinnati, Ohio 45268
(513)569-7412
INTRODUCTION
The promulgation of the "Lead and Copper Rule" by the USEPA in 1991 has forced hundreds of
water utilities nationwide to become concerned with controlling the uniform corrosion of copper plumbing
materials. The exact extent of the problem is hard to quantify, but in the first round of monitoring by the
large water systems (about 682), approximately 6 % exceeded the 1.3 mg/L action level according to an
AWWA study. The highest copper levels for these utilities appeared to be in the southeastern and
western regions of the United States, in utilities covering a considerable range of water qualities. The
cuprosolvency problem apparently increases with decreasing utility size. When medium-sized water
systems are included, large numbers of action level exceedences for copper occurred in the central and
north-central midwest, implicating areas having hard and high alkalinity ground waters of approximately
neutral pH. These water qualities are not those conventionally considered "corrosive".
Unfortunately, the regulatory monitoring data are of limited use for extracting details of copper
chemistry behavior and understanding potential copper passivation strategies. Nonetheless, several
interesting gross-scale trends have been discerned for large water systems. One example is that there
is a poor correlation between 90ft-percentile lead and copper levels. Another trend is that copper
exceedences tended to be highest at very low alkalinities (<25 mg CaCO/L) and increasingly greater
over 75 mg CaCOg/L. Finally, no action level exceedences were reported for systems having a pH above
approximately 8.
The data thus far suggest that cuprosolvency (copper solubility) will be a major concern across
the United States, especially for smaller water suppliers that are less likely to employ corrosion control
and use ground water sources. Further, the poor correlation between reported go^-percentile lead and
copper levels suggests that different control strategies for copper than those considered appropriate for
lead may need to be developed or employed by affected utilities. Understanding how copper will respond
to lead control measures and the results of other regulatory treatment requirements is therefore of
considerable interest. Indeed, a response that effectively controls lead corrosion might exacerbate
copper corrosion. Moreover, a utility must distribute aesthetically-pleasing water. A good example of the
conflicts between control of corrosion of iron mains and reducing copper corrosion rates has been given
for a study in Vancouver, BC.
In attempting to address some of the data gaps for cuprosolvency control by utilities, a variety of
experimental systems have been constructed and operated in USEPA laboratories. During these
experiments, some perplexing data was generated that appeared to either contradict some "conventional
wisdom" on copper corrosion, or showed unexpected sensitivities to important water chemistry variables
and experimental system operational protocols. These observations provided the motivation to begin
312
-------
exploring the "cuprosolvency" (copper solubility) issue in detail. The initial results of this ongoing
research are reported in the new U. S. Environmental Protection Agency publication1.
An accurate equilibrium model for cuprosolvency provides utilities and consultants with a useful
and practical tool for reliably selecting and applying corrosion control programs that will work over the
long term. A sound theoretical and practical understanding of the important factors affecting
cuprosolvency greatly obviates the need for complicated, costly, and sometimes misleading bench-top
and pilot-plant experimental studies. Small and medium-sized water systems frequently lack the
technical sophistication as well as the mechanical and financial resources needed to conduct elaborate
"demonstration" studies. For water systems able to conduct experimental treatment evaluations, an
accurate model for cuprosolvency reduces the need for including fundamental tests (e.g. pH and
carbonate concentration effects) in their evaluations. Utilities may then concentrate on determining the
need for additional experimental evaluations of inhibitor treatments or investigation of other unusual
chemistry characteristics of particular water systems.
A better understanding of roles of pH and DIG (dissolved inorganic carbon) in cuprosolvency
greatly improves the accuracy of regulatory "desk-top" corrosion optimization evaluations, thereby
providing a better assessment of long-term treatment impacts than those ascertained over a short time-
frame (i.e. weeks, months). Finally, an accurate solubility model for copper allows many utilities to
achieve the requirements of the Lead and Copper Rule without the need to conduct the treatment
studies.
A newly-emerging indirect constraint on copper corrosion control in drinking water is by
wastewater effluent guidelines and limits that are becoming increasingly stringent. Ironically, ambient
corrosion of domestic, commercial and institutional plumbing systems is now becoming a "contaminant"
of wastewater that is becoming difficult to control to an adequate degree through normal waste treatment
processes. The development of an accurate model for copper solubility allows the evaluation of
treatment alternatives for optimizing the control of cuprosolvency beyond drinking water regulatory
requirements, to achieve adequate levels at the end of wastewater treatment.
OXIDATION REACTIONS OF COPPER
Because of the positive cell potentials for copper metal oxidation, copper pipe in water containing
oxidizing agents will continue to corrode until either all of the oxygen is depleted, or until precipitated
oxide films arrest the rate of corrosion. There is some evidence that the overall transformation from Cu+
to Cu2* is essentially the rate-limiting factor, with Cu+ existing in reversible equilibrium with the Cu metal
at the pipe surface.
In drinking waters, the oxidizing agents (electron acceptors) that will cause the corrosion of
metallic copper are predominantly dissolved oxygen and aqueous chlorine species. These reactions may
be composed of several intermediate steps, any of which can be rate-controlling. Many studies have
documented (to some degree) the impact of dissolved oxygen or aqueous chlorine species on copper
oxidation and dissolution rates. Free chlorine species (i.e. HOCI, OCI", CI2) have not been conclusively
shown to affect the equilibrium solubility of copper, other than by influencing the valence state of the
copper by its presence or absence.
1 Schock, Michael R., Lytle, Darren A., and Clement, Jonathan A. Effect of pH, DIG, Orthophosphate
and Sulfate on Drinking Water Cuprosolvency. EPA Office of Research and Development, Cincinnati,
Ohio, 1995. In press.
313
-------
Oxidants may have several other potential impacts on the observed copper levels in the water
and the nature of the passivating solids on the pipe. The effect of chlorine on the oxidation rate of the
copper metal might be alteration of the crystalline characteristics and porosity of the oxide corrosion
product film produced at the pipe surface. For example, by reducing the formation of a protective
Cu2O(s) underlayer through maintaining a high EH level, or by indirectly influencing scale structure and
conductivity through chloride formation as the chlorine is reduced. However, many of the apparent
effects of chlorine on copper solubility may merely result from accelerated corrosion kinetics (rates), '
rather than changes in equilibrium conditions.
Copper plumbing can be exposed to a wide environment of EH-pH conditions. This environment
evolves as oxidation and reduction reactions occur through corrosion and deposition processes for
different lengths of stagnation times, and in different diameters of pipes. Therefore, all three normal
copper valence states (metal, +1 , +2) could reasonably occur under drinking water conditions either
geographically or with time. The relative stabilities of Cu(l) and Cu(ll) species in aqueous solution
depend very strongly on the nature of anions or other ligands present in the water.
Solubility Chemistry of Copper (I) and Copper(II)
Because the immune region for copper corrosion extends well above the minimum EH for water
stability, water does not corrode copper in the absence of oxygen or other added oxidants. With the
addition of electron acceptors to the solution (e.g. HOCI , O2), either or both of the oxide solids Cu2O
(cuprite) or Cu(OH) may form. The equilibrium solubility of cuprous copper, ST Cu(l), may be described as
the sum of the concentrations of all dissolved cuprous species:
ST.cu(i) =[Cu+]+[CuCl0] + [CuCl2] + [CuCI^] + [Cu2CI^"] + [CuNH3] + [Cu(NH)3]
in which [ ] (brackets) indicate concentration in mol/L The total solubility may then be computed by
standard techniques. Under relatively anoxic drinking water conditions, such as many well waters, pH
and chloride concentrations are the dominant controls on copper solubility. Metastable solid formation
and the rate of the oxidation reaction are important in governing copper levels at the tap. Ammonia forms
strong complexes, but little research has been done to determine its significance in drinking water.
Under oxic conditions, copper(ll) aqueous speciation is dominated by hydroxide, carbonate, and
hydroxycarbonate complexes. Weak complexes are formed with sulfate, orthophosphate and chloride.
Strong complexes are formed with ammonia, but the high concentrations of carbonate, bicarbonate and
hydroxide usually swamp out any expected impact on the speciation. Generally, copper(ll) solubility is
given by the equation:
T.(OH.co3)
[Cu(OH)° ] + [Cu(OH)
] + 2[Cu
] + 3[Cu
"
In new plumbing systems, copper(ll) solubility is generally controlled by a poorly-crystalline form
of cupric hydroxide, Cu(OH)2. This scale ages, and after some time becomes either CuO (cuprite) or
Cu2CO3(OH)2 (malachite), depending on pH and the dissolved inorganic carbon concentration. Below
about pH 7.5, orthophosphate reduces copper(ll) solubility. The effect is particularly strong in the pH
range of 6.5 to 7, and is fairly consistent with the hypothesis of the formation of a solid of approximately
the composition of Cu3(PO4)2«2H2O. Orthophosphate also appears to interfere with the normal formation
of cupric oxide, cupric hydroxide, and malachite, at low DIG concentrations.
314
-------
Little research has been done on the effect of sulfate on copper solubility in potable water.
Results of this ongoing research project so far suggest that sulfate plays a major role in the increase of
soluble copper(ll) through the formation of a metastable basic cupric hydroxide solid, such as
Cu4(OH)6SO4«2H2O. This effect appears to be particularly strong above pH 8.
PRACTICAL CONSEQUENCES OF REDOX POTENTIAL ON CUPROSOLVENCY
In water supplies maintaining oxidizing disinfectant residuals, copper oxidation and
dissolution into the water will take place until both that residual and dissolved oxygen are depleted.
However, copper levels will stop climbing after saturation equilibrium is reached with a passivating solid
(e.g. Cu(OH)2). The profile of copper concentration versus time in oxidizing systems will depend upon
the water chemistry and the relative stabilities of copper(l) and copper(ll) aqueous species. The copper
concentration in the water may continue to rise for many hours, perhaps substantially beyond the
"overnight standing" period normally used for sampling, depending on the operative kinetics in a given
system. Complexation of either copper(I), copperjll), or both, will be important in these systems.
In undisinfected water systems having low levels of dissolved oxygen, such as many
closed ground water systems, complicated copper versus time profiles may result, as the initial small
quantity of oxygen is consumed. This behavior is consistent with the observation of short-term copper
concentration maxima, followed by decrease in concentration, reported in several studies. Copper(l)
aqueous speciation and solid reactivity are extremely important in these systems. Different parts of a
distribution system, particularly ones fed by a combination of different water sources (such as multiple
wells or a mix of wells and conventionally-treated surface water), may exhibit different stagnation profiles.
For the anoxic systems, the use of aeration for VOC, radon or iron removal may cause increased
copper levels, because copper(ll) solids are virtually always more soluble than copper(l) solids.
Unanticipated violations of copper corrosion action levels may result from the installation of any new
oxidation processes for water treatment. In contrast, for oxic systems, proper aeration may strip DIG
from the water and increasing pH, both favorable effects in reducing copper(ll) solubility.
Significant variability in copper levels from site to site in the same water system will likely be
found in water after the normal number of hours of stagnation found in most sampling programs. This
results from the interaction of aqueous complexation and oxidation chemistry of copper, plus effects of
the Existence or development of passivating films in the pipe. Examination of copper concentrations in
experimental coupon cells and pipe loop systems after various standing times frequently show almost
linear increases in concentration extending far beyond the 6-16 hour standing time requirement of
regulatory sampling.
SIGNIFICANCE OF pH AND DIG IN CUPROSOLVENCY CONTROL
The relationship between DIG and copper(ll) solubility is very complex. DIG can play several
significant roles, depending upon its concentration, other water chemistry factors, age of plumbing, water
flow amount and pattern of use.
This research has shown conclusively that trends in copper concentrations resulting from uniform
corrosion are predictable for a range of appropriate water qualities. Given that field conditions frequently
do not reflect equilibrium conditions, some discrepancy between the model predictions and the results of
normal tap water monitoring programs is completely understandable. Nonetheless, using a
comprehensive equilibrium solubility chemical model,.water quality objectives for copper corrosion control
315
-------
by pH and DIG adjustment can be predicted without the necessity of complicated experimental and field
studies.
In new plumbing or at high pH, where cupric solubility is controlled by either cupric hydroxide or
cupric oxide, DIG complexes dominate copper speciation above pH 7.5, resulting in increased
cuprosolvency. The solubility enhancement effect is strongest in the DIG range of 0 to 20 mg C/L, but
the predicted soluble copper at equilibrium is still below about 1 mg/L
DIG serves to control the buffer intensity in most water systems. Therefore, sufficient DIG is
necessary to provide a stable pH throughout the distribution system for corrosion control of copper (and
lead). In treatment practice, the increase in DIG to ensure pH control through buffering will probably need
to be offset by increasing pH to maintain lowered cuprosolvency.
Possibly offsetting the solubility enhancement of copper(il) by carbonate complexes is the
possibility that moderate DIG levels would logically accelerate the formation of passivating Cu2CO3(OH)2
(malachite) films in the pH/DIC region where it would be thermodynamicaliy stable. Thus, enhancement
of conditions that would hasten the formation of Cu2CO3(OH)2(s) (malachite), CuO(s) (tenorite), or both,
relative to cupric hydroxide, would result in a net lower copper(ll) solubility, even with the existence of
some additional carbonate complexation.
In the region of pH from approximately 8.5 to 10, this solubility model developed occasionally
predicts slightly lower copper(ll) solubility than has actually been observed in some well-controlled pipe
loop and precipitation studies. The discrepancies are most likely the result of formation constant errors,
inadequacies of understanding copper(ll) aqueous complexation with hydroxide and carbonate, or the
presence of metastable copper solids associated with anions other than carbonate or hydroxide.
SIGNIFICANCE OF pH AND ORTHOPHOSPHATE IN CUPROSOLVENCY CONTROL
Because cupric hydroxide solubility has been shown to be a better estimate of
cuprosolvency tendencies in many water systems than previously presented models based on
Cu2CO3(OH)2 (malachite) equilibrium, orthophosphate in sufficient dosage is now predicted by
the models to have an ability to further reduce copper levels in the pH range of approximately
6.5 to 7. The necessary dosage is hard to quantify, but the calculations suggest that 3-5 mg
POJL orthophosphate may be necessary to achieve substantial improvements in cuprosolvency
over the DIC/pH system at approximately pH 8, but perhaps only 1-3 mg PO/L at approximately pH 7.
More investigation is still needed to ascertain objectively and consistently if any synergistic cuprosolvency
reduction effect exists from the use of blended phosphates rather than orthophosphate, as has been
suggested by one research project.
IMPLICATIONS FOR CONTROLLING HIGHEST COPPER EXPOSURES
The distinct aging process noted for recrystallization and development of cupric hydroxide, cupric
oxide, and cupric orthophosphate films raise an important caution for determining the highest exposures
to copper in drinking water. While the Lead and Copper Rule was specifically intended to bias the
sampling site selection towards locations with high relative risks for lead exposure, these sites do not
generally correspond to commensurately high risks for copper exposure, which include: newest
construction and remodeling, areas with unstable pH, and dead ends.
Changes in solubility of copper as pipes age, and films build up, has been indicated in several
studies and by practical experience. This investigation has shown how such a phenomenon may have a
316
-------
firm chemical basis that has heretofore been unappreciated. In practice, when attempting to predict the
impact of different water treatment scenarios on cuprosolvency, a fruitful strategy may be to apply
different solubility models depending on the general age of the plumbing systems involved. For example,
for older neighborhoods where Cu2CO3(OH)2 (malachite) and CuO (tenorite) have had sufficient time to
form an integral part of the passivating film, a better prediction of pH, DIG and orthophosphate dosing
impacts would be obtained using their solubility constants to predict copper levels and success of
different hypothetical control strategies. Conversely, for areas predominantly of new construction, a
"cupric hydroxide model" would be more applicable.
SIGNIFICANCE OF THE "CUPRIC HYDROXIDE MODEL" FOR DEMONSTRATION STUDIES
Utilities may encounter some complications in projecting pipe loop or coupon leaching
cuprosolvency data obtained in demonstration studies to behavior across the whole distribution
systems. In the absence of significant concentrations of orthophosphate, this research
indicates that for the short timeframes of most experimental studies conducted to satisfy Lead
and Copper Rule requirements (6 months to 2 years in most cases), cupric hydroxide will
usually be the most important solid phase, rather than CuO(s) (tenorite) or Cu2CO3(OH)2(s)
(malachite). Both pH and DIG effects will be magnified in the experimental systems, relative
to significantly aged piping in the distribution system.
Because of the slow oxidation rate relative to diffusion rates, it is unlikely that copper
concentrations are at "equilibrium" when samples are taken from experimental pipe rig
systems operated to allow "overnight" standing times, unless the systems have been operated
long enough that the passivating films are well-developed, perhaps for years.
When orthophosphate dosing is tested, overestimates of its effectiveness could be
obtained for distribution system areas having old copper plumbing with well-developed
passivating films. The experimental systems may also be very sensitive to minor dissolved
oxygen, chlorine residual, sampling, and stagnation time fluctuations, because of the highly
non-equilibrium nature of copper behavior in the 6-16 hour time period.
OTHER IMPORTANT APPLICATIONS OF EQUILIBRIUM CUPROSOLVENGY MODELS
In addition to the applications to'drinking water corrosion control, the solubility model has
considerable application to several other important environmental research areas, namely:
e aqueous speciation of copper in relation to aquatic toxicity
• estimation of the impact of domestic and institutional plumbing corrosion on wastewater loading
• estimation of the feasibility of various wastewater effluent guidelines for copper
• estimation of copper removal from wastewater by hydroxide or carbonate precipitation
317
-------
HYDRAULIC FRACTURES AS SUBSURFACE ELECTRODES:
EARLY WORK ON THE LASAGNA PROCESS
Larry Murdoch, Jiann-Long Chen, Phil Cluxton, Mark Kemper, Jim Anno and Dave Smith
Center for GeoEnvironmental Science and Technology
Department of Civil and Environmental Engineering
University of Cincinnati
1275 Section Rd.
Cincinnati, Ohio, 45237
INTRODUCTION
Fine-grained sediments present the greatest obstacle to in situ remediation at many
contaminated sites. Where fine-grained sediments are continuous, discharges from wells are
slow and contaminants are recovered at negligible rates. In contrast, at sites underlain by
interbedded sand and silt or clay, volumetric discharges from wells can be significant as water is
readily recovered from the more transmissive sands. The rate of recovery of contaminants can
be rapid as the sands are flushed during initial operations, but the clay interbeds act as persistent
sources of contaminants so that concentrations can be maintained at small but serious values for
many years.
Hydraulic fractures filled with coarse-grained sand provide high permeability layers that
increase fluid flow primarily by reducing the large losses of hydraulic head in the vicinity of a well
(1,2). These features have been shown to increase the discharge from wells by an order of
magnitude or more (2). Nevertheless, current applications of hydraulic fractures are designed to
improve the recovery and removal of contaminants from the subsurface by advective processes.
Effects related to adsorption, preferential flow and other processes can result in mass transfer
limitations that require a large number of pore volumes of fluid to be moved through porous
material before it is remediated using hydraulic flow alone.
Electrokinetics is an effective method of inducing the movement of water, ions, and
colloids through fine-grained sediments. In this process, typically a direct current is applied to
the sediment, inducing osmotic movement of water away from a positively charged anode and
toward a negatively charged cathode (3). Ions and colloids migrate toward the electrode with a
charge opposite from the one that they carry (3,4). In addition, electrolysis of water results in a
decrease in pH at the anode and an increase at the cathode (4). The acidic conditions at the
anode and the basic conditions at the cathode will propagate into the region between the
electrodes, potentially forming a sharp discontinuity in pH (4,5).
Based on the results of laboratory tests and limited field applications (7), electrokinetics
has been shown to be a promising method of recovering ionic and water-soluble contaminants.
However, the process is not without problems. Acidic conditions and electrolytic decay can
corrode some anode materials. Sharp discontinuities in pH induced within the soil mass by
electrokinetics could result in a deposition front where minerals are precipitated in soil pores,
markedly reducing permeability and inhibiting recovery. It may be possible to mitigate problems
related to pH by circulating a buffering solution across the electrodes (7).
Another drawback to conventional electrokinetic remediation is that it requires
contaminants to migrate from their initial location to an electrode and then up to the
groundsurface. In some cases, the migration path could be long or there could be stagnant
zones between wells where the rate of migration is particularly slow, both of which result in
incomplete remediation of the contaminated zone. Moreover, sharply convergent electrical
fields can result in heating and potential losses in the vicinities of electrodes, just as convergent
flow paths result in head losses in the vicinities of wells. Electroplating, or pH-related deposition
can cause contaminants to be removed from solution prior to arrival at the ground surface.
Recent investigations have shown that it may be possible to address some of the
shortcomings of electrokinetics by degrading contaminants in situ. Our colleagues have
developed a solid compound (8), which slowly releases oxygen, that can be injected into
318
-------
hydraulic fractures along with slowly dissolving nutrients to stimulate in situ aerobic
biodegradation of organic compounds in soils. In addition, it is feasible to fill hydraulic fractures
with metal catalysts, such as iron particles, which Gillum and Burris (9) have proposed as a
method of degrading a wide range of organic compounds. By degrading compounds in situ it
should be possible to shorten the migration paths of contaminants. Periodically reversing the
polarity of the field is intended to repeatedly pass contaminants through a degradation zone
while limiting the development of high or low pH conditions in the vicinities of electrodes and
reducing fouling of electrodes by precipitation. This approach to in situ remediation is the
essence of the so-called "Lasagna" process (Fig. 1; ref 10).
Hydraulic fracturing appears to be a method of improving the performance of electrodes
and of creating zones where contaminants are degraded in situ. An investigation into this
application was initiated in the summer of 1994, with pilot-scale testing scheduled for 1995 and
full-scale testing at a contaminated site anticipated for 1996.
METHODS
+ + +
+ +
Figure 1. Configuration of fractures, electrical
delivery, and fluid delivery system. 1. DC Power
supply; 2. Water delivery apparatus; 3. Graphite-
filled hydraulic fracture as cathode; 4 anode.; 5.
water flow by electroosmosis; 6. graphite contactor
on the end of power cable; 7. Well casing.
Recent work on the project
has consisted of identification or
development of methods to monitor
in situ electrokinetics and
degradation, as well as preliminary
evaluation of hydraulic fractures as
electrically and hydraulically
conductive electrodes.
Monitoring
The monitoring efforts have
assembled techniques to measure
the following parameters
temperature
moisture content
fluid pressure/suction
electrical conductivity
electrical potential
pore fluid pH
pore gas composition
The configuration shown in Figure 1
will result in strong vertical
gradients, so monitoring systems
have been designed to obtain data
along relatively close vertical spacing in order to resolve those gradients. Moreover, we have
sought to measure several parameters at each location in order to reduce the number of
subsurface devices that must be installed. This effort has produced two types of devices, the T-type
and S-type, that will be used to measure the subsurface parameters.
The T-type of monitoring device consists of a 2.5-inch (nominal) PVC pipe that fits
snugly into a reamed borehole. The wall of the pipe is penetrated by stainless steel rivets, which
are flush on the outside and protrude slightly on the inside. A probe is lowered into the pipe to
contact each rivet and measure the electrical potential of that point relative to the cathode,
thereby providing a profile of electrical potential. When it is not being used to measure potential,
the pipe is thermally insulated and a thermocouple, which contacts the inner wall, is used to
measure the temperature with respect to depth. The pipe also provides access for a neutron
probe, which has been calibrated for losses due to neutron absorption by the PVC. Neutron
probe data will be used to estimate the moisture content of soils in the vicinity of the pipe.
Some of the 2.5-inch PVC pipes are wrapped with a series of stainless steel bands
riveted to their exterior. The bands are 1 cm wide and spaced every 5 to 10 cm along the axis of
319
-------
the pipe. A probe that can contact four bands simultaneously is lowered into the pipe and the
upper and lower bands are energized with approximately 10V of low frequency AC. The
potential difference between the middle two bands is proportional to the electrical conductivity of
soil in the vicinity. This modification of the T-type device will be used to monitor changes in
electrical conductivity during electrokinetics, which result from electrochemical processes and
have been reported in laboratory experiments (11).
The S-type monitoring system is used to measure fluid potentials and composition. It
consists of a suite of sensors that are pushed horizontally into the side of a vertical borehole. A
prototype device has been fabricated that pushes a tube into the sidewall of a bore to obtain a
soil sample and create a hole suitable for either a ceramic-cup tensiometer, or a porous soil-gas
sampler. Those devices are installed at various depths along the vertical bore to obtain pore
water pressure or suction, and pore gas pressure or composition as functions of depth.
Electrical potential
We created two roughly circular, horizontal hydraulic fractures filled with electrically
conductive graphite to evaluate their effect on the electric field in the subsurface. The hydraulic
fractures were approximately 3 m in radius, 6 mm thick and contained graphite whose electrical
conductivity was approximately 190 S/m when saturated with water. They were separated by 0.9
m vertically, sandwiching roughly 25 m3 of soil between them. The fractures were enveloped in
silly clay soil whose electrical conductivity was 0.05 to 0.1 S/m, or 2000 to 4000 times less than
that of the fracture.
A potential difference of 40 V DC
was maintained between the two fractures
with a current of 30 amps. The potential
difference could be adjusted by changing
the current, with a ratio of
current:potential difference of 0.75
amps/volt observed for a wide range of
power conditions. The design of the
constant current power supply used during
the initial tests limited the potential
difference to 40 V for continuous
operation-greater potential differences
could be applied for testing over short
durations. T-type monitoring tubes were
placed at radial distances of 0.6,1.5, 2.6
and 3.5 m from the power electrodes (Fig.
2), and rivets were spaced every 7.5 to 15
cm along the length of the tubes.
Preliminary theoretical analyses
of the electric field were conducted using
both analytical solutions to the potential in
0.0°
i.O
-1.0
-2.0
-3.0
-4.C
1.0
2.0
3.0
4.0
5.'
°0.0
-1.0
-2.0
-3.0
1.0
1.0
2.0
3.0
4.0
5.0
,-4.0
Figure 2. Axisymmetric view of electrical potential
distribution in the vicinity of two charged disks
representing hydraulic fractures. Potential is with
respect to the upper fracture and is normalized to
the potential difference between the fractures.
Vertical bands are T-type casings used to measure
potential.
the vicinity of a charged circular disk and
numerical solutions to the potential in the
vicinity of a highly conductive disk. The analytical solution indicates that under ideal conditions
of a homogeneous, isotropic medium, the electrical field around two, horizontal, conductive,
circular disks beneath an insulated free surface is as shown in Figure 2.
RESULTS
The theoretical analyses indicate that there should be strongly positive vertical gradients
between the fractures. Indeed, the strong vertical potential gradients are intended to drive the
electrokinetic process. The gradients are more subtle in other regions, with weakly negative
320
-------
vertical gradients over the fractures and weakly positive gradients at radial distances beyond the
edge of the fractures (Fig. 2).
Electrical potential was measured along the T-type casings at distances of 0.6,1.5, 2.6,
and 3.5 m, the data were normalized to the applied potential difference and compared to the
results of the theoretical analysis (Fig. 3). The field data show that strong vertical electrical
potential gradients are produced between the hydraulic fractures, at least within the region
bounded by the radial extent of the fractures (about 3 m). The subtle predictions of the character
of the electrical field are represented in the field data, with gentle negative gradients overlying
the fractures and gentle positive gradients beyond the edge of the fractures. The potential in the
vicinity of the lower fracture is less than predicted by the analysis, perhaps because that fracture
is thinner than assumed in the analysis.
The field data and ratio of applied
current to potential difference given
above indicates that the vertical electrical
potential gradient available for
electrokinetics is
0.0 0.2 0.4 0.6 0.8 1.0
Normalized voltage
0.0 0.2 0.4 0.6 0.8 1.0
Normalized voltage
0.0 0.2 0.4 0.6 0.8 1.0
Normalized voltage
0.0 0.2 0.4 0.6 0.8 1.0
Normalized voltage
Figure 3. Electrical potential as a function of
depth at various distances from the power
electrode. Line is from theoretical analysis, points
are field data.
dVldz=CQIh
d)
where Q is the applied current, h is the
fracture spacing, and C is a constant that
depends on the design of the fractures.
The available results indicate that C = 0.6
to 0.9 V/amp for the design used here.
Accordingly, for an applied current of 30
amps, the potential gradient is 0.2 to 0.3
V/cm. Gradients 0.2 to 0,4 V/cm are
reported during field tests by Langeman
and others (1989), and slightly greater
gradients are described from laboratory
experiments; e.g. Bruell and others (12)
used gradients of 0.4 V/cm, whereas Acar
and others (11) report gradients of 0.5
V/cm or more. Potential gradients of 0.4
V/cm appear to be achievable over most
of the 25 m3 of soil between the fractures
by increasing the current provided by the
power supply to approximately 50 amps.
CONCLUSIONS
The field observations indicate that graphite-filled hydraulic fractures behave as broad,
sheet-like electrodes capable of providing useful electrical potential gradients distributed over a
reasonably large volume of soil. These results indicate hydraulic fractures can be used to create
the electrical field required to conduct the Lasagna process outlined by Ho and Brodsky (10).
Preliminary measurements of fluid flows suggest that the observed potential gradients will induce
electroosmotic migration of water through the region between the fractures. A pilot test designed
to provide detailed data on electroosmotic flow is underway, and those data should be available
in the near future.
REFERENCES
1. Murdoch, L.C.. A field test of hydraulic fracturing in glacial till. Proc. 15th Annual USEPA
Research Symposium, Cincinnati, Ohio, 1990.
321
-------
2. Murdoch, L.C., G. Losonsky, P. Cluxton, B. Patterson, I. Klich, B. Braswell. The feasibility of
hydraulic fracturing of soil to improve remedial actions. Final Report USEPA 600/2-91-012.
NTISReportPB91-181818. 298,1991.
3. Shapiro, A. P. and Probstein, R. F. Removal of contaminants from saturated clay by
electroosmosis. Environ. Sci. Tech. 27:283-287,1993.
4. Acar, Y.B., R.J. Gale, G.A. Putnam, J. Hamed, and R.L. Wong. Electrochemical processing
of soils: theory of pH gradient development by diffusion, migration, and linear convection.
J. Eviron. Sci. Health, A25(6), 687-714, 1990.
5. Probstein, R, F. and Hicks, R. E. Removal of contaminants from soils by electric fields.
Science 260:498-503.1993.
7. Lageman, R., Pool, W. and Seffinga, G. Electro-reclamation theory and practice. Chem. Ind.
London. 18:575-579, 1989.
8. Vesper, S. J., L.C. Murdoch, S. Hayes, and W.J. Davis-Hoover. Solid oxygen source for
bioremediation in subsurface soils. Journal of Hazardous Materials 36:265-274,1994.
9. Gillham, R.W. and D.R. Burris. In-situ treatment walls - Chemical dehalogenation,
denitrification, and bioaugmentation. Proceedings of Subsurface Restoration Conference,
Dallas, Texas, June 21-24, 66-68,1992.
10. Ho, S.V. and P.H. Brodsky. Integrated In Situ Technology for Soil Remediation-Trie
Lasagna Process. Presented at Amer. Chem. Soc. Annual Meeting, Atlanta, GA, Sept. 19-
21, p 504-506, 1994.
11. Acar, Y. B., Li, H. Gale, R. J. Phenol removal from kaolinite by electrokinetics. J. Geotech.
Eng. 118:1837-1841,1992.
12. Bruell, C. J., Segall B. A., and Walsh, M. T.. Electroosmotic removal of gasoline
hydrocarbons and TCE from clay. J. Environ. Eng. 118:68-74,1992.
For more information: Mike Roulier, EPA Project Officer
US EPA Center Hill Lab
5995 Center Hill Rd.
Cincinnati, OH 45223
513-569-7796
322
-------
HYDRAULIC FRACTURES AS
ANAEROBIC AND AEROBIC BIOLOGICAL TREATMENT ZONES
Wendy Jo Davis-Hoover, Ph. D., Michael Roulier, Ph.D., L. Taras Bryndzia, Ph. D.
Risk Reduction Engineering Laboratory (RREL), 5595 Center Hill Ave., Cincinnati, Ohio 45224
513-569-7206, 513-569-7796, 513-569-7857; fax 513-569-7879
Jonathan Herrmann,P.E., Leland Vane, Ph.D.
RREL, 26 W. M.L King Jr. Dr., Cincinnati, Ohio 45268
513-569-7839, 513-569-7799; fax 513-569-7787
Lawrence C. Murdoch, Ph. D., Stephen J. Vesper, Ph.D.
University of Cincinnati, Cincinnati, Ohio
513-556-2568, 513-556-2538
Hydraulic fractures were initially developed to be used in bedrock to enhance oil recovery. We
have modified the procedure for use in the environmental field in shallower subsurface soils to
increase the permeability of the soils to liquids, gases, and solids. This has allowed us to significantly
enhance in situ methods of soil cleaning (pump and treat, steam injection, solvent extraction) (1) and
aerobic bioremediation (2) .
We are now concentrating on enhancing movement of contaminants between hydraulic
fractures and into hydraulic fractures using electrokinetics and destruction of contaminants within the
hydraulic fractures. Bioremediation, very specific to the unusual conditions that electrokinetics
presents to the subsurface soil is being developed. This work is being done as a result of the RTDF
and in conjunction with researchers at Monsanto, Dupont, General Electric, and DOE.
Hydraulic fracturing allows for the insertion of a pancake-shaped lens of sand. The sand lens
increases the soil permeability of the area allowing a 10 fold increase in the area of influence of the
well allowing for a 10 fold increase in vapor extraction, and a 10 fold increase in liquid addition (such
as solubilized nutrients and hydrogen peroxide), which leads to a 100 fold increase in aerobic
bioremediation (1). The process also allows for other.solid addition to contaminated subsurface soils
such as solid slow-releasing oxygen which increases aerobic bioremediation (2, 3, and 4).
The Lasagna technology uses 2 outer hydraulic fractures as two dimensional electrodes in the
electrokinetic process moving contaminants into 3 inner hydraulic fractures designed to be zones of
biodegradation (5). Several experiments that the members of the research consortium have performed
have indicated that the process of electrokinetics modifies the pH of the soil to a spectrum of 2 to 10
and increases the temperature significantly. Thus as a result of this we decided to naturally select for
thermophilic organisms that will degrade contaminants at these high and low pHs .
We have been successful in obtaining thermophilic organisms that will degrade contaminants
at high and low pHs (6). These are being tested in lasagna microcosms (6). The next step is to test
them in the field.
REFERENCES
1. U. S. Environmental Protection Agency. Hydraulic Fracturing Technology: Applications Analysis
and Technology Evaluation Report. EPA/540/R-93/505, U. S. Environmental Protection Agency,
Cincinnati, Ohio, 1993. 60 + pp.
2. Vesper, S. J., L. C. Murdoch, S. Hayes, and W. J. Davis-Hoover. Solid Oxygen Source for
323
-------
Bioremediation in Subsurface Soils. Journal of Hazardous Materials. 36:265-274, 1994.
3. Vesper, S., L. Murdoch, W. Davis-Hoover. Oxygen Pellets Spike Bioremediation. Soils. May
1994. p 14-17, 1994.
4. Davis-Hoover, W. J. , L. C. Murdoch, S. J. Vesper, H. R. Pahren, O. L. Sprockel, C. L. Chang,
A. Hussain, W. A. Ritschel. Hydraulic Fracturing to Improve Nutrient and Oxygen Delivery for
In situ Bioreclamation.ln; R. E. Hinchee and R. F. Olfenbuttel (eds.) In Situ Bioreclamation.
Butterworth-Heinemann, Boston. 1991. p. 623.
5. Murdoch. L. C., J.-L. Chen, P. Cluxton, and M. Kemper. 1995. Hydraulic Fractures as
Subsurface Electrodes: Early work on the Lasagna Process, im RREL. Symposium
Proceedings, paper before this.
6. Vesper, S. J,, etal. 1995. Advances in TCE-degrading Thermophilic, Mesophilic and Alkaline-
Resistant Organisms . hr RREL Symposium Proceedings, paper after this.
FOR MORE INFORMATION: Wendy Jo Davis-Hoover, Ph. D.
RREL, 5595 Center Hill Ave., Cincinnati, Ohio 45224
513-569-7206, fax 513-569-7206
324
-------
ADVANCES IN TCE-DEGRADING THERMOPHILIC. MESOPHILIC. AND ALKALINE-
RESISTANT ORGANISMS
Stephen J. Vesper
University of Cincinnati •'''
1275 Section Road
Cincinnati, Ohio 45237-2615
513-5562538
Wendy J. Davis-Hoover
The U.S. Environmental Protection Agency
5595 Center Hill Road
Cincinnati, Ohio 45424
' 513-569-7206
Michael H. Roulier
The U.S. Environmental Protection Agency
5595 Center Hill Road
Cincinnati, Ohio 45424
513-569-7796
Lawrence C. Murdoch
University of Cincinnati
1275 Section Road
Cincinnati, Ohio 45237-2615
513-556-2568
Leland M. Vane
The U. S. Environmental Protection Agency
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
513-569-7799
Jonathan G. Herrmann
The U. S. Environmental Protection Agency
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
513-569-7839
INTRODUCTION
Trichloroethylene (TCE) is a chlorinated organic solvent that has been widely used in metal
processing, electronics, dry cleaning, paint and many other industries. The methylotrophic bacteria have
been shown to co-metabolize TCE. These organisms have been utilized as degraders of TCE in aquifers
(Semprini et al. 1987) which suggests that the use of methylotrophs is a practical solution for TCE
contamination.
325
-------
Electroosmosis is a relatively old technique used to dewater soils (Casagrande, 1949). Electroosmosis
uses a direct-current electric field to induce a motion of a liquid and dissolved ions that transports the
contaminants in the electric field. This technique is limited to fairly soluble contaminants Like benzene,
toluene, xylene, phenol, and chlorinated solvents (Bruell et al. 1992; Acar et al. 1992; Shapiro and
Probstein, 1993). This technique creates highly acidic and basic zones in the soil. The process can also
raise the temperature of the soil substantially. We are selecting methylotrophs to survive and degrade
TCE in this environment.
METHODOLOGY
Soil samples were taken at 10 major locations in Yellowstone Park and shipped back to the laboratory
the next day. When the samples arrived in the laboratory, they were placed in water baths at 50, 60, 70,
or 80°C, whichever temperature was closest to the temperature from which they came. Five grams of
sample were placed in 20 ml of "L" medium (ATCC, 1984) in serum stoppered bottles. These were then
purged with 0.6% methane in air. The bottles were then placed in water baths at the appropriate
temperature. After 10 days, the bottles were sub-sampled by placing 1 ml of the aqueous phase on L agar
plates containing 2% agar. These plates were placed in desiccators which were sealed then purged with
the methane:air mixture for 30 minutes. The desiccators were then placed in incubators at the appropriate
temperature. As different colonies appeared they were picked and placed in L medium or L medium
diluted 1:10 with water and returned to serum stopped bottles. These were grown-up in the methane in air
mixture and restreaked for purity. These isolates were then ready for screening for TCE degradation
Each isolate was grown in duplicate 60 ml serum stoppered bottles for at least one month in 20 ml of L
medium, L medium diluted 1:10 with deionized water, or AMS medium (Whittenbury et al. 1970) or
AMS medium diluted 1:10 with deionized water. The medium was adjusted to pH 2, 4, 8 or 10 with
concentrated NaOH or HC1. The bottles were periodically purged with a mixture, of 2.5% methane in air
to provide a carbon source.
After the growth period of at least one month, each isolate was tested for TCE degradation. The
methane was first removed by placing them in the laminar flow hood for 1 hour. At that point, the serum
bottle was resealed. Two microliters of a TCE stock solution were added to each bottle so that the final
concentration of TCE added was approximately 10 ppm. On the same day that the serum bottles received
the TCE, the starting concentrations of TCE were determined. TCE concentration was measured using a
GC-FID. Headspace samples (0.5ml) were removed from the serum bottles and injected into the GC. On
the fifth day, the bottles were resampled for TCE concentration. The concentration of TCE obtained from
the second analysis was then subtracted from the initial value of TCE to indicate the level of TCE
degradation or disappearance.
A model electroosmosis microcosm (EM) was developed to test the bioremediation process. This
design now allows for maintenance free operation for at least a month. This design also allows for the
manipulation of various components of the EM to test concepts for field applications. The EM chamber is
20 cm long and 10 cm wide and is made of polyethylene phthalate (PET). (The chamber is a two liter soft
drink bottle). The bottom of each EM is filled with 400 grams of sand. This acts as a reservoir for water.
On top of the sand, a mixture of 100 grams of granular graphite and 50 grams of granular activated
carbon is layered. This makes up the anode. Next a soil slurry is made. It contains 1500 grams of soil
obtained from subsurface soil environment. The soil is mixed for one hour in a Hobart mixer with 600 ml
of tap water. The whole soil slurry can be added to make a complete soil column. If a remediation zone
326
-------
is to be added, half of this mixer is then gently poured onto the anode electrode. The remediation zone is
created by adding 150 grams of granular activated carbon on top of the first layer of soil plus 50 ml of tap
water. Then the rest of the soil is added gently on top of the remediation layer. Another electrode,
identical to the first, is then poured on top of this soil. This is the cathode. On top of the cathode is added
400 grams of sand and then the rest of the chamber is filled with tap water. A Mariotte Bottle is used to
control the water level. A gas relief valve is added to the anode to remove oxygen gas generated process.
Tests were made using the EM. Some tests were run at constant 100 volts for 6 hrs. All other tests were
run at a constant 25 volts. (The soil temperature does not exceed the ambient temperature at this voltage.)
At the end of the test, the pH of the soil is taken with pH paper every 2 cm along the soil column length.
RESULTS
The screening program to date has identified 20 (Table 1) isolates that appear to have degraded more
than 20 % of the TCE in 5 days; two isolates showed more than 30% degradation in 5 days. Thirteen of
the isolates came from 50°C (total of 59) and seven came from 60°C (total of 27). Most isolates came
from more basic pH's. Ten came from pH 8 and 3 from pH 10 but six came from pH 4 conditions. We
have no pH 2 isolates that reached 20 % degradation within 5 days but we have one pH 2 isolate that was
close to 20%.
In general the low pH isolates (pH 2 and 4) degraded more TCE when grown in the AMS medium at
full strength (data not given). Similarly, the high pH isolates (pH 8 and 10) degraded more TCE in full
strength L medium. However, there were exceptions in each case.
Some EMs were run for 6 hrs at 100 volts to simulate a long field experiment. The temperature rose
as high as 82°C in these chambers. At 25 volts, the chambers stay in equilibrium with the ambient
temperature. When the EM is run at 25 V, the current and monitoring voltage both started high but
dropped quickly. The purge water started flowing immediately but slowed after about 400 hours. It took
nearly 600 hours to move 2 pore volumes (pore volume is 969 ml).
If an activated carbon layer was added in the middle of the soil column, the current started out higher
than in the soil alone configuration but the current rapidly dropped to about 35 mA. Voltage also started
very high, declined then rose again. About 3 pore volumes were moved in 300 hours. The results of the
pH analysis of the soil at the end of the experiments indicated that the pH near the anode was 4 or higher
but near the cathode the pH was as high as 12 (data not given).
CONCLUSIONS
It must be noted that we have not proven that any of these isolates are methylotrophs nor that the TCE
was biodegraded. These questions will await further analysis. It does appear that the isolates we have
obtained have the capacity to survive the extreme pH and high temperatures expected in the
electroosmosis process. So far these organisms grow very slowly and we have much to learn about the
optimal growth conditions. The design of the EM has allowed us to create conditions expected in the field
327
-------
TABLE 1. SUMMARY OF ISOLATES TESTED FROM YELLOWSTONE NATIONAL PARK,
. WYOMING
LOCATION # ISOLATES # DEGRADATION > 20% % TOTAL
A. Mammoth Hot Springs (MHS)
B. Norris Geyser Basin (NGB)
C. Madison Junction (MAD)
D. Paint Pot (PP)
E. V Road Site (VRS)
F. Near Obsidian Cliff (OBS)
G. Mud Volcano Site (MVS)
H. Above Sulfur Caldron (ASC)
I. Calcite Spring Site (CSS)
J. Artist Paint Pots Site (APPS)
Total
16
8
6
7
o
2
4
4
. 19
20
86
2
1 -
1
1
b
i
0
0
7
7
20
13
13
17
14
0
50
0
0
37
35
24
and this should be a useful laboratory model to test electroosmosis. Next, the isolates will be tested in the
EMs for biodegradation of TCE.
REFERENCES
1. Semprini, L., Roberts, V. P., Hopkins, G. D., and Mackay, D. M. 1987. A Field Evaluation of In-Situ
Biodegradation for Aquifer Restoration. U.S. EPA. EPA/600/2-87/096. Washington, D. C.
2. Casagrande, L. 1949. Electroosmosis in Soils. Geotechnique 1(3): 159-166.
3. Bruell, C. J., Segall, B. A., and Walsh, M. T. 1992. Electro-Osmotic Removal of Gasoline Hydrocarbons
and TCE from Clay. J. Environ. Eng. 118 (4): 68-74.
4. Acar, Y. B., Li, H., and Gale, R. G. 1992. Phenol Removal from Kaolinite by Electrokinetics. Geotech.
Eng. 118:(2):1837-1841.
5. Shapiro, A.P. andProbstein,R.F. 1993. Removal of Contaminants from Saturated Clay by
Electroosmosis. Environ. Sci. Tech. 27 (3):283-287.
6. ATCC. 1984. Media Handbook. American Type Culture Collection, Rockville, MD.
7. Whittenbury, R., Phillips, K. C. and Wilkinson, J. F. 1970. Enrichment, Isolation and Some Properties
of Methane-utilizing Bacteria. J. Gen. Microbiol. 61:205-218.
328
-------
INTEGRATED PNEUMATIC FRACTURING AND BIOREMEDIATION FOR THE IN-SITU
TREATMENT OF CONTAMINATED SOIL
Sankar N. Venkatraman and David S. Kosson
Rutgers, The State University of New Jersey
Department of Chemical and Biochemical Engineering
P.O. Box 909
Piscataway, NJ 08855-0909
Tel.: (908)445-4346
Thomas M. Boland and John R. Schuring
New Jersey Institute of Technology
Department of Civil and Environmental Engineering
Newark, NJ 07102
Tel.: (201)596-5849
Uwe Franks
US. EPA - ORD
Releases Control Branch (MS-106)
2890 Woodbridge Avenue
Edison, NJ 08837-3679
Tel.: (908)321-6626
INTRODUCTION
In-situ bioremediation is often limited by the rate of transport of nutrients and electron acceptors
(e.g., oxygen, nitrate) to the microorganisms, particularly in soil formations with moderate to low
permeability. An investigation was conducted to integrate the process of pneumatic fracturing with
bioremediation to overcome these rate limitations. Pneumatic fracturing is an innovative technology which
utilizes high pressure air to create artificial fractures in the contaminated geologic formations, resulting in
enhanced air flow and transport rates in the subsurface. Following the fracturing, the pneumatic fracturing
system can be used to inject electron acceptors and other biological supplements directly into the
formations to stimulate biodegradation.
Pneumatic fracturing is an innovative technology which enhances the in situ removal and
treatment of contaminants from low permeability soil and rock formations. The process may be generally
described as injecting air (or another gas) into a contaminated geologic formation at a sufficient pressure
and flow rate so that artificial fractures are created. Once established, the fractures increase the
permeability of the formation, thereby making contaminant removal and/or treatment more efficient.
Pneumatic fracturing is designed to be integrated with other in situ treatment technologies such as vapor
extraction, bioremediation, thermal injection, and pump and treat. The technology can be applied to a
range of hydrogeologic conditions including vadose zones, perched water zones, and saturated zones.
Research which led to the development of pneumatic fracturing began in 1988 with bench scale
studies and analytical modeling at the Hazardous Substance Management Research Center (HSMRC)
located at New Jersey Institute of Technology (NJIT). The first industrial site demonstration of pneumatic
fracturing was performed in 1990 at a site in Richmond, Virginia. In 1992, the technology was evaluated
under the U.S. EPA Superfund Innovative Technology Evaluation (SITE) Demonstration Program at a
contaminated industrial site in New Jersey for enhancement of a soil vapor extraction system [USEPA,
1993}. Pneumatic fracturing is now available commercially through Accutech Remedial Systems for
329
-------
applications involving soil vapor extraction and pump and treat enhancement, and is being applied to
production clean-ups. Pneumatic fracturing is a patented process, and the assignee is HSMRC.
The sequence of steps required to apply the integrated pneumatic fracturing/bioremediation
process are as follows. During the first step, the formation is pneumatically fractured by inserting a
proprietary device known as an "HQ Injector" into a drilled well. The HQ injector applies pressurized air
along a discrete two foot interval, and is subsequently repositioned at various elevations to create a
fracture network. The injector is then moved to additional bore holes and the process repeated until the
entire contaminated formation is fractured.
After the formation has been initially fractured with air, the second step is to introduce nutrients
into the fracture network to provide substrates for enhance biological activity. This is accomplished by
inserting liquid nutrients into the main injection air stream which are then dispersed throughout the already
established fracture network. By maintaining a high air to liquid ratio, the liquid supplements actually
become atomized during injection, which increases their ability to penetrate the formation. Additional
nutrient injections are made periodically to replenish the substrates, and to provide beneficial aeration.
Optionally, other biological supplements such as innoculum and/or granular media could be injected into
the fracture network.
In the final step of the process, the site is operated as in situ bioremediation cell to degrade the
contamination. A continuous air flow is maintained throughout the fracture network via a vacuum pump
connected to a central extraction well. Outlying wells vented to the atmosphere serve as a passive source
of oxygen. In the present application, air flow was purposely maintained at a low level, to minimize off-gas
treatment and to maximize biological effects.
It is impo |