IN SITU FLUSHING & SOILS WASHING
TECHNOLOGIES FOR SUPERFUND SITES
Presented at:
RCRA/Superfund Engineering
Technology Transfer Symposium
By:
Hazardous Waste Engineering Research Laboratory
U.S. Envirpnmental Protection Agency
Cincinnati, OH 45268
SEPA
RICHARD P. TRAVER, P.E.
•f £D • STAFF ENGINEER / UNIT DIVING
RELEASES CONTROL BRANCH '
MaurttoiM WM««. EnglnMrlng I
United SIMM EnvlronnMnUl frM*ctim Ag«ncy (201) 321-0977
(FTS) 34O4477
-------
IN SITU FLUSHING & SOILS WASHING
TECHNOLOGIES FOR SUPERFUND SITES
Presented at:
RCRA/Superfund Engineering
Technology Transfer Symposium
By:
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
&EPA
RiCHARD P. TRAVER, P.E.
STAFF ENGINEER / UNIT DtVWO OFFICER
RELEASES CONTROL BRANCH
Hazardous Wa
ring
Unn*d SutM ErMbwMMnut Preiwiieii Agmcy (201) 3Z1-4677
WoodMdg* AVWMM • Edlaon. NM» J«My (WC374C7*
-------
TOE DEVELOPMENT OP CHEMICAL COUNTER MEASURES
FOR HAZARDOUS WASTE CONTAMINATED SOfLj
W. D. BIIIc end J. R. Payne
JRB Associates
IfeLeaa. Virginia BIOS
A.N.TaJuriandF.J.
OO and Hazardous Materials Spills Branch
Municipal Environmental Rusweh Laboratory
U.S. Environmental Protection Agency
-- Hew Jersey OUST
ABSTRACT
The U.S. Environmental Protection Agency's
(EPA) OH and Hazardous Materials Spills Research and
Development Program in Edison, New Jersey, has de-
signed a Chemical Countermeasures- Program to evalu-
ate in situ methods for mitigating or eliminating envir-
onmental damage from releases of toxic and other
hazardous materials to the soils around uncontrolled
hazardous waste disposal sites, and from spins of haz-
ardous chemicals to stm or relatively slow-moving
surface water bodies. To date efforts nave concentrat-
ed on sods-related activities to determine whether use
of aqueous surfactants could significantly enhance the
In ritu cleanup of chemically contaminated soils with
standard water washing techniques.
Laboratory studies were performed to determine
the maximum cleanup efficiency under equilibrium
conditions using water washes and a combination of 2
percent each Hyonic PE90 (now known as NP90, Dia-
mond Shamrock), and Adsee 799 (Wltco Chemical)
surfactants and to evaluate son cleanup efficiency
under gravity flow conditions, fa general, overall sod
cleanup approaching the 90-plus percent level was
attained with intermediate molecular weight aliphatic
and aromatic hydrocarbons, polyehlorinated biphenyl
mixtures and chlorinated phenol mixtures. Results
appear to support larger scale field demonstrations, and
plans are being discussed to conduct fun-scale, con-
trolled tests at appropriate hazardous waste or apOl
sites ("dtes-of-opportunity").
The work reported herein was performed by JRB
Associates under UJ. Environmental •Protection
Agency contract No. 81-03-3113, Task J9. The
content of this publication does not necessarily
reflect the views or policies of the UJ. Environ-
mental Protection Agency, nor does mention of
trade names, commercial products, or organizations
Imply endorsement by the U.S. Government.
BACKGROUND
The Comprehensive Environmental Response,
Compensation, and Liability Act of 1980 or Superfund
recognizes the need to develop eountermeasures (mech-
anical devices, and other physical, chemical, and biolog-
ical agents) to mitigate, the effects of hazardous sub-
stances that arc released Into the environment and to
clean up Inactive hazardous waste disposal site*. One
key eountermeasure Is the use of chemicals and other
additives that are intentionally introduced into the
environment to control the hazardous substance. The
indiscriminate use of such agents, however, poses •
distinct possibility that the release situation could be
made worse by the application of an additional chemical
or other additive.
The UJS. Environmental Protection Agency's
(EPA) OO and Hazardous Materials Spills Branch in
Edison, New Jersey, has begun a support program to
define technical criteria for the use of chemicals and
other additives at release situations of hazardous sub-
stances. The criteria are to ensure that the combina-
tion of the released substance plus the chemical or
other additive, Including any resulting reaction or
change, results in the least overall harm to human
health and the environment.
The Chemical Countermeesure Program (CCP)
has been designed to evaluate the efficacy of in situ
treatment of large volumes of subsurface soils, such as
found around uncontrolled hazardous waste sites, and
treatment of large, relatively quiescent waterbodies
contaminated with spCDs of water-«plub!e hazardous
substances. For each situation, the following activities)
are planned!
• • literature search to develop the existing body of
theory and date
• laboratory studies on candidate chemicals to assess
adherence to theory and define likely candidates for
fun-scale testing
• fun-scale, eontroDed tests at a dte-of-opportunlty
This paper presents the results of the Information
search and laboratory studies for the soils-related
activities of the program.
-------
CLEANUP 117
INFORMATION SEARCH
A literature search of limited scope was per-
formed to gather information on the state-of-the-art in
chemical countermeasures. The emphasis of the search
was on the most recent and Innovative work on the
subject, and on work most likely to be fruitful for
further development. The search was primarily directed
toward readdy available publications on subjects related
to chemical countermeasures, and toward contacting
key people doing research, development, and field
implementation of chemical countermeasures.
The application of chemical countermeasure
techniques in the field has been very limited. Th* main
reasons are caution and scarcity of Information/experi-
ence. Current technology for removal of contaminants
from large volumes of sods (too large to excavate
economically) having relatively low to moderate levels
of contamination has been limited to withdrawal of
groundwater, with or without recharge to th* sod, U*.,
In situ •water washing."
Accordingly, the laboratory studies were design-
ed to determine whether adding aqueous surfactants to
recharge water used in a continuous recycle could
significantly enhance the efficiency of contaminated
sods cleanup by water washing. Based on the literature,
it was thought that surfactant mixtures would improve
th* solvent properties of the recharge water, thereby
enhancing th* removal of chemical contaminants ad-
sorbed onto sod particles. This approach was a direct
derivative of th* laboratory studies performed by th*
Texas Research Institute for the American Petroleum
Institute on the use of surfactants for enhanced gasottn*
recovery from sand (Tax** Research Institute, 19M).
LABORATORY PROGRAM DESIGN
The experimental design of the laboratory pro-
gram was formulated after reviewing th* results of
simdar Investigations identified during th* information
search. The primary purpose of th* laboratory studies
was to determine whether us* of aqueous surfactants
could significantly enhance the in situ cleanup of chem-
ically contaminated sod* by standard water washing
techniques. A secondary objective (assuming the pri-
mary goal was successful) was to obtain Information and
make recommendations for designing larger seal* tests
under controlled condition* and field test* at sit** of
opportunity.
• Before conducting the laboratory studies, four
specific Issues had to be resolved. The first Issue was to
Identify and select • suitable sod to b* used In th*
laboratory tests and Included sod characterizations and
evaluation of permeability versus compaction para-
meters. Th* second issue Involved contaminant atlee-
tion and determination of the concentrations required
for sods studies. Th* third issue dealt with surfactant
selection, surfactant solubility, compatibility with sod
type, and efficiency of pollutant removal. Th* fourth
Issue Involved th* analytical methods to b* used for
extraction and analysis for th* pollutant group* of
interest in the sods and leachatca.
SeleetlonofTestSoa
In choosing th* sod to be used In the tests, native
sods at each of 10 Region d Superfund sites were tdenti-
fied to determine the most commonly occurring sod
series. Once determined, a sod type of the same taxo-
nomic classification was located in the vicinity of the
potential larger-wale test facility that could be excava-
ted and used in the experiments. The most commonly
occurring classification was Typie Hapludults (Freehold
sod series), a fine-to-coarse loamy sod of humid
climates, containing cones of clay accumulation. In
addition to taxonomic classification, a permeabdity
rating of 10*** to 10"* cm/s was specified as a desirable
range.
Table I presents; the grain size distribution
obtained by wet sieve and pipette analyses. Approxi-
mately 95 percent of the theoretical surface area Is
represented by fines (IS percent sQt and 8 percent clay).
To determine the mineralogical composition of
the Freehold sofl, x-ray diffraction studies were under-
taken. The .results showed quartz and feldspar to be the
only measurable constituents. Quartz was the major
phase, representing at least 91 percent of the total
weight. No measurable amounts of clay minerals ap-
peared.
The total organic carbon content (TOC) of the
sod wss determined on a sample of sod prepared by
grinding and suspending in an aqueous solution of phos-
phoric add and sodium phosphate, In accordance with
EPA Method 413.1. The TOC value was 0.12 percent by
weight. This relatively low level of organic matter In
the sod Implies a relatively low adsorption potential for
organic contaminants.
The cation exehang* capacity (CEO of the sod
also was determined by the methods of Jackson UtTO),
and the results war* combined to yield th* total CEC.
The result was 14 mdUequivalents par 100 grams, an
extremely low value, confirming the absence of miner-
alogic clay la the sod.
Selection of Contaminant*
The compounds used for tasting in the laboratory
were chosen on th* basis of several criteria. They
should:
• occur frequently In high enough concentrations in
th* son surrounding Superfund sites
• present a significant hazard to human health and
th* environment
• have lost to moderate mobility and high persistence
in sod
Tabtel. Or*Jasi»»d»u^tIcn of Freehold sod by w*t
at*** and psp*tt* analys** (ModlfM ASTM O-4» using
H1 Ih**aii of ffnfli si*T* siisrt
Theorttieal
Class Size rang* Mass surf ace area
turn) (percent) (percent)
Gravel >1000 If
Sand 6* to 1000 01
SOt t to «1 IS
Clay <• •
9
M
• 1
-------
118 1984 HAZARDOUS MATERIAL SPILLS CONFERENCE
• be treatable by an existing chemical method
• have an appropriate chemical analog, If too hazard-
ous or expensive for experimentation
Data were gathered on the concentrations,
frequency of occurrence, soil adsorption, and toxicity of
waste chemicals found at Superfund sites. The inci-
dence of the various hazardous waste and waste classes
is given in Table 2. The data on soQ contaminants
indicate that the most widespread class of contaminant
is the slightly water-soluble organics, which includes
low molecular weight aromatics and halogenated hydro-
carbons. The next most common contaminant classes
are heavy metals and hydrophobie organics. Clearly,
the occurrence of phenols also is widespread.
Based on this review and analysis, three pollutant
compound mixtures were selected for use in soQ test-
Ing! (1) intermediate and high molecular weight alipha-
tic hydrocarbons and polynuclear aromatic hydrocarbons
(high oofling distillation fraction of Murban crude OH),
(2) polychlorinated biphenyl mixture In ehlorobenzenes
(Askarel8$, and (3) di-, trl=, and pent&ehlarophenols.
Selection of Surfactants
The preliminary selection of 2 percent Richonate
YLA and 2 percent Hyonic NP90 as the surfactant
mixture was based on the results of a Texas Research
Institute study (Texas Research Institute, 1979) evaluar.
ting the removal of gasoline from pure Ottawa sand.
After initial studies, however, this mixture was found to
be unsuitable due to its marked tendency to suspend the
silt- and clay-size grains (less than 63 urn in diameter),
which resettled in small pores, thereby inhibiting col-
umn flow.
Beaker studies then were conducted to evaluate
solubility properties of the surfactants and their ten-
dency to disperse the fine clay-size particles present in
the Freehold sod.
The decision was made to use a combination of 2
percent Adsee T99 and 2 percent NP90, non-ionic sur-
factants, based on the mixture's:
• high water solubility
• ability to disperse Murban hydrocarbons
• minimal suspension of fine soil particles
• lower content of compounds that eause analysis
interferences than previously tested surfactants
Table 2. Hazardous oO contaminants at Superfund rite
Table*. Hi
oil contaminants at Superfund stU
SoQ Contaminants
Number
of sites
Example
Heavy metal wast*
Chromium
Arsenic
Lead
Zinc
Cadmium
' Iron
Copper
4T
Son Contaminants
Mercury
Selenium
Nickel
Vanadium
Fly ash
Plating wastes
Other Inorganics
Cyanides
Acids
Alkalis
Radioactive wastes
Number
of sites
2
2
1
1
1
2
26
Examples
/
8
7 sulf uric acid
6 lime, ammonia
3 uranium mining and
Miscellaneous
3S
Hydrophobie organics
PCBs IS
OQ. grease 11
Volatile hydrocarbons 6
Chlorinated hydrocarbon S
pesticides
Polynuclear aromaties 1
Slightly water soluble
organics
Aromaties
Benzene
Toluene
Xylene
Other aromatic*
64
9
9
5
3
Halogenated hydrocarbons
Trichloroethylene 11
Ethylene diehloride 6
Vinyl chloride 4
Methylene chloride 3
Other halogenated IS
hydrocarbons
HydrophOlc organies 20
Alcohols 4
Phenols IS
Other hydrophQlcs 4
Organic solvents 30
(unspecified)
and other organics
purification wastes,
radium, tritium
beryllium, boron
hydride, sulfides.
hexane, Varsol
endrin, lindane, DDT,
2,4,5-T, dieldrin
styrene, naphthalene
chloroform,
trichloroethane,
tetrachloroethylene,
trichlorofluoro-
methane
methyl, isopropyU
butyl
picric acid,
pentacrdorophenol,
creosote
dioxane, bis (2-ehle
ethyl) ether,
urethane, rocket fuel
dfoxfn, dioxane, dyes,
pigments. Inks,
paints, nitrobenzene
-------
CLEANUP 119
Analytical PTC
The analysis of contaminated sods and aqueous
leaehates involved solvent extraction, liquid chromato-
graphy (fractionation into aliphatic, aromatic, and polar
fractions), and instrumental analysis by gas ehromato-
graphy (GO using flame ionization detectors (FID) and
electron capture detectors (ECO), and high performance
liquid chromatography (HPLO.
For leaehates in which aromatic hydrocarbons or
PCBs were present, EPA Method 608 was followed. The
Murban hydrocarbon contaminant extracts were ana-
lyzed by FID-GC.
The extract from the PCS contaminant leachate
was analyzed by ECD-GC without silica gel frectiona-
tion.
For the leachate containing chlorinated phenols,
EPA Method 825 was used. The leacnate was subjected
to the acid/phenol extraction step only, and then ana-
lyzed by HPLC.
Soil samples were prepared for pollutant analysis
usinf a rigorous shaker table extraction procedure that
has been shown to yield results comparable to Soxhlet
extraction*
LABORATORY KXPKRIMKNTATIOM
The laboratory experimentation was conducted in
two phases.' The first phase involved ihaker table
agitation (equilibration) to determine the sod/aqueous
surfactant partitioning of the pollutants. This compared
the maximum cleanup efficiency under conditions of
water washes and 4 percent aqueous surfactant washes
with thorough agitation. The combination of 2 percent
each of Hyonie NP90 (Diamond Shamrock) and Adsee
799 (Witco Chemical) in water was used.
After the surfactant efficiency was determined
In the shaker table tests, the sod column studies were
performed to evaluate soO cleanup efficiency under
gravity flow conditions. In the column studies, different
concentrations of the three pollutant groups were
used. The concentration of Murban hydrocarbons was
1,000 ppm in the sou; the concentration of PCBs was
100 ppm; and the concentration of chlorinated phenols
> 30 ppm.
Shaker Table Studies
Table 3 presents the experimental design for
shaker table agitation/equilibration of contaminated
for shaker table acitatton/equflibratloB
of contaminated sou*
I. «S*MtaM
fl. *—flnUem
[• «• amlyttMl Mthrfcy, L«, UM •• to MOT* f« tta ««t tu*
-------
120 1984 HAZARDOUS MATERIAL SPILLS CONFERENCE
soils with aqueous surfactant. Nine contaminated soil
samples plus three controls (not shown in the table)
were prepared. All 12 samples (80 to 100 g each) were
placed in 500 ml Teflon* screw cap Jan and subjected
to « water wash (200 ml), with the leachate from the
first and second samples analyzed after the first water
wash and the soil from the first sample sacrificed for
analysis after the first water wash. The remaining 11
samples then were subjected to a second water wash
(200 ml), with the sofl from the second jar sacrificed for
analysis, as weQ as the leachate from that sample and
the leachate from the third jar analyzed, etc.
In this manner, as illustrated In Table 3, son and
leachate were analyzed after each of the three initial
water equilibrations; after each of three water-plus-
surfactant equilibrations (200 ml each); and after each
of three final water-rinst equilibrations (200 ml each).
Eseh soil/solution mixture was agitated vigorously for a
half an hour and then centrifuged to separate the aque-
ous and soQ phases.
Soa Column Studies
The columns used in this study were glass, 7.6 cm
(3 In) inner diameter by 1S2 cm (5 ft). Both ends of the
column were sealed with nippled glass caps. A Teflon*
O-ring placed between the glass column and cap sealed
the two surfaces as they were clamped together by an
adjustable stainless steel jacket. Teflon* tubes con-
nected to the caps allowed the Introduction of the
aqueous solution and the collection of the leachate.
The test soil was prepared by spreading a uniform
layer in aluminum pans to • depth of about 1.3 cm (0.5
in) and treated with • fine aerosol spray of the contam-
inant mixture dissolved In methylene chloride. The
methylene chloride was allowed to evaporate, after
which the sod was mixed by stirring with a stainless
steel spatula.
Contaminated son was packed into the columns
using the following procedures A plug of glass wool was
pushed to the bottom of the column. About 77S g (1.7
Ibs) of soQ then was added to the column and packed to
a height of 10.2 cm (4 in) using a eontroUed-drop ham-
mer compactor designed to fit inside the column.
Following compaction of each lift, the soil was tested
with • pocket penetrometer to determine the compac-
tion. The SOU was packed to a total height of 0.92 m (3
ft) and compacted to a density of IM to 1.76 gm/cm9
UOStollOlbs/ft3).
A falling head permeability test, using modified
American Society for Testing and Materials methods
because of column design, was performed on one of the
control columns before starting the column tests. Head
level fall from the Initial starting point was measured
over time while maintaining a constant head level at the
outflow. PermeabQities (K) ware calculated from the
following standard equation *
K • (2.3 L/t)
Where* L • length of sod sample (em)
t * elapsed time (s)
hg * original head level (cm)
Permeabilities measured in this manner ranged
from 1.1 K 10"3 to 9.0 x 10"4 em/s (3.6 K 10'5 to 3.0 x
10'5 ft/s).
Figure 1 presents an overview of the column
setup during an experiment. Water or aqueous surfac-
tant was gravity fed under a constant 61 cm (2 ft) head
pressure to the top of each column via Teflon* tubing
from reservoir carboys and collected below after pass-
Ing through the column. Leachate was collected and
analyzed for pore volumes 1 through 3, pore volumes 4
through 7 and pore volumes 8 through 10 for each
treatment.
Figure 1. Overview of sofl eakmns te e »»».——. •—
designed te support weter and surfactant carboys at a
constant height above eosuans Surfactant and water
were Introduced to each eokusa through the Teflon*
tubing IB each 2t Utar (S J gal) glaes earboy above the
rack, leaehate efaiting from each eontamlnated sod
eohima wee oolleetad In the glaea carboys shown In the
-------
CLEANUP 121
RESULTS
Murban-Contaminated SoQ
The quantitative data are illustrated in Figure X
as a bar graph of total hydrocarbons present in nil and
leachate after each step in the shaker table experi-
ments. The data in the graph Illustrate that very little
cleanup of the sod occurred during the first three water
washes, but a significant reduction (down to 41 percent
of original levels) was obtained after the Initial aqueous
surfactant wash. Continued improvement in hydrocar-
bon removal was observed in the second and third equi-
libration with aqueous surfactant. Gradual removal of
surfactant and residual hydrocarbons then was observed
in the three final water washes. In general, overall
mass balance approaching 95 percent was obtained in
the shaker table experiment.
In the sod column studies with Murban, very
limited removal of aliphatic hydrocarbons from the sod
occurred even after 10 pore volumes of initial water.
After three pore volumes of aqueous surfactant, how-
ever, the soQ material was significantly cleaner, and
after the final 10 pore volumes of water rinse, the son
was effectively decontaminated.
Significant levels of aromatic hydrocarbons were
present in the soil after 10 initial water washes. After
the first three aqueous surfactant washes, however, the
aromatic components were completely removed, and an
that remained in the son were components from the
surfactant material itself.
Soft column and leachate data are shown In
Figure 3, which presents the relative contaminant
concentration in the sod and leachate after each water-
Surfactant treatment. The distillation fraction >
tration ringed from 80 to 100 percent during the first
three water washes and then drooped to about 10 per-
cent during the surfactant treatment, with 70 to 80
percent of the original hydrocarbons observed in the
aqueous surfactant leachate. The final three water
washes did not effect any additional cleanup of the soil,
and the average residual soil concentration was about 7
percent of the initial spiked distillation fraction concen-
tration.
PCB Pollutant Mixture
The Initial PCB shaker table experiments follow-
ed the same protocol as that described for the Murban
tiyorocsrooBswo
Figure 4 illustrates PCB cleanup from the shaker
table experiments. After the first surfactant wash, the
sod PCB concentration decreased to about 25 percent of
the original level, with 45 percent of the original PCBs
accounted for in the aqueous surfactant "at that point.
Additional surfactant rinses produced. even greater
cleanup of sett PCBs, with up to 6! percent of the
original PCB material present In the surfactant after
the third aqueous surfactant equilibration. An overall
removal of about 67 percent of the. original PCBs after
the three final water rinses was finally obtained.
As In the shaker table studies, very little cleanup
of the sod column was effected with the water washes,
while significant removal of PCBs was observed after
pore volumes 1-3 of aqueous surfactant. The data are
Illustrated In Figure 5, which shows the overall concen-
trations of PCBs In the son and leachate after each
successive treatment. The effect of the first aqueous
surfactant wash from the soil column was a 90 percent
reduction la PCB concentration in the sod column.
i
^
§
|
La
•w
I
§
$
^
\»
•4
|
§
^
v////////////////////////////////////.
•u
'
<
1
u
u
•u
CM
fU
••
•u
W
mi
•4
•i
m»
-------
122 198* HAZARDOUS MATERIAL SPILLS CONFERENCE
Itt
s It
1
X M
3 »f
5
5
I >•
P
1
^J= SOU
n -
•i • LJ - L—eh«l« „_,
•
1
-
•.1
1
.Ml
|
!».•
*.M
• I.I
n.
•••
I.H
1.
i
1-1 t » • 1* 11 • T • It 11
•
1
••M
I
• »
•bna .
•IUM«I
>i
^
t-W
-i
"•• n _
S .„, Ea LJ - «—*•«•
ftttallvt ConlMitntnt Concentration 1
s t t t i
1
•II
M.I
1
t
.1
•.*
•M
II. 1
+.•
1
H.l
M.
••
1 1
1
HI ftitfM M kltf V« «*« 1
M.
••
•
1.11
n« lr
,
M.I
•M
1
«
-•«
-------
CLEANUP 123
= Soil
= Laactwte
• •M
W.t M.l
». •»
t.n.
• r s-te vi
Surfactant •••h
••M
Figure*. PC8
*H-
tram mA
During the three final water rinses, the overall PCB
concentrations were reduced to less than 1 percent of
the Initial value. An overall mass balance of about 70
percent was obtained.
Chlorinated Phenols Pollutant Mixture
The overall cleanup of phenols In the soO to
illustrated in Figure ft. h is clear that before initial
treatment, about 93 percent of the added dl-, tri-, and
pentachlorophenol mixture was associated with the
soli. After the three water washes, however, the resi-
dual contamination of the chlorinated phenol group in
the aoQ had dropped to 1 percent of the amount origin-
ally present. Pore volumes 4-7 and t-10 increased the
final proportion of chlorinated phenol In the leachate to
about 70 percent of the amount added to the son origin-
ally, and the residual chlorinated phenol in the son
dropped to about 9A percent oMho value originally
introduced.
COMCLOSlOm/UCOMMXMDATiONS
The shaker table studies and the sod column
studies showed that the 4 percent aqueous solution of
surfactants was extremely effective In removing hydro-
phobic and slightly hydrophOic organlcs from the test
solL The performance of the aqueous surfactants in
removing PCBs from the son was quite similar to their
performance with the Murban distillation fraction.
However, water alone was sufficient to decontaminate
the chlorophenols-contamlnated soQ.
A small amount of aqueous surfactants sotubQbc-
ed substantial amounts of two llpophltte contaminant
mixtures from the) teat sofl. Although the surfactants
were chosen for this function, the relative ease of
removal of the contaminants from the son Is partly
because* of the soffs characteristics. The TOC of the
Freehold soQ used la the laboratory tests was 0.11
percent! this Is somewhat low, and values of 04 to 1.0
percent might be eipeoted for a son mixture of A, B,
and C horizons. At higher TOC values, organies would
be removedTrom the son less readily.
The results of the sott column tests with Murban
and PCBs paralleled the shaker table test results.
Because of their hydrophobia nature, little of the con-
taminants was removed by the initial water washes,
while the aqueous surfactants removed them from the
soQ quite efficiently. The aqueous surfactant appeared
to be somewhat more effective in the column tests than
in the shaker table tests.
-------
124 1984 HAZARDOUS MATERIAL SPILLS CONFERENCE
100
r «.
r *
0
| -
1 "
•
^M4
u.
^f
'/
s
's'
X
'/
X
X
X
' ^
J>
2
i
0 Soil
M.f
V r?
.*!
1
.M
M.I
.1
M.I
—
•t.l
-------
1984 HAZARDOUS MATERIAL SPILLS
CONFERENCE PROCEEDINGS
Prevention, Behavior, Control and
Cleanup of Spills and Waste Sites
April 9-12,1984
Nashville, Tennessee
err
Sponsored by:
ASSOCIATION OF AMERICAN RAILROADS/BUREAU OF EXPLOSIVES
CHEMICAL MANUFACTURERS ASSOCIATION
UNITED STATES COAST GUARD
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
National Agricultural Chemical
Association
National Association of Chemical
Distributors
National Association of Counties
National Audubon Society
National Solid Wastes
Management Association
National Tank Truck Carriers
Railway Progress Institute
Seaboard System Railroad
Spill Control Association of
America
Tennessee Emergency
Management Agency
Tennessee Manufacturers
Association
U.S. Army Corps of tnRinrcrt
Member*.-
American Chemical Society
American Institute of Chemical
Engineer*
American Petroleum Institute
American Society of Civil
Engineers
American Waterways Operators
Association of State and
Territorial Solid Waste
Management Officials
Federal Emergency Management
Agency
Hazardous Materials Advisory
Council
International Association of Fire
Chiefs
Nashville Department of Civil
Defense
-------
Published in: Land Disposal, Remedial Action, Incineration and Treatment
of Hazardous Waste: Proceedings of the Twelfth Annual Kc:u
Sympor, Uim aL Cine i nnat L , Ohio, April 2\-2'\, 1<)H(>.
-H(>/():r> AuiMtst lfJHf>. pi'. ^OM-1'. I/.
FIELD EVALUATION OF IN SITU WASHING
OF CONTAMINATED SOILS WITH "HATER/SURFACTANTSl
James Nash
Mason & Hanger-Silas Mason Co., Inc.
P.O. Box 117
Leonardo, New Jersey 07737
Richard P. Traver
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, New Jersey 08837
ABSTRACT
Since 1981e the Releases Control Branch of the Hazardous Waste Engineering Research
Laboratory has been developing techniques to wash contaminated soil 1n place (in situ).
The project includes: design and fabrication of the hardware to carry out the washing,
evaluation of surfactants to do the washing, determination of which geological character*
istics to use to judge the appropriateness of in situ washing, development of a monitoring
and reporting system, evaluation of two candidate sites for the field testing of the hard-
ware, and a pilot treatment study at a contaminated site.
* • ,
This paper summarizes the design and development of the j£ Situ Containment/Treat-
ment Unit (ISCTU) and the evaluation of surfactants for in situ soil"washing. The empha-
sis is on work completed at Volk Air National Guard Base. Camp Douglas, Wisconsin. The
work shows that surfactants will remove otherwise obstinate contaminants from soil even
without mechanical agitation of* the soil. However, subsequent treatments of the surfactant
laden leachate is an unresolved problem.
INTRODUCTION
In situ soil washing is the term to
describe washing of contaminated soil with*
This report Is a summary of work per-
formed 1n partial fulfillment of Con-
tract Numbers 68-03-3113 and 68-03-3203
under sponsorship of the U.S. Environ-
mental Protection Agency. The U.S. Air
Force through Interagency Agreement
IRW 57931283-01-0 with the U.S. EPA
has also sponsored much of the'work
reported here. This paper discusses
out excavating. The washing is accomplished
by applying a liquid at or near the surface
the key activities of four projects:
"Treatment of Contaminated Soils With
Aqueous Surfactants". "Retrofit of the
J_n Situ Containment and Treatment Unit",
Chemical Counter-measure Application at
Volk Field Site of Opportunity", and
"Site Characterization and Treatment
Studies of Soil and Groundwater at Volk
Field."
-------
The first washing tests were run on
a shaker table and the next test series 1n
columns. Contaminated soil Mas compacted
1n 3 1n. Increments Into 3 1n. diameter,
5 ft high glass tubes. The tubes were
fitted with nlppled glass caps at the-
bottom and top. A pressure head of 30 cm
of surfactant solution was applied to the
surface of the contaminated soil. The
soil pores were, therefore, experiencing
saturated flow of the surfactant solution.2
The soil used for the laboratory work
was a Freehold series typlc hapludult from
Clarksburg. New Jersey. It was selected
because of Its grain size distribution and
similarity to soil at CERCLA candidate
sites 1n EPA's Region II. Ten percent was
silt OP clay, eight percent gravel and 80%
coarse-to-flne sand. Its permeability of
10-* cm/sec 1s at the low end for 1n situ
washing. Nine to eleven percent oT~tne~
soil was HC1 soluble. Of the crystalline
structure, 98% was quartz and 2% was
feldspar. Only 0.12% was organic carbon
which 1s a low value and accounts. In
part, for a low cation exchange capacity.
A topped Murban crude oil 1n methyl-
ene chloride was applied to the soil.
This contaminant was selected because 1t
contained many organic types Including
aromatlcs, polynuclear aromatlcs, allpha-
tlcs, polar and non polar compounds. The
methylene chloride was allowed to evaporate
and the soil was aged prior to being
loaded Into the test columns. 'Other con-
taminants, 1n separate tests, were chloro-
phenols and a polychlorlnated blphenyl.
Gas chromatographlc analysis showed
that ten pore volumes of surfactant solu-
tions passed through the columns removed
88% of the topped Murban crude oil and 90%
of the PCB's. Using high performance
liquid chromatograQhy (HPLC). It was shown
that chlorophenols were removed with the
water alone. Surprisingly, removal 1n the
column studies, where there Is a low level
of mechanical washing, was better than re-
moval In the shaker table studies'. Start-
Ing at 1000 ppm contamination 1n the
columns, removal efficiencies as high as
98% were reported.
Control of In S1tu Washing Fluids
Accelerating the natural tendency of
a contaminant to migrate through the vadose
zone Into the groundwater 1s the basic
purpose of |n situ soil washing. In order
to do this so there 1s no adverse Impact
on an aquifer, rigid controls must be
maintained to assure the contaminant 1s
captured. The EPA's In Situ Containment
and Treatment Unit (fsTtOFwas designed
for this purpose. The drawing 1n Figure 2
represents the parameters (of an hydraulic
budget) that were considered for the
(ISCTU).3 They are: recharge Ga, discharge
Da, treatment system flow R, evapotrans-
p1ration E, precipitation P, natural
groundwater flow U}, and Induced ground-
water flow U?. Variation 1n these
qualities will change those items 1n lower
case letters; vadose zone thickness w.
mounding m, drawdown (he-hw), and radius
of Influence re (not to be confused with
the radius of capture).
* *^-T »T~ "-•»-•*.»»—-™ -.-»--• -^-^. - - -.
Figure 2. In situ parameters
-------
Figure 3 1s a simplified drawing of the
rsCTU, which Is equipped with recovery and
delivery pumps, batch mixing and propor-
tional-additive metering pumps, flow rate
controls, pressure and flow meters, and a
volatile organic stripping tower. Any
treatment of groundwater requiring more
than air stripping must be done "off-
board." A microcomputer/data logger 1s
used to monitor environmental conditions
and the effect of pumping and recharge on
the aquifer. To do this, depth gauges,
flow meters, moisture meters, and a weather
station are connected to the data logger.
A. AIM DIAPHRAGM PUMPS
I. PROPORTIONAL CHEMICAL
AOOITIVC MCTIRINQ PUMf
C. INPUT MANIFOLD
MAIN ELECTRICAL
BREAKERS
0. PROCESS MONITOR RECORDER
E. WATER PUMP
P. IATCH CHEMICAL METERING PUMP
CHEMICAL MIXING TANK
VAPOR EXTRACTION
SYSTEM
I * I r Tl ELECTRICAL 1
\ Fi 1 I I I CONTROL I
PULLOUT OPf RATORt PLATFORM
DIESEL ELECTRICAL
GENERATOR
INJECTION MANIFOLD
Figure 3. In Situ Containment and Treatment Unit
Site Selection for the Field Evaluation
In September 1984 the U.S. Air Force
and the U.S. EPA started 1n a joint effort
to evaluate 1n situ washing technology.
The primary "oBjectlve of the project was
to demonstrate full-scale feasibility.
A secondary objective was to develop a
more comprehensive strategy for^the decon-
tamination of fire-training areas of all
A1r Force and Department of Defense (DoD)
Installations. The following criteria
were used in selecting a site suitable
for full-scale soils washing research..
A site of less than one acre was desired
.to reduce soil variability and reduce
'sampling costs. Because soil washing Is
best suited for permeable soils, a sandy
site was* sought. Contaminants at the
site were to be common organic chemicals
found at many other Air Force sites.
I.e., trichloroethane, benzene, toluene,
tMchloroethylene. Officials of the
selected installation and responsible
environmental agencies would need to be
cooperative.
Preliminary screening of candidate
sites was accomplished through a review
of Air Force Installation Restoration Pro-
gram (IRP) reports. Over sixty reports and
nearly 800 sites were screened. During
the review, it was apparent that most sites
with organic chemical contamination fell.
Into two common categories: sites of fuel
spills and fire training areas.
F1re training areas were especially
suited to this research because of their
limited size and range of contaminants,
which included chlorinated solvents, fuel
components and lubricating oil. Fire
training areas are found at almost all Air
Force Installations and, because of the
long-term fuel and solvent dumping at these
sites, they have significant off-site pollu-
tion potential.
Following this careful review, a fire
training area at Yolk Field. Air National
Guard Base, Wisconsin, was selected as a
research site. Historical records indi-
cate that the Volk fire training area may
-------
have been established as early as World
War II and has routinely received waste
solvents, lubricating oil, and JP-4 jet
fuel. Although It 1s Impossible to deter-
mine the quantity.of chemicals that soaked
Into the ground versus the amount volatil-
ized and burned 1n fire training exercises,
one estimate 1s 52,000 gallons. Measure-
ment of volatile organlcs from groundwater
samples taken 1n 1980 directly below the
fire pit showed chloroform, trlchlorethane,
trlchlorethylene, benzene, toluene, and
ethyl benzene totaling above 50 mg/Uter.4
Site Studies
Two site studies were made at the fire
pit area during 1985. These studies were
conducted to thoroughly understand the
hydrology and chemistry associated with
the contamination have produced as a
by-product a great deal of data and In-
sight Into a chronic oil spill. Initially,
the character of the contamination was
.misunderstood. The original concept of
a floating layer of oil that could be
handled easily gave way to the realization
that the contamination had not remained as
a water Insoluble oil but had been trans-
formed to soluble organlcs by biological/
chemical activity. Biological activity
had been nourished by the flreflghtlng
foams used 1n the training exercises.
These f1re-f1ght1ng foams may have also
contributed directly to solub1Uz1ng the
oils. The groundwater, 25 ft below the
surface (and only 60 ft from the
pit), had up to 50 mg/llter total organic
carbon (TOC). Infrared spectrophotometrlc
(IR) scans Indicated this contamination
was 1n part esters or organic acids.
Upon emerging from the centrifugal pump
(used for a pumping test), the groundwater
frothed.5 Directly below the pit the
water table was at 12 ft. The hydraulic
conductivity was 5 x 10~2 cm/sec.
Treatment Studies of the Soil
The overall soil contamination had
the physical consistency of a medium
weight lube oil. At a one-foot depth
average oil and grease (determined by
carbon tetrachlorlde CCC143 extraction)
was 13,500 mg/kg. Deeper Into the soil.
oil and grease (O&G) values decreased.
At 5 ft, and continuing to the capillary
zone at 10 ft, O&G values were 400-800
mg/kg. Soil samples from the aquifer
taken at 15 ft produced 5000 mg/kg O&G.
The chemical composition of the CC1.4
extract also varied with depth. IR scans
of extracts of soil from 1 ft depth
match scans of parafflnlc oil. Esters or
acids of oil become more evident when
approaching the water table. Below the
water table, the oxidized oils although
present, are less prominent. This profile
is apparently a symptom of weathering.
The more soluble oxide forms have been
carried to the groundwater by percolating
rain water.
The volatile contaminants also show
evidence of weathering. In contrast to
O&G. the weathered volatlles are found
closer to the surface than to the water
table and are an order of magnitude less
abundant than O&G extracts. A relatively
high abundance of isoprenoid compounds
(includes many naturally occurring mater-
ials such as terpenes) In relation to
normal alkanes also Indicates long term/
mlcroblal degradation.6 A terpene-like
odor was noticed while taking soil samples
to determine the lateral extent of contam-
ination near the surface. Within 6
In. of approaching the clean soil and
at depths of 6 to 12 in. a
"mlnty-turpentlne" smell was reported by
the field technician.
A part of the fire training area was
prepared so that ten mini soil washings
could be conducted simultaneously. The
first foot of soil was not to be Included.
Therefore, ten 1 ft deep holes were dug
and the bottom of each hole was called
the "surface" of the test chamber. Each
"chamber" was a 14-1n. depth of soil from
the bottom of the hole down. Surfactants
tested were: an anlonlc sulfonated alkyl
ester (Pit 17), a polyethylene glycol dio-
leate (Pit 110), ethoxylated alkyl phenol/
ethoxylated -fatty acid blend (Pit 18). and
the contaminated groundwater (Pits 12,3,4.
5,9). The dloleate caused soil plugging
immediately. Compared to water, penetra-
tion rates were reduced when any surfac-
tant solutions were used. The groundwater.
-------
which has a low concentration of biologi-
cally produced surfactant, had the least
effect on the penetration rate.
The dominant contamination 1n the
soil was oil and grease, up to 16,000
mg/kg, where volatlles were less than 100
mg/kg.6 OAG measurements were therefore
used to determine the effectiveness of the
soil washing. To avoid channeling during
the pilot treatment, prewash 046 measure- .
merits were made on samples taken adjacent
to the chambers. Statistically, the O&G
measurements had a coefficient of varia-
tion (CV) throughout the test area of 35%
making 1t difficult to draw conclusions
of soil washing effectiveness. Figure 4
shows the O&G measurements after the sur-
factant wash process and the blank value.
Pit 18 was washed with the lab-developed
50/50 surfactant blend. It Is Interesting
to note that the O&G at 12-14 1n. has
Increased 24% above the blank and the
surface top layer O&G has decreased 50%.
Implying a transport of contaminant down-
ward during the seven days of washing with
14 pore volumes. Keep In mind a CV of 35%
precludes any definitive conclusion. The
expected reduction of contamination at the
12 1n. depth to 50% of the original level
was not realized.
Treatment Studies of the Groundwater
Bench scale and then pilot treatment
studies of the already contaminated ground-
water were undertaken In anticipation of
full-scale soil washing. Bench-scale
studies evaluated addition of: lime,
hydrogen peroxide, alum, ferric chloride,
,ind various water treating polymers. The
pilot treatment was run using the EPA's
Mobile Independent Chemical/Physical Treat-
ment Unit, a holding lagoon, and an air
stripper made by the A1r Force. Figure 5
1s a process flow diagram that also Indi-
cates sampling points. The three treat-
ments consistently used during the opera-
tion were lime addition, settling, and
» m
".
F
«,
• .
r
t
t
t
t
t
T
f
i
t
f
t
t
Jl
t\
i \
<\
t\
i
I
t
t
t
i
f
I /
r
t
t
f
f
f
t
m f
s; ?
si i
n
Q
^
^
V
js
\ T:
\?.
ll
\
s
V
^
^
s
s
s
^
\
1
n _n
*« f
S /
y >
\ i
n \ '
ft s >
A \it /
h\Lt
\
\
\
\
\
\
V
s
t
f
1
t
t
t
t
t
f
)
t
t
/
1
i
•
y
^
y
»l
i
V k
• • 1 » I
\
\
\
\ 't
\ 'l
M
N i
\'l
2
it •
Figure 4. Soil washing data
volatilization. Total organic carbon
(TOC). volatile organic analysis (VOA),
and suspended solids (SS) tests were
used to monitor the effect of these
treatments.
Addition of lime brought about signi-
ficant reductions In TOC. Organlcs were
removed with an Iron hydroxide to form a
floe. (Interestingly, the contaminated
groundwater had up to 56 mg/liter Iron
compared to background levels of 0.2 mg/
liter.) Volatiles were 95 to 98% removed
1n the lagoon and air stripper. Figure 6
1s a bar chart depicting the measured
level of TOC at four points In the process.
Figure 7 Is a bar chart showing the mea-
sured levels of four volatlles at three
locations In the process.
-------
miTlltS
III!
•ELL
L3
FUSM
•II
cuimu
H M55
tt
5 i
UCOOI
•in
TO HIOIIC
UCOOK
Figure 5. Well field effluent treatment process and
sampling points
M)
3D
SO
r
O ya
5 *
HO
ID
'/
y
I
y
y
/
/.
v
S
x
s
••
W]
|
s
I*
71
V.
7
t
I
1
S
1
«
1
1
71
1
1
1
IF
|
?
>
I
1
! •
I
I
1
I
\
I
t,
%
%
m
i
^
|
1
I
I
',
33
1
1
m 0U
1234
Figure 6. Four data sets showing Level of
TOG at the well field, clarifier
effluent, stripper feed, and
stripper effluent.
T«A
B£N
TOL
ETH
BEN
Figure 7. Volatiles at the well field,
stripper feed, and stripper
effluent.
-------
In anticipation of conducting a In
situ soil washing of the entire pit,
tests were run to determine control of
"the natural groundwater flow beneath the
pit. This was accomplished by a six-
member well field. In total there have
been 13 wells Installed 1n the study,
7 monitoring wells and 6 withdrawal wells.
Boring logs were kept during the drilling
operations. Split spoon samples of the
sand and weathered sandstone were used for -
chemical analysis and particle size analy-
sis. The fines content of the directly
below the pit 1s significantly lower than
1n the adjacent uncontamlnated soils -
2 to 5% versus 10 to 15%. Fines content
of soil 8 ft below the water table,
slightly down gradient, and In the plume
1s unusually high: 28% versus 10-15%.
The production wells placed In the high-
est contamination zones have the poorest
fluid yield. Paradoxically, according
to equlpotential lines constructed from
water table depths, there 1s a convergence
of flow passing beneath the pit
(see Figure 8).
•02.4S
AIR
STRIPPER
Figure 8. Treatment site showing water
table equipotential lines
This 1s directly 1n line with a pro-
duction well producing water containing
700 mg/ liter TOC at less than 2 gallons
per minute. The average for the rest of
the wells Is 260 mg/Hter at 6 gallons per
minute. The design pumping rate for each
well was 12 gpm. In spite of well yield
problems the natural gradient of 0.001
(ft/ft) was easily reversed to create a
radius of Influence of greater than 100
ft and a radius of capture greater than
the 40 ft training pit radius.
A Follow-up Electromagnetic Survey
An electromagnetic survey was con-
ducted over the ground surface surrounding
the training area to determine the measur-
able extent of the plume. The decision to
do this was based on the low conductivity
of the soil, high conductivity of the
plume (600 mlcromohs), and the low
conductivity of the background water
(20 mlcromohs). A study conducted by
the New Jersey Geological Survey7 had been
-------
able to map an organic plume from a fire
training area In a sandy aquifer. In the
report of that work, the fire fighting foam
AFFF was felt to be the conductive organic
that made the survey possible. In this
work the high Iron content of the plume 1s
considered the reason for the success of
the survey. The reason for the high Iron
content 1s the reducing conditions that
exist(ed) during biological activity at
the site. Figure 9 is a map of the plume -
based on conductivity.
•? a ' 3
The CCl_4 extract of a soil sample taken
at 12 ft at the point marked "S" 1n the
figure was Identified as an oxidized oil.
The authors wish to express their
appreciation for the cooperation, encourage-
ment and help given by a number of people
from the Wisconsin Air National Guard and
Department of Natural Resources. But
especially we wish to acknowledge Doug
Downey of the U.S. A1r Force for his
gentle persistence in directing the work
done at Volk Field.
CONCLUSION
The mechanical aspects of applying
a surfactant to soil and controlling an
underlying unconfined acquifer to capture
the wash solution have been demonstrated at
a site of opportunity. Issues that remain
to be addressed are treatment, if necessary,
of the used surfactant solutions. Isolation
of the containment from the surfactant and
developing a method to recycle the surfac-
tant.
Figure 9. Electromagnetic Survey
-------
REFERENCES
i. Texas Research, Institute, Inc.
Underground Movement of Gasoline on
Groundwater and Enhanced Recovery by
- Surfactants. September 14, 1979
American Petroleum Institute, 2101 L
Street, NW, Washington, DC.
2. Ellis, W. D., J. R. Payne, Treatment
of Contaminated Soils with Aqueous
Surfactants (Interim Report)
September 6, 1985 to EPA-HWERL.
'Releases Control Branch, Edison, NJ.
3. Waller, H. J., R. Singh, J. A. Bloom.
Retrofit of a Chemical Delivery Unit
for In Situ Waste Clean-up, EarthTech,
Inc. January 7, 1983.
Releases Control Branch, Edison, NJ.
4. Hazardous Materials Technical Center
Installation Restoration Program
Records Search prepared for 8204th
Field Training Site, August 1984
available N.T.I.S.
5. Nash, J. H.. Pilot Scale Soils Washing
and Treatment' at Volk Field ANG, Camp
Douglas WI, In preparation.
6. McNabb, 6. D.. et. al. Chemical
Countermeasures Application at Yolk
Field Site of Opportunity^
September 19. 1985 to EPA-HWERL.
Releases Control Branch, Edison, NJ.
7. Andres, K. G. and R. Crance, Use of the
Electrical Resistivity Technique to
Del 1neate a Hydrocarbon Sp111 In the
Coastal Plain Deposits of New Jersey.
Proceedings: Petroleum Hydrocarbons
and Organic Chemicals 1n Ground Water,
November 5-7, 1984 available National
Water Well Association. Dublin, OH.
-------
In Place Detoxification of
Hazardous Materials Spills
in Soil
INTRODUCTION
Spill incidents can occur in almost any known geographic
area, contaminating air, water and/or soil. Containment and
treatment technology for water spills has received the most
attention and is the furthest advanced. However, in many
instances, both water and soil are contaminated when land
spill threatens a nearby water body or the groundwater
table! The state-of-the-art of land spill cleanup has consisted
mainly of excavation and hauling to an approved landfill
site or possibly flushing of the affected area with water.
These methods are appropriate in certain circumstances.
However, when the groundwater is threatened, when a large
soil mass is contaminated or when no suitable disposal site
is available, other approaches may be needed.
It is the purpose of this effort, funded by the U.S. En-
vironmental Protection Agency under contract number
68-03-2508. to develop t mobile treatment system which
allows in place (in-situ) detoxification of hazardous mate*
rials spilled on soil. Detoxification in this context refers to
amelioration of a spill's effect by chemical reaction. The
project goals were to design and demonstrate a mobile
vehicle capable of encapsulating a 10.000 gallon land spill
in grout and treating the spilled chemicals in place by either
oxidation/reduction, neutralization, precipitation or poly-
merization. The approach to achieving the design goals was
to use direct injection of grouting material into the soil
around the contaminated area to envelop the spill and isolate
it from the groundwater. followed by detoxification by
injection of treatment agents. This paper documents the
results of the laboratory and pilot tests and the resulting
preliminary system design. The vehicle which will be fabri-
cated 4nd demonstrated during 1978 should be a part of
the EPA spill response arsenal by 1979.
Project Approach
The work was divided into five phases: I) Laboratory
Study. 2) Pilot Testing ana Design. 3) Fabrication.4) Test-
ing and Demonstration and 5) Reports. The information
obtained during the laboratory and pilot tests was used to
develop the final system design and. as anticipated, the end-
product design was modified from that origwally envisioned.
Laboratory Testing
The laboratory tests had two main objectives:
Kathryn R. Huibrcgtse
Envirex Inc.
Milwaukee,.Wisconsin
and
Joseph P. Lafornara
U.S. Environmental Protection Agency
Edison, New Jersey
and
Kenneth H. Kastman
Soil Testing Services Inc.
Northbrook, Illinois
1. To determine if fo-sfru treatment techniques could
effectively detoxify chemicals present in various soil
systems and,
2. To evaluate, choose and test, various grout types for
their potential use in spul containment.
Choice of Cbemicab and So3s
Various reagents and soil types were chosen for testing
the four types of chemical reactions: oxidation/reduction.
neutralization, precipitation and polymerization. Chemical
compounds studied as contaminants were chosen based on
the following criteria: 1) efficiency of the chemical reaction*,
2) common use of the chemical and 3) potential risk of
spillage. Treatment agent choices were based on: 1) the haz-
ardous nature of the treatment chemical. 2) its availability.
3) its handling difficulties and 4) the volume needed for
detoxification of the contaminant. Contaminant concentra-
tions were established by common shipment concentrations.
and the strength of the reactant was established to keep the
detoxification controllable. The chemical systems are
shown in Table L
Four soil types were also included in the laboratory
study. It was determined that classification of sofls by grain
size would be most advantageous, since this characteristic
often controls the sofls permeability and therefore its amen-
ability to injection of treatment agents. The four soil types
considered were day, silt, sand and gravel. In order to
simplify data interpretation, it was decided to select soils to
minimize the amount of interaction of the soils with the
chemical systems. This was justified because the objective
of the laboratory study was to evaluate the effects of a soil's
physical properties on in-ittu detoxification and it was
thought that the potential interferences from soil chemical
properties could be to mask important physical effects
which needed to be defined. Therefore, the following rela-
tively inert soi types were chosen: day-Georgia Kaolin;
silt-No. 290 Silica Hour; sand-blended Ottawa Silica Sand
(Flint shot and No. 1 Federal Fine); (ravel-trap rock. The
soil gradations were selected to be representative of the
specific soil type to be tested. For example, the amount of
clay or silt in the sand sample was minimal.
362
-------
Table I: Chemical Reaction System* Inve*tif tied
React ion Type
Oxidation/
RuUuct ion
Ueutral izat ion
F'recip! tat icn
Polyr.cr izat Ion
Ccntiininant
Compound Concentration
Sodium Hypo- 12-15* Cl
chlorite
Su If uric Acid 3f.U
Copper Sulfate 75 g/ I
Styrene 1005
Reactant
Compound Concentration
Sodiun Bisulfite*
Sod Turn Hydroxide
Sodiui* Sulf ide/
Sodiuu tiydroxirle
Persul fate
7.*
1-5N
1.0
O.I
Laboratory Reaction Feasibility Testing
The laboratory testing was subdivided into three parts:
design and fabrication of the testing apparatus and develop-
ment of the procedures: the actual performance of the tests
and evaluation of the results. Two types of testing were
performed: flow through tests in which drainage of the
system was allowed during the reaction and sealed tests
which involved direct addition of reactant to the soil with
no drainage of the soil allowed.
In order to evaluate as many of the critical variables as
possible, an experimental design was established. This
approach varied soil conditions (bulk density and water
content), contaminant loadings (as percent of the soil void
space available) and detention time (pressure). The sofl and
chemical systems were to be evaluated separately. After
initial attempts and problems involved with developing a
safe, uniform and generally applicable approach to poly-
merization in the soil, this reaction was not further evaluat-
ed. Therefore testing was limited to three reactions and
four sofl types.
The laboratory testing apparatus consisted of a 33 in.
diameter dear column supported by machined aluminum
bottom and top fittings (See Figure 1). The column was
fiUed with an underdrain support system for the flow-
through tests and a porous plate/screen cover to distribute
the chemicals placed into the column. When necessary,
regulated air pressure was used to force the reactant through
the contaminated soil. The sealed apparatus required elimi-
nation of the base and drainage holes. Columns of both
acrylic and dear FVS plastics were used since neither ma-
terial was resistant to aQ of the chemical species tested.
The testing procedure involved mixing specified amounts
of soil and water and packing this mixture to incremental
heights to achieve a specified sofl bulk density. These sofl
columns were then contaminated with liquid to fill a
certain soil void volume, the treatment agent added and
samples collected at the underdrain. If sealed tests were
performed on a system the oontaminant/reactant/soil
mixture was allowed to stand for a given time and soil core
samples were taken and analyzed.
»c*T
CTT*T
VMS xu"
ttutn
u**vtn
Fiaura I: Laboratory Testing Apparatus
Initially, flow-through testing only was to be imple-
mented. Howevej. it toon became apparent that this
approach was not feasible for the fine grained days. The
-------
364 In Place Detoxification
high pressures required to force the reactant through the
soil caused short circuiting along the column sides and no
detoxification occurred. Therefore it was decided to test t
surface treatment method (sealed tern) for the day systems.
The data collected from all laboratory testing were
evaluated and the percent of contaminant treated was cal-
culated along with the residual concentration in the treated
soil. Statistical analyses of these results using ANOVA
design and F tests were used to identify which of the vari-
ables had significant effects on the efficiency of the reac-
tion. The results indicate that both soil type and reaction
type significantly affect the degree of detoxification, along
with the three internal variables (soil conditions, detention
time and loading).
The efficiency of in-titu treatment in gravel was much
lower than with other soils (See Table II). This is a result of
most of the contaminant rapidly percolating through the
gravel prior to treatment. However, for the contaminants
entrained on the gravel, the reaction efficiency ranged from
95-99%. The overall efficiency of the neutralization reac-
tions was also lower since a pre-reactant water rinse was
required in order to reduce the heat of reaction. Precipita-
tion reactions were more efficient than anticipated. This
may be due to the blocking affect of the precipitate which
clogs some of the voids and forces the treatment agent to
flow into other contaminated areas. Redox reactions were
generally quite efficient under all conditions. The detention
time was critical for sand detoxification indicating that too
high a pumping rate will be detrimental in final treatment.
The effectiveness of sealed detoxification (surface treat-
ment) was not anticipated. As long as void saturation was
not exceeded, the treatment agent entered the fine grained
soils and mixed to a degree which detoxified most of the
contaminant. This apparent mixing in the small void sizes
was not expected. Reduced reaction efficiencies were ap-
parent for precipitation because the precipitate did block
the reactant's path into the soil. Overall, even this reaction
was quite effective. The main problem with a sealed system
is that the volumes which can be treated are limited to voids
available for the reactant.
Grout Evaluation
The second objective of the laboratory testing was to
evaluate the grout which could be used for encapsulation of
a spul. The main types of grout available include paniculate
grouts such as cement and bentonite and chemical grouts
which are mainly AcryUmide (AM-9). urea-formaldehyde
resin, lignin or silicate based materials. Partkulate grouts
are generally used in coarse grained sods since they have a
relatively high viscosity due to their suspended particles in a
Table II: Summary of Laboratory Tect Rcnlta
Soi 1 rs*£j '£2, J*1.' Jl2£.
Ti«vi *cH "low Thru
'•ird "cdor. rlo-./ Ti»ru
:«.-J r?T Ha* Thru
*flcj .'•/« rlo.i Thru
S«lt -*cij rio"* T*»ru
i«!t -*io* '^lo.' Tliru
lilt "FT ric.» Tnru
Sill .'"/•: • -lc..: Thru
Travel "cid FIc* TJ>ru
rravel "edoji rlo-r Thru
••ravel P?T r!o-< Thru
travel '-vo rlo« Thru
Clay -'cid Sealed
•
Clav 'COOK Sealed
r:av rrj Sealed
Clav -Vc Scaled
^r.*,«4 on the (I0t*' -m8"
Pantje of
Effectiveness*
3-V-52.2
1S.«.-6C.*
n. 2-85. 3
( 3.4-35.8)
3*.3-?5.9
12 -**^.3
(12 -5?.«)
3.7- 5-3
12.5-26.4
I?.S-31.3
3.7-31.3
74.6-78
96. 2-2?. 5
$6.5-87
'(56. 5-9?. 5)
mt of containment-amount
Averane
Efficiency*
22.4
37.3
42. «J
(34.2)
57.7
55.7
7'.. 4
C3.6
4.8
20.0
20.1
(15)
76.3
98
74.6 •»
82.3
not reacted^
total containment '
Significant '
Variables
Detention Time
Detention Tine
Detention Time
Soil Conditions
Loading C Soil
Conditions
Loading
Monc
Load in*
(tone
riooe
None*
Ikme
X 100
number
of Tests
12 11
12 '•
12
12
12
8
k
k
4
k
k
k
-------
water oue. i_nerrucu &iuuit &ic &cuci*u
and can be uied to pout finer grained toils. One of the
most commonly used chemical grouts is AM-9 which can
acryUmide bue which U toxic to groundwaters. Therefore
it was not considered suitable for the spill containment
application.
0.160
0.1*40
0.120
UJ
3!
ca
£ 0.100
^ 4
i «
o 0. 0300
4/1 i
<
o
C3
$ 0.0600
o
1
a
3
0
0 0. 0*»00
u.
O 1
> o
i
p. 0200
0.0
A INST. A INST. A '"ST-
GEL GEL GEL
-
A A A
INST ZM SM
•GEL • •
^"sTM^MJ^A33* /
f3Mv^8M * 12 U SOM ^»v '^ ^
*^ "^ IO HH lj FlCC \
48HR^X IOHR .^T-IOKR
"Kim V-WEAK BOM^.FLOC GEL /
"°«- ^ /TcFf^ . 0- GEL
NO GEL NOOCL 8HR
j
NO GEL NO GEL LTGCL
• • «8HR
NO GEL %HP CEt- ,NO ^h
i i •
ZOME OF OPTIMUM COMOIHA-
TIOMS FOR GEL FORMATIONS
& TIME.
KEY
AMOUNT OF SODIUM SILICATE
lit TOTAL GROUT VOLUME
33% (BY VOLUME)
26* (BY VOLUME)
16% (DY VOLUME)
II - MINUTES
H - HOURS
IttST. - INSTANTANEOUS GEL
FORMAT I Oil
FLOC. - FLOCCULATED GEL
STRUCTURE
0 0.010 0.020 0.330
RATIO OF COPPER SULFATE TO SODIUM SILICATE CY WEIGHT
Figure 2: Affect of Various Chemical Mixture* on Gel Formation for Silicate Grout
-------
366 In Place Detoxification
Evaluation indicated that bentonite/cement or silicate
grouts would be most feasible for spill containment. De-
pending on both the soil and chemical characteristics, one
may be more applicable than the other. Both systems are
environmentally acceptable, since the bentonite is a natural
day product and may eventually resorb into the soil and
the silicate grout may break down with time; thus long
term adverse effects will be minimized.
There are several silicate grout formulas in general usage.
The silicate grout used in this survey was formed using a
mixture of sodium silicate, sodium bicarbonate and a
copper sulfate catalyst. Extensive laboratory testing was
performed to establish the most feasible dosages. The
results are plotted in Figure 2. It is anticipated that this
type of presentation will be included in the final systems
operation and maintenance manual with instructions for
choosing an appropriate mix. Chemical tests to determine
the grout's resistance to treatment chemicals were also
performed. The results indicated that the silicate grout
while resistant to bisulfite, hypochlorite. sodium sulfide
and copper sulfate. had very low resistance to acids and
relatively low resistance to bases. This was expected because
the silicate is an alkaline material and the gel is affected by
pH. When a high pH occurs, a bentonite grout would b«
recommended.
The final output of this effort was to develop an approach
for establishing a specific chemical's treatability by in-titu
techniques. This involved determining if neutralization.
oxidation/reduction or precipitation would detoxify the
hazardous material and establishing which type of grout
would be most resistant to chemical penetration. These
results will be presented in the final report and Operation
and Maintenance Manual in tabular form for quick reference.
Pilot Testing
Based on the results of the laboratory tests, two reaction
types and two soil types were chosen for pilot scale evalua-
tion. Precipitation and redox reactions were selected to
further define effect of solids formation. Sand and clay soils
were chosen so that both flow-through and sealed pro-
cedures could be tested on a larger scale. The main objec-
tives of the pilot testing were: 1) to determine if the
detoxification procedure was feasible on a larger than
laboratory scale and 2) to establish critkal parameters such
as pumping rate, injector placement and back pressure, for
consideration in the development of the final system design.
Testing Equipment and Procedures
Special test cells were constructed for the two types of
tests as illustrated in Figure 3. Both were made from coated
plywood, the larger box having heavy reinforcing. Addi-
tional tanks, pumps, tubing and mixers were procured and
used during the test operations, as needed. The test pro-
cedures for the surface and injection treatments were quite
different. The surface testing was basically similar to the
laboratory tests. The soil and water were compacted in the
PLEXIGLASS
TMK.
.o
I
SANO TEST BOX
Figure 3: Pilot Te« Cells
-------
box to a given bulk density and the specified amount of
contaminant wu spnnkled over the surface and allowed to
.migrate. After 24 hours, the reactant was sprinkled on the
soil surface and allowed to detoxify the soil for 48 hours.
Core samples were taken at specified locations in the box
and analyzed for contaminant concentration.
The flow through testing required that the box be filled
with 5600-5800 Ib of sand which wu placed and compacted
in 3.5 on layers to achieve the desired bulk density. Water
was added to yield a 5% water content. The contaminant
was again placed on the surface, and the reaction was per-
formed the same day as contamination. An injector and wet
well were placed on opposite ends of the box and then the
specified volume of reactant was forced through the injector
into the soil. After the reactant was pumped into the
system, a volume of water was injected to rinse the soil of
excess reactant. Throughout the pumping period, the wet
well was continuously emptied into a separate holding tank.
After aO liquids were pumped into the soil, core samples
were collected and analyzed for moisture content and con-
taminant concentration.
Two pilot grouting tests were also performed to aid in
choosing injector types and establishing anticipated pump-
ing pressures and to define some of the problems associated
with grouting. Various mixes of grout were pumped and the
resultant grout wall observed and tested, where possible.
Result* of the Not Tests
Data on the percent of contaminant removed in the pilot
tests are shown in Table EL This measure of extent of
reaction was based on residual concentrations found in the
sofl as opposed to the total amount of contaminant which
had reacted as calculated for the laboratory testing. This
percent reaction is generally higher *^«>i *tyi* fimt^r"'"**1*
percentages, but for a large system it is a better measure of
the overall effectiveness of detoxification. However, direct
comparisons to the laboratory results should not be made.
The effectiveness of detoxification for aD of the pilot
testa was quite high. As expected, the geometry of reactant
injection and the shape of the pilot study box affected the
detoxification. When evaluating the results of flow-through
testing, it was apparent that the detoxification was most
effective within a radius of 1.5 ft from the injector. How-
ever, detoxification effects did extend beyond this radius.
The surface treatment results reflected those predkted
from the laboratory testing. The redox reactions were very
effective, removing most of the contaminant which was
entrained in the surface layers. Precipitation reactions were
less efficient than the redox reaction. This can be attributed
to the blocking of voids by precipitate formation. Shrinkage
cracks which formed when the surface dried allowed more
effective'reaction in some of the lower layers. However, as
with the redox system, the majority of the contaminant
entrained in the surface layer was detoxified.
Evaluation of the grout test results indicated that Injec-
tion of chemical grout on an angle was possible, while
grouting near the sofl surface was not feasible because of
short circuiting caused by grouUngi pressures being larger
than the soil over burden weight. The paniculate grout
was difficult to handle in the shallow testing box and the
only injection device which proved feasible wu one with a
single outlet hole.
The pilot tests abo indicated: 1) the importance of
driving an injector directly into the soil as opposed to
boring and then placing the injector, 2) the necessity of a
wet well equipped with a self priming pump for liquid
removal, 3) the need for pumping systems equipped for
pressures up to 80 psi., 4) the requirement for volumes of
rinse water was not as critical as originally anticipated. S)
the back-pressure caused by higher void volume loadings of
contaminant reduced the forward flow rate significantly
and 6) the neutralization chemicals could be added using a
multi-holed injector (which allowed for much faster treat-
ment). It was determined that pilot test grout gel times
were shorteT than in the lab and that the chemical grout
injection could be controlled by the volume added while
the particulate grout addition was best regulated by pres-
sure in the injection lines.
Prototype Design
rVemninary Design
After the pilot tests were completed, the design of the
prototype system was begun. Much of the information
obtained throughout both the laboratory and pflot tests
significantly influenced the design. A process and instru-
mentation diagram is shown in Figure 4 and a layout is
shown in Figure S. The system provides much flexibility for
spul cleanup. The grout or chemicals are to be mixed in
alternate batches in the two 1500 gal fiberglass tanks.
Batching eliminates potential problems associated with
exact mixing of grout constituents at the point of injection
and thereby allows closer system control.
Two pump types were included. For grouting, positive,
displacement pumps wOl provide the most control and the.7
simplest operation, however they wen not sufficiently
chemically resistant for chemical injection which wfll be
accomplished by the sir pumps, available in Hasteuoy C It
was also determined that multiple pumps instead of exten-
sive manifolding of injectors would allow more control of
the volumes pumped into the soil. If necessary, the injectors
can be manifolded in pairs to allow higher pumping rates,
however this approach may not always be feasible when
difficult so3 conditions are encountered. The volume of
liquid added is to be metered and totalized, since in most
instances the chemical solutions wiD be added until a cal-
culated amount is pumped into a specified area. The injector
win then be withdrawn a certain distance and the pumping
process repeated.
The vehicle wfll be equipped with a diesel-electric genera-
tor and an air compressor. An •"air-hammer" type device
wfll be used to drive the injectors (1M in. OD. 1 in. ED) into
the ground. Separate multiholed injectors wfll be used for
chemical addition. Since the cost of chemical resistant
injectors would be excessive, standard steel pipe injectors
wfll be replaced when they corrode to the point where they
an no longer usable. AO components would be accessible
either on the vehicle or from the aid*. The controls will be
centralized on a panel permanently mounted on the truck.
Accessory equipment wfll include standard test apparatus
-------
Jos Ln nice ueioxinciuoa
to measure fofl conditions and chemical concentration!, Cosu are presently being developed and this design may
well points for use as wet wells, some small air pumps to be modified depending upon the complete economic con-
empty wet wells, and a surface holding tank. siderations.
Table HI: Summary of PQot Testing Results
Test Condi tlons
Test
:io. Media Containment '/.Lotdlnij Location
1 sand llaOCl 25 top
aid
bot
H sand riaflCl SO top
•Id
bot
Z sand CuSOij 25 top
mid
bot
3 sand CuSOl, SO top
mid
bot
7 clay MaOCl 25 top
•14
bot
8 clay KaOCI SO top
mid
bot
9 clay CuSOi, 2$ top
mid
bot
fo clay CuSQij 50 top
mid
bot
It Rcnoval 100 * • (concentration of contaalnant In -
sol I before treatment
- Results
Avg Cont Avg Percent Avg Percent
Cone Renoval(Tot) P.enoval ( In j)
Cl S0j
•ill B7 ?9"T73
20C6 55.2 95.7?
337* 97.6 95.76
Cl SOj
2535 5T.3 95.92
2218 100 55. ?0
6606 97.3 99.89
Cu Cu
10*0 7T.5 8«».S
1253 85.2 85.7
1262 88.7 88.2
Cu Cu
2096 75". 7 8O
5791 9*.9 97.5
8530 96.5 97.5
Cl
20306 57.7
413 85.0
28 60.7
Cl SO,
20306 97.9 95.9
*I3 82.3 99.85
28 85.7 99.32
i
Cu
8197 7?.8
2653 99.5
86 76.7
Cu
8197 70.6
2653 76.0
86 7*.8
concentration after)
treatment
concentration of contaminant In •*
soil before treatment '•
-------
•8
•;»
IT
'•>
Ficure 4: Process tad InftiumenUtiom Diafram of Prototype Unit
De«gn UniUtioM and DecUoa Matrix
The limitatioas of in-titu detoxiftcatioa techniques
either through surface treatment or direct injection of grout
and chehiicals must be understood before the prototype
equipment is used. When a land spul occurs, altemaUrt
approaches should be evaluated and the most time and cost-
effective approach for the specific situation chosen. la
order to determine if fn-tttu detoxification is most efficient.
a decision matrix wfll be prepared. This matrix wul present
an approach for evaluating the feasibility of grouting and
chemical injection, as weO as surface containment and treat-
ment. Among the critical variables an type of chemical
spuled, interaction with the soil, the soil's "grouubOity"
(permeability, void loading, geometry, water table level.
etc.X soil volume contaminated, feasibility of excavation
and availability of treatment supplies and manpower.
This equipment wfll not be applicable to all land spiOs.
However, then an many situations in which it wul be a
feasible technique. The surface treatment approach may be
desirable in many cases even if the spuled soO is to be
removed and transported to • landfill. This pretreatment
wul protect equipment and may even allow redefinition of
the removed sofl as non-hazardous. Grouting in and of itself
wfll be feasible even when direct chemical treatment is not
possible. Construction of a grout layer will protect the
-------
ground water if excavation is incomplete or if rain rimes the
area. Although grouting will be limited to relatively coarse
grained und and gravel materials, it is these sous that allow
permeation of the contaminant through the sol structure
and into (he groundwater.
Design Changes
Several changes have been made in the initial design con-
cept. Most significant is the addition of a surface treatment
technique for fine grained sous. Polymerization was limited
to a few possible materials and was determined to be too
dangerous to implement in a Reid situation. The pilot tests
indicated that it was critical to meter liquid flows indivi-
dually so the original design which included a high capacity
pump with extensive manifolding of injectors was changed
to include a larger number of lower capacity pumps with
much less manifolding.
It was also determied that the pumping rates for chemi-
cal injection should be relatively low to allow effective
reaction. Therefore the overall time required for treatment
will be longer than anticipated.
CONCLUSIONS
I. An In-place treatment technique has been shown to be
an effective land spill cleanup on a laboratory and pilot
scale basis.
2. Grouting technology appears to be an effective method
to contain ipills and thereby minimize potential ground-
water contamination.
3. Where small grained soils (tflts and day) preclude the use
of injection equipment, a surface treatment using a
diluted reactant provides an efficient way to detoxify
land spills of applicable hazardous materials.
4. In order to establish the most time and cost-effective
method for land spill cleanup, the limitations of the in-
place detoxification as well as specific spill variables
must be considered.
S. A stepwise approach to containment by grout injection,
followed by chemical treatment seems to provide the
most flexible treatment system.
ACKNOWLEDGMENT
The work oa which this paper is based was performed
under Contract 6S-03-2458 with EPA's Oil and Hazardous
Materials Spills Branch, Industrial Environmental Research
Laboratory (Cincinnati) Edison, New Jersey."
IWmiifiiiiiiiiniiJ
Figure 5: Preliminary Layout of Prototype Unit
-------
control of
HAZARDOUS MATERIAL SPILLS
Proceedings of the 1978 National Conference on
CONTROL OF HAZARDOUS MATERIAL SPILLS
April 11-13,1978
Miami Beach, Florida
Sponsored by:
United States Environmental Protection Agency; United States
Coast Guard; Hazardous Materials Control Research Institute
In Participation with:
Oil Spill Control Association of America
-------
V>EPA
FACT SHEET
United States
Environmental Protection
Agency
April 1982
In-Situ Containment/Treatment System
EPA'j Off let of Research and Development (ORD) has recently completed con-
Uructton of a Mobile In-Situ Containment/Treatment Unit designed for field
use to detoxify sol It which have Deen contaminated by hazardous ••terlils
fro* spills or uncontrolled hazardous waste sites. EPA develops such equip-
ment to actively encourage the use of cost-effective, advanced technologies
during cleanup operations. Once an Item of hardwire Is complete. It Is tested
under field conditions. After testing, the plans, specifications and other
Information are made available publicly for the purpose of encouraging com-
mercialization of the new technology. Numerous syste*s, Including • Mbtle
w«ter treatment unit and a Mobile laboratory, have been developed by QUO,
«ere duplicated by the private sector, and are MM available commercially.
When spills, or hazardous substance releases fro* waste sttesM. contaminate
soils and threaten nearby surface water or groundwater, an effective method
of treating tne soil Is needed. Excavation and hauling of contaminated soil
to a secure landfill Is one solution. However, this approach Is not practi-
cal for those Incidents where a large volume of soil Is Involved. An alter-
nate commercially feasible approach Is to flush the soil In place with water.
ORD Is developing an Innovative, Improved method for treating contaminated
soils In place at reduced cost. In terms of dollars per pound of contaminant
removed. The technique employs flushing with additives and detoxification
by chemical reaction.
The mobile In Situ Containment/Treatment Unit, shown left, Is mounted on a
13.1-m (43-ft) drop deck trailer and Includes: a dieset electric generator,
an air compressor, mixing tanks, hoses, a solids feed conveyor, pipe Injec-
tors, soil testing apparatus, and accessory Items. In-situ containment Is
accomplished by direct Inaction of grouting material Into .the soil around
the contaminated area In order to Isolate the released chemicals. The chemi-
cals are then treated In place by flushing, oxidation/reduction, neutraliza-
tion or precipitation. Specially prepared solutions Of wash water can be
delivered into highly contaminated soil through 16 Injectors. A vacuum well-
point withdrawal system (not shown) creates an artificial hydraulic gradient
which draws the wash solution from the Injectors through the contaminated
soil thereby collecting water-soluble contaminants In the solution. The with-
drawal system has granular activated vapor-pnase carbon packs for removal of
organic vapors released during the withdrawal operation.
The collected chemically contaminated wash solution Is processed through a
mobile water treatment unit, where contaminants are removed. Fresh chemical
additives are then Introduced Into the cleansed wash solution which is re-
injected Into the contaminated area. This process Is continued until •
point of diminishing returns Is reached.
For further Information, contact Kr. Frank J. Freestone or Nr. Richard P.
Traver, Municipal Environmental Research Laboratory, Oil $ Hazardous Mater-
ials Spills Branch, Edison, New Jersey 08837. Telephone numbers are: (201)
321-6632/6677 (Commercial); 340-6632/6677 (FTS).
-------
TREATMENT OF SOILS CONTAMINATED WITH HEAVY METALS
William 0. Ellis. Thomas R. Fogg
Science Applications International Corporation
8400 Westpark Drive
McLean. VA 22102
Anthony N. Tafuri
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison. NJ 08837
ABSTRACT
The U.S. Environmental Protection Agency's Hazardous Waste Engineering Research Lab-
oratory has initiated a program to evaluate in situ methods for mitigating or eliminating
environmental damage from releases of toxic and other hazardous materials to the soils
around uncontrolled hazardous waste disposal sites. As part of this program, various re-
agents suitable for the in situ washing of heavy metal contaminants from soil were tested
at laboratory scale. The work was performed on a soil from an actual Superfund site near
Seattle. WA. The soil contained five toxic heavy metals often found In hazardous waste
site soils:-cadmium, chromium, copper, lead, and nickel.
The tests demonstrated that sequential treatment of soil with ethylenedlamlnetetra-
acetic add (EDTA), hydroxylamlne hydrochlorlde, and citrate buffer was effective 1n re-
moving metals from soil, and all were necessary for good cleanup. The EDTA chelated and
solubllized all of the metals to some degree; the hydroxylamlne hydrochlorlde reduced the
soil Iron oxide-manganese oxide matrix, releasing bound metals, and also reduced insoluble
chromates to chromium (II) and (III) forms; and the citrate removed the reduced chromium
and additional acid-labile metals. The best removals observed were: cadmium, 98 percent;
lead. 96 percent; copper, 73 percent; chromium, 52 percent; and nickel. 23 percent.
INTRODUCTION
The U.S. Environmental Protection
Agency's (EPA) Hazardous Waste Engineering
Research Laboratory (HWERL) Initiated a
program to develop In situ chemical methods
for mitigating or eliminating environmen-
tal damage from releases of hazardous ma-
terials at chemical spill sites and around
'hazardous waste disposal sites. As part
of this program. Science Applications
International Corporation (SAIC) ,'under
EPA Contract No. 68-03-3113. Investigated
chemical methods for In^ situ cleanup of
heavy-metal-contaminated soil.
Toxic heavy metals are frequently
found in soil at uncontrolled hazardous
waste sites, including lead (15 percent of
sites surveyed), chromium (11 percent).
cadmium (8 percent), and copper (7 percent)
(Ellis and Payne. 1983).
-------
Based on these results, an optimum
treatment sequence was designed. Then
column tests of the optimum treatment
sequence were conducted.
The column studies evaluated metal re-
moval under gravity flow conditions, with
analysis of soil and duplicate analysis of
leachate after each treatment. A three-
agent sequential extraction was tested
using five pore volumes of the optimum
concentration and pH for the EDTA solution
to eemove most metals, followed by hydrox-
ylamine hydrochloride to reduce any hexa-
valent chromium to trlvalent, and to
reduce any soil iron or manganese oxides
to release any bound metal. Citrate
buffer was then used as a final acidic
leaching agent. The same metal-contamin-
ated soil was used for all tests; all
initial concentrations for each metal were
the same (see Table I).
Samples were analyzed for trace ele-
ments by atomic adsorption spectrophotom-
etry (AAS) using flame or graphite fur-
nace procedures. Analyses by the method
of standard additions were routinely
performed along with standard calibrations.
When the two calibration curves deviated
significantly, calculations of sample
concentrations were based upon the stan-
dard addition calibration; when they were
the same, a combination of the standard
addition/standard calibration was used.
Sample blanks and National Bureau of Stan-
dards (NBS) standards were analyzed in the
same manner as the samples.
TABLE 1. SINGLE AGENT SHAKER TABLE EXTRACTION EFFICIENCIES
Soil Metals (ppm)
Cd
47
Cr
Cu
N1
219
T14
Pb
2^480
EDTA (0.1 M 9 pH 6)
% Extracted
114
24
62
14
106
Hydroxylamine hydrochloride
(0.1 M In acetic acid)
% Extracted 86
32
43
20
80
Citrate buffer (0.1 M 9 pH 3)
% Extracted 77
24
48
14.5
65
Pyrophosphate (0.1 M)
% Extracted
5.4
9.6
29
2.9
9.7
DPTA (0.005 H in
0.1 M trlethanolamlne)
% Extracted
59
48
67
-------
Conventional remedial methods for
.sites containing heavy metals Include
"excavation followed by land disposal and
groundwater punping and treatment. The
use of excavation and land disposal 1s
meeting with Increased opposition not only
because of high cost but also because the
contaminated soil is slnply transferred to
another location. Also, pump and treat-
ment methods are costly and are not effec-
tive for removing contaminants sorbed to
the'soil. In situ treatment of toxic
metals in soTl and groundwater offers a
potentially cost-effective remedial
alternative. However more research is
needed before the ji£ situ methods can be
implemented in the field.
The objective of this project was to
select the most promising in situ treat-
ment method for metals and evaluate the
method through laboratory studies. The
study was limited to methods suitable for
J£ situ treatment of cadmium (Cd), chro-
mium (Cr), copper (Cu), lead (Pb), and
nickel (Ni). These metals are found
frequently at hazardous waste sites and
are among the most toxic. Methods that
are effective with these metals might
also be suitable for treating other heavy
metals found at hazardous waste sites.
Potential in situ treatment methods
for metals Include methods that immobilize
the metals 1n soil by means such as preci-
pitation and methods that solubillze and
remove the metals from the soil. Methods
that solubillze and remove the metals
offer an advantage over immobilization
methods because the need for long-term
monitoring is eliminated. Immobilization
methods, on the other hand, simply reduce
the concentration of dissolved species.
The potential exists for resolubilization
of the metals through subsequent natural
chemical reactions; therefore, the site
must be continually monitored.
Methods for mobilizing metals in
soils involve the use of dilute weak
acids, bases, or aqueous solutions of
.chelating agents. Considerable research
on a laboratory scale has already been
conducted on the use of chelatlng^and
other complexlng agents for selectively
removing Metals from soil.
This research demonstrated different
degrees of extractabllity of any given
heavy metal from soil. The extractabillty
has been described according to which type
of extraction agent will remove the bound
metal which corresponds to a specific soil-
metal binding mechanism or the chemical
state of the metal. For example, soluble
heavy-metal salts are extractable with
water; metals bound to the soil organic
fraction are extractable with aqueous
alkaline buffers such as tetrasodlum
pyrophosphate ("tetrapyrophosphate"); and
metals occluded in the iron and manganese
oxide fraction of the soil are released
by reduction of the oxides with hyroxy-
lamlne hydrochloride. These techniques,
if developed further, could be used for
the cleanup of contaminated soil at
hazardous waste sites.
Laboratory Task Description
Laboratory studies were conducted to
determine whether In situ cleanup of heavy-
metal-contaminated soil by treatment with
chelating solutions or acidic buffers was
possible. The soil used In the studies
was collected from the Western Processing,
Inc. Superfund site, near Seattle. WA.
Previous analysis of this soil (Repa, et ,/
al, 1984) had shown high levels of cadmium,
chromium, copper, and lead (>10 pom). ' '-
The laboratory task consisted of:
(1) soil characterization; (2) laboratory
equilibration (shaker table) experiments
designed to evaluate treatment methods
(i.e., single agent treatment vs sequential
treatment with several agents) for metal
removal; and (3) soil column tests to
evaluate cleanup efficiency under gravity
flow conditions.
Based on a review of the literature,
the chelating agent ethylenediaminetetra-
acetic acid (HOTA), the reducing agent
hydroxylamine hydrochloride, and the
acidic citrate buffer were Identified as
suitable agents for testing. Shaker table
equilibration studies were conducted in
which various combinations of the above
treatment agjnts (10:1: w/w agent solution:
soil), either singly or 1n sequence, were
shaken with the contaminated soil in a
closed container on a vibrating platform.
-------
RESULTS AND DISCUSSION
Soil Characterization
Soil permeability measured in the
laboratory was approximately 5 x \Q-$
cm/sec. The grain size distribution was
determined by wet and dry sieve procedures
and pi pet analyses on organic-free soil
after a hydrogen peroxide wash. Approxi-
mately 75 percent of the soil was in the
silt and clay range. This probably caused
the rather slow percolation rate. X-ray
diffraction analysis showed alpha-quartz
and feldspar to be the only measurable
constituents of the soil. No measurable
amounts of crystalline aluminum oxide
forms were present. The total carbon
content of the soil averaged 16.400 ^ 709
ppm by weight (1.64 percent). This Tnter-
mediate level of carbon corresponds to the
phenols and other organic compounds found
in the soil.
The cation exchange capacity (CEC) of
the soil was also determined. The results
were 13 and 8.2 milliequivalents per 100 g
for bulk and organic-free soil, respec-
tively. These results are quite low and
indicate an absence of mineralogic clay in
the soil. The pH and Eh measurements
(made i-n triplicate) yielded an average
soil pH of 7.39 and an Eh of +0.198 v
(electron potential, pe - +7.01), reveal-
Ing a neutral, slightly oxidizing soil.
The iron and manganese oxide mean concen-
trations were 15.000 and 291 ug/g, respec-
tively. The carbonate results yielded an
average value of 1.42 meq/g as bicarbonate.
The results of the determination of
heavy metals of interest in Western Pro-
cessing soil were as follows (in ug/g):
cadmium (47). chromium (349). copper
(219). iron (30,200), manganese (1,690).
nickel (214), and lead (2,480). These
values were compared with the concentra-
tions of'the metals in the treatment
solution to assess percent removal of
metals by the treatment.
Shaker Table Studies
In the single'shaker table extractions
using EOTA at Different concentrations and
pH values, the 0.1 H solution was much
more effective in metal removal than the
0.01 M solution. The pH trends, however.
were not so clear cut. A pH of 6 was
chosen as the optimum because it afforded
slightly better chromium removal than that
obtained at pH 7 or 8; EDTA 1s more ion-
ized at pH 6. This pH and concentration
combination was used in subsequent
studies.
The results of the EOTA, hydroxylamine
hydrochloride, acidic buffer, and diethyl-
enetriamine pentaacetic acid (DTPA) single-
method shaker table extractions (Table 1)
showed that EDTA was the best single
extraction agent for all metals. However,
hydroxylamine hydrochloride was more
effective at chromium extraction.
Results of the two-agent sequential
extraction (Table 2) indicated that the
EOTA was much more effective in removing
metals than the weaker agents often used
to characterize the mechanism of binding
of metals to soils. Thus, weaker extrac-
tion techniques (magnesium chloride,
potassium fluoride, acetate buffer, tetra-
pyrophosphate) can be eliminated if just,
an EDTA solution Is used.
The results of the three-agent sequen-
tial extraction studies (Table 3) showed
that, compared to bulk untreated soil,
this extraction scheme removed nearly all
the lead and cadmium, 73 percent of the
copper, almost 52 percent of the chromium,
and only 23 percent of the nickel. Over-
all, this scheme was shown to be better
than three EOTA washes, better than switch-
ing the order of EDTA and hydroxylamine
hydrochloride, and much better than simple
water washes, in subsequent three-agent
tests. However, the EDTA washing alone
might be used with only a slight decrease
in removal efficiency.
-------
TABLE 2. TWO-AGENT SEQUENTIAL SHAKER TABLE EXTRACTION EFFICIENCIES
Soil Metals (pom)
EOTA (0.1 M (a pH 6)
% Extracted
Magnesium chloride (1 M)
% Addnl. Extracted
EOTA (0.1 M (» pH 6)
% Extracted
.Potassium fluoride (0.5 M)
% Addnl. Extracted
EOTA (0.1 M9 pH 6)
Cd
47
83.6
1.02
95.3
1.17
X Extracted 119
Acetate buffer (1 M 9 pH 5)
I Addnl. Extracted
EOTA (0.1 M 0 pH 6)
% Extracted
Tetrapyrophosphate (0.1 M)
% Addnl. Extracted
TABLE
2.36
75.3
23.9
Cr
349
24.4
0.11
28.9
0.37
24.3
2.36
24.2
5.59
Cu
219
77.6
2.22
56.4
1.27
76.3
1.18
59.6
3.11
N1
214
10.8
1.47
11.6
0.47
10.7
1.89
9.72
0.99
3. CUMULATIVE SHAKER TABLE
THREE- AGENT SEQUENTIAL
Soil Metals (ppm)
1) EOTA (0.1 M P pH 6)
2) Oeionized water
3) Hydroxylamine hydro-
chloride (0.1 M in
• acetic acid)
4) Deionized water
5) Citrate buffer
(0.1 M f> pH 3)
(•Total % Extracted)
Cd
47
87.2
92.5
96.3
96.6
98.4
EXTRACTION
Cr
349
24.6
27.5
34.0
34.5
51.9
EFFICIENCIES
Cu
219
63.0
67.4
69.8
70.1
73.0
(%)
N1
214
13.8
15.4
19.8
20.6
23.0
Pb
2.480
84.6
0.29
85.3
0.85
117
1.41
98.2
..
1.20
i '
Pb
2,480
87.1
92.6
94.8
94.9
96.4
-------
Column Studies
The results of the metals extraction
achieved during column tests are shown 1n
Table 4.
The pattern of removal for each metal
was somewhat unique. Lead appeared to
be removed easily by the EDTA; further
removal occurred with citrate. Cadmium
was removed by EDTA and also by hydrox-
ylamine hydrochloride; removal was slight-
ly improved with the other treatments.
Copper was removed only by EDTA; the other
treatment methods had little effect on
removal. The data indicated a generally
hiyh extraction efficiency for EDTA. The
analysis of metal remaining in soil versus
pore volume and type of treatment indica-
ted that lead and cadmium concentrations
in soil decreased steadily from the begin-
ning of treatment to the end. The pattern
for the other metals was similar, but with
slight differences, probably due to random
sampling-or analytical errors. Chromium
appeared to exhibit a pattern of migration
from the top to the middle of the column,
followed by rather ineffective removal.
Nickel showed a similar trend. These
latter results suggest that more pore
volumes of each treatment solution (e.g..
10 rather than 5) would improve the re-
moval, probably to the level of extraction
efficiency achieved in the shaker table
tests.
CONCLUSIONS
The results of the shaker and soil
column studies permit a number of con-
clusions about the potential feasibility
of in situ cleanup of soil contaminated
with heavy metals*
The Cleanup Efficiency of the Soil Treat-
ment Agents
The various treatment-agent tests
showed that there are definite differences
in efficiency of the agents that vary with
the heavy metal.
TABLE 4. THREE-AGENT SEQUENTIAL EXTRACTION EFFICIENCIES:
SOIL COLUWi TESTS
Soil Metals (ppm)
Cd
Cr
Cu
219
N1
Pb
2,480
Water
X Extracted by water
0.2
0.1
EO.TA (0.1 M 9 pH 6)
% Extracted by agent
60.5
12.2
47.1
6.8
60.1
Hydroxylamine hydrochloride
(O.I M in acetic acid)
X Extracted by agent 23.8
8.9
n.7
H.7
2.3
Citrate Buffer (n.l M 9 pH 3)
1 Extracted by agent 3.6
12.2
0.2
4.8
8.8
Water Wash
X Extracted by water
0.4
1.1
0.1
0.5
0.5
Total X Extracted:
88.5
34.4
48.1
20.8
71.8
-------
The preliminary tests of single heavy-
met-al treatment agents provided the opti-
mum concentration and optimum pH for EOTA
treatment. The more concentrated solution,
O.l M EDTA, 1s clearly more effective. A
pH of 5 is probably as effective as pH 6,
but either 1s more effective than pH 7 or
above.
The two-agent tests demonstrated that
weaker agents do not remove any of the
metals of Interest more efficiently than
EDTA alone.
The three-agent tests demonstrated
that EOTA, hydroxylamine hydrochlorlde, and
citrate buffer are all necessary for good
cleanup of the soil. The EDTA chelates
and solubilizes all of the metals to some
degree; the hydroxylamine hydrochloride
probably reduces the Iron oxide-manganese
oxide matrix, releasing bound metals, and
also reduces Insoluble chromates to chro-
mium (II) and (III) forms; and the citrate
removes the reduced chromium and addi-
tional acid-labile metals. The chelating
agent/reducing agent/acidic citrate buffer
combination appears to be very effective
in heavy-metal cleanup.
The three-agent test with just EDTA
demonstrated that cleanup of cadmium and
chromium Is significantly better with the
sequential EDTA/hydroxylamlne/ citrate
than with three treatments of EDTA alone.
However, EDTA alone appears to be suffi-
cient for removing the lead and copper;
although the nickel removal was poor with
EDTA alone, the treatment with all three
agents showed no better removal.
The three-ayent test with hydroxyla-
mine hydrochloride first, followed by EDTA
and then citrate, demonstrated that the
use of a chelating agent following the re-
duction step does not improve the cleanup.
Effects of the Soil Characteristics on the
Cleanup Efficiency"
The efficient cleanup of the heavy-
metal contamination in the soil was prob-
ably facilitated by the low cation ex-
change capacity (CEC) of the soil. How-
ever, the presence of Iron and manganese
oxides apparently Interferes with heavy
metal removal by EDTA; reducing these
oxides was necessary to remove all the
cadmium.
Feasibility Studies Using Shaker and
Column Tests
The shaker studies were quick and
effective screening tests for estimating
treatment-agent efficiency. The column
tests, although more difficult and time-
consuming more closely represent the
behaviour that might be expected If the
agents were used for In situ cleanup of an
actual contaminated sTte"The column tests
model cleanup under gravity flow conditions
through soil with a permeability somewhat
similar to the native soil. If time had
permitted longer soil column tests, extrac-
tion efficiencies would probably .have been
similar to the shaker table test results.
Both the shaker and column tests are very
useful for studying the feasibility of / ,
potential soil cleanup agents. ...
REFERENCES
1. Ellis, M. D., J. R. Payne, and 6. 0.
McNabb. 1985 Treatment of Contamina-
ted Soils with Aqueous Surfactants.
EPA/600/S2-85/129 U. S. Environmental
Protection Agency.
2. Repa, E. W., E. F. Tokarski. and R. T.
Eades. 1984. Draft Final Report.
Evaluation of the Asphalt Cover at the
Western Processing, Inc.. Superfund
Site. EPA Contract l68-03-3li3. U.S.
EnvTronmental Protection Agency.
-------
STUDENT PAPER COMPETITION
*
To encourage student participation in the Association and to
recognize outstanding research at New England colleges and
universities, NEWPCA recently held its fourth annual student
paper competition. Judges under the direction of Mr. William
VanBlarcom reviewed a number of entries and selected four
finalists who presented their papers at the January 28 session of
the NEWPCA 1985 winter meeting. Based on the quality of the
written papers and the oral presentations, judges selected
Camille C. Connick as winner of the $200 cash award. Other
finalists, each of whom received a $100 cash prize, were Robert C
Backman, Northeastern University (The Treatment of Dairy
Wastewater by the Anaerobic Up-Flow Packed Bed Reactor);
AkbarJohbri, University of Rhode Island (A Pilot Study of the
Responses of Powered and Granular Activated Carbon in the
Removal of Shock Loadings of Synthetic Organics); and Bid
Alkhatib, University of Rhode Island (Treatment of a Combined
Petrochemical Industrial Waste Stream for Reuse).
Presented herein is Ms. Conflict's winning entry. Copies of
the other finalists'papers are available from NEWPCA.
•
MITIGATION OF HEAVY METAL MIGRATION IN SOIL
BY CAMILLE C. CONNICK*
INTRODUCTION
The uncontrolled or accidental contamination of the environ-
ment with hazardous materials through chemical spills and
hazardous waste site releases necessitated the enactment of the
Comprehensive Environmental Response Compensation and
Liability Act of 1980 (CERCLA) often called Superfund. The pur-
pose of one Superfund program, the U.S. Environmental Protec-
•GudusU Student, DepL of Civil Engineering. Northeastern University.
Boston, Massachusetts.
C.C. CONNICK 5
tion Agency's (USEPA) Chemical Countermeasures Program
(CCP), is to investigate in-situ chemical methods for mitigating or
eliminating environmental damage from releases of toxic and
other hazardous materials to the soils around uncontrolled hazar-
dous waste disposal sites and from spills of hazardous chemicals
to still or relatively slow moving surface water bodies. Because it
is recognized that the environmental impact of hazardous
material spills and releases can be worsened by adding chemicals
indiscriminately, the CCP is intended to provide guidance and
define appropriate treatment techniques.
This investigation of in-situ treatment of soils contaminated
by heavy metals was performed as a joint research project with
the USEPA and Northeastern University. The study involved
the determination of adsorption isotherms for the heavy metals
and specified soil, as well as the desorption behavior of the metal
using water rinses, water and surfactant rinses, and water plus
chelating agent rinses. The first phase consisted of shaker table
agitation (equilibration) to determine maximum adsorption of
metal to soil. The second phase involved the use of soil column
studies to evaluate the maximum adsorption/desorption of the
metal. A simulated spill of heavy metal-laden liquid for soil con-
tamination was followed by successive treatment rinses under
gravity flow conditions to determine removal efficiencies. In-
fluent and effluent pH, metal content, permeability rates and
variations, and chemical oxygen demand (COD} were monitored
during the study to determine metal removal efficiencies and the
occurrence of unanticipated reactions.
The results of this research and results from a similar study
investigating the use of in-situ treatment of soil contaminated
with hazardous organic constituents are to be used as the basis
for development of pilot scale testing in a chemical additive treat
ment tank at USEPA's Oil and Hazardous Materials Spills En
vironmental Test Tank (OHMSETT) facility in Leonardo, NJ.
BACKGROUND INFORMATION & LITERATURE REVIEW
The soil used in the chemical countermeasure study wa:
selected based on its frequency of occurrence at Superfund site;
in New Jersey and also its availability for excavation in or
-------
METAL MIGRATION IN SOIL
uncontaminated condition. The soil selected for the research was
Typic Hapludult of the Freehold Series. It is described as fine to
coarse loamy, low clay content (< 16%) and a high content (>
15%) of fine, medium, and coarse sands, plus coarse fragments up
to three inches. Only soil from the B horizon was intended to be
used so as to attempt to model soil containing contaminant
releases which are greater than two feet underground Such
releases usually affect large volumes of soil, making excavation
and land treatment methods and equipment uneconomical and
physically impractical.
The characteristics of soil have a tremendous effect on the ef-
ficiency of various treatment processes for contaminant removal
Grain size, specific gravity, density and water content
characterizations determine available void volume, soil porosity,
and permeability which directly affect both pollutant and treat-
ment considerations. Buffering capacity and soil pH affect
neutralization and possibly precipitation reactions resulting in
enhanced or decreased water solubility of products. High organic
soils (such as peat) have a higher affinity for non-polar organics,
which can affect in-situ treatment with surfactants and/or
solvents. A high cation exchange capacity (CEC) observed in
some clays and fine silts can attenuate treatment of metals and
metal salts. A high mineral content can affect neutralization and
redox treatment of acid spills. In some cases, treatment of a
caustic spill with acid might increase resolubilization of inherent
metal species. Interfering reactions can result in a need for a
greater volume of the treatment reagent, increasing wastewater
treatment requirements.
A complete chemical and physical analysis of the soil was
performed prior to the start of the studies by JRB Associates1.
The mineralogical composition of Clarksburg soil was determined
using X-ray detraction studies. Quartz is the major phase,
representing at least 98 percent of the total weight. No
measurable amounts of clay minerals appeared in the sample
which accounts for the relatively low CEC of 8.6 mg Na/100
grams. The organic carbon analysis showed only 0.12 percent.
C.C. CONNICK
The average permeability when compacted to a density of 107
Ibs/cu ft was 1.5 X 10"' cm/sec. The natural moisture content waa
10 to 12 percent.
Metal Contaminant!
The heavy metals (Cd, Cu, Pb, Ni, Zn) selected for use in the
reseach were chosen based on frequency of occurrence in soil at
USE PA Region II Superfund sites and concern for toxicity to
human health and the environment The metals Cd, Cu, Pb, Ni
and Zn were detected in soil at 4,3,7,3 and 6 of 60 sites reviewed,
respectively, at concentrations ranging from 2.000 to 30,000
ppm. The toxicity of these metals in such elevated concentrations
is apparent when compared to the acceptable* concentrations
specified by USEPA water quality criteria and the reported Rat
Oral LDM of these cations (Table 1).
Table 1. WATER QUALITY AND TOXICITY LIMITS
USEPA Water Quality
Criteria, ppm
0.01
1.
0.06
0.0134
6.
Rat Oral LDM
mg/kg
88 (CdCIJ
266 (CuCl)
106 (NiCLJ
350 (ZnCIJ
Chemical Countermeasurea
A literature review was performed to investigate the avail-
able methods for in-situ treatment of contaminants. Three types
of treatment were reported for either removing or fixing con-
taminants in soil including: use of surfactants to solubilize and
flush contaminants; use of chemicals for in-situ metal precipita-
tion; and use of chelating agents for metal extraction. __.
Surfactants were reported as being successful in the recovery
of gasoline from soils and as having the ability to solubilize
organic materials that were previously only solubilized in organic
-------
8
METAL MIGRATION IN SOIL
solvents'. Several analyses were performed by JRB Associates'
to determine the appropriate concentration of surfactant/water
solution which would provide adequate pollutant extraction effi-
ciency and yet not inhibit soil-column flow. A mixture of two non-
ionic surfactants, one percent Adsee 799 (Whitco Chemicals) and
one percent NP90 (Diamond Shamrock) in tap water was chosen
as the chemical countermeasure to be tested for in-situ treat*
ment. Both surfactants, Adsee 799 and NP90, are considered
non-toxic. They are often used for treating farmland to enhance
percolation of fertilizers and irrigation waters. The surfactants
are biodegradable and the potential for excessive accumulation or
hazardous effects is minimal, which further enhances their ap-
plicability for in-situ removal of organic contaminants. The high
organic content of the surfactant allows one to monitor its con-
centration in soil leachate by performing analyses such as the
COD determination of organic content.
The use of sodium sulfide for in-situ metal precipitation and the
use of ethylenediaminetetracetic (EDTA), a chelating agent for
metal extraction were reported as successful in fixing and remov-
ing heavy metal contaminants in soil. Chelating agents are com-
pounds or ligands (generally organic) that coordinate or bond a
metal ion in more than one position. This bonding of the metal
ion. in most cases results in its deactivation. The metal is no
longer able to react chemically and is, therefore, made less toxic*.
Competition from hydrogen ions usually occurs at low pH levels.
A decrease in pH always produces a shift towards disaasbciation
of the complex ion (an increase in free metal concentration).
Organic chelating agents may be divided into two classes, se-
questrant* and precipitates. Sequestrants form chelate com-
plexes which are soluble in water; therefore, the compound still
remains distributed throughout the water body although in a less
toxic form.
EDTA is a sequestering agent used in metal cleaning, preser-
vation of canned fruits and vegetables, leather tanning, and in
medical treatment of Zn, Fe, Ni, Pb, and Hg poisoning. EDTA is
generally applied as a soluble sodium salt along with a buffer
solution such as ammonia ammonium nitrate to maintain a pH of
C.C. CONNICK
9 to 10. Since the effectiveness of the chelating agent EDTA i:
pH dependent, the buffer solution was prepared so as to mainlaii
a pH of 9 to 10 when subjected to the acidity of the soil system a
the time of treatment and during the displacement of hydrogci
ions as the EDTA reacted with the metal cations in the soi
system*. A 0.144 M concentration of disodium EDTA wa
selected as the chemical countermeasure to be tested in thi
research along with the prescribed surfactant combination aup
plied by JRB Associates and tap water1.
EXPERIMENTATION METHODS AND MATERIALS
The laboratory study conducted to evaluate the effectivenes
of the chemical countermeasures included shaker table agitatio:
and gravity flow soil column studies. To insure data accuracy
replicate leachate samples were analyzed along with blan!
samples (non-contaminated soil mixed with deionized water) fo
each run during shaker table analysis and column testa. Al
glassware, plastic ware, columns, storage vials, and any ir
struments used in the study were acid cleaned (1 + 1 HNOJ arn
rinsed with deionized water where feasible. Control samples o
metal contaminants were placed in shaker table bottles and a co:
umn to evalute the extent of the cation adsorption onto the e>
perimental apparatus throughout the course of the study.
Shaker Table Studies
Four different concentrations, as shown in Table 2, wer
prepared for each metal using a solution of the sulfide or aceUt
salt of the metal with deionized water. The selection of the mete
Table 2. METAL CONCENTRATIONS FOR SHAKER
TABLE ADSORPTION STUDY
Metal (Source Compound)
Cadmium (Sulfate)
Copper (Sulfate)
Lead (Acetate)
Nickel (Sulfate)
Zinc (Sulfate)
Concentration*, mg/l
40,000
2,000
20,000
20,000
30,000
4,000
200
2,000
2,000
3,000
400
20
200
200
300
4
2
2
3
-------
0 METAL MIGRATION IN SOIL
oncentrations was based on the review of the data on average
ontaminant concentrations found iaSuperfund sites. The pur-
ose of various concentrations of the specified metals during the
daorption shaker analysis was to determine Freundlich and
.angmiur isotherms which allow determinations of compound-
pecific soil/water partition coefficients.
Seven pyrex bottles for each of the specified concentrations
f the five metals were agitated with 100 ml of the metal solution
nd 10 grams of the soil Agitation time ranged from 15 minutes
o 48 hours with samples removed at intervals of 15 rain, 30 min,
hr, 3 hr, 6 hr, 12 hr, 24 hr and 48 hr for analysis. The shaker
able was operated at 180 rpm throughout the analysis to insure
:6mplete mixing of the soil in the metal solution (Figure 1). pH
/alues of the initial metal solution prior to mixture with the soil
tnd pH of each liquid sample from the adsorption analysis were
Figure 1. SHAKER TABLE ADSORPTION STUDY
C.C. CONNICK
11
recorded. Samples removed at the specified times for each metal
and their respective concentrations were filtered using a Vacuum
Pump MUlipore Filter Apparatus and a 0.45-micron filter pad
placed in a sample vial and acidified to a pH of 2 with 1 + 1 HNO,.
Soil samples from the 48-hour time interval for each metal
and its respective concentrations were digested using the Nitric
Acid Digestion Procedure (Standard Methods, 302D, 15th Ed.)
The purpose of the digestion was to determine the maximum
quantity adsorped on the soil following the longest contact
period. Metal content of each sample was determined using a
Perkin Elmer 560 Atomic Absorption Spectrophotometer (AA).
Data from the adsorption analysis using the shaker table were
presented in the form of plots of percentage of contaminants in li-
quid samples versus time. These data were used to obtain the
adsorptive capacity of the soil at a given contaminant concentra-
tion. Plota of concentration adsorbed per unit weight versus
residual concentration were used to obtain adsorption isotherms.
Soil Column Studitt
Column tests were conducted for each of the five metal con-
taminants and a mixture of Cd, Cu, Ni, and Zn to simulate field
contamination and cleanup using the specified chemical counter
measures, under gravity flow conditions. The custom-fabricated
soil columns used in this study were 32-inch (81.28-cm) long clear
plexiglass cylinders with an inside diameter of 2.75 inches (6.985
cm). Both ends of the column were fitted with a plexiglass cap
with 1-inch (2.54-cm) diameter holes. A 2.5-inch (6.35-cm)
diameter, 0.25-inch thick perforated plastic disk was placed at
the base of each column to prevent the loss of soil during the
analysis. The caps were held in place with four nuts attached to
support rods running from column top to column bottom. Teflon
tubes connected to plastic fittings threaded into the end caps
allowed the introduction of aqueous solutions and the collection
of effluent samples. Tubes at the base of the columns were placed
into one-liter plastic containers for the collection of effluent
samples during column rinsing. An aqueous solution contaminant
or treatment rinse was introduced at the top of each column in
-------
12
MKTAL MIGRATION IN Sou.
premeasured aliquots in such a manner as to minimize the distur-
bance of the surface soil structure (figure 2).
Column Packing
A plug of soil weighing 0.73 pounds (331 grams) was brought
to the field moisture content of 11 to 12 percent and added to the
column. It was packed in 2-inch (5.1-cm) lifts using a custom-
made controlled-drop hammer compactor designed to fit inside
the column (Figure 3.)
This procedure was repeated for a total of nine lifts per column
to acheive a soil height of 18-inches (46.72-cm). a total volume of
106.9 cubic inches (1762.3 cc) and a total mass of 6.6 pounds
(2979 grams). Records were maintained for each plug of soil that
was added to %ach column. Soil weight, packing depth, number of
taps required, and compaction data (from the pocket penetro-
meter) were monitored for uniformity. The columns were packed
Figure 2. SOIL COLUMN APPARATUS
13
Figure 3. CONTROLLED-DROP HAMMER COMPACTOR
in this manner to achieve the desired density of 105 to 110 Ibs/cf
(1.68 to 1.76 gm/cc) to simulate original field conditions and the
desired permeability rates of approximately 1.6 X 10"' to 1.0 X
10"' ft/sec (6 X W4 to 3 X 10'4 cm/sec).
Determination of Quantity of Counts-measure
The treatment or cleanup of the contaminated soil was de-
fined as the number of pore volumes of water or water and
countermeasure needed to remove the desired amount of metal.
Successful cleanup was defined as the removal of enough metal to
produce a leachate from the columns which fell below EP toxicity
criteria4. EP Toxicity Concentrations for the heavy metals used
in this study are presented in Table 3. EP toxicity values are 100
times the concentration permitted by drinking water standards.
The pore volume (quantity of water within the pores of a
saturated soil sample) was calculated using the following equation:
pv
WV — 8V
-------
14
METAL MIGRATION IN SOIL
where pv = pore volume (cc); wv = whole volume of soil in col-
umn (cc); and av = solid volume of soil (cc) = (weight of soil added
to column in grama)/(specific gravity in g/cc)1.
The determination of specific gravity of the soil was
calculated following the procedure outlined in Method* of Soil
Analysis* and ASTM D864-58. The pore volume of each of the
packed columns was determined based on the above formula. The
average pore volume of the 14 packed columns was 690 ml and for
the remainder of the study this volume was used as the "treat-
ment" pore volume.
Table 3. EP TOXICITY CONCENTRATIONS
Metal
Concentration, mg/l
1
100
5
1.34
600
Cadmium
Copper
Lead
Nickel
Zinc
Column Contamination
The concentration of contaminant used In the column
analysis was chosen as the maximum concentration used in
shaker table analysis. Two columns were contaminated with each
metal. Two columns packed with soil were used as blanks. No
metal was applied to these columns, but they did receive the
treatment rinses applied to the contaminated columns.
Columns 1 to 10 received 1.915 liters of the metal con-
taminants. The tube at the base of the columns was closed off and
the contaminant poured slowly into each column through the hole
in the cap of each column. The columns were filled to the top with
the metal solution which was allowed to saturate the soil for four
days. Following this period of saturation, the metal con-
taminants were drained from the base of each column into a two-
liter collection container. The columns were then allowed to air
dry for two days to insure complete draining and simulate the
C.O CONNICK
II
drying of a spill which might occur in the field. Samples of th<
drained contaminants (leachate) were analyzed for metal concen
tration using the atomic adsorption spectrophotometer (AA). Th<
pH of the metal contaminant was recorded before and foUowinj
its passage through the soil column. A soil sample was takei
from the surface of each column and digested using the Nitri
Acid Digestion Procedure (SM 302D).
Column Treatment and Cleanup , : .
One column of each contaminant pair received only Up wate
rinses while its sister column received the chemical counter
measures, water plus surfactant (Rinse 2) and water plu
chelating agent (Rinse 6). Columns receiving only tap water wer
rinsed 15 times in pore volume aliquota (690 ml). Columns whicl
received the surfactant and EOT A solution received a total o
eight rinses, one surfactant rinse, one EDTA rinse and six U|
water rinses. Initial and final pH, metal content, and COD wer.
recorded for each rinse.
RESULTS AND DISCUSSION
From the shaker table analysis, plots of adsorbance versu
time were prepared for each concentration of each metal Figure
shows an example of cadmium adsorption. From each plot, th
final adsorbance was estimated and presented as total percent ac
sorbance and total mg metal adsorbed per gram ofsoU as well a
the equivalent (nvmoles) metal adsorbed per granrof soil (Tabl
4). The shaker table results were used to estimate a "minimum
contact time between soil and contaminant to achieve a heavil.
contaminated soil and to determine if the time to reach equilibr,
um is a function of initial contaminating concentration. Data ii
dicated that six hours of agitation achieved maximum adsorptio
values for the contaminant concentration tested, with a longe
time needed for the lower concentrations. The shaker table dat
were also used to generate adsorption isotherms, a graphical pret
entation of the mass of metal adsorbed per gram of soil versus th
residual metal contaminant concentration in the contact solution
Table 4 (showing the format of data generated for each meta
-------
16
MKTAL MIGRATION IN SOIL
lUO -I
f
3
10 •
INIIIM. CAOniur coNUNriuTioh no M/L
it
TIM • MOMS
II
M
Figure 4. CADMIUM ADSORPTION
- SHAKER TEST ANALYSIS
Table 4. SHAKER TEST RESULTS - CADMIUM ANALYSIS
A) 24 Hour TVil — Liquid Sam pit Analytit
Initial Final Rtduette
Cone, Cone, Cone,
No mj/l * mg/l _ mg/l
1 30000 26000 4000
2 2200 1300 900
3 320 176 146
4 26 12 13
B) Soil Samplt Digtttion Analyili
No
1A
2A
3A
4A
Digtittd
Initial Samplt
Cone, Cone,
mg/l ma/I
30000 7000
2200 160
320 160
26 2.26
Soil
Samplt
Man,
3.423
3.667
3.041
3.224
Adiorbanet,
mg/g
133
2.33
2.46
0.041
in
Adiorbanct,
mg/g
Equiu
Cone,
Rtmovtd,
mg/l
13300
233
246
4.07
40
9
1.46
0.13
Equiu
Cone,
Rtmaining,
mg/l
16700
1967
74
20.9
Adiorbanc€,
ptrctnt
13
41
46
62
Ad$orbanct,
ptrctnt
44
11
23
84
C.C. CONNICK
Table 4. SHAKER TEST RESULTS - CADMIUM AN ALYS1
(CONTINUED)
O Summary
Conctntrotion
Rtmaining,
No mgTI mM/l
1 26000
2 1300
3 176
4 12
1A 16700
2A 1967
3A 74
4A 21
231.3
11.6
1.56
0.107
148.6
17.6
0.668
0.187
Log-Cone
Rtmaining,
mM/l
2.36
1.06
0.19
-0.97
2.17
1.24
-0.181
-0.731
Adiorbanct,
ms't.
40
9.0
1.46
0.13
133
2.33
2.46
0.041
mM/l
0.356
o.oao
0.013
0.0012
1.183
0.021
0.022
0.00036
Log
Abs.
mM/l
-0.44
-1.10
-1.89
-2.94
0.01
-1.68
-1.66
-3.44
summarizes the data required for isotherm generation based on 1
quid sample analysis. Part B presents the results of the digeste
soil samples. Part C is a representation of data in Part A and I
expressed in units necessary for plotting the two types <
isotherms.
A comparison of the percent adsorption columns in Part.
and Part B of the summary tables showed that the digested so
samples consistently varied from the corresponding 1
quid/leachate samples. The soil sample analysis consistently ii
dicated a lower value for total metal adsorbed than did tt
filtrate analysis. An explanation for this trend is that the sc
digestion process does not remove all the metal adsorb*
therefore, total adsorbance is underestimated by the soil samp
analysis.
The isotherms developed were prepared using the Freundlic
(Figure 5) and Langmiur equations (Figure 6). The Langmiur a
sorption isotherm equation1 can be derived from simple ion e
change considerations, assuming that only one type of adsor
tion site is involved and that only simple heavy-metal catioi
take part in the exchange reaction (1-site model). The Freundlic
isotherm1 equation can be interpreted as an approximate deacri
tion of ion exchange involving one or more types of heavy met
cations and one or more types of adsorption sites (2-site mode!
-------
18
MKTAL MIGRATION IN SOIL
I*-' -
totiim HIM
— .totumi mini moiHioi evin.
Q -MIIICUI ititui tiioiMioi
O -iititviii IIIIIH tiioiMiot
t»«tit
turn
-ton* i»*n« mini)
I
itoimin I
too tot. • o.oii i lot ton. -i.ii
1 • O.lll • • I I
1
ll-l
!»• !•»
coiciifMtioi
!•'
Figure 5. FREUNDUCH ISOTHERM - CADMIUM ADSORPTION
•> .
s
*•
!
>»«ICI UIU
t • U«Ull (Mm mi till
* • ton i until mum
itotmin
lit • I/IUI.4 I l/CUC • 11.111
I • l.»ll • • I
MIINUI) MIIICUI moiMltl
• l/t« • O.IIU KUllMllS/lllll
»*•»
10'
10-1 ll-l 10° !•'
utiiuu CUCIIIIMIII . i/imumoiit/miii
Figure 6. LANGMUIR ISOTHERM - CADMIUM ADSORPTION
O.C. CONNICK
19
From the plots and their corresponding correlation coeffi-
cients, it can be seen that for all five metals the Freundlich equa-
tion corresponds well with the adsorption data generated in the
study of this soil and contaminant system. The Langmiur equation
corresponds well only with data generated from the adsorption
behavior of Pb, Ni, and Zn.
An explanation for the correlation of the data to the
Langmiur equation for only Pb, Ni, and Zn is that these ions are
not complexing in solution to the same degree as Cd and Cu and
they are adsorbing to the soil based on the mono-layer theory
with more uniform bonding strengths. Excessive complexing of
Cd and Cu in solution would cause adsorption on the soil surface
to be less uniform with varying strengths of attachment and.
therefore, be more accurately described by the Freundlich theory.
Support of this hypothesis is found in a study by B.E. BlonY
which determined that in the presence of a relatively large excess
of calcium or potassium the formation of CdCl* enabled the Cd lo
be more easily bound to the soil system due to the preference of
univalent ions over multivalent ions. The soil used by Blom was
similar in type to the Typic Hapludult soil type used in this
study, although the calcium content of the Typic Hapludult soil
was not determined. It can be hypothesized (but not proven) that
Cd was adsorbed as CdCl* in this study. During AA analysis, the
flame appeared red and yellow in color, indicating the presence ol
significant levels of calcium and sodium respectively, in the liquic
sample.
Considering the theoretical aspects of the two isotherm type*:
and the better agreement of the Freundlich equation to the dau
generated, the Freundlich isotherm was selected for use durint
soil column evaluation. The isotherm plots also contain a dotte<
line which represents a family of potential adsorbance versu.
residual concentration end points. The line was formed by select
ing a series of arbitrary final concentrations and, using th<
change from the initial concentration, calculating the unique ad
sorbance that could occur. The predicted adsorbance of the meta
in the column at the initial contaminant concentration applied i
designated at the intersection of the isotherm line by the squar
-------
20
METAL MIGRATION IN SOIL
symbol. The actual adsorbance measured for the metal by the soil
column is designated by the hexagon symbol. The optimum con-
tamination obtained in the columns was consistently lower than
that obtained in shaker tests. This is due to the greater
contaminant-to-soil ratio in the shaker test and also the improved
soil-liquid contact achieved during the agitation process, as com-
pared to the gravity flow conditions in the soil column.
Adsorption of the metal contaminants achieved by the soil
column were: Cd, 0.083 mM/g; Cu, 0.023 mM/g; Pb, 0.030 mM/g;
Ni. 0.073 mMlf. and Zn, 0.132 mM/g. These values are about 70
percent of the values predicted to be adsorbed based on the
shaker test analysis.
Soil Treatment and Decontamination
Table 5^ presents the percent removals of the metal con-
taminants by each treatment method. The tap/surfactant/EDTA
8-rinse treatment was more effective than the 16 tap water rinse
in all cases except lead. An increase of metal concentration in the
leachate following the application of the EDTA/buffer solution
indicates that EDTA is responsible for the increased removals in
these columns. Metal concentrations in surfactant leachate are
equivalent or less than the concentrations in the leachate of the
corresponding tap water rinse from its sister column, indicating
that the surfactant was ineffective in desorbing heavy metals
from soil systems. This is shown in Figures 7 and 8. The shape of
the removal curves indicates the majority of the metal is removed
in the first four to five rinses. The column receiving the EDTA
»
Table 6. TOTAL PERCENT METAL REMOVED
Tap Water Only,
15 Rinses
87
44
74
87
88
Tap/Surfactant/EDTA,
8 Rinses
100
82
63
94
93
C.C. CUNNICK
21
2
ul
u
o
10'-
A - INIIIAI CONCCNUATION
• - IAf WAUH RINSC
T
4 6 II 10
RINSE VOLUME * LITERS
12
Figure 7. CADMIUM COLUMN TEST - WATER RINSE
-------
22
METAL MIGRATION IN SOIL
C.C. CONNICK
•23
1"'
o
u
10'-
10'
• • INITIAL CONCENTRATION
A • TAP WATER RINSE
• - SURFACTANT RINSE
A- EDTA RINSt
14
RINSE VOLUME * LITERS
Figure 8. CADMIUM COLUMN TEST TAP
WATER/SURFACTANT/EDTA RINSE
solution experienced a marked decrease in permeability. This in-
dicates that the increase of the system pH due to the addition of
the EDTA buffer mixture ia causing the precipitation of the
metals, presumably as hydroxides. (Precipitants were also
observed in the leachate from the EDTA treated columns.)
EP Toxicity analysis performed in the soil following the
treatment rinses indicated that five pore volumes of tap water (or
tap water plus surfactant) were successful in reducing the metal
content of the soil contaminated by zinc, copper and lead to
within EP Toxicity limits, but only with the application of the
EDTA/buffer rinse was the soil contaminated with cadmium and
nickel reduced to levels within EP Toxicity limits. Using rain
data for the area of the soil origin, the pore volume of rinse ap-
plied was equated to 0.34 years of rain.
CONCLUSIONS
Results of this study indicate that in-situ treatment is a
viable solution for the removal of metals Cd, Cu, Pb, Ni and Zn
from contaminated soil. Care must be taken when extrapolating
the results obtained in these tests to other situations as there are
many variables which influence detoxification.
The use of the surfactant mixture as a rinse treatment for the
removal of heavy metals proved ineffective in this soil system.
The surfactant solution provided removal efficiencies com-
parable, but not superior to the tap water alone rinses.
EDTA proved effective in desorbing the metal cations from the
soil system. The columns which received only eight treatment
rinses, one of which included EDTA, indicated greater removals
of contaminant than the columns which received 15 rinses of tap
water alone. The use of EDTA appears to flush the metal from
the soil as observed from the very high metal content of the
EDTA rinse leachate in comparision to the previous tap water
rinse leachate from the same column.
A decrease in the permeability of the column is observed when
a large volume of treatment rinses is applied. This occurs in part
-------
24
METAL MIGRATION IN SOIL
because the fines are washed to the base of the column where they
accumulated and inhibit the flow. The application of the
EOT A/buffer solution increases the system pH to 9 to 10 which
induces the formation of precipitates within the column, further
decreasing the column permeability and potentially clogging it.
Maximum adsorbance of the metal by the soil under shaker
table anaylsis was obtained within the first three to six hours for
contaminant concentrations greater than approximately 20,000
mg/1. The required contact time increased to six to twelve hours
for contaminant concentrations between 20,000 mg/1 and 20 mg/1.
At contaminant concentrations less than 20 mg/1, the time to
equilibrium was as long as 18 hours.
The Freundlich isotherm appeared to be applicable for the
description of the adsorption behavior of all the soil/metal
systems in this study. This implies that the adsorptive sites in
the soil system are heterogeneous and a possible interaction
among particles in the adsorbed phase may be occurring. The
energy of this adsorption decreases logarithmically as the frac-
tion of surface covered Increases.
The Langmiur isotherm only successfully described the adsorb-
tive behavior of Pb, Ni, and Zn. The Langmiur adsorption equa-
tion is derived from simple ion exchange considerations, assum-
ing that only one type of adsorption site is involved and that only
simple heavy metal cations take part in the exchange reaction.
The fit of Pb, Ni and Zn adsorption results to the Langmiur equa-
tion may indicate that these ions are not completing in solution
to the same degree as Cd and Cu and that they are adsorbing to
the soil based on the mono-layer theory with more uniform bond-
ing strengths.
REFERENCES
1. Ellis, W.D. and J.R. Payne, "Chemical Countermeasures For
ln-Situ Treatment of Hazardous Material Releases", USEPA
Contract No. 68-01-3113, Oil and Hazardous Materials Spills
Branch, Edison, NJ, 1983.
C.C. CONNICK
25
2. "Final Report: Underground Movement of Gasoline - in
Ground Water and Enhanced Recovery by Surfactants",
Texas Research Institute, 1979.
3. Drake, E. et at, "A Feasibility Study of Response Techniques
For Discharges of Hazardous Chemicals that Disperse
Through the Water Column", US Dept. of Transportation,
Report No. CG-D-16-77, 1976.
4. US EPA, Federal Register, Vol. 45, No. 98, Rules and Regula-
tions, Appendix II, p. 33127.
6. Shepard, J., Su&marine Geology, Harper and Row Publishers.
NY. 1973.
6. Black, C.A., ed., Methods of Soil Analysis, Chemical and
Microbiological Properties, Agronomy No. 9. Part 2, 1965.
7. Metcalf and Eddy, Inc., Wasiewater Engineering: Collection
Treatment and Disposal, McGraw Hill, NY, 1972.
8. Blom, B.E., "Sorption of Cadmium by Soils", National Sci-
ence Foundation, June, 1974.
-------
VOL. 19. NO. 1
MAY 1985
OF THE NEW ENGLAND WATER
POLLUTION CONTROL ASSOCIATION
-------
' UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
^ OFFICE OF RESEARCH AND DEVELOPMENT
* "*" HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
CINCINNATI. OHIO 45268
REPLY TO:
Releases Control Branch
U.S. EPA
Woodbridge Avenue
Edison, New Jersey 08837
DATE: December 19, 1985
SUBJECT: Draft Research Project Plan: Removing Lead with EDTA Chelating
Agent from Contaminated Soil at the Michael Battery Company,
Bettendorf, Iowa
^^^^
? tttduJ-tt
[
FROM: Richard P. Traver, Staff Engineer
Releases Control Branch, LPCD, HWERL
TO: James R. MacDonald, Environmental Engineer
Site Investigation Section, Emergency Planning
and Response Branch, ESD - Region VII
THRU: Frank 0. Freestone, Chief
-Technology Evaluation Staff. RGB, LPCD, HWERL
This 1s 1n response, to your request to Ira Milder for an estimate to use
the EPA Mobile Soils Washing System at an Immediate Removal Action at the former
Michael Battery Company, Bettendorf, Iowa.
Attached Is a Research Project Plan for your review and comment. The pro-
posed project consists of the following four phases:
Phase I Preliminary Laboratory Feasibility Study for Evaluating
Potential Use of EDTA Chelating Agent for Removing Lead
from Michael Battery Soil
Phase II Laboratory Feasibility Study for Evaluating Removal of
Chelated Lead from EDTA Solution, and Preliminary Process
Design
Phase III .....i Full Scale Pilot Study
^Phase IV Field Demonstration
The objective of the proposed project 1s the development of operating proto-
cols and cost estimating procedures that could be used by Region VII to engage
the services of a commercial cleanup company or those of an existing EPA cleanup
contractor. We are flexible jregarding the extent to wMch this plan needs to be
implemented and we stand ready to discuss any modifications you might suggest to
suit your purposes.
-------
Page 1 of 14
DRAFT RESEARCH PROJECT PLAN
REMOVING LEAD WITH EDTA CHELATING AGENT FROM SOIL CONTAMINATED
WITH LEAD IN BETTENDORF, IOWA
OPTION B: On-Site Treatment/Soil Washing"
December 19, 1985
OBJECTIVE
The overall objective of this project is the development of engineering speci-
fications, cost estimates, and operating protocols for use by Region VII to
evaluate the alternative of soils washing for treatment of lead-contaminated
soil, defined as Option B under the Region VII Action Memorandum of 8/28/85.
If this alternative is subsequently implemented for a full-scale cleanup, the
treatment of substantial quantities of contaminated material at the Michael
Battery Company could be pursued under either a separate contract with a haz-
ardous material cleanup company or under the appropriate EPA Emergency Response
Cleanup Services contract.
SUMMARY AND LIMITATION OF SCOPE
The Hazardous Waste Engineering Research Laboratory's Releases Control Branch
(RCB) in Edison, NJ, has been asked by EPA Region VII to evaluate the feasi-
bility of removing lead from contaminated soils at the Michael Battery Company,
located in Bettendorf, IA. Previous work by RCB and others!-? has shown that
lead may be removed from some soils using EDTA as a chelatlng agent in an aqueous
solution to solubilize the lead, with subsequent removal and concentration of
the lead from solution. This Research Plan addresses a multi-phase engineering
feasibility study only, and does not explore other aspects of the lead-in-soil
problem at the Bettendorf Site such as: a detailed "extent of contamination"
survey, or means of solving the contamination problem other than by processing
the soils. It should be further noted that removal and treatment of contami-
nated soils may be limited to collected dust/soil from the main building, the
approximate 535 cubic yards of soil from site drainage ditches, and the approxi-
mate 300 cubic yards from around the building.
BACKGROUND
1. Site Description - The information Pertaining to the Site Description 1s
Basically a Summary of Information Provided In James
R. McDonald's Draft Action Memo of 8/28/85.
•
The Michael Battery Company operated a battery manufacturing and recycling busi-
ness in Bettendorf, Iowa, from October 1979 thru June 1983. Michael Battery
Company leased the 0.6 acre site and a 5.000-square-fbot metal building from the
present deeded owner, Jessee Roofing and Painting Company.« The site is located
-------
Page 2 of 14
in an Industrial area of Bettendorf within the floodplaln of the Mississippi
River which 1s located approximately one half mile south. Surface run-off from
the battery manufacturing operation has contaminated portions of the adjacent
property.
The subsurface geologic characteristics are the bedrock, which Is approxi-
mately 10-15 feet below the surface, and the 0-10 feet of unconsolidated
sediments which are alluvial silts, clays and fine sands. The upper surface has
received crushed limestone to level the surface and to serve as footings for the
building. The hydrology 1n the area consists of the surface water, groundwater
in the unconsolidated alluvial deposits, and the deep bedrock aquifer. The sur-
face water and storm runoff 1s largely contained 1n the Industrialized area
around the site, and 1s eventually diverted to the Mississippi River. Local
drainage from the Michael Battery Company site 1s to the south, over the adjacent
Rogan Scales property, Into a railroad ditch draining west. The runoff In the
ditch ponds and percolates Into the substrata. The Davenport Water Company has
water Intakes on the Mississippi River, 3.75 miles downstream from the site.
A. Quantity and Types of Substances Present
In February 1982, In response to a report of Illegal dumping of sulfuric add at
the site, preliminary soil and surface water samples were collected. These pre-
liminary samples Identified heavy metal contamination of both soils and surface
waters. Followup sampling conducted by EPA on July 8, 1982, detected lead con-
centrations In soil up to 5,200 ppm. In response to these sampling efforts, an
expanded EPA field Investigation was conducted in April 1984. On site monitoring
wells were Installed In June 1984. The results of the above Investigations have
Indicated that significant lead contamination exists on site. The areas of lead
contamination have been divided Into four subareas: (1) metal building; (2)
western dralnageway; (3) sump area and eastern drainage way; and (4) storage areas
around the building. The concentrations of lead and the volume of lead-contami-
nated soil/dust In each area are summarized below:
1. Interior Dust/Soil Samples
Concentrations of lead In dirt and dust collected from Inside the 5,000 square
foot metal building, ranged from 4% to 5% for EPA samples collected In June, 1984,
and from 17% to 33% for the National Institute of Occupational Safety and Health
(NIOSH) samples collected In November, 1984. Dust has settled throughout the :
building on walls, roof and floors; with notable concentrations on the roof trus-
ses and cross member supports for the walls and loft area. An estimate of the
quantity of dust/soil that could be vacuumed from the building would be approxi-
mately ten 55-gallon drums.
-------
Page 3 of 14
2. Western Drainage Samples
The western drainage from the Michael Battery Company site is directed south from
the blacktop around the building, southwest across Rogan Scales property, flow-
ing west in the drainage ditch to the southwest corner of the lot. Concentrations
of lead in this drainageway varied from 65 ppm to 31,700 ppm and averaged over
4,000 ppm. Soil samples were collected to a depth of 12 inches; if subsequent
soil sampling below the 12 inch depth reveals further lead contamination, quanti-
ties of soil to be processed could be dramatically increased. The length of the
western drainage ditch is approximately i50 feet. The surface area of the sur-
rounding contaminated drainage area is approximately 13,000 square feet. The es-
timated volume of contaminated soil, assuming an average depth of one foot, Is
480 cubic yards.
3. Sump and Eastern Drainage Samples
Drainage from the sump at the loading bay at the east end of the building was
pumped onto the shoulder of Devils Glen Road where it drained south to the drain-
age ditch beside the railroad and then drained west. Concentrations of lead in
this eastern drainage varied from 94 ppm to 9,600 pom and averaged 4,600 ppm.
The length of the Devils Glen Road shoulder from the sump to the drainage ditch
south is approximately 150 feet. The surface area of the surrounding contami-
nated area is estimated to be 1,500 square feet. The estimated volume of con-
taminated soil, assuming an average depth of one foot. Is 55 cubic yards.
4. .Storage Areas Around the Building
The highest concentration of lead found (102,000 ppm) was located outside the
backdoor where Michael Battery Company sorted lead. Other storage areas In-
cluded an area north of the blacktop adjacent to the auto parts warehouse;
concentrations range from 74 ppm to 5,300 ppm and average 1,000 ppm. A second
storage area 1s located to the west of the blacktop area; concentrations range
from 210 ppm to 2,300 ppm and average 770 ppm.
Sweeping of soil/dust from the asphalt surfaces would result 1n an estimated
five 55-gallon drums of material. The unsurfaced area on the site with potential
storage, not Including the western drainage-way, 1s estimated at 8,000 square
feet. The estimated soil volume assuming a one foot depth 1s 300 cubic yards.
5. Surface Water and Groundwater Analysis
Previous sampling efforts have documented moderate lead contamination of sur-
face drainage waters (96 ppm). No significant groundwater contamination has
been detected, however. ,.
-------
Page 4 of 14
REMEDIAL ACTION
Based on COC advisories, a clean-up level of 1,000 ppm lead In soil 1s recom-
mended. Soils which fall the E.P. Toxicity Test for lead, it is proposed, would
be handled as hazardous waste and transported to a licensed hazardous waste site
for disposal. Soils which do not fall the E.P. Toxicity Test, but which contain
lead in concentrations above 1,000 ppm, would be disposed of at a state approved
landfill.
Region VII1s Remedial Action Plan calls for cleaning the Interior of the building,
including the roofing, trusses, walls and floor of all dirt/dust. This would be
accomplished vacuuming with a High Efficiency Particulate Air (HEPA) filter fol-
lowed by pressurized water and detergent wash. The use of a chelate solution of
EDTA should be considered for the wash solution. This would allow for the col-
lected wash solution to be treated and recycled. The concentrated lead would be
either disposed of as a hazardous material, or could be sold to a metal refinery
to be reprocessed.
Region VII has proposed three action options: Option A - Dig and Haul. Option B •
Soils Washing, and Option C - On-S1te Chemical Fixation and Capping. Options A
and C are briefly summarized with a detailed explanation of Option B following.
OPTION A - DIG A HAUL
Region VII's Option A calls for excavation and off-site disposal of soil and
materials having lead concentrations In excess of 1,000 ppm. It Is estimated
that the volume of soil and lead dust would approach 900 cubic yards. It Is
presumed that 75% of this material (675 cubic yards) would not fall E.P. Toxic-
ity criteria for lead (_< 5 mg/1 In leachate) and would be suitable for disposal
In a state approved landfill. The remaining material, approximately 225 cubic
yards, 1s expected to fall the E.P. Toxicity Test and would be handled as a
hazardous waste. Disposal of this material would be carried out at an approved
Resource Conservation and Recovery Act (RCRA) disposal site. Cost estimates are
approximately $214 K if only a portion of the material must be disposed of at a
RCRA approved site. If all material must be taken to a RCRA site, the cost
estimate 1s $463 K. It should be noted that this Option does not eliminate the
contamination problem, but merely relocates it until such time that the RCRA
site material would have to be treated.
-------
Page 5 of 14
OPTION C - ON-SITE CHEMICAL FIXATION & CAPPING
Region VII has proposed a commercial chemical fixation process for on-site en-
capsulation. This approach would stabilize the contaminated soil through a pro-
prietary fixation process. The fixated soil would be replaced on-site and then
covered with a clean soils cap.
With the approximate 1125 cubic yards of material, the rough cost for on-site
chemical fixation is $100/cubic yard, or $112 K. An additional estimated $60 K
would be needed to install a clean soil cover.
No laboratory analysis has been performed evaluating the effectiveness of chemi-
cal fixation with the site specific Bettendorf Soil. A thorough bench-scale
study would be necessary in order to determine if the fixated soil would pass the
E.P. Toxidty Test for lead. It 1s also uncertain 1f the site would be usable
by the owners following the chemical fixation process.
OPTION B - SOILS "WASHING" USING EDTA
The soil decontamination process first used by RGB was at a lead-storage type
battery reclamation site in Leeds, Alabama, in 1984, at the request of Region IV.
This involved the use of a prototype "Soils Washing System" for application of
13% EDTA solution to lead contaminated soil. The lead-in-soil concentration was
reduced from 50,000 to less than 100 ppm. EDTA or ethylenedlaminetetraacetlc
acid, d1sodium or tetrasodiurn salt, 1s a commercially produced chelating agent
that. In an aqueous solution, can complex with lead to produce a water soluble
chelate. (See attached Project Summary and Fact Sheet for more detail on the de-
sign and operation of the EPA prototype Soils Washing System.)
Region V has subsequently evaluated various treatment processes for the cleanup
of a battery reclamation site at Woodville, Wisconsin. After examining the ORD
experience and conducting laboratory tests. Region V also chose washing with
EDTA as the best approach. A pilot-scale system Is now being implemented 1n the
field for treatment of battery casings.
A literature search and laboratory study, performed by JRB Associates under the
Hazardous Waste Engineering Research Laboratory's "Chemical Counter-measures Pro-
gram," also established the use of EDTA as the likely technology for the removal
of a variety of heavy metals from soils. The study noted that lead-in-soil wash-
ing with EOTA must be evaluated on a site-specific basis. An Independent study
conducted by Northeastern University, in cooperation with RCB, corroborated these
findings.
-------
A significant concern at this time is not knowing the percent of EOTA that can
be regenerated for reuse. Chelate that cannot be regenerated causes a double
expense: one, it must be replaced; and two, it must be disposed of in a safe
manner. Apparently, iron blocks the regeneration process. In the Alabama work,
iron is listed at 2,100 mg/kg and apparently, although not specifically noted,
the EOTA regeneration was only 5% through sulfide precipitation. The iron
content of the alluvial silts, clays and fine sand at the Bettendorf site is
approximately 1-2% in the form of hematite, magnatite and ilmenite. Or. Anderson
of the Geology Department of Augustana College (across the river from Bettendorf)
indicated that the Mississippi River received a "slug" of Iron from Wisconsin in
that area in the last ice age. If this 1s the case, there 1s, on the average,
three times as much Iron as there Is lead. This would be expected to cause sig-
nificant problems 1n regenerating the lead if the chelate can remove the iron
from these mineral structures. For this reason, a thorough comparison on a lab-
oratory scale basis needs to be run on both sulfide precipitation and electrodi-
alysis as means for EOTA regeneration.
SCOPE OF WORK
The response activities proposed by RGB for dealing with the lead problem in
Bettendorf consist of four phases. -Phase I will be a laboratory feasibility
study to determine if EOTA offers a reasonable chance of success for removing
lead from the type of soil matrix present at the affected Michael Battery site.
Phase II will also be a laboratory-scale engineering study geared to determine
the optimum approach and conditions"for removing chelated lead from solution
and regenerating EDTA for recycling purposes. If these phases are successful,
Phase III will be a full-scale pilot study involving approximately 100 drums
of lead-contaminated soil being shipped to Edison, New Jersey, where the QRO
Soils Washing System will be used to evaluate process performance, operating
costs, and system capacity. Additionally, Phase III will provide for any
necessary permit applications. Including a del1sting petition. Phase IV will
be a field activation with the Soils Washing System at the Michael Battery
site to demonstrate the field capability of the technology and to develop oper-
ating protocols for use by Region VII in acquiring contracted cleanup services,
if so desired.
Phase I Preliminary Laboratory Feasibility Study for Evaluating Potential
Use of EDTA Chelatlng Agent for Removing Lead from Bettendorf Soil
The objective of Phase I 1s to establish the optimum concentration of EOTA in
solution for lead removal and the percent lead reduction in the Bettendorf soil
-------
rage / or
A 2-4 kg sample consisting of a homogeneous blend of "Michael Bettendorf Site
Soil" contaminated with 2,000-330,000 ppra lead will be obtained by Region VII
by compositing samples from several hot spots. Region VII will attempt to make
this single composite sample as representative as practicable of the soils in
the hot spots in terms of organic content, soil particle size, and potentially
interfering elements such as Zn, Ba, Ti, Cr. and Fe.
It should be noted that this preliminary study is a single sample study only—
the results must, therefore, be interpreted xith great caution.Soil variabil-
ity among the hot spots could easily be obscured in the blending process needed
to obtain the single "representative" sample. Phase II will Include samples
from a greater number of locations such that an analysis of the variability of
key parameters of the soils to be treated can be made. ~
The single sample will be "washed" with EDTA solution In the laboratory to deter-
mine the effectiveness of the EDTA chelating process. Ten gram (lOg) soil por-
tions will be agitated on a "shaker table" for 30 minutes with one hundred mini-
liter (100 ml) volumes of the following percentages of EDTA (disodium salt) in
water:
0 (blank); 1.0; 2.5; 6.5; 13.0; and 25.0
Analyses will then be performed to determine the amount of lead removed by EDTA
washing and lead remaining on treated soil.
An EP Toxldty Test (40 CFR 261.24) and a qualitative analysis for all metals
present in the Bettendorf soil blend will also be performed to determine some
of the soil's characteristics.
The QA/QC program for this Phase I study will have the single sample limitation
as noted above, and will Include the following:
[a] The soil washing and analyses procedures will be performed in duplicate.
[b] At least three replicate portions of the original Bettendorf soil blend
will be analyzed to assure homogeneity.
[c] "Lead 1n Soil" analyses will be. performed using both X-Ray Fluorescence and
Acid Digestion methods.
[d] Analyses performance will be evaluated using "QA Audits" with primary em-
. phasis on Performance Evaluation Audits.
-------
Page 8 of 14
A reduction of lead content 1n soil to approximately 1,000 ppm is currently con-
sidered successful for Phase I. If unsuccessful, due to the possible presence
of interfering compounds (e.g., iron) that limit the performance of EDTA, a more
intensive laboratory effort (not fully described in this Plan) may be necessary.
This subsequent effort would seek to define alternative chelating agents or
entirely different treatment processes. If successful and an adequate reduction
of the soil lead level is achieved with EDTA, Phase II will be implemented.
It should be noted that some residual EDTA will remain on the treated soils along
with residual lead (and probably other residual substances). There is a possi-
bility that the residual EOTA could cause the residual lead to have a greater
environmental mobility than that experienced by an equivalent concentration of
lead prior to the treatment process, or the EDTA may, itself, pose some type of
toxiclty problem. While the reported 1059 of EDTA 1s 2 g/kg (rats, orally), and
toxicity does not appear to be an obvious problem, these aspects of the use of
EDTA will be investigated on a preliminary level during this Initial laboratory
study phase. Assistance from other ORD offices may be needed for answers to
these questions.
Time Frame ... 15-30 days from receipt of "representative" sample.
Cost $10,000 - $15,000
Product Letter report on the preliminary feasibility of EDTA extraction.
Phase II Laboratory Feasibility Study for Evaluating Removal of Chelated Lead*
from EDTA Solution, and Preliminary Process Design
The objective of this phase 1s to establish the optimum treatment process for the
recovery of'lead and EDTA from the "soil wash" solution and to prepare preliminary
engineering process specifications, a detailed cost estimate, a test plan, and a
schedule for Phase III.
The EDTA recovery process used by ORD at Leeds, Alabama, reacted sodium sulflde
with the EDTA-lead chelate to form a lead sulflde precipitate that was dewatered
and disposed of at a smelter. Subsequent acidification of the remaining EDTA
solution enabled substantial recycling of EDTA. An alternative treatment process
for the removal of lead from solution 1s based on electrolytic reduction and may
be potentially more cost-effective than the use of sodium sulflde. Evaluation of
final disposal or reclamation of the EDTA (e.g., solidification for storage) will
be pursued.
-------
Vag'g VOT
Additionally, further testing with a selected concentration of EDTA on several sep-
arate "representative" samples from hot spots (the inverse program of Phase I) will
be performed to determine if variability of soil parameters will cause unacceptable
treatment system performance changes among the various soils to be treated. Each
soil sample will be analyzed for particle size, organic content, presence of other
metals or other interfering compounds, and other parameters that could affect per-
formance of either the EDTA extraction or the recycle of the EDTA. This testing
will be performed with QA/QC similar to that in Phase I to assure reliability and
reproduclblHty of data. The samples will be obtained through coordination with
Region VII.
At this time, it will be necessary to assure that the soil samples received are
reasonably representative of those expected in the field. Subsequent project ele-
ments (Phases III and IV) are considerably more expensive than these laboratory
phases and rely heavily upon the precision and accuracy of the laboratory data.
Once an EDTA recovery process 1s identified, the necessary process equipment for
executing the entire treatment (lead removal with EDTA recycle) at pilot scale
(Phase III) must be Identified and-sources sought for needed equipment not now on
hand (e.g., dewatering equipment for lead sludge or electrolytic lead removal cells)
Finally, a detailed cost estimate, testing protocol. Including a Sampling and
Analysis Plan and a Quality Assurance Project Plan, and schedule for Phase III will
be prepared.
Time Frame ... 30-60 days from receipt of authorization to proceed.
Cost Laboratory work: $10,000-$!5,000
Detailed Engineering Planning: $50,000-$!35,000
Product Letter report providing the results of Phase II and detailed
planning Information for Phase III as noted above.
Phase III Full Scale Pilot Study
The objective of this phase 1s to obtain engineering Information on the unit cost,
capacity, personnel requirements, and treatment effectiveness of lead removal using
EDTA 1n the EPA soils washing system, and to provide preliminary planning Infor-
mation for Phase IV.
The study will simulate a field activation using the full-scale prototypical equip-
ment* 1n Edison, New Jersey. Equipment needed for the treatment process but not
currently on hand will be acquired or leased, whichever 1s more favorable. The
buy/lease decision will be made during Phase II, such that the estimate for Phase
III is as accurate as possible.
-------
The tests will Involve the following sequence of activities:
1. EPA and contractor personnel involved with the proposed tests will be provided
with operator training, safety training, and medical monitoring as appropriate.
2. Equipment will be set up indoors in a suitable area where the testing can be
conducted safely and in an environmentally suitable manner.
3. Initial process shakedown will be conducted using clean soils to assure that
all elements of the process function properly Individually and together. Such
normal operating activities as determining pump capacities and flow balances
among the various unit processes must be performed carefully and on clean
material. During this activity, minor process adjustments will be made to
assure appropriate system function in the absence of contaminants or treatment
chemicals.
4. Clean soil of a type reasonably similar to the Bettendorf soil will be Inten-
tionally contaminated with lead known to be In similar form and concentration
as the lead from the Michael Battery site and controlled-condition tests will
be performed, first at laboratory scale, then at pilot scale to assure that the
treatment process Is operating properly. (This 1s done to reduce the amount.of
Bettendorf soil that must be transported to Edison for the shakedown portion of
the tests as opposed to-the portion of the tests Intended for data gathering.)
This activity will assure that the treatment chemistry 1s operating properly
and that such steps as EDTA addition, addition of other treatment agents, and/
or removal/recycle of the EDTA are functioning properly.
5. Approximately 100 - 200 (55 gal.) drums of lead-contaminated soil will be ob-
tained from Region VII and used in a set of tests (probably three or four
"runs") designed to provide capacity and performance Information. The samples
contained in the drums must be "representative" to the satisfaction of the EPA
Office of Solid Waste such that the data resulting from the treatment test can
be submitted in a del 1sting petition, as noted below. Variables for the test
will Include soil feed rates, EDTA concentration, recycle system data, and other
system operating parameters. Measurements will Include Initial lead concen-
tration, final lead concentration, lead concentration In produced sludge, feed
rates, EDTA recycle effectiveness (EDTA use rate), and other chemical use rates.
Also, the number and training levels of the personnel needed for operating the
process will be determined. The. goal of these tests is to Identify the most
cost-effective treatment conditions, requiring the minimum personnel, at the
.greatest possible capacity.
-------
6. After the tests, remaining soil must be disposed of. Some soils, by design,
will not have been adequately treated and may have to be recycled to reach the
design treatment level. There is always an outside possibility that all of the
soil will fall short of the treatment goal. These soils will be either trans-
ported back to Bettendorf or sent to a hazardous waste disposal facility.
Specific arrangements for the disposition of untreated/inadequately treated
soils must be made and agreed to by all principals prior to the transportation
of the contaminated soils from Bettendorf to Edison, and should be addressed in
the Plan for Phase III prepared during Phase II.
7. The equipment and test area must be decontaminated and the decontaminating
solutions disposed of in a suitable manner as noted above for the soils.
8. The test equipment must be disassembled and returned to storage or prepared for
shipment to the field.
Concurrent with these tests, necessary permitting documentation associated with
Phase IV (and also appropriate to a full field activiation using the same process)
will be prepared. As noted above, this will Include State and Federal requirements
and will probably Include a delisting petition. Data from the pilot-scale tests
will be used in the delisting petition to demonstrate that the treated soil 1s
"nonhazardous" to the satisfaction of OSW.
Additionally, during and following these tests, preliminary planning will be con-
ducted for a field activation using the EPA prototype Soils Washing System. This
planning will Include all of the necessary logistical elements and preparations
for operating the system 1n the field for an extended period. However, because
this planning Is a significant effort, a detailed plan will not be conducted
until authorization to proceed with Phase IV is received.
Time Frame ... 3-6 Months from authorization to proceed
Cost $300,000 - $700.000
Product Interim report providing data, detailed estimates and preliminary
plans for Phase IV
Note: This Interim report will contain sufficient data for the specification of a
field operation by sources other than ORO. Therefore, Phase IV is designed
• to be an optional phase.
-------
Phase IV Field Demonstration
The objective of this phase will be to determine field-related variations to the
unit costs, lead removal performance and system reliability determined during the
pilot scale tests in Phase III. The resulting information from this phase would
be used by Region VII to specify contracted cleanup efforts using commercially
available equipment and personnel.
Pilot scale tests conducted during Phase III will be done under carefully con-
trolled conditions at Edison, NJ, with a maximum of nearby shop and logistical
support to help overcome unanticipated difficulties. Running changes can be made
relatively easily and cheaply because of the availability of extra personnel when
needed and a strong base of equipment testing capabilities. Field operations, by
comparison, require substantial advance planning to assure that the operation pro-
ceeds smoothly from mobilization through startup and Into reliable continuous
operations. Omissions or errors in the planning process, as well as uncontrolla-
ble variations such as severe weather, quickly translate Into lost time and extra
costs. Field tests are, therefore, expensive, demand the most from advance plan-
ning and preparations, and require contingencies in the planning process relative
to both time and costs. However, once these advance planning activities have
been completed, the equipment has been set up and 1s operating smoothly, continu-
ing field operations are not especially difficult.
RGB has had twelve years of field experience with operations utilizing complex
cleanup equipment for hazardous material spills and waste sites. These experi-
ences have highlighted the need for careful, sequential advance planning and ade-
quate shakedown and testing prior to committing to expensive .field activities,
**
This phase would proceed In approximately the following manner:
o Meet with Region VII to define goals, objectives, financing arrangements, oper-
ating location(s) permitting responsibilities; division of activities between
ORD and Region VII (e.g.. Region would handle legal and public affairs, ORD
would execute technical aspects of project; Regional analytical support could
be very helpful 1f available; authority to access site critical). Note that
operating location may or may not be on the site to be cleaned up—depending
upon many factors.
o Define with Region VII a project management plan, including roles and responsi-
bilities of Regional, ORD, and contractor individuals on the project. Define
lines of communication and patterns of routine reporting. This 1s critical 1
*
o Define with Region VII a desirable scope of operation, e.g., materials to be
treated during demonstration, duration of operation, operating period per day
(8. 10, 12, or 24 hours).
o Define with Region VII means to excavate and transport (If needed) contaminated
soils to treatment site and treated soils from treatment site to point of origin.
-------
o Define with State of Iowa, as needed, permitting requirements and responsi-
bilities. (This will be done preliminarily during Phase III but must be
continued during Phase IV.)
o Prepare detailed site-installation design(s), with provision for security,
power, wastewater discharge, water supply, storage of"equipment and chemicals,
personnel support trailers or other quarters, etc.
o Prepare detailed logistical support list of all necessary equipment to be
taken to the field, including spare parts and necessary tools and trouble-
shooting apparatus.
o Arrange for necessary analytical support, either through the Region, a local
laboratory, or an on-site mobile laboratory, as appropriate. Prepare a de-
tailed Sampling and Analysis Plan and a Quality Assurance Project Plan.
o Arrange for suitable ultimate disposal (hazardous landfill, smelter) of con-
centrated lead products.
o Arrange for chemical and other expendable supplies.
o Prepare detailed project plans, Ip.cluding schedule and budget, with arrange-
ments for routine reporting to compare planned progress and expenditures
against actual progress and expenditures, and management "checkpoints.-14
o Mobilize operating crews, with appropriate safety, environmental, and operator
training (may be subcontractor personnel, particularly 1f 24 hr/day, 7 day/
week operations are needed and multiple crews with rotation are used). Conduct
training on equipment set up at Edison or at Bettendorf.
o Mobilize equipment Including all necessary arrangements for transportation,
setup, and on-site shakedown.
o Execute operation, in accordance with detailed operating plans.
o Demobilize and decontaminate equipment and restore operating s1te(s) to a con-
dition suitable to owners (criteria for suitability to be agreed to prior to
mobilizing personnel and equipment at site). Return equipment to Edison and
perform restoration maintenance, as needed.
The scope of .this Phase can be highly variable. It is desirable to clean up a
smaVl site or sftes to.demonstrate the suitability of the process; however. It
is not desirable to use the ORO equipment for extended operations for the pur-
pose of cleaning up many sites. The most appropriate scope will Involve a
short proof-of-technology demonstration to obtain specifications and cost esti-
mates such that the actual cleanup Involving many "hot spots" could be executed
by a cleanup contractor.
-------
foye it ui
Time Frame ... Planning & preparations:
Field Demonstration:
Report:
1-6 months (depending on permits)
30-90 days of operations (Including
some "down time" for maintenance, etc.)
Draft delivered 90 days after completion
of field operation; final report to
management after additional 90 days.
Cost $500,000 - $2,000,000:
(depending on hours/day of operation
and degree of acceleration of the
schedule)
Products
Final Report, consolidating the work of all phases, and providing
specifications, cost estimates, and activity schedules suitable for
use by Region VII In procuring contracted services for a full-scale
cleanup using EDTA-extractlon technology.
Technical paper, providing synopsis of Final Report.
-------
REFERENCES
1. IT Corporation, "Quick Response Feasibility Testing of LCPO Removal from
Contaminated Fill Materials by Extraction with EDTA for Application to
the EERU Mobile Drum Washer Unit", June 1984 - USEPA InHouse Report
2. IT Corporation, "Comparison of Alternative Technologies for the Removal
of Lead from Contaminated Soils - Leeds, Al" July 1984 - USEPA InHouse
Report
3. Connlck, C., "Mitigation of Heavy Metal Contamination 1n Soil", January
1985, New England Water Power From Control Federation
4. SAIC, Corp., "Treatment of Soils Contaminated with Heavy Metals,"
September 1985, USEPA Draft Report
5. Castle, C. etal., "Research and Development of a Soil Washing System
for use at Superfund Sites," November 1985, Management of Uncontrolled
Hazardous Waste Sites.
6. Davles, B., "Halkyn Mountain Project Report,* April 1983, Final Report
to the Welsh Office
7. IT Corporation, "Laboratory Feasibility of Prototype Soil Washing
Concepts, December 1983, USEPA, InHouse Report
-------
FACTSHEET
Environmental Protection
Agency
September. 1985
Mobile System For Extracting Spilled Hazardous Materials From Soil
The Hazardous Waste Engineering Research Laboratory, Releases Control
Branch at Edison, NJ, has recently developed a mobile system for extracting
spilled hazardous materials from soils at cleanup sites.
Landborne spills of hazardous materials that percolate through the soil
pose a serious threat to groundwater.
Effective response to such Incidents should Include the means for removing
the contaminants and restoring the soil to Its original condition. Currently
practiced techniques, such as excavation with transfer to land fill or flushing
with water In situ, are beset with difficulties - large land area and volume of
materials Involved. An Innovative In Situ Containment/Treatment System has
been developed to treat contaminated soils. However, It Is not suitable for all
sojls and/or all chemicals.
The mobile treatment (see Illustration) has been designed for water extrac-
tion of a broad range of hazardous materials from spill-contaminated soils.
The system will: (1) treat excavated contaminated soils. (2) return the treated
soil to the site, (3) separate the extracted hazardous materials from the
washing fluid for further processing and/or disposal, and (4) decontaminate
process fluids before reclrculatlon, or final disposal. A prototype system has
been developed utilizing conventional equipment for screening, size reduc-
tion, washing, and dewaterlng of the soils. The washing fluid • water - may
contain additives, such as acids, alkalies, detergents, and selected organic
solvents to enhance soil decontamination. The nominal processing rate will
be 3.2-m' (4-yd*) of contaminated soil per hour when the soil particles are
primarily less than 2-mm In size and up to 14.4-m' (18-yd1) per hour for soil of
larger average particle size.
For further Information contact Frank J. Freestone or Richard P. Traver,
Hazardous Waste Engineering Research Laboratory, Releases Control
Branch, Edison, NJ. Telephone numbers are: (201) 321-6632/6677 (commer-
cial) or 3404632/6677 (FTS).
MAKtUr »ATIR f.
SMNT CARION
PROCESS FLOW SCHEME FOR SOIL SCRUBBER
-------
Research and Development
EPA-600/S2-83-10O Dec. 1983
SEPA Project Summary
Mobile System for Extracting
Spilled Hazardous Materials
from Excavated Soils
Robert Scholz and Joseph Milanowski
A technique waa evaluated for the
•crabbing or cleansing of excavated
soils contaminated by spilled or
relesssd hazardous substsnces.
Laboratory tests were conducted with
three separate pollutants (phenol.
arsenic trioxide. and Dob/chlorinated
biphenyls (PCB'sI) end two soils of sig-
nificantly different character
(ssnd/gravel/siK/cley and organic
loam).
The tests show that scrubbing of
excavated soil on site to an efficient
approach for freeing soils off certain
contaminants but that the effectiveness
depends on the weshing fluid (water +
eddltives) and on the aol composition
and particle-size distribution, Bssed on
the test results, a fuO-ecale. field-ue*.
prototype system wss designed.
engineered, fabricated, assembled, and
briefly tested under conditions where
lerge (>2.5 cm) objects were removed
by • bar screen. The unit to now reedy
for field demonstration*.
The system mdudee two major aofl
scrubbing component*: • water-knife
stripping snd soaking unit off novel
design for disintegrating the soB fabric
(matrix) and solubillzlng the
contaminant from the larger pertidee
P>2 mm) and an existing, but re-
engineered, four-stsge countercurient-
extractor for freeing the contaminant*
from smaller particle* (O mm). The
processing rate of the system to 2.3 to
3.8 mVhr (4 to 8 ydVhr). though the
water-knife unh (used slone) can
process 11.6 to 13.S m*/hr (IB to 18
ydVhr). The complete system requJre*
auxiltory equipment, such a* the EPA-
ORD physieal/chemical traetment
trailer, to process the westeweter foj
recyding: under some circumstances,
provision must be made to confine and
treet releesed gesea and mists.
Treatment residuee consist of
skimmings from froth flotation, fine
particles discharged with the used
washing fluids, and spent carbon. The
principal Hmhing constraint on the
trsatsbOhy of sois to day content (high
weight-percent), since breaking down
and efficiently treating consolidated
days to impractical or not economically
a tti active. Meet inorganic compound*.
almost all water soluble or readily oxt-
dizabto organic chemical*, end some
partiaty miedbie hi water organic* can
be trested with water or water ptua an
additive, • .
During Imfted leboretory extrectfa
effflcientfy
organic
inorganle aofia. whereea PCS
arsenie dung mute teneciouely to tho
eoito and were relaassd tees reedOy Into
the washing fluids. The extent to which
the system he* practical, cost-effective
utility in a particular situation cannot be
determined until preUminery. bench-
scale lab work has been performed end
acceptable limits of residual concentra-
tions In the washed eofl ere sdopted.
Laboratory testa show that eofl scrub-
bing ha* the capabttty of vastiy
speeding up the retoaae of chemical*
fromsoBa. a process that oocura very
slowly under natural leaching
Not* that thto system i
vation of the
quentiy be replaced or transported t»m
low-grade landfiM. In ohu weshing of
contaminated eofl. • process in which
the contaminated eree to teoiated for
-------
•xarnple. by grouting, and then water-
flu*h«d with removal of the wash water
at a well-point is an alternative. The
overall efficiency of the soil washing
•rstem is greater than that currently
being achieved by in situ methods.
Based on the laboratory program. •
series of steps (water-knife size
reduction; soaking; countercurrent
extraction; hydrocyclone separation;
and waste fluid treatment for reuse)
was selected as the most suitable
process sequence for the prototype
system. The system was constructed
for the U.S. (EPA) and is now being
subjected to field evaluation. However.
soils rich in humus, organic detritus.
and vegetative matter can present
special problems in the extraction of
certain hazardous substances, which
may not partition between the solid and
fluid phases to a practical and necessary
extent.
This Pro/let Summary was developed
by EPA'* Municipal Environmental Re-
tearch Laboratory. Cincinnati. OH. to
announce key findings of the research
project thmt i* fully documented in a
separate report of the feme title (tee
Project Report ordering information et
back).
Introduction
The leaching of hazardous materials
from contaminated 'soils into ground-
water is recognized as a potential threat
to the Nation's drinking water supplies.
Such situations occur as the result of
accidental spills of hazardous substances
and from releases at the many uncon-
trolled hazardous wast* disposal sites
now known to exist across the country.
Current removal/remedial technology is
largely limited to the excavation and
transfer of such soils to suitably sealed or
lined landfills where uncontrolled leach-
ing cannot occur.
Onsite treatment can be • more cost-
effective solution to the problem. In some
research projects, contaminated soils
have been isolated by injected grout.
trenched slurry walls, steel piling, etc..
and then subjected to in situ leaching.
The effectiveness of such a process is
limited by. among many factors, the
permeability of the soil in its undisturbed
state. Economic and effectiveness factors
cannot be generalized but are situation-
specific.
An alternative process is needed for
those situations in which permeability or
other factors prevent effective in-situ
leaching and where landfilling is too
costly. The proposed technology — the
subject of the current effort — consists of
excavation, onsite but above-ground
treatment of the contaminated soil, and
return of the treated soil to its original
site.Excavation of the soil from its natural
state opens a number of options for
improved separation of contaminants
through better (high energv)mixing and
the potential for using different solvents.
Such cleanups can also be carried out
more quickly than they could by the
leaching of a more or less compact
natural soil (cost factors not being consid-
ered). This engineering approach has also
made it possible, or more convenient, to
incorporate any control devices that may
be needed to reduce emissions of particu-
lates or fumes into the air column and/or
to treat the contaminated wastewaters
generated during the processing.
The purpose of this project was to carry
out appropriate laboratory studies and to
develop, design, and construct a full-
scale system capable of treating a wide
range of contaminated soils. The existing
system will be useful for the correction of
long-standing (remedial) contamination
problems (waste disposal sites), as well
as for the emergency cleanup of spills and
for the prompt removal of released
wastes.
Discussion
To meet the objectives of the program.
specific criteria were identified for the
solvent the soils, the pollutants, and the
process.
To be suitable for field use in such a
process, the solvent or extracting fluid
should have the following characteristics:
1. A favorable separation coefficient
for extraction.
2. Low volatility under ambient condi-
tions (to reduce air contamination
effects).
3. Low toxitity (since traces of
extractant may remain in the
deansed soil).
4. Safety and relative ease of handling
in the field.
S. Recoverability for reuse.
The selected solvent must be able to
separate the contaminant from the soil.
preferably using • minimum volume of
solvent so that the equipment can be kept
compact In addition, the solvent must be
readily separable from the soil fines to
allow return of the decontaminated soil to •
the site and to permit treatment and
reuse of the solvent. High volatility in the
solvent can contribute to unacceptable
losses and can. when coupled with
ftammability. exacerbate health and
safety risks for the workers.
Following a brief evaluation and
screening of potential solvents (including
-organics). consideration of all the above-
cited factors clearly indicated that water
was suitable as the primary target
solvent. The use of additives such as
acids or bases, oxidizing or reducing
agents, or wetting agents was judged to
be a reasonable approach for enhancing
removal efficiency. Though certain
organic solvents can meet most of the
solvent criteria and may have definite
advantages in specific cases, a decision
was made early in the project to limit the
investigation to water-based systems.
The range of soils that is encountered
in a cleanup situation is very broad.
encompassing fine, highly cohesive clays.
sandy soils, silts, soils high in organic
maner. etc. Though processes can be
devised to handle any or all of these
materials, certain contaminated soils do
not require exhaustive extraction and
others do not lend themselves to an
extractive process. The organic content of
a soil can affect the ease of size reduction
and the efficiency of extraction. The pH of
• soil can affect the extraction efficiency
for a particular contaminant When the
soils and contaminants have catonic or
anionic qualities, ion exchange (partition,
factors cannot be neglected. / .
For purposes of this investigation, two
soils were selected as suitable represent-
atives of many that might be encountered.
These were a granular (sandy),
essentially cohesionless inorganic soil
(containing some fine sand and about
20% clay) and a highly organic (18.4%,
mostly as peat and humus) commercial
topsoiL
Though spill situations and waste.
disposal sites may differ in many ways
(such as the portion of a contaminant that
is tightly bound to the soil versus the
amount loosely associated in the voids).
plans for the test program emphasized
the spill situation by using freshly
prepared mixtures of soil plus
contaminant. Funding was insufficient to
support work with aged or weathered
contaminated soils that are more repre-
sentative of dumpsites.
The actual process for the planned
system must include excavation and
transfer to the processing equipment.
screening to remove large (>2.6 cm)
-------
objects, size reduction to maximize soil-
solvent contact, extractive treatment.
separation of contaminated solvent from
(relatively) decontaminated soil particles.
and return of the soil (either "as Is" or
after drying) to the excavation.
Excavation can be readily handled by
conventional earthmoving and
construction machinery. Size reduction
of soils can be accomplished with
various, commercially available
equipment, including rotary scrubbers.
log washers, attrition scrubbers, and
high intensity water-knives. The
properties of each were considered, and
the water-knife was chosen as the most
versatile unit; it was also suitable for both
disintegrating clay-like lumps and for
scrubbing the loosely held contaminant
from the resulting smaller (>2 mm)
components.
For the decontamination process to be
effective with a wide range of water-
insoluble and tightly held contaminants
on small panicles f>2 mm), follow-on
multi-stage extraction was judged to be
necessary. The use of countercurrent
extraction allows several stages of
extraction with minimum solvent use.
Clearly, the final system also requires
equipment to separate fines from the
solvent, both between extraction stages
and after the last stage. Gravity
separators, clarifiers. and filters were
generally inappropriate for the planned
system; hydrocyclones were selected for
evaluation.
The three hazardous contaminants
selected for testing were phenol, arsenic
trioxide. and PCB's. These were chosen
because of the frequency with which they
are encountered in spills and the range of
physical and chemical characteristics
they offer. Laboratory tests were carried
out to assess the effects of different
water-based solvents and different pro*
cessing conditions on these three
chemicals mixed with the two soil types
noted earlier. The results of these studies
were then used to design the full-scale
prototype.
Equipment Evaluation
Size Reduction and Extraction
A seriee of tests was conducted with
the water-knives, first using • local, avail-
able, uncontaminated soil sample.
Numerous approaches to exposing the
soil to the water-knife jets were tried and
abandoned (refer to the full report). Only
when the soil was contained in •
truncated, cone-shaped, tilted rotary-
screen drum (2-mm mesh openings) was
the desired lump breaking obtained. The
first tests were performed in an 18-in.
trash basket (top ID = 15 in.; bottom ID =
12 in.) in which 50% of the bottom
sidewall (up to 8 in.) was cut away in four
sections that were overlain with various
mesh screens. (The device was re-
engineered for the actual testing.) In the
bench apparatus, approximately two-
thirds of the soil was washed out through
the screen within the first 2 min of
treatment with 4.5 L/min (1.2 gal/min) of
water at a pressure of 4.9 kg/cm2 (70 psi)
and a drum speed of 10 to 20 rpm. Further
experiments indicated that a three step
sequence was needed to achieve the best
decontamination:
1. Low-pressure wash.
2. Soaking, followed by stripping, and
3. Low-pressure fresh-water wash.
Liquid-Solid Separation
To study the separation of soil fines
from water, a full-sized hydrocydone
(227 L/min) was used with different
inflow rates (and pressures) and different
concentrations of both soils. Though the
results of these tests show that the
hydrocydone is suitable for each soil.
they also indicate that the solids were
better concentrated in the underflow
from the inorganic soil. With both soils.
the overflow contained a small but
significant amount of fines (0.7% to 3.7%).
which would require additional separation!
Passing this overflow through the -
hydrocyclone in • second treatment was
not notably affective in removing these
fine solids.
Because the hydrocyclone was too
large for routine use in the laboratory
study of contaminant removal from soil.
simply gravity settling in e beaker was
evaluated and found to represent • good
simulation of the separation achievable
with the hydrocyclone.
Extraction Tacts
Tests were carried out with the three
chemicals (all three were not used in all
experiments) to establish the following:
a) probable loading on • soil column.
b) distribution on particles of different
sizes, and
c) effect of extraction with different
aovents on particles of different
Column Loading Studies
A stock solution of the contaminant
equal in volume to the void space in the
column was added to a 15.2-cm (6.O-in.)
column of soil (various moistures and
densities) and allowed to drain for 24 hr.
The contaminant remaining in the
column was calculated on a dry weight
basis, based on the amount of fluid that
drained from the column. Modified gas
chromatographic and atomic absorption
methods (described more fully in the
report) were used. Results obtained with
the three materials are shown in Table 1.
Note the heavy loading of phenol, which
represents the situation that might exist
shortly after a spillage onto soil.
Distribution Tasts
Different procedures were used with
phenol and with arsenic trioxide to evalu-
ate their distribution on particles of
different sizes. For phenol, dry soils were
first size-classified with a sonic fraction-
ation device. Each fraction was then
wetted with a stock solution of phenol.
After 18 hr. the fractions were rinsed
with water and analyzed. For arsenic, the
scil from the column dosing tests was
dried, size fractionated, end then
analyzed. High recoveries (based on
analyses) were achieved in both cases.
With phenol, these tests indicated that
approximately 90% of the contaminant
.was absorbed (or retained interstitialty)
on the larger particles (0.6 to 2 mm*) of
the organic soil. These somewhat
unexpected results also appear to be •
consequence of nonuniform distribution
of organic* In the different particle-siza)
fractions. Tests confirmed that the fine
particles contained predominantly
organic degradation products rather then
plant tissues, which remained primarily
with the larger particles. Such
differences may make it necessary, in
some cases, to preaoek the soil for
efficient extraction.
Unexpected results were also obtained
when testing the distribution of phenol on
the inorganic soil. The relatively low
adsorption by the finer particles wee
attributed to differences in internal
porosity and chemical composition
between the targe and-small, particles
rather than the proportionately greater
surface area (calculated on a weight
basis) of the fine particles.
The results obtained with arsenic
trioxide on the organic soil were similar to
those obtained with phenoL With the
-------
inorganic soil, however, the arsenic
compound exhibited the normally
••pected relationship between panicle
tite (i.e.. surface area) and amount
adsorbed. That is to say. because of the
greater surface-to-mass ratio, more
adsorption occurs per unit weight of fines.
PCB's were not tested to any great
extent because of their low solubility and
the hazards involved in working with
them. Time and funding constraints also
influenced this decision to curtail PCS
studies.
Water-Knife Stripping Tests
Contaminated soil samples were
subjected to 1 min of stripping by the
water knife to remove particles smaller
than 2 mm. Residual contaminants on the
remaining (larger than 2 mm) particles
were then determined. The results (Table
2) show the value of additional washing
or extraction, at least for phenol and
arsenic trioxide.
Chemical/ Extraction Tests
Since water is not the optimum extract-
ant for all contaminants tested, and sine*
most of the contaminants will be
absorbed by and adsorbed on the smaller
(<2 mm) particles, a series of tests with
the following aqueous solutions was
conducted to determine whether
extraction efficiency could be improved:
water *• sulfuric acid to pH 1
water *• sodium hydroxide to pH 11
water + 7.5% sodium bisulfate
water + S.O% sodium hypochlorite
water * 1.O% TWEEN SO
water * 1.0% MYRJ 52
water * 5.0% methanol
For the inorganic soils contaminated
with phenol, all extractions were highly
efficient, with removals greater than
87%. Only for the organic soil could the
difference between solvents be
considered significant, with the sodium
hydroxide solution being the most
effective solvent. A portion of the date
presented in the report is summarized in
Table 3. The relative and actual
importance of the residual contaminant
on the soil should not be ignored, nor
should the fraction of solvent remaining
in the soil (not shown in Table 3). When
the residual level of contamination is
f. Maximum Column Lotting*
Contaminant
Organic Soil
(mg/g toil)
Inorganic Soil
(mg/g toil)
fnanol
Arsanic trioxida
KB
4S3.2
5.0'
25.6
4*3
0.75*
3.0
"At artanic (At).
Taola 2. Eflact of Wishing on Larga Partielat *
Soil
Tun
Tima
(min)
Phenol
%Rammal
At,0t
fCB
Inorganic
Organ*
tS
3O
SO
120
15
30
60
120
97.3
98.2
98.8
99.1
60.7
79.2
86.0
91.6
28.9
Sit
42.2
SZt
47.7
SS.8
S4.0
59.0
21.4
50.0
21.4
28.6
•2 to 12.7 mm
TaMa3. Solvant Extraction: Raoratantativa Singla-Waihing-Tastt*
Contam-
inant
Pnanol
>»«iO,
KB
Soir*
1
O
,
0
1
O
• Sorrant
Watar
Watar
NaOHtpHII)
Watar
HjSOt (pH 1)
Watar
HtSOt (pH 1)
Watar
1% TwaanSO
Watar
1% Twaan 8O
Initial
SoZOota
(mg/g dry
toil)
48
452
0.78
6
3
26
*£-,
98.6
77.6
88.4
42.7
88.3
75.0
85.0
24.6
37.5
48.3
23.6
Supamatant
Concantration
1.1 9O
17.600
2O.OOO
16
32
375
426
72
110
416
366
Ratioual Son
Concantration
mg/g
0.68 '
100.4 / ;
• e*-g / ;
0.43
O.11
1.28
0.75
2.66
1.68
13.2
19.5
• fxtractant to dry tolMt 10:1 (w/wf.
** / » inorganic: O * organic.
sufficiently low. the treated soil may no
longer require disposal as • hazardous
material, e.g.. in a safe landfill
Samples of phenol-contaminated
organic and inorganic soils were also
subjected to multiple extractions. These
tests demonstrated that continued
removal of phenol did occur, even when
the extractant was recovered solvent
(water) from a previous stage and already
contained phenol. Residual phenol
concentrations of 30 mg/kg (0.03 mg/g)
of soil were achieved after four
countercurrent extractions of the)
inorganic soil.
Prototype Design and
Construction
The process sequence) for full-scale)
treatment (Figure 1) was finalized, based
on the laboratory experiments. The
sequence includes Initial removal of
oversized chunks O2.6 cm), water-knife
-------
scrubbing to deconsolidate the remaining
soil matrix and to strip any contaminant
loosely absorbed on the solids (>2 mm) or
field in the void spaces of the soil, and
four-stage, countercurrent extraction
coupled with hydrocyclone separation
after each extraction stage to separate
the solids (<2 mm) from the liquid. Froth
flotation is used to give maximum mixing
of extractant and soil in each stage. The
overhead extract (mostly sorbent) from
the first stage extractor hydrocyclone
contains the highest level of dissolved (or
dispersed) contaminants and fines. This
extract must be clarified and then treated
(possibly with activated carbon) before it
is recycled.
Note that chunks (> 2.5 cm) are not
normally processable in the system
except for moderate washing on a bar
screen*; the 2.5-cm to 2-mm as well aa
the <2-mm fraction, will be used to fill in
the .excavation; all processing fluids must
be appropriately treated. All dust and
vapor emissions should be ducted to an
air cleaner or scrubber before discharge.
The basic system was constructed
according to the design shown in Figure 1.
The water-knife unit (rotary drum-
screen scrubber) consists of a tilt-skip
loader and hopper feed from which the
soil moves into a tillable 19-m(21 -ft) long
by 1.4-m (4.5-ft) ID cylinder fitted with
end piece*. wateMtnives. and a rotating
mechanism (Figures 2. 3, and 4).
Soil is metered from the tilt-skip
reservoir hopper at rate* up to 18 ydVhr
.onto •. manually washed bar screen
where >2.S-ctn (1-in.) chunk* are
rejected. The solids then pas* into the
tilted drum-screen scrubber where it is
subjected to first-stage water-knife strip-
ping, water Making, and finally second-
stage water-knife stripping using fresh or
partially recycled water. The first section
of the scrubber cylinder i» 1.3-m (4-ft)
long and is fabricated from 2-mm mesh
(HYCOR Contra-Shear screen) and
equipped with internal water-knives.
Solids then move into the 5-m (15 ft)
soak cylinder that is fined with a baffle
plate that has a 0.5-m (22-in.) center
opening through which solids pas* into •
0.7-m (2-ft) long screened, water-knife
rinse zone, fines «2 mm) pass through
the screens, a* doe* the wash water. The
coarse panicle* are voided at the end of
+2 mm Scrubbed Soil
on • 7.S- or 5-cm (3- or 2-irt.) upper »cmn I
•kip-hoppor from which large or norn»in<«Of»bt«
chunk* »r« r*kod orl. Wotrtocl chunk* tfwt DM* lh»
upper icroon« urm r«jwt*d and removed M tho
Meond (lowort bar MTMA «2.S cm (1 taj,
CbtnAir
Diiehirgt
Conuminutfd
Soil
Ovtnin
Non-Soil
Mtttritlt
'' Exhaust
from Hood
Stimmingt
to Oitpotml
t
Counter-Cutrtni
Chumicfl
Extnctor
-2 mm
JDrfing \
B.d \
Soont
Wishing
Fluid*
I
Seruobmd (H Ntodod)
Soil •- '
Runoff
Clffifiof
FMtf
0*cAwMr>
Fin** to
Oiipotsl
CltfHitd
Wfttiing Fluid*
Wutiing Fluid /tocycfer
;
Spent Cwoon
Ftgui* 1.
Ffgum 2. Fully conttmctid fcrery drum tti *»n terubbor.
the drum. The unit can be bacfcflushed as
needed. The screen* resist buildup of
fines (blinding). The actual arrangement
of the water-kntve* and other detail* of
construction are given in tho project
report
From the water-knife and soaker unit.
the slurry (<2-mm particles) is pumped to
the coiintercurrent extractor. The four-
etagc- countercurrent extraction unit
(Figure* 5 and 6) ha* been modified from
the ao-called EPA beach sand froth
-------
Tilt Skip
Hopper up to
Loud Metering
Hopper
Drum-Screen
Soil Scrubber
Hind Wtth
Large
Stones
Figure 3. Soil loeding end metering system (cross sectional side view).
Rinse
U2one
2ft.
1
P
Soak Zono M
* IS ft.
Initial
Sprey Zone
4ft.
-«
1
--
Soil Out
Inner
Cylinder
Outer Shall
A. Drum cross sect/on
,— 16 Inchn
Beffl*
Soil Surf i
Inner
Winder
Figure, 4.
Sou* Zone)
Channel Formed by
Soil end Drum Wet
B. Drum Itomotrhj
Soek tono description.
6
flotation unit.* Basically, the washing
chamber was partitioned into four
sections (3-ft long X 4-ft wide X 5-ft
deep), each of which has an aerator
agitator and a hydrocyclone with pumps
and piping. Flow of solids «2mm) and
fluid is countercurrent with clear water
being introduced at the fourth (discharge)
chamber (Figure 6). The extraction unit
has an on-board diesel generator; the
water-knife unit requires external power.
The underflow (solids-rich) slurry from
the fourth hydrocyclone is discharged to a
drying bed.
To achieve mobility, the water-knife
unit is skid-mounted for transport by
semi-trailer; the countercurrent extractor
is integrally attached to a separate semi-
trailer. Refer to Figures 2 and 5 for details.
Calculations indicate that the total
system has a throughput range of 2.3 to
3.8 rnVhr (3-5 ydVhr). but that the
water-knife unit alone can process 11.5
to 13.5 mVhr (15 to 18 ydVhr).
Conclusions
\The following conclusions can be
drawn from the work carried out during
this program and the knowledge gained
during that effort:
1. Spill-contaminated soils can be
excavated and treated onshe using
extraction with water or aqueous
solutions for many pollutants that
are frequently encountered in such
situations.
2. A system capable of decontamina-
ting 2.3 to 3.8 mVhr (3-5 ydVhr) of
soil has been designed and
constructed and it is now available
for field testing by EPA.
3. Water-knives function as • compact
efficient, and economical means or
achieving effective contact between
contaminated soil particles and
extractant.
4. Countercurrent extraction is an
affective process for removing
certain adsorbed contaminants
from soils and. for the sin of
equipment needed, hydrocyclone*
are preferred devices for separating
the extracted solids from the ex-
tractant.
•Qantt 0. Gurntt. MwMrmkM of I
•tad by Oil. EPA4U-72-O4S (Washington. O.C; US
EPA. 1*721
-------
5. Laboratory experiments demon.
• strata that soil characteristics
(particle size, distribution, organic
content. pH, ion-exchange proper-
ties, etc.) are important factors in
the removal or retention of
contaminants.
6. In addition to the actual percentage
of the contaminant removed, the
allowable level of pollutant
remaining in the soil is an important
factor in determining when
adequate decontamination has
been achieved since the final.
residual concentration affects the
options available for disposal of the
cleansed solids.
Recommendations
Based on the observations made during
this' investigation, several suggestions
are offered for future work.
1. Laboratory screening tests should
be performed on a wider range of
typical compounds and mixtures
encountered in hazardous
substance spill and release situa-
tions to ensure that appropriately
high levels of decontamination can
be achieved with this process.
2. The results of this study apply pri-
marily to spill situations. Contami-
nated soils found at waste disposal
sites may exhibit different
. extraction characteristics because
of the extended soil/contaminant
contact time and of weathering and
in situ reactions. Studies are needed
to establish whether and to what
extent such change* affect the
decontamination process.
3. Other extractant solutions should
be evaluated to determine whether
the efficiency of the process can be
improved without damaging the
equipment or increasing the
hazards to which the workers are
exposed.
4. A wider range of soils should be
examined to determine what
changes in the system are practical
to better cleanse soils with charac-
teristics (e.g.. greater conesivenesa
and adsorptive properties of clay-or-
«ilt-rtch soils) that differ signifi-
cantly from those of the soils already
tested.
Figure S.
Chemical
Additive
(If Needed)
EPA Froth Flotation System (batch cleaner) modified »* a countereurnnt
chemical extractor for soil scrubbing.
Raw
Feed
Chemicel
Additive
flf Needed)
Chemical
Additive
(If Needed)
Fresh
Water
Figure 6.
Slurry Pump
Process flow schema for soil scrubber.
Clean
Product
The full report was submitted in
fulfillment of Contract No. 68-03-2696 by
Rexnord. Inc.. under the sponsorship of
the JJ.S. Environmental Protection
Agency.
-------
Robert Scholz and Joseph Milanowski are with Rexnord Inc.. Milwaukee. Wl
53214
John £. Brugger is the EPA Project Officer (see below).
The complete report entitled "Mobile System for Extracting Spilled Hazardous
Materials from Excavated Soils," (Order No. PB 84-123 637; Cost. HI.SO.
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-437-4650
The EPA Project Officer can be contacted at:
Municipal Environmental Research Laboratory—Cincinnati
U.S. Environmental Protection Agency
Edison. NJ 08837
United States
Environmental Protection
Agency
Center lor Environmental Research
Information
Cincinnati .OH 45268
BULK RATE
U.S. POSTAGE
PAID
Cincinnati. Ohio
Permit No. 636
Official Business
Penalty for Private Use (300
• us. QOVCNNMCNT Mw
ornct ie»«- na- ica/sie
-------
United Slates
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-205 Oct. 1981
Project Summary
Guidelines for the Use of
Chemicals in Removing
Hazardous Substance
Discharges
C. K. Akers. R. J. Pilia. and J. G. Michalovic
This project was undertaken to
develop guidelines for the UM of
various chemical and biological agents
to mitigate discharges of hazardous
substances. Eight categories of miti-
gating' sgents are discussed slong
_ with their potential uses in removing
' hazardous substances discharged on
land and in waterways. The agents are
classified as follows: mass transfer
media, absorbing agents, thickening
and gelling agents, biological treat*
merit agents, dispersing sgents. pre-
cipitating agents, neutralizing agents.
snd oxidizing agents. Each of these
classes is developed in terms of the
agents' general properties, their use in
spill scenarios, evnironmental effects.
possible toxic side effects, and recom-
mended uses.
A matrix of eountermeasures has
been developed that refers to various
classes of mitigating agents recom-
mended for treatment of hazardous
substances Involved In spills in or near
a watercourse. The matrix includes a
list of hazardous chemicals, the
corresponding U.S. Environmental
Protection Agency (EPA) toxlclty
classification, and the physical prop-
erties of the chemical.
TMt Projict Summary w*i devel-
oped* 6y EfA't Munidaml Environ-
mtnttl Rttoorch Liboritory. CMIC**>
•fieff. OH. to a/wtotf/ice Aey finding* of
tht rexeascA pro/act fAaf /* fully
doeumontod In o ttpormto report oftho
same fft/e /see Project Rtport ordering
information ft b*c*l.
Introduction
The 1972 Water Pollution Control Act
Amendments gave the) U.S. Environ-
mental Protection Agency (EPA) re-
sponsibility for removing spilled haz-
srdous substances from the) environ-
ment. EPA was also made responsible)
for developing criteria to be used for
designating substances as hazardous.
Of the two criteria developed, the first
involves the potential toxic effect of •
substance on the biosphere. The second
criterion considers the probability of
spills bssed on snnual production.
methods of transporting, storage.
physical-chemical properties, and past
history. Bssed on these criteria. •
proposed list of hazardous substances
was published in the Federal Register
(VoL 4O. No. 250) on December 3O.
1975.
The responsibility EPA beers for
hazardous material spills raises many
questions about removing discharged
hazardous substances effectively. Me'ny
parameters are involved in deciding
how to countered • hazardous sub-
stance spill, and which countermeasure
(if any) to use. The guidelines developed
by this study for mitigating hazardous
materiel discharges ere to be used by
EPA in the future to expand and revise
-------
Annex X of tha National Oil and Haz-
ardous Substance Pollution Contingency
Plan. 4OCFR1S10. so that it includes
specific reference to chemical use for
spills of hazardous substances. The
guidelines also establish a method for
determining the circumstances under
which a particular mitigating agent can
b* used and those under which the use
of chemicals and other additives is
harmful to the environment.
Results
Use and Effects of
f^itigating Agents
Study results are outlined in Table 1.
which summarizes the recommended
uses for each class of agent and the
possible toxic side effects associated
with their use. The eight categories of
mitigating agents are as follows: mass
transfer media, absorbing agents.
gelling and thickening agents, biological
treatment agents, dispersing agents.
precipitating agents, neutralizing agents.
and oxidizing agents. The recommended
uses, effectiveness, and possible toxic
effects of these agents are discussed
her* briefly.
Note that the effectiveness of •
mitigating agent depends largely on the
specific spill situation. The amount of
•gent needed to counteract a hazardous
substance discharge is dictated by many
factors, including the sis* of the
watercourse, the conditions of flow, and
the possible long-term toxic effects of
7*64* 1. Mitigation Summary...
Mitigation
Category
irretrievable contaminated agents and
byproducts.
Mass Transfer Media—
Agents within this category include
activated charcoal and ion exchange
resins. Available evidence indicates
that activated charcoal and ion ex-
change resins introduced in moderate
amounts to the aquatic environment
will not in themselves be toxic. But the
desorption. of • hazardous chemical
from such mass transfer media in
natural surface water and the potential
persistence of these toxic organic
compounds in the aquatic environment
must be considered in any decision to
us* irretrievable mass transfer agents.
We can safely assume that if those toxic
compounds can b* removed from the
environment by biological processes.
they can also b* removed if bound to a
mass transfer medium. We can also
assum* that the total toxic affect of
those biodegradable materials can be
reduced if mass transfer agents can be
used to minimize acute toxicity.
Irretrievable mass transfer media
should be considered acceptable for
treating the class of materials that is
biodegradable under all conditions.
Absorbing Agents—
Th* us* of absorbing agents is
generally limited to spills of oil and
petroleum products. Natural agents
such as straw, sawdust, etc.. are
routinely' used in such cleanup*. A
Possible Toxic effectfs)
variety of synthetic absorbents are
available for mitigating both hydrophobia
and hydrophilic chemicals. These ab-
sorbents are nontoxic and do not
present a hazard to the environment in
an uncontaminated state, but desorption
of the spilled* material from both natural
and synthetic absorbents can be signifi-
cant. For this reason, the use of
absorbing agents is recommended only
in those situations in which the sorbent
can be removed from the environment.
Thickening and Gelling Agents—
Mitigating agents in this category are
actually special types of absorbents
used to immobilize the spilled material
to prevent further spread into the
environment and to condition the spill
for mechanical removal. We recommend
that these agents be used on land spills
of all liquid materials on which they are
effective. Certain agents should be
considered appropriate for treatment of
water spills of insoluble organics that
float. Thickening or gelling agents
should not be used on water spills of
materials that sink or mix into the water
column.
Biological Treatment Agents—
Biological agents have been shown to
be effective in mitigating spills of oil and
oil-derived products. Several limitations.
however, exist to the use of these
agents in the treatment of spilled
organic materials.
/tecommended Uses
Mass transfer madia
Absorbing agent*
Thickening and
galling agents
Biological treatment
Dispersing agantt
Precipitating agents
Neutralizing agent*
Oxiditing agent*
Desorption of hazardous substance -
chronic toxicity.
Desorption of hazardous substance • .
chronic toxicity. increased biological
oxygen demand.
No known toxic effects.
Biodegradable substances.
All land spill*. Insoluble organics that
float, provided absorbent can bo removed
from tha environment.
All land spills. Insoluble organics that float.
Ecological imbalance. Toxicity of da- Biodegradable substances. Spills that are
gradation product*. easily contained and monitored.
Increase in toxicity resulting from dis- Biodegradable substances.
parsed substances. Toxicity of degrada-
tion product of added agent --•—_
Toxic effect of insoluble metal salts. Removal of metal ions from solution.
Toxicity resulting from change in pH All spills involving acids or bases.
from natural conditions. Toxic metal ion
byproduct.
Toxic intermediate reaction product*. Limited to detoxification of hatardous sub-
Oxidation of natural organic materials - stances in closed system to allow control
ecological imbalance. of reaction.
-------
Considerable time is required by the
biological degradation process, which
makes it necessary to contain and
isolate the spilled material from the
environment before treatment. The
bacterial culture must also be given
sufficient nutrients and maintained in
an environment that will encourage
adequate growth. A culture maintenance
program must therefore be initiated.
Finally, no agent should be introduced
into the environment if it will cause any
significant change to the ecological
balance of the treated waterway.
• Biological agents should be considered
appropriate for treating spills of materials
that are biodegradable, but only when
conditions allow the contaminated
environment to be contained for suffi-
cient time to permit detoxification.
Other types of mitigating agents should
be used whenever possible.
Dispersing Agents—
Dispersing agents can be used to (1)
increase the rate of biodegradation of
spilled material. (2) protect aquatic fowl
by removal of oil or other organics from
surface water. (3) minimize fire hazards
by dispersing hazardous material into
the water column, and (4) prevent
shoreline contamination. Some dis-
persants are toxic, however, and care
must be exercised to prevent unneces-
sary harm to aquatic life.
Precipitating Agents—
Precipitation is • valid mitigating
technique for removing toxic metal ions
from solution. The technique generally
requires the addition of either hydroxide
or sulfide ions at elevated pH levels.
Hydroxide ions will re-enter the water
column when the pH returns to neutral.
creating the possibility of a long-term
environmental hazard. Sulfide precipi-
tation is thus recommended. At toxic
concentrations of heavy metal ions, an
insoluble metal sulfide will form and
reduce toxicity rapidly. The precipitate is
insoluble enough to reduce re-entry of
metal ions into the environment to a
nontoxie level. Further study will be
necessary, however, to determine the
long-term effect of metal salts on the
water system.
A byproduct of sulfide precipitation is
toxic hydrogen sulfide gas. To inhibit
hydrogen sulfide formation, the sodium
sulfide precipitating solution should be
stabilized with sodium hydroxide.
Neutralizing Agents—
Neutralization is an acceptable method
of treating all spills of acids and bases.
provided some method for monitoring
PH is available. Treatment should be
accomplished on land whenever possible
to prevent the spilled material from
entering aquifers or surface water. Toxi-
city associated with pH change from
normal values once the spill has entered
a waterway is critical, in which case
neutralization of the spill becomes the
primary method of treatment.
Toxicity reduction is coupled with the
return of normal pH values regardless of
the neutralizing agent; however, care
must be taken to select an agent that
produces the least toxic byproducts. All
other considerations being equal, weak
acids and bases should be selected to
neutralize a spill in preference to strong
acids and bases. This policy will
minimize the potential for overtreat-
ment. The use of solid agents should
also be avoided when possible.
Where the monitoring system is not
accurate enough to ensure treatment to
the exact pH desired, it is better to
undertreat than to risk overtreatment.
PH values between 6 and 9 are recom-
mended.
Oxidizing Agents—
Oxidizing agents are toxic to most
organisms at relatively low concentra-
tions. The reactions are difficult to
control and seldom go to completion.
thus leaving toxic intermediate reaction
products. The use of oxidizing agents
should be limited to land or water spills
that are completely contained. Further-
more, these agents should be used only
as a last resort.
Countirmttsur* Matrix
A comprehensive list of the various
types of mitigating agents and their
potential uses has been generated in
matrix format (Table 2). This counter-
measure matrix.refers to classes of
agents recommended for treating
hazardous substances involved in spills
in or near waterways. The matrix is a
comprehensive list of hazardous chemi-
cals, the EPA toxicity classification for
each, and the density and the physical
form of the pure hazardous substance.
Each chemical is also assigned a
physical/ chemical/dispersal (P/C/0)
factor, which has a range from 0.1 to 1.0
and is "....based on the solubility.
density, volatility, and associated
propensity for dispersal in water of each
hazardous substance." 40CFR60002.
December 30. 1976. The remainder of
the matrix specifies which categories of
countermeasures are effective for
controlling hazardous substances dis-
charged on the ground or in a waterway.
The full report was submitted in
fulfillment of Contract No. 68-03-2093
by Claspan Corporation. Buffalo. NY.
under the sponsorship of the U.S. Envi-
ronmental Protection Agency.
-------
T*bl» 2- Hmutdous Sut>sttne»/Count»rm»»sur» Matri*
r-rc/o
AifM •<•*
A€fft
Atftan*
At*n< *
A*t*
e
e
e
e
a
o
A
c
o
A
g
c
o
o
e
o
e
0
0
0
0
0
0
s
0
0
0.
0
0
c
0
0
0
0
0
0
C
0
0
0
0
O
0
c
e
e
e
c
c
c
c
a 7« i
f.0*f l
rod i
a» i
I.M t
iff 1
a«M i
o.wr t
a** t
I.U S
0.«M t
04 <•
2.M *
»«• s
a«o «•
1.073 S
i_» *
ij* *
z»« *
«t *
_ s
149 «
_ S
_ s
».*» x
•jr X
_ s
f^> X
»Jf X
ox x/i
"• ?
I.M X
»JO X
_ X
I.M X
2.0* X
_ S
t.at x
• «» x
t.«f X
r.jf x
_ X
aw i
i.att i
1M X
1M X
XX X
4,14 X
4.4 X
4LJ* X
M
M
V
V
SS
a
Sf
sf
st
a
H
NT
f
p
Sf
SM
SS •
SS '
ss <
SS '
ss
ss
ss
St*
ss
ss
ss
ss
SM
/M
- ss
SM
SM
ss
sx
ss
sx
SM
ss
ss
ss
SM
SM
r*r
ss
f
f
f
f
f
f
*
v«
t
*
•
•
•
•
*
•
•
•
•
-------
(continued).
MM«W
«rm tt
******
*«•**•
*OT*M»
/•'MAC
inc/wfcn*
/•/««/•«
»«•!<§
O^Mt
• !• urn uimitt
ffMMfW
J«flf«« MBrf
•«««•«•«»
^ni.
-rr
3f**L
*<»««*•"
0«t>fc»»l«
Jm^a r«-r
Cj^VV^Nt AMMM
Tl*n n riVir (t
r.fm.n ri»nic«
CMMMI «MMW
2SZ***"*
r£££»
AMMMtMMMM
*££-.
0*nM
CjrfMMIWA
C^BM
C**»j/~ ~ '
CvftM^MWriMi
CMMM*
CMvm
OMViA«M*M
£y***"? ._
Mtf
CTumj'i n»m«
On-wtr*
*:srr
Cz^sr
C^2T
C/UJj^uj*
C«M«Mfl
iuflinMI (
C^^NW^IW
O*M<
Cw«rmn»
'ij^i**"
C4«n«Muw
Ct^rwmvM
Cx>« «n>M«
tr*
C«M-
c
c
c
4
C
•
0 •
A
'C
o
e
0
0
e
e
£
e
o
f^
^
e
e
o
0
0
0
A
0
A
0
c
A
4
0
0
C
0
0
0
0
e
c
e
c
4
0
•
*
0
0
0
OMMT
tj
/•£*
*4
4.0*
114
J.W
141
_.
a«7»
f.JM
I.OI
».»
r.jo
>.M
t.M
a 74
1.00
l.Of
4.IJ
4.01
—
il
2.M
-
1104
13*
140
f.4
~~ «•
I-M
l^t
jj
l.f
!-•
1.0
_
17
• J
in
I Iff
9-9*
Z47
4.4*
if a
_
•«
1.0
'"*
^
*** .
L
—
ta
—
****•!
Firm
S
S
S
S
S
S
S
S
i
S
I
I
S
S
S
i,
1
S
-
*
S
S
S
S
a
a
a
a
- a
t
i
t
t
i/a
i
s
i
a
a
a
a
a
a
a
a
a
a
a
a
a
a
*M«r«A«nf
M«u Tltnrfir Mctfw AgtM
ffC/O Ato- fntiif tittofirir
C«t«- r««W C«(I *•««• /lotf ••<• /IfWM /•»»• /IfWM 4p«<> 4^M 4p«n>
/• * t m
f • * t • * *
IS t t m t * *
r* t m M
P i m g
f M t t t
IS t * m m
SS M t m * t
Hf * ' •• t
SS * m
SS .
f t M m
f t * m *
f m m m
M • « »
Mm m m t * m *
SS t • t
f ' *
a*. .
SS • m
* • . •
SS m • m
«m m ' m
m v m •
ss * '*
SS t
Sit t m • ' '
SU m : i
SS * : •
ss — « ----- • . ~^-
SS t m
rs t •
Sf •
a* *
a * t
SS m • m * •
SS * * • M
SU • t m • t t •
Mm m m
• • •
a • * »
Mm m ' m-
mm 9
f • m t m •
fill m
ft* *
f t m t
SS t *
SS t , i * *
* * ' '
a * t t t
9 m m m
r t t t
fit t
f * t t '
ft* t • . •
a * * * t
5
-------
(continued)
MM* Tmu/* Mf
If*
Cm*-
^/c/o
At**
Aco-
Cv6f*€ i
Ciw>c mtlut
Cuor* <«««>•
CM«r«c ttfirt
It
f
f
a
is
CrcloA*i»n9
3.4-O •€*
oo r
0-wr.
A
C
t
•
•
A
C
C
A
A
A
e
e
e
t
OwiMM
fnOMMTM
f>/>
tUtflrn
IOTA
e
e
o
e
e
e
e
e
e
e
MM
0.779
1.3*
I.I It
1.79
an
r.M
f.M
1.14
i.a
O.M
a
4.
t
*
i
S
s
s
4.
S
I
I
I
S
I
S
I
S
t
I
*
t
t
a
a
ss
a
a
ss
ss
ss
ss
ss
Sf
ss
ss
ss
ss
ss
ss
ss
a
ss
4MP
»
a
t.M
za
i.a
I.O9
1*3*
1.44
I.M
i.ao
i.it
O.JO
i.a
atti
aw
IM
7t
4LM
**
4M
1.4
T.I
t.11
AM
ua
ss
ss
ss
M
ss
ss
a
is
a
a
a
a
a
a
sx
-------
f«0/« 21 (continued)
tt*
Ai|*icW
r/c/o
C««-
Mtu limntlfe Mmjim
Acv-
*t*M
Hfti*
Agtn
Ag*M
MMW/MM
««rrur« «•/«!
M*/rur« Cw»-
MMAttrcAfcr
MM/IT*
«<*•>•
*««« «>«•«
lurimmemt
o
o
4
0
0
A
A
A
A
A
A
t
O
t
A
C
C
0
0
C
C
0
O
e
0
e
c
4
o
o
0
c
0
A
c
c
>.*•
4.0*
«J
£47
>.4f
a«7
ant
I.JM
1.0*
(.10
1.4
ija
X.M
4.M
».*4«
I.JOJ
».»•
j.*w
f.4
rlM
turn,
T*>M>
ti< mil «r«»»M
XX
XX
XX
XX
XX
XX
XX
&u
XX
XX
XX
XX
XX
XX
-------
Ttblt 2. (continued)
M»*l Truu/ir Htdt*
UMftitl
Sodium "wocnAxx*
S odium motnrtoto
Sodium OKrito
Sodium pnouHotm
mOnoHOM*
*££'*"*""
trtoosit
Sodium ulonat
Sodium lull-do
Sltnnoul tluond»
SutMHim chromtu
J^wjjuma
Sirrono
Sutunt **
2.3.S-T otmt
2.4.S-J «w*
roe
Touoothr* *o*
rrmettddetuju
Toulon*
rttffftonu
Tiichttttm
r/c/MWMMl
^EEr*"*
TtiMJi^itmjig
UrtnvmtdnM*.
(/r«nr< ftCMftf •
Uroftyt Mtmn
{Jttn\i f vtfftffft
Vlflfl OCOtOlO
MfOtHf
XtoHttl
Zoatojt
bntocflou
^mmuinam
/Ww §cfrftim<«
2i*K gofote
I'nt tn~i ft
jy, ,»(,nMt
/MV cMftrrtto
Ijttf citntdg
2i/K ttutttd*
£fv tofmotg
li* mj+ttutiol
&»*«""•
'ulfVMf**^
l**ft*M»tudm
"£££""*"
lint ti^tmUllifidM
tint tuHot*
^mo^rZ+OH
linonum mux
tmnium MOM
Zi'tonum lufoM
C«t
4
C
f
O
o
o
c
c
0
0
c
c
c
4
4
4
4
t
C
4
f
4
t
C
0
0
0
0
£
C
c
c
• e
c
c
c
c
e
c
c
c
4
c
c
c
c
e
c
c
c
c
c
o
0
0
o
o
0~«r
_ _
14
Zf 7
ZM
ze*
>.t
/•O
f.«M
2.71
_
13*
O.M»
f.«34
'-
«•
^
».«*
».J
OM
I.M
1 jj
>!l
I.IJ
Zf
Z0
r«
X20
t^f
^
aM
aM
r.a>
_
1.73*
IK
^
3.U
4ZZ
44}
1*07
f.M
4.M
2JI
Z07
^
4.M
.^
zr
J.M
ii«
^
—
-
•<•
J.12
*sr
X
X
X
X
X
J
X
X
X
X
X
t
t
X
X
x
1
t
I
I
f
4
I
I
X
X
X
X
x
X
X
1
i
_
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
•VC/0 4«t»- *«»• •«*»«•*
CM- twW Coltntit Anitnie Utont JitumoM Of/tint **«"*<"» Oudamt Oilftnuif
SM ft i
M < « t « «
« <
XX ft
X/M •
U
XX i
XX < i * •
XX « > •
a < ft i *
XX i *
*tr < « < <
M ft ft « «
«_ •
• •
- a •
a •
a • •
a « « •
M ft *
MP ft • « ft ft «
ct ft • <
XX « *
a « • •
XX • « «
X* ft < • « •
X* * ft < « •
cl « < «
A ft ft
^ C ft
•'ft *
ft* ft ft
X* « ' • , * • *
JM* ft ft ft •
XX ft • • .
XX ft < '
* ft * • ft .' ;
• ft ft ft ft - ,
f • ft ft ft
^ ft ft ft ft
^ ft ft ft <
a • • «
•* « « «
a < • < • •
•• • » • «
^ • « ••
«••<•« • '
«• • • <
•• • • • «
a • • • *
a • • • «
* « • < •
* « « <
f g g <
* • « <
^ • « <
^ • • •
41 * * '
^ • « *
-------
C. K. Alters, ft. J. Pilit. and J. G. Michalovic are with Claspan Corporation.
Buffalo. NY 14221.
Joseph P. Lafornara is the EPA Project Officer (see below).
The complete report, entitled "Guidelines for the Use of Chemicals in Removing
Hazardous Substance Discharges." (Order No. PB 82• 107 483: Cost: S9.SO.
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
For information contact John £. Bruggar at:
Oil and Hatardous Materials Spills Branch
Municipal Environmental Research Laboratory—Cincinnati
U.S. Environmental Protection Agency
Edison. NJ 08837
- f
fctt $. GOVBMMMT IttNTMG OWtt lfM/SW-On/0343
-------
WIZARD OF ID
B.C.
FEAR /wj's Misuse OF ppr
OF \&
• 'r-7
«N*«i MMMM *|iK>M«*. IN*
-------
RCRA/CERCLA
ALTERNATIVE TREATMENT
TECHNOLOGY SEMINARS
SOIL WASHING OF LEAD
CONTAMINATED SOIL
EPA Mobile Soils Washer
Leeds, Al.
-------
DRUM WASHER UNIT
-------
MOBILE SOILS WASHER
Counter-Current Extraction Unit
-------
PROCESS FLOW SCHEME FOR
COUNTERCURRENT EXTRACTOR
CHEMICAL
ADDITIVE
(IF NEEDED)
SPENT
WASHING,-*
FLUID
RAW
FEED
CHEMICAL
ADDITIVE
(IF NEEDED)
CHEMICAL
ADDITIVE
(IF NEEDED)
FRESH
WATER
SLURRY PUMP
CLEAN
PRODUCT
-------
CM£L*NT
PROCESS FLOW DIAGRAM AND SAMPLE POINT
LOCATIONS • BATTERY WASTE TREATMENT SYSTEM
MOODY, AL.-EPA/EERU
FIOCCUIANT
MAKE-UP
TANK
FILTER RINSE TANKS
-------
LEAD REMOVAL
LEEDS, ALABAMA
LEAD E.P. TOXIC1TY
MATERIAL (mg/I) (mg/I)
ILCO FEED 47,000 88
+ 2mm DISCHARGE 3,050 60
- 2mm DISCHARGE 1,300 49
-------
IN-SITU WASHING OF
JP-4 AND SOLVENTS
Volk Air National Guard Field
Camp Douglas, Wl
LOCATION OF VOLK FIELD ANG UASE AND HARDWOOD RANGE.
f \
\ S Wisconsin \ .
\l VI
,-V\"Q
-- XJ / (
-' \~i \
-------
LOCATIONS OF THE PROPOSED AREAS Al VOLK FIELD ID Ut
INVESTIGATED DURING PHASE II
Regional
Up-gridlent
MeM
ML-1, Fire Training/
Transformer Site
HL-4
Current L*ndfni
Former LtndfiU
-------
LAYOUT OF THE IN SITU CONTAINMENT/
TREATMENT UNIT (ISCTU)
A. AIR DIAPHRAGM PUMPS
B. PROPORTIONAL CHEMICAL
ADDITIVE METERING PUMP
C. INPUT MANIFOLD
O. PROCESS MONITOR RECORDER
f. WATER PUMP
f. BATCH CHEMICAL METERING PUMP
BREAKERS *L GROUT MIXING EQUIPMENT CHEMICAL MIXING TANK
VAPOR EXTRACTION
SYSTEM
fuLLOUT OPeHATOR.S PLATFORM
INJECTION MANIFOLD
VOLK FIELD
EVALUATION CRITERIA
•Reduction
- Total Organic Carbon
- Volatiles
- Oil And Grease
- Chemical Oxygen Demand
- Biological/Chemical Oxygen
Demand
-------
Localions ol Exislinp GtoumJwatcc Monitoring Bo.e Holes at
.Fire Department Training Area.
G
o-ir
N
o
ES
Y
O
Legend
Boundary ol Te»ioift5
Aie*
Bore Hole Location
Groundwatei
Direction
DESCRIPTION OF VOLK ANG
FIRE TRAINING PIT
• Diameter: 75 Feet
• Depth (To Water Table): 12 Feet
• Surface Area-. 4.400 Sq. Feet
• Volume Of Soil: 53.000-Cu. Feet
-------
FACT SHEET
United States
Environmental Protection
Agency
September. 1985
Mobile System For Extracting Spilled Hazardous Materials From Soil
The Hazardous Waste Engineering Research Laboratory, Releases Control
Branch at Edison, NJ, has recently developed a mobile system (or extracting
spilled hazardous materials from soils at cleanup sites.
Landborne spills of hazardous materials that percolate through the soil
pose a serious threat to gjoundwater.
OVI.ICSI/.E
NOK SOIL
MATCH!*!.*
AND DMIMIS
MAKEUP KATF.II ^_
Effective response to such Incidents should Include the means for removing
the contaminants and restoring the soil to Its original condition. Currently
practiced techniques, such as excavation with transfer to land fill or flushing
with water In situ, are beset with difficulties - large land area and volume of
materials Involved. An Innovative In Situ ContalnmentfTreatment System has
been developed to treat contaminated soils. However, It Is not suitable for all
soils and/or all chemicals.
The mobile treatment (see Illustration) has been designed for water extrac-
tion of a broad range of hazardous materials from spill-contaminated soils.
The system will: (1) treat excavated contaminated soils, (2) return the treated
soil to the site, (3) separate the extracted hazardous materials from the
washing fluid for further processing and/or disposal, and (4) decontaminate
process fluids before reclrculatlon, or final disposal. A prototype system has
been developed utilizing conventional equipment for screening, size reduc-
tion, washing, and dewaterlng of the soils. The washing fluid • water • may
contain additives, such as acids, alkalies, detergents, and selected organic
solvents to enhance soil decontamination. The nominal processing rate will
be 3.2-m1 (4-yd*) of contaminated soil per hour when the soil particles are
primarily less than 2-mm In size and up to 14.4-m'(1B-yd') per hour for soil of
larger average particle size. '
For further Information contact Frank J. Freestone or Richard P. Traver,
Hazardous Waste Engineering Research Laboratory, Releases Control
Branch, Edison, NJ. Telephone numbers are: (201) 321-6632/6677 (commer-
cial) or 34f>6632/6677 (FTS).
SPENT CAtlllON
JCESS FLOW SCHEME FOR SOIL SCRUBBER
-------
CALVIN AND HOBBES
WrtW A&E VCU
GO\NGTODKSS
UP kS TOR •
•' HAU.OWEEH?
^s^mmm^m:,
>^:^B^te^iiiifef
-------
ESTIMATED TIME REQUIRED FOR
VOLK IN-SITU WASHING
• Recommended Application Rate
3 Inches Per Day
• 3 Inches Applied On Surface
Fills 10 Inches Of Soil Column
* 14.4 Days Per Pore Volume Or
144 Days Washing To Achieve
80% Removal
MASS BALANCE
What Goes In...
... Must Come Out.
-------
SURFACTANT QUANTITIES REQUIRED
TO REMOVE CONTAMINANTS AT
VOLK ANG
1.5% Surfactant Solution Required
Ten Pore Volumes Used In Pilot
Lab Study To Achieve 80% Removal
One Pore Volume At Volk Fire Training
Pit = 16,000 Cubic Feet
SURFACTANT QUANTITIES REQUIRED
AT VOLK ANG (Continued)
• 10 Pore Volumes = 9.984.000 Ibs.
* 9.984,000 Ibs. x .015 = 150.000 its.
Of Surfactant
-------
TEMPERATURE EFFECT.
TVJR8IOITY v«
3
^
Z
£
0
m
O.-* O.B 1-2 1.8
-1.2 -O.« -0.4
1O
z
L*
-------
TURBIDITY vs TIME
WASHING SURT*CTAMTS_
o BLENO
-------
FLOW DIAGRAM FOR ISCTU
TO ATMOSPHERE
CONTAMINATED
GROUNDWATER
CHEMICAL OR GROUT
• ADDITIVES
INJECTION PUMPS
AND MANIFOLD
TO PHYSICAL-
. CHEMICAL
TREATMENT
UNIT (OPTIONAL)
REINJECTION. SURFACE
APPLICATION OR DISPOSAL
WATER TREATMENT PROCESSES
AT VOLK ANG
* Lime Precipitation
« Clarification
* Aeration
-------
CONTAMINATION LEVELS AT
VOLK ANG
Soil:
Oil & Grease 500.-25.000 mg/fcg
Groundwater:
Volatiles 10-20 mg/l
TOC 100-700 mg/l
AMOUNT OF CONTAMINATION AT
VOLK ANG FIRE TRAINING PIT
* Estimated 52.000 Gallons "Unburnr
After 35 Years Of Operation
* Remaining Contamination Averages
0.2%. Equivalent To 1.700 Gallon
s
SOIL TREATMENT AT VOLK FIELD
* In-situ Washing
- Water
- Surfactants
- Treated/Contaminated
Groundwater
-------
SIMPLIFIED PROCESS FLOW DIAGRAM
SOIL SCRUBBER
AIM CKANCN
CONTAMINATfO
SOIL
SOIUStZf
CLASSIFICATION
SVSTfM
COUNTCft CUNMNT
CHfMICAl
. (XIKACTON
OVfKSIZt
MATIKIALV
OtWIS
STfNT
on
OHVIMC no
NICTCtCO
SOLVtNT
nccvcitii
__ MCCIAIMIO
"* SOILS
SOUtfZATU
MLmOff
C1C
WASTI
SLUOC*
-------
EPA Report Number
November 1987
ROUGH DRAFT
INTERIM REPORT
INVESTIGATION OF FEEDSTOCK PREPARATION AND HANDLING
FOR MOBILE ON-SITE TREATMENT TECHNOLOGIES
by
William F. Beers, P.S.S.
Roy F. Weston, Inc.
EPA Ohmsett Facility
NWS Earle .- Waterfront
Leonardo, New Jersey 07737
68-03-3450
Work Assignment 087208
Richard P. Traver, P.E.
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
Edison, New Jersey 08837
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
2096B
-------
SECTION 1
BACKGROUND
Under the Comprehensive Environmental Response, Compen-
sation, and Liability Act of 1980 (CERCLA), the current
National Contingency Plan (NCP) that implements it, and SARA
(1986) requirements, response actions at hazardous waste sites
must reduce the threat of uncontrolled wastes into the envi-
ronment. In the 1984 Resource Conservation and Recovery Act
(RCRA) Amendments, Congress clearly showed its intent to
minimize the volume of solid waste disposal in landfills. This
policy would mandate a major change in the current practices at
CERCLA sites of removing the hazardous waste material and
burying it elsewhere without any prior treatment.
The policy of the Office of Solid Waste and Emergency
Response (OSWER), responsible for implementing the 1984 HSWA
requirements, is to discourage the excavation and reburial
"disposal" philosophy of CERCLA waste and debris/ and to
encourage technologies to eliminate or reduce the hazardous
character of the waste materials. On-site treatment
technologies that destroy or reduce contaminant levels achieve
more positive control than containment techniques. Off-site
disposal to engineered and protected landfills will only be
allowed in the future when no destruction technology is
available, or for "pretreated" soil and debris materials
complying with Best Demonstrated Available Treatment (BDAT)
levels as promulgated under the impending 1988 Land Ban
legislation. In addition, as landfill disposal becomes more
expensive and as hazardous waste transportation is more
stringently regulated, on-site waste destruction or volumetric
2096B
-------
reduction technologies will be far more desirable—if they are
technologically feasible, environmentally safe, and
economically viable.
In order to destroy or reduce the hazardous character of
any contaminated material, the treatment technology selected
must receive a "feedstock" with a predetermined range of
physical/chemical characteristics to ensure reliable treatment
efficiencies and cost effectiveness. The types of contaminated
materials normally identified and discussed in remedial
investigation/feasibility study (RI/FS) reports are primarily
materials such as soils, sludges, and liquids. The debris
component is not addressed unless the primary contaminated
matrix is a mixture of materials (i.e., building demolition
debris or sanitary landfill type wastes, such as household
trash and garbage).
The land disposal rules, which are scheduled to be enacted
in November 1988, will address feedstock and site debris as
well as contaminated soil under the Land Ban legislation.
Following the review of numerous Records of Decision (RODs)
and RI/FS's, there is a lack of historical site-specific data
quantifying and qualifying Superfund debris. Few, if any RODs
or RI/FS's factor in the operational considerations of han-
dling, segregating, sizing, site excavation, and feedstock
delivery to various recommended mobile on-site technologies
such as biological degradation, chemical treatment (K-Peg),
solidification/stabilization, incineration, low temperature
thermal desorption, and physical treatment (soils washing). It
is critical that an engineering and economic evaluation of the
types of debris and their impacts on these technologies be
2096B
-------
performed, if any form of on-site treatment is ever to be
successfully executed. The current HWERL work assignment
addresses these issues. The objectives of the work assignment
are to:
• Categorize Superfund-related solids, sludges, sedi-
ments, and debris according to excavation, handling,
and separation problems. Data will be provided on
frequency of problem occurrences.
• Provide a written summary on the state-of-the-art
isolation/separation technologies of debris from
feedstock-excavated soils, sediments, and sludges.
• Identify specific handling areas for a detailed
engineering analysis of feedstock preparation for the
following six candidate on-site treatment technologies:
- Incineration
- Low temperature desorption
- Chemical treatment (K-Peg)
- Solidification/stabilization
- Physical treatment (soils washing)
- Biological degradation
• Provide recommendations for future research needed on
technologies that have a high probability of success
and that are applicable to frequently occurring
debris-handling problems.
2096B
-------
SECTION 2
CONCLUSIONS
Debris is anything that cannot be handled by the treatment
process. In general, any material larger than the 1/4 to 6-inch
range presented in Table 1 would be considered debris for all
six technologies being reviewed. Decontamination of debris is
not always possible because of its material nature (absorption
of the contaminant), or because the contaminant should not be
diluted, as in the case of dioxin. Debris cannot always be
subjected to analytical testing to determine its hazard classi-
fication or its level of cleanliness. Debris is only currently
categorised by the participating regulatory agencies as to its
hazardous or nonhazardous status, and handled on a case-by-case
basis. Disposal is currently based on the type of debris,
guantity, contaminant involved, and local/regional regulatory
concerns.
2096B
-------
SECTION 3
RECOMMENDATIONS
RECOMMENDATIONS FOR PROPOSED RULEMAKING
1. Classify material as debris based on the size require-
ments of the recommended technology.
2. Segregate debris for decontamination, recycling and
reuse, incineration, treatment, or land disposal.
3. Treat each site debris situation on a case-by-case
basis with the disposal determination made by the
local regulatory authority (i.e., county, state, or
EPA region).
RECOMMENDATIONS FOR IMMEDIATE RESEARCH NEEDS
1. Modify reporting and site investigations under RI/FS
programs to quantify and qualify the forms and amounts
of debris as presented in Appendix A on both a percent
weight and volume basis.
2. Conduct an engineering review and evaluation of exist-
ing applicable vendor technologies for segregation of
soil and debris for further processing and feedstock
preparation.
2096B
-------
3. Conduct a pilot evaluation of selected feedstock
processing equipment utilizing a standardized nonhaz-
ardous debris matrix as presented in Appendix A.
4. Preparation of a guidance document for use by EPA
Remedial Project Managers, On-scene Coordinators,
contractors, and emergency response personnel identi-
fying mobile and transport separation equipment,
sources, costs (lease/purchase, operation and main-
tenance), debris applications, and anticipated per-
formance.
2096B
-------
SECTION 4
DEBRIS DEFINITION
The six on-site technologies under review include: incin-
eration, low temperature desorption, chemical treatment
(K-Peg), .solidification/stabilization, physical treatment
(soils washing), and biological degradation. Each technology
requires that the feedstock material (soil and debris) be
delivered with predetermined consistencies so that the selected
treatment "hardware" can function and perform reliably in order
to efficiently and cost-effectively destroy or reduce the
contaminants of interest. To. accomplish this task, the contami-
nated material, which may be in the form of soil, sludge,
liquid, or debris, must be prepared by either of the following
means:
• . Physical preprocessing of oversize material (e.g.,
crushing, shredding, screening, separation, dewater-
ing, etc.).
• Chemical preconditioning, such as neutralization or
reduction/oxidation.
Debris can be defined as oversize materials that cannot be
handled by the selected treatment hardware, and may, in fact,
damage the processing equipment.
The types of debris and contaminated materials found at
Super fund sites vary considerably and range in size from
clay-sized particles to large contaminated tanks and buildings.
Debris can be grouped into the following nine general cate-
gories:
2096B
-------
• Cloth
• Glass
• Metals (ferrous/nonferrous)
• Paper
• Plastic
• Rubber
• Wood
• Construction/demolition materials (e.g., concrete,
brick, asphalt)
• Electronic/electrical devices
The nine categories of debris were determined by interviews
with various EPA Regional Superfund Site Managers; EPA Environ-
mental Response Team members; EPA, TAT, REM, FIT consultants; .
and the EPA HWERL Technical Project Managers for various treat-
ment technologies. A detailed breakdown of specific items found
in each debris category is located in Appendix A.
Along with the wide range in the types of debris, the
quantities of debris at sites also vary considerably. It has
been "unofficially" reported through interviews, that the
debris at sites varies on a volumetric basis from less than 1
percent to greater than 80 percent. This is attributed to sites
where demolition debris or sanitary landfill wastes have been
co-disposed with hazardous materials.
A preliminary assessment of each of the six on-site treat-
ment technologies was conducted to determine the maximum size
of debris and material that could be allowed to undergo the
treatment process. The maximum debris size for each technology
based on this preliminary assessment is indicated in Table 1.
8
2096B
-------
TABLE 1. MAXIMUM DEBRIS SIZE/TECHNOLOGY
Maximum debris size Technology
1-2 inches Biological degradation
1 inch Chemical treatment (K-Peg)
6 inches Incineration
1/4 inch Low temperature desorption
2 inches Physical treatment (soil washing)
6 inches Solidification/stabilization
Debris larger than the maximum allowable size must be
segregated from the feedstock material and handled separately.
This oversized material must then be either treated separately
or reduced in size to allow the debris to be refed to the
treatment equipment. \
In addition to debris removal, feedstock preparation may
also include other preparatory steps for the treatment process
to be effective. Feedstock requirements will vary with each
technology and contaminant under consideration. The types of
other feedstock factors that must be identified and evaluated
when considering one of the six technologies include:
• Contaminant concentrations
• pH adjustment
• Moisture content
• Oxidation/reduction status
• Temperature range
• Salt concentrations
• Any special requirements
2096B
-------
The range of contaminant concentrations found in the waste
to be treated must be known to prevent "shock" loading of the
treatment process and also to ensure that the technology
treatment process can handle the contaminant concentrations
identified. Biological degradation is adversely affected by
"shock" loading of toxics. Dechlorination (K-Peg) treatment
processes become noncost effective when contaminant concen-
trations exceed high levels of certain contaminants because of
excessive sodium requirements.
pH adjustment of the wastes may be necessary to reduce cor-
rosion potential; air impacts in incinerators and to ensure
proper growth of microorganisms in the biological degradation
process.
Moisture content also affects certain treatment tech-
nologies because excess moisture can adversely affect reaction
rates and energy input requirements.
The temperature of the waste is an important factor in
rotary kiln incineration because of the potential for thermal
shock due to the moisture content and low temperature of the
waste.
Salt concentrations in the waste under consideration affect
biological degradation processes and immobilization/fixation
treatment. Excessive salts retard or prevent biological growth,
and, in the case of fixation technologies, salts interfere with
the setting and curing times of cement.
Each treatment technology may have other special handling
requirements for various wastes, and these need to be identi-
fied in a detailed engineering analysis of each technology.
10
2096B
-------
SECTION 5
CURRENT DEBRIS HANDLING PRACTICES
The preliminary information collected on debris indicates
that the current handling procedures at hazardous waste sites
range from "elaborate separation and recycling" to "no
separation." Processed material and debris are then handled in
one of the following ways:
• Sent to a secure landfill for ultimate disposal.
• Decontaminated to levels allowing disposal in a
municipal landfill.
• Treated material used for construction foundation
bedding.
• Recycled/reelaimed as a recoverable resource.
• Delisted to a nonhazardous status.
Current debris handling practices have been determined by:
• Technology feedstock requirement.
• Type of contamination.
• Type of debris (size, shape, phase, form, Btu, and
recycle value).
11
2097B
-------
• Quantity of debris (percent volume or weight).
• "Clean-up" standards or target levels (Federal, state,
local, private).
• Potential for decontamination of the debris.
A list of the types of debris and their handling history at
29 Superfund sites is shown in Table 2.
12
2097B
-------
TABLE 2. DEBRIS HANDLING AT SUPERFUND SITES
Site name
1. Kane & Lombard
Recommended
EPA Major clean-up
Contact region contaminant alternative
Charles Kufs III Organics, Incineration
metals Soil washing.
Containment
In-situ vitrification
(ISV)
Debris
types
Concrete
Rocks
Metals
Debris
handling
Presorting
and
shredding
2. Ambler Asbestos Frank Finger
III Asbestos,
CaC03
3. Myers Property Victor Velez II
Organics,
metals
U)
4. Fried Industries Victor Velez II Organics
5. Roebling Steel George Anastos II
Metals.
organics,
asbestos
ISV
Containment
Capping
Off-site
land disposal
None
reported
Solidification/stabi- Pebbles
lization Boulders
Biological degradation Wood
Soil washing Bolts
Off-site land disposal
(untreated waste)
Biological degrada-
tion
Low temperature
thermal stripping
Incineration
Soil washing
FS not done; RI in
progress
Partial emergency
removal action
Drums
Tires
Shredded rubber-
Shredded plastic
Concrete
Baghouse dust
Buildings and
metals
Wire, cables
6. L.A. Clark
Ralph Shapot III Organics
High temperature
thermal stripping
Metals, cyanides
Solidification/ Railroad ties
stabilization Rails, wood
Biological
degradation
(in-situ)
Soil washing
Containment
Concrete,
rocks
2097B
-------
TABLE 2. (CONTINUED)
Site name
EPA
Contact region
Major
contaminant
Recommended
clean-up
alternative
Debris
types
Debris
handling
7. Morgantown
8. Southern MO
9. Cryochem
Ralph Shapot III Organics, Capping
metals Incineration
Jay Motwani III Organics. Biological degrada-
dioxins tion
Incineration
Soil washing
ISV activated
R. Purcell III Organics Work plan stage
Tires
Refrigerators
Wood
Concrete
Cloth
Railroad ties
Rails, wood
Concrete, rocks
No debris
Separation
10. Shaffer —
11. Montgomery Bros. T. Massey
12. Bridgeport Oil 0. Lynch
13 Swiss vale J. Downey
15 Allied-Hopkins —
III PCBs •
III Organics ,
II Oil
Water
III Oioxins,
PCBs
5 Toxaphene,
DDT,
xylene
Methanol extraction
Off-site disposal
Incinerate lagoon
contents
Off-site disposal
in secure land-
fill and recycling
Incineration
Off-site disposal
Tires
Large stones
Drums
Residential
trash
Wood, drums
Tanks
Buildings
Buildings
Metals
Drums
Railroad ties
Rails
Concrete pad
Blocks
Vibratory
screen-
set aside
Off-site
disposal
Clean tanks
Dioxins to
secure land-
fill; steel
decontami-
nated and
recycled to
steel mill
Rails decon-
taminated for
re-use
Railroad ties,
concrete to
secure land-
fill
2097B
-------
TABLE 2. (CONTINUED)
Site name
EPA Major
Contact region contaminant
Recommended
clean-up
alternative
Debris
types
Debris
handling
16. Baird & McGuire Ms. Sanderson
17. Metaltec/ M. Rusin
Aerosystems, NJ
I Cresote.
dioxins
18. Syncon
19. Delaware City G. Chodwick
20. Drake Chemical T. Legel
II TCE
Incineration
Off-site disposal
Heat treatment
Rotary dryer
E. Finnerty II
Pesticides Off-site
PCBs. disposal
metals On-site capping
Tanks
Wood buildings
Masonry
No debris
Large stones
Buildings
Tanks
Piping, heat-
coils
21. Colemon Evans C. Teepen IV PCP
III PVC. TCE Off-site disposal No debris
Reuse of recoverable
product
III Organics Off-site disposal Furniture
and inor- • Piping
ganics
Incineration Miscellaneous
Metal-recycled
Wood-shredded
and incinerated
Masonry-Off-site
disposal
Screening of
stones/rocks
Buildings and
tanks-decontam-
inated for
future use
Piping, etc,
-Off-site
disposal
Reuse of
recoverable
product
Off-site
disposal
Separation
with shredding
and recycling
of metals
22. Hollingsworth E. Zimmermen
23. MowGray J. Trudeau
Engineering
24. Sapp Battery E. Moore
IV TCE, metals Vacuum extraction None
IV PCBs Solidification None
IV Lead,
cadmium
Solidification
Battery cases Crushing
2097B
-------
TABLE 2. (CONTINUED)
Site name
25.
26.
27.
28.
29.
LaSalle
Electrical
Met amor a
Landfill
Geneva
Industries
United
Creosoting
Denver/ROBCO
Recommended
EPA Major clean-up
Contact region contaminant alternative
B. Cat ti che V PCBs
3. Tanaka V VOCs,
metals
D. Williams VI VOCs,
PCBs,
PAHs
0. Williams VI PCPs.
PAHs
J. Brink VIII Radiation
Incineration
Incineration
Off-site disposal
On-going
investigation
Off-site disposal
Wood
Debris
types
Roots,
sticks, stones
Tanks
Prefabicated
buildings
Cracking tower
Houses
Miscellaneous
masonry
Debris
handling
Screening
Off-site
disposal
Clean
Wipe samples
Recycle
Separation
of materials
2097B
-------
SECTION 6
DEBRIS DECONTAMINATION
Once contaminated debris has been separated from the hazard-
ous waste material undergoing treatment, it must either be
disposed of in a secure landfill, stored for future approved
treatment (i.e., dioxin-contaminated material) or decontami-
nated. The determination that debris is contaminated is gener-
ally an assumption that is made with little or no analytical
testing. In some instances monitoring devices such as an Hnu
organic vapor analyzer or a geiger counter are utilized to
determine if a particular object is contaminated with volatile
organic compounds or is radioactive.
Decontamination of debris is possible for contaminants .that
are water soluble and can be washed, rinsed, or are removed
when the associated contaminated soil is cleaned off. Insoluble-
and inorganic (heavy metal)-contaminated fine soil material can
sometimes be successfully separated from debris by high
pressure washing or vibratory separation, allowing the over-
sized material to be safely disposed of. Some contaminants,
such as dioxin, are not generally considered for decontamina-
tion and are designated for interim storage, awaiting either
incineration or alternate approved treatment.
Impervious debris, such as steel, brass, or copper, is
generally decontaminated and recycled, when possible.
17 .
2097B
-------
In most instances debris cannot be subjected to current or
proposed testing procedures (EP toxicity (extraction procedure
toxicity testing), TWA (total waste analysis), and TCLP (toxic
contaminant leaching potential)) to determine if it is hazard-
ous due to the type, form, and surface areas involved. Such
determinations are generally made by the participating regu-
latory parties at the regional, state, and local levels.
Debris that is determined to be nonhazardous can be dis-
posed of as industrial or municipal trash in a sanitary land-
fill. Debris that is deemed hazardous by the regulatory parties
involved must then be incinerated, decontaminated, or otherwise
disposed of in a secure landfill.
18
2097B
-------
SECTION 7
DEBRIS HANDLING EQUIPMENT
Oversize material removal, disaggregation, and material
sorting can be accomplished by physical preprocessing. The
specific type of preprocessing required is dependent on the
technology under consideration and the contaminant involved.
Oversize material removal and debris sorting is generally
accomplished by vibrating or static screens. Magnetic
separation and flotation separation are also utilized in
conjunction with the required screens to obtain a feedstock
with a predetermined consistency.
Grizzlies and hammermills are.used to remove a small amount of
•oversized material from fines. Shredders are used for size
reduction.
A list of preprocessing equipment vendors was assembled and
is included in Appendix B. The applicability of several speci-
fic types of preprocessing equipment should be considered when
completing the detailed engineering analysis of the feedstock
preparation and handling for the six mobile on-site
technologies.
19
2097B
-------
BIBLIOGRAPHY
Ananth, K. P., Shannon, L. J., and Schong, M. P., "Environmental
Assessment of Waste-to-Energy Processes: Source Assessment
Document," Sponsored by Industrial Environmental Research
. Laboratory, Cincinnati, Ohio, Report No. EPA 600/7-77/091,
August. 1977.
Arthurs, J. and Wallin, S. C., "Cadmium in Emissions from
Incineration of Waste at Local Authority Disposal Sites in the
United Kingdom," Proceedings Fourth International Cadmium
Conference, Munich, West Germany, 2-4 March 1983.
Bernheisel, J. F. and Hickert, M. T., "Status of Aluminum and
Glass Recovery by Mechanical Separation," Department of
Energy, Washington, D.C., Report No. DOE NBB-OOA1, September
1983.
Bridgewater, A. V., "Waste Incineration and Pyrolysis," Paper
presented at the EC Congress, Packaging, Recovery and Reuse,
Government and Industrial Viewpoints, Utrecht, Netherlands,
23-24 October 1979, pp 99-115.
Burke, J., "Cannon Shredding of Municipal .Solid Waste'for the
Preparation of Biological Feedstock," Department of Energy,
Washington, D.C., Report No. DOE CE/20088-01, April 1981.
Camp Dresser & McKee, Inc., "Mobile Treatment Technologies for
Superfund Wastes," EPA Contract No. 68-01-7053, September 1986
Castle, C., Bruck, J., Sappington, D., and Erbaugh, M.,
"Research and Development of a Soil Washing System for Use at
Superfund Sites," National Conference on Management of
Uncontrolled Hazardous Waste Sites, 1985.
Colon, F. J., and Kruydenberg, H., "The Mechanical Separation
of Municipal Refuse into Useful Components and Their Appli-
cation as Raw Materials in Industry," Presented at Recycling
World Congress, Basel, Switzerland, 6-8 March 1978.
Crane, G., Elefritz, R. A., Kay, E. L., and Lounan, J. R. ,
"Scrap Tire Disposal Procedures," Rubber Chemistry and
Technology, V51, N3, July-August 1978, pp 577-599.
Dellinger, B., Durnall, D. S., and Hall, D. L., "Laboratory
Determination of High Temperature Decomposition Behavior of
Industrial Organic Materials," Proceedings 75th APCA Annual
Meeting, New Orleans, Louisiana, June 1982.
Dobson, G. R. .and Webb, M., "Economic Assessment of the Energy
Potential of Non-Metallic Fragmentiser Waste," Wassen Spring
Laboratory, England, Report No. LR-480 (MR) - M, 1984.
20
2095B
-------
Ellis, W. , Payne, J., and McNabb, D., "Treatment of Contam-
inated Soils with Aqueous Surfactants," SAI Corporation,
EPA Contract No. 68-03-3113.
Ellis, W. and Fogg, T., Task No. 29, "Interim Report - Treat-
ment of Soils Contaminated with Heavy Metals," SAI Corpora-
tion, EPA Contract No. 68-03-3113.
"Energy Recovery from Refuse-Derived Fuel: Precombustion Pro-
cessing of Refuse Offers a Number of Advantages," Depart-
ment of Energy, Washington, D.C., August 1985.
Esposito, M. P., McArdle, J., Greber, J. S., and Clark, R.,
"Guidelines for. Decontaminating Buildings, Structures and
Equipment at Superfund Sites," 5th National Conference on
Management of Uncontrolled Hazardous Waste Sites, 1984.
Fitzpatrick, V. F. , Buelt, J. L., Oma, K. H., and Timmerman, C.
L., "In Situ Vitrification - A Potential Remedial Action
Technique for Hazardous Wastes," 5th National Conference on
Management of Uncontrolled Hazardous Waste Sites, 1984.
Ford, J., "Handling of Waste Stream Sludges," Process Biochem-
' istry, Vol. 12, No. 5, June 1977, pp 16-17.
"Incineration Disposes of Refinery Wastes," Oil and Gas
"Journal, 17 November 1975, V73, N46.
Institute for Industrial Research and Standards, "Retrospective
Search on Production of Ethanol from Cellulose," Informa-
tion Technology Group, Dublin, Ireland, Report No.
NP5901166, October 1984.
Jones, J. , "Converting Solid Wastes and Residues to Fuel,"
Chemical Engineering, 85(1), January 1978, pp 87-94.
Kavaska, K. E., "Military Wastes-to-Energy Applications," Aero-
space Corporation, Germantown, Maryland, Report No. ATR-80
(8374)-!, November 1980.
Lie, L. X., "Wastewater Treating at Lanzhar," Hydrocarbon
Processing, Vol. 64, No. 6, June 1985, pp 78-79.
"Magnetic Separation of Materials," NTIS, November 1985.
McArdle, J. L. , Arozarera, M. M. , Gallagheo, W. E., and
Optaken, E. F., "Treatment of Hazardous Waste Leachate,"
National Conference on Hazardous Wastes and Hazardous
Materials, 1986.
21
2095B
-------
"Mobile System for Extracting Spilled Hazardous Materials from
Excavated Soils," Reports prepared by Rexnord, Inc.,
Milwaukee, Wisconsin for U.S. EPA Municipal Environmental
Research Laboratory, Report No. EPA 600/2-83, October 1983.
Munoz, H., Cross, F. L. , and Tessiture, J. L., "Comparison
Between Fluidized Bed and Rotary Kiln Incinerators for
Decontamination of PCS Soils/Sediments at CERCLA Sites,"
National Conference on Hazardous Wastes and Hazardous
Materials, 1986.
Nack, H., Lilt, R. D., and Kim, B. C., "Cofixing Coal with Waste
Materials," First Annual Pittsburgh Coal Conference, 17-18
September 1984.
National Waste Processing Conference, 7th Proceedings, Boston,
Massachusetts, Proceedings 1976, Published by ASME, Solid
Waste Processing Division, New York, New York, 1976.
Noland, J. W., McDevitt, N. P., and Koltuniak, D. L., "Low
Temperature Thermal Stripping of Volatile Compounds,"
National Conference on Hazardous Wastes and Hazardous
Materials, 1986.
"Reclamation Center Saves Valuable Landfill Space.by Shredding
Refuse," Solid Wastes Management Refuse Removal Journal,'21
(9), September 1978, pp 28-32.
Roy F. Weston, Inc., "Heavy Metal Contaminated Spil Treatment
Conceptual Development," Interim Technical Report, Prepared
for USATHAMA, Aberdeen Proving Ground, Maryland, February
1987.
Roy F. Weston, Inc. "Economic Evaluation of Alternative Options
for Low Temperature Thermal Stripping of Volatile Organic
Compounds from Soil," Prepared for USATHAMA, Aberdeen
Proving Ground, Maryland, June 1986.
Rullens, W. H. and Assink, J. W., "Extraction as a Method for
Cleaning Contaminated Soil: Possibilities, Problems and
Research," 5th National Conference on Management of
Uncontrolled Hazardous Waste Sites, 1984.
Scofield, F., Levin, J., and Beeland, G., "Assessment of Indus-
trial Hazards Waste Practices, Paint and Allied Products
Industry Contract Solvent Reclaiming Operations, and
Factory Application of Coatings," Report No. EPA 530/
SW-1190, September 1975.
Shea, G. J., "TRV Waste Analysis and Preliminary Operating
Parameters for a Radioactive Waste Slagging Pyrolysis
Incinerator," Sponsored by Department of Energy, Contract
No. EY-76-C-07-1570, March 1978.
22
2095B
-------
Shih, C. C., Cotter, J. E., and Dean, D., "Comparative Cost
Analysis and Environmental Assessment for Disposal of
Organochlorine Wastes," Sponsored by Industrial Environ-
mental Research Laboratory, Research Triangle Park, North
Carolina, Report No. EPA 600/2-78/190, August 1978.
Smith, H. V., "Some Criteria for the Successful Commercial
Recycling of Heterogeneous Plastic Waste," Conservation and
Recycling, V2, N2, 1978, pp 197-201.
Spoorex, P. A., "Stabilization/Solidification Alternatives for
Remedial Action," 6th National Conference on Management of
Uncontrolled Hazardous Waste Sites, 1985.
Stoddard, T. L. and Short, J. J., "Full-Scale Rotary Kiln
Incinerator Field Trial: Phase I, Verification Trial Burn
on Dioxin/Herbicide Orange Contaminated Soil," National
Conference on Hazardous Wastes and Hazardous Materials,
1987.
Store, R., "Municipal Solid Wastes and Their Disposal," Fifth
Los Alamos Life Sciences Symposium, Los Alamos, New Mexico,
October 1977.
Tantore, F. J. and'Richardson, A. W., "Wastewater Residuals
Represent Major Handling Costs, Also Opportunity to Recoup
Investment," Wasteworld News, Vol..2, No. 5, September
1986, pp 14-17.
Taylor, M. R. G. and Patrick, P. K. , "Hazardous Wastes Manage-
ment in Hong Kong - Summary of a Report and Recommenda-
tions," Agricultural, Industrial and Municipal Waste
Management in Today's Environment Conference, 17-18 April
1985.
Thompson, J. D., "Low-Level Waste Institutional Waste Incinera-
tor Program," Department of Energy, Washington, D.C.,
Contract No. EY-76-C-07-1570, April 1980.
Triegel, E. K. , Kolmer, J. R., and Ounanious, D. W., "Solidifi-
cation and Thermal Degradation of TNT Waste Sludges Using
Asphalt Encapsulation," National Conference on Management
of Uncontrolled Hazardous Waste Sites, 1983.
U.S. EPA Handbook, "Remedial Action at Waste Disposal
Sites," Report No. EPA 625/6-86/006, October 1985.
Vrable D. L. and Engles, D. R., "Transportable Circulating Bed
Combustor for the Incineration of Hazardous Waste," 6th
National Conference on Management of Uncontrolled Hazardous
Waste Sites, 1985.
23
2095B
-------
Young, D. A., "Biodegradation of Waste Coolant Fluid," Depart-
ment of Energy, Washington, D.C., Report No. BDX-613-3174,
January 1985.
24
2095B
-------
APPENDIX A
DEBRIS IDENTIFICATION
Cloth Paper
- Rags - Books
- Tarps - Magazines
- Mattresses - Newspaper
- Cardboard
- Packing
Glass ° Plastic
- Bottles - Buckets
- (white, brown, green - Pesticide
clear, blue) containers
- Windows - Six-pack
retainer rings
- Thin plastic
sheets
- Plastic bags
- Battery cases
Ferrous Metals Rubber
- Cast iron - Tires
- Tin cans - Hoses
- Slag - Insulation
- Battery cases
Nonferrous Metals Wood
- Stainless steel - Stumps and
- Aluminum leaves
- Brass - Furniture
- Copper - Pallets
- Slag - Plywood
- Railroad ties
Metal Objects Electronic/Electrical
- Autos/vehicles - Televisions
- 55-gallon drums/containrs - Transformers
- Refrigerators - Capacitors
- Tanks/gas cylinders - Radios
- Pipes
- Nails
- Nuts and bolts
- Wire and cable
- Railroad rails
- Structural steel
25
2095B
-------
Construction Debris
- Bricks
- Concrete blocks
- Asphalt
- Stones and rocks
- Reinforced concrete pipe
- Wood
- Steel beams
- Asbestos insulation and roofing/siding shingles
- Fiberglass insulation
- Fiberglass tanks
26
209 5B
-------
SUPERFUND STANDARD ANALYTICAL REFERENCE MATRIX PREPARATION
AND RESULTS -OF PHYSICAL SOILS WASHING EXPERIMENTS
by: M. Pat Esposito*, Barbara Bruce Locke, and Jack Greber
PEI Associates, Inc.
Cincinnati, Ohio 452A6
and
Richard P. Traver
U.S. EPA, HWERL
ABSTRACT
In response to the RCRA Hazardous and Solid Waste Amendments of 1984 pro-
hibiting the continued land disposal of untreated hazardous wastes, the EPA
has instituted a research program for establishing best demonstrated and
available technologies for RCRA and Superfund wastes. Under Phase I of EPA's
Superfund research program, several projects were initiated under which a
surrogate soil containing a wide range of chemical contaminants was prepared
for use in bench-scale and pilot-scale performance evaluations of five dif-
ferent treatment technologies. This paper covers one of the projects in which
the surrogate test soil was developed and bench-scale soil washing treatabi-
lity studies were completed. This work was conducted by PEI Associates under
EPA Contract No. 68-03-3413 during 1987. This paper has been reviewed in
accordance with the U.S. Environmental Protection Agency's peer and admini-
strative review policies and approved for presentation and publication.
Formerly with PEI, now with Bruck, Hartman & Esposito, Inc., Cincinnati,
Ohio.
-------
INTRODUCTION
The RCRA Hazardous and Solid Waste Amendments of 1984 prohibit the con-
tinued land disposal of untreated hazardous wastes beyond specified dates.
The statute requires the U.S. Environmental Protection Agency (EPA) to set
"levels or methods of treatment, if any, which substantially diminish the tox-
icity of the waste or substantially reduce the likelihood of migration of
hazardous constituents from the waste so that short-term and long-term threats
to human health and the environment are minimized." The legislation sets
forth a series of deadlines beyond which further disposal of particular waste
types is prohibited if the Agency has not set treatment standards under Sec-
tion 3004(m) or determined, based on a case-specific petition, that no further
migration of hazardous constituents will occur for as long as the wastes
remain hazardous.
In addition to addressing future land disposal of specific listed wastes,
the RCRA land disposal restrictions also address the disposal of soil and
debris from CERCLA site response actions. Sections 3004(d)(3) and (e)(3) of
RCRA state that the soil/debris waste material resulting from a Superfund-
financed response action or an enforcement authority response action imple-
mented under Sections 104 and 106 of CERCLA, respectively, will not be subject
to the land ban until November 8, 1988.
Because Superfund soil/debris waste often differs significantly from
other types of hazardous waste, the EPA is developing specific RCRA Section
3004(m) standards or levels applying to the treatment of these wastes. These
standards will be developed through the evaluation of best demonstrated and
available technologies (BDAT). In the future, Superfund wastes in compliance
with these regulations may be deposited in land disposal units; wastes exceed-
ing these levels will be banned from land disposal unless a variance is is-
sued .
In early 1987, EPA's Hazardous Waste Engineering Research Laboratory, at
the request of the Office of Solid Waste, initiated a research program to
evaluate various treatment technologies for contaminated soil and debris from
Superfund sites. Under Phase I of this research program, which was conducted
from April to November 1987, a surrogate soil containing a wide range of
chemical contaminants typically occurring at Superfund sites was prepared for
use across the board in the bench-scale or pilot-scale performance evaluations
of five available treatment technologies: 1) soil washing, 2) chemical treat-
ment (KPEG), 3) thermal desorption, 4) incineration, and 5) stabilization/fix-
ation. This report covers those segments of Phase I related to development of
the surrogate soil and experimental bench-scale tests on the potential effec-
tiveness of physical soil washing as a treatment technology.
PROCEDURES
SARM PREPARATION
The surrogate soil is referred to throughout the text as SARM, an acronym
for Synthetic Analytical Reference Matrix. More than 30,000 pounds of clean
(uncontaminated) SARM was prepared after considerable research into the types
of soils found at Superfund sites nationwide. The final composition selected
-------
consisted of 30 percent by volume clay (a mixture of montmorillinite and
kaolinite), 25 percent silt, 20 percent sand, 20 percent topsoil, and 5 per-
cent gravel. The components were air-dried to minimize moisture and then
mixed together in two 15,000-lb batches in a standard truck-mounted 6-yd3
cement/concrete mixer.
A prescribed list of chemicals found to be widely and frequently oc-
curring at Superfund sites was then added to the clean SARM in a series of
smaller-scale mixing operations utilizing a 15-ft3 mortar mixer. The organic
chemicals added included ethyl benzene, 1,2-dichloroethane, tetrachloroethyl-
ene, acetone, chlorobenzene, styrene, xylene, anthracene, pentachlorophenol,
and bis(2-ethylhexyl) phthalate. Salts or oxides of the following metals were
also added: lead, zinc, cadmium, arsenic, copper, chromium, and nickel.
Because concentrations of contaminants in soils vary widely, four different
SARM formulas containing either high or low levels of organics and metals were
prepared for use in subsequent treatability tests using the five technologies
named. Table 1 presents the target contaminant concentration of the four
SARMs prepared. Reserves of each SARM were also packaged and archived for
future use. The archived samples are being stored at EPA's R&D facility in
Edison, New Jersey.
PHYSICAL SOIL WASHING EXPERIMENTS
As part of the performance evaluation of soil washing as a potential
treatment candidate, samples of each SARM were physically washed in a series
of bench-scale experiments designed to simulate the EPA-developed Mobile Soils
Washing System (MSWS). This system can extract certain contaminants from
soils, which reduces the volume of the contaminated portion of the soils. The
MSWS is expected to be an economic alternative to the current practice of
hauling contaminated soils offsite to a landfill and replacing the excavated
volume with fresh soils.
Specifically, this project was designed to simulate the drum screen
washer segment of the MSWS as described by J.S. Shum in the Operation and
Maintenance Manual(1). This segment of the MSWS separates the +2 mm soil
fraction from the -2 mm soil fraction (fines) by use of a rotary drum screen.
A high-pressure water knife operates at the head of the system to break up
soil lumps and strip the contaminants off the soil particles. Both the design
of the MSWS and the.design of the bench-scale experiments to simulate the MSWS
for cleanup of the SARMS samples are based on the following assumptions, which
underlie the volume reduction approach of physical soils washing:
1. A significant fraction of the contaminants are attached to the silt,
humus, and clay particles.
2r •-—The silt and clay are attached to the sand and gravel by physical
processes (primarily compaction/adhesion).
3. Physical washing of the sand/gravel/rock fraction will effectively
remove the fine sand, silt, and clay-sized (less than 0.25 mm)
materials from the coarse material.
-------
TABLE 1. TARGET CONTAMINANT CONCENTRATIONS .FOR SARKS
(nig/kg)
Analyte
Anthracene
Bis(2-ethylhexyl)
SARM IV
SARM I SARM II SARM III
(High
organic, (Low organic, (Low organic, (High organic,
low metal) low metal) high metal) high metal)
Volatiles
Acetone
Chlorobenzene
1 ,2-Dichloroethane
Ethyl benzene
Styrene
Tetrachl oroethyl ene
Xylene
Semivolatiles
6,800
400
600
3,200
1,000
600
8,200
680
40
60
320
100
60
820
680
40
60
320
100 .
60
. 820
6,800
400
600
3,200
1,000
600
8,200
6,500
650
650
6,500
phthalate
Pentachl orophenol
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
2,500
1,000
10
20
30
190
280
20
450
250
100
10
20
30
190
280
20
450
250
100
500
1,000
1,500
9,500
14,000
1,000
22,500
2,500
1,000
500
1,000
1,500
9,500
14,000
1.000
22,500
-------
4. The contaminants will be removed to the same extent that the silt
and clay are separated (i.e., increasing the efficiency of the
washing process will directly increase the removal efficiency for
the majority of the contaminant mix).
These assumptions were tested by evaluating different wash solutions in
bench-scale shaker-table experiments. The wash solutions chosen.for evalua-
tion included 1) a chelant solution (tetrasodium salt of EDTA, Dow Chemical
Versene 100 ), and 2) an anionic surfactant solution (phosphated formulation
from Procter and Gamble, Institutional Formula Tide ). Different pH and
temperature conditions were evaluated for the wash solutions. Organic sol-
vents and oxidizing agents were considered, but were found not to be viable
soil-washing solutions because of material handling problems associated with
these compounds, especially when used in a field situation. Following the
shaker-table wash, the soil was wet-sieved to separate the fines from the
coarse material. Although the EPA MSWS only separates the soil into +2 mm and
-2 mm size fractions, three size fractions (+2 mm, 250 ym to 2 mm, and -250
urn) were investigated in this study to determine if an intermediate size
fraction (medium to fine sand) could be cleaned effectively, thereby increas-
ing the volume reduction potential. For determination of the effectiveness of
the soil-washing techniques in reducing the volume of contaminated material,
each of the resulting soil fractions was subsequently analyzed for total
organics and metals by standard Gas Chromatography Mass Spectrometry (GC/MS)
and Inductively Coupled Plasma (ICP) techniques (SW-846, 3rd ed.) and for
leachable constituents by Toxicity Characteristic Leaching Procedures (TCLP).
RESULTS
SARM PREPARATION
Results of physical tests conducted on the clean SARM are summarized in
Table 2. These test results indicate that the synthetic soil is characteris-
tic of a slightly alkaline sandy loam with moderate clay and organic content
and a relatively high cation exchange capacity. Such a soil, when contami-
nated, should present a reasonable challenge to any applied treatment tech-
nology.
Chemical analyses of samples of the four SARMs were conducted before
treatment to verify contaminant levels and moisture content. Table 3 contains
the average concentrations obtained for each analyte in each of the four
SARMs. All numbers reported by each laboratory conducting the analyses (five
separate analytical laboratories performed these analyses) were included in
calculating the averages.
If the target contaminant- levels (Table 1) are compared to the actual
levels found (Table 3), the recovery efficiencies obtained are the highest and
most consistent for the metals, followed by the volatiles and the semivola-
tiles. Generally, the SARMs containing the higher concentrations of volatiles
and semivolatiles showed better corelation between the target and the actual
contaminant levels. The results for the lower organic contaminated SARMs
(SARM II and III) seem to indicate either that a greater portion (relative to
the high organic SARMs) of the indicator organics added to the soil were lost
through one or more routes (e.g., volatilization, adsorption), or alterna->
tively, that the lower concentrations of the organics were more difficult to
reliably detect and quantitate.
-------
TABLE 2. PHYSICAL CHARACTERISTICS OF CLEAN SARM
Average3 Range
Cation exchange capacity, meq/100 g
Total organic carbon, %
PH
Grain size distribution, weight %
Gravel (>4.75 mm)
Sand (4.75 mm - 0.075 mm)
Silt (0.074 mm to 0.005 mm)
Clay (<0.005 mm)
132.7 77.5 to 155
(10)
3.2 2.7 to 3.9
(6)
8.5 8.0 to 9.0
(6)
3 2 to 4
(6)
56 54 to 58
(6)
28 27 to 30
12 11 to 14
(6)
a Values in parentheses indicate number of samples analyzed.
-------
TABLE 3. ANALYTICAL PROFILE OF SARMS: AVERAGE CONCENTRATION
FOUND UPON ANALYSIS3
(mg/kg)
Analyte
SARM I
SARM II
SARM III
SARM IV
(High
organic, (Low organic, (Low organic, (High organic,
low metal) low metal) high metal) high metal)
Volatiles
Acetone
Chlorobenzene
1,2-Dichloroethane
Ethyl benzene
Styrene
Tetrachl oroethyl ene
Xylene
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachl orophenol
Metals
Arsenic
Cadmi urn
Chromium
Copper
Lead
Nickel
Zinc
Moisture, %
4,353 (9) 356 t
316 (9) 13 (
354 (9) 7 {
J 358
5 11
J 5
3,329 (9) 123 (8) 144
2)
2)
2)
12)
707 (9) 42 (8) 32 (2)
408 (9) 19 (8) 20 (2)
5,555 (9) 210 (8) 325 (2)
5,361 (9) 353 (7) 181 (3)
1,958 (9) 117 (7) 114 (3)
254 (9) 22 (7) 30 (3)
18 (10) 17 (7) 652 (4)
22 (8) 29 (6) 2,260
24 (8) 28 (6) 1,207
231 (10) 257 (8) 9,082
(2)
:A)
(4)
236 (10) 303 (8) 14,318 (4)
32 (10) 38 (8) 1.489 (4)
484 (8) 642 (6) 31,871
20 (7) 11 (7) 19
(4)
(3)
8,030 2)
330 2)
490 2)
2,708 (2)
630 (2)
902 (2)
5,576 (2)
1,920 (3)
646 (3)
80 (3)
500 (4)
3,631 (2)
1,314 4)
10,503 4)
14,748 4)
1,479 4)
27,060 (4)
26 (2)
Values in parentheses indicate number of samples analyzed.
-------
PHYSICAL SOIL WASHING EXPERIMENTS
During the initial phase of these experiments, pH and temperature varia-
tions were evaluated as well as different chelant and surfactant concentra- ,
tions. Experiments were also run to determine the optimum reaction time for
both the chelant and surfactant solutions. In all cases, a 10:1 wash solution-
to-soil ratio was utilized. Temperature ranges from 78° to 120°F were found
to have little effect on the contaminant reduction efficiencies. Adjustment
of the pH of the surfactant solution from 5.0 to 12.0 resulted in no appreci-
able change in the organic contaminant removal efficiencies. Also, reducing
the pH of the chelant solution from its natural pH of 12 to 8.0 produced no
additional metal removal.
Reaction times of 5, 15 and 30 minutes were evaluated in a series of
trial tests for the chelant and surfactant solutions in order to select the
optimum reaction time for all subsequent testing. Figures 1 and 2 present the
reaction time results for a 1:1 molar ratio (moles of tetrasodium EDTA to
moles of total metals present in the SARM) chelant wash of metals from SARM
III, and for a 0.1 percent (by weight) surfactant wash of organics from SARM
I, respectively. The concentrations used for evaluation of the reaction times
were the lowest concentrations of both chelant and surfactant chosen for
overall evaluation in this study. As shown in Figure 1, no significant addi-
tional metal chelation occurred for SARM III after 15 minutes for any of the
six metals. Therefore, a 15-minute reaction time was chosen for all of the
subsequent chelant wash tests. As shown in Figure 2, no similar completion of
reaction was evident for the organic contaminants (as total organic halogens);
their concentration in the wash water continued to increase over the entire
30-minute interval. Therefore, 30 minutes' was chosen as the reaction time for
all subsequent surfactant washes. Longer reaction times were not evaluated
because reaction times in excess of 30 minutes are typically too costly in
scale-up operations.
Next, surfactant concentrations of 1.5, 0.5, and 0.1 percent (by weight)
were evaluated in a series of 30-minute washing tests of SARM I to determine
the optimum organic contaminant removal efficiency achievable. The tests
showed that the 0.1 percent solution was least effective, and that the 1.5
percent and 0.5 percent concentrations were essentially equal; the results
obtained for the 1.5 percent solution did not indicate sufficient additional
contaminant reduction over the 0.5 percent solution to justify the higher
surfactant concentration. Thus the 0.5 percent surfactant solution was chosen
as the optimum wash concentration for subsequent organics removal tests. Two
molar ratios (moles of tetrasodium EDTA to total moles of metals present in
the higher metal SARM-SARM III) were evaluated for metals removal—1:1 and
3:1. The 3:1 EDTA molar ratio solution exhibited consistently higher removal
efficiencies for the metals, particularly in the middle soil fraction (250 urn
to 2 mm); therefore it was chosen for further study in all subsequent metal
removal tests.
During the second phase of these experiments, the optimum conditions for
reducing organic and metal contamination (as determined in the initial phase
of the soil experiments and discussed in the preceding paragraphs) were ap-
plied to all four SARMs and compared with a baseline plain water wash for each
SARM. Tables 4 through 7 present the results of these final washings. In
-------
2000
•o- Arsenic a
•*- Cadmium
•*• Chromium
•*- 'Copper
•*- Lead
•o- Nickel
-*• Zinc
Time, mln.
Arsenic and nickel overlap in this figure.
Figure 1. Reaction time -1:1 molar chelant wash, SARM
I?
i
•o- TOX
10 20
Tlmt, mln.
a Total organic halogens
30
Figure 2. Reaction time - 0.1% surfactant wash, SARM I
-------
TABLE 4. SOIL WASHING RESULTS: SARM
(ppm)
I (HIGH ORGANICS, LOW METALS)
Contaminant
Initial
concentration
Water Mash
>2 ra> 250 wm to 2 m <250 vm
0.5J Surfactant Mash
>2 mn
250 wm to 2 mm <2SO
Volatile organlcs
Acetone 4.353 10 20 140 22 8.0 50
Chlorobenzene 316 0.028 0.28 160 0.30 1.0 31
1,2-Dtchloroethane 354 <0.023 0.1B 24 0.15 0.32 6.0
Ethyl benzene 3,329 0.13 1.4 2300 2.3 8.5 680
Styrene 707 NDD NO 400 <0.17 NO 96
Tetrachloroethylene 408 0.009 0.12 250 0.20 0.81 49
Xylene 5,555 0.38 3.2 1800 4.0 14 820
Total volatile organic >99.9X 99.8* 66.2% >99.8t 99.81 88.5X
reduction
Senlvolatlle organlcs
Anthracene
61s(2-ethylhexy1)
phthalate
Pentachlorophenol
Total semi volatile
organic reduction
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Total metal reduction
5.361
1,958
254
IB
22
24
231
236
32
484
6.5
4.0
66
98. 9*
3.0
7.3
1.5
10.6
11.1
3.2
44.8
92.2
3200
92
26
56. 2%
5.2
11.3
2.6
30.5
28.8
7.8
106
81.6
1400
1600
53
59. 7%
18.6
28.8
43.4
387
402
35.1
726
NRC
3.3
<6.1
8.4
>99.8t
4.5
6.9
3.0
11.0
10.1
5.1
47.9
91.5
2500
100
4.6
65.61
5.8 '
11.0
3.0
34.6
40.1
6.8
101
80.7
2700
1600
NO
43.21
19.1
26.2
46.8
384 ,
420
31.6
647>
NR
a Fro* Table 3.
b NO - not detected.
c NR • no reduction In overall contamination.
-------
TABLE 5.
SOIL WASHING RESULTS: SARM II (LOW ORGANICS, LOW METALS)
(ppm)
3:1 Molar chelant wash
Contaminant
Mater Mash
Initial
concentration >2 nm 250 urn to 2 mm
>2 m 250 pm to 2 mm <250
0.51 Surfactant wash
>2 iim 250 pm to 2 mm <250pm
Volatile organlcs
Acetone
Chtorobeniene
1,2-Dlchloroethane
Ethyl benzene
Styrene
Tetrachloroethylene
Xylene
Total volatile organic
reduction
Semlvolatlle organlcs
Anthracene
Bis(2-ethylhexyl)
phthai ate
Pentacltlorophenol
Total semlvolatlle
organic reduction
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
356
13
7
123
42
19
210
353
117
22
17
29
28
257
303
38
642
0.50
0.002
N0b
0.014
0.016
ND
0.040
99.91
3.2
27
ND
93.91
2.5
6.0
<0.88
5.0
4.0
4.0
21.0
>96.7t
Total metal reduction
From Table 3.
b NO « not detected.
c NR • no reduction In overall contamination.
0.31
0.013
<0.004
0.082
0.13
<0.004
0.31
>99.91
180
46
6.8
52.71
4.2
10.2
4.0
25.4
69.0
7.2
107
82.71
0.50
<0.23
ND
0.14
0.25
<0.22
0.52
830
370
4.6
NRC
24.8
55.6
90.4
652
710
68.6
1380
NR
0.58
<0.004
NO
0.005
<0.006
ND
0.021
8.8
40
ND
90.11
3.9
2.0
1.6
8.2
6.2
4.2
28.3
95.91
1.2
0.006
0.003
0.058
0.066
<0.004
0.20
>99.81 >99.91 >99.81
210
44
5.1
47.31
4.4
4.0
3.4
15.6
12.6
7.0
63.6
91.61
2.7
0.020
0.003
0.13
0.12
0.009
0.44
99.61
660
260
ND
NR
12.6
7.5
69.7
238
110
43.0
546
21.91
0.46
0.002
ND
0.009
0.010
ND
0.028
99.91
1.6
28
2.4
93.51
3.0
4.8
2.7
9.0
8.5
3.2
25.8
95.71
0.75
0.002
0.004
0.015
<0.013
ND
0.040
>99.9I
120
32
7.8
67.51
3.6
9.4
3.5
28,6
31.8
6.8
112
85.11
1.8
ND
ND
0.62
0.28
<0.30
1.3
>99.41
700
160
ND
NR
27.8
37.7
56.6
478
511
41.8
906
-------
TABLE 6. SOIL WASHING RESULTS: SARM III (HIGH ORGANICS, LOW METALS)
(ppm)
Initial
Contaminant concentration
Volatile organlcs
Acetone
Chlorpbenzene
1,2-pichloroethane
Ethyl benzene
Styrene
Tetrachloroethylene
Xylene
Total volatile organic
reduction
Semivolatile organlcs
Anthracene
Bls(Z-ethylhexyl)
phthalate
Pentachlorophenol
Total semlvolatile
organic reduction
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Total metal reduction
358
11
5
144
32
20
32S
181
114
30
652
2,260
1,207
9,082
14,318
1.489
31.871
Hater wash
* >2 mm 250 urn to 2 mm
0.74
0.008
<0.004
0.040
0.026
0.002
0.10
>99.91
<5.6
2.2
9.2
>94.8t
54.6
372
3.8
68.4
122
IB. 6
558
98. OS
1.7
0.16
0.024
1.3
<0.30
0.16
2.6
>99.3X
480 '
7.4
40
HRC
102
276
14.8
264
491
42.2
1010
96.41
3:1 Molar chelant wash
<250 urn
16
1.6
0.084
34
6.4
3.0
58
86. 7*
1.800
1,100
59
NR
1.160
746
2,590
20,800
30,600
1,570
48,200
NR
>2 m 250 urn to '2
0.96
0.011
0.002
0.054
ND°
0.006
0.091
99.9X
1.7
3.4
«6.6
96. At
36.6
290
3.2
38.6
98.1
17.5
500
98. 4 1
2.6
0.23
0.034
2.0
0.55
0.23
3.6
99. OX
540
9.4
13
NR
51.0
116
9.2
104
171
28.2
519
98.41
mn <250 wm
3.3
1.2
<0.050
20
3.0
2.2
31
>93.2X
1.800
790
<96
NR
243
110
1940
2250
1470
472
6760
78. 2t '
From Table 3.
NO • not detected.
NR • no reduction In overall contamination.
-------
TABLE 7. SOIL WASHING RESULTS: SARM IV (HIGH ORGANICS, HIGH METALS)
(ppm)
Contaminant
Volatile organics
Acetone
Chlorobenzene
1,2-Dkhloroe thane
Ethyl benzene
Styrene
Tetrachloroethylene
Xylene
Total volatile organic
reduction
Semivolatile organics
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorophenol
Total semivolatlle
organic reduction
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
Total metal reduction
Initial
concentration
8.030
328
490
2.708
630
902
5.576
1.920
646
80
500
3.631
1.314
10,503
14.748
1.479
27.060
>2 mm
5.8
0.020
0.028
0.080
NDD
<0.017
0.18
>99.9t
28
5.8
23
97. 8t
126
348
7.7
148
168
29.8
873
97.lt
Water Mash
250 urn to 2 mm
5.8
1.5
<0.34
15
<2.8
2.3
25
>99.7t
2700
34
39
NRe
no
286
29.0
467
1260
56.4
3320
90. 7 1
<250 wm
120
68
<8.6
2.000
150
120
3.200
>69.6t
•
5.200
3.100
360
NR
924
643
2.180
18.400
23,900
1,240
36.200
NR
3:1
>2 mm 250
16
0.012
<0.004
0.051
<0.026
0.006
0.11
>99.9t
40
9.6
8.4
97. 8t
6>4
-279
6.4
80.6
103
19.4
558
98.11
Molar chelant Mash
iim to 2
21
1.4
<0.54
12
NO
1.4
23
>99.7t
1700
70
22
32.31
91.7
210
29.8
332
272
70.7
4730
90.31
mm <250 ym
180
99
40
1000
200
170
1700
81. 8t
3300
2800
<180
NR
180
107
1480
1990
1360
284
5160
82. 2 1
>2 mm
14
0.076
0.10
0.52
0.17
0.048
0.86
99.91
2.4
<5.6
38
>98.3t
30
308
5.9
63.1
68.4
14
462
98. «
0.5t Surfactant Mash
250 ym to
15
0.94
<0.30
7.4
NO
1.1
26
>99.7t
1800
26
42
29.4t
110
336
32.5
446
818
62.9
3040
91.81
2 mm <250 ym
53
22
4.4
300
54
45
460
95. Ot
5,800
1,500
100
NR
538
739
1.500
11,100
15.000
618
25,400
7.3t
From Table 3.
NO • not detected.
NR • no reduction In overall contamination.
-------
general, the cleaning results of the water wash, the 3:1 molar chelant wash,
and 0.5 percent surfactant wash for the +2 mm soil fraction did not differ
significantly. As hypothesized, the silt and clay particles appeared to be
attached to the sand and gravel primarily by physical processes such as com-
paction and adhesion. These physical attractions are often related to the age
of the soil and the contact time between the contaminants and soil particles.
Because the SARM was a freshly prepared soil that had not been compacted,
weathered, and aged, the physical forces of attraction are believed to have
been relatively weak, a condition more typical of a spill site soil than an
older soil found at an abandoned CERCLA site. Consequently, the water wash
was as effective in cleaning the +2 mm soil fraction as the water-plus-addi-
tive solutions were.
Removal of contaminants from the medium-grained fraction (250 ym to 2 mm)
appears to entail both physical and chemical processes. By nature, this
middle soil fraction, which is composed of medium to fine sand, does not
absorb contaminants to the degree that clays and silts do. It has more sur-
face area, however, and should be somewhat harder to clean than the coarse +2
mm fraction. A comparison of the water wash with the 3:1 molar chelant wash
showed that the chelant wash reduced the residual concentration of metals in
the medium soil size class for each SARM subjected to the chelant wash (SARM
II, III, and IV). This trend is especially apparent in the data for SARM II
(Table 5) where the total residual metal reduction increased from 82.7 percent
for the water wash to 91.6 percent after the chelant wash. The organics show
less variation among experimental runs in this soil size class. For the most
part, water was as effective as the surfactant wash for reducing the level of
organic contamination. The one anomaly was anthracene, which showed very high
concentrations in the medium soil class. The anthracene evidently was not
fully dissolved before it was added to the SARM; flakes of what was believed
to be anthracene were observed on the 250 pm screen during the washing experi--
ments.
Reduction of contaminants appears to be affected more by the use of
different wash solutions in the fine soil fraction (less than 250 urn) than in
the other soil fractions. Contaminants are typically bound by both chemical
and physical processes in fine soil fraction. As shown in Tables 5 through 7,
the chelant wash significantly reduced metal contamination in the fine soil
fraction. This reduction is particularly evident in Tables 6 and 7, which
present the results for the SARMs initially high in metal content. Although
the spent wash water was not analyzed, it can be assumed that the chelant
effectively mobilized the metals into solution. Similarly, the surfactant
wash significantly reduced the volatile organic contamination in the fine soil
fraction, as evidenced by the results shown in Tables 4 and 7 for the high-
organic-content SARMs. Again, the wash water was not analyzed; however, it
can be assumed that the surfactant successfully mobilized the organics into
solution.
The trends indicated by the results of the TCLP analysis were similar to
those shown in Tables 4 through 7. In general, reduction efficiencies ranging
from 93 to 99 percent were obtained in the TCLP analysis of volatile organics,
semi-volatile organics, and metals for the top two soil fractions (+2 mm and 2
mm to 250 urn). Most of the TCLP contaminants present in the +2 mm soil frac-
tions dropped below the proposed regulatory limit given in the Federal Register,
Volume 51, No. 114, June 13, 1986. In the SARMs containing lower levels of
-------
metals (specifically SARM I and II), the middle soil fraction (2 mm to 250 urn)
also exhibited concentrations below the proposed TCLP levels.
CONCLUSIONS AND RECOMMENDATIONS
SARM PREPARATION
The preparation of a standard synthetic surrogate soil with physical
characteristics and contaminant levels representative of a wide range of
conditions typically found at Superfund sites was successfully completed. The
surrogate or SARM was subsequently utilized in evaluating the relative effec-
tiveness of five selected treatment technologies (physical soil washing,
chemical treatment, stabilization, low temperature thermal desorption, and
incineration), and a soil treatability data base has now been established.
Further studies comparing the treatability results that were obtained
with the SARM to results from similarly designed studies using actual site
soils are needed to further supplement the data base. Also, future studies in
which the SARM is used to evaluate the relative effectiveness of other pro-
posed treatment technologies at Superfund sites would be valuable.
PHYSICAL SOIL WASHING EXPERIMENTS
The soil washing results from this study appear to support the basic
assumptions underlying the volume-reduction approach to site remediation—that
a significant fraction of the contaminants in contaminated soils are attached
to the smaller sized particles or fines (i.e., silt, humus, and clay) and that
the coarse material can be cleaned and returned to the site by physically
washing and separating it form the fines. The data indicate that water alone
can efficiently remove a significant portion of the contamination from the +2
mm soil fraction. Contaminant removal from the middle (2 mm to 250 jam) soil
fraction and the fine (<250 vim) soil fraction, however, can be generally
enhanced by chelant and surfactant solutions. Addition of a chelant to the
wash solution can improve metal reduction efficiencies for both the medium and
small particle size fractions. Addition of a surfactant to the wash solution
can lead to higher organic removals (compared with the water wash) from the
fine particles. In general, water appears to be more effective in mobilizing
the organics into solution than'in mobilizing the metals.
In the preliminary bench-scale experiments, it was determined that the
SARM was approximately 13 percent (by weight) coarse material (i.e., >2 mm),
29 percent medium-grained material (250 pm to 2 mm), and 58 percent fines
(<250 urn). Therefore, the data presented in Tables 4 through 7 indicate
achievement of at least a 13 percent weight reduction of contaminated material
with a water wash alone. Addition of a chelant solution resulted in a 42
percent reduction by weight of the metal-contaminated SARM, and use of the
chelant and surfactant solutions resulted in lower metal and organic contam-
ination, respectively, in the fine particles.
The mix of contaminants in Superfund soils often lends itself to an
extraction or washing treatment technology such as that demonstrated in this
study. Although promising results have already been achieved at the pilot
scale at a number of lead-contaminated Superfund sites, additional research is
needed to demonstrate the cost-effectiveness of soil washing for full-scale
-------
treatment of a wide range of metal- and organic-contaminated soils. Specif-
ically, most of the research conducted to date has involved demonstration of
the operation of various pieces of equipment for pretreatment and extraction
of the contaminants from the soil and for post-treatment of the extractant.
The effective separation of the wash solution from the soil, the recycling of
the regenerated wash solution, and the concentration/destruction of the con-
taminants, however, have not been demonstrated at a large-scale pilot facil-
ity (2). The following is a listing of areas in which future work is needed
with respect to the development of soil washing as a full-scale, viable treat-
ment option for Superfund soils:
1. Laboratory feasibility studies for evaluating removal of contaminants
from the wash water.
2. Laboratory-scale physical soil washing studies using actual Superfund
soils containing a mix of metal and organic contamination. (The first
study of this type is currently funded and should begin in the spring of
1988.)
3. Evaluation of sequential wash solutions for reducing combined organic and
metal contamination.
4. Additional pilot-scale studies on the use of the EPA Mobile Soil Washing
System.
5. Bench-scale feasibility studies evaluating stabilization/solidification
effectiveness as a treatment train option for the concentrated fines
remaining after soil washing.
6. Evaluation of feed stock preparation methods for the EPA Mobile Soil
Washing System.
REFERENCES
1. Shum, J. S. Drum Screen Washer Operation and Maintenance Manual. Prepared
for the U.S. Environmental Protection Agency, Hazardous Waste Engineering
Research Laboratory, Releases Control Branch, by Mason & Hanger-Silas
Mason Company, Inc., under Contract No. 68-03-3203. February 1987.
2. Dietz, D. H., et al. Cleaning Contaminated Excavated Soil Using Extraction
Agents (Draft). Prepared for the U.S. Environmental Protection Agency,
Hazardous Waste Engineering Research Laboratory, by Foster Wheeler Corpora-
tion, under Contract No. 68-03-3255. September 1986.
-------
88-6B.5
RESULTS OF TREATMENT EVALUATIONS OF CONTAMINATED SOILS
Pat Esposito*
Judy Hessling
Barbara Bruce Locke
Michael Taylor, Ph.D.
Michael Szabo
PEI Associates, Inc.
11499 Chester Road
Cincinnati, Oh.io 45246
Robert Thurnau
Charles Rogers
Richard Traver, P.E.
Edwin Barth
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
26 W. Martin Luther King Drive
Cincinnati, Ohio 45268
INTRODUCTION
The RCP.A Hazardous and Solid Waste Amendments of 1984 prohibit the
continued land disposal of untreated hazardous wastes beyond specified dates.
The statute requires the U.S. Environmental Protection Agency (EPA) to set
"levels or methods of treatmert, if any, which substantially diminish the
toxicity of the waste or substantially reduce the likelihood of migration of
hazardous constituents from the waste so that short-term and long-term
threats to human health and the environment are minimized." The legislation
sets forth a series of deadlines beyond which further disposal of untreated
wastes is prohibited. Specifically, Sections 3004(d)(3) and (e)(3) require
solid/debris waste material resulting from a Superfund-financed response
action or an enforcement authority response action implemented under Sections
104 and 106 of CERCLA,** respectively, to become subject to the land ban on
November 8, 1988.
In response to this mandate, the EPA Office of Solid Waste and Emergency
Response (OSWER) is developing standards for the treatment of these wastes.
These standards will establish treatment levels through the evaluation of
readily available treatment technologies. In the future, Superfund wastes
meeting these levels or standards may be deposited in land disposal units;
otherwise, they will be banned from land disposal unless a variance is issued.
EPA's Office of Research and Development has initiated a research program to
Formerly with PEI Associates, Inc., now with Bruck, Hartman, & Esposito,
Inc.
Comprehensive Environmental Response, Compensation, and Liability Act
-------
88-68.5
identify and evaluate readily available treatment technologies for contam-
inated Superfund soils.
Under Phase I of EPA's research program, which was conducted from April
to November 1987, a surrogate soil containing a wide range of chejnical con-
taminants typically occurring at Superfund sites was prepared and" subjected
to bench- or pilot-scale performance evaluations using the following treat-
ment technologies: 1) physical separation/volume reduction (soil washing),
2) chemical treatment (specifically, KPEG), 3) thermal desorption, 4) in-
cineration, and 5) stabilization/fixation. This report covers the formu-
lation and development of the surrogate soil; it also highlights the results
of the five treatment evaluations. It is worth noting that virtually all of
the analytical data underlying this research were developed using EPA-SW846
methods. Detailed project reports covering the findings of each study are
available through EPA's Hazardous Waste Engineering Research Laboratory in
Cincinnati (see acknowledgments for contact names).
PREPARATION OF SURROGATE SOIL (SARM)
SARM, an acronym for synthetic analytical reference matrix, is the
term used throughout this text to refer to the synthetic soil. The decision
to use a synthetic soil was driven by several factors. First, RCRA permit
regulations restricting off-site treatment of hazardous wastes, such as
contamination Superfund site soils, limited the planned research program.
Second, there Was a strong desire for the test soil to be broadly represen-
tative of a wide range of soils and contaminants, and it was felt that no
single site soil could adequately satisfy this need. Third, large quantities
of a horogeneous test material were needed for the-research program, parti-
cularly for incineration, which was to be ^yjluaced using pilot-scale equip-
ment (requiring thousands of pounds of feed stock). Fourth, it was important
to have contaminants present in the soil at sufficient levels to determine at
least 99 percent reduction efficiencies. Fifth, the contaminants had to
include both metals and organics, and the organics had to include compounds
representing a wide variety of structural types (e.g., both chlorinated and
nonchlorinated alphatics and aromatics, volatiles and semivolatiles, etc.).
Sixth, the soil with its mix of contaminants had to present a reasonable
challenge to the technologies of interest.
The basic formula for the SARM soil was determined from an extensive
review of 86 Records of Decision (ROD's) and a parallel independent study of
the composition of eastern U.S. soils. The recommendations of both sets of
data came to almost the same conclusion: 30 percent by volume of clay (mont-
morillinite and kaolinite), 25 percent silt, 20 percent sand, 20 percent top
soil, and 5 percent gravel. These components were assembled, air-dried, and
mixed together in two 15,000-1b batches in a standard truck-mounted cement
mixer.
Also, as part of the background work, the ROD's were studied to deter-
mine the occurrence, frequency, and concentration of more than 1000 contami-
nants found on Superfund sites. The objective of this effort was to identify
contaminant groups, and indicator chemicals for those groups, that were most
representative of CERCLA wastes.
-------
88-6B.5
The three basic contaminant groups identified as being frequently found
in Superfund site soil and debris were volatile organics, semi volatile organ-
ics, and metals. The selection of specific compounds to serve as representa-
tive analytes for each contaminant group was based on an analysis of specific
site contaminants and their occurrence, as well as the physical and chemical
properties of each compound, including:
0 Molecular structure
0 Vapor pressure
0 Heat of vaporization
0 Heat of combustion
Solubility
° Henry's Law constant
0 Partition coefficient
0 Soil adsorption coefficient
Health effects and toxicity were also taken into account during the
selection process.
As a result of this research effort, a list of target contaminant com-
pounds was developed that represented the most frequently occurring hazardous
compounds at Superfund sites, and that also provided a challenging test
matrix for all five treatment technologies. The final list of chemical
contaminants chosen for the SARM studies is as follows:
Volatile organics Metals
Ethyl benzene Lead
Xyler.e Zinc
1,2-Dichloroethane Cadmium
Tetrachloroethylene Arsenic
Acetone Copper
Chlorobenzene Chromium
Styrene Nickel
Semivolatile organics
Anthracene
Pentachlorophenol
Bis(2-ethylhexyl)phthalate
The final step in this research process was to examine the levels at
which these chemicals have been found at Superfund sites and to select concen-
trations that would be representative of contaminated soils and debris. The
EPA compiled average and maximum concentrations of each selected chemical and
calculated the percentage of each compound within its group. From these
data, target contaminant concentrations were devised for formulating four
different SARM preparations:
SARM 1: High levels of organics (20,800 ppm volatiles plus 10,000 ppm
semivolatiles) and low levels of metals (1,000 ppm total metals).
-------
88-68.5
SARM 2: Low levels of organics (2,080 ppm volatiles plus 1,000 ppm semi-
volatiles) and low levels of metals (1,000 ppm total metals).
SARM 3: Low levels of organics (2,080 ppm volatiles plus 1,000 .ppm semi-
volatiles) and high levels of metals (50,000 ppm total metals).
SARM 4: High levels of organics (20,800 ppm volatiles plus 10,000 ppm
semivolatiles) and high levels of metals (50,000 ppm total metals).
Table I presents the selected target levels for each of the contaminants in
each of the four SARM's.
More than 28,000 pounds of SARM samples were prepared through a series
of small-scale mixing operations utilizing commercial stocks of chemicals,
the clean SARM soil, and a 15-ft3 mortar mixer. Batches of each SARM were
prepared in 500-lb quantities sufficient to meet the needs of each treatment
technology. Only a few pounds of each SARM was necessary for most of the
technologies because they were conducted at bench scale; however, incinera-
tion was evaluated at pilot scale, and therefore required thousands of pounds
of SARM to serve as feed stock for the testing. More than 200 Ib of each
SARM was also reserved, packaged, and archived for future use. The archived
samples are currently being stored at EPA's R&D facility in Edison, New
Jersey, and are available to serve as standard test material for future
treatability studies.
A number of chemical and physical analyses of the basic SARM soil and
the four spiked SARM formulas have been conducted to verify their composition
prior to treatability testing.- Results of the physical and chemical analyses
are compiled in Tables II through IV. Toxicity characteristic leaching
procedure (TCLP) data were also generated during the study, but space limita-
tions prevent their being presented here. These data can be found in the
individual EPA project reports.
METHODOLOGY AND RESULTS OF TREATMENT EVALUATIONS
Physical Separation/Volume Reduction (Soil Washing)
As part of the performance evaluation of this technology, samples of
each SARM were physically treated in a series of bench-scale washing experi-
ments designed to simulate the EPA-developed pilot-scale Mobile Soils Washing
System (MSWS). This system physically separates contaminated fines from
coarse soil material, which effectively reduces the volume of the contaminat-
ed portion of the soils. The MSWS is expected to be an economic alternative
to the current practice of hauling contaminated soils offsite to a landfill
and replacing the excavated volume with fresh soils. The use of a soil
washing system also performs the task of feedstock preparation for other
subsequent treatment technologies by prescreening the soil into a "smooth"
homogenous feed.
Specifically, this project was designed to simulate the drum-screen
washer segment of the MSWS. This segment separates the >2-mm soil fraction
(coarse material) from the <2-m\ soil fraction (fines) by use of a rotary
drum screen. A high-pressure water knife operates at the head of the system
4
-------
88-6B.5
TABLE I. TARGET CONTAMINANT CONCENTRATIONS.FOR SARMS.
(mg/kg)
Analyte
SARM IV
SARM I SARM II SARM III
(High
organic, (Low organic, (Low organic, (High organic,
low metai) low metal) high metal) high metal)
Volatiles
Acetone
Chlorobenzene
1,2-Dichloroethane
Ethylbenzene
Styrene
Tetrachloroethylene
Xylene
Semivolatiles
Anthracene
Bis(2-ethylhexyl)
phthalate
Pentachlorophenol
Inorganics
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
6,800
400
600
3,200
1,000
600
8,200
6,500
2,500
1,000
10
20
30
190
280
20
450
680
40
60
320
100
60
820
650
250
100
10
20
30
190
280
20
450
680
40
60
320
100
60
820
650
250
100
500
1,000
1,500
9,500
14,000
1,000
22,500
800
400
600
200
000
600
8,200
3,
1,
6,500
2,500
1,000
500
1,000
1,500
9,500
14,000
1,000
22,500
-------
TABLE II. Results of Clean Soil Matrix Analyses/
Sample and Batch Numbers
Sample
Batch
Cation exchange
capacity, meq 100/g
TOC, %
pH, S.U.
Grain size
distribution, %
Gravel
Sand
Silt
Clay
1
1
117.5
3.2
8.0
3
55
29
13
2
2
152.5
3.9
9.0
2
57
30
11
3
2
150
3.0
8.5
4
58
27
11
4
1
150
3.8
8.5
3
54
30
13
5
1
77.5
2.8
9.0
2
56
28
14
6 7 8 9 10
22112
150 155 80 147.5 147.5
2.7 -b - -
8.0
3
57
27 - -
13
Average
133
3.2
8.5
3
56
28
12
The clean SARM was also analyzed for all contaminants on the Hazardous Substances List to determine
background contamination, if any. Organic analyses showed no volatile or semivolatile compounds at
the micrograms/kilogram level; metals analyses showed appreciable quantities of iron, potassium,
aluminum, calcium, and magnesium (as would be expected), but no substantial amounts of the more toxic
metals (e.g., chrome, nickel, lead, zinc). In other words, the clean SARM was found to be free of
anthropogenic contamination.
A dash indicates that the sample was not analyzed for this parameter.
CO
CO
I
CTi
CO
-------
TABLE III. MOISTURE CONTENT OF SPIKED SARMSC
(percentage)
88-6B.5
Laboratory
IT Corp.
SARM-I
16.9
SARM-I I
6.0b
SARM-I 1 1
•»•>
SARM-IV
— —
Method
Oven-dried
(thermal desorption
for PEI)
Hittman-Ebasco 31.4
(stabilization
for Acurex)
Radian Corp. 17.1
(incineration 16.1
for PEI) 16.1
EPA - Edison 22.9
(soil washing 19.6
for PEI)
Analytical Enter-
pri ses
(KPEG for Wright
State)
8.61
16.0,
17.8;
17.6C
7 7
t
6.2E
19.3
20.6
18.6
22.1 Oven-dried
Oven-dried
Oven-dried
Oven-dried
30.1 Oven-dried
Dean Stark
distillation
Average (all values) 20.0
11.3.
7.0
17.1{
19.5
26.1
Values obtained by the oven-drying method (ASTM D2216) are expressed as
percent total moisture (i.e., water plus volatile organics); values ob-
tained by Dean Stark distillation Method (ASTM D95) represent percent
water only.
These values are for aliquots taken only from Batch 1 of SARM-II, to which
only a small amount of water was added. See footnote C.
c These values are for subsequent batches of SARM-II, which were prepared
with a higher water content, similar to that added to other SARMs.
-------
S8-6B.5
TABLE IV. ANALYTICAL PROFILE OF SPIKED SARM'S:
AVERAGE CONCENTRATIONS FOUND UPON TOTAL WASTE ANALYSIS'
(mg/kg)
Analyte
Volatiles
Acetone
Chlorobenzene
1 ,2-Dichloroethane
Ethyl benzene
Styrene
Tetrachl oroethyl ene
Xylene
Semivolatiles
SARM I
(High
organic,
low metal )
4,353 (9)
316 (9)
354 (9)
3,329 (9)
707 9)
408 9)
5,555 9)
SARM II
(Low organic,
low metal)
356 (8)
13 (6)
. 7 (8)
123 (8)
42 8)
19 8)
210 8)
SARM III
SARM IV
(Low organic, (High organic,
high metal) high metal)
358 (2
11 (2
5 (2
144 (2
32 (2
20 (2
325 (2
) 8,030 (2)
) 330 (2)
) 490 (2)-
2,708 (2)
630 (2)
902 (2)
5, 576 .(2)
Anthracene
Bis(2-ethylhexyl)
5,361 (9)
353 (7)
181 (3)
1,920 (3)
phthalate
Pentachlorophenol
Inorganics
Arsenic
Cadmi urn
Chromium
Copper
Lead
Nickel
Zinc
Moisture, %
1,958 (9)
254 (9)
18 (10)
22 (8)
24 (8)
231 (10)
236 (10)
32 (10)
484 (8)
20 (7)
117 (7) 114 (3) 646 (3)
22 (7) 30 (3) 80 (3)
17 (7) 652 (4) 500 4)
29 (6) 2,260 (2
28 (6) 1,207 (41
3,631 2)
> 1,314 4)
257 (8) 9,082 (4) 10,503 (4)
303 (8
38 8
642 (6
14,318 (4
1,489 (4
31,871 (4
14,748 4)
1,479 4)
27,060 4)
11 (7) 19 (3) 26 (2)
Values in parentheses indicate number of samples analyzed.
8
-------
88-6B.5
to break up soil lumps and strip the contaminants off the soil particles.
Both the design of the pilot-scale MSWS and the design of the bench-scale
experiments to simulate the MSWS for cleanup of the SARMS samples are based
on a set of assumptions that underlie the volume-reduction approach of treat-
ing contaminated soil, i.e.:
1) A significant fraction of the contaminants are either physically or
chemically bound to the silt, humus, and clay particles.
2) The silt and clay are attached to the sand and gravel by physical
processes (primarily compaction/adhesion).
3) Physical washing of the sand/gravel/rock fraction will effectively
remove the fine sand, silt, and clay-sized (less than 0.2 mm)
materials from the coarse material.
4) The contaminants will be removed to the same extent that the silt
and clay are separated from the sand/gravel/rock fraction (i.e.,
increasing the efficiency of the washing process will directly
increase the removal efficiency for the majority of the contaminant
mix).
These assumptions were tested by evaluating different wash solutions in
a series of bench-scale shaker-table experiments. Two wash solutions were
chosen for evaluation: 1) a chelant solution (tetrasodium salt of EDTA, Dow
Chemical Versene 100 ), and 2) an anionic surfactant solutionp(phosphated
formulation from Procter & Gamble, Institutional Formula Tide'). Organic
solvents and oxidizing agents were considered, but were found unacceptable
because of material-handling problems associated with these compounds, espe-
cially when used in a field situation. Following shakar-table washing, each
SARM soil was wet-sieved to separate the fines from the coarse material.
Although the EPA MSWS only separates the soil into >;2-mm and <2-mm size frac-
tions, three size fractions (^2-mm, 250-ym to 2-mm, and <250-ym) were investi-
gated in this study to determine if the middle fraction (medium to fine sand)
could be cleaned effectively and thereby increase the potential volume reduc-
tion. For determination of the effectiveness of the soil-washing techniques
in reducing the volume of contaminated material, each individual treated size
fraction was analyzed for residual total organics and metals by standard gas
chromatography/mass spectrometry (GC/MS) and inductively coupled plasma (ICP)
techniques (SW-846, 3rd ed.), and for leachable constituents by toxicity
characteristic leaching procedures (TCLP, Federal Register Vol. 51, No. 114,
June 13, 1986).
The soil-washing experiments were conducted in two phases. During the
initial phase, pH and temperature variations were evaluated as well as dif-
ferent wash concentrations of chelant and surfactant. Experiments were also
run to determine the optimum reaction time for both the chelant and surfac-
tant solutions. Temperature ranges from 78° to 120°F had little effect en
the contaminant reduction efficiencies. The pH of the surfactant solution
was adjusted from 5.0 to 12.0 with no appreciable change in the organic
contaminant removal efficiencies. A reduction of the pH of the chelant
solution to 8.0 produced no additional metal removal (ambient pH of the
chelant solution was 12.0).
-------
88-6B.5
The optimum chelant concentration was determined to be a 3:1 molar ratio
of tetrasodium EDTA to total contaminant metals present in the SARM. A
surfactant solution of 0.5 percent (by weight) proved to be most effective in
removing the organic contaminants. Reaction times of 15 minutes^for the
chelant solution and 30 minutes for the surfactant solution were"determined
to be optimum for allowing sufficient contact between the solution and soil
matrix.
During the second phase of these experiments, the optimum conditions for
reducing organic and metal contamination (as determined in the initial phase
of the soil experiments and discussed in the preceding paragraphs) were
applied to all four SARM's and compared with a baseline tap-water wash for
each SARM. Tables V through VII show an approximation of the effectiveness
of various treatment solutions (wash solutions) by presenting the overall
removal efficiencies observed for each size fraction and contaminant group.
These efficiencies, which are expressed as percentage reductions, were devel-
oped by dividing the residual contaminant concentration in each size fraction
by the initial concentration in the whole soil. Although this comparison is
admittedly imprecise, it is nevertheless useful for demonstrating trends and
relationships between soil fractions, contaminant types, and waste solutions.
The discussion that follows examines the data according to the results ac-
hieved for each soil size fraction. \
The data underlying Tables V through VII clearly showed the tendency for
contaminants to accumulate or concentrate in the smaller size fractions
(i.e., to'bind to the clay and silt). For nearly all of the contaminants,
the concentration increased as the size fraction decreased. This finding is
consistent with the findings of earlier soil-washing tests.1'2'3
For the ^2-mrn soil fraction (see Table V), the water wash, the 3:1 molar
chelant wash, and 0.5 percent surfactant wash were all about equally effective.
In all cases, overall contaminant removal efficiencies by group exceeded 90
percent, and volatile removals as a whole exceeded 99 percent across the
board. Semi volatile removals ranged from 90 to 99+ percent, and metals from
92 to 98 percent. Individual contaminant removal efficiencies within groups
varied somewhat. These variations are probably due to physical properties
associated with each contaminant (such as water solubility, volatility,
polarity, etc.), as well as physical properties of the soil (e.g., cation
exchange capacity, surface area) and the wash solution itself (pH, tempera-
ture, chelant, surfactant concentration, contact time, etc.). These excel-
lent results are believed to be closely related to the "freshness" of the
soil. It has been hypothesized that the physical processes of compaction and
adhesion were not highly operative in the SARM soils, which allowed the
loosely attached silt and clay particles to be easily separated from the
larger sand and gravel fractions. These physical attractions tend to be more
operative in older soils, and are especially noticeable in soils that have
experienced long periods of weathering and contact time between contaminants
and soil particles. Because the SARM was a freshly "prepared synthetic mixture,
the forces of compaction and adhesion at the time of treatment were probably
weak, a condition more typical of a recent spill-site soil than an older soil
found at an abandoned CERCLA site. Consequently, in these studies, the water
10
-------
Table V. Soil Washing Effectiveness (greater than 2-mm size fraction),
overall percentage reduction by contaminant group.
SARM I
(high organics,
low metals)
Volatiles
Semivolatiles
Inorganics
a Total waste
Water
>99.9
98.9
92.2
Surfactant
>99.8
>99.8
91.5
analysis.
Table VI.
SARM II
(low organics,
low metals)
Water
99.9
93.9
>96.7
Surfactant
99.9
93.5
95.7
Chelant
>99.9
90.1
95.9
SARM III
(low organics,
high metals)
Water Chelant
>99.9 99.9
>94.8 96.4
98.0 98.4
Soil Washing Effectiveness (250-um to 2-mm size
overall percentage reduction by contaminant group.
SARM I
Volatiles
Semivolatiles
Metals
a Total waste
Water
99.8
56.2
81.6
analysis.
Table
Surfactant
99
65
80
VII.
.8
.6
.7
Water
>99.9
52.7
>82.7
SARM II
Surfactant
>99.8
47.3
91.6
SARM IV
(high organics,
high metals)
Water Surfactant
>99.9
97.8
97.1
Jraction),
SARM III
Chelant
>99.9
67.5
85.1
Water Chelant
>99.3
0
96.4
99.0
0
98.4
Soil Washing Effectiveness (less than 250-pm size
overall percentage reduction by contaminant group.
SARM I
Volatiles
SemivolatHes
Metals
Water
66.2
59.7
0
Surfactant
88
43
0
.0
.2
Water
>99.8
0
0
SARM II
Surfactant
>99.4
0
0
99.9
>98.3
98.4
SARM IV
Water Surfactant
>99.7
0
90.7
, fraction)
a
SARM III
Chelant
99.6
0
21.9
Water
86.7
0
0
Chelant
>93.2
0
78.2
>99.7
29.4
91.8
SARM IV
Water Surfactant
>69.6
0
0
95.0
0
7.3
Chelant
>99.9
97.8
98.1
Chelant
>99.7
32.3
90.3
Chelant
81.8
0
82.2
CO
00
en
CO
Total waste analysis.
-------
88-6B.5
wash proved to be as effective in cleaning the ^2-mm soil fraction as the'
water-plus-additive solutions.
Contaminant removals from the 250-ym to 2-mm size fraction are summa-
rized in Table VI. Overall, the data show that the volatiles also were
efficiently removed from this soil category at levels exceeding 99 percent by
all wash solutions. These results are similar to those seen in the ^2-mrn
fraction. Semivolatile removal efficiencies dropped off compared with results
for the >2-mm size fraction (see Table V). Also, semi volatile removal effi-
ciencies for SARM's III and IV were markedly lower than for SARM's I and II.
Metal removal efficiencies were also somewhat lower across the board for this
size fraction compared with the ^2-mm fraction. The trend toward reduced
removal efficiencies for the semivolatiles and metals is not surprising, as
this size fraction has more surface area than the ^2-mm fraction, and also
some small amount of silt and clay particles; therefore, it has a higher
potential to adsorb and retain more contamination than the larger ^2-mm
fraction.
For the fine soil fraction (<250 urn) washing with any of the solutions
effectively removed the volatiles; conversely, none of the solutions were
found to be consistently effective in removing the semivolatiles from this
size fraction of the SARM's. Removal of metallic contaminants definitely
appeared to be enhanced somewhat by the use of the chelant. As shown in
Table VII, the chelant wash was much more effective than with the water wash
or the surfactant wash in reducing metal contamination-in the fine soil
fraction.
In summary, the results support the basic assumptions underlying the
volume-reduction approach to soil decontamination; that is, a significant
fraction of the contaminants are attached to the fines (silt, humus, and
clay), and the coarse material (sand and gravel) can be cleaned by physical
separation from the fines. The data indicate that 1) water alone can effi-
ciently remove a significant portion of both the organic and inorganic contam-
ination from the ^2-mm soil fraction, and 2) the addition of a chelant can
enhance metals removals from the middle (2 mm to 250 ym) and fine (<250 pm)
soil fractions.
Chemical Dechlorination/KPEG
Chemical dechlorination was examined as a treatment technology because
it had already been successfully demonstrated at laboratory scale with PCB-
and dioxin-contaminated soils and sludges, and was viewed as a promising
treatment technology for development to pilot scale and possibly full scale.
The KPEG dechlorination process involves the application of a potassium
hydroxide-polyethylene glycol reagent to contaminated soil at elevated tempera-
tures for a period of 2 to 4 hours, after which the reagent is decanted and
recovered and the soil is rinsed and neutralized. The reagert strips one or
more chiorine atoms from the PCB or dioxin molecule, forming an inorganic
chloride salt and a derivative of the PCB or dioxin, which, in theory, should
be less toxic than the original contaminant.
12
-------
88-6B.5
Each of the four SARM's was evaluated in this study. Although the
SARM's did not contain any PCB's or dioxins, other chlorinated species were
present, and there was interest in learning whether these compounds could be
dechlorinated. There was also interest in learning whether the process would
exhibit any removal effectiveness on the other organic and inorganic contami-
nants in the test soils.
Testing was conducted in either 500-ml or 2-liter glass reaction vessels
mounted within temperature-controlled heating mantles. In each test, either
125 or 500 g of SARM were treated with KPEG reagent at 100°C for 2 hours.
During the reaction period, the contents of the glass reaction vessel were
continually stirred at 100 rpm with a Teflon-coated stainless steel stirring
rod. The system was also continually purged with nitrogen, and the off-gases
were filtered through a Tenax/XAD-2/carbon trap system. The contents of the
traps were subsequently analyzed to establish material balances and to deter-
mine which compounds had been destroyed versus those which had simply been
volatilized. At the end of the 2-hour reaction period, the reagent was
separated from the soil by centrifugation and decantation. The soil was then
neutralized by an acid rinse followed by a plain water rinse. All rinse
solutions, soil residues, and the spent reagent were analyzed for the target
SARM contaminants.
Overall results of the KPEG tests are given in Table VIII. The analyses
show that the KPEG process was very effective in removing the volatiles from
all four SARM's. Removal rates for all volatiles exceeded 90 percent in all
tests, and most.often ranged from 98 to 99+ percent. Although material bal-
ances were generally poor, the data strongly indicated that most of the vola-
tiles were unaffected chemically by the treatment and were removed strictly
Table VIII. KPEG effectiveness on SARM's -
overall percentage reduction by contaminant group.
SARM I SARM II SARM III SARM IV
Test 1 Test 2 Test 1 Test 2 Test 1 Test 2 Test 1 Test 2
Volatiles
(all)
Semi volatiles
Anthracene
Pentachlo-
rophenol
Inorganics
(all)
99.9
91.3
98.1
44.5
98.3 »
96.3
97.7
_
98.2
75.6
91.9
39.4
96.3
-10
94.5
-
99.5
-490
99.6
49.4
97.5
-1246
99.0
_
99.9
96.0
95.8
29.3
98.1
97.0
95.4
_
As measured by total waste analysis. A negative percent reduction results
when chemical analysis of a treated residue yields a higher contaminant
concentration than the untreated material.
13
-------
88-6B.5
by volatilization processes. Notable exceptions to this were 1,2-dfchloro-
ethane and tetrachloroethylene, which appeared to have been completely de-
stroyed by the process.
Semi volatile results are available for only anthracene and pentachloro-
phenol. In the case of pentachlorophenol, the data indicate it was removed
from the soil at efficiency levels ranging from 92 to 99 percent; however,
the mass balance data indicate that it was not dechlorinated by the KPEG
reagent. Anthracene also was not destroyed. Removal efficiency data for the
compound are somewhat equivocal; in the tests utilizing SARM's I and IV,
which had starting concentrations of anthracene of 4000+ ppm, it was found to
be efficiently removed (i.e., removal rates ranged from 91 to 97 percent).
In tests involving SARM's II and III, which had much lower anthracene levels
(i.e., less than 250 ppm), no removal was observed. This may be due to
analytical limitations associated wtih recovering anthracene at these levels
in soils.
The KPEG process had only a limited effect on removing the inorganic
contaminants from the SARM's. Overall removal rates ranged from 29 to 49
percent.
Low-Temperature Thermal Desorption
The purpose of this research was to investigate the capability of a
laboratory-scale low-temperature thermal desorption technology for removing
volatile and semivolatile contaminants from the SARM's. The laboratory test-
ing program consisted of 15 separate bench-scale tests (10 in a tray furnace
and 5 in a tube furnace). Only SARM's I and II were tested at 150°, 350°,
and 550°F for 30 minutes to determine the effect of each temperature on
remova" of the contaminants. The tray furnace was used as a baseline tech-
nology to determine the overall effectiveness of thermal desorption in remov-
ing contaminants from the soil. The tube furnace was used to provide addi-
tional data on the concentration of contaminants in the off-gas in an attempt
to establish a material balance.
The first series of 10 tests involved the use of the tray furnace in
which SARM's I and II were each tested once at 150° and 350°F (four tests)
and three times each at 550°F (six tests).
-------
88-6B.5
550°F was applied. The apparent increase in metal concentrations in the
residues (as indicated in the negative reduction values) may be an artifact
in the data, due to moisture losses during heating; because the SARM's con-
tained 6 to 17 percent moisture before treatment (see Table III), the losses
tended to produce a higher metal-to-soil ratio (i.e., concentration) in the
treated residual, which results in an apparent (but unreal) increase in metal
content. A second factor that may have contributed to the change in concen-
tration of the metals may have.been a change in the matrix's ability to
retain metals after heating.
Table IX. Low temperature desorption - overall percent
reduction of contaminants by group at various test temperatures using
tray furnace and 30-minute residence time.
SARM I SARM II
150°F 350°F 550°F 150°F 350°F 550°F
Volatiles
Semivolatiles
Metals
97.8
-5.3
-9.3
99.8
41.6
-12.1
99.8
93.6
• -15.1
98.3
11.7
5.1
95.9
74.8
10.2
96.0
86.3
-7.3
a As measured by total waste analysis. A negative percent reduction results
when chemical analysis of a treated residue yields a higher contaminant
concentration than the untreated material.
In "terms of total actual residual concentrations, the following state-
ments can be made (refer to fable IV for initial concentrations prior to
treatment):
SARM I:
At 350° and 550°F, all volatiles except acetone were reduced to
less than 1 mg/kg in the treated residue; acetone residuals on the
order of 100 ppm remained, even at the highest temperature.
For the semivolatiles anthracene and BEHP, residuals remained well
above 1000 mg/kg at the 150° and 350°F temperatures, but were re-
duced to less than 20 mg/kg at 550°F. Pentachlorophenol residuals
remained high at the 150° and 350°F temperatures and were only
reduced to levels on the order of 100 ppm at the 550°F temperature.
SARM II:
As with SARM I, at 350° and 550°F, all volatiles except acetone
were reduced to less than 1 mg/kg; acetone residuals on the order
of 100 mg/kg remained, even at the 550°F temperature.
All semivolatiles were reduced to less than 100 mg/kg at 350°F and
to less than 10 mg/kg at 550°F.
15
-------
88-6B.5
Overall, the 150°F temperature was considered ineffective under the reation
conditions tested.
High-Temperature Incineration
In this segment of the test program, a series of pilot-scale test burns
was conducted with SARM's I and II only. The testing was conducted at the
John Zink testing facility in Tulsa, Oklahoma, in a rotary kiln incineration
system using a nominal feed rate of 1000 Ib/h. More than 12,000 pounds of
each SARM soil was prepared for the tests so that three 4-hour test burn runs
(for a total of six test burn runs) could be conducted on each SARM. Approxi-
mately 1 week prior to startup of the test burns, the soils were delivered to
John Zink in forty-eight 55-gallon steel drums, each containing 500 to 600 Ib
of SARM I or SARM II.
Two runs per day were conducted over the 3-day period of September 16
through 18, 1987. Runs 1,2, and 3 were conducted with SARM I (high organics,
low metals), and Runs 4, 5, and 6 were conducted with SARM II (low organics,
low metals). Equipment operations were normal throughout each run.
The process operating data collected during each test show that the
temperatures and feed rates achieved were reasonably close to the goals
(i.e., 1800°F in the kiln, 2000°F in the secondary combustion chambers, and a
nominal feed rate goal of 1000 Ib/h). Excess air was maintained at about 3
percent in the kiln and about 5 percent in the secondary chamber during both
.tests. Emissions of 02, C02, and CO were steady throughout; and CO remained
at less .than 10 ppm at all times except for a few brief excursions of 45 to
90 ppm, which lasted from 1 to 5 minutes. A total of 13,932 Ib of SARM I and
13,460 Ib of SARM II were incinerated over a course of 3 days that involved
29 hours 22 minutes of testing.
Table X presents the results of chemical analyses (total waste analyses)
of the bottom ash (i.e., SARM residue) samples collected during each test
run. Samples analyzed for semivolatiles and metals were collected as com-
posites over the course of each test; samples analyzed for volatiles were
collected as discrete samples at the beginning, middle, and end of each run
and composited at the time of analysis.
The volatile compounds styrene, tetrachloroethylene, and chlorobenzene,
and the semivolatile compounds anthracene and pentachlorophenol were not
detected in any of the ash samples. Measureable quantities of ethylbenzene
and xylene were found in the ash of both SARM's. and 1,2-dichloroethane was
found in the ash of SARM II, but the amounts were small (in the low parts-
per-billion range) and typically at levels within 2 to 3 times the method
detection limit. Acetone was found in the ash samples of all runs for both
SARM's at significant levels ranging from 190 to 790 yg/kg; these levels are
24 to 99 times higher than the method detection level (8 yg/kg).
On the average, the concentrations of acetone and phthalate found in the
ash of SARM I are similar to those found in the ash of SARM II, even though
16
-------
88-6B.5
Table X. Total Waste Analysis for SARM ash.
Parameter
VOLATILES, ug/kg
Ethyl benzene
Xylene
Tetrachl oroethyl ene
Chlorobenzene
Acetone
1,2-Dichloroethahe
Styrene
SEMIVOLATILES, yg/kg
Anthracene
Bis(2-ethylhexyl)
Method
detec-
t1 T f\n
U 1 Uii
limit
7.0
5.0
4.0
6.0
8.0
3.0
3.0
37
63'
Run 1.
NDa
ND
ND
ND
440
ND
ND
ND
1600
SARM I
Run 2
19
34
ND
ND
420
ND
ND
ND
540
Run 3
ND
ND
ND
ND
630
ND
ND
ND
740
Run 4
8
11
ND
ND
190
ND
ND
ND
950
SARM II
Run 5
ND
6
. ND
ND
210
• 5
ND
ND
710
Run 6
13
20
ND
ND
790
10
ND
ND
1300
phthalate
Pentachlorophenol
METALS, mg/kg
370
ND
ND
ND
ND
ND
Lead
Zinc
Cadmium
Arsenic
Copper
Nickel
Chromium
VOLATILE PICs, pg/kg
2-Butanone
Methyl ene chloride
2-Chl oroethyl vi nyl
ether
a ND = Not detected.
"* f ** 4. •£ BM 4 ^ A J ' 1**«1 ••«*.• 1 M «» «•
4.2
0.12
0.12
0.04
0.42
0.30
0.30
25
2.8
5.0
4.U-... n«4-U
56
217
<0.2
38
111
12
10
35
2.9
70
j*t x4 stfi^r+s+
98
227
<0.2
36
132
15
14
ND
5.4
ND
+• ^ /\r* Tin
107
250
<0.2
44
159
11
12
ND
4.2
ND
nit
14L
ND
ND
ND
ND
ND
ND
ND
ND
ND
17
-------
88-6B.5
the input waste feed levels for these compounds were roughly 10 times higher
in SARM I than in SARM II. This suggests sample contamination or carryover,
and the data for these compounds should be interpreted with caution. Signif-
icant quantities of phthalate were also found in several of the method blanks,
and phthalates are known to be commonly encountered contaminants'in sample
analysis.
The metals data for the ash samples were also interesting. Prior to the
testing, most of the metals concentrations in the ash were expected to be
elevated compared with those in the waste feed because of the combined effects
of the retention of metals in the ash and the losses of water and organics
from the feed during the incineration process. Cadmium levels in the ash,
however, were expected to be low as a result of volatilization of the metal
in the kiln at the high operating temperature of 1800°F. As expected, cad-
mium levels in the ash were quite low, at least 99.9 percent lower than the
waste feed levels. Surprisingly, all of the other heavy metal levels were
also lower in the ash (e.g., on the order of 50 to 80 percent lower) than in
the waste feed, which indicates significant volatilization or perhaps slag-
ging or condensation onto the kiln refractory. On the other hand, arsenic
levels in the ash were more than double those in the feed across the board.
The test burns successfully met all the RCRA emission requirements for
hazardous waste incineration. Stack samples collected during the SARM I and
II test burns revealed the following:
0 Particulate concentrations corrected to 7 percent 02 were below the RCRA
allowable limit of 0.08 gr/dscf for each SARM type.
0 Measured HC1 emission rates in pounds per hour were considerably less
than the RCRA allowable rate of-4.0 Ib/h for each SARM type.
0 The average stack gas concentration of CO was less than 23 ppm during
each test.
0 The destruction and removal efficiency (ORE) performance standard of
99.99 percent was achieved for all of the volatile compounds for each
SARM. The ORE data for the semivolatiles show that anthracene was
effectively destroyed, as the amount in each emission was less than the
method detection limit, and the resulting DRE's were greater than 99.99
percent. The ORE data for bis(2-ethylhexyl)phthalate showed that only
three of six sample runs met the 99.99 percent criteria. Sample contam-
ination (background level) problems may have been responsible for the
poor DRE's in the other three runs.
SOLIDIFICATION/STABILIZATION
This project evaluated the performance of solidification as a means of
treating the SARM soils. Tests were conducted on all four SARM's using three
commonly used solidification agents or binders: Portland cement (Type 1),
lime kiln dust, and a 50:50 mixture by weight of lime and fly ash. At 7, 14,
18
-------
88-6B.5
21, and 28 days after the SARM's and binders were mixed, samples of the
solidified material were subjected to unconfined compressibility strength
(DCS) testing. Samples that achieved a UCS minimally greater than 50 psi or
that showed the highest UCS below 50 psi after 14 and 28 days were subjected
to total waste and TCLP analyses.
Results of the testing showed that the UCS tended to increase with time
as the samples cured. Portland cement produced the strongest, hardest, and
most consistent product, followed by kiln dust and lime/fly ash. The lime/
fly ash samples required several weeks of curing before they finally set.
The amount of moisture in the SARM's seemed to be an important factor in
solidifying the soils. Offgassing of volatile compounds from the stabilized
samples occurred during mixing and continued throughout the curing period.
Table XI presents an overall summary of the 28-day samples when analyzed
for TCLP and TWA. The percentage reduction values in this table represent
the total amount of contaminants found in the untreated soil samples (or TCLP
extract) less the total amount of contaminant found in the stabilized sample
(or TCLP extract) divided by the amount in the untreated soil (or TCLP ex-
tract) x 100 percent. The data have been adjusted to account for changes in
soil volume and contaminant concentration caused by the addition of the
binders.
The results fail to indicate either dramatic or consistent treatment
efficiencies. Volatiles in the SARM's were reduced, but the reductions are
attributed to volatilization losses (offgassing) during mixing and curing
rather than binding within the stabilized matrix." Metals were less prevalent
in the treated sample TWA and TCLP extracts, which indicates reduced mobility
after stabilization/solidification. The percent reductions seldom exceeded
90 percent, however, and generally tended to range-from 0 to 20 percent to 60
to 75 percent. Overall, kiln dust and lime/fly ash produced the best contam-
inant-reduction results.
SUMMARY AND CONCLUSIONS
The research program produced a valuable and interesting new data base
outlining the kinds of results that can be achieved by treating a synthetic
contaminated soil at bench and pilot scale. This paper only highlights key
portions of the data base; it is by no means complete. Detailed reports
covering the complete findings of each study are available through EPA's
Hazardous Waste Engineering Research Laboratory in Cincinnati (see Acknow-
ledgments).
Preparation of the SARM's is viewed as a particularly valuable segment
of the research because this had never before been attempted on such a large
(volumetric) scale. Methods of mixing both the basic clean soil and the
contaminated material were developed and found to produce a quality product
with good homogeneity. This allowed each of the treatment technologies to
operate with a high degree of assurance that the starting materials were
essentially identical from one test to another.
19
-------
Table XI. Summary of Effectiveness of Stablization/Solidification Agents
After 28-Day Cure
(% Reduction)3
PCb
SARM I
KDC L/FAd
SARM
PC KD
II
L/FA PC
SARM III
KD L/FA
SARM
PC KD
IV
L/FA
TCLP
Volatiles 73.9 97.6 75.8 42.2 82.0 88.0 68.3 >93.6 90.3 -45 77.5 73.7
Semivolatlles 67.2 >98.8 >96.5 >53.8 >68.5 >72.8 -47.0 82.2 54.8 -139 57.2 88.6
Inorganics >82.1 >75.0 >92.3 >83.0 >71.2 >92.1 99.4 83.7 54.3 85.3 66.4 68.3
TWA
o Volatiles - 98.5 83.7 86.9 99.7 97.0 76.9 98.1 92.0 58.5 95.3 83.4
Semivolatiles - 87.7 80.2 33.1 38.0 27.5 -101 -37.5 -32.2 24.2 28.3 47.9
Inorganics - 43.8 56.6 -13.6 9.7 28.1 28.5 73.2 82.3 32.3 60.5 73.9
a A negative percent reduction results when analysis of the treated residual (or extract of a treated
resid'.-^l) yields a higher contaminant concentration than the untreated material.
PC = Portland cement.
c KD = Kiln dust.
d LFA = Lime/fly ash.
00
00
I
co
en
-------
88-6B.5
A rank-order summary of the effectiveness of each treatment technology
on the four SARM's, is presented in Table XII. The thermal technologies
effectively reduced the organic fractions (>99.6%) when measured by TWA. The
chemical treatment (KPEG) operated on the semivolatile fraction with greater
than 90 percent reduction effectiveness. Greater than 98 percent of the
volatile organic compounds were removed, but this was likely due to volati-
lization during the test runs. Soil washing was the best metals reduction
technique across all the SARM's, averaging 93 percent. Soils washing was
also very effective in reducing the semivolatile compounds (averaging about
87%) and the volatiles (99%). Stabilization generally ranked behind the
other technologies, as expected, since it does not remove metals, but im-
mobilizes them. For stabilization, TCLP is a better measure of treatment
effectiveness than TWA.
Phase II of the CERCLA Research Program was initiated in 1988 and is
continuing. Soils from actual Superfund sites have been collected and are
being tested for treatment effectiveness using the same bench-scale proce-
dures as in Phase I. Results, which are expected to be available in late
1988, will be compared with those produced on the SARM's.
ACKNOWLEDGMENT
Phase I of this CERCLA Research Program was funded in its entirety by
the U.S. Environmental Protection Agency, Office of Research and Development,
Hazardous Waste Engineering Research Laboratory, Cincinnati, Ohio. The work
was conducted by the following contractors:
Project
SARM preparation
Physical soil
washing
Dechlorination/
KPEG
Thermal desorp-
tion
Contractor
PEI Associates, Inc.
Cincinnati, Ohio
PEI Associates, Inc.
Cincinnati, Ohio
Wright State Univer-
sity
Dayton, Ohio (subcon-
tractor to PEI Asso-
ciates)
IT Corporation
Knoxville, Tennessee
(subcontractor to
PEI Associates)
EPA contract EPA Project Officer
68-03-3413
Work Assign-
ment 0-7
68-03-3413
Work Assign-
ment 0-7
68-03-3413
Work Assign-
ment 0-6
68-03-3389
Work Assign-
ment 0-5)
Richard P. Traver
Richard P. Traver
Charles J. Rogers
Robert C. Thurnau
21
-------
Table XII. Overall BOAT Phase I treatment efficiency summary.'
Percent Percent Percent Percent
SARH 1 reduc- SARM II reduc- SARM III . reduc- SARM IV reduc-
(high organ)cs, low metals) tion (low organic:, low metals) tion (low organics, high metals) tlon (high organic;, low metals) tion
ro
ro
VOLATILES
Incineration
>99.99 Incineration
Soils washing + 2 mm water >99.99
Chemical treatment KPEG 99.96
No. 1
Soils washing + 2 mm surfac- 99.82
tant
Soils washing 2 mm to 250 99.82
um surfactant
Soils washing 2 mm to 2SO 99.8
um water
Low temperature thermal 99.79
desorb at 350°r
Low temperature thermal 99.78
desorb at S50°F
Solidification - kiln dust - 98.5
28 days
Chemical treatment KPEG 98.3
No. 2
SEHIVOLATIIES
Incineration >99.98
Soils washing + 2 mm sur- >99.8
factant
Soils washing •» ', mm water >98.9
Chemical treatment KPEG 97.0
No. 2
Soils washing - all frac-
tions - water
Soils washing - all frac-
tions - chelate
Soils washing - all frac-
tions - surfactant
Solidification - kiln dust
28 days
Low temperature thermal
desorb at 150°F
Chemical treatment KPEG
test No. 1
Solidification - lime/fly
ash
Chemical treatment - KPEG
No. 2
Low temperature thermal at
500" f
>99.98 Soils washing + 2 mm water >99.9
>99.9 Soils washing + 2 mm chelate 99.9
>99.7 Chemical treatment KPEG No. 1 99.5
99.7 Soils washing 2mm to 250 \m 99.3
water
99.7 Soils washing Znin to 250
um chelate
99.0
98.70 Solidification - kiln dust - 98.1
28 days
98.2 Chemical treatment KPEG No. 2 97.6
97.0 Soils washing <250 pm
chelate
98.2
96.3 Solidification lime/fly ash - 92.0
28 days
96.17 Soils washing <250 um water 86.7
Incineration >99.87 Chemical treatment KPEG 99.6
No. 1
Soils washing + 2 mm water 93.9 Chemical treatment KPEG No. 2 99.0
Soils washing + 2 mm sur- 93.5 Soils washing + 2 mm chelate >96.4
factant
Chemical treatment KPEG 99.98
No. 1
Soils washing * 2 mm water >99.9
>99.9
>99.9
Soils washing t 2 mm
chelate
Soils washing + 2 mm
chelate
90.1 Soils washing + 2 mm water >94.8
Soils washing + 2 mm
surfactant
Soils washing 2 mm to 250 >99.7
iim surfactant
Soils washing 2 mm to >99.7
250 vim chelate
Soils washing 2 mm to >99.7
250 um water
Chemical treatment KPEG 98.1
No. 2
Solidification kiln dust - 95.3
28 days
Soils washing <250 um 81.8
chelate
Soils washing + 2 mm sur- >98.3
factant
•
Soils washing * 2 mm 97.8
chelate
Chemical treatment KPEG 96.2
No. 2
Chemical Treatment KPEG 92.9
No. 1
(continued)
-------
Table XII (continued)
ro
Percent
SARM I reduc-
(high organics, low metals) tion
Percent Percent • Percent
SARM II reduc- SARM III reduc- SARM IV reduc-
(low organics, low metals) tion (low organics, high metals) tion (high organics, low metals) tion
Chemical treatment KPEG 95.6
No. 1
Low temperature thermal 94.6
desorb at 250°F
Soils washing 2 mm to 250 82.3
urn surfactant
Solidification lime/fly ash 80.2
Low temperature thermal 88.73
desorb at 350°F
Chemical treatment KPEG B3.8
Test No. 1
Soils washing 2 mm to 250 vm 67.5
surfactant
Soils washing 2 mm to 250 um 22.7
Solidification lime/fly
ash - 28 days
Soils washing 2 mm to
250 um chelate
Soils washing 2 mm to
250 u"i surfactant
Solidification kiln dust -
47.9
32.3
39.4
28.3
surfactant
Solidication kiln dust
80.2
Soils washing <250 um water 59.7
METALS
water
Soils washing 2 mm to 250 um 47.3
chelate
Chemical treatment KPEG 42.3
No. 2 .
28 days
Soils washing + 2 mm water 92.2 Soils washing + 2 mm water >96.7
Soils washing + 2 mm surfac- 91.5
tant
Soils washing 2 mm to 250 vm 81.6
water
Soils washing 2 mm to 250 um 75.5
surfactant
Solidification lime/fly ash - 56.6
28 days
Solidification kiln dust - 40.2
28 days
Soils washing + 2 mm 95.9
chelate
Soils washing + 2 mm sur- 95.7
factant
Soils washing 2 mm to 91.6.
250 um chelate
Soils washing 2 mm to 85.1
250 um surfactant
Soils washing 2 mm to 82.7
250 um water
Incineration
38.7
Incineration
64.3
Chemical treatment KPEG 39.4
No. 1
Soils washing 2 mm to 250 um
chelate
Soils washing 2 mm to 250 um
chelate
Soils washing 2 inn to 250 um
water
Soils washing 2 mm to 250 um
water
Solidification lime/fly ash
Soils washing <250 m chelate
Solidification kiln dust - 28
days
Chemical treatment No. 1
98.4 Soils washing + 2 mm sur- 98.4
factant
98.4 Soils washing + 2 inn
chelate
98.1
98.0 Soils washing + 2 mm water 97.1
91.8
90.7
78.2 Solidification lime/fly ash 73.9
73.2 Solidification kiln dust 60.5
49.4
96.4 Soils washing 2 mm to
250 um surfactant
82.3 . Soils washing 2 mm to
250 um water
a Based on total waste analyses.
-------
R8-6B.5
Project
Incineration
Contractor
PEI Associates, Inc.
Cincinnati, Ohio
EPA contract
68-03-3389
Work Assign-
ment 0-7
EPA Project Officer
Robert C. Thurnau
Stabilization
Acurex Corporation
Durham, North Carolina
68-03-3241
Work Assign-
ment 2-18
Edwin F. Earth
NOTICE: This report has been reviewed by the Hazardous Waste Engineering
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessar-
ily reflect the views and policies of the U.S. EPA, nor does men-
tion of trade names or commercial products constitute endorsement
or recommendation for use.
REFERENCES
Castle, C., et al. 1985. Research and Development of a Soil Washing System
for Use at Superfund Sites. In: Proceedings of the 6th National Confer-
ence on Management of Uncontrolled Hazardous Waste Sites, Washington,
D.C., November 4-6, 1985. Hazardous Materials Control Research Insti-
tute, Silver Spring, Maryland.
Rayford, R., R. Evangelista, and R. Unger. 1986. Lead Extraction Process.
Prepared for the U.S. Environmental Protection Agency, Emergency.Branch,
by Enviresponse, Inc., under Contract No. 68-03-3255.
Scholz, R., and J. Milanowski. 1983. Mobile System for Extracting Spilled
Hazardous Materials From Excavated Soils. EPA-600/2-83-100.
24
------- |