&EPA
United States
Environmental Protection
Agency
EPA/540/S5-91/003
March 1992
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
Technology Demonstration
Summary
BioTrol Soil Washing
System for Treatment of a
Wood Preserving Site
A SITE Program demonstration of one
configuration of soil washing process
developed by BioTrol, Inc., was carried
out at the MacGillis and Gibbs wood
treatment facility in New Brighton, MM.
The processing train, so called BioTrol
Soil Washing System (BSWS), con-
sists of three units: .
• The Biojrol Soil Washer (BSW) -
a volume reduction process, which
uses water to separate contami-
nated soil fractions from the bulk
of the soil.
• The BioTrol Aqueous Treatment
System (BATS) - a biological water
treatment process.
• The Slurry Bioreactor (SBR) - a Bio-
Trol biological slurry treatment
process conducted in an EIMCO
BIQLIFT1"1 reactor.
At the MacGillis and Gibbs site, where
pentachlorophenol (penta) and poly-
nuclear aromatic hydrocarbons (PAHs)
are the contaminants of primary con-
cern,'the BSW separated the feed soil
into relatively uncontaminated sandy
fraction that accounts for the majority
of the feed soil, contaminated woody
fractions, and a small fraction of fine
clay and silt particles. Contaminant re-
moval, defined by the difference be-
tween the concentrations of penta in
the feed soil and the washed soil, was
between 87% and 89% in tests with soil
with low penta content (130 mg/kg) and
high penta content (680 mg/kg). For
total PAHs the removal efficiencies were
83% and 88% in the two tests. The
process is particularly attractive where
the washed soil material would meet
site-specific regulatory requirements for
return to the site without further treat-
ment.
The BATS, using a penta-specific Fla-
vobacterium, degraded between 91%
and 94% of the penta in the process
water from the two soil washer tests.
PAH concentrations were below detec-
tion limits and removals could not be
determined.
The contaminated fines are only a
small fraction of the original feed soil
(<10%) at this site but contain over
50% of the original penta mass. Bio-
logical treatment of the slurry of these
fines in the SBR required a longer-than-
expected acclimation period. Conse-
quently, it can only be estimated that
removal levels for penta and PAHs of
Printed on Recycled Paper
-------
over 90% can be attained once steady-
state operation is achieved.
Costs were estimated for an inte-
grated, commercial-scale system and
also for each process operating inde-
pendently. Incineration of woody mate-
rial segregated by the soil washing ac-
counts for the major portion (76%) of
the costs.
This Summary was developed by
EPA's Risk Reduction Engineering
Laboratory, Cincinnati, OH, to announce
key findings of the SITE program dem-
onstration that is fully documented in
two separate reports (see ordering in-
formation at back).
Introduction
The Superfund Innovative Technology
Evaluation (SITE) Program was instituted
in 1986 to promote the development and
application of innovative technologies to
Ihs remediation of Superfund and other
sites contaminated wilh hazardous wastes.
The National Priorities List (NPL) includes
56 sites contaminated with penta and creo-
sote-derived PAHs from wood preserving
practices. Soil washing and subsequent
biodegradation of the organic contaminants
at such sites potentially could be an at-
tractive means for remediation of such
sites and, consequently, an appropriate
topic for investigation under the SITE Pro-
gram,
This Project Summary highlights the re-
sults of an evaluation of a specific ar-
rangement of the three technologies of
the BSWS. The system consists of mul-
tiple stages of physical abrasion, attrition,
flotation, and washing of excavated soil in
the BSW. This is accompanied by biodeg-
radation of solubilized contaminants in the
BATS and btodegradation of contamina-
tion adhering to fines in the SBR. While
the BATS was compatible with the BSW
in capacity, the SBR unit provided by
Eimco was considerably smaller.
The site selected for the evaluation is a
wood preserving facility in New Brighton,
MN, where creosote and penta were used
for several decades. The facility, owned
and operated by the MacGillis and Gibbs
Company, is currently using the newer
chromated copper arsenate (CCA) pre-
servative and improved technology to mini-
mize site contamination.
Process Description
Soil previously excavated from two ar-
eas on the site was used to provide test
soils with two levels of penta contamina-
tion. The weighted average concentrations
of panta in the two soils were 130 and
680 mg/kg.
In the proprietary BSW process, the soil
is first screened to remove large material.
It is then subjected to slurrying with water,
froth flotation, attrition/classification, thick-
ening, and dewatering using a combina-
tion .of vibrating screens, a mixing trommel,
flotation and attrition units, hydrocyclones,
spiral classifiers, and dewatering centri-
fuges. The test units are mounted on a
42-ft semi-trailer with drop-down sides. It
is readily transportable to a site for evalu-
ation. With this sequence, the soil is seg-
regated into a large volume of relatively
uncontaminated sandy material called
washed soil, contaminated woody debris,
and a much smaller fraction of contami-
nated fines. The washed soil is dewatered
and returned to the site if in compliance
with required cleanup levels, which were
not known for this site.
Process water from the BSW is treated
at a rate of 10 L/min (3 gpm) in the BATS,
a multiple-stage fixed-film bioreactor. In-
digenous bacteria are supplemented with
a penta-specific Flavobacterium to accel-
erate biodegradation. The system is first
acclimated for about 2 wk by recycle of
contaminated water. Nutrients are added
and the pH and temperature are adjusted
automatically as needed. Air is injected at
the base of each bioreactor cell through a
series of sparger tubes. The design of the
BATS is such that a minimum of operator
attention is required.
The fines produced by the BSW can be
dewatered and disposed of by conven-
tional technology such as incineration. An
attractive alternative is aerobic biodegra-
dation in a 3-stage, stirred SBR operating
in a cascade mode. A growth of both
indigenous and a penta-specific
Flavobacterium are acclimated for at least
1 wk to prepare the system but it was
found that the acclimation period is de-
pendent on penta concentration. For the
demonstration, slurry was diverted to a
storage tank during the soil washing of
the more highly contaminated soil (680
mg/kg penta) and then fed to the small
scale SBR unit (180 L capacity) over a
14-day test period at a flow rate of 24 ml/
min. Membrane diffusers provide the nec-
essary air for aerobic treatment. Tempera-
ture and pH are automatically controlled.
Woody debris, also containing signifi-
cant levels of penta and PAHs, is contain-
erized for disposal by conventional tech-
nology such as incineration.
Test Program
Tests established that most of the con-
taminants (penta and PAHs) are associ-
ated with the fine particle fraction of the
soil, making the BSW process well suited
for concentrating the contaminants in a
small fraction of the feed soil. Particle size
and contaminant distribution for the vari-
ous output fractions from the BSW pro-
cess are also provided.. However, these
results cannot be related directly to the
operation of the BSW.
The BSW is designed to operate con-
tinuously. For this demonstration, tests
were carried out using two soil piles. The
first test used a feed soil with a weighted
average penta concentration of 130 mg/
kg (average penta: 112.8±19.8 mg/kg) and
lasted about 50 hr. About 11,000 kg of
contaminated soil was treated. The sec-
ond test was carried out over about 150
hrs, with a total of about 18,000 kg of feed
soil containing a weighted average of 680
mg/kg penta (average: 657.8 ± 228 mg/
kg) treated. Average throughput rates in
the two segments of the study were 220
and 160 kg/hr, respectively.
During each test, the different output
solids and the process water were weighed
and sampled for extensive analyses, in-
cluding penta, PAHs, metals, dioxins, oil/
grease, TOG, chloride, organic halide, sol-
ids content, etc. All sampling and analy-
ses during the demonstration program fol-
lowed approved EPA or Standard Method
protocols. Flow rates for the input and
output streams were calculated using mea-
sured differences in the output over time
increments.
The water used to slurry the soil for
processing consisted of varying ratios of
municipal water, an aqueous solution of
polymer used to thicken the fines stream,
and treated effluent from the BATS. The
volumes of these were also measured to
establish an overall material balance. The
water from the BSW process was ana-
lyzed for parameters similar to those noted
above, both as produced and before it
was introduced into the BATS. The efflu-
ent from the BATS was also carefully ana-
lyzed to establish contaminant removals
attributable to the biological action.
Similarly, the contaminated fines slurry
was analyzed before and after treatment
in the SBR. In this case, because of the
solids in the slurry (~11% solids), the solid
and liquid phases of the slurry were ana-
lyzed separately for the critical param-
eters (penta and PAHs).
Results
Predemonstration particle size and con-
taminant analyses indicated that the fines,
constituting 5% to 8% of the combined
output mass, contained approximately 30%
of the penta and PAH contamination in
the output streams. .However, the proce-
dures used in these tests do not allow the
results to be related directly to the segre-
gation achieved in the BSW.
-------
BloTrol Soil Washer
, Wide fluctuations in the feed soil rate
were encountered during the soil washing
of both soil piles due to mechanical prob-
lems and the consistency of the soils.
Nevertheless, the BSW succeeded in seg-
regating the feed soil into relatively
uncontaminated washed soil, contaminated
woody debris, and a slurry of contami-
nated fines amounting to only a small
percentage of the feed soil mass. Con-
siderable penta is also dissolved into the
process water.
Removal efficiency was defined by the
developer as:
% Removal efficiency = -
100 x (1 - washed soil conc.^
feed soil cone.
On the basis of weighted average penta
concentration in the washed soil versus
penta in the feed soil, the removal-effi-
ciency was 89% in the test with the low
penta concentration' soil and 87% in the
test with the high penta concentration.
Within each test of the soil washer, the
concentration of penta retained by the
washed soil remained relatively constant
(9% and 11% of output, respectively). In
addition, the mass and penta content of
the washed soil remained relatively uni-
form, regardless of fluctuations in feed
soil flow rate or input penta concentration,
suggesting that maximum removal had
been achieved from that soil. In the two
tests, the bulk of the penta in the output
streams was found in the fine silt and clay
fraction (34% and 27%, respectively), the
process water (40% and 34%, respec-
tively), and the coarse and fine woody
material (14% and 29%, respectively).
Weighted values were used for stream
masses and penta concentrations in cal-
culating masses because sampling inter-
vals varied and different masses were col-
lected during the different sampling peri-
ods; consequently standard deviations
could not readily be calculated and would
not have the usual meaning. Arithmetic
averages and standard deviations for
stream flows, masses, and penta concen-
trations which indicate that the two calcu-
lation methods do not provide greatly dif-
ferent results are provided in the report.
Mass balances for input and output
streams and, particularly for penta con-
centrations, varied widely in the two tests
with considerably higher combined masses
found in the output streams than in the
input. While the explanation for this is not
certain, there is some reason to believe
that the soil washing improves the acces-
sibility of the solids during the solvent
extraction step of the analyses. Summa-
ries of the results for the two tests are
presented in Tables 1 and 2. The full
report provides more,extensive data on
concentrations and masses.
Results with PAHs paralleled those ob-
served for penta except that much smaller
concentrations were found in the process
water, which is consistent with the lower
solubility of the PAHs. The silt and clay
fraction contained 61% of the PAHs in the
low penta soil washing test and 55% in
Table 1. Results of Low Penta Soil Washing Test
the high penta soil washing test while the
washed soil retained slightly more PAHs
than penta (18% and 15%, respectively).
Removal efficiencies in the two tests were
calculated to be 83% and 88%.
In addition to the copper, chromium,
and arsenic expected from the CCA wood
treatment now in use, analyses were car-
ried out for a number of other metals.
Removal efficiencies for copper, chromium,
Input
Feed Soil
Municipal Water
Thickener Sol'n
Total
Output
Washed Soil
Coarse Oversize
Fine Oversize
Fine Particle Cake
Combined Dewatering
Effluent -.
Total
% Removal Efficiency
As-is
wt.
kg
1.00
5.45
0.82
7.27
1.18
0.14
0.06
0.22
5.00
6.60
Penta
cone
ppm
130
0
0
14
170
96
270
14
175.5
89
mg
130
0
0
130
16.5
23.8
5.8
59.4
70.0
mass'
%*
100
0
0
9
13
3
' 34
40
Total PAHs
cone
ppm
247
0
0
42
309
208
778
0.5
83
mass'
mg
247
0
N/A
247
49.7
-44.9
12.3
170
'. 2.5
279
%*
100
18
16
4
61,
1
' Mass refers to the mass of contaminant in the as-is weight of each fraction obtained from treatment
of 1 kg of feed soil.
2 Refers to percent of total input or output, respectively.
Table 2. Results of High Penta Soil Washing Test
Input
Feed Soil •
Municipal Water
Thickener Sol'n
BATS Effluent
Total
Output
Washed Soil
Coarse Oversize
Fine Oversize
Fine Particle Cake
Combined Dewatering
Effluent
Total
%Removal Efficiency
As-is
wt.
kg
1.00
0.22
0.75
4.12
6.09
1.31
0.18
0.06
0.22
4.50
6.27
cone
ppm,
680
0
0
2.8
691.5
87
1400
900
1300
80
87
Penta
mass'
mg %*
680
0
0
11
114
252
54
286
360
1066
98.3
0
0
1.7
11
24
5
27
34
cone
ppm
404
0
0
0.2
48
535
463
1064
1.9
88
Total PAHs
mass'
mg
404
0
N/A
0.9
405
61.8
92.2
25.8
230.5
8.6
419
%2
99.7
0.3
15
22
6
55
2
' Mass refers to the mass of contaminant in the as-is weight of each fraction obtained from treatment
of 1 kg of feed soil. - ,
2 Refers to percent of total input or output, respectively.
-------
and arsenic were in the range of 48% to
70% (based on concentrations in feed soil
and washed soil). Barium, lead, and mer-
cury were the only metals found at signifi-
cant concentrations in the output streams.
Once again, the fine particle cake con-
tained the bulk of the mass of each metal
for which sufficient data were available.
Similar results were obtained for Total
Organic Carbon (TOG) and Total Recov-
erable Petroleum Hydrocarbons (TRPH),
with removal efficiency from the feed soil
of 84% and 94%, respectively, in the test
Using the low penta concentratbn soil and
81% and 92%, respectively in the test
with the high penta concentration soil. It
should be noted, however, that the ranges
on which these averaged values are based
are broad, perhaps reflecting the variabil-
ity in the sampling of the soil washing.
Analyses were also carried out on the
feed soil and each output stream for vari-
ous polychlorodibenzo-p-dioxins and
furans(CDD/COFs). The octachtorodibenzo
dfoxin (OCDO) isomer constituted about
65% of the congeners found at measur-
able concentrations in each soil, and the
ratios of congeners were essentially con-
stant for all streams. Again, most of the
CDD/CDFs concentrated in the Fine Par-
ticle Cake and distribution among the out-
put streams was similar to that found for
the PAHs, Removal efficiencies, based on
washed soil and feed soil concentrations
of total CDD/CDFs, were 92% in the test
with the tow penta concentration and 97%
in the test with the high penta soil.
BloTroI Aqueous Treatment
System
Both tests of the BATS were carried out
at flow rates of about 10 L/min. Calcu-
lated on the basis of flow-weighted mass
data, the weighted penta concentrations
in the BATS using process water from the
low penta concentratbn soil washer test
decreased from about 14 to 1.3 mg/L,
equivalent to a removal of about 91%. In
the test using process water from the high
penta concentration soil washing, weighted
influent penta concentrations of 44 mg/L
were reduced to 3 mg/L, or a 94% re-
moval. Analytical data for free chloride
production and organic chloride consump-
tion were insufficient to establish that the
toss of penta occurs by mineralization to
chloride, carbon dioxide, and water. Insuf-
ficient data for PAHs above the detection
limits precluded the estimation of removal
efficiencies on the basis of these tests;
only acenaphthene was found regularly in
influent and effluent. Analyses for metals
indicated that there was some decrease
during the course of biotreatment. Since
none Is expected, it may be presumed
that metals are adsorbed in/on the bio-
mass in the reactor.
Slurry Bioreactor
The SBR operated at 24 mL/min (0.38
gal/hr), which provided a retention time of
about 5.2 days. Problems were encoun-
tered with the feed of slurry due to cold
ambient temperatures, with variability in
feed concentrations (probably due to poor
mixing in the storage tank), and with inad-
equate nutrient feed for the unexpectedly
high penta concentrations during the ini-
tial days of operation. The solid phase of
the slurry that was introduced to the SBR
contained an average of 2570 ± 506 mg/
kg penta while the liquid phase contained
only 59 ± 19 mg/L. Because of these
problems and the delay in system re-
sponse to corrective actions, acclimation
was not achieved before sampling was
initiated and initial effluent concentrations
of penta and PAHs were higher than ex-
pected. Only toward the end of the 14 day
test period were the anticipated penta and
PAH removal levels of 90% reached for
both solid and liquid phases of the slurry.
PAHs were not detected in the liquid
phase of the influent or the effluent. Even
for the solid phase, many of the PAH data
were at or below detection and could only
be estimated. The high penta concentra-
tions in all the samples contributed to the
analytical difficulties and high detection
limits. Removal efficiencies calculated for
each of seven PAHs using measured or
estimated concentrations in the solid
phases confirmed that removal was pro-
ceeding slower than expected and only
attained values in the 90% or higher range
after about 7 days. For some of the PAHs,
removal efficiency never reached this level
(Table 3), and for others the highest level
was reached early in the treatment and
then decreased over the course of the
study. These anomalous results remain
unexplained.
Costs
Basic operating and equipment capital
costs were provided by the vendor and
estimates were made concerning the frac-
tion of feed soil that would be returned to
the site as washed soil (assuming clean-
up levels are met), that which would be
woody debris requiring off-site disposal,
and the fraction that would be clay/silt
fines amenable to slurry biodegradation.
Certain other assumptions were made
based on experiences during the demon-
stration project and certain other cost fac-
tors were assumed to be the responsibility
of the site owner/operator. The reasoning
used in making these estimates or omit-
ting a particular cost category are dis-
cussed in the report.
Costs were first estimated for a 0.25 to
0.5 ton/hr pilot scale system similar to that
used in the demonstration but operating
for only 2 wk and without the costs related
to the SITE evaluation. Assuming lease of
all equipment, the cost estimated was
$6.50/kg or $6000/ton of soil treated. Mo-
bilization and demobilization and lease
costs are major contributors to the high
cost of such an evaluation.
The cost to treat similarly-contaminated
soil in an area such as the MacGillis and
Gibbs site with an 18.2 metric ton/hr (20
ton/hr) .commercial soil washer coupled
with three 100 gpm BATS units operating
in parallel and three parallel SBR trains
capable of treating 23 gpm of slurry was
estimated at about $168/ton of soil. Dis-
posal of the contaminated woody material
segregated during soil washing accounts
for the major portion (76%) of this cost.
Applicability to Other Sites
Based on the demonstration and other
information provided by the vendor, the
BSWS appears to be attractive for soils
where: (1) there is a small fraction of
fines, (2) the fines retain the bulk of the
contaminants, and (3) the fines can be
Table 3. PAH Removals in the Slurry Bio-Reactor
PAH
Acenaphthene
Fluoranthene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(b)ftuoranthene
Benzo(a)pyrene
Avg. Cone.'
On Influent Solid '
mg/kg
53 + 44
277 + 144
363 + 121
63+.41
89 + 47
69 + 36
48 + 41
Max.
%
89
99
94
87
93
75
80
Removal
Day of Test
7
14
14
14
14
14
6
On Day 14
%
75
99
94
87
93
75
71
'Includes measured and estimated values in averages.
4
-------
segregated from the bulk of the soil by the
BSW process. If the soil contains exces-
sive fines (over about 25%), the benefits
of soil segregation by particle size begin
to become less evident.
Water soluble contaminants such as
penta tend to dissolve in the aqueous
process stream and subsequently can be
treated biologically in the BATS. Water-
insoluble contaminants such as PAHs (as
well as adsorbed penta) can be removed
from the slurry of fines by biodegradation,
as in the SBR, or can be disposed of by
conventional means such as incineration.
While no experimental work was done
to evaluate such options, it may be fea-
sible to customize the system for removal
of other contaminants, such as by adding
surfactants, adjusting pH, or adding sol-
vents.
Conclusions
For properly selected soils, the BSWS
can segregate a large fraction of relatively
uncontaminated washed soil. Whether this
soil can be returned to the site after de-
watering with no further treatment will be
dependent on cleanup requirements; this
will have a major impact on the cost-
effectiveness of the process.
Soil character, moisture content, par-
ticle size distribution, and contaminant con-
centrations and solubilities all are factors
in the efficiency and the operability of the
soil washer. For example, the soil washer
is most effective when the soil contains
25% or less of fine material such as clay
and silt.
Based on the decrease in contaminants
in the washed soil relative to the feed soil,
contaminant removal efficiencies of just
under the vendor's claimed 90% can be
achieved by the soil washing process for
penta, PAHs, and even dioxins.
Subsequent biotreatment of contami-
nated soil washer process water in the
BATS can achieve over 90% degradation
of penta. Degradation of PAHs could not
be determined in this study.
The small fraction of fines, even con-
taining very high levels of adsorbed penta,
can be subjected to biodegradation in the
SBR if adequate acclimation time is al-
lowed. Significant removal of various PAHs
is also achieved in the SBR, but to a
much more variable level. For maximum
cost-effectiveness, the treated fines should
then meet regulatory cleanup requirements
and be suitable for return to the site.
The cost to treat 1 ton of feed soil in a
20 ton/hr soil washer, combined with the
cost to biodegrade the contaminated fines
and associated process water before it is
recycled and to incinerate woody debris,
is approximately $168/ton. Incineration of
woody material is the major cost factor.
•&U.S. GOVERNMENT PRINTING OFFICE: 1992 - 648-080/40194
-------
-------
-------
The EPA Project Manager, Mary K. Stinson, is with the Risk Reduction
Engineering Laboratory, Edison, NJ 08837 (see below)
The complete report, entitled "Technology Evaluation Report: BioTrol Soil
Washing System for Treatment of a Wood Preserving Site," consists of two
volumes:
"Volume /" (Order No. PB92-115 310-V1; Cost: $35.00, subject to change)
discusses the results of the six-week field demonstration.
"Volume II", Part A (Order No. PB92-115 328-V2-Pt A; Cost: $43.00, subject to
change) and Part B (Order No. PB92-115 336-V2-Pt B; Cost: $43.00,
subject to change) contains the technical operating data - laboratory
analytical results, etc.
Both volumes of this report will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
A related report, entitled "Applications Analysis Report: BioTrol Soil Washing
System for Treatment of a Wood Preserving Site," discusses the applica-
tions of the demonstrated technology.
The EPA Project Manager can be contacted at:
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
United States
Environmental Protection
Agency
Center for Environmental
Research Information
Cincinnati, OH 45268
BULK RATE
POSTAGES FEES PAID
EPA
PERMIT NO. G-35
Official Business
Penally for Private Use $300
EPA/540/S5-91/003
------- |