United States
                   Environmental Protection
                   Agency	
                                  Risk Reduction
                                  Engineering Laboratory
                                  Cincinnati OH 45268
&EPA
                   Research and Development
                                  EPA/600/SR-92/068  Sept. 1992
 Project  Summary
Pilot-Scale Incineration  of
PCB-Contaminated
Sediments from  the  New Bedford
Harbor  Hot Spot Superfund Site
                    W. E. Whitworth and L. R. Waterland


                     A  detailed test program was per-
                    formed at the U.S. Environmental Pro-
                    tection Agency's (EPA's) Incineration
                    Research Facility (IRF) to define the
                    incineration characteristics of contami-
                    nated marine sediments from the Hot
                    Spot in  New  Bedford Harbor, a
                    Superfund site near New Bedford, MA.
                    The sediments at this site are contami-
                    nated with PCB concentrations of 4,000
                    to over 200,000  mg/kg, as well as with
                    trace metals, chiefly cadmium, chro-
                    mium, copper, and lead, at concentra-
                    tions up to several hundred mg/kg. The
                    test program was designed to evaluate
                    the effects  of incineration  operating
                    conditions on the composition of the
                    discharge streams.
                     Three incineration tests were per-
                    formed using native sediments spiked
                    with pure PCB  Askarel transformer
                    fluid. The third test also  included a
                    period  of  operation  with  native
                    (unspiked) sediment alone. Spiking was
                    performed to increase the sediment
                    PCB content from nominally 6,000 to
                    nominally 46,000 mg/kg, a  level that
                    allowed an unambiguous determination
                    of whether a regulatory level PCB de-
                    struction and removal efficiency (ORE)
                    of 99.9999% could be achieved For the
                    three spiked sediment tests, plans were
                    to vary kiln exit gas temperature from
                    816 to 982 °C (1,500 to 1,800 °F). Aver-
                    age  test temperatures achieved  were
                    824°C and 984  °C (1,516°C and 1,803
                    °F).
                     Test  results show that greater than
                    99.9999% PCB ORE (100 • [1- flue gas
                    emission rate/feedrate]) was achieved
                    at both kiln temperatures with the af-
                    terburner operated at 1,208°C (2,206°F).
                    However, with a kiln solids residence
                    time of 0.5  hr, the treated sediments
                    (kiln ash) were still PCB-contaminated.
                    In tests with the PCB-spiked sediment
                    feed, incinerated without dewatering,
                                  kiln ash contained 128 to 245 mg/kg of
                                  PCBs. The PCBs in the kiln ash dis-
                                  charge accounted for between 0.08%
                                  and 0.14% of the PCBs fed, regardless
                                  of kiln temperature. For a native
                                  (unspiked) sediment  feed  incinerated
                                  without dewatering, the kiln ash con-
                                  tained 100 mg/kg of PCBs, which ac-
                                  counted for 0.44% of the PCBs  intro-
                                  duced in the sediment feed. The wet
                                  scrubber system discharge  flue gas
                                  contained low levels of polychlorinated
                                  dibenzo-p-dioxins (PCDDs) and higher
                                  levels of dibenzofurans (PCDFs), chiefly
                                  total  tetra-CDF (TCDF),  penta-CDF
                                  (PeCDF), and hexa-CDF (HxCDF). The
                                  2,3,7,8-tetra-CDD (2,3,7,8-TCDD) toxic-
                                  ity equivalent emissions were in the
                                  nominal 0.05 to 0.10 ng/dscm range.
                                    Of  the contaminant trace  metals,
                                  chromium and  copper were relatively
                                  nonvolatile. The kiln ash discharge ac-
                                  counted for nominally 80% to 90% of
                                  the discharged amount of these met-
                                  als. These fractions were not affected
                                  by kiln temperature in the range tested.
                                  Cadmium and lead exhibited relatively
                                  volatile behavior, and increasingly so
                                  at the higher kiln temperature. At low
                                  kiln temperature the kiln ash discharge
                                  accounted for 53% of the lead and 61%
                                  of the cadmium discharged. At higher
                                  kiln temperature  these fractions de-
                                  creased to the nominal 10% to 20%
                                  range for cadmium and the 20% range
                                  for lead. Scrubber exit flue gas frac-
                                  tions (cadmium and lead) and scrubber
                                  liquor  fractions (cadmium) increased
                                  accordingly. Neither treated sediments
                                  nor the scrubber liquor were toxicity
                                  characteristic (TC) hazardous wastes
                                  based on their cadmium, chromium, or
                                  lead concentrations. However, the lead
                                  content of scrubber liquor toxicity char-
                                  acteristic leaching procedure (TCLP)
                                  leachate  was near the TC regulatory
                                  level. This suggests that the scrubber
                                                                   Printed on Recycled Paper

-------
 liquor discharge from a wet scrubber
 could be a TC hazardous waste In the
 Incineration of hlgher-than-tested lead-
 containing sediments, or under ex-
 tended scrubber operation at minimum
 blowdown.
  Test  results suggest that Incinera-
 tion would be an effective treatment
 option for the she sediments. However,
 sediment dewaterlng before Incinera-
 tion, and/or incinerating at longer kiln
 solids residence  times might be re-
 quired to yield a treated sediment not
 contaminated by PCBs. If a wet scrub-
 bar were used for particulate and acid
 gas control, the  scrubber blowdown
 discharge might require further treat-
 ment to stabilize teachable lead levels.
 In addition, lead levels In flue gas emis-
 sions from a wet scrubber air pollution
 control system (ARCS) might be of con-
 cern.
    This  Project Summary was devel-
 oped by EPA's Risk Reduction Engi-
 neering Laboratory, Cincinnati, OH, to
 announce key findings of the research
 project that Is fully  documented In a
 separata report of the same title (see
 Project Report ordering Information at
 back),

 Introduction
   EPA Region 1 Is conducting the reme-
 dial design (RD) for the remediation of a
 Superfund site  located in New Bedford
 Harbor near New Bedford, MA. According
 to the record of decision (ROD) document
 of 1990, the EPA has Identified approxi-
 mately 10,000 yd3 of contaminated sedi-
 ment in a 5-acre area  of the harbor. This
 area has been Identified as the Hot Spot
 Operable  Unit, and incineration of the
 dredged sediment has been selected as
 the treatment option.  EPA Region 1 re-
 quested that test burns be conducted at
 EPA's  IRF to support the RD for this
 Superfund site.
   The primary objective of this test pro-
 gram was to obtain data to support the RD
 plans and specifications. Therefore, the
test conditions were designed to evaluate
the effectiveness of varying incinerator op-
 erating conditions  in the destruction of
 PCBs and  other pollutants. Specifically,
thetest prog ram attempted to answerthese
questions:
  •  Can  incineration effectively destroy
    PCBs to the required regulatory level
    ORE of 99.9999%  In the flue gas
    emission?
  •  What is the distribution of the con-
    taminant trace metals in the discharge
    streams during incineration of the sedi-
    ment?
  • What are the effects of incineration
    excess  air and temperature on or-
    ganic constituent destruction and met-
    als distributions, including the leach-
    ability of the metals from the kiln ash?
  • What is the effectiveness  of the air
    pollution control system (APCS), with
    design and operating characterization
    similar to those of the IRF venturi/
    packed-column  scrubber, in  collect-
    ing particulate and trace metals?
  • Can the treated sediment (i.e., kiln
    ash) from the incinerator be disposed
    of as nonhazardous solid waste?
   The test program consisted of a set of
three incineration tests in the rotary kiln
incineration system (RKS) at the IRF. These
tests were aimed at evaluating PCB de-
struction and the fate of contaminant trace
metals in the sediment as functions of kiln
temperature and kiln excess air  level.

Test Program

Test Facility
   A process schematic of the RKS is
shown in Figure 1. The IRF RKS consists
of a primary combustion chamber, a transi-
tion section, and a fired afterburner cham-
ber. After exiting the afterburner, flue gas
flows through a quench section followed
by a primary APCS. The primary APCS for
these tests consisted of a venturi scrubber
followed by  a packed-column  scrubber.
Downstream of the primary APCS, a
backup secondary APCS, comprised of a
demister, an activated-carbon  adsorber,
and a  high-efficiency particulate (HEPA)
filter, is in place.
                             Test Waste Description
                                Eight 30-gal drums of sediments were
                             dredged from the Hot Spot area of New
                             Bedford Harbor for these tests. A charac-
                             terization sample representing each drum
                             was shipped to the IRF for pretest analy-
                             ses.  These  samples were subjected to
                             proximate, PCB, and hazardous constitu-
                             ent trace  metals analyses. The results of
                             these analyses showed that the average
                             total PCB concentration of the  eight drums
                             was 5,300 mg/kg  as received. The level
                             required in an RKS feed to be able to just
                             establish 99.9999% ORE at a typical RKS
                             feedrate of 68 kg/hr (150 Ib/hr) is 5,100
                             mg/kg, just below the average dredged
                             drum characterization sample level. Con-
                             sequently, it was decided to spike the test
                             sediment to higher PCB concentrations to
                             provide a margin in the ability  to establish
                             99.9999% ORE. The material used to spike
                             the sediments was an Askarel  transformer
                             fluid  comprised  of roughly 75% Aroclor
                             1242 and 25%  Aroclor  1254. Resulting
                             spiked feed levels were 3.48% Aroclor 1242
                             and 1.11% Aroclor 1254.
                                For the test program,  all eight drums of
                             sediment  were shipped to the  IRF, where
                             they were combined to form one test feed
                             material. Before testing, the combined sedi-
                             ments  were repackaged into l.5-gal
                             fiberpack containers for feeding to the RKS
                             via the ram feeder system. The PCB spike
                             was added to the sediments  during this
                             packaging. In addition to spiked sediments,
                             a number of fiberpack drums were pre-
                             pared without the PCB  spike for testing
                             using only the native sediment.
                              Single-stage Ionizing
                                 Wet Scrubber
   Transfer
   Duct
       Ash-
       Bin
                                     Scrubber Liquor
                                     Recirculation
       Afterburner
        Air
   Natural
   Gas,
   Liquid
   Feed
                                                                  Atmosphere
                                                           Stack
                                                          
-------
Test Conditions
   The test series was designed to evalu-
ate the effects of incinerator operating con-
ditions on PCB destruction and trace metal
distributions in the  incinerator discharge
streams. The operating parameters to be
varied were kiln exit gas temperature and
kiln excess air (exit flue gas  O2) Three
tests were to cover the range of target kiln
exit flue gas temperatures of 816 and 982
°C (1,500 and 1,800 °F) and target kiln exit
O2 levels of 6% to 10%. For all tests, the
operating conditions noted in Table 1 were
to be held at the nominal values noted in
the table.
   Table 2 summarizes the actual incin-
erator exit temperatures and flue gas lev-
els, including their ranges  and averages
for each test  during  flue gas sampling.
These are compared  with the respective
target conditions. During Test 3, unspiked
native sediment was fed to the kiln for a
period of time to collect kiln ash associated
with native sediment  feed  only. Test 3a
represents the period of native sediment
feed; Test 3b represents  the period  of
spiked sediment feed for the third test.
   For all tests, the average kiln exit gas
temperature was within 8°C (16°F) of the
respective target temperature.  The actual
O2  levels at the kiln  exit were generally
higher than the target  concentrations. The
higher O2 levels experienced resulted from
higher than expected air inleakage into the
kiln chamber  because of the  inability  to
tightly secure a rotating kiln seal. The mini-
mum O2 achievable was 9% at the kiln exit.
The maximum O2 tested was 11.2%. As a
practical matter, these two  levels present
comparable combustion environments.
Consequently, it was  not possible to test
kiln excess air as  a variable.
                                Table 1.  Incinerator System Operating Conditions Held Constant

                                  Operating Condition                                     Value
                                Kiln solids residence time
                                Total sediment feedrate
                                Scrubber blowdown rate
                                Venturi liquor flowrate
                                Venturi pressure drop
                                Packed tower liquor flowrate
                                Scrubber liquor temperature
                                                                0.5 hr
                                                                68.2 kg/hr (150 Ib/hr)
                                                                0 Umin (0 gpm) or minimum operable
                                                                76 L/min (20 gpm)
                                                                6.2 kPa (25 in WC)
                                                                115 Umin (30 gpm)
                                                                43PC (12CPF)
                                Sampling and Analysis
                                Procedures
                                   The scope of the sampling effort under-
                                taken during this test program is illustrated
                                in Figure 2, in which the sampling locations
                                are identified. Specifically, the  sampling
                                effort during each test consisted  of:
                                  • Collecting samples  of the  sediment
                                    feed  from  each drum  received,  as
                                    well as a composite sample from the
                                    mixing  container in which  all sedi-
                                    ments received  were blended before
                                    the sediments  were packaged  into
                                    feed fiberpacks.
                                  • Collecting a composite sample of the
                                    kiln ash.
                                  • Collecting a composite sample of the
                                    scrubber liquor.
                                  • Continuously  measuring O2 levels in
                                    the kiln  exit and afterburner exit  flue
                                    gases;  O2,  CO, CO2,  NO  and  total
                                    unburned hydrocarbon (TUHC) levels
                                    at  the  venturi/packed-column scrub-
                                    ber exit; and O2,
                                    in the stack.
                                       CO, and CO, levels
                                               • Sampling  flue gas at the scrubber
                                                 system exit for PAHs and PCBs.
                                               • Sampling  flue gas at the scrubber
                                                 system exit for PCDDs and PCDFs.
                                               • Sampling  flue gas at the scrubber
                                                 system exit for volatile organics.
                                               • Sampling  flue gas upstream of the
                                                 scrubber system for particle size dis-
                                                 tribution.
                                               • Sampling flue gas  upstream  and
                                                 downstream of the scrubber  system
                                                 for particulate and  trace metals using
                                                 a variation of EPA  Method 5 modified
                                                 for multiple metals  capture.
                                               • Sampling  downstream of  the scrub-
                                                 ber system  and at the stack down-
                                                 stream of the secondary APCS for
                                                 particulate and HCI using Method  5
                                                 to comply with permit  requirements.
                                                An aliquot of the composited sediment
                                             feed and  each test's kiln ash  was sub-
                                             jected to  the  TCLP leaching  procedure
                                             and analyzed for cadmium, chromium, cop-
                                             per, and lead. Waste feed samples, kiln
                                             ash samples, and scrubber liquor samples
                                             were analyzed separately for PCBs, PAHs,
 Table 2. Actual Versus Target Operating Conditions for the New Bedford Harbor Tests

                                     Temperature, °C fF)
                                                                                             Flue gas O2, %
Test

1
2
3a
3b
Date

3/15/91
3/19/91
3/21/91
3/21/91
Target

816(1,500)
982 (1,800)
982 (1,800)
982 (1,800)
Minimum

755(1,391)
945(1,733)
939(1,723)
968(1,775)
Maximum
Kiln exit
907(1,665)
1,022 (1,871)
1,010 (1,850)
999 (1,831)
Average

824 (1,516)
984 (1,803)
981 (1,797)
985 (1,805)
Target

6
6
10
10
Range

5.8 to 13.9
6.6 to 12.0
7.2 to 11.5
7.0 to 11.4
Average

11.2
9.0
9.3
10.0
 1
 2
 3a
 3b
3/15/91
3/19/91
3/21/91

3/21/91
1,204 (2,200)
1,204 (2,200)
1,204 (2,200)

1,204 (2,200)
1,195(2,183)
1,201 (2,194)
1,194(2,181)

1,198 (2,188)
Afterburner exit

 1,221 (2,229)
 1,213 (2,216)
 1,226 (2,239)

 1,214(2,217)
1,208 (2,206)
1,208 (2,206)
1,208 (2,206)

1,208 (2,206)
7
7
7

7
3.2 to 8.2
3.6 to 7.4
2.9 to 7.6
4.3 to 8.0
6.4
6.0
6.4

7.0

-------
  T
Kiln
\
fc-


After-
burner

^


Quench
Section
w

Venturi/PacKea
Column
Scrubber
1
te~


Demister

^

Carbon
Bed

fr.

HEPA
Filter

                                                   5         6
                                 Venturi/     Continuous    Method      Method      Method
Sam-    Sediment       Kin    Packed-Column  Flue Gas      0010        0010        0030
pKng      Feed        Ash    Scrubber Liquor Monitoring    (PCBsand    (PCDDs/     (Volatile
Point  (Composite)   (Composite  (Composite)    (CEMs)      PAHs)       OCDFs)     Organics)
                                                 Particle Size
                                                  Distribution
                                                   (Cascade
                                 Method   Method
                                 5 (Par-    5 (Par-
                                 ticulate,   ticulate,
                                                   Impactor)   Trace Metals)   HCI)
 1
2
3
4
5
6
7
      X
      X


      X
      X
X
                        X
X
X


X
X
X
Figure 2. Sampling matrix.
and cadmium, chromium, copper, and lead.
The composite sediment feed sample was
ateo subjected to proximate (moisture, ash
content, and heat content) analysis and
ultimate (C, H, O, N, S, Cl) analysis.

Test Results
   The results of the test program are
discussed in the subsections that follow.
Test results are grouped by analyte class.

Proximate and Ultimate
Analysis Results
   The proximate and ultimate analysis
results for the composite sediment sample
analyzed are presented in Table  3. The
high moisture content of the  test sedi-
ments  Is consistent with their marine ori-
gin. Table 4 summarizes the cumulative
sediment weight fed for each test and the
total amount of kiln ash collected. As indi-
cated in the table, between 25% and 30%
of the sediment weight fed for a given test
was collected as kiln ash. This fraction
agrees quite well with the ash content of
the sediment obtained by proximate analy-
sis shown in Table 3.

PCB, Semivolatile and Volatile
Organic, and Dioxin/Furan
Analysis Results
   Table 5 summarizes the PCB contents
of each incineration test sample. As noted
In the table, the spiked sediment feed con-
tained  3.48%  Aroclor 1242 and  1.11%
Arocior 1254. The kiln ash resulting from
the incineration  of the sediments (both
spiked and native), without dewatering, had
substantially reduced, though still  signifi-
cant, PCB contents. The  kiln ash  for the
spiked wet sediment feeds contained be-
tween 96 and 177 mg/kg of Aroclor 1242,
and between 32 and 84 mg/kg of Aroclor
1254. Interestingly, within the range of the
variability of the data, the higher kiln tem-
perature tested for Tests 2 and 3 did not
result In significantly lower kiln ash PCB
concentrations than the lowertemperature
tested in Test 1. The kiln ash resulting from
native sediment feed incinerated without
watering  also  contained significant PCB
levels, 57 mg/kg Aroclor 1242 and 44 mg/
kg Aroclor 1254.
   No scrubber liquor sample contained
detectable PCB at practical quantitation
Table 3.  Proximate and Ultimate Analysis Re-
         sults for the Composite Sediment
         Feed Sample


Proximate Analyses

    Moisture, %                  63.9

    Ash. %                      28.5
    Volatile matter, %              5.9

    Fixed carbon, %               1.4
    Higher heating value, kJ/kg    2,200

                     (Btu/lb    (948)
Ultimate Analysis, %

    C                           11.1
    H                            1.2

    N                            0.4
    S                            0.6
    Cl                           0.9
           limits (PQLs) of  1 ug/L for Aroclor 1242
           and 0.3 ug/L for Aroclor 1254. The scrub-
           ber exit flue gas contained  low, though
           measurable, levels of both PCB formula-
           tions in Tests 1 and 2.
              Table 6 summarizes the degree of PCB
           decontamination achieved in each test, in
           terms of the fraction of the amount of PCB
           introduced in the incineratorfeed accounted
           for by the resulting kiln ash. As shown in
           the table, about 0.1% of the Aroclor 1242
           and  about 0.1% to  0.2%  of the Aroclor
           1254 fed in spiked sediments  was ac-
           counted for in the kiln ash produced. The
           remaining 99.8% to  99.9% was  removed
           and  largely destroyed, as  discussed be-
           low.  Higher fractions of feed PCBs were
           present in the kiln  ash from the  native
           sediment test: 0.3% for Aroclor 1242 and
           0.9% for Aroclor 1254.
              The  data shown in Tables  5  and 6
           confirm that incineration under the condi-
           tions tested was not  sufficient  to com-
           pletely decontaminate the sediments. The
           incineration temperatures tested, 820 to
           980 °C (1,500 to 1,800 °F), were typical of
           those that have resulted in  successful de-
           contamination, as was the kiln solids resi-
           dence  time  (0.5 hr). The New Bedford
           Harbor marine sediments,  however, con-
           tained substantial moisture; the composite
           sediments tested were 64% moisture. Evi-
           dently, with such high moisture content,
           solids bed temperatures were  not raised to
           levels needed for more complete PCB de-
           struction in  the residence time available.
           Longer solids residence times, 1  hr in the
           IRF's experience, would have likely allowed
           more complete PCB decontamination. Fur-

-------
 Table 4.   Sediment Feed and Ash Collected
                                                           Ash Collected


1
2
3a
3b

Test
(3/15/91)
(3/19/91)
(3/21/981)
(3/22/91)

Total Sediment Fed,
kg (Ib)
284 (625)
284 (625)
229 (504)
318 (700)

Weight,
kg (Ib)
85 (187)
76 (168)
62 (136)
80 (175)
Fraction of
Feed
%
30
27
27
25
ther testing would be required to verify this
supposition.
   Table 7 combines the scrubber exit flue
gas PCB concentrations noted in Table 5
with sediment feedrate and flue gas flow-
rate data to give the PCB DREs achieved
for the tests. ORE is defined as:
        100 • emission rate/feedrate)
As shown in the Table 7, greater than the
regulation-required 99.9999% PCB ORE
was achieved for all three tests.
   No PAH compounds analyzed for were
detected in any sample at PQLs of 50 mg/
kg in sediment feed, 1.3 mg/kg in kiln ash,
20 ug/L in scrubber liquor, and 6 ug/dscm
in scrubber exit flue gas.
   Results of the scrubber exit flue gas
PCDD/PCDF measurements showed that
total TCDD, PeCDD, HxCDD, and hepta-
CDD (HpCDD) levels were in the nominal
0.01 to 0.02  ng/dscm range for all three
tests, with octa-CDD (OCDD) levels in the
0.03 to 0.06 ng/dscm range. Flue gas PCDF
levels  were significantly greater,  ranging
up to about 2.8 ng/dscm for total TCDF.
The 2,3,7,8-TCDD toxicity equivalents cor-
responding to the isomer concentrations
measured from nominally 0.05 to 0.10 ng/
dscm over the three tests.

Trace Metal Discharge
Distributions
   Two primary objectives of the test pro-
gram  were (I) to evaluate the fate of the
contaminant trace metals cadmium, chro-
mium, copper, and lead in the incineration
treatment of the New Bedford Harbor sedi-
ments and  (2) to  investigate whether
changes in incineration conditions affected
the distribution of these metals in the incin-
erator discharges. Table 8 summarizes the
concentrations of the test metals in sedi-
ment samples and in each of the incinera-
tor discharge streams.
   The concentrations  of chromium and
copper in resulting kiln ashes were higher
than the composite sediment feed sample
for all tests and were higher for cadmium
and lead for the low-kiln-temperature test
(Test 1). This  reflects the weight reduction
in converting sediment to kiln ash during
incineration. However, the cadmium and
lead concentrations in kiln ash were sig-
nificantly lower in the high-kiln-tempera-
ture tests (Tests 2 and 3) when compared
to the Test 1 kiln ash concentrations and
were also lower than the corresponding
sediment feed concentrations. Flue gas
cadmium and lead concentrations, both in
the afterburner exit flue gas and the scrub-
ber exit flue gas, were generally higher for
Tests 2 and 3 than for Test 1. Both trends
are the result of the volatile behavior of
these two metals. The extent of volatiliza-
tion of these metals was evidently higher
in Tests 2 and 3 than in Test 1, giving rise
to lower kiln ash, and generally higherflue
gas, concentrations of these metals for the
highertemperature tests.
   Table 8 also notes the TCLP regulatory
limit for the three TCLP metals determined.
Comparing composite feed  and kiln ash
Table 5.  PCB Analysis Results
         Sample
  Spiked sediment feed, %
TCLP leachate and scrubber liquor metal
concentrations to the TCLP regulatory lev-
els shows that neither the composite sedi-
ment feed nor any test's kiln ash would be
a TC hazardous waste based on teachable
cadmium, chromium, or lead concentra-
tions.
   The scrubber liquor cadmium and chro-
mium concentrations noted in Table 9 are
below  the  TCLP regulatory level for all
three tests. In contrast, the scrubber liquor
lead concentrations exceed the regulatory
level for all three tests. However, the scrub-
ber liquor metal concentrations noted  in
the table are for the total scrubber liquor,
which contains suspended solids. A true
TCLP  leachate  was prepared from the
three-test composite scrubber liquor that
was held in a storage tank until all analy-
ses were completed. This leachate was
analyzed for lead and found to contain 4.2
mg/L, less than the TCLP regulatory level.
Thus, the scrubber liquor for these tests
was also not a TC hazardous waste.
   The metal concentrations  shown  in
Table 8 can be  combined with feed soil
and discharge stream mass flowrate infor-
mation to better show how the metals dis-
tribute among the discharge streams as a
function of incineration condition. These
distributions are summarized in Table 9.
The distribution fractions in Table 9 have
been normalized to the total amount of
each metal measured in all the discharge
streams analyzed. Thus, these normalized
values represent fractions that would have
resulted had mass balance closure in each
                                                     PCB Concentration
   Aroclor 1242
                       Aroclor 1254
  Test 1
    Kiln ash, mg/kg
    Scrubber liquor, ug/L
    Scrubber exit flue gas, ug/dscm

  Test 2
    Kiln ash, mg/kg
    Scrubber liquor, ug/L
    Scrubber exit flue gas, ug/dscm
3.48

 133

0.76


  96

0.54
                            84
                           <0.3
                           0.22
                            32
                           <0.3
                           0.21
Test 3b
Kiln as/7, mg/kg
Scrubber liquor, ug/L
Scrubber exit flue gas, ug/dscm
Composite native feed, mg/kg
Test 3a
Kiln ash, mg/kg
Scrubber liquor, ug/L

177
<1
<0.26
4,850

57
<1

68
<0.3
<0.09
1,300

44
<0.3

-------
Tabt* ft  PCS Decontamination Effectiveness

          Parameter
   Aroclor 1242
                        Aroclor 1254
    Sediment feed
        Concentration, %
        Amount fed, kg
    KUnAsh
        Concentration, mg/kg
        Amount discharged, g
        Fraction of amount fed, %

Tost 2
    Sediment feed
        Concentration, %
        Amount fed, kg
    Kin Ash
        Concentration, mg/kg
        Amount discharged, g
        Fraction of amount fed, %

Toatto
    Sediment teed
        Concentration, %
        Amount fed, kg
    Kiln Ash
        Concentration, mg/kg
        Amount discharged, g
        Fraction of amount fed, %

Teat 3b
    Sediment feed
        Concentration, %
        Amount fed, kg
    Kiln Ash
        Concentration, mg/kg
        Amount discharged, g
        Fraction of amount fed, %
      3.48
      9.87


       133
      11.3
      0.11
      3.48
      9.87


       96
       7.3
      0.07
     0.485
      1.11


        57
       3.5
      0.32
      3.48
     11.05


       177
      14.1
      0.13
           1.11
           3.15


            84
            7.1
           0.22
           1.11
           3.15


            32
            2.5
           0.08
          0.130
           0.30


            44
            2.7
           0.91
           1.11
           3.52


             68
            5.3
           0.15
T*W«7. PCBDREs
           Parameter
 Test 1
(3/15/91)
 Test 2
(3/19/91)
 Test 3b
(3/21/91)
Sediment feed
    Sediment feedrate, kg/hr                     69.5            69.5           69.3
    Aroclor 1242 foodrate, g/hr                  2,420          2,420           2,410
    Aroclor 1254 feedrate, g/hr                    773            773            771
Scrubber exit flue gas
    FhiB gas flowrate, dscm/mln                  33.9            32.9           30.6
    Aroclor 1242 concentration ug/dscm          0.76            0.54           <0.26
                emission rate, ug/hr              1.6             1.1           <0.5
                ORE, %                 99.999936       99.999956      >99.999980
    Aroclor 1254 concentration ug/dscm          0.22            0.21           <0.09
                emission rate, ug/hr              0.5             0.5           <0.2
                ORE, %                  99.99994        99.99994       >99.99998
case been 100%. Use of distribution frac-
tions  normalized in this  manner allows
clearer data interpretation, because they
remove variable mass balance closure as
a source of test-to-test data variability.
   Actual mass balance closures achieved
around the conventional incineration sys-
tem portion of the RKS ranged from 52%
to 103% for  cadmium, 72%  to 79% for
chromium, 71% to 86% for copper, and
38% to 66% for lead. These levels  are
considered good compared with past ex-
perience on achieving trace metal mass
balance closures from a variety of com-
bustion sources, including incinerators.
Typical mass balance closure  results from
this past experience have been, at best, in
the 30% to 200% range.
   Several  interesting  observations
emerge from the data in Table 9. Chro-
mium  and copper exhibited relatively
nonvolatile behavior. The kiln ash dis-
charge represented the predominant
fraction of the discharged amount for
these two metals: kiln ash  accounted
for 88% to  92% of the  chromium and
82% to 89% of the copper.  These dis-
tributions  were not affected by kiln
temperature in the range tested (i.e.,
824 to 985 °C [1,516 to 1,805 °F]).
   In contrast, cadmium  and  lead exhib-
ited relatively volatile behavior. At the low-
kiln temperature test  (Test I) conditions,
the kiln ash accounted for 53% (lead) to
61% (cadmium) of the metals measured in
the discharges. Even at this relatively low
incineration  temperature, a significant
amount of each metal evidently vaporized
in the kiln and was carried into the after-
burner and downstream  to augment the
amount entrained in flyash carried out of
the kiln.
   Evident vaporization was enhanced at
the high-temperature-test (Tests 2 and 3)
condition. For these  tests, the kiln  ash
accounted for significantly decreased frac-
tions  of cadmium and lead measured in
the discharges, 8% to 19% for cadmium
and  19% to 23% for lead. Of the metals
measured and accounted for, scrubber exit
flue gas and scrubber liquor fractions were
higher than the kiln ash  fractions at  this
high-temperature condition. Indeed, 36%
to 42% of the cadmium  discharged and
42% to 55% of the lead measured in the
discharges escaped the incineration sys-
tem and the venturi/packed-column scrub-
ber.
   Table  10  summarizes the  apparent
scrubber collection efficiencies calculated
for each metal measured in the test pro-
gram. In  calculating  apparent  collection
efficiency, it is assumed that  the flowrate
of metals at  the scrubber inlet can be

-------
Table 8. Trace Metals Analysis Results
Sample Cd
Sediment feed
Composite, mg/kg 7.4
Composite TCLP leachate, mg/L 0.11
Test 1 (3/15/91)
Kiln ash, mg/kg 9.5
Kiln ash TCLP leachate, mg/L 0.26
Afterburner exit flue gas, ug/dscm 42.3
Scrubber exit flue gas, ug/dscm 34.9
Scrubber liquor, mg/L 0. 14
Test 2 (3/19/91)
Kiln ash, mg/kg 2.7
Kiln ash TCLP leachate, mg/L 0.046
Afterburner exit flue gas, ug/dscm 77.6
Scrubber exit flue gas, ug/dscm 55.3
Scrubber liquor, mg/L 0.27
Test 3a (3/21/91)
Kiln ash, mg/kg 2.3
Test3b (3/21/91)
Kiln ash, mg/kg 2.0
Kiln ash TCLP leachate, mg/L 0.043
Afterburner exit flue gas, ug/dscm 51.6
Scrubber exit flue gas, ug/dscm 79.0
Scrubber liquor, mg/L 0.73
TCLP regulatory level, mg/L 1.0
• = Not a TCLP metal.


Table 9. Normalized Trace Metal Distributions
Test 1 2
(3/15/91) (3/19/91)
Kiln exit temperature, °C 824 984
(°F) (1,516) (1,802)
Kiln exitOs, % 11.2 9.0

Cr

161
0.041

376
0.048
219
158
1.9

434
0.030
136
73.2
1.4

367

357
0.03
126
83.9
1.3
5.0




3b
(3/21/91)
985
(1,805)
10.0

Cu Pb
-
308 236
0.066 1.2

608 277
6.3 0.71
571 1,030
421 903
4.5 8.8

828 75.6
3.01 0.17
768 1,814
436 1,273
2.6 5.4

785 96

721 62
3.0 0.41
519 984
750 2,020
3.4 5.6
—' 5.0








Apparent distribution, % of metal measured


Cadmium
Kiln ash 61 19
Scrubber exit flue gas, 23 42
Scrubber liquor 16 39
Total 100 100

Chromium
Kiln ash 88 92
Scrubber exit flue gas, 4 2
Scrubber liquor 8 6
Total 100 100
Copper
Kiln ash 83 89
Scrubber exit flue gas, 6 5
Scrubber liquor 11 6
Total 100 100
Lead
Kiln ash 53 23
Scrubber exit flue gas, 17 42
Scrubber liquor 30 35
Total 100 100



8
36
56
100


92
2
6
100

82
10
8
100

19
55
26
100








obtained by summing the flows in the two
scrubber discharge streams: the scrubber
exit flue gas and the scrubber liquor. Ap-
parent scrubber collection efficiency is de-
fined as the scrubber liquorfraction divided
by the sum of the scrubber liquor fraction
and the scrubber exit flue gas fraction. The
data in Table 1 0 show that, at the low-kiln-
temperature test condition, apparent col-
lection efficiencies for chromium, copper,
and lead were comparable, and in the
nominal 65% to 70% range. Collection effi-
ciency for cadmium was lower at 41%. At
the high-kiln-temperature test conditions,
cadmium and chromium efficiencies were
relatively unchanged. Copper and lead col-
lection efficiencies decreased.
Paniculate and HCI Emissions
Data
For the three tests, flue gas particulate
levels at the scrubber exit ranged from 70
to 1 01 mg/dscm (0.03 to 0.04 grains/dscf),
corrected to 7% O2. These levels were
below the 180 mg/dscm at (0.08 grains/
dscf), 7% O2, hazardous waste incinerator
performance standard.
The sediments incinerated during this
test program contained 0.85% chlorine. "
Measured HCI concentrations at the scrub-
ber exit ranged from 0.2 to 2.4 ppm, with
corresponding emission rates ranging from
0.7 to 7.2 g/hr. These emission rates were
less than the hazardous waste incinerator
performance standard floor of 2 kg/hr. The
scrubber system HCI collection efficien-
cies ranged from 98.8% to 99.9% of the
chlorine fed.
Conclusions
Test conclusions are as follows:
• Greater than 99.9999% ORE of the
PCBs in the site sediments can be
achieved in the scrubber exit flue gas,
as calculated by the regulatory defini-
tion, at incineration temperatures of
both 824°C (1,516°F) and 984°C
(1 ,803°F) in a rotary kiln with an af-
terburner operated at 1,208°C
(2,206°F). However, when sediments
that had not been dewatered were


Table 10. Apparent Scrubber Collection Efficiencies
Test

7 2 3b
(3/15/91) (3/19/91) (3/21/91)
Kiln exit temperature, "C 824 984 985

Kiln exitOz, %

Cadmium
Chromium
Copper
Lead
(°F) (1,516) (1,803) (1,805)
1 1.2 9.0 10.0
Apparent Scrubber Collection Efficiency, %
41 48 50
69 79 62
66 53 33
64 45 23
•U.S. Government Printing Office: 1992— 648-080/60068

-------
   incinerated at a kiln solids residence
   time of 0.5 hr, the treated sediments
   (kiln ash) were still PCB-contaminated.
   In tests with a PCB-spiked sediment
   feed, the kiln ash discharge accounted
   for between 0.08% to 0.14% of the
   FOB fed regardless of kiln tempera-
   ture.  For a native  (unspiked) sedi-
   ment feed, the kiln ash accounted for
   0.44% of the PCBs feed.
   Of the contaminant trace metals, chro-
   mium and copper were relatively non-
   volatile. The kiln ash discharge ac-
   counted for nominally 80% to 90% of
   the measured discharged amounts of
   these  metals. These fractions were
   not affected by the range of kiln tem-
   peratures tested.
   Of the contaminant trace metals, cad-
   mium  and lead exhibited  relatively
   volatile behavior, and increasingly so
   at the high kiln exit gas temperature
   (984°C [1,803°F]). The kiln ash dis-
charge  contained 53% of the lead
and 61% of the cadmium accounted
for in the discharges at the low kiln
temperature (824°C [1,516°F]). These
fractions decreased to the nominal
10% to  20% range for cadmium and
the 20% range for lead at the high
kiln temperature.  Scrubber exit flue
gas fractions (cadmium and lead) and
scrubber liquor fractions  (cadmium)
increased accordingly. Mass balance
closures achieved were 52% to 62%
for cadmium in two of the three tests
performed, and ranged from  38% to
66% for lead over the three tests.
Based on recovered scrubber liquor
-and scrubber exit flue gas concentra-
tions, apparent scrubber collection ef-
ficiencies were in the nominal 65% to
70% range for  chromium, copper, and
lead at the low kiln temperature, and
lower, at 41%,  for cadmium. Cadmium
    and chromium collection efficiencies
    were apparently unaffected  by in-
    creased kiln temperature,  although
    copper and lead collection efficiency
    decreased to the 33% to 53% range
    for copper and the 23% to 45% range
    for lead.
  • Neither treated sediments nor the
    scrubber liquor discharges would ex-
    hibit the TC based on their cadmium,
    chromium, or lead concentrations.
   The test results suggest that incinera-
tion would be an effective treatment option
for the site sediments. However, sediment
dewatering  before incineration, or inciner-
ating at higher kiln solids residence times
(perhaps up to 1 hr), might be required to
yield a treated sediment not contaminated
by PCBs.
   The full  report was submitted in fulfill-
ment of Contract No. 68-C9-0038 by Acurex
Corporation under the sponsorship of the
U.S. Environmental Protection Agency.
   W. E. Whttworth and L R. Waterland are with Acurex Corporation, Jefferson, AR
     72079.
   R.C. Thurnau is the EPA Project Officer (see below).
   The complete report,  entitled "Pilot-Scale Incineration of PCB-Contaminated Sedi-
     ments from the New Bedford Harbor Hot Spot Superfund Site," (Order No. PB92-
     178953/AS; Cost: $26.00; subject to change) will be available only from:
           National Technical Information Service
           5285 Port Royal Road
           Springfield, VA 22161
           Telephone: 703-487-4650
   The EPA Project Officer can be contacted at:
           Risk Reduction Engineering Laboratory
           U.S. Environmental Protection Agency
           Cincinnati, Ohio 45258
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268

Official Business
Penalty for Private Use
$300
                                                       BULK RATE
                                                 POSTAGE & FEES PAID
                                                           EPA
                                                     PERMIT No. G-35
EPA/600/SR-92/068

-------