EPA/600/R-92/051
March 1992
OPERATIONS AND RESEARCH AT
THE U.S. EPA INCINERATION RESEARCH
FACILITY: ANNUAL REPORT FOR FY91
By
L. R. Waterland
Acurex Corporation
Environmental Systems Division
Incineration Research Facility
Jefferson, Arkansas 72079
EPA Contract 68-C9-0038
EPA Project Officer: R. C. Thurnau
Waste Minimization, Destruction, and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
Printed on Recycled Paper
-------
NOTICE
The information in this document has been funded wholly or in part by the United
States Environmental Protection Agency under Contract 68-C9-0038 to Acurex Corporation. It
has been subjected to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
u
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial products and
practices frequently carry with them the increased generation of materials that, if improperly
dealt with, can threaten both public health and the environment. The U.S. Environmental
Protection Agency is charged by Congress with protecting the Nation's land, air, and water
resources. Under a mandate of national environmental laws, the Agency strives to formulate
and implement actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life. These laws direct the EPA to perform
research to define our environmental problems measure the impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning, implementing,
and managing research, development, and demonstration programs to provide an authoritative,
defensible engineering basis in support of the policies, programs, and regulations of the EPA
with respect to drinking water, wastewater, pesticides, toxic substances, solid and hazardous
wastes, and Superfund-related activities. This publication is one of the products of that
research and provides a vital communication link between the researcher and the user
community.
This document reviews the accomplishments at the Incineration Research Facility
(IRF) in Jefferson, Arkansas, during Fiscal Year 1991. In that twelve-month period, five
major test programs were completed at the facility. Three major EPA Program/Regional
Office programs were supported through test activities: the hazardous waste incinerator
regulation development program within the Office of Solid Waste (OSW); the land disposal
restriction regulation development program within OSW; and the Superfund site remediation
program within the Office of Emergency and Remedial Response (OERR) as administered by
EPA Regions 1, 2, and 3. A sixth test program was done for Region 2, but was not as
extensive in scope as the others. The report outlines all efforts completed or ongoing at the
facility during FY91.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
m
-------
ABSTRACT
The U.S. Environmental Protection Agency's Incineration Research Facility (IRF) in
Jefferson, Arkansas, is an experimental facility that houses two pilot-scale incinerators and the
associated waste handling, emission control, process control, and safety equipment; as well as
onsite laboratory facilities.
During fiscal year 1991, five major test programs were completed at the facility: tests
to establish residue characteristics from the incineration of spent potliners from aluminum
production (K088) for the Office of Solid Waste (OSW); an evaluation of the incinerability of
five contaminated soils from the Drake Chemical Superfund site for Region 3; an evaluation of
the incinerability of PCB-contaminated marine sediments from the New Bedford Harbor
Superfund site for Region 1; a parametric evaluation of the fate of trace metals in a rotary kiln
incinerator equipped with a Calvert high-efficiency scrubber system; and an evaluation of the
incinerability of arsenic-contaminated soil from the Chemical Insecticide Corporation Superfund
site for Region 2. A sixth test program consisted of an evaluation of the effectiveness of
low-temperature thermal desorption in decontaminating wastes from the Caldwell Trucking
Superfund site for Region 2, but the program was not as extensive in scope as the others. In
addition, the results of a test program completed in FY90, an evaluation of the thermal stability-
based principal organic hazardous constituent (POHC) incinerability ranking for the OSW, were
reported. The report outlines all efforts completed or ongoing at the facility during FY91.
IV
-------
TABLE OF CONTENTS
Page
FOREWORD, iii
ABSTRACT , ,... jv
FIGURES .;.''. viii
TABLES ix
1 INTRODUCTION . . . 1
2 PARAMETRIC TESTING TO EVALUATE THE PROPOSED POHC
INCINERABILITY RANKING .... .............. 4
2.1 TESTPROGRAM .r .................. 5
2.1.1 Synthetic Waste Mixture 5
2.1.2 Test Conditions 7
2.1.3 Sampling and Analysis Procedures 8
2.2 TEST RESULTS 10
2.2.1 Test 1—Baseline Incineration Conditions 14
2.2.2 Test 2—Thermal Failure (Quenching) 14
2.2.3 Test 3—Mixing Failure 15
2.2.4 Test 4—Matrix Failure 15
2.2.5 Test 5—Worst-Case Combination , 16
2.3 CONCLUSIONS . 16
3 INCINERATION TESTS OF SPENT POTLINERS FROM THE PRIMARY
REDUCTION OF ALUMINUM (K088) 18
3.1 TEST PROGRAM 18
3.1.1 Test Conditions 19
3.1.2 Sampling and Analysis Procedures 21
-------
TABLE OF CONTENTS (continued)
Section
3.2
Page
TEST RESULTS ' , 24
3.2.1 Proximate, Ultimate, and Silica Analysis Results 24
3.2.2 Cyanide and Semivolatile Organic Analysis Results 26
3.2.3 Trace Metal Analysis Results 29
3.2.4 Sulfide and Fluoride Analysis Results , 34
3.2.5 Flue Gas Particulate and HC1 Emissions 37
3.3
CONCLUSIONS 37
INCINERATION OF CONTAMINATED SOILS FROM THE DRAKE
CHEMICAL SUPERFUND SITE 40
4.1
TEST PROGRAM 41
4.1.1 Test Conditions 42
4.1.2 Sampling and Analysis Procedures 44
4.2
TEST RESULTS 46
INCINERATION OF PCB-CONTAMINATED SEDIMENTS FROM THE
NEW BEDFORD HARBOR SUPERFUND SITE 49
5.1
TEST PROGRAM 50
5.1.1 Test Conditions 50
5.1.2 Sampling and Analysis Procedures 51
5.2
TEST RESULTS 54
5.2.1 Proximate and Ultimate Analysis Results 54
5.2.2 PCB, Semivolatile and Volatile Organic, and Dioxin/Furan
Analysis Results 54
5.2.3 Trace Metal Discharge Distributions 58
5.2.4 Particulate and HC1 Emissions Data 63
5.3
CONCLUSIONS 65
VI
-------
TABLE OF CONTENTS (continued)
Section
6
10
11
12
Page
FATE OF TRACE METALS IN THE ROTARY KILN SYSTEM WITH A
CALVERT FLUX/FORCE CONDENSATION SCRUBBER 67
6.1
TEST PROGRAM .... . .... ...... .... 69
6.1.1 Synthetic Waste Mixture :...•. .'. ... ... . . . . . .rv . . :'.-v~ . . 71
6.1.2 Test Conditions ..; 71
6.1.3 Sampling and Analysis 75
6.2 TEST RESULTS . 77
INCINERATION OF ARSENIC-CONTAMINATED SOILS FROM THE
CHEMICAL INSECTICIDE CORPORATION SUPERFUND SITE
7.1 TEST PROGRAM
80
80
7.1.1
7.1.2
Test Waste Description ........;. ...... ... 81
Test Conditions 84
7.1.3 Sampling and Analysis 84
7.2 TEST RESULTS .. ^.. ^ ..:....:.......: i I/. .'....• 86
: , • - 'w ' ' '
FACILITY PHYSICAL PLANT IMPROVEMENTS 87
8.1 OFFICE SPACE . 87
8.2 INCINERATION SYSTEM IMPROVEMENTS '. 87
8.3 FLAMMABLE CHEMICAL STORAGE BUILDING 88
8.4 BENCH-SCALE THERMAL TREATABILITY TEST UNIT 88
HEALTH AND SAFETY, ENVIRONMENTAL COMPLIANCE, AND
PERMIT ADMINISTRATION '. . . . . 92
9.1 TOXIC SUBSTANCE CONTROL ACT RESEARCH AND
DEVELOPMENT PERMIT . . . . . 92
9.2 RCRA FACILITY INVESTIGATION ......... . 93
THIRD-PARTY TESTING . . 94
EXTERNAL COMMUNICATIONS 95
PLANNED EFFORTS FOR FY92 . 101
REFERENCES 103
Vll
-------
FIGURES
Number
1 Schematic of the IRF rotary kiln incineration system.
2 Sampling matrix
3 Kiln exit POHC DREs for Test 1
4 Kiln exit POHC DREs for Test 2. ...
5 Kiln exit POHC DREs for Test 3
6 Kiln exit POHC DREs for Test 4.
7 Kiln exit POHC DREs for Test 5
8 Sampling matrix
9 Test sampling locations.
10 Sampling matrix.
11 Afterburner exit particle size distributions
12 Schematic of the Calvert Scrubber System
13 Test sampling locations
14 Sampling matrix
15 TTU: external configuration
16 TTU: internal configuration
Page
. 6
. 9
11
11
12
12
13
22
47
53
65
70
76
85
90
91
vui
-------
TABLES
Number
1 Synthetic waste mixture composition .....' • 7
2 Target test conditions • 8
3 Target test conditions , , 20
4 Actual versus target operating conditions for the K088 Tests . . . 20
5 Waste feed and ash collected • • • • > •. • 21
6 Trace metals determined 24
7 Proximate and silica analysis results 25
8 Test program composite proximate component distributions 25
9 Cyanide analysis results 27
10 Test program composite cyanide distributions 28
11 Cyanide DREs . . .. . : • • • • • 28
12 BDAT trace metal analysis results 30
13 Nonhazardous constituent metal analysis results 32
14 Test program composite BDAT trace metal distributions 35
15 Test program composite nonhazardous constituent metal distributions 35
16 Sulfide and fluoride analysis results for waste feed and kiln ash samples 36
17 Test program sulfide and fluoride distributions 38
18 Flue gas particulate levels 38
19 Stack gas HC1 emissions 38
20 Target test conditions 43
21 Target versus actual operating conditions for the Drake chemical soil tests . . 45
IX
-------
TABLES (continued)
Number page
22 Soil feed and ash collected .......=.. 45
23 Incinerator system operating conditions held constant......; ;. 50
24 Actual versus target operating conditions for the New Bedford Harbor tests . 52
25 Proximate and ultimate analysis results for the composite sediment feed
sample 55
26 Sediment feed and ash collected 55
27 PCB analysis results 56
28 PCB decontamination effectiveness 57
29 PCB DREs 58
30 Flue gas PCDD/PCDF analysis results 59
31 Trace metals analysis results 60
32 Normalized trace metal distributions 62
33 Apparent scrubber collection efficiencies 64
34 Flue gas particulate levels 64
35 Flue gas HC1 levels 64
36 Test organic liquid mixture compositions 72
37 Metal spike concentrations 73
38 Target test conditions 74
39 Test conditions held constant 75
40 Actual versus target operating conditions for the calvert scrubber trace
metal tests 73
-------
TABLES (continued)
Number Page
41 Synthetic waste fed and ash collected 79
42 Soil characterization sample analysis results 82
43 Soil characterization sample TCLP leachate analysis results 83
44 IRF program reports and presentations in FY91 . . . . . . 96
45 Visitors to the IRF 98
XI
-------
-------
SECTION 1
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) Incineration Research Facility
(IRF) in Jefferson, Arkansas, is an experimental facility that currently houses two pilot-scale
incinerators (a rotary kiln incineration system and a k'quid injection incineration system) and
their associated waste handling, emission control, process control, and safety equipment. The
IRF also has onsite laboratory facilities for waste characterization and analysis of process
performance samples.
The objectives of research projects conducted at the IRF have been and continue to be
as follows:
• To develop technical information on the performance capability of the hazardous
waste incineration process to assist EPA Regional Offices and state environmental
agencies in the review, assessment, and issuance of reasonable and responsible
permits for regulated hazardous waste incineration facilities, and to assist waste
generators and incinerator operators in the preparation of permit applications
• To develop incinerator system performance data for regulated hazardous wastes
to support current Resource Conservation and Recovery Act (RCRA) incinerator
regulations and performance standards, and to provide a sound technical basis for
any necessary future standards
• To promote an understanding of the hazardous waste incineration process and
develop methods to predict the performance of incinerators of varying scale and
design for the major classes of incinerable hazardous wastes as a function of key
process operating variables
• To develop methods of improving reliability and control of the incineration
process, including the use of destruction and removal efficiency (DRE) surrogates
• To provide a means of conducting specialized test burns (particularly for high
hazard or special waste materials such as Superfund site wastes) iri support of
specific Regional Office permitting or enforcement actions and Regional Office
or private party Superfund site remediation efforts
• To test the performance of new and advanced incinerator components and
subsystems, and emission control devices
-------
Fiscal year 1991 (FY91, October 1, 1990 through September 30, 1991) saw the return
of production-paced incineration testing to the IRF after the final completion of a major facility
ejqpansion and reconfiguration construction effort begun in FY89. Beginning in December 1990
and continuing through August 1991, incineration testing proceeded virtually uninterrupted.
Over this time period, five test programs were completed. These test programs focused on the
objectives above.
Three major EPA Program/Regional Office programs were supported through test
activities in FY91.
• The hazardous waste incinerator regulation development program within the
Office of Solid Waste (OSW), via testing to evaluate the fate of trace metals fed.
to a rotary kiln incinerator equipped with a Calvert Flux Force/Condensation
Scrubber system
• The land disposal restriction regulation development program within OSW, via
incineration testing of spent potliner waste from the primary reduction of
aluminum (listed waste K088) to develop treatment residuals characteristics
• The Superfund site remediation program within the Office of Emergency and
Remedial Response (OERR) as administered by
— EPA Region 3, via incineration treatability testing of contaminated soil from
the Drake Chemical Superfund site in Lock Haven, Pennsylvania
— EPA Region 1, via incineration treatability testing of PCB-contaminated
sediments from the New Bedford Harbor Superfund site in New Bedford,
Massachusetts
— EPA Region 2, via incineration treatability testing of arsenic-contaminated
soil from the Chemical Insecticide Corporation Superfund site in Edison,
New Jersey
In addition, the results of a series of tests to evaluate the principal organic hazardous constituent
(POHC) thermal-stability-based incinerability ranking, completed in support of OSW in FY90,
were assembled and reported in FY91. Also, the test planning documents (test plan and quality
assurance project plan [QAPjP]) for two additional Superfund site contaminated soil treatability
tests in support of Regions 2 and 5 were completed.
Activities completed during FY91 are discussed in more detail in the following sections.
Section 2 presents the results of the POHC incinerability ranking evaluation tests. Section 3
presents results of the K088 incineration residuals characterization tests. Section 4 discusses
results of the Drake Chemical Superfund site treatability tests. Section 5 presents results of the
New Bedford Harbor Superfund site treatability tests. Section 6 discusses the Calvert scrubber
trace metal tests. Section 7 discusses the Chemical Insecticide Corporation (CIC) Superfund site
treatability tests. Section 8 describes various facility improvement and capability enhancement
activities completed during FY91. Section 9 discusses environmental compliance and permit
administration activities. Section 10 outlines third-party test solicitation efforts. Section 11
-------
discusses external communication activities associated with the facility and its operation.
Section 12, the final section, presents an outline of plans for activities to be completed in FY92.
In addition, FY91 saw the completion of a brief series of scoping tests for EPA Region 2
to evaluate low-temperature thermal desorption as a treatment approach for contaminated soil
from the Caldwell Trucking Superfund site in Fairfield, New Jersey. Results from these tests are
not discussed in this report. Because this series of .tests was scoping in nature and performed
only to supply qualitative data to Region 2, no formal test report was prepared.
-------
SECTION 2
PARAMETRIC TESTING TO EVALUATE
THE PROPOSED POHC INCINERABILITY RANKING
One of the primary functions of the IRF is to conduct research activities for OSW in
support of regulation development and implementation. One major regulatory issue of high
priority during 1990 concerned the evaluation of an incinerability ranking system for POHCs.
Such a system was developed over the past several years by the University of Dayton Research
Institute (UDRI) under contract to EPA's Risk Reduction Engineering Laboratory (RREL).
The 1981 hazardous waste incinerator regulations require that an incinerator undergo
a trial burn performance test in order to become permitted to operate. This trial burn is
required to show that the incinerator is capable of achieving the mandated 99.99 percent POHC
ORE- In trial burn planning, the incinerator operator is required to select POHCs using two
criteria: concentration in the waste and difficulty to thermally destroy, or "incinerability." The
incinerability ranking included in the 1981 regulations was based on compound heat of
combustion.
The heat of combustion ranking has several acknowledged deficiencies, however. Thus,
EPA initiated studies to define or develop alternate, more suitable incinerability ranking
approaches. One such approach is the thermal-stability-based POHC ranking, developed by
UDRI. This ranking is based on the temperature required to achieve 99 percent destruction at
2 s residence time under oxygen-starved conditions as measured in laboratory experiments. As
of early 1990, the fundamental basis supporting the ranking approach had been documented and
sufficient information to rank the organic hazardous constituents had been collected. Since it
is based only on laboratory-scale data, evaluation of the thermal stability POHC incinerability
ranking under actual incineration conditions became a high-priority research need during 1990.
The test program described in this section was designed to develop the data to evaluate
the POHC incinerability ranking at the pilot scale. The specific objective of the test program
was to measure the ORE of a number of POHCs under each of several modes of incinerator
operation, and compare relative POHC DREs as a function of incineration conditions and feed
characteristics. The comparison would facilitate a determination of how relative POHC DREs
compared with expectations based on the thermal stability ranking.
In the tests, a mixture of 12 POHCs with predicted incinerabilities spanning the range
from the most-difficult-to-incinerate class to the least-difficult-to-incinerate class was tested. This
"POHC soup" mixture was combined with a clay-based sorbent solid matrix and packaged into
fiberpack drums for incineration testing in the rotary kiln incineration system (RKS) at the IRF.
-------
The drums containing the soup/clay mixture were batch fed to the RKS via a fiberpack drum
ram feeder.
A series of five incineration tests was performed during which incinerator operating
conditions and test mixture composition were varied. Specific test program variables were:
• Kiln temperature
• Feed batch charge mass
• Feed composition, specifically H/C1 ratio
One test was performed under typical operating conditions with a baseline mixture composition.
The other tests varied the above in an attempt to simulate various modes of incineration failure:
thermal failure, mixing failure, feed matrix effects, and a worst-case combination of these.
2.1 TEST PROGRAM
The IRF's RKS was used for this test program. A process schematic of the RKS is
shown in Figure 1. The IRF RKS consists of a primary combustion chamber, a transition
section, and a fired afterburner chamber. After exiting the afterburner, flue gas flows through
a quench section followed by a primary air pollution control system (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.
2.1.1 Synthetic Waste Mixture
Twelve POHCs were selected for inclusion in the synthetic waste mixture employed in
the test program. The incinerabiHty ranking groups 333 POHCs into 7 stability classes from most
stable (class 1) to least stable (class 7). UDRI recommended that two compounds from each
class be included in the test mixture and provided a list of candidates for selection. The selection
of compounds from this candidate list was guided by sampling.and analysis, compound
compatibility, compound availability, and safety considerations.
The compounds selected for the test mixture are listed in Table 1. The table also notes
the composition of two test mixtures containing the POHCs. Test mixture 1 was the baseline
test mixture. The POHC concentrations in test mixture 2 represent adjustments to relative
POHC concentrations to yield a mixture with decreased H/C1 ratio.
The mixtures incinerated in the test program were prepared using commercially-available
pure chemicals and materials. Test material formulation consisted of adding weighed quantities
(1.3 kg, 3 Ib total) of the mixture of the twelve organic constituents to a weighed quantity (2.3 kg,
5 Ib) of an absorbent clay. The clay/organic mixtures were packaged into 1.5-gal fiberpack
drums lined with polypropylene bags, the mouths of which were closed with wire ties.
-------
06-Z90 QS3
I
•*•*
2
I?
1
E
a>
u
cn
-------
TABLE 1. SYNTHETIC WASTE MIXTURE COMPOSITION
Concentration (wt %)
Mixture 1 Mixture 2
Component high H/C1 low H/C1
Benzene
Chlorobenzene
Tetrachloroethene
1,2,2-Trichloro-
1,1,2-trifluoroethane
(Freon 113)
Benzenethiol
Nitrobenzene
Hexachlorocyclohexane
(Lindane)
Hexachloroethane
1,1,1-Trichloroethane
p-Dimethylaminoazobenzene
(methyl yellow)
Nicotine
N-nitroso-di-n-butyl amine
H/C1 (molar)
8
8
8
8
8
8
10
10
10
10
10
2
3.6
4
4
33
4
4
4
5
25
5
5
5
2
1.2
T99(2), "C"
1,150
990
890
780
725
655
645
585
545
-400
<320
<320
Stability
Rankb class
3
22
43
92
122
150/151
159
213
233
268
286 to 289
316 to 331
1
1
2
3
3
4
4
5
5
6
7
7
"Temperature required to achieve 99 percent destruction in 2 s.
blncinerability rank in list range from most refractory (No. 1) to most labile
(No. 333).
2.12 Test Conditions
The variables for the test program were the H/C1 ratio in the synthetic waste feed, kiln
temperature, and synthetic waste feed charge mass. Five tests, specified to be conducted with
various combinations of these parameters, were selected to evaluate the relative incinerability
of the POHCs. The target test matrix is shown in Table 2. Test 1 represented a baseline, or
normal, set of incinerator operating conditions. Test 2 attempted thermal failure by decreasing
the kiln exit temperature to a target of 649° C (1,200°F). To further promote thermal.failure,
0.9 kg (2 lb) of water was added to each waste feed charge for Test 2. In Test 3, mixing failure
was attempted by doubling the drum charge mass from 3.6 to 7.3 kg (8 to 16 lb). This doubled
charge mass was introduced at half the baseline frequency, so as to maintain the overall waste
feedrate equal to that for the other test conditions. Test 4 was designed to investigate the effects
of reducing the H/C1 ratio (matrix failure) in the waste feed. Test 5 combined the three failure-
promoting conditions to produce a "worst-case" condition by operating with the kiln exit
temperature at a target of 649° C (1,200° F), introducing the waste at double the baseline charge
mass, and by using a low H/C1 ratio waste mixture.
-------
TABLE 2. TARGET TEST CONDITIONS
Test
1
2
3
4
5
Klin H/CI,
molar
22.8
22.8
22.8
15.7
10.3
Kiln exit
Kiln
Afterburner
exit
temperature, exit O2, temperature,
°C(°F) % °C(°F)
871 (1,600)
649 (1,200)
871 (1,600)
871 (1,600)
649 (1,200)
10.4
12.6
10.4
10.4
13.2
982 (1,800)
982 (1,800)
982 (1,800)
982 (1,800)
982 (1,800)
Afterburner
exit O2,
%
9.0
9.1
9.0
9.2
9.2
Organic/clay
per fiberpack,
kg(lb)
3.6 (8)
4.5 (10)"
3.6 (8)
3.6 (8)
3.6 (8)
Feed
regimen
drums/
charge
1
1
2
1
2
Charge/
hr
12
12
6
12
6
"0.9 kg (2 Ib) water added per fiberpack.
For all tests, the average kiln exit temperatures were within 14° C (26° F) of the
respective target temperatures. However, actual O2 levels in the kiln exit flue gas were generally
higher than the target concentrations. The higher O2 levels were generally the result of higher
than expected air in-leakage into the kiln chamber.
2.13 Sampling and Analysis Procedures
The scope of the sampling effort undertaken during this test program is illustrated in
Figure 2, in which the sampling locations and the corresponding sample collection methods are
identified. Specifically, the sampling effort during each test consisted of:
• Obtaining a sample of the POHC/clay feed mixture by compositing the contents
of three waste fiberpack drums randomly selected during the test
• Obtaining a sample of the scrubber blowdown liquor composited from grab
samples taken at hourly intervals over the test period
• Obtaining a composite sample of the kiln ash from the ash collection bin at the
end of the test
• Continuously measuring O2, CO, CO2, and unheated total unburned hydrocarbon
(TUHC) concentrations in the flue gas at the kiln exit; O2 concentrations at the
afterburner exit; O2, NOX, unheated TUHC, and heated TUHC concentrations at
the scrubber exit; and O2, CO, CO2, and heated TUHC concentrations in the stack
• Sampling flue gas at the kiln exit, scrubber exit, and stack for the semivolatile and
volatile POHCs using Method 0010 and Method 0030, respectively
-------
68-frW QS3
J
OH
S
<:
CO
io a e
11 =
15 IS T3
•
a «
*? "5
"a o
® .-
•5 S
"sy
*"* S
K ^
Unheated
TUHC
g
do
u u
•S -a
eo
.S «
fl
5 ex
X
X
XX
X
X
X
XX
XX
X
X
X X
XX XX
X
X
X
H
e
1
-------
The laboratory analysis procedures used to characterize the samples collected over the
test program included:
• Analyzing the composite feed, kiln ash, and scrubber blowdown samples from each
test for volatile and semivolatile organic constituents
• Analyzing Method 0010 train samples from each test for semivolatile organic
hazardous constituents
• Analyzing Method 0030 train samples from each test for volatile organic
hazardous constituents
Semivolatile organic analyses were performed by Method 8270. Solid samples, including
waste feed and kiln ash, and Method 0010 samples were Soxhlet-extracted by Method 3540 in
preparation for analysis. Liquid samples were liquid-liquid extracted by Method 3510.
Method 0030 (VOST) sample analysis was by thermal desorption purge and trap GC/MS
(Method 5040) analysis with ah ion trap detector.
22 TEST RESULTS
The POHC measurements at the kiln exit are the most relevant with respect to
evaluating the incinerability ranking in that the incineration failure conditions tested involved
varying kiln operation. Thus, incineration failures achieved would be most evident and best
measured at the kiln exit.
Figures 3 through 7 show kiln exit POHC DREs measured in bar chart form. The
POHCs are ordered along the horizontal axis by their thermal stability index ranking from
predicted most stable (benzene) to least stable (N-nitroso-di-n-butyl amine). The vertical axis
is the quantity [-log (l-DRE/100)] for each POHC, which represents the "number of 9's" of
POHC destruction. A value of 1 signifies 90 percent DRE, a value of 2 signifies 99 percent
DRE, and so on. Each bar represents the measured DRE for the corresponding POHC. Where
flue gas analysis indicated that the particular POHC was below its detection limit, a stacked bar
format is used to convey this information. The height of the bottom bar of the stack represents
the DRE calculated using the practical quantitation limit (PQL). The combined stacked bar is
extended to the top of the chart (six "9's" DRE) as a visual reminder that the POHC was not
detected and that the measured DRE was greater than that computed using the PQL.
The bar graphs shown in Figures 3 through 7 show DREs based on the feed formulation
data. As noted in Section 2.1.3, feed samples were collected and analyzed for the POHCs. Thus,
DREs could be calculated based on feed POHC analysis results. However, analyzed POHC
concentrations in feed samples were substantially lower than the concentrations corresponding
to the POHC quantity used to form the POHC mixtures. On average, only between 12 and
29 percent of the volatile organic constituents and between 27 and 82 percent of the semivolatile
organic constituents in the prepared mixtures could be accounted for in the feed analyses. One
POHC, nicotine, was not detected in the Test 3 feed. Fiberpack weights measured during the
tests rule out mass evaporative loss as the explanation for differences between prepared and
analyzed concentrations. Thus, it is believed that what was prepared indeed was fed. One
possible explanation is that the organic liquid constituents were so tightly adsorbed to the porous
10
-------
5 -
uJ
cc
9
3 -
2 -
1 -
4455
STABILITY CLASS
DENOTES POHC NOT DETECTED
Figure 3. Kiln exit POHC DREs for Test 1.
o
o
s
tr
5 -
4 -
3 -
1 234455
STABILITY CLASS
DENOTES POHC NOT DETECTED
Figure 4. Kiln exit POHC DREs for Test 2.
11
-------
ft
DC
-------
1 234455
STABILITY CLASS
DENOTES POHC NOT DETECTED
Figure 7. Kiln exit POHC DREs for Test 5.
clay that the sample preparation procedures associated with Methods 8240 and 8270 analyses
could not quantitatively free the organic constituents for detection in the analyses. The validity
of this hypothesis was not examined, however. Nevertheless, test conclusions were supported by
DRE calculations using both the prepared formulation and test feed sample analysis
concentrations. For clarity, only the feed formulation evaluations are shown in Figures 3
through 7, and discussed in the following.
In addition, Table 1 noted that benzenethiol was selected as one of the class 3 POHCs
in the POHC mixture. However, while performing tests to verify that stable synthetic waste
organic feed mixtures could be prepared, it was discovered that benzenethiol quite rapidly and
completely reacts, in the presence of the other organics and the clay matrix, to form diphenyl
disulfide, a class 6 compound. As a result, diphenyl disulfide, not benzenethiol, was actually fed
to the incinerator. Thus, the DRE for diphenyl disulfide is shown in the Figure 3
through Figure 7 bar charts, and its bar location corresponds to its class 6 incinerability order.
The following discusses relative POHC DREs measured for each test in turn.
13
-------
2.2.1 Test 1—Baseline Incineration Conditions
The incinerator operating conditions for Test 1 represented baseline incineration
operation, which, from past experience would result in acceptable POHC destruction. As shown
in Figures, kiln exit DREs were 99.99 percent or greater for all POHCs. Benzene,
chlorobenzene, tetrachloroethene, Freon 113 and 1,1,1-trichloroethane were quantitatively
measured at the kiln exit and their corresponding DREs are shown by the single bars. The
remaining POHCs were not detected at the kiln exit and their respective DREs are represented
by the two-segment stacked bars, the significance of which was discussed above.
The high POHC DREs confirmed that this baseline incinerator operating condition was
indeed capable of satisfactorily destroying even the predicted most difficult to incinerate POHC,
benzene. A weak correlation might exist between DRE and the POHC incinerability ranking'
in that, except for 1,1,1-trichloroethane which had a DRE of about 99.99 percent, DREs for
POHCs ranked in class 4 and above were higher than the class 3 and below POHCs. It bears
emphasis, however, that the measured POHC DREs only varied by a small degree (from
99.994 percent to 99.9997+ percent). This, coupled with the lack of gross incineration failure
to broaden the incinerability response, could explain the inability to establish a clear correlation
between DRE and the incinerability ranking index from the baseline test data.
It is interesting to note that 1,1,1-trichloroethane, a POHC ranked in class 5 and
believed to be relatively easy to incinerate, had a measured DRE substantially lower than
similarly ranked POHCs. One possible explanation is that 1,1,1-trichloroethane is a common
product of incomplete combustion (PIC), and can be formed during the incineration process,
most directly from hexachloroethane, another component of the POHC mixture.
222 Test 2—Thermal Failure (Quenching)
Test 2 was intended to simulate a thermal failure condition through incineration quench.
This was accomplished by lowering the kiln temperature from nominally 871° C (1,600° F) to
649° C (1,200° F) via two means. A measured amount of water contained in a polyethylene bag
was added to each waste feed fiberpack drum and the kiln was fired at very high air/fuel ratio.
These two actions in combination would be expected to create conditions conducive to the
formation of cold POHC-containing pockets of gas which would escape the kiln prior to being
destroyed.
Figure 4 presents the kiln exit POHC DREs for this test. The data clearly indicate that
this test condition resulted in significantly different POHC DREs compared to the baseline test.
A wide range of POHC DREs was observed, from less than 99 percent for Freon 113 to greater
than about 99.999 percent for diphenyl disulfide, methyl yellow, nicotine, and N-nitroso-di-n-butyl
amine. The low DREs for several POHCs confirmed that incineration failure did occur during
this test.
With the exception of a few anomalies (discussed below), a general correlation between
DRE and incinerability ranking seems apparent for this test. The observed DREs for the class 3
to 7 POHCs appeared to follow the incinerability ranking predicted behavior. Some POHC-to-
POHC variability existed within this sub-group of POHCs. Lindane exhibited a higher DRE than
the neighboring ranked POHCs.
14
-------
A considerably more significant deviation from incinerability ranking expectation
involved the DREs of the four most-stable ranked POHCs within classes 1,2 and 3. The relative
DREs measured for these four POHCs were in an order opposite to the ranking predictions.
The relative extent of incineration failure for these four POHCs was not in accordance with
expectations from the thermal stability ranking.
223 Test 3—Mixing Failure
One of the modes by which POHCs can escape an incinerator undestroyed results from
the lack of adequate mixing between POHC and oxidizer. For Test 3, the weight of each waste
charge to the kiln was doubled, while the hourly waste feedrate was maintained at a level
consistent with the other four tests. The doubled waste charge was thought to increase the
likelihood of creating oxygen-deficient pockets of POHCs in the kiln chamber. The expectation
was that if the oxygen-deficient conditions persisted through the kiln, undestroyed POHCs could
escape the kiln chamber.
However, as the data in Figure 5 show, no clear failure was apparent during this test.
All POHC DREs were greater than 99.99 percent and exhibited trends similar to those observed
for the baseline condition test (Test 1). DREs for nitrobenzene and the group of POHCs ranked
easier to incinerate were high. Within this group, only 1,1,1-trichloroethane was detected in the
kiln exit flue gas sample at a level above its PQL. The other less-stable ranked POHCs were not
detected in the kiln exit flue gas.
The four most difficult to incinerate POHCs, benzene, chlorobenzene, tetrachloroethene
and Freon 113 were present in the kiln exit flue gas sample at levels corresponding to between
99.99 and 99.999 percent DRE.
No correlation between POHC DREs and POHC incinerability ranking was apparent
for this test. DREs for the volatile POHCs (detected in the flue gas) were comparable to the
DREs associated with the PQLs for the semivolatile POHCs (not detected in the flue gas).
2.2.4 Test 4—Matrix Failure
This test (Test 4) attempted to cause incineration failure by decreasing the H/C1 ratio
in the organic feed to the kiln. The H/C1 ratio in the feed waste for Test 4 was 1.2, as compared
to a H/C1 ratio of 3.6 for the baseline Test 1.
The kiln exit POHC DREs for this test are shown in Figure 6. These were uniformly
high: all exceeded 99.99 percent. As in Test 3, no correlation between POHC DRE and
incinerability ranking was apparent because the POHC DREs were uniformly, high.
One possible explanation for the inability to achieve POHC DRE failure in this test is
that the actual H/C1 ratio in the kiln environment as a whole was quite different from that in
the waste feed itself. This was so because the auxiliary fuel for the burner, in this case natural
gas, was a significant additional source of hydrogen. If this source of hydrogen is included, the
H/C1 ratio in the total kiln environment for this test would be 15.7, which may be considerably
higher than the H/C1 ratio required to cause DRE failure.
15
-------
2.2.5 Test 5—Worst-Case Combination
This last test was conducted to present the most challenging combination of the
mechanisms tested in terms of POHC destruction failure. The kiln was operated at the reduced
temperature of 640° C (1,184°F) to induce thermal failure; the waste feed charge size was
doubled to promote mixing failure; and the chlorine content in the feed waste was elevated to
promote matrix failure. It should be noted that, in a departure from Test 2 procedures, no water
was added to the feed waste fiber drums, because doing so would introduce a quantity of
hydrogen that might nullify any potential elevated chlorine (matrix failure) effect.
Figure 7 shows the POHC DREs for this test condition. In this test, eight of the twelve
POHCs were detected in the kiln exit flue gas. This was in contrast to only five POHCs being
found at concentrations above their PQLs during baseline Test 1. The four predicted most easily
incinerated POHCs, namely, diphenyl disulfide, methyl yellow, nicotine, and N-nitroso-di-n-butyl
amine were not found above their PQLs. Assuming that these POHCs were present at their
respective PQLs would lead to computed POHC DREs greater than 99.998 percent for these
four POHCs. These DREs, being higher than those measured for the remaining POHCs, were
consistent with their incinerability ranking indices.
The DREs for the eight quantifiable POHCs ranged from over 99 to almost 99.999
percent. Lindane had the highest DRE at 99.9989 percent. Freon 113 and 1,1,1-trichloroethane
exhibited the lowest DRE at about 99.8 percent. While no monotonic correlation between
POHC DREs and incinerability ranking order existed for this test, a weak relationship may have
existed for the class 3 through class 7 POHCs. As observed in Test 1, 1,1,1-trichloroethane
exhibited a DRE significantly below that of its neighboring ranked POHCs.
It is interesting to note how well the relative POHC DREs of this "worst-case" test
compared to those observed for Test 2. Recall that Test 2 simulated only quench failure. The
relative DREs for these two tests exhibited quite similar patterns although two differences could
be noted. One difference is the absolute DRE levels, which for this test were generally higher
than those observed for Test 2. The other difference relates to the DREs for benzene and 1,1,1-
trichloroethane. The DREs for benzene and 1,1,1-trichloroethane were nearly two "9's" higher
for Test 2 than for Test 5.
23 CONCLUSIONS
Conclusions from the tests include the following:
• The baseline operation condition resulted in effective POHC destruction. Kiln
exit POHC DREs were in the 99.99 percent range for the volatile POHCs in the
test mixture. Semivolatile POHCs were not detected in the kiln exit flue gas;
corresponding lower bound DREs were generally greater than 99.999 percent.
« Neither the mixing failure nor matrix failure attempts resulted in incineration
failure. Kiln exit POHC DREs were comparable to those measured in the
baseline test for all POHCs.
16
-------
• The thermal failure and worst-case tests resulted in kiln POHC destruction failure.
For both tests, kiln exit POHC DRE ranged from 99 percent or less for Freon 113
to greater than 99.999 percent for the highest ranked (least stable) semivolatile
POHCs.
?•" - ; - .
• For the incineration failure tests, there was general agreement between observed
relative kiln exit POHC DRE and thermal stability incinerability ranking
expectations. However, two deviations occurred for both tests.
•— The class 1 compounds (benzene and chlorobenzene) and the class 2
compound (tetrachloroethene) were less stable (had greater kiln exit DRE)
than the class 3 compound Freon 113.
— 1,1,1-trichloroethane was apparently more stable in the baseline and worst-
case tests than its class 5 ranking would suggest, when compared to the other
class 5 and the class 4 compounds; production of 1,1,1-trichloroethane as a
PIC could account for this observation.
Test results were documented in the test report:
• Lee, J. W., W. E. Whitworth, and L. R. Waterland, "Pilot-Scale Evaluation of the
* Thermal Stability POHC Incinerability Ranking," draft June 1991.
Test results were also presented in a poster presentation:
• Carroll, G. J., "Pilot-Scale Evaluation of an Incinerability Ranking System for
Hazardous Organic Compounds," presented at the 17th Annual Hazardous Waste
Research Symposium, Cincinnati, Ohio, April 1991
and in a technical paper:
• Lee, J. W., L. R. Waterland, and G. J. Carroll, "Evaluation of the Thermal
Stability POHC Incinerability Ranking in a Pilot-Scale Rotary Kiln Incinerator,"
paper 91-343, presented at the 84th Annual Meeting of the Air & Waste
Management Association, Vancouver, British Columbia, June 1991.
17
-------
SECTION 3
INCINERATION TESTS OF SPENT POTLINERS
FROM THE PRIMARY REDUCTION OF ALUMINUM (K088)
OSW is under Congressional mandate to ban the landfill disposal of specific, listed
hazardous wastes that can be treated to reduce their volume and/or hazardous characteristics.
Because incineration is an appropriate treatment technology for many such wastes, OSW
(J. Labiosa, L. Rosengrant, coordinators) has requested that the incinerability of these wastes
be evaluated, and the characteristics of the incineration residues be determined. The
incineration residues of interest are the kiln bottom ash and the scrubber system blowdown
discharge.
This test program at the IRF evaluated the incinerability and characterized the
incineration residuals of spent potUners from the primary reduction of aluminum, listed waste
K088. Specific objectives of the test program were to:
• Determine whether incineration of K088 can be accomplished while complying
with the hazardous waste incinerator regulations
• Characterize the waste feed and incineration residuals in terms of their
concentrations of the hazardous constituents found on the best demonstrated
available technology (BDAT) list
The incineration tests were completed on January 15, 16, and 17, 1991. Results of the test
program are discussed in this section.
3.1
TEST PROGRAM
In the electrolytic reduction process for the production of aluminum, alumina (A12O3)
contained in bauxite ore is dissolved in molten cryolite (3NaF • A1F3). An electric current is
passed through this molten bath, reducing the alumina to aluminum and evolving oxygen (O2).
The electrodes for passing the current are carbon. Thus, the electrolytic cells are large carbon-
lined steel boxes, the carbon lining being the cathode, with carbon anodes suspended in the bath.
The O2 released in the reduction combines with the carbon at the electrodes to form CO2. Thus,
the ^electrodes (the cathode potliner and the anodes) are consumed during the process and are
periodically replaced (the anodes much more frequently than the cathode potliner). The spent
cathode potliner comprises the subject waste for these tests.
18
-------
This potliner waste consists of the remaining carbon cathode being replaced, along with
residual ore and cryolite material left in the cell when the spent potliner is removed. Thus, its
major constituents are carbon, aluminum, alumina, silicon and iron oxides (from the bauxite ore),
and sodium and aluminum fluoride (from the cryolite). In addition, other ore and steel cell
elements, including trace quantities of several hazardous constituent trace metals, are present.
The potential presence of high cyanide levels is the basis for listing this waste as toxic. Thus, for
these tests, cyanide was considered the POHC requiring 99.99 percent DRE in the incineration
treatment of the waste.
3.1.1 Test Conditions
The tests for this test program were performed in the RKS at the IRF. Like the POHC
incinerability ranking tests, the waste was packaged into 1.5-gal fiberpack drums and fed to the
kiln via the fiberpack drum ram feed system. For these tests, each fiberpack contained nominally
5.7 kg (12.5 Ib) of waste. The single-stage ionizing wet scrubber system shown in Figure 1 was
used in place of the venturi/packed-column scrubber system for the K088 tests. Three tests were
conducted, each under the same set of incinerator operating conditions.
The planned incinerator operating conditions for each test are summarized in Table 3.
These conditions were chosen to maximize the opportunity for waste carbon burnout. However,
past experience with K088 has shown that the waste can agglomerate and form slag at
temperatures in the range of the target kiln temperature for these tests shown in Table 3. Slag
formation was to be avoided as an operational consideration for these tests.
Table 4 summarizes the actual incinerator exit temperatures and flue gas levels,
including their ranges and averages for each test during flue gas sampling. These are compared
with the respective target conditions. The average kiln exit temperature for Test 1 was observed
at 17° C (30° F) higher than target and allowed because there was no evidence of slagging. The
average kiln exit temperature for Test 2 was within 2°C (3°F) of target temperature. The kiln
exit temperatures for Test 3 were held to an average of 11° C (19° F) below target because there
was evidence of slagging at target conditions. The average afterburner exit temperatures were
within 3°C (5°F) of the target temperatures for all tests. The average kiln O2 levels were within
1.6 percent of the target O2 levels for all tests. The afterburner exit O2 levels were somewhat
higher than targeted levels, although average afterburner exit O2 levels were within 2.6 percent
of target O2 levels.
Table 5 summarizes the total amount of waste fed to the RKS during each test, and the
cumulative amount of kiln ash discharged over each test's duration. As shown in the upper
portion of Table 5, the amount of ash discharged from the kiln varied from 60 percent of the
amount of waste fed during Test 1 to 48 percent for Test 3. During Test 1, there was no
evidence of waste slagging. During Test 2, there was some visual evidence of waste slagging late
in the test. The waste fed for Test 2 was of slightly smaller average size than for Test 1. In
Test 3, which fed a waste of even smaller average size, there was evidence of slagging essentially
throughout the test. The relative amounts of ash collected as a fraction of the waste weight fed
in each test agrees with the relative degree of slagging observed. The largest amount of ash, as
a fraction of waste weight fed, was discharged in Test 1 with no observed slag formation. The
least amount of ash was discharged in Test 3, with observed slagging throughout the test.
19
-------
TABLE 3. TARGET TEST CONDITIONS
Total waste feedrate
Kiln temperature
Kiln exit flue gas O2
Afterburner temperature
Afterburner exit flue gas O2
Scrubber blowdown flowrate 0 L/min (total recycle)
Scrubber liquor flowrate 230 L/min (60 gpm)
Scrubber pressure drop 1.5 kPa (6 in WC)
Scrubber liquor pH 7.5 to 8.0
68 kg/hr (150 Ib/hr)
980°C (1,800°F)
10 percent
1,090° C (2,000° F)
7 percent
TABLE 4. ACTUAL VERSUS TARGET OPERATING CONDITIONS FOR THE K088 TESTS
Temperature, °C (°F)
Test
Date
Target
Minimum
Maximum
Average
Flue
gas O2, %
Target Range
Average
Kiln exit
1
2
3
1/15/91
1/16/91
1/17/91
982 (1,800)
982 (1,800)
982 (1,800)
962 (1,764)
954 (1,749)
922 (1,692)
1,062
1,016
1,012
(1,944)
(1,861)
(1,853)
999 (1,830)
984 (1,803)
972 (1,781)
10
10
10
9.6
9.9
9.7
to
to
to
12.3
12.8
12.9
11.0
11.6
11.4
Afterburner exit
1
2
3
1/15/91
1/16/91
1/17/91
1,093 (2,000)
1,093 (2,000)
1,093 (2,000)
1,088 (1,990)
1,091 (1,995)
1,088 (1,990)
1,102
1,103
1,102
(2,016)
(2,017)
(2,015)
1,096 (2,005)
1,096 (2,005)
1,096 (2,005)
7
7
7
8.2
7.9
8.4
to
to
to
10.0
9.9
10.4
9.0
9.2
9.6
20
-------
TABLE 5. WASTE FEED AND ASH COLLECTED
Test
1 (1/15/91)
2 (1/16/91)
3 (1/17/91)
Total
Slag recovered15
Total ashc
Total waste
fed, kg
(Ib)
282 (621)
233 (513)
267 (587)
782 (1,721)
Ash collected8
Weight, kg Fraction of feed,
(Ib) %
168 (370)
125 (275)
129 (284)
422 (929)
121 (266)
543 (1,195)
60
54
48
54
69
aAsh collected in collection drum during test.
bSlag removed from kiln walls after test program completion.
cSum of total ash collected during test program and slag
removed from kiln walls after test program completion.
The consequence of the slag formation experienced is greater ash holdup in the kiln,
with formed slag adhering to the kiln wall. After the test program was completed, a buildup of
slag on the kiln walls was clearly evident. The slag layer covered the entire length of the kiln
with an average thickness of 1.3 cm (0.5 in). The slag buildup was manually removed (chipped
off) after test program completion. A total of 121 kg (266 Ib) of slag was removed, as indicated
in Table 5. When this quantity is added to the ash collected in the ash collection drums during
the test, the total ash plus slag collected accounted for an increased 69 percent of the total waste
fed over the test program. The relationship between this total recovered kiln ash and the ash
content in the waste is discussed in Section 3.2.1.
3.1.2 Sampling and Analysis Procedures
The scope of the sampling effort undertaken during this test program is illustrated in
Figure 8, in which the sampling locations and the corresponding sample collection methods are
identified. Specifically, the sampling during each test consisted of:
• Obtaining grab/composite samples of the waste feed from each drum as the waste
was packaged into fiberpack containers
• Obtaining samples of the kiln ash generated from each test
• Obtaining samples of the recirculating scrubber liquor for each test
• Continuously measuring O2 levels in flue gas at the kiln exit and at the afterburner
exit; O2, CO, CO2, NOX and TUHC levels at the scrubber exit;- and O2, CO, and
CO2 levels in the stack gas
21
-------
16-Z03 QS3
2g
^§
T
CARBON
BED
I
DEMISTER
i \
IONIZING WET
SCRUBBER
I
QUENCH
SECTION
f
1
AFTER-
BURNER
f
1
•z
2
i
—
—
T
(O
CO
1/1 |s
ill
2 « 8
3 2
f § i
^ EL
0)
o *S *S
^T W M
M ao
J JfJ
1=1
Ionizing wet
scrubber
liquor
il
^ »rt
|i
OA
.S *.
— ^
es
X
X
X X
XX XX
x
X
X
»-H CS CO •* >O vo t^
.
£
"S.
00
SJ
1
22
-------
• Sampling flue gas downstream of the scrubber for cyanide using a Method 5
variation of a NIOSH Method 7904 sampling train, for fluoride using an EPA
Method 13B train, and for paniculate and trace metals using an EPA Method 5
train
• Sampling the stack gas for cyanide using a Method 5 variation of a NIOSH
Method 7904 sampling train and for particulate and HC1 using an EPA Method
5 train
For each test, 48 fiberpack drums of waste wers packaged the day before the test.
During packaging, a nominal lOOg aliquot of the contents of each of the first 36 fiberpacks was
taken. Each of these aliquots was ground to a nominal size of 1 mm or less. These ground
aliquots were then combined in groups of 12, so that three composite waste feed samples per test
day (nine total) resulted, each corresponding to a set of 12 fiberpacks of feed packaged for a test
day. Throughout each test, kiln ash discharged to the ash pit was continuously removed by an
auger system and conveyed to a series of initially clean 55-gal drums. Two ash samples
corresponding to two of the three feed samples were collected each day.
Throughout the three-test program, the scrubber system was operated at total recycle
(zero blowdown). This mode allowed any combustion gas contaminants to build up to maximum
(worst-case) concentrations in the scrubber liquor over the three-test program. Two samples of
the recirculating scrubber liquor were collected for cyanide analyses each test day. These two
samples were collected to correspond to two of the three feed sample sets incinerated each test
day. After all tests were completed, the scrubber system was drained to a storage tank. A final,
full test program composite scrubber liquor sample was collected while the system was being
drained. This composite sample was analyzed for the other test program constituents of interest.
The sample collection procedures resulted in two waste samples, with corresponding kiln
ash and scrubber liquor samples, each test day, to provide six waste/kiln ash/scrubber liquor
combinations over the three tests. A third waste feed sample (the initial set of waste fiberpack
drums fed) was also analyzed to provide a further measure of the degree of waste composition
variability. One set of flue gas characterization samples was collected for each test. One total
test scrubber liquor sample was collected as well.
The test program composite scrubber liquor sample was filtered using the TCLP
filtration procedure for solids-containing liquid waste. An aliquot of the kiln ash was subjected
to the TCLP leaching procedure. Waste feed samples, kiln ash samples, kiln ash TCLP leachate
samples, and both the filterable solids and filtrate fractions of the scrubber liquor were analyzed
separately for:
• The BDAT semivolatile organic constituents
• The BDAT trace metals listed in Table 6
• The nonhazardous constituent metals listed in Table 6
• Fluoride
• Cyanides (total and amenable)
23
-------
TABLE 6. TRACE METALS DETERMINED
BOAT List
trace metals
Nonhazardous
constituent metals
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Vanadium
Zinc
Aluminum
Calcium
Cobalt .
Iron
Lithium
Magnesium
Manganese
Molybdenum
Phosphorus
Potassium
Sodium
Strontium
Tin
Exceptions to this were that kiln ash TCLP leachates were not analyzed for the BDAT
semivolatile organic constituents, and the scrubber liquor solids were not analyzed for fluoride.
Waste feed, kiln ash, and scrubber liquor filtrate samples were also analyzed for sulfide; and
waste feed and kiln ash samples were subjected to proximate (ash, moisture, fixed carbon, and
volatile carbon), ultimate (C, H, O, N, S, Cl), and silica (SiO2) analyses.
32 TEST RESULTS
Test results are presented in this section. In many of the data tables that follow, sample
analysis results are grouped by test day (Test 1, 2, and 3), with separate results grouped for each
of the two waste feed/kiln ash pairs within each test day (with scrubber liquor added for cyanide
analysis). These two pairs are denoted the "a" pair and the "b" pair in the tables. Results for
the third waste feed sample (not paired with a kiln ash residue sample), denoted as waste
sample "c," are also given.
3.2.1 Proximate, Ultimate, and Silica Analysis Results
Table 7 summarizes the proximate and silica results for the waste feed and kiln ash
samples. These data can be used to calculate the average proximate composition for the waste
feed and kiln ash for each test day. Further, weighted averages can be calculated by weighting
each test day's average composition by the amount of waste feed/kiln ash collected that day,
shown in Table 5.
Using these total test program weighted average concentrations, proximate component
distributions for the incineration tests performed can be calculated. These distributions are
summarized in Table 8. The data in Table 8 show that roughly half the carbon content (both
24
-------
TABLE 7. PROXIMATE AND SILICA ANALYSIS RESULTS
Concentration, wt %
Sample
Fixed Volatile Total fixed
carbon matter plus volatile Moisture Ash Silica
Test 1 (1/15/91):
Waste la
Kiln ash la
Waste Ib
Kiln ash Ib
Waste Ic
33.0
21.5
34.7
25.4
31.3
2.4
2.1
2.5
1.3
3.1
35.4
23.6
37.2
26.7
34.4
0.1
0.1
0.02
0.1
0.2
64.6
76.3
62.7
73.2
65.4
3.4
7.3
3.3
4.0
3.7
Test 2 (1/16/91):
Waste 2a 34.2 3.5
Kiln ash 2a 27.7 4.0
Waste 2b 34.8 4.8
Kiln ash 2b 26.1 3.7
Waste 2c 33.8 3.6
Test 3 (1/17/91):
37.7
31.7
39.6
29.8
37.4
0.2
0.1
0.2
0.1
0.2
62.1
68.2
60.2
70.1
62.4
5.6
5.5
5.3
5.4
5.7
Waste 3a
Kiln ash 3a
Waste 3b
Kiln ash 3b
Waste 3c
37.4
28.9
36.4
20.3
38.8
5.4
2.3
5.4
3.6
4.0
42.8
31.2
41.8
23.9
42.8
0.1
0,1 .
0.2
0.1
0.2
. 57.1
. 68.7
58;0
76.0
57.6
5.9
5.6
5.5
5.7
6.3
TABLE 8. TEST PROGRAM COMPOSITE PROXIMATE COMPONENT DISTRIBUTIONS
Parameter
Waste feed:
Composite concentration1", %
Amount fed, kg (Ib)
Kiln ash:
Composite concentration13, %
Amount discharged0, kg (Ib)
Fraction of amount fed, %
Total
99.9
781 (1,719)
99.9
543 (1,194)
69
Fixed
carbon
34.9
273 (601)
24.8
135 (297)
49
Volatile
matter
3.8
30 (66)
2.7
15 (32)
48
Ash
61.2
478 (1,052)
72.4
393 (865)
82
Silica"
5.0
39 (86)
5.6
30 (67)
78
aSilica is a component of the ash fraction.
bMoisture excluded.
Includes 121 kg (266 Ib) of slag removed after the test program; assumes slag has same
composition as composite ash collected during the test program.
25
-------
fixed and volatile fractions) of the waste was destroyed (oxidized) during the incineration of the
waste. Roughly 80 percent of the ash content of the waste was discharged as kiln ash; the
remaining 20 percent of the ash was either volatilized out of the kiln or entrained as particulate
in kiln exit flue gas. Roughly 80 percent of the waste silica (a component of the ash fraction of
the waste) remained with this kiln ash as well. This distribution (80 percent kiln ash, 20 percent
entrainment/volatilization escaping the kiln) is in line with past IRF testing experience.
322 Cyanide and Semivolatile Organic Analysis Results
Table 9 summarizes the cyanide concentrations in all test program samples analyzed.
The waste incinerated during the first test day (Test 1) contained an average of 5,200 mg/kg
total cyanide, roughly 97 percent of which was amenable to chlorination. The waste fed during
the second and third test days (Tests 2 and 3) contained less cyanide, an average of about
3,500 mg/kg. In addition, less (about 85 percent) of the cyanide in the Test 2 and 3 waste feeds
was amenable to chlorination.
Kiln ash cyanide content varied from test to test, and from sample to sample within a
test day. Measured levels were in the 60 to 330 mg/kg range. Interestingly, the kiln ash cyanide
levels were generally higher for the lower-cyanide-content Tests 2 and 3 wastes than for the
higher-cyanide-content Test 1 waste. The fraction of kiln ash amenable to chlorination also
varied from test to test, and from sample to sample within a test day. Amenable cyanide
fractions ranged from 22 to 73 percent.
Kiln ash TCLP leachates contained measurable cyanide levels ranging from 0.5 to
0.6 mg/L for the Test 1 ash, to 3 to 4 mg/L for the Tests 2 and 3 ash. Amenable cyanide
fractions ranged from <0.2 percent (no amenable cyanide in the TCLP leachate) to > 100 per-
cent (more amenable cyanide measured in the leachate than total cyanide). No cyanide was
detected in the scrubber liquor filtrate or solids from any test. Cyanide was also not detected
in any scrubber exit flue gas or stack gas sampling train sample.
In an approach similar to that followed in Table 8 for the distribution of the proximate
components from the incineration of K088, Table 10 summarizes the composite test program
cyanide distributions. As shown in Table 10, the composite test program kiln ash is estimated
to have accounted for 2.7 percent of the waste total cyanide fed, and 1.6 percent of the amenable
cyanide fed. Alternatively, 97.3 percent of the waste total cyanide and 98.4 percent of the waste
amenable cyanide was removed by the incineration process and destroyed. (Table 9 shows that
cyanide was not detected in either scrubber exit flue gas or stack gas, nor was cyanide detected
in any test scrubber liquor filtrate or solids.)
To assess the degree of cyanide destruction, the cyanide DREs were calculated using
the regulatory definition:
DUE = 100 x (1 - flue gas emission ratelfeedrate)
Table 11 presents the data required to calculate cyanide DREs corresponding to the detection
limits noted in Table 9 for combustion flue gas. As shown in Table 11, cyanide DREs were
26
-------
TABLE 9. CYANIDE ANALYSIS RESULTS
Sample
Test 1 (1/15/91):
Waste la, mg/kg
Kiln ash la, mg/kg
Kiln ash la TCLP leachate, mg/L
Scrubber liquor filtrate la, mg/L
Scrubber liquor solids la, mg/kg
Waste Ib, mg/kg
Kiln ash Ib, mg/kg
Kiln ash Ib TCLP leachate, mg/L
Scrubber liquor filtrate Ib, mg/L
Scrubber liquor solids Ib, mg/kg
Waste Ic, mg/kg
Scrubber exit flue gas, ftg/dscm
Stack gas, pg/dscm
Test 2 (1/16/91):
Waste 2a, mg/kg
Kiln ash 2a, mg/kg
Kiln ash 2a TCLP leachate, mg/L
Scrubber liquor filtrate 2a, mg/L
Scrubber liquor solids 2a, mg/kg
Waste 2b, mg/kg
Kiln ash 2b, mg/kg
Kiln ash 2b TCLP leachate, mg/L
Scrubber liquor filtrate 2b, mg/L •
Scrubber liquor solids 2b, mg/kg
Waste 2c, mg/kg
Scrubber exit flue gas, /ig/dscm
Stack gas, ng/dscm
Test 3 (1/17/91):
Waste 3a, mg/kg
Kiln ash 3a, mg/kg
Kiln ash 3a TCLP leachate, mg/L
Scrubber liquor filtrate 3a, mg/L
Scrubber liquor solids 3a, mg/kg
Waste 3b, mg/kg
Kiln ash 3b, mg/kg
Kiln ash 3b TCLP leachate, mg/L
Scrubber liquor filtrate 3b, mg/L
Scrubber liquor solids 3b, mg/kg
Waste 3c, mg/kg
Scrubber exit flue gas, /tg/dscm
Stack gas, /ig/dscm
"NA = Not analyzed. If total CN not
Total CN
5,290
60
0.51
< 0.005
<5
4,550
130
0.58
< 0.005
<5 .
5,880
<0.16
<0.14
3,200
90
3.1
< 0.005
<5
3,500
330
3.2
< 0.005
<5
3,500
<0.15
<0.14
3,390
260
4.0
< 0.005
<5
3,500
130
3.3
< 0.005
<5
3,800
<0.16
<0.15
detected, amenable CN ane
Amenable
CN '
5,120
36
0.21
NAa
NA
4,430
40
1;17
NA
NA
5,780
NA
NA
2,700
20
< 0.005
NA
NA
2,970
170
< 0.005
NA
NA
3,000
NA
NA
2,930
190
1.6
NA
NA
2,900
90
< 0.005
NA
NA
3,100
NA
NA
ilysis not perfor
Fraction
amenable, %
96.8
60.0
41.2
97.4
30.8
>100
98.3
84.4
22.2
<0.2
84.9
51.5
<0.2
85.7
86.4
73.1
40.0
82.9
56.3
<0.2
81.6
med.
27
-------
TABLE 10. TEST PROGRAM COMPOSITE CYANIDE DISTRIBUTIONS
Parameter
Total CN Amenable CN
Waste feed:
Composite concentration, mg/kg 4,120 3,720
Amount fed, kg 3.22 2.91
Kiln ash:
Composite concentration, mg/kg 160 86
Amount discharged8, g 87 47
Fraction of amount fed, % 2/7 1.6
"Includes 121 kg (266 Ib) of slag removed after test program;
assumes slag has the same composition as the composite ash
collected during the test program.
TABLE 11. CYANIDE DREs
Parameter
Waste feed:
Waste feedrate, kg/hr
Average CN concentration, mg/kg
CN feedrate, g/hr
Scrubber exit flue gas:
CN concentration, [ig/dscm
Flue gas flowrate, dscm/min
Flue gas CN emission rate, ng/hr
CN DRE, %
Stack gas:
CN concentration, iig/dscm
Flue gas flowrate, dscm/min
Flue gas CN emission rate, ng/hr
CN DRE, %
Test 1
(1/15/91)
72
5,240
377
<0.16
29.5
<280
> 99.99993
<0.14
33.0
<280
> 99.99993
Test 2
(1/16/91)
68
3,400
231
<0.15
34.6
<310
> 99.99987
<0.14
35.5
<300
> 99.99987
Test 3
(1/17/91)
72
3,560
256
<0.16 •
35.2
<340
> 99.99987
<0.15
36:6
<330
> 99.99987
28
-------
greater than 99.99993 percent at both the scrubber exit and at the stack for Test 1 and greater
than 99.99987 percent for Tests 2 and 3.
As noted in Section 3.1.2, all nine waste feed, six kiln ash, and the one test program
composite scrubber liquor and solids filtrate samples, were analyzed for the semivolatile organic
BDAT list compounds. None of these compounds was found in any kiln ash sample at
compound-specific practical quantitation limits (PQLs) of 0.4 to 4 mg/kg, in the scrubber liquor
filtrate sample at compound-specific PQLs of 20 to 200 |ig/L, and in the scrubber liquor solids
sample at compound-specific PQLs of 130 to 1,300 mg/kg. No waste feed sample contained any
semivolatile organic constituent at compound-specific PQLs of 0.8 to 8 mg/kg, with the exception
of waste feed 3c which contained benzo(b)fluoranthene at 4.0 mg/kg and benzo(a)pyrene at
3.6 mg/kg. The PQLs of the scrubber liquor solids were high because the solids content of the
scrubber liquor was very low (0.154 g/L), so very little sample was available for extraction.
323 Trace Metal Analysis Results
Table 12 summarizes the BDAT list trace metal analysis results for all samples analyzed.
Table 13 presents a similar summary for the nonhazardous constituent metals analyzed. In both
tables, scrubber exit flue gas concentrations are shown as ranges in many cases. Two sampling
train samples were analyzed to give the total flue gas concentration: the probe wash and filter
catch sample, and the combined impinger solutions sample. In many cases, one or both samples
contained no detectable metal. In these instances, the flue gas concentration is shown as a
range; the lower bound corresponds to assuming a zero metal concentration in the sample or
samples having the not-detected metal, and the upper bound corresponds to assuming the metal
was present at the detection limit in the sample or samples having the not-detected metal.
The data in Table 13 show that the major elemental constituents of both the waste and
resulting kiln ashes were aluminum, calcium, and sodium. These were present in waste and kiln
ash samples at levels greater than 1 percent. Sodium alone accounted for about 20 to 30 percent
of individual waste and kiln ash samples.
The data in Table 12 show that major BDAT trace metals present were barium, nickel,
and thallium. These were present in waste and kiln ash samples at levels in roughly the 100 to
200 mg/kg range. No antimony, lead, mercury, selenium, or silver was found in any waste
sample, nor, with the exception of lead, in any kiln ash, scrubber exit flue gas, or scrubber liquor
sample. No data are shown for lead in the la and Ib kiln ash samples, the three test flue gas
samples, and the three test composite scrubber liquor filtrate sample. No lead was detected in
any waste feed sample. The measurable levels found in the kiln ash samples, the flue gas
samples, and the composite scrubber liquor sample are attributed to contamination.
Table 12 also notes the TCLP regulatory level established to define toxicity characteristic
hazardous waste per the TCLP for those metals having a regulatory level tabulated in
40 CFR 261, Appendix II. Comparing kiln ash TCLP leachate and scrubber liquor filtrate
concentrations to the regulatory levels shows that neither the kiln ash from any test nor the
program composite scrubber liquor would be a toxicity characteristic hazardous waste due to
their leachable hazardous constituent trace metal content.
29
-------
TABLE 12. BOAT TRACE METAL ANALYSIS RESULTS
Sample
Test 1 (1/15/91):
Waste la, mg/fcg
Kiln ash la, mg/kg
Kiln ash TCLP leachate la, mg/L
Waste Ib, mg/kg
Kiln ash Ib, mg/kg
Kfln ash TCLP leachate Ib, mg/L
Waste Ic, mg/kg
Scrubber exit flue gas, Mg/dscm
Test 2 (1/16/91):
Waste 2a, mg/kg
Kiln ash 2a, mg/kg
Kiln ash TCLP leachate 2a, mg/L
Waste 2b, mg/kg
Kfln ash 2b, mg/kg
Kiln ash TCLP leachate 2b, mg/L
Waste 2c, mg/kg
Scrubber exit flue gas, Mg/dscm
Test 3 (1/17/91):
Waste 3a, mg/kg
Kiln ash 3a, mg/kg
Kiln ash TCLP leachate 3a, mg/L
Waste 3b, mg/kg
Kiln ash 3b, mg/kg
Kiln ash TCLP leachate 3b, mg/L
Waste 3c, mg/kg
Scrubber exit flue gas, Mg/dscm
3 test scrubber liquor:
Solids, mg/kg
Filtrate, mg/L
TCLP regulatory level, mg/L
Sb As
<10 11.0
<10 11.4
<0.2 0.016
<10 9.8
<10 8.6
<0.2 0.034
<10 8.8
<96 7.1-11.6"
<10 7.4
<10 10.0
<0.2 0.033
< 10 9.0
< 10 7.7
<0.2 0.020
<10 8.2
<93 2.5-6.9
< 10 13.4
<10 15.9
<0.2 < 0.010
<10 12.9
<10 9.7
<0.2 0.096
<10 11.2
<78 3.9-7.5
<40 <2.0
<0.5 <0.03
5.0
Ba
199
216
31.8
191
227
31.2
195
12.9-17.4
190
221
29.4
176
209
10.5
180
26.2-30.6
205
238
19.1
221
235
23.9
193
12.2-15.8
9,740
17.8
100
Be
26
28
0.020
24
26
0.026
25
3.7-8.2
20
25
0.015
22
26
0.013
21
1.0-5.5
19
25
0.015
17
29
0.025
17
1.1-4.7
2.1
<0.03
Cd
2.6
<3.9
<0.03
2.3
<3.6
<0.03
2.6
4.5-17.9
2.4
3.3
<0.03
2.2
3.3
<0.03
2.4
3.8-17.0
2.2
3.3
<0.03
2.2
3.7
<0.03
2.3
4.6-15.2
<7
<0.1
1.0
Cr
19
27
0.16
23
<20
<0.06
23
13.6-35.9
26
15
0.15
23
13
<0.05
22
13.6-35.7
27
15
<0.05
23
16
<0.05
26
12.5-30.2
<10
<0.2
5.0
Cu
25
33
0.05
22
23
0.14
22
10.3-19.2
24
22
0.36
23
19
0.41
30
7.4-16.3
40
31
0.45
41
35
0.73
36
6.2-13.2
6.5
1.1
Pb
<5
a
<5
<5
—
<5
<5
<5
<5
<5
—
<5
<5
<5
<5
<5
—
<20
—
5.0
*— Denotes sample contaminated, analytical result compromised. (continued)
bRange corresponds to assuming zero metal concentration in a sample having nondetectable metal to
assuming metal concentration at the detection limit in a sample having nondetectable metal.
30
-------
TABLE 12. (continued)
Sample
Test 1 (1/15/91):
Waste la, mg/kg
Kiln ash la, mg/kg
Kiln ash TCLP leachate la, mg/L
Waste Ib, mg/kg
Kiln ash Ib, mg/kg
Kiln ash TCLP leachate Ib, mg/L
Hg
<0.1
<0.1
< 0.002
<0.1
<0.1
< 0.002
Ni
94
142
<0.05
71
85
<0.05
Se Ag
<10 <1.0
<10 1.0
<0.2 <0.02
< 10 < 1.0
< 10 < 1.0
<0.2 <0.02
Tl
133
194
1.0
123
197
1.2
V
42
66
0.24
44
57
0.44
Zn
43
38
9.5
49
35
8.2
Waste Ic, mg/kg
Scrubber exit flue gas, Mg/dscm
Test 2 (1/16/91):
Waste 2a, mg/kg
Kiln ash 2a, mg/kg
Kite ash TCLP leachate 2a, mg/L
Waste 2b, mg/kg
Kiln ash 2b, mg/kg
Kiln ash TCLP leachate 2b, mg/L
Waste 2c, mg/kg
Scrubber exit flue gas, fig/dscm
Test 3 (1/17/91):
Waste 3a, mg/kg
Kiln ash 3a, mg/kg
Kiln ash TCLP leachate 3a, mg/L
Waste 3b, mg/kg
Kiln ash 3b, mg/kg
Kiln ash TCLP leachate 3b, mg/L
Waste 3c, mg/kg
Scrubber exit flue gas, jtg/dscm
3 test scrubber liquor:
Solids, mg/kg
Filtrate, mg/L
TCLP regulatory level, mg/L
91
NAC 2.6-24.8 <96
121 83 41
6.7-95.7 3.5-25.7 97.2
<0.1 55
<0.1 88
<0.002 <0.05 <0.2 <0.02
< 0.002 <0.05 <0.2 <0.02
NA 2.4-24.5 <96
108
169
1.2
101
174
1.2
33
32
0.32
37
34
0.56
48
37
8.7
44
50
5.7
110
22
46
5.4-93.7 2.9-25.0 320
<0.1
<0.1
< 0.002
<0.1
<0.1
< 0.002
115
142
<0.05
93
159
<0.05
< 10 < 1.0 92
< 10 < 1.0 158
<0.2 <0.02 1.0
<10 , <1.0 102
< 10 < 1.0 181
<0.2 <0.02 1.8
36
37
0.50
35
40
0.68
42
40
7.0
59
23
5.6
<0.1 83
NA 2.2-19.9 <78 <8.0
<0.1 <10 <40, <5.0
< 0.002 <0.2 <0.5 <0.05
0.2 1.0 5.0
100 36 45
7.0-78.0 2.9-20.6 262
<40
<0.5
<0.2
7,250
5.9
°NA = not analyzed.
31
-------
TABLE 13. NONHAZARDOUS CONSTITUENT METAL ANALYSIS RESULTS
Sample
Test 1 (1/15/91):
Waste la, mg/kg
Kiln ash la, mg/kg
Kiln ash TCLP leachate la, mg/L
Waste Ib, mg/kg
Kiln ash Ib, mg/kg
Kiln ash TCLP leachate Ib, mg/L
Waste Ic, mg/kg
Scrubber exit flue gas, /ig/dscm
Test 2 (1/16/91):
Waste 2a, mg/kg
Kiln ash 2a, mg/kg
Kiln ash TCLP leachate 2a, mg/L
Waste 2b, mg/kg
Kiln ash 2b, mg/kg
Kiln ash TCLP leachate 2b, mg/L
Waste 2c, mg/kg
Scrubber exit flue gas, ng/dscm
Test 3 (1/17/91):
Waste 3a, mg/kg
Kiln ash 3a, mg/kg
Kiln ash TCLP leachate 3a, mg/L
Waste 3b, mg/kg
Kiln ash 3b, mg/kg
Kiln ash TCLP leachate 3b, mg/L
Waste 3c, mg/kg
Scrubber exit flue gas, /ig/dscm
3 test scrubber liquor:
Solids, mg/kg
Filtrate, mg/L
"Range corresponds to assuming zero
AI
86,500
62,500
672
80,700
123,000
642
75,600
4,910-5,000
73,200
104,000
664
71,400
106,000
693
76,400
4,230-4,320
54,200
96,900
570
68,700
112,000
1,070
57,600
3,800-3,880
5,450
25
Ca
14,700
13,600
8.6
14,300
16,100
9.6
18,600
6,950
15,300
15,900
8.4
15,100
13,200
6.4
16,200
7,390
16,800
14,200
8.1
13,200
15,200
6.9
12,400
6,690
2,550
640
metal concentration in a
Co
16
16
0.09
18
14
0.13
12
<15
10
14
0.24
11
12
0.23
13
<15
16
17
0.30
17
17
0.28
15
<11
8.3
<0.1
Fe
2,630
4,300
0.9
2,560
3,400
1.4
2,950
148-166
2,770
3,820
0.9
2,770
3,620
1.1
2,810
101-119
3,020
4,120
1.6
2,300
4,370
1.3
3,050
80.9-95.1
123
0.74
Li
5,620
6,500
18
5,710
5,700
22
5,530
1,290-1,310
8,250
7,180
30
8,520
5,940
22
7,810
1,560-1,610
7,390
6,640
23
6,470
8,030
23
7,610
1,190-1,230
<20
4.0
Mg
557
769
3.7
472
585
3.6
• 573
509-531
537
659
3.6
556
646
2.1
539
448-470
528
645
3.3
499
743
2.8
508
420-438
455
4.5
sample having nondetectable
Mn
19
37
0.30
18
27
0.32
27
2.5-11.4a
24
29
0.27
21
29
0.20
23
2.1-11.0
22
31
0.23
23
27
0.27
22
2.1-9.2
<4
<0.17
(continued)
metal to assuming metal concentration at the detection limit in a sample having nondetectable metal.
32
-------
TABLE 13. (continued)
Sample
Test 1 (1/15/91):
Waste la, mg/kg
Kiln ash la, mg/kg
Kiln ash TCLP leachate la, mg/L
Waste Ib, mg/kg
Kiln ash Ib, mg/kg
Kiln ash TCLP leachate Ib, mg/L
Waste Ic, mg/kg
Scrubber exit flue gas, /tg/dscm
Test 2 (1/16/91):
Waste 2a, mg/kg
Kiln ash 2a, mg/kg
Kiln ash TCLP leachate 2a, mg/L
Waste 2b, mg/kg
Kiln ash 2b, mg/kg
Kite ash TCLP leachate 2b, mg/L
Waste 2c, mg/kg
Scrubber exit flue gas, #g/dscm
Test 3 (1/17/91):
Waste 3a, mg/kg
Kiln ash 3a, mg/kg
Kiln ash TCLP leachate 3a, mg/L
Waste 3b, mg/kg
Kiln ash 3b, mg/kg
Kiln ash TCLP leachate 3b, mg/L
Waste 3c, mg/kg
Scrubber exit flue gas, iig/dscm
3 test scrubber liquor:
Solids, mg/kg
Filtrate, mg/L
Mo
<3
<3
0.10
<3
<3
0.09
<3
<29
<3
<3
0.06
<3
<3
0.07
<3
<29
<3
<3
0.10
<3
<3
0.11
<3
<23
<12
<0.2
P
170
<0.5
0.71
262
148
<0.5
94
148-370
231
151
<0.5
228
193
<0.5
234
85-306
250
148
4.4
275
238
<0.5
240
77-255
857
25
K
728
1,560
19
566
110
21
717
1,100-1,110
728
1,200
46
857
1,150
46
839
3,500
1,000
1,290
45
795
1,520
54
855
1,060-1,070
4,460
19
Na
176,000
287,000
6,420
196,000
207,000
6,850
210,000
91,100
199,000
215,000
8,660
203,000
243,000
9,660
208,000
108,500
186,000
183,000
7,770
61,300
209,000
9,000
176,000
30,100
19,600
2,950
Sr
137
148
0.42
138
144
0.44
133
77-86
156
144
0.38
148
143
0.29
149
70-79
144
153
0.39
147
178
0.33
144
66-73
196
0.42
Sn
<10
12
<0.2
<10
<10
<0.2
<1Q
<96
<10
<10
<0.2
<10
<10
<0.2
<10
<96
<10
<10
<0.2
<10
<10
<0.2
<10
<78
<40
<0.5
33
-------
The data from Table 12 are combined with appropriate waste and discharge stream
flowrate data to give the distributions of the BDAT list trace metals among the three incinerator
discharges (kiln ash, scrubber liquor, and flue gas) shown in Table 14. The distributions shown
in Table 14 are in percentages of the amount of each metal introduced in the waste feed, and,
thus, represent trace metal mass balances. The mass balances have been calculated over the
entire three-test program for two reasons:
• A significant quantity of slag was removed from the kiln after test program
completion; arbitrarily allocating slag quantities among the three tests would be
subject to challenge (the mass balances in Table 14 assume the slag removed had
the test program composite trace metal composition)
• A single test program composite scrubber liquor sample was analyzed; again,
arbitrarily allocating the scrubber liquor metal quantities to individual tests would
be subject to challenge
Table 15 represents the mass balance summary, analogous to Table 14, for the nonhazardous
constituent metals.
In both Tables 14 and 15, fractional distributions are shown as ranges for the scrubber
liquor and scrubber exit flue gas discharges of several metals. This is because scrubber liquor
and flue gas contained nondetectable quantities of many metals, and flue gas concentrations for
several metals were expressed as ranges as discussed above.
t
The data in Tables 14 and 15 show that, with the exception of phosphorus, metal mass
balance closures achieved for the test program ranged from 45 to 132 percent. This level of
closure is considered quite good when compared to past experience in measuring metal
distributions in discharges from combustion sources. This past experience shows that closures
in the 50 to 200 percent range represent about the best achievable. Phosphorus closures would
not be expected to be near 100 percent, as phosphorus is most likely oxidized to P2O5. The
oxidizing acid impingers in the metals train supplying the flue gas phosphorus concentrations
would not be expected to be very efficient at collecting P2O5.
Most of each metal fed was overwhelmingly accounted for by the kiln ash discharge for
all metals except barium and zinc (phosphorus, discussed above, is not a metal). A substantial
fraction of these two metals was found in the scrubber liquor, and, for zinc, in the scrubber exit
flue gas. The scrubber liquor also contained moderate fractions of calcium, potassium, and
copper. The scrubber exit flue gas contained moderate fractions of potassium and cadmium.
3.2.4 Sulfide and Fluoride Analysis Results
Table 16 summarizes the sulfide and fluoride analysis results for all samples analyzed
for these two analytes. As shown in the table, waste sulfide levels ranged from 0,006 to
0.080 percent. Kiln ash levels were comparable and ranged from 0.010 to 0.081 percent. Kiln
ash TCLP leachate levels were generally less than 10 mg/L.
34
-------
VO
CO
oo
s
in
10
m
en
m
ON
in
0
oo cn
2
o\
0
V
CN
VO
vd
oo
VO
vo O O
00
O\
is: v v
vo r-»
oq
vd '"J
o\ o o
a
o
1*
o3
S ?
o
00
«n
s
oq
cs
vo
s
0
o\
C3
o\
vq
oo
m
vd
f-
OX
ca
S
35
-------
TABLE 16. SULFIDE AND FLUORIDE ANALYSIS RESULTS
FOR WASTE FEED AND KILN ASH SAMPLES
Sample
Test 1 (1/15/91):
Waste la, %
Kiln ash la, %
Kiln ash TCLP leachate la, mg/L
Waste Ib, %
Kiln ash Ib, %
Kiln ash TCLP leachate Ib, mg/L
Waste Ic, %
Scrubber exit flue gas, mg/dscm
Test 2 (1/16/91):
Waste 2a, %
Kiln ash 2a, %
Kiln ash TCLP leachate 2a, mg/L
Waste 2b, %
Kiln ash 2b, %
Kiln ash TCLP leachate 2b, mg/L
Waste 2c, %
Scrubber exit flue gas, mg/dscm
Test 3 (1/17/91):
Waste 3a, %
Kiln ash 3a, %
Kiln ash TCLP leachate 3a, mg/L
Waste 3b, %
Kiln ash 3b, %
Kiln ash TCLP leachate 3b, mg/L
Waste 3c, %
Scrubber exit flue gas, mg/dscm
3 test scrubber liquor:
Filtrate, mg/L
Suffide
concentration
0.014
0.081
9
0.014
0.003
7
0.014
a
0.022
0.022
4
0.006
0.025
<0.4
0.020
—
0.080
0.035
6
0.018
0.010
11
0.019
—
50
Fluoride
concentration
2.72
6.80
85
1.30
4.05
122
5.92
, 17.4
4.33
3.48
78
3.48
2.44
668
5.33
4.75
4.98
2.54
46
4.39
1.34
42
3.51
135
0.3
*— Denotes not measured.
36
-------
Waste fluoride levels ranged from 1.3 to 5 percent. Kiln ash levels were also comparable
and ranged from 1.3 to 6.8 percent. Kiln ash TCLP leachate fluoride levels were generally less
than about 100 mg/L, although one leachate contained almost 670 mg/L of fluoride.
Table 17 takes the calculated composite test program weighted average concentrations
to calculate the fraction of waste feed sulfide and fluoride accounted for in the kiln ash
discharge. As shown in Table 17, the kiln ash sulfide accounted for 94 percent of the amount
present in the waste feed. Thus, apparently very little of the waste feed sulfide content was
destroyed by incineration. The kiln ash discharge accounted for about 64 percent of the waste
feed fluoride. The scrubber liquor and scrubber exit flue gas concentrations noted in Table 16
correspond to less than 0.01 and 4.1 percent of the fluoride fed over the test program. Mass
balance closure for fluoride (sum of the amounts accounted for in the kiln ash, scrubber liquor,
and scrubber exit flue gas relative to the amount fed) was 68 percent.
3.2.5 Flue Gas Particulate and HC1 Emissions
Table 18 summarizes the flue gas particulate levels at the two locations measured, the
scrubber exit flue gas and the stack. Stack particulate levels were uniformly lower than scrubber
exit flue gas levels. All levels measured were below the hazardous waste incinerator performance
standard of 180 mg/dscm at 7 percent O2.
Table 19 summarizes the stack HC1 levels and emission rates measured. All emission
rates were significantly below the hazardous waste incinerator performance standard of 1.8 kg/hr
(41b/hr).
33 CONCLUSIONS
Test results confirm that K088 can be incinerated in compliance with the hazardous
waste incinerator performance standards under the test conditions employed. Test data show
that the DRE for cyanide, the POHC in the waste, present at an average level of 4,120 mg/kg,
was greater than 99.999 percent. This exceeds the performance standard of 99.99 percent.
Furthermore, particulate emissions in the scrubber system exit flue gas were 110 to 130
mg/dscm, corrected to 7 percent 62, which meets the performance standard of 180 mg/dscm at
7 percent O2. HC1 emissions were very low, consistent with the low chlorine content of the
waste.
The kiln ash discharge still contained measurable cyanide, however, at a test average
level of 160 mg/kg. The degree of cyanide decontamination of the waste achieved (100 [1 -
amount discharged in kiln ash/amount fed]) was 97.3 percent. The kiln ash TCLP leachate
contained between 0.5 and 4.0 mg/L total cyanide. Cyanide was not detected in scrubber liquor.
Care must be exercised in incinerating K088 due to relatively low slagging temperatures.
The material tested in these tests formed slag in the kiln at temperatures just above the test
temperature of 980° C (1,800° F). Incinerator operation for these tests was to be at" kiln
temperatures below the slag formation threshold. However, despite attempts to avoid slag
formation, this proved to be a problem. About 20 percent of the kiln ash collected for the test
series was in the form of slag removed from the kiln walls after test program completion.
37
-------
TABLE 17. TEST PROGRAM SULFIDE AND
FLUORIDE DISTRIBUTIONS
Parameter
Sulfide
Fluoride
Waste feed:
Composite concentration, % 0.023 3.97
Amount fed, kg (Ib) 0.180 (0.396) 31.0 (68.3)
Kiln ash:
Composite concentration, % 0.031 3.63
Amount discharged3, kg (Ib) 0.168 (0.370) 19.7 (43.4)
Fraction of amount fed, % 94 64
'Includes 121 kg (266 Ib) of slag removed after test
program; assumes slag has the same composition as the
composite ash collected during the test program.
TABLE 18. FLUE GAS PARTICULATE LEVELS
Flue gas participate, mg/dscm at 7% O2
Test 1 (1/15/91)
Test 2 (1/16/91)
Test 3 (1/17/91)
Scrubber exit
121
110
131
Stack
78
100
70
TABLE 19. STACK GAS HCI EMISSIONS
Test 1 (1/15/91)
Test 2 (1/16/91)
Test 3 (1/17/91)
Stack gas HCI
Concentration
mg/dscm ppm g/hr
1.27 0.84 2.50
0.44 0.29 0.95
0.04 0.03 0.10
38
-------
Other test conclusions include:
• About a 30-percent reduction in waste weight occurs with incineration; the kiln
ash discharge was about 70 percent of the amount of waste feed
• The waste contained about 35 percent fked carbon, 4 percent volatile carbon,
61 percent ash, and 5 percent silica (an ash component). Roughly 50 percent of
both the fixed and volatile carbon content of the waste was destroyed during
incineration. Roughly 80 percent of the ash and silica introduced in the waste
feed was discharged as kiln ash; the remaining 20 percent was likely either
volatilized or entrained into kiln exit flue gas.
• An average of 90 percent of the waste total cyanide was analyzed to be amenable
cyanide; an average of 54 percent of the kiln ash cyanide was analyzed to be
amenable cyanide
• The waste was essentially free of semivolatile organic BDAT list constituents, as
were the kiln ash and scrubber liquor
• The major metal constituents of both the waste and kiln ash were aluminum,
calcium, and sodium; the major BDAT list trace metal contaminants of both the
waste and kiln ash were barium, nickel, and thallium, which were present at levels
in the 100 to 200 mg/kg range
• Hazardous constituent trace metal concentrations in TCLP leachates of waste feed
and resulting kiln ash were sufficiently low that neither the waste nor the
incineration kiln ash would be toxicity characteristic hazardous wastes
• The kiln ash discharge accounted for the major proportion of the waste feed metal
content for all metals except barium and zinc. A substantial fraction of these two
metals was found in the scrubber liquor, and, for zinc, in the scrubber exit flue gas.
The scrubber liquor contained moderate fractions of calcium, potassium, and
copper. The scrubber exit flue gas contained moderate fractions of potassium and
cadmium. Overall metal mass balance closures achieved (sum of the discharge
metal measured in the kiln ash, scrubber liquor, and scrubber exit flue gas as a
percentage of the amount of metal fed in the waste) ranged from 45 to
132 percent.
• About 94 percent of the waste sulfide content was accounted for by the kiln ash
discharge; apparently very little sulfide destruction occurred. About 64 percent
of the fluoride fed in the waste was accounted for by the kiln ash discharge.
Roughly 4 percent of the fluoride fed was accounted for in the scrubber exit flue
gas. Overall fluoride mass balance closure was 68 percent.
Test results were reported in the report:
• Whitworth, W. E., J. W. Lee, and L. R Waterland, "Pilot-Scale Incineration Tests
of Spent Potliners from the Primary Reduction of Aluminum (K088)," draft
August 1991.
39
-------
SECTION 4
INCINERATION OF CONTAMINATED SOILS FROM
THE DRAKE CHEMICAL SUPERFUND SITE
Another of the primary missions of the IRF is to support Regional Offices in evaluations
of the potential of incineration as a treatment option for contaminated soils and sediments at
Superfund sites. One priority site is the Drake Chemical site in Lock Haven, Pennsylvania.
EPA Region 3 and the U.S. Army Corps of Engineers (USAGE) (R. Schrock, Region 3;
D. Johnson, USAGE) requested that test burns be conducted at the IRF to support evaluations
of the suitability of incineration as a treatment technology for the contaminated soils and
sediments at the site.
^According to site investigation data, the soils and sediments at the Drake site are
contaminated to varying degrees with various organic constituents and several hazardous
constituent metals. With respect to incinerability evaluation, the primary objective was to
determine whether treating the soils and sediments by incineration would generate a residue
environmentally suitable for disposal (redeposit), without further treatment, at the Drake site
during full-scale remediation. Therefore, one primary concern was whether incineration could
effectively destroy the organic contaminants in fhe soils and sediments. Equally important was
what the fate of the trace metals in the soils and sediments would be when subjected to
incineration.
This test program was designed to evaluate the effectiveness of varying incinerator
operating conditions on organic contaminant destruction and the effects of these varied
conditions on the distributions of the trace metals in the discharge streams. Particulate and HC1
emission measurements were taken to establish the required flue gas cleanup performance
necessary to safely treat the soils and sediments. Analysis of the kiln bottom ash (treated
soil/sediment) indicated whether the ash would be suitable for redeposit without further
treatment, or what additional treatment would be required before its final disposal. Specific
questions answered in this test program were:
• Can rotary kiln incineration effectively destroy the organic contaminants in the
site soils and sediments?
• Will the kiln ash from incineration of the site soils and sediments have
characteristics that will allow it to be redeposited onsite, without further
treatment?
40
-------
« Can the incineration treatment of the site soils and sediments be performed in
compliance with the hazardous waste incinerator performance standards?
* What is the fate of the contaminant trace metals in the incineration of the site
soils and sediments?
• What are the effects of incineration temperature on contaminant metal fate and
kiln ash characteristics?
As originally conceived, this test program was to have consisted of an initial phase of
nine tests and an optional phase of four additional tests. The results from the initial-phase
testing specifically'the toxicity characteristics exhibited by the incinerator residuals, guided the
decision as to whether the optional testing would be needed. The initial-phase incineration
testing, modified to consist of five, not nine, tests (see Section 4.1.1), was conducted at the IRF
in January and February 1991. The IRF RKS, equipped with the venturi/packed-column
scrubber primary APCS (see Figure 1), was used for these tests. The toxicity characteristics of
all test program samples were demonstrated to be below regulatory threshold levels. These
results led to the conclusion that the optional testing would not be necessary to meet the stated
program objectives.
Results of the test program are discussed in the subsections that follow.
4.1 TEST PROGRAM
The Phase III record of decision (ROD) document for the Drake Chemical site indicates
that about 252,000 yd3 of contaminated soils and sediments will be excavated and decontaminated
onsite by a transportable rotary kiln incinerator. The ROD further indicates that these materials
are contaminated with varying levels of organic compounds and several hazardous constituent
trace metals, including arsenic, barium, cadmium, chromium, lead, and mercury.
For the test program, 17 55-gal drums of the contaminated site soils were excavated and
shipped to the IRF for testing. These 17 drums consisted of:
• Two drums of soil, one from each of two "organic contamination hot spots,"
denoted as locations O-l and O-2
• Two drums of lagoon sediment from locations denoted as L-l and L-2
• Seven drums of soil from the "inorganic contamination, hot spots" (one from each
of three locations denoted as M-l, M-2, and M-3; and two from locations denoted
as M-4 and M-5)
• Six drums of composite soil representing general site characteristics from locations
denoted as G-l, G-2, G-3, G-4, G-5, and G-6
41
-------
4.1.1 Test Conditions
..As "ot?d above» the objective of the proposed test program was to evaluate the
suitability of incineration as a treatment technology for the contaminated soils and sediments at
the Drake Chemical Superfund site. Table 20 outlines the nine originally planned test
conditions.
Analyses of samples of the soils actually excavated, however, revealed that several soils
contained very low concentrations of contaminants, and that conducting incineration testing with
these would therefore likely produce data of limited usefulness. As a result of these analyses,
the scope of the test program was modified to more adequately meet the original program
objectives. The modified test conditions are included in Table 20. The specific changes
included:
• Deleting the original test conditions 4, 5, 8, and 9
• Conducting Tests 3b and 7 with a kiln temperature of 538° C (1,000° F) instead
of the originally planned 982°C (1,800°F) and 816°C (1,500°F), respectively
• Reassigning the soil selection for each test, as shown in Table 20
• Spiking each soil from locations L-2 and O-2 with naphthalene and
1,4-dichlorobenzene to 3,000 and 130 ppm, respectively, and using these spiked
soils for Tests 6 and 7; and designating naphthalene and 1,4-dichlorobenzene as
the surrogate POHCs
Of the five retained conditions, Tests 1, 2, 3a, and 3b were designed to study the fate
of the inorganic contaminants (trace metals). Tests 1 and 2 studied the distribution of the trace
metal contaminants throughout the incinerator system. These tests also provided information
on the concentrations of trace metals in the kiln and flue gas flyash TCLP leachates. These data
are important as they may affect whether additional treatment of the incinerator solid residual
streams is needed. Tests 3a and 3b investigated the effects of kiln temperature on the trace
metal concentrations in the kiln ash and scrubber liquor streams. All these tests were conducted
with the soils in their original, as-received, form. Tests 6 and 7 were designed to study the
destruction of the organic contaminants. The destruction of the spiked POHCs became the
principal indicator of the effectiveness of incineration under these tests conditions.
Tests 1, 2, 3a, and 3b were performed with the high-inorganic-contaminant soils from
locations M-2, M-5D, and O-l, respectively. Tests 6 and 7 followed, using the spiked location
L-2 lagoon sediments and the location O-2 high-organic-contaminant soil.
The five tests were conducted from January 30, 1991 through February 7, 1991. Each
test consisted of incinerating one 55-gal drum of the contaminated soils over a 4- to 5-hr period.
Tests 3a and 3b were performed in one day, with sufficient time allowed in between subtests to
achieve steady-state operation at the target kiln temperature of 816° C (1,500° F) for Test 3a and
538°C (1,000°F) for Test 3b. Test soils were fed to the kiln via the fiberpack drum ram feeder
system. Each fiberpack contained 4.5 kg (10 Ib) of soil. One fiberpack was charged into the kiln
every 5 min, resulting in soil feedrates of nominally 55 kg/hr (120 Ib/hr).
42
-------
TABLE 20. TARGET TEST CONDITIONS
=========
Revised
X-
Test
no.
1
2
,3a
3b
4a
4b
5a
5b
6
7
8
9
r a *^UKM**J
Test
material
M-5
M-4
M-l
M-l
M-2
M-2
M-3
M-3
O-l
O-2
L-l
L-2
Kiln
temperature,
°C (°F)
816(1,500)
816 (1,500)
816 (1,500)
982 (1,800)
816 (1,500)
982 (1,800)
816 (1,500)
982 (1,800)
816 (1,500)
816(1,500)
816 (1,500)
816(1,500)
Kiln Solids
Test Test temperature, residence
no. material °C (°F) time, hr
1
. 2
3a
3b
Deleted
Deleted
Deleted
Deleted
6
7
Deleted
Deleted
• •-—
M-2 816 (1,500) 0.5
M-5D 816 (1,500) 0.5
0-1 816 (1,500) 0.5
0-1 538(1,000) 0.5
L-2 816(1,500) 0.5
O-2 538 (1,000) 0.75
•
Organic
spiked
No
No
No
No
Yes
Yes
=====
For all tests:
Total waste/soil feedrate
Afterburner temperature
Kiln exit flue gas O2
Afterburner exit flue gas O2
Scrubber blowdown rate
Venturi liquor flowrate
Venturi pressure drop
Packed tower liquor flowrate
Scrubber liquor temperature
55 kg/hr (120 Ib/hr)
1,093° C (2,000° F)
11 percent
7 percent
0 L/min (0 gpm) or minimum operable
76 L/min (20 gpm)
6.2 kPa (25 in WC)
115 L/min (30 gpm)
49° C (120°F)
43
-------
Table 21 compares the actual test operating conditions to the respective targets for the
tests performed. As shown in the table, the average kiln temperatures were within about 15° C
(25° F) of the target temperatures for all of the tests. Afterburner temperatures were
maintained within 3°C (5°F) of the 1,093°C (2,000°F) target for all tests. Both kiln exit and
afterburner exit flue gas O2 levels were somewhat higher than target levels for all tests, however,
because of excessive air inleakage into the kiln resulting from an inability to tightly secure the
rotating kiln seals.
Table 22 summarizes the total amount of soil fed to the RKS for each test, and the
cumulative amount of kiln ash discharged over each test's duration. As shown, except for Test 7,
the weight of ash discharged was generally about 70 percent of the weight of soil fed to the kiln.
4.1.2 Sampling and Analysis Procedures
Because the objectives of Tests 1, 2, 3a, and 3b were different from those of Tests 6
and 7, different sampling and analysis procedures were employed for each test group. However,
several procedures were performed for all tests, including:
• Obtaining a composite sample for the soil feed from each drum before the soil
was packaged into fiberpack containers
• Collecting a composite kiln ash sample
• Collecting a composite scrubber liquor sample
• Continuously measuring O2 concentrations at the kiln exit; O2, CO, CO2, and
TUHC at the afterburner exit; O2, CO2, and NOX at the scrubber exit; and CO and
TUHC at the stack
• Sampling the flue gas at the stack for HC1 and particulate using a Method 5
sampling train
The above were the only sampling procedures employed for Tests 3a and 3b. In
addition to the above, the following were performed for Tests 1 and 2:
• Sampling the flue gas at the afterburner exit (i.e., upstream of the scrubber) for
particulate load and for trace metals (excluding mercury) using a Method 17
sampling train, modified with multiple metals train impingers
• Determining the particle size distribution of the afterburner exit flue gas
particulating using an Anderson cascade impactor train
• Sampling the flue gas upstream and downstream of the scrubber for mercury using
a Method 101A train at each location
• Sampling the flue gas downstream of the scrubber system for particulate and trace
metals (excluding mercury) using a Method 5 sampling train modified for multiple
metals capture
44
-------
TABLE 21. TARGET VERSUS ACTUAL OPERATING CONDITIONS FOR THE DRAKE
CHEMICAL SOIL TESTS
Test
no.
1
2
3a
3b
6
7
Kiln
Afterburner
Target Actual average Actual Actual
temperature, temperature, Target O2, average O2, Target O2, average O2,
° f* (° ¥T"^ * f1 (^ \^\ tffi tyf) ^7o ' /o~
816 (1,500) 826
816 (1,500) 823
816 (1,500) 829
538 (1,000) 546
816(1,500) 822
538 (1,000) 553
(1,519)
(1,513)
(1,524)
(1,015)
(1,512)
(1,027)
TABLE 22. SOIL
Test Test soil
1 (1/30/91)
2 (2/5/91)
3a (1/31/91)
3b (1/31/91)
6 (2/6/91)
7 (2/7/91)
M-2
M-5D
O-l
O-l
L-2
0-2 «
11.0
11.0
11.0
11.0
11.0
11.0
FEED AND
Total soil
fed, kg (Ib)
240 (529)
232 (512)
112(246)
113 (249)
240 (529)
209(460)
13.3 7.0
13.1 7.0
13.8 7.0
17.0 7.0
12.7 7.0
15.4 7.0 .',
ASH COLLECTED
Total ash collected
Weight, Fraction of
kg (Ib) soil fed, %
173 (381) 72
177 (390) 76
74 (163) 66
83 (183) 73
187 (411) 78
183 (404) 88
. 8.7
9.2
9.2
11.8
9.3
9.9
45
-------
In addition to the sampling performed for a}l tests noted above, the following was
performed for Tests 6 and 7:
• Sampling the flue gas downstream of the scrubber system for semivolatile POHCs
using a Method 0010 sampling train
Figure 9 shows the system sampling locations sampled and identifies the sampling
procedures used.
During each test, kiln ash was continuously removed from the ash hopper, via an ash
auger system, and deposited in a 55-gal drum. Scrubber liquor was completely recycled (zero
blowdown) for the duration of each test. After flue gas sampling was completed for each test
(no flue gas sampling was performed for Test 3), soil feed to the kiln was suspended. The
incinerator was operated, firing natural gas auxiliary fuel for a period of 2 hr, or until the kiln
was visually clear of ash, whichever was longer. After this period, the ash collection drum was
removed and kiln ash samples taken. In addition, the entire volume of scrubber system liquor
was transferred to a storage tank. During this transfer, a scrubber liquor sample was collected
from a tap in the transfer line. The scrubber system was then recharged with fresh makeup and
restarted in preparation for the next test.
In addition to analyzing flue gas sampling trains for their sampled analyte set, the
following were performed for Tests 1, 2, 3a, and 3b.
• Analyzing the soil feed and kiln ash samples for trace metals (arsenic, barium,
cadmium, chromium, copper, mercury, nickel, lead, selenium, silver, and zinc)
• Analyzing the soil and kiln ash TCLP leachates for trace metals
• Analyzing the scrubber liquor samples for trace metals
In addition, for Tests 1 and 2 the following was performed:
• Subjecting an aliquot of the afterburner exit particulate collected with the
Method 17 sampling train to the TCLP and analyzing the leachate for trace metals
For Tests 6 and 7, the analysis protocol included:
• Analyzing the soil feed, kiln ash, and scrubber liquor samples for volatile and
semivolatile organic contaminants
42 TEST RESULTS
The test program was completed on February 7, 1991. Sample analysis, data reduction,
and test results interpretation continued through September 1991. Test report drafting was
nearing completion at the end of FY91.
46
-------
Q<
a>
1
o
e
o
S
V)
3
3
C
C
3
(A
3
13
•z
z
1
a
I
ii!
• * i°
*!»
55 "3, «
lilt
•8 ^ S |
illl
w .§ 1 2
S.-0
f 1?
S. g. «
M
Hit
u Jj C g
2 **
*
« u
p
D
0
Z
8
8
•<5*
In
«* t.
H
V3
Kjl 55
13 «^jj 5
O 3 ,«
V3 «5J *-
e
1
.S
"a.
O
< <
<
u
<
0 X
<
X X
X
X X
X X
XX XX
X
X
X
1 &
E"> "O o
00 « I «
1 = J H- '1 M
,111115
I 1 1 ^ 1 8 I
c/5
1
«
^C
"a,
i
I
ON
gj
M
Sxi
s
•o
M *!' «
las
3 ii n ii
| X < 0
47
-------
Preliminary test conclusions include:
• Organic contaminants in the test soils can be destroyed to greater than
99.99 percent DRE. Naphthalene, spiked into test soils at 3,000 mg/kg for two
tests (Tests 6 and 7) was destroyed at a DRE of greater than 99.995 percent.
1,4-dichlorobenzene, spiked into the test soils at 130 mg/kg for the same two tests,
was not detected in incineration flue gas; detection limits corresponded to a DRE
of greater than 99.89 percent. No treated soil (kiln ash) contained detectable
levels of organic contaminants.
• Particulate levels in the flue gas at the exit of the venturi/packed-column scrubber
APCS were less than 20 mg/dscm at 7 percent O2, in compliance with the
hazardous waste incinerator performance standard of 180 mg/dscm at
7 percent O2
• None of the soils tested, or the kiln ash resulting from their incineration, would
be considered a toxicity characteristic (TC) hazardous waste due to its leachable
trace metal contents
• No test scrubber liquor (including suspended solids) would be considered a TC
hazardous waste due to trace metal concentrations. However, lead concentrations
in test scrubber liquors were at levels over 50 percent of the TC regulatory level
in some cases. This suggests that the scrubber liquor discharge from a wet
scrubber APCS could become a TC hazardous waste in the incineration of "hot
spot" lead-containing soils, or under scrubber operation at minimum blowdown.
• The flyash collected at the afterburner exit (upstream of the wet scrubber APCS)
would be a TC hazardous waste due to leachable chromium and lead
concentrations. This suggests that the collected particulate from a dry ACPS
(such as a fabric filter) would be a TC hazardous waste.
48
-------
SECTION 5
INCINERATION OF PCB-CONTAMINATED SEDIMENTS FROM
THE NEW BEDFORD HARBOR SUPERFUND SITE
EPA Region 1 is conducting the remedial design (RD) for the remediation of a
Superfund site located in New Bedford Harbor near New Bedford, Massachusetts. According
to the record of decision (ROD) document of 1990, the EPA has identified approximately
10,000 yd3 of contaminated sediment in the harbor. Incineration of the dredged sediment is the
selected treatment option. In support of the RD, incineration technologies will be examined to
determine the optimum equipment configuration and incinerator operating parameters for the
waste material. The EPA (M. Sanderson, Region 1;K. Howe, USAGE, coordinators) requested
that test burns be conducted at the IRF to support the RD for this Superfund site.
The sediment at the New Bedford Harbor site is contaminated with from 4,000 ppm to
more than 200,000 ppm of PCBs. In addition to PCBs, other contaminants are present, including
polynuclear aromatic hydrocarbons (PAHs), and trace metals (copper, chromium, lead, and
cadmium). The primary objective of this test program was to obtain data to support the RD
plans and specifications. Therefore, the test conditions were designed to evaluate the
effectiveness of varying incinerator operating conditions in the destruction of PCBs and other
pollutants. Specifically, the test program attempted to answer these questions:
• Can incineration effectively destroy PCBs to the required destruction and removal
efficiency (DRE) of 99.9999 percent?
• What is the distribution of the contaminant trace metals in the discharge streams
during incineration of the sediment?
• What are the effects of incineration excess air and temperature on organic
constituent destruction and metals distributions, including the leachability of the
metals from the kiln ash?
• What is the effectiveness of the air pollution control system (APCS) in collecting
particulate and trace metals?
• Can the treated sediment (i.e., kiln ash) from the incinerator be disposed of as
non-hazardous solid waste?
The test program consisted of a set of three incineration tests in the RKS at the IRF.
These tests were aimed at evaluating PCB destruction, and the fate of contaminant trace metals
49
-------
in the sediment, as functions of kiln temperature and kiln excess air level. The tests were
performed in the RKS with the venturi/packed-column scrubber as the primary APCS. The
following subsections document the findings of the test program.
5.1
TEST PROGRAM
Eight 30-gal drums of sediments were dredged from the Hot Spot area of New Bedford
Harbor for these tests. A characterization sample representing each drum dredged was shipped
to the IRF for pretest analyses. These samples were subjected to proximate, PCB, and
hazardous constituent trace metals analyses. The results of these analyses showed that the
average total PCB concentration of the eight drums was 5,300 ppm, as received. The level
required in an RKS feed to be able to just establish 99.9999 percent DRE at a typical RKS
feedrate of 68 kg/hr (150 Ib/hr) is 5,100 ppm, just below the average dredged drum
characterization sample level. Consequently, it was decided to spike the test sediment to higher
PCB concentrations to provide a margin in the ability to establish 99.9999 percent DRE. The
material used to spike the sediments was an Askarel transformer fluid comprised of roughly
75 percent Aroclor 1242 and 25 percent 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. Prior to testing, the combined sediments were
repackaged into 1.5-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 prepared without the PCB spike for testing using
only the native sediment.
5.1.1 Test Conditions
The test series was designed to evaluate the effects of incinerator operating conditions
on PCB destruction and trace metal distributions in the incinerator discharge streams. The
operating parameters to be varied were kiln exit 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 percent. For all tests, the
operating conditions noted in Table 23 were to be held at the nominal values noted in the table.
TABLE 23. INCINERATOR SYSTEM OPERATING CONDITIONS HELD CONSTANT
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 L/min (0 gpm) or minimum operable
76 L/min (20 gpm)
6.2 kPa (25 in WC)
115 L/min (30 gpm)
49°C (120°F)
50
-------
Table 24 summarizes the actual incinerator exit temperatures and flue gas levels,
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 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 due to the inability to tightly secure a rotating kiln seal. The
minimum O2 achievable was 9 percent at the kiln exit. The maximum O2 tested was 11.2 percent.
As a practical matter, these two levels present comparable combustion environments.
Consequently, it was not possible to test kiln excess air as a variable.
5.12 Sampling and Analysis Procedures
The scope of the sampling effort undertaken during this test program is illustrated in
Figure 10, in which the sampling locations and the corresponding sample collection methods are
identified. Specifically, the sampling effort during each test consisted of:
• 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, NOX, and TUHC levels at the venturi/packed-column scrubber exit;
and O2, CO, and CO2 levels in the stack
• Sampling flue gas at the scrubber system exit for PAHs and PCBs
• Sampling flue gas at the scrubber system exit for polychlorinated dibenzo-p-dioxins
(PCDDs) and dibenzofurans (PCDFs)
• Sampling flue gas at the scrubber system exit for volatile organics
• Sampling flue gas upstream of the scrubber system for particle size distribution
• 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 scrubber system and at the stack downstream of the
secondary APCS for particulate and HC1 using Method 5 to comply with permit
requirements
In addition, grab/composite samples of the sediment feed from each drum were taken before
the sediments were mixed, and a grab/composite sample of the mixed sediment feed was taken.
51
-------
>^J
£"*
1
o
s
o
§
t"™*4
i
M
£
o
f*<
s
o
B
>5<
o
o
2
§
O
i
irS
AL VERSUS
D
8
«<£
S
d>3
1
H
^
cf
E
bO
2
I
o
§
fifl
s
a
H
S)
2
§•
<3
Maximum
I
:=
s
o
I
-2
08
Q
1
S
•^A
••4
i
CN o co o
J^ O\ O\ O
o\ o 10 •>*
1— 1 T— ( T— 1 T— 1
0000
oq vq 01 o
VO VO O O
T—1 t-
T— 1 O ON O
r-T r-T T-T T-H"
V- : J v J *- J *^ J
. . i tr.
oi oo oo oo
oo o\ o\ c\
/*~s y*~N
fill
f ||l
T-I co co r> in o\ oo
r^" o\ o\ o\
g^^ 0s 0s o*1
O C3 C3
r- T r-T T-T T-T
VO Ol Ol Ol
T-I oo oo oo
oo o\ o\ o\
T-H T— 1 T— 1 T— 1
o\ o\ o\ o\
ID ON T— ' T— i
co co co co
« JD
^ ^ co co
^,
'8
1
<* o •* o
CN •* vq o
oo t^* r^* oo
o o o o
CN vq o\ co
co co oi ••*
t^- t- t^ c~
0000
OI OI CN CN
of of of of
OO OO OO OO
o o o o
T-T T— T T-T T-T
O\ vo O*\ f^
O) OJ OI OI
of of of of
T-I CO vo "3-
OI T-I OI T-H
CO •<* T-I OO
co o co oo
T— 1 T— 1 T— 1 T— 1
of of of of
If) T-H •*+• OO
O\ O OS O\
f*"^ c**^ t— ' i f— i
o o o o
OJ.OJ.OJ.OJ.
C3 C3 C3 O
01.01.0^01.
T-H~ T-T T-T T—T
T— 1 T— 1 T— 1 T— 1
OS OS ON O\
IO O\ T— 1 r-l
T-I T-I OI 01
CO CO CO CO
,-H a ja
co co
52
-------
S
li§
§5 ® —
'S -J§ •§ £
« s 8 t5
"3 5 « «
•S S a p.
•a s ^
00C8'"
1l«^
l§y^
o *• a
w s
B j-
~
X X
X
X
X
X
X
X
C3
S
OJD
I
x
x
X
X
X
to ^ in vo f»
53
-------
Each test was run with scrubber system operating at, or near, total recycle (zero
blowdown). At the conclusion of each test day, the incinerator was operated on natural gas for
2 hr after waste feed cessation. After the 2-hr cleanout time, the scrubber system was drained
to a storage tank. A composite scrubber liquor sample was collected from the collection tank
after draining was complete, and a composite kiln ash sample was collected from the ash
collection system drum.
The sample collection procedures resulted in eight individual sediment samples and one
composite sediment sample. One sample of the PCB spike material was taken, resulting in a
total of 10 feed samples for the test series. One set of kUn ash samples was taken for each test.
This included the period of time with native sediment only as the feed. No flue gas
characterization was done during the native sediment feed period; however, one set of flue gas
characterization samples was collected during spiked sediment feed periods for all three tests.
Scrubber liquor samples were collected for all tests.
An aliquot of the composited sediment feed and each test's kiln ash was subjected to
the TCLP leaching procedure and analyzed for cadmium, chromium, copper, and lead. Waste
feed samples, kiln ash samples, and scrubber liquor samples were analyzed separately for PCBs,
PAHs, and cadmium, chromium, copper, and lead. The composite sediment feed sample was
also subjected to proximate (moisture, ash content, and heat content) analysis, and ultimate (C,
H, O, N, S, Cl) analysis.
5.2
TEST RESULTS
The results of the test program are discussed in the subsections that follow. Test results
are grouped by analyte class.
52.1 Proximate and Ultimate Analysis Results
The proximate and ultimate analysis results for the composite sediment sample analyzed
are presented in Table 25. The high moisture content of the test sediments is consistent with
their marine origin. Table 26 summarizes the cumulative sediment weight fed for each test and
the total amount of kiln ash collected. As indicated in the table, between 25 and 30 percent 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 analysis shown in Table 25.
523, PCB, Semivolatile and Volatile Organic, and Dioxin/Furan Analysis Results
Table 27 summarizes the PCB contents of each incineration test sample. As indicated,
the spiked sediment feed contained 3.48 percent Aroclor 1242 and 1.11 percent Aroclor 1254.
The kiln ash resulting from the incineration of the sediments (both spiked and native) had
substantially reduced, though still significant, PCB contents. The kiln ash for the spiked sediment
feeds contained between 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
temperature tested for Tests 2 and 3 did not result in significantly lower kiln ash PCB
concentrations than the lower temperature tested in Test 1. The kiln ash resulting from native
sediment feed also contained significant PCB levels.
54
-------
TABLE 25. PROXIMATE AND ULTIMATE ANALYSIS RESULTS
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
TABLE 26. SEDIMENT FEED AND ASH COLLECTED
Ash collected
1
2
3a
3b
Test
(3/15/91)
(3/19/91)
(3/21/91)
(3/22/91)
Total sediment fed,
kg(lb)
284 (625)
284 (625)
229 (504)
318(700)
Weight,
kg(lb)
85 (187)
76 (168)
62 (136)
80 (175)
Fraction of
feed,
%
30
27
27
25
55
-------
TABLE 27. PCB ANALYSIS RESULTS
PCB concentration
Sample
Aroclor 1242 Aroclor 1254
Spiked sediment feed, % 3.48
Test 1:
Kiln ash, mg/kg 133
Scrubber liquor, |ig/L < 1
Scrubber exit flue gas, jig/dscm 0.76
Test 2:
Kiln ash, mg/kg 96
Scrubber liquor, ng/L < 1
Scrubber exit flue gas, jig/dscm 0.54
Test 3b:
Kiln ash, mg/kg 177
Scrubber liquor, ng/L < 1
Scrubber exit flue gas, ng/dscm <0.26
Composite native feed, mg/kg 4,850
Test 3a:
Kiln ash, mg/kg 57
Scrubber liquor, |ig/L < 1
1.11
84
<0.3
0.22
32
<0.3
0.21
68
<0.3
<0.09
1,300
44
<0.3
No scrubber liquor sample contained detectable PCB at practical quantitation limits
(PQLs) of 1 (ig/L for Aroclor 1242, and 0.3 |ig/L for Aroclor 1254. The scrubber exit flue gas
contained low, though measurable, levels of both PCB formulations in Tests 1 and 2.
Table 28 summarizes the degree of PCB decontamination achieved in each test, in terms
of the fraction of the amount of PCB introduced in the incinerator feed accounted for by the
resulting kiln ash. As shown in the table, about 0.1 percent of the Aroclor 1242 and about 0.1 to
0.2 percent of the Aroclor 1254 feed in spiked sediments was accounted for in the kiln ash
produced. The remaining 99.8 to 99.9 percent was removed and largely destroyed, as discussed
below. Higher fractions of feed PCBs were present in the kiln ash from the native sediment test:
0.3 percent for Aroclor 1242, and 0.9'percent for Aroclor 1254.
The data shown in Tables 27 and 28 confirm that incineration under the conditions
tested was not sufficient to completely 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 decontamination, as was the kiln solids residence time (0.5 hr). However, the New
Bedford Harbor marine sediments contained substantial moisture, over 60 percent. Evidently,
with such high moisture content, solids bed temperatures were not raised to levels needed for
more complete PCB destruction in the residence time available. It is suspected that longer solids
residence times, perhaps on the order of 1 hr, would allow essentially complete, or certainly more
complete, PCB decontamination.
56
-------
TABLE 28. PCB DECONTAMINATION EFFECTIVENESS
Parameter
Test 1:
Sediment feed
Concentration, %
Amount fed, kg
Kiln ash
Concentration, mg/kg
Amount discharged, g
Fraction of amount fed, %
Test 2:
Sediment feed
Concentration, %
Amount fed, kg
Kiln ash
Concentration, mg/kg
Amount discharged, g
Fraction of amount fed, %
Test 3a:
Sediment feed
Concentration, %
Amount fed, kg
Kiln ash
Concentration, mg/kg
Amount discharged, g
Fraction of amount fed, %
Test 3b:
Sediment feed
Concentration, %
Amount fed, kg
Kiln ash
Concentration, mg/kg
Amount discharged, g
Fraction of amount fed, %
Aroclor 1242
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
Aroclor 1254
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
57
-------
Table 29 takes the scrubber exit flue gas PCB concentrations noted in Table 27, and
combines them with sediment feedrate and flue gas flowrate data to give the PCB DREs
achieved for the tests. As shown in the table, greater than the regulation-required
99.9999 percent 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 ng/L in scrubber liquor, and 6 ng/dscm in scrubber exit
flue gas.
Results of the scrubber exit flue gas PCDD/PCDF measurements are summarized in
Table 30. As shown, total tetra-CDD (TCDD), penta-CDD (PeCDD), hexa-CDD (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 TCDF. The 2,3,7,8-TCDD
toxicity equivalents corresponding to the isomer concentrations noted are also given in Table 30.
As shown, the scrubber exit flue gas PCDD/PCDF levels corresponded to 2,3,7,8-TCDD toxicity
equivalents ranging from nominally 0.05 to 0.10 ng/dscm.
5.23 Trace Metal Discharge Distributions
Two primary objectives of the test program were to evaluate the fate of the contaminant
trace metals cadmium, chromium, copper, and lead in the incineration treatment of the New
Bedford Harbor sediments, and to investigate whether incineration conditions affected the
distribution of these metals in the incinerator discharges. Table 31 summarizes the
concentrations of the test metals in sediment samples and in each of the incinerator discharge
streams.
TABLE 29. PCB DREs
Parameter
Sediment feed:
Sediment feedrate, kg/hr
Aroclor 1242 feedrate, g/hr
Aroclor 1254 feedrate, g/hr
Scrubber exit flue gas:
Flue gas flowrate, dscm/min
Aroclor 1242 concentration, |ig/dscm
emission rate, ng/hr
DRE, %
Aroclor 1254 concentration, iig/dscm
emission rate, jig/hr
DRE, %
Test 1
(3/15/91)
69.5
2,420
773
33.9
0.76
1.6
99.999936
0.22
0.5
99.99994
Test 2
(3/19/91)
69.5
2,420
773
32.9
0.54
1.1
99.999956
0.21
0.5
99.99994
Test 3b
(3/21/91)
69.3
2,410
771
30.6
<0.26
<0.5
> 99.999980
<0.09
<0.2
> 99.99998
58
-------
TABLE 30. FLUE GAS PCDD/PCDF ANALYSIS RESULTS
Analyte
Total TCDD
2,3,7,8-TCDD
Total PeCDD
1,2,3,7,8-PeCDD
Total HxCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
Total HpCDD
1,2,3,4,6,7,8-HpCDD
OCDD
Total TCDF
2,3,7,8-TCDF
Total PeCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
Total HxCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
Total HpCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
2,3,7,8-TCDD equivalents
Scrubber exit
flue gas concentration,
ng/dscm
Test 1 Test 2
(3/15/91) (3/19/91)
0.016
0.003
0.005
0.002
0.011
0.001 '
0.001
0.002
0.008
0.0083
0.043
2.79
4.60
0.78
0.075
0.089
0.13
0.035
0.013
0.016
0.001a
0.005
0.0 lla
0.002
0.013
0.104
0.010
0.003
0.010
0.001a
0.008
0.0013
0.001
< 0.001
0.023
0.013
0.060
1.78
2.80
0.30 '
0.033
0.030
0.070
0,020
0.008
0.010
0.001
0.020
0.010
0.003
0.010
0.052
Test 3b
(3/21/91)
0.008
0.003
0.005
< 0.003
0.008
< 0.003
< 0.002
< 0.003
0.019
0.008
0.030
1.48
2.70
0.43
0.035
0.046
0.15
0.038
0.014
0.022
< 0.002
0.002
0.0223
0.005a
0.016
0.062
"Estimated maximum possible concentration.
59
-------
TABLE 31. TRACE METALS ANALYSIS RESULTS
Sample
Sediment feed:
Composite, mg/kg
Composite TCLP leachate, mg/L
Test 1 (3/15/91):
Kiln ash, mg/kg
Kiln ash TCLP leachate, mg/L
Afterburner exit flue gas, fig/dscm
Scrubber exit flue gas, jig/dscm
Scrubber liquor, mg/L
Test 2 (3/19/91):
Kiln ash, mg/kg
Kiln ash TCLP leachate, mg/L
Afterburner exit flue gas, ug/dscm
Scrubber exit flue gas, jig/dscm
Scrubber liquor, mg/L
Test 3a (3/21/91):
Kiln ash, nig/kg1
Test 3b (3/21/91):
Kiln ash, mg/kg
Kiln ash TCLP leachate, mg/L
Afterburner exit flue gas, ug/dscm
Scrubber exit flue gas, ug/dscm
Scrubber liquor, mg/L
TCLP regulatory level, mg/L
Cd
Cr
Cu
7.4
0.11
161
0.041
308
0.066
2.3 367
785
1.0
5.0 -a
Pb
236
1.2
9.5
0.26
42.3
34.9
0.14
376
0.048
219
158
1.9
608
6.3
571
421,
4,5
277
0.71
1,030
903
8.8
2.7
0.046
77.6
55.3
0.27
434
0.030
136
73.2
1.4
828
3.01
768
436
2.6
75.6
0.17
1,814
1,273
5.4
96
2.0
0.043
51.6
79.0
0.73
357
0.03
126
83.9
1.3
721
3.0
519
750
3.4
62
0.41
984
2,020
5.6
5.0
Not a TCLP metal.
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 going from sediment to kiln
ash during incineration. However, the cadmium and lead concentrations in kiln ash were
significantly lower in the high-kiln-temperature 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 scrubber exit flue gas were generally higher for Tests 2 and 3 than for Test 1 as well.
Both these trends are the result of the volatile behavior of these two metals. The extent of
volatilization of these metals was evidently higher in Tests 2 and 3 than in Test 1, giving rise to
60
-------
lower kiln ash, and generally higher flue gas, concentrations of these metals for the higher
temperature tests.
Table 31 also notes the TCLP regulatory limit for the three TCLP metals determined.
Comparing composite feed and kiln ash TCLP leachate and scrubber liquor metal concentrations
to the TCLP regulatory levels shows that neither the composite sediment feed nor any test's kiln
ash would be a toxicity characteristic (TC) hazardous waste based on leachable cadmium,
chromium, or lead concentrations.
The scrubber liquor cadmium and chromium concentrations noted in Table 31 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 scrubber 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 which was held in
a storage tank until all analyses were completed. In accordance with the procedure, a sample
of the scrubber liquor was filtered and the percent solids content was measured. If the solids
content is less than 0.5 percent, the procedure specifies discarding the solids and using the
filtrate as the TCLP leachate. As this was the case, the filtrate 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 31 can be combined with feed soil and
discharge stream mass flowrate information to better show how the metals distribute among the
discharge streams as a function of incineration condition. These distributions are summarized
in Table 32. The distribution fractions in Table 32 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 case been 100
percent. Note that the sum of the normalized values (the totals) for each metal is indeed
100 percent in Table 32. Use of distribution fractions 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 system
portion of the RKS ranged from 52 to 103 percent for cadmium, 72 to 79 percent for chromium,
71 to 86 percent for copper, and 38 to 66 percent for lead. These levels of mass balance closure
are considered excellent when viewed in light of past experience on achieving trace metal mass
balance closures from a variety of combustion sources, incinerators included. Typical mass
balance closure results from this past experience have been, at best, in the 30 to 200 percent
range.
Several interesting observations emerge from the data in Table 32. The first is that, in
these tests, chromium and copper exhibited relatively nonvolatile behavior. The kiln ash
discharge represented the predominant fraction of metal discharged for these two metals. The
kiln ash accounted for 88 to 92 percent of the chromium discharged and 82 to 89 percent of the
copper discharged. Further, these distributions were not affected by kiln temperature in the
range varied (i.e., 824° to 985°C [1,516° to 1,805°F]).
61
-------
TABLE 32. NORMALIZED TRACE METAL DISTRIBUTIONS
Test
Kiln exit temperature, °C
(°F)
Kiln exit O2, %
1
(3/15/91)
824
(1,516)
11.2
2
(3/19/91)
984
(1,803)
9.0
3b
(3/21/91)
985
(1,805)
10.0
Distribution, % of metal measured
Cadmium
Kiln ash
Scrubber exit flue gas
Scrubber liquor
Total
Chromium
Kiln ash
Scrubber exit flue gas
Scrubber liquor
Total
Copper
Kiln ash
Scrubber exit flue gas
Scrubber liquor
Total
Lead
Kiln ash
Scrubber exit flue gas
Scrubber liquor
Total
61
23
16
100
88
4
8
100
83
6
11
100
53
17
30
100
19
42
39
100
92
2
6
100
89
5
6
100
23
42
35
100
8
36
56
100
92
2
6
100
82
10
8
100
19
55
26
100
In contrast, cadmium and lead exhibited relatively volatile behavior. At the low-kiln-
temperature test (Test 1) conditions, the kiln ash accounted for 53 (lead) to 61 (cadmium)
percent of the metals discharged. Even at this relatively low incineration temperature, a
significant amount of each metal evidently vaporized in the kiln and was carried into the
afterburner and downstream to augment the amount entrained in flyash carried out of the kiln.
The extent of evident vaporization was enhanced at the high-temperature-test (Tests 2
and 3) condition. For these tests, the kiln ash accounted for significantly decreased fractions of
cadmium and lead discharged, 8 to 19 percent for cadmium and 19 to 23 percent for lead.
Scrubber exit flue gas and scrubber liquor fractions were higher than the kiln ash fractions at this
high-temperature condition. Indeed, 36 to 42 percent of the cadmium discharged and 42 to
55 percent of the lead discharged escaped the incineration system and the venturi/packed-
column scrubber.
62
-------
Table 33 summarizes the apparent scrubber collection efficiencies calculated for each
metal measured in the test program. In calculating apparent collection efficiency, it is assumed
that the flowrate of metals at the scrubber inlet can be obtained by summing the flows in the two
scrubber discharge streams: the scrubber exit flue gas and the scrubber liquor. In other words,
apparent scrubber collection efficiency is defined to be (scrubber liquor fraction)/(scrubber
liquor fraction plus scrubber exit flue gas fraction). The data in Table 33 show that, at the low-
kiln-temperature test condition, apparent collection efficiencies for chromium, copper, and lead
were comparable, and in the nominal 65 to 70 percent range. Collection efficiency for cadmium
was lower at 41 percent. At the high-kiln-temperature test condition, cadmium and chromium
efficiencies were relatively unchanged. Copper and lead collection efficiencies decreased.
52.4 Participate and HC1 Emissions Data
Table 34 gives the particulate levels measured at the scrubber exit. For the three tests,
flue gas particulate levels at the scrubber exit ranged from 70 to 101 mg/dscm (corrected to
7 percent O2). These levels were below the 180 mg/dscm (at 7 percent O2) hazardous waste
incinerator performance standard.
The sediments incinerated during this test program contained 0.85 percent chlorine.
Table 35 summarizes the levels of HC1 measured at the scrubber. Measured HC1 concentrations
at the scrubber 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 HC1 collection efficiencies ranged
from 98.8 to 99.9 percent of the chlorine fed.
Figure 11 illustrates the results of the particle size measurements performed at the
afterburner exit. In performing the particle size sampling, it was observed that the cascade
impactor sampling nozzle collected noticeable amounts of particulate on its inside wall. This
particulate adhered to the probe wall, and could not be recovered and weighed accurately. Thus,
it is believed that a significant amount of the flue gas particulate could have been left
unaccounted for in the afterburner exit particle sizing measurements.
The data in Figure 11 show that the afterburner exit size distributions were comparable
for Tests 2 and 3b, which were performed at comparable incineration conditions. The size
distribution for the low-temperature test (Test 1) was coarser. Nonetheless, even for Test 1,
more than 50 percent of the measured particulate matter at the afterburner exit was smaller than
0.6 \im. For Tests 2 and 3, more than 70 and 80 percent, respectively, of the captured
particulate was smaller than 0.6 jim. One explanation for the finer particles in Tests 2 and 3
could be the higher kiln temperature during these two tests, 985° C (1,805° F), compared to
824° C (1,516° F) during Test 1. This observation is consistent with the expectation that high
combustion temperature can favor the formation of fine particles via the vaporization and
condensation mechanism.
63
-------
TABLE 33. APPARENT SCRUBBER COLLECTION EFFICIENCIES
Test
Kiln exit temperature, °C
(°F)
Kiln exit O2, %
1
(3/15/91)
824
(1,516)
11.2
2 3b
(3/19/91) (3/21/91)
984 985
(1,803) (1,805)
9.0 10.0
Apparent scrubber collection
Cadmium
Chromium
Copper
Lead
41
69
66
64
efficiency, %
48 50
79 62
53 33
45 23
TABLE 34. FLUE GAS PARTICULATE LEVELS
Test
Scrubber exit flue
gas participate,
mg/dscm at 7% O2
Test 1 (3/15/91)
Test 2 (3/19/91)
Test 3b (3/21/91)
70
82
101
TABLE 35. FLUE GAS HC1 LEVELS
Parameter
Test 1
(3/15/91)
Test 2
(3/19/91)
Test 3b
(3/21/91)
Cl feedrate, g/hr
Scrubber exit
Flue gas HC1 concentration,
591
591
589
mg/dscm
ppm
Flue gas emission rate, g/hr
System collection efficiency, %
0.31
0.2
0.68
99.9
3.60
2.4
7.24
98.8
1.14
0.8
2.10
99.6
64
-------
99.98
g 93
CD
°-84
I
JO 70
3
O 5°
30
0.5
1.0
2.0 5.0
Particle size (microns)
Figure 11. Afterburner exit particle size distributions
10.0
20.0
53 CONCLUSIONS
Test conclusions are as follows:
• Greater than 99.9999 percent DRE of the PCBs in the site sediments can be
achieved at incineration temperatures of both 824° C (1,516° F) and 984° C
(1,803° F) in a rotary kiln with an afterburner operated at 1,208° C (2,206° F).
However, with a kiln solids residence time of 0.5 hr, the treated sediments (kiln
ash) are still PCB-contaminated. In tests with a PCB-spiked sediment feed, the
kiln ash discharge accounted for between 0.07 and 0.22 percent of the PCB fed in
the spiked feed regardless of kiln temperature. For a native (unspiked) sediment
feed, the kiln ash accounted for 0.32 to 0.91 percent of the PCBs introduced in the
sediment feed.
• The APCS discharge flue gas from the incineration of the sediments contained low
levels of PCDDs and higher levels of PCDFs, chiefly total TCDF, total PeCDF,
and total HxCDF. The 2,3,7,8-TCDD toxicity equivalent levels were in the
nominal 0.05 to 0.10 ng/dscm range. PCDD/PCDF emissions were not affected
by kiln temperature.
65
-------
• Of the contaminant trace metals, chromium and copper were relatively
nonvolatile. The kiln ash discharge accounted for nominally 80 to 90 percent of
the discharged amounts of these metals. These fractions were not affected by the
range of kiln temperatures tested.
• Of the contaminant trace metals, cadmium and lead exhibited relatively volatile
behavior, and increasingly so at the high kiln temperature (984° C [1,803° FJ).
The kiln ash discharge accounted for 53 percent of the lead and 61 percaftt of the
cadmium discharged at the low kiln temperature (824° C [1,516° F]). These
fractions decreased to the nominal 10 to 20 percent range for cadmium and the
20 percent range for lead, at the high kiln temperature. Scrubber exit flue gas
fractions (cadmium and lead) and scrubber liquor fractions (cadmium) increased
accordingly.
• Apparent scrubber collection efficiencies were in the nominal 65 to 70 percent
range for chromium, copper, and lead at the low kiln temperature, and lower, at
41 percent, for cadmium. Cadmium and chromium collection efficiencies were
apparently unaffected by increased kiln temperature, although copper and lead
collection efficiency decreased to the 33 to 53 percent range for copper, and the
23 to 45 percent range for lead.
• Neither treated sediments nor the scrubber liquor discharges would exhibit the TC
based on their cadmium, chromium, or lead concentrations.
No conclusions regarding the effects of variations in kiln excess air levels on organic
constituent destruction and metals distributions were possible, as little variation in kiln excess
air level could be operationally achieved.
^The test results suggest that incineration would be an effective treatment option for the
site sediments. However, sediment dewatering prior to incineration, or incinerating at higher
kiln solids residence times (perhaps up to 1 hr), might be required to yield a treated sediment
not contaminated by PCBs.
Test results were reported in the following report:
• Whitworth, W. E., and L. R. Waterland, "Pilot-Scale Incineration of PCB-
Contaminated Sediments from the New Bedford Harbor Superfund Site," draft
September 1991.
66
-------
SECTION 6
FATE OF TRACE METALS IN THE ROTARY KILN SYSTEM WITH A
CALVERT FLUX/FORCE CONDENSATION SCRUBBER
Risk assessments have suggested that trace metal emissions from some incinerators
treating waste streams with high levels of metals could pose unacceptable risks to human health
and the environment. Despite their importance, available field data on the fate of trace metals
from hazardous waste incinerators are limited. Data describing the effects of incinerator
operation and waste composition on trace metal fate are particularly lacking. Data to evaluate
the effectiveness of typical APCSs for collecting flue gas metals are also needed.
In response to these data needs, two extensive test programs were completed using the
RKS at the IRF in FY88 and FY89. Both test programs were performed to support regulations
development by EPA's OSW (S. Garg, R. Holloway, coordinators). These tests quantified the
distribution of metals among the discharge streams of the RKS, and identified the effects of kiln
temperature, afterburner temperature, and waste feed chlorine content on* these discharge
distributions.
For both test programs, kiln temperature was varied from nominally 816 to 927° C (1,500
to 1,700°F); afterburner temperature was varied from nominally 982 to 1,204°C (1,800 to
2,200° F); and waste feed chlorine content was varied from 0 to nominally 8 percent. The main
difference between the two programs was the primary APCS used. A venturi/packed-column
scrubber was used in the first test program (FY88); a single-stage ionizing wet scrubber was used
in the second test program (FY89).
The feed for both test programs consisted of a synthetic waste containing organics
premixed with a clay absorbent material. Toluene, tetrachloroethylene, and chlorobenzene were
used as the organic compounds, and the waste feed chlorine content was altered by adjusting
their ratio in the mixture. The synthetic waste was continuously fed to the kiln by a screw
feeder. The test metals were combined in an aqueous solution and spiked onto the solid
material as it was fed to the kiln. Five hazardous constituent metals were used: arsenic, barium,
cadmium, chromium, and lead. In addition, the nonhazardous constituent trace elements
bismuth, copper, magnesium, and strontium were also incorporated into the test feed for
comparison and to provide data to evaluate a predictive numerical model under development
in another effort within RREL.
The results of the two completed test programs have shown the following:
67
-------
• Cadmium and bismuth are relatively volatile, averaging less than 40 percent
discharged to the kiln ash. The other metals are relatively nonvolatile, averaging
greater than 75 percent discharged to the kiln ash. Lead was the only exception,
exhibiting volatile behavior in the first test program and relatively nonvolatile
behavior in the second.
• Increased kiln temperature in the presence of feed chlorine caused increased
volatility of bismuth, cadmium, and lead. Data were not obtained to evaluate the
effects of kiln temperature in the absence of feed chlorine.
• Afterburner temperature did not affect metal partitioning among the scrubber exit
flue gas and scrubber liquor discharge streams. There was no conclusive evidence
indicating an effect on scrubber collection efficiency for metals.
• The effect of the waste feed chlorine content on the partitioning of metals was not
consistent between the two test programs. Data from the first test program
indicated increased volatility with increased chlorine for copper and lead. Data
from the second test program showed that waste feed chlorine content did not
affect metal discharge distributions within limits of data variability established by
replicate test conditions.
• Further sample analysis has shown that Method 3050 may not be sufficiently
aggressive to fully liberate the metals from the clay feed and kiln ash. Metal
recoveries achieved with sample preparation by fusion methods were about twice
the recoveries achieved by Method 3050.
• Relative metal volatilities agreed with expectations based on metal volatility
temperatures with the exception of arsenic, which was much less volatile than
expected for feeding arsenic as As2O3. Dissolving As2O3 in solution results in
AsO3" anion, which could be much less volatile than As2O3 or perhaps adsorbed
onto the clay matrix. The other spiked metals were added to the feed as soluble
nitrates dissolved as soluble cations.
• The four nonhazardous constituent metals behaved similarly to the hazardous
constituent metals; no unique information was obtained.
• Flue gas sampling using a Method 0030 volatile organic sampling train (VOST)
provided no significant information on PIC formation.
The IRF trace metal research program continued in FY91 with the completion of a third
parametric test program. For the FY91 program, a Calvert Flux Force/Condensation Scrubber
pilot plant was installed as the primary APCS. Mercury was added for a total of 10 test metals.
In addition, the particulate load and particle size distribution were measured after the quench
and at the scrubber exit in an attempt to determine fractional particulate collection efficiencies.
Scrubber pressure drop was also added as a test variable to investigate its effect on fractional
particulate collection efficiencies by particle size range.
68
-------
As in the past trace metal test programs, the synthetic waste was a toluene-based organic
liquid mixture added to a clay absorbent material. The chlorine content of the synthetic waste
was varied by changing the relative amounts of toluene, chlorobenzene, and tetrachloroethylene.
The test metals were premixed in an aqueous solution and spiked onto this solid material during
feeding to the kiln.
The test variables for the FY91 tests were kiln temperature, waste feed chlorine content,
and scrubber pressure drop. Afterburner temperature (a test variable in past programs) was not
a test variable for these tests and was held constant. The concern over metal emissions from
incinerators has created an interest in determining whether operating the primary combustion
chamber at reduced temperature can cause greater metal retention in the kiln ash by reducing
metal volatilization and entrainment. To study this mode of operation, three low-kiln-
temperature tests were specified. Unlike the two previous trace metals test programs, three tests
were performed with the kiln temperature varied with no chlorine in the waste feed. These tests
were designed to provide data on the effects of kiln temperature in the absence of chlorine.
Test data collected included feed material composition, incinerator process variables,
and discharge stream analysis results. The sampling and analysis protocol was specified to track
the distributions of metals among the RKS discharge streams (incinerator ash, scrubber liquor,
and flue gas). Sampling and analysis for volatile organics in the flue gas was also performed to
demonstrate that the RKS is operated at conditions suitable for organic waste destruction.
In summary, a 6-week test program was performed, aimed at identifying:
• The distribution of metals among kiln ash, scrubber liquor, and flue gas discharge
streams
• The effects of kiln temperature and waste feed chlorine content on metal fate
• The efficiency of the Calvert scrubber for collecting flue gas metals and particulate
by particle size range
• The effects of scrubber pressure drop on metal collection efficiencies and
particulate collection efficiencies by particle size range
6.1
TEST PROGRAM
This test program was also conducted in the IRF RKS. For the program, the RKS
APCS was modified by installing a Calvert Flux Force/Condensation Scrubber pilot plant as the
primary APCS. The skid-mounted components of the system supplied by Calvert were the
condenser/absorber, Calvert Collision Scrubber, entrainment separator, wet electrostatic
precipitator, caustic tank and injection pump, and induced draft (ID) fan, as shown in the
scrubber process schematic in Figure 12. The flue gas quench, heat exchanger, and secondary
APCS currently in place at the IRF were used.
The RKS quench at the IRF was modified to create a two-stage quench. Provision for
injecting fresh water in the first stage and recirculated water in the second stage was installed.
This arrangement was designed minimize the addition of fine particles to the flue gas by reducing
69
-------
I
I
3
I
70
-------
the spray dryer effect, which occurs when recirculated quench liquor containing dissolved solids
is used to quench hot flue gas. In the normal RKS quench configuration, makeup water is added
to the recirculating liquor storage tank for the quench and scrubber systems. For these tests,
however, the quench system plumbing was modified so that most of the makeup water could be
added as the fresh water supply to the first stage of the quench.
After quenching, the flue gas was directed to the Calvert scrubber. Flue gas exiting the
Calvert scrubber ID fan was directed to the secondary APCS before being discharged to the
atmosphere.
6.1.1 Synthetic Waste Mixture
These tests were conducted with the same base feed material used in the previous trace
metal test programs. The clay absorbent was combined with an organic POHC mixture in the
ratio of 0.4 kg organic liquid to 1 kg clay absorbent. This ratio produced a free flowing solid,
similar to the unspiked clay absorbent. For these tests, the chlorine concentrations were
nominally 0,1, and 4 percent of the combined feed. Desired feed chlorine content was achieved
by adding chlorobenzene and tetrachloroethylene to toluene. The organic liquids were combined
as noted in Table 36, then mixed with the clay absorbent in a portable cement mixer. The feed
mixture was stored in a 55-gal drum with lid until needed. A twin-auger screw feeder was used
to continuously feed the mixture to the kiln during testing.
As in the previous test programs, the trace metals were fed in an aqueous solution. The
test metals were arsenic, barium, bismuth, cadmium, chromium, copper, lead, magnesium,
mercury, and strontium. Previous analyses showed that chromium and magnesium were present
in the clay matrix at approximately 53 ppm and 2.4 percent, respectively. These concentrations
were high enough to provide sufficient feed quantities of these metals that they were not
included in the aqueous metals spike solution. Table 37 notes the specific metal compound
chosen for spiking and the aqueous spike solution concentrations used.
The aqueous trace metals solution was prepared in 5-gal glass containers, from which
it was continuously metered into the solid matrix at the screw feeder. The spike solution was
added at a nominal flowrate of 2 L/hr using a separately controlled gear pump. Inputs of the
solid matrices and the metals solution were carefully monitored and recorded using scales.
Previous analyses showed that barium, cadmium, copper, and strontium were present
in the clay absorbent at the concentrations given in Table 37. These background concentrations
were included in the calculations to determine the total metal concentrations in the integrated
waste feed, shown in the last-column of Table 37. Chromium and magnesium concentrations in
the integrated feed mixture were approximately 38 ppm and 1.7 percent, respectively.
6.1.2 Test Conditions
As noted above, the test variables were the kiln exit temperature, the waste feed chlorine
content, and the scrubber pressure drop. Each was varied over three target levels, as specified
in Table 38. The test program consisted of 11 test points. Kiln temperatures were nominally
538°, 816° and 927°C (1,000°, 1,500° and 1,700°F). Waste feed chlorine content was
nominally 0, 1, and 4 percent. The scrubber pressure drop for Tests 1 through 9 was held
71
-------
Ol
fc
0
H
So
1
o
U
1
S
a
1
a
1
0
EH
1
NO
CO
S
1
«r!
a"
5 «
frQ *-_-*
« •*?
*3a
fiff
1 iS
fil ^ §1
•3
0
S
<1)
1
f
1*
o
U
0
w
o
S
e
1
2
5
g
O)
^
1
1
^3
S
V
1
1
2
2
.a
S
X~>
^
0
06"
^~*>
rt-
c4
•*
o
oo
o\
o
o
S
*— I
1
n
oo
o
T— 1
0
o
T-4
o
0
o\
oq_
^~4
vq
^
tN
vq
T-5
co
^
o
o
o
T"~t
o
o
s^
^i.
r—(
^.
CN
"*
O
00
cq
rH
O\
O
o
0
o
T— 1
G,
I
OH
rH
S
3
•S
S
o
O
VO
K
^
oo
C3
in
CO
00
-------
CO
0
S
e!
H
CJ
fc
o
o
3
S
CO
1
t>
3
2§
£
1
g
?
fcl
••
•^
*m
5
4
^
£M
§
1
S
1
&
CO
B
33
„««*_,
3 es fa B
IB •*•* **
S g^
1 8 *
3
e"
?-S 2 g
Ilia
A 3
CQ M
3
•si
5 1| a.
2 -w ^
8 S M
o §
3
•o
e
f
|
"« "«* J
2 "S "S35
1
3
i
m o» 01 r Qj HH
^H ^ ^ ^^^
s
T-I t- O\ VO
T-H oi cj T-5 cs
T— 1
B
V
5
to
c3 —
2-2 S -
S ,« o e 3 ^
"S? § .2 I -3 §
1 p 1 I ^ s ^
« g <; pq (j i_) i^;
l*"^ t^
o »o
,O OI
vo u->
V M
°. '"i
CO Ox
w "*'
q^ OM
3? w
IO ^
CO ^
II
S u
t> r-
oi oi
T— 1 T-t
Other Metals
Bismuth
Copper
o
Tt
O
0)
r-3
C^l
N
CO
t-.
oi
T-H
Strontium
|
3
0
«
a,
C/}
1—
l-l
a o
•2 S
_3 ^~^
O J~
W3 *™^
••-» ^
2 in
. cd
"1 "^
«J <*-!
13 c
•a a)
C -Q
°^ o
t« on
73 jD
C rt
H
8^
is
0} O
U^ ^^
JJ "^B
.S ^
BC3 \o
So
w
O *-*
en 2
/*^ F^
z <8 *•*
52^"
"Add sufficien
bSynthetic was
feedrate of 2
73
-------
TABLE 38. TARGET TEST CONDITIONS
Test No.
1
2
3
4
5
6
7
8"
9
10
11
Kiln exit
temperature,
°C (°F)
538 (1,000)
816 (1,500)
927 (1,700)
538 (1,000)
816 (1,500)
927 (1,700)
538 (1,000)
816 (1,500)
927 (1,700)
816 (1,500)
816 (1,500)
Feed mixture
chlorine content,
%
0
0
0
1
1
1
4
4
4
4
4
Calvert scrubber
pressure drop,
in WC
50
50
50
50
50
50
50
50
50
35
70
"Two baseline tests were performed at test condition 8 with the clay/
organic liquid mixture, but without the aqueous metals spike solution.
constant at nominally 50 in WC. Test points 10 and 11 were at the same conditions as test
point 8, but with scrubber pressure drops of nominally 35 and 70 in WC, respectively.
Two baseline tests were also performed. These tests were designed to clarify if there
was a hysteresis effect caused by test-to-test carryover of metals. The clay/organic liquid mixture
was used, but the aqueous metals spike solution was not included. One test was conducted
before the test series began to establish baseline conditions for metals present in the incinerator
system. The second baseline test was performed after the test series was completed. The two
baseline tests were performed at test condition 8.
Table 39 summarizes the RKS operating conditions that were held constant. All tests
were performed at the same nominal afterburner exit flue gas O2 (7.5 percent), afterburner exit
temperature (1,093°C [2,000° F]), and synthetic waste feedrate (63.5 kg/hr [140 Ib/hr] of which
18 kg/hr [40 Ib/hr] was the organic liquid matrix). For all tests, the kiln rotation rate was held
constant to provide a solids residence time of nominally 1 hr.
74
-------
TABLE 39. TEST CONDITIONS HELD CONSTANT
Clay/organic liquid feedrate
Afterburner temperature
Afterburner exit flue gas O2
Kiln solids residence time
Quench blowdown rate
63.5 kg/hr (140 Ib/hr)
1,093° C (2,000° F)
7.5 percent
Ihr
Minimum operable
6.1 J Sampling and Analysis
Figure 13 identifies the sampling point locations. For the 11 tests and the two baseline
tests, the sampling effort included the following:
• Obtaining a composite sample of the clay/organic liquid mixture
• Obtaining two composite samples of the aqueous metals spike solution (except
baseline tests)
• Obtaining two composite samples of the kiln ash
• Obtaining a scrubber liquor composite sample
• Sampling the flue gas at the quench and scrubber system exits for particulate, HC1,
and trace metals (excluding mercury), using a Method 5 train modified for
multiple metals capture
• Sampling the flue gas at the quench and scrubber exits for mercury using a
Method 101A train
• Determining the particle size distribution of flue gas particulate in the quench and
scrubber exits using an Anderson cascade impactor train
• Obtaining samples of the flue gas at the scrubber exit using a Method 0030
sampling train
• Sampling the flue gas at the stack for HC1 and particulate using a Method 5 train
• Continuously measuring O2 concentrations at the kiln exit; O2, CO, CO2, and
TUHC (unheated) at the afterburner exit; O2, CO2, and NOX at the scrubber exit;
and O2, CO, and TUHC (unheated) at the stack
75
-------
,
1 CO
^JL
f Ty \
1 £< )
v /
x^ ./
T
<£
Q-. i
\\\ *
u.
1
f
^^ ^^
IE
o
4
i r-
i- °-
i
°co
1
1
•x. z
oo
LU o
—^ 111
Oco
CD
I
1 ^
i CC
DC LU
LU Z
"^ m
f
1 «••••
•z.
CM
k
1
lfl|
0
«|B
Is
M
S>
SJ
E
£
2
i
(A
=
i
3
* e
D
•S S «
cs ^3 v »3
« 8.3.1:
<3 .§ 1 .a
111
•B r 1
S §L«
^
10 n. 2
J~ .2 3
- I |-|
S £ .1
•a
1 U
o"
8
o
u
6C
2i
IS,
a
C3
CA
>> V
1 H
1 a
"I* i .§)
0) ^
*M "O
« B
O eg
2" ao
s 1
ro S
Nil
II
n u
«
•2 J2
qj CA
||
«M B
,e
-------
In addition, three composite samples of the feed clay absorbent material were collected for
metals analysis to confirm earlier data on background metal concentrations. Sampling for
volatile organics by Method 0030 (VOST) was performed at the stack during Test 8 and the two
baseline tests to meet trial burn requirements.
Flue gas sampling train samples were analyzed for their method specific analytes, e.g.,
metals train samples for the test program trace metals, and Method 0030 samples for the test
program volatile POHCs. Synthetic waste feed mixtures and kiln ash samples were analyzed for
test program volatile POHCs. Feed clay absorbent, aqueous metal spike solution, kiln ash, and
scrubber liquor samples were analyzed for the ten test program trace metals. The ASTM lithium
tetraborate fusion method was used to digest solid samples for analysis. In addition, kiln ash
TCLP leachates were prepared and analyzed for the test program trace metals.
6.2
TEST RESULTS
The 13-test program was conducted from late May through mid-July 1991. Table 40
summarizes the actual kiln temperatures achieved for each test and compares them to test target
levels. In general, average achieved kiln exit gas temperatures were within about 17° C (30° F)
of test target temperatures for the tests. Table 41 shows the cumulative amount of synthetic
waste fed for each test and the amount of kiln ash collected. As shown, between 72 and
88 percent of the synthetic waste clay (ash) fraction was collected as kiln ash. The remaining
12 to 28 percent of the synthetic waste clay fraction was ostensibly carried out of the kiln as
entrained particulate in the combustion flue gas. These fractions are comparable to past IRF
experience in feeding this synthetic waste mixture.
Sample analyses and test data reduction were underway at the close of FY91 and are
continuing. Test results will be assembled and reported in FY92.
77
-------
TABLE 40. ACTUAL VERSUS TARGET OPERATING CONDITIONS FOR THE CALVERT
SCRUBBER TRACE METAL TESTS
Kiln temperature, °C (°F)
Test
Target
Minimum Maximum
Average
Baseline (5/29/91)
1
2
3
4
5
6
7
8
9
10
11
Repeat
baseline
(6/5/91)
(6/6/91)
(6/13/91)
(6/18/91)
(6/19/91)
(6/21/91)
(6/25/91)
(6/28/91)
(7/9/91)
(7/11/91)
(7/16/91)
(7/18/91)
816 (1,500)
538 (1,000)
816 (1,500)
927 (1,700)
538 (1,000)
816 (1,500)
927 (1,700)
538 (1,000)
816 (1,500)
927 (1,700)
816 (1,500)
816 (1,500)
816 (1,500)
751 (1,384)
469 (876)
741 (1,366)
811 (1,491)
501 (934)
780 (1,436)
887 (1,629)
441 (825)
803 (1,477)
912 (1,673)
806 (1,483)
751 (1,383)
781 (1,437)
912 (1,673)
594 (1,102)
856 (1,573)
954 (1,750)
581 (1,077)
902 (1,656)
968 (1,775)
568 (1,055)
834 (1,533)
981 (1,798)
842 (1,547)
844 (1,552)
917 (1,682)
831 (1,528)
541 (1,006)
819 (1,507)
909 (1,669)
555 (1,031)
842 (1,547)
919 (1,686)
543 (1,010)
817 (1,502)
944 (1,731)
829 (1,524)
827 (1,521)
834 (1,534)
78
-------
TABLE 41. SYNTHETIC WASTE FED AND ASH COLLECTED
Synthetic waste fed, kg (Ib)
Test
Baseline
1
2
3
4
5
6
7
8
9
10
11
Repeat
baseline
(5/29/91)
(6/5/91)
(6/6/91)
(6/13/91)
(6/18/91)
(6/19/91)
(6/21/91)
(6/25/91)
(6/28/91)
(7/9/91)
(7/11/91)
(7/16/91)
(7/18/91)
Clay/organic
liquid
313 (691)
279 (614)
271 (597)
331 (730)
299 (660)
255 (563)
305 (674)
269 (593)
281 (619)
264 (582)
284 (627)
298 (657)
343 (756)
Clay fraction
247 (545)
221 (487)
216 (475)
266 (587)
249 (550)
206 (453)
240 (528)
203 (448)
214 (471)
210 (463)
236(519)
230 (507)
265 (585)
Kiln ash collected
Weight, Fraction of
kg (Ib) clay fed, %
219 (482)
176 (388)
166 (365)
202 (446)
179 (396)
153 (338)
186 (409)
169 (373)
166 (367)
154 (340)
178 (392)
181 (398)
204 (450)
88
80
77
76
72
75
77
83
78
73
75
78
77
79
-------
SECTION 7
INCINERATION OF ARSENIC-CONTAMINATED SOILS
FROM THE CHEMICAL INSECTICIDE CORPORATION SUPERFUND SITE
The Chemical Insecticide Corporation (CIC) site in Edison Township, New Jersey, was
formerly used to manufacture and formulate pesticide products. The results of the remedial
investigation and feasibility study (RI/FS) show that the soils at the site are highly contaminated
by organic pesticides and arsenic. Dioxin (i.e., 2,3,7,8-tetrachlorodibenzo-p-dioxin) has also been
found in some soil samples collected during the RI/FS at concentrations up to 1.8 ng/kg (ppb).
Thermal treatment has previously been demonstrated to be an effective means of destroying
organic pesticides, dioxin, and other organic compounds. The finding of high concentrations of
arsenic in the soils at the CIC site has raised the question of whether a thermal treatment unit
treating soil from the site, and operating under conditions capable of attaining a 99.9999-percent
DRE for dioxin and other organic contaminants, can also reduce arsenic concentrations to
acceptable levels in the stack emissions. To address this question, EPA Region 2 (J. Josephs,
coordinator) requested that incineration testing be performed at the IRF.
7.1
TEST PROGRAM
The objective of the test program performed was to obtain data to support feasibility
study (FS) efforts in evaluating incineration as a possible remedial alternative. To attain this
objective, the test program was designed to:
• Maintain critical operation parameters in the effective range, ensuring a dioxin
DRE of 99.9999 percent and minimizing arsenic air emissions
• Determine whether the incinerator can attain a 99.96-percent removal efficiency
(RE) for arsenic, where RE is defined as:
RE = 100 (1 - [flue gas arsenic emission rate]/[arsenic in soil feedrate])
• Determine the characteristics and arsenic content (at a minimum) of all effluent
streams resulting from the thermal treatment process based on the test equipment
employed
• Serve as a model for determining important operating parameters to be used for
projecting comparable full-scale performance and operational costs
All tests in this program were performed in the RKS at the IRF.
80
-------
The high-efficiency scrubber system for this program was the Calvert Flux
Force/Condensation Scrubber System used in the trace metal fate test program described in
Section 6.
7.1.1 Test Waste Description
The CIC site is located in Edison Township, Middlesex County, New Jersey. From 1958
through 1970, a variety of pesticides was formulated and distributed for both commercial and
military applications by CIC. The CIC product list included a wide range of insecticides,
fungicides, rodenticides, and herbicides. One specific product was 2,4,5-trichlorophenoxyacetic
acid (2,4,5-T), which might have contained tetrachlorodibenzo-p-dioxin (TCDD) as a
contaminant.
Pesticide manufacturing activities, combined with process-water storage lagoons and
poor housekeeping, led to the widespread chemical contamination of this site. As part of the
RI/FS, approximately 170 surface/shallow subsurface and 200 soil boring samples were collected
and analyzed. Analytical results of these samples indicate that the major contamination
constituents at the CIC site are organochlorine pesticides, chlorinated herbicides, and hazardous
constituent trace metals.
Four drums of soil were excavated from the site in February 1991 for use in this test
program. Characterization samples from each drum were forwarded to the IRF for pretest
analyses. Each sample was analyzed for the eight TCLP trace metals, antimony, beryllium, and
thallium; and for the organochlorine pesticides and chlorinated herbicides known to be site
contaminants. In addition, TCLP leachates of each sample were prepared and analyzed for trace
metals and organochlorine pesticides.
Soil sample analysis results are summarized in Table 42. TCLP leachate analysis results
are summarized in Table 43. The data in Table 42 show that soil with an average arsenic
content of about 900 mg/kg was excavated for testing. This soil was also contaminated with an
average of 2 mg/kg of p,p'-DDD, 3 mg/kg of p,p'-DDE, 26 mg/kg of p,p'-DDT, and 9 mg/kg
of chlordane. Despite having arsenic, barium, chromium, and lead contamination levels of
several tens to over 1,000 mg/kg, the soils were not toxicity characteristic (TC) hazardous wastes
based on the TCLP leachate analysis results summarized in Table 43.
A composite of the four-characterization samples received was also subjected to
proximate (ash, moisture, and heating value) analysis, with the following results:
• Ash content:
• Moisture content:
• Heating value:
• pH:
87 percent
13 percent
Will not burn
6.1
81
-------
TABLE 42. SOIL CHARACTERIZATION SAMPLE ANALYSIS RESULTS
Concentration, mg/kg
Constituent
Trace metals:
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Thallium
Drum 1
12
1,260
63
0.25
2.3
17
55
7.8
<4.4
<0.33
164
Drum 2
10
771
52
0.25
1.5
17
80
8.9
<4.7
<0.35
185
DrumS
24
875
59
0.35
2.1
18
104
10.5
<4.5
<0.34
145
Drum 4
33
784
60
0.35
1.9
18
103
8.6
<6.6
<0.45
156
Average
20
922
59
0.33
2.0
18
86
9.0
<6
<0.5
163
Organochlorine pesticides:
«-BHC
y-BHC
p,p'-DDD
p.p'-DDE
p.p'-DDT
Chlordane
<0.8
<0.8
3.0
4.3
33
11
<0.8
<0.8
1.9
2.8
32
<8
<0.8
<0.8
1.5
2.3
21
<8
<0.8
<0.8
1.8
3.1
19
<8
<0.8
<0.8
2.1
3.1
26
9
Chlorinated herbicides:
2,4-D
2,4,5-T
Silvex
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
82
-------
TABLE 43. SOIL CHARACTERIZATION SAMPLE TCLP LEACHATE ANALYSIS RESULTS
Concentration, mg/L
Constituent
Trace metals:
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
Thallium
Drum 1
0.175
3.0
0.92
0.002
0.015'
0.051
0.062
< 0.002
<0.08
<0.01
0.116
Drum 2
0.058
2.1
0.48
0.002
0.014
0.078
0.052
< 0.002
<0.08
<0.01
0.018
Drum 3
0.077
1.5
0.03
<0.001
0.012
0.072
0.051
< 0.002
<0.08
<0.01
0.076
Drum 4
0.113
1.2
0.08
< 0.001
0.014
0.216
0.051
< 0.002
<0.08
<0.01
0.029
TC
regulatory
level,
mg/L
—
5.0
100
—
1.0
5.0
5.0
0.2
<1.0
^<5.0
—
Organochlorine pesticides:
«-BHC
Y-BHC
p,p'-DDD
p,p'-DDE
p,p'-DDT
Chlordane
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.010
<0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.0 10
<0.001
< 0.001
< 0.001
< 0.001
< 0.0025
<0.010
<0.001
< 0.001
< 0.001
< 0.001
< 0.001
-------
7.1.2 Test Conditions
The test program performed to address the objectives of the project consisted of four
tests. Each test entailed incinerating one 55-gal drum of contaminated soil over a 4- to 5-hr time
period. As noted above, a primary test objective was to confirm the ability of a rotary kiln
incineration system equipped with a high-efficiency APCS to achieve 99.96 percent arsenic RE.
Preliminary results from the first test showed that arsenic RE was just less than 99.96 percent,
so a fourth test was performed in which the soil was mixed with lime, at a blend ratio of 0.5 Ib
lime/10 Ib soil.
For each test, the soil was fed to the kiln via the fiberpack drum ram feed system at a
rate of 12 fiberpack drums/hr (1 fiberpack drum every 5 min). Each fiberpack drum held
approximately 4.6 kg (10 Ib) of soil. Therefore, the soil feedrate was nominally 55 kg/hr
(120 Ib/hr).
All tests were performed at a kiln gas temperature of nominally 980° C (1,800°F) and
an afterburner gas temperature of nominally 1,200° C (2,200° F). These conditions are
comparable to those that have resulted in 99.9999 percent dioxin and PCB DRE in past tests.
Thus, these conditions are anticipated to meet or exceed the requirements for destroying dioxin
to levels greater than 99.9999 percent. In addition, these conditions are typical of full-scale
incinerator operation.
For all tests, the kiln rotation rate was at a value corresponding to a kiln solids residence
time of 0.5 hr. The Calvert scrubber system pressure drop was nominally 12 kPa (50 in WC),
and scrubber system blowdown was held to the minimum operable level.
7.13 Sampling and Analysis
The sampling and analysis procedures employed for all tests are outlined in Figure 14.
For all tests, the sampling protocol entailed:
• Obtaining a composite sample of the soil samples from each drum before the soil
was packaged into the fiberpack containers
• 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, NOX, and TUHC levels at the scrubber exit; and O2, CO, and CO2
levels at the stack
• Sampling flue gas at the scrubber system exit for semivolatile organics, arsenic,
and particulate and HC1
• Sampling at the stack downstream of the secondary APCS for arsenic and
particulate and HC1
84
-------
i-6-eso asa
£S
*§
I
CARBON
BED
f
CALVERT
SCRUBBER
f
QUENCH
SECTION
t
1
AFTER-
BURNER
t-
z
S
4
CO
S
5 «3
so
•g 8
"i-S §
£ -2 s
3 B^
OJ
•s
1/1 "
g)
!i
is.
t/3-
X X
'X X
X
X X
X X
VO
85
-------
The composite soil sample for each test was obtained by taking thief samples from each
shipment drum at three locations in the drum cross section just prior to packaging the soil into
the fiberpack drums. These three samples were combined to form one composite waste feed
sample per test.
During each test, kiln ash was continuously removed from the kiln ash pit via a transfer
auger, and deposited into an initially clean 55-gal drum. After all test ash was deposited in this
drum, representative kiln ash samples were taken by thief sampling in at least three locations
across the collection drum cross section. These three ash samples were then combined to form
one composite sample.
As noted above, each test was run with the scrubber liquor loop operating at minimum
blowdown. At the end of each test, the incinerator was operated (firing natural gas) for at least
12 hr after stoppage of soil feed, or until the kiln was visually clear of ash, whichever was longer.
After this period of time, the scrubber system was drained to the blowdown collection tank.
During the entire period the system was being drained, a sample was continuously taken from
a tap.
The soil composite sample for each test, and each of the kiln ash and scrubber liquor
samples, were subjected to analysis (still in progress) for the TCLP trace metals (arsenic, barium,
cadmium, lead, mercury, selenium, and silver). The soil and kiln ash from each test were
subjected to TCLP extraction. The resulting extracts were subjected to analysis (still in progress)
for the eight TCLP trace metals. In addition, the soil leachates were subjected to analysis (still
in progress) for the organochlorine pesticides. Flue gas sampling train samples were subjected
to analysis (still in progress) for these procedure-specific analytes, as noted in Figure 14.
7.2
TEST RESULTS
This test program was completed in early August 1991. Sample analysis and test data
reduction were underway at the close of FY91 and are currently in progress. Test data
evaluation and reporting will be completed in FY92.
86
-------
SECTION 8
FACILITY PHYSICAL PLANT IMPROVEMENTS
Several modifications and improvements to the IRF physical plant were completed
during FY91. These are briefly outlined in the following subsections.
8.1
OFFICE SPACE
Since FY85, the IRF operating staff has worked in some combination of single- and
double-width office trailers adjacent to the main facility building. After the completion of the
new, 15,200 ft2 main facility building in FY891, and completion of the major RKS reconfiguration
and upgrade effort in FY902, efforts focused on providing for more appropriate office space for
the IRF staff.
Plans to procure 5,040 ft2 of modular office space were finalized in late FY90. Approval
to purchase the planned complex was received in early FY91. The complex was subsequently
installed, and completed and occupied in March 1991.
8.2 INCINERATION SYSTEM IMPROVEMENTS
Most aspects of a major RKS reconfiguration and system upgrade were completed
during FY902. However, a few efforts extended into FY91, as discussed in the following
paragraphs.
The first effort involved replacing the kiln support and drive mechanism of the RKS.
It had been acknowledged for some time that the former performance of the rotary kiln drive
mechanism was not optimal. To improve system performance, the design of a mechanically
sturdy trunion and roller system with an improved drive mechanism was completed in early
FY91. The retrofitting of this system to the RKS took place from late December 1990 through
early January 1991, and was completed in time for the K088 waste incinerability test program
discussed in Section 3.
The second effort involved expanding the operating functions attended to by the
automatic process control system installed in early FY90. The first phase in the implementation
of automatic process control included developing the controls for the burner management system
and implementing extensive data acquisition functions. This phase was completed in June 1990,
and the controls were used during the POHC incinerability ranking tests discussed in Section 2.
87
-------
The next phase of automatic control included bringing remaining process parameter
sensors online; implementing the required automatic waste feed cutoff interlock functions
mandated by the IRF's modified hazardous waste management permit; enhancing the system's
data gathering and recording features; and refining the burner controls for automatic gradual
startup and shutdown. This effort was initiated in early FY91 and completed, with the controls
largely functional, in April 1991.
83 FLAMMABLE CHEMICAL STORAGE BUILDING
Following the recommendations that resulted from a facility environmental audit, an
adequate building for the IRF to store flammable chemicals in was procured and installed during
FY90. Several modifications to this building were required prior to its intended use. These were
completed, and the storage building commissioned for use in April 1991. With the addition of
the building, flammable liquids, previously stored outdoors under a tent and inside containment
traps, can now be stored in an appropriate enclosed space.
8.4
BENCH-SCALE THERMAL TREATABILITY TEST UNIT
The past two years have seen a significant amount of pilot-scale incinerability testing of
Superfund site wastes and contaminated soils conducted at the IRF. In FY91, it became clear
that there was a need to supplement the pilot-scale testing capabilities at the IRF with smaller,
bench-scale capabilities that could offer screening information at less expense. The specific need
for such capabilities arose out of the Superfund Technical Assistance Response Team (START)
program recently initiated within RREL.
The purpose of the START program is to assist EPA Regional Offices in determining
the best methods for remediating Superfund sites. In a typical project, a Regional Office sends
an analysis of the site waste, and related data, to START. Using this information, START
selects the possible remedial technologies applicable to the site. Once START identifies the
technologies, a screening process, called the Remedy Screen, is undertaken to select those
treatment technologies likely to be the most effective.
Remedy Screen testing, at present, is envisioned to be bench-scale, relatively inexpensive,
and capable of yielding proof-of-concept information. In the Remedy Screen process, the
Regional Office sends a quantity of the site waste material to the location designated to perform
the Remedy Screen testing. Should the results of the testing suggest that a given treatment
technology is an appropriate and likely remedy for the site, larger scale pilot-scale testing may
be warranted.
The IRF is the most logical location to conduct the thermal treatment Remedy Screen
testing, as the waste management skills and the pilot-scale testing capabilities are already in place
at the facility. Thus, in response to a START program request, the conceptual design of a
bench-scale thermal treatability test unit (TTU) was initiated in March 1991. The conceptual
design of the unit focused on the unit's being capable of:
• Simulating an incinerator operation in terms of solids retention time and mixing
• Providing for continuous feed of the waste to the combustion zone
88
-------
• Operating over a wide range of temperatures in order to simulate high- and low-
temperature thermal treatment
- Temperature range of 260° to 980° C (500° to 1,800° F)
— Gas fired to expose waste to a flame
• Generating sufficient gas volume to allow for a variety of gas-phase analyses to be
conducted, including analysis for volatile and semivolatile components and trace
metal components
• Allowing for collection of the ash residue stream for analysis
After examining several alternatives, modifying a small pathological incinerator of
standard design was selected as the most cost-effective approach to achieving the above
capabilities. Figure 15 illustrates the TTU unit resulting from this approach. The base
combustor system is a standard-design, National Incinerator, Inc., Model P-50 incinerator, with
the following modifications:
• The supplied burners were modified to accept modulating control valves and
controllers to permit accurate temperature control without the undesirable side
effects of on-off control
• Two load channels were attached to the sides of the main incineration chamber,
as shown in Figure 16. The channels allow for the insertion of ceramic sample
trays that are gradually fed through the combustion chamber by electrically driven
rollers. The feed rate (and the solids retention time in the combustion zone) is
adjusted by variable speed control of the rollers. Sample trays can be inserted one
"' after the other in order to make continuous incinerator simulation possible.
• The stack was equipped with sampling ports
Approval to proceed with procuring and installing the TTU was received in April. The
incinerator unit was ordered in May, and delivered in late July. Hookup and refractory curing
were completed in August. The material feed and removal system was installed in September.
Unit shakedown testing was underway at the end of FY91. Thus, the IRF will have bench-scale
TTU testing capability very early in FY92.
89
-------
Figure 15. TTU: external configuration.
90
-------
•a
I
91
-------
SECTION 9
HEALTH AND SAFETY, ENVIRONMENTAL COMPLIANCE,
AND PERMIT ADMINISTRATION
All aspects of the administration of the IRF's hazardous waste management permit were
routinely accomplished throughout the year. Special mention of the IRF's response to the new
training requirements for workers at hazardous waste management facilities, under
29 CFR 1910.120, is deserved. In response to these new regulations, an IRF staff training plan
was developed and implemented during FY91. All staff members were given the required level
of training over the course of the year, and required refresher training is routinely scheduled.
That all aspects of environmental compliance are properly administered is evidenced by
the results of the facility's biannual environmental audit by EPA's Safety, Health, and
Environmental Management Division (SHEMD). The results of this audit, which was conducted
in August 1991, noted that the facility's environmental compliance program is of exemplary
quality.
Two additional areas of environmental compliance and permitting that deserve
discussion are covered in the following subsections.
9.1 TOXIC SUBSTANCE CONTROL ACT RESEARCH AND DEVELOPMENT
PERMIT
During the initial planning for the incineration tests of the PCB-contaminated sediments
from the New Bedford Harbor Superfund site (discussed in Section 5) in early FY91, it became
clear that some form of PCB permit for the facility would be required under the Toxic
Substances Control Act (TSCA) regulations before the tests could proceed. After investigating
several approaches, it was decided that the most appropriate and expedient approach would be
to obtain a research and development (R&D) permit.
Accordingly, a TSCA R&D permit application was prepared and submitted to EPA's
Office of Toxic Substances in November 1990. The permit, in the form of an Approval to
Conduct R&D Tests to Dispose of PCBs, was received in late February 1991. With this annually
renewable permit in place, the IRF can now perform testing with both hazardous wastes
regulated under RCRA and PCB wastes regulated under TSCA.
92
-------
92 RCRA FACILITY INVESTIGATION
A modified hazardous waste management (RCRA Part B) permit for the facility, jointly
administered by the Arkansas Department of Pollution Control and Ecology (ADPCE) and EPA
Region 6, became effective in September 1990. The corrective action module of the permit,
included in accordance with regulations in response to the Hazardous and Solid Waste
Amendments of 1984 (HSWA), requires a RCRA Facility Investigation (RFI) to determine the
nature and extent of releases of hazardous wastes, including hazardous constituents and
hazardous substances from identified solid waste management units (SWMUs). The permit
outlines five Facility Investigation tasks.
The first task requires the completion of a "Preliminary Report: Description of Current
Conditions." This report was completed and submitted in December 1990. The second task
requires the completion of an RFI Work Plan describing the approach and procedures for the
initial phase of the RFI. This document was completed and submitted in March 1991.
93
-------
SECTION 10
THIRD-PARTY TESTING
^ The Federal Technology Transfer Act allows for the use of government facilities and
equipment in joint projects with private-sector third parties. The IRF represents a unique facility
with capabilities unavailable anywhere else in the United States. Furthermore, the hazardous
waste incineration research and testing arena is quite active. Thus, the potential demand for
such third-party joint projects is expected to become quite significant.
The RREL policy established during FY89 was to encourage, and even solicit third-
party use of the IRF. In fact, the IRF operations and research contract specifically provides for
this type of usage, and efforts to identify appropriate joint third-party projects have proceeded.
• CVnn this 6nd> a facility caPabmties brochure was prepared, printed, and distributed
in FY90. The purpose of the brochure is to outline the capabilities of the IRF and the type of
testing that has, and can, be performed, and to solicit third-party inquiries. This brochure was
reissued in color in FY91. FY90 also saw the preparation and release of a videotape highlighting
one of the test programs completed during that year. A solicitation-of-interest mailer was
distributed during FY91, and preview copies of the videotape distributed to mailer respondents
FY91:
Three firm proposals to perform test programs for third-party users were prepared
in
• A proposal to evaluate the incinerability of contaminated soils via a parametric
test program for a Department of Defense-administered Superfund site
• A proposal to evaluate the incinerability of wastes from a petroleum refinery
Superfund site being remediated by a private-sector potentially responsible party
(PRP)
• A proposal to evaluate kiln particle carryover from the incineration of surrogate'
mixed waste to support a Department of Energy incineration system design effort
All thee proposals were in active evaluation at the end of FY91, with the expectation that at least
one of the three will proceed.
94
-------
SECTION 11
EXTERNAL COMMUNICATIONS
During FY91, six reports were prepared and submitted, and 11 technical papers were
presented. These are listed in Table 44. This level of external communication and technology
transfer is comparable to levels experienced over the preceding 4 years, and testifies to the high
level of important research being supported at the IRF.
Table 45 lists some of the visitors to the IRF during FY91. The length of the list attests
to the visibility of the work being performed at the IRF to the incineration research community.
95
-------
Waterland, L. R., Operations and Research at the U.S. EPA Incineration Research
Facility, Annual Report for FY90," draft January 1991, revised February 1991
published as EPA/600/9-91/010, April 1991
Fournier, Jr., D. J., and L. R. Waterland, "The Fate of Trace Metals in a Rotary Kiln
incinerator with a Single-Stage Ionizing Wet Scrubber," revised February 1991, final
• Vocque, R. H., and L. R. Waterland, "Pilot-Scale Incineration of Contaminated Soil
from the Purity Oil Sales and McColl Superfund Sites," revised February 1991 final
May 1991 '
• Lee, J. W., W. E. Whitworth, and L. R. Waterland, "Pilot-Scale Evaluation of the
Thermal Stability POHC IncinerabiUty Ranking," draft June 1991 ,
• Whitworth, W. E., J. W. Lee, and L. R. Waterland, "Pilot-Scale Incineration Tests of
Spent Potlmers from the Primary Reduction of Aluminum (K088)," draft August 1991
• Whitworth, W. E., and L. R. Waterland, "Pilot-Scale Incineration of
PCB-Contaminated Sediments from the New Bedford Harbor Superfund Site "
preliminary draft September 1991
Papers and Presentations:
r, Jr., D. J., L. R. Waterland, and G. J. Carroll, "Size Distributions of Trace
Metals in Flue Gas Particulate from a Pilot-Scale Rotary Kiln Incinerator," presented
at the American Flame Research Committee 1990 Fall International Symposium on
NO Control, Waste Incineration, and Oxygen-Enriched Combustion, San Francisco
California, October 1990
Waterland, L. R., C. King, R. C. Thurnau, and M. K. Richards, "Incinerability Testing
of an Arsenic-Contaminated Superfund Site Soil," presented at the Pacific Northwest
International Section of the Air & Waste Management Association 1990 Conference
Portland, Oregon, November 1990 '
Waterland, L. R., D. J. Fournier, Jr., J. W. Lee, and G. J. Carroll, "Trace Metal Fate
in a Rotary Kiln Incinerator with an Ionizing Wet Scrubber," Waste Management
Vol. 11, p. 103, 1991 : 5 '
Waterland, L. R., C. King, M. K. Richards, and R. C. Thurnau, "Incineration
Treatment of Arsenic-Contaminated Soil," Remediation, p. 227, Spring 1991
Waterland, L. R., D. J. Fournier, Jr., J. W. Lee, G. J. Carroll, and R. C. Thurnau
ine Fate of Trace Metals in a Rotary Kiln Incinerator-Tests with Two Different
Scrubber Systems," presented at the Second International Congress on Toxic
Combustion Byproducts: Formation and Control, Salt Lake City, Utah, March 1991
submitted for publication to Combustion. Science, and Technology
~~ =
(continued)
96
-------
TABLE 44. (continued)
Papers and Presentations (continued):
• Wall, H. O., "The Incineration of Lead-Contaminated Soil Related to the
Comprehensive Environmental Response, Compensation and Liability Act
(CERCLA) (Superfund)," in Proceedings of the Seventeenth Annual RREL
Hazardous Waste Research Symposium, EPA/600/9-91/002, April 1991
• Fournier, D. J., Jr., L. R. Waterland, J. W. Lee, and G. J. Carroll, "The Behavior of
Trace Metals in Rotary Kiln Incineration: Results of Incineration Research Facility
Studies," in Proceedings of the Seventeenth Annual RREL Hazardous Waste
Research Symposium, EPA/600/9-91/002, April 1991
• Carroll, G. J., W. E. Whitworth, J. W. Lee, and L. R. Waterland, "Pilot-Scale
Evaluation of an Incinerability Ranking System for Hazardous Organic Compounds,"
presented at the Seventeenth Annual RREL Hazardous Waste Research Symposium,
Cincinnati, Ohio, April 1991
• Lee, J. W., D. J. Fournier, Jr., and R. C. Thurnau, "U.S. EPA Incineration Research
Facility Update," presented at the Seventeenth Annual RREL Hazardous Waste
Research Symposium, Cincinnati, Ohio, April 1991
• Waterland, L. R., C. King., R. H. Vocque, M. K. Richards, and H. O. Wall,
"Pilot-Scale Incinerability Evaluation of Arsenic- and Lead-Contaminated Soils from
Two Superfund Sites," presented at the 1991 Incineration Conference, Knoxville,
Tennessee, May 1991
• Lee, J. W., L. R. Waterland, W. E. Whitworth, and G. J. Carroll, "Evaluation of the
Thermal Stability POHC Incinerability Ranking in a Pilot-Scale Rotary Kiln
Incinerator," paper 91-34.3, presented at the 84th Annual Meeting of the Air &
Waste Management Association, Vancouver, British Columbia, Canada, June 1991
97
-------
TABLE 45. VISITORS TO THE IRF
Person
T. Bucci
T. Rice
G. Martin
D. Brown
A. Dorobati
G. McClure
R. Kennedy
J. Baudain
J. Calhoun
D. Pauley
B. Fontaine
P. Holland
R. McDuffee
J. Palmateer
F. Bales
R. Loftin
R. Wilkins
P. Murphy
A. Dorobati
H. Huppert
T. Wagner
R. Turner
J. Labrosa
C. Guffey
H. Hittner
D. Backfinch
A. Sykes
Affiliation
Pathology Associates, Inc.
ADPCE
ADPCE
ADPCE
ADPCE
University of Arkansas
Medical Sciences
Gentry & Associates
Clean Ventures
Pine Bluff Arsenal
Pine Bluff Fire
Department
Jefferson County Sheriffs
Department
White Hall Police
Department
ADPCE
Jefferson County Office
of Emergency Services
USAGE
USDA
Clean Ventures
ADPCE
ADPCE
SAIC
SAIC
EPA/RREL
EPA/OSW
Reynolds Metals
Alcoa
Noranda
Acurex
Date
10/15/90
10/31/90
11/9/90
11/16/90
11/27/90
12/6/90
12/6/90
12/13/90
12/19/90
1/3/91
1/11/91
1/15/91 '••'
through
1/18/91
1/15/91
1/16,17/91
1/18/91
1/23/91 .
through
1/25/91
Purpose of visit
Discuss incineration of
pathological waste
Facility familiarization tour
Facility familiarization tour
Facility tour
Facility tour
Facility tour
Permit-mandated facility
familiarization for emergency
planning
Witness Caldwell Trucking tests
Facility tour
Facility tour
Facility familiarization tour
Witness K088 tests
Witness K088 tests
Witness K088 tests
Witness K088 tests
Facility internal QA review
(continued)
98
-------
TABLE 45. (continued)
^M^^^BS— -"^^••"^^— l^^^-" 1
Person
J. Claysori
H. Wong
R. Schrock
R. Wilson
M. Billedean
J. Sudnick
L. Tate
K. Bisgard
B. Heineman
G. McGill
R. DeCesare
E. Cole
S. Arnold
R. Schluite
J. Camara
T. Wesson
B. Laswell
D. McCormick
K. Yates
E. Fleming
E. Lagoutte
J. Camara
S.-C. Yung
J. Huff
C. Westerfield
G. Cathije
T. Ho
H. Lee
Affiliation
El Dorado Engineering
JMM Associates
EPA/Region 3
JMM Associates
NCTR
Four Nines
USAGE
USAGE
EPA/Region 6
Texas Water Commission
U.S. Bureau of Mines
U.S. Bureau of Mines
U.S. Bureau of Mines
U.S. Bureau of Mines
Calvert Environmental
Calvert Environmental
ICF Kaiser Engineers
IGF Kaiser Engineers
Acurex
ADPCE
Rhone Poulenc
Calvert Environmental
Calvert Environmental
Calvert Environmental
Mobay Chemical
3M
Lamar University
Lamar University
Date
1/29/91
, * i
through
2/7/91
1/30,31/91
1/31/91
2/7/91
3/18/91
3/21/91
4/10/91
4/18/91
4/29/91
through
5/3/91
5/1/91
5/9/91
5/15/91
5/17/91
5/20/91
5/30/91
6/4/91
- r
Purpose of visit
Witness Drake Chemical tests
Witness Drake Chemical tests
Witness Drake Chemical tests
Witness Drake Chemical tests
Facility tour
Facility tour, discuss potential
Superfund site test program
• . - " '
Facility tour, discuss potential
Superfund site test program
Facility tour
Install Calvert scrubber
Facility tour
Facility internal safety review
Facility familiarization tour
Facility tour
Witness Calvert scrubber
shakedown
Witness Calvert scrubber trace
metal tests
Facility tour, trace metal research
discussion '
(continued)
99
-------
TABLE 45. (continued)
Person
B. Russell
D. Banks
R. Fish
D. Brown
M. Greet
G. Thompson
D. Brown
K. Modearis
T. Yoder
P. Weggel
K. Crabtree
K. Roser
R. Shipmer
Affiliation
Calvert Environmental
Calvert Environmental
Garver and Garver
Garver and Garver
Pine Bluff Arsenal
Pine Bluff Arsenal
ADPCE
Booz, Allen, Hamilton
Booz, Allen, Hamilton
EPA/SHEMD
EPA/PMD
EG&G
CE O&MS
Date
6/27/91
6/28/91
8/19/91
through
8/21/91
9/3/91
through
9/5/91
9/30/91
Purpose of visit
Witness Calvert scrubber trace
metal tests
Annual hazardous waste inspection
Biennial facility environmental
audit
EPA property audit
Facility tour
100
-------
SECTION 12
PLANNED EFFORTS FOR FY92
One test program was completed in the second quarter of FY91, the Drake Chemical
site tests discussed in Section 4, for which reporting efforts were underway at the end of FY91.
In addition, two major test programs were completed in the fourth quarter of FY91, the Calvert
scrubber trace metal tests discussed in Section 6 and the CIC site tests discussed in Section 7,
for which sample analyses and data evaluation efforts were underway at the end of FY91. All
remaining test sample analyses, test data reduction and interpretation, and test reporting efforts
for these three test programs will proceed through to completion in FY92.
With respect to test activities, three firm test programs are planned for FY92:
• Incinerability testing of contaminated sludges from the Bofors-Nobel Superfund
site in Region 5 (D. H. Ellison, Region 5, L. Janis, USAGE, coordinators); testing
is planned to begin in October 1991
• Incinerability testing of contaminated soils from the Scientific Chemical Processing
Superfund site in Region 2 (P. Evangelista, Region 2, R. Koustas, RREL/Edison,
coordinators); testing is tentatively planned to begin in November 1991
• Testing to evaluate incinerator emissions associated with repeated incinerator
waste-feed-cutoff (WFCO) episodes to support the Regional incinerator permit
writers (S. Sasseville, coordinator); testing is tentatively planned to begin in
December 1991
Other candidate test programs under discussion that could be conducted later in FY92
(January 1992 on) include:
• One of the third-party test programs discussed in Section 10
• Testing of the fate of trace metals in the RKS using a spray dryer/baghouse
combination for air pollution control. A test matrix similar to that employed in
the Calvert scrubber tests discussed in Section 6 is contemplated if an appropriate
spray dryer/baghouse system can be rented or fabricated at acceptable cost.
• A parametric test series to evaluate the effect of feed metal form on trace metal
fate in the RKS. Alternative feed metal forms other than the aqueous solution
co-fed with a clay-based hazardous waste analog include an aqueous metal
101
-------
solution atomized into the kiln burner flame, and mixed metal oxide powders fed
with the clay-based hazardous waste analog.
A parametric field test series to evaluate a POHC surrogate soup for possible trial
burn applications.
Further evaluation testing of low-temperature thermal desorption as a Superfund
site remediation technology
Parametric testing of a synthetic Superfund soil matrix to support the Superfund
Program Office
102
-------
REFERENCES
1.
2.
Waterland L. R, "Operations and Research at the U.S. EPA Incineration Research
Fac^ Annual Report for FY89," EPA/600/2-90/012, March 1990.
Waterland L. R., "Operations and Research at the U.S. EPA Incineration Research
S£ Annual Report for FY90," EPA/600/9-91/010, April 1991.
*U.S. GOVERNMENT PRINTING OFFICE: 1 9 9 2 .6-» 8 . 0 0 y. 0 7 9 9
103
-------
------- |