905-D-95-002C

                             S B^^^^K^ uj
                               \TfX
                               REGION 5
Risk Assessment for the Waste Technologies Industries (WTI)
 Hazardous Waste Incinerator Facility (East Liverpool, Ohio)
                  DRAFT — DO NOT CITE OR QUOTE
                             Volume III:
CHARACTERIZATION OF THE NATURE AND MAGNITUDE OF EMISSIONS
                         Prepared with the assistance of:
                    A.T. Kearney, Inc. (Prime Contractor: Chicago, IL);
    with Subcontract support from: ENVIRON Corp. (Arlington, VA), Midwest Research Institute (Kansas City, MO)
             and EARTH TECH, Inc. (Concord, MA) under EPA Contract No. 68-W4-0006
             NOTICE: THIS DOCUMENT IS A PRELIMINARY DRAFT.
    It has not been formally released by the U.S. Environmental Protection Agency as
       a final document, and should not be construed to represent Agency policy.
             It is being circulated for comment on its technical content.

-------
V..,-
                                  VOLUME ra

                            CHARACTERIZATION OF THE
                       NATURE AND MAGNITUDE OF EMISSIONS
                                                        External Review Draft
         Volume III                                        Do Not Cite or Quote

-------
                                VOLUME ra

                        CHARACTERIZATION OF THE
                  NATURE AND MAGNITUDE OF EMISSIONS

                               CONTENTS

                                                                           Page

     I.  INTRODUCTION                         ,                           1-1

        A.  Overview of Routine Emissions                                     1-1
        B.  Peer Review Group Comments                                      1-3

     II.  DATA USED IN CHARACTERIZING EMISSIONS                      II-l

        A.  Waste Profile Data                                               II-l
        B.  Overview of Stack Testing at the WTI Facility                        II-4
        C.  Key Assumptions for Emissions Characterizations                      II-8

    III.  INCINERATOR STACK EMISSIONS                                  III-l

        A.  Substances of Potential Concern in Stack Emissions                    III-l
        B.  Development of Chemical-Specific Stack Emission Rates                III-l
            1.   Chlorinated Dioxins and Furans (PCDDs/PCDFs)                  III-2
            2.   Other PICs and Organic Residues                               III-3
            3.   Metals                                                      IH-6
            4.   Acid Gases                                                 111-10
            5.   Particles                                                   III-l 1
        C.  Determination of Emissions Partitioning                             III-11
            1.   Partitioning of Emissions Between the Vapor
                and Particle Phases                                          III-11
            2.   Distribution of Constituents Emitted on Particles                  111-13
        D.  Key Assumptions for Incinerator Stack Emissions                     III-13

    IV.  FUGITIVE EMISSIONS                                              W-l

        A.  Potential Emission Sources                                        IV-1
        B.  Substances of Potential Concern in Fugitive Emissions                  IV-2
            1.   Substances  of Potential Concern in Fugitive Organic
                Vapor Emissions                                              FV-2
            2.   Substances  of Potential Concern hi Fugitive Ash Emissions          FV-3
        C.  Development of Fugitive Emission Rates                            IV-3
            1.   Tank-Related Emissions from the CAB System                    IV-3
            2.   Other Organic Fugitive Emissions                               FV-4


                                                            External Review Draft
Volume III                               i                    Do Not Cite or Quote

-------
                            CONTENTS (continued)
             3.  Fugitive Ash Handling Emissions
         D.  Key Assumptions for Fugitive Emissions

     V.  UNCERTAINTY IN EMISSIONS CHARACTERIZATION

         A.  Uncertainties in Stack Emissions Characterization
             1.  Uncertainty Associated with Metal Emissions
             2.  Uncertainties Due to Uncharacterized Stack Emissions
             3.  Uncertainties Associated with Chromium
             4.  Uncertainties Associated with Laboratory Contamination
         B.  Uncertainties Introduced by Process Upset Emissions
         C.  Uncertainties in Fugitive  Emissions Characterization

    VI.  REFERENCES
                IV-5
                IV-6

                V-l

                V-l
                V-2
                V-7
                V-9
                V-9
               V-10
               V-14

                VI-1
                                    TABLES

Table II-1:    Average System Removal Efficiencies (SREs) Measured
             at the WTI Facility                                                  II-9
Table II-2:    Chlorinated Dioxin/Furan Stack Emissions at WTI Facility              11-10
Table II-3:    Key Assumptions for Chapter II                                      11-11
Table III-l:   Substances of Potential Concern in Stack Emissions                     III-14
Table III-2:   Estimated Average and High-end Stack Emission Rates for
             Dioxin and Furan Congeners                                         111-16
Table III-3:   Estimated Average and High-end Emission Rates for Products of
             Incomplete Combustion (PICs) and Residues of Organic Compounds     III-17
Table III-4:   Compounds Anticipated to be Emitted in Very Low  Quantities
             for which Emission Rates are Not Developed                          111-24
Table III-5:   Estimated Average Metal Emission Rates                              111-25
Table III-6:   Estimated Average Acid Gas and Particle Emission Rates               111-26
Table III-7:   Key Assumptions for Chapter III                                     111-27
Table IV-1:   Fugitive Substances of Potential Concern                               IV-7
Table IV-2:   Estimated Total Fugitive Organic Vapor Emissions Rates                 IV-8
Table IV-3:   Key Assumptions for Chapter IV                                      IV-9
Table V-l:    Observed Variation in the Control Efficiency of Selected
             Metals During the May 1993  Trial Bum                               V-16
Table V-2:    Possible Variation in Predicted Metals  Emissions Due to
             Uncertainty hi Input Data                                           V-17
Volume III
                                          11
External Review Draft
Do Not Cite or Quote

-------
                                        CONTENTS (continued)
l
                                               FIGURES

            Figure II-1:   Total Chlorinated Dioxin/Furan Emissions vs. Chlorine Feed Rate        11-12
            Figure II-2:   TEQ Emissions vs. Chlorine Feed Rate                                11-12
            Figure III-1:  Pathways Available for Toxic Metals hi the WTI Incinerator             111-31
            Figure IV-1:  Locations of Stack, Fugitive Organic Vapor & Ash
                         Emission Sources                                                   IV-10
                                           APPENDICES
           Appendix III-l:   Emissions Estimation Methodology and Background
           Appendix III-2:   Summary of Measured Dioxin/Furan Congener Emission Rates
           Appendix III-3:   Products of Incomplete Combustion Analyzed for and Detected in the
                            Trial Burn and Performance Tests
                                                                           External Review Draft
           Volume III                               iii                     Do Not Cite or Quote

-------
                                I.  INTRODUCTION
A.  Overview of Routine Emissions
     An initial step of the WTI human health and ecological risk assessments is to evaluate
the nature and magnitude of atmospheric emissions during operations of the WTI facility.
This involves identifying potential emission sources, characterizing the composition of
emissions from these sources, and developing emission rates for the substances of potential
concern from each significant source.
     Routine operations at the WTI facility will result hi emissions from  several locations
within the facility, including particles and vapors in stack gases generated by the incinerator,
fugitive organic vapors emitted during waste processing and storage, and particles released
during handling of ash produced by incineration.  The general approach  used to characterize
routine emissions associated with these sources during normal operations is described in this
volume, with details provided in Appendix III-l.  The characterization of potential emissions
during on-site and off-site accidents is described in Volume VII.
     Normal combustion processes in hazardous waste incinerators will result in the release of
stack gas emissions into the atmosphere.  These emissions  will consist primarily of
combustion gases, such as carbon dioxide (CO^, carbon monoxide (CO), water, nitrogen,
and  oxygen.  However, despite the high temperatures typical of hazardous waste incinerators,
a fraction of the organic compounds hi the waste feed can  still pass through the incineration
process without being combusted.  In addition to the uncombusted residues of the waste feed,
fragments of the partially combusted organics from the feed may be emitted along with
organic chemicals formed through reactions in the combustion or post-combustion zones.
These compounds, known as products of incomplete combustion (PICs),  can be different in
chemical structure from the original organic compounds in the waste feed. Other potentially
hazardous substances, such as  metals, may also be present  hi the stack emissions.
     In the risk assessment, substances of potential concern in the stack gases are classified as
follows:

     •   Polychlorinated dioxins and furans (PCDDs/PCDFs).  which are believed to  be a
        product of incomplete combustion of some types of hazardous waste;
                                                                  External Review Draft
Volume III                                I_l                      Do Not Cite or Quote

-------
    •   Other organic chemicals, including PICs other than PCDDs/PCDFs, and residues of
        organic chemicals present in the feed that are not completely combusted within the
        incinerator;

    •   Metals, which may be present hi the waste but can not be destroyed by combustion;

    •   Acid Gases, such as nitrogen oxides (NO,), sulfur oxides (SOJ, and hydrogen
        chloride (HC1), which are formed during the combustion process; and

    •   Particles, which may be entrained hi the stack gas during waste combustion, or
        formed as stack gases cool hi the post-combustion zone of the incinerator.

    Comprehensive stack testing, including the collection of several sets of data on
PCDD/PCDF emission rates, was performed at the WTI facility to provide site-specific
estimates of organic emissions. For the risk assessment, these site-specific measurements are
supplemented by developing a waste feed chemical composition profile (based on wastes
received or projected to be received by WTI during its first year of operation) and applying
an incinerator destruction and removal efficiency (ORE)  (based on testing at the WTI
facility) to estimate emissions for chemicals not analyzed for during the stack testing.
Particle emissions and acid gas emissions are also estimated based on stack testing at the
WTI facility.  For metals, data on system removal efficiencies (SREs) derived from testing at
the WTI facility are used along with the projected waste feed composition to estimate
emissions. Where SRE data from WTI tests are not available, thermodynamic considerations
of metal behavior are used to extrapolate from metals for which test data are available to
other metals that were not analyzed for at WTI.  In this  manner, through a combination of
testing and predictive modeling,  incinerator stack gas emission rate estimates are developed
for the risk assessment.
    In addition to  incinerator stack emissions,  atmospheric releases may occur from fugitive
sources during waste unloading,  processing, and storage, and during the handling of
incinerator ash. The following potentially significant sources of fugitive organic vapor
emissions have been identified through an analysis of the WTI facility design and operating
procedures and a review of the operating experience at similar incineration facilities:

     1)  Carbon adsorption bed (CAB) system, which controls emissions from  tanks hi the
        organic waste tank farm building and from the container processing area;
     2)  Seals, valves, and flanges associated with tanks and piping hi the organic waste tank
        farm building that are vented from the building;

                                                                  External Review Draft
Volume III                                 1-2                      Do Not Cite or  Quote

-------
     3)   Wastewater holding tank' and
     4)   On-site truck wash station.

     In addition to these sources of organic vapor emissions, the bag filter used to control
 particle emissions during the loading of fly ash from the incinerator air pollution control
 system into trucks is identified as the primary source of fugitive ash handling emissions.
 Substances of potential concern that may be emitted from these fugitive emission sources are
 identified through an evaluation of the specific processes resulting in the release, and a
 review of the waste feed composition profile.  Emission rates for fugitive sources are
 developed using the waste profile information and predictive models.
     Process upsets may periodically occur during normal operation of the WTI facility.
 These upsets may include, but are not limited to:  1) unplanned plant shutdowns, 2)
 interruptions hi control measure systems, 3) interruptions in air flow,  and 4) kiln
 overpressure events.  Although brief periods of high emissions may result from these events,
 the magnitude of emissions from process upsets is not expected to be significant compared to
 the quantity of routine facility emissions.  Chapter V describes the identified process upsets
 and then- relative contributions to total facility emissions.

 B.   Peer Review Group Comments
     As part of the review process for the WTT Risk Assessment Project Plan (U.S. EPA
 1993a), a peer review panel provided several recommendations on the identification of
 chemicals emitted from the WTI  facility and the estimation of the magnitude of emissions.
 Peer Review Panel recommendations that are specifically addressed in the analysis described
 in this volume  include the following:

     •  Additional Compounds to Be Evaluated - The Peer Review Panel recommended that
        the Risk Assessment should include several constituents  not identified  in the Project
        Plan, including polycyclic aromatic hydrocarbons (PAHs) and two additional metals:
        copper and aluminum.

     •  Characterization of Organic Chemical Emissions - The Peer Review Panel
        recommended that additional, more comprehensive stack testing be performed at the
        WTI facility to provide site-specific organic emission estimates.  In addition, it
        recommended that a waste feed chemical composition profile be developed and  an
        incinerator destruction and removal efficiency (based on testing at the  WTI facility)
        be applied to estimate emission rates for individual chemicals in the waste feed  that
        were not measured directly in the stack testing.

                                                                   External Review Draft
Volume III                                1-3                      Do Not Cite  or  Quote

-------
    •   Characterization of Dioxin/Furan Emissions - The Peer Review Panel recommended
        that additional data on dioxins and furan emission rates be collected at the WTI
        facility to provide site-specific data over an extended time period.

    •   Development of Metal Emission Rates - The Peer Review Panel recommended that
        the existing data derived from the trial burns be used along with thermodynamic
        considerations of metal behavior to extrapolate the testing results to other metals that
        were not tested at WTI.

    •   Consideration of Fugitive Emissions Sources - The Peer Review Panel recommended
        that an evaluation of fugitive emissions at the WTI facility from sources other than
        the incinerator stack be conducted.

    •   Uncertainty of Emission Rates - Several members of the Peer Review Panel
        emphasized the need for a comprehensive evaluation of uncertainty with respect to
        emission rate estimations, and recommended the incorporation of sensitivity analyses
        of the modeling results and the selection of input parameter values.

    The manner hi which these comments are addressed is described hi the following
sections of this volume.
                                                                  External Review Draft
Volume III                                1-4                      Do Not Cite or Quote

-------
                                H.  DATA USED IN
                        CHARACTERIZING EMISSIONS
 A.  Waste Profile Data
     Wastes arrive at the WTI facility in a variety of forms (e.g., bulk, drummed) and
 physical states (e.g., liquids, solids, and sludges).  Each type of waste that is incinerated can
 affect incinerator stack emissions.  Pumpable wastes typically have the highest volatile
 content, and thus represent the most significant source of fugitive vapor emissions.  Although
 a list of specific substances hi the pumpable and non-pumpable wastes handled by the facility
 is not maintained by WTI, the facility does require waste profile sheets for each waste
 stream, which indicate the physical form and chemical content of each waste stream routinely
 handled at the facility.
     To identify specific substances likely to be in the wastes handled by the facility,  a data
 base has been developed from projections based on the information hi the waste profile
 sheets provided by WTI  (see Chapter  II of Appendix III-l for a more complete discussion).
 These profiles are completed by the generator prior to a waste being received by WTI to
 document the expected composition of the waste and to receive acceptance from WTI.  The
 profiles include the following information:

     •   Waste composition (within expected ranges);
     •   Physical state ("liquid", "liquid, solid/liquid mix",  "solid",  "solid/liquid  mix",
         "solid, solid/liquid mix");
     •   Specific handling instructions;
     •   Hazardous waste codes;
     •   Anticipated annual volume; and
     •   Approximate concentrations for a list of specific analytes tested by the  generators.
         These analytes include metals, certain anions (e.g., chloride, fluoride, and bromide),
         and certain organics  (e.g., PCBs,  dichlorofluoromethane, and
         trichlorofluoromethane).

 Prior to approving a waste for acceptance, WTI reviews the data hi the profile and verifies
 they are within permit conditions and  other operating constraints.  Once a waste is accepted
by WTI, the profiles must be maintained by WTI, and checked against manifests as wastes

                                                                  External Review Draft
Volume III                                II-l                      Do Not Cite  or Quote

-------
are shipped to the facility.  In addition, "fingerprint" analyses are performed by WTI at the
time of waste receipt to verify that the waste is consistent with the waste profile.
    To develop estimated annual average chemical concentrations for the overall waste
stream received by the WTI facility, information from 78 waste profile sheets is coded into a
data base (to protect confidentiality), and individual waste streams are then ranked based on
projected total annual volume.  These 78 profile sheets provide information on the wastes
actually received at WTI during the first nine months of operation, and are used along with
waste profile receipt data to develop a list of wastes to be received at the WTI facility.
    In estimating annual average concentrations hi the overall waste stream, projected
quantities for individual waste streams are derived from the waste profile sheets.  Individual
waste streams that each comprise less than 0.5 percent of the total volume are deleted from
the data base. The deleted waste streams constitute 4.45  percent of the total waste projected
to be received by the WTI facility, leaving 95.55  percent of the total waste stream in the data
base.
    Data on volumes of individual waste streams  actually received during the first nine
months of operation at the WTT facility are also compared against the projections contained
in the waste profile sheets.  As would be expected, there  is considerable variability between
the projected volumes for individual waste streams and the actual volumes received.  For
many waste streams, the projected volumes differ from the actual volumes by more than a
factor of 10,  and in several cases the difference is more than a factor of 100. However,
when evaluated over the course of a year the projections appear to provide a reasonable
estimate of the overall waste feed.  For example,  the projected volumes for the top 10% of
the individual waste streams (i.e., the 8 waste streams with the largest projected volumes)
make up approximately 55 % of the total projected waste. These 8 streams  also make up a
large portion (about 40%) of total waste actually received during the first nine months of
operation.
    The approximate constituent breakdown for each individual waste stream is estimated
based on information provided by the generators hi the  waste profile sheets. Waste stream
constituents are generally reported  in the waste profile sheets as ranges (e.g., 0 to 30 percent
of total waste stream).  In the risk assessment, the upper bound of the range for each waste
stream constituent  is assumed to be a conservative estimate of the constituent content La the
waste stream.  These upper bound  concentration estimates for individual constituents are then
summed to characterize the composition of the overall waste stream.  Using this
methodology, the combined upper  bound percentages for  individual constituent in the waste
usually exceed 100 percent.  In such cases, the individual constituent concentrations are
normalized to total 100 percent, keeping the relative amounts of each constituent the same as
hi the waste profile description.

                                                                  External  Review Draft
Volume III                                H-2                      Do Not Cite or Quote

-------
     As discussed in detail in Chapter II of Appendix III-l, a number of refinements are in
 the data base prior to estimating concentrations of specific chemicals in the facility waste
 streams.  These refinements include:

     •   Eliminating chemicals that can not be adequately characterized, including a
         compound listed as "2,3-dibromophosphate" and a compound listed under a
         confidential trade name that can not be identified. Such chemicals represent less
         than 1 % of the volume of waste stream in the data base.

     •   Eliminating "ash" for which a specific composition can not be estimated and which
         should be removed by ash handling systems.  Ash constitutes approximately 5 % of
         the waste volume.

     •   Eliminating "miscellaneous" substances such as  "grit", "dirt", "rust",  "trash",
         "tyvek", and "absorbent", which are not expected to contribute to facility emissions,
         and can not be adequately characterized.  "Miscellaneous" substances  represent 3%
         of the volume of the total waste stream.

     •   Eliminating "lithium batteries", which represent about 1 % of the waste, and can not
         be adequately characterized for the purposes of this risk assessment.

In addition, identical constituents reported under different names are consolidated and
reported under one listing, and isomeric compounds (e.g., ortho-, meta- and para- isomers of
xylene) are summed and reported under a single listing.  Although some waste streams are
eliminated from consideration, as noted earlier, the remaining data base of waste streams is
prorated to account for full thermal capacity of the unit.
    Concentrations of individual constituents in the overall waste data base are adjusted using
a "correction factor" calculated using die analytical results for the limited set of metals,
anions and other analytes hi the  waste profile sheets.  For example, each waste profile sheet
contains analytical results for chloride, as well as concentration ranges for all chemicals hi
the waste stream that contain chlorine.  After the quantities of all chemicals containing
chlorine are estimated using the  ranges reported by the generators, the corresponding amount
of chlorine is compared to the total quantity of chloride indicated in the analytical results
presented in the waste profile sheets. A correction factor is then calculated as  follows:
                                                                   External Review Draft
Volume III                                II-3                      Do Not Cite or Quote

-------
                                 mollyr of Cl, analytical results                    ,^_^
                                 mol/yr of Cl, generator ranges

Compounds containing chlorine are then multiplied by this factor to ensure consistency with
the analytical results. Correction factors for the other specific analytes reported in the waste
profile sheets are calculated and applied hi the same manner,
    Finally, based on physical state descriptions provided by the generators hi the waste
profile sheets, the waste stream constituent list is divided into five "composite" waste
streams:  (1) liquid; (2) liquid,  solid/liquid mix; (3) solid; (4) solid/liquid mix; and (5) solid,
solid/liquid mix.  Of these,  three categories  ("liquids", and "liquid,  solid/liquid mix" and
"solid/liquid mix")  are designated pumpable based on physical characteristics such as
viscosity.  The remaining two categories ("solid" and "solid, solid/liquid mix") are
considered to be non-pumpable.

B.  Overview of Stack Testing at the WIT Facility
    To the extent possible, testing performed at  the WTI facility is used to identify
substances of potential concern in the stack gases, and the estimated chemical-specific
emission rates.  Two major  stack emission measurement programs have been performed at
the WTI facility under the RCRA permit:

    •   Trial burns conducted hi March 1993 (ENSR 1993) and February 1994 (ENSR
        1994a); and

    •   Incinerator performance tests conducted hi August 1993 (WTI 1993), February 1994
        (Entropy 1994), April 1994 (ENSR  1994b),  August 1994 (ENSR 1994c; ENSR
        1994d), and December 1994 (ENSR 1995).

    The purpose of the trial burn program was to: (1) demonstrate that the incineration
system met permit requirements for organic  destruction and removal efficiency (DRE);  (2)
demonstrate that hydrogen chloride, chlorine, and particle emissions would meet permit
requirements while  operating under various worst-case operating conditions; (3) establish
system removal efficiencies  for  seven specific metals; and (4) define the range of allowable
operating conditions for the  incineration system.  As  with testing performed at similar
facilities, the trial burns at the WTI facility utilized engineered waste feeds specifically
synthesized to represent reasonable worst-case combustion or emission conditions.
                                                                  External Review Draft
Volume III                                II-4                      Do Not Cite or Quote

-------
     In addition to the trial burns, WTI conducted several performance tests.  These
 performance tests differed from trial burns in that they were conducted while the incinerator
 was operating under normal conditions and burning routine wastes, rather than synthetic,
 "engineered" waste streams.  Performance tests are believed to be more representative of
 day-to-day operations, whereas trial burns are meant to evaluate reasonable worst-case
 conditions.
     The original trial burn at the WTI facility was conducted in March 1993.  This trial burn
 consisted of testing at three different incinerator operating conditions, with three test runs
 performed under each test condition, for a total of nine test runs.  All nine test runs were
 performed using synthetic wastes with the maximum total chlorine feed rate and maximum
 heat rate (in lower heating value).  Specifics of the three test conditions are as follows:

     •  Condition 1 was designed to establish compliance under high kiln temperature,
        maximum kiln aqueous waste feed rate, maximum kiln organic liquids feed rate, and
        maximum toxic/carcinogenic metals feed rate.  The results from this test condition
        were intended to establish system removal efficiencies for toxic and carcinogenic
        metals, to allow the development of metals feed rate permit limits.

     •  Condition 2 was designed to establish compliance at low kiln temperature,  minimum
        secondary combustion chamber (SCC) temperature, maximum SCC aqueous liquids
        feed rate,  and maximum pumpable ash feed rate to the kiln.

     •  Condition 3 was also conducted at low kiln temperature, and was designed to study
        operation at maximum solids feed rate to the kiln.

    In the trial burn, WTI was required to demonstrate a 99.99% destruction and removal
efficiency (DRE) for the following principal organic hazardous constituents ("POHCs"):
chlorobenzene, trichloroethene, 1,2,4-trichlorobenzene, and carbon tetrachloride.  The  DRE
is defined as:
                                      Emission Rate of POHC,
                                        Feed Rate of POHC.
    During the trial burn, three of these four POHCs were fed during each test run. Carbon
tetrachloride and chlorobenzene were fed under each of the three conditions; trichloroethene
                                                                  External Review Draft
Volume III                                II-5                     Do Not Cite or Quote

-------
was fed under Condition 1 and 2; 1,2,4-trichlorobenzene was fed under Condition 3.  In
addition, all of the runs for the three conditions included sampling and analysis for
PCDD/PCDF and other organic compounds hi the stack gas.
     The March 1993 trial burn also included the measurement of emissions of seven metals:
antimony, arsenic, beryllium, cadmium, chromium, lead, and mercury.1  Measurements for
the seven tested metals were performed only under Condition 1,  i.e. the condition intended to
maximize metal emissions.  Using the emission rates detected during three runs under
Condition 1,  and the concentrations of metal in waste feed during each run, system removal
efficiencies (SREs) were measured  for each of the seven metals.  The  SRE is defined as:
                          SRE.   1                        i                    (11-3)
                              '        Feed Rate of Metcdi
The average SREs measured hi the trial burn are presented hi Table II-1.
    Early in April 1993, WTI representatives notified U.S. EPA that preliminary test results
indicated that the required DRE had not been demonstrated for carbon tetrachloride during
Condition 2 of the trial burn.  U.S. EPA evaluated the test results and determined that the
failure had been caused by carbon tetrachloride present hi a very dilute aqueous stream fed
into the SCC.  U.S. EPA subsequently imposed restrictions to ensure that the incineration
system would be operated hi compliance with the requirements of its permit, including
prohibiting aqueous waste feeds to  the SCC.  In February of 1994, WTI conducted a second
trial burn with a revised Condition 2, excluding aqueous waste feed to the SCC.  In this
second trial burn, all required performance standards were met during revised Condition 2,
which consisted  of four test runs.
    In June 1993, U.S.  EPA expressed concern over elevated PCDD/PCDF emission rates
recorded during  the March 1993 trial burn.  In order to reduce the PCDD/PCDF emissions,
WTI requested a permit modification to install an enhanced carbon injection system (ECIS).
The ECIS injects dry activated carbon into the flue gas stream at two locations, providing a
removal mechanism for organic substances through adsorption onto carbon particles. The
airborne carbon  particles are subsequently removed from the flue gas with the fly ash hi the
electrostatic precipitator (ESP) component of the air pollution control system.  As a condition
of approval of the installation of this new system,  U.S.  EPA required quarterly stack testing
    1   Stack sampling was not performed for the three other regulated metals — barium,
        silver, and thallium — based on WTI's agreement to accept permit conditions for
        these metals assuming no removal by the incineration or emission control system.

                                                                  External Review Draft
Volume III                              H-6                      Do Not Cite or Quote

-------
 for PCDDs/PCDFs and particles for the first year of operation, with annual testing
 thereafter. These "ECIS performance tests" included a minimum of five test runs, while
 burning typical wastes with varying total chlorine content. The modified permit required that
 WTI comply with an average2 stack gas PCDD/PCDF concentration of 30 ng/dscm (total
 sum of the tetra-through-octa congeners), a value recommended in U.S. EPA's May 1993
 Draft Strategy for the Combustion of Hazardous Waste (U.S. EPA  1993b).
    The ECIS was installed at WTI hi July 1993. The initial ECIS performance .test
 conducted in August 1993, and all subsequent performance tests and trial burns,
 demonstrated compliance with the PCDD/PCDF permit limit of 30  ng/dscm. The August
 1994 performance test included two extra runs to gam additional data on the potential effects
 of increased chlorine feed rate on PCDD/PCDF emissions, and the  December 1994
 performance test consisted of a total of 11 test runs over a range of chlorine feeds and two
 different kiln temperatures.  PCDD/PCDF congeners emission rate measurements used in the
 Risk Assessment are presented in Appendix III-2, and are summarized hi Table II-2.
    Both the chlorine feed rate and dioxin/furan emission rate data from the trial burns and
 August 1994 performance test are presented hi Table II-2.  As shown in Table II-2,
 PCDD/PCDF emissions were measured  over a range of total chlorine feed rates from
 approximately 400 Ib/hr to 3,300 Ib/hr.  A slight trend hi the dioxin/furan emissions and
 chlorine feed rate at the WTI  facility is suggested based on these data (see Figures II-1 and
 II-2).  However, regression analyses indicate low statistical significance hi this trend (r2  =
 0.084 for Total Emissions and r2  = 0.188 for TEQ Emissions). Thus,  no clear relationship
 between total chlorine feed rate and PCDD/PCDF emissions is demonstrated.
    In addition to PCDD/PCDF,  WTI voluntarily conducted stack sampling for an extended
 list of volatile and semivolatile organic products of incomplete combustion (PICs) during all
 runs of the August 1994 performance test and nine runs of the December 1994 performance
 test.   In addition to the  17 homologs of PCDD/PCDF, the August 1994 PIC testing  included
 a search for 36 volatile  compounds, 52 semivolatile compounds (eight compounds were
 analyzed by both volatile and semivolatile methods), nine of the 10 categories of  PCBs, and
 four specifically designated pesticides (specific compounds analyzed for hi the August 1994
 PIC testing are indicated hi Appendix  III-3). The December 1994 performance test analyzed
 for the same PICs, plus decachlorobiphenyl, which had not been analyzed for hi the August
 test due to analytical interferences. It should be noted that the original  March 1993 trial  burn
 included searching for a list of 24 organic PICs and POHCs, and that stack samples  from the
 February 1994 trial burn were analyzed for 39 volatile PICs and 64  semivolatile PICs.
    2    Average over the five runs of the ECIS performance test.

                                                                External Review Draft
Volume III                               II-7                     Do Not Cite or Quote

-------
However, since these organics data were collected while burning synthetic trial burn waste
instead of real waste, they were deemed less representative than the August 1994 PIC data.
    Altogether, the trial burns and performance tests conducted at WTI after installation of
the ECIS have yielded 37 test runs of PCDD/PCDF emission data and 16 test runs of PIC
data.  Because of the timing of this risk assessment, the results of the December 1994 testing
can not be evaluated along with the earlier test results.  Therefore this risk assessment relies
on the 26 PCDD/PCDF test runs and seven comprehensive PIC test runs conducted from
August 1993 through August 1994.

C.  Key Assumptions for Emissions Characterizations
    The key assumptions used in performing emission characterizations for the WTI facility
are summarized in Table II-3.  This table indicates the basis for the assumptions listed,  the
estimated relative magnitude of the assumptions' effect on the overall risk assessment, and
the direction of the effect,  if known.  These assumptions are further discussed in Chapter V
of this report.
                                                                  External Review Draft
Volume III                               H-8                     Do Not Cite or Quote

-------
TABLE H-l
Average System Removal Efficiencies (SREs) Measured at the WTI Facility
Metal
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
SRE (%)
99.986
99.977
> 99:991
99.987
> 99.9993
99.990
6.54
Source: March 1993 trial burn (ENSR 1993), average of three runs.
Volume III
II-9
External Review Draft
Do Not Cite or Quote

-------
TABLE H-2
Chlorinated Dioxm/Furan Stack Emissions at WTI Facility"
Date
Pre-ECIS*
March 1993
Post-ECIS"
August 1993
February 1994
February 1994
April 1994
August 1994
Type of Testing
Trial Burn
Performance Test
Performance Test
Trial Burn
(Re-run of
Condition 2)'
Performance Test
Performance Test
Tests
Performed
Condition 1
Condition 2
Condition 3
Run 1
Run 2
Run 3
Run 4
Run5
Run 1
Run 2
Run 3
Run 4
Run 5
Runl
Run 2
Run 3
Run 4
Runl
Run 2
Run 3
Run 4
Run 5
Run 1
Run 2
Run 3
Run 4
Run5
Run 6
Run?
Chlorine
Feed Rate
(Ib/hr)
3090
3117
3102
2386
2573
2351
2387
1823
2530
2103
1790
1970
1500
2958
3304
3231
3109
2459
2234
2004
2109
1389
1049
411
414
390
1904
1017
399
Total
Emissions
(ng/dscm)b
210
66.4
115
6
7
39
6
6
8.8
3.2
4.8
5.0
4.9
6.2
4.9
3.7
3.6
4.5
4.8
3.4
2.5
3.5
1.7
0.7
1.2
1.0
1.4
1.5
1.9
TEQ
Emissions

-------
TABLE H-3
Key Assumptions for Chapter II
Assumption
In applying the waste profile data, individual waste streams
that comprise less than 0.5% of the total volume are deleted
In estimating the constituent content in a waste stream, the
upper bound of the reported range for each waste stream
constituent is used. To prevent the combined percentage
from exceeding 100%, the constituent contents of the waste
streams are normalized.
The waste feed data are based on waste profile sheets for the
first year of operation.
Basis
The small quantities of these waste streams limit their effect,
so this simplifying assumption focuses the assessment on the
waste streams that present the most significant health hazard
Conservative estimate. Professional judgment based on
facility design and operation, and predicted waste
characterization.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics and receiving patterns.

Magnitude
of Effect
low
low
medium

Direction
of Effect
underestimate
overestimate
unknown
Volume III
II-ll
External Review Draft
Do Not Cite or Quote

-------


s
es
Dioxin/Fui
ig/dscm)
|B
.£ —
o JE
5!
"3
5




Afi -.
35 -
30 -
25 -
20 -
15 -
10 -
5 -

(

FIGURE D-l:
Total Chlorinated Dioxin/Furan Emissions vs. Chlorine Feed Rate




• / • • *• * * +•
* * •**
) 500 1000 1500 2000 2500 3000 35
Chlorine Feed Rate (lb/hr)*








00


FIGURE D-2:
0.3 -I
L °-25 -
S
gg
5 0.2 •
B 0.15 -
i o.i -
a
Cx
£ 0.05 •
0 -
(
TEQ Emissions vs. Chlorine Feed Rate
.




* t *
• •
% ***•*•*





I

) 500 1000 1500 2000 2500 3000 3500
Chlorine Feed Rate (lb/hr)*
 * Feed rate for chlorine in waste feed to incinerator (lb/hr).
 ** Polychlorinated dioxin/furan emissions expressed as Toxicity Equivalents (TEQs), in nanograms
 per dry standard cubic meter (ng/dscm). Results are from the 26 post-ECIS runs.
Volume ffl
11-12
External Review Draft
 Do Not Cite or Quote

-------
                   ffl.  INCINERATOR STACK EMISSIONS
 A. Substances of Potential Concern in Stack Emissions
    An initial list of substances of potential concern in the stack gases has been developed
 for the purposes of the risk assessment using the analytical results from the trial burns and
 performance tests. This list is supplemented by adding substances recommended for
 inclusion by the peer review committee (U.S. EPA 1993c) and PICs recommended for
 inclusion by U.S. EPA (1994a) in the Implementation Guidance for Conducting Indirect
 Exposure Analysis at RCRA Combustion Facilities. This listing, which consists of 174
 organic residues and PICs, 17 PCDD/PCDF congeners, 15 metals, 3  acid gases, and
 particles is presented in Table III-l.  All of the 17 PCDD/PCDF congeners and 3 acid gases
 were detected hi the WTI  stack testing, as were 7 of the 15 metals and 32 of the 174 organic
 residues and PICs.

 B.  Development of Chemical-Specific Stack  Emission Rates
    Due to the different sources of emission information used to characterize stack emissions
 in this risk assessment, and because of the different mechanisms associated with the
 generation  of different categories of pollutants,  different approaches must be utilized in the
 treatment of source data.   Statistical approaches are used where possible, as described below:

    •   For those chemical constituents where a reasonable number of detailed stack test
        runs were conducted (e.g., PCDD/PCDF and many "nondioxin PICs"),  a statistical
        approach is used to develop representative "average" and  "high end" values.

    •   For those chemical constituents where representative test data are not available and
        emissions must be estimated, conservative high-end estimates are consistently used.
        Since, in these cases, only one value is available (i.e., the highest estimate), a
        statistical approach can not be used. Thus, certain tables  hi this document report
        both the  "average" and "high-end" values for certain constituents as being the same
        (or, more precisely, the "average"  value is set equal to the "high-end" value).  This
        approach is deemed to be a conservative way of treating these estimated  values.
                                                                 External Review Draft
Volume III                               III-l                     Do Not Cite or Quote

-------
    •   Unlike organic PICs, stack emissions of metals are directly related to estimated
        input feed.  Because these emissions vary with the feed material, estimates of the
        quantities of metals in the feed material are essential to the calculation of emissions.
        However, since these feed estimates are based on conservative single-value
        summations of projected quantities over the first year of WTTs operation, a
        statistical approach can not be used for the metals emissions. The approach used is
        deemed to be conservative, i.e., it is expected that actual emissions of metals would
        be less than the values used in the risk assessment.

The specific approaches used to develop stack emission rate estimates for the WTI facility
are discussed separately below for PCDDs/PCDFs, other PICs and organic residues, metals,
acid gases, and particles.

    1.   Chlorinated Dioxins and Furans (PCDDs/PCDFs)
        The precise mechanism by which PCDDs/PCDFs result in incinerator stack
    emissions is not completely understood. While  it is possible that PCDDs/PCDFs may
    already be present in trace concentrations hi some hazardous waste feed streams, it is
    generally acknowledged that these compounds are more typically formed in the
    combustion or post-combustion zone of the incinerator through reactions involving
    chlorinated organic compounds.
        PCDD/PCDF emissions from WTI were first evaluated as part of the March 1993
    trial burn, conducted under three  different sets of operating  conditions.  The 2,3,7,8-
    tetrachlorodibenzo-p-dioxin (TCDD) equivalent  (TEQ) emission  rates for the  three
    conditions ranged from 20 ng/sec to 64 ng/sec.  During a three-day performance test
    conducted in early August 1993 after installation of the ECIS, PCDD/PCDF
    measurements were collected under five sets of operating conditions, each at least four
    hours in duration.  Total TEQ emission rates for the five runs ranged from 1.7 ng/sec to
    3.8 ng/sec, with an average of 2.4 ng/sec,  more than an order of magnitude below the
    emissions recorded from the March 1993 trial burn. These results indicated that the
    ECIS had significantly reduced PCDD/PCDF emissions from the facility.
        Additional trial burn and  performance tests  including  PCDD/PCDF measurements
    were conducted in February, April, and August 1994.  The results of these tests
    confirmed the effectiveness of the ECIS in reducing PCDD/PCDF emissions  from die
    facility. The total TEQ emission rates for the performance  tests and trial burn (total of
    nine runs) in February 1994 ranged from 0.55 ng/sec to 1.2 ng/sec, with an average of
    0.91 ng/sec.  The total TEQ emission rates for the five runs hi April 1994 ranged from
    0.34 to 0.49 ng/sec, with an average of 0.43 ng/sec, and  the total TEQ emission rates

                                                                 External Review Draft
Volume III                               HI-2                     Do Not Cite or Quote

-------
    for the seven runs in August 1994 ranged from 0.15 to 0.30 ng/sec, with an average of
    0.24 ng/sec. Thus, the dioxin/furan emission rates from the August 1994 performance
    test are approximately 100-fold lower than the original March 1993 trial burn results.
    Appendix ffl-2 presents the dioxin and furan sampling results from the individual trial
    burns and performance tests.
        Since repeated testing at the WTI facility has confirmed the effectiveness of the
    ECIS, estimates of PCDD/PCDF emission rates hi the risk assessment are based on
    emission data from the post-ECIS installation tests  taken through August .1994.  These
    data are summarized hi Table ffl-2.  Specifically, average  emission rates for the 17
    PCDD/PCDF congeners (listed hi Table DI-2) are  calculated as the arithmetic mean of
    the emission rates measured hi the 26 post-ECIS installation test runs taken through
    August 1994. Individual PCDD/PCDF congeners  not detected during a specific run are
    assumed to be present at one-half of the detection limit for the congener during that run.
        The variability of the PCDD/PCDF emission measurements is evaluated by
    estimating high-end emission rates.   High-end emission rates are estimated based on the
    95 percent upper confidence limit (UCL)3 of the arithmetic mean of the  26 post-ECIS
    installation performance test runs (assuming a normal distribution) or the maximum
    detected value, whichever is smaller, in accordance with U.S. EPA guidance for
    calculating the likely upper bound on mean data (U.S. EPA 1992b). In estimating high-
    end emission rates, PCDD/PCDF congeners that were not detected in a  specific run are
    conservatively assumed to be present at the detection limit for the congener hi that run.
    The estimated average and high-end emission rates  for the  17 dioxin and furan congeners
    are listed hi Table ffl-2.

    2.   Other PICs and Organic Residues
        Sampling of organic residues and PICs emitted from the WTI facility was conducted
    during the March 1993 and February 1994 trial burns; testing for PICs was also
    conducted hi August and December 1994.  As previously indicated, the  results of the
    3    The 95% UCL is defined as:

                      95-« UCL = mean + tOM.„_,
                                             VV»

        where:      t   =  Student-t statistic at 0.05 level of significance with n-1 degrees
                           of freedom
                   s   =  sample standard deviation
                   n   =  number of samples

                                                                External Review Draft
Volume ffl                              ni-3                    Do Not Cite or Quote

-------
   December 1994 performance tests were not available in time for inclusion in the risk
   assessment.  Appendix III-3 summarizes which of the 174 organic residues and PICs,
   listed in Section III.A above, were analyzed for during trial burns and performance tests,
   and which of these compounds were detected.  Table III-3 presents estimated emission
   rates for the PICs and organic residues, and indicates chemicals for which these rates are
   based on stack measurements from the August 1994 performance test.
       As noted previously, DREs for POHCs were also determined during the March 1993
   and February 1994 trial  burns.  The emission rates of the POHCs measured during the
   trial burns are not expected to be representative of emission rates during normal
   commercial operations because the trial burns were conducted using an engineered feed,
   spiked with POHCs at high concentrations.  The PIC emission rates and POHC DREs
   measured during the trial burns, however, are  used to estimate residual organic
   compound and PIC  emissions during normal commercial operations.
       Given the conditions under which the trial burns were conducted, the primary source
   of PIC emission rate estimates for the risk assessment is the extensive sampling of
   organics conducted during the August 1994 performance tests.  This program consisted
   of collecting samples during seven runs conducted during routine operation of the
   facility.  Samples collected during each of the seven runs were analyzed for a total of 93
   organic stack gas constituents in addition to individual congeners of PCDD/PCDF.
       For the 93  stack constituents analyzed for in the August 1994 tests, average emission
   rates in the risk assessment are estimated as the arithmetic mean of the seven runs from
   the August 1994 sampling. Nineteen PICs or  organic residues were detected in
   measurable quantities in at least one of the seven runs during the August 1994
   performance test. Seven compounds ~ methylene chloride, carbon disulfide,
   chloroform, carbon tetrachloride, bromodichloromethane, toluene, and bis(2-
   ethylhexyl)phthalate — were detected hi measurable quantities hi all seven runs.
   Compounds that were analyzed for but not detected are assumed to exist at one-half the
   detection limit hi the stack gas. High-end emission rates of the PICs and organic
   residues  are estimated based on the 95 percent UCL of the arithmetic mean of the
   measured PIC emission  rates, or the maximum detected concentration, whichever is
   lower.  In estimating high-end emission rates, compounds that were not detected are
   assumed to be present at the detection limit.
       A number  of the organic compounds listed in Table III-l were not analyzed for
   during the August 1994 testing.  For these compounds, emissions are estimated based on
   the following techniques:
                                                                External Review Draft
Volume III                               III-4                    Do Not Cite or Quote

-------
        (1) Use of measurements from the trial burns - During the March 1993 and
            February 1994 trial burns, WTI identified and measured emission rates for a
            number of PICs.  Because the trial burns were conducted using spiked,
            engineered waste feeds, these results are deemed less representative than
            subsequent performance tests that involved combustion of actual commercial
            waste streams.  Nonetheless, for PICs detected during the trial burns but not
            analyzed for during the August 1994 PIC sampling, trial burn results are used to
            estimate emissions.  Compounds that were analyzed for but not detected during
            the trial burn are assumed to be emitted at one-half the detection limit  for the
            compound.

        (2) Application of calculated DRE  to typical waste profile - An upper bound DRE is
            estimated based on the DRE values measured during the three runs of Condition
            2 of the March 1993 trial burn. (Condition 2 was selected because it resulted in
            the lowest DREs,  corresponding to the highest POHC emissions). The average
            of nine DRE values (three runs, three POHCs tested in each) is used to develop
            the upper-bound DRE.  An estimate of uncombusted organic emissions is
            determined from the organic feed rate and the estimated DREs using the
            following calculation:


                                E   Fx\l   £5*1                           (ffl-D
                                          I      100J
            where:
                E       =  emission rate, g/s
                F       =  chemical feed rate, g/s
                DRE    =  destruction and removal efficiency,  %

            The estimated worst-case DRE  is applied to the feed rates of organic compounds
            identified from chemical characterization data contained in waste profile sheets
            provided by WTI (described in Chapter II).

        Estimated emission rates are developed using the two approaches described above
    for the organic compounds not analyzed for during the August 1994 PIC testing.  For
    each chemical, the higher of the emission rates estimated using the two approaches listed
    above is selected.  Using this procedure, only a single emission rate is estimated for each
                                                                 External Review Draft
Volume III                               HI-5                     Do Not Cite or Quote

-------
    chemical; consequently, these values are used to represent both average and high-end
    emission rates.  The estimated average and high-end emission rates for the PICs and
    residual organic compounds from the stack are presented in Table III-3.
        It should be noted that 31  of the original list of 174 organic compounds of potential
    concern were not reported hi the waste profile sheets or analyzed for in stack emissions
    at WTI. Thus, emission rates  can not be estimated for these 31 compounds using either
    of the approaches listed above. It is assumed, therefore, that these compounds are not
    emitted at significant levels. These PICs are listed hi Table III-4.  Additionally, it
    should be noted that the August 1994 PIC testing yielded an emission rate for
    "m/p-xylene", but not for the individual isomers.  It is conservatively assumed,
    therefore, that the emission rates estimated for "m/p-xylene"  combined apply to both
    m-xylene and p-xylene individually.

    3.   Metals
        While the wastes to be treated  at the WTI incinerator are predominantly organic hi
    nature, inorganic substances, such as metals, are expected to be present in many waste
    streams.  Metals that are present hi the  incinerator feed may evaporate at the high
    temperatures in the rotary kiln and, subsequently, condense to form aerosols of
    submicron particles hi the cooler, later stages of the incineration process.  Although most
    of the metals would be captured  hi solid form along with slag and ash, a fraction of each
    metal is expected to escape the emission control systems and be vented to the atmosphere
    via the stack.
        Emission rates are developed for the ten metals regulated at the WTI facility
    (antimony, arsenic, barium, beryllium,  cadmium, chromium, lead, mercury, silver, and
    thallium) and five additional metals (aluminum, copper, nickel, selenium, and zinc).
    The behavior of each metal within  the incinerator is evaluated based on equilibrium
    modeling performed for the WTI facility (see Chapter III of Appendix III-l).  The model
    indicates that for the specific operating  conditions of interest, emission rates can be
    estimated based on the system removal  efficiency (SRE) data compiled from the trial
    burn and the projected waste feed data for the WTI facility.
        The general equation used to calculate metal emission rates for the incinerator stack
    is the following:

                                   £,.  (1 SRE)xF.                            (IH-2)
                                                                  External Review Draft
Volume III                                III-6                     Do Not Cite or Quote

-------
     where:
         EJ       =   annual average stack emission rate for metal i, Ib/yr
         FJ       =   annual feed rate for metal i, Ib/yr
         SRE    =   system removal efficiency, %/100

     The trial burn conducted at the facility in March 1993 prior to installation of the ECIS
     provided SREs for seven metals (antimony, arsenic, beryllium, cadmium, chromium,
     lead, and mercury). Trial bum data are not available, however, to estimate SREs for the
     remaining eight metals evaluated in this risk assessment (aluminum, barium, copper,
     nickel, selenium, silver, thallium, and zinc).  For metals where direct SRE
     measurements were made during the trial burns, the average SRE value from the various
     sampling runs is used.  For metals not analyzed hi the March 1993 trial burns, SRE
     values are extrapolated from the trial burn data for the metals that were tested,
     considering the results of thermodynamic modeling, as described below.
         The behavior of metals hi the incinerator train is modeled based on mechanistic
     theories of metal reactions and particle formation (Barton et al. 1990).  Figure III-l
     illustrates the pathways  metals may take through the WTI incinerator.  As illustrated hi
     Figure III-l,  metals present hi the waste feed may first volatilize,  become entrained as
     panicles hi the combustion gas stream, or enter the slag.  Complex oxidation and
     reduction reactions can then occur between metals and other reactive elements in the
     combustion gases hi the primary and secondary chamber, creating newly formed metal
     species with different physical/chemical properties than the metals introduced with the
     waste feed.  When metal speciation is expected to occur, the worst case scenario for
     oxidation state is assumed.  For example, all chromium emissions are assumed to be hi
     the Chromium VI oxidation state (the most toxic form of chromium).  Chromium
     speciation is further discussed hi Chapter V.
        As the combustion gas cools after it exits the secondary combustion chamber, a
    portion of the metals will condense  to form new particles, or condense on the surfaces of
     the entrained ash particles.  The formed particles collide with one another and with the
    entrained ash.  Investigations of combustion  systems have found that any particles
     smaller than about 0.1 j*m quickly coagulate, while those larger than 1 /*m do not (Linak
    and Wendt 1993).  Thus, two groups of particles typically enter the ah" cleaning system:
    one group ranges hi size from 0.1 to 1 jim and is formed from the metals that vaporized
    and subsequently condensed (fine particles);  the second group is generally hi the range of
     1 to 10 pm hi diameter  and consists of the material entrained hi the incinerator (coarse
    particles).
                                                                  External Review Draft
Volume III                                III-7                     Do Not Cite or Quote

-------
       After exiting the boiler, the cooled combustion gases enter the air pollution control
    (APC) system.  The mechanism of metal removal in the APC  system differs from metal
    to metal, and is largely a function of two characteristics: vapor solubility and particle
    size.  A detailed description of the methods used to model metal behavior in the APC
    system is contained hi Appendix III-l. The primary assumptions and determinations
    made in conducting the analyses include the following:

       •  The compounds PbC^ and CrO2Cl2 are commonly predicted to form at low
           temperatures by the thennodynamic programs, but are not typically found in
           combustion gases (Linak and Wendt  1993).  Thus, these two compounds are
           excluded from consideration during the thennodynamic modeling.

       •  MAEROS, a computer model which simulates the behavior of suspended
           particles (Gelbard 1980),  is used to examine the evolution of an aerosol in the
           WTI incinerator in studying coagulation processes.  Based on the modeling
           results, all metals that vaporize and subsequently condense form particles with
           diameters of about 0.5
        •   All of the ash in the non-pumpable waste is assumed to be incorporated into the
            slag. Little of the slag is entrained due to its viscous nature.  Since the ash for
            the non-pumpable waste is incorporated into the slag, it is concluded that none
            of the ash for the non-pumpable waste would be entrained.

        •   All of the pumpable waste is injected through a sludge lance, a slurry burner, or
            a liquid burner. These devices cause the wastes to separate into small drops
            that form solid particles as volatile compounds are released from the drops.
            Most of these particles are entrained. Thus, it is assumed that all of the ash in
            the pumpable wastes is entrained.

        •   Based on classical condensation theory, it has been determined that very little
            condensation occurs onto the surface of the larger entrained particles
            (Friedlander 1977;  McNallan et al. 1981).

        Based on the modeling, metals in the incineration system are  classified into one of
    four categories depending on their expected behavior in the incinerator system: 1)
    insoluble vapor; 2) soluble vapor; 3) fine particles (< 1 /un); and 4) coarse particles (1-
    10 ftm).  Soluble and insoluble vapors are  created by metals that vaporize in the

                                                                  External Review Draft
Volume III                               III-8                    Do Not Cite or Quote

-------
     incinerator, but do not condense in the quench system.  Fine particles vaporize hi the
     incinerator but subsequently condense hi the quench system.  Coarse particles are largely
     created by atomization of the waste and do not originate as vapors. The metals
     evaluated hi the risk assessment are categorized as follows:
Form of Metal
Insoluble vapor
Soluble vapor
Fine particles
Coarse panicles
SRE Measured in WTI Trial Burn
Mercury
-
Arsenic, Antimony, Beryllium,
Cadmium, Lead
Chromium
SRE Not Measured in WTI Trial Burn
'• -
Selenium
Barium, Copper, Nickel, Silver, Thallium
Zinc
Aluminum
        The SREs measured for the seven metals of potential concern tested hi the trial burn
    are shown hi Table III-5.  Based on the modeling, chromium is used to estimate the SRE
    for aluminum (coarse particles) and arsenic is used to estimate SREs for barium, copper,
    nickel, silver, thallium, and zinc. Arsenic is conservatively selected to represent metals
    on fine particles because it had the lowest measured SRE during the trial burn.  No
    metal analyzed for during  the WTI trial burn is classified as a soluble vapor.  Thus,
    another  monitored substance that is present hi the flue gas as a soluble vapor must be
    identified to establish the ability of the  flue gas cleaning  system to remove this class of
    material. Of the two substances  identified, SOX and HC1, SOX generally exhibits lower
    solubilities and lower removal efficiencies than HC1.  Thus, the removal efficiency of
    SOX is selected as a conservative  estimate of the ability of the flue gas cleaning system to
    capture soluble vapors.
        It should be noted that an SRE of zero is assumed for mercury, i.e., there would be
    no mercury removal hi the APC.  This is based on the very low SRE (< 10%) measured
    for  mercury hi the March  1993 trial burn, prior to installation of the ECIS.  However,
    the  ECIS may significantly increase the SRE for mercury.  For example, a medical
    waste incinerator hi Morristown,  NJ, observed a mercury SRE increase from an average
    of =  30% (based on three runs)  to an average of =90% (based on three runs) after
    installation of an APC (Radian 1992).  However, it is conservatively assumed hi the risk
    assessment that the ECIS would not enhance mercury removal hi the APC  system.
        Waste feed data for the 15 metals of potential concern are developed based on waste
    profile sheets and feed rates provided by WTI for the first year of operation at the
    facility,  as discussed hi Chapter II of this volume.  Because data from the first year of
    operation may not represent the maximum operating capacity of the system, the
    estimated metal feed rates are prorated to account for the maximum heat input of the
Volume III
III-9
External Review Draft
Do Not Cite or Quote

-------
    incinerator. Therefore, the metal feed rates are multiplied by the ratio of the maximum
    heat input rate based on the design of the kiln to the heat input rate derived from the
    waste profile data sheets to develop mayimuTri predicted metals feed rates, which are
    listed hi Table III-5.  The maTimum permitted heat input is 97.8 million BTU per hour
    (MMBTU/hr) on a lower heating value basis. This corresponds to a higher heating
    value of approximately 121 MMBTU/hr. The annualized heat input rate derived for the
    waste profile is 29.5 MMBTU/hr on a higher heating value basis.  Corresponding metal
    emission rates, calculated using the measured or estimated SRE values along with the
    maximum predicted metal feed rates, are also listed in Table III-5.
        Emission rates of metals are strongly influenced by the metal composition in the
    feed to the incinerator, and the variability hi the SREs for the individual metals in the
    incineration system.  Since the feed composition of metals is highly variable, high-end
    emission rates are not distinguished from average emission rates (i.e., high-end estunates
    of metal emission rates are developed  and are used to represent average metal emission
    rates).

    4.  Acid Gases
        A variety of acid gases, including hydrogen chloride (HC1), nitrogen oxides (NOJ,
    and sulfur oxides (SOJ, may be formed during  the incineration of hazardous wastes.
    The extent to which acid gases are generated and released from an incinerator is
    primarily related to waste composition, incinerator design and operation, and the
    effectiveness of pollution control equipment.
        WTI measured emissions of HC1 during the March 1993 trial burn (nine runs) and
    the February 1994 trial burn (four runs). The average emission rate of HC1 was
    calculated to be 0.25 Ib/hr (0.032 g/sec) using 12 of the 13 trial burn runs (one run in
    the February 1994 trial burn was prematurely terminated and was not used to estimate
    the HC1 emission rate).  This emission rate is considerably below the regulatory limit of
    4 Ib/hr (40 CFR 264.343b). The HC1 control efficiency was not substantially affected
    by the installation of the ECIS.
        During the trial burns, the total chlorine feed rate to the incinerator ranged between
    2,960 and 3,300 pounds per hour.  According to WTI,  typical chlorine  feed rates rarely
    exceed 800 pounds per hour during actual operation (personal communication, G.
    Victorine 1995). Thus, the HC1 emission rates, which are related to the chlorine feed
    rates, may  be significantly overestimated.
        WTI continuously monitors stack  emissions of NOX and SOX and makes the results
    available on an electronic bulletin board. Average emission rates of NOX and SOX are
                                                                  External Review Draft
Volume III                               HI-10                     Do Not Cite or Quote

-------
                estimated based on the mean of one randomly selected month of recent continuous
(^ ^            monitoring data (February 23, 1995 to March 21, 1995).
                    Average emission rates of the three acid gases considered in this assessment are
                presented hi Table III-6.

                5.  Particles
                    Particle emission rates were measured by WTI during the March  1993 trial burn
                (nine runs), February 1994 trial bum (four runs), and four performance tests (22 runs
                combined).4  The average particle emission rates is estimated as the arithmetic mean of
                these runs, respectively.  The estimated average particle emission rate is 0.07 g/sec. As
                discussed hi Appendix III-l, almost all the particles are less than  10 /*m in diameter.
                The estimated average emission rate of particles is shown hi Table III-6.

            C.  Determination of Emissions Partitioning
                Substances hi the stack gas will generally be present hi either the vapor phase (e.g.,
            volatile organic compounds, acid gases, and mercury) or the particle phase (e.g., metals and
            condensed organic compounds).  As the gases exit the stack, substances hi the vapor phase
            will either remain hi the vapor phase or become adsorbed to particles that are present hi the
^          stack gases and the atmosphere.  This partitioning between phases is based on  chemical-
 ^~"/        specific parameters, and is an important factor hi estimating the rate at which compounds
            deposit out of die atmosphere onto soil, surface water, and plants.  Due to their
            physical/chemical properties, such as vapor pressure, acid gases remain hi the  vapor phase,
            metals (with the exception of mercury) remain entirely as particles, while most organics tend
            to partition between the particle and vapor phases.  Some of the polycyclic aromatic
            hydrocarbons (PAHs) witii very low vapor pressures (e.g., dibenzo(a,h)anthracene and
            indeno(l,2,3-cd)pyrene) are assumed to be entirely hi the particle phase.

                1.  Partitioning of Emissions Between the Vapor and Particle Phases
                    The degree to which a vapor-phase chemical will attach to suspended  particles is a
                function of the amount of suspended particles hi ah" and the vapor pressure of the
                chemical. The fraction of a chemical that is adsorbed to particles is estimated using a
                theoretical model presented by Junge (1977), whereby:
                4    Particle sampling was conducted during 23 performance test runs.  However, the
                    particle concentration in one performance test was reported as zero, presumably due
                    to either sampling or laboratory error. Thus, only the results from the 22 tests with
                    non-zero particle concentrations are used in the risk assessment.

                                                                              External Review Draft
            Volume III                               HI-11                    Do Not Cite or Quote

-------
                                   =     CXST
                                     P°
    where:
        <£   =  fraction of organic chemical adsorbed to particles, unitless
        ST  =  particle surface area per unit volume of air, cnrVcm3
        p°  =  vapor pressure, atm
        c   =  molecular weight and heat of condensation factor, atm-cm

        For particles, surface area per unit volume of air (87) is assumed to be 3.5 x 10"6
    cm2/cm3, which corresponds to a value for "background plus local sources" (Bidleman
    1988). This value is the most appropriate value available for this parameter given the
    commercial/industrial area around the WTI facility.  Other values of particle surface area
    per unit volume cited by Bidleman (1988) include 4.2 x 10"7 cm2/cm3 for clean
    continental background; 1.5 x 10"* cm2/cm3 for average background; and 1.1 x 10"5
    cm2/cm3 for urban air.  The molecular weight and heat of condensation factor (c) does
    not vary much between compounds and is estimated to be 1.7 x 10"* atm-cm (Junge
    1977; Bidleman 1988).
        For compounds with melting points higher than ambient temperature (i.e.,
    compounds that are solid at ambient temperature, such as dioxin- and furan-like
    compounds and some semi-volatile organic compounds),  Bidleman (1988) and U.S. EPA
    (1994a) recommend that the sub-cooled liquid vapor pressure, PL, be used as the vapor
    pressure in the above equation, because it compares more favorably with field data.  The
    value of PL is estimated using the following equation:
                                    L  __   /m-
                                   P         RT
    where:
        PL  =  sub-cooled liquid vapor pressure, atm
        Ps  =  crystalline solid vapor pressure, atm
        ASf =  entropy of fusion, atm-m3/mole-K
        Tm  =  melting point, K
        T   =  ambient air temperature, K

                                                                 External Review Draft
Volume HI                              m-12                    Do Not Cite or Quote

-------
        R   =  universal gas constant, atm-m3/mole-K

        Bidleman (1988) reports that ASf/R can be satisfactorily estimated as 6.79.  Partition
    factors calculated for substances evaluated in the human health and ecological risk
    assessments are presented hi Volume V.

    2.  Distribution of Constituents Emitted on Particles
        Two separate approaches are used for organics and metals hi determining the
    physical  distribution of a chemical on or within a particle in the stack gas.  It is assumed
    that a portion of the organic compounds, which would be hi the vapor phase exiting the
    combustor, adsorb onto the outer surface of airborne particles as condensation occurs in
    the cooler regions of the post-combustion zone (i.e., surface distribution).  Metals are
    assumed  to be homogeneously dispersed throughout the entire particle (i.e., mass
    distribution) because they may form particles themselves rather than condensing onto
    existing particles.  Therefore, for organics, deposition on particles is a function of the
    total surface area of particles emitted from the stack; whereas, for metals, deposition is a
    function  of the total mass.  The type of distribution (mass versus surface area) that
    occurs is accounted for hi the air dispersion modeling, as described hi Volume IV.

D.  Key Assumptions for Incinerator Stack Emissions
    The key  assumptions used hi predicting and analyzing incinerator stack emissions for the
WTI facility are summarized hi Table III-7. This table indicates the basis for the
assumptions listed, the estimated relative magnitude of these assumptions' effect on the
overall risk assessment, and the direction of the effect, if known.  These assumptions are
further discussed hi Chapter V of this report.
                                                                  External Review Draft
Volume III                               111-13                    Do Not Cite or Quote

-------
                                  PICs and Residual Organk Compounds
  Acenaphthene
  Acenaphthylene
  Acetaldehyde
  Acetone
  Acetophenone
  Acrolein
  Acrylomtrile
  Anthracene
  Benzaldehyde
  Benzene
  Benzole acid
  Benzotrichloride
  Benzo(a)anthracene
  Benzo(a)pyrene
  Benzo(b)fluoranlhene
  Benzo(e)pyrene
  Benzo(g,h,i)perylene
  Benzo(j)fluoranthene
  Benzo(k)fhioranthene
  Benzyl chloride
  Biphenyl
  Bis(2-chloroethoxy) methane
  Bis(2-chloroethyl)ether
  Bis(2-chloroisopropyl)ether
  Bis(2-ethylhexyl)phthalate
  Bromochloromethane
  Bromodichloromethane
  Bromoethene
  Bromofonn
  Bromomethane
  Bromodiphenylether, p-
  Butadiene, 1,3-
  Butanone, 2- (MEK)
  Butylbenzylphthalate
  Carbon disulfide
  Carbon tetrachloride
  Chlordane
  Chloro-3-methylphenol, 4-
  Chloroacetophenone, 2-
  Chloroaniline, p-
  Chlorobenzene
  Chlorobenzilate
  Chloroethane
  Chloroform
  Chlorometnane
  Chloronaphthalene, beta
Chlorophenol, 2-
Chlorodiphenylether, 4-
Chloropropane, 2-
Chrysene
Cresol, m-
Cresol, o-
Cresol, p-
Crotonaldehyde
Cumene
2,4-D
4,4'-DDE
Dibenz(a,h)anthracene
Dibenz(a,h)fluoranthene
Dibromo-3-chloropropane, 1,2-
Dibromochloromethane
Dichloro-2-butene, cis-1,4-
Dichloro-2-butene, trans-1,4-
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine, 3,3'-
Dichlorobiphenyl
Dichlorodifluoromethane
Dichloroethane, 1,1-
Dichloroethane, 1,2-
Dichloroethene, 1,1-
Dichloroethylene, trans-1,2-
Dichlorofluoromethane
Dichlorophenol, 2,4-
Dichloropropane, 1,2-
Dichloropropene, cis-1,3-
Dichloropropene, trans-1,3-
Diethylphthalate
Dimethoxybenzidine, 3,3'-
Dimethylphenol, 2,4-
Dimethylphthalate
Di-n-butylphthalate
Di-n-octyl phthalate
Dinitritoluene, 2,6-
Dinitro-2-methylphenol, 4,6-
Dinitrobenzene, 1,2-
Dinitrobenzene, 1,3-
Dinitrobenzene, 1,4-
Dinitrophenol, 2,4-
Dinitrotoluene, 2,4-
Dioxane, 1,4-
Ethyl methacrylate
Ethylbenzene
Ethylene dibromide
Ethylene oxide
Ethylene thiourea
Fluoranthene
Huorene
Fbnnaldehyde
Furfural
Heptachlor
Heptachlorobiphenyl
Hexachlorobenzene
Hexachlorobiphenyl
Hexachlorobutadiene
Hexachlorocyclohexane, alpha-
Hexachlorocyclohexane, beta-
Hexachlorocyclohexane, gamma-
  (a.k.a. Lindane)
Hexachlorocyclopentadiene
Hexacnloroethane
Hexachlorophene
Hexane, n-
Hexanone, 2-
Hexanone, 3-
Ihdeno(l ,2,3-cd)pyrene
Isophorone
Maleic hydrazide
Methoxychlor
Methylene bromide
Methylene chloride
Methylnaphthalene, 2-
Methyl-tert-butyl ether
Methyl-2-Pentanone, 4-
Monochlorobiphenyl
Naphthalene
Nitroaniline, 2-
NitroanUine, 3-
NitroanUine, 4-
Nitrobenzene
Nitrophenol, 2-
Nitrophenol, 4-
N-Nitroso-di-n-butylamine
N-Nitroso-di-n-propylamine
N-Nitrosodiphenylamine
Nonachlorobiphenyl
Octachlorobiphenyl
Volume III
             111-14
          External Review Draft
           Do Not Cite or Quote

-------
Substances
Pentachlorobenzene
PentacMorobiphenyl
Pentachloronitrobenzene
Pentachlorophenol
Phenanthrene
Phenol
Phosgene
Propionaldehyde
Pyrene
Quinoline
Quinone
Safrole
Dioxin Congeners
2,3,7,8-TetraCDD
1,2,3,7,8-PentaCDD
1,2,3,4,7,8-HexaCDD
1,2,3,6,7,8-HexaCDD
1,2,3,7,8,9-HexaCDD
1,2,3,4,6,7,8-HeptaCDD
OctaCDD

Aluminum
Antimony
Arsenic
Barium
Beryllium
TABLE m-1 (continued)
of Potential Concern in Stack Emissions
Styrene Trichlorobiphenyl
Tetrachlorobenzene, 1,2,4,5- Trichloroethane, 1,1,1-
Tetrachlorobiphenyl Trichloroethane, 1,1,2-
Tetrachloroethane, 1,1,1,2- Trichloroethene
Tettachloroethane, 1,1,2,2- Trichlorofluoromethane
Tetrachloroethene Trichlorophenol, 2,4,5-
Tetrachlorophenol, 2,3,4,6- Trichlorophenol, 2,4,6-
Toluene Trichloropropane; 1,2,3-
Toluidine, o- Vinyl acetate
Toluidine, p- Vinyl chloride
Trichloro-l,2,2-TFE, 1,1,2- Xylene, m-
Trichlorobenzene, 1,2,4- Xylene, o-
Xylene, p-
Foran Congeners
2,3,7,8-TetraCDF 1,2,3,7,8,9-HexaCDF
1,2,3,7,8-PentaCDF 2,3,4,6,7,8-HexaCDF
2,3,4,7,8-PentaCDF 1,2,3,4,6,7,8-HeptaCDF
1,2,3,4,7,8-HexaCDF 1,2,3,4,7,8,9-HeptaCDF .
1,2,3,6,7,8-HexaCDF OctaCDF
Metals
Cadmium Mercury (inorganic and organic)
Chromium (hexavalent and Nickel
trivalent) Selenium
Copper Silver
Lead Thallium
Zinc
Acid Gases
Hydrogen chloride
Total nitrogen oxides (NO,)
Total sulfur oxides (SO,)
Paniculate Matter
Respirable (PM10)
Total
Notes:
TEE - trifluoroethane MEK - methyl ethyl ketone
CDD - chlorodibenzo-p-dioxin PM,0 - paniculate matter < 10 microns
CDF - chlorodibenzofuran
Volume III
ffl-15
External Review Draft
Do Not Cite or Quote

-------
TABLE m-2
Estimated Average and High-end Stack Emission Rates for
Dioxin and Furan Congeners
Congener
Emission Rate (g/sec)
Average
High-end
Dioxin Congeners
2,3,7,8-TetraCDD
1,2,3,7,8-PentaCDD
1,2,3,4,7,8-HexaCDD

1,2,3,6,7,8-HexaCDD

1,2,3,7,8,9-HexaCDD
1,2,3,4,6,7,8-HeptaCDD
OctaCDD
1.08 x 10-"
6.78 x 10-"
8.95 x 10-11
1.66 x 10-10
1.09x 10-10
1.24 x 10-9
6.15x10*
2.16x 10-"
9.46 x 10-"
1.25 x 10-10
2.18x 10-10
1.55 x 10-10
1.69 x 10-9
9.80 x 10-9
Furan Congeners
2,3,7,8-TetraCDF
1,2,3,7,8-PentaCDF
2,3,4,7,8-PentaCDF
1,2,3,4,7,8-HexaCDF
1,2,3,6,7,8-HexaCDF

2,3,4,6,7,8-HexaCDF

1,2,3,7,8,9-HexaCDF
1,2,3,4,6,7,8-HeptaCDF

1,2,3,4,7,8,9-HeptaCDF
OctaCDF
8.77 x 10-11
3.45 x 10-10
4.67 x 10-10
1.43 x 10-9
1.33 x 10-9
1.50 x 10-9
2.93 x 10-10
9.30 x 10-9
1.22 x 10-9
1.89 x 10-8
l.lSxlO-10
4.35 x 10-10
6.04 x 10-10
1.85 x 10-9
1.71 x 10-9
1.96 x 10-9
3.85 x lO'10
1.30 x 10-8
1.80 x 10-9
3.62 x 10-8
Notes:
CDD - chlorodibenzo-p-dioxin
CDF - chlorodibenzo-p-furan
Volume III
111-16
External Review Draft
Do Not Cite or Quote

-------
TABLE m-3
Estimated Ayerage and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Acenaphthene
Acenaphthylene
Acetaldehyde
Acetone
Acetophenone
Acrylonitrile
Anthracene
Benzene
Benzole acid
Benzotrichloride
Benzo(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(g ,h, i)perylene
Benzo(k)fluoranthene
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
Bromodichloromethane
Bromoform
Bromomethane
Bromodiphenylether, p-
Emission Rate (g/sec)
Ayerage
6.69 x 10^
6.69 x 10*
3.01 x lO4
2.90 x 10-3
2.93 x 104
2.02 x 1O4
5.50 x 10^
1.47 x 10-5
1.13 x 10-5
3.20 x lO'5
5.50 x 10*
5.50 x 10*
5.50 x 10*
5.50 x 10-*
5.50 x 10*
6.69 x 10*
1.33 x lO"5
6.69 x 10-6
3.72 x 10-5
1.03 x lO4
5.50 x 10*
4.90 x 104
6.69 x 10-*
High-end
6.69 x 10*
6.69 x lO*
3.01 x 104
2.90 x lO"3
2.93 x 10-»
2.02 x lO4
l.lOxlO'5
2.63 x lO'5
l.lSxlO-5
3.20 x 10-5
LlOxlO'5
l.lOxlO-5
LlOxlO"5
l.lOxKT5
l.lOxlO-5
6.69 x 10-6
1.33 x 10-5 .
6.69 x 10*
5.23 x lO'5
1.53 x 104
LlOxlO"5
9.80 x 104
6.69 x 10-*
Source
a
a
a
a
a
a
b
b
a
a
b
b
b
b
b
a
a
a
b
b
b
b
a
Volume III
111-17
External Review Draft
Do Not Cite or Quote

-------
TABLE m-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Butanone, 2-
Butylbenzylphthalate
Carbon disulfide
Carbon tetrachloride
Chlordane
Chloro-3-methylphenol, 4-
Chloroaniline, p-
Chlorobenzene
Chlorobenzilate
Chloroethane
Chloroform
Chloromethane
Chloronaphthalene, beta-
Chlorophenol, 2-
Chlorodiphenyl ether, 4-
Chrysene
Cresol, m-
Cresol, o-
Cresol, p-
Crotonaldehyde
Cumene
2,4-D
4,4'-DDE
Emission Rate (g/sec)
Average
S.HxlO-5
5.50 x 10*
8.91 x l(rs
1.58 x 1O4
5.50 x 10-7
6.69 x 10*
6.69 x 10*
5.50 x 10*
3.68 x 10-5
4.90 x 104
2.66 x 104
2.45 x 104
6.69 x 10*
5.50 x 10*
6.69 x 10-6
5.50 x 10*
5.50 x 10*
5.50 x 10*
5.50x10*
1.39 x 1O4
5.50 x 10*
3.88 x 10-5
5.50 x 10-7
High-end
7.40 x 10-5
l.lOxlO-5
9.46 x 10-5
2.75 x 104
1.10x10*
6.69 x 10*
6.69 x 10*
l.lOx ID"5
3.68 x 10-5
9.80 x 104
4.07 x 1O4
4.90 x 1O4
6.69 x 10*
l.lOxlO-5
6.69 x 10*
l.lOxlO-5
l.lOx 10-5
l.lOx 10-5
l.lOx 10-5
1.39 x 1O4
l.lOx 10-5
3.88 x 10-5
1.10x10*
Source
b
b
b
b
b
a
a
b
a
b
b
b
a
b
a
b
b
b
b
a
b
a
b
Volume III
111-18
External Review Draft
Do Not Cite or Quote

-------
TABLE m-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Dibenz(a,h)anthracene
Dibenzo(a,h)fluoranthene
Dibromochloromethane
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine, 3,3'-
Dichlorobiphenyl
Dichlorodifluoromethane
Dichloroethane, 1,1-
Dichloroethane, 1,2-
Dichloroethene, 1,1-
Dichloroethene, trans- 1,2-
Dichlorophenol, 2,4-
Dichloropropane, 1,2-
Dichloropropene, cis-1,3-
Dichloropropene, trans- 1,3-
Diethylphthalate
Dimethoxybenzidine, 3,3'-
Dimethylphenol, 2,4-
Dimethylphthalate
Di-n-butylphthalate
Di-n-octyl phthalate
Emission Rate (g/sec)
Average
5.50 x 10-6
5.50 x 10*
2.63 x 10-5
5.50 x 10-6
5.50 x 10-6
5.50 x 10-6
3.33 x 10-5
4.68 x lO"8
2.45 x 1O4
1.25 x lO'5
1.25 x ID"5
1.25 x lO"5
1.25 x lO"5
5.50 x 10-*
1.25 x 1C'5
1.25 x 10-5
1.25 x 10-5
1.69 x lO"5
1.15 x 1O4
5.50 x 10-6
5.50 x 10-*
1.57 x 10-5
5.50 x lO^6
High-end
LlOxlO-5
l.lOxlO-5
2.63 x 10-5
l.lOxlO-5
l.lOxlO-5
l.lOxlO-5
3.33 x ID'5
8.22 x 10-*
4.90 x 104
2.50 x 10-5
2.50 x 10-5
2.50 x lO'5
2.50 x 10-5
l.lOxlO-5
2.50 x lO'5
2.50 x lO'5
2.50 xlO-5
3.60 x lO'5
1.15x 1O4
l.lOxlO'5
l.lOxlO-5
2.04 x ID'5
LlOxlO-5
Source
b
b
a
b
b
b
a
b
b
b
b
b
b
b
b
b
b
b
a
b
b
b
b
Volume III
111-19
External Review Draft
Do Not Cite or Quote

-------
TABLE m-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Dinitrotoluene, 2,6-
Dinitro-2-methylphenol, 4,6-
Dinitrophenol, 2,4-
Dinitrotoluene, 2,4-
Dioxane, 1,4-
Ethyl methacrylate
Ethylbenzene
Ethylene dibromide
Ethylene oxide
Ethylene thiourea
Fluoranthene
Fluorene
Formaldehyde
Furfural
Heptachlor
Heptachlorobiphenyl
Hexachlorobenzene
Hexachlorobiphenyl
Hexachlorobutadiene
Hexachlorocyclohexane, gamma-
(Lindane)
Hexachlorocyclopentadiene
Hexachloroethane
Emission Rate (g/sec)
Average
5.50 x 10-*
5.50 x 10-6
5.50 x 10*
5.50 x 10*
4.94 x 104
2.45 x 1O4
4.98 x 104
l.lSxlO4
3.05 x 10-5
1.46 x 10-10
5.50 x 10-6
6.69 x HT6
6.07 x 104
5.50 x 10-6
5.50 x 1C"7
1.40x10*
5.50 x 10*
1.40x lO'8
1.01 x 1O4
5.48 x lO'5
5.50 x KT6
5.50 x 10^
High-end
l.lOxlO-5
l.lOxlO'5
l.lOxlO-5
i.ioxio-5
4.94 x 104
4.90 x 1O4
7.53 x 104
l.lSxlO4
3.05 x 10-5
1.46x 10-10
l.lOxlO-5
6.69 x ICT6
6.07 x 19*
l.lOxlO-5
l.lOxlO6
2.80 x 10-8
l.lOx 10-5
2.80 x 10-8
1.01 x 1O4
5.48 x 10"5
l.lOxlO-5
l.lOxlO-5
Source
b
b
b
b
a
b
b
a
a
a
b
a
a
b
b
b
b
b
a
a
b
b
Volume III
111-20
External Review Draft
Do Not Cite or Quote

-------

TABLE m-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Hexachlorophene
Hexanone, 2-
Indeno(l ,2,3-cd)pyrene
Isophorone
Maleic hydrazide
Methoxychlor
Methylene chloride
Methylnaphthaiene, 2-
Methyl-tert-butyl ether
Methyl-2-Pentanone, 4-
Monochlorobiphenyl
Naphthalene
Nitroaniline, 2-
Nitroanilhie, 3-
Nitroaniline, 4-
Nitrobenzene
Nitrophenol, 2-
Nitrophenol, 4-
N-Nitroso-di-n-butylamine
N-Nitroso-di-n-propylamine
N-Nitrosodiphenylamine
Nonachlorobiphenyl
Octachlorobiphenyl
Emission Rate (g/sec)
Average
3.20 x 10-5
6.43 x 10-5
5.50 x 1O*
6.69 x 10-6
1.15x 1O4
5.50 x ID"7
3.96 x 1O4
4.18 x 10-5
1.25 x 10-5
1.25 x 1O5
1.67 x 10-8
5.50 x 10-6
6.69 x 10-6
6.69 x 10*
6.69 x 10*
5.50 x 1O*
6.69 x 1O6
5.50 x 10*
1.21 x 1O4
6.69 x 10*
6.69 x 1O*
1.40x10*
1.40x 1O8
High-end
3.20 xlO5
6.43 xlO'5
LlOxlO5
6.69 x 10*
l.lSxlO4
1.10x10*
6.19X1O4
4.18X10'5
2.50 x 1O5
2.50 x 1O5
2.99 x 10-8
l.lOxlO'5
6.69 x 1O6
6.69 x 1O*
6.69 x 10*
l.lOxlO5
6.69 x 1O*
l.lOxlO5
1.21 x 1O4
6.69 x 1O*
6.69 x 10-6
2.80 x 1O8
2.80 x 1O8
Source
a
a
b
a
a
b
b
a
b
b
b
b
a
a
a
b
a
b
a
a
a
b
b
Volume III
111-21
External Review Draft
Do Not Cite or Quote

-------
TABLE ffl-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Pentachlorobenzene
Pentachlorobiphenyl
Pentachloronitrobenzene
Pentachlorophenol
Phenanthrene
Phenol
Pyrene
Safrole
Styrene
Tetrachlorobiphenyl
Tetrachloroethane, 1,1,1,2-
Tetrachloroethane, 1,1,2,2-
Tetrachloroethene
Tetrachlorophenol, 2,3,4,6-
Toluene
Trichloro-l,2,2-trifluoroethane, 1,1,2-
Trichlorobenzene, 1,2,4-
Trichlorobiphenyl
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Trichloroethene
Trichlorofluoromethane
Trichlorophenol, 2,4,5-
Emission Rate (g/sec)
Average
4.76 x 10-5
1.40x 10-*
3.37 x 10-5
5.50 x 10*
6.69 x 10*
5.50 x 10*
5.50 x 10-*
USxlO4
2.25 x 10-5
1.40 x 10-8
5.50 x 10*
5.50 x 10*
5.13xlO-5
6.80 x 10*
6.13x10-*
3.30 x 104
5.50 x 10-6
3.02 x ID"8
1.25 x \OS
1.25 x ID"5
1.86 x 10-5
2.45 x 104
5.50 x 10*
High-end
4.76 x 10-5
2.80 x 10-8
3.37 x 10-5
1.10 x 10-5
6.69 x 10*
l.lOx 10-5
i.iox ler5
1.15 x 104
4.04 x 10-5
2.80 x 10-*
l.lOxlO-5
l.lOxlO"5
8.02 x 10-5
6.80 x 10*
1.03 x 1C'3
3.30 x 1O4
l.lOx 10-5
5.80 x lO"8
2.50 x 10-5
2.50 x 10-5
3.09 x 10-5
4.90 x 1O4
1.10 x 10-5
Source
a
b
a
b
a
b
b
a
b
b
b
b
b
a
b
a
b
b
b
b
b
b
b
Volume III
111-22
External Review Draft
Do Not Cite or Quote

-------
TABLE m-3 (continued)
Estimated Average and High-end Emission Rates for Products of Incomplete
Combustion (PICs) and Residues of Organic Compounds
Substance
Trichlorophenol, 2,4,6-
Vinyl acetate
Vinyl chloride
Xylene, m-°
Xylene, o-
Xylene, p-c
Emission Rate (g/sec)
Average
5.50 x 10-6
6.43 x ID'5
2.45 x 104
3.80 x 1CT4
5.50 x 10-6
3.80 x KT4
High-end
l.lOxlO"5
6.43 x lO'5
4.90 x KT4
5.64 x 1O4
l.lOxlO-5
5.64 x 10-4
Source
b
a
b
b
b
b
Notes:
* Emission rate based on March 1993 and February 1994 trial burn results and waste profile information.
In these cases, the average and high-end estimates are the same because the estimation method used in
this process results in a high-end estimate, which is conservatively assumed to apply to the average case
as well.
b Emission rate based on August 1994 PIC testing results.
c The emission rate for the mixed isomer "m/p-xylene" estimated from the August 1994 PIC testing is
conservatively assumed to apply to both m-xylene and p-xylene.
Volume III
111-23
External Review Draft
Do Not Cite or Quote

-------
                                    TABLE ffl-4
   Compounds Anticipated to be Emitted in Very Low Quantities for which Emission
                              Rates Are Not Developed
                 Acrolein                            Dinitrobenzene,  1,3-
               Benzaldehyde                          Dinitrobenzene,  1,4-
              Benzo(e)pyrene                     Hexachlorocyclohexane, alpha-
            Benzo(j)fluoranthene                  Hexachlorocyclohexane, beta-
              Benzyl chloride                              Hexane, n-
                 Biphenyl                                Hexanone, 3-
            Bromochloromethane                       Methylene bromide
               Bromoethene                                Phosgene
              Butadiene,  1,3-                           Propionaldehyde
           Chloroacetophenone, 2-                           Quinoline
             Chloropropane, 2-                              Quinone
       Dibromo-3-chloropropane, 1,2-              Tetrachlorobenzene,  1,2,4,5-
         Dichloro-2-butene, cis-1,4-                        Toluidine, o-
        Dichloro-2-butene, trans-1,4-                       Toluidine, p-
           Dichlorofluoromethane                    Trichloropropane,  1,2,3-
            Dinitrobenzene, 1,2-
 Note:

 These compounds were on original list of possible organic constituents of concern, but
 not reported in Waste Profiles or analyzed for in WTI stack emissions.
                                                                External Review Draft
Volume III                              111-24                   Do Not Cite or Quote

-------

Metal
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
TABLE m-5
Estimated Average Metal Emission Rates
Measured SRE
(percent)
NA (99.99932')
99.986
99.977
NA (99.977b)
99.9907
99.987
99.99932
NA (99.977b)
99.99
0°
NA (99.977b)
99.68
NA (99.977")
NA (99.977")
NA (99.977b)
Feed Rate
(Ib/hr)
140
0.24
1.3
5.3
0.0028
0.96
0.83
3.2
3.4
0.011
0.17
1.2
0.52
1.7
4.2
Feed Rate
(g/sec)
18
0.030
0.16
0.67
0.00035
0.12
0.10
0.41
0.44
0.0014
0.022
0.15
0.065
0.15
0.54
Emission Rate
(g/sec)
2.4 x 10-4
4.2 x 10^
3.7 x lO'5
1.5 x 104
3.3 x lO'8
1.6 x 10-5
7.1 x lO'7
9.4 x 10'5
4.3 x 10-5
1.4 x 10-3
5.0 x 10-6
4.7 x 10-4
1.5 x 10'5
3.4 x 10-5
1.2 x 1O4
Notes:
System removal efficiency (SRE) determined from March 1993 trial burn
(ENSR 1993)
NA: Not applicable;. SRE not determined in March 1993 trial burn (ENSR 1993).
* Estimated based on chromium SRE.
b Estimated based on arsenic SRE.
c Assumed to be zero although very low, non-zero efficiency was measured.
Volume III
111-25
External Review Draft
Do Not Cite or Quote

-------
TABLE ffl-6
Estimated Average Acid Gas and Particulate Matter Emission Rates
Substance
Hydrogen Chloride
Nitrogen Oxides
Sulfur Dioxides
Particles
Average Emission Rate
(g/sec)
0.032
2.4
0.091
0.07
Source: WTI monitoring data; February 23, 1995 to March 21, 1995.
Volume III
111-26
External Review Draft
Do Not Cite or Quote

-------
                                                      o
TABLE UI-7
Key Assumptions for Chapter in
Assumption
All stack chemicals of potential concern have been
identified and included
Emission rates are estimated based on performance tests
and trial burns and not on long-term emissions data
Average emission rates for the 95th UCL of the individual
dioxin and furan congeners are based on the arithmetic
mean of 26 post-ECIS runs
Non-detected chemicals equal one-half the detection limit
when estimating average emission rates and equal the
detection limit when estimating high-end emission rates
Long-term PIC emission rates are based on the arithmetic
mean of seven runs from the August 1994 sampling
Basis
The list of COCs have been developed from U.S. EPA
guidance documents and stack testing during trial burns and
performance tests. Additional substances have been added
based on peer review committee recommendations.
Long-term data are not available because the facility has
only had limited operation. The trial burn data are derived
from subjecting the incinerator to extreme conditions not
encountered on a regular basis. The results from the trial
burn and performance test indicate a decrease in emissions
of dioxins/furans over the 1-2 year operating period.
The use of average data when a declining trend is apparent
overstates long-term emissions.
U.S. EPA (1989 a,b) guidance was relied upon for the
average case and a conservative assumption based on
professional judgment was used for the high-end case
The uncertainty of emission estimates decreases as the
number of data points increases. The assessment uses all of
the data that are available.
Magnitude
of Effect
low
medium
low
low
medium
Direction
of Effect
underestimate
unknown
overestimate
overestimate
unknown
Volume HI
111-27
External Review Draft
Do Not Cite or Quote

-------
TABLE III-7 (continued)
Key Assumptions for Chapter III
Assumption
For PICs not analyzed for during the August 1994 testing,
emission rates are estimated based on:
Earlier trial burns. These burns are less
representative than performance tests of actual
operating conditions
The feed rate and a worst-case DRE. The feed
rate is developed from waste profile sheets for
the first year of operation and the DRE is
calculated for three POHCs
The maximum of the estimated emission rates is
used.
This estimated emission rate is used as both the high-end
and average emission rate
If no emission rate can be estimated for a chemical, the
chemical is dropped from consideration assuming it is not
emitted at significant levels
Metals emissions are estimated from trial burns, one year
of waste feed data, and thermodynamic modeling
The trial burn during which metal SREs were calculated
was conducted prior to installation of the ECIS. These
SREs are used to estimate metal emission rates.
Basis
The estimation method includes conservative assumptions so
emission rates are not likely to be underestimated.
The chemicals that have been dropped are not likely to be
emitted in significant quantities, if at all, and are considered
in evaluating the uncharacterized fraction.
Best available data. Professional judgment based on a
review of information on facility design and operation, and
predicted waste characteristics.
The ECIS is not designed to appreciably reduce metal
emissions, so SREs measured pre-ECIS should be similar to
post-ECIS. An exception may be mercury, for which
removal may be enhanced by the ECIS.
Magnitude
of Effect
low
low
medium
low
Direction
of Effect
overestimate
underestimate
unknown
overestimate
Vo'Mtne III
 ?   \
External Review Draft

Do Not Cite or P-\te

-------
                                                         TABLE IU-7  (continued)
                                                    Key Assumptions  for Chapter III
                     Assumption
                        Basis
Magnitude
 of Effect
Direction
of Effect
 The thermodynamic modeling to describe metal behavior
 contains several assumptions including:

           PbC14 and CrO2C12 are excluded
           All metals that vaporize subsequently condense
           to form particles with diameters that are 0.5/im
           All pumpable waste is entrained and no ash
           from non-pumpable waste is entrained
           Little condensation occurs onto the surface of
           larger entiained particles
Professional judgment was relied upon based on a review of
information on facility design and operation, and predicted
waste characteristics.
     low
 unknown
 As part of the modeling, the following assumptions are
 applied:

           Cr is an appropriate surrogate for Al
           As is an appropriate surrogate for Ba, Cu, Ni,
           Ag, Tl, Zn
           SO2 is an appropriate  surrogate for Se
           Hg has a zero SRE
The surrogates were selected based on the expected behavior
of the metals, which are a function of their
physical/chemical characteristics. The trial burn results
indicate SREs in a relatively narrow range for most metals,
so these assumptions are not expected to have a significant
effect.
     low
  unknown
 The metals feed rates are prorated to account for the
 maximum heat input of the incinerator.
Conservative assumption. Professional judgment was relied
upon based on a review of information on facility design and
operation, and predicted waste characteristics.
     low
overestimate
 HC1 emission rates are based on the average of 12 trial
 burn runs
Trial burns were conducted at elevated total chlorine feed
rates producing conservative estimates of HC1 emission
rates.
     low
overestimate
 Metals (other than Hg) are emitted from the stack in
 particle form
Metals are generally non-volatile and those that volatilize in
the high temperature of the rotary kiln will condense to form
aerosols in the cooler, later stages of the incineration
process.
     low
  variable
Volume HI
             111-29
       External Review Draft
       Do Not Cite or Quote

-------
TABLE HI-7 (continued)
Key Assumptions for Chapter in
Assumption
The vapor/particle partitioning equation (Equation III-3)
and associated constants (ST, c, and AS/R) are used to
describe the partitioning phenomena, with the exception of
chemicals with low volatility such as metals and some
PAHs.
Compounds partition either via surface area distribution or
mass distribution, but not in combination.
Assumptions regarding metal modeling:
Thermodynamic equilibrium is maintained
throughout the incineration and flue gas
cleaning system
All important compounds are present in the
thermodynamic data base
The reactor outlet temperature is the
temperature to which the metals are exposed
Chromium is assumed to exist completely in the hexavalent
state
Products with low vapor pressures are not likely to exist in
the vapor phase
Basis
U.S. EPA guidance indicates that the model is a reasonable
estimation of partitioning of organics in air. However, field
data show that the model may under-or overpredict
vapor/particle partitioning for PAHs and organochlorines.
Simplifying assumption based on theoretical considerations.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Conservative assumption because hexavalent Cr is more
toxic than trivalent.
Professional judgement based on a review of information on
facility design and operation, and predicted waste
characteristics.
Magnitude
of Effect
nigh
low
medium
low
high
Direction
of Effect
variable
variable
unknown
overestimate
unknown
-rne III
V
External Review Draft
Do Not Cite or o   te

-------
  Metals In
  Pumpabte
   Wastes
  Metals In
Non-Pumpable
   Wastes
                            ROTARY KILN
SCO
  INITIAL
QUENCHING
  ZONES
                                                                                             APCE
                                                         STACK
                                                                                                    Fly Ash
                                                                                                (Condensed
                                                                                                 Vapors)
                                                                                      Residual Vapors
Volume
III
PATHWAYS
AVAILABLE
FOR
TOXIC
METALS
IN
THE
WTI
INCINERATOR
External ^Review Draft
Figure
111-1

-------
                           IV.  FUGITIVE EMISSIONS
A. Potential Emission Sources
    During normal operations, atmospheric emissions may occur from a variety of sources
other than the incinerator stack. These releases, which are collectively termed "fugitive"
emissions hi this risk assessment, typically involve waste unloading, processing and storage,
and the handling of ash generated by the incinerator.  Potentially significant fugitive emission
sources at the WTI facility are identified based on information hi the WTI facility permit
application (WTI 1982), the facility permit (U.S. EPA 1983), an evaluation of the types of
waste handled at each stage of facility operations, and a facility site visit.  Through this
process, five potentially significant fugitive emission sources are identified at the WTI
facility for evaluation hi the risk assessment:

    •   The carbon adsorption bed (CAB) system, which receives organic vapors vented
         from tanks hi the organic waste tank farm, from operations hi the container
         processing building,  and other potential sources of fugitive organic compound
         emissions.  Vapors from the storage tanks are collected and vented to  the incinerator
         when it is operating, and to the CAB when the incinerator is not operating.
         Emissions from the container processing building may result from such activities  as
         sampling of waste-containing drums, puncturing of drums prior  to incineration and
         pumping of liquids from drums.  Fugitive emissions within the container processing
         building are collected via localized hoods and vented to the incinerator when it is
         operating, and  to the CAB when the incinerator is not operating.

    •    Seals, valves, and flanges associated with storage and process tanks inside the
         organic waste tank farm building. Although WTI has a program hi place to detect
         and repair leaking valves and flanges (WTI 1982), some leaking may be expected
         over the life of the facility.

    •    The wastewater holding tank that collects storm water runoff from process areas
         within the facility. There are three different wastewater systems at WTI:  A, B,  and
         C.  System "A" collects non-contaminated storm water from such areas as roofs and
         the employee parking lot, and the water is discharged directly to the Ohio River.
                                                                  External Review Draft
Volume III                              IV-1                      Do Not Cite or Quote

-------
        System "B" collects storm water from "nonactive" process areas such as sumps and
        plant roadways where contamination is possible but not normally expected.  "B"
        water is retained in three 200,000-gallon tanks and is tested prior to discharge.
        System "C" collects water from active process areas such as diked tank areas,
        washdowns, and other areas where some contact with hazardous waste can
        reasonably be expected.  "C" water is stored in one 250,000-gallon, open-top tank
        prior to treatment.  This tank may be a source of fugitive emissions.

    •   The on-site truck wash station.  The truck wash station is a building covered and
        enclosed on two sides and open to the atmosphere on the ends.  Waste trucks can
        drive through the building for washing, during which volatile releases may occur.
        This station is presently used approximately once per calendar year, but could
        potentially be used more frequently  hi the  future.  In the risk assessment, the wash
        station is assumed to be continuously operated.

    •   Routine fugitive ash releases which  may be associated with the bag filter that is used
        to control emissions during the loading of flyash from the electrostatic precipitator
        (ESP) into trucks.

    The locations of the four organic vapor emissions sources and the fugitive ash emissions
source are shown in Figure JTV-1.

B.  Substances of Potential Concern in Fugitive  Emissions

    1.  Substances of Potential Concern in Fugitive Organic Vapor Emissions
        As discussed in Chapter II of this volume, wastes received at the WTI facility may
    be either pumpable or non-pumpable.  Pumpable wastes typically have the highest
    volatile content, and thus represent the most  significant source of fugitive vapor
    emissions.  The composite pumpable waste stream contains more than 300 chemicals,
    which are ranked based on estimated annual  feed (Ibs/year) calculated using the waste
    profile sheet information.  To focus the  fugitive emissions analysis on the substances
    most likely to be present in vapor releases during normal operations, this list is truncated
    to include only those chemicals with the largest annual volumes (i.e., the chemicals
    which taken together constitute 90 percent of the total mass of pumpable waste).  This
    value (90 percent) provides an approximate two order of magnitude difference between
    the compounds with the highest and lowest feed rates on the truncated list.  If
    compounds with feed rates an additional order of magnitude lower were included, this

                                                                   External Review Draft
Volume III                                IV-2                     Do Not Cite or Quote

-------
    would result in almost 130 additional compounds, and would account for 99.8 percent of
    the total pumpable feed.  Evaluating over 220 chemicals would not have focused the
    fugitive emissions assessment to a significant degree.  The resulting 96 compounds
    represent the substances of potential concern for the vapor fugitive emissions analysis,
    and are shown in Table IV-1.

    2.  Substances of Potential Concern in Fugitive Ash Emissions
        The combustion of waste materials typically results hi the production of solid
    residues (i.e., ash).  Fugitive particle emissions may result from the subsequent
    collection, handling, and disposal of this ash. The solid incinerator residue of greatest
    concern with respect to fugitive emissions is the flyash, collected by the  ESP in the air
    pollution control system because it is produced in relatively large quantities, generally
    has a very fine consistency and thus is subject to atmospheric entrainment,  and contains
    potentially hazardous metals (WTI 1995).
        Substances of potential concern associated with fugitive flyash emissions are
    identified based  on chemical analyses conducted by WTI (WTI 1995).  In 1994, monthly
    samples of flyash were collected from the ESP at the WTI facility. The samples of
    flyash were analyzed for  80 volatile and semi-volatile organic compounds,  total and
    amenable cyanide5, and 9 metals. None of the 80 organic compounds  were detected in
    any of the  12 flyash  samples, and thus organic compounds are not identified as
    substances  of potential concern in fugitive ash emissions.  The metals that were detected
    hi at least one sample of  ash are selected as substances of potential concern and are
    identified hi Table IV-1.  None of the 12 samples included detectable levels of all 9
    metals. Total cyanide was also detected hi  the flyash samples, and thus is  selected as a
    substance of potential concern.

C.  Development of Fugitive Emission Rates

    1.  Tank-Related Emissions from the  CAB System
        Emission rates from  the tanks hi the tank farm that are vented to the CAB system
    are estimated using U.S.  EPA tank calculation program (TANKS2). The TANKS2
    program uses physical/chemical properties of the waste constituents, such as molecular
    weight, vapor pressure over a range of temperatures, and concentration in the waste, hi
    deriving  emission rates.  In estimating the overall  composite physical/chemical properties
    5    Amenable cyanide is subject (or amenable) to chlorination, and is the most toxic
        form of cyanide.

                                                                  External Review Draft
Volume III                               PV-S                      Do Not Cite or Quote

-------
    of the entire waste stream, data for the 12 constituents expected to be present in the
    highest volume as determined from the waste profile for the first year of operation, is
    used.  Thus, waste feed properties are assumed to be reflective of the 12 constituents
    that comprise approximately 60 percent of the pumpable waste feed.  The total waste
    feed throughput to the tank farm is based on the maximum heat input rate (121 million
    BTU/hr HHV) of the incinerator (ENSR 1993).
        Contributions from the container processing building to the CAB system are based
    on the estimated number of drums received at the facility (=45,000/year), the number
    of drums sampled (»4,500/year), the number of drums repackaged ( = 16,000/year), and
    the container processing rate (as listed in Appendix III-l). Emissions are assumed to be
    equivalent to releases from a leaky valve with heavy liquids, and calculated using  U.S.
    EPA emissions factors (U.S. EPA 1992a).
        Vapor emissions from the tanks (and the container processing building) are typically
    vented to the incinerator for combustion, or to the CAB system when the incinerator  is
    not operating.  Based on WTI's first year of operation, the incinerator operated S3
    percent of the time.  It is assumed that the CAB system effectively contrails 90 percent
    of the organic vapor emissions from the tanks and the container processing building,
    based on average efficiency data that has been compiled for carbon systems (U.S.  EPA
    1992a).

    2.  Other Organic Fugitive Emissions
        Total emissions from the seals, flanges, and other sources of vapor leaks in the tank
    farm are estimated by summing emissions from the individual facility components and
    sources  (i.e., pump seals, in-line valves, and flanges) based on emissions factors for
    these sources (U.S. EPA 1992a).  These vapors would be released through four vents on
    the tank farm building.
        Total organic vapor emissions from the wastewater tank are calculated using mass
    transfer correlations and emission equations  developed for wastewater treatment systems
    (U.S. EPA 1992a).  For purposes of these calculations,  the design average throughput of
    15,000 gal/day from the WTI permit application (WTI 1982)  is used. The wastewater
    tank emission equations require physical/chemical properties data for the chemical
    substances in the wastewater to develop a total mass emission rate.  It is not possible to
    input chemical-specific parameters for all the substances that may be present in the
    wastewater;  therefore, chemical-specific  parameters for toluene are used to represent
    total VOC behavior. Toluene is selected because it is one of the top ten constituents by
    weight in the waste feed accepted by the facility based on waste profile sheets maintained
    by WTI, and because its critical physical/chemical properties  are representative of the

                                                                  External Review Draft
Volume III                               IV-4                     Do Not Cite or Quote

-------
    other substances received in large quantities in the pumpable wastes. Toluene has a
    similar molecular weight, vapor pressure, aqueous solubility,  and other
    physical/chemical parameters as other major components in the waste feed.  In addition,
    toluene is the constituent projected to be received in the third largest quantity (4.0% of
    the pumpable waste); other chemicals projected to be received in large quantities are
    hydrocarbon, unspecified (16.8%), cresol (5.2%), methyl ethyl ketone (3.5%), methanol
    (3.1%), and acetone (2.9%).  Total organic waste emissions from truck washing are
    based on emissions factors developed by U.S. EPA (1992a) for releases of heavy and
    light liquids from valves.
        Total organic vapor emission rates for the four sources of fugitive organic emissions
    are listed  hi Table IV-2.  As shown hi Table IV-2, fugitive organic vapor emissions
    from the tank farm building, wastewater tank, and truck wash are estimated  to be 2,126
    Ibs/yr, 202 Ibs/yr, and 9.9 Ibs/yr, respectively.   The estimated 224 pounds of organic
    vapor emitted annually from the CAB system are divided among the four types of tanks
    in the tank farm (i.e., blending, holding, pumpout, and reception) and the container
    processing building.  Of the 224  pounds of emissions from the CAB, the tanks hi the
    tank farm and the container processing building contribute 212 pounds and 12 pounds of
    organic vapors, respectively.

    3.  Fugitive Ash Handling Emissions
        The flyash generation rate at the WTT facility is estimated to be 5,300 tons per year,
    based on the estimated flyash generation rate in WTI's permit application (WTI 1982).
    Actual flyash emissions from calendar year 1994 totaled approximately 4,000 tons
    (Victorine 1995).  Fugitive emissions of flyash from the ESP may occur during transfer
    of ash into covered trucks prior to disposal.  Emissions generated during the loading
    process are controlled by a fabric filter, with a fraction of the flyash escaping capture as
    fugitive emissions.  An uncontrolled ash emissions factor of 0.107 Ib/ton flyash dial was
    empirically developed by Midwest Research Institute (Muleski et al. 1986) from field
    testing a coal-fired power plant equipped with an ESP is used, assuming that the flyash
    from coal burning and hazardous waste combustion are similar. This empirical
    emissions factor, however, is developed for  flyash with an average moisture content of
    29 percent (moisture was added to the flyash for control). Since the flyash at WTI
    would be  expected to have a negligible moisture content (due  to its temperature above
    the boiling point of water and the absence of moisture addition), the emissions factor is
    increased  by a factor of ten to 1.07 Ib/ton based on the assumed relationship between
    water content and credibility.  In addition, it is  assumed that the control efficiency of the
    fabric filter that controls emissions during truck loading would reduce the  uncontrolled

                                                                   External  Review Draft
Volume III                                IV-5                     Do Not Cite or Quote

-------
    emissions by 99.5 percent (U.S. EPA 1992a). The resulting ash emission rate based on
    these factors is estimated to be 28 Ib/yr.

D.  Key Assumptions for Fugitive Emissions
    The key assumptions used in analyzing fugitive emissions at the WTI facility are
summarized in Table FV-3.  This table indicates the basis for the assumptions listed, the
estimated relative magnitude of the assumptions' effect on the overall risk assessment, and
the direction of the effect, if known.  These assumptions are further discussed in Chapter V
of this report.
                                                                  External Review Draft
Volume III                               IV-6                     Do Not Cite or Quote

-------

                                                           TABLE IV-1
                                          Fugitive Substances of Potential Concern
                                                  Fugitive Organic Vapor Emissions*
                Acetone
                Acetonitrile
                Acetophenone
                Acetylaminofluorene, 2-
                Acrylonitrile
                Alcohols
                Aliphatic hydrocarbons
                Aniline
                Benzene
                Benzenedicarboxylic acid, 1,2-
                Benzidine
                Benzoquinone, p-
                Benzo(a)pyrene
                Butanol
                Butanone, 2-
                Butyl acetate
                Calcium chromate
                Carbon
                Carbon disulfide
                Carbon tetrachloride
                Chlorobenzene
                Chloroform
                Chlorinated paraffin, oil, wax
                Chrysene
                Creosote (coal tar)
                Cresol
                Crotonaldehyde
                Cumene
                Cyclohexane
                Cyclohexanone
                Dibenz(a,h)anthracene
                Dibromoethane, 1,2-
Dicnlorobenzene
Dichlorodifluoroethane
Dichlorodifiuoromethane
Dichloroethane, 1,1-
Dichloroethene
Diethyl stilbestrol
Diethylphthalate
Dimethyl sulfate
Dimethylamine
Dimethylbenzidine, 3,3'-
Dimethylhydrazine
Dimethylphenol, 2,6-
Dimethylphthalate
Dinitrotoluene
Dioxane, 1,4-
Epichlorohydrin
Ethanol
Ethoxyethanol, 2-
Ethyl acrylate
Ethylbenzene
Fluoranthene
Formaldehyde
Formic acid
Furfural
Heptane
Hydrazine
Indeno(l ,2,3-cd)pyrene
Isobutanol
Isopropanol
Isosafrole
Maleic anhydride
Methanol
Methyl metnacrylate
Methylbutadiene, 1-
Methylcholanthrene, 3-
Methyl isobutyl ketone
•Naphthalene
Naphthylamine,  1-
Naphthylamine, 2- '
Nitrobenzene
Nitrophenol, 4-
Nitropropane, 2-
N-nitrosodiethanolamine
N-nitrosodiethylamine
N-nitrosodi-n-butylamine
N-nitrosopyrolidine
Phenol
Phthalic anhydride
Picoline, 2-
Pyridine
Resorcinol
Tetrachlorobenzene, 1,2,4,5-
Tetrachloroethane, 1,1,1,2-
Tetrachloroethene
Tetrahydofuran
Toluene
Toluene diisocyanate
Toluenediamine
Trichloro-l,2,2,-TFE, 1,1,2-
Trichlorobenzene
Trichloroethane, 1,1,1-
Trichloroethene
Trichlorofiuoromethane
Xylene
                                                       Fugitive Ash Emissions1*
                            Arsenic
                            Barium
                            Cadmium
                              Lead
                              Nickel
                            Selenium
                              Silver
                             Cyanide
                Notes:

                '    Developed from list of "pumpable" waste streams handled by WTI during first year of operation.  These
                     substances represent approximately 90 percent of the total pumpable waste stream.  Pumpable waste
                     stream constituents list developed as described in Appendix ffl-1.
                b    Based on metals detected in fly ash samples collected by WTI (1995).
                TFE -     trifluoroethane
\
              Volume III
              IV-7
          External Review Draft
           Do Not Cite or Quote

-------
TABLE IV-2
Estimated Total Fugitive Organic Vapor Emissions Rates
Source
Tank Farm Building*
Wastewater Tank
Truck Wash Building
Carbon Adsorption Bed System (total)
From Tanks:
Blending
Holding
Pumpout
Reception
From Container Processing
Total All Sources:
Estimated Emissions
Ib/yr (g/sec)
2,126 (3.06 x lO'2)
202 (2.91 x 10-3)
9.9(1.42x10-*)
224.3 (3.23 X 1(T3)
212.2 (3.06 x 10-3)
57.2 (8.23 x 10"1)
57.2 (8.23 x 10"4)
47.7 (6.86 x 10-4)
50.1 (7.21 x 10-*)
12.1 (1.74x10")
2,562 (3.68 x 10"2)
Note:
a Tank farm building emissions are from the leaky valves and flanges which, for the purposes of this risk
assessment, are assumed to all occur from the vents on the roof of this building.
Volume III
IV-8
External Review Draft
Do Not Cite or Quote

-------
TABLE IV-3
Key Assumptions for Chapter IV
Assumption
All fugitive chemicals of potential concern have been
identified and included, even though the list is limited to
pumpable wastes (e.g., nonpumpable waste may also be a
source of fugitive emissions).
The composite liquid waste stream list is truncated to include
only the chemicals in the top 90% by mass.
All fugitive emission sources have been identified.
The 12 monthly flyash samples used to determine the fugitive
chemicals of concern and amounts are representative of
actual conditions. Because organic compounds were not
detected in these samples, they are not considered.
Additionally, the chemicals on the analyte list include all the
chemicals that are likely to be present.
The same chemical composition is used for all fugitive
sources.
The contributions to the CAB system are based on the
estimated number of drums received at the facility, the
number of drums sampled and repackaged, and the container
processing rate.
Emissions from the CAB system are estimated as equivalent
to releases from a leaky valve with heavy liquids using
appropriate emission factors.
The wastewater tank is treated as a non-aerated sump using a
default wind speed and an average throughput. Toluene is
used as a surrogate for VOC behavior.
Emissions from the truck wash are equal to releases of heavy
and light liquids from valves.
The emissions factor for flyash from coal burning is applied
to incinerator flyash emissions but increased by a factor of 10
because of negligible moisture content.
Basis
Non-pumpable wastes are handled separately from pumpable
wastes and because they are generally not volatile, they are
not likely to result in fugitive emissions.
Simplifying assumption to focus assessment.
A site inspection was conducted to identify all significant
sources of fugitive emissions.
Best available data.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on relationship between water
content and erodibility.
Magnitude
of Effect
low
low
low
low
low
low
low
low
low
low
Direction
of Effect
underestimate
underestimate
underestimate
underestimate
unknown
unknown
unknown
unknown
unknown
unknown
Volume III
,V-Q
External Review Draft
Do Not Cite or 0'">te

-------
                  o\acad\013999GM3999GM
                                                                                       &       \
                                                                                         ^"'' ^^ Waste Water   v
                                                                                         Waste Wal.r    Tanki
                                                                                         Treatment
                                                                                           Bldg.
                                                                                                 Truck Unloading
                                                                                                     Station
                                                                                            Organic Waete
                                                                                             Tank Farm
                                                              ^'--"''           «*$
                                                              >
-------
/
\
                   V.  UNCERTAINTY IN EMISSIONS CHARACTERIZATION
                As previously discussed in this volume, an important initial step hi the risk assessment is
            estimating emissions from both the incinerator stack and fugitive sources.  To reduce
            uncertainty in estimating incinerator emissions, a comprehensive stack testing program was
            performed at the WTI facility. However, variabilities hi the stack testing results and waste
            feed composition must still be taken into account in evaluating the results of the stack test.
            Furthermore, predictive models,  some with potentially significant uncertainties, are used to
            supplement the stack testing, and also as the basis for predicting fugitive emission rates.  The
            primary sources of uncertainty associated with the emissions characterization are described in
            the following sections.

            A.  Uncertainties in Stack Emissions Characterization
                The primary uncertainties associated with stack measurements from the WTI facility are
            associated with the sampling and analytical techniques, which are subject to rigorous quality
            assurance/quality control procedures, and differences between the waste feed during the tests
            and  waste feeds that might be received during long-term operation.  Changes in facility
            design or operation may also result in emissions which differ significantly from those
            observed during facility tests.
                Sampling and analysis uncertainty is relatively small given the rigorous quality
            assurance/quality control procedures employed during the testing at WTI.   Some of the
           variability and uncertainty hi the  stack testing has also been captured by repetitive sampling
            that  has been conducted as part of the quarterly performance test during which  the
           constituents of primary concern, PCDD/PCDF, have been measured.
               Variability hi both  metal and organic emissions can occur as a result of changes hi
           operating conditions of the incinerator (e.g., kiln temperature).  However,  the trial burns
           have been conducted to provide conservative measures of emissions.  Emissions of metals are
           generally more sensitive to relatively small changes hi incinerator operating conditions;
           consequently, the following sections provide a summary of the sensitivity of metal emissions
           to several incinerator parameters.
                                                                            External Review Draft
           Volume III                               V-l                     Do Not Cite or Quote

-------
   1.  Uncertainty Associated with Metal Emissions
       Although stack testing performed at the WTI facility reduces the uncertainty
   associated with estimated metal removal rates, the variability hi metal feed rates remains
   an important source of uncertainty.  Furthermore, it is necessary to extrapolate SRE data
   from the seven metals tested in the WTI trial burn (antimony, arsenic, beryllium,
   cadmium, chromium, lead, and mercury) to eight metals not included hi the trial burn
   (aluminum, barium, copper, nickel, selenium, silver, thallium, and zinc.)  The
   uncertainty of such extrapolation is reduced through thermodynamic modeling performed
   to better understand metal behavior in the incinerator. The assumptions used hi the
   modeling exercise,  and sensitivity of the model to the various input parameters, are
   described below.
       The modeling reflects the current state of the art for understanding metals behavior
   hi waste incineration systems.  However, several assumptions inherent hi the model
   fundamentally limit its capability to precisely predict metal emissions.  Some of the most
   important limiting assumptions are:

        •   Thermodynamic equilibrium is maintained throughout the incinerator and  flue
            gas cleaning system. This assumption is required because hi most cases, the
            needed reaction rates are not known. However, hi some cases  it  is  clear that
            equilibrium is not maintained. For example, the equilibrium model predicts that
            at ambient conditions, nearly all chromium present with chlorine  will form a
            volatile hexavalent compound.  However, field measurements confirm that little
            chromium is present hi the stack as hexavalent chromium. The discrepancy
            probably arises because the rate of formation of the hexavalent form of
            chromium becomes very  slow at lower temperatures.  In this example, the
            equilibrium model prediction that all chromium is present hi the hexavalent form
            is conservative because hexavalent chromium is more toxic than other ionic
            forms of chromium. It is possible, however, that assuming equilibrium
            conditions is not conservative for all metals.

        •   The formation of metal complexes will not significantly affect metal emissions.
            Complex reactions between ash components and certain metals are possible.
            The complexes that form may have different volatilities than other forms of the
            metal, and are not generally present hi the thermodynamic data base used hi the
            model.  The potential effects of these complex reactions on predicted emission
            rates are unknown.
                                                                  External Review Draft
Volume III                               V-2                     Do Not Cite or Quote

-------
         •  The reactor outlet temperature adequately characterizes the temperature to which
             the metals are exposed. The temperatures in the incinerator vary significantly,
             and the behavior of the metals, as demonstrated by the sensitivity analysis
             described in the Appendix III-l, can be strongly influenced by temperature.
             However, for most metals of concern, no change hi behavior is expected within
             the typical temperature range of the WTI incineration system.

         These assumptions introduce unquantifiable uncertainties, but are believed to be
     appropriate given the base of information available.  Stack testing from the WTI facility
     suggests that the metal emission rates developed for the risk assessment are
     representative, and are consistent with theoretical modeling considerations.
         It is possible to evaluate the impact of variations in the input parameters used by the
     model on the predicted emissions.  Site-specific input parameters include such data as
     combustion chamber temperature and waste feed composition.  General, non-site-specific
     information  includes  thermodynamic data, gas viscosities, and similar parameters. The
     non-site-specific data are generally more precisely known than the site-specific data,  and
     thus are believed to be less likely to contribute significantly to the overall uncertainty in
     the predictions.  Because of this, a sensitivity analysis is performed to evaluate the
     impact of variations in site-specific input parameters.  The following site-specific
     parameters are used hi the model:

         •   Waste composition and feed rate
                 Chlorine
                 Trace metals
         •   Incinerator system conditions
                 Temperature
                 Availability of oxygen
                 Entrainment
                 Quench  temperature
         •   Control device efficiency
                 Vapors
                 Fine Particles
                 Coarse Particles
                                                                    External Review Draft
Volume III                               V-3                      Do Not Cite or Quote

-------
       The effects of reasonable variations hi waste feed rate, combustion chamber
    temperature, availability of oxygen, waste chlorine concentration, quench temperature
    and entrainment rates are examined hi Appendix III-l. With the exception of waste feed
    rate, each of these parameters will affect the behavior of only a few of the metals.
    However, the impact on the metals affected can result hi changes hi emission rates of
    more than an order of magnitude, as discussed below.

       a)   Waste Composition and Feed Rate
            Three feed rates are examined for each of the 15 metals hi the sensitivity
       analysis:

            •   Base case (maximized heat input feed rates)
            *   High feed rate (one order of magnitude higher than base case)
            •   Low feed rate (one order of magnitude lower than base case)

       When the feed rates of all of the metals of interest are varied by an order of
       magnitude, emission rates change proportionately for each metal,  indicating that this
       variable has a strong impact on predicted emission rates (with the predicted system
       removal efficiencies remaining constant).

       b)   Conditions in Incinerator System
            Four combustion chamber temperatures are examined — 1,000°C, 1,100°C,
       1,200°C and  1,400°C. A temperature of 1,200°C is  used as the base case.  These
       temperatures represent an approximate maximum  variation possible during sustained
       operation given the physical constraints of the system.  Of the 15  metals examined,
       only the emissions of beryllium, copper and nickel are affected by the change hi
       temperature.  At the two lower temperatures, no beryllium vaporizes hi the
       combustion chamber.  The emissions decrease from 1 x l(r7 g/s hi the base case to
       less than 1  x  10"8 g/s at the two lower temperatures. Vaporization of copper and
       nickel is reduced at 1,000°C.  The emissions of both  of these metals is
       approximately one order of magnitude lower at 1,000°C than at any of the other
       temperatures.  At the higher temperatures, beryllium, copper and nickel are
       predicted to vaporize and condense forming fine particles.  At the lower
       temperatures, the metals will remain with the ash.
            To  assess the affect of the availability of oxygen on the predicted metals
       behavior, the effects of two additional values of the primary chamber ah* to waste
       stoichiometric ratio are examined.  In general, varying the stoichiometry of a

                                                                  External Review Draft
Volume III                                V-4                     Do Not Cite or Quote

-------
         combustion system for values around 1.0 has the greatest impact on emission rates.
         As the ratio moves away from 1.0, the incremental effects observed with additional
         changes in the value decrease dramatically.  Typically, most of the variation occurs
         between values of 0.8 and 1.2.  Thus, the effect of these two values of
         stoichiometric ratio on emission rates is examined.
             At the lower oxygen concentration, the vaporization of beryllium is lower than
         in the base case due to the formation of reduced forms that are less volatile than the
         oxide.  The predicted beryllium emissions are approximately one order of magnitude
         lower when the stoichiometrie ratio is 0.8 than when it was 1.0. The predicted
         nickel emissions are about one order of magnitude lower when a stoichiometric ratio
         of 1.2 is used.
             The impact of chlorine is examined.  Emissions are predicted for  a waste that
         does not contain chlorine and the results are compared with the maximum heat input
         case.  Copper, nickel, and selenium are affected by the removal  of all chlorine from
         the system. In the absence of chlorine, copper and nickel no longer vaporize in the
         primary combustion chamber. When the  vaporization in the primary chamber is
         eliminated, the emissions drop by about one order of magnitude.  Selenium, in
         contrast, still vaporizes in the primary chamber when no chlorine is present.
         However,  the vapors will condense in the quench if there is no chlorine present.
         This results in a decrease in emissions by  about  an order of magnitude.
            Quench temperature can have a strong effect on the behavior of some metals.
         To quantify this impact and identify the metals influenced most significantly by the
         quench temperature, two temperatures in addition to the base value of 150°C are
         examined.  While it is unlikely that lower temperatures could be achieved, higher
         temperatures may  result from system failures or  operator errors.  Thus, 200° and
         400°C are selected because they best illustrate the  impact of quench temperature on
         the behavior of the metals of interest. It is extremely unlikely that the system would
         ever operate at these higher temperatures for any extended time.  Cadmium,
         antimony,  thallium, and zinc are affected  by the quench temperature.  As the
         temperature increases,  condensation of these metals decreases.  Emissions increase
         correspondingly due to the low capture efficiency for vapors. Thallium emissions
         are one order of magnitude higher than the base  case for both of the higher quench
         temperatures.  Zinc and cadmium emissions are  one order of magnitude higher than
         the base case at 400°C only.  Antimony emissions are slightly higher than the base
         case at the highest quench temperature.
            The final parameters examined are the entrainment rates.  The  following values
         are selected to represent reasonable extremes based on facility design and operation:

                                                                  External Review Draft
Volume  III                               V-5                      Do  Not Cite or Quote

-------
            •   A reduction of the entrainment of the ash in the pumpable wastes by 50
                percent; and
            •   An increase in the entrainment of the ash in the non-pumpable wastes by 50
                percent.

        Decreasing the entrainment of ash in the pumpable waste is found to result in a
        decrease in emission rates for aluminum and chromium while the rates for other
        metals are unchanged. Increasing the entrainment of the ash in the non-pumpable
        waste causes a small increase in emissions for chromium and aluminum.

        c)   Control Device Efficiency
            The uncertainty in the predicted removal efficiencies can be estimated by
        calculating the standard deviation observed during the three test runs conducted
        during the trial burn. Table V-l summarizes the control efficiencies used and the
        associated standard deviations. Because the control efficiencies are estimated using
        data obtained during the trial burn hi a series of tests repeated over a short time
        using a well controlled simulated waste, it is likely that the values hi Table V-l
        represent  the smallest possible variability.  In operation over long time periods,  it is
        likely that the variation in waste composition, emission values, and perhaps control
        efficiencies would be much greater.

        d)   Results of Sensitivity Analysis
            Based on the values hi Table V-l and the sensitivity study, it is possible to
        determine the potential range of variation that may be expected hi the predicted
        emission rates. Table V-2 summarizes this analysis.  The minimum value is
        produced  by using the model to predict emissions when all the data are at the values
        which produce the lowest predicted emissions. Thus, this prediction uses the lowest
        reasonable entrainment rates, primary chamber temperature, quench temperature,
        chlorine concentration, and metals feed rates. The impact of oxygen concentration
        is relatively small and is  not included in this analysis. In addition, the control
        efficiency is  assumed to be greater by one standard deviation than the average from
        the March 1993 trial burn. One standard deviation is selected as sufficiently
        representative of the probability variations hi the control efficiencies. Variations hi
        control efficiencies are very small compared to the possible variations hi other
        parameters such as metals feed rates and have little impact on this analysis.  The
        maximum value is obtained using a similar technique. The highest reasonable
        entrainment rates, primary chamber temperature, quench temperature, chlorine

                                                                   External Review Draft
Volume III                                V-6                      Do Not Cite or Quote

-------
         concentration, and metals feed rates are used simultaneously.  The control
         efficiencies are assumed to be less by one standard deviation than the 1993 trial burn
         averages.
             Variation in the feed rate accounts for most of the variation observed in the
         predicted emission rates for each metal.  For this analysis, it is assumed that actual
         feed rates could deviate from the maximum heat input feed rate by as much as an
         order of magnitude,  based on the following observations:

             •   The concentration information in the data base used to determine the waste
                 composition is generally based on the results of a single analysis of the
                 waste. Trace metals concentrations in wastes typically vary widely.

             •   The data base incorporates many assumptions about the type and quantity of
                 the wastes which would be available for incineration.

             •   Normalization and other data processing is required to place all data on a
                 common basis.  Such processing adds potential sources of variation.

         Because metal emissions from hazardous  waste incinerators are very dependent on
         the site-specific feed rates of those metals, feed rate limits are  generally established
         under operating permits to control maximum emissions.  Permit limits can be used,
         where necessary, to restrict the upper bound of the metal emission rates and hence
         limit the impact of this kind of uncertainty.
             It should be emphasized that modeling of the behavior of metals in waste
         incinerators is still very approximate.  Current models are best used for predicting
         trends and estimating the impact of changing operating conditions.  They are not as
         successful at precisely predicting metal emission rates.  Furthermore, the model is
         capable of predicting metal species present hi emissions and predicting the phase of
         each metal emitted, as shown in Appendix III-l. However, the results of this
         modeling exercise are not used hi the risk assessment to predict specific metal
         species or phases given the high degree of uncertainty in the model results.

    2.   Uncertainties Due to Uncharacterized Stack Emissions
         Stack gas samples collected during  the trial burn and performance tests were
    analyzed for a list of conventional analytes. A fraction of the sample may contain
    organic compounds such as methane and ethane, or other aliphatic or aromatic
    hydrocarbons, which are  not conventional analytes.  This fraction has been termed the

                                                                   External Review Draft
Volume III                                V-7                      Do Not Cite or Quote

-------
    "uncharacterized fraction."  In some cases, the uncharacterized, non-methane portion of
    the emissions may be considerably larger than the characterized fraction.  In the absence
    of any information on the fate and transport characteristics or toxicity of these unknown
    compounds, there is some uncertainty hi the risk assessment associated with the
    incomplete characterization of emissions from the stack.
        For the purposes of uncertainty analysis in the risk assessment, the amount of
    organic emissions that might not have been measured is estimated using the following
    procedure.  First, the recorded total hydrocarbon (THC) stack emission levels from the
    March 1993 trial burn are evaluated and adjusted upward (by a factor of 2.87) to  account
    for the fact that THC analyzers typically underestimate organic emissions (U.S. EPA
    1988a).  As discussed in Appendix III-l, the following factors account for the
    understatement of THC measurements:  (1) organics potentially condense in the sampling
    line moisture trap; and (2) several compounds are biased low by the flame ionization
    detector within the THC analyzer. A correction is also required to account for the
    difference in molecular weight between the calibrating gas,  methane, and the  actual
    average molecular weights of the organic emissions.  Taking these factors into
    consideration, the resulting adjusted THC value is a measure of the "total organics" in
    the combustion stream.
        The adjusted THC value is then compared to the total characterized fraction PIC
    emissions, as described in Appendix III-l.  The comparison of the characterized fraction
    organic emissions to the adjusted THC indicates that an estimated 60% of the non-
    PCDD/PCDF PIC emissions could potentially remain uncharacterized. Thus, non-
    PCDD/PCDF PIC emissions from the incinerator could potentially be as much as 2.5
    times the measured amounts.
        Within the uncharacterized fraction of the THC,  it is possible that various
    halogenated dioxins and furans may be present. Stack testing conducted during
    performance tests at the WTI  facility has focused on emissions of PCDD/PCDF, i.e., the
    chlorinated dioxins and furans.  There is evidence in the literature that brominated
    dioxins and furans could have comparable toxicities to the chlorinated dioxins and furans
    (Mason et al. 1987).  However, it is anticipated that only a small fraction of the dioxins
    and furans emitted from the WTI  incinerator stack will be brominated, because a review
    of the waste feed profile from the first year of operations indicates that less than five
    percent of the halogenated wastes  are brominated.  Other halogenated dioxins and furans
    would be expected to pose a lower risk than either the chlorinated or brominated forms,
    based on toxicity and waste feed composition.
        In the uncertainty analysis of the risk assessment, the emission rate for                    -
    PCDD/PCDF is multiplied by a factor of 1.5 to conservatively account for brominated

                                                                 External Review Draft
Volume III                                V-8                     Do Not Cite or Quote

-------
     dioxin-like compounds.  This factor is believed to be conservative because brominated
     waste is much less prevalent than chlorinated waste at WTI.  It should be recognized that
     the factor of 1.5 is based primarily on professional judgment, rather than any specific
     data.

     3.   Uncertainties Associated with Chromium
         The oxidation state of chromium can be a critical  issue in characterizing metal
     emissions in a risk assessment.  Chromium compounds can exist  in any of three
     oxidation states:  elemental, trivalent, or hexavalent.  The toxicity of chromium varies
     considerably from one form to another.  Hexavalent chromium is the most hazardous
     form of chromium based on its carcinogenic potential. In contrast, elemental chromium
     and  trivalent chromium have not been shown to be carcinogenic in either humans or
     laboratory animals.  Incinerator test results suggest that chromium can exist in more than
     one  of the oxidation states in stack emissions (U.S. EPA 1990a).
         The importance of hexavalent chromium is considered in U.S. EPA's proposed rules
     for owners and operators of hazardous waste incinerators (U.S. EPA  1990b).  Although
     U.S. EPA states in the proposed rules that both the hexavalent and trivalent forms of
     chromium are expected to be present in incinerator flue gas, the worst-case assumption,
     that  100 percent of the chromium present in the flue gas is in the hexavalent form, is
     adopted hi this risk assessment.

     4.  Uncertainties Associated with Laboratory Contamination
        The possibility of laboratory contaminants must be considered when evaluating
     analytical results, especially for common laboratory contaminants, such  as phthalates
     (ATSDR 1993). In this assessment, six phthalate compounds are considered as
     substances of potential concern, including bis(2-ethylhexyl)phthalate,
     butylbenzylphthalate, diethylphthlate, dimethylphthalate, di-n-butylphthalate, and di-n-
     octylphthalate.  The presence of these compounds has been documented in many types of
     laboratory equipment,  including plastic tubing, plastic gloves, rubber, and teflon sheets
     (U.S. EPA 1988b; Giam et al. 1975), and can lead to  the contamination of
     environmental samples. While efforts are generally made to reduce laboratory
     contamination by thorough cleaning, phthalates are often reported hi many laboratory
    blanks (Thuren  1986; U.S. EPA 1988b).  Consequently, U.S. EPA (1988b) has reported
     that common phthalate esters generally can not be precisely measured at concentrations
    below 2 ppb due to blank contamination.
                                                                  External Review Draft
Volume III                               V-9                     Do Not Cite or Quote

-------
B.  Uncertainties Introduced by Process Upset Emissions
    As discussed in Appendix III-l, a variety of process upsets may be expected during
normal operation of the WTI facility. However, potential process upset emissions are not
included in the risk assessment calculations due to:  (1) the significant uncertainties
associated with characterizing emissions during these short-duration events; (2) the
expectation that the magnitude and duration of such potential emissions would be quite
limited; and (3) the measures in place at WTI to reduce the frequency and impact of such
emissions.  As discussed below, several different situations have been identified that could
lead to process upset emissions, although the magnitude of these releases is not anticipated to
be significant when compared to routine facility emissions or releases during the types of
accidents evaluated in Volume VII.

    •  Plant Startup.  The RCRA permit for the WTI facility prohibits injection of
       hazardous waste to the incinerator until the kiln reaches the minimum temperature
       demonstrated during the trial burn, and until the incinerator is within the operating
       envelope defined by the permit. Therefore, wastes can not legally be introduced
       until the incinerator is within the  complying operating conditions. Similarly,
       planned shutdowns would involve shutting off waste feed and properly burning out
       residue before termination of incinerator operation.  However, unplanned shutdowns
       can occur; some of the potential problems associated with these conditions are
       discussed below.

    •  Emergency Vent Stack Releases.  Unlike most rotary kiln incinerators, the WTI
        facility does not have an emergency relief vent stack.   Instead, an automatically
       activated emergency electrical generator allows continued operation of the induced
        draft (ID) fan (albeit at a lower speed) and all major pumps in the event of a
        complete electrical power failure.  Therefore, no emergency vent stack emissions
        are anticipated.

    •   Interruptions in Water Supply to the Scrubber System. The packed bed and
        venturi scrubbers installed at the WTI facility are used both for HC1 control and
        additional fine particle control (note that the scrubbers follow a 3-stage ESP which
        collects most of the particles).  The  scrubber system would not be expected to
        contribute significantly to the control of organics emissions, but might have a minor
        impact on emissions of those organics which are adsorbed onto the surface of the
        activated carbon injected into the duct work. Therefore,  failure of water supply to
                                                                   External Review Draft
Volume III                                V-10                    Do Not Cite or Quote

-------
        the scrubber would not be expected to have a significant impact on organics
        emissions.

        Under the RCRA permit for the WTI facility,  any failure of the scrubber system
        would trigger an instantaneous automatic waste feed cutoff. Pumpable wastes would
        rapidly cease burning and therefore no longer represent a source of emissions.
        Solid-form wastes, such as drummed wastes and bulk wastes, may continue to burn
        in the kiln for several minutes (up to perhaps 30 minutes).  Therefore, any
        assessment of potential emissions  after an unplanned waste feed cut-off would need
        to consider the potential emissions generated by solid-form wastes which may have
        been charged into the incinerator immediately prior to the process upset event.

        Because solid-form wastes are not atomized but, rather, are burned in a more
        quiescent manner,  these wastes would not be expected to generate as high a particle
        loading in the gas stream or liberate as great a quantity  of metals (if the waste
        contained metals) as die pumpable wastes.  This is because: (1) the atomized liquid
        wastes generally are better dispersed and have more surface area available to absorb
        heat, vaporize metals, and then transfer metals to the gas stream; and (2)
        atomization tends to mechanically generate  a considerable amount of paniculate
        matter, as well as generate more vapors which can subsequently condense/nucleate
        into particulate matter.   In addition, atomization more vigorously disperses both of
        these types of particulate matter hi the combustion stream.  Furthermore,  since die
        vast majority of the mass flow of  particle emissions is caught in the ESP and not in
        the scrubber, even complete failure of the wet scrubber  system would not be
        expected to increase overall particle emissions  dramatically.

        Volatile metals emissions may also be partially collected in the wet scrubber system;
        therefore, emissions of such metals might be expected to rise slightly for the first 5
        to 30 minutes of a scrubber water failure event, if 1) solid-form waste continued to
        burn in die kiln after a scrubber failure, 2)  die solid waste happened to contain a
        significant quantity of volatile metal(s),  and 3) the metal species were predominantly
        in a soluble vapor form.  But since solid-form wastes  generally liberate considerably
        less of then- volatile metals than do pumpable wastes,  and since solid-form wastes
        are not the majority of total wastes at WTI, this scenario is  not expected to create
        the potential for a significant  additional  source of volatile metals emissions.
                                                                    External Review Draft
Volume III                                V-ll                     Do Not Cite or Quote

-------
       Because the scrubber vessel is rubber-lined, failure of water cooling could cause the
       rubber liner and plastic bed media to melt and be destroyed by the hot combustion
       gases.  Accordingly, WTI has designed several levels of redundancy into the
       quench/scrubber water supply system to prevent such a scenario. Because of this
       redundancy, complete failure of water supply to the quench/scrubber is considered to
       be an infrequent event.

       In summary, because of the low expected frequency of such a scrubber water
       failure, because of the nature of the waste combustion,  and because of the relatively
       minor reduction in overall control efficiency expected during a scrubber water
       failure, this type of  event is unlikely to significantly affect overall facility emissions.

   •  Interruptions in Air Flow.  An ID fan operates hi the incineration train prior to
       entry of flue gases into the stack. Fan failure due to  catastrophic mechanical failure
       or due to power failure would be expected to result hi positive pressure hi the
       combustion chamber. Pumpable wastes would be cut off immediately and would  not
       create,  therefore, any further emissions.  If solid waste had been charged to the kiln
       shortly before such an event, partially burned organics  could be emitted through the
       kiln seals, since this would generally be the path of least resistance to the
       atmosphere once the kiln is under positive pressure.

       As mentioned earlier, power failure at the WTI facility does not necessarily result in
       complete loss of the ID  fan.  An  emergency 1,000 KW generator is present as
       backup in the event of a power failure, and this generator can keep the ID fan
       running, although at reduced speed, to keep the system under negative pressure.

       In spite of this back-up system, a fan-related release did occur hi December of 1993
       when a software logic error caused a complete fan shut-off after a quench water feed
       system failure.  Although the water supply was, hi fact, maintained to the quench
       unit via backup, the control  system shut off the ID fan. During that event, the
       operators reported that visible emissions (taken to be an indication of particle
       emissions and perhaps emissions  of unburned organic chemicals) were observed
       from the kiln seals for three to five minutes.

       On October 31, 1994, kiln gases were again accidentally released for up to ten
       minutes.  This event was also reportedly caused by a software logic error.  In this
       case, the control system incorrectly interpreted a blown fuse as an indication that the

                                                                   External Review Draft
Volume III                                V-12                      Do Not Cite or Quote

-------
         waste heat boiler had gone dry.  This tripped the ID fan motor circuit completely
         off, and the operators were not able to quickly override the automatic control system
         to get it running again.  In this instance, the kiln gases escaped by flowing
         backwards through the vapor recovery system piping, rupturing a section of flexible
         piping, and escaping through the rupture.

         In response to these events, WTI has taken measures, including computer logic
         changes and the addition of manual overrides, to eliminate situations such as these
         and others which might result in the ID fan completely tripping off.  Under these
         circumstances, releases from  interruptions  in ah* flow are anticipated to be infrequent
         events.

     •   Kiln Overpressure Events.  Events which cause the kiln to "overpressure," that is,
         to be held at higher than atmospheric pressure for more than a few seconds can
         result hi the release of partially burned waste material through the kiln seals.  Kiln
         overpressures may be caused  by chunks of solidified ash, or clinker, falling into the
         slag quench tank, which is located directly beneath the secondary combustion
         chamber. This causes a sudden release of steam to travel back into the secondary
         combustion chamber,  causing an overpressure event. An overpressure event of this
         nature is less likely to cause a release of waste constituents than an event associated
         with a feed or waste flow anomaly, since a steam-related overpressure is most likely
         to release steam to the atmosphere.  An example of a waste flow anomaly is the
         intermittent charging of containerized high BTU solid and liquid wastes in a batch
         mode that can cause temperature and pressure excursions.

         Kiln overpressure events trigger automatic  waste feed cutoffs (AWFCOs);
         consequently, as described earlier,  emissions associated with these events are not
         expected to occur over extended durations.  Nevertheless, due to  the frequent
         occurrence of kiln overpressures (WTI 1994), a detailed evaluation of these events
         has been conducted as described hi Appendix III-l.

         The analysis of AWFCOs conducted as part of this risk assessment, and described hi
         the Appendix of this report, does not identify reliable estimation techniques or
         empirical studies quantifying the nature or  magnitude of emissions from
         overpressure releases hi  the scientific literature.  However,  since these emissions
         occur from the kiln seals,  PCDD/PCDF are unlikely to be associated with these
         releases given current understanding of the conditions conducive to PCDD/PCDF

                                                                   External Review Draft
Volume  III                               V-13                     Do Not Cite or Quote

-------
        formation.  PCDD/PCDF formation is maximized in the post-combustion zone of
        the incinerator if the flue gases are allowed to reside at temperatures of 450° to
        750°F.

        In summary, emissions from AWFCOs associated with kiln overpressures are
        difficult to estimate, but are unlikely to contain significant concentrations of dioxins
        and furans.  Nevertheless, the inability to quantify emissions associated with
        frequent AWFCO events results in some uncertainty in the risk assessment.

C.  Uncertainties in Fugitive Emissions Characterization
    The major sources of uncertainty associated with estimation of fugitive emissions from
routine storage activities are summarized below:

    •   The amount of waste handled by the facility on an annual basis;

    •   Number and type (bulk vs.  containers) of truckloads of waste received by the
        facility;

    •   Because the pumpable waste stream for the waste profiles data base is used to
        calculate emissions, all uncertainties associated with the waste profile data base are
        inherent to  these calculations;

    •   The tank emissions calculation software is based on empirical equations using default
        values for parameters such as seasonal temperature, wind speed, and meteorological
        data.  Also, since the actual waste handling pattern between tanks may vary greatly
        depending on the volume of pumpable wastes received and the available storage
        tanks, it is  assumed that the entire volume of pumpable wastes passes through each
        type of tank once, rather than all tanks sequentially;.

    •   The relative percent of time the tank and container pumpout emissions are vented to
        the incinerator vs. the carbon adsorption bed could vary from the 53% to 47% ratio;

    •   Uncertainty relative to the wastewater tank includes the throughput of stonnwater,
        type and concentration of contaminants in the water (both are dependent on rainfall
        and spills or leakage of water), and uncertainties associated with the empirical
        equations used to calculate  the emissions;
                                                                   External Review Draft
Volume III                               V-14                     Do Not Cite or Quote

-------
     •   For container releases and truck washing, uncertainty is based on number of
         containers or vehicles processed, type of waste contained, and use of relatively non-
         specific, generic emission factors to calculate emissions; and

     •   For fugitive emissions, uncertainty is associated with the number of pumps, flanges,
         valves and seals that are in operation at any one time, and the use  of relatively non-
         specific emission factors to calculate emissions.

     It is believed that there is no general trend in these uncertainties which would result in
an over- or underestimation of fugitive emissions.
01-3999G:PCCOOB56.WS1
                                                                    External Review Draft
Volume III                                V-15                     Do Not Cite or Quote

-------
TABLE V-l
Observed Variation in the Control Efficiency of
Selected Metals During the May 1993 Trial Burn
Class of Material
Insoluble vapors
Soluble vapors
Fine particles
Coarse particles
Control Efficiency
(percent of weight)
6
99.68
99.977
99.997
Standard Deviation
4
0.2
0.006
0.002
Volume III
V-16
External Review Draft
Do Not Cite or Quote

-------

TABLE V-2
Possible Variation in Predicted Metals Emissions
Due to Uncertainty in Input Data



Metal
Aluminum
Barium
Copper
Nickel
Selenium
Silver
Thallium
Zinc
Predicted Emission Rate (g/s)
Based on
Most Probable
Data
2.4 x ID"4
1.5 x 104
9.4 x 10-5
5.0 x 10-6
4.7 x 1O4
1.5 x 10-5
3.4 x 10-5
1.2 x 1O4


Minimum
2.2 x Kf6
1.2 x 10-5
1.0 x 10-7
3.1 x lO'9
2.6 x 10-6
1.2 x 10*
2.6 x 10*
9.5 x 1CT6


Maximum
5.2 x lO'3
1.9 x ID'3
1.2 x 10-3
6.3 x 10-5
4.7 x lO'3
1.9x 10-4
4.7 x lO'3
1.7x ID'2
Volume III
V-17
External Review Draft
Do Not Cite or Quote

-------
                              VI.  REFERENCES
Agency for Toxic Substances and Disease Registry (ATSDR).  1993.  Toxicological Profile
    for Di(2-ethy1hexyl)phthalate.  TP-92/05. U.S. Department of Health and Human
    Services; Public Health Service. April.

A.T. Kearney,  Inc.  1993.  Test report for particle size distribution study conducted at Waste
    Technologies Industries, Inc. in East Liverpool, Ohio during March 15-17,  1993 trial
    burn.  A.T. Kearney, Inc., Chicago, Illinois.

Barton, R.G., W.D. Clark, and W.R. Seeker.  1990.  Fate of metals in waste combustion
    system.  Combustion Science and Technology, 74.

Bidleman, T.F.  1988. Atmospheric processes.  Environ. Sci. Technol. 22(4):361-367 and
    22(7):726-727.

ENSR  Consulting and Engineering (ENSR).  1993. Final trial burn report for the rotary kiln
    incinerator, Waste Technologies Industries, East Liverpool, Ohio.  Document number
    7136-001-800.  May.

ENSR  Consulting and Engineering (ENSR).  1994a.   Waste Technologies Industries,  East
    Liverpool, Ohio:  Final trial burn report for condition 2 - February 1994.  Document
    number 7136-007-400.  April.

ENSR  Consulting and Engineering (ENSR).  1994b.   Waste Technologies Industries,  E.
    Liverpool, Ohio:  August 1994 quarterly test emission results for PCDDs/PCDFs and
    paniculate matter. Document number 6933-660.  May.

ENSR  Consulting and Engineering (ENSR).  1994c.   Waste Technologies Industries,  E.
    Liverpool, Ohio:  August 1994 quarterly test emission results for PCDDs/PCDFs and
    paniculate matter. Document number 7115-660.  September.

ENSR  Consulting and Engineering (ENSR).  1994d.   Waste Technologies Industries,  E.
    Liverpool, Ohio:  3rd quarter 1994 sampling program for products of incomplete
    combustion (PICs).  Document number 7153-660. October.

ENSR  Consulting and Engineering (ENSR).  1995.  Waste Technologies Industries, E.
    Liverpool, Ohio:  4th quarter 1994 sampling program for products of incomplete
    combustion (PICs).  Document number 7289-860. January.
                                                              External Review Draft
Volume III                             VI-1                    Do Not Cite or Quote

-------
Entropy Environmentalists, Inc. (Entropy).  1994.  Volume 1: Stationary Source Sampling
    Report, Reference No. 12585,  WTI - Von Roll,  Inc. East Liverpool, Ohio; Emissions
    testing for: Lead, particulate, PCDD/PCDF; Incinerator stack.  February.

Friedlander, S.K.  1977.  Smoke, dust and haze. New York:  John Wiley and Sons.

Gelbard, F.   1980.  MAEROS.  Aerosol Science and Technology, 3,

Giam, C.S., H.S. Chan, and G.S. Neff.  1975.  Sensitive method for determination of
    phthalate ester plasticizers in open-ocean biota samples.  Anal. Chem. 47:2225-2228.

Hahn, J.L., H.P. VonDemfange, R.J. Jordan and J.A. Finney, Jr.  1985.  Air emissions tests
    of a Deutsche Babcock Anlagen dry scrubber system at the Munich North refuse-fired
    power plant.  Presented at the 78th Annual Meeting of the Air Pollution Control
    Associate, No. 85-76B.1, June.

Junge, C.E.   1977.  Basic considerations about trace constituents in the atmosphere as related
    to the fate of global pollutants. In Fate of Pollutants in the Air and Water Environments,
    Part I, ed.  LA. Suffet, 7-025.  New York:  Wiley and Sons.

Lemieux, Paul M., William P.  Linak, Joseph A. McSorley,  and Jost O. L. Wendt. 1992.
    Transient suppressions packaging for reduced emissions  from rotary kiln incinerators.
    Combust. Sci. and Tech. 85: 203-216.

Linak, W.P., and J.O.L. Wendt.  1993.  Toxic metal emissions  from incineration:
    mechanisms and control. Progress in Energy and Combustion Sciences, 19.

Mason,  G., T. Zacharewski,  M.A. Denomme, L. Safe, S. Safe.  1987.  Polybrominated
    dibenzo-p-dioxins and related compounds:  Quantitative  in vivo and in vitro structure-
    activity relationships. Toxicology 4: 245-255.

McNallan, M.J., G.J. Yurek, and J.F. Elliot.  1981.  The formation of inorganic particles
    by homogeneous nucleation hi gases produced by the combustion of coal.   Combustion
    and Flame, 42.

Muleski, G.E. and F.J. Pendleton, Midwest Research Institute; W.A. Rugenstein.  1986.
    Measurement of fugitive emissions in a coal-fired power plant. In Proceedings: Sixty
    Symposium on the Transfer and Utilization of Particulate Control Technology, Volume 3.
    The Detroit Edison Company. November.

Radian Corporation.  1992.   Medical Waste Incineration Emission Test Report.  Volume 1,
    pp.  2-24  - 2-25.

Thuren, A.  1986.  Determination of phthalates hi aquatic environments.  Bull. Environ.
    Contam.  Toxicol.  36:33-40.
                                                                External Review Draft
Volume III                              VI-2                     Do Not Cite or Quote

-------

 U.S. Environmental Protection Agency (U.S. EPA).  1983.  Hazardous waste management
     permit.  Waste Technologies Industries.  EPA Identification # OHD980613541.  U.S.
     Environmental Protection Agency, Region V.

 U.S. Environmental Protection Agency (U.S. EPA).  1988a. Measurements of particulates,
     metals, and organics at hazardous waste incinerators.  Office of Solid Waste. Draft
     Report.  November.

 U.S. Environmental Protection Agency (U.S. EPA).  19885. Methods for the Determination
     of Organic Compounds in Drinking Water; Determination of organic compounds in
     drinking water by liquid-solid extraction and capillary column gas chromatography/mass
     spectrometry-method 525.  EPA-600/4-88/039; U.S. EPA Environmental Monitoring
     Systems Laboratory; Cincinnati, OH.

 U.S. Environmental Protection Agency (U.S. EPA).  1989a. Guidance on metals and
     hydrogen chloride controls for hazardous waste incineration.  Volume IV of the
     hazardous waste incinerator guidance series. Draft. Office of Solid Waste, Waste
     Treatment Branch, Washington, D.C.

 U.S. Environmental Protection Agency (U.S. EPA).  1989b. Background information
     document for the development of regulations for PIC emissions from hazardous waste
     incinerators.  Draft Final Report.  Office of Solid Waste. BBSP-S0002.  Washington,
     D.C.

 U.S. Environmental Protection Agency (U.S. EPA).  1990a. Operations and research at the
     U.S. EPA incineration research facility.  Annual report for FY89B. Risk Reduction
     Engineering Laboratory, Office of Research and Development, Cincinnati, Ohio.
     EPA/600/9-90/012.

 U.S. Environmental Protection Agency (U.S. EPA).  1990b. Standards for owners and
     operators of hazardous wastes incinerators and burning of hazardous wastes in boilers
     and industrial furnaces.  Fed.  Reg. 55:17862-17921.

 U.S. Environmental Protection Agency (U.S. EPA).  1991a.  Burning of hazardous waste in
     boilers and industrial furnaces: final rule.  Fed. Reg. 56:7134-7240.

 U.S. Environmental Protection Agency (U.S. EPA).  1992a.  Compilation of air pollutant
     emissions factors,  Volume I: Stationary point and area sources (and supplements).
     Research Triangle Park.  AP-42.

 U.S. Environmental Protection Agency (U.S. EPA).  1992b.  Supplemental Guidance to
     RAGS;  Calculating the Concentration Term. May.
                                                                External Review Draft
Volume III                              VI-3                    Do Not Cite or Quote

-------
U.S. Environmental Protection Agency (U.S. EPA).  1993a.  WTIphase E risk assessment
    project plan, EPA ID number OHD98061354L Region V, Chicago, Illinois.  EPA
    Contract No. 68-W9-0040, Work Assignment No. R05-06-15.  November.

U.S. Environmental Protection Agency (U.S. EPA).  1993b.  EPA draft strategy for
    combustion of hazardous waste hi incinerators and boilers. Environment Reporter -
    (May 21):   157.

U.S. Environmental Protection Agency (U.S. EPA).  1993c.  Report on the technical
    workshop on WTI incinerator risk issues. Risk Assessment Forum, Washington, D.C.
    EPA/640/R-94/001.  December.

U.S. Environmental Protection Agency (U.S. EPA).  1994a.  Implementation guidance for
    conducting indirect exposure analysis at RCRA combustion units. Memorandum from M.
    Shapiro, Director, Office of Solid Waste.  EPA/530/R-94/021.  Revised April 22.

U.S. Environmental Protection Agency (U.S. EPA).  1994b.  WTI Risk Assessment: Stack
    Organic Compound Emissions.  Memorandum from Gary Victorine.  December 20.

Waste Technologies Industries (WTI).  1982. Application to the  United States Environmental
    Protection Agency.  November.

Waste Technologies Industries (WTI).  1993. Letter from HJ. Dugan, Environmental
    Manager to D.  Schregardus, Ohio EPA, containing preliminary testing results and
    certification. August 30.

Waste Technologies Industries (WTI).  1994. Report of AWFCO incidents to Ohio EPA,
    1994.

Waste Technologies Industries (WTI).  1995. Letter from D.E. Apple, Environmental
    Specialist to G. Victorine, U.S. EPA Region V, containing incinerator ash analyses.
    January 23.

Victorine,  G.  1995.  Personal communications.
                                                               External Review Draft
Volume III                             VI-4                    Do Not Cite or Quote

-------
          APPENDIX IH-1

EMISSIONS ESTIMATION METHODOLOGY
        AND BACKGROUND

-------
                               APPENDIX ffl-1

        EMISSIONS ESTIMATION METHODOLOGY AND BACKGROUND

                                CONTENTS
CHAPTER I. INTRODUCTION	1-1

      A.  Scope  	1-1
      B.  Concerns to be Addressed  	1-1
      C.  Approach	1-2

             1.    Waste Profile Data base	1-2
             2.    Estimation of Metals Emissions	 1-2
             3.    Estimation of Organic Emissions	,  . 1-2
             4.    Estimation of Emissions from Other Activities	1-2

      D.  Overview of Limitations	1-3
      E.  Organization  	1-3
CHAPTER H.  WASTE PROFILE DATA BASE	H-l

      A.    Introduction	II-l
      B.    Data Base Development	n-1

            1.     Summary Information	n-2
            2.     Anticipated Constituent Composition	n-2
            3.     Analysis Results	n-2

      C.    General Data Base Refinement	n-3

            1.     Estimation of waste composition in ranges	II-3
            2.     Nomenclature of identical compounds and the
                   presence of isomers	n-4
            3.     Incomplete analysis results	n-5

      D.    Potential Sources of Error  	H-9
      E.    Limitations   	n-10
      F.    Uncertainty  	EHO
Volume HI                                                       External Review Draft
Appendix ni-1                            i                       Do Not Cite or Quote

-------
                              CONTENTS
                                  (Cont'd)
CHAPTER ffl. ESTIMATION OF UNMEASURED METALS EMISSIONS AND
             EVALUATION OF METALS BEHAVIOR AT THE WTI
             INCINERATOR  	ffl-l

      A.    Concerns to be Addressed	 ni-1
      B.    Approach  	ni-1
      C.    Estimation Technique Selection	m-1

            1.    Review of Raw Analytical Data From Trial Burn	ni-2
            2.    Review of Historical Data	DI-3
            3.    Modeling of Metals Emissions	IQ-4
            4.    Conclusion	m-4

      D.    Emissions Estimation	ffl-4

            1.    Introduction  	m-4
            2.    Model Development  	ffl-5
            3.    Model Application  	  ffl-15

      E.    Uncertainty Analysis	 ffl-19

            1.    Modeling Assumptions	  ni-19
            2.    Data	  ffl-20
            3.    Removal Efficiencies	  ffl-21
            4.    Uncertainty Estimates	  ffl-22

      F.    Speciation	  ffl-23

      G.    Other Topics  	  DI-23

            1.    Aluminum Toxicity	  ffl-23
            2.    Chromium Valence State	  ffl-24
            3.    Emissions from Scrubber Water 	  ffl-25

      H.    Conclusions	  ffl-25

CHAPTER IV. ESTIMATION OF ORGANIC EMISSIONS FROM
             THE WTI INCINERATOR	IV-1

      A.    Introduction	IV-1
      B.    Estimation of Organic Emissions 	IV-2
Volume HI                                                     External Review Draft
Appendix III-l                           ii                      Do Not Cite or Quote

-------
                                CONTENTS
                                     (Cont'd)
             1.     Estimated PIC Emissions	IV-4
             2.     Estimation Procedure for the Uncharacterized
                   Fraction	IV-6

      C.     Effect of Control Device	....:.,	IV-8
      D.     Emissions of PCDD/PCDF .	IV-9
      E.     Uncertainties   	  IV-10

CHAPTER V. ESTIMATION OF EMISSIONS FROM OTHER SOURCES	V-l

      A.     Evaluation of Emissions from Automatic Waste
             Feed Cutoffs   .	V-l

             1.     Concerns to be Addressed	V-l
             2.     Approach	V-l
             3.     Frequency/Emissions Estimate for AWFCOs	V-3
             4.     Uncertainty	V-13
             5.     Conclusions	V-14

      B.     Estimation of Fugitive Emissions from
             Routine Operations	V-15

             1.     Introduction  	V-15
             2.     Site Description	V-15
             3.     Review of Releases at TSDFs	V-16
             4.     Identification of WTI Sources of Releases  	V-18
             5.     Emission Estimation Calculations	V-19
             6.     Compound  Specific Emissions  	V-25
             7.     Uncertainty	V-27

      C.     Emissions from Ash Handling	V-28

             1.     Concerns to be Addressed	V-28
             2.     Approach	V-28
             3.     Estimation of Emissions	V-29
             4.     Uncertainty   	V-32

CHAPTER VI.  REFERENCES  	VI-1
Volume III                                                        External Review Draft
Appendix III-l                            Hi                       Do Not Cite or Quote

-------
                                 CONTENTS
                                     (Cont'd)
                                     TABLES
                                                                             PAGE
Table II-l    Assumptions Made During Data Base Refinement	11-12
Table II-2    Analytical Values for Total Waste Streams
             (Pumpable & Non-Pumpable)   	:	n-13
Table n-3    Analytical Correction Factors   	n-14
Table n-4    Key Assumptions for Chapter H	H-15
Table ni-1    Metals System Removal Efficiencies Measured
             During the WTI Trial Burn	  IH-27
Table ffl-2    Waste Feed Compositions Used in the Modeling	  HI-28
Table IH-3    Comparison of Predicted Feed Rates with the Permit Limits	  ffl-29
Table ffl-4    Classification of Metal Vapors Used to Determine Control Efficiency  .  IH-30
Table ffl-5    Control Efficiencies Used in Model	  IH-31
Table IH-6    Predicted Metals Emission Rates  	  m-32
Table ffl-7    The Observed Variation in the Control Efficiency of
             Selected Metals During the March 1993 Trial Burn	  m-33
Table ffl-8    Possible Variation in Predicted Metals Emissions Due to
             Uncertainty in Input Data	  ffl-34
Table m-9    Key Assumptions for Chapter HI  	  m-35
Table IV-1    POHCs, PICs and THC Measured During the WTI Trial Burn	  IV-11
Table IV-2    Source of Emission Rates	  IV-12
Table IV-3    Key Assumptions for Chapter IV	  IV-13
Table V-l    Positive  Pressure and Total Automated Waste Feed
             Cut-Offs (AWFCOs) Reported by Month at WTI   	V-33
Table V-2    DRE Results from the 1994 Trial Burn Condition No. 2	V-34
Table V-3    Summary Provided by Permittee of Equipment/Procedural Failure
             (Unrelated to Incinerator) Which Resulted in Spills or Releases	V-35
Table V-4    Summary of Releases at TSDFs	V-36
Table V-5    Emissions from Organic Waste Storage Tanks (Human Health -
             Constituents  of Concern)	V-37
Table V-6    Emissions from Organic Waste Storage (Ecological)	V-38
Table V-7    Summary of Estimated  Emissions from Routine Operations  	V-39
Table V-8    Key Assumptions for Chapter V	V-40

                                     FIGURES


Figure ffl-1   System penetration observed during the WTI Trial burn
             for As, Be, Cd, Pb and Sb. The metals are listed in
             order of increasing volatility	  ffl-37
Volume m
Appendix III-l
IV
External Review Draft
Do Not Cite or Quote

-------
                                  CONTENTS
                                       (Cont'd)
                                     FIGURES
Figure ffl-2  Pathways available for toxic metals hi the WTI incinerator	  ffl-38
Figure ffl-3  Schematic diagram of modeling approach used	  ffl-39
Figure ffl-4  The predicted evolution of the particle size distribution
             hi 1400 K gases   	  ffl-40
Figure ffl-5  The predicted evolution of the particle size distribution
             in 400K gases	  ffl-41
Figure ni-6  Relative rates for homogenous condensation and heterogenous
             condensation onto 0.1 fan particles and 10 /mi particles  	  111-42
Figure ffl-7  Comparison of the model's predictions and the March trial burn
             results from the WTI incinerator  	  ffl-43
Figure ffl-8  The impact of temperature on the predicted metals emissions
             and the observed SREs  	  ffl-44
Figure EI-9  The impact of the quantity of ah- used on the emissions rates
             and the observed SREs of the metals of interest   	  ffl-45
Figure ffl-10 The impact of chlorine on the predicted metals emissions rates and SREs ffl-46
Figure HI-11 The impact of quench temperature on the predicted metals emissions
             rates and SREs	  HI-47
Figure ni-12 The impact of entrainment rates on the predicted metals emissions
             rates and SREs	  ffl-48
Figure IV-1  Components of organic compound emissions estimate	  IV-18
Figure IV-2  Procedures for estimating emissions of organic compounds  	  IV-19
Figure IV-3  Procedures for estimating emissions of organic compounds
             (estimation steps with uncertainties highlighted)   	  IV-20
Figure V-l    Slag and ash handling diagram - WTI facility	V-40
Volume HI                                                            External Review Draft
Appendix III-l                               V                         Do Not Cite or Quote

-------
                        CHAPTER I.  INTRODUCTION
A. Scope
       This appendix presents estimations of emissions from various activities at the WTI
facility in East Liverpool.  These estimations are performed in support of the risk assessment
at the WTI facility.  Emission estimates are performed for metals and organics emissions
from the WTI incinerator, and for fugitive emissions from other activities at the  site.
Subsequent to development of emissions estimates, additional stack testing of PICs were
performed.  Data from these tests are used in the risk assessment (see Volume HI); however,
the discussion on emissions estimates is presented to further verify the test data.

B.  Concerns to be Addressed
       The 1993 Peer Review Panel raised a series of comments regarding the emission
characterization  section of the original 1993 project plan for the WTI risk assessment (U.S.
EPA 1993b).  These comments were directed primarily toward characterizing certain types of
emissions from the WTI incinerator for which test data did not exist.  Additional concerns
directed toward organic emissions noted that a potentially  large fraction of organic emissions
from the incinerator were uncharacterized, and that the waste feed used hi the WTI trial burn
was not representative of the type of waste normally burned in the incinerator. These
concerns called for the development of a chemical composition profile representative of
wastes routinely burned at the facility, and the estimation  of compound-specific emissions to
represent the uncharacterized fraction of organic emissions from the incinerator.
       Concerns were also directed towards characterization of metals emissions  including
defining the emission rates of metals for which there are no test data, identifying the
chemical form (speciation) of the metals, and predicting the impact of changes hi parameters
such as kiln temperature on metals emissions.
       Furthermore, concerns were raised about transients due to non-steady state operation,
system upsets that could cause a waste feed cutoff, fugitive emissions due to leaks and spills,
and catastrophic events such as fires and natural disasters.
Volume m                                                             External Review Draft
Appendix III-l                               1-1                         Do Not Cite Or Quote

-------
C.  Approach
      The comments summarized above led to the development of an assignment to estimate
emissions.  The assignment is broken down into four major areas, each described below.
       1.     Waste Profile Data Base
             Development of a waste profile data base to represent wastes received or
       projected to be received by the facility over the course of the first year of operations.
       The data base, and ultimately two surrogate waste streams representing pumpable and
       non-pumpable wastes, is developed based on projected waste volumes from individual
       waste sources.  The surrogate waste streams are then used in the estimation of
       emissions from the other tasks (primarily metals).

       2.     Estimation of Metals Emissions
             Estimation of metals emissions that were not measured during tests at the
       facility, and evaluation of metals control efficiency, particle size distribution, and
       reentrainment from scrubber water.  Conservative estimates of system metals removal
       efficiencies and metals speciation are developed as part of this task.

       3.     Estimation of Organic Emissions
             Estimation of organic emissions representing the unmeasured fraction of
       emissions from the incinerator.  A target list of potential compounds is identified, and
       a method developed to characterize the unmeasured fraction based on organic
       analytical data from tests performed at the facility.  Subsequent to this exercise,
       additional stack testing was performed providing the bulk of the data used in the risk
       assessment.

       4.     Estimation of Emissions from Other Activities
             Estimation of frequency of occurrence and emissions from automatic waste
       feed cutoffs, estimates of fugitive emissions from leaks, spills, and routine waste
       handling activities at the facility, and estimation of fugitive particulate emissions from
       ash handling activities at the site are presented in this task.
             The assignment is presented  in two phases, particularly since the Peer Review
       Panel had suggested a number of techniques for estimating emissions, and other
       techniques for estimation were also  available.  Therefore, the first phase is to evaluate
Volume III                                                             External Review Draft
Appendix IH-1                               1-2                         Do Not Cite Or Quote

-------
       alternative approaches to complete the assignments and the second phase is to
       implement the favored approach.

D.  Overview of Limitations
       As in any estimation technique, there are uncertainties involved with the emissions
estimated under this assignment. These uncertainties and key assumptions made in the
process will be discussed in greater detail in each of the following sections of this appendix.
Many of the limitations encountered are similar, in that they stem from not having complete
knowledge or information about wastes fed to the kiln, behavior of the wastes or combustion
process within the kiln, or behavior of the combustion gases hi the air pollution control
system.  Still other uncertainties stem from the assumptions made within the estimation
techniques, or the use of one technique rather than another.  However, the majority of
uncertainties should either have minimal impact on the overall calculated risk, or should err
on the side of overestimating the risk.

£.  Organization
       This appendix is organized as follows.  Chapter n describes activities relating to the
waste profile data base. Chapter HI covers estimation of metals.  Chapter IV deals with
organic estimates, and Chapter V includes activities relating to automatic waste feed cutoffs,
waste handling  activities, and ash handling.  References are included as Chapter VI. Each
section contains a description of the peer review comments, approach, emissions estimations,
and a discussion of uncertainties.  The Attachments to all sections are included following
Chapter VI.
Volume III                                                              External Review Draft
Appendix ffl-1                               1-3                          Do Not Cite Or Quote

-------
               CHAPTER H.  WASTE PROFILE DATA BASE

 A.    Introduction
       As an initial step hi the emissions analysis, an evaluation is performed of waste feed
 material.  Because materials fed to the incinerator during the initial trial burn may not be
 representative of a typical waste feed during facility operations, it was suggested that waste
 manifests from the facility's first year of operation be used to develop a waste feed chemical
 composition profile.  This chapter describes the approach taken to develop the waste profile
 data base.

 B.     Data Base Development
       To develop the waste feed profile, it is necessary to evaluate the waste composition of
 individual incoming waste streams to the facility.  Waste manifests, which were a source of
 data suggested by the Peer Review Panel, list incoming waste by  U.S. EPA hazardous waste
 codes (as  defined in 40 CFR Part 261) and  were determined to lack appropriate level of
 detail. Similarly, "fingerprint" analyses performed by WTI at the time of waste receipt
 provide data important to waste acceptance  and handling procedures, but do not provide
 sufficiently detailed analysis of the chemical composition of the incoming waste stream.
       A third alternative source was selected to develop the waste profile data base.  The
 U.S. EPA obtained waste profile sheets for the first nine months of operation at WTI, and
 waste profile receipt data for the first year of operation.  A sample waste  profile sheet is
 included as Attachment 1. Seventy-eight waste profiles are evaluated and entered hi the data
base.  The profiles are completed by the generator prior to any wastes being received by
WTI to document the expected composition of waste streams and  to receive acceptance
approval from WTI.  The profiles identify waste composition within expected ranges,
physical state of wastes; specific handling instructions and hazardous waste codes.  Prior to
approving a waste for acceptance, WTI reviews  these data to verify the anticipated wastes are
within their operating constraints (e.g., permitted conditions).   Once a waste is approved by
WTI, the profiles are submitted to the Ohio EPA for approval and are subsequently
maintained on file to verify incoming wastes.  As wastes are shipped to WTI, the manifests
are checked against the profiles to determine acceptability of wastes.  A quick analysis, often
referred to as a fingerprint analysis, is normally  performed to verify that the waste is
consistent with the waste profile.
       To formulate the data base, the waste profiles were first reviewed for content to
determine which data were needed.  Data selected include:  waste profile number (which

Volume HI                                                              External Review Draft
Appendix III-l                              rj-1                         Do Not Cite or Quote

-------
identifies the generator), sampling date, annual estimated volume for the waste stream,
physical state (i.e., solid, liquid) special handling requirements, constituent content in the
form of ranges, and sample analysis results for selected constituents, primarily metals.
Paradox™ is used as the spreadsheet software since large amounts of data need to be sorted
quickly and because of the  ease with which data can be converted to Lotus and Dbase files,
should the data need to be manipulated using other software.
       Because the waste profiles contain proprietary information regarding the generator and
the generator's processes, the waste profile sheets are claimed by WTI as Confidential
Business Information (CBI). A unique surrogate waste ID code is used for each waste
stream in place of the original WTI source codes, regardless of the number of streams from
the same source.  The surrogate codes are assigned randomly and are not related hi any way
to the original WTI source  codes. In this  way, a detailed nonconfidential master list of the
waste streams is created for use in the risk assessment,  while the required confidentiality of
the generator information is still maintained.
       From the waste profiles, three datasets are created as described below:
       1.     Summary Information
             Data from pages 2 & 3 of the waste profile include:
             •      Surrogate waste code;
             •      Anticipated annual volume;
             •      Physical  state; and
             •      Special handling requirements.

       2.     Anticipated  Constituent Composition
             Data from pages 4 & 5 of the waste profile include:
             •      Surrogate waste code;
             •      Chemical constituents; and
             •      Range of composition for each constituent (by percent).

       3.     Analysis Results
             Data taken from actual analytical results  (page 10 of waste profile) include:
             •      Surrogate waste code;
             •      Anions (%s, measured values);
             •      Metals (measured values);  and
             •      Physical  Properties.

       The three data sets are  cross-linked through the  common surrogate code for
appropriate linkages as described below.

Volume HI                                                              External Review Draft
Appendix HI-1                               H-2                        Do Not Cite or Quote

-------
       The anticipated volume, reported in pounds per year are totaled for each source, and
the total over a one year period is assumed to be representative.  This means that the total
wastes for a given source are taken to represent an input rate to the incinerator,  from that
source, in pounds per year (Ibs/yr).  The rank of each waste stream is computed based on
anticipated volume.  The wastes are then sorted by the anticipated volume of each waste
stream.
       The input, in Ibs/yr, from each source, is then divided by the total input volume in
Ibs/yr to give a percentage of the total input.
       After sorting the data base by volume, those sources of wastes which contribute less
than 0.5% of the total feed are deleted to streamline the data.   This results in a reduction in
number of input streams from 78 to 42 (41 %) but reduces the total amount of waste volume
carried throughout the remainder of the analysis by only 4.45%.  The resulting uncertainty
due to this reduction is deemed insignificant, and results in a more manageable data base1.
       The second and third WTI data sets,  including analytical values, are then appended to
the first by the common surrogate codes.  This results hi the "parent file" for subsequent
calculations  of constituent feeds and elemental breakdown, as well as a list of waste streams
containing multiple packing types and physical states  of each stream.

C.     General Data Base Refinement
       In an effort to evaluate the data base, the various streams  hi the parent file are
analyzed by sorting by volume, percent volume of total, and surrogate codes.  The files are
then sorted by constituents of each stream.  In this analysis, three factors impact the overall
usability of the data:  1) the estimation of waste composition by ranges; 2) the nomenclature
of identical compounds and the presence of isomers and 3) some  analysis results listed are
incomplete.  Each of these factors are discussed below and steps taken to reduce their impact
are described.

       1.      Estimation of waste composition in ranges.
              In the waste profiles, the concentration of the individual components (i.e.,
       constituents) of the waste stream are  often estimated as a range.  For example, the
    1   It should be recognized that metal emission rates ultimately used in the risk
       assessment were prorated up to account for the maximum heat input to the
       incinerator.  Similarly, the annualized volume of each organic compound, whose
       emissions were estimated using the waste profile data base, was adjusted upwards by
       a factor of 4.1 to reflect maximum throughput of waste through the incinerator.
       Thus, the exclusion of certain wastes as discussed above did not significantly affect
       the emission rates used in the risk assessment.
Volume III                                                             External Review Draft
Appendix ffl-1                               H-3                        Do Not Cite or Quote

-------
       amount of a constituent in a waste stream might be reported as "1,2-Dichlorobenzene
       0-30%".  To ensure conservative bias, the upper bound of the percentage range for
       each constituent is selected as the constituent's concentration.  However, this
       conservative assumption introduced an uncertainty in that waste constituent volumes
       now totals greater than 100%  of the total waste stream volume.
              To illustrate this factor, suppose one waste stream contains four constituents.
       Their estimated concentrations for each constituent is reported as a range of 0-30% of
       the total waste stream volume. When selecting the upper bound for each constituent,
       (in this illustration 30% each), the waste stream volume now totals 120% of the
       original volume.
              To account for this factor a prorating method is applied when the total of the
       upper bounds of estimated ranges is greater than 100%. The upper range value for
       each constituent is multiplied by the reciprocal of the total.  In the example described
       above, each constituent's upper range value (30%) is multiplied by 100/120 resulting
       hi 25% as the new value.  The total for the example waste stream would then be
       100% of its input in pounds/year, and each constituent is assumed to be input at 25%
       of the total input rate for that  stream in pounds per year.
             In this scheme, each estimate is forced to the highest value possible, while
       maintaining the ratio of the various constituents of each stream relative to each other.
       After the prorating is complete, the percentages are totalled and indicated a 1.99%
       reduction in the total waste volume (possibly due to rounding errors).
             The selected prorating  method keeps the relative amounts of each constituent
       constant with each other; however, it introduces a  source of uncertainty within the
       data base.   This method ascribes a "point-value" to each constituent to identify a
       percentage of the waste stream which is then multiplied by the total anticipated annual
       volume of waste input. The result provides a  point value for the volume of each
       constituent.  However, there is no way to determine the actual value of the
       percentage.  Since the waste data represent anticipated annual amounts and thus an
       average of the total waste feed stream, and the conservative maximum range is
       applied, the uncertainty introduced by this method is deemed acceptable.
       2.     Nomenclature of identical compounds and the presence of isomers.
             The next level of review identifies several cases of synonyms and isomeric
       compounds. Since the data originates from multiple sources, different constituent
       names are utilized to represent the same compound. In the cases of synonyms, the
       conventional name or IUPAC accepted name is selected.  All identical constituents

Volume ffl                                                             External Review Draft
Appendix III-l                              H-4                         Do Not Cite or Quote

-------
   N               reported under other names are renamed to the conventional name and summed.  The
V  /               resulting renaming and grouping of isomers under the most commonly used name or
                   IUPAC designation, (followed by an asterisk if renamed) reduces the number of
                   compounds from over 500 to about 320.
                         An example is one waste source listed methyl benzene as a constituent, and
                   another listed the identical constituent as toluene. Therefore, the methyl benzene is
                   renamed as toluene for consistency.  Once this review is performed for all
                   constituents in the parent file, it is re-sorted by name.
                         Similarly, in the case of isomeric compounds, only the root name is used. An
                   example is the ortho,  meta-, and para- isomers of xylene, which are all renamed
                   "Xylene".  The data base  is then again sorted alphabetically by name.

                   3.     Incomplete analysis results.
                         Thirteen analysis results listings are incomplete and contain  "not analyzed"
                   entries for some analytes or "Not Detected" entries.  These entries  do not include
                   detection limits or elaborate on testing methods and are entered as values of zero in
                   subsequent calculations.
                         Upon application of the waste stream data base to the metals analysis  (see
("~^               Chapter HI),  a further refinement of the data base is necessary.   To calculate the
 '                  molar ratios of elements fed to the incinerator the modeling software requires the
                   waste streams be presented hi two discrete waste streams:  pumpable and non-
                   pumpable feeds.  To create these two waste streams, the wastes are combined by thek
                   physical  state assigned by the generator.  The physical states are reported as  one  of
                   five types on the waste profile.  Wastes reported as  Liquid; Liquid, Solid/Liquid Mix
                   and Solid/Liquid Mix are  aggregated to create the pumpable waste  stream. Those
                   wastes reported as Solid and Solid,  Solid/Liquid Mix are aggregated to represent the
                   non-pumpable waste stream.  Several inconsistencies are noted at this stage of the
                   data analysis  as described  below.

                         •     Lithium Batteries—listed in the packaging field as  "drummed
                               solids"—are sometimes designated as  liquids in physical state field of
                               the original WTI data base.  Similarly, several wastes identified as
                               "solid" in the field labeled physical state are identified as bulk liquids
                               in the packaging field.  Often, the packaging does not seem to match
                               the physical state. Materials listed as Liquids in the  physical state
                               column are also called out as drummed solids, bulk solids, and "other".
\  /                           Materials listed as solids hi the physical state field are sometimes listed

            Volume III                                                             External Review Draft
            Appendix III-l                              JQ-5                        Do Not Cite or Quote

-------
                    as bulk liquids or drummed liquids in the packaging field.  For
                    consistency, the simplest field (physical state) is selected when
                    determining the ultimate category (pumpable or non-pumpable) for the
                    waste stream.  Since only one physical state is listed for each stream,
                    there could not be any contradiction.  Further justification for this
                    approach is the apparent consistency of the reported physical state with
                    the composition of each waste stream.  In other words, on a case-by-
                    case examination of the waste streams, the physical state column
                    usually seems reasonable for the type of waste specified and is selected
                    as the primary criterion for pumpability or non-pumpability.

                    The molecular formula of one of the listed constituents of one of the
                    streams could not be determined. This compound is listed as "2,3-
                    dibromo-phosphate" and amounts to 462 Ibs/yr as estimated before
                    correction for analyte contribution.  It is left out of the mass balance.
                    Another compound, listed by confidential trade name, can not be
                    identified in the literature, amounting to 4,604 Ib/yr (uncorrected
                    values).

                    The data base contains many  instances of inputs which can not be
                    accurately speciated; this category is called Miscellaneous ("Misc.").
                    It contains mainly absorbent,  dirt, rust, grit, trash, tyvek, debris, sand
                    filter media, personal protective equipment and wood  chips.  The
                    consequence is an incomplete mass/energy balance. This is a more
                    significant portion of the waste; uncorrected weight hi the pumpable
                    stream is nearly one million Ib/yr.

                    "Ash" percentage is one of the results of analysis of the wastes since
                    ash is often a listed constituent.  These values are left out of mass
                    balance calculations, since there is no way to determine the composition
                    of the  ash from the data available. No precise number can be attached
                    to the percentage of ash, because it is not measured for all waste
                    streams.  Many waste streams listed "NA" for the percentage ash. It is
                    also notable that ash is assumed to be the predominant product of
                    combustion of the Miscellaneous ("Misc.") category of waste.
Volume III                                                               External Review Draft
Appendix ni-1                               II-6                         Do Not Cite or Quote

-------
              •      Misspellings in constituent names are corrected to the best of our
                     ability.  This may lead to misidentification in some isolated cases where
                     different compounds differ only slightly in spelling and no correct
                     choice is evident for a misspelled name.

              At this point, each file is sorted by constituent names, and like constituents
       summed to determine the total amounts of each constituent in each file.  The resulting
       files provide a preliminary estimate for the amounts of individual constituents in each
       of the pumpable and non-pumpable streams.  The data base is further refined by
       calculating values for the input  rates of the individual constituents which both conform
       to the anticipated chemical composition percentages, and yield amounts of each
       constituent predicted by the analysis results.  The analysis results do not always agree
       with the amounts of compounds projected in the anticipated chemical composition
       shown in the waste profile.  In  order to revise the data to conform to analytical
       results, the following operations are performed on the data bases:

              •      Ash (5% of total volume) is deleted from calculations, because it is
                     assumed to be removed by ash handling systems, and since it is
                     primarily a physical parameter rather than a waste constituent;

              •      The miscellaneous category (3 %)  is deleted from calculations, since it
                     contains mainly absorbent, dirt, rust,  grit, trash, tyvek, debris, sand
                     filter media, personal protective equipment (PPE), and wood chips, for
                     which it is assumed that ash is the primary combustion product;

              •      The lithium batteries (1 %) are deleted from subsequent calculations,
                     since insufficient information existed to properly speciate them; and

              •      The confidential tradename for which no data are available and "2,3-
                     dibromophosphate" (<1%) are deleted.

              These assumptions delete an additional 10% of the volume of waste fed to the
       incinerators, leaving a total of over 83% of the total volume. The volume  (Ibs/yr)
       input columns  in each file are totaled excluding water, giving 16,270,747 Ibs/yr for
       the pumpable stream and 5,068,920 Ibs/yr for the non-pumpable stream.  All
       subsequent calculations are performed without changing those values, ensuring a
       "mass-balance". While the exact mix of individual compounds (i.e., relative % of

Volume III                                                             External Review Draft
Appendix HI-1                              n~7                         Do Not Cite or Quote

-------
      total volume) fed to the kiln did change, neither the specific compounds nor the total
      volume in pounds per year change.
             The two large data bases (representing the pumpable and non-pumpable waste
      feed streams) resulting from this approach are treated identically to each other hi
      order to determine their anion and metal contributions.  A mass-balance approach is
      used to assure reasonable agreement between the total mass of wastes known to be fed
      and the sum of the weights of the elements as determined by spreadsheet analysis (see
      discussion below). All calculations performed change only the ratios of species to
      each other, not the overall mass input to the kilns. Over 83% of the total volume in
      Ibs/yr originally  calculated is accounted for hi these two data bases. The final  results
      of these two data bases are presented hi Attachment 2.
             To perform the mass balance approach, the next steps are to 1) determine the
      molecular formulas of the species input to the kiln, and 2) use those formulas hi
      conjunction with the mass balance approach to determine better estimates of the
      amounts of each species input to the kirn.
             First, molecular formulas are entered for each of the materials listed hi  the
      data base.  The molecular formulas are obtained through CD-ROM data bases; the
      Registry of Toxic Effects of Chemical Substances Data Base from the National
      Institute for Occupational Safety and Health (RTECS  from NIOSH, updated 1994),
      the Hazardous Substances Data Bank from the National Library of Medicine (HSDB
      from NLM, updated 1994), and the Oil and Hazardous Materials/Technical Assistance
      Data System from the Environmental Protection Agency (OHM/TADS  from EPA,
      1985).  Other sources of molecular formulas are the Alfa Catalog of Research
      Chemicals (1993-1994 edition), the Aldrich Catalog Handbook of Fine  Chemicals
      (1994-1995 edition), and the Handbook of Environmental Data and Organic
      Chemicals 2nd ed. (Versheueren, 1983).  Assumptions made during this portion of
      the data analysis are listed hi Table II-1.
             At this tune,  analytical values presented hi the third WTI dataset (the analysis
      results) are used to calculate the estimated volume in  pounds per year of each of the
      metals and anions. The analytes and then: calculated values are presented hi Table n-
      2.
             These values are then used to calculate analytical correction factors for  each
      constituent of the waste feeds containing one or more of the analytes.  The correction
      factors are used to adjust the waste profile to address discrepancies between the
      analytical data presented on metals and anions and the estimated percentages for
      specific constituents presented in the waste profiles.   These factors are  used to
      recalculate the total Ibs/yr of each constituent in order to force the resulting revised

Volume III                                                             External Review Draft
Appendix ffl-1                              II-8                        Do Not Cite or Quote

-------
        Ibs/yr value to yield the analytically determined amount in moles of the analytes. This
        is necessary in order to reconcile the difference between the amount of each chemical
        species calculated to be input to the kilns using the prorated percent values and the
        amounts of chemical species calculated to be input to the kiln by applying the
        analytical results to the volume of total materials fed to the kiln. This is a significant
        source of uncertainty.  The consequences of not adjusting the input quantities to
        reflect the analytical results include the overestimation of elemental inputs for many
        analytes, and would therefore skew the modeling based on those calculations.

        Correction factor = (moles/yr analytical value) / (moles/yr determined from %  total constituent)

              Once the amounts of compounds containing analytes are revised, their mass is
        totaled and subtracted from the total mass of then: respective  stream, pumpable  or
        non-pumpable.  This resulting mass is  then used to create  a different correction factor
        for the remaining constituents, bringing the  total mass back to its original total (the
        mass balance approach).
              Certain compounds  contain more than one of the analytes for which
        subtraction correction factors are developed. For example, some compounds contain
        both chlorine and bromine.  In such cases, only one correction factor  is applied. For
        compounds containing both chlorine and another analyte, the  factor for chlorine is
        always used to adjust the values.  Errors due to constituents containing multiple
        analytes are < 2%  in the non-pumpable stream and < 1% in the pumpable stream.
        Table n-3 lists the correction factors for the analytes.

D.      Potential Sources of Error
        These revisions (recalculation of input  rates  to reflect the analytical results, the
dropping out of the data base of ash, miscellaneous, trade name compounds,  and 2,3-
dibromophosphate) are mainly intended to reduce the miscalculation  potentially introduced
early in data analysis  due to the assignment of a point-value to constituents which are
identified originally by a range percent, and to reconcile the profile ranges with analysis
results.  For example, if an individual  shipment of  waste is labeled "1,2-Dichlorobenzene,
0 - 30%", it will have an actual 1,2-Dichlorobenzene content within that range.  However,
since the waste received for destruction might contain several compounds (and perhaps even
1,2-Dichlorobenzene under various names), the percentages when added up by upper range
percents could total more than 100%. In such  cases, each constituent of the waste will  have
been prorated and assigned a point-value, with the point values adding to 100%.  The ranges
Volume IE                                                             External Review Draft
Appendix III-l                              H-9                         Do Not Cite or Quote

-------
of percent content of most waste shipments seem to be consistently on the high side (possibly
due to caution on the part of the shippers), especially in the case of chlorinated compounds.
       Once the volumes of constituents containing analytes such as chlorine are revised to
reflect actual analytical values rather man arbitrary percentages of waste volume, the
reported quantities of these constituents in the data base drop. This drop is significant,
varying from several percent in the case of Lithium-containing constituents, to nearly two
orders of magnitude in the cases of Silver-containing and Thallium-containing constituents.
       Other sources of potential error include the omission of ash, batteries, and
miscellaneous categories in the elemental analysis.  The deletion of waste streams comprising
less than 0.5% of waste volumes early in data analysis may also sacrifice some
representativeness.  In addition, 13 analysis results listings are incomplete and contain "Not
Analyzed" entries for some analytes or "Not Detected" entries.  "Not Detected" entries do
not include the limits of detection or elaborate on testing methods and are entered as values
of zero  in subsequent calculations.

£.  Limitations
       Potential sources of uncertainty in the WIT data base are due  to the listing of
constituents in conjunction with percentage ranges of content rather than analytical values.
The lack of standardization in naming wastes further complicates the effort to form a
complete picture of the wastes incinerated over a period of one year.  The revision of
volumes of waste fed to reflect the analytical values available decreases, but does not
eliminate, these concerns.
       The fact that the data base containes waste input projections for a period of one year
means that the analysis represents average values.  Daily or hourly deviations from the
"average" waste stream could lead to significantly different results upon incineration.
       The clarification of data derived from shipping information, as well as the revision of
theoretical waste feed values to reflect analytical results, makes the current version of the
data a more realistic representation and more useful for modeling efforts.  It is hoped that the
revisions also make the data base more meaningful and more comprehensible for future uses
than the original collection of waste profile sheets and analytical results.

F.     Uncertainty
       Key assumptions in the development  of the data bases are listed hi Table H4.  The
uncertainty associated with the volumes of constituents determined in the data base hi its final
form result from the following:
        •      The reconciliation of Analysis Results with the anticipated constituent
              composition for compounds containing correction factors between 0 and 0.1

Volume  III                                                              External Review Draft
Appendix III-l                               11-10                         Do Not Cite or Quote

-------
               (see Table n-3).  Very few compounds are affected; however,  compounds
               containing iodine, non-amenable cyanide, selenium, silver, antimony, thallium,
               and chromium are affected;

        •      The prorating of constituent ranges in isolated specific cases where the upper
               bounds of the constituent ranges for a waste stream totalled between 200% and
               1000% before prorating;

        •      The reconciliation of Analysis Results with the anticipated constituent
               composition for compounds containing with correction factors between 0.1 and
               0.5 (see Table II-3);

        •      The reduction in the size of the data base due to dropping those waste streams
               representing >0.5% of the original total waste input;

        •      The choice of physical state rather than packaging information as criterion for
               determining pumpability versus non-pumpability;

        •      The deletion of unidentifiable compounds;

        •      The deletion of the ash and "abs" categories; and
        •      The reconciliation of the analysis results with anticipated constituent
               composition for those compounds containing analytes with correction factors
              between 0.5 and 1.0 (see Table n-3).

        In general, the sources of uncertainty are due to limitations in the way hi which data
are reported hi the original waste profile, and normalization and other data processing to
make the data  base usable.  It is not anticipated that the resulting uncertainty is greater than
one order of magnitude for most chemicals.  It should be recognized however, that it is not
possible to quantify the uncertainty accurately, since it is  necessary to project  the types of
waste feeds likely to be received by the WTI facility over an extended period  of tune hi the
future.
Volume ffl                                                              External Review Draft
Appendix ffl-1                              H-l 1                         Do Not Cite or Quote

-------
                                   TABLE H-l
           ASSUMPTIONS MADE DURING DATA BASE REFINEMENT
 Description in Original Database
 •Antimony salt
 Arsenic salt
 Barium salt
 Calcium chromate
 Calcium salt
 Chlorinated Fluorocarbons
 DCFM
 He (hydrocarbons)
 Ketones
 Mixed organics, alcohols and amines
 PCBs
 TCFM
 Acetates
 Anhydride
 A,a-dimethyl benzyl hydroperoxide
 Chrome compounds
 Chlorinated Paraffin oils & waxes
 Creosote
 Notes
MRI Assumptions
SbCl3
AsCl3
BaCl2*2H20
CaCrO4
Ca«(OCl)2
Dichlorofluoromethane
C8H18 (octane)
Methyl Ethyl Ketone
(Ethanol +ethanolaniine)/2
PentachlorobiphenyP
Trichlorofluoromethane
CjI^Oj (acetate ion)
Acetic anhydride
Cr2O3 (Cr(ni)oxide)
Ci4H12O0_jN0-2c
       Per OHM/TADS
       Per HSDB; C-12 = 60% chlorine, C-23 =43% Chlorine.  Interpolated for
       general elemental ratio.
       Per HSDB; creosote =
              12% phenanthrene
              06% aliphatic hydrocarbons (assumed dodecane)
              62% polycyclic hydrocarbons (assumed naphthalene)
              09% nitrogenated hydrocarbons (assumed acridine)
              09% hydroxy-functional polycyclic hydrocarbons
              (assumed dihydroxynaphthalene)
Volume IE
Appendix ffl-1
n-i2
External Review Draft
Do Not Cite or Quote

-------
  Table H-2.  Analytical Values for Total Waste Streams (Pumpable & Non-Pumpable)
Analyte
Total water
Total Cl
TotalS
Total Fluoride
Total Ca
Total Bromide
Total Na
Total K
Total P
Total Zn
Total Cu
Total Ba
Total Li
Total Pb
Total Cr
Total Se
Total Cd
Total Tl
Total Ag
Total Ni
Total Sb
Total As
Total Be
Total Hg
Total I
Total
Total Ib/yr
3,496,624
1,758,883
75,805
28,305
46,493
75,373
16,618
8,932
5,447
9,071
6,951
11,318
353
7,371
1,782
2,498
2,054
2,474
1,105
365
510
302
6
24
0
5.558.661
Atomic Mass
Units
. 18.0
35.5
32.2
19.0
40.1
79.9
23.0
39.1
31.0
65.4
63.5
137
6.94
207
52.0
79.0
112
204
108
58.7
122
74.9
9.01
201
127

Total
g moles/yr
8.80e+07
2.25e+07
1.07e+06
6.76e+05
5.26e+05
4.28e+05
3.28e+05
1.04e+05
7.98e+04
6.29e+04
4.96e+04
3.74e+04
2.31e+04
1.61e+04
1.55e+04
1.43e+04
8.29e+03
5.49e+03
4.65e+03
2.82e+03
1.90e+03
1.83e+03
3.05e+02
5.36e+01
O.OOe+00
1.14e+08
Volume m
Appendix ffl-1
n-is
External Review Draft
Do Not Cite or Quote

-------
                           Table H-3. Analytical Correction
                                       Factors
               Total I
               Total CN non-amen
               Total Be
               Total Hg
               Total Se
               TotalCd
               Total Ag
               Total Ni
               Total sulfides
               Total As
               Total Li
               Total Sb
               Total CN
               Total Tl
               Total PCBs
               Total TCFM
               Total Cr
               Total Zn
               Total bromofonn
               Total DCFM
               Total Na
               Total P
               Total Cu
               Total Pb
               Total K
               Total Ba
               Total fluoride
               Total S
               Total Ca
               Total Bromide
               Total Cl
               Total water
Correction Factor

0.000
0.000
1.000
0.140
0.0013
0.045
0.036
1.000
1.000
0.032
1.179
0.049
1.000
0.02
1.000
1.000
0.039
2.303
1.000
1.000
1.836
0.467
7.199
0.123
9.244
0.305
1.000
0.231
2.947
0.015
0.463
1.000
Volume m
Appendix ffl-1
     n-i4
External Review Draft
Do Not Cite or Quote

-------
Table H-4
Key Assumptions for Chapter II
Assumption
In applying the waste profile data, individual waste streams
that comprise less than 0.5% of the total volume are deleted
In estimating the constituent content in a waste stream, the
upper bound of the reported range for each waste stream
constituent is used. To prevent the combined percentage
from exceeding 100%, the constituent content of the waste
streams are normalized.
The waste feed data is based on waste profile sheets for the
first nine months of operation.
Basis
The small quantity of these waste streams limit their effect,
so this simplifying assumption focuses the assessment on the
waste streams that present the most significant health hazard
Conservative estimate. Professional judgment based on
facility desing and operation, and predicted waste
characterization.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics and receiving patterns.
Magnitude
of Effect
low
low
medium
Direction of
Effect
underestimate
overestimate
unknown
Volume 1H
Appendix III-l
H-15
External Review Draft
 Do Not Cite or Quote

-------
       CHAPTER HI.  ESTIMATION OF UNMEASURED METALS
  EMISSIONS AND EVALUATION OF METALS BEHAVIOR AT THE
                              WTI INCINERATOR

A.     Concerns to be Addressed
       The portion of the March 1993 trial burn that focused on the emissions of metals was
designed and conducted using an appropriate protocol, but the emissions of several metals of
interest in the risk assessment were not measured. Thus, an appropriate method for
estimating the emissions was needed.

B.     Approach
       Different methods for estimating the emissions of the unmeasured metals are
evaluated and an appropriate modeling procedure is selected.  The modeling approach chosen
addresses the issues raised in the peer review comments.  It predicts the reactions and phase
transformations of metals in the incinerator using thermodynamic equilibrium calculations.
Condensation processes are modeled. The particle dynamics program MAEROS is used to
predict the evolution of the particle size distribution. Particle size specific estimates of the
removal efficiencies of the system are determined using the March 1993 trial burn data. The
modeling approach is used to conservatively estimate the emissions of the unmeasured
metals.
       The model is used to examine the impact of several operating parameters on the
predicted emissions. The model is capable of predicting metals speciation and  the distribution
of the metals across the range of particle sizes.

C.     Estimation Technique Selection
       The goal of this  effort is to provide reasonable estimates of the emissions of the
metals for which there are no trial burn data for use in the risk assessment. Emissions data
are not reported for three classes of metals. The first class consists of the regulated metals
for which the facility elected to comply with Adjusted Tier I feed rate limits. These metals
are silver (Ag), barium  (Ba), and thallium (Tl). The second class consists of toxic metals
identified in the regulations for Boilers and Industrial Furnaces burning hazardous wastes
(BIFs) but not included  in current guidance or regulations for dedicated hazardous waste
incinerators. These metals are nickel (Ni) and selenium (Se). The third class contains
unregulated metals that may be important in multi-pathway risk assessments. These metals
are aluminum (Al), copper (Cu), and zinc (Zn). It is useful to also consider the metals that
Volume III                                                          External Review Draft
Appendix ffl-1                             fll-l                        Do Not Cite or Quote

-------
are measured in the trial bum — antimony (Sb), arsenic (As), beryllium (Be), cadmium (Cd),
chromium (Cr), lead (Pb), and mercury (Hg). These metals provide information on the
validity of the estimation procedures since the actual emissions rates are known.

      Three potential estimation techniques were evaluated hi this preliminary analysis:
      •      Review of raw analytical data;
      •      Review of historical data; and
      •      Modeling.

      Each of these techniques is discussed below.

      1.     Review of Raw Analytical Data From Trial Burn
             The trial burn report provided information on the seven target metals for
      which Tier HI permit limits were needed.  Analytical laboratories, however, often
      generate at least semiquantitative data on non-target metals, when inductively coupled
      argon plasma (ICAP) emission spectroscopy is used for analysis. Instrumentation
      used for ICAP analysis is generally capable of simultaneously measuring over 30
      metals or other inorganic elements,  but the accuracy of those measurements depends
      on how specific standards and other quality control (QC) steps are used.
             It was thought possible that the ICAP raw data from the trial burn included
      information on at least some of the other eight metals of interest.  A detailed list of
      raw data and laboratory procedural items was obtained to determine if this was a
      method to estimate metals emissions.  Although the data may  have been only
      semiquantitative, they might still have been superior to values obtained from other
      emission estimation methods.  The major limitations of this search were:

             •       The waste feeds apparently were not analyzed for metals; and
             •      "Synthetic" wastes were developed from "pure" materials and used for
                    the trial  burn (along with spiking of  metal compounds and other toxic
                    constituents), so that many of the eight untested metals of interest may
                    not have been present hi the trial burn feeds.

             In spite of the above limitations, a quick review of the stack emissions raw
      ICAP data for any  of the eight metals of interest seemed appropriate because:

             •      Waste feed samples were obtained by the Ohio EPA and may have
                    been available for analysis;

Volume HI                                                            External Review Draft
Appendix Ifl-1                              flI-2                        Do Not Cite or Quote

-------
              •     Depending on'the purity of the spiking compounds used, there may
                    have been sufficient levels of some of the metals of interest present as
                    impurities to allow some evaluation of control performance; and
              •     Some of the feed components may have contained some of these metals
                    (e.g., the cellulose-based batch solid waste may have contained Cu or
                    Zn).

              If any of the metals of interest were found in the stack samples, then trial burn
       feed rate information for those metals could have been obtained or  estimated, as
       possible. For instance, there may have been published data on typical levels of trace
       metals in certain feed materials.
              However, appropriate data were not found to be available. Thus, it was not
       possible to use this  technique to obtain the necessary estimates.

       2.      Review of Historical Data
              A brief search for relevant historical data on relative metals  control
       performance was performed.  The main objective was to obtain, for a hazardous
       waste incinerator (or possibly other combustion device), metals control data (e.g.,
       input rates, emission rates, system removal efficiencies, and/or collection efficiencies
       across specific air pollution control devices).  Data were needed for any of the eight
       metals of interest, as well  as many of the remaining seven metals which were reported
       for the WTI trial burn. Ideally, the data should have been from an  incinerator
       equipped with similar air pollution control equipment (APCE) to the WTI incinerator.
              Several potential sources of control performance data were identified, where
       the metals measured included at least some metals which were measured during the
       March 1993 WTI trial burn and some which were not measured.  Unfortunately, none
       of these tests were performed on an electrostatic precipitator (ESP)-equipped
       hazardous waste incinerator.   Furthermore, no hazardous waste incinerators were
       identified with an APCE train similar to that at WTI (i.e., boiler, spray dryer,
       activated carbon injection system, ESP, quench, packed bed scrubber).  In fact, no
       incinerators with even a general sequence of a spray dryer followed by an ESP were
       identified, with one possible exception. Von Roll's Biebesheim facility in Germany
       was believed to have been outfitted to this general system, however, waste feed data
       were not available.  Thus, it was concluded that sufficient historical data to estimate
       emissions were not available.
Volume III                                                              External Review Draft
Appendix IH-1                              HI-3                         Do Not Cite or Quote

-------
       3.     Modeling of Metals Emissions
             Modeling can be used as a "tool" to combine relevant empirical data with
       general knowledge of the chemistry and physics involved to predict how the metals of
       interest behave. Several modeling approaches have been used to examine the behavior
       of metals hi waste incineration systems. These range from the comprehensive
       computational model of the behavior of metals in waste combustion systems that was
       developed by Barton et al. (1990) for the EPA's Risk Reduction Engineering
       Laboratory to semi-empirical approaches such as that developed by Biswas et al.
       (Biswas et. al 1992). The majority of the models assume that chemical equilibrium is
       maintained for the metals within the waste. The two main weaknesses of this approach
       are—detailed thermodynamic data on all species and complexes that may possibly
       form are required and may not be available; and, the model does not account for
       condensed phase non-idealities.  However,  the only successful general modeling
       approaches available are based on the equilibrium approach.

       4.     Conclusion
             Based on the review of different estimation techniques, the application of a
modeling procedure provides the best estimates of metals' behaviors. An approach sunilar to
that developed for the EPA based on the assumption that chemical equilibrium is maintained
throughout the incinerator system is deemed to be  the most appropriate.

D.     Emissions Estimation
       1.     Introduction
             This section consists of seven subsections as follows:

             •      Model Development
                    This subsection summarizes the current understanding of the processes
                    controlling metals' behavior  as they apply to the WTI incinerator. The
                    conceptual model and computational techniques used to estimate
                    emissions from the WTI incinerator are described. Assumptions and
                    simplifications used are discussed.

             •      Model Application
                    This subsection describes the data used and the resulting emissions
                    estimates. The sensitivity of the emissions estimates to variations in
                    selected parameters are investigated.
Volume III                                                            External Review Draft
Appendix III-l                              III-4                       Do Not Cite or Quote

-------
             •      Uncertainty Analysis
                    This subsection describes and quantifies the uncertainty associated with
                    the predicted emissions.

             •      Speciation
                    This subsection discusses issues related to prediction of the metal
                    species emitted.

             •      Other Topics
                    This subsection addresses the three additional limited topics related to
                    the assessment of the risk associated with the emission of metals from
                    the WTI facility.
                                 Evaluate the health risks associated with aluminum
                                 emissions.
                                 Model the ratio of Cr(III) to Cr(VI) in the incinerator
                                 stack and at the receptor.
                                 Evaluate the potential for trace impurities hi the scrubber
                                 water to become stack emissions.

             •      Conclusions
                    This section  summarizes the  predicted emissions estimates and the
                    limitations of the estimates. Recommendations for improving the
                    estimates are listed.

      2.     Model Development
             a.     Background - Trial Burn Data
                    Examination of the March 1993 trial burn data included in the trial
             burn report (WTI, 1993) provides insight into the behavior of the unmeasured
             metals. Emissions data for metals were only reported for the test conditions
             intended to maximize  the emission of metals (that is, Runs 1, 2, and 3).
             Combustion chamber temperature was maximized at this condition. The average
             temperatures for each  test  ranged from 2160° to 2200T (1180° - 1200°C). This
             temperature was measured at the inlet to the secondary combustion chamber
             (SCC) and thus is not the  actual primary chamber temperature. However, the
             measured temperature should be closely related to the actual primary chamber
             temperature.
Volume III                                                             External Review Draft
Appendix ffl-1                              HI-5                        Do Not Cite or Quote

-------
                    The system removal efficiencies (SREs) measured during the trial burn
             are summarized in Table III-l. SRE is defined as:
                                   SRE=1_Emssion
                                            Feed
                    "Emission" is the emission rate of a given metal on a mass basis and
             "feed" is the feed rate  of the metal on a mass basis.

             Three sets of metals are defined based on the SRE data.

             1.     Mercury (Hg). The SREs for mercury are very low (4.40 to 10.59 %).
                    Mercury is known to be a vapor even after the flue gases have been
                    cooled to ambient conditions.  The control system in place at the WTI
                    incinerator when the trial burn was conducted was  not effective at
                    removing mercury vapors.
             2.     Antimony (Sb), Arsenic (As), Beryllium (Be), Cadmium (Cd), and Lead
                    (Pb). The SREs for these metals are similar (99.97% to 99.995%). This
                    implies that similar mechanisms controlled the behavior of these metals.
             3.     Chromium (Cr). The SREs for chromium are very  high (99.99996%).

                    More information is obtained  by looking at the SREs for metals in
             Group 2 in greater detail. Figure III-l is a graph of the system penetration of
             each of these metals (system penetration=l-SRE). The metals are ranked in
             order of increasing penetration. The ranking agrees with the volatility ranking
             systems developed by the EPA and investigated at the Incineration Research
             Facility (Waterland and Fournier  1993).  This indicates that the observed SREs
             are influenced  by the volatility of the metals.

             b.     Current Theory
                    To understand the modeling approach used, it is useful to understand the
             mechanisms thought to control the behavior of metals in waste combustion
             systems. Several researchers have been investigating this behavior. Data
             generated by these programs were reviewed by Barton et.  al., (1993) Seeker
             (1990) and Linak and  Wendt (1993). These reviews synthesized a  mechanistic
             understanding  of the phenomena that control the behavior of metals in waste
Volume III                                                           External Review Draft
Appendix ffl-1                              IQ-6                         Do Not Cite or Quote

-------
             combustion systems. The complexity of the proposed mechanisms has increased
             over time. A simplified discussion of the current mechanistic understanding as
             it applies to the WTI incinerator follows. The sources cited above contain a
             more complete discussion of the general mechanism.
                    Figure III-2 illustrates the pathways metals may take through the WTI
             incinerator. Many interrelated mechanisms control the pathway a given metal
             will take. These mechanisms are dependent on the physical form and chemical
             composition of the original waste, the chemical speciation of the metal within
             that waste,  and the way in which the metals are dispersed within waste. The
             strong influence of physical state implies that the pumpable and non-pumpable
             wastes should be considered separately.
                    Most of the inorganic material present in the non-pumpable waste is
             expected to form slag (Eddings and Lighty 1992). The slag flows through the
             kiln and into an ash quench pit where it is quickly cooled to near ambient
             temperature. A small fraction of the slag may be entrained by the flue gases.
             Non-volatile metals present in the non-pumpable waste will remain in the slag.
             Most volatile metals present will vaporize. However, some of the volatile
             metals may react with the slag to form refractory complexes which remain with
             the slag (Eddings and Lighty  1992).
                    Pumpable waste is injected through a sludge lance, a slurry burner or a
             liquid burner. These devices break the pumpable waste stream into  a cloud of
             small drops. A large fraction of the ash formed by these drops will be entrained
             by the flue gases. The size of the particles formed will be a function of the size
             of the drops and the physical distribution of the ash within the waste. The non-
             volatile metals and the metals that interact strongly with the ash remain with
             the entrained particles. Most volatile metals present in the pumpable waste will
             vaporize.
                    Non-volatile metals from either the pumpable wastes or non-pumpable
             wastes may interact with components of the waste or the flue gas to form
             volatile compounds. Two types of reactions have been observed. In the first
             type,  reactive elements released during the incineration process interact with
             metals (Quann and Sarofinn 1982). The resulting species may be much more
             volatile than the original species causing the metal to vaporize. The second
             class of reactions occurs in the high temperature, reducing environment formed
             near the burning waste. This environment forms in nearly all waste combustion
             systems even if the incinerator is operated at overall excess oxygen conditions
             (Quann and Sarofinn 1982). The reactions principally involve the reduction of

Volume HI                                                             External Review Draft
Appendix ffl-1                              HI-?                        Do Not Cite or Quote

-------
            metal oxides. The newly formed species may have a different volatility than the
            original species.
                   After the primary chamber, flue gases contain metals on entrained
            particles and in the vapor state. Additional reactions may occur in the flue
            gases as a result of variations in temperature and gas  composition. These
            reactions may result hi the formation of compounds with different volatilities
            than the original compounds and subsequent vaporization or condensation of
            material. In addition, reactions may be promoted by the addition of reactive
          .  solids to the flue gases (Uberoi et al. 1990) such as activated carbon.
                   As the flue gas cools, metals will condense both homogeneously to form
            new particles and heterogeneously on the surfaces of the entrained ash particles
            (Senior and Flagen 1982). Homogeneous condensation (nucleation) produces
            particles that are much less than 1 fan in diameter (Friedlander 1977).
            Heterogeneous condensation tends to favor small particles due to their high
            surface area to volume ratio (Linak and Peterson 1984). Thus, the small
            particles have higher concentrations  of the volatile metals than the original
            waste. The concentrations of metals  such as silver, cadmium, lead and
            antimony in particles emitted from waste combustion facilities have been found
            to be 30 to  100 times higher than would be expected if no vaporization and
            condensation occurred.
                   The  particles formed collide  with one another and with the entrained
            ash. When the particles stick together, these collisions result hi coagulation of
            smaller particles into larger particles. Small particles  are extremely mobile, are
            present hi large numbers  and coagulate quickly. As the particles become
            larger, then- mobility and number concentrations decrease resulting hi a lower
            coagulation rate. Investigations of combustion systems have found that any
            particles smaller than about 0.1 um  quickly coagulate, while those larger than 1
             urn do not (Linak and Wendt 1993). Thus, the small particles formed by
            homogeneous condensation quickly  coagulate to form 0.1 um particles at which
             time the coagulation rate slows dramatically. Larger particles do not coagulate
             significantly in the incinerator. Due to these mechanisms, two groups of
             particles typically enter the ah" cleaning system.  One group ranges in size from
             0.1 to 1 um and is formed from the metals that vaporized and subsequently
             condensed.  The second group is larger than 10  um in diameter and consists of
             the material entrained in the incinerator (Linak and Wendt 1993).
                    The specific mechanisms responsible for removal  of metals from flue
             gases vary  from device to device. The principal mechanisms responsible for

Volume III                                                             External Review Draft
Appendix HI-1                              IE-8                        Do Not Cite or

-------
              removal of metals in any given device are generally well known and are
              described hi the many texts in the area (e.g., Wark and Warner 1981). Three
              features of these mechanisms are important to the development of an
              appropriate model. First, particle collection efficiency is a function of particle
              size. Second, most flue gas cleaning devices are least efficient at capturing
              solid particles that range hi size from about 0.1 /tm to 1 /*m (Wark and Warner
              1981). Third, vapor collection is a function of the solubility of the vapor in the
              fluid used.

              c.     Modeling Approach for WTI Incinerator
                    A Legrangian approach (the control volume moves with the flow of
              material through the reactor) is used as the basis for the model of the behavior
              of metals at the WTI facility. Figure III-3 illustrates the modeling approach.
              Two feed streams enter the primary combustor — pumpable wastes and non-
              pumpable wastes. Data on the bulk composition, concentration of metals and
              feed rate of each waste stream are provided to the model. In addition, the
              conditions (temperature and oxygen concentration) are provided. The primary
              chamber model uses these data to predict any reactions that may occur and the
              subsequent vaporization of any metals. The primary chamber model also
              predicts the quantity of material that is entrained. The specific techniques used
              are described below. The  data produced by the primary chamber model include:

              •      Quantity of each metal hi the residuals (slag);
              •      Quantity of air required;
              •      Quantity of each metal which is in the flue gases as  vapors;
              •      Entrained particles; and
              •      Flue gas flow rate and composition.

                    The flue gas  then flows into the secondary combustion chamber (SCC)
              carrying any entrained particles and vapors with it. The model is capable of
              predicting any physical or chemical changes that may occur in the SCC.
              However, the conditions in the SCC at the WTI facility are  similar to those hi
              the primary chamber so no changes are expected to occur. Thus, no calculations
              are performed for the SCC.
                    The processes which occur hi the flue gas cleaning system are simulated
              by two models.  The chemical and physical transformations that occur  are
              simulated by the quench model. The removal of particles and gases  from the

Volume III    .                                                        External Review Draft
Appendix III-l                              HI-9                        Do Not Cite or Quote

-------
             flue gases is simulated by the flue gas cleaning model.  These calculations are
             performed sequentially though the processes are simultaneous. Creation of a
             model that would account for both of these processes simultaneously is beyond
             the scope of this effort.  Previous work by Barton et al. (1990) indicates that
             performing the calculations sequentially probably has little impact on the final
             predictions and significantly reduces the computational effort and data required.
                   The flue gas composition predicted by .the primary chamber model is
             combined with data on the conditions in the quench by the quench model. This
             model predicts any reactions and condensation that occur when the flue  gases
             are cooled. The model predicts the quantities of metals present as vapors, fine
             particles (< 1 um) and coarse particles (>1 urn). The specific procedures used
             are discussed below.
                   The final calculation uses data on the efficiency of the flue gas cleaning
             system at removing each class of material from the flue gases to determine the
             emissions. Ideally, in the absence of measured emissions for metals not tested
             during the trial burn, models based on theoretical considerations  would be used
             to predict the control efficiencies. However, no suitable models are available
             for the devices present at the WTI incinerator. Thus, for metals not tested in
             the WTI trial burns, efficiencies derived from the trial burn data for metals
             which are tested are used as discussed below.

             d.    Waste Composition
                    The composition of each waste feed, including the concentration of the
             metals of interest, is one of the most important parameters in predicting the
             emission rates. The chemical composition profiles described in Chapter II are
             used. Data are available for all of the metals of interest except Al.
                    Although a wide variety of waste feed scenarios can be envisioned, a
             simplified approach is used for this task. A time-averaged (that is, annualized)
             waste composition for each of the two feed streams is used. These values are
             based on the potential annual production rates of each component of the
             facility's total waste stream. The composition of the pumpable and non-
             pumpable waste streams are summarized in Table III-2. The total waste feed
             rates and individual metal feed rates  determined are compared to existing
             RCRA permit limits. This comparison is summarized in Table III-3. The
             approach  assumes that each waste type will be fed at the  maximum rates
             envisioned and appears  to be reasonably conservative. However, it ignores
Volume III                                                              External Review Draft
Appendix ffl-1                              111-10                        Do Not Cite or Quote

-------
              short-term variations in the feed rate of any given metal. Details on the
              development of the waste composition are provided in Chapter II.
                    The annualized average feed rate described above is below the
              maximum allowable heat input for the system. Thus, a second case is examined
              in which the wastes would be fed at the maximum heat input rate allowed
              based on the design of the kiln (121 MBTU/hr on a higher heating value basis).
              This maximum feed rate is 4.10 times greater than the annualized average feed
              rate. The composition of each waste stream is not altered for this case.

              e.     Incinerator Operating Conditions
                    The values for key operating conditions in the primary chamber are:
                    Temperature = 1200°C; and
              •      Air to waste stoichiometric ratio =1.0 (no excess oxygen available).

                    Research indicates that the greatest potential for the vaporization of
              metals occurs at the flame front (Barton et al. 1990). At this location the
              temperature is maximized and there is no excess oxygen available (the feed to
              air stoichiometric ratio is 1). The temperature measured at the entrance to the
              SCC is assumed to be the maximum temperature.

              f.     Chemical and Physical Transformations
                    The chemical and physical transformations that each metal undergoes in
              the incinerator and in the quench are determined assuming that equilibrium is
              continuously maintained. Two computer programs capable of determining the
              equilibrium composition of a mixture are used. The programs are CETPC (an
              equilibrium calculation computer program developed  by NASA (Gordon and
              McBride 1994)) and HSC Chemistry for Windows Version 2.0 (a commercial
              equilibrium calculation program marketed by ARSoftware Corp.). The
              programs are used to determine the vapor pressure of each metal in the
              incinerator and in the quench. These thermodynamic  codes function by
              minimizing the Gibbs  free energy of a system given the elemental composition,
              temperature, pressure,  enthalpy and possible resultant species. A major
              limitation  is that erroneous results may occur if all important resultant species
              are not included (either by choice or due to lack  of sufficient thermodynamic
              data).
                    It should be noted that chemical equilibrium will  probably not be
              achieved in all sections of the incinerator due to  kinetic limitations. Studies

Volume III                                                            External Review Draft
Appendix ni-1                             IH-11                        Do Not Cite or Quote

-------
             conducted at the University of Utah and the EPA, among others, indicate that
             behavior not predicted by equilibrium models may occur in some systems
             (Linak and Wendt 1993; Eddings and Lighty 1992) due to the formation of
             solid solutions or refractory complexes. Thermodynamic data for these systems
             are not generally available. However, equilibrium-based models are effective
             for most systems and are the best approaches available.
                    Two kinetic limitations are incorporated into the reaction modeling. The
             compounds PbCl4, and CrO2Cl2 are commonly predicted to form at low
             temperatures by the thermodynamic programs. These predictions are probably
             not correct because of kinetic limitations (Linak and Wendt 1993 and Ihara et
             al. 1983). However, due to the lack of a validated test method for measuring
             these two species, test data confirming this is not yet available. Inclusion  of
             these compounds can make lead and chromium appear to be much more
             volatile than they are generally observed to be. Thus, these two  compounds are
             excluded from consideration during the thermodynamic  modeling.

             g.     Entrainment
                    No acceptable entrainment models are identified. Li  (1974) and Tackie
             et al. (1990) have examined entrainment in kilns containing granular material.
             Their models are semi-empirical and can not be directly adapted to a slagging
             kiln. Thus, an engineering assessment of the entrainment potential is required.
             It is assumed that no  entrainment of the slag occurred. All of the pumpable
             waste is injected through a sludge lance, a slurry burner or a liquid burner.
             These devices break the wastes into small drops that form solid particles as the
             volatile material is driven off.  Most of these particles are entrained. It is
             assumed that all of the pumpable ash is entrained.

             h.     Particle Dynamics
                    Two assumptions simplify the particle  dynamics considerations.

             •      All material that vaporizes and subsequently condenses, forms particles
                    with diameters of about 0.5 um; and
             •      No material condenses  onto the  surface of the entrained particles.

                    These assumptions are  based on an examination  of coagulation and
             condensation processes as described below.
Volume III                                                            External Review Draft
Appendix HI-1                             ffl-12                        Do Not Cite or Quote

-------
                    Coagulation. Coagulation processes nave been studied extensively over
              the past decade and many of the basic mechanisms are relatively well
              understood. Several models have been developed and some of the more recent
              approaches are summarized by Seigneur et al. (1986).  The MAEROS model
              was developed by Gelbard (1980) and is widely accepted. It accurately tracks
              the evolution of many types of aerosol over time. Its main limitations are that it
              does not account for condensation processes and sometimes fails to arrive at
              acceptable solutions for very short time periods (on the order of 0.1 s).
                    MAEROS is used to examine the evolution  of an aerosol that is  similar
              to that which would form in the WTI incinerator. It is assumed for the
              evaluation that only very small particles (0.01 to 0.05 um in diameter) are
              present initially. These represent the nuclei formed  by the initial wave of
              homogeneous condensation.  Figure III-4 illustrates the evolution of the aerosol
              hi a gas at 1400 K. In 0.5 s, nearly all the nuclei have formed particles  0.1 to
              0.5 um hi diameter. After 120 s, the aerosol has stabilized with all of the mass
              that started as nuclei having shifted to 0.5 to 1 um. Figure III-5  illustrates the
              aerosol evolution predicted if the gases are at 400 K (approximately the quench
              temperature). The growth of the nuclei is retarded.  After 120 s, only about half
              of the mass has grown to the 0.5 to 1 um range. These two cases define the
              coagulation rate  extremes that may be observed.
                    These findings are substantiated by other models and by experiments
              (Linak and Wendt 1973). In typical  gas residence times in an incinerator,
              bimodal particle distributions are usually generated. One mode occurs around
              0.5 um and is the result primarily of vaporization, condensation  and
              coagulation processes. A second mode occurs  above 1  um and is -the result of
              entrainment.

                    Condensation.  The relative importance of homogeneous condensation,
              condensation onto fume nuclei, and  condensation onto entrained  particles can
              be assessed using classical condensation theory. Nucleation rates can be
              estimated using the Becker-During equation (Friedlander 1977). Heterogeneous
              condensation can be estimated using diffusion and gas kinetics considerations as
              described by McNallan et al. (1981). At conditions typical of the incinerator,
              the relative rate of each type of condensation is shown hi Figure III-6.
              Heterogeneous condensation on the surfaces on fume particles is the
              predominant form of  condensation. When fume particles are present, almost no
              material condenses by either of the other two modes.

Volume III                                                             External Review Draft
Appendix III-l                              HI-13                        Do Not Cite or Quote

-------
             i.     Flue Gas Cleaning
                   The flue gas entering the cleaning system contains vapors, fine particles
             (0.1 to 1 urn) and coarse particles (>10 um). The efficiency of the control
             system at removing each class of material is assumed to be different. However,
             due to the complexity of the pollution removal system and the limited resources
             available, it is not possible to calculate the efficiency of the system for each
             class of material based on fundamental principles. Data from the trial bum are
             used to determine the needed efficiencies.
                   Two classes of vapors can be identified. Soluble vapors dissolve easily
             in the scrubbing solutions used in the flue gas cleaning system while insoluble
             vapors do not. The observed control efficiency for mercury hi the trial burn is
             assumed to be the flue gas cleaning system efficiency for removal of all
             insoluble metal  vapors. Mercury is generally present in flue gases in a form
             that is probably insoluble (Vogg et al.  1986). In addition, it is known to be
             present as a vapor. No metal mat is both soluble and known to be present as a
             vapor was used hi the trial burn. Thus, another monitored material that is
             present in flue gas as a soluble vapor must be identified to establish the ability
             of the flue gas cleaning system to remove this class of materials.  Two likely
             compounds are  identified - HC1 and SOX.  Of the two, SOX generally exhibits
             lower solubilities and lower removal efficiencies than HC1.  Thus, the removal
             efficiency of SOX is selected as a conservative estimate of the ability of the flue
             gas cleaning system to capture  soluble  vapors.  The relative solubility of each
             metal is determined based on the compound predicted to form by the
             thermodynamic calculations and summaries of compound solubility (Weast
             1974). Table III-4  summarizes  the soluble and insoluble metals.
                   The control efficiency for arsenic observed hi the trial bum is assumed
             to be the control efficiency for all fine particles (that is particles with diameters
             of 0.1 to 1  u,m). The assumption is justified by the following observations:

                   Arsenic  is volatile and it is likely that all of the arsenic present hi the
                   waste  during the trial bum vaporized; and
             •     Thermodynamic calculations indicate that arsenic will condense hi the
                   flue gas cleaning system. As discussed above, all material which
                   condenses will probably be present as fine particles.

                   Based on these observations, it is probable that all of the arsenic
             originally present in the waste  entered the flue gas cleaning system as fine

Volume HI                                                              External Review Draft
Appendix ffl-1                             111-14                        Do Not Cite or Quote

-------
             particles. Thus, the control efficiency of arsenic should be the same as the fine
             particle removal efficiency.
                    The trial burn data from chromium is used to determine the flue gas
             cleaning systems removal efficiency for coarse particles. Chromium is expected
             to be the most refractory metal examined hi the trial burn. Thus, the only
             mechanism which will result hi chromium being present hi the flue gas
             cleaning system is entrainment. As discussed previously, entrainment results in
             the presence of coarse (>10 mm) particles. However, chromium is divided
             between the pumpable waste and the non-pumpable waste. To be consistent
             with the assumptions that the ash from non-pumpable waste is not entrained
             and that all of the ash hi the pumpable waste is entrained, it is assumed that
             only the chromium from the pumpable waste would be present in the flue gas
             as entrained particles at the entrance to the flue gas cleaning system. Thus, the
             coarse particle control efficiency would be:

                                           Cr                                      (2)
                                           ~Crt
          CrE is the mass emission rate of chromium hi the trial burn and CrFp is the feed
          rate of chromium hi the pumpable waste stream. Table III-5 summarizes the
          control efficiencies used for each class of material.

   3.     Model Application

          a.      Preliminary Validation
                 To determine if the modeling procedure produces reasonable predictions, it
          is first used to predict the behavior of metals under the conditions used  hi the trial
          burn. The waste feed composition is summarized in Table III-2. The results of the
          comparison are summarized hi Figure  III-7. The predictions for arsenic, mercury
          and chromium are very close to the actual values. However, this result is to be
          expected because the control efficiencies are based on these metals. The predictions
          for all of the other metals observed during the trial bum also agree well with the
          observed values. The greatest difference is observed for beryllium where the
          prediction is about twice the observed value. In all cases, the predicted emissions
          are higher than the observed emissions indicating that the predictions are likely to
          be conservative.

Volume III                                                              External Review Draft
Appendix ffl-1                              HI-15                        Do Not Cite or Quote

-------
                It must be noted that this comparison is not a rigorous validation of the
         prediction procedure but is rather a preliminary gross check on the realism of the
         predictions. A more extensive validation of the procedure is needed to increase the
         confidence hi the results. However, a similar predictive approach has been
         compared with data from boilers and industrial furnaces (BIFs) and has been found
         to agree with measured values to within an order of magnitude over a broad range
         of conditions (Clark etal 1994).

         b.     Predictions
                The procedures are used to predict emissions from the incinerator under the
         assumptions described above.  Table III-6 summarizes the results. Emissions
         predicted for the annualized average waste feed rate and the maximum heat input
         feed rate cases are included in Table III-6. For the metals that were measured
         during the trial burn, it is possible to calculate predicted emissions using the SREs
         determined in the trial burn. The emissions predicted using that technique are also
         summarized in Table III-6.  It should be noted that extrapolation of trial burn data
         hi that manner is not generally conservative. However, the modeling results
         indicate that for the metals measured hi the trial burn, the specific incinerator
         conditions expected and the specific quench conditions used, the SRE values do not
         change with the metals' feed rate. Thus, for this limited case, the model indicates
         that extrapolation of the trial burn SRE data is conservative.

         c.     Sensitivity Analysis
                The sensitivity of the predictions to variations hi assumed values is
         examined. For each parameter examined, alternative values are selected and used as
         input  for the model. The results produced are  compared with the base case
         presented above. The following parameters are investigated:

          •     Feed rate;
          •      Combustion chamber temperature;
          •      Combustion chamber stoicbiometry;
          •      Waste chlorine concentration;
          •      Quench temperature; and
          •      Entrainment rates.

                 The behavior of each  of the 15 metals listed hi Table III-6 is examined.
Volume IE                                                             External Review Draft
Appendix ffl-1                              ffl-16                       Do Not Cite or Quote

-------
                 1.     Feed Rate
                        The impact of feed rate on the predicted behavior of each of the 15
                 metals is evaluated. Three feed rates for each of the 15 metals are examined
                 as follows:

                 •      Base case (the maximized heat input feed rates);
                        High feed rate (one order of magnitude greater than the base case);
                        and
                 •      Low feed rate (one order of magnitude lower than the base case).

                        Emission rates change proportionately for each metal and the
                 predicted SREs are constant. This result does not imply that, in general,
                 SRE is not affected by feed rate. It only implies that the feed rates
                 examined are  much higher or much lower than needed to saturate the flue
                 gases with any of the metals. If the feed rates are close to those needed to
                 saturate the gases, small changes in feed rate may have a significant impact
                 on the predicted SREs.

                 2.     Combustion Chamber Temperature
                        Four combustion chamber temperatures are examined — 1000°C,
                 1100°C, 1200°C and HOOT (1830T, 2010°F, 2200°F and 2550°F). The
                 temperature of 1200°C is used in the base case predictions reported above.
                 Figure III-8 summarizes the emission rates and system penetration predicted
                 for each chamber temperature. The emissions of Be, Cu and Ni are affected
                 by the change in temperature. At the two lower temperatures,  no beryllium
                 vaporized in the combustion chamber. The emissions decreases from 1 x
                 10'7 g/s in the base case to less than 1 x 10~8 g/s at the two lower
                 temperatures.
                       Vaporization of copper and nickel is reduced at 1000°C.  The
                 emissions of both of these metals is approximately one order of magnitude
                 lower at 1000°C than at any of the other temperatures. At the  higher
                 temperatures, Be, Ni and Cu are predicted to vaporize and condense
                 forming fine particles. At the lower temperatures, the metals will remain
                 with the ash. Thus, the changes in the SREs reflect the transfer of these
                 metals from the fine particles to the coarse particles.
Volume HI                                                             External Review Draft
Appendix HI-1                              ffl-17                       Do Not Cite or Quote

-------
                3.     Combustion Chamber Stoichiometry
                      Two primary chamber stoichiometric ratios are examined in addition
                to the base case in which the ratio is 1.0. The additional ratios are 0.8 and
                1.2.  Figure III-9 summarizes the results of the calculations. Both emission
                rates and system penetration are shown. Be and Ni are the only metals
                effected by the change in the quantity of air present. At low oxygen
                concentrations, the vaporization of Be is lower than in the base case due to
                the formation of reduced forms that are less volatile than the oxide. The
                predicted Be emissions are approximately one order of magnitude lower
                when the stoichiometric ratio is 0.8 than when it is 1.0. The predicted
                nickel emissions are about one order of magnitude lower  when a
                stoichiometric ratio of 1.2 is used. The SRE for nickel increases from
                99.977% to 99.997% at the higher oxygen  concentration.

                4.     Waste Chlorine Content
                      The impact of chlorine is examined. Emissions are predicted for a
                waste that did not contain chlorine and the results are compared with the
                maximum heat input case. The comparison is illustrated in Figure III-10.
                Cu, Ni, and Se are effected by the removal of all chlorine from the system.
                In the absence of chlorine, copper and nickel no longer vaporize in the
                primary combustion chamber. As in the previous two analyses, when the
                vaporization in the primary chamber is eliminated, the emissions drop by
                about one order of magnitude and the SREs increase dramatically.
                      Selenium, in contrast, still vaporizes in the primary chamber when
                no chlorine is present. However, the vapors condense in the quench if there
                is no chlorine present. This results in a decrease in emissions by about an
                order of magnitude and an increase  hi the predicted SRE from 99.7% to
                99.98%.
                      These results do not imply that the  other metals do not form
                chlorides that are more volatile than the oxides. Some, such as lead and
                cadmium, do form volatile chlorides (Barton et al. 1990). However, the
                temperature in the primary chamber is high enough to vaporize the oxides
                of these metals and the quench chamber is cool enough to condense the
                chlorides. Thus, the predictions are unaffected by the formation of the
                chlorides.
Volume III                                                            External Review Draft
Appendix III-l                             IE-IS                       Do Not Cite or Quote

-------
                 5.     Quench Temperature
                        Two quench temperatures are examined in addition to the base value
                 of 150°C. The additional values are 200°C and 400°C.  Figure III-ll
                 summarizes the comparison. Emission rates and system penetrations are
                 shown. Cd,  Sb, Tl and Zn are affected by the quench temperature. As the
                 temperature increases, condensation of these metals decreases. Emissions
                 increase correspondingly due to the low capture efficiency for vapors. Tl
                 emissions are one order of magnitude higher than the base case for both of
                 the higher quench temperatures. Zn and Cd emissions are one order of
                 magnitude higher than the base case at 400°C only. Sb emissions are
                 slightly higher than the base case at the highest quench temperature. When
                 the emissions rates increase, the predicted SREs decrease.

                 6.     Entrainment
                        The final parameters examined are the entrainment rates assumed for
                 the slag and the pumpable waste. In addition to a pumpable waste
                 entrainment rate of 1.0 and a non-pumpable waste entrainment rate of 0.0, a
                 pumpable waste entrainment rate of 0.5 and a non-pumpable entrainment
                 rate of 0.2 are examined.  Figure 111-12 summarizes the results. Cr and Al,
                 the two least volatile metals, are affected by the changes. Decreasing
                 entrainment of pumpable material decreased emissions and increased SRE.
                 Increasing the entrainment of non-pumpable wastes increased emissions and
                 decreased SRE slightly.

£.        Uncertainty Analysis
          This section briefly discusses the uncertainty in the predicted emission rates of
those metals for which no trial bum data are reported.

          1.     Modeling Assumptions
                 The model used to make the predictions reflects the current state of the art
          for estimating metals behavior hi waste incineration systems. However, several
          assumptions are  inherent hi the model which fundamentally limits its capability to
          precisely predict the emissions of metals. Some of the most important of the
          limiting assumptions are:
Volume III                                           -                 External Review Draft
Appendix III-l                              ffl-19                        Do Not Cite or Quote

-------
          a.      Thermodynamic equilibrium is maintained throughout the incineration
                 and flue gas cleaning system. This assumption is required because in most
                 cases the needed reaction rates are not known. Because sufficient kinetic
                 data are not available, it was not possible to precisely quantify the  impact
                 of assuming equilibrium is maintained. However, for most metals it is
                 unlikely that the assumption had any impact on the predicted emissions.
                 The quantity that vaporizes at incineration conditions is often independent
                 of the chemical form of the metal. Similarly the quantity which condenses
                 in the quench is also often independent of the chemical form. However,
                 care must be taken in applying this assumption for there are a few  specific
                 metals and condition where the chemical form can have  a significant
                 impact.

          b.      All important compounds are present in the thermodynamic database.
                 Recent experiments show that complex reactions between ash components
                 and certain metals are possible. The complexes which form have different
                 volatilities than the other forms of the metal and are not generally present
                 hi the thermodynamic data base. It is not believed that this has a major
                 impact on the predicted emissions hi this case.

          c.      The reactor outlet temperature adequately characterizes the
                 temperature to which the metals are exposed. In fact, the temperatures in
                 the incinerator vary significantly. The behavior of the metals, as
                 demonstrated by the sensitivity analysis can vary significantly with
                 temperature. Comparison of the emissions measured in the trial burn with
                 the model's predictions indicates that representative conditions have been
                 selected and that the model did not under predict emissions.

                 These assumptions are necessary and are  the best that could be made given
          the base of information available. However, they limit the fundamental precision of
          the model and render a rigorous error analysis premature.

          2.      Data
                 It is possible to evaluate the impact of variations hi the data used by the
          model on the predicted emissions. The data fall into  two classes:
Volume III                                                            External Review Draft
Appendix III-l                             111-20                        Do Not Cite or Quote

-------
          •      Site-specific data; and
          •      General data.

                 Site-specific data include such items as combustion chamber temperature
          and waste feed composition. General data includes thennodynamic data, gas
          viscosities, and similar parameters. The general data are more precisely known than
          the site-specific data and contribute little to the potential error in the predictions.
          Because of this, the impact of site-specific data only are examined. The following
          parameters are used in the model:

          •      Waste Composition and Feed Rate
                       Chlorine
                       Trace metals

          •      Conditions in Incinerator System
                       Temperature
                       Availability of oxygen
                       Entrainment
                       Quench temperature
          •       Control Device Efficiency
                       Vapors
                       Fine Particles
                       Coarse Particles

                 The effect of reasonable variations in several of these parameters are
          examined. These are waste feed rate, combustion chamber temperature,  availability
          of oxygen, waste chlorine concentration, quench temperature and entrainment rates.
          With the exception of waste feed rate, each of these parameters will affect the
          behavior of only a few of the metals. However, the impact on the metals affected
          can be significant (changes  in emission rates of more than an order of magnitude).

          3.      Removal Efficiencies
                 Control device removal efficiencies for vapors,  fine particles and coarse
          particles are determined based on the results of the March 1993 trial burn as is
          described above. The control efficiency for insoluble metals was estimated based

Volume III                                                              External Review Draft
Appendix ffl-1                              ffl-21                        Do Not Cite or Quote

-------
         on the March 1993 trial burn collection efficiency of mercury. The control device
         efficiency for collecting soluble metals was estimated based on the trial burn values
         for SOX. The efficiencies for fine and coarse particles are estimated based on the
         trial bum values for arsenic and  chromium.  The uncertainty in the predicted
         removal efficiencies is estimated by calculating the standard deviation observed
         during the three test runs conducted during the trial burn. Table III-7 summarizes
         the control efficiencies used and the associated standard deviations. Because the
         control efficiencies are estimated using data obtained during the trial bum in a
         series of tests repeated over a short time using a well controlled simulated waste, it
         is likely that the values in Table III-7 represent the smallest possible variability.  In
         operation over long times, it is likely that the variation in waste composition,
         emission values and perhaps control efficiencies would be much greater.

         4.     Uncertainty Estimates
                Based on the values in Table III-7 and the sensitivity study discussed above,
         it was possible to determine the  range of variation that may be expected in the
         predicted emissions rates. Table  III-8 summarizes this analysis. The minimum
         value is obtained by using the model  to predict emissions when all the data are at
         the values which produced the lowest predicted emissions. Thus, this prediction
         used the lowest reasonable entrainment rates, primary  chamber temperature, quench
         temperature, chlorine  concentration, and metals feed rates. The impact of oxygen
         concentration is relatively small  and is not included in this analysis. In addition, the
         control efficiency is assumed to  be one standard deviation greater than the average
         from the March 1993 trial burn. The  maximum value  is obtained using a similar
         technique. The highest reasonable  entrainment rates, primary chamber temperature,
         quench temperature, chorine concentration, and metals feed rates are used
         simultaneously. The control efficiencies are assumed to be one standard deviation
         less than the March 1993 trial burn averages.
                The feed rate  of each metal has a strong impact on the predicted emission
         rate. Variation in the  feed rate accounts for most of the variation observed in the
         predicted emission rates for each metal. For this analysis, it is assumed that actual
         feed rates could deviate from the maximum heat input feed rate by as much as an
         order of magnitude. This wide range of possible variation hi metals feed rates is
         based on the following observations:
Volume HI                                                              External Review Draft
Appendix ni-1                              HI-22                        Do Not Cite or Quote

-------
          a.     The concentration information in the data base used to determine the waste
                 composition is generally based on the results of a single analysis of the
                 waste. Trace metals concentrations in wastes typically vary widely.

          b.     The data base incorporates many assumptions about the type and quantity of
                 the wastes which would be available for incineration.

          c.     Normalization and other data manipulation is required to place all data on a
                 common basis. These manipulations are potential sources of variation.

                 Modeling of the behavior of metals in waste incinerators is still
          approximate. Much work is focused on refining the models, but it will be several
          years before more precise models are available. The current models are best used
          for predicting trends and estimating the  impact of changing operating conditions.
          They are not as successful at predicting  absolute emission rates.

F.        Speciation
          The model generates some very approximate predictions of the metal  species which
may form hi the combustion system. However, because of the many assumptions required to
produce these  predictions they are not considered to be sufficiently reliable to be included in
this analysis. The predictions are summarized in Attachment 3.1 for  informational purposes
only. Great care should be taken in any  quantitative use of this information.

G.        Other Topics
          Three additional, specific topics are addressed. These are:

          •       Evaluate the health risks associated with Al  emissions;
                 Model the ratio of Cr(III) to Cr(VI) in the incinerator stack at the receptor;
                 and
          •       Evaluate the potential  for trace impurities hi the scrubber water to become
                 stack emissions.

          1.     Aluminum Toxicity
                 The Combustion Engineering Work Group of the Peer Review Panel placed
          Al on its list of metals that should be included hi the multi-pathway risk
          assessment.  The group stated  that Al might play a role hi biological metabolisms or
          may interfere with the action of other metabolic metals. Though the toxicity of Al

Volume III                                                             External Review Draft
Appendix ni-1                              ni-23                        Do Not Cite or Quote

-------
         is low, the group felt that the emissions would be high enough to warrant attention.
         A brief literature review was conducted to determine the relative importance of Al
         emissions due to its direct toxicity. While Al is one of the metals thought to
         catalyze the formation of dioxins, consideration of this indirect effect is not part of
         this assignment.
                The major toxic effect of Al is from the inhalation of pure dust, which
         .causes pulmonary fibrosis and neurological disorders. These disorders are also
         associated with Al exposure during renal dialysis. It is postulated that Al is
         associated with the formation of Alzheimer's disease although it is not thought to
         be causative on its own. Excess Al is deposited hi bone and induces a form of
         anemia.
                Some epidemiological studies have shown that all  of the health effects
         mentioned above are more common in areas with higher levels of Al in the
         drinking water. Al is thought to be more biologically available hi these areas.
         Usually the human gut excludes Al. The total quantity hi the body is typically 25
         mg and the daily intake varies from 10 to  100 mg. Antacids, deodorants and  some
         processed foods contain high levels of Al.  Additional exposure may arise from the
         use of aluminum cooking utensils.
                Little appears to be known about the absorption of Al from  inhaled dust.
         Incinerator emissions of Al would presumably.be as fine particles. Fine particles
         themselves could be of significant concern but the aluminum on them would
         probably be in the form of aluminosilicates or oxides in which case the Al is not
         easily bioavailable. Furthermore, based on the refractory nature of aluminum
         compounds and the high control efficiency observed for entrained material,
         airborne concentrations of aluminum compounds from the incinerator system are
         expected to be extremely low at this facility.
                It appears that Al emissions should be of lower priority than other metals
         but can not be neglected.  The major problem hi the assessment of Al related  health
         risks is that little analytical data on the Al concentration in waste feeds are
          available. Without accurate feed data,  prediction of emissions is difficult.

          2.     Chromium Valence State
                 The comments prepared by the Peer Review Panel referred  to an EPA-
          developed model for evaluating the distribution of Cr between Cr+3 and Cr*6  in the
          stack and at the point of deposition. As a  first step, the Kearney Team contacted
          Don Oberacker at the U.S EPA's suggestion as well as George Huffman, both of
          EPA/RREL hi Cincinnati, Ohio. Neither person was aware of such an EPA model.

Volume III                                                            External Review Draft
Appendix ffl-1                             IH-24                        Do Not Cite or Quote

-------
                 The Kearney Team did, however, identify a report produced by the
          Research Triangle Institute for the California Air Resources Board that presents the
          results from both laboratory and field studies on the conversion of Cr*3 to Cr*6 in
          the atmosphere. The half-life of Or*6 was found to be on the order of one day.
                 The EPA reviewed the research efforts on this issue and suggested that no
          further effort be spent on searching  for a speciation model. However, the modeling
          performed on this work assignment  did predict the expected chemical form for
          each metal.  Almost all of the stack emissions would be expected to consist of Cr*6
          compounds based on strictly thermodynamic considerations. However, this
          depended heavily on the formation of CrO2Cl2. Data from studies of the corrosion
          of steel suggests that the oxychlorination of chromium produces CrCl2 and CrCl3
          and that these reactions occur slowly requiring several hours to convert significant
          amounts of Cr2O3 to the chlorides. If CrO2Cl2  is excluded from consideration as for
          the predictions reported above, nearly all of the Cr present is predicted to be
          present as Cr+3 compounds. However, in the risk assessment all chromium
          emissions are conservatively assumed to be Cr*6. It is  also assumed in the risk
          assessment that no atmospheric reduction of Cr*6 would occur.

          3.     Emissions from Scrubber Water
                 It is thought that emission of trace impurities from the scrubber water may
          contribute to the facility's total emissions. Data on scrubber water composition
          were requested for this activity, but only limited data were available. Due to lack
          of sufficient data, no further effort was applied to this  issue. The U.S. EPA
          believes this to be a minor potential source of emissions because this feed stream
          normally consists largely of carbon-treated storm water runoff.

H.        Conclusions
          Emission rates of the metals that were not  measured during the trial burn are
estimated. The estimates are based upon current understanding and scientific principles.
Realistically conservative assumptions are made when required and are summarized in Table
III-9. The predicted emissions are summarized in Table III-6. As with any theoretical model
where many assumptions are required, the predictions should always be used with caution.
Sensitivity analyses indicated that the predictions for most of the metals are not strongly
effected by changes in the assumptions.  The emission rates of a few metals are affected by
each varied parameter, but with no clear trend toward increased or decreased predicted overall
emissions.  Additional data would help refine the emission estimates. Kinetic modeling should
also be considered for future productions. A computer model such as the HCT program

Volume HI                                                             External Review Draft
Appendix ffl-1                              III-25                        Do Not Cite or Quote

-------
developed at Lawrence Livermore National Laboratory may be useful in refining predictions
of the chemical species formed.
 Volume ffl                                                              External Review Draft
 Appendix III-l                              ffl-26                        Do Not Cite or

-------
              Table 1H-1. Metals System Removal Efficiencies Measured
                            During the WTI Trial Burn
.
Metals

System Removal Efficiency, percent

Run 1
Sb 99.977
As 99.97
Be > 99.988
Cd 99.993
Cr 99.9996
Pb 99.995
Hg 4.40

Run 2
99.989
99.98
99.992
99.986
> 99.9990
99.989
4.62

Run3
99.993
99.98
99.992
99.982
> 99.9990
99.987
10.59

Average
99.986
99.98
> 99.991
99.987
> 99.9993
99.990
6.54
Source:  May 1993, WTI Trial Burn Report
Volume III
Appendix III-l
m-27
External Review Draft
Do Not Cite or Quote

-------
                          Table m-2.  Waste Feed Compositions Used in the Modeling


Element


Trial
Pumpable

Bum
Non-
Pumpable


Total

Input Rate, mol/min
Annualized Average
Pumpable

Non-
Pumpable

Total

Maximum Heat Input
Pumpable

Non-
Pumpable

Total

Nonmetals
C
H
O
N
Cl
F
Br
S
P
Si

-
-
-
-
-
-
-
-
21.6
.
-
-
-
-
-
-
-
-
84.5
3700
6600
551
0
766
0
0
47.3
0
106
.
-
-
-
-
-
-
-
-
23.6"
_
-
-
.
-
-
-
-
.
23. 6C
1180
2029
166
40
43
1.29
0.81
2.04
1.15
47.2
_
.
.
_
.
_
.
.
_
94.3
.
_
_
_
_
_
_
_
_
94.3
4840
8320
681
165
175
5.27
3.34
8.36
4.72
189
Nontoxic Metals
Ca
K
Na
Fe
Li
Toxic Metals
Al
As
Sb
Ba
Be
Cd
Cr
Cu
Pb
Hg
Ni
Se
Ag
Ti
Zn
Water
0
0
0
0
0

25.6
0.385
0.600
0
0.252
0.787
6.2
0
3.67
0.0075
0
, 0
0
0
0
-
0
0
0
0
0

0
0
0
0
0
0
18.5
0
0
0
0
0
0
0
0
-
0
0
0
0
0

25.6
0.385
0.600
0
0.252
0.787
24.7
0
3.67
0.0075
0
0
0
0
0
2040
0.980
0.177
0.143
4.75"
0.043

4.90'
0.031
0.0035
0.0728
0.0006
0.0006
0.0293
0.0922
0.0307
0.00009
0.0031
0.00041
0.00075
0.00312
0.0288
0
0.019
0.024
0.480
4.75°
0.0004


4.90C
0.00002
0.0001
0.00005
0
0.0152
0.0004
0.0022
0.00007
0.00006
0.0023
0.0268
0.00810
0.00732
0.0912
0
1.000
0.201
0.623
9.50
0.0434


9.80
0.0311
0.0036
0.0728
0.0006
0.0158
0.0297
0.0944
0.0308
0.00015
0.00564
0.0272
0.00885
0.0104
0.120
256
4.03
0.709
0.585
19.0d
0.178


19.6°
0.128
0.0144
0.291
0.0023
0.0024
0.120
0.378
0.126
0.00017
0.0128
0.00169
0.00308
0.0128
0.118
0
0.0772
0.0995
1.97
19.0°
0.0015


19.6C
0.0001
0.0004
0.0002
0
0.0623
0.0016
0.0090
0.0003
0.00025
0.00922
0.110
0.0322
0.0300
0.374
0
4.10
0.808
2.56
38.0
0.180


39.2
0.128
0.0148
0.291
0.0023
0.0648
0.121
0.387
0.126
0.00042
0.0220
0.112
0.0362
0.0428
0.492
1050
a       From waste constituents (not from air).
b       Assumed 50 wt.-pct. of ash to be Si.
c       Assumed amount in non-pumpable streams (from ash) to be the same as in the pumpable streams, i.e., assumed same total ash feed rate in non-
        pumpable streams as in pumpable streams (5,300 g/min for maximum heat input case) and same concentration in ash.  This was done because most of
        the non-pumpable streams were not  analyzed for ash content.
d       Assumed 20 wt.-pct. of ash to be Fe.
e       Assumed 10 wt.-pct. of ash to be Al.

Source: May 1993 Trial Bum Report.
        Volume III
        Appendix III-l
m-28
External Review Draft
Do Not Cite or Quote

-------
                          Table HI-3.  Comparison of Predicted Feed Rates
                                       with the Permit Limits




Metals
Sb
As'
Ba
Cd
Cr
Cu
Pb
Hg
Ni
Se
Ag
Tl
Zn

Permit Limit


Type
Tier HP
Tier HI
Adj. Tier I"
Tierffl
Tier m
NAe
Tier III
Tier in
NA
NA
Adj. Tier I
Adj. Tier I
NA

Feed Rate
' (lb/hr)
9.65
3.81
260
0.3
169
NA
100.4
0.2f
NA
NA
16
1.6
NA
Annualized Average Case9


Feed Rate
(lb/hr)
0.0582
0.309
1.29
0.0007
0.203
0.793
0.841
0.0027
0.0417
0.285
0.126
0.282
1.04
Fraction of
Permit Limit
(%)
0.6
8.1
0.5
0.2
0.1
NA
0.8
1.4
NA
NA
0.8
18
NA
Other Components
Cl
NA
2700
,201
7.4
Maximum Heat Input*


Feed Rate
(lb/hr)
0.234
1.27
5.30
0.003
0.834
3.25
3.45
0.011
0.171
1.17
0.517
1.16
4.25
Fraction of
Permit Limit
. %
2.4
33.3
2.0
0.9
0.5
NA
3.4
5.5
NA
NA
3.2
73
NA

823
30.5
Solid Waste8
Feed Rate
NA
10500
285
2.71
1170
11.1
a     Assumes 24 hour/day, 365 days/year operation.
b     Higher heating value in waste profile used to determine total heat input rate for the annualized
      average case of 29.5 MBtu/hr.  Maximum design heat input rate is 121 MBtu/hr, or 4.10 times
      higher than the average annual case.  Thus, the maximum rates are obtained by multiplying the
      average annual rates by 4.10.
c     Tier IE limits are assumed to be based on the average feed rate during the March 1993 trial burn,
      test condition 1.
d     Adjusted Tier I limits are based on an assumed dispersion coefficient of 1.5 j*g/dscm/g/sec.
e     NA  = Not applicable.
f     Calculated using the average feed rate during the March 1993 trial burn although little removal
      occurred and the emission rates were unacceptably high.
g     "Solid waste" is assumed to include wastes termed "solids" and "solids,  solid/liquid mix" in the
      waste profiles.

Source:  May 1993 Trial Burn Report.
      Volume m
      Appendix III-l
m-29
External Review Draft
Do Not Cite or Quote

-------
                    Table ffl-4.  Classification of Metal Vapors Used
                            to Determine Control Efficiency
Soluble Metals
As
Sb
Ba
Cd
Cu
Pb
Ni
Se
Ag
Tl
Zn
Insoluble Metals
Hg
Al
Cr
Be






Volume in
Appendix ni-1
m-so
External Review Draft
Do Not Cite or Quote

-------
                         Table m-5.  Control Efficiencies Used in Model
Material

Insoluble vapors
Soluble vapors
Fine Particles ( < 1
Coarse Particles  (> 10
Control Efficiency,  %

6
99.68
99.977
99.9973
     Calculated from May 1993 Trial Bum Report

     Volume HI
     Appendix in-1
  m-si
External Review Draft
Do Not Cite or Quote

-------
                     Table HI-6. Predicted Metals Emission Rates
Metals


Tier m Metals
Sb
As
Be
Cd
Cr
Pb
Hg
Other Metals
Al
Ba
Cu
Ni
Se
Ag
Tl
Zn
Predicted Emission Rates Using
Model, g/s
Annualized
Average Basis

1.7 x 10*
9.0 x 10*
2.0 x 10'8
6.8 x 10*
6.8 x 10-7
2.4 x ID'5
3.3 x lO"4

5.9 x 10-5
3.7 x ID'5
2.3 x 10-5
1.2 x 10-6
1.1 x 104
3.7 x 10*
8.2 x 10-6
3.0 x ID'5
Maximum Heat
Input Basis

6.9 x 10*
3.7 x 10-5
8.1 xlO-8
2.8 x 10-5
2.8 x 10*
1.0 x 104
1.4 x 10-3

2.4 x IQr*
1.5 xlO4
9.4 x 10-5
5.0 x 1CT6
4.7 x IQ4
1.5 x lO'5
3.4 x 10-5
1.2X10-4
Predicted Emission Rates Using
Trial Burn SREs, g/s
Annualized
Average Basis

1.0 x 10*
9.0 x 10*
8.1 x lO'9
3.8 x 10*
1.7 x lO'7
1.1 x 10-5
3.4 x 1C"4









Maximum Heat
Input Basis

4.2 x 10*
3.7 x 10-5
3.3 x 10-8
1.6 x 10'5
7.1 x 10-7
4.3 x lO'5
1.4 x lO'3









Volume III
Appendix III-l
m-32
External Review Draft
Do Not Cite or Quote

-------
            Table HI-7. The Observed Variation in the Control Efficiency of
                   Selected Metals During the March 1993 Trial Burn

Class of Material
Insoluble vapors
Soluble vapors
Fine particles
Coarse particles
Control Efficiency,
percent of weight
6 .
99.68
99.977
99.997

Standard Deviation
4
0.2
0.006
0.002
Volume III
Appendix III-l
m-33
External Review Draft
Do Not Cite or Quote

-------
            Table IH-8.  Possible Variation in Predicted Metals Emissions
                          Due to Uncertainty in Input Data


Metal
Al
Ba
Cu
Ni
Se
Ag
Tl
Zn
Predicted Emission Rate, g/s
Based on Most Probable
Data
2.4 x lO"4
1.5 x 104
9.4 x lO'5
5.0 x lO"6
4.7 x KT1
1.5 x 1C'5
3.4 x lO"5
1.2 x 10"4
-
Minimum
2.2 x lO"6
1.2 x lO'5
1.0 x 10-7
3.1 x 10'9
2.6 x 10*
1.2 x 10*
2.6 x lO"6
9.5 x 10-6

• Maximum
5.2 x 10'3
1.9x lO'3
1.2xlO-3
6.3 x 10's
4.7 x lO'3
1.9 x 10"4
4.7 x 10-3
1.7xlO'2
Volume HI
Appendix ni-1
m-34
External Review Draft
Do Not Cite or Quote

-------
                                                                                                                                 r
Table ffl-9
Key Assumptions for Chapter HI
Assumption
Metals feed rates can be estimated based on one year of
waste data
Annual average waste feed rate should be adjusted to
provide the maximum permitted heat input for the
incinerator
Modeling is the most appropriate method for estimating the
emissions of the metals which were not measured in the trial
burn
Thermodynamic equilibrium is maintained throughout the
incineration and flue gas cleaning system
All important compounds are included in the thermodynamic
database
All elements in the incinerator are intimately mixed
No condensed phase non-idealities occur
The air to waste stoichiometric ratio which best
characterized the region where metals vaporize is 1.0
Incinerator outlet temperature adequately characterizes the
temperature to which the metals are exposed
PbCl4 and CrO2Cl2 will not form
All metals that vaporize and subsequently condense are
found on particles 0.5 fan in diameter
Basis
Best available data
Most conservative reasonable assumption. Produced highest
metals feed rates which can be reasonably predicted.
Best available data
Most high temperature reactions are very fast. No detailed
kinetic data are available
Best available data
Best available data
Best available data
Metals are most likely to vaporize in the hottest regions of
the incinerator. These regions generally occur at the flame
front. The stoichiometric ratio at the flame front is 1.0
Best available data
Field data indicates that these thermodynamically stable
compounds are not present in significant quantities
Laboratory data and theoretical calculations indicate that
condensing vapors concentrate on particles between 0. 1 and '
1 /*m
Magnitude
of Effect
high
high
high
medium
medium
low
medium
low
low
medium
low
Direction of
Effect
either
either
either
either
either
either
overpredict
either
underestimate
underestimate
either
Volume III
Appendix III-l
                                                                 111-35
External Review Draft
Do Not Cite or Quote

-------
Table ffl-9
Key Assumptions for Chapter III
Assumption
All pumpable waste is entrained
No non-pumpable waste is entrained
Chromium capture efficiency is an appropriate indicator of
flue gas cleaning system's ability to capture coarse particles
Arsenic capture efficiency is an appropriate indicator of the
flue gas cleaning system's ability to capture fine particles
SOX capture efficiency is an appropriate indicator of the flue
gas cleaning system's ability to capture soluble vapors
Basis
Pumpable waste is atomized to form drops which
subsequently dry. The solid residuals form particles which
are too small to settle out of the gas stream
Non-pumpable waste is incorporated into slag. Most slag
retained in incinerator by viscous forces.
Chromium is a refractory metal that modeling indicates will
not vaporize significantly. Thus, chromium was probably
present only on the coarse particles reaching the cleaning
system
As a volatile metal that will vaporize at any condition
reasonably likely to occur in the incinerator and is expected
to condense in the flue. Thus arsenic was probably present
on only the fine particles reaching the flue gas cleaning
system. The capture efficiency was the lowest measured in
the trial burn
SOX will be present in the flue gases as a vapor. SO, are
soluble. The capture efficiency of SOX is lower than other
soluble vapors which were measured in the trial burn.
Thus, use of SO, is conservative
Magnitude
of Effect
low
tow
low
low
low
Direction of
Effect
overestimate
underestimate
overestimate
either
either
Volume HI
Appendix III-l
                                                                111-36
External Review Draft
Do Not Cite or Quote

-------
    0.030
    0.025
_  0.020

-------
   Metals In
  Pumpabte
   Wastes
   Metals In
Non-PumpaWe
   Wastes
                                                                                                                        STACK
                                                                                                   Fine Fly Ash
                                                                                                   (Condensed
                                                                                                     Vapors)
                                                                                         Residual Vapors
                                    Figure III-2.  Pathways available for toxic metals in the WTI incinerator.
                       Volume HI
                       Appendix III-l
                                                                 111-38
External Review Draft
Do Not Cite or Quote

-------
1 ™" ~ ~ ~ • ""• ™ • "» •" — f~ "' "t- ™ «- I I I !•• — _i
[ Auxiliary Fuel | Non-Pumpable Wastes
j Feed Rate J | Feed Rate
1 Composition • j Composition
i ! | Metals Feed Rates
\
Metals in Residuals
\ \

| Pumpable Wastes ,
| Feed Rate !
| Composition ,
| Metals Feed Rates j

Primary Chamber
Reaction
Vaporization
Entrainment
^=^~^=^
Air
Feed Rate
	 "si
Conditions
| Temperature
Stoichiometry
1



f ^
^ZTzr— ^
Metals in Rue Gases
Vapor
Entrained Particles

Quench
Reaction
Condensation
^-^
Metals in Flue Gases
Vapor
Fine Particles
Entrained Particles
illtMllltimillHMIIIIIIIIIIIIIIIIIIUMIItllllMMMIM

/ —^
m^
| Conditions • .
| Temperature .
I Stoichiometry .
J 	 I
^^fc
Flue Gas
Flow Rate
Composition
^*"
— • Input
•"-" Models
	 Output .
Removal Efficencies
   Soluble Vapors
   Insoluble Vapors
   Fine Particles
   Coarse Particles
Flue Gas Cleaning
                         Figure III-3.  Schematic diagram of modeling approach used.
           »T_1	lit
                                                                                    ,,.,,™,.! !)„..:„... r»—r.

-------
                                &
  Figure
         Hl-4.
External Review Draft
Volume HI

-------
            0   o
                                     ^
     Figure
            lil-5-
Volume HI

-------
1x10°
1x104
1x103
^ 1X102
(T
.1 1x10°
75
C .-]
^3
"1X10-2
1x10'3

1X10"4
^ . .^ ^*"5






-

—






'

^
V
-









'
f '<•
%%-•"
1. O
^.^
K
%%C v
N -.
•.
"^ -•
%
*tf [f ,
V^A">^ %"^
.- x

-

"
_
-
-


Conditions
Temperature, K 1400
Saturation Ratio (P/P ) 1000
eq
0.1 /AD Particle Concentration.
gr/dscf 100
10 pm Particle Concentration
gr/dscf 100





          Homogenous   Heterogenous   Heterogenous
          condensation   condensation    condensation
                          on 0.1pm       on 10pm
                          particles        particles
Figure ffl-6.  Relative rates for homogeneous condensation and heterogeneous condensation onto
              0.1 fim particles and 10 fan particles.
Volume III
Appendix III-l
m-42
External Review Draft
Do Not Cite or Quote

-------
   1x10"
    1x10"'
 jo
 O)
 "
 cc
 o
 11x10"
 LU
    1x10
    1X10"1
       ,-5
               Sb
As
Be
Cd
Cr
Pb
Hg
 Figure ffl-7.  Comparison of the model's predictions and the March trial burn results from the
               WTI incinerator.
Volume III
Appendix ni-1
                m-43
                                 External Review Draft
                                 Do Not Cite or Quote

-------
1x10u -
1x1 0~1 "
1x10'* '
^i
"S IxlO"4
CE
c
•i 1X10'5
OT
m 1X10"6
1x10'7
1x10'8
1x10^



i
1 * *
: !•
: »'
i .
t
t .
t
i
f
t
i .
•.

^^ 1000
rn HOG

;- .
5 .^


(I
I
' *• !
' £ "' I
' !•: *•
.'

°C H 1200°C(Base)
°C -ISS 1400°C

i ;

r: I 1: p i; \ I-: :;
'• i ' v ' i • i :: '.
^ ;:? < * '. > •; •.
'l •!,•• ! S " •! ..
: • ( . " < 5 . i :': < S .
* • '!?.'•: f '• ? S .
i 1: 1 • i Si i h •!
:: mi !• S : i !- i !; -
- H I: j • \ K M- •:
• I: I- i : \ I- n: ••

-
t *-
^ - '' \ : '•
' . i . i ' '

|i||!|;i ; ;:
' '. 't '•:-.'• 't ''. •''••:
I :\^ '• ;: ':

             Al    As  Sb  Ba  Be   Cd   Cr  Cu  Pb  Hg   Ni   Se  Ag   Tl   Zn
     1x10°
     1x10
     1x10'
   Q.
   E
     1x10'
   CO
     1x10"
     1x10"'
                                                                JLJe
              Al   As   Sb   Ba   Be  Cd   Cr   Cu   Pb  Hg   Ni   Se  Ag   Tl   Zn
Figure HI-8.  The impact of temperature on the predicted metals emissions and the observed
              SREs.
Volume III
Appendix III-l
m-44
External Review Draft
Do Not Cite or Quote

-------
        1x10'
       1x10
        1x10
      
-------
   1x10°
 £
 co


 11X10-4
 JD

 09


 E 1x10'5
 ID
   1X10"
   1x10
   1x1 cr
       '7
                      O  Base



                      D  No Chlorine

     t_
            Al   As   Sb  Ba   Be   Cd   Cr   Cu   Pb  Hg   Ni   Se  Ag   Tl   Zn
1x10"
1X10"1
c
.2
Slx10'2
i
s. .
E
|1X10'3
1X10"4















rn









i-

1
1



































1

i
1

















































































































I




























Ifl









:|

1
f












1









'.;.;

i
I


























V















































1
x;
1
1












n






:*:••

•:•:•
•:i:
;

;;




































•?:

'*
'•V.
































:•?;
;.;.;
i


















-


—




1 •

I
*;
             Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag   Tl   Zn
  Figure ni-10.   The impact of chlorine on the predicted metals emissions rates and SREs.
Volume III

Appendix ni-1
m-46
External Review Draft

Do Not Cite or Quote

-------
         1x10°
        1x10"
        1x10"'
      •
      75

      c 1X10"4
      o
      '55
      —     c
      E1X10'5
        1x10M
        1x10
        1x10''
            ,-7
                      150°C

                      (Base)


                      200°C



                      400°C.
                Al   As   Sb   Ba   Be   Cd   Cr   Cu  Pb  Hg   Ni   Se   Ag   TI   Zn
        1x10°
            1-1
        1x10
.f

£ Ix10"2
CD
      E
      I
        1x10~
        1x10"'
                           150CC
                           200°C



                           400eC
                Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb  Hg   Ni   Se  Ag   TI   Zn
Figure ffl-11.  The impact of quench temperature on the predicted metals emissions rates and

               SREs.
Volume HI

Appendix m-1
                                   m-47
External Review Draft

Do Not Cite or Quote

-------
        1x10°
       1x10'
       1x10'
     jo
     o>1x10-3-
     £
     "5
           ,-2  -
       1X10"1
           ,-7
       1x10
        1X10"1
        1x10°
       1X10'1
      |1x10'2
     "5
     "3
     |lx10'3
      
-------
    CHAPTER IV. ESTIMATION OF ORGANIC EMISSIONS FROM
                           THE WTI INCINERATOR

A.     Introduction
       This section describes the steps taken to estimate emissions of organic compounds
from the WTI incinerator.  At the time of the original WTI project plan (U.S. EPA 1993b)
and Peer Review (U.S. EPA 1993c), there were very few stack test data available for
nondioxin products of incomplete combustion (PICs) at WTI, and it was proposed to estimate
these emissions for use in the risk assessment.  While developing a viable estimation
technique, the U.S. EPA also talked with WTI representatives about the importance of
conducting actual  stack testing for nondioxin PICs.  In late August 1994, and again in
December 1994, WTI did perform a very comprehensive series of PIC speciation tests,
including a total of 16 test runs. When the results of the August  1994 PIC tests were
received, the  U.S. EPA decided to use these actual data in lieu of the estimates which were
being prepared. However, it was decided that the estimates would still be important for
filling in gaps in the August 1994 measured data, and for putting the measured data into
perspective.  Therefore, a master list of estimated emissions was assembled, and the
methodology  for this estimated list is described in this section.
       The master list of estimated emissions is based partly on calculations from waste feed
information, and partly on earlier and less detailed (and potentially less representative) test
data obtained during the trial burn testing.  The list of estimated emissions is subsequently
used as a basis for calculating the possible magnitude of the potential "uncharacterized
fraction" of nondioxin organic PIC emissions, an element necessary for the uncertainty
analysis of this risk assessment.
       To the extent that the results of actual sampling and analysis (including detection limit
values hi nondetect situations) from the August 1994 test series are available, this risk
assessment preferentially uses those actual values.  In situations where an analyte was not
detected during that series of tests, U.S. EPA assumes that emissions of that PIC exist at 1/2
the detection  limit for the  typical case, and that they  exist at the full detection limit for the
high-end case.  Only when these data sources are not available for a particular hazardous
constituent is  information from the master list of estimated emissions employed in the risk
assessment.
       The results of the December 1994 PIC testing were not received hi time to be
numerically averaged into the risk assessment calculations.  However, those data were
reviewed and found to be  comparable to the August 1994 data. A U.S. EPA compilation of
the data from both those tests is available from Region 5.
Volume III                                                           External Review Draft
Appendix ffl-1                             IV-1                       Do Not Cite or Quote

-------
       The following discussion is divided into four sections.  The first section deals with the
approaches the U.S.  EPA evaluated and eventually uses for creating the list of estimated PIC
emissions from the WTI stack, and for estimating the potentially uncharacterized fraction.
The second section addresses the related issue of potential effects of the WTI air pollution
control devices on these emissions.  The third section discusses the most appropriate way of
estimating dioxin emissions for use in the  risk assessment (dioxin/furan emissions were
addressed separately  because of the different formation mechanisms involved).  The fourth
and final section discusses uncertainties in the estimated data.

B.     Estimation of Organic Emissions

       Estimating Emissions and Uncharacterized Fraction
       Data on organic compound emissions that WTI reported as being quantitated during
the trial burns are limited to eight compounds, plus dioxins and furans, as shown in Table
IV-1.  Specifically, there were four principal organic hazardous constituents (POHCs) and
four products of incomplete combustion (PICs).  During the trial bum, a test waste stream
containing the POHCs was made up specifically for the purposes of a worst case test of the
incinerator, and this  resulted in a chemical composition that is not deemed representative of
the annual waste profile for the WTI facility. Because of this, the measured POHCs  in the
                                                                                           •-,__
stack gas are  not necessarily realistic estimates of day-to-day emissions. But since the
concentrations of PICs in stack gas are not as directly related to waste composition as are
POHCs, the facility-specific trial burn PIC emission data are considered to be the reasonable
estimates for  the PIC compounds.
       Total hydrocarbon (THC), as methane,  was reported at slightly less than 1 ppm
during the March 1993 trial burn test.  If  the maximum concentration values for the
measured compounds on Table IV-1 are compared to the THC concentration, a large portion
of the THC mass appears to be identified. However, one important factor needs further
evaluation because it might change this conclusion significantly.  Specifically, the response
factor of chlorinated compounds to the flame ionization detector (FID) used for THC
measurement is lower than for hydrocarbons (e.g., carbon tetrachloride has about half the
THC response of methane).  If most of the PIC emissions happen to be chlorinated
compounds, the THC values might be expected to be biased low.  Related to this concern is
the observation that the concentrations of  carbon tetrachloride and tetrachloroethylene varied
widely across the different trial burn test conditions at WTI, but the observed variation in the
THC values was very small.  This issue will be discussed later in this Chapter.
       Because so few PICs were identified during the March 1993 trial burn, several            ^
approaches were considered to further identify or estimate emission levels of specific

Volume III                                                             External Review Draft
Appendix IH-1                              IV-2                        Do Not Cite or Quote

-------
compounds.  Perhaps the simplest approach would have been to assume that all of the
organic mass identified by the THC analyzer comprises the most toxic species, i.e.,
dioxins/furans.  This is the most conservative approach, but it would also be extremely
unrealistic and would likely result in an overestimation of the risk by at least several orders
of magnitude. The most accurate technique for further identifying PICs would be additional
stack sampling and analysis for a long list of target compounds, but this approach is very
resource intensive and time consuming.  In addition, it is likely that even the best sampling
and analytical program would still not provide complete knowledge of all of the organic
components of the stack emission.  Within this range of approaches, several additional
approaches were considered to provide a more realistic estimate. These approaches are
described below.
       The first approach was an idea suggested by the Peer Review Panel to gather
historical waste composition data from WTI and to apply combustion chemistry knowledge to
predict PICs that would be emitted. This approach would require a specific model or
procedure which could both predict the compounds emitted and estimate the emission rates.
       This approach was investigated further.  Phone contacts were made with incineration
or combustion experts to solicit their views on estimating PIC emissions from knowledge of
the waste composition or by other methods.   The experts contacted included William Linak,
EPA/AEERL; Robert Thurnau and Gregory Carroll, EPA/RREL; and Daniel Chang,
University of California/Davis.  The preliminary consensus from these experts was that no
method had, to date, been shown to quantitatively predict PIC emissions from knowledge of
waste composition. Most of these respondents suggested that specific data for the facility, if
available, or historical data on hazardous waste combustion might provide the best basis for
emission estimates.
       This investigation did identify data where the identity (although not the quantity) of
several PICs had been predicted from  knowledge of waste burned in research tests (viz., one
full-scale hazardous waste incinerator and an EPA pilot-scale incinerator (U.S. EPA 1991,
U.S. EPA 1992).)  Although sampling and analysis did indeed detect these  predicted PICs hi
the incinerator stack gas, the specific compounds predicted were those PICs generally
regarded as  being very common in combustion systems.  These data suggested that it may be
possible to predict the likely identity of several major PICs based on major waste
constituents, but it was concluded that this technique would  not be a better predictor than
Volume HI                                                            External Review Draft
Appendix III-l                             IV-3                        Do Not Cite or Quote

-------
simply using known common PICs.  Also, the specific procedure used for these predictions
was not reported.
       A second approach which the U.S. EPA and its consultants evaluated was the
evaluation of raw data from the gas chromatograph/mass spectrometer (GC/MS) analyses
performed as part of the WH trial burn.  These analytical data would be used to further
evaluate the presence or absence of compounds of interest by allowing  the U.S. EPA to
determine if the GC/MS instrument recorded any additional evidence of volatile and/or
semivolatile compounds hi the stack gas samples.  For compounds present in high enough
quantities, a rough estimate of emission rates could potentially be determined.  For
compounds not found, minimum quantitation limits could be estimated, allowing the
assumption that the compounds are not present above a specified level.
       A second potential use of the GC/MS data would be to estimate the portion of the
mass associated with different classes of compounds. If a specific percentage of the mass
could be associated with a class of compounds that seldom contains toxic compounds, then
this percentage of the overall mass of organic emissions could be deducted from the total
mass assumed  hi the risk assessment.  Although this approach would not result hi a complete
characterization of all organic compounds emitted, it could potentially allow more realistic
assumptions to be made about the unidentified mass.
       Because a complete file  of the raw GC/MS data could not be obtained within the time
available, the GC/MS data approach was only partially implemented.  The most important
information to come from this element
of the evaluation was an estimate of the practical detection limits.  These were later used hi
estimating PIC emissions.
       A third approach which the U.S. EPA evaluated was the use of available historical
emission data from other hazardous waste incinerators.  This would not provide a complete
site-specific characterization, but it was thought that it might allow more realistic
assumptions than would the use of certain more theoretical approaches. This approach was
eventually deemed less reliable than the other methods and was therefore dropped.
       Based on this initial analysis, it was decided to take the following approach to develop
an estimate of the organic nondioxin PICs and to also estimate the potential magnitude of the
"uncharacterized fraction" of the total mass of organic emissions.

       1.      Estimated PIC Emissions
              The trial burn reports for the March 1993 and the February 1994 (retest of
       condition 2) trial burns  indicate that of the PICs identified for sampling and analysis,
       only benzene, chloroform,  and tetrachloroethylene are detected in quantifiable
Volume III                                                            External Review Draft
Appendix HI-1                              IV-4                        Do Not Cite or Quote

-------
       amounts.1  In addition to these measured PICs, WTI submitted summary data from
       an additional PIC analysis which had been conducted on samples collected during the
       February 1994 trial burn. These data include results of GC/MS analysis of VOST
       and semivolatile samples collected during the retest.  The average of runs 2, 3, and 4
       of this trial burn is used to determine the emission rate for each analyzed compound
       under test condition 22. Nondetect compounds are assumed to exist hi the stack gas
       at half of the detection limit. While WTI provided detection limits for the volatile
       compounds, detection limits are not reported for the semivolatile compounds.
       Detection limits for these compounds are estimated at 2 jig based on back-up GC/MS
       data files received from WTI (Sigg  1994b).
             Residual POHC emissions are calculated based on the annualized waste profile
       and on the destruction and removal  efficiency (DRE) measured during the trial burns.
       This technique is essentially a simplified version of an approach suggested by the Peer
       Review Panel. The annualized waste profile discussed hi Chapter n is used to
       estimate emission rates of each compound on an annual basis, after that profile is
       corrected3 to reflect projected operation at full thermal capacity 365 days per year
       and 24 hours per day.  Of these compounds, only the compounds that are on the U.S.
       EPA's list of target PICs (U.S. EPA 1994 and Mercer 1994) are considered hi
       calculating residual emissions. Estimated emissions are calculated hi grams per
       second (g/s).  For those organic compounds whose emissions were estimated using the
       waste profile data base, the annualized volume of each compound is  adjusted upwards
       by a factor of 4.1 to reflect maximum throughput of waste through the incinerator.
             Of the three test conditions in the March 1993 trial burn, the DREs  for
       condition 2 are lowest4, and these values are also slightly lower than the DREs
       1  Of the three test conditions for the March 1993 trial burn, condition 2 (a low kiln
temperature test) had the highest emissions for these three compounds.  Condition 2 of the
March 1993 trial burn demonstrated higher emissions of tetrachloroethylene than the
February 1994 retest of condition 2.  The February 1994 retest had higher emissions of
benzene, and emissions of chloroform were about the same for both tests.

       2   Because run 1 was stopped prematurely, the U.S. EPA judged that the run 1 test
results might not be representative.

       3   This correction was necessary in order to "scale up" the waste feed rate to full
capacity from the relatively small quantity actually received during the first year of
operation.

       4  In fact,  it was condition 2 of the March 1993 trial burn where one of the  POHCs
failed to achieve the required DRE of 99.99% during two of the three test runs.
Volume III                                                             External Review Draft
Appendix m-1                              IV-5                        Do Not Cite or Quote

-------
      measured during the February 1994 retest of condition 2.  The average of nine values
      (three runs, three POHCs) from condition 2 measured during the March  1993 trial
      burn is used to represent a reasonable worst case DRE.  This DRE is then applied to
      the emission rates of organic compounds identified in the waste profile (and included
      on the U.S. EPA's target list of PICs) to calculate residual emissions of these
      compounds.
            A compound specific estimate is made for 129 of the compounds  on EPA's list
      of target PICs (including PCDD/PCDF TEQ as one compound).  The breakdown of
      the number of compounds selected from each source of data is shown in Table IV-2
      below.  Stack gas concentrations that correspond to the estimated emission rates
      presented hi  Attachment 4, range up to 400 ng/L for tetrachloroethylene.  These
      values are hi  the same range as historical measurements  for hazardous waste
      incinerators (USDOC 1984).
            Attachment 4 to this Appendix presents the complete list of compound-specific
      emission rates that were compiled from the various sources along with the final
      estimated emission rate for each compound.

      2.    Estimation Procedure for the Uncharacterized Fraction
            The approach to estimating the potential magnitude of the uncharacterized
      fraction of organic PIC emissions is based on comparing the total mass of organic
      compound emissions, as derived from recorded total  hydrocarbon (THC) values, with
      the mass that would result from the emissions estimated in the  above subsection.
      Total hydrocarbon emissions measured during the trial burn are mathematically
      increased to account for compounds thought to be biased low by the design of the
      typical THC analyzer, and the resulting values are believed to be a reasonable
      conservative (high-end) estimate of total mass of organic compounds.
            Total hydrocarbon emissions measured during the March 1993 trial burn are
      about the same for all three test conditions.  The emissions range from 0.68 ppm to
      0.89 ppm, measured as methane. The average of the three test runs under condition 2
      (0.70 ppm) is used in this analysis to be consistent with the earlier selection of
      condition 2 as the worst case for DRE.  Total hydrocarbon measurements are only
      available from one other test (viz., the February 1994 trial burn retest of condition 2).
      The THC recorded during the February 1994 test is much lower  than 0.70 ppm, and
      it is therefore judged more conservative to use the values recorded hi March 1993.
             The selected THC value is reported in units of ppm and pounds per hour of
      methane in the trial  burn  report.  This mass/time value is  converted  to units of grams
      per second (g/s), converted to the average molecular weight of compounds hi the

Volume III                                                           External Review Draft
Appendix ni-1                             IV-6                        Do Not Cite or Quote

-------
       emissions estimate (versus the molecular weight of methane, the calibrating gas), and
       adjusted to account for known differences in response of the flame ionization detector
       (FID) to the various organic compounds in the emissions list when compared to that
       of methane. The Ib/hr values reported for the THC monitor are calculated from a
                                           i
       volume (molar) concentration that is directly measured by the monitor and the
       molecular weight of the compound (methane) used to calibrate the monitor, which
       converts the value to a mass rate. To cpnvert this value to Ib/hr of the estimated
       emissions, the Ib/hr value, as methane, is multiplied by a factor representing the
       molecular weight of the estimated emissions divided by the molecular weight of
       methane. An additional correction is necessary for FID response because an FID
       responds differently to different compounds  depending on the molecular structure of
       the compound (U.S. EPA 1979).  To simplify the molecular weight conversion and
       FID response correction, they are base<^ on the molecular weights and structures  of
       the compounds that make up 90 percent of the mass of emissions.  The result is an
       increase in the THC emission rate by a factor of 3.08, (of which 2.72 is the
       molecular weight correction) relative to the actual measured value.
              Furthermore, the THC value in g/s is increased by a factor of 2.87 to account
       for the organic compound mass that is typically not measured by such a THC
       monitor.  The THC monitoring system used at WTI, like most such monitors, has  an
       ice temperature condensate trap on the sample line. The  factor of 2.87 is based on
       research findings that THC measurements using such a THC monitor  are typically
       less than half the measurements obtained using a simultaneous alternate total organic
       mass measurement procedure.  This difference is explained in the findings of that
       research as being largely due to volatile compounds, likely water  soluble, that
       condensed with water hi the  condensate trap on the sampling line  for the THC
       monitoring  system (U.S.  EPA 1988).
              The result of this  analysis is that the  uncharacterized fraction is approximately
       60 percent5 of the total mass of organic stack emissions.  Figure IV-1 presents the
       possible components of the organic compounds emissions estimate.  The  first column
       in Figure IV-1 shows the total mass of organic compound emissions that is
       5   It should be emphasized that because the THC value used in this analysis has been
conservatively increased by several factors to account for potential instrument bias, and
because it has been observed at other facilities that much of the THC recorded by  an FID-
type monitor is due to non-toxic methane and ethane, this estimated total organic emission
rate would represent a reasonable upper bound on the emissions of toxic PICs.  While it is
unlikely that the fraction of uncharacterized toxic emission is this high, such an upper bound
is useful for the purposes of examining uncertainty.
Volume III                                                             External Review Draft
Appendix III-l                              IV-7                         Do Not Cite or Quote

-------
      represented by the adjusted THC value.  This total organic mass is divided into
      several components for the purpose of this analysis, and these are qualitatively
      described in the second column of Figure IV-1.  These components include estimated
      emissions from residuals of waste compounds, PIC emissions measured during the
      trial burn, de minimus estimates for compounds below detection during the trial burn,
      the potential uncharacterized mass fraction, and dioxin and furan emissions measured
      during compliance tests. The third column in Figure IV-1 presents the number of
      compounds based on each  category.  The fourth column presents the percent of total
      mass represented by each category of compounds in the emissions estimate.
            At the time when this evaluation was originally underway, the uncharacterized
      fraction was conservatively assumed to be similar in composition and potency to the
      known carcinogenic fraction  of the nondioxin PICs.  This approach would have
      included multiplying each  of the estimated emission rates for each identified nondioxin
      carcinogen by a correction factor (sometimes referred to as "prorating") to bring it up
      to a level which would account for the "unknown" mass.  For example, if 50 percent
      of the total mass of organic emissions  was unaccounted for,  the estimated emission
      rate for each identified carcinogen would have been multiplied by 3.  However, as
      this evaluation developed and evolved, it was eventually decided to not prorate the
      emissions hi this way because it would overemphasize carcinogenic impact and ignore
      toxic (i.e., non-carcinogenic) impacts. Therefore, the decision was made to instead
      discuss the impacts of any potentially unknown mass in the uncertainty section of the
      risk assessment.
             Figure IV-2 presents a detailed flowchart of the steps taken in estimating PIC
      emission rates and hi comparing this estimate to the adjusted THC value.

C.    Effect of Control Device
      In evaluating organic compound emissions from the WTI facility, it is important to
consider the potential effect of the control device on these emissions.
      The effect of the control devices on nondioxin organic compounds is most associated
with two factors. First, particle removal will reduce the emissions of heavier organic
compounds that may have condensed on particles. Second,  most organic compounds have
the ability to adsorb onto the activated carbon which is injected for dioxin control.  Neither
of these factors can easily be quantified,  but both factors could result hi a reduction hi PIC
emission rates.
      No data are known at this time which show the amount of removal for specific
organic compounds (except dioxin)  that might result from carbon injection.  Some data,
however, are found for THC levels on a medical waste incinerator equipped with carbon

Volume HI                                                            External Review Draft
Appendix ffl-1                              IV-8                        Do Not Cite or Quote

-------
injection in a spray dryer/fabric filter air pollution control system (U.S. EPA 1992).  Data
are available before and after the control device for three test runs with and without the
carbon injection.  Inlet levels of THC ranged from 1  to 6 ppm.  Outlet levels are generally
about half the inlet levels, either with or without operation of the carbon injection system.
These data show that there is no large removal effect of a carbon injection system on the
bulk of combustion THC emissions.  Since no other data are found which could be used to
evaluate the potential removal of specific nondioxin compounds by the control device, a
conservative assumption of no removal is used in this risk assessment.
       It is not deemed necessary at this point to evaluate the theoretical effect of the air
pollution control train on the emissions of dioxins/furans, because the effectiveness of the
pollution control system in collecting these materials is demonstrated at WTI through
extensive  stack sampling and analysis.

D.     Emissions of PCDD/PCDF
       Estimating polychlorinated dibenzodioxin/polychlorinated dibenzofuran
(PCDD/PCDF) emissions is discussed separately because the factors that affect the emission
levels are significantly different than those for most other organic compound PICs. In
particular, it is presently believed that most PICs are formed hi the combustion chamber of
an incinerator, but that most PCDD/PCDFs are formed in the ductwork and pollution control
devices which follow the combustion unit.
       As with estimating emission levels for other compounds, the best estimates will be
based on measured emission levels.
       When this risk assessment project first got underway, there was insufficient emission
data to determine whether the newly installed enhanced carbon injection system (ECIS)
would reliably reduce the emissions of dioxins and furans.  Since that time, additional testing
has been performed. As of this writing, there have been 37 dioxin/furan test runs since the
installation of the ECIS.
       Because the WTI facility will be operating with the ECIS hi place, and because
repeated testing has confirmed the effectiveness of the ECIS, the risk assessment is based on
emissions data from the post-ECIS installation tests.  Specifically, average emission rates for
the 17 dioxin  and furan congeners (see Volume HI, Table ffl-2) are calculated as the
arithmetic mean of the emission rates measured in the 26 post-ECIS installation test runs
conducted between August 1993 and August 1994. Dioxin and furan congeners not detected
during a specific run are assumed to be present at one-half of the detection  limit for the
congener  during that run.  It is believed that in this way, emissions of dioxins and furans will
not be understated.
Volume III                                                             External Review Draft
Appendix ffl-1                              IV-9                        Do Not Cite or Quote

-------
E.     Uncertainties
       Figure IV-2 presented the procedure used to estimate the emission rates for organic
compounds for the WTI incinerator.  Figure IV-3 is identical to Figure IV-2,  except that
several data sources and steps in the procedure are numbered.  These numbers correspond to
the key assumptions listed in Table IV-3.  For each factor hi Table IV-3, the magnitude of
effect is categorized as low, medium, or high.
       The types of key assumptions are exemplified by the two factors identified relative to
item 1 in Table IV-3 (THC data) which may lead to uncertainty hi the analysis.   The first
factor is the representativeness of the incinerator operation during  the test when the THC
data were collected (i.e., does long term operation differ from operation during the test, such
that the long term THC data would differ from the value used for the emissions estimate?)
The second factor is the uncertainty associated with the measurement method. Most of the
factors in Table IV-3 are similar to these two factors.
       The measurement uncertainty associated with emission rate estimates for compounds
reported as not detected in analytical data sets (items 7 and 9) are  judged high.  This is
because the true value could range from zero to the detection limit; an infinite factor.  These
emission rates, however, are very small and may not be significant. The other factors
judged to have high uncertainty are items 10 and 14. The uncertainty associated with these
items relates to the uncertainty of identification of compounds, not to the magnitude of the
emission rate. Uncertainty associated with identification of compounds is likely the greatest
area of uncertainty hi the organic compound emissions estimate.
 Volume III                                                             External Review Draft
 Appendix ffl-1                             IV-10                        Do Not Cite or Quote

-------
               TABLE IV-1.  POHCs, PICs, AND THC MEASURED
                        DURING THE WTI TRIAL BURN
Analyte
POHCs:
Monochlorobenzene
Carbon tetrachloride
Trichloroethylene
1 ,2,4-Trichlorobenzene
Identified PICs:a
Benzene
Chlorobenzene
Tetrachloroethylene
Methylene chloride1"
SUBTOTAL (POHCs and PICs):
THC (as methane):
Concentration range,
ng/L
<2.5
4.4-220
<2.5
0.64
5.7-39
5.3-35
< 2.5-490
< 2.5-39
< 17-820
450-590
(0.68-0.89 ppm)
        aTarget PICs included seven compounds (three
         volatile compounds), plus an additional
         12 compounds in the class of tentatively
         identified compounds (TICs). Only compounds
         which were detected/identified are included.
        "Designated as a TIC.
VOLUME in
APPENDIX in-1
IV-11
EXTERNAL REVIEW DRAFT
  DO NOT CITE OR QUOTE

-------
                                                           Table IV-2
                                                   Sources of Emission Rates
Source
Estimated emissions from residuals of waste compounds
2*--DREs applied to waste profile data, DREs from condition 2 of 3/93 trial burn
PICs measured during trial burn
lA*-Condition 2 of 3/93 trial burn
lB*»2/94 retest of trial burn condition 2
3A*-Other volatile PICs, 2/94 retest of condition 2
4A*--Other semivolatile PICs, 2/94 retest of condition 2
Total

De minimus estimates for compounds below detection during trial burn
3B*--Other volatile PICs, 2/94 retest of condition 2 (nondetected)
4B*--Other semivolatile PICs, 2/94 retest of condition 2 (nondetected)
Total

Total dioxins and furans measured during compliance tests, TEQ
No. of values
54
2
1
12
3.
20
14
40
54
1
"•--Designation used in Attachment 3.
Volume III
Appendix III-l
IV-12
External Review Draft
Do Not Cite or Quote

-------
                                                                V  J
TABLE IV-3
Key Assumptions for Chapter IV
Assumption
THC values are representative of incinerator operation and
appropriate method of measure was used(l)
Factor accurately reflects sample loss(2)
Factor accurately reflects instrument response (3)
Trial burn data (Condition 2 of 3/93) are representative of
incinerator operation and waste composition and appropriate
method of measurement was used (4)
Trial burn data (Condition 2 of 2/94) are representative of
incinerator operation, waste composition and appropriate
method of measurement was used (5)
Semivolatile and volatile PICs reported (detects and
nondetects) on 7/1/94 are representative of incinerator
operation and waste composition and appropriate method of
measurement was used (6,7,8,9)
Waste profile database is representative of compounds in
waste stream (10)
DREs from 3/93 trial burn are representative of incinerator
operation and appropriate method of measurement was used
(11,12)
Waste profile database is representative of concentrations of
targeted compounds (12)
Basis
Best available data
Best available data
Best available data
Conservatively high values selected from best available data
Conservatively high values selected from best available data
Conservatively high values selected from best available data
Best available data
Conservatively high values selected from best available data
Best available data
Magnitude
of Effect
medium
low
low
medium
medium
medium to
high
high*
medium to
low
medium
Direction of
Effect
variable
variable
variable
variable
variable
variable
variable
variable
variable
Volume III
Appendix III-I
IV-13
External Review Draft
Do Not Cite or Quote

-------
                                                              TABLE IV-3
                                                   Key Assumptions for Chapter IV
                       Assumption
                        Basis
Magnitude
 of Effect
Direction of
   Effect
  Dioxin/furan data from 8/93 test are representative of
  incinerator and APC operation and waste composition and
  appropriate method of measurement was used (13)
Conservatively high values selected from best available data
medium to
   low
  variable
 nign relative magnitude out smau ausoiuic magnitude
Volume III
Appendix III-I
          IV-14
       External Review Draft
       Do Not Cite or Quote

-------
Toiai Mass of Organic
Compound Emissions
 Total Hydrocarbons
 (THC) Measured
 During Trial Burn*
CategonWbff Compounds in
Emissions Estimates	
                                        Estimated Emissions
                                        from Residuals of Waste
                                        Compounds
                                        PICs Measured During
                                        Trial Burn
  Deminimus Estimates for
  Compounds Below
  Detection During Trial
  Burn
                                        Unknown Fraction
                                        Assumed Similar in
                                        Composition to Known
                                        Carcinogenic Fraction
                                         Dioxins and Furans
                                         Measured During
                                         Compliance Tests
 Number of
Compounds
                                 54
                                 20
    54
                               Unknown
                                 Tetra
                                through
                                 Octa
                               Congeners
PercentV. Total
     Mass
                     20.9
                     19.9
       0.6
                      58.6
                    2.3E-6
     •Reported value increased to account for compounds not detected or not accurately measured.

                       Figure IV-1.  Components of organic compound emissions estimate.

              Volume III
                             External Review Draft

-------
                             3 PIC'S Irom 3/93 Inal
                               burn. Condition 2
                                                      3 PIC's Irom 2/94 lelasl ol
                                                        (rial burn. Condition 2
                            Volatile PIC's reported
                           7/1/94 by WTI (dala Irom
                            2/94 trial burn relesl)
THC values lrom3/B3 trial
    burn. Condition 2
                                                                                                               Semivolalile F'ICs reported
                                                                                                                                                                           compounds in annualized
                                                                                                                7/1/94 by wn (dala Irom
                                                                                   Divide reported detection
                                                                                   Imlls by two lor all lesl
                                                                                   run/compounds where no
                                                                                    values were reporled
Calculate average over
three lesl runs, convert
       to g/sec
Calculate average over
three lest runs lor each
 PIC, convert to a/sec
                             Calculate average over
                             Ihree test runs lor each
                              PIC. convert to a/sec
                                                                                                                                                                                                        Calculate average ORE
                                                          over nine reported values
Divide detection limit by 2
                                                                                                                                                                                                       Calculate residual POHC
Multiply THC by laclorol
                                                                                  Calculate average lor each
                                                                                                               Calculate average lor each
                                                                                                                                                                                                       emissions lor compound
Multiply THC by laclorol
                                                                                                                Usl emission rates (g/sec)
                                                                                                                 lor all compounds; celecl
                                                                                                                compounds me present In
                                                                                                                more than cmt data source
                                                                                 Total the emission rates
                                   mass ol uncharactertzed
                                                                                        Figure IV-2.  Procedures for estimating emissions of organic compounds.
                                                                         Volume III
                                                                         Appendix III-l
                                                                                                                           IV-16
                                                                                                                                                              External Review Draft
                                                                                                                                                              Do Not Cite or Quote

-------
                            3 PIC's liom 3/93 maj
                              burn. Condition 2
                                                   3 PIC's liom 2/94 relest ol
                                                    trial burn. Condition 2
                          Volatile PIC's reported
                         7/1/94 by WTI (data (ram
                           2/94 trial burn relesl)
THC values Iron, 3/93 trial
   burn. Condition 2
                                                                                                         Semlvolalili. PICs reoorled
                                                                                                                                                                 compounds m annuafiznd
                                                                                                          7/1/94 by ml (data Irom
                                                                              Divide reported detection
                                                                               Mis by wo lor all lest
                                                                              run/compounds where no
                                                                                values were reported
                                                                                                                                                                                            Calculate average ORE
Calculate overage over
three lest runs, convert
      log/sec
Celcuiate average over
three test runs lor each
 PIC, convert to g/sec
                           Calculate average over
                           three test runs lor each
                            PIC, convert to g/sec
                                                       over nine reported values
Divide detection (mil by 2
                                                                                                                                                                                            Calculate res dual POHC
 Multiply THCcy teetotal
                                                                                                         Calculate average lor each
                                                                              Calculate average lor each
                                                                                                                                                                                            emissions lor compounds
 Multiply THC by lactorol
                                                                                                           Us! emission tales (g/sec)
                                                                                                           lor ell compounds; select
                                                                                                          compounds are present In
                                                                                                          more that on* data source
                                                                              Total the emission rales
                                                                       ' Figure IV-3.  Procedures for estimating emissions of organic comppounds (estimation steps
                                                                        with uncertainties highlighted
VOLUME  III
APPENDIX  3-1
                                                                       Volume III
                                                                       Appendix III-l
                                                                                                                                                      External Review Draft
                                                                                                                                                      Do Not Cite or Quote
                                                                                                                      IV-17
                                                                                                                                                                                           WORKING DRAFT  3  -  NOT  FOR RELEASE
                                                                                                                                                                                                               DO  HOT  CITE  OR QUOTE

-------
                 CHAPTER V.  ESTIMATION OF EMISSIONS FROM OTHER
                                                SOURCES

           A.     Evaluation of Emissions from Automatic Waste Feed Cutoffs

                  1.     Concerns to be Addressed
                        One of the critical factors affecting health impacts is the assessment of
                  emissions during abnormal operation and accidents, including transients due to non-
                  steady state operation, system upsets that could result in a waste feed cutoff, fugitive
                  emissions due to leaks and spills, and catastrophic accidents such as fires and natural
                  disasters.  The Peer Review Panel recommended an alternative approach be developed
                  to address these scenarios,  if necessary.

                  2.     Approach
                        The approach for the analysis of emissions from automatic waste feed cutoffs
                  (AWFCOs) is a two step process. The first step is to develop an estimated number of
f                 AWFCOs per year for the WTI facility.  To the extent possible, the estimate of
                  occurrences is based on the facility reports of AWFCOs and general non-compliance,
                  as supplied by U.S. EPA Region 5.   This step includes an examination of facility
                  information to identify any trends in the  occurrence rate or types of cutoffs which
                  might have an impact on emissions to the atmosphere.
                        This initial step also includes a review of available AWFCO data from other
                  operating incinerator facilities to determine if any of the data could be applied to
                  develop the WTI-specific estimate.
                        Once an estimate of AWFCO frequency is developed, the second proposed
                  step is to develop an estimate of the chemical composition of the emissions.  Although
                  established estimation techniques or  studies dealing directly with nonroutine emissions
                  from AWFCOs could not be identified, several sources of data are tentatively
                  identified to develop this estimate.
                        These potential sources were evaluated for their usefulness and are briefly
                  described below:
           Volume III                                                           External Review Draft
           Appendix ffl-1                              V-l                        Do Not Cite Or Quote

-------
                    A primary source of information is the February 1994 trial burn
                    results.  According to U.S. EPA Region 5 personnel, during Run 1 of
                    this test, an AWFCO occurred, due to an exceedance of the
                    temperature setpoint.  The results of this run are included in the trial
                    burn, and are compared with three "normal" runs performed as part of
                    the same condition.
             •      Another source of information involves the waste stream reported by
                    the facility to be in the combustion chamber when the waste feed cutoff
                    occurred.  This information might be used to draw conclusions about
                    unburned constituents in the fugitive release.
             •      An experimental study was performed at the Incineration Research
                    Facility (IRF) in Jefferson, Arkansas.  The purpose of this study was to
                    simulate and quantify emissions from an incinerator operating under
                    "upset" conditions, much the same way  an incinerator might be
                    operating prior to or during a cutoff.  This research report was, in part,
                    never finalized because some of the analytical equipment broke down
                    during the experimental testing, and resources were not available to
                    repeat the testing at a later date.   Specifically, the total hydrocarbon
                    (THC) analyzer broke down, and the experiments were completed
                    without the benefit of that equipment or the resulting THC emission
                    data. Although the research concluded that emissions do not
                    significantly increase during an AWFCO, some have questioned
                    whether this conclusion  can be supported without the availability of
                    THC emission data (Whitworth 1992).  Because of these questions, the
                    results or conclusion of  this research have not been used here.
             •      In 1988, the Midwest Research Institute (MRI) conducted a study at a
                    local incinerator.   In part, the purpose of this study was to assess the
                    relationship between CO and HC during upset conditions hi the
                    incinerator. It was thought that this test could yield some information
                    about organic emissions during upset conditions.
             •      In the late 1980s, MRI was involved in  a study commonly referred to
                    as the "TME", or  total mass emissions study. Although one of the
                    objectives of this study was to determine the effect of upset conditions
                    on mass emissions from an incinerator,  this study in general did not
                    show a great difference  in emissions, probably because the facility

Volume HI                                                             External Review Draft
Appendix ffl-1                              V-2                         Do Not Cite Or Quote

-------
                               could not be forced into enough of an upset condition to yield usable
^                             data (U.S. EPA 1987).

                        In addition to gathering the data for the initial research, additional information
                 was sought by contacting several people considered to be experts in the combustion
                 emissions field.  Those contacted are as follows:
                        Larry Wateriand. Acurex Corporation:  Mr. Waterland is responsible for the
                 operation of the IRF, under contract to U.S. EPA.  He was also responsible for the
                 previously described test at the IRF, and has many years of experience in the field of
                 combustion research.  Mr. Waterland was unaware of any  other studies or data
                 relating to quantifying emissions from AWFCOs.
                        Dr. Paul Lemieux and Dr.  William P. Linak.  EPA RTP:  Both Dr.  Lemieux
                 and Dr. Linak are considered to be experts in the field of combustion research.  They
                 provided a series  of published articles on research conducted based on a 250,000 Btu
                 rotary kiln simulator.  This information is evaluated in this study and is discussed
                 below.
                        The above-described approach provides an estimate of the number of AWFCO
                 events per year at WTI and to quantify emissions during an AWFCO.  A detailed
/"               discussion of the process is described below followed by a summary of conclusions
 "               and a discussion of uncertainties inherent in the analysis.

                 3.  Frequency/Emissions Estimate for AWFCOs

                        a.     Estimated AWFCO Frequency
                               As identified previously, the first step in this analysis is to develop an
                        estimate of the annual number of AWFCOs at the WTI facility (WTI 1994).
                        To develop this estimate  Table V-l summarizes a log of AWFCO events
                        submitted  by the WTI facility to Ohio and U.S. EPA. A more complete
                        summary on all AWFCO events is presented hi Attachment 5. As shown in
                        Table V-l, 396 separate AWFCO events occurred,  for varying reasons, over a
                        span of approximately 9 months.  Extrapolating from the occurrence of 396
                        AWFCOs  over a period of 9 months, an annual frequency is estimated at 528
                        AWFCOs  per year (396 X 12/9 = 528).  Of the 396 reported AWFCOs, 108
                        are reported to have been associated with overpressure releases from the kiln.
                        Extrapolating as hi the previous case, it is assumed that 144 of the total
 ^~ •-•;
(          Volume in                                                            External Review Draft
^         Appendix ni-1                              V-3                        Do Not Cite Or Quote

-------
            number of AWFCOs at the facility would cause a kiln overpressure which
            could result in a release through the kiln seals.
                   The information in Attachment 5 is analyzed further to determine if any
            trends could be discerned which might bear an impact on emissions.
            However, it is apparent that no single parameter consistently caused the
            events. A variety of failed equipment or unforeseen circumstances are
            identified as the cause of the AWFCOs. Furthermore, no data are available
            that could lead to conclusions about the type of wastes present in the kiln
            during each of the AWFCOs.  Of the 108 reported positive pressure events, 56
            are reported by the facility to have been the result of feed or waste flow
            problems.  This converts to  approximately 74 per year.
                   To verify the estimated annual AWFCO frequency of 528, or 44 per
            month, these estimates are compared with a report prepared by a joint
            EPA/OSHA Task Force established to evaluate worker health and safety
            compliance at 29 hazardous  waste incinerators nationwide.  The report
            provides the number of AWFCOs for a 30 day period for each of the 29
            inspected facilities, ranging  from no AWFCOs to > 13,000 for a facility with
            4 incinerators. Of the 29 units, 5 had no AWFCOs during the 30 day period
            and information was not available for 6. Of the remaining  18 units, 13 had
            greater than 44 AWFCOs per month.  Even though the value of > 13,000
            AWFCOs per month seems  anomalously high, the broad range of this study
            without consideration of the potential anomaly indicates that the use of facility-
            specific data provides the most accurate AWFCO estimate.  Therefore,  data
            from other operating facilities  is not incorporated in the effort to estimate the
            WTI AWFCO emissions.

            b.    Effect of AWFCOs  on Emissions
                   For purposes of this  evaluation, AWFCO events are broken into two
            major categories; AWFCOs associated with kiln overpressures, and all  other
            types of AWFCOs.  Each category of AWFCOs are analyzed to determine
            their potential impact to emissions. AWFCOs  associated with kiln
            overpressures have been treated separately because these events can in some
            cases cause an uncontrolled "fugitive" release via puffing at the rotary kiln
             seals or other points.
Volume HI                                                            External Review Draft
Appendix HI-1                              V-4                        Do Not Cite Or Quote

-------
                               The effect of other types of AWFCOs on emissions is less certain.
                         Most other types of AWFCOs  tend to have more potential impact on stack
                         emissions (such as the potential for a low temperature AWFCO to result in
                         higher organic emissions in the stack or the potential for a high temperature
                         AWFCO to result in higher metals emissions in the stack) as opposed to the
                         fugitive emission concerns associated with kiln overpressure events.  Fugitive
                         emissions from kiln seals, since they are uncontrolled and are not moderated
                         by the air pollution control system, are seen as having a greater potential to
                         result hi significant emissions.  The results of this initial analysis are presented
                         below.
                               (1)    Non-overpressure AWFCOs
                                      In an attempt to identify and quantify the effect of AWFCOs on
                               the incinerator stack emissions, data generated during the Condition 2
                               retest performed at WTI are evaluated.  Combustion experts with U.S.
                               FJPA's Air and Energy Engineering Research Laboratory, the contractor
                               for U.S. EPA's Incineration Research Facility at Jefferson, Arkansas,
                               and  internal combustion experts at Midwest Research Institute (a
                               consultant to U.S. EPA on this project) were also contacted during  the
                               preliminary research phase. Unfortunately, no published data,
v-                             estimation techniques or studies were identified that directly address
                               nonroutine emissions from AWFCOs.  One experimental study
                               designed to simulate and quantify emissions from an incinerator
                               operating under upset conditions was identified.  This  study was
                               performed at the IRF; however, the results were not conclusive
                               (Whitworth and Waterland  1992).
                                      In the absence of direct data establishing a link between
                               AWFCOs and stack emissions, several other sources of potentially
                               useful test results are identified.  A primary  source of information
                               which provides actual data on the WTI incinerator performance during
                               an non-overpressure AWFCO is the Condition 2 retest performed at
                               WTI in February 1994. According to U.S. EPA Region 5 personnel,
                               an AWFCO occurred approximately ten minutes before the end of Run
                               1 of the test, due to exceedances of the CO and minimum  temperature
                               setpoints caused  by  plugging hi the high BTU waste feed lance.  The
                               plugging caused  observable flow disturbances and increasing CO

(            Volume HI                                                             External Review Draft
V          Appendix ffl-1                              V-5                         Do Not Cite Or  Quote

-------
                    readings (i.e., process upset) for the final hour of the four-hour
                    sampling event.  This AWFCO is included in the tabulation of total
                    AWFCOs in Table V-l. Runs 2 through 4 occurred without mishap.
                    The test results for Run 1 are therefore compared with the other runs to
                    determine whether the AWFCO which occurred in Run 1  caused an
                    identifiable increase in emissions. Table V-2 provides a comparison of
                    destruction and removal efficiency (DRE) achieved during the four
                    runs.  The DRE standard was met for all four runs, and no significant
                    differences in DRE were evident between Run 1 and the rest of the
                    runs.  In addition, dioxin emissions were not significantly higher in
                    Run 1 than in Runs 2-4 (WTI1993).  Note, however, that testing
                    stopped after the cutoff occurred and test results reflect an averaging
                    throughout the sampling period.  Nevertheless, no apparent difference
                    is noted.  The highest HC value recorded during the test is 0.10 ppm
                    and the HC monitor continued to operate during the AWFCO.
                    Insufficient data on PICS are available from this test to draw any
                    conclusions about specific compounds.  All other standards tested for
                    were achieved.  Based on the above information, there is no evidence
                    that the AWFCO that occurred in Run 1 affected the performance of
                    the incinerator in achieving the emission standards.
                          One other test study with potentially useful results  was
                    identified.  This study was conducted at a rotary kiln incinerator (U.S.
                    EPA 1987). It's purpose was to qualitatively and quantitatively study
                    the characteristics of incinerator effluents under both steady state and
                    transient upset operating conditions.  In this study, transient CO spikes
                    exceeding 700 ppm were induced by tripling the organic waste feed rate
                    for a period of 7 seconds every 30 minutes.  Emissions were measured
                    while feeding waste containing 10% carbon tetrachloride,  both for
                    steady state operation and for operation with the transient  CO  spikes.
                    A large number of volatile and semivolatile compounds were sampled
                    and analyzed. Increases in emissions were measured for methane,
                    methylene chloride (by a factor of approximately 47) and  benzene (by a
                    factor of approximately 15). However, the final concentrations during
                    the transient conditions were still within the range normally seen from
                    hazardous waste incinerators. No semivolatile compounds showed

Volume HI                                                            External Reyiew Draft
Appendix HI-1                              V-6                        Do Not Cite Or Quote

-------
                    significant increases in emissions. The conclusion drawn from this
                    study is that no increase in emission concentrations were observed for
                    most compounds as a result of operating the incinerator under transient
                    upset conditions.
                           It is important to note that this study did not include AWFCOs,
                    the use of which might be more conservative since the AWFCO limits
                    the upset condition by stopping waste feeds.  When coupled with the
                    results of the February  1994 WTI test, it is apparent that the effect of
                    nonover-pressure AWFCOs on incinerator emissions at WTI has not
                    been quantifiable in the past.  Although there are not a lot of data
                    available, the effect does not appear to be significant and results in
                    emission concentrations which are within "normal" range of incinerator
                    emissions.

                    (2)    Overpressure AWFCOs
                           AWFCOs associated with kiln overpressures are of particular
                    importance due to the possibility of escape of unburned or partially
                    burned organics from the kiln seals.   As previously noted, 108
                    AWFCO events hi the 9 month reporting period were associated with
                    kiln overpressures. However, 44 of the 108 events were associated
                    with chunks of solidified ash, or clinker, falling into the slag quench
                    tank, which is  located directly beneath the secondary combustion
                    chamber.   This results hi a sudden release of steam backing into the
                    secondary combustion chamber, causing an overpressure event.  An
                    overpressure event of this nature is less likely to cause a release of
                    waste constituents than  an event associated with a feed or waste flow
                    anomaly,  since a steam related overpressure is most likely to release
                    steam to the atmosphere. Therefore, 64 (or 85, on an annualized basis)
                    of the 108 overpressure events have  the potential to release unburned or
                    partially burned organics from the kiln seals.
                           Little information is available on the duration of these events, or
                    if these overpressure events were associated with an observable release
                    from the kiln seals. However, two events  are reported to  have been
                    observed as visible releases from the kiln seals (Victorine  1994).
                    These two events are estimated to have resulted in emissions from the

Volume III                                                              External Review Draft
Appendix ni-1                              ,  V-7                         Do Not Cite Or Quote

-------
                    kiln seals of three second and three minute durations, respectively.
                    These durations are observer estimates, since no mechanism is hi place
                    to record the duration or magnitude of an event.
                          Although insufficient information exists to determine accurately
                    an average duration for a puff event,  certain general assumptions can
                    be made.  A major cause of kiln overpressure is the intermittent
                    charging of containerized high BTU solid and liquid wastes hi a batch
                    mode.  Once charged, the containers  rapidly release volatile
                    components into the gas phase.  This rapid volatilization and
                    subsequent combustion can cause temperature and pressure excursions
                    hi the kiln, resulting hi kiln overpressure and  "puffing."  Assuming a
                    sufficient excursion to exceed permit  limitations for temperature, CO,
                    or particularly kiln pressure, an AWFCO will occur, stopping all
                    hazardous waste feeds to the kiln. Assuming that the excursion was
                    also severe enough to cause external puffing to occur at the seals, this
                    puffing will occur only for that period of time that  the kiln remains
                    under positive pressure. Once the system with the  organic vapor
                    released by the existing charge hi the kiln, the kiln will return to
                    negative pressure and the release will stop.  Unless a significant failure
                    such as loss of electrical power to the system has occurred, this
                    equilibration process takes only a few seconds.  Therefore,  the
                    assumption can be made that hi many cases puffing releases will be
                    limited to a few seconds hi duration.
                          With respect to the cause of the AWFCO which resulted in the
                    reported three  minute release, the quench system pump failed, causing
                    the backup quench pump to activate.  However, a problem  with the
                    backup quench pump check valve caused the computer to interpret a
                    quench system failure and shut down the induced draft (ID) fan. This
                    failsafe is designed to protect the downstream APC equipment from
                    being exposed to hot combustion gases.  The operators were unable to
                    override the computer system and restart the ID fan.  Since the ID fan
                    could not be quickly restarted, the event was, hi essence, a significant
                    failure as previously described.  This event reportedly caused WTI to
                    initiate changes to prevent the occurrence of this type of software
                    problem hi the future.

Volume III                                                             External Review Draft
Appendix III-l                               V*-8                       Do Not Cite Or Quote

-------
                                      No estimation techniques or empirical studies quantifying
                               emissions from overpressure releases have been identified in the
                               scientific research.  However, a series of research studies have been
                               performed by the U.S. EPA Air and Energy Research Laboratory on
                               various aspects of transient emissions from rotary kiln incinerators
                               using a rotary kiln incinerator simulator.  This 250,000 BTU/hr pilot-
                               scale device simulated the operational features of full sized units hi
                               terms of volumetric heat release, gas-phase residence tune and
                               temperature profile.  This series of studies was designed to evaluate
                               incinerator performance under upset conditions.  Specific research
                               included the effects of parameters such as type and quantity of wastes
                               fed, kiln temperature, and kiln rotational speed on the occurrence of
                               transient puffs within the rotary kiln  (Linak, et al. 1988,  Wendt, et al.
                               1988, and Lemieux et al. 1990).  The term "puff1 in this application is
                               used to denote the rapid  release of volatile waste components into the
                               gas phase,,and does not necessarily mean that the gas phase
                               components are released to the environment.  Additional  research
                               evaluated the means of minimizing transient puffs by controlling waste
                               container  packaging and  operating a rotary kiln at low rotational speed
^-                             and low temperature (Lemieux et al.  1992).
                                      Building on these previous research studies, the most recent
                               research study (Lemieux et al. 1993) addressed emergency stack vent
                               (ESV) issues related to rotary kiln incinerator using the rotary kiln
                               simulator. This research examined optimum  settings of kiln operating
                               parameters to minimize PIC emissions during a simulated ESV  opening
                               event.   A series of 12 tests were performed, varying kiln air flow rates
                               and rotational speeds.  A surrogate perfonnance  indicator was
                               developed to evaluate performance of the simulator based on CO, THC
                               and soot generation rates weighted as a factor of stoichiometric oxygen
                               requirements. This surrogate indicator represented a measure of the
                               relative degree of poor combustion occurring hi the unit, and alleviated
                               the need for costly, time consuming  sampling and analyzing of
                               individual compounds.  However, a limited number of compounds were
                               tested using tedlar gas bag  samples.  Gas bag sampling was performed
                               for one run which was a baseline "puff" test  and one run performed

(            Volume III                                                             External Review Draft
V          Appendix III-l                              V-9                        Do Not Cite Or Quote

-------
                    under the simulated ESV opening conditions.  The concentrations of
                    VOCs from these two tests were of the same order of magnitude,
                    leading to the conclusion that the concentration of organic PICs emitted
                    during an ESV opening appear comparable to those emitted from the
                    rotary kiln prior to secondary combustion during normal batch
                    operation.
                          Concentrations of the 19 organic compounds analyzed during the
                    two runs described above ranged from non-detectable to a high of 1100
                    - 1900 ng/L for methylene chloride.  Toluene was present in
                    concentrations from 100 - 830 ng/L, and most other compounds  ranged
                    from 10 -100 ng/L, including benzene at 29 ng/L.  While it is
                    tempting to use these analytical results to attempt to draw conclusions
                    about possible concentrations of compounds emitted during an
                    overpressure release at an incinerator such as WTI, the authors of the
                    study specifically cautioned against doing  so, by stating the following:

                                 Measurements made on the rotary kiln
                                 incinerator simulator are  not intended to be
                                 directly extrapolated to full-scale units. It
                                 is, for example, very difficult to scale up
                                 some of the important gas-phase mixing
                                 phenomena from the simulator, where, for
                                 instance, stratification is known to be
                                 significant  (Cundy et al.  1989). The
                                 purpose of the simulator  is to individually
                                 examine the fundamental phenomena that
                                 occur hi full-scale units, and to gain an
                                 understanding of the qualitative trends that
                                 would be found hi a full-scale rotary kiln.
                                 In no way should it be inferred that the
                                 concentrations of pollutants from this
                                 apparatus would be the same as those from
                                 full-scale units (Lemieux, et al. 1993).

                          Given the possible similarity in emissions from puffs and ESV
                    openings, it is pertinent to examine the results of an analysis performed
                    on ESV openings  (Kroll et al. 1992).  An unsteady-state computer
                    model was developed to estimate combustion gas flow, temperature and
Volume HI                                                             External Review Draft
Appendix III-l                              V-10                       Do Not Cite Or Quote

-------
                    hydrocarbon concentration versus time for ESV events. The modeled
                    incinerator was similar to the WTI unit hi size and design, and was
                    operated under similar conditions with respect to waste feed rate,
                    temperatures and gas flow rates. The results of this hypothetical
                    simulation indicated that assuming an initial hydrocarbon concentration
                    represented by benzene (from waste) in the 700 ppm range in the
                    system, the concentration of benzene at the ESV would drop to  < 100
                    ppm in under 2 minutes, and to near zero in 5 minutes.  Note that these
                    values are the result of a computer simulation, and a number of
                    assumptions were made relative to the incineration system, ambient
                    conditions, combustion chamber conditions and hydrocarbon vapor
                    concentration.   However, this analysis illustrates the expected rapid
                    decrease in concentration of volatile constituents over the duration of a
                    puff episode.
                           A simpler, hypothetical analysis is presented to illustrate the
                    potential order of magnitude of emissions which could be expected as a
                    result of a kiln overpressure episode. To conduct the hypothetical
                    analysis, a waste constituent is first selected for the analysis.  The
                    constituent must be a common component of the waste and also volatile
                    with known health effects to be illustrative. Based on waste profile
                    data, benzene is selected for this analysis. Based on information on
                    individual waste streams provided in the WTI waste profile data base,
                    the maximum concentration of benzene which could be expected to be
                    present hi any waste stream, either liquid or solid, is 10%.  Only one
                    of the 74 waste streams in the data base has a higher potential
                    concentration of benzene, at 0% to 25%, and this waste stream
                    comprises less than 2%  of the total annual projected volume fed to the
                    incinerator.  Conversely, a number of waste streams list a potential
                    maximum benzene content of 9% or 10%, so 10% is selected for this
                    hypothetical illustration.
                           To further the  illustration, it is assumed that the incinerator
                    waste feed is a combined maximum of 8000 Ib/hr liquids and solids,
                    and that a drum weighing 500 Ibs is charged, causing a pressure
                    excursion and AWFCO, and subsequently, an overpressure release.
                    From this scenario, several assumptions are made about the

Volume in                                                              External Review Draft
Appendix m-1                              V-ll                        Do Not Cite Or Quote

-------
                    concentration of benzene in the kiln and the relative contributions from
                    the various waste streams.  Under normal operating conditions in a
                    rotary kiln incinerator, pumpable wastes are atomized and rapidly
                    dispersed in the gas stream.  During an AWFCO event, all feeds are
                    immediately cut off and the liquid waste present in the kiln prior to cut-
                    off is destroyed in a matter of seconds. Bulk solid wastes have a much
                    longer residence time in the kiln after the AWFCO occurs.  The
                    resulting difference hi residence time is illustrated below by calculating
                    relative concentrations at the time the drum is charged.  The relative
                    contribution of benzene from each stream is as follows:

                          Liquids and bulk solids:
                                 (8000 - 500)lb/hr * (0.1) 73600 sec/hr =  0.21 Ib/sec;
                                 and

                          Drum:
                                 500 Ib * (0.1) = 50 Ib, instantaneously charged.

                          Compared to the instantaneous 50 Ib contribution from the
                    drum, the 0.21 Ib/sec contribution from the liquids and bulk solids
                    (over several seconds maximum time) is not significant.  At the point
                    of AWFCO initiation, the liquid and bulk solids feeds will cease,
                    leaving the drum as the remaining contributor of benzene to the kiln.
                          Next, it is assumed that the benzene volatilizes and is perfectly
                    mixed within the kiln chamber, and that 50% is destroyed hi the kiln.
                    This value, which can not be accurately measured, is selected as a
                    conservative estimate since engineering judgment would indicate actual
                    destruction efficiency would be expected to be much higher. To
                    estimate the potential amount of benzene which could be released hi
                    this hypothetical case, calculations are based on the assumption that the
                    vapor generated by the drum release is equal to  the amount  of gas
                    released from the kiln through the seals. Assuming a kiln operating
                    temperature of 1800°F, the benzene volume hi the incinerator is
                    calculated as follows:
Volume III                                                            External Review Draft
Appendix UI-1                             V-12                       Do Not Cite Or Quote

-------
                  50 » • "-0-50) .  37V?lmole .  <1800°F * Wf>  - 528*' fam vapor
                     78 Ib/mole                          520°F
                                      Based on an internal volume of 7,600 ft3 for the incinerator, this
                                volume of benzene would represent, and therefore displace 7% of the
                                kiln's volume.   If 7% of the kiln's contents escapes as part of the
                                overpressure release, the following amount of benzene will be released:

                                             (50 lb)(0.50)(0.07)  = 1.75  Ib benzene emitted

                                      At an estimated 85 AWFCOs per year potentially causing
                                overpressure releases, a maximum of 149  Ib of benzene would be
                                released.  In a real world setting, this release would be composed of a
                                number of different unburned and partially burned organics.   Benzene
                                is selected to represent the range of compounds emitted.
                                      This hypothetical illustration is not intended to present a means
                                of calculating emissions from this type of  release, but to provide a
                                sense of the  order of magnitude of emissions that might occur as a
                                result of an overpressure release.  Most of the assumptions hi this
                                illustration are made to create a sufficiently simple system to allow the
                                emission calculation to be performed.  Estimating emissions  for the
                                range of compounds that could be expected to be present is beyond the
                                scope of this illustration.

                  4.     Uncertainty
                         The uncertainty inherent hi estimating releases associated with AWFCOs is too
                  great to allow the estimation  of emissions from these events.  These uncertainties
                  include the following:

                  •      Lack of any empirical or theoretical data showing whether or not a  release
                         occurs as a result of a non-overpressure AWFCO;
{           Volume HI                                                             External Review Draft
V          Appendix ffl-1                              V-13                       Do Not Cite Or Quote

-------
       •      Lack of empirical data or a suitable estimation method to quantify releases
              from overpressure AWFCOs.  Specific uncertainties in this case include type
              and degree of destruction of wastes present in the kiln at the moment the
              AWFCO occurs, and duration (quantity) of release (vs. duration of AWFCO
              event);

       •      Calculations performed in this section illustrate that any such potential
              emissions appear to be very small when compared to overall potential fugitive
              emissions from this facility. Because of the small fraction of overall risk
              attributed to fugitives, and because of the small fraction of fugitives attributed
              to kiln seals, the impact of potential kiln seal emissions is deemed negligible;
              and

       •      lack of information necessary to determine the exact level of positive kiln
              pressure necessary to overcome the seals and cause a significant leak, or the
              percentage of overpressure events which actually result in a leak.

       A summary of the key assumptions used in this  analysis are listed in Table V-8.

       5.      Conclusions
              Based on the analysis described above, the following conclusions can be
       reached.

       •      The projected annual frequency for AWFCOs at WTI  is 528, with a projection
              of 85 of those events involving overpressure releases with the potential to
              release unburned or partially burned waste constituents from the kiln seals.

       •      Increased emissions associated with AWFCOs not related to kiln overpressures
              are probably not significant.

       •      AWFCOs associated with kiln overpressures do have the potential for
              emissions via overpressure releases.  These releases may be similar to organic
              emissions from an ESV opening. No direct method of estimating emissions
              from these events is identified, although  the magnitude of these events  is
              expected to be relatively small when compared to the  total fugitive emissions

Volume in                                                              External Review Draft
Appendix HI-1                              V-14                        Do Not Cite Or Quote

-------
              identified in this risk assessment (See Table V-7, page V-39).  This order of
              magnitude is incorporated by the use of conservative assumptions hi estimating
              routine emissions.  Although this release does not occur at stack height,
              resulting in a potentially different dispersion factor, the release would be at a
              far higher temperature (~ 1800°F), resulting  in a much faster plume rise.

B.     Estimation of Fugitive Emissions from Routine Operations

       1.     Introduction
              This section describes the estimation of organic emissions from a variety of
       routine, non-incinerator activities at the WTI site.  The term fugitive is used, because
       emissions from most of these activities are not directly associated with a stack.  The
       activities include routine waste handling hi tanks and containers, and fugitive
       emissions associated with drips and leaks from pumps, flanges,  seals and valves.
       Originally, this section was to cover estimation of emissions from accidental spills,
       however, this portion of the analysis evolved into an Accident Analysis presented hi
       Volume YE of the Risk Assessment.  However, information relating to routine spills
       and storage accidents is retained hi this section for informational purposes.

       2.     Site Description
              Various activities at the facility are examined to identify operations most likely
       to lead to releases of fugitive emissions and to determine a reasonable approach to
       estimating releases due to routine operation.  Physical operations at WTI include
       transport of waste onto the site, transferring, mixing, and blending (WTI 1982).
              Reportedly, WTI normally handles on average,  approximately one dozen
       trucks of incoming waste per day and does not normally receive more of one type of
       truck than another (e.g., tank trucks of bulk liquids versus  semitrailers of drums).
       Therefore, it is assumed that half of the trucks (6) are tankers.  Bulk liquid wastes are
       pumped out the bottom of the truck.  No railcar shipments are currently possible, but
       WTI has reportedly been considering installation of railcar unloading facilities.
              Trucks entering the facility are sent to the truck holding area on the west end
       of the facility so  that truck contents can be identified.  Liquid organic wastes are
       pumped from tanker trucks into a waste reception tank hi the organic waste tank
       farm,  located on  the southeast section of the facility. Sludges are unloaded into
       sludge reception tanks. Liquid materials which arrive hi containers are pumped into

Volume III                                                            External Review Draft
Appendix HI-1                              V-15                        Do Not Cite Or Quote

-------
      the three container pump-out tanks located outside the drum processing building.
      Reception tanks allow for waste decanting and settling of suspended solids.  Lighter
      fractions are pumped to appropriate holding tanks.  Sludges, slurries or highly viscous
      material (which have settled to the bottom of reception tanks) are pumped to sludge
      holding tanks.
             Holding tanks are used for accumulating working inventories of pumpable
      waste to be fed to the incineration system.  Holding tanks are used in the following
      number: two for accumulation of waste sludges, two for accumulation of aqueous
      waste, six for accumulation of all other waste liquids.  Waste sludges can be pumped
      to either the blending tanks or directly to the incineration system.
             Blending tanks are used to combine and mix compatible pumpable wastes into
      a uniform mixture to maximize the consistency and stability of the combustion
      process. Each of the three blending tanks is equipped with an agitator.  In some
      cases, recirculation loops are used in lieu of agitators. The  waste mixture in the
      blending tanks is pumped to lances in the front wall of the incineration system.
             Organic storage tanks described in the part B permit application (WIT,  1982)
      consist of carbon steel, stainless steel,  lined carbon steel or stainless steel, and
      fiberglass reinforced plastic.  Reception tanks have a 7,000 gallon capacity, holding
      tanks have a 20,000 gallon capacity, and blending tanks have either a 10,000 gallon
      or 20,000 gallon capacity.

      3.     Review of Releases at TSDFs
             Historical reports of releases  at WTI and at other TSD facilities provide
      perspective as to the operations at the WTI site and the possibilities for routine release
      of emissions.  A summary of a historical review is presented below.

             a.     Incidents at WTI
                    First, an analysis of all the non-incinerator related incidents reported
             from November 1993 through May 1994 at the WTI facility is reviewed to
             identify operations or to see if the incidents are linked to specific accidents
             (Victorine 1994).  A total of eleven incidents  were reported.  Eight are
             categorized as tank-related incidents, two are categorized as drum sampling or
             repackaging incidents, and one is reported  as a spill.  On-site release volumes
             from drum sampling/repackaging and simple spills ranged from negligible
              fumes to 10 gallons of waste.

Volume III                                                              External Review Draft
Appendix ffl-1                             V-16                        Eto Not Cite Or Quote

-------
 ,^                            Of the eight reported tank-related incidents, three are the result of tank
(^                       overfilling, two are the result of improper maintenance, one is the result of an
                         accident to the tank.  A failed check valve and incompatibility of waste with
                         tank material accounts for the remaining two tank-related incidents.  For
                         incidents where volumes are reported, on-site releases as a result of these
                         incidents range from 8 to 18 gallons.
                                On-site releases as a result of non-tank related incidents which did not
                         involve incinerator operation (i.e., during "routine operations") range from
                         fumes liberated during repackaging to 10 gallons.  Table V-3 provides a
                         breakdown of the tank related incidents reported from November 1993 through
                         May 1994 at the WTI facility.  Non-tank related spills at WTI range from
                         vapors released during repackaging to 10 gallons.
                                Based on the above information, a wide range of spill volumes could be
                         expected as a result of releases that could occur at the WTI facility. Based on
                         the reported spill quantities and the lack of any information showing that any
                         of these spills were highly volatile, it is assumed that minor routine spills are
                         relatively small sources of emissions (i.e., compared to overall fugitive organic
                         emissions)  at this plant and hence are negligible for the purposes of this risk
.                         assessment.  Instead,  spills are evaluated in the Accident Analysis included as
  ^                     Volume YE of the Risk Assessment.

                         b.     Incidents at Other Facilities
                                A look at a profile on reported incidents from Hazardous Waste
                         Treatment, Storage, and Disposal Facilities (TSDFs) (Mangino 1992) shows,
                         in general, that a majority of non-incinerator related incidents are due to tank
                         releases which also account for the largest reported spill volumes.  Of the 15
                         identified non-incinerator related incidents at TSDFs, 12 are tank related.  For
                         incidents where volumes are reported, release volumes range from 40 to
                         20,000 gallons. Of the 12 tank related incidents, 7 are related to tank
                         transfers, 1 occurred  during blending, 2 are attributed to arson, 1 is attributed
                         to tank rupture, and no reason is given for the final reported tank release.  The
                         remaining three non-incinerator incidents are related to repackaging for
                         incineration (2) and equipment failure. Volumes of releases from these
                         incidents ranged from 55 to 75 gallons. The FJ*A study concluded that the
                         major causes of releases from tank systems are unrelated to the characteristics

C           Volume III                                                               External Review Draft
^          Appendix III-l                               V-17                         Do Not Cite Or Quote

-------
              of the material stored in the tanks, assuming that the stored material is
              compatible with the material of construction of the tank system.   Therefore, it
              is likely that hazardous material storage tanks do not have significantly
              different rates of failure (rupture/explosion) regardless of what kind of material
              is stored.
                     A significant cause of releases (not counting operator errors) from
              above ground tank systems is found by U.S. EPA to be from ancillary
              equipment failures (including failures of pumps, flanges, couplings,
              interconnecting hoses, and valves).
                     Since incinerator explosions documented at other facilities in the past
              have tended to be either steam explosions (such as a waste heat boiler
              exploding, or a large slag ball falling into a quench tank) or explosions of
              small quantities of energetic waste, no significant amounts of hazardous
              constituents  were believed to have been released to the atmosphere during
              these events. For this reason incinerator explosions are  not evaluated  in this
              assignment.
                     Reported releases  at TSDFs, hi general, are evaluated hi order to
              determine typical causes associated with releases and typical volumes released.
              These parameters are summarized hi Table V-4.  As  shown hi Table V-4, spill
              volumes as a result of tank failures range from 40 gallons to an estimated
              20,000 gallons at TSDFs. The majority of these incidents involved releases
              between 2,000 gallons and 6,000 gallons with a mean spill volume of
              approximately 5,500 gallons.  The largest spill identified is 638,000 gallons,
              and resulted from a railcar spill (Mangino  1992).

       4.      Identification of WIT Sources of Releases
              Based on the information provided above, and evaluation of the processes at
       the WTI site, the following routine waste handling sources are selected for evaluation
       of fugitive  releases.

              a.      Organic Waste Tank Farm
                     The organic waste tank farm is enclosed hi a building located on the
              southeast section of the facility.  The tank  farm houses the blending, holding,
              and  reception tanks.  All tanks are vented to the incinerator when operating,
              and to  a carbon adsorption bed when the incinerator is not operating.

Volume ffl                                                               External Review  Draft
Appendix m-1                               V-18                        Do Not Cite Or Quote

-------
              b.     Fugitive Emissions
                    For the purposes of this risk assessment, fugitive emissions include
              emissions associated with drips, leaks, and vapor releases from equipment
              transporting the waste. This equipment includes pumps that move the waste
              through pipes between the tanks and to the incinerator, as well as seals and
              flanges in the piping.  Although WIT has a program in place to detect and
              repair leaking equipment of this type, some emissions may still occur.  The
              bulk of this equipment is located within the organic tank farm building,  and
              releases from this building are to the atmosphere through building vents, rather
              man to the incinerator or carbon adsorption bed.

              c.     Container Processing
                    Drums which contain liquids are pumped into the container pumpout
              tanks located outside the drum processing building. The pumpout tanks are
              vented to the incinerator when operating, and to a carbon adsorption bed when
              the incinerator is not operating.

              d.     Open Waste Water Tank
                    This tank is  also referred to as the "C" water tank.  It is open to the
              atmosphere and is used to collect stormwater runoff from areas of the facility
              that are expected to contribute small amounts of contamination to the runoff.
              However, there is some potential that the collected runoff could contain small
              amounts Of contamination.

              e.     Truck Wash
                    The truck wash is a building enclosed on two sides and open to the
              atmosphere on the ends to allow trucks to drive through.  The facility
              management reports that the truck wash has rarely been used.

       5.      Emission Estimation Calculations

              a.     Organic Waste Tank Farm and Pump Out Tanks
                    For purposes of estimating emissions, all tank calculations are grouped
              together.  This includes the blending, holding,  pumpout, and reception tanks.
              Throughput values for emission calculations are based on the waste profile

Volume III                                                             External Review Draft
Appendix m-1                              V-19                        Do Not Cite Or Quote

-------
             data base.  The total base facility throughput (pumpable and non-pumpable) is
             29,880,000 Ib/yr, or 15,000 TPY.  Based on facility estimates, the total
             expected throughput for the facility at full operation is 70,000 TPY.
             Therefore, throughputs used in various calculations have been scaled up by a
             factor of 4.67 (70,000/15,000).
                    Uncontrolled emissions from all these waste tanks (not the waste water
             tank) are calculated using U.S. EPA's Storage Tank Emission Calculation
             Software, Version 2.0, November, 1993. A multicomponent waste stream
             mixture is created based on the top 12 constituents (by weight) found in the
             pumpable waste stream of the MRI waste profile data base.  These 12
             components are prorated to form a simplified base input stream for the Tanks
             program.  However, these compounds are not very toxic or volatile. Hence,
             additional compounds are selected based on toxicity, volatility and quantity
             predicted hi waste stream and added to the program at their un-prorated
             values, to provide the most realistic emission estimates possible for these
             selected compounds.  Because the toxicity of a given compound may be
             different for human receptors than ecological receptors, two iterations of the
             tanks program are performed using two lists of selected compounds. The first
             listed selected compounds relating to human risk and the second listed selected
             compounds relating to the ecological-risk assessment.  These compounds are
             listed in Table V-5 and V-6 respectively. In this way, estimated risks are
             adequately evaluated for each assessment.  Please refer to Volume V (Human
             Health Risk Assessment).
                    Attachment 6 contains an example printout from the tanks program.
             This printout shows all assumptions and default values used hi the program
             itself.

             b.     Fugitive Emissions from Equipment Leaks
                    Fugitive emissions are calculated by summing the emissions of all
             components below.

                    Emission Factors  (Ib/yr)
                    Pump Seal           0.047
                    Valve (In-Line)      0.00051
                    Release             0.23

Volume III                                                             External Review Draft
Appendix III-l                              V-20                        Do Not Cite Or Quote

-------
                    Example calculation for pump seal leakage:

                    n * 0.047 Ib X 24 hr * 250 day
                        hour     day     year

                    Where n = the number of pumps hi the process.

                    Assumptions used to Calculate Fugitive Emissions

                    •     From drawing on page 16-25 of WTI Part B Permit application
                          (WIT1982), there are 37 major pumps.  Assumed eight would
                          be working at any given tune.

                    •     Average emission factors for heavy liquids will apply to the
                          sources.

                    •     Eight lines would be running at any given time with 4 valves
                          per line (32).

                    •     One pressure relief valve would be in operation at any given
                          time.

                    •     Operations of processing 24 hr/day for 250 day/yr.

             c.     Container Processing

                    Assumptions used to Calculate Emissions from Containers

                    •     55 gallon drums will be representative of the majority of
                          packaged  waste received (i.e.,  range from 1 qt to 85 gal sizes).

                    •     Random sampling will be performed on  10% of all drums.

                    •     Repackaging will be done on 35% of drums.
Volume III                                                             External Review Draft
Appendix ffl-1                              V-21                        Do Not Cite Or Quote

-------
                    •      Emissions are equivalent to releases from a leaky valve with
                           heavy liquids.

                    •      Facility receives 6 truckloads per day for 250 days per year.

                    •      Approximately 30 drums per truckload are processed for 8
                           hr/day.

                    •      Average fugitive emission factors for the Synthetic Organic
                           Chemical Manufacturing Industry (SOCMI) are used.  These
                           factors take into account a leak frequency determined from field
                           studies in the SOCMI.  Light liquids have a vapor pressure
                           greater than 0.1 psia @ 100°F.

                    •      Processing averages 15 min (0.25 hr) to repackage drum
                           (Repackage 35%  + Sample 10% =45%)

             d.     Open Waste Water Tank
                    Emissions from the open top stormwater runoff tank are calculated
             using the mass transfer correlations and emissions equations provided hi AP-42
             for waste water treatment systems.  For purposes of these calculations, the
             tank is treated as a non-aerated sump  of the same dimensions. The
             recommended default values for average wind speed and temperature are used,
             and the design average throughput of 15,000 gal/day from the WTI permit
             application is also used. Although the emissions are reported as VOCs, the
             calculations are performed using toluene as the constituent at an average
             concentration of 10 ppm. Toluene is selected because it ranks among the top
             ten constituents by weight in the composite pumpable waste stream from the
             waste profile data base, and it's value for diffusivity is in the midrange of the
             top ten constituents for the pumpable  waste stream.

                    Individual liquid and gas phase mass transfer coefficients
                    k, (m/s)  = 1.0 x 10* + 144 x 1O4 (U*)2-2  (ScJ-0-5; U* < 0.3
                    k, (m/s)  = 1.0 x 10* + 34.1  x 1O4 U* (SqJ0-5; U*  > 0.3
                    For U10 > 3.25 m/s and F/d  < 14
                                                                                           ^
Volume III                                                              External Review Draft
Appendix ffl-1                              V-22                         Do Not Cite Or Quote     v

-------
                           where:
                                 U* (m/s) = (0.01)(U10)(6.1 + 0.63(U10))°-5
                                 ScL  = /tL/(pLDw)
                                 F/D = 2 (A/ir)0-5
                                 K,  = liquid phase mass transfer coefficient
                    kg (m/s) = (4.82 x 10-3)(U10)078 (ScG)^67

                           where:
                                 ScG =
                                 dXm) = 2(A/ir)°-5
                                 kg  = gas phase mass transfer coefficient

                    Overall mass transfer  coefficients

                    K = (k, Keq kg + k,)

                           where:
                                 Keq =  (H/(RT)

                    Air emissions

                    N(g/s) =  K CL A

                           where:
                                 CL(g/m3) = Q Co/(KA + Q)

                    Parameters and Constants

                    -Uj0 = 4.47 m/s                  Wind speed at 10 m above the
                                                      liquid surface

                    -Dw -*• 8.6 xlO"6 cm 2/5            Diffusivity of constitutent hi water
                    (selected toluene as midrange)
Volume HI                                                               External Review Draft
Appendix ffl-1                               V-23                         Do Not Cite Or Quote

-------
                    -/*L = 8.93 x 1Q-3 g/cm           Viscosity of water

                    -PL = 1 g/cm3                    Density of water

                    de =  33 ft = 10.1 m             Effective diameter

                    Ma =  1.81  x 10"4 3/cm-s           Viscosity of air

                    pa =  1.2 x 10"3 3/cm3             Density of air

                    Da = 0.87 cm 3/5 for toluene      Diffusivity of constituent hi air

                    H = 0.00668 atm m3/gmol        Henry's Law Constant of
                                                     constituent

                    R = 8.21 x 10"5 atm nWgmol °F   Universal gas constant

                    T = 298 K (25 °C)                Temperature of water

                    A = 80.1 m2                     Waste water surface area

                    C0 =  10 ppm or 100 ppm         Initial concentration of constituent
                                                     in the liquid phase

                    Q = 15,000 gal/day              Volumetric flow rate

                    Q = 6.57 x 104 m3/s

             e.     Truck Wash
                    Emissions from Truck Washing are calculated using the emission
             factors below:

                    Emission Factors (Ib/hr);
                    Valve (light) = 0.016; and
                    Valve (heavy) = 0.00051.

Volume III                                                             External Review Draft
Appendix IH-1                              V-24                        Do Not Cite Or Quote

-------
                    An example calculation from truck washing is as shown:

                     Ihr X n trucks X 250 day X 0.60 X 0.016 Ib
                    truck      day       year               hour

                    Where n = number of trucks carrying volatile liquids.

                    •    •  Assumptions Used to Calculate Emissions from Truck Washing:
                    •      60% of trucks entering facility contain washable wastes (i.e., all
                           tankers + 10% of all container bearing trucks which may have
                           hauled leaky containers);
                    •      One-third of these trucks contain light liquids;
                    •      One-third of these trucks contain heavy liquids;
                    •      One-third of these trucks contain non-volatile solids with
                           negligible emissions;
                    •      All washing is done under a hooded work station.  Emissions
                           from truck washing  are equivalent to emissions from one leaky
                           valve;
                    •      Clean out time is one hour per truck; and
                    •      Normally handles 12 trucks/day.

       6.     Compound Specific Emissions
             Uncontrolled fugitive vapor emissions from tanks hi the tank farm building are
       calculated using U.S. EPA's Storage Tank Emission Calculation Software (TANKS2),
       Version 2.0, November 1993.  The program uses physical/chemical properties of the
       waste stream constituents, such as molecular weight, vapor pressure, and
       concentration, in deriving emission rates.  Although the actual waste streams stored in
       the tanks at WTI may change over time,  a representative, multicomponent waste
       stream was used hi the TANKS2 program to represent the top  12 constituents1 (by
       weight) projected to be found hi the pumpable waste stream received by WTI based
       on facility waste profile sheets.
    1   The top 12 constituents by weight are estimated to comprise 60 percent of the total
       pumpable waste stream, and include octane (used to represent unidentified
       hydrocarbons), cresol, methanol, methyl ethyl ketone, toluene, cyclohexanone, ethyl
       acetate, 1-butanol, xylenes, methyl isobutyl ketone.

Volume III                                                             External Review Draft
Appendix ffl-1                               V-25                        Do Not Cite Or Quote

-------
            In the HHRA (Volume V) and SERA (Volume VI), additional surrogate
      chemicals are selected to represent the volatility and toxicity of fugitive vapor
      emissions from the WTI facility.  Because the endpoints differ in these two
      assessment (i.e., human and ecological populations, respectively), the surrogate
      chemicals also differ. Table V-5 and V-6 reflect compounds of concern and the
      estimated, compound-specific emission rates for the HHRA and SERA respectively.
            The total estimated fugitive vapor emissions from the tanks in the tank farm
      building were estimated to be 212 pounds per year, as shown hi  Table V-7.  Fugitive
      vapor emissions from the individual types of tanks contained within the tank farm
      building are also indicated hi Table V-7, as well as the total estimated emissions from
      the other sources of vapor emissions  (e.g., open waste water tank, container
      processing, truck wash and fugitive emissions).  Techniques are  not available to
      predict compound specific emissions  rates from the other sources of vapor emissions.
      Therefore, the results of the tank farm/CAB modeling are extrapolated to the other
      three fugitive organic vapor emission sources by assuming that the chemical
      composition of fugitive emissions (expressed as a weight fraction) will be the same
      for all of the identified  fugitive emission sources.  Thus, weight fractions of
      individual constituents (ECOCs) derived from the above analysis of tank farm
      emissions are multiplied individually by the total estimated fugitive emission rates (all
      chemicals) for each of the sources of fugitive organic emissions.  In this fashion,
      compound specific emission rates are estimated for each source of fugitive vapor
      emissions.
            Controlled emissions are estimated based on the following. As stated hi the
      permit application and verified by U.S. EPA, emissions from all tanks and the
      container processing facility are vented to the incinerator when operating, and to a
      carbon adsorption system when the incinerator is not  operating.  Data on the number
      of hours of incinerator  operation are derived from reports from WTI to the Ohio EPA
      regarding waste feed cutoff occurrences. These data  show that the incinerator
      operated approximately 53% of the time over a period from December 1993 through
      July 1994.  Therefore,  it is assumed that the control efficiency for 53% of the
      emissions from the above operations is 99.99%. For the other 47% of emissions
      routed to the carbon adsorption system, a control efficiency of 90% is assumed based
      on average efficiency data contained hi the Evaporative Loss Section of the
      Compilation of Air Pollution Emission Factors, AP-42 (U.S. EPA 199
Volume III                                                            External Review Draft
Appendix m-1                             V-26                       Do Not Cite Or Quote

-------
3a).  Although the figure of 47% is considered to be on the high side, this conservative value
is used in part to account for the observation that at those times when the incinerator is
operating at less than full capacity, a small fraction of the flow from the vapor recovery
system may be directed to the CAB system to ensure sufficient negative pressure in the vapor
recovery system.

       7.     Uncertainty
             Following are the major sources of uncertainty associated with estimation of
       fugitive emissions from routine storage activities.

       •     The amount of wastes handled by the facility on an annual basis (low
             uncertainty factor).

       •     Number and type (bulk vs. containers) of truckloads of waste received by the
             facility (low uncertainty factor).

       •     Because the pumpable waste stream from the waste profile data base is used to
             calculate emissions, all  uncertainties described hi Section n.5 associated with
             the waste profile data base are inherent to these calculations (high uncertainty
             factor).

       •     The tank emissions calculation software is based on empirical equations using
             default values for parameters like seasonal temperature, default windspeed,  and
             meteorological data.  Also,  since the actual waste handling pattern between
             tanks is unknown, it is  assumed that the entire volume of pumpable wastes
             passed through each type of tank once, rather than all tanks sequentially
             (medium uncertainty).

       •     The relative percent of  time the tank and container pumpout emissions are
             vented to the incinerator vs. the carbon adsorption bed could vary  from the
             53% to 47% ratio (low uncertainty).

       •     Uncertainty relative to the waste water tank includes the throughput of
             stormwater, type and concentration of contaminants in the water (both are
             dependant on rainfall and spills or leakage of waste), and uncertainties

Volume in                                                             External Review Draft
Appendix ffl-1                              V-27                        Do Not Cite Or Quote

-------
             associated with the empirical equations used to calculate the emissions (low
             uncertainty).

       •     For container releases and track washing, uncertainty is based on number of
             containers or vehicles processed, type of waste contained, and use of relatively
             non-specific, generic emission factors to calculate emissions (medium
             uncertainty).

       •     For fugitive emissions, uncertainty is associated with accurate counting of the
             number of pumps, flanges, valves and seals based on plan drawings contained
             in the permit application, and the use of relatively non-specific emission
             factors to calculate emissions (medium uncertainty).

             It is believed that there is not any general trend hi these uncertainties which
       would result hi an overall underestimation of fugitive emissions, and conservative
       assumptions are generally made throughout this risk assessment.  These assumptions
       are summarized hi Table V-8.

C.     Emissions from Ash Handling

       1.     Concerns to be Addressed
             There was concern over potential exposure to fugitive emissions of incinerator
       ash.  These fugitive emissions could be caused by routine handling of residues from
       the combustion of hazardous waste at the WTI facility.  This section provides the
       details of the procedures used to estimate the potential magnitude of fugitive ash.

       2.     Approach
             The purpose of this section is to characterize emissions from ash handling
       activities at the WTI facility. The ESP ash  handling activities are of primary
       concern.
             Fugitive emissions from ash handling are estimated by applying  a series of
       emission factors which were developed based on empirical test data generated at a
       pulverized coal-fired power plant employing an ESP for paniculate matter control.
       Factors were available for complex airborne particle size distribution and were based
       on tests of flyash loadout from the ESP (Muleski et al. 1986).

Volume HI                                                             External Review Draft
Appendix ffl-1                              V-28                        Do Not Cite Or Quote

-------
             Although these factors were developed based on flyash from the combustion of
      coal, fugitive emissions are dependent more on the physical form of the flyash than
      on the feed (or waste) stream being combusted.  Specific information on the
      characteristics of the WTI ash will be incorporated into the estimation of emissions
      (See Volume V).  The necessary information includes texture of the ash, moisture
      content, production rate, and a physical description of the ash handling process,
      including the control device(s) producing the ash.

      3.     Estimation of Emissions
             This section provides an estimate of paniculate matter emissions from the slag
      and ash handling activities at the WTI site, and a description of the technique used to
      calculate the paniculate matter emission rate.

             a.     WTI Ash Handling System
                    In order to evaluate WTFs ash handling system, a diagram of the ash
             handling process, based on information provided hi WTI's RCRA Part B
             Permit application, is provided as Figure V-l.  Each step in the ash handling
             process is then identified based on the potential for ash to be emitted to the
             environment.  Those processes which are uncovered, vent to the  environment
             or could potentially leak are identified in the diagram with an "F".
                    In general, two separate groups of ash handling activities  are identified.
             These include bottom ash loadout from the secondary chamber and fly ash
             loadout from the air pollution control equipment.  Stack emissions of
             particulate  matter have been previously characterized, and have not been
             included in this estimation.
                    Each of the groups of ash handling activities is evaluated  in detail to
             determine the potential for fugitive emissions. The ash from the secondary
             chamber falls into a quench tank, where the ash will be thoroughly wetted due
             to the submerged feed chute design within the quench tank.   The ash is then
             transferred via wet conveyor to portable drop boxes which are on trucks.
             After each box is filled it is closed and hauled by a licensed independent
             hauler to a disposal facility.  Since the ash is wet and double sealed in the
             trucks, no  fugitive emissions are  expected to occur either in the form of
             airborne emissions from the ash or fluid leakout which could dry and be
Volume m                                                              External Review Draft
Appendix III-l                               V-29                        Do Not Cite Or Quote

-------
             tracked out by track traffic.  Therefore, emissions from this source are
             estimated to be negligible.
                    Ash handling activities associated with the ESP flyash are a potential
             source of fugitive emissions. These activities include transfer of the ash into
             covered trucks vented to a baghouse. The ash is transferred through a series
             of hoppers, drag chains and a bucket elevator.  Note that the drag chains and
             bucket elevator are not marked as sources of fugitive emissions on Figure V-l.
             Testing and evaluation of fugitive emissions at facilities  that employ these
             types of devices has shown that these devices are not a source of fugitive
             emissions, if properly maintained.  In addition, transfer  of fly ash into covered
             tracks is not considered a source  of fugitive emissions, since dust-laden air
             displaced by the ash in the trucks is vented back into the baghouse.
             Therefore, emissions are estimated only for the bag filter.

             b.     Description of Estimation Technique
                    Particle emissions associated with flyash loading  and unloading are
             estimated using an empirical emission factor based on a  field testing program
             conducted at a coal fired power plant equipped with an ESP (Muleski and
             Pendleton 1986). This study was conducted hi part to characterize the
             emissions of paniculate matter from the loading of flyash into tracks.
             Although the combustion of coal  and hazardous wastes are dissimilar activities,
             flyash generated from similar control devices is expected to behave similarly
             under the same conditions, with respect to fugitive emissions.  In general,
             particle behavior is dependent more on the physical form of the flyash than on
             the feed (or waste)  stream being combusted.
                    The emission factor determined during the empirical study is adjusted
             to reflect the fact that the flyash was wetted prior to loading, with an average
             moisture content of 29%, whereas the WTI flyash is not wetted, and is
             assumed to have a negligible moisture content.  The average factor determined
             during the study is  0.107 Ib/ton flyash, and the recommended adjustment,
             based on scientific judgement, is  a factor of 2 to 10 (10  being conservatively
             high). For information purposes, both factors are used in the following
             calculations to provide a range  of expected emission rates.
Volume HI                                                              External Review Draft
Appendix III-l                              V-30                        Do Not Cite Or Quote

-------
                         c.     Emission Estimate
                                The following assumptions are used in calculating fugitive emission
                         rates:

                         •      The flyash generation rate from the ESP is assumed to be 5,300 T/Y.
                                This value is based on both the original estimation of flyash generation
                                rate provided in the Part B permit application, and the total ash content
                                of the "generic" waste streams created from the waste profile (Chapter
                                II).  Both values are found to be approximately the same. Since a major
                                portion of ash fed to the combustor is converted to bottom ash, it is
                                likely that this value is a conservatively high estimate of the actual
                                flyash generation rate.

                         •      The control efficiency for the fabric filter controlling emissions from
                                the silo is assumed to be 99.5%, based on AP42 factors for typical
                                collection efficiencies of paniculate matter control devices, for the
                                particle size range of 2.5 to 10 um.
/                        The emission calculation then becomes:
V,  ,
                         low range:    (5300 T/Y)(0.107 lb/T)(2) = 1,134 Ib/yr, uncontrolled

                         high range:   (5300 T/Y)(0.107 lb/T)(10) = 5,671  Ib/yr, uncontrolled

                         Taking into account the fabric filter:

                         low range:    (1,187 lb/yr)(l-0.995) =  5.7 Ib/yr

                         high range:   (5,671 lb/yr)(l-0.995) =  28.4 Ib/yr

                         d.     Conclusion
                                 Based on the above calculations,  a conservative estimate of fugitive
                         paniculate matter emissions  from the WTI facility is 28 Ib/yr.
            Volume III                                                              External Review Draft
            Appendix IH-1                               V-31                        Do Not Cite Or Quote

-------
      4.     Uncertainty
             The primary uncertainties associated with this calculation are the flyash
      generation rate and the conversion from wetted flyash to dry flyash. Key assumptions
      used are highlighted in Table V-8.  However, this uncertainty is expected to be low,
      within the stated range of emissions.
Volume m                                                              External Review Draft
Appendix ffl-1                              V-32                        Do Not Cite Or Quote

-------

                                                       Table V-l

                     POSITIVE PRESSURE AND TOTAL AUTOMATED WASTE FEED CUT-OFFS (AWFCOs)
                                             REPORTED BY MONTH AT WTI

Description
Feed/Flow
Positive Pressure
AWFCOs
Other Positive
Pressure AWFCOs
Clinker/Quench
Positive Pressure
AWFCOs
Total Positive
Pressure
AWFCOs
Total AWFCOs
Per Month
Date
11/93
8


3

_


11



59
12/93
1


1

_


2



35
1/94
5


_

2


7



30
2/94
8


_

1


9



26
3/94
3


_

1


4



33
4/94
9


_

10


19



73
5/94
17


3

14


34



85
6/94
5


1

10


16



24
7/94
_


_

6


6



31
Totals
56


8

44


108



396
Source:     Waste Technologies Industries, Report of AWFCO Incidences to Ohio EPA, 1994. All positive pressure AWFCOs identified
           by facility.
Volume III
External Review Draft
r»« M«( «"<:*<> rw /•>•.«*<»

-------
          Table V-2.  DRE RESULTS FROM THE 1994 TRIAL BURN CONDITION NO. 2

CCU
MCB
TCE
Run 1
99.9972%
99.9998%
99.9997%
Run 2
99.9943%
99.9997%
99.9996%
Run 3
99.9964%
99.9997%
99.9995%
Run 4
99.9972%
99.9994%
99.9995%
        Source: 1994 Trial Burn Report
Volume HI
Appendix III-l
V-34
External Review Draft
Do Not Cite or Quote

-------
             Table V-3. SUMMARY PROVIDED BY PERMITTEE OF EQUIPMENT/PROCEDURAL FAILURE (UNRELATED TO INCINERATOR)
                                                WfflCH RESULTED IN SPILLS OR RELEASES
Matrix Released
liquid
liquid
liquid
liquid
liquid
liquid
liquid
liquid
air
air
air
Volume Released
10 gal
not given
not given
not given
not given
8 gal
5 gal
18 gal
not given
not given
not given
Material
hi-BTU waste
aqueous
aqueous
aqueous
quench water
acid/solvent sludge
waste
hazardous waste
unknown
cumene
hydroperoxide
hazardous waste
Cause
unknown
transfer
transfer
tank failure
accident
accident
improper
maintenance
improper
maintenance
misidentified or
unstable waste
misidentified or
unstable waste
misidentified or
unstable waste
Description
not given
Tank overflow during transfer
Tank overflow during transfer
Failed check valve on tank
Slag fell into quench tank, caused it to leak
Rupture disk blown due to overfilling
Disconnected hoses which were under pressure
Failure to reconnect a vent line on a sludge pump after
servicing
Tank lining failure due to incompatibility of tank with waste
Three 5-gallon pails liberate odors during repackaging
Six drums were opened for sampling and started smoking
          Source:  Draft Memorandum from G. Victorine, US EPA Region V to WTI project file. Compilation of Reports on Downwash, Procedural/Equipment
          Failures, and Test Results, May 19,  1994.
Volume III
Appendix III-l
V-35
External Review Draft
Do Not Cite Or Quote

-------
                                                        Table V-4.  SUMMARY OF RELEASES AT TSDFs1
Size
Class
Small



Large














Matrix
Released
Liquid
Liquid
Liquid
Solid
Liq/Air

Liquid
Liquid

Liquid

Liquid

Liquid
Liquid
Liquid
Liquid
Liquid
Liquid
Volume
Released
40 gal
75 gal
55 gal
55 gal*
200 gal

300 gal
9,600 gal

30,000 gal

4,000 gal*

638,000 gal
6,000 gal
20,000 gal
3,000 gal
2,000 gal
2,200 gal
Released
From
Tank
Pipe
Drum
Drum
Tank

Tank
Hose

Hose

Tank Truck

Railcar
Tank
Tank
Tank
Tank
Tank

Material
Unknown
Cool water
Characteristic waste
Organic peroxide
Acid

Unknown
Solvent & Wastewater
Contaminated waters

Acetone

Unknown
Maleic anhydride
Toluene
Solvents
Hydrochloric acid
Ammonia


Cause
2
5
6
6
6

2
2

2

2

2
3
3
5
11
11

Description
Spilled while filling a tank
Recycle cooling water system ruptured
Mixed several incompatible wastes
Explosion and spilled drum during consolidation of materials
Failure to identify corrosives as reactive caused tank boil over
and plume of chlorinated gas
Spilled while filling tank
Hose detached from pump during transfer

Hose broke during transfer and pump continued for a short time
Vapors ignited while the operator was checking tanker level.
Three other tank trucks also caught fire
Spilled during a railcar transfer
Spilled from storage tank— no reason given
Large spill occurred when a tank ruptured
Pipe connecting blend tank to incinerator corroded through
Arson, 2 employees arrested
Arson, 2 employees arrested

Assumed volume.
    Key to Cause:
            1 Unknown
            3 Tank failure

            5 Equipment failure

            7 Inclement weather
2 Transfer

4 Improper maintenance

6 Misidentified/unstable waste

8 Equipment design
            9 Operating conditions    10 Utility failure

            11  Arson

    1 Memorandum from M. Mangino,  USEPA Region V to A. Anderson.  Summary of Incidents at Incinerators and Commercial TSDFs.  Preliminary draft profile on reported
    incidents, March 1982 through October 1992 (Mangino 1992)
Volume HI
A	^ J2_ ₯₯T
                                                                                                         Ex    4 Review Draft

-------
       Table V-5. EMISSIONS FROM ORGANIC WASTE STORAGE TANKS (HUMAN
                     HEALTH - CONSTITUENTS OF CONCERN)
Compound
Octane
Cresol
Methanol
MEK
Toluene
Acetone
Cyclohexanone
Ethyl Aciylate
Butanol
Xylene
MffiK
2-Nitropropane
Acrylonitrile
Carbon Disulfide
Carbon
tetrachloride
Dibromoethane
Dichloroethylene
Formaldehyde
Hydrazine
Pyridine
Total by Tank
Type (Ib/yr)
Total by Tank
Type (g/sec)
Emissions (Ib/yr)
Blending
Tanks
6.13
0.020 .
10.11
8.97
3.04
20.38
0.31
2.73
0.28
0.68
1.04
0.397
0.48
1.44
0.97
0.0475
1.41
11.99
0.0306
0.50
70.94
1.02E-03
Holding
Tanks
6.13
0.020
10.11
8.97
3.04
20.38
0.31
2.73
0.28
0.68
1.04
0.397
0.48
1.44
0.97
0.0475
1.41
11.99
0.0306
0.50
70.94
1.02E-03
Pump Out
Tanks
5.09
0.017
8.40
7.45
2.53
16.93
0.26
2.27
0.23
0.56
0.86
0.330
0.40
1.20
0.80
0.0395
1.17
9.96
0.03
0.42
58.94
8.48E-04
Reception
Tanks
5.34
0.018
8.81
7.82
2.65
17.77
0.27
2.38
0.24
0.59
0.91
0.346
0.42
1.26
0.84
0.0414
1.23
10.45
0.027
0.43
61.84
8.89E-04

Total by
Compound
22.68
0.075
37.42
33.22
11.26
75.46
1.14
10.10
1.025
2.51
3.85
1.47
1.78
5.33
3.58
0.176
5.23
44.38
0.113
1.862
262.7
3.78E-03
Emissions
by
Compound
(g/sec)
3.26E-04
1.08E-06
5.38E-04 .
4.78E-04
1.62E-04
1.09E-03
1.64E-05
1.45E-04
1.47E-05
3.60E-05
5.54E-05
2.11E-05
2.57E-05
7.67E-05
5.15E-05
2.53E-06
7.53E-05
6.38E-04
1.63E-06
2.68E-05


Volume III
Appendix III-l
V-37
External Review Draft
Do Not Cite or Quote

-------
           Table V-6. EMISSIONS FROM ORGANIC WASTE STORAGE (ECOLOGICAL)
Compound
Octane
Cresol
Methanol
MEK
Toluene
Acetone
Cyclohexanone
Ethyl Acrylate
Butanol
Xylene
MffiK
Chloroform
Benzene
Formaldehyde
Dimethylamine
Hydrazine
Acetonirrile
Carbon disulfide
Dimethylhydrazine
Total by Tank Type
db/yr)
Total by Tank Type
(g/sec)

Blending
Tanks
6.17
0.021
10.2
9.04
3.06
20.5
0.31
2.74
0.279
0.37
1.05
1.41
1.31
11.97
5.32
0.031
0.566
1.44
0.404
76.59
1.10E-O3
Emissions (Ib/yr)
Holding
Tanks
6.17
0.021
10.2
9.04
3.06
20.5
0.31
2.74
0.279
0.37
1.05
1.41
1.31
11.97
5.32
0.031
0.566
1.44
0.404
76.59
1.10E-03
Pump
Out
Tanks
5.13
0.017
8.46
7.52
2.55
17.1
0.58
2.28
0.232
0.307
0.869
1.17
1.09
9.96
4.43
0.025
0.471
1.20
0.336
63.70
9.16E-04
Reception
Tanks
5.38
0.018
8.88
7.89
2.67
17.9
0.27
2.39
0.243
0.322
0.911
1.23
1.14
10.45
4.64
0.027
0.494
1.26
0.352
66.82
9.61E-04

Total by
Compound
22.85
0.076
37.7
33.5
11.3
76.0
1.15
10.2
1.03
1.37
3.87
5.22
4.84
44.35
19.71
0.113
2.10
5.33
1.50
283.7
4.08E-03
Emissions
by Compound
(g/sec)
3.29E-04
1.10E-06
5.54E-04
4.82E-04
1.63E-04
1.09E-03
1.65E-05
1.46E-04
1.48E-05
1.96E-05
5.57E-05 ^
7.51E-05 V1
6.96E-05
6.38E-04
2.83E-04
1.63E-06
3.01E-05
7.67E-05
2.15E-05


Volume III
Appendix HI-1
V-38
External Review Draft
Do Not Cite or Quote

-------
        Table V-7. SUMMARY OF ESTIMATED EMISSIONS FROM ROUTINE OPERATIONS
Source
Reception tanks
Pumpout tanks
Holding tanks
Blending tanks
Open wastewater tank
Container processing
Truck wash
Fugitive emissions2
TOTAL
Estimated Emissions
(lo/yr)
50.1'
47.7'
57.21
57.21
202
12.1
9.9
2,126
2,562
Estimated Emissions
(g/sec)
7.2x10-4
6.8x10-4
8.2x10-4
8.2x10-4
2.9x10-3
1.7x10-4
1.4x10-4
3.1x10-2
3.69x10-2
             1       These estimates  are calculated  based on control by venting to the incinerator  or carbon
                    absorption.
             2       These emissions include flanges, seals, pumps, valves, etc.
Volume ffl
Appendix ffl-1
V-39
External Review Draft
Do Not Cite or Quote

-------
TABLE V-8
Key Assumptions for Chapter V
Assumption
Many kiln overpressure events causing AWFCO's result in
puffing releases that are limited to a few seconds in duration
Calculations can be performed to illustrate the potential
order of magnitude of release from kiln seals during
overpressure release
All fugitive chemicals of potential concern have been
identified and included, even though the list is limited to
pumpable wastes (e.g., nonpumpable waste may also be a
source of fugitive emissions)
The composite liquid waste stream list is truncated to
include only the chemicals in the top 90% by mass
All fugitive emission sources have been identified
Six tanker trucks of bulk liquids are received per day
The 12 monthly flyash samples used to determine the
fugitive chemicals of concern and amounts are representative
of actual conditions. Because organic compounds were not
detected in these samples, they are not considered.
Additionally, the chemicals on the analyte list includes all
the chemicals that are likely to be present.
The same chemical composition is used for all fugitive
sources
Tank emissions calculated based on simplified 12 component
waste stream with selected additional toxic/volatile
components
Basis
Container charging of high BTU waste can cause rapid
release of gaseous volatiles, resulting in kiln positive
pressure. Kiln returns to equilibrium and negative pressure
within a few seconds, ending release
Professional judgement based on specific kiln parameters and
waste feed characteristics
Non-pumpable wastes are handled separately from pumpable
wastes and because they are generally not volatile, they are
not likely to result in fugitive emissions
Simplifying assumption to focus assessment
A site inspection was conducted to identify all significant
sources of fugitive emissions.
Best available data
Best available data
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Simplifying assumption to adapt to software constraints
Magnitude
of Effect
low
low
low
low
low
low
low
low
low
Direction of
Effect
underestimate
unknown
underestimate
underestimate
underestimate
unknown
underestimate
unknown
variable*
Volume III
        x III-l
External Review Draft
  Do not cite o^  ^ote

-------
TABLE V-8
Key Assumptions for Chapter V
Assumption
A series of simplifying assumptions were made to allow
calculation of emissions from container processing, truck
wash, and leaks from pumps/flanges/seals
Emission calculations for open wastewater tank were based
on toluene
Controlled emission estimates for tanks are based on venting
to incinerator 53% of time and CAB 47% of time
The contributions to the CAB system are based on the
estimated number of drums received at the facility, the
number of drums sampled and repackaged, and the container
processing rate
Emissions from the CAB system are estimated as equivalent
to releases from a leaky valve with heavy liquids using
appropriate emission factors
The wastewater tank is treated as a non-aerated sump using
a default wind speed and an average throughput. Toluene is
used as a surrogate for VOC behavior.
Emissions from the truck wash are equal to releases of
heavy and light liquids from valves
The emissions factor for flyash from coal burning is applied
to incinerator flyash emissions but increased by a factor of
10 because of negligible moisture content
Basis
Simplifying assumptions based on site information and
available empirical estimation techniques
Simplifying assumption based on empirical estimation
technique
Based on incinerator on-line data for first year of operation
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on a review of information on
facility design and operation, and predicted waste
characteristics.
Professional judgment based on relationship between water
content and credibility.
Magnitude
of Effect
low
low
low
low
low
low
low
low
Direction of
Effect
variable*
variable*
overestimate
unknown
unknown
unknown
unknown
unknown
Volume III
Appendix III-l
V-41
External Review Draft
  Do not cite or quote

-------
                                                                                                       "Negligible"
Waste
                                                                             Atmosphere
                                                                                79 T/Y
                                                                    17.000 T/Y
                                                                   Slag and Ash
                              Figure V-l.  Slag and ash handling diagram— WTI facility.
  Volume III
  Aooendix III-l
V-42
External Review Draft
Do Not Cite or Quote

-------
                                   CHAPTER VI.  REFERENCES
           Barton, R.G., W.D. Clark, and W.R. Seeker, "Fate of metals in waste combustion
           systems,"  Combustion Science and Technology, 74, (1990).

           Biswas, P., W.Y. Lin, and C.Y. Wu, "Fate of metallic constituents during incineration,"
           presented at 1992 Incineration Conference, Albuquerque, NM, (1992).

           Clark, W., R.G. Rizeq, D.W. Hansell, and W.R. Seeker, "Analysis ofBJF compliance test
           metals emissions results," presented at 1994 Incineration Conference, Dallas,  TX, (1994).

           Cundy, V. A., T. W. Lester, A. M.  Sterling, A. N. Montestruc, J. S. Morse, C. B. Leger,
           and S. Acharya,  (1989a), Rotary Kiln Incineration III,  An In Depth Study—Kiln
           Exit/Afterburner/Stack Train and Kiln Exit Pattern Factor Measurements During Liquid CC14
           Processing, J. Air Poll. Cont. Assoc., 39, 944-952.

           Eddings, E.G. and J.S. Lighty, "Fundamental studies of metal behavior during solids
           incineration," Combustion Science and Technology, 85, (1992).

/"---       ENSR Consulting and Engineering.  1993. Final Trial Burn Report for the Rotary Kiln
V         Incinerator, Waste Technologies Industries, East Liverpool, Ohio.  Document number 7136-
           001-800.   May.

           ENSR Consulting and Engineering.  1994a. Waste Technologies Industries, E. Liverpool,
           Ohio; Final trial burn report for condition 2 - February 1994.  Document number 7136-007-
           400. April.

           ENSR Consulting and Engineering.  1994b. Waste Technologies Industries, E. Liverpool,
           Ohio; April 1994 quarterly test emission results for PCDDs/PCDFs and particulate matter.
           Document number 6933-660.  May.

           Entropy Inc.  1994.  Stationary Source Sampling Report for WTI-Von Roll Inc. East
           Liverpool, Ohio. Reference No. 12585.  February

           Federal Register.  1986.  Volume 51, 15422-15430. July 14.

           Friedlander, S.K. 1977.  Smoke, Dust and Haze. John Wiley and Sons,  New York.

           Gelbard, F., "MAEROS," Aerosol Science and Technology, 3, (1980).

           Gordon, S. and B.J. McBride, Computer Program for  Calculation of Complex Chemical
           Equilibrium Compositions and Applications: I. Analysis, NASA Reference Publication 1311,
V.        (1994).

           Volume III                                                          External Review Draft
           Appendix III-l                            VI -1                       Do Not Cite or Quote

-------
Kroll, Peter J., and Robert C. Chang.  Unsteady-State Model of Incinerator ESV Emissions.
For presentation at the 85th Annual Meeting & Exhibition, Air & Waste Management
Association, Kansas City, Missouri, June 21-26,  1992.

Lemieux, Paul M., William P. Linak, Joseph A. McSorley, Jost O. L. Wendt, and James E.
Dunn.  Minimization of Transient Emissions from Rotary Kiln Incinerators.  Combust.  Sci.
and Tech., 1990. Vol. 74, pp. 311-325.

Lemieux, Paul M., William P. Linak, Joseph A. McSorley, and Jost O. L. Wendt.
Transient Suppression Packaging for Reduced Emissions from Rotary Kiln Incinerators.
Combust. Sci. and Tech.,  1992.  Vol. 85, pp. 203-216.

Lemieux, Paul M., William P. Linak, Carin DeBenedictis, Jeffrey V. Ryan, Jost O. L.
Wendt, and James E.  Dunn.  Operating Parameters to Minimize Emissions During Rotary
Kiln Emergency Safety Vent Openings.  Presented at Third International Congress on Toxic
Combustion By-Products, Cambridge, MA, June 14-16,  1993.

Li, K.W.  1974.  Application of Khodorov's and Li's entrainment equations to rotary coke
calciners.  AIChE Journal: 20(5).

Linak, W.P. and J.O.L. Wendt, "Toxic metal emissions from incineration: mechanisms and
control," Progress in Energy and Combustion Science, 19, (1993).

Linak, William P., Joseph A. McSorley, Jost O. L. Wendt, and James E. Dunn. Hazardous
Waste Management—On the Occurrence of Transient Puffs in a Rotary Kiln Incinerator
Simulator. JAPCA, 1988.  Vol. 37, No. 8.

Linak, W.P., and T.W. Peterson, "Effect of coal type and residence time on the submicron
aerosol distribution from pulverized coal combustion," Aerosol Science and Technology, 3,
(1984).

Mangino, M., Summary of Incidents at Incinerators and Commercial TSDFs. Memorandum
re: Preliminary draft profile on reported incidents,  March 1982 through October 1992.

McNallan, M.J., G.J. Yurek, and J.F. Elliot,  "The formation of inorganic particles by
homogeneous nucleation in gases produced by the combustion of coal," Combustion and
Flame, 42, (1981).

Mercer, M. 1994., Personal communication with M. Mangino regarding PIC Target List.

Midwest Research Institute.  1994.  Revised Feedrates Worksheet - WTI pumpable feeds.
August 19.

Muleski, G.E. and F.J. Pendleton, Midwest Research Institute; W.A. Rugenstein.  1986.
Measurement of Fugitive Emissions in a Coal-Fired Power Plant.  Proceedings: Sixty
 Volume III                                                           External Review Draft
 Appendix HI-1                             VI - 2                      Do Not Cite or Quote

-------
Symposium on the Transfer and Utilization of Paniculate Control Technology, Volume 3.
The Detroit Edison Company. November.

Performance Testing Results for the Enhanced Carbon Injection System.  August 30, 1993.

Quann, R.J., and A.F. Sarofinn, "Vaporization of refractory oxides during pulverized coal
combustion," presented at 19th Symposium (International) on Combustion, (1982).

Seeker, W.R., "Waste Combustion" presented at the Twenty-Third Symposium
(International) on Combustion, Paris, France, (1990).

Seigneur, C., A.B. Hudischewskyj, J.H. Seinfeld, K.T. Whitby, E.R. Whitby, J.R. Brock,
and H.M. Barnes, "Simulation of aerosol dynamics: a comparative review of mathematical
models," Aerosol Science and Technology, 5, (1986).

Senior, C.L., and R.C. Flagan, "Ash vaporization and condensation during combustion of a
suspended coal particle," Aerosol Science and Technology, 1, (1982).

Sigg, F., 1994a, Memorandum and attachments to D. Canter, U.S. EPA submitting PIC data
from WTI.  July 1, 1994.

Sigg, F., 1994b, Memorandum and attachments to G. Victorine, USEPA, Submitting back-
up data for PIC monitoring.  July 29, 1994.

Tackie, E.N., A.P. Watkinson,  and J.K. Brimacombe, "Mathematical modeling of the
elutriation of fine materials from rotary kilns," Canadian Journal of Chemical Engineering,
68, (1990).

Uberoi, M., W.A. Punjak, and F. Shadman, "The kinetics and mechanism of alkali removal
from flue gases by solid sorbents," Progress in Energy and Combustion  Science, 16, (1990).

U.S. Department of Commerce (USDOC).  1984. Performance Evaluation of Full-Scale
Hazardous Waste Incinerators.  Volume 2:  Incinerator Performance Results.  National
Technical Information Service Publication No. PB85-129518. November.

U.S. Environmental Protection Agency.  1979.  Guideline Series—Measurement of Volatile
Organic Compounds.  Office of Air Quality Planning and Standards. EPA-450/2-78-041.
OAQPS No. 1.2-115.  September

U.S. Environmental Protection Agency (USEPA).  1987.  Total Mass Emissions from a
Hazardous Waste Incinerator. Risk Reduction Engineering Laboratory.  June.

U.S. Environmental Protection Agency (USEPA).  1987b. Estimating Releases and Waste
Treatment Efficiencies for the Toxic Chemical Release Inventory Form ".  EPA 560/4-88-002.
December.
Volume III                                                          External Review Draft
Appendix ni-1                    '        VI-3                      Do Not Cite or Quote

-------
U.S. Environmental Protection Agency (USEPA).  1988. Measurements of Particulates,
Metals, and Organics at a Hazardous Waste Incinerator.  Office of Solid Waste.  Draft
Report.  November.

U.S. Environmental Protection Agency (USEPA).  1990. Compilation of Air Pollution
Emissions Factors.  Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina.  AP-42.  4th Edition as revised by Supplements A-F.  September.

U.S. Environmental Protection Agency (USEPA).  1991. Evaluation of Compliance with
On-Site Health and Safety Requirements at Hazardous Waste Incinerators.  May 23.

U.S. Environmental Protection Agency (USEPA).  1992. Medical Waste Incineration
Emissions Test Report for Morristown Memorial Hospital, Morristown, NJ. Research
Triangle Park, North Carolina. Volume 1.  Tables 2-50 and 2-51, pp.  2-78 and 2-79.IV.
February 18, 1992.

U.S. Environmental Protection Agency (USEPA).  1993a. Storage Tank Emissions
Calculations Software, Version 2 (TANKS2). Version 2.0. November.

U.S. Environmental Protection Agency (USEPA).  1993b. WTI Phase II Risk Assessment
Project Plan, EPA ID number OHD9806I354L  Region V, Chicago, Illinois. EPA Contract
No. 68-W9-0040, Work Assignment No. R05-06-15.  November.

U.S. Environmental Protection Agency (USEPA).  1993c. Report on the technical workshop
on WU incinerator risk issues.  Risk Assessment Forum, Washington, D.C., EPA/640/R-94-
001. December.

U.S. Environmental Protection Agency (USEPA).  1994a. Implementation guidance for
conducting indirect exposure analysis at RCRA combustion units.  Memorandum from M.
Shapiro, Director, Office of Solid Waste.  Revised April 22.

U.S. National Ocean and Atmospheric Administration (USNOAA).  Climatology in U.S. No.
81.

Victorine, G.  1994. Draft Memorandum to Project File: Compilation of Reports on
Downwash, Procedural,  and Equipment Failures and Test Results. May 19.

Vogg, H., M. Braun, M. Metzger,  and J. Schneider, "The specific role of cadmium and
mercury hi municipal solid waste incineration," Waste Management and Research, 4, (1986).

Von Roll Industries (VRI).  1994.  PIC data from WTI~collected during the Condition 2
rerun of the trial burn.  Memorandum and attachments from F. Sigg, VRI to D. Canter,
USEPA. July 1.
Volume HI                                                          External Review Draft
Appendix III-l                            VI - 4                      Do Not Cite or Quote

-------
Von Roll Industries (VRI).  1994. Back-up data for PIC monitoring conducted during the
condition 2 rerun of the trial burn.  Memorandum and attachments from F. Sigg, VRI to G.
Victorine, USEPA.  July 29.

Wark, K., and C.F. Warner, Air Pollution: Its Origins and Control, Harper and Row, New
York, (1981).

Waste Technologies Industries (WTI).  1982.  Application to the United States Environmental
Protection Agency. September 4; revised November 11.

Waste Technologies Industries (WTI).  1993.  Final trial burn report for the rotary kiln
incinerator.  May.

Waste Technologies Industries (WTI).  1994.  Report of AWFCO Incidences to Ohio EPA.

Waterland, L.D. and DJ. Fournier, Jr., "Potential surrogate metals for incinerator trial
burns," presented at 1993 Incineration Conference, Knoxville, TN, (1993).

Weast, R.C., Handbook of Chemistry and Physics, 54th Edition, CRC Press, Cleveland, OH,
(1974).

Wendt, Jost O., and William P.  Linak.  Mechanisms Governing Transients from the Batch
Incineration of Liquid Wastes in  Rotary Kilns.  Combust. Sci. and Tech., 1988. Vol. 61, pp.
169-185.

Whitworth, W.E. Jr., and L.D.  Waterland.  1992. Evaluation of the Impacts of Incinerator
Waste Feed Cutoffs. Acurex Environmental Corporation, Inc. Research Facility, Jefferson,
Arkansas 72079, RREL, ORD U.S. EPA. August.
Volume III                                                            External Review Draft
Appendix IH-1                             VI - 5                       Do Not Cite or Quote

-------
ATTACHMENT 1

-------
                                                    i.  Waste Profile No.  	
                                                    ii. Cheek here if this is a recertific»cion
                         WASTE  TECHNOLOGIES  INDUSTRIES
                                WASTE PROFILE SHEET
 •ENERATOR INFORMATION

  .  Generator Name:
   Generator Address
   Generator USEPA ID-
   Generator  Contact/Phone:
 BILLING INFORMATION

 .. Billing Name  	

   Billing Address:_
   Customer Service Contact/Phone;
 SHIPPING INFORMATION


 . PACKAGING:  Bulk Solid 	  Bulk Liquid  X   Drum Liquid X   Drum Solid  X
   'Check as many as
     ropriate.)
                Drum Liquid/Solid Mixture  X   Other Container 	


    Type of container per U.S.EPA manifest instructions;CF DF DM DW TC  TP TT	


 = . ANTICIPATED ANNUAL VOLUME (pounds/year) ;   330,000.00


 ': . HMIS  Code:     HEALTH 2*   FLAMMABLE  3   REACTIVE 0   PPE K6


 5. SPECIAL HANDLING INFORMATION


         SARANEX SUIT	


   None                                                                  '
   Is this a dust hazard?  Yes      No  X
SAMPLING INFORMATION

7. A. Was this sampled by the Generator? Yes 	  No  X   If no, answer B and C.

   B. Sample Source:N/A	_____^__

   C. Date Sampled:	
     Sampler's Name/Company:,
Printed:  03/16/94   Appendix HI-1                               External Release Draft
                     Attachment 1                                 Do Not Cite or Quote

-------
                                                        i.  Waste Profile No
                           HASTE  TECHNOLOGIES  INDUSTRIES
                                  HASTE  PROFILE SHEET
PROPERTIES AMD COMPOSITION

3.  Process 'Generating Waste:OPERATION
9. Waste Name:  -
10.  Identify ALL USEPA listed and  characteristic waste  code numbers (D,F,K,P,U) :
     D001,  D002, D004, POPS, D006, D007, D008,  D009,  D010,  DQ11,  D018
     D019,  D021, D022. D023, D024, D025, D026,  D027,  D028,  D029,  D030
     D035,  D036, D038, D039, D040, F001, F002.  F003,  F004,  F005	
     To list additional USEPA waste code numbers,  use additional pages and check  here.	

11.  Physical  State @70F:
     A.  Solid 	 Liquid 	  Sludge 	  Solid and Liquid Mixture 	
     B.  If liquid, are  there multilayers?  Yes X   No 	  N/A 	
     C.  Free Liquid Range 0	 to 20 %
     D.  If there  is a liquid or sludge present, is this waste pumpable?  Yes 	  No  X
        a. If no, can this  waste be heated to improve flow?  Yes 	  No  X
        b. If yes,  will the solids pass through a 1/8-inch screen?  Yes 	  No  	

12.  A.  pH:  s2.0 	  >2.0 to  <12.5 	  a!2.5  	   Not Applicable 	
     B.  Strong Odor  Yes  	   No 	   Describe	
13. A.  Liquid Flash Point: <73°F	  a73°F	  alOO'F	  al40«F	  z200°F	  N/A
    B.  Boiling Point: <100°F 	  alOO°F  X

14. PCBs 	  'If yes,  concentration	ppm,   Pyrophoric 	  Explosive 	-t
    Radioactive 	  Shock. Sensitive 	  Oxidizer 	  Carcinogen  X   Infectious 	
    Asbestos 	  Dioxins 	  Gas 	  Bromoform >500 ppm  X
    Dichlorodifluoromethane >500 ppm  X   Trichlorofluoromethane >500 ppm  X  None 	

15. Benzene   Yes  X   No   '
    If yes,  a) Concentration g100.OOP ppm or rag/1
           b) Does the waste contain water in an amount greater  than or equal to 10%? Yes __  No 	
           c) Is this waste stream subject to the control requirements of 40 CFR 61.340 to 61.358?
             Yes    No
Printed: 03/16/94

                      Appendix m-1                                  External Release Draft
                      Attachment 1                                    Do Not cite or

-------
                                                               i.  Waste  Profile No.
                              HASTE TECHNOLOGIES INDUSTRIES
  f                                  WASTE PROFILE  SHEET
  V .•
 16.  CHEMICAL  COMPOSITION: List ALL constituents (using specific chemical names) present in any
       concentration and forward available analysis. TOTAL COMPOSITION MUST EQUAL OR EXCEED 100%.

       Does your waste contain Antimony, Arsenic.  Barium, Beryllium, Cadmium,  Chromium, Lead, Mercury, Silver
     -  or Thallium?  Yes 	 Ho X

       If yes, include the metals that your waste  contains in the constituent  list below and specify a representative
       concentration range for each metal.   .                                        '•            .
 c
Constituents
PAINT RES IN , P IGMENTS , ADHES IVE
Range
10-100
Units
Percent
S , POLYMERS , INK , OIL
1.1 DICHLOROETHYLENE
1.1.1 TRICHLOROETHANE
1,1,2 TRICHLORO 1,2,2 TRIFL
0-1
0-50
0-50
Percent
Percent
Percent
UOROETHANE
1.1,2 TRICHLOROETHANE
1,2 DICHLOROETHANE
1,4 DICHLOROBENZENE
2,4 DINITROTOLUENE
2 - ETHOXYETHANOL
2-NITROPROPANE
ACETIC ACID
ACETONE
^ARSENIC
BARIUM
BENZENE
BROMOFORM
CADMIUM
CARBON DISULFIDE
CARBON TETRACHLORIDE
CHLORINATED FLUOROCARBONS
CHLOROBEN2ENE
CHLOROFORM
CHROMIUM
CITRIC ACID
CRESOLS
CRESYLIC ACID
CYCLOHEXANONE
DI CHLORODI FLUOROMETHANE
ETHYL ACETATE
ETHYL BENZENE
ETHYL ETHER
FORMIC ACID
0-50
0-1
0-1
0-1
0-50
0-50
5-25
0-50
0-1
0-1
0-9
0-500
0-1
0-50
0-50
0-50
0-50
0-1
0-1
5-25
0-100
0-50
0-50
0-500
0-50
0-50
0-50
5-25
Percent
Percent
Percent
Percent
Percent
Percent
Percent
Percent
Percent
Percent
Percent
ppm
Percent
Percent **
Percent
Percent
Percent
Percent **
Percent
Percent
Percent
Percent
Percent
ppm
Percent
Percent
Percent
Percent
     List  additional constituents  on  following  page,  check here  and  attach.  	

17.   Mark all extremely hazardous  substances  *1% with  **,  or  check the following:   None 	

f  ited:  03/16/94
                        Appendix m-1
                        Attachment 1
External Release Draft
Do Not Cite or Quote

-------
                                                   i.  Waste Profile  No,

                        WASTE TECHNOLOGIES INDUSTRIES
                               WASTE PROFILE SHEET

16. CHEMICAL  COMPOSITION:  List ALL constituent* (using specific chemical names) present in any
     concentration and forward available analysis. TOTAL COMPOSITION MOST EQUAL OR EXCEED 100%.
                    Constituents
    ISOBUTANOL
 Range
0-50
 Units
                                                            Percent
    LEAD
                                                 0-1
           Percent
    MEK
                                                 0-50
           Percent
    MERCURY
                                                 0-259
    METHANOL
0-50
                                                            Percent
    MIBK
                                                 0-50
           Percent
    N-BUTYL ALCOHOL
0-50
Percent
    NITROBENZENE
0-100
Percent
                                                                       **
    0-CRESOL
0-1
                                                            Percent
                                                                       **
    ORTHO DI CHLOROBENZENE
0-50
Percent
    PHOSPHORIC  ACID
5-25
Percent
    PYRIDINE
0-50
                                                            Percent
    SELENIUM
                                                 0-1
           Percent
    SILVER
0-1
                                                            Percent
    SULFURIC ACID
5-25
Percent
                                                                       **
    TETRACHLOROETHYLENE
0-500
PP"1
    TOLUENE
                                                 0-50
           Percent
    TR I CHLOROETHYLENE
0-50
Percent
    TRICHLOROFLUOROMETHANE
0-500
    WATER
0-40
                                                            Percent
    XYLENE
0-50
                                                            Percent
Printed:  03/16/94
                    Appendix IH-I
                    Attachment 1
             External Release Draft
             Do Not Cite or Quote

-------
                                                           i.  Waste Profile  No.
                            WASTE TECHNOLOGIES INDUSTRIES
                                   WASTE  PROFILE SHEET
  8.  Check ONE:  This  Waste  is a:   Wastewater
                                                           Nonwastewater
  .9.  It thi* waste is subject to any California list restrictions, enter the letter from belov(either A or B froo section 20} next
      to each restriction that is applicable:
      HOCs
               PCBs
Acid
Metals
Cyanides
                                                              None  X
  7ASTE IDENTIFICATION TABLE  (not  applicable 	}


  .'0 .  Identify ALL Characteristic and Listed USEPA hazardous waste numbers that apply (as defined by 40 CFR 261} .  For
      each waste number, identify the subcategory (as applicable, check none, or write in description from 40 CFR
      268.42. and 268.43).
t*f.
He.
A. U.S. IPA
JlaxardouB
Haste Code It)
B. Subcategory
Enter the Subcategory De*crlpclon -
It Hoc Applicable.
Sicily Check Hone
	 OeaeriPtien 1 »on« 1
e. Applicable Treatment
Standard*
Performance -Baaed:
OjecJt »* Applicable
Jt«.4i(a) IJt«.41(a)l
Specified Technology:
If Applicable. Enter tbe
40 CR 2CI.43
Table 1 Treatwnt
Code(e)

0. How Miut
the Macte >e
Managed?
later the Letter
1
2
3
4
/•"
"V
6
7
8
9
10
11
D001
D001
D002
D004
, D005
D006
D007
0008
D009
D010
D011
Ignitable Liq. 261. 21 (a) (1) Low TOC
< 10% total organic carbon
Ignitable Liq. 261.21 (a) (1) , High
TOC >or= 10% total org. carbon
Acid, Alkaline, and other based on
261.22





Low Mercury; less than 260 mg/kg
mercury





None
None
None
None
None

None
None
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No INCIN
No INCIN
No INCIN
No
No
No
No
No
No
No
No
A
A
A
A
A
A
A
A
A
A
A
To list additional USEPA waste numbers and categories, use following page and check here;  X

Indicate in Section 20D how the waste must be managed under the land disposal restrictions :

       A.  RESTRICTED WASTE REQUIRES TREATMENT
       B.  RESTRICTED WASTE TREATED TO  PERFORMANCE STANDARDS
       C.  RESTRICTED WASTE SUBJECT TO  A VARIANCE
       D.  NO APPLICABLE MANAGEMENT STANDARDS
       E.  EXEMPT UNDER 40 CFR 268.42

Pr'^ted:  03/16/94
I
                       Appendix HL-1
                       Attachment 1
                                                                     External Release Draft
                                                                     Do Not Cite or Quote

-------
                                                       i. Waste  Profile No.
                          WASTE TECHNOLOGIES INDUSTRIES
                                 WASTE PROFILE SHEET

4ASTE  IDENTIFICATION TABLE (Continuation) (not applicable 	)

20. Identify ALL Characteristic and Listed USEPA hazardous vaste numbers that apply (a* defined by 40 CFR 261). For
    each waste number, identify the aubcategory (as applicable,  check none, or write in description from 40 CFR
    26S.42. and 268.43).
let.
NO.

A. O.I. IPX
tuurdoui
NMt* COdCU)

S. Subc»t*gory
Bntcr the Subc»tegory Description -
If Not Applicable.
Siiply Check None

Standard!
FtrfnrMnrr Beied
Cheek as Applicable
5ii.«iM I14*.4j(al

SpecxZied Technology:
It Applicable, lour the
40 CFI 1(1.43
Table 1 Treatacnt
Coded)
	 SZT7! 	

D. Hov Muat
(he Maete Be
Managed?
Inter the Letter
From Belov

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
D018
D019
D021
D022
D023
D024
D025
D026
D027
D028
D029
D030
D035
D036
D038
None
None
None
None
None
None
None
None
None
None
None
None
None
None
None
No
No
No
No
No
No
No
No
No.
No
No
No
No
No
No
No
NO
No
No
No
NO
No
NO
NO
NO
No
No
NO
NO
NO
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
NONE
Indicate in Section 20D how the waste must be managed under the land disposal restrictions:

      A.  RESTRICTED WASTE REQUIRES  TREATMENT
      B.  RESTRICTED WASTE TREATED TO PERFORMANCE  STANDARDS
      C.  RESTRICTED WASTE SUBJECT TO A VARIANCE
      D.  NO APPLICABLE MANAGEMENT STANDARDS
      E.  EXEMPT UNDER 40 CFR 268.42
Printed:  03/16/94
                       Appendix m-1
                       Attachment 1
External Release Draft
Do Not Cite or Quote

-------
                                                        i.  Waste  Profile  No.
                           WASTE TECHNOLOGIES INDUSTRIES
                                  WASTE PROFILE SHEET
WfcSTE IDENTIFICATION TABLE  (Continuation)(not  applicable 	)

20 .  Identify AU. Characteristic and Lilted USEPA hazardous waste number* that apply (•• defined by 40 CFR 261) . For
     each waste number, identify the subcategory (as applicable,  cheek none, or write in description from 40 CFR
     261.42. and 26B.43).
»«f
No.
A, O.I. »>A
Hazardous
Naate -Code!*)
• . Subeateeery
Inter the Subcategory Description -
It Not Applicable.
Sicily Owek Now
I 	 DMcnption 	 	 	 1 	 HOBS — 1
Standard!
rer
-------
                                                               i.  Waste  Profile No.
                              WASTE TECHNOLOGIES  INDUSTRIES
                                      WASTE  PROFILE SHEET
TRANSPORTATION

21. TRANSPORTATION AND HANDLING  INFORMATION

   •  A.  Is this a DOT Hazardous Material? Yes  X    No 	
     B.  Proper Shipping  Name WASTE FLAMMABLE LIQUID,  NOS
     C.  Hazard Class or  Division Explosive (1.4C)	
         Package Group 	;	;	 I.D.
     D.  Additional Description (MEK.XYLENE)	
     E.  CERCIA Reportable  Quanity(RQ)  and units (Ib,  kg)	1.00 Pounds
     F.  Constituent or Waste Code RQ  is  based on 	
CERTIFICATION

22.  GENERATOR OR DESIGNEE  CERTIFICATION

     1 hereby certify chat all information submitted in item* 1. 4, 7. t.  9, 10, 12*. 13. 14, 15 and 16 contains true and
     accurate descriptions of this waste.  The signature on incoming Land Disposal Restriction Forms at the time of
     waste receipt will certify the information in item 18, 19 and 20. Any sample submitted is representative as defined
     in 40 CFR 261 - Appendix I-or by using an equivalent method. All relevant  information regarding known or
     suspected hazards in the possession of the generator has been disclosed. I authorize WTI to obtain a sample
     from any waste shipment for purposes  of recertification. This information  is the generator's best estimate and
     is not used as a limitation upon HTI's receipt of waste shipments or  quantities in excess of these estimated
     amounts.                                                                                          /
                   Signature                                       Date
            Printed  (or typed)  name                             Company
                                                                      Title
Printed:  03/16/94
                        Appendix HI-1                                      External Release Draft
                        Attachment 1                                        Do Not Cite or Quote

-------
Initial
Generator's  Name:
                                    Waste Analysis Decision

                        Renewal      Date:
                                                                                        i. Haace Profile ,vo
 WASTE PRODUCT  REVIEW INFORMATION


 1 .  is incineration either the required treatment technology or the basis for any treatment standard for the identified
     waste codes is)?  Yes _X_  Ho 	

 2 .  Is this waste profile a lab pack?  Yes  '   No JC_

 3 .  Does WTI intend to normally use this waste as the high BTO feed for the incineration process?  Yes 	 No  x

 4 .  Were supplemental analyses performed? Yes X  Ho 	
     If yes, supplemental analysis results are attached.

 5,

 6.
Are there any MSDS sheets attached? Yes .. No X
APPROVAL SIGMATORXS

Laboratory

Safety

Envi ronmental



Date : / /

Date: / /

Date : / /

Date: / /
Operations
Regulatory Agency Use only; OEPA
	 Accepted 	 Conditional Acceptance 	
Conditional acceptance upon Haste Management Alternatives Plan
for six 16) months from this date. / /
Accepted, Haste Management Alternatives Plan approved.
Conditions for acceptance or reasons for denial :
Acceptance Denied
submission. Decision effective |
{OEPA initials)
(OEPA initials)



Signature . Date / /
Title


 Printed:  03/16/94
                     Appendix m-1
                     Attachment 1
                                                                     External Release Draft
                                                                     Do Not Cite or Quote

-------
                                                              i.  Waste  Profile No.
                              WASTE TECHNOLOGIES  INDUSTRIES
                                      WASTE  PROFILE SHEET
 23 .  WASTE ANALYSIS RESULTS . Is this a miscellaneous special waste?  Yes _  No _
1 	 	 - -
1
Identification
Measured
Value
Possible Haste Codes
total
Constituent
Analysis
TCLP
    ANIONS
    BROMIDE, %
    CHLORIDE, *
    FLUORIDE, *
    IODINE, %
    SULFUR/ *
    SULFIDES, ppm
    CYANIDES (Amenable),
    CYANIDES (Won-Amenable),  ppm
 <0.04
 3.1
 <0.05
 N/A
 .2
 <20
 <20
 B.  METALS
    ANTIMONY, ppm
    ARSENIC, ppm
    BERYLLIUM, ppm
    BARIUM, ppm
    CADMIUM, ppm
    CALCIUM, ppm
    CHSOKIUM, ppm
    COPPER, ppm
    LEAD, ppm
    LITHIUM, ppm
    MERCURY, ppm
    NICKEL, ppm
    PHOSPHORUS,  ppm
    POTASSIUM,  ppm
    SELENIUM, ppm
    SILVER, ppm
    SODIUM, ppm
    THALLIUM, ppm
    ZINC, ppm
<25
<2.5
<0.03
6
<5
2330
170
8.9
740
<5
<0.5
98
<100
<2
<5
31
<35
400
DO04 ARSENIC

D005 BARIUM
D006 CADMIUM

D007 CHROMIUM

D008 LEAD

D009 MERCURY
DO10 SELENIUM
D011 SILVER
 CONCENTRATION
SOLID     LIQUID
Mg/Kg      Mg/1

           5.0
 100

2,000
  20

 100

 100

  4
  20
 100
           100
           1.0

           5.0

           5.0

           0.2
          1.0
          5.0
   PHYSICAL PROPERTIES
   FLASHPOINT *F
   BTU/lb
   WATER, %
   VISCOSITY,  cp
   ASH,  %
   DENSITY, g/ml
   PH
   SETTABLE SOLIDS,  %
   PCB(S),  ppm
   FREE LIQUID
78
15800

<24
15
.984
9.4
28
YES <50
FAIL
D. REACTIVITY
   Nater
   Acid
   Alkali
E.  Dichlorodif luoromethane (%) <0.01   Trichlorof luoromethane (%) <0.01   Bromof orm (%) <0 .01
Printed:  03/16/94  Appendix HI-1
                        Attachment 1
                                External Release Draft
                                 Do Not Cite or Quote
                                                  10

-------
ATTACHMENT 2

-------
                                 Pumpable Feeds, WTI Facility*
4/5/95
c
c
^ Constituent
he*
Cresor(cresyiicacid)
Toluene

MEK*
Methanol
Acetone* (Methyl ketone)
Cydohexanone
Ethyl acrytate
Butanor
Xylene
MIBK
Tetrachlorobenzene*
Nitrobenzene
Ethyl benzene
Pyridtne
2 Ethoxyethanot*
Alcohols
2 Nitropropane
Isobutanol
Dichlorobenzene*
Creyslicaod
heptane*
Benzene
Trichloroethane*
Carbon
Cydohexane
Chior. paraffin oils and waxes
Tetrahydrofuran
Diethyl phthalate
Creosote
1,4Dioxane
Carbon tetrachloride
Formaldehyde
Trichloroethytene
Cumene
Ethanol
Naphthalene
Chloroform*
Tetrachloroethytene
1.1,2 Trichloro 1^2 triflouroethane
Phenol
Dinitrotoluene*
Acetonitrile
Chlorobenzene
Isopropanol*
Methyl methacrylate
TCFM
Formic acid
Acetophenone
Mateic anhydride
Dichlorodifluoromethane
. Furfural
Resorctnol
^ Benzidine

* Bo^AM1 MM UM*+A VWMniM*. *M^ 4lW» UA«r l%f ft
Total Ib/yr
3208730
998281
770291
676259

586938
555858
482451
466761
464645
448321
422393
410043
382090
364159
354015
351715
338208
321555
238633
206838
178823
178323
174406
153251
149376
144739
141435
125396
122429
110180
107045
104285
100677
100350
99450
98523
92408
90589
88399
85377
84824
79191
78284
76207
72266
71012
69874
69352
66350
59443
58810
57915
57438
55116
Appendix UL-1
«.««»„ Attachment 2 1
Percent of Total
16.8
5.2
4.0
3.5

3.1
2.9
2.5
2.4
2.4
2.3
22
2.1
2.0
1.-9
1.8
1.8
1.8
1.7
1.2
1.1
0.9
0.9
0.9
0.8
0.8
0.8
0.7
0.7
0.6
0.6
0.6
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.3
0.3
External Rel
Do Not Citi

-------
                             Pumpable Feeds, WTI Facility*
                                                 4/5/95
Constituent
Calcium eliminate
srylonitriie
Oimethytphenol*
Epiehlorohydrin
N-nitrosodiethanolamine

Dtmetnyipntnaiate
Toluenediamme
• Tetrachloroethane*
Toluene diisocyanate
dicMoroethene*
Chlorinated fluorocarbons (assumed dichlorodifluoroethane)
Carbon disutfide
Phthalic anhydride
Dimethylamtne

i iiuiiuiuucimsiic
Dimethyl phthalate
2 Acetylaminofluorene
1 ,2 Benzenedicarboxytic acid
butyl acetate*
NMntfOsopyrrolidine
2 Napnthytamine
Hydrazine
Crotonaldehyde
Dimethyl sulfate
Dichtoroethane* (Ethylktene dichioride)
1 Naphthylamine
Aniline
osafrote
Oimethyfhydrazine*
dibromoethane*
N-nrtrosodietnylarnine
Benzo(a)pyrene
Chrysene
N-nJtrosodt-n-butytamine
Ruoranthene
Indeno (1.2,3-cd) pyrene
3 Methylcholanthrene
P-benzoquinone
2 Picoline
3,3' Dimethytoenzidrne
-* ~l f f- nib«MH«i«Hlb«BHi«MiMfli /i4il*«««^ tfi n\ i*Htl**>if^nn\
i ,2,9,0 Dioenzaninracene (aioenz (a,c> arnnracene)
P-nitrophenol
1 Methybutadiene
in ii * aft
£. Aceryiaminoiuiorene
Diethyl stitoestrol
Dihydrosafrole
P-dirnethytaminoazobenzene
Paraldehyde
K|_ni^mcfMrwfnA(hvkiff9thflnA
I^I^IIIUUSH^I 1^ • IdJ lyiUI 6U MM 16
N-nitroso-n-methylurea
Safrote
Phenacetin
Oibromomethane* (Methytene bromide)
Nitro-o-toluidirte
Total Ib/yr
54606
54259
53872
52628
51860
51672
51594
50480
50350
49317
49180
45647
44878
44654
44001
41680
40513
40427
39330
38548
38548
38412
37304
37304
36854
36583
36020
35777
34261
33724
33339
33257
33256
32482
32012
32012
32012
32012
32012
32012
32012
32012
32012
31430
31397
4
-------
                             Pumpabte Feeds, WTI Facility*
                  4/5/95
Constituent
4 Bremophenyt phenyl ether
diphenylhydrazme*
Streptozotocin
Reserpine
Amitrole
Cyanogen bromide
Mateic hydrazide
Hydrogen fluoride
Auramine
N,N-diethy1hydrazine (1,2-diethylhydrazine)
3,3' Dimethoxybenzidine
L-serine
Ethytene dibromide
22 Bioxirane
Acetates
Acjylamide
Ethyl methanesulfonate
Methacryionitrite
Malononitrite
Barium salt
Daunomycin
Acrylic acid
Ethylene glycol
Hexachloro 1 ,3 butadiene*
Dichloraphenol*
Acetaldehyde
Porysiloxanes
Hexachloroethane
2 Methoxyethanol
Chlordane
Lead salt
Dibromoethane
Calcium salt
Chlorophenor
Butenal*
Citric acid
Undane
Benzenesutfony! chloride
Dipropytamine
N-butytamrne
2 Heptanone
Amyl acetate
trichtoroethane
Perrtachlorobenzene
dichloropropene*
2,4 D salts and esters
Acetic acid
2 Chloroethyl vinyl ether
1 ,2 Benzanthracene (benzo (a) anthracene)
4,4* Methylenebts(2'>chloro)antltne
Chloromethyl methyl ether
dichlorobutene*
Pentachloroethane
3,3* Dtchlofobenzidtne
Total Ib/yr
30768
30768
30768
30768
30768
30768
30768
30768
30768
30768
30768
30768
30768
30768
30645
30001
30001
28808
28803
27079
26507
26033
24861
24646
23783
23690
20735
20218
19610
19053
18302
18004
17714
15574
15323
15043
14695
14464
13087
13073
13073
13073
12855
12570
11594
10405
10180
9275
9169
8929
8928
8928
8928
8928
Percent of Total

0.2
0.2
0.2
0.2
02
02
02
.02
02
02
02
02
02
02
02
02
02
02
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Based on waste profiles for first year of operations.
                                  Appendix ffl-l
                                  Attachment 2
External Release Draft
 Do Not Cite or Quote

-------
                              Pumpable Feeds, WTI Facility*
                                                  4/5/95
Constituent
b-chlomaphthatene
-chloro-m-cresol
Hexachloropropene
chloral (Trichloreacetaldehyde)
Antimony satt
Arsenic salt
Ethyl methacrytate
Pronamide
Tris (2.3 dibromopropyl) phosphate
1.2 Dibromo-3-diehloropropane
Aeetyl chloride
Benzotrichloride
Chtorambucil
Chtomapharme
_, . . . nu 1 ^ ^ fkv aT af*h ftg ffih ftrilt— 3 ^. . .1. i
Melphalan
O-toluidine hydrochloride
DDT
Benzamine, 4 chtoro 2 methyl-
Dichtoromethoxy ethane
Dimethylcarbamoyl chloride
Hexachlorophene
Thallium chloride
ODD
Dichloroisopropyl ether
Pentachtoronitrobenzene
Kepone
etones*
rtopanot*
Thiourea*
Ethytene Oxide* (Oxirane)
Heptachtor
Tribromomethane
Copper
Mixed organics (ale., amine, etc.)
Ethyl ether (Diethyt ether)
Cydoheptane
octanone*
2 Hexanone
Decanes
3 Pentanone
2 Pentanone
Styrene
2 Octanone
Phosphoric acid
Furfuran (furan)
Anhydride
Di-n-propylnitrosamine
Acetatdehyde* (Ethanal)
Vinyl chloride
Trichlorofluoroethane
Dibenzo(a,i)pyrene
D o-diethul-s-RiethvldithiaBhaSDhiite
tethanethiol (Thiomethanol)
Total Ib/yr
8928
8849
8714
8708
8687
8687
8683
8669
8659
8582
8582
8582
8582
8582
0 8582
8582
8582
8582
8582
8582
8582
8582
8494
8367
8367
8327
7819
7554
7304
7102
6906
6904
6816
6787
6651
6577
6537
6537
6537
6537
6537
6537
6537
6537
6465
6139
6113
5504
5371
5230
4896
4604
4449
4364
Percent of Total
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Based on waste profiles for first year of operations.
Appendix ni-1
Attachment 2
External Release Draft
 Do Not Cite or Quote

-------
                               Pumpabte Feeds, WTI Facility*
                  4/5/95
„ Constituent
Vtethytthiouracil
Saccharin and salts
Selenium sutfide
1,3 Propane suttone
Lasiocarpine
Thiram
Trypan blue
Thioacetamide
Water
TrichtorophenoP
Sutfuric acid
Bis(2)chtoroethyl ether (Dicriloroethyl ether)
Lead subacetate
Zinc
Hexachlorocydopentadiene
Lead acetate
Matonitrite hydroazide*
Chrome compounds
2,3,4,6 Tetrachlorophenol
4,4' Mettiytenebis(2chloro)benzeneamme
Potassium hydroxide
Benzal chloride
Sodium bicarbonate
Chtorobenziiate
Diallate
Lithium amide
™t n «i t -i r li 1 MM«MKAMMl
•"entacnioropnenoi
' Bromofbrm*
Pyrene
1 2 Etnanediylbiscarbamodithioic acid
Potassium-t-butoxkle
Mitomycin C (Azirino)
Methoxychlor
Barium
Lead
Endrin
N-nrtroso-n-ethylurea
Creosol
Phosphorus sulfide
Thioacetamine*
Lead phosphate
Chromium
Thallium carbonate
Thallium acetate
Thallium nitrate
••*• MKt Li • t »,_ -j. MK' «_i • _ 1 - .1 _»
oiSv^ijcnioroisopropyi emei (mcniorotsopropyi cuierj
2,4 DichlorophenoxyacetJc acid (2,4-D)
2,4,5 TP(Silvex)
Sodium sulfide
Toxaphene
Pthulor»> biiudithincarhainie aeki
Silver
Methyl ethyl ketone peroxide
Cadmium
Total Ib/yr
4276
4276
4276
4276
4261
4233
4066
4003
3537
3481
2462
2302
2286
2171
2170
1969
1965
1869
1823
1823
1663
1334
1304
1188
1188
1158
1155
959
909
909
817
767
669
493
472
455
428
393
331
273
260
244
230
230
224
214
127
127
111
110
107
94
86
76
Percent of Total
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
00
w.w
0.0
0.0
0.0
• Based on waste profiles for first year of operations.
                                    Appendix ni-1
                                    Attachment 2
External Release Draft
 Do Not Cite or Quote

-------
                                   Pumpable Feeds, WTI Facility*
                                                   4/5/95
Constituent
Arsenic
Selenium dioxide
PCBs
Cacodyfcacid
Mercury
Selenium
NitrophenoP
Dioctyl pnthatete
Diehloromethane * (Methytene chloride)
Hexachlorobenzene
Dipropyfnitrosarnine
N-prbpylamine (1-aminopropane)
atMWVMifMm
IIWI1IWIWIIII
N-methyt-n'-nitro-n-nrtrosoguanidine
Phenanthrene
Ethyl carbamate (urethane)
Methyl chlorocarbonate
7,12 Dimethyl Denz(a)arrthracene
Azaserine
Sym-trinitfobenzene
Ethylene fhiourea
Methapyritene hydrochloride
2 Cntoronapntnalene
Bis(2 chloroethoxy)methane
Bis(2)chloroethyl ether CDichloroethyl ether)
Dimethytcarbamyl chloride
Benz(c)acridine
£ Propanediol
1,3Pentadiene
1 ,2,7,8 Dibenzopyrene
1 .2.3,4 Diepoxybutane
4 Chtoro-o-toiuidine hydrochloride
Hydrochloric acid
lodomemane* (Methyl iodide)
Total of cmpds remaining after revision
Summary of Values
. l_l i^, Mf]nf4c f^nnlaininn analuffiu
total Ibs pumpabte feed, tess ash & balls
mass%
ash Ibs
total (with ash)
Total Ib/yr
45
24
23
15
10
3
3
3
2
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
19143596
Theoretical
94441 
-------
                             Non-pumpable Feeds, WTI Facility*
                  4/5/95
Constituent
Xytene
Isobutanol
he*
• i- , Ingt „
•nexane
Toluene
MEK
MIBK
Acetone
Ketones*
Pyridine
trichloroetnane
Ethyl acetate
Aniline
cresol*
Ethyl benzene
2 Nitropropane
Benzene
Formaldehyde
Chlorinated fluorocarbons
Tetrahydrofuran
Chlorobenzene
AcetaWehyde
N-nitrosopyrrolidine
Ben2(c)acridine
Cacodylic acid
Mitomycin C (Azirino)
Nitrobenzene
Cydohexane
2 Ethoxyethanol
Acrylic acid
Dimethylhydrazine*
Ethyl metnacrytate
Di-n-prepylnitrosamine
Dipropytamine
Lasiocarpine
Lead phosphate
Lead acetate
Chlorinated polymer resins (PVC)
Butanol
Dichlorobenzene*
Phthalic anhydride
Cresylicacid
AcetonitrHe
Total Ib/yr
688977
522564
348828
329211
323755
262653
228362
221834
191136
125649
106504
100705
99145
82414
66116
65636
65614
62035
57022
53530
53359
51724
51213
51174
49898
49898
49564
47888
47706
47358
45364
44091
44088
44088
44088
44088
43445
41458
33098
31743
29782
29367
28848
Percent of Total
12^
9.3
6.2
5.8
5.7
4.7
4.1
3.9
3.4
2.2
1.9
1.8
1.8
1.5
1.2
1.2
1.2
1.1
1.0
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.9
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.7
0.6
0.6
0.5
0.5
0.5
* Based on waste profiles for first year of operation.
                                     Appendix ni-1
                                     Attachment 2
External Release Draft
 Do Not Cite or Quote

-------
                            Non-pumpable Feeds, WTI Facility*
                                                 4/5/95
Constituent
Chrome compounds
Dimethylamine
** _i i.
wydonexanone
1,4Dioxane"
1 Naphthytamtne
1 2 Benzenedicarboxylic acid
2 Naphthylamine '
Dimethyl sulfate
Sodium hydroxide
Antimony salt
Arsenic salt
Methane!
Carbon disuffide
Acetic acid
Tetrachloroethylene
Trichloroethylene
1,1,2 Trichloro 122 triflouroethane
Formic acid
Carbon tetrachtoride
Acetophenone
Dimethytphthalate
Calcium salt
Leadsatt
Vinyl chloride
Citric acid
Benzal chloride
Methoxychtor
dichtoropropene*
Chlorobenzilate
Diallate
Tetraehloroethane*
Daunomycjn
Chloroform-
4.4' Methylenebis(2-chloro)aniline
Carbamicacid
Trichlorobenzene*
Isosafrole
Toluene diisocyanate
Toluenediamine
dinitrotoluene*
Furfural
Dimethyl phthalate
Total Ib/vr
27658
26512
26421
25502
25488
25488
25488
25488
22298
21564
21564
20534
18709
18662
17608
16928
14391
13188
12794
12338
11062
10980
10290
9926
9919
9311
9311
9168
6669
8669
7213
7086
6865
5012
5011
4648
4612
4545
4545
4305
3816
3269
Percent of Total
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.3
02
02
02
02
02
02
02
02
02
02
02
02
02
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Based on waste profiles for first year of operation.
Appendix ffl-1
Attachment 2
                                                          8
External Release Draft
Do Not Cite or Quote

-------
                            Non-pumpable Feeds, WTI Facility*
                  4/5/95
Constituent
Methacryionitnle
SulfuricackJ
Furfuran (furan)
Resorcinol
Hexachtoro 1,3 butadiene*
ChlorodJbromomethane
Chromic acid
Bfomofoiiu*
Acetates
Potassium-t-butoxide
Acryiamide
Bis(2)chloroethyl ether f Dichloroethyl ether)
Ethyl methanesuKbnate
Ethytene Oxide* (Oxirane)
Mal»ir anhvrlride
N-nitrosodiethanolamine
N-nitrosopiperidine
P-dimethylaminoazobenzene
Thallium nitrate
Thiourea
Tiypanbiue
Diethyl phthalate
Diethyl stilbestrol
Dihydrosafrote
Kepone
dtchloroethane*
Heptachlor and hydroxide
Dichlorornethoxy ethane
Chtordane
•dichloroetnene*
Pentachlorophenol
Pentachtoronitrobenzene
^— *— ^ —*-! — — .1 	 — 	 «•
i euacnioroDenzene
P-chloro-m-cresol
Calcium sulfide
Thiram
Dibromoethane
Methanethiol (Thiometrianol)
Hexachloropropene
Benzenesulfonyl chloride
ODD
ronamioe
Thallium chloride
Total Ib/yr
3269
2894
2552
2540
2505
2138
1961
1818
1403
1353
1276
1276
1276
1276
17TR
l£/w
1276
1276
1276
1276
1091
1082
1047
1024
1024
894
881
858
813
761
723
707
705
EO4
581
500
438
422
403
339
251
251
251
251
251
Percent of Total
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
on
u.u
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
f\ A
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Based on waste profiles for first year of operation.
                                   Appendix ffl-1
                                   Attachment 2
External Release Draft
 Do Not Cite or Quote

-------
                           Non-pumpabte Feeds, WTI Facility*
                                                 4/5/95
Constituent
Thioacetamide
Benzamine. 4 chtoro 2 methyl-
Pentachlorobenzene
DichloropnenoP
4-Chlore-o-toliBdine fBenzamine, 4 chtoro 2 methyl-)
N-nitroso-di-rHethanolamine
N-nitroso-n-ethyiurea

Tns (2.3 atbrornopropyij pnospnate
Toxaphene
Ethyl ether (Diethy) ether)
2 Acetylaminofluorene
DicMorodifluorornethane
5 Nitro-o-toluidine
Bis(2 ethythexyl)phthalate
Safrote
Totals
Total Ib/yr
250
206
206
201
201
144
144
75
20
19
11
4
3
3
3
5637846
Percent of Total
0.0
0.0
0.0
0.0
0.0
0.0 .
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0
Based on waste profiles for first year of operation.
Appendix ffl-l
Attachment 2
                                                          10
External Release Draft
 Do Not Cite or Quote

-------
ATTACHMENT 3

-------
ATTACHMENT 3: PREDICTION OF SPECIATION

Table 3-1  of this attachment summarizes the major species of each metal (those which
account for 1 mole percent or more of the total quantity of the metal present) of interest
predicted to exist at each of the conditions examined. Many assumptions are required to
produce these predictions so great care should be taken hi other use of these data. The
primary assumptions include:

       1.     Tkermodynamic equilibrium is continuously maintained throughout the
       combustion system.
             This implies that  there are no kinetic limitations to the reactions. However, as
       combustion gases cool, it is expected that kinetic limitations increase hi importance
       and at some temperature become dominant. Thus it is unlikely that the species
       predicted to form at low temperatures using thermodynamic modeling would actually
       form in the time in which the metals remain hi the incinerator and the flue gas
       cleaning equipment.

       2.     All elements are intimately mixed and may react with all other elements.
             Some segregation will occur.  For instance, gas phase elements will have
       limited access to the slag. However,  insufficient knowledge is available to refine the
       assumption.

       3.     No condensed phase interactions occur.
             This assumption is a simplification because recent testing indicates that there
       are significant interactions between some metals and the bulk inorganic substrate
       (Barton et al.) "The behavior of metals during the thermal treatment of mixed
       wastes," presented at the 1995 Incineration Conference, Seattle, WA, 1995). In
       slagging systems, non-ideal solutions of metals hi the slag will occur.

       4.     Data for all species which may form are included in the thermodynamic
       databases used.
             The data bases are extensive;  however, exclusion of even one key species can
       have a strong impact on the equilibrium predictions.

       5.      The species CrO2Cl2 will not form.
             The assumption is made to account for some kinetic limitations.
       Thermodynamic calculations predict that large quantities of CrO2Cl2 will form as the
       flue gases cool. However, it is strongly suspected that the reactions converting Cr2O3
       into CrO2Cl2 are slow (little CrO2Cl2 is observed hi the flue gases). To account for
       this observation, CrO2Cl2 is excluded from consideration.
Appendix III-l                                                         External Review Draft
Attachment 3                                    1                      Do Not Cite or Quote

-------
       6.     The waste feed concentrations derived are correct.
              The implications of this assumption are discussed at length in the description
       of modeling uncertainty analysis.

    It should be noted that the predictions pertain only to the specific conditions listed. Local
temperatures and oxygen concentrations can vary significantly in an incinerator.
 Appendix ni-1                                                             External Review Draft
 Attachment 3                                  2                           Do Not Cite or Quote

-------
         Table 3-1.  Speciation Results From Thermodynamic Modeling
T
SRa
a
Element
Al




As


Sb



Ba




Be


Cd






Cr

Cu





Pb







Hg

C

mol/min
^Chernical'-Xf?
.^Formv > • -!-

AI203
AI2Si05
AICI
AlClg

AsaOg
AsO

Sb2Os
Sb02
SbO

BaCI2
Ba(OH)2
BaS
BaO

Be(OH)2
BeO

Cd
CdCI
CdCI2
CdO
CdSO4
CdOSiO2

Cr2O3

CuCI
Cu3CI3
Cu2O
Cu
CuO

PbO
PbCI
PbCIa
PbS
PbO2
Pb304
Pb

HgSO4
1400
1
175
;PSl?l

94
3
1
2

0
100

0
0
100

54
16
0
30

5
95

89
1
7
3
0
0

100

36
0
62
2
0

73
10
1
0
0
0
16

0
1200
1
175
..:-••-•>"•

96
4
0
0

0
100

0
0
100

74
3
0
23

2
98

22
1
68
6
0
3

100

5
0
95
0
0

82
10
2
0
0
0
6

0
1100
1
175
Portion

96
. 4
0
0

0
100

0
0
100

82
2
0
16

0
100

3
0
87
6
0
4

100

1
0
99
0
0

93
4
1
0
0
0
2

0
1000
1
175
1200 1200
0.8
175
of element havinc

95
5
0
0

0
100

0
0
100

87
0
0
13

0
100

0
0
90
6
0
4

100

0
0
100
0
0

100
0
0
0
0
0
0

0

96
4
0
0

0
100

0
0
100

85
2
8
5

0
100

27
1
64
5
0
3

100

0
. 0
0
100
0

0
1
0
1
0
0
98

0
1.2
175
150
1.5
175
200 400
1.5 1.5
175 175
1200
1
0
150
1.5
0
the listed form, (mole percent):

96
4
0
0

0
100

0
0
100

67
4
0
29

0
100

21
1
69
6
0
3

100

8
0
92
0
0

91
5
3
0
0
0
1

0

76
24
0
0

100
0

1
99
0

88
12
0
0

0
100

0
0
0
0
100
0

100

0
0
0
0
100

0
0
0
0
100
0
0

1

79 88
21 12
0 0
0 0

100 100
0 0


100 100
0 0

88 88
12 12
0 0
0 0

0 0
100 100

0 0
0 0
0 0
0 0
100 100
0 0

100 100

0 0
0 0
0 0
0 0
100 100

0 100
0 0
0 0
0 0
0 0
100 0
0 0

2 0

96
4
0
0

0
100


0
100

0
4
0
96

2
98

72
0
0
19
0
9

100

0
0
100
0
0

96
0
0
0
0
0
6

0

76
24
0
0

100
0

1
99
0

0
100
0
0

0
100

0
0
0
0
100
0

100

0
0
0
0
100

0
0
0
0
100
0
0

0
Appendix HI-1
Attachment 3
External Review Draft
Do Not Cite or Quote

-------
              Table 3-1.  Speciation Results From Thermodynamic Modeling
                                     (Continued)



Ni




Se


Ag



Tl



Zn


HgCI2
Hg
HOO

NCI
NiCfe
Ni
Ni3S2

SeO
Se02

Ag
AgCI
Ag2SO4

Tl
TICI
TI2S04

Zn
ZnSO4
0
99
1

1
1
98
0

9
91

1
99
0

0
100
0

100
0
0
99
1

0
1
99
0

2
98

0
100
0

0
100
0

100
0
0
99
1

0
0
100
0

1
99

0
100
0

0
100
0

100
0
0
98
2

0
0
100
0

0
100

0
100
0

0
100
0

100
0
0
99
1

0
0
100
0

2
98

0
100
0

0
100
0

100
0
0
99
1

7
0
93
0

2
98

0
100
0

0
100
0

100
0
99
0
0

0
0
100
0

0
100

0
0
100

0
0
100

92
8
98
0
0

0
0
100
0

0
100

0
0
100

0
0
100

92
8
100
0
0

0
0
100
0

0
.100

0
0
100

0
76
24

92
8
0 100
99 0
1 0

0 0
0 0
100 100
0 0

2 0
98 100

100 0
0 0
0 100

100 0
0 0
0 100

100 92
0 8
SRa
= Air to  waste  stoichiometric  ratio.
      Appendix III-l
      Attachment 3
                                                                External Review Draft
                                                                Do Not Cite or Quote

-------
                             ATTACHMENT 4
WORKING DRAFT-NOT FOR RELEASE                     .            DO NOT CITE OR QUOTE

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
CAS No.
Chemical Name
Emission Rates, g/s
1A 1B 2 3A 3B

4A 4B
Maximum
Emission
Rate, g/s
CARCINOGENS

764-41-0
79-34-5
79-00-5
75-34-3
75-35-4
95-50-1
107-06-2
78-87-5
106-99-0
95-50-1
106-46-7
88-06-2
121-14-2
606-20-2
91-94-1
107-02-8
107-13-1
71-43-2
50-32-8
205-99-2
111-44^*
39638-32-9
117-81-7
75-27-4
75-25-2
85-66-7
56-23-5
57-74-9
67-66-3
74-87-3
PCDD/PCDF TEQ
(trans) 1 ,4-Dichloro-2-butene
1 ,1 ,2,2-Tetrachloroethane
1 , 1 ,2-Trichloroethane
1,1-Dichloroethane [WTI PIC LIST)
1,1-Dichloroethylene (vinylidine Chloride)
1 ,2-Dichlorobenzene
1,2-Dichloroethane
1,2-Dichloropropane (WTI PIC LIST]
1,3-Butadiene
1 ,3-Dichlorobenzene
1 ,4-Dichlorobenzene
2,4,6-Trichlorophenol
2,4-Dinftrotoluene
2,6-Dinitrotoluene
3,3'-Dichlorobenzidine
Acrolein
Acrylonftrile
Benzene
Benzo(a)pyrene
Benzo(b)fluoranthene
Bis(2-chloroethyl)ether [WTI PIC LIST]
Bis(2-chloroisopropyl)ether [WTI PIC LIST]
Bis(2-ethylhexyl)phthalate
Bromodichloromethane
Bromoform
Butylbenzyl phthalate
Carbon tetrachloride
Chlordane
Chloroform
Chloromethane


2.15E-04 1.29E-05
1.29E-05
1.29E-05
1.29E-05
8.90E-04 3.31E-05 .
1.29E-05
1.29E-05

1.29E-05
1.29E-05



3.33E-05

2.02E-04
3.59E-04 7.17E-03 8.95E-04
1.24E-04

1.33E-05
7.98E-07
1.12E-08
1.23E-04
1.04E-05 9.01 E-05

4.37E-04 6.05E-04
7.39E-05
6.10E-04 4.08E-04 3.63E-04
4.22E-05






2.66E-05



6.69E-06
1.02E-05
6.69E-06
6.69E-06
6.69E-06
6.69E-06



6.69E-06
6.69E-06
6.69E-06
6.69E-06
4.52E-05


6.69E-06




2.38E-09
O.OOE+00
2.15E-04
1.29E-05
1.29E-05
1.29E-05
8.90E-04
1.29E-05
1.29E-05
O.OOE+00
1.29E-05
1.29E-05
6.69E-06
6.69E-06
6.69E-06
3.33E-05
O.OOE+00
2.02E-04
7.17E-03
1.24E-04
6.69E-06
1.33E-05
6.69E-06
4.52E-05
1.23E-04
9.01 E-05
6.69E-06
6.05E-04
7.39E-05
6.10E-04
4.22E-05
  Appendix III-l
  Attachment 4
External Release Draft
 Do Not Cite or Quote

-------


CAS No.
218-01-9
124-48-1
106-93-4
75-21-8
86-73-7
50-00-0
76-44-8
118-74-1
87-68-3
77-47-4
67-72-1
78-59-1
75-09-2
86-30-6
1336-36-3
127-18-4
79-01-6
108-05-4
75-01-4
95-53-4
106-47-8
106-49-0
58-89-9

542-75-6
764-41-0
156-80-5
542-75-6
630-20-6
71-55-6
76-13-1


Chemical Name
Chrysene
Dlbromochloromethane [WTI PIC LIST]
Ethylene dibromlde
Ethylene oxide
Fluorene [WTI PIC LIST]
Formaldehyde
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroethane
Isophorone [WTI PIC LIST]
Methylene chloride
N-Ntoosodiphenylamine [WTI PIC LIST]
Polychlorinated biphenyls (209 congeners)
Tetrachloroethylene
Trichloroethylene
Vinyl acetate
Vinyl chloride
o-Toluidine
p-Chloroaniline
p-Toluidine
-Hexachlorocyclohexane (Lindane)
NON-CARCINOGENS
(cis)1 ,3-Dichloropropene
(cis) 1 ,4-Dichloro-2-butene
(trans) 1 ,2-dichloroethylene
(trans) 1 ,3-Dichloropropene
1,1 ,1 ,2-Tetrachloroethane
1,1,1-Trichloroethane
1 ,1 ,2-Trichloro-1 ,2,2-trifluoroethane

Emission Rates, g/s
1A 1B 2 3A 3B
1.24E-04
2.63E-05
1.15E-04
3.05E-05

6.07E-04
2.57E-05
6.23E-09
1.01E-04
8.09E-06
7.54E-05

5.20E-04

1.05E-02 1.76E-04 3.95E-04 .
4.37E-04
6.43E-05
5.65E-05 1.29E-05



5.48E-05

1.29E-05

1.29E-05
1.29E-05

4.45E-05
3.30E-04

Emission
4A 4B Rate, g/s
6.69E-06 1.24E-04
2.63E-05
1.15E-04
3.05E-05
6.69E-06 6.69E-06
6.07E-04
2.57E-05
6.69E-06 6.69E-06
6.69E-06 1.01E-04
6.69E-06 8.09E-06
6.69E-06 7.54E-05
6.69E-06 6.69E-06
5.20E-04
6.69E-06 6.69E-06
O.OOE+00
1.05E-02
4.37E-04
6.43E-05
5.65E-05
O.OOE+00
6.69E-06 6.69E-06
O.OOE+00
5.48E-05

1.29E-05
O.OOE+00
1.29E-05
1.29E-05
O.OOE-I-OO
4.45E-05
3.30E-04
Appendix III-l
           4
External Release Draft
 Do Not Cite or ^ ote

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
CAS No.
96-18-4
95-94-3
120-82-1
96-12-8
99-65-0
123-39-1

58-90-21
95-95-4
94-75-7
120-83-2
105-67-9
51-28-5
78-93-3
532-27-4
95-57-8
75-29-6
591-78-6
91-57-6
95-48-7
88-74-4
88-75-5

119-90-4
99-09-2

59-50-7

100-01-6
100-02-7
83-32-9
208-96-8
Chemical Name
1 ,2,3-Trichloropropane
1 ,2,4,5-Tetrachlorobenzene
1 ,2,4-Trichlorobenzene
1 ,2-Dibromo-3-chloropropane
1 ,3-Dinitrobenzene
1,4-Dioxane
1-Methyl-2-pentanone [WTI PIC LIST]
2,3,4,6-Tetrachlorophenol
2,4,5-Trichlorophenol
2.4-D
2,4-Oichlorophenol
2,4-Dimethylphenol
2,4-Dinftrophenol [WTI PIC LIST]
2-Butanone [WTI PIC LIST]
2-Chloroacetophenone
2-Chlorophenol
2-Chloropropane
2-Hexanone [WTI PIC LIST]
2-Methylnaphthalene [WTI PIC LIST]
2-Methylphenol [WTI PIC LIST]
2-NKroaniline
2-NKrophenol [WTI PIC LIST]
2-chloronaphthalene
3,3'-Dimethoxybenzidine
3-Nftroaniline [WTI PIC LIST]
4,6-Dinrtro-2-methylphenol [WTI PIC LIST]
4-Chloro-3-methylphenol [WTI PIC LIST]
4-Methyl-2-Pentanone
4-Nrtroaniline [WTI PIC LIST]
4-Nftrophenol
Acenaphthene [WTI PIC LIST]
Acenaphthylene [WTI PIC LIST]
Emission Rates, g/s
1A 1B 2 3A 3B 4A 4B


1.45E-04
O.OOE+00

4.94E-04

6.80E-06
6.69E-06
3.88E-05
6.69E-06
6.69E-06
6.69E-06
6.43E-05

6.69E-06

2.44E-05 6.43E-05
4.18E-05
6.69E-06
6.69E-06
6.69E-06
8.91 E-11 6.69E-06
1.15E-04
6.69E-06
6.69E-06
6.69E-06
6.43E-05
6.69E-06
6.69E-06
6.69E-06
6.69E-06
Maximum
Emission
Rate, g/s
O.OOE+00
O.OOE+00
1.45E-04
O.OOE+00
O.OOE+00
4.94E-04
O.OOE+00
6.80E-06
6.69E-06
3.88E-05
6.69E-06
6.69E-06
6.69E-06k
6.43E-05
O.OOE+00
6.69E-06
O.OOE+00
6.43E-05
4.18E-05
6.69E-06
6.69E-06
6.69E-06
6.69E-06
1.15E-04
6.69E-06
6.69E-06
6.69E-06
6.43E-05
6.69E-06
6.69E-06
6.69E-06
6.69E-06
  Appendix III-l
  Attachment 4
External Release Draft
 Do Not Cite or Quote

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
CAS No.
75-07-0
67-64-1
98-86-2
120-12-7
100-52-7
56-55-3

191-24-2

207-08-9
65-85-0
96-07-7
100-44-7
92-52-4
111-91-1

590-60-2
74-83-9
101-55-3
75-15-0
106-90-7
510-15-6
75-00-3
7005-72-3
4170-30-3
3547-04-4
117-84-0
53-70-3
84-74-2
75-71-8

84-66-2
Chemical Name
Acetaldehyde
Acetone [WTI PIC LIST]
Acetophenone
Anthracene
Benzaldehyde
Benzo(a)anthracene
Benzo(e)pyrene
Benzo(g,h,i)perylene
Benzo(j)fluoranthene
Benzo(k)fluoranthene
Benzoic acid {WTI PIC LIST]
Benzotrichloride
Benzyl chloride
Biphenyl
Bis(2-chloroethoxy)methane [WTI PIC LIST]
Bromochloromethane
Bromoethene
Bromomethane
Bromophenyl phenylether [WTI PIC LIST]
Carbon disulfide [WTI PIC LIST]
Chlorobenzene
Chlorobenzilate
Chloroethane [WTI PIC LIST]
Chlorophenyl phenylether [WTI PIC LIST]
Crotonaldehyde
DDE
Di(n)octy) phthalate
Dibenz(a,h)anthracene
Dibutyl phthalate [WTI PIC LIST]
Dichlorodifluoromethane
Dichlorofluoromethane [D & O PIC LIST]
Diethyl phthalate
Emission Rates, g/s
1A 1B 2 3A
3.01 E-04
2.90E-03 5.86E-04
2.93E-04


3.42E-05





3.20E-05


8.91 E-11


2.28E-05

2.40E-04 1.83E-04
4.83E-04 6.39E-05
3.68E-05


1.39E-04




2.19E-04

4.60E-04

3B 4A 4B



6.69E-06

6.69E-06

6.69E-06

6.69E-06
1.13E-05



6.69E-06



6.69E-06



1.29E-05
6.69E-06


6.69E-06
6.69E-06
2.50E-05


2.23E-05
Maximum
Emission
Rate, g/s
3.01 E-04
2.90E-03
2.93E-04
6.69E-06
O.OOE+00
3.42E-05
O.OOE+00
6.69E-06
O.OOE+00
6.69E-06
1.13E-05
3.20E-05
O.OOE+00
O.OOE+00
6.69E-06
O.OOE+00
O.OOE+00
2.28E-05
6.69E-06
2.40E-04
4.83E-04
3.68E-05
. 1.29E-05
6.69E-06
1.39E-04
O.OOE+00
6.69E-06
6.69E-06
2.50E-05
2.19E-04
O.OOE+00
4.60E-04
   Appendix III-l
   A'  \ment 4
External Release Draft
 Do Not Cite or    te

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
CAS No.
131-11-3
100-41-4
96-45-7
75-34-3
206-44-0
70-30-4
193-39-5
123-33-1
72-43-5
71-55-6
78-93-3
106-87-2
74-95-3
924-16-3
91-20-3
96-95-3
621-64-7
608-93-5
82-68-8
87-86-5
85-01-8
108-95-2
75-44-5
123-36-6
78-87-5
129-00-0
91-22-5
106-51-4
94-59-7

100-42-5
106-88-3
Chemical Name
Dimethyl phthalate
Ethylbenzene
Ethylene thiourea
Ethylidenedichloride
Fluoranthene
Hexachlorophene
lndeno(1 ,2,3-cd)pyrene
Maleic hydrazide
Methoxychlor
Methyl chloroform
Methyl ethyl ketone
Methylcyclohexane
Methylene bromide
N-Nitroso-di-n-butylamine
Naphthalene
Nitrobenzene
Nftroso-di-n-propylamine [WTI PIC LIST]
Pentachlorobenzene
Pentachloronftrobenzene
Pentachlorophenol
Phenanthrene [WTI PIC LIST]
Phenol
Phosgene
Propionaldehyde
Propylene dichloride
Pyrene [WTI PIC LIST]
Quinoline
Quinone
Safrole (5-(2-Propenyl)-1 ,3-benzodioxole)
Selenium
Styrene
Toluene
Emission Rates, g/s
1A 1B 2 3A 3B
1.68E-04
1.60E-03 2.69E-05
1.46E-10

1.19E-04
3.20E-05
1.19E-04
1.15E-04
3.72E-05
7.13E-05
3.50E-03


1.21E-04
3.45E-04
1.61E-03

4.76E-05
3.37E-05
6.94E-06
3.39E-10
3.16E-04



3.39E-06


1.15E-04
1.28E-08
2.44E-05 1.66E-05
4.08E-03 1.96E-04

4A 4B
6.69E-06



6.69E-06

6.69E-06







1.02E-05
6.69E-06
6.69E-06


6.69E-06
6.69E-06
6.69E-06



6.69E-06






Maximum
Emission
Rate, g/s
1.68E-04
1.60E-03
1.46E-10
O.OOE-fOO
1.19E-04
3.20E-05
1.19E-04
1.15E-04
3.72E-05
7.13E-05
3.50E-03
O.OOE-fOO
O.OOE+OO
1.21E-04
3.45E-04
161E-03
6.69E-06
4.76E-05
3.37E-05
6.94E-06
6.69E-06
3.16E-04
O.OOE-fOO
O.OOE+OO
O.OOE-fOO
6.69E-06
O.OOE-fOO
O.OOE-fOO
1.15E-04
1.28E-08
2.44E-05
4.08E-03
   Appendix III-l
   Attachment 4
External Release Draft
 Do Not Cite or Quote

-------
Attachment 4.  Estimated Emission Rates for Organic Compounds
CAS No.
75-69-4

1319-77-3
1330-20-7
110-54-3
1319-77-3
528-29-0
1330-20-7
1319-77-3
100-25-4
1330-20-7
319-84-6
91-58-7
319-85-7
Chemical Name
Trichlorofluoromethane
m,p-Xylene
m-Cresol
m-Xylene (m-Dimethyl benzene)
n-Hexane
o-Cresol
o-Dinftrobenzene
o-Xylene (o-Dimethyl benzene)
p-Cresol
p-Dinttrobenzene
p-Xylene (p-Dimethyl benzene)
-Hexachlorocyclohexane
-Chloronaphthalene
-Hexachlorocyclohexane
Maximum
Emission Rates, g/s Emission
1A 1B 2 3A 3B 4A 4B Rate, g/s
4.80E-04 4.80E-04
1.10E-04 1.10E-04
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
3.37E-05 3.37E-05
6.69E-06 6.69E-06
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
             lotal
Z.38E-OZ
                                                                                                                                   4.30E-02
Sources:
1A-3 Measured PIC's, Condition 2 of March '93 trial burn
1B--3 Measured PIC's, February '94 retest of trial burn Condition 2
2-DRE's applied to waste profile data, DRE's from Condition 2 of March '93 trial burn
3A--Other volatile PIC's, February '94 retest of Condition 2 (measured)
SB-Other volatile PIC's, February '94 retest of Condition 2 (non-detected, emission rates are based
  on detection limits)
4A--Other semi-volatile PIC's, February '94 retest of Condition 2 (measured)
4B--Other semi-volatile PIC's, February '94 retest of Condition 2 (non-detected, emission rates are based
  on detection limits)
   Appendix III-l
   Af   liment 4
                               External Release Draft
                                Do Not Cite or     te

-------
ATTACHMENTS

-------
                             ATTACHMENT
            WASTE FEED CUTOFF PERFORMANCE CATEGORIES
                   WASTE TECHNOLOGIES INDUSTRIES
                         EAST LIVERPOOL, OHIO
CLNK   Waste Feed Cutoffs (WFCOs) under this category are caused by
        clinkers falling into the slag quench, typically resulting in elevated SCC
        pressure.

CTRL   This category is comprised of WFCOs caused by a failure of control
        equipment, including the Bailey system and the Continuous Emissions
        Monitoring System (CEMS).  Examples of causes under this category
        include: CEMS analyzer failure; CEMS computer failure; CEMS
        calibration; and transmitter (e.g., thermocouple) failure.


ESP     This category of WFCOs are caused by conditions in the Electrostatic
        Precipitator (ESP).
FEED
FLOW
LMT
Under this category, the characteristics of the waste fed to the
incinerator contributed to the WFCOs.


This category covers WFCOs caused by the flow from the lances.
WFCOs under this category may be caused by:  variations in flow;
inadequate flow; loss  of flow; switching tanks; plugged lances, pump
failure; clogged strainers; and lance purges.

For this category, either no clear cause for the WFCO is provided or the
WFCO is caused by a prior WFCO.
MAIN   WFCOs under this designation are caused by maintenance activities.


MISC   This category covers WFCOs not included in other categories.


SCRB   Under this category, WFCOs are caused by conditions in the scrubber.


SD     WFCOs under this category are associated with the spray dryer. These
        typically involve a loss of flow from the spray dryer's atomizers.
Appendix ffl-1
Attachment 5
                                                External Release Draft
                                                Do Not Cite or Quote

-------
                               WASTE FEED CUTOFF PERFORMANCE SUMMARY
                                    WASTE TECHNOLOGIES INDUSTRIES
                                          EAST LIVERPOOL, OHIO

Description
CLNK
CTRL
ESP
FEED
FLOW
LMT
MAIN
MISC
SCRB
SD
Total
Hazardous Waste Hours
Avg. Hours Between WFCOs
Date
11/93
0
8
3
14
9
11
1
6
2
5
59
477
8.08
12/93
0
5
1
3
14
6
0
4
2
0
35
295
8.43
1/94
3
4
1
5
12
1
0
3
1
0
30
352
11.73
2/94
1
2
1
4
15
0
1
0
0
2
26
407
15.65
3/94
1
10
1
4
5
1
2
5
3
1
33
360
10.91
4/94
12
10
0
16
20
4
1
2
3
5
73
532
7.29
5/94
14
2
0
25
22
8
0
11
2
1
85
451
5.31
6/94
11
1
0
7
1
1
0
3
0
0
24
255
10.63
7/94
6
5
0
8
6
3
0
2
1
0
31
334
10.77
 Source: Waste Technologies Industries, Report of AWFCO Incidences to Ohio EPA, 1994.

 Note: Table 1 provides specific information on positive pressure AWFCOs as Identified by the facility.
Appendix III-l
A1"  iment 5
External Release Draft
Do Not Cite or/" ote

-------
ATTACHMENT 6

-------
                                                            TANKS PROGRAM  2.0
                                                EMISSIONS  REPORT  - DETAIL  FORMAT
                                     TANK IDENTIFICATION AND PHYSICAL CHARACTERISTICS
                                                                                                              10/16/95
                                                                                                              PAOE 1
Identification
   Identification No.:
   City:
   State:
   Company:
   Type of Tank:
B&H
Pittsburgh
PA
WTI
Vertical Fixed Roof
Tank Dimensions
   Shell Height (ft):               34
   Diameter (ft):                   10
   Liquid Height (ft):              34
   Avg. Liquid Height (ft):          33
   Volume (gallons):             19978
   Turnovers:                     739
   Net Throughput (gal/yr):     14617282

Paint Characteristics
   Shell Color/Shade:        White/White
   Shell Condition:          Good
   Roof Color/Shade:         White/White
   Roof Condition:           Good

Roof Characteristics
   Type:                   Cone
   Height (ft):                  0.00
   Radius (ft) (Dome Roof):        0.00
   Slope (ft/ft) (Cone Roof):    0.0000

Breather Vent Settings
   Vacuum Setting (psig):          0.00
   Pressure Setting (psig):        0.00
Meteorological Data Used in Emission Calculations:  Pittsburgh, Pennsylvania
          Appendix III-l
          Attachment 6
                                                                                              External Release Draft
                                                                                               Do Not Cite or Quote

-------
                                     TANKS PROGRAM 2.0
                             EMISSIONS  REPORT - DETAIL FORMAT
                              LIQUID CONTENTS OF  STORAGE TANK
            10/16/95
            PAGE 2
                               Liquid
Daily Liquid Surf. Bulk
Temperatures (deg F) Temp. Vapor Pressures
Mixture/Component Month Avg. Min. Max. (deg F) Avg. Min.
All 51.75
Octane (-n) '
Cresol (-0)
Toluene
Methyl alcohol
Methyl ethyl ketone
Acetone
Water
Cyclohexanone
Ethyl acrylate
Butanol-(l)
Methyl isobutyl ketone
Chloroform
Benzene
Formaldehyde
Hydrazine
Nitropropane(2-)
Acetonitrile
Carbon disulfide
Dimethyl hydrazine (1,1)
Dimethylamine
Xylene (-0)
47.02 56.48 50.32 0.5354
0.1228
0.0011
0.2540
1.1066
0.8537
2.3582
0.1919
0.0411
0.3753
0.0383
0.1582
1 .9471
0.9248
14.7000
0.0991
0.1532
0.8929
3.9755
1.4881
14.7000
0.0525
0.4818
0.1054
0.0008
0.2176
0.9464
0.7394
2.0752
0.1612
0.0350
0.3251
0.0304
0.1329
1.7048
0.8057
14.7000
0.0825
0.1312
0.7775
3.5542
1.2900
14.7000
0.0439
Vapor Liquid
(psia) ' Mol. Mass
Max. Weight Fract.
Vapor
Mass
Fract.
Mol.
Weight
Basis for Vapor Pressure
Calculations
0.5969 44.933
0.1426
0.0015
0.2956
1.2896
0.9824
2.6724
0.2293
0.0482
0.4321
0.0480
0.1874
2.2164
1.0583
14.7000
0.1195
0.1751
1.0224
4.4364
1.7111
14.7000
0.0626
0.2777
0.1018
0.0666
0.0508
0.0585
0.0481
0.1575
0.0417
0.0404
0.0402
0.0365
0.0040
0.0078
0.0045
0.0017
0.0143
0.0035
0.0020
0.0015
0.0020
0.0388
0.0752
0.0003
0.0373
0.1240
0.1102
0.2502
0.0667
0.0038
0.0334
0.0034
0.0127
0.0172
0.0159
0.1459
0.0004
0.0048
0.0069
0.0175
0.0049
0.0648
0.0045
114.23
108.14
92.13
32.04
72.10
58.08
18.00
98.20
100.11
74.12
100.20
119.39
78.11
30.02
32.05
89.09
41.05
76.13
60.10
45.08
106.17
Option
Option
Opt i on
Option
Opt i on
Opt i on
Option
Opt i on
Opt i on
Opt i on
Opt i on
Option
Option
Option
Option
Option
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
1
2:
2:
2:
2:
2:
1
2:
2:
2:
2:
2:
2:
1
1
1
2:
2:
2:
1
2:

A=6.9110,
A=6.9540,
A=7.8970,
A=6.9742,
A=7.1170,

A=7.8492,
A=7.9645,
A=7.4768,
A=6.6720,
A=6.4930,
A=6.9050,



A=7.1190,
A=6.9420,
A=7.4080,

A=6.9980,

8=1435.500,
8=1344.800,
8=1474.080,
6=1209.600,
8=1210.595,

8=2137.192,
8=1897.011,
8=1362.390,
8=1168.400,
6=929.440,
8=1211.033,



6=1314.400,
6=1169.110,
6=1305.910.

6=1474.679,

C=165.160
C=219.480
C=229.130
C=216.000
C=229.664

C=273.160
C=273.160
C= 178. 770
C=191.900
C= 196. 030
C=220.790



C=230.000
C=24 1.590
C=225.530

C=213.690
Appendix III-l
N" I|iment6
External Release Draft
Do Not Cite or   Nte

-------
                                                                 TANKS  PROGRAM  2.0
                                                     EMISSIONS  REPORT  -  DETAIL  FORMAT
                                                         DETAIL  CALCULATIONS  (AP-42)
                                                                                       10/16/95
                                                                                       PAGE 3
Annual  Emission Calculations

Standing  Losses (Ib):                          5.5107
  Vapor Space Volume (cu ft):                   78.54
  Vapor Density (Ib/cu  ft):                    0.0044
  Vapor Space Expansion Factor:               0.045101
  Vented  Vapor Saturation Factor:              0.972409

Tank Vapor Space Volume
  Vapor Space Volume (cu ft):                   78.54
  Tank  Diameter (ft):                              10
  Vapor Space Outage (ft):                       1.00
  Tank  Shell Height (ft):                          34
  Average Liquid Height (ft):                      33
  Roof  Outage (ft):                             0.00

Roof Outage (Cone Roof)
  Roof  Outage (ft):                             0.00 .
  Roof  Height (ft):                            0.000
  Roof  Slope (ft/ft):                          0.00000
  Shell Radius (ft):                               5

Vapor Density
  Vapor Density (Ib/cu  ft):                    0.0044
  Vapor Molecular Weight (Ib/lb-mole):        44.933135
  Vapor Pressure at Daily Average Liquid
  Surface Temperature (psia):                 0.535352
  Daily Avg. Liquid Surface Temp.(deg. R):      511.42
  Daily Average Ambient Temp. (deg. R):         509.97
  Ideal Gas Constant R
    (psia cuft /(lb-mole-deg R)):               10.731
  Liquid  Bulk Temperature (deg. R):            509.99
  Tank  Paint Solar Absorptance (Shell):           0.17
  Tank  Paint Solar Absorptance (Roof):            0.17
  Daily Total Solar Insolation
  Factor  (Btu/sqftday):                      1069.00

Vapor Space Expansion Factor
  Vapor Space Expansion Factor:               0.045101
  Daily Vapor Temperature Range (deg.R):         18.91
  Daily Vapor Pressure  Range (psia):          0.115095
  Breather Vent Press.  Setting Range(psia):       0.00
  Vapor Pressure at Daily Average Liquid
  Surface Temperature (psia):                 0.535352
  Vapor Pressure at Daily Minimum Liquid
  Surface Temperature (psia):                 0.481849
.  Vapor Pressure at Daily Maximum Liquid
  Surface Temperature (psia):                 0.596943
  Daily Avg. Liquid Surface Temp,  (deg R):      511.42
  Daily Min. Liquid Surface Temp,  (deg R):      506.69
  Daily Max. Liquid Surface Temp,  (deg R):      516.15
  Daily Ambient Temp.  Range (deg.R):            19.20
Appendix III-l
Attachment 6
External Release Draft
 Do Not Cite or Quote

-------
                                                             TANKS  PROGRAM  2.0
                                                  EMISSIONS  REPORT -  DETAIL  FORMAT-
                                                      DETAIL  CALCULATIONS  (AP-42)
                 10/16/95
                 PAGE  4
Annual Emission Calculations
Vented Vapor Saturation Factor
  Vented Vapor Saturation Factor:            0.972409
  Vapor Pressure at Daily Average Liquid
  Surface Temperature (psia):                0.535352
  Vapor Space Outage (ft):                      1.00

Withdrawal Losses (Ib):                    1738.5739
  Vapor Molecular Weight (Ib/lb-mole):       44.933135
  Vapor Pressure at Daily Average Liquid
  Surface Temperature (psia):                0.535352
  Annual Net Throughput (gal/yr):            14617282
  Turnover Factor:                           0.2077
  Maximum Liquid Volume (cuft):                  2670
  Maximum Liquid Height (ft):                     34
  Tank Diameter (ft):                            10
  Working Loss Product Factor:                   1.00

Total Losses (Ib):                          1744.08
          Appendix III-l
          Af  lament 6
External Release Draft
 Do Not Cite or f    te

-------
                                                    TANKS PROGRAM 2.0
                                           EMISSIONS  REPORT  -  DETAIL  FORMAT
                                           INDIVIDUAL TANK EMISSION TOTALS
                                                                                       10/16/95
                                                                                       PAGE 5
Annual Emissions Report
Liquid Contents
Losses (Ibs.):
Standing   Withdrawal
                                                   Total

Octane (-n)
Cresol (-0)
Toluene
Methyl alcohol
Methyl ethyl ketone
Acetone
Water
Cyclohexanone
Ethyl acrylate
Butanol-(l)
Methyl isobutyl ketone
Chloroform
Benzene
Formaldehyde
Hydrazine
Nitropropane(2-)
Acetonitrile
Carbon disulfide
Dimethyl hydrazine (1,1)
D i methyl ami ne
Xylene (-0)
5.51
0.41
0.00
0.21
0.68
0.61
1.38
0.37
0.02
0.18
0.02
0.07
0.09
.0.09
0.80
0.00
0.03
0.04
0.10
0.03
0.36
0.02
1738.57
130.70
0.44
64.88
215.58
191.51
434.97
115.92
6.57
58.14
5.90
22.14
29.87
27.66
253.67
0.65
8.40
11.98
30.49
8.56
112.74
7.82
1744.08
131.12
0.44
65.09
216.26
192.12
436.34
116.29
6.59
58.33
5.92
22.21
29.96
27.75
254.47
0.65
8.43
12.02
30.59
8.59
113.10
7.84
Total:
   5.51
1738.57
1744.08
         Appendix III-l
         Attachment 6
                                                                         External Release Draft
                                                                          Do Not Cite or Quote

-------
                                                            TANKS  PROGRAM  2.0
                                                 EMISSIONS REPORT  - DETAIL FORMAT
                                      TANK  IDENTIFICATION AND  PHYSICAL  CHARACTERISTICS
                                                                                                               10/16/95
                                                                                                               PAGE 1
Identification
   Identification No.:
   City:
   State:
   Company:
   Type of Tank:
P/0
Pittsburgh
PA
UTI
Vertical Fixed Roof
Tank Dimensions
   Shell Height (ft):                8
   Diameter  (ft):                    8
   Liquid Height (ft):               8
   Avg. Liquid Height (ft):           7
   Volume (gallons):              3008
   Turnovers:                    4906
   Net Throughput  (gal/yr):     14652507

Paint Characteristics
   Shell Color/Shade:        White/White
   Shell Condition:          Good
   Roof Color/Shade:         White/White
   Roof Condition:          Good

Roof Characteristics
   Type:                   Cone
   Height (ft):                  0.00
   Radius (ft) (Dome Roof):        0.00
   Slope (ft/ft) (Cone Roof):     0.0000

Breather Vent Settings
   Vacuum Setting (psig):         0.00
   Pressure Setting (psig):        0.00
Meteorological Data Used in Emission Calculations:  Pittsburgh,  Pennsylvania
           Appendix III-l
           Pf Ifunent 6
                                                                                               External Release Draft
                                                                                                Do Not Cite oif '~te

-------
                                      TANKS PROGRAM 2.0
                             EMISSIONS REPORT  -  DETAIL  FORMAT
                              LIQUID CONTENTS  OF STORAGE TANK
            10/16/95
            PAGE 2
                               Liquid
Daily Liquid Surf. Bulk
Temperatures (deg F) Temp.- Vapor Pressures
Mixture/Component Month Avg. Min. Max. (deg F) Avg. Min.
All 51.75
Acetone
Acetom'trile
Benzene
Butanol-(l)
Chloroform
Cresol (-0)
Cyclohexanone
Dichloroethane (1,1)
Dimethyl ami ne
Ethyl acrylate
Methyl alcohol
Methyl ethyl ketone
Methyl isobutyl ketone
Octane (-n)
Toluene
Trichloroethane (1,1,1)
Trichloroethylene
Water
Xyiene (-0)
47.02 56.48 50.32 0.
2.
0.
0.
0.
1.
0.
0.
2.
14
0.
1.
0.
0.
0.
0.
1.
0.
0.
0.
4227
3582
8929
9248
0383
9471
0011
0411
3582
.7000
3753
1066
8537
1582
1228
2540
3182
6356
1919
0525
0.3685
2.0752
0.7775
0.8057
0.0304
1.7048
0.0008
0.0350
2.0759
14.7000
0.3251
0.9464
0.7394
0.1329
0.1054
0.2176
1.1548
0.5469
0.1612
0.0439
Vapor Liquid
(psia) Mol. Mass
Max. Weight Fract.
Vapor
Mass
Fract.
Mol.
Weight
Basis for Vapor Pressure
Calculations
0.4850 49.815
2.6724
1.0224
1.0583
0.0480
2.2164
0.0015
0.0482
2.6693
14.7000
0.4321
1.2896
0.9824
0.1874
0.1426
0.2956
1.5015
0.7360
0.2293
0.0626
0.0486
0.0035
0.0078
0.0406
0.0040
0.1030
0.0422
0.0016
0.0020
0.0408
0.0513
0.0592
0.0369
0.2807
0.0674
0.0074
0.0045
0.1593
0.0392
0.. 2904
0.0079
0.0183
0.0039
0.0197
0.0003
0.0044
0.0096
0.0745
0.0388
0.1439
0.1281
0.0148
0.0873
0.0434
0.0247
0.0072
0.0775
0.0052
58.08
41.05
78.11
74.12
119.39
108.14
98.20
98.97
45.08
100.11
32.04
72.10
100.20
114.23
92.13
133.42
131.40
18.00
106.17
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 1
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 2:
Option 2:
Option 2:
Option 1
Option 2:
A=7.1170,
A=7.1190,
A=6.9050,
A=7.4768,
A=6.4930,
A=6.9110,
A=7.8492,


A=7.9645,
A=7.8970,
A=6.9742,
A=6.6720,

A=6.9540.
A=8.6430,
A=6.5180,

A=6.9980,
8=1210.595,
6=1314.400,
8=1211.033,
8=1362.390,
8=929.440,
8=1435.500,
8=2137.192,


8=1897.011,
8=1474.080,
8=1209.600,
8=1168.400,

8=1344.800,
8=2136.600,
6=1018.600,

6=1474.679,
C=229.664
C=230.000
C=220.790
C=178.770
C= 196. 030
C=165.160
C=273.160


C=273.160
C=229.130
C=216.000
C= 191. 900

C=219.480
C=302.BOO
C= 192. 700

C=213.690
Appendix III-l
Attachment 6
External Release Draft
Do Not Cite or Quote

-------
                                                                 TANKS  PROGRAM  2.0
                                                     EMISSIONS  REPORT  -  DETAIL  FORMAT
                                                         DETAIL  CALCULATIONS  (AP-42)
                                                                                       10/16/95
                                                                                       PAGE  3
Annual  Emission Calculations

Standing Losses (Ib):                           3.1075
  Vapor Space Volume (cu ft):                    50.27
  Vapor Density (Ib/cu ft):                     0.0038
  Vapor Space Expansion Factor:               0.045138
  Vented Vapor Saturation Factor:              0.978090

Tank Vapor Space Volume
  Vapor Space Volume (cu ft):                    50.27
  Tank  Diameter (ft):                               8
  Vapor Space Outage (ft):                        1.00
  Tank  Shell Height (ft):                           8
  Average Liquid Height (ft):                       7
  Roof  Outage (ft):                              0.00

Roof Outage  (Cone Roof)
  Roof  Outage (ft):                              0.00
  Roof  Height (ft):                             0.000
  Roof  Slope (ft/ft):                          0.00000
  Shell Radius (ft):                               4

Vapor Density
  Vapor Density (Ib/cu ft):                     0.0038
  Vapor Molecular Weight (Ib/lb-mole):        49.814730
  Vapor Pressure at Daily Average Liquid
  Surface Temperature (psia):                 0.422662
  Daily Avg. Liquid Surface  Temp.(deg.  R):       511.42
  Daily Average Ambient Temp.  (deg. R):         509.97
  Ideal Gas  Constant R
    (psia cuft /(lb-mole-deg R)):               10.731
  Liquid Bulk-Temperature  (deg. R):             509.99
  Tank Paint Solar Absorptance (Shell):           0.17
  Tank Paint Solar Absorptance (Roof):             0.17
  Daily Total Solar Insolation
  Factor  (Btu/sqftday):                      1069.00

Vapor Space  Expansion Factor
  Vapor Space Expansion Factor:               0.045138
  Daily Vapor Temperature  Range (deg.R):         18.91
  Daily Vapor Pressure Range (psia):          0.116535
  Breather Vent Press. Setting Range(psia):        0.00
  Vapor Pressure at Daily  Average Liquid
  Surface  Temperature (psia):                 0.422662
  Vapor Pressure at Daily  Minimum Liquid
  Surface  Temperature (psia):                 0.368506
  Vapor Pressure at Daily  Maximum Liquid
  Surface  Temperature (psia):                 0.485042
  Daily Avg. Liquid Surface Temp, (deg R):       511.42
  Daily Min. Liquid Surface Temp, (deg R):       506.69
  Daily Max. Liquid Surface Temp, (deg R):       516.15
  Daily AmbierT imp. Range (deg.R):             19.20
Appendix III-l
Attachment 6
8
External Release Draft
 Do Not Cite orf~  ^te

-------
                                                             TANKS PROGRAM  2.0
                                                 EMISSIONS  REPORT  -  DETAIL  FORMAT
                                                     DETAIL  CALCULATIONS  (AP-42)
                10/16/95
                PAGE 4
Annual Emission Calculations
Vented Vapor Saturation Factor
  Vented Vapor Saturation Factor:             0.978090
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                0.422662
  Vapor Space Outage (ft):                     1.00

Withdrawal Losses (tb):                    1269.4712
  Vapor Molecular Weight (Ib/lb-mole):       49.814730
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                0.422662
  Annual Net Throughput (gal/yr):             14652507
  Turnover Factor:                           0.1728
  Maximum Liquid Volume (cuft):                  402
  Maximum Liquid Height (ft):                      8
  Tank Diameter (ft):                             8
  Working Loss Product Factor:                  1.00

Total Losses (Ib):                 .         1272.58
          Appendix III-l
          Attachment 6
External Release Draft
 Do Not Cite or Quote

-------
                                                   TANKS PROGRAM 2.0
                                           EMISSIONS  REPORT -  DETAIL FORMAT
                                           INDIVIDUAL TANK EMISSION TOTALS
                                                                                      10/16/95
                                                                                      PAGE  5
Annual Emissions Report
Liquid Contents
Losses (Ibs.):.
Standing   Withdrawal
            Total

Acetone
Acetonitrile
Benzene
Butanol-(1)
Chloroform
Cresol (-0)
Cyctohexanone
Dichloroethane (1,1)
D i methyl ami ne
Ethyl acrylate
Methyl alcohol
Methyl ethyl ketone
Methyl isobutyl ketone
Octane (-n)
Toluene
Trichloroethane (1,1,1)
Trichloroethylene
Uater
Xylene (-0)
3.11
0.90
0.02
0.06
0.01
0.06
0.00
0.01
0.03
0.23
0.12
0.45
0.40
0.05
0.27
0.13
0.08
0.02
0.24
0.02
1269.47
368.70
10.05
23.21
5.00
25.06
0.37
5.58
12.14
94.58
49.26
182.63
162.59
18.78
110.87
55.08
31.38
9.20
98.36
6.62
1272.58
369.60
10.08
23.26
5.01
25.12
0.37
5.59
12.17
94.81
49.38
183.08
162.98
18.82
111.15
55.22
31.46
9.22
98.60
6.64
Total:
   3.11
1269.47
1272.58
         Appendix III-l
         tf \hment6
                                                                        External Release Draft
                                                                        Do Not Cite or/

-------
                                                            TANKS  PROGRAM 2.0
                                                 EMISSIONS  REPORT -  DETAIL  FORMAT
                                     TANK  IDENTIFICATION  AND  PHYSICAL CHARACTERISTICS
                                                                                                              10/16/95
                                                                                                              PAGE  1
Identification
   Identification No.:
   City:
   State:
   Company:
   Type of Tank:
Rec
Pittsburgh
PA
UTI
Vertical Fixed Roof
Tank Dimensions
   Shell Height (ft):               19
   Diameter (ft):                    8
   Liquid Height (ft):              19
   Avg. Liquid Height (ft):          18
   Volume (gallons):              7145
   Turnovers:                    2065
   Net Throughput  (gal/yr):     14652507

Paint Characteristics
   Shell Color/Shade:        White/White
   Shell Condition:          Good
   Roof Color/Shade:         White/White
   Roof Condition:           Good

Roof Characteristics
   Type:                   Cone
   Height (ft):                  0.00
   Radius (ft) (Dome Roof):        0.00
   Slope (ft/ft) (Cone Roof):    0.0000

Breather Vent Settings
   Vacuum Setting  (psig):          0.00
   Pressure Setting (psig):        0.00
Meteorological Data Used in Emission Calculations:  Pittsburgh, Pennsylvania
          Appendix III-l
          Attachment 6
                                                   11
External Release Draft
 Do Not Cite or Quote

-------
                                      TANKS  PROGRAM  2.0
                             EMISSIONS REPORT - DETAIL FORMAT
                              LIQUID CONTENTS OF STORAGE TANK
            10/16/95
            PAGE 2
                               Liquid
Daily Liquid Surf. Bulk
Temperatures (deg F) Temp. Vapor Pressures
Mixture/Component Month Avg. Min. Hex. (deg F) Avg. Min.
All 51.75
Acetone
Acetonitrile
Benzene
Butanol-(1)
Chloroform
Cresol (-0)
Cyclohexanone
Oichloroethane (1,1)
Oimethylamine
Ethyl acrylate
Methyl alcohol
Methyl ethyl ketone
Methyl isobutyl ketone
Octane (-n)
Toluene
Trichloroethane (1,1,1)
Trichloroethylene
Water
Xylene (-0)
47.02 56.48 50.32 0.
2.
0.
0.
0.
1.
0.
0.
2.
14
0.
1.
0.
0.
0.
0.
1.
0.
0.
0.
4227
3582
8929
9248
0383
9471
0011
0411
3582
.7000
3753
1066
8537
1582
1228
2540
3182
6356
1919
0525
0.3685
2.0752
0.7775
0.8057
0.0304
1.7048
0.0008
0.0350
2.0759
14.7000
0.3251
0.9464
0.7394
0.1329
0.1054
0.2176
1.1548
0.5469
0.1612
0.0439
Vapor Liquid
(psia) Mol. Mass
Max. Weight Fract.
Vapor
Mass
Fract.
Mol.
Weight
Basis for Vapor Pressure
Calculations
0.4850 49.815
2.6724
1.0224
1.0583
0.0480
2.2164
0.0015
0.0482
2.6693
14.7000
0.4321
1.2896
0.9824
0.1874
0.1426
0.2956
1.5015
0.7360
0.2293
0.0626
0.0486
0,0035
0.0078
0.0406
0.0040
0.1030
0.0422
0.0016
0.0020
0.0408
0.0513
0.0592
0.0369
0.2807
0.0674
0.0074
0.0045
0.1593
0.0392
0.2904
0.0079
0.0183
0.0039
0.0197
0.0003
0.0044
0.0096
0.0745
0.0388
0.1439
0.1281
0.0148
0.0873
0.0434
0.0247
0.0072
0.0775
0.0052
58.08
41.05
78.11
74.12
119.39
108.14
98.20
98.97
45.08
100.11
32.04
72.10
100.20
114.23
92.13
133.42
131.40
18.00
106.17
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 1
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 2:
Option 2:
Option 2:
Option 1
Option 2:
A=7.1170,
A=7.1190,
A=6.9050,
A=7.4768,
A=6.4930,
A=6.9110,
A=7.8492,


A=7.9645,
A=7.8970,
A=6.9742,
A=6.6720,

A=6.9540,
A=8.6430,
A=6.5180,

A=6.9980,
8=1210.595,
8=1314.400,
8=1211.033,
8=1362.390,
8=929.440,
8=1435.500,
8=2137.192,


8=1897.011,
8=1474.080,
8=1209.600,
8=1168.400,

8=1344.800,
8=2136.600,
8=1018.600,

8=1474.679,
C=229.664
C=230.000
C=220.790
C=178.770
C= 196. 030
C=165.160
C=273.160


C=273.160
C=229.130
C=216.000
C=191.900

C=219.«80
C=302.800
C= 192. 700

C=213.690
Appendix III-l
A  ^unent 6
External Release Draft
Do Not Cite or (  \te

-------
                                                                 TANKS  PROGRAM  2.0
                                                     EMISSIONS  REPORT  -  DETAIL  FORMAT
                                                         DETAIL  CALCULATIONS  (AP-42)
                                                                                       10/16/95
                                                                                       PAGE  3
Annual  Emission Calculations

Standing  Losses (Ib):                          3.1075
  Vapor Space Volume (cu ft):                   50.27
  Vapor Density (Ib/cu ft):                    0.0038
  Vapor Space Expansion Factor:               0.045138
  Vented  Vapor Saturation Factor:              0.978090

Tank Vapor Space Volume
  Vapor Space Volume (cu ft):                   50.27
  Tank  Diameter (ft):                               8
  Vapor Space Outage (ft):                       1.00
  Tank  Shell Height (ft):                          19
  Average Liquid Height (ft):                      18
  Roof  Outage (ft):                             0.00

Roof Outage (Cone Roof)
  Roof  Outage (ft):                             0.00
  Roof  Height (ft):                            0.000
  Roof  Slope (ft/ft):                          0.00000
  Shell Radius (ft):                               4

Vapor Density
  Vapor Density (Ib/cu ft):                    0.0038
  Vapor Molecular Weight (Ib/lb-mole):        49.814730
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                 0.422662
  Daily Avg. Liquid Surface Temp.(deg. R):      511.42
  Daily Average Ambient Temp. (deg.  R):         509.97
  Ideal Gas Constant R
    (psia cuft /(tb-mole-deg R)):               10.731
  Liquid  Bulk Temperature (deg. R):             509.99
  Tank  Paint Solar Absorptance (Shell):           0.17
  Tank  Paint Solar Absorptance (Roof):            0.17
  Daily Total Solar Insolation
  Factor  (Btu/sqftday):                      1069.00

Vapor Space Expansion Factor
  Vapor Space Expansion Factor:               0.045138
  Daily Vapor Temperature Range (deg.R):         18.91
  Daily Vapor Pressure Range (psia):           0.116535
  Breather Vent Press.  Setting Range(psia):       0.00
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                 0.422662
  Vapor Pressure at Daily Minimum  Liquid
  Surface Temperature (psia):                 0.368506
  Vapor Pressure at Daily Maximum  Liquid
  Surface Temperature (psia):                 0.485042
  Daily Avg. Liquid Surface Temp,  (deg R):      511.42
  Daily Min. Liquid Surface Temp,  (deg R):      506.69
  Daily Max. Liquid Surface Temp,  (deg R):      516.15
  Daily Ambient Temp.  Range (deg.R):             19.20
Appendix III-l
Attachment 6
13
External Release Draft
 Do Not Cite or Quote

-------
                                                             TANKS PROGRAM 2.0
                                                 EMISSIONS  REPORT  -  DETAIL  FORMAT
                                                      DETAIL  CALCULATIONS  (AP-42)
                 10/16/95
                 PAGE  4
Annual Emission Calculations
Vented Vapor Saturation Factor
  Vented Vapor Saturation Factor:             0.978090
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                0.422662
  Vapor Space Outage (ft):                      1.00

Withdrawal Losses (Ib):                    1331.68H
  Vapor Molecular Weight (Ib/lb-mole):       49.814730
  Vapor Pressure at Daily Average  Liquid
  Surface Temperature (psia):                0.422662
  Annual Net Throughput (gal/yr):             14652507
  Turnover Factor:                           0.1813
  Maximum Liquid Volume (cuft):                  955
  Maximum Liquid Height (ft):                     19
  Tank Diameter (ft):                             8
  Working Loss Product Factor:                   1.00

Total Losses (Ib):                          1334.79
          Appendix III-l
          A    %mi»nt fi
External Release Draft
 Do Not Cite or    \e

-------
                                                    TANKS PROGRAM 2.0
                                           EMISSIONS REPORT -  DETAIL FORMAT
                                           INDIVIDUAL  TANK  EMISSION  TOTALS
                                                                                       10/16/95
                                                                                       PAGE 5
Annual Emissions Report
Liquid Contents
Losses (Ibs.):
Standing   Withdrawal
            Total

Acetone
Acetonitrile
Benzene
Butanol-(1)
Chloroform
Cresol (-0)
Cyclohexanone
Dichloroethane (1,1)
Dimethylamine
Ethyl acrylate
Methyl alcohol
Methyl ethyl ketone
Methyl isobutyl ketone
Octane (-n)
Toluene
Trichloroethane (1,1,1)
Trichloroethylene
Water
Xylene (-0)
3.11
0.90
0.02
0.06
0.01
0.06
0.00
0.01
0.03
0.23
0.12
0.45
0.40
0.05
0.27
0.13
0.08
0.02
0.24
0.02
1331.68
386.77
10.55
24.34
5.24
26.28
0.39
5.85
12.73
99.22
51.67
191.58
170.55
19.70
116.31
57.78
32.92
9.65
103.18
6.95
1334.79
387.67
10.57
24.40
5.26
26.34
0.39
5.87
12.76
99.45
51.80
192.03
170.95
19.74
116.58
57.92
33.00
9.68
103.42
6.97
Total:
   3.11
1331.68
1334.79
         Appendix III-l
         Attachment 6
                                   15
                                                              External Release Draft
                                                              Do Not Cite or Quote

-------
APPENDIX ffl-2

-------

APPENDIX ffl-2
Measured Dioxm/Furan Congener Emission Rates
August 1993 ECIS Performance Test
Dioxm/Furan Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
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
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Emission Rate (g/sec) - August 1993 ECIS Performance Test
Rnnl
<2.32e-ll
1.72e-10
2.106-10
2.86e-10
2.10e-10
1.616-09
6.47e-09
1.94e-10
8.08e-10
9.70e-10
2.48e-09
2.64e-09
2.96e-09
4.15e-10
1.24e-08
1.67e-09
1.83e-08
Run 2
< 2.186-11
1.656-10
2.02e-10
3.406-10
2.23e-10
1.546-09
5.846-09
1.38e-10
7.43e-10
1.066-09
3.346-09
3.026-09
3.87e-09
7.96e-10
1.59e-08
2.28e-09
2.39e-08
Run3
< 3. 07e-ll
2.32e-10
3.72e-10
5.23e-10
4.26e-10
5.28e-09
5.39e-08
2.37e-10
8.08e-10
1.51e-09
5.12e-09
4.58e-09
5.39e-09
1.136-09
5.93e-08
9.16e-09
2.69e-07
Run 4
<3.72e-ll
1.65e-10
2.23e-10
2.92e-10
1.916-10
l.Ole-09
3.246-09
l.Ole-10
7.446-10
1.28e-09
3.78e-09
3.14e-09
3.67e-09
6.91e-10
1.28e-08
1.816-09
1.70e-08
Run5
<3. 18e-ll
1.67e-10
2.37e-10
3.83e-10
2.21e-10
l.Sle-09
5.17e-09
7.546-11
8.08e-10
1.24e-09
3.77e-09
3.29e-09
4.15e-09
7.00e-10
1.45e-08
2.16e-09
1.67e-08
Volume III
Appendix III-2
-1-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-2
Measured Dioxin/Furan Congener Emission Rates
February 1994 Trial Burn
Dioxin/Furan Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
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
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Emission Rate fe/sec) — 1

Runl
<2.86e-ll
6.83e-ll
1.18e-10
2.23e-10
1.18e-10
1.32e-09
4.046-09
l.lle-10
4.67e-10
5.29e-10
2.02e-09
1.88e-09
1.88e-09
3.76e-10
1.25e-08
1.606-09
1.46e-08

Run 2
<3.98e-ll
6.486-11
9.446-11
1.756-10
l.Ole-10
l.Ole-09
2.43e-09
1. 15e-10
3.98e-10
4.86e-10
1.62e-09
1.55e-09
1.69e-09
3.44e-10
l.Ole-08
1.28e-09
8.77e-09
February 1994 Trial Burn
Run 3
<2.28e-ll
5.516-11
8.066-11
1.686-10
8.066-11
8.06e-10
2.226-09
8.066-11
2.89e-10
3.97e-10
1.34e-09
1.28e-09
1.48e-09
2.69e-10
8.06e-09
l.Ole-09
7.39e-09
Run 4
<2.31e-ll
5.956-11
7.276-11
1.78e-10
8.59e-ll
8.59e-10
1.78e-09
1.52e-10
3.23e-10
4.49e-10
1.39e-09
1.26e-09
1.39e-09
2.84e-10
7.93e-09
8.596-10
5.22e-09
Volume III
Appendix III-2
-2-
External Review Draft
Do Not Cite or Quote

-------
f" "'•
\
APPENDIX ra-2
Measured Dioxin/Furan Congener Emission Rates
February 1994 ECIS Performance Test
Dioxin/Furan Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
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
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Emission Rate (g/sec) - February 1994 ECIS Performance Test
Run 2
<5.90e-ll
1.27e-10
2.09e-10
4.25e-10
3.51e-10
5.75e-09
2.616-08
2.31e-10
6.19e-10
6.79e-10
1.42e-09
1.42e-09
1.57e-09
4.636-10
8.216-09
1.12e-09
7.316-09
Run 3
< 2.206-11
3.42e-ll
7.36e-ll
1.21e-10
6.07e-ll
7.59e-10
2.28e-09
7.296-11
2.43e-10
2.66e-10
9.11e-10
8.35e-10
9.11e-10
1.67e-10
5.16e-09
4.93e-10
3.34e-09
Run 4
<4.03e-ll
9.126-11
1.37e-10
2.05e-10
1.82e-10
2.136-09
7.606-09
1.22e-10
4.036-10
3.87e-10
1.06e-09
1.066-09
1.146-09
2.58e-10
6.08e-09
6.466-10
4.03e-09
Run5
< 4.58e-ll
6.996-11.
9.026-11
1.65e-10
8.276-11
9.01e-10
2.48e-09
6.76e-ll
3.38e-10
3.916-10
1.28e-09
1.28e-09
1.50e-09
3.16e-10
8.27e-09
8.27e-10
5.56e-09
Run 7
<6.72e-ll
1.36e-10
1.43e-10
2.276-10
1.81e-10
1.21e-09
3.17e-09
2.27e-10
5.89e-10
5.51e-10
1.36e-09
1.286-09
1.286-09
2.79e-10
6.65e-09
6.19e-10
3.32e-09
           Volume III
           Appendix III-2
-3-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-2
Measured Dioxin/Furan Congener Emission Rates
April 1994 ECIS Performance Test
Dioxin/Furan Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
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
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Emission Rate (g/sec)

Runl
< 6,27e-ll
<4.01e-ll
< 6.56e-ll
<1.17e-10
<1.02e-10
8.02e-10
5.54e-09
8.026-11
1.60e-10
2.62e-10
6.34e-10
7.296-10
7.296-10
1.24e-10
7.29e-09
9.48e-10
2.116-08
Run 2
<3.56e-ll
<4. 17e-ll
<7.78e-ll
8.346-11
< 6. 12e-ll
9.45e-10
7.78e-09
5.346-11
1.28e-10
1.726-10
6.67e-10
6.67e-10
6.67e-10
1.39e-10
8.90e-09
9.45e-10
2.22e-08
- April 1994 ECIS Performance Test
Run3
<3.79e-ll
<4.12e-ll
< 6. 13e-ll
1.06e-10
< 5.296-11
8.36e-10
3.57e-09
3.066-11
l.lle-10
1.84e-10
6.69e-10
6.69e-10
7.24e-10
1.28e-10
6.696-09
7.80e-10
l.lle-08
Run 4
<4.23e-ll
<5.26e-ll
< 8. 13e-ll
1.036-10
<7.04e-ll
5.42e-10
2.06e-09
4.126-11
1.35e-10
2.17e-10
7.04e-10
7.04e-10
8.136-10
1.52e-10
5.96e-09
5.96e-10
7.04e-09
Run5
<3.37e-ll
<4.64e-ll
<7. 18e-ll
9.396-11
< 5.306-11
4.52e-10
1.71e-09
3.316-11
1.27e-10
1.32e-10
6.63e-10
6.63e-10
4.646-10
9.946-11
4.696-09
3.426-10
4.86e-09
Volume III
Appendix III-2
-4-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX III-2
Measured Dioxin/Furan Emission Rates
August 1994 ECIS Performance Test
Dioxin/Furan Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD

1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
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
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Emission Rate (g/sec) - August 1994 ECIS Performance Test
Run 1
<2.16e-ll
<4.21e-ll
<1.02e-10
<8.53e-ll
<9.10e-ll
6.26e-10
4.15e-09
< 5.29e-ll
1.19e-10
1.31e-10
3.816-10
3.81e-10
4.386-10
1.146-10
3.36e-09
5.57e-10
5.52e-09
Run 2
< 5.226-11
<4.49e-ll
< 6. 18e-ll
< 5.226-11
<5.51e-ll
2.58e-10
6.18e-10
<3.93e-ll
< 7.866-11
8.996-11
2.926-10
2.706-10
2.87e-10
<3.71e-ll
1.85e-09
2.30e-10
< 8.99e-10
Run 3
<3.33e-ll
<4.55e-ll
< 1.05e-10
< 8.32e-ll
<8.87e-ll
3.94e-10
1.226-09
<3.33e-ll
8.886-11
1.33e-10
3.72e-10
3.666-10
4.05e-10
8.326-11
2.94e-09
3.05e-10
2.33e-09
Run 4
< 2.756-11
<3.65e-ll
< 6.736-11
< 8.416-11
<5.61e-ll
3.08e-10
6.73e-10
< 5.056-11
1.12e-10
1.686-10
3.98e-10
3.936-10
3.93e-10
< 5.55e-ll
2.36e-09
2.97e-10
< 1.126-09
RunS
< 4.086-11
<5.31e-ll
<6.71e-ll
< 5.596-11
< 6. 15e-ll
3.80e-10
1.736-09
<3.69e-ll
l.Ole-10
1.12e-10
3.97e-10
3.91e-10
4.02e-10
1.12e-10
3.13e-09
4.08e-10
3.80e-09
Run 6
<4.49e-ll
<2.59e-ll
< 8.636-11
< 6.906-11
< 7.486-11
4.77e-10
1.55e-09
< 5. 12e-ll
1.32e-10
1.67e-10
4.776-10
4.66e-10
4.54e-10
1.21e-10
3.05e-09
3.39e-10
2.07e-09
Run 7
<3.43e-ll
<4.72e-ll
< 1.076-11
< 8.426-11
< 8.996-11
5.34e-10
2.47e-09
< 4.27e-ll
1.29e-10
1.69e-10
5.28e-10
4.94e-10
4.66e-10
< 7.866-11
3.826-09
4.66e-10
< 6. 18e-09
Volume HI
Appendix III-2
-5-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX ffl-3

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
Substance
Acenaphthene
Acenaphthylene
Acetaldehyde
Acetone
Acetophenone
Acrolein*
Acrylonitrile
Anthracene
Benzaldehyde*
Benzene
Benzole acid
Benzotrichloride
Benzo(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(e)pyrene*
Benzo(g4i,i)perylene
Benzo(j)fiuoranthene*
Benzo(k)fluoranthene
Benzyl chloride*
Biphenyl*
Bis(2-chloroethoxy)methane
Bis(2-chloroethyl)ether
Bis(2-chloroisopropyl)ether
Bis(2-ethylhexyl)phthalate
Bromodichloromethane
Bromodichloromethane*
Bromoethene*
Bromoform
Bromomethane
Compounds Detected in
Stack Emissions



Feb-94





Mar-93, Feb-94, Aug-94
Feb-94













Feb-94, Aug-94
Feb-94, Aug-94


Feb-94
Feb-94
Compounds Analyzed for but Not
Detected in Stack Emissions
Feb-94
Feb-94




Mar-93
Feb-94, Aug-94




Feb-94, Aug-94
Feb-94, Aug-94
Feb-94, Aug-94

Feb-94, Aug-94

Feb-94, Aug-94


Feb-94
Feb-94
Feb-94




Aug-94
Aug-94
Volume III
Appendix III-3
-1-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Bromodiphenylether, p-
Butadiene, 1,3-*
Butanone, 2- (MEK)
Butylbenzylphthalate
Carbon disulfide
Carbon tetrachloride
Chlordane
Chloro-3-methylphenol, 4-
Chloroacetophenone, 2-*
Chloroaniline, p-
Chlorobenzene
Chlorobenzilate
CHoroethane
Chloroform
Chloromethane
Chloronaphthalene, beta-
Chlorophenol, 2-
Chlorodiphenyl ether, 4-
Chloropropane, 2-*
Chrysene
Cresol, m-
Cresol, o-
Cresol, p-
Crotonaldehyde
Cumene
2,4-D
4,4'-DDE
Dibenz(a,h)anthracene
Dibenzo(a,h)fluoranthene
Compounds Detected in
Stack Emissions


Aug-94

Feb-94, Aug-94
Aug-94







Mar-93, Feb-94, Aug-94
Feb-94














Compounds Analyzed for but Not
Detected in Stack Emissions
Feb-94


Feb-94, Aug-94


Aug-94
Feb-94

Feb-94
Aug-94

Feb-94, Aug-94

Aug-94
Feb-94
Aug-94
Feb-94

Feb-94, Aug-94
Aug-94
Aug-94
Aug-94

Aug-94

Aug-94
Feb-94, Aug-94
Aug-94
Volume III
Appendix III-3
-2-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Dibromo-3-chloropropane, 1,2-*
Dibromocbloromethane
Dichloro-2-butene, cis-1,4-*
Dichloro-2-butene, trans-1,4-*
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine, 3,3'-
Dichlorobiphenyl
Dichlorodifluoromethane
Dichloroethane, 1,1-
Dichloroetbane, 1,2-
Dichloroethene, 1,1-
Dichloroethene, trans-1,2-
Dichlorofluoromethane*
Dichlorophenol, 2,4-
Dichloropropane, 1,2-
Dichloropropene, cis-1,3-
Dichloropropene, trans-1,3-
Diethylphthalate
Dimethoxybenzidine, 3,3'-
Dimethylphenol, 2,4-
Dimethylphthalate
Di-n-butylphthalate
Di-n-octyl phthalate
Diiiitrotoliiene, 2,6-
Dinitro-2-methylphenol, 4,6-
Dinitrobenzene, 1,2-*
Dinitrobenzene, 1,3-*
Compounds Detected in
Stack Emissions

Feb-94


Feb-94

Feb-94

Aug-94










Feb-94, Aug-94



Feb-94, Aug-94





Compounds Analyzed for but Not
Detected in Stack Emissions




Mar-93, Aug-94
Mar-93, Feb-94, Aug-94
Mar-93, Aug-94
Mar-93, Feb-94

Aug-94
Feb-94, Aug-94
Mar-93, Feb-94, Aug-94
Feb-94, Aug-94
Feb-94, Aug-94

Feb-94, Aug-94
Feb-94, Aug-94
Mar-93, Feb-94, Aug-94
Mar-93, Feb-94, Aug-94


Feb-94, Aug-94
Feb-94, Aug-94

Feb-94, Aug-94
Feb-94, Aug-94
Feb-94, Aug-94


Volume III
Appendix III-3
-3-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Dinitrobenzene, 1,4-*
Dimtrophenol, 2,4-
Dinitrotoluene, 2,4-
Dioxane, 1,4-
Ethyl methacrylate
Ethylbenzene
Ethylene dibromide
Ethylene oxide
Ethylene thiourea
Fluoranthene
Fluorene
Formaldehyde
Furfural
Heptachlor
Heptachlorobiphenyl
Hexachlorobenzene
Hexachlorobiphenyl
Hexachlorobutadiene
Hexachlorocyclohexane, alpha-*
Hexachlorocyclohexane, beta-*
Hexachlorocyclohexane, gamma- (a.k.a.
Lindane)
Hexachlorocyclopentadiene
Hexachloroethane
Hexachlorophene
Hexane, n-*
Hexanone, 2-
Hexanone, 3-*
Indeno(l ,2,3-cd)pyrene
Componnds Detected in
Stack Emissions





Feb-94, Aug-94






















Compounds Analyzed for but Not
Detected in .Stack Emissions

Feb-94, Aug-94
Feb-94, Aug-94

Aug-94




Feb-94, Aug-94
Feb-94

Aug-94
Aug-94
Aug-94
Feb-94, Aug-94
Aug-94
Feb-94



Feb-94, Aug-94
Mar-93, Feb-94, Aug-94


Feb-94

Feb-94, Aug-94
Volume III
Appendix III-3
-4-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Isophorone
Maleic hydrazide
Methoxychlor
Methylene bromide*
Methylene chloride
Methylnaphthalene, 2-
Methyl-tert-butyl ether
Methyl-2-Pentanone, 4-
Monochlorobiphenyl
Naphthalene
Nitroaniline, 2-
Nitroaniline, 3-
Nitroaniline, 4-
Nitrobenzene
Nitrophenol, 2-
Nitrophenol, 4-
N-Nitroso-di-n-butylamine
N-Nitroso-di-n-propylaniine
N-Nitrosodiphenylamine
Nonachlorobiphenyl
Octachlorobiphenyl
Penlachlorobenzene
Pentacbiorobiphenyl
Pentachloronitrobenzene
Pentachlorophenol
Phenanthrene
Phenol
Phosgene*
Propionaldehyde*
Compounds Detected in
Stack Emissions




Feb-94, Aug-94
Feb-94


Aug-94
Feb-94



















Compounds Analyzed for but Not
Detected in Stack Emissions
Feb-94

Aug-94

Mar-93

Aug-94
Feb-94, Aug-94

Aug-94
Feb-94
Feb-94
Feb-94
Feb-94, Aug-94
Feb-94
Feb-94, Aug-94
Mar-93
Feb-94
Feb-94
Aug-94
Aug-94

Aug-94

Feb-94, Aug-94
Feb-94
Feb-94, Aug-94


Volume III
Appendix III-3
-5-
External Review Draft
Do Not Cite or Quote

-------
APPENDIX m-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Pyrene
Quinoline*
Quinone*
Safrole
Styrene
Tetrachlorobenzene, 1, 2, 4, 5-*
Tetrachlorobiphenyl
Tetrachloroethane, 1,1,1,2-
Tetrachloroethane, 1,1,2,2-
Tetrachloroethene
Tetrachlorophenol, 2,3,4,6-
Toluene
Toluidine, o-*
Toluidine, p-*
Trichloro-l,2,2-trifluoroethane, 1,1,2-
Trichlorobenzene, 1,2,4-
Trichlorobiphenyl
Trichloroethane, 1,1,1-
Trichloroethane, 1,1,2-
Trichloroethene
Trichlorofluoromethane
Trichlorophenol, 2,4,5-
Trichlorophenol, 2,4,6-
Trichloropropane, 1,2,3-*
Vinyl acetate
Vinyl chloride
Xylene, m-/p-
II Xylene, o-
Componnds Detected in
Stack Emissions




Feb-94, Aug-94




Mar-93, Feb-94, Aug-94

Feb-94, Aug-94




Aug-94
Feb-94

Aug-94
Feb-94





Feb-94, Aug-94
Feb-94
Compounds Analyzed for but Not
Detected in Stack Emissions
Feb-94, Aug-94





Aug-94
Aug-94
Feb-94, Aug-94






Aug-94

Aug-94
Feb-94, Aug-94

Aug-94
Feb-94, Aug-94
Mar-93, Feb-94, Aug-94

Feb-94
Feb-94, Aug-94

Aug-94
Volume III
Appendix III-3
-6-
External Review Draft
Do Not Cite or Quote

-------
                                     APPENDIX m-3
                     Products of Incomplete Combustion Analyzed for
                  and Detected in the Trial Burns and Performance Tests
                                        (continued)
  Notes:
      Mar 93 - March 1993 trial burn
      Feb 94 -  February 1994 trial burn
      Aug 94 - August 1994 PIC testing

      These compounds were not listed in the waste profile sheets or analyzed for in the stack emissions.  Thus,
      incinerator stack emission rates are not developed for these compounds.
01-3999O:PCCOOB56.W51
Volume III
Appendix III-3
-7-
External Review Draft
Do Not Cite or Quote

-------