United States                                                    EPA-905-R97-002c
Environmental Protection Agency                                      May 1997

 WASTE MANAGEMENT
   Risk Assessment for the Waste Technologies Industries (WTI)
    Hazardous Waste Incineration Facility (East Liverpool, Ohio)
VOLUME III:
Characterization of the Nature and Magnitude of Emissions
                   U.S. Environmental Protection Agency - Region 5
                       Waste, Pesticides and Toxics Division
                             77 West Jackson Blvd.
                              Chicago, IL 60604
                           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

-------
                                  VOLUME III

                          CHARACTERIZATION OF THE
                    NATURE AND MAGNITUDE OF EMISSIONS
                                 CONTENTS

                                                                             Page

      I.  INTRODUCTION	  I-1

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

     H.  DATA USED IN CHARACTERIZING EMISSIONS	U-1

         A.  Waste Profile Data	D-1
         B.  Overview of Stack Testing at the WTI Facility	II-4
         C.  Key Assumptions for Emissions Characterizations	D-8

    in.  INCINERATOR STACK EMISSIONS 	ffl-1

         A.  Substances of Potential Concern in Stack Emissions  	HI-1
         B.  Development of Chemical-Specific Stack Emission Rates	ffl-1
             1.   Chlorinated Dioxins and Furans (PCDDs/PCDFs)	ffl-2
             2.   Other PICs and Organic Residues	HI-3
             3.   Metals	ffl-6
             4.   Acid Gases  	m-10
             5.   Particles	 ffl-11
         C.  Determination of Emissions Partitioning	IB-11
             1.   Partitioning of Emissions Between the Vapor
                and Particle Phases	HI-12
             2.   Distribution of Constituents Emitted on Particles  	BQ-13
         D.  Key Assumptions for Incinerator Stack Emissions	IE-14

    IV.  FUGITIVE EMISSIONS	IV-1

         A.  Potential Emission Sources	FV-1
         B.  Substances of Potential Concern in Fugitive Emissions 	FV-2
             1.   Substances of Potential Concern in Fugitive Organic
                Vapor Emissions	IV-2
             2.   Substances of Potential Concern in Fugitive Ash Emissions	IV-3
Volume IB                              -i-

-------
                                    CONTENTS
                                       (continued)
                                                                                   Page

          C.  Development of Fugitive Emission Rates	IV-3
              1.   Tank-Related Emissions from the CAB System	IV-3
              2.   Other Organic Fugitive Emissions	IV-4
              3.   Fugitive Ash Handling Emissions  	IV-5
          D.  Key Assumptions for Fugitive Emissions	IV-6

      V.   UNCERTAINTY IN EMISSIONS CHARACTERIZATION  	  V-1

          A.  Uncertainties in Stack Emissions Characterization	  V-1
              1.  Uncertainty Associated with Metal Emissions	  V-2
                 a)  Waste Composition and Feed Rate	  V-4
                 b)  Conditions in Incinerator System  	  V-4
                 c)  Control Device Efficiency	  V-6
                 d)  Results of Sensitivity Analysis 	  V-6
             2.  Uncertainties Due to Uncharacterized Stack Emissions	  V-8
             3.  Uncertainties Associated with Chromium  	 V-9
             4.  Uncertainties Associated with Laboratory Contamination	 V-9
         B.  Uncertainties Introduced by Process Upset Emissions  	  V-10
         C.  Uncertainties in Fugitive Emissions Characterization	  V-14

    VI.  REFERENCES 	VI-1


                                     TABLES

Table II-1:    Average System Removal Efficiencies (SREs) Measured
             at the WTI Facility	E-9
Table H-2:    Chlorinated Dioxin/Furan Stack Emissions at WTI Facility	11-10
Table II-3:    Key Assumptions for Chapter H  	II-1 1
Table ffl-1:   Substances of Potential Concern in Stack Emissions  	HI-15
Table EQ-2:   Estimated Average and High-end Stack Emission Rates for
             Dioxin and Furan Congeners	ffl-17
Table IH-3:   Estimated Average and High-end Emission Rates for Products of
             Incomplete Combustion (PICs) and Residues of Organic Compounds	ITJ-18
Table ffl-4:   Compounds Anticipated to be Emitted in Very Low Quantities
             for which Emission Rates are Not Developed	111-25
Table ffl-5:   Estimated Average Metal Emission Rates	111-26
Table HI-6:   Estimated Average Acid Gas and Particle Emission R'ates	ITI-27
Table III-7:   Key Assumptions for Chapter IE	HI-28
Volume TTT
                                         -11-

-------
                                    CONTENTS
                                      (continued)
                                                                                  Page
 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	TV-9
 Table V-1:    Observed Variation in the Control Efficiency of Selected
              Metals During the May 1993 Trial Burn	  V-16
 Table V-2:    Possible Variation in Predicted Metals Emissions Due to
              Uncertainty in Input Data	  V-17
                                    FIGURES

Figure II-1:   Total Chlorinated Dioxin/Furan Emissions vs. Chlorine Feed Rate  	11-12
Figure H-2:   TEQ Emissions vs. Chlorine Feed Rate	n-12
Figure HI-1:   Pathways Available for Toxic Metals in the WTI Incinerator	01-32
Figure IV-1:   Locations of Stack, Fugitive Organic Vapor & Ash
             Emission Sources	IV-10
                                 APPENDICES
Appendix ffl-l:
Appendix Hl-2:
Appendix HI-3:
Emissions Estimation Methodology and Background
Summary of Measured Dioxin/Furan Congener Emission Rates
Products of Incomplete Combustion Analyzed for and Detected in the Trial
Burn and Performance Tests
Volume
                                         -m-

-------
                                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 in 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 1H-1.  The characterization of potential emissions
during on-site and off-site accidents is described  in Volume VH.
    Normal  combustion processes in hazardous waste incinerators will  result in the release of
stack gas emissions into the atmosphere.  These emissions will cor^ist primarily of combustion
gases,  such as carbon dioxide (CO2), carbon monoxide (CO),  water, nitrogen, and oxygen.
However, despite the high temperatures typical of hazardous waste incinerators, a fraction of
the organic compounds in 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 in the stack emissions.
    In the risk assessment, substances of potential concern in the stack  gases are classified as
follows:
                                                               f*
    •   Polychlprinated dioxins and furans  (PCDDs/pCDFsX which are believed to be a
        product of incomplete combustion of some types of hazardous  waste;

Volume m                                1-1

-------
         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 in the waste but cannot be destroyed by combustion:

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

     •    Particles,  which may be entrained in the stack gas during waste combustion, or
         formed as stack gases cool in 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 (DRE) (based on testing at the WTI facility) to
estimate emissions  for chemicals not analyzed for during the stack testing.  Panicle 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 in the
         organic waste tank farm building and from the container processing area;
     2)   Seals, valves, and flanges associated with tanks and piping in the organic waste tank
         farm building that are vented from the building;

Volume III                                 1-2

-------
     3)  Open Wastewaier 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 reviev,
 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 in
 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 their relative
 contributions to total facility emissions.

B.  Peer Review Pane) Comments
    As part of the review process for the WTI 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 (USEPA  1993c) 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)
Volume HI                                1-3

-------
         be applied to estimate emission rates for individual chemicals in the waste feed that
         were not measured directly in the stack testing.

     •    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 in which these comments are addressed is described in the following sections
of this volume.
Volume HI                                 1-4

-------
                                II. DATA USED IN
                       CHARACTERIZING EMISSIONS
A.  Waste Profile Data
    Wastes arrive ai 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 in 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 in the waste profile sheets
provided by WTI (see Chapter n 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).                                ''"
Volume IE                                H-l

-------
 Prior to approving a waste for acceptance. WTI reviews the data in 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 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 in 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 WTI facility are  also compared against the projections contained in die
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 in 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 in 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

 Volume ffl                                 II-2

-------
 100 percent, keeping the relative amounts of each constituent the same as in the waste profile
 description.
     As discussed in detail in Chapter II of Appendix III-l. a number of refinements ha\e been
 made in the data base prior to estimating concentrations of specific chemicals in the facilit}
 waste streams.  These refinements include:

     •    Eliminating chemicals that cannot be adequately characterized, including a compound
         listed as "2,3-dibromophosphate"  and a compound listed under a confidential trade
         name that cannot 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 cannot 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 cannot 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 cannot be
        adequately characterized for the purposes of this risk assessment.

In addition, identical constituents reponed under different names have been consolidated and
reported under one listing, and isomeric compounds (e.g.,  ortho-,  meta- and para- isomers of
xylene) have been summed and reponed under a single listing.  Although some waste streams
have been eliminated from consideration, as noted earlier,  the remaining data base of waste
streams has been 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 the analytical results for the  limited set of metals, anions
and other analytes in the waste profile sheets. For example, each waste profile sheet contains
analytical results for chloride, as well as concentration ranges for all chemicals in the waste
stream that contain chlorine.  After the quantities of all chemicals containing chlorine are
estimated using the ranges reponed by the generators, the corresponding amount of chlorine is
Volume III                                II-3

-------
 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:
                  mollyr of Cl,  analytical results
                  mollyr 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 in the same manner.
    Finally, based on physical state descriptions provided by the generators in 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 WTI 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 in March 1993 (ENSR 1993) and February 1994 (ENSR
        1994a); and

    •   Incinerator performance tests conducted in 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

 Volume III                                II-4

-------
 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.
     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-trichIorobenzene, and carbon tetrachloride.  The DRE is
defined as:
                       Emission Rate of POHC
          DRE  -  1  -
                         feed Rate of POHC
Volume m                                115

-------
     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
 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 in 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:
                       Emission Rate  of Metal
            SRE. =  1 - 	'-                                   (II_3)
                         Feed Rate  of Metal.                                     v
The average SREs measured in the trial burn are presented in Table D-l.
    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 in a very dilute aqueous stream fed
into the SCC.  U.S.  EPA subsequently imposed restrictions to ensure that the incineration
system would be operated in 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
   1    Stack sampling was not performed for the three other regulated metals — barium, silver, and thallium
        based on WTl's agreement to accept permit conditions for these metals assuming no removal by the
        incineration or emission control system.

Volume ffl                                II-6

-------
 removal mechanism for organic substances through adsorption onto carbon particles  The
 airborne carbon panicles are subsequently removed from the flue gas with the fly ash in 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 quarter!) stack testing
 for PCDDs/PCDFs and panicles 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 WT1 comph
 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 in 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 gain 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 in 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 in 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 in 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 in this trend (r = 0.084
 for Total Emissions and r = 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 in the August 1994 PIC
testing are indicated in Appendix III-3).  The December 1994 performance test analyzed for
                Average over the five runs of the ECIS performance test.

Volume III                               II-7

-------
 the same PICs, plus decachlorobiphenyl, which had not been analyzed for in 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 semi volatile PICs.
 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 was not
 evaluated along with the earlier .test results.3  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 n-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.
              The results of the December 1994 PIC testing were not received in 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 tests is available from
              Region 5.

Volume HI                                H-8

-------
TABLE II-l
Average System Removal Efficiencies (SREs) Measured at the WTl 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
II-9

-------
TABLE II-2



Date
Pre-ECIS"
March 1993


Post-ECIS"
August 1993




February 1994




February 1994



April 1994




August 1994






Notes:
Chlorinated Dioxin/Furan Stack Emissions at WTI Facility''


Type of Testing

Trial Burn



Performance Test




Performance Test




Trial Burn
He-run of
Condition 2)'

'erformance Test




'erformance Test







Tests
Performed

Condition 1
Condition 2
Condition 3

Run 1
Run 2
Run 3
Run 4
Run 5
Runl
Run 2
Run 3
lun 4
Run 5
Run 1
Run 2
Run 3
tun 4
Run 1
Run 2
Run 3
Run 4
Run 5
Run 1
Run 2
Run 3
Run 4
Run5
Uin 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
+ 14
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
(ng/dscm)c

2.45
0.82
1.08

0.10
0.12
0.27
0.11
0.11
0.045
0.019
0.028
0.029
0.035
0.084
0.072
0.057
0.056
0.037
0.036
0.035
0.032
0.033
0.017
0.010
0.016
0.017
0.018
0.021
0.021
TEQ
Emission Rate
(g/sec)

6 4E-08
2.0E-08
3.0E-08

1.7E-09
2.1E-09
3.8E-09
2.1E-09
2.2E-09
1.2E-09
5.5E-10
7.8E-10
8.3E-10
9.6E-10
1.2E-09
l.OE-09
8.3E-10
8.5E-10
4.9E-10
4.6E-10
4.3E-10
4.5E-10
3.4E-10
2.6E-10
1.5E-10
2.3E-10
2.4E-10
2.4E-10
2.8E-10
3.0E-10

1 All emissions testing was performed with a kiln temperature in the range 2, 150-2, 200° F.
s Measurements were taken of all dioxin and furan congeners containing four to eight chlorine s3ubstituents with
chlorines in the 2-, 3-, 7-, and 8- positions; the reported value represents the sum of these congeners.
c The terra- through octa- chlorinated dioxin and furan congeners are expressed on the basis of toxiciry equivalents
(TEQs); the reported value is derived by multiplying the measured emissions of each congener by its toxic
equivalency factor (TEF) and then summing the results.
d An enhanced carbon injection system (ECIS) was added to the incinerator after the results of the trial burn (Pre-
ECIS) demonstrated that dioxin and furan emissions levels were higher than expected. All subsequent operations
and emissions testing occurred with the ECIS in place (Post-ECIS).
' WTI did not achieve a 99.99% DRE in Condition 2 of the trial burn and therefore repeated this Condition in
February 1994.
Volume III
H-10

-------
TABLE II-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 hound of the reported range for each waste stream
constituent is used. To prevent ihe combined percentage from
exceeding 100%, the constituent contents of the waste streams
are normalized.
The waste feed data are based on waste profile shee's 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.
x
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
Vnlnmp  III
H-ll

-------
FIGURE D-l:
40
i 35
f ? 30
"•* &
li 25
If 20
s e
J5 15
5 J I0.
H 5 -
0 -
(
Total Chlorinated Dioxin/Furan Emissions vs. Chlorine Feed Rate
•



* „ •* '.*••*** *••"





) 500 1000 1500 2000 2500 3000 3500
Chlorine Feed Rate (Ib/br)*

FIGURE D-2:
0.3 •
L 0.25-
i
it 0-2 "
e
IB-'
S 0.15 -
o
•9
«9
E 0.1 -
a
« 0.05 -
0 -
0
TEQ Emissions vs. Chlorine Feed Rate
•


* •
•
•• . '•* **
t % ••'•





500 1000 1500 2000 2500 3000 3500
Chlorine Feed Rate Qb/hr)*
 * Feed rate for chlorine in waste feed to incinerator (Ib/hr).
 ** Polychlohnated dioxin/furan emissions expressed as T ixicity Equivalents (TEQs), in nanograms
 per dry standard cubic meter (ng/dscm). Results are from the 26 post-EQS runs.'
Volume
D-12

-------
                  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 panel (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 in 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 being 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 cannot be used.  Thus, certain tables in 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.
Volume III                               III-l

-------
     •    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 WTI's operation, a statistical approach
         cannot 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 being '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 panicles.

    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 in 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 pan 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 the
    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 in April 1994 ranged from

Volume  TTT                               III-2

-------
     0.34 to 0.49 ng/sec, with an average of 0.43 ng/sec. and the total TEQ emission rates 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 in the risk assessment are based on emission
     data from the post-ECIS installation tests taken through August 1994.  These data are
     summarized in Table ffl-2., Specifically, average emission rates for the 17 PCDD/PCDF
     congeners (listed in Table ffl-2) are calculated as the arithmetic mean of the emission rates
     measured in 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)4 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 in that run. The estimated
    average and high-end emission rates for the 17 dioxin and furan congeners are listed in
    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 in August and December 1994.  As previously indicated, the results of the
                The 95% UCL is defined as:

                      95% UCL = mean  -
        where:       t    =   Studem-t statistic at 0.05 level of significance with n-1 degrees of freedom
                    s    =   sample standard deviation
                    n    =   number of samples

Volume m                               ffl-3

-------
    December 1994 performance tests were not available in time for inclusion in the risk
    assessment.5  Appendix III-3 summarizes which of the 174 organic residues and PICs,
    listed in Section in. A above, were analyzed for during trial burns and performance tests,
    and which of these compounds were detected.  Table HI-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 calibrated 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 in all seven runs. Compounds that were analyzed
    for but not detected are assumed to exist at one-half the detection limit in 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.
       5      The results of the December 1994 PIC testing were not received in timej^o 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 tests is available from
             Region 5.

Volume HI                                III-4

-------
         A number of the organic compounds listed in Table III-l were not analyzed for dunnc
     the August 1994 testing.  For these compounds, emissions are estimated based on the
     following techniques:

         (1) Use of measurements from the trial burns - During the March 1993 and Februan
             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) was 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 tne
            following calculation:

            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).
Volume III                               III-5

-------
         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 approache.s listed above
     is selected.  Using this procedure, only a single emission rate is estimated for each
     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 in the waste profile sheets or analyzed for in stack emissions ai
     WTI. Thus, emission rates cannot 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 in 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 congeners.  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 WTJ incinerator are predominantly organic in
    nature, inorganic substances, such as metals, are expected to be present in many waste
    streams. Metals that are present  in the incinerator feed may evaporate at the high
    temperatures in the rotary kiln and, subsequently, condense to form aerosols of submicron
    panicles in the cooler, later stages of the incineration process. Although most of the
    metals would be captured in 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.
Volume III                                III-6

-------
        The general equation being used to calculate metal emission rates for the incinerator
     stack is the following:

                    E. = (l-SRE) x Ft                                          (III-2)
    where:
        Ej      =  annual average stack emission rate for metal i, Ib/yr
        F,      =  annual feted 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 burn 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 in the March 1993 trial burns, SRE values  are
    extrapolated from the trial burn data for the metals that were tested, considering the
    results of thennodynamic modeling, as described below.
        The behavior of metals in 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 in
    Figure ffl-l, metals present in the waste feed may first volatilize, become  entrained as
    particles in 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 in 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 in the
    Chromium VI oxidation state (the most toxic form of chromium). Chromium speciation is
    further discussed in 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

Volume III                               ffl-7

-------
     entrained ash.  Investigations of combustion systems have found that any panicles smaller
     than about 0.1 ^m quickly coagulate, while those larger than 1 ^m do not (Linak and
     Wendt 1993). Thus, two groups of panicles typically enter the air cleaning system: one
     group ranges in size from 0.1 to 1 /^m and is formed from the metals that vaporized and
     subsequently condensed (fine panicles); the second group is generally in the range of 1 to
     10 ^m in diameter and consists of the material entrained in the incinerator (coarse
     panicles).
        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 panicle  size.
     A detailed description of the methods being used to model metal behavior in the APC
     system is contained in Appendix ffl-1.  The primary assumptions and determinations made
     in conducting the analyses include the following:

        •    The compounds PbCl4 and CrO2Cl2 are commonly predicted to form at low
            temperatures by the thermodynamic programs, but are not typically found in
            combustion gases (Linak and Wendt 1993).  Thus, these two compounds have
            been excluded from consideration during the thermodynamic 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 panicles are entrained.  Thus, it is assumed that all'Of the ash in the
            pumpable wastes is entrained.
Volume HI                               m-8

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

         Based on the modeling, it has been determined that metals in the incineration system
     can be classified into one of four categories depending on their expected behavior in the
     incinerator system: 1) insoluble vapor; 2) soluble vapor; 3)  fine panicles (< 1  ^m): and 4}
     coarse particles (1-10 pm).  Soluble and insoluble vapors are created by metals that
     vaporize in the incinerator, but do not condense in the quench system.  Fine panicles
     vaporize in the incinerator but subsequently condense in the quench system.  Coarse
     panicles are largely created by atomization of the waste and do not originate as vapors.
     The metals evaluated in 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 in the trial burn
    are shown in Table ffl-5.  Based on the modeling, chromium is used to estimate die 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 in 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, SO, 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 in the APC. This is based on the very low SRE (< 10%) measured
    for mercury in 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 in Morristown, NJ, observed a mercury SRE increase from an average of =
Volume III
m-9

-------
    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 in the risk assessment
    that the ECIS would not enhance mercury removal in the APC system.
        Waste feed data for the 15 metals of potential concern have been developed based on
    waste profile sheets and feed rates provided by WTI for the first year of operation at the
    facility, as discussed in Chapter n 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 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 maximum predicted metals feed rates, which are listed in Table III-5.
    The maximum 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 HI-5.
       Emission rates of metals  are strongly influenced by the metal composition in the feed
    to the incinerator, and the variability in 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 estimates
    of metal emission rales are developed and are used to represent average metal emission
    rates).

    4.  Acid Gases
       A variety of acid gases, including hydrogen chloride  (HCI), nitrogen oxides (NO,),
    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 HCI during the March 1993 trial burn (nine runs) and the
    February 1994 trial burn (four runs).  The average emission rate of HCI 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  HCI emission
    rate). This emission rate is considerably below the regulatory limit of 4 Ib/hr (40 CFR
Volume III                               HMO

-------
V
                   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 rareh
                   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 NO, and SO, and makes the results
                   available on an electronic bulletin board. Average emission rates of NOX and SO, are
                   estimated based on the mean of one month of recent continuous monitoring data (Februarx
                   23,  1995 to March 21, 1995).
                       Average emission rates of the three acid gases being considered in this  assessment are
                   presented in Table III-6.

                   5.   Particles
                       Particle emission rates were measured by WTI during the March 1993 trial burn (nine
                   runs), February 1994 trial burn (four runs), and four performance tests (22  runs
                   combined).6  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 in Appendix ffl-l, almost all the particles are less than 10 ^m in diameter.  The
                  estimated average emission rate of particles is shown in Table III-6.

              C.  Determination of Emissions Partitioning
                  Substances in the stack gas will generally be present in 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 in the vapor phase will
              either remain in the vapor phase or become adsorbed to particles that are present in the stack
              gases and the atmosphere.  This partitioning between phases is based on chemical-specific
              parameters,  and is an important factor in estimating the rate at which compounds deposit out
              of the atmosphere onto soil, surface water, and plants. Due to their physical/chemical
              properties, such as vapor pressure, acid gases remain in the vapor phase, metals (with the
              exception of mercury) remain entirely as particles, while most organics tend to partition
                              Panicle 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.
              Volume III                                III-11

-------
 between the particle and vapor phases. Some of the polycyclic aromatic hydrocarbons (PAHs)
 with very low vapor pressures (e.g., dibenzo(a.h)anthracene and indeno(1.2.3-cd)pyrene) are
 assumed to be entirely in the paniculate 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 in air 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:
                            c X
                        P   + (c  * ST)
    where:
        4>   ==   fraction of organic chemical adsorbed to particles, unitless
        ST-  =   panicle surface area per unit volume of air, cnr/cm3
        p°   =   vapor pressure, atm
        c   =   molecular weight and heat of condensation factor, atm-cm

        For particles, surface area per unit volume of air (Sj) 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 1O6 cnrVcm3 for average background; and 1.1 x 10'5 cnrVcm3 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  1O4 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
Volume III                              III-12

-------
     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:

                      P,    bs,(Tm-T)
                                RT
    where:
        PL   =  sub-cooled liquid vapor pressure, atm
        Ps   =  crystalline solid vapor pressure, atm
        ASf  =  entropy of fusion, atm-nrVmole-K
        Tm   =  melting point, K
        T    =  ambient air temperature, K
        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 in Volume V.

    2.  Distribution of Constituents Emitted on Particles
        Two separate approaches are used for organics and metals in 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 in 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 emined 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 in the air dispersion modeling, as described in Volume IV.
Volume IE                               IE-13

-------
D.  Key Assumptions for Incinerator Stack Emissions
    The key assumptions being used in predicting and analyzing incinerator stack emissions
for the WTI facility are summarized in Table ffl-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 in Chapter V of this report.
Volume HI                               HI-14

-------
                                                TABLE III-l
                        Substances of Potential Concern in Stack Emissions
                                    PICs and Residual Organic Compounds
   Acenaphthene
   Acenaphthylene
   Aceialdehyde
   Acetone
   Acetophenone
   Acrolem
   Acrylonitriie
   Anthracene
   Benzaldehyde
   Benzene
   Benzole acid
   Benzotrichloride
   Benzo(a)anthracene
   Benzo(a)pyrene
   Benzo(b)fiuoranihene
   Benzo(e)pyrene
   Benzo(g,h,i)perylene
   Benzo(j)fluoranthene
   Benzo(k)fluoranthene
   Benzyl chloride
   Biphenyl
   Bis(2-chloroethoxy) methane
   Bis(2-chloroethyl)ether
   Bis(2-chloroisopropyl)ether
   Bis(2-ethylhexyl)phthalate
   Bromochloromethane
   Bromodichloromethane
   Bromoethene
   Bromoform
   Bromomethane
   Bromodiphenylether, p-
   Butadiene, 1,3-
   Buianone, 2- (MEK)
   Butylbenzylphthalate
   Carbon disulfide
   Carbon tetrachloride
   Chlordane
   Chloro-3-methylphenol, 4-
   Chloroacetophenone, 2-
   Chloroaniline, p-
  Chlorobenzene
   Chlorobenzilate
  Chloroe thane
   Chloroform
  Chloromethane
  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)fiuoranthene
  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, IX-
  Dichlorobenzidine, 3,3'-
  Dichiorobiphenyl
 Dichlorodifluoromethane
 Dichloroe thane, 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-butylphthalaie
 Di-n-octyl phthalate
 Dinitriioluene, 2,6-
 Dinitro-2-methylphenol, 4,6-
 Dinitrobenzene, 1,2-
 Dimtrobenzene, 1,3-
 Dinitrobenzene, 1,4-
 Dinitrophenol, 2,4-
 Dinurotoluene, 2,4-
 Dioxane. 1.4-
  Ethyl methacn'late
  Ethylbenzene
  Ethylene dihronude
  Ethylene oxide
  Ethyiene thiourea
  Fluoranthene
  Fiuorene
 Formaldehyde
 Furfural
 Heptachlor
 Heptachlorobipheny 1
 Hexachlorobenzene
 Hexachlorobiphenyl
 Hexachlorobutadiene
 Hexachlorocyclohexane, alpha-
 Hexachlorocyclohexane, beta-
 Hexachlorocyclohexane, gamma-
  (a.k.a. Lindane)
 Hexachlorocyclopentadiene
 Hexachloroe thane
 Hexachlorophene
 Hexane, n-
 Hexanone, 2-
Hexanone, 3-
Indeno(l ,2,3-cd)pyrene
Isophorone
Maleic hydrazide
Methoxychlor
Methylene bromide
Methylene chloride
Meihylnaphthalene, 2-
Methyl-tert-buryl ether
Methyl-2-Pentanone, 4-
Monochlorobiphenyl
Naphthalene
Nitroanilme, 2-
Nuroaniline, 3-
Nitroaniline, 4-
Nitrobenzene
Nitrophenol, 2-
 Nitrophenol, 4-
 N -N itroso-di-n-buty lamine
 N-Nitroso-di-n-propylamine
 N-Nurosodipheny lamine
 Nonachlorobipheny)
 Octachlorobiphenv 1
Volume III
              III-15

-------
TABLE III-l (continued)
Substances of Potential Concern in Stack Emissions
Pentachlorobenzene
Pentachiorobipheny 1
Pentachloron j trobenzene
Pentachlorophenol
Phenanthrene
Phenol
Phosgene
Propionaldehyde
Pyrene
Quinoiine
Qumone
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
Sryrene
Tetrachlorobenzene, 1.2.4.5-
Tetrachlorobiphenyl
Tetrachloroethane . 1,1.1.2-
Tetrachloroethane. 1.1,2.2-
Tetrachloroethene
Tetrachiorophenol. 2,3.4.6-
Toluene
Toluidme, o-
Toluidine, p-
. Trichloro-l,2.2-TFE, 1.1,2-
Tnchlorobenzene, 1,2,4-
Tnchlorobiphem 1
Tnchloroethane . 1 . 1 . i -
Tnchloroethane . 1.1.2-
Tnchloroethenc
Tnchlorofluoromethane
Tnchlorophenol, 2.4.5-
Trichlorophenol, 2.4.6-
Trichloropropane. 1.2.3-
V'myl acetate
Vinyl chloride
Xylene. m-
Xylene, o-
Xylene, p-
Furan Congeners
2,3,7, 8-TetraCDF
1,2,3, 7.8- PemaCDF
2,3.4.7,8-PentaCDF
1,2,3,4,7.8-HexaCDF
1,2,3,6,7,8-HexaCDF

1.2,3.7.8,9-HexaCDF
2. 3.4,6.7. 8-HexaCDF
1.2.3.4.6,7.8-HeptaCDF
123478 9-HeptaCDF
OctaCDF
Metals
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium (hexavalent and
trivalent)
Copper
Lead
Mercury (inorganic and organic)
Nickel
Selenium
Silver
Thallium
Zinc
Acid Gases
Hydrogen chloride
Total nitrogen oxides (NO,)
Total sulfur oxides (SO,)
Participate Matter
Respirable (PM10)
Total
Notes:
TFE - trifluoroeihane MEK - methyl ethyl ketone
CDD - chlorodibenzo-p-dioxin PM10 - paniculate matter < 10 microns
CDF - chlorodibenzofuran "*"'
Volume III
III-16

-------
TABLE III-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
l.OSxlO-11
6.78 x 10-n
8.95 x 10-n
1.66x10-'°
1.09x10-'°
1.24xlO-9
6.15x 10-9
2.16x 10-"
9.46x10-"
1.25 x 10-'°
2.18x lO'10
1.55x 10-10
1.69x 10-9
9.80 x 10-9
Furan Congeners
2,3J,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-"
3.45 x 10-10
4.67 x lO'10
1.43xlO-9
1.33x 10-'
l.SOxlO-9
2.93 x 10-'°
9.30 x lO"9
1.22 x ID'9
1.89X10-8
1.15x10-'°
4.35 x 10-'°
6.04 x 10-'°
1.85x 10-9
1.71 x 1C'9
1.96xlO'9
3.85 x 10-10
l.SOxlO'8
l.SOxlO'9
3.62 x lO'8
Notes:
CDD - chlorodibenzo-p-dioxin
CDF - chlorodibenzo-p-furan
Volume
m-17

-------
TABLE III-3
Estimated Average 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
Benzotrichloiide
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)
Average
6.69 x 1CT6
6.69 x 10-6 .
3.01 x 10*
2.90 x lO'3
2.93 x 10-4
2.02 x 10^
5.50 x 10-*
1.47 x 10-5
1.13 xlO"5
3.20 x 1C'5
5.50 x 10*
5.50 x 10-6
5.50 xlO-6
5.50 x 10-6
5.50 x 10-6
6.69 x lO"6
1.33 x 10-5
6.69 x 10-6
3.72 x 10-5
1.03x 10^
5.50 x 10-*
4.90 x 1O4
6.69 x 10*
High-end
6.69 x 10-6
6.69 x 10*
3.01 x ICT1
2.90 x 10°
2.93 x 10-4
2.02 x 1O4
l.lOx 10-5
2.63 x lO'5
1.13 xlO-5
3. 20 xlO'5
l.lOxlO-5
l.lOx 10-5
l.lOx lO'5
l.lOx 10-5
l.lOx 10-5
6.69 x 10*
1.33 xlO-5
6.69 x 10-6
5.23 x lO'5
1.53 x 10"
l.lOxlO"5
9.80 x 10"
6.69 x 10-6
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 ffl
III-18

-------
TABLE III-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
5.14X10'5
5.50 x 10-6
8.91 x ID'5
1.58x10^
5.50 x lO'7
6.69 x IQr6
6.69 x 10"6
5.50 x 10-*
3.68 x lO'5
4.90 x 1O4
2.66 x IQr4
2.45 x 10^
6.69 x 10-6
S.SOxlO-6
6.69 x 10-*
5.50 xlO-6
5.50 x 10-6
5.50 x 10-6
5.50x10^
1.39x10^
5.50x10^
3.88xlO-5
5.50 x lO'7
High-end
7.40 x 10-5
l.lOx 10'5
9.46 x 10'5
2.75 x 10-4
l.lOx 10-6
6.69 x 10^
6.69 x 10^
l.lOxlO-5
3.68 x lO'5
9.80 x 10-4
4.07 x 10^
4.90 x 10^
6.69 x 1O*
l.lOxlO-5
6.69 x 10^
l.lOx lO'5
l.lOxlO'5
l.lOxlO-5
l.lOxlO-5
1.39x10^
l.lOxlO-5
3.88 x lO'5
l.lOxlO-6

Source
b I
b 1
b
b
b J
a
a I
b I
a
b |
b I
b
a
b
II
b
b
b
b 1
a
b
a
b
Volume HI
ra-i9

-------
TABLE III-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-
IDichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine, 3,3'-
DichlorobiphenyJ
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
5.50x10-*
2.63 x 105
5.50 x 10-*
5.50x10-*
5.50 x 10"6
3.33 x 10-5
4.68 x 10-8
2.45 x KT*
1.25xlO-5
1.25x 10-5
1.25xlO'5
1.25 x 10-5
5.50 x 10-6
1.25x 10-5
1.25 x 10-5
1.25x lO'5
1.69xlO-5
1.15x 1O4
5.50x10-*
5.50 r. 10-6
1.57xlO-5
5.50 x 10-6
High-end
l.lOx 10-s
l.lOx 10-5
2.63 x 10-5
l.lOx 10-5
l.lOx lO'5
l.lOx 10-5
3.33 x lO'5
8.22 x lO'8
4.90 x 10^
2.50 x 10'5
2.50 x 10-5
2.50 x lO'5
2.50 x lO'5
l.lOx lO'5
2.50 x 10-5
2.50 x 10-5
2.50 x lO"5
3.60 x lO'5
1.15x 104
LlOxlO-5
LlOxlO'5
2.04 x 10-5
l.lOx 10-5

Source
b
b
a
b
b
b
a
b
b
b
b
b
b
b
b
b
b
b
a
b 1
b I
b
b 1
Volume HI
in-2o

-------
TABLE III-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
5.50x10^
5.50 x 10-6
4.94 x 10"
2.45 x 10"
4.98 x 10"
1.15x 10"
3.05 x 10-5
1.46x10-'°
5.50x10^
6.69 x 10-6
6.07 x 10"
5.50 x 10^
5.50 x 10-7
1.40xlO-8
5.50 x 10-6
1.40xlO'8
1.01 xlO"
5.48 x lO'5
5.50 x KT6
5.50 x 10"6
High-end
l.lOx lO'5
l.lOx lO'5
l.lOx 10'5
l.lOxlO5
4.94 x 10"
4.90 x 10"
7.53 x 10"
1.15x10"
3.05 x lO'5
1.46x lO'10
LlOxlO'5
6.69 x 1O*
6.07 x 10"
l.lOx 10-5
l.lOx 10"6
2.80 x 10 "
l.lOx 10'5
2.80 x Kr8
1.01 x 10"
5.48 x 10's
l.lOxlO'5
l.lOx 10's

Source
b 1
b ||
b 1
b 1
a ))
b I
b 1
a jj
a 1
a
b
a
a
b
b
b
b
b
a
a
b
" 1
Volume
HI-21

-------
TABLE III-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
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-propylamine
N-Nitrosodiphenylamine
Nonachlorobiphenyl
OctachlorobiDhenvl
Emission Rate (g/sec)
Average
3.20x 10-5
6.43 x ID'5
5.50x10*
6.69 x 10*
1.15x10^
5.50 x ID'7
3.96 x 10^
4.18xlO-5
1.25xlO-s
1.25X10'5
1.67x 10-8
5.50x 10*
6.69 x 10-*
6.69 x 10*
6.69 x 10*
5.50x 10*
6.69 x It)"6
5.50 x 10*
1.21 x 1O4
6.69 x 10*
6.69 x 10*
1.40x lO'8
1.40x lO'8
High-end
3.20 x lO'5
6.43 x lO'5
l.lOx lO'5
6.69 x 10*
1.15x 10"4
l.lOx lO'6
6.19x 10^
4.18xlO'5
2.50 x 10-5
2.50 x lO'5
2.99 xlO"8
l.lOx lO'5
6.69 x Kr6
6.69 x 10*
6.69 x 10*
l.lOxlO'5
6.69 x 10*
l.lOxlO-5
1.21 x KT*
6.69 x 10*
6.69 x 10*
2.80 x 10-*
2.80 x 10-8

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 HI
HI-22

-------
TABLE III-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
Tetrachiorobiphenyl
Tetrachloroethane, 1,1,1,2-
Tetrachloroethane, 1,1,2,2-
Tetrachloroethene
Tetrachlorophenol, 2,3,4,6-
Toluene
Trichloro-1 ,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"8
3.37 x lO'5
5.50 x 10-6
6.69 x 10"6
5.50x10-*
5.50x10-*
1.15x10"
2.25 x 10-5
1.40X10-8
5.50x10-*
5.50 x 10-6
5.13X10'5
6.80 x 10-*
6.13x10"
3.30x10"
5.50 x 10"6
3.02 x 10-8
1.25xlO-5
1.25xlO-5
1.86xlO-5
2.45 x 10"
5.50x10-*
High-end
4.76 x ID'5
2.80 x lO'8
3.37 x 10'5
l.lOx ID'5
6.69 x 10-6
l.lOx 10'5
l.lOx lO'5
1.15x10"
4.04 x lO'5
2.80 x 10-8
l.lOxlO-5
l.lOxlO'5
8.02 x 10-5
6.80 x 1O*
1.03 x 10°
3.30 x 1O4
l.lOx lO'5
5.80 x 10-8
2.50 x lO"5
2.50 x 10-s
3.09 x lO'5
4.90 x ID"4
l.lOx lO'5

Source
a ||
b |
a ||
b I
a ||
b 1
b ||
a ||
b |
b ||
b I
b ||
b 1
a
b
a
b I
b
b
b It
b I
b I
b 1
Volume
ifl-23

-------
TABLE III-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-c
Xylene, o-
Xylene, p-c
Emission Rate (g/sec)
Average
5.50x 10-6
6.43 xlO-5.
2.45 x 10"
3.80 x 10"
5.50 xlO-6
3.80x 10"
High-end
l.lOx ID'5
6.43 x lO'5
4.90 x 10"
5.64 x 10"
l.lOx 10-5
5.64 x 10"
Source
b
a
b
b
b
b
Notes:
       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.
       Emission rate based on August 1994 PIC testing results.
       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
HI-24

-------
                                      TABLE III-4
     Compounds Anticipated to be Emitted in Very Low Quantities for which Emission
                                Rates Are Not Developed	
                   Acrolein
                Benzaldehyde
               Benzo(e)pyrene
             Benzo(j)fluoranthene
               Benzyl chloride
                  Biphenyl
             Bromochloromethane
                Bromoethene-
               Butadiene, 1,3-
           Chloroacetophenone, 2-
              ChJoropropane, 2-
        Dibromo-3-chloropropane, 1,2-
          DicUoro-2-butene, cis-1,4-
         Dichloro-2 -butene, trans-1,4-
           Dichlorofluoromethane
             Dinitrobenzene, 1,2-
                Dinitrobenzene. 1.3-
                Dinitrobenzene. 1.4-
           Hexachlorocyclohexane. alpha -
           Hexachlorocyclohexane, beta-
                    Hexane. n-
                  Hexanone. 3-
                Methylene bromide
                    Phosgene
                 Propionaldehyde
                    Quinoline
                    Quinone
           Tetrachlorobenzene, 1,2.4.5-
                  Toluidine, o-
                  Toluidine, p-
             Tr ichloropropane, 1.2.3-
  Note:
  These compounds were on original list of possible organic constituents of concern, but not
  •eported in Waste Profiles or analyzed for in WTI stack emissions.	
Volume III
111-25

-------
TABLE III-5
Estimated Average Metal Emission Rates
Metal
Aluminum
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Measured SRE
(percent)
NA (99.99932a)
99.986
99.977
NA (99.977b)
99.99Q7
99.987
99.99932
NA (99.977b)
99.99
0°
NA (99.977b)
NA (99.68d)
NA (99.977")
NA (99.977b)
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-
4.2 x 10-°
3.7 x lO'5
1.5 x 10^
3.3 x 10-8
1.6x 10-5
7.1 x 10-7
9.4 x 10'5
4.3 x lO'5
1.4x 10-3
5. Ox 10'6
4.7 x 10-*
1.5 x 10-5
3.4 x 10'5
1.2x10^
Notes:
System removal efficiency (SRE) determined from March 1993 trial burn
(ENSR 1993)
NA: Not applicable; SRE not determined in March 1993 u.al 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.
" Estimated based on SO, SRE. ||
Volume
m-26

-------
                                    TABLE III-6
           Estimated Average Acid Gas and Particulate Matter Emission Rates
                 Substance
              Average Emission Rate
                     (g/sec)	
             Hydrogen Chloride
                      0.032
               Nitrogen Oxides
                      2.4
               Sulfur Dioxides
                     0.091
                  Particles
                      0.07
  Source:  WTI monitoring data; February 23, 1995 to March 21, 1995.
Volume IH
ffl-27

-------
TABLE III-7
Key Assumptions for Chapter III
Assumption
All stack chemicals of potential concern have been identified
and included
Emission rates are estimated based on performance tests and
(rial 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
of26post-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 III
                                                       111-28

-------
                                                          TABLE III-7 (continued)
                                                     Key Assumptions for Chapter III
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 ORE.  The feed
          rate is developed from waste profile sheets for
          the first year of operation and the ORE 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	         	
                                                                                 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.	
                                                          Magnitude
                                                           of Effect
 Metals emissions are estimated from trial burns, one year of
 waste feed data, and thermodynamic modeling
 Best available data. Professional judgment based on a review
 of information on facility design and operation, and predicted
 waste characteristics.
 The triakburn during which metal SREs were calculated was
 conducted prior to installation of the ECIS.  These SREs are
 used to estimate metal emission rates.
 The ECIS is not designed to appreciably reduce metal
 emissions, so SREs measured pre-ECIS should be similar to
 post-EClS.  An exception may be mercury, for which removal
 may be enhanced by the ECIS.	
Vr.lnrm> III
                                                                       HI-29

-------
TABLE III-7 (continued)
Key Assumptions For Chapter HI
Assumption
The thermodynamic modeling to describe metal behavior
contains several assumptions including:
PbC14 and CrO2CI2 are excluded
All metals that vaporize subsequently condense to
form particles with diameters that are O.Sfim
All pumpable waste is entrained and no ash from
non-pumpable waste is entrained
Little condensation occurs onto the surface of
larger entrained particles
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 metals feed rates are prorated to account for the
maximum heat input of the incinerator.
HCI emission rates are based on the average of 12 trial burn
runs
Metals (other than Hg) are emitted from the stack in particle
form
Basis
Professional judgment was relied upon based on a review of
information on facility design and operation, and predicted
waste characteristics.
The surrogates were selected based on the expected behavior
of the metals, which are a function of their physical/chemical
clnracteristics. 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.
Conservative assumption. Professional judgment was relied
upon based on a review of information on facility design and
operation, and predicted waste characteristics.
Trial burns were conducted at elevated total chlorine feed rates
producing conservative estimates of HCI emission rates.
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.
Magnitude
of Effect
low
low
low
low
low
Direction
of Effect
unknown
unknown
overestimate
overestimate
variable
Volume III
                                                       111-30

-------
TABLE III-7 (continued)
Key Assumptions for Chapter III
Assumption
The vapor/particle partitioning equation (Equation 111-3) 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 metais 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
high
low
medium
low
high
Direction
of Effect
variable II
variable
unknown
overestimate '
unknown
Volume III
                                                       111-31

-------
          _
  Melals In
  PumpaWe
  Wastes
  Melals In
Non-Pumpable
  Wastes
1
ROTARY KILN SCO
1
1

1
1
INITIAt '
QUENCHING!
ZONES |
!
1
1 „

APCE

Fine Fly Ash
(Condensed ^
Vapors) .,
_ _ • _J. _ _ • • * .^^Ix1
1
1
1
1 X
ly^
>X^

                            Vapors
                                                                                           STACK
                                                                                          >••« tlv II.,. ^ :„:„
 Volv   ill    PATHWAYS AVAILABLE  FOR TOXIC METALS IN THE WTI INCINERATOR

-------
                           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 in 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 have been identified based on information in the WT! 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 have been identified at the WTI
facility for evaluation in the risk assessment:

    •    The carbon adsorption bed (CAB) system, which receives organic vapors vented from
        tanks in the organic waste tank farm, from operations in 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 in 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

Volume m                                IV-1

-------
         C.  System "A" collects non-contaminated storm water from such areas as roots and
         the employee parking lot. and the water is discharged directly  to the Ohio River
         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-gaiion 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 reasonabh
         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 in 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 IV-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 tnfncated 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)

Volume III                               IV-2

-------
     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 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 in the production of solid residues
     (i.e., ash).  Fugitive panicle 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 (WTI1995).
        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
    cyanide6, 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 in at least one sample of
    ash are selected as substances of potential concern and are  identified in Table IV-1. None
    of the 12 samples included detectable levels of all 9 metals. Total cyanide was also
    detected in 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 in 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,
                Amenable cyanide is subject (or amenable) to chlorination, and is the most toxic form of cyanide.

Volume HI                               IV-3

-------
    vapor pressure over a range of temperatures, and concentration in the waste, in deriving
    emission rates. In estimating the overall composite physical/chemical properties 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 Bru/lir
    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 ffl-1). Emissions are assumed to be
    equivalent *o 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 53 percent of
    the time.  It is assumed that the CAB system effectively controls 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 maybe 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

Volume in                               IV-4

-------
     in the waste feed accepted by the facility based on waste profile sheets maintained by
     WTI, and because its critical physical/chemical propenies are representative of the 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 in Table IV-2.  As shown in Table FV-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 in 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 WTI 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 that 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 consent (due to its temperature above the boiling
    point of water and the absence of moisture addition), the emissions factor is increased by a

Volume HI                               FV-5

-------
    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 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 IV-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.
Volume III                                IV-6

-------
                                               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-
 Dichlorobenzene
 Dichlorodifluoroethane
 Dichlorodifluoromethane
 Dichloroethane, 1,1-
 Dichloroethene
 Diethyl stilbestrol
 DiethylphthaJate
 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( 1,2,3-cd)pyrene
Isobutanol
Isopropanol
 Isosafrole
 Maleic anhydride
 Methanol
 Methyl methacrylate
 Methylbutadiene.  1-
 Methylcholanthrene. 3-
 Methyl isoburyl 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
Trichlorofluoromethane
 Xylene
                                          Fugitive Ash Emissions'*
               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 III-l.
  b     Based on metals detected in fly ash samples collected by WTI (1995).
  TFE -     trifluoroethane
Volume
              IV-7

-------
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 AH Sources:
Estimated Emissions
Ib/yr (g/sec)
2. 126 (3.06 x 10 :)
202 (2.91 x 103)
9.9 (1.42 x 10-)
224.3 (3.23 x 10°)
212.2 (3.06 x 10 })
57.2 (8.23 x 10")
57.2 (8.23 x 10")
47.7 (6.86 x 10-)
50.1 (7.21 x 10-")
12.1 (1.74x 10-)
2,562 (3.68 x 102)
Note:
1 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 in
IV-8

-------
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 fly ash 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
arroronriate 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 fly ash emissions but increased by a factor of 10
Basis

-------
                                                                                    Watt*  Water
                                                                                    Troatmont
                                                                                       Bldg.
                                                                                            Truck Unloading
                                                                                                Station
                                                                                        Organic Watt*
                                                                                        Tank Farm
              Existing
            Malntonaneo
               Bldg.
ruW^^v/B^A    -

  9'..-r"  ESP     \''''
,--'"  Stdek           x " >|n^
                                                            ln«tntrator
                                                              HVAC
t Tank  farm
2 Wattowaltr  tank
3 Truck wath
4 Carbon adtorptlon bed
5 A«h loading
     Volume III
NOTE: Not to tool*
                                  LOCATION  OF STACK,  FUGITIVE OBHANIC VAPOR, AND ASH  EMISSION  SOURCES
Bond  en VonRoll Drawing! o-OI-t-00001
                          ••IMaMHiiMi

                          Figure
                                                                                                                     IV

-------
       V. UNCERTAINTY IN EMISSIONS CHARACTERIZATION
    As previously discussed in this volume, an important initial step in 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 in the srack 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, have been
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 in 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 in both metal and organic emissions can occur as a result of changes in
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 in incinerator operating  conditions;
consequently, the following sections provide a summary of the sensitivity of metal emissions
to several incinerator parameters.
Volume m                                V-l

-------
    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 in 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 in 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 being used in 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 in
    waste  incineration systems.  However, several assumptions inherent in 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 in most cases, the
            needed reaction rates are not known. However, in 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 in 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 in 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 in the thermodynamic data base used in  the model.
            The potential effects of these complex reactions on predicted emission rates are
            unknown.

Volume m                               V-2

-------
         •   The reactor outlet temperature adequately characterizes the temperature to which
             the metals are exposed. The temperatures in the incinerator van' 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 in 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 sue-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 has been performed to evaluate the
    impact of variations in site-specific input parameters.  The following site-specific
    parameters are used in 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
Volume ffl                                 V-3

-------
        The effects of reasonable variations in waste feed rate, combustion chamber
    temperature, availability of oxygen, waste chlorine concentration, quench temperature and
    entrainment rates are examined in Appendix ffl-1.  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 in changes in emission rates of more than an
    order of magnitude, as discussed below.

        a)   Waste Composition and Feed Rate
            Three feed rates have been examined for each of the 15 metals in 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 in
       temperature. At the two  lower temperatures, no beryllium vaporizes in the
       combustion chamber. The emissions decrease 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 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.
                                                               f
            To assess the affect of the availability of oxygen on the predicted metals
        behavior, the effects of two  additional values of the primary chamber air to waste
Volume III

-------
                    stoichiometric ratio are examined. In general, varying the stoichiometry of a
                    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 stoichiometric 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 heal 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
 V_y              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
V                  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.

           Volume ffl                               V-5

-------
             The final parameters examined are the entramment rates. The following values
         are selected to represent reasonable extremes based on facility design and operation:

             •   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 bum.  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 in a series of tests repeated over a short time using
        a well controlled simulated waste, it is likely that the values in 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 in Table V-l and the sensitivity study, it is possible to
        determine the potential range of variation that may be expected in 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 in the control efficiencies. Variations in control efficiencies

Volume III                                V-6

-------
         are very small compared to the possible variations in 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 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 in emissions and predicting the phase of
        each metal emitted, as shown in Appendix ffl-1.  However, the results of this
        modeling exercise are not used in the risk assessment to predict specific metal species
        or phases given the high degree of uncertainty in the model'Tesults.
Volume III                                V-7

-------
    2.  Uncertainties Due to Uncharacterized Stack Emissions
        Slack 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 "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 in 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 ffl-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 ffl-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 dhd 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

Volume  III                               V-8

-------
     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 toxiciry 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 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 in 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; Giarn 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 in many laboratory blanks (Thuren 1986; U.S.

Volume IE                                V-9

-------
     EPA 1988b).  Consequently. U.S. EPA (1988b) has reported that common phthalate
     esters generally cannot be precisely measured at concentrations below 2 ppb due to blank
     contamination.

B.   Uncertainties Introduced by Process Upset Emissions
     As discussed in Appendix ffl-1. 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 VH.

    •   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 cannot 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

Volume IE                               V-10

-------
         collects most of the panicles). 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 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 the 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) at^mization
        tends to mechanically generate a considerable amount of paniculate matter, as well as
        generate more vapors which can subsequently condense/nucleate into paniculate
        matter.  In addition, atomization more vigorously disperses both of these types of
        paniculate matter in the combustion stream.  Furthermore, since the 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
        panicle 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 the kiln after a scrubber failure, 2) the 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 their
         volatile metals than do pumpable wastes, and since solid-form wastes  are not the

Volume ffl                                V-ll

-------
         majority of total wastes at WTI. this scenario is not expected to create the potential
         for a significant additional source of volatile metals emissions.

         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 ED fan operates  in 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 in positive pressure in 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 in 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,  in 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.

Volume HI                                V-12

-------
         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 thar the
         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 air flow are anticipated to DC 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
        in 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 in Appendix ffl-l.
                                                                f^
        The analysis of AWFCOs conducted as part  of this risk assessment, and described in
        the Appendix of this report, has not identified reliable estimation techniques or

Volume m                               V-13

-------
         empirical studies quantifying the nature or magnitude of emissions from overpressure
         releases in 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 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.

         An additional exercise was performed by EPA to estimate the possible magnitude of
         increased emissions of organic compounds from process upsets at WTI.  This exercise
         utilized continuously recorded values of total hydrocarbons ("THCs") in the
         combustion gas at WTI to compare known emissions during the stack test to
         emissions at all other operating times.  Because it used one year of continuous
         emission data, it allowed the estimation of the cumulative effect of all deviations from
         normal operation.  The details of this technique are presented hi Chapter n of Volume
         Vin of this risk assessment.  The exercise concluded that, cumulatively, process
        upsets would not result in more than a 30 percent increase in the emission of organic
        compounds.   This result has been referenced in the Uncertainty Analysis of Volume
        V.

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 #ie available storage
Volume ffl                               V-14

-------
         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 stormwater,
         type and concentration of contaminants hi 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;

     •    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:WP\3533 l.WPD
Volume HI                               V-15

-------
                                    TABLE V-l
                    Observed Variation in the Control Efficiency of
                   Selected Metals During the May 1993 Trial Burn
      Class of Material
Control Efficiency
(percent of weight)
Standard Deviation
  Insoluble vapors
  Soluble vapors
  Fine particles
  Coarse particles
        6
      99.68
     99.977
     99.997
        4
       0.2
      0.006
      0.002
Volume IE
      V-16

-------
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 104
1.5x10-"
9.4 x 10-5
5.0 x 10-6
4.7 x 10-4
1.5 x 10'5
3.4 x 10-5
1.2x 10"


Minimum
2.2 x 10-6
1.2 x lO'5
l.Ox 10-7
3.1 x 10-9
2.6 x 10-6
1.2 x 10-6
2.6 x 10-*
9.5 x 10-6


Maximum
5.2 x 10-3
1.9x ID'3
1.2x ID'3
6.3 x lO'5
4.7 x ID'3
1.9 x 10J
4.7 x ID'3
1.7x 10°
Volume III
V-17

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

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 quaner 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 quaner 1994 sampling program for products of incomplete
    combustion (PICs).  Document number 7289-860.  January.
                                                           if*
Entropy Environmentalists, Inc. (Entropy).  1994.  Volume 1: Stationary Source Sampling
    Repon, Reference No.  12585,  WTI - Von Roll, Inc.  East Liverpool, Ohio; Emissions
    testing for: Lead, paniculate, PCDD/PCDF; Incinerator stack.  February.

Volume HI                              VI-1

-------
 Friedlander. S.K.  1977. Smoke, dust and haze   New York-  John Wile\ 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

 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.
     Pan I, ed. I.A. Suffet, 7-025. New York:  Wiley and Sons.

 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.  Polybrommated
     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 in 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 Paniculate 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 in aquatic environments. Bull. Environ.
    Contain. Toxicol. 36:33-40.

U.S. Environmental Protection Agency (U.S. EPA).  1983. Hazardous waste management
    permit.  Waste Technologies Industries.  EPA Identification # OHD980613 541. U.S.
    Environmental Protection  Agency, Region V.

U.S. Environmental Protection Agency (U.S. EPA).  1988a. Measurements ofparticulates,
    metals, and organics at hazardous waste incinerators.  Office of  Solid Waste.  Draft
    Report.  November.
Volume III                               VI-2

-------
 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 chromatogranin -mas*
     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). 1990a.  Operations and research ai 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;.  1993a.  WTIphase II risk assessment
    project plan, EPA ID number OHD980613541.  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 in 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).  1993d.  Test report for particle size
    distribution study conducted at Waste Technologies Industries, Inc. in East Liverpool,
    Ohio during March 15-17, 1993 trial burn.  Prepared by A.T. Kearney, Inc., Chicago,
    Illinois.

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.

Volume TTT                              VI-3

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

Waste Technologies Industries (WTI).  1993.  Letter from H.J. 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.
Volume ITI                               VT-4

-------
          APPENDIX III-l

EMISSIONS ESTIMATION METHODOLOGY
        AND BACKGROUND

-------
                                APPENDIX III-l

         EMISSIONS ESTIMATION METHODOLOGY AND BACKGROUND

                                  CONTENTS

                                                                           PAGE

 CHAPTER I.  INTRODUCTION 	  I-1

       A.  Scope 	  I-1
       B. Concerns to be Addressed 	  I-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  	 1-2
       E. Organization	 1-3

CHAPTER E.  WASTE PROFILE DATA BASE	H-l

       A.     Introduction  	0-1
       B.     Data Base Development	II-1
             1.     Summary Information  	H-2
             2.     Anticipated Constituent Composition	D-2
             3.     Analysis Results	0-2
       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  	U-5
       D.     Potential Sources of Error 	II-9
       E.     Limitations	II-10
       F.     Uncertainty	II-10

CHAPTER ED. ESTIMATION OF UNMEASURED METALS EMISSIONS AND
           EVALUATION OF METALS BEHAVIOR AT THE WTI
           INCINERATOR	ffl-1

       A.     Concerns to be Addressed	HI-]
       B.     Approach  	ni-1
Volume III
Appendix III-l

-------
                                   CONTENTS
                                       (Cont'd)
       C.     Estimation Technique Selection  	ni-1
              1.     Review of Raw Analytical Data From Trial Burn	  ni-2
              2.     Review of Historical Data	ni-3
              3.     Modeling of Metals Emissions  	ni-4
              4.     Conclusion	III-4
       D.     Emissions Estimation	ni-4
              I.     Introduction  	HI-4
              2.     Model Development	ni-5
              3.     Model Application	ID-15
              4.     Uncertainty Analysis 	    Ill-19
                    a.     Modeling Assumptions 	IE-19
                    b.     Data 	m-20
                    c.     Removal Efficiencies	III-21
                    d.     Uncertainty Estimates  	ni-22
             5.      Speciation	ffl-23
             6.      Other Topics	IQ-23
                    a.     Aluminum Toxicity  	HI-23
                    b.     Chromium Valence State	ffl-25
                    c.     Emissions from Scrubber Water 	ID-25
             7.      Conclusions	ffl-26

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

      A.     Introduction  	IV-1
      B.     Estimation of Organic Emissions	IV-2
             1.     Estimation of Organic Emission  	IV-2
             2.     Estimated PIC Emissions	IV-5
             3.     Estimation Procedure for the Uncharacterized Fraction 	FV-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
Volume III
Appendix III-l

-------
                                   CONTENTS
                                        (Cont'd)
              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-14
              1.     Introduction  	  V-14
              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-28
             4.     Uncertainty  	  V-31

CHAPTER VI. REFERENCES  	VI-1
Volume III
Appendix III-1                               iii

-------
                                    CONTENTS
                                        (Cont'd)
                                       TABLES
                                                                                PAGE
 Table II-1     Assumptions Made During Data Base Refinement	II-12
 Table n-2     Analytical Values for Total Waste Streams
              (Pumpable & Non-Pumpable)	11-13
 Table U-3     Analytical Correction Factors  	11-14
 Table E-4     Key Assumptions for Chapter n  	II-15
 Table HI-1     Metals System. Removal Efficiencies Measure
              During the WTI Trial Burn 	ffl-27
 Table m-2     Waste Feed Compositions Used in the Modeling  	ffl-28
 Table HI-3     Comparison of Predicted Feed Rates with the Permit Limits  	 Ifl-29
 Table ni-4     Classification of Metal Vapors Used to Determine Control Efficiency  .... 111-30
 Table HI-5     Control Efficiencies Used in Model  	IE-31
 Table ffl-6     Predicted Metals Emission Rates	ID-32
 Table HI-7     The Observed Variation in the Control Efficiency of
              Selected Metals During the March 1993 Trial Burn  	ID-33
 Table ni-8     Possible Variation in Predicted Metals Emissions Due to
             Uncertainty in Input Data	 HI-34
Table HI-9    Key Assumptions for Chapter IE	IH-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-32
Table V-2    DRE Results from the 1994 Trial Burn Condition No.  2  	 V-33
Table V-3     Summary Provided by Permittee of Equipment/Procedural Failure
              (Unrelated to Incinerator) Which Resulted in Spills or Releases  	 V-34
Table V-4     Summary of Releases at TSDFs	 V-35
 Table V-5     Emissions from Organic Waste Storage Tanks (Human Health -
              Constituents of Concern)	 V-36
 Table V-6     Emissions from Organic Waste Storage (Ecological)  	 V-37
 Table V-7     Summary of Estimated Emissions from Routine Operations  	 V-38
 Table V-8     Key Assumptions for Chapter V  	 V-39
Volume HI
Appendix III-l
IV

-------
                                     CONTENTS
                                         (Cont'd)
                                        FIGURES
                                                                                  PAGE
Figure HI-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  	  111-37
Figure ni-2    Pathways available for toxic metals in the WTI incinerator  	111-38
Figure ni-3    Schematic diagram of modeling approach used	ID-39
Figure HI-4    The predicted evolution of the particle size distribution
              in 1400 K gases  	  ffl-40
Figure III-5    The predicted evolution of the particle size distribution
              in 400K gases	  ID-41
Figure ffl-6    Relative rates for homogenous condensation and heterogenous
              condensation onto 0.1 ^urn particles and 10 /^m particles  	Ill-42
Figure HI-7    Comparison of the model's predictions and the March trial bum
              results from the WTI incinerator	  ID-43
Figure ni-8    The impact of temperature on the predicted metals emissions
              and the observed SREs	ID-44
Figure ffl-9    The impact of the quantity of air used on the emissions rates
              and the observed SREs of the metals of interest 	ffl-45
Figure IE-10   The impact of chlorine on the predicted metals emissions rates and SREs ..  IH-46
Figure III-l 1   The impact of quench temperature on the predicted metals emissions
              rates and SREs	ITJ-47
Figure ID-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-14
Figure IV-2    Procedures for estimating emissions of organic compounds	IV-15
Figure IV-3    Procedures for estimating emissions of organic compounds
              (estimation steps with uncertainties highlighted)	IV-16
Figure V-l     Slag and ash handling diagram - WTI facility	  V-41
Volume III
Appendix III-l

-------
                         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 oroject 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 in 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  in 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.

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.
                                                           >••
 Volume HI
Appendix HI-1                               1-1

-------
        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 paniculate 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
       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 assumptipns made in the process
will be discussed in greater detail in each of the following sections of this appendix.  Many of the
Volume ffl
Appendix III-1                                1-2

-------
 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 or the combustion gases in 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 II describes activities relating to the waste
 profile data base. Chapter III 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 HI
Appendix III-1                                 1-3

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

 A.    Introduction
       As an initial step in 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 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 ba?°  The 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 in 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 identifies the
generator), sampling date, annual estimated volume for the waste stream, physical state (i.e.,
Volume m
Appendix m-1                                II-1

-------
 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 Obase 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 in any way to the original WTI source
 codes. In this way, a detailed honconfidential 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 data sets are created as described below:
       1.     Summary Information
             Data from pages 2 & 3 of the waste profile include:
             •      Surrogate waste code;                                                     ^
             •      Anticipated annual volume;                                                 v
             •      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.                                                                   v^
Volume m
Appendix m-1                               IT-?

-------
        The anticipated volume, reported in pounds per year are totalled 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 in 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 in 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 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
   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 ffl
Appendix ffl-1                               II-3

-------
       selected as the constituent's concentration. However, this conservative assumption
       introduced an uncertainty in that waste constituent volumes now total 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 in 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 indicate 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 reported under other
       names are renamed to the conventional name and summed. The resulting renaming and
       grouping of isomers under the most commonly used name or IUPAC designation,
Volume HI
Appendix ffl-1                               II-4

-------
        (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 in two discrete waste streams: pumpable and non-pumpabie feeds.  To create
       these two waste streams, the wastes are combined by their 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 aic 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 in the physical
                    state field are sometimes listed as bulk liquids or drummed liquids in the
                    packaging field. For consistency, the simplesHield (physical state) is
                    selected when determining the ultimate category (pumpable or non-
Volume m
Appendix m-l                                II-5

-------
                     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 apparenr 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 (unconnected 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 in 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.

                     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.
Volume ffl
Appendix IH-1                                II-6

-------
              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 totalled 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 total volume) fed to the
       kiln did change,  neither the specific compounds nor the total volume in pounds per year
       change.
Volume ffl
Appendix m-1                               II-7

-------
              The two large data bases (representing the pumpable and non-pumpable waste
       feed streams) resulting from this approach are treated identically to each other in 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 in these two data bases.  The final results of these two data bases  are
       presented in 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 in
       conjunction with the  mass balance approach to determine better estimates of the amounts
       of each species input to the kiln.
             First, molecular formulas are entered for each of the materials listed in 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 in Table II-1.
             At this time, analytical values presented in 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 their calculated values are presented in Table II-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 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
Volume HI
Appendix III-1                               H-8

-------
        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
       totalled and subtracted from the total mass of their 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 II-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 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.
Volume HI
Appendix m-1                                II-9

-------
       Once the volumes of constituents containing analytes such as chlorine are revised to
 reflect actual analytical values rather than 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.

 E. Limitations
       Potential sources of uncertainty in the WTI 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 contains 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  in Table n-4.  The
uncertainty associated with the volumes of constituents determined in the data base in 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  (see Table II-3).
              Very few compounds are affected; however, compounds containing iodine, non-
              amenable cyanide, selenium, silver, antimony, thallium, And chromium are affected;
Volume ffl
Appendix III-1                                11-10

-------
        •      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
               n-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-pumpabiliry;

       •      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 H-3).

       In general, the sources of uncertainty are due to limitations in the way in which data are
reported in 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 time in the future.
Volume III
Appendix in-1                                11-11

-------
                                    TABLEIM
            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
 Assumptions
 SbCl3
 AsCl3
 BaCl2-2H2O
 CaCrO4
 Ca«(OCl)2
 C2H2C12F2
 Dichlorofluoromethane
 C8Hlg (octane)
 Methyl Ethyl Ketone
 (Ethanol+ethanolamine)/2
 Pentachlorobiphenyla
 Trichlorofluoromethane
 C2H4O2 (acetate ion)
 Acetic anhydride
 CiftHie^
 Cr2O3 (Cr(m)oxide)
 CiyHjyC^
 CI4H12O0 iNa2c
        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 m
Appendix III-1
n-i2

-------
Table n-2. Analytical Values for Total Waste Streams (Pumpable & Non-Pumpable)
1














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 As
Total Ni
Total Sb
Total As
Total Be
Total He
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.C
112
204
108
58.7
122
74.9
9.01
201
127

Total 1
g moles/yr I
8.80e+07
2.25e+07
1.07e+06
6.76e+05 I
5.26e+05
4.28e+05
3.28e-K)5 I
1.04e-K)5
7.98e+04
6.29e-K)4
4.96e+04
3.74e+04
2.31e+04
1.61e+04
1.55e+04
1.43e+04
8.29e+03
5.49e-K)3
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 in
Appendix HI- 1 11-13


-------
                           Table II-3. Analytical Correction
                                       Factors
                    Analytes

               Total I
               Total CN non-amen
               Total Be
               Total Hg
               Total Se
               Total Cd
               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
               Totals
               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 in
Appendix HI-1
     n-i4

-------
Table II-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 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
^=as^====:
Direction of
Effect
========
underestimate
overestimate
unknown
===== 	
Volume III
Appendix III-1
H-15

-------
        CHAPTER m. 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 (BEFs) 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 are measured in the triai burn — antimony (Sb), arsenic (As), beryllium

 Volume III
Appendix III-1                              III-l

-------
(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 in this preliminary analysis.
       •       Review of raw analytical data;
       •       Review of historical data; and
              Modeling of metals emissions.

       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 in 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 bum 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 in 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:
Volume III
Appendix III-1                               III-2

-------
               •      Waste feed samples were obtained by the Ohio EPA and may have been
                     available for analysis;
               •      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 ffl
Appendix IE-1                               IH-3

-------
        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 in
        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 similar
       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 six  additional 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 ID
Appendix III-1                                III -4

-------
                     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+3 to Cr*" in the incinerator stack and at the
                           receptor.
                           Evaluate the potential for trace impurities in 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 IH
Appendix m-1                               III-5

-------
                    The system removal efficiencies (SREs) measured during the trial bum
              are summarized in Table III-l. SRE is defined as:
                                           Emission                             ^
                                           —    ,
                                             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 HJ-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 witlj 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., (1990) Seeker (1990) and
                                                             »"
             Linak and Wendt (1993). These reviews synthesized  a mechanistic understanding
Volume HI
Appendix ffl-1                               III-6

-------
              of the phenomena that control the behavior of metals in waste 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 HI-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
Volume m
Appendix ffl-1                               III-7

-------
              reactions principally involve the reduction of 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 in 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 ^m in diameter (Friedlander 1977).
              Heterogeneous condensation tends to favor small particles due to their nigh
              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 in coagulation of smaller
              particles into larger particles. Small particles are extremely mobile, are present in
              large numbers and coagulate quickly.  As the particles become larger, their mobility
              and number concentrations decrease resulting in a lower coagulation rate.
              Investigations of combustion systems have found that any particles smaller than
              about 0.1  urn quickly coagulate, while those larger than 1 um 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
              air cleaning system. One group ranges in size from 0.1 to 1 urn and is formed from
              the metals that vaporized and subsequently condensed.  The second group is larger
              than 10 jim 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 removal of
Volume HI
Appendix ni-1                               III-8

-------
              metals in any given device are generally well known and are described in 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 in size
              from about 0.1 jum 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 \VTI 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 ffl-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 in  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 in the primary
              chamber so no changes are expected to occur. Thus, no calculations are performed
              for the SCC.
                    The processes which occur in 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 flue gases is
Volume ffl
Appendix m-1                               III-9

-------
              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 urn) 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 ni-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 short-term variations in the feed rate of any
Volume HI
Appendix ffl-1                              III-10

-------
              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).
Volume ffl
Appendix ffl-1                              HI-11

-------
                    It should be noted that chemical equilibrium will probably not be achieved
             in all sections of the incinerator due to kinetic limitations. Studies 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^pf the entrained particles
Volume IH
Appendix III-1                               III-12

-------
                     These assumptions are based on an examination of coagulation and
              condensation processes as described below.

                     Coagulation. Coagulation processes have 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 ffl-4 illustrates the evolution of the aerosol in a gas at 1400 K. In 0.5 s,
              nearly all the nuclei have formed particles 0.1 to 0.5 um in diameter. After 120 s,
              the aerosol has stabilized with all of the mass that started as nuclei having shifted
             to 0.5 to 1 urn. Figure ni-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 |im
             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  1993). In typical gas residence times in an incinerator, bimodal
              particle distributions are  usually generated. One mode occurs around 0.5 jam 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 in Figure III-6. Heterogeneous
                                                            t*
              condensation on the surfaces on fume particles is the predominant form of
Volume HI
Appendix m-1                              III-13

-------
             condensation. When fiime particles are present, almost no material condenses by
             either of the other two modes.

             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 burn 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 in 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 that
             is both soluble and known to be present as a vapor was used in 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 in the trial burn is assumed to
             be the control efficiency for all fine particles (that is particles with diameters of 0.1
             to 1 urn). The assumption is justified by the following observations.

             •      Arsenic is volatile and it is likely that all of the arsenic present in the waste
                    during the trial bum vaporized; and
             •      Thermodynamic calculations indicate that arsenic will condense in the flue
                    gas cleaning system. As discussed above, all material which condenses will
                                                           **'*
                    probably be present as fine particles.
Volume ffl
Appendix ffl-1                              HI-14

-------
                     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 particles  Thus, the
              control efficiency of arsenic should be the same as the fine particle removal
              efficiency.
                     The trial bum 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 in the trial bum. Thus, the only mechanism
              which will result in chromium being present in 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 in 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:
          CrE is the mass emission rate of chromium in the trial burn and CrFp is the feed rate of
          chromium in the pumpable waste stream. Table ni-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 in the trial bum.
          The waste feed composition is summarized in Table ffl-2. The results of the
          comparison are summarized in Figure ffl-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 ^yell with the observed
          values. The greatest difference is observed for beryllium where the prediction is about
Volume HI
Appendix ffl-1                              HI-15

-------
           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.
                  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 in 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 et al. 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 IH-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 in that manner is not generally
          conservative. However, the modeling results indicate that for the metals measured in
          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 in 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 stoichiometry;
           •      Waste chlorine concentration;
           •      Quench temperature; and
           •      Entrainment rates.
Volume III
Appendix ffl-1                               III-16

-------
                 The behavior of each of the 15 metals listed in Table III-6 is examined

                 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).
                          t s
                       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 ths predicted SREs.

                 2.     Combustion Chamber Temperature
                       Four combustion chamber temperatures are examined — 1000°C,
                 1100°C, 1200°C and 1400°C (1830°F, 2010°F, 2200°F and 2550°F). The
                 temperature of 1200°C is used in the base case predictions reported above.
                 Figure IH-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"* 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 ID
Appendix ffl-1                               III-17

-------
                 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 IEI-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 HI-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 in 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 ffl
Appendix ni-1                              IJT-18

-------
                  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-l 1 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 pumpabie 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 ni-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.

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

          a.      Modeling Assumptions
                 The model used to make the predictions reflects the current state-of-the-art for
          estimating metals behavior in waste incineration systems. However, several
          assumptions are inherent in 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 ffl
Appendix m-1                               III-19

-------
                 1.     Thennodynamic 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 the assumption
           that 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.

                2.     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 in the thermodynamic
          data base. It is not believed that this has a major impact on the predicted emissions in
          this case.

                3.     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.

          b.    Data
                It is possible to evaluate the impact  of variations in the data used by the model
          on the predicted emissions.  The data fall into two classes:
Volume HI
Appendix ni-1                              111-20

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

                  Site-specific data include such items as combustion chamber temperature and
           waste feed composition. General data includes thermodynamic 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).

          c.     Removal Efficiencies
                 Control device removal  efficiencies for vapors, fine particles and coarse
          particles are determined based on the results of the Marcfi 1993 trial burn as is
Volume HI
Appendix m-1                              FI-21

-------
          described above. The control efficiency for insoluble metals was estimated based 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 bum. 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.

          d.     Uncertainty Estimates
                Based on the values in Table ffl-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
          bum 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 in metals feed rates is based on the
          following observations:
Volume HI
Appendix m-1                              111-22

-------
                   1)     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.

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

                  3)     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.

    5.      Speciation
           The model generates some very approximate predictions of the metal  species which
    may form in 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 for informational purposes
    only. Great care should be taken in any quantitative use of this information.

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

           •      Evaluate the health risks associated with Al emissions;
           •      Model the ratio of Cr+3 to Cr"* in the incinerator  stack at the receptor; and
           •      Evaluate the potential for trace impurities in the scrubber water to become
                 stack emissions.

                 a.     Aluminum Toxicity
                        The Combustion Engineering Work Group  of the Peer Review Panel
                 placed Al  on its list of metals that should be  included in the multi-pathway risk
Volume ffl
Appendix III-1                               IH-23

-------
                  assessment. The group stated that Al might play a role in biological
                  metabolisms or may interfere with the action of other metabolic metals.
                  Though the toxicity of Al 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 in 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 in these areas.
                 Usually the human gut excludes Al. The total quantity in the body is typically
                 25 mg and the daily intake vanes 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 in 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.
Volume ni
Appendix ffl-i                               111-24

-------
                 b.     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 EPA's suggestion as well as George  Huffman,
                 both of EPA/RREL in Cincinnati, Ohio. Neither person was aware of such an
                 EPA model.
                        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**in the atmosphere. The half-life of Cr*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"* 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 reposed 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.

                 c.     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 EPA believes this to be a minor potential source of emissions
                 because this feed stream normally consists largely of carbon-treated storm
                 water  runoff.
Volume m
Appendix ffl-1                               111-25

-------
           7.     Conclusions
                  Emission rates of the metals that were not measured during the trial bum are
           estimated. The estimates are based upon current understanding and scientific
           principles. Realistically conservative assumptions are made when required and are
           summarized in Table ID-9. The predicted emissions are summarized in Table III-6. ,\s
           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 developed at
           Lawrence Livermore National Laboratory may be useful in refining predictions of the
           chemical species formed.
Volume ID
Appendix m-1                              111-26

-------
               Table ffl-1.  Metals System Removal Efficiencies Measured
                             During the WIT 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

Run 3
99.993
99.98
99.992
99.982
>99.9990
99.987
10.59

Average J
99.986
99.98
>99.991
99.987
>99.9993
99.990
6.54 ||
Source:  May 1993 WTI Trial Burn Report
Volume m
Appendix ni-1
m-27

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


Element

Nonmetals
C
H
O
N
Cl
F
Br
S
P
Si
Input Rate, mol/min
Trial Burn
Pumpable



.
-
_
_
_
.
_
_
21.6
Non-
Pumpable

_
-
-
-
.
-
-
-
-
84.5

Total


3700
6600
551
0
766
0
0
47.3
0
106
Annualized Average
Pumpable


.
-
-
-
-
-
-
-
-
23. 6b
Nontoxic Metals
Ca
\-rtl
K
Na
Fe
Li
Toxic Metals
Al
As
Sb
Ba
Be
Cd
Cr
Cu
Pb
Hg
Ni
Se
Ag
Tl
Zn
Water
o
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

4.90e
0.031
0.0035
0.0728
00006
0.0006
0.0293
0.0922
0.0307
0.00009
0.0031
0.00041
0.00075
0.00312
0.0288
0
Non-
Pumpable

-
-
-
-
-
-
-
-
-
23. 6C

0.019
0.024
0.480
4.750
0.0004


4.90*
000002
0.0001
0.00005
0
0.0152
00004
0.0022
0.00007
0.00006
0.0023
0.0268
0.00810
0.00732
0.0912
0

Total


1180
2029
166
40
43
1.29
0.81
2.04
1.15
47.2

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
Maximum Heat Input
Pumpable


-
-
-
-
-
-
-
-
-
94.3

4.03
0.709
0.585
19.0"
0.178


19.6C
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
Non-
Pumpable

-
-
-
-
-
-
-
-
-
94.3

0.0772
0.0995
1.97
19.0C
0.0015


19.6°
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

Total


4840
8320
68!
165
175
5.27
3.34
8.36
4.72
189

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
        From waste constituents (not from air).
        Assumed 50 wt.-pct. of ash to be Si.
        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.
        Assumed 20 wt.-pct. of ash to be Fe.
        Assumed 10 wt.-pct. of ash to be Al.
Source: May 1993 Trial Burn Report
         Volume III
         Appendix HI-1
ffl-28

-------
                           Table III-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 IIIC
Tier III
Adj. Tier I"
Tier III
Tier III
NAC
Tier III
Tier III
NA
NA
Adj. Tier I
Adj. Tier I
NA

Feed Rate
(Ib/nr)
9.65
3.81
260
0.3
- 169
NA
100.4
0.2f
NA
NA
16
1.6
NA
Annualized Average Case1


Feed Rate
(Ib/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
Maximum Heat Input" I
1

Feed Rate
(Ib/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
<7C |j
2.4
33.3
2.0
0.9
0.5
NA
3.4
5.5
NA
NA
3.2
73
NA ||
Other Components ||
Cl
NA
2700
201
7.4
823
30.5 1
Solid Waste* ||
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 averag*
       case of 29.5 MBtu/hr. Maximum design heat input rate is 121 MBtu/hr, or 4.10 times higher than th
       average annual case.  Thus, the maximum rates are obtained by multiplying the average annual rates
       by 4.10.
c      Tier HI limits are assumed to be based on the average feed rate during the March 1993 trial burn, tes
       condition 1.
d      Adjusted Tier I limits are based on an assumed dispersion coefficient of 1.5 Mg/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 wasi
       profiles.

Source: May 1993 Trial Burn Report
       Volume III
       Appendix IH-1
ra-29

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







Volume III
Appendix III-l
ffi-30

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

Insoluble vapors
Soluble vapors
Fine Panicles (< 1
Coarse Particles (> 10 /xm)
Control Efficiency,  %

6
99.68
99.977
99.9973
     Calculated from May 1993 Trial Burn Report
     Volume III
     Appendix III-l
  ffl-31

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


Tier III 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

1.7x10-*
9.0 xlO-6
2.0 x'lO-8
6.8 x 10^
6.8 x 10~7
2.4 x lO'5
3.3x 104

5.9 x ID'5
3.7 x ID'5
2.3 x 10-5
1.2x10*
1.1x10-*
3.7 x 10-6
8.2 x 10-6
3.0 x 10-5
Maximum Heat
Input Basis

6.9 x 10-6
3.7 x 10-5
8.1 x 10-"
2.8 x 10-5
2.8 x 10-*
l.Ox 10-4
1.4x lO'3

2.4 x 10-4
1.5x10-*
9.4 x lO'5
5.0 x 10-6
4.7 x 10-*
1.5xlO-5
3.4 x 10-5
1.2x 10-*
Predicted Emission Rates Using
Trial Burn SREs. g/s
Annualized
Average Basis

l.Ox 10-6
9.0 x 10-6
8.1 x 10-9
3.8 x lO'6
1.7xlO-7
1.1 x ID'5
3.4 x 10-*









Maximum Heat
Input Basis

4.2 x ID'6
3.7 x 10-5
3.3 x 10-8
1.6x 10s
7.1 x 10-7
4.3 x ID'5
1.4x 10'?









Volume III
Appendix III-l
                   ffl-32

-------
            Table III-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
IL-33

-------
              Table III-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 10^
1.5x 10-4
.. 9.4 x ID'5
5.0 x 10-6
4.7 x IQ-4
1.5 x ID'5
3.4 x 10'5
1.2x 10-4

Minimum
2.2 x 10-*
1.2x lO'5
l.Ox 10-7
3.1 x 10-9
2.6 x 10-*
1.2 x 10-*
2.6 x 10-*
9.5 x 10-*

Maximum
5.2 x 10 3
1.9x 10 '
1.2x 10 3
6.3 x 10'5
4.7 x 10-'
1.9x 10J
4.7x 10°
1.7x 10 :
Volume III
Appendix HI-1
HI-34

-------
                                                     Table III-9
                                               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
bum
Thermodynamic equilibrium is maintained throughout the
incineration and flue oas cleaning system 	 J 	
AH important compounds are included in the thermodynamic
database 	
All elements in the incinerator are intimately mixed
No condensed chase 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 	 	 	
PbCVand CrO,Cl, will not form
All metals that vaporize and subsequently condense are found
on particles 0.5 /urn in diameter
All pumpable waste is entrained
	 	 	 	 	
Basis
Jest available data
t/iost conservative reasonable assumption. Produced highest
metals feed rates which can be reasonably predicted.
3esi 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
urn
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
Magnitude of
Effect
high
high
high
medium
medium
low
medium
low
low
medium
low
low
-^ 	 ,
Direction of II
Effect |l
either II
either II
	
either |
II
either II
II
either U
D
either ||
II
overpredict II
either
underestimate
underestimate
either
overestimate
	 	
Volume III
Appendix 111-1
111-35

-------
Table III-9
Key Assumptions for Chapter III
Assumption
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
SO, capture efficiency is an appropriate indicator of the flue gas
cleaning system's ability to capture soluble vapors
Basis
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; is
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 bum
SO, will be present in the flue gases as a vapor. SO, are
soluble. The capture efficiency of SO, is lower than other
soluble vapors which were measured in the trial bum Thus,
use of SO, is conservative
Magnitude of
Effect
low
low
low
low
Direction of
Effect
underestimate
overestimate
either
either
Volume III
Appendix 111-1
111-36

-------
     0.030
    0.025
_  0.020
0)
o
 I  0.015

"2
"55
 0)
°-  0.010
             Source Ma% 199? TnaJ Burn Report
    0.005
    0.000
Run 1

Run 2

Run3

                                 i
                    I

                                 1
                Be
   Pb
Cd
Sb
As
 Figure ffl-1.  System penetration observed during the WTI trial bum for As, Be, Cd, Pb and Sb.
              The metals are listed in order of increasing volatility.
Volume ID
Appendix fll-l
                                        ffl-37

-------
  Melals in
  Pump a We
   Wastes
  Melals In
NonPumpaWe
   Wastes
                                                                          INITIAL
                                                                        QUENCHING
                                                                          ZONES
                                                                                                 Fine Fly Ash
                                                                                                 (Condensed
                                                                                                   Vapors)
                                                                                       Residual Vapors
                                    Figure III-2.  Pathways available for toxic metals in the W11 incinerator
                       Volume III
                       Appendix 111
II 38

-------
I Auxiliary Fuel
I    Feed Rate
I    Composition
I
I Non-Pumpable Wastes
|    Feed Rate
|    Composition
j    Metals Feed Rates
           j Pumpahle Wastes
           |    Feed Rate
           |    Composition
           |    Metals Feed Rales
| Conditions
|    Tcmpcruluic
|    Stoichiomctry
                                         Primary Chamber
                                            Reaction
                                            Vaporization
                                            Entrainmcnl
] Metals in Residuals
i
!
1

1 Conditions ,
1 Temperature ,
1 Stoichiometry j
1 	 1


Air
Feed Rate
	 "sT 	

'

Metals in Rue Gases
Vapor
Entrained Panicles

Quench
Reaction
Condensation
r ^
^
Flue Gas
Flow Rate
Composition

                                                                                                    ..i
                            Metals in Flue Oases
                               Vapor
                               Fine Particles
                               Entrained Particles
                                        Input
                                        Models
                                        Output
I Removal Efficencies
I   Soluble Vapors
I   Insoluble Vapors
I   Fine Particles
I   Coarse Particles
Rue Gas Cleaning
                         Figure III-3.  Schematic diagram of modeling approach used

-------
1.0
          Figure III-4.  The predicted evolution of the particle size distribution in  1400 K  gases
       Volume III
       Appendix Ill-l
III

-------
Hl-S   The predicted evolution of the particle si?e distribution  in 400K pases

-------
1X105
1x104
1x103
H> 1X102
5 1X101
CD
rr
I 1x10°
To
in
-a 'x'u
o
0 1X10'2
1x10'3
1x10"*
1x1 0~5
-









.... ;.










•:•: '.
. ':.-.- :..'.".• ' •
-

-
-

Homogenous Heterogenous Heterogenous
condensation condensation condensation
onO.lpm on 10pm
partides partides
Conditions
Temperature. K 1400
Saturation Ratio (P/P ) 1000
eq
0.1 pm Partide Concentration
gr/dsd 100
10 pm Partide Concentration
gr/dscf 100




Figure ffl-6.   Relative rates for homogeneous condensation and heterogeneous condensation onto
              0.1 /xm particles and 10 /xm particles.
Volume III
Appendix III-l
HI-42

-------
    1x10
        ,-1
    1x10
        ,-2
 JO
  O)
  oT
  o
  v>    _..
 •g 1x1 cr4
 LU
   1x10''
   1x10**
              Sb
As
Be
                          v f  Trial Bum

                             Prediction
Cd
Cr
Pb
Hg
 Figure IE-?.  Comparison of the model's predictions and the March trial burn results from the
              WTI incinerator.
Volume III
Appendix III-l
                m-43

-------
   1x10
          Al   As   Sb   Ba   Be   Cd   Cr  Cu  Pb   Hg   Ni   Se   Ag   Tl    Zn
  1x10L
     1-1
  1x10
S Ix10'2

  1X10"4
  1x10"'
         rA i i-j
          Al   As  Sb  Ba   Be   Cd   Cr  Cu   Pb   Hg  ,Wi   Se  Ag   Tl   Zn
 Figure ffl-8.  The impact of temperature on the predicted metals emissions and the observed
              SREs.
 Volume III

-------
     1x1 Ol
    1x1 0~2
  .J51X10"3
  O:

  11X10-4
  •§ 1X10'5
  v>
    1X10
    1X10
    1X10"
    1x10"
        '7
   Stoiaometnc
      Ratio
  ffl  0.8
  El  1.0
  D  1.2

            Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag    Tl    Zn
   1x10°
   1x10''
 21X10"2

 |ixi
-------
1x10°




1x10'1




1x10'2
£
1C
•      K
E1X10'5
  1x10H
  1x10
  1X10"1
      ,-7  .
                     D  Base



                         No Chlorine
          Al   As   Sb   Ba   Be  Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag   TI   Zn
1x10°
1x10"1
.1
£ 1x10'2
S.
11X10-3
CO
1x1 0~*




-
m



i
i













^
V













s.














;














.f
k:









Ifl



''•(
• •y
f




1



i
1













•!•!•
|













::-
S




1


••
.'!
i
•


























.;|;
:•;
i





-
-
-
1
'•;
            Al   As   Sb  Ba  Be   Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag   TI    Zn
     Figure IE-10.  The impact of chlorine on the predicted metals emissions rates and SREs.
   Volume m

   Appendix ffl-1
                                         m-46

-------
   1x10'
  1x10
          Al   As   Sb   Ba   Be  Cd   Or  Cu   Pb  Hg   Ni   Se   Ag    Tl   Zn
  1x10°
  1x10
g
S1X10"2 I-

I
E
31
  "
  1X10-3
  ixior
  1x10"'
                     150°C
                     (Base)
                     200°C
                     400°C
          Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag   T!    Zn
Figure m-11.
               The impact of quench temperature on the predicted metals emissions rates and
               SREs.
Volume HI
Annendi* TTT-1
                                         m-47

-------
    1x1 Ou
   1x10'
       "2
   1x1 0
 °>1x10~3

 E1X10'
   1x10M
   1x10
       ,-7
Base
50 % Entrainment of Pumpable

20% Entrainment of Nonpumpable
            Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb   Hg   Ni
   1x10°
   1x10"
 |1x10'2
 "
 "3
   1x10'3
 CO
    1X10'5
 Base

 50 % Entrainment of Pumpable

 20% Entrainment of Nonpumpable
                                                   Ag   Tl   Zn
            Al   As   Sb   Ba   Be   Cd   Cr   Cu   Pb   Hg   Ni   Se   Ag   Tl   Zn
Figure ffl-12.   The impact of entrainment rates on the predicted metals emissions rates and SREs.
  Volume III
  Appendix III-l
                      m-48

-------
    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 niaster 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 in 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, 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 in time  to be
 numerically averaged into the risk assessment calculations. However, those data were
Volume III
Appendix III-l                              IV-1

-------
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.
       The following discussion is divided into four sections.  The first section deals with the
approaches the 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

      1.     Estimating Emissions and Uncharacterized Fraction
             Data on organic compound emissions that WTI reported as being quantitated
      durirg 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       v_
      trial burn, 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      v
       be biased low. Related to this concern is the observation that the concentrations of

Volume ill
Appendix III-l                               iV-2

-------
                     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
                     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/rarans.  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 identify ing PICs would
                                         j
                     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.
fs	\s
(                           The first approach was an idea suggested by the  Peer Review Panel to gather
v	,/
                     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  in the incinerator stack gas, the specific compounds

             Volume III
             Appendix III-l                              IV-3

-------
      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 simply using known common PICs.
      Also, the specific procedure used for these predictions was not reported.
             A second approach which the 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 WTI 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 in 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 in the risk assessment.  Although this approach
      would not result in 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 in estimating PIC emissions.
             A third approach which the 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.

Volume III
Appendix III-l                              IV-4

-------
       2.     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
       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 in the stack gas at half of
                           f
       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 Mg 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 in 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 EPA's list of target
       PICs (U.S. EPA 1994 and Mercer 1994) are considered hi calculating residual
       emissions.  Estimated emissions are calculated in 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.
       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 tetrachloroethyiene 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 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.
Volume III
Appendix III-1                              IV 5

-------
             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
       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 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. Stack
       gas concentrations that correspond to  the estimated emission rates presented in
       Attachment 4, range up to 400 ng/L for tetrachloroethylene. These values are in 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.

       3.     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).
       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
Appendix III-1                              FV-6

-------
       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 in 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 in the
       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 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 convert 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 based 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 in 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
       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

Volume III
Appendix III-l                              IV-7

-------
       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 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 fiiran emissions measured during compliance tests.  The third
       column hi 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 hi the emissions estimate.
              At the tune 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 in 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 hi the uncertainty section of the
       risk assessment.
              Figure F/-2 presents a detailed flowchart of the steps taken in estimating PIC
       emission rates and in comparing this estimate to the adjusted THC value.

C.     Effect of Control Devices
       In evaluating organic compound emissions from the WTI facility, it is important to
consider the potential effect of the control devices on these emissions.  The effect of the
control devices on nondioxin organic compounds is most associated with two factors.  First,
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
Appendix III-l                               W-8

-------
 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 in a reduction in 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 injection in a spray
 dryer/fabric filter ah* 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 FCDD/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 fonned in 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 in* place, and because
repeated testing has confirmed the effectiveness of the ECIS, the risk assessment is based on

Volume III
Appendix III-l                               F/-9

-------
 emissions data from the post-ECIS installation tests.  Specifically, average emission rates for
 the 17 dioxin and furan congeners (see Volume ffl, 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.

 E.     Uncertainties
       Figure IV-2 presented the procedure used to estimate the emission rates for organic
 compounds  for the WTI incinerator. Figure FV-3 is identical to Figure FV-2, except that
 several data sources and steps hi the procedure are numbered. These numbers correspond to
 the key assumptions listed in Table IV-3.  For each factor in 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 in 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 FV-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 in  the organic compound emissions estimate.
Volume III
Appendix III-1                               IV-10

-------
                 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 chloride6
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
< 2.5^90
< 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 III
Appendix III-l
rv-ii

-------
                                                                Table IV-2
                                                        Sources of Emission Rates
  Source
                                                                                                                            No. of vah
  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
   1 A*-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
                                                                    54
  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
                                                                    12
                                                                     5.

                                                                    20
                                                                    14
                                                                    40
  Total dioxins and furans measured during compliance tests, TEQ
 "•-Designation used in Attachment 3.
Volume III
Appendix IH-1
IV-12

-------
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)
Dioxin/ftiran data from 8/93 test are representative of
incinerator and APC operation and waste composition and
appropriate method of measurement was used (13)
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
Conservatively high values selected from best available data
Magnitude
of Effect
medium
low
low
medium
medium
medium to
high
high*
medium to
low
medium
medium to
low

Direction of
Effect
variable
variable
variable
variable
variable
variable
variable
variable
variable
variable
*High relative magnitude but small absolute magnitude
Volume HI
Appendix III-I
IV-13

-------
Total Mass of Organic
Compound Emissions
Total Hydrocarbons
(THC) Measured
During Trial Burn*
Categories of 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
Percent of 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 HI
             .	«:. in.|

-------
                            3 PIC's horn 3/93 liial
                             bum. Condition 2
                                                                                                              Eillrnaiad samlvolaiila
                                                                                                               dalactlon Vmils from
                                                                                                              backup data la 7/1/94
                                                                                                                  raporlad PICi
                                                     3 HC's horn 2/94 ralMI ol
                                                      l/ial bum. Condition 2
 Volalila PIC'l raporlad
7/1/84 toy IATI (dau Irom
  2/94 uia faun lalMl)
  Liu ol ma|« organic
compounds in annuahzeri
      wajla pralile
                                                                                                             SamiralalUa PICf npoiwd
TIIC valuai liom 3/B3 Irlul
                                                                                                             7/1AMbyWII|dalaliam
                                                                                Divldt lipotud daiactlon
                                                       Cakulala avarag* ovar
                                                       Ihtaa Ian luni lor «ach
                                                        PIC, convail lo o/iac
                                                                                                                                                                                                   Calculala av»ag> ORE
Calculala aval aga ovaf
Ihtaa nil luni lot aach
 PIC. convail lo g/iac
                                                                                 ImlubyrtrotaalllMI
                                                                                                                                                                                                 oval nina rapoilad values
                                                                                                                                                                                                                                Calculala 1EQ avwaga
ihraa ml tuns, oonvail
                                                                                                                                         Divldt dalacllon imM by 2
                                                                                lun/oompoundi whan no
                                                                                                                                                                                                   Calculala casldual POHC
                                                                                 Calculala ivataga lor aach
Middy 1HC by laclor ol
                                                                                                             Calculala avwaga lor aach
                                                                                                                                                                                                   amisslonl lor compounds
Multiply THC by laclor ol
                                                                                                              UU anHsslonialai (g/iacl
                                                                                                               »> all conuoundi; idact
                                                                                                              compoonoa aia pratanl In
                                                                                                              mora tun ma daia lourca
                                   Subtract loial amlsslon
                                                                                Total Iht •million ratal
                                   ma» ol uncharactarizad
                                                                                       Figure 1V-2.  Procedures for estimating emissions of organic co npounds.
                                                                        Volume III
                                                                        Appendix 111 I
                                                                                                                            IV-15

-------
List ol majof ofganlc
compounds in annuahtwd
wasle prolil*


                                                              Figure IV-3. Procedures for estimatini; emissions of organic comppound:; (estimation steps
                                                              with uncertainties highlighted.
                                                             Volume 111
                                                             Appendix 111-1
VOLUME  III
APPENDIX 3-1
                                                                                                      IV-16

-------
       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
       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 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 m
Appendix ffl-1                               V-l

-------
                     A primary source of information is the February 1994 trial burn results.
                     According to EPA Region 5 personnel, during Run 1 of this test, an
                     AWFCO occurred, due to an exceedance of the minimum 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 in 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
Volume in
Appendix HI-1                               V-2

-------
                     difference in emissions, probably because the facility could not be forced
                     into enough of an upset condition to yield usable data (U.S. EPA 1987a).

              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 Waterland. Acurex Corporation: Mr. Waterland is responsible for the
        operation of the IRF, under contract to 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 in 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 in the
             previous case, it is assumed that 144 of the total number of AWFCOs at the facility
Volume m
Appendix ffl-1                              V-3

-------
              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.
                    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
Volume HI
Appendix m-1                              V-4

-------
              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 in significant emissions. The results
              of this initial analysis are presented below.
                     0)    Non-overpressure AWFCQs
                            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 EPA's Air and
                     Energy Engineering Research Laboratory, the contractor for EPA's
                     Incineration Research Facility at Jefferson, Arkansas, and internal
                     combustion experts at Midwest Research Institute (a consultant to EPA on
                     this project) were also contacted during the preliminary research phase.
                     Unfortunately, no published data, 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
                    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 in the high BTU waste
                    feed lance.  The plugging caused observable flow disturbances and
                    increasing CO 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
Volume m
Appendix m-1                               V-5

-------
                     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  1987a).
                     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 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
Volume ffl
Appendix m-1                               V-6

-------
                     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 unbumed or partially bumed
                     organics'from the kiln seals. As previously noted,  108 AWFCO events in
                     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 in 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 kifc seals (Victorine 1994). These two events are
                     estimated to have resulted in emissions from the kiln seals of three second
                     and three minute durations, respectively. These durations are observer
                     estimates, since no mechanism is in 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 in a batch mode.  Once
Volume HI
Appendix ffl-1                                V-7

-------
                     charged, the containers rapidly release volatile components into the gas
                     phase.  This rapid volatilization and subsequent combustion can cause
                     temperature and pressure excursions in the kiln, resulting in 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 in 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 in many cases puffing releases will be limited to a few
                     seconds in 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, in 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 in the future.
                          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 in terms of volumetric heat release,
                     gas-phase residence time and temperature profile.  This series of studies
Volume ffl
Appendix ffl-1                               V-8

-------
                     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 "puff" 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 incinerators using the rotary kiln
                     simulator. This research examined optimum settings of kiin 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 performance 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 in 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 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
Volume III
Appendix ffl-1                               V-9

-------
                     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
                                  (Cundyetal. 1989). The purpose of the simulator is
                                  to individually examine the fundamental phenomena
                                  that occur in full-scale units, and to gain an
                                  understanding of the qualitative trends that would be
                                  found in 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 hydrocarbon
                    concentration versus time for ESV events.  The modeled incinerator was
                    similar to the WTI unit in 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
Volume m
Appendix ffl-1                              V-10

-------
                     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 in 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 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 in 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) /3600 sec/hr = 0.21 Ib/sec; and
Volume HI
Appendix ffl-1                               V-ll

-------
                            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 in 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 in 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 in the incinerator is calculated as follows:
      50 Ib * (1-0.50)    ,-OA3/   »    (1800°F + 460°F)    .__ r3 ,
      	—;—	  * 319ft*/mole * -^	'-  = 528//3 benzene vapor
         78 Iblmole                           S20°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
Volume m
Appendix HI-1                              V-12

-------
                      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 in 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 in 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;

              •       Lack of empirical data or a suitable estimation method to quantity 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.
Volume ffl
Appendix LI-1                               V-13

-------
              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 identified in this risk assessment (See Table V-7, page
                     V-39).  This  order of magnitude is incorporated by the use of conservative
                     assumptions  in 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 in 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
Volume III
Appendix ffl-1                               V-14

-------
        of the analysis evolved into an Accident Analysis presented in Volume VII of the Risk
        Assessment. However, information relating to routine spills and storage accidents is
        retained in 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 in the organic waste tank farm, located on
       the southeast section  of the facility. Sludges are unloaded into sludge reception tanks.
       Liquid materials which arrive in containers are pumped into 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.
Volume m
Appendix ffl-1                               V-15

-------
              Organic storage tanks described in the Part B Permit Application (WTI 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.
                    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
Volume m
Appendix m-1                                V-16

-------
               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 dud hence are negligible for the purposes of this risk assessment. Instead,
               spills are evaluated in the Accident Analysis included as Volume VII 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 EPA study
              concluded that the major causes of releases from tank systems are unrelated to the
              characteristics 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 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.
Volume m
Appendix ffl-1                               V-17

-------
                     Reported releases at TSDFs, in general, are evaluated in order to determine
              typical causes associated with releases and typical volumes released.  These
              parameters are summarized in Table V-4. As shown in 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 WTI 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 in 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.

              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 WTI 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 than 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
Volume HI
Appendix ffl-1                               V-18

-------
                          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 stonnwater 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

 v s                      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 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 in waste stream and
                                                                          /-•*
v .,/
            Volume ffl
            Appendix ffl-1                               V-19

-------
              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 in 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

                   Example calculation for pump seal leakage:

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

                   Where n = the number of pumps in the process.

                   Assumptions used to Calculate Fugitive Emissions

                   •       From drawing on page 16-25 of WTI Part B Permit Application
                           (WTI1982), there are 37 major pumps. Assumed eight would be
                           working at any given time.
Volume ffi
Appendix m-1                               V-20

-------
                     •      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.

                     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.

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

                    •      Facility receives 6 truckioads 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
                           than0.1psia@100°F.
Volume ffl
Appendix ffl-1                               V-21

-------
                    •      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 in 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 difiusivity 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 10"4 (U*)12 (ScJ"0-5; U* < 0.3
                    k, (m/s) = 1.0 x lO"6 + 34.1 x lO"4 U* (ScJ"03; U* > 0.3
                    For U10 > 3.25 m/s and F/d < 14

                          where:
                                 U*  (m/s) = (0.01)(U10)(6.1 + 0.63(U10))05
                                 SCL = uL/(pLDw)
                                 F/D = 2 (A/Ti)0-5
                                 K,  = liquid phase mass transfer coefficient

                    k, (m/s) = (4.82 x  10-3)(U10)07' (Sc0)^7 (dj*»

                          where:
                                 ScG=u;(paDt)
                                 de(m) = 2(A/:r)0-5
                                 kg = gas phase mass transfer coefficient
Volume HI
Appendix m-1                              V-22

-------
                     Overall mass transfer coefficients
                            where:
                                  Keq = (H/(RT)>

                     Air emissions

                     Nfe/s^KQA

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

                     Parameters and Constants

                     -U,0 = 4.47 m/s                    Wind speed at 10 m above the liquid
                                                      surface

                     -Dw - 8.6 x 10"* cm 2/5             Difiusivity of constitutent in water
                     (selected toluene as midrange)

                     -UL = 8.93 x 10"3 g/cm              Viscosity of water

                     -pL = 1 g/cm3               Density of water

                     de = 33 ft = 10. 1  m          Effective diameter

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

                     p, = 1 .2 x 1 0'3 3/cm3        Density of air

                     D, = 0.87 cm 3/5 for toluene Difiusivity of constituent in air
Volume ffl
Appendix DM                               V-23

-------
                    H = 0.00668 atm mVgmol
             Henry's Law Constant of constituent
                    R = 8.21 x 10'5 atm mVgmol °F      Universal gas constant
                    T = 298K(25°C)

                    A = 80.1m2
            Temperature of water
            Waste water surface area
                    C0 = 10 ppm or 100 ppm
                    Q= 15,000 gal/day
                      = 6.57xlO-4m3/s
            Initial concentration of constituent in
            the liquid phase

            Volumetric flow rate
             e.      Truck Wash
                    Emissions from Truck Washing are calculated using the emission factors
             below:
Volume ffl
Appendix HI-1
                    Emission Factors (Ib/hr);
                    Valve (light) = 0.016; and
                    Valve (heavy) = 0.00051.

                    An example calculation from truck washing is as shown.

                    1 hr 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);
V-24

-------
                     •      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 in 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 in the
       TANKS2 program to represent the top 12 constituents1 (by weight) projected to be found
       in the pumpable waste stream received by WTI based on facility waste profile sheets.
             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 in Table V-7.  Fugitive
       vapor emissions from the individual types of tanks contained within the tank farm building
       are also indicated in 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
   1    The top 12 constituents by weight are estimated to comprise 60 percent of the total
       pumpabie 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 m
Appendix m-1                              V-25

-------
       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 in the permit
       application and verified- by 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 in the
       Evaporative Loss Section of the Compilation of Air Pollution Emission Factors,  AP-42
       (U.S. EPA 1993a).  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).
Volume ffl
Appendix III-1                              V-26

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

                     For container releases and *mck 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).
Volume ffl
Appendix ffl-1                               V-27

-------
                     It is believed that there is not any general trend in these uncertainties which
              would result in an overall underestimation of fugitive emissions, and conservative
              assumptions are generally made throughout this risk assessment.  These
              assumptions are summarized in 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).
              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.
Volume m
Appendix ffl-1                               V-28

-------
              a.     WTI Ash Handling System
                     In order to evaluate WTI s ash handling system, a diagram of the ash
              handling process, based on information provided in WTTs 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 Ioadout from the secondary chamber and fly ash ioadout
              from the air pollution control equipment.  Stack emissions of paniculate 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 tracked out by truck 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 trucks 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.
Volume HI
Appendix ffl-1                               V-29

-------
              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 in part to characterize the emissions
              of paniculate matter from the loading of flyash into trucks.  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.

              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 n).  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
Volume ffl
Appendix EC-,                               V-30

-------
                            collection efficiencies of paniculate matter control devices, for the
                            particle size range of 2.5 to 10 um.

              The emission calculation then becomes:

              low range:     (5300 T/YXO. 107 lb/TX2) - 1,134 Ib/yr, uncontrolled

              high range:    (5300 T/Y)(0.107 lb/TX10) - 5,671 Ib/yr, uncontrolled

              Taking into account the fabric filter:

              low range:     (1,187 Ib/yrXl-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 WIT facility is 28 Ib/yr.

      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
Appendix BI-1                               V-31

-------
                                                     Table V-l

                     POSITIVE PRESSURE AND TOTAL AUTOMATED WASTE FEED CUT-OFFS (AWFCOi)
                                           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 identifier by
           facility.
Volume III
Appenf

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

CC14
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 ffl-1
V-33

-------
Table V-3. SUMMARY PROVIDED BY PERMITTEE OF EQUIPMENT/PROCEDURAL FAILURE (UNRELATED TO INCINERATOR) WHICH
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 was'e
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
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.

The data included in this analysis were current through 1994.  A subsequent review of reported WTI Equipment/Procedural Failures through 01/01/97 was
performed. Although the updated data were not included in the risk assessment analysis, the review indicated no significant trends or impacts to the analysis
resulted from the updated data.
Volume III
Appendix III-l
                                                                 V-34

-------
      Size
      Class
     ••—•——

      Small
      Large
" "' ' —••••••••• ,±J
Matrix
Released
T iniiiH
i .iiitim
Liquid
Liquid
Solid
»«_ — — — —^m—
T in/ Air
l^iV^/iTkU
Liquid
Liquid

Liquid
Liquid

Liquid
Liquid
Liquid
Liquid
Liquid
Liquid
S^^SSSS^SSSK^— S^B
Volume
Released
40 aal
O
75 gal
55 eal
JJ (5**1
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
* »*r
SSSSSSSSSSSSSS^SSS^^SS
Released
From
Tank
Pipe
Drum
Drum
Tank

Tank
Hose

Hose
Tank Truck

Railcar
Tank
Tank
Tank
Tank
Tank
•— as— igi^^^gg^a

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 II
Mixed several incompatible wastes |
Explosion and spilled drum during consolidation of materials |
Failure to identify corrosives as reactive caused tank boil over and Q
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 I
Spilled during a railcar transfer 1
Spilled from storage tank — no reason given
Large spill occurred when a tank ruptured
Pipe connecting blend tank to incinerator corroded through 1
Arson, 2 employees arrested
Arson, 2 employees arrested
           Assumed volume.

    Key to Cause:
            1 Unknown

            3 Tank failure

            5 Equipment failure

            7 Inclement weather

            9 Operating conditions

            11 Arson
2 Transfer

4 Improper maintenance

6 Misidentified/unstable waste

8 Equipment design

10 Utility failure
    1  M       dum 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 III

-------
        Table V-5. EMISSIONS FROM ORGANIC WASTE STORAGE TANKS (HUMAN
                       HEALTH - CONSTITUENTS OF CONCERN)
Compound
Octane
Cresol
Methanol
MEK
Toluene
Acetone
Cyclohexanone
Ethyl Acrylate
Butanol
Xylene
MIBK
2-Nitropropane
Acrylonitrilc
Carbon Disulfide
Carbon
tetrachloride
Dibromoethane
Dichloroethylene
Formaldehyde
Hydrazine
Pyridine
Totolbv Tank Type

-------
            Table V-6. EMISSIONS FROM ORGANIC WASTE STORAGE (ECOLOGICAL)
Compound
Octane
Cresol
Methanol
MEK
Toluene
Acetone
Cyclohexanone
Ethyl Acrylate
Butane!
Xylene
MIBK
Chloroform
Benzene
Fotmaidehyde
Dimethylamine
Hydnzine
Acetonitrile
Carbon disilfide
Dimethylhydnzine
Total bv Tank Type
(IWyr)'
Total by Tank Type
(Bfrec)
Emissions (lb/yr)
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-03
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
S.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
O.U3
2.10
5.33
1.50
283.7
4.08E-03

Emissions
by Compound
(g/sec)
1
II
3.29E-04
1.10E-06
5.54E-04 I
4 82E-04
1.63E-04
1.09E-03
1.65E-05
1.46E-04
1.48E-05 I
1.96E-05
5.57E-05
7.51E-05
6.96E-05
6.38E-04
2.83E-04
1.63E-06
3.01E-05
7.67E-05
2.15E-05





Volume HI
Appendix ni-1
V-37

-------
         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 wain
Fugitive emissions2
TOTAL
Estimated Emissions
(Ib/yr)
50.1'
47.7'
57.2'
57.2'
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 HI
Appendix III-1
V-38

-------
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
A series of simplifying assumptions were made to allow
calculation of emissions from container processing, truck wash,
and leaks from pumps/flanges/seals
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
Simplifying assumptions based on site information and
available empirical estimation techniques
Magnitude of
Effect
low
low
low
low
low
low
low
low
low
low
Direction of
Effect
underestimate
unknown
underestimate
underestimate
underestimate
unknown
underestimate
unknown
variable*
variable*
Volume III
Appendix III-1
V-39

-------
TABLE V-8
Key Assumptions for Chapter V
Assumption
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 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
Direction of
Effect
variable*
overestimate
unknown
unknown
unknown
unknown
unknown
Volume III
Apppr»dix III-1
V-40

-------
                                                                                                        "Negligible"
                                                                               Atmosphere
                                                                                 79T/Y
Waste
     SEV rinl OMM4
                                                                      17,000 T/Y
                                                                     Slag and Ash
                               Figure V-l.  Slag and ash handling diagram—WTI facility.
  Volume III

-------
                          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, MM, (1992).

 Clark, W., R.G. Rizeq, D.W. Hansell, and W.R. Seeker, "Analysis of BIF 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.
 Achaiya, (1989a), Rotary Kiln Incineration III. An In Depth Study—Kiln Exit/Afterburner/Stack
 Train and Kiln Exit Pattern Factor Measurements During Liquid CClj 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 Bum Report for the Rotary Kiln
 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 paniculate 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).
Volume IH
Appendix ffl-1                             VI-1

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

 Ihara, Y., H. Obgame, and K. Sakiyama, "The Corrosion Behavior of Chromium in Hydrogen
 Chloride Gas and Gas Mixtures of Hydrogen Chloride and Oxygen at High Temperatures",
 Corrosion Science, 23, No. 2, pp. 67-181, 1993.

 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).
Volume ffl
Appendix ni-1                             VI-2

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

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

 Muleski, G.E. and FJ. Pendleton, Midwest Research Institute; W.A. Rugenstein. 1986.
 Measurement of Fugitive Emissions in a Coal-Fired Power Plant. Proceedings.  Sixty
 Symposium on the Transfer and Utilization of Particulate Control Technology, Volume 3.  The
 Detroit Edison Company.  November.

 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 Idlns," 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 (USEPA). 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).  1987a.  Total Mass Emissions from a
Hazardous Waste Incinerator.  Risk Reduction Engineering Laboratory.  June.
Volume ffl
Appendix m-1                              VI-3

-------
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.

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 Morristawn Memorial Hospital, Morristawn, 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 OHD980613541. 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
WTI incinerator risk issues. Risk Assessment Forum, Washington, D.C., EPA/640/R-94-001.
December.

U.S. Environmental Protection Agency (USEPA). 1994. 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 in
municipal solid waste incineration," Waste Management and Research, 4, (1986).

Von Roll Industries (VRI).  1994. Back-up data for PIC monitoring conducted during the
condition 2 rerun of the trial bum. Memorandum and attachments froih F. Sigg, VRI to G.
Victorine, USEPA. July 29.
 Volume HI
 Appendix III-1                             VI-4

-------
 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 D.J. Fournier, Jr., "Potential surrogate metals for incinerator trial burns"
 presented at 1993 Incineration (Conference, Knoxville, IN,  (1993).

 Weast, R.C., Handbook of Chemistry andPhysics,  54th Edition, CRC Press, Cleveland, OH,
 (1974).

 Wendt, lost 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 m
Appendix HI-1                              VI-5

-------
ATTACHMENT 1

-------
                                                     i. Waste  Profile No.  	
                                                     11. Cheek here if tJu* i* • reeerti£ic«tion
                          WASTE TECHNOLOGIES INDUSTRIES
                                WASTE  PROFILE SHEET
  ENERATOR INFORMATION

  .  Generator  Name:  	
   Generator  Address
   Generator  USEPA ID-
   Generator  Contact/Phone:
 -ILLING INFORMATION

  . Billing Name 	

   Billing Address:_
   Customer Service  Contact/Phone
-DIPPING  INFORMATION

~~^ PACKAGING:  Bulk Solid 	  Bulk Liquid  X   Drum Liquid  X  Drum Solid  X
   (Check 
-------
                                                         i.  Waste  Profile No
                           WASTE  TECHNOLOGIES INDUSTRIES
                                   WASTE  PROFILE SHEET
 -KOPERTIES AND COMPOSITION

 5.  Process Generating Waste:OPERATION
  .  Waste Name:  .
 10.  Identify ALL USEPA listed and characteristic waste code numbers (D, F,K, P,U) .-
     D001,  D002,  D004,  POOS,  D006,  D007,  D008,  D009,  DQ1Q,  DOll,  D018	
     D019,  D021,  D022,  D023,  D024 ,  D025,  D026,  D027.  D028,  D029,  D030	
     D035,  D036,  D036,  D039,  D040,  F001,  F002,  F003.  F004,  F005	
     To list  additional USEPA waste code  numbers,  use additional pages and check here.	

 11.  Physical State O70F:
     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 	
                                                                                           jftf-
 12.  A. pH:  s2.0	>2.0  to  <12.5 	   z!2.5 	   Not Applicable 	
     B. Strong Odor  Yes  	  No  	   Describe	  v^

 13 .  A. Liquid Flash  Point:  <73*F 	  z73»F 	  zaOO°F 	  *140°F 	  *200°F 	  N/A
     B. Boiling Point: <100°F 	  zlOOeF  X

 14.  PCBs 	  If yes,  concentration	ppm,   Pyrophoric 	  Explosive	
     Radioactive 	  Shock Sensitive 	  Oxidizer 	  Carcinogen  X   Infectious 	
     Asbestos 	 Dioxins  	  Gas 	  Bromoform >500 ppn>   X
     Dichlorodifluoromethane  >500 pptn  X   Trichlorofluoromethane >500 ppm  X   None 	

 15.  Benzene  Yes   X    No 	
     1' yes,  «) Concentration <100,OOP ppm or mg/1
           b) Does che w»te contain Mter in »n «nount greater than or equal Co iO%? Y«» __  No ___
           c! Is this vasce stream aubjecL to the control requirement* of 40 CFK 61.340 to 61.SSI?
             Yes    No
Printed:  03/16/94
                     Appendix EQ-1
                     Attachment 1

-------
  . D
                                                             i.  Waste  Profile  No.

                          WASTE TECHNOLOGIES  INDUSTRIES
                                  WASTE  PROFILE SHEET

CHEMICAL  COMPOSITION:  List AU constituents  (using specific chemical naws) present in  any
 concentration and forward available analysis. TOTAL COMPOSITION MIST EOOAL OR EXCEED 100%.
        Does your waste contain Antiaony, Arsenic. Barium. Beryllium.  Cadaiia. ChraBivm. Lead. Mercury,
        or Thallium?  Yes 	 No X

        If yes, include the metals that your waste contains in the constituent list below and specify a representative
        concentration range for each octal.
Constituents
PAINT RESIN , PIGMENTS , ADHESIVE
S . POLYMERS , INK , OIL
1,1 DICHLOROETHYLENE
1,1,1 TRICHLOROETHANE
1.1.2 TRICHLORO 1,2,2 TRIFL
Range
10-100

0-1
0-50
0-50
Units
Percent

Percent
Percent
Percent
UOROETHANE
1.1,2 TRICHLOROSTHANE
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
CHLOROBENZENE
CHLOROFORM
CHROMIUM
CITRIC ACID
CRESOLS
CRESYLIC ACID
CYCLOHEXANONE
D I CHLOROD I FLUOROMETHANE
ETHYL ACETATE
ETHYL BENZENE
ETHYL ETHER
FORMIC ACID
List additional constituents on
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
following page, check here and attach. X
x7.   Mark all  extremely  hazardous  substances  al% with  **, or  check  the following:   None
Printed:  03/16/94
                          Appendix IH-1
                          Attachment 1

-------
                                                             i.  Waste  Profile  Nc.

                              WASTE TECHNOLOGIES  INDUSTRIES
                                     WASTE  PROFILE SHEET

 16.  CHEMICAL COMPOSITION: U»t *"• constituent* (using specific chemical names! present ir. any
       concentration and forward available analysis. TOTAL COMPOSITION MOST EOUAL OR *•" M" 100%.
                         Constituents
Range
Units
ISOBUTANOL
LEAD
MEK
MERCURY
METHANOL
MIBK
N- BUTYL ALCOHOL
NITROBENZENE
0-CRESOL
ORTHO D I CHLOROBENZENE
PHOSPHORIC ACID
PYRIDINE
SELENIUM
SILVER
SULFUR 1C ACID
TETRACHLOROETHYLENE
TOLUENE
TRI CHLOROETHYLENE
TRI CHLOROFLUOROMETHANE
WATER
XYLENE
0-50
0-1
0-50
0-259
0-50
0-50
0-50
0-100
0-1
0-50
5-25
0-50
0-1
0-1
5-25
0-500
0-50
0-50
0-500
0-40
0-50
Percent
Percent
Percent
ppm
Percent
Percent
Percent
Percent * *
Percent **
Percent
Percent
Percent
Percent
Percent
Percent * *
ppm
Percent
Percent
ppm
Percent
Percent
Printed:  03/16/94
                       Appendix E3-1

-------
                                                          i.  Waste  Profile No.
                            WASTE TECHNOLOGIES  INDUSTRIES
                                   WASTE PROFILE SHEET
   8.  Check ONE:  This  Waste  is a:   Wastewater
                                                        Nonwastewater
   9 .  It this ••ice is subject to any California UK restriction*, enter the letter tram belovleither A -r B Iron section 20 next
      co each restriction that Is applicable:
      HOCs
             PCBs
                         Acid
                                Metals
Cyanides
                                                       None  X
  .'ASTE IDENTIFICATION  TABLE  (not applicable 	)

  .0 .  Identify ALL Characteristic and Listed DSEPA hazardous waste timbers that apply laa defined by 40 era 2(1)  For
      each Haste number, identify the subcategory (a* applicable, cheek none, or write in description fro* 40 CFR
      2CS.42. and 268.43) .
no
A O.I IM
•ait* Cat* 1*1




sneer 1

I. Stibcateemy
If Hot Applicable.
Si^iy Check Nooe

c . Jkppl kCAftlc TraecaeM
StaMaro*
OM«k as Applicable

Specified TecftAoieey
If Applicable, totar tka
•e era iti 4i
Table i Tretuut
eodvlei
11* .42
D sov nun 1
j
1 D001 Ignitable Liq. 261.21 (a) (1) Low TOC No No INCIN A
< 10% total organic carbon
2   D001   Ignitable Lig. 261.21(a)(1).High
   	TOC >or» 10% total org.carbon
         NO    NO
                                                                     INCIN
D002   Acid,  Alkaline, and other based on
        261 .22
                                                       No
                                                          No
                       INCIN
      D004
                                             None   Yes  No
   5  D005
                                             None  Yes  No
   6   D006
                                             None  Yes  No
   7   D007
                                             None  Yes  No
   8   D008
                                             None  Yes  No
   9   DO09   Low Mercury; less than 260 mg/kg
              mercury	
                                                    Yes   No
 10   D010
                                             None   Yes   No
 11   D011
                                             None   Yes   No
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

Printed: 03/16/94
                      Appendix m-1
                      Attachment 1

-------
                                                          i.  Waste Profile  No.
                             WASTE  TECHNOLOGIES  INDUSTRIES
                                    WASTE PROFILE SHEET

  VASTS IDENTIFICATION TABLE (Continuation) (not applicable 	)

  20 .  Identify AU. Characteristic and Listed USEPA haiardouj va«te nuaber* that apply «oTy
Inter UM Subcatroory Dticription •
If MI Applicant*.
Si^ly Owek MOM
DcicriBtiBB i Nam
StanOam
»•« araun-bMd :
Ckcek M applicable
iu.4iia> |]«a.«i(al
Sp*c>Ii*d TaeaooiaTr
If Applicabl*. tmtmi tk*
40 cn >*a.4j
Table 1 TTMtant
Ced«(«i
j*a.«j
C HIM Hut
tbr •••<• **
tun^tvl- j
t»c*r cite Letter j
rrw •*!«• '
12
13
14
Ib
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
0038
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 2CD 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
^rint«d:  03/16/94
                       Appendix ffl-l
                       A <-r^r>ViT-nant- 1

-------
                                                          i. Waste Profile  No.
                            WASTE TECHNOLOGIES  INDUSTRIES
                                   WASTE PROFILE SHEET
 WASTE IDENTIFICATION  TABLE  (Continuation)(not applicable 	}

 20   Identity AU. Characten.tic and Li.ted USEPA haxardou* «a«e number, that apply la. defined by «C CTS a«i:  For
     	K -..r. m,«~r  ia«,tifv the .ufecaterorv '•• applicable,  cheek none, or write in description tro. 40 CF*
each va.te namfter. identify the .ufecategory '•• applicable
2«(.42. and 2SS.431
1
1 * v.i trt
Mo !«»«t« Coa.m
27 D039
28 D040
29 F001
30 F002
31 F003
32 F004
33 F005
Inter Uw tuocttmforr DmcnptloB -
None
None
None
None
None
None
None


OMCK i
mTTG
No
NO
NO
NO
NO
No
No
stMUrei
i IpeeiliW T.crj.clorv .
IK **pltc«M* Int.r tM

No
No
Yes
Yes
Yes
Yes
Yes
t IM» Hu«t
CK» »••!• »•
NONE
NONE
A
A
A
A
A
To list additional USEPA waste numbers  and  categories, use following page and check  here:_

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 DQ-l
                     Attachment 1

-------
                                                                 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,  KOS
      C.  Hazard  Class or Division Explosive  (1.4C)
          Package Group _ __
      D.  Additional Description (MEK,XYLENE)
      E.  CERCLA 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

      I hereby certify  chit  ell information submitted in itema 1, 4. 7. I, », 10, 12a, 13. 14. IS and 1C cone.*in* true and
      «ccur*te description*  of this waste. 7J» fi.gnac.ur* on incoming Land Di«po»»l *«»triction Font* »c the time of
      wasce receipt will certify the information in ICOTM It. 1> and 30. Any •••pie autanictad la r«pr«a«ntative as deiined
      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 MTI to obtain a canqple
      (rent any waste shipment for purposes of recertification. Thi* information is the generator's best estimate and
      is not used as a  limitation upon NTT's receipt of vasee •Aipxmca or quantities in excess of Uteae estimated
     amounts.
                    Signature                                       Date
            Printed  (or typed)  name                             Company
                                                                      Title
Printed:  03/16/94
                        Appendix JH-1
                        Attachment 1

-------
 X   Initial
Generator's  Name:
            Waste Analysis Decision

Renewa 1      Date:	
                                                                                         ;  Hascr
 WASTE PRODUCT  REVIEW IHPORMATION
 I.  Is incineration either the required treatment technology or the basic for «ny treatment standard ler ;.ie joert
     waste cooes Is!7  tea  X   Mo __

 2 .  le this waste profile a lab pack?  Yes 	 Mo _X_

 3 .  Does WTI intend to normally use this waste as the high BTO feed for the incineration process?  Yes 	  He  x

 4 .  Here supplemental analyses performed?  Yes X  Mo 	
     II yes. supplenental analysis results are attached.
Are there any USDS sheets attached? Yes 	 Wo _X_
APPX0V&Z* SZ GK&TUMtS

laboratory

Safety

Environmental



Date : / /

Date : / /

Date : / /

Date : / /
Operations
Regulatory
	 Accepted
Conditional
fcr six 16)
Agency Use only; OEPA
	 Conditional Acceptance 	
acceptance upon Haste Management Alternatives Plan
•OTCBS from this date. / /
Accepted. Haste Kanaaenent Alternatives Plan approved.
Conditions
for acceptance or reasons for denial :
Acceptance Denied
submission. Decisicn effective |
(OEPA initials)
(OEPA initials)



Signature
Title

Date


/ /

Printed:  03/16/94
                     Appendix ffl-1
                     Attachment 1

-------
                                                             i.  Waste Profile  No.
                              WASTE  TECHNOLOGIES  INDUSTRIES
                                      WASTE PROFILE SHEET
  23.  WASTE ANALYSIS RESULTS.  Is this a miscellaneous special waste?  Yes  	  No 	


Identification

Measured
Value


Possible Waste Codes
Total
Constituent '
Analysis

TCLF

     ANIONS
     BROMIDE.  %
     CHLORIDE. %
     FLUORIDE, V
     IODINE,  %
     SULFUR.  %
     SULFIDES, ppm
     CYANIDES  (Amenable) ,  ppm
     CYANIDES  (Non-Amenable), ppm
<0-04
3.1
<0.05
N/A
.2
<20
<20
 B. METALS
D. REACTIVITY
   Water
   Acid
   Alkali
                                       CONCENTRATION
                                      SOLID     LIQUID
                                      Mg/Kg
ANTIMONY, ppm
ARSENIC, ppre
BERYLLIUM, ppr
BARIUM, ppm
CADMIUM, ppm
CALCIUM, ppm
C&tQVIiM. ppm
COPPER , ppm
LEAD, pptn
LITHIUM, ppm
MERCURY, ppm
NICKEL, ppm
PHOSPHORUS, ppm
POTASSIUM, ppm
SELENIUM, ppm
SILVER, ppm
SODIUM, ppm
THALLIUM, ppm
ZINC, pptn
C. PHYSICAL PROPERTIES
FLASHPOINT T
BTU/lb
HATER, *
VISCOSITY, cp
ASH, %
DENSITY, g/ml
pH
SETTABLE SOLIDS, %
PCB(s) , pptn
FREE LIQUID
<25
<2.5
<0.03
6
<5
2330
170
G .9
740
<5

-------
ATTACHMENT 2

-------
                                  Pumpable Feeds, WTI Facility*
4/5/95
"ionstituent
.1C*
Cresol* (cresyiic acid)
Toluene
MEK*
Methanol
Acetone* (Methyl ketone)
Cydonexanone
Ethyl acrytate
Butanor
Xytene
MJBK
Tetraehlorobenzene*
Nitrobenzene
Ethyl benzene
Pyridine
2 Ethoxyethanor
Alcohols
2 Nitropropane
Isobutanol
Dichlorobenzene*
Creyslicacid
heptane*
Benzene
Trichloroethane*
Carbon
vckshexane
/ilor. paraffin oils and waxes
Tetrahydrofuran
Diethyl phthalate
Creosote
1,40ioxane
Carbon tetrachloride
Formaldehyde
Trichioroethytene
Cumene
Ethanol
Naphthalene
Chloroform
Tetrachtoroethytene
1 ,12 Trichtoro 1^2 triflouroethane
Phenol
Dinitrotoiuene*
Acetonrtrite
Chlorobenzene
Isopropanol*
Methyl methacrylate
TCFM
Formic acid
Acetophenone
Mateic anhydride
DtchlorodrBuoromethane
curfural
ssoranol
Benzidine
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
65350
59443
58810
57915
57438
55116
Percent of Total
16.8
5.2
4.0
3.5
3.1
2.9
2.5
2.4
2.4
2.3
2^
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
* Based on waste profiles for first year of operations.
                                       Appendix ni-1
                                       Attachment 2

-------
                                  Pumpable Feeds, WTI Facility*
                                                  4/5/95
Constituent
             Total Ib/yr     Percent of Total
Calcium chromate
aytonttrile
t)imethytphenof
bptcniOfonyunn
N-nitrosodiethanoiamine
Dimethylphlfuiidte
Toluenediamine
• Tetrachloroethane*
Toluene diisocyanate
dichtoroethene-
Chlorinated fluorocarbons (assumed dichtorodifluoroethane)
Carbon disutfide
Phthafc anhydride
Dimethyiamine
Trichkxobenzene*
Dimethyl phthalate
2 Acetyiaminofluorene
1 2 Benzenedicarboxylic acid
butyl acetate*
tJ •! n it frtfififH/mn Iki i no
i viiiu w&w^yi loiiuiric
2 Naphthyiamine
Hydraane
Crotonaldehyde
Dimethyl sulfate
Dichloroethane* (Ethylidene dichloride)
1 Naphthylamine
Aniline
osafrote
Oimethylhydrazine*
dfcromoethane*
tit nltl-r- r ,-• fllnttlul l»»l'l
N-nitrosooietnyiamine
Benzo(a)pyrene
Chrysene
N-nitrosodt-rv-butylamme
Fluoranthene
Indeno (1.2,3-cd) pyrene
3 Methyteholanthrene
2Picofaie
3,3' Dimethyfcenzidine
1^,5,6 Dibenzanthracene (dibenz (a,c) anthracene)
r> t't h i
• i.iuopnenoi
1 Methyttxitadiene
2 AcetyJaminofluorene
Diethyl stitoestrol
Dihydrosafrote
^ ~*~ *" jl Liwi.*BL_i«.i«.«jLWji.n«ji.n_ii-
* ^JU 1 rCU ty Ml I III HMUOUtSI UJStlK
ParaUehyde
N-nttroso-n-methylurethane
N-nitroso-n-methylurea
Safrote
Phenacetin
Dibromomelhane* (Methytene bromide}
Nrtro-o-toiuidin6
54606
54259
53872
C^COO
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
31397
31245
31222
30854
30854
30790
30769
30769
30768
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.3
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
0.2
0.2
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02
'02
   Based on waste profiles for first year of operations.
Appendix ni-1
Attachment 2

-------
                                 Pumpable Feeds, WTI Facility*                          4/5/95
Constituent
4 Bromophenyl phenyl ether
diphenylhydrazine*
Streptozotocm
Reserpine
Amitrole
Cyanogen bromide
Mateic hydrazkte
Hydrogen fluoride
Auramine
N,r4-diethylhydrazine (1 ,2-diethylftydrazine)
3,3' Oimethoxybenzidine
L-serine
cthyiene dibromide
22 Bioxirane
Acetates
Acrytamide
Ethyl methanesulfonate
Methacrytonitrile
Malononitrite
Barium salt
Daunomycin
Acrylic acid
Ethytene glycol
Hexachloro 1,3 butadiene*
-* Dichlorophenol*
*cetaldehyde
^ Porysitoxanes
Hexachloroethane
2 Methoxyethanol
Chlordane
Lead salt
Dibromoethane
Calcium salt
Chtorophenor
Butenal*
Citric acid
Lindane
Benzenesulfonyl chloride
«.- i
LJipfopyutfntne
N-butytamme
2 Heptanone
Amyl acetate
trichioroethane
Pentachlorobenzene
dichloroptopene'
2,4 D sate and esters
Acetic acid
2 Chloroethyl vinyl ether
1 .2 Benzanthracene (benzo (a) anthracene)
4,4' Methytenebis(2-chloro)aniiine
Chlorometnyl methyl ether
dichlorobutene*
•< Pentachtoroetnane
3,3' Dichlorobenzidine
Total Ib/yr
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
14696
14464
13087
13073
13073
13073
12855
12570
11594
10405
10180
9275
9169
8929
8928
8928
8928
8928
Percent of Total
0.2
02
02
02
02
02
02
0.2
0.2
0.2
0.2
0.2
02
02
02
02
0.2
0.2
0.1
0.1
0.1
0.1
0.1
C.'
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
,, o.o
0.0
0.0
                                     Appendix ffi-1
Based on waste profiles for first year of operations.  Attachment 2

-------
                                Pumpable Feeds, WTI Facility*
                                                    4/5/95
Constituent
b-chtornaphthaiene
-chtoro-m-cresol
Hexachloropropene
chloral (Trichloroacetaldehyde)
Antimony salt
Arsenic salt
Ethyl methacryiate
Pronarnide
Tris (2,3 dibromopropyl) phosphate
1,2 Dibromo-3-dichtoropropane
Acetyl chloride .
Benzotnchloride
Chlorambucil
Chtomaphazine
Cydophosphamide (2H-1,3,2 Qxazaphosphorin-2-amine)
Melphalan
O-toluidine hydrochloride
DDT
Benzamine, 4 chloro 2 methyl-
Dichloromethoxy ethane
DimethytcarbamoyI chloride
Hexachlorophene
Thallium chloride
DDD
Dichloroisopropyl ether
Pentachloronitrobenzene
Kepone
etones*
r^ropanol"
Thiourea*
Ethytene Oxide* (Oxirane)
Heptachtor
TribrornornGthdrre
Copper
Mixed organics (ale., amtne, etc.)
Ethyl ether (Diethyl ether)
Cydoheptane
octanone*
2 Hexanone
Decanes
3Pentanone
2 Pentanone
Styrene
2 Octanone
Phosphoric acid
Furfuran (furan)
Anhydride
Di-n-propylnitrosamine
Acetaktehyde" (Ethanal)
Vinyl chloride
Trichtorofluoroetrtane
Ditaenzo(a,i)pyrene
("\ /s_^4^ jn^fKu M i4K MM\h0%g>%hafa>
w,w~uicmiyi"tf*uicuiymuiiuf7nu9^nau;
tethanethiol (Thiomethanol)
Total Ib/yr
8928
8849
8714
8708
8687
8687
8683
ft£CQ
oooy
flCCO
OOwS
8582
8582
8582
8582
8582
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
AAACk
*^^9
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
ft n
w.U
' 0.0
Based on waste profiles for first year of operations.
Appendix ffl-1
Attachment 2

-------
                               Pumpable Feeds, WTI Facility*                          4/5/95
Constituent
.demylthiouracil
Saccharin and salts
Selenium sulfide
1,3 Propane suttone

Lasiocarpine
Thiram
Trypan blue
Thioacetamide

Water
Trichtorophenol*
Suffuric acid

ois(Z)ciuoroeinyi etner ( uicnioroeuiyi emer;
Lead subacetate
Zinc
Hexachtorocydopentadiene
Lead acetate
Matonrtrfle hydroazide*
Chrome compounds
2,3,4,6 Tetrachlorophenot
4,4' Methylenebis(2chloro)benzeneamine
Potassium hydroxide
Benzal chloride
Sodium bicarbonate
Chlorobenzilate
Diallate
' ithium amide
entachlorophenol
r* t
oromoforrn
Pyrene
1,2Ethanedrytbiscarbamodithioicac3d
Potassium-t-butoxide
Mitomycin C (Azirino)
Methoxychtor
Barium
1 r+f+fl
Lead
Endrin
N-nitroso-n-ethylurea
Creosol
Phosphorus sulfide
Thioacetamine'
Lead phosphate
Chromium
Thallium carbonate
Thallium acetate
Thallium nitrate
~K ^ . J tx. %
Bis(2)cnloroisopropyl ether (Dienioroisopropyi etner)
2,4 Dichloroohenoxyacetic acid (2,4-D)
2,4,5 TPfSilwex)
Sodium suffide
Toxaphene
Ethylene bis-dHhiocarbamic acid
Silver
ethyl ethyl tetone pcroxids
Oddmuifn
Total Ib/yr
4276
4276
4276
4276
4261
•V&W 1
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
0.0
0.0
"oo
0.0
                                    Appendix IH-1

                          .   _^     Attachment 2
Based on waste profiles for first year of operations.

-------
                              Pumpabte Feeds, WTI Facility*
                                                   4/5/95
Constituent
Arsenic
>etentum dioxide
PCBs
CacodySc acid
Mercury
Selenium
Nitrophenol*
Dioctyl phtnalate
Dichloromethane * (Methytene chloride)
Hexachlorobenzene
Dipropylnitrosamine
N-propyiamine (1-aminopropane)

ifmiiuiiiiiiil
N-fnethyt-n'-nrtro-n-nitrosoguanidine
Phenanthrene
Ethyl carbamate (urethane)
Methyl chtorocarbonate
7.12 Dimethyl benr(a)anthracene
Azaserine
Sym-trinitrobenzene
Ethytene thiourea
MethapyriJene hydrochloride
2 Chloronaphthalene
Bis(2 chtoroethoxy)methane
Bis{2)chtoroethyl ether CDtchtoroethyl ether)
Dirnethylcarbamyl chloride
Benz(c)acridine
,2 Propanediol
1 ,3 Pentadiene
1 .2,7,8 Oibenzopyrene
1, 2,3,4 Oiepoxybutane
4 Chtoro-o-toluidine hydrochloride
Hydrochloric acid
lodomethane* (Methyl wdide)
Total of cmpds remaining after revision
Summary of Values
total IDS cmpds containing analytes
total tos pumpabte feed, less ash & baits
mass %
ash IDS
total (with ash)
mass % (pumpabtes)
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
0
0
19143596
Theoretical
2544153
19127250
64
1498252
20625502
69
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
100.0
Actual
2507702





Based on waste profiles for first year of operations.
Appendix ffl-1
Attachment 2

-------
                            Non-pumpabie Feeds, WTI Facility*
                              4/5/95
     Constituent
Total (b/yr    Percent of Total
Xytene
Isobutanol
he*
Hexane
Toluene
MEK
MIBK
Acetone
Ketones*
Pyridine
trichloroethane
Ethyl acetate
Aniline
oesol*
Ethyl benzene
2 Nitropropane
Benzene
Formaldehyde
Chlorinated fluorocarbons
Tetrahydrofuran
Chlorobenzene
AcetakJehyde
N-nitrosopyrroSdine
Benz(c)acridine
Cacodylic add
Mitomycin C (Azirino)
Nitrobenzene
Cydohexane
2 Ethoxyethanol
Actylicacid
Donethylhydraane*
Ethyl methacrytate
Dt-n-propytnitrosamme
Dipropytarnine
Lasioearpine
Lead phosphate
Lead acetate
Chlorinated polymer resins (PVC)
Butanol
Dichlorobenzene*
Phthatic anhydride
Cresylicaead
Acetonitrite
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
12.2
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 ffl-l
                                   Attachment 2

-------
                            Non-pumpable Feeds, WTI Facility*
                                                 4/5/95
Constituent Total Ib/yr Percent of Total
Chrome compounds
Dimethyiamine
Cydohexanone
1,40ioxane*
1 Naphthylamine
1.2 Benzenedicarboxylic acid
2 Naphthylamine
Dimethyl sutfate
Sodium hydroxide
Antimony salt
Arsenic salt
Metnanol
Carbon disutfkte
Acetic acid
Tetrachloroethylene
Trichloroethytene
1,1,2 Trichloro 1 22 triflouroethane
Formic acid
Carbon tetrachtoride
Acetophenone
Dimothvlnhthalatp
uiicuiyif/iiiiioioic
Calcium salt
Lead salt
Vinyl chloride
Citric acid
Benzal chloride
Methoxychtor
dtchloropropene*
Chtorobenzilate
Diallate
Tetrachloroethane*
Daunomyon
Chloroform*
2,4 Dichlorophenoxyacetic acid (2,4-D)
4,4* Methytenebis<2-chtoro)anaine
Carbamicacid
Trichlorobenzene*
Isosafrote
Toluene diisocyanate
Toluenediamine
dmitrotoluene*
Furfural
Dffnethyi phthalate
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
8669
8669
7213
7086
6865
5012
5012
5011
4648
4612
4545
4545
4305
3816
3269"'
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
0.2
0.2
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 IU-1
Attachment 2

-------
                           Non-pumpable Feeds, WTI Facility*
                              4/5/95
     Constituent
Total Ib/yr   Percent of Total
Metnacrytonrtrite
Sulfuricacid
Furfuran (furan)
Resorcinol
Hexachtoro 1.3 butadiene*
— • — *k
ChtorodibnSrtiofTieuiane
Chromic acid
Bromoform*
Acetates
Potassium-t-butoxide
Acrytemide
Bis(2)chloroethyt ether* fDichloroettiyl ether)
_ 	
Ethyl metnanfesunoiiaie
Ethytene Oxide* (Oxirane)
Mateic anhydride
J'-iL.
N-niuosodietnanoiamine
N-nitrosopiperidine
P*dimethvlaminoazobenzene
Thallium nitrate
Thiourea
Trypanblue
Diethyl phthalate
Diemyl sttbestrol
Dihydrosafrote
Kepone
dtchloroethane*
Heptachlor and hydroxide
Dichloromethoxy ethane
Chtordane
dichtoroetnene*
Pentachtorophenol
Pentachloronitrobenzene
Tetrachlorobenzene*
P-chtoro-m-cresol
Calcium suffide
Thiram
rVliiiiiiin nth ^nn
utDromoetnane
Methanethiol (Thiomethanol)
Hexachtoropropene
Benzenesulfonyl chloride
ODD
Dm* utn
Procurnide
Thallium chloride
3269
2894
2552
2540
2505
2138
1961
1818
1403
1353
1276
1276
1276
1276
1276
1276
1276
1276
1276
1091
1082
1047
1024
1024
894
881
858
813
761
723
707
705
581
500
438
422
403
339
251
251
251
251 '
251
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
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 operation.
                                   Appendix III-l
                                   Attachment ?

-------
                            Non-pumpable Feeds, WTI Facility*
                                                  4/5/95
Constituent
Thioacetamide
Benzamine, 4 chloro 2 mettiyt-
Pentachlorebenzene
DichlOFOphenot*
4-Chloro-o-toluidine CBenzamine, 4 chloro 2 methyl-)
N-nitroso-di-n-ethanolamine
N-nitroso-n-ethyfurea
Tris (2,3 dibromopropyl) phosphate
Toxaphene
Ethyl ether (Diethyl ether)
2 Acetytaminofluorene
Dichlorodifluorometnane
5 Nitro-o-toluidine
Bis<2 ethytoexyl)pnthaJate
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
Sased on waste profiles for first year of operation.
Appendix ni-1
Attachment 2
                                                         10

-------
ATTACHMENT 3

-------
 ATTACHMENTS: PREDICTION
 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.      Thermodynamic 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 in 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 hi 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
Attachment 3

-------
       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
Attachment 3

-------
           Table 3-1. Speciation Results From Thermodynamic Modeling
T
SR"
a
Element
Al




As


Sb



Ba




Be


Cd






Cr

Cu





Pb







Hg

c

mol/min
Chemical
Form

AI203
AlaSiOs
AICI
AlClg

ASjOs
AsO

Sb205
SbO2
SbO

BaCI2
Ba(OH)2
BaS
BaO

Be(OH)2
BeO

Cd
Cdd
CdCfe
CdO
CdSO4
CdOSiO2

Cr203

Cud
CuaClg
CuzO
Cu
CuO

PbO
Pbd
PbCfe
PbS
PbOz
Pb304
Pb

HgS04
1400
1
175


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
100(
1
17J
3 1200
0.8
5 175
120(
1.J
17;
3 150
I 1.5
5 175
200 40(
1.5 1.i
175 17£
3 1200
5 1
5 0
150
1.5
0
of element having the listed form, (mole percent)

95
C
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
"00

27
1
64
5
0
3

100

0
0
0
100
0

0
1
0
1
0
0
98

0

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
o a

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 III-l
Attachment 3

-------
              Table 3-1.  Speciation Results From Thermodynamic Modeling
                                   (Continued)



Ni




Se


Ag



Tl



Zn


HgCfe
Hg
HgO

NCI
NiCfe
Ni
Ni3S2

SeO
SeOa

Ag
AgCI
Ag2S04

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 ffl-1
      Attachments

-------
ATTACHMENT 4

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
CAS No.
Chemical Name
Emission Rates, g/s
1A 1B 2 3A 3B

4A 4B
MckXIOlUltk
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
111-44-4
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-Dinitrctoluene
2,6-Dinitrotoluene
3,3'-Dichlorobenzidine
Acrolein
Acrylonitrile
Benzene
Benzo(a)pyrene
Benzo(b)fluoranthene
Bis(2-chloroethyl)ether [WTI PIC LIST]
Bis(2-chloroisopropy!)ether [WTI PIC LIST]
Bis(2-ethylh8xyl)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 3.31 E-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

-------
Attachment 4. Estimated Emission Rates for Organic Compounds
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
Oibromochloromethane [WTI PIC LIST]
Ethylene dibromide
Ethylene oxide
Fluorene [WTI PIC LIST]
Formaldehyde
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroetharie
Isophorone [WTI PIC LIST]
Methylene chloride
N-Nitrosodiphenylamine [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-dichloroethy lene
(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 IB 2 1A "o5
1.24E-04
2 63E-01;
b. UVIIw WVJ
1.I5E-04
3.05E-05

6.07E-04
2.57E-05
6 23E-OQ
V.fcwt \J+3
1 01F-04
1 .v 1 t *J*t
8.09E-06
7 54E-Oei
1 .«™Tt \J\J
5.20E-04


1.05E-02 1.76E-04 3.95E-04
4.37E-04

6.43E-05
5.65E-05 1 29E (W
1 .C9C~UD


5.48E-05

	 	 	 — 	 	 — —
1.29E-05
1 2QF-OC
I *b^^ VSO
1 ?QFJ1«;
1 >b«7C^JQ
4.45E-05
3.30E-04

=— «— — — .^— —
4A 4B
6.69E-06



6.69E-06


6.69E-06
6.69E-06
6.69E-06
6.69E-06
6.69E-06

6.69E-06







6.69E-06


• 	 	 	
• 	 	






Maximum
Emission
Rate, g/s
1.24E-04
2.63E-05
1.15E-04
3.05E-05
6.69E-06
6.07E-04
2.57E-05
6.69E-06
1.01E-04
8.09E-06
7.54E-05
6.69E-06
5.20E-04
6.69E-06
O.OOfc+00
1.05E-02
4.37E-04

6.43E-05
5.65E-05
O.OOE-fOO
6.69E-06
O.OOE+00
5.48E-05
—

1.29E-05
O.OOE+00
1.29E-05
1.29E-05
O.OOE+00
4.4SE-05
3.30E-04
  Appendix III-l
  Attachment 4

-------
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
119-904
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-Dinrtrobenzene
1,4-Dioxane
1 -Methyl-2-pentanone [WTI PIC LIST]
2,3,4,6-Tetrachlorophenol
2,4,5-Trichlorophenol
2,4-D
2,4-Dichlorophenol
2,4-Dimethylphenol
2.4-Dinitrophenol [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-Nitroaniline
2-Nitrophenol [WTI PIC LIST]
2-chloronaphthalene
3,3'-Dimethoxybenzidine
3-Nrtroaniline [WTI PIC LIST]
4,6-Dinitro-2-methylphenol [WTI PIC LIST]
4-Chloro-3-methylphenol [WTI PIC LIST]
4-Methyl-2-Pentanone
4-NKroaniline [WTI PIC LIST]
4-Nitrophenol
Acenaphthene [WTI PIC LIST]
Acenaphthylene [WTI PIC LIST]
Emission Rates, g/s
1A IB 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
8-91E-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
Rato, 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-06
6.43E-05
O.OOE+00
6.69E-06
O.OOE+00
6 43E-05
v/«^WI™t \J*J
4 18E-05
~. 1 \J\^ \J*J
6.69E-06
6.69E-06
6.69E-06
6.69E-06
1 i*>F.fvi
I.I OC-UH
6.69E-06
6.69E-06
6.69E-06
fi 43P n*\
Q.HOC.-UO
6.69E-06
6.69E-06
6.69E-06
6.69E-06
  Appendix III-l
  Attachment 4

-------
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 v
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
Benzole 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)octyl 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
Maximum
Emission
3B 4A 4B Rate, a/s
3.01 E-04
2.90E-03
2.93E-04
6.69E-06 6.69E-06
O.OOE+00
6.69E-06 3.42E-05
O.OOE+00
6.69E-06 6.69E-06
O.OOE+00
6.69E-06 6.69E-06
1.13E-05 1.13E-05
3.20E-05
O.OOE+00
O.OOE+00
6.69E-06 6.69E-06
O.OOE+00
O.OOE+00
2.28E-05
6.69E-06 6.69E-06
2.40E-04
4.83E-04
3.68E-05
1-29E-05 1.29E-05
6.69E-06 6.69E-06
1.39E-04
O.OOE+00
6.69E-06 6.69E-06
6.69E-06 6.69E-06
2.50E-05 2.50E-05
2.19E-04
O.OOE+00
2.23E-05 4.60E-04
    Appendix III-l
    Attachment 4
4
f

-------
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 thlourea
Ethylidene dichloride
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
Pentachforonrtrobenzene
Pentachlorophenol
Phenanthrene [WTI PIC LIST]
Phenol
Phosgene
Propionaldehyde
Propylene dichloride
Pyrene [WTf PIC LIST]
Quinollne
Quinone
Safrole (5-(2-Propenyl)-1 ,3-benzodioxole)
Selenium
Styrene
Toluene

^•^vi^^^^MlMiH
Emission Rates, q/s
1A IB 2 3A 	
1.68E-04
1.60E-03 2.69E-05
1.46E-10

1 1QF f\A
1 . 1 Z7C-U*r
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 R1 F m
1 .U 1 C-Uu
4.76E-05
3.37E-05
6.94E-06
3.39E-10
31 6E-OA
• 1 vlfV0*T


3.39E-06


1.15E-04
1.28E-08
2.44E-05 1.66E-05
4.08E-03 1.96E-04



OD 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+00
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+00
1.21E-04
3.45E-04
1.61E-03
6.69E-06
4.76E-05
3.37E-05
6.94E-06
6.69E-06
3.16E-04
O.OOE+00
O.OOE+00
O.OOE+00
6.69E-06
O.OOE+00
O.OOE+00
1.15E-04
1.28E-08
2.44E-05
4.08E-03
  Appendix III-l
  Attachment 4

-------
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-DinKrobenzene
o-Xylene (o-Dimethyi benzene)
p-Cresol
p-Dinitrobenzene
p-Xylene (p-Dimethyl benzene)
-Hexachlorocyclohexane
-Chloronaphthalene
-Hexachlorocyclohexane
Maximum
Emission Rates, g/s Emission
1A lb z 3A 3B 4A 4B Rate q/s
4.80E-04 4 8QE-04
~ . OW^"\J*T
1.10E-04 i 10E.Q4
I • 1 V/fc" W*T .
O.OOE+00
O.OOE+00
OOOE+00
• W^ I \J\J
O.OOE+00
O.OOE+00
o Q7C nc
O.Or CL-UO g Q7P.f)<%
6.69E-06 6.69E-06
OOOE-4-00
V. V/Wl^ f\J\J
O.OOE+00
O.OOE+00
O.OOE+00
O.OOE+00
             Total

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
   Attachment 4

-------
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    Under this category, the characteristics of the waste fed to the
        incinerator contributed to the  WFCOs.


FLOW   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.

LMT     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 III-l
Attachment 5

-------
                                 WASTE FEED CUTOFF PERFORMANCE SUMMARY
                                        WASTE TECHNOLOGIES INDUSTRIES
                                              EAST LIVERPOOL, OHIO

Description
CLNK
CTRL
ESP
FEED
FLOW
LMT
MAIN
MISC
SCRB
SO
Total
Hazardous Waste Hours
Avg. Hours Between WFCOs

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
36
295
8.43

1/94
3
4
1
5
12
1
0
3
1
0
30
352
11.73
Date
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
Soured:  Waste Technologies Industries, Report of AWFCO Incidences to Ohio EPA, 1994.

The data included in this analysis were current through 1994. A subsequent review of reported WTI Equipment/Procedurtl Failures
through 01/01/97 was performed. Although the updated data were not included in the risk assessment analysis, the review indicated
no significant trends or impacts to the analysis resulted from the updated data.
 Appendix IH-1
 Attachment 5

-------
ATTACHMENT 6

-------
                                                             TANKS  PROGRAM  2.0                                                  10/16/95
                                                  EMISSIONS  REPORT -  DETAIL FORMAT                                        PAGE  ^
                                      TANK  IDENTIFICATION AND  PHYSICAL  CHARACTERISTICS
 Identification
   Identification No.:     B&H
   City:                 Pittsburgh
   State:                PA
   Company:              WTI
   Type of Tank:          Vertical Kixed 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

-------
TANKS PROGRAM 2.0
EMISSIONS REPORT - DETAIL FORMAT
LIQUID CONTENTS OF STORAGE TANK
Liquid
Daily Liquid Surf. Bulk
Temperatures (deg F) Temp. Vapor Pressures
Mixture/Component Month Avg. Hin. Max. (deg F) Avg. Min.
All 51.75 47.02 56.48 50.32 0.5354

Octane (-n)
Cresol (-0)
Toluene
Methyl alcohol
Methyl ethyl ketone
Acetone
Water
Cyclohexanone
Ethyl acrylate
Butanol-(1)
Methyl isobutyl ketone
Chloroform
Benzene
Formaldehyde
Hydrazine
Nitropropane(H-)
Acetonitri le
Carbon disulfide
Dimethyl hydrazine (1,1)
Dimethylamine
Xylene (-0)
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
10/16/95
PAGE 2
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
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

.0585
.0481
.1575
.0417
.0404
.0402
.0365
.0040
.0078
.0045
.0017
.0143
.0035
.0020
.0015
.0020
.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 1
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 2:
Option 1
Option 1
Option 1
Option 2:
Option 2:
Option 2:
Option 1
Option 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,
8=1209.600,
8=1210.595,

8=2137.192,
6=1897.011,
8=1362.390,
8=1168.400,
8=929.440,
8=1211.033,


8=1314.400,
8=1169.110,
8=1305.910,
8=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=241.590
C=225.530
C=213.690

Appendix III-l
tf  \ment 6

-------
               /
               V
                                                                  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:             0.97H409

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 Sola? 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

-------
                                                             TANKS  PROGRAM  2.0                                                  10/16/95
                                                  EMISSIONS  REPORT -  DETAIL  FORMAT                                        PAGE  4
                                                      DETAIL  CALCULATIONS  (AP-42)
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
          A^  \nent 6

-------
                                                   TANKS PROGRAM 2.0
                                           EMISSIONS  REPORT -  DETAIL FORMAT
                                           INDIVIDUAL TANK EMISSION  TOTALS
                                                                                     10/16/95
                                                                                     PAGE 5
Annual Emissions Report
Liquid Contents
Losses (Lbs.):
Standing   Withdrawal
Total

Octane (-n)
Cresol (-0)
Toluene
Methyl alcohol
Methyl ethyl ketone
Acetone
Water
Cyclohexanone
Ethyl acrylate
Butanol-(1)
Methyl isobutyl ketone
Chloroform
Benzene
Formaldehyde
Hydrazine
Ni tropropane(2-)
Acetonitri '.e
Carbon disulfide
Dimethyl hydrazine (1,1)
Dimethylamine
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

-------
                                                              TANKS PROGRAM 2.0                                                   10/16/95
                                                  EMISSIONS  REPORT  -  DETAIL  FORMAT                                        PAGE  1
                                      TANK  IDENTIFICATION AND PHYSICAL  CHARACTERISTICS
Identification
   Identification No.:     P/0
   City:                 Pittsburgh
   State:                PA
   Company:               UTI
   Type of Tank:          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
           Attachment 6

-------
TANKS PROGRAM 2.0
EMISSIONS REPORT - DETAIL FORMAT
LIQUID CONTENTS OF STORAGE TANK
Liquid
Daily Liquid Surf. Bulk
Temperatures (deg F) Temp. Vapor Pressures
Mixture/Component Month Avg. Min. Max. (deg F) Avg. Min.
All 51.75

Ace ton 1 1 r i I e

uenzene
Butanol • ( 1 )
Chloroform
Cresol ( ~o)
Cyclohexanone
Dichloroethane (1,1)
Dimethylamine
Ethyl acrylate
Methyl alcohol
Methyl ethyl ketone
Methyl isobutyl ketone
Octane (-n)
Tot UGHC
Trichloroethane (1,1,1)
Trichloroethylene
M-j-f- pp
Wd i Cl
Xylene (-o>
47.02 56.48 50.32 0.4227
2.3582
0.8929
0.9248
0.0383
1.9471
0.0011
0.0411
2.3582
14.7000
0.3753
1.1066
0.8537
0.1582
0.1228
0.2540
1.3182
0.6356
0.1919
0.0525
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
1.1546
0.5469
0.1612
0.0439
Vapor Liquid
(psia) Mol. Mass
Max. Weight Fract.
Vapor
Mass
Fract.
Mol.
Weight
10/16/95
PAGE 2
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
Opt i on
Opt i on
Opt i on
Opt i on
Option
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
Opt i on
2:
2:
2:
2:
2:
2:
2:
1
1
2:
2:
2:
2:
1
2:
2:
2:
1
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,
6=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=219.480
C=302.800
C= 192. 700
C=213.690
Appendix III-l
Attachment 6

-------
                                                                   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):       49.8K730
  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 Ambient Tjemp.  Range  (deg.R):             19.20
               "  "
Appendix III-l
Attachment 6

-------
                                                              TANKS  PROGRAM  2.0                                                   10/16/95
                                                  EMISSIONS  REPORT  - DETAIL  FORMAT                                        PAGE  4
                                                       DETAIL  CALCULATIONS  (AP-42)
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):                     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

-------
                                                  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
Acetoni tri le
Benzene
Butanol-(1)
Chloroform
Cresol (-0)
Cyclohexanone
Oichloroethane (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
Water
Xylene (-0)
Total •
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
3.11
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
1269.47
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
1272.58
         Appendix III-l
         / "nchment 6

-------
                                                              TANKS  PROGRAM 2.0                                                   10/16/95
                                                  EMISSIONS  REPORT  -  DETAIL  FORMAT                                        PAGE  1
                                      TANK  IDENTIFICATION AND PHYSICAL  CHARACTERISTICS
I denti ficat ion
   Identification No.:     Rec
   City:                 Pittsburgh
   State:                 PA
   Company:               WTI
   Type of Tank:          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

-------
TANKS PROGRAM 2.0
EMISSIONS REPORT - DETAIL FORMAT
LIQUID CONTENTS OF STORAGE TANK
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
Acetonitri le
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)
47.02 56.48 50.32 0.
2.
0.
0.
0.
1.
0.
0.
2.
4227
3582
8929
9248
0383
9471
0011
0411
3582
14.7000
0.
1.
0.
0.
0.
3753
1066
8537
1582
1228
0.2540
1.3182
0.6356
0.1919
0.0525


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
Cpsia) Mol. Mass
Max. Weight Fract.
Vapor
Mass
Fract.
10/16/95
PAGE 2
Mol. Basis for Vapor Pressure
Weight 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 Option
41.05 Option
78.11 Option
74.12 Option
11,9.39 Option
168.14 Option
98.20 Option
98.97 Option
45.08 Option
100.11 Option
32.04 Option
72.10 Option
100.20 Option
114.23 Option
92.13 Option
133.42 Option
131.40 Option
18.00 Option
106.17 Option

2:
2:
2:
2:
2:
2:
2:
1
1
2:
1 .
2:
2:
1
2:
2:
2:
1
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,
6=1474.080,
8=1209.600,
8=1168.400,
8=1344.800,
B=21?6.600,
6=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=219.480
C=302.800
C= 192. 700
C=213.690

Appendix III-l

A^chment 6
 (   \
12

-------
                                                                   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 /(lb-mole-deg R)):                10.731
  Liquid Bulk Temperature (deg.  R):              509.99
  Tank Paint  Solar Absorptance (Shell):            0.17
  Tank Paint  Solar iftbsorptance (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):   t    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

-------
                                                             TANKS  PROGRAM 2.0                                                  10/16/95
                                                  EMISSIONS  REPORT  -  DETAIL  FORMAT                                       PAGE  4
                                                      DETAIL  CALCULATIONS  (AP-42)
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):                    1331.6814
  Vapor Molecular Weight (tb/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
          /f  ^hment 6                                                   14.

-------
                                               TANKS PROGRAM 2.0
                                       EMISSIONS REPORT - DETAIL FORMAT
                                       INDIVIDUAL TANK EMISSION TOTALS
                                              10/16/95
                                              PAGE 5
Annual Emissions Report
Liquid Contents

Acetone
Acetonitri le
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)
Trichlcroethylene
Water
Xylene (-0)
Losses (Ibs.
Standing
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
Withdrawal
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. n
6.95
Total
1334.79
387.67
10.57
24. 4(1
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

-------
      APPENDIX III-2

 SUMMARY OF MEASURED
      DIOXIN/FURAN
CONGENER EMISSION RATES

-------
APPENDIX 111-2
Measured Dioxin/Furan Congener Emission Rates
August 1993 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 1993 ECIS Performance Test
Run 1
<2.32e-ll
1.72e-10
2.10e-10
2.86e-10
2.10e-10
1.61e-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
I.67e-09
1.83e-08
Run 2
<2.18e-ll
1.65e-10
2.02e-10
3.40e-10
2.23e-10
1.54e-09
5.84e-09
1.38e-10
7.43e-10
1.06e-09
3.34e-09
3.02e-09
3.87e-09
7.96e-10
1.59e-08
2.28e-09
2.39e-08
Run 3
<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
1.13e-09
5.93e-08
9.16e-09
2.69e-07
Run 4
< 3.726-11
1.65e-10
2.23e-10
2.92e-10
1.91e-10
l.Ole-09
3.24e-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
1.81e-09
1.70e-08
Run5
< 3. 18e-ll
1.67e-10
2.37e-10
3.83e-10
2.216-10
1. 51e-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-

-------
APPENDIX III-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 (g/sec) — February 1994 Trial Burn
Run 1
<2.86e-ll
6.83e-ll
1.18e-10
2.23e-10
1.18e-10
1.32e-09
4.04e-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
1.60e-09
1.46e-08
Run 2
< 3. 98e-ll
6.486-11
9.44e-ll
1.75e-10
l.Ole-10
l.Ole-09
2.43e-09
l.lSe-10
3.98e-10
4.86e-10
1.62e-09
1.55e-09
1.696-09
3.44e-10
l.Ole-08
1.28e-09
8.77e-09
Run 3
<2.28e-ll
5.516-11
8.06e-ll
1.68e-10
8.066-11
8.06e-10
2.22e-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
5.95e-ll
7.27e-ll
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
8.59e-10
5.22e-09
Volume III
Appendix III-2
-2-

-------
APPENDIX III-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
2.61e-08
2.31e-10
6.19e-10
6.79e-10
l.42e-09
1.42e-09
1.57e-09
4.63e-10
8.21e-09
1.12e-09
7.31e-09
Run 3
<2.20e-ll
3.42e-ll
7.36e-ll
1.21e-10
6.07e-ll
7.59e-10
2.28e-09
7.29e-ll
2.43e-10
2.66e-10
9. lie- 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
2.13e-09
7.60e-09
1.22e-10
4.03e-10
3.87e-10
1.06e-09
1.06e-09
1.14e-09
2.58e-10
6.08e-09
6.46e-10
4.03e-09
Run 5
< 4.58e-ll
6.99e-ll
9.026-11
1.65e-10
8.27e-ll
9.01e-10
2.486-09
6.76e-ll
3.38e-10
3.91e-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
1.436-10
2.27e-10
1. 81e-10
1.216-09
3.17e-09
2.27e-10
5.89e-10
5.516-10
1.36e-09
1.28e-09
1.28e-09
2.79e-10
6.65e-09
6.19e-10
3.32e-09
Volume III
Appendix III-2
-3-

-------
APPENDIX HI-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) - April 1994 ECIS Performance Test
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
7.29e-10
7,29e-10
1.24e-10
7.29e-09
9.48e-10
2.11e-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
1.72e-10
6.67e-10
6.67e-10
6.67e-10
1.39e-10
8.90e-09
9.45e-10
2.22e-08
Run 3
< 3.796-11
< 4. 12e-ll
< 6.136-11
1.06e-10
< 5.296-11
8.36e-10
3.57e-09
3.06e-ll
l.lle-10
1.84e-10
6.69e-10
6.69e-10
7.24e-10
1.28e-10
6.69e-09
7.80e-10
l.lle-08
Run 4
<4.23e-ll
< 5.26e-ll
<8. 13e-ll
1.03e-10
< 7.04e- 11
5.42e-10
2.06e-09
4.126-11
1.356-10
2.17e-10
7.04e-10
7.04e-10
8.13e-10
1.52e-10
5.96e-09
5.96e-10
7.04e-09
RunS
<3. 37e-ll
<4.64e-ll
< 7.186-11
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
4.64e-10
9.946-11
4.69e-09
3.42e-10
4.86e-09
Volume III
Appendix III-2
4-

-------
APPENDIX HI-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
!,2,3,>,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
3.81e-10
3.81e-10
4.38e-10
1.14e-10
3.36e-09
5.57e-10
5.52e-09
Run 2
<5.22e-ll
<4.49e-ll
<6.18e-ll
< 5. 22e-ll
<5.51e-ll
2.58e-10
6.18e-10
<3.93e-ll
<7.86e-ll
8.99e-ll
2.92e-10
2.706-10
2.87e-10
<3.7le-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
1.22e-09
<3. 33e-ll
8.88e-ll
1.33e-10
3.72e-10
3.66e-10
4.05e-10
8.326-11
2.94e-09
3.05e-10
2.33e-09
Run 4
<2.75e-ll
<3.65e-ll
< 6.73e-ll
<8.41e-ll
<5. 61e-ll
3.08e-10
6.73e-10
<5.05e-ll
1 12e-10
1.68e-10
3.98e-10
3.93e-10
3.93e-10
<5. 55e-ll
2.36e-09
2.97e-10
< 1.12e-09
Run 5
< 4.086-11
<5.31e-ll
<6.71e-ll
< 5.59e-ll
<6.15e-ll
3.80e-10
1.73e-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
< 4.496-11
< 2.596-11
< 8.63e-ll
< 6. 90e-ll
< 7.486-11
4.77e-10
1.55e-09
<5. 12e-ll
1.32e-10
1.67e-10
4.77e-10
4.66e-10
4.54e-10
1.21e-10
3.05e-09
3.39e-10
2.07e-09
Run?
< 3.436-11
<4.72e-ll
< 1.07e-ll
< 8.426-11
< 8. 99e-ll
5.346-10
2.47e-09
< 4.276- 11
1.29e-10
1.69e-10
5.28e-10
4.946-10
4.666-10
<7. 866-11
3.826-09
4.66e-10
< 6. 18e-09
Volume III
Appendix III-2
-5-

-------
          APPENDIX III-3

PRODUCTS OF INCOMPLETE COMBUSION
ANALYZED FOR AND DETECTED IN THE
TRIAL BURN AND PERFORMANCE TESTS

-------
APPENDIX III-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(g,h,i)perylene
Benzo(j)fluoranthene*
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
Feh-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 HI
Appendix III-3
-1-

-------
APPENDIX III-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)
Butyibenzylphthalate
Carbon disulfide
Carbon tetrachloride
Chlordane
Chloro-3-methylphenol, 4-
Chloroacetophenone, 2-*
Chloroanilme, p-
Chlorobenzene
Chlorobenzilate
Chloroethane
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
Diben7o(a h'Jfluoranthene
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-

-------
APPENDIX III-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Dibromo-3-chloropropane, 1,2-*
Dibromochloromethane
Dichloro-2-bulene, cis-1,4-*
Dichloro-2-butene, trans- 1,4-*
Dichlorobenzene, 1,2-
Dichlorobenzene, 1,3-
Dichlorobenzene, 1,4-
Dichlorobenzidine. 3,3'-
Dichlorobiphenyl
Dichlorodifluoromeihane
Dichloroethane, 1,1-
Dichloroethane, 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
Dinitrotoluene, 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-

-------
APPENDIX III-3
Products of Incomplete Combustion Analyzed for
and Detected in the Trial Burns and Performance Tests
(continued)
Substance
Dinitrobenzene, 1,4-*
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, alpha-*
Hexachlorocyclohexane, beta-*
Hexachlorocyclohexane, gamma- (a.k.a.
Lindane)
Hexachlorocyclopentadiene
Hexachloroe thane
Hexachlorophene
Hexane, n-*
Hexanone, 2-
Hexanone, 3-*
Indenof 1 2 3-cd^nvrene
Compounds 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

Feh-94 Anj-94
Volume III
Appendix III-3
-4-

-------
APPENDIX III-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-
Methyi-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-propylamine
N-Nitrosodiphenylamine
Nonachlorobiphenyl
Octachlorobiphenyl
Pentachlorobenzene
Pentachlorobiphenyl
Pentachloronitrobenzene
Pentachlorophenol
Phenanthrene
Phenol
Phosgene*
Propionaldehvde*
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-

-------
APPENDIX HI-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-1 ,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-
Xvlene o-
Compounds 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 HI
Appendix III-3
-6-

-------
                                       APPENDIX III-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.	
Volume III
Appendix III-3
-7-

-------