Uniltx] States      lixti i:>ti i.il Lnyii i a lory
Agency        Cinrmnati OH '


Research and Development
F.PA 600 0 84 01 h
July 1984
Incineration and
Treatment of
Hazardous Waste

Proceedings of the
Ninth Annual
Research Symposium

-------
                                                       EPA-600/9-84-015
                                                       July 1984
        INCINERATION AND TREATMENT OF HAZARDOUS WASTE
      Proceedings of the Ninth Annual  Research Symposium
           at Ft. Mitchell, Kentucky,  May 2-4, 1983
Sponsored by the U.S. EPA, Office of Research and Development
         Municipal  Environmental  Research Laboratory
         Solid and Hazardous Waste Research Division
                             and
         Industrial  Environmental Research Laboratory
              Energy Pollution Control  Division
           Edited by:  Incineration Research Branch
   Industrial  Environmental  Research Laboratory-Cincinnati
                       Project Officer

                        Ivars J. Licis
         Industrial Environmental  Research Laboratory
              Energy Pollution Control Division
                   Cincinnati, Ohio  45268
         INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
              OFFICE OF RESEARCH AND DEVELOPMENT
             U.S. ENVIRONMENTAL PROTECTION AGENCY
                   CINCINNATI, OHIO  45268

-------
                                         NOTICE


    This document has been  reviewed in  accordance  with  U.S.  Environmental  Protection
Agency policy and approved  for publication.   Mention  of trade  names  or  commercial  pro-
ducts does not constitute endorsement or  recommendation for  use.

-------
                                        FOREWORD
     When energy and material  resources are extracted,  processed,  converted,  and used,
the related pollutional  impacts on our environment and  even on our health often require
that new and increasingly more efficient pollution control  methods be used.   The Indus-
trial  Environmental Research Laboratory Cincinnati (lERL-Ci)  assists in  developing  and
demonstrating new and improved methodologies that will  meet these  needs  both  efficiently
and economically.

    These Proceedings present the results of completed  and  on-going research  concerning
the incineration and treatment of hazardous wastes.   The information will  inform those
who own, operate,  design, or regulate hazardous waste incineration and treatment facili-
ties of current government-sponsored research in this area.  For further information  on
this subject, interested parties should contact the  Incineration Research  Branch,  Indus-
trial  Pollution Control  Division, Industrial Environmental  Research Laboratory,  USEPA,
Cincinnati, Ohio 45268.
                                                            David  G.  Stephan
                                                                Director
                                               Industrial  Environmental Research  Laboratory
                                                               Cincinnati
                                            m

-------
                                         PREFACE


    These Proceedings  are  intended  to  disseminate up-to-date  information on extramural
research projects concerning  land disposal,  incineration, and treatment of hazardous
waste.   These projects are funded by the  U.S. Environmental Protection Agency's Office
of Research and Development and  have been reviewed  in accordance with the requirements
of EPA's Peer and Administrative Review Control System.

-------
                                 ABSTRACT
     The Ninth Annual Research Symposium on land disposal, incineration and
treatment of hazardous wastes was held in Ft. Mitchell, Kentucky, on May 2,
3, and 4, 1983.  The purposes of the symposium were (1) to provide a forum
for state-of-the-art review and discussion of ongoing and recently completed
research projects dealing with land disposal, incineration, and treatment of
hazardous wastes; (2) to bring together people concerned with hazardous
waste management who can benefit from an exchange of ideas and information;
(3) to provide an arena for the peer review of the Solid and Hazardous Waste
Research Division's and the Energy Pollution Control Division's research
programs on hazardous waste management.  These Proceedings are a compilation
of papers presented by the symposium speakers.

     The symposium proceedings are being published as two separate documents.
In this document, Incineration and Treatment of Hazardous Waste, seven technical
areas are covered.  They are as follows:

     (1)  Incineration Emissions Measurement Methods
     (2)  Lab Scale and Pilot Scale Thermal Decomposition Research
     (3)  Evaluation of Emissions from Full-Scale Hazardous Waste Incinerators
     (4)  Hazardous Waste Incineration in High Temperature Industrial Processes-
          Boilers and Kilns
     (5)  Methods for Conducting Environmental and Economic Assessment of
          Hazardous Waste Incinerators
     (6)  Innovative Hazardous Waste Control Technology
     (7)  Biological Degradation of Hazardous Waste

-------
                                  CONTENTS


                                                                      Page

           SESSION B:   HAZARDOUS  WASTE  INCINERATION  AND TREATMENT


                 INCINERATOR EMISSIONS  MEASUREMENT METHODS
Development of a Volatile Organic  Sampling  Train  (VOST)
     Gregory A.  Jungclaus,  Paul  G.  Gorman,  George  Vaughn, George W.
     Scheil, FredJ.  Bergman
     Midwest Research Institute	
The Feasibility of Hydride Generation  Inductively Coupled
Plasma Spectroscopy for Analysis  of  Volatile Metals
     M.P. Miller, P.H.  Chinn,  B.G. Snyder,  A.K. Wensky
     Battelle Columbus  Laboratories	 28

Speciation of Halogen and Hydrogen Halide  Compounds  in
Gaseous Emi ssions
     David A. Stern, Barbara M. Myatt, Joseph F. Lachowski,
     Kenneth T. McGregor
     GCA Corporation	 33

Dioxin Collection from  Hot Stack  Gas Using  Source Assessment
Sampling System and Modified Method  5  Trains - An Evaluation
     Marcus Cooke, Fred DeRoos, Bruce  Rising
     Bettel le Columbus  Laboratories	 42

Stack Sampling and Analysis of Formaldehyde
     Kevin J. Beltis,  Anthony  J.  DeMarco,  Virginia A. Grady,
     Judith C. Harris
     Arthur D. Little,  Inc	 56
          LAB SCALE  AND  PILOT  SCALE  THERMAL  DECOMPOSITION RESEARCH
Factors Affecting the Gas-Phase  Thermal  Decomposition  of
Chlorinated Aromatic Hydrocarbons
     Barry Dellinger, Douglas  L. Hall,  Wayne A.  Rubey,
     Juan L. Torres
     University of Dayton  Research  Institute....	 65
                                    VI 1

-------
Laboratory-Scale Flame Mode  Study  of Hazardous Waste
Incineration
     W.R.  Seeker, J.C. Kramlich, M.P.  Heap
     Energy and Environmental  Research Corporation	   79

The Packaged Thermal  Reactor System:   Development and
Application
     Wayne A. Rubey,  John  L.  Graham,  Barry  Dellinger
     University of Dayton  Research Institute	   95

Incinerability Characteristics of  Selected  Chlorinated
Hydrocarbons
     David L. Miller, Vic  A.  Cundy,  Richard A.  Matula
     Louisiana State  University	  113

Status Report USEPA Combustion Research Facility  (CRF)
     Frank C. Whitmore, C.F.  Fowler,  R.W. Ross
     Versar, Inc	  129


    EVALUATION OF EMISSIONS  FROM FULL-SCALE HAZARDOUS WASTE  INCINERATORS
A Profile of Existing Hazardous Waste  Incineration Facilities
     Edwin L. Keitz, Leo J.  Boberschmidt
     The MITRE Corporation ............................................  137

Particulate and HC1  Emissions  from Hazardous  Waste Incinerators
     Paul Gorman, Andrew Trenholm
     Midwest Research Institute .......................................  151

Emission Test Results for a  Hazardous  Waste Incineration  RIA
     Andrew Trenholm, Paul Gorman
     Midwest Research Institute .......................................  160

Fluidized-Bed Incinerator Performance  Evaluation
     Robert R. Hall, Gary T. Hunt,  Mark M. McCabe
     GCA Corporation ..................................................  171
  HAZARDOUS WASTE INCINERATION  IN  HIGH  TEMPERATURE  INDUSTRIAL  PROCESSES  -
                             BOILERS  AND  KILNS


Full-Scale Boiler Emissions  Testing of  Hazardous Waste Co-firing
     Carlo Castaldini,  Howard B. Mason, Robert J. DeRosier,
     Bruce C.  DaRos
     Acurex Corporation ............... . . .............................. 180
                                     vi n

-------
Subscale Parametric Studies on the Combustion  of  Hazardous  Waste
     Carlo Castaldini, Andrew R. Garman,  Jeffrey  M.  Kennedy,
     Howard B. Mason, C. Dean Wolbach
     Acurex Corporation	  194

Evaluation of Hazardous Waste Incineration in  a Cement  Kiln at
San Juan Cement Company
     James A. Peters, Thomas W. Hughes
     Monsanto Research Corporation	  210


      METHODS FOR CONDUCTING ENVIRONMENTAL AND ECONOMIC ASSESSMENT OF
                        HAZARDOUS WASTE INCINERATORS
Automated Methodology for Assesing Inhalation  Exposure  to  Hazardous
Waste Incinerator Emissions
     F.R. O'Donnell, G.A. Holton
     Oak Ridge National  Laboratory	 225

Operation and Maintenance Cost Relationships  for  Hazardous Waste
Incineration
     Robert J. McCormick
     Acurex Corporation	 235

Retrofit Cost Relationships for Existing  Hazardous  Waste Incineration
Facil ities
     Robert J. McCormick
     Acurex Corporation.	 248
               INNOVATIVE HAZARDOUS WASTE  CONTROL TECHNOLOGY
Full-Scale Demonstration of Wet Air Oxidation  as  a  Hazardous Waste
Treatment Technology
     Dr. William Copa, James Heimbunch,  Phillip Schaefer
     Zimpro, Inc	 267
                 BIOLOGICAL  DEGREDATION  OF  HAZARDOUS WASTE
Engineering Genes in Yeast for Biodegradations
     John C. Loper,  Jerry B.  Lingrel,  Vernon  F.  Kalb
     University of Cincinnati	 274

Ninth Annual Symposium Attendees  List.	 282

-------
                             DEVELOPMENT OF A VOLATILE ORGANIC
                                   SAMPLING TRAIN (VOST)
                   Gregory A. Jungclaus, Paul G.  Gorman,  George Vaughn,
                           George W. Scheil, and Fred J.  Bergman
                                Midwest Research Institute
                               Kansas City, Missouri  64110

                                     Larry D. Johnson
                       Industrial Environmental Research Laboratory
                         USEPA, Research Triangle Park, NC  27711

                                      David Friedman
                                   Office of Solid Waste
                               USEPA, Washington, DC  20460
                                         ABSTRACT
     The hazardous waste incineration regulations include the requirement that, for se-
lected principal organic hazardous constituents (POHCs), a destruction/removal efficiency
(ORE) of ^ 99.99% must be achieved.  In order to calculate meaningful DRE values,  reliable
sampling and analysis methods must be available.  This paper reports on the development
and evaluation of a volatile organic sampling train (VOST) for the collection of volatile
POHCs from stack gas.  The VOST is a method designed by the USEPA as an alternative to the
use of integrated gas bulbs and bags.  The paper includes data concerning the collection
and analysis of four volatile POHCs during the laboratory evaluation, descriptions of the
equipment, a description of a field version of the VOST, procedures followed to minimize
sample contamination in the field, and conclusions and recommendations from the study.
1.0  INTRODUCTION

     The results of previous hazardous
waste incineration trial burns have sug-
gested that volatile principal organic
hazardous constituents (POHCs) and vola-
tile products of incomplete combustion
(PICs) may be important components in the
incineration effluents.  The sampling
technique described in a recent sampling
and analysis document  (1) for volatile or-
ganic compounds involves the collection
and analysis of integrated gas bulb and
bag samples.  However, the authors of that
report recognized that the gas bag tech-
nique suffers several drawbacks, including
the need to position the gas bag in a bulky
evacuated sampling box, bag leakage pro-
blems, adsorption losses of sample compo-
nents, contamination problems, and low
sensitivity when the bulb or bag is
analyzed using a gastight syringe sampling
technique.

     To address the need to develop a bet-
ter sampling and analysis technique for
volatile POHCs, personnel from the En-
vironmental Protection Agency (EPA) dis-
cussed concepts for a volatile organic sam-
pling train (VOST) with several contracted
laboratories.   One concept was adopted for
development to provide a method to collect
a sufficient quantity of volatile POHCs to
enable calculation of destruction/removal
efficiencies (DREs) as high as 99.999°/0 for
incinerators whose waste feed contains as
little as 100 ppm of a POHC.

     MRI was selected by EPA to carry out
a laboratory study to develop and evaluate
the sampling train concenpt.   Following
the laboratory evaluation of the VOST (2),

-------
a field version of the VOST was designed
and built by MRI,  and is currently being
evaluated under field sampling conditions.

     This paper describes how the labora-
tory evaluation was performed, presents
the results of the evaluation, describes
the field version of the VOST, and presents
conclusions and recommendations based on
the results to date.
2.0  VOST CONCEPT

     The VOST concept to be evaluated bas-
ically consisted of a system designed to
draw sample gas at a flow rate of 1 liter/
min through two traps in series.  The first
trap contained Tenax and was preceded by a
gas cooler/ condenser and followed by an
impinger for condensate collection.  A sec-
ond trap containing a section of Tenax and
a section of charcoal was located after
the the impinger.  The purpose of the sec-
ond trap was to collect very volatile POHCs
(e.g., vinyl chloride), which have low
breakthrough volumes and may break through
the Tenax trap.  In addition, the concept
involved replacing both pairs of traps with
fresh traps at selected intervals (i.e.,
every 20 min or 20 liters of sample) over
a 2^h sampling period.  There were two
basic reasons for changing the traps at
selected intervals:

        At sample volumes of greater than
        20 liters, some of the very vola-
        tile POHCs may break through both
        the front and backup adsorbent
        traps.

        The changing of the traps allows
        an initial analysis of one pair of
        traps.  Analysis of a single pair
        of traps lowers the possibility of
        collecting too much sample and
        overloading the GC/MS system.  How-
        ever,. • if the POHCs are not detected
        or are present at low levels in
        the single pair, the option exists
        of combining the contents of the
        remaining pairs of traps onto one
        pair of traps with a concomitant
        increase in sensitivity.  The ad-
        vantage of the seond option for
        samples with low POHC concentra-
        tions is given below.

      If a hazardous waste incineration  fa-
 cility is achieving a DRE of 99.999% for a
POHC that is present in the waste at a con-
centration as low as 100 ppm, the resulting
concentration of that POHC in the flue gas
will be approximately 0.1 (Jg/m3 or 0.1 ng/
liter.  Sampling 20 liters of that gas will
collect only 2 ng of the POHC on a single
pair of traps.  Since 2 ng may not be de-
tectable by GC/MS analysis, the concept
required collection of several (e.g., five)
additional pairs of traps and the desorp-
tion of their contents onto another pair
of traps, thereby providing a total of
10 ng for GC/MS analysis.

     It was anticipated that, when the
VOST system is used in the field, one will
not know whether pairs of traps should be
analyzed individually or if the contents
of several pairs should be desorbed onto
one pair.  That is, if the concentration
of a selected volatile POHC in the efflu-
ent is low (e.g., 0.1 to 1.0 ng/liter),
several pairs may need to be desorbed onto
one pair to achieve sufficient analytical
sensitivity.  However, if the concentration
is high, the pair of traps should be an-
alyzed individually, since desorption of
the contents of several pairs of traps onto
one pair of traps would make the quantity
even larger and saturate the GC/MS de-
tector.  Therefore, the intent was to use
the VOST to collect six pairs of sample
traps, but with one pair being analyzed
first, individually, to determine the
amount of selected POHCs present.  Then,
if warranted, the contents of some or all
of the remaining pairs could be desorbed
onto one pair for analysis, or other pairs
of traps could be analyzed individually to
check the variability in the stack gas com-
position with time.

     The selection of a Tenax front trap
and Tenax/charcoal backup trap was based
on several factors including the authors'
previous experience with adsorbents and
information in the literature, primarily
from work done at Research Triangle Insti-
tute  (3).  Tenax alone is not a very good
adsorber for very volatile organic com-
pounds such as chloromethane and vinyl
chloride.  Charcoal is a good adsorber for
the very volatile organics, but compounds
that  are less volatile are not easily de-
sorbed from charcoal.  Thus  the dual trap
configuration was considered the most ver-
satile for providing efficient sample col-
lection  and recovery of  all volatile or-
ganics.

-------
     The plan was developed to evaluate
the VOST concept that consisted of the fol-
lowing:

        Set up an experimental system to
        generate a wet gas stream prepared
        with four volatile POHCs at each
        of four different concentration
        levels as described in Section 3.1.

        Construct three identical VOSTs
        that would simultaneously draw gas
        from the synthetic gas stream.

        Set up equipment for conditioning
        traps and for thermally desorbing
        the contents of several pairs of
        traps onto one pair.

        Set up equipment for analyzing
        traps by GC/MS.

     After the above equipment had been
set up  and made operational, the plan con-
sisted  of carrying out a series of 10
tests.  The test runs included:  tests at
each of four concentration levels, repli-
cate tests at one level, blanks, and a test
(also  at the replicate level) where the
gas contained HC1.  The purpose of this
last test was to determine if HC1, which
is present in many incinerator effluents,
had any effect on the analysis results.
The order of the tests was randomized to
prevent bias from affecting the results.
The sequence of the tests in this plan was:
Test
      Level
(concentration of
  POHCs  in gas)
   1     III  (10  ng/JZ)
  10
        0
        III  (10  ng/£)
        I  (0.1 ng/£)
        II  (1.0  ng/£)
        IV  (100  ng/£)
        0
        II  (1.0  ng/£)

        II-HC1 (1.0 ng/£)
    Comment

Exploratory run
to check system
Blank run
                    Blank  run
                    Duplicate  of
                    Run  5
                    Duplicate  of
                    Run  5  with HC1
                    in gas
                    Blank  run
The equipment used  in  carrying out the
tests  is  described  in  the next section.
3.0  LABORATORY EVALUATION OF THE VOST

     This section contains descriptions of
the equipment and procedures used in the
laboratory evaluation of the VOST includ-
ing:

     3.1  Sample Gas Generator System.

     3.2  Sampling Train Design.

     3.3  Trap Conditioning Equipment.

     3.4  Analytical Procedures.

     3.5  Results.

     3.6  Summary and Interpretation of
Results.

3.1  Gas Generation System

     As shown in Figure 1, the gas gen-
eration system consisted of 1/2-in. (1.27
cm) stainless steel tubing to carry vapor-
ous N£- from a liquid N2 tank through a
heater, where the N2 was heated to about
300°F  (149°C).  At that point, the N2 was
rendered "wet" by vaporizing deionized/
charcoal-filtered water fed through a
quartz  tube heater.  Also near that point,
the liquid containing the four POHCs was
pumped  by a syringe pump  (5 ml/h) into the
hot N2  stream where the liquid immediately
vaporized to a gas.

     The liquid injected by the syringe
pump was a solution of the four POHCs,
vinyl  chloride, carbon tetrachloride, tri-
chloroethylene, and chlorobenzene in meth-
anol.

     The concentrations of each of the four
POHCs  tested are  listed in Table  1.

     The solution with the highest con-
centration (Level IV) was prepared first,
then aliquots were serially diluted with
methanol to prepare the three lower con-
centration solutions.  These same solu-
tions  were used as calibration  standards
for the subsequent analyses.

     Following the steam  and POHC solution
injection point,  the gas  stream entered  a
sampling manifold, with a perforated  dis-
persing plate  at  the inlet.  Gases were
drawn  from this manifold  into the three
sampling trains.  After the manifold, the
hot gas (11 to 12 liters/min) passed
through a series  of impingers (for water

-------
                                            l/Min
       Syringe Pump
       for Injecting
       Liquid Containing
       Volatile  POHCs
                   (5 ml/fir)
                     (149°C)
                      300° F
Heat Traced &
Insulated Lines
                                                   +• Sampling Train No. 1
-*• Sampling Train  No. 2


-*• Sampling Train  No. 3
                                                              Exhaust
                                                 Water-Removal
                                                 Impingers (3 )
                              Vaporizer
                              Furnace
        ~5 ml/ mi
Figure  1.   Schematic diagram  of  laboratory apparatus used to generate  and sample
    a simulated stack gas containing known concentrations of volatile  POHCs.

-------
      TABLE 1.   POHC CONCENTRATIONS TESTED DURING LAB EVALUATION OF THE VOST
  Level
    Concentration
in methanol solution
       (ng/ml)
Expected cone.
in gas stream
    (ng/£)
Expected amount
 on each pair
 of traps (ng)
I
n
ni
IV
84
840
8,400
84,000
0
1
10
100
.1
.0
.0
.0
2
20
200
2,000
removal) and on through a pump and dry gas
meter.

     During each of the tests, the gas gen-
eration system operated quite well, and
all readings were consistent from run to
run.  The water content of the gas stream
ranged from 35 to 37 volume percent, as
measured by the impingers and gas meter.

     During Run 9, HC1 was added to the
water at a level of 1.3 g of HC1 per liter
of water to provide an HC1 concentration
in the gas of about 0.5 g/Nm3 (normal cubic
meter).  This is the HC1 concentration
estimated to occur in the effluent from an
incinerator burning a waste containing 157o
Cl and equipped with a wet scrubber operat-
ing at the relatively low HC1 removal ef-
ficiency of 95%.

     The gas flow rate in the gas generator
system and the POHC syringe pump injector
rate were used to compute an "expected
value" for the quantity of each POHC in
the Tenax traps.  It was not feasible
within the scope and time frame of this
project to quantify the actual concentra-
tion of the POHCs in the gas produced by
the gas generator system.  An independent
analysis of the spiked gas stream would
have been desirable but very difficult to
accomplish; however, the subsequent VOST
data gave little or no reason to believe
that the actual gas stream concentrations
were significantly different than the com-
puted "expected values."

3.2  Sampling Train Design

     Figure 2 shows the VOST configuration
that was evaluated.  The train consisted
of:

        A sampling line (1/4-in., 0.64 cm,
        Teflon tubing in the test system);
                                      First condenser;

                                      Tenax trap;

                                      Impinger (for condensate removal) ;

                                      Second condenser;

                                      Tenax/charcoal trap;  and

                                      Other sampling components (rotom-
                                   eter,  pump, dry gas  meter).

                                   Except for the  Teflon sampling line,
                              most of the components were made of glass,
                              including the traps.  However,  the fittings
                              at the inlet and outlet of each trap were
                              stainless steel.

                                   When the trains were initially as-
                              sembled, two problems developed.  First,
                              the 5/8-in. (1.58 cm) stainless steel
                              Swagelok fittings for the inlet traps were
                              designed to slip over the glass trap, but
                              some of the traps had an outside diameter
                              slightly larger than the inside diameter
                              of the Swagelok nut.  Thus, all the nuts
                              had to be drilled out.  Secondly, some of
                              the glass traps were out-of-round.   This
                              meant that  some traps, when inserted into
                              the sampling train,  would not leak-check
                              unless the  fittings  were tightened with
                              wrenches.  Several tubes broke before they
                              could be tightened enough to pass a leak-
                              check.  Since checking for leaks and cor-
                              recting of  leaks can take considerable
                              time, the adsorbent  traps were  redesigned
                              for field use as described in Section 4.0.

                                   Before each run, the gas generation
                              system was  started up and allowed to op-
                              erate for about 1 h.  During that time,
                              traps were  connected in the three trains
                              and leak-checked.  All three trains were
                              then started and operated for 20 min at

-------
Glass Wool
Partic
Filter
 Stack
;ulate
\

t
:k
Test
em )
Teflon
Probe
	 ^f
i
\
/
\~
                      Condensate
                      Trap Impinger
                                                                      Vacuum
                                                                      Indicator
                                                       Tenax
                                                       Trap
                                                       Charcoal Backup
                                                                            '
Empty    Silica Gel
                      Note:  Tenax & Tenax/charcoal traps were 1.6 cm in diameter
                             & 10 cm long
                                                                 Exhaust
                                                                 1 l/min
                                                                                          Pump
                                                 Dry Gas
                                                 Meter
Note:  3 trains as shown above
       were operated each test
       day.

       Both traps were changed
       every 20  minutes over
       2 hour period .
                                  Figure 2. Volatile organic sampling  trian (VOST) ,

-------
about the same rate (1 liter/rain).'*  All
three sampling trains were then shut off
and the traps removed and placed in pre-
marked container tubes.   Another pair of
traps were then inserted in each train and
leak-checked before starting the next 20-
min sampling period.   A run was considered
complete after six pairs of traps had been
used in each train.  During each run, ice
water was circulated through the con-
densers.  Thermocouples, located against
the surface of the condenser outlet tubes,
indicated that the gas temperature enter-
ing the first trap was in the range of 60
to 80°F (16 to 27°C).  (The train with the
longest Teflon sampling tube yielded the
lowest temperature.)

     Overall, the train configuration
caused no particular difficulty, except
for the leak-check problem described above.
However, using this train configuration to
sample a "wet" gas stream saturates the
first trap with condensate.  This caused
no problems in the sampling but did re-
quire development of special procedures
for analyzing the wet traps, as discussed
in Section 3.4.

3.3  Trap Conditioning Equipment

     The trap conditioning/desorption ap-
paratus, purchased from Nutech''"* (Model No.
322), served two purposes for the VOST
evaluation.  First, it was used to condi-
tion traps prior to use, by heating them
at 250°C for 4 h with an estimated flow of
30 ml/min of purified nitrogen gas through
each trap.  Second, it was used to ther-
mally desorb the contents from each of sev-
eral low-level pairs of traps onto one pair
of traps for GO/MS analysis.  The purpose
of this desorption/adsorption was, in ef-
fect, to further concentrate the samples
from the sampling train.

     A  schematic diagram of the condition-
ing/desorption apparatus is shown in Fig-
ure 3,  along with the trapping system that
was added at the outlet to re-adsorb the
contents from the desorbed pairs of traps.
    Gas flow rates in liters per minute
      refer to normal conditions of 20°C,
      1 atm (dry basis).
    Nutech Corporation, 2806 Cheek Road,
      Durham, NC  27704.
     When four traps were being desorbed
(which is the capacity of one section of
the desorption apparatus), the carrier gas
(N2) exits the desorption chamber hot, but
cools rapidly.  However, when the traps
being desorbed are wet, the cooling is not
nearly so rapid because of the steam that
must be condensed.  Thus, it was necessary
to use a condenser at the outlet of the
conditioning equipment, in front of the
first trap (Tenax).   An impinger was also
required to remove the condensate before
the desorbed gas passed into the second
trap.  As a result,  the re-adsorption sys-
tem of traps at the outlet of the desorp-
tion equipment is equivalent to the sam-
pling train itself.   Also, the condensed
steam again wets the first trap, so the
need to analyze a wet trap still remains.

     When using the Nutech conditioning
apparatus to desorb several pairs of traps
(e.g., five pairs), the conditioner was
first heated to its normal operating tem-
perature of 250°C.  Four  traps were then
dropped into the chambers and allowed to
remain there for 10 min  (with the total N2
carrier flow of 120 ml/min passing thr'ough
the four traps).  These four traps were
then removed and four more traps inserted,
repeating the procedure until all five
pairs had been desorbed onto the one pair
at the outlet.  This pair, or any pair an-
alyzed individually without first being
desorbed, was then spiked with an internal
standard and analyzed using the equipment
and procedures described  in Section 3.4.

3.4  Analytical Procedures

     The analytical procedures described
below include cleanup of  the Tenax and
charcoal prior to packing into traps, prep-
aration of the traps, conditioning of the
traps prior to sampling,  spiking of the
traps with an internal  standard following
sampling, GC/MS analysis  of the traps, and
data reduction.

3.4.1  Tenax  and  Charcoal Cleanup--

     The Tenax  (35/60 mesh) and SKC pe-
troleum-based charcoal  (Lot No. 104) were
initially prepared by Soxhlet extraction
for  24 h with methanol  and then with pen-
tane.  The sorbents were  then dried in a
vacuum oven at  100°C for  6 h prior to  load-
ing  into empty  traps,  each engraved with  a
unique number.

-------



A
c
r










D
S^












• \\
u









N2
100 ml/Min

x Liquid
N2





111 ' o *i ' ' rt *
i'ii i'ii i / 1 1 i / \ i
!/_>! !' u \f •• i' >i
r'-^j-^r^j^j-^l^j


Trap Conditioner
& Desorber *
Ice
Water
Ter
Tra








— b
r
tax
P
4


I
i

><

V
1

Impinger


Condenser


^
/

' 	 ^
S? Tenax Trap
>< Charcoal
I
4
Vent

Figure 3.  Schematic diagram of trap conditioner/desorption apparatus.

-------
3.4.2  Preparation of Traps--
     The 10- x 1.6 cm glass traps with one
nippled end (to facilitate removal of the
traps from the desorption apparatus with
tweezers), available from the Nutech
Corporation, were used for the VOST evalu-
ation.   A minimum amount of pre-extracted
and oven-dried glass wool was used in each
of the glass tubes to hold the sorbents in
the glass traps.   The all-Tenax traps con-
tained about 1.6 g of Tenax, and the Tenax/
charcoal traps contained about 1 g of Tenax
and 1 g of charcoal (two-thirds Tenax by
volume).

3.4.3  Trap Conditioning--

     The traps were thermally conditioned
prior to use, using the Nutech Model 322
thermal conditioning unit.  The condition-
ing gas (nitrogen or helium) was purified
by passing through a U-trap containing a
5-angstrom (5 x 10 8 cm) molecular sieve
with the U-trap immersed in liquid nitro-
gen.  The temperature of the conditioning
unit was adjusted to 240 to 250°C, and the
flow rate of gas through each trap was
estimated to be about 30 ml/min.  However,
only the sum of the flow through four of
the traps, which was set at 120 ml/min,
could actually be measured with the Nutech
conditioning unit.  The traps were condi-
tioned for at least 6 h prior to their
first use in the VOST evaluation and for
at least 2 h more prior to use in sampling.
The actual flow through each trap may be
lower due to the fact that some of the con-
ditioning gas may flow around rather than
through the traps.  Also residual pentane
was observed during several subsequent
analyses of the traps, suggesting that the
conditioning step was not completely ef-
fective.

     Following conditioning, each trap was
transferred to a clean 25- x 150-mm screw-
cap test tube engraved with the same unique
number as engraved on the trap.  The traps
were then ready for sample collection or
spiking experiments.

3.4.4  Spiking of Traps with Internal and
Calibration Standards--

     Prior to GC/MS analysis, all Tenax
and Tenax/charcoal adsorbent trap samples
and standards were spiked with 25 ng of
perfluorobenzene  (PFB) internal standard
using  the flash vaporization technique in
which the spiking solution is vaporized
and carried onto the trap with a carrier
gas.  The glass traps were attached to the
injection port (160°C) of a GC with a 5/8-
in. (1.58 cm) stainless steel Swagelok nut
containing Teflon ferrules.  The Swagelok
fitting was connected to the GC column con-
nection via a reducing fitting.  The helium
flow through the traps was set to about
50 ml/min.  The gas flow through the trap
was turned on and off using the shutoff
valve on the side of the Varian 1400 GC.

     The spiking solution was loaded and
expelled from the syringe using the solvent
flush technique to ensure that the standard
solution would be completely expelled from
the syringe.  To use this technique, the
needle of a 5.0 )jl syringe was filled with
clean methanol.  The methanol was then ex-
pelled leaving methanol only in the syringe
needle.  Then air was drawn into the sy-
ringe to the 1.0 pi mark followed by a 25-
ng/(jl methanolic solution of the PFB to
the 2.0 pi mark.  The gas flow was turned
on through the trap and the syringe needle
inserted through the GC septum port.  The
contents of the syringe were then slowly
expelled over about a 15-s period.  At the
end of about 25 s, the gas flow through
the trap was shut off and the syringe re-
moved.  All POHC calibration standards were
spiked using exactly the same procedure.
The total flow of gas through the traps
during spiking was thus only about 25 ml.

3.4.5  GC/MS Analysis of the Traps--

     To analyze the traps, the contents of
the wet traps (dry traps in the case of
method blanks, field blanks, and calibra-
tion standards) were thermally desorbed
using a stream of carrier gas into a water
column (1 to 5 ml); this is a component of
the EPA Method 624 purge-trap-desorb GC/MS
analysis system.  A schematic diagram of
the apparatus is shown in Figure 4.  The
sample trap was dropped into the desorp-
tion chamber and desorbed at a flow rate
of  100 ml/min for 10 min at 180°C.  The
desorbed compounds passed into the bottom
of the water column, were purged from the
water, and then were collected on an ana-
lytical adsorbent trap also containing
Tenax and charcoal.  The compounds were
then desorbed from the analytical adsorbent
trap into the GC/MS system per EPA Method
624.

-------
CD
         N
                Thermal
                Desorption
                Chamber
                                        t
 Flow During
                                           Desorption
                                Flow to
                                GC/MS    p|ow During
                              Frit
 /
Heated
Line
                                                              He or N2
Analytical Trap
with Heating Coi I
(0.3 cm diameter
by 25 cm long)
                                                             Vent
                                      Purge
                                      Colum
           (T)  3%  SP-2100  (1cm)

           (2)  Tenax  (7.7cm)

           (?)  Silica  Gel  (7.7cm)

           (T)  Charcoal (7.7cm)
        Figure  4.   Schematic  diagram of trap desorption/analysis system.

-------
     The normal routine for analyzing a
set of traps from each of the laboratory
VOST evaluation runs was to analyze a cal-
ibration standard on Tenax, a calibration
standard on Tenax/charcoal, and then to
intersperse calibration standards about
every fourth sample.  Blank Tenax and blank
Tenax/charcoal traps (conditioned traps
spiked with internal standard) were also
analyzed when the samples from the blank
VOST train were analyzed.  The same POHC
solution used to spike the wet gas in the
VOST runs was used  to prepare the calibra-
tion standards for  quantification of the
POHCs on the traps.

     The problem of analyzing the wet sam-
ple traps was overcome by desorbing the
contents of the wet traps into an aqueous
purge and trap apparatus.  Since the purge
and trap technique  initially appeared to
offer minimal risk  of losing or affecting
the very small amounts of each compound to
be quantified (i.e., 2 to 10 ng of POHC),
and was basically consistent with an ac-
cepted EPA method,  it was used in this
evaluation.  The Nutech apparatus was ini-
tially tested in the normal cryogenic trap-
ping configuration, but the desorbed water
froze and clogged the analytical system.
Other wet trap analysis techniques were
considered but not  investigated because of
lack of time and possible associated pro-
blems .

3.4.6  Data Reduction--

     The POHCs in the samples were quanti-
fied using the internal standard technique.
The area of the masses of m/z 62 for vinyl
chloride, m/z 117 for carbon tetrachloride,
m/z 130 for trichloroethylene, m/z 112 for
chlorobenzene, and  m/z 186 for the per-
fluorobenzene internal standard were used
to calculate response factors from analy-
sis of the 8.4- and 84-ng calibration stan-
dards according to  the equation:
     The amounts of the POHCs in the sam-
ples were then calculated according to:
     Response Factor  (RF)  =
                                   (Cs)
where:   A   =  The  area  of  ion  for the POHC.
        A.,_  =  The  area  of  the  ion for the
                 internal standard (m/z  186).
        C.,_  =  The  amount of  internal stan-
                 dard  (25 ng).
         Cg  =  The  amount of  POHC in the
                 calibration  standard (gen-
                 erally  8.4 or  84 ng).
     Amount of POHC =
                           (cis)
                      (AIS) (RF)
     The reproducibility of the internal
standard area counts (m/z 186) was very
good with an average variation of only
about 10% for a daily batch of samples at
a given multiplier setting.  There did not
appear to have been any particular dif-
ference in perfluorobenzene area counts
with the type of trap;  e.g., wet traps or
Tenax/charcoal traps did not give lower
area counts than dry traps and Tenax-only
traps.  This is in contrast to the POHC
vinyl chloride which gave higher response
on Tenax than on Tenax/charcoal traps dur-
ing analysis of calibration standards.
The reason for this is  as yet undetermined.
There is no reason to believe that the
vinyl chloride broke through the Tenax to
the charcoal because only about 15 ml of
gas passes through the  trap during spiking.

3.5  Results

     It was noted in Section 2.0 that a,
total of 10 runs were made during this
evaluation, involving four different con-
centration levels of the POHCs and three
blank runs.  After each run, the traps from
each of the three trains were analyzed and
results reported for the traps as pairs
(i.e., Tenax trap plus  Tenax/charcoal trap)
Not all pairs were analyzed, especially at
the higher concentration levels, where it
was not necessary to desorb several pairs
onto one pair.  In most cases, at the lower
concentration levels, one pair of traps
from each train were analyzed individually,
while all remaining pairs from each train
were desorbed onto a single pair for analy-
sis.  Several calibration standards were
analyzed along with the traps from each
run in order to be able to quantify a wide
concentration range.

     Results were initially calculated
after each run based on calibration stan-
dards analyzed with the samples from each
run.  After all runs and analyses had been
completed, the results were recalculated
based on the average mass spectrometric
response factor of all the appropriate cal-
ibration runs using the internal standard
technique described earlier in Section
3.4.6.  These response factor data are
                                            11

-------
shown in Table  2.   The  data  in  Table  2  also
show transfer efficiencies determined for
desorbing several  pairs of traps  with re-
adsorption onto a  single pair.  Since the
transfer efficiency for vinyl chloride  was
relatively low  (49%),  the reported values
for vinyl chloride were corrected for this
low transfer efficiency.   Also, the data
for the four POHCs were blank-corrected,
as discussed below.

     Three blank runs  were  carried out
using the gas generation system and three
VOSTs,  but without any injection  of the
solution containing the POHCs into the  sys-
tem.  The results  for  these  blank runs  are
shown in Table  3 and include analyses of
single pairs, and  several pairs combined
onto one pair.   As can be seen  in Table 2,
most of the blank values are relatively
low, but are still significant  relative to
the run at the  lowest concentration level
where the expected amount of any  POHC on
each pair was only about 2 ng.   In this
regard, the blank values for carbon tetra-
chloride in Runs 7 and 10 are higher than
the expected value.  Thus, it was not pos-
sible to blank-correct the carbon tetra-
chloride results obtained in the  lowest
level run  (Run 4), which makes  it difficult
to  make any  definitive conclusions about
using the VOST train for detecting such
low levels of  carbon tetrachloride.

     The problem with  the high carbon
tetrachloride  blanks was evident  after
Run 7,  and therefore another blank run was
made  (Run  10), after the gas generation
system  and the trains  were purged with
vapor  from the liquid  N2 tank at room tem-
perature  for 24 h.  However, the  blank
carbon  tetrachloride values were  again
found  to be  high  in Run  10.   Other blank
traps were analyzed which had not been ex-
posed to the gas  generation system, but
had been exposed  to room (laboratory) air,
and no  POHCs were  detected in these blanks
(i.e.,  < 0.5 ng).   The absence of carbon
tetrachloride  in  the blanks suggested that
the high blank values  for carbon tetra-
chloride resulted  from within the gas gen-
eration system or  the  sampling trains and
was  not a  result  of any  subsequent analyt-
ical procedures or contamination from the
ambient room air.

     Except  for the carbon tetrachloride
data from the  lowest level run, all uncor-
rected  and blank-corrected results were
tabulated, with the corrected values being
used to compute the results as  a  percentage
of the expected value.  These tabulated
data are summarized in Table 4.  The data
in Table 3 provide information on results
computed as averages but do not show the
range in results.  The compounds are dis-
cussed individually below.

3.5.1  Vinyl Chloride--

     Figure 5 (for vinyl chloride)  shows
all results, blank-corrected and corrected
for the 49% transfer efficiency when trans-
ferring the contents of several pairs  of
traps onto one pair.

     The results for vinyl chloride at the
0.1 and 1.0 ng/liter gas phase concentra-
tions appear to be similar, with total re-
coveries when analyzing single pairs rang-
ing from 48 to 95% of the expected  value.
When combined pairs were used the recov-
eries ranged from 48% of the expected  value
up to 148%.  Conversely, at the 10  ng/liter
level where only single pairs were  analyzed,
all except one data point are greater  than
the expected value, ranging from 100 to
180% of the expected value.  This is a
rather wide range, but vinyl chloride  is
very volatile, and it is commonly recog-
nized that analyses for this compound  are
difficult.

     At the highest concentration level
(Level IV, 100 ng/liter gas-phase concen-
tration, 2,000 ng/pair of vinyl chloride
expected on the  traps), the results were
consistently low  (~ 48% recovery).  Al-
though nearly all of the other POHCs were
consistently found on the first Tenax  trap
of any pair, most of the vinyl chloride
was found on the backup Tenax/charcoal
trap.  The data  thus  suggest that break-
through or irreversible adsorption  of  the
vinyl chloride occurred at the highest con-
centration level.  Thus in any further
testing, one should be aware that this may
occur when using the VOST method  at high
concentrations of vinyl chloride.

3.5.2  Carbon Tetrachloride--

     The results  for  carbon  tetrachloride
are shown  in Figure 6.  These  data  are all
blank-corrected  except  for the data at the
lowest concentration  level.  As  a  conse-
quence, the  data at the 0.1  ng/liter  level
exhibit some very high values, which  un-
doubtedly  are  not representative.

-------
                   TABLE 2.  GC/MS RESPONSE FACTOR AND THERMAL DESORPTION COLLECTION
                               EFFICIENCY FOR FOUR VOLATILE POHCs

Compound
Vinyl chloride
Carbon tetrachloride
Trichloroethylene
Chlorobenzene
Avg
RF a
0.140
0.197
0.490
0.338
a % RSDb
± 0.042 30
± 0.069 35
± 0.068 14
± 0.065 19
c
n
15
13
14
14
RFf
Type of following
trap desorption
T/Cd 0.069
Te 0.143
T 0.435
T 0.331
Desorption
transfer
n efficiency
2 49
2 73
2 89
2 98

Mass spectrometric response factor relative to
perf luorobenzene .
RSD = Percent relative standard deviation, which equals
p
n = Number of determinations
d T/C = Tenax/charcoal
P
T — AIT T^T-.OV ^^-^n ( '
(~ 70:30
i £ ,»•*
(includes 8.4 and
v/v) trap.

a ~ Mean x 100.

84 ng calibration standards).






Contents of calibration standard thermally desorbed onto type of adsorbent
trap in previous column.

-------
           TABLE 3.  TABULATION OF DATA FROM BLANK RUNS

No. of
Blank combined
run No . pairs



7
7
7
10
10
10
Avg for single pairs

2 4
2 4
7 5
7 5
7 5
10 5
10 5
10 5
Average per pair
for combined pairs


Vinyl
chloride 1
Single i


< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
< 0.5
0
Combined
5.4
4.8
4.7
1.1
Lost
< 0.5
< 0.5
< 0.5
0.5


Amount detects
Carbon
:etrachloride
sairs data

No single pairs
12.5
3.6
3.5
1.2
3.7
< 0.5
4.1
pairs data
< 0.5
< 0.5
35
49
17
53
50
54
6.8


;d (ng)
Trichloro-
ethylene


analyzed
0.7
0.2
< 0.5
3.4
< 0.5
2.0
1.1

2.3
2.5
5.0
6.6
6.3
4.8
2.2
3.0
0.9



Chloro-
benzene



0.8
0.6
< 0.5
< 0.5
< 0.5
< 0.5
0.2

8.0
6.4
8.1
10.0
4.0
1.8
3.6
6.2
1.3

Amounts calculated based on average  response  factor determined from
all standards (response factor for VC was  based on analyzing stan-
dards on Tenax/charcoal traps; the other compounds were based on
analyzing standards on Tenax traps).

Less than values were assumed to  be  zero in order to compute an
average.

-------
                                                       TABLE  4.    VOST  DATA SUMMARY


Run
No.
2
7
7
10
10
4
4

5
5
8
8
9
9
3
3
6
6
Expected
concentration
of each compound
(ng/litrr)a
0
0
0
0
0
0.1
0.1

1.0
1.0
1.0
1.0
1.0 (with HC1)
1.0 (with HC1)
10.0
10.0
100.0
100.0
cartridge
analyzed
(i.e.,
replicates )
2
3
3
3
3
2
3

3
3
3
3
3
3
7
-
6

No. of b
pairs
combi ned
for analysis
4
0
5
0
5
0
5
(4 in 1 set)
0
4
0
4
0
4
0
0
0
0

Average amount .found , in ng, - blank
and (% of average expected ^
corrected ,
alue)
expected Vinyl chloride Carbon tetrachloride Trichloroethylene Chlorobenzene
value Single
(ng) pairs
0
0
0
0
0
1.9
9.1

19
77
20
79
20
77
193
-
2,020

-
< 0.5

< 0.5

1.5(79%)


12(63%)

17(85%)

19(95%)

Combined Single
pairs pairs
5.1

2.9

< 0.5

10.1(111%)


37(48%)

115(146%)

61(79%)
274(142%)
870(43%)

6.5

1.6

4.2(221%)d


'9(47%)

11(55%)

8(40%)

Combined Single
pairs pairs
< 0.5

34

52

16(176%)d


68(88%)

89(113%)

85(110%)
136(70%)
2,180(108%)

0.3

1.8

1.5(79%)


22(116%)

23(115%)

19(95%)

Combined Single Combined
pairs pairs pairs
2.4

6.0

3.3

8.8(97%)


83(108%)

83(105%)

81(105%)
210(109%)
2,660(132%)
7.2
0.5
7.4
< 0.5
3.9
1-8(95%)
9.5(104%)

29(153%)
101(131%)
21(105%)
91(115%)
18(90%)
74 (96%)
204(106%)
"
2,050(101%)

   Gas  volume  in liters refers  to dry standard conditioas  (20°C,  1 atm).




   Cartridge pairs refers to one Tenax cartridge and  an  associated Tenax  +  charcoal cartridge.




C  Data tor vinyl chloride include  correction for 49% transfer efficiency  when several pairs  are desorbed onto one  pair  (per Table 1).




   All  values  are blank-corrected except for carbon tetrachloride in Run  4,  due to large average blank value (per Table 2).

-------
260
240

220


200
~D
1! 18°
u
4>
x- 16°
UJ
«-*_
o
•£ 140
u
4)
^ 120
_, o
o>
i nn
D IUU
-
_

—


—

-

—

a
~

n
~


1

> 1 10
-a
4)
S 80
D
*!
60


40
20
0

O

D
O
-


—
-

rl r- v <~ ! 5 P) n fr n k
UC V tJ 1 I L/LIILI ^
Expected value
near 2 ng/pair











D
D
0 D



D

1
1
8 100

° a

o o
o
o
o D

D
%


^ Irvrl II Dntn fel
Expected value
near 20 ng/pair








O
o
0





o


o
0 1
° 1
1000











4 1 f"T 1 1 11 Hnln b,
Expected value
near 200 ng/pair


O Single Pair Data
D Combined Pairs Data
Note: Data have been
blank-corrected.













Expected Value (ng)
1
10,000






0

°c£?


4 Level IV Data
Expected value
near 2000 ng/pair
Figure 5. Vinyl chloride test results.

-------
260

240

220


200

„ 120
D
(U
-5 100
O
D
-

~~


—


D
O
-


_




—

D

> 1 10
-a
 1 1 IB Onf n k
Expected value
near 200 ng/pair


O Single Pair Data
D Combined Pairs Data
Note: Level II, III, and IV
data have been blank-corrected.
Level 1 data were not blank-
corrected because correction
was large relative to measured
value and would, in several
cases, have resulted in negative
values .





O



O
00 Expected Value (ng)
	 — e 	 1 	 "-
o 10,000












^ |n|(n| jyy p^ t
Expected value
near 2000 ng/pair
Figure 6. Carbon tetrachloride test results.

-------
     Data at the 1.0 ng/liter level are
similar to that found for vinyl chloride,
in that all data for single pairs are less
than the expected value, but data for com-
bined pairs range from 60 to 170% of the
expected value.  This phenomenon is not as
yet explainable, but most probably relates s
to the quantity present on any trap being
analyzed and the characteristics of the
purge-trap-desorb and GC/MS analysis method.
As originally conceived, the intent would
be to rely on results for combined pairs
at low levels, which does seem to be sup-
ported by the data.

     At the 10 ng/liter level, data for
all the single pairs, except one, were less
than the expected value, ranging from 36
to 120%.  Thus, a result for any single
pair might be quite low, but it is antici-
pated that, in any field testing, results
would be based on the average of the analy-
sis of several pairs, which in this case
would have yielded an average value of 70%
of the expected value.  It is evident in
Figure 6 that, at the highest concentra-
tion level  (100 ng/liter), all the single'
pair results were quite close to the ex-
pected value.

     In summary, the results for carbon
tetrachloride do not indicate any major
deficiency  in the VOST method, except for
the relatively high blank values and their
effect on results at the lowest concentra-
tion level.  If these blank values were
due to the  gas generation system, then high
blank values might not be a problem in any
field testing.  However, if the high blanks
somehow resulted from the sampling trains,
further work would be needed to determine
how trains  should be cleaned and prepared
prior to each test to minimize blank pro-
blems.

3.5.3  Trichloroethylene--

     Results for trichloroethylene, given
in Figure 7 (blank-corrected), show a much
narrower range  at  all concentration levels
than did the  results for vinyl chloride or
carbon tetrachloride.   The extremes varied
from 70% of the  expected value  (at the 0.1
ng/liter level  for  combined pairs), up to
slightly above  140% of  the expected value
 (at the  100 ng/liter level  for single
pairs).  These  results  appear  to be quite
good for this  compound  using  the VOST
method.
3.5.4  Chlorobenzene—

     Results for chlorobenzene, given in
Figure 8 (blank-corrected),  are not as nar-
row as for trichloroethylene, but do show
decreasing variability with increasing con-
centration levels.  Again, data at the low-
est concentration level showed the greatest
deviation from the expected value, ranging
from 36% up to 173% of the expected value.
However, at the next higher concentration
level (1.0 ng/liter), data for combined
pairs ranged from 88% to about 140% of the
expected value.  At 10 ng/liter, the re-
sults for the single pairs were about the
same, ranging from 70% to about 140% of
the expected value.  As is evident in Fig-
ure 8, the range at the highest level was
very narrow.

3.6  Summary .and Interpretation of Results

     The preceding sections have shown
that, at the two lowest concentration
levels, the results for combined pairs of
traps might range from 38 to 173% of the
expected value (excluding higher values
for carbon tetrachloride at the lowest
level, which were not blank-corrected).
However, the data presented earlier in
Table 4 show that if three trains (or
three runs) are used, the average for
combined pairs may range from  48% up to
146% of the expected value.  At the two
highest concentration levels,  the average
for several pairs analyzed individually
ranged from 70% up to 142% of  the expected
value (excluding a 43% average value for
vinyl chloride at the highest  concentra-
tion level where it appears that break-
through or irreversible adsorption oc-
curred) .

     If one assumes,  for  simplicity of
number, that the average value from three
tests may span a range of 50 to 150%, it
is possible to determine  the implications
on a subsequent calculation of DRE based
on that range, as explained in the two
scenarios given below.

     In the first scenario, one may be try-
ing to determine DRE  for  an  incinerator
that is actually  achieving 99.999% for a
POHC present in the waste at the  low  con~
centration  of  100 ppm.  As mentioned  ear-
lier, the approximate resulting true  con-
centration  of  that POHC in the stack
effluent would be about 0.1  ng/liter.  If
this gas is sampled  over  a 2-h period using
                                            18

-------
 OJ
_D

 D


-o
 0)

 O
 0)
 Q-
 X
LU
 0)
Q_
-D
 0)
260
240

220

200
180
160

140


120

100

80

60
40
20
0




_





-




— D

10
1
A

8
D




•4fl LC V C 1 B 0 Q f Q fc
Expected value
near 2 ng/pair










°n
^j
a
D
o a
oo m .
o a n 1
°~ " Dioo
o






4 Level il Data t
Expected value
near 20 ng/pair












Q
D
O
°0 I
O 1
1000







4 Level III Data »
Expected value
near 200 ng/pair

O Single Pair Data
D Combined Pairs Data
Note: Data have been
blank- corrected.



o
0
93
cP


Expected Value (ng)
10,000







^ Level IV Data
Expected value
near 2000 ng/pair
Figure 7. Trichloroethylene test results.

-------
rss
o
260
240

220

200
V
D
O
S 180
o
 1 ° 10
-o
0)
5 80
a
a>
60
40

20
0



•~




0


•4 	 1 f>v/f» 1 1 Hnfrn 	 ^
^l LC VCIIL'MIU F^
Expected value
near 2 ng/pair










o

o


0 D
D
a
U

o
a n
n
n00 100
o a








•4 \K ••«! II Dntn b
Expected value
near 20 ng/pair















o
0


o

n

o 1000


o







Expected value
near 200 ng/pair

O Single Pair Data
° Combined Pairs Data
Note: Data have been
blank-corrected.














O
o
Expected Value (ng)
O
o 10,000
0








L*_ 1 1 l\ / r»
^ 	 Level IV Uata 	
Expected value
near 2000 ng/pair
                                                   Figure 8.  Chlorobenzene  test  results.

-------
five pairs of traps each time, at 20 rain
each pair and a flow of 1-liter/min, the
results from combining five pairs should
be 10 ng.  However, if the average for
three tests (three runs) at this low con-
centration ranged from 50 to  15070, the re-
ported value would be between 5 and 15 ng.
As a result, the computed DRE would be:
     Average amount
        detected

            5
           10
           15
Computed
 DRE (%)

 99.9995
 99.9990
 99.9985
      In this  first  scenario  it  is  clear
that  the  sampling/analysis method  does al-
low an accurate  determination of DRE, and
minimizes  the need  to  report a  DRE value
as "greater than" 99.99% when it is  actu-
ally  achieving 99.999%.

      As a  second scenario, the  situation
might be  that the waste again contains 100
ppm of another volatile POHC, but  the in-
cinerator  is  actually  achieving a  DRE of
99.99%.   In this case, the amount  of POHC
present from  combining five  pairs  of traps
should be  100 ng.   Since the average for
three tests at this level may again  range
from  50 to 150%  of  the true  value, the re-
ported value  might  be  as low as 50 ng or
as high as 150 ng  (i.e., 100 ng ±  50).  As
a  result,  the computed DRE would be:
      Average  amount
         detected

            50
            100
            150
Computed
 DRE (%)

 99.995
 99.990
 99.985
      From  the  above,  it  can  be  concluded
 that:

         The  sampling/analysis method  does
         provide  assurance  that  the  com-
         puted  DRE  is  accurate to  the  same
         decimal  place as the true DRE, even
         if the true DRE  is as high  as
         99.999%.

         The  computed  DRE could  be as  low
         as 99.985% for an  incinerator that
         is actually achieving 99.99%.

      This  second conclusion  is  vitally im-
 portant  since  current regulations stipulate
 a DRE  of 99.99%.   Data obtained in  this
                 project make it appear unlikely that a
                 computed DRE would be below 99.985% for an
                 incinerator that is actually achieving
                 99.99%.
4.0  DEVELOPMENT OF A FIELD VERSION OF THE
       VOST

     Following the successful laboratory
evaluation of the VOST described above,
the VOST concept was chosen to be evalu-
ated under field sampling conditions along
with integrated gas bags.  However, the
laboratory version of the VOST was not
deemed appropriate for field use for the
following reasons:

        The difficulty in changing traps
        under field conditions.

        The lack of ruggedness of the sam-
        pling train.

        The high potential for contamina-
        tion of the outside surfaces of
        the traps in the hostile environ-
        ment of the stack and from han-
        dling the traps.

As a result of the need for a more rugged
VOST with protected traps, several pos-
sible VOST and trap designs were considered
and evaluated.  This paper will only de-
scribe the final field version of the VOST
which is being used at all trial burns con-
ducted by MRI.

4.1  Trap Design

     Figure 9 shows the various components
of the field adsorbent traps used in the
VOST.  The following items should be noted.

        The dimensions of the glass tube
        remain the same except that neither
        end is nippled  (10-cm x 1.6-cm ID
        glass tube).

        The amount of Tenax and Tenax/char-
        coal remains the same.

        The Tenax and charcoal are held in
        the tubes with a fine-mesh screen
        held by a C-clip both made from
        stainless steel.  These supporting
        materials hold the adsorbents more
        uniformly inside the tubes than
        the glass wool used during the lab-
        oratory evaluation.  This  results
        in a lower likelihood of channeling
                                            21

-------
Figure 9.  Components of field adsorbent traps for the VOST.

-------
        and lower retention of water in
        the trap.  The stainless steel sup-
        ports were found not to cause any
        degradation of volatile POHCs from
        thermal desorption during analysis.

     •   The glass tube containing the ad-
        sorbents is held within a larger
        diameter outside tube using Viton
        0-rings.  The purpose of the out-
        side glass tube is to protect the
        outside of the adsorbent-containing
        tube from contamination.

        The glass tubes are held in a
        stainless steel carrier.  The glass
        tubes each butt up against Viton
        0-rings which are held in machiaed
        grooves  in each metal end piece.
        A set of three cylindrical rods
        are secured into one of the end
        pieces and fasten to the other end
        piece with threads and nuts, thus
        sealing  the glass tubes.

        The end pieces, which are fitted
        with a 1-in.  (2.54 cm) female nut,
        are capped during transport and
        storage with an end-cap which also
        seals with a Viton 0-ring.

4.2  VOST Design

     A photograph of the field version of
the VOST is shown in Figure 10.  The upper-
most section of  glass tubing attaches to
the probe which  is inserted into the stack
to collect the sample.  The hot wet stack
gases, which are drawn into the VOST by
the air pump in  the lower right-hand part
of the photograph, are cooled in the first
spiral condenser at the upper left.  The
bottom portion of the open case is filled
with ice water which is continually circu-
lated by a small water pump.  The condensed
water and stack  gas then pass down through
the front Tenax  trap where most of the or-
ganics are adsorbed except those with very
low breakthrough volumes; e.g., vinyl
chloride.  The condensed water collects in
the Erlenmeyer flask-shaped impinger and
is continually purged by the sampled gas.
Any volatile POHCs which pass through the
front Tenax adsorbent trap with the water
are then purged  from the water and pass
upward through the Teflon® tube, down
through the second spiral condenser and
through the backup Tenax/charcoal trap
where they are adsorbed.  The gas is then
dried in the silica gel tube and passes
into the dry gas meter for volume measure-
ment.  When not in use, the VOST folds up
inside the portable case for easy trans-
port.

     The field VOST is generally used as
described-in the laboratory evaluation;
i.e., one pair of traps is sampled for 20
min at a flow rate of 1 liter/min.  The
first trap pair is then removed and a new
pair inserted for sample collection.   A
total of six pairs of traps are collected.
The changing of the trap pairs is greatly
facilitated by using the field carrier.

     A "slow VOST" is also being evaluated
during which only two or three pairs of
traps are used for sample collection.  The
slow VOST, which generally samples only
5-10 liters of stack gas sample over a
longer sampling period, has the following
advantages:

        The lower sample volume reduces
        the likelihood of breakthrough and
        serves as a check on breakthrough
        for the regular VOST.

        A more integrated sample is ob-
        tained.  This is very advantageous
        in situations where the stack gas
        composition changes during the in-
        cineration test.

The main disadvantage of the "slow VOST"
is its decreased sensitivity

4.3  Trap Preparation Procedures

     During the development and evaluation
of the field VOST, it was discovered that
the  sorbent traps were sometimes severely
contaminated with volatile organic com-
pounds.  Several possible sources of con-
tamination were identified such as ambient
air, contaminated metal carriers, 0-rings,
and  the adsorbents.  In order to prevent
contamination, a series of stringent trap
preparation procedures were tested and
adopted which have proved very effective
in eliminating the contamination  for field
sampling with the VOST.  These procedures
are  discussed below.

4.3.1  Preparation of Tenax and Charcoal--

     New Tenax and charcoal  is Soxhlet-
extracted with methanol for  16 h, and  dried
in a vacuum oven at 50°C prior to packing
into tubes.  The Tenax and charcoal  in
                                            23

-------
Figure 10.   Photograph of the field version of the VOST.
                            24

-------
packed tubes are not routinely reextracted
following sampling and analysis unless very
high concentrations (i.e., micrograms) of
sample components are collected.

4.3.2  Preparation of 0-Rings—

     The Viton 0-rings are thermally con-
ditioned in a vacuum oven at 200°C for 48 h
prior to use.  This procedure removes vola-
tile solvents which may be present in the
0-rings and could outgas later.

4.3.3  Preparation of Metal Parts—

     The metal parts (including the stain-
less steel carriers, end plugs, C-clips,
and screens) are subjected to sonification
in a warm non-ionic soap solution, rinsed
with distilled water, air-dried, and heated
in a muffle furnace at 400°C for 2 h.

4.3.4  Preparation of Glass Tubes--

     The glass tubes are cut from new glass
tubing, fire-polished, and annealed.

4.3.5  Packing--

     The Tenax and charcoal are packed into
the glass tubes in an organic-free labora-
tory (laboratory air filtered through char-
coal) .

4.3.6  Trap Conditioning--

     The traps are conditioned as de-
scribed in Section 3.4.3.  However, two
different conditioning periods are used of
at least 4 h each.

4.3.7  Trap Assembly--

     The conditioned traps are assembled
into the metal field carriers in the same
organic-free room where the adsorbents are
packed into the glass tubes.

4.3.8  Leak Checking--

     The assembled field traps are checked
for leaks by removing one of the end caps
and attaching the trap to a source of
organic-free nitrogen gas at 30 psi  (2.1
kg/cm2).  The trap is then immersed in dis-
tilled water to check for the appearance
of bubbles.
4.3.9  Trap Monitoring--

     Following trap assembly and assurance
that the traps do not leak, each trap as-
sembly is attached to a manifold (capacity
of 10 traps).  Organic-free nitrogen is
passed through each trap at a flow rate of
30 ml/min while the traps are heated to
190°C.  The flow through each trap is se-
quentially monitored with a flame ioniza-
tion detector to check for emission of
volatile organics from the trap assembly.
Most traps show no organic emissions,
while others need to remain on-the condi-
tioner for several hours until the emis-
sions from the trap are reduced to less
than a detectable level (< 2 ppb).

4.3.10  Trap Storage--

     When the traps are shown to be clean
with the flame ionization detector, they
are capped and stored under ice water until
they are used for sampling.  The traps are
also placed back under ice water after sam-
pling until they are analyzed by GC/MS.
The ice water serves to keep the'traps cold
which slows aging of the Tenax; i.e., the
gradual transfer of compounds such as
benzene and toluene from within .the poly-
meric Tenax matrix to the surface of the
Tenax where these compounds can be ther-
mally desorbed during analysis and con-
tribute to high background levels.  The
water also protects the traps from vola-
tile organic compounds in the ambient at-
mosphere which could collect on the out-
side of the trap assembly and contaminate
the adsorbents during disassembly just
prior to analysis.  A summary of the trap
preparation procedures is shown in Figure
11.
5.0  CONCLUSIONS AND RECOMMENDATIONS

     The conclusions and recommendations
based on this evaluation of the VOST are
presented below.  Some of the conclusions
are preliminary and could change upon fur-
ther evaluation of the VOST.  We also ex-
pect that the precision and accuracy of
the method will improve during further
evaluation.

        This laboratory evaluation demon-
        strated that the overall concept
        for the VOST is valid, and that
        combining several pairs of traps
        onto one pair of traps for analy-
        sis is advantageous when the POHCs
                                            25

-------
 O - rings
 Tenax 	
 Charcoal
• Glass Tubes-
 Metal Parts—I
 End Plugs •
 C - clips
 Screens
  200°C Vacuum
  48 Mrs
Thermally Condition
(250°C, 4 hrs) x 2
     Alconox
     Ultrasonic
     Dl Rinse
                       Alconox
                       Ultrasonic
                       Dl Rinse
           Store Under Ice Water
Store  in Clean
Container
 Culture Tubes
 in VGA Lab
Oven Dry,
Store  in Closed
Container
                          400 °C Oven
                          2 hrs
                     Check with
                     GC/FID
                 190°C Oven
                with N2 Flow
                      Figure 11.  VOST  trap cleanup  procedure.

-------
are present at low levels.  The
field work thus far, however, sug-
gests that the levels of volatile
POHCs are high enough that combin-
ing the contents of several pairs
of traps onto one pair is gen-
erally not necessary.

The VOST method does overcome the
problem of reporting a DRE value
of > 99.99% for an incinerator
which is actually achieving 99.999%.

Results of the laboratory evalua-
tion indicate that a reported value
may be as low as 46% or as high as
146%, of the expected value (based
on the average of three runs when
several pairs from each sampling
train are combined onto one pair).

Therefore, for an incinerator that
is achieving a DRE of 99.999% (100
ppm concentration of the POHC in
the waste), the VOST method does
permit determination of DRE to the
third decimal place, but with re-
sults that could range from as low
as 99.9985 to as high as 99.9995.

The VOST method does not ensure
that DRE results can always be ac-
curately computed to the third
decimal place.  In fact, if an in-
cinerator is actually achieving a
DRE of 99.990%, the average re-
sults reported for three tests
could have a deviation of 99.990
±  0.005%.

In this evaluation, results for
vinyl chloride and carbon tetra-
chloride show the most variabil-
ity, especially at lower concen-
trations .

The presence of HC1 in the gas
being sampled did not appear to
have any serious effect on the
VOST results.

The problem of analyzing wet traps
can be satisfactorily overcome by
desorbing the contents of the sam-
ple collection traps into a purge-
trap-desorb GC/MS analytical sys-
tem.

Stringent trap preparation pro-
cedures are required to eliminate
the risk of contaminating the traps
prior to use.
        Separate traps (blanks)  should be
        exposed to air in the field in
        order to determine the level of
        compounds on the traps due to ad-
        sorption of the compounds during
        handling of the traps and their
        insertion into/removal from the
        VOST apparatus.
6.0  ACKNOWLEDGMENTS

     Much of the work discussed in this
paper was funded under contract with the
U.S. Environmental Protection Agency (EPA
Contract No. 68-01-5915).   The work was
performed under the direction of Dave
Friedman of EPA/OSW and Larry Johnson of
EPA/IERL-RTP who provided counsel in all
phases of the work.
7.0  REFERENCES

1.    Rechsteiner, C., J.  C.  Harris, K.  E.
     Thrun, D. J. Sorlin, and V  Grady.
     1981.  Sampling and Analysis Methods
     for Hazardous Waste Incineration,
     A. D. Little, Inc.,  in support of
     Guidance Manual for Evaluating Permit
     Applications for the Operations of
     Hazardous Waste Incineration Units,
     EPA Contract No. 68-02-3111, EPA/IERL,
     Research Triangle Park, North Carolina

2.    Jungclaus, G., and P. Gorman.  1982.
     Draft Final Report,  Evaluation of a
     Volatile Organic Sampling Train, Mid-
     west Research Institute, EPA Contract
     No. 68-01-5915.

3.    Krost, K. J., E. D.  Pellizzari, S. G.
     Walburn, and S. A. Hubbard.  1982.
     Collection and Analysis of Hazardous
     Organic Emissions.  Anal.  Chem.,
     S(4):810-817

4.    EPA Method 624   Purgeables.  1979.
     U.S. Environmental Protection Agency,
     Federal Register 44:69532-69539.
                                    27

-------
                    THE FEASIBILITY OF HYDRIDE GENERATION  INDUCTIVELY
                       COUPLED PLASMA SPECTROSCOPY  FOR ANALYSIS  OF
                                      VOLATILE METALS
                M. P. Miller, P. M. Chirm, B. G.  Snyder,  and  A.  K.  Wensky
                             Battelle Columbus Laboratories
                                  Columbus, Ohio  43201
The detection limits normally obtained by  inductively  coupled  argon plasma spectroscopy
(ICAP) for arsenic, selenium, antimony, and mercury were  reduced  by factors of  between
50-500 by use of a commercially available  hydride  generation system (Applied Research
Laboratories).  Coupling of the hydride generator  to the  spectrometer  required  minor
modifications.  Detection limits obtained  in standard  solutions of  As,  Se,  Sb,  and Hg
were 0.2, 0.2, 0.1, and 0.2 yg/1, respectively.  Analysis in some samples  is difficult
due to hydride suppression, particularly from  aluminum and iron.  Sample  clean-up
procedures can be used to eliminate some interferences.
INTRODUCTION

     The method of inductively coupled
argon plasma spectroscopy  (ICAP)  affords  a
rapid, accurate, and precise means for
multielemental analysis of up to  70 ele-
ments in a wide variety of sample matrices.
The detection limits obtained by  this
method are sufficient for most environ-
mental studies, and the relative  freedom
from matrix interferences provides reliable
quantitative analyses.  However,  the
detection limits obtainable by the conven-
tional pneumatic nebulizers commerically
available are inadequate for some elements,
specifically the Group !IVA, VA, and VIA
metalloids.  Of particular concern in
environmental analyses are low level
determinations of arsenic, selenium, and
antimony.  Furthermore, detection of
mercury at sub part-per billion levels is
impossible by ICAP using pneumatic nebuli-
zation.

     The method of hydride generation has
been used for a number of years to deter-
mine sub-ppb levels of As, Se, and Sb when
used in conjunction with atomic absorption
spectroscopy (AAS).  The application of
the AAS method is relatively slow and
expensive since a single element  is
analyzed during each run.  Furthermore,
an aliquot of 25-100 mL of sample is
usually required for each determination,
thus limiting the technique to cases where
large volumes of sample are available.
Recently, hydride generation methods have
been employed in conjunction with induc-
tively coupled plasma spectrometers to
provide rapid, multielemental ultratrace
analysis of hydride-forming metals.
Thompson et al (1) demonstrated the use of
a prototype system and examined the effect
of varying the acid media and optimizing
the operating conditions for several
elements.  Robbins et al (2) examined the
utility of a microwave induced plasma
system for analysis of complex samples  and
found good recoveries in NBS orchard
leaves when using the method of standard
additions.  The hydride generation-ICAP
technique was applied to the analysis of
arsenic, antimony, and bismuth in herbage
after a dry ash procedure with magnesium
nitrate (3).  Wolnik et al increased the
utility of the technique while decreasing
the total analysis time per sample by
designing a tandem nebulizer, permitting
simultaneous introduction of the sample
with a pneumatic nebulizer and with the
hydride system (4).

     The work conducted by these and other
researchers demonstrated the feasibility
and promise of the hydride ICAP technique,
As a result, most ICAP manufacturers began
                                           28

-------
actively investigating the market poten-
tial of a commercial hydride generation
unit designed specifically for ICAP
use.  This paper describes the set-up,
interfacing, and application of a
commercially available unit manufactured
by Applied Research Laboratories, the
first commercially available hydride
generator produced for use with ICAP.  The
modifications required for interfacing
to a different manufacturer's spectro-
meter are given.  A comparison of results
obtained using different  acid media and
the optimization of instrumental para-
meters  are discussed.  Preliminary inter-
ference studies, both  from other hydride
forming elements and from elements
known to suppress the  formation of
hydrides in  atomic absorption studies,
are included.
EXPERIMENTAL SECTION

Spectrometer

     The ICAP/polychromator used in this
study was a Jarrell-Ash Model 975 Atom
Comp, equipped for simultaneous analysis
of 30 elements.  For this study, results
were taken from four channels; arsenic,
selenium, antimony, and mercury.  The
spectrometer was equipped with the
Spectrum-Shifter background correction
capability.  The argon coolant and sample
flowmeters were replaced with mass flow
controllers to provide more precise flow
regulation.

Hydride Generator

     The hydride generation system used
was an ARL Model 341 Continuous Hydride
Generator, manufactured in England.  The
system consists of a three head peristal-
tic pump, a sample and reagent mixing
block, and a phase separator in which the
liquid and gas phases are separated after
reduction with the sodium borohydride.

     In normal operation, the hydride
generator works in the following fashion.
A stream of sample is drawn into the
mixing block by one channel of the peri-
staltic pump at about 6.0 mL/min.   One
reagent channel contains the sodium
borohydride reagent which is mixed with
the sample stream.  The second reagent
channel can contain hydrogen peroxide,
which is required for hydride determina-
tion of tin.  In this study, the hydrogen
peroxide stream was not required, rather
deionized water was fed through this
channel.  Upon mixing with the sample,
the sodium borohydride solution causes
rapid formation of gaseous hydrides and
excess hydrogen.  The gases and the solu-
tion enter the phase separator, where the
solution exits to a drain and the gases
are swept into the ICAP torch via an argon
carrier gas stream of approximately
1.0 L/min.

Mercury Determination

     Although mercury does not form a
hydride under the conditions generated,
the reducing environment developed by the
borohydride solution is sufficient to
reduce mercury to the metallic state (5).
The gas generated from the metallic
mercury is swept into the plasma in a
manner analogous to that used in cold
vapor atomic absorption spectroscopy.  Thus,
an added benefit of the hydride generation
procedure is the concurrent determination
of mercury.

Reagents

     Deionized water was obtained from a
Barnstead Nanopure® system, with a reading
of 15 Mfi or higher.  The nitric, hydro-
chloric, and perchloric acids used were
Baker Ultrex® grade.  Sodium borohydride
was obtained from Alfa Chemicals.   A two
percent w/v solution of borohydride was
prepared and stabilized with a trace of
NaOH.  The solution was filtered through
Whatman No. 1 filter paper prior to use.
The borohydride solution was prepared
fresh daily.

Hydride Generator Interfacing

     The ARL Model 341 hydride generation
system was designed to connect directly
to the ARL line of ICAP spectrometers.
Some modifications were required for use
with the Jarrell-Ash Model 975 spectro-
meter system in our laboratories.   First,
the hydride unit is equipped with an
interlock to prevent generation of hydro-
gen in the torch chamber if the torch is
not lit.  The interlock consists of an
optical sensor which must be aimed
directly at the ignited torch.  If the
torch is extinguished, the sensor
immediately stops the peristaltic pump,
                                           29

-------
preventing the further generation of
hydrogen.  In the Jarrell-Ash spectrometer,
the optical sensor was easily mounted
inside the torch housing, immediately
adjacent to the image focusing mirror.
A small hole drilled in the mirror access
door was used to pass the cable from the
spectrometer to the hydride generation
unit.

     In order to transfer the hydride
generated into the base of the torch, a
special glass apparatus was designed.
Originally, a length of tubing equipped
with a .glass ball joint was connected from
the outlet of the generator to the base
of the torch.  However, in order to pro-
vide a stable mounting system, the device
shown in Figure 1 was constructed.  This
apparatus consists of a glass transport
tube connecting the hydride generator
output to the base of the ICAP torch.  The
transport tube is surrounded by and
attached to a glass housing which serves
to support the torch assembly.  This
housing fits in the J-A holder for the
expansion chamber used with the conven-
tional pneumatic nebulizer.  Thus,
replacing the expansion chamber with the
hydride transport system and connecting
sample flow lines converts the ICAP from
hydride to conventional.  This interchange
requires less than five minutes.
                           X
     Initial work with the hydride ICAP
system resulted in a poor signal to noise
ratio for arsenic, and poor stability of
the plasma itself.  This problem was the
result of a build up of pressure in the
phase separator due to a blockage from
excess liquid.  In order to prevent this
problem, the drain system was modified,
incorporating a pressure by-pass system
which prevents blockage of the phase
separator.  The by-pass system is shown
in Figure 2.  Once this system was in use,
the plasma stability improved signifi-
cantly, greatly improving the hydride ICAP
performance.

Optimization

     Optimum operating conditions were
determined for the four elements selected
 (As, Se, Sb, Hg) by comparison of the
signal to noise ratios obtained under each
set of conditions.  The parameters in the
optimization procedure included the plasma
gas  flow rate, the carrier gas flow rate,
 the  plasma forward power, and the lateral
and vertical torch positions.  The optimum
set of compromise conditions is given in
Table 1, selected for maximum signal to
noise for all four elements.

Acid Matrix Optimization

A brief literature search indicated an HCL
matrix should be most suitable for multi-
elemental hydride analysis.  A preliminary
study was conducted using 0.5 percent,
5 percent, and 10 percent HC1 solutions
as the matrix for the elements of interest.
The best signal to noise was obtained with
the 5 percent HC1 solution for As, Se, and
Sb.  Initial studies indicated the HC1
matrix was unsuitable for use with mercury,
while a 5 percent HN03 matrix was reliable.
However, once the pressure bypass system
had been installed on the phase separator,
stable determinations for mercury were
obtained in a 1 percent HC1 medium.  This
matrix was used for further studies.  The
instrument was calibrated using a blank
and a single standard containing 200 ppb
of all four elements of interest.  Studies
showed this to be near the upper end of
the linear calibration range for all
elements.
RESULTS AND DISCUSSION
Interelement Interferences
     The instrument was calibrated as
described above.  Subsequently, single
element solutions prepared at 200 ppb were
run to determine whether interferences
occurred for any of the remaining three
elements.  The results of this study are
given in Table 2.  These data indicate
that no significant interelement inter-
ferences occur among these four elements,
thus the simultaneous analysis of these
four elements in solution by hydride
generation ICAP is feasible.

Calibration Curve and Precision Studies

     Standards were prepared containing
all four elements over a concentration
range of 5-200 ppb.  The calibration
curves were run on three separate occa-
sions and the results for each element
at each concentration were averaged.  The
average and relative standard deviations
for the observed concentrations are given
in Table 3.  The low relative standard
deviations obtained by these triplicate
                                           30

-------
determinations of mixed standards indicate
the precise analytical capabilities of
this method.  Furthermore, a comparison of
the observed average concentrations with
the known concentrations indicate the
accuracy of the technique in the ideal
case, i.e. standards in 1 percent HC1.

     A least square fit was performed on
the data given in Table 3 for each of the
four elements.  The parameters describing
the best fit line are given in Table 4 for
each element.  Note that the coefficient
of correlation over the tested concentra-
tion range is 0.9998 or greater in all
cases.  Further, the y intercept is near
zero in all cases.  These data correspond
to a linear dynamic range of at least
0-200 ppb.  However, analysis of samples
containing 200 ppb levels of the
analytes has indicated that a long rinse
time (>3 min) is required between samples
to prevent carryover.  Samples containing
above 200 ppb can easily be analyzed by
conventional pneumatic nebulizer ICAP.

     The most frequently used method for
reporting detection limits for ICAP is to
calculate the analyte concentration
giving a signal equivalent to three times
the standard deviation of the blank.
Based upon this convention, the detection
limits obtained by the hydride generation-
ICAP method are as given in Table 5.

Analysis of Complex Samples

     Initial studies with complex samples
have concentrated on the analysis of fly
ash samples.  Unfortunately, significant
interferences eliminated the possibility
of conducting the analyses.  High levels
of aluminum, iron, and/or copper in the
samples are suspected to be the primary
source of interference.  It should be
noted that these elements routinely cause
significant suppression of the hydride
signal obtained in atomic absorption, and
the method of standard additions is
frequently required for these determina-
tions .  Future work here will examine
potential clean-up steps to eliminate the
requirement for use of standard addition
methods.

     The most promising cleanup technique
is the use of a cation exchange resin to
remove the interfering elements from
solution prior to hydride generation-ICAP
analysis of the sample.  In a recent paper,
Jones et al (6) described the use of
Chelex 100 resin to clean up samples prior
to analysis by hydride generation atomic
absorption.  At pH of 5.3, they found that
most metals were retained on the resin,
while the metalloids passed through the
resin bed quantitatively, thus the inter-
fering elements were removed easily.
Earlier work in our laboratories on sea
water agrees with this finding.  Future
work will include application of the
Chelex 100 resin cleanup procedure to
samples of various types, including
industrial effluents and incinerator
byproducts.
REFERENCES

1.  Thompson, M.,  B. Pahlavanpour,  S. J.
    Walton, and G. F. Kirkbright, Analysts,
    June 1978, Vol. 103, pp. 568-579.

2.  Robbins, W. B., J. A. Caruso, and
    F. L. Fricke,  Analyst, Jan. 1979,
    Vol. 104, pp.  35-40.

3.  Pahlavanpour,  B., M. Thompson,  and
    L. Thorne, Analyst, April 1981, Vol.
    106, pp. 467-470.

4.  Wolnik, K. A., F. L. Fricke, M. H. Han,
    and J. A. Caruso, Anal. Chem.,  1981,
    5J^, pp. 1030-1035.

5.  Hatch, W. R.,  and W. L. Ott, Anal.
    Chem. , 1968, 4_0, 2085.

6.  Jones, J. W.,  S. G. Capar,  and T. C.
    O'Haven, Analyst, April, 1982,  Vol.
    107, pp. 353-375.
                                           31

-------
DISCLAIMER

     Although the research described in
this article has been funded wholly by the
United States Environmental Protection
Agency through Contract No. 68-02-3628 to
Battelle's Columbus Laboratories, it has
not been subjected to the Agency's required
peer and policy review and therefore does
not necessarily reflect the view of the
Agency and no official endorsement should
be inferred.
AKNOWLEDGEMENT

     The work presented in this article
was funded wholly by the United States
Environmental Protection Agency through
Contract No. 68-02-3628 to Battelle's
Columbus Laboratories.   Dr. Larry D.
Johnson and Mr. Frank E. Briden provided
invaluable technical guidance and
generated crucial research ideas.  Their
support is grea'tly appreciated.
                                           32

-------
                            SPECIATION OF HALOGEN AND HYDROGEN
                           HALIDE COMPOUNDS  IN GASEOUS  EMISSIONS
                                      David A. Stern
                                     Barbara M. Myatt
                                    Joseph F.  Lachowski
                                    Kenneth T.  McGregor

                            GCA Corporation/Technology Division
                               Bedford,  Massachusetts   01730
                                         ABSTRACT

     A sampling and analytical method for the speciation and quantification of hydrogen
halide (HX) and halogen  (Xo) emissions  in a gas  stream was evaluated  in the
laboratory.  Analyte gases of certified purity were  introduced into a mixing and
sampling manifold  system, dynamically diluted with air, and sampled into a series of
midget impingers.  Quantification of analyte  species was performed by ion
chromatography.  The methodology was also evaluated  in the presence of S02 and NOX
matrix gases.  The procedures were  effective  in  selectively absorbing HX and X£ from
the gas stream providing  for the speciation of these gases.
INTRODUCTION

     The proper disposal of hazardous
waste  is one of the  important environ-
mental problems of this decade.   In  order
to address  this issue, the United  States
Congress passed the  Resource Conservation
and Recovery Act  (RCRA) in 1976 which
empowered the Environmental Protection
Agency to promulgate  regulations
concerning  the management of hazardous
waste.  To  this end,  the EPA established
standards for waste  generators,
transporters, and operators of hazardous
waste  facilities  (40  CFR 260-267).

     Of the waste disposal methods
presently employed,  incineration  is
emerging as a means  of ultimate disposal
in a safe,  cost-efficient, and
environmentally sound manner [Bonner
(1)].  However, an incinerator burning
hazardous waste must  achieve a
destruction and removal efficiency (DRE)
of 99.99% for each principal organic
hazardous consituent  (POHC) designated in
its permit.  Moreover, if the waste
contains more than 0.5% chlorine,  then
99% of the  hydrogen  chloride produced
during incineration must be removed  from
the exhaust gas (40  CFR 264).  To
demonstrate compliance with these
regulations, sampling and analytical
methods which can quantitate POHCs and
HC1 emissions are required.
     Substances classified as hazardous
(Appendix VIII, 40 CFR 261) encompass a
wide variety of materials having a
correspondingly wide range of combustion
characteristics.  This often results in
incinerator emissions that are
substantially different from site to site
and from burn to burn.  Of the 375
constituents listed in Appendix VIII,
about 100 (or nearly 27%) contain
chlorine.  Additionally,  12% of the
Appendix VIII constituents contain sulfur
and 39% contain nitrogen.  Thus, when
these materials are combusted, a complex
gas stream will be produced which will
likely contain both SOX and NOX.

     Analytical methods for halogen
compounds have not been critically
evaluated in this matrix nor have they
been designed to permit speciation.  As
shown in Table 1, the methods commonly
used for measurement of halogen species
are not analyte specific and are therefore
prone to significant bias.  For example,
in the determination of chlorine by the
o-tolidine colorimetric method, serious
positive errors will occur if bromine,
ozone, or other oxidizing agents are
present in the gas stream [Ruch(4)j.  The
matrix gas,  SO?, will drastically mask
the end point in the mercuric nitrate
titration of chloride (for HC1) causing
positive errors of 100 fold [Cheney and
Fortune (2)].
                                             33

-------
                  TABLE 1.  COMMON METHODS FOR HALOGEN/HALIDE ANALYSIS
Analyte
Absorbing
Solution
Sampling
Rate Method
Interferences
    C12, Br2     NaOH
 1 1pm     Colorimetric
C12,  Br2

  HX



  HX
                 H20
                NaOH
 1 1pm

 1 1pm



1-3 1pm
Colorimetric

Potentiometric,
Volumetric,
Turbidimetric

Potentiometric,
Volumetric
C102, most oxidizing
agents (OA)

C102, OA

Other halides present
Other halides present;
other acidic gases
(S02, H2S)
The purpose of the investigation
reported here is to develop and evaluate
a sampling and analytical scheme for
halogen and hydrogen halide speciation
and quantification that can be applied to
incinerator flue gas streams. The
effects of SOX and NOX as matrix
components are investigated.
Equation 1.

Equation 2.

Equation 3.

D d £i r- f~ -\ i-i
HX(g)

x2(g)

x2(g)

„-.
-------
                            TABLE  2.   PROPOSED  SAMPLING TRAIN
Impinger
#1
Impinger
n
Impinger
#3
Impinger
#4
   Absorber

   Analytes Collected
H20
H20
                                NaOH(O.lM)     NaOH(O.lM)

HC1,  HBr, HF    HC1, HBr, HF    C12, Br2      C12, Br2
Other Species
of Interest
Analytical
Finish
SOX, NOX SOX, NOX SOX, NOX SOX, NOX
ION CHROMATOGRAPHY
     All analytical determinations for
F~, Cl~ and Br~ were performed on a
Dionex Model 14 Ion Chromatograph (Dionex
Corp., Sunnyvale, CA).  Normal operating
procedures employed an anion separator
column and a fiber suppressor with
conductivity detection.  The eluent was
0.003 M NaHC03 and 0.0024 M Na2C03
at a flow rate of 2.3 ml/min.  Part per
million (ppm) sensitivities were easily
obtained.

Procedure

     To evaluate the proposed sampling
train and analytical finish for its
ability to differentiate and quantitate
the halogen/halide species, a laboratory
investigation was performed.  Each
gaseous analyte was introduced into an
all glass manifold system (Figure 1),
diluted with clean/dry laboratory air,
sampled using midget impingers, and
analyzed by 1C.  Analyte combinations at
different concentrations with and without
a matrix gas background of S02 and
NOX were also investigated.  Gas
sampling into the midget impinger train
was performed at either 0.50 1pm or 1.00
1pm for 30 minutes using calibrated
rotameters for flow control.  Using
slightly higher sampling rates, although
possible, ran the risk of carry over and
impinger stem blow out and were therefore
not used in this study.

Materials

     Analyte gases of HC1, HBr, and C12
were custom blended by Matheson Gas
Products (Gloucester, MA) as dilute
                    mixtures in nitrogen.   Their
                    concentrations  were 1.04%,  0.083%,  and
                    0.075% (v/v)  respectively.   Custom
                    blended matrix  gases of S02 (1.59%)  and
                    NOX (4.82% NO,  0.05% N02)  in nitrogen
                    were used to  simulate  matrix background
                    effects (Matheson Gas  Products).

                         Corrosion  resistant gas regulators
                    and flow meters (Matheson Gas Products)
                    were used in all cases.  To minimize
                    analyte gas adsorption, connections
                    between gas cylinder,  manifold,  and
                    impingers were  made with Teflon tubing
                    (1/4" o.d.).  Diluent  air was laboratory
                    compressed air  passed  through Drierite ..
                    and activated carbon prior to
                    introduction into the  manifold.   All
                    reagents used in this  study were ACS
                    reagent grade or better.  The high purity
                    water was equivalent to ASTM Type I water.

                    RESULTS AND DISCUSSION

                         Prior to using the manifold system,
                    a series of preliminary tests were
                    performed to determine qualitatively the
                    clean air purging efficiency of C12 and
                    Br2 from water  impingers.   Four
                    Greenburg-Smith impingers were connected
                    in series; the  first two contained water,
                    and the last  two contained sodium
                    hydroxide.  The first  impinger was spiked
                    with bromine liquid to produce concen-
                    trations of either 1,10 or 100 mg/1.
                    After 20 minutes of clean air purging at
                    500 ml/min (0.5 1pm),  the impinger
                    solutions were  tested for the halide
                    anion using silver nitrate solution.
                    Similar experiments were performed using
                    chlorine-water  in the first impinger.
                                             35

-------
                                                GASEOUS
                                               POLLUTANT

                                             INTRODUCTION
                                                                               MIXING CHAMBER
VENT
CO
CTl
                              CYLINDER-
                              GASES
                                                   Figure  1.   Mixing  and sampling manifold system.

-------
     In each case, the free halogen will
form the halide and the hypohalite ions
which should precipitate upon addition of
AgN03.  In the NaOH impingers, the pH
was adjusted to approximately 8.0 by
addition of concentrated nitric acid to
prevent precipitation of the hydroxide as
AgOH and thus possibly confound the
results.  No adjustments were necessary
in the water impingers.  Table 3 shows
the results for these trials.

     These preliminary results which
suggested successful clean air purging
were sufficiently encouraging to warrant
proceeding with more definitive testing.
Individual test streams of HCl/air,
HBr/air, and Cl2/air were generated by
metering these analytes from gas
cylinders of certified concentration into
the glass mixing and sampling manifold
system  (Figure 1).  The analytes were
dynamically diluted to the desired
concentration by introducing clean, dry
laboratory air.  For example, an HC1 test
stream  of 130 ppm (v/v) was generated by
metering in 10,400 ppm HC1 (in nitrogen)
at 0.050 1pm with air dilution at a rate
of 3.95 1pm.  Total test stream flow was
therefore 4.00 1pm.

     The analytes in question were
sampled at 0.50 1pm and 1.00 1pm into a
series  of midget impingers.  Impingers 1
and 2 each contained 15 ml of water, and
impingers 3 and 4 each contained 15 ml of
0.10 M  NaOH.  All analytical
determinations were performed by ion
chromatography.

     Excellent overall recoveries were
obtained for HCl/air and Cl2/air test
atmospheres (Table 4), demonstrating the
combined ability to generate and
quantitate these process streams.
HBr/air streams demonstrated poor
recoveries suggesting other forces at
work; surface adsorption, sample
degradation, analytical anomalies.

     As shown in Table 4, significant
amounts of residual chlorine remained
trapped in the water impingers.
Concentrations of chloride after 30 min
of purging, 130, showed no difference
relative to TQ.  This experiment was
repeated several times to confirm the
"non-purging" phenomenon.  Table 5
presents absorption efficiency data for
HC1.  For these stream concentrations,
99.8% of the HC1 is collected in the
first water impinger.

     A calculation of the chlorine
collection efficiency, for Cl2/air gas
streams, revealed that the water
impingers collected 37% (average, n = 4)
of the chlorine whereas the alkaline
impingers absorbed 63% as summarized in
Table 6.  Clearly, for speciation
purposes this distribution appeared to
preclude the use of this sampling train
system.   However, from an examination of
Equation 2, it is expected that the
disproportionation of Cl2> and hence
its solubility in water would be
significantly repressed in low pH media.
This suggested that lowering the pH of
the aqueous medium would result in more
favorable Cl2 purging.  In a typical
application, e.g., incineration, the gas
stream will probably contain high levels
of HC1.   Therefore it was expected that
HC1 extant in the flue gas would serve to
enhance the speciation both by lowering
the pH and by common ion and mass action
effects.

     To test this hypothesis, a gas
stream containing 260 ppm HC1 and 19 ppm
Cl2 (in air) was generated and sampled
according to procedures previously
discussed.  The results for two
independent trials are shown in Table 7.
The theoretical amounts were calculated
assuming all of the generated HC1 was
collected in water impingers (1 and 2)
and all of the generated Cl2 was purged
out of water and absorbed in the sodium
hydroxide impingers (3 and 4).  The
somewhat high chloride content (~118%
recovery) found in the water impingers
was not surprising and most probably
reflected the uncertainties of the HC1
rotameter.  The Cl2 recoveries of 81%
and 68% indicated fair speciation
efficiency.

     Further inspection of the data shows
impinger collection characteristics which
provided for even better speciation
efficiency.  Since essentially 100% of
the HC1 is collected in the first water
impinger (Table 5), it is proposed that
the chloride found in the second impinger
originated from Cl2 and not from HC1.
Because the pH of this water impinger was
found to be 4 (as opposed to pH  1 in  the
first water impinger), it is expected
that clean air purging will not  be as
                                            37

-------
       TABLE 3.  HALOGEN SPARGING EFFICIENCY

(Bromine  concentration of 1.0 mg/1 in impinger No. 1
         sparged for 20 min at 500 ml/min)*

Absorber
Color
Initial
Final
AgNO Test
Impinger
#1
H2°
Orange
Colorless
Negative
Impinger Impinger
#2 #3
H 0 NaOH
Colorless Colorless
Colorless Colorless
Negative Positive
*Similar results were obtained using 10 and 100 mg/1 Br2 and 1, 10
100 mg/1 C12.


cone.

TABLE 4. ANALYTE

rate, time,
Run Analyte ppm (v/v) 1pm min
1 HC1 130
2 HC1 130
3 HC1 130
4 C12 19
5 C12 19
6 C12 19
7 C12 19
8 HBr 10
9 HBr 10
ND = not detected
* expressed in term
1.0 30
1.0 30
0.50 30
1.0 30
1.0 30
1.0 30
0.50 30
1.0 30
0.50 30

s of the halide anion

RECOVERIES FROM PROCESS STREAMS
yg Collected*
yg analyte Imp #1 Imp #2 Imp #3 Imp #4
generated* H20 H20 NaOH NaOH
5640 5570 8 ND ND
5640 5430 22 ND ND
2820 2805 8 ND ND
826 234 110 459 21
826 228 114 558 13
826 176 88 635 20
413 110 65 240 21
980 116 8 ND ND
490 45 4 ND ND


Impinger
#4
NaOH
Colorless
Colorless
Positive
, and



7.
Total Recovered
5578 99
5452 97
2878 102
824 100
913 110
919 111
436 106
124 13
49 10


TABLE 5. IMPINGER ABSORPTION EFFICIENCY FOR HC1

HC1 Q t" TOflTTl
concentration
ppm
130
130
130
800
Ave.


Imp #1
H20
99.9
99.6
99.7
99.9
99.8
Percent HC1 absorbed

Imp #2 Imp #3
H20 NaOH
0.1 0
0.4 0
0.3 0
0.1 0
0.2


Imp #4
NaOH
0
0
0
0

                      38

-------
           TABLE 6.  WATER AND SODIUM HYDROXIDE SCRUBBER EFFICIENCY FOR C12*

                             yg Cl~ Collected (% collected)
                                            Run Number

                                          5            6
                                 Ave %
Impingers
H20
NaOH
Total (found)
Total (theoretical)

344 (42)
480 (58)
824
826

342 (37)
571 (63)
913
826

264 (29)
655 (71)
919
826

175 (40)
261 (60)
436
413

37
63


    *C12 concentrations in gas stream was 19 ppm.
                         TABLE  7.   HC1/C12  SPECIATION EFFICIENCIES*
                                                        cr
                                   Combined
                             Water  impingers  (1+2)
                                     (HC1)
               Combined Sodium hydroxide
                   impingers (3+4)
                        (C12)
                            Trial  1
Trial 2
Trial 1
Trial 2
Found
Theoretical**
% Collected
13,980
11,700
119
13,670
11,700
117
685
851
81
575
851
68
      *HC1  and  C12  stream concentrations were  260 and 19 ppm  (v/v) respectively.

     **Calculated amounts assuming complete speciation.
effective in driving out the chlorine
from this impinger.  Using chloride data
from impingers 2, 3, and 4 to calculate
C12 recovery resulted in an average
speciation of 95% as seen in Table 8.
This supports the hypothesis that some
    remains in impinger 2.
     However, because the carryover of
HCl from impingers 1 to 2 and thus the
resultant acidity of impinger 2 cannot be
predicted with great certainty, this
method of calculation is not recommended.
To circumvent this apparent drawback and
to further investigate the effect of pH
on C12 solubility and speciation
     efficiency,  it was decided to  acidify
     impingers 1  and 2 with H2S04  to pH ~1
     prior to sampling.

          In addition, S02 and NOX gases
     were introduced into the process stream
     to document  the speciation of ,BC1/C12
     when these matrix gases are present.
     Table 9 shows the distribution of the
     S02/NOX matrix gases, present as
     dissolved anions, in the impinger train.
     Quantification of these species was  not
     necessary but it was apparent from
     inspection of the ion chromatograms  that
     S02 produced sulfate and lesser amounts
     of sulfite,  whereas, NOX formed mostly
                                           39

-------
                    TABLE 8.   HC1/C12 SPECIATION EFFICIENCY*
                                           ug cr

Found

Trial 1
Trial 2
Imp #1
H20
13,800
13,500
Imp #2
H20
182
165
Imp #3
NaOH
675
570
Imp #4
NaOH
11
5
TotaK
868
740
2,3,4)

Theoretical
% Collected
Trial 1
Trial 2
11,700**
118
115
851**
102 ,
87 (ave.,
95)
 *HC1 and C12 stream concentrations were 260 and 19 ppm (v/v), respectively,
  and sampled at 1 1pm.

**For HC1 and C12 respectively.
           TABLE 9.  MATRIX GAS SPECIE DISTRIBUTION FOR S02 AND NOX*
Impinger
#1
(H20)
2_
so4



Impinger
n
(H20)
2-
N02 , S03 ,
2-
NO, , SO.
3 4
Impinger
#3
(NaOH)

N02 , N03
2-
SO.
4
Impinger
#4
(NaOH)

N03 , S042



     *S02 and NOX stream concentrations were 250 and 600 ppm, respectively.
      HCl and C12 stream concentrations were 530 and 19 ppm, respectively.
nitrite and  some  nitrate.  Using a  "slow"
anion  separator column, excellent resolu-
tion of analyte species was achieved.

     Table 10 demonstrates the  speclation
efficiency for HC1/C12 with prior
acidification of  implngers 1 and 2  In  the
S02/NOX matrix.   As expected,
differentiation between HC1 and C12
species was  successful.  In this trial
excellent recovery of HCl and C12 was
achieved with 99.4% of the HCl  collected
In  the first impinger.  The slightly high
recoveries of HCl and Cl2 resulted  from
small  fluctuations in the rotameter float
positions which permitted larger actual
flow rates of these gas species.
     In summary, it has been shown that
it is possible to differentiate between
HCl and C12 species in a laboratory-
generated air stream.  The next phase of
this project should be carried out using
a combustion gas stream with a gas-
particle separator to determine their
effects on the separation and analysis.
For effective speciation using this
absorption train, it is important that
the water impingers should be acidified
to pH 1 prior to sampling.  The major
components of most process streams are
S02 and NOX and these gases, as
demonstrated in this work, do not
interfere with the ion chromatographic
determinations.
                                           40

-------
               TABLE 10.  HC1/C12 SPECIATION EFFICIENCY IN S02/NOX STREAM*
                                                  yg cr
Impinger
#1
(H20, pH~l)
Impinger Impinger
#2 #3
(H20, pH~l) (NaOH)
Impinger
#4
(NaOH)
Found
Theoretical
% Collected
28,100 160
23,000
123
735
826
110
173


     *HC1 and 019 stream concentrations were 530 and 19 ppm, respectively.
      SO-) and NOX stream concentrations were 250 and 600 ppm, respectively.
DISCLAIMER

     The research described in this
article has been funded wholly by the
U.S. Environmental Protection Agency
through Contract No. 68-02-3129 to GCA
Corporation, Technology Division, and It
has been subjected to  the Agency's
required peer and policy review.
However, it does not necessarily reflect
the views of the Agency and no official
endorsement should be  inferred.

REFERENCES

1.   Bonner, T. A., et al., 1981,
     Engineering Handbook for Hazardous
     Waste  Incineration, SW-889, U.S.
     EPA, Cincinnati,  Ohio.
2.    Cheney,  J.  L.  and C.  R.  Fortune,
     1979.   Evaluation of  a method for
     measuring hydrochloric acid in
     combustion source emissions.   The
     Sci.  of  Total  Environ.,  13:9-16.

3.    Holm,  R.  D.,  and S.  A. Barksdale.
     1978,  Analysis of Anions in
     Combustion Products.   In:   Ion
     Chromatographic Analysis of
     Environmental  Pollutants.
     E.  Sawicki,  J. D. Mulik, and
     E.  Wittgenstein (Eds.).   Ann Arbor
     Science  Publishers,  Ann Arbor,
     Michigan, pp.  99-110.

4.    Ruch,  E.  1970, Quantitative Analysis
     of  Gaseous Pollutants; Ann-Arbor-
     Humphrey Science Publishers,  Inc.,
     Ann Arbor,  Michigan,  pp. 65-66.
                                            41

-------
                           DIOXIN  COLLECTION FROM HOT STACK GAS
                             USING SOURCE  ASSESSMENT SAMPLING
                              SYSTEM AND MODIFIED METHOD 5
                                  TRAINS—AN EVALUATION
                       Marcus Cooke, Fred DeRoos, and Bruce Rising
                             Battelle's  Columbus  Laboratories
                                  Columbus, Ohio  43201

            Merrill D. Jackson, Larry D. Johnson, and Raymond G. Merrill, Jr.
                       Industrial  Environmental Research  Laboratory
                          U. S. Environmental Protection Agency
                      Research Triangle Park, North Carolina  27711
                                         ABSTRACT
     Dynamic gas-phase spiking  was  used to demonstrate  the  collection efficiency of  two
EPA  source  sampling  systems  for  tetrachlorod,ibenzo-p-dioxins.   The  Source Assessment
Sampling  System and  Modified  MethoQ  5  trains  were  used  to  collect  a  representative
sample administered  in  trace quantities into  the  hot  exhaust gas from a flue gas source
capable  of  simulating  incinerator  stack  conditions.    Nine experiments  were performed
with .the  two sampling systems to measure the overall method  recoveries  for varied levels
of  1,2,3,4-tetrachlorodibenzo-p-dioxin.   High  resolution  glass  capillary  column  gas
chromatography/high  resolution  mass  spectrometry  techniques were  us'ed to  analyze  the
collected samples.   Recoveries  were  considered quantitative  for all spiking  experiments,
except one,  demonstrating the  usefulness  of the  SASS  and MM5  trains  in  collecting  and
analyzing low levels of dioxins  in hot, gaseous combustion emissions.
INTRODUCTION

     It  is critical,  in  measuring  trace
quantities    of   organic    species   in
combustion  gas  streams,   to   use   well
documented  procedures   that   have   been
evaluated  for  specific  applicability  to
the source studied.   Since the analytical
process    must    begin   with   sampling,
defensibility   of  any  combustion   stream
study    requires    appropriate   quality
control  checks  at  every  stage  in  sample
taking and sample handling.

     In  this  study  two  sampling  systems
were  chosen  for  evaluation  as  research
tools  to measure  combustion flue gas,  the
Source  Assessment Sampling  System  (SASS)
train  and  the   Modified  Method   5   (MM5)
train.    These   two  EPA  sampling  systems
have  been used  in  sampling  many  source
types  (1-2).
     In   this   study   the   two  sampling
systems,  SASS  and  MM5,  were  challenged
with     low     levels     of     1,2,3,4-
tetrachlorodibenzo-p-dioxin      (1,2,3,4-
TCDD)  in  the gaseous  discharge  of a gas-
fired  research combustor.    Generating a
continuous  flue gas  stream which contains
known  levels  of   a   test   compound   is  a
challenging  task.    The  flue  gas  tested
must  be  taken  from  an authentic  source
discharge,  and   the   spiking  system  for
administering  the  compound  of  interest
must  ensure  complete  mixing  in  the  gas
stream  and must,  in  its  design,  prevent
deposition  or decomposition.

     This    paper   describes   a    study
designed  to spike hot flue  gas  with  low
levels   of   1,2,3,4-TCDD,    collect    the
spiked  flue  stream  using   two   standard
sampling    systems,    then   measure    the
overall recovery  by  a  high resolution  gas
                                           42

-------
chromatography/high  resolution mass spec-
troraetry (HRGC/HRMS)  technique.   The goal
of  this  study was  to  determine  the effi-
ciency  of  collecting  and  retaining   low
levels   of    dioxins   from   combustion
discharges.    Richard  and  Junk   (3)   de-
scribed  a  dynamic  injection system   for
spiking  an incineration  flue  stream with
a   polychlorinated   biphenyl   standard.
Recently Petersen  et  al.  (4) described  an
experiment  to spike  a hot  diesel engine
exhaust  stream  with  perdeuterated poly-
nuclear  aromatic  hydrocarbon  (PAH) stan-
dards  to  determine  if  a  significant loss
of   PAH  occurred   during  sampling   and
analysis.

      In  this  study low levels of  1,2,3,4-
TCDD  were  continuously  administered into
the  hot  flue gas  stream  at  a  range   of
concentrations   which   simulated   the   low
level  discharges  reported  in  the  litera-
ture  on  incineration.    Overall   recovery
of  1,2,3,4-TCDD  was  used  to  assess   the
efficiency of the  two EPA source  sampling
systems  to collect, retain, and stabilize
ultra-trace   TCDD   levels   for  subsequent
analysis.
TEST  STRATEGY

      A  pilot  plant  furnace  was  used  to
generate  the  simulated  incinerator  flue
gas.   A  low  flow  rate pump  was  used  to
inject  a  TCDD   sample   solution   into   a
slipstream  of the  flue  gas  where  the gas
temperature    was   approximately    260°C
(500°F).   Flue gases were then  sampled  by
the   stack   sampling   equipment.     The
sampling  temperatures,  rates, and  volumes
were  typical  of  those used  in a  MM5  or
SASS  apparatus.    Since TCDDs  are  highly
toxic  compounds,   it  was  imperative  that
none  of   the  sample  escaped  into  the
environment.   To  prevent  such  emissions,
all  gases  spiked with  1,2,3,4-TCDD  were
passed    through    the   sampling    train
followed  by  a  charcoal filter  trap  used
as  a  secondary filter.   A schematic  dia-
gram  (Figure  1)   shows  the   experimental
setup:  natural  gas combustor outlet; hot
gas   spike   injection;   sampling   train;
charcoal  safety  filter;   and  gas   moving
system.
COMBUSTION AND SAMPLING  SYSTEMS

     In  this  program,  two sampling  trains
(SASS  and  MM5)  were   used.    The   SASS
equipment was operated  in accordance with
procedures  outlined   in  the  EPA  Level  1
manual  (5).   Because a  relatively  clean
fuel  (natural  gas)  was  used  to  produce
the   flue   gas    stream   sampled,   some
modifications were made  to  the   standard
SASS   apparatus.      The   normal   stack
sampling probe and cyclones  were replaced
with    borosilicate     glass    connecting
tubing.     Since   the  flue   stream  was
essentially    particulate    free,    the
stainless  steel  cyclones  were  removed.
The  solutions  that  would   normally  be
contained  in  impingers   1   and   2   were
eliminated,  although  the impingers  were
left  in the system  as condensate traps.
Impinger   3   contained   a   charge   of
activated  charcoal to  serve  as  a safety
filter  in  the   event  that  1,2,3,4-TCDD
escaped  collection   in   the   resin   bed.
Impinger  4  contained a  charge of silica
gel to  protect the sampling  pumps  and dry
gas  meter.    The  modified SASS  sampling
apparatus is illustrated in Figure 2.

     The  major   difference   between  the
SASS  and  MM5 trains  is  in  the  sampling
rate  and  volume  of  gas collected.   This
gas  volume  allows  the  use  of a  smaller
quantity  of  XAD-2 in the  MM5  train.   The
filter  is  typically  held  at  a  higher
temperature  in  the  SASS  train.   The  MM5
train  and location  of gas-phase  spiking
are shown in Figure 3.

     A  development  of  this  study  was  a
gas-phase   spiking   system   that   could
inject   liquid   solutions  of   the   test
material  into the  hot  flue gas  stream.   A
fused  silica  capillary line  was  used  to
transfer   the   dioxin  solution   from  a
precision metering pump  to  the  flue  gas
stream.    Condensation  was   prevented  by
jacketing  the  capillary   at  the  point
where heating occurred.   By  circulating a
refrigerant  around the  transfer  line  up
to  the  point  where  the  capillary  dis-
charged  into  the   hot  flue  gas, premature
vaporization  and   condensation  were  pre-
vented.   The primary solvent  was acetone
with a  keeper,  0.5 percent  (v/v)  decane,
added  in  order  to wet the  final segment
of  transfer  line  and  prevent  deposition
of  the  small  amount  of   1,2,3,4-TCDD in
the capillary tip.

     As  shown  in  Figure  4,  a  slipstream
was  drawn  from   the  primary  combustion
discharge  into   the  sampling  apparatus.
The   transfer   line  in   which   solution
spiking   occurred   was   insulated   and
                                            43

-------
Furnace
 Outlet
                  Spike Injection,
                     System
                                                                                        To
                                                                                     Atmosphere
Charcoal
  Filter
                 Figure  1.  TCDD  Injection and  Sampling Apparatus.

-------
         From
       TCDD Spike
     Injection Syitem
                 Hoi
                 Filter
Centralized Temperature
 and Pressure Readout
Figure  2.   Source Assessment Sampling  System  (SASS).
                                45

-------
                        From
                     TCDD Spikl
                    Injection System
          Coaling
          Impinger
 Drying       Gu
Impinger       Meur
Figure  3.   Modified Method 5 Sampling System  (MM5).
      Figure 4.  TCDD  Spike Injection System.
                          46

-------
externally    heated   to   maintain    the
required  gas  stream temperatures from  the
point  of  spiking  into  the samplers  (SASS
and  MM5).     A  summary  of   operational
parammeters   for   the   flue   gas  sampling
systems  is  shown  in Table  1.   For both
the  SASS  and  MM5   tests   the   flue   gas
temperature  at  the  injection  point   was
260°C  (500°F).   The filter  temperature  in
the  SASS  was   held  at  204°C  (400°F),
whereas the MM5 was held at  120°C  (250°F)
filter  temperature.  The  operating upper
temperature limit  for the XAD-2  traps  was
set  at 20°C  (68°F).   Water  at  ice  tem-
perature  was   circulated in  the jacketed
condensers  shown  for  both  the  SASS   and
MM5  trains.   As  illustrated in Figure  2,
a  double  condenser  section  was  necessary
to achieve 20°C in  the  SASS  train.

     Level 1  guidelines  set  the  total  dry
gas  volume  sampled  by  the  SASS at  30  m3.
Relatively   large   sample   volumes  were
taken  for MM5  tests to ensure  collection
times  similar  to   actual  field  operation
of   the   sampling   system  in   incinerator
applications.     A   total   of  10    tests
(numbered  1  through  10)  were  conducted
during  this  program.   The  experimental
outline  is  shown  in Table  2.   The  first
series  consisted  solely of SASS  equipment
and  Level 1  operational procedures, while
the  second  series  employed MM5  equipment.
The  first test  (Test  1) was  a blank  run
to   identify    native   concentrations   of
1,2,3,4-TCDD   and   2 , 3 , 7 ,8-TCDD,  produced
either  by  the  furnace,   or  by  sample
handling.   Test   2  was  a  blank  test   in
which  a  known  quantity  of  1,2,3,4-TCDD
was  injected  into  the  sample train;  how-
ever,  a dry heated air  stream was used  as
the  sampled  gas.    Tests  3-6 employed  the
SASS sampling  train with the flue gas  and
variable  spike  levels  of   1,2,3,4-TCDD.
The  MM5  tests  (7-10) were  completed  in a
similar  manner using  hot  combustion   gas
starting  with  the  highest   1,2,3,4-TCDD
concentration   and  progressing   to    the
lowest  concentration.   Two  sets  of tests
(3 and  4,  9  and 10) were  run at the same
concentrations  to  determine  experimental
reproducibility.

     The  TCDD   concentrations  shown   in
Table  2 were  the  targeted concentrations.
Actual  concentrations  varied somewhat  for
each   test.     Prior  to  beginning  each
experiment  all  glassware  was  thoroughly
cleaned with  methylene  chloride and baked
in  a  clean   oven  at  400°C   (752°F)   for
12 hrs.   The  samnling  equipment  was  then
assembled  in  the  combustion  laboratory
and  attached  to  the   tube  furnace.   The
sampling  systems  were  leak-checked prior
to commencing  an  experiment.   Leak rates
for  the  SASS  train  were typically 8.5  x
10~4-  m3/min  at 635  torr.   To  begin  the
experiment  the gas   sampling   pumps  were
started   and    temperatures   allowed   to
equilibrate.    After  sampling  rates  were
set  and   operating  temperatures  reached,
the  precision metering  pump  was turned on
and  set at an  injection  rate of 0.17-0.18
mL/min.    Table  3  gives  a   summary  of
actual system  conditions  during the spike
experiments .
CHEMICAL ANALYSES

     The solvents  used  were Distilled-in-
Glass   grade   hexane,   benzene,   carbon
tetrachloride,    and  methylene   chloride
(Burdick and Jackson Laboratories,  Inc.,
Muskegon,  Ml).     The   adsorbents   used
(solvent-rinsed  and  activated  immediately
prior to use)  were alumina and silica gel
(Biorad Laboratories, Richmond,  CA).   The
standards used  were tetrachlorodibenzo-p-
dioxin-13C12   (2 ,3,7,8-TCDD"13C12)   (KOR
Isotopes,  Cambridge,  MA),  and  1,2,3,4-
tetrachlorodibenzo-p-dioxin      (1,2,3,4-
TCDD) (Ultra Scientific,  Inc.,  Hope,  RI).
All  analytical  glassware  was  washed  with
soap and water,  rinsed  with reagent grade
acetone, and baked  at 400°C  for  a  minimum
of  12   hours  prior  to  use.   The  XAD-2
resin,   precleaned   grade   (Supelco,  Inc.,
Bellefonte,   PA)  was extracted  for  24 hrs
with methylene  chloride  and  dried  prior
to   being    packed  into    the   sampling
modules.    Test  batches  of  resin  were
extracted  and  blanks   checked  by  glass
capillary GC-FID prior  to use.

Sample Extraction and Cleanup

     XAD-2   resin   samples  were  Soxhlet
extracted   for   16  hrs   with  methylene
chloride.   The  125  g  resin samples  from
the  SASS  experiments were spiked  with  2
ng  of  internal  standard  (2 , 3,7,8-TCDD-
  G12-)   Pri°r to extraction, while  the 22
g  resin  samples from the  MM5  experiments
were spiked  with 0.8 ng  of 2, 3, 7,8-TCDD-
  C\2°   The  extraction  solvent volume was
1  L  for the SASS  samples and  200  mL for
the MM5 samples.

     Methylene   chloride   extracts   were
concentrated to  approximately  10 mL using
a   Kuderna-Danish    concentrator.      The
                                            47

-------
           TABLE 1.   OPERATING PARAMETERS FOR THE SAMPLING SYSTEMS
           Fuel
      Natural Gas
Sampling Volumes
   SASS
   Modified Method 5
30 m3 (1060 ft3)
6.0 m3 (200 ft3)
Flue Gas Sampling Rate
   SASS
   Modified Method 5
0.11 nu/min (4 scfm)
0.02 m /min (0.8 scfm)
Spike Solution Injection Rate
Approximately 0.2 mL/min
Filter Temperatures
   SASS
   Modified Method 5
204°C (400°F)
120°C (250°F)
XAD-2 Resin Bed Temperatures
   SASS
   Modified Method 5
<20°C (68°F)
<20°C (68°F)
                                    48

-------
                             TABLE 2.   EXPERIMENTAL OUTLINE
Test
Number
1
2
3
4
5
6
7
8
9
10
(a) Flue gas
Test Number
Gas
Flue Gas
Hot Air
Flue Gas
Flue Gas
Flue Gas
Flue Gas
Flue Gas
Flue Gas
Flue Gas
Flue Gas
TCDD(a)
Concentration1
0
1000
1000
1000
100
10
1000
100
10
10
concentration (pg/m ) at 20°C.
TABLE 3. SYSTEM OPERATING
1
2345
Sampling
Apparatus
SASS
SASS
SASS
SASS
SASS
SASS
MM5
MM5
MM5
MM5
CONDITIONS
6 7
Proposed
Sample
Blank
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD
1,2,3,4-TCDD

89 10
Gas Temperature,
                    260    233    263    254    257    251    258    258    256     262
Filter
  Temperature, °C   203    212    201    193    187    208    130    128    129     129

XAD-2(b)
  Temperature, °C   5.00   5.00   6.10   6.67   8.33   7.22   6.67   6.28   5.44   6.11

02,  percent         5.0    5.0    5.1    4.9    4.7    4.9    4.7    5.0    5.1     4.9

Sampling Rate,
  m3/min            0.113  0.112  0.112  0.114  0.116  0.117  0.021  0.021  0.021  0.021
          (c)
  Sampled, irr
Gas Volume
28.7   30.2   30.6   30.2   30.7   30.7    5.50    5.54    5.69    5.53
Solvent Injection
  Rate, mL/min      0.18   0.17   0.17   0.17   0.17   0.18   0.17   0.17   0.18   0.17

(a)   Gas temperatures measured at the point of injection.
(b)   Temperatures measured at XAD-2 outlet.
(c)   Dry volumes at 20°C.

                                           49

-------
extracts were  further  concentrated  after
transfer to  tubes  (with  drawn  out  tips)
using three  2  mL hexane  rinses  and  care-
fully   solvent-exchanged   into   5   mL  of
hexane.

     All extracts  were  cleaned  up  using
column  chromatography  with  a  combination
of  silica,  modified silica,  and alumina.
The  extracts  were  first  passed  through a
multilayered   silica   column   containing
silica,  44%  sulfuric acid on  silica,   and
33%  1  molar  sodium hydroxide on  silica.
The  acid, neutral,  and  basic silicas were
placed  in  layers.   The  column was pre-
rinsed  with   hexane,   and   extract   was
added,  followed  by  column  elution with
1:1  hexane/benzene.  The  total  eluent  was
collected,    concentrated,   and   solvent-
exchanged  into  5  mL  of  hexane.     The
extract  was  then  added to  a  column con-
taining  5  g  of precleaned  and  activated
alumina.    This  column  was  sequentially
eluted  with  hexane,  hexane/carbon  tetra-
chloride   (1:1),    and   hexane/methylene
chloride   (1:1).     The  hexane/methylene
chloride   fraction   was    collected   and
gently  taken   nearly   to  dryness.     The
extract  was  dissolved  in  20   mL   of  n-
decane  and  stored at  0°C until  analyzed.
The  sequential  cleanup procedure is shown
in Figure 5.

HRGC/HRMS Procedure

     The  extracts  were  analyzed  by com-
bined  HRGC/HRMS.   A VG  Model  7070H HRMS
directly  coupled  to  a Carlo Erba  Model
4160 GC was used for  the analyses.   Data
were acquired  by  a  VG  Model  2035  fore-
ground/background  data  system.    The  MS
was  operated  in the  electron   ionization
mode at  a   resolution  of   10,000-12,000
(M/AM,  10%  valley).

     Five  ion  masses were monitored using
the  VG  Digital MID  unit  which  included a
prototype  20-bit   digital-to-analog con-
verter  for  accelerating voltage  and elec-
trostatic  analyzer  voltage  control.   The
monitored  masses   were:     319.8965   and
321.8936,  the  most  intense peaks   in  the
molecular  ion  cluster  of  native   TCDDs;
along   with  331.9368   and   333.9338,   the
most intense peaks  in the  molecular  ion
cluster  of  the  internal standard.    A peak
for  perfluorokerosene,  318.9792, was used
as  a lock  mass  by the  data  system to con-
trol mass focus.
                                 Scmplc/Hexine
     Figure 5.  Schematic Diagram of  the
                Sequential TCDD  Cleanup
                Procedure.

     The   chromatographic   column  was   a
30 m, DB-5  coated,  fused silica  capillary
interfaced  directly into  the ion  source.
Helium  carrier  gas was used  at  a  flow
velocity  of  21 cm/sec.   Sample  injection
was made  at  150°C with a 3-min   hold time
followed  by  a linear  temperature increase
of  30°C/min  to  280°C.    The  final  tem-
perature   was  maintained   for   12   min.
Under  these  conditions  2,3,7,8-TCDD-13Ci2
and  1,2,3,4-TCDD  had  retention   times  of
approximately  13  min each.   Recent papers
by Harless et al.  (6-7) demonstrated  the
sensitivity  and  selectivity  obtainable by
the  combination   of  HRGC   and   HRMS   to
accurately  measure  low  levels  of  TCDD
isomers.

Calibration Procedures

     Calibration  was  based  on  comparing
responses  from known  amounts of 1,2,3 4-
TCDD  (the  analyte  of  interest)  with   a
                                             50

-------
known  amount  of  2,3,7,8-TCDD-13Ci2  (the
internal  standard).

     The   1,2,3,4-TCDD   standard   was   the
same  as  that  injected  in  the   hot   gas
stream.   All  solutions  of this  material
were  based  on  the  same  gravimetrically
prepared  stock  solution.    The   internal
standard   material,    2,3,7,8-TCDD-13Ci2»
was  obtained  in  solution.   However,   its
concentration   was   checked  with   gravi-
metrically  prepared  standards  of  native
2,3,7,8-TCDD  from four  sources:   EPA  (R.
Harless,  EPA/RTP) ,  University of  Nebraska
(Lincoln),  McMaster  University   (Toronto,
Canada) ,  and  Monsanto  Research   Corpora-
tion (Dayton,  OH).

     The    instrumental   response  ratio
between   the  analyte  material   and   the
internal    standard   was   obtained    by
analysis  of  a  portion  of  the  flue   gas
injection solution  that was  spiked  with
the  internal  standard  solution.   Since
the  concentration of analyte was  based on
comparing  its  response   to  a   labeled
isomer,  calibration was  carried   out  with
every  analysis run.

Data Reduction,  Validation,  and Reporting

     Both stack  gas spikes  and  internal
standards were  used  in  this  study,   and
recoveries  are reported  for  each.  It is
important to  note  the  difference  between
them  and how  they  were  used.   The spike
was  added   to the  stack  gas  slipstream
before  the  sampling train.    This value
was  used  to  calculate   the  collection
efficiency  of  the  sampling  system.    The
internal  standard was  added to  the resin
after  sampling  was  completed.     Its  pur-
pose  was  to   determine   the  extraction,
cleanup,  and  analytical  recovery.    The
collection  efficiency  was  corrected   for
the recovery of  the  internal  standard.

     The  amount of recovered 1,2,3,4-TCDD
was calculated by  the relationship:
ng 1,2,3,4-TCDD =
        Area m/e 322 (1,2,3,4-TCDD)
     Area m/e 334 (2,3,7,8-TCDD-13c12)
     	2.5  ng*	
     Fraction of Sample Extracted
* = ng of 1,2,3,4-TCDD added.
     Data  validity  was  ensured  through
application of the following procedures:

     •  Daily   tuning   the   mass   spec-
        trometer  for  resolution  and  area
        response.

     •  In every  run, the  internal  stan-
        dard must  be  from 1/2  to  2  times
        the    expected    response    (the
        response  observed in  a  test  mix
        of known concentration).
RESULTS AND DISCUSSION

     The   analytical   determinations   in
Test 1  (system  blank  using  flue  gas)  and
Test  2  (system  spike  using heated  air)
were completed before  the  actual  SASS/MM5
challenge  tests  were  performed.   In  Test
1,   the   system   blank,   no   detectable
1,2,3,4-TCDD or  2,3,7,8-TCDD was  found at
the  minimum  detectable   amount   (1   pg)
injected.   For  Test  2,  spiked  hot  air,
subsamples  were  analyzed   representing
several   regions  of   the  SASS   train,
regions  where  TCDDs   might  condense  or
accumulate.      The   results   of   those
analyses   are  shown  in   Table  4  where
recoveries  are  shown  to total  about  94%.
An  instrument  fault  occurred in  the  com-
puter  system while  computing  the  XAD-2
extract  data,  and  the  quantification of
the  reported  70%  resin   trap  recovery
could  not  be   verified   by  exact  area
computation.   This value  is  approximate
but  represented  the  major  portion  of  the
1,2,3,4-TCDD  and showed  a  high  recovery
overall  for the system spike.
CHOICE OF SPIKE LEVELS FOR TESTS 3-10

     A   series    of    technical   review
meetings  included  consideration  of  pro-
posed  spike  levels in relation  to source
discharge  levels  and  analytical  sensi-
tivity.   These  data  have  received  much
attention,     and  to   date  no  definitive
study  of incinerator  discharge levels  for
TCDDs  has  been  performed.   The proposed
analytical  system is   capable  of subpico-
gram   sensitivity  for   individual  TCDD
isomers, and spiking  levels  of 1000, 100,
and  10  pg/m-*  were  proposed.   Based  on
typical  collection volumes  for  the SASS
and  MM5   trains,   the  total  weight  of
1,2,3,4-TCDD spiked  ranged  from 60 pg to
30  ng  as  shown  in  Table  5.    A  syringe
volume   of   50  mL  was   used  with   the
                                            51

-------
   TABLE  4.   EPA INCINERATION SIMULATION RECOVERIES—TEST 2
Amount of
1,2,3,4-TCDD
Recovered (pg)
50
ND(a)
260
4,800
17,500
Total Recovery
Percent
Recovery SASS Test
(%) Section
0.2 Injector
liner
Heated flue
1.1 Filter
23 Transfer
to XAD-2

line
70('b-) XAD-2 resin
,94

(a)   Not  detected.
(b)   Approximate.
         TABLE 5.   TCDD  LEVELS  ADDED TO HOT STACK GAS
Spike Level
High
Medium
Low
TCDD Concentration
Cpg/m3)
1000
10.0
10
Total TCDD
SASS (30 m3)
30 ng
3 ng
300 pg
Total TCDD
MM5 (6 m3)
6 ng
600 pg
60 pg
                              52

-------
precision  metering  pump  to  deliver   the
dioxin  spike   solution;  and  since   gas
collection  rates  of  6.8  m3/hr   for   the
SASS  train  and  1.3  m3/hr  for   the   MM5
train  were  used  in  sample  collection,
1,2,3,4-TCDD   spike   solutions  of  1-600
ng/L,     liquid    concentration,    were
required.      This   corresponds    to   an
approximate  delivery  rate  of  0.17-0.18
mL/min.    This  rate  was  high enough  to
flush   the   transfer  probe   and   prevent
condensation in  the  capillary.

     The  choice  of an optimum solvent  was
based  on  delivery rate,  viscosity, solu-
bility  of  1,2,3,4-TCDD,   boiling  point,
XAD-2  retention,  and demonstrated  purity.
The  solvent system  developed  was  acetone
with  0.5%  (v/v)  decane  as  a   "keeper."
Decane  also   served  to   flush   residual
1,2,3,4-TCDD   from  the    transfer  probe.
The  acetone acted  as a volatile  carrier
for  the  1,2,3,4-TCDD and was  not  retained
in the XAD-2 resin bed.
RECOVERY STUDIES—TESTS  3-10.

     Tests  3-10  were  designed  to  chal-
lenge  both  the SASS and MM5 sampling  sys-
tems  at gas-phase  concentrations  expected
to  cover  a  range from  high levels to  the
minimum levels   detectable by   available
analytical  methodology.   Results of  this
series   of   experiments   are   given   in
Table  6 which includes  the dry air  spike
experiment  (Test  2).

     These  data  show good  internal  con-
sistency  and  high  recoveries   for   each
experiment,  except  Test  10,  where a  229%
recovery  was  observed.    No experimental
evidence  exists  to explain  this  unusual
result.   On  balance  the  data  show  that
both  the  SASS and  MM5  trains  work  well
with  the  isotope dilution HRGC/HRMS  ana-
lytical technique   to   recover   TCDDs  as
modeled by  recovery  of 1,2,3,4-TCDD  from
hot flue gas streams.

     It is  noted  that the  distribution  of
1,2,3,4-TCDD  among  the  heated  and cooled
sections  of the  SASS and  MM5  trains may
vary  as  a  function   of   the  combustion
source  tested.   Dioxins may have a  high
adsorptive  affinity  for  the  particulate
matter  found  in  some  sources.     It  is
interesting  to note  in  Test  2,  the dry
air  experiment,   that  a major  portion  of
the spiked  1,2,3,4-TCDD was  found in the
cold  regions  of  transfer  tubing  leading
into  the  resin  trap  (23%   in   transfer
line).    To  investigate  the  effect  of
moisture condensation  on  the operation of
the  sampling  systems  two  separate  inves-
tigations  were  performed:   for  the SASS
samples  (Tests  3-6),   the   heated  "front
half"  of the  train (probe  rinse,  filter
housing, filter,  and heated  transfer line
wash)  was   analyzed  separately   from  the
cooled  "back  half"  of the  train (cooled
transfer   line  washes,   resin  bed,  and
condensate).    In  all four  experiments,
less  than  1%  of  the  spiked 1,2,3,4-TCDD
was  found  in the  heated  "front  half"  of
the  SASS  train.    A  second  series  of
experiments  were  performed  to  determine
the  possibility  of  TCDD breakthrough from
the  resin   trap.    For  this  purpose  the
condensates  from   Tests   3-6,   the  SASS
challenge  tests,  were  analyzed separately
to   determine   if   any  of   the   spiked
1,2,3,4-TCDD   was   carried   through  the
resin  bed  with condensed  water.     After
extraction  and  concentration, no measur-
able   levels  of   1,2,3,4-TCDD  could  be
detected in the condensates  from  Tests  3-
6,  thus demonstrating the  effectiveness
of XAD-2 to  trap TCDDs even in  moisture-
laden combustion gas streams.
SUMMARY

     This   study   demonstrated   a   high
dioxin  collection  efficiency  using  two
stack  sampling  techniques,  the  SASS  and
MM5 trains.   In  eight  combustion tests,  a
system  blank,  and  a  spiking  experiment
with hot  dry  air,  recoveries  of  1,2,3,4-
TCDD  were  quantitative   except  in  one
instance.   The  spiking  device developed
in   this   study   incorporated   several
features  that  make  it useful  for spiking
a broad  range of test  compounds directly
into  hot  flue gas  streams  in order  to
generate  a  stream of  known concentration.
Such  a  spiking  device   is  necessary  to
evaluate   the   performance  of  candidate
sampling   and  analysis   procedures   for
stationary source testing.
ACKNOWLEDGEMENTS

     The  authors  would  like  to  acknowl-
edge several  individuals  who participated
in   the   research  program:     Harry  G.
Leonard,  Thomas  C.  Lyons, Jr.,  and James
J.  McNeely,  who  operated the  combustion
system; Daniel  G.  Aichele and  Michael E.
Larson, for  chemical analyses;  and Joann
                                            53

-------
TABLE 6.  1,2,3,4-TCDD RECOVERIES FROM TESTS 2-10
Test Number
2 (Dry Air)
3
4
5
6
7
8
9
10
Sampler
System
SASS
SASS
SASS
SASS
SASS
MM5
MM5
MM5
MM5
Liquid Spike
Volume (mL)
49
51
42
50
51
47
47
49
47
1,2,3,4-TCDD
Liquid
Concentration
(pg/mL)
500
500
50
5
5
15
12
8
8
Expected
Concentration
(ng)
24
25
21
2.5
0.25
7.1
0.56
0.39
0.38
Calculated
Concentration
(ng)
23
18
17
2.9
0.28
8.0
0.57
0.47
0.87
Percent
Recovery
=94
73
83
117
113
113
101
120
229

-------
Pelino  and  Maria B. Dean  for  secretarial    4.
assistance.   This  study was  supported  by
the  United  States  Environmental  Protec-
tion  Agency  under  Contract  68-02-2686,
Task Directive 131.
DISCLAIMER

     Mention  of  trade  names  or  products
does not  constitute  endorsement or recom-
mendation  for  use  by  the  United  States
Environmental Protection Agency.
REFERENCES

1.  Estes,  E.  D.,   F.  Smith,  and D.  E.
    Wagoner.      Level    1   Environmental
    Assessment    Performance   Evaluation.
    EPA-600/7-79-032   (NTIS   PB   292931),
    U.  S.     Environmental     Protection    6.
    Agency,  Research  Triangle Park,  NC,
    1979.

2.  Cooke,  W.  M., J. M.  Allen,  and R. E.
    Hall.    Characterization  of Emissions
    from   Residential    Wood   Combustion
    Sources.     In:     Residential  Solid    7.
    Fuels,   J.  A.  Cooper  and D. Malek,
    eds.   Oregon Graduate Center,  Beaver-
    ton,  OR,  1981.   pp.  139-162.

3.  Richard,   J.   J.,  and  G.  A.  Junk.
    Polychlorinated   Biphenyls  in Efflu-
    ents from  Combustion of  Coal  Refuse.
    Environmental   Sci.   Tech.,   15(9):
     1095-1100,  1981.
Petersen, B.  A.,  C. C.  Chuang,  T.  L.
Hayes,  and  D. A.  Trayser.   Analysis
of PAH  in Diesel  Exhaust  Particulate
by  High  Resolution  Capillary  Column
Gas  Chromatography/Mass  Spectrometry.
In:   Proceedings  of the Sixth  Inter-
national   Symposium   on   Polynuclear
Aromatic  Hydrocarbons,  M.    Cooke,  A.
J.  Dennis,  and  G. L.   Fisher,  eds.
Battelle  Press,  Columbus,  OH,   1982.
pp. 641-653.

Lentzen,  D. E.,  D. E. Wagoner,  E.  D.
Estes,  W. F.  Gutknecht,  and  L.   D.
Johnson.   IERL-RTP Procedures Manual:
Level   1   Environmental   Assessment
(Second  Edition).     EPA-600/7-78-201
(NTIS PB  293795),  U.  S.  Environmental
Protection Agency, Research  Triangle
Park, NC, 1978.  259 pp.

Harless,  R.  L.,  E. 0.  Oswald,   R.  G.
Lewis,  A.  E.   Dupuy,  Jr.,  D.   D.
McDaniel, and H.   Tai.   Determination
of  2,3,7,8-Tetrachlorodibenzo-p-diox-
in in Fresh Water  Fish.   Chemosphere,
11(2): 193-198, 1982.

Harless,  R.  L.,  E. 0.  Oswald,   M.  K.
Wilkinson, A.  E.  Dupuy,  Jr.,  D.   D.
McDaniel, and  H.  Tai.    Sample  Prepa-
ration  and   Gas   Chromatography-Mass
Spectrometry     Determination      of
2,3,7,8-tetrachlorodibenzo-p-dioxin.
Anal. Chem.,  52:1239-1245,  1980.
                                            55

-------
                        STACK SAMPLING AND ANALYSIS OF FORMALDEHYDE
                                     Kevin J. Beltis
                                    Anthony J.  DeMarco
                                    Virginia A. Grady
                                     Judith C.  Harris
                                  Arthur D. Little, Inc.
                              Cambridge, Massachusetts 02140
ABSTRACT

A collection medium of potential use for the determination of formaldehyde in stationary
combustion research and development projects was evaluated in a laboratory test atmosphere.
The medium consists of a porous polymer sorbent, i.e., XAD-2®, coated with 5% by weight of
2,4-dinitrophenylhydrazine-hydrochloride (2,4 -DNPH-HC1).  Formaldehyde is retained on this
medium by adsorption and formation of the 2,4-dinitrophenylhydrazone derivative.  Analysis
was conducted by reverse phase HPLC with an acetonitrile/water eluant.

The laboratory evaluation consisted of:
     • Sampling/Analysis of a formaldehyde test atmosphere
     • Sampling/Analysis of a simulated stack matrix
     • Analysis of spiked samples
     • Evaluation of storage stability
     • Independent analysis.

The precision (relative standard deviation) associated with the sampling/analysis of the
formaldehyde atmosphere with the test medium was observed to be 7.1%.  The measured
formaldehyde concentration was within ± 1% of the "true" concentration as determined by
the chromotropic acid method.

Samples collected from the stack matrix were spiked with formaldehyde.  Observed recovery
was 107%.  Samples stored for seven days showed an increased apparent formaldehyde content
of about 17%.
INTRODUCTION

The possible presence of designated high-
probability carcinogens (HPCs) in the
effluent of a stationary combustion source
is dependent on the nature of the fuel or
other feedstock and on the combustion
efficiency.  Complete combustion of mate-
rials in these sources requires sufficient
air (oxygen), turbulent mixing, and a high
temperature maintained over n long period
of time.  Non-optimum combustion conditions,
such as inadequate amounts of air within the
combustion zone or within localized subzones,
will result in unburned or only partially
burned organic material.  The vaporous
material can include acids, aldehydes, and
other more complex organic species.  These
species are of particular concern during
the evaluation of combustion modification
techniques since care must be taken that
such modifications do not lead to addition-
al environmental problems.

In a literature survey for the EPA(l) which
focused on sampling/analysis (S/A) of the
21 designated HPCs, formaldehyde was among
a few HPCs selected a a "special needs"
group.  These "special needs" compounds
were categorized as requiring immediate
sampling and analysis method attention
before being selected as candidates for
                                            56

-------
combustion-source studies.  The EPA survey
report concluded that the previously recom-
mended S/A procedure for formaldehyde (i.e.,
using 3-methyl-2-benzothiazolone hydrazone
[MBTH]) should be replaced since sulfur dio-
xide (802), which is found in stationary
combustion source effluent,  negatively in-
terferes with the aldehydes measurements in
that method (2).  The report suggested an
alternative approach involving the formation
of the 2,4-dinitrophenylhydrazone derivative
of the aldehyde.  Two methods based on this
recommendation are currently in use, but
have not been evaluated for use in stack gas
analysis.  The first, a liquid impinger
method, has been routinely used in the past
for diesel exhaust analysis of carbonyls
(2,3).  More recently a second method using
2,4-dinitrophenylhydrazine•hydrochloride
coated XAD-2®  (termed a chemosorbent) in a
solid sorbent tube has been developed (4,5).
Because of its ease of operation, the chemo-
sorbent method was selected for evaluation.

In evaluating the candidate S/A method for
collection of formaldehyde in a simulated
stationary combustion source effluent,
several points were considered.  Of prime
consideration was the accuracy of the test
method versus that of an accepted indepen-
dent method (chromotropic acid) (6).
Another concern was the comparison of the
results from the simulated stack gas ana-
lyses versus the data from clean air back-
ground analyses, indicating possible inter-
ferences or chemical reactivity.  Lastly,
the stability of stored samples was ex-
plored.  In short, the test method was
expected to be able to provide a viable
collection medium for formaldehyde in the
simulated combustion matrix in terms of
ease of  operation, precision and accuracy.
EXPERIMENTAL

An atmosphere simulating a stationary com-
bustion source effluent was generated in
the laboratory.  The atmosphere consisted
of a mixture of gases at concentrations
which might be expected in the stack gas.
The gas matrix generated was 10% C02, 1000
ppm NOX, 100 ppm SOX with 68% relative
humidity (RH).  The formaldehyde level
produced was approximately 5 ppm by volume.

Simulation of the stack gas was achieved by
addition of component gases into a vapor
dilution system (Figure 1).  The system
had a designed air flow of 100 L/min at the
inlet.  The air was pretreated to remove
any particulate matter, organics or
water.  The standardized component gases
were then individually added into the
dilution system.

Formaldehyde was added first to the clean
air.  A 3.3% formalin solution in water
was added to the system at the rate of
.041 mL/min.  Theoretically, this would
have produced a formaldehyde atmosphere
of 14 yg/L.  However, due to the poly-
merization of formaldehyde to paraformal-
dehyde in the system, an atmosphere of
approximately 40% of the theoretical con-
centration, or about 5.6 ug formaldehyde/L
air, was generated.

Further downstream the background gases
were introduced to simulate a stationary
combustion source.  Nitric oxide (NO) was
delivered into the system as a 10% mixture
in nitrogen at 10 mL/min.  Sulfur dioxide
(S02) and carbon dioxide (C02) were added
as pure gases at the rate of 100 mL/min
and 11 L/min, respectively.  The total
dry gas through the system, therefore, was
approximately 111 L/min.  The water was
added at the rate of 1.4 mL/min and
vaporized into the system (Figure 2) pro-
ducing .10 g H20/g dry gas.  At an ambient
temperature of 20°C this was equivalent to
a RH of approximately 68%.

Replicate samples were collected for ana-
lysis of formaldehyde by both the test and
independent methods simultaneously.  In
the chromotropic acid determination method,
collection was made in a 1% sodium bi-
sulfite solution in midget bubbler pairs
connected in series.  Each impinger con-
tained 15 mL of the bisulfite solution.
The DNPH chemosorbent method collection
was accomplished through a 5% coating of
2,4-dinitrophenylhydrazine-hydrochloride
on XAD-2® packed in glass tubes.

The chemosorbent was prepared in the labo-
ratory by coating the DNPH-HC1 on to the
resin, as follows:

• The DNPH-HC1 was made by dissolving 2,4-
  dinitrophenylhydrazine in hot 4N HC1.
• The hydrochloride produced was crystal-
  lized by cooling and recrystallized
  again from fresh 4N HC1.
• The hydrochloride was then dissolved in
  a 9:1 mixture of ethanol:HCl (cone) to
  produce a 5% coating on the XAD-2®.
• The solution was placed in a rotary
                                            57

-------
                                           A P Gauge
                                                        0-2
tn
CO
                                       Syringe Drive
                                                                                          = Critical Flow Orifice (CFO)
                                     Figure   1
Apparatus for Vapor  Generation/Dilution/Sampling  System

-------
Add Gas
  or
Aerosol
                    Main Line
                                                                 Water Reservoir
                                                                      i  r
                                                            Copper Coil
                                                Tube Furnace
Liquid
               Figure  2        Apparatus for  Generation of High Humidity Atmospheres

-------
  evaporator with clean XAD-2® and the
  solvent removed.
• The dry chemosorbent (ca 200 mg/tube) was
  then packed into 10 cm x 6 mm OD (1 mm
  wall) glass tubes, with glass wool  plugs.

The flow through each of the sample ports
was controlled by a critical flow orifice,
calibrated with a typical sampling device
in line.  All orifices were used on a
single vacuum system to ensure simultaneous
sampling.  Collection through the impingers
was at approximately 1 L/min for one hour.
Collection through the chemosorbent was at
approximately .2 L/min for 20 minutes.

Immediately after sampling, half of the
samples were prepared for analysis, while
the rest were sealed for storage stability
testing.  The sodium bisulfite collection
pairs were combined and diluted to 50 mL,
from which an aliquot was taken and ana-
lyzed according to the APHA chromotropic
acid method (#116). The chemosorbent samples
were removed from'their tubes and desorbed
with acetonitrile in small vials and ana-
lyzed directly by HPLC using an acetoni-
trile/water eluant and 365 nm UV detector.

Accuracy of the methods was checked by
standard addition.  One-third of the sam-
ples for the stored chemosorbent set were
spiked with an amount of a dilute formalin
solution equal to the amount expected from
the generator collection.  This standard
addition method provided a measure of
quality control and absolute accuracy of
the test method by confirming compound
identification and recovery data.

Precision of the test method was determined
by agreement of the replicate formaldehyde
concentration results.  Accuracy of the
chemosorbent method was further checked by
comparison with the independent reference
method, i.e.,  chromotropic acid deter-
mination of formaldehyde.

Storage stability of the samples was
checked by sealing half of the collected
samples in their respective sampling devices
and setting them aside for a period of
seven (7) days in the dark at ambient condi-
tions.
RESULTS

The evaluation of the DNPH/XAD-2® method
consisted of two independent studies.  The
chemosorbent was first evaluated in a clean
air environment, i.e., no interferring back-
ground gases, at low humidity (<5% RH).   Two
atmospheres of formaldehyde were generated
under identical conditions and the results
pooled (see Table 1).  The results show
excellent agreement between the DNPH'HCl
coated XAD-2® chemosorbent and the chrom-
tropic acid reference methods.  An atmos-
phere containing the formaldehyde at a
slightly higher concentration with 10% C02
1000 ppm NOX, 100 ppm SOX, and 70% RH was
then produced. Again, two separate
collections were made.  Half of the samples
from each collection were set aside for
storage stability analysis.  The remaining
samples were analyzed within  24 hours of
collection.  The pooled results from the
"next day" analyses  (short-term storage),
shown in Table 2, again indicate excellent
agreement between the chemosorbent and the
chromotropic acid reference methods, 6.17
yg/L and 6.12 yg/L respectively.  These
results indicate that the acidic nature of
an unscrubbed combustion effluent should
produce no bias on the collection and
analysis of formaldehyde using the chemo-
sorbent method.

The stored samples (long-term storage)
indicate the same agreement.  As shown in
Table 3, the samples stored for seven days
at ambient conditions in their sealed
sampling containers had concentrations of
7.45 yg/L by the chromotropic acid method
and 7.20 yg/L by the chemosorbent method.
These results represent a statistically
significant increase in formaldehyde
concentration, for the four pairs of chro-
motropic acid results and nearly the same
for the seven pairs of chemosorbent results
(one outlier pair from the  eight analyzed
was disregarded).  Entrainment of parafor-
maldehyde from the generation system into
the sampling devices may account for the
increased formaldehyde concentration over
time.

The evaluation of spike recovery performed
on a number of the storage samples shows
results confirming the excellent agreement.
Four of the sample tubes were spiked with
20 yg of formaldehyde in solution, prior to
the collection on the generator system.
Recovery data were calculated assuming a
theoretical collection based on the mean
measured formaldehyde level and the col-
lected volume of vapor.   The percent recov-
ery was found by subtracting the amount of
formaldehyde expected from the generating
system alone from that actually recovered.
Using this method, a mean recovery of 107%
                                            60

-------
TABLE 1: FORMALDEHYDE COLLECTION - "CLEAN AIR" RESULTS  (Combined from two sample runs)


Sample No.
1
2

3
4
5
6
7
8
Mean
Std. Dev.
C.V.


Sample No.
1
2
3
4
5
6
7
8
9
10
Mean
Std. Dev.
C.V.

Total CH20
(yg)
80.88
77.00

87.63
92.00
65.64
72.43
29.40
66.06




Total CH20
(yg)
53.72
49.34
51.20
53.26
48.71
57.36
58.79
53.76
61.99
53.22



Chromotropic Acid

Volume Sampled CH20 Cone
(L) (yg/L)
15.63 5.17
15.57 4.95

18.82 5.54
15.30 6.01
15.63 4.20 \
15.57 4.65 1
15.82 1.86*
15.30 4.32



DNPH/XAD-2Q (Chemosorbent)



Run 1



Run 2


4.97
.654
.131

Volume Sampled CH20 Cone
(L) (yg)
10.74 5.00 \
10.58 4.66 /
11.04 4.63 /
11.46 4.65 1
10.55 4.62 '
10.74 5.34 \
10.58 5.56 I
11.04 4.87 /
11.46 5.41 \
10.55 5.05 /






Run 1




Run 2


4.98
.355
.071
Source: Results  in above  table obtained from a report prepared for U.S. Bureau of Mines
*Values omitted  from  data summary-
Note: All samples have been corrected for blanks which accounted for less than 3% in each
      of the samples.
C.V.  Coefficient of  Variation (= Relative Standard Deviation)


TABLE 2: FORMALDEHYDE COLLECTION IN SIMULATED STACK GAS - "NEXT DAY" ANALYSES
Sample No.
    1
    2
    3
    4
  Mean
  Std. Dev.
  C.V.
Total CH20
   (yg)
     349
     441
     410
     455
Chromotropic Acid
      Volume Sampled
           (L)
           59.5
           76.5
           59.5
           76.5
CH20 Cone
   (yg)
     87
     76
     89
     95
                                                           6.12
                                                            • 521
                                                            .083
                                            61

-------
                                   TABLE  2  (Continued)

                                 PNPH/XAD-2®  (Chemosorbent)

Sample No.
1
2
3
4
5
6
7
8
Mean
Std. Dev.
C.V.
Total CH20
(UR)
16.4
19.7
27.9
17.4
3.32
16.9
23.9
15.5



Volume Sampled
(L)
2.66
3.75
3.73
2.45
2.66
3.75
3.73
2.45



CH20 Cone
(UR/L)
6.17
5.26
7.47
7.08
1.25*
4.50
6.39
6.32













6.17
1.02
.165
*Values omitted from data summary-
Note:  All samples have been corrected  for  blanks  which accounted for less than 2% in each
      of the samples.
TABLE 3.   FORMALDEHYDE COLLECTION IN SIMULATED  STACK GAS  - STORAGE STABILITY DATA
Chromotropic Acid

Sample No.
1
2
3
4
Mean
Std. Dev.
C.V.


Sample No.
1
2
3
4
5
6
7
8
Mean
Std. Dev
C.V.
Total CH20
(PR)
108
353
141
411



DNPH/XAD-2<»
Total CH20
(yg)
30.0
28.6
21.9
6.57
26.8
26.3
19.2
5.90



Volume Sampled
(L)
15.72
54.72
15.72
54.72



3 (Chemosorbent)
Volume Sampled
(D
3.69
3.68
2.54
1.12
3.69
3.68
2.54
1.12



CH20 Cone
(PR)
6.87
6.45
8.92
7.51




CH20 Cone
(UR/L)
8.11
7.78
8.62
5.87
7.27
7.14
7.56
5.27









7.45
1.10
.148











7.20
1.12
.156
Note: All samples have been corrected for blanks which accounted for less than 1% in each
      of the samples.
                                            62

-------
was determined  (see Table 4).  Allowing for
variation in the actual collections on the
vapor generator, these results indicate a
probable recovery of 100%.
CONCLUSIONS

The DNPH/XAD-2® chemosorbent appears to be
a viable method for the sampling and ana-
lysis of formaldehyde in incineration or
combustion source environments.  The chemo-
sorbent is relatively simple to prepare and
remains stable for several months if stored
in the dark.  The sorbent tube has proven
to be a very convenient method to use in
the field in terms of ease of handling and
in shipping.  The chemosorbent has the add-
ed feature of producing a slight color
change from yellow/brown to yellow as the
hydrazone is formed, allowing the user to
know if breakthrough is imminent.  The tube
is easily desorbed for  analysis, and  the
results are specific for formaldehyde.
Overall, the method appears to be a prime
candidate for. formaldehyde S/A, and could
possibly, be extended to provided specific
methods for other carbonyls.  Additional
testing may be required for applicability
at higher stack temperatures.
ACKNOWLEDGEMENT

The authors wish to thank the EPA, Dr.
Raymond G. Merrill, and Dr.  Larry Johnson
for its encouragement in this study under
EPA Contract No. 68-02-3627; Rose E.
Fasano for technical assistance; and  Paula
Sullivan for her editoral comments.  The
authors would also like to thank Mr.
Kenneth Menzies, and Ms. Kathleen Thrun for
their continual expert assistance.
DISCLAIMER

The research described in this article has
been partially funded by the United States
Environmental Protection Agency through
Contract 68-02-3627 to Arthur D.  Little,
Inc.  It has been subjected to the Agency's
required peer and policy review.   Approval
does not signify that the contents neces-
sarily reflect the views and policies of
the Agency, nor does mention of trade names
or commercial products constitute endorse-
ment or recommendation for use.
 XAD-2®  is  a  registered  trademark of  Rohm & Haas, Inc., Philadelphia, PA.


 TABLE 4: FORMALDEHYDE COLLECTION IN  SIMULATED GAS - SPIKE RECOVERY DATA (7.2 yg/L Atmos-
         phere)
Sample
No.
Volume
Sampled
(L)
1
2
3
4
Mean
Std. Dev.
C.V.
3.
5.
3.
5.



47
29
47
29



Expected
CH?0
(yg)
25.
38.
25.
38.



,0
.1
.0
1



Recovered
CH70
(yg)
44
62
44
60



.4
.2
.5
.9



Differences
Due to Spike
(yg)
19,
24.
19.
22,



.4
.1
.5
.8



Spike
CH70
(yg)
20
20
20
20



Percentage
Recovered
(%
97.
120
97.
114
107
11.

)
0

5


7
109
Note: All  samples have been corrected for blanks which accounted for less than 2% in
      each of  the samples.
                                            63

-------
REFERENCES

1.   Grady,  V.A.,  1981,  Candidate Methods  for Sampling and Analysis of Twenty-one High
     Probability  Carcinogens.  Environmental Protection Agency,  TSS/IERL,  Research
     Triangle Park,  NC  Contract  68-02-3627.

2.   Menzies, K.T.,  Beltis,  K.J.,  Fasano,  R.F.,  1982,  Comparison of aldehyde methods,
     SAE Technical Paper presented at  West Coast International  Meeting, San Francisco
     California.

3.   Dietzmann, H.E.,  Smith, L.R., Parness,  M.A.,  Fanick,  E.R.,  1979,  Analytical
     Procedures for Characterizing Unregulated Pollutant  Emissions from Motor Vehicles,
     Environmental Protection  Agency,  600/2-79-019.

4.   Menzies, K.T.,  Beltis,  K.J.,  Wong,  C.M.,  In Preparation, Development of Sampling
     and Analytical Methods  for Toxicants  in Diesel  Exhaust Streams,  U.S. Department
     of the  Interior,  Bureau of Mines, Contract  J0308005.

5.   Andersson, G.,  Andersson, K., Nilsson,  C-A.,  Levin,  J.O.,  1979,  Chemosorption of
     formaldehyde  on Amberlite XAD-2®  with 2,4-dinitrophenylhydrazine,  Chemosphere,
     10, 823-827.

6.   Altshuller, A.P., Miller, D.L., Sleva,  S.F.,  1961, Determination  of  formaldehyde
     in gas  mixtures by  the  chromotropic acid  method,  Anal.  Chem.,  33,  621-623.

7.   Katz, M., Editor, Methods of  air  sampling,  American  Public  Health Association.

8.   Kuwata,  K., M.  Verbori, Y. Yamasaki,  1979,  Determination of aliphatic and aromatic
     aldehydes in  polluted airs as their 2,4-dinitrophenylhydrazones by HPLC,  J.  of
     Chromatographic Sciences, 17, 264-268.
                                           64

-------
                FACTORS AFFECTING  THE  GAS-PHASE
                   THERMAL DECOMPOSITION OF
               CHLORINATED AROMATIC  HYDROCARBONS

                        Barry  Dellinger
                        Douglas  L. Hall
                        Wayne  A. Rubey
                        Juan L.  Torres
                     University  of Dayton
                      Research Institute
                 Environmental Sciences Group
                       Dayton, OH  43469

                       Richard A.  Carnes
             U.S. Environmental  Protection  Agency
               Combustion Research Facility/NCTR
                      Jefferson, AR  72079
                           ABSTRACT
     In this presentation, we  report  the  results of laboratory
studies concerning  the high-temperature  gas-phase thermal decom-
position of seven different chlorinated  benzenes.  The generally
observed trend is toward  increasing thermal stability with
increased chlorine  substitution.   The relationship of thermal
decomposition to time and temperature can be adequately
described by first  order  kinetics  and application of the
Arrhenius equation.  Studies of  the effect of oxygen concentra-
tion on the thermal  decomposition  of  hexachlorobenzene and
pentachlorobenzene  suggest that  the susceptibility to oxygen
attack is reduced with increased chlorine substitution.  Data
also suggest that the decomposition mechanism changes with
temperature.

INTRODUCTION                     much  more desirable.  Under the
                                 current  interim final rule  (1)
     Regulating the  disposal of  on incineration, the incinerator
hazardous waste as  promulgated  operator must show that the fac-
by the Resource  Conservation    ility can destroy those waste
Recovery Act (RCRA)  of 1976 has  constituents most difficult to
made some traditional methods    incinerate.  In theory the per-
of disposal, such as  impounding  mit writer will select compounds
and landfilling, less desirable  within the mixture that are of
and consequently, incineration  sufficient toxicity, concentration,

                               65

-------
and thermal stability so as to
be designated as principal
organic hazardous constituents
(POHCs) .   It must then be shown,
possibly by trial burn, that the
designated POHCs can be
destroyed by the particular
incineration system to a
destruction and removal effici-
ency (DRE) of 99.99%.  Further-
more,  the specific operating
conditions must be established
under which the 99.99% DRE is
achieved.

     A univeral ranking proce-
dure which readily allows deter-
mination of thermal stability
and designation of POHCs is of
interest to everyone, but
primarily to the permit writer.
The incinerator operator is by
necessity concerned with both
economical operation and the
determination of the optimum
waste- disposal conditions.  It
is generally acknowledged that
incineration is a complex pro-
cess and our present knowledge
does not permit the designation
of POHCs and overall operating
conditions without some
uncertainty.

     One approach to establish-
ing optimum incineration condi-
tions is to perform the.necessary
determinations by trial and
error using the actual incinera-
tion system; however, due to the
difficulty of controlling all
of the parameters affecting
full-scale incineration, this
approach can be inefficient,
incorrect, and time consuming.
Extrapolation of data from one
facility to another is very
difficult  (2) due to differences
in wastes, basic design charac-
teristics, and the inability to
develop simple physical models
for complex systems.
     An alternative approach is
to generate data in the labora-
tory using well characterized
and controlled conditions.  The
challenge in this approach is
to determine which incineration
variables are most significant
with respect to destruction
efficiency  (DE) and design
suitable laboratory experiments
which permit evaluation of these
variables in a manner that is
amenable  to eventual scale up
conditions.

     Since  the nominal condi-
tions within the flame zones of
large field-scale incinerators
are usually sufficient to decom-
pose any gas-phase waste';com-
ponents, our approach has been
to address  the 'exceptional'
molecules which escape this
highly reactive environment.
Although temperatures are much
higher in the flame than non-
flame region of the incinerator,
molecules will experience a
much greater residence time in
the non-flame region.  (It is
this region, i.e., the non-
flame region of the afterburner
that provides the final thermal
decomposition environment which
significantly contributes to
the attainment of an overall
system DRE  of 99.99%.)

     In this presentation we
address a class of compounds
known both  for their toxlcity
and thermal stability, the
chlorinated hydrocarbons  (CHCs).
We report the results of our
studies which were designed to
gain insight into the factors
which control the non-flame
thermal decomposition of these
compounds.  In this work we
have attempted to separate the
chemical parameters from the
physical parameters.  We then
                               66

-------
studied the effect of the
chemical parameters on thermal
stability of  the  test compounds.
Specifically, we  have addressed
the effect of reaction atmo-
sphere  (oxygen  concentration) ,
mean residence  time,  and tem-
perature on various chlorinated
benzenes.  A  simple mechanistic
model is proposed to account
for the observed  form of the
thermal decomposition profiles.

Experimental

     The affect of mean resi-
dence time and  temperature on
the thermal decomposition of
chlorinated benzenes was deter-
mined on the  thermal decomposi-
tion unit-gas chromatographic
(TDU-GC) system which was
designed and  built with funding
provided by the US-EPA
(Cooperative  Agreement No. CR-
807815-01-0).   A  block diagram
of the  TDU-GC is  shown in
Figure  1.  The  data pertaining
to the  effect of  oxygen concen-
tration on thermal decomposition
behavior was  obtained using the
thermal decomposition analytical
system  (TDAS),  also designed
and built with  funding provided
by the  US-EPA (Grant No.
R805117-01-0).  These units
are quite similar in design and
operation and data compares very
favorably between the two sys-
tems.   Their  major difference is
that the  TDAS  utilizes an LKB
2091 GC/MS for  data acquisition.
(3)  Kinetic  data from the TDAS
was obtained  using the GC in
combination with  the total ion
current (TIC) detector of the
mass spectrometer.

     Both thermal reactor assem-
blies are constructed of 1 mm
nominal I.D.  fused quartz tubing
in a race track configuration
(3.5 cycles,  1  meter in length).
This quartz  tubular reactor
construction  was  chosen to min-
imize the possibility of wall
reactions while simultaneously
providing a  very  narrow resi-
dence time distribution and a
square-wave  high-temperature
exposure profile.  (4)

     High vapor pressure liquid
phase samples were  prepared at
concentrations  of approximately
10 ppm in air.   (Precise
quantitation  is not required
since data reduction employs a
difference method involving
the comparison  of the partially
decomposed sample peak size
with a non-decomposed quanti-
tation peak.)   These samples
were slowly  injected into a
flowing air  carrier stream for
transport through the reactor.
Low vapor pressure  liquid phase
samples were  injected onto
quartz wool  in  a  temperature
programmable  insertion chamber.
The temperature was then
increased at  a  rate of 12°C/min.
from 0° to 250°C  such that the
sample was slowly volatilized
for transport.  Solid phase
samples tested  using the TDU-GC
were dissolved  in cyclohexane
and then injected into the
insertion chamber which was
subsequently  heated at a pro-
grammed rate.   Solid phase sam-
ples that were  tested using the
TDAS were prepared in solution
and then deposited on a glass
probe whereupon the solvent was
evaporated.   The  probe was then
placed into  the insertion cham-
ber of the TDAS and heated at a
programmed rate.   This process,
which involves  a  random entry-
controlled thermal  exposure
(RE-CTE) of  sample  molecules,
is described  more thoroughly in
a previous paper. (3)   This mass
                                67

-------
                       THERMAL DECOMPOSITION UNIT
CAPTURE
  OF
EFFLUENT
PRODUCTS
         \
CONTROLLED
 HIGH
TEMPERATURE
 EXPOSURE
          HIGH TEMPERATURE TRANSFER
                        MULTIFUNCTIONAL
                       GAS CHROMATOGRAPHIC
                        INSTRUMENTATION
    Figure  1.   Block  Diagram of  the  Thermal  Decomposition
                  Unit-Gas Chromatograph  (TDU-GC)  System.
                                  68

-------
transport procedure is designed
to insure that the sample  con-
centration is very dilute  and
in an oxygen rich environment.

     Gases used in this  study
were purchased from Ai r Products ,
Inc. and contained less  than 2
ppm total hydrocarbons.  All
gases were passed through  a
molecular sieve trap  before
introduction into the instru-
mentation systems.

Re s ul ts

     A homologous series of six
chlorine substituted  benzenes
were examined to determine their
thermal stability in  an  oxida-
tive environment.  Figure  2
depicts the thermal decomposi-
tion profiles at a mean  resi-
dence time of two seconds,
tr = 2.0 seconds, for each of
the six chlorinated benzenes
studied.  From these  plots one
can ascertain that each  of the
compounds is very stable with
little decomposition  occurring
at exposure temperatures up to
about 550-650°C.  Once the
decomposition begins, it reaches
the 99.99% DE level with a fur-
ther temperature increase  of
approximately 200°C.  Table 1
presents the temperatures
corresponding to the  onset of
thermal decomposition and  the
extrapolated temperatures  nec-
essary for 99.99% DE.

     Data have also been
obtained at tr = 1.0, 4.0, and
6.0 seconds.  The family of
decomposition profiles at  dif-
ferent residence times for
hexachlorobenzene  (HCBz) is
depicted in Figure 3.  It  is
evident from this plot that
increasing the HCBz residence
time in the reactor   reduces
the temperature required  for  a
given destruction efficiency.

     First order kinetic  rate
expressions have been  found to
adequately describe  the time
versus temperature relationship
for such thermal decomposition
data by applying the Arrhenius
equation to the temperature
dependence of the first order
rate constant (5,6) .   However,
previously obtained  data (7) for
2,2',4',5,5'-pentachlorobiphenyl
(2,2',4' ,5,5'-PCB) clearly
indicated a pronounced depen-
dence of the thermal decomposi-
tion behavior on the oxygen
concentration.  Specifically,
the disappearance of the  PCB
was found to be first  order in
oxygen (5).  This behavior indi-
cates that the observed first
order kinetics is actually
a more complex phenomenon
involving reaction with oxygen,
possibly as the rate control-
ling step in the thermal  decom-
position.  The disappearance
kinetics then, are actually
pseudo first order in  the con-
centration of the parent
because of the great excess of
oxygen concentration over that
of the test sample.

     Figures 4 and 5 depict the
thermal decomposition  profiles
for hexachlorobenzene  and
pentachlorobenzene at  tr  = 2.0
seconds for various  concentra-
tions of oxygen in nitrogen
from 2.5% to 70%.  Over this
range, it is apparent  that
increasing the oxygen  concen-
tration promotes the decomposi-
tion of the parent compound.
However, the effect  is more
dramatic for pentachlorobenzene
than hexachlorobenzene.   In
conjunction with previous
observations with  the
                               69

-------
LJ
   100
UJ   m
or   !U
a
o:
UJ   ,n
o_   IAJ
   0.
  0.01
O MONOCHLOROBENZENE
A 1,2-DICHLOROBENZENE
O !,2,4-TRICHLOROBENZENE
D 1,2,3,5-TETRACHLOROBENZENE
V PENTACHLOROBENZENE
O HEXACHLOROBENZENE
      0   50
                        I
                                 I
500      600     700     800
   EXPOSURE TEMPERATURE, °C
                                         900
1000
             Figure 2.   Thermal  Decomposition Profiles for
                        Selected Chlorobenzenes in Air at
                        2.0  Seconds Mean Residence Time.

-------
                     TABLE 1

       THERMAL  DECOMPOSITION PARAMETERS
   FOR SELECTED CHLORINATED BENZENES  IN AIR
         AT  2.0 SECONDS RESIDENCE  TIME
                       Temperature  for     Extrapolated
                      Onset of Thermal    Temperature  for
                        Decomposition        99.99% ORE
Compound                      °C                  °C

Benzene
Monochlorobenzene
1 , 2-Dichlo.robenzene
1,2, 4-Trichlorobenzene
1,2,3, 5-Tetrachlorobenzene
Pen tachlo robe nzene
Hexachlorobenzene
610
520
615
620
620
675
615
760
790
790
795
800
840
880
                        71

-------
       100
-j
ro
     UJ
     o:
     LJ
     o
     ct:
     LU
     CL
                                                                        O
       0.
      0.01
.HEXACHLOROBENZENE


    O  ?r= 1.0


    D  tr = 2.0


    A  fr =4.0


    O  tr =6.0



   i    I     i    I
          0
                00
                                             I
                           I
600               700

       EXPOSURE TEMPERATURE,°C
800
900
                  Figure  3.   Thermal  Decomposition  Profiles  for  Hexachlorobenzene

                              in Air at Mean Residence Times  of 1.0,  2.0,  4.0,

                              and 6.0  Seconds.

-------
   100
S   10
LU
O
DC
S   1-0
X
O
UJ
   0.1
  0.01
       O
                                  07O.4
                                  021.0
                                  D1QO
                                  A  2.5
650
 700        750       800

EXPOSURE TEMPERATURE,  °C
050
900
               Figure 4.   Thermal Decomposition Profiles for Hexachlorobenzene
                         in 2.5,  10, 21,  and 70% Oxygen in Nitrogen.

-------
   10O
£  10
UJ
O
cc
uj   1.0
X
O
UJ
   0.1
  0.01
      0
650
       tr = 2.0 sec.
                                    07O.4
                                    021.0
                                    D1QO
                                    A 2.5
 700       750      800       850

EXPOSURE TEMPERATURE,  °C
900
               Figure 5.   Thermal Decomposition Profiles  for Pentachlorobenzene
                         in 2.5, 10, 21, and 70% Oxygen  in Nitrogen.

-------
2, 2 ' , 4 ' ,5 , 5 '-PCB,  some  potenti-
ally useful and  interesting
models can be hypothesized.

Discussion

     The  development of a model
for the prediction of thermal
destruction efficiencies of haz-
ardous organic compounds is of
obvious utility.   An equation
has been  developed to predict the
required  temperature for 99.99%
DE. (6)

T99>99=503Ea(ln  0.109 Atr)~l  (1)

Where:  Tgg.gg = the temperature
        required for 99.99% DE,
        °K
T99.99 of  880°c  (Ea  =  41  kcal/
mole and A = 2 x  108s~l.

     The decomposition of the
parent at  the  temperatures
studied  can in principle be due
to four general mechanisms:
unimolecular decomposition,
attack by  molecular  oxygen,
chain reactions involving mole-
cular decomposition  products,
and chain  reactions  also  involv-
ing oxygen.  The  four  possible
decomposition  pathways may be
treated with two  global mechan-
isms, one  dependent  upon  the
oxygen concentration (oxidation)
and a second,  dependent on the
bond strengths of the  parent
compound  (pyrolysis).
        Ea=  the  measured activa-      The         dependence of
        tion energy,  kcal mole 1 the ^composition  of the  parent
        A =  the  measured         A, may be expressed  in  terms of
        Arrhenius  coefficient,   the global reaction  with
        s~l                       oxygen:
        tr =  the  mean residence
        time,  s~l

This expression  can be used to
accurately determine the tempera-
ture required for 99-99% DE,  at
various high  temperature zone
residence times,  for compounds
for which the kinetic parameters
Ea and A are  known.  The model
explicitly assumes that the dis-
appearance of the parent com-
pound is first order with respect
to the parent compound and zeroth
order for all other species.
For each compound tested thusfar,
thermal decomposition data taken
between Er -  1-0  and 6.0 seconds
(with more than  one percent of
the original  sample destroyed)
has fit first order kinetics  with
correlation coefficients  (r^)
greater than  0.95.  The calcul-
ated T99.99 at tr = 2.0s for
hexachlorobenzene is 888°C, com-
pared to the  extrapolated
A + nC>2 (xs)
            k2
                   products  (2a]
while the pyrolysis  is given
simply by:

   kl
A - >  products            (2b)
This results  in  the  overall
kinetic expression  for the
disappearance of the parent:
                              3)
               k2[A]  [02Jn
where:
         [A]  is  the  concentration
             of  the  parent,

         [02]  is the oxygen
             concentration,  and

        k]_  and  k2  are the global
             rate constants for
             pyrolysis and oxida-
             tion respectively.
                                 75

-------
Figure 6 presents the experi-
mentally determined reaction
order with respect to oxygen as
a function of temperature  for
pentachlorobenzene and hexa-
chlorobenzene.   The dependence
upon the oxygen concentration
for pentachlorobenzene is  much
more pronounced than for hexa-
chlorobenzene,  the latter  being
almost constant.  In the case of
pentachlorobenzene, the decom-
position would appear to be
dominated by oxygen attack at
low temperatures while it  is
much less oxygen dependent at
higher temperatures.

     The implication of the data
thus far obtained on 2 , 2 ' , 4 ' , 5 , 5 '-
PCB  (which was first order in
oxygen at 704°C), pentachloro-
benzene, and hexachlorobenzene
is that the susceptibility to
oxygen attack is a function of
both the number of hydrogens
present in the parent compound
and the reactor temperature.
This is consistent with a  mech-
anism of oxygen attack through
abstraction of a hydrogen  radi-
cal.  The rate of the pyrolysis
pathway may be controlled  by the
strength of the weakest bond. (8)
This suggests that the initial
decomposition of HCBz .(which
contains no hydrogens) is  depen-
dent on the carbon chlorine bond
strength while in pentachloro-
benzene and 2,2',4 ' ,5,5'-PCB the
initial decomposition results
from hydrogen abstraction  by
molecular oxygen.

Summary

     The study of the phenomena
occurring in the non-flame
afterburner region of an incin-
eration system is only one part
of the very complex process we
know as incineration.  However,
one must recognize  that the
afterburner is  the  component of
the system which  allows it to
achieve the desired destruction
efficiencies.   If all  gas phase
matter were exposed to the nomi-
nal conditions  of a high-
temperature kiln, multiple hearth
fluidized bed,  or liquid injec-
tion incinerator, they would be
destroyed with  greater effici-
ency than 99.99%.   Unfortunately
due to the heterogeneous nature
of industrial organic  wastes and
various nonuniform  transport
properties through  certain por-
tions of incineration  systems,
there is breakthrough  of
'exceptional' molecules which
experience less destructive
conditions.  Thus it is up to
the non-flame high-temperature
region of the system to destroy
these residuals and in fact,
control the ultimate DE of the
incinerator.

     Our data indicate a complex
relationship between destruction
efficiency, time, temperature,
and reaction atmosphere in this
region.  The time/temperature
relationship can  be adequately
modeled with first  order kin-
etics.  However,  this  relation-
ship of time and  temperature to
the reaction atmosphere is not
so simple.

     Our observations  suggest
that excess oxygen  concentra-
tions does not  appear  to be as
important for the thermal
degradation of  highly  chlorin-
ated species as for hydrocar-
bons.  This effect  is  most
pronounced at high  temperatures.
Based on these  thermal decom-
position studies, the  optimum
combination may be  very com-
pound specific  for  chlorinated
waste materials.
                                76

-------
    1.0
 CM
o
t_

-------
Credits                          4 .

     This work was performed
under the sponsorship of the
US-EPA through Contract 68-03-
2979 and Cooperative Agreement
CR-807815-01-0.

Acknowledgement                  5 .

     We gratefully acknowledge
our colleague, Mr. John M.
Duchak, for his assistance in
the generation of some of the
experimental data.

REFERENCES

1.  Federal Register, Vol. 47,
    No. 122, Thursday, June 24,  6.
    1982.

2.  L. Weitzman, "Scale-Up
    Criteria for Incinerators,"
    presented at the 73rd APCA
    Meeting, Montreal, Quebec,
    June, 1980.                  7-

3.  J. L. Graham, W. A. Rubey,
    and B. Dellinger, Determina-
    tion of Thermal Decomposi-
    tion Properties of Toxic
    Organic Substances, pre
    sented at the Summer
    National Meeting of the
    AICHE, Cleveland, OH,
    August, 1982.
W. A. Rubey,  Design
Considerations  for a Thermal
Decomposition Analytical
System  (TDAS),  EPA-600/2-
80-098, U.  S. Environmental
Protection  Agency,
Cincinnati, OH,  August,  1980.

B. Dellinger, D.  S.  Duvall,
D. L. Hall, W.  A.  Rubey,
and R. A. Carnes,  Laboratory
Determination of High-
Temperature Decomposition
Behavior of Industrial
Organic Materials,  presented
at the 75th APAC
Meeting, New Orleans, LA,
June, 1982.

K. C. Lee,  H. J.  Jahnes, and
D. C. Maculey,  "Thermal
Oxidation Kinetics  of
Selected Organic Compounds,
JAPCA, Vol. 29,  No.  7,
pp. 749-751, July,  1979.

D. S. Duvall, and W. A.
Rubey, Laboratory  Evaluation
of High-Temperature  Destruc-
tion of Polychlorinated
Biphenyls and Related
Compounds, EPA-600/2-77-228,
U. S. Environmental
Protection Agency,
Cincinnati, OH,  December,
1977,

W. Tsang, and S.  Shaub,
Chemical Processes  in the
Incineration of Hazardous
Materials, presented at  the
ACS Symposium on
Detoxification  of Hazardous
Wastes, New York,  NY,
August, 1981.
                               78

-------
             LABORATORY-SCALE FLAME MODE STUDY OF HAZARDOUS WASTE INCINERATION
                                        W. R. Seeker
                                       J. C. Kramlich
                                         M. P. Heap

                       Energy and Environmental Research Corporation
                                          18  Mason
                                      Irvine,  CA  92714

                                         C. C. Lee
                             Industrial Waste Combustion Group
                              Environmental Protection Agency
                                          ABSTRACT

A research program to study the flame-mode incineration of hazardous waste liquids in
laboratory scale reactors is presented.  The objective of this study was to supply the
flame-mode data that will assist in the evaluation of the applicability of various
approaches to ranking ease of incinerability.  Two reactors, each emphasizing different
aspects of liquid injection incinerator performance, were utilized.  Five common liquid
waste compounds were selected for testing:  benzene, chlorobenzene, chloroform, 1,2-
dichloroethane, and acrylonitrile.  In the Microspray Reactor, monodisperse waste drop-
lets were injected into  laminar postflame gases to study the combined effects of droplet
evaporation, droplet-flame stabilization, and flame-zone chemistry on the compound destruc-
tion efficiency.  A turbulent flow reactor used a commercial spray nozzle to inject an
auxiliary fuel, containing dilute test compound, into a commercial spray nozzle to inject
an auxiliary fuel, containing dilute test compound, into a swirl-stabilized turbulent-
spray flame.  In addition to the droplet and chemical effects considered in the microspray
reactor, the turbulent flow reactor includes the effect of turbulent mixing on destruction
efficiency.  This paper  outlines the program goals in relation to the incinerability rank-
ing procedure, presents  the experimental design, and discusses data.
INTRODUCTION

     Permitting procedures for hazardous
waste incinerators are defined by the Re-
source Conservation and Recovery Act (RCRA).
.A permit to operate is issued after a trial
burn has been executed or other appropriate
test data obtained which demonstrated that
the incinerator satisfactorily eliminated
the hazardous compounds when operated under
specified conditions.  Satisfactory elimi-
nation is defined in terms of destruction
and removal efficiency (DRE).  However,
since most hazardous waste streams contain
many compounds, a trial burn which involves
the measurement of all of them would be
prohibitively expensive.  Consequently, the
trial  burn involves the measurement of a
subset of compounds (the principal organic
hazardous constituents--POHCs) which are
present in the input stream.  If the DRE of
these POHCs is 99.99 percent or greater,
then a permit to operate is granted.  Thus,
the burden of responsibility rests with the
permit writer who must select the subset of
compounds (POHCs) based upon concentration
and incinerability.  This paper summarizes
a project which was carried out to examine
methods of ranking incinerability.

     Several  procedures have been proposed
to rank incinerability (2), namely:

t  The heat of combustion.
•  The auto-ignition temporation  (AIT).
                                            79

-------
•  A computational  approach based upon AIT,
   compound structure,  and other compound-
   dependent parameters (10,11).

•  The temperature necessary for a  given
   destruction level  within a given time
   under dilute premixed conditions
   (T99-99) (4,6,7,10,11).

•  Susceptibility of  the compound bond
   structure to attack  by flame radicals
   (16).

These procedures have their merits  but fail
to take into account  all the conditions
which may exist in actual  incinerators.
The heat of combustion, for example,  of  a
particular compound may be insignificant
if it is present in small  quantities  and
is mixed with an auxiliary fuel.  Some of
these procedures do not consider reactions
that occur in flames.  The times and  tem-
peratures which exist under nonflame exper-
imental conditions may  be inappropriate
for large-scale diffusion flames.

     Incinerability defines the relative
difficulty with which a compound can  be
destroyed by incineration.  If during  the
trial burn it is demonstrated that  those
compounds which are most difficult  to
destroy have a ORE greater than 99.99  per-
cent, then it is assumed that all the  more
easily incinerated compounds will have
been satisfactorily eliminated.  Thus,
there is a need for a system which  will
rank hazardous compounds according  to  their
incinerability and be applicable to all
incinerator types and operating conditions.
If the ranking system is not generally
applicable, then a condition could  exist
wherein a POHC was eliminated satisfactor-
ily but other hazardous compounds in  the
waste stream were not destroyed suffici-
ciently.  Under these circumstances,  a
trial burn designed to  measure only the
POHC would have incorrectly demonstrated
the satisfactory operation of the incin-
erator.

     Because of the nature of flames,
waste compounds which experience a  flame
environment are rapidly and effectively
destroyed.  In a liquid injection inciner-
ator, the waste is destroyed by heat  and
radical attack produced by the combustion
of the waste and an auxiliary fuel  (oil  or
natural gas) in a turbulent diffusion  spray
flame.  Under ideal flame conditions  (uni-
form mixing and high  temperatures)  the
concept of incinerability has little
significance since all hazardous  compounds
will be completely destroyed.  Nonflame
thermal decomposition data obtained  under
dilute premixed conditions (4,6,7,10,11)
indicate that temperatures much lower  than
those encountered in typical  incinerator
flames will destroy all the organic  hazard-
ous waste compounds which have been  tested
to date.  For example, nonflame data indi-
cates that chlorobenzene would decompose
to 99.99 percent of its original  concentra-
tion in 1 sec at 1038°K (11).  At tempera-
tures more representative of  flames (approx-
imately 2000°K), the time required to
obtain the same destruction level is much
smaller (
-------
following reasons:

     1.   Atomlzation Parameters.  When the
         waste material is injected as a
         liquid which must be atomized,
         poor destruction efficiency can
         result from inappropriate atomiza-
         tion.  Droplets which are too
         large may be produced or their
         trajectory may be such that they
         penetrate the flame zone and igni-
         tion does not occur.

     2.   Mixing. Parameters.  In a turbulent
         diffusion flame the reactants are
         supplied separately and reactant
         contacting takes place via turbu-
         lent mixing.  Poor mixing can re-
         sult in  low destruction efficien-
         cies because the waste material
         may not  be mixed with oxygen before
         it escapes from the flame region.

     3.   Thermal  Parameters.  The destruc-
         tion efficiency may be low because
         flame temperatures are low.  This
         can occur if the calorific value
         of the waste/auxiliary-fuel mix-
         ture is  low, heat removal rates
         are high, or quench rates are high
         due to mixing with excessive excess
         air levels.

     4.  Quenching Parameters.  The react-
         ants can be quenched before destruc-
         tion is  complete by heterogeneous
         or homogeneous phenomena.  Mixing
         can cause quenching as explained
         above, or the flame may contact
         a relatively cool surface.

 Consequently, it  is essential to investi-
 gate the concept  of incinerabilityin flames
 under conditions  which could account for a
 failure to completely destroy the waste com-
 pound and are typical of real systems.

     The primary  goal of the study described
 in this paper was to compare the inciner-
 abil ity ranking procedures which have been
 proposed with those measured under flame
 conditions typical of liquid injection in-
 cinerators.  The  approach utilized was to
 measure the exhaust compound concentration
 under different simulated failure modes and
 to compare the ordering of the  compounds to
 those given by several incinerability rank-
 ing procedures.   Failure conditions were
 simualted for all the parameters expected
 to influence incinerator performance; i.e.,
 atomization, mixing, thermal, and quenching.
To simulate all of them required two reac-
tors.  A microspray reactor consisting of
a laminar premixed flat flame into which
test compounds were injected was used to
investigate thermal parameters.  A subscale
turbulent diffusion spray flame was used
to investigate atomization, mixing and
quenching parameters.   Secondary goals
included the generation of fundamental
flame mode destruction data necessary to
compare flame and nonflame decomposition
and to act as a guide for the types of
experimentation which are needed to estab-
lish an incinerability ranking which could
account for different modes of failure.

EXPERIMENTAL APPROACH

     Extensive investigations are being
carried out at the University of Dayton
Research Institute under EPA sponsorship
to define the kinetics of waste decomposi-
tion in postflame regions (4,6,7).  The
emphasis of the present study was on the
flame zone itself and the impact of fail-
ure conditions associated with mixing,
thermal, quenching, and atomization param-
eters on the relative destruction of five
compounds.  These compounds were selected
because they represented a broad range of
incinerability as defined by existing rank-
ing procedures.  The study was restricted
to conditions typical  of liquid injection
incinerators—no attempt was made to include
phenomena associated with waste destruction
in beds.  Two flame reactors were used to
study destruction efficiency under differ-
ent limiting conditions:

     1.  A microspray reactor allowed
         destruction efficiency to be
         investigated under controlled
         conditions, using well  dispersed
         droplets in a high-temperature
         laminar flow background gas.

     2.  A turbulent flame reactor allowed
         destruction efficiency to be in-
         vestigated under more realistic
         conditions and the conditions
         could be exaggerated to simulate
         different failure modes.

This section briefly describes these reac-
tors, the analytical procedures, and the
test compound selection.  A complete
description is available elsewhere (8).
                                            81

-------
Microspray Reactor

     The microspray reactor was employed  in
this program in order to investigate  single-
droplet reactions without limitations asso-
ciated with turbulent mixing.  Data from
this reactor could be easily interpreted  in
terms of how thermal parameters influenced
compound incinerability in the absence of
atomization, mixing, or quenching param-
eters.

     In the microspray reactor, droplets
are injected through a laminar premixed
hydrocarbon flat flame and react in the
postflame gases whose temperature and com-
position can be controlled.  If the reactor
is to simulate the conditions which prevail
in a liquid spray diffusion flame, it must
satisfy the following criteria (1):

•  The gas-phase temperature must be  in the
   range of typical flame temperatures (up
   to 2000°K) and be well-defined and con-
   trollable.

•  The fuel must be in the form of droplets
   in the size range (20-200 ym) of typical
   commercial atomizer performance.

•  The droplets should be well-dispersed  in
   a gas-phase environment which is similar
   to the recirculated products in the near
   field of a turbulent diffusion flame in
   order to similate the contacting between
   liquid and gas.

The microspray reactor used in this study
is similar to the flame reactors used pre-
viously in studies of the thermal decompo-
sition of pulverized coal particles and
heavy fuel oil droplets (9,14,15) and the
characterization carried out in the pre-
vious studies confirmed that it satisfied
the criteria listed above.

   The microspray reactor used in this
study is shown schematically in Figure 1.
Droplets were injected through the center
of a square flame holder constructed  of
ceramic honeycomb with dimensions 8.9 x
8.9 cm.  The flame reactant mixture,  either
Hp/air/N2 or CH4/air/N2 was distributed to
trie honeycomb through a sintered metal plate
and burned as a thin flat flame which was
supported on the flame holder.  The test
compound droplet stream was injected through
a 1.3 cm diameter opening in the burner
center.  Almost monosized 38 ym droplets
were generated with a Berglund and Liu
vibrating orifice technique and the droplet
stream was dispersed  by  passing the stream
through a second  orifice and coflowing dis-
persion gas  (9).  The composition of the
dispersion gas was  adjusted so that the
stoichiometry including  the droplets, was
by 100 cm long, stainless steel duct.  The
exhaust gases from  the chimney were direc-
ted through  a mixing  baffle tube prior to
a gas sampling port in order to mix the
products uniformly.

     The microspray reactor was used to
investigate  the impact of three control-
lable parameters  on the  destruction effi-
ciency of five hazardous waste compounds:
temperature, stoichiometry, and droplet
composition  (pure compounds versus mix-
tures).  One potential incinerator failure
mode is that the  droplet environment (e.g.,
temperature) may  not  be  sufficient to cause
ignition or  decompose the waste constitu-
ents.  In the microspray reactor, the reac-
tion temperature  was  controlled by varying
the amount of diluent nitrogen temperature
as measured  with  a  type  R (Pt/Pt-30% Rh)
thermocouple on the center line of the duct
are presented in  Figure  2a.   Radial  tem-
perature profiles (Figure 2b)  indicated
that the gas-phase  temperature across the
duct was relatively uniform except near the
edges.  These measurements were made with
the droplet  dispersion flow present but
                                 Monodisperse
                                 Droplet Flame
                                   10-cm-Square
                                  Stainless Steel
                                    Chimney
                                Hydrocarbon
                                Flat Flame
                                Ceramic Honeycomb

                               Porous Stainless
                                   Steel
                              Dispersion Orifice
                            Orifice in
                           ibrating Crystal

                   Purge Drain
 Main     /
 Burner Flow'
Dispersion Flow
             B
      Test Liquid
    Figure  1.   Details  of microspray
                reactor.
                                             82

-------
                                                 150011
                                                 1000
             1600
             1400  -
             1200
             1000
                                                 1500
                                                 1000
                                                 71501
                                  tJlOOf
                                                 £1501
                                                  100(
                          .2    .3    .4    .5
                           Time (sec)
                           I     I     I     I
  0    10    20    30    40   50
        Axial Distance  (cm)

(a)  Gas phase temperatures on center
    line of microspray reactor.   For
    different flame conditions.
                                                 1500
                                                 1000
                                                                     i—r
                                                           *1.3 cm from flame holder


                                                       -I  I  I t  t  I—1—4
                                                                     +



                                              6.4 cm from flame holder


                                          I  I  I  t  I  I—I  t  I  I  (  |—|—\
                                                            11.4 cm from flame holder


                                                      I  I  I  I  I I I  I  I  I  I  I  I  I  I
                                        ^^
                                       -/3    16.5 cm from flame holder
                                                      I  I  t  I  I  I  I  I  I
                                                             I  I  I  I  I
                                                            44.4 cm from flame holder
                                                                   1
                                                                        L
                                                                           1
                                                             Radial Distance (cm)

                                                   (b)  Gas phase temperature as a function
                                                       of axial and radial distance for one
                                                       flame condition.  Axial Distance is
                                                       measured from top of flat flame holder.
              Figure  2.   Gas  temperature measurements in microspray  reactor.
without droplets  present.  When droplets
were injected  they  were confined to the cen-
tral zone of the  duct (maximum diameter,
3 cm) and were therefore all subject  to  the
same thermal history.  The destruction data
was determined as a function of the flame
temperature which was obtained by extrapo-
lating the axial  temperature measurements
to the burner  face.

     Compound  effects were also investigated
for two extremes  in stoichiometry, rich  and
lean.  For rich conditions, the flat  flame
was operated with deficient oxygen (stoichi-
ometry ratio = 0.83)  and the gas used  to
disperse the test compound droplets was iden-
tical to the flat flame feed gas.  Under
these conditions,  the droplets evaporated
and thermally  decomposed with no oxygen
available for  oxidation.  For lean condi-
tions, the flame  stoichiometry was adjusted
to provide 10  percent more oxygen than was
necessary to burn both the flame gas  and  the
test compounds.   In this case, the droplets
were dispersed with air.
                                        When droplets were  injected into a
                                   rich  flame, they were  virtually invisible
                                   to  the naked eye.  However,  when droplets
                                   were  injected into a sufficiently hot
                                   (>850°K) lean flame, the individual drop-
                                   let flames became visible.   Droplets igni-
                                   ted due to heat transfer from the flat
                                   flame gases and flame  radical  attack on the
                                   vaporized material from  the  droplet.  The
                                   ignited droplets appeared as blue streaks
                                   due to the chemiluminescence resulting from
                                   flame reactions around individual droplets.
                                   The visual appearance  of droplet flames was
                                   utilized as an indication that the flat
                                   flame temperature was  sufficient to ignite
                                   the test compound droplets.

                                   Turbulent Flame Reactor  (TFR)

                                        The second reactor  consisted of a tur-
                                   bulent diffusion spray flame contained in a
                                   cold-wall chamber and  is shown schemati-
                                   cally in Figure 3.  The  subscale flame was
                                   designed to simulate many aspects of commer-
                                   cial  practice relevant to liquid injection
                                               83

-------
                       (Sample Port

                     _t-
                     "1   P^-Mixing Chamber
                                        TABLE  1.   NORMAL OPERATING  CONDITIONS
                                             FOR  TURBULENT-FLAME  REACTOR


e


K

[




r-
£$*
i^
f- ''.T
V?, ^i
»^ " Water Inlet


a Water Outlet
^,. 	 ^-Stainless Steel Water-
Cool ed Chamber
— -^^Turbijlent
U» Quartz
	 U i ndow

S^i'Ai^— -Q"31"1
How Control -^-V'^g-'E •"•
. Baffle Llkl 	 Spray
( n.T 	 Swirl Nozzle
Slower Venturi U ^
/" > "•' !y
f (\ \ [i II i !_
V ^ y ra L — -"^_:-_v-v."
Spray Flame


Pressurize
/
^ ^
Fuel
Reservoi r
r"^j


d NO
x- '


CO
       Figure 3.  Turbulent  flow reactor.

 incinerators  including swirl, recirculation,
 a broad droplet  size distribution and a  high
 variation  in  droplet number density.  Since
 the flame was  contained in a cold-wall  cham-
 ber, the flame gases were rapidly cooled.
 Consequently, there  was  little opportunity
 for postflame destruction of the waste  com-
 pounds and the destruction  efficiencies were
 only dependent upon  flame conditions.

        The waste compounds  were mixed  with
 auxiliary fuel (either heptane or diesel
 oil) and atomized using  a commercial  pres-
 sure jet spray nozzle  (Delavan UDA-6QO).
 Air entered through  interchangeable swirl
 vanes placed in the  annular  space around
 the nozzle and the burner throat.   The re-
 sulting swirl-stabilized  flame was contained
 in  a water-cooled stainless  steel  cylindri-
 cal  chamber 30 cm in diameter  and  90 cm
 long.   The lower section  of  the  chamber was
 equipped  with  an ignition port,  observation
 ports,  and a castable  refractory quarl  to
 prevent corner recirculation.   Combustion
 products  from  the chamber were  homogenized
 in  a mixing chamber prior to sampling.

        The  normal  operating  condition  (see
 Table  1)  for the  TFR was selected  based
 upon commercial  practice, visual flame  sta-
 bility, low exhaust  CO (<75 ppm) and total
 hydrocarbon  (<20  ppm) and high destruction
 efficiency  of  hazardous waste compounds.
 Under these conditions, extremely high de-
 struction efficiencies  (>99.995 percent)
 for  all of  the test  compounds were measured
 Several variations from the normaloperat-
 ing  conditions were  investigated, which had
 little  influence  on  the high flame destruc-
tion efficiency of the hazardous waste con-
tunIU/MnS Ihe?e i^luded auxiliary fuel
type (No  2 fuel oil),  burner throat veloc-
ity and test compound concentration (up  to
                                                        NOZZLE
                                                        NOZZLE PRESSURE

                                                        AUXILIARY FUEL

                                                        FUEL FLOW RATE

                                                        AIR FLOW RATE

                                                        BURNER THROAT VELOCITY

                                                        BURNER SWIRL NUMBER

                                                        EXCESS AIR

                                                        BURNER HEAT RELEASE RATE

                                                        TEST COMPOUND CONCEN-
                                                         TRATION

                                                        EXHAUST CONCENTRATIONS
                                           DESTRUCTION EFFICIENCY
                                             OF TEST COMPOUNDS
                                                                         Oelavan Pressure Jet, Hollow
                                                                         Cone, 60°Spray Anqle Model
                                                                         UDA-60-1.S
                    103 psig

                    Heptane

                    1.4 gm/sec (11 Ib/hr)

                    14.3 I/sec (1830 ft3/hr)

                    7.1 m/sec (23.3 ft/sec)

                    1.0

                    301

                    38 kW (131,000 Btu/hr)


                    31 by Mass

                    CO: 75 ppm
                    Total Hydrocarbons: 20 ppm


                    >99.995I
                                     25  percent).   Variations which  did result
                                     in  significant deterioration  in both flame
                                     performance and destruction efficiency and
                                     were used  to  investigate compound effects
                                     under  failure conditions included:
                                     •
                                     •
                                     •
                                     •
                                     t
Low excess  air
High excess  air

High excess  air and low load

Poor atomization quality
Quench coils within flame
                                     Test Compounds

                                        Five compounds,  listed in Table  2,  to-
                                     gether with  some  of their properties were
                                     selected as  representative of liquid organ-
                                     ic hazardous wastes.   All the compounds  are
                                     listed in the 1980  RCRA regulations, Part
                                     261, Appendix VIII  (16).  The compounds
                                     were chosen  to  represent a broad range of
                                     incinerability  based  on the most commonly
                                     proposed ranking  procedures.  They  cover
                                     greater than 90 percent of the range in
                                     heats of combustion for the listed  compounds
                                     (.13 to 10.14 kcal/gm).   Since a direct
                                     comparison between  nonflame thermal decom-
                                     position rankings and the flame mode  destruc-
                                     tion was an objective of this study, com-
                                     pounds were  selected  for testing for which
                                     nonflame data was available.  In addition,
                                     the selection also  took into account the
                                     National  Bureau of  Standards ranking sys-
                                     tem, a range of auto-ignition temperatures
                                     and a variety of molecular structures.   Two
                                     of the compounds are  aromatic, one  is  a
                                     highly chlorinated  methane,  another is a
                                     chlorinated ethane and  one contains nitrogen.
                                             84

-------
                       TABLE 2.  TEST  COMPOUNDS AND  COMPOUND PROPERTIES
COMPOUND
Chloroform
1 ,2-Dichloroethane
Benzene
Acrylonitri le
Chlorobenzene
Boi 1 ing
Point
.(K)
355
356
352
351
405
T99.99
§
290 ©
931
1006
1002
1037
Autoigni tion
Temperature
(K) ©

686
836
754
911
Heat of
Combustion
(kcal/gm) @
0.75
3.0
10.03
7.93
6.6
Heat
Rank
®
18
28
4

3
Evaporation
Time
(msec) ©
9.0
11.9
14.4
20.9
10.8
                     (T)  Temperature required for 99.99% DE at t  1  sec
                     ©
                         Data as reported by Lee
                                            (11)
                                                               (11)
                         Data from Guidance Manual, Mitre Corp.

                         Incinerability Ranking Based on Bond Strength and  Susceptability to
                         Radical  Attack from Tsang,(15)    Rank 1 is most difficult to incinerate.
Calculated time for 100 urn droplet to evaporate at 1000°K

Data after Duvall
                                                                      (8)
Measurement  Techniques

     The  measurement of waste concentration
in the  sample stream for  both reactors in-
volved  the  adsorption from  a  known amount
of exhaust  gas onto Tenax organic sorbent
(8,13).   The exhaust gas  sample was first
passed  through a heated (200  C) particulate
filter, and  then through  a  water-cooled
Tenax cartridge held at 20°C  (see Figure 4a).
The cartridge was on 8 cm long by 1.3 cm
diameter  Pyrex tube packed  with 1 gm of
Tenax.  After adsorption, the cartridge was
placed  in a  specially designed aluminum
block heater immediately  upstream of the GC
column  (Figure 4b).  The  column employed
was a 1-meter-long 0.3 cm OD  Teflon tube
packed  with  acetone-washed  Porapak-Q.  A
temperature  of 150°C was  found to be suffi-
cient to  desorb the test  compounds into the
packed  column.  The GC oven temperature was
subsequently temperature-programmed to
120°C to  separate the compounds.

     The  use of Tenax for concentrating the
sample  and then thermally  desorbing provided
the necessary rapid turnover  of samples
with sufficient separation  and sensitivity.
The breakthrough volumes  of all the test
compounds were directly measured  and were
found to  be  greater than  the  utilized sam-
ple volumes  (8).  Benzene and 1,2-dichloro-
ethane  were  not separable by  the  Porapak-Q
column  and  hence mixtures containing both
                        compounds were  avoided.  Under  the normal
                        operating conditions specified  in  Table 1,
                        the exhaust  concentration of  individual com-
                        pounds was measurable down to 1.5  ppb.   This
                        corresponds  to  a destruction  efficiency in
                        the TFR of greater than 99.995  percent.
                                                        Dry Test Meter
                                      Oven
                                                 Tenax Cartridge
                             (a) Sampling and adsorption.
                        Helium Carrier
                                    GC Oven


                                    Desorption Block
Porapak-Q
 Column
                           A.C.  Power

                                     Variable Transformer

                             (b)  Desorption and analysis.
                                                      Thermocouple Readout
                        Figure 4.  Tenax-GC sampling and  analysis
                        technique for measuring exhaust  concentra-
                        tions of compounds  selected for  testing.
                                              85

-------
MICROSPRAY  RESULTS

   The microspray was  used to investigate
the  impact  of  thermal  parameters for three
conditions:

•  Fuel-lean--oxygen available to oxidize
   test compounds.
•  Fuel-rich—no oxygen  available to oxi-
   dize test compounds.
•  Mixtures of  compounds and pure compounds.

The  other failure mode parameters (atotniza-
tion, quenching, and mixing) cannot be
effectively investigated in the microspray
reactor and were investigated in the turbu-
lent flame  reactor.

     Figure 5  presents data for two mixtures
of four compounds shown  separately in Fig-
ure  5a and  5b.  In  these tests,  38 ym drop-
lets of the two mixtures were injected sep-
arately into a  lean  (10  percent excess
oxygen) H2/Air/N2 flame  with different flat-
flame temperatures.  Exhaust concentrations
of the individual test compounds were mea-
sured and the data  are shown in Figure 5
in terms of the fraction of each compound
remaining versus the measured flat-flame
temperature.  This  temperature is deter-
mined by extrapolating the axial  tempera-
ture measurements to the burner face and
is the highest temperature of the flat-
flame gas.  Under these  excess oxygen con-
ditions, droplet flames  were observed for
both mixtures for flat-flame temperatures
in excess of 850°K.  However,  droplet igni-
tion was observed at slightly lower temper-
atures for one mixture probably due to the
substitution of compounds.   When  the flat-
flame temperature is greater than the igni-
tion temperature of the  specific  compound
mixture, the exhaust concentration of the
test compounds were below the detection
limit of the analytical  technique which
indicated a destruction  level  in  excess  of
99.995 percent.

     Calculations using  nonflame  kinetics
(8) indicate that almost no  decomposition
should occur below 800°K for the  residence
times (M  sec) available in  the  microspray
reactor.  However,  as  shown  in Figure 5,
significant destruction  was  measured at
flat-flame temperatures  below  800°K.   This
destruction at low flat-flame  temperatures
is probably due to a local  increase in tem-
perature around droplets  and flame radical
attack.   Above the visual  droplet ignition
point,  all the compounds  were  destroyed,
                Visual Droplet
                  Ignition
          700      800       900
          Flat-Flame Temperature (K)

 (a) Mixture containing dichloroethane,
    chlorobenzene, chloroform, and
    acrylonitrile

                  Visual Droplet
                      Igni tion
                                1000


                             D Chlorobenzene
                             A Benzene

                             • 1,2-Dichloroethane
                             O Chloroform
                             A Acrylonitrile
         700      800       900 '
          Flat-Flame Temperature (K)

(b) Mixture containing benzene, chlorobenzene,
   chloroform, and acrylonitrile.
                                1000
Figure 5.   Fraction of test compound remain-
ing in exhaust when 38 ym droplets of mix-
tures of compounds  were injected into lean
(10% excess oxygen) H2/air flames as func-
tion of flame temperature.
but below the  ignition temperature the
fraction destroyed  depended upon the com-
pound.  At low flat-flame temperatures for
the mixture containing dichloroethane,
chlorobenzene  had the highest concentra-
tion in the exhaust followed by dichloro-
ethane, chloroform,  and acrylonitrile.  At
flat-flame temperatures just below the
droplet ignition point, again chloroben-
zene was found to be the most difficult
compound to be eliminated but the other
compounds showed some rearrangement in
ranking; but the effect of compound type
is small.  When benzene was substituted
for dichloroethane  (Figure 5b), chloroben-
zene remained  the most prominent compound
in the exhaust followed by benzene, chlo-
roform, and acrylonitrile.  Again, just
below ignition there was some reordering
of compounds with chloroform becoming the
                                            86

-------
easiest to eliminate.  These data indicate
that for single droplet oxidative condi-
tions where the temperature is too low for
droplet ignition, a particular order of
compounds does exist in terms of the frac-
tion remaining in the exhaust.  This order
is chlorobenzene, benzene, 1,2-dichloro-
ethane, chloroform, and acrylonitrile.
However, this order changes as the tempera-
tures reach the ignition point.  The order-
ing just below ignition is identical to the
ordering suggested by Tgg^g and auto-
ignition temperature (see'Table 2).

     When 38 ym droplets of pure compounds
were injected into oxygen-rich lean flame
products, the droplets were observed to
ignite at different temperatures.  For
example, visual ignition for chloroform
droplets was observed at 860°K, while di-
chloroethane ignited at 850°K, acryloni-
trile at 800% and chlorobenzene at 740°K.
Benzene had the lowest ignition temperature
and was observed to ignite at temperatures
below 600°K.  For pure compounds, the de-
struction is controlled by droplet ignition.
The observed ignition temperature does not
agree with any proposed incinerabi1ity
ranking procedures, although the heat of
combustion criteria is almost the same
with the exception that acrylonitrile and
chlorobenzene are reversed.

     The absence of oxygen was the third
failure mode investigated with the micro-
spray reactor.  Droplets of equal molar
mixtures of compounds were injected into
fuel-rich (stoichiometric ratio   0.83)
H2/air/N2 flames of different temperatures.
In  these tests, the oxygen was rapidly and
completely consumed by the hydrogen in the
flat flame so that no oxygen was available
to  oxidize the test compounds.  The frac-
tion of each compound remaining in the
exhaust as a function of the flat-flame
temperature is shown in Figure 6.  Even
with mixtures, the temperature required to
destroy the compounds, 1050°K, was found
to  be very similar to the Tgg 99 tempera-
tures of the individual compounds  (920 to
1037°K; see Table 2) and were much higher
than those required if droplet ignition
occurred (Figure 5).  The fractional de-
struction was strongly dependent upon flame
temperature.  A difference between the com-
pounds was observed only at a temperature
just below the flat-flame temperature
required for complete destruction.  At that
temperature, the compound that was most
predominant was chlorobenzene, followed by
benzene, chloroform, and acrylonitrile.
                      A A
1
- D
A
- A
0
-,
1 do;
*:
Chlorobenzene
aenzene
Acryl oni tn "e
Chloroform
1 i
•" ' 1 1
' A
A
£, , 1
        500       300      :oco      ":oo
                Flame Temperature (<)

Figure 6.   Fraction of test compound remain-
ing in exhaust when 38 ym droplets of mix-
tures of compounds were injected into rich
(stoichiometric ratio = 0.83) Hg/air flame
as a function of flame temperature.  Incin-
erability order at 1050°K is chlorobenzene,
benzene, acrylonitrile, and chloroform.
This ranking was identical to that measured
for the low-temperature oxidation data (Fig-
ure 5).  The nonflame Tgg.gg did identify
the temperature range required for complete
destruction and the most predominant com-
pounds (chlorobenzene and benzene); however,
acrylonitrile and chloroform are reversed
from the Tgg gg ranking.  Pure compounds
were not tested under fuel-rich conditions
although similar results to the mixture
data is expected.

TURBULENT FLAME REACTOR RESULTS

     The turbulent flame reactor was oper-
ated and tested under a number of condi-
tions.  However, many of these conditions
resulted in high destruction efficiency of
all the test compounds.  Only those param-
eters resulting in significant deteriora-
tion of destruction efficiency are presented
here.  Data on high destruction levels are
presented elsewhere (8).  The conditions  in-
vestigated in the turbulent flame reactor
which had a strong influence on destruction
efficiency were primarily associated with
three failure parameters:

•  Atomization parameters — poor atomization
   qua!ity.
•  Quenching parameters — quenching on  cold
   surface.
•  Mixing parameters--!ow excess air
                    --high excess air
                    --low heat-release rate

Those parameters found  to be of  less
                                            87

-------
importance included burner velocity,  fuel
type (No. 2 fuel oil) and concentration  of
hazardous waste compounds (from  3 to 25 per-
cent).

     It was generally found that exhaust
concentration measurements of  carbon  monox-
ide and total hydrocarbons were  good  indi-
cators of flame performance and  compound
destruction efficiency.  The exhaust  CO
level in particular appeared to  be  well
correlated with the exhaust concentration
of the test compounds.  This result was
expected since the high heat removal  rates
in the TFR emphasized flame performance
over postflame reaction.  Since  CO  is an
intermediate in the oxidation  of hydrocar-
bons to C02  (13), it  is directly linked
with combustion efficiency.  Therefore,  an
examination of the relative CO levels for
each failure condition  indicates the  over-
all combustion efficiency which  can be com-
pared to the destruction efficiency of the
hazardous waste compounds.  The  relation-
ship between exhaust  CO, total hydrocarbons
measured by the flame ionization detector,
and destruction efficiency measured for  a
mixture of compounds  is shown  in Figure  7.
The maximum ORE (>99.995 percent) was mea-
sured at 30-40 percent  excess  air,  which
corresponded to the minimum  in both exhaust
CO and hydrocarbon.
   ,-.2000
   3 1500 -
     1000
     500
                                        0.03
                O CO
                  Hydrocarbons as

                Q Test Compound
             125     150    175     200

                Percent Theoretical Air
                                      225
Figure 7.  Exhaust CO  and total hydrocarbons
and fraction of test compound  remaining in
exhaust as a function  of theoretical  air
(constant air velocity, variable load,  equal
molar mixture of chloroform, benzene,  chlo-
robenzene, and acrylonitrile added  3  per-
cent by weight to heptane).
      Figure 8 presents data obtained  with
the  TFR  at  high heat-release rates  (164,000
Btu/hr).  Very high destruction levels
(>99.995  percent) were measured for all com-
pounds at 20 percent excess air at  this
heat-release rate with the exception  of
benzene.  It is possible that benzene was a
product  of  incomplete combustion of either
the  auxiliary fuel  or one of the test com-
pounds (e.g.,  chlorobenzene).   The  actual
source of the benzene whether it is a pro-
duct  of  incomplete combustion or an indica-
tion  of  incomplete benzene destruction, has
not  been  determined.  Benzene is a possible
intermediate in the formation of soot which
was  observed in the flame in the form of
luminosity,  especially at low excess  air
levels.   Because of the relatively  large
amounts  of  heptane present (97 percent)
only  a small  conversion of heptane to ben-
zent  is  required to account for the exhaust
levels of benzene measured at this low
excess air  condition.  However, the benzene
could also  be the result of a simple  trans-
formation of chlorobenzene.
         4000
         3000
         2000 -1
                                                        1000
                125    150   175    200
                  Percent Theoretical Air

                    O Chloroform

                    A Acrylonitrile
                    A Benzene

                    O Chlorobenzene
                                     225
                                                         99
                                                       99.99
                                                      99.999
                                                                   (b)  Destruction and
                                                                      Removal Efficiency
            100
                 125   150    175   200
                  Percent Theoretical Air
                                     225
Figure 8.  Impact of  theoretical air  on  CO
and D_RE from turbulent flame reactor.  Incin-
erabil ity order at 150 percent T.A.  is chloro-
form, acrylonitrile, benzene, and chloroben-
zene  (constant load - 164,000 Btu/hr; vari-
able air flow rate and burner velocity; equal
molar mixture of compounds added 3 percent by
weight to heptane).

-------
     At higher excess  air  levels (>150 per-
cent theoretical air)  the  exhaust concen-
trations of CO and  the test compounds in-
creased.  This is probably due to lower
flame temperatures  and increased quenching
which can occur when  large amounts of un-
heated air are present.  The lowest ORE
level obtained for  these high heat-release
rates (164,000 Btu/hr) was 99.9 percent.
The compound differences were small but
measurable at 150 percent  theoretical air.
Chloroform was the  most predominant test
compound  in the exhaust followed closely by
acrylonitrile and benzene  with chloroben-
zene having the highest destruction effi-
ciency.   This particular order, which was
found to  exist for  a  number of failure con-
ditions tested with the turbulent flame
reactor does  not  agree with any of the pro-
posed rankings  in Table 2, although the
heat of combustion  did identify the most
predominant compound  (chloroform).

     The  data obtained at  low heat-release
rates  (90,000-155,000 Btu/hr) are shown  in
Figure  9.  This data  set was achieved by
lowering  the  fuel  flow rate from the nomi-
nal  operating  conditions  (Table 1) while
maintaining  the  air flow  constant.  This
drop  in load  and  increase  in theoretical air
resulted  in  significant increase in the
fraction  of  the  waste compounds in the ex-
haust.  Under this  failure condition, chlo-
 roform  and benzene  had similar high exhaust
concentrations  followed by 1,2-dichloro-
 ethane  and similar low-exhaust concentra-
 tions  for acrylonitrile and chlorobenzene.

    The  data  presented in  Figure 10 indicate
 that atomization  parameters  had a  signifi-
 cant impact  upon compound destruction.   In
 these  tests,  a nozzle designed for 1.5  gal/
 min was operated at .75 gal/min dropping the
 pressure  from 161  psig  to 40  psig.   This in-
 creases the  mean droplet  size  (5) and affects
 fuel  air  mixing causing some  of the  large
 droplets  to escape the  flame.  The highest
 compound  exhaust concentrations were  mea-
 sured  under these poor atomization conditions.
 However,  the order of compounds was  found to
 be identical  to other failure  conditions for
 the TFR such as high  excess  air  at high loads,
 low excess air at low loads, and quench coils.
 The chloroform was found  to  be  the most pre-
 dominant  compound followed  by  benzene,  acry-
 lonitrile, and chlorobenzene.

    A water-cooled copper  coil  was  placed di-
 rectly within the flame in  the TFR to pro-
 vide an extreme case  of flame quenching in
 order to investigate  destruction  efficiencies
                Firing Rate (103 Btu/hr)
                150    120
                   Mixture containing
                   Benzene, Chlorobenzene
                   Chloroform and
                   Acrylonitrile
                    Chloroform

                    Benzene

                    AcrylonitriVe
                    Chlorobenzene
            .02
Mixture containing
Qichloroethane,
Chlorobenzene, Chloro-
form and Acrylonitrile.

   Chloroform

   Acrylonitrile

   Chlorobenzene

   1,2-Oichloroethane
                       150      200
                  Percent Theoretical Air

Figure 9.  Impact of theoretical airand load
on fraction of test compounds remaining in ex-
haust oftubulent flame  reactor(constant air
velocity, variable load 155,000-90,000 Btu/
hr; equal molar  mixture of compounds added
3 percent by weight to  heptane).
under this mode of  failure.   In  this fail-
ure condition test,  the  coil  acted to cool
the flame and supplied a surface area for
reactants to quench.  The presence of the
quench surface  increased both CO and the
test compound concentration  (see Figure 11).
The order of the  compounds was similar to
other failure conditions with chloroform
being the most  predominant and chlorobenzene
the least predominant  compound in the ex-
haust.  However,  the positions of acryloni-
trile and benzene were  reversed  from the
order found  in  other failure modes.

DISCUSSION

     The combustion of  hydrocarbon fuels  in
turbulent diffusion flames results in rela-
tively high  flame zone  temperatures (between
1600 and 2000°K)  and residence times are  on
the order of 0.1  seconds.  If the waste com-
pounds investigated in  this  study experience
                                               89

-------
        4000


        3000



        2000


        1000
I     I    I
Carbon Monoxide

Off Design
            "Standard"
              Nozzle
          "100      200        300
              Percent Theoretical Air
         0.5



         0.4


         0.3


         0.2



         0.1
(b)  '
Test Compound Data with
•Standard Nozzle and Off ,
Design Nozzle
 "Standard"
   Nozzle
                    ,-  Chlorobenzene
                    -f	I    I
           100      200       300
              Percent Theoretical Air
Figure 10.   Impact  of  atomization qua! ity on
CO and fraction of test compounds remaining in
exhaust of turbulent flame reactor.   Inciner-
ability order is chloroform, benzene, acry-
lonitrile, and chlorobenzene (constant air
velocity, variable load 155,000, 55,000 Btu/
hr; equal molar mixture of compounds added
3 percent by weight to heptane).
                                            5000


                                            4000


                                            3000


                                            2000



                                            1000
                                                                (a)  Carbon Monoxide
                                                                           Data
                                               100       150       200
                                                  Percent Theoretical Air


                                       (b)  Test compound data with and without cooling coil.
.06
.05
.04
j> .03
.02
.01
0
1 IO 1 1
— /Chloroform
- ^
\jnitri 1e
; -




i Uncooled Data
i1^ -4
~ \ Benzene /' ~|
\ ^ £r ^~~-' -Chlorobenzene
SXL-, U-] K ' I
                                              100       150       200
                                                  Percent Theoretical Air
                                    Figure  11.   Impact of cooling coil  placed
                                    in flame  on CO and fraction of test compound
                                    remaining  in  exhaust of turbulent  flame
                                    reactor.  Incinerability order is chloroform,
                                    acrylonitrile, benzene, and chlorobenzene
                                    (constant  air velocity; load= 114,000 Btu/
                                    hr; equal molar mixture of compounds added
                                    3 percent  by  weight to heptane).
these conditions, then  they  would be quanti-
tatively destroyed.  The  results of this
study agree with this hypothesis.  Turbulent
diffusion spray flame and  a  laminar reactor
burning single droplets were capable of de-
struction efficiencies  greater than 99.995
percent.  In the case of  the turbulent flame
reactor under optimized conditions (stable
flame, low CO and total hydrocarbon),  the
compounds were destroyed  mainly in the flame
because postflame decomposition was minimized
due to the fact that the  flame was contained
by cold walls.  Consequently,  it can be con-
cluded that a flame is  an  extremely effi-
cient mode of destroying  waste compounds and
the concept of incinerability  under these
conditions has little value.   If everything
is destroyed, it is not possible to rank com-
pounds in terms of difficulty  or ease  of de-
struction.  Consequently,  a  series of  exper-
iments were designed to assess incinerabil ity
                                    under  several  limiting conditions  which
                                    might  typify the failure mode  of practical
                                    liquid injection incinerators.

                                         The  microspray reactor  investigated
                                    those  conditions associated  with single
                                    droplet combustion in the absence  of com-
                                    plications  due to turbulent  mixing.   It was
                                    selected  in order to study thermal  effects
                                    separated from turbulent mixing  and  atomi-
                                    zation.   The temperature required  to ignite
                                    droplets  of hazardous waste  under  oxygen-
                                    rich conditions in the laminar premixedflat
                                    flame  reactor was found to be  low  (850°K)
                                    in comparison to typical flame temperatures
                                    (1500-2000°K).  Above the ignition tempera-
                                    ture,  the droplets were visually observed
                                    to ignite and the compounds  tested were
                                    quantitatively (>99.995 percent) destroyed.
                                    Even in the absence of oxygen, the micro-
                                    spray  data  indicated that low  temperatures
                                             90

-------
HOOD K) were required  for complete destruc-
tion even without droplet ignition.   These
data were consistent with the  high  destruc-
tion efficiencies achievable  in  a  turbulent
diffusion spray flame environment  of the
TFR.  The TFR was operated  at  high  heat re-
moval rates by operating with  water-cooled
walls in order to minimize  postflame reac-
tions and mixtures up to 25 percent by
weight of the test compounds were  investi-
gated.  Even in the absence of significant
postflame decomposition, destruction effi-
ciencies which corresponded to the  detec-
tion limits of the analytical  systems
(99.995 percent) were achieved for  all  the
compounds tested.  In the turbulent flame
reactor, a direct relationship was observed
between overall combustion  efficiency as
indicated by exhaust CO and hydrocarbon
emissions and the destruction  of the test
compounds.  Conditions which minimized the
CO concentration in the exhaust  gases also
maximized destruction efficiency.   Under
all failure conditions investigated,  exhaust
CO concentration increased  when  the test
compound concentration  increased.   These
results suggest the feasibility  of  using
exhaust CO and potentially  total hydrocar-
bons to monitor the performance  of  liquid
injection incinerators once the  conditions
giving the maximum destruction efficiency
have been defined.

     The incinerability or  ordering of the
compounds was found to depend  on the actual
failure condition which caused the  ineffi-
ciency.  When both the microspray  and the
turbulent flame reactor were operated under
conditions which simulated  failure  modes of
practical incinerators, measurable  differ-
ences in the destruction efficiency of the
five test compounds were obtained.   For
example, chlorobenzene was  the most diffi-
cult to eliminate in the microspray when
the temperature was too low to ignite the
droplets, but was the least difficult to
eliminate for a variety of  failure  condi-
tions in the TFR such as poor  atomization
qua!ity.

     Figure 12 presents a series of bar
graphs which allow a comparison  between in-
cinerability as defined by  the various fail-
ure modes and the rankings  indicated by pro-
cedures based upon Tgg.gg, heat  of combustion,
the NBS method, and AIT.  The  bar graph
shows the concentrations measured  in the
experiment normalized so that  the most  pre-
dominant compound shows full-scale  and  the
lesser concentrations are expressed  as  a
percentage of that maximum  concentration.
               •ION-FLAME
               TEMPERATURE
    INCINERABILITY WRINGS T59.99
    CHLOROFORM

    ! ,2 OICHLOROE7HANE

    3ENZENE

    ACP.YLONITHILC

    QILCROBFJJZE.1E
X8S FLAME  AUTOIGNnlOH
WJKJ.NG   | TEKPE3ATTJRE
    CHLOROFORM
    1 ,1 OICHLOROETKAIIE
    AL'RYLONITTLZ

    -.InLCROBENZETIE
roBBULQIT trLAHE
1,2-OTCHLOm} ETHANE
6ENEZEHE
CHLOR08EHZE.NE
TURBULENT FLAME
CHLOROFORM
1 ,2 aiCHLOROETriANE
EL'IZLNE
HIGH
EXCESS MR
=t
1
HIGH 1
EXCESS AIR |
^
LOU
OCESS AIR
n
3
LOW
E:(C£S3 AIR
P
POOR
ATmtZATTCH
	 1
^^J
No. 2 FUEL OIL
?OOR
ATOMIZATION
P
OUEIICH COIL
D
^_l
,'lo.2 -UEL DIL
ATGMIIATTON
r
Figure 12.  Comparison  of  proposed  ranking
techniques and concentration  measured  in the
experiments under flame  failure  conditions
normalized so most predominant compound  shows
full scale.
This approach gives an  indication  of the
measured magnitude of the  difference in de-
struction efficiency between  compounds.   A
comparison of these relative  concentration
measurements with proposed  incinerability
ranking techniques demonstrates  that none of
the proposed techniques agree with the  data
for all failure conditions.   However,  some
of the ranking procedures  were found to be
appropriate for specific failure conditions.
For example, the nonflame  thermal  destruc-
tion (Tgg^gg) and AIT procedures both agreed
with the compound concentration  measurements
when the temperature was below droplet  igni-
tion temperature and under  oxygen-deficient
conditions.  Heat of combustion  was  found to
correlate the pure compound data when  the
microspray was operated below droplet  igni-
tion temperature.  In most  instances,  chlo-
roform was the most difficult compound  to
incinerate for the failure  conditions  inves-
tigated with the TFR, and  this was intici-
pated by only one of the four ranking  tech-
niques—heat of combustion.

     Although measurable differences in the
destruction efficiency of  the five test com-
pounds were obtained, the  differences  were
not large under any of the  conditions tested.
For the most part, the variation in  the con-
centration (between highest and  lowest) of
                                            91

-------
the compounds in the exhaust was typically
of the order of five,  although variations
larger than ten were measured under some
circumstances.   This suggests that the se-
lection of POHC may not be very critical
because the differences between compounds
are small.  If the permit writer selects
three compounds based  upon two or more rank-
ing techniques, and it is demonstrated that
their ORE is greater than 99.99 percent,
then it is very unlikely that any other
compounds will  be destroyed to a lesser de-
gree.  Nonetheless, to be assured that the
incinerator is  quantitatively destroying
all compounds requires measurement of the
most difficult to destroy compounds under
all potential failure  conditions.

     This study has identified the differ-
ences between compound destruction effici-
ency caused by failure conditions associ-
ated with the flame zone.  High destruction
efficiencies have been demonstrated in the
flame alone.  However, many incinerators
are equipped with postflame hold-up zones
and after-burners in order to achieve addi-
tional thermal  decomposition of compounds
which escape the flame zone.  In order for
an incinerator to fail to destroy a com-
pound, the material must both escape the
flame and the temperature be too low in the
postflame hold-up zone to destroy the com-
pound (less than Tgg gg).  The differences
in the concentration'of compounds in the
exhaust of the incinerators is associated
with both the flame and nonflame zones.
The thermal decomposition which occurs in
the postflame zone can alter the ranking in
the exhaust.  However, this occurs only if
the temperature in the postflame zone is
between the Tgg go of the two compounds.  For
example, the TFR data  indicated that chloro-
form with a Tgg_gg of 930°K is the most lik-
ely compound to'escape the flame and chlo-
robenzene is the least likely with a Tgg gg
of 1038°K.  If the postflame zone tempera-
ture is less than 930°K, then the flame
zone ordering will prevail in the exhaust.
If the temperature is  greater than 1038°K,
then both compounds are quantitatively de-
stroyed.  However, if  the temperature is
between 930 and 1038°K, then chloroform is
destroyed leaving chlorobenzene intact.
Hence the postflame rank will prevail if,
and only if, the temperature in the post-
flame is between the two compounds; in this
case, a 100 C temperature range.  The tem-
perature in the postflame zone is not uni-
form and the temperatures referred to above
are minimum temperatures for a residence
time of one second.
     It was not the purpose of this  study
to ascertain why destruction efficiency
under flame conditions can be compound and
failure mode specific.  More detailed mea-
surements are needed to provide further
information.  It could be associated with
flame inhibition due to the presence of
halogens which are known to reduce burning
rates.   Under quenching conditions,  these
effects could be enhanced.  The formation
of products of incomplete combustion (PICs)
as a consequence of the partial destruction
of the waste compound, was not investigated.
An alternate method of assessing inciner-
ability could be based upon the potential
to form PICs which are themselves hazardous.

CONCLUSIONS

1.  Under optimum conditions, flames are
    capable of destroying hazardous waste
    compounds with very high efficiencies
    (greater than 99.995 percent) without
    the need for long residence time high-
    temperature postflame zones or after-
    burners.

2.  Reduced flame destruction efficiencies
    are the result of operation under some
    failure mode such as poor atomization,
    poor mixing, or flame quenching.

3.  Incinerability, or ordering of compounds
    in  terms of their relative destruction
    efficiency,  is dependent on the actual
    failure condition which caused the
    inefficiency.

4.  Optimum conditions for destruction of
    hazardous waste compounds in turbulent
    diffusion spary flames correspond to
    minimal exhaust CO and total  hydrocar-
    bons.

5.  No  one incinerability ranking system
    appears to predict correctly the rela-
    tive destruction efficiency of the five
    compounds tested for all failure condi-
    tions investigated.  However, several
    rankings did correctly predict relative
    DE  for specific failure conditions.

6.  More data is required on other compounds
    and other failure conditions more appro-
    priate to different types of hazardous
    waste incinerators to fully determine
    the limitations of incinerability rank-
    ing systems and develop an appropriate
    incinerability ranking methodology.
                                            92

-------
ACKNOWLEDGEMENTS

     This study was supported by the U.S.
Environmental Protection Agency through
Prime Contract 68-02-3113 to JRB Associates,
Task 24-1 to Energy and Environmental Re-
search Corporation.  C. C. Lee was the EPA
Project Officer and V.  S. Engleman was the
JRB Project Monitor.  Members of the EPA
Technical Advisory Committee who assisted
in program guidance were:  A. F  Sarofim,
Massachusetts Institute of Technology;
F. W. Marble, California Institute of Tech-
nology; R. M. Fristrom, The Johns Hopkins
University; B. Bellinger, University of
Dayton Research Institute; W. Tsang,
National Bureau of Standards; R. A. Carnes,
Industrial Environmental Research Labora-
tory—Cincinnati EPA; and E. P. Crumplier,
Office of Solid Waste, EPA.

     We wish to acknowledge the contribu-
tions of E. M. Poncelet and H.  D. Crum, who
played important roles in the development
of the analytical approach and  experimental
construction and operation.  Discussions
with J. H. Pohl aided the compound selec-
tion and many other aspects of  the work.

     Although this work was supported by
the  Environmental Protection Agency, the
views contained herein do not necessarily
reflect the views of the agency and no offi-
cial endorsement should be inferred.
 REFERENCES
 1.
Badzioch, S.  1967.
tion in Combustion
Thermal Decomposi-
of Pulverized Coal,
 2.
by N.  A.  Field, D.  W.  Gill, B.B.Morgan
and P. G. W.  Hawksby.   The British Coal
Utilization Research Association, Lea-
therhead, England.

Cudahy,  1981.   Incinerability, Thermal
Oxidation Characteristics and Thermal
Oxidation Stability of RCRA Listed Haz-
ardous Wastes.
 3.   Bellinger,  B.  1982.   Personal  Communi-
     cation.

 4.   Dellinger,  B.,  D.  S.  Duvall,  D.L. Hall,
     and W. A. Rubey,  1982.   Laboratory  De-
     terminations  of High  Temperature  Decom-
     position  Behavior of  Industrial Organic
     Materials.   75th  Annual  Meeting of  the
     APCA.  New  Orleans, LA.
                                               10.
                                               11.
                                           12,
                                           13.
Dietrich, V. E. 1979.  Dropsize Distri-
bution for Various Types of Nozzles.
In Proceedings of the 1st International
Conference on Liquid Atomization and
Spray Systems.  The Fuel Society of
Japan, Tokyo, Japan,  p. 69.

Duvall , D. S. and W. A. Rubey, 1976.
Laboratory Evaluation of High-Tempera-
ture Destruction of Kepone and Related
Pesticides.  Technical Report UDRI-TR-
76-wl, University of Dayton Research
Institute, May 1976.  EPA 600/2-76-299.

Duvall , D. S. and W. A. Rubey, 1977.
Laboratory Evaluation of High-Tempera-
ture Destruction of Poly-chlorinated
Biphenyls and Related Compounds.   EPA
600/2-77-228.

Kramlich, J. C., M. P. Heap, E.  Ponce-
let, G.  S. Samuelson, and W. R.  Seeker,
1983.  Laboratory-Scale Flame Mode Haz-
ardous Waste Thermal Destruction Re-
search.   Final Report Task 24-1, Con-
tract No. EPA 68-03-3113.

Kramlich, J. C,, G. S, Damuelsen, and
W. R. Seeker, 1981.  Carbonaceous Par-
ti cul ate Formation from Synthetic Fuel
Droplets.  Western States Section of
the Combustion Institute.  Fall  Meet-
ing, Tempe, Arizona.  WSS/CI-81-52.

Lee, K.  C., J. L.  Hansen, and D.  C.
Macauley, 1979.  Predictive Model of
the Time/Temperature Requirements for
Thermal  Destruction of Dilute Organic
Vapors,   72nd Annual Meeting of the
APCA.  Cincinnati, OH, 6/79.

Lee, K.  C., N. Morgan, J. L. Hansen,
and G. M. Whipple, 1982.  Revised Model
for the Prediction of the Time-Temper-
ature Requirements for Thermal Destruc-
tion of Dilute Organic Vapors and its
Usage for Predicting Compound Destruc-
tability.  75th Annual Meeting of the
APCA, New Orleans, June 1982.

Parson, J. S. and S. Mitzner, 1975.
Gas Chromatographic Method for Concen-
tration and Analysis of Industrial
Organic Pollutants in Environmental
Air and Stacks.  Env. Sci. Tech., 9_,
p. 1053.

Seeker, W. R. , M. P. Heap and T. J.
Tyson, 1981.  Gas Phase Chemistry.
Volume I of  Final Report  for EPA 68-02
-2631.
                                            93

-------
14.   Seeker,  W.  R.  and M.  P.  Heap, 1982.
     Flame Combustion Processes.   Volume
     II  of Final  Report for Contract EPA 68-

15.   Seeker,  W.  R., G. S.  Samuelsen, M.  P.
     Heap, and J.  D.  Trolinger,  1981,   The
     Thermal  Decomposition of Pulverized
     Coal  Particles.   The  18th Symposium
     (International)  on Combustion.   The
     Combustion  Institute, Pittsburgh,  PA,
     p.  1213.

16.   Tsang, W.  and  W.  Shaub,  1981.  Chemi-
     cal  Processes  in  the  Incineration  of
     Hazardous  Waste.   National  Bureau  of
     Standards.   Paper presented  to  Ameri-
     can  Chemical Society  Symposium  on
     Detoxification of Hazardous  Wastes,
     New  York,  August  1981.

17.   U.S.  EPA,  1980.   Hazardous  Waste Con-
     solidated  Permit  Regulations.   Federal
     Register,  pp.  33132-33233,  Monday,
     May  19.
                                           94

-------
              THE PACKAGED THERMAL  REACTOR SYSTEM:
                   DEVELOPMENT AND  APPLICATION

                         Wayne A. Rubey
                         John L. Graham
                         Barry Dellinger
                      University of Dayton
                       Research Institute
                  Environmental Sciences  Group
                       Dayton, OH 45469

                        Richard A.  Carnes
              U.S. Environmental Protection Agency
                Combustion Research Facility/NCTR
                       Jefferson, AR 72079
                            ABSTRACT

     In view of the urgent need  for thermal  decomposition data to
guide the safe incineration of hazardous  organic materials and
industrial wastes, the University  of  Dayton's  Environmental
Sciences Group has developed a simple  Packaged Thermal Reactor
System  (PTRS).   The PTRS can rapidly,  easily,  and safely deter-
mine the relative thermal stability for organic substances under
a wide variety of temperatures,  residence times,  and gaseous
atmospheres.  In addition, the PTRS is quite sensitive,  as the
level of 99.99% destruction can  often  be  measured directly.  Also
the PTRS is small and quite compact,  making  it easy to use in
almost any installation.

     In this paper the design, development,  and application
aspects of the prototype PTRS are  discussed.
INTRODUCTION

     As controlled  high-
temperature incineration plays
an increasing role in the per-
manent disposal of hazardous
organic wastes there is a grow-
ing need for basic thermal
decomposition data concerning
these materials.  Closely
related to this is  the avail-
ability of  suitable analytical
instrumentation which can  pro-
vide these data.

     Since  1969,   the
Environmental Sciences Group  of
the University of   Dayton
Research Institute  has been
actively involved in the design
                                95

-------
and development of thermal
instrumentation systems.  This
paper will discuss the various
aspects of the latest in a ser-
ies of thermal instrumentation
systems designed and developed
at the University.  This new
system is referred to as the
Packaged Thermal Reactor System
(PTRS) .

BACKGROUND

     The early thermal decomposi-
tion research activities at the
University of Dayton were aimed
at developing special instru-
mentation to study the thermal
decomposition properties of
organic polymeric materials.
In 1974, the emphasis shifted
to developing special thermal
systems which could address the
thermal decomposition behavior
of industrial organic wastes,
toxic Organic materials, and
other environmentally sensitive
substances.  The first such
system was the Discontinuous
Thermal System (DTS).  This was
a relatively simple instrumen-
tation assembly consisting of
a heatable sample inlet region,
a high-temperature quartz tubu-
lar reactor, an effluent trap
for collection of unreacted
parent compounds along with
various reaction products, and
a detached programmed-
temperature gas chromatograph.
The DTS provided valuable
information with respect to the
thermal decomposition of vari-
ous pesticides, e.g., Kepone,
Mirex, and DDT. [1]

     From an instrumentation
standpoint, this particular sys-
tem was important for another
reason.  Specifically, it laid
the groundwork for more sophis-
ticated studies by identifying
those variables which  are  of con-
cern to laboratory-scale  thermal
decomposition studies.  These
variables are listed in Table 1.
Also, while conducting experiments
using the DTS the  importance of
the formation of products  of
incomplete combustion  (PICs)  was
clearly revealed.[1]   This impor-
tant finding led to the develop-
ment of the Thermal Decomposition
Analytical System  (TDAS).
  \-
     In principle, the TDAS is
similar to the DTS in  that it
consists of a thermally program-
mable inlet region that is con-
nected to a quartz tubular reac-
tor with a cryogenic in-line trap
for capture of the effluent
products from the  reactor.   In
reality, however,  the  TDAS is a
far more sophisticated system and
has evolved to the point  that it
bears little resemblance  to its
predecessor.  The  TDAS, unlike
the DTS, is a truly continuous
system.  Connected downstream
of the cryogenic trap  is  an LKB
2091 GC-MS which is used  for
analyzing the effluent products.
This unit also has a dedicated
minicomputer which processes the
data and aids in the interpreta-
tion of the mass spectra.

     With its GC-MS and dedicated
minicomputer, the  TDAS proved
to be a powerful tool  for the
investigation of gas phase ther-
mal decomposition  behavior.
Over the past four years,  this
unit has been used to  study the
thermal decomposition  of  numerous
pesticides, industrial organic
wastes, and pure compounds.  The
TDAS is a very sophisticated
system and by its  nature  it is
also a relatively  expensive
system to assemble and maintain.
With that in mind, we  set out to
design an instrument with
                                96

-------
            TABLE 1

           GAS  PHASE
THERMAL DECOMPOSITION VARIABLES
      EXPOSURE TEMPERATURE

         RESIDENCE TIME

   COMPOSITION OF ATMOSPHERE

    TEMPERATURE VARIATIONS

       CHAMBER PRESSURE

   RESIDENCE TIME DISTRIBUTION
              97

-------
capabilities similar to the TDAS
but which incorporated recent
technological advances at a much
reduced cost.  This led to the
development of the Thermal
Decomposition Unit-Gas
Chromatographic  (TDU-GC)
system[3] .

     As with the TDAS, the TDU-
GC incorporated several improve-
ments over its predecessor.  In
particular, the TDU-GC's sample
inlet region was designed to be
more versatile with respect to
the different types of samples
which can be analyzed.  The most
significant change was replac-
ing the LBK 2091 with a Varian
Vista series programmed tempera-
ture high-resolution gas chro-
matograph.   The versatility and
operating characteristics of
this particular GC, particularly
its ability to operate at cryo-
genic temperatures, expanded the
analytical  capabilities of the
system significantly.

     Since  it became operational
in late 1981, the TDU-GC has
shown that it can perform many
of the same tasks as the TDAS.
The major difference between the
two instrumentation assemblies
is that while the TDU-GC can
detect the  presence of PICs, to
actually identify them would
require further analysis.
However, a  TDU-GC system may be
assembled for about one-third
the cost of a TDAS.

     As work with the TDAS and
the TDU-GC  systems progressed,
researchers became aware of the
enormous demand for data con-
cerning the thermal stability of
both pure compounds and complex
industrial  organic wastes mix-
tures.  Events and discussions
at a recent hazardous waste
incineration  conference  [4]
clearly pointed  out  the  urgent
need to determine  thermal
decomposition behavior  for a
very large number  of toxic
organic compounds.   Eventually,
hundreds of different organic
compounds may need to be tested.
Accordingly, analytical  instru-
mentation is needed  which  can
rapidly screen the thermal
decomposition properties of
various principal  organic  haz-
ardous constituents  (POHCs)
prior to subjecting  them to  con-
trolled high-temperature incin-
eration.  In view  of the urgency
of this situation, a design  was
conceptualized for a relatively
simple instrument  called the
Packaged Thermal Reactor System.

Concept for the  Packaged Thermal
  Reactor System

     The goals for the PTRS  were
to design a system which could
rapidly and safely determine the
destruction efficiency  (DE)  for
an organic material  at a given
residence time,  exposure tem-
perature, and flowing gaseous
atmosphere.  This  instrument
should also be small, compact,
and relatively portable so that
it can be used as part of  a
mobile field unit  or as part of
a permanent installation.
Finally, it should be an inex-
pensive unit to  assemble and
maintain.  The concept of  such
an instrument is presented in
block diagram form in Figure 1
and is shown schematically in
Figure 2.

     The schematic shown in
Figure 2, describes  one version
of the originally  proposed
PTRS.[5]  The various compon-
ents within the  dashed-line
boundary represent the packaged
                                98

-------
UD
/ /\ /
SELECTIVE
DETECTION
OF EFFLUENT
PRODUCTS
s—
-
)
CONTROLLED
HIGH
TEMPERATURE
EXPOSURE

} 	
-
/"
/
UNIFLOW
DISPERSION
OF
SAMPLE

yi / /
/ 	
-
)
SAMPLE
INSERTION
VAPORIZATION
OR
DEGRADATION /
        Figure  1.   Block Diagram  of the Packaged Thermal  Reactor System.

-------
PURE COMPONENTS
OR WASTE SAMPLES

   TYPICALLY
l.0fj.l INJECTIONS
                   HEATED
                   UNIFLOW
                 DISPERSION
                  CHAMBERS
                             SPECIAL
                             PURPOSE
                             REACTORS
COOLANT
  u
                                          ELECTROMETER
                                          AMPLIFIER
                                          ELECTRONIC
                                          INTEGRATOR
                                                    PACKAGED
                                                    REACTOR
                                                    ASSEMBLY
                                                      VENT
GASEOUS
CARRIER:
 e.g., AIR
      N
        0
                                         POTENTIOMETRIC
                                           RECORDER
     Figure  2.   Conceptual  Schematic  of the Packaged Thermal
                 Reactor  System.
                                100

-------
reactor assembly-  The compon-
ents outside the boundary are
typically found in almost every
chemical instrumentation labora-
tory.  An analyst who has been
conducting gas chromatographic
analyses could readily operate
this unit.  In addition, once
this unit has been installed in
a laboratory A the thermal decom-
position data can be obtained
very quickly.  A single deter-
mination should take less than
ten minutes; and an extremely
small sample can be used to
minimize safety problems.

     In Figure 2, it is seen
that localized cooling is an
option that can be used for
condensing and concentrating
organic substances that have
passed through the high-
temperature reactors.  The same
localized cold regions of the
transport path can also be
heated rapidly to release the
trapped organic substances and
thereby pass them on downstream
to the exit location as narrow
concentrated zones.

     The design of the packaged
reactor will also permit the
use of interrupted flow techni-
ques.  Such an arrangement per-
mits the use of atmospheres
with differing oxygen concen-
trations.  Also, with this
reactor assembly, tests can be
made with almost any type of
flowing gas, e.g., air, oxygen,
humid atmospheres, hydrogen, or
even low levels of chlroine.

     Before going further into
the details of the uses for the
PTRS, it would be desirable to
discuss the characteristics of
one of the major components of
this system which is often
overlooked, namely the hydrogen
flame ionization detector  (HFID).
The Hydrogen Flame lonization
  Detector

     One of the most useful
detectors that can be used with
a packaged thermal reactor sys-
tem is the HFID.  This particular
device came into existence in
1958 as a sensitive detector for
sensing extremely low levels of
hydrocarbon.   It was apparently
conceived and studied simultane-
ously by several researchers
[6,7].  Essentially, this
detector has undergone only minor
revision since its introduction,
and these refinements have been
with respect to enhanced elec-
tronics with accompanying signal
handling techniques.

     Today the HFID is probably
the most readily available gas
chromatographic detector in
laboratories throughout the
world.  It is also the major
component in total hydrocarbon
analyzers.  The HFID has excel-
lent sensitivity in that it can
respond to 10~12 grams per
second of an organic substance.
This detector can be used for
sensing a broad range of concen-
trations as it has a linear
dynamic range of between 10^ and
10 "7 (depending upon design and
associated electronics).  For
organic compounds, the HFID is a
destruction device, as all gas
phase species must pass through
a flame that operates at approxi-
mately 2,200°C.  This is a bene-
ficial characteristic with
respect to the safety of person-
nel using this device with hazard-
ous organic materials.

     The hydrogen flame ioniza-
tion detector responds to all
compounds containing CH groups,
but fails to respond to a num-
ber of common inorganic com-
pounds.  The chemicals in the
                                101

-------
following list exhibit essen-
tially no response to the HFID.

rare gases
hydrogen
oxygen
nitrogen
oxides of carbon
oxides of nitrogen
ammonia
water
silicon tetrachloride
silicon tetrafluoride
trichlorosilane
hydrogen sulfide
hydrogen chloride
sulfur dioxide
carbonyl sulfide
carbon disulfide

     A very interesting observa-
tion can be made concerning the
response to hydrocarbon and
halocarbon compounds.  However,
it will not respond to the final
combustion products of hydro-
carbons', i.e., H20, CC>2, HCl,
etc.  This is an important
observation with respect to
incineration studies and the
associated evaluation of the
completeness of thermal decom-
position or combustion.

     If one admits a pure com-
ponent sample to the transport
channel* of the packaged thermal
reactor and senses the total
effluent with an HFID and its
associated electronics, one can
obtain total response counts for
the unreacted sample.  Then, if
the same quantity of sample is
subjected to another channel set
for higher temperature exposure,
the response of the surviving
*Channel A of Figure 2 would
have adjustable transport tem-
peratures of between 200°C and
500°C.
organic species can be  measured.
Thus, the extent of decomposi-
tion, combustion, etc.,  for a
given high-temperature  exposure
can be readily evaluated.   In
short, by using a packaged
thermal reactor system,  the cal-
culation for DE  (in percent) can
be obtained as follows:
DE = 100
(1)
Where R represents the  inte-
grated response from the  respec-
tive reactor channel  (see
Figure 2).  This is a very  con-
venient method for determining
the relative thermal decomposi-
tion behavior of POHCs.

     There are several  addi-
tional features of the  HFID
which make it especially  suited
as the primary detection  device
for a PTRS.  The detector is
not affected to any great
extent by subtle changes  in flow
rate or other physical  para-
meters which often have profound
effects on other types  of gas
chromatographic detection
devices.  Also, the response
from the detector is practically
independent of the simultaneous
elution of non-responding spe-
cies with compounds that  do
respond to the HFID.  For
example, the detection  of low
levels of organic compounds in
the presence of massive quanti-
ties of C02 has little  effect
upon the resultant output signal
from this device.  In short, the
presence of H20, HCl, and large
concentrations of other non-
responding substances will  not
adversely affect the detection
of low-level concentrations of
organic materials.
                                102

-------
     The HFID is also ideally   as X2  in  Figure  2,  is a special
suited for conducting thermal   high-temperature channel that
decomposition tests in pyroly-  differs  greatly  from the other
tic atmospheres such as nitro-  channels  in  that the reactor
gen,  helium,  argon, and so      itself is  of a dramatically dif-
forth.   This detector can per-  ferent geometry.   The reactors
mit studies using low levels of in Channels  A and Xi are narrow-
water in the carrier and can    bore  quartz  tubes that have been
also function very nicely with  formed into  helical coils.   These
carbon dioxide or hydrogen as   narrow-bore  gas  flowpaths pro-
the carrier gas [8,9].  Now let duce  residence time distribu-
us return to the discussion of  tions  which  are  narrow and
the PTRS and its application.   symmetrical.   However,  Channel X2
                                generates  a  broad and highly
Application of the PTRS         skewed residence time
                                distribution.
     Figure 2 shows that the
PTRS reactor unit is in fact          The  configuration of the
made up of three sub-units,     PTRS  as  shown in Figure 2 is
each of which is a miniature    intended  primarily for experi-
thermal reactor system in its   mentally  determining the DE of
own right.  The first channel,  organic  materials.   From the
referred to as Channel A in     basic  drawing of the packaged
Figure 2, is a low to moderate- thermal  reactor  system (Figure
temperature reactor and this    2), it is  evident that the  DE can
channel is maintained at        be readily obtained by merely
approximately 300°C to obtain   comparing the integrated
quantitative transport of the   responses  from Channel A and
sample to the detector.  Since  Channel  X]_ [See  Equation (1) ,
no thermal degradation of the   page  10].  This  calculation
sample occurs in this channel,  would apply  for  the thermal
it is used to obtain the total  testing  of pure  organic sub-
integrated response for an      stances  as well  as complex mix-
unreacted inserted sample.      tures  of  hydrocarbons.   However,
This, of course, corresponds to when  a multi-family chemical
a DE of 0%.                     sample is  to be  subjected to
                                thermal  decomposition studies,
     The second channel, desi-  e.g.,  a  waste stream sample that
gnated as X]_ in Figure 2, is a  contains  both hydrocarbons and
conventional narrow-bore high-  chlorinated  hydrocarbons, a dif-
temperature quartz tubular      ferent detection  mode could be
reactor.  Physically this       used,  as shown in Figure  3.   In
channel is identical to Channel this situation,  a parallel
A except in this case the tern-  arrangement  using the HFID  and
perature of the reactor may be  the Hall Electrolytic Conductivity
as high as 1,000°C.  By compar- Detector (HECD)  can provide the
ing the total integrated        same  type of information, i.e.,
response obtained from this     the  determination of destruction
channel to that of Channel A,   efficiency.   The HECD is basic-
the DE may be easily calculated, ally  an  elemental detection
                                device that  responds selectively
     The third channel, noted   to the chlorine, nitrogen,  and

                               103

-------
 ©
 ©
                                              ELECTROMETER
                                               AMPLIFIER
                                               ELECTRONIC
                                               INTEGRATOR
                                              POTENTIOMETRIC
                                                RECORDER
Figure 3.  Conceptual Schematic of the Packaged Thermal
           Reactor System Equipped with Parallel Detectors
                                104

-------
sulfur generated during the
course of these thermal expos-
ures.  The procedure under
which these particular analyses
would be conducted is slightly
different in that cold trapping
is used for these determinations.

     Another modification of
this PTRS is shown in Figure 4.
In this version there is no
detector in the system as it is
now a preparative unit.  Very
simply, a collection trap has
been placed at the exit of the
assembly, and the samples that
have passed through the selected
reactors can now be trapped for
further examination and analy-
sis using other types of
analytical techniques, such as
liquid chromatography, high-
resolution gas chromatography,
and gas chromatography-mass
spectrometry.  Since highly
toxic decomposition products can
be generated during certain high-
temperature exposures a protec-
tion trap has been included in
this preparative unit.

     The availability of a pre-
parative system, such as shown
in Figure 4, has considerable
value in that sizable quantities
of special thermally prepared
samples could be readily
obtained.  This capability would
permit sophisticated analyses of
important collected products,
i.e., products of incomplete
combustion.  In addition, such
a preparative thermal unit could
be used to generate samples
which could eventually be tested
for special toxicological pro-
perties, e.g., mutagenic
activity, etc.

     The Prototype PTRS

     The primary objective
associate^ with designing  and
fabricating a "prototype"  PTRS
was to evaluate the validity of
the basic concept.  More speci-
fically, it was necessary  to
determine if it was possible
(and practical) to construct
a reasonably compact multi-
channel assembly  for measuring
the DE of organic materials
under a wide range of test
conditions.

     The PTRS as  shown in
Figure 5 is a two compartment
instrument consisting of a
reactor assembly  unit and  a con-
trol unit.  The control unit
contains all of the gas flow
controllers, temperature con-
trollers, and temperature  moni-
toring equipment.  The reactor
assembly houses all of the
reactors, dispersion chambers,
the detector, and all of their
associated furnaces and heaters.
A schematic of the interior of
the reactor assembly unit  is
shown in Figure 6.

    The operation of the PTRS
may best be explained by fol-
lowing through each step out-
lined in Figure 1 while
referring to Figures 5 and 6.
The first step is to select and
adjust the appropriate carrier
gas.  The PTRS was designed to
operate with a 50 psig source
of any non-corrosive gaseous
atmosphere.  The  carrier enters
a distribution manifold within
the control unit, which supplies
gas to all six of the flow con-
trol valves  (see  Figure 6).
The control valves used in the
PTRS operate by maintaining a
fixed pressure drop across a
laminar  flow element.  By  using
different restrictor elements,
these controllers may be oper-
ated over a very  large range  of
                                105

-------
 ©
 ©
PROTECTION
   TRAP
Figure 4.  Conceptual Schematic  of the Packaged Thermal
           Reactor System  Equipped with a Preparative
           Unit.
                                106

-------
Figure 5.  Photograph of the Prototype Packaged Thermal
           Reactor System.  Note:  The Reactor Assembly
           is to the Left (front) and the Controller
           Assembly is to the Right (rear).

-------
o
oo
                                                       LOW TEMPERATURE
                                                      /REACTOR (LTR)
                                                       DISPERSION
                                                       CHAMBER  (DC-II
                                                                              HIGH TEMPERATURE
                                                                              REACTOR (HTR)
DISPERSION
CHAMBER (DC-2)
                                                                       HTR
                                                                                                        SIMULATION
                                                                                                        REACTOR (SR)
DISPERSION
CHAMBER (DC-3)
                              Figure  6.   Schematic of  the Reactor  Assembly of  the  Prototype
                                           Packaged Thermal Reactor  System.   Note:   The
                                           Inverted Triangles  Denote Carrier Gas Inlet
                                           Points.

-------
gas flow rates.

     Figure 6 shows that there
are two carrier gas lines sup-
plying each channel, one purges
the dispersion chamber, the
other is used to supply auxili-
ary make-up carrier.  By con-
trolling the ratio of the flow
rates within these two lines the
operator can control the inlet
sample concentration profile
to the reactor without affect-
ing the total flow of carrier.
By using extremely small dis-
persion chamber purge rates the
inlet profile to the reactor
will be very dilute and may be
hundreds of seconds wide.
Since the mean residence time
in the reactor is typically 2.0
seconds, this results in a
concentration profile within
the reactor which is nearly
uniform and at all times is
highly dilute.  This arrangement
gives the operator complete con-
trol over the conditions under
which the experiment is to be
performed.

     After the carrier gas has
been adjusted so the total flow
to each reactor will give the
desired residence time, the
operator must set the various
heaters to the desired tempera-
ture.  In all there are ten
heaters to be set and 21 thermo-
couples are used to monitor the
conditions within the reactor
assembly.  The dispersion cham-
bers, transport lines, Channel
A reactor, and HFID heaters
may be heated from ambient to
400°C.  In normal operation
these are set between 200 °C and
300°C.  The cryogenic traps may
be controlled from -130°C  (by
purging with chilled nitrogen
gas)  to +300°C.  Under normal
conditions these are  also  set
between 200°C and  300°C.   The
reactors of Channels  X-|_  and X2
are normally operated from
300°C to 1,000°C.

     Now that the  level  of dis-
persion, the residence times,
and the various exposure tem-
peratures have been set, the
instrument is ready for  the
ignition of the HFID  and the
conducting of a thermal  decom-
position test.  Since  the  PTRS
is fitted with standard  GC
septa, normal syringe  injection
techniques are used.   Sample
sizes vary from 0.1 to 5. Dpi
for liquid samples, and  from
1.0 to lOOyl for gas  phase
samples.  Upon injection the
sample is deposited at the
entrance to the dispersion
chamber.  These chambers have
been packed with fine  mesh
Pyrex glass beads.  (48%  void
volume.)  This assures complete
mixing of the sample  with  the
gaseous carrier.

     As the sample is  swept
from the dispersion chamber,  it
is augmented by the make-up
carrier and transported  to the
reactor.  To reduce the  resi-
dence time within  this trans-
port region the tubing bore is
reduced from 1.0 mm I.D, to
0.5 mm I.D.  The two  narrow-
bore tubular reactors  are
constructed of 2.0 mm I.D.
fused quartz tubing.   To give
a reasonable path  length and
still have small over-all
physical dimensions,  these two
reactors are configured  in the
form of helical coils.   The
actual path length of each is
about 40 cm while  the  coils
are only 8 cm in length  and
3 cm diameter.  The Channel X2
                                109

-------
reactor is in the form of a
1.0 cm I.D. sphere  with the
entrance and exit tubes located
tangential to the sphere and on
opposite sides.  The exit of
each reactor is once again of
the small 0.5 mm I.D. quartz
tubing.  Through the combina-
tion of the small bore inlet
and outlet tubes the small
physical size of the reactors,
and the use of small (10 cm)
ceramic heaters, the sample is
exposed to nearly square-wave
thermal pulse, the duration of
which may be accurately calcu-
lated and easily controlled.

     The effluent products from
the reactor are carried down-
stream where they are either
cryogenically trapped  (this
may be necessary to concentrate
the effluents when extremely
dilute inlet streams are used)
or transported directly to the
HFID.  When the effluent trap-
ping mode is selected,  the trap
is maintained at cryogenic tem-
peratures until all of the
condensable effluent has been
captured.  The trap may then be
heated at 300°C/min. to release
the effluent products and they
are  quickly  passed into the
detector.  After obtaining the
total integrated responses from
Channel A "and the respective
high-temperature channel, the
DE can then be readily
calculated.

     Figure 7 shows several
examples of thermal decomposi-
tion curves generated using the
PTRS.  All of these data were
taken with a mean residence
time of 2.0 seconds in air.
Note that in all cases the
thermal decomposition was fol-
lowed very near to the 99.99%
level of destruction.
Summary and Future Work

     The prototype PTRS  unit has
demonstrated that a  small  thermal
decomposition system can be  built.
Thus far, research results  indi-
cate that the PTRS will  meet or
surpass all goals with respect
to its performance.   The specific
features of this prototype unit
are summarized below.

•  The PTRS is small  and com-
   pact.  The entire  unit  is
   smaller than some  gas
   chromatographs.

•  The PTRS is easy  to use.   A
   technician with GC experi-
   ence can operate  the  unit.

•  The PTRS is fast.  Experience
   has shown that a  single deter-
   mination may be taken in  as
   little as four minutes.

•  The PTRS is versatile.   The
   unit can analyze  almost any
   material which responds to
   the HFID.  Also,  the  PTRS
   can operate at temperatures
   up to 1,000°C and  with  any
   non-corrosive atmosphere.

•  The PTRS is sensitive.   In
   most cases the 99.99% DE
   may be measured directly.

•  The PTRS is safe.  Typically
   .25yl gas samples  are used.

     Further work on  the present
PTRS unit will explore its capa-
bilities and limitations.   After
a complete evaluation of the
prototype PTRS  (in each  of its
various modes) it will be  used to
obtain thermal decomposition
profiles for organic  substances
which are of environmental
concern.
                               no

-------
                Ob
               90
            o

               99
            u_
            LJ
            O

            h-
            O

            cc
              99.9
            LU
            Q
             9999
            99.999
                                       ~!    I    I    T
   A ANILINE

   Q NITROBENZENE


   El TETRACHLOROETHYLENE


|=  Tr  20 sec, Air
                    400      500     600     700     800

                                EXPOSURE TEMPERATURE,°C -
                                         900
                                                                  1000
Figure  7-   Three  Examples  of  Typical  Packaged Thermal Reactor

             System Data.
                                     Ill

-------
     Future activities will
include the development of a
second generation PTRS that will
emphasize compactness, porta-
bility, and versatility with
respect to measuring DEs.

Credits

     This work was performed
under the sponsorship of the
US-EPA through Cooperative
Agreement CR-807815-01-0

Acknowledgements

     We gratefully acknowledge
the technical assistance of
Ira B. Fiscus and his col-
leagues in the Design and
Development Group.  We are
specially indebted to Richard
A. Grant for his fabrication
suggestions and his expert
scientific glassblowing.
REFERENCES

1.  Duvall, D. S. and Rubey,
    W. A., Laboratory Evalua-
    tion of High-Temperature
    Destruction of Kepone and
    Related Pesticides, Report
    for U. S. Environmental
    Protection Agency, EPA-600/
    2-76-299, December, 1976.

2.  Rubey, W. A., Design Con-
    siderations for a Thermal
    Decomposition Analytical
    System (TDAS),  Report for
    U. S. Environmental
    Protection Agency, EPA-600/
    2-80-098, August, 1980.

3.  Dellinger, B.,  Duvall, D. S.,
    Hall, D.  L., Rubey, W. A.,
    and Carnes, R.  A.,
    Laboratory Determination of
    High-Temperature  Decomposi-
    tion Behavior of  Industrial
Organic Materials, Paper
presented at the  75th Annual
Meeting of Air Pollution
Control Association, New
Orleans, June, 1982.

Hazardous Waste Incineration
Conference, Jointly Sponrored
by American Society of
Mechanical Engineers and
U.S. Environmental Protection
Agency, Williamsburg, VA,
May, 1981.

Rubey, W. A., A Packaged
Thermal Reactor System  for
Characterizing Thermal
Stability of Organic
Substances, University  of
Dayton Research Institute
Technical Memorandum, UDR-
TM-81-30, August, 1981.

Harley, J., Nel, W., and
Pretorius, V., Nature,
181:177, 1958.

Mcwilliam, I.G., and Dewar,
R. A., Nature, 181:760, 1959.

Sevcik, J., Detectors in
Gas Chromatography.
Flsevier Scientific,
Amsterdam, 1976.

Schupp, 0. E., Gas
Chromatography, Inter-
science, New York, 1968.
                                112

-------
                             INCINERABILITY CHARACTERISTICS OF
                             SELECTED CHLORINATED HYDROCARBONS
                              Ninth Annual Research Symposium
                              Land Disposal,  Incineration and
                               Treatment of Hazardous Waste
                                      David L.  Miller
                                       Vic A.  Cundy
                                     Richard A.  Matula

                           Department of Mechanical Engineering
                              Hazardous Waste Research Center
                                Louisiana State University
                                  Baton Rouge,  Louisiana
Abstract

     This paper presents an overview of the fimdamentals of liquid fuel incineration and
the results of shock tube experiments investigating the oxidation of pure chlorinated
hydrocarbons.   In these experiments the ignition behavior of chlorinated methanes,  ethanes
and benzene were compared with the parent hydrocarbons.  The results indicate that  it is
no more difficult to ignite the chlorinated hydrocarbons which were studied than their
analogous hydrocarbons.  This appears to be in contradiction with practical experience in
incinerators.   Preliminary results with spectroscopic measurements of infrared emission
from carbon monoxide (CO) and carbon dioxide (CO,.,) indicate that the oxidation of CO to
CO^ is inhibited during the combustion of chlorinated hydrocarbons.  In the future, these
spectroscopic  studies will be continued and combined with results from flat flame burner
experiments.
INTRODUCTION

     Incineration in a properly designed
and operated facility has been recommended
as a preferred control technology for
combustible organic hazardous wastes [22].
In principle,  a completely efficient
incinerator converts organic hazardous
waste material via high temperature ther-
mal oxidation to carbon dioxide (C0«),
water (H20), and low volume inert ash
material.   During the incineration of
halogenated hazardous waste streams,
significant quantities of the halogen
acids will be  formed and, if the stream
contains dissolved minerals, hazardous
salts may be included in particulate
emissions.  In order to optimally design
incinerators which perform according to
current United States Environmental Pro-
tection Agency (USEPA) regulations, it is
necessary to understand the thermal de-
struction behavior of organic hazardous
wastes.  Developing this knowledge is
complicated by the fact that there are
several different chemical environments
present in an incinerator.  In each of
these regions the rate of conversion of
organic wastes into products is controlled
by different physical and chemical
mechanisms.  An appreciation of these
differences is also essential for the
proper design and modeling of an inciner-
ator.
                                           113

-------
     This paper presents an overview of
the fundamentals of liquid fuel inciner-
ation and the results of shock tube exper-
iments investigating the oxidation of
several pure chlorinated hydrocarbons.  In
these experiments the ignition behavior of
chlorinated methanes, ethanes and benxene
were compared with the parent hydrocar-
bons .
HAZARDOUS WASTE INCINERATION

Overview

     A large proportion of organic hazard-
ous wastes occur as liquid streams.  The
incineration of liquid wastes is an ex-
tremely complex process with numerous
physical and chemical phenomena involved.
The present state-of-the-art does not
allow complete characterization of an
incineration system nor the a priori
prediction of the destruction and removal
efficiency (DRE) for a given set of con-
ditions.

     A schematic of the subprocesses
occurring in a typical turbulent flame,
such as that which may be encountered in
liquid injection incineration, is shown in
Figure 1.  The diagram is representative
of phenomena occurring during  the  combus-
tion of either hazardous wastes  or conven-
tional fuels.  A number of phenomena
including carry over of liquid droplets,
vapor carry over due to low temperature,
vapor carry over due to unmixedness,
inadequate residence time within the  flame
zone, etc., all may contribute to  unsatis-
factory system performance.

     Generally, the combustion processes
in an incinerator may be considered to
occur in the following three regimes:

     o    Regime 1: Combustion of  fuel and
                    oxidizer that  have
                    been mixed on  a
                    molecular  scale.  When
                    these reactants are
                    brought to a tempera-
                    ture above the igni-
                    tion temperature,
                    flame reactions occur
                    and complete and  in-
                    tense combustion  is
                    achieved.

     o    Regime 2:  Thermal reactions of
                    fuel and/or fuel
          LIQUID INJECTION
                                      HEAT TRANSFER
              OXIDATION REACTIONS
              HW + 02
              HW + HW
              HW + FLAME RADICALS
              HW + SOOT
          Figure 1.   Diagram of the phenomena occuring in a turbulent flame.
                                           114

-------
                    fragments  in  the
                    absence  of/or with  a
                    limited  quantity  of
                    oxidizer.   This com-
                    bustion  regime occurs
                    within the central
                    core  of  the flame in a
                    liquid hazardous  waste
                    burner.  In this
                    regime the fuel is
                    vaporized,  pyrolized
                    and may  result in the
                    formation  of  soot and
                    other high molecular
                    weight aromatic hydro-
                    carbons  or chlorocar-
                    bons.

     o    Regime  3:  Reactions  of  fuel
                    and/or fuel fragments
                    in the presence of
                    oxidizer at a tempera-
                    ture  too low  to cause
                    ignition.   Complex
                    processes  and reac-
                    tions may  occur when
                    some  of  the liquid
                    fuel  is  mixed with  hot
                    oxidizer and  combus-
                    tion  products where
                    the temperature is
                    insufficient  to cause
                    ignition.

     It is reasonable  to  expect high
destruction efficiencies  of  hazardous
waste compounds that are  processed by a
flame front.   However,  waste materials
which escape the  flame front and  are
processed in Regimes 2 or 3  are likely  not
to be adequately  destroyed unless the
residence times at  high temperatures  are
substantial.

     An understanding  of  the relative time
scales associated with the incineration of
a pure liquid hazardous waste  may be
obtained by considering the  following
sequential processes to occur:

     o    Insertion of liquid  droplets
          through a liquid injector into
          the incinerator

     o    Heat transfer between hot gases
          and the liquid  droplet  until  the
          droplet reaches its  boiling
          temperature

     o    Vaporization of the  liquid
          droplet
          Mixing of the waste vapor with
          hot oxidizer

          Chemical reaction of waste/oxi-
          dizer to destroy the waste
     In an incinerator designed to destroy
liquid wastes a liquid spray nozzle is
generally employed to promote vaporization
of the waste by finely atomizing the
liquid.  This process increases the ex-
posed surface area of the waste by forming
droplets having an average size generally
in the range of 20-400 micrometer (pm) in
diameter.

     An understanding of the drop size
distribution of organic waste sprays is
essential in evaluating the effectiveness
of an incinerator system.  In particular,
it is critical to know the fraction of the
total spray that is contained in large
diameter droplets.  This is due to the
fact that the time required to vaporize a
droplet is dependent on the square of its
diameter.  Hence, a drop with a diameter
three times greater than the average
droplet diameter requires approximately a
factor of 10 longer to vaporize than the
time required for the average droplet to
vaporize.

     Generally, liquid spray nozzle manu-
facturers do not have detailed drop size
distribution data.  In many cases the only
data available is a specification of the
Sauter Mean Diameter (d^ ) as given below:
                       Sffl
           In.d.
     1   = 1 L x
      sm   ,-   , 2
           In.d.
           .  i i
(1)
where:    n. = number of drops in the
spray with diameter d. ((Jm).
Unfortunately, d   does not provide infor-
mation on the distribution of droplet
diameters in the spray.  Various drop size
distribution equations have been
developed, but none have been universally
accepted  [12, 21].

     The  importance of droplet diameter
distribution in a liquid spray can be
readily illustrated by considering a
simplified hypothetical spray having a
                                           115

-------
total of 1000 droplets with only three
diameters as shown in Table 1.  With this
hypothetical spray distribution, a Sauter
Mean Diameter of 251 )Jm is calculated.
However, the determination of the mass
distribution among the three droplet
diameter groups is of significance to the
incineration process.  The mass (m.) of a
droplet with diameter (d.) is given by:
     m. =
                                   (2)
where:    p = the fluid density.
The corresponding mass fraction (MF.) of
the spray contained in each droplet dia-
meter group can be calculated from:
                                                 TABLE 1.  WASTE MASS DISTRIBUTION  IN
                                                    A HYPOTHETICAL SPRAY POPULATION
                                                           OF 1000 DROPLETS

Initial Droplet
Diameter, d
(Mm) °

100
500
750

Number of
Drops, n.

900
6
4
Initial Mass
m.n.
M_ 11
T> — V1
F zm.n.
i i
0.29
0.22
0.49

           m.n.
     MF. -
       i
            n.n.
             i i
                                   (3)
The time to complete droplet evaporation
(t ) is given by
Referring to Table 1, 99% of the droplets
in this hypothetical spray have a diameter
of 100 pm.  These droplets, however,
represent less than 30% of the total
liquid mass associated with the spray.   It
should also be noted that the large dia-
meter droplets represent only 0.4 percent
of the number of droplets but almost  50
percent of the waste mass.
Vaporization

     The evaporation and/or combustion of
single droplets has received considerable
attention.  A comprehensive model of these
phenomena has not emerged; however, for a
wide variety of systems it has been shown
that the instantaneous diameter (d) of an
evaporating fuel droplet is given by the
relation:
      p    p
     dz = d/ - X t
           o    v
                                   (4)
where,    d  = initial diameter (milli-
               meters, mm)

           t = evaporation time (s)
          A.,, = eva
               (mm^/s).
           -v   	goration rate constant
                                                         r
                                                          v
                                                                                   (5)
                                               A number of factors including:
                                                         ambient temperature
                                                         liquid physicl properties
                                                         local gas velocity
                                                         presence of dissolved and sus-
                                                         pended solids
affect the numerical value of A.  .  For a
wide variety of liquids the numerical
value of A.  in air with a temperature in
the range ?f 1300 K ranges between 0.25 to
2.5 mm /s. [16].

     The evaporation times for various
droplet diameters as a function  of the
evaporation rate constant are summarized
in Table 2.  Inspection of these results
shows that the time required for evapora-
tion may range from 1 x 10   s to approxi-
mately 4 s.  These results clearly indi-
cate that evaporation of large liquid
droplets may be an important consideration
in determining the destruction efficiency
of liquid wastes in incinerators.  For
example, if the group of droplets having
the size distribution of Table 1 is ex-
posed for 0.5 s to an incineratorpenviron-
ment where A.  is equal to 0.25 mm /s all
of the 100 (jm droplets will vaporize.
                                           116

-------
TABLE 2.  EVAPORATION TIME AS A FUNCTION
          OF INITIAL DROPLET DIAMETER
          THE EVAPORATION RATE CONSTANT
                                  the larger, less numerous droplets which
                                  might be produced during atomization.


Initial

Diameter, d
(|jm)
50
100
500
750
1000
Evaporation Rate
Constant, A
' V
(mm /s)
0.25 2.5
0.01s 0.001s
0.04s 0.004s
1.0s O.ls
2.25s 0.225s
4.0s 0.4s

                                               Chemical Kinetics

                                                    After the liquid wastes have evapo-
                                               rated and mixed with the oxidizer, they
                                               must then be destroyed by chemical reac-
                                               tions.  Chemical kinetic considerations
                                               can be used to determine the rate of
                                               chemical reaction.  Consider the following
                                               elementary chemical reaction:
                                                    aA + bB -> cC + dD
                                                                      (6)
                                               where     A, B, C, D represent chemical
                                                         compounds, and

                                                         a, b, c, d the stoichiometric
                                                         coefficients.
However, applying Equations 2 and 3, the
500 urn and 750 |jm droplets will be reduced
to 350 [Jm and 660 pm, respectively.
Assuming that the 100 pm diameter family
of droplets has vaporized, one can
compute the non-vaporized mass fraction of
the original spray from the following
equation:
where   MF..  = mass of family i remaining
               at time t divided by total
               intial mass
                                  It has been shown that the rate of reac-
                                  tion of Compound A can be represented by:

                                              = -k(T)[A]a[B]
                                   (7)
                                  where:    [A] = concentration of compound A
                                                 (moles/cc)

                                          k(T) - chemical rate constant for
                                                 reaction (6) at temperature
                                                 T

                                             t = time.

                                  The rate  constant, k(T), for a given
                                  reaction  is conveniently represented in
                                  terms of  the Arrhenius equation:
         n.  = number of non-vaporized
               droplets in family i at
               time t
                                       k(T) = A exp(-E /RT)
                                                      8
                                   (8)
         mit
= mass of non-vaporized
  droplets in family i at
  time t (see eqn. 4)
As shown in Table 3, after 0.5 s in the
incinerator environment, 99% of the
initial droplets in the spray have been
vaporized while 41% of the initial mass
remains in the liquid state.   From this
example it is clear that an incinerator
must be designed with consideration for
where:     A = frequency factor

          E  = activation energy
               (cal/gm-mol)

           R = universal gas constant
               (1.98 cal/gm-mole-K)

           T = temperature (K).

The rate equations represented by Equa-
tions (7) and (8) indicate that the
                                           117

-------
               TABLE 3.  MASS DISTRIBUTION IN A HYPOTHETICAL  SPRAY  AFTER 0.5 s
                         RESIDENCE TIME IN AN INCINERATOR ENVIRONMENT-



Initial
Droplet
Diameter, d
(Mm) °

100
500
750

Initial
Number
of
Drops ,
n.
i

990
6
4


Droplet Diameter
After 0.5 s
Residence
Period, d (|Jm)

0
354
661

Drops
Remaining
After 0.5 s
Residence
Period, n.

0
6
4
Mass Remaining
After 0.5 s
Residence
Period,
n. . m.
ti " 2n.m.
i i
0
0.07
0.34

           *(\v = 0.25 nun /s)
destruction of a vaporized compound in an
incinerator is a function of residence
time, species concentrations, and temper-
ature.  Since the rate constant, k(T), is
exponentially dependent on temperature,
the rate of chemical reactions are
strongly dependent on the temperature of
the incinerator.

     In any incinerator system, a host of
elementary chemical reactions are occur-
ring simultaneously.  It is beyond the
scope of this paper to discuss the de-
tailed chemical reactions occurring in an
incinerator.   However, it is possible to
obtain an overview of the time scale of
reactions in an incinerator by considering
the rate of thermal destruction of a
compound (C)  which is highly diluted in a
constant temperature environment.  The
rate of reaction of compound C may be
approximated  by the following pseudo first
order kinetic expression:
                                                    C -*• products
= -k[C]
      dt
                                    (9)
where:    [C] = concentration of the
               compound at time t

           k = rate constant at the
               reaction temperature

Integrating Equation 9 yields:
     ln([C]/[C]Q) = -kt
                         (10)
where:  [C]  - the initial concentration
               of compound A at time
               t = 0 s.

Combining Equations (8) and (10) and
simplifying yields:
                                                    [C]/[C]  = expt-At exp(-E /RT)]
                                                           O                 a
                                           118

-------
Ueiimng the destruction efficiency, DE
     DE = 1   [C]/[C],
        = 1   exp[-At exp(-E /RT)]
                            3
yields
(I   DE) = exp[-At exp(-E /RT)]    (11)
                         3
Taking the natural logarithm of both sides
of Equation (11) and simplifying yields:
     t = exp(Ea/RT) [- ln(1A  DE)] ,
leading to
ln(t) = E /RT + ln['ln(1.' DE)]    (12)
         a              A
The use of Equation (12) to determine the
time/temperature relationships for a
specific first order chemical reaction may
best be demonstrated by another example.
In order to effect a 99.99% chemical
destruction efficiency, we must set DE in
Equation (12) to 0.9999.  In many in-
stances the Arrhenius frequency factor (A)
for first order decomposition reactions is
of the order of 1014 s"1  [2].  In
Figure 2, the log.,, t required to destroy
a compound to 99.99°/<, DE is shown versus
1/T for an A of 1014 s  {  and a series of
overall activation energies.  Inspection
of these results indicates that the time
required to destroy hazardous waste com-
pounds with a DE of 99.99% is strongly
dependent on the incinerator temperature
and the Arrhenius parameters of the pseudo
first order rate constant.  Typical para-
meters for the thermal destruction of a
number of dilute organic  vapors in air are
given in Reference 24.

     With these simplifications the time
scales for the incineration of liquid
hazardous wastes may be approximated.  The
waste stream is fed to an injector which
atomizes the waste into small droplets
with a range of sizes that are inserted
into an incinerator environment; pressure
slightly sub-atmospheric, temperature
~1200K.  Heat transfer from the incine-
rator environment raises  the temperature
                                          TEMPERATURE.
                                     20OO  1700    i-JOO
              Figure 2.  The logarithm of the  time  required  for  99.99%  DRE
                         of a coupound via a first  order  decomposition
                         reaction versus temperature.
                                           119

-------
of the droplet until it reaches its boil-
ing point where it begins to vaporize.

     Consider a droplet with an initial
diameter of 251 Mra> tne Sauter Mean dia-
meter of the hypothetical spray of
Table 1.   The minimum and maximum times
for complete vaporization, calculated from
Equation 5, are listed in Table 4.

     Once in the vapor phase the waste
must mix with the oxidizer for combustion
to occur.  Assuming pseudo first order
kinetics, the chemical time required to
achieve a DE = 99.99% can be calculated
from Equation 12.  Table 4 presents these
calculations for hazardous wastes within
an incinerator operating at 1200K for
various values of activation energy, Ea.
The time shows a strong dependence on E
ranging from 0.007 s for E  = 60 kcal/mol
to 115,000 s for E  = 100 tcal/mol.  While
the vaporization time of this example is
small compared to the chemical kinetic
time, droplets with diameters > 251 pm
which could contain a majority of the mass
in the spray will take significantly
longer to vaporize.

     This simple example calculation
illustrates the importance of basic re-
search on the incineration process, from
atomization to chemical kinetics.  There
are approximations for the time scales of
varporization and chemical kinetics but
these depend on information, droplet size
distribution and activation energies, not
presently available.  Approximations of
the time scales required for heat transfer
and mixing do not exist.  Faced with these
difficulties the incinerator community has
developed measures to rank the relative
destructibility of hazardous wastes.
Incinerability Ranking Criteria

     In order to incinerate organic haz-
ardous wastes a permit from the USEPA must
be obtained.  The permit writer examines
the components of the waste stream and
selects one or more compounds, termed the
principal organic hazardous constituents
(POHC's).  A trial burn of the waste or
data in lieu of a trial burn must demon-
strate a 99.99 percent destruction and
removal efficiency (DRE) of each POHC
before a permit is granted.  The USEPA
guidance manual to permit writers suggests
that hazardous waste compounds in the
greatest concentration in the waste
stream and waste components which  are  the
most difficult to destroy be designated  as
POHC's.

     In order to select the appropriate
POHC's,  it is necessary to develop  an
incinerability ranking scale for hazardous
waste compounds.  Presently, the USEPA is
suggesting the use of the heat of  combus-
tion (AH ) as the measure of a compound's
incineraSility in its guidance manual  to
permit writers [25].

     The choice of AH  as the incinera-
bility ranking parameter has generated
controversy in the technical community.
Numerous alternative approaches have been
suggested to determine the relative incia-
erability ranking of compounds.  These
are:

     o    Chemical kinetic considerations
          [24],

     o    Autoignition temperature  [7],

     o    Thermal oxidation and thermal
               decomposition under non-
               flame conditions[8, 9],

     o    Linear regression models based
               upon autoignition tempera-
               ture and structural con-
               sideration [18] and

     o    Toxicity.

     The relative strengths (+) and short-
comings (-) of various incinerability
ranking parameters are summarized  in
Table 5 [10].  Research presently being
conducted in this area may ultimately  lead
to a more effective methodology for deter-
mining the relative incinerability of
hazardous waste compounds.
EXPERIMENTAL STUDIES
Introduction

     In order to develop  information
concerning the high temperature  combustion
kinetics of organic hazardous waste com-
pounds, the reaction of selected pure
chlorinated hydrocarbons  (CHC) and oxygen
mixtures when exposed  to  high temperatures
in a shock tube has been  investigated.   In
the operation of a shock  tube, a one-
                                           120

-------
                               TABLE 4.   INCINERATION  TIME  SCALES
             Complete Vaporization

             d    251 pm (dsm of hypothetical
                  spray)
                                   Pseudo  First  Order Kinetics

                                        DE =  99.99%, A    1014 s'1

                                        T  = 1200K  (1700F)
             t(  s
                       A .
                                                              t, s
                                                                          E ,  kcal/mol
            0.26

            0.026
0.25

2.5
      0.007

     27.6

115,000
 60

 80

100
                   TABLE 5.  SUMMARY OF STRENGTHS (+) AND SHORTCOMINGS (-) OF
                             PROPOSED  1NC1NEKABTUTY RANKING PARAMETERS.*
          Parameter
                                                 Strengths
                                                               and Shorii-omings
Heat of Combustion
     Enthalpy change  from
     reactants to products
Autoignition Temperature
     Temperature at which
     compound ignites
     spontaneously in air

Chemical Kinetics
     Reaction path used as
     measure of incinerability
Thermal Decomposition
     Experimental, nonflame
     determination of
     destruction efficiency

Multiple Linear Regression
     Relates physical and
     chemical properties to
     decomposition

Other Methods
            +  Ddta available or can be calculated
            +  Relates  to heat  release ami  tempera lure  rise
              Some apparent inconsistencies  in ranking
              Correlation with other methods

            +  Correlates with  thermal decomposition data
            +  Extensive data available
            -  Value varies with experiment
              Some compounds do not autoignite

            +  Considers destruction as a rate process
            +  Considers both unimolecular  and bimolecular processes
              Does not consider physical processes
              Limited  kinetic  data available

            +  Simple experimental system
            +  Can get  99-99 percent destruction efficiency directly
              Does not consider flame reactions
              Validity for incineration unknown

            +  Considers both AI and structure
            +  Several  variables included
              Many adjustable  parameters
              Coefficients obtained from thermal decomposition

            *  Heat of  formation
            -'  Gibbs free energy
            *  lonization potential
            *  Flash point
            ""  Combustion ignition delay
            *  Thermal  decomposition in a flame environment
^Source:  Reference [10]
                                             121

-------
dimensional shock wave is caused to pro-
pagate within a tube filled with a poten-
tially reactive gas sample.  This shock
wave compresses the gas sample, thereby,
heating the sample to a temperature high
enough to initiate reactions.  The pro-
gress of the reaction can be monitored in
real time by several optical techniques
and/or dynamic pressure measurements.   In
a slightly modified configuration, a gas
sample can be heated for a short period of
time, then quenched and analyzed later by
gas chromotographic or other techniques.

     The principle advantages of the shock
tube technique are three fold [11, 19].

     1)   Any gas phase compound can be
          studied; including pure hazard-
          ous waste compounds, with or
          without oxidizer.

     2)   The passage of the shock wave
          through the experimental gas is
          equivalent to moving the gas
          from a reactor at room tempera-
          ture to another reactor at a
          specified high temperature in a
          time on the order of 10
          seconds.

     3)   The typical reaction time in a
          shock tube (~2 milliseconds) is
          far shorter than the time for
          chemical species to diffuse to
          the walls and hence wall effects
          are negligible.

     In the initial study, the ignition
delay times of selected pure CHC's and
oxygen mixtures were measured in the
conventional shock tube facility of the
Combustion Laboratory of Louisiana State
University.  Such studies have been useful
in determining the combustion character-
istics of a wide variety of hydrocarbon
fuels [17, 5, 26, 3].  Computer simula-
tions with hydrocarbon oxidation
mechanisms proposed in the literature [27]
indicate that at the time of ignition the
fuel concentration is over 99 percent
destroyed.  Ignition delay times therefore
seemed promising as a measure of compound
destruction.

     Ignition delay time is defined as the
interval between the initial exposure of
the CHC oxidizer mixture to a step func-
tion change in temperature and the
occurence of the principal exothermicity
of the reaction, which  is  signified  by a
sudden increase in temperature  and pres-
sure.  The duration of  this  ignition delay
time is determined by the  overall kinetics
of the combustion reactions.

     In order to compare the  ignition
delay times and thereby the  combustion
mechanisms of selected  CHC's  with their
hydrocarbon analogs, stoichiometric  fuel-
oxygen mixtures, diluted with argon  (Ar) ,
using (1) methane and its  chlorinated
derivatives, (2) ethane, 1 , 1 , 1-trichloro-
ethane, and 1,2-dichloroethane,  (3)  ethene
and trichloroethene, and (4)  benzene  and
monochlorobenzene were  shock  heated  over
the temperature ranges  listed in Table 6.

     When studying hazardous  wastes which
contain halogen compounds  as  well as
hydrogen and carbon, it is necessary  to
define a methodology for determining  the
stoichiometric oxygen or air  required for
complete combustion of  the compound.
Incineration temperatures  are generally
high enough to favor the conversion of
almost all of the H and Cl in the haz-
ardous waste to hydrochloric  acid (HC1).
This indicates that when determining  the
stoichiometric oxygen requirements for the
chlorinated hydrocarbons it should be
assumed that maximum conversion of avail-
able H and Cl to HC1 occurs.   Any remain-
ing Cl is assumed to form molecular chlo-
rine (C12) and any remaining H is assumed
to form water (H^O) .  Based on this
reasoning the stoichiometric  reaction
equation for methyl chloride  (CILC1)  with
CL is given by
CH3C1 + 1,5
                       HC1 + H20.   (13)
This definition of stoichiometry was used
to determine the test mixture compositions
for all the compounds in Table 6.

     A useful correlation equation for
ignition delay time is of the form [20,
6]:
T = ATB exp (E/RT)  [fuel]3[02]b[Ar]c
where,
                                    (14)
           T = ignition delay time  ((Jsec)
                                           122

-------
           T  = temperature (K)

           R  = gas  constant (cal/gmmole-K)
                                      o
          []  = concentration (moles/cm )

 A,B,E,a,b,c  - empirical constants.

This equation provides a means  of compar-
ing shock tube data,  but it is  not
directly applicable to real combustion
systems since the CL  and Ar dependencies
must be replaced with an air dependence.
A correlation equation of this  nature is
also derivable from a reaction  mechanism
[14].

     Since many experimental data are
required to determine the six empirical
constants in Eq. 14 for each fuel of
interest, a simpler comparison  scheme for
ignition delay data was employed in this
study [4].   Initial test gas compositions
were carefully chosen to allow  a direct
comparison of the apparent activation
energies of the fuels studied without
having to explicitly determine  the concen-
tration dependencies a, b, and  c.  In
these experiments stoichiometric fuel-
oxygen mixtures, diluted in argon, with
approximately equal carbon atom concen-
trations were studied.  The pressure
behind the reflected shock wave was held
nearly constant at 2.0 atm for  all experi-
ments.  Within these limitations, the
measured ignition delay times and the
apparent activation energies for the fuels
studied may be compared directly.


Experimental Facility

     A 76.2 mm diameter stainless steel
shock tube was employed (Figure 3) in this
study.  The ignition delay time was char-
acterized as the interval from  the arrival
of the shock at the end wall of the tube
until the sudden rise in pressure due to
the onset of the principle reaction
exothermicity as measured by a  PCB 113A
piezoelectric pressure transducer.  The
output of the transducer was displayed on
a Nicolet Explorer III digital  oscillo-
scope and recorded.  Ignition delay times
could be read with an accuracy  of 5-10%.

     Incident shock speeds were measured
using piezoelectric pressure transducers
to trigger the start and stop channels of
interval timers.  Transit times between
transducer stations were measured to
within ± Ips and the velocity linearly
extrapolated to the end wall.  The initial
temperature and pressure behind the re-
flected shock wave were computed according
to the standard conservation equations
assuming no reaction behind the incident
shock.  Thermodynamic data for enthalpy
and specific heat required for the calcu-
lations were taken from the JANAF tables
[13] and the API Project 44 tables [1] and
extrapolated to the 1500-2500 K range.

     The liquid chlorocarbons were puri-
fied by bulb-to-bulb distillations dis-
carding the first fraction which may
contain low-boiling point impurities and
the last fraction which may contain high-
boiling point impurities.  The vapor
pressures of the purified liquids were
used to prepare the mixtures.  The
methane, methyl chloride, ethane and
ethylene were Matheson CP grade gases and
the Ar and 0,., Matheson pre-purified grade
gases.  The test mixtures were prepared by
following standard manometric procedures.
Because of the low exothermicity from the
oxidation of the chlorinated hydrocarbons
it was necessary to use relatively large
mole fractions of fuel and oxidizer in
order to measure ignition from the pres-
sure traces.  The mixtures were stored in
stainless steel tanks at approximately 1.1
atm and allowed to mix for a mimumum of 36
hours before use.
Results

     The natural logarithms of the meas-
ured ignition delay times vs. the recipro-
cal of the initial experimental tempera-
ture for the GI and both the €„ and C,
fuels are plotted in Figs. 4 and 5,
respectively.  Where possible the data
were fit by least squares analysis to an
expression of the form
     In t = InA + (E /RT)
                    3
(15)
where
      T = ignition delay time  (|Js)

      R = universal gas constant
          (cal/gmmole-K)

      T = initial experimental temperature
          (K)
                                           123

-------
           TABLE 6.  TEST CONDITIONS, FUEL-OXIDIZER MIXTURES, AND LEAST SQUARE PARAMETERS.

10%
10%
.0%
10%
10%
5%
5%
5%
55
')%
Text Mixtures
CH4 + 20% 02 + Ar
CHjCl + 15% 02 + Ar
CHjCl + 10% 02 + Ar
CHC13 + 10% 02 + Ar
CCl^ + 10% 02 + Ar
C,H, + 17.5% 0, + Ar
26 2
1,2-C2H4C12 + 12.5% 02 + Ar
1,1,1-C,H?C1, + 10% 02 + Ar
C,HA + 1S% 0, + Ar
C.HC1 + 10% 0, t Ar
P (a tin)
2.0
2.0
2.0
2. 1
2. 1
1.7
1.8
2 3
1.6
I . 7
T Range (K)
1330-1544
1280-1570
1220-1398


1155-1287
1188-1540
1288-1618
1103-1307

In A ( sec) E/R x 10~3(K)
-5.5 ± 1.4 17.6 ± 2.1
-3.0 ± 0.7 14.5 + 3.0
-9.0 ± 2.0 19.7 ± 2.9


-7.8 + 3.7 17.4 ± 4.4
-b.O + 0.4 15.0 ± 0.5
-0.02 +1.2 9.5+1.7
-10.4+3.2 19.0+3.6

     Wolloce ond Tiernon
     FA 145 Monomeler
Vent
        Edwards ED 500
        Roughing Pump
 O  Argon or Nitrogen

-*— Hydrogen
                                              Onygen
                              Televac Thermocouple Guoge   Edl
-------
  A,  E    empirical constants calculated
          by least squares.

The least squares lines are also shown in
the figures.  The reflected shock pres-
sure, the apparent activation energies
divided by R (Ea/R in Eq. 15), fuel/
oxidizer compositions, and temperature
range for each set of data are shown in
Table 6.
    Examination of Fig.  4 indicates that
H,  and CH.-.C1 have similar ignition delay
imes and that dichloromethane (CH0C1 ')  is
CH
times
                                  2
more easily ignited.  Data for chloroform
(CHC1-) and carbon tetrachloride (CC1,)
exhibit considerable scatter, but they
tend to cluster between the results for
CH, and CH0C10.
  4       22
     Inspection of Fig. 5 indicates that
ethylene (C?H,) is more readily ignited
than any of tne C~ fuels studied.  The
data for trichloroethene (C^HCl,,) are too
scattered to obtain a good least squares
parameters.  However, they tend to scatter
around the C^H, results.  The ignition
delays for 1,1,1-trichloroethane (CpH-Cl,,)
are longer than those for ethane (C^H^),
but the ignition delays of 1,2-dichloro-
ethane (C?H,C1«) are similar to those for
CpHx.  The measured ignition delays for
benzene (C,.H,) and chlorobenzene (C£HtCl)
        ,0.0..-,               03
are nearly identical.
Discussion

     The present results seem to indicate
that it is no more difficult to ignite the
chlorinated hydrocarbons which were
studied than their analogous hydrocarbons.
This appears to be in contradiction with
practical experience in incinerators where
it has been reported that chlorinated
hydrocarbons are difficult to destroy.

     A possible explanation for this
difference is that there is a large dif-
ference in the "strength" of the ignition
between hydrocarbons and progressively
more substituted chlorinated hydrocarbons.
Once an ignition occurs in the hydrocar-
bons the reactions are exothermic enough
to raise the temperature and completely
destroy the fuel; whereas, in contrast the
less exothermic CHC ignition may not
provide sufficient temperature rise to
produce the destruction.
     Actually, a  comparison of the C-H
and C-C1 bond dissociation energies, 415
and 280 kJ/mol, respectively, indicates it
is reasonable to expect that the chlo-
rinated compounds will more readily decom-
pose to produce the radical pool required
to initiate ignition.  This may partially
explain the relative ignition delay times
of the compounds studied.  In addition,
recent results obtained near room tempera-
ture [15], indicate that the rate of
hydroxyl radical (OH) attack on CH^Cl,, was
faster than OH attack on other chlorinated
methanes or on CH,
                 4

     The conventional explanation of
ignition in hydrocarbon oxidation involves
the oxidation of carbon monoxide (CO) to
carbon dioxide (CCO by OH radicals.  If
this occurs in CHC oxidation one may ask
why is there less exothermicity.   In order
to examine this important step in the
oxidation process,  the infrared emission
from CO (4.78(jm) and CO  (4.25|Jm) has been
measured for a series or shock tube exper-
iments with 1.5% fuel (CH,, CC14 and
CH«C1«) and stoichiometric oxygen.   The
analysis of this data indicates that while
CO is produced just as readily in the CHC
experiments as in CH, , CO,, is not pro-
duced.   This inhibition or the oxidation
of CO to CO,-, during the combustion of
CHC's may be an important cause of poor
incinerability characteristics of chlo-
rinated hydrocarbons.  This hypothesis is
in agreement with earlier work which
showed that chlorine inhibited carbon
monoxide flames [23]
Future Work

     The spectroscopic study of the ap-
parent inhibition of the carbon monoxide
conversion to carbon dioxide during the
oxidation of CHC's is presently continuing
in the conventional shock tube of the
Louisiana State University Combustion
Laboratory.  Future directions include the
use of a single-pulse shock tube, also in
the laboratory at LSU to study the product
distributions from the pyrolysis, oxida-
tion and reduction of selected CHC's.

     Studies of the behavior of co-fired
CHC's will be performed in the flat flame
burner facility of LSU's Hazardous Waste
Research Center.  Initially, the net
reaction rate profiles of stable species
will be determined during the combustion
                                            125

-------
     8.25
     7.50
     6.75
  £36.00
  CO
    5.25
    4.50
     3.75
    3.00
       3.0
                  TEMPERATURE,   ( K )
              1700   1500      1300
                                         I 100
               6.0
                        7.0
                      I04/T,
                                8.0
                                        9.0
                                                 10.0
Figure  4.   The natural logarithm of ignition delay
            times versus reciprocal temperature for
            C, hydro  and chlorocarbons.
                                                                               8.25
                                                                               7.50 -
                                                                               6.75
    6.00
                                                                               5.25 -
                                                                               4.50
                                                                               3.75
                                                                               3.00
                                                                                  3.0
                  TEMPERATURE.  ( K )
              1700   1500     1300        1100
                                                                                          6.0
                                  / A C2H6
                                  20l,2-C2H4CI2
                              U/   3 • l,l,l-C2H3Clj
                                  ^ a C2H,
                                   O CjHCI,
                        7 0
                      I04/T,
                                                                                                           8.0
                                                                                                                   9.0
                                                                                                                            10.0
Figure  5.   The natural logarithm of ignition delay
            times versus reciprocal temperature for
            C2 and  Cg  hydro and  chlorocarbons.

-------
process.   Similar to the shock tube
studies,  the overall goal of the flat
flame research is to eventually determine
the fundamental chemical kinetic reactions
which determine the combustion character-
istics of chlorinated hydrocarbons.
ACKNOWLEDGEMENT

     This report has  been reviewed by
the         ,  U.  S.  Environmental Pro-
tection Agency, and  approved for publica-
tion.  Approval does  not signify that the
contents necessarily reflect the views and
policies of the U.  S.  Environmental Pro-
tection Agency, nor  does mention of trade
names or commercial  products constitute
endorsement or recommendation for use.

     The research described in this
article has been funded wholly or in part
by the United  States  Environmental Protec-
tion Agency through  Cooperative Agreement
No. CR809714010 to  the Hazardous Waste
Research Center at  Louisiana State Uni-
versity.

     The authors would like to acknowledge
the assistance of Mr.  S. M. Courter, Ms.
J. Hayes, and  Ms. S.  Early during the
course of this study.   Further apprecia-
tion is extended to  Professor E. J.
Dantin, director of  the Hazardous Waste
Research Center at  Louisiana State Uni-
versity.
REFERENCES
 1.  "American Petroleum Institute Project
     44",  issued by Thermodynamic Research
     Center,  Texas A&M University, College
     Station,  Texas.

 2.  Benson,  S.  W.,  Thermochemical
     Kinetics,  2nd Edition,  John Wiley &
     Sons,  NY,  NY (1976).

 3.  Bowman,  C.  T.,  Combust.  Flame
     25:34(1975).

 4.  Burcat, A.,  Farmer,  R.  C.,  Espinoza,
     R.  L.  and Matula,  R.  A.,  Combust.
     Flame  36:313(1979).

 5.  Burcat, A.,  Lifshitz,  A.,  and
     Scheller,  K., Combust.  Flame 16:29
     (1971).
 6.  Burcat, A., Lifshitz, A., Scheller,
     K., and Skinner, G. B., "Thirteenth
     Combustion Symposium," Salt Lake
     City, 1971, p. 745.

 7   Cudahy, J. J., Sroka, L., and Toxler,
     W., "Incineration Characteristics of
     RCRA Listed Hazardous Wastes," EPA
     Contract No.  68-03-2568, Final
     Report, IT Enviroscience, Inc., July
     1981.

 8.  Duvall, D. S., Rubey, W. A., and
     Mescher, J. A., "High Temperature
     Decomposition of Organic Hazardous
     Waste," Proceedings of the Sixth
     Annual Research Symposium:  Treatment
     and Disposal of Hazardous Waste, U.S.
     EPA, MERL, EPA-600/9-80-010, March
     1980, pp.  121-131.

 9.  Duvall, D. S., Rubey, W. A., and
     Mescher, J. A., "Applications of the
     Thermal Decomposition Analytical
     System (TDAS)," paper presented at
     the Annual APCA Meeting, Montreal,
     Quebec, June 1980.

10.  Engleman,  V  S., Personal Communica-
     tion, Science Applications,  Inc., La
     Jolla, CA, November 1982.

11.  Gardiner,  W.  C., Jr., Rates  and
     Mechanisms of Chemical Reactions, The
     Benjamin/Cummins Publishing  Co.,
     Menlo Park, Calif. , 1972.

12.  Griffin, E. and Muraszew,  A., The
     Atomization of Fuels, John Wiley &
     Sons, NY,  NY (1953).

13.  JANAF Thermochemical  Tables, D. R.
     Stull and H.  Prophet, Ed., NSRDS-NBS
     37(1971).

14.  Jachimowski,  C. J., Combust. Flame
     29:55(1977).

15.  Joeng, K.  and Kaufman, F., J. Phys.
     Chem. 86:1808(1982).

16.  Kanury, A. Murty,  Introduction to
     Combustion Phenomena, Gordon and
     Breach, New York,  N.Y. (1975) Chapter
     5.

17   Kogarko, S. M., and Borisov, A. A.,
     Bull. Acad. Sci. USSR 8:1255 (I960).
                                           127

-------
18.   Lee,  K.  C.,  Morgan,  N.,  Hansen,  J.
     L.,  and  Whipple,  G.  M.,  "Revised
     Model for the Prediction of the
     Time   Temperature Requirements  for
     Thermal  Destruction of  Dilute Organic
     Vapors,  and  Its Usage for Predicting
     Compound Destructibility," Paper No.
     82-5.3,  75th Annual ACPA Meeting, New
     Orleans, LA, June 1982.

19.   Lifshitz, A., ed., Shock Waves in
     Chemistry, Marcel Dekker, New York,
     N.Y., 1981.

20.   Matula,  R. A., Gangloff, J. H.,  and
     Maloney, K.  L., "Symposium on Hydro-
     carbon Combustion," American Chemical
     Society  Meeting,  Dallas, 1973, p.
     355.

21.   NACA 1300, Basic Considerations  in
     the Combustion of Hydrocarbon Fuels
     with Air.  Chapter 1.  Atomization
     and Evaporation of Liquid Fuels  by
     Graves,  C. C. and Bahr,  D. W. ,
     (1957).

22.   Oppelt,  E. T., Civil Engineering-ASCE
     72 (Sept 1981).

23.   Palmer,  H. b. and Serry, D. J.,
     Combust. Flame 4:213 (1960).

24.   Tsang, W. and Shaub, W., "Chemical
     Processes in the Incineration of
     Hazardous Materials," National Bureau
     of Standards, paper presented at the
     American Chemical Society Symposium
     on Detoxification of Hazardous
     Wastes,  New  York, August, 1981.

25.   USEPA:  "Presentation of a Method for
     the Selection of POHC's in Accordance
     with the RCRA Interim Final Rule,
     Incinerator  Standards," January  23,
     1981, Office of Solid Waste, Aug. 13,
     1981.

26.   Vermeer, D.  J., Meyer,  J. W., and
     Oppenheim, A. K., Combust. Flame
     18:327(1972).

27   Westbrook, C. K., Combustion Science
     and Technology, 20 (1979).
                                           128

-------
                                  STATUS REPORT
                        USEPA COMBUSTION RESEARCH FACILITY
                                      (CRF)

                                Richard A.  Carnes
                       U.S. Environmental  Protection Agency

                                  F. C. Whitmore
                                   C. F. Fowl er
                                    R.  W.  Ross
                                   Versar,  Inc.
INTRODUCTION

     After extensive planning, design-
ing and a protracted period of budget
analysis, the USEPA Combustion Re-
search Facility (CRF) is on the thres-
hold of hazardous waste incineration
research.  It now becomes extremely
important for the EPA and its contrac-
tor to conduct research that will  be
cost effective, useful  to the Agency
fora better understanding of the com-
plex processes occurring during the
incineration of a hazardous waste.
Concomitant to those objectives the
CRF has the mission to  safely conduct
experiments using actual industrial
waste streams and to develop a system
of accurate and precise sampling and
monitoring of the process and to ini-
tiate an on-site analytical operation
that will provide the sophisticated
analyses required by legal  procedures
and citizens acceptance in  a reason-
able sonable time period.

OBJECTIVES

     Principal  objectives of the CRF
are to carry out pilot  scale test
burns on hazardous wastes,  be they
liquid, thixotropic, sludge and/or
solid, that have been previously
studied in a laboratory scale sys-
tem.  This will  allow the extrapola-
tion of data from the laboratory to
pilot scale systems and to indicate
differences between idealized labor-
atory studies and the results from
real world equipment.  The CRF will
constantly strive in all  its experi-
mental designs to study the effects,
either real or perceived, of impor-
tant incineration parameters such as
residence time, turbulence, tempera-
ture, nozzle type and configuration,
atomization techniques, etc., on the
destruction efficiency of hazardous
wastes, and the POHC and PICs asso-
ciated with the thermal  destruction
of the waste.

     The CRF will provide a vehicle
whereby the performance of different
design and configuration in commer-
cially available burners and burner
types relative to Destruction Effici-
ency (DE) and other associated incin-
eration operating requirements and/or
variables.  Ultimately the CRF will
provide a mechanism by which the basic
performance of different air pollution
                                     129

-------
control  devices can be evaluated in
order to permit the intercomparison of
apparently comparable devices under a
set of standard conditions.

     The CRF will  establish  a detail-
ed test protocol  for each incoming
waste for study.   Early experiments
will  look at the effects that a vari-
ety of operating parameters  have on
the basic combustion process.  Then
each parameter will  be evaluated as
to its contribution to the DE; those
found to have minimal  effect will  be
dropped from the test design.  Those
found to have a major impact will  be
investigated further.   In this way we
will  be moving toward an understand-
ing of process reliability without
process design change.  Another as-
pect of incineration that will  play
an integral  role in the CRF studies
is to determine the minimum  instru-
mentation required by an incinerator
such that should the process go out
of specification appropriate control
measures come into action to achieve
compliance without a total  shutdown
occurring.  In this regard we will  be
striving for overall  system  reliabi-
lity so that it can be transferred to
the field with assurance that it is
safe to both man and his environment.

     An important  aspect of  the CRF
incinerator lies in the on-line in-
strumentation that allows the detail-
ed evaluation of the operation of
each of the subunits that make up the
incinerator.  The  subunits in ques-
tion have been defined as the waste,
the waste feed and injection system,
the kiln, the afterburner and the air
pollution control  system.  The value
of this manner of thinking lies in
the ability to alter the operating
parameters of each subunit in such a
way as to introduce non-compliance of
the entire system.  In essence, such a
capability will  allow a definition of
the critical parameters of each sub-
unit and thereby assist in the
definition of the allowable range of
each variable and in the determination
of the necessary on-1 ine corrective
measures required to assure compliance
of the entire system.  This approach
might be termed "failure mode anal-
ysis" (although failure mode analysis
is a part of our program, there will
be no dangerous emissions to the envi-
ronment due to a downstream carbon
bed/HEPA filter system) of a hazardous
waste incinerator.

     Integrated throughout the entire
CRF operating philosophy is a sound
program of health and safety along with
system sampling and analysis.  All  CRF
personnel are required to have exten-
sive medical  examinations prior to the
onset of hazardous waste research.
These will  be supplemented with annual
physical exams and special testing as
warranted to specific waste streams.
The CRF has, in its basic design,
safety of operation, as evidenced by
explosion proof glass in the operation
room, which will  allow visual  observa-
tion.  The wall  adjacent to the ana-
lytical  laboratories has been rein-
forced with steel  and concrete.  The
roof of the incinerator room has two
large ventilation fans so as to pre-
vent combustible gas accumulation and/
or dangerous and toxic fume collec-
tion.

     The CRF has as a vital  part of
its overall  operating mission timely
and cost effective analysis of combus-
tion gas for Principal  Organic Haz-
ardous Constituents (POHCs) and Prod-
ucts of Incomplete Combustion (PICs)
along with residue analysis for haz-
ardous noncombustibles.  The labora-
tories are located on the periphery
of the incinerator room and broken in-
to four separate rooms.  There is a
waste characterization lab where all
incoming wastes will  be physically
characterized prior to incineration.
This will insure against incompatible
materials going to the incinerator and
                                     130

-------
will  provide insight into material
handling requirements and operating
parameter boundaries.  There are
two laboratories concerned with or-
ganic chemical  analysis from combus-
tion gas before and after scrubber.
Here is the heart of the analytical
operations and  it is in these labor-
atories that our turn around will  be
critical to getting the results in  a
timely fashion.  Present thinking
dictates evaluation of several  ana-
lytical procedures for time/cost
savings, however, the results must
be as accurate  and reproducible as
established ones or the new proce-
dures will be discarded.

     Thus there will be some at-
tempts at analytical methods deve-
lopment all in  the direction of
speeding up results without losing
precision and accuracy.  The last
laboratory presently has a Thermal
Decomposition Unit - Gas Chromato-
graph (TDU-GC)  which will  be used to
develop basic combustion parameters
for each waste  while looking for PIC
formation.  The TDU-GC provides the
first basic link in our overall  quest
for scale-up criteria for process de-
sign without getting into a basic de-
sign program.

CRF FACTS
1.  Construction Cost - $385,796.52.

2.  Physical  Dimension - 61'  x 51'
    (see Figures 1 and 2).

3.  Operation and Maintenance con-
    tract awarded to Versar,  Inc.
    April  1,  1982 for 18 months for a
    total  cost of $1,417,927.

4.  Option 1  for 12 months  for
    $975,705.

5.  Option II for another 12  months
    for $1,039,325.
6.  CRF personnel on-site, 13 includ-
    ing secretarial  and professional
    staff.

PLANNING TESTING PROGRAMS AND METHODS

     The set of experiments to be pre-
sented are designed  to test the opera-
tion and stability of the high temper-
ature zones of the incinerator, the
adequacy of the feed system, the air
pollution control  system, the on-line
operation of the control  and monitor-
ing systems, and the organic samplers.
In addition, the engineering staff
will  derive hands-on experience with
the use of safety equipment and with
the effects that such equipment have
on the performance of the various ac-
tivities in and about the incinerator.
Finally, these preliminary tests
should suffice as a  test  run on the
sample handling, record keeping and
laboratory operations prior to the
need to deal with more hazardous waste
materials, i.e., as  an operational  test
of the QA/QC Program at the CRF.

     The results of this  series of ex-
periments will  be of direct interest to
the staff of the CRF in that they will
test the procedures  and facilities in
a live performance.   The  results will
further serve as a basis  for judgement
of the reliability and utility of fu-
ture efforts in that the  proposed ex-
periments will  form  a base line for the
newly upgraded system against which the
effect of future changes  can be com-
pleted.

     Although the explicit details of
the proposed experiments  will  be dis-
cussed in more detail below, it is ap-
proprite to discuss  the general  philos-
ophy of these experiments at this time.
Much of the ancillary equipment in-
cluding the organic  samplers as well as
the upgraded incinerator  still  have not
been tested under waste feed condi-
tions.  These experiments are speci-
fically designed to  accomplish this
                                     131

-------
required testing.   In addition,  the
original  equipment manufacturer-sup-
plied air pollution control  system
(ARCS) has been replaced prior to the
onset of these experiments.   Clearly,
the experimental  tests will  be of
value in testing the new ARCS.  Fur-
ther, many of the  personnel  at the
CRF have never used safety equipment
under any but training activities.
The proposed HCB incineration  experi-
ments will  provide hands-on  experi-
ence with a substance that is  only
minimally toxic, but requires, due to
its physical  properties, minimal
safety controls.

     As has been pointed out,  the EPA
incinerator has been provided  with the
capability of conventional  stack sam-
pling under rigorous isokinetic  sam-
pling.  In addition, facilities  have
been provided for  hot zone sampling in
the kiln transfer  duct and within the
afterburner output duct.

     Sampling at the stack will  util-
ize EPA Methods 1  and 2 for the  deter-
mination of the gas flow parameters.
Sampling for residual  HCB will  be con-
ducted using standard EPA Method 5
with toluene as the stripping  solu-
tion.  Upon completion of the  sampl-
ing run, the entire sampling train
will be transferred to the laboratory
for sample extraction and train  clean
up.

     To estimate the duration  of sam-
pling it is only necessary to  note
that, with a feed  rate of 100  gm/hr
and a ORE of 99.99 percent,  the  emis-
sion rate (q0) should approximately
be:
     q  = 1  ^ gm/hr

The flue gas generation rate is  appro-
imately 400 SCFM so that the residual
HCB concentration, C0, in  the flue
gases woul d be:
     CQ = 4.17 x ID'7 gm/SCF

If this stream is sampled at the rate
of 10  1/min for one hour, then a total
mass, M0, of HCB would be collected,
where:

     M  = 600   x 4.17 x 10'7
          28.34

             8.8 x 10-6 gms

Introducing the assumption that reduc-
tion of the extract to 10 ml  and the
subsequent injection of 5 ul  into the
GC would result in an absolute mass,
M*, injected of:

     M* = 8.83 x 10-7 x 5 x 10~3
        - 4.415 x 10-9
gms
     Since the absolute sensitivity of
the electron capture detector is said
to be of the order of 1 x 10~9 gm (ac-
tual  laboratory data from the CRF sug-
gests that the actual detection level
is at least an order of magnitude
lower than this estimate), clearly the
sample taken for 1 hour is more than
adequate for the detection of a ORE of
99.99 percent.  Hence, stack sampling
will  require a 1 hour sampling period.

     In addition to sampling for the
POHC (in this case, HCB) in the stack
as well as in both the kiln transfer
line and in the afterburner exit line,
a number of additional parameters will
be sampled.  The listing of these
parameters, the method by which moni-
toring will be accomplished and the
details of monitoring appears as Table
1 for stack measurements.  Tables 2
and 3 provide the same information
for sampling rates at the kiln trans-
fer line and afterburner exit duct,
respectively.
                                     132

-------
     A number of additional  miscell-
aneous parameters will  be monitored
during the test burns.   These are
1isted in Tables 4 and  5.

EXPERIMENTAL DESIGN

     It is the explicit purpose of
this set of experiments to test the
newly completed upgrading of the CRF
incinerator and its ancillary equip-
ment.  For this reason, a rather ex-
tensive series of tests has been
proposed.  It will  first be essential
to test the liquid and  pressure in-
tegrity of the Andersen 2000 ARC
equipment which will  be carried out
using neutral water so  as to mini-
mize possible cleanup problems
should leaks be detected.  It is
further required that the effective-
ness of the demister be determined
(as will  be discussed below, the
effectiveness of the demister will
determine the significance of future
stack testing at the CRF).  Finally,
the overall  system must be tested
including the Bendix analyzer, the
organic samplers, the feed system
and the control  and monitoring sys-
tems.  This latter series of tests
will be carried out using HCB as the
test material.  It is estimated that
a series of light experiments will be
required to ascertain the necessary
information on the detailed operation
of the system and of its components.
The detailed measurements that will
be taken are indicated  in Table 6
which follows the justification of
each of the experiments.

EXPERIMENT I
with the caustic solution.  In addi-
tion to the testing of the integrity
of the piping it is also of interest
to determine the effectiveness of the
demister which will be accomplished by
a stack determination of aerosol  water.
Finally, both the temperature and the
ambient pressure will  be determined at
the important points throughout the
system.  It is anticipated that this
experiment will  require a full day for
its completion.

EXPERIMENT 2

     In experiment 2, the Andersen will
be operated at pH 9 in order to test
the adequacy of the pH control system,
the use of caustic in the absence of
HC1 , the residual  solids concentration
in the blowdown, the effect of caustic
on the C02 concentration in the stack
gases and a further test of the demis-
ter through a determination of the Na
concentration in the stack gases.  This
experiment is also expected to require
a full  day of testing.

     The accepted method for the deter-
mination of HC1  emission require the
collection of the HC1  in an aqueous me-
dium followed by quantitation using
a chloride ion specific electrode.
Clearly, NaCl  is going to be produced
within the scrubber and, if there is
significant carry over of NaCl parti-
cles, the resulting Cl ion concentra-
tion in the stack sample will  give er-
roneous estimates of the HC1  emission.
Clearly, the presence of sodium in a
stack sample, as in experiment 2 can
only result from particulate carry
over which indicates inadequate opera-
tion of the demister.
     The Andersen system has been just
been installed and therefore requires  EXPERIMENT 3
that the integrity of the many piping
junctions be tested.  Since there is
the possibility that leaks will  be
detected it is desired that testing
be accomplished with water rather than
     In experiment 3 further exerciz-
ing of the general  system will  be ac-
complished.  In addition, the ability
of the feed system to feed the 1:1
                                      133

-------
water/glycerine solution at a rate of  EXPERIMENT 7
1  gal/hr will  be determined.   This ex-
periment should be completed  in 4
hours.

EXPERIMENT 4

     As in experiment 3,  all  system
will  be subjected to additional  test-
ing with the feed system  operating at
near its maximum rate of  10 gal/hr.
This experiment should require appro-
ximately 3 hours for completion.

     It should be noted that  the first
four experiments will  also  be used to
determine the possibility of  contamin-
ants being introduced into  the organ-
ic samplers from the sample lines due
to extraction by the hot  gases.   In
the next members of the proposed set
of experiments, the system  will  be
challenged using HCB as the POHC.

EXPERIMENT 5

     This experiment has  as its goals
the challenging of the system and its
ability to achieve the required ORE of
99.99 percent for HCB.  In  addition,
the kiln temperature has  been select-
ed so that there should be  very little
decomposition of the HCB  therein.
This has been introduced  in order to
determine the efficiency  of the heat-
ed sample lines to prevent  the con-
densation of HCB prior to the im-
pingers.  It is further designed to
determine the efficiency  of the
Andersen system to deal with  low
concentrations of HC1  in  the  flue
gases.

EXPERIMENT 6

     This experiment has  the  same
general  goals as experiment 5 with
a  ten-fold increase in HCB  concen-
tration and the corresponding ten-
fold increase in HC1  concentration.
     This experiment has, as its goal,
a determination of the effect of kiln
temperature on the overall ORE of the
system for a compound such as HCB.

EXPERIMENT 8

     Experiment 8 is designed to deter-
mine the effect of AB exit temperature
on the ORE attainable by the system.
Further, since lowering the AB exit
temperature results in lowering the
flue gas flow rate (in ACFM), this
experiment will  also test the effect
of  a reduced input flow rate on the
scrubber efficiency for the removal  of
HC1 .

ANTICIPATED RESULTS

     This set of experiments has been
designed for the purpose of testing all
facets of the newly upgraded incinera-
tor.   The results of these experiments
should serve to establish a base line
on the operation of each of the basic
elements of the present system and to
establish the normal  operating charac-
teristics of these elements.

STATUS

The CRF has in-place an approved Qua-
lity Assurance/Quality Control  project
description, project responsibility,
QA objectives, sampling/measurement,
sample identification and custody,
calibration procedures and frequency,
analytical  methods, data reduction,
validation and reporting, internal
quality control  checks, performance
and  system audits, preventive main-
tenance, routine procedures used to
assess data precision, accuracy and
completeness, corrective action and
quality assurance reports to manage-
ment.  This document can be obtained
by requesting a copy from the EPA.
                                     134

-------
     The Health and Safety Manual  is
an integral  part of all  CRF activi-
ties and has been in place since
September 1982.  This manual  is bro-
ken into the following components:
administrative requirements,  general
facility safety rules and guidelines,
CRF respiratory protection program,
medical  surveillance program, health
and safety training program,  facility
security, personal  protective equip-
ment, safety equipment, mechanical
and electrical  equipment and  com-
pressed gases, emergency responses,
response to  fire emergencies, chemi-
cal spills,  response to loss  of
electricity  or water, waste trans-
port, handling and testing, and
storage, use and disposal  of  labora-
tory chemicals.  This manual  is
available upon request also from the
USEPA.

     A document entitled "Utilization
of the USEPA Combustion Research Fac-
lity" has been prepared that  describes
in detail  much of the facility and  its
operating philosophy.  At the present
time, it is  being reviewed and up-
dated and is not publically available.

FUTURE

     By mid-Spring of 1983, physical
start-up activities at the CRF should
be in complete action.  Immediately
following the start-up the action
plans call for a test burn using the
HCB test plan as previously described.
Following this, the rotary kiln system
will be subjected to a POHC "soup"  for
system and analytical testing.  During
this period  of time a complete RCRA
Part A and B hazardous waste  facility
permit application will  be developed
and formally submitted to Arkansas
authorities  for their approval.   It
is anticipated that during this RCRA
permit application procedure  a spe-
cial short term permit will be re-
quested in order to permit actual
hazardous waste research activities
to continue.

     Recent decisions by EPA officials
have resulted in the design plans for
fabrication and shakedown of a research
pilot-scale liquid injection incinera-
tor to accompany the present rotary
kiln system.  When complete the CRF
will possess the capability to conduct
hazardous waste research on technology
that presently covers about 90 percent
of all  field capability.  This will
surely make the CRF and its accompany-
ing TDU-GC and analytical  laboratories
one of the most comprehensive research
facilities for this particular activity
i n the worl d.

     Anticipated benefits from CRF ac-
tivities cover such operating parame-
ters as mixing, residence time distri-
bution effects, refractory type and
thickness, injection/atomization tech-
niques, feed rate, parameter control
systems and concepts such as operation
conformance to predetermined operat-
ing conditions.  The CRF will  con-
stantly strive to develop simplified
analytical procedures so as to reduce
the cost burden of test burns and
shorten the turnaround time from burn
to results.  We are looking to dev-
eloping in-line combustion as sampl-
ing and analysis for controlling
basic operating parameter.

     Since operations have begun at the
CRF it has been determined that the
facility qualifies as a major RCRA fa-
cility and must submit a formal  Part  A
and B facility permit application.
This will  entail  a significant effort
on behalf of the Versar Senior staff
and the on-site EPA project officer.
Al so since we are in an R&D mode of op-
eration, the tests must be conducted
in an upset mode. Therefore,it has been
determined that down stream of the nor-
mal stack emission, there will be a
carbon bed/HEPA filter system to insure
                                     135

-------
there are no toxic or hazardous emis-
sions to the environment.   This was
not in the original  plans  for the
CRF but has been included  as  opera-
tions can truly study hazardous
waste incineration and develop op-
erating conditions based on less
than optimal  conditions studied at
CRF.  Studying incineration at up-
set conditions is novel  and will
assist the regulators significantly
in establishing standard operating
conditions for field units and
should provide insight for scale-up
criteria to design engineers  study-
ing scaling of incinerator systems.
                                     136

-------
      A  PROFILE  OF EXISTING HAZARDOUS  WASTE INCINERATION FACILITIES
                              Edwin L.  Keitz
                            Leo J.  Boberschmidt
                           The MITRE Corporation
                         1820 Dolley Madison Blvd.
                          McLean,  Virginia  22102
                               Dr.  C.  C.  Lee
                   U.S.  Environmental  Protection Agency
                    Office of Research and Development
              Industrial Environmental Research Laboratory
                          Cincinnati,  OH  45224
                                 Abstract

     The incineration of hazardous wastes has been receiving increasing
attention since the implementation of the Resource Conservation and
Recovery Act of 1976 (RCRA).   The MITRE Corporation is under contract to
the U.S. Environmental Protection Agency (EPA) to assist in the
development of a hazardous waste incineration (HWI) data management system
A major part of this effort centers on the verification which began in
late 1981 and is still underway, a profile of hazardous waste incineration
in the United States has been developed.

     The principal approach was to follow-up and verify data extracted
from the RCRA Part A applications submitted to EPA by facilities who
indicated incineration as one of their process codes.  Information was
obtained from 514 of the 566 such facilities listed in the EPA's
automated Hazardous Waste Data Management System on 30 November 1981 plus
23 facilities identified outside the data base.   Facility spokesmen were
asked to verify the existence of hazardous waste incinerators at their
facilities, design characteristics, operational parameters and the types
and quantities of hazardous waste incinerated.

     This paper summarizes the results of the study.   A total of 284
operational hazardous waste incinerators were verified at 219 facilities.
Projection of these figures for 'the entire population results in an
estimate of approximately 350 operational incinerators at 270 facilities.

     Other data presented include geographical location of facilities,
types and capacities of incinerators, combustion zone temperatures and
residence times, types and quantities of wastes burned, and the use of
heat recovery and air pollution control devices.
                                   137

-------
INTRODUCTION

     The U.S. Environmental
Protection Agency (4)(EPA) estimates
that 57 million tons of organic
hazardous wastes are generated
annually in the United States.
Current estimates indicate that
perhaps 40 to 50 percent of this
waste can be disposed of by using
thermal destruction technologies.
EPA regards incineration as a
principal technology candidate for
destroying hazardous waste.  Since
the Congress enacted the Resource
Conservation and Recovery Act of
1976 (RCRA), incineration has been
included among those hazardous
waste disposal technologies that are
regulated by the Agency.

     In 1980, the U.S. EPA (3)
promulgated regulations requiring
every facility which is treating,
storing or disposing of hazardous
waste to file Part A of the RCRA
permit application form.  The data
submitted on these forms was stored
in a computer information system
entitled "Hazardous Waste Data
Management System" (HWDMS) which
is operated in each of the 10 EPA
Regions.  The incineration Research
Branch of EPA's Industrial Environ-
mental Research Laboratory in
Cincinnati has expanded HWDMS and
is developing the Hazardous Waste
Control Technology Data Base (HWCTDB)
to manage detailed incineration
engineering data, trial burn data
and related information.

     The information presented in
this paper is based on part of the
data assembled for the HWCTDB
project.  The topics to be discussed
include a profile of existing
incinerator facilities and a profile
of incinerator manufacturers.
PROFILE OF EXISTING INCINERATION
FACILITIES

     The existing facilities  data
discussed in this paper were
assembled originally in support of
the Regulatory Impact Analysis
Program for proposed regulations
concerning hazardous waste incin-
eration (Keitz et al. 2).  The
approach to the data assembly
began with preparation of a list
of all known facilities which
might have one or more operational
hazardous waste incinerators.  As
of the 30 November 1981 cutoff date
established for this list, 612 such
facilities were identified.  The
HWDMS contained 566 of these and 46
were identified from other sources.
However, at that time it was known
that some of the Part A applications
had not yet been entered into HWDMS.
Based on later information obtained
from HWDMS in July 1982 (estimated
100% complete), it was calculated
that the list of 612 facilities
was approximately 90 percent
complete.

     Initial telephone contacts
with many of these facilities
showed that a significant number
did not have an operational
hazardous waste incinerator.  Of
the 612 facilities, a total of 537
facility spokesmen indicated
whether or not their facility had
an operational hazardous waste
incinerator and provided varying
amounts of additional information.
The summary findings discussed here
are based on the information
verified by these 537 facilities.

     Table 1 shows the hazardous
waste incineration status of these
537 facilities divided into EPA
regions.  A total of 284 operational
HW incinerators were identified at
219 facilities.  Thus only 40.8% of
                                    138

-------
                                            TABLE 1

             STATUS OF HAZARDOUS WASTE INCINERATION  FACILITIES  IN EACH  EPA REGION3
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
TOTAL
Operational
Facilities
10
22
23
46
29
62
8
5
14
_0_
219
Operational
Incinerators
12
28
30
59
31
95
8
5
16
_0
284
Facilities With
No Operational
Incinerators
34
46
48
52
44
28
11
9
14
_0
286
Under
Construction
3
4
5
6b
3
4b
3
1
3
_p_
32b
Status
Unknown
1
1
0
0
0
0
0
1
0
0_
3
Sample
Size
48
73
76
102
76
93
22
16
31
_£
537
alnformation obtained from an estimated 81% of the  HWI  facility  population.
^Three facilities have both an operational unit and a unit  under construction;  two  in
 Region IV and one in Region VI.

-------
the facilities contacted verified
having an operational HW incinerator.
Inspection of the list of facilities
in the HWDMS data base in July 1982
showed that there were 128 facilities
not previously contacted during the
telephone campaign.   If it is
assumed that these facilities have
the same verification rate as those
contacted earlier, then an additional
52 facilities should have 68
operational HW incinerators.  It is
therefore estimated that there are
approximately 350 operational HW
incinerators at 270 facilities in
the United States.

     Table 2 shows the number of
operational HW incinerators by
type.  Of the 264 incinerators
whose type was specified, 208
(79 percent) are capable of burning
liquids by injection.  Twenty-nine
units (11 percent) are capable of
burning bulk wastes (solids or
liquids).  The remaining types are
mostly special purpose units such
as steel drum reconditioning burners
or military ammunition disposal
units.

     Table 3 shows the design
capacities of operational HW
incinerators.  Design capacities
were reported for 180 incinerators
burning liquids and 44 incinerators
burning solids.  The median design
capacity of incinerators burning
liquids is 150 gallons per hour
with most units (86 percent) not
exceeding 1000 gallons per hour.
Incinerators burning solids tend to
have smaller capacities with the
median being approximately 650
pounds per hour (equivalent to 78
gallons of water per hour).

     Table 4 shows the temperature
and gaseous residence time for
operational HW incinerators.  Gaseous
residence times were reported for
104 incinerators.  The median
combustion temperature was
approximately 1800°F  (980°C), and
median gaseous residence time was
slightly under 2 seconds.  It
should be noted, however, that
residence times can be calculated
by many techniques and the
respondents were not asked to
indicate the technique.  Therefore,
care should be exercised in
interpreting this data.

     Table 5 shows the major wastes
burned and the number of facilities
reporting these wastes.  Most of
the wastes reported are liquids,
principally spent non-halogenated
solvents and aqueous solutions of
corrosives, reactives or ignitables.
About 600,000 tons per year of
wastes were actually weak aqueous
solutions containing only a few
percent of the hazardous substance
reported.  These solutions accounted
for 59 percent by weight of all
wastes reported.  The most
frequently reported waste was the
non-listed ignitable waste with
high heat content (less than 6000
Btu per pound).  This waste was
reported for 69 incinerators.  The
largest single category of waste
by weight was non-halogenated
solvents (EPA Code F003) accounting
for 223,000 tons per year at 18
incinerators.
PROFILE OF THE HAZARDOUS WASTE
INCINERATOR MANUFACTURING INDUSTRY

     During February and March 1981,
incinerator manufacturers were
contacted in order to determine
those marketing hazardous waste
units (Frankel et al.,  1).  Four
directories were used to provide
names, addresses and telephone
numbers of manufacturers, specif-
ically:
                                    140

-------
      TABLE 2  TYPE AND NUMBER OF OPERATIONAL HW INCINERATORS3
Type
Liquid Injection
Hearth with Liquid Injection
Fume with Liquid Injection
Rotary Kiln with Liquid Injection
Combination System
Rotary Kiln (Solids Only)
Hearth (Solids Only)
Ammunition and Explosives
Drum Burner
Other0
Total Specified
Total Not Specified
TOTAL
Number
136
33
24
10
5
1
23
12
7
J.3
264
20
284
Percent of
Total
Specified
51
12
9
4
2
1
9
5
3
J>
100

Information obtained from an estimated 81% of the HWI facility
   population .
 Includes interconnected multiple units (e.g., rotary kiln in
   series with  liquid injection unit)-
clncludes such  items as fluidized bed incinerators.
                                  141

-------
                                                   TABLE 3




                         DESIGN CAPACITY OF OPERATIONAL HAZARDOUS  WASTE  INCINERATORS3
ro
Incinerators Burning
Capacity Number of
(Gallons/Hour) Incinerators
0-50
51 - 100
101 - 200
201 - 300
301 - 500
501 - 1000
1001 - 2000
2001 - 10,000
Total Specified
Unspecified
TOTAL
48
28
22
22
12
23
17
8
180
28
208
Liquids
Incinerators Burning Solids
Percent of Capacity Number of
Total Specified (Pounds/Hour Incinerators
27
16
12
12
7
13
9
4
100

0 - 100
101 - 300
301 - 500
501 - 1000
1001 - 2000
2001 - 5000
5001 - 10,000
10,001 - 20,000
Total specified
Unspecified
TOTAL
4
5
7
12
6
7
1
2
44
17
61
Percent of
Total Specified
9
11
16
27
14
16
2
5
100

       alnformation obtained from an estimated 81% of the HWI facility population.

-------
                                                    TABLE 4

                            REPORTED MAXIMUM TEMPERATURE AND GASEOUS RESIDENCE TIME
                                 FOR OPERATIONAL HAZARDOUS WASTE INCINERATORS3
(A)
Residence Time
(seconds)
<1.0
1.0 - 1.9
>2.0
Not Specified
TOTAL

1600°F
7
8
8
13
36

1600°F
-1900°F
2
17
3
42
64
Maximum
1901°F
-2200°F
0
6
32
12
50
Temperature
2200°F
4
5
7
7
23

Not
Specified
1
4
0
106
111

Total
14
40
50
180
284
       alnformation obtained from an estimated 81% of the HWI facility population.

-------
                                            TABLE 5

             NUMBER OF HW INCINERATORS REPORTING MAJOR WASTES AND QUANTITY BURNED3
Number of Incinerators Reporting
EPA Waste
Code
DO 01
DO 01
DO 01
D001
DO 01
D002,D003
F001,F002
F003
F005
K011
K016-K020
KQ49
P063
U220
Description
Ignitables (High Btu, High HW)b
Ignitables (Low Btu, Low HW)b
Ignitables (High Btu, Low HW)b
Ignitables (Waste Light Oils)
Ignitables (Unspecified)
Corrosives and Reactives
Spent Halogenated Solvents
Spent Non-Haolgenated Solvents
Spent Non-Hal ogenate'd Solvents
Acrylonitrile Production Bottoms
Hvy Ends, Cl Chemical Production
Slop Oil Solids, Petroleum Ref.
Discarded Hydrocyanic Acid
Discarded Toluene
Waste
69
19
6
3
7
32
18
22
24
3
10
1
6
3
Quantity
60
17
3
2
5
30
11
18
21
3
10
1
5
3
Quantity
Short Tons/Year
39,578
140,015
5,870
1,610
21,190
191,895
17,945
233,120
18,253
120,000
38,165
5,000
135,325
10,801
alnfonnation obtained from an estimated 81% of the HWI facility population.
bHigh Btu =  6000 Btu per pound.
 High HW  = major portion of waste is hazardous (e.g., organic liquids).
 Low HW   = major portion of waste is non-hazardous (e.g.,  contaminated water).

-------
     •   1981  Chemical Engineering
        Catalog

     •   February 1981 Buyer's Guide,
        Pollution Equipment News

     •   1981  Catalog and Buyer's
        Guide,  Pollution Equipment
        News

     •   1980-81 Directory and
        Resource Book, Air Pollution
        Control Association


     Hazardous  waste incinerator
manufacturers were asked to
voluntarily provide information
about the types of incinerators
manufactured, the approximate number
of units sold between 1969 and 1981,
and design and  operating information.
1969 was selected as a cutoff date
based upon several manufacturers'
estimates that  12 years of useful
service may be  expected from a
hazardous waste incinerator.  A
summary of the  number of manufactur-
ing companies and the number of
incinerators  in service classified
by type is presented in Table 6.

     Liquid injection incinerators
are most prevalent with 64.0 percent
of the market,  hearth incinerators
comprise 20.8 percent of the units
sold, and 12.3  percent of the
incinerators  are rotary kilns.
These three types account for 97
percent of the  units manufactured.

     Of the 57  companies identified
as marketing hazardous waste
incinerators, 28 have sold no units
in the United States.  Apparently
many of the companies that have not
sold an incinerator are anticipating
a large market  growth.  Of the 23
companies marketing liquid injection
incinerators, eight have sold none
to date; eight  of the 17 companies
offering rotary kiln incinerators
have sold none  to date; and  five
of the nine companies offering
fluidized bed incinerators have
sold none to date.  All hearth
incinerator manufacturers have sold
at least one unit.  Most of  the
companies offering  innovative
incineration technology have not
sold any units  to date.

     Incinerator capacities  may
be related by the thermal input or
the mass input  to the combustion
chambers.  The  ranges and typical
values of the capacities of  the
major types of  incinerators  are
.presented in Table  7.  Hearth
incinerators generally have  the
smallest capacity of the major
types, although rotary hearths can
be constructed  with capabilities
up to 170 million Btu/hr.  Typical
rotary kiln and liquid injection
incinerators have approximately
the same capacity.  Although the
largest incinerator listed in
Table 7 has a capacity of 150
million Btu/hr, some manufacturers
have received requests to bid on
facilities as large as 300 Btu/hr.
 COMPARISON  OF  OPERATIONAL  DATA WITH
 MANUFACTURERS'  DATA

      Table  8 presents  a  comparison
 of  the  number  of  HW incinerators
 reported  by manufacturers  and
 existing  HW facilities.  A total
 of  284  operational  HW  incinerators
 were  identified at  219 facilities.
 If  a  projection of  these facility
 figures is  made to  account for the
 estimated 128  facilities from
 which data  were not obtained,  the
 total operational HW Incinerator
 population  would be approximately
 350 at  270  facilities.  This figure
 agrees  very Well with  the  335
 operational units reported by
 manufacturers.   In  contrast.
                                    145

-------
    TABLE 6.  NUMBER OF HAZARDOUS WASTE INCINERATORS SOLD IN THE
               UNITED STATES.
Number of
Type of Manufacturing
Incinerator Companies
Liquid Injection
Fixed Hearth
Rotary Kiln
Fluidized Bed
Multiple Chamber Hearth
Pulse Hearth
Rotary Health
Salt Bath
Induction Heating
Reciprocating Grate
Infrared Heating
Open Drum
Total
23
12
17
9
2
1
1
2
1
1
1
1
Hazardous Waste
Incinerators
Sold
219
59
42a'b
9
7
2
2C
0
0
1
lc
0
342
Percent of
Total
64.0
17.3
12.3
2.6
2.0
0.6
0.6
—
—
0.3
0.3
—
100.0
^Includes five units in construction.
^Includes one oscillating kiln.
clncludes one unit in construction.
                                  146

-------
                       TABLE  7

DESIGN CAPACITIES OF HAZARDOUS WASTE INCINERATOR TYPES
              (From Manufacturers'  Data)
Incinerator Type
Liquid Injection
Hearth
Rotary Kiln
Fluidized Bed

Range
Ib/hr
30 - 24,500
25 - 2,500
1200 - 2080
	
Mass Capacity
Statistical
Value and
Population
Median of 43
Average of 48
Average of 2
1 Only
Thermal Capacity
Typical
Value
Ib/hr
1,600
810
1,600
31,000
Range
106 Btu/hr
0.125 - 130
3-9
1 - 150
8.5 - 67
Statistical
Value and
Population
Median of 50
Average of 4
Median of 34
Average of 5
Typical Value
106 Btu/hr
8
4.9
10.3
45.5

-------
   TABLE 8.   COMPARISON OF NUMBER OF HAZARDOUS WASTE INCINERATORS
              REPORTED BY MANUFACTURERS  AND HWI FACILITIES
                     Reported by HWI
                   Facilities Contacted
              Actual Number     Projection for          Reported by
	Reported	Total Population	Manufacturers

Operational
Incinerators      284               350                      335

Units Under
Construction       32                40                        7

Total Reported    316               390                      342
                                 148

-------
existing facilities reported 32
units under construction which is
much higher than the 7 reported by
the manufacturers.

     Table 9 presents a comparison
of the types of operational HW
incinerators reported by manufac-
turers and HW facilities.  The
manufacturer's data and the
projected total existing population
agree extremely well for the liquid
injection and hearth type inciner-
ators.  However, the rotary kiln
population reported by manufacturers
is more than double the number
reported by facilities.  Reasons for
the discrepancies may include:  some
of the units sold since 1969 may no
longer be in use or may now burn
non-hazardous wastes; some manufac-
turers might not have been given
enough information to know whether
the customer's wastes are hazardous;
or, some manufacturers may not know
if the customer's wastes' are
regulated under RCRA or the Toxic
Substances Control Act.  All of
these may cause high or low estimates
of the number of incinerators.
SUMMARY

     This paper has presented a
brief look at hazardous waste
incineration from two views:  namely,
existing facilities, and the
incinerator manufacturing industry.
A large majority of the wastes
incinerated are liquids with aqueous
solutions predominating.  A wide
variety of incinerators were shown
to be in operation throughout the
nation with reasonable agreement
between data reported by operational
facilities and the incinerator
manufacturing industry.  In addition,
the data tends to substantiate some
concepts which heretofore were
mostly assumptions.  These include
the locations, types and capacities
of hazardous waste incinerators.
On the other hand, some concepts
previously assumed will require
modification.  Among these are
the total number of facilities
incinerating hazardous waste and the
nature of the wastes incinerated.
In many cases the decision to
operate a hazardous waste inciner-
ator appears to have been selected
only when other choices such as
material recovery, recycling,
energy recovery or other disposal
methods were not cost-effective.
REFERENCES

1.  Frankel, I., N. Sanders and
    G. Vogel.  1982.  Hazardous
    Waste Incineration—Profile of
    Manufacturers.  MTR-82W31,
    The MITRE Corporation, McLean,
    Virginia.  38 pp.

2.  Keitz, E., L. Boberschmidt,
    D. 0'Sullivan and N. Sanders.
    1982.  Hazardous Waste
    Incineration—A Profile of
    Existing Facilities.  WP-82W84,
    The MITRE Corporation, McLean,
    Virginia.  82 pp.

3.  U.S. Environmental Protection
    Agency.  1980.  Part A of
    Hazardous Waste Application
    Requirements:  paragraph  122.24
    and Form 3.  45FR33543-33588,
    May 19,  1980.  Washington, D.C.
    46 pp.

4.  U.S. Environmental Protection
    Agency.  1981.  Engineering
    Handbook for Hazardous Waste
    Incineration.  EPA SW-889,
    Industrial Environmental
    Research Laboratory, Cincinnati,
    Ohio 45268.
                                    149

-------
TABLE 9.  COMPARISON OF THE TYPES OF OPERATIONAL HW INCINERATORS
          REPORTED BY MANUFACTURERS AND HWI FACILITIES
                     Reported by HWI
              	Facilities Contacted	
              Actual Number     Projection for          Reported by
	Reported	Total Population	Manufacturers

Liquid
Injection        159(a)            213                    219
Hearth            56^ b'             75                     70

Rotary Kiln       13(b'c)           17                     37

Fluidized Bed      45                      9

Other             31(d)             42                     (e)

Type not
Specified         20                _J)	

Total
Operational      284                 352
(a) = includes fume/liquid units.
(b) = includes units both with and without liquid injection.
(c) = includes 2 rotary kilns in combination units.
(d) = includes 3 combination units not having a rotary kiln.
(e) = this category not obtained from manufacturers.
                                 150

-------
            PARTICULATE  AND  HC1  EMISSIONS  FROM HAZARDOUS WASTE INCINERATORS
                                       Paul Gorman
                                     Andrew Trenholm
                               Midwest Research Institute
                              Kansas  City,  Missouri  64110

                                      Don Oberacker
                          U.S.  Environmental Protection Agency
                                 Cincinnati, Ohio  45268

                                        Ben Smith
                          U.S.  Environmental Protection Agency
                                 Washington, D.C.  20460
                                        ABSTRACT

EPA regulations  place  limits  on particulate and HC1 emissions from hazardous waste incin-
erators.   MRI  has  collected particulate and HC1 emissions data from a variety of such in-
cinerators.   This paper  evaluates the data,  control  system performance,  and probable
mechanisms and relationships affecting emissions.  Results  indicate  that alkaline scrub-
bing and  particulate emissions  may be related.
INTRODUCTION

     EPA regulations  (under RCRA)  for haz-
ardous waste incinerators  require  that par-
ticulate emissions be  no  more  than  180
mg/Nm3 (0.08 gr/dscfj corrected to 7°/0 02 ,
and that chloride removal  efficiency  be no
less than 99% if chloride  emissions exceed
1.8 kg/h (4 Ib/hJ.  As  a result, most haz-
ardous waste  incinerators  which  handle
waste containing chlorinated compounds are
equipped with particulate/HCl removal sys-
tems.  This equipment usually involves some
type of wet scrubbing device, the  most com-
mon being packed towers, venturi scrubbers,
or electrified scrubbers.   In  some cases
the scrubbing medium  is  water only; but in
other cases, the scrubbers utilize recircu-
lated water with the addition of  alkaline
materials (caustic or lime)  to  neutralize
absorbed HC1.

     This paper emphasizes HC1 and particu-
late emissions,  and factors related to the
control of  these emissions.  Data  and re-
sults discussed in the  paper are primarily
based  on  recent work  done  for EPA  by
Midwest Research Institute (MRI).
     The next section summarizes how these
emissions are determined and  the  implica-
tions of the measurement methods which may
not be immediately obvious.
MEASUREMENTS OF HC1 AND PARTICULATE
  EMISSIONS

     The EPA  Modified Method 5  sampling
train, as  shown  in Figure 1,  is used  to
collect samples  for particulate  and  chlo-
ride emission measurements.  The probe, cy-
clone, and  filter  are  the  only  components
used to quantify particulate emissions, and
the only components used to quantify chlo-
ride emissions are the contents of the con-
densate and KOH impingers.

     The probe cyclone and filter compo-
nents are  maintained  at a temperature of
121° ± 14°C (250° ± 25°F) during sampling.
Therefore,   any droplets  carried  over from
the scrubber would be  evaporated after they
entered the sampling probe.  If those drop-
lets contained dissolved HC1, not combined
                                            151

-------
                                                         , Condenser
             Filter Used in All Runs
                        V
ui
ro
 Probe    .JL
a     &4              II
                     Cyclone
                                                                                                                                            Exhaust
                                     Implngor   4 Liter
                                     (Reversed) Botrle
                                   I           (Condemate)
                                                  Empty
                                                  (Condensate )
                                                                                                             Dry Got
                                                                                                             Meter
                                                 Figure 1.   Schematic Illustration  of Sampling Train

-------
with any alkali salts, then the HC1 would
be returned  to  the gaseous state and would
pass through the  filter and be captured by
the condensate or  KOH and quantified as
part of the  chloride  emissions.  On the
other hand,  if  the droplets contained sus-
pended or dissolved solids  (e.g., Nad or
CaCl2),  these could form particulate matter
when the droplets  evaporated,  which should
be captured  on  the filter and  quantified as
part of the  particulate  emissions.   As a
result,  droplet  carryover  may  increase
either  the  chloride  or  the  particulate
emissions,  or both, depending on the com-
position of  the droplets.  This aspect of
the  sampling method  is pertinent to the
discussion of HC1 and particulate emission
data  presented  in subsequent  sections.
HC1 EMISSIONS

     During  the  combustion process, HC1
emissions  are  formed from chlorine  com-
pounds in  incinerator waste feeds.   It  is
the function of scrubbing systems to absorb
this gaseous HC1 from the combustion gases
and to achieve 99% removal, the requirement
under  RCRA (unless emissions  are  below
1.8 kg/h).

     MR I has tested several hazardous waste
incinerators burning high  Cl  wastes that
utilize some type of wet scrubbing system.
Data for six of these facilities are given
in  Table  1  and  show  that at least  five  of
the six  achieved  the required  99%  HC1  re-
moval.  The  first  facility listed in Ta-
ble 1  (Plant A) achieved  99%  efficiency
when operating normally with caustic addi-
tion  to  maintain  the scrubbing liquid at
pH  4,  but  HC1  removal  was below  99% when
the pH of  the  scrubbing liquid was quite
low (pH 1).

     It would appear from  the  data in  Ta-
ble 1  that Plant B achieved 99% HC1 removal
even  though  the pH of the  liquid  effluent
was low  (pH 2).   However,  the pH  of this
effluent was raised to a much higher value
(~ pH  10) by addition of lime, after which
the solution was  recirculated  back to  the
scrubbers.  Thus, the pH of the liquid fed
to  the scrubbers  was  about  pH 10  even
though the  outlet  liquid  had a low pH of
about  2.
     An HC1 removal efficiency  of  99%  was
achieved by three  plants  that used water
only with  no  alkaline  addition to  the
scrubbing liquid.  Two  of these  (Plants  E
and F)  used once-through water,  which means
that water usage was relatively large.   The
other plant (Plant D)  utilized water recir-
culation but  incorporated  three  packed
towers  in series, with the fresh makeup wa-
ter being  fed to  the  top of the  third
tower.   Liquid  effluent from this  third
tower was then fed to  the second tower, and
so on.

     Based on the  data presented,  it  is
clear that 99% HC1 removal can be achieved
by  wet  scrubbing systems without  neces-
sarily using alkaline  scrubbing solutions
In  fact, theoretical  calculations  of HC1
removal efficiency, as  outlined  in Bonner
et  al.   (1), have been carried out by MRI
using average  design and operating charac-
teristics for scrubbers employed at hazard-
ous waste incinerators.  These theoretical
calculations,   based  on  a normal  water
makeup and drawoff rate of the recirculat-
ing liquid, have shown that 99% HC1 removal
is  achievable  using water only.  However,
in water-only systems, the pH of the liquid
would be  low, and  this  acidity must be
taken into account  in  selecting  construc-
tion materials.

     On the basis of the theoretical calcu-
lations, one might question why Plant A  did
not  achieve  99% removal  even when  the
caustic addition system malfunctioned, re-
sulting in a  scrubbing liquid with  low pH.
MR I  testing  was done  during the  period
without the mist eliminator,  so,  based on
the particulate  emission  data,  it  is be-
lieved  that there  was  excessive  carryover
of  scrubbing  liquid.  Any HC1 dissolved  in
droplets  carried over from  the  scrubber
would cause an increase in the measured  HC1
emissions.

     For systems with  alkali addition, any
droplet carryover of the scrubbing  solution
could result  in  an increase  in the  measured
particulate emissions  instead of  an  in-
crease  in the HC1 emissions.  Alkaline ad-
ditives  and/or soluble reaction products
(NaCl or  CaCl2) , as well  as  any suspended
solids,  can adversely affect the measured
particulate emissions, especially  if  the
                                            153

-------
               TABLE  1.  SUMMARY OF DATA ON PARTICIPATE CONCENTRATION AND HC1 REMOVAL EFFICIENCY
                           FOR INCINERATORS BURNING WASTES CONTAINING CHLORINATED COMPOUNDS

Plant Scrubbing system
A
B
C
D

E
F
75-cm water column (w.c.) AP
venturi -> sieve tray tower
Packed tower -> 1-stage electri-
fied scrubber (ES)
Packed tower -> packed tower -»
2-stage ES
300 cm w.c. AP venturi -» 3 packed
towers in series
1 packed tower
1 packed tower
Alkaline Avg HC1 removal Avg particulate
scrubbing Mist efficiency and pH emissions
media eliminator of scrubber effluent mg/Nm3 (gr/dscf) at 7% 02
Caustic See last > 99% at pH 4;
column < 99% at pH 1
Lime Yes > 99% at pH 2
Caustic Yes > 99% at pH 5
H20 only Yes > 99%*

Once-through Yes > 99% at pH 1
H20
Once- through Yes 5 99% at pH 1.5
H20
1,520 (0.666) (no mist
eliminator)
220 (0.095) (with mist
eliminator)
78 (0.034)
150 (0.066)
23 (0.010)

200 (0.088)
4 (0.002)

pH of final effluent was 2 but is not comparable to other systems since this plant used three scrubbers in series
with feed-forward water flow.

-------
system is not equipped  with a good mist
eliminator.   This subject is  discussed  in
the following section.
PARTICULATE EMISSIONS

     Achieving  the  particulate  emission
limit of 180  mg/Nm3  (0.08 gr/dscf),  cor-
rected to 7% 02, may be more difficult than
achieving 99% HC1  removal.   The level of
particulate emissions is influenced by sev-
eral factors  that  are  represented by the
equation:


    Particulate = (A + Q)(l   Eff) + C
     emissions

where:    A -  Amount of ash in waste feed
               that  becomes  suspended in
               combustion gases

          Q =  Solids produced by quenching
               with recirculated scrubbing
               solution

        Eff =  Particulate  removal effi-
               ciency  of control  system

          C =  Particulate resulting  from
               droplet carryover


Important aspects of each of the factors in
the  above  equation are discussed  in  the
following paragraphs.

Combustion

     All wastes  contain some  "ash,"  the
amount  of  which can vary widely.  Solid
waste usually has the highest ash content.
During the combustion process, a portion of
this  ash is  suspended in the  combustion
gases, which  is  the  origin  of  some of the
particulate matter that must be  removed to
achieve a stack concentration of 180 mg/Nm3
(0.08 gr/dscf).

     None of the tests conducted by MRI in-
volved  determination of  the particulate
concentration  in the very hot combustion
gases prior  to quenching  or scrubbing.
However, if an incinerator burned a liquid
waste containing 4% ash and, if  all the ash
entered the combustion gas stream, the re-
sultant  particulate  concentration in the
combustion gases  would be  about 2,000
mg/Nm3 (1 gr/dscf) .
Quenching

     Quenching  cools  the hot  combustion
gases before they enter the particulate/HCl
control system.  Water sprayed into the gas
stream to achieve this cooling effect  may
also achieve some particulate control.   The
amount sprayed into the hot gas stream usu-
ally exceeds  the amount  evaporated  in  the
quenching process;  thus,  some particulate
may  be  removed by  impaction with water
droplets.  In some plants, however, part of
the  recirculated solution from the  scrub-
bers is  used  for  quenching.   Some of the
water droplets  sprayed into  the quenching
section will be totally evaporated and any
solids dissolved in  those droplets would
become part of the particulate that enters
into, and must  be removed by, the  control
system.

Control Systems

     The type and configuration of control
systems installed for particulate/HCl  re-
moval vary  widely from  plant  to  plant.
They may involve use of a venturi  scrubber
followed by one or more packed towers, or
packed towers  only.  Two  plants tested by
MRI  also used  electrified scrubbers  (ES).
Basically,  these devices  consist  of a high
voltage section for charging the  particles
in the gas  stream.   These charged  particles
are then collected by a series of  grounded
vertical plates with a continuous  flow of
liquid down  the surface  of  the  plates.

     Some of the variations  in control sys-
tem  configuration are shown  schematically
in Figure 2.  The systems also vary in op-
eration.   For example,  the venturi scrubber
at one plant  was designed to operate at a
pressure drop of 70 to 100 cm water column
(w.c.)  (30  to 40  in.  w.c.)  while  one  at
another plant  operated as high as 300 cm
w.c.  (120 in.  w.c.).   All of the  plants
tested by  MR I used  at least one  packed
tower, except  for  Plant  A,  which used a
sieve tray  tower.   The packed towers  or
tray towers are primarily intended for HC1
removal  and  characteristically have low
pressure drops  of  10  to   25 cm w.c. (5 to
10 in. of H20).  Packed towers may provide
some particulate removal, but low pressure
drop systems of this type have poor partic-
ulate removal  efficiency  for  small  parti-
cles  (<  5 |Jm) .   Only one of the  tests
conducted by  MR I  included determination
of  particle size at  the inlet to the
                                           155

-------
     Plant A •
                              Venturi
                              Scrubber
            Sieve
            Troy
            Tower
     Plant B
                                             Electrified
                                             Scrubber
     Plant C •
                              Packed
                              Tower
               Electrified
               Scrubber
     Plant D
                              Venruri
                              Scrubber
                              Packed
                              Tower
     Plant F
                               X
                               Packed
                               Tower
•Quench
 Section
Figure  2.   Schematic Diagram of Control  Systems  at Various Plants
                                     156

-------
particulate  removal system  (after quench-
ing).   This  test showed that at least 90%
of the particulate  (by  weight)  was less
than 2 pra in size.   This strongly suggests
that unless  the  waste  feed had low ash con-
tent,  packed towers alone  could  not achieve
the particulate  limit of  180 mg/Nm3 (0.08
gr/dscf)  and  that  other control devices
(e.g., venturi scrubbers or ES)  are neces-
sary for  that purpose.

     It is possible that,  no matter how ef-
ficient  such  particulate  removal  devices
are, the  particulate limit may be exceeded.
If a particulate removal device  is followed
by a packed  tower (for HC1 removal),  drop-
lets  carried  over  from  the packed tower
could contain dissolved salts which would
adversely affect particulate test results
(as discussed in "HC1  Emissions" section).
For this  reason most of the plants listed
in Table 1 were  equipped with mist elimina-
tors.  Tests at  Plant  A showed the dramatic
difference between operating with and with-
out a mist  eliminator,  and  suggest that  a
mist eliminator  can reduce particulate con-
centrations   by  more than  1,200 mg/Nm3.

     One caution, however, is that informa-
tion  in  Maroti  (2)  states  that  the quoted
amount of carryover from mist eliminators,
ranging  from  40  to  400  mg/m3  (0.02 to 0.2
gr/acf),   is  much lower  than limited field
and laboratory data indicate (400 to 4,000
mg/m3).  If droplet carryover were as high
as  4,000  mg/m3  (2.0 gr/acf)  and the drop-
lets  contained only 1% dissolved salts, the
resultant particulate concentration would
be 50 mg/m3  (0.02 gr/acf),  representing at
least one-fourth of the allowable limit of
180 mg/Nm3  (0.08 gr/dscf).

Discussion  of Data

      Particulate emission data presented in
Table 1  for the  six incinerators studied
show  that four facilities met the particu-
late  limit  of 180 mg/Nm3  (0.08 gr/dscf)
corrected to  7% 02, while two did not.  The
data  for Plant  A clearly demonstrated a
drastic  reduction in particulate emissions
after the mist  eliminator was  installed,
but emissions were still slightly above  the
limit of 180 mg/Nm3 (0.08 gr/dscf).  It  is
not known whether  this  was due to inade-
quate performance  of  the  venturi  scrubber
or  the need for  even  better mist  elimina-
tors .
     Plant E is the other of the two plants
which exceeded the particulate limit.  Con-
sidering  that  Plant E was equipped  with
only a packed tower, it is not too surpris-
ing that  it  did  not meet the particulate
limit.  It is somewhat surprising, however,
that the emissions were close to the limit.
The liquid waste  fed  to this incinerator
was from an associated chemical operation,
so it may  have  been low in  ash  content.
Similarly, it  is  believed  that the waste
feed  for  plant F  was  low in ash content.

     Based on  the above findings, it  is
evident that, unless the waste feed is very
low in  ash content, some  type  of  particu-
late removal device is needed at hazardous
waste incinerators in order to achieve the
limit of 180 mg/Nm3 (0.08 gr/dscf) and that
packed  towers  alone are probably  not  ade-
quate for meeting that  goal.  It  is  not
possible from the data at hand to ascertain
exactly what the particulate removal system
must  be for incinerators burning  wastes
with  high ash  content,  although  data  from
Plant A indicate  that a  70-cm  (30-in.) AP
venturi may not be sufficient.  This is not
inconsistent with control  device perfor-
mance,  considering the very small particle
size of 90% < 2 |Jm.  The data also indicate
that  incinerators  burning  low  ash content
wastes  may achieve  the  limit without any
particulate removal device.

      It is  apparent  from  the data that
packed  towers  require installation of  ef-
fective mist  eliminators.    It  had been
postulated that the need for effective mist
eliminators  would be more  critical  for
plants  using alkaline scrubbing for HC1 re-
moval,  since any  carryover of this solution
would  adversely  impact  particulate emis-
sions.  In  an  effort  to investigate this
aspect, available data were  compiled  on
metals  content  of the emitted  particulate
for Plants  B,  C,  and D.   Plants  B and C
used  recirculated alkaline scrubbing solu-
tion  while  Plant  D used water only.   The
available data  on the metals detected in
the particulate (by acid digestion and ICAP
analysis)  are   shown  in Table 2.  Both
Plants  B  and  C  indicated significantly
higher  Na concentrations than did  Plant D.
In Plant  C  the higher Na  concentration is
presumably  related  to the use of  caustic
scrubbing.   Although Plant  B  used  lime
scrubbing, there  is reason to believe  that
the lime  slurry contained  a  relatively high
                                            157

-------
 TABLE 2.   COMPARISON OF METALS CONTAINED
           IN STACK PARTICULATE (%)



Al
Ca
Fe
K
Na
Pb
Plant
Runs 6 ,
0.
0.
1.
1.
15.
4.
•p-'-
7, 8
1
9
5
8
3
8
Plant
Runs
0.
0.
0.
0.
18.
8.
C
1-3
2
8
2
4
6
4
Plant D
Run 2
ND
ND
4.8
ND
6.4
1.9

  TABLE 3.  PERCENTAGE OF MASS  INPUT FOR
            EACH METAL THAT  IS  EMITTED
            AS PARTICIPATE
                                                               Output rate as percent
                                                               of input rate (%)	
                                                       Plant B
                     Plant C
Plant D
ND = Not detected.
Al
Ca
Fe
K
Na
Pb
1.6
4.1
9.7
120
49
5.9
0.6
3.9
1.1
16
15
17
ND
ND
0.7
ND
0.7
1.1

"  Analysis of particulate collected in
   tests prior to plant modifications
   which reduced the particulate emis-
   sions .

concentration of Na.   The results for Plant
B do not show higher Ca content, however,
as might be expected for lime scrubbing.
Insufficient data are available to explain
this inconsistency.

     It should be noted in Table 2 that the
total of the major metals detected for each
plant is on the order of 20 to 30% of the
total particulate.  But these data repre-
sent only elements, not compounds.  Thus,
if the particulate contained 15% Na,  then
the equivalent percentage of Nad would be
38%, which is a significant portion of the
cotal particulate.

     The MRI tests also included analyzing
for metals in the waste feeds.  The partic-
ulate and waste feed data were then used to
calculate the output rate of each metal
(g/min) in the particulate emissions  and
the input rate of each metal in the waste
feeds.  The data were expressed as the per-
centage of input that was emitted from the
stack as particulate (Table 3).  Thus, the
higher the number, the greater the portion
emitted (i.e., 100% would indicate that the
amount emitted equals the input amount).

     Data in Table 3 reflect differences
from plant to plant in particulate removal
efficiency.  For Plants B and C, the  data
show a relatively large percentage of Na
being emitted in comparison to other  metals
(Al, Ca, Fe).   What is even more evident  in
ND = Not detected in particulate samples.

Table 3 is that a large percentage of the K
was being emitted.  One possible explana-
tion for these findings is that the Na and
K emissions include contribution of those
two very soluble components from carryover
of alkaline scrubbing liquids used in
Plants B and C.  Apparently substantiating
this fact is that at Plant D, which did not
use alkaline scrubbing, there are no large
differences in the percentage of each metal
emitted.
CONCLUSION

     MRI's analysis of information and data
obtained in tests at six hazardous waste
incinerators was used to assess HC1 removal
and particulate emissions.  The data show
that 99% HC1 removal efficiency was
achieved at five of six plants, and was
also achieved at the sixth plant  (Plant A)
when it was operating normally.  Three of
the plants that achieved 99% efficiency
used water only, which shows that alkaline
scrubbing is not necessarily required to
achieve 99% HC1 removal efficiency.

     Two of the six plants tested did not
achieve the particulate emission  limit of
180 mg/Nm3 (0.08 gr/dscf), corrected to 7%
02.  One of these used only a packed tower.
The other used a 70-cm (30-in.) AP venturi
scrubber.   Considering the probable small
size of the particulate (as determined at
one plant), there is reason to believe that
the particulate limit will not be easy to
achieve.
                                           158

-------
     Particulate  control  devices better
than a 70-cm AP venturi may be necessary,
except in those situations where the waste
feed has  a low ash content.   Another impor-
tant aspect  of meeting the  particulate
limit is  that  effective mist  eliminators
must be  used  in conjunction with packed
towers or other alkaline scrubbing devices.
Available data  on metals  analysis of par-
ticulate  emissions indicate that carryover
of soluble alkali metals  (Na,  K) from alka-
line scrubbing systems may represent a sig-
nificant  portion  of the particulate emis-
sions .
REFERENCES

1.   Bonner, T.,  et al.  November  1980.
     Engineering Handbook for Hazardous
     Waste Incineration.  Draft  prepared
     for  the  Environmental  Protection
     Agency by Monsanto Research Corpora-
     tion, Dayton,  Ohio, EPA Contract  No.
     68-03-2550.   p. 4-58.

2.   Maroti, L. A.  August 1982.   Entrap-
     ment in Wet Stacks.  GS-2520,  Electric
     Power  Research Institute.   p.  2-2.
                                           159

-------
               EMISSION TEST RESULTS  FOR A HAZARDOUS WASTE INCINERATION RIA
                                      Andrew Trenholm
                                        Paul Gorman
                                Midwest Research Institute
                               Kansas  City,  Missouri  64110

                                      Benjamin Smith
                           U.S.  Environmental Protection Agency
                                  Washington, D.C.   20460

                                    Donald  Oberacker
                           U.S.  Environmental Protection Agency
                                  Cincinnati, Ohio   45268
                                         ABSTRACT
The Environmental Protection Agency  is  preparing a Regulatory Impact Analysis  (RIA) of
hazardous waste incineration.   Data from 20 tests were gathered for this analysis includ-
ing tests conducted by Midwest  Research  Institute (MRI)  at eight  incinerators.   The scope
of MRI's test  program  is  described and  the results are summarized.  The test covered a
range of  waste and  incinerator  types,  combustion temperatures, and  residence  times.
Principal organic hazardous constituents  (POHC)  were  identified and  quantified in the
waste feeds and stack  effluents to determine destruction and removal efficiencies  (DRE).
A number of products  of  incomplete combustion (PIC)  were  also  identified and quantified
in the stack effluent.  Other measurements made were POHC in liquid and  solid effluents;
chlorides and  metals in all effluents; and particulate, hydrocarbon, and carbon monoxide
in the stack effluent.
INTRODUCTION

     The Resource Conservation and Recovery
Act  (RCRA)  has  resulted in the Environ-
mental  Protection  Agency (EPA) enacting
hazardous waste  regulations affecting all
who generate, store, transport, treat,  and
dispose of such wastes.   While several  dis-
posal and treatment options are available,
incineration has received a great  deal of
attention because  it is  a technology that
is presently available and provides perma-
nent disposal of many organic  wastes.  As
part of EPA's program,  the Office of Solid
Waste  (OSW)  is  required under Executive
Order 12291 to conduct a Regulatory Impact
Analysis  (RIA)  to  ascertain the  costs and
benefits associated with various approaches
to regulating hazardous waste  incinerators.
Integral  parts  of   this  process  are the
definition  of  current  performance,  and
evaluation of the effect of various incin-
erator design and operating parameters  on
the destruction  of hazardous  organic pol-
lutants and  the  effectiveness of control
devices for the removal of particulate and
HC1 emissions.   EPA's  Office  of Research
and Development  (ORD)   is  also actively
pursuing  the advancement  of  knowledge
about hazardous waste incineration by pro-
viding technical  support  to OSW and  look-
ing at longer range research  needs.

     Over the  last year MRI has conducted
tests at  hazardous  waste  incinerators  to
provide  data  for  EPA's programs.   The
tests  were  similar to  incinerator  trial
burns, but  had  a broader scope.   A  wide
variety of samples were collected at eight
incineration facilities.
                                           160

-------
Table 1 summarizes  the types of  samples
collected  and  the  analyses performed.
Emphasis was on  measurement of  principal
organic hazardous  constituents   (POHCs),
products of incomplete combustion (PICs) ,
chlorides,  and particulates.  Samples of
all input and  effluent  streams  were col-
lected  at every  site  and archived  after
specified  analyses were performed.

     Composite  samples of  the wastes and
grab samples of  liquid  and ash  effluents
were collected.   Stack sampling  methods
included a Modified EPA Method 5  for  par-
ticulates,  semivolatile  POHCs  and  PICs,
and chlorides;  two methods  for collecting
volatile POHCs  and PICs,  an integrated
gas bag and the  volatile organic  sampling
train (VOST);  and continuous gas  analyzers
for 02, CO, C02 ,  and  total hydrocarbons.
Analyses of  POHCs and PICs  were by gas
chromatography/mass spectrometry  (GC/MS),
and  metals  were  analyzed by inductively
coupled argon  plasma  (ICAP)  or   atomic
absorption (AA).  Details  on the  sampling
and analysis methods  used  were  presented
by Dr.  Gregory A. Jungclaus  in an earlier
paper at this conference.

     This paper  is  an overview of  MRI's
testing  results.   It  characterizes the
incinerator sites  and wastes that were
            burned and  summarizes  the results of the
            tests.   Specific  characteristics or  re-
            sults  are  not presented  for each  site
            separately  due  to confidentiality agree-
            ments  with some  of the  sites  tested.
            SITE CHARACTERIZATION

                 The  eight  incinerators  tested are
            characterized by  a wide variety  of  incin-
            erator types, waste  types,  and operating
            conditions.  Table 2  shows  the distribu-
            tion  of  incinerator  types  and  control
            devices.    All   eight   incinerators  had
            liquid injection  burners.   Four injected
            both  aqueous  and  organic liquid wastes.
            Five of  them had  a  rotary kiln or hearth
            for solids (two did not feed solids during
            the test),  and  one  ducted gaseous  waste
            directly to the incinerator.   Three of the
            incinerators did not have any air pollutant
            control  devices.   The  other  five  had
            packed scrubbers for HC1 control, and four
            of those  five also had a particulate con-
            trol device.

                 The eight  facilities  can  be further
            classified as on-site and off-site, four of
            each.   On-site facilities are those located
            at the waste generation site that dispose
            of wastes  from  a  single  firm or process.
                       TABLE 1.  SAMPLING AND ANALYSIS OF INCINERATORS
             Samples collected
Analyses performed
   at all sites
Analyses performed
 at selected sites
Waste feed



POHCs
Chloride
Heating value
Water
Metals
Ash
Viscosity

             Stack gas
             Scrubber liquid
POHCs/PICs
Chloride
Particulate
Continuous monitoring
  (02, C02, CO, HC)
Metals
                          POHCs
                          Chloride
                          Metals
             Ash
                          POHCs
                          Chloride
                          Metals
                                            161

-------
TABLE 2.   DISTRIBUTION OF INCINERATOR TYPES
          AND CONTROL DEVICES
Incinerator type
Number of facilities
Liquid injection
Rotary kiln
Hearth
Gas injection

Control device

None
HC1 scrubber
Particulate control
         2
         3
         1

Number of facilities

         3
         5
         4
 Off-site facilities are commercial dispos-
 ers that handle wastes from a large number
 of sources.  Figure 1 shows the distribu-
 tion  of  capacities for  each of  these
 classes.   The  on-site facilities ranged
 in capacity from 50 to 1,400 kg/h (100 to
 3,000 Ib/h) of waste feed or 1 to 33 GJ/h
 (1  to 31  million Btu/h)   The  off-site
 facilities  ranged  from  600  to  6,000  kg/h
 (1,300  to  13,000 Ib/h)  or 9 to  78 GJ/h
 (9  to 74 million Btu/h).   Besides having
 higher  capacities,  off-site  facilities
 generally  incinerated a  much  greater
 variety of wastes.    Solid  and aqueous
 liquid wastes  were  more frequently han-
 dled  at off-site facilities, and both the
 physical state and chemical composition of
 the wastes  were  much  more  varied  than  at
 on-site facilities.
 WASTE CHARACTERIZATION

      Overall,  three  of  the  eight  facili-
 ties handled  only organic liquid waste;
 two  handled organic and  aqueous  liquid
 waste;  and  the remaining three handled
 organic  and aqueous  liquids and drummed
 solid wastes.

      Figure 2  shows  the variation in some
 waste  characteristics by  type of waste.
 The  heating values were below 4,600 kJ/kg
 (2,000  Btu/lb)  for aqueous  liquids, and
 ranged  from  14,000 to 37,000  kJ/kg  (6,000
 to 16,000 Btu/lb)  for organic liquids and
 from  7,000   to 23,000 kJ/kg  (3,000  to
 10,000  Btu/lb)  for solid  wastes.  The or-
 ganic  liquids  and many  of the solids  had
 a  high enough heating value  to  sustain
combustion without auxiliary  fuel.   Percent
chloride  in  the wastes ranged  up  to 25%,
with the  highest values occurring for or-
ganic liquids  (chlorinated  solvents).   Lit-
tle ash  was  found in  aqueous  liquids;  in
organic liquids values  ranged  up  to  9%;  and
solids had the most ash with values  around
17%.  Obviously, water  content  is  high  for
the aqueous  wastes, but it  also ranged up
to 50 to  60% for the organic  liquids and
solids.
COMBUSTION  PARAMETER   CHARACTERIZATION

     Values for design and operating param-
eters varied widely from facility to facil-
ity.  Every  facility had  unique design or
operating  features.   Multiple  combustion
chambers were  common,  some with separate
waste feeds.   Example  arrangements  were
parallel chambers  (e.g.,  a kiln for  solid
wastes  and  a  separate  liquid  injection
unit) or  primary and secondary chambers
in series.

     Table 3 shows the range  of values  for
three key parameters:   combustion tempera-
ture, residence  time,  and percent excess
air or percent oxygen.   Temperatures ranged
from 820° to 1100°C (1500° to 2100°F).  The
values presented are those recorded on  each
facility's instruments.   They represent a
fair picture of  the range  of  values; how-
ever, when comparing values  from site  to
site it is important to consider the loca-
tion of the  thermocouple  because  the tem-
perature reading  can vary  several  hundred
degrees depending  on its  location.   Resi-
dence times  varied from  0.2  to 6.5 s.
Again, the range is representative, but in-
dividual values are difficult to interpret.
Values  were  calculated  from  combustion
chamber volumes and average flow rates.  At
times a section  of  the  stack  or duct  con-
tained gases at high temperature and might
legitimately have  been  considered  part of
the combustion volume.   At facilities with
multiple chambers at different  temperatures
it could be difficult to select the proper
time/temperature  to consider.  As  Table 3
shows,  excess  air values  fell within  a
range of 60 to 130% and corresponding  oxy-
gen content ranged from 8  to  12%.
                        ORGANIC EMISSIONS

                             From  this varied population  of in-
                        cinerators  with  different designs  and
                                            162

-------
                                               Waste Feed Rate, 1000 Pounds/Hour
                                               1    234    567
                                               I    I     I    I    I     I    I
 10
J
11
 L
 12    13
J	I
                    ONSITE
                    OFFSITE
01
CO
                                                  Heat Input Rate,  10°Btu/Hour*

                                      0    5    10   15   20   25   30   35   40   45   50   55   60   65    70   75
                                      I	I	I	I	I	L_^	I     I     I     I     I    I     I     I     I
                    ONSITE
                    OFFSITE
                           *1  pound/hour - '1.2  kj 1 oj;rani.s/linur ;  1.0^  litu/limir = 0.9 5 C. i ga joul es/liou
                                        Figure 1.  Distribution of Incineration Capacities

-------
HHv,  1000 Bru/Lb
                                                                                 10
                                                                     11
                                                                     12
                                                                    _L
                                                                     13
                                                                      I
                                                                     14
                                                               15
                                                               I
                                                              16
                                                              J
   Aqueous Liquid
   Organic Liquid
   Solid
Chloride, %
                                           10
                                      15
                                      I
                                            20
                                                  25
                                                        30
   Aqueous Liquid
   Organic Liquid
   Solid
Ash, %
                                           10
                                      15
                                            20
                                                  25
   Aqueous Liquid
   Organic Liquid
   Solid
Water, %
0
L
10
 I
20
 I
30
 I
                                           40
50
60
 I
70
 I
80
 I
90
 I
100
 I
   Aqueous Liquid
   Organic Liquid
   Solid
                  *1,000 Btu/lb  =  2,300 kJ/kg
                                          Figure 2.  Waste Characterization

-------
   TABLE 3.   SELECTED PROCESS PARAMETERS
                           TABLE 4.  POHCs FOUND IN WASTE STREAMS*
Temperature,  °C (°F)
Residence time, s
Excess air,  %
Oxygen,  %
820-1100 (1500-2075)
      0.2-6.5
       60-130
        8-12
         POHC
  Number of
waste streams
operating conditions,  a large number  of
data points were  gathered for many  com-
pounds.   Samples  were analyzed  for  all
Appendix VIII   compounds.   Table 4 shows
a list  of the  POHCs most  often  identified
and the frequency with which each compound
was found in  the  22 waste streams burned
for all facilities.  This table lists com-
pounds  contained  on EPA's list of POHCs
that were found at 100 ppm or greater con-
centrations  in a  given waste.   They fre-
quency  of  occurrence  of  carbon  tetra-
chloride (CC14) and trichloroethylene (TCE)
are artificially  high because  these  com-
pounds  were spiked  into the  waste at most
sites.  Excluding these two compounds,  the
most  frequently   encountered  POHCs  were
toluene and tetrachloroethylene.  Twenty-
one other POHCs were  found in one waste
stream  each.

     Destruction  and  removal efficiencies
(DREs)  were  calculated for all POHCs iden-
tified  in the  wastes  at each site.   Fig-
ure 3 displays the  overall range of DREs
measured for  all  compounds and for  a few
specific compounds  where  data were avail-
able for several  sites.   Each data point
represents a  test average for one POHC.
The results  span a wide  range  of DREs,
though  about  35%  of the data points fall
between 99.99  and 99.999% ORE.   Another
20% are just  below 99.99% DRE.  As  with
the  entire  data  set,  the  results   for
specific  compounds  also  cover  a  wide
range.

     The lowest  DREs  shown  in Figure 3
tended  to occur  when  one or both of two
factors were  present.   First,  there  is a
trend  in  the   data  that indicates lower
DREs will result  when POHC concentrations
in  the  waste   are  low; below 1,000  ppm.
This might reflect deviation from the com-
mon assumption that combustion of POHCs is
a first order  reaction, since DRE does not
vary with input  concentration  for a  first
order reaction.   Second,  lower DREs  tended
to occur for POHCs that were identified as
Toluene                          17
Tetrachloroethylene              10
Trichloroethylene                10
Carbon tetrachloride              8
Naphthalene                       7
Chloroform                        6
Methylene chloride                6
Methyl ethyl ketone               6
Phenol                            6
Benzene                           4
Butyl benzyl phthalate            4
bis-(Ethyl hexyl) phthalate       4
Chlorobenzene                     4
1,1,1-Trichloroethane             4
Aniline                           3
Benzyl chloride                   3
Diethylphthalate                  2
Phthalic anhydride                2
21 other POHCs including:         1 each
  Amines
  Chlordane
  Chlorobenzenes
  Chloromethane
  Chloroethanes
  Cresol(s)
  Dimethyl phenol
  Dodecanol
  Hexachlorobutadiene
  Hexachlorocyclopentadiene
  Isocyanates
  Methylene bromide
  Methyl pyridine
                         ""  Total number of waste streams = 22.

                         PICs at one or more sites.   PIC mechanisms,
                         discussed below,  might  increase emission
                         rates resulting in a lower calculated DRE.
                         Seventeen of  the 20  points  on Figure 3
                         that are below  99.99% DRE  were for cases
                         where one or  both of these factors were
                         present.  The other  three  points were be-
                         tween 99.97 and 99.99%.

                              For this program PICs were defined as
                         EPA-listed  compounds  found in  the  stack
                         effluent bot  not  found  in  the  waste above
                         100 ppm.  Some  PICs  were  found at  every
                         site.   Table  5  lists the  PICs  found  and
                         shows the number  of  sites  at  which they
                         were found and the concentrations measured
                         for each test run.
                                           165

-------
                       99
                        L_
               DRE, %
 99 9       99.99      99.999     99.9999    99.99999
__L	I	I	L__	I
All POHCs
       •—£t  •:  •*<•••«•<  .*
CCL
TCE
Tok
                                Figure 3.  DRE Ranges

-------
                         TABLE  5.   PICs  FOUND  IN STACK EFFLUENTS

PIC
Benzene
Chloroform
Bromodichlorome thane
Dibromochlorome thane
Bromoform
Methylene chloride
Naphthalene
o-Nitrophenol
Phenol
Bromochlorome thane
Carbon disulfide
Methylene bromide
2 ,4,6-Trichlorophenol
17 other PICs
Number
of sites
5
5
3
3
2
2
2
2
2
1
1
1
1
1 to 2
Concentrations
(ng/L)
< 0.6-112
l-> 1,000
3-32
1-12
0.2-10
2-27
5-100
25-50
4-22
14
32
18
110
< 10

     Thirteen different PICs  were found at
a concentration of at least 10  ng/L at one
or more sites.  Seventeen other PICs were
found  at  concentrations below  10 ng/L.
There are  several  possible  ways to explain
the presence of PICs  in the stack effluents.
One is that they actually are PICs or com-
pounds formed in the  complex  combustion re-
actions.   A second possibility is that the
compounds  were present in the waste below
detection  levels and were  destroyed at a
relatively low DRE.   A compound present in
the waste  at a concentration  just below the
detection  level of 100 ppm and subject to
99.9% DRE  would be present  in the stack ef-
fluent at  a detectable  level  of  about
10 ng/L.   Thus, it would be  identified as
a PIC.  A  third possibility  is that the
compounds  may have been  introduced  to  the
incineration system from some source other
than tne waste.   An example is  chlorinated
compounds  in city  water used  for scrubbers.
Some of the data indicated  this possibility
was very  likely for some compounds at spe-
cific  sites.  One  or any combination  of
these possibilities could affect the emis-
sion of PICs in any given case.

     Figure 4 compares the  relative concen-
trations   of  POHCs  and PICs measured  for
each test  run.  The concentrations of both
cover a wide range, but the two ranges are
similar.   PIC  concentrations are grouped
more  than the  POHC  concentrations are in
the 1- to  10 ng/L range, with 33 of the 48
points in  that range.
     At selected plants the concentrations
of POHCs were measured in scrubber and ash
effluents.  The data are very limited, but
they indicate that most POHCs were not de-
tected  in these effluents.   When POHCs
were detected  they  tended  to be toluene,
phenol, or naphthalene  at  low concentra-
tions.  Calculations  generally  show that
the quantities  emitted  in  these  effluents
are  small compared  to  the  quantities
emitted from the stack.
OTHER POLLUTANT EMISSIONS

     Carbon monoxide and total hydrocarbon
levels were  also monitored  continuously
during the tests.   Table 6  shows  the  con-
centrations  measured.   The  ranges  were
broad, but the  values  were  relatively low
at  most  sites,  below  15 ppm for  CO  and
below 10 ppm  for  THC.   At  some sites  an
additional test run was made at  altered
combustion conditions;  e.g., lower  tem-
perature  or   different  combustion  air-
flows.  For  one of  these runs,  CO and THC
levels  as  high as  4,000  and 350 ppm,
respectively, were  observed.

     Particulate  concentrations  varied
widely from  site  to site depending  on the
presence and  efficiency of control devices,
the  amount of  ash  suspended in  the  gas
stream, and  other factors.   Table 7  shows
the  range  of average  results for sites
with  and without control devices.  Sites
                                           167

-------
                                                                     Concentration, ng/L

                0.1                            1.0                            10                            100
         PiCs
CTl
00
         POHCs
                                                   Figure 4.  POHC and PIC Concentration in Stack Effluent

-------
TABLE 6.   CARBON MONOXIDE AND TOTAL HYDRO-
          CARBON CONCENTRATIONS IN STACK
          EFFLUENTS
Pollutant  Range of values (test averages)
   CO        1   570 ppm:  most values
                             below 15 ppm
   THC     < 1   60 ppm:  most values
                             below 10 ppm
                         metals  frequently detected.  As the table
                         shows,  nickel,  chromium, barium, and lead
                         were  the metals most often found  in the
                         waste   streams.   Concentrations  varied
                         widely  from barely detectable up to about
                         1%  for  lead in  one  waste.   Stack emissions
                         are also shown on  the  table  and are ex-
                         pressed as  a percentage of the input rate
                         for each metal.  This value is likely in-
                         fluenced by the degree  of  suspension  and
                         control of  particulates  and possibly by  the
                         selective removal  of  specific  metals due to
                         differences in  particle  size  or solubility
                         of  the  metal  compounds  in  scrubber  liquids.
  TABLE 7   PARTICULATE CONCENTRATIONS IN
            STACK EFFLUENT--
                        Range of values
                        (test averages)
 Four incinerators
 without a control
 device
   60  -  900 mg/dscm

(0.03    0.39 gr/dscf)
 Four  incinerators       20  - 400 mg/dscm
 with  control devices  (0.01   0.17 gr/dscf)
 ~';  Corrected to 7% oxygen.

 with  control  devices generally had  lower
 emissions,  though the  ranges  overlapped
 considerably.   This  overlap reflects  the
 variety  of  interacting  factors  that  affect
 particulate emission levels.

      Table  8 shows the  average  test  results
 for chloride emission and  chloride removal
 efficiency.   At  all  sites the emission
 rates were  near or below 2 kg/h (4 Ib/h).
 All  of  those  sites  where  significant
 amounts  of  chlorinated compounds  were
 burned   have  scrubbers  and  typically
 achieved 99% efficiency of better.

      At  four  of the sites  waste  and par-
 ticulate emissions  were  analyzed  for
 selected metals.   Table  9  shows some
CONCLUSION

     Combustion of hazardous waste  in  in-
cinerators is  a  complex  process.   A very
large  number  of  interacting  parameters
potentially  have  an  effect  on  the DRE  and
emission  levels.  Thus,  to  detect trends
or  relationships  is  difficult, and will
require further  data analysis.  Some con-
clusions  that  can be drawn at this time,
however, are:

     1.  A minimum DRE for hazardous waste
incinerators under the combinations of op-
erating conditions encountered is at least
99%.

     2.   DREs  below  99.99% often  result
when the  concentration  of a POHC in the
waste is  less  than 1,000 ppm.

     3.   DREs  below  99.99% occur  more
often  for compounds  typically identified
as PICs than for  other POHCs.

     4.   Chloroform  in  city water  used as
makeup  for scrubbers can  contribute   to
stack  emissions  and  result  in  a  lower  DRE
for that  compound.

     Further data analysis  is  being con-
ducted which will expand on the conclusions
from this data base.  EPA expects  to  have
a  background document with the results of
this analysis  available  by the end of  the
year.
                                            169

-------
                      TABLE 8.  CHLORIDE RESULTS
                               	Emission rate             Efficiency
                               Tb/hkgTh               (%)
Three incinerators
without control devices      0.007-4.3       0.003-1.9

Five incinerators
with control devices           0.3-2.3         0.1-1.1          98-99.9
              TABLE 9.   METALS FOUND IN WASTES AT FOUR SITES
                                     Waste feed               Stack
                Number of       concentration range,      emissions as a
   Metal      waste streams             Hg/g             percent of input
 Nickel            12               0.01-6.670                0.2-53
 Chromium          10                0.1-2.170                0.4-35
 Barium             9                < 1-1,460                0.1-6
 Lead               7                  1-9,830                  2-20
 Cadmium            4               0.01-224                  0.1-2
 Antimony           4               < 10-64                   < 1-7
 Selenium           2              < 100-380                     6
                                  170

-------
                 FLUIDIZED-BED INCINERATOR PERFORMANCE EVALUATION
                          Robert R. Hall, Gary T. Hunt, and Mark M. McCabe
                                       GCA/Technology Division
                                          Bedford, MA 01730
                                           John 0. Milliken
                             Industrial Environmental Research Laboratory
                                 U.S. Environmental Protection Agency
                                  Research Triangle Park, NC 27711
                                             ABSTRACT

  A sampling and analysis program was conducted to assess the performance of a fluidized-bed hazardous waste
incinerator. Union Chemical, Inc., designed and built the incinerator to destroy the organic portion of the wastes
produced by their solvent recycling business. During the test program, 5.2 kg/min (11.5 Ib/min) of wastes contain-
ing 9.6 percent chlorine were burned. The principal volatile chlorinated organics in the feed were 1,1,2-trichloro-
1,2,2-trifluoroethane (3.0 percent), 1,1,1-trichIoroethane (3.4 percent), trichloroethylene (2.4 percent), and tetra-
chloroethylene (3.3 percent). Flue gas samples were collected, for analysis of volatile organics, with the Volatile
Organic Sampling Train and Tedlar bags. Particulate and HC1 emissions were  sampled with a modified Method 5
train. Results are presented in this paper.
INTRODUCTION

  Fluidized-bed incinerators have been used in the
petroleum and paper industries, for the processing
of nuclear wastes and by municipalities for sewage
sludge disposal.  Solids,  liquids, and gases  can  be
burned in fluidized beds. However, the technology is
most applicable to wastes that are difficult to burn
in a simple liquid injection incinerator or to waste
streams that  do  not require a rotary kiln system.
Sludges or slurries  with high  concentrations   of
suspended solids, high water content, low heating
value, or high viscosity  are good  candidates for
fluidized-bed incineration. Short-term fluctuations  in
the above properties can be tolerated by a fluidized
bed without major adverse effects on operating effi-
ciency.


  A field sampling program was conducted on June
30 and July 1, 1982, to assess the performance of a
fluidized-bed hazardous waste incinerator. The tests
were conducted on an 0.82 m (32 in.) diameter incin-
erator under  normal operating conditions. The pri-
mary  objectives of the test program were to:
  1. Determine whether or not the incinerator was
    achieving at least 99.99 percent destruction/
    removal  efficiency  for  selected  difficult-to-
    incinerate, volatile,  chlorinated organic com-
    pounds;

  2. Measure HC1 removal efficiency; and

  3. Measure particulate  emissions.

Three replicate tests were planned.


SITE DESCRIPTION

  Union Chemical Company, Inc., is a commercial
solvent  reprocessing/waste  disposal  business
located  in  Union,  Maine. Waste solvents  are re-
ceived in tank trucks or  drums from generators in
the New England  area. The electronics industries
located in Massachusetts and southern New Hamp-
shire are the major sources of spent solvents. The
2-acre site includes an analytical laboratory; tank
truck loading and unloading facilities; drum  han-
                                                 171

-------
dling and storage areas; thin film evaporators for
reprocessing spent solvents; tanks for handling fuel,
spent solvents, clean solvents, and water; and the
fluidized-bed incinerator. The thin film evaporators
are  used to reprocess nonchlorinated and  chlor-
inated solvents.

  Residues  from the evaporators and  solvents not
suitable  for reprocessing are  burned either in  a
boiler (which helps  heat the evaporators) or in the
fluidized-bed incinerator.  Any chlorinated wastes
are burned in the incinerator, after blending to limit
the chlorine content to less  than about 12 percent.
Potential water runoff from the site is collected and
fed to the incinerator to destroy trace organic com-
pounds.

  Figure  1  is a simplified schematic of the  incin-
erator and the associated emission control systems.
An oil-fired  burner  is used to start up  the  incin-
erator. When  the freeboard temperature reaches
650°C (1200°F), flammable nonchlorinated solvents
are fed to the bed and the oil burner is turned off.
Incineration of chlorinated wastes is started when
the freeboard temperature reaches 1090°C (2000°F).
Typical temperatures  measured, by existing plant
instruments, during the test program, are shown in
Figure  1. Temperature modulation is achieved by
controlling the ratio of combustion air to waste feed,
the dirty water feed rate, and the distribution of
combustion air  between  the  primary  supply and
each of the overfire air injection heights.

  The fluidized-bed combustor is a refractory lined
cylindrical  vessel with a height of 7.3 m (24 ft) and
an inside diameter that varies from 0.81 m (32 in.) to
1.1  m (42 in.). The bed material consists of about
0.6 m (2 ft) of less that 16 mesh silica sand. Combust-
ible wastes are fed to the bed through a single noz-
zle. Dirty water is added  separately. The tempera-
ture profile in the combustor and experience with
other fluidized-bed combustors  indicate that volatil-
    Typical Temperatures
    T, = 743° C
    T2 = 1,180° C
    T3 = 1,220° C
    T4 = 827° C
    T5 = 643° C
    T6 = 68° C
    T7 = 68° C
         Overfire
         Air
                  Freeboard
                 Combustion
                  Chamber
                    Figure 1. Schematic of Union Chemical fluidized-bed incineration system.
                                                 172

-------
ization occurs in the bed while combustion begins in
the bed and is completed in the freeboard.

  During the test program, waste feed and dirty
water feed rates averaged  5.2 kg/min (11.5 Ib/min)
and 4.7 kg/min (10.4 Ib/min), respectively. Heat input
from the waste feed was 2.2 MW (7.6 x 106 Btu/hr).
Oxygen concentration, after the addition of overfire
air, was estimated to  be 9.5 percent based on the
measured stack concentration, stack flow, and dilu-
tion  airflow. Under these  conditions, the  average
gas residence time is estimated at  IVz  seconds
above 1100°C (2000°F). This residence time includes
the combustor and ash dropout.

  Dilution air is added to the reactor vessel to begin
cooling the flue gases. Lime  was formerly added
with the dilution  air for HC1 removal but did not
prove to be effective.  Most of the particulates are
collected  in  the Fisher Klosterman  XQ  high-effi-
ciency cyclone. The combustor ash dropout, reactor,
cyclone, and the top of the quench tower are refrac-
tory  lined.

  Lime slurry is sprayed into the top of the quench
tower to control HCl  emissions. In addition, there
are several spray nozzles that are turned on and off
as needed to control temperature.

  The final control device is a Celicote FRP cross-
flow  wet scrubber containing 1.2 m (4 ft) of Teller-
ette packing. During the test program, some NaOH
was added to the  cross-flow scrubber for final HC1
removal.
FEED STREAMS

  Drums of solvent wastes were premixed and their
contents stored in a continuously stirred 5700-liter
(1500-gallon) tank. The same waste feed was used for
each of the three IVz- to 2-hour test runs. During
each run, samples of the waste feed and dirty  water
feed were  collected, in  volatile organic analysis
(VGA)  vials, at 20-minute intervals. Results were
based on the analysis of one waste feed composite
and three dirty  water samples for each run. The
waste feed rate  was determined from the rate of
change of the level in the storage tank, and the dirty
water feed rate was determined from a flow meter.

  The  waste  feed   samples  were  analyzed  for
volatile organics by a method similar to AlOlb
(Sampling and Analysis  Methods  for  Hazardous
Waste  Incineration) [Harris et al. (1)].  The  waste
feed was dispersed in purified tetraglyme (tetra-
ethylene glycol dimethyl  ether), followed by dilution
in water and standard purge and trap GC/MS pro-
cedures. Dirty water samples were analyzed for vol-
atile organics by standard purge and trap GC/MS
procedures (EPA Method 624, 40 CFR 136).
  Analysis for chlorine, in all streams, was done by
ion  chromatography. Waste feed composites were
prepared for chlorine analysis by Parr oxygen bomb
combustion as outlined in ASTM D808-63. The dirty
water samples did not require preparation for chlo-
rine analysis.

  Results for the waste feed and dirty water feed
analyses are presented in Table  1. The average
waste  feed  contained  3.0 percent 1,1,2-trichloro-
1,2,2-trifluoroethane,  3.4  percent  1,1,1-trichloro-
ethane, 2.4 percent trichloroethylene, and 3.3 per-
cent tetrachloroethylene. The quantity of these com-
pounds in the dirty water was not significant when
compared to the waste feed, as shown in Table  1.
The chlorine content of  the waste feed averaged
9.64 percent and the dirty  water  contained 0.466
percent chlorine.
FLUE GAS EMISSIONS

Particulate and HCl

  A modified Method 5 train was used to determine
particulate  and  HCl   emissions.  The  modified
Method 5 train included impingers that contained
NaOH to trap HCl. The flue gas sampling location
was more than 8 duct diameters downstream and 2
duct diameters upstream from the nearest flow dis-
turbance.  Therefore, only 12 traverse points were
needed as flow disturbances were not significant.
Integrated gas samples  were collected in  Tedlar
bags for  02 and C02 analyses in accordance with
EPA  Method 3. Ion chromatography was used to
measure chlorine in the impingers.

  The average particulate concentration was 1100
mg/dscm (0.49 gr/dscf)  corrected to 7 percent C^-
Recent discussions with Union Chemical represen-
tatives indicate that they have been able to reduce
particulate emissions to 180 mg/dscm (0.08 gr/dscf)
by modifying the quench tower exit and the cross-
flow scrubber.

  The average HCl removal efficiency of the incin-
erator system was  99.73 percent, as shown in Table
2. Mass emissions averaged 0.087 kg/hr (0.19 Ib/hr) of
HCl. These removal efficiencies and emission rates
met with a margin  of safety, the regulatory require-
ments of 99 percent efficiency or 1.8 kg/hr (4 Ib/hr).


Volatile Organic Compounds

  Sampling and analysis for very low concentra-
tions  of volatile organic compounds in a flue gas
matrix at a hazardous waste site required the appli-
cation of some nonstandard or developmental meth-
ods. An integrated gas sample train using Tedlar
bags  and  the  Volatile  Organic  Sampling Train
(VOST) were used.
                                                 173

-------
      TABLE 1. FEED RATES AND COMPOSITIONS-MAJOR CHLORINATED VOLATILE ORGANICS
                                  Run 1
                   Run 2
                                                                                     Run 3
                              See
                           footnotes
g/min
   See
footnotes
g/min
                                       See
                                    footnotes
  g/min
Average
 (g/min)
Waste Feed
 Total                     13.1a         5,940
  1,1,2-Trichloro-
   1,2,2-trifluoroethane      2.8°          166
  1,1,1-Trichloroethane       3.3b          196
  Trichloroethylene          2.2°          131
  Tetrachloroethylene       3.0°          178
  Chlorine                  9.7b          576

Dirty Water Feed
 Total                     14.7C         6,660
  1,1,2-Trichloro-
   1,2,2-trifluoroethane      0.0004°        0.03
  1,1,1-Trichloroethane       0.0032°        0.21
  Trichloroethylene          0.0005°        0.03
  Tetrachloroethylene       0.0002°        0.01
  Chlorine                  0.452b        30.1

Total Feed
           10.3a
            2.6°
            3.4b
            2.4b
            3.0b
            9.5b
            8.34C

            0.00036
            0.00376
            0.00056
            0.00026
            0.466b
             4,670

              122
              159
              112
              140
              445
             3,780
                0.01
                0.14
                0.02
                0.008
               17.6
           11.0a
            3.7D
            3.4b
            2.6b
            4.0b
            9.7fc
            8.26C

            0.00046
            0.00346
            0.00066
            0.00026
            0.480b
4,990

  185
  170
  130
  200
  484
3,750
5,200

  158
  175
  124
  173
  502
4,730
    0.01
    0.13
    0.02
    0.007
   18.0
    0.02
    0.16
    0.02
    0.008
   21.9
1,1,2-Trichloro-
1,2,2-trifluoroethane —
1,1,1-Trichloroethane —
Trichloroethylene —
Tetrachloroethylene —
Chlorine —

166
196
131 -
178
606

122
159 -
112
140
463

185
170
130
200
502

158
175
124
173
524
aPounds per hour based on volumetric flow and a measured specific gravity of 1.12 g/cm3.
 Analysis of composite sample reported as percent.
^Pounds per hour based on volumetric flow and specific gravity of 1.0 g/cm3.
 Analysis of two samples, reported as percent by weight.
e Analysis of three samples, reported as percent by weight.
                                       TABLE 2. CHLORINE RESULTS

Waste Feed
Total, g/min
Chlorine,
percent
g/min
Dirty Water Feed
Total, g/min
Chlorine,
percent
g/min
Flue Gas
Total, m3/min
Chlorine,
/ig/in3
g/min
Removal Efficiency, percent
Run 1

5.940

9.70
576

6,660

0.452
30

76.fi
11,900
0.91
99.85
Run 2

4,670

9.53
445

3,780

0.466
18

105
19,800
2.08
99.55
Run 3

4,990

9.69
484

3,750

0.480
18

94.5
13,100
1.24
99.75
Average

5,200

502

4,730

22

92.0
1.41
99.72
                                                      174

-------
           Glass Wool
                                               Teflon Line
        Stainless Steel
           Probe
                      Glass
                    Condenser
                       Unit
                                                                                          To Pump
                                                             L
                                              Ice Bath        '— Tedlar Bag

                              Figure 2. Integrated gas sampling train.
  Two parallel integrated gas sampling trains were
used as shown in Figure 2. The 20-liter bags were
filled at a constant rate over the test runs. Because
sample degradation was considered to be a potential
problem, the Tedlar bag samples were returned to
GCA's laboratory and analyzed within 48 hours for
run 1, 8 hours for run 2, and 24 hours for run 3. Con-
densate collected before the bags was reserved for
analysis by standard purge and trap  GC/MS tech-
niques. Initial gas analyses were conducted by gas
chromatography with  electron capture  detection
(GC/ECD).  Each  of  the  gas samples was then
pumped over Tenax cartridges and reserved for
thermal desorption GC/MS analysis. Because very
low concentrations of  organics were  anticipated,
only one Tenax cartridge was used for each bag.
The objectives of Tenax cartridge GC/MS analyses
were to confirm the  identity of the four principal
volatile organics,  provide  qualitative and   quan-
titative measurements  of  other  volatile constit-
uents,  and provide data on  both  groups of com-
pounds should the GC/ECD fail to  meet the  detec-
tion limit objectives  because of interference from
other organic compounds or water vapor.

  The GC/ECD analysis was conducted with a 1-per-
cent SP-1000 Carbopack column and an argon/meth-
ane (95/5) carrier gas. A temperature programmed
analysis was  chosen for  the GC/ECD analysis in
order  to provide separation for the lower boiling
point components  while maintaining a reasonable
run time for the higher boiling components. At the
lower end of the calibration curve, temperature pro-
gramming was not compatible with the electron cap-
ture detector. As a consequence, the instrumental
detection limits were higher  than  expected. The
results of the integrated gas sampling train analyses
show that  l,l,2-trichloro-l,2,2-trifluoroethane was
less than 350 /ig/m3, trichloroethylene was less than
1300 /^g/m3, and tetrachloroethylene was less than
750 fig/m3. The above values are strictly a function
of the GC/ECD operation; actual emissions could be
far  below the instrumental detection limits noted
above. During run 1,1,1,1-trichloroethane was below
the detection limit of 200 /j.g/m3 while during runs 2
and 3, 260 to 540 ^ig/m3 were detected. Note that
blank values for the bag samples were not obtained
and that subsequent tests at other sites have shown
that contamination is possible.

  The destruction/removal efficiency, based on the
GC/ECD analysis, for l,l,2-trichloro-l,2,2-trifluoro-
ethane was  greater than 99.980 percent, trichloro-
ethylene was  greater  then  99.903  percent,  and
tetrachloroethylene was greater then 99.960  per-
cent,  as shown in Table  3. Average  destruction/
removal efficiency  for 1,1,1-trichloroethane  ap-
peared to be 99.983 percent. The GC/ECD detection
limits did not allow us to determine whether or not
the incinerator was achieving 99.99 percent destruc-
tion/removal  efficiency. In current field sampling
programs at other sites, we are using a column that
does not require temperature programming and are
achieving detection limits that are one to two orders
of magnitude better, depending on the compound.

  Tenax cartridges that were used to adsorb organ-
ics  in the  Tedlar  bag  samples  were thermally
desorbed into the GC/MS system. However, data
were  not obtained for the four major constituents
because an  early eluting  compound  exceeded the
mass  spectrometer capacity, causing  the system to
shut down. This problem is discussed in more detail
later.

  The second flue gas sampling and analysis meth-
od consisted  of one of the first field trials of the
Volatile  Organic Sampling Train (VOST). This meth-
od involves the collection of organics from the flue
gas on  Tenax  and Tenax/charcoal cartridges, as
                                                 175

-------
                       TABLE 3. GC/ECD ANALYSES OF TEDLAR BAG SAMPLES -
                    CHEMICAL FLOWS AND DESTRUCTION/REMOVAL EFFICIENCIES
                                                Run 1
                        Run 2
                    Run 3
                   Average
l,l,2-Trichloro-l,2,2-trifluoroethane
  Feed rate, g/mina
  Emissions, ftg/m3
            g/min"
Destruction/Removal efficiency, percent

1,1,1-Trichloroethane
  Feed rate, g/mina
  Emissions, /tg/m3
            g/minb
Destruction/Removal efficiency, percent

Trichloroethylene
  Feed rate, g/mina
  Emissions, /ig/m3
            g/minb
Destruction/Removal efficiency, percent
        166
      <350C
        < 0.027
       > 99.984
        196
      <200C
        <0.015
       > 99.992
       131
    <1,300C
        <0.10
       > 99.924
   122
  <350C
   < 0.037
   > 99.970
   159
   325
     0.034
    99.979
   112
<1,300C
   <0.14
  > 99.88
   185
  <350C
   < 0.033
   > 99.982
   170
   480
     0.045
    99.974
   130
<1,300C
   <0.12
  > 99.907
   158
  <350C
   < 0.032
   > 99.980
   175
   270-335
     0.029
    99.983
   124
<1,300C
   <0.12
  > 99.903
Tetrachloroethylene
Feed rate, g/mina
Emissions, /tg/m3
g/minb
Destruction/Removal efficiency, percent

178
<750C
< 0.057
> 99.967

140
<750C
< 0.079
> 99.944

200
<750C
< 0.071
> 99.965

173
<750C
< 0.069
> 99.960
a Based on Table 1.
 Based on the indicated concentrations and gas flows of 76.6 m3/min for run 1, 105 m3/min for run 2, and 94.5 m3/min for run 3.
c Worst case estimate of emission concentration based on detection limit of the GC/ECD system that was used for analysis.
   Sampling Probe
    Ice Water
   Condenser
        Tenax
      Cartridge
 Ice Water
 Condenser
Tenax/Charcoal
   Cartridge
                                                Drying
                                                Tube
                                                                                                 Pump
                        Midget
                      Impingers
                               Figure 3. Schematic of volatile organic sampling train.
                                                    176

-------
shown  in  Figure  3. Each set of cartridges was
replaced with fresh cartridges at 20-minute inter-
vals. Analysis was  conducted by thermal desorption
GC/MS.

  Two types of blank  cartridges were analyzed.
Laboratory or method  blanks consisted of Tenax
and Tenax/charcoal cartridges that were sealed in
screw-cap  culture  tubes and transported with  the
sample cartridges.  The laboratory or method blanks
were not  opened  at the site. Field-biased  blanks
were also  sealed in culture tubes during transport
but were opened for about 30 seconds, and handled
onsite in a manner similar to the handling of  the
sample cartridges. During this program, and other
early EPA-sponsored tests of this sampling train,
problems  were encountered with field  contamina-
tion  of blank samples. The field-biased blanks that
were handled onsite in a manner similar to the test
cartridges showed  high blank values for all test com-
pounds. In some cases,  the blank values exceeded
the sample values. The method  blanks that were
never opened outside the laboratory showed high
concentrations of  1,1,1-trichloroethane. Analytical
results are shown  in Table 4.

  Although the VOST samples cannot  be  used to
quantify emissions, they can  be used  to develop
some useful worst case calculations of emissions and
to place lower bounds on the destruction/removal
                 efficiency. For this purpose, it is assumed that the
                 chemicals on the sample cartridges represent actual
                 emissions. A blank correction is not applied. This is
                 a worst  case calculation because  it assumes that
                 there is no sample contamination although the blank
                 sample cartridges do show high levels of contamina-
                 tion.

                   The  results of the  above calculation are shown in
                 Table 5 for three of the test compounds. Results are
                 not reported for 1,1,1-trichloroethane because some
                 of the  blanks and samples exceeded the maximum
                 calibration range of  the GC/MS. The  VOST train
                 results indicate that  destruction removal efficiency
                 for  l,l,2-trichloro-l,2,2-trifluoroethane was  greater
                 than 99.993 percent,  trichloroethylene was greater
                 than 99.985 percent, and tetrachloroethylene was
                 greater than 99.994 percent.

                   Methylene chloride measurements were also con-
                 ducted but have been affected by especially severe
                 contamination.  The  VOST blanks  and samples all
                 contained quantities of methylene chloride above
                 the calibration range of the GC/MS. The Tedlar bag
                 samples  and the condensate  collected prior to  the
                 bags also contained  relatively large  quantities of
                 methylene chloride. Data on heat of combustion indi-
                 cate that the destruction efficiency of methylene
                 chloride should be similar to the other four test com-
                 pounds  and,  therefore,  emission  concentrations
                            TABLE 4. RESULTS OF VOST TRAIN ANALYSES
Run Number
Method Blank
Field Biased Blank
lAb
lBb
icb
Total Run 1
2Ab
2Bb
2Cb
l,l,2-Trichloro-l,2,2-
trifluoroethaoe
|ng/tube)a
<200
3,700
2,450
1,700
2,000
6,150
1,800
2,140
2,100
Trichloroethylene
Ing/tube)"
<200
> 9,000
5,900
4,400
1,400
11,700
5,030
6,400
2,620
Tetrachloro-
ethylene
(ng/tube)a
<200
7,400
2,860
1,900
460
5,220
2,700
2,970
890
  Total Run 2

    3Ab
    3Bb
    3Cb

  Total Run 3
6,040

1,300
3,600
3,530

8,430
14,050

 3,870
 4,900
  650

 9,420
6,560

2,000
3,300
 200

5,500
 During each run, a total of 60 liters of gas was sampled. Therefore, the total ng per run can be converted to iig/m° by dividing
 by 60.
 Because the blank values were so high, these data represent uncorrected results or worst case emission measurements.
                                                 177

-------
                           TABLE 5. GC/MS ANALYSES OF VOST SAMPLES -
                   CHEMICAL FLOWS AND DESTRUCTION/REMOVAL EFFICIENCIES
                                              Run 1
                Run 2
  Run 3
Average
l,l,2-Trichloro-l,2,2-trifluoroethane
  Feed rate, g/mina                             166
  Emissions, ^g/m3                           < 100C
            g/minb                             < 0.0079
Destruction/Removal efficiency, percent          > 99.9952

Trichloroethylene
  Feed rate, g/mina                             131
                122
               <100C
                 <0.0106
                > 99.9913
                 112
  185
<140C
  <0.013
 > 99.9928
  130
  158
<120C
  <0.010
 > 99.9931
  124
Emissions, /ig/m3
g/minb
Destruction/Removal efficiency, percent
Trichloroethylene
Feed rate, g/mina
Emissions, ^g/m3
g/minb
Destruction/Removal efficiency, percent
<195C
< 0.015
> 99.989

178
<87C
< 0.0066
> 99.9963
<234C
< 0.025
> 99.978

140
<110C
<0.011
> 99.9918
<157C
<0.015
> 99.989

200
<92C
< 0.0087
> 99.9957
<200C
<0.018
> 99 .985

173
<96C
< 0.0088
> 99.9946
fBased on Table 1.
 Based on the indicated concentrations and gas flows of 76.6 m3/min for run 1, 105 m3/min for run 2, and 94.5 m3/min for run 3.
c Worst case estimates of emission concentrations based on VOST train results that have not been corrected for high blank values.
should have been less than 1 percent of the values
that appear to have been measured. If the  methyl-
ene chloride was being emitted from the incinerator,
then  significant concentrations should  have  been
found in the scrubber water, but none was detected.
In addition, methylene chloride has often caused
problems in analysis of water samples because clean
samples may become contaminated from methylene
chloride  in industrial and/or  laboratory  environ-
ments. There is sufficient doubt about the validity of
the methylene chloride results such that useful con-
clusions cannot be developed at this time.
          SUMMARY OF RESULTS
          AND CONCLUSIONS

            A summary of the destruction/removal efficiency
          results for both the integrated gas samples (Tedlar
          bags and GC/ECD analysis) and the VOST is pre-
          sented in Table 6. Conclusions are presented below:

            • The measurement of low concentrations of vola-
              tile chlorinated organic compounds in the  incin-
              erator flue gas  presented  some unexpected
               TABLE 6. SUMMARY OF DESTRUCTION/REMOVAL EFFICIENCY RESULTS11
    Compound
Destruction/Removal
     efficiency
     (percent)
   Emission concentrations

 Mg/m3                ppb
Tedlar Bag Samples
l,l,2-Trichloro-l,2,2-trifluoroethane
1,1,1-Trichloroethane
Trichloroethylene
Tetrachloroethylene
Volatile Organic Sampling Train
l,l,2-Trichloro-l,2,2-trifluoroethane
1,1,1-Trichloroethane
Trichloroethylene
Tetrachloroethylene

> 99.980
99.983
> 99.903
> 99.960

> 99.9931
d
>99.98iT
> 99.9946

<350b
310

-------
  problems. Detection limits of the GC/ECD sys-
  tem were too high to quantitate flue gas concen-
  trations  or  to  demonstrate 99.99 percent de-
  struction. Therefore, GC/ECD detection limits
  were used to place upper bounds on the emis-
  sion concentrations  and lower bounds on  the
  destruction efficiency.  VOST blanks  showed
  high  concentrations of volatile organic com-
  pounds relative to the samples. The VOST sam-
  ple results were used, without blank correction,
  to place  upper  bounds on the emission concen-
  trations  and to place lower bounds on the de-
  struction efficiency.

• Analysis of integrated gas samples by GC/ECD
  shows that destruction/removal efficiency for
  l,l,2-trichloro-l,2,2-trifluoroethane  exceeded
  99.98 percent.  VOST results indicate that dt
  struction/removal efficiency  exceeded 99.99
  percent.

• Analysis of integrated gas samples by GC/ECD
  shows that destruction/removal efficiency for
  1,1,1-trichloroethane was about 99.98 percent.
  This  result does not include correction for  bag
  blank concentrations.

• Analysis of integrated gas samples by GC/ECD
  shows that destruction/removal efficiency for
  trichloroethylene  exceeded  99.90  percent.
  VOST results indicate that destruction/removal
  efficiency exceeded 99.985 percent.

• Analysis of integrated gas samples by GC/ECD
  shows that destruction/removal efficiency for
  tetrachloroethylene  exceeded  99.96  percent.
  VOST results indicate that destruction/removal
  efficiency exceeded 99.99 percent.
  « Methylene  chloride measurements  were  in-
    conclusive.  Extreme sample  contamination is
    suspected.

  • HC1 removal efficiency averaged 99.72 percent,
    and emissions averaged 0.19 Ib/hr.

  • Particulate  emissions averaged 1100 mg/dscm
    (0.49 gr/dscf) corrected to 7 percent On.


ACKNOWLEDGMENTS

  Funding for this program was  provided  by the
U.S. Environmental Protection Agency's Industrial
Environmental Research Laboratory at Research
Triangle Park, NC, under Contract No. 68-02-2693.

  The authors would also like to acknowledge the
excellent cooperation of Raymond  Esposito (Presi-
dent of Union Chemical), John Demaria (Vice Presi-
dent of Union Chemical), and the Union Chemical
operating  staff. Their willingness to let us conduct
this research  program  at their  plant  and their
patience in answering questions about their  facility
were appreciated.
REFERENCE

  1. Harris, J. C., D. J. Larsen, C. E Rechsteiner,
    and K. E. Thrum. Sampling and Analysis Meth-
    ods for Hazardous Waste Incineration (First
    Edition), Draft Report prepared by Arthur D.
    Little, Inc., for the U.S. Environmental Protec-
    tion  Agency under  Contract  No.  68-02-3111,
    Technical  Directive  No.  124,  February  1982
    (Final Report to  be published by EPA).
                                              179

-------
                           FULL-SCALE BOILER EMISSIONS TESTING
                               OF HAZARDOUS WASTE COFIRING
        Carlo Castaldini,  Howard B.  Mason,  Robert J. DeRosier, and Bruce C. DaRos
                                    Acurex  Corporation
                             Mountain View, California  94039
                                        ABSTRACT

     Cofiring of certain hazardous wastes with heat recovery in industrial boilers may be
a promising way to comply with Resource Conservation and Recovery Act (RCRA) provisions
for safe waste disposal  while recovering useful  energy.  To explore the feasibility and
possible environmental  side effects of cofiring, the Incineration Research Branch of the
U.S. EPA is sponsoring  full-scale field sampling and analysis tests of boiler waste
cofi ring.

     Several  boiler design types and waste compositions are being tested to identify
compatible waste/boiler combinations for cofiring.  The sampling protocol includes
continuous monitor measurements of stack concentrations of 02, CO, C02, NOX, and total
hydrocarbons; modified  EPA Method 5 measurements of particulates, and grab samples of
fuel, waste and ash streams.  Principal organic  hazardous constituents (POHC's) are
absorbed onto two organic sorbent traps built into sampling trains.  During post-test
chemical analyses, gas  chromatography/mass spectrometry (GC/MS) is used on all inlet and
outlet samples to allow computation of the destruction and removal efficiency (ORE) of
selected POHC's in the  waste.  Three boilers have been tested:  a 10,000 Ib/hr
wood-waste-fired watertube unit cofired with creosote waste residue; an 8,400 Ib/hr
gas-fired firetube unit cofired with alkyd resin wastewater; and a 230,000 Ib/hr gas-fired
watertube unit cofired  with phenolic residue wastes.  Analytical results for the three
tests are discussed in  this paper.  Most of the  hazardous compounds present in large
concentrations in the waste were destroyed and/or removed to an efficiency in the vicinity
of 99.99 percent.  Destruction and removal efficiencies were generally lower for the
wood-fired unit than for the two gas-fired units.
CONCLUSIONS

     The initial  three tests discussed here
were designed to explore the operational
feasibility and waste destruction
efficiency for routine "as found" boiler
operation typical  of current industrial
cofiring practice.  As such, the boilers
exhibited excursions in excess combustion
air, waste flowrate and load.  No
significant operational problems
attributable to cofiring were observed
during the 1-week duration of each boiler
test.  Boiler thermal efficiency was
generally unaffected by cofiring except in
the one case where the waste had a water
content greater than 90 percent.  This
water resulted in increased latent heat
losses out the stack which reduced boiler
efficiency by 10 percent.

     The destruction and removal
efficiencies varied from 99 to
99.999 percent depending on the boiler
tested and the composition and
concentration of the principal organic
hazardous constituents.  In several cases,
it appeared that compounds present in
higher concentrations exhibited higher ORE
than compounds present in trace quantities.
                                           180

-------
The mass weighted ORE for RCRA Appendix
VIII compounds varied from 99.98 percent
for a wood-fired boiler to 99.998 percent
for a gas-fired watertube.  The combustion
conditions, particularly high excess air,
low load and transient operation, were not
necessarily conducive to efficient waste
destruction.  Higher ORE may be achievable
if a boiler is specifically tuned for waste
cofi ri ng.

     The test results indicate  the
potential for achieving a ORE of
99.99 percent for certain boiler/waste
combinations.  The range of conditions
tested  so far, however,  is narrow.  The
organic waste heat of combustion for all
tests was approximately 9.2 kCal/g,
although one waste was mixed with water
yielding a heat of combustion for the
water-organic waste blend of 0.02 kCal/g.
Any further testing should broaden the data
base to other boiler/fuel/waste
combinations.  Particular data  needs are:
chlorinated wastes with low heats of
combustion; oil- and coal-fired boilers;
and evaluation of the effects of transient
operating conditions on ORE.

INTRODUCTION

     Thermal destruction of wastes by
direct incineration or by cofiring in
boilers, furnaces, or kilns can be an
economical alternative to landfills or
chemical treatment.  Direct incineration of
hazardous wastes is regulated by  Part  264
of the RCRA which was adopted in  January
1981 although boiler cofiring is  currently
exempted from RCRA provisions.  The
incineration rules limit atmospheric
emissions in order to minimize the
environmental effects of solid or liquid
waste disposal.  Specifically, principle
organic  hazardous constituents  (POHC's) in
the waste must exhibit a destruction and
removal  efficiency (ORE) of _>99.99 percent,
defined  as :                 ~~

    ORE  = (Mfeed   Mstack)/Mfeed x 100

where

Mfeed    mass rate of flow of a POHC in the
         waste

^stack   stack mass emission rate of a POHC

POHC's are those compounds listed in
appendix VIII of the May  1980 RCRA
amendments which are present in significant
concentrations in the waste.  This
specification of ORE gives credit for POHC
removal  in ash streams although the ash may
then require evaluation.  For chlorinated
wastes,  the RCRA rules also require control
of chloride emissions.

     Since many hazardous wastes have
significant heating value, they have a
potential  value if cofired with
conventional fuels in boilers for heat
recovery.   This, in principle, can be an
attractive alternative to incineration
while meeting the RCRA goal  of resource
recovery.   The benefits  of cofiring include
partial  replacement of conventional fuels
and the use of existing  boilers in
industries where incinerators are not
available.  Although boiler cofiring is
currently not regulated, the need for
inclusion in RCRA provisions is currently
being studied by EPA.  To support this
study, the tests discussed in this paper
are being conducted as part of the thermal
destruction RSD program managed by EPA's
Incineration Research Branch in
Cincinnati .

     The overall objective of the test
program discussed here is to evaluate the
operational feasibility  and destruction
efficiency of boiler cofiring for hazardous
waste disposal.  The desired output is a
tabulation of boiler design types, fuels,
and operating conditions which will achieve
a satisfactory ORE for a given hazardous
waste.  Toward this goal, field
measurements of ORE are  being made on
selected boiler-waste combinations.  To
limit the field measurement effort to a
manageable size, a parallel  pilot-scale
parametric test program is being conducted
to screen combustion characteristics and
waste characteristics which influence
thermal  destruction.  This pilot-scale
screening is essential because the existing
population of boiler designs and operating
conditions is very diverse, i.e., the
candidate boiler design-fuel-waste
combinations of potential interest number
in the thousands.  Accordingly, the
pilot-scale tests will be used to guide
field test site selection and to help
evaluate the post-test results.  Conducting
the pilot and field efforts in parallel
rather than in series is desirable because
the pilot-scale tests cannot completely
simulate all parameters thought to
influence thermal destruction.  Thus, the
field tests provide a calibration or
judgment aid on the translation of
                                           181

-------
pilot-scale results to full-scale ORE.  The
pilot-scale effort is documented in a
separate paper presented at this
symposium(l).

FIELD TEST PROGRAM

Site Selection

     Initial test sites were selected to
obtain a preliminary evaluation of waste
destruction feasibility over a broad range
of boiler designs and waste
characteristics.  As pilot-scale results
become available, subsequent site selection
will be focused on specific boiler-waste
combinations needed to efficiently
establish the limits of destruction
efficiency.  For the preliminary site
selection, boiler operators firing
hazardous wastes were identified and
facility specifications were obtained.
These candidates were screened based on the
following criteria:

     o   Availability and accessibility of
         the boiler for testing

     «   Proportion of hazardous
         constituents in the waste
         according to RCRA Appendix VIII
         specifications

     »   Estimated degree of difficulty in
         destroying POHC's efficiently

     »   Degree to which the boiler-fuel
         type is representative of future
         widespread use for waste cofiring

     An earlier EPA study(2) was used for
guidance in evaluating the third and fourth
criteria.  Candidates ranking high in at
least three of the criteria were evaluated
further by a pretest site survey.  Where
exact waste compositions were not
available, waste samples were obtained and
analyzed.  From this procedure, three sites
were selected:

     9   Site 1 -- 4,500 kg/hr
         (10,000 Ib/hr) Keeler watertube
         fired with wood bark and cofired
         with creosote waste

     «   Site 2 -- 3,800 kg/hr
         (8,400 Ib/hr) Cleaver Brooks
         firetube fired with natural gas
         and cofired with alkyd resin paint
         waste
     •   Site 3 --  105,000  kg/hr
         (230,000 Ib/hr) Babcock and Wilcox
         watertube  fired with natural gas
         and cofired with phenol wastes

Design and operating characteristics are
summarized in Table 1.  Tests at these
sites were conducted during the period of
March to July 1982.  Chemical analyses of
the samples from these tests are available
and discussed here.  Additional boiler
tests are in progress.

Sampling and Analysis Protocol

     Four test runs were made at each site:
a baseline test with conventional fuel  only
and no waste, and triplicate cofiring runs
with the boiler operated at approximately
the same conditions.  The baseline test was
run to establish the difference in
emissions due to cofiring.  For all runs,
the boiler was operated as  in routine
practice.  No attempt was made to constrain
the boiler to a fixed load  or excess air
level beyond routine practice.  This was
done so that the DRE's measured would be
representative of real-world operation
rather than exemplary conditions.

     The sampling protocol  was designed to
obtain POHC concentrations  at all influent
and effluent streams.  Grab samples of the
waste feed, and (for the wood-fired boiler)
the fuel, bottom ash, and particulate
collector hopper ash were taken throughout
each of the 5-hour  test runs and composited
for analysis.  Continuous monitor
measurements of 02, NO, CO, C02, and
unburned hydrocarbons were  made for the
flue gas.  Sampling was done upstream of
the preheater, if present.  The continuous
monitors as specified in Table 2 were
housed in a mobile  sampling laboratory.
These measurements  are useful to augment
control room board  data as  a continuous
record of boiler operation.  NOX and CO, in
particular, are quite sensitive to
combustion conditions, and  NOX (decreasing)
or CO (increasing)  may correlate with
conditions which promote POHC
breakthrough.

     Organic constituents were sampled
using two separate  trains extracting flue
gases from ports in the stack.  Low
molecular weight volatile organics were
sampled using a Tenax train depicted in
Figure 1.  Flue gases were  drawn at
200 cc/min through  the Tenax tube
                                           182

-------
                  TABLE 1.   BOILER DESIGN AND OPERATING CHARACTERISTICS
              Site
  Design  characteristics:

  9  Manufacturer/boiler
     type
  9  Heat  input  capacity,
     MWt  (1015 Btu/hr)

  a  Rated steam capacity,
     103  kg/hr (103 lb/hr)

  *  Air  pollution
     control  device

  Operating characteristics:

  *  Primary fuel

  a  Waste fuel
9  Waste fuel  heat of
   combustion, AH
   Kcal/g (Btu/lb)

•  Waste fuel  moisture
   content,  percent

»  Operating steam load,
   percent of  capacity

*  Waste fuel  heat input,
   percent of  total

»  Excess combustion
   air,  percent

»  Volumetric  heat
   release rate,  MW^/m3
   (103  Btu/hr-ft3)

*  Estimated furnace
   bulk  residence time, t" sec

»  Outlet streams sampled
Keeler CP-308      Cleaver Brooks    Babcock and Wilcox
watertube-stoker   250-hp firetube   watertube balanced
balanced draft                       draft
3.5
(12)

4.5
(10)

Multiclone
Woodwaste (hog)

Creosote sludge


9.28
(16,700)


24t


100


40


100
                                  300
                                  (29)
                                                   2.9
                                                   (10)

                                                   3.8
                                                   (8.4)

                                                   None
Natural gas

Alkyd resin
wastewater

0.02*
(35)
                                                     >99
                                                     25
                                                     30
                   750
                   (72)
                                                     O.Rtt
                  82
                  (280)

                  105
                  (230)

                  None
                                                                     Natural gas

                                                                     Phenoli c resi due
                                                                       9.2
                                                                       (16,600)
                                  Flue gas, bottom   Flue gas
                                  ash, flyash
                  25


                  37-39


                  85


                  78
                  (7.5)


                  >2tt


                  Flue gas
* Typical,  although  highly variable as result of nonhomogenei ty of organics in wastewater.

t Portion  of moisture content is attributed to highly volatile organics.

'"'Gross  estimate to indicate relative magnitude of bulk residence time in each boiler,  t
 calculated by V/0  at a furnace temperature of 1,650K (2,500°F).
                                           183

-------
                   TABLE 2.   CONTINUOUS  MONITORING EQUIPMENT
Instrument
NO
NOX
CO
C02
TUHC
02
Principle of
operation Manufacturer
Chemil uminescence Thermo Electron
Nondispersi ve ANARAD
infrared (NDIR)
Nondispersi ve ANARAD
infrared (NDIR)
Flame ionization Beckman
Fuel cell Teledyne
Instrument
model Range
10 AR 0-100 ppm
0-500 ppm
0-1,000 ppm
0-5,000 ppm
500R 0-1,000 ppm
AR500 0-20 percent
400 0-1,000 ppm
0-5 percent
0-25 percent
                                   ROTOMETER
        STAINLESS
        STEEL ..
        PROBE \
SOURCE
 QUICK
 STAINLESS
 STEEL
 DISCONNECTS
n
  TENAX
  POLYMER
 /-PACKED
/ TUBE
'(2mm I.D.
 6mm O.D.
 x 25 cm)
                 D
                                                                         200 CC/MIN
                                                         GAS METER
               Figure 1.   Porous polymer vapor  sampling  method,

                                      184

-------
containing approximately 2g of sorbent
until  the total  sample gas volume exceeded
2 liters.  Higher molecular weight organics
(>Cy with boiling range >110°C) were
sampled using an XAD-2 organic sorbent
module positioned at the back of the EPA
Method 5 train as shown in Figure 2.  A
quantity of about 65g XAD-2 was packed in
the glass module through which flue gas was
extracted isokinetically at approximately
200 ml/s.  A total  sample volume of >5 dry
standard cubic meters (dscm) was
extracted.

     In the analytical laboratory, the
Tenax traps were thermally desorbed at the
head of a packed column gas chromatograph
(GC) by heating to 180°C.  Volatile
compounds desorbed from the Tenax traps
were subsequently separated by heating the
GC column and were detected in a mass
spectrometer  (MS) operating in a continuous
scanning mode.  With this technique
detection limits were normally 50 ng per
sample.  Modified Method 5 samples  (probe
catch, filter, XAD, and impinger solutions)
were extracted separately for base/neutral
and acid semivolatile and nonvolatile
compounds using methylene chloride  in a
soxhlet apparatus.  Extracts were combined
and concentrated.  Organic analyses were
performed by  GC/MS using a fused silica
capillary column (FSCC).  Waste fuel
organics were extracted using liquid/liquid
extractors also with methylene chloride,
concentrated, and analyzed by capillary
GC/MS.  Detection limits for semi and
nonvolatile organics were generally 1 to
 5  ng per sample.  Analyses of duplicate,
blind  spiked, and blank samples were
performed as  part of the quality
assurance/quality control  (QA/QC)
protocol .
 FIELD TEST RESULTS

 Site 1 Results

     The wood-fired boiler at site  1 was
 operated at full load with the creosote
 waste contributing approximately  40 percent
 of the total heat input to the boiler  and
 wood contributing approximately 60  percent.
 Table 3 summarizes semivolatile and
 nonvolatile organic flowrates measured  in
 the flue gas and combined ash streams
 during the baseline test.  The compounds
 listed are those which showed a significant
 concentration in either the  flue  gas for
 the baseline test, or in the creosote
waste.  Organics which are part of the RCRA
Appendix VIII hazardous compound list are
grouped separately from other semivolatile
and nonvolatile organics.  Hazardous
organics detected in the flue gas during
combustion of wood only were found to be
phenol, 2,4-dimethylphenol, naphthalene and
nitrobenzene.  For the most part, these
compounds were also found in each ash
stream.  Additionally, pentachlorophenol
was also detected in the flyash.  This is
attributed to wood partially contaminated
with creosote.  Due to the relative
magnitude of the stream, organic flowrates
in the flue gas were much higher than in
either or combined ash streams.

     Table 4 summarizes organic compound
flowrates and flue gas-based DRE's
calculated for one of the three cofired
test runs.  Results from the other two test
runs are similar to the ones shown here.
DRE's were calculated based on flue gas
emission rates over and above those
attributable to wood only.  That is, the
difference in flue gas emissions between
the cofiring test and the baseline test was
used in calculating the actual ORE of
compounds in the feed.  The weighted
average ORE for Appendix VIII compounds was
less than 99.99 percent.  Only phenol was
found to exceed 99.999 percent destruction.
DRE's of most other organics also ranged
between 99.9 and 99.99.  A trend toward
lower DRE with lower waste concentration
may be deduced from these results.

     Volatile organic analyses showed the
presence of toluene in all three cofired
tests with emission flowrates varying from
27 to 480 yg/s (8 to  140 pg/dscm).  Other
volatile compounds appearing in only one or
two tests were benzene at 450 ug/s,
ethylbenzene at 330 Mg/s, and chlorobenzene
at 14 yg/s.

Dioxin Testing

     Additional site  1 analyses were done
to screen for the presence of chlorinated
dibenzodioxins since  these compounds are
often associated with creosote and penta-
chl orophenol .  The overall screen was posi-
tive, so Battelle Columbus Laboratories was
contracted to quantitate specific isomers
and to speciate 2,3,7,8 TCDD.  Results
for one of the runs are summarized in
Table 5.  In the waste, dioxins tended to
predominate in the higher isomers -- hepta-
CDD and octa-CDD.   In the flue gas, the
reverse was evident; dioxin concentration
                                            185

-------
CB
CTl
                                               STACK T.C.
                                                                       CONVECTION OVEN
                                                                                                 - FILTER
                           STAINLESS
                           STEEL  SAMPLE
                           NOZZLE
                                                                                                    1/2" TEFLON LINE

                                                                                                                GLASS BALL JOINTS
                                                                                                                                    WATER  JACKETED CONDENSER
GLASS  LINED
STAINLESS  STEEL
PROBE  ASSEMBLY
                                              S-TYPE PITOT TUBE
                                                                                                                                               THERMOREGULATOR/
                                                                                                                                               RECIRCULATOR
                                                                                                                                               0.1°C
                                                                                	I
                                                                                            COURSE ADJUSTMENT
                                                                                            VALVE
                                               GAS METER  T.C.
                                                                                                                                           CONDENSER  WATER BATH
                                                                                                                                           (60°C)  /-BACK-UP FILTER
                                                                                 HEAVY WALL
                                                                                 VACUUM LINE

                           ORIFICE AH
                           MAGNEHELIC
                           GAUGE
                                                                      VACUUM PUMPS
                                                                      (10 ft3/M EACH)
              PROPORTIONAL
            . TEMPERATURE
              CONTROLLERS
|	CONJROJ^MODULE	D_RY_TEST_METER	|
                                                                                                                                  SILICA GEL DESICCANT -

                                                                                                                                                ICE BATH
                                                              Figure 2.   Modified  EPA  Method  5  train.

-------
            TABLE 3.  SITE 1 -- BASELINE TEST  EMISSIONS  (WOOD  WASTE  ONLY)
             Compound
                                               Flue  gas
 ppb
ug/s
Bottom ash
 (ug/kg)
                                                                                 Elyash
RCRA Appendix VIII

  Naphthalene
  Pentachlorophenol
  Phenol
  2,4-Dimethyl phenol
  4-Nitrophenol
  Fluoranthene
  Nitrobenzene

Other semivolatiles and nonvolatiles

  Biphenyl
  Pyrene
  Phenanthrene
  Fluorene
  Anthracene
  Methylpyrene
  Dibenzofuran
  Cyclopenta(d,e,f)phenanthrene
  Acenaphthene
  Acenaphthylene
  Diphenylamine
  2-Nitrophenol
  4-Chloro-m-cresol
  Methylnaphthalene
  Methylphenol
  Hydroxymethoxybenzaldehyde
  Benzoic acid
 1.5
 Nn
18
 0.83
 ND
 ND
 0.28
 25
 <3.2
220
 13
 <0.6
  4.5
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.90
0.60
0.07
1.7
3.0
3.1
<6.4
<0.6
<0.6
<0.6
<0.6
<6.4
0.6
<6.4
<0.6
<3.2
<6.4
16
11
1.3
24
58
49
   320
    NO
    80
    NO
    ND
    ND
    60
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          ND
                          20
                          ND
                         380
                          ND
1
  200
1,200
  100
   40
   40
   ND
   ND
                              ND
                              ND
                             ,300
                              20
                              20
                             180
                             200
                              ND
                              20
                              20
                              40
                              ND
                              ND
                              40
                              ND
                              ND
                              ND
   Not detected.  For flue gas, the concentration was  below  0.03  ppb  for  most  compounds
   except for pentachlorophenol , 4-nitrophenol,  biphenyl,  methylpyrene,
   cyclopenta(d,e,f)phenanthrene and acenaphthyl ene  for  which  ND  is  less  than  0.3 ppb.
   For ash streams ND is generally less than  2.0
                                         187

-------
           TABLE 4.  SITE 1 -- DRE'S FOR COFIRING WOOD  WASTE  AND  CREOSOTE
                                                             Flue  gas
Compound
Creosote
waste
(ug/ml )*

ppb

ug/s
DRE
(percent)
  RCRA Appendix VIII

    Naphthalene                                6,000
    Pentachlorophenol                           2,200
    Phenol                                        800
    2,4-Dimethylphenol                            360
    4-Nitrophenol                                 200
    Fluoranthene                                   60

    Weighted average DRE                       . .  .

  Other major semivolatiles and nonvolatiles"*"

    Biphenyl                                   8,400
    Pyrene                                     8,400
    Phenanthrene                                7,000
    Fluorene                                   5,000
    Anthracene                                 3,000
    Methylpyrenes                               1,600

    Weighted average DRE                       . .  .

  Other minor semi volati les and nonvolatiles^

    Dibenzofuran                                  880
    Cyclopenta(d,e,f)phenanthrene                620
    Acenaphthene                                  160
    Acenaphthylene                                140
    Diphenylamine                                 120

    Weighted average DRE                       . .  .
2.0
0.77
0.20
<0.19
0.79
0.83
<0.30
0.54
2.9
1.5
0.58
0.04
1.5
0.08
1.2
<0.03
0.06

35
28
2.5
<3.2
15
23
<6.3
15
70
34
14
1.3
34
2.5
26
<0.6
1.3

 99.997
 99.975
>99.999
>99.98
 99.85
 99.2

>99.98
>99.999
 99.996
 99.980
 99.986
 99.991
 99.998

 99.992
 99.92
 99.992
 99.67
>99.991
 99.98

 99.93
*To obtain feedrate in yg/s multiply yg/ml concentration by  50 ml/s.

tMajor compounds whose measured concentration was greater than  1,000  ppm  (1,000
 in the creosote waste.

"^Minor compounds with <1,000 ppm concentration'.
                                        188

-------
tended  toward  the  lower isomers  -- tetra-
CDD and penta-CDD.   This suggests  the
possibility  that during the combustion
process the  higher isomers  in  the  waste  are
reduced to  lower isomers.   This  would
account for  the  low ORE (negative  in the
case of tetra-CDD)  for the  lower isomers.
The overall  ORE  for dioxins was  comparable
to the  ORE  for the waste as a  whole.

     Continuous  monitor results  for site 1
showed  large variations in  operating
conditions  over  the test run.   Flue gas  (^
levels  varied  from 6.2 to 16.7 percent with
a corresponding  rise in CO  from  230 to
>1,000  ppm.   These abnormally  high levels
of excess air  will  probably produce some
quenching in the flame which would also
promote breakthrough of organics.   The ORE
values  exhibited in these tests  may be
improved if  the  boilers were tuned for
higher  efficiency.

Site 2  Results

     The site  2  firetube boiler  was
operated at  a  load of 2,000 Ib steam/hr,
which is approximately a quarter load.
This off-load  operation would  produce
opposing effects on ORE of  reduced
temperature  and  longer residence time.
Table 6 summarizes emissions measured
during  the  baseline test run with  only
natural gas  combustion.  Detected
semivolatile and nonvolatile organics  were
found to have  flowrates less than  4 pg/s
corresponding  to flue gas concentrations
between 1 and  4 ppb.  Toluene  and
naphthalene  were the only two  POHC's
detected.  Table 7 summarizes  the  organics
flowrates and  DRE's for one of the three
cofired test runs  using alkyd  resin waste
water.   Heat input attributable  to the
waste was less than 1 percent  based on
waste composition  generally consisting of
about 99 percent water.  Here, weighted
average DRE's  of the POHC's and  other
organics in  the  wastewater  generally
exceeded 99.99 percent.  The other two
cofired test runs  show similar semivolatile
and nonvolatile  organic emission results.
Volatile organics  detected  during this  test
were methylene chloride and hexane.
Methylene chloride  is a suspected contami-
nant because it was not detected in the
other cofiring tests.  Hexane is a product
of incomplete natural gas combustion
through flame quenching.

Site 3 Results

     The 230,000 Ib/hr watertube boiler at
site 3 was operated off-load at
60,000 Ib/hr with four of the six burners
in service.  During cofiring the phenolic
residue was injected through one of the
four burners.  The other three operated on
gas only.  Combustion air was fed through
all six burners.  Phenolic residue
contributed approximately 40 percent of the
total heat load to the boiler.  To improve
the heat absorption profile in the boiler
at low load, the excess air was increased
to a level corresponding to 10 percent 02-
Baseline test results in Table 8 show that
phenolic, phthalate and benzene compounds
were the major organic emissions with flue
gas concentrations generally less than
2 ppb.  ORE during cofiring, summarized in
Table 9, generally exceeded 99.99 percent,
even at this off-spec operation.  For many
of the compounds, including phenol  the
major POHC, the DRE exceeded 99.999.

ACKNOWLEDGMENTS

     This work was sponsored by the U.S.
Environmental Protection Agency under
contracts 68-03-3043 and 68-03-3176.
George L. Huffman and Robert A. Olexsey
were the Project Officers, and assisted in
all phases of the project.

REFERENCES

1.  Wolbach, C. D., et al., May  2-4,
    1983.  Subscale Parametric Studies on
    the Combustion of Hazardous Wastes.
    Presented at the Ninth Annual Research
    Symposium on Solid and Hazardous
    Wastes, Cincinnati, Ohio.

2.  McCormick, R., et al., January 1982.
    Engineering Analysis of the  Practice of
    Disposing of Hazardous Wastes in
     Industrial Boilers.  Draft report on
    Task SCA04, EPA Contract
    No. 68-03-3043.
                                            189

-------
               TABLE 5.   SITE 1 DIOXIN RESULTS

Isomer
2,3,7,8 TCDD
Tetra-CDD
Penta-CDD
Hexa-CDD
Hepta-CDD
Octa-CDD
Emissions
Waste
ND
ND
0.10
18
180
170
(pg/s)
Flue gas
ND
0.14
0.044
0.024
0.017
0.014

Bottom ash
NA
ND
ND
ND
0.0048
0.014

ORE
(percent)
ND
. . .
56.0
99.87
99.99
99.99
  Total  COD
370
0.24
0.020
99.94
NA -- Not analyzed.
ND -- Compound concentration was below the detectable limit
      in the waste (0.0032 yg/s) or in the bottom ash
      (0.000088 yg/s).
    TABLE 6.  SITE 2 -- BASELINE TEST (NATURAL GAS ONLY)
                                            Flue gas outlet
               Compound
                           ppb
                            yg/s
  RCRA Appendix VIII

    Toluene                                 1.7        2.3
    Naphthalene                             1.3        2.4
    Pentachlorophenol                        ND        <0.23

  Other semivolatiles  and nonvolatiles

    Xylenes                                 2.4        3.7
    C3-Alkylbenzenes                        0.20       0.32
    C4-Alky1benzenes                        0.09       0.18
    Benzoic acid                            ND        <0.45
    Benzaldehyde                            0.60       0.90
    Ethylbenzaldehyde                        3.7        2.4
    1,2-Benzenedicarboxylic acid            ND        <0.45
    Methyl-1,1-biphenyl                     ND        <0.05
    Methylnaphthalenes                      0.18       0.36
ND = Not detected.  Detection limit corresponds to 0.06 ppb
     for pentachlorophenol, 0.26 ppb for benzoic acid,
     0.20 ppb for 1,2-benzenedicarboxylic acid and 0.02 ppb
     for methy-1,1-biphenyl .

                             190

-------
           TABLE 7.  Site 2 -- ORE'S FOR COFIRING NATURAL  GAS  AND  ALKYD  RESIN
Compound
Al kyd resin
wastewater
(yg/ml )*
Flue gas
ppb yg/s
ORE1"
(percent )
  RCRA Appendix  VIII

    Toluene

    Naphthalene

    Pentachlorophenol


    Weighted  average DRE


  Other semivolatiles  and nonvolatiles

    Xylenes

    Benzoic  acid

    C3-A1kyl benzene

    C/I.-A1 kyl benzene

    Methyl -1,1-bi phenyl

    1,2-Benzenedicarboxylic acid

    Methyl naphthalenes

    Benzaldehyde


    Weighted  average ORE
0.3-550 (186)

0.07-1.7 (0.62)

0.02-0.29 (0.18)
268-2,760 (1,770)

310-1,190 (870)

2-405 (145)

0.10-46 (16)

0.-0.7 (0.25)

0.-0.32 (0.20)

0.-0.17 (0.06)

0.-0.08 (0.03)
 1.2     1.6    >99.999

 1.2     2.1    >99.8

<0.05   <0.2    >97
                                     >99.995
 3.4     5.1     99.997

<2.7    <4.6    >99.98

 0.77    1.3     99.98

<0.02   <0.04   >99.99

<0.02   <0.02   >99.5

<0.15   <0.35   >94

 0.26    0.53    91

 0.80    1.2     70
                                     >99.990
*Range  in  concentration  is  based on analysis results of three separate waste residue
 samples  collected  during the test.  Concentration in parenthesis represents arithmetic
 average.   To  obtain  feedrate in yg/s multiply by 33.5 ml/s.

TORE  is  based  on  average concentration of compound in the alkyd resin wastewater.
                                           191

-------
    TABLE 8.  SITE 3 -- BASELINE TEST  (NATURAL GAS ONLY)

Compound
RCRA Appendix VIII
Phenol
Oibutyl phthalate
Bis(2-ethylhexyl )phthalate
Other semi vol ati les and nonvolatiles
C3-A1 kyl benzene*
C3-A1 kyl benzene*
Benzene acetaldehyde
2,2' -Methyl enebi s phenol
4,4' -Methyl enebi sphenol
Pyrene
Cl5H16°*
C15H160*
C18H20
Cl6H18°
Emi
ppb

0.63
0.32
1.7

2.1
ND
ND
0.01
0.04
ND
ND
ND
ND
ND
ssions
ug/s

32
49
370

140
<8
<8
1.6
5
<2
<8
<8
<8
<8
*Different isomers.

ND   Not detected.   Detection limit corresponds to about
     0.10 ppb or 8 yg/s.
                             192

-------
    TABLE 9.   SITE 3 -- ORE'S FOR COFIRING NATURAL GAS AND PHENOLIC RESIDUE

Compound
RCRA Appendix VIII
Phenol
Dibutyl phthalate
Bis(2-ethylhexyl )phtha!ate
Weighted average DRE
Other major semi vol atiles and
Cg-Al kyl benzene''"'"
C3-A1 kyl benzene"1"1"
Benzene acetaldehyde
Cl6H18°
C18H20
C15H16Ott
C15H16Ott
Weighted average DRE
Other minor semi vol atil es and
2,2' -Methylenebisphenol
4,4' -Methylenebi sphenol
Pyrene
Weighted average DRE
Phenol ic
residue
(ug/ml )*

S3, 000
120
31
nonvol atiles1"
100,000
45,000
5,000
>38,000
>34,000
7,500
3,000
. . .
nonvol atiles*
460
370
4
. . .
Flue gas
ppb ug/s

0.30 28
1.4 220
1.2 260

0.93 61
•C0.10 <8
<0.12 <8
<0.07 <8
<0.06 <8
<0.07 <8
<0.07 <8
. . . . . .

<0.07 <8
<0.07 <8
0.05 5
... ...

DRE
(percent )

>99.999
99.4
>99.98
>99.998

>99.999
>99.999
>99.999
>99.999
>99.999
>99.999
>99.999
>99.999

>99.993
>99.991
99.5
>99.990
*To obtain feedrate in ug/s multiply concentration in ug/ml by 250 ml/s.



'Major compounds whose measured concentration was greater than 1,000 ppm.



^Different isomers.



#Minor compounds with <1,000 ppm concentration.
                                       193

-------
                     SUBSCALE PARAMETRIC STUDIES ON THE COMBUSTION
                                  OF HAZARDOUS WASTES
                Carlo Castaldini,
                           Howard
      Andrew R. Garman, Jeffrey M. Kennedy,
      B. Mason, C. Dean Wolbach*
       Acurex Corporation
Mountain View, California  94042
                                       ABSTRACT

     Thermal  destruction of wastes by cofiring with conventional fuels in boilers,
 urnaces, or kilns have several  attractions.  The ability of such devices to destroy
wastes to environmentally acceptable levels is not well documented.  This study looks at
the effects of operational  variables of a subscale furnace on time-temperature regimes
in order to estimate the impact of the variables on destruction efficiencies in
full-scale units.  Over 50  different combinations of firing rate, excess air rate,
waste-to-fuel  ratio, waterwall  surface area, burner swirl setting, and waste type have
been investigated.

     Preliminary information suggests that excess air rate and firing rate will play the
most significant parts in establishing destruction efficiencies.
INTRODUCTION

     Thermal destruction of wastes by
direct incineration or by cofiring with
conventiooal fuels in boilers, furnaces,
or kilns is one of the most effective
methods currently available for disposal
of hazardous organic material.  While
direct incineration of hazardous wastes
is regulated by Part 264 of the Resource
Conservation and Recovery Act (RCRA) as
adopted in January 1981, boiler cofiring
is currently exempt from RCRA
provisions.  However, the potential for
boiler cofiring regulations is being
evaluated by the Environmental
Protection Agency (EPA).  To support
these efforts, EPA's Incineration
Research Branch (IRB) in conjunction
with the Office of Solid Waste is
conducting a research and development
programs on incineration effectiveness
and regulatory impact analyses of which
the tests discussed in this paper are a
part.
                        The overall objective of this test
                   study is to evaluate the effectiveness
                   of boiler cofiring as a means of
                   hazardous waste destruction and removal.
                   This evaluation will correlate such
                    arameters as boiler design types,
                   fuels, and operating conditions which
                   will achieve a  desired destruction and
                   removal efficiency (ORE) for a given
                   principal organic hazardous component
                   (POHC) in the waste.  POHC's are  those
                   compounds listed in Appendix VIII of the
                   May 1980 RCRA amendments, and ORE is
                   defined as:
                          ORE =  (Mfeed   Mstack)/Mfeed  *  100
                   where
 ""Principal Author
                          Mfeed    Mass  flowrate  of a  POHC
                                   in  the  fuel

                          ^stack   Stack mass  emission rate
                                   of  a  POHC

                         Although  it  is  intuitively  known
                    that  certain boiler  parameters will have
                    a  major  influence on  waste  destruction,
                    there is  no available collected  body of
                                           194

-------
data to estimate the magnitude of these
influences,  either separately or
collectively.   Time, temperature, and
turbulence are the fundamental physical
characteristics affecting destruction
efficiency.   These in turn are determined
by such operating parameters as burner
configuration  and swirl, excess air rate,
firing rate, fuel-to-waste ratio, waste
effects on flame temperatures, cooling
surface-to-combustion volume ratio, etc.

     The purpose of the pilot-scale
parametric study is to estimate marginal
destruction  efficiencies as a function of
selected operational  parameters and to
attempt to find methods of correlating
the results  of pilot tests to full-scale
units.  This in turn will  ultimately
assist industry and regulators in
permitting the cofiring of wastes in
boilers and  minimizing the need for
full-scale field tests.

     This paper briefly describes the
pilot-scale  parametric test program being
conducted, the results obtained to date,
and the facility and experimental
equipment employed.  It should be noted
that this is part of a larger integrated
study being  conducted by Acurex to first
                    screen the combustion  and  waste
                    characteristics which  influence  thermal
                    destruction, then compare  with direct
                    in-field measurements  of ORE on  select
                    boiler-waste combinations,  and lastly
                    prepare a semi empirical computer model
                    of thermal destruction.
                    PROGRAM DESCRIPTION

                         The current pilot-scale parametric
                    test program is divided  into four tasks:

                         «   Baseline time-temperature tests

                         «   Single-compound parametric
                             tests

                         a   Multiple compound parametric
                             tests

                         o   Mathematical simulation

                         The variables under study  in each
                    of the test series are summarized in
                    Table 1.  When completed, a total of
                    74 independent conditions will  have been
                    tested, with 58 replicated.  All
                    destruction efficiency samples  are to be
                    taken in either duplicate or triplicate,
                         TABLE 1.  SUMMARY OF TEST PROGRAM VARIABLES
              Task
       Baseline tests
       Single-compound
       tests
       Multiple compound
       tests
  Independent variables
>»  Flame basket shape
»  Load
<»  Excess air
'»  Waterwall surface-to-
   volume ratio

a  Waste/fuel ratio
3  Flame basket shape
»  Load
9  Excess air
a  Waterwall surface-to-
   volume ratio

9  Six compounds
9  Four extreme conditions
   selected  from single-
   compound  tests
    Dependent variables
,»  Temperature profile
<»  Residence times
o  Exit gas concentrations
a  Velocity profile
   Temperature profile
   Residence times
   Exit gas concentrations
   Destruction efficiencies
   Same as for single-
   compound tests
                                           195

-------
and test compound materials are mixed
with the base fuel (No. 2 distillate oil)
before firing.

     The purpose of the baseline tests
is to define the thermal  environments of
the test facility under various operating
conditions, to quantify the effects of
machine operational variables on the
thermal environment, and to gather data
to assist in the development of a
mathematical simulation.   The 10
baseline firing conditions are summarized
in Table 2 which include two firing rates
(0.23 and 0.37 MW), two excess air rates
(10 and 25 percent), three swirl settings
(maximum, median, and minimum) and two
waterwall surface-to-volume ratios (0 and
0.12).

Compound Tests

     The facility ORE was studied using
chlorobenzene under the same firing
conditions as the baseline tests.
Chlorobenzene was fired at two
waste-to-fuel ratios (10 and 5 percent
v/v).  In the multiple compound tests,
                                   five additional compounds will be fired
                                   at each of four conditions (10 percent
                                   v/v at conditions I, IV, VII, and X).
                                   The compounds are tentatively:

                                        o   Acrolein

                                        »   Pentachlorophenol

                                        «   1,2-Dichloroethane

                                        9   Ethyl acrylate

                                        o   Acrylonitrile

                                   (Note:  At the time of preparation the
                                   feasibility of firing mixtures was being
                                   studied)

                                        A mathematical  simulation of the
                                   furnace zone is being assembled.  The
                                   simulation will predict a time-temper-
                                   ature profile for the bulk gases, and
                                   estimate destruction efficiency.  It is
                                   expected that a boundary layer
                                   destruction efficiency calculation will
                                   be included.
              TABLE 2.   BASELINE FIRING CONDITIONS
                Firing rate
Condition   MW (million Btu/hr)
                                               Excess
                                              Air rate    Swirl    Waterwall
                                                 (%)     setting     ratio
I
II
III
IV
V
VI
VII
VIII
IX
X
0
0
0
0
0
0
0
0
0
0
.37
.37
.23
.23
.23
.23
.37
.37
.23
.23
(1
(1
(0
(0
(0
(0
(1
(1
(0
(0
.25)
.25)
.8)
.8)
.8)
.8)
.25)
.25)
.8)
.8)
10
25
10
25
25
25
10
25
10
25
7
7
7
7
5
2
7
7
7
7
.5
.5
.5
.5
.0
.5
.5
.5
.5
.5
0
0
0
0
0
0
0
0
0
0






.25
.25
.25
.25
                               196

-------
RESULTS AND DISCUSSION

     Time, temperature, and turbulence
(mixing) are the key to understanding the
phenomena of thermal destruction.  The
detailed calculation of destruction
performance becomes impractical when
attempts are made to apply them to "real
world" situations such as the destruction
of wastes in boilers.  Therefore
semi empirical  procedures are necessary  if
one wishes to estimate how changes in
operational variables such as firing
rate, excess air rate, flame turbulence,
and waste-to-fuel ratio aeffect both the
thermal history of the waste and the
destruction efficiency of the unit.

     The facility used in this study is
shown schematically in Figure 1 and
described in detail  in the following
section.  It is basically a subscale
furnace that approximates linearly a
44-MW packaged "D"-type watertube boiler.
Since it is important in estimating
destruction efficiency outside the flame
to characterize the thermal  environment,
the tests to date have concentrated on
the specific parameter effects on
temperature profiles.  These are
discussed in the following subsections.
        Temperature  Profi1es

              For  each  firing  condition,
        temperature  profiles  were  taken  at five
        cross  sections  (locations  A through E
        shown  in  Figure  1).   A  typical  set of
        profiles  as  shown  in  Figure 2 for
        condition  IV indicates  a boundary layer
        temperature  zone within 4  inches  of the
        furnace wall.  Axial  profiles are shown
        in  Figure  3.   The  importance of  this
        boundary  layer cannot be over emphasized
        at  this time,  since it  represents greater
        than  43 percent  of the  volume or  thermal
        environment  of the furnace.   Because
        temperature  and  velocity plays  such a
        large  role in  volume  flowrate and
        destruction, this  zone  will  be
        significant  when considering destruction
        efficiencies for this furnace.
        Preliminary  calculations indicate that
        approximately  20 percent of the  volume
        flow  may  be  in this region.

        Gross  Effects  of Firing Conditions on
        Temperature  Profiles

               The center!ine axial  temperature
        profiles  for the four basic  firing
        conditions are shown  in Figure 4.   Over
        the range  of conditions tested,  the
       CONVECTIYE
       SECTION
                                               FIREBOX

                                                C
> 	 A
( }
i
i
i
i
1 O O (J
i
i
L _i H



i '
ii
ii
_LI J
) O O C



1
> O O C
I



::
JJ J
) O C



! n
it i '
ii i
) O (



1 '1 11
'1 11
h n
l U. li.
:> o o







                                                                           BURNER
                                                                           FACE
                      447
                   CM.
_290_ _


 114
                                               90
_L68_

 66
                                                                           CM.
                                                             4?
                                                                          0 IN.
        Figure 1.  Simplified schematic of subscale  furnace  showing  locations  of
                   temperature measurements.
                                           197

-------
            2,300 -
                                                                        - 1,250
                                                                         .1,200
                                                                          1,000
                                     12      17       12     8

                                       DISTANCE  FROM WALL (IN.)
        Figure  2.   Cross-section  temperature  profiles for  condition  IV,
               2,200 -
                2,000
                1 ,bOO
                1,500
                1,400
O -- CONDITION IV
A -- CONDITION IV BURNING 10 PERCENT
    V/V CHLOROBENZENE
                                                                    1,200
                                                                    1,150
                                                                    1,100
                                                                    1,050
                                                                    1,000
                                 40      60     80    100      120

                                  DISTANCE FROM FRONT WALL (IN.)
                                                                    850
                                                                    800
                                                                    140
Figure 3.   Axial  temperature  profiles for  walls,  in boundary  layer,  and  along
             center line  (range  of measured  values  shown  by  "error"  bands).
                                           198

-------
                 2,500
                                  DISTANCE FROM FRONT WALL (IN.)
                 2,000
Figure 4.   Typical centerline  axial  temperature profiles  for four  baseline conditions.
                                                                              1,200   E
         2,100
         2,000 -
         1,900
                             8               16  16             8

                                RADIAL DISTANCE FROM WALL (IN.)
Figure 5.   Effects  of excess  air on temperature profiles (0.8  x  106 Btu/hr  and maximum
            swirl).
                                             199

-------
furnace front end temperatures varied
over approximately 270°F while the exit
temperatures varied over 300T-  Mean
residence times varied from approximately
3.4 to 5.5 seconds.  The heat intensity
in the flame zone varied from approxi-
mately 140,000 to 220,000 Btu/ft-3 hr.
This is equivalent to the range seen in
industrial boilers.

Effect of Excess Air on Temperature
Profiles

       There were no qualitative
surprises from the effect of excess air
on temperature profiles (see Figures 4
and 5).  Increasing excess air lowered
the temperature profiles.  At the higher
firing rate (1.25 x 106 Btu/hr) a change
of excess air from 10 to 25 percent
decreased the axial temperature about
100°F at the front of the furnace and
160°F at the rear.  At the lower rate
(0.8 x 106 Btu/hr) the decrease at the
front was again about 100°F, but at the
rear only about 50°F.

       Qualitatively these results will
have opposing effects on destruction
efficiencies.  At the higher excess air
rate the increased oxygen content will
promote destruction while the decreased
residence time and temperature will
dampen destruction rates.  Modeling has
not yet been conducted  to  determine  if
there is a peak efficiency with  excess
air rate.

Effect of Swirl Setting on Thermal
Profile

Burner swirl setting, a measure  of
the ratio of axial to radial air
momentum, has a marked  impact on  flame
shape but only a moderate  impact  on
temperature profile (see Table 3  and
Figure 6).  The flame shape was  estimated
from visual  observations and empirical
values for flame volume.  Table  3 shows
the visual estimates of flame length and
diameter under conditions  IV
(0.8 x 106 Btu/hr).  The calculated flame
length using the observed  flame  diameter
for two different geometries is  also
shown.  The volume of a flame is
empirically stated to be -0.2 m^/io^
watts* or -0.0586 m3/106 Btu.  For
simplicity of calculations, a flame is
usually considered to be a cylinder.
However, observations tend to favor a
truncated cone or frustrum.  A frustrum
*Lucas, D. M. and Toth, H. E., "The
 Calculation of Heat Transfer in the Fire
 Tubes of Shell Boilers"; Journal of the
 Institute of Fuel; October, 1972.
              TABLE 3.   FLAME GEOMETRY ESTIMATION FOR BASELINE CONDITION IV
                 Flame diameter   Flame length
                      (cm)             (cm)
        Swirl        (visual          (visual
       setting     estimate)        estimate)
     Cylinder
       flame
    length (cm)
  (calculated)*.t
   Frustrum
     f 1 ame
  length (cm)
(calculated)*^
7.5

5
2.5
75

50
42
±7

±7
±7
16
\
50
66
+ 5

±8
±8
11

25
37
±2

±7
±12
18

43
63
±3

±12
±20
     *Volume   (0.2 rn^/lQ^ watts)(0.29307 watts/Btu/hr)(0.8 x 1Q6 Btu/hr)
               0.0469 m3

     tLength   4V/^D2, where  D   flame diameter

     ^Length = 48V/7TTD2, for D} = 2D2
                                          200

-------
          2,300 -
          2,200
          2,100 -
          2,000 _
          1,900
                                                                         - 1,100
                                           16 16

                             CROSS-SECTION DISTANCE FROM HALL
Figure 6.  Effects of burner swirl setting on temperature  profiles  (0.8  x  10^  Btu/hr  and
           25 percent excess air).
with DI - 2D2 fits our observed flame
lengths quite well .

     Swirl  settings  for this burner can
range from 0 to 8.  The swirl settings of
7.5 (condition IV),  5.0 (condition V),
and 2.5 (condition VI) represent a short,
bushy flame, an intermediate flame, and a
long thin flame,  respectively.  From the
observed profiles in Figure 6, the effect
of swirl is to steepen the temperature
gradient as one goes to higher settings
(shorter flames).  However, the overall
profiles range ±25°F at the centerline,
or marginally outside measurement error
range.

     The intuitively expected results
would include a reversal  of the order of
the temperature profiles if, indeed, the
thermal gradient does change.  That is,
the firing condition giving the higher
temperature profile  at the front of the
fire box should give the lower profile at
the rear of the box.  This considers only
the bulk heat content of the gases and
does not take into account changes that
may arise from changes in the mass
flowrates due to the shorter length of
heat release.  Modeling of this
phenomenon has not been completed at this
time.

Effect of Waste on Temperature Profile

     For baseline condition IV (nominal
0.8 x 106 Btu/hr heat input, 25 percent
excess air, and swirl setting of 7.5),
the effect of adding 10 percent v/v
chlorobenzene in the fuel appears to be
less than the noise of the various
measurements (see Table 4).  The heat
content of the fuel is 135,000 Btu/gal
while for chlorobenzene it is 112,000
Btu/gal.  A 10 percent v/v mixture will
have a heat content of 133,180 Btu/gal,
representing a 1.7 percent decrease  in
the rate of heat input.  The accuracy of
preparing the mixture is ±4 percent  while
the accuracy of the feedrate is
±5 percent.  The precision of the
temperature measurements is ~±20°F for
points near the walls and ±10°F for
points in the bulk gas.  The conclusion
is that the chlorene radical is not
measurably changing the thermal profile
beyond the flame zone.
                                           201

-------
            TABLE  4.   EFFECT OF 10 PERCENT CHLOROBENZENE ON TEMPERATURE PROFILE
                      (FIRING CONDITION IV)
                                           Temperature (°F)
                           Cross-section
            Cross-section D
Distance
from wal 1
(in.)
0
1
2
3
4
8
12
Condition IV
1,968
2,113
2,135
2,172
2,194
2,186
2,199
Condition IV
(10% waste)
2,027
2,120
2,135
2,168
2,185
2,189
2,205
Condition IV
1,680
1,823
1,946
2,019
2,041
2,064
2,095
Condition IV
(10% waste)
1,781
1,870
1,969
2,002
2,045
2,060
2,102
Effects of Waterwalls on Temperature
Profiles

     Addition of waterwall  panels
replaces about 8 percent of the
refractory surface area.  The waterwalls
are approximately 6 in.  wide and run the
length of the furnace on either side.
The total surface area of the waterwall
is 8 ft2.  The two waterwalls are each
composed of three panels.  The panels  are
individually supplied with water, and  the
inlet and outlet temperatures are
monitored.  The total water flowrate is
also measured.  The walls extract
approximately 400,000 Btu/hr from the
furnace.

     An example of the effect of the
waterwall on longitudinal temperature  is
shown in Figure 7.  Not only does the
waterwall lower the overall temperature
profile, but it also steepens the slope.
From heat flow calculations the surface
temperature of the walls is estimated  to
be 77°C (170°F).  This correlates with
the observed condensation on the rear
panels, the implication being that the
surface temperature of the panels range
from above to below the flue gas dew
point (~55°C or 130°F).

Destruction Efficiencies As a Function of
Operating Variables

     At the time of submittal of this
paper, thirteen destruction efficiency
tests had been concluded.  Of these,
eleven were with chlorobenzene, and two
were with a mixture of carbon tetra-
chloride, chloroform, and methylene
chloride.  A summary of the results is
shown in Tables 5 and 6.  The relative
effects are displayed in Figures 8
and 9.

     It is readily apparent that the
effect of the waterwalls is to decrease
the destruction efficiency by 1.5 to 2
orders of magnitude.  Whether this
effect is caused primarily by flame
cooling or by cooling of the thermal
oxidation environment is not known at
this time.  It is postulated that the
primary effect is from the latter based
on the (unsubstantiated) assumption that
only 90 to 99 percent of the destruction
occurs within the flame.  Other
                                          202

-------
      2,500 -
      2,000 -
      1,500
      1,000 -
                                                                     160
                                     Axial  distance (in.)
Figure 7.  Changes in temperature profiles with (run 21) and without  waterwalls
           (runs  11  and  19)  for condition IV, (0.8 106 Btu/hr, 25  percent  excess
           air).  Runs  11  and 19 were performed 1 month apart.
        TABLE 5.  DESTRUCTION  OF  CHLOROBENZENE UNDER VARIOUS OPERATING
                  CONDITIONS
Run no.
12
14
23
20
22
11
19
21
16
10
Condition
I
II
II Wt
III
III W
IV
IV
IV W
45%*
45%
Ib/hr In
6.2
5.9
7.0
4.0
2.3
4.6
4.0
2.7
3.9
4.4
Ib/hr Out x 106
1.2 to 1.6
0.10 to 0.52
14 to 50
0.2 to 1.7
6.9 to 27
1.7 to 2.3
1.0 to 10
>76 to >180
0.18 to 0.87
0.4 to 2.2
MO/M!* x io7
1.9 to 2.6
0.17 to 0.89
20 to 72
0.5 to 4.2
30 to 140
3.8 to 5.1
2.6 to 26
>280 to >670
0.5 to 2.2
1 to 5
      *Range of three  test  points
      tWith waterwall s
          percent excess  air  rate  and 0.8 million Btu/hr

                                       203

-------
     TABLE 6.   DESTRUCTION OF  CHLORINATED METHANES
                 UNDER  TWO FIRING CONDITIONS
Run no.
25


24


Condition Compound
III W CH2C12
CHC13
CC14
IV W CH2C12
CHC13
CC14
Mg/M^xlO6
25t
22t
140 to 175
52 to 74
4.9 to 15
10 to 47
    *Range of  three test points

    tSingle point
       10"
      10
       10
        -7
       10
        -8
                8
                O -- C.8 MMETl'/HR
                     WITH WATERWALLS

                • -- 1.2 MMBTU/HR
                     WITH WATERWALLS

                D -- 0.8 MMBTU/HR
                     WITHOUT WATERWALLS

                • -- 1 .2 MMETLi/HR
                     WITHOUT WATERWALLS
a

a
                10
                           EXCESS AIR (',)

Figure  8.   Chlorobenzene breakthrough  versus  excess  air
                              204

-------
            III,
                      CH19HOPENZENE

                      M[THVL[N[ CHLORI3E

                      CHLOROrORM

                      CARBON TETRACHLORIDE
CHLORC3ENZENE

METIIYLENE CHLORIDE
CHLOROFORM

CARBON TETRACHLORIDE
 ©*

©*
                           10"
                                        10
                                         -6
                                                    10
                                                      -5
                                                                 10"
                                                                             10
                                                                               -3
                *SINGLE DATA POINT
       Figure 9.  Relative  destruction  efficiencies under two firing conditions
                  (chlorobenzene  data  from runs 21 and 22.  Data for chloromethanes
                  from  runs  24  and  25).
correlation with operating  parameters  may
be present, but insufficient  data  is
available at this time.  The  majority  of
data at this time has been  taken without
waterwalls which pushes the detection
limits (Ca 1-10 ng/trap).   Thus the error
associated with calculating destruction
efficiencies (sampling and  analytical
measurements) blurs smaller effects.
Comparisons of chlorobenzene  breakthrough
versus excess air rate is shown in
Figure 10.

Comparison of Chlorobenzene Breakthrough
jo NO Emissions and Temperature"

     Plots of chlorobenzene breakthrough
versus NO emissions and temperature at
one point are shown in Figures 10  and  11.
Although no absolute correlation can be
discerned for the NO plot,  it appears
that an upper bound on breakthrough may
be present.  The upper bound  behaves or
would be expected from theory.  That is,
because NOX production is proportional  to
temperature, and breakthrough is inversely
proportional to temperature,  as NOX
increases, breakthrough should decrease.
This is what is observed.   The relation-
ship of breakthrough to temperature is
                          more clearly  seen  in  Figure  11.   Again,
                          this does not  reflect the  true  situation
                          because it does  not correlate  the true
                          spatio-temperal  temperature  history to
                          breakthrough.  Further data
                          interpretation is  underway to  look at
                          these aspects  in more detail.

                          Future Efforts

                               Additional  testing by sampling in
                          the firebox  is projected.   Both
                          center!ine and boundary layer  sampling
                          will be done.  In  addition,  coal, gas,
                          and residual  oil will  be used  as fuels.
                          Finally, alternate injection methods
                          including multiple burners and  centerline
                          injection is  postulated.

                          Conclusions

                               A subscale  test  bed for studying the
                          thermal destruction of organic  materials
                          as a function  of boiler operating
                          parameters has been established and
                          characterized.   Ten baseline operating
                          conditions have  been  studied covering
                          changes in firing  rate, excess  air rate,
                          and burner swirl settings.  Destruction
                          efficiency tests for  chlorobenzene are
                                           205

-------
                          40    60    80   100   120   140    160

                                        PPM NO
          Figure 10.   Chlorobenzene  breakthrough versus  NO  emissions.
;g
- £j
-
: c
-I



p

O 10', EXCESS AIR
D 20 EXCESS Alt!
O 455 EXCESS AIR

D
a
8
>
-
—
:
i

I?
I
i
i
i i

0
3 °

>


o
I
o
9
?
A
                           1300  IbOO  1700   1900  21.00  2300  2MO
                                    TEMPERATURE CF) »T D 1
Figure  11.   Chlorobenzene breakthrough versus  center!ine  temperature  at cross
             section D-2.
                                        206

-------
                TABLE 7.   ACUREX PILOT-SCALE RESEARCH FURNACE CAPABILITIES
     Combustion Chamber
     Corrective Section
     Air Supply
   33 in. I.D. x 10 ft long
   2- and 3-ft sections
   Viewports
   2-in. diameter access ports for sampling or air staging
   Refractory lined (3,000°F)
   Optional  cooling surface
   Coal-fired 1 to 1.5 million Btu/hr
   Gas-fired 2.5 million Btu/hr
   Oil or COM 2 million
   Horizontal orientation
e  Removable heat exchange sections -- tailor flue
   temperature profile
«  Water cooled
   Preheated to 600°F
   Nominal 8 psi supply
   Optional 125 psi supply
   Up to 150 percent excess air
   Measurement and control of airflow
     Burner
     Flue
«  Variable swirl research burner

   --  Multiple fuel capability
   —  Variable axial fuel tube placement
   --  Capability to change velocity
   --  Capability to change quarl design

»  Capability for multiple burner design

•  I.D. fan
e  Baghouse
•  Potential for flue gas recirculation
being conducted.   Preliminary results
indicate the potential  to predict the
thermal  environment and possibly an upper
limit on destruction efficiency.

EXPERIMENTAL

Pilot-Scale Furnace

     The furnace  facility is shown
schematically in  Figure 1 with
capabilities listed in  Table 7.   The main
combustion  firebox is a horizontal,
refractory-lined,  cylindrical tunnel,
33 inches in diameter and 12 feet in
length.   The firebox can accommodate a
set of stainless  steel  water-cooled panels
to simulate gas temperatures in  a watertube
boiler.   At the firebox exit is  a
                    10-foot vertical convective cooling
                    section where U-shaped watertube bundles
                    extract heat from the burnt gases.  An
                    induced draft fan and damper at the
                    15-foot stack allow precise variation of
                    firebox pressure.

                         The burner is mounted on the front
                    wall of the furnace.  It can be manually
                    adjusted to operate over a large-angle of
                    swirl geometries.  The axial position of
                    the fuel injection tube can be varied
                    within a water-cooled quarl to ensure
                    flame stability.  Fuel oil is injected
                    through a Delavan pressure-atomizing,
                    hollow-cone spray nozzle.
                                           207

-------
Fuel/Air Monitoring

     Liquid fuel flow is measured to
±2 percent absolute accuracy via a
Fischer & Porter rotameter.  Most of the
uncertainty is associated with viscosity
changes in the oil  as the temperature
rises.  Viscosity/temperature
relationships for the fuel used were
supplied by Truesdail Laboratories, Los
Angeles, California.  A large fraction of
the fuel is recirculated to the fuel
supply drums.  The balance is pumped
through the fuel tank.  The oil-pressure
behind the nozzle, calibrated against
diesel oil volume flowrates, gives an
independent verification of the firing
rate.  Agreement between pressure gauge
and rotameter is generally better than
3 percent.

     Diesel oil/chlorobenzene mixtures
are prepared in 50-gallon quantities in
steam-cleaned drums.  Diesel oil volume
is measured by a Neptune "Red Seal" Model
no. 137, positive displacement totalizing
flowmeter to ±1 percent accuracy.
Chlorobenzene is hand-pumped through a
Tuthill "Fill-Rite" Model no. 800A
totalizing flowmeter to ±2 percent
accuracy.  The mixture is recirculated
through the fuel feed system for 1 hour
to ensure adequate mixing.

     Airflow is monitored continuously
by a hot-wire anemometer mounted in a
venturi tube.  The air is preheated to
350°F  to produce entrance velocities
characteristic of full-scale oil-fired
furnaces.

     A  stainless steel probe downstream
of the convective section extracts gas
samples for online major species
analysis.  The samples are cooled, dried,
and sent to an emissions bench where
electro-optical analyzers determine the
levels of C02, 02, CO, NOX, and unburned
hydrocarbons.  C02 and Q£ concentrations
give a  second verification of input
stoichiometry.  Airflow measurement
and flue gas C02 determination of excess
air levels agree within 4 percent.

Temperature Measurement

     Since a  steady-state thermal
condition requires several days of
furnace operation, temperature data must
be collected  during  transient operation.
To ensure repeatable temperature  (and
hence destruction efficiency)  data,  the
furnace is fixed on natural gas for
12 hours overnight.  The gas firing  rate
is fixed at 0.7 million Btu/hr and the
stoichiometry at 50 percent excess air.
The time is adequate for the furnace gas
and walls to cool down to  temperatures
repeatable within ±50°.  The appropriate
fuel oil conditions are set the following
morning, and oil is fired  for 3 hours
±15 minutes.  Using this approach,
temperature data are not biased by
position on the heat-up curve.

     During oil-firing, 12 key furnace
wall and flue gas temperatures are
monitored by Chrome!/Alumel
thermocouples.  Traverses  with a
ceramic-sheathed piatinum/platinum-
10-percent Rhodium thermocouple provide
gas phase axial and radial profiles in
the firebox.  Thermocouple radiation
losses are accounted for by calibration
against a section pyrometer.

Organic Sample Collection

     Samples for organic analysis are
collected in the flue gas  duct between
the convective heat recovery section and
the induced draft exhaust  fan.  Duct gas
temperatures are nominally 150°C
(approximately 300°F).   The sampling
train consisted of a 1/4-in. SS probe, a
Teflon heat traced sample  line, a Tenax
trap (1.5 x 12 mm containing
approximately 1.5g Tenax)  cooled  in an
ice bath, a dry knock-out  trap, a wet
impinger containing 0.1N NaOH solution, a
silica gel impinger, a diaphragm  pump,
and a calibrated rotameter.  The  gas
temperature from the Tenax trap is
monitored, and sampling rates are
200 ml/min for 10 to 20 minutes.  The
specific retention volume  for
Chlorobenzene on Tenax at  120°C is given
as 2.3 m3 (A. D. Little, Characterization
of Sorbent Resins for Use  in Environ-
mental Sampling, U.S. Department  of
Commerce PB 284 347, March 1978,  p.  22).

Organic Sample Analysis

     Tenax tubes are desorbed  in  a
reverse direction  at 200°C with
approximately  100 ml N2 onto a  1  percent
SP-1000 carbopak B 60/80 mesh  gas
chromatographic column held at 60°C.
Following the  thermal desorption  of  the
                                          208

-------
trap the temperature is ramped 20°C/min
to 220°C and held for 20 minutes the
detector is a Hall electrolytic
conductivity detector set in the halogen
mode.

     The detection limit of the method is
0.1 ng per Tenax trap.  Excellent
reproducibility (RSD of 6.5 percent) was
obtained at concentrations of 100 ng
chlorobenzene spiked directly onto the
Tenax trap.  The recovery averaged
92 percent at 100 ng, 85 percent at 10 ng
per trap and 70 percent at 1.0 ng spiked
onto the Tenax trap and desorbed as
indicated above.  The detector response
was found to be linear in the range of
1  to 100 ng range.  Reproducibility of
retention time was tO.l percent for a
series of five traps spiked with
chlorobenzene at 1 to 100 ng, analyzed
sequentially.
ACKNOWLEDGMENTS

     This work was performed under  EPA
Contract no. 68-02-3176; Task 31.   The
Project Officer is George Huffman.
Sampling was conducted by M. Murray while
analytical work was carried out  under the
supervision of Dr. V. Lopez-Avila.
Technicians contributing significantly  to
the effort were R. Grose and W.  G.
Hellier.  Critical technical input  was
given by Dr. L. Waterland of Acurex
Corporation and members of the  IRB  staff.
A  special note of appreciation  goes to
the Acurex technical publications  staff.
                                           209

-------
                        EVALUATION  of HAZARDOUS  WASTE INCINERATION
                        in  a CEMENT KILN at  SAN  JUAN CEMENT COMPANY

                              James A.  Peters, Thomas W.  Hughes
                                Monsanto Research  Corporation
                                      Dayton Laboratory
                                     Dayton, Ohio  45407
                                            and
                                    Robert E.  Mournighan
                           U.S.  Environmental  Protection  Agency
                                     Cincinnati, Ohio
INTRODUCTION

     An attractive alternative to  hazard-
ous waste incineration  which  makes use
of a waste's heat content is  cofiring of
hazardous waste in high temperature indus-
trial processes.   Many  such  processes,
which include cement and dolomite  kilns,
glass furnaces, steel  furnaces, and some
industrial  boilers, provide  conditions of
temperature and residence time similar to
those required for dedicated  hazardous
waste incinerators.  In addition to savings
derived from the  heat value,  the use of an
existing industrial process  requires less
capital to process a given amount  of haz-
ardous waste than does  a new  incinerator.

     Because of their large  process cap-
acities and energy use, cement kilns, in
particular, appeared to be an excellent
example of this concept.  They typically
operate at temperatures over 1,260°C
(2,300°F), gas residence times are in ex-
cess of 1.5 seconds at  or above this tem-
perature, and the combustion  zone  is highly
turbulent.  The alkaline environment in a
cement kiln absorbs acid gases such as
hydrogen chloride  (HC1), and the need for
exhaust gas scrubbing systems is elimina-
ted.

     Because of the need to gather more
data on cement kiln incineration of
hazardous waste,  the Environmental Pro-
tection Agency (EPA) Industrial Environmen-
tal  Research Laboratory-Cincinnati conduc-
ted a comprehensive demonstration program
at the San Juan Cement Company in Dorado,
Puerto Rico, from October 1981 until  Febru-
ary 1982.   The purpose of the demonstration
program was to evaluate the ability of a
cement kiln to destroy wastes, to generate
data for the purpose of formulating permit-
ting criteria, and to evaluate a kiln's
ability to remove the HC1 combustion prod-
uct.  A sampling program was conducted dur-
ing the burning of hazardous waste.  The
primary goal of the program was to evaluate
the effects of various waste parameters on
the destruction efficiency of Principal
Organic Hazardous Components (POHCs), and
the change in emissions of particulate
matter, hydrocarbons, carbon monoxide (CO),
sulfur dioxide (S02), nitrogen oxides
(NOX), and hydrogen chloride (HC1).  Sec-
ondary objectives were to detect and
quantify the Products of Incomplete Com-
bustion (PICs), chlorinated dibenzofurans,
chlorinated dibenzodioxins, trace metals,
and heavy organics.

     A number of cement plants have been
used to test cofiring of hazardous wastes,
e.g., St.  Lawrence Cement Company in Mis-
sissauga, Ontario, in 1974-1976; Peerless
Cement Company in Detroit, Michigan, in
1976; Stora Vika Cement Plant near
Stockholm, Sweden, in 1978; and Marquette
Cement in Oglesby, Illinois, in 1981.  The
                                           210

-------
data from these tests indicated that cement
kilns,  when properly operated, could de-
stroy many of the organic chemical com-
pounds  in the wastes burned under the oper-
ating conditions of these tests.

     A need remained within EPA for criter-
ia to be used in deciding whether or not
to regulate cement plants that burn wastes,
particularly those including highly chlor-
inated hydrocarbons.

     The facility at the San Juan Cement
Company was chosen for the test because of
its availability, the willingness of the
company to cooperate with the Puerto Rico
Environmental Quality Board (PRJCA) and
U.S. EPA Region II in obtaining permits,
and the suitability of the facility for a
test.

Facility and Process Description

     San Juan Cement Company has  operated a
cement plant in Barrio Espinosa in Dorado,
Puerto Rico, since 1970.   The location of
the plant is about 27 kilometers  (km) west
of San Juan.  The plant is dedicated to the
manufacture of portland cement.  Annual pro-
duction averages 4.082 x 108kilograms (kg)
(450,000 tons) per year and the plant em-
ploys 350 workers; it is  the second lar-
gest cement plant in Puerto Rico.

     A cement kiln is the heart of the
cement process, as depicted in Figure 1.
A cement kiln is a large steel horizontal
tube with refractory linings.   Such kilns
may be up to 7.6 meters (m) in diameter and
over 232 m long; at San Juan Cement, kiln
#2 is 137 m (450 feet) long with  an outer
shell diameter of 3.05 m (10 ft)  and 0.305
m (12-inch) thick walls.   The kiln rotates
slowly  (75 rotations per hour) and has a
gentle  slope (0.3 m/10 m length)  to allow
material  to pass through  by gravity.  Ce-
ment kilns operate countercurrently; i.e.,
solid materials travel in one direction and
hot gases plus dust emissions travel in the
opposite direction.   A slurry of  30% to 40%
water (typically 35% to 39%) and  finely
crushed rock is fed into  the kiln at the
upper end; at the opposite end of the kiln
is a powerful oil  fire.   At San Juan Cement
kiln #2,  No. 6 fuel  oil  is burned at about
1.51  x  10~3 m3/s (24 gpm), a heat input of
approximately 62 x 10" watts  (212 million
Btu/hr).  As the raw material passes slowly
through the kiln (1 to 4 hours), it dries;
then, at a temperature of 550°C (1,020°F),
calcination starts (C02 is extracted from
the calcium carbonate in the  feed); finally,
it approaches the hot burning zone of the
kiln.   In the burning zone, 1,500°C(2,700°F)
temperatures calcine and fuse the raw ma-
terials creating a complex calcium silicate
aluminoferrite mineral  substance called
"clinker," which is discharged from the
lower end of the kiln and cooled by large
fans in the clinker cooler [1].   The clink-
er production rate at this plant ranged
from 28 to 33 metric tons/hr  (31 tO 36 ton/
hr).  The addition of about 6% gypsum to
milled  clinker completes the  process in the
production of portland cement.

     Exhaust gases from kiln  #2  pass
through a baghouse-type dust  collector
where entrained particulate matter is re-
moved.   The cleaned exhaust gases are then
released to the atmosphere through a single
stack.   The baghouse employed at kiln #2
consists of 1,536 fabric filters, each 9.3
m in length.   The efficiency of the bag-
house in removing particulate matter from
the gas stream was reported to be 99.8 per-
cent.   It was not an objective of this  pro-
gram to determine the baghouse collection
efficiency.

Environmental  Design

     The sampling and analytical program
was designed to identify all  major pollu-
tants from the burning of the hazardous
wastes  available for this program and to
quantify their respective emission rates,
investigate the chlorine material  balance
of the cement process,  determine burning
rate limits as related to product accep-
tability and refractory lining integrity,
and determine the destruction and removal
efficiencies (DREs) of the principal organ-
ic hazardous constituents (POHCs)  in the
waste fuels.   The POHCs chosen for this
program were the three chlorinated com-
pounds   known to be present in the waste
fuel mixture:

(1) Methylene chloride (dichloromethane),
    CH2C12; higher heating value (HHV)  =
    3058 Btu/lb
                                            211

-------
                                                                    STACK GASES
                                                                (PARTICULATES + VAPOR>
            FUEL OIL
    HAZARDOUS WASTE
PRIMARY AIR (AMBIENT)
                   SECONDARY
                   AIR (HEATED)
CEMENT
CUNKER
PRODUCT
BAGHDUSE
DUST
                Figure 1.   Schematic  diagram of cement kiln
                             burning hazardous waste.
   Inputs

   •  fuel oil
   •  hazardous waste
   •  slurry  feed
 (2)  Chloroform (trichloromethane),  CHC13;
 HHV    1349 Btu/lb

 (3)  Carbon tetrachloride (tetrachlorometh-
 and),  CC14; HHV = 432 Btu/lb

      Emission measurements  included partic-
 ulate  matter, carbon monoxide,  carbon  diox-
 ide  (C02), sulfur dioxide,  nitrogen oxides,
 hydrogen chloride, total gaseous  hydrocar-
 bons  (THC), total chlorinated hyrdocarbons,
 methylene  chloride, chloroform, carbon  tet-
 rachloride, trace metals in particulate
 matter, and organics with special attention
 given  to dioxins  and furans in the bag-
 house  flyash.  Chlorine content of  the
 baghouse flyash and cement  clinker  also
 was  monitored.  The waste fuels and fuel
 oil  used to fire the cement kiln  were  anal-
 yzed  for principal organics, trace  metals,
 ash,  chlorine, nitrogen, and sulfur content

      Table 1 summarizes the test  matrix of
 the  demonstration program wherein the
            Outputs

            •  clinker  product
            •  baghouse dust
            •  stack emissions, including
              - particulates
              - HCl vapor
              - unburned chlorinated
                hydrocarbons

             waste feed rate to the  kiln  and  the chlor-
             ine content of the waste were  varied.
             Table 2 summarizes the  overall test program
             and shows each collection method and anal-
             ytical method.  A Quality Assurance Project
             Plan was prepared and reviewed prior to the
             program.  A full description of  the quality
             assurance/quality control (QA/QC)  results
             involving replicates, splits,  blanks,
             spikes, and reference standards  is provi-
             ded in the report final.

             RESULTS AND DISCUSSION

                  A detailed summary of the waste fuel
             composition of each of the six waste batch
             shipments  is given in Table 3.  A seventh
             waste batch was burned and tested;  it  was
             a mixture of Batches 4 and 6.

                   Five  comprehensive baseline tests (no
             waste fuel burned) were completed in order
             to determine the difference in emissions
             when waste fuel was burned and when  it was
                                          212

-------
      TABLE 1.   TEST MATRIX OF WASTE FEED RATE AND  CHLORINE  CONTENT
Approximate waste
   feed  rate to
kiln,  m3/s  (gpm)     6.5
                                   10.1
                                         Chlorine  in waste, wt %
18.7
21.4
22.9   32.0    35.1
8
9
1
1
1
2
2
3
3
3
.39
.46
.10
.26
.89
.21
.71
.15
.47
.79
x
x
X
X
X
X
X
X
X
X
10-
10"
10"
10"
10"
10"
10"
10"
10"
10"
5
5
4
4
4
4
4
4
4
4
(1
(1
(1
(2
(3
(3
(4
(5
(5
(6
.33
.50
.75
.00
.00
.50
.30
.00
.50
.00
)
) x x
) X X
) X X
) x
) X X X
) x
) X X
) x
\ »
    x =  Conditions tested.
not burned.   Four of the baseline testing
days involved EPA Method 5 testing, and the
fifth test was  a SASS run.  The SASS was
used to identify PICs and quantify dioxins
and dibenzofurans.

     When  waste fuel was burned, ten com-
prehensive tests on seven different waste
fuel batches  were completed.  These in-
cluded SASS  runs on waste batch numbers 3,
4, and 6;  and EPA Method 5 runs on seven
waste fuel burn tests.  POHC testing and CO
monitoring were conducted on fourteen addi-
tional  tests.

     Measurements of several conventional
pollutants were made repeatedly during the
program to determine the difference in
emissions  between baseline operation (no
hazardous  waste fed to the kiln) and waste
fuel burns.   Table 4 presents the compari-
sons for particulate, NOX, S02, total hy-
drocarbon, and  HC1 emissions using the t-
test to determine statistically significant
difference.
        In  each case, pollutant concentrations
   were  used  rather than emission rates be-
   cause the  volumetric flowrate varied as
   much  as  30%, and hence added another vari-
   able  to  the statistical analysis.

        High  carbon monoxide emissions are  an
   indicator  of incomplete combustion in the
   cement kiln.  When the kiln combustion was
   stable,  CO emission levies stayed below  10
   ppm.   However, any process fluctuations  or
   change in  kiln variables caused a momentary
   excursion  in CO emissions to levels greater
   than  1,000 ppm (0.1%), even during baseline
   testing.

        Figure 2 illustrates the approximately
   one-hour transition period and the change
   in  CO emissions resulting from the intro-
   duction  and burning of waste fuel  in a
   kiln.   From 0800 hours to 0905 hours, the
   kiln  was not burning waste fuel, and CO
   emission levels hovered at 0 to 40 ppm,
   indicative of stable kiln operation.  As
   the waste  fuel was introduced, the CO
                                          213

-------
TABLE  2.   SAN JUAN  CEMENT  COMPANY  WASTE  FUEL  DEMONSTRATION
               BURN  SAMPLING  AND  ANALYTICAL PROGRAM
        Parameter measured
                                           Sampling
                                            nethod
                             Analytical method
  Stack Samples
  Participate natter
    Metals on particulate
    Organics on particulate
  Opacity
  Sulfur dioxide
  Nitrogen oxides
  Carbon monoxide
  C02  and 02
  Hydrogen chloride
  Total gaseous hydrocarbons
  Total chlorinated hydrocarbons
  Three chlorinated species (POHCs)
  Organic compound speciation

  Ambient Air

  Particulate natter

  Process Water

  Organics (3 species)
  No.  6  Fuel Oil

  Btu  content
  Chlorinated methanes
  Sulfur content
  Trace  metals
  Principal organics

  Waste  fuel

  Btu  content
  Moisture content
  Total  chlorine
  Total  nitrogen
  Total  sulfur
  Trace  metals
  Principal organics
  PCBs and pesticides
  Ash  content

  Solid  Waste (kiln dust)

  Principal organics
    Furans and dioxins
  Chlorine content
  E.P  toxicity
    Furans and dioxins
    Trace Betels
EPA Method 5
EPA Method 5
EPA Method 5
EPA Method 9
EPA Method 6
EPA Method 7
EPA Method 10
Integrated bag cample
Impinger  train
Direct to analyzer
Integrated bag samples
Integrated bag samples
SASS train
EPA Method  5
I CAP
Extractions and GC/MS
  for principal organics,
  dioxins,  and furans
EPA Method  9  (on site)
EPA Method  6
EPA Method  7
HEIR continuous analyzer
EPA Method  3  (on site)
Specific ion  electrode
Continuous  FID
GC/EC (on site)
GC/EC (on site)
GC/MS
High volume  gas sampler    EPA-Appendix 6 FR 121:0105
Integrated sample
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
Grab
EPA priority pollutant
  nethodology, GC/MS
ASTM D240-64
GC/EC
ASTH D-3177
ICAP
GC/MS
ASTH D240-64
GC/MS
ASTM D8081
Kjeldahl
ASTM D129
ICAP
GC/MS
GC/MS and GC/EC
ASTM D482-IP4
Extraction and GC/MS

ASTM D-806

Extraction and GC/MS
ICAP
                                             214

-------
         TABLE 3 .    SUMMWARY OF  WASTE  FUEL  ANALYSES  FOR CEMENTO
                       SAN  JUAN DEMONSTRATION  BURN  (volume basis)
Compound
Water
Methanol
Ethanol
Acetone
2-Propanol
Methylene chloride (POHC)
Hexane isomers
3-Hethylpentane
Hexane
Chlorofom (POHC)
Ethyl acetate
Methyl acetate
Carbon tetrachloride (POHC)
Benzene
Hexamethyl disiloxane
Toluene
Acrylonitrile
Methyl ethyl ketone
Cj-benzene isoner
Cg-benzene isom«r
Sec-butyl ethylbenzene
Xylene isoners
Dijnethylphenol isomer
1 , 1 ' - ( 1 , 2-ethanediol)bis-
4-«ethoxyb«nrene
Unknowns
PCBs, ppm
Pesticides', ppi
Properties
Btu content, Btu/lb
Specific gravity
Chlorine content, wt %
Ash content, wt %
Batch 1,
vol %
<1 0
10.4
0.8
14.2
4.7
24.4
3.9
5.4
19.8
1.0
4.0
ND
0.8
0.4
0.1
0.2
ND
NA
NA
NA
NA
NA
NA

NA
8.9
<50
<100

11,188
NA
32.0
0.30
Batch 2, Batch 3,
vol % vol \
4.1
7.1
3.2
12.2
5 2
16 9
3.2
4.6
17.3
0.8
14.0
NAb
0 6
0 4
ND
0.1
1.0
NA
NA
NA
NA
NA
NA

NA
9.3
<100
<100

11,198 11
NA 0.
22.9
0.20
4.
13
8.
11.
5.
12.
1.
2.
7.
3.
9.
0.
1.
0.
3
9
6
2
3
0
8
7
2
4
0
4
4
2
ND
0.
1
0
1.
.02
.1
.08
.33
ND
1
0
0

0
14
<100
<100

,022
9948
21.4
0.38
.23
.24
.04

.23
.3







Batch 4,
vol %
8.9
6.2
4.7
10.5
4.5
12.1
1.5
3 2
8 5
5.4
6 6
<1 0
10.2
0.3
NA
<0.5
<0.7
1.1
0.5
NA
1.7
NA
NA

NA
11.9
<100
<100

10,099
0.9885
35.1
0.23
Batch 5,
vol %
23.0
10.9
16.8
4.6
3.1
1.4
ND
ND
5.9
4.0
3.5
ND
7.8
0.1
NA
ND
1.0
NA
0.9
ND
NA
NA
ND

NA
17.9
<100
<100

4,546
1.0092
18.7
0.31
Batch 6, Batch 4/6,
vol % vol \
2.0
ND3
5.6
2.2
1.6
5.0
1.5
3.6
15.7
0.1
22.7
ND
0.01
0.05
NA
ND
ND
NA
2.2
23.8
NA
ND
ND

NA
13.9
<100
<100

13,098 NA
0.9163 0.
6.5 10.1
0.046 NA
NA
NA
NA
NA
NA
5.1
NA
NA
NA
0.9
NA
NA
1.5
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA

NA
NA
<100
<100


9410


Density,
q/mL
1.
0
0
o.
0
1
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0

0








.000
.7914
.7983
.7899
.7855
3266
.6532
.6645
.6603
.4832
.9003
.9330
.5940
.8787
.8923
.8669
.8060
.8054
.90
•9°e
.90
•87e
.90*

.90*








TLV.C
«q/m3

260
1,900
2,400
500
360d


1,800,
50f
1,400
610,
35f
35d
a
375
5f
590d

"(J

435d


-








*ND = not detected, generally <0.1% by volume.  Components were quantified in volume % because external standards
 were prepared on a volume basis.
 NA = not analyzed.
 Threshold liait value for workplace air.
 NO TLV assigned to this compound  or isomer.
 Estimated values.
 Suspected or known carcinogen.
gAs per priority pollutant list.

-------
          TABLE 4.   COMPARISON OF POLLUTANT  LEVELS BETWEEN
                     NORMAL  OPERATION AND WASTE FUEL  FIRING  IN
                     CEMENT  KILN NO.  2,  SAN JUAN CEMENT COMPANY

Pollutant
Particulate matter
NO
X
S02
Total hydrocarbons
HC1
Mean
Baseline
93 ± 65 mg/m33
(n=4)
136 ± 83 ppm
(n=4)
279 ± 243 ppm
(n=4)
8.3 ± 2.1 ppm
(n=9)
0.82 mg/m3
(n=2)
loading
Waste firing
99 ± 65 mg/m3
(n=7)
68 ± 23 ppm
(n=9)
450 ± 245 ppm
(n=6)
12.7 ± 2.1 ppm
(n=7)
3.3 ± 1.7 mg/m3
(n=9)
Statistical significance
at 95% degree of certainty
No significant difference
Significant difference
Significant difference
Significant difference
Significant difference

      95% confidence level.
           8  «n

               0800
                    0815
                         0830
                              0845
                                  0900    W15
                                   TIME Of DAY
                                             0930
                                                  OM5
                                                       1000
                                                            1015   1030 1045
      Figure 2.   Illustration  of waste  fuel burn transition period
                   effect on carbon monoxide emissions, Run W6-SASS-1,
                   28  January  1982, San  Juan Cement Company.
levels  rose rapidly beyond 1,000 ppm for
approximately 20  min and then  returned to
levels  below 100  ppm as the kiln operation
was  stabilized.
POHC  Destruction and Removal  Efficiency

     The complex combustion chemistry for
organic materials becomes perplexing when a
                                        216

-------
broad range of organic compounds present in
a liquid waste are burned.   On a weight ba-
sis, most of the organic carbon in the
waste is oxidized to C02 in the combustion
process, but trace amounts  of organic chem-
icals survive the oxidation process and are
only partially reacted.   In addition, small
amounts of other organic chemicals must be
produced if proper conditions exist.   Oxi-
dative chemical  principles  would suggest
that all organic compounds  regardless of
their origin would be converted to C02 and
other primary products if a sufficiently
high temperature is achieved in the pres-
ence of adequate oxygen  for a time long
enough for complete oxidation to occur.  If
the conditions of temperature, time,  and
oxygen are not met, the  combustion of the
organic material in the  fuel, plus any new
organic components formed in the combustion
process, would be expected  to be incomplete
and some possibly detrimental compounds of
concern may then appear  in  the effluents
[2].  This demonstration program investi-
gated the amount of destruction of the or-
ganic compounds  in the waste (ORE of  the
POHCs) and the types of  organic compounds
(PICs) formed.

    Currently, cement kilns which burn
hazardous wastes are not regulated under
RCRA.  However,  in this  study kiln perfor-
mance in destroying waste was compared to
that required by regulation for hazardous
waste incinerators.

    The destruction and  removal  efficiency
for an incineration/air  pollution control
system is defined by the following equation:
   ORE =   in,,  out (100%)
            Win
 where  DRE
             destruction and removal
       ,,     efficiency, %
        in   mass feed rate of the princi-
             pal  organic hazardous consti-
      ,,      tuents (s) to the incinerator
       out   mass emission rate of the
             principal organic hazardous
             constituent(s) to the atmos-
             phere (as measured in the
             stack prior to discharge).

    Within the family of chlorinated  hydro-
carbons,  the monocarbon compounds are
believed to be among the most difficult to
destroy thermally.   The carbon-chlorine
bond is especially tenacious and in general,
the more C-C1  bonds present, the more dif-
ficult it will be to destroy the compound.
Accordingly, a monocarbon chlorine-contain-
ing molecule will tend to be harder to
destroy thermally than a two-carbon chlor-
ine-containing molecule.  The Gibbs free
energy of formation values at 1600°C (2421°
F) illustrate this.  For the following
compounds, the amount of energy required to
form a molecule (conversely, the "difficul-
ty" required to destroy it) will be great-
est for carbon tetrachloride and least for
trichloroethane (C2H3C13).

          CCli>CHCl3>CH2Cl2>C2H3Cl3

    Concentrations of the POHCs were meas-
ured during baseline testing (days when no
waste fuel was burned) in order to give
background or normal concentrations of
these compounds in the exhaust gas.  The
average background level was then subtract-
ed from the results obtained during a waste
fuel burn to arrive at the contribution at-
tributable to the waste burn.

      The average ORE for the POHCs for
each test run is presented in Table 5.
Methylene chloride was destroyed to at
least 99.0% efficiency, with the only ex-
ceptions being the two tests with waste
batch #5 which contained only 1.4% methy-
lene chloride.  In general, the lower the
mass feed rate of a POHC, the lower was the
DRE.
      Chloroform and carbon tetrachloride
were more difficult to destroy than methy-
lene chloride.  Also, in most waste bat-
ches, methylene chloride was the most
bountiful POHC in the waste.  Waste batches
#4 and #5 had the largest amounts of chlor-
oform and carbon tetrachloride, and the
best DRE results for carbon tetrachloride
were observed for the test runs on these
two batches.  Figure 3 illustrates the fre-
quency of obtaining a certain "number of
nines" of DRE for each of the POHCs in the
demonstration program.  In only a few in-
stances was the desired 99.99% or even
99.9% DRE obtained for the chlorinated
monocarbon compounds chosen as POHCs in
this program.
                                           217

-------
            TABLE  5.  DESTRUCTION  AND REMOVAL  EFFICIENCIES OF  POHCS
                       FOR DEMONSTRATION  BURN TESTS  AT SAN JUAN
                       CEMENT  COMPANY KILN #2

Run
number
Wl-2*
W2-la
W3-1
W3-2
W3-3
W4-1
W4-2
W4-3
W4-4
W4-5
W5-1C
W5-2
W&-1
W4/6-lC
W4/6-2C
W4/6-3
W4/6-4
W4/6-5C
Methylene
chloride
>99.997
99.995
>99.991
99.960
99.659
98.237
99.418
99.461
99.984
99.335
93.292
96.663
99.223
99.760
99.668
99.564
99.133
99.474
Chloroform
>99.842
>99.859
99.887
99.932
>99.960
98.592
99.470
99.283
98.975
99.950
98.388
96,.099
D
95.617
92.171
98.703
>99.737
97.515
Carbon
tetrachloride
99.309
>99.996
91.043
96.864
98.977
97.732
98.122
99.142
99.684
99.069
99.553
99ft460
_
94.129
99.325
94.512
92.253
95.873

                   Waste feed rate  was estimated.
                  DNot  present in waste  fuel

                  'Stack gas  volumetric  flow  rate  of 1619
                   dscmpm used.
    In one instance, run number W6-1  for
CCU,  the Wout  exceeded the W-    The  mass
feed rate to the  kiln was less than 0.11  kg
/hr.  The higher  mass emission rate observ-
ed suggests  that  CC1A is formed as a  prod-
uct of incomplete combustion from the
combustion of methylene chloride and  chlo-
roform.

    There are two probable reasons to
explain  the  low ORE results obtained  in  the
demonstration program:   (1) lack of atomi-
zation of the waste fuel, and (2) difficul-
ty of  incinerability of highly chlorinated
monocarbon compounds.  The waste fuel  in-
jection  had  to  match the fuel oil injection
pattern  in order  to prevent flame impinge-
ment on  the  inner wall  of the kiln and
preignition,  or backpuffing, of the fuel
oil  stream.   Therefore,  methods to increase
the  waste fuel  atomization were not
attempted.

    Other compounds were detected by gas
chromatography/electron capture (GC/EC)
during the  analyses for POHCs, which
eluted at retention times of 0.51 min
(CH2C12), 0.80 min  (CHC13), and 1.01 min
(CC1,,).   The most commonly seen compound
had a retention time of 0.57 min to 0.61
min, which  a post-test laboratory experi-
ment with duplicate GC conditions tenta-
tively identified as trichlorotrifluoro-
ethane.   This compound most probably was
Frror, 113 from the  laboratory's air condi-
tioner.   Another compound which was seen
in several  instances eluted at about 1.1
min and  was tentatively identified as tri-
chloroethylene, a likely PIC from chloro-
methane  combustion.  1,1, 1-trichloro-
ethane (1.54 min R. T. ), tetrachloroethy-
lene (1.68  min R. T-), acetone (2.11 min
                                           218

-------
           40
        UJ
        §30
        fe 20
        LLJ
        £ 10
                          a.  METHYUENE CHLORIDE
           40 r-
           30 -
           40 r
                             2345
                          NUMBER OF NINES  ORE

                          b.  CHLOROFORM
>
fe 20
o:
UJ
| 10
=>
Z
n
-
-
_

—
_
























J ... ,.4.
                      12345
                          NUMBER OF NINES ORE

                          c. CARBON TETRACHLORIDE
=\ ™
>
fe 20
LU
1 10
13
z n
_
-
-
-














1 1
                012345
                          NUMBER OF NINES ORE

Figure 3.  Frequency  distribution of DRE results for each POHC
                                219

-------
R.  T.), acetonitrite (2.15 min  R.  T.),  and
acrylonitrile (2.16 min  R. T.)  were never
detected by the in-field GC/EC  analyses.

    The SASS samples collected  for chlori-
nated dioxins and chlorinated dibenzofurans
also were analyzed for products of in com-
plete combustion.  Four  samples (one base-
line and three waste burning) were analyzed
for PICs.  The baseline  sample  was analzyed
because the cement plant burns  fuel  oil  as
its primary energy source.  During the
program, fuel oil accounted for 87% to 100%
of the kiln's energy requirement.   Products
of incomplete fuel oil combustion  need to
be distinguishable from  products of incom-
plete hazardous waste combustion -- hence
the baseline sample.  Three SASS samples
were collected during hazardous waste burn-
ing representing different feed composi-
tions, feed rates, and operating conditions.

    The  PICs which were  not detected during
the baseline test and can be considered at-
tributable to hazardous  waste burning are:

   • Trichloroethylene (100 to 100,000 rug/
    hr)
   • Phenol (2.4 to 11.0  mg/hr)
   • C2-naphthalene isomers (10 to 50 mg/hr)
   • C3-naphthalene isomers (14 to 46 mg/hr)

Dioxin and Dibenzofuran  Results

    One of the objectives of the program
was to determine if polychlorinated diben-
zodioxins  (PCDD) and polychlorinated di-
benzofurans  (PCDF) were emitted as products
of incomplete combustion while hazardous
waste was  being fired to the kiln.  Chlor-
inated dioxins and dibenzofurans are
believed to  be among the most toxic sub-
stances  to humans.  EPA officials required
that extensive sampling and analysis be
conducted  for these compounds during the
demonstration program.  During the course
of the program, 28 different samples were
collected  for analyses as shown below.

   • 4 SASS train samples  (particulates and
    vapors in stack gas)
   • 5  EPA  Method 5 samples (particulates
    in stack gas)
   • 11 baghouse dust samples (plant solid
    waste)
   • 8 RCRA extracts of baghouse dust sam-
    ples  (plant solid waste)
     Eight  samples were  taken  during  base-
 line  (nonburning) conditions,  and  20
 samples were taken during  hazardous  waste
 burning operations.  The SASS  train  sam-
 ples  resulted  in three  sections  for  analy-
 sis:   1) methylene chloride rinses of  the
 sampling probe, teflon  line,  filter  holder,
 and organic module, 2)  combined  filter and
 XAD-2  adsorbent resin,  and 3)  the  conden-
 sate water removed from the organic  module
 during sampling.

    The baseline (no waste fired)  SASS run
 (BS-SASS)  showed some positive,  detectable
 values of  hexachloro- and  heptachlorodi-
 benzofuran in  the adsorbent resin  extract,
 although none  could be  detected  in the
 other  portions of the train.  This caused
 the reported values to  be  less than  the
 average detection limit for the  entire
 train, which was 3.4 ng/m3.   It  is note-
 worthy that no PCDDs or PCDFs were detec-
 ted in any waste burning SASS samples  at
 a detection limit ranging  from 1.6 ng/m3
 for tetrachloro-isomers to 4.9 ng/m3 for
 octachloro-isomers.

    In the analyses of  the particulate
 catch  from EPA Method 5 runs, no detecta-
 ble quantities of PCDDs were found in  any
 of the particulate samples.   In  only one
 sample, run W3-3, 11.0  ng/m3 of  penta-
 chloro-PCDF, 26 ng/m3 of'hexachloro-PCDF,
 and 8  ng/m3 of heptachloro-PCDF  isomer
 were found.  These detectable emissions
 occurred when  the kiln was fed 2.75  x  ~\Q~U
 m3/s (4.35 gpm) of waste which contained
 21.4% chlorine.  This corresponds to a
 chlorine input of 3.5%  by weight of  total
 fuel input (fuel  oil  plus  hazardous  waste)
 which results  in the production  of off-
 spec cement clinker and a  potentially
 kiln-damaging  condition.   Excessive
 chlorine in the clinker will lengthen
 cement set time and reduce strength.  The
 Chlorine Material  Balance  section des-
 cribes how this is an operating  condition
 which  is intolerable for the cement  plant.
Thus, the generation of detectable quan-
 tities of PCDFs occurred only when oper-
 ating an "upset" or "out-of-control"
 kiln.  Under other conditions the cement
 process did not emit PCDFs, and  it did not
 emit PCDDs under any waste burning condi-
 tions.  This lack of detectable  quantities
 of PCDD and PCDF is an expected  result
 because the waste fuel  contains  no poly-
                                           220

-------
chlorophenolate percursors, the combustion
temperatures were typically well above
1000°C, and a supplemental fuel was used to
burn the wastes [2].

Chlorine Material  Balance

    The combustion of chlorinated hazard-
ous wastes as auxiliary fuel in cement
kilns results in the generation of hydrogen
chloride (HC1) in the kiln.  Hydrogen
chloride can be absorbed by the clinker
product, cement kiln lining, or baghouse
dust; it also may be emitted as part of
the particulate.  Unabsorbed HC1 will be
emitted from the stack as will  the un-
burned chlorinated hydrocarbons.  HC1 is
formed rather than chlorine gas (C12) be-
cause the conversion of chlorine gas to
hydrogen chloride is favored by high
temperatures and high water content in the
combustion gas.   Typically, the burning
zone temperature in a cement kiln is 1425°C
(2600°F) with a moisture content of about
10% to 15%.  In wet process cement kilns,
additional  moisture is added to the combus-
tion gases in the slurry drying zone of
the kiln.  Equilibrium calculations show
that 99.87% of the chlorine in  the com-
bustion gases will  be present as HC1; only
0.013% of the chlorine will be  present as
chlorine gas.  The HC1 is rapidly absorbed
by alkaline calcium, sodium, and potassium
compounds present in the clinker and dust
to form sodium,  potassium, and  calcium
chlorides.

    In addition to calcium, while it is the
major element of the cement-making process,
sodium and potassium are important impuri-
ties in the raw feed material.   They are
undesirable in the clinker and  are normally
removed by the addition of chloride to the
kiln -- either in  the form of calcium
chloride or as HC1.   The resultant sodium
chloride (MaCl)  and potassium chloride (KC1)
volatilize at the  temperature of the kiln,
condense as a particulate in the cooler end
of the kiln, and are removed by the air
pollution control  device.   The  amount of
chloride that a  given kiln can  handle is
limited by the sodium and potassium content
of its feed.   If excessive chlorides are
added to the system, these salts can form a
clinker ring in  the kiln and possibly dam-
age the refractory lining [3].
    The distribution, or fate, of the chlor-
ine (as chloride) was studied in seven"tests
by measuring the chlorine-chloride content
of the input and output streams of the kiln.
The material balance closures for chlorine
ranged from 83% to 104% with an average
closure of 92%.

    In this study, it was found that about
82% of the chlorine fed to the cement kiln
will appear in the clinker product.  The
amount of chlorine in the clinker is pre-
sumed to be dependent, primarily, upon feed
alkalinity, and will vary from one cement
plant to another.  Other cement plants with
different slurry feed composition and alk-
alinity will realize different results,
e.g., up to 92% of the chlorine was retained
in the clinker at St. Lawrence Cement in
Missisauga, Ontario, during those tests.

    Normally, the potassium contained in
the slurry feed leaves the kiln as potas-
sium sulfate (K2SO/.) and potassium oxide
and is collected in the baghouse.  Presum-
ably, HC1 reacts with K2SO<, to form KC1
salt which is trapped in the clinker.  The
kiln operates above the dew point for H2SOi,
H60°C), and H2SO<, not absorbed by dust
will leave as a vapor increasing SOX emis-
sions.  Earlier, it was shown that there
is a "statistically significant change" in
SOX emissions when burning hazardous waste.
The average SOX emission rates may indeed
be higher due to the chemistry of the in-
organic elements in the kiln; i.e., the
chloride acid gas is trapped (scrubbed)
preferentially over the sulfate acid gas.

    The amount of chlorine appearing in the
stack emissions as unburned chlorinated
hydrocarbons ranged from 0.076% to 4.3%
which represented appropriate overall de-
struction and removal efficiencies of
99.924% to 95.7% on a total chlorinated
hydrocarbons feed basis.  The tests showed
lower DREs when burning higher levels of
carbon tetrachloride.  The total absorption
of HC1 by the cement kiln process averaged
99.7% in seven tests.

    The amount of chlorine appearing in
the baghouse dust varies from 5% to 26% of
the total chlorine feed.  The amount of
chlorine appearing  in the baghouse dust  is
influenced  by the concentration of chlor-
ine in the  total fuel  (fuel oil plus
hazardous waste).
                                           221

-------
                   g
                   UJ
                   i
                   Of
                   o
                       5r
                       4 -
                   2  3
                       2 -
                       1 -
LEGEND
._ CODE
A
B
C
D
E
F
G
H
I
J
A
C
i
RUN NO.
BW-SASS-1
BW-1
BW-3
BW-2
BW
W4-1
W3-SASS-1
W4-2
W3-2
W2-1

V1 — —
1



J

/
/
/
/
1
/
/
"
1 1 1
                                  1234
                                  PERCENT CHLORINE IN TOTAL FUEL TO KILN
         Figure  4.  Effect of chlorine concentration in total  fuel
                       (fuel  oil +  hazardous  waste) on  chlorine con-
                      tent of baghouse  dust  at San Juan Cement Company.
    Figure  4  shows the dramatic increase
in chlorine concentration in baghouse dust
when chlorine comprises 3% by weight or
more of the total fuel.  The plant always
went into an  upset condition about 90 min-
utes after  the start of a trial burn which
had 3% or more of chlorine in the total
fuel.   Some key observations based on the
experience  of saturating this kiln with
chlorine are  listed below.

    The cement kiln will remain in a sta-
    ble and controlled operating condition
    when the  chlorine content of the total
    fuel (fuel oil plus hazardous waste) is
    3% by weight or less.  When the chlor-
    ine content exceeds 3%, the kiln be-
    comes operationally unstable about 90
    minutes after the start of burning
    more than 3% chlorine in the total
    fuel.   The instability takes several
    forms.  The burning zone temperature
drops from 1430°C (2600°F)  to  815°C
(1500°F) in less than  two minutes
some 80 to 90 minutes  into  the burn.
The kiln starts a rapid  loss of re-
fractory lining (coating) to cement
clinker product.  A clinker ring
forms in the burning zone.  Hydro-
carbon emissions increase  (about four-
fold) in the stack.

Recovery from this upset condition is
achieved by shutting off the hazardous
waste while maintaining  a constant
heat balance.  Remarkably,  the kiln
recovers in 12 to 15 minutes if the
waste is shut off as soon as the kiln
goes into an upset condition.   Contin-
ued operation during an  upset  would
destroy the kiln.  It is presumed  that
the excess chlorine is absorbed in the
lining and reduces its slagging temper-
ature to below 815°C (1500°F).  When
                                          222

-------
   the  lining  starts  to  slag,  all  heat  is
   used  to melt  the  lining,  not  to make
   cement.   It is  presumed that  the  kiln
   recovers  in about  15  minutes  because it
   takes  the burning  zone about  15 minutes
   to purge  itself of slagged  lining.   Re-
   covery is as  dramatic as  is the upset.

  •  Even  though the cement kiln remains
   stable up to  about 3% chlorine  in  the
   total  fuel, the product quality be-
   comes  unacceptable when the chlorine
   content exceeds 1% of the total fuel
   at the San  Juan Cement Company.

  '  Clinker rings can  be  formed by  either
   temperature excursions in the burning
   zone when not burning hazardous waste
   or by excessive chlorine  additions to
   the  kiln.   Operators  at San Juan
    Cement Company were successful  at
    breaking  clinker  rings simultaneous
   with burning  chlorinated  hazardous
   waste.   It  was  not necessary  to cease
    hazardous waste burning  in  order  to
    break a clinker ring.

    Incineration  of hazardous wastes  in
cement  kilns  also demonstrated  a  time  delay
in the release  of chlorine from the kiln.
The onset of  unstable kiln operations  some
90 minutes after  excess quantities  of
chlorine are  added to the kiln  was  just
described.   Figure 5  shows the  dynamic be-
havior of chlorine releases  from  the  cement
kiln  compared to  a constant,  short  dura-
tion, and a moderate,  but still excessive,
chlorine feed rate to the cement  kiln.  It
can be seen that  it actually  takes  about
three hours for the kiln  to  completely
purge itself  of the added chlorine.

CONCLUSIONS

    Some of the results observed  in this
demonstration program were contradictory
to results from other cement  kiln inciner-
ation tests;  e.g.,  lower DREs,  no change
in particulate  emissions, and significant
changes  in S02  and MOX emissions.  The
conclusions presented below  apply only to
this  particular kiln  and the  results  from
this  demonstration program.

1. The  inability of  this kiln  to consis-
    tently achieve 99.99% ORE (a  value
6.
7.
 which hazardous waste incinerators
 must demonstrate) of the POHCs is
 attributed to unatomized waste in-
 troduction to the kiln flame and the
 difficult incinerability of the POHCs
 These compounds (CH2C12, CHC13, and
 CC1J are occasionally employed as
 fire retardants because of their
 ability to move hydrogen atoms from
 the free-radical  branching combustion
 reactions to form HC1.   Combustion of
 chlorinated species  containing less
 chlorine may have resulted in  higher
 DREs.

 Chlorinated dioxins  and  chlorinated
 dibenzofurans are not  produced at
 detectable levels (1.6  ng/m3)  when a
 cement  kiln firing chlorinated wastes
 is  operating normally.

 A cement kiln will absorb over 99%
 (about  99.7%)  of  the  HC1  formed dur-
 ing  the  combustion of chlorinated
 hazardous  wastes.  This  absorption
 is  partitioned  between the  clinker
 and  baghouse dust.

 At  San Juan  Cement Company, approxi-
 mately 82%  of  the chlorine  fed  to the
 cement kiln  appears in the  clinker.
 This  limits  the chlorine  content  of
 the  total  fuel  to  less than 1%.
 This  may  vary  at  different  cement
 plants because  quarry alkalinity
 (ability  to  absorb chlorine) varies
 at each cement  plant.

 Achievable fuel savings are a function
 of the chlorine content of  the  waste
 and each  plant's  ability  to absorb
 chlorine.  At San  Juan Cement Company,
 a hazardous waste  containing less  than
 5% will  result  in  at least a 20%  sav-
 ings  in fuel costs.  Higher fuel  sav-
 ings may be  possible for  higher
 chlorine contents  at other plants.

 Production of salable cement product
 is possible when burning chlorinated
 hazardous wastes provided the plant's
 chlorine absorbability limit is not
 exceeded.

Atomization of the waste fuel  would
 be desirable, if a flame configura-
                                           223

-------
ce
LU
•^
»— •
VJ1
IN
I— •
INJ
	I
o
§
z
I—I
o:
I
o

500r

400 -
             300
             200
             100
               0
               800
                       CHLORINE FEED
                        TO KILN
                                           CHLORINE FEED
                                           HALTED-
                                                          CHLORINE IN CLINKER
                                                          EXITING KILN
                   1000      1200     1400      1600     1800     2000     2200     2400
                                        TIME OF  DAY
10.
II.
12.
 tion can be obtained which does not
 alter the primary fuel flame config-
 uration.

 High feed line pressure  [1,380 to
 2,070 kPa (200 to 300 psigj] is not
 required for waste injection to the
 kiln.  This pressure requirement may
 change depending on the  type of atom-
 izing nozzle used.

 There is no significant  change in par-
 ticulate emission due to burning
 chlorinated hazardous wastes.  This
 result was observed on a cement kiln
 equipped with a fabric filter air
 pollution control system.  A cement
 kiln with an electrostatic precipo-
 tator may not achieve similar results
 due to a change in dust  resistivity.

 Emissions of sulfur dioxide, total
 hydrocarbons, and hydrogen chloride
 increased significantly  when waste
 was burned.  A cement kiln with a
 higher alkalinity feed than that at
 the San Juan Cement Company may not
 have an increase in S02  emissions.

 Emissions of nitrogen oxides decreased
 significantly when waste was burned.

 There is no change in particulate am-
 bient air quality due to hazardous
 waste combustion in cement kilns.
                                                 13.   The  solid  waste (baghouse dust)
                                                      generated  by hazardous waste
                                                      burning  and  its RCRA extract
                                                      (leachate) are suitable for land-
                                                      fil1 ing.
                                    REFERENCES

                                    1.   Lauber,  J.  D.,  "Burning Chemical
                                        Wastes  as  Fuels in Cement Kilns."
                                        Journal  of the  Air Pollution Control
                                        Association,  32(7):  771-777, July
                                        1982.

                                    2.   Junk, C. A.,  and J. J.  Richard.
                                        "Dioxins not  Detected in Effluents
                                        from  Coal/Refuse Combustion."  Chemo-
                                        sphere,  10(11/12):  1237-1241, 1981.

                                    3.   Weitzman,  L.,  "Cement Kilns as Hazard-
                                        ous Waste  Incinerators."  Environmental
                                        Progress,  2(1):  10-14, February 1981.
                                             224

-------
           AUTOMATED METHODOLOGY FOR ASSESSING  INHALATION EXPOSURE
                  TO HAZARDOUS WASTE INCINERATOR EMISSIONS*
                      F. R. O'Donnell  and G.  A. Holton
                     Health and Safety Research Division
                        Oak Ridge National Laboratory
                         Oak Ridge, Tennessee   37831
                                   ABSTRACT

  The Inhalation Exposure Methodology (IEM)  is a system of  computer  pro-
  grams  developed  to estimate  atmospheric  transport and population expo-
  sure to airborne pollutants  released from  hazardous waste  incinerators.
  This  paper  outlines  the  capabilities of  IEM and discusses operation of
  the version installed  on  the  IBM system at the National Computer Center,
  Research Triangle Park, North Carolina.  Important factors affecting IEM
  exposure estimates  are  discussed.   A  six-site  comparison  shows  total
  exposed population  estimates  made  using two other methods to differ from
  IEM estimates by between  -13  and +9%.   Corresponding exposure  estimates
  show differences between  -36  and +35%.
  INTRODUCTION

       The Inhalation Exposure   Method-
  ology  (IEM) was  developed  to provide
  research   and    regulatory   offices
  within the U.S. Environmental Protec-
  tion Agency (EPA)  a means   to  assess
  the  impacts  of   stack   and  fugitive
  emissions   from   hazardous     waste
  incineration  facilities.    IEM is an
  automated set of   programs   and  data
  for estimating  ambient pollutant con-
  centrations  and   human   inhalation
  exposures in the  vicinity  of  a facil-
  ity.    Only    knowledge    of    the
  facility's  physical  dimensions, pol-
  lutant emission rates, location,  and
  site   climatology   is   required  to
  obtain exposure estimates.
        Research  sponsored  by the  U.S.
  Environmental Protection  Agency under
  Interagency Agreement  DOE  40-1174-81
  and  EPA AD-89-F-1-768-0  under Martin
  Marietta Energy  Systems,   Inc.  Con-
  tract No.  DE-AC05-840R21400 with the
  U.S. Department of Energy -
By acceptance of this article, the
Publisher or recipient acknowledges
the U.S. Government's right to
retain a nonexclusive, royalty-free
license in and to any copyright
COVerinn tha nr+:_i_
         IEM  can  be  employed in a variety
   of  ways.   It  has   been  used by the
   Office of   Solid  Waste   to  estimate
   total   population    exposures   from
   incinerators  which    have    measured
   stack emissions.  It  also can be used
   to compare  the effects on  population
   exposures   of  different stack pollu-
   tion      control        requirements.
   Applications  of these types  can pro-
   vide a measure of the  need  for  and
   effectiveness of regulating hazardous
   waste  incinerator    emissions.     In
   addition,  because IEM uses a sophis-
   ticated  air  dispersion code  which
   allows  separate modeling of  a number
   of  emission  sources  at   a   given
   facility.  the  methodology  has been
   used in comparing the  relative impor-
   tance   of  stack  emisssions  versus
   fugitive emissions on  nearby   popula-
   tions.  This  has allowed researchers
   to determine  the  relative  signifi-
   cance of various emission sources and
   to prioritize future work.

        To date, IEM has  been  used   only
   to  model  incinerator emissions,  but
   its capability of handling  area  and
225

-------
volume   sources  as  well   as  point
source  emissions   should    make    it
applicable  to  other hazardous waste
facilities.  IBM  also  provides   the
user with direct access to historical
weather station data files and  to   a
nationwide,   1980   population  data
base. The  latter   feature   has  been
used  in ground water impact  studies,
where there was a need  to   determine
the  number  of individuals  who might
be exposed to leachates from lagoons
or land disposal facilities.

     IBM has been   installed  on   the
IBM system at the EPA's National Com-
puter  Center  in   Research   Triangle
Park,  North Carolina.  It may be  run
from remote or on-site IBM-accessible
terminals.   Persons  wishing to  use
IBM should arrange  access to  the   IBM
system  and IBM through their Project
Officers.

     The methodology  consists  of   a
system of computer  programs  that uses
on-line meteorological and population
data  bases  and  user-supplied input
data to provide  various  tabulations
of pollutant concentrations  and popu-
lation exposures (O'Donnell   et  al.,
1983).  There are four groups of com-
puter programs (Fig. 1)  and  several
permanently stored  meteorological  and
population data files.  The  MET group
selects  a  meteorological   data   set
from permanently  stored  files  con-
taining  National   Oceanic and Atmos-
pheric Administration Stability Array
(STAR)  data,  and  formats it for  use
in the atmospheric  dispersion program
-  a slightly modified (ISCLTM) long-
term version of the Industrial Source
Complex   Dispersion   Model  (ISCLT,
Bowers et al.,   1979).   ISCLTM  uses
the  meteorological  data  and  other
required input to   calculate  average
ground-level  air   concentrations  of
pollutants   emitted   from   sources
located at the site of interest.  The
POP  group  selects  a  site-specific
population   distribution  from  per-
manently stored,  specially  prepared.
1980   census   population   data   and
formats   it   for   use     in     the
concentration-exposure        program
(CONEX).  CONEX takes the   concentra-
tion  estimates  from  ISCLTM and  the
population distribution from the   POP
group   and  prepares  a  variety   of
tables that allow analysis  of pollut-
ant   concentrations  and   population
exposures around the site.

     The flow  of  program   execution
and  data input in IBM are  controlled
by eight interactive executive  (EXEC)
routines, which must be run in  proper
sequence and within a span  of   a   few
days.   Each  EXEC  routine contains
appropriate job control  and program
statements  to  access  system-stored
data files and programs,  direct   the
user  in  preparation  of   a problem-
specific input data file, and run  the
program  and process program outputs.
The language used to control  an   IBM
session  is  INTERACT  (WYLBUR  6.0).
Because some of the programs must   be
run  sequentially and may have  a long
(greater than eight hour) turnaround,
four  days  are typically required to
make a complete IBM run  on the   BPA
IBM system.

     This paper gives an  outline   of
the  operation  and  capabilities   of
IBM.  Use of IBM is detailed  in   its
users'   guide   (O'Donnell et  al.,
1983).  Several factors affecting  IBM
exposure   estimates  are   also  dis-
cussed.  An indication  of   how  well
IBM  exposure  and exposed  population
estimates agree with  estimates  pro-
duced by other methods is given using
pollutant concentrations at six sites
in  different  parts  of  the   United
States.
COORDINATE SYSTEMS

     IEM uses  two  polar   coordinate
systems  -  the  "grid" system and the
"centroid" system.   Each  system   is
                                     226

-------
                          MET  GROUP
                         USE EXECO. EXECI.
                         AND EXEC2
                            ISCLTM
                          USE EXEC6 AND
                          EXEC7
 POP  GROUP
 USE EXEC3 AND
 EXEC4
                                    CONEX
                                  USE EXECS AND
                                  EXEC7
                                 OUTPUT  TABLES
                                   USE EXEC7
   Figure 1.   Schematic  representation of program group interactions in IEM.
characterized by a  set  of  radial  dis-
tances to rings  circumscribed   about
the origin and a common set  of  angles
representing sixteen  direction  vec-
tors.   The coordinate  systems  have  a
common origin.  The relative  orienta-
tion  of  the  coordinate  systems,  as
built into IEM, is  shown in  Fig.  2.

     The "grid"  system  is represented
by the solid rings  in Fig. 2.   Inter-
sections of the rings and  the   direc-
tion vectors (D1-D16) mark the  points
at which  ISCLTM  calculates  ground-
level   air   concentrations    (e.g.,
points Gl and  G2).   The  "centroid"
system  is  represented  by the  broken
rings.  Intersections of these   rings
and   the  direction  vectors   (e.g.,
point Cl) locate the centers  of  sec-
tor   segments  (represented  by   the
hatched area).  The POP program group
assigns an appropriate  number of  per-
sons to each sector segment   and   the
CONEX  program  calculates an average
air concentration  over  each   sector
segment  using  the  adjacent   "grid"
concentrations.  CONEX then  combines
the population and average concentra-
tion values to calculate exposures  in
each sector segment.
PROGRAM GROUPS

The MET (Meteorological Data) Group

     The MET group consists  of   three
sequentially  executed  computer pro-
grams, each  controlled  by   its  own
EXEC routine.  These routines provide
step-by-step guidance for preparing  a
region-specific  meteorological  data
set for use by ISCLTM.

     The first program,  SERCH   (con-
trolled  by EXECO), locates  and  iden-
tifies  meteorological  weather  sta-
tions near the site being considered.
Stations may be located by   state  or
                                     227

-------
                                                  ORNL-DWG 82-19019
                             D16
       D2
                      D15
                                                      D3
                 D14
              D13
                D12
                    D6
                      D11
               D7
             Figure 2.  The coordinate systems used in IBM.
latitude-longitude   window.   Output
from SERCH consists of the five-digit
station  number,  name  and  location
(city, state, and latitude and longi-
tude),  and  the  Federal Information
Processing  Standards  (FIPS)   state
code  of  each station located in the
search area.  The number of  stations
located  in  the  area  is also indi-
cated.  The user  notes  the  station
numbers  of  all  potentially  useful
weather stations.

     Using the noted station numbers,
the second program, DIREC (controlled
by EXEC1), supplies a description  of
the available STAR data sets for each
potentially useful  weather  station.
In  addition  to the information sup-
plied by SERCH, DIREC lists for  each
data  set  its  header  number,  tape
location information (i.e.,  starting
record number, number of records,  and
tape number), and  a  description   of
the data contained (e.g.,  time  period
covered, seasons or months   included,
etc.).   From  this  information,  the
user selects a data set  for  use   by
ISCLTM.
     The  user  then    supplies    the
information provided by DIREC  for  the
selected data set to   the   next  pro-
gram,  STAR   (controlled  by   EXEC2),
which formats the selected  data   set
for  use by ISCLTM.  This reformatted
data  set  is  stored   in   a    semi-
permanent data file.

The POP  (Population Data) Group

     The POP  group  consists   of   two
sequentially  executed computer pro-
grams, each   controlled  by  its  own
                                     228

-------
EXEC routine.  These  routines provide
step-by-step      instructions      for
transforming  1980  census population
data into a site-specific  population
distribution for  use  by CONEX.

     The first  program,  RD80   (con-
trolled   by   EXECS),   reads   grid-
recorded 1980 census  population  data
for  the  region  surrounding the site
and transforms it into a format  suit-
able  for  use by the second program.
Two types of data sets are  available
as   permanently   stored  files.   A
coarse-grid data  set  contains   esti-
mates of the number of persons resid-
ing in each cell  of   a  6'-latitude x
6'-longitude rectangular matrix.   (At
36 degrees  latitude,  each  cell   is
approximately an  11.01 x 9.01 km rec-
tangle.)  A fine-grid data  set  con-
tains similar estimates for each cell
of a 2'-latitude  x 2'-longitude  rec-
tangular matrix.  Fine-grid data sets
are available for 54  high-population
areas (see Table  3.1  of O'Donnell  et.
al., 1983).  To run EXEC3,  the  user
must  supply  the latitude and longi-
tude of the site  (the origin  of   the
coordinate  systems), the radial dis-
tance to the outermost  "grid"  ring,
and,  if applicable,  the code name  of
the fine-grid data set (listed in  the
users'  guide).   Output from RD80  is
stored as a semi-permanent data  file.

     The second program, APORT   (con-
trolled  by  EXEC4),  is an adaptation
of a computer code written by  Fields
and  Little  (1978).   After the user
supplies,  through EXEC4,  the  number
of "grid"  rings,  the distance to each
ring,  and the latitude and  longitude
of  the  origin,  APORT uses the data
file created by RD80  to  produce   the
population  data  file needed as  input
for  CONEX.   This    semi-permanently
stored  data file contains the number
of persons  located   in  each  sector
segment of the "centroid"  system.
The  Atmospheric  Dispersion  Program
(ISCLTM)

     The atmospheric dispersion  pro-
gram,  ISCLTM, is a slightly modified
version of the Industrial Source Com-
plex  Dispersion  Model  -  Long Term
(ISCLT, Bowers et al.,  1979).   This
program    calculates   the   average
ground-level air-concentration  of   i
pollutant at each "grid"  point around
the chosen site (one or more sources)
which  emits the pollutant.  Input to
ISCLTM  is  supplied  by  the   semi-
permanent  meteorological  data  file
produced by  the  MET  group  and  an
interactively  created data file that
is prepared using EXEC6.  The program
is  run  using EXEC7,  which also con-
trols program outputs.

     ISCLTM and ISCLT are essentially
identical.     All   modifications  to
ISCLT were necessitated by the  addi-
tion of two control switches; they do
not affect the calculations performed
by  the  program.  One switch directs
the program to obtain  meteorological
data  from  the  data file created by
STAR.  The second  switch  creates   a
temporary  output  file  for  use  by
CONEX.

     ISCLT is a steady-state Gaussian
plume  model  and is one of the EPA's
recommended air quality  models.   It
can  account  for  settling  and  dry
deposition  of  particles,  downwash,
and  plume  rise  as  a  function  of
downwind distance.  It can be used to
simulate  the  dispersion of nonreac-
tive gases  or  reactive  gases  that
decay  exponentially.   ISCLT has the
capability to model  numerous  point,
area,  and  volume  emission  sources
within  a  facility   simultaneously.
This  model  is  applicable for sites
having flat to  gently  rolling  ter-
rain;  receptor  elevations may be no
                                     229

-------
higher  than  the height  of  the   short-
est   stack   source  at the  site.   Ter-
rain  elevation  is   ignored   for   area
and volume sources.

      EXEC6   directs   the    user    in
preparation   of  a  permanently  stored
input data file for  use  by ISCLTM.
(See    Bowers   et   al.,  1979   for   a
detailed   explanation    of   input
requirements.)   To coordinate  ISCLTM
with  the other  IBM  progams and  to
reduce  the  quantity of interactively
input data,  many of the  input  vari-
ables   have  been assigned preselected
values  (see  Table  4.1  of  the  IEM
users'   guide).   Most  preevaluated
variables  pertain  to  the   control
switches   and   meteorological   data
descriptors.  Source, site,  and   pol-
lutant  variables  must be entered  by
the user.

      Outputs  from ISCLTM include  the
standard  ISCLT  line  printer output
and the temporary file used  by CONEX.
The   line  printer  output   lists the
input   data   and  gives   tables    of
ground-level  air   concentrations for
each  source  individually and for  all
sources combined.  The temporary  file
contains the  "grid"   system  coordi-
nates and the source-specific concen-
tration arrays.
The  Concentration-Exposure
(CONEX)
Program
     CONEX  performs  several   func-
tions.   It  (1) rewrites the source-
specific concentration estimates from
ISCLTM  into  a  variety  of  tables,
including  tables  of  concentrations
for  selectable  source combinations;
(2) converts concentrations at "grid"
points  (from  ISCLTM)  into  average
concentrations over the  sector  seg-
ments defined by the "centroid"  coor-
dinates; (3) prepares tables of  sec-
tor  segment concentrations; (4)  mul-
tiplies the sector segment concentra-
tions  by  the  number  of persons in
           corresponding  sector  segments  (from
           APORT)    to    produce   sector-segment
           exposures;  (5)  prepares tables of the
           sector-segment      exposures;      and
           (6)  calculates  and  tabulates  various
           combinations   of  the   sector-segment
           exposures.   Input  data is supplied to
           CONEX   from  ISCLTM,  APORT,  and a user
           prepared   input  data   file  that  is
           created  interactively  using   EXECS.
           Program execution  and  output are con-
           trolled by EXEC7.

               EXECS asks the  user  to   select
           the  desired   output  tables  and  to
           enter values  for  14  variables  that
           describe   the   coordinate systems and
           give the  source and  pollutant  names
           and  emission  rates  (see O'Donnell et
           al., 1983).   Some  of these   variables
           must  be   coordinated  with  those used
           in ISCLTM and  APORT.   Available  out-
           put  tables   include  a matrix of the
           number  of persons   assigned  to  each
           sector  segment;   three  sets  of five
           tables, one  set describing   pollutant
           concentrations   at  the "grid"  points,
           one  describing  average pollutant con-
           centrations  over each  sector segment,
           and  one describing  exposures in  each
           sector  segment;  and four tables that
           summarize exposures  by  source,   by
           sector  and source, by  radial band and
           source, and  by  concentration  level
           and  source.
          FACTORS  AFFECTING  IEM EXPOSURE  ESTI-
          MATES

               Factors  affecting  IEM  exposure
          estimates    include    the  calculated
          grid-point     and     segment-averaged
          ground-level   air  concentrations, the
          total  number  of exposed persons,  and
          the  assignment of air concentrations
          to  exposed  persons.   Since grid-point
          concentrations are produced by ISCLT,
          which  is documented  elsewhere (Bowers
          et  al.,  1979  and Bowers and Anderson,
          1981), this aspect of IEM is not dis-
          cussed  here.    The  remaining factors
                                     230

-------
which  interact  strongly are discussed
briefly  below.

     In  IEM,  a   segment-averaged  air
concentraton   is  calculated by taking
the  logarithmic  average of  the  con-
centrations    at    two   grid  points
defined  by  the  intersections  of  the
direction   vector  passing through the
center of  the   segment  and  the  two
 "grid"  rings that bound the segment.
For  example,  the   average  concentra-
tion  for  the  segment  shown in Fig.  2
is obtained from  the  grid point  con-
centrations Gl and G2 .  It is assumed
that all persons  assigned to  a  seg-
ment   are  exposed  to  the  average con-
centration  estimated   for  that  seg-
ment.

     Persons  are  assigned to segments
as follows.  A coarse- or fine-celled
rectangular  matrix  is  superimposed
over   the  area of  interest - the area
bounded  by  the closest  and  farthest
grid   ring.  Some  of  the  cells extend
beyond the  area of interest.    Census
(1980)   enumeration   district popula-
tions have been assigned  to each cell
of the matrix.  The number of persons
in a cell was taken to be  the  total
number   of persons in  all enumeration
districts  that lie within  the  cell.
This  total  number   of  persons  was
assumed  to be  uniformly   distributed
over  the cell area.   The segments of
interest are then   superimposed  over
the cells and persons  are assigned to
a segment in proportion to the  frac-
tion  of  each cell area  that lies in
the segment.  The  resulting   segment
population  is  assumed to  be  distri-
buted uniformly over the  entire   seg-
ment area.
     Population exposures within each
segment  are estimated by multiplying
the number of persons in the  segment
by  the  average air concentration in
the segment.   The  total  population
exposure  is  the  sum of the segment
exposures.
     There  are  other   methods   of
locating  exposed persons and assign-
ing air concentrations which can give
different  results.  To determine the
significance of such  differences,  a
comparison  was  made  between  total
human  exposure  estimates   obtained
using   IEM  and  estimates  obtained
using the Human Exposure Model (HEM),
which   uses   a  simple  atmospheric
dispersion algorithm and a  different
method  than IEM of assigning popula-
tions to air concentration  estimates
(Anderson   et   al.,  1980).   EPA's
Office of Air  Quality  Planning  and
Standards  (OAQPS)  has  used HEM for
screening-level  assessments  similar
to  those  performed  with  IEM.   An
additional comparison  was  performed
to  determine  differences  due to an
alternate,  more  detailed,   complex
method  of assigning population esti-
mates to  each  segment  (Durfee  and
Coleman,  in  press).  Durfee's esti-
mates have been used by  the  Nuclear
Regulatory  Commission  in  licensing
activities.

     To isolate the effect of obtain-
ing and matching population and expo-
sure estimates in  each  methodology,
OAQPS supplied total population expo-
sures  around  six  maleic  anhydride
plants,  which  were calculated using
the population portion of  HEM  only.
(ISCLT  was used to calculate pollut-
ant concentrations.) Results obtained
using  HEM at the six sites are shown
in the right hand column of Table  1.
In the left hand column are the total
exposures using the IEM coarse  popu-
lation  grid,  while  in  the  second
column are total  exposures  obtained
using the IEM fine grid, when one was
available for the site.

     Defining  "Base"   estimates   as
those  made using fine-grid data when
available, or coarse-grid data other-
wise,  IEM  estimates can be compared
with those obtained  using  HEM.   It
can  be  seen from the bottom half of
                                     231

-------
          TABLE 1.  TOTAL EXPOSURE ESTIMATES AT  SIX  SITES  USING  IBM
              WITH APORT (COARSE AND  FINE)  AND DURFEE POPULATION
                   DATA AMD  HEM WITH  EXTRAPOLATED 1970 DATA
     Site
IEM (coarse)   IEM (fine)   IBM (Durfee)
                                                                   HEM
                        Total exposure  (person-ug/m  )
     West  Virginia
     Missouri
     Indiana
     Illinois
     New Jersey
     Pennsylvania
    25,700
    42,900
    79,500
    28,600
    12,000
   251,000
   a
 46,100
   a
   a
 12,500
270,000
 29,000
 50,400
 77,500
 32,400
 11,000
287,000
 33,600
 47,300
 50,900
 38,700
  9,540
267,000
                Total  exposure  (percent  difference from Base)
West Virginia
Missouri
Indiana
Illinois
New Jersey
Pennsylvania
Base
-6.9
Base
Base
-4.0
-7.0

Base


Base
Base
12.8
9.3
-2.5
13.3
-12.0
6.3
30.7
2.6
-36.0
35.3
-23.7
-1.1
           No data.
           Base  = IEM fine  if available,  IEM coarse otherwise.
Table 1 that the total exposures cal-
culated  using  HEM differ by between
-36 and +35 % from those using IEM.

     These differences may arise  for
two    reasons.    First,   the   HEM
population was  an  extrapolation  of
1970  data  to  1980;   IEM  uses 1980
census population data.   As  can  be
seen  in  Table  2, the total exposed
populations estimated by HEM differed
by  between  -13  and  +7%  from  the
"base"   IEM   population   estimates.
Second,  HEM uses a different approach
to determining population  exposures.
It  assumes  that  all  persons  in a
census   enumeration   district   are
located  at  one point and calculates
the  average  concentration  at  that
point  by  extrapolation,  using  the
concentrations at  the  four  nearest
                    grid  points
                    tion point.
                  surrounding the popula-
                         A  comparison  of  total   exposures
                    obtained using Durfee's  data is  shown
                    in column  three  of Table 1.   In   this
                    case differences  in the  estimates are
                    between +13%.   Differences  between
                    the  three  sets  of IEM exposure  esti-
                    mates must  be due  to   differences  in
                    the  population  estimates produced by
                    APORT  (coarse and  fine grid)  and  by
                    Durfee.  This is verified in Table 2,
                    which   shows  percentage  differences
                    (-12  to   +9%)   in  total  numbers of
                    exposed persons  that  generally  agree
                    with  the   percentage  differences in
                    total   exposures   given   in  Table 1.
                    The  lack   of  complete   agreement is
                    explained  by the way  the three  popu-
                    lation  estimation  techniques allocate
                                     232

-------
     Site
            TABLE 2.  TOTAL POPULATIONS WITHIN 20 km OF SIX  SITES AS
             ESTIMATED  BY  APORT (COARSE AND FINE),  DURFEE,  AND HEM
APORT (coarse)   APORT  (fine)   APORT  (Durfee)
                               HEM
                       Total  number  of  persons  (thousands)
     West  Virginia
     Missouri
     Indiana
     Illinois
     New Jersey
     Pennsylvania
        56
     1,242
       110
       154
     1,424
       977
  a
1,248
  a
  a
1,470
1,063
   61
1,284
  110
  164
1,300
1,094
   52
1,338
  108
  160
1,277
1,090
             Total number  of  persons  (percent  difference  from Base)
West Virginia
Missouri
Indiana
Illinois
New Jersey
Pennsylvania
Base
-0.5
Base
Base
-3.1
-8.1

Base


Base
Base
8.6
2.8
-0.5
6.7
-11.6
2.9
-7.3
7.2
-2.1
4.1
-13.1
2.6
            No data.
            Base  = IBM fine if available,  IEM coarse otherwise.
 their  total  populations   among
 various segments used  in IEM.
              the
     These comparisons indicate   that
the choice of distributing population
and computing exposure throughout   an
impacted  area  is unlikely to intro-
duce  differences  in  the  resulting
exposure  estimate  of  greater   than
±50%.  They also show  that  although
there  is  no discernible bias intro-
duced by the APORT, HEM,   and  Durfee
methodologies   over  a  large  range
(56,000 to 1,400,000)  of  population
estimates,  use  of fine-grid data by
IEM produces slightly higher  (up  to
8%) population and exposure estimates
than those  produced  by  coarse-grid
application alone.
                    SUMMARY
                         A brief outline  of   the  opera-
                    tion,  capabilities,  and   input data
                    requirements of IEM is given.  Impor-
                    tant  factors  affecting IEM exposure
                    estimates are discussed.  A six  site
                    comparison  shows  'that total exposed
                    population estimates obtained by  two
                    other  methods  differ from IEM esti-
                    mates   by   only   -13    to    +9%.
                    Corresponding exposure estimates show
                    differences between -36 and +35%.
                                      233

-------
REFERENCES

Anderson, G. E., C.  S.  Lin,  J.  Y.
Holman,  and  J.  P.  Killus.   1980.
Human Exposure tc> Atmospheric Concen-
trations   of   Selected   Chemicals.
Report prepared  under  EPA  Contract
No.   68-02-3066.   Systems  Applica-
tions, Inc.  San Rafael, California.

Bowers, J. F., J. R.  Byorklund,  and
C.   S.  Cheney.   1979.   Industrial
Source Complex (ISC) Dispersion Model
User's  Guide (Volume 1).  EPA-450/4-
79-030.  U.S.  Environmental  Protec-
tion Agency.  Research Triangle Park,
North Carolina.

Bowers, J. F., and  A.  J.  Anderson.
1981.   An  Evaluation  Study for the
Industrial   Source   Complex   (ISC)
Dispersion  Model.  EPA-450/4-81-002.
U.S. Environmental Protection Agency.
Research  Triangle  Park, North Caro-
lina.

Durfee, R. C., and P. R. Coleman.  In
press.     Population    Distribution
Analyses for Nuclear Power Plant Sit-
ing.  ORNL/CSD/TM-197, NUREG/CR-3056.
Oak Ridge  National  Laboratory,  Oak
Ridge, Tennessee.

Fields, D.  C.,  and  C.  A.  Little.
1978.   APORT  -  A  Program  for the
Area-Based  Apport ionment  of  County
Variables  to  Cells .of a Polar Grid.
ORNL/TM-6418.   Oak  Ridge   National
Laboratory.  Oak Ridge, Tennessee.

O'Donnell, F. R., P. M. Mason, J.  E.
Pierce,  G.  A. Holton, and E. Dixon.
1983.  User's Guide for the Automated
Inhalation    Exposure    Methodology
(IBM).     EPA-600/2-83-029.     U.S.
Environmental    Protection   Agency.
Cincinnati, Ohio.  Available from the
National  Technical  Information Ser-
vice, Springfield, VA   22161,  acce-
sion no. PB83-187468.
                                     234

-------
               OPERATION AND MAINTENANCE COST RELATIONSHIPS
                     FOR HAZARDOUS WASTE INCINERATION
                           Robert J. McCormick
                            Acurex Corporation
                         Cincinnati, Ohio   45230
                                 ABSTRACT

     This paper outlines the results of  an  IRB-sponsored study  to
develop relationships between operation  and maintenance  (O&M) costs
for hazardous waste incineration facilities and the various waste-
specific, design-specific, and operational  factors that affect  these
costs.  An overview of the cost estimating  methodology is presented,
followed by a derivation of annual O&M costs  for a hypothetical
incineration facility.
BACKGROUND

     EPA is currently performing
a Regulatory Impact Analysis  (RIA)
of the RCRA performance standards
for hazardous waste incinerators.
One of the key elements of this
RIA effort is development of
representative cost data for
hazardous waste incineration,
including:

 0  Capital costs for new
   facilities designed in
   accordance with RCRA require-
   ments ,

 °  Retrofit costs for existing
   facilities to comply with  the
   RCRA standards, and,

 °  Operation and maintenance  (O&M)
   costs for new or existing  facili-
   ties .

     This cost information is also
needed by IRB to complement tech-
nical/environmental evaluations of
hazardous waste incineration  tech-
nologies, and to aid in identifying
future research priorities.

     This paper focuses on the IRB-
sponsored study of O&M costs  for
hazardous waste incineration
facilities .
OBJECTIVES

     The primary objective
of the study was to develop
relationships between O&M
costs for hazardous waste
incineration and the various
waste-speicific, design
specific, and operational
factors that affect these costs.
These cost relationships were
to be designed so that annual
and unit ($/lb, etc.) O&M cost
estimates could be calculated
for a variety of waste composi-
tions, different  incineration
system operating conditions, and
performance requirements.  This
degree of parametric cost estima-
tion capability was considered
essential for the RIA effort and
for future IRB utilization.

     The second objective of the
study was to derive and arrange
these cost relationships in
computer-ready format, so that
they could be easily programmed
for cost sensitivity analyses
purposes.
                                   235

-------
OVERVIEW
     Based on these objectives,
an O&M cost estimation model has
been developed in computer-
ready, or "workbook" format.
Sequential elements of the
model are as follows:

  (1)  Input data specifica-
       tions

  (2)   Design assumption and
        engineering calcula-
        tions

  (3)  Variable operating cost
       calculations

  (4)  Semi-variable O&M cost
       calculations

  (5)  Fixed cost calculations

  (6)  Energy recovery credit
       calculations (optional)

  (7)  Net, unit O&M cost cal-
       culation

     Input requirements for
the model include basic physical
chemical properties of the
waste(s) in question   plus
a limited number of incineration
facility design and operating
specifications.  From these
input data and numerous technical
assumptions in-line with industry
practice and standard material/
energy balance relationships,
a wide array of engineering
calculations are performed
to estimate the rates  at which
fuel, power, and other chemicals/
utilities are consumed.  These
calculated utility/chemical
consumption rates are  then
multiplied by the projected
annual utilization percentage
for the facility, and then
by unit costs for fuel, power,
etc., to estimate annual totals
for each variable cost element.
Energy recovery credits are
estimated in the same  manner.
although these credits  are
usually incorporated  after
the semi-variable O&M costs
and fixed charges to  capital
are estimated.

     Both maintenance costs
and fixed charges to  capital--
depreciation, insurance,
and taxes--are estimated
as percentages of the depreci-
able fixed capital  investment
(DFCI) for a  given  incinera-
tion facility.  As  discussed
under continuing research
at the end of this  paper,
capital cost  estimating
procedures are currently
being developed to  complement
tlhis" O&M cost estimation
model and provide DFCI  input
values for maintenance  and
fixed cost estimation.

     The final step in  the
O&M cost estimating procedure
is the unit disposal  cost
calculation.  This  is accomp-
lished by summing the annual
varialbe costs, semi-variable
costs, fixed  charges  to
capital, and  energy  recovery
credit, prior to division
by the total  quantity(s)
of waste(s) incinerated
annually.

INPUT DATA SPECIFICATIONS

     For input data specifica-
tion  purposes, wastes  are
divided into  five categories;
combustible organic liquids,
noncombustible liquids  (with-
out support fuel),  pumpable
slurries or sludges,  non-
pumpable sludges or bulk
solids, and containerized
solids.

     The following  data
are requested for each  waste
stream, all of which  must
be relegated to one of  the
                                   236

-------
above categories:

  *   Feed rate, average Ib/hr

  *   Heating value, typical
     Btu/lb

  "   Fractional content of
     carbon, hydrogen, oxygen,
     nitrogen, moisture, ash,
     and chlorine  (as received)

  *   Whether or not alkali
     metals or toxic, heavy
     metals are present in
     the ash

     Certain facility design
and operating conditions must
also be specified  to use the
model, including:

  "   Owner/operator designation

  *   Generic incinerator design-
     liquid injection, rotary
     kiln, or multiple chamber,
     hearth type

  "   Incinerator capacity,
     Btu/hr plus liquid and
     solid feed rate limitations

  "   Incinerator exit gas temper-
     ature (secondary chamber
     temperature for kiln and
     hearth incinerators )

  *   Whether or not a waste
     heat boiler is utilized
     for energy recovery

  "   Particulate emission limita-
     tions , gr/dscf

  *   HC1 emission  control require-
     ments, overall efficiency

     Operating schedule, hours/
     day and days/week

  *   Total annual  on-stream
     time, or percentage utiliza-
     tion
     This input data should
be self-explanatory with
the exception of "owner/operator
designation."  Three distinct
owner/operator scenarios
are provided so that the
economic impacts of regula-
tion can be evaluated across
different segments of the
hazardous waste incineration
user industry.  These three
scenarios are summarized
in Table 1.

     Obviously, certain
waste characteristics and
design alternatives that
can impact costs have been
passed over at this point.
First of all, wastes contain-
ing significant concentrations
of sulfur, phosphorus,  or
halogens other than chlorine
are not provided for.  This
limitation is imposed because
(a) few of these type of
wastes are listed for RCRA
purposes and (b) assessment
of air pollution control
costs would be beyond the
scope of the study if such
waste types were to be  con-
sidered .

     In a similar vein,
air pollution control device
(APCD) selection is limited
to venturi scrubbers for
particulate control and
packed bed absorbers for
acid gas (HC1) removal.
These limitations are real-
istic because venturi and
packed bed scrubbers are
the choice of most designers.
However, there is a recent
trend toward ionizing wet
scrubbers (IWS's) for parti-
culate control in large
facilities to eliminate
the substantial back pressure
associated with venturi
scrubber operation.
                                    237

-------
                                TABLE 1.   OWNER/OPERATOR SCENARIOS

Owner /Operator
Source of waste
Liquid injection
and rotary kiln
incinerator
capacities
Hearth incinerator
capacities
Liquid waste
delivery method
Liquid waste
storage practices
Liquid waste
blending (other than
storage }
Solid waste form as
delivered, stored, and
fired
Scenario A
Institution
(e.g., University
plant or refinery)
100 percent onsite
1 to 10 million
Btu/hr
1 to 10 million
Btu/hr
Small containers,
multiple onsite
sources
Bulk storage,
single tank per
waste type, minimal
segregation (high
Btu/low Btu)
No
Bulk or very small
(5-gal) containers
containers
Scenario B
Industrial Plant
(e.g. , chemical
plant or refinery)
100 percent onsite
10 to 50 million
Btu/hr
10 to 50 million Btu/hr
Pumping from
process (es )
Bulk storage, dual
tanks per waste
type, minimal
segregation (high
Btu/low Btu)
No
Bulk or small
( « 30 gal)
containers
Scenario C
Commercial waste
disposer (corporate
central or semi-
private facility)
100 percent offsite
50 to 100 million
Btu/hr
--
Tank trucks from
offsite (plus
some drummed
liquids )
Bulk storage, dual
tanks per waste
type, extensive
segregation
for blend optimiza-
' tion
Yes
Bulk, small con-
tainersor 55-gal
drums
Solid waste feed
 method
Bottom ash removal
 method

Energy recovery
Quench/scrubber
 water source
Scrubber blowdown
Administrative/
 clerical labor
 required

Frequert Waste
 analyses required

Standard operating
 schedule
                            Semiautomatic
                            Manual
                            No
Municipal supply
                            Municipal sewer
No
No
8 hr/day,
 5 days/week
                         Semiautomatic
                                                     Automatic
                                                     Optional
Onsite WWTP
 effluent, private
 well, or municipal
 supply

Onsite WWTP
 (central facility)
                         No
                         No
24 hr/day,
 7 days/week
                                                  Automatic  for  bulk
                                                   solids, semiauto-
                                                   matic  for drums

                                                  Automatic
Optional, but
 improbable

Local WWTB
 effluent, private
 well, or municipal
 supply

Onsite WWTP
 Specialized or
 central) or
 municipal sewer

Yes
                                                  Yes
24 hr/day,
 7 days/week
                                                 238

-------
     Some large facilities
also use tray tower scrubbers
rather than packed beds for
acid gas removal; however,
the differences in operating
costs are less pronounced in
this case.

DESIGN ASSUMPTIONS AND ENGI-
NEERING CALCULATIONS

     Before the input  data
listed above can be used to
estimate O&M costs, a number
of design assumptions and engi-
neering  calculations are needed
to estimate raw material and
utility consumption rates.
The design assumptions in-
corporated in the model are
too numerous to address in
this forum in their entirety;
however, the major overlying
assumptions are as follows:

  (1)  Either liquid injec-
       tion, rotary kiln, or
       multiple chamber hearth
       incinerators are used.
       Unique, exotic, or hybrid
       designs are not con-
       sidered.

  (2)  Incinerators are equipped
       with sufficient No.
       2 fuel oil firing capacity
       to reach the designated
       operating temperature
       prior to waste injection.

  (3)  Bottom ash from kiln
       and hearth incinerators
       is disposed off-site
       at sanitary or secure
       landfills, depending
       on ash component toxici-
       ties .

  (4)  Firetube and watertube
       waste heat boilers are
       the only energy recovery
       devices considered.
       Where utilized, waste
       heat boilers are located
       immediately downstream
       from the incinerators,
       reducing   combustion
gas temperatures to 550°F.

  (5 )   Waste heat boilers are
       followed by small in-
       line quenches to reduce
       gas temperatures to
       200°F upstream from
       scrubbing devices.  If
       waste heat boilers are
       not employed, larger
       quenches are used to
       achieve the same temp-
       erature reduction.
       Quench feedwater is
       fresh, rather than re-
       cycled, to limit en-
       trainment of dissolved
       solids  as submicron
       particulate.

  (6)   Three air pollution
       control system config-
       urations are considered:
       a)  Venturi scrubber for
           particulate control
       b)  Packed bed absorber
           for HC1 remov al
       c)  Venturi scrubber
           followed by packed
           bed absorber for
           combined particulate
           and HC1 control.

It is  assumed that at least one
air pollution control device
is needed for all hazardous
waste  incineration systems.
Otherwise, the waste would be
a fuel-quality material suit-
able for more profit-able
end use.

  (7)   All scrubbing systems
       have a common sump which
       receives the quench,
       venturi scrubber, and
       absorber effluents.  At
       least 5% of the combined
       effluent is discharged
       to limit solids buildup,
       with the remainder re-
       cycled to the scrubbers.

  (8)   Caustic soda solution
       is used in stoichio-
       metric quantities for
       HC1 scrubbing.
                                    239

-------
  (9)   All  systems  are  assumed
       to be  balanced draft,
       with combustion  air
       blowers  for  the  incinera-
       tors and ID  fans down-
       stream from  the  scrub-
       bing devices.

 (10)   Facility operators
       are  reasonably competent
       in controlling costs.
       For  example, no  auxili-
       ary  fuel is  burned
       if the waste will sus-
       tain combustion  with
       a reasonable excess
       air  allowance at the
       designated incinerator
       temperature.

 (11)   The  overall  system
       is adequately designed
       in terms of  safety
       interlocks and materials
       of construction  to
       prevent  catastrophic
       failure.

These  design assumptions are
typical of  good practice in
the hazardous waste incineration
user industry.   Therefore,
they provide a realistic basis
for O&M cost estimation for
the industry as a whole, which
is the goal of the  RIA. How-
ever,  these assumptions do
not reflect current practice
for all facilities, so  O&M
cost estimates  derived from
the model may not be represen-
tative for  specific facilities.
This point  is discussed further
at the conclusion  of the paper.

     Based  on these design
assumptions and the input
data previously described,
numerous engineering calcula-
tions  are performed to  determine
raw material and utility consump-
tion rates.  For the most
part,  these are standard material
and energy  balance  calculations
as summarized below.
   Front-end  storage  and hand-
   ling  equipment  operation
 -  Tank  agitator power require-
   ments
 -  Liquid  nitrogen requirements
   for tank blanketing
 -  Steam requirements for tank
   heating
 -  Liquid  waste  transfer and
   feed  pump  power requirements
 -  Fuel  consumption by solid
   waste transfer  vehicles
 -  Solid waste conveyor/
   feeder  power  requirements
 -  Atomizing  air compressor
   power requirements

'   Incinerator operation
 -  Waste component feed rates
   and gross  heat  input
 -  Low-fire fuel consumption
   for flame  stabilization
 -  Supplemental  fuel  and total
   air feed requirements
 -  Fuel  oil feed pump power
   requirements
 -  Combustion air  blower power
   requirements
 -  Combustion gas  flow and
   composition
 -  Particulate loading in
   combusion  gases
 -  Charge  stoking  and bottom
   ash handling  requirements
   for solid  waste incinerators
 -  Start up fuel "  requirments

   Waste heat boiler  operation
 -  Steam generation rate
 -  Fuel  conservation  rate

'   Quench  operation
 -  Heat  duty  and water require-
   ments
 -  Feed  pump  power requirements

'   Scrubber system operation
 -  Venturi scrubber pressure
   drop  requirements
 -  Venturi scrubbant  feedrate
   requirements
 -  Venturi scrubbant  recycle
   pump  power requirements
 -  Venturi scrubber effluent
   and exit  gas  flowrates
                                   240

-------
 - Slowdown rates and scrubber
   system makeup water require-
   ments
 - Makeup water pump power
   requirements

'   ID fan power requirements

VARIABLE OPERATING COST CAL-
CULATIONS

     The variable operating
cost elements addressed in
the model are as follows:

  (1)  Fuel

       Natural gas for flame
       stabilization
       No. 2 fuel oil for
       supplemental heat input
       No. 2 fuel oil for
       startup
       Propane for lift
       truck operation

  (2)  Power

       Liquid waste/fuel oil
       feed pumps
       Quench/scrubber system
       pumps
       Solid waste conveyors
       and feeders
       Solid waste incinerator
       mechanical requirements
       (e.g., Kiln rotational
       drive)
       Combustion air blower(s)
       Atomizing air compressor
       ID fan(s)
       Tank agitation

  (3)  Water

       Quench/scrubber system
       makeup
       Ash quenching

  (4)  Caustic soda solution
       (50 wt%) for acid gas
       scrubbing

  (5)  Liquid nitrogen for
       tank blanketing
  (6)  Steam for tank heating
       (expressed as additional
       fuel cost for the main
       plant boiler complex)

  (7)  Ash and scrubber blow-
       down disposition

     Annual costs for these
utilities and supplies are
determined from th>3 results
of the preceding engineering
calculations (which predict
hourly consumption rates),
the projected annual utilization
precentage for the facility,
and unit costs for fuel oil,
electric power, etc.  In the
first generation model, Houston
area costs alone were included.
Since then, an option has  been
designed in so that East Coast
(New Jersey), Midwest (Chicago),
or West Coast (Los Angeles)
costs can be used in place of
the Houston, Gulf Coast costs.
SEMI-VARIABLE O&M COSTS

     The semi-variable O&M cost
elements include:

  (1)  Operating labor and
       supervis ion
  (2)  Maintenance materials
       and labor
  (3)  Incoming waste analyses

     Unlike the variable cost
elements such as power and water,
these costs cannot be annualized
as a function of the total
facility on-stream time.  Labor
costs are relatively consistent
throughout the 12-month period,
depending more on (a) the hours/
day, days/week operating schedule
(b)  whether or not solid wastes
as well as liquid wastes are
handled, (c)  the degree of
automation employed, and (d) how
closely the incineration opera-
tion is tied to the waste
                                   241

-------
generating operations, if at
all.   The model presents six
scenarios for heuristic labor
cost  estimation, taking each
of these factors into account.
These scenarios are summar-
ized  in Tables 2-4.

     For most hazardous waste
inceration facilities, mainten-
ance  costs are the most diffi-
cult  to predict apriori.
These costs are influenced by
so many detailed waste-vs-equip-
ment  design considerations that
parametric evaluation is almost
impossible.  Moreover, the
major maintenance requirements
such  as refractory replacement
are affected by subjective
criteria such as adequacy of
the original design and operator
experience.  Therefore, the
simple approach is used for
maintenance cost estimation
based on percentage of the
depreciable  fixed capital
investment.  The multipliers
are 5% for liquid injection
systems, 7% for multiple
chamber hearth systems, and
10% for rotary kiln systems.

     Annual costs for analysis
of incoming wastes are wholly
dependent on the nature of the
waste generating operation(s).
These costs may run from essen-
tially zero to the burdened
labor cost for one technician,
40 hours/week.
FIXED COST CALCULATIONS

     Depreciation, insurance,
and taxes are all estimated as
percentages of the depreciable
fixed capital investment for
the facility.  Ten-year, straight
line depreciation is assumed
with essentially zero salvage
value, while insurance and
taxes are assumed to total 4%
of the depreciable fixed
capital investment.
ENERGY RECOVERY CREDIT
CALCULATION

     For the purposes of this
model, the only energy re-
covery option considered is
steam generation in a waste
heat boiler.    The value
of the steam generated is
calculated in terms of
potential fuel  savings for
the plant-wide boiler complex,
assuming equivalent steam
enthalpies as-delivered, 80%
fuel -to-steam efficiency in
the boiler complex, and that
No. 2 fuel oil is utilized.
UNIT DISPOSAL COST CALCULATION

     The unit cost for waste
disposal by incineration is
easily calculated at this
point by the relationship:

 (Annual Variable Cost) +

 (Annual Semi-Variable Cost) +
 (Annual Fixed Cost)
 (Annual Energy Recovery
  Credit)


(Annual Quantity of Waste
 Incinerated)

     This is the bottom-line
figure for comparing one
incineration O&M cost esti-
mate to another.

EXAMPLE

In order to illustrate the
inputs and outputs of the
model, the following, simpli-
fied example is provided.

Input Data Specifications

Waste description:  Mixture
 of contaminated solvents and
 process byproducts.
Normal feedrate :  2000 Ib/hr
Heating value:     8000 Btu/lb
                                   242

-------
TABLE  2.   ESTIMATED LABOR REQUIREMENTS FOR  CATEGORY A  FACILITIES

Labor
category
Process
operator
Forklift
operator
Yard
laborer
Engineering
supervisor
Liquid
Number
per
shift
1
0
0
0.25
wastes only
Number Total
of number of
shifts personnel
1 1
0 0
0 0
1 0.25
Liquid
Number
per
shift
1
1
1
0.25
and solid wastes
Number
of
shifts
1
1
1
1
Total
number of
personnel
1
1
1
0.25
TABLE  3.   ESTIMATED LABOR REQUIREMENTS FOR  CATEGORY B  FACILITIES
Labor
category
Process
operator
Forklift
operator
Yard
laborer
Engineering
supervisor
Liquid
Number
per
shift
1
0
0
0.5
wastes
Number
of
shifts
4
0
0
1
only
Total
number of
personnel
4
0
0
0.5
Liquid
Number
per
shift
1
1
1
0.5
and solid
Number
of
shifts
4
4
4
1
wastes
Total
number of
personnel
4
4
4
0.5
                                243

-------
TABLE 4.   ESTIMATED  LABOR REQUIREMENTS  FOR CATEGORY  C FACILITIES


Labor
category
Process
Operator
Forklift
Operator
Yard
laborer
Clerical
Engineering
Supervisor
Liquid
Number
per
shift

3
2

0

0
0.5
1
Administrator 1
wastes only
Number
of
shifts

1
3

0

0
1
1
1
Total
number of
personnel

9

0

0
0.5
1
1
Liquid and Solid wastes
Number
per
shift

3
2

1

1
0.5
1
1
Number
of
slhifts

1
3

4

4
1
1
1
Total
number of
personnel

9

4

4
0.5
1
1
                                244

-------
Composition: Carbon  54.5%
            Hydrogen  3.5%
            Chlorine  5.0%
            Moisture 19.5%
            Ash       0.5%
            Oxygen    9.0%
            Nitrogen  8.0%

Owner/operator designation:
 B (chemical plant)
Generic incinerator design
 type:   Liquid injection
Total thermal capacity:
 20M Btu/hr
Operating temperature:
 2000°F
Energy recovery utiliza-
 tion:   No
Maximum particulate loading:
 0.08 gr/dscf
HC1 removal efficiency:  99%
Operating schedule:  24 hr/day,
 7 day/week
Annual utilization:  80%
Location:  Houston, Texas
Estiraa ted depreciable fixed
 capital investment:  $1.5M

Raw Material/Utility Consump-
tion Rates

     Based on the  input data listed
above,  estimated raw material
and utility consumption rates for
the hypothetical facility are
summarized in Table 5.

O&M Costs

     Annual O&M costs and credits
for the hypothetical facility are
summarized in Table 6.  The unit
costs are in mid-1982/early 1983
dollars and are representative of
the Houston area.

Unit Disposal Cost

The unit disposal  cost is given
by the  expression,

Unit disposal cost, $/lb.=
Net,  annual O&M cost, $	
Total annual waste throughput, Ib.

For this hypothetical case, the
unit disposal cost is approxi-
mately $0.053/lb.

CONCLUSIONS

     The O&M cost estimation
model for hazardous waste 1-1-
cineration described and il-
lustrated in the preceding
pages should be accurate to
w-ithin ± 30-40% for all but
the most unusual cases.  This
is consistent with normal con-
ceptual design goals for cost
estimating accuracy, and
should be suitable for pur-
poses of the RIA as well as
for EPA research planning
purposes.   This model should
also be useful to waste gen-
erators making first-cut cost
comparisons between on-site
incineration and other, off-
site disposal options.
However, the model is limited
in its ability to accurately
predict site-specific costs,
so it is not recommended for
cost estimating purposes
beyond the conceptional
design stage.

CONTINUING RESEARCH

     In FY 1983, the existing
O&M cost model will be expanded
and augmented in the following
manner:

  (1)  A corollary capital cost
       estimation model for
       hazardous waste incinera-
       tion will be finalized.

  (2)  The air pollution
       control system options
       will be expanded to
       include IWS-based systems
       as  well as the more
       standard venturi scrubber/
       packed bed absorber
       system.

  (3)  Capital and annual cost/
       credit estimation models
       will be developed for
       other thermal destruction
                                    245

-------
           TABLE  5.   ESTIMATED RAW MATERIAL/UTILITY REQUIREMENTS
Item

Fuel
 Natural gas for flame
  stabilization
 No. 2 fuel oil for startup

Power
 ID fan
 Compressor
 Blower
 Pumps
 Agitators
 Total

Water

Caustic soda solution
    (50 wt %)

Liquid nitrogen

Sewer use
                                  Normal Rate
                   Total annual  quantity
    1000 scfh
     110 gal/startup


     95 hp
     70 hp
     35 hp
     20 hp
      nil
    220 hp

    110 gpm

    230 Ib/hr
      38 ft3/hr
     110 gpm
     7M ft -
   1400 gal
       1.15 Gwh

       48 M gal

       1.6 M Ib


      270 M ft3

        45 M gal
             TABLE 6.  ESTIMATED ANNUAL O&M COSTS AND CREDITS
Item

Natural gas

No. 2 fuel oil

Power

Water

Caustic soda solution
     (50 wt %)

Liquid nitrogen



Sewer

Labor

Maintenance

Depreciation

Insurance/taxes

   TOTAL
Unit Cost

$5.00/1000 ft3

  0.85/gal

  0.061/kwh

  0.25/1000 gal

  0.07/lb


  1.06/100 ft3
  (+$2,400/yr tank
   rental)

  1.29/1000 gal
  See table 3
Annual Cost

 $35,000

   1,200

  70,000

  12,000

 110,000


   5,200



  58,000

 160,000

  75,000
 150,000
  60,000

$736,400
                                  246

-------
techniques such as boiler and
cement kiln co-firing.

     Finally, an effort is
underway to validate both the
existing O&M cost model and
the other models now being
developed.
                                   247

-------
         RETROFIT COST  RELATIONSHIPS  FOR  EXISTING HAZARDOUS
                    WASTE INCINERATION FACILITIES


                         Robert J. McCormick
                         Acurex Corporation
                       Cincinnati, Ohio  45230

                               ABSTRACT

     This paper outlines the results of an IRB-sponsored study
o<- potential retrofit costs for hazardous waste incineration  facili-
ties.   Cost relationships are presented for major capital additions
or modifications that could be required to bring existing facilities
into compliance with RCRA performance regulations.  A hypothetical
retrofit cost scenario is also presented.
BACKGROUND

     EPA is currently performing
a Regulatory Impact Analysis
(RIA) of the RCRA performance
standards for hazardous waste
incinerators.  One of the key
elements of t lis RIA effort
is development  of representa-
tive cost data  for hazardous
waste incineration, including:

  *  Capital costs for new
     facilities designed in
     accordance with RCRA
     requirements,

  *  Operation  and mainten-
     ance (O&M) costs for
     these facilities, and

  *  Retrofit costs for existing
     facilities to comply with
     RCRA standards.

     This paper describes the
IRB-sponsored study of retro-
fit costs for hazardous waste
incineration facilities
PURPOSE AND SCOPE

The objective of this study
was to develop a method-
ology, and an accompanying
set of empirical cost relation-
ships, that could be used to
estimate the costs of retro-
fitting/upgrading various
components of existing
hazardous waste incineration
facilities to comply with
RCRA performance requirements
Both the methodology and
the retrofit cost relation-
ships were intended to
focus on major capital
additions or subsystem
modifications that could
be required for existing
facilities to:

  (1) Increase destruction
      and removal efficiency
      (ORE) of the principal
      organic hazardous
      constituents (POHC's)
      in the waste feed,

  (2) Reduce particulate
      loading in the stack
      gas to  -=0.08 gr/dscf,
      and/or

  (3 ) Increase HC1 removal
      to  =-99% in facilities
      burning a waste mix
      containing 0.5%
      organic chlorine.

     Because the performance
status of many incineration
facilities is unknown,
particularly with respect to
DRE, it was not possible
                                   248

-------
to predict within the framework
of this study what the actual
retrofit requirements for vari-
ous segments of the incinerator
population might be in order
to comply with RCRA standards.
In all likelihood, many exist-
ing facilities will require
no physical modification to
meet these standards.  Other
facilities may require extensive,
multiple component  modifications.
For still others, retrofit may
not be feasible because of
space or equipment design limita-
tions.  Therefore, this study
was not designed to predict
what the total retrofit costs
would be for the hazardous waste
incineration user industry
to comply with RCRA require-
ments.  Rather, the results
of the study were intended
as a cost estimating tool for
EPA decision making purposes.
The scenario envisioned for
application of the methodology
and cost relationships developed
in his study was, "If one or
more capital additions/modifi-
cations are required for
Facility XYZ to achieve RCRA
compliance, and Facility XYZ
has the following design/
operational characteristics,
what will it cost to make the
necessary modifications?"

     At the onset of the study,
it was recognized that major
capital additions or modifications
were not the only types of
retrofit costs that may be en-
countered by facilities upgrad-
ing performance.  Others in-
clude minor finetuning adjust-
ments, downtime-related costs,
and increased O&M costs.
However, these costs could
not be quantified within
the framework of this study.
OVERVIEW  OF  RESULTS

     The  results  of  this
study are expressed  in  a
series  of empirical  (graphi-
cal) relationships between
the costs for  various capital
modifications/additions and
factors that  significantly
impact  these  costs,  e.g,
capacity, materials  of  con-
struction, etc.   These
curves  were  derived  as  the
result  of queries to a  number
of vendors of  incinerator
equipment.   Costs are
developed for:

*  Combustion  system retrofit
      Burner  replacement
      Refractory  replacement
      Combustion  chamber re-
      placement
*  Quench and/or  waste  heat
   boiler addition
   Scrubber  system addition,
   replacement, or modifica-
   tion
*  Flue gas  handling system
   modification
      Fans,  stack, etc.
*  Total  system replacement

     In addition  to  the cost
curves  themselves, guidelines
are presented  to  aid the user
in determining when  particular
retrofit  activities  need to
be considered, what  types of
input data are needed to use
the various  cost  curves, and
how installation, indirect
construction  costs,  and
contingencies  can be factored
in.

     The  cost  relationships
and associated information
are designed  to cover as
broad a range  of  incinerator
•facility  scenarios as possible
although  certain  cases  may
                                  249

-------
not be addressed.   A wide range
of possible waste  compositions
are considered; hydrocarbon-
based mixtures with variable
heating values, moisture con-
tents , ash contents and composi-
tions (including alkalis),
and chlorine concentrations.
Liquid injection,  rotary kiln,
and hearth-type incinerators
are all addressed  in capacities
ranging from 1-100 M Btu/hr.
Both quenches and  steam-generat-
ing waste heat boilers with
or without economizers are
considered for gas temperature
reduction, venturi scrubbers
are assumed for particulate
control, and packed bed absor-
bers are assumed for HC1
removal.  Uncontrolled pollutant
concentrations entering the
air pollution control system
are assumed to range from
0-2 gr/dscf for particulate
and -=2 vol % for HC1.  These
ranges are believed to cover
the range of conditions ex-
perienced in existing hazard-
ous waste incineration
facilities.

     The following sections
describe the cost  relation-
ships for various  retrofit
activities in more detail.

COMBUSTION SYSTEM  RETROFIT

     The primary driving
force considered in this study
for combustion system retrofit
was to increase destruction
effeciencies (DE's) for POHC's
contained in the waste.  At
the present time,  insufficient
data is available  to relate
DE's directly to incinerator
design and operational re-
quirements.  Therefore, this
study focused on major capital
additions or modifications
that might be needed to raise
incinerator temperature above
original design specifications
and/or to increase effective
residence time, mixing ef-
ficiency, etc.

     The first potentially
major cost ite.m considered
was burner system replacement
for improved combustion
efficiency or increased
fuel co-firing capability
to elevate temperature.
The major problem encountered
in estimating the costs
for this activity was that
high-efficiency burners
capable of handling multiple
liquid waste streams plus
support fuel a.re almost
always custom designed and
fabi: Lcated.  Thus, the costs
are quite case-specific
and difficult for manufac-
turers to generalize.  The
alternative adopted for
this study was a baseline
costing approach whereby
a purchased cost vs. capacity
curve was developed for
burner systems capable of
firing waste oils.  This
curve is shown in Figure
1.  Burner auxiliaries such
as blowers, dampers, flame
safeguards, and combustion
controls are included in
the costs.  Installation
is assumed to be 50% of
purchased cost.  A major
underlying assumption is
that the burner system is
physically compatible with
the combustion chamber con-
figuration.  If not, more
extensive retrofit activities
are required as will subse-
quently be described.

     If incinerator temperature
is increased substantially
above the original design
specifications, it may be
necessary to replace the
existing refractory lining
                                    250

-------
o
o
o
   70
   60 -
   50
   40
o
0



S  30
TO
-C
(J
   20
   10
                                      I
I
             10      20      30      40       50      60      70


                               Burner capacity (million Btu/hr)
               80
90
TOO
                   Figure 1   Purchase cost of new burners  (July  1982)

-------
with a higher grade material.
For the purposes of this study
approximate refractory replace-
ment costs are estimated by
first calculating the material
requirements, then judging
the type of refractory required
and its cost, and finally,
factoring in labor costs for
removal of the old  lining
and installation of the  higher
quality material.

     The volume  of refractory
required for a given application
is estimated, in brick equiva-
lents (9 in. X 4.5 in. X 3 in.)
from the thermal capacity of
the system, typical state-
of-the-art heat release rates
and residence times for the
three generic incinerator designs
considered, typical dimensions
for these generic designs
(length:  diameter, surface:
volume), and simplified thickness
vs. temperature guidelines.
For a typical 30 M Btu/hr liquid
waste furnace, the design assump-
tions would be a 30,000 Btu/hr
ft  heat release rate, a 3:1
length-to-internal-diameter
ratio, a 4.5-6 in. inner lining
of firebrick, and a 2.5-3 in.
outer lining of insulating
refractory.

     Refractory "type" (brick
vs. castable, alumina content)
and unit cost are then estimated
based on temperature application
guidelines, plus the qualitative
presence or absence of alkalis
and/or chlorine in the combus-
tion environment.  Usually
a 45% alumina refractory is
satisfactory for low temperature
(1400-1800°F) applications.
For temperature up to 2400°F
and/or corrosive environments,
a 60-80% alumina refractory
is normally specified.  For
exotic applications above 2400°F,
a 90% alumina content is needed.
Costs range  from less than
$1 per brick equivalent to
more than $10 per brick
equivalent.

     Total material costs
are then determined by com-
bining the estimated volume
requirements in brick equiva-
lents and the dollar per
brick equivalent cost for
an appropriate refractory.
A range of installed vs.
material cost multipliers
are provided to estimate
the final installed cost,
which is affected by local
labor costs, ease of access
to the combustion chamber
interior, and other site
specific factors.  Installa-
tion-to-material-cost ratios
can range from 1 to 4.

     In many cases, it may
not be feasible to replace
only the burner system or
only the existing refractory,-
complete combustion system
replacement may be required
to significantly improve
performance.  For example,
a substantial increase in
operating temperature may
require a thicker refractory
lining to limit skin tempera-
ture.  This increased refrac-
tory thickness reduces internal
volume and residence time.
If the residence time reduc-
tion is significant enough
to impact DE, a larger shell
and, thus, a new combustion
chamber is required to provide
sufficient residence time.

     In Figures 2-4 equipment
cost vs. capacity curves
are presented for liquid
injection, rotary kiln/after-
burner, and fixed hearth/
afterburner combustion
systems.  The costs include
burner systems, as previously
described, refractory lined
shell, auxiliaries, and
controls.  Feed system costs
are not included.
                                    252

-------
1,000
                          4      6
                       Capacity, QTfnax
10        20
(mill ion Btu/hr)
50
Figure 2.  Purchase cost of liquid injection incinerators (July 1982).
                                   253

-------
     o
     o
     o
     to
     o
     u
     o
     i-
10,000






 6,000




 4,000



 3,000





 2,000








 1,000





   600





   400








   200







   100
                 I  I  I  I
                     8  10
                           I   i   i

                          20
                                             i    i  i   i  i i
40
60
100
                        Capacity,  C>  (million Btu/hr)
Figure 3,  Purchase cost of rotary kiln incinerators  (May  1982)
                                 254

-------
   1,000
     600
     400
     200
o
o
o
**•   100
o
o
3

Q_
60




40







20
      10
                   l   l   l
                            _L
                          J	III!
_L_J	L
J	L
                                                                        l  I  i
                             4     6       10        20

                               Capacity, Qy  (million Btu/hr)
                                                          40     60
                        100
       Figure 4.   Purchase cost of multiple-chamber,  hearth  incinerators

                  (July 1982).
                                        255

-------
     For new sytems, installa-
tion costs range from 25% to
100% of the  purchased cost.
A retrofit installation cost
will approach the upper end
of this range because the old
unit must be removed.

QUENCH/WASTE HEAT BOILER
ADDITION

     If air pollution control
devices (APCD's) such as venturi
scrubbers or acid gas absorbers
need to be added to existing
incineration systems to comply
with RCRA emission standards,
some means of cooling the combus-
tion gases prior to APCD entry
must also be  provided.  Two
alternatives are considered
in the study:

   (1)  Direct water-spray quench-
        ing to <200°F, and

   (2)  Waste heat boiler and
        post-boiler quench ap-
        plication to achieve
        the same temperature
        reduction.

     Separate capital cost
vs.  gas flow rate curves are
provided for high temperature
quenches and for smaller, post-
boiler quenches in Figures 5 and
6.  Costs for high temperature
quench towers are based on the
assumption of 1800-2200°F inlet
gas  temperature and, thus, int-
erior refractory lining.  Acid-
resistant design is also assumed,
although separate costs are
presented for extremely severe
service applications.  Inlet gas
temperatures of 400-600°F and acid-
resistant alloy construction are
assumed for the smaller quenches.
Installation costs are usually
30-40% of the purchased cost.
     Equipment cost vs. gas
flow rate curves are provided
for waste heat boilers in
Figure 7-  These costs are
for packaged boilers with
standard trim and controls.
Installation costs range
from 30% to as much as 200%
of the purchased cost depend-
ing on retrofit difficulty.

SCRUBBER SYSTEM ADDITION/
REPLACEMENT/MODIFICATION

     In order to meet RCRA
standards for particulate and
HC1 removal, existing hazard-
ous waste incineration facility
retrofit requirements may
range from virtually nil to
complete particulate and acid
gas scrubbing system addition.
In terms of major capital
additions or modifications,
however, four retrofit
scenarios were selected for
the purpose of this study.
These are:

   (1)  Venturi scrubber addi-
        tion/replacement for
        improved particulate
        collection,

   (2)  Conversion from once-
        through water absorp-
        tion of acid gases to
        a caustic recycle system,

   (3)  Acid gas absorption
        column addition/re-
        placement ,

   (4)  Total scrubbing system
        addition -- venturi
        scrubber and caustic re-
        cycle acid gas absorp-
                                  256

-------
no
en
—i
                      o
                      o
                      o
1/1
o
u

oj
VI

-------
                    100


                     80




                     60-
ro
en
CD
                  o
                  o
                  CD
                  O
                  u
                  o
                  S-
                  13
                  Q-
                     10
                                                  _L
I
   _L

                                                  4       6     8   10

                                                  Inlet gas flowrate, F
                                                                       TG
                           20

                    (1,000 Ib/hr)
40
60    80   TOO
                                   Figure 6.  Purchase cost of low-temperature  quenches  (July 1982).

-------
in
10
                            o
                            o
                            o
                             o
                             u
                                1,000


                                 800


                                 600



                                 400
200






100


 80


 60



 40






 20





 10
           & Watertube, severe service



           O Watertube


           Q Firetube
                                              ill   ii
                                                              I  ill
II   II
                                                                                           I  I I
                                                                                                        I	I
                                                       46810       20       40    60   80 100

                                                    Inlet Gas Flowrate, (FTG)J (1,000 Ib/hr)
                                                                     200  300
                                       Figure 7.  Purchase cost of waste heat boilers (July 1982),

-------
    tion system,  plus fan and
    stack.

     Purchased costs for complete
scrubbing systems,  including
flue gas handling equipment,
are presented in  Figure 8.
These costs are for typical
30"WC back pressure systems.
For 100"WC pressure drop systems,
the costs shown in Figure 8
can easily double.   Conversely,
the costs for 5"  WC pressure
drop systems (no  venturi
scrubber) are approximately
15% less then those presented
in Figure 8-  If  no acid gas
absorption system is needed,
the cost reduction is approxi-
mately 40%.  These guidelines,
along with the component cost
breakdown presented below,
can be used to estimate equip-
ment costs for virtually any
retrofit scenario.

  Venturi and wetted
   elbow                 9%
  Cyclonic separator and
   integral packed tower
   absorber             30%
  Caustic system        17%
  ID fan                18%
  Stack                 10%
  Ductwork and piping    6%
  Platform and founda-
   tions                 4%
  Instrumentation and
   controls              6%

     Installation costs for
scrubber system retrofit are
quite site specific, however,
50-100% of the purchased
equipment cost is a likely
range.
FLUE GAS HANDLING SYSTEM
 MODIFICATIONS

     In certain situations
particulate collection ef-
ficiency in the venturi
scrubber may be limited because
the fan capacity is insuffi-
cient to handle the combustion
gas flow at the pressure drop
necessary for good venturi
performance.  If this is
the case, then particulate
emissions can be r^di^^d
(without reducing waste   throughout)
by simply replacing the
fan.

     Purchased costs for
carbon steel and corrosion
resistant frns are presented
in Figure 9 and 10,  respective-
ly.  Fan installation costs
are relatively independent
of capacity, usually running
$20,000 to $30,000.

     In Figure  11,  cost
vs. height relationships
are presented for stacks
of various diameter.  Although
stack replacement will not
reduce emissions in itself,
this  retrofit scenario was
considered at the request
of the Office of Solid Waste
for the purposes of their
dispersion model-based risk
assessment activities.
Increased stack height re-
duces the maximum ground
level concentration of
emitted species, so the
costs for adding taller
stacks are needed to perform
cost/benefit tradeoff analy-
ses.

     It should be noted
that the costs presented
in Figure u  are purchased
costs for FRP-lined stacks
designed to receive low
temperature gas exiting
the scrubber system and
fan.   Erection costs are
approximately eaual to the
fabricated material costs
in all cases.
                                   260

-------
                         o
                         o
                         o
r-o
en
                          O)
                          -C
                          u
                          1-
                          Ol
                          on


                          CD
                                                     4    6     10       20       40


                                                    Inlet gas flowrate, (1,000 acfm)
60
100
200  300
                                 Figures.  Purchase  cost of scrubbing systems receiving  1,800°  to  2,200°F

                                            gas  (July 1982).

-------
  1,000
o
o
o
      1,000
     10,000



Gas flowrate (acfm)
100,000
   Figure   9.  Purchase cost  of carbon  steel fans  (July 1982)
                                   262

-------
    1 ,000
   o
   o
   o
      100
       10
        1,000
               AP

               +    20-in. W.C



               ©    30-in. W.C


               •    40-in. W.C



               A    60-in. W.C


               Q - 100-in. W.C



               O   160-in. W.C
      10,000



Gas  flowrate (acfm)
                                                      I    I   I  I  I  I I
100,000
Figure 10   Purchase cost  of corrosion-resistant fans (July 1982)
                                     263

-------
    50
    40
o
o
o
30
s-
-Q
   20
   10
                                          O
                                                 1 ft

                                             diameter
                      50             100


                         Stack height (ft)
                                               150
  Figure   11  Fabricated cost of FRP  stacks (July 1982)
                               264

-------
INDIRECT COSTS

     In addition to the direct
costs for equipment and instal-
lation, the indirect costs
associated with engineering,
construction, and startup must
be considered.  For the purposes
of this study, indirect costs
are estimated as percentages
of the total direct cost.  These
percentages are as follows:

  Engineering            10%
  Construction over-     10%
   head
  Construction fee        8%
  Startup                 2%

     Thus,  the indirect costs
are estimated to total approxi-
mately 30% of the direct cost
for a given retrofit activity.

EXAMPLE

     In order to illustrate
how the information presented
above can be used to estimate
costs for major retrofit activi-
ties, the following example
is provided.
Basis

     A small multiple chamber
hearth incinerator is being
used to dispose of liquid process
wastes and plant trash.  The
toxic components of the liquid
waste are not difficult to
destroy, so the unit is achieving
99.99% destruction efficiency.
However, the system was in-
stalled prior to implementation
of air emission standards, so
no pollution controls are pro-
vided.  Combustion gas is
vented directly to a refrac-
tory-lined stack.  As a result,
the unit exceeds RCRA emission
standards for both particulate
and HC1.

Retrofit Requirements

     In order to achieve
compliance,  the existing
stack must be bypassed and a
complete scrubbing system --
venturi scrubber, HC1 ab-
sorber, fan, and stack --
must be added.  The mean
particle diameter in the gas
is approximately 2 urn, so a
30" WC back pressure system
is adequate.  In addition,
quenching can be accomplished
in the venturi inlet.  Space
is available for the scrubb-
ing  system, so no special
retrofit  difficulties are
encountered.

Costs

     The combustion gas flow
from the secondary chamber is
10,000 acfm at 1600°F.  There-
fore , from Figure 8 the pur-
chased cost  for the scrubbing
system is approximately
$100,000.  Installation runs
about 50% of the equipment
cost, so the total direct
cost is $150,000.  Adding
30% for indirect costs, the
total capital expenditure is
$195,000.
CONCLUSIONS

     The study described in
this paper is a basic, first
cut effort to estimate
potential costs for hazardous
waste incineration facility
retrofit.   Because of the
many site specific factors
that impact retrofit costs,
the accuracy of the estimates
may be no better than - 50%
                                    265

-------
to 100% for some facilities.
Large discrepancies between
projected costs and actual costs
are most likely in situations
where space is limited, service
relocations are  required,
interferences are encountered,
or structural relocation is
required.  Where these problems
are not encountered, the esti-
mating methods described in this
paper may achieve conceptual
design accuracies of ± 30-40%
                                   266

-------
                    FULL SCALE DEMONSTRATION OF WET AIR OXIDATION
                      AS A HAZARDOUS WASTE TREATMENT TECHNOLOGY

                                   Dr. Milli am Copa
                                   James Heimbunch
                                   Phil lip Schaefer
                                     Zimpro Inc.
                             Rothschild, Wisconsin  54474
                                       ABSTRACT
The purpose of this paper is to summarize the demonstration of Wet Oxidation of toxic
and hazardous wastes at a full scale installation.  This work is being done at Casmalia
Resources, a commercial waste treater in California.  The report will include data on
continuous operating units of a commercial nature.  It will also include testing on
actual wastewaters produced by industrial clients in the Southwest portion of the United
States.
I.  INTRODUCTION AND SUMMARY

Wet Air Oxidation is a process which has
been used to oxidize dissolved or sus-
pended organic substances at elevated
temperatures and pressures.  The process
is thermally self-sustaining with as low
as 15 g/1 COD organic feed concentrations
and is therefore most useful for wastes
which are too dilute to incinerate
economically yet too toxic to treat bio-
logical ly.

The process has been used to treat
various wastes the last thirty to forty
years.  With the recent attention being
focused on hazardous waste, much interest
has been expressed in the use of Wet Oxi-
dation as a means of destroying and/or
detoxifying these hazardous wastes.

During the recent years much bench
scale' >2 and pilot plant^ testing has
been performed by various companies to
demonstrate the applicability of Wet Oxi-
dation on various hazardous organic wastes.
Wet Oxidation units currently detoxify
specific waste streams at several waste
generation sites^.

The purpose of this project is to demon-
strate Wet Oxidation of toxic and hazard-
ous wastes at a full scale installation
which will  be located as Casmalia
Resources,  a commercial  waste treater in
California.  The project will enable
development of data on continuous operating
units of a  commercial nature.  It will  also
enable testing on actual wastewaters pro-
duced by industrial clients in the South-
west portion of the United States.

In the operation of the  full  scale wet oxi-
dation unit, aqueous wastes selected from
classified  groups of organic  wastes will be
treated.  These classified groups will  be
cyanide wastes, phenolic wastes,  sulfide
wastes, non-halogenated  pesticides, solvent
still bottoms, and general organic waste-
waters.  A  selected waste containing com-
pounds of a specific group will  be run.
Oxidation results will then be collated
such that predictions might be made for all
the compounds in a particular group.

Each waste  to be tested  will  be selected
from those  collected by  the commercial
treater or  those supplied by  the Environ-
mental Protection Agency or its supervising
contractor.  A preliminary autoclave Wet
Oxidation test will be run where necessary
to insure compatibility  and treatability
with the existing unit.   It is expected
that a continuous test of a minimum of
eight hours will be run  for each waste.
The unit will operate at 2.3  nP/hr (10 gpm)
                                           267

-------
with a waste having a COD of up to
46 grams/liter.   Waste with higher con-
centrations will  be run at lower flow
rates.  The unit will be operated at its
designed operating temperature of 280°C
unless there is  good cause to believe
that the waste would be processed at a
less severe condition.  Samples of feed
and effluent will  be analyzed to deter-
mine effectiveness of treatment.

This paper will  report on the status of
the project.

11.  Description of Wet Air Oxidation
     (WAO) Process

The Zimpro Wet Air Oxidation unit (See
Figure 1), for this demonstration will
process aqueous  wastes at a designed
reactor temperature or 280°C, a
designed reactor pressure of 136 atm,
a liquid waste flow rate of 2.3nvVhr
(10 gpm).  Waste will be mixed with
compressed air and directed through
the cold, heat-up side of the heat
exchanger.  The  incoming waste-air
mixture exits from the heat-up side
of the heat exchanger and enters the
reactor where exothermic reactions
increase the temperature of the mix-
ture to a desired value.  The waste-
air mixture exits the reactor and
enters the hot,  cool-down side of
the heat exchanger and, after passage
through the system pressure control
valves, is directed to the separator.
In the separator,  the spent process
vapors (non-condensible gases) are
separated from the oxidized liquid
phase and are directed into a two-
stage water scrubber-carbon bed
adsorber, vapor  treatment system.

In the Wet Oxidation process, organic
substances can be completely oxidized
to yield highly  oxygenated products
and water.  For  example, organic
carbon-hydrogen  compounds can be oxi-
dized to carbon  dioxide and water, while
reduced organic  sulfur compounds (sul-
fides, mercaptans, etc.) and organic
sulfides are oxidized to inorganic
sulfate, usually present in the oxi-
dized liquor as  sulfuric acid.  In-
organic cyanides and organic cyanides
(nitriles) are oxidized to carbon
dioxide, ammonia,  or molecular nitro-
gen.  It should  be noted that oxides
of nitrogen such as NO or N02 are not
formed in Wet Air Oxidation because the
reaction temperatures are not sufficiently
high to form them.

When incomplete oxidation of organic substan-
ces occurs, the reduced sulfur and cyanide
are usually still oxidized to sulfate and
carbon dioxide-ammonia provided a sufficient
degree of oxidation is accomplished.  How-
ever, incomplete oxidation or other organic
compounds results in the formation of low
molecular weight compounds such as acetal-
dehyde, acetone, and acetic acid.  These low
molecular weight compounds are volatile and
are distributed between the process off-gas
phase and the oxidized liquid phase.  The
concentration of these low molecular weight
compounds (measured as total hydrocarbons
(THC) expressed as methane) in the process
off-gas is dependent on their concentration
in the oxidized liquid phase, which is
determined by, the degree of oxidation
accomplished, the waste being oxidized, and
the influent organic concentration of the
waste.

Ill.  Description of Commercial  Test
      irte"

The Wet Oxidation unit to be used in this
demonstration work is located at Casmalia
Resources Inc. secure chemical waste land-
fill in Santa Barbara County, CA.  This
landfill has been operating since 1972. The
site is fully permitted as a hazardous
waste treatment, storage and disposal
facility pursuant to Section 25200 of the
Health and Safety Code of the State of
California Permit No. 42-001-78 has been
amended to allow operation of the Wet Air
Oxidation unit.

The Wet Air Oxidation unit has been pur-
chased by Casmalia Resources and will  be
operated under contract by Zimpro Inc.  It
is anticipated that the unit will be used to
treat those types or liquid wastes which
will be banned from landfill on January 1,
1983 by order of the State of California.

Post-treatment of WAO effluent will be by
evaporation ponds on the treatment site.
After evaporation, any remaining residue
(metals and salts) will be solidified and
landfilled.

IV.  Applications/Costs

Wet oxidation is best applied to dissolved
or suspended organic or other oxidizable
                                          268

-------
                WAO  ELEMENTARY  FLOW SHEET
               FEED HEAT
               EXCHANGER
                                               REACTOR
   HIGH
PRESSURE
  PUMP
                                                     -START-UP STEAM
                                                     TO SCRUBBER
                                                     SEPARATOR
              AIR
         COMPRESSOR
                  Figure I.  WAD Eleirentary Flew Sheet
                                269

-------
wastes.  The process is thermally self-
sustaining at relatively low concen-
trations (15,000 ppm COD).  Since
oxidation takes place in the liquid
state, it is not necessary to evaporate
the water.  The process therefore is most
useful for wastes which are too dilute to
incinerate economically yet too toxic to
treat biologically.

Figure II helps to illustrate the appli-
cable waste concentrations.  Wet oxida-
tion operating costs are lower because
of energy requirements.  The difference
between wet oxidation and conventional
incineration for toxic waste treatment,
for example, is significant and a
function of oxygen demand.

With incineration, it is necessary to
supply not only sensible heat and heat
of vaporization of liquid, but also
for heating vapors,  combustion pro-
ducts, spent air and excess air up to
a combustion temperature of between
800 and 1369°C.  With wet oxidation,
however, the only energy required is
the difference in enthalpy between
incoming and effluent streams.  This
value is typically 33,300 kcal/m3 for
a waste low in organics, as opposed
to 1.3 x 106   2.6 x 10° kcal/m3 for
incineration.   For a waste to become
autogenous (thermally self-sustaining)
in equipment of realistic size, a COD
of approximately 15,000 ppm is required
with wet oxidation;  for thermal oxida-
tion, approximately  400,000 ppm of COD
are required.5

V.  Wastes to  be Treated

Unit Permits

The Wet Air Oxidation unit to be used
for this demonstration has been per-
mitted by the  State  of California
Department of  Health Services and the
Santa Barbara  County Air Pollution
Control  District to  treat the follow-
ing aqueous waste streams.
1.
2.
3.
4.
5.
Cyanide wastes
Phenolic wastes
Sulfide wastes
Non-halogenated  pesticides
Solvent still  bottoms
                                           Wastes Specifically Exclude

                                           1.  The unit to be used for this test is
                                               not permitted to treat chlorinated
                                               aromatics such as hexachlorobenzene,
                                               chlorobenzene, dichlorobenzene or
                                               PCB's.  Lab testing has already been
                                               completed.^»2  if it is desired to
                                               test these wastes on a full scale
                                               basis, it is possible to run these
                                               at Zimpro's facility in Rothschild,
                                               WI.
                                           2.  Acrynonitrile wastes are currently
                                               being treated at a minimum of five
                                               full scale installations.   Therefore
                                               this wastewater will not be included
                                               as part of this work.

                                           Waste Selection

                                           Based on frequency of appearance on waste
                                           reports and quantities generated, the EPA
                                           has suggested the following compounds for
                                           this demonstration.

                                           Zimpro intends to perform tests on waste-
                                           streams containing the following com-
                                           pounds.  Six waste streams will be tested
                                           in a desired twenty-four hour  run for each
                                           waste.
                                              General  Class
                                           1.   Methane
                                           2.   Ethane
                                           3.   Aromatics
                                           4.   Others
6.  General  organic  wastewaters
                   Specific Compounds
                   ~a~.  Chloroform
                   b.  Carbon tetra-
                       chloride
                   c.  Methylene chloride
                   a.  Vinyl chloride
                   b.  Dichloroethane
                   c.  1,1,1-trichloro-
                       ethane
                   a.  Benzene
                   b.  Toluene
                   c.  Xylene
                   a.  Organic and inor-
                       ganic cyanides
                   b.  Organic and inor-
                       ganic sulfur com-
                       pounds
                   c.  Phenols
                   d.  Pesticides  (non-
                       halogenated aro-
                       matics).
Zimpro is aware of available waste streams
containing a mixture of several of the
above compounds.  We expect that demon-
strations using these mixtures with analy-
sis for the specific compounds listed in
the above group will be completed.
                                           270

-------
            THERMAL ENERGY REQUIREMENTS
                            VS
          	ORGANIC  CONTENT	

          THERMAL OXIDATION WET OXIDATION
OCL
UJ


I
UJ
i
I


I
UJ

UJ
or
UJ
cc
ct
UJ
z
UJ
            MAX DEFICIT 500 BTU/GAL
          ~ 2000 °F OXDATION
40,000
                                                   20,000
                                                   10,000
                                                 - 10,000
                                                 - 20,000
                                                 - 30,000
- 40,000
                       10     30     100

                    COD G/L OR 1000 MG/L
                                         300     1000
      Figure II.  Thermal Energy Requirements VS Organic Content
                         271

-------
VI.  Sampling and Testing Program

The progrm for sampling and testing the
various liquid and gas streams associa-
ted with the Wet Air Oxidation process
is divided into screening tests and
demonstration tests.  Screening tests
are used to determine the applicability
of Wet Air Oxidation in treating specific
classes of wastes.  Demonstration tests
are used to determine the effectiveness
of the Wet Air Oxidation unit while it is
actually processing a specific class of
waste.

A.  Screening of Potential  Wastes for
    Wet Air Oxidation

Prior to processing a given class of
waste in the Wet Air Oxidation unit,
representative samples of the waste
will be obtained.  A sample of this
waste, approximately two (2) liters,
will be shipped to Zimpro's labora-
tories.  Upon receipt, the  raw waste
will be analyzed for chemical  oxygen
demand (COD), biological  oxygen
demand (BOD), pH, total  solids, ash,
soluble chloride, soluble fluoride
and specific hazardous component,
e.g., cyanide, phenol, sulfide,
chlorinated aliphatic compounds,  or
non-halogenated pesticide.   After
preliminary analyses are completed,
the waste will be oxidized  in  a
laboratory shaking autoclave at the
operating temperature and residence
time of the Wet Air Oxidation  unit.
After oxidation the autoclave  will
be cooled and the non-condensible off-
gas will  be analyzed for oxygen,  nitro-
gen, carbon dioxide, total  hydrocar-
bons, and methane.

The oxidized waste will  in  turn be
analyzed for COD, BOD, total  solids,
ash, pH,  soluble chloride,  soluble
fluoride, dissolved organic carbon
(DOC), and the specific  component,
e.g., cyanide, phenol, sulfide,
chlorinated aliphatic compounds or
non-halogenated pesticides.  The
percent destruction of the  specific
component will  be calculated along
with the autoclave  oxygen demand.

Laboratory Wet Oxidation  and analysis
will also be used to determine materi-
als of construction compatibility  and
the potential  for scale  formation  in
 the  unit.  Wastes  which  demonstrate  high
 scaling  potential  or  materials  of con-
 struction  incompatibility will  be re-
 jected for treatment  in  the Wet Air
 Oxidation  unit.

 B.   Demonstration  Period for Processing
     Wastes in the  Wet Air Oxidation  P'ro-
     cess

 A  "demonstration period" for processing
 each selected wastes  will be conducted
 after a waste has  been judged acceptable
 for  processing in  the Wet Air Oxidation
 unit.  Each waste  will be tested  during
 a  one (1)  day "demonstration period," to
 determine  the effectiveness of  the Wet
 Air  Oxidation unit.   During each  "demon-
 stration period,"  the sampling  and test-
 ing  program outlined  as  follows will be
 used.

 Upon arrival  of a  truckload of  a  screened
 and  acceptable waste, a  sample  of the
 waste will be obtained and analyzed for
 COD, pH, and specific hazardous component.

 During the Wet Air Oxidation of each class
 of waste, liquid composites of  the influ-
 ent  raw waste and  the effluent  oxidized
 waste will be obtained.  Liquid samples
 of these two streams will be taken on an
 hourly interval  and the  liquid  samples
 will be made with  1 sample going for veri-
 fication analysis  at a second laboratory
 and the other going to Zimpro's labora-
 tory for analyses.  The  raw waste and
 oxidized waste will each be analyzed for
 COD, BOD, pH, total solids, ash,  soluble
 chloride, soluble  fluoride and  specific
 component, e.g., cyanide, phenol, sulfide,
 chlorinataed aliphatic compounds, or non-
 halogenated pesticide.   In addition, the
 oxidized waste will also be analyzed for
 dissolved organic carbon (DOC) and a
 GCMS scan will  be made for any daughter
compounds.

Grab samples of off-gas  from the Wet Air
Oxi.dation unit,  sampled  after carbon
 treatment but prior to discharge to the
atmosphere, will  be obtained in Tedlar
 gas sampling bags.

The sample of Wet Air Oxidation process
 off-gas will  be analyzed for oxygen,
 nitrogen, carbon dioxide, carbon monox-
 ide, total hydrocarbon, and methane.  Gas
sampling will  be conducted when the Wet
Air Oxidation unit is operating at
                                           272

-------
steady-state.  Two replicate grab
samples will be obtained each day that
the waste is processed in the Wet Air
Oxidation unit during the "demonstra-
tion period."  One sample will be ana-
lyzed by Zimpro personnel and the
other will be available for verifica-
tion analysis by a second lab.

In addition to the above gas sampling
program, the Wet Air Oxidation unit
will be equipped with an on-line con-
tinuous total hydrocarbon analyzer
for analysis of the process off-gas
at a point before the carbon bed and
also at a point before discharge to
the atmosphere.  The results of the
continuous total hydrocarbon analysis
will become part of the Wet Air Oxi-
dation process operating record.

Analytical Procedures

The following analytical procedures,
are used by Zimpro's laboratories.

 1.  Chemical Oxygen Demand
     a.  Macro method
     b.  Macro method
 2.  Dissolved Organic Carbon
 3.  Total Solids and Ash
 4.  Soluble Chloride
 5.  Soluble Fluoride
 6.  Cyanide
 7.  Phenols
 8.  Total Sulfides
 9.  Chlorinated Aliphatic Compounds
10.  Non-Halogenated Pesticides
11.  Gas Analysis
     a.  Oxygen
     b.  Nitrogen
     c.  Carbon dioxide
     d.  Carbon monoxide
     e.  Total hydrocarbons
     f.  Methane

VII.  Status

The operating permits have been issued
as of October 1, 1982.  It is expected
that installation to be completed by
late November and that waste processing
will begin late 1982.

Bibliography

1.  Randall, T.L., Knopp, P.V., "Detoxifi-
    cation of Specific Organic Substances
    Wet Oxidation," 51st Annual Confer-
    ence; WPCF, Anaheim, October 1978.
Randall, T.L., "Wet Oxidation of
Toxic and Hazardous Comounds," Mid-
Atlantic Industrial Waste Disposal
Conference, University of Delaware,
1981.

Chowdhury,  A.K.,  Wilhelmi,  A.R.,
"Treatment  of Spent Caustic Liquors
by Wet Oxidation," 8th Annual
Industrial  Pollution Conference,
June 1980,  Houston.

DeAngelo, D.J., Wilhelmi,  A.R.,  "Wet
Air Oxidation of Northern  Petro-
chemical Company  Spent Caustic
Liquors," 1982 Spring National  AIChE
Meeting, Anaheim, California,  1982.

Knopp, P.V. & Wilhelmi,  A.R.,  "Wet
Oxidation - An Alternative  to
Incineration," Chemical  Engineering
Progress, Vol. 75, No.  8,  Page 46-52,
August 1979.
                                           273

-------
                   ENGINEERING GENES IN YEAST FOR BIODEGRADATIONS
                     John C. Loper£+, Jerry B. Lingrele and Vernon F. Kalbe
                      eDepartment of Microbiology and Molecular Genetics,
                             +Department of Environmental Health,
                          University of Cincinnati College of Medicine
                                     Cincinnati, OH 45267
                                          ABSTRACT

     In order  to obtain a yeast gene sequence which would allow the isolation of different P-450
genes for use in  specific biodegradations, we have constructed a Candida tropical is  750 genomic
library. This genomic  library  contains C_. tropicalis gene sequences stored as 5 to 10  kilobase pair
lengths of DNA inserted into the shuttle vector pABI07. The library contains DMA sequences which
will complement mutations in a Saccharomyces cerevisiae strain deficient in the biosynthesis of both
histidine  and adenine;  thus at  least those C_.  tropicalis genes can be functionally expressed  in S_.
cerevisiae. One C. tropicalis cytochrome P-450 sequence codes for an enzyme which catalyzes the
 -hydroxylation of n-alkanes.  This hydroxylation is the first step in a catabolic pathway which allows
the organism to grow on tetradecane as sole carbon and energy source.  Mutants of a closely related
yeast exist which are blocked in different steps of n-alkane  utilization, including some which may be
deficient  in  this  hydroxylase.   Consequently,  we  plan to  screen the library  for  this  gene  by
transforming these mutants with the library  DNA and  looking for  clones which  have  acquired the
ability  to use tetradecane as a carbon and energy source. In separate experiments we have indicated
that S. cerevisiae  D7  is heterozygous for the amount of P-450 since  haploid  cultures form 2:2
segregation ratios of high and low levels of P-450.
INTRODUCTION

     A group of  chemicals which cause con-
cern as risks to human health are the so-called
recalcitrant compounds (I).  These compounds
are usually  hydrophobic and are resistant  to
physical  or biological  decomposition  in the
environment.  Because  of their  chemical sta-
bility they are  available for eventual redistri-
bution  in ways  which bring them into  contact
with man.  This can happen through contami-
nation  of drinking  water sources or  through
bioaccumulation in  the  food chain.  By either
route,  the  hydrophobic nature of  many  of
these xenobiotic compounds leads to their in-
corporation into the fatty tissue  of the  body.

     If  these  recalcitrant compounds  were
entirely   inert,   and   thus   refractory   to
mammalian  metabolic  conversion, they  might
present relatively minor  problems.  However,
many act as signals  to alter  gene expression
and  cause profound  changes in the type  and
quantity  of  metabolic enzymes in  the  cells of
many organs and tissues.  The major group of
"induced" enzymes are complex oxidative  en-
zymes known as the  cytochrome P-450 mono-
oxygenases.  These  enzymes,  present  in  the
tissues  either   at  background  levels  or  at
elevated levels  following induction, can  result
in the metabolism of  recalcitrant compounds
to cytotoxins, direct  acting mutagens, carcin-
ogens, or teratogens.  However, the  different
enzymes  which  are  induced  in this  manner
include activities which detoxicate xenobiotic
chemicals as well as  those  which increase the
                                              274

-------
toxic effects of  such compounds.   In  such
cases  it  can be  stated  that  the  enzymes
necessary for metabolism of environmentally
recalcitrant  compounds  do exist,  but  in  the
wrong places.

     By combining  recent advances  in  gene
engineering, microbiology and biochemistry, it
should  be possible to  develop microorganisms
which  could  detoxicate  a  recalcitrant com-
pound  in  the  environment before  it came in
contact with man.  Any  organism  possessing
the desired activity  could be used as a source
of monooxygenase genes to be  manipulated.
For the compounds where biotransformation is
understood, the best  source could be the genes
of those organisms.  Environmental biodegrad-
ations  are primarily effected by  mixtures of
bacteria,  yeasts, molds and other fungi.  Of
these,  certain species of bacteria and yeasts
grow rapidly as  single  cells and are  well
characterized both  with  respect  to genetics
and  to  the  relevant  recombinant DMA tech-
niques,  and  so  are  best  suited  for  genetic
manipulation.   Both  possess monooxygenase
systems which include  cytochrome P-450 hem-
oproteins.   However, the  systems  found in
yeasts  are  quite  similar  to those  found in
hepatic tissues and  in other eucaryotic cells,
whereas  those  in bacteria  are   considerably
different (18).

     Based  upon these relationships we  have
initiated  the genetic  engineering of  cyto-
chrome P-450 expression  in  yeast  for the de-
gradation of hazardous organic wastes.   Our
first experiments are intended   to  obtain  a
probe for cloned P-450 sequences and to gain a
practical  understanding of the P-450 systems
which operate normally in the yeast S.  cere-
visiae.   In this  paper we report  on progress
made toward these objectives.

MATERIALS AND METHODS

Strains

     The Saccharomyces cerevisiae  haploid
strain  BWG  2-9A  (MATa  adel-100  his4-4!9
ura3-52) was  a  gift  from L.  Guarente.  A
recA+  variant of  E.  coli JW355  (F~  araDI39
 (ara, leu) 7697,   lacX74  galK galU recA rpsL
hsdR)  was a gift from  J.    Williams.    C_.
tropicalis 750 was obtained from the American
Type Culture Collection, Rockville, MD. Hap-
loids of S.  cerevisiae D7  were generated from
diploid stocks provided by E. Nestmann and by
F. Eckardt.
Enzymes

     Restriction  enzymes  were  from   New
England BioLabs.  T4 DMA ligase and bacterial
alkaline phosphatase  were from Bethesda Re-
search Laboratories.   Enzymes were used  as
recommended by the suppliers.

Preparation of DMA

     High  molecular  weight yeast DMA was
prepared by the method of Cryer  et  al.  (6).
Covalently  closed circular plasmid DMA was
isolated  from  E_.  coli  using  the  method  of
Holmes and  Quigley  (10).   In  some cases this
plasmid DMA was purified  by CsCI/ethidium
bromide banding (5).  Plasmid DMA from  yeast
was isolated using the method  of Nasmyth and
Reed  (15).   Partial digestions with the restric-
tion  enzyme Sau3A  I  were  performed as  de-
scribed (12) to  ensure maximum sequence rep-
resentation  of  restriction  fragments in  the  5
to 10 kilobase pair (kbp) length range.

Transformation Procedures

     Yeast   transformations  were  performed
using  the  procedure of  Beggs  (3) as modified
by Sherman et  al. (19).  Bacterial transforma-
tions  were performed using  a  method devised
by J.   Williams (personal communication).  One
ml of an overnight culture of the E.. coli  in  L
broth (10 g/l Bacto-tryptone, 5g/l Bacto-yeast
extract, 5  g/l NaCI)  is  used to  innoculate  50
ml of  L broth.   The culture  is  incubated  at
37°C  with shaking until  the  absorbance at 650
nm reaches  0.2. The culture is then chilled  on
ice for 10 min,  collected by centrifugation  (5
min, 3000 x g), resuspended in 4 ml sterile cold
O.I M  CaCl2, and stored 25 min  on wet ice.
The cells  are then collected by centrifugation
(5 min, 3000 x  g)  and resuspended  very  gently
(2 to 4 hr on rotating platform in  a cold room)
in 0.4 ml cold  O.IM  CaCl2.  A cold pipet  is
used to add an aliquot  of  the  cell  suspension
(0.1 ml)  to  0.02  ml  of  the  DMA  which   is
dissolved  in  lOmM Tris-CI,  O.ImM EDTA; pH
8.0.   The mixture  is incubated for 15  min  at
0°C, 2 min  at  45°C,  and then placed at 37°C.
Immediately, 1.0  ml  of  pre-warmed  37°C  L
broth is added  and the mixture incubated with
gentle shaking  for I hr.   Aliquots (O.I ml) are
then  plated  on L agar plates  supplemented
with ampicillin (25 yg/ml) and incubated  over-
night at 37°C.

Gel Electrophoresis

     Gel  electrophoresis  of  DMA  was  per-
formed as described  by  Mickel et al.    (13).
                                              275

-------
The  isolation  of  DMA  fragments  from  low
melting agarose gels followed the procedure of
Langridge et al. (II).

Vector

     The shuttle plasmid pABI07 was obtained
from L. Melnick.  It contains  pBR322 sequen-
ces which express tetracycline  and ampicillin
resistance markers and allow  replication in E.
coli.  Also present is  the S_. cerevisiae URA3
gene for use as a selectable  marker, plus an
ARSI sequence (21) which allows the vector to
replicate in yeast.

Media

     Yeast  were ordinarily grown  either on a
complex medium for routine growth (YPDG) or
a synthetic medium (SD) with the  appropriate
supplements (19).

Cytochrome Determinations

     (I. tropical is was grown in  YEPD medium
containing 20%  glucose  (22)   and  also  in
medium   containing  0.5%  tetradecane   (8).
Haploids of S. cerevisiae were grown in YEPD
medium containing 20%  glucose.   Cells were
collected  by adding an  equal volume of  ice-
cold phosphate buffered  saline (PBS) to  cul-
tures   in  the  midlog  growth  phase.   After
centrifugation (5 min, 4000 x g), the cell pellet
was washed once with PBS and  resuspended in
O.IM potassium phosphate buffer (pH 7.4) at a
concentration  of  O.I g wet weight of cells/ml.
The cytochrome P-450 and  cytochrome P-420
contents of whole cell suspensions were deter-
mined  with a Perkin-Elmer model 575 spectro-
photometer  using the method  of  Omura  and
Sato (17). Cytochrome contents (nmole/g  wet
weight of whole  cells) were  calculated using
molar  extinction  coefficients of 91  crrr'mM"'
(17) and 110 cm-'m/vH (20)  for cytochrome P-
450 and cytochrome P-420, respectively; cor-
rections  to the cytochrome P-420  values  due
to the  absorption  of  cytochromes P-450 at
420nm  were taken according to the procedure
of Guengerich (9).

RESULTS

     Our initial  experiments  have  been  di-
rected towards  two  goals.    The  first  is to
obtain  any yeast P-450  gene sequence which
may then  be  used as a probe  to isolate other
P-450  genes,  as  was   done  by   Hall   and
colleagues for the  yeast  iso-cytochrome C
genes  (14).   The  second goal   is to begin to
understand the  mechanism(s)   by  which  the
level  of  P-450 gene expression in  S. cerevisiae
is modulated so that the expression of recom-
binant genes may be controlled.

      One approach to obtaining the initial P-
450 gene was to construct a genomic library of
Candida   tropical is  750.   This  single  celled
yeast and a closely related organism Saccharo-
mycopsis lipolytica can  grow  on   tetradecane
as sole carbon and energy source.  The initial
step in this catabolic  pathway  is  catalyzed by
a  highly  inducible cytochrome P-450  enzyme,
n-alkane hyroxylase.   We  have confirmed  that
the cytochrome P-450 content in C_.  tropical is
750  increases  when  tetradecane  rather  than
glucose is used  as a carbon source (Fig. I). S_.
cerevisiae does not express this   pathway and
can not  grow on n-alkanes.  We chose initially
to prepare a genomic library for  C.  tropicalis
and to test  the  transfer of genetic information
from  this library to S. cerevisiae.

Construction of  the C. tropicalis 750 library

      The library  was made by inserting  ran-
dom restriction  fragments of C. tropicalis 750
DMA  into the plasmid pABI07 Isee  Fig. 2).  The
resultant recombinant  plasmid  mixture  was
used  to  transform  E.  coli to yield  a  large
number of different  E_.  coli clones, each con-
taining a recombinant plasmid with an inserted
fragment of C_. tropicalis DMA.   The number
of different clones needed  to assure a greater
than 99% chance  of having a given segment of
the C.  tropicalis genome is  less than  12,800
(4).

      High molecular  weight DMA  of C.  tropi-
calis  (100 kbp) was partially digested with the
restriction enzyme SauSA  I so that the number
average  molecular weight of the digested DMA
peaked in the region corresponding  to 5 to 10

kbp fragments.  This DMA  was then separated
by  electrophoresis on  a   0.8%  low-melting
agarose  flat bed  gel.   The region of the gel
corresponding to restriction fragments of  5 to
10 kbp was excised and the DNA isolated.

      DMA of the vector  pABI07, after com-
plete   digestion  with  BamH  I  and  treatment
with  bacterial  alkaline  phosphatase, was iso-
lated  in  the same manner as the  C.  tropicalis
DNA.                             ~

      The two DMAs were  then ligated  for 24
hr  at I4.5°C in  a 0.5 ml  reaction  mixture
containing 40 ug pABI07 DNA, 5  yg C.  tropi-
calis  DNA,  and  2.5 units of T4 DNA~ligase  in
                                              276

-------
         0.02
                         CO-Reduced
                             Reduced
       n
       .0
         0.02
         0.02
                         CO-Reduced
                             Reduced
           400
                       450

                       nm
                                  500
Figure  I.  Absorption spectra of C.  tropicalis
750.  Cells in log phase of growth were  re-
suspended  at  O.lg wet weight whole cells/ml,
dithionite-reduced vs.  dithionite-reduced plus
CO.  A: grown on 20% glucose; B:  grown on
0.5% tetradecane.
                                                                 EcoR I
                                                                                  HIND  III
                                   BAMH  I

                                   SAL  I
Figure 2.  The circular shuttle plasmid pABI07,
(obtained   from   L.   Melnick,   Univ.     of
Rochester).   The heavy lines correspond to the
locations  of  yeast sequences   which  allow
autonomous replication in !5. cerevisiae (ARSI)
and  complementation of a  ura3  mutation in
yeast  genomic DMA  (URA3T.The  lighter
portions  of  the circle correspond  to  pBR322
DMA sequences and contain a  DMA replication
origin which  allows the plasmid to replicate in
F_.  coli.  The  closed boxes mark the locations
of regions in the  plasmid which  code for re-
sistance  to  ampicillin  (Amp) and tetracycline
(Tet).  Foreign DMA  sequences inserted at the
BamHI  site   inactivate  the  tetracycline  re-
sistance gene.  Consequently, the phenotype of
plasmids  containing  inserted  DMA  sequences
at the BamH I  site  is Amp1" -Tets.  The plasmid
is 6.3 kbp  in length.
the buffer recommended by the supplier.  Be-
cause the pABI07 DMA had been  treated with
alkaline  phosphatase,  it  could  not self-ligate
and generate  transformants which contain no
C. tropicalis DMA sequences.

     Approximately 1/3 of this ligation mix-
ture was  used  to  transform E_.  coli which was
then plated to  yield 29,000 colonies on L-agar
plates supplemented with 25 yg/ml ampicillin.
The colonies were rinsed from the  transforma-
tion plates with L broth and collected by
centrifugation.   Part  of this  cell  population
was  resuspended  at 1.4 x  10'' cells/ml in L
broth containing  8%  dimethyl  sulfoxide  and
stored  at -80°C.  The  remainder was used to
isolate plasmid DMA, which  was  purified by
banding in CsCI, and then stored at 5°C in 10
rnM Tris-CI, O.I mM EDTA at pH 8.

     Four   random  clones  from  the  library
were grown in liquid culture.   Their plasmids
were extracted, cleaved  with  EcoR  I, and the
DMA  fragments  were  separated  by  electro-
phoresis  on a  0.8% agarose  gel.   All  4 con-
tained  inserted DMA sequences (Fig.  3).  Based
on  examination  of  an  additional  20  clones,
approximately  70%  of  the clones contain  in-
serted  DMA sequences.
Isolation of Clones  Which Complement  the
adel and  his4 Mutations

     S.  cerevisiae BWG 2-9A is auxotrophic
for histidine,  uracil, and adenine because of
mutations in its ura3, his4,  and adel genes. We
transformed this yeast strain  with  the library
                                               277

-------
Figure 3.  EcoR  I digests of random clones
from  the C.  tropical is  library.   Lanes  1-4
are plasmid   preparations   of   4  randomly
selected  clones digested with EcoR I.  Lane 5
is  pABIO?  cut with  BamH  I.   Lane 6  is  a
mixture of DNA digests:  charm 4A (digested
with BamH  I); pBR322 (digested  with EcoR I
and Rsa  I).    The  eight  strongest bands from
top to bottom are 49, 23.3, 10.4, 7.0, 4.4, 3.9,
2.1   and  1.6   kbp.      The   region    of
intense fluorescence  at the bottom of the gel
results from RNA  which is  not  removed from
the  plasmid  preparations   unless  they   are
banded in CsCI/ethidium bromide.
DNA  and  selected for  yeast  colonies  which
grew  in  a minimal medium  unsupplemented
with uracil. Only those colonies (approximate-
ly 27,000 from 3.4 \ig total library DNA) which
had received  a functional URA3 gene from the
vector pABI07 were able to grow.  It should be
emphasized that the vector itself contains the
functional  URA3  gene  and  that this  initial
transformation  simply selects  for the vector
pABI07 and  not  for a specific  C.  tropical is
insert.  We then added 30 ml  of sterile water
to the top agar  containing  the yeast  trans-
formants  from  the transformation plate,  and
ground in a Waring blender.  This material  was
washed by 3 centrifugations and resuspensions
in sterile water.
     In order to select clones containing the
C.  tropicalis  DNA  sequences  which   would
complement  the his4 and ade I mutations in S.
cerevisiqe,  aliquots  of  th~eWashed  material
were  then  plated  on minimal plates lacking
either  adenine or L-histidine.  We obtained 30
yeast colonies  that no longer required adenine
for growth  and 875 colonies that no   longer
required histidine for  growth.

     In order  to  show  that  these transform-
ants had  lost  their  auxotrophic  requirements
because they had received a recombinant plas-
mid  which  provided the  necessary  gene se-
quences, we recycled the  plasmids through E.
coli  and  then  back into  untransformed J).
cerevisiae BWG 2-9A.

     Ten  clones from each of the adenine and
histidine transformant collections  were  grown
separately  in liquid culture and their plasmid
DNAs  extracted.   The  DMAs from  the  yeast
adenine transformants and  the histidine  trans-
formants were pooled separately.   These  two
pools of DNA were then used to transform E^.
coli.    The   E.  coli  transformants  were  not
plated  but  allowed  to  grow overnight   in  L
broth supplemented with 25  ug/ml ampicillin.
The  plasmid DNA  from each E.  coli culture
was  then  extracted and used to  transform S.
cerevisiae BWG 2-9A.  Plasmid DNA originat-
ing from the yeast  adenine transformants gave
rise  to adenine transformants in yeast  after
passage through  E. coli.   The  plasmid   DNA
originating  from  the  yeast  histidine   trans-
formants also behaved in the expected  way.

     The iE. coli plasmid  preparations used in
the secondary yeast  transformations  described
immediately above  were  also  analyzed  by
electrophoresis on  a 0.8% agarose  gel  after
restriction   with EcoR  I,  see Fig. 4.   These
plasmid  preparations originating   from   either
the yeast adenine transformants (lane 3)  or the
histidine transformants  (lane 4) appear to con-
tain  more   than  one type of plasmid.   This
could arise  from  including variable portions of
the yeast genome  flanking  the  adel  and  his4
genes.

Prospects for Detection of Clones Which Code
for Expression of n-alkane Hydroxylase.

     Based   upon  these  results,  this   gene
library   should  include   clones  containing
inserted DNA  sequences  for  each  of  the
several  genes   involved   in  the   n-alkane
oxidative pathway, including  the  gene for the
cytochrome    P-450     enzyme     n-alkane
hydroxylase.   The  most  direct  approach to
                                              278

-------
Figure  4.    EcoR  I  digests  of  recombinant
plasmids which  complement the  his4 or adel
mutations in yeast.  Lane I  is an EcoR I  digest
of  total  library DMA in  pABI07.   Lane  2
contains UNA size  markers as in Fig. 3. Two
smaller bands (680  and 506 bp)  can be seen  in
this run.  Lanes 3 and 4 contain EcoR I  digests
of plasmids which  complement  the his4 (lane
3) or adel mutations in S. cerevisiae.
obtaining this  P-450 gene  is  to  transform a
recipient yeast strain lacking the function
with   the   library   of   cloned  sequences.
Recipient yeast  cells that gained  the P-450
gene  could  be easily  recognized if  the gene
conferred   the  capacity   for  growth   on
tetradecane.  However several genes  are  in-
volved in the n-alkane utilization pathway, and
it  is not  known how  many of these are  missing
in  S.  cerevisiae.   We  have  conducted  two
experiments  in   which   more  than  forty
thousand  cells of  strain  BWG2-9A received
library clones, as indicated by  expression of
URA3 on the cloning vector, but none of these
recipients grew on  tetradecane.  Fortunately
mutants   in  another  yeast  exist which  are
blocked in different steps  of n-alkane  utiliza-
tion.     These  are  the  alk   mutants  of
Saccharomycopsis lipolytica (2), a sporulating
variant of  a Candida species similar to C.
tropicalis (16).   The  phenotype of some of
these  mutants  suggests  that  they  may  be
specifically    deficient    in    the    P-450
hydroxylation  reaction.   In  collaboration with
Dr. J.   Bassel  we plan to  screen our  gene
library  by  transforming this  sub-set  of alk
mutants  with  the  library DMA  and  looking for
clones  which  have  acquired  the  ability  to
utilize tetradecanes.

Cytochrome P-450 Levels in S. cerevisiae

      We have begun an examination  of P-450
inducibility in S. cerevisiae since strains which
are differentially  inducible  for  P-450  will  be
useful as recipients in future gene  engineering
experiments.   Diploids  of  S_.   cerevisiae D7
were  sporulated, and asci were dissected  using
a micromanipulator.  Haploid cells from  each
ascospore were  grown  to log  phase  in YEPD
medium  containing  20%  glucose,  conditions
which had been  previously  shown  to  increase
the levels of  P-450  in  S.    cerevisiae  (22).
Whole cell  suspensions  of  haploid  segregants
were  used for spectral  determinations of both
cytochrome P-450 and  cytochrome P-420,  an
inactive  form of  P-450.  The results of  these
measurements, shown in Table  I, indicate that
S.  cerevisiae D7  is heterozygous for  levels of
P-450,  with  ascospores generally  segregating
2:2 (high P-450:low P-450).   Interpretation of
these ratios is complicated by the 2:2 segrega-
tions  of high  P-420:low P-420, in  that the
levels of P-420  were  not  always  in  reverse
relationship to the levels of P-450.

DISCUSSION

      We   are   using   a   gene   engineering
approach to  achieve the biodegradation  of  a
broad range  of  recalcitrant  compounds of en-
vironmental concern.   The  focus  is  on  com-
pounds which  cause their adverse effects  after
they  are taken up by cells and  become metab-
olized to more  reactive forms  — such metab-
olites eventually  effect cell  damage  or are
detoxicated by  the cell.  The  rationale  is to
modify yeasts to  do this metabolic  activation-
detoxication.    This genetic  capacity  would
then  be used  in  these yeasts;  or would  be
transferred  for  use in  related  microorganisms
occurring  in  the  environment close  to the
origin or location of the problem compound.

      We  have  shown  that  genes  from  one
genus of yeast  can be  transferred  into and
expressed in another genus of yeast  in  the case
of C_. tropicalis and S_. cerevisiae. Dickson (7)
has shown  that the  Kluyveromyces  lactis  3-
galactosidase  gene  can  be  transferred  into
and expressed in  S.   cerevisiae.   It  appears
that  the    signals    necessary   for   the
transcription and  translation of  genetic
                                               279

-------
 TABLE I.  CYTOCHROME P-450 AND P-420
        CONTENT OF S. CEREVISIAE
           STRAIN D7 HAPL01DS
Haploid
segregants£
3a
3b
3c
3d
5a
5b
5c
5d
7a
7b
7c
7d
Cytochrome
P-450
nmoles/g+
0.80
0.46
0.34
0.69
0.34
0.91
0.23
0.57
1.32
0.44
0.36
0.99
Cytochrome
P-420
nmoles/g
0.30
0.17
2.66
2.34
2.02
.34
2.83
0.21
0.49
1.98
1.32
0.64
 eHaploids  from three  asci  of  S.  cerevisioe
 were grown in YEPD medium containing 20%
 glucose.

 +nmoles/g wet weight of whole cells.
information are similar  enough between the
two  genera to  allow the isolation  of a cloned
P-450  gene from  C.   tropicalis followed by
analysis of  its expression  in S.  cerevisiae.
This cloned gene  may  then be  used  as  a
hybridization probe  to  isolate other  P-450
genes that code for enzymes involved in the
biodegradation  of recalcitrant compounds.

     The    level   of   cytochrome    P-450
expression  from the  newly  introduced  gene
must be  compatible with  the levels  of the
other components of the P-450 system such as
NADPH and   NADPH   cytochrome    P-450
reductase.   Expression   of   a  cloned  P-450
gene in yeast  could be controlled  by  using
appropriate promoter   sequences  and   other
DNA signals for expression.   This expression
can also be influenced  by introducing the gene
on a plasmid which replicates at a high  copy
number  in  yeast,  or  on a   plasmid  such as
YCpl9  which is fairly  stably  maintained at
about one copy per nucleus,  or by integrating
 the P-450  gene   into  a  site  in  a  normal
 chromosome or   chromosome    pair.      The
 segregation  of high and  low  levels  of P-450
 which  we  have  observed in  haploids  of  S.
 cerevisiae  may  also  be a   useful   tool   Tn
 controlling   the  expression  of P-450  in  our
 genetically engineered yeast.

      Our goal  is  the production of  a set  of
 yeasts each of  which  is capable  of a  safe,
 high-level biooxidation  of a  different hazard-
 ous compound or  compounds.    These  yeast
 strains  alone or  together could be tested  for
 the bioconversion  of  hazardous byproducts in
 industrial  processes  and  for  the  elimination
 of  such  problems  in  spills  or  dumps.   Since
 different species of  yeast or fungi are likely
 to  better tolerate  the  ecological conditions
 presented by  various   noxious   environments,
 these strains could be used for secondary gene
 engineering among naturally  occurring  eucar-
 yotic microorganisms,  in order  to establish
 the preferred detoxicating variants.

 REFERENCES

 I.     Alexander,  M.   1981.  Biodegradation of
      chemicals  of   environmental   concern.
      Science, 11:132-138.
 2.     Bassel, J.B.,  and  R.K.  Mortimer.  1982.
      Genetic and  biochemical  studies of  n-
      alkane non-utilizing  mutants of Saccha-
      romycopsis  lipolytica.    Curr.   Genet.,
      5:77-88.
 3.    Beggs,  J.D.   1978.   Transformation  of
      yeast  by a replicating  hybrid  plasmid.
      Nature, 275:104-109.
 4.     Clarke,  L.,  and  J.  Carbon.   1976.    A
      colony  bank  containing  synthetic  ColEI
      hybrid plasmids representative of the en-
      tire E. coli genome. Cell,  9:91-99.
 5.     Clewell,  D.B.    1972.   Nature of  ColEI
      plasmid replication in Escherichia  coli in
      the presence of chloramphenicol. J. Bac-
      teriol., 110:667-676.
 6.     Cryer,   D.R.,   R.   Eccleshall,  and  J.
      Marmur.  1975.   Isolation  of yeast  DNA.
      Methods Cell. Biol., 12:39-44.
 7.     Dickson,  R.C.   1980.   Expression of  a
      foreign eukaryotic gene in  Saccharomy-
      ces  cerevisiae:    3-galactosidase  from
      Kluyveromyces lactis. Gene, 10:347-356.
 8.     Duppel,  W.,  J.-M.  Lebeault, and  MJ.
      Coon.    1973.    Properties  of  a  yeast
      cytochrome   P-450-containing  enzyme
      system which  catalyzes  the  hydroxyla-
      tion of fatty acids, alkanes,  and drugs.
      Eur. J.  Biochem., 36:583-592.
9.     Guengerich, F.P.   1982,  Microsomal  en-
      zymes  involved   in  toxicology-analysis
                                             280

-------
     and  separation.    In:    Principles  and
     Methods  of  Toxicology.    A.   Wallace
     Hayes, ed.  Raven Press, New York.  pp.
     609-633.
 10.  Holmes,  D.S., and M. Quigley.   1981.  A
     rapid boiling method for  the preparation
     of bacterial plasmids.   Anal. Biochem.,
     114:193-197.
II.   Langridge,  J., P.  Langridge, and  P.L.
     Bergquist.  1980.  Extraction of nucleic
     acids  from  agarose gels.    Anal.  Bio-
     chem., 103:264-271.
 12.  Maniatis,   T.,  E.F.  Fritsch,   and  J.
     Sambrook.   1982.    Molecular  Cloning.
     Cold Spring  Harbor  Laboratory,   Cold
     Spring Harbor, New York. 545 pp.
13.  Mickel, S.,  V. Arena, Jr., and W. Bauer.
     1977.  Physical properties and gel  elec-
     trophoresis   behavior   of   RI2-derived
     DNAs.    Nuc. Acids  Research, 4:1465-
     1482.
14.  Montgomery,  D.L.,  D.W.   Leung,  M.
     Smith, P. Shalit, G. Faye, and B.D.  Hall.
     1980.  Isolation and sequence of the gene
     for  iso-2-cytochrome c in  Saccharomy-
     ces  cerevisiae.  Proc.  NatLAcad.  Sci.
     USA, 77:541-545.
 15.  Nasmyth, K.A.,  and  S.I. Reed.  1980.
     Isolation of  genes by  complementation in
     yeast: molecular cloning of a cell-cycle
     gene.    Proc. Natl.  Acad.  Sci.  USA,
     77:2119-2123.
 16.  Ogrydziak,    D.,    J.    Bassel,    R.
     Contopoulou,  and R. Mortimer.   1978.
     Development  of  genetic techniques  and
     the  genetic map  of  the yeast Saccharo-
     mycopsis lipolytica.   Molec.  gen.  Ge-
     net., 163:229-239.
17.  Omura,  T., and  R.  Sato.   1964.    The
     carbon   monoxide-binding  pigment  of
     liver  microsomes.    J.  Biol.  Chem.,
     239:2370-2378.
 18. . Rosazza, J.P., and   R.V. Smith.  1979.
     Microbial models for drug metabolism.
     Adv. Appl.  Microbiol., 25:169-208.
 19.  Sherman, F., G.R. Fink, and J.B.  Hicks.
     1981.  Methods in Yeast Genetics.   Cold
     Spring Harbor Laboratory, Cold Spring
     Harbor, New York.  119 pp.
20.  Shoeman, D.W.,  F.M.  Vane,  and  G.J.
     Mannering.    1972.   Induction  of   drug
     metabolism  VII:   differences in P-420
     hemoproteins  from   untreated   and 3-
     methylcholanthrene-treated  rats.   Mo-
     lec.  Pharmacol., 9:372-382.
21.  Stinchcomb, D.T., K. Struhl, and  R.W.
     Davis.   1979.   Isolation  and characteri-
     zation    of   a   yeast   chromosomal
     replicator.  Nature, 282:39-43.
22.   Wiseman,   A.,   T.-K.  Limj,   and  C.
     McCloud.   1975.   Relationship  of  cyto-
     chrome   P-450  to   growth  phase  of
     brewer's  yeast  in  l%- or  20%-glucose
     medium.  Biochem. Soc.  Trans.,  3:276-
     278.
                                             281

-------
           9th ANNUAL SYMPOSIUM ATTENDEES LIST
Ackerman, Donald G. Jr. - PhD.
TRW Inc.
Energy & Environmental Div.
23900 Hawthorne
Suite 200
Torrance, CA 90505

Adams, William R., Jr. - Mgr.
NUS Corp.
Park West Two
Cliff Mine Road
Pittsburg, PA 15275

Adaska, Wayne
Senior Engineer
Portlant Cement Assocn.
Skokie, IL 60077

Ainsworth, Brian
Schlegel Lining Tech., Inc.
200 S. Trade Center Parkway
P.O. Box 7730
The Woodlands, TX 77380

Aittola,  Jussi-Pekka - Mgr.
Stusvik Energiteknik AB
611 82 NYKOPING,  Sweden

Akers, Karol - Engr.
VA Dept. of Health
Division of Solid
  & Hazardous Waste Mgmt.
Richmond, VA 23219

Alanddin, Mohammad
KY Division of Waste Mgmt.
Ft.  Boone Plaza
ISReillyRoad
Frankfort, KY 40601

Aldridge, Wayne C.
Post Buckley Schuh & Jerrigan Inc.
889 N. Orange Ave.
Orlando, FL 32801
Allen, James L.
E.I. du Pont de Nemours & Co.
3500 Grays Ferry Avenue
Philadelphia, PA 19146

Allen, Richard
United Catalysts Inc.
P.O. Box 32370
Louisville, KY 40232

Alterman, Wayne
Certified Consultants
23 Dellforest Ct.
The Woodlands, TX 77380

Alther, George R.
International Minerals
  & Chern. Corp.
17350 Ryan
Detroit, MI 48212

Arnmon, Douglas C.
Hydrologist
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Amrnons, James T.
US Army Corps of Engineers
P.O. Box 1600
ATTN: ED-CS
Huntsville, AL 35807

Anderson, David C.
K. W.  Brown & Assoc.
707 Texas Avenue South
Suite 202D
College Station, TX 77840

Andrews, Douglas - Pres.
Andrews Engineers, Inc.
1320 South Fifth
Springfield, IL 62703
                                   282

-------
Andruacola, Evans
Sales Manager
Trane Thermal
Brook Road
Conshohocken, PA 19428

Appleton, H.
Hyton Engineering Co.
282 Maitland Avenue
Teaneck, NJ 07666

Arden, Mildred B.
Dept. of Environ. Protection
ISReillyRoad
Frankfort, KY 40601

Ardiente, Editha M.
Chemical Engineer
US EPA Region V
230 S. Dearborn St.
Chicago, IL 60604

Armstrong, Katherine
Development Engineer
Monsanto Research Corp.
P.O. Box 32
Miamisburg, OH 45342

Austin, David S.
Tech. Associate
Eastman Kodak Co.
Kodak Park Division
Rochester, NY 14650

Austin, J.A.
Supervisor
Mobil Chemical Co.
One Greenway Plaza, #1100
Houston, TX 77046

Balentine, Jack
Evir. Marketing Coordinator
Catalytic Inc.
1500 Market St.
Philadelphia, PA 19002

Ball, Roy O.
Environmental Resources Mgmt.
200 S. Prospect
Park Ridge, IL 60068

Banerji, Shankha - Prof.
University of Missouri
2037 Engineering Building
Columbia, MO 65211
Banker, Michael R.
Allen &Hoshall, Inc.
2430 Poplar Ave.
Memphis, TN 38112

Barkley, Naomi P.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Barr, William T.
Eastman Kodak Co.
E.T.S. 8th Floor B-23 Kodak Park
1669 Lake Avenue
Rochester, NY 14650

Bass, Jeffrey
Arthur D. Little, Inc.
15 Acorn Park
Cambridge, MA 02140

Bath, Thomas D.
Consultant
2555 M Street, NW
Washington DC  20037

Baugh, Thomas
Defense Property Disposal
  Service
74 N. Washington DPDA-HET
Battle Creek, MI 49016

Beecher, Norman
Assoc. Prof. Tufts Univ.
Medford, MA02155

Beggs, Thomas W.
JACA Corp.
550 Pinetown Road
Ft. Washington, PA  19034

Belk, James D. -V.P.
Welker& Assoc., Inc.
P.O. Box 937
328 Roswell St.
Marietta, GA 30061

Beltis, Kevin J.
Environmental Chemist
Arthur D. Little, Inc.
15 W.  315 B. Acorn Park
Cambridge, MA 02140
                                     283

-------
Benz, Edward Paul
Project Geologist
Paulus, Sokolowski & Sartor
67 Mountain Blvd. Ext.
Warren, NJ 07060

Beranek, Jr.
President
Beranek Associates
7442 Countrybrook Drive
Indianapolis, IN 46260

Berkowitz, Jorge H.
Bureau Chief
Dept. of Environ. Protection
8 E. Hanover St.
Trenton, NJ 08625

Bernson, Laurence
Air Monitoring Section
US EPA, Region 2
Woodbridge Ave.
Edison, NJ 08837

Belts, Stephen C.
Principal Associate
PRC Consoer Townsend
404 James Robertson Pkwy.
Nashville, TN 37219

Betzhold, Fred C. -  Mgr.
Goodyear Tire & Rubber Co.
Dept. 100D
1144 East Market Street
Akron, OH 44316

Bipes, Roger L.
Asst. Site Director
E.M. Scinece
2909 Highland
Cincinnati, OH 45212

Birch, Richard F.
Manager
Envir. Ex-Cell-o Corp.
2155Coolidge
Troy, MI 48084

Bizzoco, Francis A.
Envir. Engr. Office
Dept. of Army HQDA
  (DAEN-ECE-G)
Washington, DC 20314
Black, Michael I.
Environmental Engr.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Blaha, Frank
University of Wisconsin
1218 Engineering Bldg.
1415 Johnson Dr.
Madison, WI 53713

Blake, Peter J.
Toxic Waste Containment, Inc.
53 D Street, SE
Washington, DC 20003

Blaney, Ben
Physical Scientist
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Blanz, Robert E.
Deputy Director
Pollution Control & Ecology
P.O. Box 9583
Little Rock, AR 72219

Blick, Clifton T.
Environmental Control
E.J.DuPontCo.
P.O. Box 2042
Wilmington, NY 28402

Bodocsi, Andrew - Assoc. Prof.
University of Cincinnati
MailLoc.#71
Cincinnati, OH 45221

Boje, Rita Rae
Indiana State Board of Health
1330 W.Michigan St.
Indianapolis, IN 46206

Bond, Rick
Res. Engineer
Battelle
P.O. Box 999
Richland, WA 99352
                                    284

-------
Borner, Alan J. - Exec. Dir.
Environmental Hazards Mgmt.
  Institute
Box 283,45 Pleasant St.
Portsmouth, NH 03801

Boschuk, John Jr. - V. P.
Geotechnical Division
  Orbital Eng., Inc.
1344FifthAve.
Pittsburg, PA 15219

Boszak, Gary P.
Facilities Engr.
GMAD Warren
30009 Van Dyke
Warren, MI 48090

Bothnor, CarlH.
Env. Engr.
ARMCO, INC.
P.O. Box 600
Middletown, OH 45043

Bowen, Russell V.
Staff Engineer
ESE, Inc.
P.O. BoxESE
Gainesville, FL 32602

Brausch, Leo M. - Mgr.
Project Development
D'Appolonia
10 Duff Road
Pittsburgh, PA 15235

Breeding, David C.
Asst. Prof.
Walters State Community College
Morristown, TN 37814

Bridges, James S.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Bridle, T.R.
Environment Canada
P.O. Box 5050
Birlington, Ontario
Brogard, John N.
Env. Engr.
US EPA, Region II
26 Federal Plaza
New York, NY 10278

Brooks, Barry R.
Marketing Mgr.
Energy Inc.
P.O. Box 736
Idaho Falls, ID 83401

Brooks, John G.
Env. Supervisor
KY Dept. EPA
400 E. Gray St.
Louisville, KY 40299

Broshears,  Robert E.
NSF Fellow
Vanderbilt University
Box 6304 B
Nashville, TN 37235

Bross, Ray A.
Engineer
City of Cincinnati, MSD
1600GestSt.
Cincinnati, OH 45204

Brown, David S.
Mgmt. Mktg., Wyo-Ben, Inc.
1242 N. 28th St.
P.O. Box 1979
Billings, MT 59103

Brown, Donald
Consultant
NY-TREX, Inc.
3969 Congress Parkway
Richfield, OH 44286

Brown, K.W.
Texas A & M University
Dept. of Soil & Crop Sciences
College Station, TX 77843

Brunsing, Thomas P.
Program Mgr.
Foster-Miller, Inc.
350 Second Ave.
Waltharn, MA02154
                                    285

-------
Budde, W.L.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Budzin, Gerald
Bendix Aircraft Brake & Strut.
3520 Westmoor
P.O. Box 10
Southbend,IN 46624

Buewick, John A. - Dir.
Dev. of Environ. Affairs
Tufts University
Packard Hall
Medford,MA02155

Buice, J.E.
Process Specialist
Dow Chemical Co.
Bldg. A-1107
Freeport, TX 77541

Burkart, Joseph K.
Mech. Engr. MERL/USEPA
5995 Center Hill Road
Cincinnati, OH 45268

Burke, Kim K.
Attorney
Taft, Stettinius & Hossister
1800 First National Bank Center
Cincinnati, OH 45202

Butler, Larry-V.P.
Disposal Operations
US Pollution Control, Inc.
2000 Classen Ctr., Ste. 320
S. Oklahoma City, OK 73106

Butts, Charles
GeoEngineering, Inc.
100 Ford Rd.
Bldg. 3
Denville, NJ 07866

Butt, Karl
Regulatory Analyst
JRB Associates, Inc.
8400 West Park Dr.
McLean, VA 22102
Calouche, Samir I.
Chemist
Virginia State Dept. of Health
109 Governor St.
Richmond, VA 23219

Campbell, H.W.
Environment Canada
Wastewater Tech. Ctr.
P.O. Box 5050
Birlington, Ontario

Carcone, Eugene A.
Dir. of Loss Control Field Service
Utica Mutual Ins. Co.
P.O. Box 530
New Hartford, NY 13413

Carfora, Stephen J.
NY EPA
Division of Waste Mgmt.
120Rt. 156
Yardville, NJ 08620

Carlson, Diane, M.
State of Michigan
WQD
P.O. Box 30028
Lansing, MI 48909

Carnes,  Richard A.
Environmental Scientist
US EPA, Combustion Research Fac.
Jefferson, AR 72079

Carter, Patricia
Dir. of Public Relations
Russell & Associates, Inc.
2387 W.Monroe St.
Springfield, IL 62704

Carver,  Val B.
Process Engineer
Trade Waste Incineration
#7 Mobile Dr.
Saugey,IL 62201

Cashell, Margaret M.
Civil & Environmental Engr.
University of Cincinnati
703 Rhodes Hall #71
Cincinnati,  OH 45221
                                        286

-------
Cassidy, Paul
US EPA (WH-565E)
4th & M Streets, SW
Washington, DC 20460

Castaldini, Carlo
Project Engineer
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Castle, PaulM.
General Mgr.
W.C.Meredith Co., Inc.
P.O. Box 90456
East Point, GA 30364
Chehaske, John T. - Mgr.
Engineering & Monitoring
Engineering Science
10521 Rosehaven St.
Fairfax, VA 22030

Cherry, Kenneth F. - Mgr.
Clayton Environmental
Consultants, Inc.
25711 Southfield Rd.
Southfield, MI 48075

C'hilds, Kenneth A.
Advisor Site Remediation
Environment Canada
Ottawa, Ontario K19K8
Cavalcanti, Fernando - A.P.
Combustion Engr.
Union Carbide Corp.
P.O. Box 8361
South Charleston, WV 25303

Cawley, William A.
Acting Director
US EPA, IERL-CI
26W. St. Clair
Cincinnati, OH 45268

Chadbourne, Dr. John F.
Director of Environmental Affairs
General Portland Inc.
P.O. Box 324
Dallas, TX 75221

Charnberlin, Leland E.
Sprtdt. Envir. Activities
Harrison Radiator Div. GMA
200 Upper Mountain Road
Lockport, NY 14094

Chapman, Wayne
General Manager
NY-TREX, Inc.
3969 Congress Parkway
Richfield, OH 44286

Chawla, Ramesh C.
Associate Professor
Chemical Engineering Dept.
Howard University
Washington, DC 20059
Cho, Hak K.
Envir. Eng.
US EPA Region V
230 S. Dearborn St.
Chicago, IL 60604

Clark, Ann W.
Envir. Eng.
Rohm and Haas Co.
Box 584
Bristol, PA 19007

Clark, Lynn A.
The BF Goodrich Company
Chemical Group
6100 Oak Tree Boulevard
Cleveland, OH 44131

Clark, Scott
Dept. of Environmental Health
University of Cincinnati
Mail Loc. #56
Cincinnati, OH 45267

Clarke, W.H.
Partner, Henderson and Bodwell
3476 Irwin Simpson Road
                      . VP.
Pioneer Equities, Inc.
One Wheaton Center
Suite 1801
                                     287

-------
Clarke, William J.
President
Geochemical Corp.
162 Spencer Place
Ridgewood, NJ 07450

Claunch, C. Kenneth
President
921 Greengarden Rd.
Eric, PA 16505

Clear, Jack
Envir. Research Group, Inc.
117 N. First Street
Ann Arbor, MI 48104

Clyde, Robert
Chemical Engineer
Consultant
Box 983
Asheville, NC

Cochron, S. Robert
JRB Associates
8400 West Park Dr.
McLean, VA 22102

Coffman, Glenn N.
Sr. Engineer
Law Engineering Testing Co.
2749 Delk Road SE
Marietta, GA 30067

Coia, Michael F.
Roy F. Weston, Inc.
Weston Way
West Chester, PA 19380

Cointreau, Sandra - Pres.
Solid Waste Mgmt.
Consulting Services, Ltd.
RR1-585 Old Shortwoods Rd.
New Fairfield, CT06810

Coker, David M.
Env. Control Engineer
Aluminum Co. of America
 1501 Alcoa Building
Pittsburgh, PA 15211

Collins, T. Leo
Mgr. Env. Quality
General Electric
NottSt.
Schenectady, NY 12309
Cook, RogerS.
KY Division
  Air Pollution Control
ISReillyRd.
Frankfort, KY 40601

Cooke, Marcus
Director
Battelle-Columbus
505 King Ave.
Columbus, OH 43201

Cooper, John
Southwest Research Institute
P.O. Drawer 28510
San Antonio, TX 78284

Costello, Richard J.
P.E.  NIOSH
4676 Columbia Pkwy.
Cincinnati, OH 45202

Coulter, Royal
Peoria Disposal Co.
Hazardous Waste Landfill
Peoria, IL 61604

Cowan, Dr. Bruce M.
Project Manager
A.M. Kinney, Inc.
2900 Vernon Place
Cincinnati, OH 45219

Cox, Gary R. - Engr.
Rockwell International
   Hanford Operations
P.O. Box 800
Richland, VA 99352

Coxe, Edwin F.
Associate Vice President
Reynolds, Smith & Hills
P.O. Box 4850
Jacksonville, FL 32201

Craig, AlfredB. Jr
US EPA
Industrial Envir.  Res. Lab
26 W. St. Clair
Cincinnati, OH 45268
                                    288

-------
Crawford, Douglas M.
University of Cincinnati
MERC-SHWRD
26 W. St. Clair St.
Cincinnati, OH 45268

Crawford, James G.
Vice President
AAA Environmental Industries
5544 W. Forest Home Ave.
Milwaukee, WI 53220

Crider, Al A.
American Excel-Ltd.
1 E. Main St.
P.O. Box 510
S.Vienna, OH 45369

Crocket, Alan - Mgr.
Environmental Sciences
EG&G Idaho
P.O. Box 1625
Idaho Falls, ID 83401

Cross, Terr L.
VA Health Dept.
109 Governor St.
Richmond, VA 23219

Crumbliss, Ralph - Mgr.
Sales & Marketing
Gulf seal Corp.
601 Jefferson
Houston, TX 77002

Grumpier, Eugene P.
Chemical Engineer
US EPA
401 M Street, SW
Washington, DC 20460

Curran, Mary Ann
Chemical Engineer
US EPA
26 W. St. Clair
Cincinnati, OH 45268

D'Aquila, Margaret M.
Technical Sales Specialist
Mead Compuchem
550 W. Brompton
Chicago, IL 60657
Dalberto, Alfred
PA Dept. of Environmental
  Resources
P.O. Box 2063
Harrisburg, PA17120

Daniel, David E.
Asst. Prof.
University of Texas
Dept. of Civil Energy
Austin, TX 78712

Dantin, Elvin J.
Hazardous Waste
  Research Centers
Louisiana State Univ.
Baton Rouge, LA 70803

Davis, J.S.
Senior Geophysicist
EarthTech, Inc.
6655 Amberton Drive
Baltimore, MD 21227

Dawson, Donna
Envir. Specialist
NY EPA
120Rt. 156
Yardville, NJ 08620

Degler, Gerald H.
Sen. Engr.
Bowser-Morner
P.O. Box 81
Dayton, OH 45440

Deiss, Richard A.
Richard A. Deiss & Assoc.
R.D. 1
Alden Street Extension
Meadville, PA 16335

Dell, Lee
Principal
Dell Engineering
146 South River Ave.
Holland, MI 49423

Dellinger, Barry
Group Leader
University of Dayton
Research Institute
Dayton, OH 45469
                                    289

-------
Delph, Larry - Coordinator
Environmental Protection
Lexington Health Dept.
650 Newtown Pike
Lexington, KY 40508

DePorter, Gerald L.
Los Alamos National Lab
P.O. Box 1663
Group LS-6, MSK495
Los Alamos, NM 87545

Desai, Harish
Illinois EPA
Div. of Air Pollution Control
2200 Churchill Rd.
Springfield, IL 62704

Determann, James
KY Division of Waste Mgmt.
Ft. Boone Plaza
ISReillyRd.
Frankfort, KY 40601

Detweiler, Roy R.
Mgr. Env. Affairs
DuPont Co. Biochemicals
Barley Mill Plaza #7
Wilmington, DE 19898

DeVault, Cleve N.
Ohio State Univ.
2070 Neil Ave.
Room 470, Hitchcock Hall
Columbus, OH 43210

Dial, Clyde J.-Dir.
US EPA
Energy Pollution Control Div.
26 W. St. Clair
Cincinnati, OH 45268

Dickinson, Robert H.
Coordinator Corp.
Westraco Corp.
299 Park Avenue
New York, NY 10171

Djafari, Dr. Sirous Haji - Mgr.
D'Appolonia Waste
  Management Services
10 Duff Road
Pittsburgh, PA 15235
Dods, David A.
Graduate Student
Vanderbilt Univ. Box 6304-B
Nashville, TN 37235

Donaldson, Robert T.
Massachusetts Dept.
  of Environ. Qlty Eng
One Winter St., 8th Fir.
Boston, MA 02108

Douglas,  Jeff M.
Civil Engineer
American Fly Ash Co.
606 Potter Rd.
DesPlaines, IL60016

Downey,  Robert A.
Geologist
Indiana State Board of Health
1330 W.Michigan
Indianapolis, IN 46220

Downey, Thomas W.
Sr. Env. Spec.
NY EPA, Div. of Waste Mgmt.
120Rt. 156
Yardville, NJ 08620

Doyle, John D.
P.E. Section Chief
Dept.  of Nat. Resources
P.O. Box 1368
Jefferson City, MO 65102

Drobny, Neil L.
President
ERM-Midwest, Inc.
4621 Reed Road
Columbus, OH 43220

Duffala, DaleS.
Env. Scientist
Black  & Veatch
P.O. Box 8405
Kansas City, MO 64114

Duke, John
Engineer
Procter & Gamble
7162 Reading Rd.
Cincinnati, OH 45222
                                    290

-------
DuRoss, Frank B.
Oneida Asbestos Removal Inc.
333 South St.
Utica, NY 13501

DuRoss, James F.
Vice President
Oneida Asbestos Removal Inc.
333 South Street
Utica, NY 13501

Dzindzeleta, Mercedes - Pres.
Energy & Environmental
  Mgmt., Inc.
P.O. Box 422
Racine, WI  53401

Edwards, Linda E.
Admin. Manager
WMS Dames & Moore
644 Linn St., Ste. 501
Cincinnati,  OH 45203

Edwards, Stuart
Senior Engineer
Dames & Moore
1150W. 8th
Cincinnati,  OH 45203

Ehrenfeld, John R.
Arthur D. Little, Inc.
15 Acorn Park
Cambridge, MA 02140

Eicher, Anthony R.
Chemical Engineer
IT Corp
3333 Vine St., Ste. 204
Cincinnati,  OH 45220

Eide, Allan
Minnesota Waste Mgmt. Board
123 Thorson Community Center
7323 58th Ave. North
Crystal, MN 55428

Eisen, Paul
Wapora, Inc.
21 IE. 43rd St.
New York,  NY 10017
Eith, Anthony W.
Sr. Project Engineer
Orbital Engineering, Inc.
1344 Fifth Ave.
Pittsburgh, PA 15219

Eiam, David L.
Project Scientist
Harmon Engineering
Auburn Industrial Park
Auburn, AZ 31830

Eliades, Nick
Fort Motor Co.
17000Oakwood
RMF3016
Allen Park, MI 48101

Ellwood, Theodore R.
Industrial Hygienist
IT Corp
3333 Vine St., Ste. 204
Cincinnati, OH 45220

Ely, Robert G.
Watkins & Associates, Inc.
446 E. High St.
P.O. Box 951
Lexington, KY 40588

Emrich, Grover H.
SMC Martin Inc.
900 W. Valley Forge Road
P.O. Box 859
Valley Forge, PA 19482

Erdmann, Fred W.
Soil & Materials Engrs., Inc.
11325 Reed Hartman Hwy.
Suite 134
Cincinnati, OH 45241

Ericson, Franklyn A.
The Upjohn Co.
7171 Portage Rd.
6101-41-0
Kalamazoo, MI 49001

Esposito, Dr. R.G.
Union Chemical Co., Inc.
P.O. Box 423
Union, ME 08862
                                    291

-------
Ewing, Tom
3130 Bishop St.
Cincinnati, OH 45220

Ezell,  James D. - Mgr.
SLTC
1640Antioch
Antioch,TN37013

Falconer, Kathleen L.
Senior Engineer
EG&G Idaho, Inc.
P.O. Box 1625
Idaho Falls, ID 83401

Farhoudi, Koorosh
Division of Pollution Center
Ft. Boone Plaza
18ReillyRd.
Frankfort, KY 40601

Farnsworth, Alan
Schlegel Corp.
P.O. Box 23113
Rochester, NY 14692

Favero, David - EPS
IL Env. Protection Agency
2200 Churchill Rd.
Springfield, IL 62706

Feeley, James A.
Hydrologist
TX Dept. of Water Resources
P.O. Box 13087
Austin, TX 78711

Feldmann, John L.
KY Division of Waste Mgmt.
7964 Kentucky Drive, Ste. 8
Florence, KY 41042

Fennelly, Dr. Paul - Mgr.
Envir. Measurements Dept.
GCA/Technology Division
213 Burlington Rd.
 Bedford, MA 01730

Finucane, Matthew D.
University of Pennsylvania
 Nursing Education Building
420 Service Drive/S2
 Philadelphia, PA 19104
Fish, Douglas K.
DuPont Co.
PPD/ELD
Barley Mill Plaza
Wilmington, DE 19707

Fisher, Gerald E.
Materials Eng.
E.I. DuPont
105224 Terry Trail
Hinsdale, IL 60521

Flaig, James J.
Vice President
The H.C. Nutting Co.
5802 Beechnut Drive
Cincinnati, OH 45230

Flanigan, Jack
Plant Supt.
CloudsleyCo.
470 W. Northland Rd.
Cincinnati, OH 45240

Fleischman, Marvin - Chrm.
Dept. of Chemical
  & Environ. Eng.
Univ. of Louisville
Louisville, KY 40292

Flett, Gregory
Eder Associates
  Consulting Engrs.
85 Forest Ave.
Locust Valley, NY 11560

Flood, Jared W.
Environmental Engineer
US EPA
401 M Street, SW
Washington, DC 20460

Fochtman, Ed
Manager
Chemical Waste Mgmt., Inc.
3003 Butterfield R.
Oak Brook, IL 60521

Foss, C.B.
Bulk Petroleum Operations
Crowley Maritime Corp.
One Market Plaza
San Francisco, CA 94105
                                   292

-------
Foster, George R.
Los Alamos National Lab
P.O. Box 1663
MS-K495
Los Alamos, NM 87545

Foushee, Roy
Lexington KY Health Dept.
650 Newtown Pike
Lexington, KY 40508

Fowler, Charles F.
Health and Safety Coordinator
Versar, Inc.
Building 45
Jefferson, AK 72079

Fowler, David E.
128 Greenlawn Ave.
Findlay, OH 45840

Fox, Robert D.
IT Enviroscience
312 Directors Drive
Knoxville, TN 37923

Fralinger, Albert A. - SES
Sr. Environ.  Specialist
NJ EPA, Div. of Waste Mgmt.
RD#l,Rt. #1
Vincentown, NJ 08088

Freeman, Henry M.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Fuchs, Robert E.-V.P.
Environmental
  Consultants, Inc.
391 Newman Ave.
Clarksville, IN 47130

Fuehrer, John G. II
Fuehrer Associates
345 W. Main Street
Ephrata, PA 17522

Garcia, L.H.
Chem. Engr.
US EPA, IERL-CI
26 W. St. Clair
Cincinnati, OH 45268
Gardner, George D.
NUS Corp.
Part West Two
Cliff Mine Road
Pittsburgh, PA 15275

Gashlin, Kevin
Sr. Envir. Specialist
NY EPA
120Rt. 156
Yardville, NJ 08620

Gaynor, Charles T. II
Manager
Thomsen Associates
105 Corona Ave.
Groton, NY 13073

Gaynor, Ronald K.
Services & Safety
US Ecology, Inc.
9200 Shelbyville Road
Louisville,  KY 40222

Gee, John
University of Wisconsin
1218 Engineering Bldg.
1415 Johnson Drive
Madison, WI 53713

Georgevich, Maurice
Ind. Hyg.
NIOSH
4676 Columbia Pkwy.
Cincinnati, OH 45226

Gerbracht, Elmer K.
Technical Director
ACTS Testine Labs, Inc.
3900 Broadway
Buffalo, NY 14227

Ghia, Jay R. - Mgr.
Hazardous Waste Program
Harza Engineering Co.
150 S.Wacker Drive
Chicago, IL 60606

Gilbert, George
KY  Division of Waste Mgmt.
Ft. Boone Plaza
18ReillyRoad
Frankfort, KY 40601
                                   293

-------
Gilder, Cindy
US EPA, Region I
JFK Federal Bldg.
SWPB
Boston, MA 02203

Gillespie, Dennis P.
SCS Engineers
211 Grand view Dr.
Suite 315
Covington, KY 41017

Gillespie, Elizabeth
KY Division of Waste Mgmt.
Ft. Boone Plaza
ISReillyRd.
Frankfort, KY 40601

Givens-Reynolds, Louise C.
VA Toxics Roundtable
P.O. Box 89
Salem, VA 24153

Glysson, Eugene A.
Prof, of Civil Engineering
Civil Engr. Dept.
Univ. of Michigan
Ann Arbor, MI 48109

Gohara, Wadie F.
Development Engineer
Babcock&WilcoxCo.
Barberton, OH 44203

Goldkamp, William
Univ. of Missouri-Columbia
Dept. of Civil Engineering
Columbia, MO 65211

Gore, William D.
Vanderbilt University
Box 6304 B
Nashville, TN 37235

Gorman, Paul
Chem. Engr.
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO 64110

Gorski, Mitchel R. Jr.
Sales Manager
ThermAll, Inc.
P.O. Box 1776
Peapack, NJ 07977
Goshee, Gary B.
Environmental Engineer
US EPA, Region I
JFK Federal Bldg.
Boston, MA 02203

Gould, Cliff
Env. Prot. Spec. IL EPA
5817 Doe Circle
Westmont, IL 60559

Gower, Mike
Service Rep.
Gabriel & Associates
1814 N. Marsh Field Ave.
Chicago, IL 60622

Gracey, Charles M.
Sr. Engr. Spec.
Aerotect Liquid Rock Co.
P.O. Box 13222
Sacramento, CA 95813

Graham, John L.
University of Dayton
  Research Institute
KL 101C
Dayton, OH 45469

Grant, Kevin D.
Mgr. Regulatory Affairs
SCA Chemical Services
5 Middlesex Ave.
Somerville, MA

Grauvogel, Lawrence W.
Project Engineer
Cole Associates, Inc.
221 IE. Jefferson Blvd.
South Bend, IN 46615

Gredell, Thomas R.
Environmental Engr.
MO Dept. of Nat. Resources
P.O. Box 1368
Jefferson City, MO 65102

Green, Mark
Environmental Staff Engr.
Rust Internal Corp.
P.O. Box 101
Birmingham, AL 35201
                                   294

-------
Hall, Robert R.
Senior Staff Engr.
GCA/Technology Division
213 Burlington Rd.
Bedford, MA 01730

Haller, Ray C. - Pres.
RayHallerlnc.
Consulting Engineer
363 Bennington
Indianapolis, IN 46227

Hamlin, Wm. E.
Sales Mgr.
Arizona Refining Co.
P.O. Box 1453
Phoenix, AR 85001

Handyside, Thomas A.
Vice President
City Disposal Systems, Inc.
15 50 Harper
Detroit, MI 48211

Hanson, Eric G.
Consultant
P.O. Box 750151
New Orleans, LA 70175

Harmon, CarlB.
Principal Engineer
Watkins & Assoc., Inc.
446 E. High St.
Lexington, KY 40508

Harris, Judith C. - Mgr.
Chemical & Food Sciences
Arthur D. Little, Inc.
15-311 Acorn Park
Cambridge, MA 02140

Hartley, Robert P.
Physical Scientist
US EPA, MERL/SHWRD
26 W. St. Clair
Cincinnati, OH 45268

Hawfield, Robert A.
Post, Buckley, Schuh
  & Jernigan, Inc.
P.O. Box 106
Cola, SC 29202
Haxo, Henry E. Jr.
President
Matrecon, Inc.
2811 Adeline St.
Oakland, C A 94608

Hayes, Joe R.
Dept. of Environmental
  Resources
Rt. 2, Box 225
Bernville, PA 19506

Hazelwood, Douglas
Associate
A.T. Kearney,Inc.
P.O. Box 1405
Alexandria, VA 22313

Hedden, Kenneth F.
Environmental Engr.
EPA Env. Res. Lab.
College Station Rd.
Athens, G A 30613

Heitz, Michael W.
Engineer
Metro Sewer District
1600GestSt.
Cincinnati, OH 45204

Held, William M.
Staff Engineer
SCS Engineers
211 GrandviewDr.
Covington, KY 41017

Henz, Don J.
Assoc. Dir. of Engrg.
Pedco Environmental, Inc.
11499 Chester Rd.
Cincinnati, OH 45246

Herbst, Dr. Richard P.
Env. Coordinator
Exxon Minerals Co.
P.O. Box 4508
Houston, TX 77210

Herning, Leland P.
Envir.  Engr.
Gulf Oil Corp.-Cinn. Ref.
P.O. Box 7
Cleves, OH 45002
                                    295

-------
Herrman, Jonathan G.
US EPA
26W.St. Clair
Cincinnati, OH 45268

Hersh, Stewart
Vice Pres.
KVB Inc.
175ClearbrookRd.
Elmsford, NY 10523

Hess, Connie
Hess Environmental
  Services, Inc.
6497 Oak Park Dr.
Memphis, TN 38134

Higgins, Greg M.
Project Mgr.
Systech Corp.
245 N. Valley Rd.
Xenia, OH 45355

Hill, Ronald D.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Hillard, Ray L.
American Cyanamid Co.
Bound Brook, NJ 08801

Hinkley, David R.
Director Special Services
Central Hudson GDE Corp.
284 South Avenue
Poughkeepsie, NY 12603

Hoad, George E.
University of Connecticut
U-37
Storrs, CT 06268

Hogan,  John C.
Armco,  Inc.
703 Curtis Street
Middletown, OH 45043

Holberger, Richard
MITRE Corp.
1820 Dolley Madison Blvd.
McLean, VA 22101
Holroyd, Louis V.
Univ. of Missouri
8 Res. Pk. Dev. Bldg.
Columbia, MO 65211

Holton, Gregory A.
Oak Ridge National Lab
P.O. Box X
Oak Ridge, TN 37830

Home, Jim
NUS Corp.
900 Gemini
Houston, TX 77058

Hornig, Arthur W.
Thuyard Research
3303 Harbor Blvd.
Suite C-8
Costa Mesa, CA 92626

Horton, James F.
MERL/US EPA
5995 Center Hill Rd.
Cincinnati, OH 45268

Horz, Raymond C.
USAE Waterways Experiment
  Station
P.O. Box 631
Vicksburg, MS 39180

Houthoofd, Janet N.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Howell, S. Gary
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Hu, Alan
Rochester Institute of Technology
Rochester, NY 14623

Hubbard, Allen P.
ESE, Inc.
P.O. BoxESE
Gainesville, FL 32602
                                      296

-------
Huddleston, Walter E.
Mason & Hanger Co.
P.O. Box 30020
Amarillo, TX79177

Huffman, George L.
US EPA Center Hill Facility
5995 Center Hill Rd.
Cincinnati, OH 45268

Huggins, Andrew
SynCo Consultants Inc.
12 Bank Street
Bank Street Center
Summit, NJ 07901

Hull, John
Hull Consulting
2726 Monroe St.
Toledo, OH 43606

Hunninen, Katherine
Univ. of Cincinnati
3223 Eden Ave.
Cincinnati, OH 45267

Hunt, Gary
JRB Associates
8400 Westpark Dr.
McLeon, VA22102

Hurley, Steve
Dept. of Navy
200 Stovall St.
Alexandria, VA 22332

Huston, Arthur C.
Washington Works
E.I. du Pont de Nemours & Co.
P.O. Box 1217
Parkersburg, WV 26102

Hutzler, Neil
Michigan Technological Univ.
Dept. of Civil Engineering
Houghton, MI 49931

Hyams, Richard W.
Lockwood, Kessler & Bartlett
One Aerial Way
Syosset, NY 11791
Hyland, James
Dana Corp.
8000 Yankee Rd.
Ottawa Lake, MI 49267

lannuzz, Alphonse Jr.
NJDEP
1259R1.46
Parsippany, NJ 07054

luliucci, Robert L.
Sun Chemical Corp.
4605 Esk Ave.
Cincinnati, OH 45232

Irwin, J. Andrew
O'Brien & Gere Engineers
1304 Buckley Rd.
Syracuse, NY 13221

Jackson, C.S.
US EPA, Region IX
215 Fremont
San Francisco, CA 94105

Jackson, Danny R.
Research Scientist
Battelle-Columbus
505 King Ave.
Columbus, OH 43201

Jacobs, Philip W.
Daily &Assoc.
Engineers Inc.
3716 W. Brighton Ave.
Peoria, IL61615

Jacobson, Laurie
Research Technician
10N.E. 17th St.
Rochester, MN 55901

Jaeger, Ralph R.
Monsanto Research Corp.
Mound Lab
Miamisburg, OH 45342

Jahns, Ronald W.
Illinois EPA
2200 Churchill Rd.
Springfield, IL 62706
                                   297

-------
Jakobson, Kurt
US EPA
401 M Street, SW
Washington, DC 20460

James, Ernie
Stauffer Chemical Company
20720 S. Wilmington Ave.
Long Beach, CA 90810

James, Ruby H.
Southern  Research Institute
2000 9th Avenue, South
Birmingham, AL 35255

Janik, DavidS.
University of Cincinnati
Cincinnati, OH 45221

Jansen, Joe
Missouri Dept. of Natural Res.
P.O. Box 1368
Jefferson City, MO 65106

Jargowsky, Lester W.
Monmouth County Health
  Department
Route 9 and Campbell Court
Freehold, NJ 07728

Jarrett, Paul T.
Vanderbilt University
487 Clairmont Place
Nashville, TN 37215

Jaspers, Gregory
Camargo Associates
1329 East Kemper Road
Cincinnati, OH 45246

Jepsen, Christopher P
Technical Supervisor
American Colloid Co.
SlOOSuffieldCt.
Skokie, IL 60077

Jessee, Gene
Monsanto Company
SOON. Lindbergh
St. Louis, MO 63167
Jett, Morris E.
Schlegel Lining Tech., Inc.
200 S. Trade Center Parkway
P.O. Box 7730
The Woodlands, TX 77380

Jhaveri, Vidjut
Groundwater Decontamination
  Systems, Inc.
12 Industrial Park
Waldwick, NJ 07463

Johannesmeyer, Herman
Univ.  of Missouri-Columbia
Dept.  of Civil Engineering
Columbia, MO 65211

Johnson, Charles E.
Procter & Gamble Co.
Invorydale Technical Center
Cincinnati, OH 45217

Johnson, J.S.
CIBA-Geigy Corporation
Ardsley, NY 10502

Johnson, Larry D.
US  EPA
Research Triangle Park, NC 27711

Johnson, Thomas M.
Illinois Geological Survey
615 E. PeabodySt.
Champaign, IL 61820

Johnston, Eileen L.
Environmental Educator
505 Maple Avenue
Wilmette, IL 60091

Johnstone, John
Corp. Dir. Eng.
Knowlton Bros.
105 W. 45th
Chattanooga, TN 37410

Jones, Curtis
Kentucky Department for
  Environmental Protection
400 E. Gray St.
Louisville, KY 40202
                                       298

-------
Jones, Larry W.
Veritec Corporation
P.O. Box 8791
Knoxville, TN 37996

Joyce, William F.
Stauffer Chemical Company
Eastern Research Center
Dobbs Ferry, MN 10522

Julovich, Peter
111. Inst. of Tech. (3)
7405 W. 400 N.
Michigan City, IN 46360

Jung, Kim E.
R.L. Wurz Company
P.O. Box 223
West Chester, OH 45069

Jungelaus, Gregory
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO 64110

Kachmarsky, Dennis J.
Finkbeiner, Pettis
  & Strout, Ltd.
4405 Talmadge Road
Toledo, OH 43623

Kamphake, Lawrence
US EPA
26 W. St. Clair
Cincinnati,  OH 45268

Kelley, Mike
Waterways  Experiment Station
P.O. Box 631
Vicksburg,  MS 39180

Kelly, Ben
US Army Corps of Engineers
20 Massachusetts, NW
Washington, DC 20314

Kenning, Todd
Peoria Disposal Company
Hazardous  Waste Landfill
Peoria, IL 61604
Kerho, S.E.
KVB, Inc.
18006Skypark
P.O. Box 19518
Irvine, CA 92714

Kessler, Kimberly A.
Geotechnical & Materials
  Consultants, Inc.
1341 Goldsmith
Plymouth, MI 48170

Kim, YJ.
Chemical Engineer
US EPA Region  V
230 S. Dearborn
Chicago, IL 60604

Kimbrough, Charlotte W.
Sverdrup Technology, Inc.
Box 884
Tallahorne, TN 37388

King, Lawrence P.
Babcock & Wilcox
1562BeesonSt.
Alliance, OH 44601

Kingsbury, Bob
Regional Sales Manager
American Colloid Co.
P.O. Box 696
Laconia, NH 03246

Kingsbury, Carrie L.
Research Triangle Institute
P.O. Box 12194
Research Triangle Park,  NC 27709

Kinman, Riley N.
Dept. of Civil Engineering
University of Cincinnati
Cincinnati, OH 45221

Kithany, Subhash S.
Owens-Corning Fiberglas
Granville, OH 43223

Kjaelland, Bob
Ky Division of Waste Mgmt.
Ft. Boone Plaza
18ReillyRoad
Frankfort, KY 40601
                                    299

-------
Klee, Albert J.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Kleinhenz, Ned J.
Mgr. R & D Systech Corp.
245 N. Valley Rd.
Xenia, OH 45385

Klinger, Larry M.
Sr. Dev. Engineer
Monsanto Research Corp.
P.O. Box 32
Miamisburg, OH 45342

Klint, Steen
TRICIL LTD.
#1, Corunna, Ontario
Canada NON -160

Kmet, Peter
Wisconsin - DNR
Box 7921
Madison, WI 53707

Knauss, James D.
Dames & Moore
2551 Regency
Lexington, KY 40503

Kneuer, Paul R.
Executive Vice President
Norlite Corp.
628 S.Saratoga St.
Cohoes, NY 12047

Knowles, C.L. Jr.
Olin Corp.
120 Long Ridge Road
Stamford, CT 06904

Knowles-Porter, C.
Dames & Moore
350 W. Camino Garden Blvd.
Boca Raton, FL 33432

Knudsen, Dennis R.
Naval Surface Weapons Center
Dahlgren, VA 22448
Koczwara, Margaret K.
University of Cincinnati
Cincinnati, OH 45221

Koines, Arthur T.
US EPA
401 M Street, SW
Washington, DC 20460

Kolpa, Ronald L.
Iowa Department of
  Environmental Quality
DesMoines, IA50319

Kondas, Andrew
PADER
250KossmanBldg.
Pittsburgh, PA 15222

Koutsandreas, John D.
US EPA
401 M Street, SW
Washington, DC 20460

Kramlich, John C.
Energy and Environmental
  Research Corp.
18 Mason St.
Irvine, CA 92714

Krueger, John A. - Dir.
Dept. of Environ. Protection
State House Station #17
Augusta, ME 04333

Kuhn, Donald J.
SLC Consultants/Constructions
Box 603
North Tonawanda, NY 14120

Kuhn, Eric C.
Proctor Davis Ray Engineers
800 Corporate Drive
Lexington, KY 40503

Kush, George-V. P.
Environmental Affairs
SCA Chemical Services Inc.
5 Middlesex Ave.
Somerville, MA02145
                                     300

-------
Ladd, Donald M.
USAE Waterways
Experiment Station
Geotechnical Laboratory
Vicksburg, MS 39180

Laird, Duncan
USDA, Nat'l. Monitoring Lab
P.O. Box 3209
3505 25th Avenue
Gulf port, MS 39503-1209

Lambert, Martha E.
Research Asst.
University of Cincinnati
Mail Loc. #71
Cincinnati, OH 45221

Landreth, Robert E.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Lanford, Ed
Virginia Department of Health
109 Governor St.
Richmond, VA 23219

Lanier, William S.
US EPA
Research Triangle Park, NC 27711

Larsen, Deborah J.
Arthur D. Little, Inc.
15 Acorn Park
Cambridge, MA 02140

Lasrzal, Kenneth
Spaulding Fibre Co., Inc.
310 Wheeler
Tonawanda, NY14156

Laswell, Bruce
SRW Associates, Inc.
2793 Noblestown Road
Pittsburgh, PA 15203

Lauch, Elizabeth
Purdue University
323 Engineering Admin.
W. Lafayette, IN 47907
Lauer, William F.
Clinton Bogert Associates
2125 Center Avenue
Fort Lee, NV 07024

LaVake, Myron
Monmouth County
Health Department
Route 9 & Campbell Road
Freehold, NJ 07728

Lawson, Louis R. Jr.
Oldover Corp.
P.O. Box 27211
Richmond, VA 23261

Lee, C.C.
US EPA
26 W. St.  Clair
Cincinnati, OH 45268

Lee, Fred
E-Three Inc.
P.O. Box 155
Getzville, NY 14068

Lee, Kun-chieh
Union Carbide
P.O. Box 8361
Charleston, WV 25303

Lee, Louis W.
US EPA
Municipal Envir. Res. Lab
26 W. St.  Clair
Cincinnati, OH 45268

Lee, Lyon Y.
General Motors
3044 W. Grand Blvd.
Detroit, MI 48202

Lefke, Louis W.
US EPA
Municipal Envir. Res. Lab
26 W. St.  Clair
Cincinnati, OH 45268

Lenz, Vicki S.
US Ecology
P.O. Box 7246
Louisville, KY 40207
                                  301

-------
Leonard, Hannah H.
Kentucky National Research
& Env. Protection Cabinet
ISReillyRoad
Frankfort, KY 40601

Lewis, NormaM.
SHWRD
68W.St. Clair
Cincinnati, OH 45244

Lewis, Timothy A.
USGS
Reston, VA 22092

Lichtkoppler, Frank
Area Extension Agent
Ohio State University
99 East Eric St.
Cincinnati, OH 44077

Licis, Ivars J.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Lindsey, Alfred W.
US EPA
4th & M Streets, SW
Washington, DC 20460

Lindsey, Jerry V.
Rhone Ponbeve, Inc.
Mt. Pleasant, TN 38474

Lingle, Stephen A.
US EPA
401 M Street, SW
Washington, DC 20460

Lippitt, John M.
SCS Engineers
211 Grand view Dr.
Ft. Mitchell, KY

LiPuma, Terrance A.
Vice President
Engineering Science
10521 RosehavenSt.
Fairfax, VA 22030
Logan, Thomas J.
US EPA
Research Triangle Park, NC 27711

Logue, Edward R.
Maine EPA
Augusta, ME 04333

Loper, John C.
University of Cincinnati
Mail Loc. #524
Cincinnati, OH 45267

Lough, Chris
Pope-Reid Associates, Inc.
245 East 6th St., #813
St. Paul, MN 55101

Lowry, William F.
NJDEP
R.D. 1, Route 70
Vincentown, NJ 08088

Lu, David W.
Southwestern Ohio Air Pollution
Control Agency
2400 Beekman St.
Cincinnati, OH 45214

Lubowitz, H.R.
US EPA
13414 Prairie Avenue
Hawthorne, CA 90250

Lukey, Michael E.
Vice President
Engineering Science
10521 Rosehaven St.
Fairfax, VA 22030

Lynch, Maurice A. Jr.
Consultant
Lotepro Corp.
4248 Ridge Lea Road
Amherst, NY 14226

Lytle, Steven A. - Mgr.
Soil & Material Engineers
11325 Reed Hartman Hwy.
Suite 134
Cincinnati, OH 45241
                                      302

-------
MacDonald, Alison E.
9 Green tree Dr.
Phoenix, MD 21131
Mappes, Thomas E.
Cabot Corp.
Kokomo, IN 46901
Madden, Terry B.
University of Cincinnati
P.O. Box 23125
Cincinnati, OH 45223

Maffet, Vere - Mgr.
Research & Development
UOP Inc.
10 UOP Plaza
DesPlaines, IL 60016

Magelssen, L. Scott
Union Carbide
Box 8361
S.Charleston, WV 25303

Mahon, Joseph
Groundwater Decontamination
Systems, Inc.
12 Industrial Park
Waldwick, NY 07463

Malanchuk, Myron
EPA
26 W. St. Clair
Cincinnati, OH 45268

Males, Eric H.
ICF, Inc.
1850 K Street, NW #950
Washington, DC 20006

Malone, Philip G.
US Army Engineers
Vicksburg, MS 39180

Manning, Richard E.
Environmental Health
Colt Ind.
430 Park Ave.
New York, NY 10022

Mansfield, Charlie
Texas Instruments, Inc.
P.O. Box 1443
Mail Station 680
Houston, TX 77001
Markey, Margaret
Georgia Environmental
Protection Division
270 Washington St., SW
Atlanta, GA 30334

Marreko, Thomas R.
University of Missouri
Columbia, MO 65211

Marti, Luz R.
USDA Southeast Watershed
Res. Lab. CPES
P.O. Box 946
Tifton, GA31794

Martin, Ed
US EPA
401 M Street, SW
Washington, DC 20460

Martin, Elwood E.
Chemical Engineer
US EPA, OWPE WH527
401 M Street, SW
Washington, DC 20460

Mason, Howard B.
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Mason, Sam
Engineering Consultant A & S
5404 Peachtree Rd.
Atlanta, GA 30341

Matula, Richard A.
LA State University
College of Engr.
Baton Rouge, LA 70803

Maugham, Robert Y.
EG&G Idaho
P.O. Box 1625
Idaho Falls, ID 83401
                                   303

-------
May, James H.
US Army Engineer
Waterways Experiment Station
P.O. Box 634
Vicksburg, MS 39180

Mayer, Richard J.
The PQ Corp.
P.O. Box 840
Valley Forge, PA 19482

Mayne, Yolande C.
League of Women Voters
113 Marshall St.
Yellow Springs, OH 45387

Mayo, Francis T.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Mays, Kirk
Aramco Services Co.
1200 Smith #1426
Houston, TX 77002

McAdams, J.W.
Mobil Chemical Co.
211 College Rd. E.
Princeton, NJ 08540

McBath, Audrey
US EPA, IERL-C
26 W. St. Clair
Cincinnati, OH 45268

McCabe, Mark
GCA/Technology Division
213 Burlington Rd.
Bedford, MA 01730

McCormick, Robert J.
Acurex Corp
5658 Shadyhollow
Cincinnati, OH 45230

McCune, Harold E.
Armco Inc.
P.O. Box 600
Middleton, OH 45043
McDonnell, Gerald M.
G.M. Assoc., Inc.
8838 Spinnaker Ct.
Indianapolis, IN 46256

McDonough, James F.
Head Civil & Env. Engineer
University of Cincinnati
MailLoc. #71
Cincinnati, OH 45221

McGuire, Jerry N.
Monsanto Co.
SOON. Lindbergh
St. Louis, MO 63167

McKelroy, Rodney G.
NUS Corp.
900 Gemini
Houston, TX 77058

McLean, Loren A.
G.D. SearleLabs
4901 Searle Pkwy.
Skokie, IL 60077

McMahill, William F.
Univ. of Missouri
1020 Engineering Bldg.
Columbia, MO 65211

McNiel, Terrance J.
Michigan Dept. of
Natural Resources
P.O. Box 30038
Lansing, MI 48909

Meckes, Mark C.
Defense Property Disposal
Service
74 N. Washington DPDS-HET
Battle Creek, MI 49016

Melberg, John M.
Federal Cartidge Corp.
9th & Tyler St.
Anoka, MN 55303

Menzies, E. Miranda
Dames & Moore
1259 Garden Circle
Cincinnati, OH 45215
                                   304

-------
Merrick, Nelson J.
Aluminum Corp. of America
1501 Alcoa Bldg.
Pittsburgh, PA 15219

Merrill, Richard S.
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Meshkat, Masoud
University of Kentucky
Lexington, KY 40503

Metts, Dennis M.
University of Kentucky
Lexington, KY 40503

Meyer, G. Lewis
US EPA, Radiation
401 M Street, SW
Washington, DC 20460

Michblsbd, Donald L.
Virginia Tech.
Blacksburg, VA 24061

Michels, OtisE.
Daily & Assoc. Engineers, Inc.
3716 W. Brighton Ave.
Peoria, IL 61615

Mihelaraleis, Joseph L.
University of Cincinnati
Mail Loc. #71
Cincinnati, OH 45221

Miles, Mark
Eastman Kodak Co.
Kodak Park Division
Rochester, NY 14650

Milke, Mark
University of Wisconsin
1218 Engineering Bldg.
1415 Johnson Drive
Madison, WI 53713

Millan, RenatoC.
Wisconsin Dept. of
Natural Resources
101 S.Webster St.
Madison, WI 53707
Miller, Caryle B.
Dept. of Navy, Bldg. 212
CHESNAVFACENGCOM
Washington Navy Yard
Washington, DC 20374

Miller, David H.
Jones & Laughlin Steel Corp.
900 Agnew Road
Pittsburgh, PA 15211

Miller, Jo E.
University of Cincinnati
Cincinnati, OH 45221

Miller, Marvin P.
Battelle-Columbus
505 King Ave.
Columbus, OH 43201

Miller, Vern F.
Bow Valley Ltd.
2465 S. Estes Ct.
Lakewood, CO 80227

Mills, H. Doyle
Ky Environmental Protection
ISReillyRd.
Frankfort, KY 40601

Minkarah, Issam A.
University of Cincinnati
Mail Loc. #71
Cincinnati, OH 45221

Miullo, Nathaniel J.
US EPA Region 8
1860 Lincoln St.
Denver, CO 80925

Mixon, Forest O.
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709

Monnot, Donald R.
Envirodyne Engineers, Inc.
12161 Lackland Rd.
St. Louis, MO 63141
                                       305

-------
Monter, Louis A.
Senior Vice President
CLOUDSLEY CO.
470 W.Northland Road
Cincinnati, OH 45240

Moore, Ben
Engineer
B.H.S.,Inc.
R.R. l,Boxl!6-F
Wright City, MO 63390

Moore, Charles
Ohio State Univ.
Dept. of Civil Engineering
2070 Neil Ave.
Columbus, OH 43210

Moore, William P.
Rohm and Haas Co.
Engineering Division
Box 584
Bristol, PA 19007

Moran, Brian V.
US Army Corps of Engineers
Washington, DC 20314

Morekas, Georgeann
Duke University
301 Swift Ave. #17
Durham, NC 27705

Morrel, Mark
Fred C. Hart Assoc.
1110 Vermont Ave., NW #410
Washington, DC 20005

Morrison, Allen M.
Civil Engineering Magazine
345 E. 47th St.
New York, NY 10017

Mostara, Radmand
Penn. Dept. of Envir. Res.
Solid Waste Mgmt.
90 E. Union St.
WilkesBarre, PA 18701

Murphy, William
Sr. Env. Scientist
Pope-Reid Assoc., Inc.
245 E. 6th St.
St. Paul, MN 55101
Murray, John E.
SCA Services
3850 Lower Valley Pike
Springfield, OH 45506

Nandan, Shri
US Pollution Control, Inc.
Suite 320 South
2000 Classen Center
Oklahoma City, OK 73106

Nathan, M.F.
Crawford & Russell
17 AmpliaPl.
Stamford, CT 06904

Nechvatal, Michael
Illinois EPA
2200 Churchill Rd.
Springfield, IL 62706

Nelson, Nancy Ann
Matrecon, Inc.
P.O. Box 24075
Oakland, CA 94608

Newhof, Thomas
Prein & Newhof
3000 E. Beltline NE
Grand Rapids, MI 49505

Newland, Jim
Minnesota Waste Mgmt. Board
123 Thorson Community Center
7323 58th Ave. North
Crystal, MN 55428

Ney, Ronald E. Jr.
EPA OSW
401 M Street, SW
Washington, DC 20460

Neyer, William
NLS
P.O. Box 39158
Cincinnati, OH 45239

Nielsen, David M.
National Water Well Assoc.
500 W. Wilson Bridge Road
Worthington, OH 43085
                                     306

-------
Noel, Melbourne A.
President
M.A. Noel Consulting Inc.
5646 N. Kenneth Ave.
Chicago, IL 60646

Nowick, Henry
Monsanto Company
730 Worcester St.
Springfield, MA 01151

Nunes, Sary
Peoria Disposal Co.
Peoria, IL61604

Nutini, David L.
General Mgr.
RNK Environmental, Inc.
P.O. Box 17325
Covington, KY 17325

Oberacker, Donald A.
Sr. Mechanical Engineer
US EPA
26 W. St. Clair
Cincinnati, OH 45268

O'Bryan, Glenn A.
Regional Eng.
SCA Services, Inc.
P.O. Box 34457
Louisville, KY 40232

O'Connell, Wilbert
Sr. Res. Scientist
Battelle-Columbus
505 King Ave.
Columbus, OH 43201

O'Conner, JohnT.
Univ. of Missouri-Columbia
Dept. of Civil Engineering
Room 1047 Engineering Bldg.
Columbia, OH 45211

O'Donnell, Francis R.
Oak  Ridge National Lab
P.O. Box X
Oak  Ridge, TN 37830

Ogle, Gilbert
Sr. Staff Engineer
TRW, Inc.
8301 Greensboro Dr.
McLean, VA 22102
Ogren, Curtis W.
Chemical Waste Mgmt., Inc.
Louisville, KY 40204

Olexsey, Robert A.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Ombalski, Stephen D. - Pres.
Ombalski Consulting
Engineers, Inc.
166 West End Avenue
Somerville, NJ 08876

Oppelt, E. Timothy
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Osborne, J. Michael
Three M Company
P.O. Box 33331
Bldg. 21-2W
St. Paul, MN 55133

Osheka, J.W.
Environmental Engineer
PPG Industries, Inc.
One Gateway Center
Pittsburgh, PA 15222

Ostergren, Mark
Business Analyst
Babcock&WilcoxCo.
P.O. Box 2423
N. Canton, OH 44720

O'Sullivan, Colleen
Environmental Specialist
Hillsboro EP Comm.
1900 9th Ave.
Tampa, FL 33605

Otermat, A.L.
Res.  Chemist
Shell Dev. Co.
P.O. Box 1380
Houston, TX 77001

Otte, Les
US EPA
401 M Street, SW
Washington, DC 20460
                                   307

-------
Padden, Thomas J.
US EPA
401 M Street, SW
Washington, DC 20460

Pahren, Herbert R.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Palmer, Charlene R.
H.C. Nutting Company
4120 Airport Road
Cincinnati, OH 45226

Park, James E.
University of Cincinnati
Cincinnati, OH 45221

Parker, Beth L.
Duke University
Durham, NC 27706

Pastene, A. James
Env. Eng.
Union Carbide Corp.
P.O. Box 8361
S.Charleston, WV25064

Paul, Linda S.
NUS Corp.
Park West Two
Cliff Mine Road
Pittsburgh, PA 15275

Payme, John L.
Branch Manager
Soil & Material Engineers
11325 Reed Hartman Hwy.
Cincinnati, OH 45241

Peacock, James A. -  Mgr.
Evans Products Co., Paint Div.
P.O. Box 4098
1516 Cleveland Ave.  SW
Roanoke, VA 24016

Pendleton, Kenneth A.
K.A.  Pendleton Company
10760 Thorn view Drive
Cincinnati, OH 45241
Pennefill, Roger A.
US Nuclear Regulatory
Commission
Mailstop 623-SS
Washington, DC 20555

dePercin, Paul R.
Chem. Eng.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Perry, William H.
NIOSH
4676 Columbia Pkwy.
Cincinnati, OH 45226

Perona, Louis J.
County of LaSalle
707 Etna Road
Ohawa, IL 61350

Person, LeRoy S.
US Nuclear Regulatory
Commission
Washington, DC 20555

Peters, James A.
Monsanto Res. Corp.
1515 Nicholas Road
Dayton, OH  45418

Peters, Nathaniel
University of Kentucky
Lexington, KY 40506

Peters, Wendell
Senior Research Engineer
Southwest Research Institute
P.O. Drawer 28510
San Antonio, TX 78284

Pettyjohn, Wayne A.
Oklahoma State University
Geology Department
Stillwater, OK 74078

Pickering, Edward W.
University of Massachusetts
Amherst
47 Holyoke St.
Northampton, MA 01060
                                    308

-------
Pierce, George E.
Sen. Res. Microbiologist
Battelle Mem. Inst.
505 King Ave.
Columbus, OH 43201

Piontek, Keith
Univ. of Missouri-Columbia
1613 Wilson
Columbia, MO 65201

Pitz, Edward
E-Three, Inc.
P.O. Box 155
Getzville, NY  14068

Poff, Timothy A.
NLO, Inc.
P.O. Box 39158
Cincinnati, OH 45239

Pohland, Dr. Fred
Georgia Institute of
Technology
Atlanta, GA 30332

Polakovic, Nolton
The Bureau of Air Pollution
Control
P.O. BoxCN-027
Trenton, NJ 08625

Polcyn, Andrew
Environmental Science
and Engineering, Inc.
11665 Lilburn Park Road
St.  Louis,  MO 63146

Potzman,  Dennis W.
Wyo-Ben, Inc.
1242 N. 28th St.
P.O. Box 1979
Billings, MT 59101

Powell, Bruce
GMC-Harrison
P.O. Box 824
Dayton, OH 45401

Price, Richard H.
Hess Env. Services
Memphis, TN 38134
Prohaska, John W.
Pedco Environmental, Inc.
11499 Chester Road
Cincinnati, OH 45246

Puch, A.B.
Atlantic Richfield Pet. Prod.
400 E. Sibley
Harvey, IL 60426

Pufford, Bob
Minnesota Waste Mgmt. Brd.
123 Thorson Community Cntr.
7323 58th Ave. North
Crystal, MN 55428

Quehee, Shane
University of Cincinnati
Cincinnati, OH 45267

Raines, J. Walter
E.I. du Pont de Nemours & Co.
100 W. lOSt.
Montchanin 5625
Wilmington, DE 19898

Randolph, Brain W.
Research Asst.
University of Cincinnati
Cincinnati, OH 45221

Ransom, Randall
Environmental Specialist
Dow Corning Corp.
3901 S. Saginaw,
Midland, MI 48640

Rawe, James M.
Dept. of Civil Engineering
University of Cincinnati
Cincinnati, OH 45221

Redding, Peter M.
Vanderbilt University
Box 6304-B
Nashville, TN 37235

Reed, JohnC.
Illinois EPA
2200 Churchill Road
Springfield, IL 62706
                                     309

-------
Rehme, Elmer W.
Ohio EPA
7 E. 4th St.
Dayton, OH 45402

Reible, Dan
Asst. Prof, of Chem. Engr.
Louisiana St. Univ.
Baton Rouge, LA 70803

Reich, Andrew R.
University of Alabama
72020th St., S
Birmingham, AL 35294
Riley, Boyd T.
RyconInc.
690 Clinton Springs
Cincinnati, OH 45229

Roberto, Gerard
University of Cincinnati
Cincinnati, OH 45204

Roberts, Susan A.
Geologist
Malcolm Pirnie, Inc.
2 Corporate Park Dr.
White Plains, NY 10602
Reshkin, Mark
Assoc. Dir. for Envir. Res.
Indiana University, NW
3400 Broadway
Gary, IN 46408

Reuter, Steve P.
Engineer
Indiana State Board of Health
1330 West Michigan St.
Indianapolis, IN 46206

Rich, Charles A.
C.A. Rich Consultant
708 Glen Cove  Ave.
Glen Head, NY 11545

Richardson,  Jean
Jean Richardson & Assoc.
2709 S. 20th St.
Birmingham, AL 35209

Rickabaugh, Janet I.
University of Cincinnati
Cincinnati, OH 45221

Rickman, Bill
Manager
G.A. Technologies
P.O. Box 85608
San Diego, CA 92138

Riggenbach, Jack D.
ERM-Inc.
P.O. Box 357
West Chester, PA 19380
Robine, Deith
National Audubon Society
950 Third Avenue
New York, NY 10022

Robinson, Shird
SCA Services, Inc.
2105 Outer Loop Road
Louisville, KY 40219

Roetzer, James F.
Envirosphere Company
Two World Trade Center
New York, NY 10048

Rogers, Charles J.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Rohrer, William L.
Pope-Reid Assoc., Inc.
245 E. 6th St.
St. Paul, MN 55101

Rollins, Dixon
Sr. Engineer
New York Dept. of Envir. Con.
6274 E. Avon Lima Road
Avon, NY 14414

Ronayne, Michael
University of Kentucky
Lexington, KY 40506
                                    310

-------
Rosebrook, Dr. Donald D.
Institute Scientist
Gulf South Research Inst.
P.O. Box 14787
Baton Rouge, LA 70898

Ross, Richard D.
Trofe Incineration Inc.
PikeRd.
Mt. Laurel, NY 08054

Ross, Robert W. II
US EPA Combustion Res. Facility
NCTR Building #45
Jefferson, AR 72079

Rothenstein, Cliff L.
US EPA
401 M Street, SW
Washington, DC 20460

Roulier, MikeH.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Rowe, Walter
General Portland Company
P.O. Box 324
Dallas, TX 75240

Rubey, Wayne S.
University of Dayton
300 College Park Ave.
Dayton, OH 45469

Ruby, Mike - Asst. Prof.
Dept. of Civil Engineering
University of Cincinnati
MailLoc. #71
Cincinnati, OH 45221

Ruggles, Archie Jr.
Project Engineer
Mason & Hanger Company
P.O. Box 30020
Amarillo, TX79177

Rupa, Edward
JRB Associates
8400 Westpark Drive
McLean, VA 22102
Russell, Charles
Russell & Associates, Inc.
2387 W.Monroe St.
Springfield, IL 62704

Russell, Dwight
Texas Dept. of Water Res.
P.O. Box 13087
Capitol Station
Austin, TX 78711

Salloum,  John D.
Environmental Protection Service
Ottawa, Ontario
Canada

Samkow, Willard
Hilton Davis Chem. Co.
Cincinnati, OH 45222

Sampayo, Felix F.
Johnes & Henry Engineers Ltd.
Toledo, OH 45606

Sanning, Donald E.
US EPA, SHWRD
26 W. St. Clair
Cincinnati, OH 45268

Santoro, David S.
EA Engineering/Ecological
Analysts, Inc.
100 TechneCenter Drive, Ste. 212
Milford, OH45150

Santoro, Joseph D.
EA Engineering/Ecological
Analysts, Inc.
100 TechneCenter Drive
Milford, OH 45150

Sappington, Douglas L.
Consultant
3944 windgap Avenue
Pittsburgh, PA 15204

Sargent, T.N.
Engineering-Science
Suite 590
57 Exec. ParkS.
Atlanta, GA 30329
                                      311

-------
Sarro, William F.
US EPA, Region I
JFK Federal Bldg.
Boston, MA 02203

Sauer, Richard E.
US Ecology, Inc.
9200 ShelbyvilleRd.
Louisville, KY 40222

Sauter, S. Jay
Orange Cnty. Solid Waste Dept.
P.O. Box 14413
Orlando, FL 32857-4413

Saw, ChinC.
Int'l. Paper Co.
P.O. Box 797
Tuxedo Park, NY 10987

Sawdey, Norman J.
CMC, Harrison Radiator
Dept. 507
P.O. Box 824
Dayton, OH 45402

Sawyer, Charles J.
Syntex, Inc.
3401 HillviewAve.
Palo Alto, CA 94304

Schaefer, Betty M.
Senior Chemist
Wilson Nolan, Inc.
2809 NW Expressway, Ste. 290
Oklahoma City, OK 73112

Schaefer, Phillip T.
Zimpro, Inc.
Military Rd.
Rothschild, WI 54474

Scheben, Jackie A.
Tech. Sales Rep.
Cecos International
4879 Spring Grove Ave.
Cincinnati, OH 45232

Schmidt, Edgar H.
Ontario Waste Management Corp.
2BloorSt. W., llth Floor
Toronto, Ontario
Canada M4W3E2
Schmidt, Richard K.
Gundle Lining Systems, Inc.
1340E. RicheyRd.
Houston, TX 77073

Schneider, Carl A.
Vanderbilt Univ.
Box 2231 StationB.
Nashville, TN 37235

Schoenbeck, Melvin
El Dupont
Elastomers Lab
Chestnut Run
Wilmington, DE 19898

Schomaker, Norbert B.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Schraub, Tony
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Schreiber,  Dale E.
Poe & Assoc. of Tulsa, Inc.
10820 E. 45th St.
Tulsa, OK 74145

Schroeder, Paul R.
USAE Waterways Experiment
Station
P.O. Box 631 SESEE
Vicksburg, MS 39180

Schroy, Jerry M.
Monsanto Co.
SOON. Lindbergh Blvd.
St. Louis, MO 63011

Schuller, Rudolph M.
SMC Martin, Inc.
900 W. Valley Forge Rd.
P.O. Box 859
Valley Forge, PA 19482

Schumann, Charles E.
Southwestern Ohio Air
Pollution Control Agency
2400 Beekman St.
Cincinnati, OH 45214
                                   312

-------
Schwartz, William G.
Hazardous Waste Landfill
1113 W. Swords
Peoria, IL61604

Seeker, W. Randall
Energy & Envir. Res.
18 Mason
Irvine, CA 92714

Sehgal, S.B.
Geotechnical & Materials
Consultants, Inc.
1341 Goldsmith
Plymouth, MI 48170

Servis, David B.
Procter-Davis-Ry Engineers
800 Corporate Drive
Lexington, KY 40503

Sferra, Pasquale R.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Shack, Pete A.
Phoenix Environmental
Cons. Inc.
P.O. Box 121555
Nashville, TN 37212

Shafer, Joseph D.
Indiana State Board of Health
1330 W.Michigan
Indianapolis, IN 46206

Shaub, Walter M.
National Bureau of Standards
Washington, DC 20234

Shelley, Philip E.
EG&GWASC, Inc.
2150 Fields Rd.
Rockville, MD 20850

Shen, Almon M.
Shell Environmental
Group, Inc.
4930N.Penn.St.
Indianapolis, IN 46205
Sherman, J.S.
Radian Corp.
8500 Shoal Creek Blvd.
Austin, TX 78758

Shugart, Steven L.
Mayes, Sudderth & Etheredge
1785 the Exchange
Atlanta, GA 30339

Shultz, DaveW.
Southwest Research Institute
P.O. Drawer 28510
San Antonio, TX 78284

Shuster Ken
EPA (WH 565 E)
401 M Street, SW
Washington, DC 20460

Simms, BenM.
Mason & Hanger Co.
P.O. Box 30020
Amarillo, TX 79177

Sims, Ronald C.
Utah Water Research Lab
Utah State University
Logan, UT 84322

Singh, Rajiv
EarthTech, Inc.
6655 Amberton Drive
Baltimore, MD 21227

Skinner, Donna I.
Regional Hydrogeologist
P.A.D.E.R.
1012 Water St.
Meadville, PA 16335

Skinner, Peter N.
NYS Attorney General
Rm. 239
Justice Bldg.
Albany, NY 12224

Smalley, Carolyn J.
Day & Zimmermann, Inc.
Kansas Army Ammunition Plant
Parsons, KS 67357
                                     313

-------
Smith, Garrett A.
US EPA Region II
26 Federal Plaza
New York, NY 10278

Smith, John M. - Pres.
J.M. Smith Consulting
Engineers
7373 Beechmont Ave.
Cincinnati, OH 45230

Snow, Brad L.
Engineering Services Div.
Kerr-McGee Corp.
P.O. Box 25861
Oklahoma City, OK 73125

Socash, Stephen M.
Dept. of Environmental Res.
1012 Water St.
Meadville, PA  16335

Sokol, John Z.
Clyde E. Williams & Assoc.
1843 Commerce Dr.
South Bend, IN 46628

Speed, Nicholas A.
Brown & Caldwell
1501 N. Broadway
Walnut Creek,  CA 94596

Spigolon, S.J.
Memphis State University
Dept. of Civil Engineering
Memphis, TN 38152

Spooner, Philip
JRB Associates
8400 Westport  Drive
McLeon, VA22102

Springer, Charles
University of Arkansas
Dept. of Chemical Engineering
Fayetteviile, AR 72701

Staley, Laurel
US EPA
26 W.St. Clair
Cincinnati, OH 45268
Steffens, Chuck
Caterpillar Tractor Co.
East Peoria Plant
Bldg. KK-1
East Peoria, IL 61611

Stephan, David G.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Sterling, Harry J.
Dept.  of Civil Engineering
University of Kentucky
Lexington, KY 40506

Stern, David A.
GCA/Technology Division
213 Burlington Rd.
Bedford, MA 01730

Stidham, Leslie N. - SEC Rep.
American Thermoplastics Corp.
1235 Kress St.
Houston, TX 77020

Stoddart, Terry L.
US Air Force/Environics Lab
5724 IvyRd.
Panama City, FL 32404

Strachan, William M.
OHIO EPA
7 East 4th St.
Dayton, OH 45402-2086

Stutsman, Mark J.
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Sugarman, Peter
N.J. Dept. of Envir.
Protection Agency
P.O.BoxCN-029
Trenton, NJ 08625

Suhrer, F.C.
US EPA
215 Fremont
San Francisco, CA 94105
                                     314

-------
Sullivan, Daniel E.
Roy F. Weston, Inc.
Weston Way
West Chester, PA 19380

Swartzbaugh, Joseph T.
General Manager
Systech Corp.
245 N. Valley Rd.
Xenia, OH 45385

Talak, Anthony
PADER
1012 Water St.
Meadville, PA 16335

Talty, John
Engineer Director
NIOSH
Cincinnati, OH 45226

Tamplin, Judy
GA Envir. Protect. Div.
270 Washington St., SW
Atlanta, GA 30334

Tang, Harry K.
Planning Research Corp.
Kennedy Space Center
(PRC-1217)
Orlando, FL 32899

Tansill, S.P.
R. Stuart Royer & Assoc., Inc.
P.O. Box 8687
Richmond, VA 23226

Taylor, David R.
S-CUBED
P.O. Box 1620
LaJolla, CA 92129

Thompson, Joe D.
EG&G
P.O. Box 1625
Idaho Falls, ID 83401

Thompson, Steve R.
Divisional Vice President
Browning Ferris Ind.
P.O. Box 3151
Houston, TX 77071
Thrasher, Stephen
Bowser-Morner, Inc.
420 Davis Ave.
Dayton, OH 45403

Tibbetts, Stephen
Union Chemical Co., Inc.
P.O. Box 432
Union, ME 04862

Tite, Joseph L.
Consulting Engr.
Joseph Tite Co.
P.O. Box 366
Michigan City, IN 46360

Tong, Peter
US EPA Region V
230 S. Dearborn
Chicago, IL 60604

Totts, David
Sr. Envir. Specialist
NY EPA
120Rt. 156
Yardville, NJ 08620

Trapp,John H.
MSD DF Greater Cincinnati
1600GestSt.
Cincinnati, OH 45204

Trembley, J.W.
LAN, Inc.
1500 City West
Houston, TX 77042

Trenholm, Andrew
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO 64110

Triegel, EllyK.
Woodward-Clyde Consultants
5120 Butler Pike
Plymouth Meeting, PA 19462

Tsai, Kuo-Chun
Asst. Prof.
University of Louisville
Dept. of Chem. & Env. Eng.
Louisville, KY 40292
                                     315

-------
Tsang, Wing
Research Chemist
National Bureau of Standards
Washington, DC 20234

Tseng, Louis H.
Environmental Elements Corp.
7249 National Drive
Hanover, MD 21076

Turgeon, Marc P
EPAOSW
401 M Street, SW
Washington, DC 20460

Tussey,  Robert C.
Kenvirons, Inc.
P.O. Drawer V
Frankfort, KY 40602

Twilley, Clinton
RETECH Associates, Inc.
861 Corporate Dr.
Suite 200
Lexington, KY 40503

Tyler, Scott
Battelle PNL
Box 999
Richland, WA 99352

Tyndall, M. Frank
Howard Needles Tammen
& Bergendof f
P.O. Box 68567
Indianapolis, IN 46268

Tyro, Michael J.
General Motors Corp.
EASBldg.,GMTechCtr.
Warren, MI 48090

Ullrich, ArlieJ.
Consultant
Eli Lilly & Co.
307E.McCartySt.
Indianapolis, IN 46285

Uhl, Michael E.
Peake Operating Co.
Charleston National Plaza #423
Charleston, WV 25301
Underwood, Edward R.
US Ecology, Inc.
9200 Shelbyville Road
Louisville, KY 40222

Vanderveld, Ronald J.
DeTox, Inc.
One Wheaton Center
Suite 1801
Wheaton, IL 60187

Vakili, Hassan
Virginia Dept. of Health
109 Governor St.
Richmond, VA 23219

VanderMeulen, Joseph
Legislative Service Bureau
Michigan State Legislature
Allegan St., Farnum Bldg.
Lansing, MI 48913

Veith, Jim E. - Sr. Eng.
Soil & Material Engineers
11325 Reed Hartman Hwy.
Suite 134
Cincinnati, OH 45241

Velez, Victor G.
EBASCO Services, Inc.
2 World Trade Center
New York, NY 10463

Velie, Margaret M.
US EPA
401 M Street, SW
Washington, DC 20460

Velzy, Charles O.
Charles R. Velzy Assoc., Inc.
355 Main St.
Armonk, NY 10504

Vidmar, Kevin P.
Vanderbilt Univ.
Box 6304 B
Nashville, TN 37235

Vogt, W. Gregory
Staff Scientist
SCS Engineers
211 Grandview Drive
Covington, KY 41017
                                    316

-------
Volk, David R.
PADER
850KossmanBldg.
Pittsburgh, PA 15222

Vollstedt, Thomas
Zimpro, Inc.
Military Rd.
Rothschild, WI 54474

Wagner, Douglas H.
V.P. Operations
Solidtek
Box 888
Morrow, GA 30260

Walker, E.G.
Burns and Roe
650 Winters Ave.
Paramus, NJ 07652

Walls, James T.
Hamilton Co. Farm Bureau
2870 Markbreit Ave.
Cincinnati, OH 45209

Walsh, James
SLS Engineers
211 Grand view Dr.
Covington, KY 41017

Watkin, Andrew T.
Vanderbilt University
3102-B Wellington Ave.
Nashville, TN 37212

Watkins, David R.
Physical Scientist
US EPA
26 W. St. Clair
Cincinnati, OH 45268

Walz, Arthur
US Army Corps of Engineers
20 Mass, NW
Washington, DC 20314

Webb, George C.
Geotechnical Engineer
H.C. Nutting Co.
4120 Airport Rd.
Cincinnati, OH 45226
Webb, Thomas E.
Amer. Elec. Power Svc. Corp.
P.O. Box 487
Canton, OH 44708

Weinberger, Lawrence P.
The Aerospace Corporation
955 L'Enfant Plaza, SW
Suite 4000
Washington, DC 20024

Weishaar, Michael F.
Monsanto
800 N.Lindbergh
St. Louis, MO 63167

Weiss, Albert
Weiss Pollution Control
41001 Grand River
P.O. Box 505
Novi, MI 48050

Werner, James
Environmental Law Institute
1346 Connecticut Ave., NW
Washington, DC 20036

Werner, Steven I.
Occidental Chemical Corp.
360 Rainbow Blvd. S.
Niagara Falls, NY 14302

Westbrook, Clifton W.
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709

Westfall, Brain A.
US EPA
Industrial Envir. Res. Lab
26 W. St. Clair
Cincinnati, OH 45268

Wetzel,  David
Asst. Prof, of Chem. Eng.
Louisiana State Univ.
Baton Rouge, LA 70803

Wetzel,  Roger
JRB Associates
8400 Westpark Dr.
McLean, VA 22102
                                      317

-------
Whittle, George P.
University of Alabama
Dept. of Civil Engineering
P.O. Box 1468
University, AL 35486

Whitmore, Frank
Versar, Inc.
US EPA Combustion Research Facility
Jefferson, AR 72079

Whitney, Richard R.
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Wickline, Bob
Virginia Dept. of Health
109 Governor St.
Richmond, VA 23219

Widmer, Wilber
Dept. of Civil Engineering
University of Connecticut
Box U-37
Storrs, CT 06268

Wiggans, Kenneth E.
US Army, USAEHA
Aberdeen Proving
Ground, MD 21010

Wigh, Richard J.
Regional Services Corp.
3200 Sycamore Ct., #2B
Columbus, IN 47203

Wiles, CarltonC.
USEPAMERL
SHWRD
26W. St. ClairSt.
Cincinnati, OH 45268

Williams, Charles E.
McBride-Ratcliff & Assoc.
8800 Jameel
Suite 190
Houston, TX 77040

Willis, Dudley L. - P.E.
Resources Recovery, Inc.
108 Briar Lane
Newark, DE 19711
Withiam, James L.
D'Appolonia Consulting Eng.
10 Duff Road
Pittsburgh, PA 15235

Withrow, William
111. Pollution Control Board
309 W. Washington
Chicago, IL 60606

Wolbach, C.D.
Acurex Corp.
485 Clyde Avenue
Mountain View, CA 94042

Wolf, Fred L.
Kester Management Services
4274 Miramar Drive
Toledo, OH 43614

Wolfe, Doug
McCoy & McCoy, Inc.
85 E. Noel Ave.
Madisonville, KY 42431

Woodley, Ralph
Burke Rubber Company
2250 S. Tenth St.
San Jose, CA 95112

Worm, Brenda
Hazardous Waste Research Ctr.
CEBA3418
Baton Rouge,  LA

Yaar, A.
C.E. Williams & Assoc.
1843 Commerce Drive
South Bend, IN 46614

Yalcin, Acar
Louisiana State University
Baton Rouge,  LA 70803

Yang, Edward
Environmental Law Institute
1346 Connecticut Ave., NW
Washington, DC 20056

Yare, Bruce S.
Yare and Associates, Inc.
24 S. 77th St.
Belleville,  IL 62223
                                    318

-------
Young, Bob
DELCO MORAINE
1420 Wise. Blvd.
Dayton, OH 45401

Zak, Clarence
Sunbeam Equipment Corp.
180 Mercer Street
Meadville, PA 16335

Zaninelli, Linda
NY EPA
120Rt. 156
Yardville, NJ 08620

Zimmerman, R. Eric
ESCOR, Inc.
1845 Oak St.
Northfield, IL 60093
Zitkovic, John J.
O.H. Materials Co.
P.O. Box 551
Findlay, OH 45840

Zlamal, Frank
Slurry Systems Division
of Thatcher Eng.
7100 Industrial Ave.
Gary, IN 46406

Zralek, Robert L.
Director of Civil Systems
Waste Management Inc.
3003 Butterfield Rd.
Oak Brook, IL 60521

Zykan, James Jr.
President
B.H.S.,Inc.
R.R. l,Box!16-F
Wright City, MO 63390
                                    319
                                                       US GOVERNMENT PRINTING OFFICE 1984 - 759-102/10611

-------