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
------- |