SW-122C.4
Prepublication issue for EPA libraries
and State Solid Waste Management Agencies
DESTROYING CHEMICAL WASTES
IN COMMERCIAL-SCALE INCINERATORS
(Facility Report 5)
This final report (SW-122c.4) describes work performed
for the Federal solid waste management program
under contract no. 68-01-2966
and is reproduced as received from the contractor
Copies will be available from the
National Technical Information Service
U.S. Department of Commerce
Springfield, Virginia 22161
U.S. ENVIRONMENTAL PROTECTION AGENCY
1977 Eir/i^'j • • • '.'on Agency
-------
This report as submitted by the grantee or contractor has been technically
reviewed by the U.S. Environmental Protection Agency (EPA). Publication
does not signify that the contents necessarily reflect the views and
policies of EPA, nor does mention of commercial products constitute
endorsement by the U.S. Government.
An environmental protection publication (SW-122c.4) in the solid waste
management series.
-------
TABLE OF CONTENTS
Page No.
Table of Contents iii
List of Tables v
List of Figures vii
1. Summary 1
2. Introduction 4
3. Process Description 5
3.1 Test Facility 5
3.2 Process Parameters 6
4. Test Description 10
4.1 Waste Tested 10
4.2 Operational Procedure 11
4.3 Sampling Methods 12
4.4 Analysis Techniques 15
4.5 Operating Problems 16
5. Test Results 18
5.1 Operating Conditions and Process Data
for the PVC Waste Destruction Tests 18
5.2 Destruction Efficiency and Composition
of Combustion Zone Effluent Gas 18
5.3 Final Emissions 27
6. Waste Destruction Costs 31
6.1 Capital Investment and Total Operating
Cost Based on the 3M Company Chemolite
Incineration System 31
6.2 Capital Investment and Total Operating Cost
for the Destruction of PVC Waste in a
"Plant Scale" Incineration System 34
iii
-------
TABLE OF CONTENTS (Con't).
APPENDICES
A. Sample Volume Data and Flow Rates
B. Analytical Chemistry Details
C. On-Site Analysis of Vinyl Chloride Monomer
D. Heat Balance
E. Assessment of Environmental Impact of Destroying
Solid and Liquid Wastes at 3M Company Chemolite Plant
iv
-------
LIST OF TABLES
Table No. Title Page No.
1-1 Results Summary 2
3-1 Process Parameters for Waste
Destruction Tests 9
5-1 Operating Conditions and Process Data
For PVC Waste Destruction Tests - 3M
Chemolite Incineration System 19
5-2 Summary of Quantities of Organic Materials
in Hot Zone Effluent Sample Extracts 20
5-3 Hot Zone Emission Rates and Calculated
Destruction Efficiencies 22
5-4 Quantitative Results Obtained From On-
Line Instruments 24
5-5 Organic Chemical Species Found in Hot
Zone Effluent Sample Extracts 26
5-6 Chloride Analyses of Scrubber Water 29
6-1 Capital Investment for the 3M Company
Chemolite Plant Incineration System 32
6-2 Total Operating Cost for Incineration of
PVC Waste Based on the 3M Company Chemolite
Incineration System 35
6-3 Total Operating Cost for Incineration of
PVC Waste Based on a "Plant Scale"Incinerator 36
A-l Stack Sampling Data A-l
A-2 Hot Zone Sampling Data A-2
A-3 Estimated Total Gas Effluent Flow Rates A-3
B-l Volumes of Impinger Solutions B-10
B-2 Gravimetric Data for Probe Wash, Filter,
and Dry Impinger Samples B-ll
-------
LIST OF TABLES (Con't)
B-3 Results of Gravimetric Analyses of
Concentrated Organic Extracts B-12
B-4 Elements Identified in PVC Waste by
SSMS B-14
B-5 Results of GC Analysis of Unconcentrated
Sorbent Trap Extracts B-18
B-6 Results of LRMS Analysis of Concentrated
Pentane Extracts of Sorbent Traps B-20
B-7 Results of LRMS Analysis of Concentrated
Methanol Extracts of Sorbent Traps B-22
VI
-------
LIST OF FIGURES
Figure No. Title Page No.
3-1 Schematic of Rotary Kiln Facility 7
4-1 Hot Zone Sampling Train for PVC Tests 14
B-l Sorbent Trap Extractor B-3
B-2 Calibration Curve Used for GC Analysis
of Unconcentrated Extracts B-7
C-l Example Chromatograms for on-Site GC
Analysis of Vinyl Chloride Monomer C-3
D-l Combustion Zone Heat Balance, November
4, 1976 Test Run D-2
D-2 Combustion Zone Heat Balance, November
5, 1976 Test Run D-3
D-3 Combustion Zone Heat Balance, November
6, 1976 Test Run D-4
D-4 Combustion Zone Heat Balance, November
8, 1976 Test Run D-5
-------
FOREWORD
The tests described in this report are part of a program designed
to evaluate the environmental, technical, and economic feasibility of
disposing of industrial wastes via incineration. This objective is being
pursued through a series of test burns conducted at commercial incinerators
and with real-world industrial wastes. Approximately eight incineration
facilities and seventeen different industrial wastes have been tested under
this program. The incineration facilities were selected to represent the
various design categories which appear most promising for industrial
waste disposal. The wastes were selected on the basis of their suitability
for disposal by incineration and their environmental priority.
This report describes the test conducted at the Chemolite Plant of
the 3M Company (Cottage Grove, Minn.), which was the fifth facility of the
series. Facility reports similar to this one have been published for the
previous tests which were conducted at the Marquardt liquid injection
facility in Van Nuys, California, the Surface Combustion pyrolysis facility
in Toledo, Ohio, the Systech fluidized bed facility in Franklin, Ohio,
and the Zimpro, Inc., wet air oxidation facility at Rothschild, Wisconsin.
The facility reports are primarily of an objective nature presenting the
equipment description, waste analysis, operational procedures, sampling
techniques, analytical methods, emission data and cost information.
Facility reports are published as soon as possible after the testing has
been completed at a facility so that the raw data and basic results will
be available to the public quickly.
In addition to the facility reports, a final report will also be
prepared after all testing has been completed. In contrast to the facility
reports which are primarily objective, the final report will provide a
detailed subjective analysis on each test and the overall program.
ACKNOWLEDGEMENT
Arthur D. Little, Inc., is grateful for the cooperation of
3M Company, especially Messrs. Michael Santoro and Gordon Weum, during
these tests at the 3M Company Chemolite Incineration System. Acknowledge-
ment is also made of the extensive and fruitful interactions between ADL
and TRW personnel during the initial phases of this program. The
project is deeply indebted to Messrs. Alfred Lindsey, John Schaum, and
Eugene Grumpier of the Office of Solid Waste, U. S. Environmental
Protection Agency, for their advice and technical direction.
Vlll
-------
1. SUMMARY
The incineration tests were carried out on waste from the production
of polyvinyl chloride (PVC) at 3M Company's Chemolite incinerator system.
The waste was burned at three different conditions to determine the
effects of normal operating variables.
The 3M Company Chemolite incinerator system consists of a rotary
kiln primary chamber, a secondary combustion chamber, and a wet scrubbing
system for air pollution control. The 3M Company facility has a rated
capacity of 23 million Kcal/hr (90 million Btu/hr), and is large in
comparison with most industrial waste incinerators.
The PVC waste material tested at 3M Company was selected for
this program because of the possible hazard associated with the
presence of residual vinyl chloride monomer. The waste was found to
have a solids content of 28%, with the remainder of the mass due
to water. The solid material was primarily polyyinyl chloride. The
waste was found to contain 220 ppm of residual vinyl chloride monomer
on a wet weight basis. Only very small quantities of other organic
species were found. No trace elements were found at concentrations
high enough to cause concern for emissions of toxic metals at the
feed rates used in the tests.
The results of these tests are summarized in Table 1-1. Waste
destruction efficiencies, as measured by total quantities of chlori-
nated organics in the effluent, were estimated as greater than
99.995 percent in each of the three tests. The total destruction
efficiencies were 99.80 to 99.86 percent. The data suggest that
destruction was marginally less efficient when combustion zone gas
residence time was reduced from three seconds (3M2 and 3M3) to two
seconds (3M4).
No vinyl chloride monomer was detected in the incinerator
combustion zone effluent; the limit of detection was 0.2 ppm (v/v).
The predominant organic species identified in the combustion zone
effluents were polynuclear aromatic hydrocarbons at estimated con-
centrations of 2-5 mg/cu m. These species were found in the back-
ground (fuel oil) test as well as in the waste test samples, and are
therefore not attributed uniquely to combustion of the PVC waste.
Some members of the general class of polynuclear aromatic hydro-
carbons are known to have high toxicity (e.g. benzo[a]pyrene, a
carcinogen); those particular species were not the predominant poly-
nuclear aromatics in these test samples.
-------
TABLE 1-1
SUMMARY OF RESULTS
Background
Operating Conditions
Primary Combustion Zone
Temperature, °C (°F)
Secondary Combustion Zone
Temperature, °C (°F)
Combustion Zone Gas
Residence time, sec
Waste Feed Rate,
Metric tons/hour
Quality of Combustion Gas
Particulate, mg/m
Total Organics, mg/m
*
Chlorinated Organics,
mg/m3
Polynuclear Aromatic,
Hydrocarbons, mg/m
Hydrochloric Acid,
mg/m3
Quality of Stack Gas
Particulate, mg/m
Hydrochloric Acid,
mg/m3
Quality of Scrubber Water
Total Organics, mg/s,
Hydrochloric Acid, mg/s.
Destruction Efficiency
Total Organics, percent
Chlorinated Organics,
percent
3M1 (B)
870
(1600)
1090
(2000)
3
122
48
—
5
—
40.2
0.65
2.4
15
99.84
3M2
870
(1600)
1090
(2000)
3
0.845
427
38
< 0.02
2.3
1660
**
**
5.6
414
99.88
99.99g
3M3
870
(1600)
980
(1800)
3
0.845
357
39
< 0.02
2.5
1660
70.9
14.5
5.0
719
99.86
99.99g
3M4
870
(1600)
980
(1800)
2
0.845
378
60
< 0.04
4.5
1485
71.4
10.3
0
816
99.80
99.996
Detailed analyses were done only on the 3M1(B) and 3M3 test samples. It
has been assumed that the sane relative abundances of chlorinated organics
and polynuclear aromatics as found for 3M3 also hold for the 3M2 and 3M4
waste tests.
**
No stack sample was acquired during this test.
-------
Concentrations of hydrochloric acid in the combustion zone effluent
were estimated at about 1.5 g/cu m. The wet scrubber was found to be
> 99 percent efficient at removal of the hydrochloric acid. Stack
emissions of hydrochloric acid were < 15 mg/cu m. Total particulate
emissions from the stack were < 75 mg/cu m.
Capital and operating cost estimates were prepared for a system
of the size tested at 3M Company and for a much smaller system, i.e.
one more nearly matching the requirements of an individual PVC manu-
facturing facility.
The results of these estimates are shown below.
Incinerator Capacity (Metric Tons/Yr) 6696 335
Estimated Capital Investment ($) 7,800,000 3,900,000
Estimated Operating Costs
($/Metric Ton) 582 1767
-------
2. INTRODUCTION
The U.S. Environmental Protection Agency has sponsored a program
to evaluate the effectiveness of a variety of types of commercial
thermal destruction facilities for destroying chemical wastes. Rotary
kiln incinerator was included among the types of processes to be
evaluated because it represents a well established technology with
widespread applicability to a variety of hazardous wastes.
The 3M Company's Chemolite incineration system, located in Cottage
Grove, Minnesota, serves as a central waste disposal facility for com-
bustible liquid and semi-liquid chemical wastes from 3M Company manu-
facturing operations. The incineration system includes: the material
handling building, the rotary kiln primary combustion chamber, secon-
dary combustion chamber, air pollution control train (wet scrubber),
fan, stack, and scrubber water neutralization system. The facility
has a rated capacity of 23 million Kcal/hr (90 million Btu/hr) and is
large in comparison with most industrial waste incinerators.
The waste selected for testing at the 3M Company's Chemolite
incineration system was a stream generated in the production of
polyvinyl chloride (PVC). The estimated total annual production of
PVC waste similar to that selected for testing is on the order of
2300 metric tons. A waste of this type was selected for testing in
the program because of the possible hazard associated with the
presence of residual vinyl chloride monomer. The waste was matched
with a rotary kiln incinerator because this type of facility
can readily handle wastes with high solids content.
The sections of this report which follow describe in detail
the incineration facility and process (Section 3), the waste destroyed
and the sampling and analysis procedures used (Section 4). The
results of the tests are presented and discussed in Section 5.
Estimates of the capital and operating costs for destroying PVC wastes
in the type of incinerator tested are included in Section 6.
* Contract No. 68-01-2966
-------
3. PROCESS DESCRIPTION
The 3-M Company Chemoiite incineration system, located at its Chemolite
plant in Cottage Grove, Minnessota, serves as a central waste disposal
facility for combustible wastes from various 3-M manufacturing operations.
The incineration system was designed to be an efficient and environmentally
acceptable alternative to the land disposal of waste solvents, oils,
tars, selected sludges, and other materials.
3.1 TEST FACILITY
The 3-M Company Chemolite incineration system is divided into two
main sections, (1) the material handling building, and (2) the actual
incinerator and related equipment.
The materials handling building is designed to receive, store, and
prepare waste materials for incineration. It is equipped with a tank
truck unloading facility and five 10,000 gallon storage tanks for the
handling of bulk liquid wastes. Most of the waste material, however,
is received in 55-gallon drums. All drums are labelled as to the type
of material contained, its flammability, hazard potential, and other
characteristics pertinent to both incinerator performance and operational
safety. The drums are carefully sorted in order to maintain control
over the composition of material fed to the incinerator.
The material handling building is equipped with a specially designed
semiautomatic drum feed system which, depending on the nature of the
waste, can either empty the contents of the drum into the rotary kiln
incinerator or feed both the drum and its contents directly into the
incinerator. Generally, drums containing the more hazardous types of
materials are fed directly into the incinerator, thus assuring the complete
destruction of all material contained in the drum. The drums are fed
at preset intervals, the duration of which depends on the nature of the
waste and the desired rate of material throughput. The bulk liquid waste
(designated as "pumpable waste") is fed into the incinerator through
conventional burners with the aid of atomizing steam.
The incinerator itself consists of the following major components:
1. Rotary Kiln Primary Combustion Chamber -
3.96m (13 ft) in diameter x 10.67m (35 ft) long
2. Secondary Combustion Chamber -
7.9m x 3.05m x 3.05 m [26 ft x 10 ft x 10 ft]
3. Air Pollution Control Train - Quench elbow, quench
chamber, high energy venturi scrubber, demister.
4. Fan - 500 horsepower.
-------
5. Stack - 61m (200 ft) high x 1.53 m (5 ft) diameter
(with sampling point located 24.4m (80 ft) from the
bottom).
6. Scrubber Water Neutralization System.
The rotary kiln is fired with liquid waste which is fed through
a burner at the front of the kiln. The drummed (non-pumpable) wastes
are fed into the rotary kiln by the drum feed system previously described.
The rotary kiln is usually operated at a temperature of 815-870°C (1500-
1600°F). The average solids detention time within the rotary kiln is
roughly 2 hours, although the detention time varies considerably with
both the type of waste and the mode of incinerator operation. The average
detention time of combustion gases within the kiln is approximately 2
seconds. The rotational speed of the kiln is generally set at 0.2-0.3
rpm. At the downstream end of the kiln an ash handling system quenches
and collects ash and spent drums, which are eventually loaded into trucks
and transported a short distance to an on-site landfill.
Gases from the kiln flow through a mixing chamber to the secondary
combustion chamber. The secondary combustion chamber is designed to
burn uncombusted gases and particulate matters from the rotary kiln.
It is fired with pumpable waste and/or #2 fuel oil, depending on the
characteristics of the waste being incinerated. The secondary combustion
chamber is generally operated at 980°C (1800°F), although temperature
of over 1090°C (2000°F) can be achieved.
Gases from the secondary combustion chamber are cooled with water
in the quench elbow and quench chamber and then flow through a high energy
venturi scrubber where particulates are removed. The combustion gases
then proceed through a demister which removes entrained water particles.
From the demister the gases flow through an induced draft fan and are
exhausted to the atmosphere through the stack.
Water streams discharged from the scrubber system, demister, and
ash handling system are combined into a single wastewater stream. The
wastewater tends to be acidic, especially when burning chlorinated hydrocarbons,
and must be neutralized prior to discharge. The wastewater is neutralized
with ammonia and is then sent to a central wastewater treatment system
which accomodates other wastewater streams from the various manufacturing
operations within the Chemolite complex.
A schematic diagram of the 3-M Company Chemolite incineration system
is shown in Figure 3-1.
3.2 PROCESS PARAMETERS
The 3M Company Chemolite incineration system has been designed
primarily for the incineration of materials generally capable of support-
ing combustion, e.g., oils, tars, and waste solvents. Little, if
any, auxiliary fuel is required when incinerating such materials. The
-------
n)
ex
to
0)
4-1
SJ
o
u
•H
O
CO
&
o
•rl
4-1
-------
capacity of the 3-M Company Chemolite facility, based on the incineration
of wastes capable of supporting their own combustion is rated as follows:
• Pumpable Waste - 2270 1/hr (600 gal/hr)
• Non-pumpable Waste - 1.82 metric tons/hr (4000 Ib/hr)
• Heat Rate - 22.7 million KCal/hr (90 million Btu/hr)
The actual waste throughput depends on the composition and heat
content of the waste. Highly combustible wastes can be fed at a rate
nearly equal to the maximum capacity of the incinerator, while marginally
combustible wastes must be fed at a lower rate to accomodate the additional
heat input from the required auxiliary fuel.
The test program was designed to determine the capability of a rotary
kiln incineration system to destroy polyvinyl chloride (PVC) wastes under
different operating conditions. The main variables that were regulated
in the combustion process itself were combustion zone gas residence time and
temperature. The waste feed rate was not a variable in this test, because
it was established by the permitted heat release rate of the incinerator
and the necessity of maintaining specified temperatures. The main process
parameters for the waste destruction tests are shown in Table 3-1.
Since the higher heating value of the PVC wastes was low, the combustion
temperatures during the test were almost entirely maintained by the combustion
of large quantities of #2 fuel oil. The 2 second combustion gas detention
time for the November 8, 1976, test was maintained by increasing the
air flowrate (and hence the fuel consumption) during the test period.
This procedure was necessary since the only other way of reducing
the detention time while maintaining the specified temperatures would
be to alter the geometry of the combustion zones, and this was obviously
impossible.
-------
UJ
ro
_
00 O
a:
ex.
oo
UJ
t_J
o
CtL
o.
10
<-O 3
CT> O- 0)
I* i — O) >
O 3 Q_
Z U-
10 3
CT> Q- O>
« ""-££
S V003
• i— O
> o> >
O 3 CL
sy 1 1
^_^
1/1
VO 3
O*> O. CD
5v * «^ (Q
00 UO O 3
• r- O
> 4) >
O 3 0.
Z U.
10 r-
r~- c
0^ O
oT *"
^""* *» «r~
^! *J" O
CO . ,_
> > (O
Is.
•r- E
L. V
Q- 1—
O^ CM
^_
*^-^
O
CO
Ch
o M
o
CO
r—
'•**'
0
00
(ft
^-^
o
o co
CM
in
o
r—
•^"^
0 CO
o
CM
If)
o
•r*
*~** -4-*
c u- c
O 0 O)
• BP~ "*^J* -*^ *
•r* **•* * ^^
+J 0)
to a) o
3 i-
§-P 10 in
(O tD TJ
OS. C
(U CO
>, Q. o U
S- E -r- (U
•0 W 4-> (^
•OK- «/) —
C 3
O tt» J3 Q)
u c E E
OJ O O •<-
t/> M O 1—
in
CM 00
o — o
06 & £ *
c ^«. o
1—*-* -O Wl -i-
•r- E 0> E J-
<> ^ (y 3 4.)
fc. U_ i. 0)
>,+-* -o e
J»
•M 4->
O (O (O
Q£ OC 3
-------
4. TEST DESCRIPTION
4.1 WASTE TESTED
The waste tested at 3M Company's Chemolite Incinerator
System was a stream generated in the production of polyvinyl chloride
(PVC). The estimated total annual production of PVC waste similar
to that selected for testing was 2300 metric tons (5 million Ibs.)
in 1974. A waste of this type was selected for testing because of
the possible hazard associated with the presence of residual vinyl
chloride monomer.
A survey sample of the PVC waste was received several months
prior to testing. The waste was found to consist mainly of orange-
brown, small, grainy particles that were roughly spherical and tightly
packed. A shallow layer of liquid (which appeared to be water) was
present on the surface. The waste particles were quite densely packed,
but a stirring rod or spatula could penetrate the packed material
quite readily. On standing for several weeks at room temperature, the
waste material blackened. This sample was estimated to have a solids
content of 30-35% and a higher heating value of 1450 Kcal/Kg
(2610 Btu/lb). Elemental analysis and infrared spectroscopy on the
dried solid material indicated that it was predominantly polyvinyl
chloride. The concentration of residual vinyl chloride monomer in
the survey sample was found, by gas chromatographic head space analysis,
to be 190 ppm by wet weight of waste.
The waste material actually incinerated in the 3M Company
Chemolite system was quite similar to the survey sample. Details of
the analyses of the representative waste sample, composited from
eight of the drums of waste burned, are given in Appendix B, Section 3.
The waste was found to be 28% solids. The solid material was found
to be primarily polyvinyl chloride by infrared spectroscopy (IR) and
elemental analysis: Weight Percent
~~C H Cl N S
found for dry waste solids: 38.41 5.00 52.85 0.06 0.22
calc'd for pure PVC: 38.7 4.8 56.5
Destructing Chemical Wastes in Commercial Scale Incinerators,
Technical Summary, Volume I, PB-257-709/6 WP.
10
-------
The representative waste sample was found to contain
220 ± 55 ppm by weight (wet waste basis) of residual vinyl chloride
monomer. Other organic species identified were some aliphatic hydro-
carbons (predominantly unsaturated) and < 0.1 % by weight (wet basis)
of tetrachloroethylene.
The ash content of the waste was found to be 1.7%. No
trace elements were found by spark source mass spectroscopy (SSMS)
at concentrations high enough to cause concern for emissions of toxic
metals at the feed rates used in the tests.Calculated values of maximum
possible stack gas concentrations, assuming no removal by a scrubbing
system, were less than the OSHA standard value for all elements iden-
tified. In practice, incineration of the PVC waste would always
require a pollution control system (wet scrubber) to treat the hydro-
chloric acid emissions.
4.2 OPERATIONAL PROCEDURE
The preliminary detailed operating procedures including
test plan and safety plan were reviewed, prior to testing, by the ADL
sampling team.
The operating procedure for the test burns on PVC waste
(3M2, 3M3, 3M4) was as follows:
1. The feed to the incineration system was switched
over from the chemical waste being incinerated by
3M Company to #2 fuel oil. A two-hour purge period
on #2 fuel oil alone was allowed before feeding the waste.
2. The on-line instruments were activated and test
equipment and sampling trains were checked and
prepared for use.
3. After the purge period, waste feed was begun at
a rate of 6 drums of PVC waste per hour. Feed
rates of #2 fuel oil and air to the kiln were re-
adjusted to give the desired temperature conditions.
4. After the PVC waste had been fed to the incinerator
for at least one hour, the probes were inserted into
the hot duct and stack, and sampling was begun.
5. After the conclusion of sampling at the hot zone
and stack for the day, the incineration system was
switched over to other wastes being incinerated by
3M Company.
11
-------
For the background test, 3M1(B), step 3 was omitted. For that test,
sampling was begun as soon as possible after completion of the two
hour purge on #2 fuel oil.
4.3 SAMPLING METHODS
4.3.1 Sampling to Evaluate Incinerator Performance
Five distinct samples were taken during the waste tests. These were
• Composite sample of waste feed material.
• Sample of combustion zone effluent fed to on-line
instruments for continuous monitoring of test.
0 Grab sample of combustion zone effluent to evaluate
process effectiveness.
• Grab sample of stack gases to verify that test
program was environmentally acceptable.
• Sample of fresh and spent scrubber water.
Because there was no way to clean the ash pit prior to testing, a sample
of ash from the kiln was not acquired. No sample that represented residue
from these tests only could be obtained.
4.3.1.1 Haste Feed Sample
A single composite sample of waste feed was prepared. A
total of about 2 kg of sample was acquired by taking approximately
equal quantities of sample from each of eight drums selected at random.
The samples were in the form of cores, corresponding to the entire
depth of each drum, and were taken from several different points
within each drum.
Since this sampling required that drums of waste and their
plastic liners be opened, the drums were moved to an area outside the 3M
Company material handling building. Personnel from Arthur D. Little, Inc.,
wearing chemical cartridge respirators, sampled the drums from an
upwind position. This procedure was adopted to eliminate the possi-
bility of exposure to vinyl chloride monomer in the head space of
the open drum. (See Appendix C).
12
-------
4.3.1.2 On-line Gas Monitoring
A portion of the combustion effluent was sampled through a
ceramic probe and passed through a heated Teflon® line to a gas condi'
tioning system. The gas conditioner was designed to deliver a cool,
dry, particulate-free sample to the CO, C02, and NO analyzers. A
fraction of the sample was also supplied, untreatea, to the hydro-
carbon analyzer.
The instruments used and their ranges were:
Hydrocarbons Beckman 0.05 ppm - 10%
Model 402
Carbon Monoxide Beckman 2-220 ppm
Model 865
Carbon Dioxide Beckman 0.05 - 20%
Model 864
Oxygen Taylor 0.05 - 100%
OA 273
Nitrogen Oxides Thermo Electron 0.05 ppm - 1%
Model IDA
In addition, portions of this stream were injected at intervals into
a gas chromatograph installed in the sampling trailer for these tests.
This was done to verify that incinerator emissions of vinyl chloride
monomer did not exceed the EPA limit (for vinyl chloride and PVC
plants) of 10 ppm discharge to ambient air. Details of this procedure
are given in Appendix C.
4.3.1.3 Combustion Zone Grab Sample
The train used for collecting this three-hour sample is
shown schematically in Figure 4-1. The principal components in this
comprehensive sampling train were:
• A 12.5 mm (0.5") quartz-lined sampling probe
• A knockout trap consisting of an oversized im-
pinger to condense water
• A quartz fiber filter
*
Trademark of E. I. duPont deNemour and Company
13
-------
£
LU
oc
o
•i o ; .
> o o
1° "".
v: «
a
2
0 0 ° ° <
°°00°? 0
X'V
o ,, o o
^
a.
N
i
BE
(9
E
14
-------
A sorbent trap filled with XAD-Z^resin to
collect orgam'cs of moderate volatility
• Impingers containing aqueous sodium hydroxide
to collect acidic gases
In addition, a portion of the combustion zone effluent
was collected in gas sampling bulbs from the bypass line of the
hydrocarbon analyzer.
4.3.1.4 Stack Gas Grab Sample
The stack gas effluent was sampled isokinetically, according
to the EPA Method 5 procedure, along a twelve-point traverse. ?
The train was a typical EPA Method 5 type, the RAC Staksamplr.®**
The impingers contained aqueous sodium hydroxide to trap hydrochloric
acid. A one-hour sample was collected.
4.3.1.5 Scrubber Water
Samples of spent scrubber water were composited, using a
proportional sampling pump, over a three-hour period coincident with
the combustion zone gas grab sample. Samples of fresh scrubber water
were taken in 1 liter bottles from a convenient tap.
4.3.2 Sampling to Monitor Personnel Exposure
Ambient air samples in 5-liter Tedlarnbags were taken
from areas where drums of waste were stored or handled. The ob-
jective of this sampling was to verify that personnel were not
exposed to concentrations of vinyl chloride monomer in excess of
the OSHA standard value of 1 ppm (for 8 hour time weighted average).
Details are presented in Appendix C. No detectable amounts of
vinyl chloride monomer (i.e. < 0.2 ppm) were found.
4.4 ANALYSIS TECHNIQUES
4.4.1 Extractions and Sample Preparation
General descriptions of the techniques used are in the
* Tracemark of Rohm and Haas Company
/ The use of one traverse, rather than two, was approved by the project officer
because of the severe weather conditions during these tests
**
Tradpmark of Research Appliance Corporation
*** Trademark of E. I. duPont deNemour and Company
15
-------
Phase I Final Report of this contract. A detailed description of
the specific solvents, etc., used for the 3M Company test samples
is given in Appendix B.
4.4.2 Analytical Methods
The techniques which were chosen during Phase I of this
contract for evaluation of the effectiveness of thermal destruction
of industrial wastes were:
Low Resolution Mass Spectrometry (LRMS)
Infrared Spectrometry (IR)
Gas Chromatography
Elemental Analysis
Inorganic Analyses were done by:
X-ray Fluorescence (XRF)
Spark Source Mass Spectrometry (SSMS)
Atomic Absorption Spectroscopy (AAS)
Specific Ion Electrode Methods (SIE)
These techniques were applied to the 3M Company samples where
appropriate. Details of analytical procedures are given in
Appendix B.
4.5 OPERATING PROBLEMS
A stack gas grab sample was not acquired during the first
PVC waste test (3M2), because extra manpower was found to be re-
quired to do the hot zone sampling. This was because the hot zone
sampling train filter plugged 4 times during collection of the
three-hour sample and difficulties were encountered in disassembling
and reassembling the train. For subsequent tests, this problem
was overcome by realignment of the hot zone train monorail support.
The problem of clogging filters was encountered in all
four tests, with two filters each required for 3M1(B) and 3M3 and
four filters each for 3M2 and 3M4. In the 3M4 test, two sorbent
traps were used because the first appeared to be flooded with water
part way through the sampling.
The chemiluminescent NO/NO analyser began to malfunction
during these tests. The instrument's main power fuse would blow
out after 10-30 minutes of operation in the NO mode. However, pre-
A
Destructing Chemical Wastes in Commercial Scale Incinerators,
Technical Summary, Volume I, PB-257 709/6WP,
16
-------
liminary data showed that NO - NO for these incineration tests, so
the test was continued usingxvalues for NO only.
17
-------
5. TEST RESULTS
5.1 OPERATING CONDITIONS AND PROCESS DATA FOR THE PVC WASTE
DESTRUCTION TESTS
Operating conditions and process data for the 4 test runs are
summarized in Table 5-1. More details are presented in Appendix D.
5.2 DESTRUCTION EFFICIENCY AND COMPOSITION OF COMBUSTION ZONE
EFFLUENT GAS
Data on the composition of the combustion zone effluent
were obtained from a variety of samples and types of analyses. The
analyses, which included both quantitative and qualitative character-
izations of the effluent,and the analytical results, are described
in Appendix B. In this section of the report, the analytical results
are presented in a reduced form which facilitates overall evaluation
of the tests.
5.2.1 Quantitative Characterization
Principal criteria for assessing the effectiveness of the
incineration process for treatment of the waste are the calculated
destruction efficiencies (DE'S). Calculation of DEtotal was based
on the total quantity of material found in the orgaftic solvent
extracts of the various hot zone sampling train components. Cal-
culation of DEw5Stp was based on the quantities of chlorinated
organics found in those same extracts. The estimated quantities of
material in the various extracts are presented in Table 5-2.
In estimating the total quantity of organic material collected,
the estimates obtained by gas chromatographic analysis (of the unconcen-
trated extracts) and by gravimetric analysis (of an aliquot of concentrated
extract evaporated to dryness) have been summed. This was done because
other work in the Arthur D. Little, Inc., laboratories has shown that
many species with GC retention times similar to those found in these
samples are lost by evaporation when an organic extract is dried to con-
stant weight. The total quantity of organic material collected in the
train (Table 5-2) was converted to concentration in mg/cu m and added to
the estimated quantity of volatile material obtained from the hydrocarbon
analyzer (Table 5-4). As a result of this conservative method of estima-
tion, the total organic emissions may be slightly overestimated.
In Table 5-3, the organic emission rates are expressed in terms
of mg/standard cu m of hot zone effluent. These values are compared with
the maximum organic loading that could have occurred if no destruction
was accomplished. The latter quantity is simply the feed rate (of waste
plus fuel oil for DEtotal and of waste along for DEwaste). in mg/min, divided
by the total hot zone gaseous effluent flow, in standard cu m/min.
18
-------
TABLE 5-1
OPERATING CONDITIONS AND PROCESS DATA FOR PVC WASTE DESTRUCTION TESTS
3M CHEMOLITE INCINERATION SYSTEM
TEST RUN
• Steady State Test Period
• Primary Combustion Zone
Temperature (°F)
• Secondary Combustion
Zone
« Combustion Zone Gas
Detention Time (sec)
» Rotary Kiln Rotation
Rate (rpm)
•' Waste Feed Rate
(drums/hr)
(metric tons/hr)
• Average Fuel Oil Feed
Rate
(gal /mi n)
(MM Btu/hr)^
• Unit Fuel Energy
Requirement (Btu/kg of
waste)
• Air Feed Rate
(scfm)
• Percent Excess Air (%)
• Wastewater Flowrate
(gpm)
Nov. 4, 1976
(Fuel Oil Only)
9:30 AM-3:30 PM
1,600
2,000
3
0.2
-
-
10.3
86.89
28,100
110
1,180
Nov. 5, 1976
(Fuel Oil Plus
PVC Waste)
9:30 AM-3:00 PM
1,600
2,000
3
0.2
6
0.845
10.8
91.10
107,800
ipprox. (28, 000-
29,000)
ipprox. (100-110)
1,180
Nov. 6, 1976
(Fuel Oil Plus
PVC Waste)
9:30 AM-3-.00 PM
1,600
1,800
3
0.2
6
0.845
9.3
78.45
92,800
30,600
140
1,180
Nov. 8, 1976
(Fuel Oil Plus
PVC Waste)
9:30 AM-2:30 PM
1,600
1,800
2
0.2
6
0.845
11.8
99.54
117,800
33,800
75
1,180
(1) Based on #2 fuel oil with a higher heating value of 140,000 Btu/gal
(2) Calculated from stack gas flowrates in conjunction with fuel and waste feed rates
(3) Scrubber water plus cooling water less evaporative losses
19
-------
co
i
UJ
_l
Q.
CO
CO
O
•r"
C
to
o>
c c c c
•a
•
c
*
*
-0-0 t3
• • co •
c c c
CM
1
in
LU
|
co
^f
t—
LU
LU
O
M
H-
0
Z
t— i
co
<
Q>
LU
H-
O
i— i
•^
3
a:
o
u_
o
co
LU
h— |
1—
cr
E
n
•*->
a>
•p
c
0
o
c
tO
CT
S-
f— .
ro
O
1—
-o
OJ
4^
(O
E
•^
4^
CO
LU
O
CD
4-
°
co
>>
03-
r—
•P
0
1—
)>x
OQ
p..
(0
0
1—
•
>
fO
s_
"O
c
fO
1C
•
>
S-
CD
O
CJ3
VD r— iO O
CM CO O CO
r*- i— CM co
IO O CO CO
CM ro en ^
iv. i— r- CM
r- CO CM
^3 • • •
O O 00
V
co o r>- ^- LO
en co ^- LO CM
co r*% co
CM O (£> VO 00
co r-> CM
r—
i— r— CO f*.
• f*^ • • •
cn co co r*"*
V
10
in
P»
r—
O
1^
o
r™
VO
•
00
o> i— in
CM r—
LO r-
CO
CM
CO
to co oo o
CM r— co cn
LO i— in
r— CO in
• • •
o o *s-
V
CM
JM
cn
«t
cr
a:
CO
4-
a>
r™
a.
to
co
CM co o
_ CM C C
O ^~ IB *O
CM O -P -C
a: CM c -P
o 3: ai co
x, «*
i— t- h-
i-iCO
CO
CM CO O
i— CM C C
CMCJ -P -C
3: CM c -p
o a: co co
uT ^v. -s. -~s.
3 r- H- h-
Q. i—i CO CO
CM CO O
i— CM C C
O r— (8
O 3= CO CO
— o Q- s:
a. HI co co
o>
c
o
I/)
CO
CQ
CM
co
CO
5
o
o
20
-------
•p
c
o
o
CM
1
in
UJ
CQ
^_
o
•
CO
4J
C
8
u
•r-
C
to
o
r—
to
•p
o
1-
^3
CO
•p
to
E
•t—
•p
I/I
UJ
to
o
•r-
c
<0
s_
o
•0-JJ-
£HP
CD to
O
TJ
E C
3 tO
CO
>>
CCHc
^^^ ^^
(O ITJ
•P J-
o cu
h-
CO
*io CD
4->
O
4-
o>
"ex
E
(O
CO
c
3
cc
• • •
TD T3 "O
• • f*> •
c c " c
CM CT> CM O CO
in in cr> «* ^-
m i— in co
i— CM
CM VO CM VO VO
in in co in o-i
in r— ^" r—
t— CM
to
CO
TJ
O
in u
CO O «3- CO
o • • co
C VO CO «1- •—
i— CX
v E
(0
y-
•— o
CM CO O <*-
l— CM C C
O r- (O (O CO
CM O -P JC
Z CM C -P X
o o: co cu •!-
-v. o Q- z: TJ
U_ *^» "^ ~>. C
3f r- \- t- CU
O- HH CO CO Q.
0.
lO
^
cu
s_
CO
•p
to
01
• fm
c
•r—
fO
g
S-
co
3
TJ
to
cu
c
o
-o
CO
to
to
CO
*
*3*
^O ^"
0) CO
J-o
to to
to CM
•p s
(O CO
^J
cu
Si- 4^
•r- (0
4i5
f~—
CO T3
t- CU
Q. E
3
co to
o to
c to
•r-
to to
•r-
•«
co -P
a:
•
>t CU
*1
•o to
CO
c cu
•i- JC
E •»->
10
Xaj
w
cu s-
cu
CU 2
s_
co to
2 -p
to
to CU
CO -P
Q. CO
E +•»
ia to
en > 0-
r-^E
c: ro
O «/)
1=
•
CO
a.
E
JC
4->
en
(O
c
,r-
JC
u
<4-
o
CO
C7)
5
c
CO
o
s-
co
Q.
0)
i
to
CO
JC
4—*
T3
ra
f~
to
CO
r—
Q.
E
ro
to
>fc
40
•t—
to
c
CO
4J
c
•r-
n)
c
CD
•r-
to
£
or
rO
+->
O
4->
14-
O
^%
"~
V
II
-a
CO
4->
o
CO
4->
CO
•a
i
c
i
tl
•
•o
c
*
to
(O
to
CO
r—
a.
ra
to •
cu
CO 4->
to i/)
CO ro
JC 2
4-> UJ
o
(^
0 <4-
o
J^
4-> C
5-2
4>J
C ro
3
4-> 0
C i—
CO ra
to u
CO
I- C
Q--r-
.a"8
•o
+^ r^
u
15 =
CO O
Q. C
CX
ra to
, —
fO 4*^
•i- ro
S- J=
CO I—
4J
ra
E *
X""N
TJ CO
CO
4J X
ro fr-
c -a
•i- C
t- cu
O 0-
i— CX
"5 <~
cu
CO CU
i— tO
4— *
ra 4J
r— C
og
1 *P"
c E
O ra
C -P
c
cu o
§u
CO (O
t
•X
21
-------
co
o
o
o
o
flk
o
co
o
CO
CM
SB
«n
VO
en
en
CT>
oo
LU
HH
O
LU
CO
CVJ
o
CO
o
o
CM
o
LO
co
g
en
en
en
en
10
o
•i—
c
• ro
O) 0)
r— J-
ex o
ro TD
CO O>
•P
CO ro
CO 'I-
J_
O) O
co
LO
LU
CO
o
O -P
ro
O C
LU rO
£-cL
_J X
Z3 01
O
_l S-
50
<4-
O -P
Z X
e£ O)
-P
CO
LU CU
oo
oo
LU
LU
z:
o
M
CM
CO
co
o
CO
ro
o
O
o
CO
o
LO
co
CO
CO
en
CTi
en
en
•
en
en
OQ
CO
o
o
1,
i
co
en
01
re
co t: o
E Ol S-
^>O- T3
C" — >,
Et- .E
CO
•> O)
•P C
C CO .r-
O> 3 r—
3 I — |
•— CL C
1- 0
It- CM
LU i— tO
0 3
M CM-
=c rc cx
o
c ^
HH O S-
tO C (U
O CO ro N
•r- 3 ^ >,
C r- +J r—^
ro CX (U ro
cn s c
S- CM-^ (0
0 ^h-
O (_> Ol
&- CM C
O 3Z ro
0 +*
•o ^ c
O) Lu 0)
-(-> 3 Q-
ro 0. -v.
C I—
•r- CO
i-CO
O E to
i ^.3
JC. O> ^- ^^*
0 E CX •—
-O • CM C
0> 4-> r— ro
•P c u x:
ro Ol CM-4-1
E s 3: d>
•^r- 0 £
j % ^^ ^^^ ^^^
•*^ *f* ^^ ^^*
CO IK r- I—
LU LU t-H CO
C
•r—
CO
U
•I—
C
ro
CD
S-
C3TO
E
ro CD
+J E
O
1— •
0
E
3 II
•r- LU
X 0
(0
z: <+-
•o "~
0) «
4-> 4->
ro C
r-
C
E 01
3 3
E r— j.
•i- M- *
X H-
^ LUTO
^C^
•S-2
4-> LO
ro O •
•— -^ O
3 C
U ro II
»«• ^n
^^ w*
(O I- LU
OCC3
^—
ro
4->
O
ui~
0
•P
10
(0
U
s_
O CO
<4- (O
•O 4->
C C
3 O)
O 3
fO O)
CO C
O ••-
•T—
c -o
ro O)
CD S-
S_ ro
O O)
CX
O) ro
•P
ro T3
5- <4-
O CO
O
o.^
cno.
ro
4-> <4-
c o
O)
O r—
0) ro
CX
t»-
0) T-
ro O)
CO 3
O> (O
C >
•r-
CO Ol
22
-------
The calculated values of DE^.Q. , and DE . both suggest
that destruction of the PVC waste was complete in both tests 3M2 and
3M3. There was no significant difference in overall destruction
efficiency (DE^-i) between the PVC waste tests and the background
tests. The calculated destruction efficiency based only on chlorinated
organics (DEwaste) was greater than 99.995% for all three waste tests.
The data in Table 5-3 show decreases in both calculated
destruction efficiencies for the 3M4 test. In that test, the gas
phase detention time within the kiln was 2 sec, as opposed to 3 sec for
the other three tests (3M1(B), 3M2, and 3M3). It seems probable that
this difference, although small (99.996_vs 99.999/0 represents a real
decrease in destruction efficiency with decreased residence time.
Other quantitative characterizations of the combustion zone
effluent gas were provided by the on-line instruments and by analysis
of the hot zone sampling train impinger solutions. Data obtained
from the on-line instruments are presented in Table 5-4.
The sodium hydroxide solutions from the hot zone sampling
train impingers were analyzed for chloride ion in an attempt to
estimate the hydrochloric acid content of the combustion zone
effluent. The measured chloride ion concentrations correspond to
apparent hydrochloric acid concentrations much smaller than those ex-
pected based on the rate of PVC waste feed, as the data below indicate.
Hydrochloric Acid, mg/nv*
Est. from analysis Calc'd Maximum Based
of NaOH Impingers on PVC Feed Rate
3M1(B) 4.8 0
3M2 300. 1664
3M3 470. 1664
3M4 295. 1485
It was then recognized that some hydrochloric acid might have been
collected in the condensate that accumulated in the dry impinger.
Unfortunately,the condensate had already been evaporated to dryness, and
any hydrochloric acid present had been lost.
*
The gas bulb samples were not useful. See Appendix B.
23
-------
TABLE 5-4
QUANTITATIVE RESULTS OBTAINED FROM ON-LINE INSTRUMENTS
Concentration " '
3M1(B) 3M2 3M3 .3M4
Volatile Hydrocarbons
(as methane) 21 ppm 15 ppm 13 ppm 13 ppm
Carbon Monoxide 89 ppm 20 ppm 19 ppm 14 ppm
Carbon Dioxide 6.3% 7.0% 6.1% 6.3%
Oxygen 10.6% 11.2* 11.0% 11.4%
Nitrogen Oxide (NO) 38 ppm 74 ppm 24 ppm 29 ppm
All concentrations are on a wet gas basis, v/v. Estimated
uncertainty is 20% of the value stated in the Table.
' In the three waste tests, the on-line instruments showed
periodic fluctuations corresponding to the batch feed method
used. Values shown are means estimated by averaging individual
data points at U5 minute intervals over a 30-min period
during each test.
24
-------
Data acquired on samples from facility No. 6, after completion of the
analyses described here, showed that 83-93% of the hydrochloric acid
was recovered from the dry impinger condensate, rather than from the
basic impingers. It is presumed that the same phenomenon occurred
in the case of the 3M samples, since the train used was identical and
the two facilities were quite similar.
Since other evidence implies > 99.99% destruction of the
PVC waste, it is assumed that the calculated maximum emissions of
hydrochloric acid provide the best estimate of actual emissions
during these tests.
5.2.2 Qualitative Characterization
A summary of the organic species identified in the hot zone
effluent gases is presented in Table 5-5. More detail is given in
Appendix B. The most notable features of these data are:
t There was no evidence for the presence of vinyl
chloride monomer in any of the hot zone effluent
samples.
• Polynuclear aromatic hydrocarbons are the pre-
dominant species in both the background samples
and the 3M3 waste test samples which were sub-
jected to detailed LRMS analysis. These species
were not found in the corresponding blank samples.
There is no evidence that the emissions of poly-
nuclear aromatics is higher in the PVC tests than
in the background test. In fact, the data suggest
a lower level (2.8 mg/nr vs 5.5 mg/rn^) in the 3M3
test than in the 3M1(B), background test. That
difference may not be significant, but may instead
illustrate the cumulative uncertainty associated
with the entire sampling and analysis effort.
• The few chlorinated organic species identified are
not present in high enough concentration to be
considered environmentally hazardous.
• The alkyl phenols identified in the probe wash
and filter extract are present at significant con-
centrations in the hot zone effluent and appear to
represent PVC combustion products. These species,
however, are found in the particulate portions of
the sample and are probably fairly effectively scrubbed
from the off-gas by the pollution control system.
25
-------
TABLE 5-5
ORGANIC CHEMICAL SPECIES FOUND IN HOT ZONE EFFLUENT SAMPLE EXTRACTS
Approximate Concentration,
mg/m3 *
Species
Vinyl Chloride Monomer
Alky! Phenols
Phenol, Biphenol
Anthroquinone,
Benzanthrone
Polynuclear Aromatics,
mw 154 (Biphenyl) to
mw 254 (Binapthyl)
Chlorinated Organics
Where Found'
.»__
PWF/CH2C12
ST/Methanol
I1/CH2C12
I1/CH2C12
ST/Pentane
ST/Methanol
ST/Pentane
3M1 (B)
**
n.f.
0.14
n.f.
ttr
3.9
1.6
n.f.
3M3
**
n.f.
3.13
1.8
0.6
0.7
0.7
1.4
< 0.02
T
**
Based on LRMS abundance data (Appendix B) and total organic
emissions (Table 5-2)
See Appendix B for sample codes.
Not found
IT This sample was so highly contaminated with silicones that
other species could not be detected.
26
-------
• The phenol and biphenol are present at con-
centrations which probably do not represent appreciable
hazard (1.8 mg/m3 vs 19 mg/m3 for the OSHA Standard
for phenol).
The conclusion to be drawn from these results is that there
do not appear to be organic species produced uniquely in the combustion
of PVC waste that are known to be hazardous at the concentrations
which would be emitted in these tests. This presumes the existence
of a scrubber system at least as efficient as that used at the 3M
Company Chemolite incineration system to remove HC1 and also the
particulate alkyl phenols. However, the polynuclear aromatic
hydrocarbons that are observed for both the background test and the
3M3 PVC waste test do represent potentially hazardous emissions.
These species were found primarily in sample train components down-
stream of the filter, and might therefore not be effectively re-
moved by a wet scrubber. Some members of the general class of poly-
nuclear aromatics are known to have high toxicity (e.g. Benz[a]pyrene,
a carcinogen); those particular species were not the predominant
PNA's in these test samples.
5.3 FINAL EMISSIONS
5.3.1 Stack Gases
The total particulate loading, as measured by EPA Method 5
at the stack, and the comparable values, as measured by the hot zone
sampling train, were:
Particulate, mg/m
Stack Hot Zone
3M1(B) 40.2 122.
3M2 -- 427.
3M3 70.9 357.
3M4 71.4 378.
The estimated emissions of hydrochloric acid, based on
analysis of the sampling train impingers, were:
27
-------
Hydrochloric Acid, mq/m
Stack Hot Zone
(calc'd see p. 23)
3M1(B) 0.65 0
3M2 -- 1664
3M3 14.5 1664
3M4 10.3 1485
These data imply scrubber efficiencies for participate
removal and hydrochloric acid removal of:
Scrubber Efficiency. %
Total
Particulate HC1
3M1(B) 67
3M3 80 99.1
3M4 81 99.3
The hydrochloric acid scrubbing efficiency is thus excellent. Par-
ticulate removal is, not surprisingly, less efficient, since the
scrubber was not designed for this task.
5.3.2 Scrubber Water
There^was no evidence from gas chromatography, IR spectra,
or LRMS analyses of the methylene chloride extracts of the scrubber
water samples that any significant quantity of organic material was
present. The total quantity of material (GC estimate plus gravi-
metric determination) was 1.2 mg/1 for the fresh scrubber water and
2.4 mg/1 for the background test spent scrubber water. For the PVC
waste tests, the spent scrubber water levels were estimated as
5.6 mg/1 (3M2); 5.0 mg/1 (3M3) and 0 mg/1 (3M4). No particular
organic species were identified for these samples because the total
quantities were so low.
The scrubber water samples were also analyzed for chloride
content. The data are presented in Table 5-6. The chloride data
are in the same range as expected based on complete conversion of
chlorine in the waste feed to hydrochloric acid. This reinforces
the hypothesis made earlier that appreciable quantities of hydro-
28
-------
TABLE 5-6
CHLORIDE ANALYSES OF SCRUBBER WATER
Fresh Scrubber Hater
3MO - SI 10
Spent Scrubber Water
3M1(B) - SO 15
3M2 - SO 414*
3M3 - SO 719*
3M4 - SO 816*
These values correspond to estimated concentrations of
1430 mg/m3 (3M2); 2510 mg/m3 (3M3) and 2547 mg/m3 (3M4)
of hydrochloric acid in the incinerator effluent gas.
Calculations based on the rate of PVC waste feed imply
1664, 1664, and 1485 mg/m3, respectively, which is in
reasonable agreement with the scrubber water assays.
29
-------
chloric acid were probably present in the hot zone dry impinger
sample. The scrubber water data also confirm the implication of
the stack impinger data that the scrubber efficiency for removal
of HC1 is high (see data above).
30
-------
6. WASTE DESTRUCTION COSTS
Economic feasibility is one of the main factors in the selection or
recommendation of a waste disposal method. To enable the issue of economic
feasibility to be further examined, Arthur D. Little, Inc., has prepared
capital and operating cost estimates for the destruction of PVC manufactur-
ing waste by incineration. One estimate is based on a system similar to
the 3M Company Chemolite incineration system and on the process operating
conditions in effect during the actual testing. In addition, the cost
estimates specific to the 3M Company facility have been extrapolated down-
ward to represent a much smaller incineration system, of a size more nearly
matching the requirements of an actual PVC manufacturing facility.
6.1 CAPITAL INVESTMENT AND TOTAL OPERATING COST BASED ON THE
3-M COMPANY CHEMOLITE INCINERATION SYSTEM
6.1.1 Capital Investment
The 3M Company Chemolite incineration system was constructed during
1970 and 1971 at a cost of approximately S4.6 million. In terrrs of March
1976 dollars the capital investment would be approximately $7.8 million.
A breakdown of the major components comprising the total capital
investment is presented in Table 6-1.
6.1.2 Total Operating Cost
The total operating cost is composed of variable costs and fixed
costs. The variable costs are the expenditures incurred as a result of
the direct operation of the facility and include fuel oil, electricity,
chemicals, solid waste disposal, labor expenses (and associated overhead),
and all maintenance items. The fixed costs on the other hand, are almost
totally related to the initial capital investment and include depreciation,
cost of capital, and taxes and insurance. The owner of the facility incurs
almost all of the fixed costs regardless of the percentage of time the
facility is in operation during any given year.
The total operating costs are based on the process parameters and
operating conditions in effect during the test period. The conditions
during the November 6, 1976, test run were selected as being most appropriate
for estimating operating costs, and are presented below:
• Detention time within the incinerator = 3 seconds
• Temperature of primary combustion
zone = 870°C (1600°F)
• Temperature of secondary combustion
zone = 980°C (1800°F)
31
-------
TABLE 6-1
CAPITAL INVESTMENT FOR THE 3M COMPANY
CHEMOLITE PLANT INCINERATION SYSTEM
ITEM
CAPITAL INVESTMENT
(approx. mid-1970 dollars
ENR Index = 1375)
Equipment
Rotary Kiln $ 220,000
Secondary Combustion Chamber 125,000
Refractory 260,000
Waste Feed System 105,000
Sewer System 100,000
Stack 62,000
Fans 75,000
Mixing Chamber 75,000
Water Supply System 35,000
Scrubber Water Treatment System 120,000
Total Equipment Cost $1,177,000
Installation and Construction Costs
Structural Costs (includes material
handling building) $1,435,000
Mechanical (piping and equipment
installation) 918,000
Electrical 278,000
Instrumentation and Controls 50,000
Miscellaneous Expenses 56,000
Engineering 660,000
Total Installation and Construction Cost $3,397,000
Total Capital Investment (mid-1970 dollars) $4,574,000
Total Capital Investment in Terms of March 1976
Dollars (ENR Index = 2322) $7,800,000
Source: The 3M Company and Arthur D. Little, Inc.
32
-------
• Waste feed rate = 6 drums per hour
= 1860 Ibs/hr
= 0.845 metric tons/hour
• Waste solids concentration = 28%
• Average fuel oil consumption = 35.2 liter/minute
[Based on a higher heating (9'3 9a11ons per minute)
value of (140,600 Btu/gal)] = 19.8 million KCal/hour
(78.5 million Btu/hour)
In preparing estimates of capital and operating costs the following
premises and assumptions were made.
• Chemicals - Although the 3M Company Chemolite facility
uses ammonia for neutralization, an alkali such as soda ash
is probably more universally applicable. Therefore,
the cost of neutralization was based on the use of soda ash
rather than ammonia. The cost of neutralization was based
on a total wastewater flowrate of 4470 liter/minute
(1180 gal/minute) for scrubber water plus cooling water
less evaporation and a measured chloride ion concentration
increase (indicative of the formation of HC1) of 709 mg/1.
• Electricity - Electrical power costs were based on the 500 HP
fan plus an approximate additional power requirement of
200 HP for miscellaneous electrical equipment, and a motor efficiency
of 80%.
• Labor - Labor is based on the actual manpower requirement of
56 manhours per day.
t Maintenance Items - The cost for maintenance items is based on 3-M's
expenses for yearly maintenance expenditures. Major maintenance items
include corrosion control, refractory replacement, and instrumentation.
• Solid Waste - The only significant amount of solid waste from the
incineration of PVC waste is the steel drums containing the waste.
Although the solid waste is disposed of in a landfill on 3-M
property adjacent to the incinerator, a typical $5.00 per ton of
solid waste disposal cost was included to acknowledge the need for
all incineration systems to practice good solid waste disposal techniques.
• Operating Period - The cost estimates are based on a continuous
waste feed rate of 6 drums per hour, 24 hours per day, 330 days
per year.
33
-------
The total operating cost for the incineration of PVC waste
based on the 3-M Company Chemolite incineration system is presented
in Table 6-2. When operating continuously, it costs approximately
$580/metric ton to destroy PVC waste in the 3-M Company Chemolite
incinerator or an equivalent system.
6.2 CAPITAL INVESTMENT AND TOTAL OPERATING COST FOR THE DESTRUCTION
OF PVC WASTE IN A "PLANT SCALE" INCINERATION SYSTEM
The 3-M Company Chemolite incineration system receives a wide
variety of wastes from many 3-M Company manufacturing facilities. Furthermore,
with a rated capacity of 22.7 million KCal/hr (90 million Btu/hr) it
is among the largest size of industrial incinerators.
The estimated quantity of PVC waste, as reactor bottoms from PVC
manufacture, is 5 million Ibs per year, or 2270 metric tons per year.*
As can be seen from Table 6-2, the 3-M Company Chemolite incineration
system is capable of destroying 6696 metric tons per year of PVC waste,
which is almost three times the quantity of PVC waste generated yearly
in the U. S. It is obvious, therefore, that an incinerator of this size
cannot b*1 devoted solely to the destruction of PVC waste. Also, since there
are 23 PVC manufacturing plants in the U. S., it is clear that if individual
facilities are to employ on-site incineration as a technique for destroying PVC
waste, tne incineration systems must be much smaller than the 3-M Company
Chemolite incinerator.
An incineration system having l/20th the capacity of the 3-M Chemolite
incinerator would be within a reasonable size range for many individual plants.
As in almost any chemical process, there are economies of scale associated
with size. With regard to cost per unit of capacity, a large system
will be cheaper than a small system, both in terms of capital investment
and direct operating cost. Taking the effect of economy of scale into
account, the cost estimates based on the 3-M Chemolite incineration system
were scaled down (using appropriate scaling factors) to a size representing
l/20th of its capacity.
Capital investment for process equipment generally scales according
to the 0.6 power of the characteristic size. Labor costs are also
nonlinear but greatly dependent on the type of system. The cost of
fuel, electricity, chemicals and solid waste disposal is essentially
a linear function of size. The cost estimates for a "plant scale" incineration
system haying the same physical features as the 3-M Chemolite incinerators,
and functioning under the same process parameters and operating conditions
as the November 6, 1976, test, are presented in Table 6-3.
* Source: "Destructing Chemical Wastes in Commercial Scale Incinerators,"
Technical Summary, Volume 1, PB-257-709/6WP.
34
-------
TABLE 6-2
TOTAL OPERATING COST FOR INCINERATION OF PVC WASTE
BASED ON THE 3M COMPANY "CHEMOLITE" INCINERATION SYSTEM
BASIS: • Waste feed rate^ = 0.93 tons/hr (0.845 metric tons/hr)
• Operation - 24 hrs/day, 330 days/yr (6696 metric tons/yr of waste)
Fixed Capital Investment (FCI) - $7,800,000^
VARIABLE COSTS
Fuel Oil
Electricity
Chemicals (soda ash)
Solid Waste ^
Operating Labor
Supervision
Labor Overhead
Plant Overhead
Maintenance
Units
$7.94/MM KCal($2.00/MM Btu)
0.02/Kwh
44 metric ton
(as Na2 03)
$5.50/metric ton
$7/hr
15% operating
labor
30% labor and
supervision
70% labor and
supervision
Units Per
Metric Ton
of Waste
23.4 MM KCal
771 Kwh
0.298 metric
ton
0.068 metric
ton
$ per Metric Ton
of Waste
$185.63
15.41
13.09
0.37
19.32
2.90
6.66
15.66
37.34
Annual
Cost ($)
$1,243,000
103,200
87,700
2,500
129,400
19,400
44,600
104,200
250,000
TOTAL VARIABLE COST
FIXED COSTS
Depreciation (15% FCI)
Cost of Capital (10% FCI)
Taxes and Insurance (2% FCI)
TOTAL FIXED COSTS
TOTAL OPERATING COSTS
$296.28
158.80
105.90
21.20
$285.90
$582/metric ton
$l,984,000/hr
1,170,000
780,000
156,000
$2,106,000/yr
$4,090,000/yr
Notes: (1) The waste feed rate is the actual volume of the waste material fed to the
incinerator during the test program. The PVC waste had a solids concentration
of 28%.
(2) Capital investment adjusted to March 1976 dollars (Engineering News Record
Construction Cost Index = 2322).
(3) The only significant source of solid waste is the drums containing the
waste material.
35
-------
TABLE 6-3
TOTAL OPERATING COST FOR INCINERATION OF PVC WASTE
BASED ON A "PLANT-SCALE" INCINERATOR
BASIS: • Waste feed rate
(1)
• Operating periods
Fixed Capital Investment (FCI) $1.300.000^
VARIABLE COSTS
Fuel Oil
Electricity
Chemicals (soda ash)
0.0423 metric tons/hr (one twentieth the
capacity of the 3-M Chemolite incineration system
used in the test)
24 hrs/day, 330 day/yr (335 metric tons per year of waste)
Solid Waste
(3)
(4)
Operating Labor
Supervision
Labor Overhead
Plant Overhead
Maintenance
TOTAL VARIABLE COST
FIXED COSTS
Depreciation
Cost of Capital
Taxes and Insurance
TOTAL FIXED COSTS
TOTAL OPERATING COST
$/Un1t
$7.94/MM Kcal ($2.00/MM Btu)
0.02/Kwh
Units per
Metric Ton
of Waste
92.85 MM Btu
771 Kwh
44/metric ton 0.298 metric
(as NapCO-) ton
$5.50/metric
$7/hr
15% operating
30% labor and
70% labor and
(15% FCI)
(10% FCI)
( 2% FCI)
ton 0.068 metric
ton
labor
supervision
supervision
$ per Metric Ton
of Waste
185.63
15.41
13.09
0.37
165.38
24.78
57.01
133.13
124.48
719.28
582.09
388.06
77.61
$1,047.76
$l,767/metr1c ton
Annual
Cost ($)
62,200
5,200 " '
4,400
200
55,400
8,300
19,100
44,600
41.600
241 ,000
195,000
130,000
26,000
$351 ,000/yr
$592,000/yr
Notes: (1) The waste feed rate is the actual value of the waste material fed to the incinerator
during the test program. The PVC waste had a solids concentration of 28%.
(2) Capital Investment adjusted to March 1976 Dollars (Engineering News Record
Construction Cost Index - 2322).
(3) The only significant source of solid waste is the steel drums containing
the waste material.
(4) Based on 1 full-time operator
36
-------
The total cost of destroying PVC waste in a "plant scale" incineration
system is approximately $1770/metric ton, as compared to $580/metric ton
for the larger Chemolite facility. The proportionately higher fixed costs
and labor costs are principally responsible for the higher total cost
associated with the smaller facility. It should be recognized that some
PVC manufacturing operations might require considerably smaller
facilities, thereby further increasing the unit cost of waste incineration.
Lower destruction costs could be achieved if the PVC waste were co-incinerated
with other combustible wastes thereby increasing the incinerator size and
reducing the unit costs.
37
-------
APPENDIX A
SAMPLE VOLUME DATA AND FLOW RATES
-------
§
r— C3
Ul D-
ff oo
u
to
Total Volume 1
Sampled at STP 1
21 °C +760 |S
(Wet Basis)
s.cu. m.
Isokinetic
Average
Average
Sampling
Velocity m/sec
Average Stack
Gas Velocity
m/sec
Moisture
Content
Moisture
Content %
Carbon Dioxide
Oxygen %
Stack Pressure
mm of Hg
Stack Temp.
°C
Run Number 1
CM
r~"
in
-
00
CM
CM
i —
CM
O
CM
in
o
o
r—
10
,
•~
cr>
<£>
in
in
^
I
Date
CO
-
^
in
vo
1
co
00
A-l
-------
TABLE A-2
HOT ZONE SAMPLING DATA
Total Volume Sampled,
Run Moisture Content s.cu.m., Wet Basis at
Date Number % Volume 21°C and 760 mm Hg
11/4 3M1(B) 5.9 5.126
11/5 3M2 7.9 4.921
11/6 3M3 8.2 5.034
11/8 3M4 8.7 4.974
A-2
-------
TABLE A-3
ESTIMATED TOTAL GAS EFFLUENT FLOW RATES
HZ
Run No. SCF/Min SCM/Min SCF/Min SCM/Min
3M1(B) 38,300 1085 42,700 1200
3M2 *
3M3 44,100 1250 48,700 1380
3M4 49,400 1400 56,510 1600
*
No stack sample was acquired during this test. Since total
flow estimates are based on the stack diameter and measured
stack gas velocities, no estimate could be made for the
3M2 test. Flows should be comparable to those for the 3M3
test.
A-3
-------
APPENDIX B
ANALYTICAL CHEMISTRY DETAILS
B.I Sample Codes, Preparation, and Analysis Procedures for Samples
Returned to the Laboratory
B.2 Gravimetric and Volumetric Data
B.3 Analytical Results for Representative Waste Feed
B.4 Analytical Results for Probe Washes and Filters
B.5 Analytical Results for Dry Impingers
B.6 Analytical Results for Sorbent Traps
B.7 Analytical Results for Impingers
B.8 Analytical Results for Scrubber Water
B.9 Analytical Results for Gas Bulb Samples
-------
B.I SAMPLE CODES, PREPARATION AND ANALYSIS PROCEDURES
B.I.I. Sample Codes
Samples are identified by multi-syllable codes in which
• The first syllable identified the test run:
3M1(B) = background test on 11/4/76
3M2 = PVC test on 11/5/76
3M3 = PVC test on 11/6/76
3M4 = PVC test on 11/8/76
3MO = general sample such as blended re-
presentative waste feed
t The second syllable identifies the sampling location:
-HZ = hot zone (combustion zone)
-S- = stack
• The third syllable identifies the type of sample
train component:
-PW = probe wash
-F = filter
-II = dry impinger (See pages 13, 14 and B-2)
ST = sorbent trap
I = impinger
SI = fresh scrubber water
SO = spent scrubber water
GB = gas bulb
t A final syllable indicates the solvent used, if the
sample is an organic extract.
B.I.2. Sample Preparation Procedures
B.I.2.1. Probe Washes
The probe washes were tranferred to tared glass evaporating dishes
and the solvent evaporated on a hot plate. The dishes and contents were
then dried to constant weight in a desiccator over Drierite.vJ)*
Trademark of W.A. Hammond Drierite Company, Xenia, Ohio
3-1
-------
The residue in the evaporating dish, for the hot zone (HZ) probe
washes only, was taken up in methylene chloride. Several small portions
of solvent were used, with agitation in an ultrasonic bath to facilitate
uptake of soluble portions of the residue. Each resulting methylene
chloride suspension was combined with the corresponding hot zone filter
for Soxhlet extraction as described below.
B.I.2.2. Filters
The filters were dried to constant weight in a desiccator over
Drierite(K). The dried filters were photographed.
The filters for the hot zone only were folded and placed in pre-extrac-
ted (24 hours with methylene chloride) cellulose Soxhlet thimbles. The
methylene chloride suspension of the corresponding probe wash sample
(above) was poured through the thimble; the solvent which drained through
was added to the boiling flask of each Soxhlet apparatus. The boiling
flask was charged with 200 ml of methylene chloride. The extraction
was allowed to proceed for 24 hours.
An empty, pre-extracted thimble was extracted as a blank.
B.I.2.3. Dry Impinger
The "Dry Impinger" was an empty, standard-sized impinger, used be-
tween the filter and the sorbent trap in the hot zone sampling train
at 3M. The sample obtained from this train component consisted of a
measured volume of condensate, combined with water, acetone, and pentane
rinses of the impinger and connecting glassware.
In the laboratory, the total sample volume was measured. An aliquot
of the sample was set aside for chloride and nitrate analysis. The re-
mainder of the sample was transferred to a tared glass evaporating dish.
The liquid was evaporated on a hot plate. The dish and_contents were
dried to constant weight in a desiccator over Drierite(B).
The residue in the evaporating dish was taken up in methylene chloride.
Several small portions of solvent were used, with agitation in an ultra-
sonic bath to facilitate uptake of soluble portions of the residue. The
methylene chloride solutions were transferred to 10 ml volumetric flasks
and made to volume with fresh solvent.
B.I.2.4. Sorbent Traps
The sorbent traps were placed in the specially designed extraction
apparatus shown in Figure B-l. Each trap was extracted for 24 hours with
pentane and then for 24 hours with methanol. The extracts were not com-
bined for analysis. An unused sorbent trap was extracted; this served
as a blank.
B-2
-------
;n2o
Condenser
Flexible Teflon Coupling
250 Ml Flask
Figure B-l Sorbent Trap Extractor
B-3
-------
B.I.2.5. Impingers
Total volumes of the combined impinger solutions* and distilled water
rinses of glassware from each sampling train was measured.
A portion of each solution was acidified to pH <2 (pH paper) with
concentrated nitric acid. This acidified portion was refrigerated and
stored in a Nalgene(T?)** container.
A separate portion of each of the combined solutions was taken for
chloride analyses.
B.I.2.6. Scrubber Water
The fresh scrubber water samples collected on 4 different test days
were combined to give a single sample 3MO-SI. A portion of this solution
was acidified to pH <2 (pH paper) with concentrated nitric add and re-
frigerated. A separate portion of the combined sample was taken for
chloride analyses. The pH of the combined sample was found to be 6.
The spent scrubber water samples were not combined. An acidified
(pH <2) portion of each was stored in the refrigerator. A separate por-
tion was taken for chloride analyses. The pH's were measured.
For determination of organics, two 500-ml portions of each of the five
scrubber water samples above and of a distilled water blank were extrac-
ted with 3 x 20 ml each of methylene chloride after adjusting the ph to
7 with ammonium hydroxide. The organic extracts were dried by passing
them through anhydrous sodium sulfate.
B.I.3. Analysis Procedures
B.I.3.1. Gas Chromatographic Analysis of Vinyl Chloride Monomer in
Representative Waste Feed
Portions of the blended representative waste sample (5-8 g wet weight)
were weighed into six tared 68 ml serum bottles. The bottles were capped
with rubber serum stoppers. Three of the bottles were "spiked" by injec-
tion of 150 vn, 300 \ii or 500 yfc of pure vinyl chloride monomer through
the serum stopper. The bottles were placed in an ultrasonic bath for
15 minutes, then allowed to stand at room temperature for four days.
Not including the "dry" impinger of the hot zone sampling train. See
B.I.2.3
**
Trademark of Nalge Company, Rochester, New York
B-4
-------
After the equilibration period, the head space in each bottle was
sampled with a gas-tight syringe. Replicate 500 y«, samples were analyzed
by on-column injection. Chromatographic conditions were:
instrument = Varian 1200
detector = FID at 200°C
injector = 150°C
column = 6' x 1/8" stainless ste^l, 60/80
mesh Chromosorb 102 fl*
carrier = nitrogen
program = isothermal at 100°C
The detector response was calibrated by analysis of 500 y£ samples of
1000 ppm vinyl chloride monomer in nitrogen (calibration mixture purchased
from Supelco, Inc.).
After the gas Chromatographic analysis, the head space in the bottles
was purged with nitrogen and exhausted into an efficient hood*. The
serum caps were then removed and the bottles and contents dried to con-
stant weight in a vacuum oven at <50°C.
To calculate the vinyl chloride concentration of the waste, the
measured head space concentration in each bottle minus the concentration
expected from the spike (if any) was divided by the dry weight of sample.
The estimated mean value of head space concentration per gram of sample
was 4560 j^ 1160 ppm (v/v)/g of waste. This quantity of vinyl chloride
monomer in the head space corresponds to 800 ppm by weight of dry waste
or 220 ppm by weight of wet waste.
B.I.3.2. Gas Chromatographic Analysis of Unconcentrated Organic Extracts
The organic solvent extracts prepared, as described above, from the
probe wash plus filter, dry impinger, sorbent trap and scrubber water
from each test's sampling effort were analyzed by gas chromatography
prior to any concentration step. The total volume of each extract was
measured, and 5 y«, portions were taken for injection.
The gas Chromatographic conditions were as follows:
instrument = Varian 2700
detector = FID at 295°C
injector = glass lined, 235°C
column = 6' x 1/8" stainless steel packed with 10%
OV-101 on 100/120 mesh SupelcoportPV*
*
This step was done slowly and carefully to ensure adequate dilution of
the vinyl chloride monomer in the head space. The <68 ml of head
space contained about 1 percent by volume of monomer. The hood was
estimated to provide a dilution factor >103.
**
Trademark of Supelco, Inc., Bellefonte, Pa.
***Trademark of Johns-Manville, Denver, Colorado.
B-5
-------
program = Isothermal at 30°C for 6 m1n. Linear
program at 10°C/min, to 250°C. Hold
for 5-10 min. at 250°C
Quantitative calibration of the detector response was accomplished
by use of standard solutions of ortho dichlorobenzene. This compound
was chosen because it has a chlorine content (48% by weight) similar to
that calculated for vinyl chloride and polyvinyl chloride (56% for both
monomer and polymer). Figure B-2 shows the calibration curve obtained
for dichlorobenzene. Also shown in the figure are some data for more
highly chlorinated species, trichloroethane (79% chlorine) and per-
ch! oroethylene (85% chlorine). It was decided that use of the latter
species as models would lead to unrealistically high estimates of con-
centration, since incineration effluent components would be no more highly
chlorinated than the waste itself.
The objective of this gas chromatographic procedure is to obtain a
quantitative estimate of compounds which might be lost on concentration
of the extract prior to gravimetric anlalysis. The range of interest
was considered to be materials boiling between about 100 and 220°C.
B.I.3.3. Concentration of Organic Extracts
After gas chromatographic analysis, each of the organic extracts
was concentrated to a volume of <10 ml*. Concentration of pentane and
methylene chloride extracts was accomplished by allowing solvent to
evaporate from an open container under a gentle stream of nitrogen.
Methanol extracts were concentrated using Kuderna-Danish apparatus.
The concentrated extracts were transferred to 10 ml volumetric flasks
and the volume restored to 10 ml using fresh solvent.
B.I.3.4. Gravimetric Quantification of Organic Extracts
A 5 ml aliquot (one half of total sample) was withdrawn by volumetric
pipette from each 10 ml extract and transferred to a tared aluminum
weighing dish. The contents of the dishes were allowed to evaporate in
a hood at ambient temperature until constant weight (+_ 1 mg) was obtained
(repetitive weighings at least 6 hours apart).
B.I.3.5. Infrared Spectroscopy
A portion of each concentrated organic extract was taken for IR
analysis. Spectra were obtained using KBr micropellets and a Perkin-
Elmer Model 521 grating spectrophotometer. Peak intensities are reported
as strong (s), medium (m) or weak (w).
The I-1/CH2C12 extracts were initially 10 ml in total volume and were
not concentrated further.
B-6
-------
1 \J\J\J
>
.E
I/I
a
E
'£*
'to
rt
CO
t<0
o
*-*
-O
O
£
E
•4-^
£
Dl
1
•L
ra
a>
a.
(
: /
: /
/
/
o
f *a
iN^
/v
-------
B.I.3.6. Low Resolution Mass Spectrometry (LRMS)
A portion of each concentrated organic extract was taken for LRMS
analysis. Analyses were done using both the batch inlet and the direct
insertion probe of DuPont (CEC) 21-110B high resolution mass spectro-
meter. Because GC and IR analysis indicated close resemblance among
corresponding samples for the three PVC tests, the LRMS analyses were
done on the 3M3 extracts but not on the 3M2 or 3M4 analogs.
Both qualitative and quantitative results of the LRMS analyses are
reported. The quantitative data given are expresses as estimated percent
abundance of the indicated species, relative to the total quantity of
sample which vaporized in the mass spectrometer inlet. They are not
normalized to correct for non-volatile portions of the sample. The pre-
cision of the estimated values varies with concentration; data for major
components are probably reliable to within 50%. For components at very
low concentration, the error may be as large as a factor of five.
Mass spectrometric analysis on gas bulb samples was done by Gollub
Analytical Services Corporation, Berkeley Heights, New Jersey.
B.I.3.7. Elemental Analyses
These analyses were performed only on the waste feed and not on the
effluent samples. Elemental analysis for probable major components (C,
H, N, Cl, S) was done by Galbraith Laboratories, Knoxville, Tennessee.
Elemental analysis for trace components, especially metals, was done
using spark source mass spectroscopy (SSMS) by Commercial Testing and
Engineering Laboratories, Golden Colorado. The sample was thermally
ashed at low temperature (350°C) for 1 hour in a laboratory furnace in a
quartz crucible prior to analysis.
The ash content of the waste was determined in the Arthur D. Little,
Inc., Laboratories by ashing in a muffle furnace at 850°C.
B.I.3.8. Anion Analyses of Aqueous Samples
The stack and hot zone impinger solutions and the scrubber water
samples were analyzed for chloride.
For chloride analyses, a suitable aliquot was taken oxidized with
hydrogen peroxide, and treated with barium nitrate to precipitate sulfur
species as barium sulfate. The resulting solution was then titrated
using standardized mercuric nitrate titrant with S-diphenyl carbazone
as the endpoint indicator. Standard 0.1 N hydrochloric acid was used
as a reference standard.
B-8
-------
B.2 GRAVIMETRIC AND VOLUMETRIC DATA
Table B-l presents the measured volumes for the various impinger
solutions. These data are used in calculating the percent moisture in
the sampled streams. The total final volume of solution is also used
to estimate a mass emission concentration from the measured chloride
concentration.
Table B-2 presents gravimetric data for the three types of sample
(PW, F, and HZ-I1) that were dried to constant weight before extraction.
For each test and sampling location, the PW and F numbers in this table
are the values which are summed to estimate the total particulate loading.
Table B-3 presents the results of gravimetric analyses on the
various concentrated extracts. Also included are data for the various
blanks which were run in parallel with the samples.
B.3 ANALYTICAL RESULTS FOR REPRESENTATIVE WASTE FEED
The waste was found to consist of orange-brown solid material
(spherical particles) which appeared to be wet. The solid content of
the waste was estimated to be 27.9% (residue after drying at 50°C in
vacuum).
The elemental analysis of a dried sample of waste showed the following
composition:
WEIGHT %
%C %H %N %C1 %S
found: 38.41 5.00 0.06 53.85 0.22
calculated for PVC: 38.7 4.8 - 56.5
These data are consistent with the presumption that the solid component
of the waste was predominantly polyvinyl chloride (PVC).
The representative waste feed was found by gas chromatographic
analysis to contain 220 +_ 55 ppm by weight of vinyl chloride monomer.
Examination of the head space above a sample of representative waste
feed by LRMS confirmed the presence of vinyl chloride monomer and did
not identify any other organic species of comparable volatility.
Soxhlet extraction of dried portions of the representative waste
feed with methylene chloride indicated that about 39% by weight of the
dried material was extractable. The IR spectrum of the extract was
similar to a reference spectrum of polyvinyl chloride, with one addi-
tional weak band at 1530 cm-1. Only a small portion of the total ex-
tractable material was volatile under the conditions used for the LRMS
analysis. The species found and the relative abundances were:
B-9
-------
TABLE B-l
VOLUMES OF IMPINGER SOLUTIONS
Run
3MI(B)
Sample
HZ-n
HZ.I
S-I
Volume
Charged
Before Test
ml
0
400
200
Volume
Recovered
After Test
ml
78
605
335
Final Total Volume
Including
Rinses
ml
930
700
3M2
Hz-n
HZ-I
0
400
165
585
990
3M3
Hz-n
HZ-I
S-I
0
400
200
140
628
350
900
710
3M4
HZ-I1
HZ-I
S-I
0
400
200
575
410
920
785
No stack sample was acquired during this test.
B-10
-------
TABLE B-2
GRAVIMETRIC DATA FOR PROBE WASH, FILTER, AND
DRY IMPINGER SAMPLES
Residue After Drying
to
Run Sample Constant Weight, mg
3M1(B) = HZ-PW 333.5
HZ-F 290.8 (2 filters)
HZ-I1 39.2
S-PW 18.2
S-F 32.3
3M2 HZ-PW 940.5
HZ-F 1162.5 (4 filters)
HZ-I1 25.7
*
3M3 HZ-PW 874.1
HZ-F 922.2 (2 filters)
HZ-I-1 51.3
S-PW 14.1
S-F 77.5
3M4 HZ-PW 1015.1
HZ-F 864.9 (4 filters)
HZ-I-1 48.9
S-PW 15.2
S-F 92.9
No stack sample was acquired during this test.
B-ll
-------
TABLE B-3
RESULTS OF GRAVIMETRIC ANALYSES OF CONCENTRATED
ORGANIC EXTRACTS
Run Sample Total Weight, mg*
3M1(B) HZ-PWF/CH2C12 72.6
HZ-n/CH2Cl 2 13.0
HZ -ST/Pentane 19.8
HZ-ST/Methanol 24.8
SG/CH2C12 2.4
3M2 HZ-PWF/CH2C12 73.0
HZ-I1/CH2C12 4.6
h'Z-ST/Pentane 4.6
HZ-ST/Methanol 24.8
SO/CH2C12 5.6
3M3 HZ-PWF/CH2C12 52.6
HZ-I1/CH2C12 11.8
HZ-ST/Pentane 3.8
HZ-ST/Methanol 59.0
SO/CH2C12 5.0
3M4 HZ-PWF/CH2C12 55.2
HZ-I1/CH2C12 5.6
HZ-ST/Pentane 13.2] ,- sorbent
HZ-ST/Methanol 145.6J (* sorDent
SO/CH2C12 0.0
Soxhlet »,a,,R/un2ti2 i 2
ST Blank/Pentane 2 2.8
i.'o
1.2
B-12
-------
Probe Temp.
Relative for First
Sped es Abundance Appearance
Phthalate esters -v 2 % 50°C
Tetrachloroethylene ^ 1 50
Unsaturated Aliphatics,
mw 270-450 ^2 180
HCI, as decomposition product ^5 350
Benzene, as decomposition product ^ 6 350
Zinc chloride * 4 390
Cadmium ^ 1 390
The hydrochloric acid and benzene are clearly products of sample
decomposition on the heated probe because they first appear at a tem-
perature much higher than required for volatization of these species.
The detection of zinc chloride and cadmium by LRMS was unexpected, so
the identifications were confirmed by peak matching at high resolution.
These species appear to be present at very low concentrations in the
original PVC waste (see below). The detection of the zinc and cadmium
by LRMS results not only from the preliminary concentration of the sample
for analysis but also from the very low limits of detection for species
with distinctive isotope patterns in the LRMS.
Characterization of inorganic components of the waste included a
loss on ignition determination that implied an ash content of 1.7%.
Elements found by SSMS to be present in a dried sample of waste at con-
centrations greater than 10 ppm are listed in Table B-4. None of these
was present at high enough concentration to cause concern for emissions
of toxic metals at the feed rates actually used in the 3M tests*. The
SSMS data confirm the LRMS implication of small quantities of zinc, but
cadmium was not detected by SSMS. The limit of detection for SSMS is
estimated as 0.05 ppm. Examination of a sample of dried waste by X-ray
fluorescence again confirmed low levels of zinc but failed to detect
cadmium. Although some cadmium may have been lost from the sample during
ashing prior to SSMS (350°C for 1 hour), it is estimated that the con-
centration of this metal in the orginal, wet waste could not have ex-
ceeded 1 ppm.
*
For the actual waste feed rates and stack gas flow rates in the 3M tests,
1 ppm of a species in the wet waste would correspond to <0.01 mg/m3
of stack effluent, assuming zero scrubbing efficiency for that species.
B-13
-------
TABLE B-4
ELEMENTS IDENTIFIED IN PVC WASTE BY SSMS
Concentration, ppm (w:w)
Element
Calcium
Chlorine
Iron
Magnesium
Manganese
Phosphorus
Potassium
Silicon
Sulfur
Sodium
Chromium
Zinc
Potassium
Copper
Aluminum
Nickel
Barium
Titanium
Stronti urn
Fluorine
Lead
Molybdenum
Arsenic
Selenium
Vanadium
Boron
Bromine
Dry Weight Basis
(as measured)
MC (>100 ppm)
Wet Weight Basis
(calculated)*
MC (>28 ppm)
90
89
58
48
47
35
29
17
16
5
3
3
3
2
2
2
1
1
25
25
16
13
13
10
8
5
4
1
0.8
0.8
0.8
0.6
0.6
0.6
0.3
0.3
Assuming 28% solid content, 72% water.
B-14
-------
B.4 ANALYTICAL RESULTS FOR PROBE WASHES AND FILTERS
B.4.1. Quantitative Analysis by Gas Chromatography
Analysis of the unconcentrated methylene chloride extracts by gas
chromatography showed no peaks with retention times less than 1200 sec.
(equivalent to a n-hydrocarbon boiling point of 216°C).
B.4.2. Qualitative and Quantitative Analysis by IR and LRMS
The IR spectra of the probe wash and filter extracts for the back-
ground and the three waste test samples were all very similar. The
spectra showed:
aliphatic CH2 and CH3 (s)
carbonyl bands at 1710 and 1730 cm'1 (w)
covalent nitrate R-0-N02 or nitramine at 1620,
1255 cnr1 (w, m)
aromatic ringstretch 1500,1570 cm'1 (w)
All of these, except the aromatic ring stretching bands, were also found
in the methylene chloride solvent blank.
A detailed LRMS analysis was done only for the 3M1 (B) and 3M3
samples, since all three waste test samples appeared identical by GC
and IR. Species identified and their relative abundances within each
sample were:
Relative Abundance
3M1 (B) - 3M3 - HZ -
Species HZ-PWF/CH2C12 PWF/CH2C12
mw 326 (C21H2603) 70 23
Fatty acids Cm,16,17,18 18 18
Triglycerides 11 24
Alkylphenols 1 30
Other, unidentified 5
The fatty acids and triglycerides are most probably due to contamination
(fingerprints). The mw 326 species, tentatively identified as hydroxy
octoxy benzophenone, may also be a contaminant, although it does not
appear in the Soxhlet thimble blank. This species was also found in
samples collected at facility No. 6, Rollins Evironmental Services,
where an entirely different waste was burned. The data suggest that
this is not a product associated with combustion of polyvinyl chloride
waste, per se. The alkylphenols, on the other hand, do appear to be
attributable to the PVC waste. The quantity found corresponds to less
B-15
-------
than about 16 mg in the 3M3-PWF sample.
B.5 ANALYTICAL RESULTS FOR DRY IMPINGERS
B.5.1. Quantitative Analysis by Gas Chromatography
Analysis of the methylene chloride extracts of the 3M1(B), 3M2, 3M3,
and 3M4 - II samples by gas Chromatography showed that all four were very
similar, except that the 3M4 extract was somewhat more concentrated than
the others. Retention times of the most intense peaks in these chroma-
tograms were: 540, 850, 1030 sec. (under conditions where a n-hydrocarbon
of boiling point =216°C has a retention time of 1200 sec.). The total
quantity of material with retention time less than 1200 sec., the dichloro-
benzene calibration curve, was<0.1 mg for 3M1(B), 3M2 and 3M3 and<0.35
mg for 3M4.
B.5.2. Qualitative and Quantitative Analysis by IR and LRMS
The IR spectra of the dry impinger samples for the background and
the three waste tests were similar. The spectra showed:
silicone bands (s in 3M1(B) + 3M2, w in 3M4)
aliphatic CH2CH3
carbon yl at 1720 cm"1 (w-m)
hydroxyl (m-w)
aromatic (w)
Also observed were a weak C-0 alcohol stretch in 3M2-HZ-I1/CH2C12 and
C-0 ester bands in 3M3-HZ-I1/CH2C12.
Species identified in the 3M3 sample by LRMS, and their relative
abundances, were:
Relative Abundance. %
Species 3M3 - II
mw 116 (probably butylacetate) 11
Fatty Acids 10
Phthalates 5
mw 326 (hydroxyoctoxybenzophenone) 2
Benzanthrone 21
Anthraquinone
mw 270 9
mw 252 (not polynuclear)
Hydrocarbons, probably polynuclear 32
aromatics,
B-16
-------
Relative Abundance, %
Species 3M3 - II
mw 222 - 306
other (mw <300; not chlorinated)
The 3M1(B)-I1 sample was found to be contaminated with silicones
to the extent that no other species, except small quantities of phthalate
and anthroquinone, could be observed in the LRMS. It is presumed on
the basis of the data obtained for the probe wash and filter extracts
and for the sorbent trap extracts, that the 3M1(B)-I1 sample probably
contains quantities of polynuclear aromatics comparable to those found
for the 3M3-I1 extract.
Of the species identified in the LRMS of the 3M3-I1 extract, the
mw 116 and 326 species, the fatty acids, and the phthalates most probably
arise from sample contamination. The polynuclear aromatics and the oxy-
genated aromatics, benzanthrone and anthraquinone, are probably true
products of combustion, but it cannot be determined whether they arise
from the PVC waste or thr #2 fuel oil. No chlorinated organics were
identified in the 3M3-I1 extract.
B.6 ANALYTICAL RESULTS FOR SORBENT TRAPS
B.6.1. Quantitative Analysis by Gas Chromatography
Gas chromatographic analysis of these sorbent trap pentane extracts
showed several peaks in the range of interest (retention time less than
1200 sec.). In Table B-5, the data are summarized and quantitative es-
timates based on the dichlorobenzene calibration curve are presented.
It is seen that for both the 3M2 and 3M3 waste test samples, the quan-
tities and retention times of components are similar to those found for
the background test sample (3M(B)) and for an unexposed sorbent trap.
The total quantity of material is, in each case, <1 mg as dichlorobenzene.
For the 3M4 test sorbent traps, which were exposed sequentially during
the three hour hot zone sampling period, the quantity of material ex-
tracted appears to be considerably higher and to include a component
(RT ^720 sec.) which are not in the blank or background samples. However,
it seems likely that most of this apparent increase is due to contamina-
tion, because the quantity found is not correlated with volume of gas
sampled by each sorbent trap - 1.625 cu m for 3M4-HZ-ST-1 and 3.348 cu m
for 3M4-HZ-ST-2.
Methanol Extracts
In Table B-5 the data from GC analysis of the methanol extracts are
summarized. It is apparent that the samples from the waste test burns,
3M2, 3M3, and 3M4 are similar qualitatively and quantitatively to the
background test sample and to the extract of an unexposed sorbent trap.
B-17
-------
TABLE B-5
RESULTS OF GC ANALYSIS OF UNCONCENTRATED
SORBENT TRAP EXTRACTS
PENTANE EXTRACTS mg (as dichlorobenzene)
3MO-ST
RT. sec* Blank 3M1-B 3M2 3M3 3M4
430
660
720
900
1000
1070
1100
0.2
0.2
0.08
0.2
0.07
0.1
0.1
0.1
0.2
0.06
0.3
0.08
0.2
0.2
0.02
0.09
0.03
0.07
0.1
0.6
1.2
2.2
0.6
0.9
0.5
Total** 0.82 0.78 0.76 0.32 4.0 2.0
METHANOL EXTRACTS
660
690
720
740
780
950
1070
0.4
0.5
0.6
0.4
0.2
1.4
0.9
0.8
1.7
1.4
2.1
0.8
0.9
1.3
0.9
1.0
0.7
0.7
0.7
0.5
1.0
0.7
0.6
0.7
1.0
0.7
0.7
0.7
0.6
Total** 2.5 8.2 7.7 4.5 3.9 4.5
*Estimated retention (+ 5%) relative to leading edge of solvent front.
*
Includes several smaller peaks, not listed here, in some cases.
B-18
-------
B.6.2. Qualitative and Quantitative Analysis by IR and LRMS
Pentane Extracts
The IR spectra of these sorbent trap extracts showed aliphatic CH
stretching as the dominating feature, as indicated below:
Functional Group
aliphatic CH,CH2 s s s s
aromatic m w/m w w
carbonyl (ether, ketone) w w w/m w
hydroxyl vw w
ether C-O-C w/m w/m
silicone m m/s m w/m
The aromatic bands in the IR spectrum suggested the presence of poly-
nuclear aromatic hydrocarbons (PNA's). This was confirmed by LRMS
analysis of the 3M1(B) and 3M3-HZ-ST/Pentane extracts, as shown in
TAble B-6. The LRMS evidence shows that the PNA's, in fact, are the
predominant species in both the background sample and the waste test
sample. There are not significant differences, in either the total per-
centage of PNA's or the distribution of individual PNA's, between the
background and the waste test pentane sorbent trap extracts. The two
chlorinated species identified in the 3M3-HZ-ST/Pentane sample are pre-
sumed to be produced from the PVC waste.
Methanol Extracts
The IR spectra of the methanol extracts of the sorbent traps were
similar for all four samples, 3M1(B), 3M2, 3M2 and 3M4, and showed:
hydroxyl or N-H stretch (m/s)
aliphatic CH2,CH3 (m)
carbonyl; 1730 cnr1 (m)
secondary amide bands; 1650, 1540 crtr1 (m/s)
sulfonate or aromatic ether; 1200, 1250 cm"1 (s)
ether or alcohol; 1100, 1050, 1000 cm'1 (m/s)
The LRMS analysis of the 3M1(B) and 3M3-HZ-St/Methanol samples indicated
that much of the material was not volatilized in the probe inlet. The
LRMS data are summarized in Table B-7. The 3M1(B) extract was found to
contain a variety of polynuclear aromatic hydrocarbons which were absent
in the 3M3 waste test extract. Both extracts appeared to contain sig-
nificant and essentially equal quantities of chlorinated material that
decomposed to form HC1 at high probe temperatuees. Since no chlorinated
waste was being fed to the incineration during the 3M1(B) test, it must
be concluded that the chlorinated material is a contaminant and not a
B-19
-------
TABLE B-6
RESULTS OF LRMS ANALYSIS OF CONCENTRATED
PENTANE EXTRACTS OF SORBENT TRAPS
Relative Abundance*
mw Species 3M1(B) 3M3
178 Anthracene 16 13
192 Methyl Anthracene 11 9
202 Pyrene 35 22
204 Phenylnapthalene 15 18
206 Dimethylphenanthrene 5 2
208 Methyl Phenyl Indane 2 1
210 Diphenyl Butane 2 1
212 1 1
216 Methyl Pyrene 3 1
218 3
220 Trimethylphenanthrene 1 1
226 Benzfluoranthene 1 1
228 Chrysene 1 1
230 Terphenyl 1 3
232 Ci8Hi6 1
234 Butyl anthracene <1
236 2
238 Decahydrobenzanthracene 1
240 Dodecahydrobenzanthracene * 1
242 Methyl Chrysene <1
244 Triphenylmethane 1
252 Benzpyrene <1
254 Binapthyl <1
(Total Polynuclear Aromatics) (98) (83)
270 Unidentified 1
B-20
-------
TABLE B-6
(cont'd)
RESULTS OF LRMS ANALYSIS OF CONCENTRATED
PENTANE EXTRACTS OF SORBENT TRAPS
Relative Abundance*
mw Species 3M1(B) 3M3
Silicones 1
Pentachlorophenol 1
Hexachlorobenzene <1
*
Percent of total signal intensity for each sample.
B-21
-------
TABLE B-7
RESULTS OF LRMS ANALYSIS OF CONCENTRATED
METHANOL EXTRACTS OF SORBENT TRAPS
Relative Abundance*
mw Species 3M1(B) 3M3
Benzole Acid 17
Methyl Benzoate 1
HC1, as decomposition
product 17 40
SiF3+ (from SiFj 27
Tolyl tolyl ether 5
Fatty Acids 3
N and NO Compounds, not
identifiable, as de-
composition products 24
C8H12N202 5
Phenol 12
Biphenol 2
Napthalene 5
Biphenyl 9
178 Anthracene 4 2
192 Methyl Anthracene 1
202 Pyrene 3
204 Phenyl Napthalene 2
206 Dimethyl Phenanthrene 1
208 Methyl Phenyl Indane 2
210 Dimethyl Butane 1
212 1
216 Methyl Pyrene 1
218 . 1
220 Trimethyl Phenanthrene 1
222 1
226 Benfluoranthene 1
B-22
-------
TABLE B-7
(cont'd)
RESULTS OF LRMS ANALYSIS OF CONCENTRATED
METHANOL EXTRACTS OF SORBENT TRAPS
Relative Abundance*
nw Species 3M1(B) 3M3
228 Chrysene 1
230 Terphenyl 1
232 C18H16 1
234 Butyl anthracene 1
236 1
Percent of total signal intensity for each sample.
B-23
-------
product of PVC waste combustion. The unidentified N and NO species are
also decomposition products, which appear at much higher temperatures
(<300°C) than is consistent with their low molecular weights. Similar
species were found in samples collected at Facility No. 8; this may in-
dicate that these are contaminants. The phenol, biphenol and biphenyl
may represent products arising from combustion of the PVC waste.
B.7 ANALYTICAL RESULTS FOR IMPINGERS
The impinger solutions originally charged at the start of sample
collection were 0.1 N sodium hydroxide for all stack samples and the
3M1(B) hot zone sample, and INsodium hydroxide for the 3M2, 3M3 and
3M4 hot zone samples. The results of chloride analyses on the recovered
solutions were:
Sample Cl", ppm
3M1(B)-S-I 0.9
3M1(B)-HZ-I 25.8
3M2-HZ-I 1445
3M3-S-I 19.8
3M3-HZ-I 2550
3M4-S-I 12.7
3M4-HZ-I 1548
B.8 ANALYTICAL RESULTS FOR SCRUBBER WATER
B.8.1. Quantitative Analysis of Organic Extracts by Gas Chromatography
Gas chromatographic analysis of methylene chloride extracts of the
influent and effluent scrubber water for the four tests at 3M, showed
that there were no significant peaks present in the 3M(2), 3M(3) or 3M4-
SO/CH2C12 samples that were not also present in the 3M1(B) SO/CH2C12,
3MO-SI/CH2C12 and distilled water blank extracts. The peaks which were
observed in all of these samples (retention times of 305, 395 (largest),
860, 1030, and 1200 sec.) were clearly due to trace impurities in the
methylene chloride solvent. Concentrations, as dichlorobenzene, were
<10 ppm for all except the 395 sec. peak which corresponded to about
100 ppm.
No species which would correspond to scrubbing of organic combustion
products were found.
B. 8.2. Qualitative Analysis of Organic Extracts by IR and LRMS
The IR Spectra of the concentrated methylene chloride extracts of
the scrubber water showed the presence of bands attributable to sulfate
SOir, which was apparently introduced during drying of the extracts.
B-24
-------
Other observations were aliphatic CH2>CH3 bands in the 3M3 (vs) and
3M4 (m) samples. Also seen were hydroxyl (m/s), carbonyl or aromatic
(w/m) in the 3M4 sample.
Species found by LRMS analysis of the extracts of the scrubber
water, and their relative abundances, were:
3MO-SI 3M1(B)-SO 3M3-SO
ClttH140 (di-tolyl ether) 72 94
Nonyl Phenol 70
Di butyl Phthalate 23 5
Dioctyl Phthalate 21
Fatty Acids 3 3 <1
Napthalene 5 2
Pentachlorophenol 1
Unknown 1 1
None of the species found appears to represent a true product of combus
tion.
B.8.3. Analysis of Chloride and Measurement of pH
6
6
2
2
1
B.9 ANALYTICAL RESULTS FOR GAS BULB SAMPLES
In the GC/MS analysis of the gas bulb samples, collected from the
by-pass line of the hydrocarbon analyses, the analyst was asked to search
specifically for the possible presence of chlorinated hydrocarbons. Noe
were detected. However, the compositon found for these samples by
GC/MS suggest that the gas bulbs had leaked to the extent that the ana-
lytical results were not meaningful. This may have been due to unequal
co-efficients of thermal expansion for Teflon (stopcocks) and glass
(bulb) if dramatic temperature changes occurred during sample shipment.
Sample
3MO-SI
3M1(B)-SO
3M2-SO
3M3-SO
3M4-SO
Cl~, ppm
10
15
414
719
816
B-25
-------
Concentration % Volume/Volume
Constituents 3M1(B) 3M2 3M3 3M4
Nitrogen 78+ 79+ 78+ 79+
Oxygen 19.6 19.2 19.5 16.1
Argon 0.91 0.91 0.89 0.89
Carbon Dioxide 1.27 0.44 1.33 3.8
Hydrogen ND 0.001 ND 0.001 ND 0.001 ND 0.001
Methane ND 0.002 ND 0.002 ND 0.002 ND 0.002
C2-C6 Hydrocarbons ND 0.002 ND 0.002 ND 0.002 ND 0.002
Chlorinated Hydrocarbons ND 0.002 ND 0.002 ND 0.002 ND 0.002
ND = none detected, less than
B-26
-------
APPENDIX C
ON-SITE ANALYSIS OF VINYL CHLORIDE MONOMER
These tests at the 3M company Chemolite plant rotary kiln facility
in November 1976 involved incineration of a polyvinyl chloride (PVC)
waste, containing an estimated 200 ppm of residual vinyl chloride
monomer (VCM). The concentrations of VCM in waste handling areas
and in incinerator emissions were monitored during the test. A gas
chromatograph (GC) was installed in the ADL sampling trailer and used
for rapid analysis of VCM concentrations in various gas samples
collected during the test program. Two types of samples were analyzed.
• Grab samples in 5-liter Tedlar® bags were taken
from areas where drums of waste were stored or handled.
The objective of this sampling was to ensure that per-
sonnel were not exposed to concentrations of VCM in
excess of the OSHA standard value of 1 ppm (for 8-hour
time weighted average).
• A sample of hot zone gas from the afterburner section
of the incinerator was continuously drawn into the
sampling trailer via a heat-traced line. At intervals,
portions of hot zone gas were injected into the GC to
ensure that incinerator emissions of VCM did not exceed
the EPA limit (for vinyl chloride and PVC plants) of
10 ppm discharge to ambient air.
Both types of samples were introduced to the Varian Model 1200 GC by
means of a 5 m£ gas sampling loop. The column used was 6' x 1/8"
stainless steel packed with 60/80 mesh Chromosorb 102. The column
oven was maintained isothermally at 100°C except as noted. The in-
jector temperature was 150°C. The flame ionization detector was
maintained at 200°C. The gas chromatograms were recorded on a
Linear model 355 recorder, 1 mv full scale, at a chart speed of
2 cm/min. Calibration of detector response was achieved by injecting
5 mn samples of a standard gas mixture containing 10 ppm of vinyl
chloride in nitrogen.
Copies of representative gas chromatograms are shown in Figure C-l.
The data show that the VCM concentration was £ 0.2 ppm for all samples
taken from the following areas:
• shipping trailer before unloading waste drums
C-l
-------
• shipping trailer after unloading and clean-up
• drum storage area
• drum loading room, with drums on conveyor and rims
popped, covers on
• scrap drum landfill area
t hot zone of incinerator during waste PVC burns.
The upper limit of 0.2 ppm of VCM is set by the noise level observed.
None of the samples listed above showed significantly higher FID
response intensity at the retention time, ofVCM (1.2 - 1.3 min) than
did samples taken during the background test or from a cylinder of
zero air.
Samples taken from the head space immediately above freshly opened drums
of waste showed VCM concentrations in excess of 100 ppm, however. When
this analysis was made, at the beginning of the first waste test
(November 5), the normal 3M drum feeding procedure was modified. The
rings on the drums were popped, but the covers were not removed before
the drums were fed to the incinerator. Analysis of suFsequent samples
showed that VCM levels in the loading area were <^ 0.2 ppm when the
modified procedure was employed.
Although the analyses showed VCM concentrations well below the OSHA
limit of 1 ppm, all personnel wore chemical cartridge respirators
(with charcoal canisters) when in the drum feeding area.
C-2
-------
Rt = 1.35 min
90
80
70
60
50
40
30
20
10
3M 4 11-8-76
Hot Zone Gas: 1012 am
sens: 4x1 for 3 min
Then 16x1
Inject
\
Rt = 3.58 min
3M 4 11-8-76
lOppm VCM Standard
sens: 8 x 1
Inject
Sensitivity Changed
FIGURE C-1 EXAMPLE CHROMATOGRAMS FOR ON-SITE GC ANALYSIS OF
VINYL CHLORIDE MONOMER
C-3
-------
APPENDIX D
HEAT BALANCE
The incineration of marginally combustible waste materials, such
as the PVC waste used in this test program, requires relatively large
quantities of external heat inputs in the form of purchased fuel. In
analyzing the technical and economic feasibility of incineration as
a waste destruction technique, it is often useful to quantify the heat
inputs to the incineration system and to account for the heat released
during the combustion process. In addition, a heat balance provides
a check on the accuracy of the test parameters measured.
A heat balance simply compares the measured and/or calculated heat
(enthalpy) input to the incinerator with the heat leaving the incinerator.
While conservation of energy dictates that these two values must be
identical, normally encountered variability in measurement and system
operation usually results in a discrepancy between the two numbers.
The wider the discrepancy, the more inaccuracy is present in the measurements.
The heat (or enthalpy) inputs to the Incinerator are the ambient
air, fuel, and the waste material. Conventionally, the enthalpy of the
ambient air is set at zero. For this heat balance, the boundaries are
set collectively around the primary and secondary combustion zones in
order to avoid the difficulties introduced by having to account for quench
water evaporated and the heat content of scrubber water. Since the temperatures
of the combustion zones are well over the boiling point of water, it is
necessary to use the lower heating value of the fuel 10,150 KCal/kg (18,310
8tu/lb) since no heat of condensation is involved. The lower heating
value of PVC waste has been calculated to be approximately 740 KCal/kg
(1340 Btu/lb).
The heat outputs are contained in the combustion gases and in the
heat losses through equipment surfaces. The enthalpy of the combustion
gases can be readily obtained by the gas flowrate, composition and temperature.
The heat losses through the equipment surfaces are not easily measured,
however, for incinerators it has been found that such heat losses are
on the order of 2% of the heat inputs.
The heat balances thus calculated are shown in Figures D-l through
D-4. The heat inputs and outputs agree rather well for a test of this
type; the percent error in each case was less than 10%.
As can be seen from Figures D-l through D-4, the PVC waste contributes
very little heat to the combustion process, mostly due to the high (72%)
moisture content of the waste.
D-l
-------
0!
o.
CD
C£
Z
UJ
CD
C
o
CQ
O o
LO O CM
vo o in
•> o •
o> • cr»
CM CM r*
i ii i
> O
>> *-••-
S- «J 4-> «/>
§•0 -o «/> o>
C C 3 C
•r- (C O -Q O
J_ OEM
CL O) O
CO O
«
vo
•
o
II
UJ
o r- >-
M _l
* z
Z Kf O
2 « _,
t— Ul »-H
CO CO O
O Z
op
U.
a.
O
C£
LU
UJ
E
U
VI
o
o
i— U.
•« o
00 O
CM Ki-
ll Jl
0)
i-
3
0) 4->
-(-> (O
ia i-
i. (U
1 1
iZ i—
o
it
Zr-
^ *r-
10 O
J=
•p i—
C (1)
LU 3
U_
•
CM
S-
JZ
S- 3
«• 1 \
J^ T-*
*^ CO
V)
^ i
o vo
vo vo
""L '
«* CO
1 II
2 £
CO i—
S- (0
O •!->
i— C
U. LU
^^
•o
0
IS.
\
$
V)
to
3
O
>
Q.
a
ro
J-
^Z
^^
3
•P
CQ
VO
vo
•
00
II
fe
a.
ZE
1— 1
>;
CD
C£.
LU
LU
•£
c3
»—
D-2
-------
E
CK-
o
CO
h-
a.
t—
Z3
O
>-
CD
a:
LU
z
LU
O
O
CM
M
CTi
to
CO
CO
to
CD
c:
o
•r"
•p
CO
_a
O
CM
II
CO
•P
"3
i.
O
Lu
•
LL.
o
O
O
O
A
CM
II
CO
i.
3
•P
to
s-
V
a.
(U
h-
t-
•C
*^
•p
CO
3E
^
»a-
VO
•
CO
r>.
n
>>
S
10
•P
c
LU
•P
CO
r*
Jam
CT
^3
o
S-
x:
h-
CO
CO
o
•P
m
ot
3:
LU r*. a.
z en
O r- CO
M =3
" _l
z m a.
2 * _,
I— LU I—I
00 CO O
CO LU —I
S! > LU
O O =5
O Z Lu
E
CH
O
CO
I—
o.
Z
H- 1
CD
C£.
LU
O
O
«
00
CM
21
LU
CO
•P
(O
3
O
•r- iL
«c
B
^
Lu
O
O
«* O
21 II
CO
•P >l
nj CX i—
S- r- -r-
co ^.
S- 3
x: -P
•^ co
CO
-O S
r- S
»a- oo
r-» LO
V£>
•> LO
«fr 00
u u
Q) >> 0)
•P S -P
«J r— CO
J- « «
S -C 3
0 -P
r- C 0
U. LU >
Q.
,
ro
s-
J=
1 — ,
CO
Xi
o
VO
00
A
I—
II
CO
•p
IO
0
LU
s_
N^^
3
•P
CO
s
S
LO
•
CM
II
>»
£
m
x:
4->
C
LU
i.
^M
3
•P
co
i
CO
o
•
CO
00
II
1-
a.
z
I"H
>-
CD
0£.
LU
Z
LU
_J
«c
s
1-
a
LU
0£
CD
D-3
-------
CJ
C£
LU
=Z
LU
O
CO
CO
O)
2
o
O
O
00
II
OJ
-t->
CO
ii
10 ex
CO
E
M-
0
(/)
t— O
=» O
Q. VD
I-H 0
CO
>-
O II
Q£
LU
LU
(U
4J
ea
1
S_| r-
•r-\ U_
N» 4J ^v. _3
V) CO W -t->
X) JD CO
vo o S
u_ CM r^ vo
0 O • 00 LO
O "CO • .
«*• o •* r»* r- CM
» II II
> , O) 0) >>
a. +J •»-> ex
- I/) U. Ul
o.
•
CO
s.
f"
^LB
^«.
5
^*
CO
i
CM
•
vo
r«.
II
5
o.
fr*4
>-
ci
LU
LU
^f
1
H-
Ul
QC
CJ9
D-4
-------
E
«*-
o
i/>
O
1—
Z)
a.
I—
o
>-
CD
C£
LU
Z
LU
O
r^
ti
10
M
0)
CO
re
CD
c
o
•r-
-P
I/)
3
X)
S
O
O
CO
II
0)
•P
re
S-
3.
0
n
,_J
u.
o
o
o
00
«l
f~
II
(U
s-
3
•P
re
(U
Q.
E
0)
I—
x:
>^
3
•P
CO
2-
2
VO
^J-
r^
00
u
>>
3.
re
.e
4->
C
LU
-C
CD
3
O
L.
JC
to
to
o
—1
•P
re
CQ
CO
CO
CTl
00
CD
CC.
LU
Z
LU
\
C
>> o
>, S?-r-
s_ re -P to
re ~o "O to
LU r^ a.
z cr>
o •— co
M =>
«• _J
Z CO O-
2 s. _j
(—•-.
co .a o
s: > LU
o o =>
o z u-
CD
a:
C
E
o
10
u.
O 0
o o
CO O
«• CO
co "
CO r—
II II
0)
S-
3
0) 4->
4-> re
re s-
$- O)
^— 5J
U- 1—
o
II
% =
re o
j=
•P r-
c cn
o «*
r— •
« CO
IT) >) CX -M
re •— 10
s. re re
3 -C 3
O -P
i— C CJ
U. LU >
a.
*
PO
S- J=
JC "*^
-v» 3
10 -P
J3 CO
i—
y
0 SE
vo
co in
* •
r- CM
« •
>
-P Q.
re i—
s- re
O -P
r— C
U. LU
S-
3=
3
^J
CO
31
51
0>
CT>
•
in
CT>
u
h-
CL.
•z.
1—4
>-
U3
QC
LU
Z.
LU
•<
0
I—
LU
a:.
CD
D-R
-------
APPENDIX E
ASSESSMENT OF ENVIRONMENTAL IMPACT OF
DESTROYING SOLID & LIQUID WASTES
AT
3-M COMPANY
CHEMOLITE PLANT
COTTAGE GROVE, MINNESOTA
The rotary kiln incinerator with secondary combustion will be evaluated
for its capability of thermally destroying a sludge from the manufacture
of PVC which contain low concentrations of vinyl chloride monomer. Approxi-
mately 120 barrels of the waste will be required for tests.
This full-scale rotary kiln incinerator was started up in 1972 at the
Chemolite Plant of 3-M for the purpose of thermally destroying chemical
wastes from their numerous plants in the area. The 11' 0 x 35' long rotary
kiln is fired with pumpable liquid wastes or fuel oil when extra energy is
required.
Gases for the rotary kiln enter a mixing chamber before going to the
secondary chamber where a final burner is used to control temperature. A
modified Venturi scrubber with once-through water flow is preceded by a
quench elbow and chamber and followed by a cascade tray demister. The
Venturi operates with approximately 30" water pressure differential. The
water saturated gases discharge to an unlined carbon steel stack.
Approximately 66,000 SCFM of air flow to the incinerator.
The major purpose of this elaborate setup is to permit adequate control
of the final stack gases when feeding drums of combustible substances.
Temperature variation in the rotary kiln are wide and erratic due to the
i
possibility of rapidly releasing a large quantity of combustibles from
a nearly full drum. Observations indicate that the discharges from the
rotary kiln section can be very smoky due both to the feed conditions and the
relatively low temperatures maintained - approximately 1300°F. However,
temperature recording instruments indicated surges as high as 2200°F. To
ameliorate these effects, drums are often fed to the incinerator with covers
in place; however, holes are punched in the covers to permit the wastes to
be gradually forced from the drums. In this way, the facility decreases
E-1
-------
the frequency of a black plume from the stack. A sampling station is
installed at the 80" level of the stack. The firing rate of pumpable
wastes or fuel oil through the two burners is controlled in order to
ameliorate the wild temperature surges caused by the slugs of wastes from
the drum. The final temperature is usually controlled slightly above 1800°F
within a relatively narrow range considering the erratic fluctuations occurring
•
in the rotary section. Particulate emissions are reported to be 0.1 grains/SCF
or less and within the emission limitation placed upon the facility.
The incinerator was designed for a heat release rate of 60 x 10 Btu
per hour in the rotary section and 30 x 10 Btu per hour in the secondary
section. It is estimated that the gas residence time in the rotary section
is 2 to 3 seconds and approximately 1 second at 1800°F in the secondary
chamber. Approximately 1500 gallons/minute of water goes to the scrubber,
quench elbows, etc. The water is not recirculated. It is neutralized to a
pH of about 7.5 by the use of liquid ammonia and discharged to a wastewater
treatment plant located on the premises. The effluent from the wastewater
treatment plant goes into the Mississippi River. The NPDES permit applica-
tion for the plant has not yet been granted.
The ash from the rotary kiln along with the burned out drums falls into
a water pit from which they are removed and sent to a local landfill contractor
(Phoenix, Inc.) for off-site disposal. The ash removed every 2-3 months
from the secondary chamber is disposed of in a similar manner. The solid
residues going into landfill are tested on an intermittent basis to insure
that they meet the current regulatory requirements.
The incinerator is located away from the manufacturing facilities on a
950 acre plot of land in the midst of an agricultural area. Considerable
space for storage of drummed wastes is available in a building which also
houses the feed end of the incinerator. Approximately 20 drums per hour wastes
are fed'to the incinerator and from 500 to 1000 drums may be on hand at any
one time. Outside storage of drums is available as are storage tanks for
liquids. Liquid wastes can be fed from two 3000 gallon tanks equipped with
mixers. In addition, six interim storage tanks approximately 10,000 gallons
each are available. Provisions have been made for. preventing leaks and spills
from migrating from the area and the general appearance of the area indicates
excellent housekeeping procedures are in effect. The facility operates
24-hours per day, seven days per week. It is closed down from 1 to 2 times
per year for routine maintenance on a scheduled basis. Each of the ten
E-2
-------
operators is rcqnl.rc.tl to hold ;i powerhouse operator'5; license. Elaborate
interlock exists for controlling the important variables in the incinerator
and written safety procedures have been established. Except for the NPDES
permit currently being negotiated, all applicable operating permits have been
obtained.
The environmental impact of the incinerator will be most apparent in the
visual discharges from the stack and in the nutrient content of scrubber waters
going ultimately to the Mississippi River. However, plans are underway to
utilize an alkali other than ammonia for pH control of the scrubber water.
There is not expected to be significant impact on the 1100 people employed
at the Chemolite plant because airborne emissions would be carried over the
plant and water effluents go into a wastewater treatment system. The incinerator
'is a moderately noisy operation but only a limited number of personnel are in
the immediate area at a time. Similarly any odors are expected to be localized
and not have any noticeable effects on the nearest residents located approxi-
mately one kilometer from the plant in the direction away from the prevailing
northwest wind. No complaints of odors have ever been received because of
operation of the incinerator. The relatively remote location of the plant
indicates that the approximate 100 trucks/day (in addition to employees'
automobiles) will not have a significant impact on local traffic patterns.
The company-owned acreage is bounded on one side by the Mississippi River
and the terrain is flat to slightly rolling. Farmlands surround the site
some portion of which are leased to local farmers. There are no heavily
forested areas; consequently, the predominant local wild life is rabbits, birds,
and other small animals. The emissions from the 8" 0, approximately 200' high
stack will be widely visible as a steam plume. The effluent from the waste-
water treatment plant will be most affected by the dissolved solids contained
in the neutralized wastewaters. When ammonia is used as the neutralizing agent,
the scrubber water may serve as a nutrient source for the wastewater treatment
plant; however, it is also probable that the ammonia may cause a large load
of nutrient chemicals to enter the river and it is surmised that this may be
one of the reasons why they are replacing the ammonia. Because the effluent
enters a flowing river, there will be dust impact than if entering a lake.
It is believed that the environmental impact from air and water discharges
will be minimal and will be within acceptable limits.
E-3
-------
The storage and handling of the approximately 120 barrels of PVC wastes
will be carried out under the well-established 3-M procedures. Indoor storage
of the drums on a concrete floor immediately adjacent to the test facility is
contemplated. Any leakage or spillage will be absorbed with a solid absorbent
or otherwise put into containers for ultimate incineration. The following
components of the PVC wastes have been identified and data on their toxicity
are presented.
Components of PVr Wastes*
Vinyl Chloride Monomer (Monochloroethylene)
Inhalation - human - TCLo - 20 ppm
USDS - Air - .TWA - 1 ppm
Cl - 5 ppm
Vinylidene Chloride (Dichloroethylene)
Inhalation - rat - LCLo - 10,000 ppm/24 hr.
. Oral - dog - LDLo - 5750 rag/kg
Trichloroethylene
Oral - human - LDLo - 857 mg/kg
Inhalation - human - TCLo - 110 ppm/8 liters
USDS - Air - TWA - 100 ppm
Air - C - 200 ppm
Criteria Document Recommended Standard
TWA - 100 ppm
C - 150 ppm
Note: Above data taken from Registry of Toxic Substances - 1975 Edition
* Definition of terms
Cl - Ceiling Concentration
i
C - Continuous
TWA - Time weighted average
USDS - U. S. Occupational Standard
TCLo - lowest published toxic concentration
LDLo - lowest published lethal dose
The vinyl chloride monomer, expected to be contained in the wastes at
a level of about 100 ppm, is by far the substance of greatest concern. Because
the wastes will be fed to the incinerator in a manner which will insure minimum
E-4
-------
personal contact, it is believed that ' here is little potenti.nl for any
of these substances reaching concentration levels which would be of
occupational safety and health concern and will certainly not be of any
significant environmental impact. The disposal of residuals such as
ash and burned out barrels will follow the established 3-M procedures.
Because of the long-term and well-established 3-M procedures for
handling chemical wastes, it is believed that these tests will have
minimum potential for impacting the environment.
ya!467d
SW-122c.4
E-5
------- |