PrepubliaatJon issue for EPA
and State Solid Waste Management Agenoiee
BURNING WASTE CHLORINATED HYDROCARBONS
IN A CEMENT .KILN
This report (SU-l4?a) describee work performed
for the Office of Solid waste under contract no. 68-01-2966
and ie reproduced OB received from the contractor.
The findings should be attributed to the contractor
and not to the Office of Solid Waste.
The reader ie advised to utilize the information
and data herein wiih caution and judgement.
Copies will be available from the
National Technical Information Service
U.S. Department of Commerce
Springfield, Virginia 22161
U.S. ENVIRONMENTAL PROTECTION AGENCY
1978
-------
This report has been reviewed by the U.S. Environmental Protection
Agency and approved for publication. Its publication does not signify
that the contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of commercial products
constitute endorsement or recommendation for use by the U.S. Government.
An environmental protection publication (SW-l47c) in the sol^d waste
management series.
-------
Propublioation issue for EPA librcari.«s
and State Solid Waste Manag0m*nt Ag»nai»8
BURNING WASTE CHLORINATED HYDROCARBONS
IN A CEMENT .KILN
This report (SW-l47a) describee work performed
for -the Office of Solid waste wider contract no. 68-01-2966
and is reproduced as received from the contractor.
The findings should be attributed to the contractor
and not to the Office of Solid Waste.
The reader is advised to utilize the information
and data herein wii;h caution and Judgement.
Copies will be available from the
National Technical Information Service
U.S. Department of Commerce
Springfield, Virginia 22161
U.S. ENVIRONMENTAL PROTECTION AGENCY
1978
-------
This report has been reviewed by the U.S. Environmental Protection
Agency and approved for publication. Its publication does not signify
that the contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency, nor does mention of commercial products
constitute endorsement or recommendation for use by the U.S. Government.
An environmental protection publication (SW-W7c) in the sol^d waste
management series.
-------
ABSTRACT
An experimental program was carried out in 1975/76 at the St.
Lawrence Cement Co., Mississauga, Ontario in which waste chlorinated hydro-
carbons, containing up to about 46 weight percent chlorine, were burned
in a rotary cement kiln. The chlorinated hydrocarbons were burned in three
distinct phases of increasing difficulty of combustion. Materials burned
included mixtures of ethylene dichloride, chlorotoluene and up to approxima-
tely 50 percent polychlorinated biphenyls (PCS).
These materials were destroyed in the cement kiln with at least
99-98 percent efficiency in all cases. Emissions of high molecular weight
chlorinated hydrocarbons were not detected. Three light chlorinated hydro-
carbons, dlchloromethane, chloroform and carbon tetrachloride, were found
in the emissions in the part per billion or lower range. The quantity of
precipitator dust requiring disposal, as well as emissions of particulate
matter, increased during the test.
The chlorine input from the chlorinated hydrocarbon waste was up
to about 0.8 weight percent relative to clinker and this effectively reduced
the alkali concentration of the clinker in direct stoichiometric proportion,
A reduction in fossil fuels used while burning chlorinated hydrocarbons was
noted.
111
-------
TABLE OF CONTENTS
Page
ABSTRACT '
.TABLE OF CONTENTS Mi
List of Figures v
List of Tables vii
EXECUTIVE OUTLINE xi
SUMMARY xi i i
RECOMMENDATIONS xv
LIST OF ABBREVIATIONS xvi
1. INTRODUCTION '
2. CEMENT MANUFACTURE 2
2.1 General Principles 2
2.2 Effect of Alkalies 5
2.3 The St. Lawrence Cement Co. 6
2.3.1 Relevant unit processes - wet process kiln 6
2.3.2 Relevant unit processes - suspension preheater kiln 6
3. CONSIDERATIONS AT THE PROGRAM PLANNING STAGE 10
4. TRIAL ON THE SUSPENSION PREHEATER KILN 12
5. TRIAL ON THE WET PROCESS KILN 15
5.1 Discussion ]5
5.2 Organic Chloride Waste Burned 17
5.3 Emissions 22
5.3.1 Free chlorine and hydrogen chloride 22
5.3.2 Gaseous organic compounds 22
5.3.2.1 Desorbed samples 23
5.3.2.2 Organic solvents extracted samples 25
5.3.2.3 Results obtained by the participating laboratories 26
5.3.2.4 Estimated minimum combustion efficiencies 27
5.3.3 Particulate emissions 28
5.k Mass Balance on Wet Kiln 30
5.4.1 Significance of the mass balance 30
Vi
-------
TABLE OF CONTENTS (CONT'D)
Page
5-^.2 Chlorine and potassium retained 31
6. CONSIDERATIONS ON BURNING CHLORINATED HYDROCARBON
WASTES IN A CEMENT KILN 3k
6.1 Effect on Production 34
6.2 Alkali Reduction While Burning Chlorinated Hydrocarbon
Wastes 34
6.3 Heat Recovery from Chlorinated Hydrocarbon Wastes 37
6.k Cement Quality 38
6.5 Extrapolation to Other Kiln Types 38
6.6 Comparison of Cement Kiln Burning with Other Uses and
Disposal Methods for Waste Chlorinated Hydrocarbons 39
7. CONCLUSIONS 42
REFERENCES 43
ACKNOWLEDGEMENTS 46
APPENDIX A - QUANTIFYING, SAMPLING AND ANALYSIS OF PROCESS
MATERIALS 49
APPENDIX B - ANALYTICAL DATA, CALCULATION AND DETAILS OF
EXPERIMENT ON THE SUSPENSION PREHEATER KILN 93
APPENDIX C - RESULTS AND CALCULATIONS FOR WET PROCESS SYSTEM 107
APPENDIX D - QUALITY OF CEMENT PRODUCED 129
APPENDIX E - EQUIPMENT DESCRIPTION AND ASSOCIATED ECONOMICS 137
APPENDIX F - ONTARIO MINISTRY OF THE ENVIRONMENT EMISSION
GUIDELINES AND ANALYTICAL SUPPORT 149
APPENDIX G - LABORATORY ANALYSIS RESULTS FROM THE ST. LAWRENCE
CEMENT FACILITY TEST (TRW Systems Group) 165
APPENDIX H - DEVELOPMENT, CONSTRUCTION AND EVALUATION OF A
COLLECTION SYSTEM FOR LOW MOLECULAR WEIGHT
HYDROCARBONS (Ontario Research Foundation) 207
APPENDIX I - GC/MS/COMPUTER DETERMINATION OF CHLORINATED
HYDROCARBONS AND PCB's (Air Pollution Control
Directorate, EPS, Environment Canada) 221
-------
LIST OF FIGURES
Figure Page
1 Wet Process Kiln 3
2 Dry Process Kiln k
3 Principle of Fuller-Humboldt Suspension Preheater and
By-pass 7
k Alkali By-pass 9
A.I Schematic of the Material Balance 50
A.2 Schematic of Port Locations 59
A.3 Gas Flow Distribution at Sampling Points 60
A.k Grab Bag Sampling Equipment 61
A. 5 Particulate Sampling Train 62
A.6 Gaseous Sampling Train 65
A.7 Gas Chromatographic Profile from Flame lonization
Detector for Chlorinated Aliphatics (WBA) Sample Feed 73
A.8 Gas Chromatographic Profile from Flame lonization
Detector for Chlorinated Aliphatics plus Aromatics
and Alicyclics (WBB) Sample Feed /A
A.9 Gas Chromatographic Profile from Flame lonization
Detector for Chlorinated Aliphatics plus Aromatics,
Alicyclics and Polychlorinated Biphenyls (WBC) Sample
Feed 75
A. 10 Gas Chromatographic Profile from Electron Capture
Detector for WBB Sample Feed 76
A.11 Gas Chromatographic Profiles from Electron Capture
Detector for WBC Sample Feed and Standard Arocolor 1242 77
A.12 Gas Chromatographic Profiles from Electron Capture
Detector for Low Molecular Weight Chlorinated Hydro-
carbons and for BLB and WBC Test Samples 79
A. 13 Gas Chromatographic Profiles from Flame lonization
Detector for Low Molecular Weight Chlorinated
Hydrocarbons 80
A.\k Gas Chromatographic Profile from Electron Capture
Detector for Impinger Extract from BLA Test 3 81
Vi
-------
LIST OF FIGURES (CONT'D)
Figure Page
A. 15 Gas Chromatographic Profile from Electron Capture
Detector for Impinger Extract from WBC Test 3 82
A. 16 Gas Chromatographic Profile from Electron Capture
Detector for Impinger Extract from WBC Test 3 after
Cleanup and Separation 83
B.I Chlorine Level in Stage IV, June 3, 1975 102
B.2 Chlorine Level in Stage IV, June 10, 1975 103
E.I Schematic Diagram of Basic Chlorinated Hydrocarbon
Feed System HO
E.2 Schematic Diagram of Chlorinated Hydrocarbon Facilities \k\
F.I Representative Bar Chromatogram 156
F.2 Computer Reconstructed Bar Chromatograms for PCB Fuel
and Aromatic Fuel plus Arocolor 1242 158
F.3 Gas Chromatogram from GC/HS Analysis of Sample
PCB Fuel 159
G.I TRW Sample Coding System 169
G.2 Plan for the Combination of ORF Solvent Extracts 171
G.3 Desorption System for Chromosorb 102 Tubes (TRW) 171
H.I ORF Test Duct Schematic 212
V11
-------
LIST OF TABLES
Table
1 Composition of AUphatics (WBA) 19
2 Composition of Aromatics plus Complex (WBB) 20
3 Composition of Aromatics plus PCB's (WBC) 21
k Gas Sample Volumes and Sample Concentration Factors 2k
5 Estimated Kiln Emission Concentrations (GC-EC) for
Specific Volatile Organochlorine Compounds 25
6 Estimated Minimum Combustion Efficiencies for Each
Waste Burn 2&
7 Summary of Particulate Test Data 29
8 Accumulated Mass Balance for Chlorine 32
9 Accumulated Mass Balance for K.O 32
10 Average Reduction in K-0 Content of Clinker 35
11 Average Dust Discharged 36
12 Recovery of Btu from Chlorinated Hydrocarbons 38
A.I Process Materials Studied and Approximate Normal
Production Quantities ^9
A.2 Quantities of Aliphatic Mixture Burned Daily 51*
A.3 Quantities of Aromatic plus Complex Mixtures Burned
Daily 55
A.** Quantities of PCS Mixture Burned Daily 57
A.5 GC Analysis - System A 69
A.6 GC Analysis - System B 69
A.7 GC Analysis - System C 70
A.8 Gravimetric Dust Analyses 85
A.9 Results from Leco Induction Furnace Analyses 86
A.10 Results from Atomic Absorption Analyses 87
A.11 Least Squares Data for Calibration Lines 88
Vi 11
-------
LIST OF TABLES (CONT'D)
Table Page
B.I Percent Bypass Gas Required to Maintain Chloride Levels 95
8.2 Results from Analyses of Dry Process Kiln Raw Feeds 98
B.3 Results from Analyses of Dry Process Kiln Clinker 99
B.l» Results from Analyses of Stage IV Dusts 100
C.I Results from Analyses of Clinker Samples 109
C.2 Results from Analyses of Slurry Feed Samples 111
C.3 Results from Analyses of Discard Dust 113
C.k Results from Analyses of Return Dust 115
C.5 Btu and Chlorine Content of Chlorinated Hydrocarbons 117
C.6 Btu, S and Cl Analyses from Mo. 6 Fuel Oil 118
C-7 Daily Record of Production and Materials Consumption 120
C.8 Material Balance for Chlorine 122
C.9 Material Balance for 1^0 125
D.I Cements from Clinker Produced During Baseline Burn 131
D.2 Cements from Clinker Produced During Aromatic plus
Complex Chlorinated Hydrocarbon Burn 132
D.3 Cements from Clinker Produced During Polychlorinated
Biphenyl Burn 133
F.I MOE Specifications Applied to Waste Burns 150
F.2 St. Lawrence Cement Waste Burn Experiments - Test 1
Chromosorb Adsorption Tube Analysis 153
F.3 St. Lawrence Cement Waste Burn Experiments - Test 2
Chromosorb Adsorption Analysis 153
F.I* St. Lawrence Cement Waste Burn Experiments - Test 3
Chromosorb Adsorption Analysis 153
F.5 Fuel Sample Identification 155
F.6 Gas Chromatograph Conditions 155
ix
-------
LIST OF TABLES (CONT'D)
Table Page
F.7 MS Data from GC Analysis of Sample PCB Fuel 160
F. 8 Area Counts 161
G.I Summary of Samples Received from Each Waste Burn (TRW) 168
G.2 Organic Composition of Aromatic Waste by GC/MS 178
G.3 Trace Metals in the Chlorinated Aromatic Waste by SSMS 179
G.4 Organic Composition of PCB Waste by GC/MS 181
G.5 Trace Metals in the PCB Waste by SSMS 182
G.6 Summary of Organic Qualitative Survey Analysis of Sample
Extracts (TRW) 185
G.7 Approximate Constituent Levels of Trace Vapours Desorbed
from Sorbent Tube Samples by LRMS 188
G.8 Summary of the Interpretation of LRMS Spectra for Trace
Vapours Desorbed from Sorbent Tube Samples 189
G.9 Results and Detection Limits from GC/FID Analysis of the
Concentrated Extracts 190
G.10 Results and Detection Limits from GC/ECD Analysis of the
Concentrated Extracts 192
G.ll Results and Detection Limits from GC/ECD Analysis of the
Unconcentrated Extracts 19^
G.12 Trace Metal Semi-Quantitative Survey of Filter Digests
by ICPOES 195
G.13 Limits of Detection for Undetected Elements by ICPOES 196
G.14 Concentration of Trace Metals in Effluent Gas Particulate
Matter by AAS 197
G.15 Concentration of Trace Metals in Aqueous Samples by AAS 197
G.I6 Results of Organic Survey Analysis on Clinker Product anH
Disposal Dust Samples 199
G.17 Results and Detection Limits from GC/FID Analysis 201
G.18 Results and Detection Limits from GC/ECD Analysis 201
-------
UST OF TABLES (CONT'D)
Table
G.19 Selected Trace Metals In SLC Clinker Product and
Discard Dust Samples by SSMS 202
X1
-------
EXECUTIVE OUTLINE
Experiments using chlorinated hydrocarbon wastes in the manu-
facture of cement appear to have transformed a difficult waste disposal
problem into a solution which is not only economically and environmentally
satisfactory, but has a beneficial effect on the quality of the cement.
Safe disposal of the large volumes of the chlorinated hydrocarbon
wastes which are generated in Canada each year is a difficult environmental
problem. It is estimated conservatively that Canada generates some 25-30
million pounds annually of these highly toxic and persistent wastes. This
figure does not include many toxic compounds such as insecticides and
polychlorinated biphenyls (PCB's).
Environmentally safe disposal is difficult and expensive. Disposal
on land, or underground, requires special and expensive precautions to
prevent leaching into waterways. The favoured method of disposal, inci-
neration with recovery of hydrogen chloride, can be very costly. Without
recovery, scrubbing equipment is required to remove hydrogen chloride from
the products of combustion to control its emission. This, in turn, usually
necessitates a satisfactory disposal method for the scrubber liquid. Va-
luable fuel must be burned to maintain combustion while incinerating
chlorinated hydrocarbon wastes, as extremely high temperatures with long
residence times are required for their thermal destruction.
A long high-temperature flame is required in a cement kiln to
achieve the desired product quality. During normal" operations, the thermal
conditions that are necessary effectively consume the toxic materials. The
kiln also contains a considerable quantity of lime and thus has an intensive
"scrubbing" action.
Recovery and re-use of hydrocarbon wastes is not always felt to
be economic. The recycling of solvents is practiced when it is economically
and technically feasible. However, this entails the control of segregation,
storage, collection and ultimate treatment of the various hydrocarbon streams.
The report strongly urges that combustible liquid wastes, many of
which are persistant environmental contaminants, be combined by r.^ans of a
Xi i
-------
single economical recovery system which would collect and later use the
wastes as supplementary fuel for cement manufacture.
The experiments described In this report were monitored by
the Department of Fisheries and Environment, the Ontario Ministry of the
Environment, the United States Environmental Protection Agency, and the
Ontario Research Foundation.
For these experiments, industrial chlorinated hydrocarbon wastes
including polychlorinated biphenyls (PCB's) were burned during the
commercial manufacture of cement.
This not only utilized the thermal value of the wastes, with a
reduction in oil consumption, but the results showed that almost all the
toxic wastes were completely destroyed in the kiln. Emissions of toxic
materials into the atmosphere were negligible.
Calcium chloride is often used in cement manufacture to reduce
the alkali content of the product... The only apparent effects that the
experimental burn had on the quality of the cement were the beneficial
effects due to the incidental addition of chloride ion.
X111
-------
SUMMARY
Chlorinated hydrocarbon wastes were burned in a carefully con-
trolled experimental trial as a partial fuel at the St. Lawrence Cement
Co., Mississauga, Ontario. The experiment was conducted to determine
whether chlorinated hydrocarbon wastes could be burned in a cement kiln
without adverse effects on air pollution levels.
The wastes used included a variety of chlorinated hydrocarbons
in the series of program phases designed to progress from easily combusted
chlorinated hydrocarbons to those which are combusted only with difficulty.
The last phase consisted mainly of polychlorinated biphenyl wastes. These
materials were processed and formulated from chemical wastes as required
for the different phases and supplied by Chemtrol Pollution Services Inc.
Atmospheric emission measurements were made before, during and
after the burning of each blend of chlorinated waste. Two methods of emis-
sions sampling were used during each phase, the method normally used for
measuring emissions of particulate matter, and a sampling train designed
specially for determining emissions of organic material. All samples
from both systems were analyzed for unburned chlorinated hydrocarbons.
It was concluded that the combustion efficiency was at least 99-986
percent for the chlorinated hydrocarbons. Approximately 50 ppb of
volatile low molecular weight compounds were found in the emission samples.
There were no detectable quantities of high molecular weight chlorinated
compounds in the stack gases.
A mass balance was carried out on chlorine and potassium. This
showed that the chlorine input as chlorinated hydrocarbon was completely
reacted with the process solids.
The alkali content of the clinker showed a reduction which
corresponded exactly with the quantity of chlorine input to the system.
This agreement further confirms the data from emission testing and the mass
balance.
While burning chlorinated hydrocarbons with approximately 40
percent chlorine, a decrease in oil consumption equivalent to 65 percent
of the Btu content of the chlorinated hydrocarbons was obtained.
The only differences in the quality of clinker produced while
burning chlorinated hydrocarbons were the beneficial effects which were
expected through the reduction in alkali content.
xiv
-------
It MAS concluded that chlorinated hydrocarbon vtastes
used in cement kilns, replacing other forms of chlorine used for
reduction of alkali content. A small proportion of fossil fuel required
for cement manufacture is conserved through use of these materials.
Burning chlorinated hydrocarbon wastes is considered a valuable means of
destroying persistent and toxic forms of pollutants while recovering useful
heat values.
XV
-------
RECOMMENDATIONS
The experiments have shown that there is virtually no adverse
effect on air pollution levels by burning chlorinated hydrocarbon wastes
in a cement kiln. These wastes include polychlorInated biphenyls and other
materials which are difficult to destroy. When other methods, such as
incineration, are used hydrogen chloride and chlorine may be emitted and,
if the incinerator operation is poorly controlled, uncombusted material
may be released into the environment.
The following recommendations are the outcome of the present
report:
(1) Burning chlorinated hydrocarbon wastes in a cement kiln is considered
a valuable means of destruction of persistent and toxic materials which
are members of this family of compounds. Since flame temperatures and
retention times are similar for all cement plants regardless of type
of kiln or fuel used, chlorinated hydrocarbons should be destroyed
in any cement kiln. The feasibility of doing so in a particular kiln
installation can be determined by a technical and economic review.
(2) Due to lack of familiarity with organic chemicals, it is considered
essential that instructions on safe handling procedures be given to
cement industry personnel.
(3) Since problems such as precipitation, solidification, heat or gas release
can arise through mixing incompatible waste materials in storage tank,
it is considered advisable to obtain such materials from one reliable
source of supply at any given time.
XVI
-------
LIST OF ABBREVIATIONS
APCD Air Pollution Control Directorate (EPS)
BLA Baseline A, designates emission tests before waste
burns
BLB Baseline B, designates emission tests after waste
burns
DCM Dichloromethane
EC Electron capture (detector)
EPA Environmental Protection Agency (U.S.)
EPS Environmental Protection Service (Fisheries and Environment
Canada)
FID Flame ionization detector
GC Gas chromatography
MOE (Ontario) Ministry of the Environment
MS Mass spectrometry
ORF Ontario Research Foundation
PCB Poiychlorinated btphenyl
SLC St. Lawrence Cement Co.
TRW TRW Systems Group
Tl, T2, T3 Test one, Test two, Test three
WBA Waste Burn A (chlorinated aliphatlcs)
WBB Waste Burn B (WBA plus chlorinated aromatics and alicyclics)
WBC Waste Burn C (WBB plus poly chlorinated biphenyls)
XRF X-ray fluorescence
XVH
-------
I INTRODUCTION
In Canada each year an estimated 2J-30 million pounds of chlori-
nated hydrocarbon wastes require disposal or destruction, 17-20 million
pounds being generated in Ontario [:]. These figures are based on 5-6fc
of annual production and may be conservative. Experience in Europe
indicates that 10% of production is a more reliable estimate of waste
material.
There are waste streams from plants manufacturing, processing,
or using chlorinated hydrocarbons which must be disposed of, the method
of disposal frequently being incineration. Chlorinated wastes other than
those directly from chemical plants also present a serious disposal problem.
Among these latter chemicals are polychlorinated biphenyls (PCB's) and
insecticides which may require disposal.
Many of these wastes are toxic and persistent, and all pose a
serious disposal problem. Among the methods presently used are
incineration and land disposal. Incineration with recovery of hydrogen
chloride can be costly. Without recovery, combustion gases must be scrubbed,
thereby generating a liquid waste requiring disposal. Both incineration me-
thods require additional fuel. Deep welling and other similar methods of
disposal are environmentally unsound because of the risk of water contamina-
tion.
In cement manufacture, the kiln operates at higher temperatures
and for longer residence times than those used in incinerators for des-
truction of these waste materials. It is also common practice in the
cement industry to add chlorides to the kiln to reduce the alkali concen-
tration of the final product. Use of chlorinated hydrocarbon wastes in
a cement kiln would provide useful recovery of chlorine and energy and,
at the same time, solve a serious disposal problem.
The present research program was carried out to determine
whether waste chlorinated hydrocarbons can be burned in a rotary cement
kiln without causing adverse air pollution. The approach taken was to
analyse stack emissions for uncombusted chlorinated hydrocarbon. A mate-
rial balance on chlorine was undertaken to confirm the emission findings.
-------
2 CEMENT MANUFACTURE
2.1 General Principles
While a variety of raw materials may be used in cement manufac-
ture, materials containing calcium, silicon, aluminum and iron without
an excess of certain other elements are required. These materials are
ground to a fine powder called raw meal, the chemical composition of which
is carefully controlled by proper blending of the various materials. Nor-
mally, blending is achieved by grinding all the raw materials together
(intergrinding). Raw meals required for wet and dry processes are similar
except that the raw meal for the wet process is in the form of a slurry
containing approximately 35? water, while raw meal for the dry process
contains less than 0.5$ water.
The raw meal is fed into the kiln (see Figures 1 and 2) and is
burned in the kiln to produce an intermediate product called clinker.
The kiln slopes towards the burning zone and rotates slowly,
causing the raw material to gradually move into the burning zone. Reactions
which occur during gradual heating in the kiln are: evaporation of free
water, evolution of combined water, evolution of carbon dioxide from
carbonates, and combination of lime with silica, alumina and iron to form
the desired compounds in the clinker. These reactions require a final
material temperature of H»50°C (2650°F). Four main compounds are present
in Portland cement clinker:
Common Abbreviations
Used In The
Name Of Compound Chemical Formula Cement Industry
Tricalcium Silicate 3 CaO*SiO_ C,S
Dicalcium Silicate 2 CaO-SiO. C?S
Tricalcium Aluminate 3 CaO'Al.O. C.A
Tetracalcium Aluminoferri te 4 CaO-Al-OyFe 0. C.AF
Minor compounds are also formed In clinker, commonly magnesia
(MgO), potassium sulfate (K-SO.) and sodium sulfate (Na SO.).
Traces of other elements present in either the raw materials or
fuel are also found in clinker. Upon leaving the kiln, the clinker is
-------
www
L*
VX^y,'"'.^ "'^' •_/-
YYY
•TT
1-KILN
2-SLURRY FEED
4-PRECIPITATOR DUST SCREW
3-PRECIPITATOR
5-DUST RETURN
6- FUEL
7- CLINKER COOLER
8- CLINKER
9- FILTER
WET PROCESS KILN
FIGURE 1.
-------
YYYr
•- - i
1 RAW MEAL FEED
2 STAGE I.
3 STAGE II.
4 STAGE III.
5 STAGE IV.
6 KILN
7 CLINKER COOLER
8 CLINKER
9 FUEL
A FILTER
B PRECIPITATOR
C BY-PASS
DRY PROCESS KILN
-------
rapidly cooled to avoid undesirable crystal forms of the above compounds.
After cooling, the clinker is ground and blended, normally by intergrinding,
with gypsum to a fine powder. The final product, called Portland cement, is
the basic ingredient of concrete.
In the burning process, considerable CO- is driven from the raw
meal. Any elements not driven off are increased in the clinker in propor-
tion to the quantity of CO- evolved. Throughout most of this report, the
analyses have been reported on a natural basis, i.e. reported results are
actual concentrations present in samples as received. In some cases,
results have been reported on an ignited basis, i.e. on CO free basis,
and have been indicated as such in the report.
2.2 Effect of Alkalies
In cement manufacture, reference to alkalies implies potassium
and sodium oxide (K20 and Na20). Both alkalies are frequently combined and
reported as equivalent Na-0 for purposes of specification. The raw materials
at St. Lawrence Cement are such that Na«0 is low and practically constant
(see Table A.10). For this reason, only potassium oxide (K.O) is consi-
dered in detail in this report.
The effect of alkalies on cement quality has been well documented
[2, 3, *0. While some alkali may be desirable for early strength develop-
ment [3], an excess can be problematic. The most extensively reported of
these problems is the a 1ka11-aggregate reaction [5, 6]. Certain glassy
silicates and some dolomites react slowly with alkalies and cause expansion
and disruption of concrete. It has been found by experience that cement
containing less than 0.60% total alkalies reported as Na.O performs satis-
factorily with such aggregates. This specification is imposed with suffi-
cient frequency in ths United States that it is found as an optional spe-
cification under ASTM C-150 [?]. Strength attained and setting characte-
ristics are also related to the alkali concentration of the cement [3, 8].
It is common practice in the cement industry to add chlorides
such as calcium chloride or waste hydrochloric acid to the raw meal to
reduce alkalies [9, 10]. Alkalies are normally present as sulphates which
at kiln operating temperatures are not readily volatilized, but are retained
in the clinker. Alkali chlorides are volatile at normal kiln operating
-------
temperatures. These are evolved from the material in the kiln and carried
in the gas stream to the precipitator. The high alkali dust from the pre-
cipitator can then be discarded.
2.3 The St. Lawrence Cement Co.
The St. Lawrence Cement Co., Mississauga Plant, has a nominal
production capacity of 1,750,000 short tons per year. The company operates
two wet processes and one dry process suspension preheater kiln. General
information on the company and the plant has been published in two articles
[11, 12].
2.3.1 Relevant unit processes - wet process kiln
Each of the two wet process kilns are dumbel 1 -shaped All is
Chalmers kilns J»02' long with a diameter of 11'6", having nominal capacity
of 1050 short tons per day (see Figure 1). The chain system in the drying
zone has 57 tons of loose hung carbon steel chains with a radiation curtain
at the front (flame end) of stainless steel chains. The chain system extends
through 8?' of kiln length. The slurry feed system is a bucket wheel con-
veyor with a variable speed drive taking slurry from a constant level box.
Gases from each kiln (maximum capacity 150,000 CFM at 450 F) pass through
a six-section electrostatic precipitator. Gases from the precipitators
are exhausted via a common stack 55^' in height with 13' exit inside dia-
meter. No. 6 fuel oil is burned in a single burner at the centre of the
burner pipe. For the test, chlorinated hydrocarbons were fed just above
and to one side of centre using different size nozzles for proper atomi-
zation at different flow rates. A detailed description of the chlorinated
hydrocarbon system is given in Appendix E.
2.3-2 Relevant unit processes -suspension preheater kiln
The kiln (Figure 2) is a 1?" x 276' Traylor unit normally fired
through three nozzles with No. 6 fuel oil. For the test, chlorinated
hydrocarbons were injected via a nozzle at the centre of the triangle
formed by the three oil nozzles.
The suspension preheater (Figure 3) consists of a system of
cyclones through which hot exit gases from the kiln are drawn by a fan.
The raw meal passes through the system in counterflow to the gas. Kiln
feed is introduced into the duct between the first and second stage cyclones.
It is swept with the hot exhaust gases into the uppermost (Stage I) cyclones
-------
MEAL FEED
II
III
IV
2 STAGE
3 STAGE
STAGE
5 STAGE
KILN
KILN EXH.FAN
8 PRECIPITATOR
9 DUST RETURN
10 CONDITIONING TOWER
11 BY-PASC PRECIPITATOR
12 DUST DISPOSAL
13 DUST DISPOSAL
U WATER
MATERIAL FLOW
[> GAS FLOW
PRINCIPLE OF FULLER-HUMBOLDT
SUSPENSION PREHEATER
AND
BY- PASS
FIGURE 3.
-------
where gas and material are separated. The raw feed from the cyclone drops
into the duct between the second and third stage cyclones and is again
suspended and separated. This procedure, being swept up with hot gases
and then being dropped into the stream entering the next lowest stage,
is repeated in Stages III and IV before the partially calcined feed enters
the kiln. The average retention time of the system is approximately
30 seconds. A detailed description of the Humboldt preheater has been
given by G.A. Schroth [13] from which article these notes have been pre-
pared. The raw material entering Stage I is preheated to approximately
300°C (600°F), while the gas temperature drops from 530°C (990°F) to 3*»0°C
(650°F). At each stage, corresponding heat exhanges occur such that the
material enters the rotary kiln at approximately 800°C (1*475°F) having
been partially decarbonated. The gas temperature at the point of exit
from the kiln into the preheater is 10l»0°C (1900°F) to 1090°C (2000°F).
The St. Lawrence Cement Co. preheater is a dual Fuller-Humboldt unit
with four stages in each.
The alkali bypass system (Figure 4) fs an important unit
in relation to this study. In common with other plants using suspension
preheater systems, special measures have to be taken to reduce the buildup
of chlorides and alkalies in the system. At the St. Lawrence Cement Co.
plant, a system is in use where a fraction of the kiln exhaust gases is
removed from the preheater and passed through a conditioning tower. In
the conditioning tower, water is sprayed into the gas stream to lower the
temperature and condition the gases for precipitation. Concurrently, the
gas velocity is reduced, since the cross section of the tower is greater than
that of the bypass duct. The net effect of cooling and velocity reduction
is to divide the particulate matter carried In the gas stream into two
fractions. One fraction, of lower alkali content, Is separated in a condi-
tioning tower and returned to the raw meal silos. The other fraction, of
higher alkali content, is collected in an electrostatic pr«cip!tator, pel-
let ized and discarded.
-------
I
c
1
I
T71
T
T70
T73
ALKALI BY-PASS
FIGURE 4
PREHEATER TOWER BY-PASS HOOD 7
QUENCH AIR 8
CONDITIONING TOWER 9
PREOPfTATOR C
EXHAUST FAN PT
WATER M
AIR
650°F CAT OFF CONTROL
AIR QUENCH NOZZLE CONTROL (4®>°F)
CONTROLLER
PRESSURE TRANSMITTER
MODULATOR
-------
10
3 CONSIDERATIONS AT THE PROGRAM PLANNING STAGE
The history of suspension preheater kilns In the United States
indicates a trend to this system [!*»]• Thirteen preheater kiins were
installed in the years 1953-1955- Because of inadequate knowledge about
the basic process and effects of raw materials on this system, numerous
operating difficulties were encountered. These difficulties, possibly
combined with emphasis on alkali reduction, caused the shut down of six of
these units. From 1956 to 1969, only two suspension preheater kilns
were installed, one of which has since been shut down. However, with
increasing fuel costs and emphasis on energy conservation, 22 suspension
preheater kilns have been commissioned since 1970. Another indication
of this trend is that five of the eleven new kilns Installed in
1975 were preheater-rotary kilns [15]. In addition, three existing
rotary kilns were converted to suspension preheater units. All five
kilns planned for completion in 1976 will be preheater kilns. While the
situation in Canada is similar, there are fewer plants and such trends
are difficult to follow.
Any study concerned with future use of the technology developed
must take into consideration the suspension preheater kiln. Flame
characteristics are identical In all cement kiln systems, hence,
demonstration that destruction of chlorinated hydrocarbons occurs in
one cement kiln implies destruction in all cement kilns. Reduction of
alkalies in a straight rotary kiln system by addition of chlorides
(calcium chloride and hydrochloric acid) has been extensively demonstrated.
It was decided to conduct the experiment in the suspension
preheater kiln despite knowledge that the preheater is prone to plugging
problems at high alkali chloride levels.
The primary purpose of the study was to determine whether
chlorinated hydrocarbons were destroyed in a cement kiln. It was
considered desirable to conduct the experiment in stages with materials
of increasing difficulty of destruction. After each stage, normal
operation was resumed, thus giving time to analyze samples in order to
determine that no waste chlorinated hydrocarbons were emitted. For
this study, it was agreed to use four different formulations of industrial
chlorinated wastes:
-------
1]
I) aliphatics;
2) aromatics;
3) aliphatics plus aromatics and alicyclics; and,
k) aliphatics, aromatics, alicyclics and large complex
molecules such as polychlorinated biphyenyls (PCB's).
Generally these are filtered, processed and blended to obtain
specific formulations for control of energy value, chlorine content and
viscosity.
-------
12
k TRIAL ON THE SUSPENSION PREHEATER KILN
The first phase of the trial was done in the suspension preheater
kiln to determine whether alkali reduction could be accomplished on
this unit by burning of chlorinated hydrocarbons. Reduction of alkalies
by addition of chloride is achieved by increasing the amount of alkali
volatilization. Alkali carbonates and chlorides are almost entirely
volatilized at the normal operating temperature of the kiln. Alkali
sulphates and alkalies which are complexed in the calcium silicate and
aluminate structures are preferentially retained in the clinker.
The addition of chloride ion causes formation of the easily
volatilized alkali chloride. Intimate mixing of hot kiln gases and
countercurrent cooler raw meal causes alkali chlorides to condense on
the raw meal thereby being returned to the kiln. Alkali chlorides are
trapped between the burning zone of the kiln (volatilization) and the
bottom stage of the preheater (condensation), causing increasing concentra-
tions in the gas stream. At high concentrations of alkali chlorides,
sufficient quantities condense in Stages III and IV of the preheater to
cause plugging of this system. To alleviate this situation, the bypass
is used to withdraw a fraction of the gases leaving the rotary kiln
and with these gases a fraction of the alkali chlorides. As greater
quantities of alkali chloride are volatilized in the kiln, high with-
drawal rates through the bypass are required.
Samples of the raw meal entering the kiln from the preheater
can be taken to monitor the alkal! chloride concentration. Prior to
the burn, it was estimated that equilibrium alkali chloride concentration
at this point would be reached in approximately two hours.
Gases exhausted via the bypass have a temperature of approx-
imately 1000°C (1830°F). To cool these gases, they are mixed initially
with ambient air, then passed through a conditioning tower in which water
spray is used for further cooling and conditioning of the gases.
In the conditioning tower, the coarser fraction of the dust, which has the
lower alkali concentration, settles to the bottom of the conditioning
tower and is returned to the process. The finer fraction, which has the
higher alkali concentration, is carried with the gas stream to the bypass
precipitator. Dust collected from this precipitator is discarded.
-------
13
Initially, attempts to operate the bypass at the level required
to compensate for addition of chlorides resulted in high emission rates of
participate matter. Weighings of materials showed that 50-60 tons per
day of dust were collected in the precipitator (normally 6-10 tons per
day), while k tons per day settled in the conditioning tower (normally
20 tons per day). The reversal of solids fraction removed in the two
units and the larger amount of material carried forward to the electrostatic
precipitator was attributed to the increased quantity of ambient air used
for cooling the bypass stream. This increased gas flow gave higher
velocities in the conditioning tower, lowered its settling efficiency
and resulted in particulate entrainment in the gas stream.
The burning of chlorinated aliphatic material was begun on
June 3, 1975 with a mixture having a specific gravity of approximately
1.2 and a chlorine content of about 55% w/w.
On June 5, the bypass duct between the kiln and the conditioning
tower was plugged. The trial was suspended while the duct was cleaned.
A second attempt was begun on June 10 using the same parameters used in
the June 3 trial. On June 13, similar buildup again caused the bypass
to fail. It was apparent that the addition of chlorides to the suspension
preheater kiln could not be done without major portions of the bypass system
being rebuiIt.
Due to high particulate emissions, plugging of the bypass
system and the cost of the equipment changes envisaged to alleviate the
short comings, it was decided to continue the trial in one of the two
existing wet kilns.
The trial on the dry process kiln did show, however, that
chlorinated hydrocarbons were destroyed in this kiln. Chlorinated
hydrocarbons used are volatile at temperatures found in Stage IV (980 C),
and could not have condensed upon the material. The expected increase in
chlorine concentration of Stage IV dusts was confirmed by analyses
(Table B.A). Equilibrium concentrations of chloride at Stage IV were
obtained in approximately four hours (Appendix B). Some slight alkali
reduction was apparent. During the five days on which chlorinated
hydrocarbons were burned, the raw meal feed K_0 on ignited basis averaged
, while the average clinker K-0 was 1.27%. With the same quantity
-------
of gases being removed by the bypass, but without addition of any chlorinated
material, the raw meal feed contained on average l.Mfc K_0 and the
clinker contained on average 1.291 K.O. While the difference is slight,
the duration of the experiment was short.
Complete details and analytical results from the experiment
on the suspension preheater kiln are given in Appendix B.
-------
5 TRIAL ON THE WET PROCESS KILN
5. 1 Discussion
The test program for this section of the study was designed to
determine if kiln emissions contained any chlorinated organic compounds
when chlorinated organic wastes were burned as supplemental fuel to reduce
the alkali content of the clinker.
From a knowledge of kiln zone temperatures and gaseous product
retention times in the kiln, It was anticipated that combustion of the
waste chloride fuel would produce predominantly C02> H20, HC1 (hydrochloric
acid), Cl" (chloride ion) and a negligible amount of free C12 (chlorine
gas).
H2° + C12 * 2HC1 + i02
Kp -
= 15 at 1000°C (1832°F)
• 68 at 1500°C (2732°F)
-150 at 1500°C (3*52°F) [28].
This shows that amount of free chlorine gas decreases with:
a) increase in temperature;
b) increase in water vapor content;
c) decrease in oxygen content; and,
d) decrease or removal of hydrochloric acid.
The hydrochloric acid will react and the chlorine will be
retained as alkali chlorides in the process solids. From a
knowledge of orthodox incineration systems, the combustion conditions
in a cement kiln were expected to provide a very favourable means for
the destruction of waste chlorinated hydrocarbons. In order to determine
the effect, if any, of burning waste chlorinated organic material on
eir quality, emissions from the kiln were monitored before, during and
after three separate periods of burning the supplemental waste fuels.
Prior to each sampling period, emissions were checked for the presence of
-------
16
HC1 and C12 with detector tubes at sensitivity limits of 2 ppm and 0.5
ppm, respectively.
A mass balance for chlorine was carried out by analyzing all
feed and process samples collected during each waste burn for this
element, to confirm the emission data.
Prior to the trial on the wet process kiln, light oil was put
into the chlorinated feed system for the purpose of calibrating and
testing the proportioning and metering devices. Some light oil remained
in the system when chlorinated aliphatic material was received, causing
the chloride content to increase as the test progressed. While emissions
were tested, chlorinated aliphatic material was burned at different
rates while attempting to compensate for the changing chloride content,
the rates being equivalent to:
October 28, 1975, 0.31* chlorine relative to clinker;
October 29, 1975, 0.32* chlorine relative to clinker; and
October 30, 1975, 0.63* chlorine relative to clinker.
Due to excessive quantities of chlorine input, a ring was
formed, which required a kiln shutdown for its removal. A ring Is
caused by the buildup of material on the Inside of the kiln to such
an extent that it restricts the flow of process materials.
y
Due to production scheduling, the delay caused by the shutdown
made it necessary to drop the burning of aromatics alone and immediately
progress to the mixture of aromatic and complex molecules. The conduc-
tivity of this latter material was too low (0.3 x 10 mhos) for the
magnetic flow meter to function properly.
Control of flow rates was therefore achieved by measuring the
depth of liquid in the tank. While accurate over a long period of time,
short term control was difficult because one Inch of liquid in the tank
is 196 imperial gallons. Variable flow rates were encountered during
emission testing, the quantity added each day baing:
December 10, 1975, 0.^5 to 0.71* chlorine relative to clinker;
December 11, 1975, 0.31 to 0.51* chlorine relative to clinker; and,
December 12, 1975, 0.79* chlorine relative to clinker.
-------
17
A ring also formed during this burn but broke away after the
sampling sequence had been completed, thus avoiding another kiln shut-
down. The presence of this ring restricted the burning of the residual
quantity of this material (aromatic plus complexes) to intermittent burns.
Due to the above mentioned upset and requirements of the
production schedule, it became necessary to add 13,000 gallons of the
polychlorinated biphenyl blend to the 12,000 gallons of material remaining
in the tank. The resulting mixture contained a significant quantity of
solids and also had very low conductivity, thus requiring tank measure-
ments as the method used for flow control.
The problem with solids separation and settling became more
severe as the test progressed. This caused burner nozzle restrictions
and pluggages which, in turn, caused interruptions and irregularities in
waste 1iquid flows.
For the last day of emission testing, the nozzle was removed
completely to permit uninterrupted flow rates of polychlorinated biphenyls.
Quantities added during emission tests were:
January 7, 1976, 0.06 to 0.14& chlorine relative to clinker;
January 8, 1976, 0.13 to 0.33% chlorine relative to clinker; and,
January 9, 1976, 0.61% chlorine relative to clinker.
5.2 Organic Chloride Waste Burned
The chemical compositions of the three waste materials burned,
the labelling used to identify each burn and the time periods during
which they were fed to the kiln are given below:
Chlorinated aliphatics WBA October 23 - November V75
WBA plus chlorinated
aromatics and alicyclics WBB December 5 - December 15/75
WBB plus poly-
chlorinated biphenyls
(PCB's) WBC January 3 - January 9/76
It was anticipated that complete combustion of these materials
would occur in the kiln. However, a knowledge of the major constituents
of each of the fuels was necessary in order to determine which trace
components might be detected in the sampled emissions should incomplete
combustion occur. Previous studies on the Incineration of waste
-------
18
chlorinated hydrocarbons [16] had shown that trace quantities of low
molecular weight compounds such as carbon tetrachloride (CO.), chloroform
(CHClJ, dichloromethane (CH.C1.) were detected in the combustion product
gases. It was decided, therefore, to analyze collected emission samples
for the presence of the major components of each composite waste fraction,
and for the low molecular weight organochlorldes referred to above.
Samples of the chlorinated wastes were taken from the line
between the tank and the kiln. Typical component analyses of the three
composite wastes burned during the program are recorded in Tables 1, 2
and 3-
The quantitative values given in the tables are based upon an
assumed equal response of all the components to a flame ion i ration
detector (FID). While this assumption is not very accurate, it was
considered sufficient to provide the needed compositional data. Identi-
fication of components at concentrations greater than one percent was
achieved by use of gas chromatography-mass spectrometry (GC-MS). Identi-
fication of components at concentrations less than one percent was not
routinely performed. Analyses of samples collected on different days during
a burn showed some variation In the concentration of individual components.
However, the overall chemical composition of the feeds remained the same.
Two of the waste burn sample feeds (WBB and WBC) were also
analyzed by the Ontario Ministry of the Environment (MOE), and the TRW
Systems Group. The results obtained by these agencies are reported in
detail in Appendices F and G, respectively. The TRW Systems Group analyzed
composite samples of the daily feeds for both waste burns by GC-MS. The
MOE group analysed individual daily samples for both waste burns using
GC-FID operating conditions. The chromatographic parameters used by
the MOE group to analyse for aromatic and PCB components precluded the
identification of aliphatic organochlorine compounds such as CHC1. and
Some variation in the percent composition of the waste burn
samples was apparent between the TRW and ORF results, especially with
respect to the amounts of aliphatic organochlorine compounds present
in waste burn B (WBB). The discrepancies observed are due to the fact
that TRW received a composite sample of all daily feeds for each of the
-------
19
TABLE 1. COMPOSITION OF ALIPHATICS (WBA)
Sample Feed - October 28, 1975
Peak # in GC Profile
(cf Figure A. 6)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2k
25
26
27
Approximate
Concentration
% Identification
17.4
!.4
0.8 |
1.5
3-2
Chloropropane & propene
ethyl chloride,
dichloromethane
Chlorobutane & butene
16.5 1» 2, Dichloroethane
6.6 Carbon tetrachloride
1.0
10.5 1,1, 2-Trichloroethane
0.7
0.3
1.9
1.4
1.4
Dichloropropanes
2.4 Tetrachloroethylene
7.3 Tetrachloroethane
15.6 Chlorobenzene
2.4 -v
0.2
3.3
0.3
1.1
0.6
0.5
1.6
0.6
mul ti chlorinated
butanes, butenes
hexanes , hexenes
0.7 '
Note: No identification of compounds at concentrations of \% or less was
attempted.
-------
20
TABLE 2. COMPOSITION OF AROMATICS PLUS COMPLEX (WBB)
Sample Feed - December 12, 1975
Approximate
Peak # In GC Profile Concentration
(cf Figure A.7) %
1 2.0
2 0.1
3 0.1
If 1.4
5 1.5
6 3.5
7 0.4
8 I.I
9 1-9
10 0.2
11 0.2
12 0.4
13 0.6
14 0.5
15 0.5
16 52.2
17 0.2
18 0.1
19 0.1
20 6.3
21 2.7
22 0.1
23 1.0
24 8.5
25 4.4
26 6.9
27 1-3
Identification
Chloropropane
-
-
Chloroform
1, 2-Dlchloroethane
Carbon tetrachloride
-
-
1, 1, 2-Trichloroethane
-
-
-
-
Chlorobenzene
-
Chloro toluene
-
-
-
Hexach 1 orocyc 1 open tad i ene
Heptachlorocyclopentene
•
Pentachlorobenzene
Oc tach 1 orocyc 1 opentene
Complex associated with
Hexach 1 orocyc 1 open t ad i ene
Note: No identification of compounds at concentrations of 1% or less was
attempted.
-------
21
TABLE 3. COMPOSITION OF AROMATICS PLUS PCB's (WBC)
Sample Feed - January 8, 1976
Peak # In GC Profile
(cf Figure A. 8)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2k
25
26
27
28
29
30
31
Approximate
Concentration
% Identification
3.4 Chloropropane
0.3
1.4 1, 1 ,-Dichloroethane
1.3 Carbon tetrachloride
0.5
0.3
1.9 1,1, 2-Trichloroethane
0.1
1.9
0.5
28.5 Chlorotoluene
< 0.1
< 0.1
1.8
0.9
2.8
0.9
2.2
0.3
2.1
0.9
6.7
3.*
5.9
1.6
12.1
5.6
4.0
5.4
2.1
1.6
Hexachlorocycl open tad iene
+ complex
Dichlorobiphenyl
Trichlorobiphenyl
Tetrachlorobipnenyl
Pentachlorobiphenyl
Note: No identification of compounds at concentrations of 1% or less was
attempted.
-------
22
waste burns, and different chromatographIc columns and a different detector
system were used by the two laboratories.
Since the study was to evaluate the burning of waste chlorinated
materials of variable composition, minor differences between the
laboratories' results do not affect the conclusions of the study.
The agreement among all groups was good with respect to the
identification of components present in the two waste fuels. The presence
of chlorinated aliphatics was observed In both waste fuels. VBB samples
were found to consist primarily of chlorinated aromatic compounds, in
particular o-chlorotoluene. WBC samples were found to consist of approx-
imately 50% polychlorinated biphenyls (Aroclor 12
-------
23
to collect higher molecular wefght compounds, Including any noneombusted
starting materials in the impingers.
Components adsorbed on the sorbent were removed by thermal
desorption prior to analysis. All other collected fractions, from both
trains, were extracted with organic solvents prior to analysis. These
included probe rinse water solubles and insolubles, filters and impinger
solutions.
In considering the results obtained in this study it is
important to keep in mind the following:
- The collection efficiency of Chromosorb 102 sorbent for low
molecular weight chlorinated hydrocarbons was determined to
be better than 90%. The collection efficiencies of filters
and ice-water in the impingers of a Joy train for organo-
chloride components was not determined.
- The methods of collection, followed by thermal desorption or
solvent extraction, were designed to concentrate any organic
compounds which may have been present. Such techniques allow
detection and identification of compounds at much lower levels
than would otherwise be possible. Concentration factors for
each type of sample have been calculated for each phase of the
burn and are given in Table k. For the desorbed gas samples,
organic compounds from several cubic feet of kiln emission
gases were collected on Chromosorb 102 and then desorbed into
500 millilitres. The concentration factors are the ratios
of these two volumes. In the case of the solvent extracted
samples, molecular weights enter the calculation in the conver-
sion cf volume to weight. A complete discussion of these
calculations is given in Section A.6.4.
5.3.2.1 Desorbed samples. The concentrations of volatile organic
chlorides calculated for the kiln emissions from gas chromatography -
electron capture (GC-EC) are recorded in Table 5.
The results reported were averaged for each test series.
Dichloromethane (DCM) was tentatively identified as the major component
of these desorbed gas samples. Others tentatively identified were CHCK
-------
TABLE 4. GAS SAMPLE VOLUMES AND SAMPLE CONCENTRATION FACTORS
(Gaseous Sampling Train)
Test #
1 BLA
2 BLA
3 BLA
1 WBA
2 WBA
3 WBA
1 WBB
2 WBB
3 WBB
1 WBC
2 WBC
3 WBC
1 BLB
2 BLB
3 BLB
Date
Oct.
Oct.
Oct.
Oct.
Oct.
Oct.
Dec.
Dec.
Dec.
Jan.
Jan.
Jan.
Jan.
Jan.
Jan.
20 (1975)
21
22
28
29
30
10
11
12
7 (1976)
8
9
19
20
21
Test
Duration
(mins)
317
281
320
362
286
285
260
317
250
268
237
228
170
247
240
Volume*
Sampled
scf
2.
1.
2.
3.
2.
2.
2.
2.
2.
2.
2.
2.
1.
2.
2.
80
86
83
22
53
53
30
80
21
37
09
02
50
18
12
Sample Concentration
Factors
Desorbed Solvent**
Gas Extracted
159
Hi
160
181
1*3
143
130
159
125 .
134
119
114
85
124
120
510
500-1100
600-2000
520-1500
380
-Volume sampled per adsorbent tube.
-'-These values are based on the lowest and highest molecular weights of
compounds found in the waste1 feeds. Figures given for both baseline
series are based upon the molecular weight of dichloromethane (DCM).
-------
25
TABLE 5. ESTIMATED KILN EMISSION CfMlCENTftATllikltt
FOR SPECIFIC VOLATILE ORGANOCHLORINE COMPOUNDS
Emission Concentrations
Test Series Dichloromethane
ppb yg/m*
BLA
WBA
WBB
WBC
BLB
4.1
5.4
18.0
7.7
29.0
14.5
27.3
19.1
102.7
63.7
Chloroform
ppb pg/m'
0.004
0.015
0.038
0.069
0.018
0.020
0.080
0.190
0.345
0.090
Carbon
Tetrachlor ide
ppb yg/m3
0.0004
0.0020
0.0020
0.0060
0.0004
0.0026
0.0128
0.0128
0.0385
0.0160
and CC1,. It was found that GC-EC profiles of blank determinations
obtained with unused conditioned Chromosorb 102 were variable, and
the peaks that were present in the profiles possessed similar retention
times to the compounds of interest. Therefore, reported results are
higher than actual concentrations in the emission gases. The differing
concentrations for DCM quoted in Table S, especially with respect to BLB
versus BLA, WBA or WBB, are probably a result of this background contamina-
tion rather than real differences. It was still noted, however, that
the maximum concentration calculated as DCM in the kiln emissions was
no greater than 30 ppb even with this positive bias from background
contamination.
The MOE results confirm the ORF findings with respect to the
detection of low molecular weight chlorinated organics in the desorbed
gas samples. TRW did not find any trace of these volatile chlorinated
organics in their desorbed gas samples. However, TRW did not routinely
perform specific analyses for compounds present at concentrations below
0.1 mg/m .
5.3.2.2 Organic solvent extracted samples. Due to the much higher
gas flow rate through the particulate sampling train to maintain an
isokinetic sampling rate, the concentration factors for solvent extracted
samples for this train were greater than those collected with the gaseous
train. Calculated average concentration factors for each test series are
-------
26
given below:
BLA 7,000
WBA 6,000 - 13,000
WBB 7,000 - 23,000
WBC 7,000 - 20,000
BLB 6,000
For the waste burns, the two values given represent the highest
and lowest molecular weight species present in the waste chlorinated hydro-
carbons. Calculations for the baselines are based on the molecular
weight of dichloromethane. Subtraction of background interference based
on control blanks (Appendix A.6.3) was performed on solvent extracted
samples. Comparison was then made with the appropriate waste feed sample
chroma tog rams. It was concluded that no uncombusted waste fuel components
were present in any of the organic solvent extract samples at the
detection limits of the analytical procedure used. Representative
chromatograms together with discussion are presented in Appendix A.6.5-
TRW performed analyses on portions of all the organic solvent
extracts obtained from BLA, WBB, WBC and BLB emission samples. The
samples were specifically searched for chlorinated compounds. Also,
the Chemistry Division, Air Pollution Control Directorate (APCD),
Department of the Environment, Ottawa, Canada carried out analyses
on portions of the solvent extracts of WBC emission samples to search
for high molecular weight chlorinated hydrocarbons. Detailed reports of
the work undertaken by these groups are presented in Appendices G and I,
respectively.
The noteworthy result with respect to both reports was that in
none of the samples analyzed by either group were any chlorinated organic
compounds detected. Their results thus confirm the ORF conclusion that no
chlorinated organic residues were detectable in any of the organic solvent
extracts of the emission samples collected during the various waste
chlorinated hydrocarbon burns.
5.3.2.3 Results obtained by the participating laboratories. In gas
chromatographic analyses, program parameters define the conditions used.
The choice of detector and column makes the analysis specific to certain
-------
27
groups of compounds. Four laboratories analyzed the emissions samples
taken during this study.
One of the laboratories, APCD, was requested to search for
high molecular weight chlorinated hydrocarbons. The TRW group were not
requested to identify low molecular weight chlorinated compounds present
at less than 0.1 mg/m in the stack gases. TRW did, however, search for
PCB's using techniques designed to detect small quantities of these
compounds. ORF and MOF were assigned the task of detecting and identify-
ing all possible chlorinated organic species. Analytical design, and
hence results, reflect differences in the analyses requested of the
different laboratories.
Both ORF and MOE found low molecular weight hydrocarbons,
such as dichloromethane, to be present at microgram per cubic meter
(yg/m ) levels in the emissions. That TRW detected no such compounds
is not contradictory. The levels in the emission gases were less than
their required detection limits. Similarly, it was not within the terms
of reference for the APCD work to determine these compounds.
Results of the four laboratories can be conveniently summarized.
While burning chlorinated hydrocarbons, low molecular weight chlorinated
7
compounds were emitted at levels of a few yg/m (Table 5). None of the
participating laboratories detected any high molecular weight chlorinated
hydrocarbons from either air sampling train. At detection limits of
3 ug/m3 in the stack gases, polychlorinated biphenyls-were not found.
5.3.2.4 Estimated minimum combustion efficiencies. The maximum value
for total chlorinated hydrocarbon content in the kiln emissions was
obtained in test 2 of the WBC series. Ignoring background subtraction
for the baseline samples and the Interferences from control blanks a
maximum value of kO ppb was determined. If a collection efficiency of
80% is assumed, a maximum level of 50 ppb in the kiln emissions is
obtained. Because of the high and uncertain background levels, the
estimate is higher than actual levels in the emission gases.
An average molecular weight for each of the three composite
feeds can be obtained from a knowledge of the composition of the feed.
Using this information, together with the average fuel feed rates to
the kiln and average gas volume flow rates in the duct, the maximum
-------
28
chlorinated hydrocarbon content can be determined. These values are
presented in Table 6. Minimum combustion efficiencies were then calculated
using the estimated maximum value of 50 ppb derived above. These figures
in Table 6 are considered very conservative estimates because of the
method used to calculate them.
TABLE 6. ESTIMATED MINIMUM COMBUSTION EFFICIENCIES
FOR EACH WASTE BURN
Waste
Compos \ te
WBA
WBB
WBC
Maximum
Content
Assuming no
Combustion.
ppm g/m
550 2.40
*»70 3-37
350 3.02
Estimated Organic
in Kiln Emissions
From Sample
Chroma tog rams
ppb yg/nr*
50 177-1
50 177.1
50 177.1
Minimum
Combustion
Efficiency %
v/v
99.990
99.989
99.986
*Based on the molecular weight of dichloromethane.
5.3.3 Particulate emissions
Summary data for the particulate tests made during each test
period are provided in Table 7. For each series of tests, the particulate
loadings and emission rates are quite consistent except for the third
test in the WBA series. The very high loading obtained probably resulted
from some temporary malfunction of the precipitator. It is noticeable
that, when the chloride wastes were burned in the kiln, the particulate
emission rate increased. For the WBA series, the emission rates were
about four times the rates obtained during the baseline tests and for
the WBB and WBC series the rates were twice those obtained during the
baseline tests.
Increased emission rates were not unexpected since combustion
of the chloride wastes produces HC1 and C1-, which react with the alkali
components in the raw feed to form volatile alkali chlorides. At ths
precip!tator the particulate loading is, therefore, increased and, since
the condensed alkali chlorides are very fine and have a different
resistivity, the amount of material passing through the collector
increases. Another factor influencing emissions of particulate matter
-------
TABLE 7. SUMMARY OF PARTICULATE TEST DATA
Test #
1 BLA
2 BLA
3 BLA
1 WBA
2 WBA
3 WBA
1 WBB
2 WBB
3 WBB
I WBC
2 WBC
3 WBC
i 8LB
2 BLB
3 BLB
Date
Oct.
Oct.
Oct.
Oct.
Oct.
Oct.
Dec.
Dec.
Dec.
Jan.
Jan.
Jan .
Jan .
Jan .
Jan.
20/75
21/75
22/75
28/75
29/75
30/75
10/75
11/75
12/75
7/76
8/76
9/76
15/76
20/76
21/76
Probe
Rinse
Gain
(mg)
57.6
58.8
35.8
99.0
94.8
237.5
.143.0
86.1
145*6
62,2
43 1
8?,8
65.2
5'. .4
11.9
Fi Iter
Gain
(rag)
285.7
366.0
291.8
1242.6
1261.7
2193.8
424.1
468.4
621.6
546,3
436.5
600,0
251.7
240.1
183.5
Total
Gain
(mg)
343.3
424.8
327-6
1341.6
1356.5
2431.3
567.1
554.5
767.2
608,5
479-6
687-8
316.9
291-5
195-4
% of
Total Gain
on Filter
83
86
89
93
93
90
75
84
81
90
9i
87
79
82
94
Vo 1 ume
Sampled
(std ft3)
141.04
135.86
132.83
142.81
137.34
109.88
106.64
117-06
116.24
119.62
113-52
118.99
126.88
116.60
116,45
Flow rate
ACFM
157,000
155,000
153,000
Average
157,000
153,000
165,000
Average
148,000
162,000
161,000
Average
172,000
1 60 , 000
167,000
Average
183,000
163,000
163,000
Average
Concentration
(grains/ft3)
0.0376
0.0483
0.0367
0.0409
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1458
1524
3415
2132
0821
0731
1019
0857
0785
0652
0892
0776
0385
0386
0259
0343
Emission
Rate
(Ib/hr)
20.8
25.4
19-6
21.9
83.9
84.6
200.0
122.8
40.3
40.5
55.0
45.3
45-3
54.5
52.2
44.0
23-9
21.0
14.6
19.8
-------
30
were the kiln rings formed while burning chlorinated hydrocarbons.
Kiln rings, by their effect on gas velocities through the kiln,
Increase the amount of particulate material carried to the precipitator.
Extensive ring formation was noted during test WBA. The average
emission rate for WBA was about 3 Ib/ton of clinker produced, compared to
the Canadian Federal Government objective of 0.9 Ib/ton. The average
value for the other phases of the test was about 1.1 Ib of particulate
emitted/ton of clinker produced and that of the baselines was 0.5. This
apparent increase in the particulate emissions while burning chlorinated
hydrocarbons is believed to be partially caused by the change in resis-
tivity of the dust entering the precipitator due to its increased
alkali content. A modification in the design of the precipitator could
compensate for this change.
The higher emission rates did not significantly add to the
suspended particulate in the ambient air in the vicinity of the plant.
Computed maximum ground level concentrations using standard dispersion
equations were less than 2 yg/m for the baseline emission rates and,
during the burning of waste chlorides, did not exceed 10 yg/m . The
current Ontario standard is 100 yg/m .
$.k Mass Balance on Wet Kiln
For the wet process kiln, the input streams are slurry feed, No.
6 fuel oil and, when burned, the waste chlorinated hydrocarbons. Clinker
(the product) and a portion of the precipitator dust form the output
streams. The major portion of the precipitator dust is immediately
returned to the process. This stream, though not required for the mass
balance, was monitored for information on kiln operation while burning
chlorinated hydrocarbons. Balances of chlorine and potassium were
calculated for both baseline periods and for each type of chlorinated
waste burned. Methods of sampling and quantifying material streams are
detailed in Appendix A. Details of analytical results, calculations and
tables of daily mass balances are given in Appendix C.
5-1*.! Significance of the mass balance
In common with other material balance experiments on large
scale production systems, the material accounting in this experiment showed
apparent losses and gains when the data were expressed in percent retention.
-------
31
These should not be regarded as real system losses or gains. There are
random and systematic errors in weighing and quantifying very large
quantities of materials and in analytical results. In the case of
alkali chlorides, a cycle is developed within the kiln whereby alkali
chlorides volatilize in the hotter section of the kiln and condense in
the cooler section. Should a kiln upset occur, the cycle may be broken
by alkali chloride being retained in the clinker. Such an occurence may
be of sufficiently short duration that clinker samples taken may not be
indicative. This effect is evident from the data of October 11-14, 1975-
A power failure on the eleventh caused a three hour kiln shutdown. The
indicated very low retention of both chlorine and potassium occurred
presumably while the cycle was being re-established. Since the equilibrium
changes when burning of chlorinated materials begins or ends, this feature
can also be seen with each phase of chlorinated waste burning. At the
start of any chlorinated burn, the mass balance for chlorine indicates
very low retention while the cycle is established. At the end of the
chlorinated burn, a gain is indicated as chlorine is retained in process
solids while the new equilibrium Is established. Formation of kiln rings
also caused an apparent loss of alkali chloride since a considerable quantity
was contained within the ring.
5.A.2 Chlorine and potassium retained
The cumulative percentage of chlorine retained (Table 8) was low
in all cases throughout the study, ranging from 50.7 percent for the
chlorinated aliphatic burn to 92.2 percent for the final baseline. K20
(Table 9) closely followed the pattern of chlorine retention, although
at a different retention level, with 83-5 percent the lowest value found
during the chlorinated aliphatic burn and 97.8 percent the highest value
found during the final baseline. The reason is clearly that potassium
chloride was being lost in ring formation and kiln cycle equilibrium
fluctuations.
The cumulative percentage K20 retained was less affected than
that of chlorine by such losses because the quantity of K20 was f, ?m 3 to
20 times greater than the quantity of chlorine. Random and systematic
errors could also be expected to play a greater role in the chlorine
balance for the same reason.
-------
TABLE 8. ACCUMULATED MASS BALANCE FOR CHLORINE
Accumulated Period
7/10/75-22/10/75
23/10/75- VI 1/75
2/12/75-1V12/75
3/ 1/76- 9/ 1/76
10/ 1/76-2 I/ 1/76
Accumulation Period
7/10/75-22/10/75
23/10/75- VI 1/75
2/12/75-1V12/75
3/ 1/76- 9/ 1/76
10/ 1/76-2 1/ 1/76
Chlorinated Accumulated % of T(
Hydrocarbon Input (1b) From Chlorinated Accumulated
Burned (T|) Hydrocarbon Retention (Ib) % Retention
28,249 0
Aliphatic 134,379 83.2
Aromatic + 135,705 85.0
Complex
PCS 58,590 82.4
18,351 0
TABLE 9. ACCUMULATED MASS BALANCE
Chlorinated
Hydrocarbon Accumulated
Burned Input (tons)
240.99
Aliphatic 199.67
Aromatic + Complex 194.59
PCS 103.25
280.40
20,768
68,088
112,640
41,704
16,927
FOR K20
Accumulated
Retention (tons)
229-37
166.72
186.68
92.88
274.36
73-5
50.7
83.0
71-2
92.2
% Retention
95.2
83.5
95.9
90.0
97.8
-------
33
While the chlorine balance does show discrepancies, in all cases
while burning chlorinated hydrocarbons more chlorine was retained in process
solids than the quantity input with materials other than these wastes.
A major portion of the chlorine from the chlorinated hydro-
carbon materials is thus accounted for. Considering the uncertainties
involved with ring formation and kiln upsets, the mass balance for
chlorine confirms the finding of the emission measurements that all
chlorinated hydrocarbons are destroyed in the cement kiln.
-------
6 CONSIDERATIONS ON BURNING CHLORINATED HYDROCARBON WASTES IN A
CEMENT KILN
6.1 Effect on Production
Kiln ring formation normally decreases kiln production, apart
from any loss of production caused by downtime. However, the kiln rings
formed during this study were attributed to poor control of feed rates
of the chlorinated waste. Ring formation through chlorine addition rates
higher than desirable were previously encountered at St. Lawrence Cement
while adding waste hydrochloric acid. Adequate control would certainly
eliminate this problem and, since this is a fairly simple proposition,
kiln ring formation should not be considered a deterrent to use of this
waste in the kiIn.
During the study, average production rates (Table 10) were
1038 tons per day while not burning chlorinated hydrocarbons and 1025 tons
per day while burning these materials.
There was also an increase of approximately 20 tons per day in
the quantity of dust discarded (Table 11) while burning chlorinated
materials.
While these values may be partly due to ring formation, it is
probable that dust generated would increase due to increased volatilization
of alkali chloride, and that this would have a corresponding effect on
clinker production. Thus, if another form of chloride were being used for
alkali reduction, it is unlikely that any change would be detected by use
of chlorinated hydrocarbon wastes.
6.2 Alkali Reduction While Burning Chlorinated Hydrocarbon Wastes
As indicated earlier in this report, Na_0 is not considered in
this study because it is low and practically constant in the St. Lawrence
Cement Co. raw materials and products.
To determine the efficiency of alkali reduction, the following
points require consideration. The percent K.O reported in the slurry
feed is on the natural or "as received" basis. To determine the quantity
which would be present if none were volatilized in the burning process,
the results must be calculated on the "ignited basis", that is, recalculated
for the CO. evolved from the raw materials in the burning process. The
-------
TABLE 10.
AVERAGE REDUCTION IN 1^0 CONTENT OF CLINKER
Accumulation Period
7/10/75-22/10/75
23/10/75- 4/11/75
2/12/75-14/12/75
3/ 1/76- 9/ 1/76
10/ 1/76-21/ 1/76
C linker
Prod.
(t/day)
1056
1050
1020
1006
1020
Cl Input With
Hydrocarbon
Chlorinated % Relative to
Hydrocarbon Clinker Prod.
-
Al iphatic
Aromatic +
Complex
PCB
-
C
0.41
0.44
-
0.34
0
Slurry
Na t u ra I
Basis
0.92
0.92
0.90
0.91
0.91
Kzp (?)
Igni ted
Bas i s
1.42
1.42
1.42
1.41
1.40
K20
Cl inker
(*)
1.21
0.87
0.74
0.87
1.25
K20
Reduction
(*)
0.21
0.55
0.68
0.54
0.15
Calculated*
K20
Reduction (*)
0
0.58
0.58
0.45
0
*Based on chlorine input.
-------
36
TABLE 11. AVERAGE DUST DISCARDED
Period
7/10/75-22/10/75
23/10/75- Vll/75
2/12/75-1V12/75
3/ 1/76- 9/ 1/76
10/ 1/76-21 / 1/76
Chlorinated
Hydrocarbon
—
Aliphatic
Aromatic +
Complex
PCB
—
Average Dust
Discarded
(Tons Per Day)
22.2
42.7
74.0
62.5
45.1
percent K.O In slurry feed on the ignited basis minus the percent 1^0 in
the clinker yields the reduction through volatilization in the kiln.
Since the K.O reduction is achieved by formation of potassium chloride
(KC1), the reduction expected while burning chlorinated hydrocarbons,
based on the assumption that all chlorine is bound Into potassium chloride,
can be readily calculated. Comparison of actual and calculated values
(Table 10) are excellent. If chlorine is added to reduce alkalies, the
reduction is stolchiometric. However, the total reduction is relative to
alkali levels in the slurry and not to the baseline content in the clinker.
This result was confirmed both by this study and by previous experience
using hydrochoric acid as the source of chlorine. This may be due to
formation of potassium chloride from volatile compounds such as
potassium carbonate. On average, without chloride addition, a reduction
of 0.18 percent K.O is apparent. While burning chlorinated hydrocarbons,
the average reduction is that calculated on the basis of potassium chloride
formation plus 0.05 percent.
The extremely good correlation between quantity of chlorine
input by burning of chlorinated hydrocarbons and alkali reduction in the
clinker gives better proof than the mass balance that chlorinated
hydrocarbons are destroyed in a cement kiln. Otherwise, alkali chloride
could not be formed, and volatilization to the extent noted could not
occur. The high degree of correlation indicates that all of the chlorinated
hydrocarbon was destroyed.
-------
37
6.3 Heat Recovery From Chlorinated Hydrocarbon Wastes
Calculation of heat recovery from these materials, normally
difficult due to the low flow rates involved, was made more difficult
by the flow control problems encountered. During the aromatic plus
complex chlorinated hydrocarbon burn, the oil meter did not function
properly and assessment of heat recovery was made only over the first
two days of burning. As near as can be established, approximately 65%
of the heat of the chlorinated hydrocarbon was recovered (Table 12).
This can be expected since volatilization of potassium chloride
required energy. While the energy is recovered upon condensation, this
occurs too far along the kiln to be completely recuperated. Weber [18]
states "volatilization of alkalies consumes high grade heat in the
sintering zone and calcining zone above a material temperature of 800 C
and this heat is subsequently released only at lower temperatures. Thus,
degradation of high-grade heat takes place".
When it is considered that potassium chloride forms a cycle
within the kiln and may be volatilized several times before escaping to
the precipitator, the heat consumed by this process becomes considerable.
In the suspension preheater kiln, an indication of the number
of cycles of volatilization of alkali chlorides is given by the relation-
ship between the Stage IV chloride level and quantity of chloride input.
Since in this study, the l<20 at Stage IV was a factor of ten greater than
the K 0 input, it follows that ten cycles of volatilization occur on average
in the suspension preheater kiln. The situation in the wet process kMn,
however, is dissimilar. There is no convenient method for determining
the number of alkali cycles involved. That there will not be as many
cycles as in the suspension preheater kiln is certain when consideration
is given to the differences in the two processes.
It is assumed, for purpose of illustration, that a chlorinated
hydrocarbon with 9300 Btu/lb and k2% chlorine is being burned with an
alkali cycle of three times. For each pound of chlorinated material,
there are O.J»2 pounds of chlorine which will produce 0.88 pounds of
potassium chloride. The heat of vaporization of potassium chloride is
38,840 cal/mol [19] or 938.6 Btu/lb. The 0.88 pounds of potassi
ium
-------
38
TABLE 12. RECOVERY OF Btu FROM CHLORINATED HYDROCARBONS
Material
Aliphatic
Aromatic +
Comp 1 ex
PCB
Average
Reduction
Input With
(10 x Btu/ton
Clinker)
0.125
0.128
0.157
Average
Input With
Chlorinated
Hydrocarbon
(10^ x Btu/ton
Clinker)
0.205
0.217
0.228
% Useful
Heat From
Chlorinated
Hydrocarbon
61
59
69
chloride would require 2^78 Btu for three cycles. On this basis we
might expect to recover:
93°° 7
9300 X 10°* " 73* of the heat content of the chlorinated
hydrocarbon.
While the above is only an approximation and ignores concurrent
heat exchange processes, it indicates that the heat recovery of about
65% obtained in the present study is reasonable.
It should be emphasized that the addition of calcium chloride
would require a similar quantity of heat to volatilize alkalies. In the
case of calcium chloride, an increase in fuel quantity would be required.
6.4 Cement Quality
While the literature available [2, 3, 10] indicates that alkali
reduction is beneficial with regard to cement quality, the possibility
of unburned organic material being retained in the cement and having
a deleterious effect was considered. Cements were ground in the laboratory
with clinkers produced while burning only No. 6 fuel oil as well as
clinkers produced during the aromatic plus complex and PCB burns.
Detailed results are given in Appendix D. The results obtained indicated
the only effects to be those due to alkali reduction.
6.5 Extrapolation to Other Kiln Types
Since all cement kilns have in common the requirement that
uniformly high temperatures be maintained, the authors believe that
-------
39
chlorinated hydrocarbons will be destroyed in all types of cement kilns
regardless of fuel used. The only qualification to this statement is
that the wastes must be injected into the burning zone. While it is
advisable to atomize the chlorinated waste, there appeared to be no
difference during the PCB burn with or without a nozzle for atomization.
For fuel economy, cement kilns are operated at low excess oxygen.
This, combined with the high temperatures and hydrogen from the fuel,
will in all cases ensure that hydrogen chloride (HC1) is preferentially
formed rather than free chlorine. In all cement kilns, the lime will
readily react with the hydrogen chloride. Based upon results of the
present study, no detectable quantities of either compound are expected
to be emitted from any type of cement kiln.
Since alkali chloride is volatilized and condensed in the gas
stream as extremely fine particles, it would be expected that emission
of particulate matter would increase by an amount depending on the
efficiency of the precipitator. This would apply to any method of
chloride addition. To overcome the plugging problem caused by condensation
of these salts in suspension preheaters, design modifications to these
units need to be installed and demonstrated.
6.6 Comparison of Cement Kiln Burning with Other Uses and Disposal
Methods for Waste Chlorinated Hydrocarbons
Information on disposal methods used and actual quantities of
chlorinated hydrocarbons requiring disposal is not read!ly available in
Canada. While a conservative estimate of 25-30 million pounds of
chlorinated hydrocarbon wastes was obtained, it seems likely that this
quantity represents only wastes from plants manufacturing chlorinated
hydrocarbon products. The magnitude of the problem of disposal of these
wastes in North America can be inferred by some of the methods used.
One method makes use of ships designed for burning of these wastes at
sea, this method being described in studies monitored by the United States
Environmental Protection Agency [20, 21]. Such a method of disposal
is expensive, requires constant monitoring of temperatures within the
furnace, uses additional fuel for combustion and emits hydrogen chloride
(HC1) which is dissolved in the ocean.
-------
Another method of disposal in North America is Incineration
with production of hydrochloric acid [22],
One report from Europe [23] indicates that methods of disposal
are:
1) illegal dumping of small quantities in barrels or other
containers on uncontrolled refuse dumps;
2) deposition of larger quantities in barrels on refuse dumps
which are supposedly sanctioned for this purpose;
3) combustion in simple facilities without hydrogen chloride
scrubbing;
k) combustion of piles of barrels on remote beaches with an
offshore wind;
5) dumping barrels on the open sea;
6) dumping liquids into the sea from moving vessels;
7) separation of waste materials and recovery of useful
components;
8) combustion with recovery of hydrochloric acid; and,
9) combustion on the open sea at temperatures guaranteeing
almost complete pyrolysis.
The same report states: 'Only the last three procedures can
be considered to be not harmful for the environment. But procedures 7)
and 8) are possibly very expensive and in special cases unsustainable for
the producer". Also from the same report, it is pointed out that combustion
at sea requires extensive observation of a variety of safety procedures.
A study from France [2k] indicates that a variety of legal and
illegal means are used to dispose of chlorinated hydrocarbon wastes.
Of illegal means: "the discharge of what in general are insignificant
quantities is disposed of in drums, or by tankers, into waterways, former
quarries now used for other purposes, or discharged with unsupervised
wastes that reason suggests should be retained".
In comparing combustion in a cement kiln, with other methods
of disposal which are considered not harmful to the environment, the
following points become apparent:
Incineration of these wastes is normally done at a flame
-------
temperature of 1200°C to 1560°C [25] while cement kiln
flame temperatures are 2100°C [18] or higher.
The retention time in a cement kiln flame envelope is consider-
ably longer than the O.I seconds normally found in an
i nci nerator.
For the production of cement clinker, the temperatures cited
are necessary [26], thus removing the necessity for constant
monitoring of temperatures as required when burning in an
i ncinerator.
To prevent operating difficulties, such as kiln rings, in a
cement kiln, the amount of chlorine is restricted to
approximately O.k percent relative to clinker. Besides
maintaining constant temperature, this requirement ensures
that sufficient hydrogen is available to form hydrogen
chloride which is readily absorbed by lime.
There is always a high quantity of lime in the cement kiln
to react with hydrogen chloride and thus prevent emission
of this compound to the atmosphere.
Burning of these wastes in a cement kiln saves fossil fuels,
as opposed to the necessity of using fossil fuels to
ensure combustion of these wastes in an incinerator.
Beneficial use is obtained in a cement kiln of persistent
and toxic waste materials which normally require disposal.
-------
7 CONCLUSIONS
The concept of using chlorinated hydrocarbon wastes in
cement processing derives from knowledge of kiln operating temperatures
and residence times in comparison with incinerators capable of destroying
these compounds. The action of cement kilns as "dry-lime scrubbers" in
relation to halogens in the kiln gases was already known.
In all cases, in the present study, analyses of kiln emissions
indicated the efficiency of combustion of chlorinated hydrocarbon wastes
to have exceeded 99-98 percent. Traces of volatile low molecular weight
chlorinated hydrocarbons were found to be present at a maximum emission
concentration of 50 ppb above baseline. All other starting materials,
including polychlorinated biphenyls, were completely destroyed. There
were no detectable emissions of hydrogen chloride, free chlorine or
high molecular weight chlorinated hydrocarbons.
An increase in total particulate emissions was observed while
burning chlorinated hydrocarbon wastes. This is apparently related to
an increase in dust loading to the precipitator as indicated by the
necessity of discarding more dust while burning chlorinated hydrocarbons.
Combustion in incinerators designed for destruction of chlorinated
hydrocarbon wastes caused the emission of most of the hydrogen chloride
except for those installations using expensive hydrochloric acid recovery
or scrubbing systems. These incinerators require continous monitoring
of temperature profiles and use fossil fuels to initiate or maintain
combustion.
The present study indicates that useful recovery of about 65
percent of the heat value and approximately 100 percent of the chlorine
is attained by the burning of these wastes in a cement kiln.
A mass balance on chlorine and the effective alkali reduction
derived from the chlorine contained in the chlorinated hydrocarbon
wastes confirms the air emissions data.
Consideration of the data from this study and examination of
the general literature on cement manufacturing has led the authors to
conclude that all chlorinated hydrocarbon wastes may be used in cement
kilns without adverse effect on air pollution levels.
-------
REFERENCES
1. Personal communication, January 15, 1975-
2. Niel, E.M.M.G., "The Influence of Alkali-Carbonate on the Hydration
of Cement", Proceedings of the Fifth International Symposium on the
Chemistry of Cements, Tokyo, 1968, Published 1969.
3- McCoy, W.J. and O.L. Eshenour, "Significance of Total and Water
Soluble Alkali Contents of Cement", Proceedings of the Fifth
International Symposium on the Chemistry of Cements, Tokyo, 1968,
Published 1969.
A. Steinour, H.H., "The Setting of Portland Cement, A Review of Theory,
Performance and Control". Portland Cement Association, Research
Department Bulletin 98, 1958.
5- Lerch, W. "Studies of Some Methods of Avoiding the Expansion and
Pattern Cracking Associated With the Alkali Aggregate Reaction".
Portland Cement Association, Research Department Bulletin 3', 1950.
6. Powers, T.C. and H.H. Steinour, "An Interpretation of Published
Researches on the Alkali-Aggregate Reaction: Part 1 - The Chemical
Reactions and Mechanism of Expansion; Part 2 - A Hypothesis Concerning
Safe and Unsafe Reactions with Reactive Silica in Concrete". Portland
Cement Association, Research Department Bulletin 55, 1955-
7. Annual Book of ASTM Standards, Part 13 (1975)- American Society for
Testing and Materials.
8. Lerch, W. "The Influence of Gypsum on Hydration and Properties of
Portland Cement Pastes", Portland Cement Association, Research
Department Bulletin 12, 1946.
9. Woods, H. "Reduction of Alkalies in Cement Manufacture", Mill Session
Paper M-1^9, Portland Cement Association, Manufacturing Process
Department, 1956.
10. Woods, H., J.L. Gilliland, Jr., J.F. Weigel, B.E. Kester, and
H.A. Stevens, "Symposium on Alkali Removal and Problems". Regional
Fall Meeting of General Technical Committee, PCA, Milwaukee, Wisconsin,
-------
44
Sept. 21-24 1959. Mill Session Paper M-158, Portland Cement Associa-
tion, Manufacturing Process Department, I960.
11. Coles, C.W. and D.G. Dainton, "St. Lawrence Cement Co. Clarkson
Plant", Cement Technology 1 (2), 43, 1970.
12. Herod, B.C. "St. Lawrence Cement Expands Clarkson Operation",
Pit and Quarry. July 1968.
13- Schroth, G.A., "Suspension Preheater System Consumes Less Fuel",
Rock Products 75 (5). 70, 1972.
14. Garrett, H.M., "The Potential Promise - Prospects and Pitfalls
In Energy Conservation by the U.S. Cement Industry", Paper
presented at the Cement Chemists Seminar, Portland Cement Associa-
tion, Lincolnwood, 111., February, 1976.
15. Trauffer, W.E. "Portland Cement Outlook and Review". Pit and Quarry,
January 1976.
16. Campaan, H., Central Laboratory TNO Report No. CL 74/93. "On the
Occurrence of Organic Chlorides in The Combustion Products of an
EOC Tar Burnt by the Incinerator Ship 'Vulcanus1; A Preliminary
Investigation". Oct. 1974.
17. Gagan, E.W. "Air Pollution Emission and Control Technology - Cement
Industry", Environment Canada, Environmental Protection Service,
Economic and Technical Review Report EPS 3-AP-74-3, 1974.
18. Weber, P. (Translation), "Alkali Problems and Alkali Elimination in
Heat-Economising Dry-Process Rotary Kilns", Zement Kalk Gips, (8),
1964.
19- Perry, J.H., C.H. Chilton, and S.D. Kirkpatrick, Chemical Engineers
Handbook, 4th Edition, McGraw-Hill.
20. Marine Environmental Monitoring of "Vulcanus11, Research Burn II,
December 2-10, 1974, Preliminary Report, U.S. Environmental
Protection Agency, December 10, 1974.
21. Badley, J.H., A. Telfer, and E.M. Fredericks, "At-Sea Incineration
of Shell Chemical Organic Chloride Waste". Technical Progress
Report BRC - Corp. 13-75-F, Shell Development Company, April 1975-
-------
45
22. American Society of Mechanical Engineers, Research Committee on
Industrial Wastes Report. Incineration of Chlorinated Hydrocarbons
with Recovery of HC1 at E.I. du Pont de Nemours & Co., (Inc.)
Louisville, Ky. ASME Industry Survey, Present State of the Art,
Disposal of Industrial Wastes by Combustion, January 1971.
23- Grasshoff, K. (Translation) "Expertise Regarding the Effects of the
High Temperature Combustion of Chlorinated Hydrocarbon on Special
Vessels at High Seas", 1973-
2k. "Incineration of Industrial Chlorine Wastes on the High Seas".,
Report from the Environmental Agency (Ministere chargfe de
1'Environnement) of the Pollution and Nuisance Prevention
Administration (Direction de la Prevention des Pollutions et
Nuisances), France. 1974.
25. United States Environmental Protection Agency, Permit No. 730 DOOSC
(3) to Shell Chemical Company, Inc. and Ocean Combustion Services,
B.V. December 12, 1974.
26. Peray, K.E., and J.J. Waddel1, The Rotary Cement Kiln, Chemical
Publishing Co. Inc. New York, 1972.
27. Reynolds, L.M., "Pesticide residue analysis in the presence of
Polychlorobiphenyls (PCB's) Residue Reviews, 34, 27, 1971-
28. Santoleri, J.J. "Chlorinated Hydrocarbon Waste Recovery and Pollution
Abatement" Chem. Eng. Prog. 69 (0 68, 1973-
-------
ACKNOWLEDGEMENTS
The authors wish to thank and to acknowledge the participation
of the following people and organizations in this work.
Environment Canada
Mr. W.A. Neff (EPS/WPCD) for program coordination and invaluable
assistance in establishing details of the program management.
Mr. J. Robert (EPS/APCD) for helpful advice in sampling proce-
dures, analyses performed and review of results.
Ontario Ministry of the Environment
Dr. F. Frantisak and his staff for aid in planning the program,
analytical assistance and review of results.
Energy Mines and Resources Canada
Dr. F.D. Friedrlck for assistance at program planning stage.
St. Lawrence Cement Co.
Mr. L. Kraszewski for program coordination.
U.S. Environmental Protection Agency
A.W. Lindsey and J. Schaum (Hazardous Waste Management Division)
for program planning coordination and review of results. TRW
Systems Group and Control Pollution Services Inc., who were
contracted by EPA.
-------
APPENDIX A
QUANTIFYING, SAMPLING AND ANALYSIS OF PROCESS MATERIALS
-------
APPENDIX A
QUANTIFYING, SAMPLING AND ANALYSIS OF PROCESS MATERIALS
While analytical procedures used are identical for both wet
and dry process kilns, only the wet process kiln material streams are
considered in this Appendix. The dry process suspension preheater
kiln system is detailed in Appendix B.
The material flow through the wet kiln production system
is illustrated in Figure A.I. To obtain the material balance, six
materials were quantified, sampled and analyzed. A seventh material,
precipitator return dust, was monitored to obtain any additional operating
data this stream might yield. Table A.I lists the materials examined
with approximate relative quantities under normal production conditions.
TABLE A.I. PROCESS MATERIALS STUDIED AND APPROXIMATE NORMAL PRODUCTION
QUANTITIES
Material Approximate Quantity
Slurry Feed 15^0 dry tons/day
Clinker 1000 tons/day
Precipitator Discard Dust 0 - 140 tons/day
Precipitator Return Dust 350 tons/day
No. 6 Fuel Oil 20 gal/min
Chlorinated Hydrocarbons 1 - 2 gal/min
Kiln Exhaust Gases 160,000 ACFM
A.1 Quantifying and Sampling Slurry Feed and Clinker
There are four slurry basins, each of 6000 ton capacity,
including slurry water. The product from the slurry grinding mills is
pumped into these basins. The basins are supplied with air agitators
for the purpose of blending the slurry and maintaining a uniform suspension
of solids in liquid. Slurry is pumped from one basin at a time to the
constant level box of the slurry feed system. The material which over-
-------
1 SLURRY FEED
2 AIR
3 FUEL (OIL,CHLORINATED MATERIALS)
4 DUST RETURN
5 CLINKER
6 PRECIPITATOR DUST WASTED
7 STACK GAS
vn
o
A KILN
B PRECIPITATOR
SCHEMATIC OF THE MATERIAL BALANCE
FIGURE A. 1
-------
51
flows from the constant level box is returned to the slurry basins.
Slurry feed for the kiln is taken from the constant level box by an
Allis Chalmers "Ferris Wheel" bucket wheel feeder with variable speed
drive. From the bucket wheel feeder, the slurry flows into a small
basin with an outlet at the bottom from which the slurry goes directly
into the kiln.
The slurry feed samples were taken every two hours from the
overflow of the constant level box, and blended into a daily composite.
A portion of the daily composite then was dried at 120°C for analysis.
Slurry feed quantities were obtained by multiplying the total
number of revolutions of the "Ferris Wheel" per day by a factor. Measure-
ment of the slurry basin levels while feeding the kiln but not allowing
slurry to be fed into the basin gives an accurate quantity of slurry
fed to the kiln. A second method, performed every two hours, is to
close a valve at the outlet to the kiln of the small basin fed by the
Ferris Wheel. By determining the time to fill this basin, the quantity
fed to the kiln is known.
The clinker is fed to a bucket elevator by gravity through a
chute from the cooler. Clinker samples are taken from this chute at
intervals of two hours to form a 2A-hour composite. The clinker composite
sample is mixed, reduced by "cone and quartering" and a portion ground
for analysis.
In common with general practice in the industry, there is no
provision in the plant for continuous weighing of clinker. The water
content of the slurry feed is determined thrice daily, from which the
following calculation is made:
100 - HO
SFD - SF '
100
where SFD « quantity of slurry feed on a dry basis
SF » quantity of slurry feed including water
HO = % water in the slurry feed.
To obtain the quantity of clinker produced, the C02 which is
lost in the burning process (determined by loss on ignition of the dry
slurry feed) is deducted and a further correction is made for the quantity
-------
52
of dust discarded from the system. A total material balance may then
be written as follows:
/100 - L01 \ /100- LOI
"inker - SFD ^ ]QQ SFDj - DD ^ )00
where SFD - quantity of slurry feed on dry basis
LOI __ = % loss on ignition on slurry feed on dry basis
DD * quantity of dust discarded
LOI _ * % loss on ignition on dust discarded.
A.2 Weighing and Sampling Precipitator Dust
Dust from the precipitator is separated into two portions by a
system of screw conveyers under the precipitator. The major portion
is returned via a conveyor belt into a small holding silo from which the
dust is returned into the burning zone of the kiln (return dust). Samples
of the dust are taken at two-hour intervals from the conveyor belt to
form a 2^-hour composite. The composite is blended and a portion taken
for analysis. A minor portion of the dust (discard dust) is fed into a
second holding silo from which trucks are loaded directly. A sample
of this dust is taken from each truck load, and all samples for each day
are blended into a 24-hour composite. Each load of dust is weighed on
the truck weigh scales before being disposed of.
A.3 Measurement and Sampling No. 6 Fuel Oil
While it was considered unlikely that No. 6 fuel used in the
plant would contribute a substantial quantity of chlorine, samples of
oil were taken daily. Fuel quantities are continuously monitored and
recorded in the production data but are not given in this report.
Chlorine content of the oil was determined, and the contribution to the
chlorine mass balance due to oil was included.
A.k Measurement and Sampling Waste Chlorinated Hydrocarbons
Sampling of chlorinated hydrocarbons was carried out by withdrawing
material from the feed system twice daily and blending by vigorous mixing.
Samples were split into equal portions after blending.
-------
53
Initially, it was intended to meter chlorinated hydrocarbon
flow continuously. For this purpose, a magnetic flow meter was installed
in the line to monitor quantities of these materials. For this type of
meter to function properly, it is required that the conductivity of the
material be greater than 2 x 10 mhos. While this was the case for the
chlorinated aliphatic wastes, the conductivities of the other materials
were less than 0.3 x 10 mhos. As a result, chlorinated waste material
quantities were determined by taking measurements of the volume of
material in the storage tank. Tables A.2 through A.4 give detailed
information on calculated chlorinated waste material quantities.
A.5 Emission Sampli ng
Separate sampling trains were used to collect representative
samples of kiln emissions for analysis to determine the emission rates
of particulate material and to determine if organic chloride compounds
were present in the gas stream passing to the stack. Particulate and
gaseous samples were collected from a duct downstream of the precipita-
tor controlling dust emissions from kiln #1. The location is described
below, followed by descriptions of sampling equipment and methodology.
A.5.1 Sampling location
Emissions from both #1 and #2 wet process kilns pass through
electrostatic precipitators (Joy Manufacturing Co.) and then through
sections of rectangular breeching before mixing in a common section of
duct entering the stack. The stack, which is of height 55^-0 feet and
has an exit diameter of 13.0 feet, vents the exhaust gases to the
atmosphere at a temperature of about 400 F and a velocity of almost
1»0 feet per second.
The rectangular breeching from the precipitator makes a 90
bend and ti.en angles at about 30 upwards from the horizontal for a
distance of approximately 45 feet to the common header which leads into
the stack. This k$ foot section of dgct was considered the most suitable
for installation of sampling ports at a location which would meet both
Federal and Provincial source testing codes. It was subsequently
decided, therefore, to install five 4" diameter ports in the vertical
side of the duct at a position 12 feet upstream of the bend into the
-------
54
TABLE A.2. QUANTITIES OF ALIPHATIC MIXTURE BURNED DAILY
Date
23/10/75
24/10/75
25/10/75
26/JO/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/H/75
V 11/75
Time
From To
09:00
12:00
06:00
06:00
06:00
06:00
06:00
06:00
06:00
09:30
06:00
06:00
06:00
06:00
06:00
10:00
OFF at
12:00
06:00
06:00
06:00
06:00
06:00
06:00
06:00
09:30
06:00
06:00
06:00
06:00
06:00
10:00
17:30
17:30
Minutes
180
1080
1440
1440
1440
1440
1440
1440
210
1230
1440
1440
1440
1440
240
450
Igpm
0.5
1
1
1
1
1
1
1
1
2
2
2
2
2
2
1
Gallons
90
1080
1440
1440
1440
1440
1440
1440
210
2460
2880
2880
2880
2880
480
450
Gal Ions/Day
1170
1440
1440
1440
1440
1440
1440
2670
2880
2880
2880
2880
930
-------
55
TABLE A. 3. QUANTITIES OF AROMATIC PLUS COMPLEX MIXTURE BURNED DAILY
Date
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
6
7
8
9
10
11
12
12
13
14
15
Time
15:00
2 loads
12:15
15:00
15:30
12:00
15:00
09:00
15:30
13:30
13:20
1*4:00
Tank Difference
Measurement gallons minutes
12'
in 60930
14'
IV
13'
12'
ir
11 '
ir
10'
9'
8'
0"
Ib + 40690
Ib @ 1.
6" (7823-5886
1937
0"
6"
6"
8"
4"
0"
3"
6"
8"
1177
1177
2354
1962
785
785
1373
1766
1962
299 SG - 7823
= 1937 gal)
1275
1605
1470
1230
1620
1080
390
1320
1430
1480
Ga 1 /m i n
gal
1.519
0.733
0.801
1.914
1.211
0.727
2.013
1.040
1.235
1.326
Gal Ions per Day
Date
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
Dec.
2
5
6
7
8
9
10
From
06:00
10:00
14:00
06:00
06:00
12:15
06:00
15:00
06:00
15:30
06:00
12:00
To
13:10
12:00
06:00
06:00
12:15
06:00
15:00
06:00
15:30
06:00
12:00
06:00
Minutes
430
120
960
1440
375
1065
$1*0
900
570
870
360
1080
Igpm
(1)
1.519
1.519
1-519
1.519
0.733
0.733
0.801
0.801
1.914
1.9H
1.211
Gal Ions
430
182
1458
2187
570
781
396
721
456
1665
689
1308
Total Gallons
430
1640
2187
1351
1117
2121
1997
-------
56
TABLE A.3. (Cont'd)
Date
Dec.
Dec.
Dec.
Dec.
Dec.
11
12
13
14
15
From
06:
15:
06:
09:
15:
06:
13:
06:
13:
06:
00
00
00
00
30
00
30
00
20
00
To
15:
06:
09:
15:
06:
13:
06:
13:
06:
06:
Minutes
00
00
00
30
00
30
00
20
00
00
540
900
180
390
370
450
990
440
1000
1440
Igpm Gallons Total Gallons
1.
0.
0.
2.
1.
1.
1.
1.
1.
1.
211
727
727
013
040
040
235
235
326
326
654
654
131
785
905
468
1223
543
1326
1909
1308
1821
1691
1869
1909
-------
57
TABLE A.4. QUANTITIES OF PCB MIXTURE BURNED DAILY
Date
3/1/76
4/1/76
5/1/76
5/1/76
6/1/76
7/1/76
7/1/76
8/1/76
8/1/76
9/1/76
9/1/76
Date
3/1/76
VI /76
5/1/76
6/1/76
7/1/76
8/1/76
9/1/76
Time
14:30
12:00
09:15
14:30
09: 10
09:06
16:00
08:45
15:45
08:45
05:45
From
07:45
14:30
06:00
12:00
06:00
09:15
06:00
09: 10
06:00
09:06
16:00
06:00
08:45
15:45
06:00
08:45
Tank
Measurement
To
14:
06:
12:
06:
09 =
14:
09:
06:
09:
16:
06:
08:
15:
06:
08:
18:
30
00
00
00
15
30
10
00
06
00
00
45
45
00
:45
:15
11 ' 4"
12' 6"
9' 8"
9' 6"
8' 7"
7'H"
7'10"
7' 9"
7' 8"
7' 2"
6' 8"
Gal Ions
Minutes
405
430
360
1080
195
315
190
1250
186
414
840
165
420
855
165
570
Di f ference
gal Ions minutes
1962
1962
392
1766
1962
196
196
196
1177
1177
per Day
Igpm
1 .521
1.521
1 .521
1.539
1.539
1.244
1.577
1.366
1.366
0.473
0.195
0.195
0.467
1.154
1.154
2.180
1290
1275
315
1120
1436
414
1003
420
1020
540
Gal Ions
616
1414
548
1662
300
392
300
1708
254
196
164
32
196
987
190
1243
Gal An in
1.521
1-539
1.244
1-577
1.366
0.473
0.195
0.467
1.154
2.180
Total Gallons
2030
2210
2159
2008
614
1215
1433
-------
58
header, and positioned on a line perpendicular to the gas flow. The
port locations are shown schematically in Figure A.2.
As agreed upon by the Ontario Ministry of Environment and
Environment Canada, sampling was carried out at eight points on each
of the five duct traverses for a minimum of five minutes per point. The
sampling points were located at the centres of equal rectangular areas
in the duct as shown in Figure A.3- The numbers given in each rectangle
are typical gas velocities measured during the study and indicate the
distribution of gas flow at that location.
A.5.2 Sampling equipment
Prior to setting up equipment for particulate and gaseous
sampling, grab samples of the kiln emissions were taken using the
equipment depicted in Figure A.**. A gas sample was pulled through an
in-stack kl mm glass fibre filter in a stainless steel holder to
remove particulate material and then through a heated teflon line into
a Tedlar bag. For each bag sample a clean bag was placed in the lung
and the lid tightened to effect an air-tight seal. Valves A and C were
closed and B opened so that, with the pump on, the sampling line was
purged. Valve B was then closed and valve C opened to evacuate the
sampling lung, making sure that the vacuum did not exceed 5" Hg. At a
vacuum of V Hg, valve A was opened and the sample bag filled with
emissions from the duct. Valve A was then closed and valve B opened.
The pump was shut off and the line to the pump disconnected. Valve C
was opened to bring the lung to atmospheric pressure and the bag was
removed from the drum and quickly capped. Each bag sample was transported
to the laboratory at ORF for analysis on the same day that the sample was
collected.
A.5.2.1 Particulate train. Samples of particulate emissions were
collected using a Joy Manufacturing Company EPA (Model CU-2) train,
shown schematically in Figure A. 5. This equipment conforms to that
recommended by both the Ontario Source Testing Code [A.l] and the
Environment Canada Code [A.2]. Prior to each test a velocity traverse
was conducted across the duct through each part to determine an average
gas velocity. Gas temperatures were also recorded at each sample point.
-------
59
DUCT DIMENSIONS
(AS MEASURED) >
56-5" WIDE x 78"HIGH
SAMPLING PLATFORM.
FIGURE A.2. SCHEMATIC OF PORT LOCATIONS
-------
60
[
c
c
c
108.9
X
92.5
X
78.2
X
72.5
X
42.8
X
110.6
X
107.8
X
89.2
X
74.2
X
51.9
X
118.6
X
112.0
X
92.6
X
78.2
X
66.4
X
118.6
X
113.3
X
91-9
X
78.2
X
73.4
X
117.6
X
110.6
X
94.5
X
82.0
X
78.2
X
114.7
X
110.6
X
100.2
91.2
X
80.5
X
107.8
X
108.9
X
102.0
X
97.7
X
74.8
X
103.2
X
107.8
X
104.9
X
97.7
X
80.5
X
Velocity in feet per second
FIGURE A.3. GAS FLOW DISTRIBUTION AT SAMPLING POINTS
An Orsat analysis was made to find the concentrations of CO, C0_, 0_ and
N- in the gas stream and a gas moisture determination was carried out
using an established procedure [A.)]. With the preliminary data obtained,
the isokinetic sampling rate was calculated using known standard
equations [A.I].
For each particulate test, sampling was made for five minutes
at each of the eight points across the duct through the five ports.
Specified data [A.I] were recorded every 2.5 minutes on a test data
sheet. The high flow rates encountered at the far wall of the duct
and across the top of the duct necessitated use of probe nozzles of
diameter less than 0.25 inches. Nozzles of diameter 0.219 inches and
0.188 inches were used during certain tests.
At the conclusion of a test the glass fibre filter was
removed from its holder and placed in a labelled petri dish. The
volume of water in the impingers was measured and the contents transferred
-------
6)
VACUUM
GAUGE
IN-STACK FILTER
PROBE
HEATED TEFLON LINE
ROTAMETER
TEDLAR BAG
LUNG
PUMP
FIGURE A.A. GRAB BAG SAMPLING EQUIPMENT
-------
62
HEATED AREA -
r FILTER
HOLDER
THERMOMETER
CHECK
VALVE
r
REVERSE-TYPE
PITOT TUBE
PITOT
MANOMETER
[CYCLONE_
(OPTIONAL)
IMPINGERS_
"ICE BATH
l
THERMOMETERS^ )
BY-PASS
VALVE
VACUUM
0AUGE
\=^
»'
C>RYTEST
METER
FIGURE A.5. PARTICULATE SAMPLING TRAIN
-------
to polyethylene bottles, which were then labelled. The probe and
nozzle were rinsed with distilled water as the inside of the probe was
carefully brushed. The rinsings were collected in a polyethylene bottle.
The glassware between the end of the probe and top of the filter holder
was washed with distilled water and these washings added to the probe
rinse. The bottom of the filter holder and glassware between filter
and impingers was also washed and the contents added to the impinger
catch.
After filtering the probe rinse, drying and weighing the
insoluble particulate material, weighing the reconditioned filter, and
determining the soluble portion of particulates in the probe rinse by
aliquot evaporation, the total weight of collected particulate material
was obtained. The moisture removed from the sampled gas was found by
adding the gain in weight of the silica gel to the volume of additional
water collected in the impingers. The volume of gas sampled was corrected
to standard conditions and the particulate loading and emission rate
calculated by use of appropriate equations [A.I].
A.5.2.2 Gaseous train. It was originally intended to sample for
gaseous organic chloride compounds using the EPA-type particulate train,
with specific absorbents replacing water in the impingers. Due to the
high flow rates (0.5 - 1.0 cfm) required for isokinetic sampling, however,
it was considered that the collection efficiency of any absorbent for
volatile organic compounds would be very low. An experimental program
was, therefore, devised and carried out at ORF to evaluate collection
methods and develop a suitable sampling train for gaseous organic chlorides.
Details of this study, performed under contract to the Ontario Ministry
of the Environment, are given in Appendix H. Data obtained which are
relevant to this section are described below.
A particulate train was set up in the laboratory and heated
air was pulled through the complete system at a flow rate of 0.5 cfm.
Water was placed in the first impinger and solvents such as xylene,
toluene and decane, containing a few ppm of CHCl^, CH2C12 °r CHC12CH2C1'
were placed in the second impinger. The third impinger was left empty
and the fourth contained silica gel. After a period of three to four
hours, the time anticipated for a particulate test, the impingers were
-------
examined with respect to solvent volume and organic chloride concentra-
tion. In all cases, it was found that the loss of chloride compounds
was considerable. For the polar solvents, such as toluene or xylene,
the chloride concentration was retained but the total loss of solvent
plus chlorides was in the range of 50%-70%. For the non-polar solvents,
such as decane, the volume loss of solvent was negligible but the chloride
loss was in the range of 60%-3Q%. Even if efficient collection of
trace organic chlorides from the gas stream by specific solvents were
possible, these compounds would not be retained in solution at gas
flow rates of 0.5 cfm or higher.
Recent studies have shown that collection and retention of
trace organic compounds is possible using specific adsorbent materials
[A.3]. Inert materials, such as the Chromosorbs, are considered to
have certain advantages over activated carbons in that water vapour
present in the gas does not hinder organic compound adsorption, and
adsorbed compounds are readily removed by thermal desorption. Experi-
ments using the particulate train with Chromosorb 102 or activated
carbon in the third impinger were carried out, passing heated air
containing a few ppm of CHC1. through the system at a flow rate of 0.5
cfm. Neither adsorbent removed more than 70$ of the CHC1, initially
and, after about thirty minutes, almost all of the CHC1 was passing
through the system, it was concluded, therefore, that efficient
collection and retention of volatile organic compounds was not possible
with gas flow rates of 0.5 cfm or higher and the particulate train could
not be used for this purpose.
After further studies in the laboratory with both Chromosorb
102 and activated carbon, the sampling train for gaseous organic compounds
depicted in Figure A.6 was constructed. Tests made with an air stream
containing 15 ppm of CHC1. snowed that either adsorbent would remove
better than 95% of the CHC1. over a sampling period of four hours.
Chromosorb 102 was selected on the basis of the advantages previously
mentioned.
A 47 mm fibreglass filter was placed directly behind the probe
nozzle to remove particulate material at the duct gas temperature. The
sample gas was passed through midget impingers containing water and
-------
65
PROBE
S.S TUBING
IN- STACK FILTER
— ROTA METERS
r
.ICE BATH
^ CHROMOSORB
1 ADSORBENT TUBES
MIDGET
/IMPINGERS
WATER
SODIUM
HYDROXIDE
PRESSURE
GAUGES
PUMP
TEST
METER
FIGURE A.6. GASEOUS SAMPLING TRAIN
-------
66
caustic soda (5$ solution), respectively, to remove any HC1 and Cl»
present in the kiln emissions. After the impinger, the gas was filtered
through four Chromosorb adsorbent traps placed in parallel at a flow rate
controlled by rotameters. A flow rate of 250 ml per minute was selected
for each trap, giving a total flow of one litre per minute through the
train. The total flow over each test period was also monitored by a
wet test meter placed after the pump. Impinger solutions and adsorbent
traps were maintained at ice temperature to effect a better collection
of chloride compounds.
After each test the in-stack filter and impinger solutions were
stored in labelled containers. The adsorbent traps were removed, capped
and labelled. The probe was rinsed and the rinsings placed in a
polyethylene bottle. All samples were then transferred directly to the
ORF laboratory for analysis. The gaseous sampling train was operated
simultaneously with the particulate sampling train for each test.
A.6 Organic Chloride Analyses
The methodology used to collect kiln emission samples, in
baseline or waste burn test periods, provided four distinct types of
samples for analysis of chlorinated organic compounds. These sample
types and the analyses required were:
- chlorinated waste feeds for compositional analysis;
- grab bag samples for any chlorinated organic species;
- Chromosorb adsorbent samples for volatile low molecular
weight organic compounds; and,
- solvent extracts of filters and solutions for noncombusted
waste components.
A.6.1 Sample preparation
Waste feed. Samples of the prospective WBA feed material were
supplied in advance of the test burn for analyses. When the test period
arrived, however, insufficient quantities of some components were available
to make up the specific blend of waste feed. The composition of material
actually delivered to SLC was, therefore, different from the sample
supplied to ORF in advance. In order to obtain a true compositional
analysis of the material being supplied to the kiln, and to determine
-------
67
feed consistency, aliquots of the waste feed were collected on specific
sampling days during the burn.
Grab bag. Emission samples collected in Tedlar bags were
analyzed directly for gaseous chlorinated hydrocarbons by injecting
syringe samples taken from the bag into a gas chromatograph. After direct
analysis, the contents of each bag were drawn through a glass tube
containing Chromosorb 102 adsorbent by means of a small pump. Any trapped
organic compounds were then thermally desorbed from the adsorbent tube
into an evacuated gas jar. The gas jar contents were then subjected to
GC analysis.
Adsorbent tubes. Chromosorb 102, being readily available in
amounts required for this study, was selected as a suitable adsorbent
for low molecular weight organohalides. Experiments were performed in
the laboratory to confirm the suitability and efficiency of this adsorbent
for the collection of volatile organohalides from an air stream. Thermal
desorption of adsorbed components into an evacuated gas jar was selected
as a method for preparing samples for GC analysis. Initial tubes were
prepared with fresh Chromosorb 102 from the bottle without conditioning.
It was determined after the first baseline test period that conditioning
of the adsorbent would be necessary to reduce the amount of bleed
produced on heatjng which tended to produce excessive background noise
during GC analysis.
Adsorbent tubes were made up by packing Chromosorb 102 in glass
containers H.5 cm in length and 11.0 mm inside diameter, using silanized
glass wool plugs at either end for support. Tubes for the WBA and WBB
burns were preconditioned by heating to 200 C and passing a stream of
nitrogen at kO ml/min through them for four hours. Chromosorb 102 used
in tubes for the WBC and BLB tests was extracted with acetone in a
Soxhlet for 18 hours and then treated at 2^0°C with nitrogen for 12
hours prior to use.
Adsorbed components on the Chromosorb 102 after each test were
removed by thermal desorption into an evacuated gas jar of 500 ml
capacity. The adsorbent tube and gas jar were connected by Teflon
-------
68
tubing and the tube heated to a fixed temperature of 170 C using a
heating tape controlled by a variable transformer. When the fixed
temperature was reached the stopcock of the gas jar between the jar
and the tube was opened, and heating of the tube continued for a further
15 minutes. The stopcock was then closed, the jar removed from the
connecting tubing and taken outside the laboratory to fill up with clean
air. Previous studies of thermal desorption using this removal
technique in the laboratory have shown that recoveries in excess of 30%
may be expected for adsorbents with adsorbed organohalides.
Solution and filter extracts. Pentane or hexane was used as an
extractant for organic compounds from all filters and probe rinse
insoluble fractions, using a Soxhlet apparatus. The same solvents were
used to extract probe rinse solutions, and aqueous sodium hydroxide
and impinger solutions, using a liquid/liquid partitioning procedure.
The extracts were then dried over anhydrous Na.SO., filtered, and
concentrated by evaporation using a combination of rotary and Kontes
tube heaters. The concentrate was made up to a known small volume with
solvent for GC analysis.
Pentane was used as extractant for BLA, WBA and WBB test
samples, since its high volatility would minimize sample loss of low
molecular weight compounds during evaporation. Hexane was used as
extractant for WBC and BLB test samples.
A.6.2 Sample analysis
All samples were analyzed by gas chromatography with flame
ionization (Fl) or electron capture (EC) detectors. Gas chromatography -
mass spectrometric (GC-MS) analysis was performed on waste feed samples
to confirm the identity of major components. The various conditions
and columns used are summarized in Tables A.5 to A.7- The mass
spectrometer used was an AEI MS-30 instrument equipped with an electron
bombardment ion source. Between the GC and the MS the interface is
of all glass design with a silicone molecular membrane. GC columns similar
to those described in Tables A.5 to A.7 were used at approximately the
same instrument conditions.
-------
69
TABLE A.5- GC ANALYSIS - SYSTEM A
Co Iumn
Column Temperature
Injector Temperature
Detector Temperature
Detector
Flow Rates
Range and Attenuation
Chart Speed
- Chromosorb 102 (80/100 mesh)
6' x 1/8" SS
- 180°C
- 215°C
- 215°C
- FID and EC
- N at 40 ml/min
Air and H. adjusted for maximum
sens!tivi ty
As required
As required
TABLE A.6. GC ANALYSIS - SYSTEM B
Column
Column Temperature
Injector Temperature
Detector Temperature
Detector
Flow Rates
Range and Attenuation
Chart Speed
- 15% SE 30 on Chromosorb W (AW; HDMS;
60/80 mesh) 12' x 1/V SS
- 60°C isothermal for 20 minutes then
programmed at 10°C/min to maximum
temperature
- 190°C
- 230°C
- FID
- N2 at AO ml/min
Air and H? adjusted for maximum
sens i tivity
As requi red
As required
-------
70
TABLE A.7. GC ANALYSIS - SYSTEM C
Column
Column Temperature
Injector Temperature
Detector Temperature
Detector
Flow Rates
Range and Attenuation
Chart speed
- k% SE 30/6* QF 1 on Chromosorb W (HP;
60/80 mesh) 6' x 1/8" SS
- 200°C
- 250°C
- 230°C
- Linearized EC
- N at 25 ml/min
- As required
- As required
These conditions are used for routine PCB analyses. The same parameters
were used for WBB waste feed except that an isothermal temperature of
155 C was maintained.
A.6.3
Blanks and standards
For trace analytical studies it is necessary to avoid
contamination of samples at all stages of sample collection and
preparation. Though extreme care was taken during the study it was
not possible to pre-extract all hardware and chemical reagents used.
Due to the high sensitivity of the EC detector to many compounds, some
of the blank extracts, therefore, gave complex GC-EC profiles, which had
to be subtracted as background from the sample chromatograms. The
following blanks were obtained and analyzed for this purpose.
i) Particulate train:
i!) Gaseous train:
iii) Solvents:
(a) Particulate filter + probe rinse
filter + Soxhlet thjmble.
(b) Water
(a) In-stack filter •»• probe rinse
filter + Soxhlet thimble
(b) Water
(c) Aqueous NaOH
(a) Pentane (100 ml concentrated to
1 ml)
(b) Hexane (100 ml concentrated to
1 ml)
-------
71
iv) Polyethylene containers: (a) Solvent extracts (100 ml concentra-
ted to 1 ml)
Standard gas samples of various components of interest, such as
CC1,, CHC1,, CH Cl- and 1,2-dichloroethane, were made by injecting 50
yl aliquots of these compounds into an evacuated 500 ml gas jar through
a silicone septum. Clean air was then allowed to enter the jar to
attain atmospheric pressure. The sealed gas jar, therefore, represented
nominally a 100 ppm standard. Gas standards of lower concentration were
made by dilution of the 100 ppm standard using a similar procedure.
These standards were used immediately after preparation and were not
retained for use on the following day.
Solution standards were made by weighing accurately known
amounts of the components of interest and dissolving them in a known
volume of pentane.
A.6.** Concentration factors
Desorbed compounds from the adsorbent tubes were concentrated
into a 500 ml gas jar. For each particular test, therefore, the concentra-
tion factor would be the volume of emission passed through the tube
divided by 500 ml. The volumes sampled for each tube varied from 1.5
cubic feet to 3.22 cubic feet during the complete test program giving
a range of concentration factors from 85 to 181. Thus, a concentration
for a particular component of 1 ppm in the gas jar sample would mean a
concentration of about 10 ppb in the kiln emission, assuming a 100
percent collection efficiency, and subsequent desorption of the adsorbent.
Most of the extracted samples were concentrated into a 2 ml
volume of solvent. The following example illustrates how the concentration
factors were determined for various components detected. Assume that
CH Cl was detected at a concentration of 1 ppm in the solvent extract.
This corresponds approximately to a weight of 2 ug of the compound in
2 ml of solvent. The molecular weight of CH2C12 is 85- Therefore, 2 pg
of CH Cl at 70°C, the temperature at which the emission sample was
measured, occupies a volume of:
7- x 2k litres, or r x —— cubic feet.
85 x 106 85 x 10° 28.3
-------
72
For test WBA 1, for example, a volume of 142.8l scf was sampled. Thus,
n
6
the concentration of CH Cl- In the kiln emission is:
x x ppm or 0.14 ppb.
85 x 1(T 28.3 U2.81
The concentration factor is therefore 10 _ • 7140.
0.14
Factors were calculated for feed compounds of lowest and
highest molecular weight.
A.6.5 Samp1e chroma tog rams
Analysis for low molecular weight organohalides and noncombusted
chlorinated organic compounds present in any collected samples was made
using GC techniques only, by selecting appropriate GC parameters and
using retention time data. No specific cleanup or separation procedures
were performed in order to segregate components of interest from possible
interfering compounds because, although GC profiles obtained were quite
complex, very low concentrations of organic compounds were evident from
the peak heights obtained for the attenuations used. Estimates of specific
compound concentrations in waste burn samples were made by subtracting
blank and baseline levels and comparing with a standard for that compound.
Waste feeds. GC-FID profiles for the three waste feed materials
burned in this study are given in Figures A.7 to A.9- Identification
of the main components in the respective waste feeds are presented in
Tables 1, 2 and 3 in Section 5 of the report.
It was anticipated that the levels, if any, of uncombusted
components in the stack emissions would be very low. Therefore GC-CC
analysis was used to evaluate the organic solvent extracts of the
various impinger samples from the sampling trains for the presence of
any uncombusted components. Figures A.10 and A.11 represent GC-EC profiles
for diluted waste feed material for WBB and WBC, respectively. WBB
also contained approximately 50% o-chlorotoluene which has a weak
response to the EC detector. Therefore, in this instance the solvent
extracts were also analyzed by GC-FID in order to determine whether any
uncombusted o-chlorotoluene was present.
-------
8
>-
h-
i/v
z
LU
H-
Z
SAMPLE FEED -OCT. 28th1975
GC-SYSTEM B
12
25
-10
10 x 256
10
15
20
25
30
35 40
TIME (min)
FIGURE A. 7 GAS CHROMATOGRAPHIC PROFILE FROM FLAME IONIZATION DETECTOR
FOR CHLORINATED ALIPHATICS (WBA) SAMPLE FEED
-------
SAMPLE FEED - DEC. 12th 1975
GC-SYSTEM B
25
30
35
26
40 45
TIME(min)
FIGURE A.a GAS CHROMATOGRAPHIC PROFILE FROM FLAME IONIZATION DETECTOR
FOR CHLORINATED ALIPHATICS PLUS AROMATICS AND ALICYCLICS
(WBB) SAMPLE FEED
-------
10-
9
> 8
16
SAMPLE FEED-JAN. 8th1976
GC-SYSTEM B
10"11x256
—r—
5
10"1(i256
10^.128
10 15 20 25 30 35 40 45
50 55
TIME (min)
FIGURE A. 9 GAS CHROMATOGRAPHIC PROFILE FROM FLAME IONIZATION
DETECTOR FOR CHLORINATED ALIPHATICS PLUS AROMATICS,
ALICYCLICS AND POLYCHLORINATED BIPHENYLS (WBC)
SAMPLE FEED
-------
to
til
o-
9-
8"
7-
6-
5-
4-
3-
2-
1-
0:
I
1
py
10
9
SAMPLE FEED-DEC.12th1975
GC-SYSTEM C 8.
/
0246
*
/ 6.
5 .«./ aliquot
(Temp.155°C)
4
3-
I
I
2-
1-
0-
r
8 10 12 14 0
TIME(min)
lo-
g-
s'
A
7
07 / i 6"
5^^ aliquot
1x32
(Temp.155°C)
*•
3-
2-
/ ^"
1
/
0.7 ng /<>«•<•
5 <>i/ aliquot
1x64
(Temp.190°C)
I-
r
2468 10 02468
TIME(min) TIME (mm)
FIGURE A.10 GAS CHROMATOGRAPHIC PROFILES FROM ELECTRON CAPTURE
DETECTOR FOR WBB SAMPLE FEED
-------
10-
9
8-
7-
V- 6-
10
5-
LU
h-
Z
3-
o_
1-
1x128
STANDARD AROCLOR 1242
GC SYSTEM C
5-
3-
2-
1x256
8 10 12 0
TIME (min)
WBC-AROMATICS + PCBs
SAMPLE FEED JAN. 8th 1976
GC SYSTEM C
9
8 10
12 14
TIME (min)
FIGURE A. 11 GAS CHROMATOGRAPHIC PROFILES FROM ELECTRON CAPTURE
DETECTOR FOR STANDARD AROCLOR 12^2 AND SAMPLE FEED WBC
-------
78
Standards. Figure A.12 shows the GC-IC profile for the low
molecular weight chlorinated hydrocarbons, e.g. CH.Cl , CHC1., CC1, and
1,2-dichlorethene. Figure A.13 snows the GC-FID profile for the same
compounds to indicate the greater sensitivity of the EC detector.
A GC-EC profile for 5 ng of standard Aroclor 1242 injected
into the column is shown in Figure A.11.
Waste Burn Samples. Figure A.12 shows GC-EC profiles for
desorbed gas samples for BLB and WBC test samples. The main difference
between the samples is in the intensity of the peaks of interest.
The GC-EC profiles for all the solvent extract samples showed
the presence of small peaks. When the profiles were compared to those
obtained from the BLA samples and the various control blanks (cf. Section
A.6.3), no peaks peculiar to the waste burn samples were detected.
BLB samples became contaminated with WBC feed material and were discarded.
In most Instances, the EPA impinger extracts showed the largest
response to the EC detector. Figures A.\k and A. 15 show GC-EC profiles
for the organic solvent extracts of the EPA impinger samples obtained
for BLA-T3 and WBC-T3, respectively. Profiles obtained for the two
samples are quite similar. There is a large EC response occurring as an
unresolved peak in the early portion of the chromatographic profile.
This limited the sample size that could be used for injection, and
restricted the use of high sensitivity settings on the gas chromatograph.
It is normal practice when analysing for PCB's to perform
cleanup and separation procedures in order to segregate the PCB's from
interfering components and thus facilitate their analysis. Such a
cleanup and separation procedure was performed on WBC-T3 using Florisil
adsorbent (2?) in order to determine whether the early unresolved peak
(Figure A.15) could be removed. After cleanup, it was possible to use
a larger aliquot for injection and higher sensitivity settings on the
chromatograph. The GC-EC profile of the cleaned up WBC-T3 sample is shown
In Figure A.16,
The GC-EC profile shown in Figure A.16 was obtained on ?n
organic solvent extract of a sample that should contain PCB's if any were
present in the kiln emissions. The profile shows the presence of a few
small peaks. When compared with the Standard Aroclor profile (Figure A.11),
-------
10-
LU
GC-SYSTEM-A
GRAB SAMPLE
WBC adsorbed/desorbed
on Chromosorb 102
(1 ml aliquot )
BLB-2B
(0,5ml aliquot )
WBC-3 A
(0.2 ml aliquot
2 <. 6 02
TIME (min)
u>
8 10
FIGURE A.12 GAS CHROMATOGRAPHIC PROFILES FROM ELECTRON CAPTURE
DETECTOR FOR LOW MOLECULAR WEIGHT CHLORINATED HYDROCARBONS
AND FOR BLB AND WBC TEST SAMPLES
-------
80
9-
8-
>
h-
- 7^
LU
6'-
STANDARD
{ 1 ppm )
GC-SYSTEM-B
1,2-Dichloroethane
10"1x256
FIGURE A.13
) 5 10 15
TIME (min)
GAS CHROMATOGRAPHIC PROFILE FROM FLAME IONIZATION
DETECTOR FOR LOW MOLECULAR WEIGHT HYDROCARBONS
-------
8
5-
3-
2-
1-
81
GC SYSTEM C
injection
—i 1 1 1 1 1 1 1—i 1 1—
10 15 20
TIME (min)
FIGURE A.U GAS CHROMATOGRAPHIC PROFILE FROM ELECTRON CAPTURE
DETECTOR FOR IMPINGER EXTRACT FROM BLA TEST 3
-------
82
101
9-
8-
7-
tn
z
HI
6-
5-
4-
3-
2-
1-
1x64
GC-SYSTEM-C
injection
10
15 20
TIME (min)
FIGURE A.15 GAS CHROMATOGRAPHIC PROFILE FROM ELECTRON CAPTURE
DETECTOR FOR IMPINGER EXTRACT FROM WBC-TEST 3
-------
83
10 -{
9-
8 -I
w- *7
to
z
HI
5-
3-
2-
1-
1x32
o
GC-SYSTEM C
INJECTION
10
15 20
TIME (min)
FIGURE A. 16 GAS CHROMATOGRAPHIC PROFILE FROM ELECTRON CAPTURE
DETECTOR FOR IMPINGER EXTRACT FROM WBC TEST 3 AFTER
CLEANUP AND SEPARATION
-------
and taking into account that the extract represents a >7000 concentration
factor, if any of the peaks in Figure A.16 represent PCB components the
levels must be extremely low.
A.7 Analysis of Process Solids
For the purpose of obtaining a mass balance during the
experimental chlorinated hydrocarbon burn, it was required to carry out
chemical analyses of clinker, slurry feed and discard dust for chlorine
and potassium. Sulphur was determined on process solids and the return
dust samples were analyzed for changes which may have affected kiln
operation.
X-ray fluorescence (XRF) analysis has been extensively used
for determination of sulphur, potassium and chlorine in cement, cement
raw materials and related materials, and was selected for use in this
study.
A.7.1 Analytical procedures
Samples were prepared by grinding 10 grams of sample for 90
seconds in a SPEX Shatterbox with tungsten carbide containers and pucks.
From this material a 1.25 inch diameter pellet at 8 tons pressure was
made in SPEX aluminum sample caps.
For the X-ray fluorescence method to be applied, there must
be a linear relationship between elemental concentration (over the full
range of concentration required) and the measured fluorescence intensity.
The slope of the line representing this relationship then can be used
directly as a calibration constant, usually in the form of the ratio, counts
per second:percent element. When a calibration line is generated by
known additions of the test element to a material being analyzed ("spiking"),
the line intercept gives the percent of the element present in the
original sample prior to any additions.
In addition to linearity, reproducibi1 Ity of the calibration
constant is a prerequisite if it is to be applied across a range of
materials having some degree of matrix variability. Both conditions
were examined prior to the final development of the analytical procedure.
The calibration curves for chlorine content of clinker and
raw meal-slurry feed were prepared by addition of standard lithium
-------
85
chloride in alcohol solution to samples of these materials. Dust samples
were analyzed by the standard Volhard method for calibration of the
chlorine curve and the standard gravimetric method of precipitation with
barium for calibration of the sulphur curve (Table A.8). For calibration
of curves for sulphur in clinker and raw meal/slurry feed, determinations
were made on a Leco Induction Furnace Model 523CS with Leco Automatic
Titrator Model 517 (Table A.9). For calibration of the potassium curve,
all potassium results, as well as those for sodium, were obtained from
analyses on a Perkin-Elmer Model 460 Atomic Absorption Spectrophotometer
in the flame emission mode (Table A.10).
Examination of these samples yielded linear calibration
relationships between X-ray fluorescence count rates and percent element.
The least squares computations for each set of data gave the slopes,
intercepts and regression (correlation) coefficients listed in Table A.11
TABLE A.8. GRAVIMETRIC DUST ANALYSES
Dust Type Date
Discard 19/10/75
11/11/75
31/11/75
4/12/75
5/12/75
6/12/75
7/12/75
5/ 1/76
6/ 1/76
9/ 1/76
14/ 1/76
16/ 1/76
Return 5/12/75
13/12/75
I/ 1/76
18/ 1/76
% Cl
2.00
2.08
5.97
2.89
7.11
4.69
6.18
4.71
3-72
4.43
0.46
0.90
2.11
4.34
2.10
1.00
% 503
-
3-17
2.36
6.53
12.12
7.27
6.90
5.08
4.23
4.18
4.65
4.32
7.35
4.22
4.99
5.66
-------
86
TABLE A.9. RESULTS FROM LECO INDUCTION FURNACE ANALYSES
Material Date
Clinker 19/10/75
" 21/10/75
23/10/75
31/10/75
1/11/75
5/H/75
8/12/75
" 31/12/75
11 I/ 1/76
ll/ 1/76
20/ 1/76
Slurry Feed 13/10/75
20/10/75
27/10/75
9/11/75
6/12/75
9/12/75
13/12/75
14/12/75
31/12/75
3/ 1/76
8/ 1/76
21/ 1/76
% SO (Total S as)
0-93
1.34
0.75
0.30
0.16
0.66
0.93
1.39
1.43
1.04
1.42
0.45
0.42
0.41
0.45
0.52
0.48
0.60
0.55
0.43
0.47
0.46
0.47
-------
87
TABLE A.10. RESULTS FROM ATOMIC ABSORPTION ANALYSES
D!
Material
scard Dust
ii
ii
ii
it
ii
1 1
ii
ti
ii
Return Dust
Cl
SI
ii
ii
inker
ii
ii
ii
ii
1 1
ii
ii
ii
1 1
1 1
urry Feed
ii
ii
n
ii
1 1
1 1
1 1
1 1
1 1
Date
11/11/75
31/11/75
VI 2/75
6/12/75
7/12/75
5/ 1/76
6/ 1/76
9/ 1/76
IV 1/76
16/ 1/76
5/12/75
13/12/75
I/ 1/76
19/10/75
21/10/75
23/10/75
31/10/75
1/11/75
5/11/75
8/12/75
31/12/75
I/ 1/76
ll/ 1/76
20/ 1/76
13/10/75
20/10/75
27/10/75
9/H/75
6/12/75
9/12/75
31/12/75
3/ 1/76
8/ 1/76
21/ 1/76
% K20
6.36
10.30
7.14
8.23
9.96
8.12
6.43
7.9*
4.84
4.18
7.79
5.54
6.02
1.32
1.61
1.00
0.54
0.40
1.01
0.64
1.18
1.31
1.10
1.45
0.93
0.93
0.89
0.88
0.91
0.90
0.90
0.91
0.93
0.94
% Na20
0.32
0.39
0.36
0.44
0.45
0.43
0.34
0.39
0.35
0.34
0.42
0.41
0.41
0.23
0.25
0.20
0.21
0.21
0.21
0.20
0.21
0.24
0.22
0.26
0.19
0.22
0.18
0.20
0.19
0.22
0.18
0.22
0.19
0.22
-------
TABLE A.)]. LEAST SQUARES DATA FOR CALIBRATION LINES
Material
Raw Meal H
Element
h
Slurry Feed Potassium (as K20)
Chlorine
Clinker
Dust
A. 8
Sulphur (as SO,)
Potassium (as ICO)
Chlorine
Sulphur (as SO,)
Potassium (as K.O)
Chlorine
Sulphur (as SO,)
Correlation
Coefficient
0.9172
0.9995
0.9913
0.9829
0.9999
0.9917
0.9893
0.9548
Determination of Heat Value, Chlorine
Content in No. 6 Fuel
Slope
27128
5260
739
24980
4781
400
3166
1927
278
Content and
Intercept
-0.01
0.060
0.27
-0.01
0.011
-0.23
-0.12
-0.14
-0.10
Sulphur
The heat value was determined using standard methods on a Parr
Calorimeter Model 1241 equipped with Parr oxygen bomb Model 1108. Sulphur
was determined using the standard barium precipitation method on the
washings from the bomb. Where chlorine content was required, the
washings were analyzed for chlorine by the standard Volhard method.
A.9 Determination of Heat Value and Chlorine Content in Chlorinated
Hydrocarbons
The heat value was determined on a Parr calorimeter model 1241
equipped with Parr oxygen bomb model 1108 following standard operating
procedures for determining heat value of oil. Due to the corrosive nature
of the combustion products, it is recommended that the chlorinated
materials be diluted with standard oil. It has been found that, while
corrosion of the interior of the bomb is extensive, burning these materials
without dilution gives similar results. Time between weighing and
igniting is extremely important due to the volatile nature of some of
these compounds.
-------
89
After igniting and determining the heat value of the chlorinated
material, the interior of the bomb is thoroughly washed with distilled
water into a volumetric flask. Aliquots of the proper size for the
chlorine content expected were taken. Nitric acid is added to allow a
more distinct end point, and the chlorine is determined by the standard
Volhard titration.
REFERENCES
A.I Ontario Ministry of the Environment, Source Testing Code.
January, 1973-
A.2 Standard Reference Methods for Source Testing: Measurement of
Emissions of Particulates from Stationary Sources. Environmental
Protection Service Report EPS 1-AP-7A-1, Air Pollution Control
Directorate, Environment Canada, February, 197^.
A.3 Pellizan, E.D., J.E. Bunch, and B.H. Carpenter. Env. Sci.
Technology 9 (6), 552-560 (1975).
-------
APPENDIX B
ANALYTICAL DATA, CALCULATION AND DETAILS OF
EXPERIMENT ON THE SUSPENSION PREHEATER KILN
-------
93
APPENDIX B
ANALYTICAL DATA, CALCULATION AND DETAILS OF
EXPERIMENT ON THE SUSPENSION PREHEATER KILN
B.I Theoretical
The excellent article by Weber cited in the body of the report
describes in detail the problem of alkali reduction in a suspension
preheater kiln, and was used in preparing the following notes.
At material temperatures above 800 C in the rotary kiln,
alkalies (K 0 and Na.O) are volatilized from the material being burned
and are carried along with the exit gases to the kiln inlet. These
condense at gas temperatures below 900 C. In part, they are recaptured
by the material, particularly in the preheater, and are thus carried
back to the sintering zone, so that an internal alkali cycle is established.
The alkalies are in part also precipitated at the transition from the
preheater to the kiln.
Coatings thus formed interfere with material and gas flows
and, in the most severe cases, completely plug the system.
In preheater kilns, the raw meal is heated in the finely divided
condition in the gas stream and these particles act as condensation nuclei
for the alkalies. Between 81 and 37% of the K20 volatilized in the kiln
is trapped in the preheater.
The greater the alkali cycle, and therefore the concentration
of alkali in the gas, the more alkali will condense on the bottom
ducts and cyclone stages. If the alkali is very volatile, accretions
are likely to occur in the preheater. However, the alkali cycle can be
reduced by the provision of a bypass, that is by drawing off a portion
of the gas at the bottom of the preheater. (See Figure 3, Section 2 of
the report.)
Of alkali which enters the kiln system with the raw meal, a
proportion (e ) will be volatilized and the remaining (1- EJ) will be
discharged with the clinker. If the proportion of gas withdrawn through
the bypass is V, then of the volatilized alkalies, the proportion eyV
will be removed by the bypass. The remaining portion e. (l-V) will return
with the feed, thereby giving rise to the internal cycle in the kiln.
-------
The alkali in the internal cycle has a different volatility e,, than
alkali from the raw meal which has not yet been volatilized. By defining
total alkali including raw meal alkali into the kiln by K, then the quantity
of alkali from the cycle alone is K-l. Of this quantity, (K-l)-e -V is
removed by the bypass.
Equilibrium is attained when the total quantity of alkali input
with raw meal and fuel is equal to the quantity output by the bypass and
retained with the clinker. This is most easily determined by considering
only alkali in the cycle. Entering the cycle is the volatile raw meal
alkali « e.; leaving the cycle are a proportion in the clinker = (K-l)-(l-e )
and a proportion through the bypass - [e. + (K-l) e_)]-V.
By setting the quantities input and output from the cycle as
equal, the equation may be solved to determine the amount of gas the
bypass must remove to keep the alkali cycle from becoming excessive.
B.2 Experience at St. Lawrence Cement
At St. Lawrence Cement, samples are taken of the material
between the Stage IV cyclone and the kiln inlet (see Figure 3), this
material being called, not quite appropriately, "Stage IV dust". From
experience at St. Lawrence Cement, it was known that a chloride content
of 2.5% chlorine in this material (natural basis) is excessive. The
quantity of gas required to be withdrawn by the bypass to maintain lower
chloride levels was calculated on the assumption that the chloride
volatilities were e. * 0.99 and e * 1.0 (completely volatile). These
quantities are given in Table B.1.
While the bypass was designed to remove the quantities of
gas required for the program, attempts to reach this level resulted
in excessive emissions of particulate matter from the bypass precipitator.
These were attributed to excessive gas flows. To alleviate this problem,
measures to increase cooling water and decrease the quantity of ambient
air were undertaken, with limited success.
-------
95
TABLE B.I. PERCENT BYPASS GAS REQUIRED TO MAINTAIN CHLORIDE LEVELS
Chloride
addi tion
relati ve
to cl inker
0
0.05
0.10
0.15
0.20
0.30
0.^0
B.3
% K20
i n
cl i nker
1.31
1.2A
1.18
1.11
1 .Ok
0.91
0.77
Equation of
Bypass
1 .56% or 2,Q%
nat. ign.
2.1
1».6
7.1
9.6
12.1
17-1
22.2
Approximated T
percent required to
Stage IV Cl at:
2.0% or 2.561
nat. ign.
1.6
3.6
5.5
7.5
9.1»
11.7
17.2
ime for Equi 1 ibr ium
ma intai n
2.5% or 3-2%
nat. ign.
1-3
2.9
k.k
6.0
7-0
10.6
13-8
It was essential to the success of the program to determine
the time required to attain the equilibrium state. To obtain an approxi-
mation of the time required, and realizing that it is only an approxima-
tion since retention time in kiln and absorption phenomena influence
this time, the following approach was taken:
dt
where:
dQ
dt
Cl
C1RM " C'clinker " C1bypass
change in quantity with time
(pounds/minute)
RM
chlorine/time in with raw meal (pounds/minute)
d = chlorine/time out with clinker (pounds/minute)
clinker
bypass
ClRMand C'clinker
combined as:
chlorine/time out through bypass (pounds/minute)
are approximately constant, hence they can be
cl
- r i - r i
~ U RM clinker
-------
96
and C1U = V0_ (t)
bypass
where:
Q(t) « quantity in transition chamber at time (t) in pounds
V » bypass valve proportion.
Then:
$ - ' , - VQ (t)
dt cl
Solving for time gives:
In [I - Vd (t)] QQ
t = — minutes.
V Q
Since In 0 » °°, QF must be taken as a very small distance from
the true equilibrium.
The sampling sequence at Stage IV was planned from this model.
Q.k Sampling and Analytical Methods
Analytical methods were those described in Appendix A. Samples
of clinker, raw meal feed, bypass precipitator dust and conditioning
tower dust were taken at two hour intervals. These were then blended
into 24-hour composite samples. On the basis of the calculation in
Section B.3, samples were taken from Stage IV at 3 to 10 minute
intervals when attempting to start the chlorinated burn. These latter
samples were analyzed individually.
B. 5 Pi scussion
In both cases described, the bypass was operated to withdraw
9 percent of the gases from the preheater tower. This was the maximum
attainable due to the emission rate of particulate matter. On the first
attempt to burn chlorinated hydrocarbons, the quantity to yield 0.2^
percent chlorine relative to clinker was fed starting at 09:38, June 3,
1975. Plugging of the duct between the kiln and the conditioning tower
caused a bypass shutdown on June 5, 1975- It was not possible to
maintain feed of chlorinated material to the kiln unless the bypass was
functioning. On June 10, 1975 a second and similar attempt was started
-------
97
at 11:20. The bypass again failed due to accretion on the duct on
June 13, 1975- On the second attempt, sampling of Stage IV material
was continued for a longer period on the basis of the first set of
results.
It was realized with the failure that major revisions to the
bypass system would be required to continue this experiment on the
suspension preheater kiln.
B.6 Analytical Results
Analytical results are given in Tables B.2 through B.*». The
raw meal feeds showed a gradual increase in both the content of chlorine
and K_0. The results from the clinker analysis (Table B.3) showed that
the burning of chlorinated hydrocarbons resulted in a reduction in the
K20 content (June 3-5, 1975 and June 10-12, 1975).
The results of the analyses from Stage IV dusts (Table B.^)
show that the chlorinated hydrocarbons were destroyed in the burning
process. Chloride content of these samples increased while burning
chlorinated wastes. This finding could not have resulted without
breakdown of the hydrocarbons since the raw material is at too high
temperature at this position in the preheater for the chlorinated
hydrocarbons to condense upon it. By the methods described in Section
B.3, the approximate time to equilibrium was calculated to be about
90 minutes. It was also expected that an additional time of approximately
20 minutes would be required to travel through the kiln. The actual
curve is less steep and equilibrium is reached in about four hours.
For the June 3, 1975 burn, the curve was extrapolated to the calculated
equilibrium (Figure 8.1). On June 10, 1975 samples were taken for a
longer period. Although there are points above the calculated curve
(Figure B.2), the data from June 11, 1975 show that equilibrium was
reached; the higher values found indicate random sampling or
analytical errors.
-------
98
TABLE B.2. RESULTS FROH ANALYSES OP DRY PROCESS KILN RAW
Date
30/5/75
31/5/75
1/6/75
2/6/75
3/6/75
4/6/75
5/6/75
6/6/75
7/6/75
8/6/75
9/6/75
10/6/75
11/6/75
12/6/75
13/6/75
14/6/75
15/6/75
% Cl
0.067
0.069
0.073
0.071
0.07**
0.082
0.079
0.094
0.087
0.093
0.088
0.084
0.092
-
0.084
0.104
0.102
% SO
(Total S-'as)
0.60
0.60
0.60
0.60
0.60
0.60
0.60
0.62
0.60
0.60
0.58
0.60
0.60
NO SAMPLE
0.60
0.60
0.62
% K20
0.92
0.95
0.93
0.93
0.93
0-95
0.96
0.98
0-97
0.97
0.97
0.95
0.98
-
0.96
1.00
1.00
-------
99
TABLE B.3. RESULTS FROM ANALYSES OF DRY PROCESS KILN CLINKER
Date
30/5/75
31/5/75
1/6/75
2/6/75
3/6/75
4/6/75
5/6/75
6/6/75
7/6/75
8/6/75
9/6/75
10/6/75
10/6/75
10/6/75
10/6/75
10/6/75
10/6/75
10/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
11/6/75
12/6/75
12/6/75
12/6/75
1 2/6/75
12/6/75
12/6/75
12/6/75
12/6/75
12/6/75
12/6/75
12/6/75
13/6/75
14/6/75
15/6/75
Time
Average
1 1
1 1
1 !
II
II
1 1
1 1
1 )
1 1
1 1
1 1
1 2 : 00
14:00
16:00
18:00
20:00
22:00
Average
00:00
02:00
04 : 00
06:00
08:00
10:00
14:00
16:00
18:00
20:00
22:00
Average
00:00
02:00
04:00
06:00
08:00
10:00
12:00
14:00
16:00
18:00
Average
1 1
1 1
% Cl
0.03
0.03
0.03
0.21
0.03
0.05
0.04
0.04
0.07
0.05
0.06
0.04
0.06
0.04
0.0^
0.05
0.05
0.05
0.06
0.04
0.06
0.07
0.06
0.06
0.05
0.05
0.05
0.06
0.05
0.05
0.07
0.07
0.05
0.05
0.05
0.05
0.10
0.05
0.06
0.05
0.05
0.05
0.05
0.04
% so2
1.05
1.15
1.15
1.16
1.10
1.19
1.23
1.34
1.33
1.31
1.50
1 .22
1.38
1.00
1.07
1.30
1.03
1.11
1.29
1.18
1.36
0.98
0.96
1.10
1.08
1 .18
1.08
1.12
1.23
1 .01
1 .28
1.07
1. 12
1.11
1.10
. 1 1
.25
. 1 1
.23
. 1 1
* 1
. 14
.29
1 <•!
1 .40
1 .35
% K20
.27
1.29
1-31
1.30
1.24
1.30
-33
.44
.48
-43
.63
.29
.25
0.97
0.87
1 . 10
0.96
1.05
1 .24
1.27
1-25
1 .00
0.99
1 .10
_ _ f\
1 .08
.15
1 .13
, i
. 14
.21
f\f
.06
.29
-15
i /
. !6
- 15
* f
. 16
i O
. IO
-> £
• 36
1 - 15
t o O
i . 2o
?t -»7
- 17
i 1 O
1 . 1 0
i li f\
1 . 40
IP" f\
.52
1 f. 1
1 . 41
-------
100
TABLE B.4. RESULTS FROM ANALYSES OF STAGE IV DUSTS
Date
30/ 5/76
2/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
3/ 6/75
4/ 6/75
V 6/75
5/ 6/75
6/ 6/75
Time
-
-
09:00
09:06
09:12
09:18
09:24
09:30
09:38
09: 44
09: 47
09:50
09:53
09:56
09:59
10:02
10:05
10:08
10:11
10:14
10:23
10:32
10:41
10:54
10:59
:04
:09
:14
:19
:24
:29
:34
:35
:44
:49
:54
:59
12:04
12:09
12:14
14:00
15:30 .
09:00
14:00
-
-
% so,
% Cl (Total S as)
0.67 1.14
0.82
0.86
0.84
0.81
0.85
0.82
0.86
0.90
0.97
0.96
.02
.20
.21
.21
.22
.34
.44
• 51
.44
• 56
.79
.83
.85
1.95
2.08
2.03
2.00
2.03
2.30
2.17
2.21
2.22
2.29
2.29
2.32
2.38
2.49
2.28
2.34
2.47
2.84
1.92
1.59
0.97
1.14
.30
.23
.25
.19
.22
.18
• 30
.34
.32
.35
.38
.29
.31
.34
.30
.46
.38
.52
.29
.41
.37
.34
• 32
.46
• 59
• 35
.39
.41
.47
.36
.37
.43
.51
-57
.66
.47
.51
.46
.44
.45
.58
.42
.28
.44
.17
% K20
1.77
2.26
2.11
2.03
2.03
2.14
2.06
2.17
2.15
2.21
2.38
2.39
2.37
2.42
2.51
2.55
2.68
2.66
2.95
2.91
2.94
3.38
3.43
3.35
3.74
3.85
3.72
3.66
3-70
4.15
3.93
4.04
4.00
4.18
4.11
4.24
4.36
4.45
4.12
4.21
4.44
5-07
4.17
3.48
2.54
2.71
-------
101
TABLE B.4. (Cont'd)
Date
9/ 6/75
10/ 6/75
10/ 6/75
JO/ 6/75
10/ 6/75
10/ 6/75
107 6/75
10/ 6/75
107 6/75
107 6/75
10/ 6/75
10/ 6/75
107 6/75
107 6/75
10/ 6/75
107 6/75
107 6/75
10/ 6/75
107 6/75
107 6/75
107 6/75
10/ 6/75
107 6/75
10/ 6/75
107 6/75
107 6/75
107 6/75
JO/ 6/75
107 6/75
10/ 6/75
107 6/75
117 6/75
127 6/75
% so3
Time % Cl (Total S as)
-
09:00
09: 10
09:20
09:30
09:40
09:50
10:00
11:15
11:25
11:35
12:05
12:25
.66 1.20
.09 1.18
.17
.19
.15
.16
.16
.14
.11
.20
.46
.91
• 99
12:35 2.15
12:45 2.25
12:55 2.31
13:05 2.45
13:15 2.37
13:25 2.32
13:35 2.62
13:45 2.51
13:55 2.63
1^:15 2.57
14:25 2.87
14:35 2.74
14:45 2.85
14:55 2.77
15:05 2.79
15:25 3-02
15:35 3.28
15:45 3-09
2.58
2.42
.22
• 36
• 39
.29
.27
.29
.29
.34
.41
.49
.46
.64
.59
.66
.79
.71
.47
.67
.54
.69
.69
.81
.67
.70
.77
.52
2.41
.59
.62
• 51
.49
% K20
3-79
2.64
2-93
3-07
2.92
3-01
2-99
2-95
2.91
3-04
3-22
4.02
4.07
4.50
4.54
4.65
4.98
4.75
4.48
5.10
4.83
5.12
5.02
5.53
5.18
5-35
5.23
5.23
5.66
6.07
5.78
5.23
5-13
-------
2.80 -
2.30
UJ
z
cr
o
_i
x
o
1.80-
1,30
0.80
O
Q - CALCULATED
- ACTUAL
10:00
11:00
12:00
13:00
U:00
15:00 TIME
o
ro
CHLORINE LEVEL IN STAGE IV 3/6/1975
FIGURE B.1
-------
3,00
2.50-
LU
2
o:
O 2.00
_J
I
O
1,50
1.00
Q-CALCULATED
t)-ACTUAL
9:00
11:00
13:00
15:00 TIME
CHLORINE LEVEL IN STAGE IV 10,5/^5
FIGURE B. 2
-------
APPENDIX C
RESULTS AND CALCULATIONS FOR WET PROCESS SYSTEM
-------
107
APPENDIX C
RESULTS AND CALCULATIONS FOR WET PROCESS SYSTEM
C. 1 Mass Balance Experimentation
C.I.I Results of analyses of process materials
Analyses of process materials for the mass balance were
carried out at the St. Lawrence Cement Co. plant. Analytical results
are given in Tables C.I to C.6.
C.I. 2 Calculation of material balances
A daily record of production and material consumption is given
in Table C.7 for the period October 7, 1975 to January 21, 1976. Produc-
tion was disrupted during this period by a kiln shutdown for the removal
of a kiln ring.
The program was further disrupted by formation of a ring,
not requiring kiln shutdown, but time to reduce the quantity of aromatic
plus complex chlorinated hydrocarbon in the tank and difficulty in
scheduling deliveries. From the daily composite analysis of each material
for each element, and the total daily quantity of each material, daily
elemental quantities (as pounds or tons per 25 hour period) were determin-
ed by the relationship:
C.
where
y a M x
*ij ij 100
X.. » total daily weight of element x in material i on day j
C « concentration of element x in the composite sample material
ij
i on day j .
These data were then used to form the separate individual
elemental balance accounts given in Tables C. 8 and C-9- Due to process
fluctuations and the very low concentrations in the case of chlorine,
little significance should be attached to individual daily balance results,
except as they indicate the responsiveness of the system to major changes
in input of chlorinated hydrocarbons. Significant assessment of the
-------
108
balance is available by considering prolonged periods of plant operation.
Summary balances for this purpose are given in the body of this report
(see Section 5-A, Tables 8 and 9).
Materials entering the mass balance calculations (see Figure 1,
Section 2) were slurry feed, No. 6 fuel oil and chlorinated hydrocarbon
materials as input streams with clinker and discard dust forming the
output streams. The return dust is returned almost immediately to the
system and does not form a part of the mass balance.
-------
109
TABLE C.
RESULTS FROM ANALYSES OF CLINKER SAMPLES
Date
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
1 VI 0/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
4/11/75
5/H/75
6/11/75
7/11/75
8/11/75
9/H/75
2/12/75
3/12/75
4/12/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
10/12/75
11/12/75
12/12/75
13/12/75
14/12/75
% Cl
0.033
0.032
0.034
0.042
0.032
0.029
0.030
0.031
0.031
0.031
0.035
0.029
0.032
0.042
0.030
0.029
0.030
0.030
0.029
0.031
0.031
0.034
0.032
0.039
0.034
0.039
-
0.032
0.028
0.031
0.031
0.037
0.029
0.035
0.031
0.029
0.034
0.033
0.029
0.028
0.038
0.040
0.034
0.035
0.032
0.032
0.031
% $03
0.88
0.79
0.86
0.62
0.93
0.80
0.80
0.86
1.27
0.90
1.11
0.99
0.97
1.03
1.32
0.99
0.77
0.92
0.59
1.06
0.82
1.16
1.27
0.88
0.34
0.18
NO SAMPLE
0.26
0.30
0.64
0.70
0.71
0.68
0.72
1.13
1.39
1.02
1.23
1.18
1.18
0.91
0.71
0.47
0.70
0.92
0.75
0.73
% K20
1.13
1.12
1.14
0.88
1.27
1.08
1.09
1.15
1.47
1.16
.31
.24
.21
.28
.51
.28
.06
1.20
0.83
1.24
1.00
1.31
1.37
0.96
0.49
0.36
-
0.48
0.56
1.00
1.18
1.20
1.15
1.19
0.94
1.28
1.01
1.16
0.90
0.84
0.74
0.56
0.36
0.48
0.54
0.39
0.48
-------
TABLE C.I (CONT'D)
110
Date
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
11 1/76
8/ 1/76
9/ 1/76
10/ 1/76
ll/ 1/76
12/ 1/76
13/ 1/76
14/ 1/76
J5/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21 / 1/76
% Cl
0.029
0.029
0.029
0.033
0.030
0.035
0.031
0.029
0.030
0.028
0.028
0.030
0.030
0.029
0.030
0.029
0.030
0.037
0.028
0.028
0.029
0.029
% S03 2
.1*0 1
.48 1
.49 1
.12 c
.68 1
.14 C
0.70 (
1.14 1
0.88 (
0.46 C
0.91
0.92
0.94
1.13
.14
.12
• 38
.36
.20
.06
1.42
0.94
; K2o
.20
.34
.37
).77
.42
).78
K58
.26
).83
).46
.05
.12
.15
-35
.28
.22
.45
.42
.27
.15
.46
.12
-------
11
TABLE C.2 RESULTS FROM ANALYSES OF SLURRY FEED SAMPLES
Date
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
14/10/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/H/75
A/11/75
5/H/75
6/11/75
7/11/75
8/11/75
9/H/75
2/12/75
3/12/75
VI 2/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
% Cl
0.080
0.056
0.052
0.058
0.060
0.068
0.050
0.044
0.046
0.043
0.051
0.055
0.046
0.043
0.038
0.040
0.040
0.050
0.043
0.055
0.042
0.047
0.055
-
0.050
0.047
0.043
0.043
0.044
0.042
0.042
0.044
0.042
0.050
0.046
0.042
0.041
0.040
0.040
0.052
0.041
0.040
% S03
(Total b as)
0.48
0.48
0.45
0.46
0.46
0.45
0.50
0.44
0.44
0.43
0.43
0.43
0.44
0.42
0.41
0.40
0.40
0.40
0.41
0.42
0.43
0.41
0.43
NO SAMPLE
0.42
0.42
0.42
0.43
0.44
0.41
0.43
0.41
0.40
.0.43
0.48
0.48
0.47
0.52
0.51
0.53
0.51
0.51
% K20
0.92
0.91
0.91
0.91
0.92
0-92
0.93
0.92
0.93
0.93
0.93
0.93
0.93
0.93
0.92
0.93
0.93
0.93
0.92
0.93
0.91
0.92
0.93
-
0.91
0.91
0.92
0.93
0.92
0.Q3
0.92
0.92
0.91
0.94
0.90
0.90
0.89
0.91
0.90
0.91
0.91
0.89
-------
TABLE C.2 (CONT'D)
Date
10/12/75
11/12/75
12/12/75
13/12/75
14/12/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
11 1/76
8/ 1/76
9/ 1/76
10/ 1/76
\\l 1/76
12/ 1/76
13/ 1/76
IV 1/76
15/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21/ 1/76
% Cl
0.041
0.039
0.050
0.042
0.048
0.059
0.042
0.045
0.043
0.046
0.038
0.038
0.048
0.039
0.048
0.048
0.048
0.040
0.041
0.044
0.042
0.048
0.041
0.054
0.039
0.040
0.041
% SO,
(Total § as)
0.54
0.52
0.55
0.58
0.54
0.48
0.46
0.46
0.46
0.46
0.46
0.46
0.47
0.46
0.45
0.46
0.44
0.44
0.44
0.46
0.47
0.50
0.47
0.46
0.44
0.43
0.43
% K20
0.91
0.91
0.91
0.90
0.91
0.90
0.91
0.92
0.91
0.92
0.91
0.91
0.91
0.89
0.92
0.91
0.31
0.91
0.92
0.91
0.91
0.91
0.90
0.91
0.91
0.91
0.92
-------
113
TABLE C.3- RESULTS FROM ANALYSES OF DISCARD DUST
Date
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
14/10/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
4/11/75
5/11/75
6/11/75
7/11/75
8/11/75
9/11/75
2/12/75
3/12/75
4/12/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
% CL
0.99
1.01
0.89
0.96
0.96
J.35
2.51
1.64
-
2.40
1.88
1.86
1.77
-
-
2.08
1.41
3.08
5.11
4.28
4.01
4.39
3.89
-
—
6.79
6.62
5.85
5.82
3.09
1.35
1.07
0.75
0.83
5.33
4.14
2.64
7.14
4.76
5.75
6.24
4.76
% S03
(Total S as)
7.09
7.22
7.28
7.48
7.39
7.59
10.34
6.65
NO DUST DISCARDED
7.35
6.00
6.02
6.12
NO SAMPLE
NO DUST DISCARDED
6.23
5.65
6.05
5.56
5.49
5.22
5.15
4.84
NO SAMPLE
NO SAMPLE
3.62
3.27
3.10
2.81
4.08
4.12
4.11
3.76
4.13
5.84
6.37
6.51
12.66.
6.28
6.41
6.97
5.41
% K20
7.37
7.45
7.34
7.66
7.48
8.26
11.27
7.65
-
9.16
7.32
7.22
7.13
-
-
7-83
6.70
8.60
10.07
9.28
8.63
8.87
8.21
-
™
9.62
9-38
8.57
8.34
7.46
5-83
5.47
4.67
5.42
9-36
8.78
7.81
16.28
9.10
9.88
10.70
8.24
-------
TABLE C.3 (CONT'D)
114
Date
10/12/75
11/12/75
12/12/75
13/12/75
14/12/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
If 1/76
8/ 1/76
9/ 1/76
10/ 1/76
ll/ 1/76
?2/ 1/76
13/ 1/76
14/ 1/76
15/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21/ 1/76
% Cl
5.35
5.52
5.53
-
5.68
-
1.98
-
-
A. 55
4.58
4.03
-
3.06
4.50
1.20
1.11
J.01
1.20
0.76
0.92
0.80
0.78
0.69
0.72
0.76
0.86
% so3
(Total S as)
5-51
4.90
5.62
NO SAMPLE
5.54
NO SAMPLE
6.19
NO SAMPLE
NO SAMPLE
4.62
4.62
4.70
NO SAMPLE
5-31
4.51
5-43
5.27
5.16
5.37
4.94
4.89
4.41
4.50
4.28
5.01
4.81
4.28
% K20
8.27
8.20
8.46
-
8.65
_
6.88
_
-
7-37
7.38
6.89
-
7.25
7.60
6.07
5-79
5.54
6.10
5.16
5.15
4.62
4.68
4.49
5-32
5.19
4.84
-------
15
TABLE C.4. RESULTS FROM ANALYSES OF RETURN DUST
Date
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
14/10/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
VII /75
5/H/75
6/11/75
7/H/75
8/11/75
9/11/75
10/11/75
2/12/75
3/12/75
VI 2/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
% Cl
0.80
0.73
0.54
0.73
1.59
1.36
1.29
2.46
2.29
2.15
2.33
1.81
1.86
1.76
2.56
1.53
1.52
3.82
3-41
3.87
3.48
3.02
3-51
-
4.22
4.66
6.37
3.96
4.00
1.27
1 .00
0.79
0.54
0.44
-
4.00
2.52
2.49
2.58
1.00
0.79
0.58
3.68
% so3
(Total S as)
6.25
6.28
5.36
7.26
7.05
7.79
6.21
8.16
6.91
6.80
7-20
5.61
6.51
5.86
6.54
6.00
5.37
5-36
4.87
4.84
4.48
4.75
4.24
NO SAMPLE
2.58
2.84
3.22
2.47
2.70
3.60
3.67
3.77
3.54
3.30
NO SAMPLE
5.65
5.97
6.51
6.96
3-69
3.81
3.57
4.68
% K20
6.22
6.25
5.15
7-54
8.20
8.90
7.02
9.83
8.54
8.32
8.73
6.71
7.73
7.08
8.44
6.92
6.60
8.63
7-75
8.21
7.47
6.91
7-23
""
6.49
7.07
9.05
6.28
6.69
5.20
i f* «
4.91
I t** f\
4.92
1 r™ /^
4.50
_ f\ f\.
3.89
10.70
9.63
8_ rt
.58
8«-* **
.33
i f^\
4.94
i _ _
4.93
I x i
4.61
6m
.59
-------
116
TABLE C.4 (CONT'D)
Date
10/12/75
11/12/75
12/12/75
13/12/75
14/12/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
11 1/76
8/ 1/76
9/ 1/76
10/ 1/76
11 / 1/76
12/ 1/76
13/ 1/76
14/ 1/76
15/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21 / 1/76
% Cl
2.51
2.36
3.69
4.04
3.02
1.56
1.78
1.77
2.42
3.55
2.37
2.56
1.35
1.76
2.30
1.06
0.93
0.80
0.79
0.78
0.78
0.65
0.58
0.69
0.58
0.72
0.87
% S03
(Total S as)
3.68
3.45
4.22
4.07
4.06
5.09
5.14
A. 70
4.62
4.48
3.36
3.43
3.23
4.91
3.71
4.74
5.05
5.14
4.82
4.85
4.41
4.27
4.11
4.19
4.56
4.47
4.24
% K20
4.60
4.22
5.94
6.05
5.16
5.60
5-99
5.65
5.82
6.33
4.29
4.64
3.58
5.56
4.91
5.05
5.46
5.37
5.10
4.93
4.59
4.38
4.10
4.22
4.70
4.62
4.83
-------
117
TABLE C.5. Btu AND CHLORINE CONTENT OF CHLORINATED HYDROCARBONS (SAMPLES
FROM LIME TO KILN)
Material
Aliphatic
Aromatic +
Comp 1 ex
Polychlor inated
Biphenyl
Date
24/10/75
27/10/75
28/10/75
3/11/75
A/11/75
13/11/75
18/11/75
Average
10/12/75
11/12/75
12/12/75
13/13/75
15/12/75
Average
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
7/ 1/76
8/ 1/76
9/ 1/76
Average
Btu/!b
13400
H390
10750
8900
8590
8340
8410
9970
9530
9500
8820
9320
9310
9300
11380
11590
11880
12170
12070
12050
12000
11880
Vi scosi ty
(Centipoise Specific
% Chlorine 3 23°C) Gravity
21 .52
32.76
39.40
42.23
43.52
42.38
43.38
37.88 - 1.13
40.56
45-91
44.16
41.80
40.48
42.58 40.8 1.27
36.16
37.75
34.90
33.90
33-19
34.93
33.90
34.97 17.0 1.18
-------
118
TABLE C.6. Btu, S AND Cl ANALYSES FROM NO. 6 FUEL OIL
Date
7/10/75
8/10/75
9/10/75
10/10/75
14/10/75
15/10/75
16/10/75
17/10/75
20/10/75
21/10/75
22/10/75
23/10/75
2 VI 0/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
3/11/75
4/11/75
5/11/75
6/11/75
7/11/75
10/11/75
11/11/75
12/11/75
2/12/75
3/12/75
VI 2/75
5/12/75
8/12/75
9/12/75
JO/12/75
11/12/75
12/12/75
15/12/75
30/12/75
2/ 1/76
5/ 1/76
6/ 1/76
7/ 1/76
8/ 1/76
9/ 1/76
12/ 1/76
13/ 1/76
Btu/lb
17809
17222
17978
17942
18106
18008
17920
17988
17924
17877
17984
18189
17701
17901
17862
17970
17917
17889
18135
18077
18047
18079
18036
17979
17782
17719
17700
17676
17714
17648
17679
17732
17695
17596
17655
17504
17678
17565
17498
17433
17630
17900
17719
17990
17896
% S
2.40
2.42
2.34
2.26
2.26
2.28
2.31
2.30
2.24
2.24
2.09
2.23
2.22
2.24
2.22
1.95
2.29
.77
.62
.57
.55
.51
.56
.26
.89
2.06
2.54
2.28
2.56
2.58
2.55
2.52
2.58
2.50
2.56
2.64
2.59
2.41
1.61
2.07
2.28
2.47
2.42
1.96
1.95
% Cl
Average
7/10/75 - 22/10/75
% Chlorine = 0.028%
Average
23/10/75-12/11/75
% Chlorine - 0.064
Average
2/12/75-15/12/75
% Chlorine - 0.047
Average
30/12/75-9/1/76
% Chlorine - 0.030
-------
119
TABLE C.6 (CONT'D)
Date
IV 1/76
15/ 1/76
16/ 1/76
197 1/76
20/ 1/76
21/ 1/76
22/ 1/76
Btu/lb
17935
17952
18025
17762
17846
18005
18160
% S
1.54
2.00
1.98
1.92
1.89
1.54
J.42
% Cl
Average
12/1/76-22/1/76
% Chlorine = 0.
038
-------
120
TABLE C.7. DAILY RECORD OF PRODUCTION AND MATERIALS CONSUMPTION
Date
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
14/10/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
4/11/75
5/11/75
6/11/75
7/H/75
8/11/75
9/11/75
2/12/75
3/12/75
A/12/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
SLURRY FEED
(tons/24 hr)
1653
1676
1681
1658
1585
1325
1632
1604
1639
1646
1661
1682
1669
1669
1661
1664
1681
1686
1669
1732
1656
1652
1621
1633
1647
1658
1673
1679
1679
1692
1693
1698
1670
1701
1622
1627
1644
1653
1670
1655
1666
1664
CHLORIN. HC
(gal/24 hr)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1170
1440
1440
1440
1440
1440
1440
2670
2880
2880
2880
2880
930
0
0
0
0
0
430
0
0
1640
2187
1351
1117
2121
CLINKER
(tons/24 hr)
1043
1041
1073
1027
1029
845
1036
1047
1064
1058
1065
1081
1079
1071
1078
1046
1082
1079
1067
1112
1061
1040
1021
1032
1027
1029
1057
1030
1010
1044
1054
1058
1030
1056
1038
1042
1039
1028
1056
1074
1043
984
DISCARD DUST
(tons/24 hr)
40.44
63.44
24.84
65-66
0
18.94
6.96
23.00
0
12.90
17.86
15.32
4.93
15-97
0
45-36
11.96
19.72
20.96
16.42
18.32
42.58
41.36
37.17
56.52
63-32
38.92
80.66
107.72
72.82
60.46
59-90
73.50
64.26
20.58
18.30
37.72
61.40
38.26
0
51.36
129-78
-------
121
TABLE C.7- (CONT'D)
Date
10/12/75
11/12/75
12/12/75
13/12/75
1V12/75
15/12/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
7/ 1/76
8/ 1/76
9/ 1/76
10/ 1/76
11/ 1/76
12/ 1/76
13/ 1/76
IV 1/76
15/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21/ 1/76
SLURRY FEED
(tons/24 hr)
1676
1669
1669
1661
1653
1658
1606
1595
1608
1630
1629
1627
1616
1612
1615
1619
1621
1627
1636
1639
1618
1633
1596
1606
1635
1608
1602
1629
CHLORIN. HC
(gal/24 hr)
1997
1308
1821
1691
1869
1909
0
0
0
2029
2210
2168
1985
604
1230
1367
0
0
0
0
0
0
0
0
0
0
0
0
CLINKER
(tons/24 hr)
994
999
1024
985
956
980
1024
1035
1039
1017
1005
970
1034
944
1045
1027
1020
1011
1032
1040
1028
1046
1026
979
1026
1020
991
1016
DISCARD DUST
(tons/24 hr)
127.66
113.76
79.12
125.96
157-68
130.30
24.72
0
6.11
55.43
70.90
116.98
19.86
137.39
4.54
32.10
43-92
61.20
40.52
32.36
42.80
19-10
13.38
85.76
47.62
32.22
66.56
55.75
-------
TABLE C.8. MATERIAL BALANCE FOR CHLORINE
DATE 1
7/10/75 ;
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
1 VI 0/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
4/11/75
5/11/75
6/11/75
CHLORINE
SLURRY
•EED
2,645
,877
,748
,923
,902
,801
,632
,412
,508
,416
,694
,850
,535
,435
,262
,331
,345
,686
,435
,905
,391
,553
,783
,502*
,647
,558
,439
,444
,478
,42)
1,422
INPUT
16
OIL
81
81
82
81
78
67
81
80
81
81
81
81
81
81
80
81
185
183
183
190
182
182
183
181
186
179
179
178
179
184
182
(lb/24 hr)
CHLORINATED
HYDROCARBON TOTAL
0 2,726
0 ,958
0 ,830
0 2,004
0 ,980
0 ,868
0 ,713
0 ,492
0 ,589
0 ,497
0 ,775
0 ,931
0 ,616
0 ,516
0 ,342
0 . ,412
2,845 4,375
3,502 5,371
3,959 5,577
4,416 6,511
4,873 6,446
6,411 8,146
6,642 8,608
12,316 13,999
13,284 15,117
13,284 15,021
13,284 14,902
13,284 14,906
13,743 15,400
4,574 6,179
0 1,604
CHLORINE
CLINKER
688
666
730
836
658
490
622
649
660
656
746
627
690
900
647
607
649
647
619
689
658
707
653
805
698
803
676
659
566
647
653
RETAINED
DISCARD
DUST
801
1,281
442
1,261
0
511
349
754
0
619
672
570
174
575
0
1,887
337
1,215
2,142
1,406
1,469
3,738
3,218
3,970*
6,036*
8,599
5,153
9,437
12,539
4,500
1,632
(lb/24 hr)
TOTAL
1,489
1,947
1,172
2,097
658
1,001
971
1,403
660
1,275
1,418
1,197
864
1,475
647
2,494
986
1,862
2,761
2,095
2,127
4,445
3,871
4,775
6,734
9,402
5,829
10,096
13,105
5,147
2,285
TOTAL %
RETAINED
54.6
99-4
64.0
104.6
33-2
53-6
56.7
94.0
41.5
85-2
79-9
62.0
53.5
97.3
48.2
176.6
22.5
34.7
49-5
32.2
33.0
54.6
45.0
34.1
44.5
62.6
39-1
67.7
85.1
83.3
142.4
* Calculated from average data.
-------
TABLE 'C. 0 (CONT* 0)
<
DATE
7/H/75
8/11/75
9/H/75
2/12/75
3/12/75
VI 2/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
10/12/75
11/12/75
12/12/75
13/12/75
14/12/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
V 1/76
5/ 1/76
6/ 1/76
11 1/76
8/ 1/76
9/ 1/76
107 1/76
ll/ 1/76
12/ 1/76
13/ 1/76
CHLOR
SLURRY
rEED
,494
,403
,701
,492
,367
,348
,322
,336
,721
,366
,331
,374
,302
,669
,395
,587
,895
,340
,447
,402
,499
,236
,228
,548
,260
,554
,556
,562
,309
,344
INE INPUT
#6
OIL
183
181
182
124
133
133
132
131
131
142
132
130
134
134
133
131
85
87
87
83
82
83
83
85
83
81
109
109
109
108
(lb/24 hr)
CHLORINATED
HYDROCARBON
0
0
0
2,325
0
0
8,868
11,826
7,306
6,040
11,470
10,287
7,626
10,213
8,977
10,107
0
0
0
8,657
9,844
8,928
7,950
2,366
5,070
5,468
0
0
0
0
TOTAL
1,677
1,584
1,883
3,941
1,500
1,48?
10,322
13,293
9,158
7,548
12,933
11,791
9,062
12,016
10,505
11,825
1,980
1,427
1,534
10,142
11,425
10,247
9,261
3,999
6,413
7,103
1,665
1,671
1,418
1,452
CHLOR IN
CLINKER
783
597
739
644
604
708
678
612
601
793
787
676
699
655
630
593
594
600
603
671
603
679
641
548
627
575
571
607
619
603
IE RETAINED (l
DISCARD
DUST
1,282
1,102
1,067
2,194
1,515
1,992
8,768
3,642
0
6,410
12,355
13,660
12,559
8,751
14,208
17,906
979
0
242
4,434
6,452
10,715
1,601
10,991
278
2,889
1,054
1,359
818
111
b/24 hr)
TOTAL
2,065
1,699
1,806
2,838
2,119
2,700
9,446
4,254
601
7,203
13,142
14,336
13,258
9,406
14,838
18,499
1,573
600
845
5,105
7,055
11,394
2,242
11,539
905
3,464
1,625
1,966
1,437
1,380
TOTAL %
RETAINED
123-1
107-3
95.9
72.0
141.3
182.3
91.5
32.0
6.6
95.4
101.6
121.6
146.3
78.3
141.2
156.4
79.4
42.0
55-1
50.3
61.8
111.2
24.2
288.5
14.1
48.8
97.6
117.6
101.3
95.0
-------
TABLE C.8 (CONT'D)
CHLORINE INPUT (lb/24 hr)
DATE
I A/
15/
16/
17/
18/
19/
20/
21/
SLURRY
FEED
1/76
1/76
1/76
1/76
1/76
1/76
1/76
1/76
,424
,372
,532
,317
,776
,254
,282
,336
#6
OIL
108
108
107
108
108
107
107
109
CHLORINATED
HYDROCARBON
0
0
0
0
0
0
0
0
TOTAL
1,532
1,480
1,639
1,425
1,874
1.361
1,389
1,445
CHLORINE
CLINKER
617
607
616
724
575
571
575
589
RETAINED (lb/24 hr)
DISCARD
DUST
650
351
214
1,338
657
464
1,012
959
TOTAL
1,267
958
830
2,062
1,232
1,035
1,587
1,548
TOTAL %
RETAINED
82.7
64.7
50.6
144.7
65.7
76.0
114.2
107.1
-------
125
TABLE C.9. MATERIAL BALANCE FOR K 0
K20 INPUT (tons/24 hr) K?0 RETAINED (tons/24 hr)
DATE
7/10/75
8/10/75
9/10/75
10/10/75
11/10/75
12/10/75
13/10/75
14/10/75
15/10/75
16/10/75
17/10/75
18/10/75
19/10/75
20/10/75
21/10/75
22/10/75
23/10/75
24/10/75
25/10/75
26/10/75
27/10/75
28/10/75
29/10/75
30/10/75
31/10/75
1/11/75
2/11/75
3/11/75
4/11/75
5/H/75
6/11/75
7/H/75
8/11/75
9/H/75
2/12/75
3/12/75
4/12/75
5/12/75
6/12/75
7/12/75
8/12/75
9/12/75
10/12/75
11/12/75
SLURRY FEED
15-21
15-25
15-30
15-09
14.58
12.18
15.18
14.76
15.24
15.31
15.45
15.64
15.52
15.52
15.28
15.48
15.63
15.68
15.35
16.11
15.07
15.20
15.08
1 5 . 02*
14.99
15.09
15.39
15.61
15.45
15.74
15.58
15.62
15.20
15.99
14.60
14.64
14.63
15.04
15.03
15-06
15.16
14.81
15.25
15.19
CLINKER
11.78
11.66
12.23
9-04
13.07
9.13
11.29
12.04
15.64
12.27
13-95
13.40
13-06
13.71
16.28
13-39
11.47
12.95
8.86
3-79
10.61
13.62
13-99
9-91
5.03
3-70
4.44*
4.94
5.66
10.44
12.44
12.70
11.84
12.57
9-76
13-34
10.49
11.92
9-50
9.02
7-72
5.51
3.58
4.80
DUST
2.98
4.73
1.82
5.03
0
1.56
0.78
1.76
0
1.18
1.31
1.11
0.35
1.27*
0
3.55
0.80
1.70
2.11
1.52
1.58
3.78
3.40
2.91*
4.42*
6.09
3.65
6.91
8.88
5.43
3.52
3.28
3.43
3.48
1.93
1.61
2.94
10.00
3.48
0
5.50
0.69
10.56
9.33
TOTAL
14.76
16.39
14.05
14.07
13-07
0.69
2.07
13.80
15.64
13.45
15.26
14.51
13-41
14.98
16.28
16.94
12.27
14.65
10.97
15-31
12.19
17-40
17-39
12.82
9.45
9.79
8.09
11.85
14.54
15.87
15.96
15.98
15.27
16.05
11.69
14.95
13.43
21.92
12.98
9.02
3.22
16.20
14.14
14.13
TOTAL %
RETAINED
97.0
107.5
91.8
93.2
89-6
87.8
79.5
93.5
102.6
87-8
98.8
92.8
86.4
96.5
106.5
109.4
78.5
93-4
71.5
95.0
80.9
114.5
115.3
85.4
63-0
64.9
52.6
75-9
94.1
100.8
102.4
102.3
100.5
100.4
80.1
102.1
91.8
145-7
86.4
59.9
87.2
109.4
92.7
93.0
* Calculated from average data.
-------
126
TABLE C.9 (CONT'D)
KjO INPUT (tons/24 hr) K?0 RETAINED (tons/24 hr)
DATE
12/12/75
13/12/75
1 VI 2/75
31/12/75
I/ 1/76
2/ 1/76
3/ 1/76
4/ 1/76
5/ 1/76
6/ 1/76
7/ 1/76
8/ 1/76
9/ 1/76
10/ 1/76
ll/ 1/76
12/ 1/76
13/ 1/76
IV 1/76
15/ 1/76
16/ 1/76
17/ 1/76
18/ 1/76
19/ 1/76
20/ 1/76
21 / 1/76
SLURRY FEED
15.19
14.95
15-04
14.45
14.51
14.79
14.83
14.99
14.30
14.70
14.67
14.37
14.89
14.75
14.80
14.89
15.08
14.72
14.86
14.52
14.45
14.88
14.63
14.58
14.99
CLINKER
5.53
3-94
4.59
12.29
13.87
14.23
7.83
14.27
7.57
6.00
11.89
8.67
4.72
10.71
11.32
11.87
14.04
13.16
12.76
14.88
13-90
13.03
11.73
14.47
11.38
DUST
6.69
10.71*
13-64
1.70*
0
0.42*
4.01
5.22
8.63
1.37
0.93*
0.33
2.44
2.66
3.54
2.24
1.97
2.21
0.98
0.62
4.01
2.14
1.71
3.45
2.70
TOTAL
12.22
14.55
18.23
13-99
13.87
14.65
11.84
19-49
16.20
7.37
21.82
9.00
7.16
13.37
14.87
14.11
16.01
15.37
13.74
15-50
17.91
15.17
13.44
17.92
14.08
TOTAL %
RETAINED
80.4
97.3
121.2
96.8
95-6
99-0
79.8
130.0
109.4
50.1
148.7
62.6
48.1
90.6
100.4
94.8
106.2
104.4
92.5
106.7
123-9
101.9
91.9
122.9
93-9
* Calculated from average data.
-------
APPENDIX D
QUALITY OF CEMENT PRODUCED
-------
129
APPENDIX D
QUALITY OF CEMENT PRODUCED
While remote, the possibility of non-volatile chlorinated
hydrocarbons not being destroyed and remaining with the clinker was
considered. It is common knowledge that excessive quantities of organic
materials can detrimentally affect setting, air entraining and compressive
strength characteristics of cement.
After preparing the daily composite clinker sample for analysis,
the quantity of sample which had not undergone size reduction was
retained. These were combined into a composite of several days, three
such composites being prepared for each of the aromatic plus complex
chlorinated hydrocarbon burn, the PCB burn and the final baseline burn.
Each composite, after being crushed to approximately one-half inch, was
blended with the correct amount of the normal production gypsum used at
St. Lawrence Cement. The clinker-gypsum blends were ground in a laboratory
ball mill. The grinding was timed and samples were taken at suitable
intervals during grinding to determine surface area.
The cements were produced to conform to the requirements of
Symbol 10 cement a* designated in the Canadian Standards Association
(CSA) Standard A5 "Portland Cements", and Type I cement as defined in
the American Society for Testing and Materials (ASTM) designation C150,
"Standard Specification for Portland Cement".
Each cement was tested in conformance to the following ASTM
methods:
C185, Test for Air Content of Hydraulic Cement Mortar;
C109, Test for Compressive Strength of Hydraulic Cement Mortars
(using 2-in. cube specimens);
C359, Test for False Set of Portland Cement (Mortar Method);
C204, Test for Fineness of Portland Cement by Air Permeability
Apparatus;
C151, Test for Autoclave Expansion of Portland Cement;
C191, Test for Time of Setting of Hydraulic Cement by Vicat
Needle.
-------
130
To ensure that differences In quality were not due to unexpected
differences in cement composition, a chemical analysis was performed on
each cement tested. Procedures followed were those normally used for
control purposes at St. Lawrence Cement.
The technique used for determination of the concentration of
each oxide was fusion with lithium tetraborate followed by x-ray fluorescence
analysis. This method is commonly used for cement analyses. With the
exception of lower K~0 content in the cements from clinkers produced while
burning chlorinated hydrocarbons, there were no significant differences
in their chemical compositions (Tables D.I, 0.2 and D.3).
Differences in physical characteristics found by testing these
cements were:
Setting Time - The water required to obtain normal consistency
was greater and setting time shorter for baseline clinker cements.
Shorter time of set and higher water requirement were attributed
to the higher alkali content of the baseline clinker.
False Set - False set was severe on the cements from clinker
produced during baseline burning. There was no indication of
false set in cements from clinker produced while burning chlorinated
hydrocarbons. The improvement was due to lower alkali content of
the cement. In this regard, the determination of false set by
ASTM Method C*»51, 'Test for False Set of Portland Cement
(Paste Method)" would not have shown as dramatic an improvement.
This latter test takes into account the water requirement of
cements, and therefore compensates for the greater quantity of
water required by the higher alkali cements.
Compress Ive Strength - Higher initial and lower ultimate
strengths are characteristics of higher alkali content cements.
The only apparent differences in cements from clinkers produced
with and without burning of chlorinated hydrocarbons were those
expected because of the differences in alkali content.
-------
131
TABLE D.I. CEMENTS FROM CLINKER PRODUCED DURING BASELINE BURN
Sample Number
Chemical Tests
Loss on Igni t ion (%)
Si02 (*)
Al^O- "
Fe20o "
CaO "
MgO "
$03
K20 "
Free CaO (%}
C3S (%)
C2S "
C3A "
Ci,AF "
Physical Tests
Fineness ?
Blaine (cm /g)
Passing 200 mesh (%)
Setting Time
N.C. Penetration (mm)
N.C. Water (%)
Initial (minutes)
Final (minutes)
False Set Q
Temperature F
Penetration (mm)
3 minutes
S minutes
8 minutes
11 minutes
Remix
Compress ive Strength
1 day (psi)
3 day (psi)
7 day (psi)
28 day (psi)
Air Content
Water (%)
Flow (%)
Air (%)
Autoclave Expansion (%)
7
1.52
20.69
5.93
2.20
63.42
2.52
2.74
1.15
0.44
50.1
21.6
12.0
6.7
3406
96.4
10.0
22.5
115
225
70
50
50
21
16
50
2050
2910
3480
4130
73-0
80.6
7.2
0.07
8
1.51
20.59
5.90
2.22
63.39
2.52
2.79
1.23
0.52
50.8
20.8
11.9
6.7
3579
97.6
9.5
23.0
106
216
73
37
7
4
4
50
2010
2990
3610
73.0
86.8
7.6
0.09
9
1.43
20.53
6.00
2.24
63.35
2.52
2.89
1.22
0.54
50.0
21 .2
12.1
6.8
3607
98.8
9.0
23.0
106
221
73
34
8
4
3
50
1910
3210
3880
4530
73.0
86.0
8.3
0.05
Aver .
1.49
20.60
5.94
2.22
63-39
2.52
2.81
1.20
0.50
50.3
21.2
12.0
6.7
3530
97-6
9-5
22.8
109
221
72
40
22
10
8
50
1990
3040
3660
4380
73-0
84.5
7-7
0.07
-------
132
TABLE D.2. CEMENTS FROM CLINKER PRODUCED DURING AROMATIC PLUS
COMPLEX CHLORINATED HYDROCARBON BURN
Sample Number
Chemical Tests
Loss on Igni tion (%)
Si02 (%)
A1203 (%)
CaO \%)
MgO (*)
so3 U)
K20 (%)
Free CaO (%)
C3S (%)
C2S (%)
C3A (fc)
C^AF (%)
Physical Tests
Fineness -
T
Blaine (cm /g)
Passing 200 mesh (%)
Setting Time
N.C. Penetration (mm)
N.C. Water (%)
Initial (minutes)
Final (minutes)
False Set
o
Temperature F
Penetration (mm)
3 minutes
5 minutes
8 minutes
11 minutes
Remix
Compress ive Strength
1 day (psi)
3 day (psi)
7 day (psi)
28 day (psi)
Air Content
Water (%)
Flow (%)
Air (%)
Autoclave Expansion (%)
4
1.63
20.37
5.85
2.26
63.19
2.51
2.81
0.66
0.50
51.8
19.4
11.7
6.9
3626
98.0
10.0
22.0
118
238
70
50
50
50
50
50
1900
3520
4590
5900
71.0
80.0
9-2
0.03
5
1.86
20.51
5.79
2.23
62.99
2.49
2.69
0.53
0.59
50.8
20.6
11.6
6.8
3562
97.4
10.0
22.0
135
250
72
50
50
50
50
50
1640
3200
4360
6050
72.0
90.5
9-0
0.04
6
1.63
20.83
5-89
2.22
63.56
2.52
2.89
0.69
0.48
49.4
22.5
11.9
6.7
3561
97-2
9-5
22.0
135
250
73
50
50
50
50
50
1890
3180
4160
5460
72.0
81.3
6.7
0.02
Aver.
1.71
20.57
5.84
2.24
63-25
2.51
2.80
0.63
0.52
50.7
20.8
11.7
6.8
3583
97.5
9.8
22.0
129
246
72
50
50
50
50
50
1810
3300
4370
5800
71.6
83-9
8.3
0.03
-------
133
TABLE D.3. CEMENTS FROM CLINKER PRODUCED DURING POLYCHLORINATED
BIPHENYL BURN
Sample Number
Chemical Tests
Loss on Ignition (%)
Si02 (fc)
A1203 (%)
Fe20o (%)
CaO 1%)
MgO (%)
S03 U)
K20 (%)
Free CaO (%)
c3s (%)
C2S (*)
C3A ($)
^ i A CT i Of 1
V* /i** * \ ^ /
Physical Tests
Fineness 2
Blaine (cm /g)
Passing 200 mesh (%)
Setting Time
N.C. Penetration (mm)
N.C. Water (%)
Initial (minutes)
Final (minutes)
False Set
Temperature F
Penetration (mm)
3 minutes
5 minutes
8 minutes
1 1 minutes
Remix
Compress ive Strength
1 day (psi)
3 day (psi)
7 day (psi)
28 day (psi)
Air Content
Water (%)
Flow (%)
Air (*)
Autoclave Expansion (%)
1
1.46
20.51
5.91
2.16
63.24
2.53
2.94
1.01
0.53
50.4
20.9
12.0
6.6
3388
96.0
10.0
22.0
153
273
70
50
50
50
50
50
1980
3080
3770
4620
71.0
88.0
7-7
0.05
2
1.58
20.72
5.86
2.18
63.27
2.56
2.64
0.74
0.57
50.0
21.8
11.9
6.6
3579
97.0
11.0
22.0
125
250
71
50
50
50
50
50
1800
3000
4000
5HO
71.0
84.5
7.9
0.06
3
1.48
20.87
5.89
2.16
63.65
2.58
2.61
0.96
0.69
50.3
22.0
12.0
6.6
3593
96.8
9-0
22.0
112
235
71
50
50
50
50
50
1990
3260
4090
5080
71.0
82.2
9.2
0.08
Aver.
1-51
20.70
5.89
2.17
63.39
2.56
2.73
0.90
0.60
50.2
21.6
12.0
6.6
3520
96.6
10.0
22.0
130
253
71
50
50
50
50
50
1920
3110
3950
4940
71.0
84.9
8.3
0.06
-------
APPENDIX E
EQUIPMENT DESCRIPTION AND ASSOCIATED ECONOMICS
-------
137
APPENDIX E
EQUIPMENT DESCRIPTION AND ASSOCIATED ECONOMICS
In order to burn waste chlorinated hydrocarbons in a cement kiln,
they must be blended to be compatible with the handling and storage system
and to achieve uniform feed to the kiln. A storage tank and feed system
with associated control devices must be installed at the kiln site.
For the program to be economically viable, the price structure
must be favourable to the originator of the wastes and to the cement
producer. The waste disposal company must be able to generate sufficient
revenue to cover shipping, handling, blending, disposal of non-fuel
fractions and receive some profit. The economics from the point of view
of the cement producer are discussed in Section E.2.
E. 1 Equipment Description
The system installed at the cement plant can vary. A simple
inflexible installation with manual controls and a used rail car for
storage would cost approximately $25,000. At the other extreme, a
sophisticated system with large storage capacity, corrosion resistant
construction, well instrumented could cost $200,000.
Although there are general similarities between cement kilns,
the flow rate of waste chlorinated hydrocarbon depends on the percentage
of chlorine in the waste, the decrease desired In the potassium and
sodium levels of the finished product and the production capacity of
the kiln. Thus, for the optimum system, the design, engineering,
installation and operation must be compatible with these parameters.
Two systems are described herein; one is the actual system
installed at the St. Lawrence Cement Co. kiln and the other has been
somewhat arbitrarily selected as a basic .system. This basic system
assumes that a quality controlled blend of chemical wastes will be used.
The fuel product would be a tow viscosity, single-phase liquid, non-
corrosive or mildly corrosive to carbon steel, not highly volatile,
having an approximate composition of: 60,000 Btu/gallon, kQ% Cl ,
and specific gravity of 1.2. Fuels with higher or lower Btu and
chlorine can be prepared for those applications where other compositions
are optimum.
-------
138
At these compositional values, a feed rate of 1 Igpm into a
1000 ton per day kiln would provide an alkali reduction of 0.^5% as K 0.
The Btu's provided would be a small percentage of the total heat requirement
of the kiln. With waste streams with a lower chlorine level, larger volumes
can be used to make up a greater fraction of the heat requirements, in
practice up to 10-15% of total Btu input.
Other than capacity and the provision that the kiln be rotary,
it makes little difference what the primary fuel or fuels are. Wet or
dry fuels can burn waste chlorinated hydrocarbons, with consideration given
to the increase in dust loading to the dust handling equipment.
E.I.I Cost of basic system
The basic system with minimum component cost estimates would
entai1:
Equipment Cost:
Tank, 10,000 gal carbon steel $ 8,000
Feed Pump, Centrifugal, 7.5 hp 1,200
Flame Arrester 100
Vent Scrubber, Activated Carbon 1,000
Tank Level Indicator 2,500
Grounding 400
Tank Berming 1,000
Unloading Piping 500
Piping, Tank to Kiln, 100' 1,000
Electrical-Combination Starter 1,500
Concrete Pad, Site Preparation 2,500
Miscellaneous, Hardware
Painting and NFPA Code Marking 2,500
Instrumentation AGO
Nozzle Assembly 200
$22,800
Installation: ^,000
Total System Cost, Installed $26.800
-------
139
A schematic representation of the system is shown in Figure E.I.
Operating Cost:
Maintenance, b% of investment/year $1,080
Operating Labour, Based on 330 days/year
Operating 2k hr/day
1 hr per shift @ $8.00 8,000
Electricity, 7-5 hp, 8000 hr
@ 3
-------
O
6-
IL
1 CHEMTROL STORAGE TANK
2 PUMP
3 TO KILN
A RETURN LINE
5 VENT PIPE
6 LEVEL GAUGE
7 FILLER PIPE
FIGURE E.1
SCHEMATIC DIAGRAM OF BASIC CHLORINATED
HYDROCARBON FEED SYSTEM
-------
1 STORAGE TANK
2 KILN FEED PUMPS
3 FILTERS
4 FLOW METER
5 CONTROL VALVE
6 TO KILN III
7 FLOW RECORDING CONTROLLER
8 TO KILN II
9 TO KILN I
10 FILTER
11 TO ODOR CONTROL SYSTEM
12 BACK TRAP
13 CAUSTIC SCRUBBER
U CHARCOAL FILTER
15 TRUCK
F.GURE £.2 SCHEMATIC DIAGRAM OF CHLORINATED HYDROCARBON
FACILITIES
-------
The total installed cost of this system was over $200,000,
comprising the following items:
Equipment and Installation:
Tank ' $25,000
Acid Brick Tank Lining 35,000
Teflon* Lined Pumps and Filters 12,000
Teflon* Lined Piping 45,000
Odour Control Equipment 5,000
Foundations 10,000
Dike 1,000
Pump House 10,000
Site Clearing and Leveling 3,000
Instrumentation 20,000
Painting 5,000
Sales Tax 5,000
Engineering 18,000
Consultant 8,000
Travel Expense 10.000
TOTAL $212,000
Operating Cost:
Maintenance and Operating Labour
(cost per year) $25,^50
Electricity (23
-------
require a reduction in the alkali oxide' (K 0 + Na_0) content of the
finished product. In this case, chlorine must be added to the process
materials, and is available from different sources in different geographical
areas. Waste hydrochloric acid from the steel industry may be used where
available; in other areas cement plants purchase calcium chloride. If
reduction of alkali is not required, then addition of chloride would
cause the generation of greater quantities of dust, thereby increasing
dust disposal costs.
Another factor making economic comparisons difficult is the
different waste streams available. In some cases, the chlorinated waste
streams may be suitable for purifying and further use. Other waste
streams, such as PCB's and insecticides banned for environmental reasons,
present a serious disposal problem. Obviously, these different materials
would not all cost the same for use in a cement kiln. However, to present
some models, the price structure used by Chemtrol Pollution Services, Inc.
for blended noncorrosive wastes suitable for use in the basic system
described above has been adopted. At present, Chemtrol plans to charge
80 percent of the fuel value and, where applicable, 30 percent of
the chlorine value for these blended "fuels". In the examples, the
following assumptions have been made: The kiln is a wet process type
producing 1,000 ton? per day with a fuel requirement equivalent to
5,150,000 Btu per ton of production.
E.2.1 Process requiring addition of chlorine
Assume in this case that the fuel cost Is $1 per million Btu and
that it is necessary to reduce the ICO content of the clinker by 0.*5 percent
This would require approximately 6,800 pounds of chlorine per day. A 32
percent solution of calcium chloride weighing 13 pounds and costing $0.1*
per Imperial gallon ($0.053 per pound of chlorine) is used. In this case,
2,565 gallons would be required at a total cost of $359-
This could be replaced by "Trol Fuel*" containing a nominal 70
percent chlorine and 65,000 Btu per Imperial gallon. At a weight of 13
pounds per gallon, 750 gallons would be required each day. Using the
trademark, Chemtrol Pollution Services Inc.
-------
above mentioned price structure, the cost per gallon would be:
'•"""a'TET^ "if -$0-'*5
Btu * B.u
TOTAL $0.20
Thus, the 750 gallons would cost $150 and would supply 48.8
mi i I ion Btu.
The total daily tangible savings then become:
Cost of Calcium Chloride $359
Cost of Normal Fuel Replaced $ kS
Less Cost of "Trol Fuel" $150
Net Savings Per Day $258
E.2.2 Process not requiring addition of chlorine
The 1,000 ton per kiln and a specific heat consumption of
5,150,000 Btu per ton of clinker produced and a fuel cost equivalent
to $1.00 per million Btu is again taken as the model. In this case,
however, no chlorine is required and the waste stream contains less than
5 percent chlorine. This material is available at $0.80 per million Btu
and Is used at a replacement rate of 10 percent of the Btu requirement
or 515.000 Btu per ton of clinker.
The fuel normally used on the kiln but replaced by this waste
stream would cost $0.5)5 per ton of clinker or $515 per day. The daily
savings can be calculated as:
Cost of Normal Fuel Replaced $515
Cost of 'Trol Fuel" 412
Net Daily Savings $103
This saving can be negated by the cost of discarding the
additional dust collected in the precipi tators (See Section 6.1), a factor
which will vary with each cement plant.
E.2.3 Total economic considerations
The above models are not intended to be all inclusive. Due
to the many variables involved, each possible application must be assessed
based upon economic factors existing in the locale. At the time the system
-------
145
was installed at the St. Lawrence Cement Co., it was necessary to use
chlorine in two wet kilns each having a capacity of 1,000 tons per year.
For the period January through September 1976, maintenance on the system
amounted to $18,760. Thus, the figure of $27,290 for yearly operating
and maintenance cost is considered accurate. Under these circumstances,
the economic picture over the period of a year becomes:
Total Savings Using "Trol Fuel"
(600 kiln operating days §
$258 per day) $154,800
Maintenance and Operating Cost 27,290
Interest on $212,000 @ 1 U/year 23,320
Net $104,190
This estimated saving, which would have given excellent payback
on the system, has not been realized. Low alkali requirements have been
falling into disfavour because such requirements are not compatible with
the goal of energy conservation. With the requirement for chlorine
addition, savings must be based only on fuel costs which, if all factors
are favourable, would total for the two kilns:
Total Savings Using "Trol Fuel"
(600 kiln operating days @
$103/day) $61,800
Maintenance and Operating Costs 27,290
Interest on $212,000 @ 1 U/yr 23,320
Net $'1,190
Even this modest sum has not been realized, for two reasons:
the cost of discarding additional dust, and the unsteady supply of these
wastes.
It is obvious from this discussion that plants requiring the
addition of chlorine are in a better position to make burning of
chlorinated wastes attractive than those plants not requiring chlorine.
Also, a less sophisticated system than that installed at St. Lawrence
Cement Co. is necessary if the only consideration is the fuel value of
these wastes.
-------
APPENDIX F
ONTARIO MINISTRY OF THE ENVIRONMENT EMISSION
GUIDELINES AND ANALYTICAL SUPPORT
-------
APPENDIX F
ONTARIO MINISTRY OF THE ENVIRONMENT EMISSION
GUIDELINES AND ANALYTICAL SUPPORT
F.1 Emission Guidelines
All industrial sources of potentially significant emissions to
the atmosphere in Ontario come under the jurisdiction of the Province's
Environmental Protection Act. The Act requires such sources to operate
under a Certificate of Approval which specifies the acceptable operating
conditions and emission rates for potential air contaminants. The
Ministry of the Environment required that the experimental waste chlorinated
hydrocarbon burn at St. Lawrence Cement Co., Mississauga meet certain
conditions specified in a Certificate of Approval. Conditional approval
was granted for each of the three stages of the waste burn (A, B, C) under
the terms of the tentative guidelines for the allowable emission rates
of organic chloride.
Pasqui11-Gifford dispersion calculations based on the guideline
emission rates from the wet kiln were used to predict the worst-case ground
level concentration of residual organic chloride (i-hour average).
Table F.I shows the specifications which were applied to each
stage of the waste burn, along with the measured emission rates.
The less stringent guideline for the first two waste burns was
met easily. Because of the large percentage of highly persistent poly-
chlorinated biphenyls anticipated in the fuel for the third waste burn,
a much more stringent guideline was given.
In the most critical case, the third burn, the calculated
approximate flue gas concentration of residual chloride, 50 ug/m (as
Cl), based on the stated guideline, corresponds to roughly 5 ppb (v/v)
of Aroclor 12^2 or 17 ppb (v/v) of dichloromethane (at ambient temperature),
the predominant individually identified component of the flue gas^
These values compare with the detection limit for PCB's of 3 vg/m or 0.3
ppb in the flue gas (ambient temperature).
Since no PCB was found in the flue gas samples and the methy-
lene chloride concentration only slightly exceeded the very stringent
guideline for organic chloride in this stage, the guideline should be
considered to have been met without question.
-------
150
TABLE F.I. MOE SPECIFICATIONS APPLIED TO WASTE BURNS
Waste
Burn
A
B
C
Column
1
0.5
0.5
0.005
2
99.
99.
99-
5
5
995
3
99
99
99
.990
.989
.986
k
5000
5000
50
5*
23-150
16-
88-
150
150
6
0.2
0.2
0.002
7
0.001-0.
0.001-0.
0.004-0.
006
006
006
Identification:
1. Guideline emission rate for residual organic chloride (g Cl/sec).
2. Estimated required percent destruction or retention of chlorinated fuel.
3. Estimated minimum percent combustion of chlorinated fuel from test data
(Table 6, Section 5-3).
*4. Approximate calculated flue gas residual chloride concentration based
on guideline emission rate (yg Cl/m ).
5. Approximate measured flue gas organic chloride concentration based on
dichloromethane and data from Table 5, Section 5-3 (yg Cl/m ). The
upper limit of the range corresponds to 50 ppb organic Cl as CH C12-
6. Calculated point of impingement i-hour average chloride concentration
based on guideline emission rate (yg Cl/m ).
7. Approximate calculated point of impingement i-hour average chloride
concentration based on dichloromethane and data from Tables 5 and 6,
Section 5-3 (ug Cl/m3).
*Since no PCB's were detected in the flue gas samples, none of the
residual organic chloride values in column 5 should be attributed to
PCB's in any of these stages.
-------
151
If the concentration of residual organic chloride in the
flue gas (column 7, Table F.I) were taken to represent PCB's in a "worst
case" (recall that no PCB's were detected), the calculated point of
impingement concentration by the Pasqui11-Gifford method is approximately
equal to typical measured ambient air concentrations of PCB's in urban
areas. There would appear to be little cause for concern when the
emission guidelines are met in this process configuration.
The above discussion relates only to the wet-kiln tests at
St. Lawrence Cement. If the same guidelines were applied to another kiln,
for example the 'ower-level, multi-stack dry kiln at St. Lawrence Cement,
the same emission rate as measured in the wet kiln tests would lead to a
significant ground level concentration. For example, even if the
stringent guideline for PCB emissions were met in this hypothetical case,
the maximum ground level concentration is calculated by virtual source
methods (applied to the dry kiln stack configuration) to be 0.07 ug/m
(as Aroclor 1242) which is 7 to 70 times the typical measured ambient
air concentration of 0.001 to 0.01 yg/m3. The point is that some caution
should be exercised in applying the results of these tests to another
stack configuration.
In the case of the third waste burn, it was estimated that the
total organic vapour concentration in the flue gas during the tests was
about 50 ppb, of which about 30 ppb was found to be dichloromethane.
Since none of the components of the original fuel mixture were detected,
however, the maximum residual concentration of Aroclor 12lt2, for example,
could be only about 1 ppb or less (noting the approximate detection limit
of 0.3 ppb). For speculation purposes only, in this worst case, 1 ppb
Aroclor }2k2 in the flue gas (at ambient temperature) would correspond
to about 11 ug/m3 PCB (or about 5 ug/m3 as chloride) and about 0.1 kg
of Aroclor \2k2 emitted per day (2k hours of continuous operation). In
reality, this quantity is an overestimate of the total emissions and is
probably not significant.
P. 2 Gas Chromatographic Analysis of Process and Emission Test Samples.
p.2.1 Chromosorb 102 adsorption tube analysis
Duplicate Chromosorb adsorption tubes for each of the waste
burns were analyzed by the Air Quality Laboratory, Laboratory Services
-------
152
Branch, Ontario Ministry of the Environment for support and cross-check
of analytical results obtained by the Ontario Research Foundation. Each
adsorption tube sample represented 60 litres of emission gas. Analysis
was done after desorption at 120°C into an evacuated glass vessel by
injecting a 1 ml gas sample into a gas chromatograph. This instrument
was equipped with a 9 foot x 1/8 inch Chromosorb 102 column held at a
temperature of 180°C, and a Sc H electron capture detector.
The results from these analyses are summarized in Tables F.2
to F.*». All Chromosorb tube samples, including those from "Baseline B",
were found to contain small but measurable amounts of six or seven compounds
and traces of a further six to eight compounds. These "measurable"
compounds were almost certainly chlorinated, iow-molecular weight
compounds which, in the tables, have been designated 'toajor" and 'toinor",
according to the size of their gas-chromatographic peaks. Of these
compounds, only chloroform had been tentatively identified and quantitatively
determined. All other compounds separated from each sample were estimated
together by relating the sum of the gas chromatographic peak areas with
those obtained from a chloroform standard. Their concentrations are
listed under "Other Compounds".
The results shown in Tables F.2 to F.4 indicate that the total
of all chlorinated compounds in the Chromosorb adsorption tube samples
corresponds to concentrations only in the low parts per billion range
in the stack emission gases.
The agreement between these results and those of the Ontario
Research Foundation and TRW is only qualitative. The generally small
amounts (ppb range) of organic chloride vapours found by ORF are
confirmed but trends from stage to stage are not reproduced. Such
disagreement is merely indicative that the limits of sensitivity of
the combined sampling and analytical methods are being approached.
F.2.2 Fuel feedstock analysis
Samples of the waste chlorinated hydrocarbon fuels used in
waste burns B and C were analyzed in the laboratory of Professor F.W.
Karasek, Department of Chemistry, University of Waterloo under a
research grant from the Air Resources Branch, Ontario Ministry of the
Environment.
-------
TABLE F.2.
153
ST. LAWRENCE CEMENT WASTE BURN EXPERIMENTS - TEST 1
CHROMOSORB ADSORPTION TUBE ANALYSIS
Waste Burn
Experiment
A
B
C
Basel ine B
Blank
(Chromosorb)
"Major"
Compounds
3
1
2
2
0
"Minor"
Compounds
1»
5
4
5
0
Chloroform
ppb
1.192
0.058
0.329
0.055
-
Other
Compounds
ppb
7.121
0.646
2.909
2.487
-
Total
Compounds
ppb
8.313
0.704
3-238
2.542
-
TABLE F.3. ST. LAWRENCE CEMENT WASTE BURN EXPERIMENTS - TEST 2
CHROMOSORB ADSORPTION TUBE ANALYSIS
Waste Burn
Experiment
A
B
c
Basel ine B
Blank
(Chromosorb)
"Major"
Compounds
4
4
3
0
0
"Minor"
Compounds
3
1
4
7
0
Chloroform
ppb
1.462
1.464
0.924
0.022
•*
Other
Compounds
ppb
8.827
5.270
12.089
0.443
Total
Compounds
ppb
10.289
6.734
13-013
0.465
TABLE F.4. ST. LAWRENCE CEMENT WASTE BURN EXPERIMENTS - TEST 3
CHROMOSORB ADSORPTION TUBE ANALYSIS
Waste Burn
Experiment
A
r
Basel ine B
Blank
(Chromosorb)
"Major"
Compounds
5
3
3
3
0
"Minor"
Compounds
0
3
8
4
0
Chloroform
ppb
2.356
0.481
0.560
0.030
Other
Compounds
ppb
15-773
1.916
2.089
5.077
Total
Compounds
ppb
18.129
2.397
2.649
5.107
-------
F.2.2.1 Preparation of fuel samples. All samples were diluted ten
times with Burdick and Jackson "Distilled in Glass" methylene chloride.
A 0.1 ml aliquot of the fuel was made up to 1.0 ml with the methylene
chloride.
Methylene chloride was chosen for the solvent based on previous
fuel samples analyzed in this laboratory.
F.2.2.2 Analysis of fuel samples. All samples were surveyed by gas
chromatography using a 10' x 2 mm i.d. glass column with a specially
prepared, high resolution column packing, referred to as Ap-20M. A
similarly prepared solution of Aroclor 1242 was also chromatographed for
comparison.
The GC conditions used are shown in Table F.6 and were closely
monitored by the use of a HP 5830 A Digital Gas Chromatograph using a
flame ionization detector. The area of the solvent peak was rejected
to facilitate comparison of GC data.
Mass spectral data were obtained using a Perkin-Elmer 900 GC
interfaced to a HITACHI RMU-6 magnetic mass spectrometer via a Biemann -
Watson effusion separator. The spectra were initiated and counted manually.
All spectra were obtained with 70 eV ionizing voltage at 250 C.
Samples 1 and 9 were analyzed using GC/MS techniques as above.
F.2.3 Results and Discussion
Samples 8 to 10, the aromatic chlorinated fuels, displayed only
seven compounds with individual integrated area percentages greater than
]% and only nine with similar area percentages greater than 0.5%- There
appear to be only three major components: one with ca. 73% area, one
with ca. 4.51, and one with ca. 6.8%. This appeared to be the case for
all the aromatic fuels. A representative bar chromatogram is shown in
Figure F.I. The peak at four minutes in the aromatic fuel was ca_. 73%-
Samples 1 to 7, the PCB fuels, displayed several peaks in
addition to those of the aromatic fuels, samples 8 to 12. Comparison of
the additional peaks with a chromatogram of Aroclor 1242 indicates that
these peaks are due almost entirely to PCB compounds (see Figure F.I).
To illustrate this observation, the GC data from the aromatic
fuel was added to the GC data from Aroclor 1242 and the computer program
-------
155
TABLE F.5. FUEL SAMPLE IDENTIFICATION
Sample No.
1
2
3
k
5
6
7
8
9
10
11
12
Chemtrol
u
u
1 1
1 1
u
1 1
Chemtrol
n
n
M
it
Comments
Line Sample (PCB's),
n u n
9
II II II
1
It II II
II II II
II U II
II II II
Line Sample (Aromati
M u n
M u n
n n n
n n M
Jan.
Jan.
Jan.
Jan.
Jan.
Jan.
Jan.
cs),
9
9
9
9
3/76
V76
5/76
6/76
7/76
8/76
, 10:00 AM
, 12:00
, 9:00 AM
, 8:1*5
9/76*
Dec.
Dec.
Dec.
Dec.
Dec.
10/75**
11/75**
12/75**
13/75**
15/75**
*Sample appeared to have water in it.
**Sample had reacted with liner of sample vial cap.
TABLE F.6. GAS CHROMATOGRAPH CONDITIONS
HP 5830-A GC
TEMP 1
TIME
RATE
TEMP 2
TIME 2
50
2kO
50 min
INJ TEMP
FID TEMP
FLOW RATE
CARRIER GAS
ATTENUATION
280°
300°
28 ml /min
HELIUM
g
2 256
-------
25
20
15
10
5
25
20
5
25
20
15
10
5
-
-
-
-
-
u,
AROM FUEL
ii ,
—
CLOR
•»
i
1242
,
r
i i
_
,
r
PCB FUEL
-
-
_
, .
, i i
1
I i
. i
,,
ZAREA 529693
f •
i
ZAREA 706087
ill
III
ZAREA 608448
,
1 i
lll.lli.l I
U!
10
20
30
40
50
60
70
RETENTION TIME (MIN)
FIGURE Fl REPRESENTATIVE BAR CHROMATOGRAM
-------
157
treated the data as though it originated from a single sample. This
computer-produced bar chromatog ram, titled "AROM + CLOR", was then plotted
in comparison with the bar chromatogram of the PCB fuel sample as shown
in Figure F.2. The data were not subjected to any manipulation other than
straight combination of the data. Inspection of the two plots shows such
a close correspondence of GC peaks between the computer reconstructed
mixture and the actual PCB fuel mixture that there is a reasonable assurance
they are the same.
A GC/MS analysis was conducted on sample 1, one of the PCB
fuels. A representative gas chromatogram with numerically identified
peaks is shown in Figure F.3- Mass spectra were obtained for these
numerically identified peaks. The probable identity, molecular weight
and/or degree of chlorination is shown in Table F.7.
Since all of the peaks are found in the PCB fuel and only some
of the peaks are found in the aromatic fuel, an estimate of the dilution
or mixing of these compounds can be derived. During the gas chromatograph
analysis, all variables were held constant; thus, absolute integrated
area counts are indicative of changes in relative concentrations. These
results are shown in Table F.8. The units are area counts and the values
4
shown are 1/10 of output values.
Peak #3 is absent in the aromatic fuel and in Aroclor 1242 but
present in the PCB fuel. The mass spectrum of peak #3 suggests
tr ichlorobenzene.
Peak #19 is present in small amounts in the chlorinated
aromatic fuels, absent in Aroclor 1242 but present in increased concen-
trations in the PCB fuels. A definitive mass spectrum of this peak was
not obtained but was suggestive of lack of halogenation and is most
likely an unsaturated hydrocarbon of high molecular weight.
F.2.4 Conclusions
The aromatic fuel is primarily o-chlorotoluene and the PCB fuel
is a mixture of Aroclor 1242 and the aromatic fuel. Area data for peak
#1 indicated an initial dilution of approximately one-third (by comparison
of area counts for aromatic fuels to area counts for sample 1) which
increased to approximately four-fifths by the end of sampling data.
-------
25
20
15
10
< 5
Ul
$
f 25
|
^ 20
15
10
5
"*
PCB FUEL ZAREA 608448
-
, ,
1
.,.11 , lll.ll.l .1 .11 II .III. It Ml J
I i i
AROM + CLOR ZAREA 1235780
III III, ,1,
10 20 30 40 50 60 7G
\n
OB
RETENTION TIME (MIN)
FIGURE F2 COMPUTER RECONSTRUCTED BAR CHROMATOGRAMS FOR PCB FUEL AND
AROMATIC FUEL PLUS AROCHLOR 1242
-------
19
15
13
16
14
17
18
11
10
PCB FUEL
3/1/76
vn
10
52 48 44 40 36 32
28 24
TIME (MIN)
20 16 12
8
FIGURE F3 GAS CHROMATOGRAM FROM GC/MS ANALYSIS OF SAMPLE PCB FUEL
-------
160
The results of the University of Waterloo group with respect
to Identification of components of the waste burn B and waste burn C
fuels are in agreement with those of ORF and TRW.
TABLE F.7. MS DATA FROM GC ANALYSIS OF SAMPLE PCB FUEL
(Reference to Figure F.3-)
Peak No. Comments
1 MW 126, 1 chlorine - spectra suggest o-chlorotoiuene
2 MW 160, 2 chlorines - spectra suggest aromatic compound -
dichlorotoluene
3 MW 180, 3 chlorines - spectra suggest trichlorobenzene
k MW appears to be 216 with 4 chlorines - possible identity
W1*,
5 MW appears to be 216 with 4 chlorines - spectra very
similar to peak #4
6-8 MW 222 2 chlorines - spectra suggest dichlorobiphenyls -
PCB's
9-12 MW 256 3 chlorines - PCB's
13-16 MW 290 4 chlorines - PCB's
17 & 18 MW 324 5 chlorines - PCB's
19 MW uncertain - no suggestion of halogenation
-------
TABLE F.8. AREA COUNTS (x 10k)
Peak No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
8 9 10 11 12 1
10/12/75 11/12/75 12/12/75 13/12/75 15/12/75 3/1/76
3860 3976 3769 3994 3842 1200
80.5 82.5 78.8 82.5 80.4 24
20
243.9 247.6 240.6 251.4 245.1 74
363.1 362.1 360.5 373.3 363.4 154
158
76
340
465
82
- - - - 217
543
265
197
287
191
102
52
7.8 8.6 8.4 8.2 8.6 177
.9
.4
.2
.]
.2
.9
.4
.6
.7
.7
.5
.0
.0
.0
.3
.0
.1
.5
2
4/1/76
1142
23-
23.
72.
148.
154.
75.
333.
454.
80.
211.
518.
271.
183.
292.
191.
107.
51.
168.
1
8
3
1
8
7
7
8
6
8
0
3
8
4
1
5
5
9
3
5/1/76
913.4
19-0
38.0
64.5
155.0
160.6
77.0
348.4
451.2
84.9
212.4
503.6
275.5
179.7
302.5
186.8
134.8
54.9
143.8
4
6/1/76
770.6
16.1
48.2
60.6
150.2
156.5
86.5
328.8
431.3
84.3
203-3
481.8
258.3
177.0
284.3
180.2
143.2
50.1
120.5
5
7/1/76
766.
16.
50.
61.
153.
160.
88.
335.
439.
85.
207.
489.
265.
178.
292.
178.
149.
51.
120.
,8
6
8/1/76
753.0
1 15.7
9
7
8
2
6
2
3
6
9
1
9
6
8
6
1
3
5
49.
58.
152.
161.
89.
336.
440.
85.
208.
491.
268.
178.
294.
188.
151.
49-
120.
7
3
0
1
4
7
6
2
9
2
6
3
5
4
4
5
9
7
9/1/76
702.4
14.6
57.8
60.4
157.6
167.8
95.1
351.3
457.2
89. 1
217.1
507.6
282.1
186.7
309.8
190.5
167.5
50.2
H5.9
-------
APPENDIX G
LABORATORY ANALYSIS RESULTS FROM THE
ST. LAWRENCE CEMENT FACILITY TEST
by
D.G. Ackerman, J.F. Clausen and C.A. Zee
TRW Systems Group
-------
165
APPENDIX G
LABORATORY ANALYSIS RESULTS FROM THE ST. LAWRENCE CEMENT FACILITY TEST
G.1 Summary
Analysis of the flue gas samples for organic composition gave
the following results:
- Hydrocarbons were not detected in the samples of flue gas with
detection limits ranging from 0.6 mg/m to 0.009 mg/m .
- Low molecular weight chlorinated organic compounds such as
methylene chloride, chloroform, and carbon tetrachloride were
specifically searched for by gas chromatography with electron
capture detection (ECD) and were generally not detected in most
of the samples. Four samples indicated the possible presence
of some of these compounds, but all at levels of less than
0.1 mg/m of flue gas.
- Polychlorinated biphenyls were searched for by GC/ECD and GC/MS
and were not found in any samples at the GC/MS detection limit
of 3 ug/m of flue gas.
Trace metal concentrations in the flue gases, as determined in
the samples taken with the EPA sampling train, were all less than 10 yg/m ,
with the exception of lead in the WBC tests. The emission of lead during
the WBC tests averaged 0.12 mg/m .
Analysis of the clinker product and discard dust samples for
organic composition gave the following results:
- Hydrocarbons were not detected in solvent extracts of any of
the clinker product or dis.card dust samples. The level of
detection by GC/ECD was 5 yg/g of sample or lower.
- Polychlorinated biphenyls were not detected in any of the
clinker or dust samples by GC/ECD. Detection limits for PCB's
in the samples were typically 0.0^ yg/g or better.
The inorganic character of the clinker products and electrostatic
precipitator discard dusts was changed very little if at all by the addition
-------
166
of the chlorinated wastes to the SLC wet kiln process. Trace metal
levels in the WBB and WBC test samples did not differ significantly
from the baseline samples.
Analytical techniques selected for analysis of flue gas
constituents which include vapours, condensables and particulate matter
were gravimetric, IR, LRMS, GC, GC/MS, SSMS, ICPOES, and AAS. These
were selected on the basis of sensitivity and selectivity criteria to
enable identification of species in the flue gas at concentrations on
the order of 0.1 mg/m , representing the threshold level of the most
toxic species as defined by OSHA and other occupational health and
safety organizations. Detection limits for many of the techniques
extend to vg/m^ levels. However, specific analyses to identify compounds
below the level of interest, 0.1 mg/m , were not routinely performed.
The techniques used were both qualitative and quantitative in nature, with
an intended accuracy range of plus or minus a factor of two to three.
G.2 Introduction
In cooperation with Environment Canada and the Ministry of
Environment of Ontario, the United States Environmental Protection Agency
supplemented the Canadian studies through participation in analyses of
samples acquired from various streams during test burns of chlorinated
hydrocarbons in the St. Lawrence Cement Co. wet process kiln. These
analyses were performed by TRW Defense and Space Systems Group, Redondo
Beach, California, under contract to the U.S. EPA. Samples were provided
to TRW through the courtesy of the Ontario Research Foundation and
St. Lawrence Cement Co., of Ontario, Canada.
Emphasis in the EPA sponsored work was directed toward analyzing
for input waste residual compounds and by-products in the clinker (product)
dust from the electrostatic precipitators, and the various component
samples from the EPA Method 5 and ORF sorbent trap trains. These analyses
were limited to the two test burns performed using chlorinated aromatic
and polychlorinated biphenyl waste blends. In addition, background anal-
ytical work was accomplished on samples acquired during the two baseline
tests in which the kiln was fired using only residual oil as a fuel.
Analyses were also performed on the two waste blends. This work was done
to supplement the Canadian laboratory results as well as to acquire
-------
167
additional evaluation data in accordance with U.S. EPA protocol used
for ongoing chemical waste incineration test programs in the United States.
A brief economic analysis was made by St. Lawrence Cement Co.
using data from both their own operations and from Chemtrol Pollution
Services Inc. (Appendix E).
G-3 Analysis Techniques
The purpose of the analyses performed on the samples from SLC
was to identify and quantify:
- known hazardous species present from the waste as determined
by pretest analysis of a sample of the waste material,
- secondary decomposition products (e.g., incomplete combustion
products which are predictable), and
- other species found to be present but which are not predicted
or otherwise expected.
A complete list of the samples received by TRW for analysis is
given in Table G.I. This table shows what portion of the total sample
collected by the trains was sent to TRW as well as the respective burns
and test numbers for the received samples. The coding system used to
uniquely identify each sample is shown in Figure G.I. These codes are
used in subsequent tabulations of data and results.
The first step in the analytical approach involved various
extraction and preparation procedures to separate the organic and inorganic
constituents and/or to concentrate the samples in a suitable form for
analysis. The extracts and concentrates were then analyzed by the
techniques to be described in this section.
G.3.1 Extractions and sample preparation
G.3.1.1 Solvent extrarts received from ORF. With a few exceptions, solvent
extractions for organic species were performed by ORF using pentane or
hexane. Subsequently, i aliquots were received by TRW and combined
according to the plan shown in Figure G.2. These combinations were
performed because:
1) the i aliquots represent relatively small gas sample
volumes, thus making constituent concentrations very low
and difficult to measure, and
-------
TABLE G.I. SUMMARY OF SAMPLES RECEIVED FROM EACH WASTE BURN
Staple Description
iPA Train
• Probe rinse
Filtered Insoluble*
Solvent mtr*ct
Aqueous solution aft*r ftltr«tton Md txtr«ctlon
• Filter
Solvent extract
Filter with perttculate
• Io»1ngtrs
Solvent extract
AqHOOttS Solution *fUr extraction
Otf Train
• In-iuck filter
Solvent extract
Filter with particuUte
t Prate rinse
Filtered tntolubles
Solvent extract
ftqeeoui solution after filtration and extraction
e NaOM Uvtnger
Solvent extract
, Aqueous solution after extraction
e Distilled water Inptnger
Solvent extract
Aqueous solution after extraction
e ChrOBOsorb 102 sorbent tubes
Portion Recelvid
out of Total
Sanple
All
1/4
All
1/4
All
1/4
AH
1/4
All
All
1/4
All
1/4
All
1/4
All
1/4
torn and Test Ho. of Seattle* Origin
IU
1
2
3
X
I
MM
1
X
X
X
X
2
X
X
X
X
3
I
X
X
X
X
I
X
X
UK
1
X
X
X
X
X
X
X
2
X
3
X
MJ
1
^
X
2
X
I
1
X
X
X
X
X
X
X
o>
CO
-------
SAMPLE CODES CONSIST OF 3-5 SECTIONS
Burn
Sampling Train
Train Component Extracted
Any Special Preparation
Test No.
First Baseline on
Primary Fuel - BLA
Waste Burn on
Chlorinated
Aromatics - WBB
Waste Burn on
PCB Blend - WBC
Second Baseline on'
Primary Fuel - BL>B
Standard EPA
Method 5 Train - EPA
Filter/Insolubles - F
Probe Rinse - PR
Caustic Impinger - NAOH
Water Impinger - h^O
Combined Impingers - I
Extracts of Acidified
Solutions - AE
ORF Designed
"Sorbent" Train- ORF
Tl. T2. or T3 if
sample Is from
only one test
Blank if sample
is a combination
of all three
tests
•FOR EXAMPLE, ALL OF THE EXTRACTED SAMPLES FROM ONE WASTE BURN WOULD BE CODED AS FOLLOWS:
WBB-ORF-FE
WBB-ORF-PRE & NAOHE
WBB-ORF-NAOH-AE-T3
WBB-ORF-H2OE
WBB-EPA-FE & PRE
WBB-EPA-IE
= Waste Burn B, ORF Train, Combined Filter Extracts from
all three tests
= Waste Burn B, ORF Train, Combined Probe Rinse and Caustic
Impinger Extracts from all three tests
= Waste Burn B, ORF Train, Extract of the acidified caustic
impinger solution from Test 3
= Waste Burn B, ORF Train, Combined Water Impinger Extracts
from all three tests
= Waste Burn B, EPA Train, Combined Filter and Probe Rinse
Extracts from all three tests
= Waste Burn B, EPA Train, Combined Impinger Extracts from
all three tests
vx>
FIGURE G.I. TRW SAMPLE CODING SYSTLi!
-------
170
2) all three tests of each waste were performed at one
nominal operating condition.
In general, following this plan, the organic concentrate obtained for
each extracted sample type (e.g., filters, impingers, solids, etc.) from
each of the two sampling trains were the combination of all three tests
of each waste. Exceptions to this plan were the BLA and aqueous samples
for which only one of the three test samples were received (see Table G.I).
At no time were EPA train samples combined with ORF train samples.
Small aliquots (2-5 ml) were first taken of the "as received"
extracts and set aside for analysis of volatile compounds that would
be lost in the next step which was to concentrate the remaining solvent
sample using Kuderna-Oanish evaporators and a steam bath.
The solvent extracts of the probe rinses, in addition to being
combined as in the plan shown in Figure G.2, were also combined with the
sample (EPA train-filter, ORF train-first impinger) whose juxtaposition
in the sampling train and similar physical characteristics permit the
combination. The rinses of the EPA train were made of the glass probe
liner in front of the filter. However, with the ORF train which only
has a short nozzle in front of the in-stack filter, rinses were made of
the U feet of probe and tubing between the filter and the first liquid
impinger. The resulting solutions were filtered through standard filter
paper and then extracted by ORF.
G.3.1.2 Solids, aqueous solutions, and filters. The solid samples
were prepared for organic analyses by extraction in a SoxMet apparatus
for 2k hours with disti1led-in-glass grade pentane. These extracts were
concentrated with Kuderna-Danish evaporators to a 10 ml volume. Prepara-
tion of the solid samples for inorganic analyses consisted of a low-
temperature plasma ashing to remove possible organic interferences.
Other samples for inorganic analysis also required some degree
of preparation. Aliquots of the aqueous impinger and probe rinse samples
were taken and acidified with nitric acid to stabilize any metals present.
Each of the insoluble residues obtained by filtering the probe rinses was
combined by ORF with its matching particulate filter for treatment as one
sample. These insolubles/fiIter samples were plasma ashed and then
-------
171
TAKE AND
COMBINE
ALIQUOTS FOR
ANALYSIS OF
VO LATHES
GC
GC/MS
SOLVENT
EXTRACT T.
COMBINE, CONCENTRATE,
AND DILUTE TO 10ml
T
TO ANALYSIS OF
ORGANIC CONCENTRATES
FIGURE G.2. PLAN FOR THE COMBINATION OF ORF SOLVENT EXTRACTS
VACUUM GAUGE
SORBENT TUBE
d t=
t
FURNACE
=T~W
A
/^
DEWAR
FLASK
FORLK
&=
<
2
^
•V
^
k
•>
~s
vl
,/c
— Q
* C VACU
PLJ Qy— •
FIGURE G.3. DESORPTION SYSTEM FOR CHROMOSORB 102 TUBES
-------
172
extracted by refluxing constant boiling aqua regia over the each sample
for two hours. The resulting acid digests were made to 50 ml for
analysis.
G.3.1.3 Sorbent tubes. The Chromosorb 102 sorbent tubes from the
ORF sampling train were prepared for analysis by a quantitative desorption
technique. The apparatus used for the desorption is shown schematically
in Figure G.3- The sorbent tube is attached with a minimum length of
Tygon tubing to a 20 ml glass sample bulb and is then heated in a small
furnace to 185° ±10°C.
With valve A closed and valves B and C open, the sample bulb
was evacuated and immersed in liquid nitrogen. Valve A was then opened
for 30 minutes to allow the contents of the sorbent tube to transfer to
the sample bulb. Valves A and C were then closed, the LN_ flask removed,
the sample bulb allowed to equilibrate to room temperature, and the
pressure recorded.
The volume of the entire manifold system, including each sample
bulb, was carefully calibrated and on the average was 68.8 cc. With this
information and the recorded pressure (assuming the temperature to be
constant), the volume of the material desorbed was calculated according
to the ideal gas law:
P V = P V
pri V2
where: Pj = the pressure measured, in mm Hg
Vj = the volume of the sample bulb and manifold, ^68.8 cc
?2 = 760 mm Hg
V2 = the calculated volume of desorbed material at 1 atm.
The desorbed material was recovered from the manifold by
reimmersing the sample bulb in LN2« Valve B was then closed and the
sample bulb removed from the desorption system for analysts while still
at negative pressure.
G.3.2 Analytical methods
The extracted and prepared samples were analyzed by various
methods. The specific analyses selected depended to some extent on the
samples' forms, which were:
-------
173
For Inorganics
- Solids
- Aqueous and acid solutions
For Organics
- Aliquots of the neat solvent extracts
- Concentrates of the solvent extracts
- Desorbed materials from the sorbent tubes
G.3.2.1 Analyses for inorganics. The inorganic composition of the solid
samples was determined by spark source mass spectrophotography (SSMS)
with electronic detection, which will detect elements present down to
a concentration of 1 ppm. This SSMS technique has an accuracy from 100-500%
and will survey the sample for all the elements except gases (i.e.,
H, N, 0, He, Ne, Ar, Kr, Xe, and Rn) and Hg because of its high volatility.
Other elements with appreciable volatilities, such as Be and Se, have far
reduced accuracies in this analysis.
The aqueous and acid solutions were first surveyed for metals by
inductively coupled argon plasma optical emission spectroscopy (ICPOES).
The ICPOES analysis determines 32 elements, including most of the toxic
elements of interest in the program, down to ppb levels, with an accuracy
of 100-200%. The purpose of this survey was primarily to check that the
metals in the test samples were present in approximately the same amounts
relative to each other as they were in the waste material. Those elements
which, from the results either of the ICPOES survey or of the analysis of
the waste material, seemed to be present at potentially toxic levels, were
determined quantitatively by atomic absorption spectrometry (AAS). The
sensitivity of this method varies from approximately 1.0-0.001 ppm for
the elements which were determined, with an accuracy between 10-50%.
G.3.2.2 Analyses for organics. Organic constituents of the test samples
were determined by a combination of the following techniques:
- infrared spectrometry (IR);
- gas chroma tography (GC) with either an electron capture (EC)
or flame ionization detector (FID);
- low resolution mass spectrometry (LRMS);
- combined gas chromatography/mass spectrometry (GC/MS).
-------
The aliquots of the neat solvent extracts were analyzed only
for volatile organochlorine compounds. EC/GC was used for this analysis
and the samples were compared to standards of CH-C^. CHC1_, and CCl^ as
well as the original waste material. The column was 183 cm x 0.635 cm o.d.,
glass, 1.5* 0V 17 and 1.95* OJ-1 on Chromosorb WHP.
The concentrates of the solvent extracts were analyzed to identify
the less volatile compounds that would not be lost by evaporation of the
extract. An aliquot of each sample was first taken and evaporated in a
clean, 7 ml weighing bottle at ambient conditions to remove the solvent.
The residue was then weighed on a micro balance to O.OOJ mg and scanned
by IR as an evaporated thin film. The sample was returned to the weighing
bottle and stored for use in the LRMS analysis. To obtain an LRMS spectra,
the sample was rinsed into the glass vessel for the solids probe inlet system
and the solvent evaporated again at ambient conditions. The IR and LRMS
analyses yield qualitative information about the classes or types of
compounds (e.g., hydrocarbons, phenols, ROM's, etc.) present as well as
an idea of the complexity and toxic nature of the concentrated sample.
The sensitivity of the LRMS solids probe technique for specific
compounds in an organic matrix can be estimated even though the mass
spectrometer is not strictly considered a quantitative instrument. At
typical mass spectrometer operating pressures of 10 torr, the relative
volatility of organic materials at normal solids probe temperatures
(100-250°C) is not an important factor. That is, the volatilities of
materials of interest are sufficiently high for adequate detection.
The solids probe of the mass spectrometer Is in such close proximity to
the ionizer that sample diffusion does not significantly reduce sensitivity.
Previous experience and contacts with other laboratories using similar
equipment indicate that if 10 micrograms of an organic compound is
present in a solids probe along with other material, it will be detected
to the extent that the ten or more strongest fragment peaks will be
recorded. Since the typical weight of sample placed in the solids probe
is 1 milligram, the required weight percentage of a given organic compound
to be detected in an organic matrix is approximately one percent. Therefore,
any compounds not detected by LRMS were assumed to be present at less than
one percent levels.
-------
175
The organic compounds in the extract concentrates were separated
and quantified by GC using both flame ionization and electron capture
detection. Columns and GC conditions are given below:
Flame ionization detection
- Varian i860, dual differential FID
- Columns: dual, 183 cm x 2 mm i.d., stainless steel, 3-5%
OV-17 on 100/120 mesh Chromosorb WHP
- Temperatures: column, ambient for 5 minutes, then ambient
to 275°C at 10°C/min; detector, 275°C; injector, 250°C.
- Flow rate: helium carrier at 30 ml/min; hydrogen at
30 ml/min.; air, 300 ml/min.
- Attenuation: 1 x 10 a/mv
Electron capture detection
- Tracer MT-150, 63Ni single ECD
- Column: 183 cm x 0.4 cm i.d., glass, 1.5* OV-17 and 1.95*
QF-1 on 80/100 mesh Chromosorb WHP
- Temperatures: column, 200°C; detector, 225°C; injector 225 C.
- Flow rates: pre-purified N2 carrier through column at 60 ml/min.;
detector purge at ^0 ml/min.
- Polarizing voltage: H»V; bucking range -2 x 10 ; input
attenuation, 102; output attenuation, X2 to X64.
G.*» Analytical Data
Data obtained from the various analyses performed will be
presented in this section in the following order:
Chlorinated Hydrocarbon Wastes Tested
- Chlorinated Aromatics
- Polychlorinated Biphenyls (PCBs)
Samples Obtained from the EPA and ORF Sampling Trains
- Organic Constituents
- Inorganic Characterization
Solid Residues and Effluents frnm the SLC Kiln Process
- Clinker Products
- Electrostatic Precipitator Discard Dusts
-------
176
G.4.1 Chlorinated hydrocarbon wastes tested
Samples of the chlorinated wastes burned at St. Lawrence Cement
were taken by ORF from the liquid waste feed tank on each day of testing.
Composites of these were sent to TRW. The analyses used to characterize
the wastes and determine the expected compounds of interest in the test
burn samples were:
- thermal content - gross heat of combustion;
- viscosity;
- specific gravity;
- loss on ignition (LOl);
- C, H, N, S, and halogens;
- infrared spectroscopy (IR);
- gas chromatography/mass spectroscopy (GC/MS); and,
- spark source mass spectroscopy (SSMS).
G.4.1.1 Primary fuel. In addition to the two wastes, a sample of the
primary fuel, bunker "C" oil, was also received. Of the analyses listed
above only the gross elemental determination (C, H, N, S, and halogens)
and a trace metal scan by X-ray fluorescence (XRF) were performed on the
oil. Physical properties were not determined since they are fairly standaid
and accurate compound identification was not necessary since the background
of compounds contributed by burning the bunker "C" oil as fuel can best be
determined by analysis of the combustion test samples from the baseline
burn.
The elemental determination on the oil gave the following
results:
- 86.87% C
- 9-89% H
- C.36% N
- 2.62% S
- 0.05% Cl (total halogens as chlorine).
-------
177
The XRF scan detected six elements at the levels listed below:
E 1 emen t Approximate Concentration (ppm)
Si >500
P 50-500
Ni 5-50
V* 5-50
Ca <25
Ti <5
G.A.I. 2 Chlorinated aroma tics. The aromatic waste was a dark brown,
low viscosity liquid that was visually free of sediment. The measured
physical characteristics were:
- thermal content - 5170 kcal/kg (9310 Btu/lb);
- viscosity - 1.09 centistokes at 38°C (100°F);
- specific gravity - 1.281 at 16°C (60°F) ;
- LOi - 99.982.
Elemental analyses performed showed the following composition:
C
H
- 0.028% N
- 0.019% S
- 49.10% Cl (total halogens as chlorine).
Analytical techniques used to determine the organic composition
included IR and GC/MS.
15.
The 1R spectrum indicated the waste to be composed primarily of
aromatic hydrocarbons as well as aliphatic alkanes and alkenes wi th a strong
response in the 600-800 cm"' region which can correspond to C-C1 bonding.
No indication of any other functional groups such as phenols, ethers, or
amines was found. Comparison of the waste spectrum with spectra from
the Sadtler indices for the compounds found by GC/MS, showed the principle
3
Element is potentially toxic - OSHA TLV of <1 mg/m^ for an eight-hour
exposure.
-------
178
constituent to be o-chlorotoluene, the most characteristic peak of which
is a strong, sharp peak at 750 cm
GC/MS
Chromatographic separation was carried out on a Finnigan GC/MS
system using two columns. One was Chromosorb 101, temperature programmed
from 30° - 220°C at 10°C/min, and the other was OV-17 temperature programmed
from 30° - 275°C at 10°C/min. The compounds shown in Table G.2 were
identified and their concentrations calculated based on relative peak
areas found from both columns and assuming equal response to the total
ion monitor of the Finnigan instrument. These results reflect the fact
that, in the course of performing the tests at SLC, the waste feed tank
was not emptied between wastes. Thus, the material for the "aromatic"
burns contains ^33% chlorinated aliphatics carried over from the previous
burn (as a cost savings to the program, TRW did not plan to analyze
samples from the chlorinated aliphatics burn). The primary constituent
of the aromatic waste is chlorotoluene (52.5$) with the remainder made up
of three other chlorinated aromatics: dichlorotoluene, octachlorocyclopentene,
and octachloronaphthalene.
TABLE G.2. ORGANIC COMPOSITION OF AROMATIC WASTE BY GC/MS
Estimated Concentration
Compound (% w/w)
Acetone '-6
Methylacetate °-3
Dichloromethane O.H
Chloroform n-'
Carbon Tetrachloride 13-9
Dichloroethane 3-1
Trichloroethane 2-5
Tetrachloroethane °-^
Trichloroethylene 1>0
Toluene K0
Chlorotoluene 52.5
Dichlorotoluene ^-6
Dimethyl Benzene (Xylene) 0-9
Octachlorocyclopentene 3-7
C,rtClQ (Octachloronaphthalene) 3-0
10 o
-------
179
SSMS
Trace metals in the chlorinated aromatic waste were determined
by spark source mass spectroscopy (SSMS). The waste was first ashed by
a sulphated dry method. A very small amount of ash, 0.014%, was obtained
of which the major elements were Ba, Fe, Na, P, S, Si, and Ti (the
sulphur most likely represents the contribution from the H-SO, added during
ashing). Other elements detected by SSMS down to 1 ppb are shown in
Table G.3- This analysis was performed to determine what potentially toxic
elements might have been present in the waste at high enough concentrations
to warrant quantitative examination of the flue gas test samples. Given
the semi-quantitative nature of the data, the SSMS results should not be
construed as a quantitative characterization of the waste material.
TABLE G.3. TRACE METALS IN THE CHLORINATED AROMATIC WASTE BY SSMS
Element
Al
Ca
Cu
Mq
«7
Zn
K
Crb
Pbb
Mn,
Hgb'°
j
Sr
Pt
B
Cob
Ni
W
Mo
a
Approximate
Concentration
(ppm)
0.7-1.5
0.7-1-5
0.7-1.5
0.7-1.5
0.7-1-5
0.6
0.4
0.3
0.07
0.07
0.06
0.05
0.05
0.05
0.03
0.01
0.008
Element
Sn
Zr
Vb
Asb
Ag.
Cdb
Ce
La
Sbb
Beb
Bi
Ge
Li
Rb
Seb
Sm
Yb
Approximate
Concentration
(ppm)
0.008
0.008
0.007
0.006
0.003
0.002
0.002
0.002
0.002
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
aThe accuracy of this technique ranges from 100 to 500 percent.
Potentially toxic metals - OSHA TLV of <1 mg/m3 for an eight-hour
exposure.
CHg was determined by a highly quantitative atomic fluorescence
technique.
-------
180
G.4.1.3 Polychlorinated biphenyls (PCB's). The PCB waste was a medium
brown, low viscosity liquid with fine suspended particulate that tended
to float to the top and cake out on the sides of the container. The
measured characteristics were:
- thermal content - 6,710 kcal/kg (12,083 Btu/lb);
- viscosity - 5.87 cent!stokes at 38°C (100°F);
- specific gravity - 1.196 at 16°C (60°F);
- LOI - 99-90%
Elemental analyses performed showed the following composition:
- 57.9**% C
- 5.23* H
- 0.018$ N
- 0.12% S
- 33.601 Cl (total halogens as chlorine).
The analytical techniques used to determine the organic composi-
tion of the waste were IR and GC/MS.
IR
The IR scan showed strong peaks at 1100 cm and 1^50 - 1^80 cm
which are characteristic of biphenyls, and a peak at 750 cm indicative
of remaining o-chlorotoluene in the waste. Other small peaks in the
spectrum matched those in the aromatic waste IR scan which substantiates
the GC/MS results discussed below.
GC/MS
Analysis of the PCB waste was performed by the same GC/MS
methods and conditions as for the aromatic waste, and the compounds that
were identified are shown in Table G.*t. The PCB waste composition showed
the effect of previous waste material left in the feed tank. It was composed
of 12% chlorinated aliphatics, 33% chlorinated aromatics, and k$% PCBs.
There were 22 separate PCB compounds identified including several isomers
of each of the multiple chlorinated biphenyls.
SSMS
Trace metals in the PCB waste were determined by SSMS. The waste
was first ashed by a sulphated dry method. A small amount of ash, 0.15%,
-------
181
TABLE G.4. ORGANIC COMPOSITION OF PCB WASTE BY GC/MS
Estimated Concentration
Compound (% w/w)
Water 0.7
Acetone 1.3
Methylacetate 0.5
Methanol 0. 4
Chloroform 1.4
Carbon Tetrachloride 6.8
Ethanol 5.2
Dichloroethane 0.6
Trichloroethane 0.7
Hexachloroethane 0.2
Trichloroethylene 1.6
Tetrachloroethylene 0.2
Xylenes 2.3
Toluene 1.0
Chlorotoluene 17-2
Dichlorotoluenes 2.4
Trichlorotoluenes 3-3
Octachlorocyclopentene 4.0
Chlorobiphenyl 0.5
Dichlorobiphenyls 8.4
Trichlorobiphenyls 15-5
Tetrachlorobiphenyls 11-15
Pentachlorobipheny1s 2-6
Hexachlorobipheny1s 1-2
Heptachlorobiphenyls 0.5
Octachlorobipheny1s 0.2
Di-n-octyl or Di-ethyl Hexyl Phthalate 4.0
CI«C'Q " Octachloronaphthalene 2.5
10 0
was obtained of which the major elements were Fe, Na, P, Pt, and Si.
Other elements detected by SSMS down to 1 ppb are shown in Table G.5-
This analysis was performed to determine what potentially toxic elements
might have been present in the waste at high enough concentrations to
warrent quantitative examination of the flue gas test samples. Given the
semi-quantitative nature of the data, the SSMS results should not be
construed as a quantitative characterization of the waste material.
G.4.2 Samples obtained from the EPA and ORF sampling trains
Samples obtained by the two trains from the WBB, WBC, and BLB
tests were received for analysis. A list of the samples derived by ORF
-------
182
TABLE G.5. TRACE METALS IN THE PCB WASTE BY SSMS
Approximate3 Approximate
Concentration Concentration
Element (ppm) Element (ppm)
K
Zn
Cu
S
Mg
Al
Ni
Mo.
Bab
Ti
B
Ca
W .
CrJ
Pbb
vb
Ag
Mn.
Sbb
Sn.
« b
As
Zr.
Cob
7-15
7-15
7
5
5
k
k
3
2
2
1
1
J
0.8
0.8
0.7
0.6
0.5
0.5
0.5
0.2
0.2
0.1
Seb
Srb c
Hgb»C
Cdb
Th
La
Nb
Bi
Li
Rb
Ta
Au
Nd
Hf
Sm
Beb
Dy
Ga
Ge
Pr
Sc
U
0. 1
0.07
0.06
0.05
0.02
0.01
0.01
0.009
0.009
0.006
0.00k
0.003
0.003
0.002
0.002
0.001
0.001
0.001
0.001
0.001
0.001
0.001
aThe accuracy of this technique ranges from 100 to 500 percent.
Potentially toxic metals - OSHA TLV of <1 mg/m3 for an eight-hour exposure
cHg was determined by a highly quantitative atomic fluorescence
technique.
from each train was presented in Table G.I. Al 1 of these samples were
scheduled to be analyzed until, in early April, ORF's review of their
analytical data brought them to conclude that the solvent extracts of the
BLB samples had been contaminated with PCB waste material [G.I] which was
confirmed by the analyses at TRW. The contamination appeared to have
occurred before the split of the samples into i aliquots. In view of
this development, analysis of the BLB solvent extract samples was
discontinued and extracts from a BLA test were sent as replacements.
However, there was no evidence of a contamination problem with
the remaining BLB samples. Thus, the filters, probe rinse insolubles,
-------
183
aqueous solutions, and Chromosorb 102 sorbent tubes from the BLB
tests were analyzed for a background reference to the corresponding
WBB and WBC samples.
All of the test samples were coded for ease of reference.
The coding used was presented in Figure G.I, which also showed the
combination of the probe rinse samples with other appropriate samples.
The reasons and methods for combining the probe rinses with either the
filter (for the EPA train) or impinger (for the ORF train) samples,
as well as the preparation and analytical techniques for all the samples
were described in Section G.3-
Where constituents found in the test samples are reported as
concentrations in the effluent gas, the sample gas volumes needed for the
calculations were taken from Section 5-3 of the report. Volumes for the
ORF train were taken from Table A, and volumes for the EPA train from
Table 7.
t
G.It.2.1 Organic constituents. Samples for the analysis of organic
composition were in the form of:
- aliquots of the neat solvent extracts;
- concentrates of the solvent extracts;
- residues from the evaporation of the extract concentrates; and
- desorbed materials from the sorbent tubes.
These samples were first surveyed for their qualitative nature by gravi-
metric, infrared spectrometry (IR), and low resolution mass spectrometry
(LRMS) techniques. Quantitative determinations and analysis for specific
compounds, such as chlorinated hydrocarbons and PCB's, were then performed
by gas chromatography (GC) and combined gas chromatography/mass spectrometry
(GC/MS). The results obtained from all of these analyses are presented
in the following paragraphs. The methods themselves are described in
Section G.3.
Qualitative data. The purpose of the qualitative or survey analysis
was to identify any material in the samples which were not present
in the original waste, and therefore were not expected. In addition,
this survey analysis searched for secondary or incomplete combustion
products which would result from the waste being converted to compounds
-------
other than CO., HO, and HC1. The qualitative data will be described
in two groups: l) Extract concentrate residues, and 2) material desorbed
from the sorbent tubes.
1) Extract concentrate residues - The results of these qualitative
analyses on the residues, obtained by evaporating aliquots of the solvent
extract concentrates, are summarized and presented in Table G.6. This
table includes:
- The identification of the extract samples according to the
coding shown in Figure G.I. This coding shows how certain
related samples have been combined.
- The Sample/Aliquot Ratio shows the total volume of extract
prepared by ORF in the numerator and the volume received by TRW
in the denominator. One-fourth of each extract was sent to
TRW.
- The amount of residue found in TRW's portion of the total
extract is expressed as milligrams (mg). Values have been
corrected for appropriate blanks.
- The Total Residue is that amount (in mg) of residue found in
TRW's portion, multiplied by the Sample/Aliquot Ratio. This
value is indicative of what may be found in the total extract
if analyzed by the same procedures.
Inspection of the IR and LRMS data revealed that the only
materials detected in the evaporated residues of the various extracts
(except for the contaminated BLB samples) were silicones, hydrocarbon
oils, fatty acids or fatty acid esters, and phthalic acid esters.
These same materials were also present in the blank and control
samples; their concentrations vary from sample to sample as estimated by
LRMS. These materials, as classes of compounds, are not toxic and would
not be considered an emissions hazard even if they were present in much
higher concentrations. These materials are not believed to have come from
the combustion gas. Samples from the EPA train did not show increased
levels of these materials consistent with more than ten-fold sample volumes.
In the analysis of trace organics, It becomes very difficult to
completely avoid these ubiquitous oils, plasticisers, lubricants, antioxidants,
-------
185
TABLE G.6. SUMMARY OF ORGANIC QUALITATIVE SURVEY ANALYSES
OF SAMPLE EXTRACTS
Sample Identification
Baseline Burn A
8LA-ORF-FE-T3
BLA-ORF-PRE+NAOHE-T3
BLA-ORF-H2OE-T3
BLA-EPA-PRE+FE-T3
BLA-EPA-IE-T3
Baseline Burn 8
BLB-ORF-FE
BLB-ORF-PRE+NAOHE
BLB-ORF-H20E
BLB-EPA-PRE+FE
BLB-EPA-IE
Waste Burn 6
WBB-ORF-FE
WBB-ORF-PRE+NAOHE
WBB-ORF-HZOE
WBB-ORF-NAOH-AE-T3
WBB-EPA-PRE+FE
WBB-EPA-IE
Waste Burn C
WBC-ORF-FE
WBC-ORF-PRE+NAOHE
WBC-ORF-H20E
WBC-ORF-NAOH-AE-T1
WBC-EPA-PRE+FE
WBC-EPA-IE
Total Extract -^-^^^
(nin^"^
^--"tRW Aliquot (ml)
200/50
200/50
100/25
300/75
100/25
600/150
600/150
300/75
900/225
300/75
600/150
400/100
200/50
N/AC
900/225
300/75
600/150
600/150
300/75
N/AC
900/225
300/75
Survey Residue
Found 1n TRW
Aliquot (mg)
3.23
N/D*
N/ir
0.12
2.09
0.34
N/D6
0.50
0.48
0.04
0.07
N/Db
N/Db
0.28
1.50
11.18
0.58
N/Db
14.83
0.23
1.12
6.25
Total Residue
(mg, corrected
for Aliquot
Factor)
12.92
N/D (<1.00)
N/D (<2.68)
0.48
8.36
1.36
N/D (<1.00)
2.00
1.92
0.16
0.28
N/D (<1.00)
N/D (<2.68)
1.12
6.00
44.72
2.32
N/D (<1.00)
59.32
OQO
.yd
4.48
25.00
"There 1s no significant difference In the results between test runs. The materials in the
residues are not believed to have come from the combustion gas.
bN/D • Not detected at levels higher than the blank.
°N/A • Not applicable; the total caustic Implnger contents from one test were used.
-------
186
etc., that can enter the test matrix in very small amounts despite
the most careful and extensive preparation of equipment and selection
of reagents. One cannot be sure whether the source of these materials
is the hardware, the reagents, the samples themselves, or a combination
of these. A large number of blank and control samples might have
distinguished the source, but due to the noncritical nature of the materials,
such an examination was clearly not warranted.
It was pointed out that the survey analysis of the BLB samples
known to contain PCB's IG.1] did, In fact, reveal their presence. Evidence
of PCB.s was clearly shown in the LRMS data for the following BLB
samples:
- ORF train, combined probe rinse extracts and caustic
impinger extracts;
- ORF train, water impinger extracts;
- EPA train, combined probe rinse extract and filter extract.
Since these PCB's were found only as estimated minor constituents in samples
whose total weights were less than one milligram, the ability of the mass
spectrometer to detect these small quantities was clearly established. The
LRMS procedures can usually detect 10 micrograms, or about one percent of
a typical 1 milligram sample. The presence of a specific compound in an
organic matrix is not normally detected by IR with adequate certainty
unless present at 10 percent or higher. The sensitivity and detection
limits of these techniques were discussed in Section G.3-2.
It is stressed that chlorinated species were specifically
searched for in all of these residue samples and none were found except
for in the contaminated BLB samples described above.
2) Materials desorbed from the sorbent tubes - The Chromosorb 102 sorbent
tubes from the ORF sampling train were thermally desorbed to recover the
collected sample according to the procedure described in Section G.3-1-
The desorbed gases were then analyzed qualitatively by LRMS using a gas
inlet system. The conclusions drawn from reduction and interpretation
of the LRMS data were that all of the sorbent tubes contained essentially
the same types of materials; however, the relative amounts of these
constituents did appear to vary from sample to sample.
-------
187
Although the LRMS is a qualitative technique, estimates can
be made through inspection of the data to provide an indication as to
whether a species is present as a trace, minor, moderate, or major component
Such a summary for the six selected sample sorbent traps that were analyzed
is presented in Table G.7. An unused sorbent trap was also analyzed by the
same procedure to estimate background contributions by the Chromosorb 102.
Unfortunately a leak in the system resulted in the loss of sample and,
therefore, usable data. Only carbon dioxide was seen in the mass spec
data at greater than trace levels.
The volatile compounds removed from the sorbent traps and
collected in the gas sample bulbs were largely the oxygen, nitrogen,
carbon dioxide, and water that are the usual species found in combustion
gases. The remainder of the compounds detected were those that could
indicate incomplete combustion, or possibly the formation of other
compounds. The levels at which these remaining compounds could be
present were estimated on the basis of LRMS instrument response for key
peaks attributed to these trace residual organics. These estimates
indicated that the C, to C5 hydrocarbons present as a major component
of these trace residual organics could be present in the sampled
combustion gases in the 1 to 10 ppm range. The minor or trace components
of these trace organics were believed to be present at very much lower
levels.
A summary of the LRMS peak patterns observed in the spectra
of the desorbed gases and the compounds assigned to these patterns is
given in Table G.8.
Quantitative data, duantitat ion of the compounds detected and identified
in the test samples was performed by GC and GC/MS techniques.
1) P*tract concentrates - The chrbmatography with flame ionization
detection was performed chiefly to analyze organic species. It should
be noted that the FID is quite sensitive to chlorinated hydrocarbons
as long as chlorination is not complete. At the electrometer settings
used, the sensitivity, or the minimum detectable quantity, was 0.002
yg/yl as benzene, naphthalene, or Aroclor 1232.
The chief result of the analyses using the FID was that none
of the concentrated samples showed peaks other than those found in the
-------
TABLE G.7. APPROXIMATE CONSTITUENT LEVELS OF TRACE VAPOURS
DESORBEO FROM SORBENT TUBE SAMPLES BY LRMS
Sorbent
Tube
Sample
WBB-T1
WBB-T2
WBC-T1
WBC-T2
BLB-T1
BLB-T2
a
Constituent Level
Cl-5
Hydrocarbons
Minor
Major
Major
Major
Major
Major
NUrome thane
Trace
Minor
Minor
Minor
Moderate
Minor
NO » N0« ,
and £
Possibly
Ethanol
Trace
Major
Major
Major
Moderate
Moderate
Benzene
Trace
Moderate
Minor
Moderate
Major
Moderate
Methyl
Chloride
NO"
Moderate
NO
NO
NO
NO
Substituted
Benzene
NO
Trace
Trace
Trace
Trace
Trace
Methyl
Siloxane
Major
Minor
Minor
Trace
Trace
Trace
QD
CD
aThe range of trace to major levels represents 0.1-10 ppm concentrations in the flue gas.
bNO - Not Detected.
-------
189
TABLE G.8. SUMMARY OF THE INTERPRETATION OF LRMS SPECTRA FOR
TRACE VAPOURS DESORBED FROM SORBENT TUBE SAMPLES
Peak Pattern
(Peaks at Atomic
Mass Units
(AMU)) Assignment
15
16
17,
27,
55,
14,
44
30,
30,
78,
50,
91
147
18
29, 41, 43,
57
16, 28, 32
46, 61
46
50, 51, 52
52, 15
, 207, 281
Methyl group, CH
Methane, CH^
Water
C -C,. hydrocarbons
£. 5
Ni trogen/oxygen
co2
Possibly ni tromethane CH, NO--
NO , NO * and/or ethanol
Benzene
Methyl chloride CH.C1
Substituted benzene ring (e.g., toluene)
Methyl siloxane
*The 30 AMU peak was very large and it is believed that nitric oxide,
nitrogen dioxide, ethanol and nitromethane may have all contributed to it.
solvent controls. Consequently, a "less than" value was assigned in
terms of benzene for the analysis of hydrocarbons, or Aroclor 1232 for
the analysis of chlorinated organics. The results are given in Table G.9.
The values presented in Table 6.9 were derived in the following
manner. The sample labeled WBB-EPA-FE+PRE, for example, totaled 225 ml
as received. A 213 ml portion of this sample was concentrated to 10 ml
in a Kuderna-Danish evaporator. The chromatogram of the concentrated
sample showed no peaks other than solvent. Since the aromatic waste
contained considerable chlorinated hydrocarbons, quantitat ion as Aroclor
1232 was considered appropriate.
Thus, WBB-EPA-FE+PRE contained <0.002 yg/yl of species as Aroclor
1232 The amount of material in the entire received sample volume of 225
ml was determined. This value was multiplied by four, since only one-quarter
-------
TABLE G.9- RESULTS AND DETECTION LIMITS FROM GC/FID ANALYSIS OF THE CONCENTRATED EXTRACTS
• Swple ID
ILA-EPA-FE+PRE-T3
-IE-T3
-ORF-FE-T3
-H20C-T3
-PRE+HAOHE-T3
WBB-EPA-FE+PRE
-IE
-ORF-FE
-H90E
-PftE+WWHE
MBC-EPA-FE*rHE
-IE
-ORF-FE
-IbOC
-PRE+NMHE
«9/wl as .
Halocarbon*
0.052
0.26
0.26
0.061
0.21
*D(<0.0038)«
0.89
0.12
0.26
0.053
0.019
(lost)
ND(<0.0038)
ND(<0.0038)
0.26
Volume of .
Sample (ml)0
75
25
50
25
50
225
75
150
50
100
225
75
150
75
150
Aliquot
Factor*
67/20
21/10
46/10
21/10
42/20
213/10
63/10
138/10
42/10
42/10
213/10
63/10
138/10
63/10
138/10
Friction of
SampleC
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
Volume of Sample
Gas. Std. mJ
3.76
3.76
0.32
0.32
0.32
9.62
9.62
0.83
0.83
0.83
9.97
9.97
0.73
0.73
0.73
i
Concentration
In Flue Gas
(•a/-3)
0.001
0.003
0.04
0.09
0.06
ND(<0. 00002)
0.004
0.006
0.02
0.003
0.00008
(lost)
N0(<0.00002)
ND(<0. 00002)
0.015
*Vo1uM of sample received by TRW.
b Aliquot Factor: for example 67 ml of th* 75 »1 BLA-EPA-Fe+PRE sample received was concentrated to 20 ml for analysis.
°Fract1on of Sample: All samples received by TRU were 251 of the total sample.
Instrument was calibrated with Aroclor 1232. However, the A rod or peak pattern
was not found 1n any sample chromatogram. High sensitivity GC/MS (3pg/m3 of flue
gas) did not detect any chlorinated species In samples WBC-ORF-FE and
UBC-ORF-H20E.
"»'• Not detected. Values 1n parentheses Indicate the detection limits.
-------
191
of the whole sample was divided by the volume, in m , of gas sampled.
A concentration value of <0.009 mg/m in the flue gas was thus derived
for the sample WBB-EPA-FE+PRE.
Since the minimum detectable quantity is constant and since both
the volume of gas sampled and the volume of extract vary, this method of
calculation gives different values of mg/m . It must be stressed that
these analyses showed no species other than solvent.
Since the ECD is much more sensitive to polychlorinated species
than the FID, the analyses using electron capture detection were performed
to determine polychlorinated compounds below the level that could be
detected by the FID analyses. At the electrometer setting used on the
electron capture detector, the minimum detectable quantities for typical
species of interest were:
- Aroclor 1232: 0.0038 ng/yl
- Naphthalene: 0.066 ug/vl
- Benzene: 0.11 yg/yl.
The most noteworthy result of the GC analyses using the ECD
was that none of the concentrated extracts showed peak patterns corresponding
to those given by Aroclor 1232 (the PCB mixture most similar to that
burned in waste burn C) or Aroclor 1242. The quantified ECD results are
given in Table G. 10.
Some of the concentrated sample extracts, e.g., WBB-EPA-FE+PRE,
showed no peaks other than solvent. Most of the extracts showed several
very small peaks. After quantitation, all but one of the samples were
below the level of interest, 0.1 mg/m3, and no further effort was expended.
One sample, WBC-ORF-FE, showed a number of peaks, some of which were large.
Another sample, WBC-ORF-H2OE, showed a large number of small peaks. Both
of these samples were analyzed by GC/MS. The GC/MS analyses of the filter
extract indicated one vanishing small hydrocarbon peak, a small phthalate
ester peak, and 10 peaks which were various trimethylsi lyl compounds
(IMS). The GC/MS analyses of the water extract showed extremely small
hydrocarbon peaks and 16 peaks that were trimethylsi lyl compounds.
Because these TMS compounds were most likely the result of some contamina-
tion, the peaks appearing in the chromatograms were considered spurious.
-------
TABLE G.10. RESULTS AND DETECTION LIMITS FROM GC/ECD ANALSIS OF THE CONCENTRATED EXTRACTS
Sample ID
M.A-EPA-FE+PRE-T3
-IE-T3
-ORF-FE-T3
-M20E-T3
-PRE+NMHE-T3
UBB-EPA-FE+PRE
-IE
-OKF-FE
-H20E
-PRE+NAOHE
WBC-EPA-FE+PRE
-IE
-ORF-FE
-H20E
-PRMMOHE
M9/u1 as
Benzene"
ND(<0.002)e
N0<
NO
NO
ND
<0.002
<0.002
<0.002
<0.002
.
-
-
-
-
_
-
-
-
"*
ufl/iil as .
Halocarbon0
.
.
.
-
ND
ND
ND
NO
ND
NO
NO
NO
ND
NO
<0.002
<0.002
<0.002
<0.002
<0.002
<0.002
<0.002
<0.002
<0.002
<0.002)
Volume of
Sample, ml4
75
25
50
25
50
225
75
150
50
100
225
75
150
75
150
Aliquot
Factor"
67/20
21/10
46/10
21/10
42/20
213/10
63/10
138/10
42/10
92/10
213/10
63/10
138/10
63/10
138/10
Fraction of
SampleC
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
1/4
Volume of
Sample Gas.
Std. m3
3.76
3.76
0.32
0.32
0.32
9.62
9.62
0.83
0.83
0.93
9.97
9.97
0.73
0.73
0.73
Detectable
limits of
Concentration
In Flue Gas,
mg/m3
ND(<0.05)
NO
ND
ND
ND
ND
ND
ND
ND
NO
NO
NO
ND
NO
NO
<0.03)
<0.3)
<0.3)
<0.6)
<0.009)
<0.01)
<0.1)
<0.009)
<0.01)
<0.1)
ND • Not detected, values 1n parentheses Indicate the detection limits.
-------
193
These GC/MS analyses did not detect any chlorinated species in
either sample (typical sensitivity of 3 yg/m of the flue gas).
2) Aliquots of neat extracts - There was concern that the process of
concentrating the sample extracts, as received in pentane and/or hexane,
might cause the loss of relatively low boiling halocarbons such as
chloroform which might be expected in the samples. Therefore, the retained
portions of the as-received extract samples were chromatographed on the
Tracer instrument. The column was operated isothermally at 70°C to increase
resolution of low boiling compounds. The results are given in Table G.ll.
Most of the samples showed no peaks other than solvent. Of those
samples showing peaks other than solvent, all quantified below the level
of interest (0.1 mg/m ), and no compound identification by GC/MS was
performed.
3) Materials desorbed from sorbent tubes - The sorbent traps were desorbed
as discussed in Section G.3-1. Portions of the desorbed vapours contained
in the sample bulbs were chromatographed isothermally at 70 C with electron
capture detection. The samples were qualitatively similar to the sorbent
trap blank (unused chromosorb 102) and the baseline B (fuel oil burn)
samples. The waste burn B samples each contained a peak which might be
CH Cl: WBB Test 1, 0.0001 mg/m3; and WBB Test 2, 0.0002 mg/m3. The
waste burn C samples were essentially the same, qualitatively and
quantitatively as the blank and baseline B samples.
G.4.2.2 Inorganic characterization. Inorganic elemental concentrations
In the flue gas were determined by analysis of the particulate filters
and the aqueous probe rinse and impinger samples. The filters (including
the filtered probe rinse insolubles) were all acid digested. Out of these
20 digests (18 test samples and two blanks), three samples were prepared
for a trace element survey analysis by combining equal aliquots from each
of the three tests of WBB, WBC and BLB using the filter samples from only
the EPA sampling train. The EPA train samples were selected on the basis
of 1) representing larger sample gas volumes, and 2) having been taken by
a method specific for accurate and representative particulate sampling.
The results of the survey analysis are shown in Table G.12. The blank
-------
TABLE G.I1. RESULTS AND DETECTION LIMITS FROM GC/ECD ANALYSIS OF THE UNCONCENTRATED EXTRACTS
BLA-EPA-FE+PRE-T3
-IE-T3
-ORF-FE-T3
-H2OE-T3
-PRE-NAOHE-T3
WBB-EPA-FE+PRE
-IE
-ORF-FE
-H2OE
-PRE+NAOHE
WBC-EPA-FE+PRE
-IE
-ORF-FE
-H2OE
-PRE+NAOHE
ng/yl as
Halocarbonc
ND(<0.0038)d
ND(<0.0038)
ND(<0.0038)
0.0051
N0(<0.0038)
ND(<0.0038)
0.40
ND(<0.0038)
ND(<0.0038)
ND(<0.0038)
ND(<0.0038)
0.10
ND(<0.0038)
0.15
ND(<0.0038)
Volume of
Sample
75
25
50
25
50
225
75
150
50
100
225
75
150
75
150
Fraction, of
Samp 1 e
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
Vo 1 ume
of Sample,
Gas, Std. m
3.76
3-76
0.32
0.32
0.32
9.62
9.62
0.83
0.83
0.83
9-97
9.97
0.73
0.73
0.73
Concentration
in Flue-Gas,
mg/m
ND(<0.0003)
ND(<0.0001)
ND(<0.002)
0.0016
ND(<0.0024)
ND(<0. 000*0
0.012
ND(<0.003)
ND(<0.0009)
ND(<0.0018)
ND(<0.0003)
0.0030
ND(<0.0030
0.061
ND(<0.0031)
Volume of sample received by TRW.
Fraction of sample. AH samples received by TRW were 25% of the total sample.
The instrument was calibrated with Aroclor 1232. However, the Aroclor peak pattern was not found
in any sample chromatogram. High sensitivity GC/MS (3 yg/m^ of flue gas) did not detect any chlorinated
species in samples WBC-ORF-FE and WBC-ORF-H20E.
NO
Not detected. Values in parentheses indicate detection limits.
-------
195
TABLE G.12. TRACE METAL SEMI-QUANTITATIVE3 SURVEY OF FILTER DIGESTS BY ICPOES
Average Concentratiorr (mg/m^)
Element
Al
Ba
B
Ca
Cd
Cr
Co
Cu
Fe
Pb
Mg
Mn
Ni
P
K
Si
Ag
Na
HO
Sr
Ti
v
Zn
WBB
0.078
0.011
0.018
2.2
0.0006
0.003
NDC
0.003
0.18
0.01*7
0.076
0.003
0.003
0.014
14.
0.082
0.0003
1.8
0.004
0.002
0.0006
0.014
WBC
0.51
0.014
0.084
11.
0.002
0.006
0.001
0.007
1.1
0.12
0.38
0.018
0.006
0.055
32.
0.051
0.003
5-5
0.013
0.008
0.002
0.042
BLB
0.26
0.012
0.038
6.5
0.001
0.004
NDC
0.003
0.73
0.044
0.21
0.013
0.002
0.044
6.7
0.093
0.0004
2.4
0.0008
0.004
0.001
0.027
aAccuracy estimated to be a factor of ±2 or better.
Calculated based on average sample gas volumes of:
WBB - 3-2 mi*
WBC - 3-3 nC
BLB - 3-4 nr
Not corrected for filter contribution
cNot detected (<0.0003 ppm)
(an acid digest of unused filters) sample was not surveyed, so the values
reported in this table are uncorrected for the filter background.
Of the 32 elements that can be determined by the ICPOES analysis,
ten were not detected in the filter digest samples. These ten elements
with their lower detection limits are listed in Table G.13. along with
a calculation of the average detectable limit for each of these elements
in the flue gas.
-------
196
TABLE G.13- LIMITS OF DETECTION FOR UNDETECTED ELEMENTS BY ICPOES
Average Detectable3
ICPOES Detection Limit in Flue Gas
Element Limit (ppb) (mg/m^)
Au
As
Be
Eu
Mo
Se
Te
Sn
W
U
5
40
1
15
11
60
65
50
90
80
0 . 00008
0.0006
0.00002
0.0002
0.0002
0.0009
0.0010
0.0008
0.0014
0.0012
Based on an average sample gas volume of 3-3 cubic meters.
The results of this survey indicated only one element, lead, was
present at potentially toxic levels in the stack. While not a problem
with the stack configuration in the present study, the concentration and
the effect of thermal dispersion should be considered when extrapolating
these results to other kilns. To be sure of an accurate measurement of
the levels of lead and certain other toxic metals which the waste analysis
indicated might be present at levels of interest, a quantitative determina-
tion by AAS was performed on five elements. The results from this analysis
are presented in Table G.I4*.
The AAS results were corrected for background levels as determined
by analysis of the filter blank sample. It should be noted that back-
ground contributions were not only from trace contaminants in the filter
materials but also in some cases from matrix effects in the mixed acid
solutions.
AAS was also used to analyze for selected elements in the
aqueous probe rinse and impinger samples. Table G.15 lists the results
from these analyses and shows that, in general, nothing significant was
found. No background or blank solutions were available so the BLB results
should be used to correct for background effects. The results for these
aqueous samples were left in ppm because the total sample volumes were
unknown.
-------
197
TABLE G.ll*. CONCENTRATION OF TRACE METALS IN EFFLUENT GAS
PARTICULATE MATTER BY AAS
Waste
Burn
WBB
WBC
BLB
Train Test
EPA ]
2
3
EPA ]
2
3
EPA 1
2
3
Concentration
Ba
<0.008a
£0.008
£0.008
0.013
£0.008
0.027
£0.007
£0.008
£0.008
Cd
£0.001
£0.001
0.001
0.001
0.002
0.003
0.001
0.002
0.001
of Element (mg/m )
Co
£0.002
<0.001
£0.002
£0.002
£0.002
£0.002
£0.002
£0.003
£0.002
Cr
<0.009
<0.002
£0.00**
£0.008
£0.008
0.003
0.001
£0.007
£0.002
Pb
0.028
0.012
0.079
0.096
0.103
0.153
0.0*40
0.062
0.015
"£", a less than or equal to sign, indicates those elements which were
detected but not significantly above background levels.
TABLE G.I5. CONCENTRATION OF TRACE METALS IN AQUEOUS SAMPLES BY AAS
Tra i n/Component
EPA-Probe Rinse
Impingers
Waste
Burn
vs.
Basel ine
WBB
WBC
BLB
WBB
WBC
BLB
Concentration of Element (ppm)
Pb
ND
ND
ND
ND
ND
ND
Cr
1.9
0.03
ND
ND
0.02
ND
Cd
ND
0.01
ND
ND
ND
ND
Co
10.2
ND
ND
0.05
ND
ND
-------
198
Analysis of the filter acid digest, probe rinse, and impinger
samples from the ORF train by AAS confirmed the presence of those elements
found in the EPA train samples. However, since the ORF train was not
operated with the intent of collecting an accurate particulate sample
(i.e., isokinetic, traverses, etc.) the AAS results for this train were
not calculated out to concentrations in the flue gas.
G.4.3 Solid residues and effluents from the SLC kiln process
Of the four types of solid samples collected during the test
program:
- clinker product;
- clinker fines;
- electrostatic precipitator discard dust; and
- cement mixes,
analyses were performed only on the clinker product and discard dusts for
the following reasons. The clinker fines were recovered from air blown
through the clinker product and were then returned to the clinker storage.
The cement mixes were simply clinker product with gypsum added. Thus,
neither of these materials represented unique samples. The following
paragraphs describe the results of analyses performed on the clinker
products (CP) and discard dust (DD) samples. If significant amounts of
toxic materials were found fn the analyzed samples, further tests on the
remaining samples would have been conducted. However, this did not prove
to be necessary.
G.4.3.1. Organic constituents. Portions of the clinker product and
discard dust samples were extracted with pentane using a Soxhlet extractor
and preextracted paper thimbles. A blank sample was prepared by running
a pentane extraction on an empty, precleaned thimble. In addition, a doped
control sample consisting of ^30 grams of clinker, to which 0-9 mg
of a known PCB mixture had been added, was also extracted for analysis.
This doped sample represented a PCB concentration ^30 ppm in the clinker
product. These samples were all analyzed both qualitatively and quantita-
tively by the same techniques described in Section G.4.2.1.
Qualitative data. Aliquots of the pentane extracts were evaporated and
the residue was weighed and scanned by both IR and LRMS. These results
-------
199
TABLE G.16. RESULTS OF ORGANIC SURVEY ANALYSIS ON CLINKER PRODUCT AND
DISPOSABLE DUST SAMPLES
Sampl e
Ident i f icat ion
BLB-CP
BLB-DD
WBB-CP
WBB-DD
WBC-CP
WBC-DD
PCB Doped
Control
Weight of
Extracted
Sample (g)
30.270
24.185
29.445
18.948
21 .816
33.112
29.998
Weight of
Residue in
Extract (mg)
< 1 . 1 6*
1.31
1.30
1.16
< 1 . 1 6*
<1. 16"
< 1 . 1 6*
Qualitative Nature of Residue
Hydrocarbon oils, phthalate
esters, traces of fatty acids
and s i 1 icones
Al 1 of the above compounds
plus polychlor i nated biphenyls
The residue obtained from the blank thimble and pentane sample was 1.16
mg. Sample values varied with some being less than 1.16 mg due to the
variability of the background. Over this range, the differences in
residue weights are not significant.
are summarized in Table G.16. In most cases, the amount of weighable residue
found in the extracts did not exceed that found in the blank. The
materials identified by IR and LRMS are indicative of the low level
contamination by greases, oils, soaps, etc., that often accompanies
trace organic analysis. Hydrocarbon oils, phthalate esters, fatty acids,
and silicone compounds were found in all the extract residues including
the blank. No indication of the chlorinated species was found in any
of the test samples.
Po'ychlorinated biphenyls were clearly detected in the PCB
doped control sample. The LRMS spectrum contained peaks at 220, 222,
290, 292, and 294 AMU. The PCB's present at 30 ppm in the doped clinker
were easily found and it is certain that much lower levels could have
been detected. It is estimated that the mass spectrometer can detect
the strongest peaks from a specifically searched for compound when
present at the 10 ng level in a 1 milligram organic residue from the
clinker. This results in a lower limit of detection by this qualitative
-------
200
survey technique of about 1 ppb In the clinker or dust. The IR survey
data was not useful in detecting PCB's at this level.
Quantitative data. The clinker and dust extracts were analyzed by GC with
both FID and ECD detection. The results are given in Tables G.17 and G.18,
respectively. Using the FID, no peaks other than solvent were found, and
the results are expressed as minimum detectable quantities. Three of the
samples showed no peaks other than solvent when analyzed with the ECD.
The three samples which had peaks other than solvent using the ECD were
quantitated at well below the level of interest.
To ascertain the recovery of PCB's from the clinker and dust
samples, a 30 gram sample of clinker was doped with 0.9 mg of Aroclor
1232, extracted and concentrated in the same fashion as the clinker and dust
samples. The ECD chromatogram of this artificial sample matched the pattern
of a neat sample of Aroclor 1232, and the recovery of PCB's from the doped
sample was calculated as 25%. This recovery factor has been applied to
the clinker and dust samples in Tables G.17 and G.18. (The recovery
factor is the reason why the values for the samples in Table G.17 exceed
several ppm. If the minimum detectable quantity using the FID is 0.002
ug/yl, then after applying the recovery factor, the concentration in the
sample cannot exceed 0.008 ug/yl- For a 10 ml extract volume and a 30 g
clinker sample, the composition is thus <2.7 yg/g-)
G.i».3.2 Inorganic characterization. The CP and DD samples were also
examined for trace metal content by spark source mass spectrometry (SSMS).
The major constituents of both types of material were determined to be
aluminum, calcium, iron, magnesium, sodium, sulphur, and silicon. Other
trace inorganic elements which were detected down to a 1 ppm concentration
are listed in Table G.19. It should be remembered that SSMS data are
semi-quantitative. The values reported should be considered as being
accurate within 500% of the true value.
The trace metal character of the clinkers and dusts did not
change significantly from one waste burn to the other. The relatively
low feed rate of the wastes, compared to that of the precursor material
and the primary fuel, appeared to make any contribution of trace metals
from the wastes negligible.
-------
201
TABLE G.17. RESULTS AND DETECTION LIMITS FROM GC/FID ANALYSIS
Sample ID
BLB-CP
WBB-CP
WBC-CP
BLB-DD
WBB-DD
WBC-DD
yg/yl as
Halocarbon
ND(<0.002)b
ND(<0.002)
ND(<0.002)
ND(<0.002)
ND(<0.002)
ND(<0.002)
Volume of
Sample (ml)
10
10
10
10
10
10
yg/g of
Sample Material
ND(<3-0)
ND(<3-3)
ND(<1».5)
ND(<3.9)
ND(<5.1)
ND(<2.9)
The instrument was calibrated with Aroclor 1232. However, the Aroclor
peak pattern was not found in any sample chromatogram.
ND
Not detected. Values in parentheses indicate detection limits.
TABLE G.18. RESULTS AND DETECTION LIMITS FROM GC/ECD ANALYSIS
Sample ID
BLB-CP
WBB-CP
WBC-CP
BLB-DD
WBB-DD
WBC-DD
CP+PCB
ng/yl asa
Halocarbon
0.059
0.11
ND(
-------
202
TABLE G.19. SELECTED TRACE METALS IN SLC CLINKER PRODUCT
AND DISCARD DUST SAMPLES BY SSMS
Concentration (ppm)
WBB
Element
As
B
Ba
Be
Ce
Co
Cr
Cu
Cs
Dy
Ga
Ge
La
Li
Mn
Mo
Nb
Nd
Ni
Pb
Pr
Rb
Sc
Se
Sm
Sr
Th
Tf
U
V
Y
Zn
Zr
CP
<1.5
30
70b
NDD
10
ND
7
1
ND
1
1
ND
7
ND
100
1
3
10
7
ND
3
7
ND
ND
ND
700
1
700
ND
30
3
30
30
DD
3
30
70
ND
30
ND
10
10
10
ND
1
ND
7
^1000
100
1
3
10
3
70
3
100
1
10
ND
300
7
700
1
30
3
30
30
WBC
CP
<1.5 '
30
300
1
70
ND
30
3
ND
3
3
ND
30
30
300
1
7
10
7
3
10
10
ND
ND
1
700
7
700
3
30
7
70
100
DD
3
10
100
ND
30
3
10
7
30
1
1
ND
10
100
300
1
3
10
7
100
3
300
ND
3
1
700
7
700
3
30
10
100
70
BLB
CP
<2.9
30
70
7
10
1
30
7
ND
1
3
1
10
100
300
3
3
10
10
ND
3
10
7
1
1
700
3
700
1
70
3
30
70
DD
1
30
30
ND
10
ND
10
3
10
ND
10
3
7
100
100
1
3
7
3
30
1
70
3
1
ND
100
1
700
ND
30
3
30
10
SSMS data generally ranges within 500% accuracy.
ND - Not Detected
ppm).
-------
203
REFERENCES
G.I Communication, Gordon Thomas, Ontario Research Foundation to
Arnold Grant, TRW Systems, 7 April, 1976
-------
APPENDIX H
DEVELOPMENT, CONSTRUCTION AND EVALUATION OF A
COLLECTION SYSTEM FOR LOW MOLECULAR
WEIGHT HALOCARBONS
by
F.J. Hopton and G.H. Thomas
Ontario Research Foundation
Mississauga, Ontario
-------
207
APPENDIX H
DEVELOPMENT, CONSTRUCTION AND EVALUATION OF A COLLECTION SYSTEM
FOR LOW MOLECULAR WEIGHT HALOCARBONS
H.1 Summary
Numerous experiments have been performed to determine the
effectiveness of absorbents and adsorbents to remove and retain low
molecular weight halocarbons from a flowing gas stream. Initial studies
using a particulate sampling train, requiring a high gas flow rate of at
least 0.5 cfm, indicated that neither absorbents nor adsorbents would
efficiently collect compounds such as chloroform (CHC1 ) present in the
gas stream at low concentrations of a few parts per billion. An inert
adsorbent, e.g. Chromosorb 102, was found, however, to efficiently remove
and retain CHC1, at much lower gas flow rates of 0.2 - 1.0 litres/minute.
A collection system for low molecular weight halocarbons was constructed
and evaluated at low flow rates over sampling periods of up to four hours
duration. Results obtained from several tests indicated that the system
had a collection efficiency of better than 90% over the longest time
period. Desorption of adsorbed halocarbons was readily achieved by thermal
treatment of the adsorbent.
H.2 Introduction
Late in 1975, an experimental program to burn waste chlorinated
organic compounds in a rotary cement kiln was carried out at the St.
Lawrence Cement Company (SLC) plant in Mississauga, Ontario. The program,
sponsored by Environment Canada, was designed to obtain information on
- the use of waste chlorinated hydrocarbons as a supplemental
fuel to the kiIn;
- the reduction of a 1kali concentration of the clinker by the
chlorine contained in the waste materials added; and,
- the effect of emissions from the kiln on air quality.
In order to collect the required information, process and emission samples
were collected during periods when different composite wastes were burned
in the kiln. The emission samples were analyzed for trace chlorinated
organic compounds.
-------
208
During preliminary discussions between members of the Ontario
Ministry of the Environment (MOE), the Ontario Research Foundation (ORF)
and SLC prior to initiating the study, some concern was expressed about
the sampling methodology for trace chlorinated organic compounds. Since
a major program requirement was to obtain the particulate emission rate
during each burn, it was hoped that the EPA Method 5 sampling train, used
to collect particulate material, could also be used for the collection
of chlorinated organic compounds. However, it was not known whether
absorbents or adsorbents were more efficient for collection of the organic
compounds, or whether either collection medium would be efficient at the
high gas flow rates required for particulate sampling.
MOE, therefore, requested that ORF complete a laboratory
investigation to evaluate suggested procedures, and to develop an optimum
method for the sampling and collection of trace chlorinated organic
compounds which might be present In the kiln emissions. This report
describes the test program carried out in the laboratory and presents the
data obtained using specific chlorinated compounds as representative
pollutants.
H-3 Background Information
Previous studies involving the incineration of chlorinated
organic compounds [H.I] have indicated that, if complete combustion to
CO , HO and Cl or HCl is not realized, then trace amounts of volatile
compounds such as CC1,, CHC1. and CH-Cl. may be present in the combustion
gases. (Though it was unlikely that any HCl or C12 produced would pass
through a cement kiln without reacting to form alkali chlorides, it was
necessary to consider these gases as possible pollutants for analysis
also.)
The EPA Method 5 train for particulate sampling is also used to
collect organic compounds in the impingers by condensation and scrubbing.
The efficiency of collection for a particular substance is dependent on
gas flow rate and volatility of the organic compound. For compounds more
volatile than water, the collection efficiency is probably close to zero.
To collect these volatile compounds, specific organic solvents have been
used in the impingers, replacing the water normally present. There is,
however, little data available on the collection efficiencies obtained.
-------
209
In recent years, organic compounds in the ambient air have been
collected using inert adsorbents [H.2]. These materials have tended to
replace activated carbon for most applications since they are unaffected
by water vapour, and recovery of adsorbed species is considered, in
general, to be easier and more efficient. Recovery of adsorbed species
may be accomplished either by thermal desorption or solvent extraction.
An application of this technique has been to collect polycyclic organic
compounds from combustion effluents [H.3].
Bearing in mind that, in the study to be carried out at SLC
only trace amounts of more volatile organic compounds were expected to be
in the kiln emissions, it was decided to evaluate both adsorbents and
adsorbents in an EPA Method 5 sampling train for collection, retention
and subsequent recovery of compounds such as CHC1, and CH CJ . If a
collection method proved to be effective for capture of these compounds,
then it would almost certainly be effective for the collection of higher
molecular weight species, such as those chlorinated compounds present in
the waste feed material.
H-i* Test Methodology
The normal gas flow rate through a particulate sampling train
to maintain isokinetic sampling for most sources is 0.5 to 1.0 cfm. A
flow rate of 0.5 to 0.7 cfm was, therefore, chosen for initial experiments
using the EPA Method 5 train, which was illustrated in schematic form as
Figure A.5 in Appendix A.
H.A.I Retention and collection studies
H.A.1.1 Absorbents. Initially, tests were conducted to find a suitable
high boiling solvent for use in an impinger system, in order to trap low
molecular weight organochlorine compounds, e.g. CHC1_, CCl^, DCE and TCE.
The following solvents were tested for their suitability as an
impinger solution for use in an EPA stack sampling train:
- reagent grade toluene;
- reagent grade xylene; and,
- reagent grade decane.
All solvents were redistilled prior to use in order to remove
or limit to usable level-, interfering peaks in the GC-EC chroroatographic
-------
210
profiles which tended to obscure measurement of the components of
Interest.
Tests were made to see if trace quantities of volatile
chlorinated compounds could be retained in a solvent such as toluene
with a high gas flow rate passing through the train for a period of four
hours, a time anticipated for tests at SLC. A sampling train was,
therefore, assembled in the laboratory with probe and oven temperature
set above 250°F to warm incoming gas.
Toluene (100 ml) containing known amounts of 1,2-dichloroethane
(EDC) and 1,1,2-trichloroethane (TCE) (main constituents of the first
waste feed in the waste chloride study at SLC) was placed in iropinger A
(see Figure A.5). 100 ml of pure toluene was placed in impinger B,
impinger C was empty and impinger 0 contained silica gel. Air was
pulled through the sampling train after cooling the impingers to ice-
water temperature, at a flow rate of 0.5 cfm. After running for four
hours, the contents of each impinger were noted and analyses made for
EDC and TCE. At specific times during each test, the parameters normally
recorded during a source sampling test, such as impinger inlet and outlet
temperatures and orifice Ap, were recorded. At the conclusion of each
test, the total volume of gas sampled was determined.
This type of experiment was repeated, with 100 ml of water in
impinger A, 100 ml of toluene containing EDC and TCE in impinger B,
impinger C empty and impinger D containing silica gel. Both experiments
were then repeated using xylene and decane as absorbents. Additional
experiments were made using decane with known quantities of CCl^, CHCl^
and CHC1, in the standard impinger solution.
H.l».1.2 Adsorbents. Though a large number of adsorbents are capable of
adsorbing a wide variety of organic compounds, Chromosorb 102 was used for
the tests described below because it was readily available, and could be
purchased without delay from various companies supplying chromatographic
materials.
In order to determine fairly rapidly if Chromosorb 102 would
adsorb volatile chlorinated compounds, the following experiment was
performed. 100 ml of decane containing known quantities of CCl^, CHCl^
and TCE was placed in impinger B of the sampling train, and 25 gm of
-------
211
Chromosorb 102 in impinger D. Impinger A contained 200 ml of distilled
water and impinger C was empty. Room air at 0.5 cfm was then pulled
through the system for one hour and at intervals of 10 minutes the decane
solution was analyzed for chlorinated organic concentration. After one
hour the Chromosorb was removed from the impinger, placed in a septum
sealed container and heated to 100°C. Head-space samples were then
analyzed for the respective chlorinated organic compounds. All three
halocarbons present in the spike had been retained by the resin.
Following this experiment, a series of tests were carried out
using the ORF test duct system, depicted in Figure H.I. This system
allows dust or gaseous pollutants to be added at a controlled rate into
an air stream flowing through a one-foot diameter duct at flow rates
of 500 to 2500 cfm. Sampling ports are located at ideal and non-ideal
positions in the system, to allow sampling of the air stream with an
EPA type sampling train or other collection equipment. A low flow of air
was allowed to bubble through CHC1- in a midget impinger and bleed into
the main air flow in the duct. By varying the carrier flow through the
CHC1V concentrations of 10 to H*0 ppb of CHC13 were obtained in the duct
air stream. These concentrations were determined both from weight loss
of CHC1 with time and direct analysis of the air flowing in the duct.
Tests were made with Chromosorb in impingers C and D but considerable
carry over of the adsorbent occurred at flow rates of 0.5 - 0.7 cfm. This
was minimized in further tests by placing a fine mesh screen on top of the
Chromosorb layer in the impinger. At a CHC13 concentration of 13 ppb,
breakthrough of this pollutant with 25 gm of Chromosorb in the third
impinger occurred within 30 minutes.
It was recognised that high volume flow rates through the
adsorbent caused not only contaminant problems with the material in the
sampling train but saturated the bed fairly rapidly. This latter occurrence
was proved by repeating the experiment with the Chromosorb packed tightly
in a collector. A final series of tests were, therefore, completed,
evaluating both Chromosorb and activated carbon in tube collectors at
much lower air flow rates of 1-2 litres/minute. The CHC13 concentration
in the air stream was analyzed before and after the collector at selected
time intervals over a period of four hours. For both adsorbents collection
of CHC1 was more than 30% efficient for the duration of each test. S.nce
-------
A
212
1 DUST FEEDER
2 BLOWER
3 SAMPLING PORTS NON IDEAL LOCATION
6 SAMPLING PORTS IDEAL LOCATION
5 SAMPLING PORTS
6 TO ATMOSPHERE
FIGURE H.I TEST DUCT SCHEMATIC
-------
213
desorption of adsorbed species from Chromosorb was readily accomplished
thermally and since, unlike activated carbon, the adsorption characteris-
tics of Chromosorb are unaffected by moisture in the gas stream, Chromosorb
102 was selected as a suitable adsorbent for the SLC waste chloride program.
H. 5 Col 1ect i on System Eva 1uat i on
This section outlines the experimental protocol undertaken to
verify the performance of Chromosorb 102 as a collection media for the
concentration and analysis (by adsorption/desorption) of low molecular
weight chlorinated hydrocarbons from an air stream under conditions
relevant to field sampling. The concentration and analysis procedure is
essentially an extension of the method used to concentrate organohalogen
compounds present in potable water [H.4].
Adsorptlve capacity values were determined. Adsorptive capacity
is defined as the amount of solute vapour retained by a given weight of
sorbent. In this study only low levels of chlorinated hydrocarbons
(1 - 10 ppb) were examined since anticipated levels of organohalides in
the sampling streams were expected to be low (i.e. ng/m ).
The sampling cartridges consisted of glass tubes (11 mm i.d. x
various lengths) containing various depths of sorbent bed supported
by plugs of silanized glass wool. The tubes were conditioned by heating
to 200°C and passing a stream of NZ through them (l»0 ml/min) for four
hours.
Known concentrations of air-solute vapour mixtures were prepared.
The solutes examined were CHClj, CCl^ and dichloroethane (DCE). The
solutes were examined both individually and as mixtures. The air-solute
vapour mixtures were prepared as follows.
Microlitre quantities of organic compounds were injected into
gas jars of known volume. The gas jar was heated to * 50°C and, after
cooling, aliquots were removed and added to Tedlar bags. The bags
initially were evacuated and then half filled with air, at which stage
the aliquot from the gas jar was injected into the bag. The bag .as
finally filled to the desired capacity (6, 20 and 40 litres).
-------
A schematic diagram of the component parts of the test
assembly is shown below.
TEDLAR
BAG
Gas Jar
The various parts of the assembly were connected to each other
by means of Teflon tubing. Differing air-solute mixture concentrations of
CHC1. were used to determine the adsorptive capacity of the Chromosorb
102. The CHC1, concentrations used were 5 ppb and 10 ppb. A flow rate
of 250 ml/min was maintained throughout the run. For these conditions
the collection efficiency of Chromosorb 102 for CHC1- was 100 percent.
At various time intervals the pump was stopped and the gas jar disconnected.
Aliquots (1 ml) were removed from the gas jar and analyzed for CHC1, by
gas chromatography (GC). The first detection of CHC1. in the gas jar
was taken as the stare in the test for the determination of breakthrough
volume, i.e. the volume of air necessary to purge the adsorbed vapour
through the cartridge.
From the breakthrough volume of the air-solute mixture of known
concentration and the weight of sorbent in the cartridge it was possible
to calculate the adsorptive capacity of Chromosorb 102 for CHCl^ u-»der the
given experimental conditions.
From the data determined with respect to the adsorptive capacity
of Chromosorb 102 for CHC1- and the knowledge that a four hour sampling
period at a flow rate of 250 ml/min would be required in the field tests,
-------
215
sampling cartridges of the following dimensions were constructed: 11 mm
i.d. x 11.5 cm long containing 6 g of Chromosorb 102. These cartridges
were used to determine desorption efficiency data.
quantity CHC1, from Chromosorb 102
Desorption efficiency = -i-- , - • - - — : - rr-i - Z — : —
K 7 CHC1_ concentration Volume of air
in air sampled sampled
The same test assembly as described previously was used to collect
CHC1 on the cartridge. The adsorbed CHCl^ was thermally desorbed into
an evacuated gas jar. The adsorbent tube and the gas jar were connected by
Teflon tubing. The tube was heated to a fixed temperature by wrapping
with heating tape controlled by a variable transformer (170 C) . After
reaching maximum temperature, the stopcock of the gas jar between the jar
and the tube was opened. Heating of the sample tube was continued for a
further 15 minutes, after which time the stopcock of the gas jar was
closed, the gas jar removed from the sample tube, taken outside the
laboratory and allowed to fill up with the cleanest possible air. The
contents of the gas jar were then subjected to GC analysis. Aliquots
(1 ml) were removed from the gas jar using an air-tight syringe and
injected into a gas chromatograph.
Gas chromatography was conducted on a Varian 1200 series
chromatograph equipped with an electron capture detector. The GC
parameters used are shown below.
Co]un1n - 2 m x 2 mm SS containing Chromosorb
102 (80/100 mesh)
Column temperature - 180 C
Injector temperature - 215 C
Detector temperature - 215 C
Detector " EC
Flow rate - "2 "31 ml/mln
Range and attenuation - as required
Chart speed - as required
The concentration of solutes present in the gas jars following
desorption was determined by measuring peak areas of the vapours and
comparing them with those of prepared standards. The initial concentration
-------
216
In the Tedlar bag was known, and, from the volume of air drawn through
the cartridge, the actual concentration contacting the cartridge was
calculated. The desorption efficiency was then calculated.
H.6 Results and Discussion
Experiments using toluene and xylene as absorbents showed that a
considerable loss of both solvent and chlorinated organic compounds occurred
after a four hour test. The solvent was found in all impingers following
the one in which the standard solution was placed. Estimated solvent
losses were kO-6Q%. Losses of CC1, , CHC1- and TCE ranged from 50 to 90%.
With decane as the adsorbent, the solvent loss was less than 10%.
However, the loss of chlorinated compounds was 60 to 100%. Due to time
restraints and the obvious problem of solvent or pollutant loss at the
high gas flow rate, this method of collection using absorbents was rejected.
As discussed in Section H.4.1.2, retention of volatile halocarbons
by the adsorbent at high gas volume flow rates was poor and the break-
through rapid. The results for the collection system finally selected
(Section H.5) are summarized below.
Adsorptive Capacity
Component yl/g (sorbent)
CHC1. 0.1
Desorption Efficiency
Tedlar bag concentration Desorption Efficiency
_ of CHC13 _ _ | _
1.25 ppb 90
96
5.0 ppb 92
HO
10.0 ppb 9^
102
(The higher concentration samples required a second thermal desorption
in order to remove all the solute from the sorbent. However, the first
desorption was always better than
-------
217
Two additional experiments were performed. They were:
a) An air-solute mixture of CC1, and DCE (5.00 ppb level;
40 I Tedlar bag) was run through the system using a back-up
gas jar in order to test whether these solutes were retained
by the Chromosorb 102. Regular checks of the gas jar using
GC to analyze for the presence of CC1, and/or DCE failed to
show any trace. Thermal desorption of the tube followed by
GC analysis gave good recoveries for CC1, and DCE (>90%).
b) An air-solute mixture of CHC1,, CC1, and DCE (2.5 ppb level;
^»0 H Tedlar bag) was prepared. This mixture was drawn
through the test system which contained two additional traps
prior to the Chromosorb 102 cartridge. One trap contained
water (50 ml) and the other 5% NaOH solution (80 ml). These
traps were included in the ORF train in order to remove
residual Cl? and/or HC1. Thermal desorption of the tube
contents gave good recoveries of the solutes (>90%).
H. 7 Conclusions
The laboratory study undertaken showed that low molecular weight
halocarbons cannot be efficiently collected by absorbents or adsorbents
from a gas stream at high flow rates (^ 0.5 cfm).
There are numerous literature references to the technique for
collecting trace quantities of component vapours by drawing air through
cartridges containing sorbent media such as polymer beads, activated carbons
or stationary liquid phases chemically bonded to solid supports [H.5].
This report has demonstrated such a procedure for the collection
of low molecular weight chlorinated hydrocarbons on Chromosorb 102. It
has been shown to give quantitative collection together with quantitative
recovery of the trapped species using a thermal desorption technique.
The sampling tube technique appears to give better results for
sampling traces of low molecular weight organochlorine pollutants in air
than impinger methods.
-------
218
REFERENCES
H.I Marine Environmental Monitoring of Vulcanus Research Burn II,
December 2-10, 197*»» Preliminary report, U.S. Environmental
Protection Agency, December 10,
H.2 Pelllzarri, E.D., Bunch, J.E., and Carpenter, B.H., Env. Sci.
Technology
-------
APPENDIX I
GC/MS/COMPUTER DETERMINATION OF CHLORINATED HYDROCARBONS AND PCB'S
Chemistry Division
Air Pollution Control Directorate
Environmental Protection Service
Environment Canada
Submitted by
Dr. R.C. Lao
-------
221
APPENDIX I
GC/MS/COMPUTER DETERMINATION OF CHLORINATED HYDROCARBONS AND PCB's
In February 19705 30 extracted samples were received from
Dr. G. Thomas of the Ontario Research Foundation. These samples were
taken from the burning of chlorinated hydrocarbon wastes in a cement kiln.
Requests were made to determine the nature and quantity of any heavy
chlorinated hydrocarbons, in particular PCB's, present in the samples.
Experiments were completed and the procedures are briefly
described as follows:
Material: All apparatus and reagents used in the experiment were examined
by gas chromatograph-flame ionization detector (GC-FID) analysis of a pure
hexane extract to ensure freedom from organochlorine residue contamination.
The chlorinated isomers (PCB's) and Aroclor mixtures were obtained from
Analabs, North Haven, Conn., U.S.A.
GC-FID; A Perkin-Elmer 990 model GC-FID with a datasystem PGP-1 was used.
The procedures for standard calibrations were given in a previous paper
(to be published as a chapter in the book "Advances in Dynamic Mass
Spectrometry" 1976). Chromatograph operating parameters were as follows:
Column 12' x 1/8" O.D. stainless steel
Column packing 6fc Apiezon L on chromosorb W 80/100 mesh
Column temperature initial 100°C programmed at *»°C/min.
to final temp. 200°C and hold
Injection temperature 250 C
Manifold temperature 250 C
Carrier gas Helium *»0 ml/min.
GC/MS/Computer: A Finnigan 1015 D GC/MS system was used with a data system
6000 series. Its performance has been studied and documented in the same
paper as mentioned above. It also includes standard PCB mass spectra and
computed reconstructed Aroclor chromatograms. The instrumental data
are as follows:
-------
222
Finnlgan 1015D GC-MS Instrumental Data
A
Instrumental Data GLC
Column 12' x i" all glass
Column packing 6% Apiezon L 80/100
mesh chromosorb W
Column temperature 225 C
Injection temperature 250 C
Carrier gas helium kO ml/min.
Sample size 5 to 7 microlitres per injection
B
Instrumental Data Ms
Filament current 100 microamperes
Electron energy 70 or 20 eV
"3 2
Operating pressure 6.68 x 10 N/m
(5 x 10"6 torr)
Scanning speed 4 seconds
Standard deviation of
spectra maximum 570
After the quadruple HS operating parameters were adjusted the
sample was injected Into the GC. The ion abundance chromatogram of the
GC effluent was acquired by scanning the mass range (40 to kQQ). The
dialogue required for mass spectrometer control, data acquisition and
obtaining the plot are given by software programs. At the end of the
GC run the computer plots a reconstructed gas chromatogram (ion abundance
chromatogram) of total ion amplitude versus the spectrum number.
Identification of these chromatographic peaks can be accomplished by
plotting the mass spectrum of a specified peak or by a limited mass
plot chromatogram which is obtained under computer control by searching
through the collected spectra and identifying spectra containing ions
with a specific m/e_ value.
-------
223
Result and Discussion
(All chroma tograms and computer reconstructed chromatograms are
kept on file at the Air Pollution Technology Centre, Department of the
Environment, Ottawa, Canada.)
Computer reconstructed gas chromatograms of Aroclor 12^2, 125**,
and 1260 were made. By focusing on specific mass/charge ratio (m/e)
peaks such as 290 (tetra-chloro biphenyls), 32A (penta-chloro) or 358
(hexa-chloro), an ion current plot for a particular PCB isomer versus
the Aroclor spectrum is obtained.
Gas chromatograms on Perkin Elmer 990 GC were done for some
blank solutions and concentrated samples (all samples were concentrated
from about 25 ml to 1 ml). No chlorinated hydrocarbons nor PCB's were
found in the samples. If they are present the concentrations are below
the sensitivity limit of GC-FID of 1 ng or less.
Mass spectra for the GC peaks of spectrum numbers 3&5 and ^7
from the reconstructed chromatogram plot of sample (WBC-T3, EPA train,
Impinger) reveals that only hydrocarbons are present in the sample. No
evidence for chlorinated compounds was found. A limit mass searching
technique has failed to give any indication of the presence of PCB's.
The computer library search of standard Aroclors did not match those
obtained from the sample.
It was concluded, therefore, that there were no heavy chlorinated
hydrocarbons or PCB's in the samples. If they were present the weights were
less than 1 ng.
yo 1545
SVJ-147c
------- |