RESEARCH REPORT
CHEMICAL COMPOSITION OF PARTICULATE
AIR POLLUTANTS FROM FOSSIL-FUEL
COMBUSTION SOURCES
(Contract No. EHSD 71-29)
to
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
March 1, 1973
O Batteiie
-------
BATTELLE'S COLUMBUS LABORATORIES comprises the origi-
nal research center of an international organization devoted to research
and development.
Battellc is frequently described as a "bridge" between science and
industry- a role it has performed in more than 90 countries. It
conducts research encompassing virtually all facets of science and its
application. It also undertakes programs in fundamental research and
education.
Battelle-Columbus ~ with its staff of 2500 - serves industry and
government through contract research. It pursues:
• research embracing the physical and life sciences, engi-
neering; and selected social sciences
• design and development of materials., products, processes,
and systems
• informatbn analysis, socioeconomic and technical eco-
nomic' studies, and management planning research.
505 KING AVENUE * COLUMBUS, OHIO 43201
-------
FINAL REPORT
on
CHEMICAL COMPOSITION OF PARTICULATE
AIR POLLUTANTS FROM FOSSIL-FUEL
COMBUSTION SOURCES
(Contract No. EHSD 71-29)
to
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
March 1, 1973
by
L. J. Hillenbrand, R. B. Engdahl, and R. E. Barrett
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
EPA Review Notice
This report has been reviewed by the EPA and approved for publication.
Approval does not signify that the contents necessarily reflect the
views and policies of the Environmental Protection Agency nor does
mention of trade names or commercial products constitute endorsement
of recommendations for use.
-------
SUBJECT:
FROM:
TO:
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
DATE: 3/29/73
CSL-RB
I"
Distribution
Enclosed is the final report from EPA Contract No.
EHSD 71-29, "Chemical Composition of Particulate Air Pollutants
from Fossil-Fuel Combustion Sources" by Battelle Memorial
Institute's Columbus Laboratories. The contents represent a good
Initial look at the fate of chemical components of parti oil ate
matter within a sampling train.
D.r Efrifce Harris
Project Officer
Process Measurements Section
Research Branch
EPA Form 1320—6 [11-71)
APR 12 1973 *
-------
TABLE 07 CONTEI7TC
Page
ABSTRACT vi
Section
I. EXECUTIVE SUMMARY 1-1
OBJECTIVES -2
SCOPE -2
SUMMARY AND CONCLUSIONS -4
Sampling Apparatus and Procedures -1
Overall Particulate Compositions and Material
Balances -10
Sulfur Compounds -13
SUGGESTED FUTURE WORK -14
II. BACKGROUND II-l
GAS AND PARTICULATE COMPOSITIONS -1
Origins of Particulate Matter -4
Inorganic Particulate -5
Nucleated Material From Supersaturated Vapors . -7
Material Added by Physical or Chemical
Adsorption -11
Condensibles Originating by Chemical Reaction . -16
Conclusions Pertaining to Particulate Formation
and Reactions in Flue Gases -20
Station 1: Ahead of Electrostatic Precipitator. -20
Station 2: After Electrostatic Precipitator . . -21
Station 3: External Filter -at 220 F -22
Station 4: Plume of Impinger -22
Sampling Train -22
III. EXPERIMENTAL PROGRAM III-l
RESEARCH PLAN -1
Phase I. State-of-the-Art Report -1
Phase II. Plan Research Program -1
Phase III. Execute the Test Program -2
-------
TABLE OF CCOTENTS (Cont'd)
Section Page
III. (Cont'd)
Experimental Equipment and Methods. III-4
Multifuel Furnace Facility -4
Design Features -4
The Multifuel Furnace -4
Simulated Boiler/Economizer Section .... -7
Electrostatic Precipitator -7
Stack -8
Operational Characteristics -8
Analytical Methods -13
C, H, N Analyses -13
Wet Chemical Analysis for S -13
Goks«5yr-Ross Procedure for S03 and S0a, .... -13
Experimental Details and Results -15
Initial Trials -15
Experimentation Using Multifuel Furnace,
Coal-Fired Operation. -24
Field Experiments -39
Effect of Drying Conditions -39
Experiments at Edgewater Power Station. . . -41
IV. DISCUSSION OF RESULTS IV-1
OVERALL COMPISITIONS AND MATERIAL BALANCES -1
SAMPLING APPARATUS AND PROCEDURES -14
Probe -14
Filter -15
Impinger -17
Sampling Train Modifications -17
Simultaneous Sampling -17
Use of the G/R Coil in Particulate Sampling . . -18
Field Trials -19
Sample Handling . .• -21
Suggested Sampling Procedures -23
Inorganic Ash (less S) -25
Sulfur Compounds -27
V. RECOMMENDATIONS FOR FUTURE WORK V-l
VI. ACKNOWLEDGEMENTS VI-1
ii
-------
TABLE OF CONTENTS (Cont'd)
APPENDICES
APPENDIX A
CONVERSION OF SO., to H2S04
APPENDIX B
WET CHEMISTRY METHODS FOR SULFUR COMPOUNDS
APPENDIX C
FILTER EFFICIENCY MEASUREMENT
APPENDIX D
FIELD TEST SAMPLES FROM OTHER EPA CONTRACTORS
APPENDIX E
KINETICS OF S02 OXIDATION
APPEM)IX F
REFERENCES
LIST OF FIGURES
Figure 1. Locations at Which Collected Samples May Reflect
Particulate-Forming Reactions II-3
Figure 2. Typical Distribution of Deposits in Oil-Fired
Boilers II-6
Figure 3. Relation of Dewpoint and S03 Content of Combustion
Gases to Sulfur Content of Oil 11-10
Figure 4. Variation of Dewpoint With HaS04 Content for Gases
Having Different Water-Vapor Contents 11-11
Figure 5. Dewpoint as a Function of HSO Concentration. . . . 11-12
2 ^
Figure 6. L°g10 Versus Loglo P/K for Simple Languir Model . . 11-14
Figure 7. EPA Particulate Sampling Train 11-23
Figure 8. Schematic Representation of Multifuel Furnace
System III-5
Figure 9. Battelle Multifuel Furnace III-6
Figure 10. Time-Temperature Profile Typical of Oil Firing . . . III-9
Figure 11. Time-Temperature Profile Typical of Coal Firing. . . Ill-10
iii
-------
TABLE OF CONTENTS (Cont'd)
FIGURES (Cont'd)
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figi
12. Optical Emission Spectrography .
13. Schematic Representation of OES Principle. . . .
14. X-ray Emission Spectrometry (Fluorescence, XRF).
15. Schematic Representation of XRF Principle. . . .
16. Goksjiyr-Ross Coil
16.1. Goksoyr-Ross Procedure
17. Schematic View of Sample Port P-l
19.
20
18. Flow Chart of Analytical Procedure for Particulate
Samples from Coal-Fired Operation of Laboratory
Furnace
Schematic Representation of the Experimental
Apparatus for Investigation of S0g and SO .
Flow Sheet for Procedures in Treatment and
• Analyses of Stack Emission Samples ....
1C. Sketch of Sampling Arrangements
2C. Particle Size Distribution
Ill-11
III-ll
III-12
111-12
III-14
III-16
III-17
111-27
111-36
111-54
C-2
C-3
LIST OF TABLES
Table A. Study of the Influence of Sampling Procedures on the
Chemical Composition of Particulate (Flow Sheet for
the Iterative Program) 1-5
Table 1. Typical Exhaust-Gas Composition From Coal-Fired
Boiler H-2
Table 2. Initial Trials, Results from Oil-Fired Laboratory
Burner 111-18
Table 3. Comparison of Glass and Quartz Filter Materials 111-20
Table 4. OES Analyses (j; 50%) for Impurities in Glass and
Quartz Filter Materials 111-21
Table 5. Effect of Drying Treatments Upon Apparent Mass of
Particulate Samples 111-23
Table 6. Coal Analyses in Multifuel Furnace Facility 111-24
Table 7. Plan for a Series of Experiments Particulate
Collections from the Laboratory Coal-Fired
Furnace System III-26
iv
-------
- TABLE OF CONTENTS (ContV)
TABLES (Cont'd)
Page
Table 8. Analytical Data for Particulate Collection from
Coal-Fired Laboratory Burner 111-29
Table 9. Comparison of C, H, N Contents Following Extraction
of Filter Portions with Benzene and Aqueous Acid . . . . Ill-33
Table 10. Oxidation of SO During Sampling of Stack Gases
from the Multifuel Furnace Under Various Conditions. . . 111-37
Table 11. Effect of Drying Conditions on Sample Weights 111-40
Table 12. Investigation of Oxidation State for Sulfur Oxides
Retained in Several Variations of the Particulate
Train 111-42
Table 13. Comparison of the Sulfur Contents of Impinger Water
Stored in Different Ways 111-45
Table 14A. General Data for Runs 49, 50, 53, 55, 56 111-47
Table 14B. Distribution of Chemical Elements Among the Portions
of Particulate Catch With a Double. Filter 111-48
Table 14C. Distribution of Chemical Elements Among the Portions
of Particulate Catch With a Single Filter
Arrangement . 111-50
Table 15. General Sample Descriptions, Treatments, and Analytical
Methods for Characterization of Stack Samples Taken at
Lorain, Ohio 111-55
Table 16. Derived Quantities and Trial Material Balances IV-2
Table 17. Summary of C, H, N Analyses for Particulate Samples. . . IV-7
Table 18. Trial Material Balances for Particulate Samples,
Edgewater Power Station IV-9
Table 19. C, H, N Analyses of Dry Material Samples from
Edgewater Power Station IV-11
Table 20. Impinger and Probe Catch for Single and Double
Filter Experiments, Lorain IV-22
Table 21. Comparison of Calculated Solubilities for S03
in Water with Those Found in Run 41 IV-28
Table 22. Estimated Concentrations for Sulfite Species in
Aqueous Solutions IV-29
Table 1C. Comparison of Quartz-Fiber Filter and MSA 1106B
Glass Fiber Filters C-l
Table ID. Chemical Analyses of Fractions of 4 Stack Samplers . . . D-2
-------
ABSTRACT
The abatement of sources of particulates emitted tc the itnbienu
air is requiring a greatly increasing national effort to try to achieve
•the established standards of ambient air quality. The results of t'riis
effort cannot be achieved most economically if reliable methods are not
available to measure the particulate emissions. The project was conducted
to determine the effect of effluent conditions and of sampling techniques
on the composition and amount of particulates measured in the gases emitted
from oil and coal-fired combustion sources. These effects are becoming
increasingly important as the efficiency of practical collection devices
approaches 100 percent. Then any extraneous effects on the accuracy of
particle measurements loom rapidly in importance whereas, heretofore, these
effects were relatively small in measurements taken on collection equipment
of low or moderate efficiency.
Since no absolute or referee methods for collection of valid
samples of particulate exist for this program, the EPA sampling train
chosen as the best available method for particulate collection. An
investigation was made of the changes in particulate catch produced by
modification of that sampling apparatus and in sample-handling procedures.
Such comparisons were made using single-point sampling procedures instead
of conventional traversing since emission testing was not a direct objec-
tive of this work. For this reason, the comparisons of individual results
on the basis of loading per unit volume of gas are not necessarily mean-
ingful.
This project began with particulate measurements in the gases
from a laboratory furnace capable of burning either residual oil or pul-
verized coal. The sampling train used was the EPA train consisting of a
heated probe, heated filter, and chilled -impingers. After the chemical
characteristics of the particulate samples had been carefully examined
under these laboratory conditions the sampling train and some modifications
of it were applied to the gases from a pulverized-coal-fired power plant.
For the evaluation of control technologies, any alteration in
the amount or composition of materials collected in the sampling train
may result in erroneous conclusions regarding the process operation and/or
chemistry and is therefore undesirable. Experimental work was conducted
to determine the effects of materials of construction, filter-media effi-
ciency, sample drying and handling procedures, and sampling component tempera-
tures on the mass and chemical composition of the collected material. In all
cases, an alteration of 1 percent or greater was considered as a guideline
for the maximum change acceptable for control technology development studies.
After a series of screening experiments, it was concluded that the major
problems were associated with alterations in the form of sulfur compounds
and the location of their collection in the sampling equipment. More de-
tailed studies were then focused on the sulfur oxides and their effect on
the collected sample.
The principal conclusions of this research revolve around the
problems of sulfur oxides in the flue gases.
vi
-------
If the sulfur content of the fuel is low, i-:y ~c~;.<~.. 0.5 percent,
then the concentration of SC2 and SO produced by coab^:. Lion is relatively
low and sample alteration from acid condensation and reccuion may be small.
However, where many practical fuels contain six or more tines that much
sulfur.(3 percent sulfur and more), alterations in the mass reactions can
be large. About 10 percent of the total SO in the gas stream was found
to be retained by the impingers. Laboratory and field experiments carried
out under this program indicate that some common conditions of impinger
sample collection and handling can produce a significant increase in the
material collected'by oxidation of a minor part of that absorbed SO .
Another potential source of sample alteration can occur in the
probe and filter because of the effect of SO in raising the dew point of
the gases. The dew point of flue gases from moderate and high-sulfur fuels
ranges from 250 to 300 F. In these experiments a filter operated at 250 F
caught over 41 percent of the total sulfur caught. Operated at 400 F, the
filter caught only 8 to 24 percent of the total sulfur caught. Thus, if it
is necessary to determine the actual distribution of sulfur species in the
flue gas, it is recommended that the filter be operated well above 250 F.
The S03 (HgS04) which is also potential particulate in the flue gases can
then be measured separately by the Shell-Thornton or equivalent method.
Usually about 1 or 2 percent of the sulfur oxide in the gases is present
as sulfur trioxide (or H SO ) in the 'typical moist gases at or near the
acid dew point.
Regardless of the type and configuration of probe-filter combi-
nation used in particulate sampling, there is always a possibility that
the filter will fail and lose part of the sample. When testing high-effi-
ciency equipment, such a loss may cause serious error but be visually
undetectable. Since measurement of particulate after high-efficiency
devices requires sampling times of several hours or more, it appears
reasonable to include a backup filter as a precaution against having to
discard the entire test at a later time when the weighing results are
known. While impingers may provide such a backup, their efficiency for
submicron particles is questionable and the chemical reactions found to
occur will complicate the data. A second heated filter as backup will
incur less uncertainty.
vii
-------
I. EXECUTIVE SUMMARY
OBJECTIVES
SCOPE
SUMMARY AND CONCLUSIONS
Sampling Apparatus and Procedure
Overall Particulate Composition
and Material Balances
Sulfur Compounds
SUGGESTED FUTURE WORK
-------
1-1
CHEMICAL COMPOSITION OF PARTICULATE AIR POLLUTANTS
FROM FOSSIL-FUEL COMBUSTION SOURCES
by
L. J. Hillenbrand, R. B. Engdahl, and R. E. Barrett
I. EXECUTIVE SUMMARY
In the current national effort to improve air quality, one of
the most difficult tasks is to reduce the level of suspended particulates
in the densely populated metropolitan and industrial areas. A 1975 pri-
mary goal for particulates in the atmosphere has been set at only 75 y,g
grams per cubic meter, annual geometic mean. Many urban areas now have a
geometric mean level well over 100 y.g. These measured levels include all
suspended particulates, liquid or solid, as collected by glass-fiber fil-
ters oh high-volume air samplers. A major effort is underway to abate
most U. S. man-made sources of these suspended particulates.
In this program of abatement of these myriad particulate sources
it is important to be able to measure their emissions accurately so that
the magnitude and rate of improvement can be reliably gaged, and the measure-
ment should logically include those emitted substances which will become
ambient particulates when measured at ambient air conditions.
However, while the high-volume ambient air sampler indiscriminately
samples suspended solids and liquid droplets as particulates in the atmo-
sphere, the abatement of their sources requires that the type of particulate
can be identified and not just lumped together. There are two reasons for
this. In the first place not all particle collectors will collect all types
of particles. Thus, different methods of abatement may be required for
various particulates. Also, if liquid and solid particulates are combined
during the process of sampling, there is the possibility that chemical
reactions will occur which could affect the accuracy of the measurement.
Hence, this project has been directed toward examination of the effect of
the methods used in sampling particulates from fossil-fuel-fired sources
on the chemical composition of those particulates..
BATTELLE — COLUMBUS
-------
1-2
OBJECTIVES
(1) To examine the chemical interactions that occur
during sampling and which influence the estimates
of particulate emissions from fossil-fuel-fired
combustion sources.
(2) To detect the influence that the operation of
abatement equipment such as electrostatic precip-
itators for particulate .control has on the nature
of the particulate sampled and emitted.
(3) To evaluate techniques for measurement of partic-
ulate loadings in the inlet and outlet streams
from such control equipment.
SCOPE
Historically, EPA has proposed that particulate matter be defined
as all solids and !those condensible materials which are liquid at standard con-
ditions of 1 atmosphere and 70 F, excepting uncombined water. Present
methods for sampling and weighing particulate emission do not necessarily
reflect particulate levels as thus defined. For many years the predominant
accepted methods for measuring fly ash used only the dry filtration or a
combination of dry filtration and wet impingement. The former fails to
collect condensible materials that may exist as vapors at the filter tem-
perature but may immediately become liquid particulates at ambient tempera-
tures, and the liquid impingers used on some particulate trains may generate
particulate by chemical reaction in the impingers, thus producing erroneous
data on particulate emission levels.
Also, the knowledge of how conditions before and after particu-
late emission control devices (such as electrostatic precipitators and wet
scrubbers) may influence particulate measurements is incomplete. As the
application of these control devices is becoming more widespread, it is
necessary that their influence on particulate emission measurements be
established by evaluating present or proposed sampling methods on operating
units. Hence, new or modified sampling methods may be required to produce
a measure of particulate loading that is consistent with the above definition
for the evaluation of particulate emission control devices.
BATTELLE — COLUMBUS
-------
1-3
Since no absolute referee methods for collection of valid sam-
ples of particulate exist for this program, a commercially available version
of the EPA sampling train was chosen as the best available method for par-
ticulate collection and an investigation was made of the changes in partic-
ulate catch produced by modifications of that sampling apparatus and in
the sample-handling procedures. Such comparisons were made using single-
point sampling procedures instead of conventional traversing since quanti-
tative emission testing was not a direct objective of this work. For this
reason the comparisons of individual results on the basis of loading per
unit volume of gas are not necessarily meaningful.
Experiments were first run sampling gases from the laboratory
furnace to determine the chemical composition of particulates caught in
the various parts of the sampling train. Then the effect of temperature
on that sample composition was explored. Different filter materials were
tried because variations in chemical composition of glass filters caused
serious uncertainties during analysis of the filter catch. A novel quartz
filter material was tried and adapted to the requirements of this program.
The sampling apparatus was then taken to a coal-fired power plant
where the effect of sulfur oxides on particulate composition was investi-
gated. The influences of sample-handling procedures on sample accuracy
were also studied.
Most of the sampling was done on coal-fired operations because
most particulate control equipment installations are on coal-fired plants.
The research nature of the objectives made it possible to conduct a major
part of the sampling program using the Battelle Multifuel Furnace burning
either residual oil or coal. Samples collected from this furnace were used
to acquire background for further sampling operation under the more diffi-
cult field sampling conditions at a power plant.
As work continually underscored the problems associated
with SO oxidation and HgS04 condensation from wet combustion gases,
special attention was given to this aspect of the problem. Although
it must be recognized that S0g dissolved in water mist in the atmosphere
BATTELLE — COLUMBUS
-------
1-4
also oxidizes to create sulfate particulate, and the SO combines with
3
water to form acid mist, the need to measure them separately still re-
mains .
Reliable field methods for handling SO./SC) mixtures in the
e 3
presence of alkaline materials are needed before the special problems
associated with scrubbers can be investigated properly only in the field
For this reason, the investigation of such equipment was delayed and
has not been included in this phase of the investigation.
The effects of electrostatic precipitators on particulate were
examined both in the gases from the Battelle Multifuel Furnace and in the
field at a coal-fired power plant, the Edgewater Station, Lorain, Ohio.
• The evolution of sampling and sample-handling procedures has
included trials of a number of variations that were planned to help assess
the importance of problematic aspects as they were recognized. •' These
trials have been used to evolve a concept for improved procedures that may
be useful in specifying the emissions that characterize fossil-fuel com-
bustion in modern practice.
Table A summarizes the steps followed in the iterative investi-
gation of sampling techniques.
SUMMARY AND CONCLUSION
Major problems associated with the sampling and measurement of
particulate emissions from fossil-fuel-fired plants either to determine
flue gas composition or to evaluate the performance of particulate control
apparatus are:
(1) Generation of sulfate by oxidation of S0a
in the impingers. This is also a potential
problem in the probe and filter,
(2) Condensation of H S04 or adsorption of SO
in the probe or filter.
BATTELLE — COLUMBUS
-------
1-5
TABLE A. STUDY OF THE INFLUENCE OF SAMPLING PROCEDURES OK THE CHEMICAL COMPOSITION OF PARTICULATE
SHEET FOR THE ITERATIVE PROGRAM)
Initial Trials,
011-Flred
BATTEL1E MULTIR'El FURNACE
First Series,
Coal-Fired
Second Series,
Coal-Fired
Field Trials
Edgewater Power Station
Coal-Fired
Items
Investigated
... ..—. probe material and .—. . ....... . ..... Probe Catch
temperature IV-5, 1 to 15
Tables 7, 8
filter material and temperature effects; --—— single and double filters; Filter Catch
temperature comparative data for temperature IV-5, IS to 16
Table 2 quartz Tables 12, 14
Tables 7. 8
—. acidity effects; 4x'Vet" acidity effects; 4x'Vet" Impinger Catch
impingers In series impingers in series IV-17
Table 10 Table 12
----------------- coll used without comparison to Impinger before Impingers; comparison G/R (Shell Coil Use
frit; before and catch, in parallel to impInger catch IV-17-20
after filter; Table 10, Figure 19 Table 12
acidity removal
Tables 7, 8
„.-------------- distribution among comparison of C/R coll comparison of G/R coll and Sulfur Oxides
portions of catch; and Impinger catches; Impinger catches IV-17 to 24, 27-33
acidity and tempera- acidity and oxidation Table 12
ture effects effects
Table 8 Table 10
filter catch eompo- probe, filter, and ..-..-.-_..-.---_.-- probe, filter, and Impinger Ash Composition
sitlon Impinger portions - portions - weight and compo- IV-25 to 27
Table 2 weight and composition sitlon ,
Table 8 Table 14
water, catch; drying analytical accounting analytical accounting of Total Catch
conditions of total catch - total catch - drying con- IV-1 to 14
Table 5 uncertainties ditions
Table 8 Tables 11,12,14
XRF and OES of filter
catch; sample drying
Tables 2, 5
C,H,N, and OES, XRF
analyses of aliquots,
•11 portions
Table 7,8, Figure 18
individual analyses for
4 Impingers in series;
organic and aqueous
washings
Table 10. Figure 19
aliquot collection OES and
XRF analyses; all portions,
dry, aqueous, and organic
layers
Table 11,12,13,14,15
Figure 20
Sampling Procedure
and Analysis
IV-17 to 24
(3) Fractionation of the composite of ash,
sulfur oxides, and organic components
during passage of the particulate through
the sampling apparatus.
(4) Retention of water, especially in those
portions of the catch that include free
sulfuric acid.
(5) Probable loss of identity of organic
materials that are collected simulta-
neously with the sulfuric acid.
BATTELLE — C O L U M B U
-------
1-6
The first two of these have received most of the attention in
this program. The work done is regarded as the first phase of the in-
vestigation that is needed to establish the effects of control apparatus
on particulate concentration levels and composition. Subsequent phases
should be directed toward the research and areas discussed in the section
entitled, "Recommendations for Future Work".
The particulate is identified as a mixture of
(1) Fly ash: solid particulate, mainly minerals,
partially reacted with the sulfur
and oxygen in the stack gas,
(2) Sulfur: present as free sulfuric acid, absorbed
and adsorbed, and combined as sulfate
and sulfite,
(3) Organic material: condensates probably ranging
from volatile materials to high-molec-
ular-weight materials such as the
benzpyrenes and metal-organic complexes,
(4) Water: combined as hydrate water in the in-
organic catch and water retained by
the free sulfuric acid,
and the experiments were planned as an investigation of the importance of
each and the problems that they contribute to the measurements of flue gas
components.
In attempting to make material balances of the metal and sulfur
in compounds in the flue gases, that is, to account for all of the catch
by chemical analysis, uncertainties of sampling and analyses were revealed
in spite of the special attention given to such problems in this program.
In turn, these uncertainties are a reflection of the limitations that should
be recognized in comparing any two sets of data. The uncertainties are ex-
perienced especially in assessing the metal analyses for the impinger por-
tions. Here, the greatest mass of the impinger catch is due to sulfur
oxides and only very small amounts of individual metal compounds are found.
Collectively, these metal compounds appear to account for 1 or 2 percent
of the total impinger catch, but such estimates are best regarded as maximum
estimates rather than typical values.
BATTGLLE — COLUMBUS
-------
1-7
Because the small amounts of sample and the lor concentrations
of most elements in both blank and sample, there are large analytical un-
certainties in these impinger portions. The elemental analyses for these
portions were obtained as an expression of the levels of concentration
found for each element and were intended to reveal any tendencies for the
heavier (toxic) metals to be carried through the filter as a result of
their chemical forms. Since both blank and sample are subject to large
analytical uncertainties in these impinger portions the analyses are given,
except for a few of the major elements, without correction for blank and
as such, represent maximum values that may be assigned to each element.
Probe and filter analyses are in general not subject to this same degree
of uncertainty except where trace amounts are reported.
The work that has been done falls roughly into the subject areas
• Sampling apparatus and procedures
• Overall particulate and compositions
• Sulfur oxide investigations
and the conclusions are grouped under the above headings.
Sampling Apparatus and Procedures
1. Three filter materials were considered: a silver filter,
MSA Glass 1106 BH, and Tissuquartz QAO 2500. Only one
run was made using the silver filter because large areas
of visible stain on the filter after the run indicated that
the anticipated reaction of the silver with the HgS04 va-
pors had indeed occurred. At the usual moist stack condi-
tions SO does not exist as a gas but is largely hyd rated
3
chemically to H2S04 vapor. In most cases the separation
of the filter catch from the filter itself is not practical.
Therefore, the chemical analyses of the catch have highest
precision and sensitivity as the filter material becomes
more precisely defined. The silver filter is best from
this standpoint because it can be dried to constant weight
BATTELLE — COLUMBUS
-------
1-8
easily and impurity elements are of little consequence
in most cases. However, the reactivity of the silver
with sulfur compounds led to a preference for the pure
and relatively inert quartz filter material.
2. The filter temperature appears to affect significantly
the composition of the filter catch. Thus, at 400 F
the filter retained substantially less sulfur oxides
than a comparable filter at 250 F. This is no surprise
if one recalls earlier research showing that the acid
dewpoint of fossil-fuel flue gases is often lower than
300 F. The greater amount of sulfur oxides found on the
cooler filter is interpreted to mean: a. that a consid-
erable portion of the sulfur collected on that filter
might have resulted from b'oth condensation and reaction
of the particulate with the SO and SO,, and b.:that the
3 **
condensation and consequent reaction is favored at the lower
temperature. Oxidation of S0g is estimated to be minor.
3. The impingers also retain on the order of 10 percent or
more of the total S0a as measured by the G/R coil* pro-
cedure; so that accurate sulfate estimates from impinger
catches require inhibition of S03 oxidation during
collection,, sample storage and evaporation, and analysis.
The experimental results appear to require use of a sep-
arate train, such as the G/R train, for accurate sulfur
oxide measurement rather than analyzing the impinger
liquid for this purpose.
A. The G/R procedure is easy to use and provides reproducible
results. EPA has supported a demonstration of the absolute
accuracy of the method for stack gas conditions. (Walden Labs.
Contract No. CPA 22-69-95, Final Report, June, 1971.)
* (Goksjiyr-Ross; also called Shell Method).
(1) References in Appendix F.
BATTELLE — COLUMBUS
-------
1-9
5. The G/R coil inserted between filter and impingers collected
a substantial portion of the total H_SO. found in that run
3 4
and was far more effective than a cooler second filter tried
in its place in one experiment. In the experiment with the
coil, if it is assumed to be as efficient as has been claimed,
the sulfate found in the impingers of that run would seem
to have been formed by oxidation in the impingers.
6. When two filters were used in series, the mass loading of
the second filter was largely due to sulfur oxides retained
there. The mass of fly-ash oxides on the second filter,
excluding sulfur oxides, was small and the analyses for
individual elements were subject to large uncertainties.
7. The high chromium and nickel contents of the catches, col-
lected in the metal probes present evidence of the accumu-
lation of corrosion products in the probe portion and, to
some degree, in the filter portion at 250 F. Sulfur oxide
retention by the metal probes was slightly higher than with
a quartz probe, which may be a reflection of the increased
reactivity of the catch due to those corrosion products.
8. On the basis of the observations to date, a sampling pro-
cedure for examination of tbe performance of particulate
control equipment might be as follows:
(a) Use a glass probe and quartz filter media with
probe and filter at 400 F or higher to minimize
condensation of SO or reactions of sulfur dioxide
in the sampling system.
(b) Use a backup quartz filter media at 400 F or higher
to prevent loss of sample if the first filter
fails.
1ATTELLE — COLUMBUS
-------
I-10
(c) Determine total S03 and S03 in the gas phase by
a procedure, such as that of Gokstfyr-Ross, operated
simultaneously with the particulate collection.
(d) Use special collection trains for components
such as PNA and certain hazardous metals present
in relatively small concentrations. These
trains would be adapted to the most suitable
absorption or reaction condition for efficient
collection of the component concerned.
(e) Attempt material balance by the use of a suffi-
cient scope of analytical methods so that the
entire mass of catch is indentified according
to its chemical characteristics.
Overall Particulate Compositions and Material Balances
9. The assignment of definite oxide compositions to each of
the chemical elements (found analytically) makes possible
trial material balances and these balances show substantial
deviation from 100 percent in some cases. The uncertainties
may be due to the assumption of a definite oxide composition.
However, it is more probable that the uncertainties arise
from sampling and sample-handling procedures as much as
from the inherent analytical difficulties.
10. The ratios of abundance of a pair of elements frequently
are not the same in all portions of a catch, and these
variations in abundance probably reflect differences in
composition for particles of different size, as well as
the presence of unburned coal in the particle and
reactions with the gas phase as the mixture cools.
11. Analytical results show a varying Si/Al ratio for probe
samples, probably because the Si and Al are contributed
by both fly ash and coke having different Si/Al ratios.
BATTELLE - COLUMBUS
-------
I-11
The experimental Si/Al ratios for pairs of probe
samples taken near a precipitator varied with the
location as shown
Run
49
50
55
56
Date
10/6/71
10/6/71
10/8/71
10/8/71
Location
E.P.
E.P.
E.P.
E.P.
inlet
outlet
inlet
outlet
Si/Al Mass Ratio
3.
20.
0.
2.
0
0
31
1
These ratios were substantially greater for samples col-
lected at the precipitator outlet than for those collected
at the inlet.
12. For the samples obtained from firing coals, only the
elements Na, K, Mg, Al, Ca, Fe, S, and presumably 0 and
Si are present in the dry catch in significant amounts
relative to the total mass of the particulate.
13. A number of heavy elements are present in amounts that
are insignificant compared to the total mass of parti-
culate, but which may need to be considered individually
for their toxicity contributions.
The total mass of elements found by chemical analysis,
but excluding sulfur, indicates that the precipitator at
the Edgewater Power Station removed 99.5 percent or better
of the relatively coarse material caught in the probe and
about 96 to 97 percent of the filter-catch portion.
BATTELLE — COLUMBUS
-------
1-12
14. Substantial percentages of both probe and filter catch
material are composed of sulfur compounds and these per-
centages are greater in the precipitator outlet streams
than in the inlet streams. The reasons for this are not
yet determined.
15. The inorganic ash portion is distributed differently among
probe, filter, and impinger than is the sulfuric oxide portion.
In the runs made using the Battelle Multifuel Furnace, less
than 2 percent of the ash was found by chemical analysis in
the impinger, whereas between 6 and 85 percent of the sulfur
oxide catch was found there. Other results obtained from
samples collected at Edgewater Power Station show substan-
tially smaller percentages of ash in the impingers. Thus,
the mass distribution of the catch usually does not corre-
spond to the distribution of any single component of the
particulate.
16. The composition of the total catch varies too much to per-
mit calculation of meaningful averages, but typical compo-
sitions for coal-burning operation based on these trial
applications of the EPA train and its modifications would
be:
Battelle Multifuel
Furnace, precipita-
tor tnlet(a)
Full-Scale Power Plant
Precipitator Precipitator
tnlet outlet
Inorganic ash
(less sulfur),
wt percent
Sulfur, as
HaSO*, wt per-
cent
Carbon (coke or
organics), wt
percent
66
31
72
12 to 18
9 to 17
37
63
not
determined
(a) Average for all runs, dry basis.
BATTEULE — COLUMBUS
-------
1-13
Thus, on the downstream side of precipitators, the sulfate can
constitute a major portion of the material collected in the probe, on the
filter, and in the impingers.
Sulfur Compounds
A considerable part of the experimentation was devoted to exami-
nation of the influence that the apparatus and conditions used for sampling
can have on the sulfur oxide content of the catch. For this purpose, the
S02 contents of samples were examined for several experimental sampling
arrangements. In the temperature range below 400 F,essentially all S03
exists in wet-gas mixtures as U^SO^ and the terms S03 and H3S04 are used
interchangeably. Analytical determinations were made of S0g, sulfate, and
total sulfur (as sulfate) where appropriate. The conclusions were as follows:
17. The sulfur contents of the probe and filter portions
were variable and it was not possible to estimate how
much of the sulfur found in these portions was due to
reaction of the collected solid with H SO condensed
2 *k
from the gas phase or to S02 physically or chemically ad-
sorbed during sampling. The total sulfur is probably a
combination of sulfur oxide retention by such reactions
plus sulfur that had already become incorporated into the
particulate prior to sampling.
18. A relatively large amount of S02 is dissolved and re-
tained in the impingers and the oxidation of only a
minor fraction of that S02 would account for the sulfate
formation. Kinetic calculations indicate that S02
oxidation in the impinger catches may be more serious
during the typical periods of storage and analysis of
the samples than is the oxidation during the 40-minute
sampling time.
19. The formation of sulfate in impinger samples by oxidation
of S0a was indicated by the observations that (a) when
extra wet impingers were added to the sampling train, as
much sulfate was found in the third and fourth impingers
as in the first two; (b) more sulfate was found in the
BATTELLE — COLUMBUS
-------
1-14
impingers than in a 6/R coil operated simultaneously;
(c) considerable sulfate was found in the impingers
operated in series with a G/R coil.
20. The sulfate analyses on any stored water sample collected
in the field should be made on aliquots stored in completely
filled bottles to avoid oxidation during storage and trans-
portation of samples. Flushing of the sample with inert
gas before sealing for storage is recommended.
21. As expected, pH has an important bearing on the process in
the impingers. Increase of impinger pH to 5.5 increased
both the BOS retained and the S03 found. The rise in S03
is believed good evidence for oxidation of the increased
amounts of S02 retained in the impingers. Lowering of pH
to 1.1 to 1.5 caused a decrease in S02 retained and the
anticipated decrease in S0o was not found. Further work is
3
needed on this matter. Approximate calculations of S0a
solubility in strong acid at 0.1 N concentration indicate
that total sulfite concentrations may be reduced by about
-one-half as compared with that for HgO solvent.
SUGGESTED FUTURE WORK
Areas for further research all relate to the observation that some
fractionation of particulate can occur in the particulate-control apparatus
as well as in the furnace and boiler itself, and studies will be needed to
establish a basis for measurement of control for each of the important
classes of material. Among these areas for further research are
(1) Practical means for field sampling and
analysis of total S03 content
(2) Sampling and analysis of small particles
that may penetrate the filter
(3) Sampling and analysis of anions, especially
halides, nitrates, and carbonates
BATTELLE — COLUMBUS
-------
1-15
(4) Study of the chemical state of selected
heavy metals and identification of sampling
difficulties for each, especially for Hg,
Ba, Sb, Pb, As
(5) Study of special sampling conditions that
are generated by scrubber-type particulate-
control apparatus
(6) Demonstration of a valid sampling method for
the evaluation of particulate-control-equip-
ment efficiencies for specific effluent-gas
components.
Conceivably, these studies will identify methods to supplement
the existing sampling train, and these must be tried under comparative
conditions in industrial applications where particulate and emission-con-
trol equipment is used. In this connection, the special problems generated
by the use of scrubber-type control apparatus should be studied immediately
so that the observations may be used to guide the other research areas
discussed here.
BATTELUE — COLUMBUS
-------
II. BACKGROUND
GAS AND PARTICULATE COMPOSITIONS
Origins of Particulate Matter
Inorganic Particulate
Nucleated Material From Supersaturated Vapors
Material Added by Physical or Chemical Adsorption
Condensibles Originating by Chemical Reaction
Conclusions Pertaining to Particulate Formation and
Reactions in Flue Gases
Station 1: Ahead of Electrostatic Precipitator
Station 2: After Electrostatic Precipitator or
Stack
Station 3: External Filter at 220 F
Station 4: Plume of Impinger
Sampling Train
-------
II-l
II. BACKGROUND
GAS AND PARTICULATE COMPOSITIONS
The inclusion of both solid and liquid materials in the defini-
tion of particulates (page 1-2) makes necessary a careful consideration of
the origins of the particulate matter that is collected and of the influ-
ences that the sampling methods may have on this material. Of especial concern
are those particulates that may form by chemical reaction and condensation
either in the boiler and stack portions of the power plant (so that the total
collection may vary as a function of sampling location) or in the sampling
train during particulate collection. All these become even more significant
as trends in the control of pollutant emissions reduce the upper limits of
permissible quantities of material.
In order to consider in more detail the chemical and physical
interactions that may occur between the gaseous and particulate components
of the exhaust gases, it is convenient to select for quantitative examination
a typical gas composition that may be produced by a power plant, as shown in
Table 1. A similar set of concentrations can be described for an oil-fired
plant. The major difference for oil firing would be a substantially higher
H20/C02 ratio and lower fly^ash loadings. However, the above gas composition
is sufficient for consideration of reactions that might form particulates.
Several combinations of sampling points and sampling train con-
figurations should also be included in consideration of particulate-forming
reactions. Figure 1 shows three sampling points that should be considered.
These points are
Point 1. Ahead of the electrostatic precipi-
tator (assumed gas temperature of
400 F)
Point 2. After the electrostatic precipitator
(assumed gas temperature of 300 F)
Point 3. In the plume after dilution to 70 F.
BATTELLE — COLUMBUS
-------
11-2
TABLE 1. TYPICAL EXHAUST-GAS COMPOSITION
FROM COAL-FIRED BOILER
Concentration
Component
HgO
COa
Fly ash before
precipitator
Fly ash after
precipitator
NO
S02
S03
Hydrocarbons
Volume
percent
4.0
15.0
0.050
0.20
0.0030
(30 ppm)
(0.0010)
,3(a)
g/m3
30
273
9.16
(4 gr/ft3)
0.458
(0.2 gr/ft3)
0.63
5.3
0.10
<0.1
Relative
Importance 0>)
0.02
0.002
--
--
0.8
0.1
4.5
>50.
(a) At 70 F, 1 atmosphere.
Values show approximate percentage of the total amount
of each component in the flue gases that, if collected
as particulate would increase the measured loading by
1 percent (based on fly ash after the precipitator.
BATTELLE — COLUMBUS
-------
II-3
Dilution
Flue gas
-400 F
Electrostatic
precipitator
70 F
FIGURE I. LOCATIONS AT WHICH COLLECTED SAMPLES MAY REFLECT
PARTICULATE-FORMING REACTIONS
-------
11-4
Obviously, sampling the plume at 70 F (Point 3) is impractical, although
the proposed definition of particulate (1-2) includes any condensed particles
at this point.
Primary variables in the sampling train that might affect reac-
tions of interest include filter temperature (either at gas temperature
with an in-stack filter or heated to 220 F) and the presence of an impinger
in the train.
The following sections consider chemical processes that may
cause the formation of particulate matter and include a discussion of the
state of the matter existing at various sampling points and of the reac-
tions that might occur on filters and on impingers.
Origins of Particulate Matter
Particulate matter can be discussed in four classes according
to source and sequence of information:
(A) Solid or molten particles, largely inor-
ganic, that existed in the hot flue gases
at the time of entry into the stack par-
ticulate-control equipment
(B) Nucleated material from supersaturated
vapors in the gradually cooling flue
gases
(C) Material added by physical or chemical
sorption
(D) Condensible material that originates
through further chemical reaction, whether
in the stack or in the sampling train, and
which adds onto or modifies the existing
particulate.
BATTELLE — COLUMBUS
-------
II-5
(A) Inorganic Particulate, Solid or Molten
Existing in the Hot Stack Gases
By the time the electrostatic precipitator is reached in the
combustion flow system, most of the condensible inorganic components
have already formed particulate materials. As the particulate production
is judged primarily on a weight basis, further interactions of the
various metal oxides with each other are of little concern so long as
the particle size is in a range to permit efficient collection.
Of more concern are those compounds formed by interaction of
the particulate metal oxides with the nonparticulate acidic gases present.
Thus, for example, one milligram of Fe203 in the stack gases may become
2.5 mg of Fe2(S04)3 by reaction with- sulfur oxides present in the hot
gases. Possible chemical interactions are
Fe303 + 3S03 + 3/2 02-«Fes(S04)3
or
Fe203 + 3H2S04-|Fe2(S04)3 + 3H20.
Similar reactions can be written for the other metal oxide components of
the fly ash.
The actual reaction process is presumably a complex of simul-
taneous competitive reaction to form sulfates and water from oxides, free
acid, and S03 . The reactions and products would vary with temperature.
This reaction process is well illustrated by the investigation of Jacklin,
Anderson, and Thompson' ' into the composition of fireside deposits in oil-
fired boilers.
Figure 2 shows the average deposit analysis for each of the five
boiler sections studied by Jacklin, et al.' ' The parallel rise of Fe 0,
and SO. in the final three boiler sections is in large part due to the chem-
J
ical interaction of HgSO. with the iron (oxide) surfaces beginning well above
560 F. The relatively large concentration of S03 in all the boiler sections
BATTEL.LE — COLUMBUS
-------
II-6
Q)
1L'
s
w
o
If,
Q)
Q.
X
UJ
0)
CO
I
il
,c
£
Convection! Economizer
FIGURE 2. TYPICAL DISTRIBUTION OF DEPOSITS IN OIL-
FIRED BOILERS (3)
-------
11-7
Indicates that the chemical retention of S03 is important in deposits even at
temperatures well above the dew point of the aqueous HgS04 mixture present in
the stack gases. In such boiler systems, the bulk of the stack gases have
been cooled 2000 F in a travel time of 6 sec or less^ and the occurrence
of supersaturation conditions for various components of the stack gases is
a driving force for reactions of the sort to form particulate.
For present purposes, any process that may influence the weight
of particulate by 1 percent or more is worthy of consideration. In this
regard, the "Relative Importance" column of Table 1 is of interest. Thus,
the column indicates that the retention in the sampling rig of only 0.1
percent of the S02 as sulfite would be important gravimetrically, similarly
0.002 percent of the C02 and 0.8 percent of the NOX would be important if
retained as acids or as salts of those acids. Thus, the reaction of ex-
tremely small quantities of these gases could contribute significantly in
particulate measurements. Hence, in experiments and analyses to predict
the possibilities of contribution by these acids to particulate measure-
ments, a high level of accuracy is needed.
Sudden cooling of the gases can generate nonequilibrium con-
ditions that can lead to formation of salts from these gases. These salts
may or may not persist during sample collection, depending on conditions.
(B) Nucleated Material From Supersaturated Vapors
in the Gradually Cooling Flue Gases
The formation of droplets from supersaturated vapors can be a
significant source of particulates. The amount so formed depends upon the
rate and mechanism of formation. The various processes are discussed here.
(A)
Following the treatment of Green and Lane, the self-nucleation
and growth of a droplet of condensible from a supersaturated vapor depends
upon the effects of the droplet size on its vapor pressure. In the absence
of an ion or dust particle to serve as the nucleus, grouping of near-neighbor
molecules of condensibles can serve as nuclei. These groupings arise during
B/VTTELLE — COLUMBUS
-------
II-8
the statistical fluctuations of position of the individual molecules in
motion in the gas phase. The groupings formed in this way become stable
droplets if the existing partial pressure, P, and the droplet radius, r,
have the relationship p 2 M
lnpoT = — »
RTpr
where Poo is the vapor pressure of the condensible over a flat surface, y
is the surface tension, M the molecular weight, and p the density of the
droplet material. This relationship leads to the important concept of a
critical droplet size of radius rc. All droplets with rrc have lower volatility and grow further. Thus, formation of droplets,
either by self-nucleation or by condensation-on prexisting particles, re-
quires that nuclei of the critical size or larger exist in the gas stream.
As the gas stream cools Poo decreases rapidly and, with a given partial
pressure of condensible, the critical radius becomes smaller so that more
droplets are created. However, during sampling, the rate of formation of
condensed material will depend on the number and size of nuclei remaining
in each portion of the sampling train.
Because the amount of H3S04 present is relatively large, the
nucleation of aqueous HgS04 droplets below the dew point is of most interest.
First of all, equilibrium calculations for the reaction S0g(g) + HgS04(g)
indicate that essentially all S0g is converted by this reaction to HgSO^
(gas) in the flue gas at 400 F and below. The calculations for this equili-
brium are given in Appendix A.
If the flue gas mixture contains 4 x 10~3 mole fraction water
and 3 x 10 mole fraction H2S04 formed by the above equilibrium reaction,
total condensation of the water will yield a liquid that contains 0.075
mole percent H2S04 in water. The liquid corresponds to approximately 0.04
M H2S04, The work of Parkinson^ ' indicates that this liquid also would
be capable of dissolving a minor amount of SOg. For 0.058 M H2SO*, which
is somewhat more concentrated than the acid assumed here, the solubility
at 70 F is about 3 x 10~* lb moles S02/ft3 acid. In metric terms this is
31 mg S0a/l of acid solution.
BATTELt-E — COLUMBUS
-------
II-9
The dew points of aqueous H2S04 vapors of the sort considered
here have been determined and summarized by several investigators. Some
(3)
of the following earlier data were summarized in 1963 by Miller, et al.
The following data for a number of earlier workers is summarized from that
report.
Data published by Rendle and Wilsdon on the relation of the
S03 content of the combustion gases and of the dew point of the gases to
the sulfur content of fuel oils are presented in Figure 3. Results of
several other investigations also have been plotted. Consequently, the
type of oil, ash content, and combustion conditions differ for the various
sets of points. Although the plot of S03 content shows considerable
scatter, it is apparent that with more than 0.5 percent sulfur in the oil,
the S03 content of the gases does not increase in direct proportion to
the increase in sulfur.
The dew-point plot brings out two important points (1) there
is a rapid initial rise in dew point with the first increment of sulfur
in the fuel. For an estimated dew point of 100 F with no sulfur (the
water dew point), an increase to 260 F (HgSO^. dew point) is found with
1 percent sulfur; (2) there is a relatively small rise in dew point as
. the sulfur in the fuel oil increases from 1 to 6 percent. Even if an
economical method of removing sulfur from oil were available, a reduction
from 6 to 1 percent would lower the dew point only from 300 to 260 F. It
would be necessary to achieve almost complete removal of the sulfur to
obtain a significant drop in dew point.
Large changes in the water-vapor content of flue gases cause
only slight changes in acid dew point. The variation of dew point with
sulfuric acid content of gases having different water-vapor concentra-
tions is shown in Figure 4, where the range from 0.5 to 15 percent water
vapor changes as the dew point changes only 30 to 40 F for the medium-to-
high acid contents indicated.
BATTELLE — COLUMBUS
-------
11-10
0
E
"o
.0 0.004
*„
0
100
0
X
x^
X"
X
8
/
/x
r
* /
/
/
'
V
X^
2*--*;
vx.
X X
•
°^i
X v
"
'
VAX""
X^^-v
X
v-
N^ ^
^^^^"
IX ,
X
X X
S Corbett and Fireday(7)
® Flint (8)
$ Taylor and Lewis ^
All other points Rendle and
Wilsdc
>n(6)
x^^»
s**^**
\T7 m
234
Sulfur Content of Oil ,%
FIGURE 3. RELATION OF DEW POINT AND SO3 CONTENT
OF COMBUSTION GASES TO SULFUR
CONTENT OF
-------
11-11
c
6
r.-.
I
500
4CO
300
200-,
0.000 0.005 0.010 0.015
FIGURE 4. VARIATION OF DEW POINT
WITH H2S04 CONTENT FOR
GASES HAVING DIFFERENT
WATER-VAPOR CONTENTS(10)
H
in Dry Gas, %
More recently, Lisle and Sensenbaugh^ ' summarized the existing data and
made additional experiments to verify the dew-point data. Figure 5 is a
summary of the results obtained by them using the method of Goksciyr-Ross. (D
These data are quoted and used as the bases for ASME PTC 19.10.
(C) Material Added by Physical
or Chemical Adsorption
In the above considerations, only supersaturated vapors are in-
volved. The situation becomes more complex when the processes of physical
and chemical adsorption are considered. The distinction between the two
is usually made by considering physical adsorption to pertain to processes
with energies of adsorption less than some critical value, e.g., 10 kcal/
mple, and chemical adsorption is represented by all processes with energies
of adsorption greater than the critical value. Because of the lower energies
BATTELLE — COLUMBUS
-------
11-12
3t\J
350
330
310
g" 290
M
2 270
o 250
230
210
ion
A
p
r.
0 El
-
-
•
-
•
^
1 111
A. Taylor dewpoirV
Muller calculated
cperimental partia
pressure measure
S
5
i iii
o X^
^
1 III
f meter
Ii
•nents
i
i ,
£f /
1
1
1
I
l
1
1
1
1
i iii
-
-
•
•
-
-
•
-
-
).OI 0.1 1.0 10 100
S03(H2S04) in Flue Gas, ppm
FIGURE 5.
DEW POINT AS A
FUNCTION OF
HSS04 CONCEN-
TRATION
involved, physical adsorption is pressure dependent and may be described
in terms of a pressure ratio, P/Pg<1.0, where P is the existing partial
pressure of the vapor and Pg is the saturation vapor pressure needed to
cause condensation of the vapor at the temperature concerned.
In the Brunauer, Emmett, and Teller (BET) theory of multilayer
physical adsorption)1' ' a model is developed based on the Langmuir ad-
sorption isotherm in which the volume of gas, v, adsorbed by pressure, P,
is related to the saturation vapor pressure of the gas, Ps by
__J* c-1 p
v(Ps-P) v^c"
where vm is the unit volume of adsorbed gas equivalent to a single mole-
cular layer on the adsorbing surface, and where, to a first approximation,
c = exp [(q! - qL) RT]
where qx is the heat of adsorption of the first layer and qL is the latent
BATTELLE — COLUMBUS
-------
11-13
heat of liquefaction of the adsorbate. Commonly, the volume of adsorbed
gas approaches vm at value of P/PS<1, and depends on the value of the ex-
cess heat of adsorption for the first layer as described by the constant
c. For example, in the physical adsorption of water vapor at ordinary
temperatures on inert organic surfaces, we might anticipate qx to be 14
k cal and qi - qi_ «* 4- kcal. Adsorption of this sort leads to monolayer
adsorption at about P/PS « 0.4 at 27 C or of the order 40 percent of satu-
ration.
Other treatments of adsorption^ ' serve to model the effects of
pressure at small degrees of surface coverage and indicate the P/PS values
of 10" or less may be needed to limit adsorption to about 1 percent of
the available surface. This persistence of adsorption at low fractions
of saturation pressure can result in considerable error in the estimation
of condensible components of the stack gases if the possibilities for
retention of such materials at the filter are not considered during the
processing and particularly the drying of samples.
Chemical Adsorption. The process of chemical adsorption is con-
veniently approached through use of a model, described by Hobson(13) for
. the effect of partial pressure and adsorption energy on the extent of ad-
sorption. Thus, in Figure 6, the fractional coverage of the surface, 6, is
described in terms of a Langmuir model of uniform surface for which the de-
sorption energy, E, for the adsorbed material is constant. The parameter
K is a function of temperature, T (K) and molecular weight, M, of the
adsorbing gas and, following Hobson, is equal to K = 1.76 x 104(MT)J/3. The
dotted lines in Figure 6 show how the surface coverage for a gas of molecu-
lar weight 28 (N2) would be expected to vary as a function of partial
pressure and energy of surface interaction (E/RT). The higher values of
E/RT, e.g., :? 16, correspond to our definition of chemical-adsorption pro-
cesses at room temperature and the range of E/RT shown covers the full range
of adsorption processes including the transition betwen physical and chem-
ical adsorption.
BATTELUE — COLUMBUS
-------
11-14
E/RT= 100
CD
g
CF>
o
26 22 -18 14 10 -6 ' -2 0 2
FIGURE 6. LOG,00 VERSUS LOGI0 P/K FOR SIMPLE
LANGMUIR MODEL
Inspection of some available adsorption data indicate that the
heats of adsorption of S02, ESSO^, and such vapors on oxides, such as fly
ash, may be anticipated to fall generally in the interval 5 to 15 kcal.
In that event, the BET model shows that monolayer coverage of all surfaces
by these vapors at room temperature occurs at vapor concentrations equal
to a few tenths of saturation. Similarly, from Hobson, fractional coverage
of a few percent of the surfaces by adsorbed vapors occurs at P/K of the
~™ 3
order of 10 , or vapor concentrations of a few parts per million at 1 atmo-
sphere. These two estimates establish the limiting concentrations of each
gas-mixture component that can participate in appreciable adsorption at the
temperature chosen. At higher temperatures the pressures required would be
somewhat higher, but the implication of high surface coverages at vapor
pressures well below saturation remains valid. Two questions are of interest:
(1) Assuming complete monolayer adsorption on
all particulate and filter surfaces, how
much is the apparent weight of particulate
altered?
BATTELLE — COLUMBUS
-------
11-15
(2) As the adsorbed layer will determine the
chemical reactivity of the particulate
emitted, what can be said about the pro-
bable nature of that adsorbed layer?
The surface area of the particulate can be estimated from
S = surface area, m3/g
g -f = density g/cc
S = —-7 , where d = particle diameter,
^* micrometers, y.
so that, for a material of density = 2.0, the particulate before the pre
cipitator would have surface of 6 6 n nn a ,
For the performance assumed in Table 1, the effect of the precipi-
tator on particulate surface area remaining in the stack gas is illustrated
by the following:
' Surface Area
m3 /g m3 /ma
particulate before precipitator, 10(i, 9.16 g/m3 = 0.30 2.74
particulate after precipitator, 0.5u, 0.458 g/m3 = 6.0 2.74 .
Note that although the precipitator removed 95 percent
of the mass and reduced the particle size twenty-fold, essentially none of the
particulate surface was removed. As the chemical importance of the partic-
ulate is apt to be related to the amount of reactive surface present, this
may be an important observation. The 2.74 ms of surface/m3 stack gas can
adsorb approximately 2.74 x 1 x 10~5 moles of vapor of low molecular weight,
e.g., H20. This would be 2.74 x 10~5 x 18 = 49.5 x Kf5, or approximately
5 x 10~ g/m gas. As the particulate weight is 0.458 g/m3, the adsorption
is negligible on a weight basis, even if monolayer amounts of each vapor
were retained.
However, the adsorption capacity of 2.74 x 10" moles/m gas is
sufficient to retain a substantial fraction of the trace organic materials
that may be present. For this reason, high-surface-area portion of the
sampling train, e.g., the particulate and the filter medium, must be examined
for such ingredients. Thus for BaP at 25 (J-g/lOOOm , a partial pressure of
P/Pg ~ 10 8 might easily produce surface coverage of 0.1 percent. This
BATTELLE — COLUMBUS
-------
11-16
would amount to 10~a moles per m3 gas compared to the 10" moles /m3 actually
found in the gas. Thus, far more of BaP or similar ingredients may be re-
tained by adsorption on the filter than would be found in the effluent gases.
In considering the chemical nature of the adsorbed films that can
form in this way, it is important that the simultaneous physical adsorption
of two or more components of the gas mixture at a surface be favorable to
chemical reaction between the adsorbed components. Thus, possibilities
exist for chemical change in stack gases during sampling. The considerations
presented here indicate that high concentrations of minor, flue-gas compo-
nents may exist on particle surfaces. Surface influences may be recognizable
experimentally by using widely different temperature conditions at the point
where the solid particulates are expected to be collected, or by doping the
stack gases to alter ambient concentrations of individual components.
(D) Condensible Material That Originates Through
Further Chemical Reaction, Either in the
Stack or in the Sampling Train
Several reactions by which S03 might react to form condensible
material, either in the stack or in the sampling train, include
Reaction 1. SOtg oxidizing to S03 at stack tempera-
tures in contact with fly ash having
catalytic properties
Reaction 2. S03 oxidizing and being retained as
S04""2 on surfaces of oxide particulate
Reaction 3. SQs absorbing on a liquid-phase water
droplet or in the impinger solution,
and oxidizing by reaction with dissolved
Considering Reaction 1, the oxidation of SO^ to S03 in the presence
of Og is rapid on V305 or Pt surfaces at about 900 F, and most kinetic treat-
ments are complicated by the need to consider the diffusion transport limi-
BATTEULE — COLUMBUS
-------
II-17
tations in the pores of the catalyst. Two approaches have been used to
estimate the rate of oxidation at stack temperature. To maximize the
estimate of the extent of reaction, the particulate may be imagined to
be retained for 1 hour on a filter in the stack gas. In addition, the
following assumptions may be made
• There are no diffusion limitations
• The fly ash is as good a catalyst as
the commercial material but the fil-
ter (or fly ash) is exposed to the
reactant gases at 400 F instead of the
usual temperature range for catalytic
oxidation of S02.
The two treatments used for the rate of oxidation of S02 to S03 in the re-
gion of kinetic control are:
(14)
According to Calderbankv '
/"moles SOp converted^ _ /-31.000 „ n"\ % ' PSOa
r \ g catalyst-sec ) ~ 6Xp \^T^ + l2'01} ' ~p ^
.75)
Pso3
where P is in atmospheres and T = °K.
According to Kodles, et al.J '
K
where T is °K, N'sare mole fractions, and K = equilibrium constant for the
conversion of S02 to S03 at temperature, T. For this evaluation a maximum
rate was calculated with the Kodles equation by putting Nc_ = 0.
oUa
BATTEULE — C O L U tVl Q U S
-------
11-18
Both estimates show 10~6g SOS or less oxidized to S03 in the
given mixture in 1 hour time per m of gas flowing. If the filter was
exposed to gas at 1 cu ft/min, then about 1.7 m3 gas would be exposed to
the filter in 1 hour. Thus, the total S02 oxidized, <2 x 10~bg, is negli-
gible compared to either 0.45gof the fly ash or the S03 already present.
The high- temperature catalysts used for commercial oxidation of S03 do not
retain sufficient activity at stack temperature to be capable of appreciable
oxidation of the S02 in the assumed gas mixture. Therefore, the quantity
of condensible formed by this reaction should not be significant in the
presence of fly ash.
A qualitative approach to Reaction 2 can be made through the
published data for the DAP-Mn process in which manganese oxide is used
as an absorbent which, in the presence of Os forms manganese sulfate. Thus-,
MnQx . iHgO + SOg + ^ (2-x)02-» MnS04 + iHgO.
In this process, x is about 1.5 to 1.8 and i is 0.1 to 1.0. The oxide has
been found to have a high affinity for S03 in the temperature range 100 -
180 C (i.e., 212 to 356 F). In that range, at least, the effect of tempera-
ture on the rate of absorption process is negligible. The authors find
that the results can be expressed by In (Yo/Y) = kmxn • t, where
Yo = concentration of S08 in inlet gas, vol. percent
Y = concentration of SOa in outlet gas, vol. percent
m = concentration of absorbent, gMnOjj/m3 gas
t = residence time of gas, sec
and then suggest suitable values of n and k are -1.5 x 10~* , respectively.
On this basis, assuming that particulate is all as active as this MnOx and
using the estimate (see Table 1) that a 0.1 percent retention of SOa would
be important,
1.5 x 10"* (1.8)"" (0.458) t
0.001 = 0.31 x 10-* t
t = 30 sec.
BATTELLE — COLUMBUS
-------
11-19
Hence, the adsorption of S0g and fixation as sulfate is conceivably im-
portant in the filter section of the sampling rig, if much of the particle
surface is as active as MnOx.
The contribution of Reaction 3 (the absorption of SO in a water
&
phase followed by oxidation by dissolved 02 to form sulfate) is difficult
to estimate because of the influence of pH on the reactions. Data obtained
under the conditions usually used for investigation of SO absorption would
2
differ greatly from those of the itnpinger. Futhermore, because the reaction
proceeds by a chain mechanism involving radical formation processes in the
water phase, many materials can serve either to accelerate or retard the
reaction. In general, the rate of sulfite oxidation in solution is described
d St g [HSO 3]
where Sfc represents the total concentration of sulfite species and g is
a rate constant that includes contributions by various free radical pro-
cesses and is dependent on the dissolved 03 concentration. In the impinger,
the constant supply of dissolved S02 from the flowing gases and the effect
of dissolved acids generate conditions radically different from those used
experimentally in the investigation of sulfite oxidation. First- order rate
constants for the oxidation by dissolved oxygen from normal air show
half-times for reaction of about 200 seconds at 35 C for well-stirred sul-
fite solutions in neutral and mildly alkaline solution. In these experi-
ments variations in initial pH in the range 6.95 to 8.20 seems to affect
the degree of oxidation (in the absence of replacement S02 of HSO§) but not
the initial rate.
BATTELUE — COLUMBUS
-------
11-20
Conclusions Pertaining to Partlculate
Formation and Reactions in Flue Gases
Our conclusions are stated by considering the several stages of
the flue-gas handling system as analogous to sampling stations from which
particulate is sampled and characterized. These stations are
Station 1: In-stack filter aheadxof the electro-
static precipitator (at Sampling Point
1 in Figure 1)
Station 2; In-stack filter after the electrostatic
precipitator (at Sampling Point 2
in Figure 1)
Station 3: External filter maintained at a tempera-
ture of 220 F
Station 4: Diluted plume at 70 F or (theoretically
the impinger).
In practice, these stations are not a true sequence. For example, a par-
ticulate sampling train normally would include a filter at the conditions
of either Station 1, 2, or 3, and may or may not include an impinger (Sta-
tion 4). However, for the purpose of this discussion, the conditions at
each station are presented as consecutive and in the order of decreasing
temperature.
Station 1: In-Stack Filter Ahead of
Electrostatic Precipitator
At this station, the ash components have essentially all con-
densed to form solid particulate. Water vapor has begun to adsorb on all
surfaces including that of the particulate, but the monolayer is quite in-
complete because of the high stack temperatures. The S03 in the stream will
be 98.7 percent converted to H^S04 at this station and the relatively low
concentration of HgSO* formed in this way would be physically adsorbed com-
petetively with the more concentrated HgO vapor. Chemical reaction of
BATTELLE — COLUMBUS
-------
11-21
with the particulate will contribute further to HgSC^ retention so that the
particle surface may hold a relatively strong concentration of sulfate or
A *\
aqueous HgSO^. The particulate provides an estimated 2.8m surface/m gas
and the adsorbed monolayer on this surface will be negligible in mass com-
pared to the mass of the particles themselves. As shown earlier, the par-
ticulate collected on an in-stack filter at this point, if retained there
for 1 hour with 1 cu ft/min flow, would generate a negligible amount of S03
from SO in this sample gas stream, even if the particulate surface equalled
the catalytic action of commercial high- temperature catalytic oxides. The
rapid cooling of hot gases prior to this station may have generated other
salts such as the alkaline earth carbonates, which will tend to be decomposed
as the sulfuric acid accumulates. The degree of this conversion is unknown.
Station 2: In-Stack Filter After the
Electrostatic Precipitator
At this point, H20 adsorption is extensive and a major part of
the surfaces is now covered by adsorbed vapor. The conversion of S03 to
H0S04 is essentially complete at this temperature (about 350 F) and rela-
tively high concentrations may be expected in the adsorbed layers. The
precipitator has removed 95 percent of the particulate mass and reduced
the average size from 10 (J. to 0.5 p., but nearly all of the initial surface
area remains. Nevertheless' the mass of the adsorbed layer is still neg-
ligible . Available data^ ' confirm the expectation that the irreversible
S02 absorption and then oxidation to sulfate by processes such as
(1) MOX • iHgO + SOj, + | (2-X) 03- MS04 + iHgO
must be considered for material retained on filters at these temperatures.
The particulate remaining which leaves the precipitator will serve as a
nuclei for the onset of condensation of the aqueous H3S04 to form droplets
of diameter greater than a critical diameter defined by the degree of super-
saturation of vapor.
BATTELLE — COLUMBUS
-------
11-22
Station 3: External Filter at 220 F
At this station, the temperature has fallen to the point where
most of the H8S04 has condensed to form aqueous liquid, if it has not
already been absorbed by the metal oxide particulate material. Dilution
of the flue gases with air would lower the HgS04 dew point somewhat, but
the equilibrium vapor concentration of RgSC^ at this temperature, 0.1 ppm,
assures that at least 90 percent of the 1^804 would be expected to condense
even at 20:1 dilution ratio. The chances that some SOg may be oxidized
and retained as sulfate on the filter catch cannot be disregarded. For
the cases of COg and NC^, significant formation of salts from these gases
by nonequilibrium reactions of the metal oxide particulate with the humid
gases appears possible.
Station 4: Plume or Impinger
The degree to which SOg, C02s and NO^ contribute to the condens-
ible particulate in the diluted stack gases (plume) is not available by
direct measurement. Calculations indicate that some contributions may
occur in an impinger in a sampling train, but the comparison of the con-
tributions under plume and sampling conditions remains a difficult problem
for experimental investigation that has not been attempted in this program.
Sampling Train
Figure 7 shows the sampling train as used by.EPA. As originally
conceived, its objective was to collect dry solid participates in the cy-
clone and filter and to cool the gases prior to the pump and meter section.
Aside from their ability to collect "condensibles", the impingers also
serve as a "backup" for the filter in case it fails from defect or faulty
installation in the holder, to collect the water from the gases for measure-
ment and to cool the gases prior to the pump and meter section.
BATTEUUE — COLUMBUS
-------
11-23
This is the basic sampling train arrangement used for particulate
sampling in this work except for research variations as noted in the Ex-
perimental Equipment and Methods section.
Probe
Filter
Impingers
1. Stainless steel, buttonhook-type probe tip
2. Stainless steel coupling
3. Probe body, 5/8-inch OD, medium-wall
Pyrex tube logarithmically wound with
25 feet 26 ga. nickel-chromium wire
4. Cyclone and flask
5. Fritted-glass fdter holder
6. Electrically heated enclosed box
7. Ice bath containing four impingers
connected in series
8. The Greenburg-Smith type impinger with
tip removal
9. Second impinger with tip
10. Third impinger with tip removed
11. Fourth impinger with tip removed and
containing approximately 175 grams of
accurately weighed dry silica gel
12. Pressure gauge
13. Check valve
. 14. Flexible-rubber vacuum tubing
15. Vacuum gauge
16. Needle valve
17. Leakless vacuum pump
18. By-pass valve
19. Dry-gas meter
20. Calibrated orifice
21. Draft gauge
22. S-type pilot tube
FIGURE 7. EPA PARTICULATE SAMPLING TRAIN
-------
III. EXPERIMENTAL PROGRAM
RESEARCH PLAN
Phase I. State-of-the-Art Report
Phase II. Plan Research Program
Phase III. Execute the Test Program
Experimental Equipment and Methods
Multifuel Furnace Facility
Design Features
The Multifuel Furnace
Simulated Boiler/Economizer Section
Electrostatic Precipitator
Stack
Operational Characteristics
Analytical Methods
C, H, N Analyses
Wet Chemical Analysis for S
Goks{Syr-Ross Procedure for S03 and S0g
Experimental Details and Results
Initial Trials
Experimentation Using Multifuel Furnace •
Coal-Fired Operation
Field Experiments
Effect of Drying Conditions
Experiments at Edgewater Power Station
-------
III-l
III. EXPERIMENTAL PROGRAM
RESEARCH PLAN
Battelle's experimental approach to this study consists of a
three-phase program:
Phase I. State-of-the-Art Review
A state-of-the-art report has identified the
present knowledge of emissions from fossil-fuel com-
bustion processes and particulate emission measure-
ment methods.
Phase II. Plan and Prepare a Test Program
An experimental plan for the experimental aspects
of the project, including the field program was re-
viewed by EPA. This field program was designed to
provide data on emission measurements as influenced by
sampling-train construction, flue-gas parameters, and
the presence of particulate emission control devices.
Because of the desirability of holding combustion
conditions steady during sampling, a Battelie-owned
Multifuel Furnace Facility was used for most of the in-
vestigation. It is capable of firing fuel oils and coal
at low rates and of generating flue gas reasonably
typical of that produced in power plants. Samples from
this facility can be obtained at lower cost and greater
convenience than can samples from power plants. Hence,
the early stages of this investigation were conducted
using this furnace. The collection of samples from
power plants was better defined and planned because of
lessons learned in preliminary runs with the furnace
at Battelle. Also, these runs produced useful data
in planning details of the field work.
BATTELLE — COLUMBUS
-------
III-2
Unfortunately, accepted absolute or referee methods for such
sampling do not exist. The method used has been to investigate the
causes for changes in particulate catch that were produced by changes in
sampling mode and sample-handling procedures. This experimental method
has required the collection of data from pairs of experiments planned to
demonstrate the effects of individual changes in experimental procedures.
Such comparisons were made using single-point sampling procedures (as
opposed to traverses) to reduce variables and sample-collection time.
Phase III. Execute the Test Program
The iterative procedure for investigation of sampling methods
is summarized in Table A on page 1-5.
a. Initial Trials Using the Laboratory Furnace
A series of samples was collected while several No. 6
residual fuel oils were fired in the laboratory furnace.
Of primary concern in these trials were probe and filter
materials and the importance of filter temperature.
b. Experimentation Using the Multifuel Furnace Facility
As a result of the above trials and a study of potential
reactions that might occur in the sampling system as dis-
cussed in the Background Section of this report, a new
series of trials was designed. These trials were intended
to determine how sampling-train variables affected the
collection of sulfur oxides within the train. The sampling
train for these trials was basically the EPA train with
modifications for specific trials. These trials utilized
the laboratory furnace but with coal-firing operation.
A second series of trials was conducted using the
laboratory furnace to further identify problems arising
from S03 retention and oxidation during sampling and
sample handling. For these trials, the sulfur oxide catch
in the various particulate sampling rigs was compared with
the sulfur oxide content of the stack gas as measured by an
BATTELLE — COLUMBUS
-------
III-3
independent method. The Goksrfyr-Ross method (G/R) was
chosen for this purpose. In planning these trials, vari-
ations in both sampling and sample-handling procedures were
used in the hope that it might be possible to distinguish
among the several ways that S0g oxidation can occur. For
example, the sulfate collected with the particulate was
expected to represent the sum of
(1) SO present in the stack gas
3
(2) S0g generated during sampling
(here 40 minutes)
(3) SO generated in the impinger solutions
O
during storage under air in bottles, and
transported to the laboratory (usually
days or weeks)
(4) S03 generated during sample evaporation
and analysis.
c. Field Experiments
The experience gained in the experimentation with the
Battelle Multifuel Furnace was used to design a series of
trials that were conducted at the coal-fired Edgewater
Power Station, Lorain, Ohio, -and which incorporated the
most useful variations that had been tried in the labora-
tory. The experimentation was divided into two subseries:
one intended to provide as much detail as possible regarding
the amount and oxidation state of the sulfur oxides retained
in the various portions of experimental sampling apparatus,
and the other intended to provide a comparison of the
amount and chemical composition of the particulate collected
from the inlet and exhaust streams of the electrostatic
precipitator.
BATTELLE — COLUMBUS
-------
III-4
Experimental Equipment and Methods
Multifuel Furnace Facility
Design Features. The multifuel furnace used for this program
was designed to generate flue gas and fly ash under conditions closely
simulating those of a power-generation station. This implies combustion
at a high enough temperature with a proper cooling schedule to produce
flue gas and fly ash having physicochemical properties similar to those
of a typical central-station boiler and its associated stack and plume.
In addition, the laboratory-scale system used in this program was flexible
enough to permit firing with either pulverized coal or residual oil.
The Multifuel Furnace Facility was operated with an electro-
static precipitator in firing coals but without an electrostatic precipi-
tator in firing oils. The time-temperature profiles in the other sections
of the facility were appropriately adjusted for the inclusion or exclusion
of the precipitator.
Figure 8 is a schematic representation of the gas combustion
and flue gas conditioning system with major sections and sampling points
•indicated. The major sections of the system are discussed below.
The Multifuel Furnace. Figure 9 is a photograph of the
Battelle Multifuel Furnace which was used on this program. This small-
scale furnace consists of a cylindrical combustion chamber approximately
15 in. in diameter by 80 in. in length. The furnace is lined with
three layers of firebrick and insulation to accommodate surface temper-
atures up to 2900 F. At the outlet, the diameter of the furnace is
reduced to 5 in. to enclose the flame, provide for normal recirculation,
limit radiation losses, and provide sufficient gas velocity to keep fly
ash suspended in the gas stream. Viewports along the axial dimension of
the furnace provide for visual access during periods of adjustment of
firing conditions.
BATTELUE — COLUMBUS
-------
III-5
Plume chamber
Burner^
Air.
Fuel'
Simulated
stacK section
G-3
P-3
P-2
Furnace
G-5
Simulated
boiler section
Electrostatic
precipitator
G-2
r
Exhaust
G = Gas sampling port
P = Particulate sampling port
FIGURE 8. SCHEMATIC REPRESENTATION OF MULTIFUEL FURNACE
SYSTEM
-------
Electrostatic
1 prjecipitator
FIGURE 9. BA.TTELLE MULTIFUEL FURNACE
-------
III-7
In normal operation of the furnace, natural-gas firing was used
to maintain system temperatures at approximately the desired levels on a
more or less continuous basis while runs were not being made. Upon
switching to either coal or oil firing, the entire system was allowed to
re-equilibrate for several hours before any data were taken.
An adjustable-flow positive-displacement pump which was pre-
calibrated was used to regulate the supply of residual oil to the furnace
at about 3 gal/hr, and preheating of the oil was provided to insure the
desired viscosity at the burner nozzle. With coal firing, a dispersion
of the fuel in air was fed to the burner (at a rate of about 30 Ib/hr)
via "a screw feeder mounted within a pressurized coal hopper.
•Simulated Boiler/Economizer Section. The simulated boiler/
economizer section of the rig was constructed of stainless steel pipe
lined with a castable refractory material. Gas velocities within this
section were modulated using a variable exhaust duct at the head of the
section. This operation also helped in maintaining desired temperatures,
and additional control over temperature profiles was gained through the
use of removable insulation on the outside of the section. Gas velocities
in horizontal portions of this section were typically 60 ft/sec, and
velocities in vertical portions were about 5 ft/sec. Temperatures
dropped from about 2600 F at the inlet to about 400 F or less at the
outlet.
Electrostatic Precipitator. An electrostatic precipitator made
by Precipitair was used for removal of fly ash between the boiler and
stack sections of the system for runs made with coal firing. Typically,
the inlet and outlet temperatures of the precipitator were 375 F and
300 F, respectively.
BATTELLE — C O U U M 8 U S
-------
III-8
Stack. The 37-ft-long simulated stack section of the rig
was constructed of 5 in. OD stainless steel pipe. Temperatures in the
stack section could be controlled through the use of heating tape and
insulation applied throughout the length of the section. For runs with
residual oils, the stack temperature was maintained at about 300 F. With
coal firing of the unit, the stack temperature was maintained at 250-
275 F.. These values are typical of central-power-station operation with
these fuels. Sampling points are available at regular intervals along
the stack section.
Operational Characteristics
The operational characteristics of prime interest to this program
concerned the time-temperature profiles obtained .in the system. Typical
time-temperature profiles for oil and coal firing are shown in Figures 10
and 11. Residence times with coal firing were similar to those with oil
firing except for the extra 5-sec residence time in the precipitator.
Temperatures in the system were also similar for both types of fuel
except for the slightly lower stack temperatures which are more typical
of coal-fired power plants.
Analytical Hethods
Figures 12 through 15 shows the methods of analysis used as described
in the following paragraphs. Experimental variations in sample handling
and in preparation for analyses are discussed in the section on the
Experimental Details and Results for each series of trials. Because the
research work was planned as an iterative program, each series of trials
employed techniques that drawn from the experience in the previous
series. For this reason the arrangements for each series of trials and
the results obtained from each are described together. The Discussion
section describes the overall relations and meanings that have been drawn
and which serve as the basis for the Conclusion section.
BATTELUE — COLUMBUS
-------
V
V4
3
4-1
to
M
-------
2800
2400
2000 -
1600 -
0)
1-1
3
4-1
(0
IJ
0)
&
0)
H
1200 -
800 -
400
Boiler/Economizer
/Air-heater
I
t-1
o
FIGURE 11. TIME-TEMPERATURE PROFILE TYPICAL OF COAL FIRING
-------
III-ll
FIGURE 12. OPTICAL EMISSION
SPECTROGRAPHY
(OES)
Battelle's emission spectrographic laboratory is equipped with a 1.5 meter, 25,000 line/inch grating instru-
ment. The spectra are recorded on photographic film in the region 2200A to 9400A and the dispersion in the
second order is 3.5A/mm. Samples for analysis are routinely handled in solid form, as powders and as liquids.
The system can detect some 70 elements qualitatively and quantitatively to a precision of about 3 percent.
However, only about 50 elements are determinable at low concentrations. Detection limits for the detectable
elements are in the range of 1-100 ppm on a bulk analysis or of the order of 10"° to 10"'grams when analyzing
a micro sample.
A strong feature of the emission spectrograph, besides being able to accept materials for analysis in a
variety of forms, is that it provides a very rapid method for qualitative and semiquantitative determination of a
large number of elements. The arc-spark chamber is equipped for handling samples in inert or oxygen-rich atmo-
spheres where useful.
Ancillary to both the optical and X-ray spectrographs are sample preparation facilities including both wet
and dry preparative techniques.
FILM
•MM
READOUT
GRATING
FIGURE 13. SCHEMATIC REPRESENTATION
OF OES PRINCIPLE
-------
III-12
FIGURE 14. X-RAY EMISSION
SPECTROMETRY
(FLUORESCENCE,
XRF)
Battelle's X-ray emission spectrometer is a vacuum instrument capable of detecting and determining ele-
ments between atomic number 12 (Mg) and 92 (U) inclusive. It is equipped with A.D.P., PET and LiF analyzing
crystals, gas flow and scintillation detectors and sample beam collimaters of 0.005" x 4" and 0.02" x 4. The
readout is on strip chart, in counts per second and in seconds per count. It can handle samples in form of chunk,
powder and liquid and is especially useful for rapid quantitative determination of elements in concentration
ranges of 0.01 percent to major amounts. The absolute sensitivity when looking at a single element is of the
order of 10"^ g. The precision of the analyses provided is of the order of 1 percent.
Ancillary to both the optical and X-ray spectrographs are sample preparation facilities including both wet
and dry preparative techniques.
SAMPLE
DETECTOR
FLUORESCENT X-RAY
COLLIMATOR COLLIMATOR
ANALYZING CRYSTAL
X-RAYS
FIGURE 15. SCHEMATIC REPRESENTATION
OF XRF PRINCIPLE
-------
Ill-13
C, H, N Analyses
Carbon, hydrogen, and nitrogen analyses were made on a Perkin-
Elmer Model 240 Elemental Analyzer. This is an automated instrument
combining the classic Pregl and Dumas techniques. The combined analyses
for C, H, and N are obtained on single samples of 1 to 3 mg of materials
ranging from high-melting organic solids to highly volatile liquids.
Wet Chemical Analysis for S
Sulfate sulfur was determined gravimetrically by precipitation
as barium sulfate in a solution which had been acidified and boiled to
expel H3S, S03 .
Total sulfur was determined gravimetrically by precipitation
as barium sulfate in a solution which has been treated with bromine and
nitric acid to convert all sulfur to sulfate. •
The details of procedure for these wet chemical methods are
given in Appendix .B .
Goks^yr-Ross Procedure (G/R) for S03 and SQ^
Figure 16 shows the Goks^yr-Ross coil. The flue gases,
reasonably cleaned of particulate by a glass wool plug, are drawn through
a glass coil and a Grade 4 sintered glass disc, both enclosed in a water-
filled jacket. The water is maintained at a temperature (60-90 C) well
below the acid dew point but is slightly above the water dew point. Practically
all the sulfuric acid condenses on the walls of the coil. Any acid mist
particles left in the gas stream are retained by the sintered glass
filter. The acid is then washed out and determined by titration with
standard sodium hydroxide solution.
BATTEUUE - COLUMBUS
-------
111-14
FIGURE 16. GOKStfYR ROSS COIL
The gas sample flows from right to left. In the exit chamber is visible
a fritted glass disc which stops most of the residual acid mist that
may escape condensation on the coil wall. (Also widely referred to as the
Shell-Thornton coil.)
-------
111-15
The gases leaving the sulfur trioxide collector are then passed
through hydrogen peroxide solution; all the sulfur dioxide is converted
to sulfuric acid which is titrated with standard sodium hydroxide solution.
Figure 16.1 shows the sulfuric acid coil in two applications: in
(a) the coil is in use in the analysis train described by Goksfllyr and
Ross, and in (b) a larger version of the coil is shown in places between
the filter and impingers of an EPA sampling train. The white material
visible in the inlet end of the coil is the sulfuric acid mist formed in
the coil during particulate sampling at Edgewater Power Station (see
Run 51 in Table 12).
Experimental Details and Results
Initial Trials
In these trials, samples were collected from the Multifuel
Furnace Facility during oil-fired operation in order to obtain information
on particulate from residual-oil fuels and to examine several alternatives
in sampling and analyses procedures that were under consideration. The
samples were collected from Port P-l, see Figures 8 and 17, using a Vycor
probe about 8 in. long. Both silver and MSA glass 1106 BH. filter
materials were tried at temperatures of 230 and about 400 F using the
EPA sampling train. The results for these experiments are summarized in
Table 2.
The filter samples were stored over indicating Drierite and
weighed several times. Because of the heterogeneous mixtures of organic
and other condensibles that might collect on the filter, each in small
amount, the most suitable handling of the filter was not known. For
these trials a procedure was employed in which the filter was divided into
halves; one for homogenizing by grinding for elemental analyses by means
of optical emission spectrometry and X-ray fluorescence and the other for
extraction by benzene and dilute sulfuric acid. The amounts of material
ATTELLE — COLUMBUS
-------
111-16
(a)
Sulfuric acid coil in analysis train with HO,, bubblers for S0g
(b)
Sulfuric acid coil adapted for use in EPA train, between filter and impingers
FIGURE 16.1. GOKS0YR-ROSS PROCEDURE
-------
111-17
i
II"
12'
4-1/2" ID
T
-2"OD pipe
Direction of flow
-Vycor
probe
FIGURE 17. SCHEMATIC VIEW OF SAMPLE PORT P-l,
SEE FIGURE 8
-------
TABLE 2. INITIAL TRIALS. RESULTS FROM OIL-FIRED LABORATORY BURNER
Run No.
Date
Materials
Probe
Filter
Oil
Temperatures
Stack
Filter
Impinger
Sample
Cas Vol. , m3
Mass , tng
Probe .-
Filter
Impinger
TOTAL
Analysis of Filter (OES)
Li
B
Na
Mg
Al
Si
K
Ca
Ti
V
Cr
Mn
Fe
Ki
Co
Cu
Sr
Zr
Mo
Sn
Ba
Pb
S (XRF)
V (XRF)
TOTAL
4
4/26
Vycor
Silver
Venezuelan 954
465
230
68
0.173
(a)
NA
22.9
NA
NA
(c)
--
Trace
0.18
0.11
0.68
0.08
__
0.11
Trace
1.20
0.12
..
0.05
0.19
Trace
0.01
__
0.01
Trace
__
Trace
0.02
NA
NA
(d)
0.2
0.2
1,3
0.3
_.
0.2
Trace
2.1
2.0
_-
0.1
0.2
Trace
Trace
__
Trace
Trace
..
Trace
Trace
NA
NA
5 6
4/26 4/27
Vycor Vycqr
MSA glass 1106 BH Silver
Venezuelan 954 Venezuelan 954
465 510 F
400-370 390
68 65
0.275 0.286
(a)
15.2
28.6
29.8
73.6
(c)
—
0.4
2.0
0.8
1.2
7
0.6
2.0-4.0
Trace
3.8
0.25
0.01
0.04
0.26
--
0.02
--
Trace
0.01
--
0.03
0.02
8.8
3.0
(b)
5.5
10.5
44.6
(d)
--
1.3
2.7
1.3
2.3
7
0.7
4.2
Trace
6.8
0.4
Trace
0.1
0.3
--
Trace
--
Trace
Trace
-.
Trace
Trace
22.0
5.4
40.7
(a)
36.7
22.6
78.1
137.4
(c)
--
__
0.18
0.11
0.68
0.08
0.12
0.11
0.01
2.4
0.12
Trace
0.11
0.19
Trace
0.01
--
0.01
0.01
0.05
Trace
0.02
NA
NA
(b)
17.3
20.4
60.3 •
(d)
—
__
0.24
0.18
1.28
0.26
0.14
0.15
0.02
4.3
0.18
Trace
0.15
0.24
Trace
0.01
--
0.01
0.01
--
Trace
Trace
NA
NA
7 8
4/28 5/4
Vycor "Vycor
MSA glass 1106 BH MSA glass 1106 BH
Mid Continent Mid Continent
490 F 435
230 400
64 71
0.422 0.422
(a)
145.1
74.0
32.6
251.7
(c)
0.16
1.00
5.00
2.00
3.00
7
1.60
5.0-10.0
--
0.96
0.31-
0.01
0.42
0.32
Trace
--
0.03
—
Trace
--
0.04
0.06
NA
NA
(b)
23.5
12.5
110.0
(d)
0.3
3.0
6.8
3.3
5.7
7
1.9
10.5
--
1.7
0.5
Trace
0.6
0.4
Trace
—
Trace
—
Trace
—
Trace
0.1
NA
NA
(a)
62.7
24.4
75.0
162.2
(c)
0.07
0.20
• 5.80
0.40
0.60
7
0.40
1.0-2.0
—
0.96
0.23
Trace
0.02
0.24
Trace
Trace
0.01
--
Trace
--
0.05
0.02
8.4
1.96
(b)
12.8
30.0
67.2
(d)
0.15
0.6
7.8
0.7
1.1
7
0.5
2.1
--
1.7
0.3
Trace
Trace
0.3
Trace
Trace
Trace
—
Trace
--
Trace
Trace
21.0
3.5
39.0
00
-------
111-19
FOOTNOTES FOR TABLE 2
(a) Filter weight arter desiccation; probe and impinger samples obtained
by evaporation H20 layer 230'F, evaporation organic layer, 75 F.
(b) Probe and impinger residues resulurried in acetone, refiltered onto
weighed Tissuquartz filter, dry 170 F.
(c) Optical emissions spectra (OES) and X-ray fluorescence spectra (XRF)
analysis; results in mg of element contained in total filter catch.
(d) Oxide weights equivalent to OES and XRF analyses; results in mg
of oxide/filter.
(e) NA indicates not obtained.
ATTELLE — COLUMBUS
-------
111-20
removed in these extractions were expected to yield estimates of organic
material (benzene soluble), and carbonate or nitrate contents (acid de-
composed) .
The high background provided by the glass filter for the ma-
jor elements of the fly ash resulted in a high degree of uncertainty in
the fly-ash analyses. For this reason a brief investigation was made of
suitability of a high-purity quartz filter as a substitute for the glass.
Tissuquartz 2500 QAO was obtained and compared with MSA glass 1106 BH
filter on the basis of weight, surface area, fiber diameter, and elemental
composition. These results are summarized in Tables 3 and 4.
TABLE 3. COMPARISON OF GLASS AND
QUARTZ FILTER MATERIALS
Filter
Tissuquartz QAO 2500
Weight taken (6 in?) , g
Dried He, 150 C, 1 hr, g
Heated in air, 450 C, 1 hr, g
Derived Properties
Moisture as received , wt 7.
Loss due to oxidation, wt %
Surface Areas. BET
After He, 150 C, ma/g
After air, 450 C, m2/g
Derived Property
Effective fiber diameter, a
0.265
0.245
0.242
8
1
4.0
6.3
0.3 - 0.5
MSA Glass
0.263
0.263
0.262
0
~0
1.7
2.1
1.1 -
1106BH
0.9
Note: Tissuquartz QAO 2500 obtained from Pallflex Products Corporation,
Putnam, Connecticut 06260
MSA Glass 1106BH obtained from Mine Safety Appliances Company,
Pittsburgh, Pennsylvania 15208
BATTELLE — COLUMBUS
-------
111-21
TABLE 4. OES ANALYSES (± 50%) FOR IMPURITIES
IN GLASS AND QUARTZ FILTER MATERIALS
Li
B
Na
Mg
Al
K
Ca
Ti
V
Cr
Mn
Fe
Ni
Cu
Sr
Zr
Ba
Pb
Micrograms /3- inch
Tissuquartz
< 3
6
30
150
150
16
Circle,
MSA Glass
--
2200
11000
4400
6600
3200
300 11000 to 22000
16
< 3
trace
< 3
30
—
16
16
—
< 3
< 30
22
< 2
6
4
220
2
10
—
22
22
4
BATTELLB — COLUMBUS
-------
111-22
The probe and impinger catch were collected by washing the
respective portions of the apparatus with water, acetone, and methylene
chloride. For the impinger portions, these washings were added to the 100
cc of water used in each of the first two impingers. In each case, the
water layer was separated from the organic layer portion and dried in an
oven at 110 C. Then, the organic portion was added to the same dishes
and dried at room temperature. After the bulk of liquid had evaporated,
the dishes were put into a desiccator to dry to a constant weight. Weigh-
ings made on successive days showed that no further loss of liquid was
occurring by evaporation. However, the weights bore no systematic relation
to the volume of gas sampled and so each of the probe and impinger residues
was reslurried in acetone and the insoluble residues were filtered onto
Tissuquartz 2500 QAO filter material. Because the fly-ash material would
be expected to have low solubility in acetone, the change in residue weight
obtained in this way provided a preliminary estimate of the nonfly ash
part of the probe and impinger portions. The results for these trials are
summarized in Table 5.
Table 5 shows that substantial, and apparently arbitrary changes
could be produced in the weight of particulate by the choice of drying con-
ditions. The acetone extraction was not valid as it removed inorganic
materials and H2S04 that should remain. The weights from this treatment
thus should'be too low. The slow and more or less continuous loss of water
as the particulate temperature was raised illustrated the difficulty in
recognizing when all free water had been removed and only combined water
remained. In the absence of a clear-cut distinction between these two water
forms, the choice among various drying temperatures was arbitrary. As free
water should be rapidly lost at 212 F, and in view of the weight changes
observed, the drying treatment at 210-220 F was preferred.
Recognition of the analytical uncertainties resulting from the
use of glass filters, and the large effects of the choice of drying con-
ditions on sample weight led to the design of a new series of trials in-
tended to permit comparison of some alternative choice of collection con-
dition. The decision corresponded in time with conversion of the Multifuel
BATTELL.E — COLUMBUS
-------
111-23
TABLE 5. EFFECT OF DRYING TREATMENTS UPON
APPARENT MASS OF PARTICULATE SAMPLES
Run 5 Probe
Filter
Impinger
Total
Run 6 Probe
Filter
Impinger
Total
Run 7 Probe
Filter
Impinger
Total
Run 8 Probe
Filter
Impinger
Total
75 F
15.2
28.6
29.8
73.6
36.7
22.6
78.1
137.4
145.1
74.0
32.6
251.7
62.7
24.4
75.0
162.2
Weight Collected
Ace tone-Extract
5.5
(28.6)(b)
10.5
44.6
17 -3 IM
(22. 6) (b'
20.4
60.3
23.5
(74.0)0>)
12.5
110.0
12 *8 rs%
(24.4)
30.0
67.2
After Drying, va) mg
175-180 F
13.4
(28.6)(b)
23.6
65.6
34.0
(22.6)
25.2
214.7
50'2 trt
(24.4)
73.4
148.0
210-220 F
11.1
(28.6)(b)
21.6
61.3
29'° (M
(22.6)
-------
111-24
Furnace to coal-fired operation (for the purposes of another program for
which the furnace was built) so that oil-fired sampling was not continued.
The first series of experiments on coal-fired operation of the Multifuel
Furnace was thus undertaken.
Experimentation Using the Multifuel
Furnace. Coal-Fired Operation
The coal fired was from Harrison County, West Virginia, see
Table 6 for analysis.
TABLE 6. ANALYSES OF COAL USED IN MULTIFUEL
FURNACE FACILITY, WEIGHT PERCENT
Coal Identification
HaO
C
H
N
S
Ash
(by difference)
3078(a)
1.3
68.8
5.01
1.2
3.5
15.4
4.4
3077(b)
1.44
76.6
5.46
1.4
3.00
7.3
4.8
(a)
3078 = Unwashed Pittsburgh bed.
3077 - Washed Pittsburgh bed.
These coals were from the Robinson Run
Mine of Consolidation Coal Company near
Shinnston (Harrison County), West Virginia.
BATTELLE — C O l_ U M B U S
-------
111-25
The .sampling conditions were chosen so as to correspond to the
array of experiments outlined in Table 7. Under the operating conditions,
the stack gas temperature at the sampling point upstream of the precipitator
was about 550-650 F and so the probe temperatures listed in Table 7 were
used as a guide for adjustment of minimum temperatures in the probe. In
those runs where 400 F probe was specified, temperatures were maintained
at 400 F or greater until the gases entered the filter holder portion of
the apparatus. In those runs where 250 F probe temperatures were called
for, the probe was allowed to cool to this temperature at a point pre-
ceding the filter-holder portion. Where the G/R-type coil was used, it
was intended to collect the H2S04 vapors that otherwise might accumulate
in downstream portions of the apparatus and so might have altered the
acidity of those portions. The acid thus collected was washed out of
the coil and combined with the probe or impinger catch, after the run was
completed, depending on whether the coil was before or after the filter
holder, respectively. In this way the coil represented a special portion
of a probe or impinger used for these research purposes.
The procedure for particulate analysis for this series of trials
is shown in Figure 18. It included C, H, and N analyses which were to be
used as the basis for interpreting the carbon-containing portion of the
particulate. For the- filter-catch portion of the sample a series of C,
H, and N analyses was included. Thus, after .obtaining a total of C, H,
and N analysis on an aliquot, a portion of the filter catch was extracted
with benzene, dried, and the C, H, and N-analyses were determined on this
portion. The relatively large amount of benzene used was expected to
remove all organic materials, and the difference in the two C, H, and N
analyses was called organic material. A portion of the residue from
benzene extraction was then extracted using aqueous HgS04 and the extrac-
tion was assumed to remove water-soluble inorganic material and decom-
pose all carbonates and nitrates. The change in C, H, and N analysis
due to this extraction was interpreted as due to carbonates and nitrates.
The C residue left after acid extraction included elemental carbon and
was referred to as such.
BATTEUL.E — COLUMBUS
-------
111-26
TABLE 7. PLAN FOR A SERIES OF EXPERIMENTAL
PARTICULATE COLLECTIONS FROM THE
LABORATORY COAL-FIRED FURNACE SYSTEM
Collection Conditions t1)
Run
Mo.
A
B
C
0
B '
P
G
H
I -
J
K
Gases
S
S
S
S
S
. S
S
S .
s<8>
StAIR
1:10(10)
S
Probe
250F 400F
316SS
Inconel
Q
Q
Q
Q
Q
Q
Q
Q
Q
Filter
230F
Q
Q
Q
Ag
G
Q
400F
Q
Q
Q
Q
Q
Oblective: Comparison of
Probe Filter
Coil' (2) (3)
No x
Mo x
No x
No x
No x
No x
No
Yes
Yes
Yes
Yes<9>
Acidity Time Dilution
(4) (5) (6) (7)
x
X
X XXX
X
X X
X
(1) The symbols are as follows: S - Stack gas; Q * Quartz or Vycor, G » Glass,
Ag * Silver.
(2) To examine acceptability of metal probes under "wet" collection conditions.
(3) To compare influence of available filter materials under 'Vet" collection
conditions.
(4)' To examine changes as major acid collection- is shifted from filter to coil
to impinger.
(5) To examine effect of residence time of particulate in sample train.
(6 and 7) To examine dilution effects - in (6) the effect of 03 and dew point on
mass collected, in (7) the effect of dew point at filter is emphasized.
(8) Total gas flow rate j that of runs A-H, but time of collection is doubled.
(9) The coil is before the filter in this case; otherwise it is between filter and impingers.
(10) Total flow rate is same as in runs A-H, K.
BATTELLE — COLUMBUS
-------
0
m
r
r
m
n
0
r
C
3
o
c
(0
Probe Sample
Coobtne washings
Record pH of water layer
Dry 212 F, evaporate
organic layer 75 F,
Dry combined residue 212 F .Weigh
Loosen residue with 81 spatula
Add few drops acetone and swab
sample onto 4 or 5 pieces of
Tissuquartz (equal to 3"
circle)
Dry 175-180 F
Ground to homogenize and
remove 3 aliquots
Put remainder into vial
OES
Analyses
CHN
Analyses
XRF
Analyses
Filter Sample
on 3" circle of Tlssuquarti
Desiccate at 75 F
Weigh, cue into halves
Designate as A, B
Weigh halves
Remove Aliquot
for analysis
Extract remainder
with benzene, dry
Remove aliquot for
analysis
Dry at 185-180 F
Extract remainder
with dil H2S04,
dry
Remove aliquot for
analysis
Put remainder into
vial
-*
r
ry
or
r
or
to
t
Weigh
Grind to homogen
Remove 2 aliquot
Put remainder in
vial
CHN OES
Analyses Analyses
J ^
XRF
Analyses
Inpinger Sample
Combine HjO and washings
Record pH of water layer
Dry 212 F, evaporate
organic layer 75 F,
Dry combined residue 212 F, Weigh
Loosen residue with nickel
spatula
\dd few drops acetone and
swab sample onto 4 or 5
pieces of Tissuquartz
(equal to 3" circle)
Dry 175-180 F
.M
h-l
K>
Grind to homogenize and
remove 3 allquots
Put remainder in vial
OES
Analyses
CHH
Analyses
XRT
Analyses
FIGURE 18.
FLOW CHART OF ANALYTICAL PROCEDURE FOR PARTICUIATE SAMPLES
FROM CCAL-FIRED OPERATION OF LABORATORY FURNACE
-------
111-28
The experimental series outlined in Table 7 was carried out
using three basic arrangements of apparatus:
For modes A-C
Probe
Filter
Impingerl
For modes H-J
Probe
Filter
Coil
Impinger
For mode K
Probe
Coil
ryj 1 4-A>*
r liter
Impinger
The results for these trials are summarized in Table 8. Their
significance is discussed under Discussion of Results.
'The benzene and acid extraction of the filter catch were dis-
continued after several had been completed. The relatively thick filter
.samples obtained from coal-fired operation tended to flake off during
extraction of some samples and thereby ruined the accuracy needed. In
addition, the total C, H, and N analyses for probe, filter, and impinger
portions, listed in Table 8, indicated that in many cases only a minor
part of the carbon in the particulate was found in the filter portion.
The results for one sample from oil-fired operation and two samples from
coal-fired operations are listed in Table 9. These results indicate the
nature of the analytical data for samples where handling difficulties
were minimal.
The data of Table 8 were used to make detailed material balances
so that the effects of these experimental variations in sampling mode
could be studied. The material balances, and conclusions are discussed in
a -later section along with the results of the other trials, see Discussion
of Results.
BATTELLE — COLUMBUS
-------
TABLE 8. AKM.YTXCU. M3A. FOR rAXXXCVXAXB COLLECTION FROM COAl-FIREO lABORATORTt 8UBXER
Run Ko. 9
Date 6/16/71
Materials
Probe Shore Vycot
Filter TissuquartzQAO 2500
Coal (•> 3078
Te=perature», F
Stack ... 630
Probe 185
O.E.S., «8 (h)
U 0.01
B 0.02
Na 0.35
Hg 0.42
Al 2.4
K 1.4
Cs 2.3
Tl 0.37
V 0.01
Cr 0.05
Kn 0.01
Fc 4.8
XI 0.02
Co
Cu 0.03
Sr 0.04
£r 0.01
Ko
AS
Sb
Ba 0.09
?b
Total 12.39
XXF. mg (1)
Al 6.9
K 3.5
C« 2.5
Tl 0.57
V 0.11
Fe 10.4
Cu 0.15
Pb
Total 24.1
S (01) 23.3
Cocbuitlcn, og (n)
C 8.0
H 4.7
H 0.4
F
202
0,02
0.10
0.3S
0.95
5.0
1.5
4,9
0.50
0.05
0.10
O.C2
5.0
0.02
0.03
0.09
0.03
0.01
— ^
0.24
18.94
9.4
4.0
8.1
1.0
0.15
12.7
0.15
35.5
y>. o
1.0
nil
0.1
I T
2.0
101 489
— 0.03
T 0.12
-- 0.74
0.10 1.47
0.14 7.5
0.04 2.9
0.32 7.5
-- 0.87
0.05
— 0.15
-- O.C3
0.18 10.0
0.04 O.C3
2.0 2.0
0.02 0.13
0.01 0.14
— 0.04
— 0.01
• » mm
— 0.33
2.S5 34.18
0.18 16.5
2.6 10.1
4.1 14.7
0.18 1.8
0.04 0.3
0.56 23.7
0.12 0.4
7.8 67.4
IU.6 07.5
13.2 22.2
3.0 7.7
0.9 1.4
10
6/17/71
Short Vycor.
Tissuquartz
3078
630
~AOO
400
74
0.516
G
P
3.9
990
0.55
0.13
5.2.
5.1
26.
19.
13.
6.4
0.26
0.90
0.13
25.
0.09
0.03
0.64
0.26
0.13
0.01
—
• 0.13
0.93
0,13
105.12
91.
40.
31.3
7.2
0.75
135.
0.82
306.1
26.9
2.9
3.2
F
232
0.12
0.11
1.14
1.7
12.
4.7
5.7
1.7
0.12
0.29
0.05
8.7
0.04
0.01
0.28
O.OS
0.05
0.01
0.01
0.12
0.40
0.06
37.41
32.5
12.
10.
2.5
0.23
40.
0.25
97.5
0.3
0.5
0.5
0.1
QA.O 2500
I
2.2
533
0.01
0.08
0.42
0.84
2.6
0.86
4.2
0.34
0.04
0.04
8.7
0.03
1.73
0.08
0.11
•»•
..
*«
0.16
20.24
4.3
2.3
3.6
0.49
0.27
10.4
0.16
21.5
41.6
4.0
11.7
0.2
T
1805
0.78
0.32
6.8
7.6
40.
25.
23.
8.5
0.38
1.23
0.23
43.
0.16
1.8
1.0
0.45
0.»
0.02
0.01
0.25
1.5.
0.19
162.77
127.8
54.3
44.9
10.2
1.3
185.4
1.2
425.1
74.8
7.4
15.4
0.3
11
6/18/71
Short Vyeor
Silver
3078
580
«AOO
245
72
P. 370
D
P
7.9
284
0.06
0.04
1.15
1.7
8.7
3.5
5.7.
1.2
0.03
0.12
0.04
11.7
0.04
0.01
0.28
0.11
0.03
0.01
.«
0.34
0.06
34.82
21.6
13.9
8.5
1.8
0.18
37.
0.23
83.2
4.3
17.7
0.4
F I
2.4
312 153
(D "
0.02
0.43
0.16
0.32
0.31
0.70
0.02
0.07
T
0.30
0.01
0.01
0.03
0.01
--
...
0.05
2.44
(1) 0.6
0.02
0.25
0.07
0.06
0.67
0.02
1.7
U). 13.9
7.2 3.0
5.3 8.9
0.3 3.9
12
6/18/71
Short Vycor
USA Glass 1106 BH
3078
550
.sAOO
230
83
0.322
Z
T P F
2.8
749 79 160
T (J?
0.01
0.38
0.69
1.47
0.37
1.65
0.14
0.02
T
1.87
0.01
• •
0.01
0.01
0.01
»•>
_.
„_
0.85
0.43
7.90
2.7 (J)
1.37
1.47
0.27
0.03
3.9
0.07
9.8
13. S (J)
27.9 9.2 (k)
14.2 2.4
4.6 0.2
I T
2.4
111 350
T
0.18
0.15
0.20
0.19
1.01
0.01
0.01
0.26
0.01
0.03
0.02
0.01
-
•
0.04
2.12
0.3
0.22
1.1
0.01
0.04
0.45
0.02
2.1
10.7
7.2
8.1
0.5
13 14
6/23/71 6/23/71
Ihconei .316 SS
Tlssuquart?QAO 2500 Tissuquartz; QAO 2500
3077 3077
630 640
240 240
230 235
72 71
0.392
B
?
3.4
567
0.35
0.26
4.3
4.3
26.
17.
8.6
2.6
0.61
2.5 •
0.17
8.7
2.6
0.02
0.16
0.86
0.09
0.03
0.09
»-
0.86
0.43
80.63
34.7
30.5
12.9
3.6
0.4
5.7
0.27
88.1
32.
4.7
0.2
0.601
F
235
0.11
0.27
1.06
2.1
10.7
1.6
7.9
1.6
0.21
0.32
0.05
8.0
0.02
0.01
0.09
0.26
0.04
0.01
0.01
1.06
0.11
35.53
21.7
4.8
9.2
2.4
0.27
26.0
0.22
64.6 '
24.1
0.4
3.7
0.1
I
3.0
77
0.01
0.35
0.12
0.14
0.25
0.83
0.02
0.01
T
0.16
0.01
0.03
0.01
0.01
0.01
--
T
0.03
2.09
0.09
0.19
0.31
0.03
0.03
0.28
0.02
1.0
7.1
12.0
6.6
0.7
T
879
0.46
0.54
5.4
6.5
37.
19.
17.
4.2
0.82
2.9
0.22
17.
2.6
0.06
0.26
1.13
0.14
Of04
0.10
1.9
0.62
118.25
56.5
35.5
22.4
6.0
0.7
32.0
0.5
153.7
63.2
17.1
10.3
1.0
A.
P
3.2
501
0.24
0.24
2.4
4.0
24.
4.0
11.9
2.4
0.32
3.2
0.32
16.
1.6
0.02
0.15
0.40
0.03
0.24
T
1.6
0.40
73.51
33.
10.
13.8
3.7
0.42
63.5
0.48
124.9
37.5
2.9
1.0
0.1
r
292
C.03
0.30
0.40
1.7
8.9
0.40
1.0
0.29
0.12
0.18
0.04
2.S
0.01
0.11
0.11
0.01
"*™
0.05
16.55
19.6
4.3
9.2
2.7
0.23
25.
0.22
61.3
46. i
0.8
6.9
0.3
I T
2.5
74 667
— 0.27
T 0.54
0.72 3.5
0.30 6.0
0.32 33.
0.25 4,5
0.83 13.3
0.01 2.7
• « 0.44
0.01 3.4
— 0.36
0.16 19.
0.01 1.6
T 0.02
0.01 0.27
0.01 0.52
O.OS
-- 0.25
0.01 1.7
0.08 0.48
0.09 52.7
0.15 14.5
0.38 23.4
6.4
0.03 0.7
0.3 88.8
0.02 0.7
1.0 187.2
6.5 SO. 5
7.0 10.7
2.3 10.2
1.0 1.4
H
M
I
N5
vO
-------
TABU J. ANVLYTXCAL DATA FOR PARTICUIATE COLLECTION FROS COAL-FIRED tABOSATORY 8URNEA (Contd)
Run fjo.
Bttt
rtaterUla
Probe
Flltir
Co»l
Teep«rature*( F
Stick
Fro be (b)
Filter
loipinger
Coil (e)
Sanple
Ca» Vol. SCF
Stapling M*4« (d)
(c)
' PH (O
Kas>, ng (g)
O.E.S., ng (h)
11
B
K*
Kg
A.1
K
Ca
11
V
Cr
Kn
Fe
HI
Co
Cu
Sr
Zr
Ko
Ag
Ba
Pb
Total
XRF, eg ' (1)
Al
K
Ca
Tl
V
Fe
Cu
PS
Total
5^ tf'R (m)
Cocbustlcn, OR (n)
C
H
N
15 •
6/23/71
Long Quartz
Tissuquartz
3077
5?0
~240
230
.75
11.9
C
P F
2.7
203
0.05
0.05
0.99
1.45
10.0
1.5
4.8
0.49
0.10
0.15
0.04
5.0
0.02
0.08
0.14
0.03
--
0.24
0.15
25.28
9.1
2.5
4.2
0.9
0.15
15.
0.08
31.9
24.6
1.6
2.6
0.2
163
0.03
0.32
0.92
1.36
9.3
0.93
6.8
0.93
0.14
0.09
0.03
4.7
0.01
0.07
0.23
0.05
0 01
0 46
0.09
26.47
13.4
2.7
6.0
1.6
0.15
15.3
0.13
39.3
23.8
0.3
3.9
0.1
QAO 2500
I T
2.3
133
0.04
0.28
0.26
0.39
0.20
0.65
0.01
0,01
0.20
0.01
0.02
0.02
0.03
--
T
0.04
2,16
0.27
0.01
0.42
0.025
0.05
0.2
„
1.0
6.7
4.3
9.1
0.6
504
0,08
0.41
2.2
3.1
20.
2.6
12.2
1.4
0.24
0.25
0.07
9.9
0.04
0.02
0.17
0.40
0.03
0.01
0.70
0.23
53,91
22.8
5.2
10.6
2.5
0.3
30.5
0.2
72.2
55.1
6.2
15.6
0.9
17
6/25/71
Short Vycor
Tissuquartz QAO 2500 .
3077
640
~400
395
72
184
0,336
U
"3.1
130
0.04
0.08
0.84
2.1
4.3
0.85
4.1
0.85
0.13
0.22
0.02
4.3
0.03
T
0.02
0.12
0.02
T
0.21
0.13
18.36
7.4
2.7
3.7
0.88
0.21
13.
0.07
28.0
9.1
2.4
0.3
0.2
132
0.05
0.37
0.51
5.3
5.3
1.05
7.8
1.6
0.05
0.16
0.05
5,3
0.03
T
0,09
0.20
0.03
0.15
T
0.52
0.11
28.67
16.8
4.2
8.4
1.95
0.24
20.4
0.13
52.1
8.
0.4
0.4
2.8
43
0.01
1,0
0.15
0.15
0.23
1.5
T
T
0,14
0.01
0.01
0.01
T
T
0.10
3.31
0.13
1.9
0.48
305
0.09
0.46
2.35
7.6
9.8
2.1
13.4
2,5
0.18
0.38
0.07
9.7
0.07
0.01
0.11
0.33
0.05
0.15
T
0.73
0.34
50,34
24.4
8.8
12.6
0.09 2.9
0.015 0.5
0.24 33.6
0.03
2.9
6.9
11.9
1.9
1.0
0.2
83.0
24.0
14.7
2.6
1.2
18
6/25/71
Short Vycor
Tissuquartz Q\t> 2500
3077
610
400
71
172
0.632
I
3.1
116
0.05
0.10
1.5
2.1
3.6
1.0
5.0
1.0
0.05
0.15
0.03
5.1
0.03
T
0.09
0.20
0.03
0.01
T
0.08
--
20.12
8.5
2.6
3.9
0.87
0.15
4.6
20.6
11.7
7.0
0.6
0.4
2.2
155 194'
0.09
0.45 0.02
t.4 0.23
5.3 0.18
6.8 0.21
2.3 0.24
6.6 0.8O
1.4 0.02
0.18 —1
.0.23 °»0l
0.05 T !
4.5 0.23
0.02 0.01
0.01
0,07 0.0'J
0.22 °«0*
'0.05 T
0.03
T
0.20 °.01
0.03 0.04
29.982.07
14.6 0.08
3.8 0.08
7.6 0.16
1.9 0.025
0.19 0.08
17.3 0.31
0.11 —
45.5 0.7
9.6 15.3
1.7 6.9
1.3 11.6
0.3 1.4
465
0.14
0.57
3.1
7.6
10.6
3.5
i2.4
2.4
0'.23
0.39
0.08
•9.8
0.05
0.01
0.19
0.46
0.08
0.04
' T
0.29
0.12
52.17
23.2
6,5
11,7
3.0
0.4
22.2
0.11
66.8
36.6
17.6
13.5
2.1
19
6/25/71
Short Vyeot
Tissuquartz
3077
570
.s-400
230
67
176
0.386
f K -F '
2.6
' 131
0.01
0.03
0.23
0.37
2.1
0.21
2.8
0.21
0.01
0.02
T
2.1
0.01
0.01
0.02
0.07
0.01
0.08
"*•
8.34
3.1
1.2
1.4
0.4
0.10
6.0
0.09
12.3
22.
1.8
3.5
0.3
118
0.01
0.42
0.28
1.2
2.9
0.41
3.8
0.41
0.04
0.03
0.01
4.2
0.01
0.02
0.07
0.02
0.01
0.20
0.08
14.12
3.1
1.3
2.7
0.63
0.09
5.9
0.10
13.8
28.6
0.5
7.2
0.1
QAO 2500
I T
2.2
70
0.01
0.35
0.11
0.14
0.10
0.52
0.01
0.01
0.09
0.01
0.01
0.02
0.02
T
0.01
0.04
1.45
0.035
0.03
0.05
0.026
0.2
0.04
0.5
6.1
3.9
5.4
1.1
319
0.02
0.46
0.91
1.7
5.1
0.72
7.1
0.63
0.05
0.06
0.01
6.4
0.03
0.02
0.05
0.16
0.03
0.01
...
0.29
0.12
23.91
6.3
2.5
4.1
1.1
0.2
12.1 .
0.2
26.6
56.7
6.2
16.2
1.5
23 24
7/14/71 7/14/71
Short Vycor Short; Vycor
Tissuquartz QAO 2500 Tissuquartz A.QO 2500
3077 '3077
750 720
230 390
230 230
74 72
0.332 0.327
C F
3.1 2.6 3.10
358 214 87 659 189 190
. 2.0 0.05 « 2.1 0.05 0.1
0.1 0.26 0.01 0.37 0,1 0.25
9.9 1.5 0.2 11.6 1.5 1.5
4.6 3.6 0.4 8.6 3.4 2.9
9.9 10.3 0.2 20.4 9.3 9.8
2.6 2.1 0.4 5.1 2.0 2.0
9.9 10.3 0.4 20.6 9.8 9.8
1.0 1.0 0,02 2.0 1.0 1.0
0.07 0.2 — 0.27 0.05 0.2
0.3 0.4 -- 0.7 0.3 0.3
0.03 0.04 — 0.07 0.03 0.03
13.2 10.3 0.2 23.7 9.8 9.8
0.03 0.05 -- 0.08 0.05 0.05
0.01 0.01 0.01 0.03 0.01 0.01
0.2 0.2 0.1 0.5 0.1 0.1
0.3 0.4 0.04 0.74 0.2 0.5
0.03 0.03 -- 0.06 0.02 0.03
0.01 0.01 — 0.02 0.01 0.02
T — ,T T
0.3 0.04 0.04 0.74 0.2 0.5
0.3 0.3 0.04 0.64 0.1 0.3
56.1 42.2 2.1 100.4 39.0 39.9
17.3 19. 0.18 36.5 14.8 19.6
5.5 4.8 0.7 11.0 4.1 5.1
6.5 7.6 0.1 14.2 5.3 8.6
1,7 1.9 — 3.6 1.36 2.0
0.2 0.18 0.03 0.4 0.14 0.18
27.6 21.5 0.21 49,3 22,5 23.6
0.35 0.09 0.03 0.5 0.2 0.15
0.6 -- -- 0.6 —
59.8 55.1 1.3 116.1 43.4 59.6
19.7 26.6 9.2 55.5 13.6 15.5
34.7 1.0 2.2 37.9 7.2 0.8
1.3 3.1 8.0 12.4 0.4 1.7
0.1 0.6 0.7 0.1 0.1
2.35
35 414
— 0.15
0.01 0.36
0.1 3.1
0.4 6.7
0.1 19.7
0.2 4.2
0.4 20.0
0.02 2.0
— 0.25
0.6
-- 0.06
0.1 19.7
-- 0.1
0.02 0.04
0.07 0,27
0.03 0.73
0.01 0.06
— 0.03
0.03 0.73
Ofl2 1 17
— 0.4
1.5 80.4
0.09 34.5
0.6 9.8
0.1 14.0
3.4
0.03 0.4
0.2 46.3
0.03 0.4
0.38 0.4
1.1 109.2
8.9 38.0
1.5 9.3
3.4 5.5
0.9 1.1
-------
111-31
TA3LE 8. ANALYTICAL DATA FOR PARTICIPATE COLLECTION FROM COAL-FIRED LABORATORY BURNER (Contd)
Run No.
Date
Materials
Probe
Filter
Coal (a)
Temperatures, F
Stack
Probe (b)
Filter
loplnger
Coil
00
Li
B
Na
Mg
Al
K
Ca
Ti
V
Cr
Mn
Fc
Ni
Co
Cu
Sr
Zr
Ho
Ag
Sb
Ba
Fb
(1)
Al
K
Ca
Ti
V
Fe
Cu
Pb
(m)
(n)
N
26
7/14/71
Short Vyeor
Tissuquartz QAO 2500
3077
680
390
395
72
.406
G
P
4.30
413
0.2
0.2
2.9
5.0
21.5
7.2
14.3
1.4
0.7
1.4
0.05
14.3
0.2
0.01
0.4
0.7
0.05
0.02
T
0.7
0.7"
0.6
72.5
29.4
11.
11.
3.2
0.64
59.
0.38
2.0
9.7
17.6
..
F
208
0.15
0.25
1.5
4.1
15.2
2.5
10.2
1.0
0.5
0.8
0.04
10.2
0.1
0.01
0.2
0.3
0.03
0.03
T
0.3
0.8
0.3
48.5
16.6
5.0
6.2
0.85
0.69
5.7
0.24
—
6.1
3.6
0.7
0.1
I T
2.25
142 768
~ 0.35
0.01 0.46
0.1 4.5
0.4 9.5
0.2 36.9
0.04 9.7
1.3 25.8
0.02 2.4
-- 1.2
-- 2.2
-- 0.09
0.3 24.8
— 0.3
-- 0.02
0.03 0.63
0.04 1.0
— 0.08
-- 0.05
-- T
0.04 1.0
0.02 1.5
0.04 0.94
2.5 123.5
0.28 46.3
-- 16.0
0.34 17.5
« 4.0
-- 1.3
0.68 65.4
0.03 0.7
— 2.0
8.8 24.6
2.0 23.2
9.3 10.0
0.7 0.8
28
7/15/71
Short Vycor
Tissuquartz
3077
570
~400
405
69
163
0.310
H
P F
4.60
149 274
0.05 0.1
0.05 0.2
0.9 1.7
0.9 1.7
6.7 11.5
2.2 5.7
4.5 8.6
0.5 1.1
0.03 0.1
0.09 0.2
0.02 0.06
9.0 11.5
0.02 0.06
0.01 0.01
0.05 0.1
0.1 0.3
0.02 0.06
0.02
-_
0.1 0.3
0.2 0.3
0.2 0.2
25.6 43.8
14.5 31.
6.1 11.8
4.5 10.7
1.1 2.4
0.085 0.22
19. 35.5
0.06 0.13
0.06 1.2
4.6 6.6
3.2 8.6
-- 0.9
0.2 0.4
29
7/15/71
Short Vycor
QAO 2500 Tissuquartz
3077
640
~400
405
65
174
0.685
I T
2.55
94 517
— 0.15
0.01 0.26
0.4 3.0 .
0.4 3.0
0.2 18.4
0.2 8.1
2.0 15.1
0.02 1.6
-- 0.13
0.01 0.3
— 0.08
0.2 20.7
— 0.08
0.01 0.03
0.02 0.17
0.04 0.44
0.01 0.09
-- 0.02
_.
0.04 0.44
0.01 0.51
0.2 0.6
3.8 73.2
0.19 45.7
2.9 20.8
0.24 15.4
-- 3.5
0.3
0.3 54.8
0.02 0.2
— 1.3
7.2 18.4
4.0 15.8
3.7 4.6
0.2 0.8
I
P F
<».bU-
340 311
0.1 0.1
0.1 0.2
1.9 1.8'
1.9 0.6
12.8 18.3
6.4 0.6
9.6 12.2
1.3 1.2
0.3 0.06
0.4 0.4
0.06 0.06
12.8 12.2
0.1 0.06
0.01 0.1
0.1 0.1
0.3 0.3
0.06 0.06
0.01 0.04
T
0.3 0.3
0.3 0.4
0.5 0.5
31.2 33.2
13.5 13.9
10.8 11.9
2.5 2.7
0.33 0.3
44. 41.3
0.14 0.28
0.07 2.1
7.1 6.0
3.5 1.3
0.7
0.1
QAO 2500
I " T
2.65
628 1.279
0.1 0.3
0.1 0.4
1.9 5.6
2.8 5.3
13.9 45.0
2.8 9.8
13.9 35.7
1.4 3.9
0.03 0.39
0.3 1.1
0.03 0.15
18.6 43.6
0.03 0.19
-- 0.11
0.1 0.3
0.4 1.0
0.05 0.17
0.01 0.06
T
0.4 1.0
0.4 1.1
0.6 1.6
156.8
10. 74.4
6.8 34.2
6.3 29.0
1.5 6.7
0.32 1.0
27.5112.8
0.17 0.6
— 2.2
74.5 87.6
1.2 6.0
8.2 8.9
0.5 0.6
31
7/15/71
Short Vycor
Tissuquartz QAO
3077
560
~AOO
230
69
174
0.389
K
P
2.9
409
0.1
0.07
2.1
2.8
21.3
7.1
14.2
1.4
0.2
0.2
0.07
14.2
0.04
0.01
0.1
0.4
0.07
0.01
--
0.4
0.4
0.2
36.9
15.8
11.1
2.6
0.27
38.4
0.15
0.08
38.0
6.4
..
0.1
F I
2.8
95 60
0.01 —
0.16 0.01
0.3 0.03
0.8 0.4
1.6 0.02
0.4 0.01
5.9 1.1
0.3 0.02
0.02 —
0.03 0.01
0.1
4.0 0.03
..
—
0.4 0.02
O.OS 0.04
0.01 ~
—
T
0.08 0.04
0.2 0.01
0.04 0.07
4.4 0.14
2500
T
564
0.11
0.24
2.4
4.0
22.9
7.6
21.2
1.7
0.22
0.24
0.17
18.2
0.04
0.01
0.16
0.52
0.08
0.01
T
0.52
0.61
0.31
81.2
41.4
2.3 0.025 18.1
3.3 0.29
-- 0.31
-- 0.05
5.8 0.48
0.1 0.02
O.C6 --
17.3 4.5
9.2 1.5
3.3 4.3
0.3 0.5
14.7
2.9
0.3
44.7
0.3
0.1
59.8
17.1
7.6
0.9
-------
111-32
FOOTNOTES FOR TABLE 8
(a) The coals used are identified as follows:
3077 Pittsburgh Bed, washed, S » 3.07.
ash - 77.
3078 Pittsburgh Bed, unwashed, S = 3.57.
ash =.13%.
(b) Probe temperature was controlled by specifying the temperature for
the coolest portion just preceding the filter chamber. The inlet
end of the probe is at stack temperature.
(c) Used to remove H2S04 vapor, but no frit was used to retain mist so
that particulate collections would not be altered. The location
of the coil is specified by the sampling mode.
(d) Sampling mode refers to the alphabetized listing of runs in Table 7
of this report. Gas samples volumes are expressed in cubic
meters at 1 atmosphere and 20 C.
(e) The letters in this row are column headings -
P « Probe
F - Filter
I » Impinger
T ** Total.
(f) pH obtained on the water layer after being combined with H20 used
for washing.
(g) Weight of the total catch after drying at 212 F.
(h) Analysis results for weights of elements found by optical emission
spectroscopy. Because of the filter material Si could not be
determined with an accuracy better than + 50 percent.
(i) The silver filter requires a different handling procedure and was
not analyzed at this time.
(j) The high background for many elements in the glass filter used in
this experiment makes this analysis meaningless in the present work.
(k) The glass filter material is not suitable for use in the C-H-N
combustion train.
(1) X-ray fluorescence analyses for selected elements.
(m) Obtained by XRF and represents all forms of sulfur present.
(n) Carbon, hydrogen, and nitrogen combustion train analyses.
BATTELLE — COLUMBUS
-------
Ill-33
TABLE 9. COMPARISON OF C, H, N CONTENTS FOLLOWING
EXTRACTION OF FILTER PORTIONS WITH BENZENE
AND WITH AQUEOUS ACID
Run Element
4<" C
H
N
9(b) C
H
N
H(b>
H
N
Total,
mg
0.19
0.12
0.02
1.0
nil
0.1
7.2
5.3
0.3
After CSHS
Extraction,
mg
0.22
0.12
0.02
1.0
5.5
0.1
7.0
0.5
0.3
After CsHa and
then Aqueous Acid,
mg
0.14
0.02
0.02
0.9
0.6
nil
6.2
1.0
0.1
^ Oil-fired operation, see experimental data, Table 2.
' 'Coal-fired operation.
BATTEL.LE — COLUMBUS
-------
Ill-34
. The methods for handling of fly ash, and'.the analytical require-
ments, had become fairly well established as a result of the earlier series
of trials. However, basic problems remained to be explored for the sulfur
oxide portion of the catch. For this reason, a series of trials was con-
ducted to determine the magnitude of S0g retention and oxidation during
sampling and during subsequent handling of the samples. This series, made
by sampling gases from the Battelle Multifuel Furnace using coal, is described
in this section. A subsequent series made at Edgewater Power Station, also
using coaL at Lorain, Ohio, utilized the experience gained in handling the
sulfur oxide portion and is described in the following section.
For these experiments, the objective was to compare the total
sulfur oxide contents of the stack gas, as determined by the Goksrfyr-Ross
procedure, with the sulfur oxide retained in a series of impingers.
More explicitly, the objective was to compare the SO, content of the stack
3
gas with the S0e found in the impingers and to note any contribution of
3
S0a to the total sulfur retained in the impingers.
The sulfate normally found in the evaporated residue of impinger
catch solutions can originate by oxidation of S02 at any of several stages
in the sampling and analysis procedure. Thus,.the sulfate found may be
the sum of
(1) SO present in the stack gas
(2) SO generated during sampling
3
(here 40 minutes)
(3) SO. generated in the impinger solutions
3
during storage under air in bottles, and
transported to the laboratory (usually
days or weeks)
(4) S0g generated during sample evaporation
and analysis.
BATTELLE — COLUMBUS
-------
III-35
The data for this series provides a comparison of the magnitude of (1),
measured by the Goks^yr-Ross procedure, with the sum of all four,
measured several ways in this work but comparable to the range of con-
ditions normally encountered in sampling processing.
The experimental procedures used can be understood with the
help of the schematic representation of the experimental plan shown in
Figure 19. Details of the configuration used for each run and data for
these trials are presented in Table 10. For these trials, only the sulfur
analyses for the components contained inside the dashed line were of
interest. The coal used was Number 3077 which is Pittsburgh Bed Coal,
washed, S = 3.0% and ash = 7% (see Table 6).
The first four impingers used in this series of experiments
each contained 100 ml of water or aqueous solution. Impinger 5 shown
in Figure 19 was empty. To determine the possibilities for oxidation of
S02 to S0g , the total catch of S03 in the four impingers was compared for
different experimental conditions, and the relative concentration of S00
3
in each of the four impingers was examined for evidence of a continuing
production of SO during collection, storage, and analysis.
3
The liquid from the impingers and the wash water were bottled
and kept separate from the subsequent washings with acetone and methylene
chloride. Only the water phase of the impingers was analyzed for sulfur
in this series. The analysis of the probe samples represents the total
sulfate and sulfite sulfur retained after evaporation of the water layer
at 210-220 F and subsequent 75 F evaporation of the organic washings.
The filter analysis represents the material as collected at 250 F and .
subsequently stored over P305 .
Sulfate in impinger solutions was determined by making the
solution strongly acid by adding 15 ml of 6 N HC1 to a 50-100-ml aliquot
and heating to drive off the SO^ . The residual sulfate was measured
BATTELLE — COLUMBUS
-------
Probe
Filter
Stack
Probe.
Filter
EPA Train (plus added impingers or coil)
Impinger
1
Impinger
2
Impinger
3
Impinger
4
G/R
Coil
Bubbler
1
Bubbler
2
Impinger
5
I
CO
cr«
Goks^yr-Ross
FIGURE 19. SCHEMATIC REPRESENTATION OF THE EXPERIMENTAL
APPARATUS FOR INVESTIGATION OF S02 AND S03
-------
TABLE 10. OXIDATION OF S02 DURING SAMPLING OF STACK GASES FROM
BATTELLE MUtTIFUEL FURNACE UNDER VARIOUS CONDITIONS.
(OXYGEN CONTENT OF STACK GASES APPROXIMATELY 4 PER-
CENT, SAMPI£ TAKEN BEFORE EliCTROSTATIC PRECIPITATOR)
Run 38
Date 9/17/71
Temperature, F
Stick 375
Probe 380
Filter ' 250
loplnger 64
G/R Coil(a) 163
Go, Volumes, a3
Coll 0.126
EPA Train 0.406
Experimental Mod«(b> I
Parttculate, mg(O og mg/m3
Filter 208 512
Probe 130 520
Remits, ng S/in3 /i\
as S03 as S02 Total
C/R Procedure<*> 15.5 2820 2836
Probe<£> 10.2 2.5 12.7
Fllter 20.5 4.2 24.7
Implngers^*'
1 12.0 49 61
2 5.3 88 93
3 6.7 92 99
4 8.8 92 101
lap Inge r Total 32.8 321 354
Ratio Implnzera
Stack 2.11 0.11 0.12
Impingera
1
2
3
4
Inplnger Total
Ratio Implngera
Stack
9.5 ' 174 184
114.5 834 950
3.78 0.35 0.40
1
41
9/17/71 .
375
400
255
67
177
0.163
0.338
I
mg mg/m
113 334
' 36.5 108
as SOn as SOJr Total
35.7 2400 2490
26.1 1.4 27.5
30.0 • 6.7 36.7
15.5 70 85
8.5 169 177
9.9 181 191
9.5 156 166
43.4 576 619
1.22 0.24 0.25
42
9/29/71
370
400
250
72
161
0.212
0.416
I
mg mg/m3
214 515
W , .W
as collected^')
as SOi as SO* Total
41 2849 2890
..
12.7 -- 11.7
4.9 34 38.8
10.2 39 49
13.4 33 46
17.0(1) 85 102
T5T5" ~T5T • 236
1.10 0.067 0.082
acldlfled(e)
10.9 31 42
6.7 35 42
13.1 36 49
15.5 62 78
1 ""•••- "•
46.2 164 211
1.12 0.059 0.074
43
9/29/71
370
410
255
74
163
0.200
0.391
IV
mg rng/n3
182 465
(d) (d)
as SOa as SO; Total
43 2S50 2890
.-
10.6 .- 10.6
15.9 19 35
14.1 42 56
16.2 37 53
23 37 60
69.2 135 ~3TO
1.46 0.046 0.071
H
-------
111-38
FOOTNOTES FOR TABLE 10
(a) Refers to temperature of water circulated through jacket of G/R
coil used for SOg (or t^SO^) collection in parallel with the
particulate train.
(b) Experimental modes for the particulate train
I First four impingers each have 100 ml 1^0. The
fifth is empty and the sixth contains Drierite.
II The first two impingers each have 100 ml 1^0 and
the second two have potassium phthalate buffer in
1^0 at pH - 5.5. The fifth and sixth impingers are
as in I.
Ill The first four impingers each have 100 ml 1^0 as
in I, but 76 mg of filter catch from Run 37 was
added to the third impinger. (See also Footnote h).
The fifth and sixth impingers as above.
IV The first four impingers each have 100 ml 0.1 N HC1.
The fith and sixth impingers as above.
(c) Filter catch collected at 250 F and stored in desiccator until
weighed. Water wash of probe dried at 212 F and organic wash added
and dried at 75 F. Gas volumes are cubic meters calculated at
20 C and 1 atmosphere.
(d) Not analyzed.
(e) The portion labeled "as collected" was analyzed within 24 hours
without any further treatment. That labeled "acidified" had
3 ml of 6 N HC1 added as soon as it.was collected and separated
from the other half of the impinger solutions. Samples analyzed
within 24 hours.
(f) The probe for the particulate collection and the inlet to the G/R
train were both at about 400 F. The S(>2 and 803 contents for the
G/R procedure are determined directly by titration; for probe and
filter the sulfate and totals were determined by XRF and the 803
determined by difference. The catch of the G/R procedure is assumed
to represent the sulfur analysis of the stack gases being sampled.
(g) Impingers 1 and 3 have open tubes, 2 and 4 are of the Greenfield-
Smith type. The total S was obtained by both gravimetric and XRF
methods. Good agreement was found for these two analytical methods.
Sulfate sulfur was determined gravimetrically and sulfite determined
by difference. Analyses for water layer only.
(h) 76 mg of filter catch from Run 37 was added to the third impinger
of this run and contributed a total of 5.4 mg S as sulfate and
6.7 mg total sulfur. .The values listed in the table are corrected
for this contribution.
(i) This impinger contained 200 ml of water initially so as to provide
evidence for the influence of the volume of water used.
(j) SQa as sulfite sulfur.
-------
Ill-39
gravitnetrically as BaS04 . Direct measurement of sulfites in impinger
and probe solutions by iodometric titration was not attempted at this time
because of the possibility that other species e.g., ferrous iron or
organic materials, might contribute to the titration value. Total sulfur
was measured as sulfate in two ways: in one method, sulfite was oxidized
by bromine and the total measured as BaS04; in the other, the solution
was made alkaline to prevent SOS loss and then analyzed as a solution by
X-ray fluorescence. These two methods gave very good agreement for the
total sulfur estimate.
The probe in the typical Goks«fyr-Ross procedure is normally a
glass tube that is almost entirely inside the stack and the filter is a
small ball of glass wool. The external part of the probe and the glass
wool are heated to 400 F. Neither of these parts was analyzed in this
series of experiments. Also, the S02 effluent from Impinger 4 in the EPA
train was not determined.
For the data of Table 10, the Goks^yr-Ross procedure was run
simultaneously with the EPA sampling train and for the same period of
time. Because the total gas flow through these two systems is different,
the results of the table are expressed as mg S (as S03 or S03) per m3
gas sampled so that results for the two procedures can be compared directly.
In each case, the total sulfur of any kind in an impinger or gas sample
can be found by multiplying the mg S/m3 given by the gas volume sampled.
Field Experiments
Effect of Drying Conditions. Particulate material collected
during sampling of exhaust from a cyclone-fired boiler of the Ohio Edison
Company at Niles, Ohio, was used to continue the investigation of the
hygroscopic nature of some samples. For these samples, the common practice
of desiccation over Drierite was abandoned in favor of P_0_ as the desiccant.
2 O
Table 11 shows data for both water and organic wash layers of
probe and impinger catch for two runs. Data are presented showing sample
BATTELLE — COLUMBUS
-------
TABLE 11. EFFECT OF DRYING CONDITIONS ON SAMPLE WEIGHTS-
COAL BURNING OPERATION, NO PRECIPITATOR •'
(Weights in Grams)
Initial
Evaporation
(b)
Desiccate,
24 hr
(c)
Heat,
18 hr
(d)
Desiccate,
25 hr
(c)
Heat,
46 hr
(d)
Desiccate,
23 hr.
(c)
Appearance
Water layer
Probe
#1
Impinger
#1
#2
Organic layer(c)
0.5888 0.5906 0.5865 0.5888 0.5866 0.5881 White and tan solid
0.8425 0.8462 0.8405 0.8452 0.8408 0.8427 White and tan solid
0.2375 0.2417 0.2290 0.2475 0.2237 0.2313 Vitreous dark brown
0.5035 0.5217 0.4917 0.5103 0.4762 0.4882 Vitreous tan, translucent
M
M
H
Probe
#1
#2
Impinger
#1
#2
0.0235
0.0193
0.0032
0.0017
0.0242
0.0202
0.0041
0.0024
0.0184
0.0144
0.0002
0.0000
0.0196
0.0155
0.0015
0.0007
0.0177
0.0151
0.0001
-0.0004(f)
0.0185
0.0161
0.0012
0.0005
Black, solid specks
Black, fine solid
Black solid
Tan solid
(a) Samples collected from Ohio Edison Power Station, Niles, Ohio, cyclone-fired boiler.
(b) For water layer, evaporate to dryness in 212 F oven; organic layer, evaporate at
room temperature in air, then heat 2 hours at 155-160 F.
(c) Over PgOs - all dishes in desiccator at one time. , .
(d) At 212 F, cool 1 hour or less in desiccator before weighing,
(e) Combined acetone and methylene chloride solvent washings.
(f) Erroneous negative weight results from difficulty of weighing small amounts of sample in dishes
weighing several grams.
-------
Ill-41
weights after repeated exposure of the samples to alternative periods
of heating and desiccation.
The solids recovered from the water layers of the probe catch
are tan and white solids similar in appearance to the filter catch and
the percentage changes in weight of these solids due to repeated heating
and desiccation periods are shown in Table 11 to be quite small (less than
1 percent). The impinger water layers are vitreous in appearance and are
slightly more sensitive to drying procedures (maximum variation in weights
were about 10 percent).
The residues from the organic layers are black, noncrystalline
in appearance and show changes in weight that are greater than 25 percent.
However, as the weights of the organic layer residues were only about
five percent of weights of the residues from the water layer, these large
variations in weight may not appreciably influence the weight of the
total sample.
Although based on only a few samples, the differences in the
effects of humidity and heat treatment on reproducibility of weighings
from Table 11 appear to correlate well with the physical appearance of
the residue. Thus, physical appearance may serve as an indicator of
sample drying and weighing difficulties and may identify samples that
will require analysis for water content.
Experiments at Edgewater Power Station. The problem of measure-
ment of the sulfur content of particulate was approached by attempting to
account for all of the sulfur in the gases and particles sampled, and to
examine the influence of variations in the sampling train configuration
on sulfur retention in the various parts of the train.
Table 12 summarizes the sulfur data for several samples collected
with various impinger configurations during field sampling on the pulverized
BATTELLE — COLUMBUS
-------
TABLE 12. INVESTIGATION OP THE OXIDATION STATE FOR SULFUR OXIDES RETAINED IN
SEVERAL VARIATIONS OF THE PARTICULATE SAMPLING TRAIN
hu> mid
Sulfur
o»m«
B*ttr»lM4
it
SOj
Total
47
SOj
Tot.?
48
80,
SO,
tet»l
51
SOj
SOj
Tout
32
SO,
Tgtti
Dot<
of
SM»II
10/3
10/5
10/5
10/7
10/7
Stick
I«p..
r
355
180
370
410
370
CM AulyiU,
Percent
HjO CO, 0,
2.5 11.5 I. 5 to
9.0
3.1 11.5 8. 5 to
9.0
2.6 U.I S.Ito
9.0
5.5 It.] 6.8CO
to 7.9
lt.0
5.1 11, 5 6. 8 to
to 7.5
12.0
Ce.
SupK
VoluM,
•»
O.i2«
0.631
0.648
0.782
0.899
rla
48.4 nil
tch(c)
14
HjO Org.
3.88 '
99.2 nil
6.0
51.2 nil
1.41
37.6 nil
Tool «5
(n) KjO,
4.87
2112
339.0 2112
9.3
2352
148.7 2352
16.2
2366
178.8. 2366
14.9<")
1157
114.5M»3 METHOD
• IllUrun 5/.5
Coll
«M
HU.r Bubblir Totil
12.4 12.4
2854 1894
<•. 37 2866 2866
12.4 |2,4
1834 285k
0.57 2866 2866
12.4 11. 4
2854 2854
0.57 2866 2866
17.3 17.3
1744 "1744
0.92 2761 1761
17. J 17.3
2744 2744
0.92 2761 2761
M
M
H
*>
N3
Stapling train trrAngeMnuU)*
M
) (32F>| H,0 (32F)|H20 -I- eeh 11,0 <32F)|
a 1 p I g4~o4
|l62F|H2° (32f)|H,0 (32F)|
-4»-P-o--| -D-4-
{He.tedl 250F 150F 110F m I HjO (32F)( HjO (32F) I I
-3-*.—q-M-H—I—\
-------
111-43
FOOTNOTES FOR TABLE 12
(a) Samples taken from the electrostatic precipitator at the outlet
stream at Edgewater Power Station, Lorain, Ohio.
(b) Impinger residues were not determined since sulfur analyses were
obtained on the solutions themselves.
(c) The headings for probe, filter, and impinger portion include the
terms dry, H£0, and organic, which pertain to the dry catch water
layer and acetone-methylene chloride layer portions of the sample
respectively. "H20" refers to a 6-7 percent H£02 aqueous solution
used in Impinger 5. The impinger water samples were made on
aliquots stored in completely filled bottles immediately after the
run.
(d) These values are the totals for the first four impingers only.
(e) These values are the grand totals for all portions of the catch.
(f) Goks«5yr-Ross procedure. Sample collected simultaneously with
particulate sample but using separate train. Although isokinetic
conditions were not attempted for this train, the quartz wool
plug used to filter particulate from this gas sample was analyzed
for total sulfur in order to attempt a total sulfur balance. The
probe was a hot, smooth glass tube that contained little material
and was not analyzed.
(g) 45.6-mg ash from Run 45 added to the third impinger in this run.
The dry residue recovered from this impinger weighed 29 mg and no
attempt has been made to correct the sulfur analysis of this im-
pinger for the solid added.
(h) Includes coil.
(i) Includes second cyclone held at 150 F.
(j) Train variations shown schematically
I - Probe including cyclone
Filter
-4/J4 — Coil, as in Goksrfyr-Ross procedure
— Q~ Impinger.
BATTELUH — COLUMBUS
-------
III-44
coal-fired boilers at the Edgewater Power Plant, Lorain, Ohio. The vari-
ations in sampling apparatus are shown schematically below the table.
During the sampling operation, two coals were fired at normal
full-load rates for the plant. Partial analyses for these bituminous
coals, as received supplied by the Ohio Edison Company are
Dates - 10/4 to 10/6/71 10/7 and 10/8/71
Moisture, 7. - 4.86 5.09
Ash, 7. - 10.91 8.49
Sulfur, 7. - 3.24 2.65
Heat Value, Btu/lb - 12,215 12,706
.The gas was sampled from the electrostatic precipitator outlet
stream: Runs 46, 47, 48 were made simultaneously using adjacent ports
in the rectangular duct leading from the precipitator, and Runs 51 and
52 were a similar pair of simultaneous runs carried out at another time.
An adjacent port was used for sampling by the Goks^yr-Ross procedure for
measuring the SO and SOg concentrations of the flue gas simultaneously
with each set of runs. From prior information obtained by plant staff
the precipitator efficiency was about 98+ percent.
In order to minimize the weathering and oxidation of impinger
water samples, 60-ml aliquots of the combined impinger water and water
washings were taken from each impinger of Runs 46, 47, 48 and stored in
completely filled bottles until analyzed. The results for the impinger
portion of Table 12 are based on those aliquots which, in general, show
about the same total sulfur contents as the portions stored conventionally
in partly filled bottles, but they exhibit substantially less sulfate
contents. The improvement in procedure represented by this change is
illustrated in Table 13. The differences in sulfate content confirm the
expectation that considerable sulfate is generated during sample storage
for transportation to the analytical laboratory.
BATTEL.LE — COLUMBUS
-------
TABLE 13.
111-45
COMPARISON OF THE SULFUR CONTENTS OF
IMPINGER WATER SAMPLES STORED IN
DIFFERENT WAYS
mg S/cc solution
Sulfate Sulfur
Total Sulfur
Run 46
Impinger
Run 47
Impinger
Run 48
Impinger
1
2
3
4
1
2
3
4
1
2
3
4
Conventional
(a)
0.019
0.028
0.090
0.039
0.002
<0.002
<0.003
0.054
0.042
0.044
0.12
0.010
Aliquot
(b)
<0.003
0.0025
0.0025
0.017
<0.003
<0.003
0.017
0.023
0.025
0.040
0.017
0.008
Conventional
(a)
0.05
0.29
0.55
0.47
0.11
0.13
0.14
0.18
0.23
0.33
0.30
0.35
Aliquot
(b)
0.11
0.52
0.45
0.43
0.08
0.12
0.22
0.20
0.15
0.27
0.23
0.33
(a) Samples stored approximately 3 weeks in partly filled bottles.
(b) 60-ml aliquots stored in completely filled bottles.
SO can be lost by evaporation and oxidation of the
sample if the seal on the bottle is not tight. This seems
to be a problem in field samples where storage times for
such samples is often longer than desirable. The major
evaporation and oxidation losses seem to be in the conventional
samples; some caps-on aliquot samples were reported to be "not
tight" after their return to the laboratory so some loss from
these is possible. Analytical uncertainties for these samples
are in the second significant figure, so these uncertainties
also contributed.
BATTELLE — COLUMBUS
-------
111-46
The problem of the collection of the fly-ash portion of the
sample was investigated in a separate series of experiments summarized
in Tables 14A, 14B, and 14C. For this purpose, the analyses were chosen
to emphasize the amounts of the individual cations collected, and, in
order to identify any problems in sample fractionation during collection,
analyses were attempted on all portions of the sampling rig regardless of
the apparent importance of the mass of the catch in that portion.
The experiments shown in Tables 14B and 14C represent pairs of
experiments made simultaneously, one on the precipitator inlet stream and
one on the outlet stream. Simultaneously, G/R collections were made on
both inlet and outlet streams. Two additional sets of G/R collections
were made to determine reproducibility (see Runs 53 and 54) but the sample
from 54A was lost.
The sample-hand ling and analytical procedures used for the
samples . from Edgewater Power Station were varied according to the objectives
for each experiment. The following is a description of those procedures
for the ten runs made in this field work. The program comprised
nearly 300 actual analyses of samples all of which required some form of
preparation and/or treatment prior to the analyses.
Each run represented a number of subsamples. These subsamples
included particulate catches on both quartz filter material and quartz
wool; dry, loose powder from the probe area; water washes and organic
washes at and near the probe area; water washes and organic washes of up
to four impingers; and some solution washes containing particulate sedi-
ments .
The analyses of these collective runs varied slightly as to
completeness of the characterizations. For the run series No. 44 (including
Runs 44, 49, 50, 55, and 56), a metal analyses was made on all subsamples
except the filtered sediments from the solutions and catches on the quartz
wool. Either total sulfur and/or sulfate sulfur were obtained on specific
subsamples. For the run series No. 46 (including Runs 46, 47, 48, 51,
BATTELLE — COLUMBUS
-------
TABLE 14A. GENERAL DATA FOR RUNS 49, 50, 53, 55, 56
Run
49
50
55
56
53A
53B
54B
Date
10/6/71
10/6/71
11/8/71
11/8/71
10/7/71
10/7/71
10/7/71
Stack
Temp., F
380
350
370
390
Gas Analysis,
percent.
H20
5.8
5.5
6.0
5.5
co2
11 to
12
11 to
11.5
11.5
to
12.5
11.5
to
12
°2
8.0 to
8.5
8.2 to
8.6
7.5
7.6 to
8
8.0 to
8.6
7.4 to
8.2
5.7 to
6.6
L Goks^yr-Ross Procedure
Sample,
m^
0.173
0.225
0.215
0.100
0.111
0.112
milligrams An?
S03
19.9.
15.5
17.2
15.2
16.8
19.8
S02
2942
2493
2486
2892
2854
2585
Total
2970
2511
2504
2907
2871
2617
Particulate Sample
Sample
Location
Electrostatic
Precipitator
inlet stream
Electrostatic
Precipitator
outlet stream
Electrostatic
Precipitator
inlet stream
Electrostatic
Precipitator
outlet stream
Electrostatic
Precipitator
outlet stream,
Port A
Gas
Volume,
m3
0.782
0.694
0.420
0.691
Probe
Temp,
F
255
250
255
255
Filter Box
Temp,
F
250
240
255
250
Electrostatic
Precipitator
outlet stream,
Port D, 53 A and 53 B run simultaneously
Electrostatic
Precipitator
inlet stream,
1.5 hours after 53A and B; Port C
-------
•••••••••^••BWaBMBB
Run «nd
Sample, ,fc.
Port Ion tb)
Run 49
Probe
dry material
H.O wash
organic wail)
Total Probe
Filter*
*1 .„
glassware
tz
Total filter
Inplngers
#1 B20
* 1 organic wa«h
#2 H20 wash
t 2 organic waih
Total for #1 and 12
#3 impinger«)
Run 50
"""""""
Probe
dry material
H2 wash
organic wash
Total Probe
Filters
*1
glassware (O
t2
Total filter
Icpingeri
*1 II 20 wash
#1 organic wash
*2 H20 wash
#2 organic wash
Total for 01 and 02
«3 imnlni-crts)
TABLE UB. DISTRIBUTION OP CHEMICAL ELEMENTS AMONG THE PORTIONS OP PARTICULAR CATCH WITH A DOUBLE PILTER ARRANGEMENT**'
Partlculate Sauo e
Sample Location
Electrostatic precl-
pitator Inlet scream
Electrostatic precl-
pitator outlet •trean
Catch.
ms/m3
2899.
191.4
2.47
3092 .'9
1593.
0.24
4.24
1597.
'
28.6
8.8
37.4
91.1
10.9
102.0
LI
1.74
).74
Bo
0.014
<.0007
<.0003
0.014
0.018
<.0004
<.0004
0.018
<. 00004
<.00004
<.00004
<. 00004
<.0004
<.0004
<.007
<.0001
<.004
<. 00004
<. 00004
<.00004
<. 00004
a
1.20
0.007
0.004
1.21
0.35
0.002
0.004
0.36
0.001
0.0001
<.0001
0.0001
0.0013
0.002
<.004
0.002
0.035
0.001
0.014
0.050
0.007
0.004
0.0004
0.0002
0.011
Na
9.2
1.3
<.04
10.5
9.18
0.39
0.04
9.61
<.018
<.018
<.018
<.018
0.14
<.035
0.14
0.17
0.071
0.035
0.276
<.002
<.002
<.002
<.002
M« ,
15.2
0.25
0.004
15.4
5.30
0.035
0.11
5.45
0.002
0.035
0.004
0.0002
0.041
0.014
0.001
0.015
0.424
0.106
0.283
0.813
0.007
0.00004
0.002
0.0003
0.0093
Al
LSI. 9
0.64
<0.004
152.5
L44.8
0.035
0.35
L45.2
0.002
0.035
0.000
0,000
0.038
0.007
0.004
0.011
5.30
0.004
0.318
5.62
0.0004
<.0004
0.000;
0.0004
o.ooi:
fl
Elemental AnalyslsW, mg/m'
5
SI Total
59.
0.071
0.32
59.
0.007
0.018
0.002
0.004
0.031
0.14
0.014
0.15
0.004
0.004
0.002
0.007
0.014
97.1
2.3C
99.4
21.9
1.55
1.17
24.6
70.63
<.14
70.63
<.14
141.2)
2059.
.
2.01
2.01
18.72
1.17
6.71
26.60
27'.5
0.57
77.7
0.07
05.8
974.
Sulfaea
(7.8)
(4.6)
(1.45)
(0.42)
K
91.8
0.18
O.04
92.0
35.3
0.03:
0.025
35.4
0.004
0.004
0.002
0.002
0.012
<.035
<,005
1.06
<.014
0.07
1.13
<.002
0.002
<.002
0.002
0.004
Ca
91.8
2.5
0.010
94.3
35.3
0.11
0.11
15.5
0.004
0.004
0.004
o.ooo;
0.012
0.106
0.014
0.12
1.24
0.07
0.85.
2.16
0.004
0.0002
0.007
0.0003
0.011
Ti
9.2
0.014
<.004
9.2
3.53
<.002
0.021
3.55
<.0004
<.0004
<.0004
<.0004
C.002
<.004
0.25
<.0001
<.021
0.25
C0004
<.0004
<.0004
<.0004
V
1.91
<.007
<.003
1.91
0.92
<.002
<.004
0.92
<.002
<.002
<.002
<.002
<,002
<.0004
0.007
<.0001
<.004
0.007
<.002
<.002
<.002
<.002
Cr Mn P« Ml
0.92
0.007
<.004
0.92
0.74
<.002
<.004
0.74
<.0004
<.0004
0.0004
0.0004
0.0009
<.002
<.004
0.011
<.0001
<.004
0.011
<.0004
<.0004
<.0004
<.0004
0.92
0.018
<.004
0.92
0.35
<.000>
0.071
0.42
<.ooo:
<.0002
<.0002
<.ooo:
<.002
<.004
0.014
0.004
<.004
0.018
<.0002
<.0002
<.0002
<.0002
459.
o.ia
<.004
459.
180.1
0.007
<.004
180.1
0.002
0.004
O.ooo:
0.002
0.008
0.004
0.007
0.011
2.72
0.004
0.35
3.07
o.ooos
0.002
0.0002
0.002
0.005]
0.60
0.007
<.004
0.61
0.18
<.002
<.004
0.18
<.002
<.004
0.011
<.004
<.004
0.011
00
-------
TABLE
Run and
Sample
Portlon(b)
Run A9 (Cont'd)
Probe
dry material
H20 wash
organic wash
glassware^*'
#2
Total filter
Imp Inge rs
#1 H20 wash
#1 organic wash
#2 H20 wash
#2 organic wash
Total for #1 and #2
#3 lmplnger'8'
Run 50 (Cont'd)
Probe
dry material
H20 wash
organic wash
PI 1 r>n r«
#1
glassware* '
#2
Total .filter
- Implngers
#1 H20
#1 organic
#2 H20
#2 organic
Total for #1 and #2
03 Imp Inge r(8>
14B. DISTRIBUTION OF CHEMICAL ELEMENTS AMONG THE PORTIONS OF PARTICULATE CATCH WITH A DOUBLE FILTER ARRANGEMENT (a) (Confd)
Co
.14
0.007
<.004
.15
.071
<.002
<.004
.071
<.002
<.002
<.002
<.002
<.002
<.004
<.007
<.0004
<.0004
<,021
<.002
<.002
<.002
Cu
.32
C004
.46
0.141
0.018
<:.oo4
0.159
0.0008
0.0004
0.0002
0.00004
0.0014
i.004
0.004
<.04
0.0001
<.04
0.0001
0.0004
O.OC07
0.0004
0.00004
0.0015
As
.32
.07
.04
.176
<.018
C.021
.176
<.002
<.002
<.002
<.002
<.04
C.04
0.071
<.014
<.021
0.071
<.002.
<.002
<.002
<.002
Sr
.32
<.07
004
0.32
0.883
<.018
0.025
0.908
<.002
<.002
<.002
<.002
<.04
C04
0.071
<.014
0.071
0.14
<.002
<.002
<.002
<.002
Zr
0.32
<.07
<.04
0.32
0.353
<.002
<.014
0.353
<.002
<.002
<,002
<.002
<.002
<.004
0.014
<.001
<.014
0.014
<.002
<.002
<.002
<.002
Mo
0.071
<.007
<.004
0.071
0.106
<.002
<.004
0.106
<.002
<.002
<.002
<.002
•
<.002
<.004
<.007
<.001
<.004
<.002
<.002
<.002
<.002
Ag
less Sulfur
1295.
5.4
0.34
1301.
419.6
0.67
0.83
421.1
0.025
0.102
0.015
o.oos
0.150
0.54
0.58
11.44
0.26
2.06
13.76
0.024
0.013
0.016
0.012
0.065
-------
TABLE 14C. DISTRIBUTION OP CHEMICAL ELEMENTS AMONG THE PORTIONS OF PARTICULATE CATCH WITH A SINGLE FILTER ARRANGEMENTS
nn and
.mple
ortion*0'
un 55
Probe
dry material
HjO wash
organic wash
Total Probe
Filter
Implnger
i'l 1(2° wash
#1 organic wash
*2 H2O wash
#2 organic wash
Total for #1 and #2
"3 Traplnger(
-------
.TABLE 14G. DISTRIBUTION OF CHEMICAL ELEMENTS AMONG THE PORTION OF
PARTICUUTE CATCH WITH A SINGLE FILTER ARRANGEMENT^) (Cont'd)
Run and
Sample
Portion* '
Run 55 (Cont'd)
Probe
dry material
H20 wash
organic wash
Total Probe
Filter
Impinger
#1 H20 wash
#1 organic wash
#2 H20 wash
#2 organic wash
Total for i'l and #2
#3 Impinger^S)
Run 56 (Cont'd)
Probe
dry material
H20 wash
organic wash
Total Probe
Filter
Impingers
#1 H20 wash
#1 organic wash
#2 H20 wash
#2 organic wash
Total for #1 and #2
#3 Impinger(s)
Elemental Analysis, (d) mg/m3 (Cont'd)
Ni
0.99
<.004
<.001
0.99
0.14
0.011
0.071
0.082
0.011
Co
0.11
<.002
<.001
0.11
0.025
<.004
<.004
<.004
<.004
<.001
<.001
<,004
<.002
<.002
•C.002
<.002
Cu
0.67
0.071 .
<.0001
0.74
<.071
0.0007
<. 00004
0.0004
0.0011
0.0022
Oll4
0.14
0.28
<.035
0.0001
0.0007
0.0003
0.00004
0.0011
As
CO .35
CO. 04
<0.001
1.31
<.004
<.004
<.004
<.004
0.001
<.014
0.001
0.035
<.002
<.002
<.002
<.002
Sr
0.32
<.04
<.01-
0.32
0.25
<,004
<.004
<.004
<.004
0.001
C.014
0.001
0.071
C.002
C.002
C.002
C.002
Zr
0,67
<.004
•C.001
0.67
<.035
<.004
<.004
<.004
<.004
0.001
•C.001
3.001
Trace
C.002
C.002
C.002
C.002
Mo
0.071
•C.004
<.001
0.071
0.035
<.004
<.004
<.004
<.004
0.001
<.001
0.001
<.004
<.002
<.002
<.002
<.002
Ag
<.18
<.004
<.0001
<.014
<.00004
<. 00004
<.00004
<.00004
0.007
0.0004
0.0074
Trace
<.OOOC4
<.00004
<. 00004
<. 00004
Cd
<.35
<.04
<.001
<.014
<.0018
<.0018
<.014
<.014
•C.004
<.001
<.001
<.001
<.001
Sn
<.35
<.02
<.007.
<.071
<.0018
<.0018
<.001
0.007
0.007
0.11
<0.0002
0.002
0.002
Sb
<.35
<.02
<.007
<.071
<.0018
<.0018
<.003
<.007
<.025
<.001
<.001
<.001
<.001
Ba
1.66
0.007
•C.001
1.67
0.032
<.00004
<. 00004
•C.00004
<.0004
0.007
0.004
0.011
0.003
0.0002
0.004
•C.0002
0.0002
0.0044
Hg
<.011
0.011
Pb
0.32
0.021
-------
111-52
FOOTNOTES FOR TABLES 14A. 14B. AND 14C
(a) Runs 49-50 and 55-56 are pairs of runs collected during simultaneous
sampling at the electrostatic precipitator inlet and outlet at the
Edgewater Power Station, Lorain, Ohio.
(b) The headings dry material, l^O wash, and organic wash used in this
column refer to the dry portion, water portion, and acetone-methylene
chloride portion of the catch, respectively.
(c) This sample was collected simultaneously with the particulate sample
but by a separate train connected to an adjacent port in the duct.
Although isokinetic conditions were not attempted for this train, the
quartz wool plug used to filter particulate from the gas sample was
analyzed for total sulfur and has been included in the total sulfur
catch reported in this column.
(d) The elements were determined largely by OES analysis with confirma-
tory analyses and S analyses by XRF. The symbol < indicates that
the element was sought but not detected in an anlysis having a de-
tection limit as shown for the aliquot used. In some impinger
samples, two values are shown for sulfur: the first and larger
number represents total sulfur, the second represents sulfate
sulfur.
(e) The total catch without sulfur included is intended to show the
distribution of the total catch of element cations of the partic-
ulate among the portions of the rig and facilitate comparison of
Electrostatic Precipitator and outlet streams.
(f) Glassware connecting the first and second filters.
(g) The #3 impinger contained 100 ml of 6 to 7 percent H2C>2 in water
for oxidation of SC>2 remaining in the gas stream. The efficiency
of this single oxidation impinger is untested but, from the per-
formance of the first bubbler in the Gokstfyr-Ross procedure, is
anticipated to be good.
BATTELLE — COLUMBUS
-------
111-53
and 52) total sulfur determinations were obtained on all subsamples and
sulfate sulfur was determined if the amount of total sulfur was sufficient
to warrant it.
Except for a confirming elemental analyses on three quartz
filter deposits by X-ray emission spectroscopy (XRF), all metallic determi-
nations on filters and solutions were obtained by optical emission spectro-
scopy (OES) techniques. Only three filters were analyzed by XRF because
sample amounts were not sufficient to be related quantitatively to pre-
viously prepared standards. The major elements were determined by OES
after dilution of the sample to adjust the elemental concentrations of
the -samples to correspond to elemental concentration of available standards,
The data obtained on the three filter samples by XRF and OES agree within
the limits of accuracies of the two methods. Low amounts of metals found
in the wash solutions indicated XRF analyses would not be useful.
Probe and impinger washes were treated according to the sulfur
analyses desired. For measuring the total sulfur of water washes, an
appropriate portion of the sample was made basic with NaOH to convert all
species of sulfur to the stable sodium salts. For the determination of
sulfur which was collected as a sulfate, the other portion of the sample
was treated with HCl to retain only the sulfate sulfur. Both portions
were subjected to XRF analyses for sulfur content.
For the organic wash solutions, where only total sulfur content
was obtained, the mixture of methylene chloride and acetone was digested
and measured for sulfur by XRF or a wet chemical method.
Figure 20 shows an analytical route outline followed in
characterizing the emission collection runs.
Table 15 shows a more definitive breakdown of sample forms,
treatment, preparation, and corresponding analyses when referred to the
accompanying legends. These include Hg analyses of the probe and im-
pinger water washes and some sulfur checks using wet chemistry.
BATTELLE — COLUMBUS
-------
Total jun Sample
Dry Probe Quartz Wool Quartz Filter Catches Impinger and
Trace-Elements-OES Total Sulfur by XRF
0 Major Elements +
> Sulfur by OES or
"• ' " XRF
m
r If Net Sample is Less Than
r 30 Milligrams
m ,
Major Elements by
0 OES by:
n 1
P (1) Dilution or
£ Sample
2 (2) GE Matrix
0
fj; Sulfur by Wet Chem.
81 or XRF
Dry (PaOs) and
Obtain Sample Wt
Low and Trace
Elements by OES
If Net Sample Wt Is Greater
Than 30 Milligrams
Major Elements
by XRF Organic Washes
Trace
Probe Samples
Volume Measurement
Filter (if Solids Present)
Treat Residue Same M
as Filter Catches M
< 7
in
•P-
Water Washes
and Major Elements by OES
Total S and/or Sulfate S by
XRF or Wet Chem.
FIGURE 20. FLOW SHEET FOR PROCEDURES IN TREATMENT AND
ANALYSES OF STACK EMISSION SAMPLES
-------
TAB1E 15. GENERAL SAMPLE DESCRIPTIONS, TREATMENTS, AND ANALYTICAL METHODS FOR
CHARACTERIZATION OF STACK SAMPLES TAKEN AT EDGEWATER POWER PLANT,
LORAIN, OHIO
44, 49,
49, 50
46. 47.
52
49, 55.
46, 47,
44, 49,
44, 49,
44, 49,
46, 47,
46. 47,
46, 47,
49, SO
49, 50,
49, .50
46, 47.
44, 49.
46. 47,
46. 47
48
48
48
4S
46. 47,
4fi, 47
48
48
48
48
44, 49.
46, 47,
44, 49.
53, 54,
Run(s)
SO, 55,
48. 51,
56
48, 51,
50, 55,
50, 55,
50, 55,
48, 51
48, 51,
48, 51
44, 55.
48, 51,
50. 55.
48. 51,
48
50, 55,
48. 51.
50, 55,
56
52
52
56
56
56
52
56
52
56
52
56
52
56
(46-47-48)
Sample DeacrtoClon
Qu»rt« Filter 1
Quartz Filter 2
Quartz Filter 1
Quartz Filter 2
Probe
Probe
Probe Wash H,0
Probe Wash H,0
Probe Wash 11.0
Probe Wash M20
Pr obe Wash 11,0
Probe Wash lljO
Probe Wash Organic
Probe Wash Organic
Probe Wash Organic
Probe Wash Organic
Inplnger Waih-H20 91 & 2
Irapinger Wash-H,0 *1 & 2 ,
Iraplnger Wash-ll^O 03 & 4
:tr.plnger Wash-HjO «3
Imjlngor Wash-lUO S3
Implnger Waah-HjO 03
Implnger Wash-H20 #4
Iniplnger Wash-HjO Aliquot *1 & 2
Implnger Wash-H20 Aliquot 03 & 4
Implnger Wash-HjO Aliquot #3
Implnger Wash-1120 Aliquot 93
Imylngor Wash-11,0 Aliquot 03
Iraplnger Wash-Dp Aliquot #4
Implnger Wash Organic 01 fc 2
Inplnger Wash Organic #1-4
Quartz Wool
(51-52) Quartz Wool
, Snrnnle Form
Solid Deposit
Solid Deposit
Solid Deposit
Solid Deposit
Loose Powder
Loose Powder
Liquid & Solid
Liquid
Solid
Liquid & Solid
Liquid
Solid
Liquid & Solid
Liquid
Solid
Liquid
Liquid
Liquid.
Liquid
Liquid, & Solid
Liquid
Solid
Liquid
Liquid
Liquid
Liquid & Solid
Liquid
Solid
Liquid
Liquid
Liquid
Solid Deposit
Solid Deposit
Preliminary
Sample
Treatment
1A1, 1B2
1A1, 1B2
1A1, 1B2
1A1. 1B2
1B2
1B2
IF
IK
1B2
IF
IK
1B2
IF
IK
1B2
. IK
IK
IK
IK
IF
IK
1B2
IK
IK
IK
IF
IK
1B2
IK
IK
IK
1B1
1B1
Additional
Desired Sample
2A, B, C
2A, B. C
2A, B. C
2A, B, C
2A, B. C
2A. B. 0
.
2A, B, C
-
2E
*•
2A, B, C
"
2E
2A, B, C
2S
2E
-
2E
-
2E
2E
2E
-
2E •
-
2E
2A, B.'C
2E
22
2E
ICE, D
ICE, D
ICE. D
ICE, D
ID, M
ID, M
1DA2
1C1
1DA2
1-J
1-DA2
1CI
1CI
"
-
*
"
»
-
•
"•
•
"
1DA2
U
IDG
IDG
Analytical
Method and
3-Ala, c
3- Ala, c
3-Ala. e
3- Ala, c
3-A2
3-A2
3-Alb
3-B2
3-Alb
3-B2
3-Alb
3-B2
3-B2
m
"
"
"
"
•
™_
™
"
"
3-Alb
3-B2
3- Bib
3-Blb
Desired
Analy-
sis ..^
2B, E
2B, E
2E
2E
2E
2E
-
2E
-
2F
"
2E
-
2E
2F
2F
m
2F
—
2F
2F
2F
*>o
ZF
"
• 2F
2E
-
-
Addi-
tional
Sample
Treat-
ID. G
ID, G
ID, G
ID, G
ID, G
ID, G
11
1CH2
1J
1CI
1CH2
1012
™
™
*
"
™
"
"
"
**
1CI
Analyt- Analyt-
ical Addl- leal
Method tlonal Method
and Desired Sample and
Tech- Analy- Treat- Tech-
3-Bla
3-Bla
3-Bla 2F 1C, HI, G 3-Bla
3-Bla 2F 1C.* HI, G 3-Bla
3-Bla
3-Bla
3-B2
3-B2
3-C
3-B2 2F 1CS2 3-82
3-B2
3-B2
~
—
—
—
—
"•
"
"
"
3-B2
H
l-l
M
Oi
Ui
-------
TABLE 15. (Cont'd)
•tttn/a^
44, 49,
49, 50
46, 47,
52
49. 55,
46. 47.
44, 49,
44. 49,
44, 49.
46. 47.
46. 47.
46. 47.
49, 50
49, 50,
49, 50
46, 47,
44, 4»,
46, 47,
46, 47,
48
48
48
48
46. 47,
46. 47
48
48
48
48
44, 49.
46, 47.
44, 4».
53, 54,
SO. 55. 56
48, 51,
S&
48, 51.
50. 55,
50. 55,
50, 55,
48, 51
48, 51,
48, 51
44. 55,
48, 51.
50. 55.
48. 51.
48
50. 55,
48. 51,
50, 55.
52
52
56
56
56
52
56
52
56
52
56
52
56
(46-47-48)
Dannie Description
Qusrts Filter 1
Quarts Filter 2
Qusrti Filter 1
Quarts Filter 2
Probe
Probe
Probe Wash H,0
Probe Wash HjO
Probe Hash 11,0
Pr-obe Vash KjO
PC obe Uash HjQ
Probe Wash HjO
Probe Wash Organic
. Probe, Wash Organic
Probe Wash Organic
Probe Wash Organic
Impinger Wath-lljO 01 6 2
Imptnger Wssh-H-0 #1 & 2
Impinaer Kash-IUO *3 & 4
• laipinger Wasli-H20 #3
Inplnger Wash-11,0 *3
Inpingcr Uash-lUO #3
Impinger Ua«h-H20 #4
Inijlnccr Wath-iUO Aliquot 11 & 2
Impinger Wash-H20 Aliquot #344
Xmplngor Wash-)t20 Aliquot il
Inptnger Wash-IIjO Aliquot #3
Impinger Wash-11,0 Aliquot #3
Implngtr Uach-HjO Aliquot 04
Irapinger Wash Organic fl & 2
Iwplngcr Wash Organic #1-4
Quartz Wool
(51-52) Quartz Wool
Sttrnplo Fotffl
Solid Deposit
Solid Deposit
Solid Deposit
Solid Deposit
Loose Powder
Loose Powder
Liquid & Solid
Liquid
Solid
Liquid & Solid
Liquid
Solid
Liquid t, Solid
Liquid
Solid
Liquid
Liquid
Liquid
Liquid
Liquid & Solid
Liquid
Solid
Liquid
Liquid
Liquid
Liquid i. Solid
Liquid
Solid
Liquid
Liquid
Liquid
Solid Deposit
Solid Deposit
Preliminary Additional
Sample Desired Sample
1A1. 1B2
1A1, 1B2
1A1, 1B2
1A1, 1B2
1B2
1B2
IF
IK
1B2
IF.
IK
1B2
IP
IK
1B2
IK
IK
IK
IK
IP
IK
1B2
IK
IK
IK
IF
IK
1B2
IK
IK
IK
IB1
m
2A, B, 0
2A, B, C
2A, B, C
2A, B. C
2A. B, C
2A. B. C
.
2A, B, C
-
2B
-
_
2A. 8, 0
2R
2A. B. C
2E
2B
-
2E
-
2E
2B
2E
-
2E
-
28
2A, B.'C
2B
2E
2B
ICE, D
ICE, D
ICE, D
ICE, D
ID, K
ID, M
1DA2
1CX
1DA2
1-J
1-DA2
1CI
1CI
-
• -
•
•
-
-
-
-
_
-
1DA2
13
IDG
IDG
Analytical
Method and
3- Ala, e
3-Ala, a
3-Ala, e
3-Ala, «
3-A2
3-A2
3-Alb
3-B2
3-Alb
3-B2
3-Alb
3-B2
3-B2
-
-
-
-
-
-
'•
-
-
™
3-Alb
3-B2
3- Bib
3-Blb
Desired
Analy-
2B, B
28, B
28
2E
2E
2E
_
2B
-
-
' 2F
•
.
2B
-
-•
2B
2F
2F
-
2F
-
2F
2F
2F
-
2F
-
2F
2B
-
.
Analyt- Analyt-
Addl- leal Addi- leal
tional Method tional >!othod
Sample and Desired Sample and
Treat- Tech- Analy- Treat- Tcch-
10. G 3-Bla
ID, G 3-Bla
ID, 0 3-Bla 2F
ID. C 3-Bla 2F
ID. G 3-Bla
ID, 6 3-Bla
11 3-82
1CB2 3-B2
1J 3-C
1CI 3-B2 . 2F
1CR2 3-82
1CK2 3-B2
-
-
-
-
-
-
-
-
-
•
1C! 3-B2
1C', HI, C 3-Bla
1C,'- HI. C 3-Bla
1CB2 3-B2
I
tn
-------
111-57
FOOTNOTES FOR TABLE 15
1. Sample Treatment Legend for Table 15.
A. Drying
1. In the presence of Pa 0_
2. A 10 ml aliquot of solution on graphite
B. Weighings
1. Gross (residue plus filter)
2. Net (residue)
C. One-half of sample
D. Aliquot (considerably less than half of the sample)
E. Pulverize and mix filter and residue in shaker
F. Sediments of washings filtered onto quartz filters
G. Mix with quartz matrix and briquette to form solid disk
H. Treat with HC1 for determination of sulfur as sulfate
1. Filter samples: place one-half of filter specimen in an
alumina crucible, fill with 1:1 HC1:HS0 solution, heat
to dryness
2. Water wash solutions: place about one-half of the total
solution in a beaker, add about 20 ml of 1:1 HC1:ILO for
each 100 ml of sample, boil for 1 hour
I. Make water solutions basic with NaOH to convert all lower sulfur
oxides to stable sodium salts
J. Total sulfur determinations in organic washes 5 add 10 ml of
bromine water and 40 ml of 1:1 nitric acid:water to"the beaker
containing the total organic wash solution; let the mixture
stand at room temperature for about 1/2 hour} heat slightly
with care until the methylene chloride has vaporized, and the
nitric acid has digested the acetone; concentrate the remaining
aqueous solution
K. Volume measurement
L. Diluted with germanium to cover major elements.
2. Desired Analyses Legend for Table 15.
A. Semiquantitative for the elements
B. Quantitative analysis for the elements
C. Trace elements (Cd, Be, Pb, As, and Hg)
D. Hg
E. Total sulfur content
F. Sulfate sulfur content.
3. Analytical Method Legend for Table 15.
A. Optical emission spectroscopy (OES)
1. Semiquantitative analyses for metals
a. Quartz standards (range—0.01 to 0.3%, sample weight—
10 mg, expected accuracy — ± 507»).
Note: procedure not applicable for Hg analyses.
BATTELUE — COUUMBUS
-------
Ill-58
FOOTNOTES FOR TABLE 15 (Cont'd)
b. Carbon Standards (conditions above apply)
c. For Hg analyses—using the same standards as above
except that a small orificed cap is used (Boiler Cap
Technique)
2. Quantitative analyses for metals—Germanium matrix technique
(range—0.01 to 10.%; sample weight ratios; 10 mg of
sample, 90 milligrams of Ge, and 100 mg of graphite;
expected accuracy-.- + 20%)
B. X-ray emission spectroscopy (XRF)
1. Solids
a. Quantitative analyses of major elements using quartz
standards and pressed disks; detectable ranee—0.05 to
10. mg; sample weight ratios—25 mg of actual residue
sample, 175 mg of quartz, and 50 mg of pressable
graphite x; expected accuracy ± 5%
b. Estimation of sulfur content in pressed samples of
quartz matrix where sample weights are either too small
for normal sample weight ratios or where the weight
of the actual sample cannot be determined
2. Solutions quantitative analyses of sulfur in wash solutions,
either water or converted organic. Range—0.05 to 5.0 mg
per ml of solution; expected accuracy—±10%.
C. Wet (classical chemistry)
1. Quantitative analyses for total sulfur, as used on some
wash solutions; determined gravimetrically after precipi-
tating as barium sulfate.
BATTELLE — COLUMBUS
-------
IV. DISCUSSION OF RESULTS
OVERALL COMPOSITIONS AND MATERIAL BALANCES
SAMPLING APPARATUS AND PROCEDURES
Probe
Filter
Impinger
Sampling Train Modifications
Simultaneous Sampling
Use of the G/R Coil in Particulate Sampling
Field Trials
Sample Handling
Suggested Sampling Procedures
Inorganic Ash (less S)
Sulfur Compounds
-------
IV-1
IV. DISCUSSION OF RESULTS
The results of several sets of experiments reported earlier have
been brought together so as to reveal overall patterns of mass distribution
in the train. The discussion of these patterns and their significances has
been subdivided into separate topics under the headings:
(1) Material balances and overall composition
(2) Sampling apparatus and procedures
(3) Inorganic ash (less S)
(4) Sulfur compounds.
OVERALL COMPOSITIONS AND MATERIAL BALANCES
Material balances have been made in an attempt to account for
all of the mass of material collected in the various research samples.
These material balances were used to evaluate the effects of sampling-train
modifications on particulate catch, to identify the extent to which the
various chemical species are controlled, and to estimate the apparent over-
all efficiency of collection.
Because of the difficulties in obtaining reproducible operation
and sampling conditions, a considerable amount of uncertainty usually exists
in interpretation of the results from the sampling of particulate in stack
gases. The changes in particulate composition and mass that occur during
sampling and analysis are reflected in the material and lend further un-
certainties to the interpretation. For these reasons, such material balances
have been attempted at several stages in the collection of experimental
data. This portion of the report summarizes the results of those trial
material balances and includes the conclusions that can be drawn concerning
the nature and distribution of the catch.
The degree of success that has been possible in accounting for
the total mass of a group of particulate samples is illustrated by Table
16 for coal-burning operation of the Battelle Multifuel Furnace. For that
furnace, the flue-gas line from which the sample was withdrawn was
BATTELLE — COLUMBUS
-------
TABLE 16. DERIVED QUANTITIES AND TRIAL MATERIAL BALANCES
Sampling Run and
Description Mode Portion
U> (b)
Quarts filter at 280 T f 9 f
F
I
T
Quarts filter at 280 F F 24 P
F.
I
T
Quarts flltar at 400 F C 10 P
F
I
T
Quarts filtfr at 400 F C 26 P
F
I
T
loeonel probe and quarts B 13 P
filter at 230 P F
1
T
Stalnlett steel probe and A 14 P
quarts filter at 230 F F
I
T
Quarts probe and quarts C UP
filter ac 230 F F
I
T
Quarts probe am) quarts C 23 P
filter at 230 F F
I
T
Estimated Product Component*, me
XRF Other
Oxides Ox Idea S102 . H2S04 C 11,0 N20S
(c) (d) (e) (f) (g) (K) (O
37 i 17 87 8 36 t
55 ' 8 23 122 1 — -
12 2 — 60 13 17 3
104 IS 40 269 22 S3 4
75 10 36 42 7
92 13 47 47 1 14 -
2 — — 27 2 30 3
169 23 83 116 10 44 3
471 66 220 62 3 27
ISO 21 79 19 1 4
33 i 10 127 4 102 1
654 92 309 228 8 133 1
180 25 7t 30 18
55 8 40 19 4 3 -
• 2 .. 1 27 2 83 2
237 33 112 76 24 86 2
136 19 84 98 5 1
100 14 53 74 33
2 -- — . 22 12 50 2
238 33 137 194 17 83 3
..192 27 80 115 3 7 -
94 13 47 143 1 61 1
2 — ... 20 7 16 3
288 40 127 278 11 84 4
49 7 22 75 2 22 1
61 8 32 73 - 35
2 - 1 21 4 79 2
112 IS 55 169 6 136 3
93 13 42 60 35
86 12 46 81 1 27
2 -- — 23 2 70 2
181 25 88 169 38 97 2
Totals ag
Sura Collected
191 186
209 202
107 101
507 489
170 189
214 190
64 35
448 414 .
869 990
274 282
231 S33
1424 . 1805 .
324 418 '
129 202
117 142
570 768
343 567
274 235
88 77
705 879
424 501
360 292
48 74
832 867
178 203
209 168
109 133
496 504
243 358
253 214
104 87
600 659
Material
Balanca,
Percent
(V)
103
103
106
104
90
112
183
108
88
97
53
79
78
64
81
74
61
116
114
80
85
123
65
96
88
• 124
82
98
68
118
119
91
Cas
Sampled,
0.463
0.327
O.S16
0.406
0.392
0.601
0.336
0.332
Concentration!
Sum, H2S04, Alh.
8ig/ra3 mg/ni3 Bg/n>3
(1)
1090 581 343
1370 3SS 841
2760 .441 2040
1410 183 940
1800 49S 1040
1390 463 756
1480 S04 S41
1810 S09 885
Distribution
Sun. H2S04, Ash,
Percent
(B) (n) (o)
38 32 37
41 45 54
21 22 9
100 100 100
39 36 44
45 41 5$
IS 23 1
100 100 100
61 36 71
19 8 24
20 56 5
100 100- 100
57 40 72
22 24 27
21 36 1
100 100 100
51 SI 59
36 38 41
13 11
100 100 100
SS 41 66
39 52 34
6 7 . —
100 100 100
38 45 43
38 43 55
24 12 2
100 100 100
42 35 50
40 48 49
18 17 1
100 100 100
t
to
-------
TABLE 16. DERIVED QUANTITIES AND TRIAL MATERIAL BALANCES (Cont'd)
Sampling Run and
Description Mode Porcion
(«) (•) (b)
Hot filter vith coll H 17 P
after filter . F
I
T
Hot filter vith coil H 28 P
after filter F
I
T
Saoe as H. but 2x I 18 P
sampling tine F
I
T
Sane *s H, but 2js I 29 P
stapling time F
I
T
Coil ahead of hot filter K 19 P
F
I
r
Coll ahead of hot filter K 31 P
F
I
T
Estimated Product Components , ne
XRF -Other
Oxides Oxides S102 I^SOj; C H^O NjOs
(c) (d) (e) (£) (g) (h) (1)
44 6 18 28 211
82 11 41 25 3 -
5 1 •- . 21 12 8 3
131 18 59 74 14 12 4
70 10 35 14 3 -- 1
143 20 75 20 ' 9 2 1
6 1 — 22 3 31 3
219 31 110 55 15 33 5
32 It 21 36 711
71 10 35 29 2 10 1
. 1 .... 47 .9 10 5
104 14 " 55 112 18 21 7
158 22 76 22 4 —
163 -23 80 .18 16
81 11 24 228 1 73 2
402 56 180 268 6 79 2
19 3 8 67 2 31 1
"'21 3 8 88 1 65
1 19 4 46 3
41 6 16 174 7 142 4
162 23 89 116 6 —
25 3 11 53 9 22 1
2 ~ — 14 2 38 2
189 26 100 183 17 60 3
Total an
Sum Collected
100 130
162 132
50 43
312 305
132 149
270 274
66 94
469 517
102 116
158 155
72 194
332 465
282 340
291 311
420 628
993 1279
131 131 '
186 118
73 70
390 319
396 409
124 95
58 60
578 564
Material
Balance,
Percent
(k)
77
123
116
102
89
99
70
91
88
102
37
71
83
94
67
78
100
158
104
122
97
130
97
102
Cas
Sampled,
m3
0.336
0.310
0.632
0.685
0.386
0.389
Concentrations
Sum, . H2S04, Ash,
ng/ni3 ng /m3 Bg/o3
U>
930 220 620
1510 181 1160
525 177 276
1450 392 933
1010 451 163
1490 470 310
Distribution
Sura, KjSO^, Ash,
Percent
(m) (-.,) (o)
32 38 33
52 34 64
16 28 3
100 100 100
28 25 32
58 36 66
14 39 2
100 100 100
32 32 33
46 26 67
22 42
100 100 100
29 8 40
29 7 .42
42 85 18
100 100 100
40 33 48
37 51 51
23 11 1
100 100 100
72 63 87
18 29 12
10 8 1
100 100 100
<
u>
-------
IV-4
FOOTNOTES FOR TABLE 16
(a) See Table 7.
(b) P = probe, F = filter, I = impinger, T = total.
(c) The sum' of.the elements analyzed by XRF x 1.54 = oxide equivalent,
S excluded.
(d) The sum of the elements analyzed by OES but not by XRF expressed
as oxides, approximate.
(e) Estimated by assuming Al and Si arise from clay minerals. Here
assume Al:Si = 1.0 and convert the Si to the equivalent weight.
(f) Assuming all S is present as HaSO.t.
(g) The carbon content found by combustion analysis.
(n) Based on the H content found by combustion. Reduce the total H by
a weight equal to 1/12 of the carbon found as an estimate of organic
H and multiply remainder by 9 to express as H20.
(i) The N content is expressed as N205.
(j) The mass of each catch portion as originally measured, see Table 8.
(k) 100 x sum
as collected
percent.
(1) The sum of XRF oxides, other oxides, and Si08 as estimated expressed
per m3 gas sampled.
(m) Calculated from the sums for each portion but the filter catch is re-
duced by omitting the H20 listed for that portion. Expressed as per-
cent of the total catch (Sum).
(n) The distribution of H2S04 as measured.
(o) The distribution of the sum of XRF oxides, other oxides and SiQg as
estimated.
BATTELLE — COLUMBUS
-------
IV-5
approximately 1 and 1/2 inch inside diameter and sample averaging
by traversing was unnecessary and impractical. Some uncertainty arose due
to the necessity for using an estimated value for the amount of Si02 in
the inorganic portion of the catch on the quartz filter. The table shows
that somewhat better consistency is found in the percentage distribution of
a component of the catch among several portions than in the actual mass of
that component found in comparable samples. Some samples show good material
balance in all portions while others do not, mainly because of poor balance
in both probe and impinger portions (see for example Runs 10, 18, 26, 28,
and 29). This scatter is believed to result from the difficulties that
arose during sample collection and in the taking of aliquots of sample resi-
dues for analysis. Later, field experiments showed oxidation and evaporation
(see Table 13) of samples during the usual method of storage in partly filled
bottles was a problem. Aliquots collected in the field and stored in com-
pletely filled bottles showed substantially less sulfate content compared
to conventi.onally stored samples.
Using these material balances, the comparison of the material
distributions for the pairs of runs representing the various sampling modes
yield the following generalizations:
A. Modes F and G differ in filter temperature,
and the use of the higher filter temperature
in G caused a substantially higher fraction
of the sulfur to be collected in the impinger.
B. Modes C and F differ in probe temperature, and,
in these runs where substantial portions of the
catch are found in the probe, no clear effect
of the temperature change is identifiable.
C. Modes C and G differ in both probe and filter
temperatures, and again a substantially higher
fraction of the S is found in the impinger for
the runs (Mode G) with the higher filter (and
probe) temperature.
D. Modes H and G differ in that a coil was used to
collect HgS04 between filter and impinger in
Mode H. Any difference in S03 retention that
may have thus resulted in the impinger portions
is masked by considerable scatter of the data.
BATTEUUE — COLUMBUS
-------
IV-6
E. Modes H and I differ in sampling time and
anount of gas sampled since the intended
half-flow rate for Mode I was not used.
Only one run in Mode I showed the expected
large particulate collection and, in that
case, substantial breakthrough of ash and
S into the impinger is indicated.
F. Modes H and K differ in use of a coil before
the filter to collect H^SC^ and in this case
the additional HgS04 retention has an effect
similar to the use of a cooler filter tempera-
ture in that the S collection in the impinger
is reduced.
Differences in total S catch for these two modes
suggests substantial differences in furnace
performance on two separate days. These dif-
ferences in total S make difficult any determin-
ation of the effect of the coil on the filter
catch.
An important observation arising from these trial balances was
the apparent presence of a substantial amount of water in the catch for
some portions. Following evaporation and desiccation of the several por-
tions of a catch from the coal-fired laboratory furnace, aliquots were
taken for C, H, and N analyses. The results were presented as part of
Table 8 and are repeated here as Table 17. These aliquots generally showed
greater H content than could be accounted for by any assumption of organic
structure using the analytical value for.the carbon content. Accordingly,
the excess H was estimated as HgO and included as such in the material
balance in Table 16. This inclusion of HgO in the balance generally im-
proved the balance for the probe and impinger portions which had been
collected as residues from water solution. On the other hand, the in-
clusion of such water in the material balance for the filter portion appar-
ently is incorrect. The handling of filter portions had consisted mainly
of prolonged storage over Drierite, followed by grinding and collection of
aliquots, and, in spite of a re-desiccation of these aliquots, an appreciable
amount of water was found in these filter samples by C, H, N analyses.
Table 16 shows that this.water content consistently caused the filter catch
material balance to total over 100 percent, but these balances were usually
close if the water was excluded. Accordingly, this water content is believed
BATTELLE — COLUMBUS
-------
IV-7
TABLE 17. SUMMARY OF C-H-N ANALYSES FOR PARTICULATE
SAMPTJ3S (Multifuel Furnace Facility, see
Table 8)
f fl)
Portion , weight percent
Run Element
9 C
H
N
10 C
H
N
11 C
H
N
10 C
H
N
13 C
H
N
14 C
H
N
15 C
H
N
17 C
H
N
18 C
H
N
19 C
H
N
23 C
H
N
P
8.0
4.7
0.4
2.9
3.2
_
17.7
_--,
0.4
9.2
2.4
0.2
4.7 .
^
0.2
2.9
1.0
0.1
1.6
2.6
0.2
2.4
0.3
0.2
7.0
0.6
0.4
1.8
3.5
0.3
34.7
1.3
—
F
1.0
nil
0.1
0.5
0.5
0.1
7.2
5.3
0.3
(b)
(b)
(b)
0.4
3.7
0.1
0.8
6.9
0.3
0.3
3.9
0.1
0.4
0.4
—
1.7
1.3
0.3
0.5
7.2
0.1
1.0
3.1
0.1
I
13.2
3.0
0.9
4.0
11.7
0.2
3.0
8.9
3.9
7.2
8.1
0.5
12.0
6.6
0.7
7.0
2.3
1.0
4.3
9.1
0.6
11.9
1.9
1.0
8.9
11.6
1.4
3.9
5.4
1.1
2.2
8.0
0.6
T
22.2
7.7
1.4
7.4
15.4
0.3
27.9
14.2
4.6
(b)
(b)
(b)
17.1
10.3
1.0
10.7
10.2
1.4
6.2
15.6
0.9
14.7
2.6
1.2
17.6
13.5
2.1
6.2
16.2
1.5
37.9
12.4
0.7
(a)
The letters in this row are column headings -
P = Probe
F = Filter
I = Impinger
T = Total.
'The glass filter used for this sample is not suitable for the C-H-N
analysis train.
-------
IV-8
to be an artifact of -sample handling and not part o' the original catch.
The subsequent investigation of desiccation properties of probe and im-
pinger samples from the cyclone-fired boiler at the Niles, Ohio, plant of
Ohio Edison Company (see Table 11) confirms that some portions are diffi-
cult to dry, especially those that include free sulfuric acid.
Trial balances from some samples collected from the pulverized
coal-fired boilers at the Edgewater Power Station, Lorain, Ohio, are
possible using the data of Table 12. These support the conclusion that
was reached for samples from the Multifuel Furnace, that the analyses of
the impingers usually show greater SO content and less total sulfur than
found by the G/R procedure. The differences between the G/R impinger
analyses are consistent with the anticipated volatilization losses for
impinger samples and with the experimental demonstrations that SO is
apparently being generated by oxidation of the substantial amounts of S03
retained in the impinger.
The material balances for the particulate samples collected from
the large pulverized coal-fired boilers are shown in Table 18. The weights
of metal oxides and sulfuric acid shown in this table are derived from the
analytical determinations of individual elements presented in Tables 14B
and 14C. Inspection of that data had indicated that the trial balances would
not account for substantial portions of the total catch. As the dry probe
portion represented the major portion of the total catches, analyses to correct
the major discrepancies in the trial balances were directed at those portions.
C, H, N analyses were obtained on the dry probe portions of the catch to de-
termine whether the inclusion of carbon contents (coke) would give a better
material balance. These analyses, listed in Table 19, were used to derive
the carbon contents listed in the material balances. For the material bal-
ances, the carbon contents of the filters were assumed to be about the same
as those of the associated probe portions.
BATTECLE — COLUMBU
-------
TABLE 18. TRIAL MATERIAL BALANCES FOR PARTICULATE SAMPLES
FROM EDGEWATER POWER PLANT, LORAIN, OHIO
Run
49
50
55
56
Portion
(a)
P
F
I
T
P
F
I
T
P
F
I
T
P
F
I
T
Sampling
Location
E.P.
inlet
E.P.
outlet
E.P.
inlet
E.P.
outlet
me An3
Metal Estimated
Oxides SiOo S02 C Sum Collection
(b) (c) (dj (e) (f) (f)
2174 200 501 2875 3093
638 1010 74 (152) 00 1874 1597
0.23 282 282 (g)
2812 1010 556 (653)(h) 5031 (g)
0.76 4.1 4.9 37.4
33.4 53 86.4 102.0
0.12 212 212 (g)
34.28 269 303 (g)
1841 275 218 2334 3482
261 220 54 (37) (h) 572 614
0.25 595 595 (g)
2102 220 924 (255) (h) 3501 (g)
9.3 6.6 15.9 21.1
9.5 5.4 6.1 21.0 20.4
0.24 132 132 (g)
19.0 5.4 145 169 (g)
Material
Balance,
Percent
93
117
(g)
101(1)
13
85
(g)
66(D
61
87
(g)
71(i)
76
102
(g)
89(D
Distribution, Percent
Sum S02(d) Oxides
57 36 56.9
37 13 43.1
6 51 0
100 100 100
1.6 1.5 2.2
28.5 19.7 97.5
70 78.8 0.3
100 100 100
67 30 79.3
16 6 20.7
17 64 0
100 100 ICO
9.5 4.5 38.2
12.5 4.2 60.8
78 91.3 i.o
100 100 100
<
VO
-------
IV-10
FOOTNOTES FOR TABLE 18
(a) The letters in this column represent
P = Probe
F = Filter
I = Impinger
T = Total.
(b) Oxides were calculated from the normal stoichiometry for each ele-
ment that was detected. This figure represents the sum of these
calculations, but omitting sulfur oxides. Si could not be determined
on filter samples and is estimated separately, see (c).
(c) For the filter samples, SiO^ was estimated from the Si content of the
accompanying probe catch.
(d) All sulfur found is expressed here as S02; this includes dissolved S02
as collected.
(e) From the C, H, N analyses of Table 5.
(f) Sum represents the total of the estimates of oxides, HaSO , and C.
The column titled Collection, represents the mass of the catch ob-
tained experimentally.
(g) The impinger solutions were analyzed as collected so that no esti-
mates of residue remaining after evaporation were obtained.
(h) Based on the C analysis of the accompanying probe sample for the
same experiment.
(i) Based on sum of probe and filter catch values only.
BATTELLE — COLUMBUS
-------
IV-11
TABLE 19. CARBON, HYDROGEN, AND NITROGEN
ANALYSES OF DRY PROBE MATERIALS,
SAMPLES FROM EDGEWATER POWER STATION
Run
49
55
46
47
48
51
52
Weight;,
C
16.6
8.8
31.3
7.0
37.9
20.1
30.0
percent
H
0.1
0.1
0.6
0.4
0,3
0.2
0.2
N
0.1
<0.1
0.3
0.1
0.3
0.1
0.2
As for the data shown in Table 16, the trial balances of Table 18
reveal substantial separation of the component classes of the overall par-
ticulate in the sampling train. This fact, and the special difficulties asso;
ciated with efficient collection of each component class, leads to the con-
clusion that the total catch is not conveniently nor accurately represented
by a single number. Rather, expression of the total catch in terms of the
amounts of the important individual chemical categories would permit the
total to be judged on the basis of the level of emission for each material
type. At present, total catch may include a summation of several distinct
material categories with different chemistry and physical form
(1) Fly ash: solid particulate partially reacted
with other ingredients of the stack
gas
(2) Sulfur: present as free sulfuric acid, dis-
solved S0a, and combined as sulfate
and sulfite
(3) Organic material: condensates probably ranging
from volatile materials to high-molec-
ular weight materials such as
benzpyrenes and metal-organic complexes
QATTELLE — COLUMBUS
-------
IV-12
(4) Water: combined as hydrate water in the
inorganic catch and water retained
by the free sulfuric acid.
In summary, the conclusions from these trial material balances are:
(1) The assignment of definite oxide composi-
tions to each of the elements (found analyt-
ically) makes possible trial material balances
and these balances show substantial deviation
from 100 percent in some cases. The uncertain-
ties reflected by these trial balances
represent uncertainties that arise from
sampling and sample-handling.procedures as
much as from the inherent analytical diffi-
culties.
(2) The inorganic ash portion is distributed
differently among probe, filter, and im-
pinger than is the sulfuric oxide portion. In
the runs made using the Battelle Multifuel
Furnace, less than 2 percent of the ash was
found by chemical analysis in the impinger,
whereas between 6 and 85 percent of the sul-
fur oxide catch was found there. Other results
obtained from samples collected at Edgewater
Power Station show substantially smaller per-
centages of ash in the impingers. Thus, the
mass distribution of the catch usually does
not correspond to the distribution of any
single component of the particulate.
(3) The composition of the total catch varies too
much to permit calculation of meaningful aver-
ages, but typical compositions for coal-burning
operation based on these trial applications of
the EPA train and its modifications would be:
BATTELLE — COLUMBUS
-------
IV-13
Edgewater Power Station
Battelle Multifuel Furnace Precipitator Precipitator
Inorganic ash (less
sulfur), wt percent
Sulfur, as H2S04,
wt percent
Carbon (coke of organ-
ics), wt percent
Precipitator Inlet
-------
IV-14
SAMPLING APPARATUS AND PROCEDURES
The major problems associated with the sampling and measurement
of particulate emissions, whether to determine emission levels or to evaluate
the performance of particulate control apparatus, are
(1) Generation of sulfate by oxidation of S03
in the impingers. This is also a potential
problem in the probe and filter.
(2) Condensation of H_SO. or adsorption of SO .
3 % fi
(3) Fractionation of the composite of ash,
sulfur oxides, and organic components
during passage of the particulate through
the sampling apparatus.
(4) Retention of water, especially in those
portions of the catch that include free
sulfuric acid.
(5) Probable loss of identity of organic materials
that are collected simultaneously with the
sulfuric acid.
Probe
Comparison runs were made for samples collected using Inconel,
316SS, and Quartz probes. In Tables 8 and 16, the Runs 13, 14, 15, and
23, present data from the comparative runs. The high chromium and nickel
contents of the catches collected with metal probes (Runs 13, 14) present
clearcut evidence of the accumulation of corrosion products in the probe
portion, and, to some degree, in the filter portion at 250 F. In Table
16 the slight increase in sulfur retention by the metal probes, in con-
trast with the quartz probe, may be a reflection of the increased reactivity
of the catch due to these corrosion products.
No clear effect of probe temperature was noted on the nature of
the catch, probably because most of the probe was always at duct tempera- •
ture (about 550 to 750 F) and only the outlet end of the quartz probe was
subject to the temperature control at 250 and 400 F.
BATTELLE — COLUMBUS
-------
IV-15
Analytical results show a varying Si/Al ratio for probe samples,
probably because the Si and Al are contributed by both fly ash and coke
having different Si/Al ratios.
The experimental Si/Al ratios for the probe samples (see Table 14B) were
Run
49
50
55
56
Date
10/6/71
10/6/71
10/8/71
10/8/71
Location
E.P.
E.P.
E.P.
E.P.
inlet
outlet
inlet
outlet
Si/Al Mass Ratio
3.
15.
0.
2.
0
0
31
1
These data show a substantial greater Si/Al ratio for samples collected at
the precipitator outlet than for those collected at the inlet. (As shown
on page 111-44) two different coals were fired on the 2 days during which
these samples were collected. We assume that the difference in coals accounts
for the inlet Si/Al ratio being so greatly different.
Filter
Filter temperatures appear to affect significantly the composition
of the filter catch. A 400 F filter temperature resulted in a substantial
decrease in sulfur oxide retention compared with that for the 250 F filter
(compare Modes F and G in Table 16). Thus, the 250 F temperature of the
filter in Mode F retained 41 and 45 percent of the total sulfur catch,
whereas the 400 F filter temperature used in Mode G caused only 8 and 24
percent of the total sulfur catch to be retained on the filter. Similarly,
the ratio of H3S04/total oxide is less on the filters of Mode G than on
those of Mode F. The decrease in sulfur oxide retention by the hotter filter
is interpreted to indicate that a considerable portion of the sulfur col-
lected on the filter results from reaction of the particulate with the
flue gases, such as condensation of H3S04 and absorption of S0g, and
these reactions are favored at the lower temperature.
BATTELLE — COLUMBUS
-------
IV-16
Thus, the use of filter temperatures above 250 F during sampling is
recommended. The exact temperature to be used for specific studies will have
to be determined by future research.
Three filter materials were considered: a silver filter, MSA
Glass 1106 BH, and Tissuquartz QAO 2500. Only one run was made using the
silver filter and large areas of visible stain on the filter after that
run indicated that the anticipated reaction of the silver with the H3S04
vapors had indeed occurred. Because the problem of sulfur oxide retention
was an important part of the program planned for this project, no further
work was done with the silver filter. The filter is particularly advan-
tageous for electron-probe microanalytical identification of individual
particle composition, but this was not emphasized in this program.
After the chemical analyses for early runs showed that the back-
ground composition for the glass filter would present analytical problems,
detailed comparison of the appearance and performance on the quartz and
glass was made. As the separation of the filter catch from the filter
itself is usually not practical, the chemical analyses of the catch
increases in precision and sensitivity as the filter material becomes more
precisely defined. The silver filter is best from this standpoint because
it can be dried to constant weight easily and impurity elements are of little
consequence in most cases. However, the reactivity of the silver with sul-
fur led to a preference for the relatively inert quartz filter material.
Comparative surface-area, fiber-diameter, and chemical-analysis
data were obtained for the Tissuquartz- and the well-established MSA glass
filters (see Tables 3 and 4). Both of these filter materials suffer from
weighing uncertainties due to their tendencies to pick up water vapor from
the air, and they are more difficult to handle than the silver. The Tissu-
quartz was easily pulverized along with the filter catch in a Polystyrene
WigL Bug (Crescent Dental Manufacturing Company) to produce a homogenized
mass that was relatively easy to analyze for almost all elements of the catch
except Si. Portions of the dry probe catch were analyzed to obtain estimates
of this element.
BATTELLE — COLUMBUS
-------
IV-17
Impinger
The mass of the impinger catch is of overwhelming importance for
those flue gas streams with low ash content, as, for example, oil-fired
operations or the downstream side of a high-efficiency collector. Follow-
ing experiments that showed extensive oxidation of S02 to sulfate in the
impingers, several trials were made to determine what improvements might
be made. Trials of modified impinger solutions (adding particulate and
also HC1) and modified sample handling (storing samples in liquid-filled
bottles rather than in partly filled, oversize bottles) have led to the
conclusion that a substantial part of the sulfate created' from S03 in the
impinger portion may be generated during a storage of the sample. The
improvement (reduced sulfate sulfur) achieved by sample storage in completely
full bottles is illustrated in Table.13.
Sampling Train Modifications
Simultaneous Sampling .
As an alternative to the impinger, several trials of the Goks^yr-
Ross procedure (G/R) were made and the sulfur oxide catch compared with im-
pinger catch made simultaneously. The details of this comparison, discussed
in a later section, are illustrated by the data of Tables 10 and 12.
The comparison shows that, in most cases, substantially less sul-
fate is found by the G/R procedure than by analysis of the impinger catch.
The comparison shows further that the impingers retain on the order of 10
percent or more of the total S02 as measured by the G/R procedure and that
accurate sulfate estimates from impinger catches will depend on the absence
of SOS oxidation during collection, sample storage and evaporation, and
analysis. The experimental results appear to favor use of a separate train
for sulfur oxide measurement.
BATTEUUE — COLUMBUS
-------
IV-18
Use of the G/R Coil in Particulate Sampling
To determine the effect of a separate HgS04 collection on the
sulfur oxide retention by the various components of the EPA sampling train,
a coil like that used in the G/R procedure was adapted for use with the
EPA train and was tried as a means for collecting H3S04 droplets. Results
are shown for trials with the coil either before or after the filter holder
using flue gas from the Multifuel Furnace, see Table 10 for these trials.
The fritted disk normally located at the exit end of the G/R coil was omitted
so that it would not interfere with the filter catch. For analysis of these
runs, the material caught in the coil was included in the probe catch if the
coil was located before the filter, and it was included in the impinger catch
if the coil was located after the filter.
Scatter in the results (see Modes H, I, and K in Table 16) made
it difficult to compare some of these runs. However, Modes K and H, which
differ in the use of the coil before and after the filter, respectively,
showed the anticipated result of a lowered sulfur oxide catch in the im-
pingers when HgS04 was removed previously.
In later work, the G/R coil was used in its conventional form as
a separate method for S0g measurement simultaneously with the operation of
the EPA sampling train. Those trials, see page 111-35 and Table 10, provided
evidence for SO oxidation to sulfate in the impinger portion and so raised
the possibility that removal of the strong mineral acid (e.g., HSS04) be-
fore the impinger could lead to greater S02 solubility in the impinger so-
lution and enhanced opportunity for oxidation to sulfate. Such oxidation
effects could be partially responsible for the scatter obtained. For this
reason another trial of the coil was made during the field sampling at the
pulverized coal-fired boiler at the Edgewater Power Plant. In these trials
the fritted disk was included and both the coil and the disk were increased
in size compared to the normal G/R coil, to accommodate the higher gas flow
rate of the particulate sampler.
BATTELLE — COLUMBUS
-------
IV-19
Field Trials
Five of the experiments on the pulverized coal-fired boilers at
the Edgewater plant at Lorain, Ohio, were planned as an investigation of
the effects of sampling method on the oxidation state and material balance
of the sulfur oxide portion of the flue gas stream. For this purpose
several variations in the sampling train were tried. These included
(1) Use of four impingers, each containing
100 ml H20, instead of the usual 2.
(2) Use of a fifth impinger, containing 100
ml of 6 percent HgO.g, to collect the re-
mainder of the S02 in the gas stream ef-
fluent from the sampling train. The effi-
ciency of this impinger has not been
demons trated.
(3) Use of a coil, with fritted disk as in
the G/R procedure, between filter and
impingers. The coil was held at 162 F.
(4) Use of a second filter instead of the
coil above, held at 110-120 F.
(5) S0_/S0, collections by the G/R procedure
3 3
made simultaneously with each of the above
experimental trials. These collections
were made from' an adjacent port but with-
out benefit of traverse.
These modifications had previously been tried using the Multi-
- fuel Furnace, see page 111-35, or were developed as a result of those trials,
These variations in sampling method showed the following effects:
(1) When extra wet impingers were added to the
sampling train, as much sulfate is found in
the third and fourth wet impingers as in the
first two, and the results imply strongly
BATTELLE — C O U U IVI B U S
-------
IV-20
that the S03 catch in the itnpingers is
largely manufactured by oxidation of dis-
solved S0a in that portion of the appa-
ratus.
(2) The S02 collected in the fifth impinger
(containing Hg02) is large and, in most
cases, is at least 10 times that found in
the water-filled impingers. However, the
sum of all sulfur oxides collected with
.five impingers is not as great as that
found by the G/R procedure. The effi-
ciency of the fifth impinger has not been
demonstrated; the smaller S0_ catch in the
e
impinger train may be due either to the
inefficiency of this H203 impinger, or to
losses that arise during handling of the
other four impinger samples.
(3) The G/R procedure is easy to use and provides
reproducible results. A reinvestigation of
the experimental methods used by GoksjSyr-Ross
under stack conditions is recommended.
(4) The G/R coil inserted between filter and im-
pingers collected a substantial portion of
the total HgSC^ found in that run and was far
more effective than the cooler second filter
tried in its place in one experiment. In the
experiment with the coil, if it is assumed to
be as efficient as has been claimed, the sul-
fate found in the impingers of that run would
seem to have been formed by oxidation in the
impingers.
(5) When two filters were used in series, the mass
loading of the second filter was largely due to
sulfur oxides retained there. The mass of fly-
BATTELLE — COLUMBUS
-------
IV-21
?.sh oxides on the second filter, excluding
sulfur oxides, was small, and the analyses
for individual elements were subject to
large uncertainties. However, these amounts
always exceeded the estimates of oxides
found in the impingers. Table 20 extracted
from Tables 14A and 14B illustrates this
conclusion by comparing the masses of the
elements (not oxides) found analytically in
the several portions of the catch for four
runs.
Sample Handling
In the experiments to determine sulfur (oxide) material balances,
the impingers were analyzed as collected and without the preliminary eva-p -
oration to dryness usually used. The evaporation step, completed by
heating the residues to 212 F, is the source of some potential errors in
interpretation of the mass of the particulate due to
• Oxidation of the S02 retained in the impinger
during its. evaporation from the portion of the
catch
• Water retention or drying
• Thermal alteration or destruction of the or-
ganic portion in the concentrated HS°4 resi-
dues that are generated.
It has been seen that substantial, and apprarently arbitrary
changes, were produced in the weight of particulate by the choice of drying •
conditions. The acetone extraction tried on the dried probe and impinger
residues (see Table 2) was not valid since it removed inorganic materials
and H3S04 that should remain. The weights from this treatment thus should
BATTELLE — COLUMBUS
-------
IV-22
TABLE 20. IMPINGER AND FILTER CATCH FOR SINGLE
AND DOUBLE FILTER EXPERIMENTS, EDGE-
WATER POWER STATION, LORAIN, OHIO
Mass,, (ms /m3)
Run
49 (dual filter,
E.P. inlet)
55 (single filter
E.P. inlet)
50 (dual filter,
E.P. outlet)
56 (single filter,
E.P. outlet)
Portion
Probe
Filter #1
Glassware and filter #2
Impingers #1 and #2
Probe
Filter
Impingers #1 and #2
Probe
Filter #1
Glassware and filter #2
Impingers #1 and #2
Probe
Filter
Impingers #1 and #2
Sulfur(a) Other
99.4 1
21.9
2.7
141 (total)
138 . 1
27.5
298 (total)
2.0
18.7
7.9
106 (total)
3.3
3.0
66.0 (total)
Elements
,301
420
1.5
0.15
,188
168
0.16
0.58
11.4
2.3
0.07
5.8
6.5
0.13
(a) Total sulfur
BATTELLE — COLUMBUS
-------
IV-23
be too low. The slow and more or less continuous loss of water as the •
particulate temperature was raised illustrated the difficulty in recog-
nizing when all free water has been removed and only combined water re-
mains. In the absence of a clear-cut distinction between these two
water forms, the choices among various drying temperatures was arbitratry.
Since free water should be rapidly lost at 212 F, and in view of the
weight changes observed, the drying treatment at 210-220 F is now preferred,
Table 11 shows that the solids recovered from the water layers of
the probe catch were tan and white solids, similar in appearance to the
filter catch. Percentage changes in weight of these solids due to repeated
heating and desiccation were quite small (less than 1 percent). However,
the impinger water layer and the organic wash layers were more sensitive
to drying procedures and showed variations of 10 percent or more. For
these samples from coal-burning operations, the overall variation was not
serious since the mass of probe and filter catch represent most of the
total catch. However, samples with a substantially lower ratio of ash to
sulfur, as from oil-fired operations, would be markedly affected by these
variations in weight and could lead to erroneous estimates of control ap-
paratus performance. Similarly, the charring or pyrolysis of the organic
portion during the evaporation procedure would inhibit characterization of
the metal-organic or toxic components of that portion of the catch.
An improved procedure would seem to be to use separate sample
trains for sulfur oxides and for PNA or for certain hazardous heavy metals.
In this way the filter system would be used to characterize the inorganic
fly-ash particulate only.
Suggested Sampling Procedures
On the basis of observations to date a sampling procedure for
examination of the performance of particulate-control equipment might be
a. Use two quartz filters in series at 350-400 F
or stack temperature, whichever is higher, so
BATTELUE — C O U U IVI B U S
-------
IV-24
as to minimize reactions of sulfur oxide with
the probe and filter catch during sampling and
to back up the first filter if defective. The
point here is that if a filter is defective,
without a back-up there is no way of knowing it,
and gross errors can occur.
b. Determine total SO^ and SO in the gas phase
by a procedure, such as that of Goksj5yr-Ross,
operated simultaneously with the particulate
collection.
c. Use special collection trains for components
such as PNA and certain hazardous metals present
in relatively small concentrations. These
trains would be adapted to the most suitable
absorption or reaction condition for efficient
collection of the component concerned.
d. Attempt material balance by the use of a suffi-
cient scope of analytical methods so that the
entire mass of catch is identified according
to its chemical characteristics.
The OES and XRF estimates of individual metals have been useful
in the present research stages of development of this sampling problem.
In future work, the fly-ash portion of probe and filter portions may be
suitably measured by difference by correcting the total catch for the mass
of sulfur oxides retained, and for the carbon and water contents where
necessary. S03 and SO emissions would be estimated from separate samples
collected for this purpose by a method such as that of Goksjiyr-Ross. The
impinger catch, if obtained, should be used to obtain estimates of indi-
vidual metals and organic species not otherwise retained in the probe or
on the filters.
BATTELLE — COLUMBUS
-------
IV-25
Inorganic Ash (less S)
A number of detailed analyses were made of the composition of
the fly ash so as to detect any conditions that caused fractionation of
this portion of the catch. Such fractionation could be of considerable
importance in those cases where heavy metals are involved and especially
if the fractionation results in exceptionally high losses of those metals
from the sample for any reason.
Differences in particle composition can arise from differences
in origin of the particles in the furnace, and from chemical and physical
adsorption of components of the stack gases as they cool. Such adsorption
processes will have characteristic temperature dependencies that can arise
from differences in volatility or from differences in reactivity of the
vapors with the existing particulate. These processes of particle forma-
tion and reaction, under essentially nonequilibrium conditions in the
rapidly cooling stack gases as they leave the furnace, can be expected
to produce a mixture of particle compositions and size distributions.
In the initial trials, the analyses of the filter deposits from
oil-burning operation of the Multifuel Furnace, see Table 2, were quite
similar for the two samples collected on silver, but the interference of
the background elements of the glass filter caused considerable scatter in
the estimates for the samples from other runs.
Runs 4 and 5, made with silver filter material, show somewhat
greater variation in the proportion of vanadium retained than expected
from the results for other major elements. The two runs differ mainly
in filter temperature, and in this case, the larger vanadium catch was
found on the hotter filter. The variation of vanadium retention with
change in filter temperature requires further investigation before a real
effect can be claimed.
For coal-burning operation, however, variation in particulate
composition with changes in sampling conditions and with changes in fur-
BATTELLE — COLUMBUS
-------
IV-26
nace operation appear-real. This is especially noticeable in Tables 14B
and 14C for samples collected at Edgewater Power Station. Here, differ-
ences in composition for the catch on two filters in series, see Table 14B,
and for samples collected from precipitator inlet and exhaust streams,
see Tables 14B and 14C, imply composition differences for particles of
different size. As the analyses show appreciable coke content for these
samples, see Table 19, it is anticipated that the differences in elemental
composition are, in part, related to these differences in coke content.
Some fractionation may be anticipated to occur in the precipita-
tor .as precipitation efficiencies vary for particles of different size
ranges. Thus, using the total mass of inorganic elements found by analysis,
but excluding sulfur, the precipitator at the Edgewater plant removed 99.5
percent or better of the relatively coarse material caught in the probe and
about 96 to 97 percent of the filter catch portion. For this coal-burning
operation, these two portions constitute 99 percent or better of the total
catch as determined by analyses.
Most of the mass of particulate is contributed by the elements
,Na, K, Mg, Al, Si, Ca, Ti, Fe, S, and presumably 0. However, a number of
heavy elements are presented in amounts that are insignificant compared to
the total mass of particulate but which may need to be considered indivi-
dually for their toxicity contributions. These are identified for coal-
burning operations in the analyses of Tables 14B and 14C, and for residual
oil fuels in Table 2.
The presence of small concentrations of nitrogen in these samples
was revealed by C, H, N analyses, see Tables 9, 17, and 19. Acid extraction
of the N from residues listed in Table 9 was interpreted to mean that the
nitrogen was present as nitrate. Similarly, the small removal of C by
aqueous acid was interpreted to imply the presence of carbonate. In Table
1,the typical flue-gas compositions were examined and it was shown that
even very minor retention of the C03 and NOjj by the particulate could lead
to detectable concentrations of carbonate and nitrate in the particulate.
There has not been time in the program to pursue this aspect of the sampling
BATTELLE — COLUMBUS
-------
IV-27
problem. Further attention should be given to investigating the anion
compositions of the particulate more thoroughly and especially to the role
of halides for their possible effects on particle reactivity, hygroscopic-
ity, and corrositivity.
Sulfur Compounds
The sulfate found in the particulate sample can originate by
condensation of HgSOA or by oxidation of S03 at any one of several stages
in the sampling and analysis procedure. Thus, the sulfate may be the sum of
(1) S03 present in the stack gas
(2) SO generated during sampling (here 40 minutes)
(In general more S03 is generated that is present
in the initial gas sampled.)
(3) SO generated during storage of the filter and
the impinger and probe solutions under air for
transport to the laboratory
(4) S03 generated during evaporation of solutions
and analysis.
In the work that has been described, attempts have been made to
distinguish among these sources and identify the most probable sources of
error due to SO generation.
3 *
The overall process of concern is
SO(gas)5 SO(dissolved)S H SO 5 H++HSO~ * 2H++SO 3 .
H80 HgO VaOg
The first step represents the dissolution of S0g in water, and the solubility
under sampling conditions can be estimated from existing data. Table 21
shows that substantial amounts of S02 should be retained in the impinger
during sampling. Results of one experiment, Run 41, indicate that the amounts
of S0g found in the impinger solution corresponded to saturation at a tem-
perature somewhat higher than actually measured. The presence of sulfuric
acid and dissolved salts in the impinger solution may contribute to this
BATTELLE — COLUMBUS
-------
IV-28
TABLE 21. COMPARISON OF CALCULATED SOLUBILITIES
FOR SOa IN WATER WITH THOSE FOUND IN RUN
41
Solubilities of S02 in H20
as mg S/100 ml HgO
32 F 70 F 86 F
Calculated for gas inlet 20.8 9.7 7.4
to'impingers
Calculated for outlet gas 15.8 7.4 5.6
from impingers
Found in impingers, 2.0 to 5.1; Temperature of
Run 41 gas at outlet = 67 F
^'Solubilities calculated using data summarized by
L. C. Schroeter, Sulfur Dioxide. Pergamon Press,
Inc. (1966), assume solubility varies linearly with
pressure of SOS gas.
discrepancy. Note also that on page 11-12 of this report, it was estimated
that dissolved S0g concentrations of about 3.1 mg/100 cc could be antici-
pated from available data for solubility of S0g in diluted H2S04 solutions.
In any case, the amount of SO dissolved is so large that even minor ten-
dencies for oxidation of the S0g could materially alter the estimate of
total SO,.
3
The fate of this dissolved SO is described by the positions of
(18)
the subsequent equilibria in the above equation. Available literature
indicates that aqueous solutions of SOS contain no detectable HgSO species,
and at pH values less than 4 the ionization of HSO" to form S0~a is negli-
3 3
gible. Thus, essentially only HSO~ ions and dissolved SO are present. On
/1Q\
this basis, Schroeter v/ writes, for 25 C,
for the first ionization constant. Hence, for the dilute solutions of con-
cern here it can be assumed that
BATTEULE — COLUMBUS
-------
IV-2 9
where the brackets represent concentration in moles /liter. If
x = concentration of HSO~ ,
3
a-x = concentration of dissolved S02 , and
a = total concentration of sulfite species,
then
a-x
1.72 x 10-2
LH+]
This relationship has been solved for several sets of values of
a [H+] and the results are presented in Table 22. In this table the con-
centration of dissolved SOS is assumed constant and determined only by the
concentration of S02 in the gas (which is not specified but is held constant)
and by the temperature of the solvent, here 25 C (77 F). Also, the dissolved
species are assumed too dilute to substantially affect the equilibrium between
[SO ] dissolved and [SO "[ in gas. [H"] is put equal to [HSO^], and "a" assumed =
0.020 moles/liter.
TABLE 22. ESTIMATED CONCENTRATIONS FOR SULFITE
SPECIES IN AQUEOUS SOLUTIONS (ASSUMING
THE CONCENTRATION OF DISSOLVED S0g
REMAINS CONSTANT)
Concentration of Species in water,
solution in moles/liter
(1)
(2)
(3)
[s
Water solvent, 0.020
m/1, total sulfite
HC1=0.1 m/1
dissolved
H S04RJ0.001
dissolved
same
same
as (1)
; Cso8]
as (1)
502]
0
0
0
dissolved
.0082
.0082
.0082
CHSO; ]
0.
0.
0.
0118
0015
0114
total
0
0
0
.0200
.0097
.0196
large amount
of acid
small amount
of acid
For the cases considered, the concentration of total sulfurous
acid species can vary by a factor of two or more as a result of the pres-
ence of strong mineral acid in the impinger solution. 0£ somewhat greater
significance is the eightfold range of HSO- in concentrations that might be
produced by acid addition since the kinetics of oxidation are related to
BATTELLE — C O I-U IVI B U S
-------
IV-30
this quantity. Consideration of the rate of oxidation requires closet
attention to a number of other factors that are discussed below.
Many factors can influence the rate of that oxidation and pre-
dictions of such rates are highly speculative. In the laboratory and
using carefully purified solution, the oxidation process has been ob-
served to be a chain reaction process in which the rate of disappearance
of total sulfurous acid species (St) by oxidation is expressed by
-d st , 8
dt
where [HSOl] = concentration of bisulfite ion
[ff*"3 = concentration of hydrogen ion
g - reaction rate constant, the
value is dependent on Og par-
tial pressure for well stirred
sys terns.
The lowered rate of oxidation found at low pH's is the consequence
of the hydrogen ion effect on that rate equation. Hence, as the pH changes
in the range from 4 to 1 a hundredfold decrease in oxidation rate is predicted
even if no change in [HSO^]is produced. This effect is responsible, at
least in part, for the belief that the oxidation ceases at low pH values.
However, for the long periods of storage that are typical for impinger samples
collected in the field, such rates can be very significant. The review and
summary of available data by Schroeter indicates that the rate constant for
the uncatalyzed reaction
HSO^ + 1/203 -? H+ + SO-a ,
is 2.0 x 10~s (moles/liter) Va/second in a well-stirred reactor. In the im-
pinger, the H"1" ion concentration is controlled by the mixture of acid and
ash ingredients which tend to buffer the mixture. Measured pH values have
generally been in the range pH = 2 to 3 after return of impinger samples to
the laboratory. Assuming a pH of 2 for the impinger sample, the above first
BATTEULE — COLUMBUS
-------
IV-31
order reaction corresponds to a half-life (i.e., time for half of the
existing St to be oxidized Lo SO" ) of slightly less than 10 hours. Higher
pH values will yield higher oxidation rates and the absence of stirring
will lead to lower rates of oxygen dissolution and, hence, slower oxida-
tion.
This calculated half-time rate, for an uncatalyzed system, indi-
cated that large amounts of S03 may be produced overnight from typical im-
pinger solutions. Further, shortening of the storage period is not a
general solution to this problem since 10 percent of the existing SO would
be oxidized in about 1 to 2 hours. This means that the SO /S0_ determina-
3 3
tion must be carried out immediately unless oxygen can be rigidly excluded
during storage. The storage of aliquots of samples in liquid-filled bottles
at Edgewater Power Station was done for this reason. A better alternative
might have been to flush the sample with an inert gas, such as Freon, using
the readily available 1-pound containers of that gas.
A further alternative would be to inhibit the oxidation by chemi-
cal means: manitrol, isopropylalcohol, and glycerine have been used for
this purpose.(18) Trials of such materials will require demonstrations that
the inhibition has been suitably completed over long periods of time (days),
and also will require assurances that the addition has not interfered with
analyses planned for other organic elements and organic species.
These considerations made it appear probable that S02 oxidation
would be encountered in the impinger train. Therefore, several trials were
planned to evaluate the degree of this problem. The two series of trials
made using the Multifuel Furnace under coal-fired operation (see Table 10)
and the field trials at Edgewater Power Station (see Table 12) utilized a
number of variations in sampling trains and sampling conditions. These in-
cluded the use of extra impingers containing water, in series with the usual
two, G/R coil insertion in series with the conventional parts of the sampling
train, and comparison collections of SO,, and S03 using the G/R procedure
operated simultaneously with the sampling train. In practically all cases,
BATTELUE — COLUMBUS
-------
IV-32
the amount of sulfate sulfut collected in the impingers substantially
exceeded that found by the G/R procedure. .Further, the sulfate sulfur
remained high even in the third and fourth impingers so that the distri-
bution of sulfate strongly implied that S03 oxidation was occurring in
the impinger train.
The amount of S0_ retained by the cold water in the impinger has
2
been shown to be high and consistent with available solubility data (see
Table 21). Retention of S02 approached 10 times the amount of S0g found.
Thus, the oxidation of only a minor part of the S02 retained by the impingers
would produce the substantially larger sulfate catch found there in com-
parison with the G/R procedure.
The kinetic calculations fo S02 oxidation indicate, however, that
the amount of oxidation that occurs during sample handling and storage may
be more serious than that occurring during the 40-minute sampling period.
The trials of buffered impinger solution at pH = 5.5 (see Table 10),
doubled both the SOg and S03 retention in one series of trials. In that
same series, a comparison run with filter-catch particles added to the im-
pingers further increased the S0g found. A subsequent trial of filter-catch
material from the Edgewater Power Station added to the impinger did not
produce a similar increase, however. The relatively large coke content of
this filter catch material may have been responsible for this difference
but no experimental confirmation is available.
It was anticipated that a highly acid solution would not absorb
much SOg. However, lowering of the impinger pH to between 1.1 and 1.5
produced only a minor decrease in S0g retention over that found in a com-
parison run (see Table 10) but the amount was substantially below that found
in earlier trials using the same furnace (compare Runs 38, 41, 42, 43 in
Table 10). These trials of the effect of lowered pH are thus regarded as
inconclusive.
The effect of changes in filter temperature (see Tables 8 and 16)
are interpreted to mean that a substantial part of the sulfur found on the
BATTELLE — C O t-U M B U S
-------
IV-3 3
filter is the result of condensation and reaction with the gas phase. The
larger sulfur oxide catch at lower filter temperature suggests reactions
following adsorption, and the relatively large fraction of sulfite sulfur
found on the filters of Table 12 indicate that S00 is involved in this ad-
fi
sorption. The discussion of the energetics of adsorption, see page 11-12
et seq., indicates that this is not a surprising result,and the work of
Ketov' ' shows that chemical interaction of gaseous S02 is observed with
alkaline earth oxides, e.g., CaO, over the temperature range 600 to 1800 F.
Thus, the mass of the filter catch may be anticipated to be high as a
result of SO and SO interactions that occur during collection.
3 3
The use of filter temperatures of about 400 F was found to be advantageous
since the sulfur retention was less than that found on filters held at the
more conventional 250 F temperature.
Similar observations of oxidation of S02 have been reported on
glass filters as a result of the alkali content of the glass. ^ A similar
test of the quartz filter has not been made; however, since the quartz does
not contain any appreciable amount of alkali oxides,its activity, if any,
in catalyzing such oxidations may be more dependent on surface contamination
by such material during use, for example, by the fly ash.
BATTEUUE — COLUMBUS
-------
V. RECOMMENDATIONS FOR FUTURE WORK
-------
V-l
V. RECOMMENDATIONS FOR FUTURE WORK
The work done under the present contract has served to identify
several areas for work to supplement the study that has been made of the
chemical composition of particulate. These areas for further research all
relate to the observation that some fractionation of particulate can occur
in the particulate control apparatus as well as in the furnace or boiler
itself, and studies will be needed to establish a basis for measurement of
control for each of the important classes of material. Among these areas
for further research are
(1) Sampling and analysis of small particles that
may penetrate the filter
(2) Sampling and analysis of anions, especially
halides, nitrates, and carbonates
(3) Study of the chemical state of selected heavy
metals and identification of sampling diffi-
culties for each, especially for Hg, Ba, Sb,
Pb, As.
(4) Study of special sampling conditions that are
generated by scrubber-type particulate control
apparatus
(5) Demonstration of a valid sampling method for
emissions from presently used particulate
control equipment in major industries.
It is recommended that a continuation of the study of S03 sampling
should start with a study of G/R procedure under stack sampling conditions.
Ultimately, agreement must be obtained between the analytical results of any
candidate method and known Values. Because a referee sampling method for
SO does not exist, the demonstration must be based on known additions to
stack mixtures. Difficulties revealed in the present program are believed to
indicate that each experimental collection must attempt to identify all sulfur
in the stream being sampled for comparison with the inlet stream content so
that coincidental adsorptions and chemical conversions of SOS to S03 may be
identified. It is suggested that a permeation tube for S03 might be possible
using oleum in thin-walled tubing.
Reinvestigation of the G/R coil leads logically to the question of
small-particle collection, in that the efficiency of the coil must depend on
the efficiency for collection of all H S04 particles that exist in the strqam
under coil conditions. In an extension of the use of that coil, it might be
operated under conditions such that appreciable supersaturation of water
vapor is generated in the stream so that the small particles will serve as
condensation nuclei in the coil. The growth of particle size obtained in
BATTEL.UE - C O U U IVI B U S
-------
V-2
this way may prove to be an efficient basis for collection of the smaller
particles and, if successful, could also serve for estimation of the sul-
fate content of ambient air samples.
Chemical problems relating to the identification of halides
(e.g., chloride ion) nitrates, carbonates, and heavy metal compounds must
be studied from a twofold point of view: first to identify these accurately
in minor concentration in existing particulate, and second to measure emis-
sions of these in gases that also contain particulate for those cases where
the total emission level is important.
Conceivably, these studies will identify methods to supplement
the existing sampling train, and these must be tried under comparative
conditions in industrial applications where particulate and. emission con-
trol equipment is used. In this connection, the special problems generated
by the use of scrubber-type control apparatus should be studied immediately
so that the observations may be used to guide the other research areas
discussed here.
DATTELLE — COLUMBUS
-------
VI. ACKNOWLEDGEMENTS
-------
VI-1
VI. ACKNOWLEDGEMENTS
The authors of this report have been assisted at many points in
the research by helpful comments, suggestions, and active experimental
support in the course of the project. Of particular help were:
Environmental Protection Agency
John 0. Burckle Original Project Officer
James A. Dorsey Program Coordinator
Robert L. Larkin Second Project Officer
D. Bruce Harris Third Project Officer
Battelle-Columbus
David W. Locklin Project Director
John F. Foster Senior Project Leader
William M. Henry Chief, Environmental and Materials
Characterization Division
Richard E. Heffelfinger. .Associate Chief, Environmental and
Materials Characterization Division
Earl J. Schulz Senior Technologist
Clarence McComis Master Technician
Dan L. Chase Senior Chemist
Bernard Paris Research Chemist
John H. Faught Senior Technician
William G. Keigley . . . .Technician
Paul D. Miller Associate Fellow
Paul R. Webb . Senior Technologist
The management of the Ohio Edison Company and their staff at
the Edgewater Plant were significantly helpful in the field investigation.
BATTEULE — C O U U IVI B U S
-------
APPENDIX
A. CONVERSION OF S03 TO H,,S04
B. WET CHEMISTRY METHODS FOR SULFUR COMPOUNDS
C. FILTER EFFICIENCY MEASUREMENT
D. FIELD TEST SAMPLES FROM OTHER EPA CONTRACTORS
E. KINETICS OF SO OXIDATION
2
F. REFERENCES
-------
APPENDIX A
CONVERSION OF SO^ TO
-------
APPENDIX A
CONVERSION OF S0_ TO H SO
*^™^"^^™""™" ' 3 ' ' 2 ™~ 4
To answer the question of how much of the SO in flue gas is
present as HaS04, consider the following equilibrium:*
H30(g)
(S03)
K = — (H SO ) — (moles vapor/ 4)
log10K = Im5j + 0.75 log10T - 0.00057T + 4.086.
(over the temperature range 598 to 756 K)
First calculate position of the above equilibrium at the highest temperature
of interest, 400 F (477 K) assuming the above expression for log1QK is valid;
log10K = 10.480 + 2.009 - 0.272 + 4.086
= -4.657
so at 400 F K = 0.220 x 10'4 = (H30) (S03)
(H2S04)
at 1 atm, 0°C, ideal gas = ~22 4Q ml a = 0.0446 m/ b
at 1 atm, 477°K ideal gas = 273 _ = 0.0256 m/jj
(477) (22.40)
In the Typical Gas Composition, Table 1, page II-2, HgO = 47,,
so HgO = 0.04 x 0.0256 = 1.024 x 10~3 mil.
Similarly, S0g = (3.0 x 10"5 x 2.56 x 10~S - X) = (7.7 x 10~7 - X) moles/ £
^ = X moles /liter
0.220 x ID'4 = (1.02 x 10-3) (7.7 K lO"7 - X)
A
X = 7.51 x 10"7; or 97.8% of S03 is H3S04
*Yost, D.M., and Russell, H.„ Jr., Systematic Inorganic Chemistry,
Prentice-Hall, Inc., N.Y., 336 pp, 1946; see also Bodenstein and
Katayama, Z Phys. Chem., .£9, p 26, 1909.
BATTELLE — COLUMBUS
-------
A-2
Similarly, at 300 F (422 K);
log10K. = 11.846 + 1.970 - 0.241 + 4.086
log10K =-6.031
K - 0.935 x 10"6
273
at 1 atm, 422 K, ideal gas is 0.446 x = 0.289 m/1
= 0.40 x 0.0289 = 1.16 x 10~3 m/
>4 = X m/fl
SO = 3 x 10~5 x 0.0289 - X = (8.66 x 10 - X) m/
«5
0.035 x 10-8 = d.16 x ICT*) (8.66 x 10 - X)
A
(0.935 x 10'6)X = 1.04 x 10"9 - (1.16 x 10~3)X
" i 1* g,^°, = 0.866 x ID'8
1.16 x 10-3
at 300 F HaS04 = 8.66 x 10"7 & 100% of S0g is present.
Hence, S0g essentially exists as H2S04 over the entire range of
conditions of interest, if the time is sufficient to allow equilibrium of
the hydration.
BATTEULE — COLUMBUS
-------
APPENDIX B
WET CHEMISTRY METHODS FOR SULFUR COMPOUNDS
-------
APPENDIX B
WET CHEMISTRY METHODS FOR SULFUR COMPOUNDS
Sulfate Sulfur
Tap sample into beaker, dilute to 200 ml and acidify with 1:1
HC1. Boil to expel HgS. (Boil for 15 minutes at least). Add to 10 ml
of 10 percent BaCl2 solution to boiling solution (add slowly). Let stand
overnight. Filter and wash the BaS04 until Cl free. Place in crucible.
Char and heat to 900 C at least an hour. Weigh BaS04 and calculate as
S04 sulfur.
Total Sulfur
Tap sample into beaker containing bromine and 1:1 HN03 acid.
Let stand 1/2 hour. Heat gently until bromine is expelled (use boiling
rods). Add 20 ml HC1 and heat until HN03 appears gone. Cool sample,
wash down the sides of the beaker, add 10 ml HC1 and take down to near
dryness. Repeat last step. Cool sample, add 10 ml 1:1 HC1 to dissolve
salts and dilute to 200 ml. Bring to a boil and ppt. sulfur by adding
10 ml of 10 percent BaCls solution. Let stand overnight. Filter and
wash until Cl free. Place in crucible. Char and heat to 900 C for at
least an hour. Weigh as BaS04 and calculate as sulfur.
BATTEUUE - COLUMBUS
-------
APPENDIX C
FILTER EFFICIENCY MEASUREMENT
-------
APPENDIX C
FILTER EFFICIENCY MEASUREMENT
by Paul R. Webb
The efficiency of quartz-fiber filters (Tissuquartz 2500 QAO)
and glass fiber filters (MSA 1106B) were compared at a face velocity and
particle diameter which would be comparable to those conditions generally
expected during use.
Figure 1C is a sketch of the sampling arrangements. For each
experiment^ two 47-mm-diameter disks were cut from a common quartz fiber
or glass-fiber filter sheet and the two identical filter disks placed in
series in the primary and backup positions of a 47-mm Millipore filter
holder. The filter holder inlet was inserted through the wall of a 16-
cubic-foot plastic sphere.
The filtering efficiencies of the filters were expected to be
on the order of 99.99 percent, therefore, the sensitivity of the technique
used for detecting particle penetration was necessarily such that about
one part in one million could be detected. This sensitivity was obtained
by generating a fluorescent aerosol, and using a fluorophotometer to de-
termine the relative fluorescent concentrations collected on the primary
and backup filters. Under the conditions used in this study fluorescent
units are proportioned to the mass of fluorescent dye particles in the
sample. Fluorescent units (F.U.) are defined as the product of the fluoro-
photometer reading and total sample volume.
The sampling continued for approximately 4 hours with additional
particles being sprayed into the sphere every hour.
At the end of the sampling period, the two quartz filters were
removed from the holder and each placed into separate beakers. The col-
lected particle samples, were then dissolved in ethyl alcohol and the rela-
tive particle mass on each filter determined with a fluorophotometer.
The fluorescent aerosol was generated from a pressurized con-
tainer having the following constituents:
Freon 12 360 g
Fluorescent Dye 0.1 g
Toluene 1.0 ml.
Atomization of this mixture into air under the experimental conditions
produces an aerosol of solid dye particles. The particle size distri-
bution of this aerosol, shown in Figure 2C, was determined with the aid
of a Battelle cascade impactor.
The fluorescent aerosol was sprayed into the sphere for only a
few seconds in order to maintain relatively low particle concentrations.
High particle concentrations tend to cause agglomeration of the particle
BATTELLE - COLUMBUS
-------
C-2
Aerosol package
16-cu ft
plastic sphere
Flow
Vacuum pump
/—Primary filter
Jj—Backup filter
I // I
(
^—*^
V
L_
i
1
_
-J Critical
flow orifice
3.0. standard
liters per
minute
FIGURE 1C. SKETCH OF SAMPLING ARRANGEMENTS
-------
C-3
33
9(
95
90
BO
IE
cn 70
0)
^ 60
m 50
5 40
*> 30
£L
0}
> ?0
ti
3
£ 10
o
5
2
i
OR
n?
n i
/
X
/
—
/
/
^
/
/
/
i
/
/
/
jf
0.2 0.3 0.40.50.6 0.8 I 2 3 4 5 6 7 8 9 10
Equivalent Particle Diameter, microns
FIGURE 2C. PARTICLE SIZE DISTRIBUTION
-------
C-4
and, hence, changes in the particle size distribution. The con-
centration of particles in the sphere was less than 105 particles/cm3 and
well below the concentration of about 10s particles/cm3 where agglomeration
would be expected to become significant.
Results and Discussion
The efficiency of any filtering media is a function of the face
velocity and particle diameter. A direct comparison of quartz-fiber fil-
ters and MSA 1106 BH glass-fiber filters has been made using the fluorescent-
dye method at a face velocity of 5.21 cm/sec and a mean particle diameter
of 0.55 u-. The average efficiencies at these conditions as can be seen in
Table 1C were 99.9932 percent and 99.9909 percent, respectively.
For comparison with the standard DOP penetration tests, the data
of Lockhart, Patterson, and AndersonC1) for the MSA 1106B glass-fiber fil-
ters have been extrapolated to a face velocity of 5.2 cm/sec where the
efficiency would be expected to be about 99.986 percent. A similar extra-
polation of the data of Smith and SurprenantC2) would indicate an efficiency
in excess of 99.99 percent for the glass-fiber filters at this face velocity.
The differences in these extrapolated efficiencies for DOP particles with
0.3 micrometer diameters and the experimental results presented herein can
be attributed to differences in the particle sizes.
The objective of this study was to obtain comparative efficiencies
and from the results it can be concluded that the quartz-fiber filters are
as efficient as the MSA 1106B glass-fiber filters and can be used effectively
as a filtering media.
(1) Lockhart, L.B., Jr., Patterson, R.L., Jr., and Anderson, W.L., "Character-
istics of Air Filter Media Used for Monitoring Airborne Radioactivity",
NRL Report 6054 (1964).
(2) Smith, W.J., and Suprenant, N.F., "Properties of Various Filtering Media
for Atmospheric Dust Sampling", Proceedings of the ASTM, Vol. 73, pp 1122-
1135 (1953).
BATTELLE — COLUMBUS
-------
TABLE 1C. COMPARISON OF QUARTZ-FIBER FILTERS AND MSA 1106B
GLASS-FIBER FILTERS USING FLUORESCENT DYE METHOD
AT FACE VELOCITY OF 5.21 CM/SEC AND MEAN PARTICLE
DIAMETER OF 0.55 u
Filter Test No.
Quartz 1
Quartz 2
MSA1106B 3
MSA1106B 4
Primary Filter,
F.U.
18.00 x 108
12.95 x 10s
11.122 x 109
19.46 x 108
Backup Filter, Penetration, Efficiency,
F.U. percent percent
0.0012 x 10s 0.0067 99.9933
0.0010 x 10s 0.0070 99.9930
Average 99.9932
0.0012 x 10s 0.0108 99.9872
0.0015 x 108 0.0077 99.9923
Average 99.9908
o
en
-------
APPENDIX D
FIELD TEST SAMPLES FROM OTHER EPA CONTRACTORS
-------
APPENDIX D
FIELD TEST SAMPLES FROM OTHER EPA CONTRACTORS
Table ID shows the analytical results for the A, C, and D
fractions of the samples submitted by W. J. Basbagill at the request
of the Sponsor. The samples are further identified in Table 4 of the
Seventh Monthly Progress Report. The determinations were metallic con-
centration by optical emission spectrography on two of the fractions;
total carbon combustion on all fractions; total acid as H S04; and to-
tal sulfur by the gravimetric method.
Carbon and total sulfur determinations were done only on
fractions containing sufficient quantities of deposits. The total acid
content as ILSO. of the fractions was determined by titrating a water
solution of the residues with standardized NaOH. Acid was found in all
fractions. Several fractions which did not show any weighable residue,
but showed the presence of titrated acid, could not be reported on a
percentage basis. Total sulfur was determined from these solutions by
the gravimetric method.
BATTELLE — C O L U IVI B U S
-------
TABLE ID. CHEMICAL ANALYSES OF FRACTIONS OF FOUR STACK SAMPLES
Sample
Detroit Edison
Detroit Edison
Detroit Edison
Illinois Power
Illinois Power
Illinois Power
Illinois Power
Illinois Power
Illinois Power
Kansas P & L
Kansas P & L
Kansas P & L
No. Fraction
A
C
D
2E A
C
D
2W A
C
D
2A A
C
D
EPA
Container
Number
20
9
21
35
15
32
26
11
25
2a/5
6
5
Weight
Milligrams
0
229.
0
0.6
83.
0.4
21.6
236.
0
26.
85.
14.5
Total
S
(a)
10.
(a)
(c)
17.
(c)
15.
12.
(a)
1.0
22.
16.
Acid,
(b)
30.
(b)
83.
50.
20.
39.
35.
(b)
0.3
67.
45.
as
C
(a)
0.2
(a)
(a)
0.8
. (a)
16.
<0.1
(a)
79.
0.3
15.0
Weight Percent of Total Sample
B Si Mg Fe Al Ag Ca
0.9 .2 <0.01 .02 <0.01 <0.01 .04
.02 .02 <0.01 .02 .02 <0.01 .1
to
(a) Not determined.
(b) Acid, as H2S04, found as milligrams were:
DE A & D - 0.61 mg and .03 mg, respectively,
IP - 2W-D - 0.04 mg.
(c) Determined, but values are suspect because of low quantity of residue.
-------
APPENDIX E
KINETICS OF SO OXIDATION
~~
-------
APPENDIX E
KINETICS OF S00 OXIDATION
Schroeter has summarized the data for the rate of oxidation of
in water solution for the case of a well-stirred solution saturated
with Og at normal air partial pressure. The rate of oxidation of S02 in
water solution is quite sensitive to the presence of other ingredients
that can either catalyze or retard the rate of this reaction. Schroeter's
data can be used to make an order of magnitude estimate of the rate and
amount of S02 oxidation to sulfate in impinger solution both during the
period of sampling of sulfurous gases and during sample storage.
For this purpose we envision the following set of reaction
processes:
(1) S02 (gas) 2 S03 (dissolved)
solvent
(2) S03 (dissolved) 4- HgO (liquid) ? Hj,S03
(3) H_S00 2 H+ + HSO- «» 2H+ + SO -2
& 3 3 3
(4) HSOg + 1/2 Og £ H+ + S04-s .
Reaction 1 represents the equilibrium between gaseous S02 and
dissolved S02 and this equilibrium is assumed to be maintained during the
sampling period. Reaction 2 represents the hydration of SO to form sul-
furous acid in the aqueous solution; this acid ionizes according to the
processes of Reaction 3. Available evidence indicates that aqueous solu-
tions of S03 contain no detectable amounts of unionized acid, i.e., HgSOg,
and according to the equilibrium data at pH values below 4, the amount
of S03~3 is negligible compared to HSCT . -Thus, only HSO^ and dissolved
SO are present in the assumed impinger solution initially.
We shall consider two cases:
1. Oxidation in the impingers during sampling. The
passage of the sample gas through the impinger
maintains the solution in a well-stirred condition,
and the continuing supply of fresh S02 (2000 ppm)
and 02 (at 4%) is assumed to maintain the solution
at saturation levels for the operating temperature.
We shall assume that the liquid in the first im-
pinger is warmer than that in the other impinger
and put it at 25 C for convenience in calculation.
2. Oxidation during storage in the presence of a
moderate amount of air, sufficient to maintain
the 03 level of the stored sample. Assume 25 C,
normal air-02 level.
BATTEUUE — COLUMBUS
-------
E-2
The review and summary of available data by Schroeter indicates
that the rate of disappearance of total sulfur dioxide species (by oxida-
tion)» _ d st
dt
for the uncatalyzed reaction is
- ii . i D«S] (1)
dt
where
[HSQ-] = concentration of HSOr ion
3 o
[H*"] ~ concentration of H+ ion
g = reaction rate constant dependent on 03 concentration in
solution and pH for a well-stirred system.
If we assume g « k [0 ] then
d S,.
- i - g
j>f/3
Then, for a specified Oa concentration and pH, the rate becomes
-l^t-k, [HSOr]
dt 3
which is a pseudo first-order rate expression.
For Case 1, [St] does not vary with time since the fresh gas
is continually replacing any portion of dissolved S0g and HSOg that is lost
by oxidation. In that event
fHSOs 1
d (sulfate) = -d St = g L . 3J dt .
Since we assumed 4% 03, g ^ 2-° x 10 6 Qr g = Q 4 x 10-6 (_S2l£) sec-1,
and [HS(^] at 25 C is expected to be about 3.0 x 10~3 molar for 2000 ppm S02
in the sample gas stream. Thus,
"^ n v in~3
(sulfate)total = 3^° (0.4 x ICT") t .
If sample collection is carried out for 1 hour, t = 3600 seconds, then
,1/2
Sulfate Formed
pH [H+] moles/1 mg S/100 ml
2 0.1 4.3 x 10~B 0.14
4 0.01 4.3 x 10""* 1.4
BATTELLE — COLUMBUS
-------
E-3
We thus expect somewhere between 0.1 and 1.4 mg of sulfur may
be retained in the first impinger of our assumed sample because of sulfate
formation during sampling.
For Case 2, there is no replenishment of S0g during storage so
that as sulfate is formed the concentration of sulfurous acid species in
solution decreases with time. From the above we have
dt
where kx is a function of dissolved 02 concentration, temperature, and pH.
At any time t, and at pH<4, [HSO^] ^ Sfc = a - x, where a is the initial
concentration of sulfurous acid species and x is the amount oxidized at
time t. Then
_dSf-dx.,v, , , a
-£ = = ki (a-x) and so ki t = lne
dt dt a-x
This is the expression for a first-order reaction and it is characterized
by the fact that the time for half of the sulfurous acid species to be
oxidized is a constant regardless of the magnitude of the acid concentration.
-ru „ 0.693 0.105 g 2.0 - in~6
Thus, t1/a - —, similarly , t.'/L0 = ~T . For ^ =
we have
PH frffi tifeo %
2 0.1 5.3 x 103 sec (1.5 hr) 3.46 x 10 sec (9.6 hr)
4 0.01 5.3 x 102 sec (0.15 hr) 3.46 x 103 sec (0.96 hr)
Thus, for solutions obeying this kinetic expression, 10 percent of the dissolved
S0g might be oxidized in 0.15 to 1.5 hr and 50 percent in about 1 to 10 hr
For samples stored in. half-filled bottles, in air, ample 0 is usually present
but the rate will be less than the above in proportion to the actual dissolved
0 level maintained. For 2000 ppm SO the water solution may contain about 10 mg
S as S02/100 ml, i.e., per impinger, at 20 C; and about 20 mg S as S02/100 ml at
0 C, oxidation of even 1/10 of this SO is usually of importance. The above
estimate that 10 percent oxidation may occur in 9 to 90 min with adequate
stirring raises the possibility that sample storage for that period of time
will be sufficient to produce 1 to 2 mg S as sulfate if stored in contact with
air. Agitation of samples during storage will probably not be sufficient to
maintain this rate, but storage periods commonly are much longer, e.g., over-
night and sometimes for days.
The above calculation should not be taken as proof that such
oxidation must occur in each case, but rather that ample opportunity exists
for such errors to arise. First of all, the rate calculations apply to S02
dissolved in pure water and other ingredients are known to both catalyze
and inhibit such oxidation. Further, sample handling methods will strongly
ipfluence the result. A sample stored in a closed impinger will retain most
or all of the dissolved SO in the presence of flue gases containing on the
order of 4 or 5 percent 0 , whereas a sample poured into a graduate for
2
BATTELLE — COLUMBUS
-------
E-4
volume estimations (as in moisture determination) will be exposed to higher
02 saturation levels but may be substantially outgassed of the dissolved
SO originally retained in the impinger liquid. Similar oxidation effects
should occur during sample evaporation in open dishes. Storage of samples
in half-filled or in completely filled bottles will further affect the
result.
Since most of these variations in procedure are not closely regu-
lated, nor is the period of storage between field sampling and analysis,
errors of the magnitude shown by the calculations may occur. These errors
are of great importance since the "manufactured" sulfate can be a substantial
part of the total particulate collected from a low-particulate loading such
as for samples collected downstream of precipitators.
BATTELLE — COLUMBUS
-------
APPENDIX F
REFERENCES
-------
APPENDIX F
REFERENCES
1. Goks^yr, H., and Ross, K., "The Determination of Sulfur Trioxide in
Flue Gases", J. Inst. Fuel 35., 177 (1962).
2. Jacklin, C., Anderson, D. R., and Thompson, H., "Fireside Deposits
in Oil-Fired Boilers. Deposit Location vs Chemical Composition",
Ind. Eng. Chem., 48, 1931, 1956.
3. Miller, P. D., Slunder, C. J., Krause, H. H., and Fink, F. W.,
"Control of Corrosion and Deposits in Stationary Boilers Burning
Residual Fuel Oil", State of the Art Report to Bureau of Yards and
Docks, Navy Department, by Battelle Memorial Institute,May 1, 1963.
4. Green, H. L., and Lane, W. R., Particulate Clouds; Dusts, Smokes,
and Mists, second edition by D. Van Nostrand Company, Inc.,
Princeton, New Jersey, 454 pp. 1964.
5. Parkison, R. V., "The Solubility of Sulfur Dioxide in Water and
Sulfuric Acid", Tappi, 3.9, 517, -1956.
6. Rendle, L. K., and Wilsdon, R. D., "The Prevention of Acid Con-
densation in Oil-Fired Boilers", J. Inst. Fuel, 29, pp.372-380, 1956.
7. Corbett, P. F., and Fireday, F., "The SO- Content of the Combustion
Gases from an Oil-Fired Water-Tube Boiler", J. Inst. Fuel, 26,
pp. 92-106, 1953.
8. Flint, D., Lindsay, A. W., and Littlejohn, R. F., "The Effect of
Metal Oxide Smokes on the SOo Content of Combustion Gases from Fuel
Oils", J. Inst. Fuel, 26, 122-127, 1953.
9. Taylor, R. P., and Lewis, A., "Sulfur Trioxide Formation in Oil
Firing", Proc. Fourth Inst. Congress on Industrial Heating, Group
II, Section 24, No. 154, Paris, Brance, 1952.
10. Matty, R. E., and Diehl, E. K., "New Method for Determining S02 and
SOg in Flue Gas", Power Engineering, 57, 87, December, 1953.
11. Lisle, E. S., and Sensenbaugh, J. D., "The Determination of Sulfur
Trioxide and Acid Dew Point in Flue Gas", Combustion, 36, (7),
12-16, 1965.
12. Brunauer, S., The Adsorption of Gases and Vapors; Physical Adsorption,
Princeton University Press, 1945.
13. Hobson, J. T., Chapter 14. "Physical Adsorption" of The Solid-Gas
Interface, E. A. Flood, Editor, Vol. 1, Dekker, 1967.
BATTELLE — C O L U IVI B U S
-------
F-2
REFERENCES
(Continued)
14. Calderbank, P. H., "Contact-Process Converter Design", Chem. Eng.
Prog., 49. (11), 585-590, 1953.
15. Kodles, B., Regner, A., Vosolobe, J., and Pour, V., "Study of the
Oxidation of S02 on a Vanadium Catalyst in the Region of Internal
Diffusion", IV Int. Congress on Catalysis, Proceedings Symposium
III, 2477-2499, 1968.
16. Uno, T., Fukui, S., Atsukawa, M., Higashi, M., Yamada, H., and
Kamei, K., "Scale-up of a SC>2 Control Process", Chem. Eng. Progress,
66 (1), 61-65, Jan., 1970.
17. Schroeter, L. C., "Kinetics of Air Oxidation of Sulfurous Acid
Salts", J. Pharmaceutical Sciences, 52^ (6), 559-563, June, 1963.
18. Schroeter, L. C., "Sulfur Dioxide", Pergamon'Press, London (1966).
19. Ketov, A. N., Pechkovskii, J. J., and Larikov, V. V., "Interaction
of Calcium Oxide with Sulfur Dioxide", Zhur. Prikladnoi Khimii 44,
1640-1644 (1971) English Edition.
BATTELUE — COLUMBUS
------- |