EPA-600/2-76-086b
October 1976
Environmental Protection Technology Series
FIELD TESTING:
TRACE ELEMENT AND
ORGANIC EMISSIONS FROM
INDUSTRIAL BOILERS
, . . ., ,.
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U. S. Environmental Protection
Agency, and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-086b
October 1976
FIELD TESTING:
TRACE ELEMENT AND ORGANIC EMISSIONS
FROM INDUSTRIAL BOILERS
by
G.A. Cato
KVB Engineering. Inc.
17332 Irvine Boulevard
Tustin, California 92680
Contract No. 68-02-1074
ROAPNo. 21BJSC-046
Program Element No. 1AB014
EPA Project Officers: R. E. Hall and R. A. Venezia
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park. NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
DISCLAIMER
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation
for use.
LIMITATIONS ON APPLICATION OF DATA REPORTED
The data cited in this report are trace element emissions. These
emission levels are suitable for use in estimating mean emissions from in-
dustrial inventories. However, they are not suitable for predicting emissions
from any one boiler or as regulatory limits or trace element emission stand-
ards.
11
-------
ABSTRACT
Sampling of four coal fired industrial boilers was conducted to de-
termine the emissions of 19 trace and minor elements and polycyclic organic
matter (POM). The emissions of the trace and minor elements were related to
total quantities of each element present in the fuel by examining the degree
of mass balance and element partitioning based on fuel input and element out-
put in furnace deposits, fly ash and flue gas vapor. The tendency of volatile
elements for enrichment of finer particulate was examined by analysis of cas-
cade impactor samples.
Measured output of elements classified as high in volatility, tended
to be less than the fuel input, attributed to possible low collection efficiency
of sampling equipment for vapor phase elements. These same elements were found
to be more highly concentrated in the fly ash opposed to furnace deposits and
to have higher concentrations in the smaller particle sizes. Elements classed
as medium or low volatility tended to be more uniformly distributed with res-
pect to both partitioning in the boilers and particle size. Mass output re-
sults for these elements frequently exceeded goal input indicating possible
sample contamination by boiler or sampling system construction materials.
The presence of four specific POM compounds was indicated in the
coal, ashes and stack gases but results were highly variable.
111
-------
IV
-------
CONTENTS
Section Page
ABSTRACT iii
FIGURES vii
TABLES ix
ACKNOWLEDGEMENT xi
1.0 INTRODUCTION 1
2.0 SUMMARY 3
2.1 Material Balance 7
2.2 Particulate Partitioning 9
2.3 Particulate Enrichment 11
2.4 Polycyclic Organic Material (POM) . 11
2.5 Conclusions 15
3.0 BACKGROUND - 17
3.1 Source of Trace Elements 18
3.2 Source of Polycyclic Organic Material 19
4.0 SAMPLE COLLECTION AND ANALYSIS 22
4.1 Sample Collection 24
4.2 Sample Analysis 32
4.3 Particulate Size 47
4.4 Material Balance ' 54
5.0 FIELD TEST, ORIGINAL COLLECTION TRAIN 59
6.0 FIELD TEST NO. 166, MODIFIED COLLECTION TRAIN 62
6.1 Boiler and Fuel Characteristics 66
6.2 Collection Train 71
6.3 Species Concentration - 72
-------
CONTENTS (Continued)
Section Page
7.0 FIELD TEST NO. 169, MODIFIED COLLECTION TRAIN 79
7.1 Test Results 79
7.2 Boiler and Fuel Characteristics 83
7.3 Collection Train 86
7.4 Species Concentrations 88
8.0 SPECIAL REAGENT COMPARISON TEST 95
8.1 Sample Collection Equipment 95
8.2 Analytical Methods and Data 99
8.3 Results 100
9.0 ELEMENT PARTITIONING AND PARTICULATE ENRICHMENT 111
9.1 Element Partitioning 115
9.2 Particulate Size 121
9.3 Particulate Enrichment 132
10.0 .REFERENCES 140
11.0 " CONVERSION FACTORS 144
VI
-------
FIGURES
Number Page
2-1 Field test site locations and location numbers 4
2-2 Material balance 8
2-3 Element partitioning of cobalt, iron and copper 10
2-4 Particulate enrichment of antimony and manganese 12
2-5 Polycyclic organic material (POM) 13
3-1 Trace element concentrations in coal 19
4-1 Trace element collection train schematic 23
4-2 Ice bath portion of collection train showing gas 25
flow direction and bubbler identification numbers
4-3 Schematic diagram of Kapton liner inside probe tube 31
4-4 General procedure for sample treatment and analysis 36
4-5 Required analyses for trace element 42
4-6 Sample processing and analysis flow chart 43
4-7 Required analyses for cascade impactor catch 43
4-8 Detail of one stage and of precutter cyclone 49
for cascade impactor
5-1 Reagent set recommended following initial tests 61
6-1 Material balance test run numbers 166-9, 166-10, 63
and 166-11
6-2 Boiler schematic showing gas flow paths and 69
sampling port location
7-1 Material balance, test run no. 169-3 80
7-2 Solid particulate emissions at baseload, coal fuel 87
8-1 Test boiler installation 97
8-2 Sampling train schematics 98
8-3 Material balances for the three sample collection 108
trains
9-1 Partitioning of moderately volatile elements 116
9-2 Partitioning of moderately volatile elements 117
VI1
-------
FIGURES (Continued)
Number . Page
9-3 Partitioning of highly volatile elements 118
9-4 Partitioning of highly volatile elements 119
9-5 Particulate size distribution, coal fuel and chain 126
grate burner
9-6 Particulate size distribution, pulverized coal fuel 127
9-7 Baseline particulate size distribution, No. 6 129
oil fuel
9-8 Baseline particulate size distribution, No. 6 130
oil fuel
9-9 Particulate enrichment by moderately volatile elements 133
9-10 Particulate enrichment by moderately volatile elements 134
9-11 Particulate enrichment by highly volatile elements 135
9-12 Particulate enrichment by highly volatile elements 136
Vlli
-------
TABLES
Number Page
3-1 Trace Element Concentrations in U.S. Crude Oil 20
3-2 Benzo[a]pyrene Emissions From Coal- and Oil-Fired 21
Boilers in the U.S.
4-1 Trace Species Distribution in Sampling Train 27
4-2 Sample Probe Lining Materials and Selection Criteria 28
4-3 Candidate Filter Material for Trace Specie Sampling 33
4-4 Approximate Elemental Contents of Several Filter Types 34
4-5 Feasibility of Analytical Methods 37
4-6 Analytical Sensitivity of Spectroscopy Techniques 38
4-7 Recommended Chemical Analysis Methods 41
4-8 Carcinogenic Polycyclic Organic Material 47
6-1 Chloride, Fluoride and Sulfate Material Balance 65
6-2 Organics Emissions 67
6-3 Boiler, Coal, and Ash Characteristics 70
6-4 Species Concentration 73
6-5 Species Concentration 75
6-6 Species Concentration 77
7-1 Chloride, Fluoride and Sulfate Emissions 82
7-2 Organics Emissions 82
7-3 Boiler, Coal and Ash Characteristics 85
7-4 Species Concentrations 89
7-5 Species Concentrations 91
7-6 Species Concentrations 93
8-1 Test Boiler and Fuel Data Summary 101
8-2 Trace Species Train Characteristics 102
8-3 Trace Element Content of Coal and Ashes 103
8-4 Trace Element Analysis of Train M 104
8-5 Trace Element Analysis of Train T 105
8-6 Trace Element Analysis of Train S
ix
-------
TABLES (Continued)
Number Page
8-7 Material Balances by Three Calculation Methods 109
9-1 Particulate Size 123
9-2 Cascade Impactor Data Summary 131
9-3 Particulate Size and Enrichment 138
9-4 Particulate Size and Enrichment 139
-------
ACKNOWLEDGMENTS
Th'e author wishes to acknowledge the assistance of Mr. Robert E. Hall,
and Dr. Ronald A. Venezia, the EPA Project Officers, whose direction and
evaluations were of great benefit.
Acknowledgment is also made to the active cooperation and advice
of Mr. W. H. Axtman of the American Boiler Manufacturers' Association and
to the ABMA members who offered a forum for discussion of the industrial
boiler program and constructive criticism. Also of assistance were the
American Petroleum Institute, the American Gas Association, and the Naval
Civil Engineering Laboratory.
Special thanks is due the following organizations who participated
in the industrial boiler program in various ways:
Amoco, Texas City, TX
Babcock and Wilcox Company, Barberton, OH
Baltimore Gas and Electric, Baltimore, MD
Commonwealth of Kentucky, Frankfort, KY
E. I. du Pont de Nemours and Co., Wilmington, DE
Eastman Kodak Company, Rochester, NY
The Firestone Tire and Rubber Company, South Gate, CA
City of Fremont, Fremont, NB
Great Northern Paper Co., Cedar Springs, GA
Industrial Combustion, Inc., Monroe, WI
International Business Machines, White Plains, NY
Keeler Co., Williamsport, PA
Kewanee Boiler Corporation, Kewanee, IL
Lever Brothers Co., Los Angeles, CA
Minnesota Mining and Manufacturing Company, St. Paul, MN
North American Rockwell, Los Angeles Div., Los Angeles, CA
XI
-------
Peabody Gordon-Piatt, Winfield, KN
Pineville Kraft Co., Pineville, LA
City of Piqua, Piqua, OH
Tennessee Valley Authority, Washington, D.C.
Texaco, Inc., Wilmington, CA
Norton Air Force Base, San Bernadino, CA
U. S. Naval Air Station, Patuxent River, MD
U. S. Navy Base, Charleston, SC
Union Electric Company, St. Louis, MO
University of California, Davis, CA
University of California, Irvine, CA
University of California, Los Angeles, CA
University of Oklahoma, Norman, OK
Village of Winnetka, Winnetka, IL
Xll
-------
SECTION 1.0
INTRODUCTION
Concern for trace element emissions has received impetus from
published reports of the widespread atmospheric dissemination of certain
of these substances, especially mercury, as the result of fossil fuel
use. ' ' The elements that commonly are considered to be among the
most toxic are beryllium, fluorine, arsenic, selenium, cadmium, lead, and
(4)
mercury. Under Part 61 - National Emission Standards for Hazardous
Pollutants, Title 40 CFR, the U.S. Environmental Protection Agency has
promulgated national emission standards for mercury, beryllium, and
asbestos. Appropriate future research programs and control technology
may be developed from field testing of trace element and organic emissions
from industrial boilers.
Emissions from industrial boilers are of concern because fuel
combustion in stationary sources accounts for 21% of the total criteria
(6)
air pollutant emissions and a significant part of these emissions may
be trace elements. The objective of the field test and analysis work was
to determine the fly ash particle size distribution and to identify and
quantify the gaseous and particulate trace species emissions from oil
and coal-fired industrial boilers.
Trace species emission measurement was added to an existing two-
year field test program to evaluate the application of combustion modifi-
(7 8)
cations to control the pollutant emissions from industrial boilers. '
Five boilers from the existing program were selected for tests, and the
additional testing of these boilers for trace species emissions yielded
data at minimum time and cost. While criteria pollutant emission testing
took precedence over the trace specie emission testing, it did not sig-
nificantly compromise the trace emission measurements. Results of
criteria pollutant emission tests are given in Reference 8.
-------
The specific trace elements and the compounds selected for
identification and quantification are as follows:
Antimony, Sb Cobalt, Co Tellurium, Te
Arsenic, As Copper, Cu Tin, Sn
Barium, Ba Iron, Fe Titanium, Ti
Beryllium, Be Lead, Pb Vanadium, V
Cadmium, Cd Manganese, Mn Zinc, Zn
Calcium, Ca Mercury, Hg Chlorides
Chromium, Cr Nickel, Ni Fluorides
Selenium, Se Sulfates
The polycyclic organic material (POM) selected for identification
and quantification are as follows:
7,12 Dimethylbenz[a]anthracene
Benzo[ajpyrene
3 Methylcholanthrene
Dibenz [a_,h_] anthracene
Be nzo [ c_] phenan thr ene
Dibenzo [a_,hjpyrene
Dibenzo[a,i]pyrene
Dibenzo[c,g]carbazole
Quantification of the polychlorinated biphenyl (PCB) emissions was
also attempted.
-------
SECTION 2.0
SUMMARY
The objective of the field test and analysis work was to quantity the
gaseous and participate emissions of certain elements and organic compounds
from oil and coal fired industrial boilers. The intent was to utilize existing
trace species sampling concepts that had been developed for the Environmental
Protection Agency by other contractors rather than to develop new ones. The
data obtained were to be used to define the need and requirements for emission
control technology development and for an appropriate research program.
The sampling method employed was a collection train consisting of a
probe, cyclone and filter for particulates, and a liquid impinger train with
eight bubblers. A trap for organics was placed between the third and fourth
bubbler. This trap contained Tenax GC as the adsorbent.
The specie emission testing was done at the six locations shown by
the circled stars in Figure 2-1. However, shortly after the field testing was
started at the first test site, it was found that the concept upon which the
design of the sample collection apparatus was based did not provide for ade-
quate collection efficiency of volatile elements, such as mercury. Additional
development of the collection procedure was needed. Consequently, the program
evolved into two parts: (a) field test of chemicals used to scrub the species
out of the flue gas to determine the best chemicals and collection method, and
(b) field tests using the chemicals and collection method selected as a result
of part (a). The final test of collection chemicals and methods was compre-
hensive and included a material balance of several of the species, as well as
a comparison of the effectiveness of several chemicals.
-------
Locations where trace element measurements were made.
(in order of test) are:
29 - St. Paul, Minnesota
30 - Chicago, Illinois
35 - Piqua, Ohio
31 - Fremont, Nebraska
20 - Rochester, New York (samples collected, not analyzed)
40 - Santa Ana, California
(7 8)
Location numbers as assigned in a previous program
Figure 2-1. Field test site locations and location numbers.
6001-48
-------
Topics of the sections that discuss test results are
as follows:
Field Test Site
Report Section No. Topic Location Number
5 Field evaluation of original 29,30
collection chemicals and procedures
6 Field test of an early collection 35
train design with crushed coal
7 Field test of an early collection 31
train design with pulverized coal
8 Final evaluation of three trains 40
(In lieu of Location 20 sample analysis)
The first trace species test was done at Location No. 29 with oil fuel.
Except for the usual equipment "shakedown" troubles, no sample collection pro-
blems were encountered. However, when the contents of the first three bubblers
were analyzed for mercury, very little was found. After study and discussion
by those involved in the program, it was concluded that the distilled water that
had been used in the first two bubblers was unsatisfactory as a mercury collector
at isokinetic collection flow rates. The collected samples were analyzed no
further.
The second trace species test was delayed pending a resolution of the
mercury collection problem. It was later performed at Location 30 to compare
the trace element collection efficiency of four different reagents.
After review of the analytical results for Locations 29 and 30, and in-
dependent results obtained by MRI, Inc. at other locations under a separate con-
tract with the EPA, it was decided that KVB, Inc. should resume trace species
sampling at Location 35. The boiler at this location was a coal-fired, chain-
grate type. It was also decided that a reagent set developed for the EPA
trace species program by TRW Systems, Inc., Redondo Beach, California, would
be used to trap the mercury. It was expected that the latter reagent set could
trap mercury at the higher pumping rate necessary for isokinetic sample col-
lection.
When the "TRW reagent", ammonium hydroxide, was used at Location 35,
the trace species train repeatedly plugged because the granular Tenax adsorbent
-------
became hygroscopic and stopped the flow of flue gas. Previously/ the Tenax
adsorbent had not picked up water.
After a number of special test runs and consultations, it was concluded
that the cause was the ammonium hydroxide reagents used to trap mercury. The
flue gas at Location 35 had a high sulfur content. Apparently the sulfur in
the flue gas reacted with the ammonium hydroxide in the first two bubblers and
formed an ammonium/sulfur compound that caused the Tenax to become hygroscopic.
As the Tenax became wet, it was no longer granular and the flue gas could not
be pumped through it. The ammonium hydroxide was replaced by a sodium carbon-
ate-based reagent in the first two bubblers to neutralize the sulfur compounds.
With the sodium carbonate reagent, the trace species tests were run with no
further problems.
The KVB field crew then went to Location 31 to test the pulverized coal
fired boiler there for trace species. No collection train problems were en-
countered at Location 31.
The next trace species test was done at Location 20 in conjunction with
combustion modification testing. There were no difficulties. However, the EPA
Project Officer directed that the samples collected not be analyzed because the
analysis of the samples from Locations 31 and 35 indicated that mercury col-
lection still was not completely satisfactory. Also the analytical laboratory
reported that the probe liner that now was part of the train design was high in
organics and possibly could cause the organic analysis to be in error. The
funds saved by cancelling the analysis were to be used for a sixth test.
After discussion with those concerned with trace species collection
train design, the EPA Trace Element Project Officer decided to cancel the trace
species test on oil fuel scheduled at Location 19. Instead, a test was planned
where the three existing toxic species train configurations could be operated
side-by-side and their collection efficiencies compared. This comparison test
was done using the small coal fired boiler located in the KVB, Inc. laboratory
in Santa Ana, California, Location 40.
-------
Investigation of collection chemicals and methods resulted in rejecting
the original concepts as unsatisfactory and perceiving the need for new con-
cepts. The original design of the collection train ultimately proved to be un-
satisfactory for collection of a sample of the volatile elements from the flue
gas when the train was operated at isokinetic flow rates. Development of a new
collection train, designated Source Assessment Sampling System (SASS) train,
has been undertaken by the U.S. Environmental Protection Agency and will be
evaluated for future trace species sampling work, rather than the configurations
that are discussed here.
The degree of success of the collection train designs was determined
by investigating three emission phenomena, (1) material balance, (2) species
partitioning and (3) particulate enrichment.
2.1 MATERIAL BALANCE
The degree of material balance for each of the elements of interest was
evaluated using information from the field tests, and the results are shown in
Figure 2-2. Mass balances were judged to be acceptable when the total quantity
of an element present in the collected ashes and stack gas was within 75 to 125%
of the amount present in the fuel. The basis for this criteria is discussed in
Section 4.1.1.
Most elements that are not expected to volatilize during combustion
tended to be balanced or overbalanced. Barium, iron, maganese and titanium
were overbalanced in some tests. Calcium balanced well in three of four tests.
Cobalt was underbalanced in all tests.
Elements that are expected to volatilize during combustion tended to
be underbalanced, particularly arsenic, selenium, cadmium, zinc, lead and tin.
However, copper and the most volatile element, mercury, were within the ac-
ceptable balance range for two and three of the four measurements, respectively.
Elements with volatility characteristics that are not clearly defined
tended to be balanced or overbalanced. These elements include chromium,
nickel, and vanadium.
-------
ELEMENT COLLECTED / ELEMENT IN COAL, %
ARSENIC
BARIUM
BERYLLIUM
CADMIUM
CALCIUM
CHROMIUM
COBALT
COPPER
IRON
LEAD
MANGANESE
MERCURY
NICKEL
SELENIUM
TIN
TITANIUM
VANADIUM
ZINC
Test
- 169-3
166-9
- 166-10
- 166-11
50
100
150
200
ACCEPTABLE
BALANCE
Figure 2-2. Material balance.
6001-48
-------
The unacceptable balances, particularly of the more volatile elements
were attributed to low collection efficiency of the bubbler liquids for these
elements. A significant portion of the elements that entered the boiler with
the coal evidently were vaporized and passed through the filters and bubblers
without being condensed. In future tests, the degree of balance could be im-
proved by using scrubber chemicals that have a higher collection efficiency
than those used in the three trains and by reducing the velocity of the flue
gas through the bubblers. Isokinetic flow rates were used for the tests in
order to get a particulate catch that was truly representative of the actual
trace element particulate distribution in the flue gas.
2.2 PARTICULATE PARTITIONING
Element partitioning theory predicts that certain elements will tend to
be concentrated in certain parts of the boiler, depending upon the temperature
of the flue gas in that part. Highly volatile elements will tend to be depo-
sited in increasing concentrations on the fly ash as the flue gas carrying it
cools in passing through the boiler. The less volatile elements will tend to
condense early or not vaporize at all, and remain in a more uniform concen-
tration throughout the boiler. Results of the field tests were in accordance
with partitioning theory.
The partitioning of low-volatility iron, moderately volatile cobalt,
and highly volatile copper is illustrated in Figure 2-3. Concentration in
terms of micrograms of cobalt per gram of particulate upstream of the dust col-
lector increased over the cobalt concentration in the coal and then remained
relatively constant in the collector and downstream of the collector. There
was little partitioning between the bottom ash, the fly ash upstream of the col-
lector and the fly ash downstream of the collector. Iron condensed in the
furnace and then maintained a relatively constant concentration in the fly ash
as it moved to the cooler downstream parts of the boiler.
The tendency of the highly volatile elements to deposit on the fly ash
as the flue gas cools is illustrated by the distribution of copper shown in
Figure 2-3. Copper concentration on the fly ash increased steadily as the
flue gas moved to successively cooler parts of the boiler.
-------
2
O
H
I
z
a
u
1
Test No. 169, Location No. 31
M
O
In Coal
Cobalt
Sample Collection Location in Boiler
Furnace
Bottom
Upstream of
Collector
In
Collector
Downstream
Of Collector
/ / f/f / /
§
100,000
50,000
8
Iron
TTTTTTT///////// W////V
o
H
6-
gjj
Z Li
W 5
O :l
Z
8
200
100
Figure 2-3.
Element partitioning of cobalt, iron, and copper by
location in the boiler.
6001-48
10
-------
2.3 PARTICULATE ENRICHMENT
A third aspect of trace element emissions that was studied was particu-
late enrichment. Particulate enrichment theory predicts that those elements
that are volatile at the 1800 K temperature of combustion tend to condense or
absorb on the smaller size fly ash. Because the finer fly ash particles have
a higher surface area per unit mass than do the coarse particles, they will con-
tain a relatively higher concentration of the more highly volatile elements.
Those elements that are less volatile in the combustion zone condense to form
the fly ash upon which the volatiles later condense.
Particulate size results from the field tests were consistent with par-
ticulate enrichment theory, as illustrated in Figure 2-4 by the increase in the
concentration of antimony at the particulajre size decreased. Results for the
less volatile manganese, on the other hand, did not exhibit a well-defined in-
crease in concentration as the particulates became smaller. A detailed discus-
sion of the results of particulate enrichment investigation is presented in
Section 9.0.
2.4 POLYCYCLIC ORGANIC MATERIAL (POM)
The measured emissions of polycyclic organic material (POM) for six runs
during two tests are summarized in Figure 2-5. An open circle indicates an
analysis that found no POM, and a shaded circle an analysis where POM was found.
For example, POM No. 2, benzo[a]pyrene, was found in the coal of Test 166 but
not in the coal of Test 169. Test 169 has two rows of circles under ashes; the
top row indicates when POM was found in the bottom ash and the lower row when
it was found in the dust collector ash. There is only one row of circles for
Test 166 because the bottom and collector ashes were combined within the boiler
and there was only one ash sample for each of the three runs.
The concentrations of the individual POM's are tabulated in Sections
6 and 7. It was found during analysis that the amount of flue gas fly ash that
was collected was inadequate for a quantitative analysis. Therefore, the dis-
cussion that follows is limited only to the POM's that were present in the bot-
tom and collector ashes and in the bubbler liquid.
11
-------
Test No. 169, Location No. 31
1000
500
1
Antimo
ny
X.
^**» ^
JITt:
^=3 Fine
Parties
late
2
O
z
w
2000
1000
0.2 0.5 1.0 2.0 5.0 10
Particulate Aerodynamic Diameter, urn
20
measured at dust collector inlet
measured at dust collector outlet
Figure 2-4. Particulate enrichment of antimony and manganese.
12
6001-48
-------
Test No.
POM-1 166
*7-12
dime thy Ibenz169
[a] anthracene
POM-2 166
*benzo[a]
pyrene 169
POM- 3 166
3 me thy 1-
cholanthrene169
POM- 4 166
*dibenz [a,h]
anthracene 169
Location
Where
Coal Ashes
O 0 0 0
000 0
O
000
000 0
0
0
0 0
0
t
0
0
t
O
0
*
0
0
*
t
0
0
t
0
0
*
0
0
O
Found
Stack
Gas
0 0
000
000
0
000
- M
000
0 0 M
0 : Analysis found no POM
0 : Analysis found some POM
M : Data missing
*Instruments calibrated for isomer shown. The presence
of these isomers relative to other isomers was not
verified.
Figure 2-5. Polycyclic organic material (POM).
6001-48
13
-------
POM No. 1 (7-12 dimethylbenz[a]anthracene) was found in the com-
bined bottom and collector ashes of two of the'runs of Test 166. The quan-
tity was 2.2 grams or less, which was less than the emissions of the other
POM's by a factor of ten or more. It also was found as a vapor in the stack
gas downstream of the collector in one run. This POM was not found at all in
Test 169.
POM No. 2 (benzo[a]pyrene) was present in the coal of Test 166 and also
in the ashes. It. was not present as a vapor in the stack gas. In Test 169 it
appeared in four of the six ash samples and in two of the three stack gas sam-
ples. The concentration decreased at the successive downstream locations in
the boiler which indicates that it was not being partitioned, as were the ele-
ments .
POM No. 3 O'-methylcholanthreneJ was not encountered at all during Test
166, but it was present throughout the boiler in Test 169. The concentration
was successively less at the downstream point in the boiler.
POM 4 (dibenz[a,h]anthracene) appeared in the coal of both tests, but
it appeared in only one ash sample downstream.
With regard to POM 1, 2 and 4, the analytical instruments were cali-
brated for the specific isomers indicated. However, it was not possible to
distinguish isomers by the methods used. Therefore, while POM-2 was assigned
as the isomer benzo[alpyrene, isomers other than the [a] isomer could have been
present in part or in total, and similarly for isomers of the other POM com-
pounds.
It was not possible to identify four desired POM by the methods used.
These four were benzo[c]phenanthrene, dibenzo[a,b]pyrene, dibenzo[a,i]pyrene,
and dibenzo[c,g]carbazole.
14
-------
2.5 CONCLUSIONS
1. Sample collection methods based on modifications of EPA Method
5 sample trains proved to be unsatisfactory for collection of volatile
metals based on mass balance criteria. This was attributed to low collection
efficiencies of the liquid bubblers in the sampling train. Reduction of
sample flows, increased bubbler size or different collection chemicals, should
be investigated as means for improved collection efficiency while maintaining
isokinetic sampling flow rate.
2. Material balances for trace elements varied from less than 10%
to over 200% recovery of the amount of element input in the fuel. Elements
that were most difficult to recover included antimony, arsenic, cadmium,
cobalt, lead, selenium, tin and zinc. An excess amount was recovered most
frequently for barium, beryllium, chromium, manganese, nickel, titanium,
and vanadium. Acceptable recovery was achieved most frequently for calcium,
copper, and iron. While recovery of mercury was judged acceptable in
several tests, the majority of the mercury was found to be contained in boiler
ashes rather than in flue gas vapor. This is contrary to results of other
studies and to the volatile nature of mercury.
3. The results for partitioning of the elements at specific points
in the boilers were consistent with expected partitioning characteristics.
The more volatile elements were deposited in increasing concentrations
on the ash as the combustion gases passed to successively cooler parts
of the boiler system. The less volatile elements were more uniformly
distributed.
4. The results for particle size enrichment were also in agreement
with expected enrichment characteristics. The more volatile elements
tended to concentrate on the smaller size particles, while the less
volatile elements were more uniformly distributed with particle size.
15
-------
5. Polycyclic Organic Matter (POM) was found to be present in
some of the coal, ash and stack gas vapor samples. Samples were analyzed
for four specific POM compounds. All four compounds were found to be
present but not in all samples. The quantity of stack gas particulate
samples collected was not sufficient for determination of POM. The
data for presence and quantity of each POM compound in different parts
of the boiler and for replicate samples were not sufficiently consistent
to allow positive conclusions regarding the formation or emissions of
these compounds.
16
-------
SECTION 3.0
BACKGROUND
Current studies of the trace species air pollutant problem indicate
that fossil fuel combustion in stationary sources such as industrial boilers
may be a major source of these pollutants on a national level. ' The burn-
ing of coal and oil in industrial boilers is of concern because these fossil
fuels contain a variety of potentially toxic trace, elements, such as mercury,
which may be discharged into the atmosphere in significant concentrations. In
addition, the incomplete combustion of coal and oil produces carcinogenic poly-
cyclic organic compounds.
The environmental impact of potentially toxic trace species pollutants
generated by fossil fuel combustion in industrial boilers may be much greater
than the emissions themselves would suggest, because some pollutants are emitted
in the form of vapors or fine particulates which are not collected by conven-
tional pollutant control devices. The small size of trace pollutants may in-
tensify any adverse health effects because particulates of this size penetrate
the natural filters of the human respiratory tract and they may not be expelled.
Moreover, these fine particulates are enriched in trace element content above
what is found in natural aerosols.
Primarily on the basis of volatility profiles, fossil fuel burning is
expected to introduce measurable quantities of arsenic, mercury, cadmium, tin,
antimony, lead and zinc to the atmosphere. A study designed to detect fall-
out from a major coal burner equipped with a precipitator showed that cadmium,
cobalt, chromium, iron, mercury, nickel, titanium and zinc were deposited in
the soil over a 3000 square kilometer area, and with the exception of mer-
cury, the soil deposit correlated with the respective metal concentrations in
the coal.
17
-------
Mercury, being more volatile, was more widely disseminated. Previous
work had indicated that mercury exists primarily as a vapor in the flue gas
and consequently as much as 90% of it bypasses any electrostatic precipitation
field shows <
(15,16,17,18)
control device. ' Other work in the field shows evidence that selenium
and arsenic may follow a similar pattern.
In an analysis of airborne coal fly ash, it was found that ten elements,
(lead, antimony, cadmium, selenium, zinc, arsenic, nickel, chromium, beryllium,
(19)
and manganese) were concentrated in the smaller size particles. Mercury,
although not studied, was expected to follow suit because of its high vola-
(20)
tility. Another study of lead and cadmium reported similar findings. At-
mospheric vanadium, as well as selenium, antimony, and zinc, arising principally
from residual fuel combustion, showed a pattern similar to coal emissions. ' '
The magnitude of fine particulate emissions, their low collection efficiency,
their low deposition rates, and their ease of dissemination in the free atmos-
phere may increase the health risk of receptors.
3.1 SOURCE OF TRACE ELEMENTS
An analysis of the origin of coal and oil fuels found that: (1) nearly
all of the coal consumed by power plants comes from two coal mining regions of
the country: the Appalacian (56%) and Interior Eastern (34%) regions; and (2)
most of the domestic crude oil originates from three statesTexas, Louisiana,
and California. Of the total amount of crude oil processed in the United States,
approximately 85% is produced domestically while the balance is imported. Of
the domestic production, 11% is produced in California, 25% in Louisiana, 34%
in Texas and 30% elsewhere. '
The information available on the trace and minor elements in coal and
oil was analyzed to determine representative concentrations in the fuel from
(23)
each producing region. The results for coal are shown in Figure 3-1. The
figure gives the extremes in average concentration of trace elements for over
90% of the beds within each coal producing region. Other investigators have
reported iron concentrations in coals ranging from 100 yg/g to 30,000 Ug/g and
titanium concentrations ranging from 100 yg/g to 1,000 yg/g. Mercury con-
centrations from 0.012 to 33 yg/g, with an average value of about 0.2 yg/g
18
-------
Element
Antme
Beryllium
Boron
Chromium
Cobolt
Copp«f
Fluoride
Lead .
M«rcury
Tin
Vcnodium
I I 1 1 1 1 ! 1
* x
L £,° * m-C.
" i-
i i i i i i i i
i i i i i i i i
a
' : 1
A
X
i 0
A O Jf
A ^
x
n
=
.
1 1 I 1 I 1 1 I
1 'x ' 1 i 1 1 1
-a ^
_ a A _ 0
. 3,
* x*
HF4
A
a ^ ^H
a
a _^
x
V=T*x a
& a A
v B
x 1 1 1 1 1 1 t I
(116)
0.1
1.0 10.0
Regional concentration, yg/g
100.0
LEGEND
Region
Symbol _
A Appolachion (A)
O Interior-Eastern (IE)
Interior-Western (IW)
O Western (W)
Southwestern (SW)
Stores Included
Pennsylvania, Ohio, West Virginia, Maryland,
Virginia, Eastern Kentucky, Tennessee,
Alabama (and Georgia)
Illinois, Indiana, Western Kentucky, Michigan
Iowa, Missouri, Nebraska. Kansas, Oklahoma,
Arkansas, Texas
Wyoming, Idaho, Utah, Colorado, New Mexico.
Arizona. Washington
Utah, Colorado, Arizona. New Mexico
A Northern Great Plains (N) Montana. North Dakota, South Dakota
X Weighted Average
Figure 3-1. Trace element concentrations in coal.
(23)
6001-48
19
-------
have been found in certain U.S. coals. ' ' ' ' The concentrations of
most trace and minor elements vary considerably from seam to seam with a coal
bed, and the variations within a bed frequently are greater than the differences
between the averages for different beds.
The trace element content of crude oil by region is listed in Table
3-1. <">
3.2 SOURCE OF POLYCYCLIC ORGANIC MATERIAL
Polycyclic organic material (POM) is formed in the combustion of coal,
oil, and natural gas fuels, or, more generally, of any compound that contains
carbon and hydrogen. The amount of POM formed will vary widely since highly
efficient combustion favors very low POM emissions and inefficient burning fav-
ors high emissions. Thus, combustion modifications that reduce the nitrogen
oxides emissions may increase the emissions of POM.
Although the mechanism of POM formation in combustion processes is com-
plex and variable, a relatively clear picture' of the overall reaction has
emerged. Chemical reactions in flames proceed by free-radical paths, and radi-
cal species containing one, two, or many carbon atoms can combine in rapid
fashion at the high temperature attained in the flame front. This pyrosyn-
thesis of pyrolysis products is a function of many combustion variables, and
TABLE 3-1. TRACE ELEMENT CONCENTRATIONS IN U.S. CRUDE OIL
(27)
Element
Antimony
Arsenic
Barium
Manganese
Nickel
Tin
Vanadium
Concentration, Mg/g
Origin of Crude
California
<0.007
<0.007
<0.06
0.018
77
<0.6
48
Louisiana
0.05
0.05
0.09
0.027
4.4
0.5
1
Texas
<0.01
<0.12
<0.14
<0.05
3.3
<1.0
1.9
Weighted
Average
<0.024
<0.08
<0.11
<0.04
16
<0.8
9.0
20
-------
one of the most important is the presence of a chemically reducing atmosphere
which is common in the center of flames. In a reducing atmosphere, radical
chain propagation is enhanced, allowing the buildup of a complex POM molecule,
such as benzo[a]pyrene. Although methane can lead to POM, the formation of
these large molecules is favored by the presence of higher molecular-weight ra-
dicals and molecules in fuels such as oil and coal as opposed to natural gas
Polycyclic organic material is emitted from a vast number of stationary
sources and some urban areas close to significant POM sources are subjected to
high atmospheric POM concentrations. Inefficient combustion of coal in hand-
fired residential furnaces produces large amounts of benzo[a]pyrene as shown
(28)
in Table 3-2. '
TABLE 3-2. BENSO[A]PYRENE EMISSIONS FROM COAL-
AND OIL-FIRED BOILERS IN THE U.S.(28)
Type of Unit
Coal
Hand- stoker residential
furnaces
Intermediate units (chain-
grate and spreader stokers)
Coal-fired steam power plants
Oil
Low-pressure air-atomized
Other
Gross Heat Output,
Joule s/hr
0.10 x 109
63-263 x 109
1.05-2.11 x 1012
0.7 x 109
0.02-22 x 109
Benzo[a]-
pyrene
Emission,
kg/year
381 000
9 000
900
1 800
1 800
6001-48
The low benzo [a]pyrene emission found in power plants, burning crushed or pul-
verized coal, relative to the emissions from the less-efficient hand-stoked
furnaces implies that, for a given fuel, the efficiency with which combustion
takes place is the controlling factor in benzo[alpyrene emissions.
21
-------
SECTION 4.0
SAMPLE COLLECTION AND ANALYSIS
The U.S. Environmental Protection Agency is developing methods to col-
lect and analyze certain metals, gases and organics that are present in the ex-
haust gases from boilers. Trace species testing was added to the Phase II of
the industrial boiler program because it was convenient and cost effective.
A survey of the current trace species measurement and analysis methods
resulted in selection of the method developed by Midwest Research Institute,
Kansas City, Missouri (MRI).
The MRI sample collection concept consisted of a gaseous sample col-
lection "train" made up of a flue gas collection probe, cyclone and filter in
a heated oven and an ice bath that contained eight bubblers and a solid filter
(29)
in series.
In addition to participate and vapor collection with the train, a sample
of the fly ash was collected with a five-stage cascade impactor to determine
the size distribution of the fly ash and the enrichment distribution by trace
species for a given particulate size. For coal fuel tests, a cyclone was ad-
\
ded to the inlet end of the impactor to collect the larger size particulate.
The trace species collection concept of MRI is shown schematically in
Figure 4-1. The sample collection and the gas flow measurement sections (de-
lineated in Figure 4-1) were standard EPA Method 5 items. All of the sample
collection hardware performed satisfactorily. However, problems were en-
countered with the chemicals that were used in the bubblers to scrub out the
metals and organics.
22
-------
Sample Collection
Section
Cyclone
Particle!
Collection
I Section I
Vapor Collection Section
3C.
Filter
/"Tenax Adsorbent
-Bubblers
S* .* » » » > »
Ill/ill
-
J
"
J
II
II
'I
HI III
1!
li
\
III
T-
To Pump ^
Figure 4-1. Trace species collection train schematic.
6001-48
23
-------
4.1 SAMPLE COLLECTION
4.1.1 Particulate and Vapor
The solid particulate and gaseous sample collection and gas flow mea-
surement sections of the collection train used were from an Emission Parameter
Analyzer Model T1050 manufactured by the Joy Manufacturing Company for use in
a standard EPA Method 5 particulate concentration measurement. The impinger
box was replaced by a separate chilled water bath large enough to accommodate
eight, rather than four, bubblers. A glass bulb was added between the'third
and fourth bubbler in which Tenax, a gas chromatograph column packing material,
was placed to act as an organic adsorbent.
It had been found during prior testing that the temperature of the
water bath must be maintained near 270 K (32°F) so that the stack gas would be
(29)
cooled to less than 300 K (60°F) by the time it reached the Tenax adsorbent.
It is necessary to collect the organic sample at less than 300 K to obtain
acceptable adsorption efficiencies. A mechanical refrigerator unit that was
capable of reducing the temperature of the water in the ice bath chest to less
than 275 K (41°F) was used.
Figure 4-2 is a schematic drawing of the top of the ice bath and shows
the physical arrangement of the bubblers and the flue gas flow within the train.
Bubblers were used in place of impingers in order to minimize flow resistance.
Sufficient material must be collected so that the sample can be analyzed
accurately and precisely. Based on a need for determining source stream con-
centrations in the neighborhood of 60 yg m (2.6 x 10 grains ft ) the sample
collection rate was maintained near 0.0196 to 0.028 m min (0.7 to 1.0 ft
min ) to collect 100 yg of the species of interest. Collection time depended
upon the flow rate necessary for isokinetic sampling and was from 60 to 90
minutes.
Those species which existed as solid particulates at stack temperatures
were collected in the glass cyclone or on the filter located in the heated
oven. The volatile or submicron size species which behaved as gases were
scrubbed out by the liquids in the bubblers. Different pollutants were col-
lected by different sections of the train, depending upon the fuel composition,
24
-------
Bubblers
Tenax
Holder
60 cm
<^J 1 rrcm Solid
Particulate
Collection
Section
To Gas Flow Measurement Section
Numeral on bubbler is the numeric designation of the individual
bubbler.
Figure 4-2 . Ice bath portion of collection train showing gas
flow direction and bubbler identification numbers.
25
6001-48
-------
and an estimate of where the metals and gases were expected to be collected
(29)
in the train is given in Table 4-1. Table 4-1 indicates that 90% or more
of the mercury and other volatile elements would be collected in the oxidation
bubbler portion of the vapor train.
POM's were expected to be concentrated in the first three bubblers and
in the Tenax adsorber. Most of the identifiable carcinogenic POM's, in tests
(29)
to date, have been recovered with the solid particulates. The Tenax ad-
sorber was located in the train to extract all of the remaining POM's from the
sample stream before it passed on to the bubblers that contained the oxidant.
A test normally consisted of three runs under identical boiler settings.
A statistical analysis performed by MRI indicated that for three test runs a
mass balance within j^ 25% tolerance for non-volatile elements could be achieved
at a 90 percent confidence level. The mass balance tolerance was based on esti-
mates of the expected input/output stream flow and concentration standard de-
viations .
4.1.2 Probe Material Selection
Standard stack sampling probes are heated tubes equipped with an outer
protective sheath and a proper size nozzle for isokinetic sample collection.
Materials of construction are typically 316 stainless steel, glass or quartz.
Teflon often is used as a lining material. However, a Kapton liner may also
be used. The relative merits of these materials as probe materials have been
reported and are listed in Table 4-2. Aluminum probes have the lowest
operating temperature limit while quartz has the largest temperature range.
Above approximately 523 K (480°F) aluminum and many of its alloys rapidly
lose their physical strength; however, their light weight makes them attractive
for particulate sampling in large diameter stacks.
Both glass and quartz are chemically inert except for the possible in-
teraction of fluorine in the form of hydrogen fluoride, and neither would con-
tribute to the contamination of trace element sampling. For sampling source
streams that require a probe more than two meters long, however, glass and
quartz become impractical because of breakage by vibration. Carbon steel,
26
-------
TABLE 4-1. TRACE SPECIES DISTRIBUTION IN SAMPLING TRAIN
(29)
Species
Group A
Al, B, Be, Ca
Cr, Cu, Fe/ Mn
Mo, Ni, Si, Ag
Sn, Ti, Zn
Group B
As, Sb, Se
Group C
Cl, F, Pb
Group D
Cd, Hg
Solid Particulate
Collection Section
Nozzle
and
Probe
<10%
5%
<5%
<5%
Cyclone
and
Filter
70%
45%
20%
<5%
Vapor
Collection Section
Moisture
Condensing
Impinger
20%
45%
60%
<5%
Oxi dative
Impingers
<5%
5%
10%
>90%
6001-48
27
-------
TABLE 4-2. SAMPLE PROBE LINING MATERIALS AND SELECTION CRITERIA
(30,31)
Probe Material or Liner
Aluminum
Carbon steel
Stainless steel (316)
Titanium
Glass (Borosilicate)
Quartz
Kapton liner
Teflon liner
Operating
Temperature, K (°F)
523
>1173
1173
>1273
1093
>1773
(480)
(1650)
(1650)
(1830)
(1500)
(2730)
723-773 (840-930)
<543
(520)
Cleaning
Difficult
Difficult
Difficult
Unknown
Difficult
Difficult
Very easy
Easy
Handling
Light weight
Heavy
Heavy
Light
Fragile
Fragile
Satisfactory
Satisfactory
Contamination/
Sample Alteration
Depends on alloy
Yes, possible Mn and
other elements
Cr, Ni, other
stainless elements
Depends on alloy
No contamination,
possible loss of F
No contamination,
possible loss of F
None*
None
M
CD
*Although the references indicate no contamination with Kapton, Kapton may result in organic
contamination.
6001-48
-------
while having a high temperature limit, is readily corroded by many stack gases
and is a source of manganese and other elements. It has not been widely used
for stack sampling. Teflon coated probes are excellent for trace element samp-
ling from a contamination standpoint, but the maximum operating temperature of
543 K is too low.
Stainless steel is the probe construction material that was favored by
KVB for field use. It has a high operating temperature limit and is readily
available commercially. However, particulate material collected using stain-
less steel probe may contain nickel, chromium, and other alloy metals.
During sampling, particulate material is deposited on the walls as the
flue gas is drawn through the probe, and the amount of material deposited on
the probe walls could be significant, approaching a sizable proportion of the
particulate material ultimately collected. This material must be washed from
the probe and added to the material washed from the probe nozzle for the complete
determination of emissions. Probe cleaning is difficult and time consuming
under field test conditions. Rinsing and brushing are inadequate for strongly
adhering particulates, leading to biased particulate loading results. For
trace species sampling, the problems are compounded since the probe cleaning
procedure exposes the collected sample and washings to contamination from the
test site surroundings.
The sample collection probe used was a standard EPA Method 5 316 stain-
less steel design with a plastic liner developed by TRW Systems, Inc., '
added. The plastic probe liner not only prevented contamination of the
sample, but it also greatly facilitated sample collection and probe cleaning.
The liner material used was a high-temperature, thermally-stable polyimide,
called Kapton, which was manufactured by DuPont. Kapton is thermally stable
in air to about 750 K (900°F) and has demonstrated ability up to 670 K
(750°F) in flue gas streams. At present, there is no known organic solvent
for the material, and it is infusible as well as flame resistant. Strong
alkali, however, will dissolve Kapton. The results of a spark source mass
spectroscopic analysis of the film material indicate that Kapton does not
represent a significant source of contamination for trace element sampling,
although there is some doubt about its suitability for organic material
sampling.
29
-------
Kapton was available as a. sheet film 0.05 mm thick. A strip
of material the length of the probe and 7.5 cm wide was cut from a
roll of the film. The 7.5 cm width was enough to provide a double
wall thickness when the liner was rolled inside the probe.
After cutting, the film strip was dried overnight in a
desiccator. The strip then was tare-weighed and its weight recorded.
Prior to a test the rolled film strip was inserted into a sampling
probe, and the retainer bushing and sampling nozzle were added.
The retainer bushing conducted stack gases and particulate past
the end of the film preventing the stack gas from getting between
the film and the probe wall. Figure 4-3 is a schematic of the
Kapton liner and bushing inside the sample collection probe.
Upon completion of sampling the probe was disconnected from
the solid particulate collection section and the nozzle was removed.
With the end capped, the probe was taken back into the mobile
laboratory and the liner was removed with a pair of forceps and
placed inside a desiccator in a clean, labeled breaker to await
reweighing.
As with the probe, possible sample contamination or alteration
problems arise with the cyclone and filter. With stainless steel cy-
clones, the possibility of stainless steel alloying elements contami-
nating the samples may be greater than contamination from the probe
because of the increased possibility of surface abrasion. A standard
EPA Method 5 glass cyclone was employed whenever possible to reduce
contamination problems.
The sampling train bubblers and connecting plumbing were another
source of sample contamination, but this contamination was minimized by
using well-washed glass containers.
4.1.3 Filter Selection
Filter materials that have been used for general particulate
sampling contain small amounts of the very trace species being
sought in the trace species testing. Extraction procedures developed
30
-------
Nozzle
=Jr~i
Kapton Liner
Stainless Collection Probe
Flue Gas Flow
<=>
Retainer Bushing
Figure 4-3. Schematic of Kapton liner inside probe tube.
(30)
6001-48
-------
by different laboratories for removing these contaminants have been
only partially successful, and only recently have a few filter materials
suitable for trace element sampling been developed.
Table 4-3 lists some of the filter materials, along with their
(30)
relevant properties, that were considered for trace species sampling.
Of the materials in this table, two were tested in the field: Gelman
Spectro Quality Type E with a siliconized surface and Tissuequartz.
Typical trace species contents of both of these materials are listed
(32)
in Table 4-4 and both were deemed to be acceptable.
However, it was found during sample analysis that when Gelman
Type E filters were used as the final stage of the cascade impactor,
the trace species background was so high relative to the concentration
of the species deposited on it that the filter catch could not be
analyzed. When the Tissuequartz material was tried in the field it
was abandoned because it was too brittle for field use.
4.1.4 Sample Processing
All glassware used to store and/or transport components of
the sampling train was cleaned by washing with detergent, rinsing
with tap water, soaking in warm nitric acid, rinsing with distilled
water, and rinsing with reagent grade acetone.
Subsequent to each run, the filters were removed from holders
and placed in wide-mouth Wheaton glass bottles for shipment to the analy-
sis laboratory. The probe tip, probe, and cyclone were rinsed with
acetone and the rinses combined in a (Wheaton) glass bottle with a Tef-
lon-lined screw cap. Absorbing solutions from the bubblers also were
transferred into separate screw-cap Wheaton glass bottles and the bub-
blers were rinsed with acetone.
4.2 SAMPLE ANALYSIS
The pollutants that were sought consist of two major groups:
polycyclic organic materials and elemental inorganic compounds. The
latter group can be subdivided into those elements which form anions
32
-------
TABLE 4-3. CANDIDATE FILTER MATERIAL FOR TRACE SPECIES SAMPLING
(30)
Filter Material
Carbon fiber
(Kreha Corp. )
Cellulose paper
(Whatman 41)
Graphite
(Poco)
Glass fiber,
spectroquality
(Gelman Instr. Co.)
Microquartz fiber
(Under development)
(Arthur D. Little Co.)
Teflon membrane
Tissuequartz
(Pol If lex Co.)
Sintered Silver
Purity
Good
except
for F
Poor
Excellent
Excellent
Excellent
Excellent
V
Excellent
Excellent
Temperature
Limit K (°F)
>770 (930)
<470 (390)
>1300 (1800)
>270 (210)
1000 (1830)
<1000 (480)
1000 (1470)
<1200 (1650)
Filtering
Efficiency
Low
Medium
High
High
High
High
High
Undetermined
Ease of
AP
Low
Medium
High
Medium
Medium
Extreme ly
High
Medium
Medium
Handling
Strength
Good
Good
Good
Good
Good
Good
Poor
Good but
corrodes
readily
6001-48
-------
TABLE 4-4. APPROXIMATE ELEMENTAL CONTENTS OF SEVERAL FILTER TYPES
(32)
Element
Be
B
F
S
Cl
Ti
V
Cr
Mn
Co
Ni
Cu
Zn
As
Se
Pb
Cd
Sn
Sb
Filter Type
Millipore
AAWP
(cellulose
ester)
mg/cm2
<0.0003
0.002
0.0003
0.006
0.0006
2.
0.0001
0.002
0.01
0. 00002
0.001
0.006
0.01
0.001
0.0001
Flotronics
FM47-.8
(silver)
mg/czn2
<0.0003
0.02
0.1
0.2
3.
0.2
0.003
0.06
0.03
0.007
0.1
0.02
0.01
<0.007
<0.007
<0.07
<0.03
<0.01
<0.01
Pallflex
Tissuquartz
rng/cm^
<0.03
0.75
0.03
1.
0.3
<0.05
<0.05
<0.05
<0.05
<0.05
0.3
<0.5
<0.5
<0.1
<0.1
<0.1
Gelman
(glass
fiber)
mg/cm2
<0.03
70.
2.
0.3
2.
<0.05
0.05
0.3
<0.05
<0.05
<0.05
40.
<0.6
<0.1
<0.1
<0.1
MSA
(glass
fiber)
mg/cm^
<0.03
60.
2.
1.
0.1
<0.05
0.1
<0.1
<0.05
<0.05
<0.05
<0.5
<0.5
<0.1
<0.1
<0.1
34
6001-48
-------
(chlorides and fluorides) and those which form cations upon sample
dissolution. Figure 4-4 depicts the general procedure used to divide the
(29)
sample into these three groups. Organic compounds were removed from
the sample by extraction with benzene. Compounds of special interest,
e.g., the highly carcinogenic benzo-[a]pyrene, is highly soluble in
benzene; whereas.elemental pollutants are insoluble in benzene except when
present as organometallic complexes, which have limited solubility in
benzene. The loss of inorganic pollutants was determined by elemental
analysis of a number of benzene extracts, and it was not found to be
significant. The benzene insoluble material was divided and analyzed
for cationic and anionic pollutants.
Elemental analysis; A number of methods have been applied
to the analysis of cationic elements. These include:
1. Atomic absorption (AA)
2. Inductively coupled plasma emission spectroscopy (ICP)
3. Newtron activation analysis (NAA)
4. X-ray fluorescence spectrometry (XFS)
5. Spark source mass spectrometry (SSMS)
6. Optical emission spectrometry (OES)
7. Electrochemical methods (ECvoltametry and potantiemetry)
The feasibility of analyzing the collected samples for each of the
elemental pollutants of interest by the methods mentioned in the
preceding paragraph is indicated in Table 4-5. A plus indicates the
analysis is feasible, a minus indicates the analysis is not feasible,
c.r.d a circled plus indicates the analysis can be performed in the Midwest
Research Institute laboratory. The criteria used to determine the feasi-
bility of each analysis were the detection limits, sample requirements,
and accuracy and precision of the method.
Detection limits; The detection limit of the method for coal
and fly ash matrices was considered in the light of the expected con-
centration of the specie. If the detection limit was above the expected
concentration, the method was indicated as not feasible. The detection
(29)
limits of several spectroscopic techniques are shown in Table 4-6.
35
-------
Sample
Extract
with
Organic
Solvent
1
Cations
Inorganic
Materials
Anions
V
/Analysis: \ /Analysis:
', Cd4"* Hg4-1-, etcJ I C\", F", S04"
Organic
Materials
Analysis: ^
POM )
Figure 4-4. General procedure for sample treatment and analysis.
6001-48
36
-------
TABLE 4-5
Elemental
Pollutant
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tellurium
Tin
Titanium
Vanadium
Zinc
Chlorine
Fluorine
. FEASIBILITY OF ANALYTICAL METHODS-
AA AA, (micro) ICP NAA-'
©^
©s/
©
©
-
©
©
©
©
©
©
©
©-x
©
©^
©
-
©
-
©
-
-
-
-
-
©
©
©
©
©
©
©
0
©
-
©
-
-
©
©
©
©
-
-
.
©
©
©
©
©
©
©
©
©
©
©
-
©
©
©
-
©
0
©
-
-
©
©
©
-
©
-
©
©
©
©
-
©
©
©
©
©
©
©
©
©
©
©
XFS OES-S/
©
- ©
© ©
- ©
-
- ©
- ©
- ©
- ©
- ©
©
© ©
- ©
© ©
-
- ©
- ©
© ©
© ©
- ©
© -
-
^
0
©
-
-
©
-
©
©
©
©
©
©
-
©
-
-
-
-
-
©
©
©
a/ The analytical procedures arc as follows: AA, atomic absorption
spectrophotometry; AA (micro), AA with carbon rod or other
flamoluss atomization; ICP inductively coupled plasma; I1AA,
neutron activation analysis; Xt'S , x-ray fluorescence spectronetry;
OES, optical emission spcctromctry; EC, electrochemical methods.
b/ NAA methods include chemical pretreabnent.
c/ OES methods include chemical pretreataient, photometric detection,
DC arc (inert atmosphere).
d/ EC methods are anodic stripping voltametry and potentiometry
(specific ion electrode for fluoride).
e/ Includes SbH , AsH , and SeH, generation and N -H -air flame.
f/ Cold vapor.
6001-48
37
-------
TABLE 4-6. ANALYTICAL SENSITIVITY OF SPECTROSCOPY TECHNIQUES
(29)
Detection Limits *
Trace Element
(cation)
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tellurium
Tin
Titanium
Vanadium
Zinc
CP
Method
(ng/ml)
200
40
0.1
0.5
2
0.07
1
3
1
5
8
0.7
200,
6.
30.
30
300
3
6
2
Conventional
"Flame
Methods
(ng/ml)
600.
1300.
300.
20.
20.
500.
98.
90.
40.
600.
160.
50.
2000.
70.
600.
300.
2000.
2200.
1200.
12.
Flameless Hydride
Methods Methods "
(ng/ml) (ng/ml)
20 4
20 4
0.18
0.02
10.
1.0
1.2
1.4
20.
1.0
0.1
20.
2.0
20.
20.
12.
-
20.
0.02
* The concentration of a solution which produces a signal twice that
of the background.
38
6001-48
-------
Sample requirements; If a method required a prohibitively
large sample size (>1.0 g of particulate) or large dilution factors
for multispecies analysis, it was indicated as not feasible.
Accuracy and precision: The accuracy and precision of the method
must be sufficient to obtain an acceptable material balance for the
specie in question. A relative standard deviation of +_ 20% was selected
arbitrarily. Each of the methods indicated as acceptable meet this
requirement in the most extreme situation (i.e., near a signal level
of twice the noise level). At normal working levels, the feasible
methods generally are within a relative standard deviation of 5%.
Spark Source Mass Spectrometry (SSMS) is not listed in Table 4-5 because
it did not meet these criteria. The NAA ratings are for methods which
include chemical pretreatment. The Atomic Absorbtion (AA) (micro), and
Inductively Coupled Plasma (ICP) methods are for one general sample
pretreatment which consists of dissolving the sample. The AA and AA (micro)
methods consist of consecutive single species determinations. The ICP
method consists of up to 20 simultaneous species determinations which
result in a multispecies scan similar to SSMS, but ICP results are
more quantitative than SSMS.
Cost considerations are not reflected in Table 4-5. Because
several chemical treatment procedures are required for NAA and OES
analysis for the twenty elements of interest,'these methods are more
costly than AA or ICP analysis. The cost for ICP analysis is the
same as AA for sample preparation, but due to the increased speed of
analysis the cost per element quantified is somewhat less. The
sensitivity of ICP is better than Flame AA and comparable to Micro
AA techniques. Accuracy and precision for ICP are intermediate to
Flame AA and Micro AA procedures. The costs for electrochemical
analysis are similar to those for AA where satisfactory methods exist.
39
-------
Gas chromatography, liquid chromatography and UV-visible spec-
troscopy have been applied to the analysis of POM. ' ' Identifi-
cation should be verified by mass spectrometry if sufficient concen-
trations are present. Verification is accomplished most conveniently
by a mass spectrometer interfaced directly with a gas chromatograph.
The methods of choice for both organic and elemental pollutant
species are described in detail in the following section.
4.2.1 Analytical Methods
The methods of analysis that were used for each of the species
are indicated in Table 4-7. The configuration of the collection train
is shown in Figures 4-1 through 4-3 and one of the sets of reagents
that was used is shown in schematic form in Figure 4-5. Several dif-
ferent sets of reagents were used in the bubblers during the program
because the reagents still were undergoing development during the pro-
gram. The set used for most of the program will be used here for il-
lustration, and the reader is referred to the individual tests for a
discussion of the particular collection reagents actually employed.
Figure 4-6 is a flow chart of the sample processing and analysis of
the bubbler reagents, the cyclone and front filter, and the Tenax ad-
sorbent .
A schematic of the components of the cascade impactor and the
analysis methods used to investigate possible selective enrichment
of certain size particulates is shown in Figure 4-7.
Elemental pollutants (cations); Inductively Coupled Plasma
Emission Spectroscqpy (ICP) and Atomic Absorption Spectrophotometry
(AAS) were used for the elements of interest.
Some of the advantages of ICP are: the low limits of detection
for most of the trace elements of interest, the large number of ele-
ments which can be analyzed, the simultaneous multi-element deter-
minations, the rapid analysis of large numbers of samples, nearly the
same precision and accuracy of atomic absorption, the use of ultra-
trace levels on yl or yg samples and the greatly decreased interference
from matrix constituents.
40
-------
TABLE 4-7. RECOMMENDED CHEMICAL ANALYSIS METHODS
(29)
Pollutant
Trace Elements (Cations)
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Cobalt
Copper
Lead
Manganese
Mercury
Nickel
Selenium
Tellurium
Tin
Titanium
Vandium
Zinc
Trace Elements (Anions)
Chlorine
Fluorine
Sulfate
Organics
POM
Method of
Analysis*
6
6
1
1
1
1
1
1
1
1
7
1
6
2
2
1
1
1
3
4
8
Minor Elements (Cations)
Calcium
Iron
1
1
*The methods of analysis are as follows:
(1) Inductively Coupled Plasma Emission (ICP);
(2) AAS, micro flameless methods;
(3) AgNO titration, electrochemical (EC) detection;
(4) EC, fluorine selective electrode;
(5) Gas chromatography (GC), electron capture detection;
(6) AAS, hydride generation method;
(7) AAS, cold vapor method.
(8) Barium perchlorate titration.
41
6001-48
-------
TRACE SPECIES COLLECTION TRAIN:
&
Silic
Gel
Probe &
Cyclone
<: (^-=
^ W^
a Empty
©
Front Nc
Filter
rrw
H202
AgN03
HN03
Bubbler Number
(^<
H2°2
AgN03
HN03
J
(0^
^J"
HO
AgNO
HN03
-\ZJ
Empty
r?i-c
Tenax
Adsorber
Collector
Component
Designation
Analysis
Method
1,2,3
4,5,6,7
Elemental
POM
Se, As, Sb, Hg
Particulates
Figure 4-5. Required analyses for trace species collection train.
6001-48
42
-------
Train
Component
and Contents
Bubbler Nos.
4, 5 & 6
AoiH nirhromafrp
Bubbler Nos.
1 & 2
Na2C03
Cyclone and
Filter
Tenax
Adsorber
-
-
Extraction Procedure Analysis Method
Elemental Analysis:
HF
Digestion Reduce With__ Hg Analysis:
\l
I
' SnCL, Cold Vapor AAS
Reduce With As, Sb, Se Analysis:
Na ,'B'H. AAS of Hydrides
Cl~ Analysis:
' *" chlondometer
MaOH
Digestion p- Analysis.
H?O *^ SO^ Analysis
Extraction Barium-Thorin
Extraction^
s column ' POM Analysis:
Chromatography i, ^,, ,-,.,,-,, ~~
Figure 4-6. Sanple processing and analysis flow chart.
CASCADE IMPACTOR
N > 3*{C
Nozzle Cyclone 5 Individual Stages Back-up
Filter
^s-s^xv^ Impactor
^"\^on5>onen t
^^^Qesignation
Analysis^v^^
Method ^^^^^
A. Elemental
B. POM
C. Particulate
Stage
Nos.
N,C 1-5 F
0
0 0
Figure 4-7. Required analyses for cascade impactor catch.
6001-48
43
-------
Some of the advantages of AAS are: the low limits of detection
for most of the trace elements of interest, the large number of elements
which can be analyzed, the low cost of highly reliable instrumentation,
and the accuracy and precision which can be obtained without highly trained
technical personnel.
The combination of ICP and AAS gave added flexibility to the analy-
sis procedures. Those elements which were of interest could be scanned by
ICP, thus eliminating the cost of spark source mass spectroscopy. Most of
the twenty elements of interest could be quantified by ICP techniques. For
those elements which could not be determined by ICP, AAS was used. This
combination of techniques allowed rapid analysis to meet the time require-
ments of the project. It also assured that a quality analysis was per-
formed on all twenty elements of interest.
The recent development of the Inductively Coupled Plasma Emission
method enables simultaneous analysis to be conducted on as little as 100
yl of solution, hence reducing the total volume of solution required to
analyze twenty elements from >_ 25 ml to 1 to 5 ml. In addition, the solu-
tion detection limits are in the nanograms per milliliter range rather
than micrograms per milliliter range required by conventional flame methods.
In view of the analysis time required per sample, the number of samples
involved with this project and the detection limits and accuracy required
of the analytical methods, Inductively Coupled Plasma Emissions Spectro-
photometry was chosen.
The sample preparation procedures for ICP and AAS depend upon
sample matrices. For this investigation, the types of matrices were
fuel, fly ash and bottom ash, sodium carbonate or ammonium hydroxide
solution, and acid-dichromate or hydrogen peroxide solutions. A
suitable preparation for these matrices is the acid-pressure decom-
position technique developed by Bernas. This technique
44
-------
has been applied to 12 of the 19 potentially toxic pollutants of in-
terest including refractory forming elements, e.g., Ba, Ti, and V. This
procedure, coupled with AAS, also has been applied to the analysis
of granite, coal, coal ash, glass, and fish tissue for 18 elements
and resulted in a relative standard deviation of approximately 5%
for trace elements. The procedure has the following advantages:
elimination of interelement and ionization interferences, elimination
of volatilization and retention losses, relatively low cost per
analysis because of the reduction in time and supervision required
and the elimination of expensive platinum ware containers.
Since fewer interference problems occur with TCP than with
AA procedures, the acid-pressure decomposition technique is directly
applicable. In this method, the samples are decomposed in a Teflon
cup and then encapsuled in a stainless steel bomb with a decomposition
medium of hydrofluoric and nitric acid. The samples are digested
for 0.5 to 3 hr at 370 K to 440 K. The higher temperature is used
for samples of higher organic content.
Two methods of sample preparation were used for sampling the
train solutions that had collected the volatile elements in flue gas.
The first, used for the mercury determination, was the cold vapor
technique. It included digestion with fresh acid-dichromate followed
by reduction with hydroxyl amine (for excess permanganate) and stannus
chloride. The second method was to treat a portion of the solution
with sodium borohydride to form the hydrides of arsenic, selenium,
and antimony prior to their determination by AAS using an argon/
hydrogen/air flame with the sample entrained within it.
Elemental pollutants (anions): Fluorine and chlorine cannot
be determined by AAS. Samples were analyzed for fluorides following
digestion under pressure and absorbed in a sodium hydroxide medium.
45
-------
After buffering, fluorides were determined with a fluoride-selective
electrode. The practical limit of detection was 10 \ig of fluoride per
gram of sample. Chlorine was determined by igniting the samples in a
bomb and titrating the aqueous washings with silver nitrate. The prac-
tical limit of detection was 0.01 to 0.03%.
Sulfates; Solid samples were extracted with hot water for 24
hours. Sulfates in liquid and extracts from solids were determined by
microtitration with barium perchlorate using a Thorin indicator. Re-
sults are reported as water soluble sulfate as SO .
Organic pollutants: The organic compounds in the particulate
matter were separated from the total particulate by extraction with
benzene. The separated organic material and the organic material
collected in the Na.CO, or NH OH impingers and the Tenax plug were
analyzed for POMj
POM: The POM in the benzene soluble fraction was separated
from PCB, aliphatic and heterocyclic compounds by column chromato-
graphy with activated silica gel as the adsorbent.
Following the isolation of the total POM, gas chromatography
with electron capture detection was used to quantify the individual
POM. Solution detection limits were in the 0.2-0.5 yg/ml range. A
Drexsil-300 packed column has been used to separate 70 POM. This
packing was used in the current program to separte the 8 POM's listed
in Table 4-8. For the current program, quantification of 5 of the 8
desired POM's* was made by comparing sample peak areas with areas ob-
tained from synthetic mixtures. Of the remaining 3 POM's, benzo[c]
phenanthrene and dibenzo[a,h]pyrene could be identified by known rela-
tive retention times. However, there was no information of GC
analysis of dibenzo[a,g]carbazole on Dexsil-300, making identification
impossible without a standard.
*Standards could not be obtained for benzo[c]phenanthrene,
dibenzo[a,h]pyrene and dibenzo[c,g]carbazole.(29)
46
-------
fABLE 4-8. CARCINOGENIC POLYCYCLIC ORGANIC MATERIALS
1.
2.
3.
4.
5.
6.
7.
8.
Compound Structure
7,12 Di.methylbenz[a_]- ^-vJsiO
anthracene QJkJkJl
CHs
DibenzCa^hJanthracene i^VvO
AjAX/
U
Benzo[cjphenanthrene f'V^
OX)
3 Methylcholanthrene fS
i^VVV*
a^-1^
Benzo[a]pyrene i^Y^
" rrry
sAxk>
Dibenzo[a_,]ijpyrene J
Dibenzo[c,g]carbazole H
I
Carcinogentclty
+ 4
i- 3
+ 3
+ 4
+ 3
+ 3
+ 3
+ 3
6001-48
47
-------
4.2.2 Quality Assurance of Analytical Methods
All samples were analyzed in duplicate to determine the pre-
cision of the analytical results. Selected samples of each type, for
each unit tested, were fortified with three different concentrations
to determine the accuracy of the analysis procedure. These experiments
also determined the percentage of recovery for the digestion or extrac-
tion procedures. Where possible, standard reference materials from
NBS were analyzed to determine accuracy of the procedures. All con-
centration determinations were taken from five-point calibration curves,
which were repeated several times daily during the analysis of the
samples. Samples were diluted or concentrated to bring them into the
linear calibration range.
4.3 PAKTICULATE SIZE
The particulate size was measured using a Brink Model "B" Cas-
cade Impactor. A low volume type impactor was selected because the
grain loading of the coal fueled boilers was relatively high. The
nominal sample flow rate of 2.8 1/min (liters per minute) was low
enough that the impactor did not readily overload. One stage of the
cascade impactor and the precutter cyclone used with coal fuel are
shown in Figure 4-8. The cyclone design was patterned after the
/ 30 )
Southern Research Institute Model SRI-4.
A Cahn Model G-2 Electro-balance with a sensitivity of 0.05
micrograms was used to weigh the collected sample. This sensitivity
was needed for the lower stage of the impactor where the collected
weight occasionally was less than 0.1 mg. To improve the accuracy of
the weighing, an aluminum foil substratum was placed in each steel col-
lection cup. The particles were collected and weighed on these light-
weight cups, and the original steel collection cups were used only as
a backing for the foil cups.
48
-------
.so a
Inlet Nozzle
#7 DRILL
-101 TANGENT
TO GORE
PRECUTTER CYCLONE
Assembled
Stage
-3 SLOTS
Single
Collection^
Cup
STAGE
Figure 4-8. Detail of one stage and of precutter cyclone for
cascade irapactor. 6001-48
49
-------
A common problem with impactors is that the particles do
not adhere to the stage surface, but strike it, rebound, and are
reentrained in the flow through the slots down to the next stage.
Reentrainment did not prove to be a problem with the cascade im-
pactor measurements however. The flue gas flow rate was reduced
from the nominal 2.8 1/min to 2.0 1/min or less, and visual ex-
amination of the collection stages by hand lens found no evidence
of scouring or reentrainment. One set of stages was further ex-
amined under an electron microscope and there was no sign of a
significant number of particulates that were larger than the aero-
dynamic diameter cut point of the preceding stage. There was, how-
ever, a considerable amount of sponge-like material that appeared
to be an agglomeration of small particles.
Back-up filters were used as the final stage of the impactor
to collect the material that passed the last impaction stage.
Binderless, glass-fiber filter material, such as high-purity Gelman
Type A Glass Fiber-Filter Webb was employed for this purpose. The
25 mm diameter circular filters were placed under the spring in the
last-stage of the impactor. The filter was protected from being cut
by the spring by a Teflon 0-ring, and a second filter disc and a
wire mesh were placed beneath the filter to act as a support.
The flow rate and nozzle size were closely coupled, and
requirements for isokinetic or near-isokinetic nozzle flow sometimes
forced a compromise on nozzle selection. The order of priorities
used by KVB in the field to determine the proper nozzle size was
(1) nozzle diameter not smaller than 2 mm, (2) last stage jet
velocity not excessive, (3) isokinetic sample collection rate, and
(4) nozzle diameter (if greater than 2.0 mm). The impactor nozzle
diameter was selected to provide as close to isokinetic collection
as possible, but a nozzle bore less than 2 mm was avoided to
forestall nozzle plugging by fly ash.
50
-------
The impactor assembly was placed inside the stack and was
heated to flue gas temperature by the flue gas itself before the
sample collection was begun. The inlet nozzle was pointed downstream
of the flow field during this heating phase to prevent the premature
accumulation of particulates in the impactor.
The flow through the impactor was measured before each use
to determine the actual cut points of the individual stages. This
flow then was maintained by monitoring it with the pressure gauges
on the EPA Method 5 control box. The pump on the control box was
used to maintain the flow. During data analysis the true cut
points were calculated for the actual gas flow rate through the
impactor and cyclones using an approximate solution to the
/ 2Q]
following two equations.
1.43 x 10
4
3
P X(I)
c s
PpQI Po C 472'°
c-n. -2k
1.23 + 0.41 EXP ["(-0.44 D5Q)/L x lo"4)l
where
D5Q x 1C"4
D is the stage cut point
50 nominal is 0.25, 0.5, 1.0, 1.5, and 2.5 urn,
H = gas viscosity (poise) ,
D = stage jet diameter (cm) ,
c
P = local pressure at stage jet (atm) ,
3 3
p = particle density (gm/cm. ) ,
Q = impactor flow rate (cfm) ,
nominal is 2.8 1/min, 46.7
cm /sec or 0.10 ft /min,
P = ambient pressure at impactor inlet (atm)
o
C = Cunningham Correction Factor,
L = Gas mean free path (cm) , and
X(I) = Number of holes per stage.
Exact solution can be achieved by iterative methods.
51
-------
The particle size reported can be either the aerodynamic diameter
based on the behavior of unit density particles, or the approximate
physical diameter based on the estimate of the true particle density.
The aerodynamic diameter is reported as the particle size of unit
density that is collected with 50% efficiency by the impactor stage.
The symbol conventionally used for aerodynamic diameter is D . In
either case, the particles are assumed to be spherical.
Measurements were made at a sufficient number of points
across the flue or smoke stack, as specified by EPA Method 5, to
make certain that a representative sample of particulates was
obtained.
When coal fuel was fired the proportion of material with
a size larger than ten micrometers was appreciable. The precutter
cyclone shown in Figure 4-8 was used to collect this material and
to prevent overloading of the upper impactor stages. The equations
and nomenclature that were used to calculate the D_ , or cut point,
(38)
for the cyclone was:
where
D = cyclone D (ym),
U = gas viscosity (poise),
p = density of particle (g/cm ),
s 3
p = gas density (g/cm ),
N = number of turns made by gas stream in the cyclone
° body and cone,
V = inlet gas velocity (cm/sec), and
c
B = width of cyclone inlet (cm).
c
52
-------
The square root relationship between D and V was used to calculate
the aerodynamic diameter of the cyclone over the range of flowrates.
It also was used to calculate the D 's of the individual stages of
the impactor. Since the cyclone had been calibrated at a known
flowrate, one can rewrite the equation as
where
D50(1)
D_ (1) - the cyclone cut point at flowrate 1, nominal is 7.3 ym
C = the cyclone calibration constant, and
/ = flowrate 1, n<
0.10 ft3/min.
V, = flowrate 1, nominal is 2.8 1/min, 46.7 cm /sec or
Then the relationship between D Q(l) and D5Q(2)' the cut points at
the two different flowrates, was
' D50(2) ' D50(1)
For particles of different mass densities, the nominal
particle aerodynamic diameter was corrected using the relationship:
where
P = the density of the calibration aerosol, (1.35 g/cc),
p = the density of the test particulate (g/cc),
D(p ) = aerodynamic diameter of the calibration aerosol (urn)
D(P«) = aerodynamic diameter of the test particulate (Um)
The density of fly ash that was used in the calculation was taken
from Reference 39 and was 1.78 g/cm for particulates less than
44 urn in diameter.
These relationships have been verified experimentally, ^38;
and they were used to calculate the true aerodynamic diameters of
the cascade impactor stages and cyclone.
53
-------
All the material from the nozzle to the outlet of the cy-
clone was included with the cyclone catch. All material between
the cyclone outlet and the second stage nozzle was included with
material collected on the first collection substrate. All adja-
cent walls were brushed off, as well as around the underside of
the nozzle. All material between the second stage nozzle and third
stage nozzle was included with that on the second collection sub-
strate. This process was continued down to the last collection sub-
strate .
4.4 MATERIAL BALANCE
A mass balance calculation was done for the three coal-
fired boilers where the emissions were analyzed for trace metal
and organic content. The method of calculation is discussed in
this section and the results of the calculations are discussed
in the following sections. The input term to the balance was
the quantity of a particular trace element that was found upon
analysis to be in the coal burned during the test. The output
terms were the amount of the element found in the bottom ash, the
dust collector ash, and the fly ash in the flue gases.
Imbalance of the volatile elements will result if the flue
gas is sampled infrequently, therefore, an attempt was made to
collect enough data from three successive test runs that each
material balance could be made three separate times. This
technique would provide an expected range of material imbalance
of 25% with 90% confidence, however, the goal of at least three test
runs was not always achieved.
A practical field test problem with industrial boilers was
to collect for measurement all of the solid material produced by
the combustion process. During boiler operation, a significant
quantity of material was deposited on the furnace bottom as refuse
and slag, on the boiler tubes as slag and soot, and in the dust
collector as ash. For the furnace bottom and dust collector, the
54
-------
effect of differences in deposit rates during the test were mini-
mized by emptying the collector before and after the test. If
this plan was not practical, the collector was emptied one-half
hour before the start of the test and again one-half hour before
completion of the test. It was assumed that the quantity of ash
would be the same as if the emptying had occurred at the beginning
and end of the test and that the quantity of ash collected was
equal to that actually produced during the test.
4.4.1 Fuel
Three fuel-related items must be considered in determining
the input to the boiler. These items are fuel-feed rate, fuel
sampling, and fuel analyses.
Fuel-Feed Rate: Few coal-fired industrial boilers were en-
countered that had facilities for conveniently determining fuel-feed
rate by weight or any other technique. However, one of the requirements
for trace species testing was that the coal feed rate be measured,
and it was possible to find three boilers that did have a provision
for weighing the amount of coal fed to the burners. An alternative
method would have been to estimate the feed rate by a heat balance.
Fuel Sampling: ASTM D492-58, "Standard Method of Sampling
Coals Classified According to Ash Content," gives guidelines for
sampling, including recommended number of sampling increments and
increment weight. However, coal sampling, like measuring coal
firing rate, is very site dependent, and the exact technique used
had to be tailored to each boiler after inspecting the facilities
available at the boiler site. For the 11 to 12% ash, Group 1 coal
that was used for the trace species tests, the standard method
requires 35 incremental samples of two pounds each for a total
weight of 70 pounds. To handle such a quantity of coal in the
55
-------
field was impractical and a compromise was made. Twenty-five small
incremental samples of a few ounces each were collected for a total
sample weight of about two pounds.
Fuel Analysis: The coal was analyzed for its chemical and
ash content and its heating value by Industrial Testing Laboratory,
Kansas City, Missouri, as part of the trace species analysis of
the collected sample. Proximate, ultimate and heat of combustion
analyses on "as received" and "dry basis" were made and reported.
The trace species content was determined by the Midwest Research
Institute, Kansas City, Missouri.
4.4.2 Bottom Ash and Slag
Bottom ash and slag are the refuse that collects on the
bottom of a furnace or is dumped into the ash pits by the action
of the grate. As in the case of fuel, three items are important
in defining the bottom ash from an industrial boiler: measuring
the quantity of ash produced, collecting a representative sample,
and analysis.
Accumulation Rate: The ash collector was emptied just
prior to the start of a run and just after the end of a run and
the contents weighed where this was possible. At Location 31,
it was not possible in actual practice, because the refuse
collection system was not readily accessible during operation.
It consisted of a large number of closed pneumatic and hydraulic
devices for collection and disposal that proved to be impracticable
to empty individually in a short time.
Ash Sampling: A large number of small samples of refuse
were collected and combined into a single composite sample for
analysis, as was done for the coal sample.
Ash Analysis: A proximate analysis of the bottom ash was
performed and reported by Industrial Testing Laboratory, and a trace
element analysis was performed by the Midwest Research Institute.
56
-------
4.4.3 Dust Collector Ash
Each of the test boilers had a dust collector or bag house
for collecting fly ash from the flue gas stream. The three
important factors needed to characterize the dust collection for
achieving an overall mass balance were the rate at which collector
ash was produced, the method of collecting a sample of the ash,
and the analysis of the ash.
Accumulation Rate: Standard methods did not exist for
determining the rate of accumulation of ash by the dust collector
nor for obtaining samples of the ash, since these data generally
. were not of concern to the operator. Hence, we encountered no
facilities specifically for measuring ash accumulation at the test
sites. At Location 31, Test 169, particulate concentrations in the
flue gas were measured, both upstream and downstream of the cyclone
dust collector, and these data were used to determine the rate of
fly ash accumulation in the dust collector. At Location 35, Test
166, measurements were made only at one point in the exhaust gas
stream, as is explained in Section 6.
Collector Ash Analysis: A large number of small samples were
collected and combined for analysis by the testing laboratory. The
same analytical methods used for bottom ash analysis were used for
analyzing collector ash for unburned carbon content, trace elements
and organics.
4.4.4 Fly Ash
Both the solid particulates and the condensible particulates
were included in the determination of the total fly ash concentration
in the flue gas. For Tests 166 and 169 a standard EPA Method 5
analyzer was used. For the special reagent comparison tests, Tests
213 and 214, the trace species collection train pictured in Figure
4-1 was employed. The impinger case had been enlarged sufficiently
to hold eight rather than four bubblers.
57
-------
4.4.6 Soot
A salient difference between the ash measurements made as part
of the trace species tests and conventional emission tests was that the
soot generated during the test was collected and included in the analysis.
The soot blowers were operated just prior to the start of a test to clear
out the soot that had accumulated on the boiler tubes and walls. Just
prior to the end of a test the soot blowers were operated again to dislodge
the soot that had collected during the test.
58
-------
SECTION 5.0
FIELD TEST, ORIGINAL COLLECTION TRAIN
The original design of the trace species train had been developed
by field research and test teams that were under contract to the U.S.
Environmental Protection Agency at the time the trace species testing was
undertaken by KVB, Inc. This original design was used in the first KVB
test at Location 29 and by the Midwest Research Institute at two of their
(42)
test sites. Reagents initially used by KVB consisted of distilled water
in the first two impingers, saturated sodium carbonate in impinger 4, and
sulfuric acid permanganate in impingers 5 and 6. MRI used these and other
reagents. A fiber filter and Tenax adsorbent were used between impingers
3 and 4. Because the vaporous mercury was not being properly scrubbed out
of the stack gas, the results were not considered to be satisfactory.
The basic mechanical design of the collection train did not change
significantly during the first two KVB tests. Bubblers were substituted
for impingers so the gas could be pumped through the train at isokinetic
speeds. The Tenax adsorber holder was changed from a cylinder to a high-
cross section spherical shape and the bubbler chemicals were changed
several times.
Four test runs were conducted at Location 30, two for inorganic col-
lection and two for organics. Reagents for the inorganics tests were sodium
carbonate in impingers 1 and 2, acid-dichromate in impingers 4 and 5, and
acid-permanganate in impinger 6. Nitric acid was used in one test and
sulfuric acid in the second. The results showed that selenium was collected
in the first impinger, antimony in the first two, arsenic in the two
dichromate impingers (4 and 5) and mercury in all five, but highest in the
permanganate (6). Other tests by MRI indicated mercury collection efficiency
(42)
was still low. For organics tests, toluene was used in the last two
impingers and a second Tenax adsorber replaced impinger 4.
59
-------
From the results of the tests at Locations 29 and 30 and by MRI,
the sampling reagents shown in Figure 5-1 were recommended by MRI for fu-
ture tests by KVB. The function of the two sodium carbonate-water bubblers
was to trap PCB's, selenium and antimony, and to maintain a PH high enough
to stabilize benzo[a]pyrene. This solution also served to trap SO , thus
A
protecting oxidizing bubblers later in the train as well as producing an
oxidizing solution for selenium and antimony vapor. The empty bubbler was
necessary to catch spray from the preceding bubbler.
A single holder containing a Tenax adsorber was placed between the
third and fourth bubblers, rather than the combination of a fiber filter
and Tenax that formerly had been used in this position. With this arrange-
ment, including a larger amount of Tenax and 200 ml of reagent in the bub-
blers, a serious pressure drop could be avoided. After the Tenax, the re-
mainder of the recommended train was as follows: two 10% nitric acid 5%
sodium dichromate bubblers, followed by one 10% nitric acid potassium
permanganate bubbler, an empty bubbler for spray from the permanganate and
finally a dry bubbler with silica gel. This mass train configuration was
expected to give a good compromise between collection of POM and volatile
metals.
Problems were still considered to exist with this recommended rea-
gent configuration, primarily the inability to assure high percentage col-
(42)
lection of mercury vapor at isokinetic flow rates. As a result, this
reagent set was abandoned without further test and an alternate set of
reagents was adopted as recommended by EPA. The sodium carbonate was re-
placed with ammonium hydroxide in the first two bubblers. A mixture of
hydrogen peroxide, silver nitrate, and nitric acid was used in bubblers
4, 5 and 6. The subsequent test at Location 35, discussed in the next
section, revealed that ammonium hydroxide caused the Tenax to plug. The
ammonium hydroxide was then replaced with sodium carbonate. This rea-
gent set was used for all subsequent tests with this particular sampling
train.
60
-------
BUBBLER
NUMBER
Probe
Cyclone
Heated Pilfer
10% Na2CO3
10% Na2CO3
Glass Wool
Tsnax Adsorbent
0
10% HNO3 - 5%
10% HNO3 - 5% Na2Cr2O7
10% HNO3 - .2% KMnO4
Empty
Silica Gel
1
2
4
5
6
7
Console
Figure 5-1. Reagent set recommended following initial tests.
6001-48
61
-------
SECTION 6.0
FIELD TEST NO. 166, MODIFIED COLLECTION TRAIN
Material balances of the elements for the three runs of Test 166
at Location 35 on a chain grate stoker firing crushed coal are shown in
Figure 6-1. The mass of an individual element collected is the mass found
upon analysis in the combined furnace bottom and dust collector ashes and
in the flue gas fly ash. The balance also includes the amount of arsenic,
mercury and selenium that was present in the stack gas as a vapor.
In general, the moderately-volatile elements, such as barium and
manganese, tended to concentrate in the hopper ashes, and the proportion
recovered was high, running from 50% to 260% of the amount of the ele-
ment that was in the coal.
The recovery of mercury was near 100% in all three runs. This
high recovery was gratifying because it had been difficult to capture
mercury heretofore. However, almost all of the mercury was recovered
in the furnace bottom and the dust collector ashes from the ash hopper,
and only about four percent was recovered from the stack gas vapor by
the bubbler reagents. Other investigators have found up to 90% of
the mercury to be in the stack gas as a vapor. This discrepancy leads
one to suspect that the reported concentrations of mercury in the hopper
ash are in error on the high side.
The recovery of the other two highly volatile elements, arsenic
and selenium was only about 40%, but they too were found predominately
on the hopper ashes and not in the stack gas as a vapor. However,
the low recovery from the stack gases may have been due to the bubbler
chemicals not effectively scrubbing out the arsenic and selenium.
62
-------
Mass of element collected/Mass of element in coal, %
(.
Arsenic Ar
Barium Ba
Beryllium Be
Cadmium Cc1.
Calcium Ca
Chromium Cr
Cobalt Co
Copper Cu
Iron Fe
Lead pb
Manganese Mn
Mercury Hg
Nickel Ni
Selenium se
Tin sn
Titanium Ti
> 50 100 150 200 25
=
'
=
1XX/
*//.
'$/
v//
vr,
rr^
v//.
\//
%
y/,
'/A
//A
w
"///
//}
V/\
^
x^!
/ / A
-?-/-*
f///\
//A
l I
Acceptable
_-.
*^ mf*
Test No.
Balance
Calculated hopper ash amount
Weighed hopper ash amount
166-9
166-10
166-11
Figure 6-1. Material balance test run numbers 166-9, 166-10, and 166-11.
63 6001-48
-------
The elements/ chromium, nickel, and titanium, were much over-
balanced. The surplus of these elements possibly was due to contamination
from the stainless steel sample intake nozzle or the steel from which the
boiler itself was made.
The results of the material balance depended upon whether a
measured or a calculated amount of hopper ash was used for the compu-
tation. The calculated amount of bottom and collector ash was a good
deal higher than the amount of ash contained in the coal, since the
hopper ash contained from 40% to 70% of carbon and other volatiles.
The amount of ash removed from the hopper and weighed was 11% to 54%
higher in turn, than the calculated ash. Since it was difficult to
get an uncontaminated sample of the hopper ash and there was considerable
uncertainty about the weights, two calculations were made, one for the
calculated hopper ash and another for the measured hopper ash.
The material balance based on the calculated hopper ash is
shown in Figure 6-1 by the solid bar and the one based on the measured
hopper ash is shown by the dashed extension to the solid bar. The
actual balance probably lies between the end of the solid bar and the
end of the dashed bar.
Making a complete material balance for chloride and fluoride
was not possible because there was an insufficient quantity
of the flue gas fly ash for analysis. A partial balance for these
two species is shown in Table 6-1 for each of the runs of Test 169.
An acceptable balance was achieved for chloride for two out
of the three runs, while the fluoride balance was successful only
for run number 10. Apparently a sizable proportion of the fluoride
vapor was not condensed by the reagents in the bubblers. About three-
quarters of the chloride that was captured was in the furnace bottom
and dust collector ashes in the hopper. It tended to partition out
with the heavy particulate. The fluoride, on the other hand, was
distributed relatively uniformly between the hopper ash and the
stack gas vapor.
64
-------
TABLE 6-1. CHLORIDE, FLUORIDE AND SULFATE MATERIAL BALANCE
Test No. 166, Location No. 35
Run
No.
9 Chloride
10
11
9 Fluoride
10
11
9 Sulfate^
10
11
In Coal
g %
9,065 100
10,756 100
6,497 100
1,877 100
1,137 100
1,081 100
80,217 100
83,952 100
54,047 100
In Measured
Hopper Ash
g %
5,302 58
754 7
5,107 79
347 18
520 46
339 31
16,950 21
36,922 44
23,679 44
In Stack Gases
Particulate
q %
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
38.4 0.05
35.2 0.04
5.27 0.01
Vapor
g %
1,910 21
2,601 24
1,656 25
711 38
657 58
438 41
790,990 986
1,120,388 1,335
961,260 1,778
Total and
Balance
g %
7,212 80*
3,355 31*
6,763 104*
1,058 56*
1,177 104*
777 72*
807,978 1,007
1,157,345 1,379
984,944 1,822
a\
en
* Partial balance only, insufficient particulate sample for total balance.
Sulfate in solids is water soluble sulfate as SO^.
Sulfate in stack gas vapor includes all sulfur compounds collected/ expressed as
6001-48
-------
The sulfate collected in the hopper ash and stack gas particu-
lates was 21 to 44% of the sulfate present in the coal. Sulfate re-
ported for stack gas vapors includes all sulfur compounds collected
in the first two impingers in the sample train. Any SO or SO present
in the gas will be indicated as sulfate.
Some of the polycyclic organic materials (POM's) were found in
the coal and in the hopper ashes. The emissions of these organics are
tabulated in Table 6-2. POM numbers 1 and 2 were detected in the hopper
ash but not in the stack gas. The presence of POM in the hopper ash was
unexpected, since the melting and/or subliming temperature of these organics
is only about 475 K and well below the temperature of the combustion gases.
The sample of fly ash collected from the stack gas was of insufficient
size for analysis. In future tests of this type, provision should be made
to collect a larger sample of the stack gas particulate.
6.1 BOILER AND FUEL CHARACTERISTICS
Boiler #6 at Location 35 was an Erie City Iron Works brand
watertube boiler installed in 1960. The superheated steam boiler
had a rated capacity of 227 GJ/hr (215,000 Ib/hr) of equivalent
saturated steam flow and final steam conditions of 755 K (915°F),
and 6.13 MPa (875 psig). The boiler was coal fired with a chain-
grate, traveling-type stoker manufactured by Combustion Engineering.
Forty-six overfire air nozzles designed to enhance mixing of the fuel
and combustion air were mounted above the chain grate, and air was
supplied to these nozzles by a separate overfire air fan. Primary -
combustion air was supplied to the windbox under the grate by a
forced draft fan. An induced draft fan maintained a balanced furnace
draft. A tubular air preheater heated the combustion air to approxi-
mately 500 K (220°F) .
The boiler load was limited to approximately 60% of rated
capacity during the test program due to undersized fans and blocked
gas passages. Consequently, the test load was established at 110 GJ/hr
66
-------
TABLE 6-2. ORGANICS EMISSIONS
Test No. 166, Location No. 35
Run POM
No . No . Name
9 1. 7,12 Dime thy Ibenz [a] anthracene
10
11
9 2. Benzo[a]pyrene
10
11
9 3. 3 Melhylcholanthrene
10
11
9 4. Dibenz [a, h] anthracene
10
11
In Coal
g %
none
detected
none
detected
none
detected
109 100
245 100
18.8 100
none
detected
none
detected
none
detected
610 100
822 100
673 100
In Hopper
Ash
g %
none
detected
2.26
1.39
37 34
92.6 38
651 3462
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
In Stack Gases
Particulate
g %
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
insufficient
sample
Vapor
g %
none
detected
trace
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
none
detected
a\
6001-48
-------
(104,000 Ib/hr) of equivalent saturated steam. Figure 6-2 is a
schematic of the boiler. The three sampling ports are shown in the
vertical flue on the upper right side of the schematic just before the
flue gas enters the horizontal breeching leading to the stack.
The characteristics of the boiler/ the coal burned and the
hopper ash collected during the three trace species tests are listed
in Table 6-3. The load varied during from about 100 to 110 GJ/hr
during the testing, and the excess oxygen listed of 10.8% is an average
value for these tests. The hopper ash is the combined ash from the
furnace bottom and the dust collector. All of these ashes were
collected in a hopper beneath the boiler and removed periodically by
running them into dump trucks. As the section of the table labeled.
"Hopper Ash Proximate Analysis" indicates, a relatively large proportion
of the hopper ash was other than inert ash from the coal, and the
hopper ash had a significant heat of combustion. Consequently, a
good deal more "ash" was collected from the hopper than originally
went into the boiler as ash in the coal.
In comparison to other coal stoker units, the nitrogen oxides
emissions of 124 ng/J from this boiler were unusually low, since the
average nitrogen oxides emissions from all coal-fired boilers tested
during this program was 290 ng/J. The pollutant emissions during the
trace species test 166-11 are summarized below. During the testing
the load varied between 100 and 110 GJ/hr.
Load, GJ/hr (103 Ib/hr) 100-110 (95-104)
Total nitrogen oxides, ng/J (ppm) 124 (203)
Carbon dioxide, % 7.2
Carbon monoxide, ng/J (ppm) 32 (85)
One of the contributing factors to the reduced nitrogen oxides
emissions was the low-intensity, desultory flame. Also, since the
testing necessarily was conducted at relatively low loads and the heat
absorption area of the unit was comparatively large, the products of
68
-------
3 Sampling Ports
Chimney
Natural Draft Damper
Dust Collector
Inlet Damper
Coal Hopper
Dust Collector
Dust Collector
By-Pass Damper
Air Heater
By-Pass Dampers
Ash Collector
I.D. Fan
Discharge
Damper
NOT TO SCALE
Figure 6-2.
Boiler schematic showing gas flow paths and sampling
port location.
6001-48
69
-------
TABLE 6-3. BOILER, COAL, AND ASH CHARACTERISTICS
Location No. 35
Test No.
Date, 1975
fuel Type
Burner Type
Test load, GJ/hr
<103 lb/hr)
Average Excess Oxygen, %
Coal Burned kg (Ib)
Test Duration sec (min)
166-9
April 7
Crushed Coal
Chain grate
100 - 110
(95 - 104)
10.8
16023 (35251)
8100 (135)
166-10
April 3
Crushed Coal
Chain grate
100 - 110
(95 - 104)
10.3
17490 (38478)
8100 (135)
166-11
April 9
Crushed Coal
Chain grate
100 - 110
(95 - 104)
10.8
11030 (24266)
S400 (90)
Coal Proximate Anal
AS RECEIVED:
Moisture, %
Ash, %
'Volatile Matter, %
Fixed Carbon, %
Sulfur, %
Beat of Combustion, J/g
(Btu/lb)
DRY BASIS:
Ash, *
Volatile Matter, %
Fixed Carbon, %
Sulfur, %
Heat of Combustion, J/g
(Btu/lb)
4.11
12.46
38.08
45,35
2.81
27,173
(11,685)
12.99
39.71
47.30
2.93
28,339
(12,181)
/sis
4.19
12.32
38.35
45.14
3.23
27,341
(11.757)
12.86
40.03
47.11
3.37
28,536
(12,271)
4.17
16.33
36.72
42.78
3.10
25,992
(11,177)
17.04
38.32
44.64
3.23
27,122
(11,663)
Hopper Ash Proximate Analysis
Moisture, %
Ash, %
Volatiles, %
Fixed Carbon, %
Sulfur, %
Heat of Combustion, J/g
(Btu/lb)
1.16
55.39
4.01
39.44
1.58
20,190
(6,203)
1.76
31.52
5.32
61.40
2.98
22,197
(9,545)
1.04
58.89.
4.30
35.77
1.79
13,330
(5,732)
6001-48
70
-------
combustion were cooled rapidly. The combination of low intensity
combustion, low firing rate and a large furnace would be expected to
result in low nitrogen oxides emissions. The baseline solid particulate
emissions were only 133 ng/J (0.315 lb/10 Btu) and were relatively
low compared to the sixteen other coal-fired boilers that were tested
during the combustion modification part of the Program (see Figure 7-2,
Test No. 165 in the following section of this report).
6.2 COLLECTION TRAIN
After consideration of the results of the prior trace species
sample collections made at Locations 29 and 30, an ammonium-hydroxide-
based reagent was provided to KVB for bubblers 1 and 2. It was ex-
pected that with this reagent one could pump at the higher rate neces-
sary for isokinetic sample collection and still scrub out the mercury.
When the ammonium-hydroxide-based reagent was used, however, the trace
species train repeatedly plugged up, because the granular Tenax adsorbent
became hygroscopic and stopped the flow of exhaust gas through it.
The Tenax adsorbent had not picked up water heretofore. A num-
ber of special test runs and consultations resulted in the conclusion
that the cause was the new reagent used to trap the mercury. The flue
gas at Location 35 had such a high sulfur content that it was neces-
sary for the field test crew to wear gas masks when in the area of the
test ports. Apparently, the sulfur in the flue gas reacted with the am-
monium hydroxide in the first two bubblers and formed an ammonium-sulfur
compound that caused the Tenax to become hygroscopic. As the Tenax became
wet, it no longer was granular and the flue gas could not be pumped through
it.
The EPA Project Officer's representative then directed KVB, Inc.
to use a sodium carbonate-based reagent in the first two bubblers. With
the sodium carbonate reagent the trace species tests were run with no
further delay. The contents of the eight bubblers that finally were
selected were the following:
71
-------
Bubbler No. 1
Bubbler No. 2
Bubbler No. 3
Adsorbent
Bubbler No. 4
Bubbler No. 5
Bubbler No. 6
Bubbler No. 7
Bubbler No. 8
2% sodium carbonate in water
2% sodium carbonate in water
dry
Tenax
10% HNO + HO + AgNO in water
10% HNO. + HO + AgNO in water
O 4b £m J
10% HNO + HO + AgNO in water
dry
silica gel
The trace species collection train configuration that was
used was that discussed and pictured in Section 3 of this report.
The three test runs that are listed below were made using the sodium
carbonate-based scrubbing reagent.
Test Run No.
Test Duration, min
Actual flow rate through
train, cm3/sec (ft3/min)
Actual volume through
train, m3 (ft3)
166-9
135
171.4 (0.3632)
1.388 (49.03)
166-10
135
209.9 (0.4448)
1.701 (60.05)
166-11
90
314.7 (0.6667)
1.700 (60.00)
6.3 SPECIES CONCENTRATION
Concentrations of the individual trace species that were found
to be in the coal, in the combined furnace bottom and dust collector
ashes and in the flue gas as fly ash are listed in Tables 6-4, 6-5, and
6-6. Concentrations of antimony-could not be determined because of a
chemical interference. The second part of each table lists the mass of
the more volatile species, the organics and the sulfates that were
scrubbed out of the flue gas stream by the bubblers and Tenax adsorbent
of the trace species train. Bubbler reagents were analyzed only for the
more volatile species and the organics.
Mercury balances are based on an assumed average mercury content
in the coal of 0.15 Ug/g. Analysis of the coal indicated a mercury con-
tent of less than 0.3 yg/g. The assumed average of 0.15 Ug/g, one-half
the detectable limit, is a typical average for general coal types as in-
(23)
dicated in Figure 3-1.
72
-------
TABLE 6-4. SPECIES CONCENTRATION
Test No. 166-9, Location No. 35
Coal, Furnace Bottom and Dust Collector Ashes, and Flue Gas Fly Ash
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM 1$
POM 2*
POM 3$
POM 4|
Sulfate §
Location of Species
Coal
Fuel
(yg/g)
Could not be a
6.0
39
1.1
2.4
3,000
15
2.3
25
15,000
8.5
25
<0.3
4.6
6.5
1.4
790
54
30
565
117
j.
i
6.8
t
38
5,000
. Hopper
Ash
(yg/g)
nalyzed - chem
11.4
108
3.2
0.1
4,800
101
4.3
77
49,000
1.8
61
0.5
37
7.0
0.7
6,600
280
20
1,100
72
t
10.4
t
t
4,800
Flue Gas Fly Ash
(yg/g)
Leal interference
250
283
20
10
19,000
250
13
270
160,000
131
125
1.2
100
280
7.2
7,400
625
500
*
*
*
*
500
Insufficient sample for analysis
None detected
Continued
POM 1: 7,12 DimethyIbenz[a]anthracene; POM 2: Benzo[a]pyrene; POM 3;
3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
Sulfate is water soluble sulfate as SO
6001-48
73
-------
TABLE 6-4. (Continued)
Test No. 166-9
Flue Gas Vapors
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1*
POM 2*
POM 3*
POM 4*
Sulfate *
1
Na CO
(flgr
0.2
11.4
<0.3
56
21,000
7,300
t
t
t
J.
1
9.2x10°
2
"*&
0.4
3.3
<0.3
4.1
2,900
1,600
t
t
t
t
0.7xl05
Bubbler Number
Tenax
(ug)
§
§
§
§
§
§
t
t
t
t
§
4
$
<0.5
<3
0.1
0.17
§
§
§
§
§
§
§
5
$
<1.0
<7
<0.1
0.26
§
§
§
§
§
§
§
6
H 0
&7
<1.0
<7
<0.1
0.29
§
§
§
§
§
§
§
* Insufficient sample for analysis
t None detected
J POM 1: 7,12 Dimethylbenz[a]anthracene; POM 2: Benzo[a]pyrene;
POM 3: 3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
§ Not analyzed for
# Sulfate includes all sulfur compounds collected expressed as
6001-48
74
-------
TABLE 6-5. SPECIES CONCENTRATION
Test No. 166-10, Location No. 35
Coal, Hopper Ash and Flue Gas Fly Ash
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM It
POM 2$
POM 3$
POM 4$
Sulf ate #
Location of Species
Coal Fuel
(yg/g)
Could not be ai
7.9
28
1.0
1.1
2,200
20
2.1
42
19,000
8.6
26
<0.3
6.1
6.7
3.4
764
50
34
615
65
T
14
t
47
4,800
Hopper Ash
(ug/g)
lalyzed - chemical i
4.0
150
5.0
0.2
5,200
100
2.1
170
53,000
25
170
<0.7
37
.5.0
0.4
2,050
320
119
1,100
69
0.3
12.3
j.
i
t
4,900
Flue Gas Fly Ash
(ug/g)
nterference
110
290
26
14
17,000
280
11
340
130,000
129
144
1.0
124
110
11
7,550
645
982
*
*
*
*
*
*
400
Insufficient sample for analysis
None detected
Continued
f
r
POM 1: 7,12 Dimethylbenz[a]anthracene; POM 2: Benzo[alpyrene;
POM 3: 3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
Not analyzed for
Water soluble sulfate as SO .
6001-48
75
-------
TABLE 6-5.(Continued)
Test No. 166-10, Location No.
Flue Gas Vapors
35
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM It
POM 2t
POM 3t
POM 4*
Sulf ate *
Bubbler Number
1
Na CO
(pg)
-------
TABLE 6-6. SPECIES CONCENTRATIONS
Test No. 166-11, Location No. 35
Coal, Hopper Ash and Flue Gas Fly Ash
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM 1$
POM 2$
POM 3$
POM 4$
Sulf ate *
Location of Species
Coal Fuel
(yg/g)
Could not be an
5.4
50
1.1
0.7
2,200
23
2.3
22
19,000
4.3
20
<0.3
6.4
7.1
1.0
886
50
26
589
98
4-
1
1.7
t
61
4,900
Hopper Ash
(yg/g)
alyzed - chemical i
3.9
305
4.9
0.1
6,200
100
3.6
50
89,000
5.1
100
0.4
46
4.8
0.81
4,100
330
29
1,100
73
0.3
59
t
t
5,100
Flue Gas Fly Ash
(yg/g)
interference
250
470
23
14
19,000
490
18
301
140,000
185
139
<3
203
150
12.3
6,100
630
856
*
*
*
*
*
*
100
* Insufficient sample for analysis
t None detected
Continued
POM 1: 7,12 DimethyIbenz[a]anthracene; POM 2: Benzo[a]pyrene;
POM 3: 3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
Not analyzed for
Water soluble sulfate
6001-48
77
-------
TABLE 6-6. (Continued)
Test No. 166-11, Location No.
Flue Gas Vapor
35
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1$
POM 2$
POM 3|
POM 4£
Sulf ate *
1
Na CO
(jjg)
<0.1
2.5
0.5
56
24,000
6,100
a.
1
t
t
t
11x106
2
NaCO
dg)
<0.1
7.9
0.1
1.2
1,500
650
t
t
t
t
3.8xl06
Tenax
(yg)
§
§
§
§
§
§
t
t
t
f
§
4
HO
<1.0
<6
<0.1
0.09
§
§
§
§
§
§
§
5
*yg?
<1.2
<8
<0.1
0.20
§
§
§
§
§
§
§
6
(ygi
<0.5
<3
<0.1
0.01
§
§
§
§
§
§
§
Insufficient sample for analysis
None detected
POM 1: 7,12 Dimethylbenz[a]anthracene; POM 2: Benzo[a]pyrene;
POM 3: 3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
Not analyzed for
Sulfate includes all sulfur compounds expressed as SO .
6001-48
78
-------
SECTION 7.0
FIELD TEST NO. 169, MODIFIED COLLECTION TRAIN
7.1 TEST RESULTS
The results of the material balance for Test Run 169-3 at
Location 31 on a boiler burning pulverized coal are shown in Figure
7-1. The mass of the individual element collected in the boiler is
the sum of the masses found in the furnace bottom, dust collector and
flue gases, both as particulate and vapor.
Most of the moderately-volatile elements, such as barium and
manganese, concentrated in the bottom and collector ashes, and the
recovery was at least 80% of the material input with the coal.
The more highly-volatile elements, such as arsenic and zinc,
were not recovered either as ash or as vapor. The recovery was only
10% to 40% of the material originally contained in the coal.
One exception was vanadium. In both this test and in Test 166
discussed previously the recovery was high, from 88% to 120% of the vana-
dium in the coal. Forty-two percent of the vanadium that was recovered
came from the collector ash. Another twenty-six percent was
recovered from the flue gas fly ash. The vanadium tended to collect
in the dust collector ash, rather than in the bottom ash, which is
consistent with its being between moderately- and highly-volatile.
The collection of highly volatile mercury was only 40% of
the quantity input with the coal and was much lower than the 100%
recovery of Test 166. There was no apparent reason why this particular
test should have been less successful in collecting mercury, since the
same collection train and procedures were used for both tests.
The amount of the elements in the bottom ash was based on
a calculated amount of bottom ash that was generated. It was not
practical to measure the bottom ash, so the amount was taken to
be the difference between the ash in the coal and the amount in
79
-------
Mass of element collected/Mass of element in coal, %
20
40
60
80
100
120
140
Acceptable Balance
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Figure 7-1. Material balance, test run no. 169-3.
6001-48
80
-------
the flue gas upstream of the dust collector. A correction was
made to account for the fact that the ash actually contained only
about 95% inerts.
As with Tests 166, 213 and 214 more nickel and titanium were
collected than was contained in the coal. This surplus of nickel
and titanium was deemed to have come from the stainless steel sample
nozzle on the collection probe even though the nozzle was very short.
If it had come from corrosion of the metal parts of the furnace, then
the iron also should have been overbalanced. However, the amount of
iron ran from 80% to 100% of the amount in the coal, and thus it
appeared that corrosion of the boiler tubes was not a significant
source of the nickel and titanium that were recovered.
It was not possible to make a complete material balance for
chloride and fluoride because there was an insufficient sample of
particulate in the flue gas downstream of the collector. The quan-
tities determined are shown in Table 7-1.
Only 11 to 26 percent of the chloride and fluoride that
was in the coal was recovered from the ashes and stack gas. Ap-
parently, these two species passed through the boiler almost en-
tirely in a vapor form. The recovery during Test 169 was much less
than the 31 to 104 percent recovery that had been achieved previously
during Test 166 as is shown in Table 6-1. Since the collection reagents
in the bubblers were the same in both instances, the reason for the low
recovery during the current test is unknown. The dearth of chlorides
and fluorides in the bubbler reagents may have been due to the forma-
tion of free chlorine and fluorine and of hydrochloric and hydro-
fluoric acids. None of these four species would have been scrubbed
out in any significant quantity by the 2% sodium carbonate solution
in the first two bubblers nor by the chemical mixture in the fourth,
fifth, and sixth bubblers. While the foregoing rationalization appears
sound, the reason why it did not apply also to the earlier Test 166 is
unclear at present.
81
-------
TABLE 7-1. CHLORIDE, FLUORIDE AND SULFATE EMISSIONS
Test Run No. 169-3, Location No. 31
Chlorine
Fluorine
Sulfate*
In Coal
g %
8645 100
864 100
67379 100
In Bottom
Ash
g %
287 3
12 1
2373 4
In Collector
Ash
q %
360 4
10 1
3336 5
Stack Gas
Particulate Vapor
9 % q %
insufficient 1608 19
sample
insufficient 77 9
sample
492 1 203283 302
Total
Recovered
9 %
2255 26
99 11
209434 311
*Sulfate in coal, ash and particulate is water soluble sulfate as SO .
Sulfate in stack gas vapor includes all sulfur compounds as SO . 4
4
TABLE 7-2. ORGANICS EMISSIONS
Test Run No. 169-3, Location No. 31
No
1.
2.
3.
4.
?OM
Name
7,12 Oioa thy Ibcnz (a) anthracene
Banzo (a Ipyzene
3 Methylcholanthrene
Dibenz (a , h ] anthracene
In Coal
o »
none
detected
none
detected
8.90 100
12.71 100
In Bottom
Ash
a %
none
detected
none
detected
0.143 2
none
detected
In Collector
Ash
-------
The sulfates were also heavily concentrated in the vapor form
and passed through the boiler. In this case, however, the train
reagents were able to scrub out the vapors and the sulfate catch in the
train was very large indeed. Three times more sulfates were collected
than were in the coal. This increase probably was due to the
combining of the sulfur and alkali metals in the coal and their being
scrubbed out in the bubblers as a sulfate along with the sulfuric acid
from the sulfur trioxide.
Two species of polycyclic organic material (POM) were found in
the coal. POM 2, benzo[alpyrene was found only in the dust collector
ash and appears to have been a product of the combustion process.
This POM had been found in the coal and the hopper ashes in Test 166.
The emission of these organics is tabulated in Table 7-2.
The table indicates that POM's were not detected in the stack
gas. Since the melting and/or subliming point of these organics is
about 475 K or about 50 K above the stack 'gas temperature, these POM's
would be expected to condense on the solid particulate and would not
be present in the gas phase. Unfortunately, there was an insufficient
sample of stack gas particulate to determine if POM's were deposited
on the particulate.
7.2 BOILER AND FUEL CHARACTERISTICS
Boiler number 6 at Location 31 was an Erie City Iron Works
brand water tube boiler with four Combustion Engineering gas and coal
type RO burners installed in 1963. The pulverized coal burners
consisted of a centrally located coal nozzle tip with internal rifling
to impart swirl to the coal and air mixture. The combustion air
entered the burner throat through circumferrentially located vanes which
provided swirl to the combustion air. The superheated steam boiler
operated at a nominal pressure of 6.31 MPa (900 psig) and temperature
83
-------
at 756 K (900°F). The nominally balanced-draft furnace operated at
-75 Pa (-0.3 inches of water gauge) and had tangent tube wall construc-
tion. The combustion air was preheated to approximately 530 K (500°F)
with a tubular air preheater.
The tests were run at a boiler load of 148 GJ/hr (140,000 Ib/hr)
of equivalent saturated steam flow with normal boiler control settings
for that load.
This location was a good one for conducting trace species
emission testing because it was possible to collect samples both before
and after the dust collector. The dust collector was a mechanical
cyclone type. A disadvantage was that it proved to be impractical to
measure that rate at which ash was deposited in the furnace bottom or
the dust collector. It was not possible in actual practice to operate
the ash collection system in a way that the amount of ash accumulated
over a specific period of time from a particular batch of coal could
be determined.
The characteristics of the test boiler and the ashes during
the three tests are listed in Table 7-3. The fixed carbon content of
the ashes could not be determined because of the high iron content of
the samples. With those samples containing large amounts of iron, the
sample gained weight during the analysis, making a fixed carbon
determination impossible.
The measured total nitrogen oxides emissions from the boiler
during the.December 1974 combustion modification testing sequence were
found to be unusually high, i.e., 562 ng/J (922 ppm, dry at 3% 0 ) at
(8)
an excess oxygen level of 7.4%. The average emissions from all
coal-fired boilers were 290 ng/J at an excess oxygen level of 8.7%.
The emissions were rechecked in May 1975 during the trace species
testing and again were high at 534 ng/J (874 ppm, dry at 3% O ) with
an excess oxygen level of 7.0%.
84
-------
TABLE 7-3. BOILER, COAL AND ASH CHARACTERISTICS
Test 169, Location No. 31
T«*t No.
Date, 1975
Fuel Type
Test Load GJ/hr U03lb/hr>
Excess Oxygen, %
Coal Burned kg (Ib)
169-1
April 18
Pulv. Coal
143 (140)
7.0
15536 (34362)
169-2
April 18
Pulv. Coal
148 (140)
7.0
89S9 (19750)
169-3
April 19
Pulv. Coal
148 (140)
7.0
10311 (22733)
Coal Proximate Analysis
AS RECEIVED:
Moisture, *
Ash, »
Volatile Matter, %
Fixed Carbon, %
Sulfur, »
Heat of Combustion, J/g
(Btfl/lb)
DRY BASIS:
Ash, t
Volatile Matter, t
Fixed Carbon, %
Sulfur, t
Heat of Combustion, J/g
(Btu/lb)
7.39
10..79
38.75
43.07
1.15
25,281
(10,371)
11.65
41.34
46.51
1.24
27,297
(11,738)
7.65
10.13
38.59
43.58
1.03
25,550
(10,987)
11.02
41.79
47.19
1.12
27,666
(11,897)
6.85
10.78
33.26
44.11
1.28
25,897
(11,136)
11.57
41.07
47.36
1.37
27,801
(11,955)
Bottom and Collector Ash Proximate Analysis
Moisture, %
Ash, %
Vola tiles, \
Fixed Carbon, t
Sulfur, t
Heat of Combustion, %
1.43
96.76
3.65
(1)
0.47
(1)
1.56
95.99
4.44
(1)
0.55
(1)
1.42
92.44
6.11
0.03
0.56
(1)
Fly Ash Proximate Analysis
Moisture, %
Ash, %
Vola tiles, %
Fixed Carbon, %
Sulfur, %
Heat of Combustion, \
0.08
93.68
1.79
(1)
0.29
239
0.08
98.36
1.45
(1)
0.27
227
0.06
99.14
2.03
(1)
0.22
82
(1) Could not be determined duo to high iron content of sample
. 6001-48
85
-------
The gaseous pollutant emissions during the trace species
Test 169-1 are summarized below:
Pollutant Emissions
Total nitrogen oxides, ng/J (ppm) 561 (918)
Carbon dioxide, % 8.2 (estimated)
Carbon monoxide, ng/J (ppm) 0
Total particulate before dust
collector, ng/J (Ib/MBtu) 3,016 (7.014)
Total particulate after dust
collector, ng/J (Ib/MBtu) 1,068 (2.482)
The particulate emissions from this boiler are relatively high.
(8)
Figure 7-2 from the final report of the combustion modification results
shows the emissions of solid particulate from all coal fuel tests run
during the industrial boiler program, and the emission during Test 169
was one of the highest levels measured.
7.3 COLLECTION TRAIN
The trace species sample collection train that was used for
test series 169 is pictured and described in Section 3.0.
The contents of the eight bubblers were as follows:
Bubbler No. 1 2% sodium carbonate in water
Bubbler No. 2 2% sodium carbonate in water
Bubbler No. 3 dry
Middle filter Tenax
Bubbler No. 4 10% HNO + HO + AgNO in water
Bubbler No. 5 10% HNO, + HO + AgNO in water
Bubbler No. 6 10% HN03 + H202 + AgN03 in water
Bubbler No. 7 dry
Bubbler No. 8 silica gel
86
-------
Test No.
17
18
19
20
26
27
. 28
31
32
42
43
73
131
134
156
165
ezr^i69
.0
..;....;...;.:..;. ....:...;.......... . . . . ....... ..... :' .: .v.,.y.-'-i: _ ;.. v.-. . ... - . "^
.::.;.: ..;.:- ...... -:- >.' V. .- . .. . . .. .-. ,:vj
.:-;;,, ,,.,,...,, ....... .. ,, ..... , . . ]
i'v-:-:- - - -.-..-. ' . .- - -" 1
""]
...... 1
i
1
j
T
I
j
1 ! 1 II II I
i i | i i i.i i ' " " i
'05 .001 .005 .01 .05 0.1 0.5 1.0 5.0 10.0
lb/106 Btu
til III 111 III 1
1. 1 1 111 III III 1
512 5 10 20 50 100 200 500 1000 2000 5001
Solid particulate emissions, ng/J
Figure 7-2. Solid particulate emissions at baseload, coal fuel.
(8)
87
6001-48
-------
The train characteristics for each of the three runs of
Test 169 are listed below:
Test Run No.
169-3
Train Operation Duration/ min
Upstream of Collector
Downstream of Collector
103
73
Actual flow volume through
train in3 (ft3)
Upstream of Collector
Downstream of Collector
1.70 (60)
0.991 (35)
7.4
SPECIES CONCENTRATIONS
The concentrations of the various species throughout the
boiler are listed in Tables 7-4 through 7-6. The concentration of
antimony could not be determined because of a chemical interference.
The second part of each table lists the mass of the more volatile
species,, the organics and the sulfates that were scrubbed out of
the flue gas stream by the bubblers and the Tenax adsorbent in the
trace species train. The bubbler reagents were analyzed only for the
more volatile species and the organics.
88
-------
oo
TABLE 7-4. SPECIES CONCENTRATIONS
Test Run No. 169-1, Location No. 31
Coal Fuel and Ashes
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM 1*
POM 2$
POM 3$
POM 4$
Sulfate #
Location of Toxic Material
Coal
Fuel
(yg/g)
Could not be
<0.5
230
0.7
0.6
6,900
14
1.5
26
7,400
8.1
57
<0.3
<58
2.1
0.7
720
79
16
580
61
t
t
0.6
1.2
3100
Bottom Ash
(yg/g)
analyzed - che
2.8
1,300
0.3
1.3
77,000
69
3
113
85,000
45
480
0.4
80
7.6
3.0
8,200
240
145
580
30
t
0.3
t
t
4900
Collector
Ash '
-------
TABLE 7-4. (Continued)
Test Run No. 169-1
Flue Gas Vapor at Dust Collector Inlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1*
POM 2*
POM 3*
POM 4$
Sulfate #
Bubbler Number
1
Na^CO,
(Bg)3
15.8
1.9
2.3
7.6
20,000
12,000
T
t
0.003
j.
i
1.9x10°
2
Na CO
(Eg)
3.7
2.4
<0.2
0.8
700
320
f
t
t
t
1.6x10°
Tenax
(ug)
§
§
§
§
§
§
360
t
t
t
§
4
H 0
($g?
1.0
8.7
1.3
§
§
§
§
§
§
§
5
H 0
(2g?
1.3
9.1
1.2
§
§
§
§
§
§
§
6
H O
(2g?
2.0
11.8
1.3
§
§
§
§
§
§
§
Flue Gas Vapor At Dust Collector Outlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM I*
POM 2 +
POM 3 t
POM 4 $
Sulfate *
Bubbler Number
1
Na CO
(Jig)
12.7
1.6
5.2
14.3
15,000
7,400
t
t
0.002
t
2.1xl05
2
Na CO
(ug)3
3.7
0.5
<0.2
0.8
100
280
tj
t
t
t
0.2xl06
Tenax
(ug)
§
§
§
§
§
§
t
0.002
t
t
§
4
^
Sample Missing
5
"$
1.6
14.9
0.6
<0.4
§
§
§
§
§
§
6
H2°2
(iagf
1.4
6.2
0.7
<0.4
§
§
§
§
§
§
** Insufficient sample for analysis
t None detected
t POM 1: 7,12 DimethyIbenz[a]anthracene; POM 2: Benzo [a]pyrene; POM 3:
3 Methylcholanthrene; POM 4: Dibenz[a,h]anthracene
§ Not analyzed for
# Sulfate includes all sulfur compounds expressed as SO .
90
-------
TABLE 7-5. SPECIES CONCENTRATIONS
Test Run No. 169-2, Location No. 31
Coal Fuel and Ashes
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium .
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM 1 *
POM 2 *
POM 3 t
POM 4 *
Sulfate^
Location of Trace Material
Coal
Fuel
(Mg/g)
Could not b
0.95
330
0.67
0.9
10,000
12
1.0
15
7,900
5.3
67
<0.3
<61
2.5
0.7
800
63
19
600
60
T
t
0.5
t
5,500
Bottom Ash
(vg/g)
Collector
Ash
(ug/g)
a analyzed - chemical interfe
0.7
1,600
0.3
1.4
75,000
92
3.5
82
61 , 000
31
340
<0.7
55
12
3.5
8 , 100
230
132
590
21
t
0.3
0.3
t
4,900
1.0
1,900
0.22
0.5
83,000
98
3.0
81
71 , 000
140
570
0.6
<61
1.5
1.6
9,400
370
67
520
69
t
t
0.3
t
6,300
Flue Gas
Collector Inlet
(ug/g)
rence
<0.5
1,400
0.33
1.8
80,000
97
3.2
95
60,000
121
370
<0.7
<61
1.4
2.1
9,500
350
89
*
*
*
*
*
*
7,000
Flue Gas
Collector Outlet
(yg/g)
2.5
3,300
0.62
4.4
93,000
149
7.1
199
61,000
71
550
<0.8
<149
9.9
5.2
8,600
400
224
*
*
*
*
*
*
200
* Insufficient sample for analysis Continued
t None detected
% POM 1: 7,12 Dimethylbenz[a]anthracene; POM 2: Benzo[a]pyrene; POM 3: 3 MethyIcholanthrene;
POM 4: Dibenzfa^janthracena
II Water soluble sulfate. 6001-48
-------
Table 7-5. Continued
Test Run No. 169-2
Flue Gas Vapor At Dust Collector Inlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1 *
POM 2 *
POM 3 *
POM 4 t
Sulf ate ff
Bubbler Number
1
NaCO
(tfg)
<0.05
0.8
0.3
40
5,300
1,100
T
t
0.048
0.002
4.1xl06
2
Na2C03
dg)
<0.05
<0.06
<0.1
0.9
100
340
T
t
0.032
t
0.7x106
Tenax
(ug)
§
§
§
§
§
§
T
t
t
T
§
4
H2°2
dgf
1.5
4.2
<0.4
<1.0
§
§
§
§
§
§
§
5
H00.
(fef
0.5
4.5
<0.4
<1.0
§
§
§
§
§
§
§
6
H2°2
(Pgf
0.2
3.3
<0.4
<1.0
§
§
§
§
§
§
§
Flue Gas Vapor At Dust Collector Outlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1 t
POM 2 t
POM 3 t
POM 4 t
Sulf ate ff
Bubbler Number
1
Na CO
(£g)3
0.14
0.7
0.3
11.6
1,600
3,700
T
0.002
0.30
t
2.4xlOb
. 2
Na CO
(ug)3
0.16
0.5
0.8
1.0
100
250
T
t
T
J.
1
0.1x106
Tenax
(pg)
§
§
§
§
§
§
i
0.002
T
1
§
4
H 0
(ug?
1.5
5.3
<0.4
<1.0
§
§
§
§
§
§
§
5
H 0
(ug?
0.5
4.8
<0.4
<1.0
§
§
§
§
§
§
§
6
H2°2
0.2
0.1
<0.4
<1.0
§
§
§
§
§
§
§
* Insufficient sample for analysis
t None detected
$ POM 1: 7,12 Dimethylbenz[a]anthracene; POM 2: Benzo[alpyrene; POM 3:
3 Methylcholanthrene; POM 4: Dibenz [a,h]anthracene
§ Not analyzed for
# Sulfate includes all sulfur compounds expressed as 304.
6001-48
92
-------
vO
TABLE 7-6. SPECIES CONCENTRATIONS
Test Run No. 169-3, Location No. 31
Coal Fuel and Ashes
Species
Antimony
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Nickel
Selenium
Tin
Titanium
Vanadium
Zinc
Chloride
Fluoride
POM 1*
POM 2t
POM 3 +
POM 4 4=
Sulfate*
Location of Trace Material
Coal
Fuel
(pg/g)
Bottom Ash
POM 4: Dibenz[a,h]anthracene
ft Water soluble sulfate.
6001-48
-------
TABLE 7-6. (Continued)
Test Run No. 169-3
Flue Gas Vapor At Dust Collector Inlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1 *
POM 2 *
POM 3 *
POM 4 4=
Sulf ate #
Bubbler Number
1
Na CO
0.10
22.3
0.03
9.1
64,000
6,900
T
f
t
i
O.SxlO6
2
Na.CO,
(5g)3
0.12
4.4
0.01
0.6
400
600
t
T
t
X
i
0.3x106
Tenax
(yg)
§
§
§
§
§
§
T
t
t
t
§
4
H 0
<5g?
<0.2
0.01
<0.04
<1.0
§
§
§
§
§
§
§
5
H2°2
<0.2
3.8
<0.04
<1.0
§
§
§
§
§
§
§
6
H.O,
efe?
<0.2
3.9
<0.04
<1.0
§
§
§
§
§
§
§
Flue Gas Vapor At Dust Collector Outlet
Species
Antimony
Arsenic
Mercury
Selenium
Chloride
Fluoride
POM 1 *
POM 2 t
POM 3 *
POM 4 $
Sulf ate*
Bubbler Number
1
Na.CO,
(Sg)3
0.14
8.9
0.01
10.2
18,000
580
t
t
t
t
2.1xl06
2
Na_CO,
,2 . 3
(yg)
0.22
3.7
0.19
0.5
200
290
T
t
t
t
0.2xl06
Tenax
(yg)
§
§
§
§
§
§
T
t
t
t
§
4
H 0
(ilg?
<0.2
0.01
<0.04
<1.0
§
§
§
§
§
§
§
5
H 0
(yg?
<0.2
5.5
<0.04
<1.0
§
§
§
§
§
§
§
6
H 0
(ygf
<0.2
5.6
<0.04
<1.0
§
§
§
3
§
§
§
* Insufficient sample for analysis
^ None detected
t POM 1: 7,12 Dime thy Ibenz [a] anthracene; POM 2: Benzo [a]pyrene;
POM 3: 3 Methylcholanthrene; POM 4: Dibenz [a, h] anthracene
§ Not analyzed for
# Sulfate includes all sulfur compounds expressed as 304.
6001-48
94
-------
SECTION 8.0
SPECIAL REAGENT COMPARISON TEST
The reagents that had been used at Location 31 and 35 were er-
ratic in their ability to scrub out the volatile elements from the flue
gas. Consequently, a special test was designed where three sample col-
lection trains containing different sets of reagents that had been de-
veloped for the EPA by three different laboratories were used simul-
taneously by KVB to collect trace elements from the same boiler. The
objective was to use one coal having a known trace element content in
a boiler where all of the ashes could be collected and weighed indi-
vidually. It then would be possible to compare directly the collection
efficiency of each of the three reagent sets.
8.1 SAMPLE COLLECTION EQUIPMENT
The comparison test was set up in the combustion research labo-
ratory of KVB, Inc. using an eighty horsepower Ben B. Hood brand fire-
tube boiler. The salient advantage of this particular boiler was that
all of the ashes generated by a test could be identified and collected.
Prior to a test, the furnace tube, firetubes and flues were cleaned
out. After a test they were cleaned out again and all of the ashes
collected for weighing and analysis. The Foster Wheeler Company burner
had been modified so measured amounts of pulverized coal could be fed
from the hopper. The burner design was such that liquid fuel additives
could be injected into the primary air just upstream of the ignition
point. This test site was assigned the test location number 40.
95
-------
The boiler, its flues and the location of the sampling ports
are shown schematically in Figure 8-1. The four sampling ports are
located in the vertical flue section on the right where the flue gas
flow is downward.
One of the three collection trains had been developed by the
Midwest Research Laboratories in Kansas City, Missouri. This train
consisted of a cyclone, a glass-fiber filter, an adsorber filter and
eight bubblers in series. The contents of the bubblers are listed
on the schematic drawings of Train "M" that is shown in Figure 8-2
This train was used by KVB, Inc. for all of the previous trace species
testing in the field and is discussed in detail in Section 3.0.
Train M was used in its complete form which included the
Tenax adsorbent that was intended primarily for the removal of organic
materials, i.e., POM. The other two trains were intended to remove
only trace elements from the flue gas stream, since it was not the
purpose of this particular test to collect and analyze organic material.
The second train was designated Train T and it was basically
an Aerotherm Acurex Corporation High Volume Stack Sampler with reagents
that had been developed by TRW Systems, Inc. in the four impingers. '
The contents of the impingers are listed in Figure 8-2.
The third train was developed by R. M. Statnick, et al of the
Process Measurements Branch, Industrial Environmental Research Laboratory
of the U.S. Environmental Protection Agency at Research Triangle Park,
(41)
NC. It is also shown schematically in Figure 8-2. This train was
designated Train S and the bubbler reagents are listed under Train S
in Figure 8-2.
96
-------
f-
2.4
1
M
0.91
\
m
MM^»
y
m
1
-~
1 0.36 m
mmt
»~
*
*- 0.27.m. 0.36 m
^^. Smoke stack Flow rate:.^"**
3.35 m/sec
(11 ft/sec)
To baghouse
h
m
Vi
«*
MM^H
^
" i
**»
2.1 m
^Sampling
Ports
I
jocations
0 46 "m About 38% of the
Ben B. Hood
80 HP Firetube
Boiler
U
fly ash
mass was
collected by the
baghouse .
U
Figure 8-1. Test boiler installation.
6001-48
97
-------
TRAIN M
TRAIN T
TRAIN S
03
Heated Probe with Heated Probe with Heated Probe with
Kapton Liner"1*""" ""
i
Heated Cyclone f
Heated Fiber
_ , filter *
Bubbler Coilection
No. Reagent
1 2% Na^CO |
2 2% Na^CO^ |
Ten ax Adsorber
4 HO. AnKO & C^T.
- " UNO 3
5 H.O^, AgNO, £ f
* - »MU,J
6 H^OT. AgNO, & |
* " HNO^J
ft Si 1 ir-s fif.1 | _ _.. 1
KapLon iii ner -
"N Heated Cyclone /'
Heated 'Fiber
* Filter c
Water Vapor Condenser
Bubbler Collection
No. Reagent
1 H2°2 bzzzr:
2 (NH^)9So°fl' A9NO a HNO 1
3 3 i
3 oS2°8. AgN03 fi HNO f" ' "-
4 (NHy,)0S000, AgNO, & HNO^
'
Kapton iiiner * '
I
"\ Heated Cyclone f
Heated Fiber
' " * Filter t-^
Bubbler Collection
INo. Reagent i
1 2%N!2C03 t==
2 KMnO fi HNO |
4 3 1
4 KMnO. & HNO, f ~
)
Pump and Pump and
Meters Meters
Pump and
Meters
Figure 8-2. Sampling train schematics.
6001-48
-------
The Train S design was developed specifically for the collection
of mercury vapor from a flue gas stream that contains a high concentra-
tion of sulfur dioxide. In order to make the sampling procedure for
mercury compatible with streams containing 0.1 to 8% v/v of sulfur
dioxide, a prescrubber was used to selectively and quantitatively remove
all of the sulfur dioxide present. Experience had shown various alkaline
scrubbing media to be effective for sulfur dioxide removal, and
so a prescrubber containing saturated sodium carbonate solution was
employed in bubbler no. 1. The nitric acid-potassium permanganate reagent
was selected for the other three bubblers because of ease of preparation
in the field and the general availability of reagents.
The coal burned for the test was especially selected for its
mercury content of from 0.10 to 0.28 Ug/g. The Peabody Coal Company
office in St. Louis, Missouri made available to the program about
2,000 kg (2 tons) of unwashed coal from their Kentucky No. 9 mine.
In order to assure a high level of mercury in the boiler feed,the
coal also was "doped" with an aqueous solution of mercuric chloride
having a strength equivalent to 0.1 microgram of mercury per gram of
coal burned. Four liters of a solution were prepared that contained
0.1263 g of mercuric chloride in distilled water. A measured amount
of this solution was sprayed under nitrogen pressure into the primary
air of the pulverized coal burner during both test runs.
8.2 ANALYTICAL METHODS AND DATA
Analytical methods; Seven types of solid samples were analyzed
for their trace element content by the Midwest Research Institute of
Kansas City, MO. The seven solids were: coal, furnace tube ash,
firetube ash, lower stack ash, duct ash, baghouse ash, and the particulate
catch from the three mass sampling trains. The solid samples were pre-
pared for analysis by digesting 0.5 g of solid with 3 ml hydrofluoric
acid and 1 ml nitric acid in a Parr bomb at 130°C for 18 hours. Boric
acid was added (1.5 g) and diluted to 25 ml.
99
-------
A portion of each impinger solution was heated with permanganate
and nitric acid prior to analysis for mercury and selenium. Sulfuric
acid was added to another portion of each impinger solution and heated
with a reflux condenser before analysis for arsenic and antimony.
Acid oxidizing impingers were analyzed directly for ten other
elements: beryllium, cadmium, chromium, copper, lead, manganese, nickel,
silver, titanium, and zinc. A portion of the sodium carbonate impingers
were acidified before analysis for these others.
Mercury was analyzed by the cold vapor technique. Antimony,
arsenic, and selenium were analyzed by atomic absorption following
addition of sodium borohydride. The volatile hydrides that were formed
were swept into a nitrogen-entrained air-hydrogen flame and analyzed.
The ten additional elements were analyzed by conventional flame-atomization
atomic-absorption procedures.
Data: A summary of the boiler operational data for Tests 213
and 214 is given in Table 8-1. This table indicates the amounts of
coal burned and the ashes collected. The operational data for each
sampling train are presented in Table 8-2.
The trace element concentrations of the coal and ashes determined by
MRI are listed in Table 8-3. The trace element content of the particulate
collected by the probe liner, cyclone and filter, by the bubblers and by the
Tenax adsorber are listed in Tables 8-4, 8-5, and 8-6. The data shown in
Tables 8-3 through 8-6 were taken from Reference 40.
8.3 RESULTS
A number of difficulties were experienced in conducting these
laboratory tests to the extent that the precision obtained was less than
was desired. For that reason, the results in Table 8-4, 8-5, and 8-6
should be used only for qualitative rather than quantitative comparisons.
The trace element analyses of the coal and ashes is not covered by this
comment and the data in Table 8-3 are accurate and useful in evaluating
relative concentrations in the ash.
100
-------
TABLE 8-1. TEST BOILER AND FUEL DATA SUMMARY
Location 40
Test No.
Date, 1975
Fuel Type
Burner" Type
Test Load, GJ/hr (103 lb/hr)
Excess Oxygen, %
Boiler Operation, rain
Coal Burned, kg (li>)
Coal Doping with Mercury
Amount HgCl solution
used, g
Amount Hg injected, g
Mercury concentration in
coal, ug/min
Mercury concentration
injected, yg/rain
Effective total Hg
concentration in coal,
lug/rain
Boiler Operation, rain
Coal Burned, kg (Ib)
Ash Collected, kg (Ib)
Furnace tube
Firetubes
Breeching
Stack and flue
Bag house
Total Ash Collected
Stack Gas Speed,
m/sec (ft/sec)
Stack Gas Flow Rate, ,
std m/sec (std ft /sec)
Stack Gas Temperature,
K <°F)
213
August 25
Pulver. Coal
Pulverizer
2.19 (2.08)
7.2
129
156 (344)
1009.7
0.02310
169
289
458
129
156 (344)
4.535 (10.0)
6.123 (13.5)
0.2211 (0.487)
0.0981 (0.216)
8.618 (19.0)
19.59 (43.20)
214
August 28
Pulver. Coal
Pulverizer
2.15 (2.04)
7.4
153
192 (425)
787.7
0.01837
163
353
516
153
19^ (425)
5.443 (12.0)
6.350 (14.0)
0.2935 (0.647)
0.4733 (1.04)
7.484 (16.5)
20.04 (44.19)
3.35 (11.0)
.193 (6.86)
454 (355)
6001-48
101
-------
TABLE 8-2. TRACE SPECIES TRAIN CHARACTERISTICS
Location 40
Test No.
Test Duration, min
Train M
Train T
Train s
Actual flow rate through
train, m3/min (ft3/min)
Train M
Train T
Train S
Actual volume through
train, m3 (ft3)
Train M
Train T
Train S
Particulate collected by
probe, cyclone and filter,
g (lb)
Train M
Train T
Train S
213
88
78
88
0.01636 (0.578)
0.03610 (1.275)
0.02070 (0.7312)
1.440 (50.86)
2.816 (99.47)
1.822 (64.35)
1.810 (0.00398)
4.652 (0.01023)
1.302 (0.00286)
214
93
47
87
0.01827 (0.645)
0.03653 (1.29)
0.02113 (0.6942)
1.699 (60.0)
1.710 (60.4)
1.710 (60.4)
1.040 (0.00229)
2.658 (0.00584)
1.098 (0.00242)
6001-48
102
-------
TABLE 8-3. TRACE ELEMENT CONTENT OP COAL AND ASHES
Element
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Manganese
Mercury
Nickel
Selenium
Silver
Titanium
Zinc
Test
No.
213
214
213
214
213
214
213
214
213
214
213
214
213
2.14
213
214
213
214
213
214
213
214
213
214
213
214
213
214
Coal
Ojq/g)
<4 (3.S)t
<4 (3.9)t
<10 (14) t
<10 (14) f
1.2
1.4
<3 (0.9)t '
<3 (0.9)t
36
44
14
15
<20 (11) t
<22 (11) t
37
34
0.14(.ll-.28)t
0.13(.ll-.28jt
21
24
<5 (2.6)t
1.7
<4 (0.06)t
<4 (0.06)t
650
630
24
22
Furnace
Ash
(ug/g)
<2
<2
<2
<2
9.8
12.5
<4
<4
1,140
1,180
118
164
30
27
319
640
0.3
1.2
610
585
4.1
<5
1.8
1.1
5,900
6,900
102
190
Fire tube
Ash
(|ig/g>
<2
<2
7
6
7.9
9.8
<4
<4 .
130
126
125
87
65
39
408
127
1.7
2.3
78
51
19
20
<0.1
<0.1
5,200
4,400
136
140
lower
Stack
Ash
(ug/g)
<2
<2
18
15
6.4
8.6
5.2
7.1
307
272
162
113
110
74
690
340
2.8
3.7
150
96
52
35
3.0
0.3
4,400
4,200
2,200
700
Duct
Ash
(yg/g)
<2
<2
<2
<2
12
11.5
<4
<4
66
154
116
97
65
61
545
305
2.3
1.5
26
60
32
22
<0.1
0.2
7,400
6,900
3,500
1,400
Baghouse
Ash
(ug/g)
<2
*
<2
*
9.5
*
<4
. *
171
*
137
*
100
*
350
*
2.7
*
90
*
16
*
0.1
*
4,600
*
130
*
* Sample was lost j only total weight was recorded.
t Kentucky coal analysis by Illinois Geological Survey.
6001-48
-------
TABLE 8-4. TRACE ELEMENT ANALYSIS OF TRAIN M
Element
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Manganese
Mercury
Nickel
Selenium
Silver
Titanium
Zinc
Test
No.
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
Particulate
(M9/9)
<5
<5
<2
<2
20
19
<3
<3
152
88
159
90
48
46
213
176
3.9
2.1
45
37
6.1
6.1
<8
<8
9,700
9,900
267
192
Bubbler 1 Bubbler 2
Na2C03 N32CO3
(U9) (ug)
0.13 <0.1
0.77 <0.1
<0.5 1.1
<0.4 0.6
<8 <8
<8 <8-
<30 <30
<30 <30
<44 <44
<44 <44
3 126
8 38
<134 <130
<130 <130
7 40
11 <3
1.5 0.5
0.1 0.3
<33 <33
<33 <33
5.2 2.9
4.5 4.3
<27 <27
<27 <27
<120 <120
<120 <120
60 <30
45 <30
Tenax
(U9)
<0.5
<0.5
<0.8
<0.8
<2
<2
<3
<3
<19
<17
<25
<25
<41
<41
<2
<2
0.6
0.4
<10
<10
<1
<1
<8
<8
<40
<40
<15
<15
Bubbler 4 Bubbler 5 Bubbler 6
H2O2 H2O2 ^2°2
(ug) (pg) (wg)
<7 <7 <7
<7 <7 <7
<0.4 0.8 <0.3
<0.3 0.3 1.9
<6 <6 <6
<6 <6 <6
<22 <22 <22
<22 <22 <22
<37 <37 <37
<37 <37 <37
<12 <12 <12
<12 <12 <12
<102 <102 <102
<102 <102 <102
82 <23 <23
<23 <23 <23
<0.01 0.3 <0.01
0.7 0.3 0.5
<26 <26 <26
<24 <26 <26
* * . *
* * *
t t t
t t t
<100 <100 <100
<100 <100 <100
<30 <30 <30
<30 <30 <30
* A spike cannot be recovered from this solution.
f Impinger solution contains silver nitrate.
6001-48
-------
TABLE 8-5. TRACE ELEMENT ANALYSIS OF TRAIN T
o
ui
Element
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Manganese
Mercury
Nickel
Selenium
Silver
Titanium
Zinc
Test
No.
213
214
213
214
213
214
213
214
213
214
213
214
213.
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
Particulate
(ng/g)
<4
<4
4.0
3.0
20
23
4.0
1.7
92
129
130
124
99
97
115
214
2.4
2.6
52
44
7.3
6.3
19
9
9,700
13,000
220
230
Condenser
-------
TABLE 8-6. TRACE ELEMENT ANALYSIS OF TRAIN S
o
a\
Element
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Manganese
Mercury
Nickel
Selenium
Silver
Titanium
Zinc
Test
Mo.
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
213
214
Particulate
(yg/g)
<22
*
<11
*
22
23
<3
<3
123
138
65
113
41
33
209
216
3.1
3.8
38
35
6.5
6.5
<4
20
11,000
10,000
241
225
Bubbler 1
Na2CO3
(Uq)
<4
0.6
1.2
6.0
<8
<8
<3
<3
54-
47
<10
<10
<130
<130
30
29
3.4
1.3
<14
<30
4.4
4.6
<20
<20
<125
<110
47
35
Bubbler 2
KMn04
-------
Unfortunately, the difficulties experienced in the train comparison
tests involved both the sampling and the laboratory analysis. The parti-
culate catch and loadings may not have been representative since the trains
were not operated isokinetically. The sampling at nearly twice the
isokinetic rate may have had some impact on the retention of trace metals
by the bubblers. However, since the trains were operated under close to
similar conditions and sample times, the data is useful for qualitative
comparisons of train performance.
The major limitations in the laboratory analysis concerned the
analysis of the coal for mercury and low concentrations of elements in
bubbler solutions. The initial analysis of the coal for mercury content
indicated a concentration of 2.6 yg/g which is higher than any large
coal source in the United States. A very poor material balance resulted
from the use of this concentration and a reanalysis of the coal was
requested. The final analysis of 0.14yg/g is more typical of the coal
used with a resulting shift from a grossly underbalanced to somewhat
overbalanced condition for mercury.
The second limitation in performing a materials balance was the
very low concentrations in the bubbler solutions. Although the trains
were operated for what were considered to be more than adequate sample
times, a large number of elemental analyses proved to be below the de-
tection limits for these species. A reexamination of this situation
revealed that this condition would likely have existed even if the samp-
ling had been conducted isokinetically.
Acknowledging the limitations outlined above, the qualitative
data comparing the material balances for the three trains are shown
in Figure 8-3 and Table 8-7. Comparisons are made only for four of
the volatile elements which had been difficult to scrub out of the stack
gases. The balances that were ahcieved with these four are indicative of
the balances that could be obtained for the entire list of elements in
Tables 8-4, 8-5, and 8-6.
107
-------
Calculation
Methods
o
CD
Train M
n Material Balance, %
0 100
1
2
3
1
2
3
1
1
2
3
Antimony |:Xx
mm
:::.
r i
Arsenic
E 1
I
Mercury
***;
ill
Selenium XxX
Ii
Hi
111
11
111
i
lii
x-x-:4
:x:x:l
''' ''i
|:|
;:x :|
11
ill
Train T
Material Balance, %
0 100
1
'- 1
- 1
1
II
:::::x
i
I
f :::
; ; ;
11
III
I!
II
III
>X-Xi
III
Hi
::!:!:;
:::;!
11;
;i;iil
Train S
Material Balance, %
) 100 20
II
fxi:;:
~ m
i
- i
il
m
I:::
F:::
[:::
I-:-:-:-
Xv
.-.-..
...
:x:::
jSjiv:
11
ill
ill
11
11
:::x:x
:::::
..*...
.*....
x:x:::
:::::::
L
~
~
...-.:
m
Acceptable
Balance
'Acceptable
Balance
Acceptable
Balance
Figure 8-3. Material balances for the three sample collection trains.
6001-48
-------
TABLE 8-7. MATERIAL BALANCES BY THREE CALCULATION METHODS
Element
and
Test No.
Antimony
213
214
Arsenic
213
214
Mercury
213
214
Selenium
213
214
Train M
Calc. Method No.
1 2 3
Material Balance
% of input
39
41
11
12
107
81
94
112
8
8
10
10
113
79
64
90
12
10
10
11
179
81
77
90
Train T
Calc. Method No.
123
Material Balance
% of input
40
11
13
16.
105
131
105
157
12
10
13
10
72
72
60
69
25
29
22
16
155
183
86
102
Train S
Calc. Method No.
123
Material Balance
% of input
31
28
16
22
138
139
116
170
39
*
20
*
139
138
63
77
24
*
20
*
148
162
45
62
* The quantity of the sample was insufficient for analysis.
6001-48
109
-------
Since it was possible to collect, weigh and analyze for trace
element content all of the ashes, as well as the fly ash from the flue
gas, three methods of calculating a material balance were used. These
were:
1. Baghouse catch was used for the flue gas fly ash content,
rather than the particulate collected by the cyclone and
front filter as is conventional in an EPA Method 5
analysis.
2. The ratio of the mass of the baghouse catch to the mass
of the particulate catch of the train was used to calculate
the total grain loading of the entire flue gas stream.
3. The ratio of the total quantity of flue gas to the quantity
of flue gas actually sampled by the train was used to
calculate the total grain loading of the entire flue gas
stream.
Material balances were calculated for each of the three trains
using the three calculation methods described above. The balances shown
in Figure 8-3 and the percentages of closure listed in Table 8-7 are
based on data taken from Reference 40. An entry of 100% indicates that
one hundred percent of the element that was input with the coal was
recovered in the boiler.
The selenium balanced in about one-half of the calculations.
A balance of from 75% to 125% was deemed to be an acceptable balance.
The mercury balanced six times out of the eighteen calculations, and
five of these were from the Train M data. The antimony and arsenic did
not balance in any instance. These metals consistently were very much
underbalanced, and it appears that the reagents were incapable of
scrubbing out antimony and arsenic vapors from the flue gas.
The mercury balances were about evenly divided between
underbalanced and overbalanced with the five successful runs of Train
M. With the same train the selenium was predominately underbalanced;
however with Trains T and S it was more evenly divided between under-
and overbalance. Calculation method 1 produced balances that were
predominately above 100%; while methods 2 and 3 produced balances
that were usually below 100%.
110
-------
SECTION 9.0
ELEMENT PARTITIONING AND PARTICULATE ENRICHMENT
Combustion of coal and oil fuels releases the trace ele-
ments that' are contained in the fuel as vapors and particulates, and
certain elements tend to be concentrated in the ash that collects in
particular parts of the boiler and/or in fly ash of a particular size.
The tendency to concentrate in the waste that collects in certain parts
of a boiler commonly is called "element partitioning" and the tendency
to concentrate in particulates of a certain size is called "particulate
enrichment." Both phenomena were observed during the testing by
XVB, Inc.
Workers in the field have observed three classes of partitioning
in a coal-fired boiler: '
Class I: Twenty elementsAl, Ba, Ca, Ce, Co, Eu, Fe, Hf,
K, La, Mg, Mn, Rb, Sc, Si, Sm, Sr, Ta, Th, and Titended
to be deposited in the bottom ash or slag in a concentration
about seven times greater than in the coal fuel. These
elements were partitioned about equally between the bottom
ash and the collector inlet fly ash. They did not tend to
concentrate in a dust collector, and the concentrations in
the inlet and outlet fly ashes were about equal.
Class II: Nine elementsAs, Cd, Cu, Ga, Pb, Sb, Se, Zn,
and Sntended not to be incorporated into the bottom ash
and the concentration in the bottom ash was about one-third
of that in the coal. The concentration in the inlet fly ash
was about two times greater than in the coal, and in the
outlet fly ash it was about eighteen times greater.
Class III: Hg, Cl, and Br remained essentially completely
in the gas phase. (F may behave similarly.)
The other elementsCr, Cs, Na, Ni, U, and Vcould not be
definitely assigned to a class based on their data but appeared to be
intermediate between Classes I and II.
Ill
-------
Volatilization-condensation or adsorption mechanisms have been
proposed to rationalize the classes of behavior observed for the various
elements.
Class I. These elements are not volatilized in the combustion
zone, but instead form a melt of rather uniform composition
that becomes both fly ash and slag. The slag is removed
directly and quickly from the combustion zone, while the
fly ash remains in contact with the cooling flue gas. The
Class I elements remain in the condensed state, and hence
show little partitioning between slag, inlet fly ash, and
outlet fly ash.
Class II. These elements are volatilized on combustion.
Since the slag or bottom ash is removed from the combustion
zone, they have no opportunity to condense on the slag or
ash. They do, however, condense or become adsorbed on the
fly ash as the flue gas cools. These elements thus are
depleted preferentially from the slag (partitioning effect),
and preferentially concentrated on the outlet fly ash compared
to the inlet fly ash (particle size effect).
Although most Class II and Class III elements have relatively low
boiling points, and most Class I elements have relatively high boiling points,
there are enough exceptions to make it unlikely that the elemental boiling
point is the major factor determining elemental partitioning. For example,
Ca and Cu behave as if they have low boiling points, while Rb,Cs, and Mg
behave as if they have high boiling points. It is more probable that the
physicochemical properties of the elements and their chemical compounds
in the coal and combustion products, the nature of the coal-burning
process, and the mechanisms occurring in the control devices all
determine the behavior upon combustion. The following six formation
mechanisms have been proposed. '
1. Trace elements in coal are present in aluminosilicates,
as inorganic sulf ides, or as organic complexes.
2. On combustion the aluminosilicates are not decomposed.
Rather, they melt and coalesce to form the slag
or bottom ash and fly ash.
112
-------
3. During the initial stages of combustion, the conditions
in a coal particle and within its immediate vicinity
are probably reducing. Under these conditions the chemical
bonding between metallic elements and sulfur in sulfide
mineral inclusions or between the elements and the
organic matrix is broken/ and these elements form
volatile species. If the elements are dispersed in the
coal organic matrix, they become initially dispersed in
the gas stream when the coal is burned. Thus, even those
elements that are not as stable as the vapor at the
combustion temperature initially enter the flue gas
stream as a vapor.
4. The elements initially volatilized or dispersed in the
flue gas stream may be oxidized to form less volatile
species which then may condense or be adsorbed on the
fly ash as the temperature of the flue gas drops.
5. Since the slag is in contact with the flue gas for a
short time, and at a high temperature, the condensation
of volatiles on the slag is not great.
6. Increasing concentrations of Class II elements in samples
collected successively downstream in and near the
particulate control devices are a result of the particle-
size-dependent collection efficiencies of these devices
in which larger particles are more efficiently captured
than smaller particles. Thus, the finer particles
penetrating the control devices contain higher concen-
trations of Class II elements than the coarser, collected
particles.
Referring to paragraph number 4 above, the metallic sulfates
are not expected to condense on the fly ash. Once the temperature
of the flue gas drops below its dew point, sulfate ions can form, and
some of these ions then will form a metallic sulfate. However, this
reaction occurs principally in the cold liquid contained in the
bubblers, since the temperature of the free stack gas is above the
dew point.
The foregoing classification scheme correlates well with other
studies of the relationship between trace element enrichment and par-
(19)
ticle size in fly ash. It was reported that As, Cd, Cr, Ni, Pb, Sb,
Se, and Zn show a clear inverse relationship between concentration and
particle size. Except for Cr and Ni (in the intermediate group), these
113
-------
elements are in Class II. No trend with size or poorly defined trends
have been reported for Al, Ca, Co, Cu, Fe, K, Mg, Mn, Si, Ti, and V.
Of these elements, Cu is in Class II and V is in the intermediate group,
but the other nine are in Class I.
A volatilization-condensation or adsorption mechanism also has
(19)
been proposed to account for the particulate enrichment behavior.
Those elements which accumulate on the smaller fly ash particles are as-
sumed to be volatile at the temperature of combustion (about 1800 K); as
the flue gas cools, the volatiles condense or adsorb on the fly ash. Since
condensation and adsorption are surface phenomena and the smaller particles
have a larger surface area per unit mass than do the larger particles, the
concentration of condensed elements should be inversely proportional to
particle size. Those elements that are not volatile in the combustion
zone form the fly ash particles upon which the volatiles condense. The
elements that are enriched on the smaller particles usually have boiling
points comparable to or less than the 1800 K temperature of the combustion
(19)
zone.
Measurements made on individual fly ash particles using an argon
ion etching technique and X-ray spectrometric analysis indicated that sev-
eral of the elements which show preferential concentration in the finer
particles also demonstrate higher concentrations at the particle surface
than in the bulk of the particles.
From the observed enrichment-volatilization patterns of the
two groups of elements, from volatility data, and from the evidence
supporting surface condensation of elements, the following particulate en-
richment mechanism is proposed : (1) At maximum furnace temperatures the
moderately volatile Class I elements remain condensed as solids or melts and
the more highly volatile Class II elements become partially or totally
vaporized. (2) As the hot, flowing combustion gases and suspended
ash particles are cooled in the boiler( Class II elements adsorb and
condense evenly on surfaces of particles consisting mostly of Group I
114
-------
elements. Because the finer ash particles have a higher surface area
per unit mass than coarser particles, they will contain relatively
higher concentrations of Group II elements. Many Group II elements
are completely condensed, but the most volatile ones (for example, Hg
and Se) do not condense completely and pass through the particulate
control devices partially in the vapor state. (3) Increasing concen-
trations of Group II elements in samples collected successively down-
stream in and near the particulate control devices are a result of
the particle-size-dependent collection efficiencies of these devices
in which larger particles are more efficiently captured than smaller
particles. Thus, the finer particles penetrating the control devices
contain higher concentrations of Group II elements than the coarser,
collected particles.
9.1 ELEMENT PARTITIONING
The partitioning of the trace elements within the boiler is
illustrated in Figures 9-1 through 9-4. Three successive runs were
made for each test and the data points plotted on these figures are
the arithmetic mean of the clement concentrations of the three runs.
The individual data for Test 166 are listed on Tables 6-4 through
6-6 and for Test 169 on Tables 7-4 through 7-6.
For Test 169 at Location 31 samples were taken of the furnace
bottom ash, the dust collector ash and of the flue gas before and after
the collector. For Test 166 at Location 35 the dust collector was
built into the backpass of the boiler and it was possible to collect
samples only of the combined furnace bottom ash and collector ash and
of the flue gas downstream of the collector.
The moderately volatile Class I elements did tend to concentrate
in the furnace bottom ash, as proposed in the opening discussion of this
section. The concentration of the elements in the bottom .ash increased
by a factor of from two to six over the concentration in the coal for
two-thirds of the elements plotted in Figures 9-1 and 9-2.
115
-------
Test 166, Location 35
en
<
rn
I
EH
z
8
§
£3
500
250
0
20"
10
0
200000"
100000
0
200"
100
0
10000~
5000
0_
Coal
Barium
X X X X X
Cobalt
Iron-
Manganese
Titanium
X X X X
Ashes
7777,
Flue Gas
Figure 9-1. Partitioning of moderately volatile elements.
6001-48
116
-------
Test 169, Location 31
en
o
H
I
EH
Z
8
pa
5000
2500
50000
0
50CT
250
0
1000H
5000
0
Coal
Barium
-Cobalt
Iron-
Manganese'
Titanium
/
Furnace
Bottom
V777/
Y777,
Y///,
Y///,
Upstream
of
Collector
7XZ
In
Collector
Y///,
Downstream
of
Collector
Y7//,
Figure 9-2. Partitioning of moderately volatile elements.
6001-48
117
-------
Test 166, Location 35
tn
en
2
O
H
o
u
400
200
0
20
10
o
400"
200
0
2001
100
2
4ob~
200
20
10
0
ioo5"
500
0
1000"
500
Coal
Arsenic-
Cadmiuia-
Copper-
Lead-
Selenium
Tin
/ / / /
Sinc-
Ashes
-/ X / /
fa«M4MMjw
Flue
Gas
y/.
Figure 9-3. Partitioning of highly volatile elements.
6001-48
118
-------
Test 169, Location 31
§
u
§
u
10
5
0
Q_
200
100
0
205"
100
Q.
20
10
2.5
Q.
200
100
0
1000"
500
0_
Coal
Arsenic
.
Cadmium-
Copper.
Lead
S / S
Selenium
Tin.
Zinc
/ / /
Vanadium
s ,. s s x_
Furnace
Bottom
Upstream
of
Collector
Downs treai
In of
Collector Collector
X
Figure 9-4. Partitioning of highly volatile elements.
6001-48
119
-------
The concentrations of the various moderately-volatile elements
in the flue gas upstream of the collector and downstream of the collector
(Figure 9-2) were about the same, as theory predicts. There was. no
tendency to concentrate in the dust collector, again in conformance with
partitioning theory.
The highly-volatile Class II elements generally did not tend
to concentrate in the combined bottom and collector ashes of Test 166
shown in Figure 9-3, as would be expected. However, five of the seven
Class II elements, copper, lead, selenium, tin, zinc, did concentrate
in the bottom ash during Test 169, as indicated in Figure 9-4.
Partitioning theory predicts that the concentration of the
light elements in the inlet fly ash upstream of the collector would
be greater than in the slag and bottom ash. Figure 9-3 shows that
this concentration did occur for four out of the eight elements,
arsenic, copper, lead and vanadium. The concentrations of selenium,
and tin actually decreased, and the concentrations of cadmium and
zinc were unchanged.
In all instances, except the lead concentration of Test 169, the
volatile elements concentrated strongly in the fly ash downstream of the
collector outlet. Thus, the Class II elements did volatize upon combus-
tion and did concentrate preferentially on the outlet fly ash, as pre-
viously proposed.
Of the Class III elements that were sought, i.e., mercury and
chlorine, only mercury could be analyzed. There was insufficient
particulate collected to analyze accurately for chlorine. The mercury
exhibited two differing behaviors: (1) during Test 166 it increased
in concentration downstream of the collector from a level of less than
0.3 yg/g in the coal to a mean of 1.2 yg/g in the flue gas particulate
and (2) during Test 169 the concentration was relatively constant
at 0.4 yg/g at all locations within the boiler. This difference in
the concentrations of mercury is deemed to have been due to the flue
120
-------
gas temperature. For Test 169 the flue gas temperature was 440 K,
while for Test 166 is was only 375 K and more mercury was condensed
out and adsorbed onto the particulates.
Vanadium had not been assigned a class because it exhibited
a behavior intermediate between Classes I and II. It behavior is
shown on Figures 9-3 and 9-4, and its behavior, indeed, is intermediate
between the highly-volatile and the moderately-volatile elements.
9.2 PARTICULATE SIZE
The particulate sizing portion of the field test included
thirty runs on ten different boilers. Half of these runs were part
of the trace species measurement task and half were part of the
combustion modification tests. The results of the combustion modifica-
tion tests are not discussed here, however References 7 and 8 contain
information on the combustion modification tests.
A low speed flow type of cascade impactor that is described
in Subsection 3.5 was used to measure particulate aerodynamic diameters.
A cyclone was added upstream of the first stage of the impactor when
coal was the test fuel. The aerodynamic diameter, D , cited here
is the diameter of a spherical particle of unit density that is
collected with 50% efficiency by the impactor stage. In order to
forestall particulate rebound and reentrainment, the flue gas flow
speed through the impactor was reduced to about two-thirds nominal,
and this reduced flow increased the aerodynamic diameters of the
stages. The points plotted on the graphs in this section are the
aerodynamic diameter cut points of the stages that actually existed
during the test.
When analyzing the impactor data it was assumed that all of
the mass caught upon an impaction stage consisted of material having
aerodynamic diameters equal to or greater than the D for that
stage, and less than the D_- for the next higher stage. For the
first stage, or the precutter cyclone when one was employed, it was
121
-------
assumed that all of the captured particulate had aerodynamic diameters
greater than, or equal to the D for the stage or cyclone, but less
than an arbitrarily large diameter of 50 urn for oil fuel and 100 ym
for coal fuel.
The proportion of particulates in three size categories for
the trace tests with oil and coal fuels are listed in Table 9-1.
The size category of five-tenths of a micrometer or less was selected
because particles that are less than 0.5 ym in size tend to be inhaled
and then exhaled, rather than deposited in the alveoli. They do not
build up readily in the airways and possibly are not a serious health
hazard.
Particulates between 0.5 ym and 50 Um in diameter are inhaled
and either filtered out in the airways, deposited in the alveoli or
exhaled in various amounts. In this size range particles 3 ym or less
are potentially the most hazardous because they tend to penetrate deeply
(45)
into the respiratory tract.
The size range of 0.4 to 0.7 ym is of interest because
particulates of this size reduce visibility due to Mie scattering
of the sunlight by the particulates. This scattering is responsible
for a hazy atmosphere.
The combustion of oil fuel produced a larger proportion of
particulates having an aerodynamic diameter less than 3 ym than did
coal fuel. Thus, more of the particulate emissions from oil is
inhaled and exhaled (24%) , retained in the respiratory passages
(67%), and involved in reduced visibility (7%) than of the emissions
from coal.
The chain grate type of coal burner of Test 166 at Location
35 produced more fine particulate (about 36%), than did the pulverizer
of Test 169 at Location 31 (about 24%). That the greater proportion of
fine particulate was from the chain grate, rather than the pulverizer,
was unexpected. The difference in particulate size between the chain
grate and pulverizer is not explained easily by the difference in the
122
-------
TABLE 9-1. PARTICULATE SIZE
Oil, FUEL
Test
No.
121-9
121-10
121-11
171- 6A
171-6B
171-8
Location
29
20
Load
CJ/hr
U03lb/hr)
76 (72)
76 <72)
76 (72)
53 (50)
S3 (50)
54 (64)
Burner
or Oil
Type
No. 6
No. 6
Ho. 6
Ho. 6
Ho. 6
Ho. 6
Proportion of Total Weight of Catch
Particles
Inhaled
Then
Exhaled
<0.5 ua
%
10
15
3
40
37
35
Particles
In The
"Fine"
^articulate
Size Rang*
<3 um
\
68
70
30
73
67
65
Particles
Reducing
visibility
by Mie
Scattering
0.4-0.7 un
%
a
19
1
2
2
2
Soot
Included
Yes
Yes
Yes
Yes
Yes
Yes
Test Conditions
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
COAL FUEL
166-5
166-6
168-7
166-9
166-10
169-1
169-2
169-3
169-4
169-6
35
31
111 (105)
111 (105)
116 (110)
106 (100)
116 (110)
148 (140)
148 (140)
148 (140)
148 (140)
148 (140)
ChCrt
Pulv.
46
25
6
5
5
3
1
1
1
<1
65
33
30
22
25
30
30
11
31
17
13
3
. 2
2
4
1
<1
3
2
<1
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Ye*
Baseline . Down-
stream of Oust
Collector
Baseline, Up-
stream of Dust
Collector
Baseline, Down-
stream of Dust
Collector
Baseline, Up-
stream of Dust
Collector
Baseline, Down-
stream of Dust
Collector
Baseline, Down-
stream of Dust
Collector
123
6001-48
-------
size of the coal fired, since the larger size crushed coal that was
fired on the chain grate would be expected to yield thelarger particu-
late, contrary to observation.
At a diameter of 10 ym and less the order was reversed. The
pulverizer produced a greater amount of particulate having a diameter
less than 10 jam, 50%, than did the chain grate, 35%. These data are
similar to other reported data where 30% of the particulate from a pul-
verizer was under 10 ym while only 10% from a spreader stoker was under
10 ym in diameter. With pulverized coal, there was a greater pro-
portion of particulate below the 10 ym size cut as compared with the
crushed coal.
There was little difference in the proximate analyses of
the two coals burned at Location 31, Test 169 and Location 35,
Test 166. The averages of the three proximate analyses of the coals
was as follows:
Location No.
Test No.
Inerts , %
Volatile Matter, %
Fixed Carbon, %
Heat of Combustion, J/g
Coal
35
166
13
38
44
26,835
31
169
11
39
44
25,576
Bottom Ash
35
166
49
4
45
31-
169
95
5
<1
However, the proximate analyses of the two ashes were far
different. The inerts content of the bottom ash from Test 169 was
much higher than that from Test 166, while the carbon content was
practically zero. The pulverizer burner completely burned out the
carbon, leaving both a bottom ash and fly ash that was almost entirely
uncombustible mineral. The chain grate of Test 166 suffered from
undersized fans and blocked gas passageways which contributed to the
45% carbon carryover. The pulverizer ash remained or agglomerated
into larger size particulates than did the ash from the chain grate.
The smallest proportion of fine particulate was 8% from a
spreader stoker tested as part of the combustion modification task.
There was no ash analysis nor was the soot blown during this test,
124
-------
so there can be no direct comparison, as in Tests 166 and 169 above.
It was not surprising, however, that a spreader stoker that burned
crushed coal partly in suspension, as did the pulverizer, and partly
on a grate, as did the chain grate, would produce relatively few fine
particulates and many large particulates.
The five individual runs on chain grate-fired coal of Test 166,
Runs 5 through 10, are depicted on Figure 9-5. All of the measurements
were made downstream of the dust collector, since the dust collector
was built into the back-pass of the boiler and the flue gas up stream
of the collector was not accessible.
With the exception of Run No. 5 the fine particulate
proportion ranged only from 25% to 33%. At 0.5 ym diameter there
was moderate scatter of the data. The relatively large proportion
of the mass on the first three stages in Run No. 6 may be due to
particulate rebound and reentrainment. The data are typical of a
situation where some of the larger particles that belonged
on the second through the fourth stages had rebounded and ended up
on the filter stage at the outlet.end of the impactor. Run No. 5
appears to be a complete anomaly.
The data from Test No. 169, Location 31 are entered on
Figure 9-6. The triangles with the base down and connected by the
solid curve are data points taken upstream of the cyclone dust
collector. The inverted triangles and dashed curves are data taken
downstream of the dust collector. The striking feature compared to
Figure 9-5 is the much smaller proportion of fine particulate.
On the average, for the five runs, the proportion of fine parti-
culate downstream of the dust collector was larger than the proportion
upstream, 26% and 21% respectively. The difference was caused, possibly,
by the dust collector's removing more of the larger particulate than the
smaller. The amount of particulates from pulverized coal that was 0.5 ym
diam. or smaller was much less across the board than the amount from
the chain grate and from oil fuel. The entire particulate size spec-
trum definitely was weighted toward the larger particulate sizes.
125
-------
100
Run No.
Fine Particulate
I
o
0.3
0.1
Test No. 166, Location No. 35
-Chain Grate Fired
0.1 0.3 0.5 1.0 3.0 5.0 10
AERODYNAMIC DIAMETER,
30 50
100
Figure 9-5. Particulate size distribution, coal fuel and chain grate
burner.
126
6001-48
-------
100
o;
en
a)
en
H c
(0
cn
o
ex,
3
D
D
U
(U
c -u
o a)
4J 05
M -H
O T3
&
0)
H
10
1.0
0.1
Rur
i Nc
i
3
2
4
6
).
/>
X
y
/
)/
If
T
*
^
/
/
^
>
/
\i
"I
7
ij
*
^
r
k 2 .
1
' r
f
"ine PC
/
Jl
-&T*
irt.
MM
I/
M
//
i
.-fj"r /i
^CHZ
yv'
r ' //
1 //
/(;
'P
J T
..,«.. ^
est
ulv
LCI
P
V7
r
Cs
N
er
il
^
/
X
o.
i:
a1
*»
f
**
:e
:e
«
««
1
r
-
p
*
6
}
pi
»
9,
FJ
p
:^§C
.**tLJ"
. ^i-
Loca
.red
un No
2,4
r 1
£1
tio
n I
M 1 1
Line Symbol Sampling
- J^T- Downstream
collector
^ Upstream of
collector
1 i
Jo
*
31,
Location
of dust
dust
0.1 0.3 1.0 3.0 10
AERODYNAMIC DIAMETER,
30
100
Figure 9-6. Particulate size distribution, pulverized coal fuel.
127
6001-48
-------
As with Test No. 166, there were two types of distribution:
one convex with a rapid increase in the cumulative proportion up
to about 2 ym, and the other s-shaped. The distribution type was not
unique to the sample collection site, upstream or downstream; both
locations had both types of distribution.
The findings of the size measurements for Tests 121 and 170
with No. 6 oil fuel are plotted on Figures 9-7 and 9-8. It is assumed
arbitrarily that 100% of the impactor catch was 50 jam or smaller in
diameter; although the largest cut point of the impactor was about
4 ym. The significant difference between these oil data and the coal
data shown in the two proceeding figures was that the proportion of
submicrori and fine particulate from oil burning was greater than that
from coal by a factor of about 10.
Test 121, Runs 9 through 11, in Figure 9-7 illustrates the
size effect of soot blowing during a test. For test runs 121-9 and
121-10 the soot blowing was timed so only the soot deposited during
the run was caught by the impactor. For run 121-11, an operational
problem caused about 18 hours of soot accumulation to be caught,
rather than the 4 hour accumulation of runs 121-9 and 121-10, and
the total catch shown in Table 9-2 was 18.99 mg. The result was
that the submicron particulates constituted only 3% of the total
catch, and a great many more large-size particulates were caught.
Apparently there was a significant growth in the size of soot par-
ticles by agglomeration over a period of time.
The proportion of the submicron size particulate 0.5 ym
or less in diameter and of the fine particulate 3.0 ym or less were
about the same from the two runs with light soot. At a diameter of
1.0 ym there was considerable difference in the proportion, but this
difference did not persist beyond 2.0 urn.
The amount of fine particulate from the oil-fueled boiler at
Location 20, Test 170 was about the same as that shown in Figure 9-7
for Test 121. The individual data points for four of the runs of
Test 170 are plotted in Figure 9-8. There was very little scatter
in these data, except for Run 5A. The filter stage of Run 5A was
damaged during the run and the data were not included in Figure 9-8.
128
-------
id
06 4>
gin
o
< .c
Oj -P
2 to
H C M
2 W 43
O «
H (!) dC
H
2
C V4
O <1J
O -H -P
05 4J
D
I «
U -H
inn
30
10
3.0
1.0
0.3
0.1
Run
-
No
L-g
/
=v
/
s,
t
ilO
p
T
t
.
11
>
y
./
/
/
/
/
/
f
\
i
<
/
i
p
k
^
_^x"
xU/
P
/
/
y
<
C*(
r-f
^*~
^
^^BHiai
MMil^HI
-J^
^W.
>
] i
v
J
FJ
Run No.
Rf9
T-
n
e !
10 -
>articu
lat
ei
' Test No. 121, Location No.
No. 6 oil, steam atomized
i ititiiii i 1111
2
9
0.1 0.3 1.0 3.0 10
AERODYNAMIC DIAMETER, ym
30
100
Figure 9-7. Baseline particulate size distribution, No. 6 oil fuel.
6001-48
129
-------
jRun No.j
& -o
tC. 0)
u -u
U -H
100
30
10
3.0
1.0
0.3
0.1
1
I 1
No
6A
e
8?
SB
.
«^>
9
>
ss
J
r
^y
V 1
MI
f
f
.. x-r4-
'.'3S£
CT
<^
Test
No.
I
F
r
^^
I i
1
s 170
6 oil
i
F
-1
'
*
Lr
1 ^n . fin . s:
ie
P
*
artici
75, Locati
steam atom
Hill
ilat
on
ize
:e
No
d
2
0
0.1 0.3 1.0 3.0 10
AERODYNAMIC DIAMETER,
30
100
Figure 9-8. Baseline particulate size distribution, No. 6 oil.
6001-48
130
-------
Table 9-2. CASCADE IMPACTOR DATA SUMMARY
Test
No.
Ill
112
130
Ul-»
121-10
121-11
139-5
1SC-J
166-3
166-5
166-6
166-7
16U-U
168-10
160-5'
162-11
162-36
169-1
1C9-2
169-1
169-4
169-6
171-6A
171-6B
171-8
170-5A
175-58
176-5
179-4
loc.
No.
27
27
28
V,
29
29
30
11
35
35
35
35
35
35
36
16
36
31
11
31
31
31
20
20
20
20
20
37
37
_^_
Fuol
Type
PS300
PS 300
No. 6
llo . o
llo. 6
No. 6
Coal
Coul
Coal
Coal
Coal
Coal
Coal
Coal
No. 2
Ho. 2
No. 2
Coal
Coal
Coal
Coal
Coal
llo. 6
No. 6
No. 6
No. 6
No. 6
Ho. 6
NO. 6
Burner
Type
Steam
Steam
Steam
Utuam
Rtnnm
Steam
Spread
I'ulv.
Grata
Grate
Grate
Urate
(irute
Grate
Stnam
Steam
Steam
Pulv.
I'ulv.
Pulv.
Pulv.
Pulv.
Steam
Steam
Steam
Steam
Steam
Steam
Steam
Test
Load
GJ-hr"1
90
90
31
'Id
74
126
53
420
116
111
111
I'.S
103
116
58
65
65
148
148
140
148
148
53
53
54
68
67
34
34
Impact.
Flow
c»3-s-l
35.4
37.7
11.8
17.8
20.0
17.9
46.5
51.5
24.6
22.2
25.5
23.6
23.6
24.1
28.3
28.3
28.3
23.7
25.7
25.7
26.4
23.7
26.1
27.6
26.1
29.3
34.5
22.6
22.7
Actual DSO of Stage No.
~ 1
tun
2.9
2.8
5.0
4.1
3.85
4.1
2.5
2.5
3.5
3.6
3.4
3.5
3.5
3.5
1.2
3.2
3.2
3.53
3.39
3.39
3.35
3.53
3.36
3.27
3.36
3.18
2.93
1.62
3.61
2
(la
1.7
1.7
3.0
2.4
2.11
2.4
1.5
1.4
2.1
2.2
2.0
2.1
2.1
2.1
1.9
1.9
1.9
2.12
2.01
2.01
2.01
2.12
2.02
1.96
2.02
1.91
1.76
2.17
2.17
3
Urn
1.2
1.1
2.0
1.6
1.54
1.6
1.0
0.96
1.4
1.5
1.4
1.4
1.4
1.4
1.3
1.3
1.3
1.41
1.36
1.36
1.34
1.41
1.35
1.31
1.35
1.27
1.17
1.45
1.44
4
u»
0.58
0.56
1.0
O.U1
0.77
0.81
0.5
0.40
0.70
0.73
0.68
0.71
0.71
0.70
0.64
O.64
0.64
0.71
O.bO
O.f.H
O.fi7
0.71
0.67
0.66
0.67
0.64
0.59
0.72
0.72
5
|4m
0.29
0.28
0.4
0.41
o. an
0.41
0.25
0.24
0.35
0.36
0.34
0.35
0.35
0.35
O.12
0.32
0.32
0.35
0.34
0. 14
0.34
0.35
0.34
0.33
O.34
0.32
0.29
0.36
0.36
"50
Cycl.
|IU
~
10.9
10.4
15.6
15.8
14.7
15.3
15.3
15.2
15.3
15.4
14.7
14.5
15.3
--
~
Cyclone, Staqn and Filter C.itch
Cyclone
mg
None
None
t.one
None
No no
None
75.1
52.6
10.39
32.3
30.01
13. 06
10.82
34.15
Nono
Nona
None
289.2
64.49
67.83
72.39
71.89
Nona
None
None
None
None
None
None
Stage Ho.
i
mg
0.368
0.038
2.064
1.00
1.82
11. /
8.33
57.7
1.056
6.54
6.18
3.22
0.200
0.768
96.9
5.60
10.6
48.02
32.55
5.49
35.20
40.63
0.588
1.052
1.060
1.792
3.824
2.812
3.112
2
mg
0.148
0.246
1.152
i.oa
2.20
2.62
5.49
34.2
0.780
2.75
2.32
4.60
0.276
3.42
7.55
0.148
3.57
92.46
26.90
2.09
32.00
20.96
0.616
O.4BO
0.501
0.852
1.660
1.616
1.372
3
mg
0.240
0.444
O.896
1.21
0.610
1.84
0.488
0.82
0.660
1.176
1.01
1.38
0.472
2.58
0.400
0.064
0.788
55.95
19.60
2.05
19.42
8.11
0.332
0.256
0.256
0.512
0.924
0.668
0.596
4
»9
0.272
0.296
0.880
0.992
1.06
1.64
0.396
6.69
0.340
0.714
1.33
1.21
2.10
1.20
0.756
1.032
0.62
33.24
4.82
2.74
6.73
1.36
0.172
0.172
0.168
0.164
0.372
0.456
0.332
5
mg
0.216
0.304
0.784
0.408
2.17
0.848
0.528
3.04
0.600
32.20
0.916
0.772
3.26
4.88
o.oon
0.024
0.034
6.48
2.15
1.50
3.04
1.02
0.140
0.068
0.108
0.256
0.312
0.312
0.196
Filter
WJ
0.864
H
0.328
0.400
0.500
0.348
0.120
0.236
0.572
33.91
13.74
1.22
1.29
1.55
O. 1611
0.192
0.20
11.91
0.600
0.336
0.200
0.156
1.160
1.172
1.120
0.288
3.158
2.700
2.344
Total
Catch
9
2.11
1.33
6.10
5.17
a. ea
18.99
90.5
163
14.60
109.8
55.51
25.46
18.42
48.51
1OS.C
70.6
}5.842
537.27
151.112
82.92
168.99
144.13
3.008
3.200
3.213
3.864
10.250
8. 564
7.952
Comments
Baseline
Filter destroyed
Baseline
Light soot
Uoht *oot
Heavy soot
Baseline
Baseline
Baseline no soot
Toxic with soot
Toxic with soot
Toxic with soot
Low NOx, no soot
Toxic with soot
lowor Load
Low NOx
n^fj^l ln<»
Toxic. Upstream dust collector
Toxic, Downstream dust collector
Toxic, Upstream dust collector
Toxic, Downstream dust collector
Baseline, Toxic
Baseline. Toxic
Baseline, Toxic
Baseline
Registers
Baseline
Low NOx
6001-48
-------
The complete cascade impactor test data for both the trace.
species tests and the combustion modification tests are tabulated in
Table 9-2.
9.3 PARTICULATE ENRICHMENT
There was an inverse relationship between the element concen-
trations and the particulate sizes for the highly-volatile elements
of Class I, as predicted by the enrichment theory discussed in the
introduction to this section. The less volatile Class I elements
exhibited little or no enrichment, again in accordance with theory.
The results of the particulate enrichment portion of the
Tests 166 and 169 are shown in Figures 9-9 through 9-12 . The enrich-
ment of the volatile elements, such as antimony, cadmium, chromium,
lead, and zinc, was pronounced during Test 169, as indicated in
Figure 9-12. During Test 166 the enrichment of cadmium, lead, and
zinc was erratic, as shown in Figure 9-11, but the general trend was
in accordance with enrichment theory.
Although the less volatile Class I elements did experience
some enrichment of the fine particulate, as shown in Figures 9-9
and 9-10, the trend was less pronounced than it was for the more
volatile Class II elements.
Figures 9-10 and 9-12 illustrate the effect of a dust collector
on particulate enrichment. The effect with both classes of element
was to increase the enrichment of the smaller size particulates, except
for copper, lead and manganese. Manganese was unaffected by the
collector, while the concentrations of copper and lead in the smaller
particulates was reduced, rather than increased, by the dust collector.
A phenomena that occurred in the particulate enrichment
graphs was the peak in concentration that occurred for certain elements
at a diameter of 1.5 ym. These elements were cadmium, chromium, lead,
and zinc. The enrichment of these four elements was somewhat bimodal
and reminds one of the well-known bimodal mass-size distribution of
(46)
urban aerosols. The two phenomena probably are not related,
however, since the urban aerosol mass-size concentration peaks at
diameters of 0.5 and 10 ym, rather than near 15 ym.
132
-------
Test 166, Location 35
CP
3.
Z
o
H
I
El
Z
I
U
H
2000
1000
10000
5000
10000
5000
50000
25000
2000(7
10000
0
2000"
1000
. 0
- Barium
Cobalt-
.Manganese.
"Titanium
Vanadium
'Beryllium'
Fine Particu
Late
0.1 0.3 1.0 3.0
PARTICULATE AERODYNAMIC DIAMETER,
Outlet from collector
10
20
Figure 9-9. Particulate enrichment by moderately volatile elements.
6001-48
133
-------
Test 169, Location 31
o>
§
H
I
8
2000
1000
.Fine Particulate
0.3 1.0 3.0
PARTICULATE AERODYNAMIC DIAMETER, \im
10
20
Inlet to collector
Outlet from collector
Figure 9-10. Particulate enrichment by moderately volatile elements,
6001-48
-134
-------
Test 166, Location 35
2000
1000
0
200"
100
Cn
500
0
4000"
EH 2000
a
o
§
EH
2
0_
10006"
5000
0_
200~
100
0
50000"
25000
Antimony
Cadmium
Chromium-
Copper.
-Lead
-Tin
-Zinc
0.1
\
\
0.3 1.0 3.0 10
PARTICULATE AERODYNAMIC DIAMETER, ]Jm
20
Figure 9-11. Particulate enrichment by highly volatile elements.
6001-48
135
-------
Test 169, Location 31
cr
3.
2
O
a
u
3
O
u
f-t
s
1000
500
Fine Particulate
i i r i i M
0.1
0.3 1.0 3.0
PARTICULATE AERODYNAMIC DIAMETER, pin
20
Inlet to collector
-^ Outlet from collector
Figure 9-12. Particulate enrichment by highly volatile elements.
6001-48
136
-------
The particulate size data for the two tests are tabulated in
Tables 9-3 and 9-4. Each test consisted of three runs with a cascade
impactor that was made up of a cyclone followed by five collection
stages, as is described in Section 4. An individual run did not
collect enough particulate for an accurate analysis, so the catch
of a given stage for all three runs were combined to yield an adequate
quantity of sample for analysis. Consequently, the entries in the
tables are an average concentration for a set of three successive runs.
No concentrations are reported for the backup filters, because
the mass of the particulate catch was so small that the concentration
of the trace species collected was indistinguishable from the
background of the filter impurities.
137
-------
TABLE 9-3. PARTICULATE SIZE AND ENRICHMENT
Test No. 169, Location No. 31
Inlet to Dust Collector
Trace
Species
Antimony*
Barium*
Beryllium"*"
Cadmium'
Chromium*
CobaltT
Copper "f"
Lead"!*
Manganese i"
Nickel*
TinT
Titanium*
Vanadium*
Zinc*
Concentration, Ug/g
Aerodynamic Diameter, \aa.
Cyclone
15 Um
<35
1,100
2.9
3.9
110
49
170
10
1,100
330
1.4
7,700
150
180
Stage 1
3.4 Um
<130
770
4.2
5.0
110
65
310
270
900
47
4.0
8,400
<850
400
Stage 2
2.1 Um
<66
900
2.6
5.2
62
65
280
260
800
40
3.0
7,000
<430
440
Stage 3
1.4 Um
<120
880
3.6
5.4
170
78
300
570
1,000
165
3.0
6,100
<760
540
Stage 4
0.69 Um
<210
<230
9.1
4.5
83
69
340
300
1,200
62
3.3
5,400 <
'1,400
580
Stage 5
0.34 Um
<430
<450
8.6
240
77
680
560
1,300
260
140
11,000
<2,800
850
Outlet From Dust Collector
Trace
Species
Antimony*
Barium*
Beryllium*
Cadmiumt
Chromium*
CobaltT
CopperT
Leadt
Manganese"'"
Nickel*
Tin"
Titanium*
Vanadium*
Zinc*
Concentration , \ig/q
Aerodynamic Diameter, um
Cyclone
15 Um
<61
570
4.3
3.9
92
63
260
200
1,300
76
2.2
6,900
<410
340
Stage 1
3.44 Um
<59
1,100
3.3
4.6
58
66
250
360
780
78
3.3
6,600
<390
450
Stage 2
2.06 Um
-------
TABLE 9-4. PARTICULATE SIZE AND ENRICHMENT
Test No. 166, Location No. 35
Outlet From Dust Collector
Trace
Species
Antimony*
Barium*
Beryllium*
Cadmium'
Chromium*
Cobalt^
Copper
Lead1"
Manganese"1"
Nickel*
Tint
Titanium
Vanadium*
Zinc*
Concentration, yg/g
Aerodynamic Diameter, ym
Cyclone
15 ym .
<140
230
12
8.1
400
135
530
580
710
450
4.1
<3,700
<430
1,200
Stage 1
3.5 ym
<1,600
<1,800
<18
23
610
2,300
1,100
1,300
400
490
84
<42,000
<1,000
4,000
Stage 2
2.1 ym
<1,600
<1,700
<18
98
740
300
1,700
5,300
1,700
<390
64
<41,000
<11,000
9,400
Stage 3
1.4 ym
<1,800
<1,900
27
170
580
1,000
1,800
11,000
11,000
1,000
69
<46,000
<12,000
34,000
Stage 4
0.70 ym
<3,900
<4,100
79
95
1,100
7,100
3,300
3,600
840
1,800
150
<99,000-
<26,000
1,200
Stage 5
0.35 ym
<3,800
<4,000
51
94
<930
420
2,900
12,000
3,300
<930
150
<96,000
<25,000
20,000
*
4.
Determined by flame atomic absorption.
Determined by graphite furnace atomic absorption.
6001-48
139
-------
SECTION 10.0
REFERENCES
1. Bertine, K.K., Goldberg, E.D., Science, 173, 233, 1971.
2. Billings, C. E., Matson, W. R., Science, 176, 1233, 1972.
3. Joensuu, 0. I., Science, 172, 1027, 1971.
4. Magee, E. M., Hall, H. J. and Varga, G. M., "Potential
Pollutant in-Fossil . Fuels," Contract Report EPA R2-73-249,
PB 225 039, June 1973.
5. Federal Register, 38 (66), 8820, April 6, 1973.
6. Lave, L. B., Freeburg, L. C., Nucl. Safety, 14 (5), 409, 1973.
7. Cato, G.' A., et al., "Field Testing: Application of Combustion
Modifications to Control Pollutant Emissions from Industrial
Boilers - Phase I," EPA 650/2-74-078-a, NTIS No. PB 238 920/AS,
October 1974.
8. Cato, G. A., et al., "Field Testing: Application of Combustion
Modifications to Control Pollutant Emissions from Industrial
Boilers - Phase II," EPA 600/2-76-086a, April 1976.
9. Goldberg, A. J., "A Survey of Emissions and Controls for 'Hazardous'
and Other Pollutants," U.S. Environmental Protection Agency, Office
of Research and Monitoring, Air Pollution Technology Branch,
November 1972.
10. Lee, R. E., Jr. and D. J. von Lehmeden, "Trace Metal Pollution
in the Environment," J. Air Poll. Control Assoc., _23_ (10): 853,
1973.
11. National Academy of Sciences, "Particulate Polycyclic Organic
Matter," ISBN 0-309-02027-1, Washington, D.C., 1972.
12. Klein, D. H., et al., "Pathway of Thirty-seven Trace Elements
Through A Coal-Fired Power Plant," Environ. Sci. and Tech., 9_,
973, 1975.
13. Joensuu, 0. I., "Fossil Flues as a Source of Mercury Pollution,"
Science, 172: 1027, 1971.
140
-------
14. Cuffe, S. T. and Gerstle, R. W., "Emissions from Coal-Fired
Power Plants: A Comprehensive Summary," Public Health Service
Publication No. 999-AP-35, 1967.
15. Bolten, N. E., et al., draft copy of Section V, "Trace Element
Measurements at the Coal-Fired Allen Steam Plant," for NSF by
Oak Ridge National Laboratory.
16. Bolten, N. E., et al., "Trace Element Measurements at the Coal-
Fired Allen Steam Plant," Progress Report, June 1971-January 1973,
for NSF by Oak Ridge National Laboratory, Contract No. W-7405-eng-
26, Report No. ORNL-NSF-EP-43, March 1973.
17. Vandegrift, A. E. and L. J. Shannon, "Particulate Pollutant System
Study, Volume IIMass Emissions," prepared for EPA by Midwest
Research Institute, May 1971.
18. Schroeder, H. A., M.D., "A Sensible Look at Air Pollution by
Metals," Arch. Environ. Health, 21, 798-806, December 1970.
19. Natusch, D. F. S., Wallace, J. R., Evans, C. A., Science 183,
202, 1974.
20. Toca, F. M., Berry, C. M., Amer. Ind. Hyg. Ass. J., 34 (9),
396, 1973.
21. Gladney, E. S., Zoller, W. H., Jones, A. G., Gordon, G. E.,
Environ. Sci. Technol., 8 (6), 551, 1974.
22. Schlesinger, M. D., and H. Schultz, "Mercury in Some Coals of
the United States, U. S. Bur. Mines Report Invest. No. RI 7609,
Pittsburgh, PA, 1972.
23. Minerals Yearbook; 1970, Volume _I_, "Metals, Minerals and Fuels,"
U. S. Bureau of Mines, p. 423, 1972.
24. Billings, C. E., et al.r "Mercury Balance on A Large Pulverized
Coal-Fired Furnace," J. Air Poll. Control Assoc., 23_ (9): 773, 1973,
25. Burton, J. S., G. Erskine, E. Jamgochian, J. Morris, R. Reale,
and W. L. Wheaton, "Baseline Measurement Test Results for the
Cat-Ox Demonstration Program," Contract No. F19268-71-C-0002,
EPA, June 1972.
26. Smith, W. S., and C. W. Gruber, "Atmospheric Emissions from Coal
CombustionAn Inventory Guide," Environmental Health Series,'
U. S. Department of Health, Education and Welfare, Public Health
Service Publication No. 999-AP-24, April 1966.
141
-------
27. Petroleum Facts and Figures, American Petroleum Institute,
Washington, D.C., p. 67, 1971.
28. Hangebrauck, R. P., D. J. von Lehmden, and J. E. Meeker, "Sources
of Polynuclear Hydrocarbons in the Atmosphere", Public Health
Service Publication 999-AP-33, Cincinnati: U. S. Department of
Health, Education, and Welfare, 48 pp, 1967.
29. Cowherd, C. Jr., and J. Spigarelli, "Hazardous Emission Characteri-
zation of Utility Boilers," Interim Report on EPA Contract No. 68-
02-0228, Task NO. 39, Midwest Research Institute, Kansas City, MO,
June 1974.
30. Technical Manual for Process Measurements of Trace Inorganic
Materials, Report on EPA Contract No. 68-02-1393 by TRW Systems
Group, Redondo Beach, CA, February 1975.
31. Flegal, C. A., et al., "Procedures for Process Measurements,
Trace Inorganic Materials," TRW Systems Group Document No. 24444-
6017-RU-OO, Redondo Beach, CA, July 1975.
32. Barrett, R. E., et al., "Assessment of Industrial Boiler Toxic
and Hazardous Emissions Control Needs," Final Report on EPA
Contract 68-02-1328, Task 8, Battelle-Columbus Laboratories,
October 16, 1974.
33. Cantuti, V., et al., J. Chromatog., 17, 60, 1965.
34. Charkraborty, B. B., and R. Long, Environ. Sci. Technol., 1_, 829,
1967.
35. Bhatia, K., Anal. Chem., 43, 609, 1971.
36. Bernas, B., Anal. Chem., 40, 1682, 1968.
37. Lao, R. C., et al., Anal. Chem., 45, 908, 1973.
s
38. Smith, W. B., et al., "Particulate Sizing Techniques for Control
Device Evaluation," Southern Research Institute, Birmingham,
Alabama, Contract Report No. EPA-650/2-74-102-a, NTIS No.
PB 245 184/A5.
39. Dorsey, J. A. and J. 0. Burckle, "Particulate Emissions and
Profiles," Chem. Engr. Progress, 67, 92, 1941.
142
-------
40. "Evaluation of Sampling Trains," Final Report Prepared for the
U. S. Environmental Protection Agency under Contract No. 68-02-
1399, Task No. 6, by Midwest Research Institute, Kansas City, MO,
December 31, 1975.
41. Statnick, R. M., et al., "Sampling and Analysis of Mercury Vapor
in Industrial Streams Containing SO ," paper ACS National Meeting,
August 1973.
42. Marcus, M., "Evaluation of Sampling Trains," Special report prepared
for the 0. S. Environmental Protection Agency under Control No. 68-
02-1399 by Midwest Research Institute, MRI Project No. 3925-C,
Task 4,' February 20, 1975.
43. Kaakinen, J. W., et al., "Trace Element Behavior in Coal-Fired
Power Plant," Environ. Sci. and Tech. 9_, 862-869, 1975.
44. Lyon, W. S., et al., Environ Sci. and Tech., 9_, 973, 1975.
45. Burchard, J. K., "The Significance of Particulate Emissions,"
J. Air Poll. Control Assoc., 24: 1141, 1974.
46. Hidy, G. M., et al., "Summary of the California Aerosol Characteri-
zation Experiment," J. Air. Poll. Control Assoc., 25; 1106-1113,
1975.
143
-------
SECTION 11.0
CONVERSION FACTORS
SI Units to Metric or English Units
To Obtain
g/Hcal
106 Btu
MBII/ft2
Hflll/ft3
Btu
103 lb/hr* or MBH
Ib/HDtu
ft
in
ft2
ft'
Ib
Fahrenheit
paig
psia
iwg (39.2*F)
* lb/hr of equivalent
From
ng/J
GJ
GJ-hrrl-m"2
GJ-hr-^-m"3
gm cal
GJ/hr
' ng/J .
m
cm
m2
m3.
Kg
Celsius
Kelvin
Pa
Pa
Pa
saturated steam
Multiply By
0.004186
0.948
0.08806
0.02684
3.9605 x 10~3
0.948
0.00233
3.281
0.3937
10.764
35.314
2.205
tF - 9/5(to)+32
tF - 1.8K - 460
piala pa)
To Obtain ppm
at 3% 0 of
Natural Gas Fuel
CO
IIC
NO or NOx
SO or SOx
Oil Fuel
CO
HC '
NO or NOX
so or SOx
Coal Fuel
CO
HC
NO or NOx
SO or SOx
Refinery Gas Fuel (Location 33)
CO
IIC
NO or NOx
SO. or SOx
Refinery Gas Fuel (Location 39)
CO
HC
NO or NOx
SO. or SOx
Multiply Concentration
"in ng/J by
3.23.
5.65
' 1.96
1.41
2.93
5.13
1.78
1.28
2.69
4.69
1.64
1.18
3.27
5.71
1.99
1.43
3.25
5.68
1.98
1.42
-------
English nnd Metric Untta to SI Units
To Obtain
ng/J
ng/J
-1 -2
GJ-hr -m
_, . -1 -3
GJ'hr -m
GJ/hr
m
cm
m
m
Kg
Celsius
Kelvin
Pa
Pa
Pa
lb/hr of equivalent
From 'Multiply By
Ib/MBtu 430
g/Mcal 239
MBH/ft2 11.356
MBH/ft3 37.257
103 lb/hr* 1.055
or 106 Btu/hr
ft 0.3048
in 2.54
ft2 0.0929
ft3 0.02832
Ib 0.4536
Fahrenheit tc - 5/9 (tp-32)
tR - 5/9 (tp-32) + 273
paig Ppa (Ppslg * 14.7X6.895X103)
psia P - (P ) (6.895X103)
pa psia
iwg (39.2"F) P - (P. ) (249.1)
saturated steam
To Obtain
ng/J of
Natural Gas 'Fuel
CO
HC
NO or HOx (as equivalent
SO2 or SOx
Oil Fuel
CO
HC
NO or NOx (as equivalent
SO2 or SOx
Coal Fuel
CO
HC
NO or NOx (as equivalent
SO. or SOx
Refinery Gas Fuel (Location 33)
CO
I!C
NO or NOx (as equivalent
SO, or SOx
Refinery Gas Fuel (location 39)
CO
MC
NO or NOx (as equivalent
nui^ipiy v,uncaii(.Eai.ion
in ppm at 3% O2 by
0.310
0.177
0.510 .
0.709
0.341
0.195
0.561
0.780
0.372
0.213
0.611
O.B50
0.306
0.175
0.503
0.700
0.308
0.176
0.506
N02)
S02 or SOx
0.703
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-086b
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
FIELD TESTING: TRACE ELEMENT AND ORGANIC
EMISSIONS FROM INDUSTRIAL BOILERS
5. REPORT DATE
October 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
G.A. Cato
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO ADDRESS
KVB Engineering, Inc.
17332 Irvine Boulevard
Tustin, California 92680
10. PROGRAM ELEMENT NO.
1AB014; ROAP 21BCC-046
11. CONTRACT/GRANT NOV
68-02-1074
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOC
Task Final; 6/74-6/78
NO PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
^.SUPPLEMENTARY NOTES Project officers for this report are R.E.Hall/R.A. Venezia, Mail
Drops 65/62, 919/549-8411, Exts 2477/2547. EPA-650/2-74-078a was the first report
of this series.
is. ABSTRACT
report gives results of the sampling of four coal-fired industrial boilers
to determine the emissions of 19 trace and minor elements and polycyclic matter
(POM). The trace and minor element emissions were related to total quantities of eac^h
element present in the fuel by examining the degree of mass balance and element par-"
titioning based on fuel input and element output in furnace deposits, flyash, and flue
gas vapor. The tendency of volatile elements for enrichment of finer particulate was*
examined by analysis of cascade impactor samples. Measured output of elements
classified as high in volatility tended to be less than the fuel input, attributed to pos-
sible low collection efficiency of vapor-phase element sampling equipment. These
same elements were found to be more highly concentrated in the flyash as opposed to
furnace deposits and to have higher concentrations in the smaller particle sizes. Ele-
ments classed as medium or low in volatility tended to be more uniformly distributed
with respect to both partitioning in the boilers and particle size. Mass output results
for these elements frequently exceeded coal input, indicating possible sample conta-
mination by boiler or sampling system construction materials. The presence of four
specific POM compounds was indicated in the coal, ashes, and stack gases, but
results were highly variable.
17.
KEY WORDS ANO DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Croup
Air Pollution
Boilers
Coal
Chemical Analysis
Organic Compounds
Polycyclic Compounds
Tests
Combustion
Air Pollution Control
Stationary Sources
Trace Elements
Polycyclic Organic
Matter
13 B
13A
2 ID
07D
07C
14B
2 IB
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. .NO. OF PAGES
157
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
146
------- |