EPA-650/2-73-031
October 1973
ENVIRONMENTAL PROTECTION TECHNOLOGY SERIES
-------
EPA-650/2-73-031
EFFECTIVENESS
OF SELECTED FUEL ADDITIVES
IN CONTROLLING POLLUTION EMISSIONS
FROM RESIDUAL-OIL-FIRED BOILERS
by
D.W. Pershing, G.B. Martin,
E.E. Berkau, andR.E. Hall
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
ROAP No. 21ADG
Program Element No. 1A2014
Prepared for
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
October 1973.
-------
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.
ii
-------
CONTENTS
Page
Acknowledgements vl
Summary vii
Introduction 1
The Additives 3
Trimex 3
PACE 5
KAP 5
Glo-Klen 6
Sodium Carbonate 6
Test Facility 9
The Boiler 9
Injection System 9
Standard Fuel 10
Analytical Procedures 10
Test Plan 15
Trimex 15
PACE, KAP, Glo-Klen 17
Sodium Carbonate 17
Discussion of Results: SO 19
A
Baseline Characterization 19
Trimex 20
PACE 24
m
-------
CONTENTS (Cont.)
Page
KAP 27
Glo-Klen 27
Sodium Carbonate 32
Discussion of Results: Other Emissions 35
Unburned Hydrocarbons 35
Carbon Monoxide 35
Nitric Oxide 35
Metallic 36
Conclusions 39
Bibliography 41
iv
-------
LIST OF TABLES
Table No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Figure No.
1
2
3
Title
Chemical Analysis of Additives
Chemical Analysis of Standard Test Fuels
Trimex Test Plan
Test Plan for PACE, KAP, and Glo-Klen
Results of Trimex Testing
Results of Particulate Analysis—Trimex
Analysis of Boiler Deposits (Trimex)
Results of PACE Testing
Results of Particulate Analysis—PACE
Results of KAP Testing
Results of Particulate Analysis—KAP
Results of Glo-Klen Testing
Results of Particulate Analysis—Glo-Klen
Additional Uncontrolled Metallic Emissions
Resulting from the Use of Trimex in a
1000-MW Boiler
LIST OF FIGURES
Title
Schematic of Test System
Analytical System
Sulfur Oxides Sampling Apparatus
(Met Chemical)
Photographs of Test Boiler After Trimex
Testing
Page
4
11
16
18
21
23
25
26
28
29
30
31
33
37
Page
8
12
14
38
-------
ACKNOWLEDGEMENTS
The authors wish to gratefully acknowledge the help of Messrs.
Nelson L. Butts and Daniel S. Watkins in carrying out the experimental
aspects of this program and of John H. Wasser in conducting the boiler
characterization.
-------
SUMMARY
The purpose of this study was to experimentally evaluate the effectiveness
of four additive materials in controlling pollutant emissions from fossil fuel
combustion. The additives considered were Trimex, PACE, KAP, and Glo-Klen.
Each material was carefully examined in a highly instrumented package boiler
over the range of typical operating conditions (e.g., combustion intensity
and residence time) for industrial and utility systems.
The test results show that Trimex, PACE, KAP, and Glo-Klen do not reduce
emissions of SOX, NO, CO, or UHC under any condition tested. Based on these
test results, the boiler operating problems, and the possibility that use of
these materials might create potentially harmful new emissions, none of the
additives can be recommended as a means of controlling pollutant emissions.
VI1
-------
INTRODUCTION
For many years numerous companies have marketed a variety of fossil-fuel
additives purported to be able to improve or change the combustion of oil
and/or coal in some beneficial fashion. In an effort to determine the effect
of these compounds on pollutant emissions, the U. S. Environmental Protection
Agency* has undertaken a detailed evaluation. The work began several years
ago with a literature survey and a series of contacts with individuals
knowledgeable in the use of fuel additives for specific situations. A
distillate oil testing program followed: some 200 different materials were
evaluated for their effect on light oil emissions. The results of the work
showed that although a few proprietary metallic additives substantially
reduced soot emissions, in no case did any additive reduce the emissions of
carbon monoxide, unburned hydrocarbons, or nitrogen oxides from distillate
oil combustion.
To evaluate the effect of these materials on SO emissions and to
/\
establish the potential of other new materials, a continuing residual oil
and coal program was initiated. While the bulk of the detailed testing
will be conducted by an outside contractor, the Combustion Research Section
has conducted an initial in-house evaluation of the most widely publicized
materials: Trimex, PACE, KAP, and Glo-Klen.
*Combustion Research Section, Clean Fuels & Energy Branch, Control Systems
Laboratory, Office of Research and Development, National Environmental
Research Center/RTP
-------
This document is a report of the effectiveness of these new materials
in controlling pollution emissions from residual oil combustion. Mention
of company and product names herein does not constitute endorsement by the
U. S. Environmental Protection Agency.
Environmental Protection Agency policy is to express all measurements
in Agency documents in metric units. When implementing this practice will
result in undue costs or lack of clarity, conversion factors are provided
for the non-metric units used in a report. Generally, this report uses
British units of measure. For conversion to the metric system, use the
following conversions:
To convert from
°F
ft
ft2
ft3
gal.
Btu
Ibs
lbs/106 Btu
grains/ft
tons
To
UC
meters
2
meters
3
meters
1
cal
9
g/106 cal
9/m3
kg
Multiply by
5/9 (°F-32)
0.304
0.0929
0.0283
3.79
252
453.6
1.80
2.29
907
-------
THE ADDITIVES
Pollutant reduction additives can be divided into two categories:
tnose that catalyze reactions which convert the pollutant to a non-toxic
state (e.g., catalytic reduction of NO to N2 and 02) and those that
chemically react with the pollutant to form specie which can be easily
collected (e.g., reaction of CaC03 with SOX to form calcium sulfate).
Catalytic additives are most desirable because of the small amount of
material required. The four proprietary additives tested in this work
are of the catalytic type according to their manufacturers. The last
additive tested, sodium carbonate, is of the latter type because it
reacts chemically with SO to form sodium sulfates.
/\
Table 1 gives the detailed chemical analyses of each proprietary
material tested. These data were determined for EPA by an independent
laboratory* using atomic adsorption, neutron activation, etc. as indicated.
In each case x-ray diffraction was also used to determine the chemical
compound(s) predominant in each additive. The final additive, sodium
carbonate, was a single compound (commercial grade) and was therefore not
analyzed. All of the materials are dry powders that are blown into the
primary combustion zone where they are reported to cause reductions in
pollutant emissions.
TRIMEX
Trimex is a fine, dry powder fuel additive manufactured by Trimex
Corporation, 495 Mandal ay Avenue, Clearwater, Florida. It is available
*Shell Development Company, Emeryville, California
3
-------
Table 1. CHEMICAL ANALYSIS OF ADDITIVES
Element
Al
B
Ca
C
Cl
Cr
» Cu
Fe
Mg
Mn
Na
Ni
0
Si
Ti
Zn
Total
Analysis method
Atomic absorption
Flame emission
Atomic absorption
Combustion
Ag-titration
Atomic absorption
Atomic absorption
Atomic absorption
Atomic absorption
Atomic absorption
Flame emission
Atomic absorption
Neutron activation
Neutron activation
Atomic absorption
Atomic absorption
Weight percent
Trimex I
7.9
0.5
1.0
—
—
—
—
2.8
1.9
—
1.9
—
55.5
27.6
0.2
—
99.3
Trimex
6.0
0.3
1.1
—
—
—
—
1.7
1.7
—
3.6
—
53.1
23.1
0.2
—
90.8
II PACE
—
—
21.1
0.5
21.2
0.4
0.5
7.8
4.3
6.2
11.5
0.2
24.3
0.1
—
0.7
98.8
KAP
0.3
___
4.4
6.1
30.1
—
—
3.6
2.5
2.2
19.6
—
23.3
5.6
—
—
97.7
Glo-Klen
6.5
_-_
0.6
—
—
—
1.4
0.8
—
4.6
—
60.0
19.3
—
—
93.2
-------
in two formulations, botn of which are mixtures of clays. The x-ray
diffraction patterns indicated that Trimex I is mostly montmorillonite,
while Trimex II is a mixture of 2 Na20 . A1203 . 2Si02 and gismondite,
CafA^SigOg) . 4 H^O. The material is recommended for both coal and oil
(1-3)
and is reported by the manufacturer to substantially reduce CO, UHC,
and S02 emissions. The manufacturer's recommended feed rate is about 1 Ib
of Trimex per 25 Ibs of sulfur in the fuel. (This is equivalent to about
1 Ib of additive for 300 gallons of a 1-percent residual oil, or to about
2.5 Ibs of additive per ton of 3-percent sulfur coal.) Trimex sells for
about $360 per ton F.O.B. in 50-ton lots.
PACE
PACE (Pure Air and Clean Environment) is a fine, red-powder fuel
additive, manufactured by Takayuki Oishi, Japan. X-ray diffraction
indicates that it is probably a mixture of NaCl, MgO, CaO, and alpha
Fe^. The manufacturer reports it is effective in reducing emissions of
smoke and S02. The manufacturer's recommended minimum feed rate is 1 Ib
of PACE to 1000 Ibs of fuel. (This is equivalent to about 1 Ib of additive
for 125 gallons of residual oil or to about 2 Ibs per ton of coal.) No
firm cost data are available because PACE is not widely marketed in the
United States.
KAP
KAP is a dry-powder fuel additive marketed by Kleen-Aire Products, Inc.,
3930 Newhall Road, Columbus, Ohio. Based on the results of x-ray diffraction
and KBr pressed plate infrared spectral analysis, KAP appears to be 50
percent salt (NaCl), 13 percent talc (Mg3Si4012H2), 10 percent CaC03, and
-------
an iron manganese silicate. KAP is recommended for use in coal, oil, and
smelting furnaces and is reported^ ' by the manufacturer to eliminate or
reduce the emission of black smoke, SO , CO, NO. and chlorine gas while
** n
increasing the combustion temperature, cleaning the combustor walls, and
increasing the burning efficiency by at least 10 percent. The recommended
feed rate is about 1 Ib of additive per 1000 Ibs of fuel. (This is equivalent
to about 1 Ib of additive for 125 gallons of residual oil or to about 2
Ibs per ton of coal.) KAP sells for about $1200 per ton.
GLO-KLEN
Glo-Klen is also a dry-powder fuel additive manufactured by Glo-Klen,
Inc., 3705 Morse Avenue, Lincolnwood, Illinois. X-ray diffraction indicates
that it is a mixture of clays, probably montmorillonite and eriomite. The
manufacturer reports that Glo-Klen is effective with anything burnable
including gas, oil, wood, and coal, and that it reduces or eliminates emissions
of smoke, SOX, N0x, CO, UHC, etc. In contrast to the other materials, it
is only necessary to spread or blow Glo-Klen into the firebox once or twice
a day according to the manufacturer. The recommended dosage is 2 Ibs per
1000 sq ft of heating surface. (This is equivalent to a continuous feed
rate of about 1 Ib of additive for 300 gallons of residual oil or about 0.75
Ibs per ton of coal.) Glo-Klen retails for about $1100 per ton.
SODIUM CARBONATE
At the conclusion of the testing with the proprietary materials, a
single very brief test was conducted with sodium carbonate since it has
been repeatedly shown to be effective^ in reducing SO emissions. The
A
test sample of sodium carbonate was commercial grade and was obtained
from a local supplier at a cost of $4.95 per 100 Ib delivered. No attempt
-------
was made to determine the absolute purity of the material. Previously,
(5\
workersv ' have utilized feed rates between 0.80 and 4.8 Ibs Na2CO, per
Ib of sulfur in the fuel. (The latter is equivalent to about 1 Ib of
additive per 2.5 gallons of 1-percent sulfur residual oil or about 290
Ibs of additive per ton of 3-percent sulfur coal.) Sodium carbonate was
included in this program to provide a "control" on the test system; that
is, to show that SO reductions are possible with the use of an additive.
A
-------
LAMINAR
FLOW ELEMENT
AIR BLOWER
oo
PRE-CALIBRATED
VIBRA-SCREW FEEDER
COMBUSTION
AIR IN
STEAM OUT
I
n
FIREBOX-FIRETUBE
BOILER (3-PASS)
OIL FLOW METER
ELECTRO-COIL
THERMO-CONTROLLER
OIL PUMP
Figure 1. Schematic of test system.
FLUE GAS OUT
J
TO ANALYTICAL
SYSTEM
1
T
OBSERVATION PORT
RESIDUAL OIL
SUPPLY TANK
-------
TEST FACILITY
The test system used in this program is shown in Figure 1. Basically
it consisted of a highly instrumented commercial boiler which had been
modified for addition of solid fuel additives in the combustion zone and
a complete set of analytical instrumentation for emissions characterization.
Each component is described in detail below.
THE BOILER
The experimental portion of this evaluation was conducted on a 54-HP
(1.8 million Btu/hr input), firebox-firetube package boiler firing residual
oil. With the exception of the additive injection system, the unit was a
typical commercial boiler with a 20-gallon per hour air atomizing gun burner
capable of modulation. The inlet oil temperature was maintained constant
at 220°F by a thermal controller. The oil flow was metered continuously,
and the flow rate was checked by calculations using excess air and stack
velocity.
INJECTION SYSTEM
The injection system used throughout the testing was designed to
conform to the recommendations of the respective additive manufacturers
and to simulate those currently in use in industrial systems. The dry-
powder additives were continuously metered into the primary combustion
zone by means of a pre-calibrated, variable-speed, vibrating screw feeder
above the burner. Periodic timing of the displacement rate for known
quantities of additive was also used to provide a check on the feeder
calibration. The powder was conveyed from the feeder to the injector by
-------
means of fluidizing air (less than 1 percent of the total combustion air).
Significant plugging problems were not encountered due to the close proximity
of the feeder and the injector. The injector was a stainless steel tube and
in nearly all cases was positioned inside the wind box of the burner so that
suspended powder was injected into the highly turbulent region of the burner
throat just before the oil nozzle. This was done to ensure maximum contact
between the additive and the combusting fuel. Observation ports were utilized
to verify that settling did not occur. During two of the Trimex tests the
injector was moved forward approximately 12 inches so that the additive powder
was blown in directly above the flame zone.
STANDARD FUEL
To provide a uniform test fuel, a large quantity of typical No. 6 oil
was obtained before the test series began. However, unfortunately the Trimex
testing extended so long that it was necessary to obtain a new oil supply
prior to the other testing. While both of these oils were obtained from the
same supplier there was a slight variation in composition as shown by Table 2.
These residual oils were used throughout the testing with the exception of
one special Trimex test in which a small quantity of high (1.74 percent) sulfur
oil was used to evaluate the effect of sulfur content on performance of this
additive.
ANALYTICAL PROCEDURES
Sampling and analytical systems (Figure 2) were similar to those used in
earlier studies by Martin et al:(6) paramagnetic oxygen analysis, flame
ionization detection for UHC, and non-dispersive infrared for CO, C02, and NO.
S02 emissions were monitored continuously with a non-dispersive infrared
analyzer. Total SOX was also determined using the SOV version of the standard
X
10
-------
Table 2. CHEMICAL ANALYSIS OF STANDARD TEST FUELS
Specie
C
H
N
S
0
Ash
Weight percent
Trimex testing
88.4
10.2
0.27
0.9
0.28
0.04
PACE, KAP, Glo-Klen, and
Na2C03 testing
87.9
10.18
0.25
0.88
0.38
0.04
11
-------
TOWETSOX -^-
WATER TRAP
ALL STAINLESS STEEL LINES
ALL TFE
TEFLON
LINES
NDIR
S02
ANALYZER
NDIR
NO
ANALYZER
HEATED MOLECULAR
SIEVE-3 A CLAY BASE
WATER
TRAP
FIBREGLAS
INDUSTRIAL
FILTER
SILICA
GEL TRAP
PARA-
MAGNETIC
°2
ANALYZER
r
STACK
GLASS
WOOL
FILTER
FLAME
10NIZATION
H/C
ANALYZER
Figure 2. Analytical system.
12
FLUE GAS
VELOCITY
PROBE
\
HEATED
PARTICULATE
FILTER
HEATED
STAINLESS
PROBE
D
FLOW CONTROL
VACUUM PUMP
SILICA
GEL
TRAP
NDIR
CO
ANALYZER
NDIR
C02
ANALYZER
-------
EPA wet chemical S0? sampling technique.^ ' In the SO version (Figure 3)
£ A
the contents of the isopropanol bubbler and the probe washings were titrated
and the resulting concentration in ppm was added to the total. The entire
technique was verified by analyzing certified span gases. Since the purpose
of this investigation was to determine the effect of the additive on total
SOX emissions, no attempt was made to separate the SCL/SCL components.
Filterable particulate matter was collected isokinetically on woven
silver filters capable of > 0.8 micron collection. Each sample was then
analyzed for chemical composition by an independent laboratory.
13
-------
GREENSBURG-SMITH
BUBBLERS, 4 REQUIRED
60-mm MED POROSITY
FRITTED DISK
CONDENSER, GRAHAM
COIL TYPE, 300-mm,
CORNING 2500
ISOPROPANOL SOLUTION
NEEDLE
VALVE
SILICA GEL
DRYING TUBE
VACUUM PUMP
ROTAMETER
(0.10 CFM)
DRY GAS METER
Figure 3. Sulfur oxides sampling apparatus (wet chemical).
FLUE GAS OUT
GLASS
WOOL
FILTER
STACK
I
-------
TEST PLAN
The experimental phase of the evaluation was directed at characterizing
the effectiveness of the additives over a wide range of conditions represen-
tative of actual boiler operations. While the additive materials were being
analyzed, the test facility was constructed, instrumented, and operated to
establish baseline performance over the range of conditions to be investigated.
The additive testing was then begun. Since each additive was tested as
nearly as possible according to the manufacturer's suggestions, test plans
were not identical.
TRIMEX
Trimex was injected continuously for 11 days before any detailed testing
was begun because the manufacturer stresses that the material must be allowed
to acclimate the boiler before it is effective. After the acclimation period
continuous testing was begun and the following variables were examined:
Air/fuel ratio (excess air) (20-40 percent)
Firing rate (load) (8-20 gal./hr)
Additive feed rate (1 Ib additive/200 Ibs fuel to 1 Ib
additive/25 Ibs fuel)
Additive injector location
Fuel composition (sulfur level) (0.9 - 1.8 percent)
Additive formulation (Trimex I and II)
Due to the limited time available, it was not possible to conduct a
complete factorial experiment evaluating all 64 variable combinations.
Table 3 shows the actual test plan used. Blank runs (no additive) were
15
-------
Table 3. TRIMEX TEST PLAN
Condition No.
Operating variables
Additive/blank
High fire
Normal excess air
Blank
High fire
Normal excess air
Additive - Nominal injection
rate
Low fire
Normal excess air
Additive - Nominal injection
rate
High fire
Low excess air
Additive - Nominal injection
rate
High fire
Normal excess air
Additive - 10 times nominal
injection rate
High fire
Normal excess air
Additive3 - 10 times nominal
injection rate
High fire
Normal excess air
Blank
Additive injected above flame zone.
16
-------
not conducted before each test condition, because the manufacturer claims
that Trimex must be used continuously to be effective. Four complete blank
tests were conducted, however, before and after the Trimex testing.
PACE, KAP, 6LO-KLEN
During the testing of PACE, KAP, and Glo-Klen, the following variables
were considered:
Air/fuel ratio (excess air) (20-40 percent)
Firing rate (load) (8-20 gal./hr)
Additive injection rate (dose)
Again it was not possible to do a complete factorial experiment.
Fuel composition (sulfur level) was not considered because the 1.8-percent
sulfur oil was no longer available (due to a local ordinance). Injector
location was not included as a variable because injection above the flame
zone led to boiler operating problems during the Trimex testing. Table 4
shows the detailed plan used with PACE, KAP, and Glo-Klen. As the table
indicates, a blank test (no additive) was conducted before each additive
test. In no case were data taken until after the test unit had been
operating at equilibrium condition (additive or blank) for at least 1 hour.
SODIUM CARBONATE
During the sodium carbonate testing it was only possible to briefly
examine the high fire, normal excess air condition before the test facility
was shut down for relocation of the laboratory.
17
-------
Table 4. TEST PLAN FOR PACE, KAP, AND GLO-KLEN
Condition No.
Operating variables
Additive/blank
Low fire
Normal excess air
Low fire
Normal excess air
Blank
Additive - Nominal injection
rate
High fire
Normal excess air
High fire
Normal excess air
Blank
Additive - Nominal injection
rate
High fire
Low excess air
High fire
Low excess air
Blank
Additive - Nominal injection
rate
High fire
Normal excess air
Blank
High fire
Normal excess air
Additive - 3 times nominal
injection rate
18
-------
DISCUSSION OF RESULTS: SO
A
In order to facilitate comparison, all data throughout the paper are
reported as ppm by volume, dry, and reduced to stoichiometric (zero percent
excess air). Due to the duration of the Trimex testing, it was necessary
to procure a second supply of standard fuel before starting the next testing.
Although both supplies were from the same producer, they were not identical
in composition and, therefore, gave slightly different baseline SO . Finally,
A
each material was tested primarily at a "nominal" feed rate of about 10
grams per minute (depending upon additive density). This rate exceeds most
of the manufacturer's recommended minimum dosages; however, it was selected
to ensure that claimed S02 reductions would be observed. Once the materials
were shown to be effective, the dosage would be reduced to determine the most
cost effective feed rate.
BASELINE CHARACTERIZATION
Characterization work had established the "normal" commercial operating
condition for this type of unit as full load (a firing rate of 20 gallons
per hour) and 40 percent excess air. The reduced-load condition was
defined to be 8 gallons per hour (40 percent excess air) and was included
in the testing to investigate the effect of residence time on additive
performance. (At this low-fire condition, the gas/additive contact time is
increased by a factor of about 2.5.) The low excess air condition was
defined to be 20 percent excess air (just above the flame-out limit of the
unit). This condition was included to determine if limiting the availability
of oxygen would enhance the effectiveness of the material.
19
-------
Testing before and after the Trimex injection confirmed the Trimex
baseline (no additive) to be 625 ppm SOX; this agrees with the theoretical
SOX concentration calculated from the analysis of the fuel. The 625 ppm
baseline level is the average of many tests and has a standard deviation
of 7 percent. Therefore, any change of more than 88 ppm (2o) from the
baseline level can be considered statistically significant.
During the PACE, KAP, and Glo-Klen tests, baseline runs were conducted
before and after each additive test at a particular condition. A total
of eight baseline determinations were conducted during the complete
testing of each of these additives; the average results were:
613 t 25 ppm (PACE)
614 ! 29 ppm (KAP)
614 + 16 ppm (Glo-Klen)
These values agree with the 615 ppm SOV predicted from the analysis for
A
this fuel.
TRIMEX
The bulk of the Trimex trialswere conducted using Trimex II as requested
by the manufacturer. The initial testing (after the acclimation period) was
conducted at the high fire boiler condition and the nominal additive injection
rate of 7 grams per minute (1 Ib additive/200 Ib fuel). As Table 5 indicates,
Trimex had no effect under these conditions. (All results are from at least
4 hours of testing at the specified condition and are the average of at least
two sets of complete emissions measurements.)
In the next test, the firing rate was reduced to 8 gallons per hour,
thus increasing the additive/combustion gas contact time by a factor of
2.5. Again, no significant change over the baseline (no additive case) was
observed.
20
-------
Table 5. RESULTS OF TRIMEX TESTING0
Test Condition
Alld
High fire (full load)
Normal excess air
Low fire (reduced load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Normal excess air
High fire (full load)
Normal excess air
Additive/
blank
Blank
Trimex
7 g/min
Trimex
7 g/min
Trimex
7 g/min
Trimex
60 g/min
Trimex6
60 g/min
SQy emission
(ppm dry, reduced to 0% excess air)
652 t 45
667
642
665
657
656
Effect0
None
None
None
None
None
None
All test results were obtained using Trimex II and the standard fuel after the acclimation period.
These results are the average of at least two separate emission determinations.
Variations of less than 45 ppm from the baseline are not statistically significant.
Average of all blank runs.
6Additive injected above flame zone.
-------
In the third series of tests, the boiler was operated at high fire
but at a low excess air level (20 percent) just above the flame-out
limit of the unit. The purpose of these tests was to determine if limiting
the availability of oxygen would enhance the effectiveness of the material.
Again, the results were negative.
During the fourth series of tests, the additive injection rate
was increased to 60 grams per minute (maximum output of feeder) to
investigate possible effects of increased feed rate. None were observed.
In the final tests, the additive injector was positioned so that it
sprayed the Trimex powder above the flame zone. No reduction was observed.
The average S0x emission during all the Trimex testing (after the
acclimation period) was 657 ppm with a standard deviation of only 18 ppm.
The average baseline level was 652 ppm. Thus, there is little doubt that
Trimex was wholly ineffective throughout the actual testing.
During the 11-day acclimation period, tests were run under several
other conditions. While the detailed results are not reported herein, in
deference to the manufacturer's claim regarding an acclimation period, the
trends are worth noting. Before the supply of Trimex II arrived for testing,
Trimex I was evaluated at both the low and high fire normal excess air
conditions. No reductions were observed. The unit was also switched
to a high sulfur residual oil (1.74 percent); again, no reduction was
observed.
Throughout the actual Trimex testing, particulate samples were
chemically analyzed. Table 6 presents these results. In no case did the
22
-------
Table 6. RESULTS OF PARTICIPATE ANALYSIS - TRIMEX
Test
condition
Low fire (reduced load)
High fire (full load)
High fire (full load)
Normal excess air
Low fire (reduced load)
Normal excess air
High fire (full load)
Low excess air
^igh fire (full load)
•"Normal excess air
High fire (full load)
Normal excess air
Additive/ Particulate loading3 Percent sulfur6 Particulate sulfur0 Fuel sulfurd
blank (grains/ft3) in Particulate (lb/106 Btu) (lb/106 Btu)
Blank
Blank
Trimex
7 g/min
Trimex
7 g/min
Trimex
7 g/min
Trimex
60 g/min
T • h
Trimex
60 g/min
0.056
0.074
0.081
0.113
g
1.25
1.32
0.4 0.0003 0.51
1.5 0.0014 0.51
f -- n 51
1 U. 3 1
2.4 0.0035 0.51
-- n 51
u. o i
2.0 0.0325 0.51
2.4 0.0415 0.51
Percent of entering
S leaving boiler as
solid6
0.06
0.3
0.7
6.4
8.1
Dry, at 32 F, and reduced to zero percent excess air.
As determined for EPA by an independent laboratory.
CInllys?sUSf th™art1cSS.SUlfUr C°nta1ned 1n the Peculate based on the firing rate, particulate loading, and chemical
Based on the chemical analysis of the oil.
eThis column shows the percent of the fuel sulfur which left the boiler as part of the particulate.
Insufficient sample.
9Not determined.
Additive injected above flame zone.
-------
sulfur content of the flue gas particulate exceed 0.04 Ibs per million
Btu which is less than 9 percent of the total sulfur entering the boiler
(0.51 lbs/10 Btu). Thus, these data give no indication that Trimex
was forming any significant amount of solid phase complex.
At the conclusion of the Trimex testing the boiler was shut down
and samples of various deposits taken. Table 7 shows the analysis of
these samples. The deposits in the front-tube cross section contained
1.8 percent sulfur—the highest concentration found anywhere in the
boiler. However, even if the average deposit level was assumed to be
1.8 percent sulfur, this would represent less than 0.3 percent of the
total sulfur input during the measured period based on a total sulfur
balance on the system. Hence, no significant amount of sulfur was
being converted to solid deposits.
Thus, there can be little doubt that Trimex was wholly ineffective
in reducing SOX emissions: no reduction in flue gas SO concentration was
A
observed, no significant sulfur was found in the particulate, and no
sulfur was found deposited inside the boiler.
PACE
The results of the PACE testing are shown in Table 8. Under no
conditions did PACE significantly reduce SO emissions. The slight
A
reductions in the low fire and high fire/increased rate tests, while
not statistically significant, may be real since both occurred at
increased additive-to-fuel ratio (above 1 Ib PACE/80 Ib fuel). Since
PACE contains substantial amounts of NaCl, MgO, and CaO, these slight
reductions would be expected due to stoichiometric chemical effects such
as the formation of CaSO..
24
-------
Table 7. ANALYSIS OF BOILER DEPOSITS (TRIMEX)
Location
Percent sulfur
Burner quarrel
Firebox floor
Firebox lower wall
Firebox upper wall
Second tube pass
Front-tube cross section
Back-tube cross section
1.0
0.1
0.3
1.4
1.2
1.8
1.4
25
-------
Table 8. RESULTS OF PACE TESTING
ro
(7t
Test condition
Theoretical emissions
Based on fuel analysis0
Baseline measurementsd
Low fire (reduced load)
Normal excess air
High fire (full load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Normal excess air
Additive/
blank
Blank
Blank
PACE
6 g/min
PACE
6 g/min
PACE
6 g/min
PACE
15 g/min
SO emission3 """"
(ppm dry, reduced to 0% excess air) Effect5
615
613 t 25
567 None
609 None
597 None
565 None
These results are the average of at least two separate emission determinations.
'Variations of less than 50 ppm are not statistically significant.
0.88 percent sulfur. Upon oxidation, this level
SKJ?"
dete™1"a"°"s ""<""ted throughout the
-------
Table 9 shows the results of the participate sampling and analysis
conducted during the PACE testing. Unfortunately, it was not possible
to take a particulate sample at every condition due to time limitations.
The data confirm that no significant amount of the entering sulfur leaves
the combustor in the solid phase.
KAP
The results of the KAP testing are shown In Table 10. As the table
indicates, under no condition tested did KAP have a statistically
significant effect on SO emissions. A slight possible reduction was
/\
observed at the high feed rate condition: this is probably due to a
stoichiometric reaction between the active metal components and SO to
A
form sulfates. Regardless of the mechanism, however, the dose rate is
economically unfeasible. The possible 7-percent reduction was achieved
at 1 Ib additive/40 Ibs oil which is equivalent to 11 cents per gallon
of oil treated.
Table 11 shows the results of the particulate analysis at four of
the test conditions. Again, the flue gas particulate contains no signi-
ficant amount of sulfur.
GLO-KLEN
The detailed results of the Glo-Klen testing are shown in Table 12.
In summary, the average SOX emission with Glo-Klen was 611 ppm. The
average SOV emission without Glo-Klen was 614 ± 16 ppm; the theoretical SO
« x
was 615 ppm. Glo-Klen had absolutely no effect on SO emissions under
A
any condition tested. (Its composition is very similar to that of Trimex
and it would, therefore, not be expected to affect SO .)
/\
27
-------
Table 9. RESULTS OF PARTICULATE ANALYSIS - PACE
Test
condition
High fire (full load)
Normal excess air
High fire (full load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Low excess air
Additive/ Particulate loading3 Percent sulfurb Particulate sulfur0 Fuel sulfurd Percent of enterina
blank (grains/ft3) in particulate (lb/106 Btu) (lb/105 Btu) S leaving as solid1
Blank 0.079 5.4 0.0056 0.50
PACE 0.105 12.3 0.0169 0.50
6 g/min
Blank 0.114 4.1 0.0061 0.50
PACE 0.100 8.8 0.0115 0.50
6 g/min
1.1
3.4
1.2
2.3
IN)
Q
00 Dry, at 32 F, and reduced to zero percent excess air.
As determined for EPA by an independent laboratory.
This column shows the total sulfur contained in the particulate based on the firing rate, particulate loadinq and
chemical analysis of the particulate.
Based on the chemical analysis of the oil.
This column shows the percent of the fuel sulfur which left the boiler as part of the particulate.
-------
Table 10. RESULTS OF KAP TESTING
ro
vo
Test condition
Theoretical emissions0
Baseline measurements
Low fire (reduced load)
Normal excess air
High fire (full load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Normal excess air
Additive/
blank
Blank
Blank
KAP
13 g/min
KAP
13 g/min
KAP
13 g/min
KAP
34 g/min
SO emissions3 .
(ppm dry, reduced to 0% excess air) Effect
615
614 t 29
597 None
597 None
597 None
572 None
aThese results are the average of at least two separate emission determinations.
Variations of less than 50 ppm are not statistically significant.
cThe residual oil used throughout this testing contained 0.88 percent sulfur. Upon oxidation this level
would result in 615 ppm of SO .
/\
This result is the average of the eight separate emission determinations conducted throughout the testing
(before and after each additive test condition).
-------
Table 11. RESULTS OF PARTICULATE ANALYSES - KAP
Test
condition
High fire (full load)
Normal excess air
High fire (full load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Low excess air
Additive/ Particulate loading3
blank (grains/ft3)
Blank 0.062
KAP 0.124
13 g/min
Blank 0.094
KAP 0.140
13 g/min
k H Percent of entering
Percent sulfur Particulate sulfur Fuel sulfur S leaving boiler as
in particulate (lb/10° Btu) (lb/106 Btu) solid6
f — 0.50
6.2 0.0100 0.50 2.0
5.0 0.0061 0.50 1.2
10.2 0.0186 0.50 3.7
CO
o
Dry, at 32°F, and reduced to zero percent excess air.
As determined for EPA by an independent laboratory.
This column shows the total sulfur contained in the particulate based on the firing rate, particulate loading, and chemical
analysis of the particulate.
Based on the chemical analysis of the oil.
p
This column shows the percent of the fuel sulfur which left the boiler as part of the particulate.
Insufficient sample.
-------
Table 12. RESULTS OF GLO-KLEN TESTING
Test
condition
Additive/
blank
(ppm
dry,
SO emissions3
reduced to Q% excess
air)
Effectb
Theoretical emissions
Blank
615
Baseline measurements
Low fire (reduced load)
Normal excess air
Blank
Glo-Klen
7 g/min
614 I 16
610
None
High fire (full load)
Normal excess air
Glo-Klen
7 g/min
620
None
High fire (full load)
Low excess air
Glo-Klen
7 g/min
607
None
High fire (full load)
Normal excess air
Glo-Klen
18 g/min
606
None
These results are the average of at least two separate emission determinations.
Variations of less than 32 ppm are not statistically significant.
"The residual oil used throughout this testing contained 0.88 percent sulfur. Upon oxidation this level
would result in 615 ppm of SO dry at stoichiometric.
/\
This result is the average of the eight separate emission determinations conducted throughout the testing
(before and after each additive test).
-------
Table 13 shows the results of the particulate analyses conducted
during the Glo-Klen testing. In no case does the total particulate
sulfur exceed 0.006 lbs/10 Btu which is less than 2 percent of the
total fuel sulfur.
SODIUM CARBONATE
At the end of the proprietary additive testing, one brief test was
conducted with sodium carbonate (Na2C03) to demonstrate that dry materials
do exist, capable of SO control. The feed rate was 1 Ib of Na0CO. per
A 23
30 Ibs of fuel. This corresponds with 1.3 times the sodium carbonate
required to react with all the SO to form Na^SO..
A t ^r
The Na2C03, injected dry at the high fire, normal excess air
condition, reduced the SO emissions from 614 ppm to 393 ppm. This
A
generally agrees with previously reported Na2C03 work^ and represents
about a 36-percent SOX reduction. This result is also strengthened by
the particulate analysis which revealed a sulfur content in excess of
20 weight percent. Unfortunately, it was not possible to total material
balance as in Tables 6, 9, 11, and 13 because the extremely high solid
loading prevented isokinetic sampling for more than 1 minute.
No attempt to optimize the utilization of sodium carbonate was made
because of the boiler operating problems (e.g., tube fouling and ESP
overloading) associated with the long-term feeding of large amounts of
the material.
32
-------
Table 13. RESULTS OF PARTICULATE ANALYSIS - GLO-KLEN
Test
condition
High fire (full load)
Normal excess air
High fire (full load)
Normal excess air
High fire (full load)
Low excess air
High fire (full load)
Low excess air
Additive/ Parti cul ate loading9
blank (grains/ft3)
Blank 0.048
Glo-Klen 0.076
7 g/min
Blank 0.091
Glo-Klen 0.136
7 g/min
h r H Percent of entering
Percent sulfur Parti cul ate sulfur Fuel sulfur S leaving boiler as
in parti cul ate (lb/10b Btu) (lb/106 Btu) solid6
5.0 0.0032 0.50 0.6
4.6 0.0046 0.50 0.9
4.9 0.0058 0.50 1.1
3.0 0.0053 0.50 1.1
Dry, at 32 F, and reduced to zero percent excess air.
As determined for EPA by an independent laboratory.
cThis column shows the total sulfur contained in the particul ate based on the firing rate, particulate loading, and
chemical analysis of the particulate.
Based on the chemical analysis of the oil.
This column shows the percent of the fuel sulfur which left the boiler as part of the particulate.
-------
DISCUSSION OF RESULTS: OTHER EMISSIONS
UNBURNED HYDROCARBONS
Unhurried hydrocarbon emissions from commercial boilers are typically
very low if the unit and burner are maintained properly. The average
baseline hydrocarbon emission for the test boiler was 0.3 ppm calibrated
as propane. No additive had any effect on this level.
CARBON MONOXIDE
The average baseline carbon monoxide emission was 29 ppm; no
additive had any effect on this level. (It should be cautioned, however,
that injecting a dry additive in such a way that it changes the combustion
aerodynamics can substantially increase carbon monoxide emissions. This
was experimentally observed during initial location of the injection system.)
NITRIC OXIDE
No additive had any effect on the measured NO emissions from the test
unit; this agrees with the general results found during previous additive
testing/ ' Emissions of NO from residual oil combustion are attributable
to two sources: thermal fixation of atmospheric nitrogen at high temperatures,
and oxidation of nitrogen in either the oil or the additive. However, none
of the additives were effective in reducing NO formation via either mechanism.
NO emissions did not increase since none of the materials contained significant
nitrogen.
35
-------
METALLIC
Use of any metal-containing additive is also undesirable from an air
pollution viewpoint because of the new emissions its use would create. For
example, if one uncontrolled 1000-MW utility boiler were to use Trimex at
the manufacturer's recommended dosage for a year, it would emit almost 300
tons of new metallic pollutants (assuming none of the additive collected in
the boiler). Table 14 gives a more detailed breakdown of these emissions.
(Since it is not clear in what form these metals would be emitted, their
toxicity cannot be established.) Each of the other materials would result
in similar emissions.
If the metallic emissions are avoided by collecting the solid emissions
prior to the stack with a baghouse or other particulate removal system, then
a potential water-pollution/sol id-waste problem results. For example,
disposal of the solids from PACE or KAP utilization would be difficult
because both additives contain large quantities of sodium. This would
almost certainly result in sodium compounds in the solid waste that were
water soluble and would contaminate the run-off from any disposal site.
In practice part of any dry, metallic additive always collects inside
the boiler and can create serious boiler operating problems. For example,
at the end of the Trimex testing the test boiler had to be taken off line
for complete cleaning due to deposits in the boiler tubes and firebox as shown
in Figure 4. Admittedly, the high-additive doses used during the testing
greatly accelerated the formation of deposits; however, there is little
doubt that extended additive injection at lower rates would result in similar
deposits.
36
-------
Table 14. ADDITIONAL UNCONTROLLED METALLIC EMISSIONS RESULTING
FROM THE USE OF TRIMEX IN A 1000-MW BOILER
Element Additional Ibs/yr
Al 231,000
B 12,000
Ca 42,000
Fe 66,000
Mg 66,000
Na 139,000
37
-------
DEPOSITS IN UPPER BOILER TUBES AFTER TRIMEX TESTING.
CLOSE-UP OF BOILER TUBES -- RIGHT-HALF HAS BEEN CLEANED.
Figure 4. Photographs of test boiler afterTrimex testing.
-------
CONCLUSIONS
This testing, as in any experimental program, has certain limitations;
it is pertinent to consider these before any conclusions regarding the
data are reached. First, the program was conducted on a 20-gallon per
hour (0.3 MW) package boiler, not a 1000-MW utility unit. There are
recognized differences in the operating characteristics of each; the tests
were designed to cover a broad range of conditions, including those typical
of larger utility systems. For example, combustion zone residence times
typically range between 1.0 and 1.5 seconds and combustion intensities
between 10,000 and 100,000 Btu/hr/ft3. During the testing the combustion
zone residence time was varied from 0.7 to 1.8 seconds and the combustion
intensity from 36,000 Btu/hr/ft3 (high fire) to 15,000 Btu/hr/ft3 (low
fire). These changes did not significantly affect additive performance.
Further, in the case of Trimex, recent independent testing'5^ on a 1-MW
size unit found material of the same composition to be completely ineffective.
Thus, there is no reason to think that the effectiveness of any of the
materials would change markedly as a function of unit size.
Second, since the test fuel was residual oil no absolute conclusions can
be reached regarding the effectiveness of the additives with other fuels,
such as coal or wood. However, in the Trimex case, literature data on coal
testing show no effectiveness;* ' doubling the sulfur content of the oil in
the present study gave no change in performance. KAP has been evaluated
(8)
with coal;v ' however, unfortunately the conclusions were "not firm." Thus,
it appears that the data can at least be considered a guide to performance
with other fuels.
39
-------
Each additive test was started at a dose rate of 10 grams per minute
(relatively high) to ensure that the claimed reductions would be observed.
The authors had planned to then reduce the dose rate to determine the most
cost effective rate. However, in every case the additive was ineffective
at this "nominal" rate, so further reductions were useless. Unfortunately,
increasing the dosage was also Ineffective in producing performance.
With these considerations in mind the following conclusions can be
drawn from this program.
1. Trimex, PACE, KAP, and Glo-Klen were injected under a wide variety
of operating conditions; under no circumstances was there any significant
reduction in S0x emissions. Further, based on the results of this and other
work, there is no reason to believe that the material would be effective in
controlling S0x emissions from any boiler or furnace burning oil or coal.
2. The additives had no effect on UHC, CO, or NO emissions at any
condition investigated.
3. Widespread use of any of the additives tested could result in tons
of new emissions with unknown toxicity.
4. Potential operating problems, such as tube clogging and corrosion,
need further investigation before major use of any "pollution control" additive
is considered.
40
-------
BIBLIOGRAPHY
1. Milner, M. R. and F. B. Johnston. Combustion Adjuvant. U. S. Patent
3,630,696, December 28, 1971.
2. Milner, M. R. Combustion Adjuvant. U. S. Patent 3,628,925, December 21,
1971.
3. Pratt, R. Emissions Level Reduced in Tests. Savannah Morning News,
p 1, March 26, 1972.
4. Air Pollution . . . What is the Answer? Kleen-Aire Products, Inc.
Bulletin on KAP.
5. Brancaccio, J. and C. V. Flash. Use of Dry Fuel Additives to Reduce S0?
Emissions from a Full Size Industrial Coal-Fired Boiler. Presented at
65th Annual APCA Meeting, Miami Beach, Florida, June 1972.
6. Martin, 6. B., D. W. Pershing, and E. E. Berkau. Effects of Fuel
Additives on Air Pollutant Emissions from Distillate Oil-Fired Furnaces. EPA
Office of Air Programs Publication No. AP-87, June 1971.
7. Determination of Sulfur Dioxide Emissions from Stationary Sources,
Federal Register, Vol. 36, No. 159, pp. 15717-15718, August 17, 1971.
8. Siltala, C. A. Emission Testing at City of Columbus Power Plant During
Addition of KAP. Environment Control Corp. Final Report, July 1971.
41
-------
BIBLIOGRAPHIC DATA
SHEET
1. Report No.
KPA-fiRO/2-73-031
3. Recipient's Accession No.
5. Report Date
October 1973-
4. Title and Subtitle
Effectiveness of Selected Fuel Additives in Controlling
Pollution Emissions from Residual-Oil-Fired Boilers
6.
7. Author(s)
D. W. Pershing. G. B. Martin. E. E. Berkau. R.E. Hall
&• Performing Organization Rept.
No.
>. Performing Organization Name and Address
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, North Carolina 27711
10. Project/Task/Work Unit No.
11. Contract/Grant No.
In-House
12. Sponsoring Organisation Name and Address
NA
13. Type of Report & Period
Covered
Final
14.
IS. Supplementary Notes
16. Abstracts
The report gives results of a study to experimentally evaluate the effect-
iveness of four additive materials in controlling pollutant emissions from fossil fuel
combustion: Trimex, PACE, KAP, and Glo-Klen. Each was carefully examined in a
highly instrumented package boiler over the range of typical operating conditions
(e. g., combustion intensity and residence time) for industrial and utility systems.
Results show that none of the four reduce emissions of SOx, NO, CO, or UHC under
any condition tested. Based on these results, the boiler operating problems, and the
possibility that their use might create potentially harmful new emissions, none of the
additives can be recommended as a means of controlling pollutant emissions.
17. Key Words and Document Analysis.
Air Pollution
Fuel Additives
Residual Oil
Combustion
Sodium Carbonates
Sulfur Oxides
Hydrocarbons
Carbon Monoxide
Nitrogen Oxide (NO)
17b. Idemificrs/Open-Ended Terms
Air Pollution Control
Stationary Sources
Trimex
PACE
KAP
17o.
Pcscriptors
Boilers
Glo-Klen
Unburned Hydrocarbons
Metallic Emissions
Particulates
17c. COSATI Field/Group
13B. 13A. 21B
18. Availability Statement
Unlimited
19. Security Class (This
Report)
UNCLASSIFIED
LAf
Cli
20. Security Class (This
Page
UNCLASSIFIED
21. No. of Pages
49
22. Price
FORM IMTIS-3S IREV. 3-72)
42
USCOMM-OC M932-P72
------- |