<>EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/7-79-179
August 1979
Analysis of Thermal
Decomposition Products
of Flue Gas
Conditioning Agents
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide-range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-179
August 1979
Analysis of Thermal
Decomposition Products of Flue
Gas Conditioning Agents
by
R.B. Spafford, E.B. Dismukes, and H.K. Dillon
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
Contract No. 68-02-2200
Program Element No. EHE624
EPA Project Officer: Leslie E. Sparks
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 Industrial Environ-
mental Research Laboratory, U. S. Environmental Protection Agency,
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U. S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
11
-------
ABSTRACT
The reactions of several flue gas conditioning agents were
investigated in the laboratory under conditions simulating those
in the flue gas train of a coal-burning power plant. The primary
purposes of the study were to characterize the chemical species
resulting from the addition of conditioning agents to the flue
gas of a coal-fired power plant and to identify potentially
hazardous chemical species originating from the agents that may
be emitted into the environment.
To investigate the flue gas conditioning agents, a labora-
tory bench-scale facility was constructed that simulated the
composition of the flue gas of a coal-fired power plant and the
temperature zones and residence times relevant to the injection
of conditioning agents.
The "Compounds investigated were sulfur trioxide, ammonia,
triethylamine, sodium carbonate, ammonium sulfate, and diammonium
hydrogen phosphate. These compounds were injected into the simu-
lated flue gas train at various gas stream temperatures ranging
from 650 to 160 °C, and samples of flue gas were withdrawn at
various locations downstream from the points of injection of the
agents and were chemically analyzed.
The most prevalent types of reactions observed in these
experiments were thermal decomposition at high temperatures,
recombination of the decomposition fragments at lower temperatures,
and addition and substitution reactions of the agents or their
decomposition products with normal components of the flue gas.
With the exception of the organic compound triethylamine,
none of the conditioning agents studied appear to be capable of
discharging a significant amount of any environmentally unaccept-
able chemical species into the environment. However, the potent
carcinogen N-nitrosodiethylamine was found in trace amounts as a
thermal decomposition product of triethylamine when this agent
was injected into the flue gas stream at 160 °C. The formation
of this nitrosamine thus represents the only significant environ-
mental threat of any decomposition or reaction product identified
during this investigation.
111
-------
This report was submitted in fulfillment of Contract No.
68-02-2200 by Southern Research Institute under the sponsorship
of the U. S. Environmental Protection Agency. This report covers
the period January 7, 1977, to March 31, 1979.
IV
-------
CONTENTS
Abstract iii
Figures vi
Tables vii
Acknowledgments „ x
1. Introduction 1
Technical Background 1
Choices of Conditioning Agents for Study 2
Design and Fabrication of Apparatus for Reaction
Studies 5
Kinetic Studies of Conditioning Agents 5
2. Conclusions and Recommendations 6
Conclusions 6
Recommendations 9
3. Acquisition and Analysis of Conditioning Agents for
Experimental Study 10
Nonproprietary Agents 10
Proprietary Agents 11
4. Design and Fabrication of a Laboratory-Scale Flue Gas
Train 16
General Considerations 16
Design of the Flue Gas Train 19
Fabrication of the Flue Gas Train 27
5. Investigation of Individual Conditioning Agents in the
Flue Gas Train 30
Sulfur Trioxide . 30
Ammonia 36
Triethylamine 49
Sodium Carbonate 61
Ammonium Sulfate 67
Diammonium Hydrogen Phosphate 78
References 98
v
-------
FIGURES
Number Page
1 Schematic diagram of flue gas train... * 18
2 Detailed illustration of the quartz 650 °C constant
temperature zone 22
3 Photograph of assembled laboratory-scale flue gas train 29
4 Current-voltage relationship for the ESP reflecting
the space charge effect from the combination of
ammonia and sulfur trioxide 45
5 Current-voltage relationship for the ESP reflecting
the space charge effect from the combination of
triethylamine and sulfur trioxide 57
VI
-------
TABLES
Number Page
1 'Dimensions of Constant Temperature Zones 21
2 Collection Efficiency and Corona Power for Various
Migration Velocities 23
3 Representative Flow Rates of the Individual Gases in
the Simulated Flue Gas Train 26
4 Determination of Sulfur Oxides without Injected Sulfur
Trioxide 33
5 Determination of Sulfur Oxides with Injected Sulfur
Trioxide „ 34
6 Determination of Injected Sulfur Trioxide 35
7 Determination of Ammonia in Flue Gas Containing No
Sulfur or Nitrogen Oxides 41
8 Determination of Ammonia in Flue Gas Containing Sulfur
Dioxide but No Added Sulfur Trioxide or Nitrogen
Oxides . . . „ 43
9 Determination of Ammonia and Sulfate in Flue Gas Con-
taining Added Sulfur Dioxide and Sulfur Trioxide but
No Added Nitrogen Oxides 44
10 Determination of Ammonia in Flue Gas Containing Added
Sulfur Dioxide and Nitrogen Oxides 46
11 Determination of Triethylamine in Flue Gas Containing
No Sulfur or Nitrogen Oxides 54
12 Determination of Triethylamine in Flue Gas Containing
Added Sulfur Dioxide but No Added Sulfur Trioxide
or Nitrogen Oxides 55
13 Determination of Triethylamine in Flue Gas Containing
Added Sulfur Dioxide and Sulfur Trioxide but No
Added Nitrogen Oxides 56
vn
-------
TABLES (Continued)
Number Pagi
14 Determination of Triethylamine in Flue Gas Containing
Added Sulfur Dioxide and Nitrogen Oxides 58
15 Determination of Anions in Flue Gas Containing Sodium
Carbonate but No Added Reactive Gases 64
16 Determination of Anions in Flue Gas Containing Sodium
Carbonate and Reactive Gases 66
17 Determination of Sulfite in Flue Gas Containing Sodium
Carbonate and Reactive Gases 66
18 Determination of Thermal Decomposition Products of
LPA-40 at 650 °C in Flue Gas Containing No Added
Sulfur or Nitrogen Oxides 71
19 Determination of Recombination Products of LPA-40 in
Flue Gas Containing No Added Sulfur or Nitrogen
Oxides 73
20 Determination of Decomposition Products of LPA-40 at
650 °C in Flue Gas Containing Added Sulfur and
Nitrogen Oxides 74
21 Determination of Recombination Products of LPA-40 in
Flue Gas Containing Added Sulfur and Nitrogen Oxides. 76
22 Determination of Recombination Products of LPA-40 in
Flue Gas Containing Added Nitrogen and Sulfur Oxides. 77
23 Dissociation Pressures of Ammonium Orthophosphate
Salts as a Function of Temperature 81
24 Recovery of Decomposition Products of LPA-445 at
650 °C in Flue Gas Containing No Added Sulfur or
Nitrogen Oxides 86
25 Distribution of Decomposition Products of LPA-445 at
650 °C in Flue Gas Containing No Added Sulfur or
Nitrogen Oxides 86
26 Recovery of Recombination Products of LPA-445 at 370,
160, and 90 °C in Flue Gas Containing No Added Sulfur
or Nitrogen Oxides 88
Vlll
-------
TABLES (Concluded)
Number
27 Distribution of Recombination Products of LPA-445 at
370, 160, and 90 6C in Flue Gas Containing No Added
Sulfur or Nitrogen Oxides 89
28 Recovery of Decomposition Products of LPA-445 at
650 °C in Flue Gas Containing Added Sulfur Dioxide
but No Added Nitrogen Oxides 92
29 Distribution of Decomposition Products of LPA-445 at
650 °C in Flue Gas Containing Added Sulfur Dioxide
but No Added Nitrogen Oxides 92
30 Recovery of Recombination Products of LPA-445 at 160
and 90 °C in Flue Gas Containing Sulfur Dioxide and
Sometimes Nitrogen Oxides 94
31 Distribution of Recombination Products of LPA-445 at
160 and 90 °C in Flue Gas Containing Sulfur Dioxide
and Sometimes Nitrogen Oxides 95
32 Ion Chromatographic Analysis of Anionic Species in
Filter Washes of Samples of Flue Gas Containing
Sulfur Dioxide and Nitrogen Oxides 95
IX
-------
ACKNOWLEDGMENTS
The cooperation of Dr. Robert P. Bennett, Vice President
and Technical Director of Apollo Chemical Corporation, Whippany,
New Jersey, in supplying Southern Research Institute with several
samples of Apollo's proprietary flue gas conditioning formulations
for study in this project is gratefully acknowledged.
Several members of the staff of Southern Research Institute
provided great assistance in the successful completion of this
project and are to be especially thanked. Mr. Don B. Hooks,
Associate Chemist, and Mr. J. Todd Brown, III, Assistant Chemist,
collected much of the experimental data summarized in this report.
Mr. Norman Francis, Senior Mechanical Engineer, participated in
the design of the flue gas train, and Mr. Josiah E. Smith, Jr.,
Senior Chemist, in its fabrication. Ms. Ruby H. James, Senior
Chemist, performed the GC-MS analyses; Ms. Patricia A. Goodman,
Associate Chemist, performed some of the GC analyses; Mr. Marion
C. Kirk, Research Chemist, obtained the mass spectra;
Ms. Christina Richards, Technical Specialist, obtained the infra-
red spectra; and Mr. David W. Mason, Associate Chemist, and
Mr. John C. Harmon, Chemical Technician, performed the ion chro-
matographic analyses. Additional analytical assistance was pro-
vided by Mr. Richard Wilson, Assistant Chemist, and Mr. Lester W.
Romine, Assistant Chemist.
Dr. William J. Barrett, Head of the Analytical and Physical
Chemistry Division, served as the overall technical supervisor of
this project.
x
-------
SECTION 1
INTRODUCTION
This report presents the results of a laboratory investiga-
tion of reactions of flue gas conditioning agents under conditions
simulating those in the gas train of a coal-burning electric power
plant. The reactions studied were of three types: thermal decom-
position at high temperatures, recombination of the decomposition
fragments at lower temperatures, and reactions with normal compo-
nents of flue gas. The principal purposes of this investigation
were to characterize the chemical species resulting from the addi-
tion of conditioning agents to the flue gas of a coal-fired power
plant and specifically to identify toxic substances originating
from the agents that can potentially undergo stack discharge to
the environment.
TECHNICAL BACKGROUND
The recent widespread use of low-sulfur coals in electric
power production has led to difficulty in achieving efficient col-
lection of fly ash in electrostatic precipitators at temperatures
around 150 °C, where most of the older precipitators operate.
This difficulty is attributable to an increase in the electric
resistivity of the ash, which limits the useful voltage and cur-
rent that can be maintained in a precipitator. A major reason
for the increase in resistivity is the lower rate of production
of the vapor of sulfuric acid (HaSO^) in the flue gas and the
correspondingly smaller quantity of this conductive substance that
is adsorbed on the surfaces of fly-ash particles.1
1. Dismukes, E. B. A Study of Resistivity and Conditioning of
Fly Ash. EPA-R2-72-087, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1972. 138 pp,
-------
Several approaches may be taken to upgrade precipitator per-
formance.2 One approach is to enlarge the precipitator and thus
increase the electrode collecting area and the residence time of
the gas stream during treatment. Another approach is to relocate
the precipitator in the gas stream to a position upstream from
the air preheater, where the gas temperature is 300 °C or greater
and the ash resistivity is low enough without surface conditioning
by sulfuric acid. Still other approaches involve steps to alter
resistivity when the precipitator operates downstream from the
air preheater. The gas temperature may be lowered from 150 °C to
around 100 °C to make the water vapor in flue gas more effective
as a surface conditioning agent; such a reduction in temperature
may be accomplished by increasing the ratio of combustion air to
flue gas passing through the air preheater and discarding the
excess air. The remaining approach, which was of basic concern
in this investigation, involves chemical conditioning of the flue
gas: that is, addition of a chemical compound to the gas stream
upstream from a precipitator to obtain the resistivity effect nor-
mally produced by sulfuric acid or to obtain some other beneficial
effect besides the modification of electrical resistivity.
CHOICES OF CONDITIONING AGENTS FOR STUDY
The most widely used method of chemical conditioning is to
add sulfuric acid from an external source and supplement the
limited quantity of this substance occurring in flue gas.1'3 The
compound may be added as the acid itself or as sulfur trioxide
(SOs); when the latter substance is added, it is converted to sul-
furic acid in the gas stream by combination with water vapor.
Injection of either substance is normally made between the air
preheater and the precipitator, at a location as near as possible
to the preheater to allow adequate time for interaction with sus-
pended fly ash. When properly designed, a system injecting either
sulfuric acid or sulfur trioxide performs satisfactorily, upgrad-
ing precipitator performance if factors other than resistivity do
not limit performance.
A wide variety of other chemical conditioning agents have
been investigated in the laboratory, in pilot plants, or in full-
scale precipitators. Many of the substances, unlike sulfuric
2. Sparks, L. E. Electrostatic Precipitator Options for Collec-
tion of High Resistivity Ash. In: Symposium on Particulate
Control in Energy Processes. EPA-600/7-76-010, U. S. Environ-
mental Protection Agency, Research Triangle Park, North
Carolina, 1976. pp 127-141.
3. Dismukes, E. B. Conditioning of Fly Ash with Sulfur Trioxide
and Ammonia. EPA-600/2-75-015, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1975. 157 pp.
-------
acid, are not normal constituents of flue gas. However, they
have usually been expected to give a similar effect on resistiv-
ity. In actuality, some of these compounds do alter resistivity,
but some operate through different mechanisms that still improve
precipitator performance.
After sulfuric acid, ammonia (NHs) is one of the most popu-
lar conditioning agents.3 It is used with a great deal of success
in Australia, for example, where it exceeds or rivals sulfuric
acid in importance. ** Frequently, ammonia is assumed to lower
resistivity by adsorption on fly-ash surfaces, but its importance
in this connection has not been clearly demonstrated. Available
research data have shown that ammonia operates by two entirely
different mechanisms: (1) by creating a favorable space-charge
effect in the gas stream within a precipitator and (2) by increas-
ing the cohesiveness of fly ash after deposition.3'5 The first
effect gives more efficient collection of ash; the second mini-
mizes loss of collected ash by reentrainment. Conditioning with
ammonia may also suppress one of the unfavorable manifestations
of high resistivity: back corona or reverse ionization occurring
in the collected ash.
Triethylamine, an organic base similar to ammonia
([CaHsJsN), has been receiving attention recently. Some investi-
gators argue that the amine acts as an agglomerating agent,
increasing the effective average size of fly-ash particles before
precipitation occurs and thus helping to overcome the basic diffi-
culty of precipitating small particles.6 It is probable, however,
that the amine acts also or mainly through the mechanisms
described for ammonia.
Sodium carbonate (NaaCOs) has recently been receiving atten-
tion as a conditioning agent for use in plants operating with
either a hot-side precipitator (temperature, around 370 °C) or a
cold-side precipitator (temperature, around 150 °C). The compound
may be added to the boiler along with the coal, in which case
4. Watson, K. S. Australian Experience with Flue Gas Condition-
ing. In: Symposium on Particulate Control in Energy Pro-
cesses. EPA-600/7-76-010, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1976. pp 189-
216.
5. Dismukes, E. B. Conditioning of Fly Ash with Ammonia. J.
Air Pollut. Contr. Assoc., 25:152-156, 1975.
6. Potter, E. C., and C. A. J. Paulson. Improvement of Electro-
static Precipitator Performance by Carrier Gas Additives and
Its Graphical Assessment Using an Extended Deutsch Equation.
Chem. Ind., 1974:532-533.
-------
sodium oxide from the carbonate compound is incorporated in the
fly ash as a resistivity-lowering constituent.7 Alternatively,
the sodium compound may be added to the flue gas just ahead of the
electrostatic precipitator , in which case it is coprecipitated
with the ash, serving as a conductive species in a mixture with
the relatively nonconductive fly ash.8
Still other compounds used for conditioning occur as con-
stituents of proprietary formulations of such companies as Apollo
Chemical Corporation. During this research, as described in Sec-
tion 3, different Apollo formulations were analyzed and found to
contain ammonium, sul fate ( [NHn ] aSOt, ) , sodium bisulfate (NaHSOiJ,
diammonium hydrogen phosphate ( [NH^ ] zHPO^ ) , and urea ([NH2l2-C=0)
as the principal constituents. It is conceivable that each of
these compounds may, to a degree, operate through each of the con-
ditioning mechanisms described above.
The conditioning agents selected for study during this
investigation were as follows:
• Sulfur trioxide
• Ammonia (NHs)
• Triethylamine ([C2H5]3N)
• Sodium carbonate (Na2C03)
• Ammonium sul fate { [NHi» ] 2SOi»)
• Diammonium hydrogen phosphate ( [NHi» ] 2HPOit )
The basis for the selection of each of these compounds except tri
ethylamine was the relatively widespread use known to occur pres-
ently. Triethylamine, the only organic compound in this group,
was selected because of its potential use and its potential con-
version to toxic organic compounds.
7. Bickelhaupt, R. E. Sodium Conditioning to Reduce Fly Ash
Resistivity. EPA-650/2-74-092, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1974. 10 pp.
8. Selle, S. J., and L. L. Hess. Factors Affecting ESP Perfor-
mance on Western Coals and Experience with North Dakota
Lignites. In: Symposium on Particulate Control in Energy
Processes. EPA-600/7-76-010, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1976. pp 105-
125.
-------
DESIGN AND FABRICATION OF APPARATUS FOR REACTION STUDIES
The first task completed during this investigation was the
design and fabrication of a laboratory-scale gas train simulating
pertinent components of the flue gas train in a full-scale elec-
tric power station. Requirements of this train were that it
should simulate the temperature-time profile of an actual flue
gas train extending from the highest temperature at which a condi-
tioning agent might be added to the lowest temperature reached in
the stack prior to gas emission to the environment. Accordingly,
an apparatus with temperatures decreasing in steps from 650 to
90 °C was fabricated. This simulated flue gas train is described
in Section 4.
The gas stream maintained in this apparatus simulated the
composition of flue gas from the combustion of low sulfur coal.
The most fundamental and consistently used gas-phase components
were 76% nitrogen, 12% carbon dioxide, 4% oxygen, and 8% water
vapor. As dictated by the objectives of individual experiments,
the following additional components were added on occasion: up
to 500 or 600 ppm of sulfur dioxide, up to 25 ppm of sulfur tri-
oxide, 100 ppm of nitrogen dioxide, and 1000 ppm of nitric oxide.
In one important aspect, however, no attempt was made to simulate
the composition of flue gas; no attempt was made to include sus-
pended fly ash as a constituent of the flue gas because of techni-
cal problems.
KINETIC STUDIES OF CONDITIONING AGENTS
The simulated flue gas train was fitted with ports through
which conditioning agents could be injected at various tempera-
tures and through which the reaction products could be withdrawn
for analysis at various locations downstream from the injection
sites. A variety of devices for sampling the reaction products
was employed. The choices of sampling devices and the methods of
analysis depended upon the type of conditioning agent injected
and the type of reaction product expected. Details of the experi-
mental procedures and results with the individual conditioning
agents are subsequently described in Section 5.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
The principal conclusions reached during this research pro-
gram are discussed below under headings for each of the condition-
ing agents studied.
Sulfur Trioxide
This agent was investigated mainly in control experiments
to determine whether this highly reactive substance could be main-
tained in the flue gas stream and its loss by wall effects mini-
mized. It was injected only at 370 °C. The compound was presum-
ably quantitatively converted to sulfuric acid vapor by reaction
with water vapor; however, no evidence of loss by other chemical
reactions was detected, and indications of wall losses were over-
come by taking the necessary experimental precautions.
Sulfur trioxide represents an environmental threat by its
escape from the stack of a power plant as a condensed mist of sul-
furic acid or as a sulfate coating on fly ash particles. The dis-
tribution of the agent between these two forms was not determined
in this investigation inasmuch as fly ash was absent from the
experimental system. However, previous investigations3/9 have
shown that the distribution probably will be controlled by the
alkalinity of the ash; at high alkalinity sulfate on the ash will
predominate, and at low alkalinity sulfuric acid mist will pre-
dominate .
Ammonia
This agent was injected at temperatures of 650, 370, and
160 °C. It was found to participate in reactions with sulfuric
acid and nitrogen oxides that are normally present in flue gas.
9. Dismukes, E. B., and J. P. Gooch. Fly Ash Conditioning with
Sulfur Trioxide. EPA-600/2-77-242, U. S. Environmental Pro-
tection Agency, Research Triangle Park, North Carolina, 1977.
65 pp.
-------
The reaction with sulfuric acid leads to the formation of bisul-
fate and sulfate salts, which are solids. The reactions with
nitrogen oxides were not fully characterized, but they apparently
lead to the formation of solid nitrate salts among other possible
products.
Depending upon the stoichiometry of ammonia and its reac-
tants present in the flue gas, ammonia may be largely converted
to solid compounds. The products may then be effectively removed
from the gas stream by collection in the electrostatic precipita-
tor. The collection of bisulfate and sulfate salts in this manner
has been previously demonstrated.3 However, if there is not a
molar equivalence of sulfuric acid present, the excess ammonia
will not react with sulfur dioxide (as an alternative to sulfuric
acid) to produce a bisulfite or sulfite salt. In this event,
some stack loss of ammonia to the environment may occur.
Triethylamine
This compound decomposes extensively if injected into the
flue gas stream at 650 °C. Only traces of the reaction products
thus produced were observed, however. The compounds observed were
ammonia, hydrogen cyanide, nitrogen dioxide, and nitric oxide; no
organic residue was identified.
It is thus practical to inject triethylamine only at a tem-
perature of about 370 °C or less. Under these conditions, the
compound remains largely intact molecularly, but it is capable of
reacting with either sulfur dioxide or sulfuric acid to form the
bisulfite, sulfite, bisulfate, or sulfate salt.
The most noteworthy product of decomposition of triethyl-
amine is the carcinogen N-nitrosodiethylamine.J° It appears that
diethylamine—either a primary decomposition product of triethyl-
amine or an impurity in the injected agent—is a precursor of the
N-nitroso compound. The concentrations of the N-nitroso compound
were low, only 70 to 200 ppb, compared with the injected concen-
tration of triethylamine of about 25 ppm. Even so, the formation
of the N-nitroso compound appears to represent the most signifi-
cant environmental threat of any decomposition or reaction product
identified during this investigation.
Sodium Carbonate
This agent was injected only into the 160 °C reaction zone,
as in the current field investigation of this compound under
Environmental Protection Agency Contract 68-02-2656. The molar
10. Magee, P. N., and J. M. Barnes. Carcinogenic Nitroso Com-
pounds. In: Advances in Cancer Research, Vol. 10, A. Haddow
and S. Weinhouse, Eds. Academic Press, New York, New York.
pp 163-246.
-------
components of this compound—sodium oxide and carbon dioxide—
exist in such a stable state that no thermal fragmentation of the
compound occurred at 160 °C. However, reactions of sodium carbon-
ate occurred with acidic flue gas—sulfur dioxide and sulfuric
acid—leading to the formation of sodium sulfite and sodium sul-
fate and the evolution of carbon dioxide. The environmental
impact of sodium carbonate and its reaction products should be
minimized by the collection of the solid reaction products in the
electrostatic precipitator.
Ammonium Sulfate
This compound was injected at 650 °C in either of two ways—
as ordinary ammonium sulfate in aqueous soltuion or as the pro-
prietary product of Apollo Chemical Corporation known as Coaltrol
LPA-40 (also an aqueous solution). The high injection temperature
was selected because of Apollo's practice of injecting LPA-40
ahead of the economizer in a full-scale plant. To the extent that
comparisons were made between the two forms of the conditioning
agent, the results seemed to be identical.
At 650 °C, ammonium sulfate decomposed primarily into its
constituent compounds: ammonia, sulfur trioxide, and water. Upon
cooling of the flue gas to 160 or 90 °C, recombination of the
molecular fragments occurred. Ammonium bisulfate appeared to be
the principal recombination product at 160 °C, and ammonium sul-
fate seemed to be predoninant at 90 °C.
The implications of these results with respect to stack
emissions are that some ammonia may be present in the stack emis-
sions and ammonium salts may also be present if these solids are
not effectively removed in the electrostatic precipitator.
Diammonium Hydrogen Phosphate
This compound was injected only as the commercial form sold
by Apollo Chemical Corporation as Coaltrol LPA-445. It was
injected at 650 °C, inasmuch as Apollo recommends high temperature
injection to produce thermal decomposition of the compound.
Decomposition of the compound to ammonia and unidentified
phosphoric acid species appeared to occur at 650 °C. Ammonia and
the phosphoric acid species that were found analytically were not
in the expected 2:1 mole ratio. In part, the discrepancy was due
to the experimental failure to inject the compounds in the correct
2:1 stoichiometric ratio; the injection rate of ammonia was more
than twice that of the phosphoric acid species on the mole basis.
However, the ratios of ammonia and phosphoric acid concentrations
found were even higher than the ratios injected. This observation
is tentatively explained as the result of wall losses of ortho-
phosphoric acid, condensed phosphoric acids, or perhaps even phos-
phorus pentoxide.
-------
Recombination of the high temperature fragmentation products
was observed at 160 to 90 °C. The most logical explanation for
the mechanism of recombination was that ammonia and phosphoric
acid recombined principally as ammonium dihydrogen phosphate (with
the mole ratio of ammonia to phosphoric acid being 1:1 rather than
2:1 as at the beginning). A considerable excess of ammonia vapor
remained in the gas stream after the solid phosphate was removed
by filtration.
It thus appears that ammonia will be emitted from the stack
of a power plant when diammonium hydrogen phosphate is used for
conditioning. If not removed by electrostatic precipitation,
ammonium dihydrogen phosphate particles will also be emitted.
RECOMMENDATIONS
Experimental studies of the type described in this report
should be pursued as new agents are brought into use. Representa-
tives of Apollo Chemical Corporation state that new proprietary
agents are being developed by that company. Nalco Chemical Corpo-
ration reportedly is active in the development of new conditioning
agents, and Betz Laboratories has disclosed its activity in this
area over the past 2 yr. Particular attention should be paid to
any formulation that contains an organic compound.
Further work should be done to confirm the most crucial
finding of the present investigation: that N-nitrosodiethylamine
is produced from triethylamine. Additional work should also be
done to determine how the operation of an electrostatic precipita-
tor or the presence of fly ash in the flue gas influences the
formation and removal of the N-nitroso compound.
Efforts should be made to modify the present laboratory
facility simulating the flue gas train of a power plant in such a
way that fly ash can be introduced and kept in suspension. If
these efforts are successful, then the system including fly ash
should be used to generate particulate samples treated by a
variety of conditioning agents. Such samples should be screened
for biological activity by the Ames mutagenesis assay technique11
and perhaps by other screening techniques for biological activity.
11. Ames, B., J. McCann, and E. Yamasaki. Methods for Detecting
Carcinogens and Mutagens with the SaImone11a/Mamma1ian-
Microsome Mutagenicity Test. Mutation Res., 31:347-364,
1975.
-------
SECTION 3
ACQUISITION AND ANALYSIS OF CONDITIONING
AGENTS FOR EXPERIMENTAL STUDY
The chemical conditioning agents acquired for study during
this investigation can best be categorized as either proprietary
or nonproprietary agents. Although both classes of conditioning
agents are obtained commercially for use in coal-burning power
plants, some of the additives used are standard industrial chemi-
cals, whereas others are proprietary formulations that are both
sold and injected by commercial vendors.
NONPROPRIETARY AGENTS
Five commercially available chemicals were acquired for
study during this investigation based on either their current
widespread use or their potential use as flue gas conditioning
agents. All of the commercial chemicals studied in this investi-
gation were reagent-grade chemicals. The chemicals used, their
purities, and their commercial sources are as follows:
• Sulfur trioxide (SOa)—Sulfuric acid, fuming (oleum),
18-24%, reagent ACS, Fischer Scientific Company,
Fair Lawn, New Jersey.
• Ammonia (NH3)—ammonium hydroxide, (NHi,OH) approxi-
mately 58%, Mallinkrodt analytical reagent, Mallin-
krodt, Inc., St. Louis, Missouri.
• Ammonia (NH3)—1.0% anhydrous ammonia (gas) in nitro-
gen, Linde Division, Union Carbide Corporation,
Birmingham, Alabama.
• Triethylamine ([C2H5]3N)—99% minimum purity (assayed
by GLC), Eastman Organic Chemicals, Eastman Kodak
Company, Rochester, New York.
• Sodium carbonate (NaaCOs)—Anhydrous powder, 'Baker
Analyzed1 Reagent, J. T. Baker Company, Phillipsburg,
New Jersey.
10
-------
• Ammonium sulfate ([NHi»] 2SOi»)—Granular, 'Baker
Analyzed1 Reagent, J. T. Baker Company, Phillipsburg,
New Jersey.
PROPRIETARY AGENTS
All of the proprietary conditioning agents acquired for use
in this investigation were obtained from the Apollo Chemical
Corporation of Whippany, New Jersey. Samples representing two
series of additives were obtained from Apollo, the LPA series and
the LAC series. The LPA series of additives are for "high temper-
ature" applications, where the conditioning agent is injected at
gas stream temperatures up to 900 °c.12~114 The LAC series of
additives are for "low temperature" applications, where the condi-
tioning agent is injected at some point between the air preheater
and the electrostatic precipitator (personal communication with
Dr. R. P. Bennett of Apollo Chemical Corporation). This corre-
sponds to gas stream temperatures of approximately 160 °C.
The agents obtained from Apollo Chemical Corporation were
Coaltrol LPA-40, Coaltrol LPA-410, Coaltrol LPA-445, Coaltrol
LAC-50, and Coaltrol LAC-51B. The compositions of these formula-
tions were not revealed to us by Apollo Chemical Corporation, but
some indication of the compositions was available from a series
of U. S. patents granted to Apollo in 1977 and 197812""15 and from
the analyses of some of the formulations by independent labora-
tories. The compositions of these proprietary formulations sup-
plied to us by Apollo were investigated in our laboratories by
several analytical techniques. The description and results of
these analyses are presented in the following sections.
12. Bennett, R. P., and M. J. O'Connor. Method of Conditioning
Flue Gas to Electrostatic Precipitator. U. S. Patent
4 043 768, August 23, 1977. Assigned to Apollo Chemical
Corporation, Whippany, New Jersey.
13. Bennett, R. P., M. J. O'Connor, A. E. Kober, and I. Kukin.
Method of Conditioning Flue Gas to Electrostatic Precipita-
tor. U. S. Patent 4 042 348, August 16, 1977. Assigned to
Apollo Chemical Corporation, Whippany, New Jersey.
14. Bennett, R. P., and A. E. Kober. Method of Conditioning Flue
Gas. U. S. Patent 4 113 447, September 12, 1978. Assigned
to Apollo Chemical Corporation, Whippany, New Jersey.
15. Kober, A. E., and I. Kukin. Method of Agglomerating Parti-
cles in Gas Stream. U. S. Patent 4 070 162, January 24, 1978.
Assigned to Apollo Chemical Corporation, Whippany, New Jersey,
11
-------
Coaltrol LPA-40
Coaltrol LPA-40 (Batch No. A217-44A) is a light brown solu-
tion with a density of 1.27 g/raL and a pH of 3.9. The qualitative
identification of the primary components was performed by infrared
spectroscopy and ion chromatography.16 The solid material result-
ing from the evaporation of the water from the formulation pos-
sessed an infrared spectrum that was identical to that of ammonium
sulfate, ([NHiJ 2SOi,) . Analysis of the formulation by ion chroma-
tography revealed that sulfate ion was the only anion of a strong
acid that was present in significant amounts.
The LPA-40 formulation was quantitatively analyzed for
ammonium ion by the phenol-hypochlorite colorimetric method17 and
by acid-base titration with sodium hydroxide solution, and the
sulfate concentration was determined by titration with barium per-
chlorate solution using Thorin as an indicator.18 The ammonium
ion concentration was found to be 5.9 mmol/g and the sulfate con-
centration 3.0 mmol/g. Within the experimental error of these
determinations, there were found twice as many millimoles of
ammonium ion as sulfate ion, consistent with the identification
of the solid residue from the evaporation of the formulation as
ammonium sulfate. The LPA-40 formulation thus consists of about
40% w/w ammonium sulfate, with water accounting for the remaining
60%.
A reddish-brown solid gradually precipitated from the LPA-40
formulation on standing. The precipitate was filtered from a por-
tion of the formulation and found to contain iron amounting to
30% w/w of the precipitate. The filtrate was found to contain
2.6 mg/mL iron in solution. The original formulation, after being
shaken well to suspend the solid material, was found to contain
2.7 mg/mL iron. The concentration of iron in the original solu-
tion was, therefore, about 0.2% w/w. Little of the iron appeared
to be lost from solution by precipitation. By spot tests with
ferrocyanide and ferricyanide,J9 both ferrous and ferric ions
were identified in the formulation. No attempt was made to deter-
mine their relative amounts.
16. Small, H., T. S. Stevens, and W. C. Bauman. Novel Ion
Exchange Chromatographic Method Using Conductimetric Detec-
tion. Anal. Chem., 47:1801-1809, 1975.
17. Harwood, J. E. , and A. L. Kiihn. A Colorimetric Method for
Ammonia in Natural Waters. Water Res., 4:805-811, 1970.
18. Fritz, J. S., and S. S. Yamamura. Rapid Microtitration of
Sulfate. Anal. Chem., 27:1461-1464, 1955.
19. Sorum, C. H. Introduction to Semimicro Qualitative Analysis.
Prentice Hall, Inc., Englewood Cliffs, New Jersey, 1967.
277 pp.
12
-------
Coaltrol LPA-410
Coaltrol LPA-410 (Batch No. A217-44B) is a clear, light blue-
green solution with a density of 1.33 g/mL and an apparent pH of
about zero. The identities of the primary components of the
formulation were suggested from a prior chemical analysis per-
formed by an independent laboratory for a confidential source.
This analysis indicated that LPA-410 was a rather complex aqueous
solution including sulfuric acid, sodium sulfate (or sodium bisul-
fate), and sodium benzenesulfonate. The formulation was qualita-
tively analyzed at Southern Research Institute by infrared spec-
troscopy and ion chromatography, and sulfate (or bisulfate) ion
was identified as the only anion present in significant amounts
in the solution.
The concentration of hydrogen ion in LPA-410 was found to
be 3.16 mmol/g by acid-base titration. The total sulfate concen-
tration was determined to be 3.05 mmol/g by titration with barium
perchlorate using Thorin as an indicator.18 The concentration of
sodium ion was found to be 3.39 mmol/g by atomic emission spec-
troscopy. The weight percent of water in the formulation was
determined physically from the loss in weight of a known volume
of LPA-410 that was evaporated to dryness in an oven at 150 °C.
Within the experimental error, the molar ratio of sodium to
hydrogen to sulfate was about 1:1:1. Thus, it appears that the
major component of LPA-410 is sodium bisulfate, NaHSOi*. The
formulation may be prepared by dissolving sodium bisulfate or an
equimolar mixture of sulfuric acid and sodium sulfate. The
approximate concentration of sodium bisulfate is 35 to 40% w/w.
No evidence of sodium benzenesulfonate was found, contrary to
earlier results found elsewhere.
The LPA-410 formulation was also analyzed by atomic absorp-
tion spectroscopy for iron and found to contain approximately
0.2% w/w iron. Ferrocyanide and ferricyanide spot tests19 on the
formulation indicated the presence of both ferrous and ferric
ions in the solution. Unlike the iron content of the LPA-40
formulation, however, the iron species in LPA-410 were all in
solution. Finally, the LPA-410 formulation was analyzed by atomic
emission spectroscopy specifically for cobalt, chromium, potassium,
sodium, nickel, tin, and zinc; measurable signals were obtained
only for sodium and potassium.
Coaltrol LPA-445
Coaltrol LPA-445 is a clear, colorless liquid with a dis-
tinct ammoniacal odor. Its density is 1.14 g/mL, and its pH is
approximately 8.2. Mass spectrometric analysis of the solid
residue remaining after evaporation of the formulation to dryness
indicated that the solid residue was an ammonium phosphate salt.
13
-------
LPA-445 was identified as an aqueous solution of diammonium
hydrogen phosphate ([NHiJ aHPCK) and the concentration of the solu-
tion determined by analyzing the formulation for ammonium ion by
the phenol-hypochlorite colorimetric method17 and for orthophos-
phate ion by ion chromatography16 and by the vanadomolybdophos-
phoric acid colorimetric method.28 The ammonium ion concentration
in the solution was found to be 3.71 mmol/g and the phosphate ion
concentration 1.87 mmol/g. Within the experimental error of these
determinations, there are twice as many millimoles of ammonium
ion as there are of phosphate ion. This corresponds to an empiri-
cal formula of (NHit)2HPOi* for the LPA-445 solute. Analysis of
the formulation by ion chromatography16 also revealed the presence
of trace amounts of sulfate ion (approximately 7.30 pmol/g or
0.08% w/w) as the only other anion of measurable concentration in
LPA-445. The LPA-445 formulation thus consists of about 24% w/w
diammonium hydrogen phosphate, with water accounting for the
remaining 76% of the weight.
Two separate samples of LPA-445 were used during the inves-
tigation of this conditioning agent (Sample No. A217-9916 and
Sample No. A217-9831). Each of the samples was analyzed individu-
ally, and the two samples were found to be identical within the
experimental limits of error of the chemical methods used in the
analyses.
Coaltrol LAC-50
Coaltrol LAC-50 (Sample No. A217-9915) is a clear, color-
less, odorless aqueous liquid with a density of 1.09 g/mL and a
pH of approximately 4.2. The solid residue resulting from the
evaporation of the water from the formulation was analyzed quali-
tatively by infrared spectroscopy and mass spectrometry. Both the
infrared and mass spectra of the solid residue were identical to
the corresponding spectra of pure urea. Thus, Coaltrol LAC-50
was qualitatively identified as an aqueous solution of urea.
The LAC-50 formulation was quantitatively analyzed by the
urea-specific p-dimethylaminobenzaldehyde colorimetric method.21
The concentration of urea in the LAC-50 solution was found to be
5.43 mmol/g (326 mg/g). Within the experimental error, this con-
centration of urea was identical to the concentration of solute
found in the solution after evaporation of the LAC-50 formulation
to dryness (338 mg/g). These results indicate that LAC-50 con-
sists of approximately 33% w/w of urea in water.
20. Kitson, R. E., and M. G. Mellon. Colorimetric Determination
of Phosphorus as Molybdivanodophosphoric Acid. Ind. Eng.
Chem., 16:379-383, 1944.
21. Watt, G. W., and J. D. Chrisp. Spectrophotometric Method
for Determination of Urea. Anal. Chem., 26:452-453, 1954.
14
-------
Coaltrol LAC-51B
Coaltrol LAC-51B (Sample No. A217-9922) is a clear, color-
less solution with a distinct ammoniacal odor. Its density is
1.14 g/mL, and its pH is approximately 8.3. Both the mass spec-
trum and infrared spectrum of the solid residue remaining after
evaporation of the formulation to dryness were identical to the
corresponding spectra of the evaporative residue of LPA-445.
These results indicate that LAC-51B, like LPA-445, is an aqueous
solution of diammonium hydrogen phosphate.
The LAC-51B formulation was quantitatively analyzed for
ammonium ion by the phenol-hypochlorite colorimetric method17 and
for orthophosphate ion by ion chromatography16 and by the vanado-
molybdophosphoric acid colorimetric method.20 The ammonium ion
concentration in the solution was found to be 3.71 mmol/g and the
phosphate ion concentration 1.89 mmol/g. Within the experimental
error, there is thus a 2:1 mole ratio of ammonium ion to phosphate
ion in the formulation, corresponding to an empirical formula of
(NHi»)2HP(H for the LAC-51B solution. No other anion of measurable
concentration other than orthophosphate was found in the formula-
tion by ion chromatography. Thus, with the exception of the
absence of trace amounts of sulfate ion, the LAC-51B formulation
is identical to LPA-445, consisting of a 24% w/w solution of
diammonium hydrogen phosphate in water.
15
-------
SECTION 4
DESIGN AND FABRICATION OF A LABORATORY-SCALE
FLUE GAS TRAIN
GENERAL CONSIDERATIONS
For the investigation of reactions of flue gas conditioning
agents, a laboratory bench-scale facility was constructed to simu-
late the pertinent components of the flue gas train in a full-
scale coal-burning power plant. Hardware components of the labo-
ratory facility represented components of a full-scale flue gas
train extending from a point upstream from the economizer to the
outlet of the stack, ranging from a maximum temperature of 650 °C
to a minimum of 90 °C. The gas mixture flowing through the train
simulated the composition of flue gas produced from the combustion
of coal. The sizes of the hardware components of the laboratory
train and the flow rate of simulated flue gas were selected to
give residence times of the gas at various locations in the train
that approximated those in a full-scale power plant. Originally,
the plans were to suspend fly ash in the gas stream to simulate
the particulate produced from the combustion of coal. However,
the anticipation of technical difficulties associated with main-
taining fly ash in suspension and interpreting the results of
reaction studies in a heterogeneous system led to abandonment of
efforts to include fly ash in the system.
Flue gas conditioning agents may be injected into the gas
stream in a full-scale plant at various points upstream from the
electrostatic precipitator. At the time the design of the labora-
tory train was being developed, it was assumed that 650 °C—
representative of a point upstream from the economizer—would be
the highest temperature required for injection. Later, several
patents appeared, * 2~lf* indicating that temperatures up to 900 °C
may sometimes be used in full-scale practice. The information
contained in these patents, however, became available after the
laboratory facility with a maximum temperature of 650 °C had been
fabricated and placed in use.
Flue gas undergoes varying degrees of cooling in traversing
the components of a flue gas train leading ultimately to the out-
let of the stack. These components include mainly the economizer
and the air preheater. Some degree of cooling may occur also in
16
-------
the stack; for the purpose of this investigation, a final tempera-
ture of 90 °C at the outlet of the stack was assumed to be appli-
cable.
Figure 1 shows the principal components of the laboratory
train in a schematic diagram. These components are itemized and
described briefly as follows:
• An electric heater for raising the temperature of
the synthetic flue gas to 650 °C.
• A heated quartz cylinder maintaining a portion of
the gas stream at 650 °C, a representative gas tem-
perature in the duct upstream from the economizer.
• A Pyrex heat exchanger representing the economizer.
• A heated Pyrex cylinder maintaining a portion of
the gas stream at about 370 °C, a representative gas
temperature in the duct between the economizer and
the air preheater.
• A Pyrex heat exchanger representing the air preheater.
• A heated Pyrex cylinder maintaining a portion of the
gas stream at about 160 °C, representing the temper-
ature in the duct between the air preheater and the
electrostatic precipitator.
• A small wire-and-pipe electrostatic precipitator
(ESP) maintained at a temperature of about 160 °C.
• A Pyrex heat exchanger simulating cooling near the
stack exit.
• A heated Pyrex cylinder maintaining a portion of
the gas stream at about 90 °C, a representative tem-
perature of the gas stream near the top of the stack.
The dimensions of the cylinders were chosen to provide gas
residence times of 2 s within each cylinder except the heat
exchangers and the ESP. The gas residence time within each of the
heat exchangers was a fraction of a second. The gas residence
time within the ESP was about 7 s. The gas residence times are
based on the flow rate of the simulated flue gas that was normally
used, about 35 L/min as calculated for 25 °C.
The major components of the synthetic flue gas were nitro-
gen, oxygen, carbon dioxide, and water vapor. Approximate concen-
trations of these components were as follows:
17
-------
HEAT
EXCHANGER
HEAT
EXCHANGER
(AIR
PREHEATER)
370 °C
ZONE
HEAT
EXCHANGER
(ECONOMIZER)
S02 NO2 NO
IN IN
N2 N2
650 °C ZONE
ELECTRIC
HEATER
N2
AIR
CO,
Figure 1. Schematic diagram of flue gas train.
18
-------
Component Concn, % v/v
Nitrogen 76
Carbon dioxide 12
Water vapor 8
Oxygen 4
In addition, low concentrations of sulfur dioxide, nitric oxide,
and nitrogen dioxide were added individually or in combinations.
The approximate concentration of sulfur dioxide was 600 ppm v/v,*
as produced from a low sulfur coal. The typical concentrations
of the nitrogen oxides were 1000 ppm for nitric oxide and 100 ppm
for nitrogen dioxide. In some experiments, up to 30 ppm of sulfur
trioxide (sulfuric acid vapor) was included in the flue gas compo-
sition. The purpose of these experiments was either to study the
behavior of sulfur trioxide as a conditioning agent or to study
the reactions of sulfur trioxide with other conditioning agents.
At a concentration of 30 ppm, however, sulfur trioxide was well
above the level to be expected in flue gas from a low sulfur coal.
DESIGN OF THE FLUE GAS TRAIN
Heater for the Flue Gas Mixture
The synthesized gas mixture was electrically heated to
650 °C before it entered the 650 °C constant temperature zone by
means of a small, efficient GTE-Sylvania electrical heater
(Sylvania Emissive Products, Exeter, New Hampshire). The GTE-
Sylvania fluid heater consisted of an approximately 20-cm length
of a patented ferrous alloy heating element in a serpentine con-
figuration installed in a stainless steel pipe. The heating
element came in direct contact with the gas mixture and heated it
quickly and efficiently to 650 °C. The fluid heater was powered
by a West 20-A silicon controlled rectifier power supply and con-
trolled with a West analog set point, solid-state temperature
controller (West Instruments Division of Gulton Industries, Inc.,
East Greenwich, Rhode Island). A stainless steel ball-joint
socket was connected to the outlet end of the heater by means of
a pipe fitting to enable the heater to be connected to the 650 °C
constant temperature zone.
* All flue gas concentrations given subsequently in this report
in parts per million or in parts per billion are on the volume
basis (v/v), unless otherwise specified.
19
-------
Constant Temperature Zones*
The four constant temperature zones designed for the appara-
tus were intended to maintain temperatures of about 650, 370, 160,
and 90 °C. These constant temperature zones corresponded to the
duct upstream from the economizer, the duct between the economizer
and the air preheater, the duct leading to the electrostatic pre-
cipitator plus the precipitator itself, and a portion of the stack
near the outlet, respectively.
With the exception of the 650 °C zone and the electrostatic
precipitator, the containment vessels of the flue gas at the dif-
ferent temperatures were constructed of interconnected Pyrex glass
cylinders. The containment vessel for the gas at 650 °C was ini-
tially constructed of borosilicate glass as well, but it was found
to devitrify and break into pieces at the intermittent tempera-
tures needed to maintain the gas stream at 650 °C. Thus, recon-
struction of this vessel from quartz proved to be necessary. The
electrostatic precipitator was constructed of stainless steel, as
described later in this report.
The significant dimensions of the constant temperature zones
are listed in Table 1. This table includes the inside diameters
of the quartz and glass cylinders used as containment vessels for
the constant temperature zones, the distance between the injection
and sampling ports of each zone, the calculated residence time of
the gas stream at the indicated temperatures at a total gas flow
of 35 L/min (expressed for 25 °C) in each zone, and the inside
diameter of the sampling tube in each reaction zone. Thermocouple
wells were made of either quartz (in the 650 °C zone) or Pyrex;
the thermocouple wells extended into the center of each reaction
zone and were connected to the reaction zones near each injection
or sampling port to measure the temperature in the center of the
gas stream near the ports. Each injection port was a side arm
ending in a ground-glass spherical ball-joint socket (S 35/20).
Each sampling port was a quartz or glass tube with a stopcock
valve (except for the 650 °C zone) that extended to the center of
the gas stream and then turned 90° to face the flow of gas. The
diameter of each sampling tube was chosen to provide isokinetic
sampling of the gas stream at a flow rate equivalent to 1 L/min
at 25 °C.
To compensate for the loss of heat as the gas moved through
the train, the 650 °C quartz constant temperature zone was wrapped
with two 3.66-m long beaded heaters (Nichrome resistance wire
insulated with ball-and-socket ceramic beads), and the Pyrex
cylinders were wrapped with electrical heating tapes. The quartz
* In later sections of this report, the constant temperature
zones are frequently referred to as "reaction zones."
20
-------
TABLE 1. DIMENSIONS OF CONSTANT TEMPERATURE ZONES
Temperature
of zone,
°C
650
370
160 ,
160 (ESP)T
90
Inside
diameter,
cm
7.0
7.5
8.5
15.2
7.6
Approximate
residence
Length,* time,t
cm s
91.4
61.0
30.5
33.0
30.5
2
2
2
7
2
Inside
diameter of
sampling tube,
cm
1.3
1.4
1.6
2.4
1.4
* This is the length between the injection port and the sam-
pling port.
t The calculated residence time is based on a total gas flow
of 35 L/min expressed for 25 °C.
I The electrostatic precipitator was maintained at a constant
temperature around 160 °C.
and Pyrex cylinders were then insulated with several layers of
Fiberfrax sheet, a fibrous ceramic material. Despite the insula-
tion, however, the maximum power output of the beaded heaters and
the heating tapes were chosen to be high enough to compensate for
the heat loss estimated if the quartz and Pyrex cylinders were
not insulated.
A detailed drawing of the quartz cylinder used to maintain
a gas temperature of 650 °C is given in Figure 2. This illustra-
tion is also applicable to the cylinders used to maintain lower
temperatures except for the difference in dimensions given in
Table 1 and for the insertion of a stopcock valve in the sampling
ports of the Pyrex cylinders.
Electrostatic Precipitator
A small, single-stage wire-and-pipe electrostatic precipita-
tor was designed for insertion into the flue gas train just down-
stream from the Pyrex cylinder used for the 160 °C constant
temperature zone. The flue gas in the precipitator was also main-
tained near 160 °C, a temperature typical of the flue gas tempera-
ture in a "cold-side" electrostatic precipitator.
The electrostatic precipitator was housed in a Pyrex cylin-
der wrapped with heating tapes and insulation. The precipitator
consisted of a 0.02-cm negative corona discharge wire mounted
along the axis of a stainless steel cylinder that served as the
21
-------
Figure 2. Detailed illustration of the quartz 650 <>C
constant temperature zone.
22
-------
positive collection electrode. The cylinder had an inside diam-
eter of 13.6 cm and was 33.0 cm long. The area of the inside
surface of the stainless steel cylinder was thus 1410 cm2.
The dimensions of the precipitator were chosen to provide a
residence time of the flue gas inside the precipitator of 6 to
7 s at a volume flow rate of 51 L/min at 160 °C (equivalent to
35 L/min at 25 °C) and, at the same time, to provide a high col-
lection efficiency at a reasonable operating power. From the
Deutsch equation,22 the efficiency of collection and the "slip"
were computed for a range of migration velocities of 1 to 7 cm/s.
The power densities required to produce this range of velocities
were then estimated from an empirically derived linear relation-
ship between these two variables for the collection of fly ash.22
From the surface area of the collection electrode of the wire-and-
pipe precipitator, the corona power was then estimated. These
quantities are presented in Table 2. The values predict that a
collection efficiency of about 99% would require a corona power
of 270 mW.
TABLE 2. COLLECTION EFFICIENCY AND CORONA POWER
FOR VARIOUS MIGRATION VELOCITIES
Migration
velocity,
cm/s
1
2
3
4
5
6
7
Collection
efficiency,
%*
81
96.4
99.32
99.87
99.975
99.9954
99.99912
"Slip",
%t
19
3.6
0.68
0.13
0.025
0.0046
0.00088
Power
density,
mW/cm2
0.06
0.13
0.19
0.25
0.31
0.38
0.44
Corona
power ,
mW
90
180
270
360
440
530
620
* The collection efficiency, C. E., is given accord--
ing to the Deutsch equation as follows:
C.E. = 100 (1 - exp(-WA/V))
where W = effective migration velocity
A = collection electrode surface area
V = volume flow rate
t "Slip" = 100 - C.E.
22. Oglesby, S., and G. B. Nichols. A Manual of Electrostatic
Precipitator Technology, Part 1—Fundamentals. PB-196 380,
National Air Pollution Control Administration, Cincinnati,
Ohio, 1970. 322 pp.
23
-------
The corona power requirement was also estimated by another
approach. For a coal-burning power plant precipitator, the ratio
of power to gas flow rate corresponding to a collection efficiency
of 99% was estimated to be about 0.30 mW/(cm3/s).22 The power
required for a total flow rate of 51 L/min was therefore estimated
to be about 250 mW, a value in good agreement with the first esti-
mate. It thus appears that the laboratory-scale precipitator
would be capable of efficient collection of fly ash when operating
at relatively low power—for example, at an applied voltage of
10 kV and a total current of 25 yA or a current density of
18 nA/cm2.
The calculated value of the specific collecting area—SCA
(the ratio of collection area to volume flow rate)—was
166 m2/(m3/s) or 842 ft2/(1000 ft3/min) at the usual value of the
flow rate. The SCA value would be characteristic of a full-size
unit operating with very high efficiency of dust collection.
Heat Exchangers
Heat exchangers were required to make three reductions in
flue gas temperature: from 650 to 370 °C, from 370 to 160 °C,
and from 160 to 90 °C. These devices corresponded to the econo-
mizer, the air preheater, and the stack, respectively.
A simple but effective heat exchanger was designed to cool
the temperature of the flue gas stream rapidly from 650 to 370 °C.
This simulated economizer consisted simply of two 30 cm lengths
of borosilicate glass tubing of 2 cm i.d. that were connected to
each other and to the containment vessels of the 650 and 370 °C
constant temperature zones by means of ball-and-socket joints.
The tube connected to the 650 °C zone was loosely wrapped with a
thin layer of insulation, while the tube connected to the 370 °C
zone was wrapped with heating tape and thinly insulated. Heat
from the 650 °C gas stream was lost to the ambient air surrounding
the tubes by free convection, and excessive heat loss was compen-
sated for by electrical heating (with the heating tape).
The design of a heat exchanger to cool the flue gas stream
from 370 to 160 °C rapidly and efficiently (and thus simulate the
action of the air preheater in a coal-burning power plant) was
based on the design of a similar heat exchanger used in a previous
project for an industrial sponsor. The heat exchanger consisted
of a modified, commercially available Allihn condenser that was
500 mm long with an outside diameter of 40 mm. The water jacket
of this condenser was filled with glycerol, and the entire con-
denser was tightly wrapped with a coil of 0.32-cm o.d. copper tub-
ing. An unmetered flow of cold tap water through the coiled
copper tubing counter-current to the direction of flow of the gas
stream was used to dissipate heat from the glycerol. The glycerol
was heated by the gas stream and served as a uniform heat transfer
24
-------
medium, minimizing the possibility of overcooling the gas stream
or the formation of "cold spots" on the inside walls of the heat
exchanger.
The heat exchanger designed for lowering the temperature of
the gas stream from 160 to 90 °C was a 30 cm length of uninsulated
process pipe (thick-walled glass tubing) of 7.6 cm i.d. A 90°
short-radius elbow process pipe connected the vertically aligned
electrostatic precipitator to the horizontally aligned 90 °C con-
stant temperature zone. The pipe was wrapped with a small heating
tape so that excessive gas cooling could be compensated for by
electrical heating.
Gas Mixing and Delivery System
Nitrogen and carbon dioxide from regulated, compressed gas
cylinders and regulated, charcoal-filtered compressed house air
were combined in a stainless steel Swagelok union-cross compres-
sion tube fitting (Crawford Fitting Company, Cleveland, Ohio).
The union-cross fitting was connected to the inlet of the gas
heater through a short length of 0.16-cm i.d. stainless steel tub-
ing by means of a compression fitting. The gases were individu-
ally metered through Matheson flowmeters (Matheson Gas Products,
Morrow, Georgia). The inlet pressure of the flue gas train was
measured just downstream from the point of mixing of the nitrogen,
air, and carbon dioxide cylinder gases with a Magnehelic differen-
tial pressure gauge (Dwyer Instruments, Inc., Michigan City,
Indiana). The combination of nitrogen, carbon dioxide, and air
was then passed through the heater previously described and raised
to a temperature of about 650 °C.
Distilled, deionized water was fed by gravity through a
Gilmont flowmeter (Roger Gilmont Instruments, Inc., Great Neck,
New York) and through a stainless steel "tee" joint, at which
point the liquid water entered the hot mixture of gases emerging
from the heater. Water vapor was thus added to the flue gas
stream by flash evaporation of liquid water.
Sulfur dioxide, nitrogen dioxide, and nitric oxide cylinder
gases were added to the gas mixture, when needed, through a union-
cross tube fitting and a "tee" joint into the hot, humidified gas
mixture just downstream from the point of injection of water.
The sulfur dioxide flow rate was measured with a Matheson flow-
meter, and the nitrogen dioxide and nitric oxide flow rates were
measured with Gilmont flowmeters. Teflon tubing was used to
transfer the sulfur dioxide and water. Nylon and stainless steel
tubing were used for the other gases.
In instances when sulfur trioxide was added to the gas
stream, the addition was made into the 370 °C constant temperature
zone. The procedure used for this addition is described later in
this report.
25
-------
A schematic presentation of the gas mixing and delivery
system is included in Figure 1, which appeared earlier in this
report. A tabulation of representative flow rates of the individ-
ual gases and of liquid water is presented in Table 3.
TABLE 3. REPRESENTATIVE FLOW RATES OF THE INDIVIDUAL GASES
IN THE SIMULATED FLUE GAS TRAIN
Metered
component
N2 (g)
Air(g)
C02 (g)
S02-N2 (g)
N0(g)
N02-N2 (g)
H20(l)
Corrected
flow rate,*
L/min
18.76
6.64
4.01
1.35
0.04
0.34
2.05 x 10~3
Component
in gas
mixture
N2
02
C02
S02
NO
N02
H20
Flow
rate,t
L/min
25.65
1.39
4.05
0.02
0.04
0.003
2.79
Concn
75.6%
4.1%
11.9%
595 ppm
1107 ppm
100 ppm
8.2%
* Measured flow rate corrected for gas density and pres-
sure and expressed as liters per minute at 25 °C and
760 mmHg pressure.
t Flow rate expressed in units of liters per minute at
25 °C and 760 mmHg pressure.
Electrical Control and Measurement Devices
Temperature Control and Measurement—
Chromel-alumel thermocouples were inserted into the quartz
or Pyrex wells located near the inlet and outlet of each constant
temperature zone to monitor the temperature of the gas stream at
these points. Each set of thermocouple wires was connected to
one of the binding posts of a mechanical selector switch, which
in turn was connected to the input jack of one channel of a
Hewlett-Packard Model 7128-A Dual Channel Recorder (Hewlett-
Packard Company, San Diego, California). A thermocouple inserted
into a stainless steel thermocouple well mounted at the outlet of
the gas heater was connected to the other channel of the Hewlett-
Packard recorder to monitor the temperature of the heated gas
mixture. The thermoelectric output of each thermocouple was read
directly from the recorder.
The temperature at the inlet of the 650 °C constant tempera-
ture zone was manually controlled by regulating the voltage
applied to the appropriate beaded heater with a variable auto-
transformer. The voltage applied to the beaded heater that
26
-------
controlled the temperature near the outlet of the 650 °C zone was
manually controlled with a solid-state voltage controller (Cole-
Parmer Instrument Company, Chicago, Illinois).
The voltages applied to the heating tapes around the heat
exchangers simulating the economizer and the gas stack were also
manually controlled with solid-state voltage controllers. The
temperatures at the inlet and outlet of the 370 °C zone, the inlet
of the 160 °C zone, the inlet and outlet of the electrostatic
precipitator, and the outlet of the 90 °C zone were controlled by
means of temperature controllers with thermistor sensors (Therma-
logic Division of Dytron, Inc., Waltham, Massachusetts), which
automatically adjusted the voltage from a solid-state voltage
controller applied to the appropriate heating tape to maintain
the required temperature.
Electrostatic Precipitator Electronics—
The electronic equipment used in conjunction with the elec-
trostatic precipitator consisted mainly of a commercially avail-
able power supply unit and electrical measurement devices. A
Peschel Model H20-Y Standard Hipot Tester (Peschel Instruments,
Inc., Cape Coral, Florida) was used to supply the negative poten-
tial to the corona discharge wire, and a Hipotronics Model 100
High Voltage Meter (Hipotronics, Inc., Brewster, New York) was
used to monitor the applied voltage. The plate current was
measured with a Keithley 610C Electrometer (Keithley Instruments,
Inc., Cleveland, Ohio). The electrometer was protected from cur-
rent surges by means of a simple "homemade" spark protector
inserted in the measuring circuit between the collecting electode
and the electrometer.
FABRICATION OF THE FLUE GAS TRAIN
Quartz and Glass Components
The quartz and glass components of the flue gas train,
including the cylinders for the constant temperature zones, the
heat exchangers, and the ESP housing, were fabricated by Mr. M. B.
Watson of the University of Alabama in Tuscaloosa, Alabama. They
were then wrapped with the required heating tapes and insulation
and assembled by Institute personnel.
Physical Configuration
The physical configuration of the flue gas train was dic-
tated by the dimensions of the major components of the gas train
(particularly the dimensions of the constant temperature zones)
and the size of the laboratory in which the apparatus was
installed. The ideal configuration would have been a linear,
vertical alignment of all of the constant temperature zones. The
27
-------
ceiling of the laboratory was not high enough, however, to accom-
modate this configuration. The major components of the flue gas
train were thus aligned as follows:
• The gas heater, the 650 °C constant temperature zone,
and the heat exchanger simulating the economizer were
all aligned horizontally just above the floor of the
laboratory.
• The 370 °C constant temperature zone, the 160 °C con-
stant temperature zone, the heat exchanger simulating
the air preheater, and the electrostatic precipitator
were all aligned vertically from near the floor of
the laboratory to near the ceiling.
• The 90 °C constant temperature zone was aligned hori-
zontally near the ceiling of the laboratory.
• The heat exchanger between the electrostatic precipi-
tator and the 90 °C zone was bent at a 90° angle to
join the vertical portion of the gas train to the
upper horizontal portion.
• A short 2-cm i.d. borosilicate glass tube with ball-
and-socket joints at each end formed a 90° elbow to
connect the simulated economizer to the bottom of the
370 °C zone, thus joining the lower horizontal por-
tion of the flue gas train to the vertical portion.
Support Frame
A steel frame for mounting the components of the flue gas
train was fabricated by Gregory-Salisbury Metal Products, Inc.,
of Birmingham, Alabama, and shipped to the laboratory in three
parts: a section (3.0 m x 0.5 m x 0.5 m) to be mounted horizon-
tally on the floor, a section (2.4 m x 0.5 m x .0. 5 m) to be
mounted horizontally on the ceiling, and a section (3.0 m x 0.5 m
x 0.5 m) to be mounted vertically between the floor and the ceil-
ing. The sections of the frame were bolted to the laboratory
floor and ceiling. Shelves were attached to the frame to accommo-
date sampling from the flue gas train. A plywood panel (1.2 m x
1.8 m) was also attached to provide a support for the temperature
controllers, voltage regulators, flowmeters, valves, and recorder.
A phtograph of the completely assembled apparatus is repro-
duced in Figure 3.
28
-------
Figure 3. Photograph of assembled laboratory-scale flue gas train.
29
-------
SECTION 5
INVESTIGATION OF INDIVIDUAL CONDITIONING AGENTS
IN THE FLUE GAS TRAIN
SULFUR TRIOXIDE
Sulfur trioxide was selected as the first agent to be inves-
tigated. The main objective in the s;tudy of this agent was to
determine how efficiently the agent could be recovered from the
flue gas train at different sampling temperatures.
Sulfur trioxide was injected into the flue gas train at the
inlet of the 370 °C zone and sampled at the outlet of this zone
and other zones of lower temperatures. No reaction of sulfur
trioxide other than that with water vapor to produce sulfuric
acid was expected.
S03(g) + H20(g) --» HaSOi, (g) (1)
Sulfur trioxide and sulfuric acid were indistinguishable by the
analytical methods used; thus, no detectable change in the chemi-
cal state of the injected conditioning agent was anticipated.
However, sulfur trioxide or sulfuric acid was expected to undergo
wall losses as the result of adsorption or condensation. Knowl-
edge of the extent of wall losses was expected to be of value in
interpreting the results of experiments with other conditioning
agents, such as ammonium sulfate, that thermally decompose to pro-
duce sulfur trioxide at high temperatures.
Injection Procedure
To add sulfur trioxide to the flue gas train, dry nitrogen
was flushed at a controlled rate over the surface of oleum (a
solution of sulfur trioxide in sulfuric acid) in a round bottom
flask. The temperature of the oleum was maintained at 33 °C by
placing the flask in a Styrofoam box containing a 60-W light bulb
as a heater; the light bulb was activated as necessary by means
of a temperature-sensitive switch. The sulfur trioxide vapor
entrained in the nitrogen was passed through a Teflon tube to a
glass probe extending into the center of the 370 °C zone of the
30
-------
flue gas train near the inlet of this zone. A side arm from the
connecting tubing permitted sampling to determine the output con-
centration of sulfur trioxide from the oleum source.
The concentration of sulfur trioxide in the oleum was
20% w/w. The vapor pressure of the sulfur trioxide over the
source was calculated to be 0.9 mmHg, which corresponds to a con-
centration of about 1200 ppm.23 Observed concentrations in the
effluent stream, however, were below this equilibrium value.
Determinations of the concentration of sulfur trioxide
evolved from the vapor generator were made by passing the effluent
stream through a bubbler containing 80% w/w isopropanol in water
and then through a glass fiber filter, which retained sulfuric
acid mist formed in the bubbler. The bubbler solution and a fil-
ter wash were titrated with 0.1 N sodium hydroxide with bromphenol
blue as the indicator. Observed sulfur trioxide concentrations
in the injection line with different rates of nitrogen flow over
the oleum were as follows: about 600 ppm at a flow rate of
0.6 L/min or about 500 ppm at a flow rate of 1.0 L/min. Reason-
ably constant sulfur trioxide concentrations were observed for
about 20 h of operating time; after then, the output began to
decrease, and the generator was recharged.
Analytical Procedures
The flue gas stream was analyzed for sulfur oxides by two
procedures. The first procedure gave the total concentration of
sulfur oxides, whereas the second procedure gave the concentra-
tions of sulfur trioxide and sulfur dioxide individually.
The first procedure involved sampling the flue gas at the
rate of 1 L/min through a bubbler containing 3% w/w hydrogen
peroxide in water. The glass sampling line was maintained above
150 °C to avoid loss of sulfur trioxide as condensed sulfuric
acid. The acid collected in the peroxide solution was titrated
with 0.1 N sodium hydroxide, and the concentration of sulfur
oxides in the gas stream was calculated from the observed amount
of acid titrated and the volume of flue gas sampled.
The second procedure was based on collecting sulfur tri-
oxide from a sample stream in a condenser maintained at a temper-
ature between 60 and 90 °C and then passing the sample stream
through a bubbler containing 3% w/w hydrogen peroxide in water to
collect the sulfur dioxide as sulfuric acid. The controlled-
23. Washburn, E. W., Ed. International Critical Tables of Numer-
ical Data, Physics, Chemistry, and Technology. McGraw-Hill
Book Company, Inc., New York, New York, 1928. Volume III,
p.304.
31
-------
temperature condensation of sulfur trioxide as sulfuric acid was
based on the collection principle usually ascribed to Goks^yr and
Ross;21* however, in our work the condenser was based on a design
developed by Maddalone.25 Sulfuric acid from the condenser was
titrated with dilute barium perchlorate using Thorin as an indi-
cator. l8 Sulfuric acid in the peroxide bubbler was titrated with
0.1 N sodium hydroxide with bromphenol blue as the indicator.
In some of the sampling experiments, particularly at the
lower flue gas temperatures, a particulate filter was incorporated
in the sampling train ahead of the condenser. The purpose was to
remove any sulfuric acid present in the gas stream as the con-
densed acid. A Teflon filter with a 2- to 5-ym pore size was
employed for this purpose in a heated Pyrex filter holder. Simi-
lar fibrous fluorocarbon filters have reported collection effi-
ciencies of about 90% for particles as small as 0.03 ym,26 and
the Teflon filter was thus expected to be adequate for the removal
of sulfuric acid mist. Sulfuric acid was removed from the partic-
ulate filter by drawing 80% w/w isopropanol through the filter
under vacuum; the acid was then titrated by the barium
perchlorate-Thorin method.
Experimental Results
Baseline Studies without Sulfur Trioxide Injection—
The first determinations of sulfur oxides in the flue gas
were made without the injection of sulfur trioxide. The purpose
of these experiments was to determine how closely the sulfur
dioxide concentration agreed with the level anticipated from the
makeup procedure for the flue gas and to determine how much oxida-
tion of sulfur dioxide to sulfur trioxide occurred in the flue gas
train. In all of the experiments, the anticipated concentration
of sulfur dioxide was 590 ppm, assuming a negligible degree of
oxidation to the trioxide.
24. Goksjrfyr, H. , and K. Ross. The Determination of Sulfur Tri-
oxide in Flue Gases. J. Inst. Fuel, 35:177-179, 1962.
25. Maddalone, R. L. Guidelines for Combustion Source Sulfuric
Acid Emission Measurements. TRW Document No. 28055-6005-
RU-00 (prepared under U. S. Environmental Protection Agency
Contract No. 68-02-2165). TRW Defense and Space Systems
Group, Redondo Beach, California, 1977. 14 pp.
26. Liu, B. Y. H., and K. W. Lee. Efficiency of Membrane and
Nuclepore Filters for Submicrometer Aerosols. Environ. Sci.
Technol., 10:345-350, 1976.
32
-------
In the initial series of experiments, the total concentra-
tion of sulfur oxides was determined. The average result obtained
by sampling at various points in the flue gas train was 630 ppm
with a relative standard deviation of 2%. This average result
was about 7% higher than the anticipated value.
In the second series of experiments, sulfur dioxide and
sulfur trioxide were collected independently. The results of
these experiments are listed in Table 4. The results for sulfur
dioxide were again higher than the expected value. The results
for sulfur trioxide ranged from values below the detection limit
(i.e., <0.1 ppm) to a value as high as 2.0 ppm. It seems reason-
able to conclude, however, that the representative background
level of sulfur trioxide was less than 1 ppm.
TABLE 4. DETERMINATION OF SULFUR OXIDES
WITHOUT INJECTED SULFUR TRIOXIDE
Sampling temp, °C
370
160
160
(ESP)
90
SO 2 ,* ppm
654
615
642
672
637
664
654?
609
598t
626
616
SO 3 , t ppm
1.6
0.1
0.3
<0.1
0.2
<0.1
2. Of
0.1
<0.1t
0.7
0.1
* The average result was 635 ppm; the expected
value was 590 ppm.
t Determined without a particulate filter.
t These results were obtained with the ESP
energized: applied voltage (V), 7.5 kV;
current density (j), ca. 30 nA/cm2.
Baseline Studies with Sulfur Trioxide Injection—
A substantial amount of experimentation was required before
the observed concentrations of sulfur trioxide in the flue gas
train approximated the levels predicted from the injection rate.
One result of this experimentation was the finding that silicone
33
-------
or Apiezon grease could not be used as a lubricant for the stop-
cocks in the sampling ports; instead of these lubricants, a light
coating of colloidal Teflon was found to be suitable. Another
result was the finding that additional insulation and heating
tapes around the sampling ports were required. Still another
result was the finding that a period of conditioning of the
system—up to 1 h—was required before the expected sulfur tri-
oxide concentrations could be obtained.
After the need for certain precautions as described above
was recognized, reasonably satisfactory results for the concentra-
tion of sulfur trioxide were obtained. The results of two series
of experiments are presented in Tables 5 and 6. Table 5 gives
the results at two sampling locations as determined without a
filter for the removal of sulfuric acid particulate. Table 6
gives the results at five sampling locations, sometimes as deter-
mined with the particulate filter. The average of the absolute
differences between observed and expected sulfur trioxide concen-
trations in these two tables is 2 ppm.
TABLE 5. DETERMINATION OF SULFUR OXIDES
WITH INJECTED SULFUR TRIOXIDE
Sampling
temp, °C
370
160
(ESP)
Observed SOa ,*
ppm
668
680
686
726
753
765
614
619
623
637
673
673
702
717
Observed S03 ,t
ppm
4
6
7
16
8
7
5
10
7
5
5
6
9
9
Expected S03,
ppm
9
11
11
12
5
5
6
12
7
5
6
7
12
10
* The average value was 681 ppm; the expected value was
about 600 ppm.
t Determined without a particulate filter.
34
-------
TABLE 6. DETERMINATION OF INJECTED SULFUR TRIOXIDE
Sampling Observed SOa, Expected 80s,
temp, ° C ppm ppm
370
160
160
(ESP)
90
9
9
5
6
6
7
10
11
11
11
11
12
8*
10*
11*
15*
17*
lit
lit
12t
14t
15
16
8
8
8
8*
8
8
6
5
6
6
9
10
10
9
9
13
13
16
9
18
16
16
10
13
17
16
16
14
14
10
14
* In five experiments with sampling at 160 °C,
about 2 ppm of the total sulfur trioxide
concentration was found on the particulate
filter. In one experiment with sampling at
90 °C, virtually all of the sulfur trioxide
was found in the particulate filter.
t These results were obtained with the ESP
energized: applied voltage (V), 7.5 kV;
current density (j), 28 nA/cm2.
35
-------
Special attention is directed to the results in Table 6
that were obtained with a particulate filter in the sampling line
or with the ESP energized. A relatively small fraction of the
sulfur trioxide was found on the filter at a sampling temperature
of 160 °C, but virtually all of the substance was found on the
filter at a sampling temperature of 90 °C. These results are as
expected from the acid dew point data; that is, 160 °C was above
the dew point, whereas 90 °C was below the dew point.27 Signifi-
cant losses of sulfur trioxide at a sampling temperature of 90 °C
are also suggestive of condensation of the acid and deposition of
acid mist on the walls of the gas train. Energization of the ESP
at the operating temperature of 160 °C caused little apparent loss
of the sulfur trioxide. This finding is also consistent with the
predicted absence of sulfuric acid mist at 160 °C.
AMMONIA
Ammonia was injected into the flue gas stream at the inlet
of the 650, 370, or 160 °C constant temperature zone and sampled
at the outlet of the 650, 370, 160, or 90 °C zone, depending upon
the injection temperature. Ammonia has typically been introduced
into power plant flues at temperatures of 150 to 370 °C. But it
was considered important to include injection at 650 °C in this
investigation, since ammonia is likely to be a decomposition pro-
duct of an agent such as ammonium sulfate or ammonium phosphate,
which are sometimes injected into the gas stream at temperatures
around 650 °C.
In the presence of sulfur oxides in the flue gas, ammonia
was expected to react mainly by combination with sulfur trioxide
and water at temperatures below 200 °C to form particulate ammo-
nium sulfate, (NHi»)2SOi», or ammonium bisulfate, NHuHSOi, : 3'5
2NH3(g) + S03(g) + H20(g) —> (NHi,) 2SOi, (s) (2)
NH3(g) + S03(g) + H2O(g) —> NH^HSO., (1 or s) (3)
27. Banchero, J. T., and F. H. Verhoff. Evaluation and Inter-
pretation of the Vapour Pressure Data for Sulphuric Acid
Aqueous Solutions with Application to Flue Gas Dewpoints.
J. Inst. Fuel, 48:76-80, 1975.
36
-------
The corresponding reaction to produce the sulfite or bisulfite was
not expected because of unfavorable thermodynaraic conditions.28
Additionally, there was the possibility that ammonia might ther-
mally decompose to some extent to yield nitrogen and hydrogen or
that it might be oxidized to nitrogen or nitric oxide and water:
2NH3(g) --» N2(g) + 3H2(g)
(4)
4NH3(g) + 302(g) —» 2N2(g) + 6H20(g)
4NH3(g) + 502 (g) --> 4NO(g) + 6H20(g)
(5)
(6)
With nitrogen oxides as well as sulfur oxides in flue gas,
ammonia may react with either or both nitric oxide and nitrogen
dioxide to form nitrogen and water:
4NH3(g) + 6NO(g) —> 5N2(g) + 2H20(g)
8NH3(g) + 6N02(g) —> 7N2(g) + 12H20(g)
(7)
(8)
In addition to those reactions given by Equations 7 and 8, ammonia
can also react with nitrogen dioxide at temperatures below the
melting point of ammonium nitrate (170 °C) to give mainly nitro-
gen, water, and ammonium nitrate, and it can react at higher tem-
peratures to produce mainly nitrogen, nitric oxide, and water.29
Ammonia can react with nitric oxide at 100 °C to give nitrogen and
traces of ammonium nitrate and ammonium nitrite.30 In general,
ammonia is stable in the presence of nitric oxide in the absence
of catalytic surfaces;31 however, in the presence of catalytic
28. St. Clair, H. W. Vapor Pressure and Thermodynamic Properties
of Ammonium Sulphites. In: Report of Investigations 3339.
U. S. Bureau of Mines, Washington, D. C., 1937. pp 19-29.
29. Rosser, W. A., and H. Wise. Gas-Phase Oxidation of Ammonia
by Nitrogen Dioxide. J. Chem. Phys., 25:1078-1079, 1956.
30. Gehlen, H. Reactions and Properties of Nitric Oxide and Its
Compounds. III. The Reaction Between Nitric Oxide and the
Alkali Salts of Nitrodisulfonic Acid. Chem. Ber., B66:292-
297, 1933.
31. Wise, H., and M. F. Freeh. Kinetics of Oxidation of Ammonia
by Nitric Oxide. J. Chem. Phys., 22:1463-1464, 1954.
37
-------
surfaces such as platinum and other transition metals or their
oxides ammonia is known to react readily with nitric oxide at
elevated temperatures, and such catalytic reactions have been
patented for the removal of nitric oxide from industrial stack
gases. 32~31»
Since the inside surfaces of the flue gas train (with the
possible exception of the electrostatic precipitator) were not
expected to catalyze the reactions of ammonia with oxygen or
nitrogen oxides, we were unable to predict with any certainty
whether or not ammonia would react with oxygen or nitrogen oxides,
particularly at 650 °C. Consequently, knowledge of the occurrence
and extent of such reactions could be obtained only through
experimentation.
Injection Procedure
Three types of ammonia generators were used in this study.
Initially, a solution of ammonium hydroxide was injected with a
syringe pump into a stream of hot nitrogen, where the solution was
volatilized and flushed into a sidearm of the flue gas train.
Considerable variation in ammonia concentration was experienced,
and thus other techniques were tried and evaluated. As one tech-
nique, pure anhydrous ammonia was injected into the flue gas with
a syringe pump. Finally, a mixture of ammonia with nitrogen
certified to contain 1.1% v/v of ammonia was metered into the flue
gas at a rate of 100 or 200 mL/min. The injection of the certi-
fied gas mixture was the most convenient, accurate, and reproduci-
ble method and was consequently the method of choice for ammonia
injection.
32. Conn, J., D. R. Steele, and H. C. Anderson. Selective
Removal of Nitrogen Oxides from Oxygen-Containing Gases
Especially from Nitric Acid Production from Ammonia. U. S.
Patent 2 975 025, March 14, 1961. Assigned to Engelhard
Industries, Inc.
33. Anderson, H. C., and C. D. Keith. Removal of Nitric Oxide
from Tail Gas after Oxidation of Ammonia to Nitric Acid.
U. S. Patent 3 008 796, November 14, 1961. Assigned to
Engelhard Industries, Inc.
34. Nonnenmacher, H., and K. Katte. Selective Removal of Oxides
of Nitrogen from Gas Mixtures Containing Oxygen. U. S.
Patent 3 279 884, October 18, 1966. Assigned to Badische-
Anilin- & Soda-Fabrik, AG, Germany.
38
-------
Analytical Procedures
During the investigation of ammonia, the flue gas stream
was sampled and analyzed for ammonia, sulfur trioxide, ammonium
ion, sulfate ion, nitrogen dioxide, and nitric oxide. Not all of
these species were sampled or determined in every experiment.
For example, no attempt was made to determine the nitrogen oxides
as reaction products when these gases were included in the makeup
of the flue gas.
Ammonia was determined by sampling flue gas through two
bubblers in series containing 0.02 N sulfuric acid and by analysis
of the bubbler solutions by one of two methods. In one method,
ammonia levels were determined with an ammonia-specific electrode
(Orion Research Inc., Cambridge, Massachusetts). The other
method, the phenol-hypochlorite colorimetric procedure,17 was pre-
ferred because of its better sensitivity and reproducibility.
The detection limit of the colorimetric method was estimated to
be about 5 x 10~6 M in 100 mL of sample solution, or 0.08 ppm in
a 15-L sample of flue gas.
Sulfur trioxide (or sulfuric acid vapor) was sampled with
the controlled condensation coil and the resulting sulfate deter-
mined, as described earlier in this report.
The collection and determination of ammonium and sulfate
salts formed in the gas stream was carried out in the following
way. To remove the particulate material from the gas stream, a
Teflon filter in a heated Pyrex holder was placed in the sampling
line in a location upstream from the bubblers that were to collect
ammonia. The filter was maintained at about 150 °C except when
samples were taken from the 90 °C zone; here it was maintained at
90 °C. Exposed filters were extracted by vacuum filtration with
either 0.02 N sulfuric acid for the determination of ammonium ion
or distilled water for the determination of both ammonium ion and
sulfate. The acid extract was analyzed for ammonium ion by the
phenol-hypochlorite method. To determine the ammonium ion in the
water extract, dilute sulfuric acid was added to a portion of the
extract, and the colorimetric procedure for ammonia was carried
out. To determine sulfate, isopropyl alcohol was added to a por-
tion of the water extract and the sulfate was determined by the
barium-Thorin titration, as described earlier.
Nitrogen dioxide was absorbed from the gas stream in Greiss-
Saltzman reagent in a bubbler and determined by colorimetry.35
The sum of both nitrogen dioxide and nitric oxide was determined
35. Recommended Method of Analysis for Nitrogen Dioxide Content
of the Atmosphere (Greiss-Saltzman Reaction). In: Methods
of Air Sampling and Analysis, M. Katz, Ed. American Public
Health Association, Washington, D. C., 1977. pp 527-534.
39
-------
by oxidizing the nitric oxide in the gas stream to nitrogen diox-
ide with chromium trioxide impregnated on firebrick and then
determining the total nitrogen dioxide.36
Experimental Results
Injection of Ammonia into Flue Gas Containing No Added Oxides of
Sulfur or Nitrogen—
Ammonia was initially injected into the flue gas mixture
containing no added oxides of sulfur or nitrogen. The purpose of
these experiments was to determine the agreement between the con-
centration of injected ammonia in the gas stream and the concen-
tration anticipated and to determine the extent of thermal decom-
position and oxidation of ammonia that occurred in the flue gas
train.
Ammonia was injected near the inlet of the 650, 370, or
160 °C constant temperature zone. When ammonia was injected at
650 °C, samples were taken near the outlet of the 370 °C zone,
the ESP (160 °C), or the 90 °C zone. When ammonia was injected
at 370 or 160 °C, samples were taken near the outlet of the 90 °C
zone.
The results of the ammonia determinations are presented in
Table 7. This table lists the ammonia concentrations found at
various sampling locations in the flue gas train and compares the
observed concentrations with the values anticipated from the
makeup procedure for the flue gas and the rate of ammonia injec-
tion. In some experiments the total observed ammonia concentra-
tion was composed of the ammonia found on a Teflon filter in the
sampling train and the ammonia found in bubblers of sulfuric acid
solution. The filter catch represented the adsorbed ammonia
recovered from the filter and filter holder inlet. The bubbler
catch represented the ammonia gas trapped in the bubbler, the fil-
ter holder outlet, and the glass connecting lines from the filter
holder to the bubbler.
The tabulated results show that when ammonia was injected
near the inlet of the 650 °C constant temperature zone, the
amounts recovered decreased slightly as the sampling temperature
was lowered and the location of the sampling point was farther
removed from the point of injection. Injection of the ammonia
at 370 °C did not improve the recovery at 90 °C, but injection at
160 °C improved the recovery significantly. This general trend
suggests that, while most of the ammonia (>85% of that injected)
36. Tentative Method of Analysis for Nitric Oxide Content of the
Atmosphere. In: Methods of Air Sampling and Analysis, M.
Katz, Ed. American Public Health Association, Washington,
D. C., 1977. pp 524-526.
40
-------
TABLE 7. DETERMINATION OF AMMONIA IN FLUE GAS
CONTAINING NO SULFUR OR NITROGEN OXIDES
Observed [NH3], ppm Expected
Injection Sampling
temp, °C temp, °C
650 370
160
(ESP)
90
370 90
160 90
Sample
No.
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
1
2
3
4
1
2
1
2
Filter
catch
_*
—
-
-
-
-
-
—
0.5
1.2
1.0
2.2
—
-
—
_
-
0.8
1.0
0.8
0.8
0.9
0.7
Bubbler
catch
29.3
28.8
27.8
26.2
25.7
27.5
27.1
25.5 (
28.2
21.0
17.8
22.4
23.4
33.2
25.5
23.9
26.4
23.3
27.0
26.2
24.7
29.6
28.2
[NH3],
Total Av ppm
29.3
28.8
27.8
26.2 27.2 28.8
25.7
27.5
27.1
25.5
28.7
22.2
18.8
24.6 25.2 28.8
23.4
33.2
25.5
23.9
26.4 25.6 28.8
24.1
28.0
27.0
25.5 26.2 30.6
30.5
28.9 29.7 30.6
* A dash indicates that no filter was included in the sam-
pling train.
passed through the flue unchanged, small amounts were lost by
adsorption on the walls of the train or by reaction at the higher
temperatures.
Since the background concentration of nitrogen was so great
in the flue gas mixture, no attempt was made to analyze the gas
stream for the small additional amounts of nitrogen that may have
been formed from the thermal decomposition of ammonia at 650 °C.
However, the concentration of nitrogen oxides in the flue gas was
determined from gas stream samples recovered from the 90 °C zone
when ammonia was injected at 650 °C. The nitrogen dioxide levels
found ranged from 0.11 to 0.18 ppm. With the nitric oxide
oxidizer in the sampling line, the results ranged from 0.14 to
41
-------
0.25 ppm of nitrogen oxides. Without the injection of ammonia,
the background concentration of nitrogen oxides was 0.01 ppm.
Thus, ammonia was apparently oxidized to nitrogen oxides at 650 °C
to some extent, but the extent of reaction was not large.
Injection of Ammonia into Flue Gas Containing Sulfur Dioxide but
No Added Sulfur Trioxide or Nitrogen Oxides—
To investigate the reactivity of ammonia in the presence of
sulfur dioxide, ammonia was added to flue gas containing about
600 ppm of sulfur dioxide but no added sulfur trioxide or nitro-
gen oxides. Although sulfur trioxide was not added to the flue
gas mixture, the background concentration of this constituent was
found to be 1 to 2 ppm. Ammonia was injected near the inlet of
the 650 to 160 °C constant temperature zone and sampled at
selected points downstream.
The results of the experiments are given in Table 8. The
principal aspect of these results differentiating them from the
results obtained in the absence of sulfur dioxide was the recovery
of a higher fraction of the ammonia on the filter of the sampling
train. The ammonia on the filter possibly may have been present
as the solid sulfate or bisulfate salt resulting from the reaction
of ammonia with the low background concentration of sulfur triox-
ide; some of the experimental information is not consistent with
this hypothesis, however, as discussed in the next paragraph. In
any event, the experimental results give no indication of an
appreciable reaction of ammonia with sulfur dioxide to form the
solid sulfite or bisulfite salt.
The activation of the ESP in one experiment yielded results
that were not consistent with the occurrence of suspended ammo-
nium sulfate or bisulfate in the flue gas train. No space charge
effect* was observed; also, no substantial reduction in the amount
of ammonia on the filter of the sampling train was observed. Con-
ceivably, the reaction of ammonia with sulfur trioxide to produce
a solid product occurred mainly on the filter rather than in the
flue gas train.
Injection of Ammonia into Flue Gas Containing Added Sulfur Dioxide
and Sulfur Trioxide but No Added Nitrogen Oxides—
With both sulfur dioxide (600 ppm) and sulfur trioxide
(approximately 30 ppm) added to the flue gas, experiments were
carried out to investigate the formation of ammonium sulfate or
ammonium bisulfate salt at the lower temperatures in the flue gas
train. To minimize complications involving the formation and
deposition of salts in the gas train, ammonia was injected near
* A space charge effect is indicated by a shift in current density
to higher values at a specified applied voltage. Such an effect
is subsequently illustrated in Figure 4.
42
-------
TABLE 8. DETERMINATION OF AMMONIA IN FLUE GAS
CONTAINING SULFUR DIOXIDE BUT NO ADDED
SULFUR TRIOXIDE* OR NITROGEN OXIDES
Observed [NH3] ,
Injection
temp, °C
650
170-f
Sampling
temp , ° C
160
(ESP)
looJ
Sample
No.
1
2
3t
1
2
3
4
Filter
catch
2.1
2.8
3.1
1.3
1.0
1.8
1.5
Bubbler
catch
9.8
16.8
21.3
28.1
30.7
29.4
31.4
ppm
Total
11.9
19.6
24.4
29.4
31.7
32.2
32.9
Expected
[NH3],
ppm
14.7
29.4
29.4
30.6
30.6
30.6
30.6
* The background concentration of sulfur trioxide was 1 to
2 ppm.
t This result was obtained with the ESP energized: applied
voltage (V), 10 kV; current density (j), 60 to 65 nA/cm2.
T Actual temperatures of the constant temperature zones
nominally held at 160 and 90 °C.
the inlet of the 160 °C zone (maintained at an actual temperature
of 175 °C in these experiments), and sulfur trioxide was intro-
duced near the inlet of the 370 °C zone. Samples were taken for
analysis at the outlets of the 160 and 90 °C zones.
The results of these experiments are presented in Table 9.
Analysis of the gas samples withdrawn at 175 °C revealed that
essentially all of the ammonia was collected on the filter appar-
ently as a salt, which appeared to be ammonium bisulfate since
the ratios of sulfate to ammonium ion were just slightly greater
than the one-to-one ratio expected. Excess sulfur trioxide passed
through the filter, presumably as sulfuric acid vapor.
The results from the samples taken from the 90 °C zone
(actually at 100 °C) disclose the drastic effect that the activa-
tion of the ESP had upon the passage of the ammonium salts out of
the stack. With the ESP not energized, recoveries of ammonia and
sulfur trioxide were high, indicating that ammonium salts and per-
haps some excess sulfuric acid aerosol were leaving the flue gas
train. With the ESP energized, recoveries were low, apparently
because the particulate material was collected by the ESP.
43
-------
TABLE 9. DETERMINATION OF AMMONIA AND SULFATE IN FLUE GAS
CONTAINING ADDED SULFUR DIOXIDE AND SULFUR TRIOXIDE
BUT NO ADDED NITROGEN OXIDES
Observed concentrations,* ppm
Sampling
temp, °C
175t
loot
Sample
No.
1
2
3
1§
2§
3
4
Filter
NH
24.
19.
17.
0.
1.
19.
29.
3
8
8
7
9
5
7
9
catch
S03
26
25
22
0
0
25
27
Bubbler catch Total
NH
1.
0.
0.
0.
0.
0.
0.
3 S03
9
1 2$
4 7T
3
2
3
3
NH
26.
19.
18.
1.
1.
20.
30.
3
7
9
1
2
7
0
2
SO 3
26
27
29
0
0
25
27
Expected
concn,
ppm
NH3
29.2
29.2
29.2
29.2
29.2
29.2
29.2
SO 3
26
27
24
24
24
30
30
* The NH3 and S03 on the filter presumably occurred as NH^"1" and
SOi,
-2
ions.
t Actual temperatures of the constant temperature zones nominally
held at 160 and 90 °C.
T The sulfate found was extracted from the filter holder outlet
and was counted as part of the bubbler catch since it probably
represents sulfuric acid vapor.
§ The ESP was energized: V = 10 kV, j = 41 nA/cm2.
The addition of about 30 ppm of ammonia to flue gas contain-
ing about 26 ppm of sulfur trioxide produced a space charge effect
in the ESP. The current-voltage curves demonstrating the shift
are presented in Figure 4. One interesting aspect of the curves
is that they merge at applied potentials near the point where
sparking occurred.
Injection of Ammonia into Flue Gas Containing Added Sulfur Dioxide
and Nitrogen Oxides--
To investigate the possible reactions of ammonia with nitro-
gen oxides, ammonia was injected into the flue gas mixture contain-
ing 600 ppm of added sulfur dioxide (plus approximately 2 ppm of
background sulfur trioxide), first in the presence and then in the
absence of approximately 1000 ppm of added nitrogen oxides in the
gas stream. The added nitrogen oxides consisted of approximately
100 ppm of nitrogen dioxide and 900 ppm of nitric oxide. Ammonia
was injected near the inlet of the 650 °C constant temperature
zone in all of the experiments at a rate estimated to give an
44
-------
100
90
80
70
CM
J 60
CC
DC
D
U
<
Q.
50
40
30
20
10
SPARKING
CURVE 1
No NH3
CURVE 2
30 ppm. NH3
I I
4 6 8 10
APPLIED POTENTIAL, kV
12
Figure 4. Current-voltage relationship for the ESP reflecting the space charge
effect from the combination of ammonia and sulfur trioxide.
45
-------
ammonia concentration of 63 ppm. The gas stream was sampled from
the outlets of all of the various reaction zones downstream from
the point of injection.
The results of the ammonia determinations in the presence
and absence of added nitrogen oxides to the flue gas stream are
given in Table 10. This table indicates that most of the ammonia
injected was recovered at all of the sampling points when the
nitrogen oxides were not present. The average of all the observed
ammonia concentrations in the absence of added nitrogen oxides
was 64 ppm, in excellent agreement with the anticipated concentra-
tion of 63 ppm. In the presence of added nitrogen oxides, however,
the observed ammonia concentrations at all sampling temperatures
averaged only 36 ppm. This represents a 44% decrease in the aver-
age ammonia recovery observed in the absence of nitrogen oxides.
The observed ammonia concentrations in the absence of added nitro-
gen oxides were approximately the same at all sampling tempera-
tures. Thus, the reaction of ammonia with nitrogen oxides appears
to have occurred predominantly at 650 °C.
TABLE 10. DETERMINATION OF AMMONIA IN FLUE GAS CONTAINING
ADDED SULFUR DIOXIDE AND NITROGEN OXIDES
Observed
Sampling
temp, °C
650
370
160
90
Without added
Individual
experiments
64
63
59
70
70
57
74
75
51
62
56
70
57
NOX
Av
65
69
56
63
[NH3] ,* ppm
With added
Individual
experiments
40
33
40
44
43
25
32
30
36
NOX
Av
37
42
29
36
* The expected [NHs] was 63 ppm.
46
-------
No ammonia determinations were made in the presence of
nitric oxide or nitrogen dioxide alone. And since the major reac-
tion products of ammonia with either nitric oxide or nitrogen
dioxide are the same (nitrogen and water), only speculation can
be offered for the exact nature of the reactions between ammonia
and nitrogen oxides in the flue gas stream.
One other possible cause for the observed effect of nitrogen
oxides on ammonia is interference by nitrogen oxides in the chemi-
cal determination of ammonia. No mention of such an interference
was made in the original reference to the determination of ammonia
by the phenol-hypochlorite method;17 and no mention is made of
interference by ammonia in the standard analytical methods for
the determination of nitrogen oxides.35"37 Nevertheless, a brief
experiment was conducted at the conclusion of the investigation
of ammonia to determine whether or not nitrogen oxides may have
interfered with the determination of ammonia. A series of ammo-
nium chloride standards equivalent to a concentration of ammonia
in the flue gas stream of approximately 55 ppm were analyzed by
the phenol-hypochlorite method. To some of these standards,
various concentrations of sodium nitrite (up to an equivalent con-
centration of NOX in the gas stream of 600 ppm) and sodium nitrate
(equivalent to an NOX concentration in the gas stream of 600 ppm)
were added. No measurable difference in the determined ammonia
concentration was observed whether or not nitrite ion had been
added to the ammonia standard. The observed concentration of
ammonia was, however, approximately 30% greater in the presence
of added nitrate ion than in its absence, but the effect of this
interference was opposite in direction to an interference that
would result in an artificially low observed ammonia concentration.
Thus, chemical reaction appears to be the most probable cause for
the apparent disappearance of ammonia when nitrogen oxides were
added to the flue gas stream.
Summary and Discussion of the Results
The addition of ammonia to flue gas that contained no added
sulfur oxides or nitrogen oxides indicated that the ammonia passed
essentially unchanged through the gas stream. The loss of some
ammonia was attributed to adsorption on the walls of the flue gas
train. The injection of ammonia at 650 °C produced traces of
nitrogen oxides (0.1 to 0.2 ppm) that were detected near the out-
let of the train.
37. Tentative Method of Analysis for Total Nitrogen Oxides as
Nitrate (Phenoldisulfonic Acid Method). In: Methods of Air
Sampling and Analysis, M. Katz, Ed. American Public Health
Association, Washington, D. C., 1977. pp 534-538.
47
-------
The injection of ammonia into flue gas that contained about
600 ppm of sulfur dioxide and about 1 to 2 ppm of sulfur trioxide
but no added nitrogen oxides gave some evidence of a reaction of
the ammonia with sulfur trioxide. However, there was no indica-
tion of a reaction between ammonia and sulfur dioxide, as pre-
dicted from thermodynamic data.28
When sulfur trioxide was added to the flue gas to produce a
concentration of about 30 ppm, all of the ammonia injected (about
30 ppm) appeared to be converted to the bisulfate salt. The
ammonium aerosol was collected by the ESP, and a space charge
effect was observed.
The addition of ammonia to flue gas that contained both
added sulfur dioxide and nitrogen oxides indicated that ammonia
reacted extensively with the added nitrogen oxides at 650 °C,
resulting in the destruction of over 40% of the injected ammonia.
From a strictly thermodynamic point of view, the reactions
of ammonia with nitric oxide and nitrogen dioxide are both highly
favorable energetically.38 Both of the reactions have very large
negative free energy changes, which give extremely large values
for the equilibrium constants of these reactions.* The free
energy of reaction of ammonia with nitric oxide (Equation 7)
ranges from -436.22 kcal at 25 °C to -445.27 kcal at 627 °C, with
corresponding equilibrium constant values ranging from 10320 to
10108 over the same temperature interval. The reaction of ammonia
with nitrogen dioxide (Equation 8) has free energy values ranging
from -697.72 to -792.39 kcal and equilibrium constants ranging
from 10slz to 10192 over the temperature interval 25 to 627 °C.
From a comparison of the magnitudes of these calculated thermo-
dynamic quantities, the reaction of ammonia with nitrogen dioxide
is seen to be more energetically favorable than the reaction of
ammonia with nitric oxide and would be expected to proceed prefer-
entially when in competition with the ammonia-nitric oxide reac-
tion.
38. Stull, D. R., and H. Prophet, Eds. JANAF Thermochemical
Tables, 2nd ed. National Bureau of Standards, Washington,
D. C., 1971.
* Calculated from the standard free energies of formation of the
reactants and products in Equations 7 and 8 that are tabulated
in Reference 38.
48
-------
Kinetically, the rate of disappearance of ammonia in the
reaction of ammonia with nitrogen dioxide has been found to follow
the rate expression:29'39
= k[NH3][N02] (9)
where k = 10 12 • 7e~2 75 ° 0/RT cc/(mol s) and the (10)
constant 27500 is given in cal/mol.
The rate expression given in Equation 9 was roughly evaluated by
use of the specific values of the rate constant calculated from
Equation 10 and the concentration of ammonia and nitrogen dioxide
found experimentally in this investigation.
The results of the calculations indicate that the reaction
of ammonia with nitrogen dioxide can account for only a small
fraction of the ammonia loss observed experimentally. Thus, ammo-
nia may react with nitric oxide as well as with nitrogen dioxide.
A rate expression for the reaction of ammonia with nitric oxide
that could be evaluated with the data obtained under the condi-
tions of this investigation could not be found during a brief
review of the literature. Thus, the extent of reaction of ammonia
with nitric oxide in the flue gas stream could not be estimated.
TRIETHYLAMINE
Triethylamine was injected into the flue gas train near the
inlet of the 650, 370, or 160 °C constant temperature zone. The
gas stream was sampled near the outlet of the ESP (160 °C) or the
90 °C zone.
Triethylamine was expected to undergo several routes of
thermal decomposition or chemical reaction. At high temperatures,
the expected thermal degradation products of triethylamine
included monoethyl- and diethylamines, ammonia, hydrogen cyanide,
and hydrocarbons. These fragments were potentially capable of
interacting or combining with components of the flue gas to pro-
duce such compounds as alkyl cyanides, hydrazines, and nitro-
samines. Another possible reaction expected was the oxidation of
triethylamine to carbon dioxide, water vapor, nitrogen, and nitro-
gen oxides. Finally, there were the possibilities of triethyl-
amine reacting with water in the flue gas mixture to form ethanol
39. Bedford, G., and J. H. Thomas. Reaction Between Ammonia and
Nitrogen Dioxide. J. Chem. Soc., Faraday Trans. 1, 68:
2163-2170, 1972.
49
-------
and primary and secondary amines or producing solid salts from
the combination of triethylamine with sulfur dioxide, sulfur tri-
oxide, and water.
Injection Procedure
Triethylamine vapor was generated by bubbling dry nitrogen
through a reservoir of the liquid amine maintained at 0 °C in an
ice bath. At a nitrogen flow rate of 30 mL/min and the usual flow
rate in the flue gas train, the expected triethylamine concentra-
tion in the flue gas was 27 ppm (based on the vapor pressure of
the compound at 0 °C). The actual concentrations in the flue gas
were calculated by analyzing the generator effluent and correcting
the results for the dilution factor in the flue gas. The results
ranged from 14 to 36 ppm but were reasonably constant during a
given day of experimentation.
Traces of diethylamine were present along with triethylamine
in the generator effluent. Analysis of the liquid triethylamine
showed that diethylamine was present at a concentration of about
0.15% w/w; analysis of the vapor showed that diethylamine was
added to the flue gas at a concentration of about 80 ppb when tri-
ethylamine was added at 21 ppm. The indicated diethylamine con-
centration was 0.27% w/w of the triethylamine concentration in
the vapor state. Thus, the value was higher in the vapor state
than in the liquid state, as expected, because of the higher vola-
tility of diethylamine.
Analytical Procedures
During the investigation of triethylamine, the flue gas was
analyzed for a variety of substances based on the possible reac-
tions of triethylamine in the flue gas train. These substances
included triethylamine, other organic amines, other organic com-
pounds (including N-nitrosodiethylamine), sulfur oxides, nitrogen
oxides, hydrogen cyanide, ammonia, and salts of protonated amines.
All of these substances were not determined in every experiment,
as indicated by the experimental results presented later in this
report.
Organic Amines--
Organic amines were collected from the flue gas in bubblers
containing 0.1 N sulfuric acid solution. An aliquot of an exposed
bubbler solution was made alkaline with sodium hydroxide solution
to regenerate the free amine, and the basic solution was then
analyzed by gas chromatography (GC).
Two GC procedures were used for the determination of organic
amines. In one method, a stainless steel column (4 m by 2 mm i.d.)
packed with 10% Carbowax 20M-5% potassium hydroxide on 60/80-mesh
Chromosorb W was coupled with a flame ionization detector. The
injection port was maintained at 110 °C, the column at 110 °C, and
50
-------
the detector at 250 °C. The helium carrier gas flow rate was
50 mL/min. The detection limit was estimated to be about 1 ppm
triethylamine in a 15-L sample of flue gas. This column, however,
would not separate diethylamine from triethylamine.
In the second GC method, which did allow the separation and
determination of mono-, di-, and triethylamines, "* ° the GC column
was a stainless steel tube (0.6 m by 2 mm i.d.) packed with 60/80-
mesh Chromosorb 103. A glass precolumn (10 cm by 2 mm i.d.) con-
taining 20/30-mesh Ascarite was attached to the analytical column.
The injection port, precolumn, and analytical column were held at
170 °C, and the thermionic detector was maintained at 275 °C.
The nitrogen carrier gas flow rate was 50 mL/min. The detection
limit for triethylamine was estimated to be less than 0.5 ppm in
a 15-L sample; for diethylamine, less than 1 ppb in a 15-L sample.
The detection limit for monoethylamine was not determined.
General Types of Organic Compounds—
In some experiments, a nonspecific sampling and analytical
procedure for organic compounds was followed. Two different sam-
pling devices were employed:
(1) A cartridge containing Amberlite XAD-2 porous
polymer maintained at about 30 °C. The dimen-
sions of the 1-g sorbent bed provided a gas
residence time comparable to that in the organic
module of the SASS sampling train.1*1
(2) A cold trap at 0 °C followed by a trap at -78 °C.
The exposed sorbent material was extracted with methylene chloride
with a Soxhlet extractor, and the extract was concentrated to
about 2 mL with a Kuderna-Danish evaporator. The concentrate was
then analyzed by GC. After flue gas was sampled through the cold
traps, the traps were first rinsed with distilled, deionized water
and then with methylene chloride. The water rinse was extracted
with methylene chloride. The methylene chloride extract and
rinse were then combined and concentrated prior to GC analysis.
In one experiment after the flue gas was sampled, gas samples were
taken from the trap at -78 °C with a gastight syringe and injected
into a GC.
40. Taylor, D. G., Ed. NIOSH Manual of Analytical Methods,
Vol. 1, 2nd ed. National Institute for Occupational Safety
and Health, Cincinnati, Ohio, April, 1977. pp 221-1 to 221-9,
41. Blake, D. Operation and Service Manual, Source Assessment
Sampling System. AEROTHERM Report UM-77-80, ACUREX Corpora-
tion, Mountain View, California, March, 1977.
51
-------
Both liquid and gas samples were analyzed by GC with a glass
column packed with 3% OV-1 on 40/60-mesh Chromosorb W. The column
dimensions were 2 m by 2 mm i.d. The carrier gas was helium with
a flow rate of 75 mL/min. Temperatures of the injection port,
column, and flame ionization detector were all in the range 250
to 260 °C. The detection limit for most organic compounds was
expected to be in the sub-parts-per-million range. A few samples
were also analyzed on the OV-1 column coupled to a thermionic
detector, which increased the sensitivity for compounds contain-
ing nitrogen to the parts-per-billion range.
N-Nitrosodiethylamine—
Although the nonspecific sampling procedures for organic
compounds were applicable to the determination of N-nitrosodi-
ethylamine, a separate sampling as well as a separate analytical
procedure was also used for this compound. Flue gas was sampled
through bubblers containing 1 N sodium hydroxide.1*2 The alkaline
solutions were extracted with methylene chloride and the extracts
were concentrated prior to analysis by GC with thermionic detec-
tion. The GC column was a stainless steel tube packed with 10%
Carbowax 20M-TPA on 60/80-mesh Chromosorb W. The carrier gas was
nitrogen, and its flow rate was 50 mL/min. The injection port
temperature was 230 °C, the column temperature 150 °C, and the
thermionic detector temperature 230 °C. Four-microgram amounts
of N-nitrosodiethylamine added to alkaline bubblers could be deter-
mined before or after exposure of bubblers to flue gas. In a
50-L gas sample, 4 yg of the compound corresponds to about 20 ppb.
The detection limit was estimated to be considerably less.
Other Flue Gas Components—
Other components of the flue gas were determined as follows:
• Sulfur oxides, by methods previously described in
this report.
• Nitrogen oxides, by methods previously described, or
on occasion, the total of nitrogen oxides by the
phenoldisulfonic acid method.37
• Hydrogen cyanide, by collection in a solution of
0.1 N sodium hydroxide and quantification with a
cyanide-specific electrode (Orion Research Inc.,
Cambridge, Massachusetts).
• Ammonia, by the phenol-hypochlorite colorimetric
method.l7
42. Fine, D. H., D. P. Rounbehler, E. Sawicki, and K. Krost.
Determination of Dimethylnitrosamine in Air and Water by
Thermal Energy Analysis: Validation of Analytical Procedures,
Environ. Sci. Technol., 11:577-580, 1977.
52
-------
• Salts formed by reaction of amines with sulfur oxides,
by collection on a Teflon filter, extraction into
water, and quantification of organic ammonium ions by
GC or sulfate by oxidation with hydrogen peroxide and
titration by the barium perchlorate-Thorin method
described earlier.
Experimental Results
Injection of Triethylamine into Flue Gas Containing No Added
Sulfur Oxides or Nitrogen Oxides—
In the absence of reactive oxides of sulfur and nitrogen in
the flue gas, triethylamine was injected into the gas stream near
the inlet of the 650, 370, or 160 °C constant temperature zone.
The gas stream was sampled near the outlet of the ESP (160 °C) or
the 90 °C zone.
When triethylamine was injected at 650 °C, most of it
appeared to decompose. Samples taken from the flue gas stream at
the outlet of the ESP indicated a 2 to 10% recovery of the tri-
ethylamine injected. These results, along with others, are pre-
sented in Table 11.
There was evidence that diethylamine was one product of the
degradation process. In Sample No. 5 taken at an injection tem-
perature of 650 °C, the concentration of diethylamine was found
to be 600 ppb, while that of triethylamine was only 400 ppb. In
Sample No. 8 taken at this injection temperature, small amounts
of an unidentified compound containing nitrogen with a GC reten-
tion time between that of monoethylamine and diethylamine was
found; the level of diethylamine in this sample was 60 ppb. No
other organic compound that could be associated with the decompo-
sition of triethylamine was found when samples were taken with
cold traps or a cartridge of XAD-2 resin.
Some inorganic compounds were also found in the flue gas at
the outlet of the ESP when triethylamine was injected at 650 °C.
Ammonia was found at concentrations averaging about 2 ppm; hydro-
gen cyanide averaged about 1 ppm. The nitrogen dioxide level was
found to be about 0.05 ppm, and the nitric oxide level was about
1 ppm.
When triethylamine was injected into the flue gas at 370 or
160 °C, the recovery at the outlet of the ESP (160 °C) or the
90 °C zone was considerably higher than that found when the amine
was injected at 650 °C (see Table 11). Although there was some
variation in results, the data indicate that most of the triethyl-
amine passed through the flue gas train unchanged.
53
-------
TABLE 11. DETERMINATION OF TRIETHYLAMINE* IN FLUE GAS
CONTAINING NO SULFUR OR NITROGEN OXIDES
Observed [TEA], ppm Expected
Injection Sampling Sample Filter Bubbler [TEA], Recovery,
temp, °C temp, °C No. catch catch Total ppm
650 160 1
(ESP) 2
3
4
5
6
7
8
^
-
0.0
0.0
0.1
0.0
0.0
< 0. 1
0.0
0.0
0.0
0.2
0.3
1.6
2.8
1.9
0.0
0.0
0.0
0.2
0.4
1.6
2.8
1.9
36.3
36.3
13.6
33.4
20.0
33.4
33.4
19.5
0
0
0
1
2
5
8
10
370 160
(ESP)
160 160
(ESP)
90
1
1
2
3
4
5
6
7
8
9
0.0 16.4
15.1
17.7
18.4
20.5
20.2
0.0 11.5
11.8
11.9
0.0 24.8
16.4
15.1
17.7
18.4
20.5
20.2
11.5
11.8
11.9
24.8
20.3
27.8
27.8
27.8
22.3
21.8
14.7
14.7
14.7
29.0
81
54
64
66
92
93
78
80
81
86
* Abbreviated as TEA in this table.
Injection of Triethylamine into Flue Gas Containing Added Sulfur
Dioxide but No Added Sulfur Trioxide or Nitrogen Oxides—
To investigate the reaction of triethylamine with sulfur
dioxide in the flue gas train, triethylamine was injected near
the inlet of the 160 °C zone into a flue gas mixture containing
about 600 ppm of sulfur dioxide but no other added oxides of sul-
fur or nitrogen. The flue gas was sampled at the outlet of the
ESP (160 °C) or the 90 °C zone.
The results of the analyses of the flue gas samples, which
are presented in Table 12, indicate that much of the triethylamine
passed unchanged through the flue gas train. The recovery of
triethylamine was usually less and the variation in results
greater, however, than those observed without sulfur dioxide in
the flue gas. There was evidence for the formation of diethyl-
amine in the flue gas. The diethylamine level in Sample No. 3 of
Table 12, which was collected at the ESP outlet at 160 °C, was
found to be about 700 ppb. This value is considerably higher than
the 80 to 100 ppb expected to result from the operation of the
triethylamine generator.
54
-------
TABLE 12. DETERMINATION OF TRIETHYLAMINE* IN FLUE GAS
CONTAINING ADDED SULFUR DIOXIDE BUT NO ADDED
SULFUR TRIOXIDE OR NITROGEN OXIDES
Observed (TEA] ,
Sampling
temp, °Ct
160
(ESP)
90
Sample
No.
1
2
3
4
5
6
Filter
catch
0.0
1.5
0.4
3.1
4.5
8.4
Bubbler
catch
14.2
17.2
19.8
11.1
15.0
19.7
ppm
Total
14.2
18.7
20.2
14.2
19.5
28.1
Expected
[TEA] ,
ppm
27.9
27.9
27.9
27.9
27.9
29.0
Recovery,
%
51
67
72
51
70
97
* Abbreviated as TEA in this table.
t Injection temperature, 160 °C.
There was one finding in these experiments that differed
significantly from the results obtained with no sulfur dioxide in
the flue gas. Some triethylamine, presumably present as triethyl-
ammonium ion, was collected by the filter in the sampling train.
In samples taken at the ESP, 8% or less of the total triethylamine
recovered was found on the filter; but in samples taken from the
90 °C zone 22 to 30% was collected on the filter. After portions
of the water washes of the filters were treated with hydrogen
peroxide to oxidize any sulfite present to sulfate, sulfate was
found to be present in amounts that provided a molar excess with
respect to the amounts of triethylamine present. The results
appear to confirm the presence of sulfur-containing salts in the
filter catch of samples taken from both the ESP and the 90 °C zone
and suggest the formation of triethylammonium sulfite or bisulfite
by the reaction of triethylamine and sulfur dioxide in the flue
gas.
Injection of Triethylamine into Flue Gas Containing Added Sulfur
Dioxide and Sulfur Trioxide but No Added Nitrogen Oxides--
When triethylamine was injected into flue gas containing
about 600 ppm sulfur dioxide and about 20 ppm sulfur trioxide, the
total recoveries of triethylamine were comparable to the results
obtained with samples taken from flue gas containing sulfur diox-
ide but no added sulfur trioxide (see Table 13). However, the
fraction of the triethylamine that was trapped by the filter, pre-
sumably as triethylammonium ion, was larger than that found with
no sulfur trioxide in the flue gas. The filter catch amounted to
67 to 80% of the triethylamine recovered at the ESP outlet and 92
to 95% of that recovered from the 90 °C zone. It is likely that
the salt formed in the gas stream was triethylammonium bisulfate
55
-------
TABLE 13. DETERMINATION OF TRIETHYLAMINE* IN FLUE GAS
CONTAINING ADDED SULFUR DIOXIDE AND SULFUR
TRIOXIDE BUT NO ADDED NITROGEN OXIDES
Observed [TEA] ,
Sampling
temp , ° Ct
160
(ESP)
90
Sample
No.
1
2
3
4
Filter
catch
9.8
9.0
17.3
18.8
Bubbler
catch
2.5
4.5
1.5
1.0
ppm
Total
12.3
13.5
18.8
19.8
Expected
[TEA] ,
ppm
23.4
23.4
25.0
25.0
Recovery,
%
53
56
75
79
* Abbreviated as TEA in this table.
t Injection temperature, 160 °C.
because the amounts of sulfate that were determined from filter
washes were in molar excess of the amounts of triethylamine deter-
mined.
A surprising result was that the level of diethylamine
apparently formed in the flue gas by decomposition of triethyl-
amine was greater than the amounts found in the earlier experi-
ments with no sulfur trioxide in the flue gas. The determination
of triethylamine in Sample No. 1 of Table 13 indicated a level of
3.1 ppm. Of this amount, 3.0 ppm was collected by the filter,
presumably as the diethylammonium ion. Diethylamine was also
found in a syrupy residue that collected and charred on the glass
surfaces of the flue gas train at temperatures near and below
160 °C. Analysis of the substance by mass spectrometry revealed
not only the presence of diethylamine but also triethylamine, sul-
fur trioxide, and diethylsulfamic acid.
The addition of 28 ppm of triethylamine to flue gas contain-
ing 600 ppm of sulfur dioxide and 20 ppm of sulfur trioxide pro-
duced a space charge effect at applied voltages just above the
potential required for the initiation of corona in the ESP (see
Figure 5). At higher voltages, the curve was shifted to higher
current densities relative to those obtained with no triethylamine
in the flue gas.
Injection of Triethylamine into Flue Gas Containing Added Sulfur
Dioxide, Nitric Oxide, and Nitrogen Dioxide—
When triethylamine was added to flue gas at 160 °C contain-
ing about 600 ppm of sulfur dioxide, 1000 ppm of nitric oxide,
and 100 ppm of nitrogen dioxide, the overall recovery of triethyl-
amine was around 50% of the amount injected (see Table 14). A
significant fraction (18 to 41%) of the total amount recovered
56
-------
fM
u
c
>
cc
cc
o
<
0.
80
70
60
50
40
30
20
10
CURVE 2
28 ppm TEA
I
4 6 8 10
APPLIED POTENTIAL, kV
12
Figure 5. Current-voltage relationship for the ESP reflecting the space charge
effect from the combination of triethylamine and sulfur trioxide.
57
-------
TABLE 14. DETERMINATION OF TRIETHYLAMINE* IN FLUE GAS
CONTAINING ADDED SULFUR DIOXIDEt AND NITROGEN OXIDES
Sample
No.
1
2
Observed
[TEA] ,
ppm
Filter Bubbler
catch
2.2
5.6
catch
10.2
8.0
Total
12.4
13.6
Expected
[TEA] ,
ppm
20.9
20.9
Recovery,
%
59
65
* Abbreviated as TEA in this table; injected and
sampled at 160 °C.
t The background concentration of sulfur trioxide
was found to be 3 to 6 ppm, although this sub-
stance was not added to the flue gas. The high
background level is attributed to the oxidation
of sulfur dioxide by the nitrogen oxides.
was found in the filter catch, presumably as triethylammonium
bisulfate. The assumed source of the bisulfate was the 3 to 6 ppm
of sulfur trioxide that was present in the flue gas, apparently as
the result of the oxidation of sulfur dioxide by nitrogen dioxide.
As in the experiments discussed earlier, there was evidence
for the formation of diethylamine in the flue gas although the
levels observed were not as high as those measured in flue gas
containing 20 ppm of sulfur trioxide. The total level of diethyl-
amine determined in Sample No. 1 of Table 14 was 170 ppb with
60 ppb of that amount in the filter catch. In Sample No. 2, the
total concentration was 210 ppb with about 100 ppb found on the
filter. The results indicate that about one-third to one-half of
the diethylamine determined was present as a diethy1ammonium salt,
probably as the bisulfate or sulfate.
Perhaps the most interesting finding in the investigation
of triethylamine was the discovery of organic compounds containing
nitrogen (other than triethylamine and diethylamine) in methylene
chloride extracts of the 1 M sodium hydroxide, XAD-2 sorbent, and
cold traps that were used to sample organic matter in the flue
gas at the ESP outlet. One of the compounds was identified as
N-nitrosodiethylamine. The GC retention time of the unknown on
two different analytical columns was identical with that of
N-nitrosodiethylamine. The concentrations of the unknown in
methylene chloride solution were near the detection limit of the
mass spectrometer, but the identity was tentatively confirmed with
gas chromatography coupled with mass spectrometry by monitoring
the molecular ion at m/e 102. A complete mass spectrum was not
obtained.
58
-------
Although only a few samples were taken from flue gas con-
taining added sulfur dioxide and oxides of nitrogen, the results
appeared to give some insight into a possible mechanism for the
formation of the nitrosamine. It was observed that a concentra-
tion of N-nitrosodiethylamine of 67 ppb was found in flue gas con-
taining no nitric oxide, and a comparable concentration (89 ppb)
was found with 750 ppm of nitric oxide in the flue gas. In both
instances, the nitrogen dioxide concentration was about 100 ppm.
Another observation was that an increase in the concentration of
diethylamine vapor in the flue gas from 60 to 110 ppb at compara-
ble levels of triethylamine produced an increase in the amount of
nitrosamine found from 89 to 211 ppb. The results appear to
indicate that the level of nitrosamine was more related to the
concentrations of diethylamine and nitrogen dioxide than to the
concentrations of triethylamine and nitric oxide. These observa-
tions seem to support a mechanism involving the reaction of
diethylamine with nitrogen dioxide to produce N-nitrosodiethyl-
amine .
In subsequent tests, it was observed that some, but appar-
ently not all, of the N-nitrosodiethylamine found in the bubblers
was formed during sampling. When bubblers were spiked with tri-
ethylamine and diethylamine and then exposed to flue gas, some
N-nitrosodiethylamine formed in the bubblers. The results indi-
cated, however, that the bubblers had to be spiked with many times
the amounts of amine vapors actually sampled from the flue gas to
produce amounts of N-nitrosodiethylamine comparable to those found
in unspiked bubblers.
Summary and Discussion of the Results
The addition of triethylamine to flue gas that contained no
added oxides of sulfur or nitrogen revealed that the amine passed
essentially unchanged when injected at temperatures no higher than
370 °C. A significant degradation of triethylamine occurred when
the compound was injected at temperatures near 650 °C. One decom-
position product was diethylamine; it was found at levels of 60 to
600 ppb in the flue gas. No other organic decomposition products
were identified. However, traces of ammonia (2 ppm), hydrogen
cyanide (1 ppm), nitrogen dioxide (50 ppb), and nitric oxide
(1 ppm) were found. No compounds were found in sufficient amounts
to account for all of the triethylamine that was decomposed.
The principal products of the thermal decomposition of tri-
ethylamine eluded detection. In a past study of the degradation
of triethylamine in a closed reactor at temperatures from 470 to
500 °C, a mechanism was proposed that involved the formation of
butane and hydrazine as intermediate products with subsequent
59
-------
decompositon chiefly to methane and nitrogen. "*3 Methane would
not have been trapped in the sampling trains employed in this
investigation. Detection of the small amounts of nitrogen that
would have been produced would have been prohibited by the high
background concentration of nitrogen in the flue gas. Butane and
hydrazine, if present, should have been collected by the cold
traps and resin; however, they were not observed in the analyses.
When triethylamine was added to flue gas containing 600 ppm
of sulfur dioxide, the most interesting result was the finding
that as much as 30% of the amine in the flue gas reacted at 90 °C
to form a bisulfite salt. At 160 °C, only 8% or less of the amine
reacted. The bisulfite (or sulfite) ion was not identified, but
sulfate was found in the filter extracts after the extracts were
treated with hydrogen peroxide. The ability of triethylamine to
produce bisulfite or sulfite under conditions where ammonia does
not produce these compounds is consistent with the higher base
strength of triethylamine.
One possible mechanism for the reaction of sulfur dioxide
with triethylamine is that expressed by Bateman et al.1*1* Accord-
ing to this mechanism, sulfur dioxide reacts to form a one-to-one
addition product with triethylamine that rapidly absorbs water on
exposure to moist air to form triethylammonium bisulfite with a
melting point of about 75 °C. The bisulfite slowly absorbs oxygen
to produce chiefly the bisulfate salt with a melting point of
about 115 °C. If the addition product only exists as an unstable
intermediate in the presence of the water in the flue gas, then
the resulting bisulfite would be expected to slowly absorb oxygen
to produce the bisulfate. Samples taken from the flue gas would
contain the bisulfite, the bisulfate, or a mixture of both.
When triethylamine was injected into flue gas containing
both 600 ppm of sulfur dioxide and 20 ppm of sulfur trioxide,
most of the triethylamine reacted to form the sulfate or bisulfate
salt. As in the previous experiments, the salt amounted to a
greater fraction of the triethylamine injected when sampled at
90 °C than when sampled at 160 °C at the ESP outlet.
An unexpected finding was that as much as 3 ppm of diethyl-
amine was produced in the gas stream. The diethylamine was con-
verted chiefly to a salt, presumably bisulfate, that was collected
at 160 °C at the ESP outlet.
43. Taylor, H. A., and E. E. Juterbock. The Thermal Decomposi-
tion of Triethylamine. J. Phys. Chem., 39:1103-1110, 1935.
44. Bateman, L. C., E. D. Hughes, and C. K. Ingold. Molecular
Compounds Between Amines and Sulfur Dioxide. A Comment on
Jander's Theory of Ionic Reactions in Sulfur Dioxide. J.
Chem. Soc., 1944:243-247.
60
-------
Diethylamine was also found along with triethylamine, sulfur
trioxide, and diethylsulfamic acid in a viscous residue that was
observed to collect and char in the gas train after the introduc-
tion of sulfur trioxide. The discovery of diethylsulfamic acid
in the residue (mp 89 °C) was totally unexpected.
The injection of triethylamine into flue gas containing
sulfur dioxide, nitric oxide, and nitrogen dioxide also gave
interesting results. A compound identified as N-nitrosodiethyl-
amine was found in samples taken from the ESP outlet. The com-
pound was found at concentrations of 67 to 211 ppb. Some of the
compound was apparently produced during the sampling process; but
the amount thus produced was minor compared to the total found.
Although only a few experiments were carried out with oxides
of nitrogen in the flue gas, a mechanism for the formation of N-
nitrosodiethylamine can be proposed from the data available. The
concentration of the nitrosamine appeared to be more dependent
upon the concentration of nitrogen dioxide and diethylamine than
nitric oxide or triethylamine in the flue gas. The mechanism of
formation of the nitrosamine therefore appeared to involve the
reaction of nitrogen dioxide and diethylamine. Because the
degradation of triethylamine or diethylamine in the flue gas was
evident, triethylamine might also be considered a precursor of
the nitrosamine.
SODIUM CARBONATE
The investigation of sodium carbonate was undertaken to com-
plement plans of the Environmental Protection Agency to test the
effectiveness of sodium carbonate as a conditioning agent in field
studies (Contract 68-02-2656) . The main objectives in the labora-
tory experiments were directed toward the identification of reac-
tion products of sodium carbonate with acidic flue gas components
such as sulfur and nitrogen oxides. Examples of the reactions
expected with sulfur oxides to produce sodium sulfite and sodium
sulfate are indicated below:
Na2C03(s) + S02 (g) — > Na2S03(s) + C02 (g) (11)
Na2C03(s) + S03(g) — > NazSO^s) + C02 (g) (12)
In the laboratory experiments, sodium carbonate was injected
near the inlet of the 160 °C constant temperature zone, and the
flue gas was then sampled near the outlet of the ESP (160 °C) or
the outlet of the 90 °C zone. Initial experiments were carried
out with no reactive oxides in the flue gas to determine the
61
-------
losses of sodium carbonate on the walls of the flue gas train.
Reactive sulfur oxides and nitrogen oxides were then added in
later experiments.
Injection Procedure
It was considered desirable in our laboratory experiments
to simulate the injection conditions of sodium carbonate into the
flue gas train that were proposed by the EPA for field tests. The
EPA plans called for the injection of sodium carbonate powder with
a mass median diameter of approximately 10 ym into flue gas
upstream from the electrostatic precipitator at a temperature of
160 °C. The rate of injection of the powder would correspond to
2.5 g of sodium oxide for every 100 g of fly ash. For a typical
fly ash concentration of 9 g/m3 (4 gr/ft3), the appropriate con-
centration of sodium carbonate would be 370 mg/m3 (at 25 °C).
In the laboratory tests, most of these conditions were
reproduced. A dry aerosol of sodium carbonate was produced by
nebulizing an aqueous solution of sodium carbonate into a quartz
tube heated with resistance wire to 200 °C. The solution was
nebulized with a Retec X70/N nebulizer assembly obtained from the
Burton Division of Cavitron Corporation, Van Nuys, California.
The nebulizer was activated with dry nitrogen gas. The dry aero-
sol was injected into the center of the flue gas stream near the
inlet of the 160 °C constant temperature zone. Nebulization con-
ditions were optimized to produce sodium carbonate concentrations
equivalent to 180 to 340 mg/m3 (expressed for 25 °C). These con-
ditions corresponded to 3.40 to 6.57 x 10~3 meq/L (at 25 °C).
The greatest apparent discrepancy between the injection
method proposed for the field tests and the laboratory injection
technique was that the particles injected by the nebulization
method were probably significantly smaller than 10 ym. Although
the particle size distribution of the liquid aerosol produced by
the nebulizer was not measured, the manufacturer's description
indicated that the mass median diameter of the liquid aerosol
generated was about 5 ym. If it is assumed that one liquid parti-
cle produced one salt particle when the water was evaporated, then
the calculated mass median diameter of the particles generated in
this investigation was approximately 0.6 ym.
Analytical Procedures
The sampling and analytical methods employed for the deter-
mination of sulfur dioxide, sulfur trioxide, nitrogen dioxide,
and nitric oxide have been described and referenced earlier in
this report, and descriptions of these methods will therefore not
be repeated here.
62
-------
Particulate material was collected from the flue gas with a
heated Teflon filter as also previously described, but the proce-
dures for analyzing the collected solids were modified in these
experiments. In some experiments, filters exposed to flue gas
were washed with distilled and deionized water, and the wash was
analyzed for one or more of the following ions: sodium, carbon-
ate, sulfate, sulfite, nitrite, and nitrate. Sodium ion was
determined by atomic emission spectroscopy. Carbonate ion (in
the filter wash and in the nebulizer solutions) was determined by
titration with dilute hydrochloric acid solutions. All of the
other anions were analyzed as a mixture by ion chromatography.16
In other experiments, filters were analyzed individually for
sulfite, nitrite, or nitrate ions. Sulfite was determined by the
West-Gaeke colorimetric procedure.1*5 The exposed filter was
washed with tetrachloromercurate solution, and the resulting
dichlorosulfitomercurate was then treated with pararosaniline to
yield the colored product.
Nitrite and nitrate were determined by modification of the
colorimetric methods used for determining oxides of nitrogen in
the flue gas.35'37 To analyze for nitrite, the filter was washed
with the absorbing reagent for the Greiss-Saltzman reaction. The
sum of nitrite and nitrate collected on a filter was determined
by first washing the filter with a solution of hydrogen peroxide
and sulfuric acid to oxidize nitrite to nitrate and then treating
the nitrate ion with phenoldisulfonic acid. The amount of nitrate
present in the original filter wash was determined by difference.
Since sulfite and nitrite salts are easily oxidized in humid
air, it was considered important to test the validity of the sam-
pling method for sodium sulfite and sodium nitrite to insure that
oxidation did not occur during sampling. Teflon filters were
impregnated with sodium sulfite or sodium nitrite and exposed to
flue gas containing added sulfur dioxide, nitrogen dioxide, and
nitric oxide. At a filter temperature of 160 °C, neither the sul-
fite nor the nitrite was stable, although a significant amount of
each salt was recovered when filters were spiked with several
hundred micrograms. At a filter temperature of 90 °C, on the
other hand, both sodium sulfite and sodium nitrite in amounts as
low as 35 ug (5 x 10"1* meq) were stable.
45. West, P. W., and G. C. Gaeke. Fixation of Sulfur Dioxide as
Sulfitomercurate III and Subsequent Colorimetric Determina-
tion. Anal. Chem., 28:1816-1818, 1956.
63
-------
Experimental Results
Injection of Sodium Carbonate into Flue Gas Containing No Added
Sulfur Oxides or Nitrogen Oxides—
Sodium carbonate was first injected into a flue gas stream
containing no added reactive oxides. The salt was injected near
the inlet of the 160 °C constant temperature zone, and the flue
gas was sampled near the outlet of the ESP (160 °C) or the 90 °C
zone. The purpose of these experiments was to determine the
recovery of sodium carbonate under conditions unfavorable to reac-
tion.
The results of the experiments are presented in Table 15.
Approximately 60 to 80% of the injected sodium carbonate was
recovered from the outlet of the electrostatic precipitator at
160 °C and approximately 60 to 65% was recovered from the 90 °C
zone. The low recoveries were mainly due to losses of sodium
carbonate on the walls of the flue gas train. A significant accu-
mulation occurred on the walls downstream from the injection point.
TABLE 15. DETERMINATION OF ANIONS IN FLUE GAS CONTAINING
SODIUM CARBONATE BUT NO ADDED REACTIVE GASES
Anions, meq/L x 1000*
Sampling Sample Observed G03-2
temp, °C No. C03"2 SOt,"2 + S03~2 Total injected
160
90
1
2
3
4
3.77
2.66
2.86
3.06
0.04
0.02
-t
-t
3.81
2.68
2.86
3.06
6.51
3.34
4.74
4.74
* Expressed for 25 °C; converted to the hypothetical
parts-per-million basis (gas by volume) by multiply-
ing by the factor 12.2.
t Not determined.
Little or no reaction of sodium carbonate was observed dur-
ing the tests with sulfur and nitrogen oxides absent from the flue
gas. Very small amounts of sulfate and sulfite were found in two
samples, apparently as the result of sulfur oxides still being
present in the gas stream after previous experiments. No detect-
able amount of nitrite or nitrate was observed. No other anion
was found at significant levels by ion chromatography. The sodium
ion concentrations were found to correspond to the carbonate con-
centrations.
64
-------
Injection of Sodium Carbonate into Flue Gas Containing Added Sul-
fur Dioxide and Nitrogen Oxides—
Sodium carbonate was next injected into flue gas containing
approximately 600 ppm of sulfuric dioxide, 100 ppm of nitrogen
dioxide, and 1000 ppm of nitric oxide. This mixture of reactive
gases produced a background sulfur trioxide concentration of about
6 ppm (0.5 x 10~3 meq/L at 25 °C) in the gas stream. The sodium
carbonate was again injected at the inlet of the 160 °C zone, and
the gas stream was again sampled at the outlet of the ESP or the
90 °C zone.
The primary reaction products found when sodium carbonate
was added to the flue gas were sodium sulfate and sodium sulfite.
The amounts of carbonate and the total amounts of sulfate plus
sulfite recovered are presented in Table 16. The sulfite concen-
trations determined in separate experiments at similar injection
rates of sodium carbonate are given in Table 17. The sulfate and
sulfite salts comprised 50 to 80% of the solids recovered. Sul-
fite levels were usually comparable to sulfate levels. However,
in Sample No. 2 of Table 17, it appears that almost all of the
sample was sulfite. The concentrations of nitrite and nitrate
found were detectable but insignificant relative to the sulfite
and sulfate levels. The highest nitrite concentration was
2 x 10~5 meq/L (expressed for 25 °C), and the highest nitrate con-
centration was 3 x 10~5 meq/L (expressed for 25 °C). The sodium
ion concentrations corresponded to the total anion concentrations.
The low recoveries of the salts in gas samples with respect
to the sodium carbonate added to the flue gas stream were princi-
pally the result of wall losses. Analyses of water washes of the
flue gas train after the experiments cited in Tables 16 and 17
revealed that significant quantities of sodium carbonate, sodium
sulfite, and sodium sulfate had accumulated. The mole fractions
of these compounds in the residue from the walls of the 160 °C
constant temperature zone and the electrostatic precipitator were
0.36 sodium carbonate, 0.05 sodium sulfite, and 0.59 sodium sul-
fate. The mole fractions in the residue from the walls of the
90 °C zone were 0.36 sodium carbonate, 0.55 sodium sulfite, and
0.09 sodium sulfate.
Summary and Discussion of the Results
In summary, sodium carbonate reacted extensively with the
sulfur oxides in the flue gas. The location of the reaction—in
the flue gas or on the collection medium—could not be identified;
but similar reactions in a power plant flue would be expected to
occur on precipitator electrodes, on the walls of the gas train,
and possibly on the surface of fly ash particles.
The formation of some of the sodium sulfate found in flue
gas samples was undoubtedly the result of the reaction of sodium
carbonate with sulfur trioxide (6 ppm was present as sulfuric
65
-------
TABLE 16. DETERMINATION OF ANIONS IN FLUE GAS CONTAINING
SODIUM CARBONATE AND REACTIVE GASES
Anions, meq/L x 1000*
Sampling
temp, °C
160
90
Sample
No.
1
2
3
4
5
6
7
C03~2
0.74
0.29
0.60
0.91
1.49
1.80
1.49
Observed
SO.,"2 + S03~2
0.80
1.37
1.31
2.26
1.54
1.03
1.03
Total
1.54
1.66
1.91
3.17
3.03
2.83
2.52
C03~2
injected
3.60
3.40
4.26
5.71
5.23
4.74
6.57
* Expressed for 25 °C; converted to the hypothetical
parts-per-million basis (gas by volume) by multiply-
ing by the factor 12.2.
TABLE 17. DETERMINATION OF SULFITE IN FLUE GAS CONTAINING
SODIUM CARBONATE AND REACTIVE GASES
Sampling
temp , ° C
160
90
Sample
No.
1
2
3
4
5
SO3~~2 observed,
meq/L x 1000*
0.54
1.86t
0.44t
0.14
0.17
CO3~2 injected,
meq/L x 1000
5.66
5.66
5.66
4.74
6.57
* Expressed for 25 °C; converted to the hypotheti-
cal parts-per-million basis (gas by volume) by
multiplying by the factor 12.2.
t Total S03~2 + S04~2 = 1.93 x 10~3 meq/L
expressed for 25 °C.
66
-------
acid vapor at 160 °C). The remainder of the sulfate found at
160 °C must have been produced by the oxidation of sodium sulfite.
The presence of sulfite in some of the samples withdrawn from the
gas stream and on the walls of the flue gas train shows clearly
that sulfur dioxide reacted with sodium carbonate. Nitrogen diox-
ide and nitric oxide were found to undergo no significant reaction
with sodium carbonate.
AMMONIUM SULFATE
Ammonium sulfate is the primary component of the first pro-
prietary conditioning formulation studied in this investigation,
Apollo Chemical Corporation's Coaltrol LPA-40. Ammonium sulfate
was always injected into the flue gas train at 650 °C in these
experiments, either as an aqueous solution of reagent-grade
ammonium sulfate or as the aqueous proprietary formulation LPA-40.
The compound was injected at such a high temperature because
Apollo normally injects LPA-40 ahead of the economizer in full-
scale power plants. The gas stream was sampled at 650, 160, or
90 °C. The primary objective in the study of this agent was to
identify and quantitate (if possible) the thermal decomposition
products and recombination products of ammonium sulfate in the
flue gas stream in both the absence and presence of added reactive
sulfur and nitrogen oxides.
The probable decomposition products at 650 °C were the gases
shown in the following equation:
(NIU)2S04 (s) —> 2NH3(g) + H20(g) + S03 (g) (13)
The occurrence of this reaction is consistent with the thermo-
dynamic data of Kelley et al."6 and Scott and Cattell.1*7 Stepwise
reversal of this reaction was expected with decreasing temperature:
S03(g) + H20(g) —» HaSCMg) (1)
46. Kelley, K. K., C. H. Shomate, F. E. Young, B. F. Naylor,
A. E. Salo, and E. H. Huffman. Thermodynamic Properties of
Ammonium and Potassium Alums and Related Substances with
Reference to Extraction of Alumina from Clay and Alunite.
Technical Paper 688. U. S. Bureau of Mines, Washington,
D. C., 1946. pp 66-69.
47. Scott, W. D., and F. C. R. Cattell. Vapor Pressure of Ammo-
nium Sulfates. Atmos. Environ., 13:307-317, 1979.
67
-------
NH3(g) + H2S(Mg) —> NHnHSCMl or g) (14)
NH3(g) + NH^HSCMl or g) —» (NHi,) 2SOi» (s) (15)
Decomposition products accompanying those shown in
Equation 13 were expected to include nitrogen, nitrogen oxides,
and sulfur dioxide. This statement is based on the findings of
Halstead1*8 in an investigation of the decomposition products of
ammonium sulfate.
Injection Procedure
An aqueous solution of ammonium sulfate or a filtered solu-
tion of LPA-40 was nebulized with a Retec X70/N nebulizer assembly
(Burton Division, Cavitron Corporation, Van Nuys, California).
The nebulizer was activated with dry nitrogen gas, and the ammo-
nium sulfate aerosol resulting from the nebulization process was
introduced into the 650 °C constant temperature zone through a
quartz tube that extended to the center of the flue gas stream.
The rate of injection was chosen so that the concentration of
ammonium sulfate in the flue gas did not exceed 41 yg/L at 650 °C,
a concentration corresponding to the upper limit specified by
Apollo.12'13 To achieve levels below this limit in the laboratory
flue gas train, a dilute solution of LPA-40 was injected. As dis-
cussed in Section 3 of this report, the LPA-40 formulation con-
tained small quantities of an uncharacterized insoluble iron com-
pound. The formulation was filtered to remove this compound and
diluted by a factor of two. This dilute solution was injected at
a nebulization rate of 0.2 mL/min. The measured concentrations
of ammonium sulfate injected into the flue gas streams ranged from
9 to 41 yg/L at 650 °C, which is equivalent to 4.2 x 10~" to
19 x 10"1* meq/L at 25 °C.
The average injection rate of ammonium sulfate was usually
determined by measuring the amount of sulfate in the nebulizer
before and after the nebulization period and by calculating the
amount of sulfate lost. Frequently, however, the injection rate
was found by determining the amount of ammonium ion lost during
the nebulization period. In some experiments, the rates deter-
mined by the two methods agreed within experimental error. In
others, however, it appeared that as much as 25% more ammonium ion
was lost than sulfate (based on the ratio of equivalents lost from
the nebulizer). The apparent origin of the excess was the decom-
position of a residue of ammonium sulfate that collected in the
48. Halstead, W. D. Thermal Decomposition of Ammonium Sulphate,
J. Appl. Chem., 20:129-132, 1970.
68
-------
outer tube of the nebulizer at high injection rates. Presumably,
the ammonium sulfate was converted to ammonium bisulfate, and
ammonia gas was released.
Analytical Procedures
Before and after each experiment, the nebulizer solution was
analyzed for ammonium sulfate. In addition, the flue gas was
analyzed for the following possible degradation and recombination
products: sulfur dioxide, sulfur trioxide, ammonia, nitric oxide,
nitrogen dioxide, ammonium bisulfate, and ammonium sulfate. When
LPA-40 was injected, the gas stream was also analyzed in some
experiments for the traces of iron that were found in the analysis
of the LPA-40 formulation.
Most of the analytical methods used were similar or identi-
cal to those described and referenced earlier in this report.
Sulfur trioxide was determined by sampling the flue gas through a
controlled condensation coil and by subsequently measuring the
collected sulfate by the barium perchlorate-Thorin titration
method18 or by ion chromatography.l6 (In some instances, ammonia
was present concurrently in the sampling stream, and the sulfur
trioxide was actually collected as ammonium sulfate and not as
condensed sulfuric acid.) Sulfur dioxide was collected in a bub-
bler containing 3% hydrogen peroxide solution, and the resulting
sulfate was determined by ion chromatography. Ammonia gas was
collected in bubblers containing 0.1 N sulfuric acid solution and
determined by the phenol-hypochlorite colorimetric method.17
In the absence of added nitrogen oxides, nitrogen dioxide
was determined by the Greiss-Saltzman procedure;35 nitric oxide
was first oxidized to nitrogen dioxide on firebrick impregnated
with chromium trioxide and then analyzed by the Greiss-Saltzman
method. In the presence of added nitrogen oxides, the phenol-
disulfonic acid method37 was used to determine the sum of nitric
oxide and nitrogen dioxide concentrations, and the Greiss-Saltzman
method was used to determine the nitrogen dioxide concentration.
The concentration of nitric oxide was then determined by differ-
ence .
Heated Teflon filters and heated quartz wool plugs were
employed for the collection of particulate material from the flue
gas. Exposed filters and plugs were usually washed with dis-
tilled, deionized water; the wash was analyzed for ammonium ion
by the phenol-hypochlorite colorimetric method and for sulfate,
sulfite, nitrate, and nitrite by ion chromatography. The solid
material that collected on the glass tubing used for connecting
the various components of the sampling trains was also dissolved
in water and analyzed for ammonium ion and the anions listed.
above. In some experiments, exposed filters were washed with
tetrachloromercurate solution, and sulfite was determined by the
West-Gaeke method."5 In other experiments, iron was determined
69
-------
by atomic absorption spectroscopy in the water washes of the
quartz wool plugs. Frequently, the plugs were also washed with
hot hydrochloric acid, and iron was determined in these washes.
Experimental Results
Injection of LPA-40 into Flue Gas Containing No Added Oxides of
Sulfur or Nitrogen—
In the investigation of LPA-40 in the absence of added reac-
tive oxides in the gas stream, the agent was always injected into
the flue gas near the inlet of the 650 °C constant temperature
zone, and the flue gas was always sampled near the outlet of the
650 °C zone. The types of experiments performed fell into two
categories: (1) the determination of the degradation products at
650 °C and (2) the determination of the recombination products at
160 and 90 °C. In the thermal degradation experiments, the sam-
pling train was maintained at a high temperature, while in the
recombination studies the sampled flue gas was allowed to cool
rapidly to 160 or 90 °C in the sampling train. The purpose of
this sampling procedure was to minimize the wall losses that would
have occurred in the flue gas train if the gas stream had been
sampled at the outlets of the 160 and 90 °C constant temperature
zones.
Thermal decomposition studies—In the thermal decomposition
studies, a sampling train was employed that kept the sampled gas
hot enough to avoid the formation of salts prior to the determina-
tion of the expected gases. The sampling train included a plug
of quartz wool enclosed in a glass tube at approximately 500 °C
followed by a sampling manifold at about 400 °C with three sam-
pling ports. One port was for the determination of ammonia, a
second for the determination of sulfur oxides, and a third for
the determination of nitrogen oxides. In some experiments the
manifold was omitted, and either ammonia or sulfur oxides were
determined.
The results of these experiments are given in Table 18. The
primary thermal degradation products were sulfur trioxide, sulfur
dioxide, and ammonia. The sulfur oxides recovered were approxi-
mately 90% sulfur trioxide and 10% sulfur dioxide. Microgram
quantities of a solid iron compound were also found but are not
listed in Table 18. Nitrogen or nitrogen oxides should also have
been produced along with sulfur dioxide by oxidation of ammonia
by sulfur trioxide. Because of the large background levels of
nitrogen in the flue gas, however, measurement of the trace
amounts of nitrogen produced by this reaction was not possible.
Nitrogen oxides were found at levels no higher than the background
concentrations of approximately 10 ppb.
All of the ammonium sulfate injected into the gas stream as
LPA-40 could not be accounted for in these experiments. Since no
other decomposition product than those discussed above was found,
70
-------
TABLE 18. DETERMINATION OF THERMAL DECOMPOSITION PRODUCTS
OF LPA-40 AT 650 °C IN FLUE GAS CONTAINING
NO ADDED SULFUR OR NITROGEN OXIDES
Concentration, meg/L x 10 OOP*
Observed
Sample
No.
1
2
3
4
5
6Af
est
7Af
7Bt
SO 3
7.62
4.31
3.75
4.40
4.78
6.51
7.91
S02
0.71
0.46
0.48
0.73
0.41
0.44
0.56
Total
SOX
8.33
4.77
4.23
5.13
5.19
6.95
8.47
NH3
5.65
5.32
3.62
4.64
2.96
8.14)
14.16)
Injected
SOu"2
6.78
6.78
6.40
6.40
4.21
9.39
10.39
NHi,*
_ j.
—
—
—
_
10.39
12.71
* Expressed for 25 °C; converted to the parts-per-
million basis (gas by volume) by multiplying the
SO3 and SOa concentrations by 1.22 and the NH3 con-
centration by 2.44.
t A dash indicates that the concentration was not
determined.
t A filter was added near the end of the sampling
train for ammonia.
the lack of mass balance must be chiefly attributed to peculiari-
ties in the injection procedure, losses of species on the walls
of the flue gas train, or artifacts in the sampling procedure.
As discussed earlier, in some experiments more ammonia than sul-
fate (expressed in equivalents) was apparently injected into the
gas stream. In addition, sulfur oxides may have been irreversibly
sorbed on quartz or glass surfaces. The affinity of glass for
sulfur trioxide was demonstrated; typically, over half of the sul-
fur trioxide recovered was collected by the quartz wool plug.
Losses of ammonia (as ammonium sulfate) probably occurred during
the collection of Samples Nos. 1 to 5 listed in Table 18. The
passage of an ammonium salt through the bubblers employed for the
collection of ammonia was evident during the collection of Samples
Nos. 6A and 7A. It was found that 9 to 20% of the total amount of
ammonia plus ammonium ion was collected on filters placed down-
stream from the bubblers.
Recombination studies—In the chemical recombination experi-
ments, flue gas was sampled at 650 °C and cooled to 160 or 90 °C
and passed through a Teflon filter heated to the same temperature.
71
-------
The filter was followed by the sampling manifold; ammonia, sulfur
dioxide, sulfur trioxide (or sulfuric acid vapor) and nitrogen
oxides were determined at separate ports.
The results of these experiments are listed in Table 19.
When the filter temperature was maintained at 160 °C, the primary
product found was ammonium bisulfate, with lesser amounts of ammo-
nium sulfate. The predominance of the bisulfate salt was indi-
cated by the low ratios of ammonium ion concentration to the anion
concentration. The average ratio expressed in terms of equiva-
lents was 0.59, which corresponds to mole fractions of 0.83 for
ammonium bisulfate and 0.17 for ammonium sulfate in the filter
catch. An average of about 4% of the total sulfur species
recovered was sulfur trioxide (or sulfuric acid vapor) and about
4% was recovered as sulfur dioxide. Ammonia was the only nitrogen
compound found at significant concentrations. Nitrogen oxides
were not above the background levels.
With a filter temperature of 90 °C, the primary recombina-
tion product was ammonium sulfate. The average ratio of ammonium
ion concentration to anion concentration expressed in terms of
equivalents was 0.89, which corresponds to mole fractions of 0.76
for ammonium sulfate and 0.24 for ammonium bisulfate in the filter
catch. Sulfur trioxide (or sulfuric acid vapor) was found to be
present at concentrations representing 0.5 to 0.8% of the sulfur
compounds, while sulfur dioxide corresponded to 4 to 6%. Since no
sulfuric acid should exist as vapor at 90 °C, the sulfuric acid
found must have occurred as fine particles of condensed liquid
that slipped through the filter. As above, ammonia was the only
nitrogen compound found in significant concentrations.
The data in Table 19 do not reveal good mass balances.
Since products other than those listed are not found, low recover-
ies must be attributed to peculiarities of the injection procedure,
wall losses, and artifacts of the sampling as in the previously
discussed results.
Injection of an Aqueous Solution of Ammonium Sulfate into Flue Gas
Containing No Added Oxides of Sulfur or Nitrogen—
An aqueous solution of ordinary ammonium sulfate (rather
than LPA-40) was injected into the flue gas near the inlet of the
650 °C constant temperature zone and was sampled near the outlet
of the same reaction zone. The flue gas stream contained no added
sulfur or nitrogen oxides. Determinations were made of the degra-
dation products present at 650 °C and of the recombination pro-
ducts at 160 and 90 °C, as described in connection with the
studies of LPA-40.
The primary thermal degradation products of ammonium sulfate
found in these experiments were the same as those of LPA-40 at
650 °C—ammonia, sulfur trioxide (or sulfuric acid vapor), and
small amounts of sulfur dioxide. The sulfur oxides recovered
72
-------
TABLE 19. DETERMINATION OF RECOMBINATION PRODUCTS
OF LPA-40 IN FLUE GAS CONTAINING NO ADDED
SULFUR OR NITROGEN OXIDES
(Samples Taken from the 650 °C Reaction Zone)
Concentration, meg/L x 10 OOP*
Observed
Filter
temp, ° C
Sample
No.
HS04~
or S0u~2
Total sulfur
SO 3 S02 species
SO,,"2
injected
160
90
1
2
3
4
5
6
7
8
7.62
3.99
6.09
5.58
6.93
13.63
13.70
15.89
0.31
0.12
0.06
0.18
0.96
0.12
0.08
0.14
0.37
0.15
0.02
0.26
0.66
0.42
0.53
0.94
14.76
9.98
19.08
Total nitrogen
NHi»+ NH3 species
160 1
2
3
4
5
6
90 7
8
4.86
2.69
2.18
3.77
4.15
8.02
12.46
13.67
6.
3.
4.
3.
7.
9.
5.
5.
50 11.37
61 6.30 )
98 7.16
51 7.29 >
22 11.37
50 17.52
27 17.73
77 19.44
injected
-t
9.85
19.30
* Expressed for 25 °C; converted to the hypothetical parts-per-
million basis (gas by volume) by multiplying the concentra-
tions of the sulfur species by 1.22 and the concentration of
NH3 by 2.44.
t NHi, not determined.
consisted of about 95% sulfur trioxide and 5% sulfur dioxide;
there was no indication that the relative amount of sulfur dioxide
changed as the result of the absence of the iron compound found
in LPA-40.
At 160 °C the predominant recombination product found was
ammonium bisulfate; at 90 °C the main product was ammonium sul-
fate, as during the injection of LPA-40. The predominance of one
sulfate salt over the other was indicated by the ratios of
ammonium ion concentrations to the sulfate and bisulfate concen-
trations. At 160 °C the average ratio expressed in terms of
equivalents was 0.58, which corresponds to mole fractions of 0.84
73
-------
for ammonium bisulfate and 0.16 for ammonium sulfate; at 90 °C
the average was 0.85, which corresponds to mole fractions of 0.70
for ammonium sulfate and 0.30 for ammonium bisulfate.
Injection of LPA-40 into Flue Gas Containing Added Sulfur and
Nitrogen Oxides—
Dilute, filtered solutions of LPA-40 were nebulized into
the flue gas stream near the inlet of the 650 °C zone. Approxi-
mately 600 ppm sulfur dioxide, 1000 ppm nitric oxide, and 100 ppm
nitrogen dioxide were present in the flue gas mixture. The gas
stream was sampled near the outlet of the 650 °C zone to determine
the thermal degradation products and, after rapid cooling from
650 °C to 160 or 90 °C in the sampling train, to determine the
chemical recombination products. Sampling was also carried out
from the outlet of the ESP (160 °C) and from the outlet of the
90 °C constant temperature zone.
Thermal decomposition studies—The results of the thermal
decomposition experiments are presented in Table 20. The primary
thermal decomposition products found were sulfur trioxide and
ammonia, the same principal degradation products that were found
in the absence of reactive gases in the flue gas stream. Because
of the large background concentrations of sulfur dioxide (approxi-
mately 500 x lO"1* meq/L) that were added to the gas stream in
these experiments, however, the small amounts of sulfur dioxide
(0.48 x 10"** meq/L) that occurred in the thermal degradation of
LPA-40 in the absence of reactive gases could not be detected.
TABLE 20. DETERMINATION OF DECOMPOSITION PRODUCTS
OF LPA-40 AT 650 °C IN FLUE GAS CONTAINING
ADDED SULFUR AND NITROGEN OXIDES
Concentration, meq/L x 10
Sample
No.
1
2
3
Observed
SO 3
12.11
4.18
2.67
NH3
10.38
2.51
3.91
000*
Expected
SO 3?
11.53
11.19
11.19
NH3
8.19
7.85
7.85
* Expressed for 25 °C; converted to the parts-
per-million basis (gas by volume) by multi-
plying the concentration of SO 3 by 1.22 and
the concentration of NH3 by 2.44.
t Based on SO<,~2 lost from the nebulizer.
t Includes the average background concentra-
tion of SO3 (3.34 x 10"* meq/L) and there-
fore exceeds the concentration of NH3.
74
-------
Recombination studies—The results of the chemical recombi-
nation experiments are listed in Tables 21 and 22. Table 21
shows the recombination products found in gas samples taken from
the 650 °C constant temperature zone in external sampling trains
at 160 and 90 °C. Table 22 shows the recombination products found
in samples taken from the ESP and the 90 °C zone. The predominant
recombination products found at 160 and 90 °C were ammonium sul-
fate salts in both types of sampling arrangements. The data
obtained in the presence of reactive gases in the flue gas stream
were, however, based on only a few determinations, had a large
scatter, and gave poor mass balances. At 160 °C the ratio of
ammonium ion concentration to anion concentration (each expressed
in meq/L) was found to range from 0.85 in one experiment (which
would indicate a predominance of ammonium sulfate) to 0.21 in
another (which would indicate a predominance of ammonium bisulfate
in an excess of sulfuric acid). At 90 °C the ratio ranged from
0.51 to 0.06. With the exception of the one value of 0.41, which
corresponds to mole fractions of 0.98 for ammonium bisulfate and
0.02 for ammonium sulfate, these low ratios indicate an excess of
sulfuric acid over any ammonium salt present. Any sulfate salt
initially formed would have been converted to the bisulfate salt
by the excess acid. Thus, the uncertainty in the data does not
permit a definite conclusion to be drawn about the distribution
of the recombination products between ammonium sulfate and ammo-
nium bisulfate.
Additional experimental findings—The only chemical species
found at significant levels in any of the experiments with added
sulfur and nitrogen oxides in the gas stream were ammonium and
sulfate ions, sulfur trioxide, and ammonia. Only negligible
amounts of nitrite, nitrate, and sulfite ions could be detected
in any of the filter washes analyzed by ion chromatography. And
because of the high background levels of nitric oxide, nitrogen
dioxide, and sulfur dioxide, the small quantities of these oxides
that could possibly be formed in these experiments could not be
determined. The lack of mass balances found in these experiments
must be chiefly attributed to wall losses of the injected species
and the thermal degradation products (or, in some cases, to desorp-
tion of previously adsorbed species back into the gas stream) and
to artifacts in the injection and sampling procedures.
At the completion of the injections of LPA-40, the various
reaction zones and heat exchangers of the flue gas train were
washed, and the washings were analyzed by ion chromatography. The
only anionic species found in other than negligible amounts was
sulfate ion. Analysis of an acidic wash of the 650 °C reaction
zone by atomic absorption spectrophotometry indicated that trace
amounts of iron had been coated on the inside walls of the reac-
tion zone.
75
-------
TABLE 21. DETERMINATION OF RECOMBINATION PRODUCTS
OF LPA-40 IN FLUE GAS CONTAINING ADDED
SULFUR AND NITROGEN OXIDES
(Samples Taken from the 650 °C Reaction Zone)
Concentration, meq/L x 10 OOP*
Observed
Filter
temp, °C
160
90
Sample
No.
1
2
3
4
5
6
HSOi,~
or S0if-2t
43.21
3.45
12.35
8.89
6.74
4.50
NH4 +
36.86
0.73
6.26
0.50
1.67
0.56
NH3
0.16
0.04
1.09
0.00
-#
0.08
Injectedt
S04~2§
24.54
-#
7.47
19.38
9.68
8.91
NH4 +
21.20
-#
4.13
16.04
6.34
5.57
* Expressed for 25 °C; converted to the hypothetical
parts-per-million basis (gas by volume) by multiply-
ing the HSOit" or SOi*"2 concentrations by 1.22 and
the NH3 concentrations by 2.44.
t Based on SOi*"2 lost from nebulizer.
t Aerosol collected on the filter. Only small addi-
tional amounts of S0i»~2 were found as SO 3 when this
substance was determined (0.47 x 10"1* meq/L in
Sample No. 1 and 1.50 x 10"" meq/L in Sample No. 3),
§ Includes the average background concentration of
SOs (3.34 x 10~4 meq/L) and therefore exceeds the
concentration of
# Not determined.
In most of the experiments with added nitrogen oxides pres-
ent in the gas mixture, the concentrations of recovered ammonia
and ammonium ion were significantly less than the injected concen-
trations of ammonium ion into the gas stream (see Tables 20, 21,
and 22). This observation is consistent with the finding obtained
in the studies with ammonia that extensive reaction occurs between
nitrogen oxides and ammonia in the flue gas train at 650 °C.
Summary and Discussion of the Results
The decomposition and recombination products of LPA-40
injected at 650 °C into a simulated flue gas stream containing no
added sulfur oxides or nitrogen oxides were essentially identical
to those products found during the injection of an aqueous
76
-------
TABLE 22. DETERMINATION OF RECOMBINATION PRODUCTS
OF LPA-40 IN FLUE GAS CONTAINING ADDED
NITROGEN AND SULFUR OXIDES
(Samples Taken from the ESP [160 °C] and the
90 °C Reaction Zone)
Concentration, meq/L x 10 OOP*
Observed
Total S0^~2
S03 SO/,"2 injectedt
Sampling Sample HSOi»~
temp, °C No. or SOi,"2
160
(ESP)
1
2
15.06
0.71
3.29
0.82
18.35
1.53
7.47
9.43
90
1.55
0.33
1.88
19.38
90
Total
NHi,'
NH3 NH.,* + NH3 injected
160
(ESP)
1
2
4.94
0.40
0.58
0.04
5.52
0.44
4.14
6.10
0.64
0.00
0.64
16.04
* Expressed for 25 °C; converted to the hypothetical
parts-per-million basis (gas by volume) by multiply-
ing the HSOi»~ or S0i»~2 concentrations by 1.22 and the
or NH3 concentrations by 2.44.
t Based on SOi,"2 lost from the nebulizer; includes the
average background concentration of SO 3 (3.34 x
10" "* meq/L) and therefore exceeds the concentration
of NH3.
solution of pure ammonium sulfate. The primary degradation pro-
ducts at 650 °C were ammonia and sulfur trioxide. Small amounts
of sulfur dioxide were also observed. Nitrogen oxides were not
produced from LPA-40 in measurable quantities. Traces of an iron
compound, however, were released into the gas stream from LPA-40.
Nevertheless, comparison of the results with LPA-40 and the
results with ordinary ammonium sulfate indicated that the pres-
ence of the iron compound in LPA-40 had no significant effect on
the recovery of any of the species found in the flue gas.
The predominant species found when the decomposition pro-
ducts were cooled to 160 °C was identified as ammonium bisulfate,
and the major species found at 90 °C was ammonium sulfate. Inso-
far as ammonium bisulfate is concerned, however, the equivalent
77
-------
mixture of ammonium sulfate and sulfuric acid may have been pres-
ent, as in the ammonium sulfate decomposition products studied by
Scott and Cattell."7
The decomposition products of LPA-40 injected at 650 °C
into a simulated flue gas stream containing added sulfur oxides
and nitrogen oxides were essentially identical to those products
found in the absence of reactive oxides in the flue gas: ammonia
and sulfur trioxide. The predominant recombination products found
at 160 and 90 °C with the reactive gases present were ammonium
sulfate salts. But because of uncertainties in the data, the dis-
tribution of the recombination products between ammonium sulfate
and ammonium bisulfate could not be determined.
DIAMMONIUM HYDROGEN PHOSPHATE
The final flue gas conditioning agent studied in this inves-
tigation was diammonium hydrogen phosphate. This agent was inves-
tigated solely as Apollo Chemical Corporation's proprietary
formulation Coaltrol LPA-445, which, as discussed in Section 3 of
this report, consists predominantly of an aqueous solution of
diammonium hydrogen phosphate. Dilute, aqueous solutions of
LPA-445 were introduced into the gas stream at 650 °C in both the
absence and presence of added sulfur and nitrogen oxides, and the
flue gas was sampled at 650, 370, 160, and 90 °C. As in the study
of LPA-40, the main objectives in this study were to identify the
thermal decomposition and chemical recombination products of
diammonium hydrogen phosphate in the flue gas stream. The LPA-445
solutions were injected at 650 °C in all of these experiments
because Apollo indicated that all of the LPA series of additives
are intended for "high temperature" application, which implies
injection of the LPA agents upstream from the economizer.
Background Information on Ammonium Phosphate Salts
Thermal Decomposition Pathways of Ammonium Phosphates--
All ammonium phosphates dissociate when heated, evolving
gaseous ammonia and leaving phosphoric acid,1*9 as illustrated by
the following equation for diammonium hydrogen phosphate:
--> 2NH3 + HaPO* (16)
49. Van Wazer, J. R. Phosphoric Acids and Phosphates. In:
Kirk-Othmer Encyclopedia of Chemical Technology, 2nd ed.,
Vol. 15, H. F. Mark, J. J. McKetta, and D. F. Othmer, Eds,
Interscience Publishers, New York, New York, 1968. pp 232-
276.
78
-------
This is the basis of the widespread use of ammonium phosphates as
flame retardants. At the normal injection temperature of LPA-445
(650 °C), the phosphoric acid resulting from the decomposition of
an ammonium phosphate could either decompose completely into phos-
phorus pentoxide (P20s) and water or polymerize into species known
as condensed phosphates.
Condensed phosphates are phosphate species containing
several phosphorus atoms and having P-O-P linkages.50 The three
most important condensed phosphates are (1) the linear polyphos-
phates, whose anions have the general formula [Pn03n+i 1 ~(n+1)
(2) cyclic polyphosphates, or metaphosphates, whose anions have
the general formula [Pn03n]~n? and (3) "infinite chain" meta-
phosphates, whose anions also have the general formula (pn°3n]~n-
The most common linear polyphosphate is the pyrophosphate ion,
whose structure is:
0 0
I I
-0-P-O-P-CT
I I
o- o~
And the most common cyclic metaphosphate is the trimetaphosphate
ion, whose structure is:
o-o
The thermal decomposition of diammonium hydrogen phosphate
was studied by Erdey, Gal, and Liptay5 l by means of thermal gravim-
etry and differential thermal analysis. These researchers found
that diammonium hydrogen phosphate thermally decomposes in a step-
wise fashion with several decomposition processes proceeding
sequentially as the temperature is raised. These processes are:
50. Cotton, F. A., and G. Wilkinson. Advanced Inorganic
Chemistry, 2nd ed. Interscience Publishers, New York, New
York, 1967. 1136 pp.
51. Erdey, L., S. Gal., and G. iiptay. Thermoanalytical Proper-
ties of Analytical-Grade Reagents—Ammonium Salts. Talanta,
11:913-940, 1964.
79
-------
< 1 RO °P
(s) -—----» NHnHzPO., (s) + NH3 (g) (17)
150 °C
±2----» NHi^PCMl) (18)
2NHlfH2POil (1) -__» (NH^)2H2P207(s) + H20(g) (19)
-------
(s)
-------
both salts are completely or nearly completely dissociated. At
90 °C triammonium phosphate is seen to be still completely disso-
ciated and diammonium hydrogen phosphate significantly dissociated.
Thus, like the thermogravimetric data, the dissociation pressure
data indicate that at 90 °C ammonium dihydrogen phosphate is
thermally stable and diammonium hydrogen phosphate partially
stable, and at 160 °C (and higher) only the monoammonium salt is
stable. These conclusions, however, assume that a sufficient
amount of salt is present to give the indicated dissociation pres-
sure and still leave a residue of the original salt in the system.
The impact of this factor on conclusions about the salts present
is discussed after the presentation of the experimental results.
Expected Decomposition and Recombination Products of LPA-445—
On the basis of the information discussed above, the injec-
tion of LPA-445 into the simulated flue gas stream at 650 °C was
expected to result in the thermal decomposition of diammonium
hydrogen phosphate into ammonia, water, phosphorus pentoxide
(which would be analyzed as orthophosphate ion in aqueous solu-
tion) , and possibly condensed phosphate species. As the tempera-
ture of the gas stream was lowered during passage through the gas
train, various ammonium orthophosphates and condensed phosphates
were expected to result from recombination reactions of the ther-
mal decomposition products. Traces of nitrogen oxides might also
occur as the result of the oxidation of ammonia in the gas stream.
In addition, substantial amounts of the ammonia formed in the
thermal decomposition of the injected LPA-445 might be destroyed
in the 650 °C zone in the presence of added nitrogen oxides in
the gas stream due to the reaction of ammonia with nitrogen oxides.
No unusual or highly toxic decomposition or reaction pro-
ducts of measurable concentration were expected to be formed in
the flue gas stream when LPA-445 was injected. The only highly
toxic compound that we could envision even possibly being formed
in the gas stream was phosphine (PHa). Phosphine is a colorless,
highly toxic gas at room temperature and has an OSHA threshold
limit value of 0.3 ppm. Phosphine could possibly be formed in
the gas stream from the reduction of phosphorus pentoxide, ortho-
phosphate ion, or some other phosphorous-containing species. Thus,
plans were made for the gas stream to be sampled and analyzed
specifically for phosphine.
Injection Procedure
Dilute, filtered aqueous solutions of LPA-445 were injected
into the flue gas stream near the inlet of the 650 °C constant
temperature zone at a nebulization rate of approximately
0.2 mL/min with a Retec X70/N nebulizer assembly.
Chemical analysis of the LPA-445 solutions in the nebulizer
before and after a series of experiments indicated that ammonium
and phosphate ions were not injected into the gas stream in the
82
-------
mole ratio corresponding to diammonium hydrogen phosphate. Two
ammonium ions would be expected for every phosphate ion (or hydro-
gen phosphate ion) injected if the salt were uniformly nebulized
from solution according to its empirical formula, (NHi,) 2HPCK.
Instead, it was found that an excess of ammonium ion was nebulized
into the gas stream. During the first half of the investigation,
when no added sulfur oxides or nitrogen oxides were present in
the gas mixture, the ratio of injected ammonium ions to injected
phosphate ions was approximately 2.5:1.0. During the later experi-
ments, when sulfur oxides and nitrogen oxides were added to the
gas stream, the Retec nebulizer was not functioning properly. In
these experiments the ratio of injected ammonium ions to injected
phosphate ions was much higher than in the earlier experiments,
averaging approximately 3.8:1.0. This phenomenon of excess
ammonia injection can be partially explained by the appreciable
vapor pressure of ammonia that is present over solutions of dianuno-
nium hydrogen phosphate, even at room temperature.55 However, it
cannot completely account for the excess ammonia injected, and no
satisfactory explanation can be given.
During the experiments with reactive sulfur and nitrogen
oxides present in the gas stream, the nebulizer outputs were sig-
nificantly less than in the earlier experiments with no reactive
gases present in the gas stream. The nebulizer outputs in the
earlier experiments of this investigation averaged 2.15 ymol/L
(expressed for 25 °C) for ammonium ion and 0.90 ymol/L for phos-
phate ion, whereas the average outputs determined during the later
experiments were only 0.95 ymol/L for ammonium ion and 0.25 ymol/L
for phosphate ion. The source of these injection discrepancies
was traced to a malfunctioning of the aging Retec X70/N nebulizer
assembly. The malfunctioning appeared to be related to stress
deformation of the plastic components of the nebulizer.
The major significance of the low nebulizer outputs in the
experiments with added sulfur and nitrogen oxides present in the
gas stream is that the input concentration of phosphate ion into
the gas stream was so low (equivalent to 6 ppm of a hypothetical
gaseous species) that phosphate ion was hardly more than a trace
species in the flue gas stream. This resulted in analytical dif-
ficulties in accurately determining the concentration of phosphate
species that were sampled from the gas stream (especially at 160
and 90 °C, where polymerization and wall losses had drastically
reduced the input phosphate concentration), and it complicated
the observation and interpretation of the recombination processes
that occurred at 160 and 90 °C.
55. Waggarman, W. H. Phosphoric Acid, Phosphates, and Phosphatic
Fertilizers. Reinhold Publishing Corporation, New York, New
York, 1952. 683 pp.
83
-------
Analytical Procedures
In the absence of added sulfur oxides and nitrogen oxides
in the gas stream, the flue gas was sampled near the outlet of
the 650 °C zone, the 370 °C zone, the ESP (160 °C), and the 90 °C
zone. In most of the experiments, the gas stream was sampled
through a particulate filter and then through a bubbler of 0.1 N
sulfuric acid. In the gas samples collected at 650 or 370 °C,
the filter was a heated plug of quartz wool. In the samples col-
lected at 160 or 90 °C, the filter was a porous Teflon disc (2- to
5-ym porosity) mounted in a heated Pyrex holder. The material
collected on a filter or in the bubbler was analyzed for ammonium
ion, orthophosphate ion, and condensed phosphate species. In
other experiments, the gas stream was sampled for nitrogen oxides.
In a few experiments, the gas stream was sampled into a bubbler
of a sodium bicarbonate-sodium carbonate buffer solution, and the
solution was subsequently analyzed for anions by ion chromatog-
raphy. (The bicarbonate-carbonate buffer solution was used as
the collection medium because this buffer solution is the eluent
commonly used in the determination of anions by ion chromatog-
raphy.)
During the experiments with reactive sulfur and nitrogen
oxides present in the gas stream, the flue gas was sampled at 650,
160, and 90 °C and was analyzed for background sulfur dioxide,
sulfur trioxide, nitric oxide, and nitrogen dioxide when LPA-445
was not being injected into the gas stream. Then the flue gas
was analyzed for ammonia, ammonium ion, orthophosphate ion, and
condensed phosphate species during injection of LPA-445. The
Teflon filters used in the sampling lines at 160 and 90 °C were
washed with distilled water and the filter washes analyzed for
anions by ion chromatography, specifically for orthophosphate and
sulfate ions.
During the course of the investigation of LPA-445, two dif-
ferent colorimetric methods were used to analyze the phosphate
species sampled from the gas stream. In the absence of reactive
gases in the flue gas stream, orthophosphate and condensed phos-
phate ions were analyzed spectrophotometrically by the vanado-
molybdophosphoric acid colorimetric method of Kitson and Mellon.20
Orthophosphate ion was determined directly, whereas condensed
phosphate ions were first hydrolyzed to orthophosphate ion by
boiling in a dilute acid solution and were then determined colori-
metrically. Due to the low concentrations of orthophosphate and
condensed phosphate ions found in the presence of reactive gases
in the flue gas stream, these species were analyzed by the
84
-------
"stannous chloride" colorimetric method.56 This method is much
more sensitive and has minimum detection limits more than 60 times
lower than the vanadomolybdophosphoric acid method.
The analytical methods for sulfur dioxide, sulfur trioxide,
nitrogen oxides, and ammonia (or ammonium ion) have been described
in detail and referenced earlier in this report, and a description
of these methods will not be repeated here.
In one set of experiments the gas stream was sampled and
analyzed for phosphine (PHs) at 160 and 90 °C in the absence of
added reactive gases and at 650 °C in the presence of added sul-
fur and nitrogen oxides to the gas stream. The method employed
for the determination of phosphine was the colorimetric method of
Dechant, Sanders, and Graul,57 which is accurate in the range of
0.1 to 1 ppm of phosphine. The method consists of sampling the
gas stream through a bubbler containing silver diethyldithio-
carbamate reagent. Phosphine reacts quantitatively with this
yellow reagent to form an orange-red complex, which is then
analyzed spectrophotometrically.
Experimental Results
Injection of Diammonium Hydrogen Phosphate (as LPA-445) into Flue
Gas Containing No Added Sulfur Oxides or Nitrogen Oxides—
LPA-445 was injected into the flue gas train at 650 °C. The
flue gas was composed of only nitrogen, oxygen, carbon dioxide,
and water vapor; it contained no added oxides of sulfur or nitro-
gen. The gas stream was sampled at 650 °C to determine the
thermal decomposition products of diammonium hydrogen phosphate
and was sampled near the outlets of the 370 °C zone, the ESP
(160 °C), and the 90 °C zone to determine the recombination pro-
ducts of the decomposition fragments.
Thermal decomposition studies—The analytical results of
the gas samples taken from the 650 °C zone are given in Tables 24
and 25. Table 24 compares the total amounts of ammonium and phos-
phate ions that were collected with the amounts injected.
Table 25 gives the quantities of ammonium, orthophosphate, and
condensed phosphate found on the filters and in the bubblers.
56. Phosphate/Stannous Chloride Method. In: Standard Methods
for the Examination of Water and Wastewater, 14th ed., M. C.
Rand, A. E. Greenberg, and M. J. Taras, Eds. American
Public Health Association, Washington, D. C., 1975.
pp 479-480.
57. Dechant, R., G. Sanders, and R. Graul. Determination of
Phosphine in Air. Am. Ind. Hyg. Assoc. J., 27:75-79, 1966.
85
-------
TABLE 24. RECOVERY OF DECOMPOSITION PRODUCTS
OF LPA-445 AT 650 °C IN FLUE GAS CONTAINING
NO ADDED SULFUR OR NITROGEN OXIDES
Concentration, ymol/L*
Injectedt
Sample
No.
1
2
3
4
5
6
7
8
9
Av
NHi» +
0.748
0.415
2.809
2.809
1.511
1.511
1.397
1.397
1.397
1.666
Total
P0t»~3
0.325
0.456
1.493
1.493
0.494
0.494
0.394
0.394
0.394
0.660
Observed
NHi» +
0.607
1.199
2.273
1.987
1.587
1.141
1.522
1.410
1.431
1.462
Total
PO,*"3
0.211
0.208
0.891
0.632
0.387
0.097
0.206
0.210
0.258
0.344
Recovery, %
NHn +
81
85
81
71
105
76
109
101
103
88
Total
po.,-3
65
46
60
42
78
20
52
53
65
52
* Expressed for 25 °C. To convert to parts per mil-
lion on the hypothetical volume basis, multiply by
24.4.
t Based on
and POi,"3 lost from nebulizer.
TABLE 25. DISTRIBUTION OF DECOMPOSITION PRODUCTS
OF LPA-445 AT 650 °C IN FLUE GAS CONTAINING
NO ADDED SULFUR OR NITROGEN OXIDES
Concentration, ymol/L*
Sample
No.
1
2
3
4
5
6
7
8
9
Av
Collected on
NH4 +
0.034
0.073
0.347
0.101
0.053
0.058
0.162
0.164
0.121
0.124
Ortho
PO.T3
0.060
0.115
0.389
0.240
0.079
0.037
0.051
0.058
0.075
0.123
filter
Condensed
po.*-3
0.025
0.009
0.098
0.050
0.082
0.010
0.071
0.024
0.000
0.041
Collected in
NHi,+
0.573
1.126
1.926
1.886
1.534
1.083
1.360
1.246
1.310
1.338
bubbler
Ortho Condensed
PO^-3 PO,,"3
0.079
0.084
0.380
0.342
0.179
0.032
0.076
0.111
0.121
0.156
0.048
0.000
0.024
0.000
0.047
0.019
0.008
0.016
0.062
0.025
* Expressed for 25 °C. To convert to parts per million
on the hypothetical volume basis, multiply by 24.4.
86
-------
Table 24 shows a marked difference in the recoveries of
ammonium and phosphate ions. For ammonium ion, the average value
was 88%; but for phosphate ion, the average value was only 52%.
This difference in recoveries strongly suggests that when LPA-445
was injected into the gas stream at 650 °C diammonium hydrogen
phosphate decomposed extensively into gaseous ammonia and partic-
ulate phosphate species. The low recovery of the phosphate
species can be attributed chiefly to wall losses of phosphate
particles due to either impingement and adsorption or to settling
out of the particles. The probability that wall losses occurred
is enhanced by the extensive polymerization of orthophosphate
ions to larger, heavier polyphosphate species that was observed
at 650 °C (see below). In addition, condensed phosphate ions are
quite reactive toward quartz1*9 at the high temperature of the
650 °C reaction zone, and additional losses of phosphate could be
attributed to such reactions of these species with the quartz
walls of the reaction zone and sampling port.
Table 25 shows that, on the average, about 90% of the ammo-
nium ions was collected in the bubbler, a result also indicating
that extensive decomposition of the diammonium phosphate to
ammonia gas occurred. For phosphate ions, roughly 50% was found
either on the filter or in the bubbler. This result suggests that
extensive volatilization of phosphate, perhaps to phosphorus
pentoxide, also occurred. On the other hand, the presence of a
significant fraction of the phosphate on the filter suggests that
part of the phosphate remained as particulate. The mole ratio of
ammonium ion to total phosphate on the filter was approximately
0.9. This ratio suggests that the filter may have collected an
ammonium phosphate solid with a mole ratio of ammonium ion to
phosphate of 1:1. The stabilities of ammonium orthophosphates
discussed, however, make the existence of an ammonium phosphate
solid appear improbable at 650 °C. Perhaps gaseous compounds were
merely adsorbed on the quartz wool in a ratio suggesting a stoi-
chiometric compound.
The decomposition studies at 650 °C also led to the follow-
ing observations:
» Approximately 25% of the original orthophosphate was
converted to condensed phosphate (see Table 26).
» Negligible quantities of ammonia were oxidized to
nitrogen oxides. The total concentration of nitro-
gen oxides was about 1% of the concentration of
ammonium ion collected on the filter and in the
bubbler.
Recombination studies--The same species were found at 370,
160, and 90 °C as were found at 650 °C; that is, ammonium ion,
orthophosphate ion, condensed phosphate ion, and traces of nitro-
gen oxides. The results of the gas sampling measurements at 370,
87
-------
160, and 90 °C are given in Tables 26 and 27. Table 26 is a com-
parison of the total amounts of ammonium and phosphate ions that
were collected with the amounts injected. Table 27 shows the dis-
tribution of the collected ammonium, orthophosphate, and condensed
phosphate ions between the particle filter and the sulfuric acid
bubbler.
TABLE 26. RECOVERY OF RECOMBINATION PRODUCTS OF LPA-445
AT 370, 160, AND 90 °C IN FLUE GAS CONTAINING
NO ADDED SULFUR OR NITROGEN OXIDES
Concentration, ymol/L*
Injectedt
Observed
Recovery , %
Sampling
temp, °C
370
Av
160
(ESP)
Av
90
Av
NH^"1"
2.221
2.221
3.040
3.040
3.040
3.154
3.154
2.838
3.615
1.707
1.622
1.622
1.179
1.949
2.560
2.685
1.511
2.328
2.271
Total
0.912
0.912
1.534
1.534
1.534
1.559
1.559
1.363
1.564
0.544
0.545
0.545
0.334
0.706
1.032
1.207
0.320
0.957
0.879
NH.+
1.443
1.608
2.310
1.727
2.243
2.861
3.022
2.173
1.870
0.946
0.858
0.914
1.065
1.131
1.155
0.603
1.044
0.882
0.921
Total
0.400
0.458
0.853
0.644
1.035
0.881
1.130
0.772
0.213
0.057
0.113
0.079
0.089
0.111
0.170
0.073
0.098
0.136
0.119
NH.+
65
72
76
57
74
91
96
77
52
55
53
56
90
58
45
22
69
38
41
Total
44
• 50
56
42
68
73
73
57
14
11
21
15
27
16
16
6
31
14
14
* Expressed for 25 °C. To convert to parts per million
on the hypothetical volume basis, multiply by 24.4.
t Based on
and POt,"3 lost from nebulizer.
The data in Table 26 show that the recovery of ammonium ion
was greater than the recovery of total phosphate ion at each of
the three collection temperatures. The recovery of ammonium ion
averaged 77, 58, and 41%, respectively, at 370, 160, and 90 °C;
the recovery of total phosphate species averaged 57, 16, and 14%
at the same temperatures. The greater recovery of ammonium ion
88
-------
TABLE 27. DISTRIBUTION OF RECOMBINATION PRODUCTS OF LPA-445
AT 370, 160, AND 90 °C IN FLUE GAS CONTAINING
NO ADDED SULFUR OR NITROGEN OXIDES
Concentration, ymol/L*
Collected on filter
Sampling
temp, °C
370
Av
160
(ESP)
Av
90
Av
NHH +
0.225
0.223
0.241
0.121
0.169
0.454
0.438
0.267
0.190
0.075
0.082
0.069
0.095
0.102
0.251
0.003
0.111
0.142
0.127
Ortho
0.256
0.284
0.386
0.229
0.474
0.545
0.580
0.393
0.140
0.034
0.068
0.042
0.058
0.068
0.122
0.053
0.072
0.100
0.087
Condensed
0.100
0.082
0.187
0.075
0.099
0.170
0.326
0.148
0.049
0.010
0.036
0.028
0.020
,0.029
0.018
0.000
0.011
0.023
0.013
Collected in bubbler
NHlt +
1.218
1.385
2.069
1.606
2.074
2.407
2.584
1.906
1.680
0.871
0.776
0.845
0.970
1.028
0.904
0.600
0.933
0.740
0.794
Ortho
0.031
0.076
0.225
0.282
0.454
0.074
0.172
0.188
0.008
0.000
0.002
0.000
0.000
0.002
0.004
0.000
0.007
0.003
0.004
Condensed
PO^"3
0.012
0.016
0.055
0.058
0.008
0.093
0.052
0.042
0.016
0.013
0.008
0.009
0.011
0.011
0.026
0.020
0.008
0.010
0.016
* Expressed for 25 °C. To convert to parts per million on
the hypothetical volume basis, multiply by 24.4.
relative to phosphate ion at all collection temperatures (includ-
ing 650 °C) seems to indicate, as discussed previously, that the
collected ammonium ion originated from a gaseous species (ammonia)
in the flue gas stream and that the collected phosphate ions orig-
inated from particulate phosphate species. However, the data also
show a general decrease in the recoveries of both ammonium ion
and phosphate ion as the physical location of the sampling point
was farther removed from the conditioning agent injection point.
The decreasing recoveries of phosphate ions can be attributed to
wall losses of particulate phosphate species, but wall losses
alone would be inadequate to explain decreasing recoveries of
gaseous ammonia. Rather, the trend of decreasing recoveries of
ammonium ion indicates that recombination of ammonia with particu-
late phosphate species occurred in the gas stream at the lower
89
-------
temperatures. The decreasing recoveries of ammonium ion at the
lower temperatures can thus be partially attributed to wall losses
of particulate ammonium phosphate salts.
The data in Table 27 show that on the average nearly 90% of
the ammonium ion collected at each temperature was found in the
bubbler solution. This indicates that the source of the ammonium
ion collected in the bubbler was gaseous ammonia. This ammonia
gas presumably originated from two separate sources: (1) the
original decomposition of diammonium hydrogen phosphate, or (2)
the injection of excess ammonia from the nebulizer.
At 370 °C, approximately 70% of the total collected phos-
phate was found on the particle filter. This percentage increased
to approximately 85% at 160 and 90 °C. These results indicate
that as the temperature was lowered the extent of recombination
of the original decomposition products into solid ammonium phos-
phate salts increased significantly. At all of the sampling tem-
peratures, approximately 75% of the phosphate ions collected on
the particle filters were orthophosphate ions.
The mole ratios of ammonium ion to orthophosphate ion and
ammonium ion to total phosphate ion collected on the Teflon fil-
ters were approximately the same at both 90 and 160 °C, 1.0:1.0
and 1.0:1.5, respectively. This indicated that the same recombi-
nation product was present at both temperatures. Thermodynami-
cally, ammonium dihydrogen phosphate appeared to be the most
probable recombination product at both 90 and 160 °C.
The amounts of phosphate ion found at 90 and 160 °C and the
equilibrium dissociation pressures of monoammonium and diammonium
phosphate at these temperatures indicated that the only species
that would be expected to be found on the particulate filter at
either of these temperatures was monoammonium phosphate at 90 °C.
These data will be used first to argue against the existence of
the diammonium salt at either temperature and then to confirm the
possible existence of the monoammonium salt at 90 °C.
The weight of phosphate ion collected on the particulate
filter at 90 °C corresponded to an average gas stream concentra-
tion of 3.0 ppm of phosphate as a hypothetical vapor. At 160 °C
the weight collected corresponded to an average of 4.0 ppm of
phosphate. The dissociation pressures of diammonium hydrogen
phosphate at 90 and 160 °C as discussed earlier are 4.57 and
257 mmHg, respectively. These dissociation pressures correspond
to 6 000 and 340 000 ppm of NH3 at 90 and 160 °C, respectively.
Thus, the amounts of phosphate collected at 90 and 160 °C were
totally insufficient to produce the indicated dissociation pres-
sures of diammonium hydrogen phosphate.
90
-------
The dissociation pressure of monoammonium dihydrogen phos-
phate is 0.05 mmHg (66 ppm) at 125 °C.51* The value is not known
at 90 °C, but it is probably low enough to be consistent with the
occurrence of the raonoammonium salt at 90 °C. The dissociation
pressure at 160 °C is also unknown, but it is certainly too high
to be consistent with the occurrence of the monoammonium salt at
160 °C. Perhaps the salt is stabilized at this temperature by
some process such as adsorption on the particulate filter.
Even at 370 °C, the analyses indicated that orthophosphate
ion was the predominant phosphate species found on the particulate
filter. This observation cannot be explained on a purely thermo-
dynamic basis, and some other mechanism such as adsorption may
have stabilized the orthophosphate at this temperature. There
were insufficient experimental data to resolve the ammonium phos-
phate recombination products found at 370 °C into specific salts.
Phosphine determination—The results showed no indication
of the presence of phosphine in the flue gas samples taken from
the outlet of the electrostatic precipitator and the outlet of
the 90 °C zone.
Injection of Diammonium Hydrogen Phosphate (as LPA-445) into Flue
Gas Containing Added Sulfur Dioxide and Nitrogen Oxides—
LPA-445 was injected into the flue gas near the inlet of
the 650 °C zone. In about half of the experiments, the gas stream
was composed of nitrogen, oxygen, carbon dioxide, and water vapor
plus approximately 600 ppm of sulfur dioxide and a background
level of approximately 2 ppm of sulfur trioxide. In the remainder
of the experiments, approximately 100 ppm of nitrogen dioxide and
1000 ppm of nitric oxide were additionally present in the flue
gas stream. The gas stream was sampled at 650 °C to determine
the thermal degradation products of diammonium hydrogen phosphate
and was sampled near the outlets of the 370 °C zone, the ESP
(160 °C), and the 90 °C zone to determine the recombination pro-
ducts of the decomposition fragments.
Thermal decomposition studies—The results of the sampling
measurements at 650 °C are listed in Tables 28 and 29. Table 28
compares the total amounts of ammonium and phosphate ions that
were collected with the amounts injected. Table 29 shows the
quantities of ammonium, orthophosphate, and condensed phosphate
ions found on the filters and in the bubblers.
On the average, approximately two-thirds of the injected
ammonium ion and a little more than one-half of the injected phos-
phate ion were recovered at 650 °C, the remainder of the injected
ions presumably being lost on the walls of the 650 °C reaction
zone and sampling port. About 86% of the ammonium ions were col-
lected in the bubbler, whereas 75% of the total phosphate ions
was found on the filter and 25% in the bubbler. These results
indicate that at 650 °C the diammonium hydrogen phosphate in
91
-------
TABLE 28. RECOVERY OF DECOMPOSITION PRODUCTS OF LPA-445
AT 650 °C IN FLUE GAS CONTAINING ADDED SULFUR
DIOXIDE BUT NO ADDED NITROGEN OXIDES
Concentration, ymol/L*
Injectedt
Observed
Recovery, %
Sample
No.
1
2
3
Av
NHi,"1"
0.625
0.625
0.625
0.625
Total
0.231
0.231
0.231
0.231
NH.+
0.412
0.371
0.408
0.397
Total
0.117
0.136
0.128
0.127
NH.+
66
59
65
63
Total
51
59
55
55
* Expressed for 25 °C. To convert to parts per mil-
lion on the hypothetical volume basis, multiply by
24.4.
t Based on analyses of NHi» + and POi,"3
nebulizer.
lost from
TABLE 29. DISTRIBUTION OF DECOMPOSITION PRODUCTS OF LPA-445
AT 650 °C IN FLUE GAS CONTAINING ADDED SULFUR
DIOXIDE BUT NO ADDED NITROGEN OXIDES
Concentration, ymol/L*
Collected on filter
Sample
No.
1
2
3
Av
0
0
0
0
NH..+
.051
.056
.063
.057
Ortho
0
0
0
0
.058
.061
.075
.065
Collected in bubbler
Condensed
POi,"3 NHj,"1"
0
0
0
0
.030
.039
.022
.030
0.
0.
0.
0.
362
315
344
340
Ortho
0.029
0.032
0.024
0.028
Condensed
0
0
0
-t
.004
.007
.004
* Expressed for 25 °C. To convert to parts per million
on the hypothetical volume basis, multiply by 24.4.
t An anomalous analytical result, indicating less total
phosphate than orthophosphate, is not reported.
92
-------
LPA-445 decomposed primarily into gaseous ammonia and particulate
phosphate species, with some volatilization of phosphate, probably
to phosphorus pentoxide, also occurring.
These results were obtained with sulfur dioxide as the only
reactive gas added to the gas stream. They are qualitatively very
similar to the results obtained in the absence of added reactive
gases. Quantitatively, however, much larger recoveries of ammo-
nium ion were found in the absence of added reactive gases
(approximately 88%) than in the presence of added sulfur dioxide,
and a larger fraction of the phosphate species was found in the
bubblers (approximately 50%). These results indicate that more
extensive thermal decomposition of diammonium hydrogen phosphate
and more extensive volatilization of phosphate ion occurred in
the absence of reactive gases than with sulfur dioxide added to
the gas stream. No satisfactory explanation for this observation,
however, can be offered.
Recombination studies—The results of the gas sampling
measurements at 160 and 90 °C are given in Tables 30, 31, and 32.
Table 30 is a comparison of the amounts of ammonium and phosphate
ions that, were injected with the amounts that were collected.
Table 31 shows the distribution of the collected ammonium, ortho-
phosphate, and condensed phosphate ions between the particle
filter and the sulfuric acid bubbler. Table 32 gives the results
of the iop chromatography analyses of the filter washes of the
flue gas sampled from the 160 and 90 °C reaction zones with all
reactive gases present in the flue gas stream.
With both sulfur oxides and nitrogen oxides added to the
flue gas stream, the same general trends in the recovery and dis-
tribution of ammonium ion and phosphate species at 160 and 90 °C
were observed as those trends observed in the absence of added
reactive gases to the flue gas stream. These results indicate
that gaseous ammonia and particulate phosphate species in the flue
gas stream recombined into ammonium phosphate salts at 160 and
90 °C, with the extent of recombination increasing as the tempera-
ture was lowered. From the thermodynamic considerations discussed
earlier in this section, the most probable recombination product
at 160 and 90 °C is monoammonium dihydrogen phosphate (NHi»H2POit) .
From the ion chromatographic analyses of the filter washes,
sulfate ion and nitrate ion were found to be present at concentra-
tions comparable to or greater than the concentrations of ortho-
phosphate. Thus, in the presence of added oxides of sulfur and
nitrogen, there appears to be a competition for the available
ammonia in the gas stream during recombination reactions at 160
and 90 °C among orthophosphate, sulfate, and nitrate ions. The
background level of sulfur trioxide in the gas stream was the
source of the sulfate species in these experiments. The back-
ground concentration of sulfur trioxide was measured and found to
93
-------
TABLE 30. RECOVERY OF RECOMBINATION PRODUCTS OF LPA-445
AT 160 AND 90 °C IN FLUE GAS CONTAINING SULFUR
DIOXIDE AND SOMETIMES NITROGEN OXIDES
Concentration, yimol/L*
Iniectedt
Sampling
temp, °C
160 (ESP)
(without NOX)
Av
160 (ESP)
(with NOX)
Av
90
(with NOX)
Av
NH4 +
0.707
1.006
0.965
0.893
1.284
0.907
1.096
1.729
0.907
1.057
1.231
Total
0.141
0.100
0.158
0.133
0.264
0.196
0.230
0.614
0.196
0.365
0.392
Observed
NHif +
0.278
0.217
0.395
0.297
0.194
0.085
0.140
0.099
0.090
0.046
0.078
Total
PO.T3
0.020
0.016
0.023
0.020
0.018
0.016
0.017
0.020
0.027
0.028
0.025
Recovery , %
NH,+
39
22
41
34
15
9
13
6
10
4
7
Total
po,,-3
14
16
14
15
7
8
8
3
14
8
6
* Expressed for 25 °C. To convert to parts per mil-
lion on the hypothetical volume basis, multiply by
24.4.
t Based on NHi,+ and POt*"3 lost from nebulizer.
average approximately 2 ppm (0.1 ymol/L, expressed for 25 °C)
with an average concentration of 600 ppm sulfur dioxide in the
gas stream. The concentration of nitrate ion in the gas stream,
on the other hand, was expected to be on the trace level. The
source of the nitrate ion is assumed to be the oxidation of ammo-
nia by the nitrogen oxides added to the gas stream, as was dis-
cussed earlier in this report in the section on the investigation
of ammonia. However, as was pointed out earlier in this section,
the average input concentration of phosphate ion into the gas
stream at 650 °C in these experiments was so small (0.25 ymol/L,
only two-and-a-haIf times greater than the average sulfur trioxide
background level) and the apparent wall losses were so great that
there is little significance to the observed competition between
the phosphate ion and the sulfate and nitrate ions for the avail-
able ammonia in the gas stream in these experiments. In the
presence of a large excess of phosphate ion relative to sulfate
and nitrate ions, the competition might be negligible.
Reaction of ammonia with nitrogen oxides—When nitrogen
oxides were present in the gas stream during the injection of
LPA-445, the ammonia recoveries at 160 and 90 °C were
94
-------
TABLE 31. DISTRIBUTION OF RECOMBINATION PRODUCTS OF LPA-445
AT 160 AND 90 °C IN FLUE GAS CONTAINING SULFUR
DIOXIDE AND SOMETIMES NITROGEN OXIDES
Concentration, ymol/L*
Collected on filter
Sampling
temp, °C
160 (ESP)
(without NOX)
Av
160 (ESP)
(with NOX)
Av
90
(with NOX)
Av
NHi,+
0.201
0.062
0.197
0.153
0.038
0.026
0.032
0.037
0.034
0.009
0.027
Ortho
po,,-3
0.011
0.007
0.014
0.011
0.009
0.010
0.010
0.016
0.020
0.017
0.018
Condensed
PO,,"3
0.001
0.005
0.003
0.003
0.002
0.002
0.001
Collected in bubbler
NHi,+
0.077
0.155
0.197
0.143
0.156
0.059
0.108
0.062
0.056
0.036
0.051
Ortho
PO,,'3
0.005
0.001
0.000
0.002
0.005
0.003
0.004
0.003
0.004
0.006
0.004
Condensed
P04-3
0.003
0.003
0.006
0.004
0.004
0.003
0.004
0.000
0.004
0.004
0.002
* Expressed for 25 °C. To convert to parts per million on the
hypothetical volume basis, multiply by 24.4.
TABLE 32. ION CHROMATOGRAPHIC ANALYSIS OF ANIONIC SPECIES
IN FILTER WASHES OF SAMPLES OF FLUE GAS CONTAINING
SULFUR DIOXIDE AND NITROGEN OXIDES
Gas stream concentration, ymol/L*
Sampling
temp , ° C
160
(ESP)
90
so,r2
0.051
0.012
0.025
0.017
N03~
0.009
0.013
0.005
0.007
Ortho
P04~3
0.010
0.010
0.017
0.019
Ortho
P04-3t
0.009
0.010
0.016
0.020
* Expressed for 25 °C. To convert to parts
per million on the hypothetical volume
basis, multiply by 24.4.
t Results obtained by the stannous chloride
colorimetric method for comparison with
the ion chromatography results.
95
-------
significantly less than those in the absence of added nitrogen
oxides. Table 30 shows that the average recovery of ammonium ion
was 34% at 160 °C with sulfur dioxide as the only reactive oxide
added to the gas stream. This table shows further that with both
nitrogen oxides and sulfur oxides added to the gas stream the
recovery of ammonium ion averaged only 13%, less than one-half
the recovery found at 160 °C with no nitrogen oxides present.
This observation is consistent with the gas stream reaction of
nitrogen oxides with ammonia at 650 °C observed during the inves-
tigations of ammonia and ammonium sulfate.
Phosphine determination—No measurable concentrations of
phosphine could be found in the flue gas sampled from the 650 °C
reaction zone during the injection of LPA-445 into flue gas con-
taining added oxides of sulfur and nitrogen.
Summary and Discussion of the Results
In the absence and in the presence of added sulfur oxides
to the flue gas stream, the primary thermal decomposition products
of the LPA-445 formulation were ammonia, ammonium ion, ortho-
phosphate ion, and condensed phosphate ion. The condensed phos-
phate species were not further characterized. Diammonium hydrogen
phosphate, the major component of LPA-445, appeared to decompose
thermally into gaseous ammonia and particulate phosphate species,
with some volatilization of phosphate also occurring. The extent
of both thermal decomposition and phosphate volatilization was
found to be less with sulfur dioxide added to the gas stream than
in the absence of sulfur dioxide; but no logical explanation for
this observation is apparent.
The major recombination products found at 160 and 90 °C in
the absence of added reactive oxides in the gas stream were ammo-
nium phosphate salts. While the experimental data obtained are
not subject to a definitive interpretation, the most probable
recombination product at 160 and 90 °C appeared to be monoammonium
dihydrogen phosphate (NHnHaPCK). With added oxides of nitrogen
and sulfur present in the gas stream, the major recombination pro-
ducts found at 160 and 90 °C were ammonium phosphate, ammonium
sulfate, and ammonium nitrate salts. The apparent competition
between phosphate ion and sulfate and nitrate ions for the avail-
able ammonia in the gas stream was probably exaggerated in these
experiments, however, due to the low input concentrations of phos-
phate ion injected into the gas stream.
Large decreases in ammonia concentrations and the appear-
ance of ammonium nitrate at 160 and 90 °C when nitrogen oxides
were added to the flue gas stream strongly indicate that gas
stream reactions occurred between ammonia and one or both of the
added nitrogen oxides.
96
-------
No phosphine could be detected, either in the presence or
absence of added reactive gases, when LPA-445 was injected and
the gas stream was specifically sampled and analyzed for phos-
phine.
97
-------
REFERENCES
1. Dismukes, E. B. A Study of Resistivity and Conditioning of
Fly Ash. EPA-R2-72-087, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1972.
138 pp.
2. Sparks, L. E. Electrostatic Precipitator Options for
Collection of High Resistivity Ash. In: Symposium on
Particulate Control in Energy Processes. EPA-600/7-76-010,
U. S. Environmental Protection Agency, Research Triangle
Park, North Carolina, 1976. pp 127-141.
3. Dismukes, E. B. Conditioning of Fly Ash with Sulfur Triox-
ide and Ammonia. EPA-600/2-75-015, U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina,
1975. 157 pp.
4. Watson, K. S. Australian Experience with Flue Gas Condi-
tioning. In: Symposium on Particulate Control in Energy
Processes. EPA-600/7-76-010, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1976.
pp 189-216.
5. Dismukes, E. B. Conditioning of 'Fly Ash with Ammonia. J.
Air Pollut. Contr. Assoc., 25:152-156, 1975.
6. Potter, E. C., and C. A. J. Paulson. Improvement of Elec-
trostatic Precipitator Performance by Carrier Gas Additives
and Its Graphical Assessment Using an Extended Deutsch
Equation. Chem. Ind . , 1974:532-533.
7. Bickelhaupt, R. E. Sodium Conditioning to Reduce Fly Ash
Resistivity. EPA-650/2-74-092, U. S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina, 1974.
10 pp.
8. Selle, S. J., and L. L. Hess. Factors Affecting ESP Per-
formance on Western Coals and Experience with North Dakota
Lignites. In: Symposium on Particulate Control in Energy
Processes. EPA-600/7-76-010, U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1976.
pp 105-125.
98
-------
9. Dismukes, E. B., and J. P. Gooch. Fly Ash Conditioning
with Sulfur Trioxide. EPA-600/2-77-242, U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina,
1977. 65 pp.
10. Magee, P. N., and J. M. Barnes. Carcinogenic Nitroso Com-
pounds. In: Advances in Cancer Research, Vol. 10,
A. Haddow and S. Weinhouse, Eds. Academic Press, New York,
New York, pp 163-246.
11. Ames, B., J. McCann, and E. Yamasaki. Methods for Detecting
Carcinogens and Mutagens with the Salmonella/Mammalian-
Microsome Mutagenicity Test. Mutation Res., 31:347-364,
1975.
12. Bennett, R. P., and M. J. O'Connor. Method of Conditioning
Flue Gas to Electrostatic Precipitator. U. S. Patent
4 043 768, August 23, 1977. Assigned to Apollo Chemical
Corporation, Whippany, New Jersey.
13. Bennett, R. P., M. J. O'Connor, A. E. Kober, and I. Kukin.
Method of Conditioning Flue Gas to Electrostatic Precipita-
tor. U. S. Patent 4 042 348, August 16, 1977. Assigned to
Apollo Chemical Corporation, Whippany, New Jersey.
14. Bennett, R. P., and A. E. Kober. Method of Conditioning
Flue Gas. U. S. Patent 4 113 447, September 12, 1978.
Assigned to Apollo Chemical Corporation, Whippany, New
Jersey.
15. Kober, A. E., and I. Kukin. Method of Agglomerating Parti-
cles in Gas Stream. U. S. Patent 4 070 162, January 24,
1978. Assigned to Apollo Chemical Corporation, Whippany,
New Jersey.
16. Small, H., T. S. Stevens, and W. C. Bauman. Novel Ion
Exchange Chromatographic Method Using Conductimetric Detec-
tion. Anal. Chem., 47:1801-1809, 1975.
17. Harwood, J. E. , and A. L. Kiihn. A Colorimetric Method for
Ammonia in Natural Waters. Water Res., 4:805-811, 1970.
18. Fritz, J. S., and S. S. Yamamura. Rapid Microtitration of
Sulfate. Anal. Chem., 27:1461-1464, 1955.
19. Sorum, C. H. Introduction to Semimicro Qualitative Analy-
sis. Prentice Hall, Inc., Englewood Cliffs, New Jersey,
1967. 277 pp.
20. Kitson, R. E., and M. G. Mellon. Colorimetric Determination
of Phosphorus as Molybdivanodophosphoric Acid. Ind. Eng.
Chem., 16:379-383, 1944.
99
-------
21. Watt, G. W., and J. D. Chrisp. Spectrophotometric Method
for Determination of Urea. Anal. Chem., 26:452-453, 1954.
22. Oglesby, S., and G. B. Nichols. A Manual of Electrostatic
Precipitator Technology, Part 1—Fundamentals. PB-196 380,
National Air Pollution Control Administration, Cincinnati,
Ohio, 1970. 322 pp.
23. Washburn, E. W., Ed. International Critical Tables of
Numerical Data, Physics, Chemistry, and Technology. McGraw-
Hill Book Company, Inc., New York, New York, 1928.
Volume III, p 304.
24. Goks^yr, H., and K. Ross. The Determination of Sulfur Tri-
oxide in Flue Gases. J. Inst. Fuel, 35:177-179, 1962.
25. Maddalone, R. L. Guidelines for Combustion Source Sulfuric
Acid Emission Measurements. TRW Document No. 28055-6005-
RU-00 (prepared under U. S. Environmental Protection Agency
Contract No. 68-02-2165). TRW Defense and Space Systems
Group, Redondo Beach, California, 1977. 14 pp.
26. Liu, B. Y. H., and K. W. Lee. Efficiency of Membrane and
Nuclepore Filters for Submicrometer Aerosols. Environ. Sci.
Technol., 10:345-350, 1976.
27. Banchero, J. T., and F. H. Verhoff. Evaluation and Inter-
pretation of the Vapour Pressure Data for Sulphuric Acid
Aqueous Solutions with Application to Flue Gas Dewpoints.
J. Inst. Fuel, 48:76-80, 1975.
28. St. Clair, H. W. Vapor Pressure and Thermodynamic Proper-
ties of Ammonium Sulphites. In: Report of Investigations
3339. U. S. Mureau of Mines, Washington, D. C., 1937.
pp 19-29.
29. Rosser, W. A., and H. Wise. Gas-Phase Oxidation of Ammonia
by Nitrogen Dioxide. J. Chem. Phys., 25:1078-1979, 1956.
30. Gehlen, H. Reactions and Properties of Nitric Oxide and
Its Compounds. III. The Reaction Between Nitric Oxide and
the Alkali Salts of Nitrodisulfonic Acid. Chem. Ber., B66:
292-297, 1933.
31. Wise, H., and M. F. Freeh. Kinetics of Oxidation of Ammonia
by Nitric Oxide. J. Chem. Phys., 22:1463-1464, 1954.
32. Cohn, J., D. R. Steele, and H. C. Anderson. Selective
Removal of Nitrogen Oxides from Oxygen-Containing Gases
Especially from Nitric Acid Production from Ammonia. U. S.
Patent 2 975 025, March 14, 1961. Assigned to, Engelhard
Industries, Inc.
100
-------
33. Anderson, H. C. , and C. D. Keith. Removal of Nitric Oxide
from Tail Gas after Oxidation of Ammonia to Nitric Acid.
U. S. Patent 3 008 796, November 14, 1961. Assigned to
Engelhard Industries, Inc.
34. Nonnenmacher, H., and K. Katte. Selective Removal of Oxides
of Nitrogen from Gas Mixtures Containing Oxygen. U. S.
Patent 3 279 884, October 18, 1966. Assigned to Badische-
Anilin- & Soda-Fabrik, AG, Germany.
35. Recommended Method of Analysis for Nitrogen Dioxide Content
of the Atmosphere (Greiss-Saltzman Reaction). In: Methods
of Air Sampling and Analysis, M. Katz, Ed. American Public
Health Association, Washington, D. C., 1977. pp 527-534.
36. Tentative Method of Analysis for Nitric Oxide Content of
the Atmosphere. In: Methods of Air Sampling and Analysis.
M. Katz, Ed. American Public Health Association, Washington,
D. C., 1977. pp 524-526.
37. Tentative Method of Analysis for Total Nitrogen Oxides as
Nitrate (Phenoldisulfonic Acid Method). In: Methods of Air
Sampling and Analysis, M. Katz, Ed. American Public Health
Association, Washington, D. C., 1977. pp 534-538.
38. Stull, D. R., and H. Prophet, Eds. JANAF Thermochemical
Tables, 2nd ed. National Brueau of Standards, Washington,
D. C., 1971.
39. Bedford, G., and J. H. Thomas. Reaction Between Ammonia and
Nitrogen Dioxide. J. Chem. Soc., Faraday Trans. 1, 68:2163-
2170, 1972.
40. Taylor, D. G., Ed. NIOSH Manual of Analytical Methods,
Vol. 1, 2nd ed. National Institute for Occupational Safety
and Health, Cincinnati, Ohio, April, 1977. pp 221-1 to
221-9.
41. Blake, D. Operation and Service Manual, Source Assessment
Sampling System. AEROTHERM Report UM-77-80, ACUREX Corpora-
tion, Mountain View, California, March, 1977.
42. Fine, D. H. , D. P. Rounbehler , E. Sawicki, and K. Krost.
Determination of Dimethylnitrosamine in Air and Water by
Thermal Energy Analysis: Validation of Analytical Proce-
dures. Environ. Sci. Technol., 11:577-580, 1977.
43. Taylor, H. A., and E. E. Juterbock. The Thermal Decomposi-
tion of Triethylamine. J. Phys. Chem., 39:1103-1110, 1935.
101
-------
44. Bateman, L. C., E. D. Hughes, and C. K. Ingold. Molecular
Compounds Between Amines and Sulfur Dioxide. A Comment on
Jander's Theory of Ionic Reactions in Sulfur Dioxide. J.
Chem. Soc., 1944:243-247.
45. West, P. W., and G. C. Gaeke. Fixation of Sulfur Dioxide
as Sulfitomercurate III and Subsequent Colorimetric Deter-
mination. Anal. Chem., 28:1816-1818, 1956.
46. Kelley, K. K., C. H. Shomate, F. E. Young, B. F. Naylor,
A. E. Salo, and E. H. Huffman. Thermodynamic Properties of
Ammonium and Potassium Alums and Related Substances with
Reference to Extraction of Alumina from Clay and Alunite.
Technical Paper 688. U. S. Bureau of Mines, Washington,
D. C., 1946. pp 66-69.
47. Scott, W. D., and F. C. R. Cattell. Vapor Pressure of Ammo-
nium Sulfates. Atmos. Environ., 13:307-317, 1979.
48. Halstead, W. D. Thermal Decomposition of Ammonium Sulphate.
J. Appl. Chem., 20:129-132, 1970.
49. Van Wazer, J. R. Phosphoric Acids and Phosphates. In:
Kirk^Othmer Encyclopedia of Chemical Technology, 2nd ed.,
Vol. 15, H. F. Mark, J. J. McKetta, and D. F. Othmer, Eds.
Interscience Publishers, New York, New York, 1968. pp 232-
276.
50. Cotton, F. A., and G. Wilkinson. Advanced Inorganic
Chemistry, 2nd ed. Interscience Publishers, New York, New
York, 1967. 1136 pp.
51. Erdey, L. , S. Gal, and G. Liptay. Thermoanalytical Proper-
ties of Analytical-Grade Reagents—Ammonium Salts. Talanta,
11:913-940, 1964.
52. Toy, A. D. F. Phosphorus. In: Comprehensive Inorganic
Chemistry, Vol. 2, J. E. Bailar, H. J. Emeleus, R. Nyholm,
A. F. Trotman-Dickenson, Eds. Pergamon Press, Elmsford, New
York, 1973. pp 389-545.
53. Passille, A. Dissociation of Ammonium Phosphates. Compt.
rend., 199:356-358, 1934.
54. Warren, T. E. Dissociation Pressures of Ammonium Ortho-
phosphates. J. Am. Chem. Soc., 49:1904-1908, 1927.
55. Waggarman, W. H. Phosphoric Acid, Phosphates, and Phos-
phatic Fertilizers. Reinhold Publishing Corporation, New
York, New York, 1952. 683 pp.
102
-------
56. Phosphate/Stannous Chloride Method. In: Standard Methods
for the Examination of Water and Wastewater, 14th ed., M. C,
Rand, A. E. Greenberg and M. J. Taras, Eds. American
Public Health Association, Washington, D. C., 1975.
pp 479-480.
57. Dechant, R., G. Sanders, and R. Graul. Determination of
Phosphine in Air. Am. Ind. Hyg. Assoc. J., 27:75-79, 1966.
103
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-179
3. RECIPIENT'S ACCESSION* NO.
4. TITLE AND SUBTITLE
Analysis of Thermal Decomposition Products of Flue
Gas Conditioning Agents
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7.AUTHORIS)
R.B.Spafford, E. B.Dismukes, and H.K.Dillon
8. PERFORMING ORGANIZATION REPORT NO.
SORI-EAS-79-267
(Project 3832-XXVII-F)
9. PERFORMING OROANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2200
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 AN.D PERIOD COVERED
Final; 1/77 - 3/79
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES ffiRL-RTP project officer is Leslie E. Sparks, Mail Drop 61,
919/541-2925.
16. ABSTRACT
The report gives results of a study of reactions of several flue gas condi-
tioning agents in a laboratory-scale facility simulating conditions in the flue gas
train of a coal-burning power plant. Primary purposes of the study were to charac-
terize the chemical species resulting from adding conditioning agents to flue gas and
to identify potentially hazardous chemical species originating from the agents that
may be emitted into the environment. The compounds investigated were sulfur tri-
oxide, ammonia, triethylamine, sodium carbonate, ammonium sulfate, and diammo-
nium hydrogen phosphate. The predominant types of reactions observed in these
experiments were thermal decomposition at high temperatures, recombination of
decomposition fragments at lower temperatures, and reactions with normal compo-
nents of the flue gas. The only significant environmental threat of any product iden-
tified during this study was the formation of N-nitrosodiethylamine as the result of
injecting triethylamine into the flue gas. This potent carcinogen was found in trace
amounts when triethylamine was injected at 160 C.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Treatment
Flue Gases
Coal
Pyrolysis
Analyzing
Sulfur Trioxide
Ammonia
Tertiary Amines
Sodium Carbonates
Ammonium Sulfate
Pollution Control
Stationary Sources
Conditioning Agents
N-Nitrosodiethylamine
Diammonium Hydrogen
Phosphate
13 B
14 B
21B
21D
07D
07B
07C
8. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
114
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
104
------- |