&EPA
United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-79-168
August 1979
Research and Development
Evaluation of
Methodology and
Prototype to Measure
Atmospheric Sulfuric
Acid
Final 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 ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-168
August 1979
EVALUATION OF METHODOLOGY AND PROTOTYPE
TO MEASURE ATMOSPHERIC SULFURIC ACID
Final Report
by
R. E. Snyder
M. E. Tonkin
A. M. McKissick
Atlantic Research Corporation
Alexandria, Virginia 22314
Contract No. 68-02-2467
Project Officer
Kenneth T. Krost
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences 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 endorse-
ment or recommendation for use.
11
-------
ABSTRACT
The objective of this study was to propose a promising methodology for
the selective analysis of ambient sulfuric acid aerosol and to determine
the feasibility of the method. The methodology was to cover the range of
o
0.25 to 50 yg/m . A review of the literature indicated that current analy-
tical deficiencies revolved around an inability to selectively analyze
the collected acid aerosol in the presence of interfering sulfates, and the
loss of sulfuric acid during sample acquisition due to the formation of
artifacts on the collection surface. Since little could be done to isolate
interfering species from the acid aerosol during sample collection, a pro-
mising approach appeared to be that of reacting the acid aerosol prior to
sample collection to form a stable adduct that could be selectively analyzed
in the presence of interfering sulfates.
Theoretical consideration of reaction kinetics and aerosol sampling
methodologies suggested that use of a volatile reagent which underwent an
acid-base reaction with sulfuric acid was a promising approach. Interest
then focused on volatile amines and amine derivatives which would react
rapidly in the gas phase with sulfuric acid aerosol to form amine sulfate
and bisulfate salts. Further consideration suggested that the adduct could
be conveniently analyzed if it decomposed to evolve S02 gas at a temperature
below that of interfering sulfates (<200 °C).
A reaction chamber was designed and constructed which allowed the
flowing aerosol sample stream to mix with various fixing reagents. Adducts
thus formed were collected on Millipore Teflon Filters of 5 and 0.5 micro-
meter pore size located at the end of the reaction chamber. A Teflon thermal
decomposition cell was constructed in which the fixed acid samples were
heated at 200°C to liberate S02. A valve arrangement connected the cell
to a flame photometric detector (FPD) or a West-Gaeke bubbler which measured
the evolved SCL- It was established with this apparatus that the gas phase
fixation process was very rapid and essentially complete during reaction
chamber residence time. In addition, using this technique, a mass balance
iii
-------
was obtained between predicted aerosol generator concentration (123 yg/m^) and
•3
measured aerosol concentration (111 yg/m ).
Ammonium sulfate, often present in ambient atmosphere was considered
to be the primary positive interference. It was demonstrated that this
interference was eliminated for H2SO, aerosols in the 0.005 to 0.3 micro-
meter size range. However, the decomposition temperatures for various
adducts made from 1.0 to 3.0 micrometer aerosols tended to overlap in some
cases with those of the ammonium sulfate salts. It was determined that
the side reaction between ammonia and sulfuric acid was suppressed on
collection surfaces as the result of forming sulfuric acid during the
collection process.
This report is submitted in fulfillment of Contract No. 68-02-2467
by Atlantic Research Corporation under the sponsorship of the U. S.
Environmental Protection Agency. The report covers the period from September
30, 1976, through November 29, 1978.
iv
-------
CONTENTS
Abstract
Figures vi
Tables viii
1. Introduction 1
2. Conclusions 3
3. Recommendations 4
4. Technical Discussion 5
General Atmospheric Constituents 5
The Sulfuric Acid Measurement Problem 15
Theoretical Development of Research Approach 18
Separation of Interfering Species 33
Summary of Theoretical Development 38
5. Experimental 41
Aerosol Generation 41
Aerosol Collection and Adduct Formation 45
6. Results and Discussion 53
Preliminary Experiments 53
Decomposition Characteristics of Adducts (Unmodified
Thomas Generator - 0.005 to 0.3 ym Droplets) .... 62
Decomposition Characteristic of Adducts (Baird
Generator, 1 to 3 ym Droplets) 74
Characteristics of Adduct Decomposition Gases 84
Modified Thomas Generator 91
Analysis of Concurrent Samples Using the Heated
Sampling Device with NH3 Fixing Reagent 96
Mass Balance 98
Volatilization of H2S04 (Modified Thomas Generator). . 99
Field Tests Ill
Summary 124
References 130
Appendices
A. Reactions in flow systems 135
B. Measurement techniques for sulfuric acid and
sulfuric acid aducts 147
-------
FIGURES
Number Page
1 Total ion current chromatogram 9
2 Three principle size ranges for ambient particle
size distribution 14
3 Filter holder 19
4 Bubbling system 20
5 Impactor 21
6 Filter system with spray jets 25
7 HoS(^4 filter collection (gas or vapor reagent) 25
8 Thomas aerosol generator 43
9 Baird aerosol generator, 44
10 Sample and fixation apparatus 46
11 Modified aerosol collection system 47
12 Revised sample decomposition chamber . 49
13 Composite DTA of potential interferences 54
14 IR spectrum of a) diethylhydroxylamine sulfate,
b) diethylhydroxylamine, and 3) sulfuric acid 56
15 IR spectrum of a) acetaldoxime sulfate, b) acetaldoxime,
and c) sulfuric acid 57
16 (PDA)2S04 and Py2S04 DTA 58
17 Early FPD recordings from samples 59
18 Effect of injecting progressively larger S0_ volumes 60
19 Quantitative FPD response to injected SO™ 61
20 Quantitative response of FPD to SO 61
21 Completeness of fixation by AAO (fixed temperature procedure). . 65
22 Completeness of decomposition of AAO/l^SO^ adduct after
5 min (A) and 15 min (B) 66
23 Typical FPD trace from AAO-treated sample 67
24 AAO-fixed filter portions after heating 5 min at 200°C 68
25 AAO-treated samples, different sampling volumes, 5 min at 200°C. 69
vi
-------
Number Page
26 Pairs of AAO-fixed samples collected simultaneously 70
27 Continuous heating procedure - FPD recordings from
H^SO, and adduct samples 73
28 Relationship of aerosol sample volume and evolved S(>2
from AAO adduct decomposition 83
29 Submicron particles fixed with ammonia ten minutes
after sample collection 116
30 Submicron particles fixed with ammonia during sample collection . 116
31 Decomposition profile of NMF-SO, adduct 117
32 Decomposition profile of AAO-SO, adduct 117
33 Composite decomposition profile - AAO-SO,, NMF-SO,, (NH,)2SO, . . 118
vii
-------
TABLES
Number ]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Organic Components of Industrial Waste Effluents
Approximate Elemental Concentrations Found in
Particulates Collected from Ambient Urban Air
Variation of Sulfuric Acid Content of Atmospheric Aerosols. . .
Comparison of Instruments . .
General Reactions of Sulfuric Acid
Decomposition Temperatures of Ambient Sulfates
Temperature of Initial FPD Response from Adducts
(Unmodified Thomas Generator)
Peak Heights of Sample Pairs Collected Simultaneously
Temperature (°C) of First FPD Response from Sulfur
Gas-Reagent Complexes
Initial H2SO4 Decomposition Temperature by H^SO,
Generation Method
Initial ^§04 Adduct Decomposition Temperatures by
H-SOA Generation Method
Summary of (NH4)2S04 Decomposition Temperatures for
Various Sample Sizes
Initial Response of FPD for (NH^SO, by Deposition Method. . .
Recovery of S02 From AAO/H-SO, Adduct (Dry Air)
Recovery of S02 From AAO/H2SO Adduct (Humidified Air)
Specificity of West-Gaeke Technique for SO-
Turbidimetric Analysis of Raw H-SO,
Sulfuric Acid/AAO Adduct Analysis (LS Filter-
Decomposition of DEA Adduct (S0? Analysis)
Percent Loss of Various Adducts Due to Reaction With Ammonia. .
Analysis of Concurrent Samples Using Ammonia Fixing Reagent . .
Analysis of Concurrent Samples Using Diethylamine Fixing Reagent
Page
6
10
13
22
27
34
63
70
72
75
76
78
79
81
82
85
86
86
87
88
91
93
93
viii
-------
Number Page
24 Concurrent and Sequential NH« Derivative Samples 96
25 Concurrent and Sequential NH3 Derivative Samples with
Number 1 Leg Heated (105°C) 97
26 Comparison of Predicted and Measured Aerosol Concentrations . . 99
27 Percent of Acid Passing Prefilters (123 yg/m3) 100
28 Background Sulfur Levels as yg H SO 101
O
29 Analysis of H2S04 Aerosol at the 12.3 yg/m3 Level
(Ambient Conditions) 102
o
30 Heated Aerosol Collection at the 12.3 yg/m Level
(Sample Rate 9 1/min) 103
3
31 Heated Aerosol Collection at the 61 yg/m Level
(Sample Rate 9 1/min) 103
32 Analysis of Samples Taken at the Prefilter (Modified
Thomas Generator - No Flame) , 105
33 Analysis of Samples Taken at the Collection Filter
(Modified Thomas Generator - No Flame) 106
34 H2SC-4 Slip Through Prefilter at 145°C (Modified
Thomas Generator - No Flame) 107
35 Quantitative Recovery of H2S04 Passed Through the
Prefilter (Modified Thomas Generator) 108
36 Passage of H2S04 Droplets Through Prefilter at 190°C
(Modified Thomas Generator - With Flame) 109
37 Passage of H2S04 Droplets Through Prefilter Using H PO,
Scrubbed Air (Modified Thomas Generator - With Flame) 109
38 Passage of H2S04 Droplets Through Prefilter at 190°C
Versus Time (Modified Thomas Generator - With Flame) 110
39 Animal Chamber Tests (Prefilter and Collection Filter) 114
40 Animal Chamber Studies - All Quantitative Samples
Taken by ARC on 7/19/78 120
41 Animal Chamber Studies - All Quantitative Samples
Taken by ARC on 7/20/78 120
42 Animal Chamber Studies - Individual Sample Comparison
Between EPA and ARC Samples 121
43 Animal Chamber Studies - Daily Average Data Summary
Comparison Between EPA and ARC Samples 122
44 Animal Chamber Studies - Average Data (Questionable
Sample Deleted) 122
45 Results of Environmental Field Tests - Rural Area 123
46 Results of Environmental Field Tests - Urban Area 123
ix
-------
SECTION 1
INTRODUCTION
The concern regarding environmental sulfuric acid aerosol has been
gradually gaining momentum over the years. This is due in part to
recent smog disasters in Meuse Valley (1), Donora (2), and London (3).
The inability to correlate mortality rates during these disasters with
measured levels of sulfur dioxide has raised serious questions as to
the real nature of the problem. Leighton (4) has documented that sulfur
dioxide, in the presence of oxygen, moisture and sunlight can form
sulfuric acid. This fact, coupled with the knowledge that high levels
of sulfur dioxide existed during these smog disasters, has made sulfuric
acid aerosol a prime suspect for responsibility for the excessive mortality
rate. Firket's (5) correlation between ambient air particulates and
mortality rates during the Meuse Valley smog disaster also suggests that
aerosols could be responsible for the large number of deaths. Sulfuric
acid is known to be a potent irritant that can cause narrowed air
passages (6,7) and, thus, be a significant health hazard to people with
respiratory difficulties.
The source of most H2SO, pollution is fuel, which often contains
significant quantities of sulfur. When fuel containing a sulfur component
is burned, sulfur dioxide is generated. The sulfur dioxide, in the presence
of oxygen, moisture and sunlight, can be further oxidized to sulfuric
acid. The process can be accelerated by metal particulates in the air
which serve as catalysts.
The current energy crisis suggests that more coal will be used in
the future as an energy alternative. Much of our coal resources has a
high sulfur content which will cause ambient sulfate levels to rise. The
advent of the automobile catalytic converter insures efficient oxidation
of S02 to H2SO, (or a sulfate salt) and will also add to the problem.
-------
Thus, the ambient sulfate situation shows no sign of improvement, but only
a steady upward trend can be anticipated.
The hazards that can result from H^SO, aerosols in ambient air make it
imperative that a reliable method for its measurement be established.
Many problems in finding a method for measuring H2SO^ in air can be foreseen.
The primary problem is the diversity of forms in which aerosols exists, i.e.,
sulfuric acid, ammonium sulfate and metal sulfates, some of which are
water soluble (Na2SO,), and some of which are insoluble (PbSO^). In
order to accurately measure sulfuric acid concentration, both the sampling
and analysis method must be capable of handling all forms of sulfate
efficiently. Other problems which affect the reliability of a method are
the collection of small particles, the loss of lUSO, by reaction with other
particulates, and interferences from various pollutants.
-------
SECTION 2
CONCLUSIONS
Sulfuric acid can react on collection surfaces with other environmental
particulates during sample collection to form other compounds and thus cause
a negative interference. This effect can be lessened by the in situ addition
of a gaseous reagent into the moving gas sample stream which will subse-
quantly react with the sulfuric acid to form a more stable adduct. Reaction
kinetics indicate that adduct formation in a moving gas stream is best
accomplished by volatile basic reagents capable of undergoing protonation.
Volatile amines, and amine derivatives appear to be ideal for the in situ
gas phase fixation of sulfuric acid.
Positive interferences during the analysis of sulfuric acid adducts
were limited to sulfur bearing species by using the sulfur specific flame
photometric detector. Interferences were further eliminated by selecting
fixing reagents which form sulfuric acid adducts that thermally decomposed
below the thermal decomposition temperature of potentially interfering
sulfates. It could not be determined if ammonium sulfate salts were totally
eliminated using the thermal decomposition/flame photometric technique
due to the absence of a suitable referee analysis method for the low levels
(0.1 - 3 yg) of sulfuric acid found in the environment. An alternate
approach for eliminating ammonium sulfate interference by volatilizing
the sulfuric acid aerosol and selectively passing it through a prefilter
was found to be inefficient at volatilization temperatures up to 200°C.
The sulfur gases SO^, COS, and H2S were found not to interfere.
The gas phase fixation/thermal decomposition methodology exhibits
good precision (C = 0.20) and good accuracy (±20%) for laboratory generated
aerosols. The laboratory generation of accurate, reproducible sulfuric
3
acid aerosols in the low yg/m range is probably at least as difficult as
sample collection and subsequent analysis. The combined lower sensitivity of
the fixation/thermal decomposition approach utilizing the flame photometric
detector is on the order of 0.05 yg H_SO,.
-------
SECTION 3
RECOMMENDATIONS
As a result of the effort at Atlantic Research Corporation under
Environmental Protection Agency Contract No. 68-02/2467, the following
recommendations are made:
1) More field tests need to be performed in order to compare the
gas phase/thermal decomposition concept with other potential
sulfuric acid measurement methodologies. Field tests must be
performed in order to evaluate the effect of ambient particle
size dispersions on thermal decomposition temperatures of both
sulfuric acid adducts and potential sulfate interferences.
2) Other derivatives should be evaluated for analysis by an alter-
nate approach should the thermal decomposition approach be
found unacceptable due to particle size effects. These deriva-
tives should be evaluated using both laboratory-generated and
field samples.
-------
SECTION 4
TECHNICAL DISCUSSION
GENERAL ATMOSPHERIC CONSTITUENTS
The air we breath is a mixture of constituents occurring from natural
life processes and the substances man generates as a result of his life
style. The latter is the basis for the current epidemiological interest in
air pollution. It appears that man is slowly poisoning himself with the
by-products of his own inventiveness.
Industrial manufacturing processes sometimes have substances associated
with them that are potentially dangerous as environmental contaminants. This
problem is especially acute in immediate geographical industrial areas, since
the concentrations may be quite high due to the close proximity of the
emanating source. These substances can enter the environment in the form
of gases, vapors, fumes, aerosols, or dusts, and hence, require a variety
of pollution monitoring and control techniques. Generally, these substances
originate from initial process materials or are by-products of the manufac-
turing process itself. As can be seen from Table 1 (8), a wide range of
chemical species and manufacturing processes are associated with potential
environmental hazards.
Fuel consumption is probably the largest source of environmental pollu-
tion. It has been estimated that the amount of sulfur introduced into the
atmosphere by energy generation (coal, gas, gasoline) is equivalent to
the combined emissions of all anaerobic processes and from volcanism (9)..
The largest single source of CO in the environment has been reported to be
from the internal-combustion engine (10).
Recent studies have shown a direct correlation between inefficient
fuel consumption and hydrocarbons found in the environment (H»12,13).
Pellizzari, et al. (14) analyzed the atmosphere in several locations using
porous polymers to concentrate trace pollutants. A chromatogram of an air
sample taken in Glendora, California, using this concentration technique
-------
TABLE 1. ORGANIC COMPONENTS OF INDUSTRIAL WASTE EFFLUENTS"!^!'
Plant
Mines, ore
treatment plants
Foundries
Iron and steel
processing
Coal production,
coking plants
Wood charcoal
production
Petroleum industry
Sulfite pulp
(sulfate)
pulp
Rayon and cellulose
Paper manufacture
Textile industry
Laundries
Composition
Humus, coal sludge, flotation
agents, particulates, gases
Cyanides, phenol, tar components,
coal sludge
Wetting agents and lubricants,
cyanides, inhibitors, hydro-
carbons, solvent residues
Humus, coal particles, cyanides,
rodanines, phenols, hydrocarbons,
pyridine bases
Fatty acids, alcohols,
particularly methanol, phenols,
carbon monoxide
Oil emulsions, naphthenic
acids, phenols, sulfonates,
hydrocarbons, sulfur gases
Methanol, cymol, furfurol,
soluble carbohydrates,
lignosulfonic acids, sulfur
gases .
Mercaptans and sulfides,
alcohols, terpenes, lignin,
resinic acids, soluble carbo-
hydrates
Xanthogenates, alkali
hemicelluloses, toxic gases
Resinic acids, polysaccharides,
mucins, cellulose fibers, flo-
tation agents, sulfur gases
Scouring and wetting agents,
leveling agents, sizers, de-
sizing Agents, fatty acids,
finishes, Trilon (nitrilo-
triacetic acid), dyes
Detergents: carboxymethyl-
cellulose, enzymes, optical
brighteners, colorants; soil:
protein, blood, cocoa, coffee,
carbohydrates, emulsified fats,
soot
-------
TABLE 1 (continued)
Plant
Leather and tanning
industry
.Natural glue and
eelatin
Sugar refineries
Starch plants
Dairies
Grease and soap
factories
Canning factories
Sauerkraut factories
Beer breweries
Fermentation
industry
.Slaughter houses
Composition
Protein degradation products,
soaps, tanning agents, emulsified
lime soap, hair
Protein degradation products,
emulsified fats and lime soaps
Sugar, plant acids, betaine,
pectin and other soluble plant
components.
Water-soluble plant components
(protein compounds, pectins,
soluble carbohydrates)
Milk components (protein, lactose,
lactic acid, fat emulsions),
washing and rinsing agents.
Glycerine, fatty acids, fat
emulsions
All types of soluble plant components
and volatile gases
Lactic, acetic and butyric acids,
carbohydrates, other soluble plant
components and gaseous constituents
Water-soluble plant components,
beer residues, rinsing agents
Fatty and amino acids, alcohols,
unfermented carbohydrates
Blood, water-soluble and emulsified
meat components, fecal matter
-------
is shown in. Figure 1, along with the identification of most of the atmos-
pheric constituents. Clearly, the majority of the identified constituents
can be attributed to fuel oils and gasoline. Nitrous oxide and sulfur
dioxide, which are also major atmospheric pollutants, are primarily
formed in combustion processes where nitrogen and sulfur are oxidized
slowly by oxygen and more rapidly by ozone.
Particulate emissions are also characteristic of many aspects of a
modern industrial society. Perhaps the most notorious source of these
emissions is the internal-combustion automobile engine. Inorganic salts
(especially lead), iron as oxides, base metals, soot, carbonaceous materials
and tars constitute the primary components of these emissions (15).
In addition, industry, agriculture, utilities and the private sector
make their own contributions to particulate pollution. While some of these
emissions are from natural sources, such as wind erosion of soils, most
are man-made. Trace metals, carbon-adsorbed gases, and various benzene-
soluble organics have all been detected in urban atmospheres. Table 2
shows the results of elemental analysis of particulates found in the atmos-
pheres of six cities (16).
Sulfuric Acid Aerosol and Sulfateg
Sulfuric acid aerosols in the atmosphere are generally believed to
be formed from the oxidation of S02 to SO., in the presence of ultraviolet
light. It is the mechanism of the S0_ oxidation that has been the subject
of much scientific speculation for years. The postulated mechanisms fall
broadly into two general classifications: Homogeneous and heterogeneous
reaction processes. The literature reveals that homogeneous gas phase
reactions (17-19) suffer from a lack of rate constant data, while the
experimental evidence on heterogeneous processes (20,21,22,23) tend to
disagree from one investigator to another. Thus, while many scientific
theories have been postulated, none has totally fulfilled the scrutiny
of scientific examination for the oxidation of environmental SO to the
corresponding acid aerosol. It would appear that both homogeneous and
heterogeneous processes, along with other factors, simultaneously influence
the rate at which S02 is oxidized by ultraviolet light. For example, it has
-------
234
216
298
TEMPERATURE PC)
A—L
96
68
83
104
128
Tfc
B
Figure 1. (A) Total ion current chromatogram of ambient air sample from Glendora, Calif. A 400 ft OV-101 SCOT programmed from 20 to 230
°C at 4 °C/min was used. (B) Total ion current chromatogram of Tenax GC cartridge blank. A 400 ft OV-101 SCOT was used-; (T47~
<1)Difluorochtoromethane: (2) 1-butene; (3) isobutane; (4) unknown; (5) unknown; (6) isopentane, trichkxofluoromethane; (7) 1-pentene, C5H8; (8) n-pentan*; (8)
•oprene; (10) methylene chloride; (11) propanal; (12) acetone; (13) unknown; (14) unknown; (15) 2-metnylperitane, 2-fluoro-2-methylpropane; (16) 3-melhyt-
pwitane: (17) C6H12; (18) 3-methyl-2-pentene + n-hexane, 2-methylfuran; (19) chloroform; (20) C6Hi4; (21) C6H,2; (22) methyl vinyl ketone (tent.), methyl ethyl
tetone: (23) 1,1.1-trichloroethane, ethyl acetate; (24) benzene; (25) carbon tetrachloride: (26) C6H,2; (27) 2.3-dimethylpentane: (28) 1,1,3,3-tetramethylcycte-
pwitane: (29) cyclohexenol isomer; (30) unknown; (31) 1-frans-2-dimethylcyclopentane, C7H,S, C7H1:, trichloroethylene; (32) n-heptane; (33) CrH12; (34) C«H,4;
(35) 2,2,3.3-tetramethyibutane; (36) 4.4-
-------
TABLE 2. APPROXIMATE ELEMENTAL CONCENTRATIONS FOUND IN PART1CUIATES
COLLECTED FROM AMBIENT URBAN AIR
(Nanogratns/M )
Ele-
ment
H
LI
Be
B
C
N
F
Na
Mg
Al
Si
P
S
Cl
K
Ca
Sc.
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Ga
Ge
As
-Se
Br
Rb
Sr
Y
Zr
Nb
Mo
Ru
Rh
Pd
AS
Cd
Cincinnati
' 3,000.
' 5.
0.2
10.
30,000.
3,000.
40.
4,000.
1,000.
2,000.
5,000.
100.
4,000.
1,000.
1,000.
6,000.
<1.
200.
10.
20.
100.
5,000.
5.
20.
300.
1,000.
5.
5.'
20.
4.
50.
10.
10.
3. .
4.
0..4
2.
<0.04
<0.04
SO. 2
<1.
2.
Denver
5,000.
20.
0.04
20.
40,000.
2,000.
400.
10,000.
1,000.
7.000.
40,000.
300.
3,000.
5,000.
10,000.
6,000.
2.
400.
20.
20.
100.
5,000.
5.
30.
500.
200.
4.
3.
5.
2.
200.
100.
100.
4.
20.
2.
5.
<0.06
"•0.2
<1.
1.
1.
St. Louis
7,000.
10.
0.4
30.
60,000.
6,000.
200.
10,000.
2,000.
3,000.
6,000.
200.
8,000.
1,000.
2,000.
20,000.
1.
300.
20.
20.
50.
5,000.
3.
20.
3,000-.
3,000.
10.
5.
50.
10.
300.
20.
50.
2.
10.
1.
3.
<0.04
<0.1
<0.4
20.
5.
Washington
3,000.
3.
0.1
10.
30,000.
2,000.
100.
1,000.
1,000.
2,000.
5,000.
30.
4,000.
300.
1,000.
3,000.
1.
200.
100.
10.
50.
2,000.
5.
50.
400.
400.
5.
4.
20.
5.
200.
5.
10.
1.
4. .
0.3
2.
<0.01
<0.03
<0.1
0.6
0.3
Chicago
10,000.
2.
<0.1
5.
100,000.
10,000.
30.
5,000.
'5,000.
2,000.
6,000.
60.
3,000.
10,000.
2,000.
10,000.
3.
300.
100.
20.
100.
4,000.
10.
40.
400.
500.
6.
7.
60.
2.
100.
20.
40.
1.
4.
0.5
3.
<1.
<0.06
<0.3
2.
3.
Philadelphia
4,000.
2.
0.05
5.
30,000.
.5,000.
20.
1,000.
•5,000.
3,000.
10,000.
50.
3,000.
4.
1,000.
8,000.
10.
400.
200.
40.
200.
4,000.
20.
100.
200.
500.
6.
<0.4
10.
2.
20.
20.
40.
2.
10.
0.3
2.
<0.03
<0.01
<0.3
0.6
1.
10
-------
TABLE 2. APPROXIMATE ELEMENTAL CONCENTRATIONS FOUND IN PARTICULATES
COLLECTED FROM AMBIENT URBAN AIR (Continued)
Ele-
ment
In
Sn
Sb
Te
1
Cs
Ba
La
Ce
Pr
Md
Sm
EU
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Hf
Ta
W
Re
Os
Ir
Pt
Au
Hg
Tl
. Pb
Bi
Th
U
Cincinnati
<2.
100.
8.
SO. 2
0.5
0.4
50.
2.
3.
1.
3.
0.4
<0.2
<0.6
<0.l
<0.3
<0.1
<0.3
<0.06
<0.2
<0.1
<0.04
<0.1
0.2
<0.02
<0.02
<0.02
<0.03
<0.1
<0.04
<0.03
1,000.
0.5
0.1
0.2
Denver .
<3.
10. v
5.
<0.2
2.
1.
300.
20.
20.
4.
10.
1.
0.5
1. '
0.1
1.
0.1
0.5
<0.2
<1.
0.3
<0.6
<0.3
2.
<0.05
<0.1
<0.05
<0.1
<0.1
<0.1
0.1
3,000,
1.
. 0.6
0.3
St. Louis
<2.
, 50-
8.
5.
A
4.
. 1.
100.
2.
3.
1.
3.
0.6
<0.4
<1.
<0.2
<0.6
<0.2
<0.6
<0.1
<0.4
<0.2
<0.2
<0.3
0.5
<0.06
<0.06
<0.03
<0.1
<0.03
<0.1
<0.1
3,000.
2.
1.
2.
Washington
<0.5
20.
3.
0.3
4.
3.
100.
4.
3.
1.
3.
0.5
0.1
0.3
0.05
0.2
0.03
<0.3
<0.03
<0.3
<0.02
<0.05
SO. 05
0.1
<0.01
<0.02
<0.01
<0.02
<0.02
<0.05
£0.02
2,000.
0.3
0.05
0.05
Chicago
<1.
20.
8.
SO. 2
SO. 7
0.7
50.
5.
7.
2.
5.
2.
0.3
2.
£0.1
0.3
0.1
£0.2
sO.l
SO. 5
<0.1
<0.3
SO.l
1.
<0.05
<0.04
<0.02
<0.04
. <0.1
<0.04
SO. 01
4,000.
2.
0.4
0.2
Philadelphia
7.
1.
0.2
0.6
0.2
0.6
0.1
0.3
0.06
0.3
SO. 04
0.2
SO. 2
1.
<0.02
<0.03
<0.06
<0.03
"^0.06
<0.03
0.04
2,000.
0.6
0.2
0.2
11
-------
been shown that furnace operating conditions alone can enhance sulfur
emissions being immediately oxidized to sulfur trioxide (24).
Takahashi, et al. (25) proposed a slaplifted three-step mechanism
for formation of the acid aerosol by photooxidation of S02.
(1) Photooxidation of SCL followed by rapid combination with a
water molecule
o2 HP
S02 ~~h7* S03 * H2S°4 (vap°r)
(2) Nucleation to a critical size by combination with several water
and sulfuric acid vapor molecules
H2S04(vapor) + H20(vaPor) nucleation> OW^O^ (embryo)
(3) Growth of the embryo to a large aerosol particle through combina-
tion with additional water, sulfuric acid and other>molecules
H-SGv, H20,particulatee
(H2S04)n(H20)m(embryo) —— > aerosol particle
The photooxidation rate, the growth of the aerosol particle and thus the
sulfuric acid concentration in each particle is affected by the relative
humidity. During the growth period, sulfuric acid molecules condense on
the embryo resulting in an increase in the sulfuric acid concentration.
Simultaneously, water molecules are also condensing. The resulting particle
will grow until an equilibrium is reached between the water vapor pressure
and the ambient relative humidity. The net result is that the concentration
of sulfuric acid in each aerosol particle is a function of its radius and
the ambient relative humidity. Table 3 shows the sulfuric acid content
of an individual aerosol droplet as a function of particle radius and rela-
tive humidity. These data ace the results .of theoretical calculations (26);
however, evidence indicates that the concentrations are representative of
actual aerosols. *
Inspection of Table 3 shows that the weight percent of sulfuric acid
in an individual aerosol particle decreases as the particle size and the
relative humidity increase. For aerosols ranging from r = 0.1 to 1.0 um
12
-------
the average concentration ranges between 45 and 60 weight percent for 50
percent relative humidity and between 33 and 54 weight percent for 75 per-
cent relative humidity. These concentration ranges will be encountered
under normal conditions. However, in areas of high humidity or inversion
conditions, the concentration per particle can drop to 18 percent or below.
TABLE 3. VARIATION OF SULFURIC ACID CONTENT
OF ATMOSPHERIC AEROSOLS (26)
Particle
radius
1.0
0.5
0.2
0.1
Relative
humidity
100
75
50
25
10
100
75
50
25
10
100
75
50
25
10
100
75
50
25
10
Hole fraction
of H,SO,
0.04
£,086
0.132
0.196
0.258
0'.06
0.098
0.140
0.200
0.267
0.104
0.132
0.175
0.225
0.289
0.154
0.179
0.215
0.267
0.327
Weight Z
of H,SO,
^™"*™*™"**« ' "*l™
18.5
33.9
45.3
57.0
65.4
25.8
37.2
47.1
57.6
66.5
38.7
45.3
53.6
61.3
68.9
49.8
54.3
59.9
66.5
72.6
It is under these high humidity-inversion conditions that high sulfuric
acid concentrations in air and resulting fog disasters occur, even though
the concentration per particle is low.
The foregoing discussion suggests that collection methods should be
capable of handling particles from 1.0 ym to 0.1 ym. It now appears, how-
ever, that sulfuric acid aerosol generated in automobile catalytic con-
verters may be in the 0.01 ym size range. Whitby (27) has suggested (Figure
2) that the ambient particle distribution consists of three principle size
ranges. The ranges given by Whitby consist of particles with a mass mean
13
-------
HOT
VAPORS
T
CHEMICAL CONVERSION
OF GASES TO LOW
VOLATILITY VAPORS
CONDENSATION
PRIMARY PARTICLES
LOW
VOLATILITY
VAPOR
f
3GE
LEA
4
HOMOGENOUS
NUCLEATION
CONDENSATION GROWTH
OF NUCLEI
WIND BLOWN DUST
EMISSIONS
SEA SPRAY
VOLCANOS
PLANT PARTICLES
RAINOUT
AND
WASHOUT
TRANSIENT NUCLEI
OR
NUCLEI RANGE
•>N-
ACCUMULATION
RANGE
-FINE PARTICLES:
MECHANICALLY GENERATED
AEROSOL RANGE
- COARSE PARTICLES •
Figure 2. Three principle Size ranges for ambient
particle size distribution (27).
14
-------
diameter of 10 ym, 0.4 ym and 0.02 ym. The 0.02 ym size range has serious
health implications, since it is likely to penetrate further into the respir-
atory system causing severe chronic problems. Additionally, the smaller
aerosol radius suggests an aerosol of higher sulfuric acid content. While
these smaller aerosol particles are not likely to remain in the environment
for extended periods, they should be considered as a serious threat to
persons who are in close proximity to high traffic areas. Thus, development
of any sampling/analytical method for sulfuric acid must necessarily consider
particles in the 0.01 to 1.0 ym size range. As the burning of coals and high
sulfur content fuel increases due to the energy shortage, a corresponding
increase in the sulfate content of the ambient air can be anticipated.
THE SULFURIC ACID MEASUREMENT PROBLEM
It is likely that any atmospheric sample would contain many species
and require special consideration to circumvent potential interferences
during the analytical determination of a specific constituent within the
sample matrix. This is especially true in the case of sulfuric acid
aerosols, since sulfufic acid is well known for its propensity toward
reaction with both inorganic and organic species. The reaction between
sulfuric acid aerosol and other species can occur in the atmosphere, or
it can occur on the sample collection surface as the result of enhanced
physical contact between the acid aerosol and other enviromental consti-
tuents. By far the most difficult problem in the development of a method
for measuring sulfuric acid aerosol has been the many potential interferences.
These interference substances can be divided into two practical categories:
those which artificially enhance the measured acid value (positive interfer-
ence) , and those which diminish it (negative interference).
Positive Interferences
Positive interferences usually result from the failure of an analytical
method to distinguish between the desired compound and one which is chemically
similar. This is particularly a problem with sulfuric acid measurement,
because so many similar forms may be present in the atmospheric sample.
Since the different forms are known to have different toxicological
properties, it is necessary to be able to distinguish between them (28).
15
-------
Ammonium sulfate and bisulfate have the greatest chemical similarity
to sulfuric acid, and have constituted a severe interference with most
previous methods of analysis. These compounds are formed by an equilibrium
reaction between ambient ammonia and sulfuric acid, the extent of which
seems to be related to climatic conditions (29). In aqueous solution,
both sulfate and acidic protons (from hydrolysis of NH, ) are present,
making it difficult to distinguish them from dilute sulfuric acid. Ammonium
sulfate and bisulfate particulates in the atmosphere are generally associated
with some moisture, futher enhancing the interference by hydrolysis which
has already occurred. Some success has been achieved in separating I^SO,
from the ammonium sulfates by volatilization, since they are stable to
about 235°C (301.
Many other sulfate salts may be present, some soluble (Na^SO.) and
some insoluble (CaSO,, PbSO,). These are generally less of a problem than
the ammonium sulfates, but may still interfere with a simple sulfate
measurement. All but a few require temperatures over 500°C to cause
decomposition, and none decompose under 250°C.
Ambient sulfur gases, including sulfur dioxide, hydrogen sulfide,
carbonyl sulfide, and methyl mercaptan, can generally be separated auto-
matically during collection, if a filtration or impaction system is used.
There has been, however, some concern about the catalytic oxidation of
S07 to sulfuric acid on certain filter media, particularly glass fiber
filters (31). Other atmospheric particulates collected during sampling
may also catalyze this reaction (31,32,33).
Also of concern are other strong mineral acids (HC1, HNO,, and H PO.)
•3 j H
which may be present in an atmospheric sample (34,35). These may inter-
fere directly if the analytical method measures only acidity, or they may
interfere by reacting with sulfate salts on the collection surface to produce
sulfuric acid.
Finally, a wide variety of substances may interfere with widely used
wet chemical methods of sulfate analysis. Most of these methods are based
upon precipitation of S04~ , usually BaSOA; thus, other anion~s~which
precipitate Ba+ may interfere (PC>4~3, s"2). Some metal cations may
cause a negative interference by precipitating sulfate (Pb+2, Ca+2),
16
-------
and other species may interfere by complexing indicator reagents.
Clearly, a simple sulfate measurement is inappropriate unless the sulfuric
acid is first separated from these other species.
Negative Interferences
Interferences that cause a negative error in analytical measurement
generally result from loss of sample. Since sulfuric acid is a highly
reactive substance, such a loss may readily occur - if the acid contacts
other particulates on the collection surface. Reactions with inorganic
particulates can produce sulfate salts, which must necessarily be prevented
if negative interferences from these compounds are to be avoided.
Ambient ammonia may also cause a negative error, since the collection process
tends to mix acid and ammonia to a greater extent than would normally
occur. The product ammonium sulfates must be excluded from the sulfuric
acid measurement.
These interfering side-reactions have been the greatest single
obstacle in the development of a reliable measurement method for sulfuric
acid aerosol. Volatilization techniques for separating the acid from its
salts failed to prevent this sample loss; in fact, the loss may be increased
by the volatilization procedure which brings hot acid vapor into contact
with potentially reactive substances. Only by completely preventing these
reactions can the interferences be avoided. By using a very sensitive
analytical method which requires only a small volume of air to be sampled,
the problem can be minimized, but not eliminated. Chemical stabilization
of the acid as it is collected seems to be the only solution.
Approach to the Problem
From the foregoing discussion, it is evident that any solution to the
selective measurement problem has several special requirements. In order to
eliminate positive interferences, the method must be highly selective. Simple
acid or sulfate measurement is not adequate unless sulfuric acid is first
separated from other acids and sulfates in the atmospheric sample. To
fir event, negative Interferences., the analytical method must be highly sensi-
tive (so that only a small sample need fee collected), atid It mtist elttpley
a process for chemically stabilizing the acid as it is collected. The
17
-------
choice of stabilizing (fixing) reagent is inherently limited by the sulfuric
acid concentration in each aerosol particle as previously discussed. Of
course, the sulfuric acid adduct thus formed must be selectively analyzable
in the presence of many potentially interfering species.
It is also evident that each aspect of this problem is intimately
related to every other aspect. Ideally, a combination of fixation, sampling,
and analysis techniques, working together, would eliminate all interferences.
No one of these three aspects can be regarded as separate from the other
two, since a decision in any one area may limit possibilites for the other.
The following discussion on theoretical development of research approach is
divided into three main areas: aerosol sampling techniques, H^SO. aerosol
fixing reactions, and adduct sample analysis. Reference is made continually
to all three areas.
THEORETICAL DEVELOPMENT OF RESEARCH APPROACH
Sulfuric Acid Aerosol Sample Collection
In any analytical scheme, it is obvious that interferences must be
eliminated, or at least controlled, if meaningful quantitative results are
to be obtained. In the case of sulfurie acid measurements, this is parti-
cularly difficult, since other sulfates and acids are present in the
atmosphere, which tend to interfere with analytical tests. Ideally,
the sampling method selected would capture only sulfuric acid molecules,
and allow all others to escape. This would greatly simplify subsequent
analytical procedures. Unfortunately, no selective collection system
exists which will capture sulfuric acid aerosol and exclude other particu-
lates. The following discussion describes current aerosol sample collec-
tion techniques and how they might be applied to the fixation of sulfuric
acid aerosols.
Aerosol Collection Methods —
Filtration — Collection of aerosol particles by filtration on a
substrate is probably the most popular, as well as the most convenient and
inexpensive method in use. Submicron particles, such as those found in
aerosols, can be efficiently collected by this method.
18
-------
The method as generally employed consists of a porous filter disc
supported by a holder, as shown in Figure 3. The filter disc itself can be
made of a number of different materials, such as common filter paper, glass
fibers, wire screen, or membranes. Each type of filter material has inherent
advantages and disadvantages, some of which are immediately obvious, and
some not so obvious as shown by the following example.
Mleropot
filter
^
aSo-
•« /
Forouz
tlatt
-
Q&-&3*
_J^s.
<§" £<*><• 0
'? * ?? a
^
I
\r/
Te U.T
Figure 3. Filter holder.
Glass filters combine two desirable aspects of air sampling: high
retention efficiency for submicron particles, and high flow rates. They
have been frequently used for collection of aerosol sulfates. However,
problems were discovered when glass filters were used for sulfate collec-
tion. Lee and Wagmen (36) showed that atmospheric SC^ could be catalytically
oxidized to S0_ on the glass surface, thus giving high values for sulfate
concentration. It has also been shown that sulfuric acid is irreversibly
fixed in some glass fibers,(24), and that many brands contain high amounts
of sulfates (37). Any one of these factors can lead to erroneous results.
Barton and McAdie (38) demonstrated that difficulties with glass filters
can be overcome by pretreatment of the filters with sulfuric acid. However,
19
-------
Maddalone, et^ al_. (38), found that treated glass filters were still trouble-
some and suggested the use of Teflon or graphite filter media for the collec-
tion of sulfuric acid aerosols.
Thomas, e* al. (40), investigated Millipore, Whatman 41, Teflon,
and graphite filters and found that Teflon filters were the least desirable
with a collection efficiency of approximately 58 percent. In more recent
work, however, West, ej^ al.(41), have made use of Fluoropore Teflon filters
and the glass fiber filters (42) found unacceptable by other researchers.
Liu, e_t al^ (43), used an electrical aerosol detector to determine the
in situ collection efficiency of various filter media for specific mono-
disperse aerosols. Liu found that the Mitex LS filter media had a collec-
tion efficiency of approximately 100 percent on aerosols down to the 0.1
micron size range and a collection efficiency of approximately 90 percent
at the 0.03 micron level. It is felt that Liu's work is more valid,
because his data do not reflect the use of involved analytical chemistry
methodology to measure the aerosol actually collected on the filter's
surface. Instead, Liu measured the particle density distribution change
as the aerosol stream passed through various collection media. Thus,
it appears that the best media for the collection of H^SO, aerosol by the
filtration method is the 5 ym pore Mitex LS filter or the 0.5 ym pore
Fluoropore Teflon filter.
Washing or Scrubbing — Another class of H2SO, aerosol sampling techni-
que currently being used is the simple bubbling system shown in Figure 4.
Intake
To Air
-
L
.,
'•
1
8
00
/
J
IrSY
0°
o°o
o
c°o
vy ^^
Bubbler
It.p
Figure 4. Bubbling system.
20
-------
The method uses a liquid to facilitate the removal of aerosols from the air
stream. A Greenburg-Smith impinger containing isopropyl alcohol is sometimes
used for sampling sulfuric acid aerosols (24). The isopropyl alcohol serves
to scrub the sulfuric acid and sulfur trioxide from the aerosol, while allow-
ing the sulfur dioxide to pass through to another absorbing solution,
usually hydrogen peroxide. The major disadvantage of this type of collection
is the mist accompanying the effluent gas which can carry some of the initial
aerosol with it. Difficulty also is often found in using this type of sampl-
ing when the component to be sampled exists in a variety of forms, as occurs
in the case of sulfates. The problem is that most sulfate salts dissolve by
the same mechanism as sulfuric acid, creating interferences.
Impact ion — The principle of operation of the impactor is based upon
the differential momentum of gas and particulate matter in an aerosol stream.
The direction of an aerosol stream is radically changed at a solid surface,
and the particles are driven toward the surface and cling to it, as shown
diagramatically in Figure §. There are many instruments available which use
variations of this principle for collecting aerosols down to the submicron
range. Cascade impactors are capable of differentiating the aerosols by
particle size. This is done by using a series of impaction points with jets
of decreasing size. Inertial impactors use a rotating drum as the collection
surface, which allows the time classification of particles. Other modifica-
tions of the collecting surface, such as metal foil for acid particles and
glass plates immersed in liquid, have been used. The impaction method has
been used by Scaringelli and Rehme (24) for particulate sizing and collecting
of sulfuric acid.
Figure 5. Impactor.
21
-------
Precipitation — Precipitation methods used in air pollution sampling
include thermal and electrostatic precipitators. In thermal precipitation,
the aerosol is passed between a hot wire which repels the particles and a
cold plate where they are precipitated. Although this method is efficient,
it has a low flow rate and sample capacity. The electrostatic precipitator,
on the other hand, combines high flow rates, small pressure drop and high
efficiency. Here, the gas stream is passed between a high electrical poten-
tial which drives the particles toward a collecting electrode where they
are precipitated. Kerrigan, Snajberk and Andersen (44) used the electro-
static precipitator, the Greenburg-Smith impinger and the sintered glass
filter method to collect a sulfuric acid aerosol for comparison purposes.
The results of these tests are given in Table 4. It should be noted, however,
that the H^SO, concentration investigated was approximately one thousand
times greater than the range of current environmental importance which is
3
in the low yg/m range.
TABLE 4. COMPARISON OF INSTRUMENTS (43)
HjSO, Mist Concentration, mg/m ,
••-••- as determined by:
Sintered
Test No. Precipitator Impinger ' Glass Filter
1 18.4 18.0 17.9
16.9 17.2 17.2
2 28.0 29.0 28.6
26.8 28.9 32.2
3 24.8 26.2 24.3
27.4 25.0 22.9
* 72.3 69.8 67.9
69.1 69.5 72.2
5 30.7 31.1 33.5
31.6 31.3
Formation of H^SO, Adduct During Sampling —
Each of the aerosol collection techniques discussed above is fairly
efficient. The question here is how these techniques could be utilized in
conjunction with a reagent to form a sulfuric acid adduct. In order to
22
-------
examine the relative merits of each collection method under these conditions,
it must be assumed that a reagent which reacts specifically with sulfuric
acid has been defined. It then becomes a matter of insuring contact between
the sulfuric acid aerosol and the adduct-forming reagent during sample
collection.
One problem that is inherent in all of the above-mentioned sampling
techniques is that they tend to collect a variety of airborne particulates.
There appears to be no feasible way to mechanically separate the sulfuric
acid aerosol from other airborne particulates. If the collection medium
is simply pre-coated with the adduct-forming reagent, airborne particulates
could form a mat over it, inhibiting the subsequent adduct reaction. There
is a statistical probability that any given H SO, droplet could strike
one of these alien particles and react to form another compound. Even if
this side reaction does not occur, there exists the problem of volatilization
of the sulfuric acid aerosol. It becomes apparent that any sampling method
must necessarily segregate H~SO, from other particulates to prevent reaction
from occurring. The sampling method must allow for immediate contact of
the H«SO, with the adduct-forming reagent.
The collection surface of the sampler could be made large enough that
the probability of an H SO/ droplet finding another particle and reacting
with it is small. This does not, however, appear to be a likely solution;,
since sample collection would become a problem of interpreting a probability
factor which is controlled by the greatest variable of all: the environment.
The heavier the loading of other particulates in the sampling environment,
the greater is the probability of this inter-reaction occurring.
A large excess of adduct-forming reagent is obviously desirable to
insure sample contact, but it is equally important that the reagent be in
a matrix that prevents other airborne particulates from masking the sulfuric
acid aerosol. Several theoretical considerations to the problem of applying
the adduct-forming reagent to the H2SO, aerosol during sampling are discussed
below.
Impinger-Bubbler Method — The impinger-bubbler method probably pre-
sents the most obvious choice as far as maintaining good reagent surface
contact area. With this method, the reagent solution would serve as the .
23
-------
collecting medium and would also serve to segregate other particles from the
H9SO, aerosol. This would allow for optimum reagent surface/sulfuric acid
contact with very little probability of side reactions occurring due to
H~SO, surface contact with other particulates.
The ideal collecting solvent is one in which dilute acid (mostly
water) is immediately soluble, but sulfate salts are not sufficiently
soluble to interfere. This method might utilize an anhydrous solvent which
might help eliminate the subsequent hydrolysis of other sulfates to sulfuric
acid. In any case, the relatively large amounts of water in acid aerosols
would be difficult to remove effectively, even with an anhydrous solvent.
Thus, it is probable that a certain amount of hydrolysis is unavoidable
with the selective solvent approach. Once the sulfate is in solution,
whether it is derived from sulfuric acid or originates from some other
particulate is indistinguishable. Thus, the selective solvent approach
does not appear .to be a viable means of eliminating analytical sulfate
interferences, even though it appears promising as a means of controlling
side reactions.
Filtration, Impaction and Precipitation — Filtration, impaction and
precipitation sample collection methods tend to.create problems with air-
borne particulates matting the collection surface. It, therefore, does not
appear desirable to simply apply an adduct-forming reagent to the collection
surface prior to the actual sampling. Several novel approaches to the
application of the adduct reagent were hypothesized and are considered in
the following paragraphs.
It appears that the simultaneous spraying of the adduct-forming reagent
while the sample is being collected is a feasible approach. To accomplish
this, a pressurized aerosol bottle would be required containing the adduct-
forming reagent which could be sprayed onto the surface of the filter medium
through a metering jet. This would automatically coat all airborne parti-
culates as they are collected and greatly reduce the probability of HLSO,
aerosol reaction with other substances. Since the fixing reagent would be
sprayed as a liquid, the potential for dissolution of other sulfate salts
would, of course, exist. While this approach is more feasible than the
24
-------
bubbler method for collection, a selective solvent may still be necessary-
An elementary design for such a collection system is shown in Figure 6.
f.
Plnhole eprer Jet* or other method of
•preying liquid reafent iclutlon
To Fuap
• Filter oedium
Sheath
Figure 6. Filter system with Spray jets.
A better approach for a filter sampling system would be one in which
the H^SO reacts with the adduct-forming reagent while the reagent is in the
gas phase. This method, of course, would require a gas (vapor) with a fast
reaction time. After being reacted in the sampling stream, the H~SO adduct
^ 4
could be collected on the filter medium while the residual vapor would pass
through. A design for such a collection system is shown in Figure 7. It
appears that the gas phase titration technique would be applicable to other
collection techniques, such as impactors.
O
•
e •
• «
0 *
4
o k
r—
'D«M Vapor
or CM
•X
X
y /
Ikoith for rtltot
ith In
Vapot Ckakbaf
Figure 7. H2SO^ filter collection (gas or vapor
reagent).
-------
Sample Collection Conclusions — Filtration or impaction sample
collection techniques, coupled with gas phase adduct forming reagents,
appear to offer the greatest potential for achieving the necessary aerosol
contact. Gas phase fixation avoids the problems of solid phase fixation
by providing better acid/reagent contact, preventing matting of the collec-
tion surface by continuously applying fresh reagent, and automatically
removing excess reagent. The physical characteristics for the fixation
of sulfuric acid aerosols in a gas flow system are reviewed in Appendix
A. Selective solvent systems did not appear to be promising due to the
solubility of other sulfate species collected at the same time as the acid
aerosol.
Reaction Kinetics —
The following discussion considers the kinetics of reaction between
sulfuric acid and various reagent classes to form sulfuric acid adducts.
It is important that the adduct forming reaction takes place instantaneously
in order to prevent the formation of artifacts during sample collection.
The requirement that the fixing reagent react rapidly with sulfuric acid
before side reactions occur can ultimately be defined only by observing the
effect of possible interferences under actual sampling conditions. The final
choice of fixing reagent will depend upon its ability to react rapidly and
stoichiometrically with various concentrations of lUSO, to form an adduct
that can be selectively analyzed. Typical reactions of I^SO, are given in
Table 5 and consist of electrophilic aromatic substitution, electrophilic
addition to alkenes, ester formation, oxidation-reduction reactions and acid-
base reactions to form salts. Each of the above reaction classes are
discussed in the following sections.
Electrophilic Aromatic Substitution — One of the reaction types which
sulfuric acid undergoes is electrophilic substitution on an aromatic ring
(sulfonation). The mechanism of this substitution is postulated to be:
kl
26
-------
TABLE 5. GENERAL REACTIONS OF SULFURIC ACID
I. Electrophilic Aromatic Substitution
OH
<^- ~v
~S03H
II. Electrophilic Addition to Alkene
•Q /"""CI— f^TI—U i _ TT Cf\ --*- 13 fTI f* 13 *C
K- —Ijir—uQ~J\rt * rlrtOU. •*• lx ~l-«n._ — L» n — r
0 SOgH
III. Ester Formation
1. Primary Alcohols
ROH + H2S04 •> R-OS03H
IV. Oxidation-Reduction
1. Inorganic
A. Metal
xM + y H0SO. -»• M (SO.) +H_
y 2 4 x 4 y 2
V. Acid-Base Reaction to Form Salts
A. Amines
R-NH2
R_wu H cn -»•
2 + 24
(R3NH)2S04
27
-------
milder conditions. However, steric hindrance may become a factor and the
highly reactive carbonium ion may react with other anions.
The electrophilic addition of sulfuric acid to alkenes has some promise
as a sulfuric acid fixing reagent. The probability of success is low due
to the necessity to determine and control the reaction conditions in order
to obtain reproducible results.
Sulfate Ester Formation — Alcohols react with sulfuric acid or its
anhydride to yield acid sulfate ester:
R-CH2-CH2-OH + H2S04 -—»- R-CH2-CH2-OS02OH + H20
Secondary alcohols react faster than primary. Once the acid sulfate is
formed, it can undergo several reactions, depending upon the conditions.
If the water content is high, hydrolysis occurs and the original alcohol
is reformed. At higher temperatures and in the presence of HSO, , the acid
sulfate is dehydrated to the corresponding alkene:
R-CH0-CH0-OSO.OH + HSO " r - " R-CH=CH_ + H_SO, + HSO ~
222 4 heat 224 4
or the dialkyl sulfate ester can be formed:
Reactions of sulfuric acid with alcohols may have many undesirable charac-
teristics which make them impractical for a sulfuric acid fixing reagent.
These characteristics include:
• Reactions proceed best in concentrated sulfuric acid,
although sec-butanol acid sulfate formation has been
carried out in 60% sulfuric acid (45).
• Excess sulfuric acid is required.
• Many side reactions occur.
• Reaction products, in many cases',; are not known.
• Reaction conditions must be carefully controlled.
28
-------
Oxidation-Reduction Reactions — Sulfuric acid is not a strong oxidizing
agent, the potential for the reaction I^SO^ •*• H2S03 being only 0.20V. This
type of reaction for fixing sulfuric acid in the H_SO form is not very
promising. Sulfuric acid, however, can be fixed as the metal sulfate by
active metals which lie above hydrogen in the redox potential series. The
following oxidation reduction reaction occurs in this case:
^M + 2H+ -> - MX+ + H9(g).
X X £•
Metals which can undergo this type of reaction to produce the metal sulfate
salt include aluminum, titanium, zirconium, zinc, chromium, iron, cadmium,
cobalt, nickel, tin, lead, manganese, and many of the rare earths. These
metals are all possibilities for collecting sulfuric acid by impaction on
a foil.
The salts Ce(SO.) and TIHSO^, which decompose at 195 and 120°C,
respectively, look particularly interesting. Although metal foils are good
candidates for sulfuric acid collection, there are many parameters which should
be evaluated experimentally to determine their usefulness. These include:
• Eliminating excess water from the reaction zone by
heating the incoming air stream or foil, or both.
• Determining the catalytic effects of the foil for
oxidation of S02 to SO.^.
• Determining interferences from oxide formation.
Separating the product of the metal sulfuric reaction
from other atmospheric sulfates. This could possibly
be accomplished by selective decomposition or selective
extraction.
Acid-Base Reactions — Sulfuric acid is an excellent non-aqueous protonic
solvent. In its concentrated form, it is capable of protonating even the
weakest organic bases. Organic compounds such as ketones, carboxylic acids,
esters, ethers, amines, amides, and even some aromatic hydrocarbons, react
with concentrated sulfuric acid to form a protonated complex:
R-CO + H9SO -> R- COH + HSO/
R - * 4
29
-------
This ability to protonate and dissolve weak bases is used in organic quali-
tative analysis schemes to aid in identification of functional groups pre-
sent. It can also be used for colorimetric quantitative analysis due to the
formation of distinctly colored complexes with certain compounds (46,47).
In general, these reactions do not occur in dilute sulfuric acid solutions,
and are of little interest for a sulfuric acid fixing reagent. The excep-
tion to this is the amines. Many organic amines are sufficiently strong bases
that they will react with sulfuric acid even in very dilute solutions.
Amines react with aqueous acid solutions to give an amine salts of the
types :
(R2NH2+)2S04
The reaction is virtually instantaneous. For example, the protonation of
trimethylamine has a rate constant of k = 2.5 x 10 liter mole sec (48)
Assuming that the reaction (CH3)"N + H -> (CH3KNH+ is second order, the
time required for the H concentration to drop to half its initial value is
KC
o
where K is the rate constant and C is the initial molar concentration.
o
Using the above equation, droplets of H?SO, over a range of 0.1 to 10 molar
Q
concentration would have a half-life of approximately 0.4 x 10 to 0.4 x
10 seconds. In addition to the reaction rate advantages of acid-base
reactions, many of the low molecular weight amines have the added advantage
of being highly volatile. Thus, it appears that low molecular weight
amines have great potential as gaseous fixing reagents.
Most acid-base indicators are, themselves, weak organic acids that
show one color in the acid form and another in the anionic form, i.e.,
H In" H+ + In"
(Acid Form, (Basic Form,
Color A) Color B)
30
-------
These weak acids are frequently prepared in a salt form in order to increase
»
water solubility and enhance the acid's usefulness as an indicator:
Na+In~ -»• Na+ + In"
Thus, when the salt form is dissolved, it dissociates, and the anion form
reacts with the H present to give the characteristic color, dependent upon
the pH of the solution. The cation from the salt should then be free to
react with the anion of the acid under consideration. Gentle heating of the
collection surface during sampling should volatilize water and cause the
sulfate salt to form instantaneously. Judicious choice of the cationic
portion of the indicator salt should result in the rapid formation of low
decomposition temperature metal sulfate salts.
Many of the salt forms of these acids are prepared by simply neutraliz-
ing the free acid form with a strong base. For example, the sodium salt of
methyl red can be prepared by adding NaOH to an equal amount of the parent
methyl red acid. Theoretically, it should be possible to prepare any metal
salt by adding the metal's basic form (the hydroxide or oxide) to the acid.
This should allow the preparation of a metal salt indicator that can react
with the sulfate ion to form a sulfate adduct with a desirable decomposition
temperature.
In selecting a possible metal ion for indicator synthesis, several
criteria should be considered:
1) The ionization proceeds in two steps:
H2S04 •«-»• H+ + HS04~ k. » 1
- + = -2
HS04 «-»• H + S04 k = 1.2 x 10
Therefore, the primary anionic species present is the
hydrogen sulfate ion.
2) The metal hydroxide or oxide must be sufficiently basic
to react fairly completely during neutralization of the
indicator acid.
3) The resulting sulfate or hydrogen sulfate should have
as low a decomposition point as possible.
31
-------
4) The metal should have a valence state of one or two, as
most indicators are monobasic acids.
Thallium (I) appears to fulfill all these criteria. Its basic form,
T10H, is a strong base which should be capable of reacting with an acid
indicator. T1HSO, has a decomposition point of 120°C, with Tl^O^, S02 and
0~ as products.
The choice of indicator is much less constricted. The indicator
selected should not contain sulfur; thus, any interference from decomposi-
tion of the indicator itself is eliminated. A high pH for indicator color
change is desirable, as this implies a high pK. , and, therefore, a strong
in +
tendency to react with protons and drive the reaction toward release of Tl
and HSO ~
4
The acid-base reaction between HUSO, and the indicator should result
in the formation of T1HSO, which decomposes at a low temperature to give
SO,,. Although other acids may react, only sulfur-containing ones could
interfere.
Reaction Kinetics Conclusions — Sulfuric acid reacts with several
classes of organic compounds, but only its reactions with amines seems to
occur very rapidly. Electrophilic addition and substitution reactions
require that the acid be concentrated almost to the point of fuming, since
the active agent is really SO,,. When concentrated H~SO, is added to highly
branched liquid alkenes, and even highly activated aromatic compounds like
phenol, the reaction is imperceptibly slow. Other reactions with organic
compounds involving sulfuric acid's oxidizing or dehydrating properties
can be duplicated by nitric acid, phosphoric acid, or a mixture of the two.
The reaction of sulfuric acid with organic amines is fundamentally
different because it is essentially an acid-base neutralization. The basic
amine group accepts a proton readily and acquires its charge, thus becoming
the counter-ion to sulfate. If 100% sulfuric acid is used, the solid amine
sulfate salt is immediately formed. If more dilute acid contacts the amine,
the acid is neutralized, but the salt does not crystallize until the water
is evaporated. Thus, it appears that the sample will have to be dried
as it is collected to insure immediate fixation. In the case of certain
high-molecular weight amines, the sulfate is insoluble in water, so the
32
-------
crystals form immediately, regardless of dilution of the acid. This, again,
leads us to the intriguing possibility of using a volatile amine vapor to
fix the acid in a gas-phase reaction. Huygen (49) found that flushing a
filter containing H2SO, for two minutes with diethylamine vapor was suffi-
cient to complex the H2SO, for a colorimetric determination. This technique,
however, does not possess the required sensitivity or selectivity.
SEPARATION OF INTERFERING SPECIES
It is essential to a successful analytical method that the substance
which is actually measured be a unique product of sulfuric acid. There
does not seem to be any means of satisfying this requirement without incor-
porating at least part of the H^SO, molecule in the adduct to be measured.
The reactions of sulfuric acid are not sufficiently unique in themselves
to form a distinct non-incorporative adduct. Specifically, sulfur from the
acid must be contained in the adduct if it is to be measurably distinct
from other products of the fixing reagent.
There are various sample collection techniques for atmospheric
aerosols which do not collect ambient sulfur gases, i.e., filters. In
addition, there are analytical methodologies which are specific for sulfur.
These options include the use of Teflon filters for sample collection and
the use of the Flame Photometric Detector (FPD) for the specific measure-
ment of sulfur. The FPD and other potential measurement techniques are
discussed in Appendix B.
The elimination of sulfur gases and non-sulfur bearing species suggests
that the most serious analytical interference to be expected in the analysis
of fixed ambient H,,SO, adducts will be from particulate sulfates. One way
of limiting this interference is to form an adduct which will volatilize or
decompose at a lower temperature than other particulate sulfates. Table 6,
compiled from various sources (50-52), gives the published decomposition
temperatures of the more common ambient sulfate particulates. As can be
seen from the table, it would be desirable for the H SO, adduct to have a
decomposition temperature of o,200°C or less. The decomposition products
of the adduct, particularly SO , could then be analyzed using the sulfur
specific Flame Photometric Detector. Therefore, if selective deomposition
33
-------
TABLE 6. DECOMPOSITION TEMPERATURES OF AMBIENT SULFATES (50-52)
Decomposition
Sulfate Temperature, "C
NH.HSO. 230
4 4
235
FeS04 255
770
ZnS04 740
CdS04 1000
CoS04 989
840
SnS04 650
1000
MnS04 850
1000
34
-------
is to be considered as a means of eliminating particulate sulfate inter-
ference, then the fixing reagent must take on the added requirement that
it form an adduct with H2SO^ that can be decomposed at a temperature of
less than 200°C.
Another possibility is to form an adduct which would selectively
precipitate or dissolve in a dry organic reagent in which interfering
particulate sulfates undergo the opposite reaction. If the fixing reagent
is soluble in an organic solvent, or can be used as the solvent itself,
then the sampling procedure is simplified. This technique will probably be
most useful for organic adducts, but it may be possible to extract H-SO.
*•* *f
fixed in certain acid salts by an organic reagent which would normally
react with H2SO,. It is felt, as previously stated, that this technique
is more problematic than thermal decomposition, due to the hydrolysis of
ambient sulfate salts in the presence of water. The present discussion
will deal with both types of potential adduct candidates.
Potential Sulfuric Acid Adduct Classes
Thermal Decomposition -
Oxygen and nitrogen in organic compounds are usually sufficiently
basic to become protonated by a strong mineral acid. Formation of oxonium
salts by protonation of the -OH group usually occurs in alcohols. Concen-
trated sulfuric acid dissolves in almost any organic compound containing
oxygen such as carbonyl compounds, ethers, etc. by the same mechanism.
However, these oxonium salts are not sufficiently stable to isolate as such;
they are useful only as soluble forms of sulfuric acid in organic solvents.
Organic nitrogen (excluding nitro-, nitroso-, and quaternary ammonium
compounds) is more strongly basic than oxygen because its unshared electron
pair is more available for protonation due to the lower electronegativity of
this element. Consequently, "amine sulfates" are more stable than oxonium
sulfates and can be isolated as ionic solids. This can be demonstrated by
adding concentrated H SO to diethylamine, which produces rapid precipita-
tion of a white solid. Most amine sulfates are soluble and completely
ionized in aqueous solution. Hydrolysis of the protonated amine produces
a fairly acidic solution, so it would seem desirable to heat the sample
35
-------
collection surface to remove water and form dry amine sulfate crystals.
This would'not be necessary for an insoluble amine sulfate; however, the
thermal stability of these compounds seems to be too high for selective
thermal decomposition analysis.
Even water soluble amines seem to be at the limit of selective analyz-
ability. Diethylamine sulfate does decompose slowly at 200°C and releases
SO more quickly then ammonium sulfate at the same temperature. Other
amine sulfates, however, such as pyridinium sulfate (from pyridine),
require a higher temperature at which the inorganic sulfates interfere.
For this reason amines and amine-derivatives in which oxygen is bonded
to nitrogen : hydroxylamines and oximes:
R
!\
N-OH C=N-OH
Hydroxylamine Oxime
were considered to have the greatest potential, particularly in the use of
a gas-phase reagent to fix sulfuric acid as it is collected. The primary
effect of oxygen in these compounds is to reduce the base strength of
nitrogen. As a result, the corresponding sulfate salts tend to be less
stable and decompose at a lower temperature then the basic amine. An
analogous difference is observed in the decomposition temperatures of
ammonium sulfate and hydroxylamine sulfate:
240° 180°
Of course, many factors besides base strength are involved in the decomposi-
tion temperature, but this seems to be a useful generalization.
Insoluble Sulfates —
Several organic amines have been investigated which are distinguished
by the very low solubilities of their sulfate salts in both aqueous and non-
aqueous solvents. These compounds are listed below:
36
-------
Compound
Benzidine (53)
Formula
Solubility of
Sulfate, g/1
0.098 (25°C)
4,4'-diaminotolane (53)
2-amino-4' -chlorodiphenyl
1 , 8 -d iamino naphthalene (55)
2-aminoperimidine (55)
perimidylammonium bromide
0.059 (25°C)
0.155 (25°C)
0.222 (25°C)
0.020 (18°C)
These reagents at first appear to be promising candidates for fixation
of sulfuric acid, both because they react with dilute acid and because they
are extremely insoluble. The latter infers that the selective solubility
of interfering species would serve as a means of separation prior to the
analysis of the sulfuric acid adduct. However, it appears likely that the
amount of sulfuric acid adduct and interfering sulfates collected during
any reasonable collection period would be so small that solubility differ-
ences between the various species would be ineffectual as a means of separa-
tion. For example, the solubility of the most insoluble sulfate known,
perimidylammonium sulfate, is 0.020 g/1. This corresponds to a solubility
of 20 yg/cc or for a 10 cc extraction volume, a total of 200 yg of perimidyl-
ammonium sulfate would be soluble. Considering that the sulfuric acid
2
aerosol lower concentration of interest to this program is 0.25 yg/m ,
a sample volume of 200 m would be required to exceed its solubility limit.
Thus, at a typical aerosol sampling rate of 14 1/min, it would take approxi-
mately 240 hours to sample a sufficient volume to exceed the solubility of
perimidylammonium sulfate. It would, therefore, appear that even for the
most insoluble sulfate known, selective solubility is not an attractive
approach.
37
-------
SUMMARY OF THEORETICAL DEVELOPMENT
Sulfuric acid is a highly reactive species of significant environmental
importance. The advent of the automobile catalytic converter and the use of
high sulfur fuels will cause the level of sulfuric acid aerosol in the envi-
ronment to increase in the future. Sulfuric acid aerosols are believed
3
to exist in concentrations of 0.25 to 50 yg/m with an aerosol particle
diameter of 0.02 to 2.0 ym. Both the aerosol droplet size distribution and
the concentration of acid in the droplet are related to environmental vari-
ables such as temperature and relative humidity.
The precise health implications of current levels of ambient sulfuric
acid aerosols ar« the basis for much scientific speculation at this time.
This speculation is due, in part, to difficulties in measuring existing
sulfuric acid aerosol concentrations. These difficulties are associated with
an inability to measure sulfuric acid in the presence of interfering sulfates
and an inability to prevent artifact formation during aerosol sample collec-
tion. It appears that these difficulties might be resolved by reacting the
acid during collection to form an adduct which can be selectively analyzed in
the presence of interfering sulfates.
Sulfuric acid aerosols can be collected for analysis by a variety of
techniques. Filtration techniques using Mitex LS filters offer the advantage
of efficiently sampling particulates of various sizes, being inert, and
relatively simple to apply. Other aerosol collection techniques, such as
precipitation and impaction collection methods, are often expensive and
difficult to utilize.
It appeared that aerosol adduct formation could best be accomplished by
utilizing reagents which are sufficiently volatile that they can be used
in the gas phase during sample collection. Gas phase fixation offers the
advantage of being able to continuously supply fresh reagent to a moving
sample stream prior to actual sample collection. The primary requirement
in the formation of sulfuric acid adducts is that the reaction between the
acid aerosol and reagent be sufficiently rapid to prevent artifact formation
on the collection surface. An examiantion of the reaction classes of sulfuric
acid suggests that only acid-base reactions are sufficiently rapid to satisfy
38
-------
the kinetics requirement. Volatile organic amines appear to offer the
advantage of both gas phase fixation and the rapid kinetics of acid-base
reactions. Calculations (Appendix S) based on a model consisting of several
volatile organic amines in a gas flow reactor system indicate that this
approach for fixing sulfuric acid aerosols is theoretically feasible.
There are many analytical techniques available that show potential for
the analysis of sulfuric acid adducts. Since it is desirable to incorporate
the sulfur from the sulfuric acid aerosol into the adduct molecule, a technique
which is both selective and sensitive for sulfur would have obvious advantages.
The Flame Photometric Detector (FPD) fulfills this requirement, as it responds
primarily to sulfur (374 nm) and is sensitive to sulfur, at the nanogram level.
Thus, the FPD would eliminate all interferences except those which contain
sulfur. Sulfur-containing gases should not interfere, since they would
probably pass through the collection filter. It is possible, however, that
some sulfur gases might react with the fixing reagent and subsequently be
collected on the filter surface.
The use of the FPD as the analytical detection method suggests that the
major interfering species would be ambient particulate sulfur compounds.
There are two methods which might be utilized to eliminate these interfering
species: selective solubility and thermal volatilization/decomposition.
Selective solubility appears to be the least promising approach because of
the low levels of acid aerosol anticipated in collected samples. The litera-
ture suggests that most anticipated environmental sulfur-containing particu-
lates would decompose well above 200°C. Thus, the selective analysis of the
adduct in the presence of other sulfur-containing species has the greatest
potential for success if an adduct can be defined which volatilizes/decomposes
below 200°C.
The preceding discussion has attempted to illustrate the theoretical
manner in which the proposed method evolved from the main possibilities for
fixation, sampling, and analysis of sulfuric acid aerosol. Rapid formation
of the adduct and convenience of sampling required a gaseous amine or amine-
derivative fixing reagent. Considerations of selective analysis added the
requirement that the adduct decompose at <200°C by releasing a sulfur gas.
The following sections describe the experimental approach, the experimental
39
-------
work and the results of the evaluation of the proposed approach for the
analysis of ambient sulfuric acid aerosols.
SYNOPSIS OF PROPOSED METHOD
1) An aerosol sample is first drawn into a glass tube and then through
a Teflon filter by a vacuum pump. Simultaneously a gaseous organic
compound containing reduced nitrogen (amine, hydroxylamine, or oxime)
is added to the glass tube through sideports. Sulfuric acid in the
sample is thus fixed as a stable adduct (the sulfate or hydrogen
sulfate salt of the fixing reagent), and collected on the filter.
2) The sample is analyzed by heating the filter to a distinct temperature
(200°C) in a closed cell with non-reactive surfaces. After five minutes,
an inert carrier gas is used to flush the sulfur dioxide evolved from the
sample into a flame photometric detector (FPD). The FPD may be cali-
brated by injecting known SO- gaseous solutions or by preparing known
lUSO, standards. Alternatively, the decomposition gases can be contin-
uously purged into a West-Gaeke Bubbler and analyzed selectively for SO,,.
The absorbance of the solution is measured at 560 mu and the concen-
tration of evolved S02 determined from a calibration generated from
sodium bisulfite standards.
3) Only compounds containing sulfur are detected by the FPD, while the
West-Gaeke technique determines only SO-. Sulfate salts, including
(NH.KSO, and NH.HSO, , do not interfere because they do not decompose
significantly at 200°C. Sulfur gases, primarily S02, may be absorbed
by the fixing reagent and collected on the filter, but the compound
thus formed can be selectively removed from the sample by decomposition
at 100°C. Finally, fixation of the acid as it is collected reduces
its acidity and decreases the probability of reaction with other sub-
stances on the filter.
40
-------
SECTION 5
EXPERIMENTAL
AEROSOL GENERATION
The main requirement for evaluating any analytical methodology is the
availability of satisfactory standards. These standards must not only
contain known amounts of the species of interest, but they must also simulate
actual sample matrices. Once the proposed methodology is evaluated using
standards which conform to the above requirements, the technique can be used
with a known degree of confidence on actual samples.
The investigation of methods proposed for the analysis of ambient sul-
furic acid aerosols was a particularly difficult problem, since there was
much speculation about their actual size distribution in the environment.
Mirabel and Katz (26) predicted that ambient conditions such as temperature
and relative humidity greatly influence aerosol growth rates and the concen-
tration of sulfuric acid within individual droplets. There appears, however,
to be two primary size ranges of sulfuric acid aerosols emerging which are
thought to be representative of the environment: aerosols in the 0.01 to
0.1 ym size range and those in the 0.1 to 3.5 ym size range (27). Thus, any
attempt to evaluate an analytical methodology for the analysis of sulfuric
acid aerosols must incorporate known standards in the size range from
approximately 0.01 to 3.5 ym.
A review of the literature pertaining to aerosol generation techniques
suggested that there was no singular technique that could be used to generate
aerosol droplets over the entire range of interest to this research. There-
fore, two aerosol generators were used during this research: one covering
the size range from approximately 0.01 to 0.1 ym, and the other from 0.1 to
3.5 ym.
It was found experimentally that the generation of a known quantitative
sulfuric acid aerosol based upon mass balance calculations was not easily
obtainable within the scope of this research. These findings were not
surprising since the complex relationship between aerosol droplet size,
41
-------
humidity, temperature and aerosol droplet-sulfuric acid concentration had
previously been predicted by other investigators (26). Alternative methods
of standardizing the aerosol generators for the purpose of closing the mass
balance also generally proved inadequate because of the lack of a viable
referee method to measure the small amounts of acid aerosol generated. The
aerosol generators, however, proved adequate for the deposition of 0.5 to
5.0 yg of sulfuric acid aerosol which was anticipated in environmental
samples. Micropipets were routinely used to deliver known amounts of sul-
furic acid to filters during the course of this research for precise quanti-
tative determinations. Each of the sample generation techniques is described
below.
Thomas Aerosol Generator (0.005 to 0.3 yim droplets)
The sulfuric acid aerosol generator used in the initial experiments was
based upon the atomizer-burner model described by Thomas, et^ al^. (40). In
this model, a dilute H?SO, solution was aspirated into an H_-0~ flame,
where it decomposed to H^O and S0~. Recombination occurred to yield sulfuric
acid aerosol. The rationale for using this type of generator was that it
approximated the process by which H«SO, aerosol is thought to be formed in
industrial or automotive combustion.
The Thomas generator, illustrated in Figure 8, consisted of a Beckman
4060 Large Bore Atomizer Burner Assembly mounted at the base of a glass
stack 1.2 M high by 15.2 cm O.D. Burner gas back-pressures were set at 10.3
mm positive pressure for H~ and 517 mm positive pressure for 0,,. The aspir-
ated solution varied from 0.01N to l.ON, depending upon the experiment,
with a typical delivery rate of 2 ml per minute into the burner jet.
Barrett, ££ al.- (57) measured the aerosol distribution of an identical
system and found that 98% of the aerosol droplets were between 0.005 and
0.3 pm.
It was thought that the problem of closing the mass balance on this
generator was associated with the stack being open at the bottom, which
resulted in varying amounts of diluent gas (ambient air) being drawn into
the stack. In addition, the aerosol temperature could vary greatly, depend-
ing upon room conditions. This generator varied quantitatively by as much
as three orders of magnitude, based upon supposedly identical samples.
42
-------
Sample
probe .-•.>?:•'•
: •""•"'.•/I
H2
o:
Solution
""Figure "8. "'Thomas aerosol generator."
Baird Aerosol Generator (1.0 to 3.0 ym droplets)
The Baird Aerosol Generator illustrated in Figure 9. was a closed system
allowing diluent air to be accurately determined. It was constructed from a
nebulizer designed for use with the Baird-Atomic D.B.-2 flame photometer, a
Fisher burner with concentric glass chimneys, and a large reflux condenser
which was approximately 1.2 m long with an outside diameter of 8.8 cm. In
operation, a dilute sulfuric acid solution was passed through the nebulizer
and aerosol droplets formed. Droplets larger than 3.0 pm strike the nebulizer
housing wall and are collected in a "U" trap at the bottom. Aerosol droplets
smaller than 3.0 ym do not collide with the housing wall and are transported
up the condenser stack to the collection filter. The condenser acts as a
stack, mixer and constant temperature bath.
43
-------
Air Supply
Figure 9. Baird
generator
The generator was found to operate satisfactorily from 414 to 776 mm
positive pressure, delivering 34 1/min of air at the lower value which was
normally used for experimental purposes. During a typical run of 30 minutes,
50 ml of sulfuric acid solution was passed through the nebulizer, 48 ml of
which was collected in the return trap of the nebulizer. Typically, acid
solutions of 0.01 N to l.ON were used, depending upon the experiment. The
Baird generator was found to be reproducible within 20%, but attempts to close
a mass balance based upon the amount of acid solution consumed proved imposs-
ible. Raabes (58) determined that the size range of the aerosol generated
by the Baird nebulizer under similar conditions was 1 to 3 pm.
44
-------
Filter
Holder Fixing
T Chamber
Control
Valve"
Figure 10. Sample and fixation apparatus
inlet for the various reagents or interferences. The double-neck joint
allowed for the simultaneous addition of both acid aerosol and fixing reagent
into the reaction chamber. The fixing chamber consisted of a 20.3 cm length
of 2.54 cm I.D. Pyrex tubing with a male and female 24/40 ground-glass joint.
At the other end of the fixing chamber, a nylon filter holder was attached
which held the Mitex LS 5 ijm filters for collection of the H0SO. aerosol
/ 4
and adducts. A Cast pump was used to draw the sample stream through the
Mitex filters with the pump exhaust being fed into a Precision Scientific
Wet Test Meter which accurately measured the total sample volume. Sample
flow rate was controlled by the use of a valve located between the filter
holder and pump.
A second sampling system was also utilized in which the temperature of
the incoming aerosol gas streams could be elevated. This was done in order
to study the effect of collecting aerosols at elevated temperatures and to
determine if unfixed I^SO, aerosol could be volatilized and efficiently
passed through a prefilter.
The device, shown in Figure 11, consists of a 30 cm x 1.8 cm I.D. heavy
wall Teflon tube heated by glass-backed tape. The end of the 30 cm heated
46
-------
• 11NCH INSULATION
THERMOCOUPLE
Figure 11. Modified aerosol collection system.
section can be used with or without a Mitex Teflon prefilter which allows
the volatilized acid to pass while collecting the remaining particles.
Once through the prefilter, the volatilized acid is mixed with fixing
reagent in a 15 cm Teflon mixing chamber. The adduct is then collected on
a second Teflon filter at the end of the 15 cm section. The fixing chamber
is wrapped with a glass-backed heating tape and insulated in order to accu-
rately control temperature. The temperature range over which the sampler
operated was from ambient up to (x>200°C.
All aerosol samples, using both collection systems, were generally
collected over a period of 2 to 30 minutes, depending upon the concentration
of the aspirating solution and the method of aerosol generation. Total
volume sampled as measured by the wet test meter was typically 0.03 to 0.95
m3. Aerosol sample collection flow rates were approximately 15 1/min with
the atomizer-burner generator and 34 1/min with the Baird generator.
Fixing reagent vapor, as shown in Figure 11, was generated from a
bubbler and delivered to the entrance of the reaction chamber. The bubbler
consisted of a 50 ml Erlenmeyer flask with a reagent level of approximately
2 cm. The flow rate through the bubbler was normally held at 0.5 1/min.
Theoretical considerations (Appendix A) predicted the approximate concen-
tration of various candidate reagents in both the reagent gas stream and the
aerosol sample stream. Heating tape was used to maintain the temperature of
47
-------
the reagents at room temperature and prevent gross concentration changes
caused by solution temperature fluctuation.
ANALYSIS OF SAMPLES
Two types of samples were subjected to various analytical methodologies
during this investigation. One type of sample consisted of those prepared
using the aerosol generators or pipet delivery method and were normally
analyzed by a combination of thermal decomposition and sensitive analytical
techniques. These determinations served to indicate the usefulness of the
proposed methodology under simulated field conditions. The second type of
sample consisted of those prepared in larger quantities in order to obtain
additional information about a specific area of the proposed methodology.
Decomposition Apparatus
The device constructed to analyze :the adducts consisted basically of
a heated cell to thermally decompose the sample, and a three-way valve
arrangement by which a helium purge gas could be used to deliver the decompos-
ition products to the analytical detector. Normally, the analytical detection
was accomplished by use of a Flame Photometric Detector (FPD) or the West-
Gaeke (59) technique for the determination of S0?. Figure 12 is a sketch
of the sample decompositon chamber. The heating tube (19 cm long x 1
cm diameter) is constructed of Teflon and is surrounded by a large mass of
aluminum to reduce temperature fluctuations. The aluminum housing also serves
to maintain an even temperature profile across the cell's length. The 0-
ring seals are outside of the heated zone to eliminate the degradation problem
associated with the 0-rings in the original prototype.
Flame Photometric Detector (FPD)
The FPD used in this investigation was a Melpar Model 100 integrated
with the electrometer of a Hewlett Packard 5750 gas chromatograph. Typical
electrometer settings during this research were on a range setting of 10
and an attenuation of 32, which resulted in a detection sensitivity of
-10
approximately 10 amps. Basically, the decomposition gases were simply
purged through the FPD at 60 cc/min with the aid of helium carrier.
48
-------
Figure 12. Revised sample decomposition chamber.
49
-------
West-Gaeke S02 Analysis Technique
The West-Gaeke technique (59) is used for the analysis of SO^- In
this method, SO- is collected in an aqueous solution of 0.1 M sodium tetra-
chloromercurate where it is fixed as the stable disulfitomercurate (II)
ion. When a p-rosaniline hydrochloride-hydrochloric acid mixture and
formaldehyde are added, a violet color is produced proportional to the
amount of fixed S0« present. The absorption maximum at 560 my is used
for a spectrophotometric measurement. This method will detect the equiva-
lent of 1 yg H9SO. converted to S0? and dissolved in a 10 ml collection
^ H ^
solution.
Basically, use of the West-Gaeke technique consisted of preparing :
samples on Mitex filters using the aerosol generator or by pipet. These
samples were then heated to 200°C individually in the sample decomposition
cell and the evolved gases bubbled into 10 ml of the collecting solution
with a helium carrier gas. After collecting for 15 minutes, the solution
was treated with 2 ml 0.2% formaldehyde and 5 ml of 0.04% acid bleached p-
rosaniline dye. After allowing 30 min for the color to develop, the absor-
bancy at 560 my was compared to that of a curve prepared by using standard
solutions of sodium bisulfite in sodium tetrachloromercurate. A Beckman
DK spectrophotometer was used to measure the absorbance.
Supporting ~Instrumental Analysis
Analytical methods other than the FPD and the West-Gaeke technique
were used to obtain additional information concerning adduct decomposition.
These techniques, however, were not considered adequate for the analysis
of actual environmental samples.
Differential Thermal Analysis (DTA) - DTA was used to obtain information
about the temperatures at which candidate adducts and potential interfering
substances decomposed. An Aminco Model 442 was used for the purposes of
this investigation. Generally, samples were prepared for differential
thermal analysis by mixing H^SO, solution and adduct forming reagent together
in a ratio of approximately 1:2. The samples were then purified by conven-
t latin I. laboratory foclmlques such as filtration.
50
-------
Mass Spectrometry - Mass spectrometry was utilized to identify the gases
released by decomposing adducts. All of the adducts examined were found to
release S02 at 200°C, but no detectable SO . There was some ambiguity in
these results, however, because even unreacted H2SO, gave predominantly SO-
at this temperature. Generally, samples subjected to mass spectrometric
analysis were prepared by gas phase fixation utilizing the aerosol generators
and subsequent collection on Mitex LS filters. The filters were inserted
into a sealed 0.95 cm O.D. Pyrex tube which was then attached to the mass
spectrometer inlet with a glass stopcock. The sealed vial was then evacuated
and the stopcock opened to allow the decomposition gases to enter the mass
spectrometer. The sealed tube was slowly heated to 200°C and a light
oscilliograph used to record the spectra generated. An EAI Model 250 quad-
rapole mass spectrometer was used for these studies.
Infrared Spectroscopy - A Beckman Acculab-6 infrared spectrometer was
used on several adducts to obtain structural information indicating that
the products were the result of an acid-base reaction. Typically, samples
were prepared by allowing equal volumes of H2SO/ solution and fixing reagent
to react in miniature reaction vials (1 ml). The products were then sub-
jected to various clean-up procedures and applied to KBr salt crystals for
analysis.
Ammonia Specific Ion Electrode - An Orion Model 95-10 ammonia electrode
was used to investigate the reactivity of ammonia with H~SO, aerosol and
several of the adduct candidates. Basically, these experiments involved
forming and collecting the various adducts in the usual manner with the
aerosol generators, and subsequently exposing them in place to a dynamic
ammonia environment of 250 ppm. The samples were then removed from the
collection device and placed in 100 mis of deionized water. One ml of 10
molar sodium hydroxide was then added to the solution to generate NH«,
which was detected by the ammonia specific ion electrode. The amount of
NH~ generated was indicative of the reactivity of each adduct with NH,-
Heilige Turbidimeter - A Heilige Model 8000-TS Turbidimeter was used
for analysis of sulfate (sulfuric acid or sulfate adduct) deposited on filters
by micropipet. The filters were placed in 50 ml of deionized water. Ten
51
-------
milliliters of a sodium chloride-hydrochloric acid solution were then added,
followed by 0.29 grams of solid barium chloride which was used to form
the precipitate barium sulfate. The mixture was then stirred using a magnetic
stirrer for five minutes, placed in the appropriate turbidimeter tube and
the amount of sulfate present determined by correlating the turbidimeter
reading to a standard graph of known SO, versus the turbidimetry reading.
Drager Tubes - Drager tubes were used to approximate the concentration
of ammonia in the gas streams passing through collection filters during
adduct interference studies. Drager tubes consist of 15.2 cm x 0.95 cm O.D.
glass tubing packed with a solid material. The solid material is impregnated
with a chemical that changes color as a gas containing NEL is passed through
it. The length of discoloration as measured by a calibration scale on each
tube, is indicative of the concentration of NH_ in the gas stream. The tubes
are accurate to ± 20% of the measured value. Measurements were made by pass-
ing a known volume of the interference gas stream through the Dra'ger tube
immediately after the filter holder containing a blank filter.
52
-------
SECTION 6
RESULTS AND DISCUSSION
PRELIMINARY EXPERIMENTS
Preliminary experiments were initiated at the onset of this research
to substantiate the theoretical assumptions from which the proposed approach
evolved. These studies were to obtain information in the following areas:
• Verify the decomposition temperatures of common sulfates
thought to have relatively low decomposition temperatures.
• Establish preliminary evidence as to the nature of the
sulfuric acid/reagent reaction and the products formed.
• Establish the decomposition temperature of various
sulfuric acid adducts.
• Establish the operating parameters and sensitivity
of the flame Photometric Detector (FPD) for both unfixed
sulfuric acid and S02-
• Calibrate the Flame Photometric Detector to establish
that the aerosol generators will deposit sulfuric acid
in the range of environmental importance.
Decomposition of Sulfate Particulates
Differential Thermal Analysis (DTA) was used to obtain preliminary
information about the temperatures at which potential interfering sulfates
decompose. Figure 13 is a composite DTA recording of several common sulfates
which have relatively low decomposition temperatures. All are above 200°C,
except for the melting (sharp peak) of KHSO, at 180°C, and a slight decompos-
ition of CuSO, at 200°C. More importantly, ammonium sulfate and ammonium
hydrogen sulfate are stable to 250°C and 230°C, respectively. This indicates
that an adduct which decomposes and releases SO™ at 2QO°C may be analyzed
without interference from these substances. All samples were taken from
laboratory reagent bottles and ground to a powder prior to analysis.
53
-------
SO 100 150 200 250 300 350 '
T °C (Corrtctwl for CHROMEL AL.UMEL Thmwocouplti)
400
450
500
Figure 13. Composite DTA of potential interferences.
Fixing Reagent Reactions
The fixing reagents proposed in this research have been described as
volatile amines and amine-derivatives (hydroxylamines and oximes). One
compound in each category was evaluated, preliminarily, although this is
not to imply that the compound chosen was necessarily the most desirable one.
The compounds were:
(C2H5)2NH (C2H5)2NOH CH3CH=NOH
Diethylamine (DEA) Diethylhydroxylamine (DEHA) Acetaldoxime (AAO)
All are liquids at room temperature, but have sufficient vapor pressure to
generate a substantial vapor from a bubbler.
The functional groups of the compounds DEA, DEHA, and AAO are all
sufficiently basic to form the bisulfate salt from sulfuric acid. Base
strength decreases in the order: DEA > DEHA > AAO. When concentrated H2SO
was added to excess DEA or AAO, white crystals immediately formed and
crystallized out of solution. They were water-soluble and turned black
(decomposed) when heated. These physical properties, as well as IR spectral
54
-------
evidence, strongly suggested an ionic sulfate or bisulfate salt. DEHA
formed a very viscous yellow oil when H9SO, was added, but this oil was
-1
also water-soluble and decomposed when heated. A peak at 1650 cm in the
IR spectra of both the DEHA and AAO adducts (Figures 14 and 15, respectively)
was absent in the unreacted reagent and acid. This suggested an N-H bending
vibration from protonation of nitrogen, as postulated in the formation of
the bisulfate. The product of the reaction between DEA and H2SO, was
previously shown by Huygen (49) to be diethylamine sulfate.
Thus, it appeared that these compounds do react with sulfuric acid by
proton transfer, forming a sulfate of bisulfate salt. There was no reason
to suppose that the product would be different when formed under sampling
conditions.
Adduct Decomposition in DTA Apparatus
DTA was used as a preliminary means of establishing the temperature
at which the various adducts decomposed. DTA of the products obtained from
the reaction of l^SO, with DEA, DEHA and AAO showed that a decomposition
occurred below 200°C in each case. Release of a sulfur gas at 200°C was then
confirmed for all three adducts by heating in the decomposition chamber and
purging the gases into the FPD. Mass spectra indicated that S02 was the main
gas, and this was supported by the West-Gaeke bubbler (which does not detect
_?
S03 or S04 ).
Several other amine sulfates were also examined by DTA: perimidylammonium
sulfate (an insoluble salt) and pyridinium sulfate (from pyridine). The
DTA, as shown in Figure 16, indicated that no decomposition occurred under
250°C. Consequently, these two species were not considered desirable as
sulfuric acid adducts.
Flame Photometric Detector (FPD)
In order to insure that the amount of H-SO, aerosol being deposited on
filters by the aerosol generator was in the quantitative range anticipated in
actual environmental samples, it was necessary as a preliminary step to determine
the approximate amount of acid being deposited. This was accomplished by
comparing the response characteristics of the FPD for both H^SO, aerosol
55
-------
• - JJ
n II M it n » M » 48
-
[(C2H5)2NOH2]-H2S04-
- - gr- •^~.-j«T=fej
/-. -- _F- %-— -1-=-^
WAVEIEMGTH TM AUCKONS
5 556
11 14 16 It 20 23 30 40
2000 1800 1400 1400 1200
WAVtNUMBEt CM'1
WAVElfNCTH IN MICTOMS
800 600 SOO 400 300
4.S 5 5.5 4 t.i 7 7.5 S 9 10 II II 1.1
H2SO. i
Figure 14. IR spectrum of a) diethylhydroxylamine sulfate
b) diethylhydroxylamine and c)' sulfuric a"cid. '
56
-------
WMUMOTH M MKXONt
4.1 1 IS
a)
b)
c)
| (CH3HCNOH)2'H2S04 m
I CH-HCNOH
3
7 7i « • '0 II « 14 u 11
1400 1200 1000 KB ,«, M, ^
Figure 15. IR spectrum of a) acetaldoxime sulfate,
b) acetaldoxime, and c) sulfuric acid.
57
-------
ISO 200 250 300 350 400
T. °C (Corrected for CHROMEL ALUMEL ThcrmocouplH)
450
500
Figure 16. (PDA)2SO and
DTA.
as collected from the aerosol generator to the FPD response of known quantities
of SO™. The amount of H?SO, aerosol anticipated in environmental samples
is in the range of 0.05 to 5.0 yg, depending upon sample collection time.
Initially, when samples from the aerosol generator were run through
the FPD after being volatilized in the decomposition chamber, very peculiar
recordings were obtained. The main peak was extremely broad, requiring
several hours to tail to baseline, and was often preceded by a small sharp
peak. Figure 17 illustrates some of these recordings. It was suspected
that these recordings were artifacts of a saturated detector. As a result,
several experiments were carried out with known S09-air mixtures to clarify
the response characteristics of the FPD.
For these experiments, the sample cell was removed and the carrier line
was connected directly to the detector inlet. A silicon rubber septum
was inserted in the line for injecting S0_ with a syringe. The source of
z*
S02 was a compressed-gas lecture bottle, fitted with a valve and a silicon
rubber septum. Dilutions of S02 in air were carried out in a 305 cc siliclad-
coated glass vessel fitted with a rubber septum. Two syringes with Teflon
pressure-lock valves were used, with maximum volumes of 100 pi and 500 yl.
58
-------
ON
o
W
a,
w
-------
Several blank injections were made after each 'sample injection to inasre
that the syringes were clean of residual SC^-
Microliter amounts of pure S02 from a gas cylinder were injected in
the first experiments. Figure 18 shows the effect of injecting progressively
larger volumes of S02. There was a slight increase in.peak width and in
separation between the two peaks, but these were not proportional to changes
in the SCL level. It was apparent that all of these peaks were resulting
from a saturated detector. When these peak heights were compared to those
previously obtained from samples (Figure 17), it was evident that the aerosol
samples had been saturating the detector. It was determined that the delay
between the first sharp peak and the second broad one was due to signal
suppression by excess sulfur atoms in the flame reabsorbing the emitted
light.
o% 4
I
• i
£
a
f
30
60
90
f
120
ISO
of S0
ISO
t
210
300
Figure 18. Effect of injecting progressively
larger S02 volumes.
In order to bring the sample size down to the quantitative response
range, various dilutions of S02 were made. When microliter volumes of these
dilutions were injected, the peaks were much sharper and their heights
were related to the amount injected as Figures 19 and 20 show. There
was some difference observed between the two syringes in Figure 20, the
smaller one producing less response from the same volume of sample. This
was possibly due to the greater surface area-to-volume ratio of the smaller
60
-------
4444
O.OS 0.10 0.20 0.30 0.50 0.50 0.70 1.00
tig of S02
Figure 19. Quantitative FPD response to injected
SO,,.
c
^«
y
7
A
*>
O • 100 jil syrinjl
Q « 500 M< lyringt
0.01 0.02 0.03 0.04 O.OS «.M 0.07 0.08 0.09
Figure 20. Quantitative response of FPD"to
61
-------
syringe, causing some retention of S02 by glass surface. The response,
however, was fairly linear from 0.01 yg up to about 0.6 yg, beyond which
i
there was no further consistent increase in peak height. It was established
on other electrometer settings that the saturation level always occurred
at about 0.6 yg of S0_ and did not depend upon the electronics of the
instrument.
If the saturation level of the detector is 0.6 yg S02, the corresponding
amount of H2SO is 0.9 yg. It is, therefore, not surprising that the large
acid aerosol samples which gave the FPD recordings of Figure 17, and were
estimated to contain 20 to 30 yg H2SO, (by an independent pH measurement),
did indeed saturate the detector.
These experiments, in which SO was injected into the FPD, were crucial
in the interpretation of the FPD recordings from aerosol samples. It was
established that the peaks of Figure 17 were saturation artifacts, and, indeed,
the H?SO, aerosol samples which had been collected were too large. In
subsequent runs, sampling procedures were modified to collect smaller samples
and the resulting peaks resembled the unsaturated S02 peaks. The electro-
meter setting was at a range of 10 and attentuation at 32 during all remaining
FPD experiments, which resulted in a full-scle recorder sensitivity of
m"10
10 amps.
DECOMPOSITION CHARACTERISTICS OF ADDUCTS (Unmodified Thomas Generator -
0.005 to 0.3 ym droplet)
The FPD analysis apparatus was used to analyze fixed aerosol samples
from the Thomas aerosol generator by two major procedures. In the rising
temperature procedure, the sample filter was inserted into the decomposition
cell at room temperature. With carrier gas flowing through the cell to
the FPD, the cell temperature was then slowly raised. The rising temperature
procedure served only to determine the decomposition temperature of the
various adducts and interferences. It was not considered to be an alter-
native to the proposed methodology, which consisted of adduct decomposition
at fixed temperatures.
In the fixed temperature procedure, the sample was inserted into the
decomposition cell which was preset at 120°C. After five minutes at 120°C
the by-pass valve was switched and the evolved gases passed through the FPD
62
-------
with the aid of a helium carrier. The initial temperature of 120°C was
chosen because DTA had previously shown that all of the adducts decomposed
between 120 and 200°C. Thus, the 120°C temperature served only to volati-
lize residual unfixed H2SO^ aerosol, which, when detected by the FPD, gave
an indication of the completeness of the fixing process for each candidate
fixing reagent.
Once the 120°C step had been completed, the temperature of the decompos-
ition cell was rapidly raised to 200°C, with the carrier gas again bypassing
the cell. After five minutes at 200°C, the valve was switched to sweep
gas evolved from the sample into the FPD. The temperature of 200°C was
chosen as the second decomposition step, since DTA had shown that all of
the adducts decompose below this temperature.
Rising Temperature Procedure (Unmodified Thomas Generator)
This procedure was used to determine the temperature at which any sulfur
gas was first evolved from the sample. As the cell temperature increased,
it was marked on the x-axis of a constant-feeding strip-chart recorder. The
pen of the strip-chart recorder on the y-axis was connected to the FPD.
The temperature at which the pen rose sharply from baseline was taken as
the adduct decomposition temperature.
The results of this procedure strikingly revealed the effects of fixation
on the collected aerosol. As shown in Table 7, the evolution of S0« in
every case occurred at higher temperature with the fixed samples than with
the unfixed H SO^.
TABLE 7. TEMPERATURE OF INITIAL FPD RESPONSE
FORM ADDUCTS (UNMODIFIED"THOMAS GENERATOR)
H2SO, Alone 90
H2SO, + AAO 140
H0SO. + DEHA 190
i 4
H2S04 + DEA 190
63
-------
Multiple runs indicated that the decomposition temperatures were reproducible
to within + 10% of the values given in Table 7. This level of reproducibility
was not an indirect effect of excess reagent on the filter, since the same
results were obtained when the samples were thoroughly flushed with clean
air prior to analysis. Of the three reagents, it appears that AAO forms
the adduct which is easiest to thermally decompose.
Fixed Temperature Procedure (Unmodified Thomas Generator)
By holding the adduct in the sample cell for several minutes before
sweeping it into the FPD, the evolved S02 is concentrated. This procedure
is, therefore, more sensitive to small samples than the previous procedure.
If care is taken to collect samples that are not too large, the peaks obtained
when the gas is admitted into the FPD are sharp, and their height may be
taken as a measure of the H0SO. collected.
2 4
The results of the rising temperature procedure showing that the adducts
decompose between 120 and 200°C were confirmed by the fixed temperature
procedure. As can be seen in Figure 21, when the H~SO, aerosol (Thomas
generator) was not fixed, the FPD showed the presence of a sulfur gas at
120°C and additional heating at 200°C showed that all of the unfixed acid
had been volatilized at 120°C. When H-SO, aerosol (Thomas generator), however,
was fixed by the simultaneous addition of AAO in the fixing chamber, the
AAO adduct gave no FPD response at 120°C, but a pronounced response at 200°C.
Thus, it was demonstrated that the acid aerosol was fixed completely and
rapidly since no FPD response was detected at 120°C, even when only 15
seconds of reagent-acid contact was allowed after aerosol collection.
Similar results were obtained for the DEHA and DEA adducts.
The rate of decomposition of the AAO-adduct at 200°C was also evaluated
by the fixed temperature procedure. The procedure was to repeat the
heating period, i.e., 200°C for 5 minutes, with the valve on bypass several
times. As Figure 22 shows, the second peak was proportionately smaller
when the first heating period was 15 minutes, than when it was 5 minutes.
In other words, decomposition was more complete after 15 minutes (about
80%) than after 5 minutes (about 50%). When a particularly small (<20 ng)
64
-------
ON
Ui
Unfixed
1 hr Heating «t 120°
After 5 min
at 120°
Alter 5 min
•t 200°
Adduct
After 5 min
•t 120°
A. Unfixed H-SO.
i 4
B. H0SO./AAO adduct
2 4
Figure 21. Completeness of fixation by AAO
(fixed temperature procedure).
-------
ON
After 5 min
»t 200°
After 2nd 5 min After 3rd S min
at 200° at 200°
After!
5 min
•t 200°
After S min
at 200°
A. Decomposition after Repeated
5 Minute Heating Periods.
B. Decomposition after
Repeated 15 Minute
Heating Periods.
Figure 22. Completeness of decomposition of AAO/^SO,
adduct after 5 min (a) and 15 min (b)•
-------
sample was analyzed, decomposition seemed to be essentially complete after
the first 5 minute period, as shown in Figure 23.
200 (1)
200(2)
TEMPERATURE CO
200 (3)
Figure 23. Typical FPD trace from AAO-treated sample.
i
From these experiments, it is evident that gas-phase fixation with
these reagents occurs rapidly and completely, as indicated by a higher
thermal stability of the fixed acid on the filter. The AAO-adduct is stable
at 120°C, but is approximately 50% decomposed after 5 minutes of heating at
200°C. Finally, the FPD peak height demonstrates the amount of acid origin-
ally collected is on the order of 1 yg, which is anticipated for environ-
mental samples.
Accuracy (Unmodified Thomas Generator)
The accuracy of an analytical method is defined as the degree to which
the measurements obtained agree with the true value, as determined by an
independent method of known accuracy. Alternatively, the accuracy can be
determined by comparison of the measured values with a known standard con-
centration. Since there is no independent method of known accuracy for
analysis of atmospheric H-SO , the first approach cannot be strictly ful-
filled. Moreover, the unpredictability of this type of aerosol generator
(Thomas) makes it difficult to calculate the exact amount of acid deposited.
Thus, neither of the above approaches allows for the absolute determination
of accuracy when the Thomas generator is used as the lUSO, aerosol source.
67
-------
The Thomas generator was later modified to give more consistent aerosols and
is discussed under the Modified Thomas Generator (page 91). The present
discussion, however, will deal with the Thomas generator utilizing the
original unmodified flame burner design.
One initial indication of accuracy which can be shown is that the
measurement accurately reflects the difference between two samples of known
size relationship to one another. The simplest manner in which to produce
samples of a known size relation was to halve a sample filter, then halve
it again, etc. Figure 24 shows the FPD trace obtained from each segment
after 5 minutes at 200°C. The peak heights definitely showed the size order
of the corresponding samples, although the magnitudes were not precisely
uniform. This may have been due to nonuniform distribution of adduct on the
filter, unsymmetrical cutting, or random sample loss during sampling.
V*
1
8- *
&
e
0-
S 3
1/2 Filter
1/4 Filter
ilter
1/16 Filter
\ljjk Filter
Figure 24. AAO-fixed filter portions after
heating 5 min at 200°C.
68
-------
In addition to the above tests, several different samples were generated
by collecting different volumes from the aerosol generator (i.e., varying
sampling time). Those which had filtered the largest volume of air also
produced the largest peaks, as was expected. Figure 25 shows a plot of data
obtained from these runs.
10
'o
X
3
ill
E
0.02
04 0.06
CUBIC METERS
0.08
0.10
Figure 25. AAO-treated samples, different sampling
volumes, 5 min. at 200°C.
It was concluded from these data that peak height was an indication of
the amount of acid adduct in a sample. The only possible way to independently
verify this measurement would be to prepare samples by a different method,
with which the amount of acid placed on the filter would be known. Many
researchers have employed micropipet deposition of dilute H-SO -methanol
or acetone solutions for this purpose. This approach was adopted later in
the study, and is discussed on page 85.
Precision (Unmodified Thomas Generator)
Precision is the degree to which the values obtained by an analytical
method are reproducible. Evaluation of precision placed stringent demands
on the ability to generate known samples. Specifically, it was necessary
69
-------
to generate two or more samples containing the same amount of adduct with
as little variation as possible. The similarity in peak heights from the
identical samples was then a measure of precision. The Thomas aerosol
generator used in these experiments was known to have poor reproducibility.
The problem was approached by designing a sampling probe for the generator
so that two samples could be collected simultaneously, and would, therefore,
be identical. The data from a series of three runs are shown in Figure 26
and summarized in Table 8.
Sample
No. 1 No. 2
Run No. 1
Sample
No. I No. t
Run No. 2
Sample
No. 1 _ No. 2
Run No. 3
Figure 26.
Pairs of AAO-fixed samples collected
simultaneously.
TABLE 8. PEAK HEIGHTS OF SAMPLE PAIRS COLLECTED SIMULTANEOUSLY
Peak Height
Run
1
2
3
Sample
3.4
8.3
5.4
Sample #2
3.1
7.7
5.8
% Difference
10.0
7.2
6.9
70
-------
For each pair of samples, the variation in peak height is no more than 10%
It was concluded from this evidence that precision is no less than 90%
and is probably greater.
Interference Studies (Unmodified Thomas Generator)
Ammonium Sulfate and Bisulfate —
Differential Thermal Analysis (DTA) of ammonium sulfate and bisulfate
previously presented as Figure 13 'indicated that these two salts are stable
to at least 250°C and 230°C, respectively. The temperature at which these
compounds release a sulfur-containing vapor in significant quantities is
the subject of some dispute. Dubois, et^ a!U (60) discussed the use of
(NH,) SO, as a standard for diffusion at 195°C. Maddalone, et al. (39)
however, reported that a thermal gravimetric analysis (TGA) of (NH.^SO,
indicated no significant decomposition until 250°C, in agreement with the
present results. Erdey, et al. (61) have reported the following reactions:
NH3
Experiments with the FPD indicated that no volatile sulfur species
were released by either of these salts until at least 240°C, and there was
some ambiguity when a signal was recorded at this higher temperature, due
to background sulfur in the system. Ammonium sulfate and bisulfate, finely
ground from a reagent bottle, gave no signal at 200°C with either of the
procedures described in the preceding section. When ammonia was passed
through a filter on which the acid aerosol had been collected unfixed, as
is normally the case in environmental sampling, no signal was detected
at 200°C. Interference from this source was apparently not present.
Sulfur Gases—
Since the FPD responds to sulfur in any form, it was recognized that
the ambient sulfur gases, SO , H~S and COS, may interfere by reacting
with the gaseous reagent. They may then be collected on the filter as a
71
-------
sulfur-gas reagent complex. These sulfur gas-reagent complexes would inter-'
fere with H2SO, adduct measurement, if their thermal stability was similar
to that of the adduct.
To test this possibility, each of these gases was mixed with reagent
vapor in a glass "Y" and drawn through a filter for several minutes. The
filter was analyzed by the rising temperature procedure in order to deter-
mine both the presence of a volatile sulfur species and the temperature
at which it was produced.
As shown in Table 9, of the three gases, only SCL was significantly
collected, and in every case, the SO^-reagent complex had a very low thermal
stability compared to the H2SO,-adduct. This was independently verified by
mass spectral evidence. Figure 27 compares the FPD recordings obtained
using the rising temperature procedure for unfixed acid, AAO-fixed acid,
AAO-S02 complex, and (NH,)2SO,.
TABLE 9- TEMPERATURE (°C) OF FIRST FPD RESPONSE
FROM SULFUR GAS-REAGENT COMPLEXES
H0S COS
DEA 25 60*
DEHA 25 -
AAO 30 - -
*Very little absorbed
First, it appears that a fa±rlj high concentration of SO (>25 ppra)
must be present before a significant amount is collected. Sulfur dioxide
was being produced by the aerosol generator in low concentration (5 ppm by
Dra'ger tubes), but none was ever detected at the 120°C step during routine
runs. Secondly, even if sufficient concentration is present, the collected
complex can be rapidly and selectively removed from the sample at 100°C,
prior to H2SO,-adduct analysis.
The chemical nature of these compounds is uncertain. Grundnes, et al.
(62) have examined the formation of charge-transfer complexes in the
gas phase reaction of S02 with amines. Studies of the analogous reaction
72
-------
tu
to
Ul
ec
60 80 100 120 140 160 180
AAO/H2S04
200
20 40 60 80 100 120 140 160 180
TEMPERATURE (°C)
ui
to
O
AAO/S02
TEMPERATURE (°C)
Ul
g
Ul
ec
t
i i i i i i i i i
20- 40 60 80 100 120 140 160 180
TEMPERATURE (8C)
-------
between S0~ and NH« have found mostly sulfite and bisulfite salts (63,64).
-2
It may be possible, however, for the collected SCL to be oxidized to SO, s
if catalytic conditions are present (64). For this reason, it might be
desirable to heat the filter to approximately 60°C during collection, so that
S02~reagent complexes are immediately decomposed and removed.
Summary (Unmodified Thomas Generator, 0.005 to 0.3 ym)
Studies using the Thomas aerosol generator showed that highly volatile
amines and amine-derivative reagents could be used to fix sulfuric acid
aerosol during the collection process. Furthermore, it was found that the
fixing reaction proceeded at a sufficiently rapid rate for quantitative
fixing and was complete during reactor residence time. A correlation was
established between the sulfur gases evolved during the adduct decomposition
and the amount of sulfuric acid aerosol stream sampled. In addition, it was
found that the amount of sulfur gases evolved during adduct thermal decompos-
ition of identical samples was reproducible. The ammonium sulfate salts,
which are thought to be the most problematic interfering species in the
analysis of ambient H~SO, were found to have decomposition temperatures
that were measurably distinct from the sulfuric acid adducts. fhas, it
was demonstrated that H2SO, adducts made from aerosols in the 0.005 to 0.3 ym
size range, could be differentiated from ammonium sulfate salts.
DECOMPOSITION CHARACTERISTICS OF ADDUCTS (Baird Generator,
1 to 3 ym Droplets
The Baird aerosol generator, described on page 43 of this report, was
originally selected for this study because of a need for aerosol particles
in the 1 to 3 ym size range. It was, however, also anticipated that the
generator, because of the closed system design, would give highly reproducible
aerosols. The FPD analysis apparatus was used to analyze fixed aerosol
samples from the Baird Generator by procedures identical to those used with
the Thomas generator.
Adduct Decomposition (Baird Generator)
Initially, the adducts formed using the Baird generator were decomposed
using the rising temperature procedure in order to verify decomposition
74
-------
temperatures. The results were surprising in that the three adducts pre-
viously investigated using the Thomas generator now had average decomposition
temperatures that were different. The results are given in Table 10, along
with a comparison of results obtained from the Thomas generator and from
depositing H2S04 on Mitex filters using the micropipet technique discussed
on page 45.
TABLE 10. INITIAL H2S04 DECOMPOSITION TEMPERATURE
BY H2S04 GENERATION METHOD.
24 Thomas Baird
Adduct Generator °C Generator "C Pipet °C
AAO 140 185 195
DEA 190 200 200
DEHA 190 170 185
As can be seen from the table, the average decomposition temperature
of the AAO and DEA adducts shifted upward while the third (DEHA) shifted
downward. Of the three adducts, only that of AAO showed a substantial change
compared to the previously determined individual decomposition temperature
reproducibility of +_ 10%. Experiments were also performed, as shown in
Table 11, using a pipet to deposit the various reagents directly onto Mitex
filters. It was found that the AAO adduct decomposed at 195°C, the DEA
at 200°C, and the DEHA at 185°C. Again, only the AAO adduct showed a signi-
ficant difference in decomposition temperature as compared to those formed
using the Thomas generator.
The results of the experiments described above made it necessary to
re_evaluate some of the previous work with amine-derivative reagents. A
comparison between samples applied with a pipet and samples applied with the
aerosol generators was again made, using the three candidate reagents:
acetaldoxime (AAO), diethylamine (DEA) and the new reagent, n-methylformamide
(NMF) . The average initial decomposition temperatures of the AAO and NMF
adducts were found to be affected by the. mode of aerosol deposition as shown
in Table 11.
75
-------
TABLE 11. INITIAL H2S04 ADDUCT DECOMPOSITION
TEMPERATURES BY H-SO, GENERATION METHOD
2 4
Sample
Generation AAO °C DEA °C NMF °C
Pipet 195 200 193
Baird Aerosol 180 200 185
Thomas Aerosol 140 190 120
Examination of the above data indicated several possible reasons for
the observed decomposition temperature fluctuations. One explanation was
that the various aerosol generation methods which were selected to give
droplets of different size ranges also resulted in particles with a large
difference in free surface energy (due to significant differences in particle
size). The difference in free surface energy between particles of greatly
varying size resulted in a corresponding change in the decomposition tempera-
ture. However, the data suggest that the mode of aerosol generation was
a significant factor only when the fixing reagent contained unsaturated bonds.
For example the DEA and DEHA fixing reagents,
C2Hr
[-OH
DEA DEHA
have no unsaturated bonds and the corresponding H_SO adducts gave average
decomposition temperatures of 195°C and 180°C, respectively, regardless of
the H^SO, generation method. The NMF and AAO fixing reagents, however,
76
-------
q H
il
C = N-OH
NMF AAO
which have unsaturated bonds, consistently resulted in decomposition tempera-
tures which reflected the method of aerosol generation. The decomposition
temperature for the NMF and the AAO adducts of H2SO, using the Baird aerosol
generator averaged 185°C and 180°C, respectively, while use of the Thomas
aerosol generator results in the NMF adduct having an average decomposition
temperature of 120°C and the AAO adduct having an average decomposition
temperature of 140°C. Thus, it appeared that the chemical nature of the fixing
reagent, in conjunction with the aerosol generation method, affects the
decomposition temperature of specific H2SO, adducts. This may be due in part
to the fact that the collection temperature of the Thomas generator was
approximately 60°C, while the Baird generator aerosol was collected at
approximately 30°C. The higher temperature of the Thomas generator was
due to the use of a flame during aerosol formation.
Ammonium Sulfate Interferences
The questions associated with the decomposition of the various adducts
led to a re-investigation of the major anticipated interferent, (NH,)«SO,.
Several sets of experiments were performed to accumulate data which might
point to the reason for the decomposition temperature deviations. A series
of 0.01N (NH.)-SO, samples from the same solution were deposited in different
volumetric amounts by micropipet on Mitex filter.s. These tests were performed
to determine if different quantitative amounts of aerosol could cause the
temperature fluctuations. Results of these experiments are given in Table
12. As can be seen from the table, the initial FPD response from the
decomposition of identical (NH,)2SO, samples occurred over a range of 170°C
to 195°C. Since the samples were prepared and analyzed under identical
conditions, it would appear that the inconsistency is not related to quanti-
tative differences in the amount of (NH,)2SO^ deposited.
77
-------
TABLE 12. SUMMARY OF (NH^SC^ DECOMPOSITION TEMPERATURES
FOR VARIOUS SAMPLE SIZES.
Sample Initial FPD
Size (ul) Response (°C)
5 190
5 190
7 190
7 195
7 175
10 170
10 185
10 190
15 175
15 170
15 180
The Baird generator was used to deposit (NH ) SO, aerosols made from
0.01N and 0.001N (NH,)9SO, solutions. It was assumed that since less
(NH,)?SO, per droplet would be present, this would result in a smaller part-
icle when dry. The aerosol was prepared by simply dissolving (NH,)9SO,
into deionized water to give a 0.01N and a 0.001N solution and aspirating
these solutions through the Baird generator instead of sulfuric acid solu-
tions. The 0.01N solution was also deposited on a filter using a pipet and
subjected to the same temperature programming sequence. In addition,
(NH,)9SO, was formed using H?SO aerosol obtained from the Thomas generator.
The (NH ) SO in this case was formed by passing anhydrous NH_ through the
fixing chamber as the H^SO, aerosol was generated, or alternatively, by
first collecting the H~SO, aerosol, waiting ten minutes, and then passing
NH_ through the filter. The results of these experiments are given in
Table 13. It should be pointed out that the results given in the table are
average values obtained from a minimum of 5 individual runs for a specific
experiment.
78
-------
TABLE 13. INITIAL RESPONSE OF FPD FOR (NH,) SO
BY DEPOSITION METHOD. 424
Average Initial FPD
Response (°C) Deposition Method
240 (NH4)2S04 from reagent bottle
183 Pipet, 0.01N solution
175 Aerosol, Baird, 0.01N solution
165 Aerosol, Baird, 0.001N solution
1 t:o&
Aerosol, Thomas, 0.01N solution
o 9 n&&
Aerosol, Thomas, 0.01N solution
*(NH4)2S04 prepared by fixing H2S04 aerosol with NHL
during sample collection.
**(NH4)2S04 prepared by fixing H2S04 aerosol with NH
after sample collection.
As can be seen from Table 13, (NH )?SO, taken from a reagent bottle and
ground to a fine powder gave an average initial decomposition temperature
of 240°C.
When the Baird aerosol generator was used to deposit samples from a
solution of 0.01N (NH4)«SO,, an average value of 175°C was found to be the
initial temperature at which the aerosol decomposes. However, when the
identical solution was deposited by pipet and subjected to the same analysis
procedures, an average decomposition temperature of 183°C was obtained.
When the Baird generator was used to deposit the 0.001N (NH,) SO, solution,
the initial response temperature dropped to an average of 165°C. Use of the
Thomas generator and exposure of the generated H^SO,^ aerosol to NH« during
sample collection resulted in a further drop of the (ML)-SO initial
decomposition temperature to an average of 158°C. When the H SO aerosol is
first collected, however, and then exposed to NH,, the initial (NH4)_SO,
decomposition temperature rises to an average of 200°C. It was thought
that the lack of agreement in results obtained using the Thomas generator,
with alternate (NH4)2SO formation procedures, was due to H2S04 aerosol
droplet growth on the collection surface prior to the addition of NH,.
79
-------
An examination of the various H2SO, generation methods suggested that
each would result in particles of different size ranges being deposited.
Thus, it appeared that a particle size effect might be inducing a corres-
ponding change in decomposition temperature of the (NH,)2SO,. This possi-
bility was discussed with two independent scientists involved in sulfuric
acid aerosol research, Dr. Herbert Miller of Southern Research Institute (65)
and Dr. Willard Richards of: Rockwell International (66).- The consensus
was that the particle size deviations anticipated from the various generators
was most probably the reason for the decomposition differences. It is
interesting to note that Dr. Miller, while using the Thomas generator in
an analogous manner to prepare (NH,)2SO aerosol (collecting H2SO aerosol
on the filter, then passing NH- through it) also found no decomposition at
220°C. Dr. Richards, on the other hand, routinely used the Baird generator
to make aerosols and found that (NH.KSO, begins to decompose at approximately
160°C.
Several experiments were also performed to see if altering the flow
or temperature programming rate of the volatilization/decomposition chamber
might induce a change in the initial decomposition temperature of (NH,)_SO,.
It was thought that the difficulty in reproducing initial decomposition
temperatures for (NH,)»SO, might be due to subtle changes in these conditions.
Moderate changes, however, in the rate of heating (+ 10°C/min) or carrier
gas flow (+ 25 cc/min) did not noticeably change the initial response temper-
atures of (NH.)_SO,.
4 2 4
Thus, it was concluded that the particle size of aerosols does influence
the decomposition temperatures of the various adducts to a large degree.
This is particularly significant since the distribution of sulfuric acid
aerosols in the environment is thought io cover a large size range which would
indicate that the corresponding adduct decomposition would also vary widely.
Aerosol Stoichiometry (Baird Generator)
Experiments were performed to obtain information regarding the
reproducibility of the Bair.d generator and the quantities of SO evolved
from the AAO adducts. Four trials were made which were identical with
respect to aerosol generating parameters and the volume of H SO solution
2 4
80
-------
passed through the Baird nebulizer. The filters were heated at 200°C for
30 minutes and the evolved gases determined by the West-Gaeke method. The
results of these expeirments are given in Table 14.
TABLE 14. RECOVERY OF S02 FROM AAO/H2S04 ADDUCT (DRY AIR)
Run
No.
Aspiration
Volume (ml)
25
25
25
25
Volume
Consumed (ml)
2
2
2
2
8.0
2 25 2 7.5
3 25 2 7.0
4 25 '2 8.5
As can be seen from Table 14, the amount of H SO, collected during
each run was fairly reproducible based upon the amount of SO,., evolved
during the thermal decomposition process. Thus, it appears that the Baird
generator is fairly reproducible, as is the evolution of S0« from an identical
quantity of the AAO adduct. However, based upon an aspiration volume con-
sumption of 2 ml (0.001N H SO, solution), 98 yg of H-SO should theoretically
have been collected. Instead, an average of 7.7 yg were collected, result-
ing in an apparent collection efficiency of approximately 7.9%.
The fact that 75 ml of acid solution passed through the Baird
nebulizer in the first experiment and 25 ml in the second experiment both
resulted in 2 ml of solution being consumed could not be explained. It was
thought that the large amount of solution lost during each run might be due
to volatilization of water, since dry compressed air was used to operate the
Baird nebulizer. Therefore, the above experiment was repeated using humidi-
fied air. The air was humidified by bubbling through a column of water
which was 2.5 cm x 30.5 cm. The results of this experiment are given in
Table 15-
As can be seen from Table 15, using humidified air decreased the
volume of acid solution consumed to 1.8 ml, while the amount of adduct
collected remained proportional to the volume of acid passed through the
nebulizer. In fact, the 13.8 yg. of acid collected when 50 ml of acid was
81
-------
TABLE 15. RECOVERY OF S02 FROM AAO/IUSO, ADDUCT
(HUMIDIFIED AIR)
Run Aspiration Volume
No. Volume (ml).. Consumed (ml) ygi H?SO,
1 50 1.8 14.5
2 50 1.8 12.6
3 50 1.8 14.5
4 50 1.8 13.5
passed through the nebulizer is on the order of twice the 7.8 yg. collected
when 25 ml of acid was used. In addition, the 75 ml passed through the
nebulizer in the first experiment resulted in 21 yg. being recovered, which
is approximately three times the 7.8 yg,u collected when 25 ml were passed
through the nebulizer.
The aspiration rate of acid solution through the nebulizer of the Baird
generator was maintained at 1.7 ml/min throughout the above experiments.
The delivery of 1.7 ml/min required a continuous air flow of 34 1/min across
the nebulizer orifice. Therefore, 25 mis of acid solution aspirated through
the nebulizer used 500 1 of air, while 50 ml of acid solution used 1000 1
of air and 75 ml of acid solution used 1500 1 of air. Since the entire
aerosol stream from the Baird generator was sampled in the above experiments,
it follows,as shown in Figure 28, that AAO can be used to fix I^SO, aerosol
in a moving gas stream to form an adduct which when decomposed gives off
an amount of SO that is related to the volume of aerosol stream sampled.
However, the question of exact stoichiometric relationship between the
evolved S02 and the original H2SO, aerosol stream can only be determined
by the decomposition of a precisely known amount of H-SO, adduct.
Summary (Baird Generator, 1 to 3 ym)
It was found that the decomposition temperature of the various
H2SO^ adducts was, to some extent influenced by the particle size of the
H-SO^ droplets from which the adduct was formed. This questions of the
significance of environmental particle size on actual adduct decomposition
82
-------
24
§
E 20
0
0 12
I
§
8"
X 4
i
a.
\7
7
0.4 0.8 1.2 1.6
VOLUME OF AEROSOL STREAM SAMPLED (M3)
2.0
2.4
Figure 28. Relationship of aerosol sample volume and evolved SO-
from AAO adduct decomposition.
characteristics should be investigated using an aerosol source which
more closely duplicates actual environmental conditions. It appears that
the larger particles in a real sampling situation would most probably have
reacted with NH3 to form (NH^SO^, while the smaller particles would tend
toward H_SO, aerosol. This, of course, would ultimately depend upon
prevailing environmental conditions, i.e., humidity, ambient ammonia content,
etc. It was felt that the sampling of any pure aerosol generator stream
tends to subject the analytical methodology to equilibrium characteristics
which are not necessarily representative of actual environmental samples.
Attempts to correlate pH and sulfate level differences between the
original acid and that returned to the trap were unsuccesful due to the
concentration of the acid in the nebulizer trap being too great to accurately
measure the small amount of acid lost during aerosol generation. It was,
therefore, necessary to find an alternative method for depositing known
amounts of H^SO,.
83
-------
CHARACTERISTICS OF ADDUCT DECOMPOSITION GASES
The success of the proposed research approach ultimately depends upon
whether a reproducible stoichiometric relationship exists between the
sulfur gases evolved during adduct decomposition and the H^SO, originally
present in the sample stream. It was originally hypothesized that the gases
evolved during adduct decomposition should consist primarily of SO™. This
was a desirable characteristic since SO™ is relatively easy to quantitate
and is a fairly stable species. It would serve little,purpose to generate
a highly reactive species, such as SO.,, which would subsequently partially
react and be lost prior to quantitation. Thus, for the above reasons, SO™
was selected as the initial gaseous species by which the stoichiometry of
the proposed method would be evaluated. The West-Gaeke technique, because
of its reported selectivity and sensitivity for SO™, was chosen as the
detection method for these tests. The FPD is not specific for SO™ and would
offer little toward defining the exact stoichiometric relationship of
evolved SO™, should other sulfur gases be present.
Specificity of West-Gaeke Technique for SO™
The specificity of the West-Gaeke technique for SO™ was examined by
preparing various samples using the Baird aerosol generator and collection
techniques previously described on page 5Q. The samples examined
consisted of the following:
1. Clean, unused filter (all filters Mitex LS, 5 ym).
2. Filter through which 2 ml of deionized water and 515.4
liters of clean air had been sampled. (2 ml based on 75 ml
passed through Baird atomizer and 73 collected in trap.)
3. Same as (2) above, but "fixed" with acetaldoxime vapor.
4. Filter through which 2 ml of 0.001N H™SO, and 515.4 1 of
clean air had been sampled. (2 ml based on 75 ml passed
through Baird atomizer and 73 collected in trap.)
5. Same as (4) above, but fixed with AAO vapor.
84
-------
Each of the above samples was decomposed separately in the decomposition
chamber for a period of one-half hour at a temperature of 200°C. The gaseous
products were continuously bubbled through 10 ml of sodium tetrachloromercur-
ate and subjected to the West-Gaeke analysis for SCL. The results of these
analyses are shown in Table 16.
TABLE 16. SPECIFICITY OF WEST-GAEKE TECHNIQUE FOR S0«.
No. Sample % Absorbance yg.. SO^ ygm H2SO,
1 Clean filter 100
2 Water + air 100
3 Water + air + AAO vapor 100
4 0.001N H2S04 + air 1 00
5 0.001N H SO + air + AAO vapor 28 14 21
6 TCM blank 000
As can be seen from Table 16, the results indicate that the West-Gaeke
technique is sensitive only to S02 and does not detect H2SO, or the AAO-
fixing reagent.
Pipet Adduct Characteristics
AAO Adducts —
The results of the experiments previously described using the Baird
sulfuric acid generator indicated that it was probable that the apparent
low decomposition yields with the AAO adduct were the results of complex
decomposition products or the inability to quantitate the aerosol generator,
rather than the actual collection of fixation efficiency. This was substan-
tiated by the results obtained using the Thomas generator and the FPD method
of detection which clearly showed no residual acid remained on the collection
filter when similarly fixed. To further test the above hypothesis, however,
and determine the actual stoichiometric S02 evolution relationship of several
adducts based upon the amount of acid deposited,,a set of experiments was
devised in which a known amount of acid was applied directly to the filter
media by micropipet.
85
-------
Several Mitex Teflon filters were loaded with 600 yg of H-SO, by
pipet and the amount of ^SO determined turbidimetrically. The purpose
of these experiments was to determine if 600 yg of acid could be repro-
ducibly deposited on the filters and analyzed turbidmetrically- Several
types of filters were used to obtain background information of the differ-
ent media. As can be seen from Table 17, all filters except the glass
fiber filter demonstrated 100% recovery. It is interesting to note that
the high SO^ reading on the glass fiber medium corresponds to Dubois' (37)
findings that glass fibers contain residual sulfates.
TABLE 17. TURBIDIMETRIC"ANALYSIS OF RAW H0SO,
L jf
Run No. Filter ygm H^SO^ Applied ygm. H2SO^ Measured
1 Whatman #40 600 600
2 Blank Whatman #40 0 0
3 Glass Fibre 600 750
4 Fluoropore 600 600
5 Mitex 600 600
Once it was established that 600 ygm could be deposited and analyzed
reproducibly, several filters were loaded with 600 ygm of H_SO, and
fixed with 25 pi of AAO. The samples were dried at 60°C for one hour, and
the sulfate content determined turbidimetrically. These tests were designed
to show that sulfur is not lost during the fixation process. As can be seen
from Table 18, the results indicate that reproducibility of the fixation/
drying process was on the order of + 10%.
TABLE 18. SULFURIC ACID/AAO ADDUCT ANALYSIS
(LS FILTER - TURBIDIMETRIC ANALYSIS)
Run yg- H2SO
No. Applied
1 600
2 600
3 600
yl AAO
Applied
25
25
25
ygm H2S04
measured
650
552
552
86
-------
Four filters were then loaded with 600 ygm of H-SO, and 25 yl of AAO
and allowed to dry at 60°C for one hour. The adduct was decomposed at
200°C for 30 minutes. The evolved gases were determined by the West-Gaeke
method and by bubbling the evolved gases into 10 ml of 0.6N H20- which was
analyzed for total sulfate turbidimetrically. The H202 solution was used
to oxidize S02 in the effluent to SO," (67). As can be seen from Table 19,
it appears that additional sulfur species are generated during the decomposi-
tion process with approximately 30% of the sulfur present in the sample
being accounted for as S0~.
TABLE 19. DECOMPOSITION EFFICIENCY OF AAO ADDUCT
West-Gaeke M0
West-Gaeke AAO 600
H2°2 AAO 600
H2°2 AAO 600 296
Subsequent turbidimetric analysis of the filter material from the above
samples for sulfur indicated that all sulfur species were totally removed
during the decomposition process. Analysis of the connecting tubing between
the decomposition chamber and the collection solutions showed the presence
is
of SO,. These analyses, however, accounted for only an additional 30% of
the sulfur anticipated from the decomposition of the adduct. The odor of
ELS gas was noted in several cases during adduct decomposition, but no
effort was made to quantitate for other sulfur species. Thus, even though
the West-Gaeke analyses from both the micropipeted samples and those prepared
using the Baird aerosol generator have consistently showed a reproducible
relationship between sample size and S0_ evolution: .from the decomposed AAO
adducts, the fact that several sulfur species are present makes consistent
accurate quantitation of the adduct by this method subject to possible
perturbations. However, the stoichiometric relationship between the SO
evolved during adduct decomposition and the original acid aerosol was
reproducibly one mole of SO- to three moles of H2SO, when both sampling and
analysis were accomplished under controlled laboratory conditions. It
87
-------
would be more desirable to have an H«SO, adduct that decomposes to give a
100% stoichiometric return of SCL based upon the amount of acid originally
present in the sample.
DEA and DEHA Adducts —
It appeared that the reason for the failure of the adduct of AAO to
return 100% of the sulfur decompositon produce as S0? is due to the formation
of several species during the decomposition process. Because the presence
of several sulfur containing species in the adduct decomposition gas stream
makes it more subject to quantitation errors, it was decided to test several
other candidate H~SO, adducts for less complicated decomposition mechanisms.
It was felt, based upon structural differences, that diethylamine was the
best candidate for investigation.
The West-Gaeke technique was used to determine the stoichiometric
relationship of the SO evolved during the decomposition of the DEA adduct.
All samples were prepared by depositing 600 ug of EUSO, onto Mitex filters
by micropipet. Immediately after the H SO, was deposited, an excess of
DEA was added (25 yl) and the adduct allowed to dry at 100°C. When dry,
the filters were placed in the sample chamber and then heated to % 200°C
for one hour, while the evolved gases were purged by helium into the TCM
solution (West-Gaeke). Aliquots of the TCM were then taken and developed
in the usual manner for comparison with SO,, standards. The results of these
tests are given in Table 20.
TABLE 20. DECOMPOSITION OF DEA ADDUCT (SO ANALYSIS)
Ug H2S04
Run No. deposited
1 600
2 600
3 600
4 600
5 600
6 600
MS- S09
341
387
398
350
406
429
ygm Equivalent
H?SOA Found
522
592
610
540
621
656
Average 590
88
-------
Thus, it appears that virtually all of the H2SO, deposited on the filters
can be accounted for, based upon the S02 evolved during :DEA adduct decomposi-
tion. This is a desirable property since it simplifies analytical methodology.
Several tests were also made with diethylhydroxylamine. The adduct
was heated for one hour and yielded an average of 318 yg of H»SO . It
appears at this point that diethylamine offers the most promising recovery of
adduct decomposition products with the least problems.
In summary, the DEA adduct of sulfuric acid returns the best stoichio-
metric ratio of S02 when thermally decomposed of the three adducts examined.
Adduct Stability
As originally conceived, a desirable aspect of the proposed methodology
was the prediction that the H.,SO, adducts would be less likely to react
during the collection process than the original acid aerosol. It was also
predicted that drying of the adducts would have a tendency to suppress
side reactions on collection surfaces. The lower reactivity of the adducts
should, in effect, minimize the significance of negative interferences due
to the loss of EUSO,. Thus, any feasibility assessment of the proposed
methodology must necessarily incorporate an investigation to determine if
the adducts are less subject than EUSO, aerosol to undergoing reaction during
the collection process. Therefore, experiments were performed to compare
the reactivity of various adducts (both wet and dry) with NH., and the
reactivity of H^SO, aerosol with NH.,. Ammonia was selected as the test
species because of its reactivity with H9SO and because it is easily
£• H
exposed in a homogeneous manner to both the adducts and the acid aerosol
on the collection surface.
The reactivity of sulfuric acid aerosol with ammonia was determined
by passing a gaseous mixture of 250 ppm ammonia through the reaction chamber
of the sample collection device concurrently with H SO aerosol from the
Baird generator. The amount of NH- reacted with the sulfuric acid aerosol
was then measured with an ammonia specific ion electrode. Basically, this
procedure involved removing each sample from the collection device and
placing it in 100 ml of deionized water. One ml of 10 molar NaOH was then
added to the solution to generate NH_. The ammonia concentration as measured
89
-------
by the specific ion electrode, served as an indication of the amount of
ammonia which had reacted with the H«SO, aerosol.
Blank samples were prepared by exposing various fixing reagents to
HUSO, aerosol from the Baird generator in a manner analogous to ammonia
above. The blank samples were then measured with the ammonia specific ion
electrode. These tests served to show that the adducts themselves gave
very little response on the NH,, electrode. Identical .adducts were then v.>.
prepared using the Baird generator and subsequently exposed to a mixture of
250 ppm ammonia for exactly the same length of time as the original H^SO,
aerosol.
In the case of experiments studying the reactivity of wet aerosols,
the adducts were exposed immediately after being formed to a dynamic ammonia
environment of 250 ppm. In the case of experiments studying dried adducts,
the freshly formed adducts were removed from the sampler, placed in a closed
Petri dish and dried for approximately one day at 85°C. The dried samples
were then returned to the sampler and exposed to a dynamic ammonia environ-
ment of 250 ppm for the same period of time as the original undried samples.
Analysis of both wet and dry samples involved removing the filter from
the collection device and placing them in 100 ml of deionized water. One
ml of 10 molar sodium hydroxide was then added to the solution to generate
NH., which was detected by the ammonia specific ion electrode. The ammonia
concentration, reactor flow rate, aerosol generation rate, and ammonia
exposure time were identical in all experiments.
Ammonium sulfate was formed by passing a gaseous mixture of 250 ppm
ammonia through the reactor during H-SO, aerosol generation. The ammonia
concentration, as measured by the specific ion electrode, served as an
indication of the amount of ammonia that would react with the H9SO,. The
difference between the formation of ammonium sulfate due to reaction of
ammonia with H«SO, and with the various wet and dry adducts was considered
to be demonstrative of the increased stability of each of the adducts. It
should be remembered, however, that the experiments used concentrations
of ammonia that were large compared to environmental conditions. Nominal
environmental concentrations would have resulted in the reactivities being
much smaller in both cases due to collision probabilities. The results
90
-------
of the above series of tests are given in Table 21. As can be seen from the
table, there appears to be little significance in drying the samples.
TABLE 21. PERCENT LOSS OF VARIOUS ADDUCTS DUE TO
REACTION WITH AMMONIA
AAO Pyridine DMF 1WF
Wet 69.6 73.1 61.0 97.1
Dried 60.8 77.0 81.1 70.6
MODIFIED THOMAS GENERATOR
Difficulties were continually encountered in obtaining reproducible
H2SO^ samples at the low yg level using the Thomas flame generator. The
Baird generator, while somewhat more reproducible,also had difficulties
in that the apparent amount of acid generated could not be accounted for on
filter samples. Thus, there was a definite need for an aerosol generator
which could be used to obtain a realistic mass balance comparison between
predicted aerosol concentration and measured aerosol concentration. In
addition, it was felt, based upon previous data, that the new generator
should possess the capability of alternately generating aerosols in the size
range of both the Thomas and Baird generators. The final generator design
consisted of the basic Thomas flame atomizer with the following modifications:
1) Provision was made so that either fuel or compressed air could
be passed through the burner orifice, thereby allowing for genera-
tion of aerosols in the large or small size range at identical
operating pressures.
2) Since the burner tip orifices are extremely critical for maintain-
ing a consistent H~SO, aspiration rate, a Radiometer AutoBurrette,
Model ABU-1, was connected to the aspirator and used to feed the
H»SO, solution to the orifices in a uniform manner. The delivery
rate of the solution into the burner tip was nominally one-half
milliter per minute.
The modified Thomas generator in either the flame or flameless mode: was
used in all subsequent testing.
91
-------
HpSO, Volatilization Approach (Modified Thomas Generator)
Since it was found that adduct particle size apparently influences the
decomposition temperature of the corresponding adduct, it was thought that
a possible solution to the H2SO, measurement problem might be to volatilize
the H-SO, aerosol. H^SO, is known to volatilize at a temperature substan-
tially lower than the potentially interfering sulfates. The volatilized
acid could then be passed through a prefilter which would at the same time
collect the unvolatilized sulfur-containing particulates. The volatilized
acid could then be reacted on the downstream side of the prefilter and
collected on a secondary filter in the absence of interfering particulates.
In fact, since the ammonium sulfate salts would be collected on the prefilter,
ammonia could be used as the fixing reagent, thereby eliminating possible
equilibrium problems between ambient ammonia and the fixing reagents.
To aid in investigating the volatilization approach the aerosol stream
from the modified Thomas generator was sampled by two parallel all-Teflon
sampling legs. One sampler leg consists of a presection which can be heated,
a prefilter, a reaction tube and a final collection filter; the other con-
sists of a presection with no provision for heating, a reactor tube and a
final collection filter. Dimensionally, both sample legs are identical.
The heated leg of the sample was described on pages 46 and 47. A diagram was
presented as Figure 11.
NH., Fixing Reagent (Modified Thomas Generator - no flame; both sample legs
at ambient temperature)
Since research emphasis shifted to the volatilization of H~SO, and
subsequent passage through a prefilter, it was decided that NH., should be
investigated as a fixing reagent. Theoretically, (NH,),?SO, and other
particulates should not pose a problem since they would be stripped from
the aerosol stream by the prefilter. The results of these experiments,
shown in Table 22, indicate that 54% of the concurrently collected samples
had a Sd of less than 0.05 for the individual samples pairs while 23%
of the samples had a S, of less than 0.20 for those sample pairs and
15% of the samples had an S, of less than 0.31 for each sample pair.
All samples were decomposed at 200°C for 15 minutes. Quantitation
of these experiments was performed by using the FPD and comparing a known
92
-------
S02 calibration standard with the SO- evolved during adduct decomposition.
As can be seen from Table 22, the amount of H-SO, found in individual sample
pairs was not always in close agreement.
TABLE 22. ANALYSIS OF CONCURRENT SAMPLES
USING AMMONIA FIXING REAGENT.
Both Legs of Collection Device at Ambient Temperature
Experiment yg yg —
No. Filter No. 1 Filter No. 2 ! d
1 0.48 0.16 0.32 0.23
2 0.58 0.58 0.58 0.00
3 0.40 0.17 0.29 0.16
4 0.47 0.71 0.59 0.17
5 1.0 1.0 1.0 0.00
6 1.0 1.0 1.0 0.00
7 0.78 0.78 0.78 0.00
8 0.48 0.53 0.51 0.04
9 0.91 0.91 0.91 0.00
10 0.56 1.2 0.88 0.45
11 0.70 0.53 0.62 0.12
12 0.69 0.75 0.72 0.04
13 1.2 0.78 0.99 0.30
PEA Fixing Reagent (Modified Thomas Generator - no flame; both sample legs
at ambient temperature)
Eight initial experiments were then conducted using diethylamine fixing
reagent. The results of these experiments are given in Table 23.
TABLE 23. ANALYSIS OF CONCURRENT SAMPLES
USING DIETHYLAMINE FIXING REAGENT
Both Legs of Collection Device at Ambient Temperature
Experiment yg yg — „
No. Filter No. 1 Filter No. 2 * d
1 0.52 0.54 0.53 0.01
2 0.23 0.25 0.24 0.01
3 0.36 0.37 0.365 0.01
4 0.34 0.34 0.34 0.00
5 0.30 0.12 0.21 0.13
6 0.44 0.44 0.44 0.00
7 0.29 0.49 0.39 0.14
8 0.50 0.55 0.525 0.04
93
-------
As can be seen from the table 75% of the values were found to be within
+ 5% of the mean value for each sample pair and 25% were within 50% of the
mean value for each sample pair. Again, the occasional erratic results were
thought to be due to aerosol memory effects in both the generator and sample
collection device. In addition, in the case of the DEA adduct, it appeared
that heating at 250°C for 15 minutes was not sufficient for complete adduct
decomposition.
In order to further examine the perturbations described above, the pre-
vious experiments were repeated using ammonia and diethylamine fixing reagents.
During these experiments effort was made to insure that the generator had
stabilized by collecting a minimum six dummy runs prior to each set of samples.
In addition, the decomposition temperature was raised to 300°C for both the
ammonia and diethylamine adducts in order to speed up the decomposition pro-
cess. The results of these experiments are discussed below.
Diethylamine Derivatives (Modified Thomas Generator - no flame; both sample
legs at ambient temperature)
It was found that sporadic perturbations of as much as 50 percent occurred
between the two legs of the sample collection device even when precautions
were taken to insure that the sampling interval was constant and a decomposi-
tion temperature of 300°C was used. Further examination of the problem re-
vealed that on occasion the decomposition peaks, as measured by the FPD,
tended to tail excessively. It was noted that the tailing phenomenon always
worsened toward the latter experiments of the day. Experiments performed
using only the decomposition chamber (no sample) during the latter part of
the day revealed that considerable sulfur bearing species were still present
in the decomposition chamber. In fact, simple insertion of the sample plunger
into the decomposition chamber would, on occasion, generate a large response
on the FPD as would simply scratching the internal surface of the decomposi-
tion chamber. However, after continuous heating of the decomposition chamber
at 300°C overnight (it normally remains heated) all traces of residual sulfur
bearing species were absent. Results of these tests indicated that the di-
ethylamine derivative was not totally decomposed at 300°C during decomposition
periods of approximately five minutes. Thus, at least in part, it appeared
that the previous erratic results are due to the diethylamine adduct or
94
-------
intermediate decomposition product collecting on the decomposition chamber
walls and sample insertion plunger. The collected material then decomposes
sporadically during subsequent analyses. Due to these decomposition inson-
sistencies of the DEA derivatives emphasis was then focused on NH« as the
fixing reagent.
Ammonia Derivatives (Modified Thomas Generator - no flame; both sample legs
at ambient temperature)
The previous tests using ammonia as the fixing reagent and a decomposi-
tion temperature of 200°C also gave sporadic differences of as much as 50
percent between the individual sample collection legs. It was found that
raising the decomposition temperature of 300°C resulted in the difference
between the individual legs and their mean value on eleven separate runs
averaging approximately 7 percent with only one of the runs having a differ-
ence of more than 10 percent (13 percent).
The average amount of H_SO, found using ammonia fixing reagent with sub-
sequent thermal decomposition at 300°C and FPD analysis showed that sample
leg 1 contained an average of 1.46 yg of H_SO, while sample leg 2 contained
an average of 1.38 yg of H^SO,. The individual legs were found to agree
within 2.8 percent of the mean value for eleven separate runs.
The results of these experiments are given in Table 24 and indicate that
duplicate H-SO, aerosol samples can be reproducibly fixed and analyzed at
the low yg/m3 level, using NKL fixing reagent and thermal decomposition tech-
niques. It is apparent from reviewing the data in Table 24 that the genera-
tion/collection system had not equilibrated since successive runs clearly
show that decreasing amounts of H-SO, were being deposited on the collection
filters of both sample legs. This is attributed to the fact that the aerosol
generator stack was cleaned prior to the initiation of this series of experi-
ments .
95
-------
Leg #1
1.70
1.53
1.53
0.92
1.33
0.66
0.61
0.55
1
Leg #2
1.42
1.18
1.42
1.02
1.18
0.81
0.61
0.61
1.56
1.36
1.48
0.97
1.26
0.74
0.61
0.58
9.0
13.2
4.1
5.2
6.4
9.5
0
5.2
TABLE 24. CONCURRENT AND SEQUENTIAL NH3
DERIVATIVE SAMPLES
Modified Thomas Generator - No Flame; Both Legs of
Collection Device at Ambient Temperature
0.01 N H2S04 Solution Values in yg. - % Difference
Volume Aspirated H^SO, (based on SO,, evolved) £ from Mean Value
1.2 ml
1.2 ml
1.2 ml
1.2 ml
1.2 ml
1.2 ml
1.2 ml
1.2 ml
X 1.10 X 1.03 X 1.06 2.8
2.5 ml
2.5 ml
2.5 ml
X 2.40 X 2.33 X 2.36 1.3
ANALYSIS OF CONCURRENT SAMPLES USING THE HEATED SAMPLING DEVICE
WITH NH3 FIXING REAGENT
The logical step after determining that duplicate H-SO, aerosol samples
at the low yg level could be collected and analyzed using the dual sample
collection device was to determine what effect heating one leg of the col-
lection device might have on reproducibility. This was necessary in order to
accurately assess the use of the heated prefilter technique for the eliminat-
ion of (NH,)-SO, interference. Therefore, a series of experiments was per-
formed in which number one leg of the dual collection device was heated to
approximately 105°C while number two leg remained at ambient temperature.
It was felt that 104°C was sufficiently hot to volatilize H»SO, aerosol but
not decompose (NH.KSO,. In these experiments no filter was inserted in the
prefilter holder since the purpose of the tests was to determine if the acid
96
2.48
2.13
2.58
2.15
2.52
2.31
2.32
2.32
2.45
7.3
8.6
5.7
-------
could be heated to volatilization temperature and be quantitatively collected
and analyzed. The results are given in Table 25. The average difference
between the heated and unheated leg was approximately 7 percent with only two
sets of duplicate samples varying by greater than 10 percent (24.6 percent
and 13.1 percent). The standard deviation of sequential runs during these
tests was approximately 0.33 indicating that the aerosol generator was equil-
ibrated from run to run.
TABLE 25. CONCURRENT AND SEQUENTIAL NH3 DERIVATIVE
SAMPLES WITH NUMBER 1 LEG HEATED (105°C)
Modified Thomas Generator - No Flame
H2S04 Solution
Volume Aspirated
Values in yg.
% Difference
from Mean Value
2.5 ml
2.5 ml
2.5 ml
2.5 ml
2.5 ml
2.5 ml
2.5 ml
2.5 ml
lard Deviation
-£ — ft
Leg #1
(Heated)
1.53
2.30
1.61
1.61
1.65
1.81
1.32
1.07
X 1.61
0.36
£.
Leg #2
0.92
2.14
1.38
1.68
1.61
1.53
1.32
1.38
X 1.50
0.35
1.22
2.22
1.49
1.65
1.63
1.67
1.32
1.22
X 1.55
0.33
24.60
3.60
7.40
2.10
1.20
8.40
0.00
13.10
3.2
It was noted during these experiments that the initial heatup cycle of
the heated leg (number 1 side) resulted in a much larger amount of H»SO,
aerosol being collected on the heated leg than the unheated leg. This was
later shown to be due to H_SO, desorption from the heated sample walls of
number 1 leg. This was established by making blank runs without the H-SO,
aerosol generator in operation. After approximately five runs the heated
sample leg reached equilibrium and gave comparable results. Thus, it appears
that the teflon manifold chamber does adsorb some H^SO, which is desorbed
when heated. However, it also appears that the walls stabilize relatively
quickly to current operating conditions.
97
-------
MASS BALANCE
Theoretical H0SO. Concentration
£• H
Since the values given in Table 2 indicated that the H2SO, generator/
sample collection device had reached equilibrium it was decided to attempt a
mass balance to determine how theoretical H2SO, values compared to measured
aerosol values. The values in Table 24 were not used since they showed a
steady downward trend. The aerosol generator stack was cleaned prior to these
experiments and it is thought that the stack was re-equilibrating during those
experiments.
The lineal flow through the aerosol generator stack was measured with a
hot wire anemometer and found to be 950 ft/min and the cross section area
of the stack had a diameter of 4.25 in. This corresponds to a total volume
3
flow of 9.93 m per sample run. During the course of a run (3.8 min), 2.5 ml
of a 0.01 N H-SO, solution was aspirated into the aerosol generator stack
which corresponds to a predicted aerosol concentration of 123 ug HUSO, Per
cubic meter of air.
Measured H?SO, Concentration
The average weight of H2S04 f°und on tne collection filters as shown in
Table 25 was 1.55 yg. Since the volume of aerosol stream sampled was consis-
3
tently 42 liters (1.5 ft ) this corresponds to a measured aerosol concentra-
3 3
tion of 36.9 yg/m or approximately one third of the 123 yg/m aerosol con-
centration predicted.
It had been expected that (NH,)?SO, would thermally decompose at 300°C
to evolve S0_ in a stoichiometric ratio of one mole of SO- evolved for each
mole of (NH,)2SO,. However, previous studies with acetaldoxime and diethyl-
hydroxylamine H2SO, adducts had shown that they thermally decompose to evolve
S00 in a stoichiometric ratio of one to three and one to two, respectively.
Therefore, since the exact stoichiometric release of S02 from (NH.)-SO, had
not previously been established experiments were performed to determine the
amount of S02 evolved from a known amount of (NH,)_SO,.
Twenty microliters of a 0.001 M (NH.) SO, solution were deposited by micro-
pipet onto the surface of several Mitex filters. The filters were then ther-
mally decomposed individually at 300°C and the amount of volatile sulfur
98
-------
species evolved measured with the Flame Photometric Detector. Based on a
standard SO- calibration curve, derived by injecti known SO- gas mixtures
into the decomposition chamber, it was determined that 0.44 and 0.43 yg of
SO- were being evolved from the respective filters. Twenty microliters of a
0.001 M (NH4)2S04 solution contains 2.64 yg of (NH.)-SO,. Assuming that each
mole of (NH,)2SO, decomposed to give one mole of SO- we would expect to find
64
2.64 x " !-28 Vg of S0
evolved from the samples. However, an average of 0.435 yg were evolved from
the known samples resulting in a stoichiometric ratio of
1.28 yg SO- predicted -
0.435 yg S02 evolved °r I
Thus, each mole of (NH,)_SO, produces one third mole of SO when thermally
decomposed at 300°C. Therefore, the values given in Table 25 should be multi-
plied by 3 since they were calculated assuming a one to one stoichiometric
relationship between SO evolution and the amount of (NH.)-SO, present. The
concentration of H SO, aerosol in the sample then becomes 111 yg H^SO, /m
which is close to the theoretical value of 123 yg E^SO./m . The higher
aerosol concentration was used in order to shorten the required sampling
time. The results of these experiments are summarized in Table 26.
TABLE 26. COMPARISON OF PREDICTED AND MEASURED
AEROSOL CONCENTRATIONS
3
Predicted H_SO, Aerosol Concentration 123 yg/m
Measured H2S04 Aerosol Concentration usii|g
NHo Fixing Reagent and Thermal Decom-r -
position at 300°C 111 yg/m
VOLATILIZATION OF H2SO, (Modified Thomas Generator)
Results of initial tests to determine the volatilization of H«SO, with
2. 4
subsequent passage through the prefilter are shown in Table 27. These data
reflect the relative efficiency of the volatilized acid to pass through the
prefilter and to be collected on the sample collection filter at various
temperatures. The aerosol samples were prepared using 0.01 N H-SO, which
corresponds to a predicted aerosol concentration in the generator of 123
99
-------
yg H SO per cubic meter of air. Under these conditions, the maximum amount
of H-SO, which could be sampled, representing 100% sampling efficiency, is
4.48 yg. The total amounts sampled range from 4.31 yg (96%) to 1.05 yg
(23%). The average collection efficiency for all tests is 53%, while the
average for the low temperature tests was 68%
TABLE 27. PERCENT OF ACID PASSING PREFILTERS
(123 yg/m3)
(Modified Thomas Generator - No Flame)
Temperature of
Prefilter, °C
57
70
108
155
Sample
Flow
1/min
9.8
9.8
9.8
6.0
Total yg H2SO^
Prefilter +
Collection Filter
2.90
1.99
3.95
2.53
3.92
4.31
1.61
1.05
1.04
2.20
3.82
1.47
1.25
3.87
2.20
1.45
1.13
% of Total
Collection Filter
17
16
18
13
14
40
35
54
54
59
70
86
81
84
59
70
62
Measured
112804 Aerosol
Concentration
(yg/m3)
69.0
47.4
94.0
60.2
93.3
103.0
38.3
25.0
24.8
52.4
90.9
34.9
29.8
92.1
52.4
34.5
26.9
Collection of H^SO. Aerosol in the 10 yg/m Range (Modified Thomas Generator-
No Flame)
The data presented above for the collection efficiency of sulfuric acid
are based upon the assumption that sulfur interference levels in the lab-
oratory are a minor interference when the samples generated are in the 123
3
yg/m range.
The next series of experiments was devised to duplicate the above results
3
at a tenfold reduced concentration. At 12.3 yg/m sample generation, however,
100
-------
ambient sulfur constituents or ammonia might be a major interference. There-
fore, two preliminary experiments were performed to determine this effect.
First, the experiments with ^SO^, volatilization (123 yg/m3) at high tem-
perature were repeated, but at 200°C. At this temperature, the maximum amount
of H2S04 collected on the sample collection :filter was 89% of the total and
the average amount was 78% of the total. The total collection efficiency was
44% compared to 48% for the tests made at 155°C. This experiment confirmed
the previous tests, establishing a maximum volatilization coefficient, again
assuming (NH^^SO, is not an interference.
The second experiment was designed to insure that minimum ammonia and
ambient surfate interference from the laboratory environment was achieved.
Only after this was determined could the low level H-SO, volatilization be
attempted. Three types of samples were used to determine nominal background
sulfate interference levels. The three samples consisted of blank unused
filters (Mitex LS), filters through which untreated laboratory air passed
up the aerosol generator stack, and filters through which treated (filtered
and H»PO, scrubbed) laboratory air was passed up the aerosol stack. All samples
were subjected to thermal decomposition at 300°C with subsequent FPD detection.
The results of these experiments are given in Table 28.
TABLE 28. BACKGROUND SULFUR LEVELS AS
yg. of H2S04
Blank Filters Treated Air (1.5 ft3) Untreated Air (1.5 ft3)
0.01 to 0.14 yg 0.15 to 0.19 yg 0.21 to 0.55 yg
The individual values vary from day to day with typical values for untreated
3
air being from 0.24 to 13.1 yg/m (as the equivalent amount of H_SO, required
to give the measured background response). Thus, it appears that background
levels of sulfur-containing species would be a significant factor when com-
3
pared to the theoretical concentration of 12.3 yg/m Ho^4 aeros°l- Th£ treated
air, which was normally used for actual experiments gave a maximum background
3
equivalence response of 4 to 5 yg/m H^SO,.
Whether treated or untreated air is sampled, the FPD response is pro-
portional to the amount of air sampled; however, the yg amounts are uniformly
less for H_PO,-treated, filtered air than for the untreated air at ambient
101
-------
conditions. Thus, it appeared that the air scrubber system removed a large
portion of the background sulfur-containing constituents, but that the remainder
3
is still significant when compared to aerosols generated at the 12.3 yg/m
level. It appeared, however, that the residual sulfur species might be the
result of the H SO, generator stack gradually evolving sporatic amounts of
H»SO, aerosol previously adsorbed to the stack walls.
o
Collection of H?SO, Aerosol in the 12.3 yg/m Range (Modified
Thomas Generator - No Flame)
As stated previously, it was expected that residual sulfur species would
be a significant interference at low levels of H-SO, collection. After allow-
ing for background levels the values given in Table 29 were obtained at am-
bient conditions.
TABLE 29. ANALYSIS OF H2S04 AEROSOL AT THE
12.3 yg/m3 LEVEL. (AMBIENT CONDITIONS)
Measured 1^804 Aerosol
Total yg H^SO, Concentration yg/m3
Z " H-
0.66 15.8
0.44 10.5
0.74 17.6
The maximum amount of H-SO, to be expected on the filters, based upon the oper-
ating conditions of the generator, was 0.53 yg as ELSO, . Background samples
(taken before each sampling) ranged from 0.01 yg (as S0_) to 0.25 yg.
An additional series of experiments in the 12.3 V-g range using the heated
zone to volatilize the aerosol resulted in findings which paralleled those
3
obtained with the 123 yg/m aerosol. The results of these experiments, after
correcting for background levels, are given in Table 30.
Collection of H.,SO, Aerosol Tests in the 61 yg/m3 Range (Modified
Thomas Generator - No Flame)
Several experiments were run at an intermediate aerosol concentration
range to determine if the H^SO, collection efficiency dropped with increasing
volatilization temperature. The results of these experiments are shown in
Table 31.
102
-------
TABLE 30. HEATED AEROSOL COLLECTION AT THE
12.3 yg/m3 LEVEL (SAMPLE RATE 9 1/min)
Temperature of
Heated Leg, °C
20
110
110
110
20
Total tig H2S04
Pref liter and
Collection Filter
0.37
0
0
0
0.42
% on
Collection
Filter
0
0
0
0
0
Measured
H2S04 Aerosol -
Concentration, yg/m
8.8
0
0
0
10.0
TABLE 31. HEATED AEROSOL COLLECTION AT THE
61 Ug/m3 LEVEL (SAMPLE RATE 9 1/min)
Temperature of
Heated Leg, °C
20
50
110
110
125
125
145
145
145
Total yg H2SO^
Prefilter and
Collection Filter
2.75
2.11
1.65
1.65
1.22
1.70
0.32
0.74
0.13
% on
Collection
Filter
0
21
0
0
36
38
68
86
100
Measured
^SO^ Aerosol _
Concentration, yg/m
65.4
50.2
39.3
39.3
29.0
40.5
7.6
17.6
4.1
103
-------
Since it was previously shown that H2^°4 could be heated to 105°C and
collected with an average efficiency of 90.2% when a prefilter was not used,
• it appeared that the prefilter had a detrimental effect on total collection
efficiency. It was thought that the poor collection efficiency might be
attributable to the gas stream being cooled too much and the reaction tube
acting as a thermal precipitation or adsorption surface. When samples are
taken under ambient conditions or, in the case of the treated tests, when
only one filter is in place, the vacuum pump operates continuously, as does
the preheater. When two filters are in place, however, it is necessary to
shut down the pump momentarily while removing and analyzing each of the two
filters. This causes fluctuations in the wall temperature and possibly the
adsorption/desorption of E SO, from the sample walls.
Modification of Sampling Device
Heated Reaction Tube —
Normally, the fixing chamber was heated by the gases passing through
the preheated section. The fixing chamber runs at approximately 50 to 60°C,
while the preheater can be operated up to 180°G. In order to narrow the
temperature differential between the two sections, the fixing chamber was
wrapped with four feet of glass-backed heating tape and insulated with fiber-
glass pipe cover. The heating tape was controlled by a Variac.
Reagent Addition Tubes —
Previously all gaseous fixing reagents were added to the reaction chamber
through a single Teflon side arm. At this time, three additional Teflon
side arms were added to the fixing chamber to aid in establishing turbulent
flow. All Teflon side arms were spaced equidistant around the reaction
chamber tube, approximately 4 cm below the prefilter.
Electrostatic Charge —
It was postulated that electrostatic charge accumulation might be res-
ponsible for the sporatic perturbation of data from identical samples. There-
fore, it was decided to modify the Teflon sample tube so that static charge
build up might be prevented. This was attempted by placing three 0.005 cm
gold plated (1%) tungsten wires along the inside walls of both the preheater
104
-------
and fixing chamber. The wires were spaced equidistant and each wire exited
the tube through small holes at each end of the respective tubes. The wires
were held in place by screws partially inserted into the wire exit holed.
In addition to the above modifications, the Teflon tube used to transfer the
aerosol stream from the aerosol generator to the sampler was replaced with a
1.25 cm O.D. stainless steel tube. The gold wires in each sampler section
and the stainless steel sample tube were tied together and grounded.
Effect of Sampler Modifications (Modified Thomas Generator - No Flame)
Since previous experiments at ambient temperatures had shown occasional
perturbations, a series of experiments was run to determine if the above modi-
fications had reduced their frequency and magnitude. Table 32 shows the
results of a series of eight samples collected at the prefilter, while
Table 33 shows the results of eight samples deposited at the collection filter.
All samples taken from the collection filter were fixed in the reaction chamber
with NH_ obtained from a bubbler device containing NH.OH. Samples obtained
at the prefilter were removed from the sampler and immediately fumed over
concentrated NH.OH. All samples were thermally decomposed at 300°C and the
decomposition products passed through the Flame Photometric Detector. All
samples were collected on Mitex filters while the FPD was calibrated with SO,,
mixtures.
TABLE 32. ANALYSIS OF SAMPLES TAKEN AT THE
PREFILTER (Modified Thomas Generator - No Flame)
Sample No.
1
2
3
4
5
7
8
yg H2S04
Per Filter
3.03
3.52
2.76
,76
,79
,76
pg
2.
2,
2.
3.21
216
251
197
197
199
197
229
X 211
S, = 19.9
I Difference
From Mean
+2.4
+18.9
-6.6
-6.6
-5.7
-6.6
+8.5
105
-------
TABLE 33. ANALYSIS OF SAMPLES TAKEN AT THE
COLLECTION FILTER (Modified Thomas
Generator - No Flame)
Wg 2 4 _ , 3 % Difference
Sample No. per Filter ^2 4 m From Mean
1 2.66 190 -18.1
2 3.55 253 _9.0
3 3.18 227 -2.2
4 3.68 263 +13.4
5 2.94 210 -9.5
6 3.18 227 +2.2
7 3.25 232 0.0
8 3.52 251 +8.2
X 232
S. =24.1
d
As can be seen from Table 32, all samples taken at the collection filter
except Number 2 gave results that were well within 10% of the mean value of
211. The standard deviation for this series of experiments was 19.9, which
is within 10% of their mean value. Table 33 shows that samples collected
at the prefilter location were generally within 10% of their mean value of
232 except for Runs 1 and 4. The standard deviation for this series of
tests was 24, which is approximately 10% of the mean value. Previously
reported experiments performed at ambient conditions using the unmodified
sampler gave standard deviations which were 43% and 23% of their mean values.
o
Thus, substantial improvement at the 100 yg/m concentration is evident
due to the sampler modifications. It is felt that the few perturbations
found in the results reported here are due to residual electrostatic
accumulation. The total elimination of this effect would probably require
the use of a nuclear static eliminator.
It should be noted that the results obtained from the two series of
data discussed here were obtained on two different days. Thus, the aersol
generation and sampling systems were totally shut down between the two
series of tests. Considering that the mean results of each series of
experiments agree to within ten percent of the other, this suggests a high
106
-------
degree of reproducibility for the aerosol generator, the sampling methodology
and the analytical technique used to quantitate the results.
Volatilization of HS0 Aerosol at 145°C (Modified Thomas - No Flattie)
Several experiments were performed to determine if H0SO, aerosol could
^ 4
be passed through a Mitex LS filter at 145 °C. The results of these experi-
ments, given in Table 34, show that at 145°C the average slip through the
prefilter is 20%. Thus, it appears that H,SO droplets on the order of
£, l\
1 to 5 ym are not efficiently volatilized at 145°C.
TABLE 34. H2S04 SLIP THROUGH PREFILTER AT 145°C
(Modified Thomas Generator - No Flame)
Total yg H2SO^ yg H2S04 yg H^SO^
Recovered Prefilter Collection Filter Prefilter
2.54 1.76 0.78 31
2.57 2.27 0.32 12 4V
2.30 1.90 0.41 18
An experiment was performed at 145°C in which the acid collected on the
prefilter was exposed to an additional five minutes of the 145°C gas stream
passing through the filter (no H SO aerosol in the stream). Results of this
experiment showed that approximately 50% of the H^SO^ collected had slipped
the prefilter. This explains why varying the total flow through the sampler
in previous experiments from 9.8 1/min to 4.8 1/min did not substantially
affect the volatilization efficiency. It is apparent that the larger aerosol
droplets cannot be efficiently volatilized at nominal reactor lengths (30 cm),
sampler flow rate (1 to 10 1/min) or preheater temperature (145°C) prior
to collection on the prefilter. The advantage of this approach was to have
the H^SO, volatilized before reaching the prefilter, rather than having
2 4
the acid slowly boil off the prefilter. It must, however, be remembered
that in environmental samples, the larger aerosol droplets would probably
have been neutralized by other ambient constituents (NH3, etc.) prior to
their growth into this size range.
107
-------
Quantitative Recovery of Volat?llized H SO, (Modified Thomas Generator)
Previous tests to investigate the volatilization of H_SO/have shown
that up to 75% of the total acid sample was lost when the acid passed through
the prefilter. It was suggested that these losses were due to the combined
effects of temperature differential between the preheater and fixing portions
of the sampler and electrostatic accumulation. Therefore, since the sampler
was modified to reduce the above effects, a series of experiments was per-
formed to determine if any improvement was obtained. The results of these
tests (shown in Table 35) indicate that ^100% of the acid passing through
the prefilter was quantitatively recovered on the collection filter. A
comparison of results in Table 34 with those in Tables 31 and 32 shows
that the mean value for this series of experiments agrees with their mean
values within 5%. The standard deviation of this series of experiments was
13.1 which was within 6% of the mean value. Thus, considerable improvement
is evidenced by the modifications made to the sampler. Prefilter slip,
however, was still not sufficient to suggest that the technique was not
acceptable for aerosols where total sample size is in the low yg range.
TABLE 35. QUANTITATIVE RECOVERY OF H2SO, PASSED THROUGH
THE PREFILTER (Modified Thomas Generator)
, 3 % Difference
yg 2 4/m From Mean
Total
yg R2so4
2.54
2.57
2.30
yg H2so4
Prefilter
1.76
2.27
1.90
yg H2S04
Filter
0.78
0.32
0.41
227 +2.7
230 +4.1
206 -6.8
X 221
S = 13.1
H0SO, Aerosol Volatilization at 190°C (Modified Thomas Generator - With Flame)
In order to study H-SO, droplets in the 0.1 to 0.5 ym size range, the
modified Thomas generator was changed to the flame mode of operation. It was
thought that the slip efficiency problem might be associated with the rela-
tively large aerosol droplet size being studied (1 to 5 ym). A decision
was made that smaller H2SO, droplets might be volatilized more readily and
108
-------
thus pass through the prefliter efficiently. To determine if this were
true, a series of experiments -was performed. The results of these experi-
ments are given in Table 36 and show that at 190°C even H-SO, droplets in
the 0.1 to 0.5 ym range do not pass through the prefilter efficiently.
TABLE 36. PASSAGE OF 1*2804 DROPLETS THROUGH PREFILTER
AT 190°C (Modified Thomas Generator - with Flame)
Total H?SO. H SO _ - .
en ( \ ^4 24 i° £>J-ip
2S04 ('ygj Prefilter Collection Filter Prefilter
8.6 6.7 1.9 22.1
8.9 7.2 1.8 20.2
H PO Scrubbed Aerosol Generator Diluent Air (Modified Thomas
Generator - with Flame)
Since the difficulty was observed in passing the H2SO. aerosol through
the prefilter at 190°C, to insure that ambient ammonia was not reacting with
the aerosol before it reached the prefilter, several experiments were per-
formed in which the aerosol diluent air was passed through an H,PO. scrubber
consisting of a 9 cm O.D. x 10 cm length of phosphoric acid coated quartz
chips connected to the inlet of the diluent air pump. As can be seen from
Table 37, no significant improvement was noted in the percentage of t^SO,
passed through the prefilter as a result of using the phosphoric acid
scrubber. Thus, it appeared that ambient ammonia was not reacting with the
H?SO, aerosol prior to reaching the prefilter.
TABLE 37. PASSAGE OF H2S04 DROPLETS THROUGH PREFILTER
USING H-PO, SCRUBBED AIR (Modified Thomas
Generator - With Flame)
Total H2S°4 ^yg^
H2S04 (yg) Prefilter
7.5 6.7
7.9 7.1
Collection Filter
0.78
0.78
Percent
Slip
11
10
Prefilter
_ Temp.(°C)
144
177
109
-------
H2SO/| Aerosol Volatilization at 190°C for Extended Period
(Modified Thomas Generator - With Flame)
Several experiments were performed in which preheated air (190°C) was
allowed to pass over collected sulfuric acid samples for extended periods.
The purpose of these tests was to determine the relative time necessary
for the H2SO, collected on the filter to volatilize off the prefilter and
pass through the collection filter. As can be seen from Table 38, sub-
stantially more H~SO, passed through the prefilter when heated for an
additional 15 to 25 minutes. These experiments indicated that extended
periods were necessary to volatilize H2SO, aerosols in the 0.1 to 0.5 ym
size range at a prefilter temperature of 190°C.
TABLE 38. PASSAGE OF H2S04 DROPLETS THROUGH PREFILTER
AT 190 °C VERSUS TIME (Modified Thomas - with Flame)
Total H0SO, (yg) H_SO, (yg) ,, « 4. •
qo /• \ 24 2 4 6 Post-Heating
> 4 ^yg; Prefilter Collection Filter % Slip Time (min)
8.8 6.9 1.9 21.6 0
10.7 6.2 4.5 42.0 15
9.6 6.1 3.5 36.4 25
Cabot H-SO, Aerosol Analyzer
- —^ - - -
The Cabot H2SO, Aerosol Analyzer, which works on the principle of
volatilizing H2SO, , uses ultra-dry air and elevated temperatures to bring
about the complete vaporization of H2SO droplets over a period of 15 minutes.
Based on Cabot's results, it appears that a low ambient relative humidity
is necessary to efficiently volatilize the H?SO, aerosols. Since it was
known that the humidity of our aerosol generators runs approximately 100%,
this could be the source of the difficulties. It was, therefore, planned
to use the animal chambers (ultrafine H-SO, aerosol) at the Health Effects
Research Laboratory (HERL) to study the volatilization of H^SO,. The use
of the animal study chambers would enable the generation and control of
the H^SO, aerosol of known size (<0.1 ym) in an environment of controlled
humidity. These studies would show whether even at low relative humidity,
the acid would pass the prefilter.
110
-------
FIELD TESTS
Several field tests were conducted in order that the Atlantic Research
gas phase fixation/analysis concept might be compared with other available
H^SO, collection/measurement techniques. Animal test chambers located
within the Health Effects Research Laboratory (HERL) at the Environmental
Protection Agency, Research Triangle Park, North Carolina, were used for
the first series of field tests. The second series of field tests consisted
of actual enviornmental samples taken in locations of anticipated varying
sulfuric acid aerosol levels. Each series of field tests is discussed
in the following sections.
Animal Chamber Field Tests
The H^SO. aerosol animal chamber used for these studies was located
in the Health Effects Research Laboratory of the Environmental Protection
Agency at Research Triangle Park, North Carolina. The chamber was approxi-
mately 0.5 M wide x 0.5 M wide x 5.0 M long. The entire chamber was
constructed of plexiglass and had plexiglass tubular inserts protruding into
the main chamber area. The tubular inserts were used for exposing animals
to the H-SO, aerosol stream. The chamber was designed to operate with a
constant aerosol droplet size of approximately 0.01 to 0.08 urn (ultrafine),
o
an aerosol concentration of approximately 500 vig/m and a relative humidity
of approximately 50%. In addition, the chamber was fully instrumented
to observe in situ any sudden perturbations in the H«SO, droplet size
(mobility particle analyzer), H SO, droplet concentration (Monitor Lab
Sulfur Analyzer) and the relative humidity (Humidity Monitor). It was felt
that a closely controlled environmental system such as the animial test
chamber would allow for a much more accurate initial assessment of the
gas phase fixation concept than general environmental sampling. This was
considered to be the case since there was no viable H2SO. analysis referee
method suitable for precise field comparison. It was recognized that the
operating aerosol level of the animal chamber was considerably higher than
anticipated environmental levels, but it was felt that the positive aspectsj
of the closely controlled animal chamber environment outweighed the negative
aspect of the higher aerosol concentration.
Ill
-------
The objectives of the animal chamber studies at HERL were to investigate
the following:
1) Prefilter slippage efficiency of H^SO, droplets at various
temperatures in a controlled humidity environment.
2) H2SO, adduct decomposition characteristics.
3) Quantitative comparison of Environmental Protection Agency and
Atlantic Research results on identical samples.
4) (NH^)2SO interference.
5) Storage studies.
Sensitivity Studies —
During the animal chamber studies at HERL, an effort was made to collect
and analyze samples on site. Analysis of these samples required the use of
the Environmental Protection Agency's Bendix Total Sulfur Analyzer. Initial
experiments using the Bendix Total Sulfur Analyzer connected to Atlantic
3
Research's thermal decomposition chamber showed that at 500 yg ti^SO./m
(5 minute sample collection period), a 10 liter sample saturated the Bendix
detector. A second series of runs was made with a sample collection time
of one minute (2 liters) which also resulted in detector saturation. Adjust-
ments to the photomultipHer head voltage and the electrometer did not signi-
ficantly increase the dynamic range of the Bendix Total Sulfur Analyzer.
Calibration curves obtained on the Bendix Total Sulfur Analyzer using SO-
standards showed that this detector was approximately ten times more sensitive
than the Melpar FPD which had been utilized at Atlantic Research. Thus,
the increased sensitivity of the Bendix Total Sulfur Analyzer, coupled
•3
with the larger H^SO, loading (^500 yg/m ) necessitated that the samples
be returned to Atlantic Research for analysis. Since this possibility had
been anticipated, provision had been made to return the samples in plastic
Petri dishes.
From the above calibrations, it was determined that the sensitivity
of the combined Atlantic Research decomposition chamber/Bendix Total Sulfur
Analyzer was on the order of 0.01 yg H^SO,. Based on the original program
requirement of 0.25 yg H S0,/m , it would require a sample of 40 liters
to obtain a sufficient quantity to be detected.
112
-------
H SO Prefilter Slippage —
Studies were performed at several prefilter temperatures in order to
determine the temperature at which HLSO aerosol sampled from the animal
chambers might pass through the prefilter. The device, shown in Figure 11
(page 47) was used during these studies and was designed to raise the temper-
ature of the aerosol sample stream sufficiently to volatilize H SO,.
The device consisted of a 30 cm x 1.8 cm I.D. heavywall Teflon tube heated
by glass-backed tape. At the end of the 30 cm heated section was a Fluoro-
pore prefilter which allowed the volatilized acid to pass while collecting
the remaining particles. Once through the prefilter, the volatilized acid
was mixed with fixing reagent in a 15 cm Teflon mixing chamber. The adduct
was then collected on a second Fluoropore filter at the end of this 15 cm
section. The fixing chamber was also wrapped with a glass-backed heating
tape and insulated in order to accurately control temperature.
Animal chamber experiments were performed using the above apparatus
with prefilter temperatures ranging from 95 to 132°C. The flow rate through
the sampler was held at 2.2 liters/minute using a critical orifice with all
samples being collected for exactly two minutes. After collection, the
samples were stored in plastic Petri dishes for shipment to Atlantic Research
for analysis. Prefilter samples were manually fixed with NH» by fuming prior
to being stored in the plastic Petri dishes. Final filter samples were
fixed in situ in the fixing chamber prior to collection.
The results of these experiments are given in Table 39- The temperature
at both the prefilter and the collection filter are given in the table, as
well as the temperature midway between the sampler inlet and the prefilter
and midway between the prefilter and the collection filter. As can be seen
from the table, the largest H^SO, concentration, 33.5%, occurred with a
prefilter temperature of 95°C. The data trend suggests that the lower
prefilter temperature allows the H-SO, aerosol to pass the prefilter more
efficiently. This is somewhat perplexing since it was thought that the
higher temperatures would causa the acid aerosol to slip the prefilter more
readily. Unfortunately, temperatures below 95°C at the prefilter were not
investigated. In any case, none of the results suggest the efficient
transport of volatilized H SO, through the prefilter.
113
-------
TABLE 39. ANIMAL CHAMBER TESTS (PREFILTER AND COLLECTION FILTER)
Collection Tube Temp.
Sample No.
1 Prefilter
2 Collection
2 Prefilter
2 Collection
3 Prefilter
3 Collection
4 Prefilter
4 Collection
5 Prefilter
5 Collection
6 Prefilter
6 Collection
Preheater
Midway to
Prefilter
157
168
167
168
190
198
Tube Temp.
At
Prefilter
96
95
108
110
126
132
Midway to
Collection
Filter
80
65
72
70
80
80
At
Collection
Filter
65
55
60
60
64
66
yg H2S04
Filter
1.69
0.72
2.41
2.02
1.02
3.04
3.04
0.22
3.26
2.19
0.22
2.41
2.25
0.39
2.64
2.65
0.39
Percent .
Prefilter
Slippage
29.9
33.5
7.2
9.1
14.8
12.8
3.04
-------
Thermal Decomposition Studies —
These studies were performed to evaluate fixing reagents and their
corresponding adducts rather than H2SO, volatility. Three different H-SO,
fixing reagents were used during these experiments in the HERL animal
chambers in order to determine the decomposition temperatures of the corres-
ponding adduct formed from ultrafine (M3.05 urn) H2SO, aerosol. The reagents
used for these studies were acetaldoxime, N-methylformamide and ammonia.
Ammonia was used as a fixing reagent to determine the effect of (NH,)2SO,
interference on the other two adducts. The three fixing reagents and their
corresponding adducts are discussed below. It should be noted that all
samples were taken at ambient temperature with no prefilter installed at
the end of the volatilization chamber.
(NH.),.,SO. Studies — Two series of experiments were conducted during
this study. The first series of experiments involved fixing the H^SO,
aerosol with ammonia during the collection process, while the second series
involved waiting approximately 10 minutes before NH., fixation. Sufficient
samples were collected during each series so that thermal decomposition
information could be obtained at several temperatures. The purpose of
these studies was twofold: (1) to determine the decomposition temperature
of (NH,) SO, formed from ultrafine (0.05 ym) H2SO, droplets; and (2) to
determine the effect on decomposition temperature when the HUSO, droplets
were allowed to age an additional 10 minutes prior to forming (NH,) SO .
The data from these experiments, shown graphically in Figures 29 and 30,
indicate that there is very little, if any, difference in the amount of
SO evolved at a given decomposition temperature by either method of NH
fixation. It should be noted however, that the relative humidity within
the animal chambers was on the order of 40%, so that the aerosol droplet
growth rate was somewhat inhibited. It appears from the data shown in
Figure 30 that the H2SO fixed with ammonia ten minutes after sample collec-
tion begins to decompose at a temperature of approximately 185°C. Unfor-
tunately, no decomposition data was obtained between 170° and 196°C for
the H0SO aerosol droplets fixed immediately with ammonia during sample
^ 4
collection. At 170°C, however, there was no sign of sample decomposition.
115
-------
1.4
1.2
1.0
O
ui
> 0.1
§
UI
f
0.4
0.2
180 184
188
192 1% 200 204 208
DECOMPOSITION TEMPERATURE °C
212 216
220
Figure 29. Submicron particles fixed with ammonia
ten minutes after sample collection.
1.4
1.2
1.0
a
ui
> 0.8
0.6
0.4
0.2
.A.
170 174 178 182
190 194 198 202 206 210 .214 216 222
DECOMPOSITION TEMPERATURE °C
Figure 30. Submicron particles fixed with ammonia
during sample collection.
116
-------
NMF/AAO Studies — Several aerosol samples were reacted with N-methyl
formamide and acetaldoxime in order to determine the decomposition tempera-
ture of the associated adducts. The results of these experiments are
shown in Figures 31 and 32.
1.0
o
2
§ 0.6
1U
f
0.4
170 174 178 182 186 190 194 198 202
NMF/S04 DECOMPOSITION TEMPERATURE °C
Figure 31. Decomposition profile of NMF-SGi, adduct
1.4
0.8
0.6
8"
O2
Ol—&
172 176
180 184
192 W6 200 20*
> • SO4 DECOMPOSITION TEMPERATURE *C
212
216
Figure 32. Decomposition profile of AAO-SO. adduct.
117
-------
As can be seen from the figures, at 174°C neither the NMF nor the
AAO adducts decompose, while at 184°C significant decomposition appears to
take place. Thus, it would seem that the temperature range between 174°C
and 184°C is of primary importance since the (NH,)_SO, experiments indicated
that it did not decompose until approximately 185°C. A composite curve of
(NH,)7SO,, AAO-SO, and NMF-SO, is shown as Figure 33. It appears from the
composite curve that there may be sufficient difference in decomposition
temperatures to allow for the selective decomposition of the AAO/NMF adducts
in the presence of (NH,)~SO,.
1.4
1.2
1.0
0.8
o
tu
§
UJ
i
5
0.6
0.4
0.2
-O— — (NH4)2SO4
-£ --- NMF-S04
•D ..... AAO SO
,1
*£•§
172 176 180 184
192 196 200 204 208 212 216
DECOMPOSITION TEMPERATURE °C
Figure 33. Composite decomposition profile - AAO-SO, , NMF-SO, ,
Quantitative Comparisons —
Samples were taken from the HERL Animal Chambers on July 19 and 20,
1978, for quantitative comparison between the gas phase fixation/thermal
decomposition technique, the EPA procedure (thorin) , and the in situ
Monitor Labs Sulfur Analyzer. The Environmental Protection Agency (EPA)
and Atlantic Research Corporation (ARC) samples were taken on Milipore FH
filters, 0.5 ym pore size. The ARC samples were fixed during sample collec-
tion, removed from the sampler and immediately placed in plastic Petri
118
-------
dishes for shipment to ARC. The EPA samples were immediately placed in a
solution of 5X10 N perchloric acid for stabilization purposes. The
Monitor Labs Sulfur Analyzer was a continuous analyzer, and therefore
required no sample storage.
The ARC results from all samples taken on July 19 are included in Table
40, while the results from all the quantitative samples taken on July 20
are given in Table 41. As can be seen from the two tables, the average
daily values obtained by ARC were within 10% of one another. During the
course of this sampling, EPA obtained concurrent samples for analysis by
the thorin technique. The results of the EPA thorin analysis are shown
in Table 42, along with the ARC results from the pertinent samples and the
real time total sulfur analyzer reading. Table 43 is a summary of the
individual results given in Table 42 and reflects the daily averages for
both ARC and EPA data.
It is interesting to note in Table 42 that the FPD data suggest that
no sudden HUSO, perturbations occurred in the animal chamber during these
tests. It, therefore, seems logical that the data from EPA samples 33-138,
33-139, 33-145 and 33-147 are low. This is further substantiated by the
ARC data which suggests that the H SO, concentration was not halved during
the sampling period. The fact that both sets of data reflect one half of
the H~SO, aerosol concentration previously determined by both the other
EPA results and the ARC results suggests the possibility of a 2:1 dilution
error during the chemical analysis.
If the questionable samples are deleted from the sample matrix and
the remaining data averaged for each of the two days, the comparisons given
in Table 44 are obtained. These results show that on 7/19/78 the EPA and ARC
results agreed to within 8%, while on 7/20/78, the data agreed to within 14%.
The results also suggest that very little, if any, ARC sample loss occurred
during a one-month storage period between sample acquisition and analysis.
The samples were kept refrigerated in the original Petri dishes during
this period.
Environmental Field Tests
Samples were taken from two locations for this series of experiments.
Two sets of samples \. .re taken in a rural location of the Shenandoah Valley
119
-------
TABLE 40. ANIMAL CHAMBER STUDIES - ALL QUANTITATIVE
SAMPLES TAKEN BY ARC ON 7/19/78
us EnSO,/m
ARC Sample No. Time of Sampling 2 4
071978-1 1255 541
071978-2 1300 490
071978-5 1313 490
071978-6 1318 592
071978-12 1345 755
071978-13 1348 684
071978-14 1353 704
071978-17 1405 776
071978-24 1502 736
071978-26 1509 736
071978-31 1532 724
071978-32 1536 847
071978-34 1544 817
Average Values for 7/19/78 X 684
Sd 118
C 0.17
TABLE 41. ANIMAL CHAMBER STUDIES - ALL QUANTITATIVE
SAMPLES TAKEN BY ARC ON 7/20/78
ARC Sample No. Time yg
072078-1 0925 735
072078-8 1004 613
072078-9 1008 664
072078-12 1033 511
072078-21 1258 642
072078-27 1341 551
072078-28 1344 756
Average value for 7/20/78 X 639
Sd 9°
C 0.14
120
-------
TABLE 42. ANIMAL CHAMBER STUDIES - INDIVIDUAL SAMPLE COMPARISON
BETWEEN EPA AND ARC SAMPLES
ro
Real Time
Total Sulfur
Chamber
Date Sample Time of Concentration
Taken Sampling FPD (ppb)
7/19 12:55-1:45 PM 111
109
111
111
7/19 2:37-3:27 PM 114
113
7/20 9:18-10:08 AM 105
106
106
7/20 10:23-11:13 AM 105
7/20 1:26-2:16 PM 107
108
7/20 3:32-4:22 PM 114
ARC "g H
Sample No.
071978-1
071978-2
071978-5
071978-6
X
Cv
071978-24
071978-26
X
Sd
Cv
072078-1
072078-8
072078-9
X
072078-12
X
Sd
072078-27
170278-28
I
Sd
Samples taken fox
thermal decomposition
studies during thia
period.
(ARC)
541
490
490
592
528
48
0.09
736
736
736
0
0
735
613
664
671
62
0.09
511
-
551
756
653
144
0.22
Equivalent
EPA
Sample No.
33-138
33-139
33-140
33-141
33-142
33-143
33-144
33-145
33-146
33-147
33-148
33-149
ug H2so4/n
(EPA)
309
246
X 278
Sd 44
Cy 0.16
599
499
X 549
S
-------
TABLE 43. ANIMAL CHAMBER STUDIES - DAILY AVERAGE DATA SUMMARY
COMPARISON BETWEEN EPA AND ARC SAMPLES
7/19/78
7/20/78
ARC
541
490
490
592
736
736
EPA
309
246
599
499
ARC
735
613
664
511
551
756
EPA
603
500
603
510
281
241
597
466
X 598
S 114
Cy 0.19
X 413
S, 164
d
C 0.40
X 638
Sd 98
0.15
X 475
S, 142
Q
G 0.30
TABLE 44. ANIMAL CHAMBER STUDIES - AVERAGED DATA
(Questionable samples deleted)
7/19/78
Pg
X
Sj
ARC
598
114
0.19
pg
EPA
ARC
7/20/78
Ug H2SO /nr
549
71
0.13
X
S,
638
98
0.15
EPA
546
62
0.11
122
-------
at the Junction of Route 55 and Rural Route 616 halfway between Strasburg
and Front Royal, Virginia. The results of these experiments are shown in
Table 45.
TABLE 45. RESULTS OF ENVIRONMENTAL FIELD TESTS -
RURAL AREA (SAMPLE RATE 10 £/MIN)
Sampling Time ARC Sample No. yg H0SO,/m3 EPA Sample No. H+
— 2 4 £
10:00 AM -
1:00 PM 112278-1NMF 0 112278-1UF Neutral
5ilOPM~ H2278-2NMF 0 112278-2UF Neutral
Two sets of roadside samples were taken on a heavily traveled portion
of Interstate 522 located within the Front Royal, Virginia, city limits.
The sampling location was near the apex of a 60° hill within 15 feet of the
roadside. A sulfuric acid production plant (H2SO aerosol), a power plant
(S02), and a rayon production plant (H S, CS2) were all located within a two
mile radius of the sampling point. The results, which were expected to
show ambient sulfate levels are given in Table 46.
TABLE 46. RESULTS OF ENVIRONMENTAL FIELD TESTS -
URBAN AREA (SAMPLE SATE 10 A/min)
O -L.
Sampling Time ARC Sample No. yg H,,SO,/m EPA Sample No. H
^•SS PM ~ H2478-1NMF 6.0 112478-1UF Basic
*:?n 2! ~ 112478-2NMF 9.T' 112478-2UF Basic
b:J-U rM
Samples collected simultaneously with the ARC samples shown in Tables
45 and 56, but not treated with fixing reagent, were sent to EPA contractor
+ + =
laboratories at Research Triangle Park for analysis for H , NH^ , and SO^ .
Due to a misinterpretation, however, only H was measured. The results of
the H+ analysis are included in Tables 45 and 46. No sulfuric acid was
found in the rural area, which is confirmed by the EPA findings of a neutral
H+ concentration for the pertinent samples (Table 45). The urban area samples
showed the presence of sulfates as expected, with the EPA H readings of
basicity for each sample indicating the possible formation of anmonium
123
-------
sulfate artifact during sample collection (Table 46). Basic species
would not originally be anticipated in such close proximity to a roadside
sampling network.
SUMMARY
Gas-phase fixation of sulfuric acid aerosol by a volatile amine or
amine derivative was proposed as a convenient and effective method of
stabilizing the acid aerosol prior to sample collection. In addition, it
was suggested that the adduct could be conveniently analyzed if it decom-
posed to evolve SCL gas at a temperature below that of anticipated ambient
particulate sulfates. In order to determine at which temperature the adducts
could be selectively decomposed, various potentially interfering sulfates
were taken from reagent bottles and subjected to Differential Thermal
Analysis (DTA). All of the sulfates examined by DTA were found to decompose
above 200°C. Thus, it was concluded that the ideal fixing reagent should
react with H~SO, aerosol to form an adduct which could be decomposed below
200°C and evolve S02 gas as a product.
The diethylamine (DBA), acetaldoxime (AAO) and diethylhydroxylamine
(DEHA) derivatives of EUSO were subjected to DTA in order to determine
the temperature at which they decomposed. It was established that each
adduct decomposed at a temperature below the decomposition temperature
of anticipated interferences. A flame photometric detector (FPD) was
subsequently used to establish that volatile sulfur gases were evolved
during decomposition. Sulfur dioxide was identified as a major product
during the decomposition of each of the above adducts by mass spectrometry
and West-Gaeke analyses. Infrared spectra were also obtained for the acetal-
doxime and the diethylhydroxylamine adducts of sulfuric accid. A peak at
1650 cm in the infrared spectra of both adducts indicated an N-H bending
vibration which was attributed to protonation of nitrogen in the basic
fixing reagent. It appeared that the basic fixing reagents reacted rapidly
with sulfuric acid by proton transfer, forming a bisulfate salt. Rapid
reaction kinetics is a desirable characteristic since the tendency for sul-
furic acid to react with other species during the collection process would
be lessened.
124
-------
A Thomas aerosol generator, which reportedly supplies aerosol droplets
in the 0.005 to 0.3 micrometer size range, was used to make sulfuric acid
aerosol. The Thomas generator was used with each of the previously mentioned
fixing reagents to produce a corresponding sulfuric acid adduct. This was
accomplished by concurrently passing both fixing reagent and sulfuric acid
aerosol through a reaction chamber connected between the Thomas generator
and the sample collection filter. The Teflon filters containing the adduct
were then placed in the Teflon decomposition chamber and heated to 200°C.
It was established with this technique that the fixation process was very
rapid and essentially complete after a few seconds. In addition, it was
shown using the FPD that the sulfur gas evolved from the decomposition of
the adducts was proportional to the quantity of sulfuric acid aerosol
sampled.
Ammonium sulfate and ammonium bisulfate were considered to be a primary
interference problem due to the close chemical similarity to H-SO, and the
relatively large amount of each species thought to be present in the environ-
ment. Differential Thermal Analysis of reagent grade ammonium sulfate and
bisulfate indicated that both salts were stable to at least 250°C and 230°C,
respectively. Experiments with the FPD indicated that no volatile sulfur
species was evolved by either of these salts until at least 230°C. When
ammonia was passed through a filter on which acid aerosol from the Thomas
generator had been collected unfixed, as is normally the case in environmental
sampling, no FPD signal was detected at 200°C. Thus, interference from this
source was apparently not present for adducts prepared using acid aerosol
particles in the 0.005 to 0.3 micrometer size range.
It was recognized that ambient sulfur gases such as S0?, H-S and COS
might react with fixing reagents to form complexes which could be collected
and subsequently decomposed along with the sulfuric acid adduct. Since
the FPD responds to any sulfur species, the complexed sulfur gas could
constitute a positive interference. To test this possibility, each of the
above gases was mixed with DBA, DEHA and AAO fixing reagent and passed
through Mitex LS filters for several minutes. The filters were then
subjected to a rising temperature in the decomposition chamber to determine
both the presence of a volatile sulfur species and the temperature at which
125
-------
it was produced. It was found that of the three sulfur gases, only SC^ was
significantly collected, and in every case, the SO^-reagent complex had a
very low thermal stability compared to that of the H2SO, adduct. Thus,
the interference could be eliminated by pre-heating the filter to an inter-
mediate temperature where the S09-reagent complex would be removed before
decomposition of the H^SO adduct at a higher temperature.
A Baird aerosol generator, which reportedly supplies aerosol particles
in the 1 to 3 micrometer size range was also selected for this study
because it was thought that the closed system design would give more repro-
ducible aerosols than the Thomas generator. Initially, the adducts formed
using the Baird generator were decomposed in order to verify decomposition
temperatures. The results were surprising in that the adducts previously
investigated using the Thomas generator now had decomposition temperatures
that were different. The decomposition temperature (as determined by the
evolution of sulfur gas) of the AAO and DBA adducts shifted upward while the
DEHA adduct shifted downward. Of the three adducts, only that of AAO showed
a substantial change. One possible explanation for the shift in adduct decom-
position temperature was that the various aerosol generation methods which
resulted in aerosol particles of widely different size range also resulted
in particles with a large difference in free surface energy. This difference
in free surface energy resulted in corresponding differences in individual
adduct decomposition temperatures. However, it was observed that the most
significant shift in decomposition temperature occured with the AAO adduct
which contains an unsaturated bond. An additional study using n-methyl-
formamide (NMF) which also has an unsaturated bond was also found to exhibit
a very big difference in decomposition temperature. Thus, it appeared that
the chemical nature of the fixing reagent in conjunction with the aerosol
generation method affect the decomposition temperature of specific H.SO,
adducts. This may be due in part to the fact that the collection temperature
of the Thomas generator aerosol was approximately 60 C. while the Baird
generator aerosol was collected at approximately 30 C. The collection
temperature difference was due to the flame generation method, used in
the Thomas generator.
126
-------
The questions associated with the decomposition of the various adduets
led to a reinvestigation of the major anticipated interferent, (NH,).?SO,.
The Baird generator was used to deposit (NHJ-SO, aerosols made from 0.01 N
and 0.001 N (NH^SO^ solutions. It was assumed that since less (NH.).SO,
per droplet would be present, this would result in a smaller particle wh*n
dry. The above solutions were also deposited on filters by pipet. In
addition, (NH^SO^ was formed using the H SO, aerosol obtained from the
Thomas generator.
When the Baird aerosol generator was used to deposit samples from a
solution of 0.01 N (NH^SO^, an average value of 175° C. was found to be
the initial temperature at which the aerosol decomposes. However, when the
identical solution was deposited by pipet and subjected to the analysis
procedures an average initial decomposition temperature of 183° C. was
obtained. When the Baird generator was used to deposit the 0.001 N (NH,)_SO,
solution, the initial response temperature dropped to an average of 165 C,
Simultaneous exposure of H2SO, aeros°l produced by the Thomas generator to
NH_ resulted in a further decline of the observed initial decomposition
temperature of (NH.)-SO, to an average of 158 C. However, when the H?SO,
aerosol was first collected and then exposed to NH_ the initial (NH,)2SO,
decomposition temperature rose to an average of 220 C. It is thought
that the lack of agreement in the Thomas generator results is due to H_SO,
aerosol droplet growth on the collection surface prior to the addition of
NH«. Thus, it appears that particle size does influence the decomposition
temperatures of the various species to a large degree. This is a particularly
significant problem requiring further study since the distribution of
sulfuric acid aerosols in the environment is thought to cover a large size
range which indicates that the corresponding adduets would also deviate
considerably.
Ambient ammonia, which was considered to be the primary negative inter-
ference due to reaction with sulfuric acid on collection surfaces, was inves-
tigated during this study. As originally conceived, a desirable aspect of
the proposed method was the prediction that sulfuric acid adduets would be
less likely to react during the collection process than the original acid
aerosol. The relative reactivity between ammonia, H.SO, aerosol and various
H SO adduets was therefore determined. It was found that the reaction
2 4
127
-------
between ammonia and sulfuric acid was suppressed on collection surfaces as
the result of forming sulfuric acid adducts during the collection process.
The success of the proposed research approach also depends upon whether
a reproducible stoichiometric relationship exists between the sulfur gaaes
evolved during adduct decomposition and the H-SO, originally present in
the sample stream. It was originally hypothesized that the gases evolved
during adduct decomposition should consist primarily of SO-. Therefore,
experiments were performed to obtain information about the reproducibility
of the Baird generator and the quantity of SO- evolved at 200 C. from the
AAO adduct. The West-Gaeke technique was used instead of the FPD to measure
the evolved gases since it was reportedly specific for SO-. It was found
that the amount of H_SO, aerosol collected during identical runs was
reproducible to within +10% based upon the amount of S02 evolved during the
thermal decomposition process. Thus, it appears that the Baird generator
is fairly reproducible, as is the evolution of SO- from an identical quan-
tity of AAO adduct. Additional tests using three different volumes of the
identical H_SO, aerosol stream were sampled in the ratio of 1:2:3. The
corresponding H-SO, adducts when decomposed were found to evolve SO- gas
in approximately the identical ratio. Thus, it appears that AAO can be used
to fix H-SO, aerosol in a moving gas stream to form an adduct which can be
decomposed to give off an amount of SO- that is related to the volume of
aerosol stream sampled.
In order to determine an exact stoichiometric relationship between the
SO- evolved during adduct decomposition and the original H-SO, aerosol pre-
sent, known amounts of H-SO, were deposited on Mitex filters by pipet. An
equal volume of fixing reagent was then applied to the sample and dried.
Decomposition of the adducts at 200 C. showed that the stoichiometric
relationship between the SO- evolved and the original acid aerosol was repro-
ducibly one mole of S0_ to three moles of H?SO, when both the sampling and
analysis were accomplished under controlled laboratory conditions. Exper-
iments identical to those above for the AAO adduct were also performed for
DEHA and DBA derivatives of sulfuric acid. It was found that the DEHA
adduct evolved SO- in a ratio of one mole of SO- to two moles H-SO, and
2. i 2 4
the DBA adduct evolved S02 in a stoichiometric ratio of one mole of SO- to
one mole of H?SO,.
128
-------
The AAO, NH- and NMF adducts of sulfuric acid were found to give off
gaseous sulfur species in a stoichiometric ratio of one mole of S09 to three
moles of sulfuric acid during thermal decomposition. Using the above values,
a mass balance was made between measured H0SO. aerosol concentration and the
24
predicted aerosol concentration based on the generator's operating conditions.
Calculations predicted that the H-SO, aerosol concentration should be 123
3 ^^
yg/m while measurements using a fixing reagent coupled with thermal decom-
position found the concentration to be 111 yg/m3. In addition, field tests
using the Environmental Protection Agency's Animal Test Chambers at Research
Triangle Park resulted in close agreement between the amount found by the
gas phase fixation methodology and the Thorin technique used by the EPA.
The alternate approach of eliminating sulfate particulate interference
by selectively volatilizing sulfuric acid and passing it through a prefilter
was investigated during this study. It was found that at temperatures
up to 200°C, sulfuric acid aerosol in the low yg range could not be volati-
lized and efficiently passed through a prefilter.
In conclusion, it appears that the gas-phase fixation of sulfuric acid
aerosol by a volatile amine or amine derivative offers definite advantages
over existing techniques. Among these are ease of calibration using SO™
standards rather than H_SO,, increased sensitivity and suppression of side
reactions. However, the apparent effect of particle size on decomposition
temperature must be resolved before the method can be used to satisfactorily
eliminate ammonium sulfate interferences.
129
-------
REFERENCES
1. Firket, J., "Fog Along the Meuse Valley," Trans. Faraday Soc., Vol. 32,
1102 (1936).
2. Schrank, H., H. Heimann, G. Clayton, W. Gafaffer and H. Wexler, "Air
Pollution in Donora, Pa.," Public Health Bulletin, Washington,
No. 306 (1949).
3. Stern, A. C., ed., "Air Pollution and Its Effects" in Air Pollution;
A Comprehensive Treatise, Vol. 1, Academic Press, New York (1968).
4. Leighton, P., "Photochemical Secondary Reactions in Urban Air" in
Photochemistry of Air Pollution, Pergamon Press, London (1957).
5. Pemberton, J. and C. Goldberg, "Air Pollution and Bronchitis," Brit.
Med. ,J., 567, Sept. 4 (1954).
6. Amdur, M., "Toxicologic Appraisal of Particulate Matter, Oxides of
Sulfur and Sulfuric Acid," J. Air Pollut. Contr. Ass., 19, 683 (1969),
7. Amdur, M., R. Schulz and P. Drinker, "Toxicity of Sulfuric Acid Mist
to Guinea Pigs," Arch. Ind. Hyg. Occup. Med., 5_, 318 (1952).
8. Leithe, W., The Analysis of Organic Pollutants in Water and Waste
Water, Ann Arbor, Ann Arbor, Mich., 13 (1973).
9- Koide, M. and E. Goldberg, "Atmospheric Sulfur and Fossil Fuel Combus-
tion," J. Geophys. Res., ^6_, 6589 (1971).
10- Wolf, P., "Carbon Monoxide, Measurement and Monitoring in Urban Air,"
Environ. Sci. and Tech., 5_, 212 (1971).
11- McEwen, D., "Automobile Exhaust Hydrocarbon Analysis by Gas Chromato-
graphy," Anal. Chem.. ^8, 1047 (1966).
12. Stephens, E. and M. Price, Aerosols and Atmospheric Chemistry, ed. by
G. Hidy, Academic Press, New York, 167 (1972).
13. Raymond, A. and G. Guiochon, "Gas Chromatographic Analysis of Cfl-C
Hydrocarbons in Paris Air," Environ. Sci. and Tech.. jj, 143 (1974).
14. Pellizzari, E., J. Bunch, E. Berkley and J. McRae, "Determination of
Trace Hazardous Organic Vapor Pollutants in Ambient Atmospheres
by Gas Chromatography/Mass Spectrometry/Computer," Anal. Chem.
48, 803 (1976). ~~
130
-------
15. Habibi, K., "Characterization of Particulate Matter in Vehicle Exhaust,"
Environ. Sci. and Tech., ]_, 223 (1973).
16. Blosser, E. R., "A Study of the Nature of the Chemical Characteristics
Of Particulates Collected from Ambient Air," NTIS No. PB-220-401.
17. Urone, P. and W. Schroeder, "SC-2 in the Atmosphere: A Wealth of
Monitoring Data, but Few Reaction Rate Studies," Environ."SoI.
and Tech.. 2, 8 (1968).
18. Sidebottom, H. and J. Calvert, "Photooxidation of Sulfur Dioxide,"
Environ. Sci. and Tech., J3, 1 (1972).
19- Friend, J., R. Leifer and M. Trichon, "On the Formation of Stratos-
pheric Aerosols," J. Atmos. Sci., 30, 465 (1973).
20. Freiberg, J., "Effects of Relative Humidity and Temperature on Iron-
Catalyzed Oxidation of S0? in Atmospheric Aerosols," Environ. Sci.
and Tech., 8, 731 (1974).
21. Novakov, T., A. Barker and W. Siekhaus, "Sulfates as Pollution
Particulates: Catalytic Formation on Carbon (soot) Particles,"
Science, 186 (1974).
22. Forester, P., "Oxidation of Sulphur Dioxide in Power Station Plumes,"
Atmos. Environ., .3, 157 (1968).
23. Katz, M. , "Photoelectric Determination of Atmospheric Sulfur Dioxide,"
Anal. Chem., 22, 1040 (1950).
24. Scaringelli, R. P. and K. A. Rehme, "Determination of Atmospheric
Concentrations of Sulfuric Acid Aerosol by Spectrometry,
Coulemetry and Flame Photometry," Anal. Chem., 41, 707 (1969).
25. Takahashi, K., M. Kasahara and J. Katz, "A Kinetic Model of Sulfuric
Acid Aerosol Formation from Photochemical Oxidation of Sulfur
Dioxide Vapor," Aerosol Science, J5, 75 (1975).
26. Mirabel, P. and J. Katz, "Binary Homogeneous Nucleation as a Mechanism
for the Formation of Aerosols, Journal of Chemical Physics, 60 (3),
1138 (1974).
27. Whitby, K., "Sulfuric Acid Review Conference," Hendersonville, N. C.,
Jan. 31, 1977 to February 3, 1977.
28. Amdur, M. 0., "Aerosols Formed by Oxidation of Sulfur Dioxide:
Review of Their Toxicology," Arch. Envir. Health, 23, 459 (1971).
29. Vanderpol, A., F. Carsey, D. Covert, R. Charlson and A. Waggoner,
"Aerosol Chemical Parameters and Air Mass Character in the St.
Louis Region," Science, 190, 570 (1975).
131
-------
30. Willard, Richard L., Preprint, National Meeting Division Water, Air
and Waste Chemistry, Am. Chem. Soc. (1973).
31. Meserole, F. B., et_ &L., "Sulfur Dioxide Interferences in the Measure-
ment of Ambient Particulate Sulfates," NTIS No. PB-249-620 (1976).
32. Faith, W. L., Air Pollution Control, John Wiley and Sons, Inc., New
York, 143 (1959).
33. Taarit, Y. and J. Lunsford, "Electron Paramagnetic Resonance Evidence
for the Formation of S0~ by the Oxidation of S0~ on MgO," Journal
of Physical Chemistry, .11, 1365 (1973).
34. Vohia, K. and P. Nair, "Chemical Formation of Aerosols in the Air by
Gas Phase Reactions," J. Aerosol Sci.. ±, 127 (1970).
35. Kiang, C. and P. Hamill, "Sulfuric Acid-Nitric Acid-Water Ternary
System in the Stratosphere," Nature (London), 250, 401 (1974).
36. Lee, R. E. and J. Wagmen, "Sampling Anomaly in the Determination of
Atmospheric Sulfate Concentration," Amer. Ind. Hyg. Assn. J., 27,
(3), 266 (1966).
37* Dubois, L., T. Teichman, R. S. Thomas and J. L. Monkman, Symposium on
the Physico-chemical Transformation of Sulfur Compounds in the
Atmosphere and the Formation of Acid Smogs, Paris, December 1967,
Paper No. 8.
38v Barton, S. C. and H. G. McAdie, "Preparation of Glass Fiber Filters for
Sulfuric Acid Aerosol Collection," Environ. Sci. and Techno1., 4^
769, (1970).
39. Maddalone, R. F., A. D. Shendrikar and P. W. West, "Radiochemical
Evaluation of the Separation of H?SO, Aerosol by Microdiffusion
from Various Filter Media," Microchem. Acta, ^ (1974).
40. Thomas, R. L., V. Dharmarajan and P. W. West, "Convenient Method for
Generation of Sulfuric Acid Aerosol," Environ. Sci. and Technol.,
j3, 930 (1974).
41. Maddalone, R. F., R. L. Thomas and P. W. West, "Measurement of Sul-
furic Acid Aerosol and Total Sulfate Content of Ambient Mr,"
Environ. Sci. and Technol., 10, 162 (1976).
42. Thomas, R. L., V. Dharmarajan, G. L. Lundquist and P. W. West,
"Measurement of Sulfuric Acid Aerosol, Sulfur Trioxide and the
Total Sulfate Content of the Ambient Air," Anal. Chem.. 48. 639
(1976).
43, Liu, B. and K. Lee, "Efficiency of Membrane and Nucleopore Filters
for Submicron Aerosols," Environ. Sci. and Technol.. 10, 345 (1976).
132
-------
44. Kerrigan, J. V., K. Snajberk and E. S. Andersen, "Collection of
Sulfuric Acid Mist in the Presence of a Higher Sulfur Dioxide
Background," Anal. Chem., 3_2, 1168 (1960).
45. Rastogi, R. and K. Kishore, "Kinetics and Mechanism of Reaction
Between Alcohol and Sulfuric Acid at High Temperatures," Ind. J.
Chem., 1_, 480 (1969).
46. dos Santos, Serpa, "Spectrophotometric Determination of Tetracycline
by its Reaction with Concentrated Sulfuric Acid," Bol. Fac. Farm.,
Univ. Coimbra, Ed. Cient., 31, 29 (1972).
47. Svirevski, I. and M. Miloshev, "Color Reaction Between Substituted
Indenones and Concentrated Sulfuric Acid," Dokl. Bolg. Okad.
Nauk., _2_7_ (3), 361 (1974).
48. Ladler, K., Reaction Kinetics; Vol. II - Reactions in Solutions, The
Macmillan Company, New York, 85 (1963).
49. Huygen, G., "Simple Photometric Determination of Sulfuric Acid Aerosol,"
Atmospheric Environment, 9, 315 (1975).
30, Moeller, T., Inorganic Chemistry. 8th Printing, John Wiley (1959).
51« Weast, R. C., ed., Handbook of Chemistry and Physics, 49th ed., The
Chemical Rubber Co., Cleveland, Ohio (1968).
52• Lange, N,, ed., Handbook of Chemistry, 10th ed., McGraw-Hill Book Co.,
New York (1952).
53. Belcher, R., M. Kapel and A. Nutten, "Some Factors Influencing the
Solubilities of Amine Sulfates," Anal. Chim. Acta, 8^, 122(1953).
54. Belcher, R., A. Nutten and W. Stephen, "Substituted Benzidines and
Related Compounds as Reagents in Analytical Chemistry. XII. Reagents
for the Precipitation of Sulfate," J. Chem. Soc., 1334 (1953).
55. Stephen, W., "New Reagent for the Detection and Determination of Small
Amounts of the Sulfate Ion" Anal. Chim. Acta. 50_, 413 (1970).
56. Maddalone, R. F., G. L. McClure and P. W. West, "Determination of
Sulfate by Thermal Reduction of Perimidylammonium Sulfate,11
Anal. Chem., 47^ 316 (1975).
57. Barrett, W., H. Miller, J. Smith, Jr. and C. Gwin, "Development of a
Portable Device to Collect Sulfuric Acid Aerosol - Interim Report,"
Contractor Report EPA-600/2-77-027, Environmental Protection
Agency, Research Triangle Park", N. C. (1977).
58. Raabe, 0. G., Fine Particles; Aerosol Generation Measurement, Sampling
and Analysis, ed. by B. Y. H. Liu, Academic Press, 57-110 (1976).
133
-------
59. West, P. and G. Gaeke, "Fixation of Sulfur Dioxide as Disulfitomercurate(II)
and Subsequent Colorimetric Estimation," Anal. Chem.. 12, 1816 (1956).
60. Dubois, L., C. Baker, T. Teichman, A. Zdrojewski and J. Monkman,
"Determination of Sulfuric Acid in Air," Mikrochimica Acta (Wien),
269, 1969.
61. Erdey, L., S. Gal and G. Liptay, "Thermoanalytical Properties of
Analytical Grade Reagents: Ammonium Salts," Talanta, 11, 913.(1964).
62, Grundnes, J., S. Christian, C. Venghuot and S. Farnham, "Solvent
Effects on Strong Charge-Transfer Complexes. IV. Trimethylamine
and Sulfur Dioxide in the Vapor Phase," J. of the Am. Chem. Soc.,
93^ 20 (1971).
63. Landreth, R., R. de Pena and J. Heicklen, "Thermodynamics of the
Reaction of Ammonia and Sulfur Dioxide in the Presence of Water
Vapor." J. of Phys. Chem.. 17. 1785 (1975).
134
-------
APPENDIX A
REACTIONS IN FLOW SYSTEMS
135
-------
REACTIONS IN FLOW SYSTEMS
It was necessary that the gas phase fixation concept and the reaction
chamber design be evaluated analytically to insure that the total concept
is theoretically feasible. In order to make a theoretical assessment of
the system, approximate values were obtained fair reagent concentrations,
molecular collision frequencies, gas/liquid diffusion rates, reaction stoi-
chiometry and reaction chamber residence time. Once these values were
obtained, it was determined that the physical characteristics of the dynamic
flow system were sufficiently rapid to allow complete fixation to occur
during reaction chamber residence time.
The theoretical assessment of the proposed gas/liquid phase concept
is the subject of this Appendix. A model is assumed for the fixation
process and the separate elements of that model are examined analytically
herein. The model can be described as follows: A mixture of gaseous fixing
reagent at a given concentration in air enters into and mixes with the main
flow stream of the apparatus, the latter consisting of micron-size droplets
of sulfuric acid in air. Molecules of vapor fixing reagent (i.e., diethyl-
amine) impinge upon the aerosol droplets, are absorbed at the surface,
diffuse throughout each droplet, and react with the H«SO,. The time re-
quired for these events to occur is taken as the reactor residence time
for the purposes of this analysis. The much longer time period, during
which acid and reagent remain in contact on the filter is neglected. The
preliminary design of a physical model which provides the necessary mixing
and reaction zone is shown in Figure 1. The sample containing the acid
aerosol (symbolized by "0" on the figure) would be drawn into the device
at A, while the gaseous fixing reagent ("X") would then react in Zone C
to form the adduct (OX) which would be collected on the filter medium.
Residual gaseous fixing reagent would pass through the filter to the
pumping source.
136
-------
6*8 or Vapor Inlet (X)
OJ
th for
Vapor Chamber
Filter
Medlim
To Pu»p
Figure 1. H2SO, filter collection System (gas or vapor reagent).
-------
DISCUSSION
A. Calculation of Reagent Vapor Concentration in
the Vicinity of H0SO. Collection Filters
,
Initially, to approximate the kinetics of the proposed method,
the concentration of the various reagents in the reaction chamber must
be considered. The reagents currently of interest in this study are
diethylamine (PEA) , diethylhydroxylamine (DEHA) , acetaldoxime (AAO) ,
dimethylamine (DMA) and hydroxylamine (HA). The vapor pressure at room
temperature for each of the reagents above can be approximated from their
known boiling points.
1. Vapor Pressure of Reagents
The boiling point is the temperature at which the vapor pressure
equals atmospheric pressure. The use of Trouton's Rule:
* 21 cal/°K mole (1)
where AH is in cal/mole and T, is the boiling point in °K, enables
one to estimate the vapor pressure of non-polar liquids at a given temper-
ature when the standard boiling temperature is known. Trouton's rule
states that the molal entropy of vaporization (AS ) at the standard
vap
boiling point (T, ) of many liquids is equal to approximately 21 cal/°K
mole (1). While the proposed fixing reagents are not non-polar, Trouton's
Rule may hold well enough to permit the estimation of reagent-vapor
pressures at room temperature.
In order to compute the vapor pressure of a liquid from its known
boiling point, the Clausius-Clapeyron equation is used:
AH (T- -T,)
vap 2 1
2.303
138
-------
where P,, is atmospheric pressure when TZ refers to the boiling point,
and Pj^ is the vapor pressure of the liquid at T .
Substituting (21 * Tb) for AH^ from equation (1), converting
to exponential form, and simplifying, equation (2) becomea:
' P2
i ~ VI
"DT I
RT1 J
The value of the gas constant R is 1.987 cal/mole degree.
Using the known boiling point value (T_) for each reagent and taking
T, as 298 °K, approximate vapor pressures were computed at ambient temper-
ature from equation (3). The results are collected in Table 1. Published
values for the vapor pressure of diethylamine and acetaldoxime are
260 mm and 10 mm respectively (2) . These values are in fairly good
agreement with the calculated values in Table 1.
TABLE 1. Calculated Vapor Pressures
Reagent Known B.P. (°C) V.P. Calculated (25 °C)
Diethylamine 55 262.3 mm
Dimethylamine 7.4 1418.5 mm
Acetaldoxime 114 32.3 mm
Hydroxylamine 110 37.3 mm
Diethylhydroxylamine 125 21.9 mm
2. Concentration of Vapors
The reagent vapors which enter B of Figure 1 are produced by bubbling
air through the volatile reagents. This saturated or partially saturated
gas stream is further diluted by the sample air stream entering at point A
of Figure 1. The concentration of each reagent vapor in the gas stream
at B is simply the ratio of each reagent's vap/xr^ pressure (Table 1)
to the total pressure of the systems. Thus, if it is assumed that the
dynamic gas stream is saturated with vapor and that it is at atmospheric
pressure, the concentration becomes:
139
-------
where C, is the reagent vapor concentration in terms of volume percent.
By using the vapor pressure values of Table 1 with equation (4) , the
values shown in Table 2 are computed as -the approximate concentrations
of the various fixing agents at point B of the H2SO, collection system.
The flow of the gas stream entering the reaction chamber at B is
estimated to be 0.5 1/min, while the sample stream has a maximum flow of
approximately 34 1/min. Thus, the concentration of vapors in the reaction
zone is:
Cc=xCb C5)
where Cfe and GC represent the volume percent reagent concentration at
points B and C respectively. The values for the concentrations at C are
given in Table 3.
B. Rate of Reaction of Fixing Reagents with H^SO, Aerosol
Once the concentration of vapors in the reaction manifold has been
defined, estimates can be made to determine if the rate of various processes
are sufficiently rapid to permit complete reaction during the time the sample
and vapors transverse the reaction chamber (C) .
1. Molecular Collisions
Since the proposed method involves the reaction of a vapor with an
aerosol particle (liquid surface) , the calculation of the number of
collisions per second per unit area would serve as an indication of the
rate of reagent impingement that could be anticipated on the surface of
the aerosol droplet. The rate of molecular collisions per unit area is
given as:
140
-------
Table 2. Estimated Vapor Concentration at Inlet
(Section B of Figure 1)
Reagent
Diethylamine
*Dimethylamine
Acetaldoxime
Hydroxylamine
Diethylhydroxylamine
Volume %
34
100
4.2
4.9
2.9
Ppm
3.4 x HT
1.0 x 10*
4.2 x 10*
4.9 x 10^
2.9 x 10^
*Dimethylamine boils at 7.4°C and has a calculated vapor pressure of 1418 mm
at 25°C. Thus, dimethylamine does not require a diluent carrier, and
enters as a pure vapor at point B in the collection system.
TABLE 3.
Reagent
Diethylamine
Dimethylamine
Acetaldoxime
Hydroxylamine
Diethylhydroxylamine
Estimated Vapor Concentration in Manifold
(Section C of Figure 1)
Volume %
0.50
1.5
0.06
0.07
0.04
Ppm
5,000
15,000
600
700
400
141
-------
Collisions _ n 00_ x
2 0,230 nu (6)
cm sec
where 0.230 is the directional characteristic for molecules striking a
given surface, assuming that the molecules travel in a random direction.
The number of reagent molecules per cubic centimeter is represented by n
in equation (6) and denotes the concentration of the vapor in the reaction
zone (C) of Figure 1. The approximate concentrations of the various reagents
in the reaction zone are given in Table 3. The term u in equation (6) is
the average root mean square velocity of ideal gaseous molecules in motion
and is equal to:
/3RT
u = —— (7)
where T is temperature (298°K), m is molecular weight of each reagent, and
R is the gas constant (8.314 x 10 j
equation (6) with equation (7) gives
R is the gas constant (8.314 x 10 joules degree mole ). Combining
Collisions n OOA /3RT ,_.
T = 0.230 n (8)
2. m
cm sec
f\
Collisions/cm sec, calculated using equation 8, are given in Table 4 for
each of the candidate reagents. As can be seen from the table, on the order
20 2
of 10 collisions/cm sec can be anticipat*
concentration within the reaction chamber.
20 2
of 10 collisions/cm sec can be anticipated based upon estimates of reagent
2. Residence Time in Reaction Manifold
The amount of time available for the reaction to occur is determined
by the volumetric flow rate of the gas and the length and cross sectional
area of the reaction tube. The diameter of the tube is 2.5 cm and its
3
length is 20 cm, which yield a volume of 98 cm .
Since the flow through the reaction chamber is known to be 34 1/min
or 0.57 I/sec, the residence time in the reaction chamber (c) is:
142
-------
TABLE 4. Collisions/cm2sec at 298°K
for Candidate Reagents
n* u Collisions
Reagent m molecules/cm^ (cm/sec) cm2sec
Diethylamine 73 1.1 x 1017 3.2 x 104 8.1 x 1020
Dimethylamine 45 4.0 x 1017 4.1 x 104 3.7 x 1021
Acetaldoxime 59 1.6 x 1016 3.5 x 1Q4 1.3 x 102°
Hydroxylamine 33 2.0 x 1016 4.7 x 104 2.2 x 102°
Diethylhydoxylamine 89 1.2 x 1016 2.9 x 1Q4 7.8 x 1Q19
*Based upon vapor concentrations in Table 3.
0 OQ8 1
=0.17 seconds or % 200 milliseconds.
6.571/sec
»
3. Vapor/Aerosol Collisions During Chamber Residence
The aerosol droplets are estimated to be of the order of one micron
in diameter. The surface area of a sphere is
SA = irD2 (9)
where D is the sphere diameter. Therefore, the surface area of a one micron
-8 2
sphere is 3.14 x 10 cm .
o
The number of collisions/cm sec for each of the candidate reagents
were previously calculated and are given in Table 4. Thus, since the surface
area of an aerosol droplet one micron in diameter is known, and the molecular
collision rate for each reagent has been estimated, then the total number
of collisions of reagent vapor with an aerosol droplet is:
_ , 11.. collisions ,, - . - _-8 1 .
Total collisions = „" x 3.14 x 10 cm x 0.2 sec (10)
cm sec
143
-------
for the measured residence time in the reaction chamber. Total collisions
for each of the reagents with an aerosol droplet as it passes through the
reaction chamber are given in Table 5.
TABLE 5. Total Reagent Vapor/Aerosol Collisions
During Reactor Residence
Reagent Total Collisions
12
Diethylamine 5.1 x 10
13
Dimethylamine 2.3 x 10
Acetaldoxime 8.2 x 10
12
Hydroxylamine 1.4 x 10
Diethylhydroxylamine 4.9 x 10
4.
Molecules H?j>0, in One Micron Aerosol Droplet
The stoichiometric reaction of reagent vapors with H_SO. molecules in
an aerosol droplet requires that a sufficient number of reagent molecules
impinge on a droplet during reaction chamber residence time to completely
react with the H-SO, molecules. An assessment of this criteria requires that
a correlation be made between the total molecular collisions from the vapor
and the actual H~SO, molecules within the aerosol droplet.
The number of molecules of H SO, within an aerosol droplet of known
molarity is:
molecules H«SO, „
, „ = V x C x 6 x 10 J (11)
droplet DM v '
where V is the aerosol droplet volume, C is the molar concentration of
D 23 M
H«SO and 6 x 10 is the number of molecules per mole. Current estimates
for the concentration of H~SO, in ambient aerosols range from approximately
0.01 to 7.0 molar, depending upon the relative humidity and the aerosol
particle size. Taking the worst theoretical case, 7.0 molar H^SO, aerosol
and a one micron droplet size, the number of molecules of H~SO,/one micron
2 4
droplet becomes:
144
-------
Molecules H2SO,
"
ly droplet = 1/^dO cm) x 7 moles/* x
23 9
6 x 10 = % 2 x 10 molecules H2S04/ ly droplet
Based upon a comparison of values from Table 5 of total collisions
(%10 ) with total H2S04 molecules within an aerosol droplet (£109) , it
can be concluded that sufficient collisions occur to completely react all
H2SO , Thus, the rate of collisions of reagent molecules with the aerosol
droplets is not a controlling factor, if the "sticking coefficient" is
9 12 -3
10 /10 or greater than 10 . The term "sticking coefficient" refers to
the molecular collisions that stick to or penetrate the aerosol droplet
rather than bounce or ricochet off the surface. The reagents proposed for
this program may have "sticking coefficients" near unity, since they are
highly water soluble.
5. Characteristic Molecular Diffusion Time in a Liquid
The reaction between vapor molecules and H2SO, molecules is most likely
to take place throughout the aerosol droplet. There are several reasons for
this: first, the H2SO, is distributed throughout the droplet; secondly,
the reagents are soluble in H~0; and finally, the H?SO adducts are also
water soluble. Thus, an important criteron is the rate at which the molecular
species distribute themselves throughout the aerosol droplet, since this will
also determine the reaction rate.
The diffusion of molecular species can be estimated using:
X2 - 2Dt 0-2)
where X is a mean diffusion distance, D is the diffusion coefficient and
t is a characteristic diffusion time. Using the one micron aerosol
CO . .. ______ „._ _________________ _________ _____
dimension for X and a value of 10 cm /sec for D* in equation (12),
-4
t is estimated to be 5 x 10 sec or %1 millisecond.
* -5 2, . a ^Mroi value for the diffusion of molecular species
° 56th edition, page MO).
145
-------
Since the reactor residence time is approximately 200 milliseconds
and the diffusion time for molecular species in a one micron droplet is of
the order of one millisecond, there is sufficient time for diffusional mixing
throughout the aerosol droplet.
6. Conclusions
There are many physical processes involved in the fixation of H_SO,
molecules within an aerosol droplet. The calculations, based upon the
experimental model show that reagent concentrations in the reaction zone
are adequate to generate sufficient molecular collisions to completely
fix the H2SO, during reaction chamber residence. In addition, diffusion
of reagent vapors in the aerosol droplet is sufficiently fast to permit
fixation of the H2SO, throughout the droplet's volume during reaction
chamber residence. Thus, the theoretical examination of the physical
processes involved during gas phase fixation of H-SO, aerosol indicates
that it is a feasible approach. This, of coufse, assumes that the chemical
reaction between H2SO, and the basic amines occurs at a high rate relative
to physical processes.
REFERENCES CITED
1. Daniels, F. and R. Alberty, Physical Chemistry, 3rd ed., p. 159,
John Wiley, 1966.
2. Weast, R., ED., Handbook of Chemistry and Physics, 56th ed., The
Chemical Rubber Company, Cleveland, OH (1975).
146
-------
APPENDIX B
MEASUREMENT TECHNIQUES FOR SULFURIC ACID
AND SULFURIC ACID ADDUCTS
147
-------
MEASUREMENT TECHNIQUES FOR SULFURIC ACID
AND SULFURIC ACID ADDUCTS
Sulfuric acid undergoes reaction with many types of compounds that
could conceivably yield a sulfuric acid adduct worthy of investigation.
The sulfuric acid adducts will be of varying physical and chemical des-
criptions and each will probably require a unique analytical approach in
order to evaluate its merit. The adducts must be examined for stability,
stoichiometry, speed of formation and interference effects. Clearly,
the analysis of each adduct can be a considerable effort, since the con-
centrations will be low and the physical states varied. Various analytical
methods and their potential for measurement of sulfuric acid and
sulfuric acid adducts are discussed below.
Wet Chemistry Techniques
+ _2
The sulfuric acid molecule consists of two species, H and SO, ,
which can be used to measure its concentration. The measurement of either
species is not specific for sulfuric acid when other acids or sulfates
are present. Once the sample is dissolved, interfering species contribut-
+ -2
ing H or SO, ions cannot be distinguished from the ions generated by
sulfxiiric acid. Thus, the applicability of wet chemical techniques to the
analysis of sulfuric acid necessitates a viable separation technique that
eliminates potential interferences before dissolution.
1. Proton Measurement
Many investigators have used proton measurement techniques to deter-
mine sulfuric acid content. Most of these indicators are organic dyes
which exist in two forms: the base form possesses one color which by gain
,*
of a proton is converted to a diffecent colored acid form. An example of
this mechanism, the reaction of methyl orange with one proton, is given
below.
148
-------
[Na+
\ / " \ / "VV'>M;»J •«
•OH
Yellow in
hone solution
Pink in
acid notation
Assume that a sample of sulfuric acid was collected on a fiberglass
filter. The filter could then be sprayed from an aerosol can with a
selected indicator. The indicator could be dissolved in an organic
solvent to prevent possible hydrolysis of metal sulfates forming colored
products. The intensity of this color would serve as a preliminary
indication of the amount of sulfuric acid present, determining whether
laboratory analysis is necessary. This further analysis can be performed
by dissolving the dye in a non-colored, non-chromophoric organic solvent,
and determining the concentration spectrophotometrically.
Some indicators that have been used to determine l^SO, aerosol
include methyl red, neutral red, methyl orange, bromocresol green, methyl-
thymol (I), and bromophenol blue (2).
The acid-base indicator method would be subject to interference from
other non-volatile acids, such as H PO,, but volatile acids such as HCl
would not be retained by the filter during collection. The indicator spray,
however, must not be added prior to sampling, because the volatile acids
would react as they passed through the filter. In addition, metal sulfates
should not interfere to a significant degree, as the color formed is
due to the protons of the sulfuric acid. Since the reaction is run in
organic solvents, no dissolution of these sulfates should occur. This
technique should work on any of the collection techniques previously
described, where the volatile acids are not collected with sulfuric acid.
2. Classical Techniques for Sulfate Analysis .
The classical methods of sulfate analysis could be applied to the
problem of measuring sulfuric acid, if interfering sulfates were eliminated
during sample collection, or if the preferential release of sulfuric acid
149
-------
were realized. Essentially, this would be total sulfate analysis with
other sulfates being absent. The main thrust of the newer literature
on sulfate analysis has been aimed at modification of four main procedures:
1) gravimetry, 2) titrimetry, 3) colorimetry, 4) turbidimetry or nephelo-
metry. This has been done in order to achieve greater accuracy and to
eliminate interferences.
a. Gravimetry
The classical, though seldom used, method for analysis of sulfate
is by precipitation of barium sulfate from a hot, slightly acid solution
by the addition of barium chloride:
_ I [
SO, + Ba •> BaSO,4-
The precipitate must be digested, filtered, the paper ignited, and the
residue weighed as BaSO,. The procedure is long and tedious, and subject
to many interferences due to coprecipitation of other substances. Inter-
fering substances are largely cations such as lead, strontium, and calcium
which form insoluble sulfates, .although anions of weak acids, nitrates,
chlorates and heavy metal ions interfere if care is not taken to remove them.
There are several organic compounds, such as benzidine, 4,4'-diaminotolane
and 2-aminoperimidine, which can be used as precipitants for sulfate. How-
ever, their use has been predominately for nephelometry and turbidimetry
determinations whidh are discussed in the following paragraphs,
b. Titrimetry
There are many modifications of titrimetric methods of sulfate
determinations. Most of the methods involve the pretreatment of the solu-
tion by passing it through a cation exchanger, or addition of suitable
reagents to eliminate interferences from metal ions. Direct titrations
with a barium salt, either barium chloride or barium nitrate, have been
reported using various indicators, such as diphenylcarbazone (3), sodium
alizarinsulfonate (^). and nitrochromeazo (5), etc. Potentiometric
titrations using both barium and lead salts have been reported as well.
Probably the most satisfactory titrimetric method is an indirect determin-
ation accomplished by adding a known amount of barium strontium, or lead
150
-------
salt and titrating the excess with a suitable reagent, such as ethylene-
diaminetetraacetic acid (EDTA) (6> There are other variations of indir-
ect titrations too numerous to mention. However, in general, the titri-
metric methods used to date do not lend themselves well to the determination
of low concentrations of sulfates.
c< Colorimetry
Colorimetric methods for sulfate determinations follow the same
general outline as found in gravimetry and titrimetry in that barium
is added to precipitate barium sulfate. The methods usually involve one
of three general procedures:
1) A known amount of barium salt is added to the sulfate solution
and the excess barium is complexed with a reagent such as
methylthymol blue (7). The remaining barium is then deter-
mined colorimetrically, thus allowing the amount of sulfate
to be calculated:
I j .j. .f-j.
excess Ba + S04 -> BaSO^-l- + excess Ba
excess Ba + MTB -»• BaMTB
2) A known amount of barium is added to precipitate the sulfate
followed by addition of a known amount of potassium chromate
to precipitate the excess barium. The amount of sulfate
present is determined by measuring the absorbance of the excess
chromate:
4-f = ++
Ba + SO, -»• BaSO,4- + excess Ba
excess Ba4""1" + K2Cr04 ->• BaCr04 + excess Cr04 (yellow)
3) A known amount of the insoluble barium chloranilate is allowed
to exchange with the unknown sulfate solution to form barium
sulfate and the highly colored acid chloranilate ion. The
intensity of the color of the chloranilic acid is proportional
to the amount of sulfate present.
BaC6Cl204 + S04= -»• BaS04* + H^CL^ (violet)
151
-------
The exchange of an insoluble metal chloranilate with an anion to form
chloranilic acid and precipitate the tnetal-anion salt is a very sensitive
reaction.
d. Turbidmetry and Nephelometry
In these methods for sulfate determination, barium chloride, 4,4'-
diaminotolane, 2-aminoperimidine, or benzidine, etc. are used to form
a fine precipitate with sulfate. The basis of both measurements is the
interaction of light with a fine suspension of particles. The differences
in the methods are in the way this interaction is measured. In nephelo-
metry, the radiation that is scattered by the particles is measured
perpendicular to the axis of incident light. Turbidimetry, on the other hand,
measures the amount of radiation that passes through the suspension and
thus, the measurement is made along the axis of the incident light. Both of
these methods are very sensitive to small amounts of sulfate, if the
proper reagent is used.
All of the above methods assume that interfering sulfates have been
eliminated by selective solubility techniques prior to analysis. All
sulfates, no matter what the cationic form, will be detected by these
methods, once they are solubilized.
Potential Instrumental Methods for Measuring H/>Sp. Addy,Q.ts
There are many commercially available instruments designed to
monitor S0« and/or other volatile substances at low concentrations. The
secret of applying most types of instruments to the measurement of H«SO,
aerosols is to develop an H2SO, adduct that stabilizes the H^SO, in the
field, but can also be programmed to give off a stoichiometrically related
substance for analysis. For instance, if the volatilized or decomposed
adduct formed a sulfur containing species, then flame photometry could
be used as the analysis method. If the adduct could be made to contain
halogens, then the electron capture detector would be advantageous as
the method of analysis. A totally organic structure could besi: be deter-
mined by the flame ionization detector.
The procedure for volatilization or decomposition of this type H~SO,
adduct consists of programming or isothermally setting the sample holder
152"
-------
cell temperature sufficiently high to bring about the necessary change
in state. The sample can then be channeled, through sufficiently heated
lines, to the detector of the particular measuring instrument. Of course,
the necessary precautions must be taken to prevent condensation, once
the sample has reached the instrument.
Another possibility would be to find a solvent that selectively
dissolves the I^SO^ adduct. If the fixed H2SC>4 adduct is colored, or
absorbs energy in the ultraviolet region, its intensity could be measured
spectrophotometrically. Of course, this technique calls for an adduct
that is sufficiently selective to react only with the H-SO, portion of the
sample. Any other sample interaction might cause erroneous readings,
especially if the interference was at the same wavelength as the H-SO
absorption adduct. Phosphoric acid aeros&l could interfere with acid-base
reactions, while the presence of additional sulfates could interfere with
reagents designed to detect H_SO.. Selective solubility would also be
L. H"
beneficial as a preliminary step, to clean up the sample before injection
into a chromatographic column. Selectivity would have to be one of the
major criteria for choosing a solvent to function in this capacity.
Analytical techniques that presently appear feasible for the determina-
tion of H?SO, adducts are presented in the following sections.
1. -Flame 'Photometric Detector (FPD)
Several instrumental methods for sulfur measurement are available which
make use of atomic emission and absorption characteristics. The one which is
most convenient for measurement of a gas containing sulfur is the Flame Photo-
metric Detector (FPD). Sample gas is fed onto an air-H_ flame which causes
sulfur atoms to emit their characteristic spectrum. A 374 nm optical filter
selectively allows the sulfur light emission to reach a photomultiplier tube
(older units are 395 nm filters, but these are subject to interference from
hydrocarbons). This detector responds to any sulfur present, regardless of
the oxidation state. The main advantage is the sensitivity, which ranges from
the low nanogram range to about ly g of SO.
153
-------
Scaringelli and Rehme (8) developed a highly selective, sensitive method
for determination of sulfuric acid aerosols in the atmosphere. The aerosol
was collected by impaction or filtration, which separates the HLSO, from the
S0?. The sulfuric acid was isothermally decomposed to sulfur trioxide under
a nitrogen stream at 400°C. The liberated SO- was then converted to sulfur
dioxide by reaction with hot copper. The S0_ can be determined through spectro-
&
photometric, coulometric, or flame photometric analysis. The only sulfate
found to interfere is ammonium sulfate, which decomposes at approximately 300°C.
The sensitivities of this method for colorimetric, coulemetric, and flame
photometric analysis are 0.3, 0.03, and 0.003 yg, respectively.
Willard (9) described an instrument based upon a technique similar to
the one described above for the separation of H9SO, from other volatiles
using warm ultra-dry air instead of high temperatures. The lower temperature
and the ultra-dry air enable the H SO, to be volatilized before (NH,)2SO,,
and thus eliminates it as an interference.
The FPD appears to offer an ideal compromise between selectivity and
sensitivity.
2. Gas Chromatography
Chromatography is a physical method for the separation of the components
of a mixture. The separation of the components is accomplished by a section
of tubing packed with a stationary phase of large surface area. The stationary
phase can be a solid or a liquid; hence, the names gas-liquid and gas-solid
Chromatography. Gas-solid Chromatography is the technique of the stationary
phase (solid) to separate the sample constituents by absorption. In gas-
liquid Chromatography, a thin film is spread over an inert solid, and the
sample is partitioned in and out of this liquid film.
Qualitative analysis is accomplished by comparing the retention time
(time from the injection of the compound of interest until it is registered
by the detector) of the unknown compound with the retention time of standards.
An identical retention time between sample and standard is an indication of
the identity of the unknown species. However, qualitative analysis by Chroma-
tography is not an absolute measurement since it is possible for more than one
compound to have the same retaation time.
154
-------
Quantitative analysis is accomplished by comparing the detector response
of known concentrations with the detection response of the compound in the
sample. This comparison is generally accomplished by peak area, peak height,
or any number of methods referenced in the literature.
Butts and Rainey- (1°) made trimethylsilyl derivatives of various anions,
including sulfates, and analyzed them by separation on OV-17 and SE-30 columns.
Subsequent detection of the eluting derivatives was made by flame photometric
and flame ionization detectors. The derivatives were prepared by placing 5
to 10 mg of the desired anion in a septum-capped vial and adding 200 yl each
of dimethylformamide and bis(trimethylsilyl)-trifluoroacetamide. It appears
feasible that H-SO, adducts could be separated from analytically interfering
species using adducts of sufficient volatility and columns of acceptable
resolution.
Sulfur halides have been separated on Kel-F oil No. 3-fit)"while organic
sulfate derivatives have been analyzed on the ethyl ester of Kel-F 8114 acid,
hexadecane (12,13), squalene (14), dinonylphthalate, and on tricresyl phosphate (15)-
These are just a few of the many types of columns used to analyze various sulfur
compounds. The final column selection would ultimately depend upon the type of
H.SO, adduct formed and the specificity of the chromatographic detector utilized.
There are many types of sensitive chromatographic detectors that could
be used to detect the eluting H_SO, adduct. A list of sensitive chromato-
graphic detectors (.16) that could be utilized are presented in Table 5. Of
course, the use of any detector could be restricted by the type of H0SO.
i e. 4
adduct being analyzed.
3.~ Mass Spectrometry
Mass spectrometry probably provides more molecular structure information
than other analytical technique for the analysis of organic and inorganic
species. The information generated from the mass spectrometer is usually
sufficient to determine the structure of the species being examined empirically.
The mass spectrometer fragments a molecule to produce charged ions usually
consisting of the parent ion and ionic fragments. The ion fragments are then
sorted according to the mass-to-charge ratio. The mass spectrum is a measure-
ment of all the different ion fragments and thiir relative intensitifes. No
155
-------
TABLE 5. Gas Chromatographic Detectors- (16)
Detectability
Detector (g/sec)
Thermal Conductivity 10
.-6
Flame lonization
10
,-12
Alkali Flame
lonization
Flame Photometric
Electron Capture
10
,-15
10
,-12
10
,-14
Helium lonization 10
-14
Mass Spectrometer 10
,-15
Comments
Despite the relatively low
detectability of this
system, it is probably
the most widely used for'
virtually any kind of
substance.
Nearly a universal detector;
however, many gases of
interest to environ-
mentalists give little
or no signal.
Tremendous sensitivity for
phosphorus compounds (pesti-
cides) . Otherwise, limited
to compounds of nitrogen,
sulfur, and halogens.
Limited application to phos-
phorus and sulfur compounds.
Very useful despite its
limited sensitivity to
halogenated compounds
and other electronegative
atoms
A universal detector, but
"limited" to low tempera-
tures (<100°C) and to the
type of column used. Very
sensitive to leaks and
contamination.
Very expensive but essential
to the absolute identifica-
tion of components in many
analyses. A universal
detector.
Applicability to
CUSP, Program
Detects all volatile
substances and is rela-
tively insensitive.
Very little use for
this research.
Very sensitive to
all organic adducts.
Sensitivity propor-
tional to organic
structure size. Good
candidate for this -
research.
Fair potential for use
in this research.
Excellent potential
for this research. Very
sensitive to adducts
containing sulfur.
Specific for sulfur.
Fair candidate for
this research. Extremely
sensitive to halogenate
adducts. Sensitivity
proportional to electrot
negativity of adduct.
Very sensitive for all
substances. Severe
problems with column
bleed. Fair potential
for this research.
Extremely sensitive to
all compounds. Can
distinquish individual
species in a mixture.
Good potential for use
in this research.
156
-------
two species exhibit the same spectrum, and thus mass spectra are suitable for
positive identification of any species detected.
Qualitative interpretation of a "mass spectrum" consists of a critical
examination of the fragment ions and piecing them together to form a molecule
consistent with the original fragmentation pattern. Final identification is
usually made by matching the unknown spectrum with the spectrum of a standard
run on the same instrument. Quantitative interpretations are made by running
a standard sample of known concentration and pressure, and calculating the
response per unit of pressure based upon a particular ion fragment. The .
response of the sample can then be used to calculate the partial pressure of
the sample constituent and, hence, its concentration.
The mass spectrometer offers a large potential for use as an H-SO, adduct
detection method. The detector offers both sensitivity and selectivity with
the only requirement being that the sample exist in the gaseous state at the
temperature and pressure existing in the ion source. The mass spectrometer
is a highly sensitive^device capable of seeing nanogiram quantities of sample.
Provisions are made on most mass spectrometer inlet systems for analysis of
solid, liquid, and gas samples.
4. X-Ray Diffraction
Every crystalline substance scatters x-rays in its own unique diffraction
pattern. The resulting diffraction pattern generated by a crystal is used
to identify its specific structure. The intensity of each reflection and
the angle at which it forms, supplies the basic information for the determina-
tion of the crystalline structure and its subsequent quantitation. It appears
feasible, at this point, that x-ray diffraction could be used to quantitate
sulfuric acid adducts, provided that they are crystalline in nature and present
in sufficient quantity.
Robert, e£ al.(17) reported that graphite which has been treated with an
oxidizing agent to remove electrons will allow sulfuric acid to insert between
the carbon layers. The amount inserted was determined by x-ray diffraction.
X-»ray diffraction is not regarded as one of the more sensitive analytical
techniques.
157
-------
5. X-Ray Fluorescence
X-ray Fluorescence depends on the excitation of secondary x-rays charac-
teristic of each element by adsorption of primary x-rays. The identification
of these characteristic wavelengths and the measurement of their intensities
constitute a method for qualitative and quantitative analysis.
Since x-ray fluorescence is responsive to an element, no matter in which
state it is, it would be necessary to separate the H2SO, adduct from other
species containing the same element to be measured prior to analysis. Thus,
while x-ray fluorescence is extremely sensitive, it would not be specific for
an adduct, unless it could be insured that the element being measured in the
adduct was not present in another form. Since, under normal conditioas, x-
ray fluorescence cannot be used on gases, a selective solvent would have to be
found for the H-SO, adduct, and the interfering species filtered off prior to
analysis.
1>. Atomic Absorption
In the atomic absorption method of analysis, solutions or dilute suspensions
are atomized in a flame and reduced to elemental state. Atoms in the ground state
of a given element absorb energy from a beam of light emitted by a hollow cathode
lamp of the element sought. The instrument can be employed for flame emission
spectroscopy, where spectral emission from ionized atoms of an element is measured.
Again, the same problem exists as with x-ray fluorescence, in that all
states of a particular element are detected simultaneously. This would require
a selective solvent approach to insure that interfering ions are eliminated
prior to analysis. In addition, the H_SO, adduct would have to oontain a
metal, so that the adduct could be detected by atomic absorption.
7. Infrared Spectroscopy
The infrared spectrum is one of the most characteristic properties of a
compound. Characteristic spectra originate primarily from the vibrational
stretching and bending modes within molecules when irradiated with electro-
magnetic energy in the infrared region. It is doubtful"that"InFrared spec-
troscopy exhibits sufficient sensitivity to detect HUSO, adducts in the range
of interest to this research.
158
-------
8. ESCA (Electron Spectroscopy for Chemical Analysis)
Electron spectroscopy is performed on the inner shell electrons emitted
from samples irradiated with x-rays. The kinetic energy of the ejected elec-
trons is equal to the energy of the incident photons hv less the electron
binding energy E, . A measurement of the electron kinetic energy is thus a
means of identifying the elements in the sample and the quantity of electrons
is proportional to the concentration.
Stein (18) reported that ESCA is the most sensitive analytical technique
13 2
at the present time, able to detect 10 sulfate/sulfur atoms per cm . Un-
fortunately, this is a new procedure, the equipment is expensive, and the
technique uncommon. However, ESCA merits consideration as a possible analyti-
cal method for the H?SO, program.
9V~ Chemiluminescence Monitors
Dzubay, Rook and Stevens (19) have described a technique using a chemilumi-
nescence monitor to measure H^SCK aerosol concentration. The technique consists
of reacting the H0SO. with excess ammonia and converting the unreacted ammonia
^ ^ „ - -" . _^JL-_— "
to NO gas in a gold or copper oxide converter. Sensitivity of this method
2
for EUSO, aerosol is said to be approximately 20 vg/m . Background levels of
NO and N0_ have been shown to have an effect.
Commercial chemiluminescence instruments could probably be utilized with
minimal alterations, to monitor H2SO, aerosol, if an H2SO^ adduct could be
developed that would decompose or volatilize to give off NO or VO^'
107 Spark Replica Technique
Stickney and Quen (20) have developed a spark replica technique for
measurement of sulfuric acid natlei. This technique involves exposure of a
polycarbonate film to an acid aerosol and counting the number of defects or
holes developed in the film with a spark replica counter. Sparking is induced
through the use of the film as the dielectric material of a capacitor, and
impressing a voltage across this capacitor. The concentration of nuclei is
an exponential function of the number of counts (holes) recorded. The minimum
4 3
detectable concentration is 10 nuclei/cm . The specificity has not yet been
determined.
159
-------
Analytical Methods Conclusions
Each of the analytical techniques discussed has inherent advantages
and disadvantages for analyzing sulfuric acid adducts. Some of the
techniques discussed are fairly selective but do not exhibit the necessary
sensitivity, while for others, the converse is true. The final choice in
selecting an analysis method will depend upon the type of H2SO^ adduct to be
analyzed. Criteria such as the specificity of the adduct formation and the
chemical and physical characteristics of the l^SO^ adduct will have to be
evaluated with each potential analysis method in order to determine which
combination is best. The optimum choice will be a trade off between selective
analyzability and adduct sensitivity. The extent of this trade off will de-
pend on the nature of the particular sulfuric acid adduct under examination.
Theoretically, it appears that the sensitive, sulfur specific flame photometric
detector has the best specifity to sensitivity ratio for sulfuric acid adduct
measurement.
REFERENCES - APPENDIX B
1. Gehard, E. and F. Johnstone, "Microdetermination of Sulfuric Acid
Aerosol," Anal. Chem., 27, 702 (1955).
2. Lodge, J. , J. Ferfueson and B. Havlik, "Analysis of Micron-Sized
Particles: Determination of Sulfuric Acid Aerosol," Anal. Chem.,
32, 1206 (1960).
3. Lewis, W. M., "Determination of Sulfate and Chloride in Water by
Direct Titration, Using Diphenylcarbazone as Indicator," Proc.
Soc. Water Treat. Exam., 16, 287 (1967).
4. Ceausescue, D. and M. Asteleanu, Hidroteh Geospodaririea Apelor
Meteorol (Bucharest), 11, 278 (1966); CA, 66, 5658K (1967).
5. Basargin, N. N. and A. A. Nogina, "The Determiantion of Sulfates in
Natural and Boiler Feed Water in the Presence of Phosphates by
the Direct Titration with Barium Salts by using Nitrochromeazo
as Indicator," Zh. Anal. Khim., _22, 394 (1967).
6. Effenberger, M., Fortschr. Wasserchem. Ihrer Grenzgeb., JL, 173 (1964);
CA 67_, 36312 a (1967). ~
7. Lazrus, A., E. Lorgane and J. P. Lodge, Jr., "Automated Microanalyses
for Total Inorganic Fixed Nitrogen and for Sulfate Ions in
Water," Advan. Chem., 73, 164 (1968).
160
-------
8. Scaringelli, F. P. and K. A. Rehme, "Determination of Atmospheric
Concentrations of Sulfuric Acid Aerosol by Spectrometry,
Coulometry and Flame Photometry," Anal. Chem.. 41. 707 (1969).
9. Willard, Richard L., Preprint, Natioaal Meeting Division Water, Air
and Water Chemistry, Am. Chem. Soc. (1973).
iO. Butts, W. C. and W. T. Rainey, "Gas Chromatography and Mass Sepctro-
metry of the Trimethylsilyl Derivatives of Inorganic Ions,"
Anal. Chem., 43_ (4), April (1971).
11. Campbell, R. H. and B. J. Gudzinowicz, "Separation of Some Fluoro-
carbon and Sulfur-Fluoride Compounds by Gas-Liquid Chromatography,"
Anal. Chem., 33_, 842 (1961).
12. Dresdner, R. D., T. M. Reed, T. E. Taylor and J. A. Young, "Six and
Twelve Carbon Fluorocarbon Derivatives of Sulfur Hexafluorides,"
J. Org. Chem., Z5, 1464 (1960).
13. Dresdner, R. D. and J. A. Young, "Some New Sulfur-bearing Fluorocarbon
Derivatives," J. Am. Chem. Soc., 81^, 574-577 (1959).
14. Davis, A., A Roadi, J. Michalovic and H. Josphe, "Applications of Gas
Chromatography to Phosphorus-Containing Compounds," J. Gas
Chromatog.. lt 23 (1963).
15. Panek, K. and K. Murda, "Continuous Measurement of the Radioactivity
in the Separation of Compounds Labeled with ^°S by Gas-Liquid
Chromatography," Radiokyhimiya, _7_, 246 (1965).
16. Gosink, T. A., "Gas Chromatography in Environmental Analysis,"
Envrion. Sci. and Tech., 9^ (7), July (1975).
17. Robert, N. C., M. Oberlin and J. Mering, "Lamellar Reactions in
Graphitizable Carbons," Chem. Phys. Carbon, 10, 141 (1973).
18. Stein, H. P. and C. D. Hollowell, "Measurement of Atmospheric Sulfate,"
NTIS, LBL-2162, U.S. Department of Commerce, Springfield, VA.
19. Dzubay, T. G., H. L. Rook and R. K. Stevens, Analytical Methods Applied
to Air Pollution Measurements, ed. by Robert K. Stevens and William
F. Herget, Ann Arbor Science, Ann Arbor, Michigan, 71 (1974).
20. Stickney, J. and J. Quon, "Spark Replica Technique for Measurement of
Sulfuric Acid Nuclei," Environmental Sci. and Tech., 5_, 1211
(1971).
161
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-168
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
EVALUATION OF METHODOLOGY AND PROTOTYPE TO MEASURE
ATMOSPHERIC SULFURIC ACID
Final Report
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R. E. Snyder, M. E. Tonkin, and A. M. McKissick
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Atlantic Research Corporation
5390 Cherokee Avenue
Alexandria, VA 22314
10. PROGRAM ELEMENT NO.
1AA601B CA-Q36 (FY-79)
11. CONTRACT/GRANT NO.
68-02-2467
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP, NC
Office of Research and Development
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 9/76-11/78
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
Interim Report: EPA-600/2-77-010, January 1977
16. ABSTRACT
A method has been developed to selectively assay ambient sulfuric acid
aerosol. The method utilizes the vapor of a volatile amine (or amine
derivative) to neutralize the acid as it is collected on a Teflon filter.
The amine sulfate thus formed is thermally decomposed at 200°C to release
sulfur dioxide, or other stoichiometrically related sulfur containing
adducts, which are measured by a flame photometric detector. Immediate
chemical fixation of the acid lessens the chance of side reactions with
other substances on the filter, while the relatively low decomposition
temperatures aids in eliminating other sulfur-containing particles.
Sulfur dioxide is also collected as a reagent complex on the filter,
but is selectively removed by heating at 100°C.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
*Prototypes
COSATI Field/Group
*Air pollution
*Sulfuric acid
*Aerosol
Sulfur dioxide
Collecting methods
Chemical analysis
*Flame photometry
13B
07B
07D
14B
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)'
UNCLASSIFIED
21. NO. OF PAGES
172
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
162
------- |