EPA-600/2-77-010
January 1977
Environmental Protection Technology Series
                EVALUATION  OF  METHODOLOGY AND
                             PROTOTYPE  TO  MEASURE
                        ATMOSPHERIC  SULFURIC ACID
                                  Environmental Sciences Research Laboratory
                                      Office of Research and Development
                                      U.S. Environmental Protection Agency
                                Research Triangle Park, North Carolina 27711

-------
                 RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection Agency,  have been grouped into five series. These  five broad
 categories were established to facilitate further development and application of
 environmental technology. Elimination of traditional grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

 This report  has  been  assigned  to the ENVIRONMENTAL PROTECTION
 TECHNOLOGY series. This series describes research performed to develop and
 demonstrate instrumentation, equipment, and methodology to repair or prevent
 environmental degradation from point and non-point sources of pollution. This
 work provides  the new  or improved technology required for the control  and
 treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                               EPA-600/2-77-010
                               January 1977
  EVALUATION OF METHODOLOGY AND PROTOTYPE
   TO MEASURE ATMOSPHERIC SULFURIC ACID
                     by

                R.  E. Snyder
                 T. J. Reed
              A. M. McKissick
       Atlantic Research Corporation
        Alexandria, Virginia 22314
                 68-02-2247
              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.

-------
                                  ABSTRACT
     The objective of this research was the development of methodology
                                                                  3
to selectively assay sulfuric acid in the range of 0.25 to 50 yg/m .
Research was oriented toward identifying reagents with which sulfuric
acid would react to form sulfur bearing adducts that could be analyzed
in the presence of interfering sulfate.
     A literature search was undertaken to determine the reactions of
sulfuric acid which  might be useful for fixation.  Theoretical considera-
tions first led to the possibility of using a selective H-SO, solvent
in which sulfate salts were insoluble to collect the acid.  Isopropanol
and 1,2-dimethyoxyethane were examined and found to be unsuitable because
of the relatively high solubility of ammonium sulfate and hydrogen sulfate
salts.  Selective solubility was rejected as being a method with a low
probability of success.
     Interest was then focused on volatile amino compounds which could
react rapidly with sulfuric acid aerosol in the gas phase to form amine
sulfate and bisulfate salts.  One such compound, diethylamine, was found
to form an adduct which could be decomposed to release SO  at 200°C,
thus eliminating interference from inorganic sulfate salts.  Related
compounds (hydroxylamines and oximes) were also found to form sulfate salts,
which could be selectively decomposed.  A sulfuric acid aerosol generator
was constructed, and a sample probe was designed, which mixed the aerosol
with one of three fixing reagents:  diethylamine (DEA), diethylhydroxyl-
amine (DEHA), acetaldoxime (AAO).  The adduct thus formed was collected
on a Millipore Mitex Filter.
     A sample cell with Teflon-coated surfaces was constructed in which
the fixed acid samples were heated at 200°C to liberate SO-.  A valve
arrangement connected the cell to a flame photometric detector (FPD),
which measured the evolved SO-.  It was established with this apparatus
that the fixation process was very rapid and essentially complete after
a few seconds.  The thermal stability of the samples was raised by the
                                     iii

-------
fixing process, so that the filter could be dried by heating at 100°C
without loss of sample.
     Interference was eliminated from ammonium sulfate and other sulfate
salts by decomposing the adduct below the decomposition temperatures
of these substances.  Sulfur dioxide was collected as a reagent-complex
on the filter, but was selectively removed by heating at 100°C.  Suppression
of side-reactions on the collection surface by the fixation process has
not been experimentally demonstrated, but is likely from theoretical consid-
erations.
     This report was submitted in fulfillment of Contract No.  68-02-2247
by Atlantic Research Corporation under the sponsorhip of the U. S.  Environ-
mental Protection Agency.   The report covers the period June,  1975, through
June, 1976, and work was completed June'30, 1976.

-------
                                  CONTENTS

Abstract	iii
Figures	  vi
Tables	vii

   1.  Introduction	1
   2.  Conclusions	3
   3.  Recommendations 	  5
   4.  Technical Discussion	6
           Physical description	6
           The measurement problem 	  7
           Summary of developed method 	  9
           Theoretical development of method 	  10
   5.  Experimental	54
           Apparatus	54
           Fixing reagents 	  63
           Procedures and results	65
           Method evaluation 	  70
           Interferences eliminated. 	  76
           Summary of achievements 	  81
                                     v

-------
                                 FIGURES
Number                                                                 Page
  1    Bubbling System	   30
  2    Impactor with Rotating Drum	   32
  3    Filter System With Spray Jets	   32
  4    Absorbance vs. Sulfate Concentration 	   34
  5    H2SOA pilter Collection System (Gas or Vapor Reagent)	   40
  6    Dual Collection/Fixation System	   55
  7    Analysis Instrumentation 	   56
  8    Early FPD Recordings from Samples	   58
  9    Effect of Injecting Progressively Larger S0» Volumes 	   59
 10    Quantitative FPD Response to Injected SO-	   61
 11    Quantitative Response of FPD to SO-	   62
 12    Composite DTA of Potential Interferences 	   64
 13    (PDA)2S04 and Py2S04 DTA	   66
 14    Completeness of Fixation by AAO	   69
 15    Completeness of Decomposition of AAO/HoSO,  after
       5 Min (A) and 15 Min (B)	   71
 16    Typical FPD Trace from AAO-Treated Samples 	  ,   72
 17    AAO-Fixed Filter Portions after Heating 5 Min at 200°C ....   74
 18    AAO-Treated Sample, Different Sampling Volumes, 5 Min at 200°C   75
 19    Pairs of AAO-Fixed Samples Collected Simultaneously, 	   77
 20    Continuous Heating Procedure -FPD Recordings from
       H-SO, and Adduct Samples	   80
                                     vi

-------
                                   TABLES
Number                                                                  Page
  1    Variation of Sulfuric Acid Content of Atmospheric Aerosol. . .     8
  2    General Reactions of Sulfuric Acid 	    15
  3    Chloranilate Absorbance of Sulfate Standards 	    33
  4    Chloranilate Absorbance of Successive (NH/)2SO,
       Wash Solutions	    35
  5    Chloranilate Absorbance of Successive (NH,),?SO,/
       1,2-Dimethyoxyethane Wash Solutions	    36
  6    Sensitivity Comparison of Sulfur-Gas Measurement Methods ...    52
  7    Temperature of First FPD Signal from Adducts 	    67
  8    Peak Heights of Sample Pairs Collected Simultaneously	    76
  9    Temperature (°C) of First FPD Response from Sulfur
       Gas Reagent Complexes	    79
                                    vii

-------
                                 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.  Pricket'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 H^SO, 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 S0~ to H~SO,  (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 E^SO, 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 H^SO,  by reaction with other
particulates,  and interferences from various pollutants.

-------
                                 SECTION 2
                                CONCLUSIONS
     Gas-phase fixation of sulfuric acid aerosol by a volatile amine or
amine-derivative is a convenient  and effective method of stabilizing
an atmospheric sample.  By choosing an appropriate fixing reagent, it is
possible to form an adduct which can be selectively decomposed to release
S02 at 200°C.  Three compounds which seem to meet this requirement are
diethylamine, diethylhydroxylamine, and acetaldoxime.
     The evolved S02 can be measured with either a flame photometric
detector (FPD), or a West-Gaeke Bubbler.  The FPD is preferred, because
the higher sensitivity allows for collection of a smaller sample with a
correspondingly lower possiblity of interfering collection-surface reactions.
Sampling volume of a few cubic meters or less are called for, to prevent
collection of an overlarge sample, which will saturate the detector.  To
be quantitatively measured, 0.01 to 1.0 yg of H,,SO, must be on the filter
as a fixed adduct.
     Instrumentation for this method requires only minor modification of
existing devices.  Any filtration or impaction device can be used, if
provision is made for addition of the gaseous reagent to the sample air
stream.  The analysis instrumentation is essentially the same as devices
which have already been developed for direct volatilization of sulfuric
acid.  The uniqueness of the proposed method arises from its use of a
gas-phase fixing reaction, and its choice of reagents which offer a useful
compromise between adduct stability and selective analyzability.

-------
                                SECTION 3
                             RECOMMENDATIONS
     As a result of  the  past year's effort at Atlantic Research Corporation
under Environmental  Protection Agency Contract No. 68-02-2247, the  following
recommendations for  further work are made.
     1)  Experimental data is required on the ability of  the developed
         fixing reactions  to prevent interfering reactions with common
         atmospheric particulates, such as CaCO« and Fe20.j.
     2)  The interference  of representative  ambient levels of ammonia must
         be determined.
     3)  Heating the filter to remove water  during sampling should  be
         evaluated as a  method for further stabilizing the fixed  acid
         sample.
     4)  More background work, both theoretical and experimental,  is needed
         on the chemical properties of the adducts, specifically  their
         formation under sampling conditions, chemical stability,  and
         thermal decomposition.
     5)  New compounds of  the same categories  (amines and amine-derivatives)
         should be evaluated for the desired properties.
     6)  Re-examination  of sulfur gas interference at ambient concentration
         levels should be  carried out, with  special attention to  the
         possible oxidation of S02 under  sampling conditions.
     7)  Effects of  the  analysis system on the sample  (i.e., adsorption
         on of reaction  with reactive surfaces) must be  determined and,
         if possible,  eliminated.
     8)  Generation  of a reproducible aerosol, which has proved  impossible
         with the Thomas generator, may be attemtped with a pneumatic or
         ultrasonic  generator.
     9)   An estimate of  collection efficiency must be made in  order to
                          3
         calculate the yg/m  value of the FPD response.

-------
10)  A reliable method of calibrating the FPD needs to be demonstrated,
     either by injecting S02 or preparing known H2SO,  samples.
11)  Precision and accuracy of the proposed methodology must be
     established.
12)  Ambient samples need to be collected in areas prone to have
     H^SO, aerosol to validate field use of the methodology.

-------
                                 SECTION 4
                            TECHNICAL DISCUSSION
PHYSICAL DESCRIPTION
     Concentrations of sulfuric acid aerosol in the atmosphere generally
range from 2.4 to 48.7 yg/m3 (8).  Toxicity is most directly related to
three parameters of the aerosol:
              • Droplet size distribution.
              • Concentration of acid in the droplet.
              • Concentration of droplets in the atmosphere.
                                                         3
All three of  these values can be calculated from the yg/m  measurement, if
temperature,  humidity, and other factors are known.  The average particle
diameter of sulfuric acid aerosols has been shown to be 0.35 y (9), but
varies greatly with humidity.  Considerations of droplet size and acid
concentration are also important in the development of analytical methods.
     Several  theoretical and experimental treatments of the aerosol nuclea-
tion and growth phenomenon have been published (10-14).  Takahashi, et al.
(15) proposed a three-step mechanism for formation of the acid aerosol
by photo-oxidation of SO-:
         Photo-oxidation of SO  followed by rapid combination with a
                              JL
         water molecule,
                                          *• H2S04 (vapor)                   (1)
        Nucleation  to  a  critical  size  by combination with several water
        and  sulfuric acid vapor molecules,

        H2S04(vapor) + H20(vapor) nucleati(m'  (H^jH^(embryo)      (2)

        Growth of the  embryo to a large  aerosol particle through combina-
        tion with additional water,  sulfuric acid and other molecules,

                                    6

-------
                               H2S°4> H?°' Particulates
         (H0SO.) (H00) (embryo)	>• aerosol particle  (3)
           *•  ** n  /  m

The photo-oxidation 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 1 shows the sulfuric
acid content of an individual aerosol droplet as a function of particle
radius and relative humidity.  This data is the result of theoretical
calculations (16); however, evidence indicates that the concentrations are
representative  of the real situation.
     Inspection of Table 1 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 = 1.0 to 0.1 ym,
the average concentration ranges between 45 and 60% for 50% relative
humidity, and between 33 and 54% for 75% relative humidity.  These concen-
tration ranges will be encountered under normal conditions.  However,
in areas of high humidity and inversion conditions, the concentration
per particle can drop to 18% or below.  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 MEASUREMENT PROBLEM
     The hazards that result from sulfates in ambient air make it imperative
that a reliable method for their measurement in this complex aerosol be
established.  It is evident from the foregoing discussion that this is not
an easy task.  A reliable method must collect and measure sub-micron
sized particles containing from 90% to less than 18% sulfuric acid in water.

-------
            TABLE 1.  VARIATION OF SULFURIC ACID CONTENT
                      OF ATMOSPHERIC  ACROSOLS  (16)
Particle
 radius
   yim

  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
Mole fraction
__of_J0S04__

     0.04
     0.086
     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
of
  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

-------
The measurement range of 0.25 to 50 yg/m3 has been established as a goal
in this work.
     Even more important is the requirement that the method distinguish
between sulfuric acid and similar compounds in the atmospheric sample,
including sulfate salts  [NH.HSO,, (NH.) SO , PbSO,, CaSO,, etc.], sulfur
gases  (S02, H2S, COS, RSH), other acid aerosols (HC1, HN03> H3P04>, and
a wide variety of other organic and inorganic particulates.  In addition,
since the aerosol must be concentrated by any sampling method that is used,
it is necessary to stabilize the acid in some manner to prevent it from
reacting with other substances on the collection surface.
     No analytical method published to date has succeeded in eliminating
all of these interferences simultaneously.  The method discussed here and
supported by research during the past twelve months shows great promise
toward solving this problem.
SUMMARY OF DEVELOPED METHOD
     The purpose of the sulfuric acid program was to develop methodology
and instrumentation for the selective analysis of sulfuric acid aerosol
                                 o
in the range from 0.25 to 50 yg/m .  This has been accomplished by the
method summarized below:
     1)  An atmospheric sample is first drawn into a glass tube and then
         through a Millipore Mitex 5y filter by a vacuum pump.  Simultane-
         ously, 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 200°C in a closed
         cell with non-reactive surfaces.  After five minutes, an inert
         carrier gas flushes the sulfur dioxide evolved from the sample
         into a flame photometric detector (FPD).  The peak height of the
         FPD recording is a direct measure of the amount of acid originally
         collected, and may be calibrated by injecting known S0? solutions
         or by preparing known I^SO, standards.

-------
     3)  Only compounds containing sulfur are detected by the FPD; others
         cannot interfere directly.  Sulfate salts, including
         and NH.HSO, , do not interfere because they do not decompose
         significantly at 200°C.  Sulfur gases, primarily S0~, 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 substances on the filter.
THEORETICAL DEVELOPMENT OF METHOD
     By far the most difficult problem in the development of a method for
measuring sulfuric acid aerosol has been the many potential interferences.
In fact, this problem has largely determined the direction of research
from the inception of the^program.  Therefore, it is first necessary to
discuss in some detail the types of potential interferences and the
general requirements to eliminate them.
     Interfering substances may be divided into two practical categories :
those which artificially enhance the measured acid value; and those which
diminish it.  To the former we have assigned the term "positive interfer-
ences," and to the latter "negative interferences."
Positive Interferences
     Positive interferences usually result from the failure of an analytical
method to distinguish between the desired compound and one which is chemi-
cally similar.  This is particularly a problem with sulfuric acid measure-
ment, 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  (17).
     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 ammonia and sulfuric acid, the extent of which seems to
be related to climatic conditions  (18) .  In aqueous solution, both  sulfate
and acidic protons (from hydrolysis of NH, ) are present, making it difficult

                                     10

-------
to distinguish from dilute sulfuric acid.  Ammonium sulfate and bisulfate
particulates in the atmosphere are generally associated with some moisture,
further enhancing the interference by hydrolysis which has already occurred.
Some success has been achieved in separating H«SO, from the ammonium
sulfates by volatilization, since they are stable to about 235°C.
     Many other sulfate salts may be present, some soluble (Na^SO.) and
some insoluble (CaSO,, PbSO,).  These are generally less of a problem
then 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
SO- to sulfuric acid on certain filter media, particularly glass fiber
filters (19).  Other atmospheric particulates collected during sampling
may also catalyze this reaction (20-22).  If a fixing reagent is used to
stabilize the acid during collection, as in this method, it is necessary
to evaluate the potential interference of these gases as a result of
interaction with the fixing reagent.
     Also of concern are other strong mineral acids (HC1, HNO«, and H_PO.)
which may be present in the atmospheric sample (23-24).  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. In this method, fixation of sulfuric acid by a basic reagent
also serves to stabilize other acids and reduces the probability of these
reactions.     <
     Finally, a wide variety of substances may interfere with widely used
wet chemical methods of sulfate analysis.  Most of these methods are
                               _2
based upon precipitation of SO,   , usually as BaSO,; thus, other anions
                    +2                   -3-2
which precipitate Ba   may interfere (PO,  , S  ).  Some metal cations may
                                                          +2+2
cause a negative interference by precipitating sulfate  (Pb  , Ca  ), and
other species may interfere by complexing indicator reagents.  The complex
nature of urban aerosols has been dramatized by studies which have found
                                     11

-------
over 50 elements present  in measurable quantities (25).   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
salt particulates produce sulfate salts, which must necessarily be excluded
from the analytical measurement if positive 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.  Once again, the product ammonium
sulfates must necessarily 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.  The volatilization technique for separating the acid from its
salts fails to  prevent this sample loss; in fact, the loss is 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 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  prevent negative interferences, the method must be highly sensitive
(so that only a small sample need be collected),  and it must employ a
process for chemically stabilizing the acid  as it is collected.  Of course,

                                     12

-------
the adduct thus formed must be selectively analyzable.
     It is also evident that every aspect of this problem is intimately
related to every other aspect.  The goal was to develop a combination of
fixation, sampling, and analysis techniques which, working together,
eliminated all interferences.  No one of these three aspects could be
regarded as separate from the other two, since a decision in any one area
would limit possibilites for the others.  The remaining discussion on
theoretical development of the proposed method is divided into three
sections:  fixing reactions, sampling techniques, and sample analysis; but
reference must continually be made to all three, as was necessary during
the actual development of the method.
HpSO, Fixing Reactions
     At  the beginning of the program, the following general requirements
were established for the fixing reagent:
         1)  React between 0 and 100°C.
         2)  Reagent must be chemically stable under sampling conditions.
         3)  Reagent must react with sulfuric acid to render it in a
             chemically stable and non-volatile configuration under
             sampling conditions.
         4)  Kinetics and thermodynamics of the reaction must be such
             that sulfuric acid is fixed quantitatively, immediately,
             and nonreversibly  under sampling conditions.
         5)  Preferably, the reagent should react only with sulfuric acid.
             In reality, this may not be possible.  In this case, the
             sulfuric reaction product must be separable in some manner
             from other reaction products.  Possible recognized inter-
             ferences, depending upon reagent, are SO, , SO.,, SO , H^S,
             organic sulfur compounds, acids, particulates, and possibly
             nitrogenous compounds.
         6)  Reagent must come in physical contact with the sulfuric
             aerosol.
         7)  KcMgent must be present in sufficient quantities to assure
             that all the sulfuric acid will be fixed.
                                     13

-------
         8)   Reagent  should  be  non-toxic and non-carcinogenic.
         9)   Reagent  should  be  relatively inexpensive.
        10)   Reagent  must  form  an I^SO.-adduct which is demonstrably less
             reactive toward common atmospheric particulates on the filter
             than unfixed  H~SO..
With regard to finding an  appropriate fixing reaction, a general survey of
the reactions of sulfuric  acid  was undertaken.  Table 2 is an outline
of these reactions, as discussed  in the following sections.
Oxidation/Condensation
     Sulfuric acid is frequently  used in organic reactions where a non-
specific catalyst is required.   In general,  these reactions are neither
specific for sulfuric acid nor  quantitative  for the amount of acid present.
Feigl  (26), however, developed  a  fairly selective test for sulfuric acid
which makes use of the fact that  this acid is both an oxidizing agent
and a  dehydrating agent.
     Methylenedisalicylic acid  reacts with H0SO. at 150°C in a combined
                                            2  4
oxidation-dehydration reaction  to form the red quinoidal formaurindicar-
boxylic acid:
                  1100C
                             CH,
                                                    ILSO.  	
                                                     2  *   150°
                                        eon
                                                                          (11)
                             CH
                          (red)
                                           — 0 + 2!LO + SO
                                     14

-------
                TABLE 2.  GENERAL REACTIONS OF SULFURIC ACID

I.        Oxidation/Dehydration
                                         H2SO
               Methylenedisalicylic acid — - - > colored product          (4)
                                      H2SO
               p-Hydroxybenzaldehyde - - — — > Colored product             (5)

II.       Sulfonation

     A.   Electrophillc Aromatic Substitution
               Phenol + H2SO, - >• 0 — I- p-hydroxybenzenesulfonic acid     (6)
     B.   Formation of Sulfate Esters
                                              + H0                       (7)
                  OH                    OSO-jH
               CH2=CHCH3 + H2S04 - »• CH3CHCH3                            (8)
                                         OSO^H

III.      Precipitation of Insoluble Sulfate
               BaCl2 + H2SO, - >• 2HC1 + BaSO^-l-                           (9)
             2PDA-Br + HS0  - >• 2HBr +  (PDA)SO                        (10)
IV.       Acid Base Neutralization/Salt Formation
     A.   Formulation of Inorganic Salts
               1.  Ac id -metal reactions
               2.  Acid-salt reactions
     B.   Formation of Organic Salts
               1.  Oxouium salts
               2.  Amine and amine-derivative salts
                                     15

-------
Our goal in studying this reaction was not to fix the acid itself, but
to form a stable, stoichiometrically related derivative, which in this
case could be measured spectrophotometrically.  A minimum of  2.5 yg H2SO^
can be detected by this method.  Feigl (26) also found that p-hydroxy-
benzaldehyde undergoes a similar reaction, with formation of  a green-
colored quinoidal compound.
     Experiments were performed with p-hydroxybenzaldehyde in ethylene
glycol  to  determine selectivity of this method for sulfuric acid.   First,
3 drops of 20%  sulfuric  acid in water were added to  25 ml of  ethylene
 glycol  in  which 0.5 g of p-hydroxybenz aldehyde had been  dissolved.   The
mixture was slowly heated  to boiling, and a green color  rapidly  appeared
 when the  temperature  reached vL60°C.  When phosphoric or hydrochloric
 acids were used instead  of H_SO,, no  color developed.  Nitric acid produced
 a red color at a lower  temperature  (1QO°C) , but in combination with
 phosphoric acid, a green color appeared.  Undried ammonium sulfate crystals
 also produced a green color when added  to  the solution and heated.
      Since this reaction utilizes the combined oxidation/dehydration
 properties of sulfuric acid,  it is not  surprising that the combination of
 an oxidizing acid (HN03> and  a dehydrating acid  (H PO,)  gives the same
 product.  Considering the immense variety of  substances  in the atmospheric
  sample, it is not unlikely that two compounds with  these properties will
  be collected together.  This  finding leads to a  generalization regarding
  the fixing reaction and the specificity of the analytical method.  The
  adduct which is formed  during sampling and measured during  analysis must
  incorporate at  least part of the-sulfuric acid molecule.   It  is impossible
  to form a unique sulfuric acid adduct unless this is done,  because
  the reactions  of sulfuric acid in themselves are not sufficiently distinct
  from those of  other  acids.
      Another problem arose from  the observation that ammonium sulfate
  also gave a green color under test conditions.  There was some question
 whether this was due to a direct action of ammonium sulfate on the reagent,
  or to  the solvent ethylene glycol dissolving a significant amount  of salt/
  Subsequent work (see page 37) indicated that the second factor may be
 most important.
                                       16

-------
     Oxidation-condensation reactions do not appear to be feasible for
fixation of atmospheric H-SO,.  The reaction is not sufficiently specific
for sulfuric acid to form a unique adduct which could not be duplicated
by other substances.  Even if this interference is allowed, the problem
of preventing sulfate salts from dissolving in the collection medium
remains.  This involves a selective solvent approach, which is discussed
in detail on page 37.
Sulfonation of Organic Compounds —
     This category of reactions includes all those in which sulfur from
sulfuric acid becomes covalently bonded to an organic group.
     Electrophlic 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:

                                          k2          -    +
                                       	^-* C6H5S03  + H               (12)


As with all electrophilic aromatic substitutions, the sulfonation position
and the kinetics of the reaction are controlled by other substituents
on the ring.  With electron releasing substituents, such as -NH2» -OH,
-OCH3, -NHCOCH3> ~C6H5> and -CH3, the rate of substitution is increased
over that of benzene due to increased stabilization of the carbonium
ion intermediate.  The substitution occurs in either the ortho or the
para position.  With electron withdrawing groups, the rate of substitu-
tion is slower than with benzene, and the substitution takes place in the
meta position.  Thus, from a kinetic point of view, it would be more
desirable to use a substituted aromatic, such as phenol, in preference
to benzene.
     Sulfonation reactions differ from other electrophilic aromatic
substitution reactions in that they are reversible and k-1 % k«.  The
reaction rate for sulfonation is inversely proportional to the square of
thp wnfer rnnrentration.  Thus, the attacking species, S0~, must be in
high concentration in order to drive the reaction toward the formation
of the sulfonic acid.  Concentrations of S03 sufficient for the  sulfonation
                                     17

-------
/   \    H2S04, 0
to occur are found only in concentrated sulfuric acid (98%)  or oleum
solutions.
     At 0°C, phenol reacts with sulfuric acid to give a mixture of ortho
and parahydroxybenzenesulfonic acids (27):

                                          ortho- and parahydroxy-
                                          benzenesulfonic acid

Of course, sampling cannot be carried out at 0° if the acid has to be
heated to achieve the necessary concentration.  It was found that at room
temperature or above, phenol is rapidly oxidized by sulfuric acid to a
bright red quinoidal compound.  Sulfuric acid is very likely reduced to
SOy and lost in the process, as was seen in comparable reactions in the
preceding section.  Whatever sulfonation does occur is far from uniform,
existing as many sulfonated isomers of phenol and its oxidation products.
Resorcinol  (m-hydroxyphenol) gave comparable results.  In both cases,
the color change was also affected by nitric acid, indicating that oxida-
tion is involved, rather than sulfonation.  Aniline, another highly
activated ring, is also readily oxidized to a variety of products.
     Unfortunately, it appears that most highly activated rings are also
most easily oxidized by sulfuric acid.  Less reactive compounds, such as
benzene and toluene, generally require refluxing with concentrated H_SO,
to achieve sulfonation.  Such reactions do not appear suitable for the
purpose of fixing sulfuric acid, since the heating that would be necessary
to concentrate the acid would also cause some to be volatilized before the
relatively slow reaction could occur.
     Formation of Sulfate Esters—Sulfuric acid reacts with alkenes and
alcohols to form alkyl hydrogen sulfate and dialkyl sulfate esters.  These
compounds are hygroscopic and readily hydrolyzed to alcohols and sulfuric
acid.  Because they are so difficult to isolate, the chief concern
with them in this program was in the theoretical mechanism by which
sulfuric acid dissolves in liquid alkenes and alcohols.  The practical
possibility of using these solvents to collect sulfuric acid and separate
it from its salts is discussed more fully on page 37.  This section
contains a discussion of the chemical reactions which occur when sulfuric
                                     18

-------
acid is added to these reagents.
     Alkenes—The mechanism by which sulfuric acid dissolves in an alkene
is postulated to be:
        (1) -C=C- + rf*HSO
                        "   H
     I  I           o       II
IT- -c-c- + :oso0rf  -^ -c-c-
                                 H                    H OS03H
                          k  » k            Alkyl Hydrogen
             M      |  |    2     111      ,     Sulfate
       (2) -c-C- + -C-C-    7-^  -C-C- + -C-C-	v
                    H OS03H        H ^      ' OS03~:^
                              II        I  I
                             -C-C-OS000-C-C-
                             Dialkyl sulfate
Product distribution is dependent upon H«0 concentration, but it was
reported  (28)  that very little of the dialkyl sulfate is formed in the
presence  of any appreciable amount of H^O.
     Polymerization and hydrolysis are major interfering side-reactions.
Acid-catalyzed and free-radical polymerization of alkenes are well-known
reactions, and there is evidence (28) that this may also result from the
decomposition  of alkyl hydrogen sulfate with time in the reaction mixture.
In addition to affecting the distribution of the two sulfate esters,
water in  the mixture may also hydrolyze the esters to an alcohol.  The
acid thus liberated may then react with more alkene, a polymer of the alkene,
or the alcohol just formed.
     Consequently, dissolution of H-SO  in an  alkene yields a very complex
mixture, with  the original acid existing in a variety of chemical forms.
This fact does not disqualify alkenes as potential fixing reagents, however,
as long as the acid can be reproducibly recovered.   A problem does appear
Lo exist, however, in the slow rate at which the reaction occurs.
     The reaction of sulfuric acid with alkenes proceeds best when

                                    19

-------
electron-donating groups are next to the double bond.  Since a carbonium
ion is formed in the rate determining step, compounds capable of forming
stable carbonium ions react faster and under milder conditions.  For example,
ethylene requires 98% sulfuric acid for the reaction to proceed, while
propylene and isobutylene required 80% and 63%, respectively.
     Concentrated sulfuric acid, however, does not dissolve immediately
when added  to isobutylene at room temperature; the mixture must be agitated
for several minutes.  Styrene, an alkene with the powerful electron-donor
phenyl group next to the double bond, does not react noticeably faster.
Heating the mixture shortens the reaction time, but it also increases
the extent  of polymerization.  Unfortunately, these relatively reactive
alkenes also undergo other reactions involving a carbonium ion intermediate
more readily, including polymerization and hydrolysis of the sulfate esters.
     These  findings cast doubt on the theoretical usefulness of alkenes
as fixing reagents.  Practical problems also arise regarding the control
of water collected from the aerosol into a water-immiscible solvent.  This
will be taken up in the section on selective solvent approach (page 37).
     Alcohols—Concentrated sulfuric acid dissolves immediately when added
to an alcohol, even if the alcohol is not completely water soluble.  This
fact made the alcohols much more attractive than alkenes as potential
collection/fixation media.  In order to dissolve an alkene, sulfuric
acid must take the difficult step of forming a carbonium ion.  Dissolution
in an alcohol is accomplished simply by protonating oxygen to form an
"oxonium salt" which is soluble in the alcohol.  If the acid concentration
is high, the alkyl hydrogen sulfate ester may be formed (29):

            i i                 i  i          e       ii
           -C-C- + H0SO,  	>- -C-C- + H0SO,  	>- -C-C- + H00             ,,,..
            II      24 •<	  j  ,      24  •<	  | |      2              (16)
              OH                ®OH2                 OS03H

Under conditions of heat and high acid concentration, the alcohol  can  also
be dehydrated to an alkene.  Normal sampling conditions, however,  should
not allow this reaction to occur.
     Thus,  at first sign, alcohols seem to satisfy many of the  requirements
for a fixing reagent.  Although the adduct (a distribution of  oxonium  salt

                                     20

-------
and sulfate ester) cannot be isolated as such, in solution the acid is
diluted and its acidity reduced immediately upon collection.  The feasi-
bility of using alcohols depends primarily upon the extent to which the
acid can be separated from interfering substances.  This question was
dealt  with experimentally, and the results are discussed on page 37.
     Another approach to the alcohol-H2SO, reaction is the sulfonation
of -OH groups in carbohydrates (30-33).  This reaction has been widely
used in carbohydrate analysis, but most previous work is not applicable
to the problem at hand.  The acid must be quite concentrated in order to
react, and the product is subject to rapid hydrolysis.  Polymerization
and dehydration of the carbohydrate have been observed as side-reactions.
Analysis of cholesterol and other sterols has also been acomplished by
similar reactions with t^SO, (34-38), but once again concentrated acid
is required.  It has been determined that even concentrated acid reacts
slowly at room temperature, while heating the mixture causes extensive
dehydration and decomposition.  These compounds do not seem to be adequate
as sulfuric acid fixing reagents.
     Summary—Of the various sulfate ester-forming reactions of sulfuric
acid, only the reaction with simple liquid alcohols seemed very promising
In this case, the adduct was not strictly a sulfate ester, but was a
distribution between it and an oxonium salt.  Rather than isolating these
adducts, it seemed most advantageous to use their solubility in excess
alcohol to separate them from potential interferences, i.e., sulfate salts.
Precipitation of Insoluble Sulfate  —
     Another manner in which to fix sulfuric acid aerosol is to bring
it into contact with a reagent which causes the sulfate to be precipitated
as an insoluble salt.  The overriding advantage of an insoluble adduct
is that the fixing reaction is not greatly affected by the dilution of the
acid.  Moreover, the reaction is virtually instantaneous for even very
dilute acid.  From the standpoint of fixation alone, adduct formation by
precipitation seemed to be ideal.
     There are, however, other considerations in the development of  the
total method.  Primarily, the adduct must be selectively analyzable.  An
                                     21

-------
insoluble adduct cannot be separated from other insoluble sulfate salts by
selective solubility.  The only possibility for selective analysis is
if the adduct can be thermally decomposed, releasing S0~ for analysis,  at
a lower temperature than these other sulfates.  Unfortunately, insoluble
sulfate salts seem to have very high decomposition temperatures, which
are not surprising, if their insolubility is taken as an indication of
high stability.  Thus, although precipitation is a highly desirable
fixing reaction, it may not be suitable for other reasons.
     The most common method for precipitating sulfate is as the highly
insoluble barium sulfate, BaSO,.  A soluble barium salt, such as the
chloride, is conveniently used for this purpose.  As previously discussed,
however, there does not appear to be any way to separate the product
BaSO, from other insoluble sulfates in the sample, since it is thermally
stable to over 1500° C.
     Much attention has been paid to the formation of insoluble amine
sulfates.  The amino group in organic compounds is a moderatley strong
base which readily reacts with mineral acids to form "ammonium salts:"
                           RNH2 + H2S04 ->  (RNH3)2S04                    (17)
                          R^-NH + H2so4 ->•  (R2NH2+)2so4=                  (18)
                           R3-N + H2S04 -v  (R3NH+)2S04=                   (19)

The reaction is virtually instantaneous, and if the product salt is insol-
uble, it precipitates out of solution.  Most organic amine sulfates are
water-soluble, but 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:
                                                          Solubility  of
     Compound                    Formula              Sulfate Salt,   g/1
Benzidine  (39)             H2N^ - (^ ^2          0.098  (25°C)

4:4'-diaminotolane (39)    N fi~\— C = C—  ~^ NH         °-059  (25°C)
                                     22

-------
                                                       Solubility of
       Compound                    Formula          Sulfate Salt,  g/1

2-amino-4'-chlorodiphenyl (40)  Civ'  y	\j=/^2        0.155 (25°C)

1,8-diaminonaphthalene (41)         f\f^\              0.222 (25 °C)
2-aminoperimidine (41) or
perimidylatnmonium bromide (42)
     Unfortunately, these compounds have the same problem as BaSO, :  their
thermal stability is too high for selective analysis.  Thomas, et al. , (43)
found that the insoluble perimidylammonium sulfate must be heated to 400°C
to cause thermal reduction.  They impregnated a filter with the soluble
bromide salt of this compound and showed that the acid was collected and
stabilized as the insoluble sulfate.  During analysis at 400°C, however,
inorganic ammonium sulfate and ammonium hydrogen sulfate interfered totally.
Theoretically, this interference could be eliminated by washing soluble
sulfates out of the sample with water; but with such trace amounts of
sample to be measured, the potential handling error makes this approach
unfeasible.
Acid-Base Neutralization/Salt Formation —
     This section discusses the general salt-forming reaction of sulfuric
acid not already mentioned under another more specific category.  The
discussion is divided into reactions with inorganic and organic reagents.
     Formation of Inorganic Salts — Sulfuric acid reacts with active metals
which lie above hydrogen in the redox potential series to yield metal
sulfates.  The following oxidation reduction reaction occurs in this case:
                                                2.                     (20)
                  in Hi Is mannpr. but the rfiart.ion rate, is highly dependent
upuu cuuceuLiaLiuu oi tae acid,  ilie uieLal collection surface would then  have
to be heated in order to complete the reaction.
                                     23

-------
     Many salts will also react with  sulfuric  acid by forming a sulfate salt
and another acid.  One of three requirements must be met before a reaction
of this  type will  go to  completion.   Either the product sulfate must be
insoluble  (page 21) , or  the product acid must  be either weaker or more
volatile than  sulfuric acid.   Each condition serves to remove one product
from  the equilibrium reaction  so  that it goes  to completion.  If the product
acid  is  more volatile than sulfuric acid but water soluble,  such as the HX
acids, it may  again be necessary  to heat the collection surface in order
to remove water and product acid, thereby  driving the reaction to completion.
      Reactions of  this type are quite promising for fixation of H^SO, , but
their practical usefulness depends upon the ability to form  an adduct
sulfate  salt which can be selectively analyzed in the presence of common
atmospheric  sulfates.  As with the insoluble sulfates previously discussed,
this, in turn, depends upon the ability to decompose the sulfate adduct at
a relatively low  temperature.  A  search of the literature  found four
sulfate  salts  which decompose  between 150°C and 200°C:  manganese, eerie,
thallium,  and  hydroxylamine sulfates.  The decomposition temperatures and
liberation of  SO 2  were verified by DTA and FPD in the manner discussed on
page  6Q.
      Manganic  sulfate is formed by the reaction of sulfuric  acid with
Mri(C2H 02) or  MnB  according to the following equations:

                      302)3 +  3H2S04  - »-  Mn^SO^ + 6HC2H  02   (44)      (21)

                       2MnB +  3HS0   - *  BH  + MN(S0)        (45)      (22)
                       2MnF3 + 3H2S04 - > Mn^SO^  + 6HF                (23)
                                                     +2
When heated,  thermal  reduction to  the  more stable MN   salt occurs with
release of  S02:

                            MN2(S04)3-i|^ 2MnS04 + S02 + C>2              (24)

                 +3
Unfortunately, Mn   salts are not  very stable,  and may decompose upon stand-
ing.  Of particular concern is the possibility  that  they may catalyze the

                                     24

-------
                       -2
oxidation of SO  to S04   during sampling.  Therefore, although the
                                      +3
decomposition temperature is ideal, Mn   salts may not be suitable fixing
reagents.
     Ceric sulfate also decomposes in the desired range:

                    2Ce(S04)2 195  > Ce2(S04)  + S02 + 0                  (25)

It is formed from the reaction of  sulfuric acid with any of several Ce
salts:
                      Ce02 + 2H2S04 - »• Ce(S04)2 + 2H20     (46)        (26)

                   Ce(OH)4 + 2H2S04 - »• Ce(S04)2 + 4H20     (47)        (27)

The Ce~™ salts are more stable than Mn   salts and will not spontaneously
decompose at room temperature.  However, eerie sulfate is very slow to
decompose at 300°C; a 100 mg sample requires several hours at this tempera-
ture to completely decompose.
     Thallium hydrogen sulfate decomposes at 120°C to the normal sulfate,
releasing S02 in the process:

                                 120°
                         2T1HS04      > T12HS04 + S02 + 02    (48)        (28)

The major problem with this adduct is insuring that excessive amounts of
the normal sulfate are not formed during sampling in preference to bisulfate.
     Hydroxylamine sulfate (considered an inorganic salt) is most conven-
iently prepared from the chloride:
                         NH2OHC1 + H2S04 - >  (NH4OH)2S04 + HC1          (29)

This sulfate decomposes rapidly at 180° (explosively if a large amount is
present).  Hydroxylamine can be either an oxidizing or reducing agent,
so there is some concern that H2S04 may be lost as S02, or S02 oxidized to
SO," .  Additionally, any decomposition to NH3 would produce  (NH^) SO  and
acid would be lost from analysis .
                                     25

-------
     Despite the problems raised, inorganic sulfate adducts hold significant
theoretical promise as fixed forms of sulfuric acid.  It seems more desir-
able to use a '.salt fixing reagent than a metal foil, because of the faster
reaction rate and lower multiplicity of products.  Nevertheless, the
practical problems of impacting collected aerosol on a salt surface, and
decomposing the adducts in the presence of excess reagent salt, led to
placing more importance on the reactions discussed in the following section.
     Organic Salt Adducts—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 in alcohols
was  discussed on page 14.  In fact, concentrated sulfuric acid dissolves
in almost any organic compound containing oxygen by the same mechanism:
carbonyl compound, ethers, etc.  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  demon-
strated by  adding concentrated H-SO, to diethylamine, which produces rapid
precipitation of a white solid.
     As discussed on page 22, 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 •
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
analysis.
     Even water soluble amines seem to be at the limit of selective
analyzability.   Diethylamine sulfate does decompose slowly at 200°C and
releases S02  more  quickly than 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.

                                      26

-------
For this reason, experiments were begun with  amine-derivatives  in which
oxygen is bonded to nitrogen  : hydroxylamines and  oximes:

                RX                              R
                     N-OH                          ^C=N-OH
                V                             R2
               Hydroxylamine                       Oxime

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 than the basic amine.
An analogous difference  is observed  in the  decomposition temperatures of
ammonium sulfate and hydroxylamine sulfate:
                    240°                                180°

Of course, many factors beside  base  strength are involved in the decomposi-
tion temperature, but  this  seems  to be a useful generalization.
     The main advantage of  these  adducts over the inorganic salts discussed
in the previous section is  that they  allow greater flexibility of sampling
techniques.  Reagents  containing  the  appropriate functional groups can be
prepared as solid,  liquid,  or gaseous compounds at ambient temperatures.
The possibility of  using a  gaseous reagent to partially stabilize the acid,
even before it reaches the  collection surface, was particularly interesting.
Huygen (49) demonstrated that diethylamine vapor is rapidly absorbed by
H?SO, on a filter,  although his purpose was not to fix the acid, but to
measure total acidity  by a  wet-chemical analysis of the amount of amine
absorbed.  By applying a more selective analytical technique, both acid
fixation and selective measurement were demonstrated in our laboratories with
this vapor-absorption  method.
Summary of H2SC>4 Fixing Reactions —
     The preceding  discussion has focused on reactions of sulfuric acid
which may be theoretically  useful as  fixing reactions.  Little experimental
                                     27

-------
data was given, since this is better discussed in the context of the actual
collection and analysis procedures used.  Some conclusions, however, were
drawn regarding the usefulness of these reactions from considerations of
theoretical specificity of adduct formation and reaction rate.
     Oxidation/dehydration and other reactions where sulfuric acid  is not
incorporated in the adduct to be measured were rejected as lacking  specifi-
city.  Sulfonation of aromatic rings was too slow, required too high a con-
centration of acid, and produced too many side-products by oxidation of
activated rings.  Sulfate ester formation was too slow for alkenes, and
polymerization was seen to be a major problem.  Alcohols showed more promise
in  this  reaction, although the adduct would probably consist of more oxonium
salt  than sulfate ester.  Precipitation of insoluble sulfate was seen as
an  ideal fixing reaction, but selective analysis of the adduct seemed to
be  practically impossible.  Formation of inorganic sulfates from metal
or  other inorganic salts was regarded as promising, if somewhat difficult
procedurally, and several adducts were suggested which may be selectively
analyzed by  thermal decomposition.  Oxonium salts, from alcohols and
other oxygen-containing organic compounds, were potentially very useful  if
inorganic sulfates could be  sufficiently separated in such solvents.
Amine and amine-derivative sulfates were given the highest priority,
particularly in the use of a gas-phase reagent to fix sulfuric acid as  it
is  collected.
Collection/Fixation Methods  - Physical State of Fixing Reagent
      The theoretical  fixing  reactions discussed in the last section are
useful only  if they  can be applied  in a practical manner.  This section
describes general techniques for collecting and fixing sulfuric acid  in
terms of the physical state  of the  fixing reagent:  liquid, solid,  and  gas.
Since most of  the reagents previously mentioned are limited in  this aspect,
the present  discussion will  serve to further illustrate the relative use-
fulness  of the various fixing reactions.
      Methods for collecting  and fixing  sulfuric acid will  be  divided into
three categories.  The first category covers bubbler and  impinger methods
in  which the acid aerosol is collected  in a liquid medium.   Second  will be
                                      28

-------
methods of filtration and impaction  in which  the acid  aerosol is brought
into contact with a solid fixing  reagent on the collection  surface.  Third
will be methods in which a gaseous fixing  reagent  is mixed  with the aerosol
prior to filtration or impaction.
Liquid Phase Reagent  —
     Selective Solubility - Theory—When sulfuric  acid aerosol is collected
into a liquid medium in which  it  is  soluble,  the acid  is immediately diluted
so  that reaction with other collected substances is unlikely.  Of course,
it  is necessary that positive  interferences from other sulfates and acids
be  avoided.  For this reason,  water  is unsuitable  as a collecting solvent,
since these interfering species would dissolve and be  indistinguishable
from the original sulfuric acid.  However, a  solvent in which sulfuric acid
is  soluble, but sulfate salts  are not, would  avoid this interference.  In
fact, there are many organic solvents which fulfill this requirement on
first examination.
     In considering solubility, it is necessary to distinguish between
concentrated sulfuric acid and dilute aqueous sulfuric acid, which is the
form in which an atmospheric aerosol exists.  Concentrated  acid is immed-
iately soluble in almost any organic solvent  containing oxygen:  alcohols,
ethers, and carbonyl compounds.   The mechanism of  this solubility is seen
as  the formation of an oxonium salt, which is soluble  in the solvent. (See
page 17.)  Similarly, concentrated acid will  dissolve  immediately in many
amino compounds, although the  amine  sulfate is often insoluble in this
solvent and precipitates out of solution.  Additionally, concentrated
sulfuric acid is soluble within a few minutes in alkenes and activated
aromatic compounds, by forming sulfate esters and  sulfonic  acids,
respectively.
     Dilute aqueous acid, however, is very similar to  H^O in its solubility
properties, even if the acid constitutes 50% of the mixture.  Like water,
dilute acid is soluble only in organic solvents which  are quite polar,
particularly low-molecular weight alcohols and carbonyl compounds.  Thus,
if  the aerosol is to dissolve  immediately  in  the collecting solvent, one of
these compounds must be used.  On the other hand,  a water-immiscible solvent
may be used, if the formation  of  an  aqueous layer  can  be avoided (which

                                     29

-------
would dissolve sulfate salts and retain much of the H2SO^).  Heating the
solvent during collection to concentrate the acid is one approach, but may
cause volatilization loss of acid and troublesome side reactions  (poly-
merization, oxidation, etc.).
     The ideal collecting solvent is one in which dilute acid  (essentially
H20) is immediately soluble, but sulfate salts are not sufficientily
soluble to interfere.  This must remain true, even after the solvent has
absorbed some moisture from the air.  A typical bubbling system utilizing
the  liquid collection medium is illustrated in Figure 1.
                Intake
                               o
                               00
                               o
                             O  CO
  Bubbler
Figure 1
                                          w
                                             Trap
                                    Bubbling system
      Fixing Reactions  in Liquid  Collection Medium — Dissolution  of  the acid
 in an appropriate solvent is  a fixing  reaction  of  sorts,  since  the high
 dilution makes  subsequent interfering  reactions unlikely.  However,  it is
 conceivable that another substance may be dissolved  in  the solvent which
 will  selectively react with the  acid,  thereby further stabilizing  the sample
 and possibly making analysis  easier.
      One example of this approach is the method described on page  14, using
 p-hydroxybenzaldehyde  in ethylene glycol, although the  reaction was not
 sufficiently specific, for H2SO^ and ammonium sulfate interfered.   An
 organic  sulfate precipitation agent dissolved in methanol, such as PDA-Br
                                       30

-------
 (page 21), would immediately fix the acid as  it  is  collected, but other
 slightly more soluble sulfates would slowly form (PDA)-SO,.
     It can be seen that the same requirements hold as  in  ordinary dissolu-
 tion of the acid; it must be immediately soluble, and other sulfate salts
must be so.highly insoluble that they do not  measurably interfere.  If
 a solvent which meets these requirements cannot  be  found,  then no fixing
 reaction in a solvent can have the necessary  selectivity for sulfuric acid.
 Figure 2 illustrates the manner in which a liquid medium might be utilized
with an impaction device, while figure 3 illustrates the use of a liquid
 reagent utilized with a filter collection device.
     Experimental Evaluation of Selective Solvents—Because of the importance
 of finding a selective solvent for use in the above techniques, experiments
were undertaken to determine the solubility of ammonium sulfate in various
 organic solvents.  The barium chloranilate method developed by Bertolacini
 and Barney (50) was used with some modifications to measure total sulfate
 in solution.  Barium chloranilate is an insoluble salt  of  the highly-colored
 chloranilic acid.  On contact with sulfate in solution, the even more
 insoluble BaSO, is precipitated and acid chloranilate is quantitatively
 released.  The UV-visible absorbance of chloranilate in solution is,  there-
 fore, a measure of the amount of sulfate originally present.
     Isopropanol—Sulfuric acid aerosol has been collected in Isopropanol
using the Greenburg-Smith Impinger and various modifications of this device.
One such method, described in the Federal Register  (51), collects sulfuric
acid aerosol and sulfur trioxide into 80% isopropanol,  allowing the sulfur
dioxide to pass on into a hydrogen peroxide collection  solution.  An obvious
modification is to use absolute alcohol rather than 80%, since this would
decrease the solubility of particulate sulfates.
     First, a standard absorbance curve was obtained from  known H^SO.
standards.  A 4 ml volume of each standard was buffered by adding 1 ml
of 0.05 M potassium hydrogen phthalate.  To this was added 5 ml of isopro-
panol and the solution mixed.  Then, 0.03 g of barium chloranilate was
added, and the solution magnetically stirred  for 15 mintues.  The mixture
was centrifueed and absorbance of the solution was  read at A    of the UV
                                                            ItlclX
                                     31

-------
Rotating Drum
  Collector
                                                            To Pump
                                               Reagent  Solution
              Figure  2.   Impactor with rotating  drum.
           Pinhole  spray jets or other method of
           spraying liquid reagent solution
                      "Filter medium
                                                            To Pump
                Sheath
           Figure 3.   Filter system with spray jets,
                                32

-------
band 310-350 nm, using a Beckman DK-1 spectrophotometer.  A blank reference
solution was prepared and further diluted so that a blank sample would have
a small positive absorbance.  Table 3 gives the results of the experiment.

           TABLE 3.  CHLORANILATE ABSORBANCE OF SULFATE STANDARDS

                          -2
                   ppm SO                   Absorbance %
                      20                        1.47
                      15                        1.04
                      10                        0.72
                       5                        0.43
                       1                        0.20
                       0.5                      0.19
                    Blank                       0.19

These results agree well with the reported sensitivity limit of the method,
about 2 ppm.  Figure 4 shows the relationship is linear.
     To determine the solubility of ammonium sulfate in isopropanol, a
quantity of this salt was added to the alcohol and let stand for 15 minutes.
Afterwards, the salt crystals were removed by centrifugation and a 5 ml
portion of this alcohol was added to 4 ml water and 1 ml phthalate solution.
Analysis was carried out by the barium chloranilate method previously
described.  Absolute isopropanol was taken directly from the reagent bottle
with no additional effort to dry it, since this would more closely approxi-
mate real sampling conditions.
     In the first experiment, 1 g of ammonium sulfate taken directly from
the reagent bottle was added to 25 ml isopropanol.  To eliminate the
effects of moisture on the crystals causing hydrolysis, a second sample
of ammonium sulfate was dried at 150°C in a dessicator oven for 3 hours.
Results are shown below.
                Sample                          A(%)
            (NH4)2S04 from bottle               0.99
            (NH4)2S04 dried                     0.83
            Blank                               0.21
                                     33

-------
u>
             1.40
             1.20
             1.00
           V

           «
           ^

           <"
0.80
            0.60
            0.40
            0.20

                                                           10
                                                                     15
                                                                -2
                                                        ppm SO4
20
                                  Figure 4.  Absorbance vs.  sulfate concentration.

-------
With  reference  to  figure  4,  this  corresponds to about 10 ppm SO,"2.
      In order to determine whether this  was a direct  dissolution of ammonium
sulfate,  or  a surface moisture  effect,  'the same ammonium sulfate crystals
were  repeatedly washed  with  isopropanol and each of  the wash solutions
was analyzed for sulfate.  If a surface moisture effect was  operating,  the
measurements should  fall  to  the blank level after several washings.  Table 4
shows the results.

              TABLE  4.  CHLORANILATE ABSORBANCE OF SUCCESSIVE
                         (NH4)2S04 ALCOHOL WASH SOLUTIONS

           Sample                                A(%)
             1                                    0.72
             2                                    0.54
             3                                    0.55
             4                                    0.50
             5                                    0.57
             6                                    0.59
           Blank                                0.43

It can be seen  that  sulfate  levels did  not decrease after the first washing,
and remained significantly above  blank.   This must, therefore,  be due to
a direct  dissolution of ammonium  sulfate in isopropanol.  From  these results,
the solubility  can be estimated at 5 to 10 ppm.   Subsequent  runs gave
comparable results.
      1,2-dimethoxyethane—An effort was made to evaluate the solubility
of ammonium  sulfate  in  1,2-dimethoxyethane.   Since its  solubility in
isopropanol  was unacceptably high, it was reasoned that the  less polar ether
group might  make a more selective solvent.
      The  procedure was  to place ammonium sulfate in a test tube, add 1,2-
dimethoxyethane, shake, and  let stand for 15 minutes.   The liquid was then
separated, filtered, and  5 ml taken for analysis.  This volume  was us.ed
in place  of  the isopropanol  in  the established chloranilate  procedure.
More  ether was  added to the  same  ammonium sulfate, and  the process repeated.
A total of three washings was carried out.

                                      35

-------
     Absorbance of all three samples was measured,  first using a blank
reference containing chloranilate but no sulfate,  then against a solution
of the original blank diluted with an ether-water  mixture.  Table 5 gives
the results.

              TABLE 5.  CHLORANILATE ABSORBANCE OF SUCCESSIVE
                        (NH,)2SO,/1,2-DIMETHOXYETHANE WASH SOLUTIONS

            Sample                              A(% at 320 mu)
Blank Reference   First Washing                      0.40
                  Second Washing                     0.52
                  Third Washing                      0.75
                  Blank                              0.03
Diluted  Reference First Washing                      1.14
                  Second Washing                     1.30
                  Third Washing                      1.56
                  Blank*                             0.75

     The results  show that  all samples were significantly higher than the
blank.   Furthermore,  the sulfate concentration appeared to increase with
successive  washings,  possibly as a result of the solvent absorbing water.
No  quantitative comparison  can be made to the isopropanol calibration curve,
                                                      _2
because  a different  solvent was used; however, the SO,   concentration
must be  over  the  sensitivity limit of the technique, at least 5 ppm.
     Discussion—The solubility behavior of a salt in a given solvent can
usually  be  best described in terms of a solubility product.  If the equili-
brium equation is CA -* C  + A , then the equilibrium expression  (with
constant [CA]  omitted) is [C ] [A ] = K  .  The solubility product, K  ,
                                       P                            p
sets a limit  on the  maximum concentration of dissolved species, regardless
of  how much excess solid remains.  For this reason, it is arbitrary to
report the  data by taking the weight of dissolved species as a percent  of
the amount  originally added.  Rather, it is meaningful to report  the  sulfate
concentration of  a solution in equilibrium with excess solid, for a given
solvent  and temperature.
                                     36

-------
     From the results given above, ammonium  sulfate  seems  to be  able  to
               _2
generate an SO,   level of about 5 ppm  in  isopropanol  at 25°C.   This  corres-
               3
ponds to 5 yg/m  or 50 yg/10 ml  (since  10  ml is  about  the  smallest collection
volume possible).  Thus, in order for the  potential  measurement  elevation
to be no more than 5%, 1,000  yg of lUSO,  would  have to be collected.
This is an unrealistic requirement when the  lower goal detection limit
            3                                 "}
is 0.25 yg/m , since it would require 4,000  m of sample air.  In most cases,
however, the amount of (NH^SC^ collected will  not  reach  the maximum value,
so the interference is quantitative.
     Barton and McAdie (52) demonstrated that hydrogen sulfate salts, as
well as ammonium sulfate, are measurably extracted by  isopropanol.
Maddalone, et al. (53), also concluded  that  isopropanol is not sufficiently
selective as a solvent for sulfuric acid.
     This supports the work accomplished here, which seems to cast grave
doubt on the feasibility of the  selective  solvent approach.  Acid sulfate
salts can also protonate basic functional  groups, in this  case to form the
soluble oxonium salt.  Apparently, the  mechanism by  which  sulfuric acid
dissolves in these solvents is available to  other acid sulfates  as well.
     It is not very likely that  another solvent  will be able to  achieve
the necessary separation.  Leahy, et_ aL. (54), have  proposed benzaldehyde
as a selective solvent, but their data  is  either ambiguous because lower
sensitivity limits are not stated, or inapplicable to  this problem because
of the sample size used.  Moreover, benzaldehyde is  not very water-soluble,
so elaborate precautions must be taken  to  prevent formation of an aqueous
layer by collected water in the  aerosol.
     As a result of these findings and  considerations, we  concluded that
the selective solvent approach is not feasible.  Any solvent which rapidly
dissolves dilute H2SO,, also dissolves  enough ammonium sulfate and hydrogen
sulfate salts to constitute an unacceptable  interference.
Solid-Phase Reagent —
     Another method for fixing H2S04 is to collect it  on a filter or impactor
surface which is coated with a solid fixing  reagent.   This method has both
advantages and limitations, but  in general seems more  useful than liquid-
phase fixation.
                                     37

-------
     Advantages and Limitations—The main advantage of this approach to
sampling over the liquid-phase method is that interfering reactions are less
likely to occur in the absence of a reaction medium.  Sulfate salts were
an interference in a solvent because trace amounts dissolved and formed
more sulfuric acid adduct.   Dry sulfate particulates, however, do not
react when collected on a dry surface of the fixing reagent, since a reac-
tion medium is lacking.  Similarly, if sulfuric acid is stabilized as dry
adduct crystals on the collection surface, there is very little tendency
to react with other dry particulates that may contact the filter.
     However, some limitations also result from the reduced reagent-acid
contact with this method.  In order to minimize "matting" on the collection
surface by the collected aerosol, it is desirable to collect as small a
sample as possible.  Additionally, it is necessary that the fixing reaction
be very rapid, to reduce the^possibility of subsequently collected particu-
lates reacting with unfixed acid.  Of course, there is still a statistical
probability  that an acid droplet may strike and react with a previously
collected particle; but a rapid fixing reaction will reduce the extent of
this interference.  Finally, the collection surface must be heated to remove
water, which would act as a medium for interfering reactions, and to speed
the fixing reactions as discussed below.  It would, of course, be necessary
to show that no acid is lost by volatilization prior to fixation under
sampling conditions.
     Solid Fixing Reagents—Regardless of the type of fixing reagent used,
the adduct must be dried (by heating) before the acid is fixed in the most
stable form  possible.  However, it is desirable that the acid be somewhat
stabilized immediately upon contact with  the reagent, even before drying
has occurred.  In this respect, the three types of solid-phase fixing reac-
tions described on page 23 can  be ranged from the most to least desirable.
     The most  desirable type of fixing reaction is one in which  an insoluble
sulfate is formed, as in the  (PDA^SO, precipitation (page  21).  Here
sulfate is fixed immediately upon  contact and is not likely  to react further,
even before drying has occurred.  Drying is  still necessary,  however, because
collected water will also allow sulfate salts to be precipitated as  the
adduct.  This  reaction may be  most desirable from  the standpoint of  fixation,

                                     38

-------
but a serious problem with selective analysis of the adduct results from
the high thermal stabilities of insoluble sulfates.  If the adduct cannot
be selectively analyzed, the fixation is worthless.
     A fixing reaction in which the product acid is weak is the second most
desirable type.  An example is the reaction of sulfuric acid with a basic
amino compound to produce a soluble amine sulfate, which is a weak acid
in aqueous solution.  With this reaction, acidity of the aerosol is reduced
immediately, since the acidic protons are held more strongly by basic
nitrogen of the fixing reagent.  The aqueous solution, however, is still
somewhat acidic, so the sample must be dried to achieve the most stable
configuration.
     The final and least desirable fixing reaction is one in which the
product acid is strong and volatile, such as an HX acid.  When sulfuric
acid reacts with a soluble chloride salt, the product acid, HC1, is strong
and completely ionized in aqueous solution.  Consequently, no reduction in
acidity occurs on contact, and the sample must be heated to remove water
and volatile HC1 before stable adduct crystals are formed.  For this reason,
if amino compounds are used as fixing reagents, it is more advantageous
to use them in the basic form than as the HC1 salts.
     Conclusions—The solid fixing reagent approach was not rejected out-
right, but offered less advantages than the gas reagent method ultimately
adopted.  Matting was the principal problem, which was minimized by collect-
ing as small a sample as possible, but could not be avoided altogether.
Of the fixing reactions, the second alternative offered more possibilites
for analysis.
Gas-Phase Reagent —
     Figure 5 illustrates the use of a gaseous fixing reagent to stabilize
the acid aerosol prior to collection on a filter.  The fixing gas is
added to the sample probe at a point in front of the filter, and excess
reagent simply passes through the filter.  This method was ultimately
adopted as most useful for collection/fixation of sulfuric acid.
                                     39

-------
                      Gas or Vapor Inlet (X)
                                                                     To Pump
                            Sheath for
                          Vapor Chamber
                  Figure 5.  H2SO,  filter collection system.
                             (Gas or vapor reagent)

     Theoretical Advantages—The gas-phase reagent method combines the
advantages of the two previously discussed methods. It allows good acid-
reagent contact without providing a medium for interfering collection  surface
reactions.  If the fixing reaction is very rapid, the acid is significantly
stabilized even before contacting the filter.  Thus, the danger of acid
reacting with pre-collected particulates is reduced.  Particulate sulfates
are collected unchanged on the filter, since there is no solvent in which
they can dissolve.  Of course, it is necessary to show that ambient sulfur
gases do not become an interference as a result of interaction with the
fixing reagent.
     Potential Gaseous Fixing Reactions—This method of fixation limits
potential fixing reagents to those which can be generated in a vapor  phase
at ambient temperatures.   Inorganic salts are immediately ruled out.  (An
aerosol powder would not provide enough reagent-acid contact prior  to filter
contact, and excess reagent could not be removed.)  Of the amino compounds

-------
which are known to form insoluble  sulfates,  none  are  sufficiently volatile
to use in this manner.  The  only potential  fixing reagents  previously
discussed, which have sufficient volatility are alkenes,  alcohols,  and
amines; but the adducts formed  from alkenes, and  alcohols are  too unstable
to be useful out of solution, and  the reaction rate may be  slow.
     Volatile amines and  amine-derivatives,  however,  seem to be ideal for
this purpose.  In the basic  form,  many of the  low molecular weight  compounds
are either gaseous or liquid with  a high vapor pressure at  room temperature.
During the first phase of fixation, the gaseous reagent dissolves in the
droplet and reduces its acid strength by forming  an aqueous solution of
the amine sulfate.  Robbins  and Cadle (55)  found  that the rate of the
analogous reaction between sulfuric acid droplets and ammonia was controlled
by the speed of product diffusion  into the  droplets.  After 4 seconds at
28°C, droplets of 0.2y and 0.9y diameters were reacted 100% and 60%,
respectively.  Huygen (49) demonstrated that 2 minutes of contact with an
air stream containing diethylamine vapor was sufficient to  react completely
with microgram quantities of sulfuric acid  on  a filter.   The kinetics of
this reaction seem to be  sufficient for rapid  fixation of the acid.
     After the first phase of fixation is complete, the adduct exists as
an aqueous solution of the amine sulfate, which is still  somewhat acidic.
To further reduce the chance of interfering  collection-surface reactions,
it is probably advantageous  to  heat the filter in order to  drive off water,
leaving the adduct as dry amine sulfate crystals.  In this  form, the
sulfuric acid is stabilized  until  analysis  in  the laboratory.
     Experimental—Several preliminary experiments were carried out in
establishing the validity of the gas-phase  fixation method.  The procedure
is much simpler than the  solid  reagent technique,  since special preparation
of the collection surface is unnecessary, and  excess reagent is automatically
removed.
     Although alkenes were not  considered very promising  fixing reagents,
the first experiments were performed with propylene gas.  The gas was
passed for several minutes through a glass  fiber  filter containing  concen-
trated H?SO,.  A variety  of  chemical changes occurred on  the filter, as
higher hydrocarbons (from polymerization of  propylene) were identified and

                                      41

-------
a series of oxidation products appeared upon standing in air for several
hours.  These results confirmed the earlier conjecture that alkenes form
a variety of products in contact with sulfuric acid, without fixing the
acid in a significantly more stable form.
     Experiments then proceeded with the volatile compound, diethylamine,
in place of propylene.  A full description of experiments and results with
this and related compounds is presented in Section 6.  In general, however,
it was demonstrated that the adducts formed rapidly (less than 15 seconds
of exposure to concentrated vapor was required), were more stable than  '
sulfuric acid, and could be selectively analyzed by thermal decomposition
at 200°C.  As a result, gas-phase fixation of I^SO^ was ultimately selected
for the proposed method.
Summary - Collection/Fixation Methods  —
      The preceding discussion on collection/fixation methods attempted to
place the theoretical fixing reactions described under H^SO, Fixing Reactions
 (page 13) in a practical framework.  On this basis, several were eliminated
as impractical.  As a net result of the experiments, it was concluded that
          »
dilute sulfuric acid cannot be separated from ammonium sulfate in the low
microgram range by selective solubility.  This ruled out any attempt to
collect the acid directly into a liquid medium via a bubbler or impinger
system.  Since sulfate  esters and oxonium salts are not stable out of
solution, this finding  also ruled out alkenes, alcohols, and other oxygen-
containing organic compounds as fixing reagents.  Thus, from the original
survey of potential fixing reactions, only those which result in the forma-
tion  of stable sulfate  salts  (by precipitation or acid base reaction)
are still viable.
      Collection of the  acid aerosol onto a solid fixing reagent is still
feasible, provided one  of the three criteria discussed on page 37 apply:
the product sulfate is  insoluble, the product acid  is weak, or the product
acid  is volatile.  In any case, heating is required to dry  the collection
surface in order to minimize interfering reactions.  Potential problems
with  this method include incomplete fixation  (from  the limited acid-reagent
contactj,,matting-over of the reagent surface by collected material,  and
analysis problems associated with the  large amount  of excess reagent on the
filter.
                                     42

-------
     Gas-phase fixation holds the most promise and was ultimately adopted
as part of the proposed method.  The only volatile fixing reagents which are
suitable in speed of reaction and stability of adducts are amines and
amine derivatives.  This method 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.
Analysis Methods
     The final criterion to be considered in this discussion of the theore-
tical development of the method is  the sample analysis methodology.  Just
as the available collection/fixatiqn techniques placed constraints upon
the theoretical fixing reactions that could be utilized, so it will be
seen that the available analytical  techniques further limit the choice
of fixing reagents and collection methods.
     Relevant analytical methods include not only ways of measuring sulfuric
acid, but also means to measure any conceivable stoichiometrically related
adduct.  Of course, the choice of adducts has already been greatly limited
in previous discussion by chemical  or sampling criteria.  However, for
the sake of a logical argument, the present discussion will deal with all
of the analytical possibilities implied in the original listing of potential
fixing reactions (Table 2).
Non-incorporative H^SO, Derivatives —
     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 incorpor-
ating at least part of the H2SO, molecule in the adduct to be measured.
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.
     This conclusion was drawn from the results of experiments on the
oxidation-dehydration reaction of sulfuric acid with p-hydroxybenzaldehyde
(page 14).  Even this relatively specific reaction could be duplicated
by other substances which may be in the atmosphere.  The product does not
incorporate sulfur from sulfuric acid, so it is not a unique product.

                                     43

-------
Evidently, the reactions of sulfuric acid are not sufficiently unique
in themselves to form a distinct non-incorporative adduct; sulfur from
the acid must be included as an indication of its origin.
     As a result, the number of relevant analytical techniques is greatly
reduced.  Since the uniqueness of the adduct is based upon the incorporation
of part of the H2SO, molecule, it follows that the analysis method must
measure this portion of the adduct.  Thus, the analysis 'problem is reduced
to measuring a closely related derivative of sulfuric acid without inter-
ference from other similar compounds.
H  Measurement —
      Simple acidity measurement is not appropriate for l^SO^ analysis,
unless  it can be separated from other substances which may affect pH.
So many substances have this property that such a separation seems imposs-
ible.   This characteristic of sulfuric acid is not sufficiently unique to
serve as  the basis for  selective analysis.
   -2
SO,   in  Solution Measurement —
      Greater selectivity may be achieved by measuring the sulfate portion
of sulfuric acid, rather than its  acidic protons.  However, this is  still
inadequate, unless  the  acid can be first separated from  sulfate salts.
Selective solubility  in an organic solvent has been ruled out as an  effec-
tive  separation method  (page  29),  which eliminates both  direct collection
in a  solvent and extraction of a filter sample by a solvent as viable alter-
natives.  Sulfate measurement may  still be useful, however, if a more
effective separation  technique is  successfully employed.  This section
discusses several wet-chemical and instrumental methods  of measuring
sulfate in solution.
      Gravimetry—The  classical, though seldom used, method for analysis  of
sulfate is by precipitation of barium sulfate from a hot, slightly acid
                                                 =      i I
solution  by the addition of barium chloride:  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.  Interfering
substances are largely  cations such  as lead, strontium,  and calcium, which

                                     44

-------
form insoluble sulfates, although anions  of weak  acids, nitrates, chlorates,
and heavy metal ions interfere, if care is not  taken  to remove them  (56).
     Titrimetry—• There are many modifications of  titrimetric methods of
sulfate determinations.  Most of the methods involve  the pretreatment of
the solution by passing it through a cation exchanger, or addition of suit-
able reagents to eliminate interferences  from metal ions.  Direct titrations
with a barium salt, either barium chloride (BaCl2) or barium nitrate
(Ba[N03l2), have been reported using various indicators, such as diphenyl-
carbazone, sodium alizarinsulfonate (57), and nitrochromeazo (58), etc.
Potentiometric titrations using both barium and lead  salts have been
reported as well.  Probably the most satisfactory titrimetric method is an
indirect determination accomplished by adding a known amount of barium,
strontium, or lead salt, and titrating the excess with a suitable reagent,
such as ethylenediaminetetraacetic acid   (EDTA) (59).
     The prescribed EPA method for sulfate determination is titration of
the sample with a known Ba(C10,)  solution, using thorin as indicator.
            —2                                    ++
When all SO,   is precipitated as BaSO,,  excess Ba    forms a colored complex
with thorin.  Hevel, Protzmann, Davis and Knarr (60)  report an indistinct
endpoint using this method and prefer to  titrate  with Ba(C10,)_ solution
using solfonazo III as indicator.
     Eastman Chemical Abstracts 122 and 124 give  titrimetric methods for
sulfate determination following precipitation by  benzidine.  Belcher,
Nutten and Stephen (40) suggest 4-amino-4-chlorodiphenyl hydrochloride as
a precipitating agent to form an amine sulfate which  is less soluble than
benzidine.  The only interference is phosphate, which can be removed as
the Ca or Zn salt.  Precipitation is analyzed by  titration with standard
NaOH.
     Colorimetry—Colorimetric methods for sulfate determinations follow
the same general outline as found in gravimetry,  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 methyl-
          thymol blue (61).  The remaining barium is  then determined

                                     45

-------
         colorimetrically, thus allowing the amount of  sulfate  to  be
         calculated:

                       excess Ba4"*" + S04~ -> BaSO^ + Ba                  (30)

                            excess Ba"*^ + MTB -> BaMTB                    (31)

      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:

                                 i i      =                     i i
                              Ba   + S04  -> BaSO,-1- + excess  Ba           (32)
                          -1^1.                                _
                  excess Ba   + K2Cr04 ->• BaCr04 +  excess Cr04~ (yellow)   (33)

      3)   In the method described on page 29,  the  insoluble barium  chlorani-
          late 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 (50).
                          BaC6CL204 + S04  -> BaS04+ + H^Cl^ (violet)  (34)

     Other less familiar colorimetric procedures for the determination of
sulfates are given below:
     a.   Barabas and Kaminski (62) formed a purple-colored sulfonic acid
          derivative of pararosaniline which was analyzed by colorimetry.
          This method detects S02 as well.
     b.   Goguel (63) added excess Fe   to the sample to form the soluble
          FeS04  complex, which is determined colorimetrically.  Phosphate,
          fluoride, chloride, and other anions also complex with Fe     but
          some correction can be made.

                                      46

-------
     Turbidimetry and Nephelometry—In  these methods  for  sulfate determina-
tion,   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 nephelometry, the radiation that is  scattered by the particles is
measured perpendicular to  the axis  of indident 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.
     Potentiometry—Rebel, et al..  (64)  titrated  a sample with PB(C10.)0
                                                                    4 2  _2
using an Orion lead specific ion electrode  as  an indicator.  When all SO,
                     +2           +2
is precipitated by Pb  , excess  Pb   is sensed by electrode.
     X-ray Fluorescence—X-ray fluorescence depends upon the exicitation
of secondary x-rays characteristic  of each  element  by absorption of primary
x-rays.  The identification of these characteristic wavelengths and the
measurement of their intensities constitute a  method for qualitative and
quantitative analysis.  Elements from atomic number 12 and up in solid and
liquid matrices may be determined routinely by x-ray fluorescence.   Liquid
samples should exceed a depth which will appear  infinitely thick to the
primary x-ray beam.
     Quantitative analysis by x-ray fluorescence requires the use of a known
standard for count comparison.   Before  the  comparison of the standard and
the sample can be made, it is necessary to  correct  for matrix effects.  The
correction for matrix effects can be accomplished by diluting the sample
with a material having  a  low absorption, or by  applying an internal standard
technique.  The internal standard technique is valid only if the matrix
elements affect both the reference  line and the  analytical line in exactly
the same way.  The internal standard line and  the analytical line ratio is
not a true measurement of  the concentration of the  element to be determined,
if the reference line or the analytical line is  selectively absorbed or
enhanced by a matrix element.
     Since x-ray fluorescence is responsive to an element, no matter in what
                                      47

-------
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.
Sulfur would, of course, be the element measured,  and this method would be
a wholly-instrumental way of measuring total sulfate in solution.  Other
sulfur-specific instrumental methods require that  the sulfate-containing
solution be atomized or vaporized, and these are discussed in the following
section.
Separation from Interferences by Decomposition/Volatilization —
     None of the methods for sulfate measurement described in the preceding
section can solve the selective analysis problem alone.  In every case, it
is necessary that sulfuric acid be separated from  sulfate salts and other
potential interferences prior to sulfate measurement.  If this cannot be
done during the collection process, then it must be done as part of the
analytical procedure.
     Volatilization i(Microdiffusion) of H?SO, —The only physical property
in which sulfuric acid differs greatly from sulfate salts is volatility.
Sulfuric acid  is often considered non-volatile at  room temperature, but it
does have a measurable vapor pressure and, when heated over 100°C, can be
volatilized quite rapidly.  In contrast, sulfate salts are truly non-volatile
until their decomposition temperature is -reached,  and even the lowest-
decomposing sulfate  (ammonium hydrogen sulfate) is stable to at least 230°C.
     Dubois, £t £il.  (65) first used this property  to separate H2SO, from
glass fiber filters by microdiffusion (volatilization).  Maddalone, et al.
(53) studied the efficiency of microdiffusion from various filter media
at 100'C to 195°C using 35S-labeled H2S04 and concluded that Mitex Teflon
and Paco graphite filters had the best diffusion characteristics.  Scarin-
gelli and Rehme (66) reduced the acid vapor to S02 over hot copper, which
was subsequently measured with the flame photometric detector  (FPD) or the
West-Gaeke bubbler (65).  Dharmarajan, et al. (68) collected the volatilized
acid onto a surface coated with the sulfate-precipitation agent, perimidyl-
ammonium bromide, and reduced the resulting sulfate to SO, at  500°C for
measurement by the West-Gaeke bubbler.  Further elaboration of  the micro-
diffusion technique by a research group at the Cabot Corporation  [Richards
                                      48

-------
 (69)] led to the development  of a prototype instrument which collects  the
 aerosol onto a filter and volatilizes it into an FPD for analysis.   This
 instrument has been  further developed by Mudgett, Richards,  and  Roehrig  (70)
 into a fully automatic unit.
     The microdiffusion  technique does effectively separate  sulfuric acid
 from sulfate salts on the filter, but at the expense of another  serious
 interference.  Acid  on the filter is free to react with other collected
 particulates.  In fact,  during  the volatilization process, it may be
 expected that any substances  on the filter capable of reacting with  the hot
 acid vapor will have ample opportunity to do so.
     Selective Decomposition  of H2S04 Adduct—This deficiency of direct
 volatilization schemes is widely recognized among researchers in the field,
 as  is the need for fixation of  the acid as it is  collected in order  to
 eliminate these side reactions.  To date, the only published attempt to
 solve this problem was made by  Thomas, ^it &L^. (43) , who collected sulfuric
 acid on a glass fiber filter  impregnated with the sulfate-precipitating
 reagent, perimidylammonium bromide (PDA-Br).   Unfortunately,  the insoluble
 (PDA)2SO, thus formed was too stable to be thermally decomposed  apart from
 ammonium sulfate.  (See  Page  21.)
     In this area, the requirement of selective  analysis seems to be at
                •
 odds with that of fixation.   If the acid is not  fixed,  it may be lost
 through side reactions on the collection surface;  but,  if it  is  fixed,
 the separation from  other interfering substances  may be impossible.  West (53)
 has made the interesting suggestion that an enzymatic reduction  of (PDA)2SC>4
 may be possible, but no  work  on this technique has been published.   The
 only alternative is  to find a different fixing reagent which satisfies both
 requirements by stabilizing the acid as an adduct which can  be selectively
 decomposed.
     The solution is necessarily a compromise between adduct  stability and
 selective analyzability.  Insoluble sulfates  are  ideal adducts from  the
viewpoint of stability,  but that same stability makes their  decomposition
 temperatures too high to avoid  interference from  common atmospheric  sulfates.
The present method proposes adducts which are water-soluble,  and thus
chemically stable, but can be selectively decomposed.   It is  this compromise

                                       49

-------
solution which is unique to  the method  considered here.
     Measurement of Evolved  Gas—Once the  adduct has been selectively
decomposed, a system for measuring the  evolved gases must be employed to
complete the analytical procedure.  From the discussion on page 43, it is
evident that the system must be specific for sulfur, if interferences are
to be avoided.  With some analytical methods, however,  the oxidation state
of sulfur in the evolved gas is also an important consideration.
     When sulfuric acid is volatilized directly, the evolved gas (loosely
termed "sulfuric acid vapor") probably consists of a mixture of S02» SC>3
and H20 vapor.  The hydrated S02 and SO molecules are  extremely reactive
and may be lost during the volatilization  procedure. When the acid is
fixed as dry acid crystals,  the evolved gas favors S02  more strongly, and
it is much less reactive in  the absence of water, since no acidic protons
are generated.
     Of the two sulfur gases, the one which is favored  depends upon
temperature, pressure, oxygen content in the chamber, and the presence
of catalysts.  A mixture may be converted  entirely to S0~ (71), or to
S02 (66) by the appropriate  catalytic conditions.  With this in mind, there
are several possible methods for analyzing the evolved  gas.
     If the gas is predominantly SO-, it may be dissolved in water to form
sulfuric acid, which can then be analyzed  by a standard method for sulfate.
(See page  44.)  This is less attractive than other alternatives for several
reasons.  First, SO., is more easily lost by side reactions than SO,; speci-
fically, it has a stronger affinity for water and produces a stronger acid
when hydrated.  Secondly, many of the wet-chemical sulfate methods  (those
based upon precipitation of  BaSO^) are sensitive only to the low ppm range.
If the detection limit is 5  ppm   (or 5 yg/ml), and the collecting volume
10 ml, 50 yg of acid would have to be collected before it was detected.  With
                               3                            i
a measurement goal of 0.25 yg/m  sulfuric acid in air,  200 m  of air would
have to be sampled.  This may be reduced somewhat by concentrating  the  SO -
collecting solution, but it  is more desirable to use a method with  higher
sensitivity which will allow a smaller air sample volume.  Finally,  some of
the wet-chemical methods are subject to non-sulfur interferences which  may
be present in the decomposition gas, such as oxides of phosphorus  and

                                     50

-------
nitrogen, particularly when such a large  sample  is  taken.
     If the evolved gas consists predominantly of S09,  it can be measured
by the West-Gaeke bubbler  (67).  In  this  method, SCL  is collected in an
aqueous solution of 0.1 M  sodium tetrachloromercurate where it is fixed
as the stable disulfitomercurate (II)  ion.  When a  p-rosaniline...hydre-
chloride-hydrochloric acid mixture and formaldehyde are added, a violet
color is produced proportional  to the  amount of  fixed S02 present.  The
absorption maximum at 560  my is used for  a  spectrophotometric measurement.
This method will detect the equivalent of 1 pg H9SO,  converted to S09
                                                 £•   ^                £*
and dissolved in a 10 ml collection  solution.  Therefore, the sensitivity
is adequate.  Nitrogen dioxide  and ammonia  (72) may interfere, but will
probably not be present to a significant  extent  in  the decomposition gas.
The West-Gaeke bubbler would be the  best  method  for measuring S02 in the
decomposition gas, if more rapid instrumental methods were not available.
     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 photometric detector (FPD) .  Sample gas is fed  onto an air-HL 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 use  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 of about 1 ug of
H.SO,.  Saturation occurs  at the upper limit, because excess sulfur atoms
reabsorb the emitted light (73).
     It is evident from Table 6 that the  higher  sensitivity of the FPD allows
a much smaller sample to be collected  than with the other methods.  Thus, less
distortion of the atmospheric sample occurs by concentration on the collection
surface.  Measurement by FPD is also much more rapid  than the West-Gaeke
bubbler.  The only theoretical  advantage  of the bubbler is that it allows
the evolved gas to be rapidly separated from the sample and concentrated
elsewhere.  With the FPD,  the gas must be held for  at least several minutes
before flushing into the detector.   If loss of sample during this period  is

                                       51

-------
significant, it may be advantageous to concentrate the gas in another manner,
prior to FPD Analysis.

              TABLE 6.  SENSITIVITY COMPARISON OF SULFUR-GAS
                        MEASUREMENT METHODS
                                                        Min. Sample
                                   Min.  wt.                       -
     Method                       H2S04  (ug)             Volume (m )
S07 2 wet-chemicalt                  50                    200
  4
S02 West-Gaeke bubbler                1                      4
FPD                                0.01                   0.04

              o
*For 0.25 yg/rn  sensitivity.
tAssuming 10 ml analysis solution.

Summary of Analysis Method —
     It was determined that the analysis method must be specific for
                     -2
sulfur, either as SO,  , S02, or both.  Since selective solubility is not
a very promising approach, it was determined that selective thermal decompos-
ition of the adduct at 200°C or below was the only feasible method for
separating interferences.
     Three methods of measuring the evolved sulfur-containing decomposition
                                     _2
gas were discussed:  wet-chemical SO^   analysis, the West-Gaeke SO-
bubbler and the FPD.  Of the three, the FPD had the greatest advantage in
sensitivity and flexibility,, responding to all forms of sulfur regardless
of oxidation state.  Thermal decomposition of the adduct and FPD measure-
ment of the evolved gas was, therefore, adopted into the Atlantic Research
method.
Summary;  Theoretical Development of Atlantic Research's Method
     The preceding discussion has attempted to illustrate the 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-

                                     52

-------
derivative fixing reagent.  Considerations of selective analysis added the
requirement that the adduct decompose at 200°C by releasing a sulfur gas.
     All of these conclusions were supported by experimental work, some
of which has been briefly described.  The following sections contain a more
detailed discussion of experimental work relating to the developed method.
                                       53

-------
                                 SECTION 5
                                EXPERIMENTAL
APPARATUS
Aerosol Generator
     The sulfuric acid aerosol generator used in these experiments was
based upon the atomizer-burner model described by Thomas, et al .  (74).
In this model, a dilute H2SO,  solution is aspirated into an H2-Q2 flame,
where it decomposes to H20 and SO .   Recombination occurs to yield sulfuric
acid aerosol.
     A Beckman 4060 Large Bore Atomizer Burner Assembly was mounted at the
base of a glass stack 1.2 M (4 ft) high by 15.2 cm (6 in.) diameter.  Burner
                                                                   2
gas back-pressures were set at 1.4 kN (0.2 psi) for H_ and 1.4 kN/m  (10
psi) for 02.   The aspirated solution was 10% I^SO, by weight.
Collection of Aerosol
     A glass sample probe was positioned at the top of the stack and drew the
aerosol into a glass mixing vessel, where it was divided into two equal
streams.  Each stream was drawn through a Millipore Mitex filter in a nylon
filter holder by a large vacuum pump, and added at a point 7.6 cm (3 in.)
above the filter holder.  Figure 6 illustrates the complete system.
     Sampling was carried out for 2 to 4 minutes with about 5 ml of solution
aspirated during that period.  Total volume sampled (measured by a wet
test meter) was 0.003 to 0.1 m .  The position of the probe inlet at the
top of the stack was found to be crucial for collection of a suitable amount
of sample.  If placed in the center of the stack, the sample would be too
large and saturate the detector, while if placed outside the stack perimeter,
a measurable amount would not be collected.  At the median position, a
suitable amount of sample was collected in a few minutes.
                                     54

-------
 Particulate
 Interference
Gaseous
Interference
                                Gaseous
                              Interferences
                                       Wet Test Meter
                   Figure 6.  Dual collection/fixation  system.
                                          55

-------
Analysis of Samples
Decomposition Apparatus
     The instrument constructed to analyze the samples consisted basically
of a heated cell to thermally decompose the sample,  and a valve arrangement
by which helium carrier gas was delivered to the FPD,  either through the
cell, or through a bypass line.  Figure 7  is  a schematic diagram of the
system.
                                                    4-WAY
                                                    VALVE
                      Figure 7.  Analysis instrumentation.

                                                                 2         2
     The sample cell was machined from a block of aluminum 5.7 cm  x 5.7 cm
x  2.54 cm  (2-1/4 in. x 2-1/4 in. x 1 in.).  The large central cavity for the
sample was 1.59 cm  (5/8 in.) ID and coated with Teflon.  One end of the
cavity was tapped to accomodate an aluminum NPT fitting, and a groove was cut
around the other end to hold a Viton 0-ring.  On the breech end, an Allen
screw was placed on either side of the cavity to hold the breech cover in
place.  The cell was opened by loosening these two screws and sliding the
cover off the screw shafts.  An airtight seal was provided by the 0-ring,
and all surfaces in contact with the sample were Teflon-coated aluminum.
The cell was heated by two implanted cartridge heaters  (not shown) on
either side of the cavity.  Asbestos tape was wrapped around the cell for
insulation.
     The valve was a Valco 4-way valve constructed of Carpenters C-20 alloy,
which is recommended for reactive gases such as S0r  This valve could  be
heated to 300°C without damage.  All connecting tubing  was 0.16 cm  (1/16 in.)
                                      56

-------
ID Teflon-lined aluminum.  The  sample gas lines and valve were wrapped with
Nichrome ribbon (sandwiched between glass fiber tape) which could be heated
independently of the sample cell.
Flame Photometric Detector (FPD) —
     The FPD used in these experiments was a Melpar Model 100 with a 374 nm
optical filter.  When  samples from the aerosol generator were first run
on this system, very peculiar FPD 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 8 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 SC^-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 S02 with a syringe.  The source of
S02 was a compressed-gas lecture bottle, fitted with a valve and a silicon
rubber septum.  Dilutions of SO™ 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 pi.  Several blank injections were made after each sample injection
to insure that the syringes were clean of residual S02-
     Microliter amounts  of pure SO  from a gas cylinder were injected in
the first experiments.   Figure  9 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 S02 level.  It  was evident that all of these peaks were resulting
from a saturated detector.
     When these peak heights were compared to those previously obtained
from samples (Figure 8) , it was evident that the samples had been saturating
the detector.  The smallest volume injected, 10 pi, corresponds  to about
3 pg of SO- and even this seemed to be considerably over  saturation  level.
     In order to bring the sample size down to the quantitative  response
range, 1 cc of pure S00  was injected into the clean 305 cc dilution  vessel.
                                      57

-------
OS

 o
 X

 CO
 P.
  co
  01
  CO

  g
  ex
  CO
  0)
 of

 n
 PL.
 at
 OS
          Figure  8.   Early FPD  recordings from  samples.
                                 58

-------
Ul
                         Figure 9.  Effect of  injecting progressively larger S02 volumes.

-------
The resulting mixture was about 3300 ppm S02-   When microliter volumes of
this dilution were injected,  the peaks were much sharper and their heights
were related to the amount injected as figure  10 shows.   The response was
fairly linear from at least 0.1 yg up to about 0.6 yg,  beyond which 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 S02 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 8, and
were estimated to contain 20 to 30 yg H2SO, (by an independent pH measure-
ment) , did indeed saturate the detector.
     To look at the lower end of the scale, a new S02-air dilution was made,
which was one-tenth as concentrated as the last one (330 ppm).  Sensitivity
of the electrometer was increased by setting the range  at 10 and attenuation
at 32.  As figures 10 and 11 show, the response was fairly linear from 10 to
100 yl of the injected sample.  There was some difference observed between the
two syringes, the smaller one producing less response from the same volume
of sample.  This was possibly due to the greater surface-to-volume ratio
of the smaller syringe, causing retention of S02 by glass surfaces to be
greater.  Both sets of data are shown.
     These experiments, in which S02 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 8 were saturation  artifacts.  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.  In subsequent runs, sampling procedures were modified to
collect smaller samples and the resulting peaks resembled the unsaturated
S02 peaks.  The linearity of the response was somewhat  puzzling, since it is
supposed to be a square function of the sample size.  No explanation has
been suggested for this observation.
Supporting Instrumental Analysis  .—
     Instruments other than the FPD were used to obtain information

                                     60

-------
X
CQ
0)
CO
g
0)
•u
cfl
                4      4
              0.05    0.10
  4
0.20
                       4          4444
                     0.30        0.50   0.60   0.70      1.00
                          Ug of S02

Figure  10.   Quantitative FPD response to Injected SO-.

-------
o
-H


 O
 CO
 &

I

 en
4-1
&
 00
•rl

£


•a
0)
CM
                           V
                                                                O  = lOOjul syringe


                                                                Q  = 500 ptl syringe
               0.01       0.02       0.03      0.04      0.05      0.06       0.07       0.08      0.09
                        Figure 11.   Quantitative  response of FPD to

-------
concerning decomposition of  the  adducts, although these instruments were not
considered adequate for analysis of  real  samples.
     Differential Thermal Analysis  (DTA)—DTA was  used to  obtain  information
about the temperatures at which  candidate adducts  and potential interfering
substances decomposed.  This instrument operates by amplifying and recording
the temperature difference between  a quantity of the sample  substance and
an inert reference  (ground glass) as they are slowly heated  in a  baffle.
When an endothermic or exothermic process occurs in the sample, it is
recorded as a peak at the temperature of  the baffle.
     Figure 12 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 indi-
cates that an adduct which decomposes and releases SO., at  200°C may be
analyzed without interference from  these  substances.
     Mass Spectrometry—Mass spectrometry was utilized to  identify the gases
released by the decomposing  adducts.  All of the adducts discussed in the
next section were found to release  SC>2 at 200°C, but no detectable SO.,.
There was some ambiguity in  these results, however, because  even  unreacted
H-SO, gave predominantly S0? at  this temperature.
FIXING REAGENTS
     The fixing reagents used in this method have  been described  as volatile
amines and amine-derivatives (hydroxylamines and oximes).  One compound
in each category was evaluated,  although  this is not to imply that the
compound chosen is necessarily the  most desirable  one.  The  compounds were:
     (C2H5)2NH                     (C2H5)2NOH                 CH3CH=NOH
Diethylamine (DBA)         Diethylhydroxylamine (DEHA)    Acetaldoxime (AAO)
All are liquids at room temperature, but  are sufficiently  volatile to
generate a substantial vapor from a bubbler.
Reactions of H9SO/[
     The  functional groups  of  the compounds DBA, DEHA, and AAO  are all

                                       63

-------
50       100      150      200       250       300      350      400



             T °C (Corrected for CHROMEL ALUMEL Thermocouples)
450
500
           Figure 12.  Composite DTA of potential interferences.

-------
sufficiently basic to form the bisulfate  salt  from sulfuric acid.  Base
strength decreases in the order:  DBA > DEHA > AAO.  When concentrated
H2S04  was added to excess DEA or AAO, white crystals immediately formed
and crystallized out of solution.  They were water-soluble and turned black
when heated.  These physical properties,  as well as IR spectral evidence,
strongly suggested an ionic sulfate or bisulfate salt.  DEHA formed a very
viscous yellow oil when I^SO^ was added,  but this  oil was also water-
soluble and decomposed when heated.  A peak at 1650 cm"1 in the IR spectrum
of this oil, which was absent in the unreacted reagent and acid, suggested
an N-H bending vibration from protcmatiori of nitrogen, as postulated in
the formation of the bisulfate.  NMR did  not give  any information on the
position of the acidic proton, because of exchange with the solvent.
     Thus, it appears that these compounds do react with sulfuric acid
rapidly, by forming a sulfate or bisulfate salt.   There is no reason to
suppose that the product would be different when formed under sampling
conditions.
Decomposition of Adduct
     DTA's of the white solids obtained from DEA and AAO, as described above,
were not very informative because of the  multiplicity of peaks, both over
and under 200°C.  Release of a sulfur gas at 200°C was confirmed for all
three adducts by heating in the FPD sample cell previously described.
Mass spectra indicated that S09 was the main gas,  and this was supported
                                                           -2
by the West-Gaeke bubbler (which does not detect SO- or SO,   ) .
     Some other amine sulfates were examined by DTA:  perimidylammonium
sulfate (an insoluble salt) and pyridinium sulfate (from pyridine).   The
DTA's indicated (Figure 13) that no decomposition  occurred under 250°C.
Consequently these are probably not suitable as adducts.
PROCEDURES AND RESULTS
     The FPD analysis apparatus described on page  56 was used to analyze
fixed aerosol samples from the aerosol generator by two major procedures.
In the rising temperature procedure, the  sample filter was inserted into
the analysis cell at room temperature.  With carrier gas flowing through
the cell to the FPD, the cell temperature was  then slowly raised.  In the

                                     65

-------
50
100       150       200       250       300        350       400
       T. °C (Corrected for CHROMEL ALUMEL Thermocouples)
450
500
          Figure  13,   (PDA12S04 and Py2S04 DTA.

-------
fixed temperature procedure, the sample was  inserted at 120°C, and rapidly
raised to 200eC, with the carrier gas bypassing  the cell.  After five minutes
at 200°C, the valve was switched to sweep gas evolved from the sample to
the FPD.  Using these two procedures, different  kinds of information were
obtained.
Rising Temperature Procedure
     This procedure was used to determine the temperature at which a sulfur
gas was first evolved from the sample.  As the cell temperature rose, it
was marked on the FPD chart and the temperature  at which the pen rose sharply
from baseline was taken as the decomposition point.  Large aerosol samples
were taken when this procedure was used to insure that the beginning of
decomposition was detected, although the peaks became saturated at a higher
temperature.
     The results  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 H7SO,.

                     TABLE 7.  TEMPERATURE OF FIRST FPD
                               SIGNAL FROM ADDUCTS
             H0SO, alone                              90
              2  4
             H0SO. + AAO                             140
              /  4
             H0SO. + DEHA                            190
              2  4
             H0SO, + DBA                             190
              2  4

This 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
     By holding the evolved S02 in  the  sample cell  for  several minutes
before sweeping it into the FPD,  the gas evolved over that period  is
concentrated.  This procedure  is, therefore, more sensitive to  small samples

                                      67

-------
than the previous procedure.  If care is taken to collect samples that are
not too large, the peaks obtained when the valve is switched are sharp,
and their height may be taken as a measure of the H^SO.  collected.  This
procedure, therefore, was adopted for the quantitative analysis of samples.
     The complete technique for analyzing samples with the FPD-decomposition
cell device is described by the following steps:
     1)  Bring cell temperature to 120°C.
     2)  With the carrier gas switching valve on bypass, open the cell
         and insert the sample filter (rolled).  Reseal the cell.
     3)  Maintain the cell temperature at 120° for 5 minutes, then switch
         the valve to sample and record the peak (if any).
     4)  Switch the valve back to bypass after two minutes and rapidly
         raise the cell temperature to 200°C.
     5)  Maintain the cell at 200° for 5 minutes, then switch the valve
         to sample and record the peak.  After 20 minutes, switch back to
         bypass.
     6)  Clean the cell between runs by heating to 250°C for 10 minutes
         with the valve on sample.
     Since the adducts are all stable at 120°C, inserting the sample at
this temperature prevents any loss before the breech is closed.   Moreover,
this step dries the sample so that the SO  evolved later is less reactive.
If any unfixed H2SO, is present, it will also produce a peak at 120°C.
With the present cartridge heater arrangement, the cell temperature can be
raised to 200°C in 3 minutes.  Once this temperature is reached, Step  (5)
may be repated several times in order to estimate the completeness of  sample
decomposition each time.
     The results of the rising temperature procedure were confirmed by this
method.  When the acid was not fixed, it gave a peak at the 120°C step
and could be completely removed by heating for an hour at this temperature.
When the sample was fixed with AAO, there was no peak at 120°C  but a
pronounced peak appeared at 200°C.  I. Figure 14 illustrates these results,
which reaffirmed the thermal stabilization produced by fixation.  It also
demonstrated that the acid was fixed completely and rapidly, since no
                                      68

-------
                      Unfixed

                      H2S04
o
T-l


 O
en
Cu
                                                                H2SO4/AAO

                                                                  Adduct
0)
CO
a
o
P.  0
co  2

-------
120°C peak appeared,  even when only 15 seconds of reagent-acid contact was
allowed after aerosol collection.
     In addition, the speed of decomposition of the AAO-adduct at 200°C
was evaluated.  The procedure was  to repeat the heating period with the
valve on bypass several times, with progressively smaller peaks produced
from each sample.  As figure 15 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) sample was analyzed, decomposition seemed to be essentially complete
after the first 5 minute period as shown in figure 16.
     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 appears to be a quantita-
tive measure of the amount of acid originally collected.
METHOD EVALUATION
Accuracy
     The accuracy of an analytical method is defined as the degree to which
the measurements obtained agree with the true values, as determined by an
independent method of known accuracy.  Since there is no independent method
of known accuracy for analysis of  atmospheric H2SO,, this requirement
cannot be strictly fulfilled.  Moreover, the unpredictability of the type
of aerosol generator used makes it impossible to calculate how much acid
is deposited.
     One 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  experiments described on page 67 gave
some data of this type, since each successive peak from the same sample
was smaller, presumably due to the lessened amount of adduct each time.
It was desirable, however, to show the validity of the technique with
different samples as well, since this would be its practical use.
                                     70

-------
                       t
                         t
After 5 min
  at 200°
After 2nd 5 min
    at 200°
After 3rd 5 min
    at 200°
After 15 min
  at 200°
After 5 min
  at 200°
                         Figure  15.   Completeness  of decomposition  of  AAO/H^SO,
                                      adduct  after  5 min (A) and 15  min (B).

-------
01
C3
a
o
CX,
CO
Q>
4J
td
     120
200 (1)                   200 (2)

           TEMPERATURE (°C)
200 (3)
                     Figure 16.   Typical FPD trace from AAO-treated sample.

-------
      The simplest way to produce samples of a known size relation was  to
 cut a whole sample into halves, then one section in half again, etc.   Figure
 17 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 order magnitude was not precisely uniform.  This
 may have been due to  nonuniform distribution of adduct on the filter,
 unsymmetrical cutting, or  random sample loss during handling.
     As  a final test  of  validity,  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 18 shows data from
 these  runs.
     It  was concluded from this data that peak height was  an accurate
 measurement of  the amount  of acid  adduct in a sample.   The only possible
 way to independently  verify this measurement would be to make 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 I^SO^-methanol or -acetone  solutions for this  purpose.   There  is,
 however,  still no independent way  of verifying the  amount of  acid calculated
 to  be on the filter,  and this method is certainly less  representative of
 real atmospheric  aerosols  than the Thomas generator method.
 Precision
     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
 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 type of H2SO^
 aerosol generator used in  these experiments  was  known to have poor reproduc-
ibility from one run  to  another.   The problem was approached by designing
the generator's sampling probe  so  that  two samples could be collected
simultaneously, and would,  therefore,  be  identical.
     Several flow adjustments were required  before the rates through the
                                     73

-------
                         1/8 Filter
1/16 Filter
   t
1/22 Filter
Figure  17.   AAO-fixed filter  portions after  heating 5 min  at 200°C.

-------
0.071 m-
               0.057 m3
0.042 m3
0.028 in
    Figure 18.   AAO-treated samples,  different sampling volumes,
                 5 rain at 200°C.
                               75

-------
two holders were equivalent.   The last set of data, from a series of three
runs, is shown in Figure 19 and summarized in Table 8.

                    TABLE 8.   PEAK HEIGHTS OF SAMPLE PAIRS
                              COLLECTED SIMULTANEOUSLY
                            Peak Height
Run                 Sample //I           Sample #2           % Difference
 1                     3.3                 3.0                  9.5
 2                     8.3                 7.6                  8.8
 3                     5.4                 5.8                  7.2

For each pair of samples, the variation in peak height is less than 10%.
It was concluded from this evidence, however,  that reliability is no•less
than 90% and is probably greater.
Calibration
     Calibration of the FPD peak height to the amount of acid in the sample
is a very complex problem.  As a crude approximation, the sample peaks could
be compared to those produced by injecting known amounts of SO^.  Obviously,
the two procedures are not directly comparable.  Known samples could be
prepared by micropipet deposition, but even with this procedure, the FPD
response may not be comparable to that of real aerosol samples.
     It is also necessary to determine how much sample adduct is lost,
either by failing to decompose during the heating period, or by reacting
with surfaces in the sample cell.  And finally, in order to calculate the
    3
pg/m  figure, it is necessary to know the sampling efficiency, including
the amount of acid lost on the walls of the sample probe.
     These problems are not impossible to solve, but will require additional
research.  Presently, a rough estimate of the sample size is made by
comparison to the S09 injection peaks previously obtained.
INTERFERENCES ELIMINATED
Ammonium Sulfate and Bisulfate
     DTA's of ammonium sulfate and bisulfate previously presented indicated
                                     76

-------
O
.-I
 X

  CO
  p.
  Hi
  03
  a

  ai


  n


  0)

  •H
  4J
  td
                   Sample

              No. 1     No. 2
    Sample
No. 1    No. 2
   Sample
No. 1    No. 2
                  Run No. 1                   Run No. 2              Run No. 3


           Figure 19.  Pairs  of AAO-fixed samples collected simultaneously.

-------
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 al. (65) discussed the use of (NIL) SO  as a standard for diffusion at
195°C.  Maddalone, £t al. (53), however, reported that a thermal gravimetric
analysis (TGA) of (NH.KSO,  indicated no significant decomposition until
250°C, in agreement with our DTA's.  Erdey, et al . (75) have reported the
following reactions:
                                 _
                    (NH4)2S04          >  NH4HS04 + NH3                   (35)

                      NH4HS04 250"350°C>  NH3 + H2S04.                    (36)

     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, 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, no signal
was detected at 200°C.  Interference from this source has apparently been
eliminated .
Sulfur Gases
     Since the FPD responds to sulfur in any form, it was recognized that
the ambient sulfur gases, S02> H2S, and COS, may interfere by reacting
with the gaseous reagent.  They may then be collected on the filter as a
sulfur gas-reagent complex.  These sulfur gas-reagent complexes would
interfere with H2S04 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.

                                     78

-------
      As shown in Table 9, of the three gases, only SCL was significantly
 collected,  and in every case, the S02~reagent complex had a very low thermal
 stability compared to the I^SO^-adduct.  This was independently verified
 by mass spectral evidence.  Figure 20 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
                                          J>P-2-   -J^2—      COS
                  DEA                      25        60*
                  DEHA                     25
                  AAO                      30
*Very little  absorbed

     First, it  appears  that  a fairly high concentration of S0«  (>25 ppm)
must be present before  a  significant amount is collected.  Sulfur dioxide
was being produced by the aerosol  generator in low  concentration (5 ppm by
Drager tubes),  but none was  ever detected at  the 120°C step during routine
runs.  Second,  even  if  the necessary 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.
(76) have examined the  formation of  charge-transfer (CT) complexes in the
gas phase reaction of S02 with amines.   Studies of  the analogous reaction
between S00 and NIL,  have  found mostly sulfite and bisulfite salts (77,78).
                  *^                                                     —9
It may be possible,  however,  for the collected S02  to be oxidized to SO^   ,
if catalytic  conditions are  present  (78).  For this reason, it  might be
desirable to  heat the filter  during  collection, so  that S02~reagent complexes
are immediately decomposed and removed.
Negative Interferences
     The rationale for  fixation was  given as  the necessity of preventing
the acid from reacting  with  other  particulates on the filter.   Experimental

                                      79

-------
   20 40 60  80 100  120  140   160      180

                     TEMPERATURE (°C)
                                                   200
Ul



1
CO
Ul
GC


g
u.
                     AAO/S02
_ 1 ,-.!'
! 1 1 1 1 1 1 1
' 	 1 	
Ul
v>
Ul
tc.
                 20- 40 60 80  100  120   140   160      180

                                   TEMPERATURE (°C)

                       
-------
 evidence for the effectiveness of  the  proposed method of fixation is not
 presently available  since  research  to date  has focused on formation of
 the  adducts,  their properties and  potential  positive interferences.   In
 the  theoretical  discussion  of the  use  of volatile amines for this purpose,
 it was  suggested that the filter be  heated during collection to remove water.
 Since an  aqueous solution of an amine  (or amine-derivative)  sulfate  is still
 quite acidic, drying the adduct should form  less reactive solid crystals.
 It was demonstrated that even the  least stable of the three  adducts  from
 AAO could be heated at 102°C without decomposition,  which is more than
 sufficient for the purpose  of drying.
     A related problem concerns the  loss of  evolved  S02  during  analysis
 by reaction with particulates on the filter  and  surfaces  in  the system.
 It has been hypothesized that the dry  S0? evolved during  this procedure
 is less reactive then the acid vapor produced  by direct volatilization.
More experimental work is needed in  this area, however.
     Finally, the effect of ammonia  must be  examined  in  representative
 concentrations.  It was found that highly concentrated ammonia reacted
 with undried adduct on a filter, which is not  surprising, since ammonia
 is a stronger base than any of the proposed  reagents.  With the reagent
 in such large excess, however, low ambient levels of ammonia should not
 pose an insurmountable problem.
 SUMMARY OF ACHIEVEMENTS
     The most important achievements of the  past twelve months'  research
are summarized below:
     1)   Three candidate reagents for  fixation of H2S04 were identified.
     2)   The corresponding  adducts were isolated and their likely compos-
         ition determined.
     3)   It was established that S02 is released by decomposing  the adducts
         at 200°C.
     4)   Formation of adduct by treatment of sample with reagent vapor was
         demonstrated by two different  temperature-programming techniques.
                                     81

-------
 5)   The FPD was calibrated by injection of known SCL-air standards,
     and the quantitative range of the detector was determined.
 6)   Sample size was adjusted to give analyzable FPD peaks.
 7)   A specific procedure for analyzing samples by the FPD was estab-
     lished .
 8)   Ammonium sulfate and ammonium hydrogen sulfate were shown not to
     be interferences.
 9)   The sulfur gases, SO,, H~S, and COS, were shown not to be inter-
     ferences where a preheating step was used to remove S02~reagent
     complexes.
10)   The AAO-H2SO,  adduct was shown to be at least 50% decomposed
     after 5 minutes of heating at 200°C,
11)   Peak height of the FPD trace was shown to be an accurate measure
     of the acid on a filter.
12)   Two identical samples were collected and shown to produce the
     same FPD peak heights, with a reliability of greater than 90%.
                                 82

-------
                                 REFERENCES


  1.    Fricket,  J.,  Trans.  Farady Soc. .  32., 1192 (1936).

  2.    Schrank,  H.,  Heimann,  H.,  Clayton, G.,  Gafaffer, W.,  and
       Wexler, H., Public Health  Bulletin, Washington, No. 306, (1969).

  3.    Stern, A., Air Pollution,  Vol.  1,  Academic Press, New York (1968).

  4.    Leighton, P., Photochemistry of Air Pollution,  Pergamon Press,
       London  (1957).

  5.    Fricket,  J. ,  Trans.  Faraday Society, 32,  1192 (1936).

  6.    Amdur, M., J. Air Pollut.  Contr. Ass..  19, 683  (1969).

  7.    Amdur, M., Schulz, R. ,  Dienker, P., Arch.  Ind.  Hyg. Occup. Med., 5,
       318  (1952).                                                     ~
  8.    Altshuller, A.P., Environ.  Sci. &  Tech.,  7^:8, 709 (1973).

  9.    Ludwig, F. L. , and Robinson,  E., Atmos. Environ, ,  2^ 13  (1968).

10.    Heist, R. H. , and Reiss, H.,  Journal of Chemical  Physics. 61:2,
       573  (1974).

11.    Goodman,  F. L. , Journal of  Catalysis. 36,  159 (1975).

12.   Matteson, M.  J., and Giardian, P.  J. , Environ.  Sci. & Tech. , JJ:1,
       50 (1974).

13.    Cox, R. A. and Penkett, S. A., Atmos. Environ., 4_, 425  (1970).

14.   Bufalini, M. , Current Research, ,5:8, 685  (1971).

15.   Takahashi, K., Kasahara, M. ,  and Katz, J.  L., Aerosol Science, 6>,
       75 (1975).

16.   Mirabel,  P.,  and Katz,  J. L., Journal of  Chemical  Physics, 60:3,
       1138 (1974).

17.   Amdur, M. 0., Arch.  Environ.  Health, 23,  459  (1971).

18.   Vanderpol, A. H., eit al., Science,  190, 570  (1975).

19    Meserole  F. B. , et  al., "Sulfur Dioxide  Interferences  in the Measure-
      ment of Ambient PlFticulate  Sulfates," NTIS No. PB-249-620 (1976).

20.   Faith, W. L., Air Pollution  Control,  John Wiley  and Sons, Inc.,
      New York, p. 143 (1959).

                                      83

-------
21.   Freiberg, J., Current Research, jh8, 731 (1974).

22.   Taarit, Y. B., and Lunsford, J. H., Journal of Phys. Chem., 77:11.
      1365 (1973).

23.   Vohra, K. G., and Nair, P. N., Aerosol Science, !_, 127 (1970).

24.   Kiang, C. S., and Haas, P., Nature. 250, 401 (1974).

25.   Blosser, E. R., "A Study of the Nature of the Chemical Characteristics
      of Particulates Collected from Ambient Air," NTIS No. PB-220-401,
      (1970).

26.   Feigl, F., Spot Tests in Inorganic Analysis, 5th Ed., Elsevier,
      Amsterdam, p.317-18 (1958).

27.   Cotton, F. A., and Wilkson, G., Advanced Organic Chemistry,  3rd Ed.,
      Interscience Publishers, New York, p. 353 (1972).

28.   Pelofsky, A., Ind. Eng. Chem. Prod. Res. Develop., tt (2), 187-191
      (1972).

29.   Rastogi, P. B. , and Kishore, K. , Ind. J. Chem. . 7_, 480 (1969).

30.   Takiura, K., Yuki, H., Honda, S., Koyima, Y., and Lan-yii, C., Chem.
      Phar. Bull. (Tokyo) 18, 429 (1970).

31.   Nagasawa, K., Inoue, Y., Carbohydrate Research, 28, 103-108 (1973).

32.   Feigl, F., op. cit., p. 316.

33.   Bissett, D. L., Hanson, T. E. , and Anderson, R. L., Microchem. J., 19.
      71 (1974).

34.   Bathhurst, E. T., and Cohon, J. M., J. Chem. Soc., Chem. Commun.,
      131  (1974).                                             ~~

35.   Hewitt, T. E., and Pardue, H. L., Clin. Chem., 1^ (10), 1128  (1973).

36.   Hewitt, T. E., and Pardue, H. L., Clin. Chem., ^1(2), 199 (1975).

37.   Ibrahim, E. A., Wahbi, A. M., and Abdel-Salam, M. A., Pharmazie
      _28:4, 232 (1973).

38.   McAllister, R. A., Crook, J. N., Smith, D. C., and Mackay, C.,
      Biochem. Med.. 6:3, 267 (1972).
                                     84

-------
 39.   Belcher, R., Kapel, M.,  and Nutten, A. J., Anal. Chim. Acta. 8, 122
       (1953).

 40.   Belcher, R.,  Nutten, A.  J., and Stephen, W. I., Anal. Chem., 1334 (1953),

 41.   Stephen, W. I.,  Anal. Chim. Acta. 50, 413 (1970).

 42.   Maddalone,  R.  F.,  McClure,  G. L., and West, P. W. ,  Anal.  Chem.,  47:2,
       316  (1975).                                        	  —

 43.   Thomas,  R.  L. ,  Dharmarajan, V., Lundquist,  G.  L., and West, P  W.,
       Anal.  Chem.,  48:4, 639 (1976).

 44.   Jacobson, C.  A., and Hampel, C. A., ed., Encylopedia of Chemical
       Reactions,  IV,  Reinhold  Publishing Corp., New York,  p. 485  (1956).

 45.   ibid..  I, 691.

 46.   ibid.,  II,  576.

 47.   ibid.,  II,  587.

 48.   ibid.,  VII, 237.

 49-   Huygen.  C., Atmospheric  Environment,  j),  315-319  (1975).

50.    Bertolacini,  R.  J.,  and  Barney,  J.  E., Anal. Chem.. _29:2, 281 (1957).

 51.    Federal  Register,  3^:247, December  23, 1971.

 52.    Barton,  S.  D. , and McAdie,  H.  G., International Clean Air Congress,
       Washington, D. C., Dec.  6-11,  1970  (Paper CP-7D).

 53.    Maddalone,  R. F.,  Shendrikar,  A. D. ,  and West, P. W., Mikrochimica
       Acta  [Wein],  391 (1974).

54.    Leahy, D. ,  Siegel, R., Klotz,  P., and Newmann, L., Atmospheric
       Environment,  9_,  219  (1975).

55.    Robbins, R. C.,  and  Cadle,  R.  D., Journal of P. Chem., 62, 469  (1958).

56,    Bobtelsky,  M. , and Eisenstadter, J. , Anal. Chim. Acta, 41, 186  (1956).

57     Ceausescue, D.,  and  Asteleanu, M., Hidroteh Geospodaririea Apelor
      Meterorol [Bucharest], 11,  278  (1966) CA, 66, 5658K (1967).

58.    Basargin, N. N. , and Nogina, A. A., Zh. Anal. Khim.. 22,  394 (1967).
                                    85

-------
59.   Effenberger, M., Fortschr. Wasserchem. Ihrer Grenzgeb., JL, 173
      (1964); CA, 67., 36312 a (1967).

60.   ALCHE Symp. Ser. 700137, Recent Adv. Air Pollut.  Control, pp. 299-
      306 (1974).

61.   Lazrus, A., Lorgane, E., and Lodge, Jr., J. P., Advan. Chem., 73,
      164 (1968).

62.   Barabas, S., and Kaminski, J., Anal. Chem., _35:9, October, 1963.

63.   Goguel, R., Anal. Chem., 41:8, 1034-1038, July, 1969.

64.   Rebel, J., ALCHE Symp. Ser. 700137, Recent Adv. Air Pollut. Control,
      p. 299-306  (1974).

65.   Dubois, L., Baker, C. J., Teichman, T., Zdrojewski, A. , and Monkman,
      J. L., Microchimica Acta  [Wjen], 269 (1969).

66.   Scaringelli, F. P., and Rehme, K. A., Anal. Chem., 4.1, 707 (1969).

67.   West, P. W., and Gaeke, G. C., Anal. Chem., 28:12, 1816 (1956).

68.'   Dharmarajan, V., Thomas, R. L., Maddalone,  R. F., and West, P. W.,
      The Science of the Total Environment, 4_, 279 (1975).

69.   Richards, L. W., "A New Technique to Measure Sulfuric Acid in
      the Air," Presented Before the Division of Water, Air and Waste
      Chemistry, American Chemical Society, Dallas, Tex., Apr. 8-13, 1973.

70.   Mudgett, P. S., Richards, L. W., and Roehrig, J.  R., "A New Tech-
      nique to Measure Sulfuric Acid in the Atmosphere," in Analytical
      Methods Applied to Air Pollution Measurements, R. K. Stevens  and
      W. F. Herget, Eds., Ann Arbor Science Publishers, Inc., Ann Arbor
      (1974).

71.   Duecker, W. W., and West, J. R., The Manufacture of Sulfuric Acid,
      Reinhold Publishing Corp., New York, N. Y.   (1959).

72.   Rehme, K. A., and Scaringelli, F. P., Anal. Chem., ^7:14,  2474  (1975)

73.   Greer, D. G., and Bydalek, T. J., Environ.  Sci. & Tech..  7:2, 153
      (1973).

74.   Thomas, R. L., Dharmarajan, V., and West, P. W., Environ.  Sci. &  Tech
      8:10, (1974).                                                     	
                                     86

-------
75.   Erdey, L., Gal, S., and Tiptay, G., Talanta, 11, 913 (1964).

76.   Grundnes, J., Christian,  S. D., Cheam, V., and Franham, S. B.,
      J. of the Am. Chem. Soc..  93;. 1, 20 (1971).

77.   Landreth, R. , de  Pena, R.  G.,  and  Heicklen, J., J. of Phys. Chem.,
      79;17, 1785  (1975).

78.   Scargill, D. , Nature,  247, 101 (1974).

79.   McKay, H.A.C.,  Atmospheric Environment, ^,  7  (1971).
                                      87

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
  EPA-600/2-77-010
                                                            I. RECIPIENT'S ACCESSIOI»NO.
 4. TITLE AND SUBTITLE
  EVALUATION OF METHODOLOGY AND  PROTOTYPE TO MEASURE

  ATMOSPHERIC SULFURIC ACID
                                                         5. REPORT DATE
                                                                    January  1977
                                                         6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(s) R.  E.  Snyder
           T.  J.  Reed
           A.  M.  McKissick
                                                         8. PERFORMING ORGANIZATION REPORT NO.

                                                                    ARC 49-5664
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Atlantic Research Corporation
  5390  Cherokee Avenue
  Alexandria, VA  22314
                                                         10. PROGRAM ELEMENT NO.
                                                                    TAD-  605
                                                         11. CONTRACT/GRANT NO.
                                                                    68-02-2247
 12.SPONSORING AGENCY NAME AND ADDRESS,  .  ,
  Environmental Sciences Research Laboratory
  Office of Research and Development
  U.S.  Environmental Protection  Agency
  Research Triangle Park, N.C.   27711
                                                         13. TYPE OF REPORT AND PERIOD COVERED
                                                                    Final 6/75-6/76
                                                         14. SPONSORING AGENCY CODE

                                                                    EPA-ORD
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT

         A method is being 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 adduct, which  is  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 temperature  eliminates particulate interferences such as
   ammonium sulfate.  Sulfur  dioxide is also collected as  a reagent complex on the
  filter but is selectively  removed by heating at 100°C.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                                                                         c.  COSATI Field/Group
  *Air pollution
  *Sulfuric  acid
  *Aerosol
   Sulfur dioxide
  *Collecting methods
                  sis
                  ry
                                                                             13B
                                                                             07B
                                                                             07D
                                                                             14B
..Chemical
"Flame PRc
^Prototypes
DISTRIBUTION S
 8. DISTRIBUTION STATEMENT

             RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
      UNCLASSIFIED
                                                                       21. NO. OF PAGES

                                                                                96
                                                                          22, PRICE
EPA Form 2220-1 (9-73)
                                            88

-------