oEPA
              United States
              Environmental Protection
              Agency
              Research and Development
              Environmental Sciences Research  EPA-600/2-80-087
              Laboratory          May 1980
              Research Triangle Park NC 27711
                      In 1IDIMR¥
                                    ffHTAL FnOTTCr.TTOTj
Development of a
Portable  Device to
Collect Sulfuric
Acid Aerosol

-------
                  RESEARCH REPORTING SERIES


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

     1.  Environmental Health Effects Research

     2.  Environmental Protection Technology

     3.  Ecological Research

     4.  Environmental Monitoring

     5.  Socioeconomic Environmental Studies

     6.  Scientific and Technical Assessment Reports (STAR)

     7.  Interagency Energy-Environment Research and Development

     8.  "Special" Reports

     9.  Miscellaneous Reports

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

-------
                                           EPA-600/2-80-087
                                           May 1980
     DEVELOPMENT OF A PORTABLE DEVICE TO
        COLLECT SULFURIC ACID AEROSOL
                 Final  Report
                     by
      Herbert C.  Miller, David W.  Mason
           and William J.  Barrett
         Southern Research Institute
           2000 Ninth Avenue South
          Birmingham, Alabama  35205
           Contract No. 68-02-2468
               Project Officer

               Kenneth J. Krost
 Atmospheric Chemistry and Physics Division
 Environmental Science 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 publica-
tion.  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 consititute endorsement or
recommendation for use.
                                      ii

-------
                             ABSTRACT

      A quantitative, interference-free method for collecting
sulfuric acid aerosol on a filter was developed and field tested.
Since previous research found that severe losses of sulfuric
acid were caused by ammonia, ambient particulate material, and
other interferents, a method was needed that converts sulfuric
acid to a stable derivative for subsequent analysis.  Methods
evaluated for direct fixation of sulfuric acid aerosol were not
found to be selective.  Therefore, a sampling method was  investi-
gated that uses a combination of selective volatilization of
the sulfuric acid, prefiltration of particulate interferents,
and derivatization of the vaporized acid on an alkali-impregnated
filter.  Research on the volatilization, prefiltration, and de-
rivatization technique, and the development and field evaluation
of a prototype sampler based on this technique are described.

      This report was submitted in fulfillment of Contract No.
68-02-2468 by Southern Research Institute under the sponsorship
of the U.S. Environmental Protection Agency.  This report covers
the period from September 1976 to September 1978.
                               iii

-------
                             CONTENTS
Abstract
Figures
Tables
     1.  Introduction ....................    1
     2.  Conclusions and Recommendations ..........    3
     3.  Summary of Results .................    5
              Results of Interference Studies with
                Sulfuric Acid Collected on Filters  .....    5
              Results of Experiments Collecting Sulfuric
                Acid as a Derivative of Perimidylam-
                monium Bromide  (PDA-Br) ...........    6
              Results of Experiments in Which Sulfuric
                Acid was Collected after Volatilization
                and Pref iltration ..............    6
              Development of a  Prototype Sampler  for
                Sulfuric Acid Aerosol ............    7
              Results of Field  Evaluation of  the
                Prototype Sampler ..............    7
     4.  Experimental Apparatus and Procedures  for
           Collection of Sulfuric Acid  Aerosol  by
           Conventional Methods ...............    9
              Sulfuric Acid Aerosol Generator ........    9
              Analytical Methods ..............    18
              Filter Materials  ...............    22
     5.  Interference Studies Associated with the
           Collection of Sulfuric Acid  Aerosol  .......    29
              Collection of Sulfuric Acid Aerosol ......    29
              Interference  from Ammonium Salts  of
                Sulfuric Acid ................    35
              Interference  from Ambient Particulate
                Material — Analysis  Corrected  for
                Interference  from Ammonium  Salts  ......    35
     6.  Experimental Apparatus and Procedures  for
           Collection of Sulfuric Acid  Aerosol  by
           Derivatization Methods ..............    40
              Analytical Methods ..............    40
              Derivatization  with Perimidylanunonium
                Bromide ...................    46
              Fixation of  Sulfuric  Acid Aerosol With
                Versapor Filters ..............    49
                                v

-------
                   CONTENTS

Volatilization, Prefiltration, and Deriva-
  tization of Sulfuric Acid Aerosol 	   51
     Preliminary Apparatus with a Sample Flow
       Rate of 3 1/min	   51
     Modified Apparatus with a Sample Flow
       Rate of 14 1/min	   56
Development of Prototype Sampler for
  Sulfuric Acid Aerosol	   65
     Interference Studies 	   65
     Assembly of the Prototype Sampling
       Apparatus	   72
     Final Laboratory Evaluation of Prototype
       Sampling System	   76
Field Evaluation of Prototype Sampler for Sulfuric
  Acid Aerosol	   79
                      VI

-------
                           FIGURES
Number                                                   Page

  1     Schematic of sulfuric acid aerosol generator  .    10

  2     Plot of slope of cumulative number distribution
          vs. particle diameter for H-jSO,, generator . .   15

  3     Plot of cumulative volume  (mass) distribution
          vs. particle diameter for H2SOi» generator . .   16

  4     Plot of slope of cumulative volume distribution
          vs. particle diameter for H2SO^ generator . .   17

  5     Flame photometric detector apparatus for
          H2SOU analysis	   20

  6     FPD calibration for H2SC\ deposited on filters
          and volatilized at 200°C	   23

  7     Calibration of modified FPD for H2SO1| deposited
          on filters and volatilized for analysis at
          200°C	   24

  8     Calibration of modified FPD for H2SO^ deposited
          on filters and volatilized for analysis at
          150°C	    25

  9     Remodified  flame photometric detector apparatus
          for H2SOi» and PDA-SOu analysis	   41

  10     FPD  calibration for PDA-SO^ decomposed at 450°C   43

  11     Ion  chromatograph calibration  with  standard
          solutions of Na2SO1|  in  H2O	   45

  12     Apparatus for volatilization,  prefiltration,
          and derivatization of sulfuric  acid aerosol .   53

  13     Apparatus for removal  of  interfering  gases
          and volatilization,  prefiltration, and
          derivatization of sulfuric acid aerosol  ...   67
                              Vll

-------
                           FIGURES
Number                                                   Page
 14     Apparatus for the removal of basic and acid
          interfering  gases and the volatilization,
          prefiltration, and derivatization of sulfuric
          acid aerosol	   73
 15     Prototype apparatus for volatilization, pre-
          filtration, and derivatization of sulfuric
          acid aerosol	   74

 16     Gas diffusion denuders	   75

 17     Teflon filter holder   	   75

 18     Ambient roadsite sampling with prototype
          sampling system 	   75

 19     Schematic of Teflon filter holder containing
          the Teflon prefilter and the alkali-impreg-
          nated Teflon collection filter  	   77

 20     Hourly average concentrations of total sulfate
          and sulfuric acid on six days beside a city
          street	   81

 21     Remote sampling at interstate highway  	   82

 22     Ion chromatographic analyses of collected
          samples	   82

 23     Hourly concentrations of total sulfate and
          sulfuric acid on August 23, 1978, at an
          interstate highway   	   85

 24     Hourly average concentrations of total sulfate
          and sulfuric acid on three days at an inter-
          state highway	   86
                              viii

-------
                            TABLES
Number                                                   Page

  1     Aerosol Generator Standard Operating Parameters   12

  2     Sulfuric Acid Collected from the Generator as a
          Function of Sampling Time	   12

  3     Determination of S02 Concentration in the HjSO^
          Aerosol Generator with the West-Gaeke Method    14

  4     Data for the Calibration of the FPD Apparatus .   22

  5     Comparison of Parallel Fluoropore and Mitex
          Filters in Sampling Sulfuric Acid Aerosol . .   26

  6     Effect of Predeposited Fe203 upon Sulfuric Acid
          Collected on Mitex Filters  	   30

  7     Effect of Predeposited PbO upon Sulfuric Acid
          Collected on Mitex Filters  	   31

  8     Effect of Predeposited Ambient Particulate
          Material upon Sulfuric Acid Collected on
          Mitex Filters	   33

  9     Effect of Ambient Particulate Material on FPD
          Measurement of Sulfuric Acid Predeposited on
          Mitex Filters	   34

 10     Interference from Ammonium Salts of Sulfuric
          Acid in FPD Analysis	   36

 11     Results of FPD Measurement of Sulfuric Acid
          Collected on Mitex Filters following Collec-
          tion on Ambient Particulate Material   ....   37

 12     Results of FPD Measurements of Sulfuric Acid
          Collected on Mitex Filters prior to Collec-
          tion of Ambient Particulate Material   ....   39

 13     Preliminary Calibration Data for Ion Chromato-
          graphic Determination of Sulfate   	   44
                              IX

-------
                             TABLES
Number                                                   Page
  14     Derivatization of Sulfuric Acid Aerosol with
          PDA-Br in the Presence of Ambient Particulate
          Material	   48

  15     Evaluation of Selectivity of  PDA-Br with
          Aerosols of Sulfate  Salts  	   50

  16     Passage of Sulfuric Acid through  the  Heated
          Glass Sampling Probe	   52

  17     Passage of Sulfuric Acid through  the  Heated
          Teflon-lined Sampling Probe 	   54

  18     Passage of Volatilized Sulfuric Acid  Aerosol
          through Mitex Filters as a  Function of
          Temperature	   55

  19     Retention of Sulfate Salt Aerosols by a Heated
          Mitex Filter	   55

  20     Derivatization Efficiency of  Alkali-impregnated
          Filters for Sulfuric Acid Vapor at  110°C   .  .   57

  21     Volatilization Efficiency of  the  Heated Probe
          at  a Temperature of  120°C and a Flow Rate
          of  14 1/min	   59

  22     Prefiltration Efficiency of Mitex Filters at a
          Temperature of 130°C and a  Flow Rate of
          14  1/min	59

  23     Efficiency of Sulfuric Acid Derivatization by
          Alkali-impregnated Teflon Filters at a Flow
          Rate of 14 1/min	60

  24     Efficiency of Sampling Method for Sulfuric
          Acid Aerosol	64

  25     Sulfur Dioxide Removal by Gas Diffusion Denuder   68

  26     Sulfur Dioxide Removal by Gas Diffusion Denuder   69

-------
                             TABLES
Number                                                   Page
  27      Passage of Sulfuric Acid Aerosol through Gas
            Diffusion Denuder 	   70

  28      Passage of Sulfuric Acid Aerosol through the
            Dehumidifier  	   71

  29      Study of Interference from Ambient Particulate
            Material	   72

  30      Efficiency of Prototype Sampling Apparatus for
            Various Concentrations of Sulfuric Acid
            Aerosol	   78

  31      Ambient Sulfuric Acid Aerosol Determinations
            beside a City Street	   78

  32      Ambient Sulfuric Acid Aerosol Determinations
            from Remote Interstate Highway Sampling  .     84
                               XI

-------
                             SECTION  L

                           INTRODUCTION
      The increasing use of sulfur-bearing coal for the produc-
tion of electrical energy, the generation of significant quanti-
ties of sulfuric acid aerosol by catalytic converters in auto-
mobiles, and the appearance of new information on the irritant
effects of sulfuric acid and sulfates are factors that have com-
bined to intensify interest in the measurement of ambient sul-
furic acid aerosol.  The emission of sulfuric acid by automobiles
is a problem of special interest because of the fairly high con-
centrations found  in the immediate vicinity of roadways and the
small size of the  sulfuric acid particles generated.

      In a series  of measurements made with catalyst-equipped
automobiles at the General Motors Test Track in October, 1975,
it was  found that  most of the sulfur in gasoline was oxidized
to sulfur trioxide and appeared as sulfuric acid aerosol near
the test site.  Concentrations in the range of 3 to 6 yg/m3 were
found.  The sulfuric acid was apparently neutralized or partially
neutralized by ambient ammonia within a fairly short time and
within  a fairly short distance of travel.  Nevertheless, the
possible harmful effect on the health of persons in the vicinity
of heavily traveled roadways continues to be a matter of serious
concern.

      The potentially harmful health effects stem from certain
unique  properties  of sulfuric acid aerosol.  Sulfuric acid  is
the most severe bronchial and lung irritant among the various
species of sulfur  compounds that can occur  in  the atmosphere.
The particles derived from automobiles have a  mass mean diameter
of only a few hundredths of a micrometer and therefore can  pene-
trate deeply into  the lung and be deposited in the alveoli.
These small particles are relatively highly concentrated  in terms
of the  molarity of the sulfuric acid and are therefore stronger
irritants than particles  that have grown  in size over a period
of time by coagulation and adsorption of water vapor.

      In order to  assess  the health effects potential of  sulfuric
acid having the properties described,  it  is necessary  to  have
sampling and analytical methods that will measure  sulfuric  acid
concentrations without  interference or artifacts arising  from

-------
the presence of other substances in the sampled atmosphere or
from various environmental factors such as relative humidity
or temperature.  Because many interfering reactions are possible,
especially during the collection of a sample, the accurate de-
termination of sulfuric acid aerosol is a difficult problem that
is just now beginning to be adequately solved.

      Although sampling cannot be arbitrarily isolated from the
method of analysis subsequently used to determine the species
sought, there are certain difficulties peculiar to sampling
sulfuric acid aerosol.  In general terms, these difficulties
may be said to result from interactions of sulfuric acid with
other particulate and gaseous constituents of the ambient atmo-
sphere or the sample and from the simultaneous presence in the
atmosphere of sulfate salts and sulfur dioxide.  Reactions of
sulfuric acid with copollutants have been shown to occur on
filters, the collection medium that is generally best suited
for sampling atmospheric particulate material.  The oxidation
of sulfur dioxide to sulfate also occurs on  filters under certain
conditions.  These phenomena—and others that may be expected
to occur—often result  in failure of attempts to measure ambient
concentrations of sulfuric acid because some or all of the acid
is lost during the sampling process.  Furthermore, methods of
analysis that involve dissolving the sulfuric acid along with
other soluble constituents of the sample introduce errors re-
sulting from reactions  that occur in the solvent medium; a sam-
pling method that will  permit subsequent processing and analysis
of the sample without loss  (or addition) of  sulfuric acid is
therefore desirable.

      Work on this contract was concerned principally with pro-
viding a sampling and analytical method that will avoid errors
resulting from degradation of sulfuric acid  during or after  its
collection on a filter.  During the early months of the contract
period, efforts were made to evaluate the collection of sulfuric •
acid aerosol and effects of interferents using  flame photometric
detection.  However,  it was later concluded  that the fixation
of sulfuric acid as a stable, nonreactive product is clearly
a desirable approach, and may well provide the  only means of
avoiding interference effects.

      This Final Report describes the techniques used to gener-
ate the sulfuric acid aerosol, the analytical methods used to
measure sulfuric acid,  and  the prototype apparatus and procedures
for derivatization of sulfuric acid both as  the aerosol and  as
the prefiltered vapor.  Further, it gives the results of labora-
tory experiments designed to determine the effects of various
potential interferents  on the collection of  sulfuric acid, the
results of field sampling with a prototype sampler, and a summary
of the conclusions reached.

-------
                             SECTION  2

                  CONCLUSIONS AND  RECOMMENDATIONS
      The collection of sulfuric acid aerosol on Teflon filters,
as a first step in the measurement of ambient sulfuric acid con-
centrations, is subject to interference by alkaline gases and
by particulate material in the air sample.  Substantial losses
of sulfuric acid are caused by gaseous ammonia and ambient par-
ticulate material from an urban environment.  Since the deter-
mination of atmospheric sulfuric acid aerosol collected on a
filter over a period of time is subject to error from  interfering
reactions that may occur during sampling, the need is  evident
for stabilization of sulfuric acid during sampling.  Direct
methods of fixation of sulfuric acid aerosol were not  found to
be selective; it appeared that a better method of sampling would
involve selective volatilization of the sulfuric acid  and filtra-
tion of the ambient particulate interferents prior to  derivatiza-
tion of the vapor.

      A prototype sampler—based on the concept of volatiliza-
tion, prefiltration, and derivatization of sulfuric acid—was
constructed and demonstrated to be of practical value  for deter-
mination of ambient sulfuric acid aerosols in both laboratory
and field studies.  This methodology in conjunction with  sulfate
determinations by ion  chromatography has  the following advantages

      • Minimal losses of sulfuric acid during  sampling—less
        than 15%.

      • Good specificity—freedom frcm  the usual  sources  of  inter-
        ference in determinations of sulfuric acid,  i.e., other
        sulfur species.

      • Good sensitivity—with  a conservative limit of detection
        of  0.5 yg of  sulfuric acid per  cubic meter of  ambient
        air.

      • Excellent stability  of  collected  samples  for  1C  analysis
         in  the laboratory at a  later date.

      The following recommendations would provide a  logical ex-
tension of  this work:

-------
• Further  improve  the  conditions and apparatus for volatili-
  zation and  prefiltration  of  sulfuric acid,  emphasizing
  the  effects of  volatilization temperature,  sampling flow
  rate, and nature of  exposed  surfaces within the sampling
  device.

• Conduct  a systematic evaluation of other derivatizing
  agents and  filter media impregnated with the derivati-
  zing agents to  achieve optimal fixation of  the volati-
  lized sulfuric  acid  so that  the derivative  can be readily
  identified  and  quantitated.

• Investigate in  depth the  potential for  interference from
  ambient  gaseous  compounds and ambient particulate material,
  especially  various size aerosols of ammonium sulfate
  salts, during sampling of sulfuric acid aerosol with
  the  combined volatilization,  prefiltration, and derivati-
  zation methodology.

-------
                             SECTION  3

                        SUMMARY OF RESULTS
      The most significant results of the experimental studies
conducted during this contract are summarized in the following
paragraphs.  Details are given in the latter sections.


RESULTS OF INTERFERENCE STUDIES WITH SULFURIC ACID COLLECTED
ON FILTERS

      • Sulfuric acid aerosol was collected on Teflon  filters
        with prederosited Fe203.  Average recoveries of  sulfuric
        acid samples were 18% for "high" levels and 79%  for "low"
        levels of Fe203, as measured by flame photometric  detec-
        tion  (FPD) at 200°C.

      • Sulfuric acid aerosol was collected on Teflon  filters
        with predeposited PbO.  Average recoveries of  sulfuric
        acid samples were 19% for "high" levels and 66%  for "low"
        levels of PbO as measured by FPD at 200°C.

      • Ambient particulate material was collected on  Teflon
        filters for periods of 3 and 8 hr at 6 1/min;  the  filters
        were  then used  to collect sulfuric acid aerosol.   For
        both  levels of  particulate loading the losses  of the
        acid  approached 90% as measured by FPD with volatiliza-
        tion  at 200°C.

      • Sulfuric  acid aerosol was collected on Teflon  filters;
        the filters were  then used to  sample ambient particulate
        material  for 1  hr at  14 1/min.  The average loss of sul-
        furic  acid was  93%  as measured by FPD at  150°C.

      • Aerosols  of ammonium  sulfate salts were  found  to be a
        source of positive  interference  in the determination
        of  sulfuric acid  by FPD with volatilization at 200°C.
        Lowering  the FPD  volatilization  temperature to 150°C
        reduced the contribution  from  the ammonium salts but
        also  adversely  affected the  sensitivity  of detection
        of  sulfuric acid.

-------
RESULTS OF EXPERIMENTS COLLECTING SULFURIC ACID AS A DERIVATIVE
OF PERIMIDYLAMMONIUM BROMIDE  (PDA-Br)

      • Approximately 100 yg of ambient particulate material
        was collected on Teflon filters impregnated with  the
        PDA-Br reagent.  The filters were used to collect  sul-
        fur ic acid aerosol and then analyzed by FPD with  thermal
        decomposition at 450°C.  The results suggested  that
        sulfuric acid was completely recovered.

      • Aerosols of the ammonium sulfate salts were collected
        on Teflon filters impregnated with PDA-Br and analyzed
        by FPD at 200 and 450°C.  No response was found at 200°C,
        but the response found at 450°C indicated that  PDA-Br
        was not selective for sulfuric acid since it appeared
        that the ammonium salts were also derivatized.
RESULTS OF EXPERIMENTS IN WHICH SULFURIC ACID WAS COLLECTED AFTER
VOLATILIZATION AND PREFILTRATION

      • Sulfuric acid aerosol was volatilized at a  flow  rate
        of 3 1/min in a heated sampling probe at 110°C.   The
        vapors were then passed through a Teflon filter  heated
        to 120°C.  At least 90% of the resulting vapor passed
        the heated filter.  The barium chloranilate (BCA)  method
        was used for determination of sulfuric  acid as sulfate.

      • Ammonium sulfate aerosols were sampled  under  the same
        conditions with the heated sampling probe.   BCA  analysis
        showed that an average of 85% of the ammonium salts were
        retained by the heated  (120°C) prefilter.


      • Volatilized and prefiltered  sulfuric acid was passed
        through filters impregnated  with PDA-Br, sodium  carbo-
        nate, or sodium hydroxide at 3 1/min.   Very little of
        the sulfuric acid vapor was  collected by the  PDA-Br
        filters but an average of 60% was fixed by  the sodium
        carbonate filters and 80% by the sodium hydroxide filters.

      • An increase in the flow rate to 14 1/min was  achieved
        with a new probe.  The overall efficiency of  the volatili-
        zation and prefiltration was lowered to 70% from a pre-
        vious overall efficiency of  80% at 3 1/min.   The efficiency
        was later increased to 80% by increasing the  volatiliza-
        tion temperature to 130°C.

-------
        The amount of sulfuric acid derivatized after volatili-
        zation and prefiltration at 14 1/min and 130°C was in-
        creased to greater than 90% using a disk-membrane Teflon
        filter impregnated with approximately 60 mg of sodium
        hydroxide.  The filters were extracted ultrasonically
        and the sulfate concentration determined by ion chro-
        ma tography.

        Ammonium sulfate aerosols were sampled with the heated
        sampling probe of 14 1/min and 130°C.  Ion chromatographic
        analysis showed that greater than 99% of the ammonium
        sulfate salts were retained by the prefilter.
DEVELOPMENT OF A PROTOTYPE SAMPLER FOR SULFURIC ACID AEROSOL

     •  The sampling apparatus for sulfuric acid aerosol based
        on the volatilization, prefiltration, and derivatization
        methodology was incorporated into a prototype sampler.

     •  A gas diffusion denuder was used to remove more than
        99% of the sulfur oxide gases from the sampled airstream
        since these gases could be converted  to sulfate by  the
        alkali-impregnated filter.

     •  A second denuder was used to remove like amounts of
        ammonia from the sampled airstream to prevent the re-
        action of ammonia with the volatilized sulfuric acid
        to form sulfate salts which could be  removed by the
        prefilter.

     •  A dehumidifier was necessary to prevent the build-up
        of water vapor in the two denuders.

     •  The overall efficiency of the prototype sampler collect-
        ing laboratory-generated sulfuric acid aerosol at low
        concentrations was 83%.  The minimum  detectable level
        of the prototype sampler was approximately 0.5
RESULTS OF FIELD EVALUATION OF  THE  PROTOTYPE  SAMPLER

      •  The prototype  sampler was operated  at two  local roadside
        sites  in the Birmingham, Alabama, area.  Two control
        samples were collected  with each  sample  to determine
        the amount of  total sulfate and sulfates less than 2 wm
        in diameter present at  each sampling  site.

-------
Approximately 20% of the nonsized sulfate collected be-
side a city street with an average traffic flow of ap-
proximately 600 cars per hour was found to be sulfuric
acid aerosol.

Approximately 30% of the nonsized sulfate collected be-
side an interstate highway with an average traffic flow
of 8,000 to 10,000 cars per hour was found to be sulfuric
acid aerosol.

-------
                             SECTION  4

             EXPERIMENTAL APPARATUS AND  PROCEDURES  FOR
              COLLECTION OF SULFURIC ACID AEROSOL BY
                       CONVENTIONAL METHODS
SULFURIC ACID AEROSOL GENERATOR

Operating Parameters

      A modified version of the generator described by Thomas
e_t al. * was used for laboratory studies with sulfuric acid aerosol.
The use of this system, which was constructed and used in earlier
work at Southern Research Institute  (under EPA Contract No. 68-
02-2234) , was continued because it was shown to have certain
attributes that were considered to be desirable for the purposes
of this project.  In particular, no  other generator system could
be expected to produce predominantly submicron particles, while
simultaneously permitting flexibility for the addition of poten-
tial gaseous and particulate interferents directly to the aerosol.
It also allowed easy adjustment of concentrations, particle size,
flow rates, and other parameters.

      The generator is illustrated in Figure 1.   Its principal
component was a Beckman No. 4020 atomizer-burner  situated at
the base of a 1.22-m by 99-mm i.d.Pyrex chimney.  The fuel gas
was hydrogen and the oxidizer was a  mixture of oxygen and argon.
A dilute solution of sulfuric acid was aspirated  into the flame
from a  beaker placed underneath the  burner.  The  ends of two
13-mm i.d.  glass probes  were located close together about 10
cm below the top of the chimney.  One of  these probes had a side
arm closed with a silicone  septum through which gases could be
added at known  flow rates.  The probes were connected to stain-
less steel holders for 47-mm filters or impinger-bubblers which
were followed by 6-1/min  critical orifices and a  carbon-vane
vacuum  pump.  One side of this dual  sampling system was always
used as the reference  for determination of the concentration
of the  sulfuric acid aerosol, while  the other side was used  for
the introduction of interferent gases or  for holding an  impregnated
 *Thomas,  R.L., V.  Dharmarajan,  and  P.W.  West.   Convenient Method
 for Generation of  Sulfuric  Acid Aerosol, Environ.  Sci.  Technol.,
 8(10):930-935, 1974

-------
          1
       EXHAUST
FILTER
HOLDER
            OPTIONAL
         CONFIGURATIONS
                  SAMPLE PROBE
                    SEPTUM
REFERENCE
PROBE
I
->
,->

rr




                          CRITICAL
                           ORIFICE
                        FILTER HOLDERS
                          80%
                      ISOPROPANOL
                       IMPINGER
                                                                         TO
                                                                        PUMP
                                                           CRITICAL
                                                           ORIFICE
                 PYREX
                CHIMNEY
  FLAME' I  , ,  I
 SULFURIC
  ACID
SOLUTION
                                  SAME
                                OPTIONAL
                             CONFIGURATIONS
                Figure 1. Schematic of sulfuric acid aerosol generator
                        (not to scale).
                                      10

-------
 filter  or  a  filter  containing  predeposited  particulate  material.
 The  excess aerosol  from  the  top  of  the  chimney was  exhausted
 through  a  hood.

      The  basic  operating  parameters  for  the  aerosol  generator
 and  their  typical values are shown  in Table 1.   All of  these
 parameters were  varied experimentally in  order  to find  the  opti-
 mum  set  of conditions.   Data indicating  typical amounts of  sul-
 furic acid aerosol  collected during operation of the  laboratory
 generator  are given in Table 2.

      Aerosol samples usually  containing  20 to 60 ug  of sulfuric
 acid were  collected on 47-mm Mitex  filter disks in  the  experimen-
 tal  work.  Smaller  disks (ca.  6.3 mm) were  punched  from the ex-
 perimental filters  for analysis  in  the  useful range of  the  flame
 photometric  detector (FPD).  Each small  disk  represented about
 3.3% of  the  effective area of  the sample  that is collected  on
 a  47-mm disk.  Each FPD  analysis reported for the filters was
 calculated to give  the total on  the 47-mm filter.  Triplicate
 samples  were taken  from  each filter by  punching three 6.3-mm
 disks equidistantly from the center.  The "A" and "B" samples
 in Table 2 were  obtained simultaneously  with  a parallel sampling
 arrangement.  These data showed  that  while  a  10-min sample  col-
 lected  approximately one-half  of the  sulfuric acid  that was ob-
 tained  with  the  20-min sample, the  5-min  sample collected some-
 what less  than one-fourth.   These results suggested that a  small,
 relatively constant loss of  sulfuric  acid probably  occurred in
 the  handling of  samples  prior  to analysis.   This loss was probably
 less than  10% of the 20-min  samples but  may have been as much
 as 50%  of  the 5-min samples.   The nature  of this loss of sulfuric
 acid is  uncertain but may  be due to reaction  of the acid with
 the ambient  ammonia present  in the laboratory.   Levels  of
 ambient  ammonia were previously  determined  to be typically  on
 the order  of 4 mg/m3 (see  p. 38  of this  report).

Background Contamination

Nitrogen Dioxide Concentration—
      Measurements  of the concentration of  nitrogen dioxide in
the effluent of the  sulfuric acid aerosol generator were  made
previously under different operating  parameters  for the  gener-
ator.  At  that time, argon was not used for dilution  of  the
hydrogen-oxygen flame and effluent temperatures  were  substan-
tially higher (ca.   100°C)  Levels of  nitrogen dioxide were  found
to be as high as 6  mg/m  under these  conditions  and attempts
were subsequently made to minimize this concentration.   Addition
of argon to  the flame in conjunction  with lower  pressures of
hydrogen and oxygen  resulted in  lower effluent  temperatures of
approximately 30 to 35°C and a lower  nitrogen dioxide concen-
tration of 0.11 mg/m3.
                               11

-------
              TABLE 1.  AEROSOL GENERATOR STANDARD
                      OPERATING PARAMETERS
     • Feed solution—4 x 10"3 N H2SOH

     • Positive burner pressures—Ha, 1.8 kPa  (13.5 mmHg)
                                  02, 18 kPa  (135 mmHg)
                                  Ar, 120 kPa  (900 mmHg)

     • Aspiration rate—ca. 2 ml/min

     • Aerosol concentration—ca. 500 Mg/ma

     • Sampling rate—6 1/min

     • Sampling time—5  to 20 min

     • Typical sample size—30 to 60 yg H2SOH

     • Probe temperature—30 to 40°C
           TABLE 2.  SULFURIC ACID COLLECTED FROM THE
            GENERATOR AS A FUNCTION OF SAMPLING TIME
Sampling
time •, min
20
20
10
5
H2SO* found on parallel
47-mm filters , ygj*
A
54
60
32
8
B
59
52
26
8
* Each value is the average of triplicate analyses with the FPD
technique using the square roots of the measured peak heights
on a relative scale to determine the amount of HsSOu.  These
data exhibited a relative standard deviation of typically less
than ca. 10 to 20%.  The significance of the "A" and "B" samples
is explained on the next page.
                                12

-------
      The conditions for the operation of the aerosol generator
during these measurements were the same as those given in Table 1.
The nigrogen dioxide concentration was determined by a method
based on the Griess-Saltzman reaction as described in Intersociety
Committee Method 42601-01-68T.

      The ambient level of nitrogen dioxide was measured simul-
taneously with that of the generator and was determined to be
about 0.05 mg/m3 or approximately one-half of the level present
in the effluent of the generator.  Therefore, the nitrogen diox-
ide concentration in the generator effluent was judged to be
low enough under current operating parameters for our work.

Sulfur Dioxide Content—
      Previous attempts to use the West-Gaeke method  (Intersociety
Committee Method 42401-01-69T) to determine the concentration
of sulfur dioxide in the effluent of the aerosol generator were
not successful.*  However, because different operating parameters
were  later used for the aerosol  generator,  the measurements  were
repeated.  The results of  these  experiments are shown  in Table
3.  Essentially, these results were  identical to  those obtained
earlier;  that  is, no sulfur  dioxide  could be detected.  Moreover,
as before, spikes of sulfur  dioxide  added to the  tetrachloro-
mercurate absorbing solution either  before  or after  sampling
the generator  effluent could not be  recovered.  Thus,  it appeared
that  the  tetrachloromercurate-sulfur dioxide complex was decom-
posed by  some  component of  the effluent  produced  by  the aerosol
generator.   Due to  the fact  that sulfuric acid was  eliminated
as the cause of the interference (second  entry  in Table 3),  no
further explanation for this phenomenon  can be offered.  However,
based on  these  results and  those of  previous work,  it  can  be
concluded that  sulfur dioxide is not stable in  the  presence  of
the gaseous  combustion products  of  the  flame and  is  probably
oxidized  rapidly  to sulfate.

Particle  Distribution

      Estimates of  the  particle-size and particle-mass distribu-
tions of  the aerosol produced by the sulfuric  acid  generator
were  made with a Model  3030 Thermo-Systems  Electrical Aerosol
Size  Analyzer  (EAA) and a modified  Climet Particle  Analyzer.
These measurements  were made with the  generator  operating  under
the  conditions given  in Table 1.  The  results  of  these measure-
ments are presented graphically  in  Figures  2  through 4.
 *Barrett,  W.J., H.C. Miller, J.E. Smith, Jr., and C.H. Gwin.
 Southern Research Institute.  Development of a Portable Device
 to Collect Sulfuric Acid Aerosol.  EPA-600/2-77-027, U.S. Environ-
 mental Protection Agency, Research Triangle Park, North Carolina,
 1977.   pp 15-16.


                                 13

-------
       TABLE 3.  DETERMINATION OF S02 CONCENTRATION IN THE
        H2SOU AEROSOL GENERATOR WITH THE WEST-GAEKE METHOD
                                                  SO
2  foundy  yg
 Standard, 20 yg of S02.  (as Na2S03)                      20

 Standard, 20 yg of SO2  (as Na2S03) and 59 yg
  of H2SOi, (standard aqueous solution)                   23

 Sample collected from H2SOi» aerosol generator*           0.4

 Sample collected from H2SOi, aerosol generator
  with spike of 20 yg of SO2 (as Na2SO3) added
  to bubbler solution after sampling*                     0.4

 Sample collected from H2SOi» aerosol generator
  with spike of 20 yg of SO2 (as Na2SO3) added
  to bubbler solution before sampling*                    0.8


 * Samples were collected for 20 min at a flow rate of 6 1/min.
      Figure 2, which is a plot of the slope of the cumulative
number distribution as a function of particle diameter,  indicates
that the majority of the aerosol particles had a diameter of
less than 0.1 ym.  On the other hand, it can be ascertained from
Figures 3 and 4, which are plots of the cumulative volume dis-
tribution and the slope of the cumulative volume distribution,
respectively, both as a function of particle diameter, that ap-
proximately 80% of the mass of aerosol particles was accounted
for by those having a diameter between 0.2 and 1.5 ym.

      It was found in an earlier evaluation of the particle dis-
tribution that more than 98% of both total mass and number of
particles was associated with those smaller than 0.3 urn.  These
measurements were made during the time when operating conditions
for the generator were such that the effluent temperature was
in excess of 80°C.  A comparison of the more recent results with
those obtained earlier suggested that the overall particle dis-
tribution of the output of the aerosol generator was a sensitive
function of the flame temperature.  As a result, steps were taken
to regulate accurately the fuel (hydrogen) pressure of the atomizer-
burner in order to control the characteristics of the effluent
aerosol.

      Although the sulfuric acid generator does not produce an
"ultrafine" aerosol with the current operating parameters, it
appears that with appropriate adjustment of fuel, oxidant, and
                               14

-------
o
     0.
10.0
          Figure 2.  Plot of slope of cumulative number distribution vs. particle
                     diameter for HySQq generator.
                                    15

-------
    600
    500
n

 o
™   400
o
 X

 UJ
o

ui
O
    300
    200
    100
       0.01
                                                   O EAA

                                                   A CLIMET
                                                    DIAMETER, Mm


                         Figure 3. Plot of cumulative volume (mass) distribution vs. particle
                                    diameter for HgSOq generator.

-------
   1000
n

 o
CM
"    100
 X

 Q
      10
     1.0

       0.01
                           OEAA

                           ACLIMET
                                               I
0.1                     1.0

      DIAMETER, /im
10.0
           Figure 4.  Plot of slope of cumulative volume distribution vs. particle
                      diameter for HySQq generator.
                                       17

-------
diluent gas pressures the output can be varied to produce an
aerosol of a given distribution over a reasonably broad range
of particle sizes.  However, the generator should be operated
within constraints necessary to avoid high concentrations of
potential interferents, such as nitrogen dioxide, and excessive
effluent temperatures, which can promote vaporization of the
rulfuric acid aerosol.

      It is probable  that the relative humidity  (RH) of the gen-
erator effluent is an important factor in establishing a given
particle distribution.  The RH of  the effluent is/  of course,
dependent on the  partial pressure  of the water vapor, which is
relatively constant,  and the temperature, which  can be varied
over a relatively wide range  (e.g_., 30 to 150°C) by changing
the pressure of the burner  fuel.   At low RH  (high effluent  tem-
peratures) the particle distribution appears  to  favor the smaller
particles; however, at higher RH  (lower effluent temperatures)
the distribution  is dominated by  larger particles.  Thus, it
appears  that the  sulfuric acid particles grow more  rapidly  in
the high RH environment during their brief  residence  time
 (<1 sec)  in the generator than under lower  RH conditions.


ANALYTICAL METHODS

      The principal analytical techniques used  in  this work were
based on the barium chloranilate  (BCA) method for  the determina-
tion of  total  sulfate and the flame photometric  detection  (FPD)
of a volatilized  sulfur species.

      The procedures  described in the  literature were modified
somewhat to meet  the  needs  of this project.   Detailed protocols
for these methods are described  in the  following paragraphs.

Barium Chloranilate Method*

      For  the  determination of sulfate  by  the barium  chloranilate
method,  a  filter  sample was placed in  a  125-ml  screw-capped
Erlenmeyer  flask  containing about 25 mg  of  bariun chloranilate
and 10 ml  of 80%  isopropanol.  The sample was then mixed  for
30 min on  a rotary  shaker,  centrifuged,  and the  absorbance  of
the resulting  supernatant solution measured at  310  nm in  a  1-cm
fused-silica cell with a Beckman  Model DU spectrophotometer.
An 80% isopropanol  solution was  used  in  the reference cell.
A linear calibration  curve  was prepared  with known quantities
 *Schafer,  H.N.S.   An Improved Spectrophotometric Method for the
 Determination of  Sulfate with Barium Chloranilate as Applied to
 Coal  Ash and Related Materials.   Anal. Chem., 3_9 (14:1719-1726,
 1967.
                                 18

-------
of standard sulfuric acid solutions that were treated in the
same manner as experimental samples.  The useful range of the
method was 10 to 50 yg of sulfuric acid per filter without dilu-
tion of the sample.

Flame Photometric Detection of Volatilized Sulfuric Acid

      Much of the work under the previous EPA contract  (No. 68-
02-2234) was intended to study the effect of selected interferent
materials on the stability of sulfuric acid collected on filter
media.  Early in that study, analyses for sulfuric acid were
conducted by selective extraction of the acid with benzaldehyde
followed by a barium-Thorin microtitration of sulfate.  However,
in the later stages of the investigation, an FPD apparatus was
constructed and successfully used to measure sulfuric acid.
With this methodology the sulfuric acid was thermally volatilized
from the filter medium and quantitated with the FPD.  Analyses
were performed more rapidly with this technique and the limit
of detection was significantly lower than that of  the "wet"
methods employed earlier.  One disadvantage of the FPD  apparatus,
however, was that the results obtained with it often possessed
poor reproducibility.

Apparatus—
      Early in the current contract  (No. 68-02-2468) efforts
were directed toward optimization of the FPD technique  as  it
is applied to our program.  First,  in order to simplify the
laboratory apparatus, the Model FPD  100AT flame photometric
detector was removed from the Meloy  Laboratories Model  SA  160-
2 Sulfur Gas Analyzer.  This move facilitated  the  use of much
shorter Teflon transfer lines from  the volatilization chamber
to the heated detector block.  High  voltage for the photomulti-
plier tube  (PMT) of the FPD was provided by a  Kepco Model  ABC1500-
M power supply, and the PMT current  was measured with a Keithley
Model 414A Picoammeter and an appropriate strip-chart recorder.

      The sample chamber  for  thermal volatilization was also
modified to improve reproducibility  and  to  simplify  the necessary
manipulations during analysis.  The  glass chamber  that  was used
previously was replaced with  a Teflon-lined glass  tube  with
significantly smaller surface area  and  "dead volume".   A  shorter
Teflon  transfer line was  also made  possible by replacing  the
aluminum heater block with a  small  tube  furnace  for  maintaining
the  temperature of  the sample chamber.   The sample purging system
was  also modified  to provide  better  control of the airflow neces-
sary  to sweep the  sulfuric acid vapors  from  the  sample  chamber
to the  FPD.  A schematic  diagram of  the  FPD apparatus  is  shown
in Figure  5.
                                 19

-------
  £
6
CRITICAL
ORIFICE
             TEFLON
             SAMPLE
             CHAMBER
             200° C
                               VENT
                            FILTER
                            DISC
                                 FPD
                                 BLOCK
                                 200° C
PMT
                                               FLOWMETERS
SAMPLE
PURGE VALVE
                                                               HV SOURCE
                                                                   PICOAMMETER
                                                                    RECORDER
        Figure 5.  Flame photometric detector apparatus for H2SO^ analysis.
                                      20

-------
Calibration with Known Amounts of Sulfuric Acid—
      Several experiments were conducted to determine the practi-
cal working range for the quantitation of sulfuric acid with
the FPD apparatus.  Data for a calibration curve were obtained
by measuring the peak height of the detector response as a func-
tion of the amount of sulfuric acid added to Mitex filters from
standard solutions in isopropyl alcohol.  In theory, the response
of the FPD to a volatile sulfur species is proportional to the
square of the sulfur mass concentration in the hydrogen-rich
flame.  As a test of this relationship and for purposes of cali-
bration, the logarithms of the peak heights were plotted versus
the logarithms of the corresponding amounts of sulfuric acid
present in the standard samples.  The theoretical slope of this
plot should be 2, and a value of 2.09 was found experimentally.
The linearity of this log-log calibration covered a useful range
of approximately 0.1 to 2 ug of sulfuric acid per sample.  Also,
a relative standard deviation of 5.6% was obtained for 16 repli-
cate analyses that were performed over a period of 2 weeks on
standard samples at the 2-ug level.  These results demonstrated
that adequate sensitivity and reproducibility could be obtained
with the FPD device.

      For practical purposes, calibration curves were constructed
from plots of the square root of detector response  (peak height)
versus micrograms of sulfuric acid.  The advantage of this plot
was that the calibration was linear in the amount of sulfuric
acid.  Typical calibration data are given in Table 4.  These
data are also shown graphically in Figure 6.

      For purposes of calibration  in early experiments the square
root of the peak  height of the FPD response provided a satis-
factory linear correlation with the amount of sulfuric acid.
However, a more  rigorous treatment was later performed by  linear-
ization and  integration of the sulfur-FPD peaks.  The linearizer
incorporated into the FPD apparatus electronically obtained  the
square root of the FPD  signal and  manual planimetry provided
integration of the linearized peaks.

      The method  of calibration for the FPD varied during  differ-
ent phases of the project.   The method used for calibration  will
be noted  in  the  tables  that  give data  from FPD determinations.

      Calibration of  the  improved  FPD  apparatus with known amounts
of sulfuric  acid  on Mitex filters  produced  the curves  shown  in
Figures 7 and 8.  The volatilization  temperature  used  to  obtain
the data  in  Figure 7 was  200°C.  The  linear range of  the  calibra-
tion was  found to be  2  to 20 yg of sulfuric acid  per  47-mm filter
disk  (an accurately cut portion of the  47-mm  filter  was  used
for the analysis).  The volatilization  temperature  for  the data
in Figure 8  was  150°C,  and  the  linear  range at  this  temperature
                                 21

-------
       TABLE 4.  DATA FOR THE CALIBRATION OF THE FPD APPARATUS

       Ha SO i* added,*        Square root of FPD response, t
       	yg	arbitrary units	

           2.05                         17.0
                                        17.6
                                        17.9

           1.23                         10.4
                                        10.8
                                        10.2

           0.82                          7.1
                                         7.0
                                         6.9

           0.33                          2.0
                                         2.2
                                         2.3
   * Standard solutions of sulfuric acid in isopropyl alcohol were
     applied to Mitex filters (6.3-mm disks)  with microburet.

   t These data are the square roots of the measured peak heights
     on a relative scale.
was 4 to 15 yg of sulfuric acid per 47-iran  filter.   The  slight
reduction in sensitivity was probably due  to  less  efficient
volatilization of the sulfuric acid from the  Mitex filter  and
some adsorption on surfaces at the lower temperature.   Poor  re-
producibility was also observed in the  data for  the calibration
curve at 150 °C for amounts less than 1 yg.   The reproducibility
at lower levels was later improved with further  instrument
modifications  (see p. 40 of this report).


FILTER MATERIALS

      Mitex and Fluoropore are two types of Teflon  (PTFE)  filters
that have been used for the collection of  sulfuric  acid aerosol.
During previous contract work several measurements  demonstrated
that Mitex filters with a 5-ym pore size rating  were 99% effi-
cient in the collection of sulfuric acid aerosols  with  particle
sizes in the range of 0.3-to 0.005 pro.  However, because Fluoro-
pore filters are frequently used by other  workers  for the  col-
lection of sulfuric acid aerosol,  a comparison was  made by sam-
pling the laboratory generator simultaneously with  Mitex and
Fluoropore filters in parallel.
                                22

-------
                        1.0           1.5
                     AMOUNT OF H2SO4, fjLg
2.0
              2.5
Figure 6. FPD calibration for H^SO4 deposited on filters and
         volatilized at 200°C.
                            23

-------
 c
     19
     15
CO

O
a.
CO
Q
Q.
LL

Q
01

<
     10
Q

<

O
UJ
N
                                         I
I
       0.0        0.1        0.2          0.3        0.4

                                AMOUNT OF H2SO4 , us
          0.5
0.6
         Figure 7. Calibration of modified FPD for H^SQj deposited on filters
                  and volatilized for analysis at 200PC.
                                       24

-------
13
u)
*•
§
e-
(o
*^
!5

-------
Filter Evaluation with FPD Determinations

      Mitex filters with a 5-ym pore size rating and Fluoro-
pore filters  (1-ym pore size) that are manufactured without
the polyethylene backing were used simultaneously to collect
svlfuric acid from the generator.  Analyses of both filters
were conducted with the FPD apparatus.  The results of the com-
parison are given in Table 5.  The greater pressure drop across
the Fluoropore filter resulted in a lower flow rate; consequently,
the amount of sulfuric acid collected was less than Fluoropore
than for Mitex.  However, the calculated results for the sulfuric
acid concentration in the generator effluent were essentially
identical for the two filters within experimental error.


       Table 5.  COMPARISON OF PARALLEL FLUOROPORE AND MITEX
            FILTERS IN SAMPLING SULFURIC ACID AEROSOL*
Filter
material
Mitex
Fluoropore
Flow rate,
1/min
6.0
5.2
H2SOi» found, t
yg
61.4
55.4
H2SOi» concn ,
yg/m3
513
533
* Sampling time was 20 min.

t Each result is the average of triplicate  analyses.   Relative
  standard deviation was typically  less  than  ca.  10  to 20%.


      These results suggested that  neither  filter offers any
particular advantage over the other for  the collection of sul-
furic acid from the aerosol generator.   However,  it  is conceiv-
able that Fluoropore filters may be more efficient in  the col-
lection of the aerosol if the operating  conditions of  the gene-
rator were adjusted to give a higher proportion of "ultrafine"
particles.

      These results were based on FPD analysis of filters with
volatilization at 200°C.  Also, these filters had collected about
60 yg of sulfuric acid from the aerosol  generator, and at that
sulfuric acid level the results obtained with both filters were
essentially the same.   On the other hand, it had been observed
earlier for smaller amounts of sulfuric  acid aerosol  (6.3., less
than about 20 yg per 47-mm disk)  that the recovery upon analysis
was somewhat less than expected for Mitex filters.   It appeared
that a small, constant amount of sulfuric acid (ca.  5 ug) was
not recovered from the Mitex filters.  This amount was not signi-
ficant at the 60-yg level and fell within the uncertainity of
                                26

-------
the FPD analyses, but at the 10- to 20-pg level the loss was
substantial (25 to 50%) .

      In subsequent work this phenomenon was found to be more
pronounced for Fluoropore filters than  for Mitex.  For example,
collection of about 15 pg of sulfuric acid aerosol on Mitex
filters (measured independently on a parallel filter by the BCA
method for total sulfate) followed immediately by FPD analysis
with volatilization at 200°C gave a result of only 8 yg or about
50% recovery.  On the other hand, only  about 6 yg of the acid
was found by FPD analysis of a Fluoropore filter that had col-
lected about 33 yg of aerosol (approximately 20% recovery).

      These FPD results reported for the comparison of Mitex
and Fluoropore filters were based on volatilization at 200°C;
however, the recoveries were even lower at 150°C—the volatili-
zation temperature necessary to minimize FPD interferences from
the ammonium salts of sulfuric acid.  For example, FPD determina-
tion of 36 yg of sulfuric acid aerosol  collected on Fluoropore
gave 12 pg at 200°C  (about 35% recovery) and 3 yg at 150°C  (about
10% recovery).  With Mitex the results  were somewhat better:
for 19 yg of the aerosol, FPD determination at 200°C gave 14  M9
(about 70% recovery) and at 150°C the result was 12 yg  (about
65% recovery).  Further, it should be noted that the determina-
tions for each filter at both temperatures were referred  to  in-
dividual calibration curves.  These curves were prepared  from
FPD responses, at the corresponding temperatures, to known spikes
of sulfuric acid added to the appropriate filters from a  solu-
tion in isopropanol.

Filter Evaluation with BCA Determinations

      It was interesting also that the  same phenomenon as de-
scribed for FPD occurred with the BCA determinations of sulfate
when the filters were not agitated ultrasonically during  extrac-
tion into 80% isopropanol.  That is, at the 70-Mg level the
recoveries were about 90% without ultrasonic extraction for  both
Mitex and Fluoropore filters spiked with known amounts of sul-
furic acid, but at the  10-pg level the  recoveries were only  about
70% for Mitex and 50% for Fluoropore.   However, with ultrasonic
extraction the low level recoveries for both filters approached
100%.  These results suggested that the sulfuric acid may have
diffused into the interior pores of the filters where it  was
difficult to remove  by  extraction or thermal volatilization.
The phenomenon appears  to involve a constant amount of sulfuric
acid and the situation  seems to be worse with Fluoropore  filters.
The use of Teflon membrane filters may  have reduced, the severity
of this effect, but these were not evaluated in this study  due
to the need for high sampling flow rates  (ca. 14 1/min).
                                27

-------
      In conclusion, neither the Mitex nor Fluoropore filter
appears to be entirely adequate for the thermal volatilization-
FPD determination of sulfuric acid.  However, the results tend
to favor Mitex filters because of the substantially higher re-
coveries that were found relative to Fluoropore.  In any event,
these results further indicate problems in the conventional
volatilization FPD analysis of sulfuric acid.
                                28

-------
                          SECTION 5

           INTERFERENCE STUDIES ASSOCIATED WITH THE
             COLLECTION OF SULFURIC ACID AEROSOL
COLLECTION OF SULFURIC ACID AEROSOL

     The effect of selected interferent materials upon the
stability of sulfuric acid collected on Teflon filter disks
studied with quantitation of the acid with the FPD apparatus.
The interferents were chosen from those representative of
a roadside environment and those which have the potential
to react with sulfuric acid.  This study essentially concluded
the interference studies that were begun under the previous
EPA contract (No. 68-02-2234) concerning sulfuric acid aerosol.

Interference from
     Experiments were conducted to measure the effect of
Fe203 on the stability of sulfuric acid collected on a Mitex
filter with analysis by the FPD method.  A similar study
was performed in previous work with analyses by the benz-
aldehyde extraction technique  (specific for sulfuric acid)
and insignificant losses of sulfuric acid were found.  However,
results of experiments with analyses by FPD indicated that
substantial losses of acid occurred for the "high" levels
of predeposited Fe203 as opposed to somewhat more moderate
losses for the "low" level.  These data are given in Table
6.  The average recoveries of  sulfuric acid sampled as
aerosol following deposition of Fe2O3 were 18% for the "high"
level and 79% for the "low" level.  The FezOs particulate
was sampled from an aerosol produced by a DeVilbiss powder
blower; no effort was made to  characterize the size distribu-
tion of this aerosol.  The two levels of Fe203 were obtained
by sampling the aerosol ten times longer for the  "high"
level than for the "low" level.

     Recoveries of spikes of sulfuric acid  (in isopropanol
solution) added to filters also depended upon the level
of predeposited Fe2O3.  Average recoveries of the spikes
were- 54%  for the "high" level  and 106% for the "low"  level.
Although  the recoveries of spikes were adequate at  the  "low"
level of  Fe203, the greater losses of sulfuric acid  for
both spiked and aerosol samples at the "high" level  could
have resulted from further reaction of the  residual  acid

                               29

-------
with the additional Fe203 on the filter at the elevated tempera-
ture (ca. 200°C)  of the FPD volatilization chamber.  Accordingly,
it is possible that the lower recoveries found with FPD analyses
as opposed to those obtained earlier with the benzaldehyde
extraction technique resulted from reaction of sulfuric acid
with Fe203 during volatization in the FPD.  However, the fact
tnat the recoveries for the aerosol samples were lower than
those of the spiked samples suggested that additional losses
of acid might have occurred during aerosol sampling.
     TABLE  6.  EFFECT OF  PREDEPOSITED  Fe203  UPON  SULFURIC
               ACID COLLECTED  ON MITEX FILTERS

Fe203
added , mg
HzSO!*
added , *
yg
H2SO.,
spike ,t
yg
H2SO.»
found , §
yg

Recovery ,
%
    High  levels

     0.3             none
     0.6             none
     0.3  (0.3)#     40
     0.5            45
     0.3            52

    Low levels
(47)
none
 63

none
none
none
34

 7.2 (44)
 6.3
10.8
54

18 (96)
14
21
< 0 . 1 none
<0.1 none
<0.1 62
<0.1 64
<0.1 64
none
63
none
none
none
<0 . 1
67
44
54
53
—
106
71
84
82
  * The  amount  of  sulfuric acid added was  determined by analysis
   of a parallel  filter exposed only to the sulfuric acid
   aerosol.

  t The  spike was  added  to the  filter exposed to FezOa from a
   standard solution  of sulfuric acid in  isopropyl alcohol.

  § Each result is the average  of triplicate FPD determinations
   calculated  from  the  square  roots  of the  peak heights.   Rela-
   tive standard  deviation was typically  less than ca.  10 to 20%.

  # The  numbers in parentheses  are the results of similar  experi-
   ments in which the analyses for sulfuric acid were conducted
   with the benzaldehyde  extraction  technique.
                                  30

-------
Interference from PbO

     The evaluation of PbO as a cause of interference during
the collection of sulfuric acid aerosol was conducted in the
same manner as described earlier for FeaOa.  The results of
these experiments are given in Table 7.  Although there was
a relatively large variability in replicate values for recovery
of acid in the presence of PbO at the "high" level, it was
evident that substantial losses of acid occurred.  The same
trend was observed with PbO as with FezOa in that the losses
of acid at the "low" level of interferent were significantly
less than at the "high" level.  These results cannot be com-
pared with previous data because PbO was not included in inter-
ference studies conducted with the benzaldehyde extraction
technique.
       TABLE 7.  EFFECT OF PREDEPOSITED PbO UPON SULFURIC
               ACID COLLECTED ON MITEX FILTERS
PbO added,
mg
High levels
0.5
0.5
0.7
0.7
H2SOi,
added, yg*
none
59
62
60
found, ygt
0.1
3.9
20
12
Recovery ,

6
32
20
    Low levels

       0.1             none             0.1

       0.1               60             40              67
       0.1               58             38              66
       0.1               57             37              65
  * The  amount  of  sulfuric  acid  added was  determined by analysis
   of a parallel  filter  exposed only to the  sulfuric acid
   aerosol.

 t Each result is the average of triplicate FPD determinations
   calculated  from the square roots of the peak heights.  Rela-
   tive standard deviation was typically  less than  ca.  10 to 20%.
                               31

-------
Interference from Ambient Particulate Material

      The effect of ambient particulate material  (APM) on the
stability of sulfuric acid collected on a filter was studied in
previous work with analyses by the benzaldehyde extraction tech-
nique.  At that time, losses of sulfuric acid approaching 70%
were  found for collection of the acid aerosol on  filters pre-
viously exposed to APM.  Because the effect of ambient particulate
is important and was studied with FPD analyses only semiquantita-
tively in our earlier work, this potential interferent was evaluated
again with the improved, more nearly quantitative FPD technique.

      The APM was collected on 47-mm Mitex filters beside a busy
roadway in Birmingham.  Because typical sampling times were several
hundred min for collection of APM and only 15 min for collection of
laboratory-generated sulfuric acid aerosol, the aerosol could not
be collected simultaneously.  Rather we chose to study the effects
of collection of sulfuric acid "early" and "late" in the ambient
sampling period.

      The results of the FPD determinations of sulfuric acid
added to the filters from the aerosol generator are given in
Table 8.  The recoveries of sulfuric acid were determined for
two levels of APM—the "high" level representing an 8-h sampling
period and the "low" level, a 3-h sample.  These data demonstrated,
again, the severe losses of sulfuric acid that can occur in the
presence of ambient particulate on a filter disk.  For both the
"high" and "low" levels of particulate, the losses of the acid
approached 90%, which was somewhat more than determined by the
wet methods of analysis that were used earlier.  The greater
losses of sulfuric acid that were determined by FPD analyses could
have resulted from further reaction of the residual acid with the
particulate on the filter at the elevated temperature (ca. 200°C)
of the FPD volatilization chamber.  In any event, the addition of
a spike of sulfuric acid in isopropyl alcohol to a filter exposed
to APM also resulted in a low recovery of the acid (ca.  18%).

      On the other hand, detectable FPD responses were routinely
observed for filters containing only the APM.  The response could
possibly be due to a sulfur species other than sulfuric acid;
however, the concentration of ambient sulfuric acid calculated
from these data was approximately 1 to 2 yg/m3.  These data are
also in approximate agreement with earlier results obtained with
wet methods.

      Experiments were also conducted in which the order of
collection of APM and sulfuric acid was reversed.  First, known
amounts of sulfuric acid from the aerosol generator were deposited
on Mitex filters, and then APM was collected, either in the labor-
atory or near a street with heavy traffic.   The experimental con-
ditions and the results of the FPD analyses are presented in
Table 9.  The recoveries of sulfuric acid were much greater than
when APM was deposited first.
                                32

-------
Essentially all of the sulfuric acid was recovered from the
filters exposed at the roadside and about two-thirds was re-
covered from the filters exposed in the laboratory.
      TABLE 8.  EFFECT OF PREDEPOSITED AMBIENT PARTICULATE
             MATERIAL UPON SULFURIC ACID COLLECTED
                      ON MITEX  FILTERS
Ambient
particulate
added , mg
added,*
yg
H2SOH
spike , t
HjSO.,
found , §
Recovery ,
0.3 (0.3)#
0.3

0.1 (2.0)
0.2
0.3
none
none

68 (29)
54
65
none
 63

none
none
none
 5.3
11.1

 7.1
 5.2
 7.6
                                              (4)
                                              (10)
18

10 (34)
10
12
    LOW
                    none
                          none
                     3.9
<0.1 60
<0.1 58
<0.1 57
none
none
none
6.4
8.5
5.4
11
15
10
  *  The  amount of sulfuric acid added was determined by analysis
    of a parallel filter exposed only to the sulfuric acid
    aerosol.

  t  The  spike was added to the filter exposed to ambient partic-
    ulate material from a standard solution of sulfuric acid in
    isopropyl alcohol.

  §  Each result is the  average of triplicate FPD determinations
    calculated from the square roots of the peak heights.   Rela-
    tive standard deviation was typically less than ca. 10 to 20%

  #  The  numbers in parentheses are the results of similar experi-
    ments in  which the  analyses for sulfuric acid were conducted
    with the  benzaldehyde extraction technique.
                               33

-------
        TABLE 9.  EFFECT OF AMBIENT PARTICULATE MATERIAL
               ON FPD MEASUREMENT OF SULFURIC ACID
                  PREDEPOSITED ON MITEX FILTERS
Sampl ing
time , *
added, t
H^SO,,
found, §
Recovery , t
         rnin
       Roadside atmosphere
            5
          10
          20
          40
          60
         180
45
45
46
41
41
 0
45
44
46
40
34
 4
100
 98
100
 98
 83
       Laboratory atmosphere
40
180
300
360
300
55
58
57
57
0
36
39
39
42
6
65
67
68
74
~m
 * Sampling rate was 6 1/min at a level of APM of about 100
   ug/m3.

 t The amount of sulfuric acid added was determined by analysis
   of a parallel filter exposed only to the sulfuric acid
   aerosol.

 § Each  result of the average of triplicate FPD determinations
   calculated from the square roots of the peak heights.  Rela-
   tive  standard deviation was typically less than ca. 10 to 20%.
   The temperature of the volatilization chamber was 200°C.

 #  Explanations  of the trends of these recoveries and those
   reported in Table 8 are set forth in the following discus-
   sion.

      The two most important factors likely to influence the
results of these experiments are neutralization of the acid
by ambient ammonia and reaction of the collected sulfuric acid
with alkaline particulate materials on the filter.  It was
expected that when the sulfuric acid was deposited first,
neutralization by ambient ammonia during subsequent collection
of the APM would cause loss of the acid.   Based on an estimate
of the amount of ammonia sampled simultaneously with the APM,
partial losses of sulfuric acid were expected for sampling
times under 80 min and total losses for sampling times substan-
tially over 80 min.   The results,  however,  were not consistent
with this expectation.  If reaction of the predeposited acid
with ammonia did occur,  then it appeared that positive inter-
                                34

-------
ference was caused by decomposition of the resulting ammonium
salts at 200°C.  Another conceivable explanation of the high
recoveries is that neutralization by ammonia did not occur,
at least not completely and that the predeposited sulfuric
acid diffused into the interior pores of the filter where it
could not react with alkaline particulate material.


INTERFERENCE FROM AMMONIUM SALTS OF SULFURIC ACID

     Several uncertainties resulted from the experiments de-
scribed above in which sulfuric acid deposited on Mitex filters
was exposed to ambient particulate material  (APM).  The princi-
pal uncertainties involved recoveries of sulfuric acid that
were higher than expected based on the measured concentration
of ambient ammonia that was sampled simultaneously with the
ambient particulate material.  These results suggested the
possibility of positive interference from ammonium salts in
the FPD analysis at 200°C.  Subsequently, a study of the inter-
ferences from ammonium sulfate and ammonium bisulfate was per-
formed by applying either aqueous or alcoholic solutions of
the salts to Mitex filters followed by FPD analysis.  The results
of these experiments are given in Table 10.   Although these data
are not comprehensive and are of limited significance,  it appeared
that the FPD response was dependent on the temperature of the
volatilization chamber.   Both ammonium salts gave positive
responses at the 200°C volatilization temperature but no signifi-
cant response was observed at 150°C with either.   Because of these
findings, the evaluation of APM as a possible interferent in the
collection of sulfuric acid aerosol was repeated with attention to
the temperature of volatilization in the FPD analysis.

INTERFERENCE FROM AMBIENT PARTICULATE MATERIAL—ANALYSIS
CORRECTED FOR INTERFERENCE FROM AMMONIUM SALTS

      The evaluation of APM as an interference in the collection
of sulfuric acid aerosol was continued with experiments in which
a cyclone was used in ambient sampling to eliminate particles
greater than about 2 ym in diameter.  A description of the cyclone
is given in a report prepared by Southern Research Institute for
the Environmental Protection Agency (EPA 650/2-74-102-a, August
1975).  Our work in the past had not included size  fractionation
of APM collected for interference studies.  However, by elimin-
ating the larger particles the evaluation of interference due  to
APM becomes more realistic in relation to some of the state-of-
the-art devices that have been developed for sampling analysis
of ambient aerosols of sulfuric acid.  Also, in these experiments
the temperature of the volatilization chamber was adjusted  to
150°C to reduce interference from ammonium salts.
                               35

-------
           TABLE 10.   INTERFERENCE FROM AMMONIUM SALTS
        _ OF SULFURIC ACID IN FPD ANALYSIS _

                    ,.         FPD response as II2SOi» ,  pg*
         Ammonium salt        - - - 2 —
           added, Ug _ 200°C _ 150°C _
             0.32t                 0.12         0
             0.325                 0.08         0
          >100#                    1.14         0

         (NH4) 2 SO i,
0.32t
0.34§
»100#
0.09
0.09
0.34
0
0
0.1
 * The values are the averages of triplicate FPD determinations
   calculated from the square roots of the peak heights.   Rela-
   tive standard deviation was typically less than ca.  10 to 20%.

 t A standard solution of the salt in 80% isopropyl alcohol was
   applied to Mitex filters.

 § A standard solution of the salt in water was applied to Mitex
   filters.

 # A small crystal of the ammonium salt was placed on the filter.


Predeposition of Particulate Material

      Table 11 gives the results of FPD measurement of sulfuric
acid aerosol collected from the laboratory generator on Mitex
filters after sampling APM.  Because of the complex nature of
the interaction of APM and sulfuric acid aerosol during sampling,
these data collectively possessed poor reproducibility.  However,
the precision of the individual measurements on a given sample
was satisfactory  (typically less than 10 to 20% RSD).

      From the data obtained with the cyclone it appeared that
the extent of reaction of sulfuric acid with APM was a function
of the concentration of total ambient particulates.  In these
experiments, higher recoveries of sulfuric acid resulted when
ambient sampling was done at the lower concentration of APM.
Sampling of APM without the cyclone did not appear to reduce
the recovery of the sulfuric acid solution compared with that
observed for experiments with the cyclone.  In any event,
these results indicated that trends observed in previous
studies of interference from APM with the FPD determination
and the benzaldehyde extraction technique are valid.
                                36

-------
    TABLE  11.   RESULTS OF FPD MEASUREMENT OF  SULFURIC ACID
         COLLECTED ON MITEX FILTERS FOLLOWING COLLECTION
    	ON AMBIENT PARTICULATE MATERIAL	
                       _____

                      aerosol
Ambient particulate
added, yg*
Without cyclone
-90 (109 yg/m3)**
~90
-90
-90
-90
-90
With cyclone tt
<50 (68 ug/m3)
<50
<50
<50
<50
<50
<90 (109 yg/m3)
<90
<90
<90
<90
<90
added ,
jjgt
none
none
none
10.1
7.0
9.2
none
none
none
15.4
11.2
12.2
none
none
none
8.3
11.5
9.4
spike,
none
13.7
13.7
none
none
none
none
13.7
13.7
none
none
none
none
13.7
13.7
none
none
none
found ,
0.6
4.0
4.0
2.0
1.6
2.2
0
11.4
8.3
5.9
4.8
6.4
0.5
7.1
5.6
1.3
1.9
1.6
Recovery ,
.
29
29
20
23
24

83
61
38
43
52
-
52
41
16
9
17
 *  The  sampling rate was  14.1 1/min for 1 hr.
 t  Analysis  of  a parallel filter exposed only  to the sulfuric
   acid aerosol was used  to determine the amount of sulfuric
   acid added.
 §  The  spike was added from a standard solution of sulfuric
   acid in isopropyl alcohol.
 ft  Each result  is the average of triplicate FPD determinations
   calculated from the square roots of the peak heights.   Rela-
   tive standard deviation was typically less  than ca.  10 to
   The  temperature of the volatilization chamber was 150 L.

**  The  numbers in the parentheses are the total ambient particu
   late levels during sampling.

tt  An uncertain fraction of the particulate material in the air
   sampled was removed by the cyclone.
                                37

-------
Predeposition of Sulfuric Acid

     Table 12 gives the results of FPD measurements of sulfuric
acid when the acid aerosol was collected before sampling APM.
Calculations based on a measured concentration of ambient ammo-
nia of 3.6 mg/m3 indicated that sufficient ammonia was present
ir the 0.83 m3 of sampled air to neutralize all of the predepos-
ited sulfuric acid.  The results of the FPD analyses for this
experiment agreed with this assumption due to the low recoveries
of sulfuric acid found with the 150°C volatilization temperature.
Thus, the higher recoveries reported earlier from identical
experiments, except for volatilization interference by ammonium
salts that were formed on the filter during the sampling of
APM.  Furthermore, the use of the cyclone for size discrimina-
tion of the APM would not be expected to alter these results.
CONCLUSION

      Although the severity of the interference from APM in the
collection of sulfuric acid aerosol appeared to be dependent
upon the level of ambient particulate, the interference was
generally sufficient in any event to complicate the accurate
measurement of the concentration of sulfuric acid aerosol by
FPD.  Of equal importance is the additional interference due to
ambient ammonia that is sampled simultaneously with the sulfuric
acid.
                              38

-------
    TABLE 12.  RESULTS OF FPD MEASUREMENTS OF SULFURIC ACID
          COLLECTED ON MITEX FILTERS PRIOR TO COLLECTION
               OF AMBIENT PARTICULATE MATERIAL*
Ambient particulate
     added, ygt	
aerosol
 added,
  ug§
spike,
 yg#
H2SO.»
found,
 1 I fT^t "if
Recovery,
    %
   35 (45 ug/m3)tt
 none
            none
35
35
35
35
35
35
none
none
13.1
8.9
9.6
10.7
13.7
13.7
none
none
none
none
6.5
6.2
0.6
0.8
0.6
0.8
47
45
5
9
6
7
 * Size discrimination of APM was not used  for these experi-
   ments .

 t The sampling rate was 14.1 1/min for  1 hr.

  § Analysis  of a  parallel  filter exposed only  to the  sulfuric
   acid aerosol was used to determine the amount of sulfuric
   acid added.

  # The spike was  added  from a standard  solution  of sulfuric
   acid in isopropyl  alcohol.

 ** Each result  is the average of triplicate FPD  determinations
   calculated from the  square roots of  the  peak  heights.   Rela-
   tive standard  deviation was  typically less  than ca.  10 to  20%
   The temperature of the  volatilization chamber was  150°C.

 tt The number in  parentheses is the total ambient particulate
   level  during sampling.
                               39

-------
                             SECTION  6

       EXPERIMENTAL APPARATUS AND PROCEDURES FOR COLLECTION
        OF SULFURIC ACID AEROSOL BY DERIVATIZATION METHODS


      In light of the results of the interference studies with
the conventional collection methods, it became apparent that
it would be necessary to collect the sulfuric acid aerosol as
a stabilized derivative.  The first derivatization technique
investigated was based on the reaction of sulfuric acid aerosol
with perimidylanunonium bromide  (PDA-Br).  Later, experiments
were performed to determine if sulfuric acid aerosol could be
volatilized during sampling and solid  interferents such as am-
monium sulfates of lower volatility removed by filtration of
the heated airstream prior to derivatization of the volatilized
sulfuric acid.
ANALTYICAL METHODS

      The principle analytical techniques used  in developing
the derivatization methods were based on the flame photometric
detection (FPD) of a volatilized sulfur species, the barium
chloranilate (BCA) method for the determination of total  sulfate,
and ion chromatography  (1C) for the determination of total sul-
fate in the presence of ions which interfere with the BCA and
FPD methods.

Modified Flame Photometric Detector Apparatus

Modifications  to  the FPD Apparatus—
      Modifications were made to the FPD apparatus described
in Section 4.  The major change was the addition to the apparatus
of the second  volatilization chamber.  Figure 9 depicts the ap-
paratus in its final stage of development.  The use of two sepa-
rate chambers  allowed the determination of sulfuric acid  at vola-
tilization temperatures from 150 to 200°C and the perimidylam-
monium sulfate (PDA-SOJ derivative at its decomposition  tempera-
ture of 450°C.  Teflon  tubing was used to replace all metal
transfer lines from the volatilization chambers to the FPD block.
                                40

-------
SAMPLE
PURGE
VALVE
           CRITICAL
           ORIFICE.
           5 ml/min
                                     TEFLON
                                     TRANSFER
                                     LINES
                                                  FPD
                                                  BLOCK.
                                                  200°C
TEFLON
SAMPLE
CHAMBER. 150°C
                                                                          RECORDER
                                                                          INTEGRATOR
                                   GLASS
                                   SAMPLE
                                   CHAMBER
                                   450°C
                              FLOWMETERS
                                                    CRITICAL
                                                    ORIFICE,
                                                    220 ml/min
                                              NEEDLE VALVE
                                                 AIR
              Figure 9.  Remodified flame photometric detector apparatus for H2S04
                       and PDA-SO4 analysis.
                                          41

-------
Calibration of the FPD with PDA-SO,,--
      Following modifications of the FPD apparatus to allow vola-
tilization at 450°C, calibration data were obtained for thermal
decomposition of the perimidylammonium sulfate  (PDA-SO.J deriva-
tive of sulfuric acid.  To obtain these data, standard  solutions
of sulfuric acid in isopropanol were added to Mitex filters im-
t,regnated with PDA-Br and the resulting derivative (PDA-SO,,)
was decomposed at 450°C in the FPD.  The calibration curve  for
PDA-SO,, is shown in Figure 10.  In subsequent experiments,  sam-
ples were collected from the sulfuric acid aerosol generator
with PDA-Br filters; the results of the FPD analyses of the
filters gave the expected concentration of sulfuric acid in the
generator effluent.  The sensitivity of the derivative  method
was found to be essentially the same as that resulting  from
conventional volatilization of sulfuric acid at  200°C.

Barium Chloranilate Method

      Use of the barium chloranilate method for  determining total
sulfate was limited during many of the derivatization experi-
ments.  The derivatization compounds used  interfered positively
with the BCA method by giving erroneous high results.   The  BCA
method is also pH sensitive and could not  be used  in the presence
of strong alkali.

Ion Chromatography

      Due to the substantial degree of interference with the
BCA method, a new method of determining sulfate  was investigated.
The ion chromatographic  (1C) method for determining water  soluble
ions first described by Small et al* eliminated  the interference
problems that had complicated the BCA method.   The 1C method
also provided greater sensitivity for sulfate  than the  BCA or
FPD methods.  After a Model 10 Ion Chromatograph had been  pur-
chased from the Dionex Corporation  (Sunnyvale,  CA), experiments
were conducted to optimize the operational parameters for  the
determination of sulfate.

Operational Parameters—
      The instrumental parameters used in  the  initial experi-
mentation with the  1C for anion determinations  were the follow-
ing :

      •  Analytical column—3-mm i.d. by 500-mm anion separator.

      •  Suppressor column—6-mm i.d. by 250-mm anion suppressor.
 *Small, H., T.S.  Stevens,  and W.C.  Bauman.   Novel Ion Exchange
 Chromatographic Method  Using Conductimetric Detection,  Anal.
 Chem., 471801-1809,  1975.
                                42

-------
cc
<
cc
H
ffi
CC
LU
oo
111
cc

o
Q.
U.

0
HI

<
CC
CJ
HI
O
Ul
N

£
<
UJ
0.2           0.4


   AMOUNT OF
                                               0.6
0.8
                                            fig
 Figure 10.  FPD calibration for PDA-SO4 decomposed at 45CPC.
                             43

-------
      •  Eluent—0.003 M NaHCO3/0.0024 M Na2CO3.

      •  Eluent flow rate—138 ml/hr.

      •  Sample volume—100-yl loop.

      With these parameters the elution time for sulfate was
approximately 14 min.  This time was later  reduced to 7 min by
increasing the flow rate of the eluent to 230 ml/hr and keeping
the other parameters constant.  Calibration data for sulfate
obtained with the parameters listed above are given in Table
13 and are shown graphically in Figure 11.
            TABLE 13. PRELIMINARY CALIBRATION DATA FOR ION
               CHROMATOGRAPHIC DETERMINATION OF SULFATE
            Sulfate taken,* ppm	Relative peak areat

                   338                        650
                                              655

                   169                        310
                                              315

                    68                        Hi
                                              110

                    34                         51
                                               51

                    17                         28
                                               28

                     6-8                        10.7
                                               10.7

                     3.4                         4.1
                                                4.3

                     0.7                         1.1
                                                1.2
           * Standard solutions of NajSO^in distilled-
             deionized H2O.

           t These data are relative peak areas in
             arbitrary units.
                               44

-------
            100               200

                     AMOUNT OF SO49", ppm
300
Figure  11.  Ion chromatograph calibration with standard solutions of
                   in
                            45

-------
      In order to minimize interferences from organic materials
and suspended particulates in sulfate determinations and to ex-
tend the life of the analytical column, a precolumn  supplied
by Dionex  was added to the 1C.  Also, the use of a shorter sepa-
rator column to decrease the analysis time and a larger  (500-yl)
sample loop to increase the sensitivity for sulfate were investi-
gated but proved impractical for our purposes.

      With a relative standard deviation of less than 2%, the
reproducibility of replicate determinations with the 1C  was
judged to be very good.  Furthermore, sulfate was determined
successfully in samples containing relatively high concentrations
of substances  (e.g../ chloride and nitrate) that interfered with
analytical methods used in earlier work.

      Disregarding interferences, it was calculated that as little
as 0.1 pg of sulfuric acid per 47-mm filter disk can be  deter-
mined with extraction with ultrasonic agitation and quantitation
with the 1C.  Results of preliminary experiments indicate, how-
ever, that a significant amount of sulfate  (ca. 0.5 yg)  can be
extracted from blank Teflon filters.  It is conceivable  that
the magnitude and variability of the sulfate blank on the filter
materials may ultimately impose the lower limit upon the deter-
mination of sulfuric acid with the 1C technique.

      Several sets of instrumental operating parameters  for the
determination of sulfate with the 1C were used during the develop-
ment of the final prototype sampling system.  These changes in
operating parameters will be noted where appropriate in  the text
that follows.
DERIVATIZATION WITH PERIMIDYLAMMONIUM BROMIDE

       In light of  the  results  of  interference  studies  using  con-
ventional collection methodologies,  it  became  apparent that  the
stabilization of sulfuric  acid during sampling would be necessary.
This approach was  investigated with  a derivatization technique
based  on the reaction  of sulfuric acid  with  perimidylammonium
bromide  (PDA-Br) first adapted to this  problem by  P.W. West  and
his associates.*   Because  this methodology was reported by the
authors in their early studies to be selective for sulfuric  acid,
the reaction was studied as  a  means  for derivatization of sulfuric
acid aerosol in our work.

       With this methodology  samples  are collected  on filters
impregnated with PDA-Br.   Ideally, only the  sulfuric acid
aerosol reacts with the PDA-Br salt  to  form  the insoluble sul-
fate before the acid can be  neutralized or otherwise destroyed by
 * Thomas,  R. L.; V.  Dharmarajan,  G.  L.  Lundquist,  and P.  W.  West.
 Measurements of Sulfuric Acid Aerosol,  Sulfur Trioxide,  and  the
 Total Sulfate Content of the Ambient Air.   Anal. Chem. 4_8(4) :
 639-641,  1976.
                                 46

-------
other reactions.  For quantitation of the PDA derivative, PDA-SO,,
is decomposed at 450°C for FPD measurement of a volatilized sul-
fur species.

      With regard to the speed of the derivatization reaction,
preliminary work indicated that the fixation process was very
rapid.  For example, when Mitex filters impregnated with PDA-
Br were used to collect sulfuric acid aerosol and were quickly
analyzed  (within 1 min of sampling) by FPD with volatilization
at either 200 or 450°C, the expected FPD response was obtained
at 450°C, but no response was observed at 200°C.  This result
indicated that  the sulfuric acid aerosol reacted quickly and
completely to form the more refractory PDA-SO,,.

Interference from Ambient Particulate Material

      In  further experiments ambient particulate material was
evaluated as a  cause of interference in the derivatization pro-
cess during sample collection.  In these experiments, Mitex  fil-
ters impregnated with the PDA-Br reagent and plain  filters  (not
impregnated) were used to sample the ambient atmosphere.  Deter-
mination  of the derivatized sulfuric acid was performed  with
the FPD apparatus with decomposition of the derivative at 450°C.
The results of  these experiments are given  in Table 14.

      Approximately 100 \ig of ambient particulate material  (<2 pm)
was collected on each filter.  Known amounts of  sulfuric
acid aerosol from the generator were then deposited on the  filters
and analyzed by the FPD method with volatilization  at 450°C.
The amounts of  sulfuric acid found were estimated from separate
calibration curves  (one for PDA-SC\ and one for  H2SC\).  Because
the slope of the calibration curve for sulfuric  acid at  450°C
is greater than for PDA-S(\, it appeared  that a  given amount
of acid yields  a smaller FPD signal when  it is converted to  the
PDA-SO,, derivative.

      Although  earlier experiments had shown that an attenuated  FPD
response  is obtained when nonimpregnated  filters exposed to  ambient
particulate material and  sulfuric  acid are  heated at 150 to  200°c,
these results indicated complete recovery of sulfuric acid  at
450°C  (thermal  decomposition of the reaction product of  sulfuric
acid and  ambient particulate material).  Complete recovery  was
also obtained from  the PDA-impregnated filters.  Thus  it could
not be concluded from these results that PDA-SO,, was formed  selec-
tively in the presence of APM.  The results merely  showed that
whatever  the products of  the reactions on the filters may be,
the acid  is recoverable from treated or untreated filters at
450°C, but not  at lower temperatures.  However,  the smaller  FPD
responses for PDA-Br filters  (see FPD responses  listed  in Table
14) were  initially  taken  as a favorable  indication  that  the
                                47

-------
 TABLE 14.   DERIVATIZATION OF SULFURIC ACID AEROSOL WITH PDA-Br
        IN  THE PRESENCE OF AMBIENT PARTICULATE MATERIAL	

                        H2SO.,       Equivalent    FPD response,
                       aerosol        H2SO4         arbitrary
                      added, yg*	found , ygt	units§
Mitex filters
impregnated with
PDA-Br -1- APM#
None
28

6
32

75
350

Plain Mitex              None           10             150
  filters + APM#          32            36             600
 * FPD analysis of the appropriate parallel filter exposed only
   to sulfuric acid aerosol served to determine the amount
   added.

 t Each result is the average of triplicate determinations cal-
   culated from linearized integrated FPD responses.  Relative
   standard deviation was typically less than ca. 10 to 20%.  The
   FPD volatilization temperature was 450°C.

 § Separate calibration curves were prepared for impregnated
   filters.

 # The amount of APM on each filter disk was about 100 yg  (particle
   size, <2 ym).
PDA-SOi, derivative may have been formed preferentially on  the  im-
pregnated filters.

      It should be pointed out  that  the ammonium  sulfate salts
present in APM would also contribute  to the  FPD response at  450°C
regardless of whether they were fixed by  the PDA-Br  reagent  or
not.  The contribution from sulfate  salts  in APM  can be seen
in the FPD responses listed in  Table  14 for  the filters to which
no sulfuric acid aerosol was added.   Also, the data  for the
filters with added sulfuric acid reflect  this interference in
that the analytical results for equivalent sulfuric  acid exceeded
the amount added.

      On the other hand, if the sulfate salts were not fixed
by the derivatizing reagent, it is possible  that  they could  be
selectively removed by volatilization at  temperatures below  the
decomposition temperature of PDA-SO,,.  Several experiments were
conducted to investigate the possibility  of  prevolatilization
of interfering sulfate salts, but the preliminary results  were
                                48

-------
inconclusive.  However, before proceeding  further with  this  ap-
proach it was decided that the most  important experiments were
those that would establish whether or not  the PDA-Br  reagent
fixed the ammonium salts of sulfuric acid.

Interference from Ammonium Sulfate Salts

      Simultaneous fixation of ammonium sulfate  salts and sul-
furic acid constitutes a severe  limitation on any methodology
that depends upon selective derivatization of sulfuric  acid
aerosol.  The selectivity of  the PDA-Br reagent  for  sulfuric
acid was evaluated in a relatively simple  series of  experiments
with aerosols of ammonium bisulfate  and ammonium sulfate.

      Aerosols of these sulfate  salts were collected  simulta-
neously on plain and reagent-impregnated Mitex filters  in a
manner that was identical to  experiments with sulfuric  acid
aerosols.  The sulfate aerosols  were generated in a  laboratory
apparatus by nebulizing dilute aqueous solutions of  the appropri-
ate ammonium salt with a Beckman No. 4020  atomizer-burner.   This
is the same device that is used  for  generation of the sulfuric
acid aerosol; however, only argon  (at a pressure of  140 kPa
(20 psig))  was used to aspirate  the feed solutions and  the flame
was not employed for generation  of the sulfate salt aerosols.

      In order to determine if the sulfate salts were fixed  on
the PDA-Br filters, FPD determinations were  made on  both plain
and impregnated filters with  a volatilization temperature of
200°C.  At this temperature large FPD responses  were  observed
for ammonium sulfate and ammonium bisulfate  aerosols  collected
on plain Mitex filters.  Thus, if the salts  were not  fixed on
the PDA-Br filters, large FPD responses would be expected.   On
the other hand, if the sulfate salts were  completely  fixed and
formed PDA-SO,, in the same manner as sulfuric acid aerosol,  then
no FPD response would be observed at 200°c due to the higher
decomposition temperature of  the PDA-SOH derivative.  The results
of these experiments are shown in Table 15.  These data indicate
that both ammonium salts were partially fixed by topochemical
reaction with PDA-Br.  The estimated degree  of fixation is about
80%.  Thus, it was concluded  that the reaction of PDA-Br to  form
the PDA-SO,, derivative is not selective for  sulfuric  acid.


FIXATION OF SULFURIC ACID AEROSOL WITH VERSAPOR  FILTERS

      In earlier work at the  Institute it was shown  that sulfuric
acid aerosol reacted with Versapor filters  (Gelman Instrument
Company) to form an insoluble product that could not  be deter-
mined by methods depending upon  extraction and that  was probably
                                49

-------
         TABLE  15.   EVALUATION OF SELECTIVITY OF PDA-Br
                 WITH  AEROSOLS OF SULFATE SALTS
                                     FPD response,     Sulfate
              aerosol       aerosol     arbitrary     derivatized,
             added,  yg*   added,  yg*	unitst, §	%
Mitex with
PDA-Br
Plain Mitex

_
81
_
81
94
—
94
"
275
250
1400
4700
80
95
_
™
 *  The output  of  the aerosol generator was measured by deter-
   mining sulfate with the barium chloranilate method and
   ammonium ion with the indophenol method.

 t  Volatilization temperature of the FPD apparatus was 200°C.

 §  FPD responses  were linearized and integrated.


a sulfate.  The filter is characterized by the manufacturer as
a glass-fiber  medium with an epoxy binder.

      In experiments with sulfuric acid aerosol it was found
that the reaction product formed on Versapor filters was not
decomposed to  a volatile sulfur species at either 200 or 450°C.
This result was evidenced by negligible FPD responses to filters
exposed to sulfuric acid and analyzed in the FPD apparatus.
Although not amenable to thermal volatilization-FPD analysis,
the filter could be useful if it could be  shown not to react
with sulfate salts.

      However, in experiments identical to those described in
the previous section for PDA-Br filters it was demonstrated  that
the Versapor filters also fixed the ammonium sulfate salts as
well as sulfuric acid  by virtue of the fact that no FPD response
was seen with  the salts collected on these filters.  Because
Versapor filters did not appear to offer any advantage over  PDA-
Br impregnated filters for fixing sulfuric acid aerosol, work
with these filters was discontinued.
                                50

-------
                           SECTION 7

       VOLATILIZATION, PREFILTRATION, AND DERIVATIZATION
                    OF SULFURIC  ACID AEROSOL
     After it was demonstrated that PDA-Br filters fixed aero-
sols of sulfate salts as well as sulfuric acid aerosols, experi-
ments were performed to determine if sulfuric acid aerosol
could be volatilized during sampling and sulfates of lower
volatility removed by filtration of the heated air stream.


PRELIMINARY APPARATUS WITH A SAMPLE FLOW RATE OF 3 L/MIN

     To facilitate the thermal volatilization of sulfuric
acid aerosol, a 30-cm section of 12-mm  (i.d.) glass sampling
probe was wrapped with a resistance heating  tape.  At a flow
rate of 3 1/min, the aerosol sample from the laboratory generator
could be heated in this manner from about 35°C to as high
as 200°C.  The sulfuric acid, thus vaporized, could be passed
through a heated filter designed to remove particulate  inter-
ferents and then onto a filter where fixation of the acid
vapor could occur.

Passage of Sulfuric Acid through Sampling Probes and Filters

     In preliminary experiments, the heated  probe and a heated
filter holder with a Mitex filter were  run simultaneously
with an identical reference sampling line that was not  heated.
The airstream was cooled in a 35-cm section  of tubing following
the filter.  It then passed into an impinger-bubbler containing
80% isopropanol to collect vapors and recondensed sulfuric
acid aerosol.  Following the sampling period, any sulfuric
acid condensed on internal surfaces behind  the filters  was
rinsed  into the bubblers.  Because of the nature of the bubbler
samples and the higher  levels of sulfuric acid used in  these
preliminary  experiments,  it proved to  be practical to  use
the barium chloranilate method  for these analyses.

     Preliminary  results indicated that although only  30%
of  the  total sample of  sulfuric acid aerosol (relative  to
the reference  side) could  be accounted  for,  over 70%  of this
amount  passed  through  the  filter at 130°c.   The  recovery  was
only slightly  increased by changing from a  stainless  steel
filter  holder  to  one  constructed of Teflon.

                                51

-------
     These results suggested that the losses of sulfuric acid
probably occurred in the heated zone of the glass sampling
probe.  This possibility was investigated further by deter-
mining the amount of sulfuric acid passing through the heated
probe as a function of temperature.  In these experiments
the sulfuric acid was collected in impinger-bubblers following
both the heated and reference probes.  The results of these
experiments are shown in Table 16.  These data showed that
for temperatures sufficient to volatilize an appreciable
portion of the sulfuric acid aerosol (ca. 90 to 130°C), most
of the resulting vapor was then lost to the walls of the heated
glass probe.

          TABLE 16.  PASSAGE OF SULFURIC ACID THROUGH
                THE HEATED GLASS SAMPLING PROBE
Temperature of
heated sample
airstream, °C
76
82
84
108
129
133
200
HiSOi,
sampled ,
yg*
122
106
103
111
117
100
115
HzSO^
found ,
ugt
129
108
105
68
31
29
33
Recovery ,
%
106
102
102
61
26
29
29
 * These are the amounts of sulfuric acid that passed through
   the unheated reference probe (ca. 30°C) run simultaneously
   with the heated probe.  Analyses for H2SOif were performed
   by the barium chloranilate method.

 t These are the amounts of sulfuric acid that passed through
   the heated glass probe.


     In an effort to improve the passage of sulfuric acid
in the heated probe, a Teflon lining was inserted in the heated
zone to reduce the reactivity of the inner surface.  This
thermal volatilization apparatus is shown schematically in
Figure 12.  An identical series of experiments was then per-
formed to evaluate this modification.  The results are given
in Table 17.  In contrast, these data showed that sulfuric
acid vapor was passed efficiently through the heated zone
with the Teflon lining.  In further experiments with this
apparatus the amount of sulfuric acid vapor passing through
the Mitex prefilter was determined as a function of the

                               52

-------
                                      TEFLON
                                      FILTER
                                      HOLDER
                            HEATING
                            MANTLE
             SECOND BACK-UP
             FILTER
                \
47-MM REAGENT-IMPREGNATED
BACK-UP FILTER
            TEFLON SCREEN
                47-MM PREFILTER
   TO PUMP
   3 l/min
CO
                   STAINLESS STEEL
                   FILTER HOLDER
                                       17 CM
                             10 MM I.D. TEFLON LINE
                                                                                HEATING ZONE
                                                                                                                  H2SO4
                                                                                                                  AEROSOL
                                  30 CM
                             Figure 12.  Apparatus for volatilization, prefiltration, and derivatization
                                        of sulfuric acid aerosol.

-------
           TABLE 17.   PASSAGE OF SULFURIC ACID THROUGH THE
                HEATED TEFLON-LINED SAMPLING PROBE
Temperature of
heated sample
air stream, °C
82
102
184
204
HzSOi*
sampled,
Mg*
90
78
69
47
K2SO»
found ,
ygt
96
71
56
35
Recovery,
%
107
91
81
75
  * These are the amounts of sulfuric acid that passed through
    the unheated reference probe (ca. 30°C)  run simultaneously
    with the heated probe; analyses for H2SOi» were performed by
    the barium chloranilate method.

  t These are the amounts of sulfuric acid that passed through
    the heated Teflon-lined probe.
volatilization temperature.  The results of these experiments
are given in Table 18.  These data indicated that sulfuric
acid aerosol was volatilized efficiently at 110°C and that
at least 90% of the resulting vapor passed the initial Mitex
filter (heated to about 120°C).

Interference from Ammonium Sulfate Salts

     After establishing that sulfuric acid aerosol could be
efficiently volatilized and prefiltered at 110°C, experiments
were conducted to determine the extent of interference, if
any, from aerosols of ammonium sulfate salts under identical
conditions.  The results of these experiments are given in
Table 19.  These data indicated that prefiltration of the
heated sample air stream effectively eliminated the sulfate
salts.  Therefore, it appeared to be feasible with this tech-
nique to eliminate interference from these particulates in
the determination of sulfuric acid aerosol.

Derivatization of Sulfuric Acid Vapor with PDA-Br

     In order to stabilize a sample collected in the field,
we originally attempted the use of a back-up filter impregnated
with PDA-Br  (in a separate, unheated filter holder) to fix
sulfuric acid aerosol that recondensed following prefiltration
                                54

-------
   TABLE 18.  PASSAGE OF VOLATILIZED SULFURIC ACID AEROSOL
      THROUGH MITEX FILTERS AS A FUNCTION OF TEMPERATURE
Volatiliza-
tion
temperature ,
°C
85
100
105
110
Amount of
aerosol
sampled, yg*
83
45
94
141
H2SO,
> found, yg
on passing
prefilter prefiltert
43
12
10
10
40
33
84
131
Portion of
H2SO,,
passing
prefilter, %
48
73
89
93
* The amount of aerosol sampled was determined with a parallel
  unheated sampling probe. All analyses for HzSO^ were per-
  formed by the barium chloranilate method.

t Sulfuric acid vapor passing the prefilter was determined with
  a back-up impinger-bubbler containing 80% isopropanol.
        TABLE 19.  RETENTION OF SULFATE SALT AEROSOLS
                   BY A HEATED MITEX FILTER
Aerosol*
NtUHSOt,
(NHH) zSOn
Amount
sampled ,
ygt
24
48
Amount found, yg
on
prefilter
16
45
passing
prefilteri
5
<4
Portion of
salt retained,
76
>90
* Temperature of the volatilization zone was 110°C.

t The amount of aerosol sampled was determined with a parallel
  unheated sampling probe.  Analyses were performed by the
  barium chloranilate method for sulfate.

§ The amount of interferent slipping the prefilter was deter-
  mined with a back-up filter followed by an impinger-
  bubbler containing 80% isopropanol.
                              55.

-------
 of the vapor.   However,  in a series of experiments it was
 consistently observed that most of the vapor condensed on
 cool surfaces after prefiltration without forming an aerosol
 that could be collected  on the impregnated filter.  Therefore,
 it was of interest to determine if the vapor itself could
 be fixed by a PDA-Br filter.   If so,  the reagent-impregnated
 filter could be mounted  behind the prefilter (with an appropri-
 ate spacer)  in the same  heated filter holder.

      Accordingly,  PDA-Br  filters were used to  sample sulfuric
 acid vapor at  about 110°C and  were analyzed by FPD to determine
 their retention of the vapor as PDA-SCU.   in these experiments
 it was found that  less than 5% of the vapor was derivatized.
 These results  suggested  that any fixation reaction relying
 upon the formation of an  insoluble sulfate derivative "(e.g_.,
 PDA-SOi,  or BaSOO  may not occur rapidly  enough with sulTuric
 acid vapor to  be useful.   It appeared to be likely that  the
 o,nly suitable  type of fixation scheme for sulfuric acid  vapor
 would involve  an acid-base reaction.   Therefore,  we began
 to evaluate  the effectiveness  of back-up filters  impregnated
 with alkaline  substances,  such as NaaCOs  and NaOHr  for re-
 action with  the vapor at  elevated temperatures.   The prelimi-
 nary results were  favorable.

 Derivatization  of  Volatilized  Sulfuric Acid Aerosol with  Alkali

      After  it was  demonstrated that sulfuric acid  vapor was
 not  efficiently derivatized by filters impregnated  with
 perimidylammonium  bromide  (PDA-Br), several experiments were
 conducted  with  Mitex  filters impregnated  with  alkaline solids,
 such as  sodium  carbonate  or sodium hydroxide.   It  was antici-
 pated that the  reaction of sulfuric acid  vapor  passing the
 prefilter  would be very rapid  on the  back-up filter impregnated
 with the strongly  alkaline material.  Results  of  experiments
 in which sulfuric  acid aerosol was sampled with the combined
 volatilization, prefiltration,  and derivatization  technique
 are  given  in Table  20.  These  data showed that  sulfuric acid
 vapor  was  adequately  fixed on  Mitex filters impregnated with
 sodium hydroxide.   A  comparable  level of  sodium carbonate
 was  somewhat less  efficient in fixing the vapor.   Although
 the  alkali-impregnated filters were much  more  efficient than
 PDA-Br filters  for  derivatization  of  sulfuric  acid  vapor,
 a  significant amount  of the acid  did  slip the  impregnated
 filters.
MODIFIED APPARATUS WITH A SAMPLE FLOW RATE OF 14-L/MIN

     After it was established that sulfuric acid aerosol could
be efficiently volatilized and then prefiltered at approximately
110°C and that more than 75% of interfering sulfate salts
                             56

-------
were removed by the prefilter after passage of the aerosols
through the heated probe at temperatures as high as 130°C,
modifications were made so that the sampling rate of the
apparatus could be increased from 3 to 14 1/min to allow
shorter collection periods for ambient levels of sulfuric
acid aerosol.
            TABLE  20.   DERIVATIZATION  EFFICIENCY  OF
                 ALKALI-IMPREGNATED  FILTERS  FOR
                   SULFURIC ACID  VAPOR  AT  110°C
 Impregnated.
   alkali*
 H2SOi,
aerosol
sampled,
  ugt
                              H2SOi»  found.,  yg
   on
prefilter
 on im-
pregnated slipping
 back-up  back-up
 filter   filter!
  Vapor
collection
efficiency
of back-up
filter, %*
 Na2CO3

 low level  of     58
 impregnation

 high level of    76
 impregnation
             12
             14
                       57
             32
                       10
    30
                        85
 NaOH
low level of
impregnation

60
53
106
12
<4
23
40
43
66
9
10
17
82
81
80
 * The alkali was impregnated onto Mitex filters from saturated
   alcoholic solutions .

 f The amount of aerosol sampled was determined with a parallel,
   unheated sample probe. All analyses were performed by the
   barium chloranilate method for sulfate and reported as equiva-
   lent
 §  Slippage was determined with a secondary back-up filter
   (reagent-impregnated)  followed by an impinger-bubbler con-
   taining 80% isopropanol.

 #  The collection efficiency is relative to the amount of H2SOi,
   vapor to which the impregnated filter was actually exposed.
                                57

-------
New Sampling Probe

     A larger sampling probe was constructed from a 1.2-m by
12.5-mm i.d. brass outer tube with a longer 1.35-mm by 9-mm
i.d.  Teflon liner.  Both ends of the Teflon tube were threaded
so that filter holders or other devices could be attached
directly.   To effect the thermal volatilization of sulfuric
,cid aerosol, an 80-cm section of the Teflon-lined brass tube
was wrapped with a resistance heating tape and then insulated.
A thermocouple was attached to the surface of the probe to
monitor the temperature of the heated zone so that overheating
of the Teflon liner could be avoided.  A second thermocouple
was inserted into the Teflon filter holder behind the impreg-
nated collection filter.  This thermocouple was used to record
the actual temperature of the heated airstream during sample
collection.

     To evaluate the modifications to the sampling apparatus,
the heated probe (120°C) and the heated filter holder (130°C)
were operated simultaneously with a reference sampling line
that was not heated.  Results of preliminary experiments with
sulfuric acid aerosol are given in Table 21.  These data in-
dicated that an average of about 85% of the volatilized sul-
furic acid successfully passed through the heated zone of the
new probe.  The data presented in Table 22 show further that
an average of about 84% of the sulfuric acid vapor from the
heated zone passed through the Mitex prefilter.  On the basis
of the data from both tables an overall efficiency of about
70% was indicated for the combined volatilization and prefil-
tration process.  In earlier experiments the overall efficiency
for the volatilization and prefiltration at a flow rate of
3 1/min was found to be in excess of 80%.  Nevertheless, the
modified sampling system appeared to be adequate.

Perivatization of Volatilized Sulfuric Acid Aerosol
with Alkali at Increased Flow Rate"

     With regard to the fixation or derivatization of sulfuric
acid on an alkali-impregnated filter  (following the prefilter),
preliminary results with the new probe, shown in Table 23,
suggested that the efficiency was significantly lower at the
increased flow rate.  Low efficiency was indicated by slippage
of a substantial portion of the acid vapor past the impregnated
filter into a back-up impinger-bubbler.  Initial efforts to
improve the efficiency by increasing the level of impregnated
alkali were only partially successful.  Furthermore, inter-
ferences with the analytical method precluded conclusive evalu-
ation of the effect of increased impregnation of alkali.

     It appeared that the face velocity through the impregnated
47-mm filters may have been too great at the 14-1/min flow
rate to allow efficient reaction of the acid vapors with the

                                58

-------
   TABLE 2].   VOLATILIZATION EFFICIENCY OF THE HEATED PROBE
    AT A TEMPERATURE OF 120°C AND A FLOW RATE OF 14 1/MIN
Amount
aerosol

Of H2SOU
sampled, yg*
152
141
143
117
Amount of fUSOi*
found , y g f
130
125
99
112
Recovery
86
89
70
96
* Determined by barium chloranilate analysis of the filter
  from the reference probe.

t Determined by barium chloranilate analysis of the impinger-
  bubbler catch following the heated probe.
     TABLE 22.  PREFILTRATION EFFICIENCY OF MITEX FILTERS
                AT A TEMPERATURE OF 130°C AND
                   A FLOW RATE OF 14 1/MIN
Amount of H2SOi»
vapor sampled,
yg*
94
104
Amount of H2SOi» found, yg Portion of
on passing t^SOij passing
prefilter prefiltert prefilter, %
17 75 82
16 88 85
* This is the amount of volatilized  sulfuric  acid  aerosol  from
  the heated zone of the  sampling probe.  All  HaSOn analyses
  were performed by the barium chloranilate method.

t The sulfuric acid vapor that passed  through the  prefilter
  was determined by analysis  of an alkali-impregnated  filter
  and a back-up impinger-bubbler.
                               59

-------
     TABLE  23.   EFFICIENCY  OF  SULFURIC  ACID  DERIVATIZATION
            BY ALKALI-IMPREGNATED TEFLON FILTERS  AT A
 	FLOW RATE OF  14  1/MIN	

  Amount of   	Amount  of  HeSOit found,*  yg	    Efficiency
    HaSO^                      on          passing       of im-
    aerosol        on      impregnated  impregnated     pregnated
  sampled, yg   prefilter     filter	filter	filter,  %

       29            8           9           12            43

     104          16           28           60            32

       60          20           <2           38            <5

       50          13           25           12            68
  * Determinations of  sulfate were made by  the  barium
   chloranilate method.  H2SOi» passing the  impregnated  filters
   was collected in an  impinger-bubbler containing  80%
   isopropanol.


impregnated alkali.   In other experiments  it was found that
higher volatilization temperatures (ca.  150°C)  also decreased
the efficiency of derivatization with the  impregnated filters.

     Because the higher flow rate is desirable for shorter samp-
ling periods with ambient sulfuric acid aerosols, an effort
was made to increase the efficiency by first increasing the
level of impregnation of alkali and then by using thicker Teflon
filters for impregnation.  Unfortunately,  evaluation of the
results of these experiments was complicated by a substantial
degree of interference with the barium chloranilate (BCA) method
for sulfate.  The interference was apparently due to increased
levels of chloride that resulted from the  neutralization with
hydrochloric acid of excess alkali extracted from the impreg-
nated filters.  However, preliminary results based on BCA
analyses of the back-up bubblers containing the isopropanol
indicated that somewhat less sulfuric acid slipped the thicker
alkali-impregnated filters.

     To obtain a more meaningful evaluation of the modifications
to the sampling system, it was necessary to develop a new method
for determining sulfate extracted from the alkali-impregnated
filters.   The major problem was due to substantial interference
with the barium chloranilate method that was previously used
to measure sulfate fixed on the impregnated filters.  Because
of this interference problem, a study of the feasibility of
determining derivatized sulfuric acid in a filter extract with
ion chromatography (1C) was begun.  In addition to offering
                                60

-------
greater sensitivity for sulfate than the BCA method, the 1C
method eliminated the interference problems that had complicated
previous experiments.

Modified Extraction and Determination Procedures for
Sulfate on Impregnated Filters

     After operational parameters were established for the deter-
mination of sulfate with the 1C, experiments were performed
to study the extraction of sulfate from alkali-impregnated Teflon
filters.  In previous work with impregnated filters, isopropanol
(IPA), rather than water, was used for extraction of sulfate
due to the hydrophobic nature of Teflon.  Although IPA readily
wets the Teflon filters and efficiently extracts sulfate, an
organic solvent was not expected to be compatible with direct
analysis by 1C.  This conclusion was substantiated in a series
of experiments in which sulfate was determined in aqueous solu-
tions containing a small amount of IPA.  These conditions ap-
proximated the process of initial wetting of the Teflon filter
with a small volume of IPA and subsequent dissolution of the
extracted sulfate and IPA with aqueous extraction.  It was
found, however, that solutions containing as little as 5% of
IPA caused deleterious swelling of the ion-exchange resin in
the analytical column of the 1C.

     In order to eliminate the need for an organic solvent to
wet the filters, a method of aqueous extraction  in which the
Teflon filters are submerged in water in small vials was found
to be very efficient when performed with ultrasonic agitation.
However, because a substantial amount of excess  NaOH was extrac-
ted from alkali-impregnated filters, only a limited number
(ca.  five) of sulfate determinations could be performed with
the 1C before it was necessary to regenerate the suppressor
column.  Neutralization of the excess alkali with HCi prior
to analysis of the filter extracts appeared to extend the re-
generation interval, although a problem was encountered with
poor resolution of the resulting large chloride  and small sul-
fate peaks.  Subsequently, it was found that adequate resolution
could be obtained with the use of two anion separator  (analyti-
cal) columns in series.  With a flow rate of 184 ml/hr and with
3-mm i.d. by 500-mm  and  3-mm i.d. by 250-mm columns installed
in the 1C, a retention time of about 14 min was  observed for
sulfate.  This arrangement facilitated 1C determination of
sulfate extracted with water from alkali-impregnated  filters
following neutralization of the excess alkali  in the  extracts.

     Experiments were subsequently conducted to  investigate
the  feasibility of direct analysis of the aqueous extract by
1C without neutralization of the excess alkali.  Therefore,
several parameters were  modified to eliminate  the need  for prior
neutralization.  The amount of KOH  impregnated on each  Teflon
filter  (temporarily  used as a substitute  for NaOH)  was  reduced

                                61

-------
by one-half to approximately 60 mg/filter.  Also, the volume
of water used for extraction was increased to 4 ml/filter.
Theses changes effectively reduced the concentration of KOH
present in the extract by 75%.

     After the above changes had been made, calibration data
for sulfate were generated with alkali-impregnated filters
spiked with known amounts of standard sulfuric acid solutions.
The spiked filters were extracted with water by ultrasonic
agitation for 10 min in 4-ml glass vials sealed with Teflon-
lined closures.  The 1C parameters used for analysis of these
extracts were the following:

     Analytical column—3 mm i.d. by 500-mm anion separator
     Supressor column--6-mm i.d. by 250-mm anion suppressor
     Eluent—0.003M NaHC03/0.0024 M Na2C03
     Sample volume—100-yl loop

With these parameters, the elution time for sulfate was approxi-
mately 12 min.

Perivatization of Volatilized Sulfuric Acid Aerosol with
Alkali at Increased Flow Rate as Measured by Ion Chromatogratony

     After a quantitative method for the extraction and deter-
mination of sulfate from the alkali-impregnated filters had
been established, experiments were continued in the evaluation
of the combined volatilization, prefiltration, and derivatiza-
tion methodology for sulfuric acid aerosol.  Specifically,
emphasis was placed on improving the efficiency of the alkali-
impregnated Teflon filters for the collection of the volatilized
sulfuric acid aerosol.  As stated, previous evaluations of the
efficiency of Teflon filters impregnated with large amounts
(=50 mg/ml) of NaOH for the collection of volatilized sulfuric
acid were inconclusive.

     In further experiments Mitex filters were replaced with
a new Teflon filter purchased from Schleicher and Schuell  (TE-
35 disk-membrane).  These filters consist of a Teflon membrane
similar to Fluoropore supported by a coarse Teflon-fiber mat.
In the current filter sequence for the combined sampling method-
ology, a Mitex prefilter is followed by a second Mitex filter
that serves as a spacer.  These filters precede the NaOH or
KOH-impregnated TE-36 Teflon filter, which is oriented with
the heavy felt-like side toward the airstream.  Because the
impregnated filter and spacer are in contact, both are extracted
for the sulfate determination.

     After calibration data were generated, sulfuric acid  aero-
sol samples were collected by the combined volatilization, pre-
filtration, and derivatization methodology.  Parallel samples
were collected  (including the unheated reference probe) at


                                62

-------
14 1/min for 10 min with the temperature of the airstream of
the heated probe maintained at about 110°C (measured behind the
filter) .  In these experiments the heating mantle that was used
previously for the filter holder was not needed to maintain the
elevated temperature.

     The results of these experiments are given in Table 24.
These data indicated that an average of about 90% of the sul-
fur ic acid aerosol that was sampled was found on the filter
media (primarily on the impregnated filter with a small portion
on the prefilter).  Therefore, the slip through the impregnated
filters could have been no more than 10% of the total sulfuric
acid that was sampled.  However, as suggested by previous ex-
periments, the sulfuric acid that was not collected on the
filters  (i.e., 10%) was probably lost in the heated zone of
the probe and never reached the prefilter or impregnated filter.
These derivatization efficiency results can be compared to  a
value of  43%  that was  found earlier  (Table 23) for the same
sampling  flow rate of  14 1/min.  This increase in efficiency
was probably  due  to  the greater surface area and more effective
level of  impregnation  afforded by the thicker  felt-like Teflon
filter  (S & S TE-36).
                                 63

-------
          TABLE 24. EFFICIENCY OF SAMPLING METHOD
                 FOR SULFURIC ACID AEROSOL
Amount of
HaSO^ aerosol
added,* yg
10.7
23.9
25.0
32.0
28.0
36.0
Amount of
found, t
On
pref ilter
2.0
4.4
4.0
3.8
4.4
5.0
ug
On im-
pregnated
filter
6.8
18.8
19.0
26.2
22.4
28.6
Collection
efficiency
of im-
pregnated
f ilter, § %
78
97
91
93
95
92
Portion of
total
f^SO,, on
impregnated
filter,* %
64
76
76
82
82
80
* The amount of aerosol added was determined with a parallel,
  unheated probe with an impregnated filter. All analyses
  were performed by 1C.

t Samples were collected at 110°C at 14 1/min for 10 minutes.

§ The collection efficiency is based on the maximum amount of
  H,SOU reaching the impregnated filters  (e_.£. /
  [28.6/(36.0 - 5.0)] x 100% = 92%).

# These figures relate the amount of H2SOH found on the im-
  pregnated filter as compared to the total H2SO^ sampled
  (e.g_. ,  [28.6/36.0] x 100% = 80%).
                               64

-------
                            SECTION  8

               DEVELOPMENT OF PROTOTYPE SAMPLER FOR
                       SULFURIC  ACID  AEROSOL


     The sampling apparatus for sulfuric acid aerosol described
at the conclusion of Section 7 was considered to have properties
appropriate for its incorporation into a prototype sampler for
sulfuric acid aerosol.  It was realized, however, that steps
would have to be taken to deal effectively with various sources
of interference in the sampler.  Thus, various interference
studies as described below were undertaken  in an effort to de-
velop the changes in design that might be needed in  the final
design of the prototype sampler.


INTERFERENCE STUDIES

     Studies were needed  to determine  the level of  interference
from gaseous copollutants that  would not be  removed  by prefiL-
tration.  Sulfur oxide gases  (specifically,  sulfur  dioxide)  cbuld
be converted to sulfate by the  alkali-impregnated filter.  Basic
gases such as ammonia  could potentially  react with  the volati-
lized sulfuric acid aerosol to  form  sulfate salts which could
be removed by the prefilter.  Furthermore,  APM collected  on  the
prefilter might react  with volatilized sulfuric acid and  act
as an additional cause of  interference

Interference from Sulfur  Dioxide

      In  the  experiments with  sulfur  dioxide, the gas was  intro-
duced into the sampling probe with  a syringe pump while room
air  was  sampled  under  normal  operating conditions of 110°C and
14 1/min.  Determinations of  sulfur  dioxide were performed with
the  West-Gaeke method,  in which Greenburg-Smith  impingers con-
taining  the  tetrachloromercurate reagent solution were  placed
in the  sampling  line  after  the  impregnated  filters.  When a
sulfur  dioxide concentration  of 75  ug/m3  was sampled through
the  impregnated  filters at either 25 or 110°C,  about 75%  of  the
sulfur  dioxide was  found  in  the back-up bubblers.   The  remaining
25%  of  the sulfur  dioxide was collected on  the  impregnated filters,
and  when the filters  were analyzed  by 1C,  sulfate  equivalent
to a sulfuric  acid  aerosol concentration of 10  pm/iir was found.
                                65

-------
Therefore, based on the results of these experiments, sulfur
dioxide appeared to be a significant cause of interference with
the method in its then-existing configuration.

     The  interference from sulfur dioxide was anticipated and
two approaches that could possibly eliminate the problem were
considered.  One method involved low-temperature aqueous extrac-
tion of sulfate in the presence of sulfite on the  impregnated
collection filter.  This method has been used successfully by
other workers to demonstrate that very little sulfite exists
with sulfate in ambient particulates.  This approach depends,
of course, upon sulfur dioxide reacting on the  back-up  filter
to form sulfite with little or no oxidation to  sulfate.  However,
because of the high alkalinity of the collection filter and  the
elevated  temperature during sampling, sulfur dioxide collected
on the impregnated filter was evidently converted  to sulfate
prior to  extraction.   Thus, this method was not promising.

     The  other method involved the use of an acid  gas "denuder"
to remove  the sulfur dioxide prior to the thermal  volatilization
of the sulfuric acid aerosol.  This denuder would  simply be  an
extension  of the existing sampling line to include a portion
of tubing  coated with an alkaline solid  (e.g_.,  NaOH) .   During
sampling,  acidic gases would quickly diffuse to the walls of
the tubing and react with the alkaline material, eliminating
the interference.  Because sulfuric acid aerosol diffuses much
more slowly, it would not react and would pass  on  to the col-
lection device unattenuated.

     Experiments were therefore conducted to evaluate  the  fea-
sibility  of  using a gas diffusion denuder device  in the sampling
line for  selective removal of sulfur dioxide.   It  was calculated
from diffusion theory that at a flow rate of  14 1/min a single
5-mm i.d.  by 800-cm tube coated with sufficient reactive alkaline
material  would remove more than 99.9% of the sulfur dioxide.
However,  for practical application of the gas diffusion denuder,
a parallel arrangement of sixteen 5-mm  i.d. by  50-cm tubes  took
the place  of the single longer tube.  This device  is similar
in concept to a denuder used by R.K. Stevens* and  coworkers  at
the Environmental Protection Agency  in  Research Triangle Park,
North Carolina, for removal of ammonia.  The  sampling apparatus,
including  the denuder, for the volatilization,  prefiltration,
and derivatization of sulfuric acid  aerosol  is  shown schemati-
cally in  Figure 13.
*Stevens, R.K.,  T.G.  Dzubay,  G.  Russworm,  and D.  Rickel.   Sam-
pling and Analysis of Atmospheric  Sulfates and Related Species,
Atmos. Environ.,  12(1-3):55-68  (1978).
                                 66

-------
ON
-J
                                                                                                       AMBIENT
                                                                                                       AIR INTAKE
                           TEFLON FILTER
                           HOLDER
                                                                                   GAS DIFFUSION DENUDER
                                          	HEATING ZONE
                                          1 m l m vi \ u n \ in rvTi \mmunmr
                                              9-mm i.d. TEFLON TUBE
                 THERMOCOUPLE
                        1. MITEX PREFILTER
                        2. MITEX SPACER
                        3. IMPREGNATED FILTER
                          (S&S TE 36-0.5 TEFLON)
                                                      80-cm -
50-cm
                                                                                     16 PARALLEL 5-mmi.d.
                                                                                     GLASS TUBES
            CYCLONE
            2-pm CUT-OFF
                                Figure 13.  Apparatus for removal of interfering gases and volatilization,
                                            prefiltration, and derivatization of sulfuric acid aerosol.

-------
     For evaluation of the effectiveness of the denuder, experi-
ments were conducted with a single tube from the sixteen parallel
tubes at an equivalent flow rate of 0.88 1/min.  Sulfur dioxide
was introduced into a mixing chamber with a syringe pump and
diluted with room air.  This mixture  (sulfur dioxide concentra-
tion, 2.5 to 25 mg/m3) was then sampled before and after passing
'.hrough the denuder tube to determine the amount of sulfur di-
oxide removed by different alkaline materials coated from aqueous
solutions onto the walls of the denuder tube.  The effluent
sulfur dioxide was collected in bubblers containing a 3% hydrogen
peroxide solution and the sulfate that resulted was determined
by ion chromatography.

     The results of experiments with an uncoated denuder tube,
one coated with sodium hydroxide, and another with sodium carbon-
ate are shown in Table 25.  Sulfuric acid aerosol  (0.5  to 1 ym
MMD)  was also sampled through the denuder coated with sodium
hydroxide to determine if any losses occurred.  When a  volume
of aerosol containing 13.6 ug of sulfuric acid was sampled, the
amount found in the effluent from the denuder tube was  also 13.6
Mg, indicating that no loss occurred.
     TABLE  25.   SULFUR DIOXIDE REMOVAL BY GAS DIFFUSION DENUDER

Denuder
coating
None
Na2C03
NaOH
*
S02 added,
mg/m3
23.7
20.8
15.7
17.1
19.1
25.6
2.5
7.9
23.5
SO 2 passing denuder
tube,* mg/m3
23.7
20.9
0.3
0.09
0.08
0.08
<0.003*
<0.003|
<0.003t
SO 2 removal
efficiency, %
0
0
98.2
99.5
99.6
99.7
>99.0
>99.6
>99.9

     *   Sulfur  dioxide  was  measured by determination of sulfate
         in  the peroxide bubblers with ion chromatography (1C).

     t   Limit of detection  for  this series of experiments
        was  about 0.003 mg/m3 of sulfur dioxide.
                               68

-------
     Thus, it was determined that the calculated efficiency for
removal of sulfur dioxide (>99.9%)  was obtained with the denuder
coated with sodium hydroxide and that the sulfuric acid aerosol
passed through this denuder with essentially no attenuation (100%
recovery).  The efficiency of a denuder coated with sodium car-
bonate was also measured because of the conversion of the hydrox-
ide to carbonate that was expected to occur with ambient carbon
dioxide.  It was subsequently determined that conversion to car-
bonate would not cause a problem since the denuder coated with
sodium carbonate operated with essentially the same efficiency
(>99.5%)  as the denuder coated with sodium hydroxide.  Although
these experiments were performed with inordinately large concen-
trations  of sulfur dioxide, these results were taken as a favor-
able  indication that the denuder would eliminate interference
from  ambient levels of sulfur dioxide.

      Upon receipt of a gas diffusion denuder manufactured by
University Research Glassware  (Carrboro, N.C.), a method of
coating the 16-parallel glass tubes was developed.  A device
was constructed to rotate the denuder while heated air was being
drawn through  the denuder to speed evaporation of the coating
solutions.  Sodium hydroxide was the substance applied for re-
moving  sulfur  dioxide.

      Experiments were conducted  to determine  the efficiency of
the acid  gas denuder for removal of  sulfur dioxide.  The  results
of these  experiments are given  in Table  26.   Experiments  were
also  conducted to determine  the  loss,  if any,  of sulfuric acid
aerosol passing  through  the  acid gas denudes:.  These data are
given in  Table 27.  The  results  of the experiments with  the acid
gas denuder  show  that 99% of the sampled sulfur dioxide  was
removed while  the sulfuric  acid  aerosol  passed unattenuated.
       TABLE 26.  SULFUR DIOXIDE REMOVAL BY GAS DIFFUSION DENUDER

      S02 added*'"I"     S02 passing denuder     S02 removal
         mg/m3            tube,* mg/m3        Efficiency, %
34.5
34.5
0.34
0.38
99.0
98.9
      *  Sulfur dioxide was measured by determination of sulfate
         in hydrogen peroxide  (3%) bubblers with  ion chromato-
         graphy (1C).

      t  SO2 was metered into a 14 1/min airstream with a
         syringe pump.
                                  69

-------
       TABLE  27.   PASSAGE  OF  SULFURIC  ACID  AEROSOL THROUGH GAS
                          DIFFUSION  DENUDER

HjjSO^ aerosol H2SOH aerosol passing
added, pg*'^ denuder, ug''"
12.2 12.5
11.3 11.1
H2SOU attenuated
by denuder, %
0
1
    *  Determined with parallel reference probe.

    t  Sulfuric acid aerosol was sampled at 14 1/min for ten
       minutes.  All analyses were done by 1C.
Denuder for Ammonia

     Interference from ammonia and other basic gases was pre-
cluded by the use of a denuder for these gases as well as for
sulfur dioxide.  A gas diffusion denuder for basic gases similar
to the one for sulfur dioxide was fabricated by coating the walls
of a second denuder with phosphoric acid rather than sodium
hydroxide.  The ammonia denuder performed with similarly favor-
able efficiencies as the acid gas denuder.

Dehumidifier Ahead of the Denuders

     Because sodium hydroxide is hygroscopic, the removal of
some water vapor from the sampled airstream by the acid gas de-
nuder was expected.  However, during only a 1-hr sampling period,
an amount of water vapor was removed from a cubic meter of ambient
air  (relative humidity greater than 85%) that was sufficient
to wash more than half the coatings from both the sodium hydroxide
and phosphoric acid-coated denuders.  To eliminate this problem,
a method of dehumidifying the sampled airstream prior to passing
through the denuders was developed.

     The diffusion dehumidifier consists of an 11.5-cm i.d. by
80-cm clear acrylic tube mounted coaxially with a 20-mm i.d.
by 80-cm porous stainless steel tube.  The space between the
two tubes is filled with approximately 5 kg of an indicating
silica gel desiccant.  The porous center tube allows the aerosol
to pass through the dehumidifier while the water vapor diffuses
into the silica gel, where it is adsorbed.  The original silica
gel desiccant functioned efficiently without needing to be re-
placed or regenerated after 50 hr of operation.
                                 70

-------
     No measurements were made to determine the actual reduction
in the humidity of the sampled airstream after passing through
the dehumidifier,  but the reduction was sufficient to eliminate
the buildup of water in the denuders while sampling at relative
humidities greater than 95%.  On the other hand, experiments
were conducted to determine the loss of sulfuric acid aerosol,
if any, upon its passing through the dehumidifier.  The results
of these experiments are given in Table 28.  These data show
that the quantity of sulfuric acid aerosol was unattenuated by
the dehumidifier.
       TABLE  28.  PASSAGE OF SULFURIC ACID  AEROSOL THROUGH
   	THE DEHUMIDIFIER	

   jSO^ aerosol   H2SOM  aerosol passing    H2SO,, aerosol  passing
   added, pg*»t     dehumidif ier,  pgt         dehumidif ier,  %
175.6
124.5
200.0
108.5
120.4
171.1
129.1
180.0
109.0
121.6
98
104
90
100
101

    *  Determined  with parallel reference probe
       Sulfuric  acid aerosol was sampled at  14  1/min  for  15 rain,
       All  analyses  were  done by ion chromatography.
 Interference  from APM

      Experiments were  also  conducted  to  investigate  the  possi-
 bility  of  interference from APM deposited  on  the  prefilter  during
 sampling.   Size-fractionated APM (<2  Mm) was  collected for  1
 hr  through  a  cyclone onto  two Mitex  filters  in  parallel  lines.
 A Teflon overlay was used  with each  filter during sampling  so
 that  the APM  was deposited  on only one-half  of  each  filter.
 With  this  arrangement  of  two parallel half-filters and a combined
 flow  rate  of  14  1/min, a  face velocity was maintained  through
 each  filter that was consistent with  normal  collection conditions.

      Laboratory  generated  sulfuric acid  aerosol was  then sampled
 at  14 1/min and  120°C  using the Mitex prefilters  preloaded  with
 APM.   Because half  of  each  filter was clean  or  unexposed, each
 filter  served as  its own  reference.   Each  prefilter  and  the cor-
 responding  impregnated filter were then  cut  in  half  and  analyzed
 individually  for sulfate  by 1C.  The  results of these  experiments
 are given  in  Table  29. These data indicate  that  the relatively
                                 71

-------
           TABLE  29.   STUDY  OF  INTERFERENCE FROM AMBIENT
                       PARTICIPATE MATERIAL     	
                                        H2SO^ found on
                    Type prefilter         impregnated
        Run              used            filter, yg

         1          Unexposed                  92.0

                    APM preloaded              88.2


         2          Unexposed                  34.4

                    APM preloaded              37.2
3
Unexposed
APM preloaded
45.6
46.2

small amount of sized-fractionated APM deposited on the prefilter
during the maximum 1-hr sampling period did not interfere signi-
ficantly with the determination of sulfuric acid aerosol.


ASSEMBLY OF THE PROTOTYPE SAMPLING APPARATUS

     After the feasibility of the combined volatilization, pre-
filtr.Ttion, and derivatization methodology together with the
removal of gaseous interferents using gas diffusion denuders
had been demonstrated, a prototype sampler was constructed.
The prototype sampler is shown schematically  in Figure 14.  The
prototype sampler is mounted on an aluminum flex-frame lattice
attached to a 2-m tall open-type relay rack.  Casters added to
both the bottom and side of the support rack  allow the apparatus
to be moved easily in either a vertical or horizontal position.
A photograph of the complete prototype sampler and supporting
framework is shown in Figure 15.

      When the apparatus is in use, ambient air enters via a
3.2-mm i.d. stainless steel tube covered with a weather cap.
(The sample is collected 1-m above ground level.)  The sample
is then passed through a cyclone to remove all atmospheric par-
ticulate matter p\PM) larger than 2 pm in diameter.  This lowers
the buildup of APM on the prefilter during sampling. After the
humidity is lowered in the dehumidifier, the  sample passes through
the two gas diffusion denuders.  The denuders are shown in Figure
16.  The first denuder is coated with phosphoric acid for re-
                                72

-------
9-mm i.d. TEFLON /
U-TUBE WITH
BEND


1(1
SULFUR DIOXIDE
GAS DIFFUSION
DENUDER




^. n nn
AMMONIA
GAS DIFFUSION
DENUDER



1 1.5-cm i.d.
SILICA
GEL —• —
DEHUMIDIFIER
20-mm i.d.
POROUS STAINLESS
WEATHERCAP 	 ^
3.2-mm i.d. JJ,
STAINLESS Tf
STEEL TUBE 	 J I

r






50







cm


2.7!

50




cm




t
E


1
r1
CYCLnwc ^>-
2-nm CUTOFF \



•
s


j
!
Pr
A





,


rm










fc
y
o
or

24
J


•err
?
j
ft
\
\
fy
\
v
V
\


~s






^
\
K
\
\
g
O

/
JA

g
Sv
1


	 9-mm i.d. TEFLON TUBE



HEATING
ZONE



\ TEFLON ) 1. MITEX PREFILTER
1 ', u" ( 2. MITEX SPACER
uni ncp I 3. ALKALI IMPREGNATED
HOLDER J TEFLON F|LTER
- — THERMOCOUPLE
~ 	 te-TO PUMP
Figure 14.  Apparatus for the removal of basic and acid interfering gases
            and the volatilization, prefiltration, and derivatization of
            sulfuric acid aerosol.
                                         73

-------
    UR D,
    i/il I USION
AMMONIA
GAS DIFFUSION
DKNUDER
                                                               '>N-I.INED
                                                           HEATING ZONE
                                                           TEMPERATURE
                                                           CONTROLLF.R
Figure 15.  Prototype apparatus for volatilization, prefiltration, and
             derivatization of sulfuric acid aerosol.
                                 74

-------
Figure 16.  Gas diffusion denuders.
Figure 17.   Teflon filter holder.
     "**!
           Figure 18.  Ambient roadsite sampling with prototype
                      sampling system.
                                   75

-------
moving ammonia and other alkaline gases, and the second is coated
with sodium hydroxide for removing sulfur dioxide and other
acidic gases.

     After the larger APM and gaseous interferents have been
removed, the airstream is then heated to 130°C  in the 9-mm i.d.
uy 80-cm Teflon-lined heating zone.  The heated sample in which
the sulfuric acid aerosol has been selectively  volatilized then
passes into a Teflon filter holder containing a series of three
Teflon filters.  The Teflon filter holder is shown in Figure
17 and schematically in Figure 19.  The first filter is a Mitex
prefilter to collect the APM and unvolatilized  sulfate salts.
The third filter is the alkali-impregnated Teflon fiber-membrane
filter for derivatizing the sulfuric acid vapor.  The middle
filter is a separating filter to keep the alkali from the pre-
filter and is analyzed with the impregnated filter.  A thermo-
couple is located below the collection  filter to allow regulation
of the volatilization temperature at the point  of collection.


FINAL LABORATORY EVALUATION OF PROTOTYPE SAMPLING SYSTEM

     Experiments were conducted in which the concentration of
the laboratory-generated sulfuric acid  aerosol  sampled by the
prototype sampling system was gradually reduced toward antici-
pated ambient levels.  Four concentration levels of  sulfuric
acid aerosol were generated; the average amounts sampled  for
the different levels were 121, 55, 19,  and 7.4  yg/filter.  The
collection times for the different levels were  5, 10, 15  and
20 min, respectively.  The results of these experiments  are given
in Table 30.

     These data show that the percentage of sulfuric acid aerosol
collected on the impregnated filter and the percentage of sul-
furic acid aerosol collected on both filters increased as the
amount of aerosol sampled was decreased.  This  apparent  increase
in sampling efficiency is probably due  to the increase in sam-
pling time rather than the amount of sulfuric acid aerosol sam-
pled.  Each sampling interval requires  a short  equilibration
period to stabilize the volatilization  temperature at 130°C.
As the sampling time is increased from  5 to 20  min,  the  equili-
bration period becomes less significant.  At a  1-hr  sampling
time for ambient samples, this initial  warmup effect should be-
come insignificant.

     At the lowest level of sulfuric acid aerosol  (7.4 Mg/filter),
an average of 83% of the total sulfuric acid aerosol sampled
was collected on the prefilter and impregnated  filter.   Of  the
total amount collected, an average of 86% was found  on the  im-
pregnated filter.  However, these results are based  on a labora-
tory generated sample concentration of  26 ug/m3. Better  results
                                 76

-------
                                     HEATING ZONE
                                     TEFLON FILTER
                                     HOLDER
                                     47-mm MITEX PREFILTER
                                     47-mm MITEX SPACER
                                     47-mm ALKALI-
                                     IMPREGNATED TEFLON
                                     FILTER
                                      SEPTUM
                                        THERMOCOUPLE
                                     TO PUMP
Figure 19.  Schematic of Teflon filter holder containing the Teflon
          pre filter and the alkali-impregnated Teflon collection filter.
                       77

-------
       TABLE 30.  EFFICIENCY OF PROTOTYPE SAMPLING APPARATUS  FOR
            VARIOUS CONCENTRATIONS OF SULFURIC ACID AEROSOL
         Amount of H2SO4
        aerosol sampled,*
             Amount of H2SO^ found,t
              pg (% of total HaSO..)
                              On
                 On       impregnated
             prefilter,      filter,
               \ig  (%)        \ig  (%)
Average
Average
33.5
76.9
54.6
55.1
55.0
17.4
23.8
19.6
15.0
19.0
3.0
            26.6
 9.9
19.8
 7.1
 7.1
11.0
(54)
(41)
(50)
1551
(50)

(57)
(83)
(36)
(47)
(56)
                           Portion of
                           total H2SOH
                           collected,




Average
120.3
109.1
101.7
152.4
120.9
21.7
22.4
27.2
46.3
29.4
(17)
(21)
(27)
(31)
(24)
59.4
47.0
34.3
65.1
51.5
(49)
(43)
(34)
(43)
(42)
67
64
61
74
66
90
60
57
61
67

68
86
66
71
73



Average
7
7
8
7
.0
.2
.0
.4
0.
0.
0.
0.
84
87
85
85
(12)
(12)
(11)
(12)
5.
6.
4.
5.
40
05
10
18
(77)
(84)
(54)
(72)
89
96
65
84
    The amount of aerosol sampled was determined  with  a  parallel,
    unheated sampling probe.  All analyses were performed  by
    ion chromatography.

 t  Sulfuric acid aerosol was sampled at a flow rate of  14 1/min
    and volatilized at 130°C.

may be  obtained  when  the lower  ambient levels that are antici-
pated are  sampled.  The  minimum detectable level of the prototype
sampling system  appears  to  be approximately 0.5 pg/m3.
                                78

-------
                            SECTION  9

            FIELD EVALUATION OF  PROTOTYPE  SAMPLER  FOR
                       SULFURIC ACID  AEROSOL
     After the laboratory experiments were completed, ambient
sampling was done with the prototype sampler at two local road-
side sites in the Birmingham, Alabama, area.  A series of 1-hr
samples were first collected for 6 days beside a city street
during the period from August 6 to August 18, 1978.  The sampler
is shown at this location in Figure 18 on page 100.  The average
traffic flow during sampling was approximately 600 cars per hr.
The sampler was located approximately 3 m from the street.  Two
control samples were collected with each sample to determine
the amount of total sulfate and sulfate aerosol less than 2 ym
in diameter present at the sampling site.  The control data are
reasonably consistent with the data obtained from the prototype
sampling system.  The data from these samplings are listed in
Table 31.  A graphic comparison of the amount of total sulfate
and sulfuric acid aerosol present beside the city street is shown
in Figure 20.  Approximately 20% of the nonsized sulfate col-
lected beside the city street was found to be sulfuric acid
aerosol.  The average amounts of total ambient sulfate and am-
bient sulfuric acid aerosol collected were 1.5 and 0.3 pg/m3,
respectively.

     After the sampling beside the city street was completed,
a field test was conducted next to a  major highway.  The proto-
type sampling system was  set up in a  parking deck adjacent to
an elevated  (8 m) six-lane interstate highway on the north side
of Birmingham, Alabama.   The prototype sampler is shown collect-
ing samples at the  interstate highway site  in Figure 21.  The
samples were collected 8  m from the near side and 1.5 m above
the interstate highway.   The angle of the  incline from the center
of the  interstate highway to the sampling  system was approxi-
mately  5°.  Samples of 1-hr  duration  were  collected  for 12 hr
on August 21, 22, and  23, 1978.  On  these  dates  the  local at-
mospheric particulate  counts were 94, 108,  and 97 yg/m3,  respec-
tively, with a temperature low of 21°C and  a high of  34.5°C on
each of the test dates.   On  August  21 and  23 there  was  a  slight
breeze  of 3 to 5 km/hr crossing the  interstate highway  toward
the sampling system.   On  August 22,  however,  the wind  direction
                                 79

-------
                    TABLE  31.
AMBIENT SULFURIC ACID AEROSOL DETERMINATIONS
            BESIDE A CITY STREET
                                 Prototype samples*
                                         Control  samples*



Date Time
collected started
8-9-78 9:30
10:30
11:30
8-14-78 8:30
9: 30
10: 30
11:30
12:30
1: 30
8-15-78 8:30
9:30
10:30
11:30
12:30
1:30
2:30
8-16-78 8:30
9:30
10:30
11:30
12:30
1:30
8-17-78 9:30
10:30
11:30
12:30
1:30
2:30
8-18-78 9:30
10:30
11:30
12:30
1:30


SO,," found
on prefilter,
ng/m3
0. 32
0.11
0.41
0.89
1.20
0.90
1.41
0.55
0.60
0. 36
0.40
0.49
0.45
0.54
0.54
0.81
0.64
0.32
0.79
0.64
0.48
0.00
1.27
1.11
0.64
2.06
0.79
0.48
0.64
0.32
0.79
0.48
0.79


H2SOV found
on impregnated
filter,1" yg/m3
0.23
0.19
0.07
4.12
0.40
0.90
1.70
1.80
2.31
0.14
0.28
0.19
0.15
0.10
0.22
0.10
0.00
0.00
0.00
0.24
0.00
0.00
0.00
0.00
0.48
0.00
0.00
0.00
0.00
0.00
1.29
0.16
0.00
Total SO,,-
found on
prefilter and
impregnated
filter, wg/m'
0.55
0.30
0.48
5.01
1.60
1.80
3.11
2.36
2.91
0.50
0.68
0.68
0.50
0.64
0.76
0.91
0.64
0.32
0.79
0.88
0.48
0.00
1.27
1.11
1.12
2.06
0.79
0.48
0.64
0.32
2.08
0.64
0.79


Total SO,,= less
than 2-iim
diameter, S pg/m3
0.29
0.26
0.68
1.49
1.19
1.69
3.07
0.39
0.64
0.57
0.63
0.63
0.52
0.63
0.72
0.74
0.64
0.48
0.95
0.64
0.48
0.00
1.11
1.11
0.95
1.27
0.79
0.64
0.64
0.32
2.06
0.64
0.79


Total
SO..V
ijg/m3
0.14
0.11
0.16
3.71
3.82
7.64
8.10
0.63
1.39
0.71
0.89
0.99
1.07
0.83
0.94
0.81
0.48
0.79
1.43
0.64
0.48
0.32
1.59
1.27
2.54
1.27
0.64
1.11
0.95
1.11
2.86
0.64
1.90
* All analyses by ion chromatography.
t Volatilization temperature was 130°C.  All sampling was done at a flow rate of 14 1/min.
§ Determined with separate control filter with a 2-pm diameter cutoff cyclone.
I Determined with separate open-face filter.
                                         80

-------
                                       AVERAGE TOTAL SULFATE
                               	AVERAGE SULFURIC ACID
               STARTING TIME OF ONE HOUR SAMPLING
Figure 20.  Hourly average concentrations of total sulfate and sulfuric
           acid on six days beside a city street.

                             81

-------
      Figure 21.  Remote sampling at interstate highway.
Figure 22.  Ion chromatographic analyses of collected samples.




                           82

-------
and velocity were variable during most of the sampling period.
The average traffic flow passing the sampling site as determined
from government traffic counts was approximately 8,000 to 10,000
cars per hour.

     As before, two control samples were collected with each
sulfuric acid aerosol sample.  All samples collected were brought
back to the laboratory in individual glass vials for analysis
by 1C  (Figure 22.)  The data from these samplings are given  in
Table 32.  Figure 23 shows a graphic comparison of the concen-
trations of total sulfate and sulfuric acid aerosols collected
hourly on August  23, 1978, at the interstate highway.  Figure
24 is a graphic comparison of the average concentrations of  total
sulfate and sulfuric acid aerosols by time of day collected  from
the 3 days of field testing at the interstate highway location.
Approximately 30% of the nonsized sulfate collected beside the
interstate highway was found to be sulfuric acid aerosol.  The
average amounts of total ambient sulfate and ambient sulfuric
acid aerosol collected were 4.4 and 1.4 pg/m3, respectively.
                                83

-------
                      TABLE 32.  AMBIENT SULFURIC ACID AEROSOL DETERMINATIONS
                                   FROM REMOTE INTERSTATE HIGHWAY SAMPLING


Date Time
collected started
8-21-78 6:00
7:00
8:00
9:00
10:00
11:00
12:00
1:00
2:00
3:00
4:00
5:00
8-22-78 6:00
7:00
8:00
9:00
10:00
11:00
12:00
1:00
2:00
3:00
4:00
5:00
8-23-78 6:00
7:00
8:00
9:00
10:00
11:00
12:00
1:00
2:00
3:00
4:00
5:00


SO,," found
on prefilter,
ug/m
1.46
1.72
1.32
1.85
0.93
0.66
0.79
1.06
1.06
0.93
1.19
0.79
0.79
0.48
1.11
0.64
0.48
0.56
0.64
0.64
0.48
0.48
0.48
0.64
0.64
0.95
1.27
1.43
0.64
0.95
0.64
0.64
0.48
2.06
1.27
0.64


HjSO,, found
on impregnated
filter,1' (jg/m
3.10
4.29
0.71
3.33
0.00
1.91
2.38
3.10
1.67
1.19
1.90
1.19
1.43
1.43
1.43
1.43
3.57
3.10
0.00
2.70
0.32
0.00
2.70
1.58
0.00
0.00
0.00
3.57
1.67
0.00
2.38
0.00
0.00
1.90
0.00
0.00
Total S0»=
found on
prefilter and
impregnated
filter, wg/m
4.56
6.01
2.03
5.18
0.93
2.57
3.17
4.16
2.73
2.11
3.09
1.98
2.22
1.91
2.53
2.07
4.05
3.66
0.64
3.34
0.80
0.48
3.18
2.22
0.64
0.95
1.27
5.00
2.31
0.95
3.02
0.64
2.38
2.06
1.27
0.64
                                                                          Control samples*
                                                                      Total SO,,  less
                                                                         than 2-ym
                                                                      diameter,s pg/ro

                                                                             4.37
                                                                             6.22
                                                                             2.12
                                                                             5.03
                                                                             0.93
                                                                             2.65
                                                                             2.91
                                                                             4.10
                                                                             0.53
                                                                             2.12
                                                                             3.31
                                                                             1.98
                                                                             2.06
                                                                             1.90
                                                                             2.43
                                                                             2.22
                                                                             0.48
                                                                             0.48
                                                                             0.79
                                                                             0.32
                                                                             0.79
                                                                             0.48
                                                                             0.48
                                                                             2.06

                                                                             0.48
                                                                             0.64
                                                                             1.
,27
,24
 38
                                                                             0.95
                                                                             3.02
                                                                             0.64
                                                                            2.38
                                                                            1.94
                                                                            1.27
                                                                            0.64
            0.93

           14.29
 3.02
 3.90
 6.98
 0.64
 2.54
 1.27
 0.48
 0.79
 0.48
 0.48
 3.02

 0.64
 0.32
 2.06
 6.35
 5.08
 2.22
 4.44
2.54
 3.65
8.41
2.38
3.02
* All analyses by ion chromatography.
t Volatilization temperature was 130'C.  All sampling was done at a flow rate of 14 1/min
§ Determined with separate control filter with a 2-um diameter cutoff cyclone.
* Determined with separate open-face filter.
                                           84

-------
0

E
3
u.
_1

co

K
o

UJ
UJ
oc
UJ
                      	SULFURIC ACID
      AM   AM
    Figure 23.
 AM   AM   AM   AM  NOON  PM   PM    PM


    STARTING TIME OF ONE HOUR SAMPLING
                                                              PM   PM
Hourly concentrations of total sulfate and sulfuric acid

on August 23, 1978, at an interstate highway.
                                   85

-------
  CO
  ^
   2
  Q
  o
  o
  E
  D
  CO
  DC
  O
  CO
  _i
  <


  g
  UJ
  O

  DC
  LU
  >
                                          AVERAGE TOTAL SULFATE
                                  	AVERAGE SULFURIC ACID
                 8     9    10    11    12    1     2     3

                AM   AM   AM   AM   NOON  PM    PM   PM

                    STARTING TIME OF ONE HOUR SAMPLING
 4

PM
 5

PM
Figure 24.   Hourly average concentrations of total sulfate and sulfuric
           acid on three days at an interstate highway.
                              86

-------
4. TITLE AND SUBTITLE
9. PERFORMING ORGANIZATION NAME AND ADDRESS

  Southern Research  Institute
  2000 Ninth Avenue  South
  Birmingham,  Alabama  35205
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
  REPORT NO.
 EPA-600/2-80-037
                                                           3. RECIPIENT'S ACCESSION NO.
  DEVELOPMENT OF A  PORTABLE  DEVICE
  TO COLLECT SULFURIC  ACID AEROSOL
  Final Report
                             5. REPORT DATE
                               May  1980
                             6. PERFORMING ORGANIZATION CODE
                                                           8. PERFORMING ORGANIZATION REPORT NO.
  Herbert C. Miller,  David W.  Mason, and
  William J. Barrett
12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental  Sciences Research Laboratory
  Office of  Research and Development
  U. S. Environmental Protection Agency
  Research Triangle Park, NC  27711
                             10. PROGRAM ELEMENT NO.

                               1AA601  CA-28 (FY^78)
                             11. CONTRACT/GRANT NO.
                               68-02-2468
                                                           13. TYPE OF REPORT AND PERIOD COVERED
                               Final 9/76-9/78
                             14. SPONSORING AGENCY CODE
                               EPA/60Q/09
15. SUPPLEMENTARY NOTES
   Previous  Related Reports:
EPA-600/2-77-027,  February 1977
EPA-600/2-78-060,  March 1978
16. ABSTRACT
  A quantitative,  interference-free method  for collecting sulfuric acid  aerosol  on a
  filter was  developed and field tested.  Since previous research found  that  severe
  losses of sulfuric acid were caused  by  ammonia, ambient particulate material,  and
  other interferents, a method was needed that converts sulfuric acid to a  stable
  derivative  for subsequent analysis.  Methods evaluated for direct fixation  of
  sulfuric acid aerosol were not found to be selective.  Therefore, a sampling
  method was  investigated that uses a  combination of selective volatilization of
  the  sulfuric acid, prefiltration of  particulate interferents, and derivatization
  of the vaporized acid on an alkali-impregnated filter.  Research on the volati-
  lization,  prefiltration, and derivatization technique, and the development  and
  field evaluation of a prototype sampler based on this technique are described.
17

a.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
   *Air Pollution
   *Sulfuric Acid
   *Aerosol
   Collecting Methods
   *Filters
    Tests
                b.lDENTIFIERS/OPEN ENDED TERMS
  COS AT I Held/Group
                                             13B
                                             07B
                                             07D
                                             14B
     ITR'BIJTION STATEMCNT
   RELEASE TO PUBLIC
                19 SECURITY CLASS /This Report)
                  UNCLASSIFIED	
                f~20 SECURITY CLASS I'fhix page)
                I  UNCLASSIFIED
                                                                          21. NO OF PAGES
                                                                            99
22. PRICE
 SPA Torn! 2220-! ',FUv. 4-77)   P--.C/IOUS EDITION i s OBSOLETE
                                              87

-------