&EPA
            United States
            Environmental Protection
            Agency
            Environmental Sciences Research  EPA-600 2-80-068
            Laboratory         April 1980
            Reseaich Triangle Park NC 27711
            Research and Development
Analytical
Procedures for
Characterizing
Unregulated
Emissions from
Vehicles Usi xg
Middle-Distillate
Fuels

Interim Report

-------
                  RESEARCH REPORTING SERIES


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

    1.  Environmental Health Effects Research

    2.  Environmental Protection Technology

    3.  Ecological Research

    4.  Environmental Monitoring

    5.  Socioeconomic Environmental Studies

    6.  Scientific and Technical Assessment Reports (STAR)

    7.  Interagency Energy-Environment Research and Development

    8.  "Special" Reports

    9.  Miscellaneous Reports

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

-------
                                                   EPA-600/2-80-068
                                                   April  1980
ANALYTICAL PROCEDURES FOR CHARACTERIZING UNREGULATED
EMISSIONS FROM VEHICLES USING MIDDLE-DISTILLATE FUELS

                   Interim Report
                         by

                    Lawrence Smith
                  Mary Ann Parness
                  E. Robert Fanick
                 Harry E. Dietzmann
            Southwest Research Institute
              San Antonio, Texas 78284
               Contract No. 68-02-248?-
                   Project Officer

                  Ronald L. Bradow
 Emissions Measurement and Characterization Division
     Environmental Sciences Research Laboratory
        U.S. Environmental Protection Agency
         Research Triangle Park, N.C. 27711
     ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
         OFFICE OF RESEARCH AND DEVELOPMENT
        U. S. ENVIRONMENTAL PROTECTION AGENCY
    RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711

-------
                                DISCLAIMER

     This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication.  Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.

-------
                                FOREWARD

     The Clean  Air  Act  as  amended  in  1977  requires manufacturers  of
automobiles  to  certify  that  the emissions  of  automobiles  represent no
unreasonable risk to  the public health and safety,   EPA's  role  in enforcing
this section of the Act has  been to formulate methods  and  procedures by
which  toxic  pollutants  which might be emitted from various  kinds  of
automobile engines  could be  measured  and assessed.   The Environmental
Sciences Research Laboratory contributes to this overall  agency effort
through programs engaged in
i-
  i        studies to  identify  and  measure  toxic pollutants  in source
          emissions and in the ambient air.

          development of methods and  procedures to measure  air pollutants

          development of modeling  procedures  which permit  prediction
          of ambient  air quality impacts from source emissions data.

     This report is the second of  two similar documents relating  the
development  of  analytical  methods  for measuring trace  toxic pollutants in
mobile source exhaust gas.   The first of these reports provided fully tested
procedures for  10 toxic gases  in the  exhaust  of gasoline engines.  The
current report  deals  with  many of  the same compounds and methods  now
fully  qualified for use with distillate-fueled engines such as diesel
or  gas-turbine  powerplants.  It is intended that this  report serve as a
working guide to the  automotive industry and  to a variety  of government
and academic research institutions, providing well-tested  analytical
methods for  studying  hazardous pollutant emissions from automotive
powerplants.
                               A.H.  Ellison
                               Director
                               Environmental  Sciences  Research  Laboratory

-------
                                  ABSTRACT

This  research program was initiated with the objective of developing,
codifying and testing a group of chemical analytical methods for measuring
toxic compounds  in the exhaust of distillate-fueled engines (i.e. diesel,
gas turbine, Stirling, or Rankin cycle powerplants).  It is a part of a
larger  effort to characterize these components from a number of prototype
powerplants and, thus, represents a logical first step in the process.

Methods of collection and analysis for aldehydes and ketones, for* hydro-
gen cyanide and  cyanogen, for hydrogen sulfide, carbonyl sulfide and
organic sulfides, for ammonia and amines, for nitrous oxide, sulfur
dioxide, individual  hydrocarbons, for soluble sulfate and N-nitrosodi-
methylamine, benzo-a-pyrene, and phenols were studied in detail.  Ten
analytical procedures were developed and codified.  Interference studies
and proof-tests  in diesel engine exhaust were conducted with every
procedure and the results of these experiments are reported in detail.

All of  the procedures were found to be suitable for use in exhaust
emissions characterization studies.  The sampling parameters were found
to be adequate for the collection of trace levels of exhaust components
using standard CVS sampling techniques.  Interferences were, in general,
minimal  although there were two significant problem areas. Phthalate
ester interferes with crotorialdehyde determinations and this contaminant
must be avoided  in the procedure.  In the hydrogen sulfide method, S02
decreases the apparent sulfide, and its presence must be corrected for.
While other interferences were noted, all could be avoided with the
appropriate precautions noted in the final procedure.

Qualification tests were conducted by introducing known quantities of
these pollutants into the exhaust of a diesel engine operating on a
standard emissions test CVS tunnel.  The results of these experiments
indicated completely quantitative recovery for aldehydes and ketones,
S02, nitrous oxide, total cyanide and phenols.  Hydrogen sulfide is lost
to the extent of ten percent at normal exhaust levels.  Amines, ammonia
and organic sulfides can be lost in sampling in significant amounts in
the CVS apparatus.   These losses must be taken into account when calcu-
lating exhaust contributions.
                                    IV

-------
                                 CONTENTS
Foreword	   xii
Abstract	    iv
Figures	    vi
Tables	    X

     1.   Introduction 	    1
     2.   Aldehyde and Ketone Procedure	    3
     3.   Total Cyanide Procedure	   20
     4.   Individual Hydrocarbon Procedure 	   43
     5.   Organic Amine Procedure	   53
     6.   Sulfur Dioxide Procedure 	   74
     7.   Nitrous Oxide Procedure	   89
     8.   Hydrogen Sulfide Procedure 	   98
     9.   Ammonia Procedure	112
    10.   Organic Sulfide Procedure	127
    11.   Phenol Procedure 	  158
    12.   The Qualification Experiment 	  178
    13.   Results and Conclusions	184

References	187

Appendices

     A.   Aldehyde and Ketone Procedure	197
     B.   Total Cyanide Procedure	227
     C.   Individual Hydrocarbon Procedure 	 .251
     D.   Organic Amine Procefure	273
     E.   Sulfur Dioxide Procedure 	  294
     F.   Nitrous Oxide Procedure	317
     G.   Hydrogen Sulfide Procedure 	  334
     H.   Ammonia Procedure	359
     I.   Organic Sulfide Procedure	382
     J.   Phenol Procedure 	  415
     K.   Sulfate Procedure	440
     L.   DMNA Samplnig Procedure	460
     M.   DMNA Analysis Procedure	470
     N.   BaP Sampling and Analysis	484

-------
                                   FIGURES

Number                                                                   Page

  1     Plot of the formaldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	    8

  2     Plot of the acetaldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	    9

  3     Plot of the acetone-DNPH derivative concentration determined
          by prpcedure vs actual concentration	   10

  4     Plot of the isobutyraldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	   11

  5     Plot of the methylethylketone-DNPH derivative concentration
          determined by procedure vs actual concentration 	   12
                                                   r

  6     Plot of the crotonaldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	   13

  7     Plot of the hexanaldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	   14

  8     Plot of benzaldehyde-DNPH derivative concentration
          determined by procedure vs actual concentration 	   15

  9     Effect of elapsed time on hydrogen cyanide in clear and dark
          bags	   23

 10     The effect of elapsed time on cyanogen in clear and dark
          bags	*.....   24

 11     The effect of elapsed time on hydrogen cyanide in a blend of
          hydrogen cyanide in clear and dark bags	   25

 12      The effect of elapsed time on cyanogen in a blend of hydrogen
          cyanide and cyanogen in clear and dark bags ...'.'	   26

 13      The effect of elapsed time on hydrogen cyanide and cyanogen
          in a dark bag with humid nitrogen	   27

 14      Total cyanide calibration curve at low concentrations
          (0-2 ppm)	   33

 15      Total cyanide calibration curve at low concentrations
          (0-10 ppm)	   34

                                     VI

-------
                              FIGURES  (Cont'd)

Number                                                                    Page

 16     The effect of elapsed time  on  sample  development	    36

 17     Time-sample decay  curve  	    38

 18     Time-sample decay  curve  (exhaust  only)	    48

 19     Time-sample decay  curve  (standard only)  	    49

 20     Time-sample decay  curve  (exhaust  + standard)	    50

 21     GC peak  areas of pentafluorobenzoyl amine  derivatives  vs
          time	    59

 22     Linearity of monomethylamine GC response  (plot on  log-log
          scale)	    63

 23     Linearity of dimethylamine  GC  response  (plot  on log-log
          scale)	    64

 24     Linearity of trimethylamine GC response  (plot on log-log
          scale)	    65

 25     Linearity of monoethylamine GC response  (plot on log-log
          scale)	    66

 26     Linearity of diethylamine GC response (plot on log-log
          scale)	    67

 27     Linearity of triethylamine  GC  response  (plot  on log-log
          scale)	    68

 28     SC>2 calibration curve	    84

 29     Detector linearity curve	    91

 30     Sample decay curve (short term)  	    94

 31     Sample decay curve (long term)	    95

 32     Time-Light exposure study  (low concentration)  	   101

 33     Time-Light exposure study  (high concentration)	102

 34     Beer's Law plot for methylene  blue	103

 35     Ammonia  calibration curve  	   120
                                     vn

-------
                              FIGURES (Cont'd)

Number                                                                   Page

 36     Proposed GC flow schematic for analysis of organic sulfides
          (Step 1)	129

 37     Proposed GC flow schematic for analysis of organic sulfides
          (Step 2)	130

 38     Proposed GC flow schematic for analysis of organic sulfides
          (Step 3). . . /	131

 39     Gas chromatograph separation of several organic sulfides in
          prepared blend	133

 40     Cold trap experiment flow schematic	135

 41     Typical gas chromatograph trace of organic sulfides 	  141

 42     Typical GC separation of organic sulfides on acetone-washed
          Porapak QS column	143

 43     Typical organic sulfide separation with Tenax-GC column ....  144

 44     Organic sulfide permeation blend with secondary dilution,
          near detection limit of GC FPD system	145

 45     Organic sulfide permeation blend with secondary dilution,
          concentrated on Tenax-GC trap and thermally desorbed into
          GC FPD system	146

 46     Carbonyl sulfide linearity plot 	  151

 47     Methyl sulfide linearity plot 	  152

 48     Ethyl sulfide linearity plot. 	  153

 49     Methyl disulfide linearity plot 	  154

 50     Linearity of o-chlorophenol GC response 	  168

 51     Linearity of phenol GC response	169

 52     Linearity of, salicylaldehyde GC response	170

 53     Linearity of m-cresol and p-cresol GC response	171

 54     Linearity of p-ethylphenol, 2-isopropylphenol, 2,3-xylenol,
          3,5-xylenol and 2,4,6-trimethylphenol GC response 	  172
                                   viii

-------
                               FIGURES (Cont'd)

Number                                                                  Page

 55     Linearity of 2,3,5-trimethylphenol GC response	173

 56     Linearity of 2,3,5 ,6-tetramethyIphenol GC response	174

 57     Dilution tunnel-CVS system used in qualification experiments.  .  179

 58     Apparatus for injection of pollutant into dilution tunnel
          without exhaust  	  180

 59     Apparatus for injection of pollutant into dilution tunnel
          with  exhaust	181

 60     Modified mist generator	182
                                      IX

-------
                                   TABLES

Number                                                                   Page

  1     Physical Properties of the Aldehydes and Ketones (4,5)	    4

  2     Injection Repeatability 	    7

  3     Multiple Extractions of DNPH Solutions (All units are mg DNPH
          derivative/mS, of toluene)	   16

  4     Percent Recovery of Propionaldehyde 	   18

  5     Experiments Conducted for HCN and C2N2 Bag Stability	   22

  6     The Effect of Stopper Tip and Absorbing Reagent Concentration
          on Collection Efficiency at Room Temperature	   29

  7     The Effect of Absorbing Reagent Temperature on HCN Collection
          Efficiency	   31

  8     Calibration Curve Linearity at Several Cyanide
          Concentrations	   32

  9     Sample Injection Repeatability for Two Cyanide
          Concentrations	   35

 10     Total Cyanide Gaseous Recovery by Direct CVS Injection	   39

 11     Total Cyanide Recovery from Dilute Exhaust Without Filter or
          With Non-Heated Filter	   39

 12     Total Cyanide Recovery from Dilute Exhaust With Heater
          Filter	   40

 13     Important Facts on Individual Hydrocarbons	   45

 14     Injection Repeatability on Two Separate Occasions 	   46

 15     List of Individual Organic Amines Included in the Emissions
          Characterization Inventory	   54

 16     Mixing Procedure for Preparation of Pentafluorabenzoylamine
          Derivatives	   58

 17     Injection Repeatability Experiments 	   62

 18     Organic Amine Recovery from the CVS Dilution Tunnel Only. ...   70

                                      X

-------
                               TABLES  (Cont'd)

Number                                                                   Page

 19     Organic Amine Recovery from the CVS Dilution Tunnel With
          Exhaust	   72

 20     Interferences to SC>2 Analysis	   79

 21     Sulfate Standard Stability	   79

 22     S02 Collection Efficiency as a Function of Flowrate and
          Temperature	   80

 23     Injection  Repeatability for Ion Chromatograph  	   82

 24     Calibration Curve for Sulfur Dioxide	   83

 25     Sulfur Dioxide Recovery from CVS-Tunnel Injection  	   86

 26     Sulfur Dioxide Recovery from Dilute Exhaust by CVS-Tunnel
          Injection During Hot FTP Driving Cycle	   87

 27     Injection  Repeatability Over the Range of Detector Linearity.  .   92

 28     Nitrous Oxide Qualification Experiments - No Vehicle	   96

 29     Nitrous Oxide Qualification Experiment With Vehicle Exhaust  .  .   96

 30     The Effect of Sample Flow Rate and Absorbing Reagent
          Temperature on the Collection Efficiency	104

 31     The Effect of Individual Exhaust Components on the
          Development of Methylene Blue	105

 32     The Effect of Anions on the Development of Methylene Blue  .  .  .  106

 33     The Effect of Sulfur Dioxide Interference on the Development
          of Methylene Blue	107

 34     Hydrogen Sulfide Recovery - No Exhaust Present	108

 35     Effect of  Ferric Ion Solution  on Hydrogen Sulfide Recovery
          from Dilute Exhaust	109

 36     NH, Collection Efficiency as a Function of Flowrate and
          Temperature	115

 37     Injection  Repeatability for Ion Chromatograph  	  117

 38     Repeatability of Ammonia Standard  	  116
                                      XI

-------
                               TABLES (Cont'd)

Number                                                                   Page

 39     Calibration Curve for Ammonia 	  119

 40     Sample and Standard Stability as a Function of Time	121

 41     Ammonia Recovery from CVS-Tunnel, No Exhaust	122

 42     Ammonia Recovery from Dilute Exhaust (No Heated Filter)  ....  124

 43     Ammonia Recovery from Dilute Exhaust, Heated Filter 	  125

 44     List of Sulfur Compounds Included in the Analysis of Organic
          Sulfides	127

 45     List of Chemical and Physical Characteristics of Various
          Sulfur Compounds Potentially Present in Automotive Exhaust. .  134

 46     The Effect of Cold Trapping at -78°C on Carbonyl Sulfide and
          Methyl Sulfide at Various Concentrations, Flow Rates and
          Trap Sizes	136

 47     The Effect of Cold Trapping at -196°C on Carbonyl Sulfide
          and Methyl Sulfide with Various Trap Sizes	138

 48     The Efficiency of Various Materials Trapping Sulfides at
          Several Temperatures	139

 49     Injection Repeatability for the Organic Sulfides	148

 50     Trap Repeatability for Organic Sulfide Collection 	  149

 51     Trap-to-Trap Repeatability for Organic Sulfide Collection .  . .  149

 52     Percent Recoveries of the Organic Sulfides from the CVS
          Tunnel Only	156

 53     Percent Recoveries of the Organic Sulfides from the CVS
          Tunnel and Exhaust	157

 54     Physical Properties of Phenols Possible in Exhaust	158

 55     Extractions  with Methylene Chloride 	  161

 56     Extractions  with Ether	162

 57     Extraction Efficiency as a Function of pH of Aqueous Solution  .  160

 58     Effect of Reducing Sample Volume by Kuderna Danish Concentrator
          on Phenol  Recovery	163

                                     xii

-------
                                TABLES (Cont'd)



Number                                                                   Page

 59     Effect of Diesel Fuel on Recovery of Phenol	164

 60     Interferences to Phenol Recovery or Analysis	165

 61     Linearity Ranges of Internal Standard and of Phenols in
          Exhaust	167

 62     Injection Variability of Phenol	175

 63     Baseline Phenol Emission Levels from Mercedes 240D Diesel .  .  .   175

 64     Percent Recoveries from Injection of Phenol into Exhaust of
          Mercedes 240D Diesel	176

 65     Analytical Procedures of Emissions Characterization 	   185
                                     xm

-------
                                SECTION 1

                               INTRODUCTION


     The objective of this project is to evaluate the emissions  of regulated
and nonregulated pollutants in the exhaust of vehicles having advanced-con-
cept powerplants.  Examples of engine types which are being considered for
testing in this project include gas turbines, Stirling cycle, turbocharged
Diesel, Rankine cycle, stratified charge, and advanced Otto-cycle.   The
first phase of this project includes the development of analytical techno-
logy to provide qualitative and quantitative measurements of unregulated
exhaust products of the engines to be tested.  This report is a  summary of
the results of this phase of the project.

     Candidate analytical procedures were selected for each of the following
compounds or groups of compounds.

     aldehydes and ketones                     nitrous oxide

     hydrogen cyanide + cyanogen               sulfur dioxide

     hydrogen sulfide                          individual hydrocarbons

     organic sulfides + carbonyl sulfide       phenols

     ammonia                                   N-nitrosodimethylamine

     organic amines                            benzo-a-pyrene

                                               soluble sulfate

     The procedures selected represent an assessment of the optimum proce-
dures available at the time of this report and with the approval of the
project officer will be used to measure the appropriate unregualted pollu-
tants.  Reviews of the literature, procedural development work,  validation
experiments, and qualification experiments are discussed for ten of these
analytical procedures.

     These ten analytical procedures are listed in the following paragraphs
along with the appropriate section of the report in which they are discussed
as well as a brief description of the procedure.

-------
     Aldehydes and Ketones  (Section 2) - The collection of aldehdyes
 (formaldehyde, acetaldehyde, isobutyraldehyde and hexanaldelyde) and ketones
 (acetone and methylethyIketone) is accomplished by bubbling CVS diluted
 exhaust through glass impingers containing 2,4-dinitrophenylhydrazine  (DNPH)
 in  dilute hydrochloric acid.  The aldehydes and ketones  (also knwon as
 carbonyl compounds) react with the DNPH to form their respective phenyl-
 hydrazone derivatives.  These derivatives are insoluble or only slightly
 soluble in  the DNPH/HC1 solution and are removed by filtration followed by
 pentane extractions.  The filtered precipitate and the pentane extracts are
 combined and then the pentane is evaporated in the vacuum oven.  The remain-
 ing dried extract contains the phenylhydrazone derivatives.  The extract
 is  dissolved in a quantitative volume of toluene containing a known amount
 of  anthracene as an internal standard.  A portion of this dissolved extract
 is  injected into a gas chromatograph and analyzed using a flame ionization
 detector.

      Total  Cyanide  (Hydrogen Cyanide plus Cyanogen) (Section 3) - The
 collection  of total cyanide is accomplished by bubbling CVS diluted exhaust
 through glass impingers containing a 1.0 N potassium hydroxide absorbing
 solution.   This solution is maintained at ice bath temperature.  An aliquot
 of  the absorbing reagent is then treated with KH2P(>4 and Chloramine-T.  A
 portion of  the resulting cyanogen chloride is injected into a gas chromato-
 graph equipped with an electron capture detector (BCD).  External CN~ stan-
 datds are used to quantify the results.

      Individual Hydrocarbons  (Section 4) - For measurement of selected
 individual  hydrocarbons, methane  (CH4) , ethane ^2^), ethylene (C2H4) ,
 acetylene  (C2H2), propane  (C^EQ), propylene (C-fl^), benzene (CgHg) , and
 toluene  (CjUg), a sample of CVS diluted exhaust is collected in a Tedlar
 bag.   This  bagged sample is then analyzed for individual hydrocarbons using
 a gas  chromatographic system containing four separate columns and a flame
 ionization  detector.  The peak areas are compared to an external calibration
 blend and the individual hydrocarbon concentrations are obtained using a
 Hewlett-Packard 3354 computer system.

     Organic Amines (Section 5) - The collection of organic amines (mono-
 methylamine, monoethylamine and dimethylamine, trimethylamine, diethylamine,
 and triethylamine) is accomplished by bubbling CVS diluted exhaust through
 glass impingers containing dilute sulfuric acid.  The amines are complexed
by  the acid to form stable sulfate salts which remain in solution.  A
portion of this solution is then injected into a gas chromatograph equipped
with an ascarite loaded pre-column and a nitrogen phosphorus detector  (NPD).
External amine standards in dilute sulfuric acid are used to quantify the
results.

     Sulfur Dioxide (Section 6) - The concentration of sulfur dioxide in
dilute exhaust is determined as sulfate using a ion chromatograph.  Sulfur
dioxide is collected and converted to sulfate by bubbling dilute exhaust
through two glass impingers containing a  3 percent hydrogen peroxide ab-
sorbing solution.   The samples are analyzed on the ion chromatograph and
compared to standards of known sulfate concentrations.

-------
     Nitrous Oxide (Section 7) - For measurement of nitrous oxide, a sample
of the CVS diluted exhaust is collected in a Tedlar bag.  This bagged sample
is then analyzed for nitrous oxide using a gas chromatograph equipped with
an electron capture detector.  Calibration blends are used to quantify the
results.  Gas chromatograph peak areas are obtained using a Hewlett-Packard
3354 computer system.

     Hydrogen Sulfide (Section 8) - The collection of hydrogen sulfide is
accomplished by bubbling CVS diluted exhaust through glass impingers con-
taining a buffered zinc acetate solution which traps the sulfide ion as
zinc sulfide.  The absorbing solution is then treated with N,N-dimethyl-
paraphenylene diamine sulfate and ferric ammonium sulfate.  Cyclization
occurs, forming the highly colored heterocyclic compound methylene blue
(3,9-bisdimethylaminophenazothionium sulfate).  The resulting solution is
analyzed with a spectrophotometer at 667 nm in a 1-cm or 4-cm pathlength
cell depending upon the concentration.

     Ammonia  (Section 9) - Ammonia in CVS diluted automotive exhaust is
measured in the protonated form, NH4+, after collection in dilute 112804.
The acidification is carried out in a glass impinger maintained at ice bath
temperature.  A sample from the impinger is analyzed for ammonia in an
Ion Chromatograph and the concentration in the exhaust is calculated by
comparison to an ammonium sulfate standard solution.

     Organic Sulfides (Section 10) - The collection of carbonyl sulfide (COS)
and the organic sulfides, methyl sulfide (dimethylsulfide, (CH3)2S), ethyl
sulfide (diethylsulfide, (C2H5)2S) and methyl disulfide (dimethyldisulfide,
(^3)282) i is accomplished by passing CVS diluted exhaust through Tenax
GC traps at -76°C.  At this temperature the traps remove the organic sul-
fides from the dilute exhaust.  The organic sulfides are thermally desorbed
from the traps into a gas chromatograph sampling system and injected into
a gas chromatograph equipped with a flame photometric detector for analysis.
External organic sulfide standards generated from permeation tubes are used
to quantify the results.

     Phenols  (Section 11) - The collection of phenols (phenol; salicyl-
aldehyde; m-cresol and p-cresol; p-ethylphenol, 2-isopropylphenol, 2,3-
xylenol, 3,5-xylenol and 2,4,6,-trimethylphenol; 2,3,5,-trimethylphenol;
and 2,3,5,6,-tetramethylphenol) is accomplished by bubbling CVS diluted
exhaust through two Greenburg-Smith impingers containing 200 m& of 1 N KOH.
The phenols react with the KOH and remain in solution.  The contents of each
impinger are acidified and extracted with ethyl ether.  The samples are
partially concentrated, combined and then further concentrated to about 1 m£.
An internal standard is added and the volume is adjusted to 2 mil.  The final
sample is analyzed by the use of a gas chromatograph and concentrations of
individual phenols are determined by comparison to external and internal
standards.

     These ten analytical procedures underwent a series of validation and
qualification experiments.  The validation experiments were carried out to
determine if the sampling and instrument parameters were appropriate for the

-------
quantitative analysis of dilute exhaust.  The qualification experiments were
carried out to determine if the compounds of interest could be quantitatively
recovered from the Constant Volume Sampler (CVS)-dilution tunnel with and
without the presence of exhaust in the tunnel.

     Validation experiments included checks for sample stability, sample
collection efficiency, detector linearity, interferences, extraction effi-
ciency and repeatability, and analysis repeatability.

     Sample stability checks were performed using repeated analyses of the
same sample at intervals over a specified period of time and comparing the
results to the initial analysis.  Aldehydes and ketones  (after extraction),
total  cyanide, individual hydrocarbons, organic amines, sulfur dioxide,
nitrous oxide, ammonia, and phenols (after extraction) were found to be
 stable  for several days.  The organic sulfides and hydrogen sulfide samples
were found to be  stable for approximately one day.

     Sample collection efficiency experiments were performed by passing a
known  concentrations of sample through a series of impingers or traps and
 analyzing each impinger or trap individually for the compound of interest.
All the procedures discussed in this report have a collection efficiency of
98% or better.  Detector linearity experiments were performed by preparing
several samples of various known concentrations and plotting resulting
peak areas  (or heights) versus the concentrations.  All instruments demon-
strated linearity of response for expected concentration ranges (sample
 concentrations above the linear range must be diluted to concentrations that
fall within the linear range of the instrument).  The organic sulfides must
be monitored carefully as traps containing over 200 ng of sample fall beyond
the linear range  of the flame photometric detector.  The sample flow rate
can be lowered to prevent overloading the collecting Tenax trap.

     To determine the interferences for each procedure, known exhaust com-
ponents were introduced into the sample to determine their effect on the
resultant measurements.  Interferences were checked and documented for each
procedure.  Phthalates were found to interfere with the aldehyde and ketone
procedure and may cause erroneous results for crotonaldehyde and benzalde-
hyde.  In the hydrogen sulfide procedure, sulfur dioxide decreases the
apparent hydrogen sulfide concentration, and its presence or absence must
be  recorded.  Thiophene and ethyl sulfide can not be effectively separated
with the normal gas chromatographic operating conditions and therefore,
thiophene must be included as a possible source of error in the analysis
for ethyl sulfide.  The other procedures have interference that can be
avoided if care is taken.

     To determine extraction efficiency and repeatability for the aldehyde
and ketone and the phenol procedures,  several samples of known concentra-
tions were prepared and a number of analyses were performed.  The extrac-
tion efficiency is approximately 100 percent for the aldehyde and ketone
procedure, however the overall repeatability varies up to 15 percent at
concentrations of 0.2-2.0 mg derivative per m£ toluene.  The results of
extraction repeatability experiments for aldehyde and ketone DNPH derivative

-------
concentrations below 0.025 mg DNPH derivative per m£ toluene indicate that
the variability in the extraction process can be very significant (i.e.,
0.94 percent for benzaldehyde at 0.016 mg/m£).  This variability needs to
be taken into account when evaluating data obtained using this procedure.
The extraction efficiency for the phenol procedure is only about 68 percent
due to unavoidable problems in the drying down process.  This value is
repeatable if the extraction procedure is followed closely.  These losses
must be taken into account when analyzing data obtained from the phenol
procedure.

     To determine analysis repeatability, several samples of known concen-
trations were prepared and a number of complete analyses were performed at
each concentration.  The results of these tests were then compared to
determine analyses repeatability.  The test-to-test repeatabilities are
documented for all procedures in this report.  In most cases, repeatability
is difficult to obtain at the lower concentrations, while the repeatability
at high concentrations is easily obtained.

     The qualification experiments were performed to determine if the com-
pounds of interest could travel the length of the dilution tunnel in the
presence of dilute exhaust without significant loss by reaction with exhaust
or the tunnel itself.  The compounds were introduced at the same point at
which the exhaust enters the tunnel and were sampled at the normal sampling
point  (see Section 12).

     Qualification experiments were carried out on the aldehyde and ketone,
organic amine, sulfur dioxide, nitrous oxide, hydrogen sulfide, total
cyanide, organic sulfide, ammonia, and phenol procedures to determine the
recovery of known amounts of each pollutant from the CVS tunnel with and
without exhaust  (phenols CVS dilution tunnel with exhaust only).  Aldehydes
and ketones,: sulfur dioxide, nitrous oxide, total cyanide and phenols can
be recovered quantitatively from the CVS dilution tunnel with and without
(not done for phenols) exhaust.  There is a 10 percent loss of hydrogen
sulfide with and without exhaust present.  The organic amines, ammonia,
and the organic sulfides experience significant losses in the CVS dilution
tunnel with and without exhaust present.

     Despite the fact that the analytical procedures for the organic amines
and the organic sulfides have procedural detection limits of 2 and 0.2 ppb
respectively, the losses in the dilution tunnel could prevent the detection
of organic amines at levels lower than 20 ppb and the detection of organic
sulfides at levels lower than 10 ppb in dilute exhaust.  At ammonia levels
of 5-10 ppm there is a 25 percent loss of ammonia to the dilution tunnel and
an additional fifteen percent loss to exhaust.

     The procedures discussed in this report have been found to be the
optimum procedures at the time of .this report for collecting and analyzing
dilute exhaust samples and are recommended for use in this project.

-------
     A finalized copy of the analytical procedures discussed in Section 2-11,
the BCA sulfate procedure,  and DMNA procedure,  sampling conditions for DMNA,
and an outline for BaP collection and analysis  are included as an appendix.
The literature search, procedural development work, and validation experi-
ments for some of the compounds were carried out under another EPA Contract,
68-02-2497 (1).  The procedures discussed in this report were developed for
the measurement of pollutants in dilute exhaust.  The use of these procedures
for the measurement of pollutants in raw exhaust is not recommended without
additional validation and qualification work to document the acceptability
of the procedures.

-------
                                  SECTION  2

                       ALDEHYDE AND KETONE PROCEDURE
LITERATURE SEARCH

     The individual aldehydes and ketones that are included in this analysis
are:  formaldehyde, acetaldehyde, acetone, acrolein, propionaldehyde, iso-
butyraldehyde, methylethylketone, crotonaldehyde, hexanaldehyde, and benz-
aldehyde.  Acetone, acrolein, and propionaldehyde are not resolved from each
other under normal gas chromatographic operating conditions and all three
are reported together as acetone.  The common names, the International Union
of Chemists approved names, the chemical formulas, the molecular weights,
the melting points, the boiling points, the densities, the molecular weights
of the 2,4 dinitrophenylhydrazone derivatives, and the melting points of the
2,4 dinitrophenylhydrazone derivatives are presented in Table 1.  The alde-
hydes and ketones have a characteristicly pungent odor, are flammable, are
photochemically reactive, can cause respiratory problems, and are severe eye
irritants.  The 1976 American Conference of Government Industrial Hygienists
has recommended threshold limit values for several of the aldehydes and ke-
tones (2).  These values range from 0.1 ppm for acrolein to 1000 ppm for
acetone.  Other values listed were 2 ppm for formaldehyde and crotonaldehyde,
100 ppm for acetaldehyde and 200 ppm for methylethylketone.

PROCEDURAL DEVELOPEMENT

     A procedure, which is already in use at Southwest Research Institute,
developed by the Mobile Source Emissions Research Branch of the ESRL-EPA at
Research Triangle Park, North Carolina, was selected for the analysis of the
aldehydes and ketones (3).  This procedure involves bubbling exhaust through
glass impingers containing 2,4 dinitrophenylhydrazine  (DNPH) in dilute hydro-
chloric acid.  The exhaust sample is collected continuously during a test
cycle.  The aldehyde and ketones (also known as carbonyl compounds) react
with the DNPH to form their respective phenylhydrazone derivatives.  These
derivatives are either insoluble or only slightly soluble in the DNPH/HC1
solution and are removed by filtration followed by pentane extractions.
The filtered precipitate and the pentane extracts are combined, and the pen-
tane is evaporated in a vacuum oven.  The remaining dried extract contains
the phenylhydrazone derivatives.  The extract is dissolved in a quantitative
volume of toluene containing a known amount of anthracene as an internal
standard.  A portion of this extract is injected in to a gas chromatograph
and analysed using a flame ionization detector.  A copy of this procedure as
used by the Department of Emissions Research at Southwest Research Institute
will be included as an attachment to this report.

-------
                         TABLE  1.   PHYSICAL PROPERTIES OP THE ALDEHYDES  AND KETONES  (4,5)
                                                                                          Molecular
                                                                                            Weight
CO

Aldehyde or Ketone
Formaldehyde
Acetaldehyde
Acetone
Propionaldehyde
Acrolein
Isobutyraldehyde
Methylethylketone
Crotonaldehyde
Hexanaldehyde
Benzaldehyde

ICU Name
Methanal
Ethanal
2-Propanone
Propanal
Propenal
2-Methylpropanal
2-Butanone
transZ-Butenal
Hexanal
Benzenecarbonal
Chemical
Formvila
CH2O
CH3CHO
CH3COCH3
CH3CH2CIIO
CII2 :CHCHO
CH3CH(CH3)CHO
CH3COCH2CH3
CH3CH:CHCHO
CH3(CII2>4CHO
C6H5CHO
Molecular
Weight
30.03
44.05
58.08
58.08
56.07
72.11
72.11
70.09
100.16
106.13
Melting
Point
- 92
-121
- 95
- 81
- 87
- 65
- 84
- 74
- 56
- 26
Boiling
Point
- 21
21
56
49
53
63
80
105
128
178

Density
0.815
0.783
0.790
0.806
0.841
0.794
0.805
0.850
0.814
1.042
DNPH
Derivative
210.15
224.19
238.21
238.21
236.20
252.23
252.23
250.21
279.28
286.25
Melting
Point Der
167
168
128
156
165
182
	
190
104
237

-------
VALIDATION EXPERIMENTS

     Several experiments were carried out to determine the validity of the
DNPH procedure for the analysis of the aldehydes and ketones.  These experi-
ments included checks for:  GC injection variability, linearity of detector
response, sample stability in the DNPH absorbing solution and in toluene,
trapping efficiency of the DNPH/HC1 solution, interferences, and extraction
plus injection repeatability.

     The finalized sampling conditions used to collect the aldehydes and
ketones are listed below as is a discussion on their selection.  Two glass
impingers in series, each containing 40 m£ of 2 N HCl/2,4 dinitrophenyl-
hydrazine, are used to collect the aldehydes and ketones,  The two impingers
together trap 98+ percent of the carbonyl compounds.  This collection effi-
ciency was determined by bubbling known amounts of the aldehydes and ketones
through a series of impingers and analyzing each impinger separately.  No
advantage was found in using more than two impingers.  There was no observed
difference in analyzing the contents of the two impingers separately or com-
bined.  Since the analysis of the two impingers combined is less manpower
intensive, the two impingers are analyzed together.  During sampling, the
two impingers are kept in a 0°C ice bath.  The ice bath offers no signifi-
cant advantage in collection efficiency over room temperature, but does
provide a stable sampling temperature during the test.  The 0°C temperature
also lowers the vapor pressure of the aqueous absorbing solution and thus
prevents loss of any significant amount of water from the absorbing solution
during sampling.  The sample flow rate through the impingers is maintained
at 4 liters a minute.  This flow provides the largest amounts of sample to
flow through the absorbing reagent without loss in absorbing efficiency or
the physical loss of any absorbing reagent.  A heated filter is used to pre-
vent diesel particulate from contaminating the sampling system.  The filter
and the line connecting the filter to the dilution tunnel are heated to
375°F to prevent the aldehydes and ketones from being retained on the removed
particulate.  A Teflon line connecting the filter to impingers is heated to
175°F in order to prevent water from condensing in the sample line.  Some of
the aldehydes and ketones are water soluble, and the condensation of water
in the sample line could cause a significant loss of sample in the sample
line.

     The HCl/DNPH absorbing reagent has been found to be stable over several
days; however, to prevent the possibility of contamination or the inadver-
tant use of "old" absorbing reagent, the solution is prepared daily as
needed.

     The samples have been found to be stable for at least two days in the
absorbing reagent.  However, to prevent the possibility of contamination of
the samples by their standing in the lab for prolonged periods, the samples
are extracted, dried, and dissolved in toluene all in the same day.  Once
the sample is dissolved in toluene it is stable for relatively long periods
of time.  Samples run and re-run over a period of two weeks showed no signi-
ficant change in concentrations.

-------
     To determine the GC injection repeatability for the procedure over  a
wide range of concentrations, four standards containing 1.6, 0.2, 0.02 and
0.002 mg of each aldehyde and ketone DNPH derivative per m& of toluene were
prepared.  These are the concentration ranges expected when sampling dilute
exhaust.  Each standard was injected into the GC five consecutive times.
The concentration determined by the procedure for each of the derivatives was
averaged over the 5 runs, and a standard deviation  as well as a percent
standard deviation was calculated.  The results of these injection repeata-
bility experiments are presented in Table 2.  The injection repeatability is
good for the 1.6 mg derivative/m£ toluene standard  (percent deviation ranges
from 1.1 percent for formaldehyde to 9.6 percent for benzaldehyde) and the
0.2 mg derivative/mil toluene standard (percent deviation ranges from 0.5
percent for acetaldehyde to 5.9 percent for benzaldehyde).  At the two lower
concentrations, the standard deviation in the injection repeatability was
found to be much larger.  The 0.02 mg derivative/m& toluene standard gave
percent deviations ranging from 3.7 percent for acetone to 32 percent for
benzaldehyde.  The 0.002 mg derivative/m£ toluene standard gave percent
deviations which ranged  from 10 percent for formaldehyde to 110 percent  for
benzaldehyde.  It appears from the data that the injection repeatability is
good at higher derivative concentrations, but much more erratic at very  low
concentrations.

     To determine the linearity of the detector for the concentration ranges
of interest for each of  the derivatives, seven standard solutions were pre-
pared which contained 8.0, 4.0, 1.6, 0.8, 0.2, 0.02, amd 0.002 mg of each
derivative/mi toluene.   These standards were made by weighing out required
amounts of each derivative and dissolving them in the appropriate amount of
toluene to give the required concentrations.  The solution containing 0.2 mg
of each derivative/m£ toluene was used as the standard and the other six
solutions were compared  to this standard.  Figures 1-8  show plots of  the pro-
cedure determined concentration vs the actual concentration on a log-log
scale.  Acetone, methylethyIketone, and crotonaldehyde give linear plots
throughout tfie region of interest.  Formaldehyde, acetaldehyde, isobutyr-
aldehyde and hexanaldehyde give linear plots except at the lower concentra-
tions  (<0.02 mg derivative/m£ toluene).  Benzaldehyde gives a plot which is
not linear above 2.0 mg/m£ toluene.  The benzaledhyde-DNPH derivative is not
soluble in toluene at concentrations greater than 2.0 mg/mA.  This fact
should be taken into account if high concentrations of benzaldehyde are  ex-
pected (>5 ppm for a 23 minute sampling period at 4 A/minute) .

     An experiment was carried out to determine the extraction repeatability
for the DNPH procedure at low concentrations of DNPH-aldehyde derivatives.
One liter of DNPH absorbing solution containing small amounts of pure for-
maldehyde, acetaldehyde, acetone, methylethyIketone, crotonaldehyde, hexan-
aldehyde,  and benzaldehyde DNPH derivatives was prepared.  Seven extractions
(80 m£ for each extraction)  were carried out over a period of two weeks.
The results from the extractions are presented in Table 3.  These results
were determined in units of mg DNPH derivative/m£ of toluene.  Tha values
for each of the seven extractions, the average, and the standard deviation
are listed for each of the aldehydes and ketones.  Multiple injections were
                                     10

-------
TABLE  2.   INJECTION REPEATABILITY
DNPH Aldehyde
or Ketone
Derivative
Formaldehyde
Acetaldehyde
Acetone
Isobutyraldehyde
Methyl ethy Iketone
Crotonaldehyde
Hexanaldehyde
Renzaldehyde
1.6 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
1.567
1.727
1.617
1.561
1.575
1.781
1.682
1.710
0.018
0.022
0.018
0.017
0.029
0.053
0.105
0.165
1.1
1.3
1.1
1.1
1.8
3.0
6.2
9.6
0.2 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
0.188
0.201
0.210
0.208
0.206
0.206
0.204
0.222
0.007
0.001
0.002
0.002
0.003
0.003
0.003
0.013
3.7
0.5
0.9
1.0
1.5
1.5
1.5
5.9
                                     0.020 mg derivative/ml
                                              Standard
                                    Avg.  for     Std.     %
                                    5 Inject.    Dev.    Dev.
                                      0.022      0.002    9.1

                                      0.024      0.002    8.3

                                      0.027      0.001    3.7

                                      0.020      0.001    5.6

                                      0.024      0.001    4.2

                                      0.015      0.001    6.7

                                      0.015      0.001    6.7

                                      0.025      0.008   32
 0.002 mg derivative/ml
          Standard
Avg. for    Std.      %
5 Inject.   Ugv.    2.ev_-

  0.007    0.0007     JO
  0.002    0.0007     35
  0.003

  0.002

  0.001

  0.001
0.0007    23

0.0007    35

0.0007    70

o.ooii   no

-------
10.0 -
0.001
       0.002
0.005   0.01  0.02
Actual concentration
  0.05   0.1   0.2     0.5    i.o   2.0
(rag formaldehyde-DNPH derivative/ml toluene)
                                                                        5.0
                                                                              10.0
   Figure 1.   Plot of  the formaldehyde-DNPH  derivative concentration
            determined  by procedure vs  actual concentration.
                                        12

-------
 10.0 -
0.001
        0.002   0.005   0.01  0.02     0.05  0.1   0.2    0.5   1.0    2.0     5.0
                 Actual concentration (rag acetaldehyde-DNPH derivative/ml toluene)
10.0
     Figure  2.  Plot of the acetaldehyde-DNPH derivative  concentration
             determined by  procedure vs actual concentration.
                                         13

-------
10.0 -
0.001
        0.002    0.005  0,01  0.02 0.05     0.1   0.2     0.5   1.0   2.0
                   Actual concentration (mg acetone-DNPH derivative/ml toluene)
                                                                        5.0
 I

10.0
      Figure 3.   Plot of  the acetone-DNPH derivative concentration
             determined by procedure vs actual concentration.
                                       14

-------
     0.002    0.005  0.01  0.02   0.05    0.1   0.2    0.5    1.0    2.0     5.0
            Actual concentration (mg isobutyraldehyde-DNPH derivative/ml toluene)
10.0
Figure 4.  Plot of  the isobutyraldehyde-DNPH  derivative concentration
          determined by procedure vs  actual concentration.
                                      15

-------
0.001
        0.002   0.005  0.01  0.02    0.05   0.1    0.2     0.5    l.Q   2.0      5.0
            Actual concentration (mg methylethylketone-DNPH derivative/ml toluene)
10.0
 Figure 5.   Plot of the methylethylketone-DNPH derivative  concentration
             determined by procedure  vs actual concentration.
                                        16

-------
§
I
-H


I
-ri
U
    10.0
     5.0
     2.0
     1.0
I
IB
I
U

I
01
I
8
8
a
1
     0.5
     0.2
     0.1
    0.05
    0.02
    0.01
C  0.005
o
•H
-P
ID
g
U
   0.002 _
   0.001
                     I
                           I
                                 I
                                        I
                                              I
                                                    I
                                                           I
                                                                 I
                                                                       1
            0.002   0.005   0.01  0.02   0.05   0.1    0.2     0.5   1.0   2.0     5.0

                     Actual concentration  (mg crotonaldehyde-DNPH derivative/ml toluene)
                                                                                  10.0
      Figure 6.   Plot of the crotonaldehyde-DNPH derivative concentration
                 determined by procedure  vs actual concentration.
                                            17

-------
0.001
        0.002   0.005  0.01  0.02    0.05   0.1   0.2    0.5   1.0   2.0    5.0
                Actual concentration  (mg hexanaldehyde-DNPH derivative/ml toluene)
                                                                             10.0
    Figure 7.  Plot of the hexanaldehyde-DNPH derivative  concentration
             determined by procedure vs actual concentration.
                                       18

-------
 10.0
0.001
        0.002    0.005  0.01   0.02    0.05   0.1    0.2    0.5    1.0   2.0
                Actual concentration (mg benzaldehyde-DNPH derivative/ml toluene)
5.0   10.0
      Figure  8.   Plot of benzaldehyde-DNPH  derivative concentration
             determined by procedure  vs actual concentration.
                                        19

-------
                                TABLE 3.  MULTIPLE EXTRACTIONS OF DNPH SOLUTIONS

                                (All units are mg DNPH derivative/m£ of toluene)
ro
o
Extraction
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Average
Standard
Deviation
Form-
aldehyde
0.018
0.023
0.020
0.009
0.014
0.020
0.014
0.017
±0.005
Acet-
aldehyde
0.011
0.037
0.032
0.013
0.019
0.031
0.027
0.024
±0.010
Acetone
0.005
0.015
0.015
0.005
0.007
0.014
0.043
0.015
±0.013
MEK
0.000
0.003
0.000
0.000
0.000
0.004
0.014
0.003
±0.005
Cron ton-
aldehyde
0.003
0.004
0.003
0.002
0.003
0.005
0.004
0.003
±0.001
Hexan-
aldehyde
0.001
0.002
0.001
0.000
0.002
0.002
0.000
0.001
±0.001
Benz-
aldehyde
0.014
0.027
0.044
0.000
0.014
0.004
0.007
0.016
±0.015

-------
also carried out on each sample over the two week period.  The values ob-
tained generally did not vary more than 0.002 mg derivative/m£ toluene  (ex-
cept for two acetone DNPH values).  This finding indicates, for several of
the compounds, that a large part of the variation in values was due to the
extraction process and not the lack of injection repeatability.

     The results of these experiments indicate that the variability in the
extraction process for concentrations of  aldehyde  DNPH and ketone DNPH
derivatives below 0.025 mg DNPH derivative/m£ toluene can be very signifi-
cant (i.e., 94 percent for benzaldehyde at 0.016 mg/m£).  This variability
needs to be taken into account when evaluating data obtained using this
procedure.  At higher DNPH derivative concentrations (0.2-2.0 mg derivative
per m£ toluene) the overall test variability  (trapping, extraction and
injection) is approximately 15 percent.  This value was obtained from the
standard deviation of tunnel recovery and trapping efficiency experiments.

     The DNPH analysis for the aldehydes and ketones has given abnormally
high concentrations of crotonaldehyde and benzaldehyde in isolated occassions.
A gas chromatography-mass spectroscopy study was carried out on three samples
obtained from a gasoline powered vehicle.  The three samples either contained
abnormally high concentrations of crotonaldehyde-DNPH derivative or benz-
aldehyde-DNPH derivative or both.  The results from this study revealed that
neither crotonaldehyde nor benzaldehyde was present in the samples.  Further
gas chormatography-mass spectroscopy studies were carried out on two of the
samples to determine what compounds were present.  In both samples, the
crotonaldehyde peaks were due to a phthalate, and the benzaldehyde peaks
could not be identified.  This study revealed that the samples contained
several other phthalates as well as di-2-ethylhexyladipate (a fuel stabi-
lizer) .  Many phthalate esters  (e.g., dioctyl, dibutyl, dimethyl, etc.,)
are found in lubricants and plastics.  It is possible that the phthalate
peaks found in the above samples were due to contamination in the extraction
process (e.g., from a pipette bulb, etc.).  In subsequent testing, extreme
care will be taken to assure the samples do not come into contact with
plastics and other materials which could cause contamination.  It is also
possible that some of the phthalates which are found in small quantities in
the samples are from the exhaust  (orignating from lubricants) and are
possible interferences in the procedure.  Also, the di-2-ethylhexyladipate
appears to produce a minor interference.

QUALIFICATION EXPERIMENTS

     Qualification experiments were carried out using a Mercedes 240D vehicle.
Hot FTP (23 minute test) driving cycles were followed to generate exhaust
for the vehicle baseline emissions and for the tunnel plus vehicle experi-
ments.  Aluminum cylinders containing 1350 and 436 ppm propionaldehyde in
balance nitrogen were used as the source for aldehydes in the experiments.
The cylinders were named using the aldehyde-DNPH procedure.  The flow of
propionaldehyde into the tunnel was regulated to give concentrations of 0.5-
2 ppm propionaldehyde in the dilution tunnel.  Injections of propionaldehyde
into the tunnel without exhaust gave recoveries that ranged from 85 to  115
                                     21

-------
percent with an average of 102 percent (Table 4).  The recovery of propion-
aldehyde in the presence of vehicle exhaust and without a heated filter
ranged from 59 percent to 89 percent with an average of 76 percent (Table 4)
The recovery of propionaldehyde from the dilution tunnel in the presence of
exhaust while using a heated filter ranged from 82 to 120  percent for an  ,
average of 99 percent (Table 4).   The injections with the  vehicle were cor-
rected for the vehicle baseline emission of propionaldehyde.   If a heated
line and filter is used to remove  particulate and if propionaldehyde  is re-
presentative of the aldehydes, then it appears that there  is  little or no
loss of aldehyde in the dilution tunnel with or without vehicle exhaust.
                     TABLE 4.   PERCENT  RECOVERY  OF PROPIONALDEHYDE
       Tunnel Only
Tunnel + Vehicle
No Heated Filter
Tunnel + Vehicle
  Heated Filter
Run
1
2
3
4
5
6
Avg
Recovery %
85
115
99
96
110
106
102 ± 11
Run
1
2
3
4
5
6
Avg
Recovery
83
86
89
76
64
59
76 ± 12
Run
1
2
3
4
5
6
7
8
Avg
Recovery
86
82
120
106
85
107
103
103
99 ± 13
 RESULTS AND CONCLUSIONS

     The concentration of aldehydes and ketones in dilute exhaust can be
 determined by  (1) trapping the aldehydes and ketones in a DNPH/HC1 absorbing
 solution,  (2) removing the resulting derivative from the absorbing solution
 by filtration and extraction with pentane,  (3) evaporating off the pentane
 (4) dissolving the dried extract in toluene, and  (5) analyzing the resulting
 solution with a gas chromatograph equipped with a flame ionization detector.
 The aldehydes and ketones are effectively trapped in the absorbing solution
 at a flow rate of 4 Vminute.  The procedure has a minimum detection limit
 of approximately 5 ppb.  This carbonyl concentration in the exhaust gives a
 corrsponding concentration of 0.002 mg/m£ in toluene.
                                     22

-------
     The accuracy of the procedure in the 0.5-20 ppm concentration range for
the aldehydes and ketones in dilute exhaust is approximately 10-15 percent.
The accuracy of the procedure in the 0-0.05 ppm range is not as good and
values can vary as much as 100 percent.  The gas chromatograph system gives
a linear response for acetone, methylethylketone, and crotonaldehyde DNPH
derivative concentrations  between 0.002 and 8 mg derivative/ltd toluene and
gives a linear response for formaldehyde, acetaldehyde, isobutyraldehyde and
hexanaldehyde DNPH derivative concentrations between 0.02 and 8 mg derivative/
m£ toluene.  The benzaldehyde-DNPH derivative gives a linear response in the
0.02 to 2 mg derivative/m£ toluene concentration range.  The benzaldehyde
derivative is not soluble at concentrations greater than 2 mg/m£ toluene.

     Phthalates and di-2-ethyhexyladipate were found by mass spectroscopy
to be interferences in the procedure.  Many phthalate esters (e.g., dioctyl,
dibutyl, dimethyl, etc.) are found in lubricants and plastics, and di-2-
ethylhexyladipate is used as a fuel stabilizer.  Contamination from phtha-
lates could occur in the extraction process or in sample storage if the
sample is allowed to come into contact with plastics, a pipette bulb, a
lubricant, etc.  It is also possible that some phthalates originate from the
exhaust  (from lubricants) and are possible interferences in the procedure.
The benzaldehyde and crotonaldehyde values can be affected by these inter-
ferences .  The interfering peak in the region of benzaldehyde is usually
broad and the benzaldehyde peak, if present, can be observed on top of this
interference.  If care is taken, a reliable value can be determined for the
benzaldehyde.  Any value reported for crotonaldehyde may be artifically
high due to possible phthalate contamination.  Extreme care must be taken
when handling the sample in order to eliminate any possibility of contami-
nation after collecting the sample and before analysis.

     Propionaldehyde can be recovered quantitatively from the dilution tunnel
with or without diesel exhaust present if a heated filter is used.  If pro-
pionaldehyde is representative of the aldehydes and ketones, there is little
or no loss of the aldehydes in the dilution tunnel with or without exhaust
present.

     Overall the DNPH procedure should provide a relatively accurate method
for determining the concentration of aldehydes and ketones in dilute exhaust,
and its use is recommended for this project.
                                    23

-------
                                 SECTION 3

                           TOTAL CYANIDE  PROCEDURE
LITERATURE SEARCH

     Hydrogen cyanide is a flammable, toxic, and colorless liquid at room
temperature and has the characteristic odor of bitter almonds.  Some synonyms
for hydrogen cyanide are hydrocyanic acid, prussic acid, and formonitrile.
Hydrogen cyanide is a covalent molecule and dissociates in an aqueous solu-
tion as do the hydrogen halides.  Hydrogen cyanide (HCN) has a molecular
weight of 27.03, a boiling point of 24.70°C, and a melting point of -13.42°C.
It is a linear molecule with C-H and C.HN bond distances of 1.06 and 1.15 A,
respectively.  It is a weak monoprotic acid with a dissociation constant of
2.1 X 10~9.  This highly poisonous compound is a respiratory inhibitor and
irreversibly combines with the iron complex in the blood, stopping the oxi-
dation processes in tissue cells and causing death by asphyxiation.  Com-
mercially, hydrogen cyanide is prepared by reacting methane, ammonia, and
air over a platinum catalyst at 1000-1200°C, by the reaction of nitric oxide
and gasoline at 1400°C, the reaction of hydrocarbons, ammonia and oxygen
at 600-1500°C, and many other methods.  Reactions similar to these may be
responsible for the hydrogen cyanide produced in exhaust.

     Cyanogen is a flammable, toxic,  and colorless gas at room temperature
and like hydrogen cyanide, has the characteristic odor of bitter almonds.
Some synonyms for cyanogen are dicyan, oxalic acid, dinitrile, and oxaloni-
trile.  Pure cyanogen is stable, although the impure gas may polymerize to
paracyanogen  between 300° and 500°C or by exposure to ultraviolet light.
Cyanogen dissociates into CN radicals and can oxidatively add to lower valent
metal atoms, giving dicyano complexes.  It resembles halogens in the dis-
proportionation reaction in basic solution:

                     (CN) 2 + 20H~ —>• CN~ + OCN~ + H2O

Cyanogen (C2N2> has a molecular weight of 52.04, a freezing point of -27.9°C
and a boiling point of -21.17°C.  Cyanogen is a symmetrical and linear mole-
cule with a C-C bond distance of 1.37 A and a C=N bond distance of 1.13 A.
Its physiological effect on living tissue is similar to that of hydrogen
cyanide.  Cyanogen is prepared by many methods:  air oxidation of hydrogen
cyanide over a silver catalyst at 300-600°C, passage of hydrogen cyanide over
cuprous oxide.at ambient temperatures, reaction of hydrogen cyanide and
chlorine over a surface-active material such as activated charcoal at >700°C,
any many others.  In all cases above, cyanogen is produced from hydrogen
cyanide.  Although none of these are exactly applicable for an automotive


                                      24

-------
system, a similar process may be responsible  for any cyanogen that is produced.

     The analyses for hydrogen cyanide, cyanogen, and/or cyanide ion has been
performed by several basic analytical techniques:  titration, colorimetry,
specific ion electrode, and gas chromatography.  The Liebig determination of
cyanide ion by titration with silver ion was  discarded as a means of analysis
because of the low concentrations that were expected from exhaust samples.
Colorimetry has previously been used by SwRI  and has been found to be man-
power intensive.  An alternative procedure was sought with this factor in
mind.   The best means of analysis was with either a specific ion electrode
or a gas chromatograph.

     Three acceptable procedures were selected from the literature.  Sekerka
and Lechner  (6) reported the use of a cyanide ion-selective electrode for the
analysis of cyanide ion in waste water.  The  specific ion electrode was used
in conjunction with a colorimetric technique  to determine the reliability of
the procedure.  The samples were collected in sodium or potassium hydroxide
and analyzed potentiometrically.  The minimum detectable limit reported was
about 2 ppb.  The second technique required the use of Tedlar bag samples and
subsequent analysis with a gas chromatograph  using a nitrogen phosphorus
detector  (NPD).  The third technique reported by Valentour et al  (7) was used
with biological samples  (blood, urine, and gastric contents).  The samples
were collected in sodium of potassium hydroxide and the trapped cyanide ion
was reacted with chloramine-T to produce cyanogen chloride.  The cyanogen
chloride was then analyzed with a gas chromatograph using an electron capture
detector  (BCD).  After preliminary experiments, the final analytical procedure
selected was a significant modification of the Valentour et al procedure.

PROCEDURAL DEVELOPMENT

     Attempts to analyze hydrogen cyanide and cyanogen separately were un-
successful and the details are reported below.  The inability to analyze
hydrogen cyanide and cyanogen separately led  to consideration of several
specific procedures for the analysis of hydrogen cyanide and cyanogen in the
form of cyanide ion.

     Initially, it was decided to determine the concentration of hydrogen
cyanide and cyanogen by collecting a bag sample of the dilute exhaust and
analyzing it with a gas chromatograph using a nitrogen phosphorous detector.
This detector was selected because of its specificity to carbon-nitrogen
compounds.  Hydrogen cyanide and cyanogen can be resolved with a 6' X 1/4"
O.D. glass column packed with 100/120 mesh Porapak QS.  Isothermal column
temperature operation at 50°C and a helium carrier gas flow rate of 60 mVmin
were the column conditions.  A glass lined injector and interface were also
used to preserve sample integrity.

     Bag stability experiments with hydrogen  cyanide and cyanogen were con-
ducted to determine if sample integrity could be maintained over a short
period of time.  Bag stability is necessary due to time required to collect
the sample and the subsequent waiting period  before the sample can be analyzed.
The bag sample lifetime should be at least two hours after a sample is
collected.  Clear and aluminum foil tape covered Tedlar plastic bags were used

                                      25

-------
to conduct bag stability experiments.  Dark bags  (aluminum foil tape  covered)
were tested to determine the effect of photochemical decomposition on hydro-
gen cyanide and cyanogen.

     A list of bag stability experiments which were conducted is  shown  in
Table 5.  Each bag contained approximately one cubic foot of the  dilute gas.

            TABLE 5.  EXPERIMENTS CONDUCTED FOR HCN AND C-N  BAG STABILITY
Compound
HCN
C2N2
HCN & C N
HCN & C2N2
Clear Bags
Nitrogen Air
X X
X X
2
*

Exhaust
X
X
X

Dark
Nitrogen
X
X
X
X
Bags
Air
X
X
X


Exhaust
X
X
X

       * Blend of hydrogen cyanide and cyanogen in humid nitrogen.

 Experiments were conducted with nitrogen, air, dilute exhaust and humid nitro-
 gen.   Dilute exhaust was selected at random from bag samples generated during
 other tests, and humid nitrogen was generated by passing nitrogen through
 an impinger containing deionized water.  Hydrogen cyanide and/or cyanogen were
 then  added to each bag to give a nominal concentration of about 2 ppm.  At
 twenty to  thirty minute intervals, 5 mH of the gas was removed with a glass
 gas-tight  syringe and injected into the gas chromatograph.  The percent change
 in the concentration was then calculated using the initial injection.  Figures
 9  through 13 show the effect of elapsed time on the stability of hydrogen
 cyanide and cyanogen.  Figure 9 demonstrates the stability of hydrogen cyanide
 in clear and dark bags with a variety of atmospheres.  Peak areas for hydro-
 gen cyanide remained within the nominal range of injection variability for at
 least 80 minutes, and no definite trends were observed.  (The nominal range
 of injection variability was set at ± 7 percent and is indicated in all
 figures by a dotted line).  On the other hand, cyanogen showed a considerable
 percent loss in the clear bags with both nitrogen and exhaust  (Figure 10) .
 In the dark bag, cyanogen remained stable except in the presence of exhaust.

      Figures 11 and 12 show the effect of a blend of hydrogen cyanide and
 cyanogen in clear and dark bags with the various atmospheres.  Again, hydro-
 gen cyanide was stable within the limits of injection variability for a short
 period of  time.   In both cases, hydrogen cyanide was stable on the order of
 about  60 minutes.  Cyanogen behaved similarly to hydrogen cyanide in the
 clear  bag, but a steady decrease in concentration was observed in all atmos-
pheres with the  dark bag.

     In humid nitrogen (Figure 13) there was a 70 percent loss of hydrogen
cyanide after only 20 minutes.  After this initial loss, the level of hydro-
gen cyanide remained relatively constant.  Cyanogen under the same conditions

                                     26

-------
0)
(Jl
c
(0

u
C
0)
O
t-i
0)
CM
   is r-
   10
            Legend
           Nitrogen

           Air

           Exhaust
                                              Nominal  Range for Injection

                                              Repeatability        .
-5
   -10
   -15
                                    _L
                                          _L
               20        40         60         80

                             Time,  Minutes


                     Hydrogen  cyanide  in  clear bags
                                                   100
                             120
   15
   10
 cu
 0i
O

-P
C
0)
O
^
0)
Hi
                                          Nominal Range for Injection

                                           Repeatability
    -5
   -10
   -15'
      n.
                      _L
J_
               20        40         60        80        100

                             Time,  Minutes

                     Hydrogen cyanide  in dark  bags
                                                              120
             Figure 9.   Effect of elapsed time on hydrogen
                     cyanide in clear and dark bags.
                                    27

-------
 30
 20
 10
            Legend
           Nitrogen
        •  Air
        A  Exhaust
                                            Nominal Range for
                                            Injection Repeatability
                     40        60
                         Time, Minutes
                      Cyanogen in clear bags
                               100
                                                            120
                       „Nominal Range for
                        Injection Repeatability
-10
-20 -
-30
 40        60        80       100
     Time, Minutes

Cyanogen in dark bags
                                                            120
           Figure 10.  The effect of elapsed  time  on
                cyanogen in clear and dark bags.
                                 28

-------
   30 r-
   20  _
    10
c
(0
     o
-p
c
0)
Q

oJ  -10
d.
   -20
   -30
    30
    20
d)   10
     0
-p
c
o
o

0  -10
   -20
   -30
 Legend


Nitrogen


Air

Exhaust
                                               Nominal Range for

                                               Injection Repeatability
                                    I
                                I
               20
          40
                      80
             60


       Time/ Minutes


Hydrogen cyanide in clear bags
100
120
                                  Nominal Range for

                                  Injection Repeatability
                          J_
               20
          40
                                             80
            60


      Time, Minutes


Hydrogen cyanide in dark bags
                                100
         120
  Figure  11.   The effect of elapsed time on hydrogen cyanide in a blend

      of hydrogen cyanide and cyanogen in clear and drak bags.
                                    29

-------
a
A
u
•P
0)
u
i-l
0)
c
JS
u
c
0)
o
    so r~
    20
     10
    -10
    -20
    -30
                Legend

               Nitrogen
               Air
               Exhaust
                                                Nominal Range for
                                                Injection Repeatability
                          I
                20
     30
     20
     , n
     1C
                          40         60        80
                              Time, Minutes

                          Cyanogen in clear bags
          100
120
   -10
   -20
   -30
                                               Nominal Range  for
                                               Injection Repeatability
      T.
                         x
                                   _L
_L
       0       20        40        60       80        100      120

                              Time, Minutes

                         Cyanogen in dark bags

   Figure 12.  The effect of elapsed time  on  cyanogen in a blend
       of hydrogen cyanide and cyanogen  in clear and dark bags.

                                    30

-------
CO
           20



           10



            0



          - 10


          - 20
        C
        rfl
       £

       0
          -40
0)

Z -50

-------
showed only a slight decrease in  concentration.

     The short bag lifetime of hydrogen cyanide and  cyanogen prevent the use
of grab samples of exhaust.  At this point, the alternative procedures  were
investigated for the analysis of hydrogen cyanide and cyanogen.   These  tech-
niques required samples to be collected in an aqueous solution.
t
     At the same time the work with bag samples was  underway,  efforts to
develop a procedure using the specific ion electrode were being  conducted.
Potassium cyanide standard solutions were prepared with 0.1 M  potassium
hydroxide.   A calibration curve was to be determined by plotting the measured
potential in millivolts as a function of the log of  the cyanide  ion concen-
tration.  Instability of the potentiometrie measurement was observed in all
concentration ranges, especially  in the low concentration range.   Attempts
to  improve the electrode stability and potential drift were unsuccessful.
Efforts using the specific ion electrode were abandoned for another pro-
cedure using gas chromatography.

     A gas chromatograph procedure, which did not require bag samples for the
collection of hydrogen cyanide and cyanogen, was investigated.  This procedure
used a chemical collection of cyanide ion in sodium  or potassium hydroxide.
Initially, the analysis was to be conducted by reacting chloramine-T with the
trapped cyanide ion in an acid buffered solution to  produce cyanogen chloride.
Cyanogen chloride was then extracted by hexane and analyzed with  an electron
capture detector.  The electron capture detector was chosen because of  its
high sensitivity and selectivity to halogenated compounds and  relative  in-
sensitivity to hydrocarbons.  Problems with impurities in the  hexane caused
broad peaks with an excessive analysis time.  To eliminate the problem,  the
following items were tried:

     1.  Temperature program sequences

     2.  Column backflush

     3.  Column changes

         A.  6' X 1/4" O.D. glass column packed with 100/120 mesh Porapak QS
         B.  6' X 1/4" O.D. stainless steel column packed with 50/80 mesh
             Porapak Q
         C.  6* X 1/4" O.D. stainless steel packed with 7 percent Hallcomid
             M-18 on 90/100 mesh Anakrom ABS

     4.  Hexane purification with charcoal

     5.  Other extracting solvents (i.e., cyclohexane, etc.)

None of these proved to be satisfactory and long analysis times were the
result.

     A modification of the above procedure was tried by eliminating the
hexane layer and conducting the analysis in the same manner as described above,
except that the sample was placed in an air tight reaction vial with a  1 m&

                                      32

-------
head space and a septum cap.  A sample development period of 5 minutes was
required.  After vigorously shaking the vial for 5 seconds, 100 y£ of the
head space was injected into the gas chromatograph.  An electron capture
detector was used for the analysis.  As a result of this modification, a
rapid analysis time was achieved.  The finalized analytical procedure is
included as an attachment of this report.

VALIDATION EXPERIMENTS

     After selecting an analytical method, validation experiments were con-
ducted to determine detector linearity, detection limits, injection repeat-
ability, stability of reagents and sample, sampling parameters, etc.  Once
the validation  experiments were complete, the procedure was considered
ready for testing.

     Collection parameters were determined with a series of experiments de-
signed to check sample flow rates, absorbing reagent concentration, absorbing
reagent temperature, impingers or fritted glass bubblers, and collection
efficiency.  All of these experiments were conducted with hydrogen cyanide.

     The first experiments conducted were to determine the effects of stopper
tip, sample flow rate, the reagent concentration on the collection efficiency.
The results of these experiments are shown in Table 6.  A set of three of the
              TABLE  6.   THE EFFECT OF STOPPER TIP AND ABSORBING
     REAGENT  CONCENTRATION ON COLLECTION EFFICIENCY AT ROOM TEMPERATURE
                                          yg CN /ft  sample
  Collection    Sample
    Device       Flow

    impinger      1.0
    impinger      1.0
    impinger      1.0
   bubbler       1.0
   bubbler       1.0
   bubbler       1.0
   impinger     4.0
   impinger     4.0
   impinger     4.0
   bubbler      4.0
   bubbler      4.0
   bubbler      4.0
1.0 N KOH
Run
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
93.47
65.71
74.28
77.82
11.44
0.00
0.00
3.81
37.15
29.18
28.96
31.76
20.20
25.30
18.94
21.48
2
7.71
9.26
6.38
7.78
0.00
0.00
0.00
0.00
5-35
3.18
2.90
3.81
1.41
2.96
2.10
2.16
3
1.55
2.61
0.00
1.39
0.00
0.00
0.00
0.00
0.67
0.60
0.35
0.54
0.00
3.52
0.32
1.28
0.1 N KOH
1
50.02
60.13
49.92
53.36
4.64
1.11
2.57
2.77
29.60
27.20
27.24
28.01
15.14
7.42
16.21
12.92
2
7.08
6.06
6.16
6.43
1.40
0.00
0.00
0.47
3.53
3.24
2.69
3.15
1.85
0.85
1.78
1.49
3
6.21
0.00
5.48
3.90
2.73
0.00
0.00
1.37
0.77
1.08
0.62
0.82
0.00
0.00
0.00
0.00
                                      33

-------
same type collection devices (impinger or fritted glass tipped bubblers) were
filled with 1.0 N or 0.1 N potassium hydroxide.  A hydrogen cyanide calibra-
tion blend that contained a nominal 2 ppm concentration in a balance of nitro-
gen was passed through the absorbing reagent at 1.0 and 4.0 £/min.  Each
experiment was repeated three times.  All of these experiments were conducted
in a special blending building, which was external to the main building. This
building did not have the normal temperature controls within the building and
the ambient temperature fluctuated with the weather.  The room temperature
ranged from about 15 to 30°C during the experiments.

     After careful examination of the data, several trends can be observed.
First, in all cases, more cyanide ion was collected with the stronger absorb-
ing reagent.  Secondly, more cyanide ion was trapped with the impinger than
the fritted glass tipped bubbler.  The possible reason for this was a flow
restriction due to the fritted glass tip.  Finally, the higher flow rate
produced more consistent results with both concentrations of the absorbing
reagent.

     The next set of experiments took into accoutre the results of the first
set plus the effect of reagent temperature.  Five sets of three impingers
filled with 25 mi each of 1.0 N potassium hydroxide absorbing reagent were
used.  The sample flow rate was set at 4.0 Jl/min.  The first set of impingers
was sampled at ambient room temperature  (16-29°C) and a second set of im-
pingers was sampled at ice bath temperatures.  The sample collection effici-
ency for the ambient temperature experiments showed a high degree of vari-
ability.  The collection efficiency for the first impinger was between 70
and 100 percent.  At this temperature three impingers would be necessary to
collect the entire sample even at low concentrations.  With the ice bath, the
first bubbler was sufficient to collect the entire sample as well as giving
more consistent results.  The data for these experiments is shown in Table 7.

     Detector linearity was demonstrated for two cyanide ion concentration
ranges.  A linear response was observed in the 0 to 2 and the 0 to 10 yg
CN~/m£ ranges.  Table 8 and Figures 14 and 15 show the detector linearity.
All samples are expected to be within this concentration range.  If samples
are obtained that are not in these regions, the samples will be diluted to
a  concentration which falls within the linear response of the detector.

     Sample injection reproducibility is essential for an gas chromatography
technique which does not involve the use of internal standard.  To establish
sample injection reproducibility, two nominal cyanide ion concentrations,
2.0 and 0.2 yg/m£ , were used.  Five separate samples of each concentration
were developed and injected.  The results are shown in Table 9.

     Three separate experiments involving the sample storage and sample
stability were also conducted.  Three separate samples of known concentration
were developed for the required time and injected as usual.  At thirty-
minute intervals, 100 y£ of the remaining head space was also injected.  The
decay of the peak areas for a two-hour period is shown in Figure 16.  Five
separate samples of equal concentration were developed for varying lengths
of time.   The first sample was injected immediately, the second after 30
minutes,  the third after 60 minutes, the fourth after 90 minutes, and the

                                     34

-------
          TABLE 7.   THE EFFECT OF ABSORBING REAGENT TEMPERATURE
                       ON HCN COLLECTION EFFICIENCY
Absorbing Reagent
Temperature
Date
10/1.1/77
10/13/77
10/17/77
10/J7/77
10/17/77
GO
cn
10/12/77
10/12/77
10/12/77
10/12/77
10/12/77
Run
1
2
3
4
5
1
2
3
4
5
op
72
61
63.
73
U4
32
32
32
32
32
°C
22
16
17
23
29
0
0
0
0
0
Time
min
20
20
20
20
20
20
20
20
20
20
Flow
A/min
4.0
4.0
4.0
4.0
4.0
Avg
4.0
4.0
4.0
4.0
4.0
Avg
pg CN~/ft3 pg CN~/m3
1
48.61
64.98
35.27
30.06
24.46
40.68
55.76
51.96
57.86
54.00
51.05
54.13
2
4.39
0.00
7.02
10.33
8.09
5.97
0.00
0.00
0.00
0.00
0.00
0.00
3
0.00
0.00
3.76
3.95
2.26
1.99
0.00
0.00
0.00
0.00
0.00
0.00
Total
53.00
64.98
46.05
44.34
34.81
48.64
55.76
51.96
57.86
54.00
51.05
54.13
1
1716.6
2294.7
1245.5
1061.6
863.8
1436.7
1968.1
1834.9
2043.3
1907.0
1802.8
1911.6
2
155.0
0.0
247.9
364.8
285.7
210.7
0.0
0.0
0.0
0.0
0.0
0.0
3
0.0
0.0
132.8
139.5
79.8
70.3
0.0
0.0
0.0
0.0
0.0
0.0
Total
1871.7
2294.7
1626.2
1565.8
1229.3
1717.4
2147.1
1834.9
2043.3
1907.0
1802.8
1947.2
1
1.54
2.07
1.12
0.97
0.80
1.30
1.77
1.67
1.86
1.92
1.65
1.77
ppm CN~
2
0.14
0.00
0.22
0.33
0.26
0.19
0.00
0.00
0.00
0.00
0.00
0.00
3
0.00
0.00
0.12
0.13
0.07
0.06
0.00
0.00
0.00
0.00
0.00
0.00
Total
1.68
2.07
1.46
1.43
1.13
1.55
1.77
1.67
1.86
1.92
1.65
1.77

-------
TABLE 8.  CALIBRATION CURVE LINEARITY AT SEVERAL CYANIDE CONCENTRATIONS
Background
CN Cone .
Test Date yg/m&
1 10/11/77 9.64
4.82
1.93
0.96
0.00
2 10/12/77 9.64
4.82
1.93
0.96
0.00
3 10/03/77 1.93
0.96
0.48
0.19
0.00
4 10/04/77 1.93
0.96
0.48
0.19
0.00
GC
Attn
X256
X256
X256
X256
X256
X256
X256
X256
X256
X256
X64
X64
X64
X64
X64
X64
X64
X64
X64
X64
Sample
height
45
26
11
5
1
40
24
9
5
0
80
42
19
8
1
68
36
17
6
0
area
4650
2644
1182
583
50
4102
2448
953
518
0
7939
4146
1947
817
123
6813
3538
2027
643
0
Corrected
height
44
25
10
4
0
40
24
9
5
0
79
41
18
7
0
68
36
17
6
Q
area
4600
2594
1132
533
0
4102
2448
953
518
0
7816
4023
1824
694
0
6813
3538
2Q27
643
0
                                36

-------
                              LEGEND
CO

-•J
                 10.-
o
o
               •s
               (V
                                 10/03/77






                                 10/04/77
                             0.2        0.4       0.6        0.8      1.0       1.2       1.4




                                                   Cyanide Ion Concentration, |Jg CN~/ml
                                                                                               1.6
                                                                                                        l.H
                   Figure  14. Total cyanide calibration curve at  low concentrations  (0-2 ppm).

-------
co
CO
                                            34567

                                               Cyanide Ion Concentration, |ig/CN~/ni£
10
                    Figure 15.   Total  cyanide calibration  curve at low concentrations  (0-10 ppm).

-------
          TABLE 9.  SAMPLE INJECTION REPEATABILITY FOR
                   TWO CYANIDE CONCENTRATIONS
              Nominal
Sample

   1
   2
   3
   4
   5
   x
   sx
   Cv
   1
   2
   3
   4
   5

   x
   sx
   Cv
2.0
2.0
2.0
2.0
2.0
0.2
0.2
0.2
0.2
0,2
 GC
Attn

X256
X256
X256
X256
X256
                                      Peak
X32
X32
X32
X32
X32
Height
62
60
61
61
63
61.4
1.1
1.9
66
70
68
65
64
66.6
2.4
3.6
Area
6807
6550
6675
6703
6913
6730
137
2
6621
7095
6838
6711
6544
6762
216
3







.5
.0






.0
.2
                                 39

-------
      0  i-
C
(S

u
C
CL)
U
M
0)
CM
 (U
 cn
 •P

 d>
 U
 M
     20
40
      60
                        30           60


                                Time, Minutes

                          Sample decay with time
                                              90
                                                                 120
      10
15
      20
      25  -
                       30
                                60
90'
120
                                Time, Minutes

                  Five samples with varying development time



    Figure 16.  The effect of elapsed time on sample development.


                                   40

-------
fifth after 120 minutes.  The sample decay, as a function of time,is also
shown in Figure 16.  In both cases, the concentration of cyanogen  chloride in
the head space is dependent on the length of time in which the sample was
developed.  The third experiment involves the effect of real exhaust samples
that have been stored over a period of time.  Sample storage stability is
necessary when samples cannot be processed immediately or if confusing data
is to be checked at a later date.  A random sample was chosen and  reprocessed
periodically for 50 days.  The results are shown in Figure 17.  As a result,
samples can be stored for a period of  several weeks without adverse effects.

     The freshness and stability of the reagents is also very important for
the quantitative analysis of total cyanide.  Solutions of both chloramine-T
and the buffer were stored for various lengths of time.  Samples developed
with these stored solutions were found to be inferior to freshly prepared
reagents.  For these reasons, the reagents should be prepared daily.

     Several ions were tested for interference with the production of cyanogen
chloride or the production of other compounds with similar retention times in
the column.  Those ions tested were sulfate, phosphate, permanganate, nitrate,
carbonate, chloride, bromide, cyanate, thiocyanate, and ammonium ions.  The
potassium salts of each of these ions were prepared in 100 ppm and 1 ppm
concentrations in the presence of 4 ppm cyanide ion.  The sulfate and nitrate
salts of ammonium ion were then tried after the potassium salts of the sulfate
and nitrate ions were found not to interfere.  Aliquots of each were then
developed for cyanogen chloride.  Sulfate, phosphate, nitrate, carbonate, and
ammonium ions showed no effect on the development of cyanogen chloride in the
100 ppm or 1 ppm ranges.  Chloride, bromide, and permanganate ions produced
little or no effect at low concentrations.  At high concentrations, both
bromide and permanganate ions decreased the concentration of cyanogen chloride
produced.  On the other hand, chloride ion increased the concentration.
Cyanate and thiocyanate ions produced a positive interference at both concen-
trations.  Apparently, these two ions also form a halide in the presence of
chloramine-T with the same retention times as cyanogen chloride.

QUALIFICATION EXPERIMENTS

     Qualification experiments for the total cyanide procedure were conducted
with a Mercedes 240D.  Hot FTP  (23 minute test) driving cycles were followed
to generate exhaust for the vehicle baseline emissions and for the tunnel
(18 inch diameter) injection + vehicle experiments.  A cylinder containing
485 ppm hydrogen cyanide in balance nitrogen was used as the source for hydro-
gen cyanide.  The flow of cyanide into the tunnel was regulated to give a
concentration of 0.5 to 1 ppm hydrogen cyanide in the dilution tunnel.

     The baseline emission rate for the Mercedes 240D was-0.01 ppm.  Injec-
tion of hydrogen cyanide into the tunnel without exhaust gave recoveries
that ranged from 82 percent to 108 percent with an average of 98 percent
(Table 10).  The recovery of hydrogen cyanide in the presence of vehicle
exhaust without a filter to remove particulate from the sampled exhaust gave
recoveries that ranged from 68 to 84 percent with an average of 76 percent
(Table 11).  The recoveries ranged from 75-85 percent  (Table 11) when a non-
heated 0.5 y Fluoropore filter was used to remove particulate from the sampled

                                      41

-------
   O.SO
   0.70
   Q.60
I
a
c
0  0.50
§  0.40
   0. 30;
0)
T3
•H
C'
u
   0.20.
   0.10
      0.
                                                                       	I
                                                                       50
10
20
30
40
                                   Time, Days
                  Figure 17.  Time-sample decay curve.
                                    42

-------
                   TABLE 10.  TOTAL CYANIDE GASEOUS RECOVERY
                            BY DIRECT CVS INJECTION
Actual ppm
 Injected

    485
    485
    485
    485
    485
    485
    485
    485
 .Nominal Flow
Rate,  ft /min
HCN Blend
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
CVS
270
270
270
270
270
270
270
270
                 Run
                  1
                  1
                  2
                  2
                  3
                  3
                  3
Sample
1
2
3
1
2
1
2
3
Calculated
ppm HCN
dilute
0.60
0.60
0.60
0.61
0.61
0.60
0.60
0.60

Observed
ppm*
0.60
0.61
0.62
0.60
0.66
0.57
0.58
0.49
Average
Percent
Recovery
HCN
100
102
103
98
108
95
97
82
98 ± !
 Background  subtracted from observed concentration  (0.03 ppm)
              TABLE 11.  TOTAL CYANIDE RECOVERY FROM DILUTE EXHAUST
                    WITHOUT FILTER OR WITH NON-HEATED FILTER
                           Total Cyanide  Cone.  as HCN, ppm
Actual ppm
 Injected   Run
              Calculated
     Sample     Amount    Observed
                                non-filtered
   485
   485
   485
   485
   485
1
1
2
3
3
1
2
1
1
2
0.60
0.60
0.61
0.62
0.62
0.42
0.47
0.50
0.53
0.44
                              Corrected*
0.41
0.46
0.49
0.52
0.43
                                  Percent
                                  Recovery
68
77
80
84
69
                                                           Average 76 ± 7%
                             filtered/non-heated**
   485
   485
   485
4
5
6
1
1
1
0.60
0.61
0.62
0.46
0.50
0.54
0.45
0.49
0.53
75**
80**
85**
                                                           Average 80  -  5%
 Concentration corrected for background levels and
 vehicle baseline emissions
**
  Particulate removed from exhaust stream with
  non-heated 0.5 y Fluoropore filter

-------
 exhaust stream.   Higher recoveries  of 88-113 percent (average  99 percent)
 were obtained when a heated glass fiber filter (375°F) was  used to remove
 particulate  from the sampled exhaust stream.  (Table 12).
             TABLE 12.  TOTAL CYANIDE RECOVERY FROM DILUTE EXHAUST
                              WITH HEATED FILTER
 Actual ppm
  Injected   Run
          Total Cyanide Cone,  as HCN,  ppm

          Calculated
485
485
485
485
485
485
485
485
1
1
1
1
1
1
1
1
Sample

  1
  2
  3
  1
  2
  3
  1
  2
Amount
Observed  Corrected*
0.54
0.54
0.54
0.55
0.55
0.55
0.57
0.57
0.61
0.54
0.64
0.54
0.54
0.61
0.53
0.58
Percent
Recovery
                                                      0.58        107
                                                      0.51         94
                                                      0.61        113
                                                      0.51         93
                                                      0.51         93
                                                      0.58        105
                                                      0.50         88
                                                      0.55         96

                                                          Average  99 ± 9%
 A
  Concentration corrected for background levels and for
   vehicle baseline emissions
 RESULTS  AND CONCLUSIONS

      The measurement of hydrogen cyanide and cyanogen in dilute exhaust can
 be  conducted with a gas chromatography technique.  Cyanide ion is trapped in
 a potassium hydroxide solution and reacted with chloramine-T  to produce
 cyanogen chloride.  Injection of the cyanogen chloride determines the con-
 centration  of cyanide ion in the sample.  This procedure has a minimum
 detection limit of 0.01 ppm cyanide ion.

      The effect of interfering ions in the absorbing reagent was investigated.
 The ions investigated included sulfate, phosphate, permanganate, nitrate,
 carbonate,  chloride, bromide, cyanate, thiocyanate, and ammonium ions.
 Sulfate, phosphate, nitrate, carbonate and ammonium ions exhibited no effect
 on the cyanide ion concentration while chloride, bromide, and permanganate
 ions  interfered only at high concentrations.  High concentrations of chloride,
bromide, and permanganate ions are not expected in dilute exhaust and the
cyanide ion concentrations should not be affected by these ions.  The
presence of cyanate and thiocyanate ions affect the cyanide ion concentration,
and therefore, the definition of total cyanide must take into account the
possible existence and interference of these ions.
                                      44

-------
     As a result of preliminary testing with real exhaust, it was discovered
that two bubblers were necessary to efficiently trap the cyanide ion that
was present in exhaust.  Two factors which might necessitate the use of two
bubblers instead of one are presented below.  First, cyanogen has a much
lower trapping efficiency than hydrogen cyanide in potassium hydroxide. This
difference in trapping efficiency was discovered while naming high concentra-
tion cylinders which were to be used in the qualification experiments.
Secondly, the temperature of the exhaust stream is somewhat higher than the
temperature of the gases used in the determination of the sampling parameters.
The same breakthrough can be expected as with ambient conditions because the
sample gas is not cooled effectively by only one impinger in the ice bath.
Two impingers are therefore necessary for complete sample recovery.  The
final sampling parameters are listed below:

     1.  25 m£ of 1.0 N potassium hydroxide absorbing reagent.

     2.  Absorbing reagent held at ice bath temperature  (0°C-5°C).

     3.  Sample flow rate of 4.0 £/min.

     4.  Impingers rather than fritted glass bubblers.

     5.  Two impingers in series.

These parameters were sufficient to collect a sample from dilute exhaust
within the detection limits of the procedure.

     The measurement of hydrogen cyanide in the presence of cyanogen is dif-
ficult if wet chemical techniques are used.   In clear or dark Tedlar bags,
hydrogen cyanide is stable for at least 60 minutes,  if the humidity within
the bag is not too high.   High humidity increases the possibility of hydrogen
cyanide condensation on the walls of the bag.   Cyanogen, on the other hand,
cannot be quantitively stored in the presence  of exhaust.  Therefore,  bag
samples for the measurement of cyanogen is only a qualitative tool which can
determine if cyanogen is actually produced in exhaust.

     Injection repeatability, sample stability, and sample storage are three
basic requirements for most analytical methods.  The injection repeatability
is well within the expected nominal 5 percent limit for a gaseous syringe
injection.   The concentration of cyanogen chloride within the head space is
dependent upon the volume of the head space, the room temperature,  and con-
centration of cyanide ion present.  A 5 m£ reaction vial with a septum cap is
used in the analysis.  A total of 4 m£ of the  various solutions is added to
this vial.   When the vial is tightly capped, a 1 m£ head space remains above
the solution.  This head space remains constant unless the vial is not tightly
capped or the wrong volumes of reagents are pipetted into the vial.  Cyanogen
chloride obeys Henry's law in the head space.   Henry's laws states that the
mass of a slightly soluble gas that dissolves in a definite mass of a liquid
at a given temperature is very nearly proportional to the partial pressure of
that gas.  Henry's law holds for gases which do not chemically unite with the
solvent and is obeyed by a variety of gases in dilute solutions and all gaseous
solutions at the limit of extreme dilution.   The sample stability is maintained

                                      45

-------
for only a short time after complete development.  The sample may be stored
undeveloped in the potassium hydroxide absorbing reagent for at least three
weeks.

     When a heated filter is used to remove particulate from the sampled ex-
haust stream, 99 percent of the hydrogen cyanide injected into the dilution
tunnel can be recovered.  When a non-heated filter or no filter is used, only
76-80 percent of the cyanide can be recovered.  From these experiments, it is
recommended that a heated filter be used in the sampling system to increase
recoveries and to prevent contamination of the sampling system.

     This procedure provides a rapid and sensitive method for the analysis of
total cyanide in dilute exhaust.  The analysis of a single sample requires two
minutes for reagent addition, five minutes for sample development, and five
minutes for the total peak elution time.  Total sample processing time is
twelve minutes per sample.  The simplicity and ease of analysis makes this
procedure ideal for repetitive analysis.
                                     46

-------
                                 SECTION 4

                      INDIVIDUAL HYDROCARBON PROCEDURE
LITERATURE SEARCH

      The eight individual hydrocarbons (methane, ethane, ehtylene, acetylene,
propane, propylene, benzene, and toluene)  have been measured by innumerable
techniques.  One of the most efficient techniques for the individual determi-
nation of all of these compounds in a single analysis is with gas chromato-
graphy.  Because of its efficiency, this means of analysis was selected over
any of the other available techniques.

      Hydrocarbons are of interest as exhaust components because of their po-
tential for photochemical smog formation.   Hydrocarbons are placed into four
classes according to their participation in atmospheric reactions.  Methane,
ethane , acetylene , propane , and benzene are placed in Class I ,  the  non-
reactive category.  The Class II reactive category includes the C^, and higher
paraffins, while the Class III reactive category encompasses all of the aro-
matics except benzene.  The olefins are placed in the Class IV reactive cate-
gory.  When olefins such as ethylene react with ozone

                   0  + H C =CH2 •*• H C = 0 + HO + HCO

the precursors of photochemical smog are formed.  These free radicals then
participate in other atmospheric reactions that result in oxidant formation.

      Dimitriades and Seizinger (8) proposed a three-column system capable of
analyzing at least 22 hydrocarbons .  Two packed columns were required to re-
solve the GI and C2 hydrocarbon components and an open tubular column was
used to resolve the other components.  The complete analysis consisted of two
different sample loop sizes.  This procedure was considered time consuming
and the number of compounds to be analyzed was excessive.
      Papa et al (9)  presented a procedure for the analysis of Cj_ through
hydrocarbons in automotive exhaust.  This dual column system consisted of a
packed column with a mixture of stationary phases for the resolution of C-j_
and C2 hydrocarbons and an open tubular column.  About 200 individual peaks
were obtained from the method.  This procedure also required two sample loop
injections.  Excessively low temperatures were required for resolution of C^
and C2 hydrocarbons with this analytical system.

      Klosterman and Sigsby (10) proposed a simple analytical system for the
determination of hydrocarbons according to their potential for photochemical


                                     47

-------
smog formation.  A flame ionization analyzer was used in their work, though
this technique did not employ the use of a gas chromatograph.  A column
similar to that used by Klosterman and Sigsby was used to scrub the oxygenated
hydrocarbons and olefins from benzene and toluene by Black et al (11) .  This
method utilizes four packed analytical columns for the resolution of the de-
sired compounds.  Methane, ethane, ethylene, acetylene, propane, and propylene
are resolved with the first two columns; and benzene and toluene are resolved
with the other two.  As with the other procedures, two sample loops are re-
quired for the combined analysis of paraffins, olefins, and aromatic hydro-
carbons.  This procedure was also designed as a simple and inexpensive method
for the determination of smog related compounds.  Table 13 lists the compounds
of interest, along with chemical formulas, boiling and melting points, syn-
onyms, and molecular weights.
 PROCEDURAL  DEVELOPMENT
      The gas  chromatogarph procedure that will be used for the determination
 of  the individual hydrocarbons is similar to the procedure used by Black et
 al  (11)  and consists of a  four column system that is capable of resolving
 eight individual hydrocarbons.  Columns I and II in the system  consist of
 an  8' x  18"  stainless steel tube packed with 80/100 mesh Porapak Q and a 4'
 x 1/8" Teflon  column packed with 35/60 mesh type 58 silica gel, respectively.
 Columnlii consists of 15'  x 1/8" stainless steel tube packed with 15 percent,
 1,  2, 3-tris(2-cyanoethoxy) propane on 60/80 mesh Chromosorb PAW; and Column
 IV  consists  of a 2' x 1/8" stainless steel tube packed with 40 percent mer-
 cury  sulfate  (HgSO4) and 20 percent sulfuric acid (H2SO4) on Chromosorb W.
 Columns  II,  III, and IV are used isothermally and Column I undergoes a tem-
perature program sequence.  The primary purpose of Column I is to resolve
methane  from air, while Column II resolves C2 and C3 hydrocarbons.  Columns
 III and  IV resolve benzene and toluene from the other aromatics, paraffins,
olefins, acetylenes, and oxygenated hydrocarbons.  Three timers, four sole-
noid valves, and five six-port gas sampling valves are required to accomplish
the complicated sample flow through the columns.  When exhaust from diesel
powered vehicles is analyzed, higher molecular weight hydrocarbons have been
found to interfere with the analysis.  The compounds can be effectively re-
moved by simply passing the exhaust sample through an ice trap before it
enters the analytical system.  The actual analytical procedure is included
as an attachment to this report.

VALIDATION

      This gas chromatographic procedure has been used with much success on
a variety of projects.   The validation of this procedure consists of the
injection repeatability for all eight components of the calibration blend
and bag sample stability.  All other parameters were determined from previous
experience with this analytical procedure.

      The injection repeatability for the individual hydrocarbon procedure
was  conducted on two separate occasions.  Table 14 shows the data accumulated
on each occasion.   The injection repeatability for the two 10 mil sample loops
is not greater than ±2 percent.
                                      48

-------
                             TABLE 13.  IMPORTANT FACTS ON INDIVIDUAL HYDROCARBONS
vo
Compound
Methane
Ethylene
Ethane
Acetylene
Propane
Propylene
Benzene
Toluene
Formula
CH4
C2H4
C2H6
C2H2
C3H8
C3H6
C6H6
"7 ft
Molecular
Weight
16.04
28.05
30.07
26.04
44.11
42.08
78.12
92.15
Melting
Point
-182.48
-169.15
-183.3
- 80.3
-189.69
-185.25
5.5
- 95
Boiling
Point
-164
-103.71
- 88.63
- 75
- 42. Q7
- 47.4
80.1
110.6
Synonyms
marsh gas, methyl hydride
ethene, elayl, olefiant gas
bimethyl, dimethyl, methylmethane ,
ethyl hydride
ethyne , ethine
dimethylmethane , propyl hydride
propene, methylethylene, methyl-
ethene
benzol, phene, cyclohexatriene
methylbenzene , phenylmethane ,
                                                                    toluol, methacide

-------
TABLE 14.   INJECTION  REPEATABILITY ON  TWO SEPARATE OCCASIONS
                      Indivi dual 1'uak Ai'tJJ, Keldti ve Counts (Teat 1)
Injection
1
2
3
4
5
6
7
8
9
10
Average
Standard
Deviation ,
sx
Coefficient
of variation,
cv
Methane
7283
7527
7792
7566
7585
7387
7700
7455
7680
7618
7559.30
152.98
2.02
Ethylene
10097
10467
10621
10479
10495
10300
10575
10307
10493
10481
10431.50
154.55
1.48
Ktluiuu A,:uLy luhu
•J.VJ/ 12034
•)d7, l.:'.,-n
lOu'j'J i*!HOJ
<)ti'*2 I.:.>JB
y(joo 12u7!i
9b9o 1239B
9994 12753
9724 12460
9078 I2u24
98B5 U714
9847.20 12563.30
13c.9U J24.i)5
1 . i'i 1 .79
Individual Iv-ak AI..J, KL.-IU! ivd
Injection
1
2
3
4
5
6
Average
Standard
Deviation,
Coefficient
of variation,
C
Methane
8593
8633
8602
8640
8702
8561
8622
48.59
0.56
Ethylene
11893
11937
11993
11885
12022
11900
11938
57.20
0.48
KLl^,,,, A^tyl^ut,
luO/i 14U03
11011 14992
IKK,/ 151J9
10-J/7 151 10
1111 I 15J71
109JI. 15252
11015 15121
Jr, i. 7(1 200.07
,). ,.-. 1 . >_•
Propane t'ropylune Bunzi-jit
14H09 I!)5)h9 Uny/
l'KJ')2 I'j724 1 101 'j
l'iV.,2 I£)lb2 lib//
I.1H25 Ib'JKl 1J-J7/
1'jOlfc Ibhyy 1 1H.:4
14925 15441 14041
15147 15931 14042
14702 15629 14uhb
14933 15842 14049
14996 15S57 13943
14979.70 15795.40 13933.10
187.42 206.83 125. 3i
1.25 1.31 1..-.0
Count.a (Tust 2)
l-..,L.anu 1'iopyl «,...• 1,,,.:-.,,,,
15705 178b7 I046b
15912 18020 l.>t,l)4
15722 17818 lf-,553
15775 17920 1.^409
15825 17850 li,2n/
15879 17955 164J3
15803 17906 Ifc455
83.77 75.10 llil.jl
0.53 ».4:! ,,..'.
't'oiuent; '
1534H
IV, 7u
15406

1047b
155/u
15 /BO
15799
15459
15783
15584.33
166.20
1 .07
Tuliumu
IH546
18537
18735
1B479
18530
18562
18565
07.95
.1.47
                                  50

-------
      The bag sample stability experiment was conducted on a random sample
from an emissions test.  The sample was collected during the driving cycle
and analyzed immediately afterward.  This sample was then reprocessed peri-
odically for several days.  A bag sample of the calibration standard and a
bag sample of exhaust doped with the calibration standard were also processed
periodically.  The time-sample decay curve for each compound is shown in
Figures 18, 19 and 20.  The sample integrity can be preserved for approxi-
mately five days.

QUALIFICATION EXPERIMENTS

      The analysis of individual hydrocarbons in dilute exhaust has previously
been conducted on many projects.  On the request of the Project Officer, no
qualification experiments were conducted with the CVS for this procedure.
Also, long term experience with this procedure has given an insight into the
sample integrity for the complete analytical system.

RESULTS AND CONCLUSIONS

      The measurement of individual hydrocarbons in dilute exhaust is con-
ducted with a gas chromatography technique.  Tedlar bags are filled with
dilute exhaust during each driving cycle.   Analysis of the bag sample
requires a complicated system of four analytical columns with backflush and
temperature program capability.  Sample concentrations are determined by com-
parison to a calibration blend of all eight hydrocarbons.  The minimum detec-
table limit is 0.1 ppmC to 0.2 ppmC.  The higher molecular weight compounds
approach the higher minimum detectable limit.

      Injection repeatability and bag sample stability were demonstrated for
the system.  The largest injection variability was with methane and acetylene
and the smallest was with benzene and toluene.  A 2 percent variability can
be expected for the six compontents of the first sample loop and first two
columns  (C^ - C3), and  a 1 percent variability can be expected for the
second sample loop and second  two  columns  (benzene  and toluene).  This
agreement is much better than can be expected of a syringe sample injection.
The bag sample stability shows that dilute exhaust samples will be stable for
about five days.  Only propylene, benzene and toluene were shown to have a
large decrease in concentration over a period of time in exhaust.  With the
standard only sample, all compounds showed the same decrease in concentration.
A leak in the bag is suspected as a cause of this drastic change in concen-
tration.  However, even with a leak in the bag, the sample concentration is
stable for about five days.  Samples must be analyzed before this time to
maintain confidence in the sample concentrations obtained.  Otherwise, the
sample integrity is lost due to sample decay, bag leakage, and/or permeation
through the walls of the bag.

      This procedure provides an effective means for the analysis of indivi-
dual hydrocarbons in dilute exhaust.  A single bag requires about four minutes
to purge the sample loops, 23 minutes for total peak elution, and five minutes
to cool the oven temperature back to room temperature and reset the instru-
mentation.  The total analysis time per sample is about 32 minutes.  The auto-


                                      51

-------
 -1-30
                            Legend


                          • Methane
 +20

                            Ethylene

                            Propylene


                                                  fJotoinal Range oS

 +10 h                                          Injection Repeatability
6

3
•§  o
 -10
 -20
 -30
                                    6    7     8     9     10    11   12    13


                                 Time,  Days
        Figure 18.   Time-sample  decay  curve  (exhaust only).




                                     52

-------
.c
u
4J
c
I-
Nominal Range of
Injection Repeatability
	 ;. 	 	 	

* * 1 t
Legend
0 Methane
• Ethylene
A Ethane

• Acetylene
    -50
   -100
        I
                            10
                                      15         20



                                      Time, Days
                                                          25
                                                                    30
                                                                              35
                                                              Legend
50


„, n
1
<3
£
U
4-1
C
8
a -so
a



•
Nominal Range of
Injection Repeatability *
• 	 t

8 1 1 •
v 9 • •


-



0



i

Propane
Propylene
Benzene
Toluene




                                                                            I
   -100
                                                I
                            10
15         20


 Time, Days
                                                          25
                                                                    30
                                                                              35
           Figure  19.   Time-sample decay curve (standard only)
                                       53

-------
 + 30 f-
 +20
 +10
i
e,
 -10
 -20
 -30
  Nominal Range of

Injection Repeatability
                                                 4-
                                                    •


                                                    8
                                                    A
                           _L
                            5    6    7     8     9    10   11    12    13


                             Time, Days
     Figure  20.   Time-sample decay curve  (exhaust +  standard)





                                 54

-------
 +30 -
 .,.20
  +10
                                    Legend

                                 0  Propane

                                 A  Propylene

                                 •  Benzene

                                 •  Toluene
  Noisinal Range for
Injection Repeatability
                                    T
§-10
  -20
  -30
                                     I
                               5    6    7

                                 Time ,  days
                                                        10
                                                              11
                                                                    12
                                                                         13
Figure  20 (Cont'd) .   Time-sample decay  curve  (exhaust + standard)
                                      55

-------
mated system provides a simplified operation for an otherwise complicated
procedure and enables routine analysis for a large quantity of samples.
                                     56

-------
                                SECTION 5

                         ORGANIC AMINE PROCEDURE
LITERATURE SEARCH

    The individual amines that are included in this analysis are monomethyl-
amine, dimethylamine, monoethylamine, trimethylamine, diethylamine and tri-
ethylamine.  The chemical formulas, molecular weights, boiling points,
freezing points, and synonyms for these low molecular weight aliphatic
amines are presented in Table 15.  In general, these amines have a fish-type
odor at lower concentrations, but more of an ammoniacal odor at higher
levels.  The 1968 American Conference of Governmental Industrial Hygientists
has recommended a threshold limit value of 10 ppm.

    The measurement of individual low molecular weight amines has been con-
ducted using a variety of gas chromatograph techniques.  Hoshike (13,14)  re-
ported gas chromatographie separation of lower aliphatic amines in the free
form and as their Schiff base derivatives.  A glass column was employed to
provide a separation of 11 amines using temperature programming and a ther-
mal conductivity detector.  This work was directed toward achieving a satis-
factory separation rather than being concerned with minimum detection limits.
Sze (15), et al reported separation of methyl amines, ammonia, and methanol
using a mixture of tetrahydroxyethylethlenediamine and tetraethylenepenta-
mine.  O'Donnel and Mann (16) used Dowfax 9N9, Carbowax 400, and Carbowax
20M to separate mixtures of aliphatic amines, aromatic amines, and aliphatic
amines.  This work was performed using synthetic blends on a gas chromato-
graph with a thermal conductivity detector.  McCurdy and Meiser (17) used a
gas chromatograph with a flame ionization detector to determine fatty amines
in trace quantities.  The fatty amines were converted to trifluoracetyl
derivatives, providing a sensitivity of 0.05 ppm fatty amine in water.

    Smith and Waddington (18) used aromatic polymer beads to seperate a
wide range of aliphatic amines.  Peak tailing was found to exist because of
two types of active sites on the polymer:  simple acidic sites which can be
neutralized by treatment with base, and metal ions which must be deactivated
by addition of an involatile complexing agent.  Glass columns were used in a
gas chromatograph with a flame ionization detector.  Synthetic blends rang-
ing from GJ-CS were separated and analyzed using this approach.  In another
study Carbopak B/4 percent Carbowax 20 M/0.8 percent KOH (19) and 28 percent
Pennwalt 223/4 percent KOH (20) have been reported to give satisfactory sep-
arations of lower aliphatic amines.

    Analysis of amines as derivatives has been shown to be a valuable ana-
lytical tool to determine trace quantities (21).  Thirteen different deri-

                                    57

-------
                                  TABLE 15.  LIST OF INDIVIDUAL ORGANIC AMINES

                              INCLUDED IN THE EMISSIONS CHARACTERIZATION INVENTORY
en
oo
Name
Monomethylamine
Monoe thy lamine
Dimethylamine
Trime thy lamine
Diethy lamine
Triethylamine

(12)
(12)
(12)
(5)
(5)
(5)
Carbon
No.
1
2
2
3
4
6
Chemical Molecular
Formula Weight
CH NH 31.058
C2H NH 45.085
(CH3) NH 45.085
(CH3)3N 59.112
(C2H_)2NH 73.14
(C0H_)W 101.19
Boiling
Point, °C
-6.32
16.58
6.88
2.87
56.3
89.3
Freezing
Point, °C
-93.5
-81.0
-92.19
-117.08
-50.
-114.7
Synomyms
Methylamine ,
aminome thane
Ethy lamine,
aminoethane
None
None
None
None

-------
vatives were evaluated in terms of FID and ECD response characteristics.
"This work was limited to primary amines, and under optimum conditions amines
down to 10 picograms could easily be quantified using an ECD detector.
Clark and Wilk (22) used an ECD to evaulate the properties of halogenated
amine derivatives.  No increase in the sensitivity for the trifluoroacetyl
amine derivatives using ECD was observed.

     Hosier (23) , et al quantitatively measured aliphatic amines volatilized
from cattle feedyards.  Direct gas chromatograph injection of acid solutions
and GC separation of the pentafluorobenzoyl derivatives of the malodorous
volatiles were used in identification.  The derivatized amines were analyzed
using a gas chromatograph equipped with an electron capture detector.
     Methylamine and ethylamine were detected in irradiated beef by Burks
 (24), et al.  Several techniques, including colorimetric paper chromato-
 graphy and gas chromatography, were used in quantiflying results.  Gas chro-
 matographic determination of free mono-, di-, and trimethylamines in biolog-
 ical fluids were performed by Dunn  (25), et al.  A flame ionization detector
 was used to quantitatively separate the lower aliphatic amines.  Separation
 of mono-, di-, and trimethylamine from extracts of fish tissue was achieved
 by Gruger (26).

     Andrea  (27^, et al developed a precolumn inlet system for the gas chro-
 matographic analysis of trace quantities of short-chain aliphatic amines.
 Losses inherent in the collection and direct gas chromatograph analysis of
 field air samples containing volatile amines necessitated an indirect ana-
 lytical scheme.  A Teflon tube  (3" x 5/16" OD) was filled with 20/30 mesh
 Ascarite and placed in the injector inlet of the gas chromatograph.  Samples
 were collected in dilute sulfuric acid and aliquots were injected into the
 pre-column of the GC.  Release of the free amines was found to be sufficient-
 ly reproducible for quantification of results.  This technique avoided the
 problems encountered by Umbreit  (28), et al, and Hardy (29) when using base
 loaded columns to analyze acidified aqueous solutions of amines from fish.
 The in situ release of the free amines from their salts produced a chroma-
 tographic column that changed with every injection.  In addition, the column
 had a very short usable lifetime and lacked reproducibility after extended
 use.

     Bowen (30) described a gas chromatograph procedure for the analysis of
 aromatic amines using an adsorption technique.  Quantitative adsorption and
 desorption of aromatic amines using Tenax GC was demonstrated at the nanogram
 level.  Samples were pulled through the Tenax GC trap for specific sampling
 periods, thermally desorbed at 250°C, and analyzed in a GC with a FID.  The
 author recommended use of a NPD to increase sensitivity for aromatic amines
 on the tail of hydrocarbon solvents and eliminate venting the solvent.

 PROCEDURAL DEVELOPMENT

     From the results of the literature search it was determined that the


                                     59

-------
analysis of the amines should be conducted by the use of gas chromatography.
A Perkin-Elmer 3920B gas chromatograph was dedicated for this purpose.  This
instrument has a dual/differential electrometer and has linear temperature
programming capabilities along with a sub-ambient oven accessory.  The in-
strument has been equipped with a flame ionization detector (FID), a nitrogen
phosphorus detector (NPD) , and an electron capture detector (ECD) and can be
connected to a chemiluminescent detector.  Of the specialty detectors avail-
able for the analysis of the amines, the NPD appeared to be the prime candi-
date and initial work was carried out using this detector.  Because the amines
are notorious for tailing and reacting with metal sites, a glass lined heated
injector port and a glass interface were installed so that with the use of
a glass column, the system would be glass throughout.

     Lecture bottles of methylamine, dime thy lamine, trimethylamine, and
ethylamine, along with pure liquids of diethylamine and triethylamine, and
a Tracor Model 412 Permeation Calibration System containing permeation tubes
of all six amines were used as sources of the organic amines in the proce-
dural development experiments.  These sources allowed a method for preparing
blends of varing concentrations of the organic amines.

     Several  column packing materials (all were developed especially for the
analysis of amines) and column lengths were evaluated to determine which
could provide the best peak separation with the shortest analysis time.  The
columns evaluated included:  12' x 1/4" glass columns packed with 28 percent
Pennwalt 223  amine packing, a 6' x 4 mm (id) , 6" x 2 mm (id) and a 12' x 1/4"
glass column packed with 4 percent Carbowax 20 M and 0.8 percent KOH on
Carbopak B, and a 6' x 4 mm (id) glass column packed with 2 percent KOH on
Chromasorb 103.  Several column temperatures, programming rates, and carrier
flow rates  (helium) were tried for each of the columns.

     The best separation was accomplished using the 4 percent Carbowax 20 M
and 0.8 percent KOH on Carbopak B packing material in the 6' x 4 mm (id)
glass column.  This column would only partially separate diemthylamine and
ethylamine under conditions which gave very broad peaks at long retention
time.  In order to increase sensitivity and shorten the analysis time an
initial temperature of 130°C was chosen.  At this temperature, the dimethyl-
amine and ethylamine coalesce into a single sharp peak, and the analysis
time is under 30 minutes.

     The FID, NPD, and chemiluminescent detector were evaluated as detectors
for the organic amines.  The NPD was more sensitive than either the FID or
the chemiluminescent detector.  Also the NPD is only sensitive to compounds
containing both carbon and nitrogen, eliminating many interferences that
would be present using the FID.

     In order to analyze automobile exhaust for the organic amines, it was
found necessary to concentrate the amines in a trap or an absorbing reagent
to obtain enough sample for the satisfactory analysis.  Two collection pro-
cedures were evaluated, one in which the amines are collected in a trap
filled with 1 gram Tenax-GC packing material, and another in which the amines
are collected by bubbling the amines through an acid solution.  There was
some breakthrough of the amines through the Tenax-GC traps even at liquid

                                    60

-------
nitrogen temperatures.  Bubbling the amines through an acid solution proved
to be the superior method for collecting the amines.

     A sulfuric acid solution was found to effectively trap the amines at
room temperature.  To release the amines from the sulfuric acid solution into
the GC column, an Ascarite loaded precolumn was installed into the injector
block of the GC.  This precolumn was found to work very well in releasing
the amines into the GC column; however, the lifetime of the precolumn was
found to vary from one injection to several hundred.  The 4 percent Carbowax
20 M and 0.8 percent KOH on Carbowax B packing material in the GC column was
designed to be used with aqueous solutions and proved to be satisfactory
when used with aqueous sulfuric acid solutions.

     The usefulness of the precolumn was usually terminated by the aqueous
injections temporarily dissolving the Ascarite and the Ascarite redrying to
form a plug, thereby preventing the sample from entering the column.  Some
time was spent on trying to determine why some precolumns lasted for only
one injection while others for several hundred, but the results were incon-
clusive .

     Poor injection repeatability resulted from a variety of problems.  These
problems included previous GC injection history, glass syringe purging tech-
nique, precolumn conditioning, and column effects.  Because of the problems
mentioned above, and alternate method of analysis was evaluated.  This
method consisted of collecting the organic amines in glass impingers using
dilute sulfuric acid, converting the trapped amines to their pentafluoro-
benzoyl chloride derivatives, and analyzing for these derivatives using a
gas chromatograph equipped with an electron capture detector.  It was hoped
that this method would  (1) convert the amines to stable derivatives which
would improve sample injection repeatability, and  (2) provide improved de-
tection limits with an electron capture detector.

     This alternate procedure was found to be unsuitable for the detection
of amines at the ppb levels.  Tertiary amines  (trimethyl- and triethylamine)
cannot be detected by this procedure, and the secondary amines (dimethyl-
and diethylamine) had a low sensitivity that made detection in the ppb range
almost impossible.  The peak areas of the primary amines (methyl- and ethyl-
amine were found to be time dependent.  The GC peak areas of the methyl- and
ethylamine derivatives were plotted against the time they were allowed to
stand after initial mixing of the reagents to produce the derivatives.  The
mixing procedure is included in Table 16.  The standing time includes 2
minutes of vigorous shaking  (1 minute for the 1 minute test) plus any re-
maining time the mixture was allowed to stand at room temperature before
injecting into the GC.  The effect of elapsed time on peak area is shown in
Figure 21.  This figure shows a rapid increase in peak area followed by a
rapid decrease in the area.  In order to obtain reproducible data at ppb
levels, the injection time after mixing can vary by only seconds.  Under
normal operating conditions this  would not be possible.  Because of the
time limitation, the procedure was abandoned for the quantitative analysis
of the organic amines.
                                      61

-------
               TABLE 16.   MIXING PROCEDURE FOR PREPARATION OF
                     PENTAFLUORABENZOYLAMINE DERIVATIVES*
1.   Pipette 1.0 m£ of 0.01 N sulfuric acid containing 94 ppb methylamine
     and 220 ppb ethylamine into a 10 m£ reacti-vial.

2.   Pipette 3 m£ of toluene into reacti-vial.

3.   Pipette 1 m£ of pentafluorobenzoyl chloride  (PFBC)  solution (50 yg PFBC
     in 100 m& toluene)  into reacti-vial.

4.   Pipette 1 m£ of 10 percent aqueous potassium hydroxide solution into
     reacti-vial.

5.   Shake and allow to stand for X minutes.
* General procedure from private communication with Arvin R.  Hosier, USDA,
  ARS, P.O. Box E,  Fort Collins, Colorado  80521
                                    62

-------
                                                   Methylamine Derivative
                                                   Ethylamine Derivative
                                                                  1
     0123456789     10
        Time elapsed (in minutes) after mixing and before  injection
11
Figure 21.  GC peak areas of pentafluorobenzoyl amine derivatives vs time.
                                     63

-------
     The GC-NPD procedure using the Ascarite precolumn has been reevaluated
and most of the problems involved with its use have been solved.  The in-
consistent lifetime of the precolumn remains a problem in the procedure.
The syringes can be cleaned by purging several times with the next sample to
analyzed.  Memory effects in the precolumn and GC column are not a problem
as long as a blank is injected into the system after injection of a sample
with an amine concentration greater than one ppm.  This blank purges out
the system for the next sample.  The repeatability of the system is also
improved if a series of 4-5 injections of a solution containing 1 ppm of
each of the amines is made into the system.  This must be done each time the
instrument has been unused for periods of greater than one hour.

     The procedure chosen for the analysis of the organic amines consists of
trapping the amines in dilute sulfuric acid solution and analyzing the
solution using a GC equipped with an Ascarite precolumn, a 6' x 4 mm glass
column packed with 4 percent Carbowax 20 M and 0.8 percent KOH on Carbopax B,
and a nitrogen-phosphorus detector.  A finalized copy of the procedure is
included as an appendix to the interim report.

VALIDATION EXPERIMENTS

     Several experiments were carried out to determine the validity of the
amine procedure for the analysis of the organic amines.  The experiments
included checks for GC injection variability, linearity of detector response,
sample  stability in the absorbing solution, and trapping efficiency of the
0.01 N  sulfuric acid solution.

     The finalized sampling conditions used to collect the organic amines are
listed  below, as is a discussion of their selection.  A single glass impinger
containing 25 m£ of 0.01 N sulfuric acid is used to collect the organic
amines.  This single impinger traps 99+ percent of the organic amines at low
ppm and ppb amine concentrations.  This collection efficiency was determined
by bubbling known amounts of organic amines through a series of impingers
and analyzing each impinger separately.  No advantage was found in using
more than one impinger or higher concentrations of sulfuric acid except when
the concentration of the amines exceeded 5 ppm.  The 0.01 N sulfuric acid
concentrations was selected over higher acid concentrations  (0.1 and 0.1 N)
in order to prevent the neutralization of the ascarite in the precolumn any
sooner  than necessary.  The concentration of the organic amines in exhaust
should never approach the 5 ppm concentration; therefore, the single glass
impinger containing 0.01 N sulfuric acid should be sufficient to trap the
organic amines.  Sulfuric acid was chosen over hydrochloric acid as the
absorbing acid because of its higher boiling point.  Hydrochloric acid is
more volatile and could vaporize into the analytical column during analysis.
During sampling, the impinger is kept in a 0°C ice bath.  The ice bath offers
no significant advantage in collection efficiency over room temperature, but
does provide a stable sampling temperature during the test.  The 0°C tem-
perature also lowers the vapor pressure of the aqueous absorbing solution
and thus prevents loss of any significant amount of water from the absorbing
solution during sampling.  The sample flow rate through the impingers is
maintained at 4 liters a minute.  This flow rate provides the largest
amount of sample flow through the absorbing reagent without loss in absorbing

                                     64

-------
efficiency or the physical loss of any absorbing reagent.

     Samples have been found to be stable in the sulfuric acid absorbing
solution for months.  A two month standard containing 0.1 ppm of mono-, di-,
and triemthylamine showed no significant decrease in concentration when com-
pared to a freshly prepared standard.  The absorbing solution is also stable
over long periods of time with the only worry being contamination from any
amines which might be present in the laboratory environment.

     To determine the GC injection repeatability for the procedure over a
wide range of concentrations, three standards containing 0.01, 0.1, and 1
ppm of mono-, di-, and triethylamine were prepared.  Each standard was
injected into the GC ten consecutive times.  The area of each resulting peak
was averaged over the ten runs and a standard deviation was calculated.  The
results of the injection repeatability experiments are presented in Table 17.
Injections of the 0.01 N sulfuric absorbing solution were also made into the
GC system.  Peaks for monomethylamine and dimethylamine/monoethylamine (the
two compounds give one peak in the procedure and are analyzed together as
C2H7N) were detected in the absorbing solution and gave areas which corre-
sponded to 50 percent of the area for the monomethylamine and diemthylamine
in the 0.01 ppm standard and 20 percent of the monoethylamine in the 0.05
ppm standard.  The procedure is not as sensitive to the ethylamines as it
is to the methylamines; therefore, higher concentrations of the ethylamines
 (10, 1, 0.1 ppm) were used in the repeatability. experiments.  For the meth-
ylamines the injection repeatability improves with increasing concentration
of the methylamines.  The standard deviation for the 1 ppm standard con-
taining mono-, di-, and trimethylamine is 5-6 percent, while the deviation
for the 0.1 ppm standard is slightly higher at 7-8 percent.  The standard
deviation for the 0.01 ppm standard is even larger at 12-21 percent.  Con-
centrations at or below 0.01 ppm of the methylamines are difficult to
determine due to the poor injection repeatability and the interference from
the absorbing solution.  The injection repeatability follows no definite
trend for the ethylamines.  The standard deviations for mono-, di-, and tri-
ethylamine remain relatively constant at the three concentrations studied
 (0.1, 1, and 10):  7-8 percent for monoethylamine, 4-7 percent for dimeth-
ylamine, and 7-10 percent for triethylamine.  Concentrations below 0.05 ppm
of the ethylamines are difficult to determine due to the'broadness of the
diethylamine and triethylamine peaks and to the interference from the
absorbing solution.

     To determine the linearity of the nitrogen-phosphorus detector for each
of the amines at the concentration ranges of interest,  standard solutions
containing 0.01, 0.05, 0.1, 0.5, and 1 ppm of mono-, di-, and trimethylamine,
and 0.05, 0.1, 1, 5, and 10 ppm of mono-, di-, and trimethylamine were pre-
pared.  These were made by weighing out required amounts of each of the
organic  amine-hydrochloric acid salts and dissolving them in the proper
amount of sulfuric acid absorbing solution.  Figures 22-27 show plots of the
GC peak areas versus the concentration for each of the  organic amines on a
                                     65

-------
TABLE 17.  INJECTION REPEATABILITY EXPERIMENTS
Amine
Monomethylamine


Dime thy lamine


Trims thy lamine


Monoe thy lamine


Diethylamine


Triethylamine


Concen-
tration (ppm)
1
0.1
0.01
1
0.1
0.01
1
0.1
0.01
10
1
0.1
10
1
0.1
10
1
0.1
Average
Area
8803
2004
1305
7081
1385
1006
5778
1044
344
10,943
8189
2748
7025
3460
724
10,921
5446
1481
Standard
Deviation
533
134
276
372
96
117
276
87
46
977
626
178
483
134
32
1014
564
110
Percent
Deviation
6.1
6.7
21.1
5.3
6.9
11.6
4.8
8.3
14.2
8.4
7.6
6.5
6.9
3.9
4.4
9.3
10.4
7.4
                       66

-------
 10.Oi-
0.01
  0.001  0.002   0.005   0.01
0.02    0.05    0.1

Concentration (ppm)
0.2
0.5
1.0
        Figure  22.   Linearity of monomethylamine GC response
                      (plot on log-log scale).
                                  67

-------
10.0
 D.02
 0.01
                           I
                                                  I
  0.001  0.002   0.005
0.01   0.02    0.05    0.1

      Concentration (ppm)
0.2
0.5
1.0
         Figure 23.  Linearity of dimethylamine GC response
                       (plot on log-log scale).
                                  68

-------
    10.0
     5.0
     2.0
2  *


M  H  !-°
(IS
0)  g
u
  1  0.5
•H -P
-1-1 a
m
     0.2
     0.1
    0.05
    0.02
0.01	


 0.001  0.002
                               I
                                      I
                     0.005   0.01   0.02     0.05    0.1


                                  Concentration  (ppm)
0.2
0.5   1.0
           Figure 24.   Linearity of trimethylamine GC response

                          (plot on log-log scale).
                                      69

-------
   100.0
    50.0 -
    20.0 -
  •H
  C
    10.0 -

-------
100.0 r-
 50.0
           0.02
0.05
0.1   0.2      0.5    1.0

     Concentration (ppm)
2.0
5.0
                                                                       10.0
        Figure 26.  Linearity of diethylamine  GC  response
                       (plot on log-log  scale).
                                  71

-------
   100.0
    50.0
     20.0
     10.0
0) B
o

-H 4J
-P (0
(0
«
      5.0
2.0
      1.0
      0.5
      0.2
      0.1
          J
                         I
        0.01    0.02     0.05
                                       I
I
                         0.1   0.2      0.5     1.0

                              Concentration  (ppm)
             2.0
5.0    10.0
            Figure 27.   Linearity of triethylamine GC response
                            (plot on log-log scale).
                                       72

-------
log-log scale.  The relative GC areas  for monomethyl-,  dimethyl-  and mono-
ethylamine were corrected for the background peaks  found  in  the absorbing
solution.  Mono-, di-, and trimethylamine give  linear GC  responses  from
0.01 to 1 ppm and di-, and triethylamine give linear responses in the 0.05
to 10 ppm region.  Monoethylamine gives a linear response from 0.05 ppm to
1 ppm, but shows some deviation from linearity  in the 1.0 to 10 ppm range.

     Ammonia at concentrations between 10-100 ppm in sulfuric acid does not
give as large a peak as does a 0.01 ppm solution of monomethylamine, and
therefore does not present any major problems as an interference.  However,
the retention time (0.6 min) is close  to that of monomethylamine  (0.85 min)
and care must be taken not to confuse  one peak  for  the  other.  Acetonitrile
(CH3CN) had been found in exhaust at concentrations near  the 0.1  ppm level.
The NPD is sensitive to this compound  and gives a peak  in the chromatogram
at a retention time of 1.8 minutes.  This retention time  is  near  that of
trimethylamine (2.0 min) and care must be taken not to  confuse the two com-
pounds.  No other compounds in exhaust have been found  to be interferences
in the procedure.

QUALIFICATION EXPERIMENTS

     Qualification experiments were carried out using a Mercedes  240D vehicle.
Hot FTP  (23 minute test) driving cycles were followed to  generate exhaust
for the vehicle baseline emissions and for the  tunnel plus vehicle exhaust
experiments.  Two aluminum cylinders each containing three amines were used
in the experiments.  One cylinder contained 174 ppm monomethylamine, 132
dimethylamine, and 107 ppm trimethylamine.  The second  cylinder contained
408 ppm monomethylamine, 241 ppm dimethylamine, and 123 ppm  trimethylamine.
The cylinders were named by diluting the amine  gas  stream 300 fold with zero
air, collecting the diluted sample in  0.01 N sulfuric acid and analyzing the
sample with GC-NPD.  The baseline emission values from  the test vehicles
were found to be less than 0.005 ppm for all six amines investigated.  A
test sequence was developed to determine the injection  recovery for the
three methylamines from the CVS tunnel (18 inch diameter) without exhaust
present.  Four tests were conducted for the amines  at ppm levels  ranging
from 0.13 to 0.22 ppm.  Each test was  conducted on  a sequence basis with
a 10 minute soak with the CVS off between each  23 minute  collection interval.
The sample lines were heated to 175°F  to prevent amine  losses in  the sample
lines.  The results of these experiments are presented  in Table 18.  As
expected, the recovery of the methylamine was very  low  (0.7  - 10.2 percent)
for the amine" injections.  Dimethylamine recovery increased  from  4.6 to
27.3 percent after four consecutive injections. Recoveries  for trimethyl-
amine were more reasonable with 51.8 to 87.9 percent recovery for four con-
secutive tests.  These results are similar to those obtained in qualification
experiments for gasoline-powered vehicles  (1).

     To determine the percent of amine recovery in  the  presence of  exhaust,
a similar set of experiments was carried out.   The  Mercedes  240D  was used
to generate exhaust during the 23 minute sampling period. All other para-
meters were the same as described above and in  Table  18,  with the exception
of adding a non-heated filter in the sample line.   This filter was  used to


                                     73

-------
               TABLE 18.  ORGANIC AMINE RECOVERY FROM THE CVS DILUTION TUNNEL ONLY
Amine Injected

Monome thylamine
Monome thylamina
Monomethylamine
Monomethylamine

Dimethylamine
Dime thylamine
Dime thylamine
Dime thylamine

Trime thy1amine
Trimethylamine
Trimethylamine
Trime thylamine
ppm Nominal Flow
Amine
Injected
174
174
174
174
132
132
132
132
107
107
107
107
Rate , ft^/min
Amine Blend
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
CVS
320
320
320
320
320
320
320
320
320
320
320
320
Run
1
2
3
4
1
2
3
4
1
2
3
4
Calculated
ppm amine
dilute
0.22
0.22
0.22
0.22
0.17
0.17
0.17
0.17
0.13
0.13
0.13
0.13

Observed
ppm
0.001
0.003
0.008
0.022
0.008
0.016
0.032
0.045
0.069
0.119
0.087
0.118

Percent*
Recovery
0.7
1.4
3.7
10. 2
4.6
9.9
19.1
27.3
51.8
89.1
66.7
87.9
* All values are the average of three independent  samples.

-------
prevent particulate from contaminating the sampling system.  Only trace
amounts of amines were recovered in these experiments.  The tests were re-
peated using a heated filter and higher concentrations of the three methyl-
amines.  The heated filter and the sample line connecting the dilution tunnel
and the heated filter were heated to 375 °F.  The line connecting the heated
filter and the sampling system was maintained at 175°F.

     A new amine cylinder (higher in amine concentrations) and lower dilution
ratios was  used to give expected ppm levels of 0.21 to 0.68 ppm in the di-
lution tunnel.  The results of these experiments are presented in Table 19.
The methylamine recoveries ranged from 12.0 to 27.3 percent, dimethylamine
recoveries ranged from 18.6 to 44.2 percent and trimethylamine recoveries
ranged from 47.6 to 59.5 percent recovery.  If a heated filter is used it is
possible to detect amines in exhaust at 0.2 ppm and higher levels.  At levels
lower than 0.2 ppm, losses to the dilution tunnel and to the exhaust may
prevent the detection of the amines.

     At this time, it is uncertain as to the precise reasons for the losses,
but all possible steps have been made to preserve the integrity of the
sampling system and the sample handling prior to injection into the gas
chromatograph.  It is doubtful that any substantial improvement could be
made to the system without going to heating the tunnel, etc.  The losses of
the low molecular weight amines were not unexpected and these experiments
confirmed those fears.  In summary, methyl- and dimethylamine had low recov-
eries in the tunnel with and without exhaust present.  Trimethylamine recov-
eries were generally higher and improved with the number of consecutive in-
jections.
RESULTS AND CONCLUSIONS

     The concentration of organic amines in dilute exhaust can be determined
by collecting the amines in 0.01 N sulfuric acid and analyzing the solution
with a GC equipped with an ascarite precolumn and a nitrogen phosphorus
detector.  The amines are effectively trapped in 25 m£ of 0.01 N sulfuric
acid absorbing solution at a flow rate of 4 £/min.  For a twenty-three minute
test and a sample flow rate of 4 H/min, the procedure has a minimum detec-
tion limit of 2 ppb for each organic amine.

     The accuracy of the procedure decreases as the concentration of the
amine in the absorbing solution decreases.  At a 0.01 ppm concentration of
the organic amines in the absorbing solution, the present standard deviation
for the GC is 12-21 percent.  The absorbing solution itself gives peaks
equal to 0.005 ppm monomethylamine and dimethylamine/ethylamine.  At the
0.01 ppm level or lower it is difficult or impossible to determine the con-
centration of amines.  This concentration is equivalent to 2 ppb of the
amines in dilute exhaust (23 minute test, sampling at 4 H/min.)

     Acetonitrile and ammonia have been found in exhaust samples and give
peaks in the chromatograms.  Ammonia at concentrations of 10-100 ppm gives
a peak approximately the size of a 0.01 ppm methylamine peak.  The separation
of the ammonia and the methylamine peak is 0.25 minutes, but the two can
easily be distinguished if care is taken.  Acetonitrile has a retention time

                                     75

-------
                  TABLE  19.   ORGANIC AMINE  RECOVERY FROM THE  CVS  DILUTION  TUNNEL WITH EXHAUST
CTi
Ainine Injected

Mon ome t±iy 1 ami ne
Monomethylamine
Monomethylaraine

Dimethylamine
Dimethylamine
Dime thylamine

Trimethylamine
Trimethylamine
Trimethylamine
ppm
Amine
Injected
408
408
408
241
241
241
123
123
123
Nominal Flow
Rate, ft3/min
Amine Blend
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
CVS
300
300
300
300
300
300
300
300
300
Run
1
2
3
1
2
3
1
2
3
Calculated
ppm amine
dilute
0.68
0.68
0.68
0.40
0.40
0.40
0.21
0.21
0.21

Observed
ppm
0.092
0.156
0.186
0.074
0.139
0.177
0.098
0.112
0.122

Percent*
Recovery
12.0
22.9
27.3
18.6
34.7
44.2
47.6
54.9
59.5
       * All values are the average  ,of three  independent samples

-------
that differs from trims thy lamine by only 0.2 minutes, but the two can also
be easily distinguished.

     The amines are notorious for sticking to  metal sites.  The qualifica-
tion experiments represent another.example of this problem.  The amines had
low percent recoveries from the CVS dilution tunnel with and without exhaust
present.  The percent recovery increased directly as the number of injections
into the tunnel increased.  This phenomenon is probably due to the gradual
coating of the tunnel with the amines, thus neutralizing the number of metal
sites in the tunnel.  It is possible that if the amines are present in con-
centrations of less than 0.2 ppm, the percent recovery may be very low or
essentially zero.

     The organic amine procedure should provide a relatively accurate method
for determining the concentration of the organic amines exiting the CVS
tunnel; however, amine losses in the CVS tunnel must be taken into account
when reporting these concentrations.
                                      77

-------
                                 SECTION 6

                         SULFUR DIOXIDE PROCEDURE
LITERATURE SEARCH

     At room temperature and atmospheric pressure sulfur dioxide is a highly
irritating, nonflammable, and colorless gas.  The gas is readily detectable
at concentrations of 3-5 ppm by the human sense of smell.  Physical proper-
ties of sulfur dioxide, SO2 (sulfurous acid anhydride) include a freezing
point of -75.5°C (1 atm) , a boiling point of -10.0°C  (1 atm) , and a molecular
weight of 64.063 (12).

     The bulk of published literature regarding the analysis of sulfur di-
oxide has dealt with ambient air sampling.  With the development of instru-
mental methods of analysis, the ability to measure sulfur dioxide in sta-
tionary and mobile source exhausts now exists.  The following review of
references reveals a wide variety of analytical techniques used in the mea-
surment of sulfur dioxide concentrations.

     A frequently used method for the analysis of sulfur dioxide is a color-
imetric method.  The most commonly employed colorimetric technique is the
West-Gaeke method (31-35).  This method has been collaboratively tested, with
the lowest concentration range studied being well above the levels most fre-
quently found in rural and global background air (36).

     A modified version of the West-Gaeke method involves the collection of
sulfur dioxide in 0.1 M sodium tetrachloromercurate(II) (TCM) .  Sulfur di-
oxide reacts with the TCM to form a dichlorosulfiromercurate complex (DCSM) .
In this modified version, the DCSM resists oxidation by oxygen in the air
and oxygen dissolved in the absorbing solution.  Ethylenediamine tetracetic
acid disodium salt (EDTA) is added to the TCM absorbing solution to complex
any heavy metals that could oxidize sulfur dioxide before the DCSM is formed
(37,38) , and sulfamic acid is added to the absorbing solution to destroy any
interfering nitrite ion which might be present (39) .

     The colorimetric determination of sulfur dioxide is based upon the mea-
surement of the red-violet color produced by the reaction of DCSM with hy-
drochloric acid, pararosaniline and formaldehyde.  The effect of the para-
rosaniline dye purity on the colorimetric procedure has been reported by
several researchers (40,41).  since the dye purity does effect the results
of the colorimetric procedure, various techniques for the purification of
commercial grade pararosaniline have been published  (32,34,42), and para-
rosaniline purified especially for the colorimetric analysis of sulfur di-
oxide is commercially available (42).

                                     78

-------
     A major potential source of error associated with the West-Gaeke colori-
metric method for measuring sulfur dioxide is the widely differing collection
efficiency reported for Greenburg-Smith and midget  impingers at low sulfur
dioxide concentrations (43).  Urone, et al, investigated the collection ef-
ficiency of the TCM solution by the use of microgram quantities of sulfur
dioxide tagged with 35S (44).  In this investigation, it was found that a
series of bubblers cannot be used to determine absorber collection efficiency.
Bostrom obsered a 99 percent collection efficiency  for a concentration range
100-1000 ppb sulfur dioxide in a TCM solution  (45) .

     Work has been conducted in the development of  other colorimetric methods
for the analysis of sulfur dioxide.  Attari developed a procedure whereby
sulfur dioxide is absorbed into a solution of ferric ammonium chloride, per-
chloric acid, and phenanthroline dye  (46).  A color complex with an absorb-
ance of 510 mm was formed, and although the color developed within 10 minutes,
it tended to fade with time.  Hydrogen sulfide was  found to be an interfer-
ence in the procedure.

     Kawai used the reaction of barium chloranilate with sulfate as an in-
direct measurement of sulfur dioxide  (47).  Sulfur  dioxide was absorbed in
a solution containing hydrogen peroxide and barium  chloranilate.  Barium
chloranilate reacts with the sulfate ion  producing  a red-violet chloranilic
acid ion.  Although this method may be satisfactory for flue gas analysis, it
lacks the sensitivity required for ambient air analyses.

     Conductivity methods have been used  for continuously monitoring sulfur
dioxide in air  (48).  The conductivity of a dilute  sulfuric acid-hydrogen
peroxide reagent changes due to the absorption of pollutants.  This change
in conductivity is assumed to result primarily from sulfur dioxide absorbed
from the sampled air and oxidized to sulfuric acid.  In many cases, sulfur
dioxide is the major pollutant present; however, if other pollutants are
present, their collection efficiency and  solubility may be significantly
different than for sulfur dioxide.  Several field comparisons of conduc-
tivity with other sulfur dioxide procedures indicate a fair agreement  (49-54).
Hydrochloric acid gas, ammonia, and chlorine substantially increase conduc-
tivity.  Shikiya and McPhee found two- to fourfold  differences between
different conductivity analyzers and between conductivity and colorimetric
analyzers (51) .  Although the conductivity procedure may be acceptable for
point sources of sulfur dioxide in isolated areas,  its high potential for
positive and negative interferences limits its application.

     lodometric methods were among the first adapted for air pollution
analysis from the industrial hygiene literature.  With this method, the
sulfur dioxide is collected in an impinger containing standard NaOH absorbing
solution.  The absorbing solution is acidified and  the liberated sulfurous
acid is titrated with a standard iodine solution  (52).  Another method em-
ploys a standard iodine-potassium absorbing solution  (53).  lodometric me-
thods of analysis for sulfur dioxide generally suffer from a lack of sensi-
tivity and interferences from hydrogen sulfide.

     Adsorption sampling methods have  also been  developed  for the measure-
ment of sulfur dioxide  (55) .  Sulfur  dioxide  is  absorbed on  silica  gel,

                                     79

-------
desorbed, and reduced to hydrogen sulfide at 700-900°C over a platinum
catalyst.  The hydrogen sulfide is then absorbed in a 2 percent ammonium
molybdate solution and determined colorimetrically.  Although this tech-
nique is relatively specific for sulfur dioxide, the final colorimetric
determination by the molybdenum complex does not utilize the most sensitive
method available.

     In addition to the aforementioned techniques, sulfur dioxide has  been
measured by filtration (56-60) and static collectors (61-67).  Air samples
are passed through potassium bicarbonate impregnated filters and analyzed
for sulfate.  The collection efficiency of these filters is dependent  upon
humidity, temperature, and the atmospheric concentration of sulfur dioxide.
The lead peroxide candle static collector was developed by Wilson and  Mc-
Connell as an inexpensive method for measuring relative "sulfation" of the
atmosphere  (61).  The sulfur dioxide collection efficiency is dependent
upon temperature, relative humidity, wind speed, atmospheric concentration
of sulfur dioxide, and the length of exposure period (62) .  Ikeda determined
ambient sulfur  dioxide levels by collecting samples on active carbon filters,
washing the filters with distilled water, and titrating with barium chlorani-
late (67).

     With the advent of modern instrumental methods of analysis, specifically
gas and ion chromatography, a substantial amount of data has been published.
Most trace gas  analysis for sulfur dioxide has been conducted using gas
chromatographs  with flame photometric detectors (FPD) (68-77).  The FPD  is
highly selective for sulfur compounds and has low minimum detection limits.
Analysis for sulfur dioxide is generally performed using all Teflon or glass
systems.  Sulfur dioxide will react with active sites in the gas chromato-
graph system, making the use of inert materials essential for trace quanti-
tative analysis.  Gas chromatographs with FPD and linearizing circuitry pro-
vide a wide dynamic range for ambient and source sulfur dioxide levels.  In
some instances, the collection technique precludes the use of GC-FPD
techniques; i.e., bag sampling from dilute automotive exhaust or source
sampling.  The  use of gas chromatography would be a prime candidate if the
sample integrity could be assured in the sample acquisition and subsequent
analysis.

     A more recent development in methods of analysis for sulfur dioxide
involves the use of ion chromatography (78).  This technique involves col-
lection in a hydrogen peroxide absorbing reagent and measurement of the
resulting sulfate ion using ion chromatography.  Ion chromatography is a
specialized area of liquid chromatography which will separate and quantify
the individual  cations or anions.  This technique has been applied to  the
measurement of  sulfur dioxide in ambient air.
     Other instruments are commercially available that are reported to mea-
sure sulfur dioxide in ambient or dilute automotive exhaust.  Such instru-
ments include continuous detection by pulsed fluorescent UV and  second
derivative UV analyzers.  A pulsed fluorescent UV analyzer for sulfur di-
oxide was found to give recoveries on the order of 115-125 percent, indi-
cating that a positive interference is.present (79).  The second derivative

                                     80

-------
UV sulfur dioxide analyzer has inherent problems when being used on contin-
uous samples.  The mirrors are located in the actual cell and become coated
with various exhaust components even after  filtration of the sample.  The
mirrors will become etched and need resurfacing if the unit is used in the
presence of sulfur dioxide, sulfate ion, or other corrosive exhaust com-
ponents.  The inherent noise level, along with the consistent mirror problem,
preclude the use of second derivative UV analyzer for measuring sulfur di-
oxide on a continuous basis.

     A variation of the GC method  for measuring sulfur dioxide is the use of
a continuous analyzer using an FPD detector.  Although this approach is good
in theory, it has several problems associated with the performance of the
FPD detector.  These units were originally  designed to monitor sulfur dioxide
levels in the ambient air and adaption to automotive exhaust was not
staightforward.  Air samples had essentially the same oxygen and nitrogen
levels all of the time; however, dilute exhaust samples have variable car-
bon dioxide, oxygen and nitrogen concentrations.  The species have been
found to cause quenching effects on a FPD detector.  With the constantly
changing carbon dioxide, oxygen, and nitrogen it would be impossible to
correct for any quenching effect.  The use  of a continuous FPD analyzer for
measuring sulfur dioxide in automotive exhaust would not be acceptable un-
less the quenching effects could be eliminated.

PROCEDURAL DEVELOPMENT

     From the results of the literature search it was determined that the
analysis of sulfur dioxide should  be conducted by the use of ion chromato-
graphy.  An ion chromatograph built at Southwest Research Institute was
dedicated for this purpose.  This  instrument utilizes a modified Swagelok
reducing union for a conductivity  cell, a Hall conductivity detector, a
Milton Roy mini-pump, a Soltec multivoltage recorder, a Glenco Scientific
pulse dampener, and polyethylene cubitainers from Cole Farmer Instrument
Company for the analysis of sulfur dioxide. A minimal amount of procedural
development work was necessary for this procedure; however, several instru-
ment and sampling parameters did have to be determined.  The selection of
these parameters are discussed in  detail in the Validation Experiments
section.

     In order to analyze automotive exhaust for sulfur dioxide, a trap or
an absorbing reagent must be used  to concentrate the sulfur dioxide.  A
method which has been previously used at Southwest Research Institute for
collecting and concentrating sulfur dioxide was selected and validated for
use in this project.  This method  consists  of bubbling dilute exhaust through
a dilute aqueous hydrogen peroxide solution.  The hydrogen peroxide reacts
with the sulfur dioxide to give sulfate ion which remains in the absorbing
solution.

     The parameters selected for the analysis of sulfur dioxide are listed
below.  The sulfur dioxide from the exhaust is bubbled through two  impingers
(maintained at ice bath temperatures) in series with each impinger  containing
25 mJl of a 3 percent hydrogen peroxide solution.  The exhaust  flows through
the impingers at a rate of 4 £/min.  Two impingers together trap 99

                                    81

-------
percent of the sulfur dioxide present in exhaust.  A heated glass fiber filter
is installed in the sampling line prior to the bubblers to remove particulate
which could contaminate the separator column during analysis.  A portion of
the absorbing solution is loaded into the sample loop and injected into the
ion chromatograph.  For analysis the ion chromatograph utilized three columns
and an eluent composed of 0.003 M NaHCO3 plus 0.0024 M Na2CO3.  The eluent
flows at 30 percent of full pump capacity through a 3 x 150 mm precolumn
(this column helps prevent contamination of the separator column), a 3 x 500
mm separator column and a 6 x 250 mm suppressor column packed with AG 50W-X16
anion suppressor resin (this neutralizes the ionic effect of the eluent while
increasing that of the sample ion) .  A finalized copy of the procedure is
included as an appendix to the interim report.

VALIDATION EXPERIMENTS

     Sulfur dioxide validation experiments were performed to verify the
sampling and instrument parameters.  These experiments involved the deter-
mination of sampling flowrate, sampling temperature, kind and concentration
of absorbant and the number of bubblers required to collect 100 percent of the
sulfur dioxide.  The variables associated with the ion chromatograph that
were determined included type and concentration of eluent, injection loop
size, flowrate, injection variability, and linearity of response.  In
addition to determining sampling and instrument parameters, validation ex-
periments were performed to verify certain portions of the procedure for
sulfur dioxide analysis.  Tests for interferences, sample stability, and
standard stability were among those conducted.  Also, the method of washing
glassware was studied.

     A number of possible interferences were tested by bubbling the suspected
interfering gas at 4 £/min through three impingers in series.  Each impinger
contained  25 m£ of 3 percent hydrogen peroxide and was maintained at ice
bath temperatures  (0-5°C).  The tests lasted approximately twenty minutes
each.  The results are shown in Table 20.  In the zero air, zero nitrogen, 3
percent C02 and 100 ppmc HC tests, no detectable amount (less than 0.01 ppm
SO2) was found.  A positive interference of 0.01 ppm SO2, was found in the
100 ppm NOX test.  The greatest interference, 0.02 ppm SO2, was found in the
100 ppm CO test.  Another source of interference was caused by the sulfuric
 acid-chromic  acid bath in which the impingers were washed.   The  sulfate ion
from the sulfuric acid could not be sufficiently rinsed from the impingers,
even with repeated deionized water rinses.  For this reason, a 1:1 (v:v)
nitric acid and water solution was used to wash the impingers used in the
sulfur dioxide procedure.  The sulfate present in the hydrogen peroxide ab-
sorbing solution also causes a positive interference.  This interference
could be corrected for by subtracting the sulfate peak area of the absorbant
from the sulfate peak area of exhaust or background samples.

     Another validation experiment for the ion chromatographic method of sulfur
dioxide analysis involved determining the stability of samples and standards
over a period of time.   The sulfuric acid standards made up in filtered de-
ionized water remained stable for at least fourteen weeks.  Sulfate standards
made up from stock solutions prepared on 11/30/77 and 3/13/78 were analyzed
on 3/13/78 and the peak areas were compared.  The results, shown in Table 21
indicate that the fourteen week old standards repeated within 10 percent of
                                     82

-------
                 TABLE 20.  INTERFERENCES TO  S0? ANALYSIS
Suspended
Interference
Zero Air
Zero N
3% C02~run 1
3% CO- -run 2
100 ppmc HC
100 ppm NO
100 ppm CO
1
0.00
0.00
0.00
0.00
0.00
0.01
0.01
Bubbler
2
0.00
0.00
0.00
0.00
0.00
0.00
0.01
3
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Total SO? (ppm)
0.00
0.00
0.00
0.00
0.00
0.01
0.02
                   TABLE 21.  SULFATE STANDARD STABILITY
   Standard
Concentration
         -2
/ ny »u4 - ^
V m£ >
0.96
0.96
4.80
4.80
9.60
9.60
38.40
38.40
96.00
96.00
(Standard
Preparation
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
Attenuation
3
3
10
10
30
30
100
100
300
300
Height
(in)
4.73
4.77
7.72
7.62
5.41
5.45
7.93
7.81
6.09
6.08
Amplitude
125,844
125,080
148,259
147,300
130,852
131,036
149,864
148,668
135,666
135,638
%
Difference
0.6
0.6
0.1
0.8
0.0
                                     83

-------
the freshly prepared standards.  A study was also conducted with  a variety of
samples of different ages to determine sulfate longevity in the  3 percent hy-
drogen peroxide absorbant.  One week after collection and  initial analysis,  a
sample obtained from an SO2 exhaust recovery experiment remained  at 0.13
ppm S02.  A two week old 0.06 ppm SO2 baseline sample produced similar re-
sults.  There was no change observed in the sulfur dioxide level.  A ten
week old collection efficiency sample, however, did very from its  initial
concentration of 0.13 ppm by decreasing 7.9 percent to 0.12 ppm SO2.  This
is greater than the injection repeatability of 1.2 percent, however, within
the minimum detection limit of 0.01 ppm S02.  The samples appear  to be
stable at least two weeks but less than ten weeks, the break-off point
probably lying between four and six weeks.

     The second portion of validation testing included the determination of
SO2 sampling;parameters.  Nominal concentrations of 5 and 12 ppm  SO2 were
collected for 20 minutes in three bubblers, each containing, 25 m£  of 3 per-
cent hydrogen peroxide.  The results of these tests are shown in  Table 22.

                 TABLE 22.  SO2 COLLECTION EFFICIENCY AS A
                   FUNCTION OF FLOWRATE AND TEMPERATURE

                                                                     •* SO2
                                                                   trapped in
                                                           Bubbler  bubblers
                                                         1+2+3(ppm)  1 and 2
Flow- SO2 Concentration (ppm) and
rate Temp, (%), in bubbler
Test (£/min) (°F)
1
2
3
.,1
2
3
4
-:4
4
2
2
72
32
32
75
                                                            4.52,

                                                            6.96
                                                            6.85

                                                            6.93
98.7

99.1
97.7

98.6
                            Nominal 5 ppm  SOg
                      4.31(95.3)   0.16(3.4)   0.06(1.3),
                      6.73(96.7)   0.17(2.4)   0.06(0.1)
                      6.48(94.6)   0.21(3.1)   0.16(2.3),
                      6.71(98.8)   0.12(1.8)   0.09(1.4),

                            Nominal 12 ppm  SO2
                                  1.25(5.2)

                                  0.13(0.5)
                                  0.22(0.8)
                                  0.35(1.2)
                                  0.24(0.8)

                                  0.20(0.9),
                                  0.23(0.8)
                                  0.32(1.1)

The largest quantity of sulfur dioxide  was  retained at a flowrate of 4 £/min
at ice bath temperature.  Under these  conditions, 99.1 percent S02 was col-
lected in the first two bubblers.   It  is also desirable to prevent small par-
ti culate debris in the exhaust from entering the samples and, thus,  the  columns
of the ion chromatograph.   If this form of  contamination is allowed to col-
5
6
7
8
9
10
11
12
4
4
4
4
4
4
4
4
32
32
32
32
32
32
32
32
22.4(92.5)
24.8(99.1)
28.5(99.0)
29.3(97.8)
30.0(98.5)
21.2(97.9)
29.9(98.5)
28.9(98.5)
0.54(2.2)
0.09(0.4)
0.07(0.2).
0.30(4.0)
0.21(0.7)
0.25(1.2)
0.23(0.8)
0.12(0.4)
24.2
25.0
28.8
29.9
30.4
21.7
30.4
29.3
9.7.7
99.6
99.8
99.0
99.3
98.8
99.3
99.6
                                     84

-------
lect in the columns the liquid flow becomes hampered  causing increased back-
pressure.  A glass fiber filter in the  sampling  line  is  used to remove a
large portion of this debris from the exhaust.

     The third portion of validation testing involved the determination of
instrument parameters; eluent concentration and  flowrate, columns, sample
loop size, linearity of response and injection variability,  choice of
eluent concentration and flowrate will  depend on the columns chosen and the
species present in the samples.  Exhaust samples contain a variety of amines,
fluoride, chloride, nitrite, phosphate, nitrate  and sulfate.  Nitrate elutes
just prior to sulfate necessitating the use of an efficient  separator column.
A 3 x 500 mm glass column packed with patented resin  is  the separator column
chosen for sulfate analysis.  The suppressor column is a 6 x 250 mm glass
column with AG 50W-X16 resin.  A 0.003  M NaHCO3  and 0.0024 M Na2CO3 eluent
solution flowing at 30 percent of pump  capacity  gives good baseline resolu-
tion when a 500 y£ sample loop is used.  Another instrument factor which
needed to be determined was the injection variability.

     The ion chromatograph has an injection repeatability of 1.1 or 1.2
percent as shown in Table 23.  This is  represented by Cv (coefficient of
variation) which is the standard deviation divided by the mean and multi-
plied by 100.  The mean or average is represented by  x and standard devi-
ation by Sx.  For these calculations peak heights instead of peak areas were
used since the heights repeated much better and  with  greater precision than
the areas.
                                            yg S04~2         yg     ~
Two different standards were analyzed:  0.5 - „ - and  4.0
     The final instrument parameter determined was  the  linearity of response
 of  sulfate standards at different  attenuations.   The  sulfate standards, made
 up  from sulfuric acid and filtered deionized water, maintained linearity at
 each attenuation but the slopes became  steeper as the sensitivity decreased.
 Table 24 shows heights corresponding  to each standard used and Figure 28
 shows the graphical representation of the data.   At the 1 x 10  scale setting,
 the relative slope was 1.7,  at the 1  x  30 setting it  was 2.5 and at the 1 x
 100 setting the slope was 2.6.  It was  not necessary  to carry the curve any
 further, since no samples have been obtained that fall  in the higher con-
 centration range.  However,  it was found  that lineatity was maintained at
 concentrations from 40 to 100U9S4   .
 QUALIFICATION EXPERIMENTS

     Qualification experiments were  carried out to determine  the percentage
 of sulfur dioxide that could be  recovered at the sampling point when known
 amounts of sulfur dioxide were injected into the dilution tunnel at the point
 where exhaust enters the tunnel.  A  Mercedes 240D vehicle was used as a
 source of exhaust.  Hot FTP (23-minute test)  driving cycles were followed to
 generate exhaust for the vehicle baseline emissions and for the tunnel plus
 vehicle experiments.  Aluminum cylinders containing 887 ppm and 9098 ppm
 sulfur dioxide in balance air were used to inject sulfur dioxide into the
 CVS dilution tunnel.  The flow of sulfur dioxide into the tunnel was regu-

                                     85

-------
   TABLE 23.  INJECTION REPEATABILITY FOR ION CHROMATOGRAPH
Sample







  1




  2




  3




  4




  5




  6
  7




  8




  9




 10




 11




 12
Concentration fw $°4 ] Attenuation Height (in)
\ mJl
3.84
3.84
3.84
3.841
3.84
3.84





0.48
0.48
0.48
0-.48
0.48
0.48




/
1 x 10 5.37
1 x 10 5.43
1 x 10 5.51
1 x 10 5.45
1 x 10 5.50
1 x 10 5.53
x 5 . 46 in
S 0.06 in
x
C 1.1%
V
1 x 10 0.81
1 x 10 0.80
1 x 10 0.79
1 x 10 0.81
1 x 10 0.80
1 x 10 0.81
x 0.80 in
S 0.01 in
x
C 1.2%
V
                              86

-------
              TABLE 24.  CALIBRATION CURVE FOR SULFUR DIOXIDE
Standard             _2                                 Heights  Corrected
Concentration (W  S04  j  Attenuation   Height  (in)    to  1  x 10 scale  (in)
\ .m& /
0.48
0.96
1.48
1.92
2.88
3.84
4.80
4.80
7.68
9.60
19.20
28.80
38.40
1
1
1
1
1
1
1
1
1
1
1
1
1
x 10
x 10
x 10
x 10
x 10
x 10
x 10
x 30
x 30
x 30
x 100
x 100
x 100
0.85
1.87
2.38
3.23
4.89
6.62
8.16
2.60
4.77
6.49
3.37
6.15
8.32







7.80
14.31
19.47
33.70
61.50
83.20
                                      87

-------
    90 -
    30
     70
     60
"§>    50
     40
                                                   1 x 10
     30
•   1 x 30 corrected
    to 1 x 10
A   1 x 100 corrected
    to 1 x 10
    20
    10
               10.00      20.00      30.00
                 Concentration of sulfate  standard (
 40.00   ,
 Ug S0a~z,
                                                   nuc
                Figure 28.   S0? calibration curve.
                                   88

-------
lated to give concentrations of 1  (CVS-tunnel only) to 11 ppm  (CVS-tunnel
and vehicle exhaust).  Injections of sulfur dioxide into the tunnel without
exhaust gave recoveries that ranged from 80 to 108 percent with an average
of 98 ± 8 percent (Table 25).

     The recovery of sulfur dioxide in the presence of vehicle exhaust ranged
from 73 to 117 percent with an average of 97 ± 16 percent (Table 26) .  The
sulfur dioxide recoveries were corrected for background levels and for vehicle
baseline emissions.  Background levels ranged from 0.15 to 0.19 ppm while
the vehicle baseline emission levels averaged 7.73 ppm.  The recoveries from
the tunnel in the presence of vehicle exhaust were carried out using a heated
filter and heated sample lines.  The filter was used to prevent particulate
from contaminating the sampling system.  The filter and sample lines were
heated to prevent sulfur dioxide from being retained on the removed parti-
culate.  The recovery experiments indicate that sulfur dioxide can be quan-
titatively recovered from the dilution tunnel.

RESULTS AND DISCUSSION

     The ion chromatographic method of sulfur dioxide analysis is a simple,
sensitive and relatively rapid procedure with a minimal number of inter-
ferences.  Zero air, nitrogen, 3 percent C02 and 100 ppmc HC did not inter-
fere within the minimum detection limit of 0.01 ppm SO2.  However, 100 ppm
NOX and 100 ppm CO produced positive interferences of 0.01 and 0.02 ppm S02»
respectively.  The sulfuric acid-chromic acid bath which had been previously
used to wash the impingers also gave a positive interference for samples col-
lected in impingers washed in this bath.  The problem was averted by replacing
the sulfuric acid-chromic acid with 1:1  (v:v) nitric acid. The manufacturer
of the ion chromatograph has stated that persulfite will interfere with sul-
fate analysis and that oxylate ion will interfere if the separatory column
capacity is reduced.  No problem has been noted with these two species.

     The effect of age on sulfate standards and samples was investigated
and it was found that sulfuric acid standards remained stable for at least
fourteen weeks and the exhaust samples for at least two weeks but less than
ten weeks.  The actual lifetime probably lies between four and six weeks.
This relatively long period of sample stability allows for some leeway in
case the samples can not be analyzed immediately.  The best collection ef-
ficiency was obtained when the dilute exhaust flowed at 4 H/min through two
bubblers in series, each containing 25 m£ of 3 percent hydrogen peroxide
maintained at ice bath temperature  (0-5°C).  Heated glass fiber filters are
inserted in the  sample line to prevent  contamination of the samples and
subsequent column poisoning in the ion chromatograph.  The linearity of
response of the ion chromatograph is maintained in the sulfate concentration
range 0.5 to 100 yg SO4~2 per m£  (100 ppm).  However, changing the attenu-
ation on the ion chromatograph causes a  discontinuity in the calibration
curve.  This discontinuity is seen as a  slope change in Table  14.  The
standards analyzed at each attentuation  obviously  fall into a  linear pattern
even though the slopes differ.  A different set of standards must therefore
be run for each sensitivity setting.
                                      89

-------
                         TABLE  25.   SULFUR DIOXIDE  RECOVERY FROM CVS-TUNNEL  INJECTION
<£>
O
Nominal Flow
Rate ft3/min
Test
1
2
3
4
5
6
7
8
9
0.37
0.37
0,37
0.37
0.37
0.37
0.37
0.37
0.37
CVS
306
306
306
306
306
306
306
306
306
SO? Injected
Vol. S02
(ft3)a
8.547
8.547
8.547
8.493
8.493
8.493
8.446
8.446
8.446
Cone SO2
(ppm)
887
887
887
887
887
887
887
887
887
Calculated
SO2 Cone
ppm
1.08
1.08
1.08
1.07
1.07
1.07
1.06
1.06
1.06
Observed
SO2 Cone
ppm13
0.86
1.09
1.03
1.07
1.12
1.03
1.04
1.15
1.06
Percent
Recovery
80
101
95
100
105
96
98
108
100
                                                                                Average   98 ± 8
              Volume corrected to 1 atm pressure  and 68°F

              Corrected for background  levels of  sulfur dioxide  (0.15  ppm)

-------
       TABLE 26.  SULFUR DIOXIDE RECOVERY FROM DILUTE EXHAUST BY  CVS-TUNNEL
                      INJECTION DURING HOT FTP DRIVING CYCLE


Test
1
2
3
4
5
6
7
8
Nominal
Rate ft
S02
0.36
0.36
0.36
0.36
0.36
0.36
0.36
0.36
flow
•^
/rain
CVS
298
298
298
298
298
298
298
298
SO2 Injected
Vol. SO2 Cone SC>2
(ft3)a (ppm)
8.257 9098
8.257 9098
8.266 9098
8.266 9098
8.266 9098
8.372 9098
8.372 9098
8.372 9098
Calculated
SC>2 Cone
ppm
10.96
10.96
10.96
10.96
10.96
11.11
11.11
11.11
Observed
SC>2 Cone
ppm
7.98
11.34
9.68
10.88
11.91
8.65
12.09
13.04

Percent
Recovery
73
103
88
99
109
78
109
117
                                                                  Average  97 ± 16
Volume corrected to 1 atm pressure and 68°F
Corrected for background levels  (0.19 ppm) and vehicle baseline emissions
 (7.73 ppm) of SO2

-------
     The results of the qualification experiments indicate that most  (97-98
percent) of the sulfur dioxide that is injected into the CVS-dilution tunnel
can be recovered with or without exhaust present.
     The ion chromatographic method of sulfur dioxide analysis is a simple,
sensitive, specific, and relatively rapid procedure with few interferences.
No intermediate steps are involved, lessening the chance of sample loss or
contamination.  The ion chromatograph is sensitive to 0.01 ppm SO2 and
samples can be analyzed in 10 to 15 minutes.  The difference in retention
times between the various ions in the sample allows for definite peak iden-
tification.  Sulfate analysis on the ion chromatograph is also unaffected by
most interferences plaguing a number of other sulfur dioxide procedures.
                                     92

-------
                                 SECTION 7

                         NITROUS OXIDE PROCEDURE
LITERATURE SEARCH

     There are six common oxides of nitrogen:  nitrous oxide (N2O), nitric
oxide (NO) , nitrogen dioxide (NO2> / dinitrogen trioxide (^03) , dinitrogen
tetraoxide (N2O4) , and dinitrogen pentoxide  (^05) .  In addition to these,
there are two different oxides that have the empirical formula N03.  Both are
very reactive and have only been identified by spectroscopy as transient
species.

     Nitrous oxide is a colorless, nonflammable gas at room temperature with
a slightly sweet taste and odor.  Some synonyms are dinitrogen oxide,  nitro-
gen monoxide, hyponitrous acid anhydride, factitious air and laughing gas.
Nitrous oxide is the least reactive and noxious of the oxides of nitrogen.
At room temperature it is relatively inert; but at 500°C, it decomposes to
nitrogen, oxygen,  and nitric oxide.  At elevated temperatures, it will sup-
port combustion and oxidizes certain organic compounds and alkali metals.
Nitrous oxide, N20, has a molecular weight of 44.01, a melting point of
-90.8°C,  and a boiling point of -88.5°C.  It is a linear molecule with a N-N
bond distance of 1.128 A and a N-O bond distance of 11.84 A and is isoelec-
tronic with carbon dioxide.  When inhaled, nitrous oxide may cause hysteria,
insensibility to pain, or unconsciousness and therefore is used as anesthetic
for minor operations, including dentistry.   It is also used as a nontoxic
dispersing agent in commercial whipped cream.  Commercially, nitrous oxide
is prepared by the thermal decomposition of  ammonium nitrate, the controlled
reduction of nitrites or nitrates, the slow  decomposition of hyponitrites,
and by the thermal decomposition of hydroxylamine (12).

     The analysis of nitrous oxide has been conducted using  mass spectrometry,
infrared spectroscopy, and gas chromatography.  Of these, the most sensitive
method is gas chromatography (80).

     There are three gas chromatography methods that have been used to ana-
lyze for nitrous oxide.  Two of these require cold traps to collect and con-
centrate the sample (80).  With the third technique, grap samples are col-
lected in Tedlar plastic bags and analyzed with an electron capture detector
(81).

PROCEDURAL DEVELOPMENT

     The  gas chromatograph operating conditions and sampling system specifi-
cations were obtained from EPA-RTP (81) .  A  two column system with column

                                     93

-------
backflush and isothermal temperature operation was constructed for  the  ana-
lysis.  "She stripper column is a.2' x 1/8" stainless steel tube  filled  with
10 percent OV-17 on 80/100 mesh Gas Chrom Q.  The analytical column consists
of a 6'xl/8"  stainless steel column packed with 120/150 mesh Proapak Q.  A
series of two six-port valves and timers is used to direct the sample flow
through the columns.  The samples are then analyzed with an electron capture
detector.  Since this method has been successfully applied to automotive
exhaust, no other significant effort was applied to the procedural  develop-
ment.  A description of the analytical system and the adapted procedure is
presented as an appendix to this report.

VALIDATION EXPERIMENTS

     Several experiments were conducted to validate the system for  detector
linearity, injection repeatability, and bag sample stability.  Also, a  means
of calibrating the system using permeation tubes and calibration gases  was
investigated.  The results are reported below.

     Initially, a Tracor Model 412 Permeation System with an Ecocal permeation
assembly was used to calibrate the instrument.  The concentration of nitrous
oxide could be set by changing the diluent gas flow over the permeation tube.
This means of calibration can be used between the dynamic concentration range
of 0.23 ppm to 6.31 ppm.  However, a ghost peak was also generated  with this
permeation system.  Efforts to eliminate this extraneous peak were  unsuccess-
ful.  The permeation rate from this tube is not solely dependent on the con-
centration of nitrous oxide and would make this means of calibration diffi-
cult.  Because of the problems associated with the permeation system, a sta-
tic method of calibration was pursued.  Four cylinders of calibration gas
were obtained.  The nitrous oxide concentrations of the cylinders ranged from
1.31 ppm to 9.90 ppm.  No ghost peaks were observed and quantitative results
were obtained.

     Detector linearity over a wide range of concentrations is helpful  and
sometimes necessary when a variety of samples are to be analyzed.   This is
the case with gaseous bag samples.  No easy method to dilute the bag concen-
trations into the linear range of the instrument is available.   Therefore,
the detector must be linear in all of the concentration ranges expected.
The detector linearity for the electron capture detector was determined with
calibration gases from 1 to 10 ppm (Figure 29).  Sample concentrations  with-
in this range are linear with respect to the detector.

     With the gas sample, loop and electrical/pneumatic sample flow  control,
sample injections are not as subject to human error and are more reproducible
than syringe sample injections.  The injection repeatability for the four
calibration gas standards is shown in Table 27.  This table demonstrates that
sample injection reproducibility is reliable with the present analytical
system.

     Due to the length of time required for sample collection and subsequent
gas chromatograph analysis, nitrous oxide must be stable for at  least several
hours. The bag stability was determined by taking two random Tedlarbags filled

                                     94

-------
             10 I-
en
            I
            CL,
           .3  6
           .p
4-J
q
0)
o

o
o


•8
•H
            in
            p
            O
           •H

           3
                                                                     I
                                                                                 I
                                                        6            8            10

                                                     Peak Area  (counts x 10,000)
                                                                                  12
14
                                        Figure  29.  Detector linearity curve,

-------
     TABLE 27.  INJECTION REPEATABILITY OVER THE RANGE
                   OF DETECTOR LINEARITY
                      Concentration, ppm








Average
Standard Deviation
1.31

17935
17811
18346
18149
18319

18112
235
2.16
Area
28020
28759
28974
28931


28671
444
4.95

61146
61057
61004
61325
61448

61196
186
9.90

118085
118935
118687
119448
119810
119005
118995
599
Coefficient of          1.30       1.55       0.30       0.50
 Variation
                             96

-------
with exhaust from an emissions test.  The  two  samples were  reprocessed peri-
odically.  The first bag was processed hourly  for  five hours and the second
bag was processed periodically for several days.   The time  sample decay curve
for both samples is shown in Figures  30  and 31.  Samples may be stored for a
period of five days without adverse effects to the sample concentration.

QUALIFICATION EXPERIMENTS

     Qualification recovery experiments  were conducted for  nitrous oxide with
the dilution tunnel and with real vehicle  exhaust. An aluminum cylinder con-
taining 9000 ppm nitrous oxide in balance  nitrogen was used as the source
for nitrous oxide-  The cylinder was  named by  dilution with zero nitrogen
and the comparison of the diluted sample to a  known standard.  The exhaust
in the experiments was generated from a  Mercedes 240D over  hot FTP (23 minute)
driving cycles.  The flow of nitrous  oxide into the tunnel  was regulated to
give a concentration of approximately 10 ppm nitrous oxide  in the dilution
tunnel.   Injections of nitrous oxide into the tunnel without exhaust gave
recoveries that ranged from 95.1 to 105.6  percent  with an average of 100.5
± 3.8% (Table 28) .  The recovery of nitrous oxide  with real vehicle exhaust
ranged from 82.7 to 113.7 percent with an  average  of 99.6 ± 11.3 percent
 (Table 29) .  Recoveries from the injection into the tunnel  without exhaust
were corrected for background levels  of  nitrous oxide.  The injections with
the vehicle exhaust were corrected for the vehicle baseline emissions of
nitrous oxide as well as for the background levels of nitrous oxide.

RESULTS AND CONCLUSIONS

     The  measurement of nitrous oxide in dilute exhaust  can be conducted with
 gas  chromatography.  Dilute exhaust is  collected  in a Tedlar bag as a grab
 sample.   Sample  analysis of the bag sample with an electron capture detec-
 tor  and  comparison to a set of calibration blends  determines the concentra-
 tion in  dilute exhaust.  The minimum  detectable limit of this procedure is
 0.01 ppm.

     Detector linearity, injection repeatability,  bag sample stability, and
a static means of calibration were demonstrated for the  system.  The electron
 capture detector employed is linear over the range of sample concentrations
expected.  This enables direct sample analysis without secondary dilution.
 The  injection repeatability for an automated sampling system with a gas sam-
pling loop is excellent for the analytical procedure.  The  sample integrity
is also maintained if the samples cannot be analyzed immediately.  This en-
ables minor system repairs without holding up  testing.

     The  average CVS percent recovery is essentially 100 percent in dilute
exhaust for nitrous oxide.  No losses were observed with or without the
vehicle.  This is expected due to the inertness and stability of nitrous
oxide.  Sample integrity can be expected throughout the  entire testing pro-
cedure and sample concentrations are  not subject to the  instability of the
compound  tested.
                                     97

-------
               20
00
               10
            o
            •H
            4J
            10
            c
            0)
            o
            fi
            o
            o

            o
            OJ
            2

            fi
            •H

            0)
            Cn
            c
            u

            4J

            § -10
            u

            0)
              -20
    nominal range for

  injection repeatability
                                                       I
I
                                                       3            4

                                                     Time,  hours
                                     Figure 30.  Sample  decay curve (short term).

-------
            c
            o
              20
              10
            Q)
            O
            C
            O
            O

            O
            CN
   nominal range  for
injection repeatability
vo
to
            C
            •H

            0)
            tTl
            c
            -p
            fl
            0)
            04
              -20
                                                     Time, days
                                    Figure  31.   Sample decay curve (long term).

-------
    TABLE 28.   NITROUS OXIDE QUALIFICATION EXPERIMENTS - NO VEHICLE


                Run                Bag             Percent Recovery*

                 1                 1                    97.2
                                   2                   102.5

                 2                 1                   105.6
                                   2                   101.2

                 3                 1                    95.1
                                   2                   101.1

                                             Average   100.5 ± 3.8 percent

   * Corrected for background levels of nitrous oxide.
TABLE  29-.   NITROUS OXIDE  QUALIFICATION EXPERIMENT WITH; VEHICLE EXHAUST
                Run               Bag              Percent Recovery*

                 1                  I                     82.7
                                   2                     95.0

                 2                  1                    113.7
                                   2                    108.1

                 3                  1                     93.4
                                   2                    104.8

                                            Average     99.6 ±  11.3 percent

   *  Corrected  for  vehicle baseline emissions  and background levels of
     nitrous oxide
                                  100

-------
     This  procedure provides a rapid and sensitive method for the analysis
of nitrous oxide in dilute exhaust.  A single bag sample requires about two
minutes for sample loop purging and seven minutes for the automated analysis,
The total analysis time is about nine minutes per sample.  The automated
system provides simplicity and ease of operation and makes this procedure
ideal for routine analysis.
                                      101

-------
                                 SECTION 8

                        HYDROGEN SULFIDE PROCEDURE

LITERATURE SEARCH

     Hydrogen sulfide is a very flammable and toxic gas at room temperature
and has the characteristic odor of rotten eggs.  The chemical formula for
hydrogen sulfide (hydrosulfuric acid or sulfureted hydrogen) is H2S.  Hydro-
gen sulfide is a bent molecule with a H-S-H bond angle of 93.3° and an S-H
bond distance of 1.3455 A.  Hydrogen sulfide has a boiling point of -60.33°C,
a melting point of -85.49°C, and a molecular weight of 34.08 (12)-  It is a
very weak diprqtic acid with dissociation constants:

                                         K  = 5.7 x 10~8

                                         K  = 1.2 x 10"15

     Hydrogen sulfide may be. detected by its odor at about 1 ppm; however,
olfactory fatigue soon results and higher concentrations may not have an
unduly objectionable odor.  Death is caused by systemic poisoning and res-
piratory paralysis from exposure to high concentrations (>700 ppm) .

     Hydrogen sulfide is prepared commercially as a by-product from many
chemical processes and by the treatment of metallic sulfides with mineral
acid such as hydrochloric or sulfuric acid (12). Hydrogen sulfide produced
in exhaust is probably formed by the reduction of sulfur compounds in the
fuel.  With, an excess of oxygen, it burns to form sulfur dioxide and w.ater:
and with insufficient oxygen to form free sulfur and water:

          2H20 + 02  —>•  2H20 + 2S

Hydrogen sulfide also reacts with sulfur dioxide to form free sulfur and
water:

          2H2S + S02 —>  2H20 + 3S

This reaction may be significant if high levels of sulfur dioxide are pro
duced in exhaust.

     The analysis of hydrogen sulfide has been conducted with an entire
spectrum of analytical methods.  Some of these methods include:  surface


                                    102

-------
reactions on plates, tiles, tapes or filters, wet chemical, fluorimetry,
infrared spectroscopy, sulfur ion selective electrode, coulometry, gas chro-
matography, and colorimetric (82,83).  Most of these are not applicable to
dilute exhaust sampling but are applicable for ambient air sampling or "on
line" systems.  The best applicable means of analysis for dilute exhaust is
the colorimetric technique.

     There are two colorimetric methods available for the analysis of hydro-
gen sulfide.  These are the sodium nitroprusside method and the methylene
blue method (84-93).  The sodium nitroprusside method has a lower detection
limit of about 1 ppm.  This method was not considered sensitive enough for
the concentrations expected in dilute exhaust.  The methylene blue method,
on the other hand, has a reported lower detection limit of 1-2 ppb.

     The absorbing reagent is the key to successful analysis with this pro-
cedure.  Hydrogen sulfide is precipitated as the sulfide in the presence of
metal ions.  Cadmium and zinc hydroxide, cadium sulfate, and zinc acetate
have been used as the absorbing media.  However, several authors have re-
ported the oxidation of cadmium and the photochemical decomposition of
cadmium sulfide.  Bamesberger and Adams  (85) suggested the use of 1 percent
STRactan 10 as a stabilizer for cadmium absorbing solutions.  On the other
hand, zinc solutions do not appear to have these inherent problems.  Flamm
and James  (93) tested all of the above absorbing reagents and found zinc
acetate to be the most efficient absorbant.

PROCEDURAL DEVELOPMENT

     A procedure for the analysis of hydrogen sulfide by the methylene blue
method was obtained from the Project Officer under EPA Contract 68-03-2499.
This procedure is a modification of the technique used by Gustafsson (86) .
A buffered zinc acetate solution is used as the absorbing reagent.  This pro-
cedure was compared to the one recommended for ambient air sampling by Adams
et al  (94) which used cadmium hydroxide as the absorbing reagent.  The
selected analytical procedure is included as an attachment to the interim
report.

     The cadmium hydroxide method presented several problems.  First, sul-
fides in alkaline solutions are easily oxidized by air.  Second, cadmium
sulfide is photosensitive and solutions must be protected at all time from
exposure to light.  The use of special glassware or aluminum soil wrappings
are necessary to prevent exposure to light.  The addition of a stabilizer
such as STRactan 10 helps to minimize the effect of photochemical decompo-
sition, but special handling precautions are still necessary.  Cadmium solu-
tions are hard to work with and in addition some cadmium compounds are toxic.
Cadmium, cadmium oxide, cadmium sulfate, and cadmium sulfide were included
in a tentative carcinogen list issued by OSHA in July, 1978.  Zinc sulfide,
on the other hand, is not photosensitive, the solutions are much easier to
work with, and zinc compounds are not as toxic.  For these reasons zinc
acetate was selected as the absorbing reagent for this project.
                                     103

-------
     Several authors have reported that methylene blue may be bleached by
exposure to light.  In order to determine what effect this might have on a
developed sample, two high (18-19 yg/100 m£) and two low  (2-3 yg/100 mi)
concentration standards were prepared and developed for fifteen minutes.
One of each concentration was then exposed to light.  After developing for
15 minutes in the dark, the other two were wrapped with aliminum foil and
stored in the dark.  The absorbance of each was determined periodically for
several weeks.  The time-light exposure decay curves are  shown in Figures 32
and 33.  The  results of these experiments are discussed  in the Results and
Conclusions section.

     Hydrogen sulfide is readily volatilized from acidic  aqueous solutions.
In alkaline solutions sulfide ion may be oxidized by dissolved oxygen.,  The
pH of the buffered zinc acetate absorbing reagent is 7.0-  This reagent re-
mains at a pH of 7 even after bubbling with dilute exhaust.  Oxidation of
dissolved sulfide ion, does not occur rapidly at this pH.  After addition of
the amine solution and ferric ion to the absorbing reagent, the pH is below
2.0.  At this pH, the trapped sulfide ion reacts to form methylene blue.
Buffering of the absorbing reagent and subsequent change  of pH in the pre-
sence of the amine solution and ferric ion minimizes the  losses due to oxi-
dation or volatilization.

     There are two possible methods available for generating a Beer's Law
plot for calibration.  The first technique requires extensive reagent pre-
paration and tedious titrations. A thiosulfate solution is first standarized
against potassium dichromate.  This standarized thiosulfate solution is used
to standarize a dilute iodine solution.  The standard sulfide solution con-
centration is then determined with an iodimetric method.  Aliquots of the
standarized sulfide solution are used to generate a Beer's Law curve.

     The other method requires the use of a hydrogen sulfide permeation, tube.
The  calibration  curve is generated by bubbling a known concentration of hy-
drogen sulfide through impingers containing the absorbing solution for vary-
ing  lengths of time.  Generation of a calibration curve in this.manner takes
into account the collection efficiency of the impingers.  This method is
quick, efficient and more consistent with the way,the samples are actually
taken.  It also enables the generation of a daily calibration curve without
being manpower intensive.

     The calibration curve for methylene blue is shown in Figure 34.  This
curve was determined for a standarized suflide ion solution on.two separate
occasions and follows Beer's Law at low concentrations.   After about
70 ygS=/100 m£, the curve begins to deviate from Beer's Law-  Concentrations
of hydrogen sulfide in dilute exhaust are expected_to stay well within the
linear range of the calibration curve.

VALIDATION EXPERIMENTS

     After selecting an analytical method, validation experiments were con-
ducted to determine the necessary sampling and procedural parameters.  These
experiments included trapping efficiency, calibration curve linearity, and
interferences from dilute exhaust.  Since the methylene blue procedure is
a well documented analytical technique, only simple experiments were con-

                                    104

-------
o
a
o
CJ
           •  Dark

              Light
               10
                          20
                                    30

                                time, rain.
                                Short term
                                                        50
                                                                  60
   10
              Dark

              Light
               10
   30
time, days
Long term
                                                                  I
                                                                  oO
   Figure  32.   Time-Light exposure  study  low concentration H2S.
                                  105

-------
   25
o
2  20
o
u
!-•
Ul
   10 L,
            Light
            Dark
                                                    Projected from
                                                    long term study
                                 time , hours
                                 Short term
=i
E
,

S
                10
20         30         40
       time, days
       Long  term
                                                         50
                                                                   60
    Figure 33.   Time-Light  exposure study high concentration H2S,
                                     106

-------
1.6


1.5  -


1.4  -


1.3  -
1.0  -
0.6


0.5


0.4


0.3

0.2


0.1
                • Day 1 (11/23/77)


                • Day 2 (12/1/77)
              m^
     If'   .
      0  10   20   30   40   50   60  70   80  90   100  110  120  130  140  150  160  170 180  190  200
                               Sulfide concentration,  pg S=/100 mi
                       Figure  34.   Beer's Law plot for methylene blue.

-------
ducted to verify other procedural parameters.  Collection efficiency and
other sampling parameters were determined with a series of experiments.  The
first experiment determined the collection efficiency at room temperature
(23° to 26°C).  A 5 ppm concentration of hydrogen suflide was passed through
the absorbing reagent at 1.0 and 4.0 H/min.  The experiment was then repeated
with a sample flow of 4.0 Vmin and an absorbing reagent temperature of 0° to
5°C.  This temperature was achieved by immersing the impingers in an ice bath.
The data for this study is shown in Table 30.  Sample flow rate and absorbing
reagent temperature did not have a measurable effect on the collection ef-
ficiency.

          TABLE 30.  THE EFFECT OF SAMPLE FLOW RATE AND ABSORBING
             REAGENT TEMPERATURE ON THE COLLECTION EFFICIENCY

              Absorbing
    Test       Reagent       Sample        Percent H2S Collected per Bubbler
   Number     Temp., °C     Flow Rate         1            2            3

     1           23            1.0           98.0         1.7          0.3
     2           23            1.0           98.2         1.3          0.5
                                      Avg.   98.1         1.5          0.4

     1           25            4.0           96.4         3.2          0.4
     2           25            4.0           96.4         3.0          0.5
     3           25            4.0           95.7         3.8          0.5
     4           25            4.0           95.9         2.5          1.6
                                      Avg.   96.1         3.1          0.8

     1            0            4.0           95.0         5.0
     2            0            4.0           93.9         5.3          0.8
     3            0            4.0           92.8         6.2          0.9
     4            0            4.0           92.7         6.3          0.9
                                      Avg.   93.6         5.7          0.9

     The interferences for this procedure are also well documented.  In an
attempt to find the sources of these interferences, several experiments
were conducted.  The first experiment involved the interferences produced
with only the absrobing reagent.  A series of calibration gases were passed
through the absorbing reagent.  These gases were 495 ppm carbon monoxide,
2.0 percent carbon dioxide, nitrogen dioxide, and a 5 ppmC hydrocarbon blend.
These were compared to the background air and a blank with no gas bubbled
through it.  Each gas was bubbled at 4.0 £/min for twenty minutes and de-
veloped for methylene blue.  No interference from these gases was observed.

     A second experiment investigated the interferences from individual ex-
haust gas components on asulfide ion doped absorbing reagent.  The gases used
were carbon monoxide, carbon dioxide, compressed air, a hydrocarbon blend,
sulfur dioxide and NOX.  Each of these gases were passed through a separate
impinger filled with the doped absorbing reagent for twenty minutes at 4.0
Vmin.  These were then compared to the doped absorbing reagent after de-
velopment of methylene blue.  Table 31 shows the results of this experiment.
NOX at 3000 ppm and sulfur dioxide at 5 ppm were found to quench the produc-
tion of methylene blue.
                                    108

-------
          TABLE 31.  THE EFFECT OF INDIVIDUAL EXHAUST  COMPONENTS
                   ON THE DEVELOPMENT OF METHYLENE BLUE
     Gas
Doped absorbing
  reagent

Carbon dioxide

Carbon monoxide

Doped absorbing
  reagent

Air

Hydrocarbon

Sulfur dioxide

NOX

Doped absorbing
  reagent
NO.
  x
Sulfur dioxide
Sample

  1
  2

  1

  1

  1
  2

  1

  1

  1
  1
  2

  1
  2

  1
  2
  Gas
 Cone.,
  PPM
29,900

 2,709
   168

     5

 3,460
   315
   315

     5
     5
Absorbance

   0.716
   0.708

   0.711

   0.704

   0.552
   0.560

   0.574

   0.554

   0.461

   0.247

   0.648
   0.654

   0.666
   0.621

   0.484
   0.533
Methylene Blue
   Apparent
 Sulfide Ion
Cone. , yig/ml.

    0.646
    0.638

    0.641

    0.634

    0.492
    0.500

    0.513

    0.494

    0.408

    0.213

    0.582
    0.587

    0.599
    0.557

    0.429
    0.475
     A third experiment was designed to check the interference of anions in
the development of methylene blue.  Sodium salts of sulfate, thiosulfate,
and bisulfate ions were used.  Each of these anions was added to separate
solutions of sulfide ion doped and undoped absorbing reagent.  The solutions
were then developed for methylene blue and compared to the doped absorbing
reagent.  Thiosulfate and bisulfate was investigated to determine the speci-
fic source of the sulfur dioxide interference.  The results are shown in
Table 32.  Only bisulfate ion and thiosulfate ion caused the negative inter-
ference .

     Finally, an additional experiment was conducted to help determine the
source of sulfur dioxide interference.  Approximately 2.5 ft3 of 5 ppm sulfur
dioxide was passed through an impinger filled with the zinc acetate absorbing
reagent.  This operation was then repeated six times.  A 1 m£ aliquot of the
                                     109

-------
standard sulfide ion solution was added to two of these impingers.  To two
others, 5 m£ were added.  All six were developed for methylene blue.  These
were then compared to similar concentrations of sulfide ion solution doped
absorbing reagent that did not undergo sulfur dioxide bubbling.  These values
are shown in Table 33<,  Again, the absorbance for methylene blue was de-
creased by the presence of sulfur dioxide.
            TABLE 32.  THE EFFECT OF ANIONS ON THE DEVELOPMENT
                             OF METHYLENE BLUE
                                                         Apparent
                                                        Sulfide Ion
	An ion	   Sample       Absorbance       Cone., yg/m&

                       doped with hydrogen sulfide

Sulfate ion                 1             0.238             0.205
                            2             0.241             0.207

Thiosulfate ion             1             0.207             0.177
                            2             0.211             0.181

Bisulfate ion               1             0.239             0.206
                            2             0.231             0.198

Doped absorbing reagent                   0.241             0.207
                                  un doped

Sulfate ion                 1             0.001
                            2             0.000

Thiosulfate ion             1             0.009
                            2             0.004

Bisulfate ion               1             0.007
                            2             0.005
                                    110

-------
          TABLE  33.   THE  EFFECT OF SULFUR DIOXIDE INTERFERENCE
                   ON THE DEVELOPMENT OF METHYLENE BLUE
          Sulfide  ion
           added,  m&
               1
               1

               5
               5

               0
               0
               1

               5
         Absorbance

  Sulfur dioxide passed
through absorbing reagent

            0.067
            0.070

            0.463
            0.326

            0.002
            0.006

No sulfur dioxide passed
through absorbing reagent

            0.082

            0.583
 Apparent
Sulfide ion
Cone.,  yg/m£
   0.055
   0.057

   0.410
   0.284

   0.000
   0.000
   0.067

   0.521
QUALIFICATION EXPERIMENTS
     A Mercedes 240D was used in the qualification experiments for hydrogen
sulfide.   The baseline emission rate for this vehicle was below the detec-
tion limits for the analytical procedure.  This baseline was established
from three separate hot FTP driving cycles.  Hydrogen sulfide was injected
into the .CVS-tunnel system with and without vehicle exhaust.  The concentra-
tion of hydrogen sulfide injected into the tunnel was 909 ppm.  The flow of
hydrogen  sulfide into the tunnel was adjusted to give a diluted concentration
of approximately 1 ppm.  With the vehicle present, the hydrogen sulfide was
injected  into the raw exhaust stream as it entered the dilution tunnel.
Samples were taken from the dilute exhaust stream and passed through a
buffered  zinc acetate absorbing reagent.  The samples were treated with an
amine solution  and a ferric ion solution and then analyzed with a Beckman
spectrophotome ter.

     Injections of hydrogen sulfide into the tunnel without exhaust gave re-
coveries  that ranged from 85.0 to 96.5 percent with an average of 90.3 per-
cent (Table 34).  Initial experiments for the recovery of hydrogen suflide
in the presence of vehicle exhaust gave recoveries from 60 to 65 percent.
A second  set of experiments with injections of hydrogen suflide into the
dilution  tunnel with vehicle exhaust was carried out.  In this experiment
five samples were collected and treated with 6 m£ of ferric ion solution
while six others were treated with 2 m£ of ferric ion solution (Table 35) .
The samples that were treated with 6 m£ of ferric ion solution gave
recoveries that ranged from 84.1 to 95.6 percent with an average of
                                     111

-------
          TABLE 34.  HYDROGEN SULFIDE RECOVERY - NO EXHAUST PRESENT
Nominal Flow
Rate, ft /min
H2S
Blend
0.375
0.375
0.375
0.372
0.372
0.375
0.375
0.375
CVS
302
302
302
302
302
302
302
302
Run
1
1
1
2
2
3
3
3
Sample
1
2
3
1
2
1
2
3
Calculated
 ppm H2S
  Dilute

   1.13

   1.13

   1.13

   1.12

   1.12

   1.13

   1.13

   1.13
                                                  Observed
                                                    ppm*

                                                    0.96

                                                    0.99

                                                    1.09

                                                    0.98

                                                    1.02

                                                    1.02

                                                    1.07

                                                    1.03
Percent
Recovery
   H2S

  85.0

  87.7

  96.5

  90.0

  87.3

  90.2

  94.6

  91.1
                                                       Average   90.3 ± 3.8
 * Corrected for background levels of H2S


 90.6 percent.  The samples that were treated with 2 m£ of ferric ion solution
 gave recoveries that ranged from 57.8 to 80.8 percent with an average of
 70.7 percent.  Two m£ of ferric ion solution had previously been found to
 be  sufficient in the production of methylene blue in the presence of exhaust
 from gasoline powered vehicles.  This was also the amount of ferric ion used
 in  the previous diesel recovery tests.  The recoveries were approximately
 20  percent higher when 6 m£ of ferric ion were used.  The recovery of 90
 percent using the 6 m£ of ferric ion is also equal to the recovery from the
 tunnel when exhaust is not present.

     Additional tests were performed using different amounts of ferric ion
 solution.  The test using 2 and 6 mJi of ferric ion solution was repeated on
 eight other exhaust samples.  Four were treated with 2 m£ of ferric ion while
 the other four were treated with 6 m& of ferric ion.  The four treated with
 6 m£ gave recoveries of 19 percent higher than the four treated with 2 mi.
 Recoveries averaging higher than 90 percent could not be obtained using more
 than 6 m£ of ferric ion solution.  The recovery experiments without vehicle
 exhaust were repeated and no difference was found when 2 or 6 m£ of ferric
 ion were used.  Several laboratory experiments were conducted to try to de-
 termine what chemical species were involved in this phenomenon, however,
 no  conclusive results were obtained.
                                    112

-------
    Ninety percent of the hydrogen sulfide injected into the CVS-dilution
tunnel can be recovered from the dilution tunnel with or without exhaust
present.   Six m£ of ferric ion solution must be used to obtain maximum
recoveries when diesel exhaust is present.


       TABLE 35.  EFFECT OF FERRIC ION SOLUTION ON HYDROGEN SULFIDE
                       RECOVERY FROM DILUTE EXHAUST

Nominal Flow
Rate, ftj/min
H2S
Blend
0.367
0.367
0.375
0.375
0.377
0.377
CVS
296
296
297
297
296
296
Run
1
1
2
2
3
3
Sample
1
2
1
2
1
2
Calculated
ppm H2S
Dilute
1.13
1.13
1.15
1.15
1.16
1.16
Observed
ppm
0.75
0.83
0.66
0.79
0.86
0.93
m£ Ferric
Ion Added
2
2
2
2
2
2
Percent
Recovery
H2S
67.1
74.3
57.8
69.2
74.8
80.8
                                                            Average  70.7 ±
                                                                         7.8
0.367
0.375
0.375
0.377
0.377
296
297
297
296
296
1
2
2
3
3
3
3
4
3
4
1.13
1.15
1.15
1.16
1.16
0.94
1.04
1.07
1.10
1.02
6
6
6
6
6
84.1
91.1
93.7
95.6
88.7
                                                           Average   90.6- ±
                                                                         4.5
 * Corrected for background and baseline  levels of H2S

 RESULTS AND CONSLUSIONS

     The measurement of hydrogen  sulfide in dilute exhaust  can be  conducted
 with a colorimetric technique.  Hydrogen sulfide is trapped in a buffered
 zinc acetate solution.  Upon  treatment with N, N dimethy1-para-phenylene
 diamine sulfate, and ferric ammonium sulfate, cyclization occurs to form
 methylene blue.  The analysis is  conducted spectrophotometrically  at 667
 nm.  The minimum detectable concentration is 0.01 ppm.
                                     113

-------
     Several experiments were conducted to determine the interferences and
their sources in the analysis of hydrogen sulfide. Individual exhaust gas
components such as carbon dioxide, carbon monoxide, and hydrocarbons  show no
effect on the absorbance of methylene blue.  NOX showed a negative inter-
ference only at concentrations ten times higher than that expected in dilute
exhaust.  Sulfur dioxide also shows a negative interference at 5 ppm.  To
determine the source of the sulfur dioxide interference, several experiments
were conducted.  These experiments were discussed in a previous section.  In
all cases with sulfur dioxide present, the absorbance of methylene blue was
decreased.  Also, a broad peak was observed from 525 nm to 675 nm in the
visible region of the spectrum.  The presence of bisulfate ion and thio-
sulfate ion produced the same effect.  Bisulfate ion can be produced from
sulfur dioxide by the simplified reaction:
                        SO2 + H20 —> HS03~ + H

Thiosulfate ion forms bisulfate ion in strongly acidic solutions:

                        S203= + H+ —»• HSO3~ + S

Sulfate ion shows no interference.

     The presence of sulfur dioxide in the absorbing reagent apparently
quenches the production of methylene blue from hydrogen sulfide.  Sulfur
dioxide acts as an oxidizing agent toward hydrogen sulfide  (free sulfur is
formed) and as a reducing agent toward methylene blue, an oxidation-
reduction indicator.  Sulfur dioxide dissolves in water to form sulfurous
acid.  Since the presence of bisulfate ion produces a similar interference,
the decrease in the absorbance for methylene blue is probably due to this
reaction.  The elimination of the sulfur dioxide interference is not an easy
task.  It should be recognized, however, that an apparent decrease  in con-
centration of hydrogen sulfide is observed when sulfur dioxide is present.

     The sampling parameters used for the collection of hydrogen sulfide in
dilute exhaust were determined partly by necessity and partly by consis-
tency.  The sample flow rate of 4.0 H/min was selected to insure a  sufficient
sample for analysis although the lower flow rate showed a slightly better
collection efficiency in the first bubbler.  The use of an ice bath to cool
the absorbing reagent was selected for simplicity and consistency with the
other analytical procedures which require an ice bath.  The absorbing re-
agent temperature produces little or no effect on the collection efficiency
for ambient temperature sampling, but sample breakthrough is possible at
exhaust gas sampling temperatures greater than ambient conditions.  Two
impingers filled with buffered zinc acetate absorbing reagent are necessary
for complete sample recovery.  These parameters are sufficient to collect
sample concentrations within the detection limits of the procedure  from
dilute exhaust.

     Hydrogen sulfide qualification experiments revealed average recoveries
with or without exhaust present of 90 percent.  Other experiments revealed
that in the presence of the diesel exhaust higher recoveries are obtained
                                    114

-------
when 6 mfi- of ferric ion solution are used  instead of  the  usual 2 m£.

     The effect of light on the stability  of methylene blue was determined
for four samples over a period of weeks.   Both  concentrations were  stable
for about two days whether exposed to  the  light or kept in the dark.  The
high concentration samples required a  slightly  longer time to develop (30
minutes) than the low concentration samples.  Both concentrations exhibited
a steady decay with time after the initial development.   The sample decay was
independent of the exposure to light.   At  10 days for the low concentrations
and 20 days for the high concentrations, an increase  in the apparent concen-
tration was observed.  Inspection of the entire wavelength extinction curve
showed that the absorbance was no longer due to methylene blue but  some other
constituent in the solution.  No attempts  were  made to specifically define
the source of this absorbance.  However, the solutions were found to be
stable in the light for at least several hours  after  development.   If
samples cannot be processed by this time,  it is recommended that they be
discarded because of the difficulty in preserving the sample integrity.

     Two possible absorbing reagents were  compared to determine the best one
for the analysis.  Zinc acetate was selected rather than  cadmium hydroxide.
Ease of use, reduced toxicity, and photochemical stability were the criteria
for this selection.

     This procedure provides  a sensitive method for the  analysis of hydrogen
 sulfide in dilute exhaust.  A single  sample requires  five to ten minutes to
add the reagents, thirty minutes to develop, and three to five minutes to
 analyze the sample.  Absorbing reagent stability helps to simplify  the anal-
ysis and makes this procedure ideal for analyzing a  large number of samples.
                                      115

-------
                                  SECTION 9

                              AMMONIA PROCEDURE
LITERATURE SEARCH

     Ammonia is a colorless, corrosive, and weakly alkaline gas with a dis-
tinctive pungent odor.  It has a molecular weight of 17.03, a boiling
point of -33.35°C (1 atm) , and a freezing point of -77.7°C (1 atm) .  The
ammonia molecule is pyramidal in shape with N-H and H-H bond distances of
1.016 and 1.645 A, respectively.  The H-N-H bond angle is 106.67°.  Ammonia
is soluble in water, ethanol, methanol, chloroform, and ether.  The basic
nature of ammonia allows it to react with protonic acids to form water
soluble ammonium salts.  It also reacts to form stable metallic complexes.
Chemically, ammonia is a highly associated, stable gas with only slight dis-
sociation at 840-930°C and atmospheric pressure.  The toxicity level of
ammonia for humans is about 1700 ppm with an exposure of less than 30 minutes;
however, the 1968 American Conference of Governmental Industrial Hygienists
recommended a threshold limit of 50 ppm, the amount to which most workers
may be exposed to repeatedly, day after day, without adverse affects.  Ammonia
poisoning is not necessarily a serious health hazard though its odor is per-
ceptible at 20-50 ppm (12, 95, 96).  Commercially, ammonia is produced by
the Haber Process according to the reaction:
The reaction is carried out at 400-450°C and 200-600 atm over a specially
prepared catalyst composed of iron, potassium oxide, and aluminum oxide.
The most extensive use of ammonia in industry is in soil fertilization.  It
is also widely used to manufacture nitric acid via the Ostwald process  (12,
96) .

     A number of methods for ammonia analysis are available; however, most
of these methods are subject to interferences, especially from volatile
amines.  These interferences affect a number of colorimetric procedures.
Nessler's reagent is sensitive to formaldehyde, alcohols, organic compounds,
amines, sulfides, acetone, and aldehydes (97).  Distillation is necessary
to remove these interferences.  The indophenol method is more sensitive
than Nessler's (98), but it too suffers from contamination by formaldehyde,
SO2(10:1), Fe, Cr, Mn, and Cu (99).  During color development pH must be
carefully controlled for reliable results (98).  Another highly sensitive
procedure is the pyridine-pyrazalone method.  It is very involved and is
susceptible to interference from some cations at high concentrations (100).
A direct colorimetric method for ammonia analysis involves collection in a
neutral solvent (dioxane)  containing a quinone and subsequent absorbance


                                    116

-------
measurement at 480 nm on a spectrophotometer.   The major drawback  of this
procedure is that one of the reagents,  9-(benzenesulfonamide)-p-benzoquinone,
must be synthesized, purified, extracted with  benzene,  and recrystalized
before use (101).  As with the colorimetric methods,  the most  serious weak-
ness of titrimetry methods is the  large number of interferences.   The
Kjeldahl procedure includes a distillation  step,  but  this does not eliminate
the interference from volatile amines because  they, too, distill over.  The
classical titrimetry methods, acidimetry, complexometry, oxidimetry,  and
formal titration are also used for the  determination  of ammonia, but  they are
generally limited to 10~^ M solutions.   Several instrumental optical  methods
are used in ammonia analysis.  These include the  chloramine (102),  cupri-
ammonia complex  (103) , ninhydrin  (104) , and electroanalytical  methods.
Additionally, there is a method for direct  measurement  on a spectrophotometer
with a UV  (105, 106) or IP,  (107,  108) detector as well  as a number of in-
direct colorimetric methods  (109) .  Gas (110,  111)  and  paper  (112)  chroma-
tography have been employed successfully for some applications.  A number
of electrochemical techniques have also been developed  for ammonia analysis.
Among these are amperometry, polarography,  and coulometric acidimetry,  and
oxidation  (113,114).  Interferences again pose a  problem with  these proce-
dures.  The specific ion electrode for  ammonia is a relatively rapid  and
direct electrochemical method for  determination of ammonia. However, since
a longer equilibration time  (about 20 minutes)  is required for low NH,  con-
centrations, the ammonia gas tends to escape from the basic solution.   This
long equilibration time causes unreliable results in  the concentration  range
of interest  (115).  Volatile amines interfere  with analysis (116),  and  the
hydrophobic membrane of the electrode has been found  to deteriorate in  2 to
3 weeks  (115).

     Both  gasometric (117-119) and gravimetric (118,  120, 121, 122) tech-
niques are not sensitive enough for trace analysis, and the chemiluminescent
procedure  is more involved  than is practical (123, 124).  The  disadvantage
of an enzymatic method reacting ammonia, an a-keto ester, and  reduced nico-
tinamide adenine dinucleotide  (NADH) is the high  cost of NADH  (101) .

     Ammonia has been quantitatively measured  in  dilute automotive exhaust
using an ion chroma to graph.  This  procedure is free from many  of the  common
interferences that plague the classical methods.   The short analysis  time
 (10-15 minutes) makes it a prime  candidate  for ammonia  measurement.

PROCEDURAL DEVELOPMENT

     The procedure chosen for the  analysis  of  ammonia involves the use  of a
new type of liquid chromatograph  called an  ion chromatograph.   Ion chromato-
graphic analysis is direct, relatively  rapid (15-20 minutes),  and  sensitive
to 0.01 ppm NH3.  Heavy metals will contaminate the system, and sodium  and
potassium  ions interfere with ammonia  detection at 2  ppm and 0.5 ppm, re-
spectively.  However, the most common  and troublesome interferences,  vola-
tile amines, do not affect  ammonia analysis on the ion  chromatograph.  The
standards made up in water  and the samples  collected  in a weak acid solution
remain stable for at least  a month allowing some  delay  time before pro-
cessing.   These  advantages made the ion chromatograph the best choice as  a
means of measuring ammonia.

                                     117

-------
     Very little procedural development was necessary for this method of
analysis.  However, instrument and sampling parameters did need to be selec-
ted.  Suggested instrument variables such as type and strength of eluent,
flowrate, chartspeed, and separator column size were provided with the ion
chromatograph.  These will change somewhat with each set of columns.  Ultra
pure nitric acid and distilled water have been found to give the best base-
line and most rapid recovery from suppressor column regeneration.  The re-
generant solution is a 0.5 N NaOH solution made from reagent grade sodium
hydroxide.  Chartspeed was set at 12 in/hr, and the flowrate at about 40
percent of fullscale.  Good separation was obtained with 6 x 250 mm separator
column.  A 3 x 150 mm precolumn (packed with the same resin as the separator
column) was placed on line prior to the separator column to trap heavy metals
and particulate.  If these contaminates get past the precolumn they will
poison the separator and suppressor columns.  The precolumn can be cleaned
weekly with a strong acid solution, as can the separator column if resolu-
tion deteriorates.

     The sampling parameters were determined as part of the validation ex-
periments.  A sampling rate of 4 Vminute at ice bath temperatures was found
to be  most efficient.  Two bubblers containing 25 m£ of 0.01 N H2S04 capture
over 99  percent  of the ammonia passing through.  A filter located between
the sampling  cart and the dilution tunnel, is used to prevent diesel par-
ticulate from contaminating the sampling system.  The line connecting the
filter to the dilution tunnel and the line connecting the filter and the
sampling cart are heated to 175 °F in order to prevent water from condensing
in the sample line.

VALIDATION EXPERIMENTS

     The first validation experiment conducted involved the selection of
sampling parameters:  flowrate, collection temperature, number of bubblers,
and absorbing reagent.  The data from these collection efficiency tests is
found  in Table 36.  Ninety-nine plus percent of the ammonia was trapped in
the  first two bubblers under all test conditions.  A flowrate of 4 £/minute
was  selected to  obtain the most sample without loss of sampling efficiency
or physical loss of absorbing solution.  Sampling at ice bath temperatures
was  selected to  be consistent with other procedures; however, as seen in the
data,  room temperature sampling is also 99+ percent efficient.  Two impingers
containing 25 mi of 0.01 N H2SO4 as the absorbing solution are therefore used
to trap  99+ percent of the ammonia.  .Increasing the acidity of the absorbant
(0.06 N)  causes  interference with the ion chromatographic analysis by
broadening the eluted peaks.  Of significant importance is column contami-
nation that occurs if particulate is not filtered from the sample prior to
analysis.  To prevent this contamination, a filter in the sample line is
used.  Heated lines are used to prevent condensation of water and the loss
of ammonia in the sample line.  The column can be poisoned by heavy metals
present  in the exhaust.  These compounds adhere to the column resin and will
                                     118

-------
    TABLE 36.   NH3 COLLECTION EFFICIENCY AS A FUNCTION OF  FLOWRATE AND TEMPERATURE
14
15
                                          NH3 concentration  (ppm) and

Test

1
2
3
4

5
6
7
8
9

10
11

12
13
Flowrate
U/min)

2
2
2
2

4
4
4
4
4

2
2

4
4
percent in bubbler
Temperature ( °F)
Ammonia
32
32
32
32
Average
32
32
32
32
32
Average
74
74
Average
75
74
1
flow diluted
17.10(99.6)
19.51(99.5)
18.34(95.2)
19.59(99.5)
18.64(98.4)
16.24(99.4)
17.40(100)
14.17(98.3)
17.89(99.0)
17.79(99.9)
16.70(99.3)
17.84(100)
17.87(98.5)
17.86(99.2)
18.83(99.4)
18,05(100)
3
3
1:5 with zero nitrogen
0.06(0.4)
0.10(0.5)
0.28(1.5)
0.06(0.3)
0.13(0.7)
0.02(0.1)
0
0
0.11(0.6)
0.02(0.1)
0.03(0.2)
0
0.17(1.0)
0.09(0.5)
0.02(0.1)
0
0
0
0.64(3.3)
0.08(0.2)
0.17(0.9)
---«* ""'
0.07(0.4)
0
0.24(1.7)
0.07(0.4) --
0
0.08(0.5)
0
0.10(0.5)
0.05(0.3)
0.10(0.6)
0
Bubbler
1+2+3 (ppm)

17.16
19.61
19.27
19.68
18.93
16.33
17.40
14.41
18.06
17.81
16.80
17.84
18.14
17.99
18.94
18.05
2
2
                            Average    18/44(99.7)   0.01(0.05)   0.05(0.3)

                            Ammonia  flow  diluted 1:20  with zero nitrogen
32
32
4.32(99.7)
4.15(98.9)
   0
0.02(0.5)
0.01(0.3)
0.03(0.6)
                                                                      18.50
                            Average   4.24(99.3)    0.01(0.3)    0.02(0.4)

-------
slowly elute causing broad, unidentifiable peaks to appear periodically.   To
prevent contamination, a strong nitric acid solution  (1 N HNC>3)  is  used to .
wash the precolumn weekly.  If the separator column becomes  contaminated it
is washed similarly.  The lighter metals such as sodium and  potassium elute
within a reasonable length of time (less than 12 minutes), but they can in-
terfere with the ammonia peak because their retention times  are  close to
that of ammonia.  Sodium, present in the water supply, interferes when its
concentration exceeds 2 ygNa!"/m£.  The tolerable limit for potassium is only
0.5 yg K+/m&.  Its presence is due to the incomplete rinsing of  glassware
washed in chromic acid solution.  The absorbing solution, 0.01 N H2SC>4,  pro-
duces a small peak with the same retention time as ammonia,  but  a correction
is made for this by running a blank sample each testing day.  Filtered de- ,
ionized water interferes negligibly (<0.01 ppm) with ammonia analysis.     !

     Another variable for which validation tests were run is the ion chroma-
tograph.  The proper combination of eluent, columns, flowrate, and  sample
loop size are required to obtain optimum results.  Nitric Acid  (0.0075 N)
flowing at 200 m£/hour allows good separation between peaks  when a  3 x 150
mm precolumn, a 6 x 250 mm separator column, and a 9 x 250 mm suppressor
column are used.  These parameters will vary between column  sets, making  it
necessary to check the eluent and flowrate when columns are  changed.   A
small loop  (0.01 m£ or 0.2 m£) prevents the relatively small ammonia signal
from being overwhelmed by the large hydrogen ion peak.  An attempt  was made
to neutralize the acid collection medium with sodium and potassium  hydroxide,
but the sodium and potassium interferences were too large to make it-practical.
Injection repeatability.figures are shown in Table 37.  The  mean or average
for each set of peak heights and areas is represented by x,  the  standard
deviation by sx, and the coefficient of variation in percent by  Cv.   The
coefficient of variation serves as a comparison between injections  made on
the two days.  This value is simply the standard deviation divided  by the mean
and multiplied by 100.  Calculations are done using peak areas rather than
peak heights becuase on the whole they were more reliable.   The  average
variation in areas on the two days ran about 3.0 percent.  A comparison was
also made on the repeatability of standard preparation.  Four 0.5 U9 NH4

standards were made up using the same stock solution and analyzed on the  ion
chromatograph.  The results are shown in Table 38.
                TABLE  38.  REPEATABILITY OF AMMONIA STANDARD

                       /yg NH4+\
Sample    Concentration!	1      Attenuation   Height(in)     Area
  1
  2
  3
  4
0.50
0.50
0.50
0.50
3
3
3
3
   0.50
   0.53
   0.52
   0.50
 x 0.50
sx 0.015
Cv 2.9%
28947
29821
29011
26819
28650
 1284
 4.5%
                                      120

-------
                     TABLE 37.   INJECTION REPEATABILITY
                           FOR ION CHROMATOGRAPH
Sample
Date
  1       3-29-78

  2       3-29-78

  3       3-29-78

  4       3-29-78

  5       3-29-78
  1      5-09-78

  2      5-09-78

  3      5-09-78
Concentration Attenuation Height
, /yg mA+\
\ m£ / (Miiiho) (in)
0-50 3 0.53
0.50 3 0.54
0.50 3 0.53
0-50 3 0.53
0.50 3 0.50
x 0.53
sx 0.01
Cv 2.8%
0.50 3 0.51
0.50 3 0.51
0.50 3 0.51
x 0.51
sx 0.00
Cv 0.0
Area
30990
30374
29891
29665
28947
29973
766
2.6%
24335
23205
24785
24108
814
3.4%
                                      121

-------
The coefficient of variation for the area is 4.5 percent.  Subtracting  the
Cv for injection repeatability, the repeatability of standard preparation
 is 1.5  percent.   A 4.5  percent error then is to be expected from the instru-
 ment, and standards.   The particular combination of columns and the condition
 of the  suppressor column determines the actual repeatability.
              i


     The ion chormatograph gives a  linear response  to  ammonia at  the  sensiti-
vity settings of 3 ymtio»andl& ymho.  Table  39  lists concentrations and
corresponding heights and "arenas of points on the calibration curve.   These
values are'plotted'graphically in Figure 35.   (NH4)2SO4 standards  ranging
   1                  yg. NH4+          4.
 from about 0.4 to  30 —-j	  (ppm NH4  ) were run at the appropriate  atten-

uations, 3 or 10 ymho.  The areas recorded at 10 ymho were corrected  to 3
ymho.by multiplying by 10/3.  Both scales show  linearity Taut the slopes are
visually different with relative values of 1.6  and  1.2 for the  3 and  the 10
ymho scales, respectively.  The 3 ymho scale reamins linear from at least
0.4 to 8 yg NH4+/m& and the 10 umho scale from  8 to 30 yg NH4+/m&.

     Sample and standard stability as a function of time was another  factor
 investigated.  The sample was a background sample taken during  the three
bag FTP, SET-7, and FET driving, cycles on May 29, 1978, and standard  used for
                yg NH4+         :
 comparison  (0.5 	r—) was prepared on May 29, 1978.  It is obvious from
 the data presented in Table 40 that the sample  and  standards are stable for
 at least three weeks  (23 days) .  The drop to zero ammonia and the  jump  to
 0.02 ppm NH3 on the thirtieth and thirty-second days, respectively, are
probably due to instrument variation rather than sample degeneration.   By
 the fortieth day  (sixth week), however, the increase in ammonia concentra-
 tions is by 0.02 ppm.  At this point the sample has probably begun to lose
 integrity.  This is further confirmed by the fact that a ten week old FTP
 sample  increased from 0.47 to 0.55 ppm NH3 (17.0 percent) and that two  thir-
 teen week old FTP  and FET samples increased  from 0.08 to 0.11 ppm NH3 (37.5
percent) and from  0.17 to 0.23 ppm NH3  (35.3 percent).  It appears that after
 six weeks the sample concentration begins to increase sharply,  indicating a
 sample and standard lifetime of four to five weeks.

 QUALIFICATION EXPERIMENTS

     Qualification experiments were carried out using a Mercedes 240D vehicle.
Hot FTP  (23 minute test) driving cycles were followed to generate exhaust
for the vehicle baseline emissions and for the  tunnel plus vehicle experi-
ments.  An aluminum cylinder containing 9226 ppm ammonia in balance nitro-
gen was used as the source of ammonia in the experiments.  The  flow of
ammonia into the tunnel was regulated to give concentrations of 10-12 ppm
ammonia in the dilution tunnel.  The baseline ammonia emission  level  for the
Mercedes 240D was 0.18 ppm.  Injections' of ammonia  into the tunnel without
exhaust gave recoveries that ranged from 74.9 to 75.5 percent with an average
of 75.2 percent (Table 41).
                                     122

-------
                  TABLE 39.   CALIBRATION CURVE FOR AMMONIA
   Standard
Concentration
               Attenuation  Height  Heights corrected          Area corrected
\ m£ /
0.36
0.50
0.72
1.00
1.44
1.50
2.00
3.00
3.61
4.00
5.00
7.22
8.00
8.00
10.00
14.43
20.00
28.86
(Umho)
3
3
3
3
3
3
3
3
3
3
3
3
3
10
10
10
10
10
(in)
0.38
0.56
0.79
1.17
1.68
1.89
2.41
3.57
4.21
4.56
5.52
7.50
8.49
2.49
2.93
3.97
4.78
6.30
to 3 yniho scale Area
20,987
33,412
43,798
65,530
94,950
101,103
133,785
188,785
225,901
238,585
315,582
423,980
473,131
8.30 141,874
9.77 170,430
13.23 228,363
15.93 299,706
21.00 409,068
to 3ymho scale







•





472,913
568,100
761,210
999,020
1,363,560
                                      123

-------
   1,400,000  [—
   1,200,000  -
   1,000,000 -

-------
TABLE 40.  SAMPLE AND STANDARD STABILITY
          AS A FUNCTION OF TIME
Date of Analysis

5-29
5-30
5-31
6-01
6-02
6-05
6-13
6-20
6-27
6-29
7-07
Age of Sample (davs)

1
2
3
4
5
8
16
23
30
32
40
Concentration (ppm NH^ )
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.00
0.02
0.03
                   125

-------
                            TABLE 41.  AMMONIA RECOVERY FROM CVS-TUNNEL, NO EXHAUST
                              Ammonia Injected            Ammonia Recovered
Nominal Flow
Rate ft3/min
Test
1
2
3
N
0.
0.
0.
13.
38
38
38
CVS
310
310
310
volume NH3a
injected
(ft3)
8.775
8.756
8.745
Concentration
NH3 injected
(ppm)
9226
9226
9226
Total diluted*
volume
(ft3)
7130
7122
7139
Sample*3 /c Calculated
Concentration Amount of NH3
(ppm) recovered (ppm)
8
8
8
.55
.49
.53
11
11
11
.35
.34
.30
Percent
Recovery
75.3
74.9
75.5
                                                                                             Average   75.2
ro
en    a

      Volume  corrected to 1 atm and 68°F

      Corrected for background level of ammonia (0.06 ppm)
     c
      Each value is the average of three samples taken during each test

-------
     Two separate experiments were carried out for the recovery of ammonia in
the presence of vehicle exhaust.  In the first experiment  (Table 42) the
sample lines were heated to 175°P and unheated 25 mm Fluoropore filters
(0.5 y pore  size) were used to remove particulate.  The recoveries ranged from
53.6 to 63.5 percent with an average of 59.5 percent.  Samples taken without
the filter in place gave similar results.  In the second experiment (Table 43)
the filter (7 cm glass fiber filter) and the sample line between the filter
and the dilution tunnel were heated to 375°F.  All other sampling conditions
remained the same.  In this experiment the recoveries were lower and ranged
from 31.3 to 36.5 percent with an average of 33.1 percent.  At this time,
the reasons  for lower recovery using the heated filter are unknown.   It
was expected that higher recoveries would be obtained, as was the case with
the organic  amines (Section 5).  A twenty-five percent loss of ammonia to
the dilution tunnel can be expected when sampling for ammonia at the 5-10
ppm levels.   An additional fifteen percent of the sample will be lost due to
the presence of exhaust when using an unheated filter to remove particulate.

RESULTS AND DISCUSSION

     The ion chromatograph was chosen as the most favorable means of measuring
ammonia because of the simple, direct, and rapid processing of samples.  Most
compounds that interfere with alternate ammonia procedures do not affect
ammonia analysis on the ion chromatograph.  The selectivity and sensitivity
of this method warrants its use for the analysis of dilute automotive exhaust
samples.

     The sampling parameters providing the most efficieny collection of am-
monia were selected. Twenty-five milliliters of the absorbing solution,  0.01 N
H2S04, is placed in each of the two bubblers in series and maintained at ice
bath temperatures.  Over 99 percent of the ammonia in the dilute exhaust
flowing at 4 H/min is captured in these two bubblers.  After sample collec-
tion it is necessary to set instrument parameters to obtain good separation
in the shortest time possible.  These parameters, such as eluent concentra-
tion and flowrate, will depend on the particular column set in use.  A
6 x 250 mm separator column has been found to resolve ammonia adequately.
The 3 x 150  mm precolumn removes particulate , and the 9 x 250 mm suppressor
column neutralizes the acidic eluent.  With these columns installed, the
eluent, 0.0075 N HNO3, flowing at 30 percent of pump capacity, gives good
ammonia resolution.  A small sample loop  (100 y&) is necessary to prevent
the very broad H+ peak from the acidic absorbing solution from obliterating
the ammonia signal.  The injection variability of the ion chromatograph is
3.0 percent, and for standard preparation the variation is 1.5 percent.  The
ion chromatogarph gives a linear response for  (^4)2804 standards in the

range 0.4 to 30 V$- jp4- -, however, the attenuator is not linear between dif-
ferent sensitivitymsettings.  Therefore, a different set of standards needs
to be run at each attenuation.  The study conducted on the effect of age on
sample and standard stability showed the  lifetime to be between four and
five weeks.   Thereafter, sharp jumps in concentration of ammonia may occur.
Ammonia cannot be quantitatively recovered from the CVS-dilution tunnel with
or without exhaust present.  At ammonia levels of 5-10 ppm there is a twenty-
five percent loss of ammonia to the dilution tunnel and an additional fifteen
percent loss to exhaust.

                                     127

-------
                       TABLE 42.  AMMONIA RECOVERY FROM DILUTE EXHAUST (NO HEATED FILTER)
     Test   NH-.

      1    0.37

      2    0.38

      3    0.37
                         	Ammonia Injected	  	Ammonia Recovered
           Nominal Flow  volume NH3aConcentration  Total diluted**Sample  b'c      Calculated
           Rate ft3/min   injected     NH3 injected     volume       Concentration   Amount of NH3   Percent
CVS

302

296

296
(ft3)b
8.610

8.780

8.550
(ppm)
(ft3)
 9226

 9226

 9226
6937

6816

6797
(ppm)     recovered (ppm)  Recovery

7.01          11.44          61.3

7.55          11.89          63.5

6.22          11.60          53.6

                    Average  59.5
ro
oo
    ,  Volume  corrected to 1 atm pressure and 68°F
       Corrected for background levels (0.08 ppm)  and baseline (0.18 ppm)  levels of ammonia
       Each value is the average of three samples  taken during each test

-------
10
                         TABLE 43.  AMMONIA RECOVERY FROM DILUTE EXHAUST, HEATED FILTER
                             Ammonia Injected	  	Ammonia Recovered
Nominal Flow
Rate ft3/min
Test
1
2
3
NH3
0.37
0.37
0.38
CVS
299
299
298
volume NH-ja
injected
(ft3)b
8.435
8.592
8.699
Concentration
NH3 injected
(ppm)
9226
9226
9226
Total diluted3
volume
(ft3)
6884
6872
6852
Sample D'c Calculated
Concentration Amount of NH3
(ppm) recovered (ppm)
3.55
4.21
3.65
11.31
11.53
11.71
Percent
Recovery
31.4
36.5
31.3
                                                                                               Average  33.1
     ,  Volume corrected to 1 atm pressure and 68°F

       Includes baseline and background correction
     ^"<
       Each value is the average of three samples taken during each test

-------
     The ion chromatograph method of measuring ammonia is an effective and
efficient means of ammonia analysis in dilute automotive exhaust.  This pro-
cedure is insensitive to most of the interferences plaguing other widely
used methods.  The ion chromatograph simplifies ammonia measurements to a
one step injection, avoiding intermediate processes such as distillation,
color development, or reagent preparation.  The lengthiest portion of ammo-
nia determination is the actual analysis time.  This 12-30 minute analysis
is relatively short for such a sensitive method (minimum detection limit is
O.Olppm NH3).  Sample and standard stability as well as linearity of re-
sponse in the concentration range of interest are additional factors which
make this procedure the most desirable method of measuring ammonia in auto-
motive exhaust.
                                    130

-------
                                 SECTION 10
                         ORGANIC SULFIDE PROCEDURE
LITERATURE SEARCH
     The organic sulfides that are included in this analysis are carbonyl
sulfide, methyl sulfide, methyl disulfide, and ethyl sulfide.  The chemical
formulas, molecular weights, freezing points, boiling points, and common
synonyms are listed in Table 44.  Carbonyl sulfide is the only sulfide of
interest that is a gas at room temperature.  In general, the organic sul-
fides are malodorous compounds that produce an unpleasant odor similar to
rotton eggs.  The 1968 American Conference of Governmental Industrial
Hygienists made no recommendation for threshold limit values for these
sulfides.

              TABLE 44.  LIST OF SULFUR COMPOUNDS INCLUDED IN
                     THE ANALYSIS OF ORGANIC SULFIDES
                 Chemical   Molecular  Freezing  Boiling
Sulfur Compound  Formula     Weight    Point,°C  Point,°C
                                              Synonyms
Carbonyl Sulfide COS
Methyl Sulfide
Methyl Disulfide CH SSCH
Ethyl Sulfide
C2H5SC2H5
            60.075
            62.13
            94.20
90.19
           -138.8    -50.2    Carbon oxysulfide
            -98.27
           37.3
Dimethylsulfide
 -84.72    109.7    Dimethyldisulfide

-103.9     92.1    Diethylsulfide
     Several gas chromatographic methods have been used for the analysis of
organic sulfides originating from a wide variety of sources.  A gas chro-
matograph with a thermal conductivity detector has been used by several
workers to analyze gas odorants for mercaptans and/or sulfides (125-129);
however, none of these works were concerned with trace gas analysis.  Gas
chromatography and mass spectroscopy were used to separate and identify low
boiling sulfur compounds in crude oil (130-132).  Temperature programmed gas
chromatography was found to improve the separation of mercaptans and
sulfides (133-135).  The separation and identification of hydrogen sulfide,
sulfur dioxide, mercaptans, alkyl sulfides, and disulfides in Kraft pulp
digester blow gas and black liquor combustion products was accomplished
using gas chromatography (136).

     Carbonyl sulfide has been quantitatively measured in natural gas  (137)
and in carbonated beverages (138) by the use of gas chromatography.  The
measurement of carbonyl sulfide in carbonated beverages used an electron
                                   131

-------
capture detector and had a detection limit of 0.3 ppm.  Improved  sensitivity
in the detection of sulfur compounds in waste process gases was  accomplished
by concentrating the compounds on activated silica gel at -78.5°C,  desorption
under heat and vacuum, trapping at -96°C, and transferring to a  gas chroma-
tograph for analysis (139).

     Several columns have been used to separate sulfur compounds from nor-
mally occurring atmospheric hydrocarbons, but little success has been ob-
tained (140).  A GC-microcoulometry method eliminated the interference  from
the hydrocarbons and was sensitive to 1 ppm mercaptan (141) .  A  gas phase
chemiluminescent reaction of ozone with organic sulfides has been considered
as method of detection in monitoring low concentration of ozone  and sulfur
containing pollutants (142).

     The detection limits for the analysis of sulfur compounds were improved
greatly with the development of the Melpar flame photometric detector (FPD).
The characterization of the FPD response to several sulfur compounds was
carried out by Mizany (143) .  The FPD detector has been applied  to  low  con-
centration air pollution monitoring (72) , measurement of trace organic  sul-
fides in air (144), and soil and water anlaysis (144).  Permeation  tubes
have been used in several cases to generate continuous samples of known
concentrations of various sulfur compounds (71,72).  The use of  Teflon
throughout the gas chromatograph system has been found to minimize  absorp-
tive losses  (144) and has increased sensitivity to 10 ppb (71).

     Several columns have been evaluated at several temperatures  in conjunc-
tion with the Melpar flame photometric detector (77).  The columns  evaluated
were Chromosorb T, Carbopak B-HT-100, Chromosil 310, and Deactigel.   A number
of other sulfur compounds have been quantitatively measured from a  wide
variety of sources using gas chromatography (145-151).

PROCEDURAL DEVELOPMENT

     From the results of the literature search it was determined that the
analysis of the organic sulfides should be conducted by the use  of  a gas
chromatograph (GC) equipped with a flame photometric detector.

     A Perkin-Elmer Model 3920 B gas chromatograph was dedicated for this
purpose.  The instrument has a linearized flame photometric detector (FPD)
and a sub-ambient oven accessory.  The sub-ambient oven accessory allows
for maximum flexibility in determining GC operating conditions.

     A flow schematic of the gas chromatograph analytical system used in the
procedural development work is shown in Figures 36-38.  The sample  is purged
through the gas sampling valve sample loop (Figure 36, Step 1).   The values
are maintained isothermally at 100°C in a valve oven.  The sample is injected
into the gas chromatograph after the system has been efficiently purged
(Figure 37, Step 2).  After all peaks of interest have eluted from the ana-
lytical column, the column is backflushed and the system is readied for the
next injection (Figure 38, Step 3).
                                     132

-------
                                 Control Console
CO
CO
                                                                               Step 1.  Sample being purged
                                                                                       through injection valve
                                                                                    GC Oven
                  Figure  36.   Proposed GC flow schematic for  analysis  of organic sulfides (Step  1).

-------
                                Control Console
GO
                                                                             Step 2.  Sample injected
                                                                                     into GC system
                                                               Valve Oven
                                                                                   GC Oven
                Figure  37.  Proposed GC  flow schematic for  analysis  of organic sulfides (Step 2).

-------
                                 Control Console
CO
cn
                                                                              Step 3.  Column backflush
                                                                                    GC Oven
               Figure 38.   Proposed GC flow schematic for analysis of organic sulfides (Step  3).

-------
     The column selected for the initial time was a 6' x 1/8" Teflon column
packed with 60/80 Chromosil 310.  Several different GC operating conditions
were tried, and a preliminary set of conditions were sleected that provided
an adequate separation of the four organic sulfides of interest.  The
separation of these sulfides is presented in Figure 39.  The elution of
other sulfur containing compounds is also included.  Table 45 presents a
list of chemical and physical characteristics of various sulfur compounds
that could be present in automotive exhaust.

     A lecture bottle of carbonyl sulfide, pure liquids of methyl sulfide,
ethyl sulfide, and methyl disulfide, along with a Tracer Model 412 Permeation
Calibration System containing all four sulfides, were used as sources for
the organic sulfides in the procedural development.  Permeation tubes of
methyl mercaptan, ethyl mercaptan, hydrogen sulfide, and sulfur dioxide,
blends of hydrogen sulfide and sulfur dioxide in aluminum cylinders with
balance nitrogen, lecture bottles of hydrogen sulfide, sulfur dioxide, and
methyl mercaptan, and a pure liquid of ethyl mercaptan were used in the
interference checks.

     Two methods of sample acquisition were considered for the analysis of
the organic sulfides.  One method would be to use sample bags obtained
during the standard CVS testing.  An alternate approach would be to use a
trap packed with a material such as Tenax GC for concentrating the sample.
In this manner an exhaust sample would be pulled through the trap during
the entire test, thereby giving an effective sample volume of several liters
rather than 5-10 m£.  The use of the trap would increase the limits of
detectability by a factor of over 1000.  The collection by the use of sample
bags was discarded due to the expected low concentrations of organic sulfides
in exhaust and the large losses of methyl sulfide, ethyl sulfide, and methyl
disulfide onto the walls of the Tedlar bags at ppb levels.  Because the con-
centration of the organic sulfides is expected to be very low in exhaust, a
number of experiments were conducted to investigate various concentration
techniques that may apply to the measurement of the organic sulfides.

     The first set of experiments involved the use of a U-tube type trap and
was conducted using several trap volumes ranging in size from 5 to 20 mH.
The basic flow schematic of the sampling system is shown in Figure 40.  A
permeation gas blend of carbonyl sulfide and methyl sulfide was used in
these experiments, with the actual concentration depending on the particular
experiment.  Two flow rates through the traps were used:  12.0 m£/min (9.50
ppm COS and 4.77 ppm 0133013) and 81.2 m£/min (1.40 pprn COS and 0.71 ppm
CH3SCH3).  The traps were  maintained at -78°C during the sampling period.
The purpose of these experiments was to see if it is possible to cold trap
(at -78°C) the sulfides and then use the cold trap as a sample loop on the
gas chromatograph system.  Results of these experiments are presented in
Table 46.  Based on these results, it was apparent that the carbonyl sulfide
and methyl sulfide could not both be retained under any of the trap sizes or
concentrations investigated.  The only condition that indicated there may be
some possibility for this method was the large trap loop  (20 m&) at the
higher flow rate and lower concentration.  Even in this particular  case,
the trapping was effective only on methyl sulfide.


                                     136

-------
                    GAS CHROMATOGRAPH CONDITIONS



Perkin-Elmer 3920B w/FPD 6'  x 1/8" column packed  with 60/80 chromosi1

310. N2 at 20 mVmin., oven isothermal at 0°C for 8 min and programmed

to 140°C/min. at 32°C/min.
            ro
            K
            U
            to
            m
            ffi
            U

m
35
U
CO
CO
CO
35
U

•
I
1

0)
M
0)


tn
0!
-P
3
rH
(1)



Q)
^4
0
x;

to
0
-P
3
t-H

0)
&

in
0)
-P
3
i — t
(U

CM
o
en
                      m  co
                     35    t*"1
                      
-------
TABLE 45.  LIST OF CHEMICAL AND PHYSICAL CHARACTERISTICS OF VARIOUS
    SULFUR COMPOUNDS POTENTIALLY PRESENT IN AUTOMOTIVE EXHAUST
 Sulfur Compound

Carbonyl Sulfide

Hydrogen Sulfide

Sulfur Dioxide

Dimethyl Sulfide

Dimethyl Bisulfide

Diethyl Sulfide

Methyl Mereap tan

Ethyl Mercaptan
Chemical
Formula

COS

H2S

SO,,
                 CH3SH
Molecular
 Weight      Density
                                                     Boiling
                                                    Point, °C
  60.075   2.5300 g/H     -50.2

  34.08    1.5392 g/£     -60.3

  64.063   2.927 g/£      -10.0

  62.13    0.848 g/mJl      37.3

  94.20    1.0625 g/mi    109.7

  90.19    0.836 g/mi      92.1

           0.8665 g/m£
             48.11    0.8665 g/m£      6.2
                              62.13    0.8391 g/m£     35
Retention
   Time

   2.8

   4.2

  10.5

  17.5

  15.8



  12.5

  13.5
                                 138

-------
           Permeation
          Calibration
             System
                         Trap
co
                                                              .Valves
System
Control
Console
Perkin-Elmer
 Model 3920B
    Gas
Chromatograph
                                                                    Recorder
                                                                    Integrator
                          Liquid Refrigerant
                                 Figure 40.  Cold trap experiment flow schematic.

-------
      TABLE 46„   THE EFFECT OF COLD TRAPPING AT -78°C ON
CARBONYL SULFIDE AND METHYL SULFIDE AT VARIOUS CONCENTRATIONS,
                   FLOW RATES AND TRAP SIZES

Trap
loop,m£
5.0
5.0

10.0,
10.0

15.0
15.. 0,

20,. 0,
20.0

5.0
5.0

10. 0
10.0

15.0,
15.0

20.0
20.0

Trap
Flow
mH/min
12.0
12.0

12.0,
12.0,

12.0
12.0,

12.0)
12.0

81.2
81.2

81.2
81 ... 2

8.L.2-
81.2

81.2
81.2



in j .
1
2

1
2

1
2

1
2

1
2

1
2

1:
2

1
2


Cone
cos
9.50
9.50

9.50
9.50

9.50
9.50

9.50
9.50

1.40;
1.40

1.40.
1 .40

1.4P,
1.40

1.4Q
1.40

Trap
. , ppm
CH3SCH3
4.77
4-77
Ayg.
4.77
4.77
Avg>
4. 77,
4.77
Avg.
4.. 77
4.77-
Avg.
0.71
0.71
Avg.
0.71-
0.71
Avgv
0.71
0.71
Avg.
0.71
0.71
Avg.
Inlet
Peak
COS
84.5
84.5
84.5
84.2
84.2
84.2
84.2
84,. 2
84.2
85.0
85.0
85.0
63.0
63,5
63.2
63.0;
6.3-- 5,
63.2
63,0;
63.5.
63.2
63.0
62.0.
62;5
• Trap Exit
Height
CH3SCH3
16.1
16.0
16.0
13.3
13.4
13.4
13.3
13.4
13.4
14.6
14.4
14.5
5.2
5.2
5.2'
5.2
5.2
5.2
5.2
5.2
5.2;
5.5
5.0,
5.3
Peak
COS
84.2
84.2
84^2
84.5
84.5
84.5
84.9
84.9
84.9
84.9-
84,9
84.9
64.9
65.0
65.0
63.0
63.5
63.2
62.0
62.1
62.1.
65.0
66.0
65.5-
Height
CH3SCH3
4.1
13.8:
9.0
1.1
11.4
6.3
12.6
13.0
12.8
0.0
1D..O
5.0
5.2
4.2
4,. 7,
4,2
4.2
4.2-
5.0
4.5
418
0.0,
0.0,
0.0
                              140

-------
    One additional trap temperature was investigated prior to elimination of
cold trapping as a possible concentration technique.   In this experiment the
trap was cooled to liquid nitrogen temperature.  The results of this experi-
ment are presented in Table 47.  Only one set of concentrations (1.40 ppm
COS and 0.71 ppm CH3SCH3) was used in this experiment.  All sample loop
sizes were somewhat effective in collecting the carbonyl sulfide, with a
nominal collection efficiency ranging from 76 to 86 percent.  The collection
efficiency for methyl sulfide was slightly higher, but the repeatability
was less, probably due to the low peak heights for this species.  From these
experiments it was apparent that cold trapping, using traps up to 20 m£ in
volume, and temperatures as low as -196°C would not quantitatively remove
either carbonyl sulfide or methyl sulfide.

     Efforts were then directed toward determining the feasibility of using
short stainless steel cartridges packed with an absorbing material to con-
centration the organic sulfides.  The absorption traps are lengths of stain-
less steel tubing 2 inches in length and  3/8" OD.  The material is held in
the stainless cartridge by stainless micron inserts in each of the unions
on both ends.  Four packing materials were selected to be evaluated at four
collection temperatures.  The four packing materials that were used in this
experiment include Tenax-GC, Chromosorb 102, Porapak Q, and Chromosorb T.
The collection efficiency of these traps was evaluated at temperatures of
20°C, 0°C, -78°C, and -196°C.

     A permeation calibration gas sample  containing 1.40 ppm carbonyl sulfide
and 0.71 ppm methyl sulfide in a balance nitrogen gas was used for this
study.  The results of this study are presented in Table 48.  Three of the
four packings are essentially 100 percent efficient in removing both carbonyl
sulfide and methyl sulfide at -78°C.  These three packings were Tenax-GC,
Chromosorb 102, and Porapak Q.  All of the traps except Chromosorb T were
effective in removing methyl sulfide at all of the temperatures investigated.
Problems were encountered using trap temperatures of -196 °C.  At this tem-
perature flow  restrictions were noted in the trap as  the test proceeded.
There were also problems in desorbing traps that were  stored at this tem-
perature .

     The sulfides were thermally desorbed from the traps by connecting the
traps into the gas injection system with two quick connects and immediately
injecting the sample into the GC system and placing the traps inside a Lind-
burg furnace operating at 300°C.  The carrier gas upon injection flows
through the loop carrying the contents of the trap into the gas chromatograph.
The 300°C temperature is the temperature needed to thermally desorb the traps
without causing broad sulfide peaks which result from  gradual thermal de-
sorption.  The 300°C temperature is also  low enough to prevent the destruc-
tion of the packing material in the trap.  The packing material which gave
the most reproducible results in the desorption experiments was Tenax GC.
For this reason and its stability at the  300°C desorption temperature, the
Tenax GC packing material was selected for use in subsequent experiments.

     A Tenax trap at -76°C was used to collect the exhaust from a 1975 Model
350 CID Chevrolet engine for the 31 minutes of an FTP.  The resulting trap
                                      141

-------
    TABLE 47.  THE EFFECT OF COLD TRAPPING AT -196°C ON
CARBONYL SULFIDE AND METHYL SULFIDE WITH VARIOUS TRAP SIZES

Trap
loop ,m£
5.0
5.0

10.0
10.0

15.0
15.0

20.0
20.0

Trap
Flow
mVmin
81.2
81.2

81.2
81.2

81.2
81.2

81.2
81.2

Trap Inlet

in j .
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
Cone
COS
1.40
1.40
1.40
1.40
1,40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
. , ppm
CH3SCH3
0.71
0.71
0.71
0.71
0,71
0.71
0.71
0.71
0.71
0.71
0.71
0.71
Peak
COS
66.7
66.7
66.7
67.8
68.1
68.0
67.8
68.1
68.0
78.0
78.9
78.5
Height
CH3SCH3
5.3
5.3
5.3
5.2
5.2
5.2
5.2
5.2
5.2
5.5
5.5
5.5
Trap Exit
Peak
COS
15.8
15.8
15.8
16.0
15.2
15.6
8.0
10.8
9.4
8.0
14.0
11.0
Height
CH^SCH^
0.8
0.1
0.5
0.0
0.0
0.0
0.1
0.1
0.1
0.0
0.1
0.05
                             142

-------
            TABLE 48.  THE EFFICIENCY OF VARIOUS MATERIALS
               TRAPPING SULFIDES  AT SEVERAL TEMPERATURES
                                     Gas chromatograph Response-Peak Ht.

Inj.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
Trap
Temp. °C
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
Before
Trap
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Chromosorb 102
Ch romos orb 10 2
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chormosorb T
Chromosorb T
COS
68.2
69.5
68.9
65.0
63.5
64.3
64.8
65.0
64.9
68.2
69.5
68.9
64,0
64.2
64.1
67.2
67.2
67.2
62.0
63.6
62.8
64.0
64.2
64.1
67.2
67.2
67.2
65.0
63.5
64.3
64.8
65.0
64.9
64.0
63.0
63.5
CH^SCH^
4.2
5.3
4.8
4.5
4.7
4.6
5.0
5.0
5.0
4.2
5.3
4.8
4.9
4.3
4.6
4.8
4.8
4.8
5.0
5.0
5.0
4.9
4.3
4.6
4.7
4.5
4.6
4.5
4.5
4.5
5.0
3.9
4.5
4.8
4.8
4.8
After
COS
64.5
67.5
66.0
64.5
62.0
63.3
0.0
0.0
0.0
66.5
68.0
67.3
20.8
20.0
20.4
0.0
0.0
0.0
63.2
62.6
62.9
35.5
45.0
40.3
0.0
0.0
0.0
64.8
65.1
65.0
64.0
63.2
63.6
62.5
61.0
61.8
CH3SCH3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.9
4.5
4.2
5.0
5.0
5.0
3.0
0.0
1.5
Trap size  2" x 3/8" OD
                                    143

-------
was thermally desorbed into the GC analysis system.  The gas chromatograph
trace obtained from the desorption of the trap indicated that a  substantial
number of sulfur containing compounds were present.  The major drawback ob-
served was the exceptionally large sulfur dioxide  (802) peak.

     The concentration of sulfur dioxide in the dilute exhaust will normally
range from 2 to 5 ppm, whereas the other sulfides are present only in the ppb
range.  The sulfur dioxide is unstable and the proposed method is not de-
signed to quantitatively measure sulfur dioxide.  Since sulfur dioxide has
no quantitative interest, efforts were directed to determine if  techniques
are available that would allow sulfur dioxide removal without altering the
concentration of other sulfides.  A packing material containing  sodium bi-
carbonate (NaHC03) has been reported to be very effective for this purpose.

     Experiments have indicated that hydrogen sulfide  (I^S) was  not stable
enough to quantify with this procedure.  In order to remove the  interference
from sulfur dioxide and to remove any remaining hydrogen sulfide, the dilute
exhaust is passed through a 2" x 3/8" stainless steel cartridge  packed with
5 percent sodium bicarbonate on 45/60 mesh Chromosorb P-AW DMC before enter-
ing the organic sulfide collecting Tenax GC trap.  The sodium bicarbonate
trap effectively removes sulfur dioxide at 10 ppm levels !and hydrogen sulfide
at 1 ppm levels without affecting the organic sulfide concentrations.

     Initially, a set of gas chromatograph operating parameters  was developed
to provide separation of hydrogem sulfide, carbonyl sulfide, sulfur dioxide,
methyl mercaptan, methyl sulfide, methyl disulfide, and ethyl sulfide.  Since
hydrogen sulfide and sulfur dioxide are not of quantitative interest, the
original GC operating parameters were modified to shorten the analysis time.
The initial GC oven temperature of 0°C was maintained for four minutes and
then temperature programmed to 140°C at 32°C/minute.  The entire analysis
time was about 25 minutes.  Operation of 0°C was originally selected to allow
separation of hydrogen sulfide and carbonyl sulfide.  Since this separation
would no longer be necessary, the GC oven parameters were changed to provide
a compromise between separation and analysis time.  This new programming rate
provides for sample injection at 80°C followed by immediate programming to
140°C at 16°C/minute.  A typical trace of the organic sulfides is shown in
Figure 41.

     Several recovery experiments were conducted- using the Tenax GC traps.
These experiments were designed to determine the recovery of the organic
sulfides from the Tenax GC traps.  The recovery from these traps was very
erratic and was not satisfactory.  Initially, it was felt that the lack of
reproducibility was due to the technique employed to remove the  organic sul-
fides from the Tenax GC traps.  However, the GC analytical column was later
found to be suspect.  Contact was made with those researchers who originally
used the GC parameters to quantitatively measure organic sulfides.  Their
findings were similar to those experienced at SwRI.  When using  this column
packing near its maximum operating temperature, very erratic results were
experienced.  After looking into this more thoroughly, it was decided to use
a different column packing that would be reproducible and still  yield satis-
factory separation of the organic sulfides.

                                     144

-------
     16      14       12      10       86




                          Retention time, minutes
Figure  41.   Typical  gas chromatograph trace of organic sulfides.
                                 145

-------
     The column packing that was selected was a specially treated Porapak QS.
Although this column is reported to be stable at higher temperatures, the
separation characteristics are not as good as the Chromosil 310 column.  The
ethyl sulfide and methyl disulfide elute together,  A typical calibration
blend from a permeation system is presented in Figure 42.  This column was
also found to give inconsistent results after repeated use and a different
analytical column was sought for use in measuring the organic sulfides.
After reviewing the literature and conducting a brief cursory laboratory
study, it was found that the column which has the necessary qualifications
for the organic sulfide analysis is a 6' x 1/8" TFE Teflon column packed with
60/80 mesh Tenax GC.  A typical gas chromatograph trace using the Tenax GC
analytical column for the analysis of the four organic sulfide is shown in
Figure 43.

     In order to determine the efficiency of the collection of Tenax GC ab-
sorbing traps, a secondary dilution of the permeation calibration system was
included.  The organic sulfides were diluted from a 0.1 - 3 ppm level down
to the detection limits of the FPD.  A sample of the permeation blend after
secondary dilution is presented in Figure 44.  As noted, only two of the
four peaks are above the detection limits, although all four organic sul-
fides and the concentrations are listed at their elution time.  The concen-
tration of the organic sulfides with Tenax-GC traps appears to have tremen-
dous potential.  An example of the permeation calibration blend (with second-
ary dilution) after being sampled at a flow rate of 45 m£/min for 10 minutes
is shown in Figure 45.  Only the four individual organic sulfides are ob-
served, and no extraneous peaks (reaction products, etc.) are observed.

     A system was developed to re-condition the Tenax GC traps by purging
the traps with nitrogen at 500 m£/min for seven mniutes at 300°C.  Several
spot checks of traps that had been conditioned under these conditions indi-
cated no trace of organic sulfide carry-over from "used" Tenax-GC traps.
The procedure that is used to desorb the organic sulfides from the traps is
also very efficient in that no organic sulfides are retained in the trap
after the thermal desorption using the GC procedure.

     The Tenax GC traps have been found to effectively remove 100 percent of
the organic sulfides from a permeation calibration flow at a flow rate of
130 m£/min when the trap is maintained at -76°C.  Higher flow rates were
tried, and a breakthrough into the back-up Tenax-GC trap was observed at a
flow rate of 250 m£/min.  When this occurred, it was decided to return to
130 m£/min and use this flow rate as the primary sampling flow rate.

     A variety of other trap designs,  temperatures, and flow rates may be
equally acceptable; but for the purpose of developing a procedure with spe-
cific goals, these conditions have been selected.

     Problems have been encountered with batch to batch variation in the
Tenax-GC which have caused repeatability problems.  A procedure has been
implemented to validate each Tenax-GC batch prior to sampling as well as
each individual trap.
                                    146

-------
_-_





	

	 .









	


- •
—


	



	





-








—





	




	

- -J
J
.
















—


oi 	

-U)..
SB
IN
U'
.n3 	
3
,.s .
to---
(M

rq
UJ 	

a
,V;






	














	

- -




	



-•'•-







. ...









	




	
.::. ..



—


	

. . ..


.: :.
»•*



— — tn-




	 ••*-











P*1






— _.
_




" ^-v
N



-IrfJ 	





.l/>....,





	
|*-^




'X}






	
O^



	
o--




— . 	






	





	




... .
.._






	




.



	

H
	


I
._).





























	









—



— . .



(A'
r*r
•^%- •
R"




...


	





•


	

	



.. .

.„ 	



	



	






	

	
	


	





	
	
....
.
	





	


	




	






	

	
	


	




...
	
t^r,. peBMEAT/OM BL6ND Date NOV 1, 111!
Innrument r>
Column l.f
Pocked with
on 80//00
RunlSO®
*_fi*-P«
Inlet ISC
Detector; ISt
Hyd. 40
( '~H/f>
Recorder /
Inieclion /g
Sampling Dg




	








	
	







'.::.




—
r^>



...

—

._ —




;._.


a 3**:
.'t,^
N*
mesh
•W O upera"or fj mrnVrllfWF
'f," no I.D. Tf^1
% »t. ^/^ Li
PA8APAK O&
/V//t °C. using _3L
r. /VA Rotameter P
^CCOLOMM
1. 	 °c PpO
_pii» N/A Rotan
Mia V/A Bour
tataAf/A Rotan
_ln/min spend / n
All. indicated /O ;
vice £/)$ SflMPl



	
r-'




	
5T

	
-


(T>t
••

O
...I . ..
I
1
I
I
1
I
I
•
11
•_";;:-
1 1
1 1
.S 1 ._
. .u fj 1 	
If 1
II 1
~.-jM-—
i
|| 1
J\\ J

_cc/r
Ida. ,
nln. ^/A
Type
j. Phase
Support
Carrier
VIA
•*a'^2lO°eX'Jmtiue*
FVP8 (Ot
neter Re
neter Re
neter Re
nV.F.S.
ml net
bar!
o. -#O cc/min
a. 31O
to. ^Xl
cc/min
cc/min
50LT££ Type
//? m\
'-W6 VALVS







_





—
::...

~^

	






... ..
_.. _..


	 j
(•
|
' f
	 >


...










Actual



. ._ rfVJ

j




tj f3

      19   18  17  16   15  14  13  12  11   10   9   8   76   54    32   1




                                     Retention Time, minutes
Figure  42.   Typical GC separation of  organic sulfides  on acetone-washed Porapak QS  column.

-------
                                                                           sui-Pioes   oat. NOV Z?  1977
                                                                                   Operator P. SflUAl66»S
                                                                                           TPff
                                                                      ft.    fr" O-D-
                                                              Packed with /Vd % wt.   V/l
                                                              on   _
                                                              Run ISO ® _M&_- C- using  3O  cc/min
                                                                    osia. A04 Rotameter Rdg:
                                                                                    Typelotfierl
                                                                                 Rotameter Hdg.
                                                                                             3&O  ec/min
                                                                                                   cc/min
                                                                                                    Type
  (  >_j/fl
Recorder  /  in/min »eed
Rotameter Rdg.
   mV.F.S.
                                                            Injection_/jf2_  ul indicated   /Q  ul net
                                                              Sampling Device <5/iS  £AM?i-l»/6 VAU/E
                                     11    10   9    876
                                       Retention time, minutes
Figure  43.   Typical organic  sulfide  separation  with Tenax-GC column.
                                               148

-------
1 — • — i 	 : —
• ; . .__ _ _^
I 	 : 	 1 	 	 1 	
1 i
I 	 f— 	 >• —
• ** ' "*
1 	 " — * :

I j '
1 	 ; 	 1 	 : 	 : 	
|_ 	 : 	 J 	 ; 	 1 	
Sample PE*M SvST -7 DILUTED Date NOV 200@ a°/fllin Type
Detsctar: /6O C FrD Type (other) nff
Hyd. ^O psig /V>9 Rotameter Rdg. 4O cc/min
Air 7^ psiq AfA Rotameter Rdq. ^£ ^ cc/min
( ) Af/t psia AVJ Rotameter Rda. A/A cc/min
Recorder / < M I indicated /,'-'•

4-t ' i
s) . . , ; •
M
u

>.
r

0)
te
n
	 ; 	 _
i

.
.

•
,
: . i
                                                                .	 a)
20 19  18 17  16  15 14  13  12  11  10  9   8   7   6  5   4  3   2

                         Retention Time,  minutes


   Figure 44.  Organic sulfide permeation blend with secondary
         dilution, near  detection limit  of GC  FPD system.
1  0
                               149

-------
      Sample
                               .oat.<
Instrument Pg 'ZQZO S    Operator £.
Column   A  ft.   Vfl  0.0.      I.D.
  Packed with  /yA % wt.     fJA
    &O/30mesh  TEN A1 -
     . __~^r^~ ^« > ^ a A   /   ^ ^
  Run ISO ®  V/»°C. using
  @ 7S~ psio.  X/XJ Rotameter Rdg.
Inlet  16,0 ~Vtr~30°-7iaQ'
Detector:  1 1* O    °C
  Hvd. ^O  Piig
  Air  7O  p>ig
        Mir   / w  L>»IU  /»^r   Rotametar Rda.
        ( I  A^^  psig  fit ft   Rotameter Rda.
      Recorder  /
                                         Uq. Phase
                                 _ ^__ Support
                               cc/min. //A.    Carrier
                             Type (other)   A/A
                          Rotameter Rdn.   4/Q   cc/min
 	_../min speed   /   ™v F a  SOLT& C.
Injection  /g ml indicated   IO Ml net    IO   Al Actual
        ction /£?  ml indicated   /O »il net
        Sampling Device f?^5  SAftlfi/M
20  19  18   17 16   15  14   13  12   11  10   9    87    65    4    3   2    1

                               Retention  Time, minutes


      Figure  45.   Organic sulfide  permeation blend with  secondary
    dilution, concentrated  on  Tenax-GC  trap  and thermally  desorbed
                                 into GC  FPD system.
                                             150

-------
     The procedure chosen for the analysis of the organic sulfides consists
of (1)  collecting the organic sulfides on a Tenax GC trap at -76°c-  (2) ther-
mally desorbing the organic sulfides from the trap into the GC sampling sys-
tem;  (3) injecting the organic sulfides into the GC;  (4) analysis of the or-
ganic sulfides with a GC equipped with a sub-ambient oven accessory, a 6' x
1/8"  Teflon analytical column packed with Tenax-GC, and a flame photometric
detector; (5)  and quantifying the results with the use of permeation cali-
bration tubes.  A finalized copy of the procedure is included an an appendix
to this  report.

VALIDATION EXPERIMENTS

     Several experiments were carried out to determine the validity of the
organic sulfide procedure for the analysis of carbonyl sulfide, methyl sul-
fide, ethyl sulfide, and methyl disulfide.  These experiments included checks
for:   GC injection variability, linearity of detector response, sample stabi-
lity in the Tenax traps, trapping efficiency of the Tenax traps, interfer-
ences, and test repeatability.

     The finalized sampling conditions used to collect the organic sulfides
are listed below as is a discussion on their selection.  A 2" x 3/8" OD
stainless steel trap packed with Tenax-GC is used to collect 99+ percent of
the organic sulfides.  During sampling, the trap is kept at -76°C in a dry
ice-isopropyl alcohol slurry.  This temperature is necessary to effectively
trap the four organic sulfides from the dilute exhaust sample.  Higher tem-
peratures (greater than 0°C) allow carbonyl sulfide to break through the
Tenax-GC trap.  The other three sulfides can be effectively trapped even at
temperatures as high as 20 °C.  The sample flow rate through the trap is main-
tained at 130 m£/min.  At higher flow rates  ( 250 m£/min) , breakthrough of
the organic sulfides occurs.  A flip top filter, a Perma Pure Drier, and a
trap containing 5 percent sodium bicarbonate on 45/60 mesh Chromosorb P-AW
DMCS precede the Tenax GC trap.  The flip top filter removes particulate from
the gas stream prior to flow through the Perma Pure Drier.  If particulate is
allowed to enter the Perma Pure Drier, it could posion the drier and prevent
it from functioning properly.   The Perma Pure Drier removes moisture from
the gas stream which could freeze out in the -76°C Tenax-GC trap, thus re-
stricting or stopping flow through the Tenax-GC trap.  The 5 percent sodium
bicarbonate trap removes sulfur dioxide from the gas stream and prevents it
from collecting in the Tenax-GC trap.  The sodium bicarbonate trap will re-
move 10 ppm sulfur dioxide at a sample flow of 130 m£/min continuously from
dilute exhaust for periods up to 30 minutes.  Tenax-GC was chosen as the or-
ganic sulfide absorbing material over the other packing materials due to its
trapping reproducibility and ability to withstand desorption temperatures.

     The Tenax-GC traps can be used many times without replacing the Tenax-
GC packing material.  There is a large deviation in trapping effiency from
batch to batch of the Tenax-GC packing and each batch must be validated prior
to sampling.  Each trap is conditioned in an oven operating at  325°  ± 25°C
for one hour with a flow of zero nitrogen  (500 m£/min) passing  through the
trap.  No carry over of sulfides in the Tenax-GC traps has been found from
test to test.  This lack of carry over indicates that the desorption process
removes 100 percent of the sulfides collected on the  trap.

                                     151

-------
     The sample traps must be stored at -76°C before desorption and analysis
or carbonyl sulfide will be lost from the traps.  The other three sulfides,
methyl sulfide, ethyl sulfide, and methyl disulfide, are stable in the traps
overnight at room temperature,  in most cases, all traps are run between one
half hour and three hours after sampling.  The traps are capped after sam-
pling with miniature quick connects to prevent condensation of water and
other compounds into the trap before analysis.  After the traps have been
desorbed, they are again capped to prevent contamination before they are used
in sample collection again.

     To determine the GC injection repeatability for the organic sulfide
procedure, a permeation standard containing 1.95 ppm carbonyl sulfide,
3.31 ppm methyl sulfide, 0.84 ppm ethyl sulfide, and 0.20 ppm methyl disul-
fide was injected into the GC analytical system six consecutive times.  The
results of this injection repeatability experiment are presented in Table 49.
The percent deviation varies from 1 percent for methyl sulfide to 6 percent
for methyl disulfide.  This deviation appears to increase with decreasing
concentration or organic sulfide.

         TABLE 49.  INJECTION REPEATABILITY FOR THE ORGANIC SULFIDES

                          Average GC     Standard       Percent
          Compound	     Peak Area     Deviation     Deviation

      Carbonyl Sulfide       29596          420           1.4

      Methyl Sulfide         52325          449           0.9

      Ethyl Sulfide          11201          254           2.3

      Methyl Disulfide        3951          243           6.2

     To determine the test-to-test repeatability for the procedure two ex-
periments were carried out.  In the first experiment, organic sulfides from
a diluted permeation blend were collected on a Tenax-GC trap, desorbed into
the injection system, and injected into the analytical gas chromatograph
system.  This sequence was repeated 5 times using the same Tenax-GC trap and
the resulting GC peak areas for each of the organic sulfides were averaged
over the 5 tests.  Standard deviations and percent deviations were also de-
termined for the organic sulfide GC peak areas.  The results of this experi-
ment are presented in Tabel 50.  Standard percent deviations ranged from
7 percent of methyl sulfide to 10 percent for carbonyl sulfide and methyl
disulfide.  The second experiment was identical to the first experiment
except that 5 different traps were used to collect the organic sulfides
instead of using the same trap 5 times.  Table 51 shows the results of this
experiment.  Standard percent deviations ranged from 13 percent for methyl
sulfide to 26 percent for ethyl sulfide.

     To determine the linearity of the detector for the concentration ranges
of interest, a permeation system containing permeation tubes of all four
sulfides was used to generate varying concentrations of the sulfides.


                                      152

-------
TABLE 50.  TRAP REPEATABILITY FOR ORGANIC SULFIDE COLLECTION
Test 1
Test 2
Test 3
Test 4
Test 5
Average
Standard
Deviation
Percent
Deviation
TABLE 51.
Test 1
Test 2
Test 3
Test 4
Test 5
Average
Standard
Deviation
Percent
Deviation
COS
Area
20241
22052
20092
22343
17378
20421
±1985
9.7%
TRAP- TO- TRAP
COS
Area
47590
46830
44465
50590
25788
43053
±9896
23.0%
Me2S
Area
71938
66098
74326
63777
65494
68327
±4548
6.7%
REPEATABILITY
Me2S
Area
68995
43429
66874
63440
76180
65784
±8332
12.7%
Et2S
Area
34417
35156
38123
30508
32516
34144
±2864
8.4%
FOR ORGANIC
Et2S
Area
25716
20549
17283
30144
16631
22065
±5773
26.2%
Me2S2
Area
41346
39537
41683
34147
34321
38207
±3718
9.7%
SULFIDE COLLECTION
Me2S2
Area
13839
11487
10069
15901
10338
12327
±2491
20.2%
                              153

-------
     Figures 46-49 show plots of the GC peak areas vs. the nanograms of each
sulfide injected into the GC system.  Carbonyl sulfide and ethyl  sulfide
give linear responses in the 1-200 ng region, methyl sulfide  gives  a  linear
response in the 1-120 ng region, and methyl disulfide gives a linear response
in the 1-55 ng (higher levels of methyl disulfide were not tried) region.
Above 120 ng of methyl sulfide and 200 ng of carbonyl sulfide, the  detector
is not linear, with the peak area not increasing proportionally with the
weight of sulfide injected.  The concentration range of sulfides in dilute
exhaust which would fall in this linear range with the current sampling
technique is 0.1 to 25 ppb.  If the concentration of sulfides in the dilute
exhaust exceeds a concentration of 25 ppb, a lower sample flow rate will
have to be used in order to collect  a smaller amount of the sulfides.

     Hydrogen sulfide, sulfur dioxide, thiophene, methyl mercaptan, and
ethyl mercaptan are sulfur containing compounds that could interfere with
the organic sulfide procedure.  Sulfur dioxide is present in exhaust at
levels which would obscure all other compounds in the GC procedure  if it is
not removed before it enters the Tenax-GC trap.  A 5 percent sodium bicar-
bonate trap preceding the Tenax-GC trap effectively removes sulfur  dioxide
from the exhaust without affecting the concentration of the organic sulfides.
Hydrogen sulfide at levels of less than one ppm do not pose a problem with
the procedure as no breakthrough of hydrogen sulfide into the Tenax-GC trap
is detected by the GC-FPD.  However, if hydrogen sulfide is present at
concentrations of 4 ppm or greater, some hydrogem sulfide is collected on
the Tenax-GC trap and is detected by the GC-FPD.  If this higher concentra-
tion of hydrogen sulfide is present the GC parameters can be modified to
prevent hydrogen sulfide from interfering with the analysis of carbonyl
sulfide.  An oven temperature program which consists of holding the oven
temperature at 0°C for 4 minutes and then programming to 140°C at 8°/minute
will separate hydrogen sulfide and carbonyl sulfide by nearly two minutes.
This program does extend the analysis time for 25 minutes to 45 minutes if
the time it takes to recool the GC oven to 0°C is included.  Methyl and
ethyl mercaptan have yet to be detected in exhaust.  If present, the GC
operating conditions will separate these compounds from the sulfides of
interest.  With the present operating conditions, thiophene has a retention
time that differs from ethyl sulfide by only seconds.  Thiophene and ethyl
sulfide have not been effectively spearated by changing the GC operating
conditions and therefore, thiophene must be included as a possible  source
of error in the analysis for ethyl sulfide.

QUALIFICATION EXPERIMENTS

     Qualification experiments were carried out using a Mercedes 240D.  Hot
FTP (23minute test)  driving cycles were followed to generate exhaust for the
vehicle baseline emissions and for the tunnel injection plus vehicle experi-
ments.  An aluminum cylinder containing 4-8 ppm of each of the organic sul-
fides in balance nitrogen was used as the source for the organic sulfides.
The cylinder was named by comparing GC peak areas with the GC peal;  areas of
the organic sulfides generated by the permeation system.  The flow  of organic
sulfides into the tunnel was regulated to give a concentration of 5-10 ppb
of each of the organic sulfides in the dilution tunnel.  Injections of the


                                      154

-------
en
en
            360 f—
           320  -
                       100     200       300     400     500     600      700


                                                             GC Peak Area X0.01
                                                                                800
                                                                                        900
                                                                                                1000
                                                                                                        1100
1200
                                         Figure  46.  Carbonyl  sulfide linearity plot,

-------
cn
           o
           at
           I
              360  -
              320  -
              280  -
              240  -
200 ~
              160  -
120 _
              80
              40
                                                                           I
                                                                                                    I
                                                                                                             I
                         100      200     300      400      500      600     700


                                                                GC Peak Area X0.01
                                                                                   800
                                                                                           900
                                                                                                   1000
                                                                                                           1100
                                                                                                                   1200
                                          Figure 47.   Methyl sulfide linearity plot.

-------
              360 I-
en
                         100       200     300      400      500      600      700
                                                               GC Peak Area X0.01
                                                                                  800
                                                                                          900
                                                                                                  1000
                                                                                                          1100
                                                                                                                  1200
                                          Figure  48.  Ethyl  sulfide  linearity plot%

-------
Ol
CO
                                  100
200              300

         GC Peak Area X0.01
                                                                                 400
                                                                                                500
                                                                                                                600
                                       Figure  49.  Methyl disulfide linearity plot .

-------
organic sulfides into the tunnel without exhaust  gave  recoveries that varied
from approximately 93 percent for methyl disulfide  to  115 percent for  ethyl
sulfide (Table 52) .  An interfering peak  in the GC analysis for methyl
sulfide voided the tunnel recovery  experiments  for  this  compound.  The per-
cent deviation of the recovery percentages  ranged from 23-60 percent.  This
value is higher than the expected 25% due to the  trap-to-trap variations
found in the validation experiments.  The recovery  of  the organic sulfides
with real exhaust varied from 7 percent for ethyl sulfide to 57 percent for
carbonyl sulfide (Table 53).  The baseline  emissions of  carbonyl sulfide and
inethyl sulfide were erratic and of  equal magnitude  to  the carbonyl sulfide
and methyl sulfide injected into the tunnel.  This  variation of carbonyl
sulfide and methyl sulfide from the vehicle, along with  tunnel memory  for
carbonyl sulfide and methyl sulfide and trap-to-trap variations, made  the
percent recovery calculations very  difficult and  thus  gave  the resulting 37
and 55 standard deviations.  Baseline emissions for ethyl sulfide, and methyl
disulfide were insignificant and did not affect the recovery experiment.

     There is  little loss  of the organic  sulfides in  the CVS tunnel  with-
out exhaust, however a 40 to 90 percent loss with exhaust in the CVS tunnel
can be expected.  These losses must be taken into account in determining
organic sulfide  concentrations when using this  procedure.

RESULTS AND DISCUSSION

     The concentration of organic sulfides  can  be determined by: 1)  trapping
the sulfides in a Tenax-GC trap at  -76°C, 2) thermally desorbing the sulfides
from the Tenax GC trap into the GC  injection system,  3)  injecting the organic
sulfides into the analytical GC system, 4)  analyzing the organic sulfides
with a gas chromatograph equipped with a  flame  photometric  detector, and
5) quantifying the results by comparison with standards  generated by a per-
meation system.  The organic sulfides are effectively  caught in the  Tenax-
GC trap at a flow rate of 130 mJl/minute.  The procedure  has a minimum  detec-
tion limit of approximately 0.2 ppb.

     The accuracy of the procedure  in the 0.2 to  25 ppb  range is on  the order
of 25 percent due to trap-to-trap repeatability.   The  FPD gives a linear re-
sponse for the organic sulfides in  the 0.2  to 25  ppb  range. If the  concen-
trations of the organic sulfides exceed this range in  dilute exhaust,  a lower
sampling flow rate (less than 130 m£/minute) must be  used to keep the  detec-
tor response in the linear range of the detector.

     Sulfur dioxide, hydrogen sulfide, and  thiophene  are possible interfer-
ences in the procedure.  Sulfur dioxide is  removed by  the use of a sodium
bicarbonate trap, hydrogen sulfide  can be  separated in the  GC system by
changing oven parameters; however,  thiophene remains  an  interference to the
procedure.  The ethyl sulfide concentration is  affected  by  this interference.
This is a significant loss of the organic sulfides in  the CVS tunnel with
exhaust.  These losses must be taken into account in  determining the concen-
tration of the organic sulfides when using  this procedure.
     Overall, the organic sulfide procedure should provide  a relatively ac-
curate method for determining the concentrations  of the  organic sulfides in
dilute exhaust, and its use is recommended  for  this project.

                                     159

-------
           TABLE 52.   PERCENT RECOVERIES OF THE ORGANIC SULFIDES
                         FROM THE CVS TUNNEL ONLY
     Carbonyl Sulfide (8 ppb)
         Percent Recovery	

              83
             127
              82
             132
              70
                              Methyl Sulfide  (6 ppb)
                                 Percent Recovery

                                 Results voided
                                 due to inter-
                                 fering peak in
                                 the GC analysis
Average
% Recovery

Standard
Deviation
 99%
 28%
      Ethyl Sulfide (5 ppb)
        Percent Recovery

              98
             139
             120
             132
              85
Average
% Recovery

Standard
Deviation
115%
 23%
                            Methyl Disulfide  (10 ppb)
                            	Percent Recovery	

                                       31
                                      100
                                       43
                                      100
                                      190
Average
% Recovery

Standard
Deviation
                                       93%
                                                    60%
                                   160

-------
          TABLE 53.  PERCENT RECOVERIES OP THE ORGANIC SULFIDES
                    FROM THE CVS TUNNEL AND EXHAUST
     Carbonyl Sulfide (8 ppb)
         Percent Recovery	

              20
              42
             119
              52
              54
                              Methyl Sulfide (6 ppb)
                                 Percent Recovery

                                       12
                                       27
                                      144
                                       18
                                       34
Average
% Recovery
57%
Average
% Recovery
47%
Standard
Deviation
37%
Standard
Deviation
55%
      Ethyl Sulfide (5 ppb)
        Percent Recovery

               2
               5
              15
               5
              10
                            Methyl Disulfide (10 ppb)
                            	Percent Recovery

                                        5
                                        5
                                       25
                                        1
                                       17
Average
% Recovery

Standard
Deviation
 7%
 5%
Average
% Recovery

Standard
Deviation
11%
                                       10%
                                     161

-------
                                 SECTION 11

                              PHENOL PROCEDURE
LITERATURE SEARCH

     Phenols are compounds of the general formula ArOH, where Ar is phenyl
or substituted phenyl.  Phenols differ from alcohols in having the hydroxyl
group, -OH, attached directly to an aromatic ring.  Ring substitution by
alkyl, hydroxyl or carbonyl groups creates the variety of different phenols
possible in automotive exhaust.  Phenols generally have high vapor pressures,
are colorless and, except for phenol, are insoluble in water.  Some of the
physical properties of phenols possible in exhaust are shown in Table 54
below,

        TABLE 54.  PHYSICAL PROPERTIES OF PHENOLS POSSIBLE IN EXHAUST
       Phenol
Phenol
Salicylaldehyde
m-cresol
p-cresol
2,3-xylenol
3,5-xylenol
p-ethyIphenol
2-i sopropyIphenol
2,4„6-trimethyIphenol
2,3,5-trimethyIphenol
2,3,5,6,-tetramethyIphenol
Internal standard-
o-chloiqphenol
Molecular
 Weight

   94.11
  122.13
  108.15
  108.15
  122,17
  122.17
  122.17
  136.20
  136.20
  136.20
  150.22

  128.56
 Boiling
Point,PC

   182
   197
   202
   202
   218
   220
   219
   213
   221
   230
   247

   175
Freezing
Point,°C

   43
   -7
   12
   35
   75
   68
   47
   15
   72
   92
  118

    9
                                                                Density,
1.0722
1.1674
1.0336
1.0178

0.9680

1.012
1.2634
The slightly acidic nature of phenols (Ka =< 10~  ) makes them soluble in
aqueous hydroxides yet not acidic enough to be soluble in a bicarbonate so-
lution.  This property allows phenols to be separated from non-acidic com-
pounds by collection in base and from organic acids by their insolubility
in bicarbonate.  The acid base equilibrium that occurs is shown below.
                                    162

-------
     The purpose of measuring phenols in exhaust is to determine if they are
present in sufficient quantities to cause health problems.  A number of pro-
cedures have been published that are used for determining concentrations of
phenols.  These include colorimetric or spectrophotometric methods, gas
chromatography, liquid chromatography and derivatization with subsequent
analysis by gas chromatography.  Colorimetric and spectrophotometric methods
of phenol analysis (152-154) are primarily used for total phenol measurement.
This method is unacceptable because individual phenol concentrations are
desired.  One liquid chromatography procedure investigated includes the
formation of fluorescent dansyl phenol derivatives which are subsequently
analyzed by a liquid chromatograph  (LC) equipped with a fluorescence detec-
tor  (155).  The procedure is too time consuming to warrant its use.  Another
liquid chromatograph technique is difficult to set up, is very involved
chemically and suffers from interferences in one of the reagents (156).
Several procedures were available in which derivatives of phenols were pre-
pared for analysis on a gas chromatograph  (GC) .  In one method, phenols were
alkylated over an aluminum phosphate catalyst, acetylated and analyzed on a
GC  (157) .  This procedure is very involved and recoveries of phenol are not
high.  Several methods in which ester  (158-160) and ether  (161-163) deri-
vatives of phenols are produced were found in the literature search.  The
production of ether derivatives of phenols and analysis by GC seemed to be
a promising method for determining the concentrations of individual phenols.
A number of GC methods not involving derivatization were also studied.  Some
of  these listed a variety of columns and GC instrument parameters for phenols
analysis  (164-173).  Phenols can be sampled from exhaust in several ways.
Activated carbon filters have been used to absorb phenol from aqueous samples
 (174) and from air (175) .  However, a more suitable sampling procedure for
'dilute exhaust involves collection in a hydroxide solution in impingers.
Several authors have suggested this means of removing phenols from exhaust.
Collection of phenols in aqueous hydroxide is usually followed by wet chemi-
cal workup and analysis by GC.  Aqueous phenol samples are treated with a
variety of steps including acidification, extraction with an organic solvent,
distillation and extractions to remove impurities  (176-179).

     The procedures that appeared to be the most promising are those using
a GC for phenols analysis.  Samples can be collected in impingers containing
aqueous KOH and workup can be accomplished by  forming ether derivatives or by
extracting with ether  (176) .

PROCEDURAL DEVELOPMENT

     The procedure chosen for the collection and analysis of phenols required
a considerable amount of procedural development.  Extraction, analytical and
sampling parameters needed to be determined prior to exhaust sample pro-
cessing.

     The first factor investigated, extraction efficiency, was found to de-
pend on a number of variables.  Type of solvent, number of  solvent extrac-
tions, pH of aqueous sample and method of  solvent evaporation all affected
the extraction efficiency.  Two sets of spiked phenol samples were extracted
with two solvents, methylene chloride and ethyl ether.  Between  one and  five


                                     163

-------
consecutive extractions were performed and each set of samples using each
solvent.  The amount and percent of phenol recovered by these extractions
is shown in Table 55 and 56.  These figures indicate that of the phenol
recovered, most or all of it is recovered in the first two solvent extrac-
tions.  However, four times as much phenol is captured in the second extrac-
tion with methylene chloride (8.0 percent) as is captured in the second
extraction with ether (1.9 percent).  The average recoveries calculated for
the two solvents are probably low due to the fact that the averages include
cases when only one solvent extraction was performed.  Taking into consider-
ation the large difference in extraction efficiencies between the two sol-
vents (67.6 percent with ethyl ether and 49.9 percent with methylene chlo-
ride) , ether was chosen as the organic solvent for extracting exhaust samples.
It is possible that the slightly lower boiling point of ether compared to
methylene chloride  (34°C vs 40°C) allows it to be boiled off at a lower
temperature, thus preventing the evaporation of the lower boiling phenols.

     Another factor influencing extraction efficiency was investigated.  This
was the pH of the phenol spiked aqueous solution.  The extraction efficiency
was found to be unaffected by the pH of the solution when the pH was neutral
or acidic  (pH < 7) .  Table  57 lists the amount and percent phenol recovered
when the spiked aqueous solution was varied from a pH of one to seven.

            TABLE 57.  EXTRACTION EFFICIENCY AS A FUNCTION OF pH
                             OF AQUEOUS SOLUTION

              pH of         Phenol        Cone.       Percent
             Aqueous       Recovered      Spike       Phenol
             Solution        yg/m£        yg/m£      Recovered

                1             34            67         51.8
                2             36            67         53.8
                3             32            67         47.8
                4             28            67         41.8
                5             34            67         51.8
                6             44            67         65.7
                7             32            67         47.8

     The fourth factor affecting extraction efficiency that was studied was
the means of solvent removal and sample concentration.  The method producing
the highest phenol recoveries involved a two step process using a Kuderna
Danish concentrator heated by a steam bath (45°C) for initial volume reduc-
tion and a desiccating chamber modified for dry nitrogen flow for final con-
centration.  Several sample concentrating techniques were tested before it
was determined that erratic phenol recoveries occurred when samples were
dried solely by heating in a Kuderna concentrator.  Phenol recoveries for
several samples evaporated to 0r 1/2, 1, 2 and 5 m£ in the Kuderna concen-
trator are listed in Table 58.  The trend toward increasing phenol recovery
with larger final volumes in the Kuderna concentrator is apparent from the
data presented in Table 58.  Concentrating samples to a desired volume,
however, proved to be a difficult task using the Kuderna concentrator.  Due
to the tapered tip on the concentrator, the solvent level changed rapidly
when the volume decreased to 5 m£ and less.  It was necessary, therefore, to

                                     164

-------
                           TABLE 55.  EXTRACTIONS WITH METHYLENE CHLORIDE
Phenol Recovered, yg/m£
Extraction #
Sample
1
2
3
4
5
6
7
8
9
S! 10
11
12
13
14
15
16
17
18
19
20
21

1
16
24
26
24
8
23
17
36
41
37
24
_
_
—
_
-
-
_
-
_
-

2
2
2
3
-
-
-
-
-
-
-
-
9
24
37
-
-
-
—
-
-
-

345
0 - -
0 - -
0 - -
_
_
_
_
_
_
_
_ _ _
_ — _
_ _ _
_ _ _
21
21
33
24
25
32
33

Cone.
Spike
yg/m£
29
29
29
54
54
54
54
67
67
67
67
67
67
67
67
67
67
67
67
67
67



1
55.2
82.8
88.2
44.4
14.8
42.6
31.5
53.7
61.2
55.2
35.8
_
_
-
_
-
-
_
-
-
-

Percent Phenol Recovered Total
Extraction # Percent
2345 Recovered
6.9 0 - - 62.1
6.9 0 - - 89.7
10.3 0 - - 98.5
- - - - 44.4
14. 8a
42.6
- - - - 31.5
- - - - 53.7
- - - - 61.2
- - - - 55.2
- - - - 35.8
13. 4b - - - 13. 4a
35. 8b - 35.8
55. 2b - 55.2
31. 3b - - 31.3
31. 3b - - 31.3
49. 2b - 49.2
35. 8b - 35.8
37. 3b - 37.3
47. 8b 47.8
49. 2b 49.2
Average 49.9
Samples taken inadvertently to near dryness.
All extractions combined to give a total concentration.

-------
                              TABLE 56.  EXTRACTIONS WITH ETHER

Phenol
Recovered, yg/m£
Extraction #
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

1
38
23
6
40
31
38
42
49
45
52
43
18
29
26
30

2
1
4
0
0
1
0
1
-
-
-
-
-
-
-
-

3 4
0 0
6 3
0
-
-
-
-
-
-
-
-
-
-
-
-

Cone.
Spike
yig/m£
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54

Percent Phenol
Recovered
Extraction #
1
70.4
42.6
11.1
74.1
57.4
70.4
77.8
90.7
83.3
96.3
79.6
33.3
53.7
48.2
55.6

2
1.9
7.4
0
0
1.9
0
1.9
-
-
-
-
-
-
-
-

3
0
11.1
0
-
-
-
-
-
-
-
-
-
-
-
-

4
0
5.6
0
-
-
-
-
_ 'I
-
-
-
-
-
-
-
Average
Total
Percent
Recovered
72.3
50. Oa
11. lb
74.1
59.3
70.4
79.7
90.7
' 83.3
96.3
79.6
33.3
53.7
48.2
55.6
67.6
a Phenol contamination in all four extractions; total percent recovered represents
  extractions 1 and 2.
  Sample inadvertently taken to near dryness.

-------
       TABLE  58.   EFFECT OF REDUCING SAMPLE VOLUME BY KUDERNA DANISH
                       CONCENTRATOR ON PHENOL RECOVERY
         Evaporative
         Volume,  m£_

               5
               2
               1
              1/2
               0
Final
Volume
   5
   2
   2
   2
   2
 Phenol
Recovered
   15
   32
   23
   20
   14
Cone.
Spike
yg/m£

 21
 52
 52
 52
 52
 Percent
 Phenol
Recovered

  71.4
  61.5
  44.2
  38.5
  26.9
find another method of sample concentration that did not require constant
attention.   The second drying method attempted involved the use of the
Kuderna concentrator and a tray of heated sand equipped with a dry nitrogen
outlet.   The samples were concentrated to 5 m£ in the Kuderna concentrator,
transferred to a 10 m£ beaker and then further concentrated with a stream of
dry nitrogen (while being gently heated with the sand) .  This method was  un-
successful  due to water condensation on the beaker and nitrogen blowing sand
into the beaker.   The warm sand tray was abandoned as a means of drying
phenol samples in favor of a desiccating chamber modified for the flow of
dry nitrogen.  The samples were concentrated to 5 m£ in the Kuderna concen-
trator and  transferred to 10 m& beakers as was done previously.  The samples
were then concentrated to approximately 1 m£ in the desiccating chamber by
directing dry nitrogen  into the beakers with a gas manifold.  Water conden-
sation was  no longer a problem because the molecular sieve/silica gel absor-
bant in the chamber absorbed any moisture that was present.  The drying pro-
cess could  also be easily observed through the glass window and stopped when
necessary.   This last procedure was the one adopted for the concentration of
extracted phenol samples in ether due to its simplicity and lack of inter-
ferences.

     A second parameter (in addition to extraction efficiency) affecting  the
workup of phenol exhaust samples was investigated.  This factor was chemical
interferences to phenol recovery.  The source of interferences could be
contaminants in the various reagents used in sample collection or extraction
or interfering  exhaust compounds trapped in hydroxide solution along with
the phenols.  Several blank extractions were performed with methylene chlo-
ride and with ether using all solutions that would normally be used for
exhaust sample extraction.  None of the samples produced measurable levels
of phenols.   Possible interfering compounds in exhaust that may be absorbed
into the scrubber solution, 1 N KOH, are neutral hydrocarbons and organic
acids.  A set of tests were performed in which 1 N KOH samples spiked with
diesel fuel were extracted in several ways.  In the first experiment 1 y£
of diesel fuel was added to acidified 1 N KOH spiked with phenol and the
resulting sample was extracted and analyzed.  The second test was conducted
similarly to the first except that a cyclohexane extraction was performed on
the basic solution to remove neutral hydrocarbons before acidification, ex-
traction and analysis.  In the third extraction an acidified 1 N KOH solution
was spiked  with 1 yfc of diesel fuel (no phenol) , extracted and analyzed.
                                     167

-------
Several regular extractions were also performed on phenol spiked 1 N KOH
samples.  The results, shown in Table 59 indicate that diesel fuel does not
interfere with phenol recovery when present by itself.  However, when both
phenol and diesel fuel were present, an approximate 16 percent phenol loss
occurred.  The loss increased to 30 percent when a cyclohexane extraction
was performed to remove diesel fuel.  Additional extractions were performed

           TABLE 59.  EFFECT OF DIESEL FUEL ON RECOVERY OF PHENOL

                                  Phenol      Phenol     Percent
                                 Recovered    Added      Phenol
Sample Extracted (in 1 N KOH)      \ig/mH      yig/m£     Recovered

1.  Phenol + diesel fuel            21          30        70.0
2.  Phenol + diesel fuel +
    cyclohexane extraction          16          29        55.2
3.  Diesel fuel                      0           0        	
4.  Phenol                          24          28        85.7

on an actual exhaust sample and on phenol spiked 1 N KOH samples to determine
the validity of the data obtained for Table 59.  One half of an aqueous ex-
haust sample was acidified, extracted and analyzed.  The remaining half of
sample was first extracted with ethyl ether to remove netural hydrocarbons.
Then the ether was extracted with 0.5 N NaOH to recover any phenol extracted
into the ether.  The aqueous portions were combined, acidified, extracted
and analyzed the same as the first half.  The same amount of phenol was re-
covered from each half of the sample.  Also, neither of the sample halves
contained compounds that could interfere with the GC analysis of the phenols.
See Table 60.  Apparently, either no neutral hydrocarbons survive the normal
extraction process or else these compounds are eluted under the solvent peak
during GC analysis of the phenol sample.  The two 1 N KOH samples spiked
with phenol that were extracted for neutral hydrocarbon removal had an aver-
age phenol recovery 70 percent less than samples spiked and extracted nor-
mally.  The data obtained from the experiments conducted to determine neutral
hydrocarbon interference produced conflicting results.  However, since the
exhaust sample showed no evidence of interference from such compounds, it
was decided not to incorporate a neutral compound removal step into the pro-
cedure.

     Interference, to phenol recovery or analysis due to the presence of
organic acids was also studied.  The modification to the procedure for re-
moval of organic acids included an additional NaHCO-j extraction of the ether
containing the phenols and a back extraction of the aqueous NaHCO-, layer
with ether.  Organic acids are more soluble in an alkaline aqueous solution
than in ether.  However, phenols being acidic, tend to be drawn into the
NaHCO3 layer along with the organic acids.  The back extraction with ether
was to recover phenols that may have been extracted into NaHCO3.  Two phenol
spiked 1 N KOH samples were extracted following the modified procedure to
determine the presence of organic acids.  The phenol level of the phenol
spiked 1 N KOH samples averaged 567 percent higher than samples extracted
normally.  Half of each exhaust sample was extracted normally and half was


                                     168

-------
           TABLE 60.   INTERFERENCES TO PHENOL RECOVERY OR ANALYSIS

                                 Phenol Recovered    Percent Phenol
Phenol Spiked 1 N KOH Samples         yg/m.g, __      Recovered3
Normal extractions                       3
        *..'
Hydrocarbon modified samples             1                 33

Organic acid modified samples-"          17                567
a
  Percent phenol recovered is relative to the samples extracted normally.


     Exhaust                            Phenol Recovered  Percent Difference
     Sample       Extraction Method          ug/m£	    Between Halves

       1        1/2 normal                    1
                1/2 HC modified               1                  °

       2        1/2 normal                   35
                1/2 org. acid modified       35                  °

       3        1/2 normal                   41
                1/2 org. acid modified       35


extracted with the modification to the procedure.  The concentration of phe-
nol recovered from each sample half of the first exhaust sample was the same.
The second exhaust sample did not agree as closely as the first, though, pro-
bably due to an error in the extraction process.  The sample half extracted
for removal of organic acids yielded 15 percent less phenol than the sample
half extracted normally.  Since neither exhaust sample showed any evidence '
of interference from organic acids, the procedure was not modified for re-
moval of organic acids.  The results of the organic acid interference tests
are shown in Table 60.

     The next set of parameters that needed to be determined were those
governing the analytical portion of the phenol procedure.  The instrument
parameters for the gas chormatograph (GC) and an analytical column for se-
paration of phenols needed to be selected.  Also, phenols recovered from
exhaust needed to be identified and the response factors calculated.  Sev-
eral different columns were installed in a Perkin-Elmer 3920 GC equipped
with a flame ionization detector (FID).  These included an  SE-30 WCOT glass
capillary column, a 10 percent OV-101 on 100/120 mesh Gas-Chrom Q Teflon
column, a 20 percent DECS on 80/100 mesh Chromosorb W-HP Teflon column and
a Teflon column packed with 10 percent OS 138/H3P04/SP-1200 on 100/120 mesh
Chromosorb W AW.  The last column packed with 10 percent OS 138 provided
the best separation of phenols of all columns tested.  A variety of tem-
perature programming sequences were experimented with on the GC.  The most
efficient separation of solvent peak from phenols plus an  analysis time of


                                     169

-------
less than an hour were obtained with a temperature program of 4°/min  from
70°C to 170°C.  The temperature is initially held isothermally for two minutes
at 70°C.  The injector and interface temperature are maintained at 200°C.

     Phenols in exhaust samples were identified by comparing the retention
times to individual standards and standard blends.  The concentrations of
phenols in the blend used as the external standard need to be close to the
phenols concentrations found in exhuast due to the fact that both retention
times and response factors vary with concentration.  Response factors, which
correct for the different responses of each phenol to the FID, are calculated
from the concentration and counts of each phenol in the external standard
relative to o-chlorophenol (also in the external standard).  The concentra-
tions of phenols in exhaust samples are computed by comparing the area of
each phenol to the appropriate response factor and to the area and concen-
tration of the internal standard, o-chlorophenol.  Using an external and an
internal standard proved to be the easiest and most accurate method of cal-
culating phenol  concentrations in exhaust.

     The last set of parameters that needed to be determined as part of the
procedural development were those relating to the sampling of exhaust for
phenols.  The procedure chosen for phenols analysis required that phenols be
present in an aqueous solution at the start of the extraction process.  Since
phenols are acidic and therefore soluble in base, it was decided that dilute
exhaust would be bubbled through 1 N KOH in glass impingers.  The phenols
collected in this manner could be extracted directly.  The number of impin-
gers and the flowrate of dilute exhaust passing through the impingers that
would trap the most phenols needed to be determined.  Initially, experiments
were conducted with three tapered tip impingers connected in series.   Each
impinger contained 25 m£ of 1 N KOH chilled to ice bath temperature.   Exhaust
was pumped through the impingers at 4 £/min.  The samples thus obtained were
extracted and analyzed for phenols, however, no phenols were found.  In an
effort to trap more phenols dilute exhaust was passed through larger Green-
burg-Smith impingers at a higher flowrate (0.7-0.8 ft3/min).  Each of the
three impingers contained 200 m£ of 1 N KOH instead of 25 mH.  Measurable
levels of phenols were extracted from exhaust under the latter conditions.
Additional tests regarding the choice of sampling parameters is shown in the
Validation Experiments section.

VALIDATION EXPERIMENTS

     Several experiments were performed to show that the phenol procedure is
a valid method for processing exhaust samples containing phenols.  The sam-
pling parameters providing the best trapping efficiency were determined.
Dilute exhaust is allowed to flow at 0.7-0.8 ft3/min through two Greenburg-
Smith impingers in series.  Each impinger contains 200 mi of 1 N KOH chilled
to ice bath temperatures (0-5°C)..  From the results shown ir. Tables 55 and 56
it is obvious that no phenol is captured in bubblers two and three.  However,
several other phenols (salicylaldehyde, m-cresol, p-cresol, 2,3-xylenol,
3,5-xylenol, etc.)  are found in small quantities in the second impinger.  For
this reason, two bubblers are used to collect phenols.  Before pacsing through
the impingers the dilute exhaust flows through a heated sample line (375°F),


                                     170

-------
a Pallflex filter and another sample line heated to 175°F.  Tests were con-
ducted with and wirhout a filter.  The sample line without a filter was
heated to 175°F from the CVS to the impingers.  Results from the qualifica-
tion tests in  Tables 62 and 63 show that from 4% to 60% more phenol is re-
covered from filtered exhaust than from unfiltered exhaust.  For this reason
exhaust is filtered before sampling for phenols.

     The next  set of parameters that needed to be determined were those that
would give the highest recovery of phenols from the extraction process.  It
was found that two ether extractions of the contents of impinger one, and
one ether extraction of the contents of impinger two gave good recoveries.
Also, better results were obtained when the final drying step was done with
dry nitrogen instead of with heat.  Extraction efficiency of the phenols
procedure is approximately 68%.

     Two experiments were performed to validate the analytical protion of
the phenol procedure.  The first involved the analysis of phenol stnadards
in the concentration ranges expected in exhaust samples.  Calibration curves
were drawn from the data and they are shown in Figures 50-56.  The linearity
ranges of the internal standard and of the phenols found in exhaust vary
between 0-50 yg/m£ and 0-200 yg/m£.  The range for each phenol is listed in
Table 61 below.  The concentrations of phenol recovered from exhaust are well
within the linearity range of each phenol.

            TABLE 61.  LINEARITY RANGES OF INTERNAL STANDARD AND
                            OF PHENOLS IN EXHAUST

          	Phenol	   Linearity Range (yg/m£)

          o-chlorophenol                            0-120
          phenol                                    0-50
          salicylaldehyde                           0-120
          m-cresol and p-cresol                     0-80
          p-ethylphenol, 2-isopropylphenol,
               2,3-xylenol, 3,5-xylenol and
               2,4,6-trimethylphenol                0-200
          2,3,5-trimethylphenol                     0-100
          2,3,5,6,-tetramethylphenol                0-120

     Injection variability was studied as another validation test for the
phenols procedure.  A 12.2 yg/m£ phenol standard in methylene chloride was
injected five consecutive times.  The area of each injection is  shown in
Table 62.  The standard deviation is 107 and the percent variations is 2.20%
for the five injections.
                                     171

-------
  60
   50
   40
o
o
   30
u
   20
   10
                60          120         180


                        Concentration (ug/m£)
240
300
        Figure 500  Linearity of o-chlorophenol GC response,
                                172

-------
   60
   50
   40
o
o
 •
o

X

(8

-------
   60 ,-
   50
   40
o
o

o

X

(0

-------
   60.-
o
o
(0

0)
•s
0)
n<

u
                40
  80          120



Concentration (yg/m£)
160
                                                                 200
    Figure 53.  Linearity of m-cresol and p-cresol GC response,
                                175

-------
   140 t-
   130
   120
   110
   100
    90
g    80
o
 •
o

*    70
QJ
a

-------
o
o
 I
o

X

tfl
0)
0)
dl

u
c
   60,-
   50
   40
30
20
   10
                20           40           60


                       Concentration
                                                  80
100
   Figure 55.  Linearity of 2,3,5-trimethylphenol GC response.
                               177

-------
  60
  50
  40
o
o
 •
o
X
10
  30

-------
                TABLE  62.   INJECTION VARIABILITY OF PHENOL

                           Sample         Area
                             1
                             2
                             3
                             4
                             5

                           Average
                  Standard Deviation
                   Percent Variation
QUALIFICATION
     The phenol procedure was qualified by injection of an aqueous phenol
solution into  the exhaust of a Mercedes 240D diesel during three successive
FTPh driving cycles.   The percent recoveries from the tests represent the
amount of  phenol  that is expected to survive the trip through the dilution
tunnel to  the  sampling impingers.  The test sequence consisted of FTPh
driving cycles with ten minute soaks in between.  Base line phenol emission
levels were  measured for three consecutive FTPh driving cycles and during
three additional  FTPh driving cycles phenol was injected into the exhaust.
Two sets of  impingers sampled dilute exhaust during each test.  The sample
line leading to the first set was heated to 175°F and no filter was used.
The second sample line was heated to 375 °F up to a Pallflex filter and 175°F
from the filter to the second set of impingers.  The average results from
the three  baseline emission tests showed that filtered exhaust produced a
higher phenol  concentration (24 ug/m3) than unfiltered exhaust (12
The data is  found in Table 63 below.

               TABLE 63.  BASELINE PHENOL EMISSION LEVELS FROM
                            MERCEDES 240D DIESEL

                           yg/m3 Phenol	
              Test     Unfiltered     Filtered     Difference

               1           10            29            19

               2           10

               3           15            19             4

              Avg          12            24

The difference between filtered and unfiltered exhaust was also apparent in
the results from the injection of phenol into exhaust.  The data in Table 64
shows that the filtered exhaust yielded 60 percent and 17 percent more phenol
than unfiltered exhaust.  No phenol was recovered from the filtered line
second phenol injection.  This was probably due to sample loss during the

                                    179

-------
   TABLE 64.  PERCENT RECOVERIES FROM INJECTION OF PHENOL INTO EXHAUST OF
                            MERCEDES 240D DIESEL

                                   Unfiltered    Filtered    Difference

     First Phenol Injection            52.4        112.7        60.3

     Second Phenol Injection           57.7        	

     Third Phenol Injection            76.4         93.7        17.3

     Average                           62.2        103.2

extraction procedure.  A gradual trend towards increasing phenol recoveries
appears to occur with samples that flowed through the unfiltered sample line.
This may be due to phenols being initially absorbed onto particulate coating
the sample line.  The particulate removes phenol from the gas stream until
it is saturated.  Gradually less phenol is absorbed and therefore, more is
recovered in the impingers.  What appears to be a trend, however, may also
be the expected variability in recoveries.  The greatest difference in un-
filtered recoveries is 24 percent and in filtered recoveries it is 19 percent.
The average phenol recovery of unfiltered samples is 62.2 percent and the
average phenol recovery of filtered sample is 103.2 percent.  Assuming all
phenols in exhaust can be removed with similar efficiencies, quantitative
recoveries of phenols in exhaust diluted by the CVS can be expected.

     The injection of phenol into the exhaust of the Mercedes was accom-
plished by means of a Baird atomizer attached to an opening on the CVS tunnel.
An aqueous phenol solution (0.7 g/m£)was dripped into the funnel of the atom-
izer from a 50 m£ buret.  Air pressure applied through the side arm of the at-
omizer sprayed the phenol solution into the tunnel where it mixed with exhaust.
Any solution that was not dispersed into the tunnel was captured in an
Erlenmeyer flask containing the mister.  This remaining portion of phenol
solution was extracted and analyzed as usual.  The amount of phenol injected
was calculated by subtracting the micrograms of phenol in the remaining
phenol solution from the micrograms delivered from the buret.  Percent re-
coveries were computed by comparing the amount of phenol recovered to the
amount injected.

RESULTS AND CONCLUSIONS

     The method chosen for measuring phenols in dilute exhaust involves col-
lection in aqueous KDH, extraction with ether and analysis on a GC equipped
with a flame ionization detector.  Dilute exhaust is bubbled at 0.8 ft3/ndn
through two Greenburg-Smith impingers each containing 200 m£ of 1 N KOH
chilled to ice bath temperatures.  The exhaust is heated to 375°F and is
filtered through a Pallflex filter to remove particulate.  The phenol samples
are acidified, extracted two consecutive times with ethyl ether and concen-
trated.   The extracts from impingers one and two are combined, further con-
centrated and spiked with the internal standard, o-chlorophenol, before
analysis with the GC.  The temperature programming sequence starts with an

                                    180

-------
isothermal hold at 70°C for two minutes  followed by programming to 170°C
at 4°/min.  Total GC analysis time is about  30 minutes.  The injector and
interface temperatures are maintained at 200°c.  A Teflon column packed
with 10% OS 138/H3PC>4/SP-1200 on Qiromosorb  W AW is used for separating
phenols.  One microliter of the external standard and  1 yjl of each sample
is injected into the GC.  The data obtained  from the GC computer system is
used to calculate concentrations of phenols.

     The linear range of each phenol found in exhaust  and of the internal
standard was determined.  The concentrations of phenols fall well within the
linear range.  Should a sample be too concentrated it  can be diluted volu-
metrically to a level within the linear  range.

     Several factors contribute to the overall recovery of phenols from ex-
haust.  These include the stability of phenols traveling from automobile to
impingers and the trapping and extraction efficiency of phenols.  The results
from the qualification tests indicate that approximately 100 percent of
phenol injected into exhaust is recovered.   One hundred percent phenol is
also captured in two impingers connected in  series.  Extraction efficiency,
however, is only about 68 percent.  This low value is  probably due to losses
encountered in the drying process.  Injection variability of phenol into the
GC was only 2.2 percent for a series of  five injections.  Similar results
are expected for the other phenols found in  exhaust.

     Several methods for the determination of phenols  in automobile exhaust
were combined and adapted to the needs of this project.  The resulting pro-
cedure used to measure phenols is sensitive  to about 1 yg/m£.  The phenols
in  order of elution are phenol; salicylaldehyde; m-cresol and p-cresol;
p-ethylphenol, 2-isopropylphenol, 2,3-xylenol,  3,5-xylenol and 2,4,6-
trimethylphenol; 2,3,5-trimethylphenol and 2,3,5,6-tetramethylphenol.  Over-
all this procedure should provide a relatively accurate method for determin-
ing the  concentrations of the phenols in dilute exhaust, and its use is re-
commended for this project.
                                     181

-------
                                 SECTION 12

                        THE QUALIFICATION EXPERIMENT


     Qualification experiments were carried out to determine what fraction of
the unregulated pollutants entering the dilution tunnel could be  recovered at
the sampling point.  A constant flow of each unregulated pollutant was
injected from a pressurized cylinder into the dilution tunnel-CVS system at
the point raw exhaust normally enters the dilution tunnel  (Figure 57).   The
CVS diluted samples were extracted from the dilution tunnel-CVS system with
a multiport sampling probe at a point after the orifice plate on  the tunnel
and before the CVS system.  All qualified unregulated pollutants  were sampled
at this point except for nitrous oxide which was taken as  a bag sample at the
CVS  (Figure 57 ).

     Experiments were carried out with and without diesel  exhaust present in
the dilution tunnel-CVS system.  A Mercedes 240D driving over a hot FTP  (23-
minute test) driving cycle was used to generate diesel exhaust for the
experiments.  Baseline emission levels of each pollutant from the Mercedes
240D were measured in order to correct recovery values for pollutants present
in the exhaust.

     The gaseous unregulated pollutants were injected into the dilution
tunnel by the system shown in Figures 58 and 59.  The pollutant passed
through a needle valve to regulate flow, a flowmeter to monitor flowrate,
and a dry gas meter to measure the injected volume of pollutant,  before
entering the dilution tunnel.  A thermocouple was used to  monitor the
temperature of the injected gas, and a magnehelic gauge was used  to monitor
the pressure of gas passing through the injection system.  This pressure was
positive and generally recorded 0-2" of water.  The phenols were  injected
into the dilution as a water solution using the modified mist generator  shown
in Figure 60.  The test sequence developed to determine pollutant recovery
consistrd of a 23-minute continuous sampling period  (pollutant injected  with
or without exhaust present) followed by a 10-minute soak period with the CVS
off  (no pollutant injected).  During this time, impingers, bags or traps
were changed to collect the next sample.  After the soak period the test
sequence was repeated until three to four sampling periods were completed.
During each sampling period, three replicate impinger samples  (aldehydes,
total cyanide,  organic amines, sulfur dioxide, ammonia, hydrogen  sulfide—
two for phenols), or two trap samples (organic sulfides) or one bag sample
(nitrous oxide) were taken.

     Nominal injected pollutant flows into the tunnel were 0.35 cu ft/min
while nominal CVS flows were 300 cu ft/min.  This gave an  approximate 850
to 1 dilution.   Percent recoveries were determined by analyzing the recovered
                                      182

-------
         Filter
          Box
           I
                              Dilution Tunnel
                             16" long x 18" diameter
                                                                           orifice
                                                                           plate
oo
co
¥
                                                                                     CVS
                                                                                   System
                                                                      4"  diameter
         Unregulated pollutant
         injected hero with or
         without raw exhaust
                                                                  Sample Probe
                                                                for  collection of
                                                              Aldehydes & Ketones
                                                              Total Cyanide
                                                              Organic Amines
                                                              Sulfur Dioxide
                                                              Hydrogen Sulfide
                                                              Ammonia
                                                              Organic Sulf.i.des and
                                                              Phenols Samnlos
Bag sample
   for
Nitrous Oxide
                    Figure 57»  Dilution tunnel-CVS system used in  qualification experiments.

-------
CO
-P.
                          Gas Temperature
                          Digital Readout
                               Ol8]0|
                    Regulating
                      Valve
               Cylinder
                Source
                  of
               Pollutant
                                      Magnehelic
                                      (pressure)
1|213|4|5
                                               Gas Volume
                                             Digital  Readout
                                                                        Dilution
                                                                         Tunnel
                                                                                           -Filter Box
            Figure 58.  Apparatus for injection of pollutant  into dilution tunnel without exhaust.

-------
                            Gas Temperature
                            Digital Readout
                                0|8|D|
                                Flowmeter
oo
01
                     Regulating
                       Valve
                                                  Dry
                                                  Gas
                                                 Meter
o
                                       Magnehelic
                                       (pressure)
                Cylinder
                 Source
                   of
               Pollutant
                 Raw
                 Vehicle  ——]
                 Exhaust
       12  345
                                               Gas Volume
                                             Digital Readout
                                                                          Dilution
                                                                           Tunnel
                                                     Filter  Box
               Figure 59.  Apparatus for injection  of pollutant into dilution tunnel with exhaust.

-------
                                    Phenol/water
                                     solution in
               Compressed
                air in
oo
  Into
Dilution
 Tunnel
                                                             Modified 2000 mi
                                                             Erlenmeyer Flask
                                      Figure 60.  Modified mist generator.

-------
diluted sample,  multiplying by the CVS dilution and dividing by the actual
injected pollutant concentration.
                                       187

-------
                                 SECTION 13

                           RESULTS AND CONCLUSIONS
     To determine the suitability of the analytical procedures initially
selected for dilute exhaust analysis, validation and qualification experi-
ments were carried out.  The validation experiments determined if the sam-
pling and instrument parameters were appropriate for the quantitative analysis
of dilute exhaust.  The qualification experiments determined if the compounds
of interest could be quantitatively recovered from the CVS tunnel with and
without the presence of exhaust in the tunnel.  The analytical procedures
to be used in this project are listed in Table 65 along with methods of
sampling and analysis.  Table 65 also lists the validation and qualification
experiments that were carried out.

     The sampling parameters for all procedures were found to be adequate
for the collection of each of the unregulated emissions.  All samples, with
the exception of the organic sulfides  and hydrogen sulfide are stable for
several days and can be stored and rerun within hours after sampling to pre-
vent loss of sample integrity.  All instruments demonstrate linearity of re-
sponse for expected concentration ranges (sample concentrations above the
linear range must be diluted to concentrations that fall within the linear
range of the instrument) .  The organic sulfides must be monitored carefully
as traps containing over 200 ng of sample fall beyond the linear range of
the FPD.  The sample flow rate can be lowered to prevent overloading the
Tenax trap.  Test-to-test repeatabilities for all procedures are documented
in this report.  In most cases, repeatability is difficult to obtain at the
lower concentrations, while the repeatability at high concentrations is
easily obtained.  Interferences were checked and documented for each proce-
dure.  Phthalates were found to interfere with the aldehyde and ketone pro-
cedure and may cause erroneous results for crotonaldehyde.  In the hydrogen
sulfide procedure, sulfur dioxide decreases the apparent hydrogen sulfide
concentration, and its presence or absence must be recorded.  The other
procedures have interferences that can be avoided if care is taken.

     Qualification experiments were carried out on the aldehyde and ketone,
organic amine, sulfur dioxide, nitrous oxide, hydrogen sulfide, total cyanide,
organic sulfide, ammonia and phenol procedures to determine the recovery of
known amounts of each pollutant from the CVS tunnel with and without exhaust
(phenols CVS tunnel with exhaust only).  Aldehydes and ketones, sulfur di-
oxide, nitrous oxide, total cyanide and phenols can be recovered quantita-
tively from the CVS tunnel with and without  (not done for phenols) exhaust.
There is a 10 percent loss of hydrogen sulfide with and without exhaust
present.  The organic amines, ammonia, and the organic sulfides experience
significant losses in the CVS tunnel with and without exhaust.  These losses
must be taken into account when determining the concentration of these com-

                                    188

-------
                          TABLE 65.  ANALYTICAL PROCEDURES FOR EMISSIONS CHARACTERIZATION
00
vo
Compounds
Aldehdyes and Ketones
Organic Amines
Sulfur Dioxide
Nitrous Oxide
Individual Hydrocarbons
Hydrogen Sulfide
Hydrogen Cyanide + Cyanogen
Carbonyl Sulfide + Organic Sulfides
Ammonia
Sulfate
DMNA
Phenols
BaP
Sampling
Impinge rs
Impinge rs
Impingers
Bags
Bags
Impingers
Impingers
Traps
Impingers
Filters
Traps
Impingers
Filters
Analysis
DNPH
GC-NPD
Ion Chrom.
GC-ECD
GC-FID
Meth. Blue
GC-ECD
GC-FPD
Ion Chrom.
BCA
GC-MS @ RTI
GC-FID
Fluorescence
Validation
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Not required
Not required
Yes
Not required
Qualification
Yes
Yes
Yes
Yes
Not required
Yes
Yes
Yes
Yes
Not required
Not required
Yes
Not required
                                                            @ EPA

-------
pounds in exhaust.

     The procedures discussed in this report are effective in collecting and
analyzing dilute exhaust samples and are recommended for use in this project.
                                    190

-------
REFERENCES
 1.  Dietzmann, H. E., Smith,  L.  R.,  Parness,  M. A., and Fanick,  E.  R.,
    "Analytical Procedures  for Characterizing Unregulated Pollutant
    Emissions from Motor Vehicles."   Report to Environmental Protection
    Agency under Contract No. 68-02-2497, February, 1979.

 2.  Muir, G. D.  (Ed.), Hazards in the Chemical Laboratory, The Chemical
    Society, London,  1977.

 3.  Stump, F., Oxygenated Compounds  in Automobile Exhaust-GAS CHROMATO-
    GRAPHIC PROCEDURE, MSERB-ESRL-EPA, Research Triangle Park, North
    Carolina.

 4.  Roberts, R. M., Giblert,  J.  C.,  et al; An Introduction to Modern
    Experimental Organic Chemistry;  Holt, Rinehart and Winston,  Inc.
    1974, pg. 470-471.

 5.  Weast, R. C.  (Ed.), Handbook of  Chemistry and Physics, 54 Edition,
    Division of the Chemical  Rubber  Co., 1973.

 6.  Sekerka, I., and  Lechner, J. F., Water Res., Vol. 10, pg. 479,  1976.

 7.  Valentour, J. C., Aggarwal,  V.,  and Sunshine, I., Anal. Chem.,  Vol. 46,
    pg.  924, 1974.

 8.  Dimitriades, B.,  and Seizenger,  D. E., Environ. Sci. Tech.,  Vol. 5,
    No.  3, pg. 223, 1971.

 9.  Papa, L. J., Dinsel, D. L.,  et al; J. Gas Chromatog., Vol. 6,  pg. 270,
    1968.

10.  Klosterman, D. L., and  Sigsby, J. E., Environ. Sci. Tech., Vol. 1,
    pg.  309, 1967.

11.   (a)  Black, F. M., High, L. E., and Sigsby, J. E., J. Chromatog. Sci.,
    Vol.  14, pg. 257, May 1976.    (b) Communication between Mobile  Source
    Emissions Research Branch, ESRL-EPA, Research Triangle Park, and
    Southwest Research Institute.

12.  Braker, W., and Mossman,  A.  L.,  Matheson Gas Data Book, 5th Edition,
    East Rutherford,  N.Y.,  1971.

13.  Hoshika, Y., Anal. Chem., Vol. 48, pg. 1716, 1976.


                                     191

-------
14.  Hoshika, Y.,  J. Chromatog., Vol. 115, pg. 596, 1975.

15.  Sze, Y. L., and Borke, M. L., Anal. Chem., Vol. 35, pg. 240,  1963.

16.  O'Donnell, J. F., and Mann, C. K., Anal. Chem,, Vol. 36, pg. .2097, 1964.

17.  McCurdy, W. H., Jr., and Reiser, R. W., Anal. Chem., Vol. 38, pg. 795,
     1966.

18.  Smith, J. R. L., and Waddington, D. J., Anal. Chem., Vol. 40, pg. 522,
     1977.

19.  "Amine Analysis," Bulletin 737A, Supelco Co., Belefonte, Pa., 1973.

20.  "Penwalt 223 Amine Packing," Alltech Associates, Arlington Heights, 111.

21.  Hoffat, A. C., and Horning, E. C., Anal. Lett., Vol. 3, pg. 205, 1970.

22.  Clarke, D. D., Wilk, S., et al; J. Gas Chromatog., pg. 310, 1966.

23.  Hosier, A. R., Andre, C. E., et al; Environ. Sci. Tech., Vol. 7, pg.
     642, 1973.

24.  Burks, R. E., Jr., Baker, E. B., et al; J. Agr. Food Chem., Vol. 7,
     pg.  778, 1959.

25.  Dunn, S. R., Simenhoff, M. L., et al; Anal. Chem., Vol. 48, pg. 41,
     1976.

26.  Gruger, E. H., Jr., J. Agr. Food Chem., Vol. 20, pg. 781, 1972.

27.  Andre, C. E., and Hosier, A. R., Anal. Chem., Vol. 45, pg. 1971, 1973.

28.  Umbreit, G. R., Nygren, R. E., and Testa, A. J., J. Chromatog., Vol. 43,
     pg.  25, 1969.

29.  Keay, N. J., and Hardy, R., J. Scie. Food Agr., Vol. 23, pg. 9, 1972.

30.  Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.

31.  West, P. W., and Gaeke, G. C., Anal. Chem., Vol. 28, pg. 1816, 1956.

32.  Adams, D. F., Corn, M., et al; "Methods of Air Sampling and Analysis,"
     (Tentative Method of Analysis for Sulfur Dioxide Content of the
     Atmosphere (colorimetric), No. 42401-01-69T), pg. 447, Amer. Publ.
     Health Assn., Washington, D. C., 1972.

33.  "ASTM Standards - Water, Atmospheric Analysis," ASTM Designation
     D291470T, Part 23, Am. Soc. Test. Mat., Philadelphia, Pa., 1972.

34.  US EPA, Fed.  Register, Vol. 36, No. 158, 15492, 1971.


                                     192

-------
35.  Robinson, E. ,  and Robbins,  R. C., "Sources, Abundance, and Fate of
    Gaseous Atmospheric Pollutants," Final Report, SRI Project 6755
    Supplement,  Stanford Research Inst., Palo Alto, Calif., 1968.

36.  Stern, A. C.  (Ed.), Air Pollution, 3rd Edition, Academic Press  Inc
    N.Y.,  1976,  Vol.  3, pg. 214.

37.  Zurlo, N.,  and Griffini, A. M., Med. Lav.., Vol. 53, pg. 330, 1962.

38.  Scaringelli,  F. P., Elfers, L. , et al; Anal. Chem., Vol. 42, pg. 1818,
    1970.

39.  Pate,  J.  B.,  Ammons, B. E., et al; Anal. Chem., Vol. 37, pg. 942,  1965.

40.  Pate,  J.  B.,  Lodge, J. P.,  Jr., et al; Anal. Chem., Vol. 34, pg. 1660,
    1962.

41.  Scaringelli,  F. P., Saltzman, B. E., et al; Anal. Chem., Vol. 39,
    pg.  1709,  1967.

42.  Hartman-Leddon, Philadelphia, Pa.

43.   Stern, A.  C.  (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc.,
     N. Y., 1976,  Vol. 3, pg. 214.

44.   Urone, P.,  Evans, J. B., and Noyes, C. H., Anal. Chem., Vol. 37,
    pg.  1104,  1965.

45.   Bostrom,  C. E., Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.

46.   Attari, A., Igielski, T. P., et al; Anal. Chem., Vol. 42, pg. 1282,
     1970.

47.   Kawai, T.,  Netsu Kanri, Vol. 22, pg. 20, 1970.

48.   Thomas, M.  D., Ivie, J. 0., et al; Ind. Eng. Chem., Anal. Ed., Vol. 18,
     pg.  383,  1946.

49.  Yocum, J.  E., Richardson, R. L., et al; Proc. 49th Ann. Meet. Air
     Poll.  Contr.  Assn., Pittsburgh, Pa., 1956.

50.   Kuczynski,  E. R. , Environ.  Sci. Tech., Vol. 1, pg. 68, 1967.

51.   Shikiya,  J. M., and McPhee, R. D., 61st Annual Meeting, Paper No.  68-72,
     Air Poll.  Contr.  Assn., Pittsburgh, Pa., 1968.

52.   Jacobs, M.  B., The Chemical Analysis of Air Pollutants, Chapter 8,
     Wiley  (Interscience), New York, N.Y., 1960.

53.   Jacobs, M.  B., The Analytical Chemistry of Industrial Poisons, Hazards,
     and Solvents, Chapter 9, Wiley  (Interscience), New York, N.Y., 1949.
                                     193

-------
54.  Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.

55.  Stratman, H., Mikrochim. Acta., Vol. 6, pg. 688, 1954.

56.  Pate, J. B., Lodge, J. P., Jr., et al; Anal. Chim. Acta., Vol.  28,
     pg. 341, 1963.

57.  Lodge, J. P., Jr., Pate, J. B., et al; Amer. Ind. Hyg. Assn. J., Vol.
     24, pg. 380, 1963.

58.  Forrest, J., and Newman, L., Atmos. Environ., Vol. 7, pg. 561,  1973.

59.  Huygen, C., Anal. Chim. Acta., Vol. 28, pg. 349, 1963.

60.  Harding, J., and Schlein, B., "Nuclear Techniques in Environmental
     Pollution," SM-142a/0, IAEA, Vienna, Austria, 1971.

61.  Wilson, B. H., and McConnell, F. J., J. Soc. Chem. Ind., Vol. 53,
     pg. 385, 1934.

62.  Huey, N. A., J. Air Poll. Cont. Assn., Vol. 18, pg. 610, 1968.

63.  Thomas, F. W., and Davidson, C. M., J. Air Poll. Cont. Assn., Vol. 11,
     pg. 1, 1961.

64.  Stalker, W. W., Dickerson, R. C., et al; Amer. Ind. Hyg. Assn. J.,
     Vol. 24, pg. 68, 1963.

65.  Hickey, H. R., and Hendrickson, E. R., J. Air Poll. Cont. Assn.,
     Vol. 15, pg. 409, 1965.

66.  Harding, C. I., and Kelley, T. R., J. Air Poll. Cont. Assn., Vol. 17,
     pg. 545, 1967-

67.  Ikeda, T., J. Hyg. Chem., Vol. 13, pg. 41, 1967.

68.  Stevens, R. K., Mulik, J. D., et al; Anal. Chem., Vol. 43, pg. 827,
     1971.

69.  Stevens, R. K., O'Keeffe, A. E., et al; Environ. Sci. Tech., Vol. 3,
     pg. 652, 1969.

70.  Ronkainen, P., Denslow, J., et al; J. Chromatog. Sci., Vol. 11, pg. 384,
     1973.

71.  Pecsar, R. E., and Hartmann, C. H., J. Chromatog. Sci., Vol. 11, pg.
     492, 1973.

72.  Stevens, R. K., and O'Keeffe, A. E., Anal. Chem., Vol. 42, pg.  143a,
     1970.

73.  Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070, 1972.


                                     194

-------
74.   Bruner,  F.,  Ciccioli, P., et al; Anal. Chem., Vol. 47, pg. 141, 1975.

75.   DeSouza,  T.  L.  C.f Lane, D. C., et al; Anal. Chem., Vol. 47, pq  543
     1975.

76.   Bruner,  F. ,  Ciccioli, P., et al; J. Chromatog., Vol. 99, pg. 661,  1974.

77.   Bremner,  J.  M., and Banwart, W. L., Sulfur Inst. Journal, Vol. 10,
     pg.  6,  1974.

78.   Mulik,  J. D., Todd, G., et al;  "Ion Chromatography Determination of
     Atmospheric  Sulfur Dioxide," Symposium on Ion Chromatograph Analysis
     of Environmental Pollutants, EPA, Research Triangle Park, North
     Carolina, April, 1977.

79.   Dietzmann,  H. E., "Protocol to Characterize Gaseous Emissions as a
     Function of  Fuel and Additive Composition," Environmental Protection
     Technology Series under Contract 68-02-1275, September, 1975.

80.   Stern,  A. C. (Ed.), Air Pollution, Third Edition, Volume III, Academic
     Press,  New York, 1976, pg. 275.

81.   Private communication between R. B. Zweidinger and Frank Black of EPA,
     Research Triangle Park, and Harry Dietzmann of Southwest Research
     Institute.

82." Bethea, R.  M.,  J. Air Poll. Cont. Assoc., Vol. 23, pg. 710, 1973.

83.  Bamesberger, W. L., and Adams, D. F., Tappi, Vol. 52, pg 1302, 1969.

84.  "Hydrogen Sulfide in Air Analytical Method," H.L.S., Vol. 12, pg.  362,
     1975.

85.  Bamesberger, W. L., and Adams, D. F., Environ. Sci. Tech., Vol. 3,
     pg. 258, 1969.

86.  Gustafsson,  L., Talanta, Vol. 4, pg.  227, 1960.

87.  Moest,  R. R., Anal. Chem., Vol. 47, pg. 1204, 1975.

88.  Buck,  M., and Gies, H., Staub, Vol. 26, pg. 27,  1966.

89.  Jacobs, M. B.,  Braverman, M. M., and  Hochheiser, S., Anal. Chem.,
     Vol. 29, pg. 1349, 1957.

90.  Cave,  G. C.  B., Tappi, Vol. 46, pg. 1, 1963.

91.  Budd,  M. S., and Bewick, H. A., Anal. Chem., Vol. 24, pg. 1536, 1952.

92.  Bostrom, C., Air and Water Poll. Int. J., Vol. 10, pg. 435, 1966.

93.   Flamm,  D. L., and James, R. E., Environ. Sci. Tech., pg.  159,  1975.

                                     195

-------
 94.  Adams, D. F., Bamesberger, W. L., et al; J. Air Poll. Cont. Assn.,
      Vol. 18, pg. 145, 1968.

 95.  Stecher, P. G. (Ed.), The Merck Index, 8th Edition, Merck and Company,
      Inc., Rahway, N.  J., 1968, pg. 64.

 96.  Masterton, W. L., and Slowinski, E. J., Chemical Principles, 3rd
      Edition, W. B. Saunders Company, Philadelphia, Pa., 1973.

 97.  Carter, R. V., et al; Air Pollution Engineering Source Evaluation of
      Ammonia Oxidation Plant Number 10, Holston Army Ammunition Plant,
      Kingsport, Tennessee, Army Construction Engineering Research Lab.,
      Champaign, 111.,  1974, pg. 10.

 98.  Healy, T. V., and Pilbeam, A., Ammonia and Related Atmospheric
      Pollutants at Harwell Part 1.  Techniques for hourly sampling and
      analysis of pollutants, Applied Chemistry Division, Atomic Energy
      Research Establishment, Harwell, Berkshire, 1972.

 99.  Kothny, E. L., Intersociety Committee, Health Lab. Sci., Vol. 10,
      pg. 115, 1973.

100.  Okita, T., and Kanomori, S., Atmos. Environ., Vol. 5, pg. 622, 1971.

101.  Kramer, D. N., and Seen, J. M., Anal. Chem. , Vol. 44, pg. 395, 1972.

102.  Zitomer, F., and Laumbert, J. L., Anal. Chem., Vol. 34, pg. 1738, 1962.

103.  Blinn, R. C., and Gunther, F. A., Anal. Chem., Vol. 29, pg. 1882, 1957.

104.  Williams, D. D.,  and Miller, R. R., Anal. Chem., Vol. 34, pg. 225,
      1962.

105.  Gunther, F. A., Barkley, J. H., et al; Anal. Chem., Vol. 28, pg 1985,
      1956.

106.  Kolbezen, M. J.,  Eckert, J. W., et al; Anal. Chem., Vol. 36, pg. 593,
      1964.

107.  Pierson, R. H., Fletcher, A. A., et al; Anal. Chem., Vol. 28, pg. 1218,
      1956.

108.  Burns, E. A., The Analysis of Exhaust Gases Trapped from Ablating
      Nozzles, TRW Report No. 9840-6001-TU 000, Redondo Beach, Calif., 1964.

109.  Howell, J. H., and Boltz, D. F., Anal. Chem., Vol. 36, pg. 1799, 1964.

110.  Burns, E. A., Unpublished results, 1963.

111.  Jenkins, R. W., Jr., Cheek, C. H., et al; Anal. Chem., Vol. 38,
      pg. 38, pg. 1257, 1966.
                                      196

-------
 112.  Miettinen, J. K.,  and Virtanen, A. I., Ann. Acad. Aci. Fennicae,
      Ser. A, II, No.  41,  1951,  through Chem. Abst. ,  Vol. 46, pg. 11500,
      1952.

 113.  Arcand, G. M.,  and Swift,  E. H., Anal. Chem., Vol. 28, pg. 440, 1956.

 114.  Krivis, A. F.,  Supp, G.  R., et al; Anal. Chem., Vol. 35, pg. 2216,
      1963.

 115.  Gilbert,  T. R.,  and Clay,  A. M. , Anal. Chem., Vol. 45, pg. 1758, 1973.

 116.  Mertens,  J.,  Van den Winkle, P., et al; Bull. Soc. Chem. Belg. ,
      Vol.  83/n°  1-2/, pg. 26, 1973.

 117.  Furman, N.  H. (Ed.), Scott's Standard Methods of Chemical Analysis,
      5th Edition,  Van Nostrand, N. Y., 1958, Vol. 1, pg. 2336.

 .118.  Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
      Green, London,  1961, pg. 1092.

 119.  Steyermark, A., Quantitative Organic Microanalysis, Blakestone, N. Y.,
       1951, pg. 56.

 120.  Sadek, F. S., and Reilly,  C. N., Anal. Chem., Vol. 31, pg. 494, 1959.

 121.  Crane, F. E., and Smith, E. A., Chemist-Analyst, Vol.  49, pg. 38, 1960.

 122.  Crane, F. E., and Smith, E. A., Chemist-Analyst, Vol.  52, pg. 105,
       1963.
v
 123.   Hodges,  J. A., Bell, J. P., et  al; Joint Conf. on  Sensing Environmental
       Pollutants Collected, Technical Papers, 1972, Vol. 1,  71-1067.

 124.   Sigsby,  J. E., Jr., Black, F. M., et  al; Environ.  Sci. Tech., Vol. 7,
       pg. 51,  1973.

 125.   Spencer,  C. F., Baumann, F., et al; Anal.  Chem., Vol.  30, No. 9,
       pg. 1473, 1958.

 126.   Sumner,  S., Karrman, K. J., et  al; Mikrochim Acta, pg.  1144,  1956.

 127.   Liberti,  A.,  and Cartoni, G. P.,  Chim.  e ind., Vol.  39,  pg.  821,  1957.

 128.   Ryce, S.  A.,  and Bryce, W. A.,  Anal.  Chem.,  Vol.  29,  pg.  925,  1957.

 129.   Desty, D. H., and Whyman, B. H. F., Anal.  Chem.,  Vol.  29,  pg.  230,
       1957.

 130.   Coleman,  H. J., Thompson, C. C.,  et al; Anal.  Chem.,  Vol.  30,  pg.  1592,
       1958.
                                      197

-------
131.   Desty, D. H., and Harbourn, C. L. A., Div. of Analytical and Petroleum
      Chemistry, Symposium on Advances in Gas Chromatography, 132nd meeting,
      ACS, New York, N. Y., September, 1957.

132.   Amberg, C. H., Can. J. Chem., Vol. 36, pg. 590, 1958.

133.   Sullivan, J. H., Walsh, J. T., et al; Anal. Chem., Vol. 31, pg. 1827,
      1959.

134.   Dal Nogare, S., and Bennett, C. E., Anal. Chem., Vol. 30, pg. 1157,
      1958.

135.   Harrison, G.'F., Knight, P., et al; Second Int'l Symposium on Gas
      Chromatograph, Amsterdam, May, 1958.

136.   Adams, D. F., and Koppe, R. K., Tappi, Vol. 42, pg. 601, 1959.

137.   Schols, J. A., Anal. Chem., Vol. 33, pg. 359, 1961.

138.   Hall, H. L., Anal. Chem., Vol. 34, pg. 61, 1962.

139.   Adams, D. F., Kippe, R. K., et al; Tappi, Vol. 43, pg. 602, 1960.

140.   Feldstein, M., Balestrieri, S., et al; J. Air Poll. Cont. Assn.,
      Vol. 15, pg. 215, 1965.

141.   Fredericks, E. M., and Harlow, G. A., Anal. Chem., Vol. 36, pg. 263,
      1964.

142.   Kummer, W. A., Pitts, J. N., Jr., et al; Environ. Sci, Tech., Vol. 5,
      pg. 1045, 1971.

143.   Mizany, A. I., J. Chromatog. Sci., Vol. 8, pg. 151, 1970.

144.   Rasmussen, R. A., AM. Lab, Vol. 4, pg. 55, 1972.

145.   Brody, S. S., and Chaney, J. E., J. Gas Chromatog., Vol. 4, pg. 42,
      1966.

146.   Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070, 1972.

147.   Elliott, L. F., and Travis, T. A., Soil Sci. Soc. Amer. Proc., Vol. 37,
      pg. 700, 1973.

148.   Francis, A. J., Adamson, J., et al; Bull. Ecol. Res. Comm.,  (Stockholm),
      Vol. 17, pg. 485, 1973.

149.   Lewis, J. A., and Papavizas, G. C., Soil Biol. Biochem., Vol. 2,
      pg. 239, 1970.
                                    198

-------
150.   Lovelock,  J.  E.,  Maggs, R. J., et al; Nature, London, Vol  237
      pg.  452,  1972.

151.   Ronkainen,  P.,  Denslow, J., et al; J. Chromatog. Sci., Vol  11
      pg.  384,  1973.

152.   Stanley,  T. W., Sawicki, E., Johnson, H., and Pfaff, J. D.,
      Mikrochim,  Acta,  Vol. 1, pg. 48, 1965.

153.   Umbreit,  G.,  and Houtman, R. L., Journal of Pharmaceutical Sciences,
      Vol. 56,  pg.  349, 1967.

154.   Afghan,  B. K.,  Belliveau, P. E., Larose, R. H., and Ryan, J. P.,
      Analytica Chimica Acta, Vol. 71, pg. 355, 1974.

155.   Lawrence, J.  F.,  Journal of Chromatographic Science, Vol. 17, pg. 147,
      1979.

156.   Wolkoff,  A. W., and Larose, R. H., Journal of Chromatography, Vol. 99,
      pg.  731,  1974.

157.   Sheffer,  H. E., Perry, R. L., Thimineur, R. J., Adams, B. T., Simonian,
      J. L., Zutty, N.  L., and Clendenning, R. A., Ind. Eng. Chem. Prod.
      Res. Develop.,  Vol. 10, pg. 362, 1971.

158.   Shulgin,  A., Analytical Chemistry, Vol. 36, pg. 920, 1964.

159.   Argauer,  R. J., Analytical Chemistry, Vol. 40, pg. 122, 1968.

160.   Heenan,  M. P.,  and McCallum, N. K., Journal of Chromatographic Science,
      Vol. 12,  pg.  89,  1974.

161.   Kawahara, P.  K.,  Analytical Chemistry, Vol. 40, pg. 1009, 1968.

162.   Cohen, I. C., Norcup, J. , Ruzicka, H. A., and Wheals, B. B., J.
      Chromatog., Vol.  44, pg. 251, 1969.

163.   Seiber,  J. N.,  Crosby, D. G., Fouda, H., and Soderquist, C. J.,
      J. Chromatog.,  Vol. 73, pg. 89, 1972.

164.   Bartle,  K. D.,  Els tub, J., Novotny, M., and Robinson, R. J., Journal
      of Chromatography, Vol. 135, pg. 351, 1977.

165.   Ma,  T. S., and Speigel, D. , Microchemical Journal, Vol. 10, pg. 61,
      1966.

166.   Mindrup,  R.,  Jr., Industrial Research/Development, pg. 79, August,
      1978.

167.   Chriswell, C. D., Chang, R. C., and Fritz, J.  S., Analytical Chemistry,
      Vol. 47,  pg.  1325, 1975.


                                     199

-------
168.  Braithwaite, B., and Penketh, G., Analytical Chemistry, Vol. 36, pg.
      185, 1964.

169.  Landault, C., and Guiochen, G.,  Analytical Chemistry, Vol. 39, pg. 713,
      1967.

170.  Fisher, G. E., and Neerman, J. C., I and EC Product Research and
      Development, Vol. 5, pg. 288, 1966.

171.  Bark, L. S., and Clarke, K. F.,  J. Chromatog., Vol. 48, pg. 418, 1970.

172.  Di Corcia, A., Journal of Chromatography, Vol. 80, pg. 69, 1973.

173.  Preston, S. T., A Guide to the Analysis of Phenols by Gas
      Chromatography, PolyScience Corporation, Niles, Illinois, 1966.

174.  Eichelberger, J. W., Dressman, R. C., and Longbottom, J. E.,
      Environmental Science and Technology, Vol. 4, pg. 576, 1970.

175.  Yrjanheikki, Erkki, Am. Ind. Hyg. Assoc. J.,  Vol. 39, pg. 326,  1978.

176.  Brown, R. A., Searl, T. D., King, W. H., Jr., Dietz, W. A., and
      Kelliher, J. M., Rapid Methods of Analysis for Trace Quantities of
      Polynuclear Aromatic Hydrocarbons and Phenols in Automobile Exhaust,
      Gasoline and Crankcase Oil, Final Report for CRC - APRAC Project
      CAPE-12-68, Esso Research and Engineering Company, Linden, N.J., 1973.

177.  Barber, E. D., Sawicki, E., and McPherson, S. P., Analytical Chemistry,
      Vol. 36, pg. 2442, 1964.

178.  Grouse, R. H., Garner, J. W., and O'Neill, H. J., J. of G.C., pg.  18,
      February, 1963.

179.  Spears, A. W., Analytical Chemistry, Vol.  35, pg. 320, 1963.
                                    200

-------
         APPENDIX A





ALDEHDYE AND KETONE PROCEDURE
             201

-------
            THE MEASUREMENT OF ALDEHYDES AND KETONES IN EXHAUST
     The aldehydes and ketones that are included in this analysis are:
formaldehyde, acetaldehyde, acetone (acetone, acrolein, and propionaldehyde
are not resolved from each other under normal operating conditions and all
three are reproted together as acetone), isobutyraldehyde, methylethylketone,
crotonaldehyde, hexanaldehyde, and benzaldehyde.  The measurement of the
aldehydes and ketones in exhaust is accomplished by bubbling the exhaust
through glass impingers containing 2,4 dinitrophenylhydrazine  (DNPH) in
dilute hydrochloric acid.  The exhaust sample is collected continuously
during a test cycle.  The aldehydes and ketones (also known as carbonyl
compounds) react with the DNPH to form their respective phenylhydrazone
derivatives.  These derivatives are insoluble or only slightly soluble in the
DNPH/HC1 solution and are removed by filtration followed by pentane extrac-
tions.  The filtered percipitate and the pentane extracts are combined and
the pentane is removed by evaporation in the vacuum oven.  The remaining
dried extract contains the phenylhydrazone derivatives.  The extract is
dissolved in a quantitative volume of toluene containing a known amount of
anthracene as an internal standard.  A portion of this dissolved extract is
injected into a gas chromatograph and analyzed using a flame ionization de-
tector.  The detection limits for this procedure under normal operating con-
ditions are on the order of 0.005 ppm carbonyl compound in dilute exhaust.

SAMPLING SYSTEM

     Two glass impingers in series, each containing 40 m£ of 2NHCl-2,4
dinitrophenylhydrazine, are used to collect exhaust samples for the analysis
of the aldehydes and ketones.  A flow schematic of the sample collection
system is shown in Figure 1.  The two impingers together trap approximately
98 percent of the carbonyl compounds.   The temperature of the impinger is
maintained at 0-5°C by an ice water bath, and the flow rate through the im-
pinger is maintained at 4 &/minute by the sample pump.  A dry gas meter is
used to determine the total flow through the impinger during a given sampling
period.  The temperature of the gas stream is monitored by a thermocouple
immediately prior to the dry gas meter.  A drier is included in the system
to prevent condensation in the pump, flowmeter, dry gas meter, etc.  The
flowmeter in the system allows monitoring of the sample flow to insure pro-
per flow rates during sampling.  When sampling diesel fueled vehicles, a
heated filter, located between the on-off solenoid valve and the dilution
tunnel, is used to prevent diesel particulate from contaminating the sampling
system.  The filter and line connecting the filter to the dilution tunnel
are heated to 375 °F in order to prevent the aldehydes and ketones from being
retained in the filter and sample line.  The Teflon line connecting the
                                    202

-------
ro
o
GO
    Sample
     Probe
     Dilute
    Exhaust
                                                                         Gas Temperature
                                                                         Digital Readout
                                                                              Plowmeter
                   Regulating
                     Valve
     Ice Bath
Temperature Readout
                                               Dry
                                               Gas
                                              Meter

  Gas Volume
Digital Readout
                       Figure 1.  Aldehyde and ketone sample collection flow schematic.

-------
heated filter and the solenoid valve is heated to ~175°F in order to prevent
water from condensing in the sample line.  Several views of the  sampling
system are shown in Figure 2.

ANALYTICAL PROCEDURE

     The analysis of the aldehydes (formaldehyde, acetaldehyde,  isobutyr-
aldehyde, crotonaldehyde, hexanaldehyde, and benzaldehyde) and of the ketones
(acetone and methylethylketone) in dilute exhaust is accomplished by col-
lecting these carbonyl compounds in a hydrochloric acid  (HCl)/2,4 dinitro-
phenylhydrazine  (DNPH) solution as their 2,4 dinitrophenylhydrazone  deri-
vatives.  The derivatives are removed from the HCl/DNPH absorbing solution
by filtration and/or extractions with pentane.  The filtered precipitate and
the pentane extracts are combined and the volatile solvents are  removed.
The remaining extract contains the phenylhydrazone derivatives.   The deri-
vatives are then dissolved in a quantitative volume of toluene containing a
known amount of anthracene as an internal standard.  This solution is ana-
lyzed by injecting a small volume of the solution in to a gas chromatograph
equipped with dual flame ionization detectors.  From this analysis and the
measured volume of exhaust sampled, the concentration of the carbonyl com-
pounds in exhaust can be determined.  The analysis flow schematic for the
aldehydes and ketones is shown in Figure 3.  A detailed description  of the
procedure follows.

     The aldehdyes and ketones are trapped in solution by bubbling a known
volume of dilute exhaust through two glass impingers connected in series,
with each impinger containing 40 m& of a 2 N HC1 solution saturated with
DNPH.  The sampling temperature and barometric pressure are recorded Curing
this bubbling period.  The carbonyl compounds in the exhaust react with the
DNPH to form slightly soluble or insoluble 2,4 dinitrophenylhydrazone deri-
vatives.  The two impingers together collect 98+ percent of the  carbonyls
that are present in the exhaust.  The impingers are removed from the sam-
pling cart and are allowed to stand at room temperature for at least one hour
before proceeding to the filtration and extraction steps.  Figure 4  shows
two impingers containing the HCl/DNPH absorbing solution after being removed
from the sampling cart.

     Under normal operating conditions the contents of the two impingers
are combined and analyzed as one sample.  If either of the two impingers
contain a precipitate they are first subjected to a filtration step.  If
no percipitate is present, this filtration step is omitted and the extrac-
tion step, described later in the procedure, is the first step.

     For the filtration step, the contents of the two impingers  are  poured
through a fritted glass filter into a flask under vacuum  (Figure 5).  The
two impingers are rinsed with small portions of deionized water.   This wash
water is also poured through the fritted glass filter.  The precipitate in
the filter is then washed with a few m£ of deionized water.  The fritted
filter is then removed from the flask containing the 80 m£ of absorbing
reagent and the water washings.  The flask is then set aside for the ex-


                                    204

-------
               Front View
                                                   Digital
                                                   Readout
                                                   Flowmeter
                                                   Regulating
                                                   Valve
        Close-up of Upper Front
Figure 2.  Aldehyde and ketone sampling system
                       205

-------
  Solenoid
Impinger
Ice Bath
                               Close-up of  Impingers  (Side View)
                                                         Solenoid
                                                         Filter
                                                         Ice  Bath
                                                         Drier
                                                         Dry Gas Meter
                                                         Pump
                 Rear View
      Figure  2  (Cont'd).   Aldehyde  and kstone sampling system,
                                 206

-------
       CVS
    glass impingers
 derivative filtered
 and extracted with
 pentane from absorber
filtered ppt combined
with pentane extract
and solvent removed
extract dissolved in
toluene containing
anthracene as an
internal standard
          I
 sample analyzed
in gas chromarograph
      with FID
   A/D converter
    Recorder
Hewlett-Packard 3354
  Computer System
Figure 3.  Aldehyde and ketone analysis flow schematic.
                          207

-------
Figure 4.  Impingers containing HC1/DNPH abosrbing solution.
                            208

-------
I >
c
c
                                   Figure  5.   Filtration of absorbing solution.

-------
traction step.  The fritted glass filter containing the precipitate  is  con-
nected to a dry flask.  The two impingers that had previously contained the
filtered precipitate are then each washed with small portions of methylene
chloride.  The methylene chloride dissolves any solid residue which  was not
removed by the water wash.  These methylene chloride washings are poured
into the fritted glass filter containing the precipitate.  After the pre-
cipitate has been dissolved by the methylene chloride, a vacuum is applied
to the flask and the methylene chloride solution is pulled through the
filter into the flask.  Another small amount of methylene chloride is poured
through the filter into the flask to wash the filter.  The methylene chloride
solution is now saved until the extraction step is complete.

     The extraction step is carried out as follows.  The contents of the
two impingers  (if no precipitate is present) are transferred to a 250 m£
separatory funnel.  The impingers are each washed with small portions of
deionized water which is also added to the separatory funnel.  If a  pre-
cipitate was found in the impingers the contents of the flask containing
the filtered absorbing reagent and the water washings from the filtration
step are transferred quantitatively to a 250 m£ separatory furinel.   The
flask is washed with a small portion of water, and this water, is added  to
the separatory funnel.  Forty m£ of pentane is now added to the separatory
funnel containing the 80 m£ of absorbing reagent and water washings.  The
funnel is stoppered and shaken for five minutes in an automatic shaker,
Figure 6.  The shaker is stopped and the funnel is vented.  After the two
phases are allowed to separate, the lower phase is collected in a second
separatory funnel.  The remaining phase is transferred to a third 250 m£
separatory funnel.  A second 40 mJl portion of pentane is added to the al-
ready once extracted absorbing solution.  The funnel is again stoppered,
shaken for 5 minutes and vented.  After the phases have separated, the
lower phase is again collected in another separatory funnel.  The upper
or pentane layer is combined with the pentane layer from the first extrac-
tion.  A third 40 m£ portion of pentane is added to the twice extracted
absorbing solution and the extraction process repeated.  After the third
extraction, the lower layer is discarded and the pentane layer is combined
with the pentane layers from the first two extractions.  Any absorbing
solution which might have been accidently transferred with the pentane
layers is drained off.  Deionized water (25-50 m£) and sodium bicarbonate
(1/4-1/2 gram) is added to the 250 m& separatory funnel containing the  120
m£ of pentane extract.  The funnel is stoppered and manually shaken  for
30 seconds.  The phases are allowed to separate and the lower water  phase
is drained off.  Another 25 m£ of deionized water is added and the shaking
is repeated.  After the phases have separated, the water is drained  off
insuring that all traces of water are removed.  The contents of the  funnel
are then combined with the methylene chloride solution which was saved  from
the filtration step.

     The flask containing the methylene chloride solution and the pentane
extracts is then placed in a vacuum oven, Figure 7, operating at 50-60°C
and 65" water vacuum until the pentane and methylene chloride have been
removed.  At this time only the dried phenylhydrazone derivative remain.
                                    210

-------
Figure 6.  Automatic shaker.
  Figure 7.  Vacuum oven,
           211

-------
     Each time a series of samples are  collected,  a blank containing 80 mJi
of HCl/DNPH solution is extracted and dried  in the same manner as the samples.
This accounts for any aldehydes or interferring compounds which might be
found in the reagents used for extraction.

     Two m£ of toluene which contain a  quantitative amount of anthracene
C\£.05 mg/m£ toluene) as an internal standard  is pipetted into the flask
containing the dried phenylhydrazone derivatives.   The flask is then placed
• in a sonic bath until all of the residue is  dissolved.   After the precipi-
tate has dissolved, the solution is transferred to a 1/2 dram vial (Figure 8).
At this point the derivative is ready for injection into the gas chromato-
graph system.

     The gas chromatograph system used  to analyze  the toluene solution con-
taining the pherylhydrazone derivatives is shown in Figure 9-  The system
consists of a Varian 1700 GC, and A/D converter, and a recorder.  The GC
is equipped.with dual columns and dual  flame ionization detectors with a
single differential amplifier.  The columns  consist of 24 x 1/8 inch O.D.
stainless steel tubing packed with 6.7 percent Dexsil (polycarboranesiloxane)
300 GC on DMCS treated and acid washed, 60/80  mesh Chromosorb G.  The carrier
gas is helium which flows through the columns  at a rate of 40 mi/minute.
The optimum hydrogen and air flow rates are  35  m&/minute and 500 mil/minute,
respectively.  The column temperature,  after injection of the sample, is
programmed from 120°C to 300°C at 8° a minute.   In a chromatogram of a
standard sample (Figure 10) containing  anthracene  and the phenylhydrazone
derivatives of formaldehyde, acetaldehyde, acetone,  isobutryaldehyde, methyl-
ethylketone, crotonaldehyde, hexanaldehyde,  and benzaldehyde, the first peak
eluted is toluene followed by anthracene, and  then the derivatives of for-
maldehyde, acetaldehyde, acetone, isobutyraldhyde,  methylethylketone, cro-
tonaldehyde , hexanaldehyde, and benzaldehyde.   Data obtained from the five
repetitive injections of the standard derivatives  in toluene showed a max-
imum standard deviation of 4.56 percent for  benzaldehyde and a minimum stan-
dard deviation of 0.87 percent for formaldehyde.   The computer printout of
the standard, Figure 10, is shown in Figure  11.  This printout gives the
retention time, area, and the name of each peak.   The printout also gives
the concentration of each of the derivatives in mg/m£.   The concentration  is
calculated by the computer from response factors which are determined daily-
Each day a  standard containing known amounts  of the derivatives and anthra-
cene is injected into the GC.  From the anthracene and derivative areas the
computer calculates a response factor F.  The  F factors are used in all sub-
sequent runs during the day to determine the concentration of the derivatives.
This response is calculated from the following equation:


           Response Factor (F)  = Anthracene  Area x mg/m£ Derivative
                                 Derivative  Area   mg/m£ Anthracene
                                    212

-------
            Figure 8.  1/2 dram vials.
Figure 9.  Aldehyde and ketone analytical system.







                       213

-------
                                                SOTOUAldehyde & Ketone Stdjfa
                                                Imtrunmit ynT-< »
                                                Column  2   ft. 1/8   P.P.
 1.  Injection
 2.  Toluene
 3.  Anthracene
 4.  Formaldehyde
 5.  Acetaldehyde
 6.  Acetone
 7.  Isobutyraldehyde
 8.  Methylethylketone
 9.  Crotonaldehyde
10.  Hexanaldehyde
LI.  Benzaldehyde
               m. ry>xsii ago GC
               Qiromosorb G
            °C mimi 40 _cc/mln. Helium
                Rounwttr RMdii
• Id  aUa
     120  "CISOIof  -
      Oj_i_
/min. Htld
   min. lothwl
   "C.
   "C
                    min.. Prog to	"CM
                  Rounwtw Rdg,
                  Rouiratw Rdg.
                  Rounwur ftdg.
                  1   mV.F.S. so
                           14       12        10       8

                             Retention time, minutes
                  Fiugre  10.   Chromatogram of  standard.
                                             214

-------
REPORT:     14.11  CHANNEL:  11

SAMPLE: STANDARD      INJECTED AT 11:18:27 ON  MAR

ISTD METHOD:  DNPH11
                                                                        1978
              ACTUAL  RUN  TIME:  30*008 MINUTES

              ISTD-RATIO:     .050..RMG/ML     STD-AMT:
                                               0500
                               SAMP-AMT:
                                     1 .0000
                 RT
            AREA
MG/ML
ro
7.26
9.81
11.71
12.64
13.28
13.70
14.69
16.08
19.00
9638
11 159
13355
17898
16448
16469
11 167
15988
10525
BB
BB
BV
VV
VV
VV
VV
VV
BB

.203
• 202
.203
.202
.201
. 199
.202
. 198
              TOTAL  AREA =
                    122648
     NAME

&ANTHRACENE
#FORMALDEHYDE
#ACETALDEHYDE
#ACETONE
#I SO-BUTYRALDEHYDE
#MEK
#CROTONALDEHYDE
#HEXANALDEHYDE
#BENZALDEHYDE

 TOTAL MG/ML   =
                              1.610
                               Figure 11.  Computer printout of standard .

-------
Typical response factors for each of the derivatives are listed below:

                      Factor            Name	

                      1.0000        Anthracene
                      3.1043        Formaldehyde
                      2.7736        Acetaldehyde
                      2.2366        Acetone
                      2.4160        isobutyraldehyde
                      2.3332        MethylethyUcetone
                      3.4174        Crotonaldehyde
                      2.342 8        Hexanaldehyde
                      2.9329        Benzaldehyde

When the response factor is known a concentration in mg/mH for each of the
derivatives can be found.  This concentration, along with the volume of
sampled exhaust is then used to calculate the concentration of the carbonyl
compounds in exhaust.  Figures 12 and 13 show a typical sample chromatogram
and accompanying printout respectively.

CALCULATIONS

     This procedure has been developed to provide the user with the concen-
trations of the aldehydes (formaldehyde, acetaldehyde, isobutyraldehyde,
crotonaldehyde, hexanaldehyde, and benzaldehyde) and ketones (acetone and
methylethylketone) in exhaust.  The results will be expressed in Vg/m3 of
exhaust and ppm for each carbonyl compound.  The equations for determining
the concentrations in yg/m^ and ppm are derived in the following manner.

     The first step is to correct the volume of exhaust sampled to a stan-
dard temperature, 68°F and pressure, 29.92"Hg, by use of the equation

                  P    v V      P       V
                   exp x  exp _  corr x  corr
                      T             T
                       exp           corr
     V     = experimental volume of gas sampled in ft
     VGXP  = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
     P     = experimental barometric pressure
     P  P  = 29.92"Hg
      corr
     T     = experimental temperature in °F + 460
     TexP  = 68°F + 460 = 528°R
      corr

     Solving for V     gives:
                  corr

                             P    ("Hg) x V     (ft3) x 528°R
                     v     =  e*P	exp
                      corr        T    (°R)   29.92" Hg
                                   exp
                                    216

-------
 1.  Injection
 2.  Toluene
 3.  Anthracene
 4.  Formaldehyde
 5.  Acetaldehyde
 6.  Isobutyraldehyde
 7-  Methylethylketone
 8.  Crotonaldehyde
 9.  Hexanaldehyde
10.  Benzaldehyde
    24
22
12    10
                          Figure 12.  Sample chroamtogram.
                                       217

-------
               REPORTS     20      CHANNEL:  11

               SAMPLE: RCI            IMJECTED AT   15:41:05 ON MAP.   1*  1976

               ISTD METHOD:  DNPHll
               ACTUAL RUN  TIME:   30.017 MINUTES

               ISTD-RATIO:     .050*R MG/ML     STD-AMT:
                       0500
                       SAMP-AMT:
1.0000
                           AREA
MG/ML
ro
!-»
oo
7. 15
7.95
10.03
11. 88
12.83
13.76
14.85
15.95
16.77
19.20
20.23
23.52
25.03
25.48
8604
435
8877
575
463
1594
2630
146
675
648
1912
217
13
84
BV
VE
BB
BV
W
W
vv
vv
vv
VB
BV
VV
VB
BB

• 003
. 186
.010
.007
.022
.053
.002
• 004
. 012
• Oil
.001
7-6E- 5
4.9E- 4
               TOTAL AREA =         26874

               PROCESSED DATA  FILE: *PRC11
     NAME

&AMTHP.ACENE

^FORMALDEHYDE
#ACET ALDEHYDE
# I SO -BUTYRALDEHYDE
                                                #CR0TONALDEHYDE
                                                #HEXAN ALDEHYDE

                                                #BENZALDEHYDE
          TOTAL MG/ML =        .310

          RA¥ DATA FILE:  *RAW11
                               Figure 13.  Computer printout of sample

-------
 The next step converts the  volume from cubic feet to cubic meters by

of the conversion factor;! cubic meter is equal to 35.31 cubic feet.
use





                   V      ,    Pexo  ("H9)X vAvn  (ft)x  528°R
                                ex           exp      _

                                      29.92" Hg x  35.31  ft3/m3





                                                             (Equation 1)



     The next step converts the mg/m£ of  derivative  determined by the com-

puter to mg of carbonyl collected in the  two impingers.   To  obtain mg of

derivative, the concentration  (from the computer printout) in  mg/mJl is

multiplied by the volume of toluene used  to dissolve the solid extract.
          mg derivative = Conc_   (mg/m£)  x Vol     (m£)
                              L/63T
     To find mg of carbonyl compound per  sample  the mg of derivative are

multiplied by the ratio of the molecular  weight  of the carbonyl derivative

over the molecular weight of its phenylhydrazone derivative.




                .    ,       ..            mol. wt. carbonyl
          mg carbonyl = mg derivative x — ; - — — - — : — *-. —
                                        mol. wt. derivative
          = Cone,,   (mg/m£) x Vol   ,   (m&)    mol. wt.  carbonyl
                Der   *'         Tol        x — ; - - — - — : — f-r—
                                              mol. wt.  derivative



     To obtain the number of TO of  carbonyl compound  the  mg of  carbonyl are

 multiplied by the conversion factor,  1000  yg/mg.




     Vig carbonyl =* Cone     (mg/m£)    Vol   1 (m£)   mol. wt.  carbonyl

                       Der                        mol. wt.  derivative


                 x 1000 yg/mg

                                                             (Equation  2)



     The concentration of the carbonyl compound in  exhaust  can  now be  found

 in iug/m3 by dividing equation 2 by  equation 1.




                      Cone     (mg/m£) x Vol  . (m&lxmol.  wt. carbonyl
                3         Der              Tol         _
   pg carbonyl/m   =  - - - (1   } x y - (.ft3)  x  528°

                               exp           exp



                      1000  ug/mg xT     (°R) x  29.92" Hg x  35.31  ft3/m3
                   x  _ _       exp
                                      mol. wt.  derivative




                                                             (Equation 3)







                                     219

-------
     To find the concentration of each carbonyl compound  in ppm,  the den-
sities of carbonyls are needed.  At 29.92" Hg and 32°F, one mole  of gas
occupies 22.4 liters.  This volume is corrected to 68°F from  the  equation.
                                 v
                         V  = 22.4
                         T^ = 32 °F + 460 = 492 °R

                         V  = volume at 68°F
                         T  = 68°F + 460 = 528°R
     Solving for V gives :

              V x.,T     22.4 x 528
Since one mole of gas occupies 22.04& at 68°F, the density can be  found in
g/H by dividing the molecular weight in g/mole by 24.04 8,/moIe.

          ,ori  < /(M _ mol. wt. (g/mole)
          den  (g/Jl) -- 24.04Vmole

The density in yg/m£ can be found by converting g to yg and yg and £ to
m& as follows :

             o    mol. wt. g/mole   1 x 10  yg/g   mol. wt. x 1000
     den yg/rn*  -- 24.04Vmole   X 1 x 1Q3 ml/A X       24~04

                                                             (Equation 4)

To obtain the concentration of each carbonyl in ppm, the concentration in
yg/m3 is divided by the density in yg/m&

                    ppm = yg/m3 T yg/m& = — r
                                          m

Using Equations 3 and 4 gives the ppm concentration in the form of the raw
data.
           Cone    (mg/m£) x Vol    (m£) x mol. wt. carbonyl x 1000 yg/mJl
     ppm = - B§! - — — -
               P    ("Hg) x V    (ft )  x 528° x mol. wt. derivative
                exp          exp

               Texp (°R)  X 29'92" H9 x 35'31 ft3/m3 x 24.04
                              mol. wt carbonyl x 1000
                                    220

-------
          Conper (mg/m£)  x VolToi (m£) * Texp  (°R) x 29.92" Hg

                     Pexp ("Hg) X Vexp  (ft3) X 528°


                35.31 ft3/m3 x 24.04 A/mole
                  mol . wt . derivative
                                                             (Equation 5)

At this point,  the concentration can be expressed in yg/m   (Equation 3) and
ppm (Equation 5)  at 68°F and 29.92" Hg from the raw data.

Hewlett-Packard Calculations

     In order to insure maximum turnaround in a minimum time period a Hewlett-
Packard 67 program was developed to calculate the aldehyde  and ketone concen-
tractions in yg/m3 and ppm from the raw data and phenylhydrazone derivative
concentrations (from computer printout) .  This program is presented in
Figure 14.

Sample Calculations

     Assume exhaust samples were collected in glass impingers for each por-
tion of a three bag 1975 FTP.  Raw data for these tests is  presented in
Figure 15.  Calculations were performed using the HP-67 programs and manual
calculations .

Manual calculation for driving cycle FTP-1 :
                          Cone     (mg/mJl) x Vol     (m£) x mol. wt. carbonyl
     yg/m  formaldehyde                         °
                             1000 yg/mg x T     (°R)  x  29.92" Hg

                           x              528°R

                               35.31 ft3/m3
                           X mol. wt. derivative

                          0.186 mg/ml, x 2m£ x  30.03  g/mole x 1000 yg/mg
                                 29.80" Hg x 3.196 ft3 x  528°G
                             535°R x 29.92" Hg  x  35.31  ft3/m3
                           X         201.15 g/mole

                        = 597.5 yg/m3
                                                  n
                          ppm formaldehyde = yg/m  v  density yg/mJl

                                        mol. wt.  (formaldehyde)  x 1000
                        density yg/m£ = 	24.04&~
                                     221

-------
        I S
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEYEN1RY KEY COO£ ,,,U^M^
l!OJ





010
	
	
	

0»





	


030









040









OaO





"
I LUL..& _
U
0
n
0
5
X
X
S'I'O 1 	
R/s 	
. 4 	 .
	 t 	
	 0. 	
RCL 1

i
ll 1/X
R/S
X
STY) •}
R/S
RCL 2
	 X 	
.
,
d
_3
X
R/S
1
2
4
q

B/e;
RCL 2
X
0
,
1

ft
Y
D /J
1

^
2
.
R/S
ur-i. •}
V
XUIii 11.
. L1JJ._
00
no
DO
2i
71
84
. 21.
84"
04
	 ua 	
bl
34 01
35 52
81

84
71
1! IK:

34 u^
00
83
(id
03
71
84
(it
02
04
nq
81
84
34 02
71
00
83
01

iifi
71
H4
01

in
02
si
84
14 02
71
Input S.imj.
«- Vol . ,
rt1
Input barometer ,"H9
Input :ijn,[,l
e 'IVmp,°F
h,| ui V..I. lulurne,
lUpIlt !ii-l iiii'.
i)utput \\^/
Output pj'-ti
uut/put liq/u
Output ppm.

0 1 2.3
SO 31 S2 S3
A
3 C
, In niLj/in?
In tny/mS.
Ot>u





1 	
070






ObO
-





090









100









MO


-.Q. .
_ 2...
	 4 	
X
R/S
2
4
1
R/S
	 ECi_2_
0
.
2
8
6
X
-is/s.

0
u 	
	 £ 	
R/S
HCI. 2
x
0

2
8
6
X
R/S
3
0
0
r>

R/S
RCL 02
X
0
.
2
8
0
X
R/S
2
9
1
6
bJ
-.- Ui
71

02

- 01
. ._ Hit
81
B4
...34 	 02
"OO
83
02
08
ll 1 1
.. n ..
- ii-i . .
uu .
.. ..JU 	
HI
84
34 U2
71
00
HI
02
08
06
71
84
03
00
00
00
81
U4
34 02
7)
00
83 j
02
08
00
71
84
02
09
01
06
output v^/ra-*
Output ppm, In
,9/mt
Output ppiu, In
Output My/m
Output ppiu/ In
Output iiy/in
REGISTERS ._ . „ . .
4
34

5 6 /
S5 36 S7
s
D
y y
Sti b4
F. I
Figure 14 (Cont'd).  HP-67 program form,
                  223

-------
                    Program I
SIEP  KtYENTHY  KEY CODE
                     COMMtNIl     SltP  KEYtNTHY  KtY COIJt
                                                   CUMMIN1&
" ~


	
'~
1."

-- —








	





14.1








150






	
lou

1
1 _




R/S
_RCL 02 _
X
0

3
5
9
R/S
4

tl
6

k/s
kri, 2
X
	 0 	
7
].
X
R/S
4
j
1
5

R/S
h RTN







	














ftl 	
84
3-1 02
71
00
83
03
05
09
	 LL 	
84
0
0
i 	
1
06
0
&
81
34 02
71
0
	 a
0
0 	 	
i 	
3 	
7
01
71
84
04
04
01
05
8
]
84
^S5^L_






	














Output. H'iii, ^" I-'J/111^
J
Output i.^/m
Output
output
Pf-iu, limy
j

Output ppni




If'U
	





	
ttJG
	




— • —







200








;rd —









J.'U




LABELS
A
^ t>
1
1:
c
c
^
7
11
.1
J
8
L
'•'
.1
y


	















	


















	
















	

-



-








-
_




	


















	 . 	
.











FLAGS
u
1
2
J

SET STATUS
FLAGS TRIG LlISP
ON Ol [-
U II 1 i
i I i , j
i 1 ! i J
J 1 i i .
DLu 1 111 i
I'iltAl) i • SU
HAD i ! 1 r*.
n
       Figure 14 (Cont'd).   HP-67 program form.
                        224

-------
SWRI PROJECT NO.	TEST NO.	TEST DATE:



FUEL:	CVS NO.	TUNNEL SIZE:	DRIVER:
SAMPLE COLLECTION BY:_



GENERAL COMMENTS:	
_CHEMICAL ANALYSIS BY:
                               VEHICLE:
                                                           MILES:
                           CALCULATIONS BY:
Test No.
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Vol. Toluene ml
Formaldehyde Der Cone mg/ml
Formaldehyde Cone yg/m3
Formaldehyde Cone ppm
Acetaldehyde Der Cone mg/ml
Acetaldehyde Cone yg/m3
Acetaldehyde Cone ppm
Acetone Der, Cone mg/ml
Acetone Cone yg/m3
Acetone Cone ppm
I-Bu Aldehyde Der Cone mg/ml
I-Bu Aldehyde Cone yg/m3
I-Bu Aldehyde Cone ppm
MeEt Ketone Der Cone mg/ml
MeEt Ketone Cone yg/m3
MeEt Ketone Cone ppm
Cro- Aldehyde Der Cone mg/ml
Cro-Aldehyde Cone yg/m3
Cro-Aldehyde Cone ppm
Hex- Aldehyde Der Cone mg/ml
Hex-Aldehyde Cone yg/m3
Hex-Aldehyde Cone cpm
Benzaldehyde Der Cone mq/ml
Benzaldehyde Cone ya/m3
Benzaldehyde Cone ppm
1
FTP-1
3.196
29.80
75
2
0.186
598
0.479
0.127
559
0.305
0.121
663
0.274
0.022
141
0.047
0.098
630
0.210
0.086
541
0.186
0.031
250
0.060
0.093
775
0.176
2 3 4 5 fi
FTP-2
1.625
30.02
80
2
0.105
665
0.532
0.092
798
0.436
0.098
1060
0.439
0.011
139
0.046
0.084
1060
0.353
0.074-
917
0.314
0.018 -
286
0.069
0.081
1330
0.301
FTP-3
2.010
29.02
96
2
0.201
1100
0.881
0.157
1170
0.639
0.161
1500
0.621
0.028
305
0.102
0.097
1060
0.353
0.076
811
0.278
0.030
411
0.099
0.097
1370
0.310
SET-7
3.730
29.25
85
2
0.312
891
0.713
0.282
1100
0.600
0.285
1390
0.575
0.023
131
0,044
0.198
1130
0.377
0.105
587
0.201
0.027
194
0.047
0.121
897
0.203
HFET
8.241
29.95
83
2
0.732
921
0.737
0.612
1060
0.579
0.595
1280
0.530
0.051
128
0.043
0.252
634
0.211
0.286
705
0.242
0.078
246
0.059
0.232
757
0.171
NYCC
1.070
29.50
89
2
0.142
1410
1.130
0.102
1390
0.759
0.105
1780
0.737
0.009
179
0.060
0.075
1490
0.497
0.072
1400
0.480
0.011
275
0.066
0.081
2090
0.473
                Figure 15.   Aldehyde  collection sheet,
                                     225

-------
                           tnol. wt formaldehyde = 30.03 g/mole


                                  30.03 g/mole x 1000   ,_...
                        density = 	24.04&	=

     ppm = 597.5 yg/m3 4 1249 yg/m£ = 0.478 mVm3 = 0.478 ppm

The calculations for acetaldehyde, acetone, isobutyraldehyde, methylethyl-
ketone, crotonaldehyde, hexanaldehyde, and benzaldehyde are carried out in
the same manner by substituting the appropriate derivative concentrations
and molecular weights into the above formulas.  These calculations give the
following concentrations:

     acetaldehyde,       561 yg/m  and 0.306 ppm
     acetone,            663 yg/m  and 0.274 ppm
     isobutyraldehyde,   141 yg/m3 and 0.047 ppm
     methylethyIketone,  630 yg/m3 and 0.210 ppm
     crotonaldehyde,     541 yg/m3 and 0.186 ppm
     hexanaldehyde,      250 yg/m  and 0.060 ppm
     benzaldehyde,       775 yg/m  and 0.176 ppm

Note:  The values used in these calculations are picked from a range of tem-
peratures, derivative concentrations, etc. to validate the calculations and
may not be representative of expected raw data.  The calculations are pre-
sented to confirm the manual and HP-67 calculations give the same results.
This was confirmed for six sets of calculations.

LIST OF EQUIPMENT

     The equipment required for the analysis of aldehyde and ketones is
divided into three groups:  sample acquisition, sample preparation, and
sample analysis.  Manufacturer, stock number and any pertinent descriptive
information are listed.

Sample Acquisition

     1.   Glass impingers. Ace Glass Products, Catalog #7530-11, plain
          tapered tip stoppers with 18/7 arm joints and 29/42 bottle joints.

     2.   Flowmeter, Brooks Instrument Division, Model 1555, tube size
          R-2-15-C, graduated 0-15, sapphire float, 0-5 £/minute range.

     3.   Sample pump, Thomas Model 106 CA18, capable of free flow capacity
          of 4 £/minute.

     4.   Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
          capacity.

     5.   Regulating valve, Nupro 4MG, stainless steel.
                                    226

-------
 6.    Teflon tubing,  United States Plastic Corporation, 1/4" OD x
      1/8"  ID and 5/16" ID x 1/8" ID.

 7.    Teflon solenoid valve. The Fluorocarbpn Company, Model DV2-144NCA1.

 8.    Drying tube, Analabs, Inc., Catalog #HGC-146, 6" long, 1/4" brass
      fittings.

 9.    Miscellaneous Teflon nuts, ferrules, unions, tees, clamps,  connec-
      tors , etc.

10.    Digital readout for dry gas meter.

11.    Miscellaneous electrical switches, lights, wirings, etc.

12.    Six channel digital thermometer, Analog Devices, Model #2036/J/1.

13.    Iron/Cons tan tan type J single thermocouple with 1/4" OD stainless
      steel metal sheath, Thermo Sensors Corporation.

14.    Variable autotransformer, Staco Inc., Type 3PN 1010.

15.    Heating sleeve wrapped with insulation and insulation tape.

16.    Class A, 20 m& volumetric pipets.

17-    Class A, 1000 m£ volumetric flask.

18.    Teflon coated stirring bar.

19.    Hot plate-stirrer, corning, PC-351.

20.    Stainless steel heated filter assembly - 7 cm, Scott, capable  of
      temperature to 204°C, includes 2 heaters, adjustable thermostat
      switch, stainless steel insulated covers and sample bypass solenoid
      valves.

21.    Glass microfiber filter discs, Reeve Angel 934-AH, Whatman, 7  cm
      diameter.
22.   Flexible, heavy insulation heating tape, Briskeal***, width-1/2  inch,
      length-48 inches.

23.   Temperature Controller, Athena, 100-600°F.

24.   Heated TFE Teflon hose, Technical Heaters Inc, 5' x 1/4", tempera-
      ture limit 400°F.
                                227

-------
Sample Preparation

     1.   Fritted glass filters, Ace Glass Company, porosity  D, ASTM 10-20
          microns pore size, 24/40 ground glass joint, vacuum takeoff.

     2.   Constant temperature vacuum oven, National Appliance Company.

     3.   Pump for oven, Thomas Industries, Model 907CA18 2.

     4.   Flasks, 125 m£ capacity, 24/40 ground glass joints.

     5.   Separatory funnels, 125 m£.

     6.   Separatory funnels, 250 m£.

     7.   Separatory funnel shaker, Burrell Corporation, Wrist-Action ©type
          with appropriate funnel holders, Model 75.

     8.   Ring stands, labels, holders, tubing, vacuum tubing, fittings and
          clamps needed for equipment manipulation.

     9.   Wash bottles, 500 m£.

    10.   Graduated cylinders, 50 m£.

    11.   Vials, Kimble, 1/2 dram.

    12.   Vacuum pump, Sargent-Welch.

Sample Analysis

     1.   Varian 1700 gas chromatograph equipped with dual flame ionization
          detectors in differential operation, and a linear temperature
          programmer.

     2.   Soltec Model B-281 1 mv recorder.

     3.   Hewlett-Packard Model 3354 gas chromatograph computer system with
          remote teletype printout.

     4.   Syringe, 10 mi, Hamilton Company, #701.

     5.   Dual columns, 24 x 1/8" ID, stainless tubing packed with 6.7
          percent Dexsil 300 GC on Chromosorb G 60/80 mesh, DMCS treated
          and acid washed.

LIST OF REAGENTS

     A list of the reagents used in the determination of the  aldehydes and
ketones in exhaust is provided along with chemical formula, molecular weight,
purity, manufacturer, and catalog number.


                                    228

-------
    1.   Hydrochloric acid,  HC1,  36.46 g/mole,  concentrated (37%),  analyt-
         ical reagent, Mallinckrodt,  Cat. #2612.

    2.   Pentane, C5Hi2,  72.15 g/mole, Distilled in glass (bp 35-37°C),
         Burdick and Jackson Laboratories, Inc.

    3.   2,4 DinitrophenyIhydrazine (2 ,4-DNPH) , (NO2) 2C6H3CH=N -NH2,
         210.149 g/mole,  Aldrich analyzed, Aldrick, Cat. #D19,930-3.

    4.   Sodium Bicarbonate, NaHC03,  84.00 g/mole, Mallinckrodt,  Cat.  #7412.

    5.   Anthracene, C14H10, 178.24 g/mole, K and K Laboratories, Cat.
         #10714.

    6.   Toluene, C6H5CH3,  92.14 g/mole Baker Analyzed Reagent, Baker
         Cat. #3-9460.

    7.   Methylene  Chloride, CH2C12,  84.93 g/mole, Reagent ACS, Eastman,
         Cat. #13022.

PREPARATION  OF ABSORBING  SOLUTION

    To prepare the  absorbing solution, 163 m& of concentrated HC1 and  2.5  g
of 2,4-DNPH  crystals are  added to a one liter volumetric flask containing
about  500 m& of deionized water.  The flask is diluted to mark and stirred
for several  hours at room temperature with an automatic stirrer/Teflon
coated stirring bar  to dissolve the DNPH.  Fresh absorbing solution  is  pre-
pared  daily  as needed.

PREPARATION  OF TOLUENE/ANTHRACENE SOLUTION

    Toluene containing approximately 0.05 mg anthracene per m£ of toluene
is used to dissolve  the dried phenylhydrazone extracts.  This solution  is
made by adding 100 mg of  anthracene to a two liter volumetric flask  and di-
luting to mark with  toluene.

PREPARATION  OF PHENYLHYDRAZONE DERIVATIVES

    In order to obtain response factors for each of the phenyhydrazone
derivatives  to anthracene, pure derivatives were prepared from their re-
spective  aldehydes and ketones.  These derivatives were made by adding  each
of the carbonyl compounds separately to a 2NHCl-DNPH solution.  The result-
ing orange to red precipitates were filtered and dried.  The derivatives
were then recrystallized  from hot absolute ethanal.  The melting points
for each  of  the derivatives  were compared to literature values before use.
A GC trace was also  made  on  each of the derivatives to further check the
purity.
                                    229

-------
PREPARATION OF STANDARD SOLUTION OF PHENYLHYDRAZONE DERIVATIVES AND ANTHRACENE

     A standard containing the phenylhydrazone derivatives and anthracene in
toluene is prepared to obtain a response factor of each of the derivatives
to anthracene.  The solution is made by dissolving weighed amounts of an-
thracene and each of the derivatives in a quantitative volume of toluene.
These solutions contain 0.05 mg anthracene per m£  of  toluene  and 0.2 mg  of
each  derivative  per m& of  toluene.

REFERENCES

     This procedure is taken from the procedure:  "Oxygenated Compounds  in
Automobile Exhaust-Gas Chromatograph Procedure"  by Fred Stump, ESRL,
Environmental Protection Agency, Research Triangle Park, North Carolina.
                                   230

-------
      APPENDIX B




TOTAL CYANIDE PROCEDURE
         231

-------
                THE MEASUREMENT OF TOTAL CYANIDE IN EXHAUST
     The measurement of total cyanide (hydrogen cyanide and cyanogen) in
dilute exhaust is accomplished by bubbling exhaust through glass impingers
containing a 1.0 N potassium hydroxide absorbing solution.  The cyanide
reacts with the potassium hydroxide to form a stable salt which remains in
solution.  Upon completion of the test, an aliquot of the absorbing solution
is treated with monopotassium dihydrogen phosphate buffer and Chloramine-T.
The reaction of cyanide and Chloramine-T in the presence of the buffer re-
leases a gas, cyanogen chloride.  For analysis, a portion of this cyanogen
chloride gas is injected into a gas chromatograph equipped with an electron
capture detector (ECD).  External cyanide standards in 1.0 N potassium
hydroxide are used to quantify the results.  The detection limit for this
procedure is less than 0.01 ppm.

SAMPLING SYSTEM

     Two glass impingers in series, with each containing 25 m£ of 1.0 N
potassium hydroxide, are used to collect exhaust samples for analysis of
cyanide.  A flow schematic of the sample collection system is shown in
Figure 1.  The two glass impingers, when maintained at ice bath temperature
(0-5°c), collect 99+ percent of the hydrogen cyanide and cyanogen.  The flow
rate through the impinger is maintained at 4 &/minute by the sample pump.
A dry gas meter is used to determine the total flow through the impinger
during a given sampling period.  The temperature of the gas stream is
monitored by a thermocouple immediately prior to the dry gas meter.  A
flowmeter in the system allows continuous monitoring of the sample flow.
A drier is included in the system to prevent condensation in the pump,
flowmeter, dry gas meter, etc.  When sampling from diesel fueled vehicles,
a heated filter, located between the on-off solenoid valve and the dilution
tunnel, is used to prevent diesel particulate from contaminating the sampling
system.  The filter and line connecting the filter to the dilution tunnel are
heated to 375 °F in order to keep hydrogen cyanide and cyanogen from being re-
tained on the removed particulate.  The Teflon line connecting the heated
filter and the solenoid valve is heated to ~175°F in order to prevent water
from condensing in the sample line.  Several views of the sampling system
are shown in Figure 2.

ANALYTICAL PROCEDURE

     The analysis of total cyanide (hydrogen cyanide and cyanogen) in exhaust
is accomplished with the use of a gas chromatograph equipped with an :electron
capture detector (ECD).  This detector is highly sensitive to halogens and
                                   232

-------
                                                                        Gas Temperature
                                                                        Digital Readout
ro
CO
GO
                On-Off
               Solenoid
                 Valve
                           Flowmeter
    Sample
     Probe
                                                                     Regulating
                                                                       Valve
    Dilute
   Exhaust
  Ice Bath
 Temperature
Digital Readout
    Dry
    Gas
   Meter

  Gas Volume
Digital Readout
                       Figure 1.   Total cyanide sample  collection flow schematic.

-------
              Front View
                                                    Digital
                                                    Readout
                                                   Flowmeter
                                                   Regulating
                                                   Valve
        Close-up of Upper Front
Figure 2.  Total cyanide sampling system
                  234

-------
  Solenoid
Impinger
 Ice Bath
                           Close-up of impingers  (Side View)
                                                        Solenoid
                                                        Pump
                  Rear View




         Figure  2  (Cont'd).  Total cyanide sampling system,
                               235

-------
halogenated compounds.  In this procedure the cyanide  ion (CN~)  is reacted
with Chloramine-T  (sodium paratoluene sulfonchloramide)  to  form cyanogen
chloride (CNCl) which can be detected at low concentrations by the BCD.
Because of the sensitivity of the BCD to the halogenated cyanogen chloride,
cyanide can be detected at low concentrations in exhaust by this procedure.
A detailed description of this procedure follows.  An  analysis schematic for
the procedure is shown in Figure 3.

     During each test cycle a portion of the diluted exhaust is bubbled
through two impingers in series, with each impinger containing 25 m£ of  l.ON
potassium hydroxide.  "The temperature of the impingers  is maintained at
0-5°C by an ice water bath, and the flow rate through  the impinger is main-
tained at 4 Vwinute throughout the test cycle.  Upon  completion of each
driving cycle, the impingers are removed and the coritent of each are trans-
ferred to a separate 30 mJi ploypropylene bottle and capped.   A 1 mS, aliquot
is removed from qne of the bottles and placed in a 5 m& Glass Reacti-vial.
A 2 m£ aliquot of 1.0 Npotassium dihydrogen phosphate buffer is then added
carefully down the side of the vial.  This adjusts the  pH to neutral or
slightly acid.  A 1 mH aliquot Qf Chloramine-T is then  carefully added down
the side of the vial to the buffered solution.  Turbulent addition of this
reagent can cause premature release of cyanogen chloride.   The cap with  a
septum top is immediately screwed tightly into place.   The  resulting solu-r
tion is then set aside for 5 minutes.  This allows the  Chloramine-T to
react completely with the trapped cyanide ion.  The vial is then vibrated for
5 seconds to release cyanogen chloride into the gas phase.   With a gas-tight
syringe a 100 p£ sample of the head space is removed through the septum  top
and immediately injected into the gas chromatograph.  This  procedure is  then
repeated for the second impinger.  Some of the steps in this procedure are
shown in Figure 4.

     A Perkin-Elmer 3920B gas chromatograph with an BCD is  used to analyze
the sample.  A 6' x 1/8" stainless steel column packed  with 100/120 mesh
Porapak Q is used to separate the cyanogen chloride from other compounds
in the sample.  The carrier gas, 95% argon-5% methane,  flows through the
column at a flow rate of 40 n\£/minute.  The column temperature is isothermal
and maintained at 140°C.  Oxygen, carbon dioxide, and water elude from the
column before cyanogen chloride.  The sample peak area  is determined with a
Hewlett-Packard Model 3354 computer system with a remote teletype printout.
The peak area is compared to the peak  area of a standard cyanide ion solution
which is developed in a manner similar to that of the  sample.   Figure 5  shows
the analytical system with gas chromatograph detector,  integrator,  and
recorder.

     This procedure provides a rapid and sensitive method .for analyzing  total
cyanide in exhaust without extensive wet chemical work  up.   The analysis time
is on the order of about 5 minutes after injection into the gas chromatograph.
The sensitivity of the BCD extends the minimum detectable limit to less  than
0.01 ppm cyanide ion with the specified flow rates, absorbing solution vol-
ume, syringe size, vial size, and reagent quantities.   This limit can possibly
be extended by changing these parameters.  The simplicity and rapid data
                                    236

-------
            CVS
     Glass Irapinger
 Excess Reagent
saved as needed
     Aliquot buffered
      with KH2PO4
       Chloramine-T
          added
      CNC1 in head gas
       analyzed in GC
         with BCD
  A/D Converter
             \
       Recorder
 Hewlett-Packard
   3354 GC
 computer system
Figure 3.  Total cyanide  (HCN + C H  ) analysis  flow schematic
                             237

-------

Step 1.  Glass reacti-vial with septum cap
             Step 2.   Aliquot removal
Figure 4.  Various steps in sample collection and
       analysis of total cyanide in exhaust.
                        238

-------
              Step 3.   Reagent addition
               Step 4.  Sample shaking
Figure 4 (Cont1 d).  Various steps in sample collection and
          analysis of total cyanide in exhaust.

                         239

-------
               Step  5.   Head  gas  removal
                Step  6.   Sample  injection
Figure 4 (Cont'd).  Various steps in sample collection and
          analysis of total cyanide in exhaust.
                            240

-------
IVj
                              Fiugre 5.  Total cyanide analytical  system.

-------
turnover makes this procedure ideal for repetitive analysis.  A  gas  chroma-
tograph trace for a cyanide standard is shown in Figure 6.

CALCULATIONS

     This procedure has been developed to provide the user with  the  concen-
tration of total cyanide in exhaust.  The results will be expressed  in yg/m3
of exhaust and ppm.  The equations for determining the concentrations in
pg/m3 and ppm are derived in the following manner.

     The first step is to correct the volume of exhaust sampled  to a stan-
dard temperature, 68°F, and pressure, 29.92"Hg, by use of the equation


                       p      v      P       v -    -'-, •,-;•
                        exp x  exp =  corr x  corrt;
                          T               T
                           exp             corr

       V     = experimental volume of gas sampled in ft
       V     = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
       P     - experimental barometric pressure
       PSXP  = 29.92"Hg
       mcorr
       T     = experimental temperature in °F + 460
       TexP  = 68°'F + 460 = 528°R
        corr
 Solving for V'•    gives:
              corr

                    Pg   ("Hg) x Vgx  (ft3) x 528°R

             corr ~    T    (°R) x 29.92"Hg
                        exp
     The next step converts the volume from cubic feet to cubic meters by
use of the conversion factor; 1 cubic meter is equal to 35.31 cubic feet.


                         P    ("Hg) x V   (ft3) x 528°R
                          3OVT^          _ „__
                     „     GA.IJ          tJAM
             v    vm )  = ——^	£
              C0rr                 x 29.92"Hg x 35.31 ft3/Ht3
                                                                (Equation 1)
     The next step is to find the concentration of total cyanide in
Since the gas chromatograph BCD has a linear response in the concentration
of concern, then the following equation holds.

                      Csam (Ug/m£)   Cstd
                         sam            Astd
                                    242

-------
                                       Liq. Prim
                                        Support
                 C«iing_4U _ce/min. Ar/5» CH. Orr*
  on 13J/12O mnn pnBAPAK
  Run ISO e
  »  dn P*«
  htld «	"C ISO tor
  hddfor
Inlel
Dtttctor
  Hyd  '•"  pug
  Air  UA  psig
  (  I  NA  P»ia  NA    Rolnmnr Hdo.
               C. Huatad ShainU.cs Stujal   T»o«
                       Rotwnvur Rdg.
                       Rotamnir Rdg.  NA
       cc/min
       cc/min
      _ cc/min
Sol tec  Tvo.
  R«cordtr  1  in/miniDMd    1   rnV.F.S.
                     65432
                     Detention time, min.
Figure  6.    Typical trace  for  a  standard.
                              243

-------
      C    = concentration of the sample in yg/m£

      Asam = GC peak area of sample in relative units

      CS , = concentration of the standard in yg/m£

      AS   = GC peak area of standard in relative units
       std


Solving for C    gives:
      3      sam



                           C    (yg/m£)    A

            sam         =  	r	
                                 A . ,
                                  std



      The Csam (yg/m£)  in solution is corrected for any necessary dilution by

 multiplying by the dilution factor,  D.F.



                ,   , n,    C ., (yg/m£)  x A    x D.F.
           C    fyg/m£)     std  ^        sam
            sam         = 	

                                     std



      To obtain the total amount of yg of total cyanide in the absorbing

 solution, the absorbing reagent volume is multiplied by the concentration

 to give:




      yg sample = C    (yg/m£)  x Abs.  Vol. (m£)
                   sam


                  c *.* (yg/m£)x A    x D.F. x Abs. Vol. (m£)
                =  std           sam
                                   Astd
                                                             (Equation 2)
      To obtain yg sample/m ,  Equation 2 is divided by Equation 1 to give:



                 ->    C . , (yg/m£)  x A    x D.F. x Abs. Vol. (m£)
             ,  / 3     std            sam
      yg sample/m  = 	
                               A . , x P    ("Hg)  x 528°
                                std    exp



                      T    x 29.92"Hg x 35.31 (ft3/m3)


                               V    (ft3!

                                e3!P                           (Equation 3)



      To find the concentration of total cyanide (as HCN) in ppm, the density

 of hydrogen cyanide is needed.  At 29.92"Hg and 32°F, one mole of gas oc-

 cupies 22.4 liters.  This volume is corrected to 68°F from the equation



                             V      Vi
                             T
           Vj_  = 22.4£

           T!  = 32°F  4-  460  = 492°R

           V  = volume  at 68°F

           T  = 68°F  +  460  = 528°R
                                     244

-------
Solving for V gives:

              Vn v T     22.4  x 528

                                     = 24.045,
     Since one mole of gas occupies 22.04£ at 68°F, the density can be found
in g/£ by dividing the molecular weight in g/mole by 22.04 £/mole
                          mole.  wt.  g/mole
density (g/£) = m^' ™~  
-------
       1T*
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEY ENTRY KEY CODE rouu^c








010









OiO









030









040









050






2
t
R/S
X
STO 1
R/S
4
6
0
+
RCL 1
f
R/S
X
STO 2
RCL 2
R/S
X
R/S
X
R/S
-'.
R/S
X
STO 3
R/S
RCL 2
X
R/S
X
R/S
V
R/S
X
R/S
RCL 3
+
R/S
1
•
0
3
8
7
X
1
1
2
4
f
R/S
h RTN



02
81
84
71
33 01
84
04
06
00
61
34 01
81
84
71
33 02 .
34 02
84
71
84
71
84
81
84
71
33 03
84
34 02
71
84
71
84
81
84
71
84
34 03
61
84
01
83
00
03
08
07
71
01
01
02
04
81
84
35 22





In Barometric "Hg

In Sample Temp. °F
In Sol. Vol. , mi!.
In Dilution Factor

In Std Cone ug/mJl

In Standard Area

In Sample Area,
Bubbler #1
Out Sara. Cc
Bubbler #1
In Di 1 utic
In Std Cone
Input Std.
In Sample t
Bubbler
Ou-t Sam. Cc
Bubbler #2
Out Cone, V
Output ppm
, pg/m3-
n Factor
, ug/mfc
Area
ire a,
82
me,
., pg/mj
ig CN~/m3
HOI
	

060









070









060









090









100









110


























































...






















































REGISTERS
0123
SO SI S2 S3
A i
i C
4
S4

5 6 7
S5 S6 S7
D
8 9
S8 S9
E 1
Figure 7  (Cont'd).  HP-65 program form.
                 247

-------
Hewlett-Packard Calculations

     In order to insure maximum turnaround in a minimum time period, a
Hewlett-Packard 67 program was developed to calculate the total cyanide
concentration in yg/m3 and ppm from the raw data.  This program is
presented in Figure 7.

Sample Calculation

     Assume exhaust samples were collected in glass impingers for each por-
tion of a three-bag 1975 FTP.  Raw data for these tests are presented in
Fiugre 8.  Calculations were performed using the HP-67 program and manual
calculations.
Manual Calculations for Driving Cycle Cold-FTP

     For Bubbler #1

                 C  . (yg/m£) x A   x D.F. x Abs. Vol.
        _ - , 3    std            sam
     yg CN /m  = 	
                            AstdX  Pexp ("Hg)


                  Texp x 29.92"Hg x 35.31 ft3/m3
                X       528°R x v    (ft3)
                                 exp

               = 5.0 yg/m& x  ISQQ x i x 25 mfc
                       2000 x 29.19"Hg

                x (460° + 70°) x 29.92"Hg x 35.31 ft3/m3
                            528o x 3.453 ftJ


                   986 yg CN~/m3

•Hie concentration in bubbler #2 is calculated in the same manner using the
appropriate dilution factor, standard concentrations, standard area, and
sample area:

    For Bubbler #2
   yaCN~/m3 - 1  ^g/m  x  500  x  i  x  25  ml.
   • *    /m
              x  (460°  +  70°)  x  29.92"Hg x  35.31 ft3/m3
                           528°  x  3.453 ft-3
               131 yg CN~/m3
                                    248

-------
                SWRI PROJECT NO.11-1234  TEST NO.   001
                                           TUNNEL SIZE:   18"   DRIVER:   R.R.
TEST DATE:  11-1O-79 VEHICLE: Practice



                     MILES:    1000
FUEL: EM-237 CVS NO.  3                 	       	      	



SAMPLE COLLECTION BY: D'E'B-    CHEMICAL ANALYSIS BY:  ' ' 'CALCULATIONS BY:



GENERAL COMMENTS:
                Test No.
ro
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Absorb . Rea . Vol . , m£
Dilution Factor, Bubbler #1
Std. Cone UgCN~/m& Bub. #1
Std. Area - Bubbler #1
Sample Area - Bubbler #1
Sample Cone pgasT/m3 , Bub #1
Dilution Factor, Bubbler #2
Std. Cone ygCN~/m£ Bub. #2
Std. Area - Bubbler #2
Sample Area - Bubbler #2
Sample Cone ugCN~/m^,Bub#2
Total Cone. ygCN~/m3
Total Cone, ppm HCN
Cold FTP
3.453
29.19
70
25
1
5
2000
1500
986
1
1
1000
500
131
1117
1.03
Hot FTP
3.486
28.66
75
25
5
2
1500
1800
3213
1
1
2000
1000
134
3347
3.09
SET-7
3.508
29.33
80
50
10
1
3000
2500
4374
1
2
3000
500
175
4549
4.20
HFET
1.926
29.40
85
50
2
1
5000
4500
1732
2
0.5
2000
200
96
1829
1.69
NYCC
1.525
29.10
90
25
1
2
10,000
9000
1115
I
I
3000
1900
392
1508
1.39
Backgroui
15.826
29.04
77
25
1
1
58
4000
2000
1
0.5
2000
500
7
37
0.03
d





                                  Figure 8.  Raw data sheet for total cyanide.

-------
The concentrations from the two bubblers can be added  for  a total concentration:

     Total yg  CN~/m3 = cone  (bubbler #1) + cone  (bubbler  #2)
                      = 986 yg CN /m  + 131 yg CN /m

                      = 1117 yg CN~/m3

       ppm CN~(as HCN)  = yg HCN/m  -r density yg/m£
                                   / n   Mol. Wt. (HCN) X 100
                         density yg/mJ6 = 	--4 ^ Q4j
                                             24J

                        Mol.  Wt.  HCN = 27.026 g/mole

                                  .  „   27.026 x 100   nn_. ,,  . „
                        density yg/m£ = 	-. -	= 1124 yg/mJ6
                                           ^4r • V/4


                        yg HCN/m3 =  yg CN~/m3 x 1.039 yg HCN/yg CN~

                                  =  1117 x 1.039 = 1161

                     ppm CN'(HCN) =  1161 T 1124 = 1.03

Note:  The values used in these calculations are picked from a range of tem-
peratures, standards, dilution factors, etc. to validate the calculations
and may not be representative of expected raw data in all cases.  These cal-
culations are presented to confirm that manual and HP-67 calculations give
the same results.  This was confirmed on six sets of calculations.

LIST OF EQUIPMENT

     The equipment required in this  analysis is divided into three basic
categories:  sample acquisition,, sample preparation, and sample analysis.
Manufacturer, stock number and any pertinent descriptive information are
listed.

Sample Acquisition

     1.   Sample pump, Thomas model  106 CA18, capable of free flow capacity
          of 4 £/minute.

     2.   Glass impingers, Ace Glass Products, catalog no.  7530-11 29/42
          bottle joints, 18/7 arm joints

     3.   Flowmeter, Brooks Instrument Division, Model 1555, Tube size
          R-2-15-C, graduated 0-15,  sapphire float, 0-5 Vminute range.

     4.   Regulating valve, Nupro 4MG, stainless steel
                                     250

-------
    5.   Dry gas meter, American  Singer Corporation,  Type AL-120, 60 CFH
         capacity

    6.   Teflon tubing, United  states  Plastic Corporation,  1/4" OD x 1/8" rn
         and 5/16" OD x 1/8"  ID

    7.   Teflon Solenoid Valve, The Fluorocarbon Company, Model DV2-144NCA1

    8.   Miscellaneous Teflon nuts, ferrules,  unions, tees, clamps, connec-
         tors , etc.

    9.   Drying tube, Nalgene Corporation,  10 cm length x 1/2 in. dianeter

   10.   Digital readout for  dry  gas meter

   11.   Miscellaneous electrical switches,  lights, wirings, etc.

   12.   Six channel digital  thermometer, Analog Devices, Model #2036/J/1.

   13.   Iron/Constantan type J single  thermocouple with 1/4" OD stainless
         steel metal sheath,  Thermo Sensors  Corporation

   14.   Stainless steel heated filter assembly -  7 cm; Scott, capable of
         temperature to 204°C,  includes 2 heaters, adjustable thermostat
         switch, stainless  steel  insulated  covers  and sample bypass solenoid
         valves

   15.   Glass microfiber filter  discs, Reeve Angel 934-AH, Whatman, 7 cm
         diameter

   16.   Flexible heavy insulation heating  tape, Briskeal®, width-1/2 inch,
         length-48 inches

   17.   Temperature Controller,  Athena, 100-600°F

   18.   Heated TFE Teflon  hose,  Technical  Heaters, Inc., 51 x 1/4", tem-
         perature limit 400°F.

Sample Preparation

     1.   Glass gas syringe, Teflon tipped plunger, 100 yl,  Pressure-Lok
         Series A-2, Alltech  Associates

     2.   Glass Reacti-vials,  5  m&, Pierce Chemical Company

     3.   Class A, 1 mJl volumetric pipets

     4.   Class A, 2 m£ volumetric pipets

     5.   Class A, 25 m& volumetric pipets
                                    251

-------
     6.   Class A, 50 mi volumetric flask

     7.   Class a, 100 mi volumetric flask

     8.   Class A, 250 mi volumetric flask

     9.   Class A, 500 mi volumetric flask

    10.   Class A, 1000 mi volumetric flask

    11.   Vortex-Genie, Scientific Industries, Inc. Model K-550-G

Instrumental Analysis

     1.   Perkin-Elmer Model 3920B gas chromatograph equipped with a lin-
          earized electron capture detector  (BCD)

     2.   Soltec Model B-281 1 mv recorder

     3.   Hewlett-Packard Model 3354 gas chromatograph computer system
          with remote teletype printout

LIST OF REAGENTS

     This procedure requires the sample collection In glass iiopingers using
a 1.0 N potassium hydroxide absorbing reagent.  After collection, a buffer
potassium phosphate monobasic is added to control the pH followed by
Chlonnaine-T to convert the CN~ to cyanogen chloride.  Potassium cyanide is
used as the CN~ standard in 1.0 N KDH.  The reagents are listed below along
with the manufacturer and quality.

     1.   Potassium phosphate monobasic, formula -weight = 139.09, chemical
          formula = KH2PO4, ACS Analytical Reagent Grade, crystals, Mallinc-
          krodt Code 7100.

     2.   Potassium hydroxide, formula weight = 56.11, chemical formula -
          KOH, ACS Analytical Reagent Grade, pellets, Mallinckrodt Code 6984

     3.   Potassium cyanide, formula weight = 65.12, chemical formula KCN,
          ACS Analytical Reagent Grade, granular, Mallinckrodt Code 6881

     4.   Chloramine-T (sodium para-toluene sulfonchloramide trihydrate),
          formula weight = 282.70, chemical  formula = p-CI^Cgl^SC^NClNa'ai^O,
          Assay (by titration) 96% minumum, Eastman, crystals, Eastman
          Code 1022

PREPARATION OF REAGENTS

Primary Standard - the primary standard is prepared by dissolving 0.602
grams of KCN in 500 mi of 1.0 N KOH.  This is equivalent to 500 ppm HCN
(500 tig HCN/m£) or a 481 ppm CN~ (481 yg CN-/m£).  Additional standards are
prepared from the primary standard.  A typical dilution to prepare a 0-10
pg CN~/m£ calibration curve is as follows:
                                    252

-------
    m£ of 481  yg CN /m£      Final Diluent      QJ- concentration,
     Primary Standard         Volume, m£             yg/CN~/m£

         1.000 m£                 50.0 m£               9.62
         4.000 m£                250.0 m£               7*.70
         1.000 m£                100.0 m£               4.8!
         3.000 m£                500.0 m£               2.89
         1.000 m£                250.0 m£               1.92
         1.000 m£                500.0 m£               0.96

Buffer Solution - A 1.0 M KH2PO4 buffer solution is prepared by dissolving
13.609 g  KH2PO4 in 100 m£ of deionized H20.  The buffer solution should be
prepared  daily.

Absorbing Reagent - The absorbing reagent is a l.ONKOH solution.  This
solution  is  prepared by dissolving 56.11 grams of KOH in 1000 m£ of deionized
water.

Chlormaine-T -  The Chlormaine-T converts the -CN~ to CNC1.  This reagent is
prepared by  dissolving 250 mg in 100 m£ of deionized water.  This reagent
is the most  critical in this procedure and should be prepared daily.
                                     253

-------
REFERENCES

Braker, W. and Mossman, A.L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N.J., 1971, pg. 301

Epstein, J., Anal, Chem., Vol. 19, pg. 272, 1947

Collins, P.P., Sarji, N.M., and Williams, J.F., Tobacco Sci., Vol. 14,
pg. 12, 1971

Collins, P.F., Sarji, N.M., and Williams, J.F., Breitrage zur Tabakforschung,
Vol. 7, pg. 73, 1973

Sekerka, I. and Lechner, J.F., Water Res., Vol. 10, pg. 479, 1976

Ito, C., Sakamoto, H. and Kashima, T., Kyoritsu Yakko Daigaku Kenkyu Nempo,
Vol. 18, pg. 29, 1973

Frant, M.S., Ross Jr., J.W., and Riseman, J.H., Anal. Chem., Vol. 44
pg. 2227, 1972

Riseman, J.H., Am. Lab., Vol. 4, pg. 63, 1972

Fleet, B., and von Storp, H., Anal. Chem, Vol. 43, pg. 1575, 1971

Fleet, B., and von Storp, H., Anal. Lett., Vol. 4, pg. 425, 1971

Conrad, F.J., Talanta, Vol. 18, pg. 952, 1971

Vickroy, D.G., and Gaunt Jr., G.L., Tabacco Sci., Vol. 174, pg. 50, 1972

Burke, James, "Ion Selective Electrode-Direct Technique, "Ford Motor Company
procedure for measuring HCN in dilute automotive exhaust, 1976

Llenado, R.A. and Rechnitz, G.A., Anal. Chem., Vol. 43, pg. 1437, 1971

Nebergal, W.H., Schmidt, F.C., and Holtzclaw Jr., H.F., College Chemistry
with Qualitative Analysis, 5th edition, D.H. Heath and Co., Lexington, Mass.,
1976, pg. 684

Woolmington, K.G., J. Appl. Chem., Vol. 11, pg. 114, 1961

Isbell, R.E., Anal.  Chem., Vol. 35, pg. 255, 1963

Myerson, A.L. and Chludzinski Jr., J.J., J. Chromatog. Sci., Vol. 13, pg.
554, 1975

Valentour,J.C., Aggarwal, V., and Sunshine, I., Anal. Chem., Vol. 46, pg.
924, 1974

Runge,  H., Z. Anal.  Chem., Vol. 189, pg. Ill, 1962

                                    254

-------
           APPENDIX C





INDIVIDUAL HYDROCARBON PROCEDURE
              255

-------
           THE MEASUREMENT OF INDIVIDUAL HYDROCARBONS  IN EXHAUST


     This procedure was developed to measure individual  hydrocarbons in di-
lute automotive exhaust.  The term, individual hydrocarbons  (IHC)  is used to
define the collection of compounds:  methane (Cttg.) , ethane  (C2Hs) ,  ethylene
(C2H4) , acetylene (C2H2) , propane (C3H8) , propylene (€3%), benzene  (CQHQ) ,
and toluene (07%).  Dilute exhaust is collected in Tedlar bags  during a
test cycle and analyzed with a gas chromatographic system containing four
separate columns and a flame ionization detector  (FID).   The peak  areas are
compared to an external calibration blend and individual hydrocarbon concen-
trations are analyzed with a Hewlett-Packard 3354 computer system.   The
analysis flow schematic is shown in Figure 1.

ANALYTICAL SYSTEM

     The analysis for individual hydrocarbons is conducted with  a Varian
Aerograph Series 1400 gas chromatograph using a flame  ionization detector
(FID).  Four separate packed columns are used to resolve these individual
compounds.  An elaborate system of timers, solenoid valves, and  gas sampling
valves are used to direct the flow of the sample through the system.   The
actual analytical system is shown in Figure 2.

     The first two columns are used to resolve air, methane, ethylene,  ethane,
acetylene, propane and propylene, respectively; while  columns III  and IV re-
solve benzene and toluene.  Column I consists of an 8' x 1/8" stainless steel
tube packed with Porapak Q 80/100 mesh.  This column is  primarily  used to re-
solve methane from air.  It undergoes temperature programming from 25°C to
100°C at 12°/min.  Column II consists of a 4' x 1/8" Teflon column  packed
with 35/60 mesh type 58 Silica gel.  C2 and C3 hydrocarbons are  resolved
with this column.  It is held isothermal at room temperature  (20°C).   The
third column is used to resolve benzene from the other aromatics, paraffins,
olefins, and acetylenes.  It consists of a 15' x 1/8"  stainless  steel column
packed with 15 percent 1, 2, 3-tris (2-cyanoethoxy) propane on 60/80 mesh
Chromosorb PAW.  This column is held isothermal at 100°C at the  end of the
temperature program sequence.  Column IV is a 2' x 1/8"  stainless  steel tube
packed with 40 percent mercury sulfate  (HgSO^ and 20  percent sulfuric acid
(H2S04)  on Chromosorb W.  This column resolves benzene and toluene from the
oxygenated hydrocarbons such as aldehydes and ketones.   It is also held iso-
thermal at room temperature for the entire analysis sequence.  All samples
pass through a 6" x 0.01" capillary restrictor before  entering the detector.
Helium is the carrier gas with a column flow of 52 m£/minute.

     The temperature program sequence is accomplished  with the oven of the



                                    256

-------
    Vehicle
        CVS
  Tedlar Bags
        Gas

    Ch roma t o g r aph
A/D Converter
      Recorder
     HP 3354
  Computer System
                                                       Teletype
                                                       Printout
Figure 1.  The analysis flow  schematic for individual hydrocarbons.

                                257

-------
I 3
01
00
                             Figure 2.   Analytical system for individual hydrocarbons.

-------
 gas chromatograph.   Columns I and III are in this oven although Column I is
 the only one used during the temperature program.  The gas sampling valves
 are contained  in a Bendix Valve Oven.  The temperature is maintained at
 100°C.  Columns  II and IV are external to this oven for isothermal room tem-
 perature operation.

      Samples as  well as backgrounds are collected in Tedlar bags during the
 driving cycle.   The sample is purged through two 10 m£ sample loops for
 four  (4) minutes.  Samples from diesel fueled vehicles are passed through an
 ice trap before  entering the sample loops.  The ice trap removes high mole-
 cular weight compounds that can interfere with later analyses.  The ice trap
 consists of 8  feet of 1/4 inch stainless steel tubing submerged in an ice
 bath.  The initial configuration of the analytical system is shown in Figure
 3. Upon injection, gas sampling valve A is activated by solenoid valve G,
 the temperature  program sequence is started, and the first timer begins to
 count 680  seconds.  The temperature program sequence for Columns I and II
 starts  at  25°C and increases at 12°/min to a final temperature of 100°C.
 Columns I  and  III are held isothermal at this temperature for the remainder
 of the  analysis.  The configuration of the analytical system is shown in
 Figure  4.   The sample in the first loop passes through Columns I and II and
 into  the detector.  The peaks (in the order of elution) are air, methane,
 ethylene,  ethane, acetylene, propane and propylene.  After 670 seconds, the
 second  step begins with solenoid H activating gas sampling valve B.  The
 second  timer begins to count down 120 seconds.  At this time, the sample
 trapped in the second 10 m£ sample loop is channeled through Column III.
 The analytical system configuration is shown in Figure 5.  After 120 seconds,
 step 3  begins.  The third timer starts counting down 480 seconds and gas
 sampling valves  C and J are activated by solenoid valve E.  Columns I and
 II are  backflushed through a capillary restrictor to the vent and Columns
- Ill and IV are directed to the detector.  The configuration is shown in
 Figure  6.   After 480 seconds, solenoid valve F activates gas sampling valve
 D.  Column III is backflushed through a capillary restrictor to the vent.
 The final  configuration is shown in Figure 7.  The last two peaks in the
 order of elution  are benzene and toluene.  Upon elution of the last peak,
 the system is  reset to the initial position.

      A time/temperature system operation sequence is presented in Figure 8.
 The solid  line on this graph represents the gas chromatograph oven tempera-
 ture  during the  temperature program sequence.  The time at which each step
 begins  is  also represented on the graph.

      Figures 9 through  13  illustrate  a simplified version  of the  flow of  the
 carrier gas and  sample  through the  gas sampling valves.  Figure 9  shows the
 configuration of the  gas sampling valves in the sample loop purge position.  The
 sample  is pumped out  of the  sample  bag and through the sample loops.  Column
 III is  flushing  and Column IV is backflushing to the  vent  and Columns I  and
 II are  directed  to  the  detector.  At  the  start of an injection, the position
 of gas  sampling  valve A changes and the trapped sample is  directed to Columns
 I  and II.  Columns  III  and IV remain  in the flushing mode.   Figure 10 illus
 trates  the analytical configuration upon injection.   Step  2  begins with  the
                                     259

-------
                  PRESSURE
ro
cr>
                                   SOLENOID VALVE
SAMPLE IN


SAMPLE LOOP A
                                                                                               HELIUM

                                                                                                   CAPILLARY RESTRICTOR
                                                                                                  SEISCOR VALVE
                                                                                   MifiJ COLUMN IV
                                   Figure  3.  Initial  analytical system  configuration.

-------
ro
                 PRESSURE
                                   SOLENOID VALVE
SAMPLE IN

SAMPLE LOOP A
                                                                                              HELIUM
                                                                                                  CAPILLARY RESTRICTOR
                                                                                                 SEISCOR VALVE
                                                                                        COLUMN IV
                          Figure  4.   Step 1 solenoid G activation  of gas  sampling valve A.

-------
ro
CT»
ro
                 PRESSURE
                                 SOLENOID VALVE
      SAMPLE IN


      SAMPLE LOOP A
                                                                                             HELIUM

                                                                                                 CAPILLARY RESTRICTOR
                                                                                                SEISCOR VALVE
                                                                        *••••••••••Ja«••••••!•»•••*••


                                                                               liQMfti COLUM
COLUMN IV
                          Figure 5.   step 2 solenoid  H activation  of gas  sampling valve B.

-------
ro
CTl
CO
               PRESSURE
                                SOLENOID VALVE
 SAMPLE IN


 SAMPLE LOOP A
                                                                                            HELIUM
                                                                                                CAPILLARY RESTRICTOR
                                                                                               SEISCOR VALVE
                                                                       • ••••••*•••J«•••**•••••••*••• »•


                                                                              [MM]  COLUMN
IV
                        Figure 6.   Step  3  solenoid E  activation of gas sampling valves C and J.

-------
                  PRESSURE
ro
en
                                  SOLENOID VALVE
SAMPLE IN


SAMPLE LOOP A
                                                                                              HELIUM

                                                                                                  CAPILLARY RESTRICTOR
                                                                                                 SEISCOR VALVE
                                                                                              IV
                                    Figure  7.  Final analytical system configuration.

-------
en
                 120,-
                 100  -
                  80 -
              S   60
              111
                  40
                  20
                                                                    STEP 2  STEP 3
                                                                     STEP 4
ISOTHERMAL OPERATION-
 STEP 1
                                                            12        15
                                                            Time,  Minutes
                                                                               18
                                                                                        21
                                                                  24
                                                                                                            27
                                                                                                                     30
                                            Figure  8.   System operation  sequence.

-------
                                                                                                     FID DETECTOR
CT)
CT>
            SAMPLE IN
               HELIUM
           SAMPLE OUT
              HELIUM
                                                                                                           RESTRICTOR
                             Figure 9.   Sample purge position prior to sample injection.

-------
                                                                                                        FID DETECTOR
po
            SAMPLE IN
               HELIUM
            SAMPLE OUT
                                                                                                               RESTRICTOR
               HELIUM
                                         Figure  10.  Step 1 sample loop A injected.

-------
                                                                                                      FID DETECTOR
ro
01
oo
           SAMPLE  IN
              HELIUM
          SAMPLE OUT
              HELIUM
                                                                                                             RESTRICTOR
                                        Figure 11.   Step 2 sample loop B injected.

-------
                                                                                                        FID DETECTOR
ro
               SAMPLE IN
                 HELIUM
              SAMPLE OUT
                 HELIUM
                                                                                                               RESTRICTOR
                             Figure 12.   Step  3  simultaneous solenoid C and  J activation.

-------
                                                                                                      FID DETECTOR
po
—i
o
            SAMPLE IN
               HELIUM
           SAMPLE OUT
                                                                                                            RESTRICTOR
              HELIUM
                                      Figure 13.   Step 4 backflushing of column III.

-------
trapped sample in sample loop B routed  to Column  III.   Figure  11 demonstrates
Step 2.  Gas sampling valves C and J are simultaneously switched to backflush
Columns I and II and direct the sample  in sample  loop  B through Column III
and Column IV and then to the detector  in Step 3.   Finally,  step 4 results
in the backf lushing of Column III to the vent.  The system remains in this
configuration until it is reset to the  initial purge position.  A sample
chromatogram for the calibration blend  of individual hydrocarbons is shown
in Figure 14.

CONTROL SYSTEM

     The control of the five Seiscor gas sampling valves is  accomplished by
ATC timers and ASCO electric solenoid valves.   With the solenoid valve in
the de-energized configuration, the gas sampling  valve is in one position.
An electrical impulse from the timer energizes the solenoid  valve.  The
pressure difference created in the gas  sampling valve  changes  the position.
Figure 15 illustrates the de-energized  and  energized configuration for sole-
noid valve G and gas sampling valve A.  This accomplishes the  complicated
column flow sequence required for this  analysis.   The  system repeatability
is ± 1.8 percent.

CALCULATIONS

     The concentration of the individual hydrocarbons aro compared to a cali-
bration blend with known concentrations of  each of the components.  The peak
areas  are determined with a Hewlett-Packard Model 3354 computer system and
printed out on a remote teletype unit.  The concentration of the sample is
determined with the equation:

     Astdn _ Asamn
     Cstdn   Csamn

where  Astd = the peak area of component n in the  calibration blend
       Cstd = the concentration of component n in  the calibration blend
       Asam = the peak area of component n in the  sample
       Csam = the unknown concentration  of component n  in the sample
       n    = the component of interest  (i.e.,  methane, ethylene, ethane
             acetylene, propane, propylene, benzene, or toluene)

If the equation is solved of Csam, the  result is:

                     Cstd
                                                                       ,
               Astdn

Example  1:
     A bag  sample was  taken from the exhaust stream of a vehicle during a
driving  cycle.   The peak area of the calibration blend for methane was 11893
area counts with a concentration of 13.616 ppmC.  The peak area of methane
in the sample was 8593 area counts.  Calculate the concentration of methane
in the exhaust.
                                    271

-------
ro
»-sJ
ro
                      22  21  20   19  18  17  16  15  14  13  12  11  10   9   8   7   6   5   4   3   2   10
                                                       Retention time, min.
                              Figure 14.  Calibration blend for  specific hydrocarbons.

-------
                   de-energized configuration
                     energized configuration
Figure 15.   Solenoid valve G with gas  sampling  valve A
                             273

-------
     Csamn -- Adtdn

          ,  ^ •• %    (8593) (13.616)
     Csam (methane)  = - (11893) -

     Csam (methane)  = 9.84 ppm C

Example 2:

     In the sample analysis, the peak area for toluene was  10973  area counts.
The peak area for the calibration blend was 18546 area counts  for toluene
with a concentration of 21.272 ppm C.  Calculate the concentration of toluene
in the sample.

             As am., x Cstd-
     C5amn=            ^
                  .    (10973) (21.272)
     Csam (toluene)  =
     Csam (toluene) = 12.59 ppm C

Note:  "These sample calculations are presented as a example  only  and are not
       necessarily representative of expected values in exhaust.

EQUIPMENT

     This analysis is performed using a gas chromatograph equipped with a
flame ionization detector (FID) .  The equipment required is  divided into two
categories.  The major items in each category are listed below:

Gas Chromatograph Detection

     1.   Varian Aerograph Series 1400 Gas Chromatograph

     2.   Leeds and Northrup 1 mv Recorder

     3.   Hewlett-Packard Model 3354 Computer System

     4.   Hewlett-Packard Model 1865A A/D, Converter

Control Console System

     1.   Seiscor Model VIII Gas Sampling Valve

     2.   ATC Timers, Model 325A346A10PX

     3.   ASCO 4-way Midget Solenoid Valve, No. 8345B
                                    274

-------
 4.   Thomas  Sample  Pump, Model 106 CA18

 5.   Brooks  Flowmeter,  R-2-15-AAA with glass float, 0-15 Scale

 6.   Miscellaneous  stainless steel and brass nuts, ferrules,  unions,
     tees, connectors,  etc.

 7.   Miscellaneous  stainless steel, copper, and Telfon tubing

 8.   Bendix  Valve Oven

 9.   Column  I,  8' x 1/8" SS, 80/100 mesh Porapak Q

10.   Column  II, 4'  x 1/8" Teflon, 36/60 mesh Type 58 Silica Gel

11.   Column  III, 15' x 1/8" SS, 15 percent 1, 2, 3-tris (cyanoethoxy)
     Propane on 60/80 mesh Chromosorb PAW

12.   Column  IV, 2'  x 1/8" SS, 40 percent Hg;so4 on Chromosorb  W

13.   Miscellaneous  electrical switches, wiring, lights, etc.
                                275

-------
 REFERENCES

 Altshuller, A.P. and Bufalini, J.J. Environ.  Sci.  Tech.,  Vol. 5, pg. 39,
 1971.

 Federal  Register, Vol.  37, No. 221, 24270-77  Nov.  1971.

 Klosterman, D.L. and Sigsby, J.E., Environ. Sci0 Tech., Vol.  1,  pg 309,
 1967.

 Rasmussen, R.A. and Holdren, M.W., Chromatog,  Newslett.,  Vol. 1, No. 2,
 pg.  31,  1972.

 Dimitriades, B. and Seizenger, E.D., Environ.  Sci.  Tech.,  Vol.  5, No.  3,
 pg.*223,  1979.

 Purcell,  J.E. and Gilson, C.P-, Chromatog. Newslett., Vol. 1, No. 2,
 pg  45,  1972.

 Stuckey,  C.L., J. Chromatog. Sci., Vol. 7, pg. 177,  1969.

 Papa,  L. J., Dinsel, D.L., and Harris, W. C., J. Gas  Chromatog.,  Vol. 6,
 pg.  270,  196,8.

Black, F.M.,  High,  L.E.  and Sigsby, J.E.,  J. Chromatog.  Sci.,  Vol. 14,
pg. 257,  May  1976.

Private communication between Mobile Source Emissions Research Branch,  ESRL-
EPA, Research Triangle Park and Southwest  Research  Institute.
                                    276

-------
       APPENDIX D




ORGANIC AMINES PROCEDURE
         277

-------
                THE MEASUREMENT OF ORGANIC AMINES IN EXHAUST
     The organic amines that are included in this analysis are:  monomethyla-
mine, dimethylamine, trimethylamine, monoethylamine, diethylamine,  and
triethylamine.  Dimethylamine and monoethylamine are not resolved  from  each
other under normal operating conditions and are reported together  as C2H7N.
The measurement of organic amines in exhaust is accomplished^ by bubbling
the exhaust through glass impingers containing dilute sulfuric acid.  The
amines are complexed by-the acid to form stable sulfate salts which remain
in solution.  The exhaust sample is collected continuously during  the test
cycle.  For analysis, a portion of the sulfuric acid solution is injected
into a gas chromatograph equipped with an ascarite  loaded pre-column and a
nitrogen phosphorus detector (NPD).  External amine standards in dilute
sulfuric acid are used to quantify the results.  Detection limits  for this
procedure are on the order of 0.002 ppm in dilute exhaust.

SAMPLING SYSTEM

     A glass impinger containing 25 m£ of 0.01N sulfuric acid is used to
collect exhaust samples for the analysis of the organic amines.  A  flow
schematic of the sample collection system is shown  in Figure 1.  The single
glass impinger is sufficient to collect 99+ percent of the organic  amines.
The temperature of the impinger is maintained at 0-5°C by an ice water  bath,
and the flow rate through the impinger is maintained at 4Jl/minute by the
sample pump."  A dry gas meter is used to determine  the total flow through!
the impinger during a given sampling period.  The temperature of the gas
stream is monitored by a thermocouple immediately prior to the dry  gas  meter.
A drier is included in the system to prevent condensation in the pump,  flow-
meter, dry gas meter, etc.  The flowmeter in the system allows continuous
monitoring of the sample flow to insure proper flow rates during the
sampling.  When sampling from diesel fueled vehicles, a heated filter,
located between the solenoid valve and the dilution tunnel, is used to
prevent diesel particulate from contaminating the sampling system.  The
filter and the line connecting the filter to the dilution tunnel are heated
to 3,75°F in order to keep the organic amines from being retained on the
removed particulate.  The Teflon line connecting the heated filter  and  the
solenoid valve is heated to ~175°F in order to prevent water from  condensing
in the sample line.  Several views of the sampling  system are shown in
Figure 2.

ANALYTICAL PROCEDURE

     The analysis of the organic amines (monomethylamine, dimethylamine,
trimethylamine, monoethylamine, diethylamine, and triethylamine) is


                                      278

-------
ro
•vj
vo
    Sample
     Probe
                                                                         Gas Temperature
                                                                         Digital Readout
                  Heated
                  Filter
              On-Off,
             Solenoid
                            Valve
Plowmeter
V
                                                          Sample
                                                           Pump
   Heated
    Lines
             Dilution
             Tunnel
     Dilute
    Exhaust
                                        Ice  Bath
                                      Tempe rature
                                    Digital  Readout
                                                                    Regulating
                                                                     •••' Valve
 Dry
 Gas
Meter
                                                                                             HTalakHBl
               Gas Volume
              Digital Readout
                       Figure 1.   Organic amines sample collection flow schematic.

-------
       Front View
                                            igital
                                           Readout
                                           Flowmeter
                                           Regulating
                                           Valve
Close-up of Upper Front
              280

-------
Solenoid
Impinger
Ice Bath
                                                         Solenoid
                                                         Filter

                                                         Ice Bath
                                                         Drier
                                                         Dry Gas Meter
                                                         Pump
            Figure 2  (Cont'd).  Organic amines sampling system,
                                281

-------
accomplished by trapping the amines in sulfuric acid and  analyzing the sample
with a gas chromatograph equipped with an NPD.  The NPD is  highly sensitive
to organic nitrogen compounds and relatively insensitive  to inorganic nitrogen
compounds.  The analysis flow schematic for the organic amines  is shown in
Figure 3.  .A detailed description of the procedure follows.

     For the analysis of the organic amines, dilute exhaust is  bubbled through
a glass impinger containing 25 m£ of 0.01N sulfuric acid.   Upon completion of
each driving cycle, the impinger is removed and the contents are transferred
to a 30 m& polypropylene bottle and capped.  The amines,  as their sulfate
salts, can be stored in solution for long periods of time without decomposition.

     A Perkin-Elmer 3902B gas chromatograph equipped with an ascarite loaded
pre-column, a Teflon interface, and a nitrogen phosphorus detector (NPD) is
used to analyze the sample.  A 10 y£ portion of the sample  is injected into
the gas chromatograph  (GC).  In the ascarite pre-column,  Figure 4, the amines
are released from their sulfate salts into the GC column.   The  column is a
6" x 4 mm glass column containing Carbopack B coated with 4 percent Carbowax:
20 M and 0.8 percent KOH.  The column effectively separates the amines, with
the exception of ethylamine and dimethylamine, which are  reported together
as total C2H_N.  The carrier gas is helium which flows through  the column
at a rate of 30 mi/minute.  The column temperature is 130°G for 4 minutes
and then programmed to 170°C at a rate of 32° a minute.   In a chromatogram
of a standard sample containing all six of the amines, Figure 5,  the first
peak is monomethylamine, followed by the combined peak of dimethylamine-
and monoethylamine, C H N, and then by peaks of trimethylamine,  diethylamine,
and triethylamine.  To quantify the results, the sample peak areas are
compared to peak areas of standard solutions.  Figure 6 showa the analytical
system with gas chromatograph, detector, A/D converter, and recorder.

CALCULATIONS

     The procedure has been developed to provide the user with  the concentra-
tion of the organic amines (monomethylamine, total dimethylamine and mono-
ethylamine as C2H7N, trimethylamine, diethylamine, and triethylamine)  in
exhaust.  The results will be expressed in yg/m3 of exhaust and ppm for each
of the amines.  The equations for determining the concentrations in yg/m3
and ppm are derived in the following manner.

     The first step is to correct the volume of exhaust sampled to a standard
temperature, 68°F, and pressure, 29.92"Hg, by use of the  equation*

                     P    x V       P     x V
                      exp     exp _  corr	corr
                        T               T
                         exp             corr
                                     282

-------
        CVS
           \
        Glass
      Impinger
 Unused Sample
saved as needed
     Sample analysis
  in gas chromatograph
with ascarite pre-columr
        and NPD
 A/D Converter
      Recorder
Hewlett-Packard
     3354
Computer System
    Figure 3.  Organic  amines  analysis flow schematic
                            283

-------
Figure 4.  Ascarite pre-column,
             284

-------
Sample  injection
1.4 ppm methylamine
1.6 ppm ethylamine  +
dimethylamine
1.5 ppm trimethylamine
0.9 ppm diethylamine
0.9 ppm triethylamine
                            Date January
                       Operator
                       4mm    l.D.
Instrument
Column
                %wt.Carbowax  2QM+0.8%Liq. Phase
            mesh  Carbopack B      » KOH Support
         NA   °Cusina  30  cc/min. Helium   Carrier
 1  GO  psig
held®   130
                     Rotameter Reading
        /min. Held for
           min. (other)	
                Heated-Glass  Lined
                          min.. Prog to	 C at
  held for
Inlet  20
         psig	NA	Rotameter Rdg.	  3
         psig	NA    Rotameter Rdg. _  10 Q
         psig	 Rotameter Rdg.
                      1   mV.F.S.
                            ul net
Recorder  1   in/min speed
Injection  10  ul indicated
  Sampling Device   10  Ul
                     ~8	       6.4         2
                   Retention time, min.
      Figure  5.  Chromatogram of amine  standard.

                             285

-------
•
 '
                                 Figure  6.   Total  amine  analytical  system.

-------
         vexp  = experimental  volume of gas sampled in ft3
         vcorr = volume of  gas sampled in ft3 corrected to 68°F  and  29.92"Hg
         pexp  = experimental  barometric pressure
         Pcorr = 29.92"Kg
         Texp  = experimental  temperature in °F +• 460
         Tcorr = 68°F  +  460  = 528°R
    Solving for Vcorr  gives:

                 Pexp  ("H9)  x Vexp (ft3)  x 528°R
         v
          corr        Tex  (OR)  x 29.92"Hg
    The next step  converts the volume from cubic feet to cubic meters by
use of  the conversion  factor;  1 cubic meter is equal to 35.31 cubic  feet.

                    pexp("H9)* vexp (ft3) x 528°
         vcorr(m3) =—
                    Texp x 29.92"Hg x 35.31 ft3/m3
                                                                (Equation 1)

    The next step  is  to find the concentration of each of the amines in
yg/m£.  Since the gas  chromatograph NPD has a linear response in the con-
centration of concern, then the following equation holds.
                      Csam (ug/mA)     Cstd (ug/m£)
                                          Astd

     Csam = concentration of the sample is yg/m£
     Asam = GC  Peak area of sample in relative units
     cstd = concentration of the standard in yg/nd
     Astd = GC  Peak area of standard in relative units

     Solving for C    gives:

                             cstd
     The Csam(yg/irJl)  in solution is corrected for any necessary dilution by
multiplying by the dilution factor, D.F.

                             Cstd  (ug/m£) x Asam x D.F.
                           = - -- — -
                                      Astd
                                     287

-------
     To obtain the total amount in ug of each amine in the absorbing solution,
the absorbing reagent volume  is multiplied by the concentration to give:

     Ug sample = Csam  (yg/mA)  x Abs.  Vol.  (mjl)

                 Cstd  (yg/m&)  x Asam  x D.F.  x Abs.  Vol. (m£)
                                 Astd

                                                               (Equation 2)


     To obtain yg  sample/m3, Equation 2  is divided by Equation 1 to give:

                 Cstd  (ug/m£) x Asam  x o.F.  x Abs.  Vol. (mfc)
     Ug samp/m3 =
                                            528°

                 T    x 29.92"Hg x  35.31  (ft3/m3)
                  -exp
               x -
                          Vexp  (ft3)
                                                               (Equation 3)
     To find the concentration of each amine  in ppm,  the densities of the
amines are needed.  At 29.92"Hg and 32°F, one mole  of gas occupies 22.4
liters.  This volume is corrected to 68°F from the  equation

                          V _ Vj
                          T ~ T]_

              Vj_ = 22.4fc
              Ti = 32°F + 460 = 492°R
              V  = volume at 68°F
              T = 68°F + 460 = 528°R

     Solving for V gives:

                Vj_ x T   22.4 x 528

            V = 	= 	= 24.045,
                  Tl         492


Since one mole of gas occupies 22.045, at 68°F, the  density can be found in
g/£ by dividing the molecular weight in g/mole by 22.04

                  mol. wt.  g/mole
                = 24.04
                                     288

-------
The density in ug/m£ can be found by converting g to ug and £ to mJl as
follows :

                  mol. wt. g/mole   1 x 106ug/g   moi. wt. x looo
      den Ug/m£ = - x --- - .. __
                  24.04  Vmole      1 x I03m£/Jl        24.04

                                                                (Equation 4)

To obtain the concentration of each amine in ppm, the concentration in
is divided by the density in ]ag/m£

      ppm = iag/m3 T yg/m£ = — =•
                            irr

Using Equations 3 and 4 gives  the ppm concentration in the form of the raw
data.
            24.04(£) x cstd (Vig/mfc)  x Asam x o.F. x Abs. Vol. (m£)
            - -
                Mol. Wt.  (g/mole)  x  IQOO x Astd x pexp  (»Hg)


                      Texp(°R)  x 29.92"Hg x 35.31 ft3/m3
                              528°R x vexp (ft3)
                                                                (Equation 5)
 At this point, the concentration can be expressed in yg/m3  (Equation 3)  and
 ppm (Equation 5) at 68°F  and 29.92"Hg from the raw data.

 Hewlett-Packard Calculations

      In order to insure maximum turnaround in a minimum time period, a
 Hewlett-Packard 67 program was developed to calculate the organic amine
 concentrations in yg/m3 and ppm from the raw data.  This program is
 presented in Figure 7.

 Sample Calculation

      Assume exhaust samples were collected in glass impingers for each
 portion of a three-bag 1975 FTP.  Raw data for these tests  are presented
 in Figure 8.  Calculations were performed using the HP 67 program and
 manual calculations.
                                      289

-------
        tser I ii si rufi ions
  "raanis Amines in Exhaust
STEP
°1
3?
31
1
;
3
4
5
6
7
*
3
10
LI
12
i ^
14
i ^
•
H
13
J.9 .
>C
21
22
23
24
25
;e
2"
;::
3"

NSTRUCT1ONS
Switch to on; =wii7n to run
Feed side I of cari in from right to left
Set decimal place
Input Sample Volusie
Input - Barometric Pressure
Input - Sample Temperature
Input Dilution Factor
Input - Absorbing Reagent Vol.
Input - Standard Cone. CH3NH2
Input - Standard Area CH3NH2
Input - Sample Area CHiNH?
Output Samp le ~or.c . CH }NH7
Output - Sample Cone. CI^MH?
Input - Standard zone. Co^N
Input - Standard Area CT-^N
Input - Dannie Area ~->H7N
Output Sample "one . oH-rN
Output 3'^niple '7criC . 7;H-N
._Inp_ut_ -_ 3tar.dar<3 Ocnc.. jCHjJjN
Input - ritr.adari Area <~H3)3N
Incut - Sample Area (CH-0 jN
_ -QUtpj^t _ - 5ampJ.e Cone . -,' CH^)_^N
Output - Sample Cone. (013)3^
Input - Standard, Cone. CC^H^-jNH
Input - Standard Area (C2H5)2MH
Input - Sample Area (C2H5)2NH
Ou^ut - Sample Cone. (C->H5)^NH
Output - Sample Cone- (C-jH^jj^H
Input - Standard tore. 'C^Hq) -,>J
Input_' ..Standard Area '^^S1 ^!
Input - Sample Area -C--.H-J ,N'
Output - Sample Cone. >'C-,Hri ,n

INPUT
DATA UNITS



"f^T*
"Hg
3C<

mi
ug/m?




Ug/mJl


.g/nul 	

	
yg/mf.



liri/mJ.

KEYS


9 ^Cl
A
R/'S
H/5
R/S
R/S
R/S
VS
3/S
R/S
;
R/3 •
R/S i
R/S '
R/S
R/S
R/S
R/S
R/S .
R/S !
R/S :
R/S i ,
R/S ' ;
R/S
R/3 1
R/3
R/ .3
n •:"^;
OUTPUT
DATA. UNITS











ug/m3
Dom



-g/^T"" ~
ppm

	 ..- — ..
'^g/m
"ppn " "~"'



•*Z? "~
ppm
__ -
-cr/mj
I
i
i
Figure 7. HP-67 user instructions.
                290

-------
STEP








".'-:









C20









030




KEY ENTRY
0
•
5
0
0
y
i/~
V
5 TO 1
R/3
4
6
n
a.
srr. 1
:n x" *y

h 1/x
R/S
X
R/S
X
^TT} ?
R/S
PCL 2
X
R/S
-r
R/S
X
s/s
1
2





•XG









050




	
^
f
"/3
RCL 2
X
R/S
=•
R/S
X
R/S
1
q
7
^

R/S
RCL 2
H
•3 / -
" "p/i 	
KEY CODE
OO
33
05
on
,-,o
n
84
71
"3 Tl
34
:14
Ofi
TO
Si
34 01.
35 52
81
35 62
84 •
7L
84
71
33 02
34
34 02
71
34
81
84
71
34
01
32
09
02
91
34
34 02
71
84
31
84
71
34
01
08
07
05
31
84
34 02
Jl
-14
. ...__x± ..


IP. cut-
Incut
COMMENTS STEP KEY ENTRY KEY CODE COMMENTS
ft 3
Barometer, "Ha
-Sample TemD,3F
Input Dilution Facto

Input Abs Sol Vol,

In Std Cone ug CH3NH
In St
In 5'
out :
Out '
In 34
In S
In S<
Out v
•lut I
In 3
"n .T
.d Area CH^NH2
line Area CH-NH-,
:onc ja/cn3 CH3NH
:onc ppm '7H3NH"-)
:d :onc -g/nt
:d Area C7K7N
imo Area C2H7N
:onc ug/mi
C2H7N
rone ?ptn C^^'i^
:d Cone jg/rai
tri -Xr^a CCH-j} V
^rr, \r:
?/s
4
2
3
1
81
34
71
34
04
02
00
•>9
f 31
p; ! s/s ^4
'-. ~TM " r 22
1
!

| i





TOO








i '•;
--
1












" 	 ".-"
5 r
r
r)













... —
Out Cone 'jg/ra3
iCH3)3M
"ut "j-nc -pm ic:?-}^';
In ftd Area
Ir. S^TO Aroa
Out "one '..g/mJ
f°2H5' 2N'H
Out Cone ppm
(C2H5)2NH
In Std Cone _ig/mi?,
(C2H5) 3N
In Std Area (C2Hs)3^
In Samp Area
^"2^5) 3N
Out Cone ug/tn^
:ut Cone r;^n
I' :' |'
e" " """ " »' 	
Figure 7  (Cont'd).  HP-67 program :form.
                  291

-------
SWRI PROJECT NO.	



FUEL:       CVS NO.
SAMPLE COLLECTION BY:_



GENERAL COMMENTS:
                        TEST NO.
TUNNEL SIZE:
                _TEST DATE:_



                  DRIVER:
                            CHEMICAL ANALYSIS BY:
                                                          VEHICLE:
                                   MILES:
                                                      CALCULATIONS  BY:
Test NO.
Driving Cycle
Volume Ft3
B.P." Hg
Temp°F
Dilution Factor
Absorb. Rea. Vol. Ml
Stan. Cone. CH-,NH,yg/ml
Stan. Area CH3NH2
Sample Area CH3NH2
Sample Cone. CH-,NH9 ug/m3
Sample Cone. CH,NH, ppm
Stan. Cone. C2H7N ug/ml
Stan. Area c?H7N
Sample Area C3H-7N
Sample Cone. C7H7N ug/1"3
Sample Cone. C-jH^N ppm
Stan. Cone. (CH,)7N ug/«l
Stan. Area (CH-,),N
Sample Area (CH3) ^N
Sample Cone. (CH^) -^N ug/m3
Sample Cone. (CH-,) 7N ppm
Stan. Cone. (C2H5) 2NH yg/ml
Stan. Area (C3Hc;)7NH
Sample Area (C,Hq),NH
Sample Cone. (c,H^),NH ug/m3
Sample Cone. (C->H^) 7NH ppm
Stan. Cone. (C-jH,;) 3N ug/ml
Stan. Area (C-,HO 7N
Sample Area (C2H^) -;N
Sample Cone. (C^Hs) •$} ua/m3
Sample Cone. (C2Hs) 3N ppm
FTP-1
3.196
29.30
75
1
25
0.05
1000
640
8.99
0.007
0.10
1500
1000
18.7
0.010
0.50
1500
1020
95.5
0.039
0.90
1560
1760
285
0.094
1.60
3120
2850
410
0.-098
FTP-2
1.625
30.02
80
5
25
0.10
2000
1000
138
0.107
0.20
3000
1800
332
0.177
1.00
3000
2800
2580
1.050
1.50
3000
4321
5980
1.970
2.00
1260
780
3430
0.814
FTP- 3
2.010
29.02
96
10
50
0.02
3000
1880
119
0.092
0.30
6000
5000
2380
1.270
0.20
4000
2100
1000
0.407
0.10
1870
2000
1020
0.335
0.10
4000
3160
753
0.179
SET-7
3.730
29.25
85
2
50
1.00
10000
800
79.9
0.062
2.00
15000
8500
1130
0.603
2.00
2000
1600
1600
0.650
1.00
1420
1070
753
0.247
0.50
6210
3100
249
0.059
HFET
8.241
29.95
83
1
75
0.10
4000
2000
16.5
0.013
0.20
2000
1500
49.5
0.026
1.50
1000
875
433
0.176
0.50
1900
1870
162
0.053
0.60
3110
2000
127
0.0^0
NYCC
1.070
29.50
39
1
75
0.10
5000
2700
141
0.109
0.20
3000
2000
348
0.186
1.00
2000
1630
2130
0.865
0.40
3100
2810
946
0.311
0.20
5620
4020
373
0.089
      Figure  8.   Organic  amine sample  collection sheet.
                                  292

-------
Manual calculations for driving cycle FTP-1

               Cstd(yg/m£) x Asam x D.F.  x Abs.  Vol.  (m£)
yg/m3 CH3NH2 = 	_—_
                          Astd X Pexp  <"Hg)

                     Texp x 29.92"Hg x  35.31  ft3/m3
                   x 	
                          528°R x vexp  (ft3)

               (0.05 yg/m£) x 640 x l x 25
                     1000 x 29.80"Hg

                     (460 + 75) x 29.92"Hg  x  35.31  ft3/m3
                   x	———-
                              528°R x  3.196 ft3

             =8.99 yg/m3

PPM CH3NH2   - yg/m3 -f density yg/mJl

                             Mol. Wt.  (CH3NH2) x  1000
             density yg/mJl =
                                    24.04£

             Mol.  Wt. CH3NH2 = 31.058 g/mole

                        31.058 x 1000
             density =	= 1292 yg/m£
                           24.045,

ppm =8.99 yg/m3 -f 1292 yg/m£ = 0.007 m£/m3 = 0.007 ppm

The calculation for C2H7N, (CH3)3N, and (C2H5) 3N are carried out in the same
manner by substituting the appropriate standard concentrations, areas and
molecular weights  into the above formulas.  These calculations give the
following concentrations:   C2HyN, 18.7 yg/m3 and 0.01 ppm;  (CH3)3N, 0.05
Hg/m3 and 0.039 ppm; (C2H5)NH, 285 yg/m3 and 0.094 ppm; and  (C2H5)3N,
410 yg/m3 and 0.098 ppm.

Note:  The values  used in these calculations are picked from a range of
temperature,  standards, dilution factors, etc., to validate the calculations
and may not be representative of expected raw data.  These calculations are
presented to  confirm the manual and HP-67 calculations give the same results.
This was confirmed on six sets of calculations.
                                      293

-------
LIST OF EQUIPMENT AND REAGENTS

     The equipment and reagents for the analysis of the organic amines are
divided into two groups.  The first involves the sample acquisition and the
second the instrumental analysis of the sample once it has been obtained.
Manufacturer, stock number and any pertinent descriptive information are
listed.  The preparation of the absorbing solution, the ascarite pre-column
and the primary standards are also discussed.

Sampling

 1.  Glass impingers, Ace Glass Products, Catalog #7530-11, plain tapered
     tip stoppers with 18/7 arm joints and 29/42 bottle joints.  -

 2.  Flowmeter, Brooks Instrument Division, Model 1555, tube size R-2-15-C,
     graduated 0-15, sapphire float, 0-5 &/min range.

 3.  Sample pump, Thomas Model 106 CA18, capable of free flow capacity of
     4 5,/min.

 4.  Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
     capacity.

 5.  Regulating valve, Nupro 4MG, stainless steel.

 6.  Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8" ID
     and 5/16" OD x 1/8" ID.

 7.  Teflon solenoid valve, the Fluorocarbon Company, Model DV2-144NCA1.
                                               «
 8.  Drying tube, Analabs Inc., Catalog #HGC-146, 6" long, 1/4" brass
     fittings.

 9.  Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, connectors,
     etc.

10.  Digital readout for dry gas meter.

11.  Miscellaneous electrical switches, lights, wirings, etc.

12.  Six channel digital thermometer, Analog Devices, Model #2036/J/1.

13.  Iron/Constantan type J single thermocouple with 1/4" OD stainless
     steel metal sheath, Thermo Sensors Corporation.

14.  30 m£ polypropylene sample storage bottles, Nalgene Labware, Catalog
     #2006-0001.

15.  Sulfuric Acid, H2S04, formula weight - 98.08, Certified 1N by Fisher
     Scientific Company, #SO-A-212.
                                     294

-------
16.  Class A,  10  mJl volumetric pipet.

17,  Class A,  1000 m& volumetric flask.

18.  Stainless steel heated filter assembly -  7 cm, Scott, capable of
    temperatures to 204°C, includes 2 heated, adjustable thermostat
    switch,  stainless steel insulated covers  and sample bypass solenoid
    valves.

19.  Glass microfiber filter discs, Reeve Angel 934-AH, Whatman, 7 cm
    diameter.

20.  Flexible, heavy insulation heating tape,  Briskeat®, width - 1/2 inch,
     length  - 48 inches.

21.   Temperature Controller, Athena, 100-600°F.

22.   Heated  TFE Teflon hose, Technical Heaters Inc.,  5' x 1/4",
     temperature limit 400°F.


Instrumental Analysis

  1.  10 ]il syringe, Pressure-Lok, Precision Sampling  Corporation.

  2.  Perkin-Elmer Model 3920 B gas chromatograph equipped with an ascarite
     loaded pre-column, a Teflon interface, and a nitrogen phosphorus
     detector  (NPD).

  3.  Soltec Model B-281 mv recorder.

  4.  Hewlett-Packard Model 3354 gas  chromatograph computer system with
     remote teletype printout.

  5.  Glass insert for 1/8" and 1/4"  heated injectors  (used as pre-column),
     Perkin-Elmer,  #009-1958.

  6.  Ascarite, 20-30 mesh, Arthur H. Thomas Company,  Catalog #C049-U86.

  7.  Teflon tubing  (used for interface in gas  chromatograph), Analabs
     Inc., 1/6" OD  x 0.03" ID, #HGC-024.

  8.  Methylamine Hydrochloride, CH3NH2'HC1,  formula weight = 67.52,
     crystals, Eastman #116.

  9.  Dimethylamine  Hydrochloride,  (CH3)2NH-HC1,  formula weight = 81.55,
     crystals, Eastman #94.

 10.  Trimethylamine Hydrochloride,  (CH3)3N-HC1,  formula weight = 95.57,
     crystals, Eastman #265.
                                      295

-------
11.  Ethylamine Hydrochloride, C2H5NH2-HC1, formula weight =  81.55,
     crystals, Eastman #731.

12.  Diethylamine Hydrochloride, (C2H5)2NH'HC1, formula weight =  109.6,
     crystals, Eastman #2090.

13.  Triethylamine Hydrochloride, (C2H5)sN-HCl, formula weight = 137.65,
     crystals, Eastman #8535.


Preparation of Absorbing Solution

     The absorbing solution  (0.01 N ^804) is prepared by diluting 10 mi of
1 N sulfuric acid (certified Fisher Scientific Company) to 1  liter with
deionized water.

Preparation of Ascarite Pre-Column

     The ascarite pre-column is prepared by packing a Perkin-Elmer glass
insert for a heated injector block with 20-30 mesh ascarite and plugging
both ends with 1/4" of glass wool.   The packed pre-column is  shown in
Figure 4.  This packed pre-column is then inserted into the heated injector
and held in place against the column with a spring loaded metal tube and the
septum cap.

Preparation of Primary Standards

     The primary standards for the organic amines are prepared by dissolving
a weighed amount of the amine-hydrochloric acid salt in 0.01  N sulfuric acid
and diluting the resulting mixture to the proper volume with  0.01 N sulfuric
acid.  Standards less than 10 ppm are prepared by diluting higher concen-
tration standards with 0.01 N sulfuric acid.
                                      296

-------
REFERENCES

Broker,  W. and Mossman, A. L.,  Ma the son Gas Data Book, 5th Edition, East
Rutherford, N. J. , 1971, pg.  385^           — — —    -

Weast,  R.  C. ,  (Ed.), Handbook of Chemistry and Physics, 54th Edition, The
Chemical Rubber Co., Cleveland,  Ohio,  1973.

Hoshika, Y. , Anal. Chem. , Vol.  48,  pg. 1716, 1976.

Hoshika, Y., J. Chromatog., Vol. 115,  pg. 596, 1975.

Sze, Y. L. and Borke, M. L.,  Anal.  Chem., Vol. 35, pg. 240, 1963.

O'Donnell, J. F., and Mann, C.  K. ,  Anal. Chem., Vol. 36, pg. 2097, 1964.

McCurdy, W. H. Jr., and Reiser, R.  W. , Anal. Chem., Vol. 38, pg. 795, 1966.

Smith, J. R. L. and Waddington, D.  J., Anal. Chem., Vol0 40, pg. 522, 1977.

"Amine Analysis," Bulletin  737A, Supelco Co., Belefonte, Pa., 1973.

"Penwalt  223 Amine Packing,"  Alltech Assoc., Arlington Heights, Illinois.

Moffat, A. C. and Horning,  E. C., Anal. Lett., Vol. 3, pg. 205, 1970.

Clarke, D. D., Wilk, S.,  and  Gitlow, S. E., J. Gas Chromatog., pg. 310, 1966.

Hosier, A. R. , Andre, C.  E. ,  and Viets, F. G. Jr., Environ. Sci. Tech.,
Vol. 7, pg. 642, 1973.

Burks, R. E. Jr., Baker,  E. B., Clark, P., Esslinger, J., Lacey, J. C. Jr.,
J.  Agr. Food Chem., Vol.  7, pg. 778, 1959.

Dunn, S.  R., Simehoff, M. L., Wesson,  L. G. Jr., Anal. Chem., Vol. 48,
pg. 41, 1976.

 Gruger, E. H. Jr., J. Agr.  Food Chem., Vol. 20, pg. 781, 1972.

Andre, C.  E. and Hosier,  A. R. , Anal.  Chem., Vol. 45, pg. 1971, 1973.

Umbreit,  G. R. , Nygren, R.  E. ,  and Testa, A. J., J. Chromatog., Vol. 43,
Pg. 25, 1969.
     , J. N. and Hardy,  R.,  J. Sci. Food Agr., Vol.  23, pg. 9, 1972.

 Bowen, B. E., Anal.  Chem.,  Vol. 48, pg. 1584, 1976.
                                       297

-------
       APPENDIX E




SULFUR DIOXIDE PROCEDURE
          298

-------
               THE MEASUREMENT OF SULFUR DIOXIDE  IN EXHAUST


     The  concentration of sulfur dioxide, S02,  in automotive exhaust can be
determined as  sulfate using an ion chromatograph.  Sulfur dioxide exhaust
samples are collected in two glass bubblers, each containing 3 percent
hydrogen  peroxide.  The temperature of the  absorbing  solution is kept at 0°C
by means  of an ice water bath.  The bubbled samples are  analyzed on the ion
chromatograph  and compared to standards of  known  sulfate concentrations.

SAMPLING  SYSTEM

     Two  glass impingers in series, each containing 25 m£ of a 3 percent hy-
drogen peroxide solution, are used to collect exhaust samples for the analy-
sis of sulfur  dioxide.  A flow schematic of the sample collection system is
shown in  Figure 1.  The two impingers together  trap approximately 99 percent
of the sulfur  dioxide.  The temperature of  the  impinger  is maintained at
0-5°C by  an ice water bath, and the flow rate through the impinger is main-
tained at 4 £/minute by the sample pump.  A dry gas meter is used to determine
the total flow through the impinger during  a given sampling period.  The
temperature of the gas stream is monitored  by a thermocouple immediately prior
to the dry gas meter.  A drier is included  in the system to prevent condensa-
tion in the pump, flowmeter, dry gas meter, etc.   The flowmeter in the system
allows continuous monitoring of the sample  flow to insure proper flow rates
during sampling.  When sampling diesel fueled vehicles,  a heated filter,
located between the on-off solenoid valve and the dilution tunnel, is used
to prevent diesel particulate from contaminating  the  sampling system.  The
filter and line connecting the filter to the dilution tunnel are heated to
375°F in  order to prevent sulfur dioxide from being retained in the filter and
sample line.  The Teflon line connecting the heated filter and the solenoid
valve is  heated to 175 °F in order to prevent water from  condensing in the
sample line.  Several views of the sampling system are shown in Figure 2.

ANALYTICAL PROCEDURE
     Sulfur dioxide in dilute exhaust  is  collected in two  impingers connected
in series with each impinger  containing  25  mi  of 3 percent hydrogen peroxide.
The  temperature of the impinger is maintained  at 0-5°C by  an  ice water bath.
The  flowrate through the impinger is adjusted  to 4 2,/minute with a regulating
valve and the sample pump.  After sampling  is  completed, the  absorbing solu-
tion in each bubbler is transferred to a 30 m£ polypropylene  bottle and  capped.
The  samples should be analyzed within  four  or  five weeks after  collection.
Approximately 2 m£ of the sample is loaded  into the  ion  chromatograph sample
loop and injected.  The injection inserts the  sample  loop  volume  (0.5 mfc) of
sample into the instrument.  An analysis  flow  schematic  and pictures of  the
ion  chromatograph are shown in Figures  3  and 4.  The  ion chromatograph

                                      299

-------
CO
o
o
     Sample
     probe
                                                          Sample
                                                           Pump
     Dilute
    Exhaust
                                                                         Gas Temperature
                                                                         Digital Readout
                                                                              Flowmeter
                     *
                                                                    Regulating
                                                                      Valve
    Ice Bath
Temperature Readout
                                                                                                Dry
                                                                                                Gas
                                                                                               Meter

  Gas Volume
Digital Readout
                               Figure 1.  SO  sample collection flow schematic.

-------
          Front View
                                                Digital
                                                Readout
                                               Flowmeter
                                               Regulating
                                               Valve
      Close-up of Upper Front
Figure 2.  SO2 sampling system.
             301

-------
Solenoid
Impinger
Ice Bath
                          Close-up of Impingers (Side View)
                                                      Solenoid
                                                      Filter

                                                      Ice Bath
                                                       Pump
            Rear View
        Figure 2  (Cont'd).  S02  sampling  system.
                           302

-------
       CVS
          \
      Glass

     Impinger
    Unused
    sample
    saved as
    needed
Sample analysis in
Ion Chromatograph
with conductivity
      cell
A/D Converter
          I
   Recorder
Hewlett-Packard
    3354
Computer System
         Figure 3.  SO-  analysis flow schematic.
                           303

-------
                   Suppressor	1
                     Column
                                     Conductivity
                                       Detector
                                         Conductivity
                                           Detector
Figure 4.  SO^ ion chromatograph,
               304

-------
utilizes two  columns,  the separator and the  suppressor.   The  3 x 500 mm
analytical  column and a 3 x 150 mm precolumn are packed with  a patented resin
composed of a strong base anion exchanger'  in the bicarbonate  form.  The
analytical  column separates the anions before entering the  suppressor column
The 6 x 250 mm glass suppressor column packed with AG 50W-X10, a strong acid'
cation exchanger, neutralizes the ionic effect of  the eluent  while increasing
that of the sample ion.  The column packing  is in  the hydrogen form so that
in the presence of the eluent (NaHCO3 and  Na2CO3) , H2CO3  is generated and when
sulfate is  introduced, ^864 is formed.

                    NaHCO  + Resin-H ^=^ Resin-Na + H CO

                    Na+ Anion" + Resin-H ^^ Resin-Na + H+ Anion"

The acid  being more conductive than the hydrogen carbonate  produces a signal
on the conductivity meter.  This can be interpreted as peak height from a
trace or  as peak area measured by the Hewlett-Packard 3354  computer system.
Figures 5 and 6 show two representative chromatograms produced in the analy-
sis of a standard and a sample.  The ion chromatograph operates at room tem-
perature  at a maximum pressure of 500 psi.

CALCULATIONS

     This procedure has been developed to  provide  the user  with the concen-
tration of sulfur dioxide in exhaust.  The results will be  expressed in
 Ug S02/m3  of exhaust  and ppm.   A stepwise derivation of  the  equations used in
calculating the concentrations is provided as well as a copy  of a Hewlett-
Packard 67 calculator program  (Figure 7) .   This program is  designed to reduce
the  amount of time required to do the calculations manually.  For illustra-
tion, two examples using  information from  the data sheet  (Figure 8) are
 included at the end of this section.

     The first step in the calculations is to correct the volume of exhaust
 sampled to a standard temperature, 68°F, and pressure,  29.92"Hg, by use of
 the  equation:

          PxVx = Pspec x  V
           Tx      Tspec

 Solving for V gives:

              PxVx  x Tspec
             Tx+460   Pspec

     where V = volume of  gas sampled at  specified temperature
               and pressure  (ft  )
          Px = experimental pressure ("Hg)

          Vx = experimental gas volume collected (ft )

          Tx = experimental temperature (°F)
                                       305

-------
 12   10   8    6420
   Retention time, minutes
Figure 5.  Sample chromatogram.

             306

-------
                      Standard  Date 11-30-79
 Attenuation 30x1
 Eluent 0.003 M NaHCO3 + OeQQ24 M
 Flowrate 167 nd/hr  Chartspeed 12 in/h
 Loop Size 0.5 m£   Packing AG 50W-X16
 Analytical Column Pre column 3x150 mnTglass,
  Analytical column 3x500 mm glass"         '
 Packing Strong base anion exchanger in  the
  bicarbonate form
 Supressor Column 6x250 mm glass
       15        10       5
     Retention time, minutes

Figure 6.  Standard Chromatogram

              307

-------
           I Sut Absorbing reagent volume
Input Standard Concentration
Input Standard Area
Input Sample Area
Output Safliple Concentration
Output Sample Concentration














.



-- 	 -


INPUT
DA f A UNITS



ft3
"H.J
°F

mi
MgSO42-/i«

























KtVS














1
]
1 y
; A
i R/sl
i H/S|
1 R/sl
i R/si
I R/S|
1 R/sl
1 R/sl
1 R/sj
1 1
1 a 1
1
1
|Scl
1


i
1
i
1


!
IKON

1





1






1 1



















1 i


,
i
i












i
1








1



I
























OUTPUT
DATA, UNITS











]iy SO-;/Ul
Pi 'in





















Figure  7.   HP-67 user instructions.
                 308

-------
STEP KEYENTRY KEY CODf COMMH,,* iltP Kty EMHY KEY CO,*
- 	 '
	 -
	
:'.IU
	


it JO



	

OJO









U40
' V








-J'M






tI£L.A. .
3

r_i 	 ."."
1~JL ~.~".
2
	 B 	
	 X 	
2
9
....2 	
•>

R/S
_Sm_iIl. 	
6
0
-f
KCL 1
h X ?y _
h 1/X
R/S
X
R/S
X
R/S
X
R/S

R/S
X
0
•
6
6
7
X
R/S
0
•
0
0
0
3
6
^
x
_ R/S
il ii ii..
.. .ui. .....
_. U5__
sa
.. !il. .
01
«i 	
	 05 	
o;>
	 i)iL 	
.._.?! 	
OJ
_fy. _
	 ca 	
ai
u-4
H4
	 ud 	 _.
Ou
00
61
34 01
35 52
8) ~^
35 62
84
71
84
71
84
71
Si 4
81
84
71
00
81
06
06
07

84
00
B3
00
00
00
03
On
07
71
H4
In^ul ?>.iiu[.
>; v,-,l,
ft3
Input, barometer, "liq
Input. Saitii>lu Temp,
0F

Input Dilution Factc


Input Abs Suln Vol ,
Input Std
P9 SU42'
Input Std
Input Samp
Output Con
.tl ^'^II
o i a J
30 Sl S2 S3
A 1
3 C
mil
Cono
'/M
Area
le Area
c,Vig ,S02/
mj
.1
^•4

ot.o
	
l) 'll


	
UtlO


	

c



uao









100








_


U KB!
	 ^ .

	
	 	



	




























JS j.:
-- - --



	







	






	











5 6 /
S5 S6 SI
D

~ a •'

t 1
Figure 7  (Cont'd).  HP-67 program form.
                  309

-------
     SwRI  Project No.
     Fuel:
Test No.
Test Date:
   Vehicle:
                           CVS No.
      Tunnel Size
            Driver:
     Sample  Collection By:
     General Comments:
Chemical Analysis By:
	  Miles:
 Calculations  By:
Driving
Cycle
FTP-1
FTP-2
FTP-3
SET- 7
HFET
NYCC

BG

Sampling Conditions
Volume
Ft3
1.189
2.040
1.250
3.240
1.690
1.410

2.000
1
B.P.
"Hg
29.47
29.48
29.49
29.50
29.50
29.51

29.45
2
Temp.
OF
74.7
75.0
74.5
75.0
74.7
74.9

74.9
3
Dilution
Factor
1
1
1
1
1
1

1
4
Absorb.
Reagent
Volume m&
25
25
25
25
25
25

25
5
Standard
UgS04-2/m£
:10
30
10
40
20
15

1.0
6
Area
3200
4500
3200
4200
5700
4900

3000
7
Sample
Area
4000
3500
4300
4500
5500
5000

2700
8
ugSO2/m3
6360
6930
6500
8000
6670
6630

273
9
ppm
2.34
2.54
2.39
2.94
2.45
2.43

.0.10
10
CO
I—»
o
                                           Figure  8.   SO2  data  sheet.

-------
      Tspec =  specified temperature = 68°F = 528 °R

      Pspec =  specified pressure = 29.92 " Hg

    Next, the  quantity of SO2 in this volume of gas is calculated.  The ion

chromatograph measures the amount of sulfate  (from S02) in yg S°4~2  Assuming

linearity between  concentration of sulfate and peak area, a standard of known
sulfate  concentration is compared to a bubbled exhaust sample.

        Cst =    Csamp
      AREAst  AREAsamp

                Cst x AREAsamp
        Csamp = - _____ -

                                                            _2
    where    Csamp = concentration of SO."2 in sample ^9   4
                                                         ml
                                                               _2
                Cst = concentration of SO."2 in standard
                                                            mil

           AREAsamp = area of sample  (relative units)


             AREAst = area of standard  (relative units)

 Converting to yg SO /m£:
                                64.06 g/mole SO
                                96.06 g/mole  SO,"

     If the sample has been diluted the dilution  factor, DF, needs to be
 included.

                        0.667 Csamp x DF

 This represents the amount of SO2 in one m£ of  absorbing solution.  The
 volume of absorbing solution is multiplied by this  last quantity to give the
 amount of SO  collected.
                        0.667 Xsamp x DF  x  absorb,  vol.  (m£  ) = yg
 The concentration of SO2 in Ug 3°?  in the  exhaust  sample  is obtained by
 dividing yg SO  by gas volume  (corrected  from ft3 to  m3) .
                                      311

-------
Concentration of SO ,
                      ug so2
                       yg SO     x AREAsamp
      yg SO          ' - £. —             x DF x absorb .  vol . ,  m£
0.667 -    X -  -
      yg SO             AREAst
                 Px,  "Hg x Vx,_ ft3  x     528°R
                  Tx, °F + 460           29.92" Hg
                               -2
         yg so           yg SO
   0.667 	^j x cst, 	—JT	  x AREAsamp x DF x absorb, vol.,
         yg so4
                                              3
                              Px,  "Hg x Vx,  ft

                                                       3
                   (Tx,  °F + 460) x 29.92" Hg x 35.31 ——
                                                      ft3
                             528°R x AREAst

To find the S02 concentration in ppm the density of the gas at the specified
conditions is needed.  The density of S02 at  32°F and 29.92"  Hg  is 2.927
g/lit.  This can be corrected to 68°F C492°R)  and 29.92" Hg using Charles"
version of the ideal gas law:


                                 V     V
                                       T
where V± is the volume at 32 °c  (£) = 1.0
      T-L is 32 °F = 492 °R
      V  is the volume at 6S°F  (£)
      T  is 68°F - 528°R
      v =  (1.0 i)  (528°R)
                492°R
Density at 68°F and 29.92" Hg = 2'927 3
                                x • U / 
-------
Sample  Calculation

     The two examples will be calculated from information recorded on the
data sheet (Figure 8) .  This information does not necessarily represent
actual  experimental data but serves as  a means of confirming calculations
done by hand with those done with the Hewlett-Packard Calculator.

Example 1

     Assume 1.189 ft  of exhaust was collected in 25  nd of 3 percent H202
in the  FTP-1 driving cycle at a barometric pressure of 29.47" Hg (cor-
rected) and temperature of 74.7°F.  The sampling and  analysis were both
performed according to the procedure outlined in a previous section.  An
excess  of sample injected into the 0.5  m& sample loop produced a peak area
                                                in yg  S04~2
of 4000 counts and the corresponding standard, - j — * — , an area of
3200 count.                                        m£


                       yg so2            yg so "2
                 0.667 - -  x Cst, - 5 -  x AREAsamp
                             -
             3
     yg SO./m  =
          2                  Px, "Hg x Vx, ft3
     DF x absorb, vol., m& x  (TX,  °F  +  460)  x   29.92" Hg x 35.31  m
                                                                  ft3
                                  528°R x  AREAst
     0.000367 x Q.667 x 1Q x  4QQQ  x  i  x 25  x (74.7 + 460)  x  29.92 x 35.31
                            29.47  x  1.189 x 528 x 3200



                        = 6360 yg  SO2/m


                                        *3             m Q
                ppm SO0 = 6360 yg  SO_/m  x  0.000367
                      •'rt   *-" — w  j-i -y  '-"~-ijf —*    —	—    11/T



                        =2.34  ppm SO
 Example 2
     Assume that in the SET-7  driving cycle dilute automotive exhaust was
 collected in 25 m£ of 3 percent H202  according to the procedure described
 previously.  The sampling conditions  under which the 3.240  ft  of exhaust
 was collected were 75.0°F and  29.50 " Hg.  An area of 4200 counts  was pro-
                       n
 duced by the 40 yg S(?4~  standard and the exhaust sample yielded  an  area of
                                     313

-------
4500 counts.   Inserting these values into the same equation used in Example
1 gives concentrations of 8000 yg S02/m3 and 2.94 ppm S02.

LIST OF EQUIPMENT

     The equipment required for the SC>2 determination is divided into four
categories:  Sampling, Analysis, Water Filtration and Sample preparation.
Manufacturer, stock number and any pertinent descriptive information are
listed.

Sampling

     1.   Glass impingers. Ace Glass Products, Catalog #7530-11, plain tapered
          tip stoppers with 18/7 arm joints and 29/42 bottle joints.

     2.   Flowmeters, Brooks Instrument Division, Model 1555, R-2-15-C,
          sapphire ball, 0-5 lit/min range, graduated 0-15.

     3.   Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
          capacity.

     4.   Digital readout for dry gas meter.

     5.   Sample pump, Thomas, Model #106 CA18 3, 4 lit/min.

     6.   Drying tube, Analabs Inc., Catalog #HGC-146, 6" long, 1/4" brass
          fittings.

     7.   Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8" ID
          and 5/16"OD x 3/16" ID.

     8.   Teflon solenoid valve, The Fluorocarbon Company, Model #DV2-144N
          Cal.

     9-   Miscellaneous Teflon nuts, ferrules, unions, tees, connectors and
          clamps.

    10.   Miscellaneous electrical switches, lights, wiring, etc.

    11.   Regulating valve, Nupro 4M6, stainless steel.

    12.   six channel digital thermometer, Analog Devices, Model #2036/J/1.

    13.   30  m£ polypropylene sample storage bottles, Nalgeie Labware,
          Catalog #2006-0001.

    14.   Iron/Constantan type J single thermocouple with 1/4" OD stainless
          steel metal sheath, Thermo Sensors Corporation.

    15.   Stainless steel heated filter assembly-7 cm; Scott, capable of tem-
          perature to 204°C, included 2 heaters, adjustable thermostat switch,


                                    314

-------
         stainless steel  insulated covers and sample bypass solenoid valves.

   16.   Glass raicrofiber filter discs,  Reeve Angel 934-AH, Whatman, 7 cm
         diameter.


   17.   Flexible heavy insulation heating tape,  Briskeat®, width-1/2 inch
         length-48 inches.                                                '


   18.   Temperature Controller, Athena, 100-600°F.


   19.   Heated TFE Teflon hose, Technical Heaters, Inc.,  5' x 1/4" tem-
         perature limit 400°P.


Analysis


     1.   Conductivity  cell,  modified swagelok reducing union, Catalog
         #SS-200-6-l,  approximate volume, 4.5 yl.


     2.   Conductivity  detector,  Hall, Tracor 700.


     3.   Multivoltage  recorder,  Soltec,  Model #B-281 H.


     4.   Mini-pump, Milton Roy,  series 196-0066-033, 46/460 m£/hr capacity.


     5.   Pulse  dampener,  Glenco Scientific, Catalog #PD 1000.

     6.   Polyethylene  cubitainers, Cole  Partner Instrument  Company, Catalog
         #6100-20, 1 gallon.


Water Filtration

     1.   Filtration apparatus, Millipore, Catalog #XX 15 047 00.

     2.   Filters, Millipore, Catalog ttGSWP 047 00, 0.22 micron pore size.


Sample preparation

     1.   3 cc disposable  syringes, Becton-Dickson, Catalog #5585.


     2.   Class A, 1 m& volumetric pipets.

     3.   Class A, 2 m£ volumetric pipets.


     4.   Class A, 3 m£ volumetric pipets.


     5.   Class A, 4 m£ volumetric pipets.


     6.   Class A, 5 m£ volumetric pipets.


     7.   Class A, 10 n>Jl volumetric pipets.
                                     315

-------
     8.    Class A,  20 mH volumetric pipets.

     9.    Class A,  25 rtd volumetric pipets.

    10.    Class A,  50 m£ volumetric pipets.

    11.    Class A,  100 m£ volumetric pipets.

    12.    Class A,  100 m£ volumetric flasks.

    13.    Class A,  1000 m£ Volumetric flasks.

    14.    Class A,  2000 mH volumetric flasks.

    15.    Mohr pipet, 1 mi graduated 1/10.

LIST OF REAGENTS

     A list of reagents used in determination of SO2 is provided indicating
purity,  manufacturer and catalog number.  The function of each reagent in
the procedure is also given.

     1.    Water - deionized and filtered through 0.22 micron filter.

     2.    Primary standard - Sulfuric acid, H2SO4, certified 0. IN, formula
          weight = 98.08, ACS reagent grade, Fisher Scientific Company
          #SO-A-212.
     3.   Absorbant - Stabilized 30 percent hydrogen peroxide,
          weight = 34.01, analytical reagent grade, Mallinckrodt #5239.

     4.   Eluent - Sodium bicarbonate, NaHCO-^, formula weight = 84.01, ACS
          analytical reagent grade powder, Mallinckrodt #7412.

     5.   Eluent - Sodium carbonate, Na2CO3, formula weight = 105.99, ACS
          analytical reagent grade anyhdrous power, Mallinckrodt #7521.

     6.   Regenerant - Sulfuric acid, 112804, formula weight = 98.08, ACS
          analytical reagent grade, Mallinckrodt #2876.

PREPARATION OF REAGENTS

     Water is prepared by filtering deionized water through a 0.22 micron
Millipore filter and storing in polyethylene bottles.  All solutions and
dilutions are made up to volume with water prepared in the above manner.

Primary Standard

     The stock solution is prepared by diluting 20 mJl of certified 0.1 N
                   f\
H2S04 (4800 yg   4   to 1000 m£  with water.  The resulting solution contains
              m£

                                    316

-------
           -2
96 -° yg   4   •  More dilute standards are prepared by pipetting 0.5,  1,  1.5,
       m£
2, 3, 4, 8, 10, 20, 30, 40, 50,  60,  70, 80, and 90 m£ of the stock  solution
into 100 m£ volumetric flasks  and making up the volume.  -These standards  re-
main stable for at least  fourteen weeks.  ^11 glassware used in the prepara-
tion of standards should  be washed with 1:1 (v:v)  nitric acid and then rinsed
copiously with tap water  with  a  final rinse of filtered deionized water.
 Absorbing Solition  (3 percent

     100 m& of 30 percent  H.2Q2 i-s diluted to 1 liter with water.

Eluent  (0.003  M NaHCO^  + 0.0024 M
     Concentrated stock solutions of sodium bicarbonate and sodium carbonate
are prepared in the following manner.  For a 0.6 M solution of sodium bicar-
bonate,  50.41 g of solid sodium bicarbonate is dissolved in 1000 mi of water.
For a 0.48  M solution of sodium carbonate, 50.88 g of solid sodium carbonate
is dissolved in 1000 m£ of water.  The carbonate solution used as the eluent
is made  up  by pipetting 10 m£ of each stock solution into a 2 liter volumetric
flask and diluting to mark with water.

Regenerate  (1 N H2SO4) *

     56  mJl  of 95 percent H2SO4 is diluted to 2 liters with water.

     *A total of 4 liters of each of these solutions are prepared to fill
the 4 liter reservoirs in the chromatograph .
                                      317

-------
REFERENCES

Braker, W. and Mossman, A.L.,  Matheson Gas Data Book, 5th Edition, East
Rutherford, N.J., 1971, pg. 513.

West, P.W. and Gaeke, G.C., Anal. Chem., Vol. 28, pg. 1816, 1956.
                                 \
  /
Adams, D.F., Corn, M., Harding, C.I., Pate, J.P., Plumley, A.L., Scaringelli,
F.P., and Urone, P., "Methods  of Air Sampling and Analysis,"  [Tentative
Method of Analysis for Sulfur Dioxide Content of the Atmosphere  (colormetric),
No. 42401-01-69T], pg. 447, Amer. Pub. Health Ass., Washington, D.C., 1972.

"ASTM Standards - Water, Atmospheric Analysis," ASTM Designation D291470T,
Part 23, Am. Soc. Test. Mat.,  Philadelphia, PA., 1972.

US EPA, Fed. Register, Vol. 36, No. 158, 15492, 1971.

Robinson, E. amd Rabbins, R.C., "Sources, Abundance, and Fate of Gaseous
Atmospheric Pollutants," Final Report, SRI  Project 6755 Supplement,
Stanford Research Inst., Palo  Alto, Calif., 1968.

Stern, A.C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc., N.Y.,
1976, Vol. 3, Pg. 214.

Zurlo, N. and Griffini, A.M.,  Med. Lav-, Vol. 53, pg. 330, 1962.

Scaringelli, F.P., Elfers, L., Norris, D., and Hochheiser, S. Anal. Chem.,
Vol. 42, pg. 1818, 1970.

Pate, J.B., Ammons, B.E., Swanson, G.A., Lodge, Jr., J.P., Anal. Chem.,
Vol. 37, pg. 942, 1965.

Pate, J.B., Lodge, Jr., J.P.,  and Wartburg, A.F., Anal. Chem., Vol. 34,
pg. 1660, 1962.

Scaringelli, F.P., Saltzman, B.E., and Frey, S.A., Anal. Chem., Vol. 39,
pg. 1709, 1967.

Stern, A.C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc., N.Y.
1976, Vol. 3, pg. 214

Urone, P., Evans, J.B., and Noyes, C.H., Anal. Chem., Vol. 37, pg. 1104,
1965.

Bostrom, C.E. Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.

Attari, A., Igielski, R.P., and Jaselskis, B., Anal. Chem, Vol. 42,
pg. 1282, 1970.
                                    318

-------
Kawai,  T.,  Netsu Kanri, Vol. 22, pg. 20,  1970.

Thomas, M.D.,  Ivie, J.O., and Fitt, T.C.,  Ind. Eng. Chem., Anal, Ed.,
Vol. 18, pg. 383, 1946.

Yocum,  J.E., Richardson, R.L., Saslaw,  I.M.,  and Chapman, s., Proc. 49th
Ann. Meet.  Air Poll. Contr., Ass., Pittsburgh, Pa.,  1956.

Kuczynski,  E.R. , Environm. Sci.  Tech.,  Vol. 1, pg. 68, 1967.

Shikiya, J.M. and McPhee, R.D.,  61st Annual Meeting,  Paper No.  68-72, Air
Poll. Conr. Ass., Pittsburgh, Pa.,  1968.

Jacobs, M.B., The Chemical  Analysis of Air Pollutants,  Chapter  8, Wiley
 (Interscience), New York, N.Y.,  1960.

Jacobs, M.B., The Analytical Chemistry of Industrial  Poisons, Hazards, and
 Solvents, Chapter 9, Wiley  (Interscience), New  York,  N.Y.,  1949.

 Katz,  M., Anal. Chem.,  Vol. 22,  pg. 1040, 1950.

 Stratman, H. , Mikrochim, Acta.  Vol. 6, pg. 688,  1954.

 Pate,  J.B., Lodge,  Jr., J.P.,  and Neary, M.P., Anal. Chem. Acta., Vol.  28,
 pg. 341, 1963.

 Lodge, Jr., J.P., Pate, J.B.,  and Huitt, H.A.,  Amer.  Ind.  Hyg.  Ass., J.,
 Vol. 24, pg.  380,  1963.

 Forrest, J. and Newman, L., Atmos. Environ. Vol. 7, pg. 561,  1973.

 Huygen, C., Anal.  Chim, Acta.  Vol. 28, pg. 349, 1963.

 Harding, J.,  and Schlein,  B.,  "Nuclear Techniques in Environmental  Pollution,'
 SM-142a/9,  IAEA,  Vienna, Austria, 1971.

 Wilsdon, B.H. and McConnell,  F.J., J.  Soc. Chem. Ind. Vol. 53,  pg.  385,  1934.

 HueyN.A.,  J. Air Poll. Cont.  Ass., Vol. 18, pg. 610, 1968.

 Thomas, F.W.  and Davidson, C.M., J. Air Poll. Cont.  Ass., Vol.  11,  pg.  1-
 1961.

 Stalker, W.W.,  Dickerson,   R.C., and Kramer, G.D., Amer. Ind. Hyg. Ass.,  J.,
 Vol. 24, pg.  68, 1963.

 Hickey, H.R.  and Hendrickson, E.R., J. Air Poll. Cont. Ass., Vol. 15,
 Pg. 409, 1965.

 Harding, C.I.,  and Kelley, T.R., J. Air  Poll. Cont.  Ass., Vol. 17, pg. 545,
 1967.

                                     319

-------
Stevens, R.K.,  Mulik, J.D., O'Keeffe, A.E., and Krost, K.J., Anal. Chem. ,
Vol. 43, pg. 827, 1971.

Stevens, R.K.,  O'Keefe,  A.E., and Ortman, G.C., Environ. Sci, Tech.,
Vol. 3, og. 652, 1969.

Ronkainen, P.,  Denslow,  J., Leppanen, O., J. Chroraatog. Sci,, Vol. 11,
pg. 384, 1973.

Pescar, R.E., and Hartman, C.H., J. Chromatog. Sci., Vol. 11, pg. 492, 1973.

Stevens, R.K.,  and O'Keefe, A.E., Anal. Chem., Vol. 42, pg. 143a, 1970.

Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I., Anal. Chem.,
Vol. 44, pg. 2070, 1972.

Bruner, F., Ciccioli, P-, and DiNardo, F., Anal. Chem., Vol. 47, pg. 141,
1975.

Bruner, F., Ciccioli, P., and DiNardo, F., J. Chromatog. Vol. 99, pg.  661,
1974.

Bremner, J.M.,  and Banwart, W.L., Sulfur Inst. Journal, Vol. 10, pg. 6,
1974.

Mulik, J.D., Todd, G., Estes, E., and Sawicki, E., "Ion Chromatography
Determination of Atmospheric Sulfur Dioxide," Symposium on Ion Chromato-
graphic Analysis of Environmental Pollutants, EPA, Research Triangle Park,
N.C., April, 1977.

Dietzmann, H.E., "Protocol to Characterize Gaseous Emissions as a Function
of Fuel and Additive Composition," Environmental Protection Technology
Series under Contract 68-02-1275, September, 1975.

DeSouza, T.L.C., Lane, D.C., and Bhatia, S.P-, Anal. Chem., Vol. 47,
pg. 543, 1975.
                                   320

-------
       APPENDIX F




NITROUS OXIDE PROCEDURES
          321

-------
                THE MEASUREMENT OF NITROUS OXIDE IN EXHAUST
     This procedure was developed to measure nitrous oxide  (N20)  in  dilute
gasoline and diesel exhaust.  Standard CVS bag samples are  analyzed  for N20
using calibration blends to quantify the results.  Gas chromatograph peak
areas are obtained using a Hewlett-Packard 3354 computer system.   This tech-
nique has a minimum detection limit of less than 0.01 ppm.  The total system
schematic for the analysis of N20 in exhaust is shown in Figure 1.

ANALYTICAL SYSTEM

     The analysis for N20 in exhaust is conducted with a gas chromatograph
system using a Perkin-Elmer Model 3920B electron capture detector.   The
system employs two pneumatically operated electrically controlled Seiscor
valves, an analytical column, and a stripper column.  The gas chromatograph
separation is obtained at room temperature.  A special control console was
fabricated to house the entire system except for the electron capture de-
tector.

     A stripper column is included as a precautionary measure to  prevent
 unwanted heavier molecular weight exhaust  species  from entering the  ana-
 lytical  system.  Figure  2  (Step  1) illustrates the gas chromatograph flow
schematic with ttie gas sampling valve in the purge position  and the backflush
valve is foreflushing to the analytical column.  Figure 3 (Step 2) illu-
strates the flow schematic when the gas sampling valve is actuated and the
backflush valve still in the foreflush position to the analytical column.
Once the N20 peak foreflushing has eluted, the backflush valve is activated
and the heavier molecular weight species retained on the stripper column are
backflushed to vent, as shown in Figure 4.  A summary of the individual steps
is presented below:

           Gas Sampling Valve	           Backflush  Valve	
        Position        Function             Position       Function
           off      purge CSV w/sample          off      foreflush to
                                                          analytical column

  2        on       sample injected             off      foreflush to
                                                          a-ialytical column

  3        on       sample injected             on       backflush to vent

     Under normal conditions, it is not necessary to backflush the calibra-
tion standards since they are free of contaminants that would interfere with
                                    322

-------
Vehicle
CVS
                                                       Tedlar bags
                                                          Gas
                                                      Chromatograph
                           HP 3354 GC
                         Computer System
                           Teletype
                           Printout
     Figure 1.   Total system flow schematic  for the analysis
                   of nitrous oxide in exhaust.
                               323

-------
        Stripper Column
   (2' x 1/8" SS, 10% OV-17
    on 80/100 Gas Chrom Q)
       Seiscor Valve
(backflush configuration)
                                                         Perkin-Elmer
                                                           3920  B
                                                       Electron Capture
                                                           Detector
                                       Analytical  Column
                                        (61 x  1/8"  SS,  80/100 Porapak Q)
                                                              Auxiliary
                                                             Carrier Gas
                      Capillary
                      Restricter
           »«•»•••••»••*•*••»••»••••«••*••••••%

                o	o
                                                              Carrier
                                                                Gas
                                       Seiscor Valve
                                        (Gas Sampling  Configuration)
Flowmeter
                                                       Female
                                                    Quick-Connect
                         {   |	Sample or
                               Calibration
                                   gas in
       Figure 2.  Flow schematic of nitrous oxide analytical system
                   (Step 1 - Purge of sample loop of GSV).
                                    324

-------
        Stripper Column
    (2' x 1/8"  SS, 10% OV-17
    on 80/100  Gas Chrom Q)
                         Seiscor Valve
                  (backflush configuration)
                                                          Perkin-Elmer
                                                            3920  B
                                                        Electron Capture
                                                            Detector
                                               I
                                         Analytical Column
                                     ••»  (61  x 1/8" SS, 80/100 Porapak Q)
                       Capillary
                       Restrictor
 10 ml
sample
 loop
                                                               Auxiliary
                                                              Carrier Gas
                                                               Carrier
                                                                 Gas
           mmmmmmmm
1—T
mmmmmmmmmmmmmmmmmmmmmmmmmt
                       Seiscor Valve
                       (Gas Sampling Configuration)
 Vent
                                                 Regulating
                                                   Valve
                                 Pump
                                                            f"l	 Sample or
                                                            .       Calibration
                                                        Female         gas in
                                                     Qui ck-Connec t
 Flowmeter
       Figure 3.  Flow schematic of nitrous oxide analytical  system
         (Step 2 - Inject  sample or calibration gas into system).
                                    325

-------
         Stripper Column
    (21  x 1/8"  SS,  10% OV-17
     on  80/100  Gas  Chrom Q)
                               Seiscor Valve
                        (backflush configuration)
                                                         Perkin-Elmer
                                                           3920  B
                                                       Electron Capture
                                                           Detector
                                       Analytical Column
                                        (61 x  1/8" SS,  80/100  Porapak Q)
                                                             Auxiliary
                                                             Carrier Gas
                       Capillary
                       Restrictor
 10 ml
sample
 loop
d-
-Q
                                                              Carrier
                                                                Gas
                             • ••
                             Seiscor Valve
                             (Gas Sampling Configuration)
 Vent
                                               Regulating
                                                 Valve
                                Pump
 Flowmeter
                                                       Female
                                                    Quick-Connec t
                                                      —Sample or
                                                      Calibration
                                                          gas in
         Figure 4.  Flow schematic of nitrous oxide analysis system
                (Step 3 - Backflush OV-17 stripper column).
                                    326

-------
the analytical column.  A typical  gas  chromatograph  trace  for a calibration
blend is shown in Figure 5.  A baseline  separation is  obtained and the N,0
peak area is obtained using a Hewlett-Packard 3354 GC  computer system.

     On gasoline and diesel samples  it is  necessary  that the backflush is
inlcuded in the analysis to prevent contamination  of  the analytical column
Using the system described, 60 seconds are allocated to the foreflush posi-
tion and 360 seconds are allowed for backflushing the  stripper column.  A
typical gasoline-CVS sample GC trace along with the  gas chromatograph'
operating conditions is presented  in Figure 6.
CONTROL SYSTEM

     The control of the two  Seiscor valves if accomplished by ATC timers and
ASCO electric solenoid valves.   The electrical schematic  for the control of
the Seiscor valves using these  timers  and electric solenoid valves is shown
in Figure 7.  The flow schematic for vacuum and pressure  lines to the Seiscor
valve are presented in Figures  8-10.

SAMPLE CALCULATIONS

     The quantification of ^0  in exhaust is based on a direct comparison
of the N20 in exhaust with a calibration blend of a known ^0 concentration.
Two basic assumptions are made  in these calculations that should be con-
sidered with other systems.   The first assumption is that the electron cap-
ture detector has a linearized  output  and that measurements are made within
the working range of the system.  These two parameters were verified for this
procedure using instruments  previously described.  Working within the linear
range of a given gas chromatograph equipped with an electron capture detec-
tor, the following relationship is true.

     Let Csam = ppm concentration of ^0 in sample

         Cstd = ppm concentration of ^0 in standard

         Asam = area of ^0  peak in sample

         Astd = area of N20  peak in standard

         Astd _ Asam
         Cstd ~ Csam

     Solving for Csam

     Csam = Asam x Cstd
               Astd
                                     327

-------
                             Operator B. Fanick
                        O.D. 0.063" I.D.    S.S
Instrument PE
Column
                                            Liq. Phase
                                             Support
                                              Carrier
     Packed with
     on  80/100
Porapak
                        Rotameter Reading
             /min. Held for
                 min.
                           min.. Prog to	 C at
  held for
Inlet  Room
                           Rotameter Rdg.
                           Rotameter Rdg.
                           Rotameter Rdg.
  Recorder 0.5  in/min speed
  Injection 10   ul indicated                _
     Sampling Device Gas Sampling Valve
                             mV.F.S.L+N-H
                               ul net
                                 L	 j.	il
    12   11    10     9    8     7    6  ~  5     4     3
                         Retention  time,  minutes
Figure 5.   Typical N2O calibration blend gas  chromatograph
                                                                       0

                                                                     trace.
                                     328

-------
Sample Gasoline Exhaust-CVS Date
Instrument PE 3920B Operator 5
11-10
-77
. FanicK
Column 6 ft. 1/8" O.D.O . 06 3 I.D. R . s .
Packed with - %wt. NA
on«U/lUU mesh PORAPAK Q
Run ISO @ 20 °C usina cc/min.
@ psig NA Rotameter Reading


Cii4/Ar

held @ °C ISO for min.. proq to
@ °/min. Held for min.. Proq to
held for min. (other)
Inlet Room <>c
Detector 325 °C BCD Type (other)
Hyd NA psig NA Rotameter Rdo.
Air NA psjg NA Rotameter Rdq.
( ) NA psig NA Rotameter Rdg.
Recorder 1 in/min speed 1 mV.F.S.
Injection 10 ul indicated 10 ul net
°Cat


^one
NA
HA
14A
Type
Liq. Phase
Support
Carrier

°C
°/min

Type

cc/min
cc/min
cc/min
Soltac Type
10
ul Actual
Sampling Device Gas Sampling Valve
12345678




           Retention  time,  minutes




Figure 6.  Typical gasoline-CVS exhaust sample.
                       329

-------
                  AC (-)
CO
co
o
T'l
  2  SJ
  3  !
  4  i

                            12
                            14
                            15
                            16 j
                            ATC
                           Timer
                                                                                       «NC



%






— -
I
•-••"1
1 1
il
4
;
Q



— <
, ,„•/"

L
^
J
\
^ 1




1









GSV
solenoid




g S
S :



I
|re
                11
                12
                14
                15
               I 16

               ATC
               Timer
                                                                                          backflush
                                                                                          solenoid
                        Figure  7.   Electrical  schematic  for nitrous oxide analysis  system.

-------
Vacuum
      cap
                	n_.
                      E2
Air Pressure (30 psi)
   P   V
    CSV
  Seiscor
                                                             CSV off
Vacuum
Air Pressure  (30 psi)   T
                                     cap
   ?   V

 Backflush
  Seiscor
Backflush  off
        Figure 8.  Flow schematic in electric solenoid valves
                      (Both valves de-energized).
                                 331

-------
 Vacuum
       cap
 Air  Pressure  (30 psi)
                                      cap
   P   V
    CSV
   Seiscor
  CSV  on
Vacuum
Air Pressure (30 psi)
Backflush off
         Figure 9.  Flow schematic in electric  solenoid valves
                 (CSV energized, backflush  de-energized).
                                  332

-------
Vacuum
Air Pressure  (30 psi)
                                                                CSV  on
Vacuum
cap
Air Pressure  (30 psi)   t
                                cap
                                                                P   V

                                                              Backflush
                                                               Seiscor
                                                       Backflush on
         Figure  10.   Flow schematic in electric solenoid valves
                         (Both valves energized).
                                   333

-------
Example 1:

     A 4.95 ppm N20 (in nitrogen) calibration blend was found to give 5291
area counts for the N20 Peak-  ^ exhaust sample was found to give 2674
area counts for the N20 peak.  Calculate the N20 in the exhaust sample.
            Asam x Cstd
     Csam =    Astd	

            2674 x 4.95
     Csam =
               5291

     Csam =2.50 ppm N20

Example 2:

     A 1.13 ppm N20 (in nitrogen)  calibration blend was found to give 1208
area counts for the N20 peak.  An exhaust sample was found to give 534 area
counts for the N20 peak.  Calculate the concentration of N20 in the exhaust
sample.

            Asam x Cstd
            534 x 1.13
     Csam =
               1208

     Csam =0.50 ppm N20

EQUIPMENT

     This analysis is performed using a gas chromatograph equipped with an
electron capture detector.  The detector, detector heater controls, electro-
meter, recorder and GC integrator are major electronic components in the
detection system.  A control console was fabricated to house the mechanical
hardware items that are necessary for the proper operation of the N20 analy-
sis syptem.  Figure 11 illustrates the complete analytical system for mea-
suring N20 in exhaust.  The major items that are included in each of these
systems is listed below:

Gas Chromatograph

     1.   Perkin-Elmer  Model 3920B gas chromatograph

     2.   Linerarized electron capture detector (BCD)

     3.   Leeds and Northrup Model W 1 mv recorder

     4.   Hewlett-Packard Model 3354 GC computer system
                                    334

-------
                                                                      A/D Converter
.,,
                                                                                            Recorder
                                                                 __Control Console
                                                                                          Calibration
                                                                                           Standards
                                 Figure 11.  Nitrous oxide analytical system.

-------
     5.   Hewlett-Packard Model 1865A A/D Converter

Control Console System

     1.   Seiscor valve - gas sampling configuration

     2.   Seiscor valve - backflush configuration

     3.   ATC timers, Model 325A364A10PX (2 ea)

     4.   Analytical column, 6' x 1/8" SS, 120/150 Porapak Q

     5.   Stripper column, 21 x 1/8" SS, 10% OV-17 on 80/100 Gas Chrom Q

     6.   ASCO -solenoid valve, Model 834501 (2 ea)

     7.   Brook flowmeter, R-2-15-A w/SS float,  0-150 scale

     8.   Metal Bellows MB-155 pump

     9.   Female quick-connect, stainless steel

    10.   Nupro Model 2M stainless steel regulating valve

    11.   Stainless steel tubing (0.01"ID)  for capillary restrictor

    12.   Miscellaneous stainless steel, copper and Teflon tubing (1/8"  and
          1/16")

    13.   Miscellaneous stainless steel and brass unions, tees,  etc.

    14.   Bud Classic II control console cabinet

    15.   Miscellaneous electrical on-off- switches
                                   336

-------
REFERENCES

Braker,  W. and Mossman, A.L., Matheson Gas Data Book. 5th Edition, East
Rutherford, N.J., 1971, pg. 431^

jay, B.  and Wilson, R., J. Appl. Phys., Vol. 15, No. 2, pg. 298, 1960.

De Grazio, R., J. Gas Chromatog., Vol. 3, pg. 204, 1965.

F & M Applications Chromatogram, 1942, F & M Scientific Division of
Hewlett-Packard Corp., Avondale, Pa.

Porapak Brochure, Waters Associates, Framingham, Mass.

Bethea, R., and Adams, F., J. Chromatog., Vol. 10, pg. 1, 1963.

Leithe, W., The Analysis of Air  Pollutants, Humphrey Sci. Pub., Ann Arbor,
Michigan, 1970, pg. 176.

Private communication between Dr. R.B. Zweidinger and Frank Black of EPA and
Harry Dietzmann of SwRI.

LaHue,  M.D.,  Axelrod,  H.D., and Lodge Jr.,  J.P.,  Anal.  Chem.,  Vol. 43,
pg. 1113,  1974.
                                     337

-------
        APPENDIX G




HYDROGEN SULPIDE PROCEDURE
            338

-------
              THE MEASUREMENT OF HYDROGEN  SULFIDE  IN EXHAUST


     The measurement of hydrogen sulfide in dilute automotive exhaust is
accomplished by bubbling dilute exhaust through  glass impingers containing
a buffered zinc acetate absorbing solution.   The hydrogen  sulfide reacts
with the zinc acetate to form zinc sulfide which remains in solution.  The
exhaust sample is collected continuously during  a  test cycle.  Upon com-
pletion of the test, the absorbing solution is treated with N,N dimethyl-
para - phenylene diamine sulfate and ferric  ammonium sulfate.  This reaction
produces a highly colored heterocyclic compound, methylene blue (3,9 - bis-
dimethylaminophenazothionium sulfate).   This colored solution is analyzed
with a spectrophotometer at 667 nm in a 1  cm or  4  cm pathlength cell.  The
results are quantified by comparison  to a  standard curve.  The minimum de-
tectable concentration is 0.01 ppm.

SAMPLING SYSTEM

     Two glass impingers in series, each containing 50 mH  (10 m£ buffered
zinc acetate solution and 40 m£ freshly vacuum boiled deionized water) of
buffered zinc acetate absorbing solution,  are used to collect exhaust
samples for the analysis of hydrogen  sulfide. A flow schematic of the
sample collection system is shown in Figure 1.   The two impingers together
trap approximately 99+ percent of the hydrogen sulfide. The temperature
of the impinger is maintained at 0-5°C by  an ice water bath, and the flow
rate through the impinger is maintained at 4 £/minute by the sample pump.
A dry gas meter is used to determine the total flow through the impinger
during a given sampling period.  The  temperature of the gas stream is
monitored by a thermocouple immediately prior to the dry gas meter.  A drier
is included in the system to prevent  condensation  in the pump, flowmeter,
dry gas meter, etc.  The flowmeter in the  system allows continuous moni-
toring of the sample flow to insure proper flow  rates during sampling.
When sampling diesel fueled vehicles, a filter,  located between the on-off
solenoid valve and the dilution tunnel, is used  to prevent diesel parti-
culate from contaminating the sampling system.   The lines  connecting the
filter to the dilution tunnel and the filter to  the solenoid valve are
heated to 175 °F.  Several views of the sampling  system are shown in Figure 2.

DESCRIPTION OF METHOD

     Hydrogen sulfide is collected in a buffered zinc acetate absorbing re-
agent from dilute automotive exhaust.  The highly  colored  species, methylene
blue, is generated by the addition of N,N  dimethyl-para-phenylene diamine
sulfate and ferric ion.  Methylene blue has an absorbance  maxima in the red
region of the visible spectrum.  The extinction  wavelength curve for methy-
lene blue is shown in Figure 3.

                                    339

-------
co
CD
       Sample
       Probe
                                                                          Gas Temperature
                                                                          Digital Readout
                                                             Sample
                                                              Pump
                                                                              IQIBIOI
                                                                                F lowme t e r
                         Dry
                         Gas
                        Meter
*
                                                                      Regulating
                                                                        Valve
       Dilute
       Exhaust
   Ice Bath
 Temperature
Digital Readout
                        Gas  Volume
                      Digital Readout
                           Figure 1.   Hydrogen sulfide  sample  collection flow  schematic.

-------
                            Front View
                                                           Digital
                                                           Readout
                                                           Flowmeter
                                                           Regulating
                                                           Valve
              Close-up of Upper Front
Figure 2.  The dilute exhaust sampling system for hydrogen sulfide.
                                341

-------
                                   Close-up of Impingers (Side View)
                                                               Pump
                    Rear View
Figure 2 (Cont'd).  The dilute exhaust  sampling  system for hydrogen sulfide.
                                    342

-------
2.0
 0.6
 0.4
 0.2
    350
400
450      500      550      600
      Wavelengths, nanometers
                                                         650
                                                     700
                                                                          750
          Figure 3.  Extinction-wavelength curve  of methylene blue.
                                     343

-------
     The analytical procedure for the determination of hydrogen  sulfide  in
dilute automotive exhaust consists of two major areas.  The  first  is  the
standarization and calibration of the standard sulfide ion solution.   The
second area includes the sample acquisition and color development  of  the
hydrogen sulfide sample.  Each will be discussed in detail below.  The
analysis flow schematic for this procedure is shown in Figure 4.

Standardization and Calibration

     The concentration of the standard sodium sulfide solution  (approximately
0.03 M sodium sulfide in deionized water) is determined by an iodometric me-
thod.  To three  (3) Erlenmeyer flasks, 10 m£ of the absorbing solution and
50 m£ of the sulfide solution are added and the resulting solution is  mixed.
Into one of the flasks 5 m£ of the 0.01 N iodine solution and 10 m£ of con-
centrated hydrochloric acid are added.  The resulting solution is  immediately
titrated to the starch endpoint with a standarized thiosulfate solution.
This procedure is then repeated for the remaining two flasks and for  two
blanks prepared with only the absorbing reagent and 50 mi of vacuum boiled,
deionized water.  The excess iodine in the solution is reacted with the
thiosulfate and the amount of sulfide ion present can be back-calculated.

     The thiosulfate solution used to titrate the sulfide solution is  stan-
dardized against primary standard grade potassium dichromate.  The potassium
dichromate is dreid in an oven at 150° to 200°C for 1 to 2 hours.  A weighed
0.10 to 0.15 g (0.001 mole)  portion of dried potassium dichromate  is placed
in a 500 m& Erlenmeyer flask and dissolved in 50 m£ of deionized water.  A
freshly prepared solution of 3 g (0.02 mole)  of potassium iodine,  5 mil of
6 N hydrochloric acid and 50 m& of deionized water is then added.  This
solution is gently swirled,  covered with a watch glass, and allowed to stand
for five (5) minutes. 'The sides of flask are then washed with deionized
water followed by approximately 200 m£ of deionized water.  The resulting
solution is titrated with the thiosulfate solution.  As the end point  is
approached, about 5 m£ of starch indicator is added.  The solution is blue
from the starch-iodine complex before the end point is reached.  At the end
point, there is a change in color from blue to green due to the production
of Cr (III) ion.  This standarized thiosulfate solution can then be used to
standarize the dilute iodine solution as well as determine the sulfide ion
concentration in the sulfide standard solution.

     A concentrated iodine solution (~0.1 N)  from which the dilute iodine so-
lution is prepared is standarized against primary standard grade arsenic tri-
oxide by an iodometric titration.  A weighed portion of 0,15 to 0.20g (0.001
mole) arsenic trioxide is placed into an Erlenmeyer flask.  Then 10 to 20 m£
of 1.0 M sodium hydroxide are added to dissolve the solid.  With a small
piece of blue litmus paper as the indicator,   1.0 M hydrochloric acid  is
added dropwise until the arsenic trioxide solution is slightly acidic.
About 1.0 g (0.01 mole)  of sodium bicarbonate is slowly added to prevent
the loss of solution due to the effervescence of carbon dioxide.   The  re-
sulting solution is diluted to about 100 m£.    About 2 m£ of the starch
indicator is added and the solution is then titrated with the 0.1  N iodine
solution.  During the titration, the top of the buret should be  covered  to
minimize the volatilization of iodine.  The standarization needs to be done

                                     344

-------
    CVS
      I
  Glass
 Impinger
   Permeation
      Tube
  Reagent
  Addition
   KCr2O7
   Primary
   Standard
  AS?03
  Primary
  Standard
     I
                                 I
   Color
 Development
  Thiosulfate
Standardization
Absorbance
  Reading
   Iodine
Standardization
                Calibration
                 Curve of
              Sulfide Standard
        Figure  4.   Hydrogen sulfide analysis flow schematic.
                                 345

-------
only once for the concentrated solution.

     The dilute iodine solution (-0.01 N) prepared by diluting the  concentra-
ted iodine solution, is standarized against the previously standarized thio-
sulfate solution.  To a beaker containing 25 m£ of the dilute iodine  solution,
10 m£ of concentrated hydrochloric acid is added and the resulting  solution
is immediately titrated with the standarized thiosulfate solution.  Just
before the endpoint, starch is added to serve as an indicator.  The dilute
iodine solution should be standarized daily.

     A Beer's Law curve is determined by adding 0.1, 0.25, 0.5, 1.0,  2.0,
3.0, and 5.0 mJl of the sulfide standard solution to seven separate  100 mjl
volumetric flasks with each containing 10 m£ of the absorbing reagent.  A
blank of the absorbing reagent is also prepared.  A 10 m£ portion of  the
amine solution is carefully poured down the side of each flask taking care
not :to introduce air bubbles. Then, 2 mH of the ferric ion solution  are
added and gently swirled.  The contents are then diluted to 100 m£ with
vacuum boiled, deionized water and placed in the dark to develop for  30
minutes.  This procedure should be repeated for the remaining flasks  and the
blank.  The absorbance is read at 667 mm in a spectrophotometer against the
reagent blank.  The concentration of the samples are used to determine a
best fit plot of the Beer's Law curve.  In the higher concentration ranges,
the curve is nonlinear and does not necessarily follow Beer's Law.

     Another more efficient means of generating a Beer's Law curve requires
the use of permeation tubes as a calibration standard.  The calibration
curve is generated by passing a diluent gas (preferably nitrogen)  over the
permeation tube and through a set of two impingers containing the buffered
zinc acetate absorbing reagent.  The length of time sampled is proportional
to the concentration in the set of impingers.   Not only is this a quick and
efficient means of calibration, but it also takes into account the collection
efficiency of the sampling technique.  The samples are developed by the tech-
nique described above.  The absorbance is read at 667 nm and this absorbance
is used to determine the calibration curve.  This best fit plot of the Beer's
Law curve is used to determine the concentration of the samples.  A quick
and easy method of providing a daily calibration for the instrument can be
obtained with this method.

Sample Acquisition and Color Development

     To two clean glass impingers, 10 m£ of acetate buffer and 40 m&  of vac-
uum boiled, deionized water are added.  These impingers are connected in
series in the sampling system.  During each test cycle, a portion of  the
diluted exhaust is bubbled through the absorbing reagent at a flow  rate of
4.0 5,/min.  Upon completion of each driving cycle, the impingers are  re-
placed with fresh ones.  To each of the collected samples, 10 m£ of the amine
solution is added through the top of the impinger and gently swirled.  Then
6 m£ (2 m& for samples from gasoline powered vehicles is sufficient)  of the
ferric ion solution is added in the same manner and mixed for 30 seconds.
The solution is quantitatively transfered from the impingers to a 100 m£
volumetric flask and diluted to 100 mJl with vacuum boiled, deionized  water.
                                    346

-------
Ifce color  development is complete in 30 minutes.  The proceduer is repeated
with the remaining samples.  After 30 minutes, the absorbance is measured on
a spectrophotometer at 667 nm against a reagent blank and the concentration
determined from the calibration curve.  The one and  four cm curvettes are
shown in Figure 5 and the spectrophotometer is shown in Figure 6.

     This  procedure is well documented and commonly  used for the analysis of
hydrogen sulfide.  With a 4 cm cell, the minimum detectable quantity is on
the order  of 0.01 ppm.  Extensive wet chemistry is involved to establish the
calibration curve but the sample analysis time is minimal after color develop-
ment.

CALCULATIONS

     This  procedure has been selected and developed  to determine the quan-
tity of hydrogen sulfide in dilute exhaust.  A Hewlett-Packard 67 program was
developed to reduce the time required for manual calculations.  The deri-
vation of the equations are given below and a copy of the steps in the pro-
gram are shown in Figure 7.

Derivation of Equation

     The first step is to correct the volume of exhaust sampled to a stan-
dard temperature, 68°F and pressure, 29.92"Hg, by use of the equation

                        P    XV      P     x V
                         exp    exp _  corr    corr
                           T              T
                            exp            corr

          V     = experimental volume of gas sampled in ft
          vexp  = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
          P     = experimental barometric pressure
          pexp  = 29.92"Hg
          TCOrr = experimental temperature in  °F + 460
          T6XP  = 68°F + 460 = 528°R
           corr

      Solving for V     gives:
                  corr
v
                  P     ("Hg) X V    (ft3)  x 528°R
                =  exp    *'     exp _ _
           corr       T     (°R)  x 29.92"Hg
                       exp
     The next step converts the  volume from cubic feet to cubic meters by
 use  of the conversion f actor; 1 cubic meter is equal to 35.31  cubic feet.

                      P     ("Hg)  x V    (ft3) x 528°R
          „     , 3.    exp    *'     exp __
           corr   ' ~ T    x  29.92"Hg x 35.31 ft3/m3
                       exp
                                                             (Equation 1)
                                     347

-------
         One and four cm pathlength curvettes
                Reading the absorbance
Figure 5.  The analysis of hydrogen sulfide in exhaust.
                          348

-------
; •
                                 Figure 6.   Beckman Model 25 spectrophotometer.

-------
        ls
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEY ENTRY KEY CODE Cnuut»,<
	






CjlO



	

	

».M



	
	

JJO









04U









ObJ






2
-:
R/S
X
STP_] 	
4
A
,1
+
RCL 1

X
STO 2
BfT. •>
R/R
X
1
•
0
6
3
X
STO .J 	
RCL 2
	 X 	
.
o
h
,
X
R/S
RCL 3
-t-
R/S
1
4
3
4
^.
R/S
h kTO










02
81
84

J101_ ..
04
(16
O'l
fcj.
34 01
81
71
13 u2
JiUZ 	
„ H4
71
01
63
Oil
06
0)
71
1.) 01
.....
34 02 s
01

ULI
ilu
111
71
84
34 OJ
61
84
01
04
03
04
81
84
.35 22










ill nmni/ie VOJ, It-
Ill baiometric, "Hg
Ii. bam. le TBmp, "F
Tn Total Vol of Sol
In Sample Cone.
Bubbler itl.
-Out S
-------
     The concentration of sulfide ion in ygS  /m£ is obtained from the Beer's
Law Plot of the absorbance.  The concentration of hydrogen sulfide in
ygH2S/m& is obtained by multiplying the concentration in ygS  /m& by the
ratio of the formula weight of hydrogen sulfide to the formula weight of
sulfide ion.

               / o     ~2- , n   formula weight HoS(yg/y mole)
          ygH9S/m£ = ygS  /m£ x - - - - .  . ^ A ,   . - r— r
          Hy * '       ^         formula weight S^~(yg/y mole)

                                34.080 yg H2S/y mole
                              X 32.064 ygS2-/y mole

                   = ygS2~/m£ x 1.063
                                      yg &

     To obtain the total amount in yg of hydrogen sulfide, the concentration
in ygH2S/m£ is multiplied by the total volume of solution (TVS) .  This is
the volume to which the absorbing solution, amine solution, and ferric ion
solution have been diluted with deionized water.  In this case the volume
is 100 m£.
          ygH2S = ygH2S/m£ x TVS

                = ygH2S/m£ x 100 m£

                = ygS2~/m£ x 1.063 Uq"§f  x 100
                                                            (Equation 2)

     To obtain ygH2S/m3, Equation 2 is divided by Equation 1 to give:
  ygH2s/m3  =
WS   x 1.063 yg*^! x 100 m£ x T    x 29.92"Hg x 35.31  ft3/m3
                                                             (Equation 3)
                             P    ("Hg)  x V    (ft3) x 528°
                              exp    3     exp
     To find the concentration of hydrogen  sulfide  in ppm, the density of
hydrogen sulfide at 68° and 29.92"Hg is needed.  The fifth edition of the
Matheson Gas Data Book lists the density of hydrogen sulfide at 0°C and 1
atmosphere at 1.5392 g/£.   if the volume of l£ of gas is corrected for
temperature,
                      r ODOrj
                   (32'F + 460

it can be divided into the weight of the gas to give the density at 68°F

          1.5392 g
          1>073 £  = 1-434 g/£ = 1434
                                    352

-------
     To obtain the concentration of hydrogen sulfide in ppm, the concentra-
tion in yg/mj is divided by the density in yg/m£
          ppra = yg H2S/m  T yg/m£ =
                                    m
                                                             (Equation 4)
     At this point, the concentration can be expressed in ygH2S/m3 (Equation
3) and ppm (Equation 4) at 68°F and 29.92"Hg from the raw data.

Sample Calculations

     Assume that a set of six samples was taken  from dilute exhaust during
several different driving cycles.  The volume of dilute exhaust sampled
during the first test was 3.185 ft3 at a barometric pressure of 29.41"Hg and
a temperature of 77°F.  The analysis was conducted and Bubbler #1 was found
to contain 9.2 ygS2~/m£ while Bubbler #2 contained 0.8 ygS2~/m£

     For Bubbler #1

       ,   ygS2~/m£ x 1.063 ^"2- x 100 m£  x T    x 29.92"Hg x 35.31 ft3/m3
      / 3                    ygs               exp
VlgH2S/m  = - « - * - - -
                                P     ("Hg)  xV     (ft  ) x 528°R
                                 exp          exp
              f\

       9-2 ygS  /m& x 1.063
          29.41" Hg x 3.185 ft3

       100 m& x (460 + 77°F) x 29.92"Hg x 35.31 ft3/m3
     X                     528°R

     = 11218 yg/m3

     The concentration for Bubbler #2 is calculated in the same manner using
                                    n _
 the appropriate concentration in ygS  /m£

     For Bubbler #2

     ,3   0.8 ygS2~/m£ x 1.063 ygHpS/ygS2"
 UgH2S/m  = 	29.41" Hg x 3.185 ft?*
                                                      3   3
           100 ml, x (460 + 77°F) x 29.92"Hg x  35.31 ft /m
         X                       528°R

         =975 yg/m3

 Total yg H2S/m3 = Cone (Bubbler #1) + Cone  (Bubbler #2)

                = 11218 + 975

                = 12193 yg/m3
                                     353

-------
     ppm H-S = ygH S/m  T density yg/m£

             = 12193 ygH2S/m3 4- 1434 Ug/m3

             = 8.50 ppm

The values for the six sets of sample data  are  given  in Figure 8 along with
the data sheet.  The calculations were carried  out manually and with the
HP-67 program.  The program will be used to calculate both ygH2S/m3 and ppm
H2S and to insure rapid data turnaround.  The values  used in these calcula-
tions were picked at random and may  not represent the expected values in
exhaust.

LIST OF EQUIPMENT

     The equipment required in this analysis is divided into two basic cate-
gories.  The first category involves the sample 'acquisition using the glass
impingers.  The second category contains equipment related to the analysis
of the sample once it has been obtained. The  individual items in each cate-
gory are listed below:

Sample Acquisition

     1.    Sample pump, Thomas Model  106  CA18,  capable of  free flow capacity
           of 4
      2.    Glass  impingers,  Ace Glass Products, Catalog No.  7530-11,  29/42
           bottler joints,  18/7 arm joints.

      3.    Flowmeter,  Brooks Instrument Division,  Model 1555, tube  size
           R-2-15-C,  graduated 0-15, sapphire float, 0-5 £/min range.

      4.    Regulating valve, Nupro 4MG, stainless  steel.

      5.    Dry gas meter,  American Singer Corporation, Type  Al-120, 60 CFH
           capacity.

      6.    Teflon tubing,  United States Plastic Corporation, 1/4" OD x 1/8"
           ID and 5/16" OD x 3/16" ID.

      7.    Teflon solenoid valve, The Fluorocarbon Company,  Model DV2-144NCS.1.

      8.    Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, con-
           nectors, etc.

      9.    Drying tube, Analabs, Inc., Catalog No. HGC-146,  6" long, 1/4"
           brass  fittings.

     10.    Digital readout for dry gas meter.

     11.    Miscellaneous electrical switches, lights, wriings, etc.

                                     354

-------
                                                               TEST DATE:  11-10-79 VEHICLE:  Practice
SWRI PROJECT NO. 11-1234   TEST NO.	001                 	


FUEL: EM~237 CVS NO.   3     TUNNEL SIZE:  18"    DRIVER:  D.A.T.    MILES:
                                                                                             1000
                 SAMPLE COLLECTION  BY:  G-°-


                 GENERAL COMMENTS:
                                 CHEMICAL ANALYSIS  BY:  W.M.S-CALCULATIONS BY:
CO
on
cn
Test No.
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Total Vol. of Solution, mi
o
Sample Cone-Bub. #1, ygS /mfi,
Sample Cone-Bub . #1 , ygH2S/m3
Sample Cone-Bub. #2, ygS2~/m£
Sample Cone-Bub. #3, pgH2S/m3
Total Sample Cone, ygH2S/m3
Total Cone, ppm H2S


1
Cold FTP
3.185
29.41
77
100
9.2
11200
0.8
975
12200
8.50


2
Hot FTP
3.486
28.66
75
100
7.3
8310
6.2
228
8540
5.95


3
SET-7
3.508
29.33
80
100
10.4
11600
1.1
1230
12800
8.95


4
HFET
1.926
29.40
85
100
3.1
6340
0.3
614
6960
4.85


5
NYCC
1.525
29.10
90
100
1.0
2630
0.1
263
2900
2.02


6
Backgrour
15.826
29.04
77
100
0.5
124
0.1
24.8
149
0.10


                              Figure  8.   Raw data sheet for hydrogen sulfide analysis.

-------
   12.   Six-channel digital thermometer, Analog Devices,  Model 2036/J/l.

   13.   Iron/Constantan Type J, Thermo  Sensors  Corporation,  single thermo-
         couple,  1/4" OD stainless  steel metal sheath.

   14.   Modified 25 mm A-H Microanalysis filter holder, Millipore, Catalog
         #XX50 020 00.

   15.   Fluoropore 25 mm  filters,  Millipore, Catalog #FHLP 025 00, 0.5
         micron pore size.
    16.    Flexible  heavy  insulation heating  tape, Briskeat ,  width  1/2 inch,
          length  48 inches.

    17.    Temperature  Controller, Athena,  100-600  °F

    18.    Heated  TFE Teflon  hose, Technical  Heaters, Inc., 5'  x 1/4", tem-
          perature  limit  400°F.

Sample Preparation

     1.    Pipet-aid, Order No. JX-7290-

     2.    Class A,  10  m£  volumetric pipets.

     3.    Class A,  15  m£  volumetric pipets.

     4.    Class A,  20  m£  volumetric pipets.

     5.    Class A,  25  m£  volumetric pipets.

     6.    Class A,  100 m£ volumetric  flask.

     7.    Class A,  500 m£ volumetric  flask.

     8.    Class A,  1000 m£ volumetric flask.

     9.    Class A,  25  m£  burets with  Teflon  Stopcock.

    10.    Class A,  100 m£ graduated cylinder.

    11.    Micro burets, .5 m£ Teflon stopcock.

    12.    Erlenmeyer flask,  250 m£.

    13.    Erlenmeyer flask,  500 m£.

    14.    Dropping  bottle, 60 m£ ground  glass pipet and rubber bulb.

    15.    Beaker, 100  m£.
                                    356

-------
    16.   Repipet dispenser, 2 m£.

    17.   Repipet dispenser, 10 m£.

    18.   Centrifuge tube, 15 m£.

    19.   Watch glass.

Instrumental Analysis

     1.   Beckman Model 25 spectrophotometer with recorder.

     2.   One cm pathlength disposable curvettes, CI regular.

     3.   Beckman UV silica cell, 40mm pathlength, No. 580016.

LIST OF  REAGENTS

     This procedure requires the sample collection in glass impingers with
a buffered  zinc acetate absorbing solution  (0.25 M zinc acetate and 0.10 M
sodium acetate) .  An amine reagent  (0.005 M, N,N dimethyl-para-phenylene
diamine  sulfate and 3.50 M sulfuric acid) and a ferric ion solution (0.25 M
ferric ammonium sulfate and 0.5 M sulfuric acid) are mixed with the absorb-
ing solution.  Sodium sulfide is used as the standard.  An iodometric
titration is used to standarize the sodium sulfide.  All of the reagents
used in preparing these solutions are listed below along with the manu-
facturer and the quality.

     1.    Zinc acetate, dihydrate formula weight = 219.49; chemical formula =
          Zn(C,,H 0 ) -2H O, crystal, "Baker Analyzed" reagent.
              ^ O ^ ^   &

     2.    Sodium acetate, anhydrous formula weight = 82.03; chemical formula=
         NaC2H3O2, analytical reagent grade, powder, Mallinckrodt Code 7372.

     3.   N, N dimethyl-para-phenylene diamine sulfate, formula weight =
          370.47; chemical formula = (NH2C6H4N(CH3)252*H2S°4) ' 98 Percent
         minimum by titration and spectro analysis, Eastman Code 1333.

     4.    Sulfuric acid, formula weight = 98.08; chemical formula = H,,SO4,
         ACS analytical reagent grade, Mallinckrodt Code 2876.

     5.   Ferric ammonium sulfate, formula weight = 482.19; chemical formula=
          Fe(NH4) (S04)2*12H20, ACS analytical reagent grade, crystals, Mallin-
          ckrodt Code 5064.

     6.    Sodium sulfide, formula weight = 240.18; chemical formula =
         Na S-9H 0, ACS analytical reagent grade, crystals, Mallinckrodt
         Code 8044.

     7.    Iodine, formula weight = 253.81; chemical formula = I,» ACS
         analytical reagent grade, resublimed, Mallinckrodt Code 1008.


                                    357

-------
     8.    Hydrochloric acid,  formula weight = 36.46; Chemical formula =
          HCl,  ACS reagent,  assay 36.5-38.0 percent HC1, Eastman Code 13061.

     9.    Sodium thiosulfate, formula weight = 248.18; chemical formula =
          Na2S2O3*5H2O, ACS  analytical reagent grade, crystals, Mallinckrodt
          Code  8100.

    10.    Starch soluble;  chemical formula=  (C6H10O5)n, certified ACS,
          powder, Fisher Code S-516.

    11.    Arsenic trioxide primary standard, formula weight = 197.82;
          chemical formula = AS203, powder, ACS analytical reagent, Mallin-
          ckrodt Code 3668.

    12.    Potassium iodine,  formula weight = 166.01; chemical formula =
          KI, compacted crystal, "Baker Analyzed" reagent.

    13.    Sodium hydroxide pellets, formula wieght = 40.00; chemical formula=
          NaOH, caustic soda, ACS analytical reagent grade, Mallinckrodt
          Code  7708.

    14.    Sodium bicarbonate, formula weight = 84.01; chemical formula =
          NaHCO_, powder,  ACS analytical reagent grade, Mallinckrodt Code 7412.

    15.    Sodium carbonate anahydrous powder, formula weight = 105.99;
          chemical formula = Na2CO3» ACS analytical reagent grade,  Mallin-
          ckrodt Code 7521.

    16.    Potassium dichromate,  crystal, primary standard, formula weight =
          294.22; chemical formula = K Cr O , "Baker Analyzed" reagent.

    17.    Litmus test paper,  blue, reagent ACS, Fisher Code 14-875.

    18.    Glycerol, formula  weight = 92.10; chemical formula =
          HOCH2CH(OH)CH2OH,  ACS  analytical reagent, Mallinckrodt Code 5092.

PREPARATION OF REAGENTS

     The various reagents needed for the analysis of hydrogen sulfide can be
divided into two separate categories.  The first category includes the
reagents for sample acquisition  and color development.  The other category
encompasses the standard solutions and solutions used for standardization.
The chemicals used to make these solutions are ACS analytical reagent grade
with the exception of potassium  dichromate and arsenic trioxide which are
a primary reagent grade.

Sample Acquisition and Color Development

     Absorbing Reagent - The absorbing reagent is prepared by dissolving
54.9 g (0.25 mole) of zinc acetate, 8.2 g  (0.10 mole) of sodium acetate,
and 40 m£ (0.56 mole) gylcerol,  in vacuum boiled, deionized water and di-

                                    358

-------
luting to 1 liter. A 2 ml portion  of 0.05  M sodium sulfide  solution is added
dropwise to the diluted solution with vigorous  shaking.   This removes traces
of heavy metals by precipating them as their insoluble sulfides   This
solution is then set aside overnight.  The resulting solution is filtered
through a fine textured paper with the first 50 m£ portion  being discarded
This solution is stable for at least one  (1)  week.  A gradually developing"
cloudiness is of no consequence.

     Amine Solution - A 0.005 M solution is prepared by  dissolving 0.93 g
of N,N dimethyl-para-phenylene diamine sulfate  in 75 m£  of  deionized water.
Then, 197 m£ of concentrated sulfuric acid is slowly added  and the mixture"
is allowed to cool.  The resulting solution is  diluted to 1 liter with
deionized water.  This solution is stable  for about six  (6)  months.

     Ferric Ion Solution - The ferric ion  solution is prepared by dissolving
120.6 g (0.25 mole) of ferric ammonium sulfate  in 750 m£ of deionized water.
A 27 m£ portion of concentrated sulfuric acid is added and  the solution is
allowed to cool before diluting to 1 liter with deionized water.  The solu-
tion is stable for one (1) month.

Standards and Standarization Solutions

     Sulfide Standard Solution - Approximately  8 g (0.03 mole) of sodium
sulfide is rinsed with vacuum boiled, deionized water to remove traces of
sulfide from the surface of the crystals.   The  crystals  are then dissolved
in vacuum boiled, deionized water  and diluted to 1 liter to give a concen-
trated solution.  A more dilute solution of sulfide ion  is  prepared by
diluting 10 m& of the concentrated solution to  1 liter with vacuum boiled,
deionized water.  This solution is standardized by iodometric titration.
The approximate concentration is 10 ppm sulfide ion by weight.  This solu-
tion should be prepared immediately before use.

     Thiosulfate Standard Solution - Approximately 2.5 g (0.01 mole) of
sodium thiosulfate is dissolved in 500 m£  of freshly vacuum boiled, de-
ionized water in a 1 liter volumetric flask. A 0.1 g (0.001 mole) portion
of sodium carbonate is added and the solution is stirred until dissolved.
The solution is then diluted to 1  liter and stored in the dark.  Addition
of a small amount of sodium carbonate pervents  the formation of hydrogen
sulfite ion from thiosulfate ion in the presence of acid.  The addition of
such substances as chloroform, sodium benzoate, or mercury  (II) iodine in-
hibits the growth of bacteria.  This solution is stable  for several weeks
but should be discarded if it becomes turbid.  It is standarized with a
potassium dichromate primary standard.

     Starch Indicator - The indicator is prepared by making a paste of 0.5 g
of soluble starch in 2 or 3 m£ of  boiling  water.  The resulting slurry is
slowly poured into 50 mJl of  boiling deionized  water and heated until clear
(about 2-3 minuted) .  The solution is then cooled and centrifuged for several
minutes.  The supernatant liquid is decanted into a clean,  60 mJl reagent
bottle equipped with a pipet.  This aqueous starch solution will decompose
due to bacterial action in several days.   This  can be prevented by storing


                                     359

-------
the indicator under sterile conditions, the addition of mercury  (II) iodide
to inhibit the bacterial action, or by preparing fresh daily.

     Iodine Solution - About 25 g  (0.15 mole) of potassium iodide is dis-
solved in 10 m& of deionized water.  Then, 12.7 g  (0.05 mole)  of iodine
crystals are added and dissolved with occasional stirring.  The  solution is
filtered and diluted to 1 liter with deionized water.  The resulting solu-
tion is approximately 0.1 N and is used to prepare the 0.01 N solution.  The
0.1 N solution is standarized against arsenic trioxide.  The dilute iodine
solution is prepared by diluting 10 m& of the 0.1 N  solution to 100 m&
with deionized water.  This solution is standarized with a previously stan-
darized thiosulfate solution.  The dilute iodine solution is prepared and
standarized daily because of the volatility of iodine and oxidation by dis-
solved oxygen.
                                     360

-------
REFERENCES
Stecher,  P.G.,  (Ed.), The. Merck Index. An Encyclopedia of Chemicals and Druas
8th Edition,  Merck and Co., Inc. Rahway, N.Y., 1968, pg. 545.                '

Braker, W.  and Mossman, A.L. , Matheson Gas Data Book, 5th Edition  East
Rutherford, N.J., 1971, pg.               ~     "
Bethea, R.M. ,  J. Air Poll. Cont. Assoc., Vol. 23, pg. 710, 1973.

Bamesberger, W.L. and Adams, D.F., Tappi, Vol. 52, pg. 1302, 1969.

Falgout, D.A.  and Harding, C.I., J. Air Poll. Cont. Assoc., Vol. 18, pg. 15,
1968.

Biles, B., Brown, C., Nash, T. , J. Phys, E.:  Scientific Instruments, Vol.  7,
pg. 309, 1974.

Wellinger, R.  and Fiever, P.M., Am. Ind. Hyg. Assoc. J. , Vol. 35, pg. 730,
1974.

Natusch, D.F.S., Sewell, J.R. , and Tanner,  R.L. , Anal. Chem., Vol. 46, pg.
410, 1974.

"Trace Sulfue  in Hydrocarbons by the Raney  Nickel Method," Analytical Method
Information,  Analytical and Informational Division, Exxon Research and
Engineering Co., September 1975.

Pierce, R.W.,  ISA Tran., Vol. 13, pg.  291,  1974.

Axelrod, H.D., Gary, J.H., Bonelli, J.E., and Lodge, J.P., Anal. Chem.,
Vol. 41, pg.  1959, 1969.

Natusch, D.F.S., Klonis, H.B., Axelrod, H.D., Teck, R. J. , Lodge, Jr., J.P.,
Anal. Chem.,  Vol. 44, pg. 2067, 1972.

Bock, R. and Puff, H., Z. Anal. Chem., Vol. 240, pg. 381. 1968.

Nauman, R. and Weber, C., Z. Anal. Chem., Vol. 253, pg.  Ill, 1971.

Ehman, D.L. ,  Analy. Chem., Vol. 48, pg. 918. 1976.

Hseu, T.M. and Rechnitz, G.A. , Anal. Chem., Vol. 40, No. 7, pg. 1054.

Hseu, T.M. and Rechnitz, G.A. , Anal. Chem.,  Vol.  40, No.  7, pg. 1054, 1968;
see also correction:  Ibid. Vol. 40, No. 11, pg. 1661, 1968.

Baumann, E.W., Anal. Chem., Vol. 46, No. 9, pg.  1345, 1974.

Blanchette, A.R., and Copper, A.D. , Anal. Chem., Vol. 48, pg. 729, 1976.
                                     361

-------
"Analysis of H S and SO  in the ppb Range,"  Bulletin 722C, Supelco, Inc.,
Belefonte, Pa., 1975.

Setter, J.R., Sedlak, J.M., Blurton, K.F., J. Chromatog. Sci,, Vol. 15,  pg.
125, 1977.

Greer, D.G. andBydalek, T.J., Environ. Sci. Tech., Vol. 7, pg. 153, 1973.

Obermiller, E.L. and Charlier, G.O., Anal. Chem. , Vol. 39, pg. 397, 1967.

Bethea, R.M. and Meador, M.C., J. Chromatog. Sci., Vol. 7, pg. 655, 1969.

Jones, C.N., Anal. Chem., Vol. 39, pg. 1858, 1967.

Wilhite,  W.F. and Hollis, O.L., J. Gas Chromatog., Vol. 6, pg. 84, 1968.

Thornsberry Jr., W.L., Anal.  Chem., Vol.  43, pg. 452, 1971.

Adams, D.F., Jensen, G.A., Steadman, J.P., Koppe, R.K., and Robertson, T.J.,
Anal. Chem., Vol. 38, pg.  1094,  1966.

Martin, R.L. and Grant, J.A.,  Anal. Chem., Vol.  37, pg. 644, 1965.

Adams, D.F., Bamesberger, W.L.,  and Robertson, T.J., J. Air Poll. Cont.
Assoc., Vol. 18, pg. 145,  1968.

Levaggi,  D.A., Siu, W., and Fedlstein, M., Advances in Automated Analysis,
Vol. 8, pg. 65, 1972.

"Hydrogen Sulfide in Air Analytical Method," H.L.S., Vol.  12, pg. 362, 1975.

Bamesberger, W.L. and Adams,  D.F., Environ. Sci. Tech., Vol. 3, pg. 258, 1969.

Gustafsson, L., Talanta, Vol.  4, pg. 227,  1960.

Moest, R.R., Anal. Chem., Vol. 47, pg, 1204, 1975.

Buck, M.  and Gies, H., Staub,  Vol. 26, pg. 1966.

Jacobs, M.B., Braverman, M.M., and Hochheiser, S., Anal. Chem., Vol. 29, pg.
1349, 1957.

Cave, G.C.B., Tappi, Vol. 46,  pg. 1, 1963.

Budd, M.S. and Bewick, H.A., Anal. Chem., Vol. 24, pg. 1536, 1952.

Bostrom,  C., Air and Water Poll. Int. J., Vol. 10, pg. 435, 1966.

Flamm, D.L. and James, R.E., Environ. Sci. Tech. pg. 159,  1975.
                                    362

-------
   APPENDIX H





AMMONIA PROCEDURE
      363

-------
                    THE MEASUREMENT OF AMMONIA IN EXHAUST
      The measurement of ammonia in dilute automotive exhaust is accomplished
 by bubbling  dilute  exhaust through glass impingers containing dilute sulfuric
 acid.  Ammonia is complexed by the acid to form a stable sulfate salt which
 remains in solution.  The exhaust sample is collected continuously during a
 test cycle.   A sample from the impinger is analyzed for ammonia by the use
 of an Ion  Chromatograph.  The concentration of ammonia is calculated by
 comparison to a standard.

 SAMPLING SYSTEM

      Two glass impingers in series, each containing 25 m£ of 0.01 N sulfuric
 acid, are  used to  collect exhaust samples for the analysis of ammonia.  A
 flow schematic of  the sample collection system is shown in Figure 1.  The two
 impingers  together  trap approximately 99+ percent of the ammonia.  The temp-
 erature of the impinger is maintained at 0-5°C by an ice water bath, and the
 flow rate  through  the impinger is maintained at 4 Vminute by the sample pxanp.
 A dry gas  meter is  used to determine the total flow through the impingers
 during a  given sampling period.  The temperature of the gas stream is
 monitored by a thermocouple immediately prior to the dry gas meter.  A
 drier is  included  in the system to prevent condensation in the pump, flow-
 meter, dry gas meter, etc.  The flowmeter in the system allows continuous
 monitoring of the  sample flow to insure proper flow rates durinq sampling.
 When sampling diesel fueled vehicles,  a filter,  located between the on-off
 solenoid valve and  the  dilution tunnel,  is  used  to prevent diesel particulate
 from contaminating  the  sampling system.   The  line connecting the filter to the
 dilution tunnel and the line connecting the filter to the solenoid valve are
 heated to  175°F in  order to prevent water from condensing in the sample lines.
 Several views of the sampling system are shown in Figure 2.

 PROCEDURE

      Ammonia  in exhaust is  collected in two impingers connected in series with
 each impinger containing 25 mJl of  0.01 N H2SO4.   These two impingers trap 99+
 percent of the ammonia.  After the  acidification of ammonia which takes place
 at  ice  bath temperatures (0-5°C),  the  ammonia samples are poured into poly-
 propylene bottles and stored.   The  samples  are then ready for NH4+ analysis
 on  the  ion chromatograph (Figure 3).   Approximately 2 m£ of sample are used
 to purge a 0.1 m£ sample loop,  after which  0.1 m£ of sample is injected into
 the  eluent stream.    Separation of  ions occurs in the separator (analytical)
 column.  The background conductance of the  eluent (0.0075 N HNO3)  is neutral-
 ized  in the suppressor  solumn.   The 9  x 250 glass cation suppressor column is
packed with AG- X10, a  strong base  ion exchange  resin in the hydroxide form.
A patented resin containing a sulfonic acid cation exchanger is packed into

                                     364

-------
                                                                        Gas  Temperature
                                                                        Digital  Readout
CO
cr>
en
     ample
     Probe
                  Regulating
                    Valve
             Heated Lines
                        Off  Solenoid Valve
                                                                              Flowmeter
     Dilute
    Exhaust
Ice Bath Temperature
  Digital Readout
                                              Dry
                                              Gas
                                             Meter
                                                                                             hla|3|q|5l
                                            Gas Volume
                                          Digital Readout
                               Figure  1.  NH  sample collection flow schematic.

-------
                Front View
                                                Digital
                                                Readout
                                                Flowmeter
                                                Regulating
                                                Valve
Close-up of Upper Front
    Figure 2.  Ammonia sampling system
                  366

-------
Solenoid
Impinger
Ice Bath  _
                              Close-up of Impingers  (Side View)
                                                          Dry Gas Meter
                                                          Pump
                   Rear View
            Figure 2  (Cont'd).  Ammonia  sampling  system.
                                367

-------
     Polyethylene  Storage
             Bottle
                                       Suppressor
                                        Column
                                          Ion
                                      Chromatograph
Recorder
    Figure 3.  NH  ion chroma.tograph-
                   368

-------
the 6 x 250 mm glass separator  column and the 3 x 150 mm precolumn    After
the cations are separated in the  analytical column they pass into the sup-
pressor column in the dilute nitric  acid eluent.  The hydroxide  form  of the
suppressor resin neutralizes the  acid and then converts the cations to their
hydroxides.

          HNO  + Resin - OH^=± Resin - NO  + H 0

          Cation  + NO3  + Resin  - OH ^   ~- Resin - N03 + Cation"1"  + OH~


      The  conductivity cell produces  a  signal  for  the  species of interest,
 NH4OH,  but doesn't "see" the neutralized eluent,  deionized water.  Con-
 ductance  is interpreted as a recorder  trace  (chromatogram) or as peak area
 by the Hewlett-Packard 3354 computer system.   Figure  4  depicts the analysis
 flow schematic of the entire ammonia procedure.   Two  chromatograms produced
 by the analysis of a sample and a standard are shown  in Figures 5 and 6.
 After the 12-20 minute analysis the  collection conditions and areas of sample
 and standard are used to compute NH3 concentration from a Hewlett-Packard 67
 program.

 CALCULATIONS

      The purpose of this procedure is  to determine the  concentration of
 ammonia in automotive exhaust.  To do  this, ammonia is  converted to the
 protonated form, NH4+, which is measured on the ion chromatograph.  The cal-
 culations involve correcting the measured concentration of NH^+ to NH3 at a
 desired temperature and pressure.  These calculations are carried out in a
 minimum amount of time by using a Hewlett-Packard 67  calculator program.   A
 copy of the program is shown in Figure 7.  Information  from  the data sheet
 (Figure 8)  is entered into the calculator and the ammonia concentration in
  W y*3 and ppm NH 3 is computed.  For  illustration, two examples using infor-
 mation from the data sheet will be included at the end  of this section.

      The first step in the derivation  of the  equation for the calculation of
 ammonia concentration involves comparison of  the  sample to a standard of
 known concentration.  A standard with peak size close to that of the  sample
 is selected.

                             PAo=            pAsa  x cst
                               sa
                           = c^T               PAst

 where PAst = peak area of the standard

       PAsa = peak area of the sample
                                              yg NH4+
       C«,4.   = concentration of the standard  (—-~	)
       • O l>                                       iUA/

                                           yg NH4+
       Csa   = concentration of the sample  (	—)
                                      369

-------
        CVS
           I
      Glass

     Impinger
                     Unused Sample
                       Saved as
                        Needed
 Sample Analysis
 in Ion Chromatograph
with Conductivity Cell
                        A/D
                      Converter
           I
       Recorder
                    Hewlett-Packard
                        3354
                    Computer System
         Figure 4.
NH  analysis flow schematic.
                           370

-------
Sample NYCC B-2 23111 6-19-79
Attenuation lOumho
Date 8-9-79
Eluent 0.01 N HNO-^  Sample Loop  Size 0.2 mil
Eluent Flowrate  138 mjt/min  Chartspeed  12
Pre column  3 x  150 mm,  glass
Separator  Column 6 x  250  mm,  glass
Precolumn and Separator Column Packing Patented
 resin
Suppressor Column 9  x 250 mm,  glass
Suppressor Column Packing AG IX- 10
         15           10
             Retention time, minutes

         Figure 5.  Sample  chromatogram.


                       371

-------
  Sample 2.0
2 Attenuation 3yimho
  Eluent 0.0012 N H->SC>4 Sample Loop Size  0.1 mi
  Eluent Flowrate 167 ml/hr Chartspeed  12  in/hr
  Analytical Precolumn 3x150 mm glass
  Analytical Column 6x250 mm glass
  Column Packing Patented resin-sulfonic acid
CO                cation exchanger
  Suppressor Column 9x250 mm glass
 '• Packing AG 1X-10
 15          10           5           0
      Figure  6.   Standard chromatogram

                    372

-------
             Insii unions
8,
Aiuuoiiia In Exhaust ^*
srgp
j
*
°>
'J3
1
2
3
4
i 5
b
7
i !»
f
y
10


I

I








1



INSTRUCTIONS
Switch to on; Switch to t ui»

Peed Card in from riyht tu lull
Set Decimal place
Input Sample Volume
Input Barometric Pressure
Input Sample Temperature
Input Dilution Factor
I n£u t Absorb i ng Reagen t Vo 1 uine
Input Standard Concentration
Input Standard Area
Input Sample Area
Output Sample Concentration
Output Sample Concentration











. . -
. _ ..
-
"


INPUT
DATA UNITS




ft3
"ttq
T

ml
ligNH^/me








.

......



- -
.... . .





KtVb
1
! i
I
,| -!>,-, \
I A :
\ K/S; | ,
| R/S ! 1
1 R/S . '
i K/S' '
R/Si i
1 R/S|
R/si :
1 R/S! :
II
h i 1 RTH
i! 1
i i i
i i
i i
i i i
i ii
I i
il i
1 1 i
i
i
i I
1 M
i •!
: I i
1 ' '
1 ! i
1 ! '
1
1 1
1 !
i i
OUTPUT












M'lNllj/iu-1
I'I'M

















Figure  7.  HP-67 user instructions.
               373

-------
STEP  KEY ENTRY   KEY CODE
                        COMMENTS
                                    STEP  KEY ENTRY  KEY CODE
                                                            COMMENTS
001







010









020









030









040









050






f T.HT. A 	
3
5

1
1
i
5
2
a
X
?
q
.
9
2
T
R/S
X
STO 1
H/S
4
6
0
•f
RCL 1
h X? v
i
h 1/X
R/S
X
R/S
X
R/S
X
R/S

R/S
X
0
.
9
4
4
X
R/S
0
•
0
0
1
4
1
X
R/S
h RTN
31 25 .IX,
03
05
83
ni
01
fll
05
02
08
71
n->
nq
83
09
02
81
84
71
33 01
84
04
06
00
61
34 01
35 52
81
35 62
84
71
84
71
84
71
84
81
84
71
00
83
09
04
04
71
84
00
83
00
00
01
04
01
71
34
35 22
Input Sample Vol, ft3
Input Baron
Input 5dir£>l
Input Dilut
Input Aba £
Input Stand
ug NH4
Input Stand
later, "Hg
e Temp.
°F
ion Factoi
oln Vol,
mi
Cone ,
VmJt
Area
Input Sample Area

Output Cone, M9 NHj/m-*


0 1 2
SO SI S2
A B
3
S3
C










060









070









oeo









090









too









10

























































































































REGISTERS
4
S4

567
SS S6 S7
D !
s a
SB S9
; r
       Figure 7  (Cont'd).   HP-67 program form.
                              374

-------
This equation  gives the concentration of NH4+  in  the  sample.  To convert
     NH
to     3 the ratio of the formula weights of NH3 to NH4+ is multiplied
                                                                        by
-sa*
                    yg
                              17.03(
                     yg NH3
                     y mole
                sa
                              18.04(
                                          4
                                                          yg NH.
                                                           mi
                                    y mole
     The next step involves the determination of the amount of NH3 collected
in the bubbler.   This is obtained by multiplying the volume of absorbant
(absorb, vol.)  and a dilution factor  (DF) by  the concentration of ammonia
collected.
                         ' yg NH.
              0.944 C
                     sa
                            mi
                                  x absorb, vol.(mi)  x DF = yg NH3
     To find the concentration of NH3  in exhaust  the volume of gas collected
 is corrected to the specif ied temperature and pressure  (68°F = 528°R and
 29.92 "Hg) .     The volume as read from the digital  readout in cubic feet is
 converted to cubic meters by dividing by 35.31 f t3/m3.

                  VOL, ft3 x B.P. "Hq    x  	528°R
                         3   29.92 "Hg      (TEMP,
                 35.31 £|—
                                             (TEMP, °F + 460)
                       m-
 where VOL = volume of gas collected,  ftj

     B.P. = collection pressure,  "Hg

     TEMP = collection temperature,  °F

 The concentration of NH, is  then  calculated by  dividing yg NH3 by  the volume
 of gas.
Cst
yg NH *
mi J
x PA
CZ fH
	 v nh^nrh vol . m'r X IYF
PAst
VOL, ft3 .. B.P. "Hq 528°R
35.31
                            29.92  "Hg
                                          (TEMP,  °F + 460)
                      rn"
0.944 x cst
\
yg NH4+
mi / X
x al^sorb. vol, w
i x DF
                              VOL,  ft  x PAst
                                       375

-------
                35.31 ft3 v 29.92   " Hg    tTEMP, °F + 460)
                      1?  _
                             B.P.,  "  Hg x 528°R


                                 yg NH3
                               =    -3
                                   m


     The concentration of ammonia can also be expressed in ppm NH^ by taking
into consideration the density of the gas at the desired  conditions  (68°F,
29.92 "Hg). .    The fifth edition of the Matheson Gas Data Book lists the
specific gravity of ammonia gas at 70°F and 1 atm pressure at 1.411 £/g.  The
inverse of the specific gravity gives a density of 0.709  y.  If the volume of
l£ of gas is corrected for temperature,


                                    528°P
                      V = X
                                 (70°F H- 460)

it can be divided into the weight of the gas to give the density at 68°F
                        0..996 £


                              = 0.00141
                                        yg NH3
                                                                  yg NH,
When the inverse of density is multiplied by the concentration in 	j--
the concentration is given in ppm NH3.

                    m              yg NH3   m

Sample Calculation

     The two examples will be calculated from information recorded on the
data sheet  (Figure 8) .  This information does not necessarily represent
actual experimental data but serves as a means of confirming calculations
done by hand with those done with the Hewlett-Packard Calculator.

Example 1

     Assume that in the FTP-2 driving cycle that 2.652 ft  of dilute exhaust
is collected in 25 m£ of 0.01 N H2S04 at 74.2°F and 29.42  "Hg.   when the
undiluted sample is injected into the Ion Chromatograph it yields a peak area
                     yg NH4+
of 3800 counts.  A 9 - T - standard similarly injected produces an area of
4800 counts.
                                     376

-------
   SwRI Project No.
   Fuel:
Test No.
                Test Date:
                                          Vehicle:
   Sample Collection By:
   General Comments:
                         CVS No.
      Tunnel Size
                                                              Driver:
Chemical Analysis By:
                                                   Miles:
Calculations By:

Driving
Cycle
FTP-1
FTP -2
FTP-3
SET-7
HFET
NYCC

BG

Sampling Conditions
Volume
Ft3
1.546
2.652
1.625
4.212
2.197
1.890

2.600
1
B.P.
"Hg
29.40
29.42
29.44
29.45
29.47
29.49

29.50
2
Temp.
oF
74.0
74.2
74.5
- 74.3
74.7
74.7

74.4
3
Dilution
Factor
1
1
1
1
1
1

1
4
Absorb.
Reagent
Volume m£
25
25
25
25
25
25

25
5
Standard
ygNH^/mfc
4
9
5
15
9
8

1.5
6
Area
3500
4800
3600
4500
5000
5200

3300
7
Sample
Area
2500
3800
2700
3500
3500
3500

2500
8
ygNH3/m3
1590
2300
1980
2370
2460
2440

374

ppm
2.23
3.25
2.79
3.35
3.46
3.44

0.53

CO
                                             Figure 8.  NH  Data Sheet.

-------
     The equation needed to calculate the concentration of ammonia follows:
          yg NH3 _
                   °'944
                         x PAca x absorb, vol,
                                                                 X
             3
            m
            DF
               VOL,  ft3 x PAst
x 35.31 ft3/m3 x 29.92   " Hg x (TEMP, °F 4- 460)

             B.P.,  "Hg x 528°R
        0.944 x 9 x 3800 x 25 x  l  x 35.31  x  29.92  x  (74.2 -I- 460)
                         2.652 x 4800 x 20.42 x 528
                                       Ug NH3
                                = 2300 	5—
                  ppm NH3 = 2300 x 0.00141 =3.25 ppm NH3
Example 2
     Assume that 1.890 ft3 of dilute automotive exhaust was collected in 25
m£ of 0.01 N I^SC^ during the NYCC driving cycle.  The sampling conditions
during this test were 74.7°F and 29.49 "Hg.    When the undiluted sample was
injected into the ion chromatograph the peak produced had an area of 3500
counts.  When an 8 ^g   ^  standard was injected it produced an area of 5200
counts .              m&

     The same equations in Example one are used to give concentrations of
NH3 of 2440        and 3.44 ppm NH3
              m3
EQUIPMENT
     The equipment section lists the equipment used in the ammonia procedure.
It is divided into four sections corresponding to each major division in the
procedure:  Sampling,  Analysis,  Water Filtration and Sample Preparation.
For convenience the item,  manufacturer, model number and any additional
pertinent information are  included.

Sampling

     1.  Glass impingers,  Ace Glass Products, Catalog #7530-11, plain
         tapered tip stoppers with 18/7 arm joints and 29/42 bottle joints.

     2.  Flowmeters, Brooks Instrument Division, Model 1555, R-2-15-C,
         sapphire ball,  0-5 lit/min range, graduated 0-15.


                                    378

-------
     3.   Dry gas meter,  American Singer Corporation, Type AL-120,  60  CFH
          C3.p3.cxty •

     4.   Digital readout for dry gas meter.

     5.   Sample pump,  -Thomas, Model #106 CA18, 4 lit/min free flow capacity.

     6.   Drying tube,  Analabs Inc., catalog #HGC-146,6" long, 1/4" brass
          fittings.

     7.   Teflon tubing,  United States Plastic Corporation, 1/4"  OD x  1/8"
          ID and 5/16"  OD x 3/16" ID.

     8.   Teflon solenoid valve, The Fluorocarbon Company, Model  IDV2-144N
          CAl.

     9.   Miscellaneous Teflon nuts, ferrules, unions, tees, clamps and con-
          nectors,  etc.

     10.   Miscellaneous electrical switches, lights, wiring, etc.

     11.   Regulating valve, Nupro 4MG, stainless steel.

     12.   Iron/Constantan type J, Thermo Sensors Corporation, single thermo-
          couple,  1/4"  OD stainless steel metal sheath.

     13.   Six  channel digital thermometer, Analog Devices, Model  #2036/J/1.
     14.   30 m£ polypropylene sample storage bottles, Nalgene Labware, Catalog
           #2006-0001.

     15.  Modified 25 mm A-H microanalysis filter holder, Millipore, Catalog
          #XX50 020 00.

     16.  Fluoropore 25  mm filters, Millipore, Catalog #FHLP 025 00, 0.5
          micron pore size.
                                                          fn\
     17.  Flexible heavy insulation  heating tape, BriskeaM  width-1/2",
          length-48".

     18.  Temperature Controller, Athena, 100-600°F.

     19.  Heated TFE Teflon hose, Technical Heaters Inc., 5' x 1/4", tem-
          perature limit 400°F.
Analysis

      1.  Dionex Model 10 Ion Chromatograph.
                                      379

-------
      2.   Multivoltage recorder, Texas Instruments, Model #PS02W6A.

      3.   Polyethylene cubitainers, Cole Farmer Instrument Company, Catalog
           #6100-20, 1 gallon.

Water Filtration

      1.   Filtration apparatus, Millipore, Model #XX 1504700.

      2.   Filters, Millipore, Model #GSWP04700, 0.22 micron pore size.

Sample Preparation

      1.   3 cc disposable syringes, Becton-Dickson, Model #5585.

      2.   15 mH disposable polypropylene cups, Cole-Farmer Instrument
           Company, Catalog #6006-10.

      3.   Class A, 1 m£ volumetric pipets.

      4.   Class A, 2 m& volumetric pipets.

      5.   Class A, 3 m£ volumetric pipets.

      6.   Class A, 4 m£ volumetric pipets.

      7.   Class A, 5 mJZ, volumetric pipets.

      8.   Class A, 10 m£ volumetric pipets.

      9-   Class A, 20 m£ volumetric pipets.

     10.   Class A, 25 m£ volumetric pipets.

     11.   Class A, 50 m£ volumetric pipets.

     12.   Class A, 100 m£ volumetric pipets.

     13.   Class A, 100 m£ volumetric flasks.

     14.   Class A, 1000 m£ volumetric flasks.

     15.   Class A, 2000 m& volumetric flasks.

     16.   Mohr pipet, 1 m£ graduated 1/10.

LIST OF REAGENTS

      The reagents used in the analysis of ammonia are presented in this sec-
tion.  In addition to the function of each reagent, the purity, manufacturer
and catalog number are also listed.


                                    380

-------
     1.   Water-deionized and filtered through a 0.22 micron filter.

     2 .   Standard
               Ammonia  sulfate,  (NH4)2S04, formula weight = 132  146  ACS
               analytical reagent grade, granular, j.T. Baker Chemical Co.
               TT\J /-7^ *

     3 .   Absorbant
               Sulfuric acid, H2SO4, formula weight = 98.08, ACS analytical
               reagent  grade, Mallinckrodt #2876.

     4.   Eluent
               Nitric Acid,  HN03, formula weight = 63.01, ACS analytical
               reagent  grade (Ultrex) , J.T. Baker, #1-4801.

     5 .   Regenerant
               Sodium hydroxide,  NaOH, formula weight = 40.00, ACS analytical
               reagent  grade, pellets, Mallinckrodt #7708.

PREPARATION OF REAGENTS

     The water used in  making all solutions and dilutions is prepared by
filtering deionized water through a 0.22 micron filter.  After filtration the
water is stored in polyethylene bottles.

Standard ( (NH4) 2804 )

     The stock solution is prepared by diluting 0. 3660 g of (NH4)2S04 to
1000 m£ with water.  This yields a solution with a concentration  of 100
yig NH^ .   Less concentrated  standards are made up by diluting portions of

the stock solution to 100 mJl  with water using volumetric glassware.

absorbing Solution  (0.01 N
     The absorbant  is  prepared by diluting 20.0 m& of the certified 1.000 N
sulfuric acid to 2000 mH with water.

Sluent  (0.0075 N HNOs) *

     A 1 N HNO3 stock  solution is prepared by diluting 62.5 m£ of concen-
trated  nitric acid to 1000 mH with water.  The eluent is prepared by further
diluting 15 m£ of  the stock solution to 2000 mH with water.

Regenerant  (0.5 N  NaOH)*

     40.00g of NaOH is dissolved in water in a 2 liter volumetric flask and
diluted to  volume.

*4 liters of each  of these solutions are prepared and stored in labeled
 polyethylene cubitainers .
                                     381

-------
 REFERENCES

 Braker,  W.  and Mossman, A.L.,  Ma the son Gas Data Book,  5th  Edition,  East
 Rutherford, N.J., 1971, pg. 17.

 Furman,  N.H.,  (Ed.), Scott's Standard Methods of Chemical Analysis,  5th
 Edition,  Van Nostrand, N.Y., 1958, Vol. 1, pg. 636.

 Pifer, C.W. and Wollish, E.G., Anal. Chem. Vol. 24, pg. 519,  1952.

 Kolthoff, I.M. and Stenger, V.A., Volumetric Analysis,  Interscience, N.Y.,
 1947, Vol.  2, pg. 125.

 Furman,  N.H.,  (Ed.), Scott's Standard Methods of Chemical Analysis,  5th
 Edition,  Van Nostrand, N.Y., 1958, Vol. 1, pg. 632.

 Pierce,  W.C., Haenisch, E.L., and Sawyer, D.T., Quantitative  Analysis,
 Wiley, N.Y., 1958, pg. 260.

 Kolthoff, I.M. and Stenger, V.A., Volumetric Analysis,  Interscience, N.Y.,
 1954, Vol.  2, pg. 168.

 Milner,  O.I. and Zahner, R.J., Anal. Chem., Vol. 32, pg. 294,  1960.

 Conway,  E.J., Micro-Diffusion Analysis and Volumetric Error,  2nd Edition,
 Crosby-Lockwood, London, 1947.

 Altman,  C.J. G., de Heer, B.H.J., and Hermans, M.E.A.,  Anal.  Chem., Vol. 35,
 pg.  596,  1963.

 Vogel, A.I., Quantitative Inorganic Analysis, 3rd Edition,  Longmans, Green,
 London,  1961, pg. 254.

 Blinn, R.C. and Gunther, F.A., Anal. Chem., Vol. 29, pg. 1882,  1957.

 Kolthofi, I.M. and Stenger, V.A., Ind. Eng. Chem., Anal. Ed.,  Vol. 1,
 pg.  79,  1935.

 Kolthoff, I.M. and Belcher, R., Volumetric Analysis, Interscience, N.Y.,1957,
 Vol. 3, pg. 582.

 Welcher,  F.J., Organic Analytical Reagents, Van Nostrand, N.Y., 1947, Vol. 1,
 pg.  376.

 Rowe, D.J.,  Gas Journal, Vol.  265, pg. 49, 1951.

Burns, E.A., The Analysis of Exhaust Gases Trapped from Ablating Nozzles,
 TRW Report No.  9840-6001-TU 000, Redondo Beach, Cal., 1964.

Nyman, C.J.  and Johnson, R.A.,  Anal. Chem., Vol. 29, pg. 483,  1957.
                                    382

-------
Norton,  O.R.  and Mann, C.K., Anal. Chem., Vol.  26, pg.  iiso, 1954

Clear, A.J.  and Roth, M., in Treatise on Analytical  Chemistry. Kolthoff  I M
andElving,  P.J. (Eds.), Interscience, N.Y.,  1961, Part Il,"vol. 5  pg  284

Kolthoff,  I.M., Stricks, W., and Morren, L.,  Analyst, Vol.  78, pg. 405, 1953.

Laitinen,  H.A. and Woerner, D.E., Anal. Chem., Vol. 27, pg. 215, 1955.

Arcand,  G.M.  and Swift, E.H., Anal. Chem., Vol. 28,  pg.  440, 1956.

Krivis,  A.F., Supp, G.R., and Gazda, E.S., Anal.  Chem.,Vol. 35, pg. 2216, 1963.

De Ford, D.D., Johns, C.J., and Pitts, J.N.,  Anal. Chem., vol.  23,pg. 941, 1951.

Segal N.S. and Wodley-Smith, R., Anal. Chem., Vol. 38,  pg.  829, 1966.

Sambucetti,  C.J., Anal. Chem., Vol. 38, pg. 105,  1966.

Barendrecht, E. and Janssen, N.G.L.M., Anal.  Chem.,  Vol. 33, pg. 199, 1961.

Burns, E.A., Uipublished results, 1963.

Jenkins Jr., R.W., Cheek, C.H., and Linnenbom,  V.S., Anal.  Chem., Vol. 38,
pg. 1257,  1966.

Miettinen, J.K., and Virtanen, A.I., Ann. Acad. Sci. Fennicae, ser. A. II.,
No. 41, 1951; through Chem.  Abst. Vol.  46, pg. 11500,  1952.

Gilbert, T.R. and Clay, A.M., Anal. Chem., Vol. 45,  pg.  1757,  1973.

Le Blanc,  P.J. and Sliwinski, J.F., Am.  Lab., Vol. 5, pg. 51,  1973.

Thomas, R.F. and Booth, R.L., Environ. Sci. Tech., Vol.  7,  pg. 523, 1973.

Woodis Jr.,  T.C. and Cummings Jr., J.M., JAOAC, Vol. 56, pg. 373, 1973.

Stockdale, D., Analyst, Vol. 84, pg. 667, 1959.

Sadek, F.S.  and Reilley, C.N., Anal. Chem., Vol.  31, pg. 494,  1959-

Crane, F.E.  and Smith, E.A., Chemist-Analyst, Vol. 49,  pg.  38, 1960.

Crane, F.E.  and Smith, E.A., Chemist-Analyst, Vol. 52,  pg.  105, 1963.

Vogel, A.I., Quantitative Inorgrnic Analysis, 3rd Edition,  Longmans, Green,
London, 1969, pg. 566.

Furman, N.H.,  (Ed.), Scott's Standard Methods of  Chemical Analysis, 5th
Edition, Van Nostrand, N.Y., 1958, Vol.  1, pg.  2336-
                                     383

-------
  Vogel, A.I., Quantitative Inorganic Analysis,  3rd Edition, Longmans,  Green,
  London, 1961, pg. 1092.


  Steyermark, A., Quantitative Organic Microanalysis, Blackstone, N.Y.,  1951,
  pg. 56.


  Taras, M.J., in Colorimetric Determination of Nonmetals, Boltz, D.F.,  (Ed.),
  Interscience, N.Y., 1958, pg. 84.


  Miller, G.L. and Miller, E.E., Anal. Chem. , Vol. 20, pg. 481, 1948.


  Thompson, J.F. and Morrison, G.R.,  Anal. Chem., Vol. 23, pg. 1153, 1951.


  Kruse, J.M. and Mellon, M.G., Anal. Chem., Vol. 25,  pg. 1188, 1953.


  Prochazkova, L.,  Anal. Chem. , Vol.  36,  pg. 865, 1964.


  Scheurer, P.G. and Smith, F., Anal. Chem., Vol. 27,  pg. 1616, 1955.


 Bolleter, W.T.,  Bushman, C.J., and  Tidwell,  P.W., Anal. Chem., Vol. 33,
 pg. 592, 1961.


  Zitomer, F. and Lambert, J.L., Anal.  Chem.,  Vol. 34, pg. 1738, 1962.


 Williams, D.D. and Miller, R.R.,  Anal.  Chem.,  Vol.  34,  pg.  225,  1962.


 Howell, J.H.  and Boltz, D.F., Anal. Chem., Vol.  36,  pg. 1799, 1964.


 Gunther, F.A., Barkley, J.H., Kolbezen,  M.J.,  Blinn, R.C.,  and Staggs,
 E.A.,  Anal. Chem., Vol. 28,  pg. 1985. 1956.


 Kolbezen M.J., Eckert,  j.w.,  and  Wilson,  C.W.,  Anal. Chem., Vol.  36, pg. 593,
 1964.


 Pierson,  R.H., Fletcher, A.A., and  St.  Clair Gantz,  E., Anal.  Chem. Vol. 28,
 pg.  1218, 1956.


 Vandeveene, L. and Oudewater, J., Centre Beige D'Etude  et de  Documentation
 des  Eaux, Vol. 352, pg.  127,  1973.


 Banwart, W.L., Tabatabai, M.A. and  Bremner, J.M., Comm.  Soil  Sci. Plant Anal.,
 Vol. 3, pg. 449, 1972.


 Byrne, E. and Power, T., Comm. Soil Sci. Plant Anal., Vol. 5, pg. 51, 1974.


Me William,  D.J.,  and Ough, C.S.,  Amer.  J. Enol. Viticult, Vol. 25,  pg.67, 1974.


Sigsby Jr.,  J.E.,  Black, F.M., Bellar, T.A., and Klosterman, D.L.,  Environ.
Sci. Tech.,  Vol.  7, pg. 51, 1973.
                                     384

-------
Christian,  G.D.,  Knoblock, E.G., Purdy, W.C., Anal.  Chem., Vol. 35, pg. 2217,
1963.

Zweidinger, R.B., Tejeda, S.B., Sigsby Jr.,  J.B.,  and Bradow, R.L., "The
Application of Ion Chormatography to  the  Analysis  of Ammonia and Alkyl Amines
in Automotive Exhaust," Symposium on  Ion  Chromatographic Analysis of Environ-
mental Pollutants, EPA, Research Triangle Park,  N.C., April 1977.
                                       385

-------
        APPENDIX I





ORGANIC SULFIDES PROCEDURE
           386

-------
              THE MEASUREMENT OF ORGANIC  SULFIDES  IN EXHAUST
     The  measurement of organic sulfides;   carbonyl sulfide  (COS), methyl
sulfide  (dimethylsulfide,  (O^^S),  ethyl  sulfide (diethylsulfide,  (C2H5)2S)
and methyl disulfide (dimethyldisulfide,  (CH3)2S2)  in exhaust  is accomplished
by passing the exhaust through Tenax GC traps  at -76 °C.   The organic sulfides
are removed from the exhaust by the  traps  at this temperature.  The exhaust
sample is collected continuously during the test cycle.   The organic sulfides
are thermally desorbed from the traps into a gas chromatograph sampling
system and injected into a gas chromatograph equipped with a flame photometric
detector  for analysis.  External organic sulfide standards generated from
permeation tubes are used to quantify the  results.   Detection  limits are on
the order of 0.1 ppb.
SAMPLING SYSTEM

     A Tenax GC trap is, used to collect  exhaust samples  for  the analysis of
the organic sulfides .  A flow schematic  of  the  sample  collection system is
shown in Figure 1'.  The trap collects  99+ percent of the sulfides at flows up
to 130 mil/min.   Several views of the sampling system are shown in Figure 2.
The various components of the sampling system and their  functions are listed
below.
Item       Component
        NaHCO  trap
        Tenax-GC trap
                Description
        Perma-Pure Drier
        Sample Pump
2" x 3/8" OD x 0.035" wall stainless steel
cartridge packed with 5 percent NaHC03 on
45/60 mesh Chromosorb P  (this trap removes
SO2 from the exhaust sample).

2" x 3/8" OD x 0.035" wall stainless steel
cartridge packed with preconditioned 60/80
mesh Tenax-GC  (this trap  collects and con-
centrates the organic sulfides).

Model PD-62512S Perma-Pure Drier  (this dryer
removes the moisture in the  exhaust without
jeopardizing the sample integrity) .

Model MB-158 Metal Bellows Vacuum/Compressor
Pump.  The sample pump pulls the exhaust sample
through  flip-top filter  and Perma-Pure Drier
and forces the sample under  pressure through
the remainder of the system.
                                     387

-------
CO
00
00
                   Flip-Top
                    Filter
   Perma Pure
     Drier
          Tunnel
Flowmeter
                                                   Regulating
                                                     Valve
                                                                              Sample
                                                                               Pump
                                                                                                  Tenax-GC
                                                                                                   Trap
Low Temperature
    Bath
                         Figure   1,   Organic  sulfide  sample  flow  schematic.

-------
Figure 2 .   Several views of the organic sulfide sampling system.
                             389

-------
Figure 2. (Cont'd).  Several views of the organic sulfide
                      sampling system.
                            390

-------
  5    Low temp bath      A. constant low temperature bath is obtained by
                          using a CO2-isopropyl alcohol slurry.  A bath
                          temperature of -76 to -78°C is obtained with
                          this bath.

  6    Flip-top filter   A 7.0 cm stainless steel flip-top filter is
                          included to remove all particulate from the
                          exhaust sample prior to its entry into the
                          Perma-Pure Drier.

  7    Regulating valve  A Nupro SS-4MG regulating needle valve is used
                          to control the exhaust flow through the NaHCO
                          and Tenax-GC traps.                          3

  8    Flowmeter          A Brooks Model 1550 flowmeter with R-2-15-AAA
 .                         ('SS float) is used to determine the exhaust
                          sampling rate.

TRAP  PREPARATION

     The  Tenax traps used for the collection of the organic sulfides are
prepared  by  filling a 2" x 3/8" OD  (0.028" wall) stainless steel tube with
approximately  1 gram of 60/80 mesh Tenax-GC.  Stainless steel fritted discs,
50y,  3/8" OD are placed at each end of the trap to hold the Tenax in the
trap  while allowing a gas flow through the trap.  Nut and ferrules, 3/8",
and Swagelok 3/8"  x 1/8" stainless steel reducing unions hold the fritted
discs in  place and allow the trap to be inserted into the sampling system.
A 1/8" stainless steel cap is placed on each end to prevent moisture and
other unwanted compounds from collecting in the trap.  Figure 3 shows a view
of the completed trap and of its components.  When a trap is ready to use,
1/8"  Swagelok  nuts and ferrules are used to connect 1/8" stainless steel
tubing to each end of the trap.  At the other end of each piece of the 1/8"
stainless steel tubing a miniature male quick connect is added with 1/8"
Swagelok  nuts  and  ferrules (Figure 4) .   The trap can now be connected by
the miniature  quick connects to the sampling system or the desorbing system
with  ease.

ORGANIC SULFIDE TENAX-GC TRAP CONDITIONING PROCEDURE

     The  analysis  of organic sulfides in dilute automotive exhaust requires^
Tenax-GC  traps that have been properly conditioned.  It is absolutely essen-
tial  that each Tenax-GC trap undergo the identical conditioning.  This will
insure that  there  are no residual compounds in the trap from a previous
sample, or in  the  case of a fresh trap, to remove any residual solvents.
Accurate  quantitative data is directly dependent on performing the condi-
tioning procedure  in a consistent manner according to the procedure outlined
in this section.
                                      391

-------
         Figure 3.   Tenax-GC trap


Figure 4.   Tenax-GC trap with quick connect.
                   392

-------
     The  Tenax-GC traps can be used repeatedly provided they are properly
conditioned.   Conditioning of the Tenax-GC  traps  is  accomplished by purqinq
with zero nitrogen at a selected temperature  for  a specified time period at
a given flow rate.  A system was developed  that was  capable of conditioninq
two Tenax-GC traps simultaneously.  A  flow  schematic of this system is
illustrated in Figure 5.   Several views of the Tenax-GC sample conditioning
system are presented in Figure 6.   This system has  been shown to reproducibly
condition Tenax-GC traps to a negligible level of organic sulfides.

     The organic sulfide trap conditioning  procedure is listed below in a
step-wise sequence.

     1.  Turn on the furnace and adjust the temperature to 325°C
         ± 25°C.  The furnace should be allowed to stabilize for
         at least 15 minutes before the first trap is  conditioned.
         The temperature readout on the furnace should be verified
         at least once a week with a digital  thermocouple.

      2.   Connect the traps according to the flow schematic.   Insert
          the traps  into  the furnace  (with no  nitrogen) and allow  the
          traps  to equilibriate  for 5 minutes.

      3.   Turn on the nitrogen and adjust the  flow to 500 mJl/min.

      4.   After  the  60-minute conditioning period, the trap is  removed
          from the furnace  and is allowed to cool to room temperature
          with nitrogen flowing through the trap.

      5.   After  the  trap returns to room temperature, turn off  the
          nitrogen and remove the trap.  The system is then ready  for
          conditioning the  next two traps.

 NOTE 1:   A master log should be maintained on the conditioning of each
          trap.   It  is the  .responsibility of the individual who is con-
          ditioning  the traps to keep a record of all traps that have
          been conditioned, who  conditioned them and the date that they
          were conditioned.  All traps should be permanently  identified
          to enable  keeping of these records.

 NOTE 2:   Tenax-GC traps should not be placed in the furnace  if the tem-
          perature is in excess  of 375°C.   This is the manufacturer's
          maximum recommended temperature and should not be exceeded.
          If this is allowed to happen, the chemical and physical  pro-
          perties of the Tenax-GC may be altered thereby affecting the
          trapping characteristics of the Tenax-GC.

 ANALYTICAL PROCEDURE

      The  analysis of the organic sulfides  (carbonyl sulfide, methyl sulfide,
 ehtyl sufide, and methyl disulfide)  in dilute exhaust is accomplished *Y
 collecting the  organic sulfides in Tenax traps at -76°C.   The  orgamc sulfides


                                      393

-------
                    Drier
Zero

 N2
                             Regulating
                               Valves
                        Tenax
                               Flowmeters
Traps
                L
                        Furnace  @  325+ 25°C
           Vent
       (500 ml/min)
                   Vent
                (500 ml/min)
    Figure  5.  Flow schematic for conditioning Tenax-GC traps
            for organic sulfide analysis (dual system).
                              394

-------
Figure  6.   Several views of Tenax-GC trap conditioning system
                             395

-------
Figure  6 (Cont'd).  Several views of Tenax-GC trap
                conditioning system .
                       396

-------
are thermally desorbed from the  trap into a gas chromatograph sampling system
The organic sulfides are analyzed by injecting the desorbed sample into a qas'
chromatograph equipped with a  flame photometric detector.  A standard blend
containing known amounts of the  four organic sulfides is injected into the qas
chromatograph to quantify the  results.   From the GC analysis of the sample
the analysis of the standard blend, and the measured volume of exhaust sampled
the concentration of the organic sulfides in the exhaust can be determined
The analysis flow schematic for  the organic sulfides is shown in Figure 7 ' A
detailed description of the procedure follows.

     The organic sulfides are  removed from the exhaust stream by trapping them
in Tenax GC traps.  A sample pump pulls dilute exhaust from CVS through a
flip-top filter, to remove particulate  from the exhaust sample, and then
through a Perma Pure Drier which selectively removes moisture from the exhaust
sample.  The moisture must be  removed to prevent ice from forming in the Tenax
traps.  The ice formation would  plug the trap and prevent the exhaust sample
from passing through the trap.   After exiting the Perma Pure Drier, the
exhaust sample passes through  a  sodium bicarbonate trap which removes any
interfering SC>2 in the exhaust sample.   The exhaust sample is then pulled
 through the Tenax  GC  trap which removes the organic sulfides from the exhaust.
 The trap  is held at -76°C with a dry ice-isopropyl alcohol slurry.  After the
 organic sulfides have been removed by the Tenax trap, the exhaust passes
 through the sample pump,  a regulating valve and a flow meter before exiting
 the sample system.  The  needle valve regulates the flow through the system
 which is  monitored by the flow meter.  A constant flow of 130 m£/min is main-
 tained throughout  the test.   The Tenax trap is disconnected from the sampling
 system at the  two  miniature quick-connects and the two male miniature quick-
 connects  are capped.   The trap remains in the -76°C dry ice-isopropyl slurry
 until it  is desorbed  by  the gas chromatograph injection system.

      The  Tenax trap is removed from the dry ice-isopropyl slurry, the liquid
 is quickly wiped  from the trap,  the the caps are removed from the ends of the
 trap. The trap is connected into  the gas injection system with the two quick
 connects, (Figure  8)  and the sample is immediately injected and placed into
 the Lindberg furnace  operating at  300°C  (Figure 9) .  The carrier gas upon
 injection flows through  the loop carrying the contents into the gas chromato-
 graph where the sulfides are separated and identified by their retention
 times. After  the  peaks  of interest have passed through the column in the gas
 chromatograph,  the system is  backflushed to remove any high molecular weight
 impurities that could interfere with later analysis  (Figure 10) .

      The  gas chromatograph system used  to  analyze  the  organic  sulfide  sample
 is shown  in Figure 11.  The system consists  of  a Perkin-Elmer  3920B  GC, and
 A/D converter  and a recorder.   The figure  also  shows  the  control  system, the
 Lindberg  furnace,  and the Bendix valve  oven.   The  GC  is  equipped with  a
 linearized flame  photometric  detector which has a  high sensitivity to  sulfur
 containing compounds.  The column consists of 6'  x 1/8"  Teflon tubing  packed
 with 60/80 mesh Tenax GC.  The  carrier gas is helium which flows  through the
 column at 30 m£'min.   The optimum hydrogen and air flows  are  40 m£/min and
 360 mVmin, respectively.  The column temperature,  after injection of the
 sample  is programmed from 30°C to 140°C at 8°  a minute.   In  a chromatogram
                                       397

-------
       CVS
      Tenax
     GC traps
          I
  Sample Analysis
in gas chromatograph
      with FPD
A/D Converter
                                              I
      Recorder
Hewlett-Packard
     3354
Computer System
  Figure  7.   Organic  sulfide  analysis  flow schematic.
                        398

-------
     Analytical Column
   (6* x 1/8" TFE,  60/80 Tenax-GC)
      • ••••••••••••••••a mmmm/tm m mm

      .	r*.   /
      Seiscor Valve
(backflush configuration)
Vent
                                                         Perkin-Elmer
                                                            3920B
                                                       Flame Photometric
                                                           Detector
                               Seiscor Valve
                         (Gas Sampling Configuration)
                                                               Carrier
                                                                Gas
                                                          Tenax Trap
                                                          300°C  furnace
                                            SS miniature
                                           quick connect
                                             Regulating
                                               Valve
                                                        Female
                                                       Quick-Connect
        Flowmeter
     Figure 8 .  Flow  schematic of organic sulfide analysis system
                  (Step 1  -  connect Tenax trap in CSV).
                                  399

-------
      Analytical Column
    (6*  x 1/8"  TPE,  60/80 Tenax-GC)
      Seiscor Valve
(backflush configuration)
                                                          Perkin-Elnter
                                                             3920B
                                                        Flame Photometric
                                                            Detector
                                Seiscor Valve
                           (Gas Sampling Configuration)
Vent
                   '—T
                   • •••»•••»• •«'•»% •§•'!•••• «1» * •» • *
                                                                 Carrier
                                                                  Gas
                                                            Tenax Trap
                                             __    I     :            :
                                             •^TjM-^M^J     •«••••»••<••»•»•••••
                                               \L          300°C furnace
     SS miniature
    quick connect
                                               Regulating
                                                 Valve
                                Pump
         Flowmeter
                                                          Female
                                                        Quick-Connect
       Figure  9.   Flow schematic of organic sulfide analysis system
            (Step  2 - inject Tenax trap contents into GC system).
                                   400

-------
           Analytical Column
     (61 x 1/8" TFE,  60/80 Tenax-GC)
       mmmmmmmmmmmm mmmmm mmm/m mmmmmmmmmmmmmmm

       	-P  I      A
      Seiscor Valve
(backflush configuration)
Vent
                                                         Perkin-Elmer
                                                            3920B
                                                       Flame Photometric
                                                           Detector
                               Seiscor Valve
                       (gas sampling configuration)

                         /
                    Carrier
                     Gas

                  Tenax Trap
              *»•» mmmmw^mmmmmmmm |
                    *      •
                                                        300°C furnace
                                          SS miniature
                                         quick connect
                                             Regulating
                                                Valve
         Flowmeter
                                                        Female
                                                     Quick-Connect
        Figure 10-   Flow schematic of organic sulfide analysis system
                     (Step 3 - backflush analytical column).
                                    401

-------
<
' • I
                                 Figure 11.  Organic sulfide analytical system.

-------
of a standard sample  (Figure  12)  containing the four sulfides, the first
peak eluted is carbonyl sulfide,  followed by methyl sulfide, ethyl sulfide and
methyl disulfide.  The GC sulfide peaks are recorded on a Soltec dual channel
recorder (1 mv) and peak areas  and retention times are obtained from the
Hewlett Packard GC computer system. A Metronix Dynacalibrator operating at
40°C and containing permeation  tubes of carbonyl sulfide, methyl sulfide,
ethyl sulfide, and methyl disulfide is used to supply standard concentrations
of the organic sulfides.  The permeation rate of each tube is monitored by
monthly weighings.  Zero nitrogen is used to dilute the permeation gases to
the concentrations desired.   The  10 m& sample loop is purged with this per-
meation gas for 10 minutes  (Figure 13) and the 10 m£ of permeation gas is then
injected into the 'GC  and analyzed (Figure 14) .  From the standard peak areas,
the exhaust sample peak areas and the volume of exhaust sampled, the concen-
tration of the organic sulfides in exhaust can be determined.

CONTROL SYSTEM

     A control system was developed to systematically control the flow of
the two Seiscor valves.  This control is accomplished by ATC electric timers
and ASCO electric solenoid valves.  This system employs one Seiscor valve in
a gas sampling valve  (GSV) configuration with the second Seiscor valve in a
backflush configuration.  These valves are pneumatically operated and
electrically controlled.  The electrical schematic for the control of the
Seiscor valves using  the ATC  timers and ASCO electric solenoid valves is
shown in Figure 15. The flow  schematic for vacuum and pressure lines to the
Seiscor valve are presented in  Figures 8-10 and 13-14.   The Seiscor valves
have been found to operate much more dependably if a vacuum assist is included
in the valve actuation controls.

CALCULATIONS

     This procedure has been  developed to provide the user with the concen-
trations of the organic sulfides  (carbonyl sulfide, methyl sulfide, ethyl
sulfide, and methyl disulfide)  in exhaust.  The results will be expressed in
yg/m3 of exhaust and  ppm for  each of the sulfides.  The equations for
determining the concentrations  of  yg/m3 and ppm are derived in the following
manner.

     The first step is to  find  the volume of exhaust sampled from the flow
rate and the sampling time  by the equation:

            Vol exp  (m£) =  F.R..,.  (mVmin) x Ti  (sec)/60 sec/min

            Vol exp  (m£) =  volume of gas sample in m£

          F.R.j  (mVmin) =  flow rate of exhaust sample in itd/min

                Ti  (sec) =  sampling time in minutes

               60 sec/min =  conversion of sample time in seconds to minutes
                                                                [Equation 13


                                      403

-------
Sample Permeation Blend	DaiJtov.  29. 1977
in.irumant PE  392OB	Operator P. Saunders
Column 6    ft. 1/8   61T
  Packed with  NA   % wt.
  on 60/80   meih  TENAX-GC
  Run ISO ® NA   °C mine 30	cc/min.
                                        Liq. Phase
                                         Support
                                          Carrier
  e 75   otig  MA    Rotameter Reading
  held ®       °C I SO far	min., prog to
         /min. Held for
  held for
Inlet	
Detector  160
  Hvd  40  prig
  Air   70  psig
  (  I   NA  Olio
                      Rotameter Rdg.
                      Rotameter
                      Rotameter Rdg.
       	                      -   NA    cc/min
Recorder  1  in/min weed    1  mV.F.S.Soltec    Type
Injection 10  ul indicated   10    ul net   A°    ul Actual
  Sampling Device  Gas Sampling Valve
        ±=S::-4-v---i
                             13   12    11   10    9    8

                             Retention Time, minutes
        Figure 12.    Chromatogram of  organic  sulfide  standard.
                                           404

-------
      Analytical  Column
(61 x 1/8" TFE,  60/80 Tenax GC)
       Seiscor Valve
(backflush configuration)
                                   Seiscor Valve
                             (Gas Sampling Configuration)
                                                              PE 3920B
                                                           Gas Chromatograph
                                                              with FPD
                                                                  Carrier  Gas
                                                      SS miniature
                                                     quick connect
                                                         10 ml sample  loop
  Vent
                                                  Regulating
                                                    Valve
                                     Pump
                                                              |—|  Calibration
                                                                    gas  in
    Flowmeter
                                                            Female
                                                         quick-connect
       Figure  13.   Flow schematic of  organic  sulfide  calibration  system
                     (Step 1 - purge of  sample loop of CSV).
                                      405

-------
      Analytical  Column
(6'  x 1/8"  TFE,  60/80 Tenax GC)
       Seiscor Valve
(backflush configuration)
                                                              PE 3920B
                                                           Gas Chromatograph
                                                              with FPD
                                  Seiscor Valve
                             (Gas Sampling Configuration)
                                                                  Carrier Gas
  Vent
                      '—T
                      •••••••••••••••••••••••••••^
                                                      SS miniature
                                                           connect
                                                        X
                                                         10 ml sample loop
                                                  Regulating
                                                    Valve
                                     Pump
                                                              |—\ Calibration
                                                                    gas  in
    Flowmeter
                     Female
                  quick-connect
       Figure  14.   Flow schematic of organic sulfide calibration system
               (Step  2  - inject calibration gas into GC system).
                                     406

-------
'  2
  3
  4
  9
 11
,12
"I14
Jl5



^•N

r-~T
...... _|
1 ">
I
\ A ~ 	 *
\ *± 	
1 ,
A -*
1 „
^

V,


p 	






n^v
solenoid
|


8l "3
a sr6
,n
                                                                                  Aux
                     -111
                      I
                      [12
                     -*14
                      I
                      115
                      "U6
                      i	
                                              backflush
                                               solenoid
          ATC
         Timer
 ATC
Timer
Figure  15.   Electrical schematic for organic sulfide analysis system .

-------
     The next step is to correct the volume of exhaust sampled  to  a  standard
temperature, 68°F, and pressure, 29.92" Hg, by use of the equation

                         p    xy      P     x V
                          exp    exp _  corr    corr
                            T              T
                             exp            corr

          V     = experimental volume of gas sample in m£
          vcorr = volume °f 9as sample i-n m^ corrected to 68°F
                    and 29.92" Hg
          pexp  = experimental barometric pressure
                =29 92" Ha
                  «-^  «y
          Texp  = experimental temperature in °F + 460
          Tcorr = 68°F + 460 = 528°R
    Solving for Vcorr gives:

                  P    ("Hg) xv    Cft3) x 528°R
          V        exp	exp	
           corr =	
                     Texp (°R) x 29.92" Hg
                                                               [Equation 2]

    Substituting Vol exp (md) from Equation 1 into Equation 2 gives:

                  P	("Hg) x F.R.T (rnVmin) x Ti (sec) x 528°R
      corr
                       Tovr, (°R)  x 29.92" Hg x 60 sec/min
                        exp

     The next step converts the volume from m£ to cubic meters by use of the
conversion factor;  1 cubic meter = 10  m&.

   V     (m3)   P    ("Hg) X F-R-T (mVmin) x Ti (sec) x 528°R
    corr      =  exP	i	•	
                Texp (°R)  x 29.92" Hg x 60 sec/min x 1Q6 md/m3
                                                              [Equation 3]

     The next step is to find the yg of each of the sulfides in the Tenax trap.
Since the FPD has a linear response in the region of concern, then the fol-
lowing equation holds:
                                A       A  . .
                                 sam     std

                = yg sample in Tenax trap
           A    = GC peak area of sample in relative  units
          Ugstd = yg standard
                = GC peak area of standard in relative  units

     Solving for yg sample gives :
                  yg .  , x A
          ug        std
            sam       A
                       std                                   (Equation 4)
                                      408

-------
     The  yg of standard for each of  the  organic sulfides  is  determined from
the permeation rates of the permeation tubes containing each of the sulfides,
the flow  rate of the diluting gas, and the volume of gas  sampled  (10 m£).

           ua std =     P.R.  (ng/min)  x  10
                    F*R'II  (m^/min>  x  10°°
                                                               [Equation 5]

     P.R. (ng/min) = permeation rate of  permeation tube at 40° in ng/min
             10 m£ = volume of calibration gas injected
    F.R.jj (m£/min)= flow rate of  diluting gas for permeation  system
        1000 ng/yg = converts ng to  yg,  one yg equals 1000 ng

      Substituting yg std from Equation  5 into Equation 4 gives:
                         P.R.  (ng/min)  x  10 m£ x A
                   _  	sam	
            yg sam - F.R     (m£/min) x 1000 ng/yg x Agtd

                                                               [Equation 6]

     To obtain yg sample/m3,  Equation  6  is divided by Equation 3 to give:
                    P.R.  (ng/min) x 10  m£ x ASS^ x T    (°R)
         yg sam/m3 =	—	
                    F.R.I:E (mVmin x 1000 ng/yg x Astd x Pexp ("Hg)

                       29.92" Hg x 60 sec/min x 1Q6 mjj,/m3
                       F.R.  (m£/min) x Ti (sec) x 528°R
                                                               [Equation 7]

      To  find the  concentration of each sulfide in ppm, the densities of the
 sulfides are needed.   At 29.92" Hg and 32°F, one mole of gas occupies 22.4
 liters.  This  volume  is corrected to 68°F from the equation

                                     V _ Vi_
                                     T " Tj

                       VT = 22.4
                       TI = 32°F + 460 = 492°R
                        V = volume at 68 °F
                        T = 68°F + 460 = 528°R

      Solving  for V gives:
                            V^  22.4^528
                        v -   T          492

      Since one mole of gas occupies 22.04* at  68°F, the density can be  found
 in g/S, by  dividing the molecular weight in g/mole by 24.04

                            mol. wt. g/mole
                den
                                       409

-------
     The density in yg/m& can be found by converting g to yg and  H  to m£
as follows:
                           mol.  wt. g/mole   1 x lp6yg/g _ mol. wt. x IQQQ
               den yg/mJt =  24.04 Vmole     1 x lo3mA/£        24.04

                                                               [Equation 8]

     To obtain the concentration of each sulfide in ppm, the concentration in
yg/m3 is divided by the density in yg/m&
                              . 3 .      . „   mJl
                     ppm = yg/m  r  yg/m£ = ^3-

     Using Equations 7 and 8 gives the ppm concentrations in the  form of the
raw data.
                P.R. (ng/mJl) x 10 mH x A    x T     (°R) x 29.92"  Hg
                                        sam    exp
          ppm =
                 F.R.   (mVmin x IQOO ng/yg x Astd x Pexp  ("Hg)
                    .I]:
                         60 sec/min x 1Q  m&/m
                 F..R.  (mVmin)xTi (sec) x 528°R
                 _ 24.04 Vmole _
                 mol. wt. (g/mole) x 1000 yg-Jl/g-m&
                                                               [Equation 9]

At this point, the concentration can be expressed in yg/m   (Equation 7) and
ppm (Equation 9)  at 68°F and 29.92" Hg from the raw data.

Hewlett-Packard Calculations

     In order to insure maximum turnaround in a minimum time period, a
Hewlett-Packard 67 program was developed to calculate the organic sulfide
concentrations in yg/m3 and ppm from the raw data.  This program is presented
in Figure 16.

Sample Calculation

     Assume exhaust samples were collected in Tenax traps for  each portion of
a three-bag 1975 FTP.  Raw data for these tests are presented  in Figure 17.
Calculations were performed using the HP-67 program and manual calculations.

Manual calculations for driving cycle FTP-1
         3       P.R. (ng/m£) x 10 m£ x A    x T     (°R) x  29.92" Hg

     yg/m  COS -   F.R.I3. (mVmin) x 1000 ng/yg x Astd x Pex f  ("Hg)

                 x _ 60 sec/min x IQ  mJl/m _
                   F.R.  (m£/min) x Ti (sec) x 528°R

                 667.5 ng/m£   10 m& x 3Q20Q x 535°R x 29.92"  Hg
                   580 mVmin x 1000 ng/yg x 18514 x 29.80" Hg

                 x  60 sec/min x ip6
                   130 m£/min x 504 x 528°R

                                     410

-------
              iiisiniciioiis
           ukGAtllC SULFIDES IN EXHAUST
STEP
Ul
Q jj
0 i
1
i
a
4
b
i
7
8
9
)0
11
12
13
14
1 ~~>
It,
17
Ib
1'J
1 20
21
22
23
24
25.






INSTRUCTIONS
SwiJ.cH UQ >iui switcJi to. run
?'^§d card in from right to left
Set dec ima.i Piai'fi
ln£y£ Samjjle Flow B^Le
Ijjjjut Sampling Time
lnp.ut Barometric Pressyi:^
lutut Salable Teniifijaliirs.
Input OUutian Qas flow fjjr tarmeaiiun staiiaaias
Int^ut Permeation Rate COS
Input Standard Area COS
Input Srfrgpl tj Area COS
Output CojiceutxdLiiiix .COS
Output Concentration COS
Input Permeation Rate (CH^^S
ItlPUt Sta.Hd^rd Aras (QLljlaa
Input Sample Area tCHj)2^
UuLput LoucBUtratiou (Ca3)aS
QUUtUL CilULMl!.rat_LQU ICtijJjS
Input Permeation Kate (C,Hc)-,S
r _ - . .^- . , — - ,~_.-~ - X 3-4#-
?!ieut Standard. Area (C^Js^S
InjJUL Sdfitple Art;a tC^H^) jS
Output Cuiicentration (€2^1^)2^
Output Concentration {^2^5^ 2s
40BHt Permeation Rate (CH^J^S^
ioaut. .Standaxd. Arfia.. iCHj) 232
^OE^t Sample Area t^i3i2§2
Output Concentration .tCH3]2S2
Oytp^t Concentration (CJi ) -S
3 2^2
"I" ""."_""""'.". "" - -.
	 	 __ .. _ — 	
. - 	 	 -
;'_;_ ;__" 	 ;""
. . . —
INPUT
DATA UNITS



mi/miti
Sue
"Hu

ffli/BJ n .
. ..na/oin




nQ/min




ntj/min




nq/min


	


	
i:

	 	

KEVS
1
1
Ig sci
IA
! R/S
IR/S
IR/S
R/S
R/S
R/S
R/S
1 R/S

IR/S
IR/S
IR/S
IR/S ||
1 II
iK/s'll
IR/S
IR/S II
IB/S
1
IR/S
IR/S
IR/S 11
IR/S
t
1
|h RTN
1
1 1
1 1
[ 1
1
1 1
1
1 1
1 i
1 1
1
1
OUTPUT
OtH UNITS











M/rc3
Pl'Bl



•mj
Wffi



U'J/n.'3
ppm



H9/.«3
PiJJD






Figure 16.  HP-67 user instructions.
                 411

-------
STEP  KEY ENTRY   KEY CODE
                        COMMtNTS
                                    STEP  KEY ENTRY  KEY CODE
                                                            COMMENTS
001








010









020









030









WO









050







0
SO
A
f I.BL A
3
4
0
0


R/S
X
R/S
X
STO 1
R/S
4

0
+
RCL 1
h x 4y

h 1/x
R/S
;
STO 2
B«
prj 7
X
R/S

R/S
X
R/S
2
4
9
9
.-.
R/S
RCL 2
X
R/S
!
R/S
X
R/S
2
5
8
4
T
R/S
R[~T. 03
X
R/S
'
R/S

1
Si

31 25 11
m_
04
00
00
00
81
84
71
84
71
33 01
84
04
06
00
61
34 01
35 52
81
35 62
84
81
33 02
84
14 02
71
84
81
84
71
84
02
04
09
09
81
84
14 02
71
84
ai
84
71
R.I
n?
nq
08
04
81
84
34 02
71
84
81
84

2
S2
3
In svui'lu Flow.mlt/min






In Sampling Time, se*

In barometer, "Hg


In Sample Temp, °f








InDil Gas flow m £/ni.n


In Perm RdLe COS , ng/nii


III Std Area COS

In Sample Area COS

3ut pg/m COS





3ut ppm COS
In PR (Cn3)2S 119/mil


In Std Area (CH,)2S

In Samp Area (CHj) 2S

Ou>t 119/81 (CH>)2S




Jut ppm (CH^^S
In PR (C-jHcJ^S, nq/nu


In Std Area (C2H5)2S

tn Samu Area (CoH^l-iS
REGIE
3 4
S3 *
C



"°l)









070









080
1








090









100






11


110


>TERS
5
SS
D
	 X 	
......K/S. .,
3
7
, .. .5 	
2
i
R/-S
Ijf'T. ? 	
	 * 	

i
R/S
X
R/S
^
4
1
8

R/S
h RTN



































6
S6

71
H4
LU
._ . U2. - ..
_U6 	
02
ai.
H4
14 112

ftil 	
ai
Kd
71
84
03.
09
01
08
HI
an
IS 22















•



















7
S7
i

>ut HU/BI* (("• H.)-S
2 b 2




Hut t>£'IH [C ,Hs) ,h
J ^ '!' ''"'"'"

In Std Area (0)3) ^s->
"
In Std Area (CH3)2S2

Out p9/m3 (CHj)2S2





!)ut ppm (CH3)2S2




































8 9
sa S9
I
       Figure  16  (cont'd).   HP-67 program form
                            412

-------
SWRI PROJECT NO.	



FUEL:       CVS NO.
SAMPLE COLLECTION BY:



GENERAL COMMENTS:
TEST NO.
                        TUNNEL SIZE:
                JTEST DATE:



                   DRIVER:
VEHICLE:
                                                           MILES:
   CHEMICAL ANALYSIS BY:
                              CALCULATIONS BY:
Test No.
Driving Cycle
Sample Flow Rate, m£/min
Sampling Time, sec
B.P., "Hg
Temp., °F
Dilution Gas Flow, mS/min
Permeation Rate COS, ng/min
Standard Area COS
Sample Area COS
Sample Cone. COS, yg/m3
Sample Cone. COS, ppm
Permeation Rate (CH3) 2S ng/min
Standard Area (CK^) 2s
Sample Area (013)23
Sample Cone . (CH3) 25 , yg/m3
Sample Cone. (013)2$, ppm
Permeation Rate (C7He;)2S, ng/min
Standard Area (C2H$)2S
Sample Area (CjHs^S
Sample Cone. (€2X5) 2s' yg/m3
Sample Cone. (C2Hs)2S, ppm
Permeation Rate (013)952, ng/min
Standard Area (013)232
Sample Area (CH3)2S2 	


1 2 3456
FTP-1
130
504
29.80
75
580
667.5
18,514
30,200
17.5
0.00700
1061
41,006
50,716
21.1
0.00816
445
6971
31,649
32.5
0.00865
133.5
2315
2011
1.86
0.000475
FTP-2
110
867
30.03
80
500
667.5
20,112
40,100
17.1
0.00683
1061
38,100
40,381
14.4
0.00558
445
7017
34,650
28.2
0.00751
133.5
2210
3120
2.42
0 00061"
FTP- 3
150
505
29.02
96
600
667.5
21,238
33,162
14.9
0.00598
1061
35,100
49,162
^B^|HHMflBfla^Ba^^^M^HA
21.3
0.00824
445
6844
32,111
29.9
0.00798
133.5
2763
6372
4.41
0.00113
SET-7
90
1397
29.25
85
650
667.5
16,542
20,122
6.29
0.00252
1061
41,000
38,142
7.65
0.00296
445
6015
7022
4.03
0.00107
133.5
1651
1561
0.978
0.00250
HFET
120
765
29.95
83
450
667.5
15,962
16,269
10.2
0.00406
1061
45,610
54,753
19.0
0.00736
445
7113
7914
7.39
0.00197
133.5
1814
1418
1.56
0.000391
NYCC
170
1200
29.50
89
575
667.5
23,146
32,641
5.08
0.00203
1061
35,122
41,611
^W^^^^^HVB^^^^
6.78
0.00262
445
7099
17,416
5.89

133.5
2917
2372
0.586
0.000149
        Figure  17.   Organic sulfide  sample  collection  sheet,
                                      413

-------
               =17.5 yg/m

       ppm COS = yg/m  * density yg/m£

                 mol. wt. (COS)  x loop
  density yg/m = 	24.04£

  mol. wt. COS = 60.08 g/mole
                 60.08 g/mole x 1QQQ
       density = 	24^04 ft/mole   ' = 24" Ug/m
                          ^                         —3     3          —3
           ppm =17.5 yg/m  * 2499 yg/m£ = 7.00 x 10   m*/m  =7.00X10  ppm

The calculations for methyl sulfide, ethyl sulfide, and methyl disulfide
are carried out in the same manner by substituting in the appropriate
permeation rates, standard areas, sample areas, and molecular weights
into the above formulas.  These calculations give the following concentra-
tions: (CH3)2S, 21.1 yg/m3 and 0.00816 ppm;  (C2H5)2S, 32.5 Ug/m3and 0.00865
ppm; and  (CH3)2S2, 1.86 yg/m3 and 0.000475 ppm.


NOTE:  The values used in these calculations are picked from a range of tem-
peratures, pressures, etc. to validate the calculations and may not be
representative of expected raw data.  These calculations are presented to
confirm that manual and HP-67 calculations give the same results.  This was
confirmed for six sets of calculations.

LIST OF EQUIPMENT

     The  analysis for the organic sulfides is performed using a gas chromato-
graph equipped with a flame photometric detector.  The gas chromatograph,
control console, sample collection, trap conditioning and trap preparation
are the basic functions in the analysis.  The major equipment required for
each function is listed below.

Gas chromatograph and control console

     1.  Perkin-Elmer Model 3920B gas chromatograph equipped with a linearized
         flame photometric detector (FPD) and subambient temperature pro-
         grammer .

     2.  Soltec dual channel recorder, Model B-281, 1 mv recorder.

     3.  Hewlett-Packard Model 3354 GC computer system with remote teletype
         printout.

     4.  Hewlett-Packard Model 1865A A/D converter.

     5.  Metronix Dynacalibrator Model 220-R for generation of organic
         sulfide standards.
                                     414

-------
    6.  Bendix valve oven.

    7.  Lindberg Furnace/Heavy Duty Model  55035.

    8.  Seiscor valve - gas sampling  configuration

    9.  Seiscor valve - backflush  configuration.

   10.  ATC timers. Model 325A346A10PX  (2  ea.).

   11.  Analytical column, 6' x  1/8"  Teflon,  60/80 Tenax-GC.

   12.  ASCO solenoid valve, Model 834501  (2  ea.).

   13.  Brooks flowmeter, R-2-15-A  w/ss  float, 0-150 scale.

   14.  Metal Bellows MB-158 pump.

   15.  Female quick-connect, stainless steel.

   16.  Nupro Model 2M stainless steel  regulating valve.

   17.  Miscellaneous stainless  steel,  copper and Teflon tubing (1/8"
        and 1/6").

   18.  Miscellaneous stainless  steel and  brass unions, tees, etc.

   19.  Bud Classic II control console  cabinet.

   20.  Miscellaneous electrical on-off switches.

Sampling

    1.  Perma Pure drier, Model  PD 625  12S (17").

    2.  Brooks flowmeter, R-2-15-AAA, SS float, 0-150 scale.

    3.  Metal Bellows MB-158 pump.

    4.  Tenax-GC trap.

    5.  Sodium bicarbonate trap.

    6.  Miscellaneous stainless  steel and  Teflon tubing (1/16", 1/8"
        and 1/4").

    7.  Miniature stainless steel  Swagelok female quick-connects.

    8.  Miniature stainless steel  Swagelok male quick-connects.

    9.  Miscellaneous stainless  steel and  brass unions, tees, etc.


                                   415

-------
    10.  Stainless steel 7.0 cm flip-top filter.



    11.  Reeve Angel Type AH 7.0 cm fiber glass filters.



    12.  Nupro Model 4M stainless steel regulating valve.



    13.  Dewar flask, 1 quart capacity.



Trap Preparation and Conditioning System



     1.  Lindberg Furnace/Heavy Duty Model 55035.



     2.  Brooks flowmeter, R-2-15-A, glass float, 0-150 scale.



     3.  Nupro  Model 4M brass regulating valve.



     4.  Swagelok 3/8" stainless,steel reducing unions  (2 per trap).



     5.  Stainless steel fritted discs, 50y,  3/8" OD  (2 per trap).



     6.  Stainless steel tubing, 2" x 3/8" OD (0.028" wall) (1 per trap).



     7.  Tenax-GC, 60/80 mesh (about 1 gram per trap).



     8.  Miscellaneous stainless steel and Teflon tubing (1/8" and 1/4").



     9.  Miscellaneous stainless steel and brass unions, tees, etc.



    10.  Asbestos gloves, pair.



    11.  Refillable plexiglass, gas drier, 6".



    12.  Nupro in-line filter, brass, Model 4.



Expendables



     1.  Zero hydrogen gas (GC).



     2.  Zero air (GC).



     3.  Zero helium carrier gas (GC).



     4.  Nitrogen (Perma Pure drier).



     5.  Isopropyl alcohol,  CH CHOHCH .
                              J       j



     6.  Sodium bicarbonate.



     7.  Dry  ice.



     8.  Zero nitrogen (permeation system) .




                                    416

-------
REFERENCES


Spencer, C. P., Baumann,  P.,  and Johnson, J. P., Anal. chein.,  Vol.  30,  No.  9
pg. 1473, 1958.                                                              '

Sumner, S., Karrman, K. J.,  and Sunden, V., Microchim Acta,  pg.  1144, 1955.

Liberti, A. and Cartoni,  G.  P., Chim. e ind., Vol. 39, pg.  821,  1957.

Ryce, S. A. and Bryce, W. A.,  Anal.  Chem.,  Vol. 29, pg. 925, 1957.

Desty, D. H. and Whyman,  B.  H. P.,  Anal. Chem., Vol. 29,  pg. 230, 1957.

Coleman, H. J., Thompson, C.  C. and Rail, H. T., Anal. Chem.,  Vol.  30, pg.
1592, 1958.

Desty, D. H. and Harbourn, C.  L. A., Div. of Analytical and  Petroleum Chem-
istry, Symposium on Advances in Gas Chromatography, 132nd meeting,  ACS, N.Y,
N. Y., September 1957.

Amberg, C. H., Can. J. Chem.,  Vol.  36, pg.  590, 1958.

Sullivan, J. H., Walsh, J. T., and Meritt Jr., C., Anal.  Chem.,  Vol. 31,
pg.  1827, 1959.

Dal  Nogare, S. and Bennett,  C. E.,  Anal. Chem., Vol. 30,  pg. 1157,  1958.

Harrison, G. P., Knight,  P.,  Kelly,  R. P.,  and Heath,  M.  T., Second Int'l.
Symposium on Gas Chromatography, Amsterdam, May 1958.

Adams, D. F. and Koppe, R. K., Tappi, Vol.  42, pg. 601, 1959.

Schols, J. A., Anal. Chem. Vol. 33, pg. 359, 1961.

Hall, H. L., Anal. Chem., Vol. 34,  pg. 61,  1962.

Adams, D. F., Koppe, R. K.,  and Jungroth, D. M., Tappi, Vol. 43, pg. 602,
1960.

Feldstein, M., Balestrieri,  S., and Levaggi, D. A., J. Air  Poll. Cont. Assn.,
Vol. 15, pg. 215, 1965.

Fredericks, E. M. and  Harlow,  G. A., Anal.   Chem. Vol. 36,  pg. 263, 1964.

Mizany, A. I., J. Chromatog. Sci • ,,  Vol. 8, pg. 151, 1970

Rummer, W. A., Pitts Jr., J. N., and Steer, R.  P., Environ.  Sci. Tech., Vol.5,
pg.  1045, 1971.

Rasmussen, R. A., Am.  Lab, Vol. 4,  pg. 55, 1972.
                                      417

-------
Brody, S. S. and Chaney, J. E., J. Gas Chromatog., Vol. 4, pg. 42,  1966.

Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I., Anal. Chem.,
Vol. 44, pg. 2070, 1972.

Elliot, L. F. and Travis, T. A., Soil Sci. Soc. Amer. Proc., Vol. 37, pg. 700,
1973.

Francis, A. J., Adamson, J., Duxbury, J. M., and Alexander, M., Bull. Ecol.
Res. Comm.,  (Stockholm), Vol. 17, pg. 485, 1973.

Lewis, J. A. and Papavizas, G. C., Soil Biol. Biochem., Vol. 2, pg. 239, 1970.

Lovelock, J. E., Maggs, R. J., and Rasmussen, R. A., Nature, London, Vol. 237,
pg. 452, 1972.

Ronkainen, P., Denslow, J., and Leppanen, O., J. Chromatog. Sci., Vol. 11,
pg. 384, 1973.
                                     418

-------
   APPENDIX J





PHENOLS PROCEDURE
       419

-------
                              PHENOL PROCEDURE
     Phenols  (phenol; salicylaldehyde; m-cresol/p-cresol; p-ethylphenol/
 2-isopropylphenol/ 2,3-xylenol/3,5-xylenol/2,4,6-trimethylphenol;  2,3,5-
 triroethylphenol; and 2,3,5,6,-tetramethylphenol) in automotive exhaust can
 be sampled and quantitatively analyzed with a gas chromatograph  (GC) equipped
 with a flame  ionization detector.  Dilute exhaust is passed through two
 Greenburg-Smith impingers in series, each containing 200 m£ of 1 N KOH
 chilled in an ice bath.  The contents of each impinger are acidified and
 extracted with ethyl ether.  The samples are partially concentrated, com-
 bined and then further concentrated to about 1 m£.  An internal standard
 is added and  the volume is adjusted to 2 m£.  The final sample is  analyzed
 by the use of the GC and concentrations of individual phenols are  determined
 by comparison to external and internal standards.  The minimum detection
 limit is about 1 yg/m&.

 SAMPLING SYSTEM

     A schematic of the phenols sampling system is shown in Figure 1.  As
 seen in the illustration, exhaust from the automobile is first diluted by
 the constant  volume sample (CVS).  The dilute exhaust entering the sampling
 probe is then filtered by a heated (375°F) Pallflex filter of porosity
 1-100 ym to remove particulate from the gas stream.  Next, a Thomas sample
 pump draws the exhaust through a heated sample line at about 0.8 ft^/min.
 Both the filter and sampling line are heated to prevent phenol loss to
 condensation.  The sample pump then pulls the warm exhaust through two
 impingers, each containing 200 m£ of 1 N KOH chilled to ice bath temperatures
 (0-5°C).  Wet exhaust exiting from the impingers passes through a molecular
 sieve/silica  gel dryer before flowing through the sample pump, flowmeter and
 dry gas meter.  The needle valve on the flowmeter controls flow through the
 sampling system and the dry gas meter measures the volume of gas in cubic
 feet that passes through the impingers.  Gas temperature is measured by an
 iron-constantan thermocouple and can be monitored by a digital readout.
 Pictures of the phenol sampling cart are shown in Figure 2.

 PROCEDURE

     The flow schematic for the analysis of phenols is shown in Figure 3.
 This diagram describes sample treatment from collection to analysis.  Diesel
 exhaust is first diluted in the constant volume sampler.  Phenols  present
 in the diluted exhaust are captured in two glass Greenburg-Smith impingers
 connected in series and chilled in an ice-water bath.  Once collected, the
 samples are quantitatively transferred to 250 m£ polyvinylchloride storage
bottles.


                                    420

-------
ro
       Sample
       Probe
                             Greenburg
                               Smith
                             Impingers
Filter
       Dilute
       Exhaust
                                                                            Gas Temperature
                                                                            Digital Readout
                                                                                Flowmeter
                                              Sample
                                               Pump
                                                                                    Dry
                                                                                    Gas
                                                                                   Meter
                              Figure 1.  Phenols sample collection flow schematic.

-------
                                                    Heated
                                                    Sampling
                                                    Line

                                                    Flowmeter
Ice Water
  Bath
Greenburg-Smith
   Impinger
                                                  Dry Gas Meter
                                                  Sampling Pump
        Figure  2.   Phenols  sampling system.
                       422

-------
       CVS
   Greenburg-
     Smith
   Impingers
     Extraction
     with Ether
      Bubblers
      1 and 2
      Combined
     Sample
Concentrated and
Internal Standard
      Added
 Sample Analysis
 in GC with FID
    Recorder
 A/D Converter
Hewlett-Packard
     3354
Computer System
       Figure  3.   Phenols analysis  flow schematic.
                          423

-------
     The entire workup procedure for phenols is carried out under a vented
hood to prevent ether vapors from escaping into the room.  Explosions can
occur when handling ether, therefore,care needs to be taken not to heat
samples to dryness.  When ready for processing, the sample is poured into
a 500 m& separatory funnel with 1 N KOH washings of the storage bottle.
Thirteen milliliters of 50% H2SO4 is carefully pipeted into the funnel con-
taining the sample and the flask is swirled and very gently shaken with
venting until thoroughly mixed.  Acidity is checked with litmus paper.  Next,
200 m£ of ethyl ether is added, again with swirling and gentle shaking and
venting.  When venting is no longer necessary the separatory funnel is shaken
for two minutes and the two layers are allowed to separate.  The bottom
aqueous layer is drawn off into a second 500 m& separatory funnel and set
aside.  Anyhdrous Na2SC>4  (9.4 g) is added to the phenolic ether mixture
the first funnel with swirling and very gentle shaking and venting until
venting is not needed.  The contents are shaken for two minutes to remove
traces of water.  The dry ether and phenol mixture are transferred to a Kuderna
Danish concentrator with washings.  A boiling chip is added and the solvent
volume is reduced to about one-fourth by heating in a 45°C water bath.  The
Kuderna concentrator is then set aside.

     While the first ether portion is being heated, a second extraction is
performed on the aqueous layer in the second separatory funnel.  First, 100
mH of ether is added to the second funnel with swirling and gentle shaking
and venting.  When venting is unnecessary, the flask is shaken for two minutes.
After the two layers have separated the bottom aqueous layer is drawn off and
discarded.  Anhydrous Na2SC>4 (4.7 g) is added to the ether and phenols mix-
ture and the funnel is swirled, gently shaken and vented.  The flask is sha-
ken for two minutes and the contents are then transferred with washings to
the Kuderna concentrator containing the first ether extraction.  A second
boiling chip is added and the sample is concentrated to approximately 5 m£
by means of the water bath.  The concentrator is cooled to room temperature
and the remaining sample is transferred to a 10 m£ beaker for the final eva-
poration.

     The last drying step is carried out is a desiccator box modified for
nitrogen flow.  A tray of molecular sieve and silica gel  absorbs moisture
that condenses on the beaker during the drying process.  The nitrogen flow is
directed into the 10 m£ beakers containing the phenol sample by a six position
manifold,,  The samples are placed under one of the curved needles on the mani-
fold and the solvent is evaporated to about 1 m£ by a stream of dry nitrogen.
When the sample has warmed to room temperature (about 15 minutes) the con-
centrate is transferred to a 2 m£ volumetric flask with small ether washings.
The sample is then spiked with 100 y£ of 300 yg/m£ o-chlorophenol, the inter-
nal standard.,  The volume is adjusted to 2 m£ with ether and the sample is
labeled, sealed with Teflon tape and refrigerated until samples are ready for
analysis.  Several pictures of the equipment used in the workup of phenol
exhaust samples are shown in Figure 4.

     Phenol samples and standards are analyzed by a Perkin-Elmer 3920B gas
chromatograph (GC) equipped with a flame ionization detector.  The column
used to separate phenols is a 6' x 1/8" Teflon column packed with 10%
OS-138/H3PO4/SP-1200 on 100/120 mesh Chromosorb W AW.  The carrier gas, zero
nitrogen, flows through the column at 50 m£/min.  The temperature of the

                                    424

-------
                     .





   Extraction  Apparatus
Sample Drying Chamber







         foruorkup of






        425

-------
 column is programmed from 70-170°C at a rate of 4°C per minute with  an initial
 hold at 70 °C for 2 minutes.  The purpose of temperature programming  is to
 prevent the solvent peak from obscuring the phenol peaks and  to allow better
 separation of phenols that elute at higher temperatures.  The temperature of
 the injector and interface is maintained at 200°C.  Samples and standards
 are usually analyzed at attention of XI X8 or XI X16 for 30 minutes.
 Injection volume is 1 y£.  A picture of the analysis system is shown  in
 Figure 5.

     The external standard is injected first and response factors are  de-
 termined from the concentrations and areas of each phenol.  Then samples are
 injected and analyzed.  Calculations are performed using sampling information
 and data obtained from the GC analyses.  A chromatogram of the external
 standard is shown in Figure 6 and a chromatogram of a diesel  exhaust  sample
 is shown in Figure 7.

 CALCULATIONS

     The information obtained from the sampling system and the GC analysis
 of phenol samples is used to calculate the concentration of phenols in  ex-
 haust.  The mathematical steps were programmed into a Hewlett-Packard 67
 calculator for rapid data turnover.  A copy of this program is shown in
 Figure 8.  The concentration of individual phenols is determined by comparing
 the area of each phenol to the area of an  internal "standard, JO-chlorophenol.

      Different phenols do not  give the same  response  to the flame ionization
 detector.   Therefore,  a correction needs  to be incorporated into the  calcu-
 lation to account for this difference  in  response.   This correction,  termed
 the response factor,  F, is determined  by  analyzing  the external  standard
 each day before sample analysis.   The  area  per concentration unit of  each
 individual phenol is  compared to  the per  concentration unit of o-chlorophenol,
 the phenol used as the internal standard.
                                C     A
          Response factor (F) = -£*• x   ocpx
                                A     C
                                 px     ocpx

     where   C     = concentration of individual phenol in external standard,
              px     yg/m£
             A     = area produced by individual phenol in external standard,
              PX
                     counts
             A     = area produced by o-chlorophenol in external standard,
                     counts
             C     = concentration of o-chlorophenol in external standard,
              oqpx
Response factors provide the means of calculating the concentration of the
various phenols in the exhaust sample relative to only one phenol, the in-
ternal standard.

     The concentration of any particular phenol in the exhaust sample in
yg/m£ is :


                                     426

-------
p>
[\J
- a
                               Recorder      A/D Converter
     Gas

Chromatograph
                                     Figure 5.  Phenols analytical system.

-------
ro
co
                                o~chlorophencu
                                pnenoi
                                        i athvde
                                r—ere sol an;'- j.—ere sol
                                r—ethylf/nenol  i-ipopropy
                                i , 3-xylerioI, 3,5-xvienol ,
                                   ,6-trimt'thylphencI
                                2,3,5~trimethylpheno
                            7.  2,3,5,6-tetramethylpheno]
                                                                                                         Sample	External- Standarc
                                                                                                         Initrument  F -F  ?9I; "'F     Operato  Hal Baylor
                                                                                                                               1
                                                                                                                                         Tf-f Ion _ Type
                                                                                                                                              id Phase
                                                                                                                                              Support
                                                                                                                                              nCerner
                                    Packed with  1
                                       100/i
                                    Run ISO •  -
                                                      _  _cc/min _ Nitro
                                                     Hotametei Readme
                                          70   °CISO«or  2   min..proBto
                                           /mm. Hrtd tot 3
                                    held for  ~   mm  (olher I
                                   Inlut
                                                                                                                                              170  "C
                                                                                                                                 .. Prog 10 _^__ C at  -  /mm
                                                                                                              200    UC.	Glass  lined	
                                                                                                         Petectof2QO    °C FID	TVpe lotherl	~
                                                                                                           Hyd      PIIQ   -    Rotemeter HdQ- _ ~
                                                                                                                   PMO        Rotemeter RdB
                                                                                                                             Rotameter Rdg
                                                                                                                   n/mm epeed  1   mV.F.S
                                                                                                                   ul indicated
                                                                                                         injection	-
                                                                                                           Sempling Device
                                                   20
    15                     10

Retention  time,  minutes
                                                     Figure  6.   Typical  phenols  external  standard.

-------
ro
                       25
20             15            10
       Retention time, minutes
                                  Figure 7.  Typical diesel-CVS exhaust sample.

-------
     UMT  Insiriiciions
   PHENOLS
               12-17-79
                      MAP
                            HP-6 7
STEP
Ol
"2
°3
1
2
3
5
6
7
8
9
lu
il
12
13
14
15

lib
17

1H
ly
2i>
21
22
23
24
25__
26



-
	
INSTRUCTIONS
Switch to on, switch to run
Peed card in from right to left
Set decimal place
Input - Sample volume
Input - Baronetric pressure
Input - Sample temperature
Input - Cone, of internal standard
Input - Area of internal standard
Input - Phenol area
Input - F (phenol)
Output - Phenol Cone. .
Input - Salicylaldehyde area
Input - P(salicylaldehyde)
Output - Salicylaldehyde cone.
Input - m-and p-cresols area
Input - P{m-and p-cresols)
Output - m-and p-cresols cone.
Input - p-ethylphenol , 2-isopropylphenol,
2,3-and 3,5-xylenols, 2,4 ,6-triroethylphenol are
Input - F (p-ethylphenol, 2-isopropylphenol ,
2,3-and 3,5-xylehols, 2,4/6-trimethylphenol
Output - p-ethylphenol, 2-isopropylphenolr
2,3-and 3,5-xylenols, 2,4,6,-trimethylphenol co
Input - 2, 3,5,-trimethylphenol area
Input - F(2,3,5,-triirethylphenol)
Output - 2, 3,5,-trimethylphenol cone.
Input - 2,3,5,6,-tetramethylphenol area
Input - F(2,3,5,6,-tetrajnethylphenolj
output - 2,3,5,6,-tetramethylphenol eonc.
Input - additional phenol area
Input - F(additional phenol)
Output - additional phenol eonc.



... 	 	 	
— — — ^— ^— — ' 	 — — .
INPUT
DATA/UNITS

TitTIZ
"Hg
°F
pg/m£
counts
counts

counts

counts

	
counts
	
tc.
counts

counts


counts
	



.._. 	
	
KEYS

IDSP Q
A
|h/S |
IR/S |.|
|R/S
l«/s
IR/S
IR/S
IR/S
IR/S
IR/S
IR/S
IB/S

IB/S
IH/S

IR/S
IR/S
IR/S
IR/S

IR/S
[R/S
1 h OTH



1
II 1
OUTPUT
DATA/UNITS

	




Mg/m3

M9/J53.

IHH/5 ._

• i~
lig/m



K^/SC

M'J/m3

tiu/m3





Figure 8.  HP-67 user instructions,
              430

-------
STEP WtM< KEY CODt COMMENTS STEP KEY ENTflY KtY COOL MUU.MT,

-jr:
._ 	 ^_ .,_
. _ 	
1 U1L.A..
_ „
j 	
c.
...2 	
	 -X- 	
	 2. 	 	
M

2

"-'" . I 0
^
1 	 I 	 K/s._._
r x



ujO
	
	



Ma









C*>





STO 1
4
6
	 0
HC'L 1 	
	 ll 1/X
	 X 	

STO 2

X
R/S
X
R/S
R/S
X
RCL 2
X
R/S
R/S
X
FVI. 2
X

R/S
X
RCL 2
3] 25 11
.3 US
	 in .. .
oa 	 	
	 21.....
02 	
_ BJL_ ..
02
81
-siQ
81
	 ai 	
71
3J 0)
04

.-.-. UU 	
aJ.
til
	 21 . .
M.1
81
_.il_02 	
84

H4
7 j

S4
71
34 02
71
84
84
71
_.24.._02 	
71
fl-1 *»_
«4 K
"x^
ni.ut-:;ami.. IL
VoJ ft^
Input h.P. "H,j
lli['US -. aluj. Iclllp "F
Jn[^uf-i oiu;
|i'
ll I-Ul - JU'd
InpuL-ph al
Input.-ph f
Outpljl -ph
iui. -.sal von
Input -pup, 2
i. I,S-X,2,4,
iJ,5-X,2,4,
0 1 2 J
SO SI S2 S3
A
B C
,... -[^
/ni«
Q<, I-
JOI5C
ire a
•
iri|»b',2.3-
6 Imp F
REGl*
4 "
S4


IUHJ
	



	
ud'i


— ._
	

	




1 00








a(-
	
X
.r_vs"..""
X
Rc-'l." 2
X ~
- K/S"
R/S
X
' R/s'J
X
- !"-' R'!'N

^

	
	


	
	 .
















	
	
71
a-l *^.
H-J * ~~*
71 "V
1-1 i).1
'1
"""71
M OJ
)4 1)2
- -— • --
JS 22

	
...





	



.
_

.„


TEHS
5 6 '
S5 So ^ •
D
cone
'til-.' , 1 , ^,-v Hill. Jli: j
In-.;, 1,',,-ttmp F
^HI -J , j/i-t i inp cone
hi-.1 , l.'j ,t.-ti.inp acua
In-J, l,'j,,,-i,,ni|i K
| 'til-Ill '1 J I 1. ,1
-iii-a-iJi Liuual

it
_ 	
— L 	
Figure 8  (Cont'd).  HP-67 program form.
                 431

-------
                       C     x A   x F
                        ocps    sa

                       A
                        ocps




     where   A     = area produced by the phenol in the sample, counts

             Asa   = area produced by o-chlorophenol in the sample, counts

             cocps _ concentration of o-chlorophenol in the sample, yg/mJl
              ocps


The extraction process involves concentrating the phenols from 200 mfi, in the

impingers to 2 m£ in the final sample.  The weight in Ug of the phenol in

the impingers is given by



                       C     x A   x F x V
                        ocps    sa        sa
     where   V   = final sample volume, m&.
              ^t 21
                                                            Equation 1



     To determine the concentration of phenols in dilute exhaust the volume

of exhaust passing through the impingers has to be measured.  This volume,

measured at ambient conditions, is corrected to standard conditions (68°F

and 29.92" Hg) .



                       P   xV     Px.xV^
                        am    am _  stp _ stp

                       T   + 460 ~    T
                        am             stp



     where   P    = barometric pressure at ambient, "Hg

             V    = volume at ambient, ft3
              am
             T    = temperature at ambient, °F

             P    = standard barometric pressure, 29.92 "Hg

             Vs p = volume at 68°F and 29.92 "Hg
              stp
             T .   = standard temperature, 68°F + 460 = 528°R.
              stp
Rearranging and converting ft3 to m3,
                          P         T ^           V
                      _    am        stp    x _ am

                  stp ~   P    X(T   + 460)  "  35.31
                           stp    am
                                                            Equation 2
     Dividing the weight of the phenol captured in the impingers in Equation

1 by the volume of dilute exhaust passed through the bubblers in Equation 2

yields the concentration of the phenol in dilute exhaust in yg/m3.
                  C

                  - — — x A   x F x V
                  A        sa        sa
                   ocps _ _
             P         T ^           V
              am        stp           am
             Pstp       35.31
                                     432

-------
                    C     Ug/m& x A    counts x F x V    mS,
                     ocps ,  _ sa, _ sa,
                               A      counts
                                ocps ,

                   29.92 "Hg x(T     °F +  460) x 35.31  ft3/m3
                           ^    am,
                X  P    "Hg x 528°R  x V     ft3
                    am,                am,
                                                            Equation 3

Equation 3 yields the concentration  of phenols that the Hewlett-Packard 3354
computer identifies as separate peaks.  However,  two peaks  found in several
exhaust samples may contain more  than one phenol,  in which  case an average
response factor is used to calculate the  concentration. For  example, m-
cresol and p-cresol elute at the same time.   The response factor is determined
by comparing the area of  the  cresols peak per total concentration unit of
the two cresols to the area per concentration unit of  o-chlorophenol.  this
method is the same used  for only  one phenol except that the concentrations
of all the phenols under the  peak are  added together.   In  the case of multi-
ple phenols under a  single peak,  the calculated concentration represents the
total concentration  of all  the phenols eluting at that retention time.  One
additional peak  represents  more  than one phenol.   The  phenols eluting under
the same peak are p-ethylphenol ,  2-isopropylphenol,  2,3-dimethylphenol, 3,5-
dimethylphenol  and  2,4,6-trimethylphenol.  All of these compounds are not
necessarily found in automotive  exhaust, but since separation of them
 is not  possible using the specified equipment and ^^\^^'^
 can only be assumed that they are all present   individual ph enols ^ at have
 been  recovered from exhaust are phenol,  salicylaldehyde , 2,3,5,-tnmethyl
 phenol and 2,3,5,6,-tetramethylphenol.

 Sample  Calculation
  trations from raw data.  Experimental values are
  the calculations are performed to give answers in

  Example 1
                                            ^4-1-7 dfn ft3 of dilute exhaust is
      During an FTPh driving cycle, ™™f*JL %l\ KOH.  ^ient tern-
  passed through two impingers containing 200 mx, o   ^ ^ ^ respectively.
  perature and pressure  were ^ted ^be          .^^ standard  was
  Before final concentration 100 y£ of the 200 ^/      ^ standard, at a
  added and  the volume was adjusted to 2 n*.  i      duced an area response
  concentration of 10 pg/ml  in the ^™\S^'Jl°^ °phenol are listed
  of 2159 counts.   The  response factor and area ot eac
  on the following page.
                                       433

-------
SWRI I-ROOECT NO.	



FUEL:	CVS NO.	



SAMPLE COLLECTION BY:_



GENERAL COMMENTS:	
                        TEST NO.
TUNNEL SIZE:
                _TEST DATE:_



                   DRIVER:
     CHEMICAL ANALYSIS BY:
                                                             VEHICLE:
                                    MILES:
                                CALCULATIONS BY:
Test No.
Driving Cycle
Volume, ft3
B.P., "Hg
Temp. °F
Final Sample Vol. (mi)
Cone. Internal Std. (ug/mH)
Area Internal Std.
Area Phenol
F Phenol
Ug/m Phenol
Area Salicylaldehyde
F Salicylaldehyde
Ug/m3 Salicylaldehyde
Area m-and p-cresols
F m- and p-cresols
Ug/m3 m- and p-cresols
. * . -- .
-X, 2,4,6-trrap
F pep, 2ispp, 2,3-S 3,5-X,
2,4,6-trmp
Ug/m3 pep, 2ispp, 2,3-53, 5
-Xr 2r4rfiJ.-1-nnp
Area 2,3,5-trmp
F 2,3,5-trmp
yg/mj 2,3,5-trmp
Area 2,3,5 ,6, -temp
F 2, 3, 5, 6, -temp
Ug/m3 2, 3, 5, 6, -temp
Area additional phenol
F Additional phenol
yg/m additional phenol

1
FTPc
17.582
29.27
77
2
10
2400
1373
0.8354
20
1457
0.5079
13
1023
0.4232
8
412
0.7778
6
876
0.5010
8
1167
0.6342
13
--
-
-

2
FTPh •
17.483
29.27
79
2
10
2159
2209
0.8354
36
6401
0.5140
64
1227
0.5105
12
2029
0.7864
31
1443
0.4766
13
2382
0.6140
29
-
-
- -

3
SET 7
18.400
29.27
79
2
10
2262
1070
0.8379
16
2078
0.4980
18
986
0.4657
3
1569
0.7680
21
2785
0.5284
26
1771
0.5865
18
-
-
-

4
FET
10.435
29.27
78
2
10
2109
879
0.3348
25
6006
0.5136
103
1500
0.5033
25
1283
0.8107
35
2235
0.4546
34
4067
0.6676
91
-
-
-

5
NYCC
8.126
29.27
80
2
10
2387
500
0.8299
16
2868
0.5200
57
3274
0.4451
55
2433
0.7776
72
2567
0.5160
50
3852
0.5744
84
-
-
-

6
85 kph
18.567
29.27
80
2
10
2182
758
0.8350
12
3543
0.5092
33
2921
0.4079
22
4155
0.8091
61
2488
0.5518
25
981
0.6020
11
-
-
—

                   Figure 9.   Phenol  data sheet.
                                   434

-------
                Phenol                  p
                	•—	                   Area

         Pheno1                     0.8345                2209
         Salicylaldehyde            0.5140                6401
         m-cresol and p-cresol      0.5105                1227
         p-ethylphenol,
           2-isopropylphenol,
           2,3- and 3,5-xylenol,
           2,4,6-trimethylphenol     0.7864                2029
         2,3,5-trimethylphenol      0.4766                1443
         2,3,5,6-tetramethylphenol  0.6140                2382

     The  appropriate values are plugged into Equation 3  for each phenol.

Phenol

         10 yg/ml, x 2209 counts x 0.8345 x2.0m£
                         2159  counts

         29.92 "Hg x (79°F +  460)  x 35.31 ft3/m3
              29.27 "Hg x 528°R x 17.483 ft3

       =  36 yg/m  phenol

Salicylaldehyde

         10 yg/m£ x 6401 counts x 0.5140 x 2.0 n&
                      2159 counts

         29.92 "Hg x (79°F +  460)  x 35.31 ft3/m3
       X      29.27 "Hg x 528°R x 17.483 ft3

       =  64 yg/m3 Salicylaldehyde

m-cresol  and p-cresol = 12 yg/m3

E-etfaylphenol,  2-isopropylphenol, 2,3- and 3,5-xylenol and2,4,6-triinethylphenol

         10 yg/m£ x 2029 counts x 0.7864 x 2.0 m&
                      2159 counts
       x
29.92 "Hg x (79°F + 460) x 35.31 ft3/m3
  29.27 "Hg x 528°R x 17.483 ft-*
       =   31  yg/m3 p-ethylphenol,  2-isopropylphenol,  2,3- and 3,5-xylenol,
                   2,4,6-trimethylphenol

1/3,5-trimethylphenol  = 13 yg/m3,

2,3,5,6-tetramethylphenol = 29 yg/m3

                                    435

-------
Example 2

     Suppose that during the FTPc driving cycle 17.582 ft  of dilute exhaust
measured at 29.27 "Hg and 77°F passed through two impingers.  The sample
was processed and analyzed following the phenols procedure.  The 10 yg/m£
internal standard gave a responce of 2400 counts.  The response factors and
areas of the other phenols are listed below.
         Phenol                  F

    Phenol                    0.8354
    Salicyaldehyde            0.5079
    m-cresol and
     p-cresol                 0.4232
    p-ethylphenol,
     2-isopropylphenol,
     2,3-and 3,5-xylenol
     and 2,4,6-trimethyl-
     phenol                   0.7778
   2,3,5-trimethylphenol      0.5010
   2,3,5,6-tetramethylphenol  0.6342
Phenol
          10 yg/m& x 1373 counts x 0.8354 x 2.0 m£
                        2400 counts


       x  29.92 "Hg x (77°F + 460)  x 35.31 ft3/m3
             2g>27 ,,Hg x 528oR x 17.582

       =  20 yg/m3 phenol
Salicylaldehyde
x
          10 yg/m£ x 1457 counts x 0.5Q79 x 2.0 m&
                     2400 counts

          29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
             29.27 "Hg x 528°R x 17.582
       =  13 yg/m3 salicylaldehyde

m-cresol and p-cresol

         10 yg/m£ x 1023 counts x 0.4232 x 2.0 m
                       2400 counts

         29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
            29.27 "Hg x 528°R x 17.582 ft3

      =  8 yg/m3 m-cresol and p-cresol
                                                      Area

                                                      1373
                                                      1457

                                                      1023
                                                       412
                                                       876
                                                      1167
                                     436

-------
?-ethylphenol, 2-isopropylphenol, 2,3-andJ^-xylenoj^and 2 ,4,6,-trimethylphenol

          10 yg/m£ x 412 counts x 0.7778 x 2.0 m£
                       2400 counts

       x  29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
             29.27 "Hg x 528°R x 17.572 ft3

       =  6 yg/m3 p-ethylphenol, 2-isopropylphenol, 2,3-and 3,5-xylenol
                  2,4,6,-trimethylphenol

2,3,5-trimethylphenol = 8 yg/m3

2,3,5,6-tetramethylphenol = 13 yg/m3

LIST OF EQUIPMENT

     The equipment needed for collection, workup and analysis of phenols in
exhaust is listed below in separate sections.  The manufacturer, catalog or
model number and description are given for each entry.

Sampling

     1.   Greenburg-Smith glass impingers, Houston Glass Fabricating, Catalog
          #310610-0028, arm joints 28/15, bottle joint 45/50.

     2.   Ground glass socket joint with arm modified to 5/16",  Houston
          Glass Fabricating, Catalog #285045, 28/12 socket.

     3.   Ground glass ball joint with arm modified to 5/16", Houston Glass
          Fabricating, Catalog # 285040, 28/12 ball.

     4.   L-shaped glass connecting adapter, Houston Glass Fabricating,
          Catalog #015639, male and female size 28/12.

     5.   U-shaped glass connecting adpater, Houston Glass Fabricating,
          Catalog #160719 size 28/15 male-male socket joint, 2.25" center
          to center length.

     6.   Thomas ball and socket joint clamp, Houston Glass Fabricating,
          Catalog # 285100 size 28.

     7.   Class A, 2000 m£ volumetric flask.

     8.   Flowmeters, Brooks Instrument Division, Kynar, Sho-Rate "150" with
          R-6-15-B metering tube, stainless steel ball, 1-92 CFH range,
          graduated 1-100.

     9.   Dry gas meter, American Singer Corporation, Type Al-120, 60 CFH
          capacity.


                                     437

-------
    10.    Sample pump,  Thomas  Model #727CA39,  1 ft3/min free flow capacity.

    11.    Drying tube,  Analabs Inc.,  Catalog #HGC-146, 6" long.

    12.    Teflon tubing,  United States  Plastic Corporation, 5/16" OD x 1/8"
          ID and 3/8"  DD  x 1/4" ID.

    13.    Miscellaneous Teflon nuts,  ferrules, unions/ tees, clamps and
          connectors/  etc.

    14.    Miscellaneous electrical  switches, lights,  wiring, etc.

    15.    Miscellaneous Swagelok fittings.

    16.    Athena temperature controller,  Technical Heaters Incorporated,
          Model 6000,  100-600°F range,  110  volts.

    17.    Heated sample line.  Technical Heaters Incorporated, Catalog
          #LP-212-8-5,  5'  length, 13/32"  hose  with 1/2" tube end stainless
          steel fittings.

    18.    Pallflex Fiberfilm filters, Pallflex Products Corporation,
          Catalog #T60A20, 70  mm diameter.

    19.    250 m£ polyvinylchloride  sample storage  bottles, Nalgene Labware,
          Catalog #2000-0008.

    20.    Iron/constantan type J single thermocouple, Thermo Sensors  Corp.

Workup

     1.    Pear shaped 500  mJl separatory  funnel, Houston Glass Fabricating,
          Catalog #260145, with Teflon  stopcock.

     2.    Class A, 3 and  10 m£ volumetric pipets.

     3.    Safety bulb  for pipetting,  Markson Science  Supplies, Catalog #E-8074.

     4.    Miscellaneous glass  beakers.

     5.    250 mJl glass  graduated cylinder.

     6.    Basic indicating litmus paper.

     7.    Ring stands.

     8.    Kuderna Danish  Concentrator,  Ace  Glass  Incorporated, Catalog
          #6708-03 and  6708-35*

     9.    Boileezers boiling chips, Fisher  Scientific Company, Catalog
          #B-365.


                                     438

-------
   10.   Hot plate with  heat control.

   11.   2000 m£ beaker  for water bath.

   12.   10 m£ glass beakers with pouring spout.

   13.   Disposable transfer pipets, Curtin Matheson Scientific Incorporated,


   14.   Class A,  2 m£ volumetric flasks with hexagonal base, Fisher
         Scientific Company, Catalog #20814B.

   15.   Boekel desiccating cabinet modified for nitrogen flow, Curtin
         Matheson  Scientific Incorporated, Catalog #076-190.

   16.   6 position gas  manifold for sample concentrating,  Alltech Associ-
         ates, Catalog #9555.

   17..  Zero grade nitrogen

   18.   Teflon Lab-Tape,  Fisher Scientific Company, Catalog #14-831-300A,
         13 mm width.

Analysis

     1.   Perkin-Elmer  Model 3920B gas chromatograph equipped with a flame
         ionization detector.

     2.   Soltec Model  B-281 1 mv recorder.

     3.   Hewlett-Packard Model 3354 GC computer system with remote teletype
         printout.

     4.   Hewlett-Packard Model 1865A A/D Converter.

     5.   5 yJl Hamilton liquid syringe, Alltech Associates,  Catalog #N-75.

List  of Reagents

     All  compounds  used in sample acquisition and workup are listed in this
section.  Formula weights,  grade of purity, manufacturer and catalog number
are listed for each reagent.

     1.   Potassium  hydroxide, KOH, formula weight = 56.11, Mallinckrodt, 85%
         analytical reagent grade pellets, Catalog #6984.

     2.   Ethyl ether, anhydrous,  (CH3CH2)20, formula weight  =74.12,
         Mallinckrodt,  ACS analytical reagent grade, Catalog #0848.
                                    439

-------
     3.   Sodium sulfate,  anhydrous,  Na2S04,  formula weight = 142.04,
         Mallinckrodt,  ACS analytical reagent grade, granular, Catalog
         #8024.

     4.   Sulfuric acid, H2S04,  formula weight = 98.08, Mallinckrodt, ACS
         analytical reagent grade, Catalog #2876.

     5.   o'-chlorophenol,  Cl (C6H4OH) ,  formula weight = 128.56, Eastman Kodak
         Company, analytical reagent grade,  Catalog #1087.

     6.   Phenol,  C6H5OH,  formula weight = 93.11, Mallinckrodt, ACS analytical
         reagent grade, loose crystals, Catalog #0028.

     7.   Salicylaldehyde, 2-HOCgI^CHO, formula weight=122.13, Eastman Kodak
         Company, analytical, reagent grade. Catalog #225.

     8.   m-cresol, CH3  C6H4OH,  formula weight = 108.14, Aldrich Chemical
         Company, 99+%  Gold Label, Catalog #C8,572-7.

     9.   p-cresol, CH3C6H4OH, formula weight = 108.14, Aldrich Chemical
         Company, 99+%  Gold Label, Catalog #C8,575-1.

    10.   p-ethylphenol  (4-ethylphenol), C2H5C6H4OH, formula weight = 122.17,
         Aldrich Chemical Company, 97%, Catalog #E4,420-5.

    11.   2-isopropylphenol, (CH3)2 CHC6H4OH, formula weight = 136.19,
         Aldrich Chemical Company, 97%, Catalog #12,952-6.

    12.   2,3-dimethylphenol (2,3-xylenol), (CH3)2  C6H3OH,  formula weight=122.17,
         Aldrich Chemical Company, 97%, Catalog #D17,400-9.

    13.   3,5-dimethylphenol (3,5-xylenol), (CH3)2  C6H3OH,  formula-weighty
         122.17,  Aldrich  Chemical Company, 99.9+%  zone refined,
         Catalog #15,085-1.

    14.   2,4,6-trimethylphenol, (CH3)3C5H2OH, formula weight = 136.19,
         Aldrich Chemical Company,  99%, Catalog #T7,900-6.

    15.   2,3,5-trimethylphenol, (CH3)3C5H2OH, formula weight = 136.19,
         Aldrich Chemical Company, catalog #T7,860-:3.

    16.   2,3,5,6-tetramethylphenol,  (CH3)4 CgHOH,  formula weight = 150.22,
         Aldrich Chemical Company, 90+%, Catalog #17,877-2.

Preparation of Reagents

Absorbing solution, 1 N KOH

     132.02g of KOH pellets are dissolved in deionized water in a 2000 ml
volumetric flask.  When the solution cools to room temperature the volume is
adjusted to 2000 ml.


                                    440

-------
50% H2SO4


     Concentrated H2SO4 (500 mfc) is slowly added to 500 m£ of chilled
deionized water with swirling and  shaking.   The solution is stored in a
glass bottle.

Internal Standard - 300 yg/m£ o-chlorophenol

     Approximately O.OSOOg  of o-chlorophenol is added to a 100 ml volumetric
flask and filled to volume  with ethyl  ether.  This solution is sealed with
Teflon tape and refrigerated.

External standard stock solution

     The external standard  is prepared by blending the following phenols in
a  1000 m& volumetric flask  and  diluting with CH2C12.  This solution is
sealed with Teflon tape and refrigerated.

                                          Wt.  (mg)

                  Phenol                      150
                  Salicylaldehyde             300
                  m-cresol                   150
                  p-cresol                    50
                  p-ethylphenol              100
                  2-isopropylphenol          150
                  2,3-xylenol                 100
                  3,5-xylenol                  50
                  2,4,6-trimethylphenol      100
                  2,3,5-trimethylphenol      100
                  2,3,5,6-tetramethylphenol   200
                  o-chlorophenol             300

 Dilute  external  standard

      10 m£  of the stock external standard to diluted to 100 m£ with CH2C12.
 This solution is sealed with Teflon tape.
                                      441

-------
 REFERENCES

 Brown,  R. A.,  Searl,  T.  D.,  King Jr.,  W.  H.,  Dietz,  W. A. and Kelliher,
 J. M..,  Rapid Methods  of  Analysis for Trace Quantities of Polynuclear
 Aromatic Hydrocarbons and Phenols in Automobile Exhaust, Gasoline and
 Crankcase Oil,  Final  Report  for CRC - APRAC Project  CAPE-12-68, Esso
 Research and Engineering Company, Linden,  New Jersey, 1973.

 Carter, M.  J.  and Huston,  M.  T., Environ., Sci. Tech., Vol. 12, pg. 309, 1978.

 Preston, S. T., A Guide  to the Analysis of Phenols by Gas Chromatography,
 PolyScience Corporation,  Niles, Illinois, 1966.

 Bartie. K.  D.,  Elstub, J., Novotny, M. and Robinson, R. J., Journal of
 Chromatography, Vol.  135,  pg. 351,  1977.

 Ma,    T. S. and Speigel,  D., Microchemical Journal, Vol. 10, pg. 61,  1966.

 Sheffer, H. E., Perry, R.  L., Thimineur,  R. J., Adams, B. T., Simonian,
 J. L.,  Zutty,  N. L. and  Clendenning,  R. A., Ind.  Eng. Chem. Prod. Res.
 Develop., Vol.  10, pg. 362,  1971.

 Yrjanheikki, Erkki, Am.  Ind.  Hyg. Assoc.  J.,  Vol.  39, pg. 326, 1978.

 Stanley, T. W., Sawicki,  E.,  Johnson,  H.  and  Pfaff,  J. D.,  Mikrochim,  Acta,
 Vol.  1, pg. 48, 1965.

 Umbreit, G. and Houtman,  R.  L.,  Journal of Pharmaceutical Sciences, Vol. 56,
 pg. 349, 1967.

Lawrence, J. F., Journal  of Chromatographic Science,  Vol. 17,  pg. 147,  1979.

 Wolkoff, A. W.  and Larose, R. H., Journal  of  Chromatography,  Vol. 99,  pg. 731,
 1974.
                                               *
 Afghan, B.  K.,  Belliveau,  P.  E., Larose,  R. H. and Ryan, J. F., Analytica
 Chimica Acta, Vol. 71, pg. 355,  1974.

 Mindrup, Jr., R.,  Industrial  Research/Development, pg. 79,  August,  1978.

 Eichelberger, J.  W.,  Dressman,  R. C.  and  Longbottom, J. E., Environmental
 Science and Technology,  Vol.  4,  pg. 576,  1970.

 Barber, E.  D.,  Sawicki,  E. and  McPherson,  S.  P., Analytical Chemistry,
 Vol.  36, pg. 2442, 1964.

 Heenan, M.  P. and McCallum,  N.  K.,  Journal of Chromatographic Science,
 Vol.  12, pg. 89,  1974.

 Grouse, R.  H.,  Garner, J.  W.  and O'Neill,  H.  J.,  J.  of G.C.,  pg. 18,
 February, 1963.


                                     442

-------
Kawahara, F. K., Analytical Chemistry,  Vol.  40,  pg.  1009,  1968.

Kushnir, I., Barr, P. A. and  Chortyk,  O.  T.,  Analytical  Chemistry  Vol  42
pg. 1619, 1970.                                                 *'    *

Spears, A. W., Analytical  Chemistry,  Vol. 35, pg.  320, 1963.

Chriswell, C. D., Chang, R. C.  and Fritz, J.  S., Analytical Chemistry
Vol. 47, pg. 1325, 1975.

Braithwaite, B. and Penketh,  G.,  Analytical  Chemistry, Vol. 36, pg. 185, 1964,

Shulgin, A., Analytical Chemistry, Vol. 36,  pg.  920, 1964.

Argauer, R. J., Analytical Chemistry,  Vol. 40, pg. 122,  1968.

Landault, C. and  Guiochen, G.,  Analytical Chemistry, Vol.  39, pg. 713, 1967.

Fisher,  G. E. and Neerman, J. C., I and EC Product Research and Development,
Vol. 5,  pg. 288,  1966.

Cohen,  I. C. Norcup,  J.,  Ruzicka, H.  A. and Wheals,  B. B., J. Chromatog.,
Vol. 44, pg. 251, 1969-

Bark,  L. S. and Clarke,  K. F.,  J. Chromatog., Vol. 48, pg. 418, 1970.

Di Corcia, A., Journal  of  Chromatography, Vol. 80, pg. 69, 1973.

Seiber,  J. N., Crosby,  D.  G.  Fouda, H. and Soderquist, C.  J., J. Chromatogr.,
Vol. 73, pg. 89,  1972.
                                       443

-------
   APPENDIX K




SULFATE PROCEDURE
        444

-------
               Determination of Soluble  Sulfates  in Automobile
                Exhaust by Automated HPLC Modification of the
                         Barium Chloranilate  Method

                                     by

                  Silvestre B. Tejada, John E.  Sigsby, Jr.
                            and Ronald L. Bradow
                  Mobile Source Emissions Research Branch
                 Environmental Sciences  Research  Laboratory


1.    Principle and Applicability

1.1   Principle

      Automotive exhaust is vented  into  a dilution tunnel where it is mixed
with a flowing stream of cool filtered air.   In the tunnel, the SO3 reacts
rapidly with water in the exhaust to form sulfuric acid aerosols.  The aero-
sols are allowed to grow to filterable size range and are collected on fluoro-
carbon membrane filter downstream of the tunnel via isokinetic probes mounted
in the following aerosol stream.  Particulate sulfate salts are collected as
well.

      Sulfuric acid on the filter is converted  to ammonium sulfate by expo-
sure to ammonia vapor.  The soluble sulfates  are  leached from the filter with
a measured volume of 60% isopropyl  alcohol  -  40%  water solution (60% IPA).  A
fixed volume of the sample extract  is injected  into a high pressure liquid
chromatograph (HPLC) and pumped through  a column  of strong cation exchange
resin in Ag+ form to scrub out the  halides  (Cl~,  Br~) , then through a column
of strong cation exchange resin in  H+ form  to scrub out the cations and con-
vert the sulfate to sulfuric acid and finally through a reactor column of
barium chloranilate crystals to precipitate out barium sulfate and release
the highly UV absorbing chloranilate ions.  The amount of chloranilate ions
released is equivalent to the sulfate in the  sample and is measured by a
sensitive liquid chormatograph UV detector  at 310-313 nanometers.  All the
reactions and measurement take place in  a flowing stream of 60% IPA.  The
scrubber and reactor columns also function  as efficient filter media for any
solid reaction products formed during passage of  the sample through the
column system.

1.2   Applicability

      The method as specified is applicable to  the determination of soluble
sulfates in automobile exhaust.  It may  be  used for the analysis of sulfates
in samples where the sulfates can be leached  out  with water or aqueous IPA


                                      445

-------
solution.  Aqueous extracts must be made up to 60% IPA before  they  can be
analyzed.

2.    interferences

2.1   Cationic interferences are removed by the strong cation  exchange resin
      in H+ form.

2.2   Halide interferences are removed by the strong cation exchange resin in
      Ag+ form.  Other anions which form insoluble salts with  silver are also
      removed.

2.3   Sulfide is measured quantitatively as sulfate.

2.4   Anions which form strong acids after passage through the cation exchanger
      in H+ form interfere positively.  Nitrate at 60 ygs/m£ gives  an apparent
      sulfate response corresponding to 8 ygs/m£.

2.5   Presence of anionic interference is manifested by a negative peak
      immediately preceding the positive apparent sulfate peak.

2.6   Organics with absorption bands at 310-313 will interfere positively.

3.    Range, Sensitivity and Precision

3.1   The absolute amount of SO4= normally injected into the system is between
      0 - 12.5 ygs.  For a typical sample injection volume of  0.5 m£, this
      translates to a concentration working range of 0 ?• 25 ygs/mJl.  Working
      range can be extended by using a smaller sample injection volume for
      concentrated samples or conversely, by using a larger injection volume
      for dilute samples.

3.2   Minimum detectable quantity of SO^ is in the low nanogram range.  There
      are commercially available liquid chromatograph UV detectors capable of
      detecting 5 nanograms of sulfate at a signal to noise ratio of about 5.
      Figure 1 shows typical recorder response for a 0.5 m£ injection of
      sulfate samples at trace concentration levels (0.01 - 0.1 ygs/m&.

3.3   Precision better than 3% at 0.5 yg/m£ SO^  level and better than 2%
      between 1 and 20 ygs/m£ for four repetitive runs have been attained.
      Table 1 and Figure 2 show typical reproducibility obtained with the
      automated BCA system.

4.    Apparatus

      A schematic of the principal components of the automated BCA  set up is
      shown in Figure 3.

4.1   Hardware

4.1.1  Basic System


                                    446

-------
4.1.1.1  High pressure  liquid chromatoqraph pump (LP).   The pump must be
         capable of delivering liquids at flow rates of at least 3 mVmin  at
         pressures as high as 1200 psi.  Liquid pumps capable of delivering
         pulseless and  constant liquid flow are recommended for good  quanti-
         tation.  Most  HPLC pumps in the market are adequate.

4.1.1.2  UV detector  (6)  equipped with low dead volume  (8 y£) flow-through
         cell and a grating, prism or appropriate interference filter to iso-
         late a narrow  radiation band centered at 310-313 nanometers.  For
         low noise and  long life, a UV detector equipped with low pressure
         mercury lamp and a 313 nanometer narrow band pass interference filter
         is recommended.

4.1.1.3  A two position,  six port, high pressure, low dead volume sample in-
         jection valve  (SV) .  This must be equipped with interchangeable
         external loop  (L) .  Two loop sizes are desirable:  a 0.5 m£  volume
         for the 0-20 yg/m£ range and a 100 y£ volume for 0 - 100 ygs/m£.
         range.  The  sample valve must be equipped with an external handle
         for manual operation or an air actuator for remote and/or automatic
         operation.   These valves are commercially available and have  pressure
         ratings as high as 7000 psi.

 4.1.1.4  Recorder.  Must be multi-range, with chart speed as low as 10 minutes
         per inch.  Dual channel is preferable so that chromatogram can be
         simultaneously recorded at two different sensitivities.

 4.1.1.5  Strong cation  exchange resin column (CX-Ag+),  4 mm I.D. by 1/4 inch
         O.D. by 6  inches long stainless steel column packed with chromato-
         graphic grade, strongly acidic cation exchange resin in silver (Ag+)
         form.

 4.1.1.6  Strong cation  exchange resin column (CX-H+), 4 mm I.D. by 1/4 inch
         O.D. by 9  inches long, stainless steel column  packed with strongly
         acidic cation  exchange resin in hydrogen  (H+)  form.

 4.1.1.7  Barium chloranilate column  (BCA) , 4 mm I.D. by 1/4 inch by 1 inch
         long  stainless steel column packed with crystalline barium chlor-
         anilate .

 4.1.1.8  1/4"  to  1/16"  stainless steel reducer preferably fitted with 5 micron
         pore  size  frit - for column inlet and end fittings.

 4.1.1.9  1/4"  to  1/4" stainless steel unions.

 4.1.1.10  1/16" to  1/16" low dead volume stainless steel couplings to inter-
           connect  CX-Ag+ to CX-H+ to BCA columns.

 4.1.1.11  1/4" and 1/16" nuts and ferrules.

 4.1.1.12  Reservoir (LR) for the solvent  (60% IPA) .
                                      447

-------
4.1.2  Options - These items are needed for automating the basic system.

4.1.2.1  Integrator, for measuring peak areas.  Recommended unit must have
         baseline tracking capability and possibly with built-in calculation
         accessory.  The integrator extends useful dynamic range of detector
         response for a given sample loop size beyond that of the strip chart
         recorder.

4.1.2.2  Peristaltic pump, (PP), to draw sample from its container and load
         it into sample loop.  Silicone pump tubing is recommended.

4.1.2.3  Automatic sampler (AS), available from commercial sources.  This is
         needed if the number of samples to be analyzed is large and manpower
         is limited.

4.1.2.4  Three timer relays to control pump, sampler, injection valve and
         integrator and provide automatic reset for cyclic operations.

4.1.2.5  Prepackaged sampler systems for HPLC application are commercially
         available.

5.    Principle of Operation

      Solvent  (60% IPA) in reservoir LR (Figure 3) is continuously pumped
by an HPLC LP through a column of strong cation exchange resin in silver
form, CX-Ag"^7 then through a column of strong cation exchange resin in hydro-
gen form, CX-H+, then through a reactor column of barium chloranilate, BCA,
and finally through a flow-through cell of a UV detector, 13, and on to waste.
CX-AG+ removes the halides (Cl~, Br~, F~)  and other anions which precipitate
with silver; CX-H+ removes metallic cations and converts the sulfate to sul-
furic acid, and the BCA reacts with the sulfate to form barium sulfate pre-
cipitate and a soluble UV absorbing dye, chloranilic acid and its ions.  Back-
ground absorbance at 310-313 nanometers is continuously measured and monitored
on a strip chart recorder.

      Sample is introduced into the system without flow interruption by means
of a two-way six port low dead volume sample injection valve SV.  In load
position "A" (see Figure 3) the peristaltic pump, PP, draws the sample from a
cuvette in the automatic sampler AS and pumps it into port 4 filling external
sample loop L_ then through port 5 to waste.  (Sample loop loading may also be
accomplished by pushing the sample through port 4 by means of a syringe.) The
high pressure liquid flow comes in through port 1, bypasses the loop L, comes
out of port 2, and continues on through the three columns and the flow-through
cell of the UV detector 13 and on to waste.

      After loop L_ is loaded with sample, injection valve SV switches to in-
ject position "B".  The high pressure stream purges the loop and pushes the
sample through the cation exchangers and then through the BCA column where
the color reaction takes place.  The BaS04 Precipitate is retained in the
column while the acid chloranilate is carried by the flowing liquid through
the detector system for colorimetric measurement.
                                      448

-------
 6.     Apparatus

 6.1   Pipette, volumetric, 1, 2,  4,  5,  8,  25,  50,  100 m£

 6.2   Pipette, measuring, 1, 2, 5, 10 m£

 6.3   Automatic burette, 25 m£

 6.4   Volumetric flasks, 10, 25, 50, 100, 500, 1000, 2000 m£

 6.5   Bottles, polypropylene, with screw caps, 30, 60, 125, 250,  500, 1000

 6.6   Microbalance

 6.7   Vortex test tube mixer

 6.8   Centri fuge

 6.9   Magnetic mixer

 6.10   Magnetic bars

 6.11   Graduated cylinders

 6.12   Automatic dispenser pipet,  5,  10,  20  mH (optional)

 6.13   Automatic burette,  10 m£  (motor driven, optional)

 6.14   Ammoniation chamber (Figure  5)

 7.     Reagents

 ^•1    Isopropyl  alcohol  (IPA), spectro quality grade or equivalent

 7.2   Water, doubly deionized, distilled.

 '•^   60% IPA. Add  4 parts water to 6 parts IPA by volume.   Store  in tightly
      capped bottles.

 7.4   Barium chlo rani late, suitable for sulfate analysis.   Must be crystalline,
      granular, preferably with average granule length of about 200 microns.
      Finer particles cause excessive column pressure drop.

 7.5    Cation exchange resin, chromatographic grade, strongly acidic, hydrogen
      form,  100 - 200 mesh.

7.6    Hydrochloric acid (4N) .  Add 30 mi concentrated hydrochloric acid to
      6.0 mH  of deionized water.

7.7    Ammonium sulfate, primary standard
                                     449

-------
7.8   Silver nitrate (IN).   Dissolve 17 grams silver nitrate in deionized
      water and make up to  100 m£.   Store in the dark in an amber colored
      reagent bottle„

8.    Procedure

8.1   Column Preparation

8.1.1  Barium chloranilate  column.   In order to prepare a full column with
       minimum dead volume, connect two lengths of 4 mm I.D., 1/4" O.D.
       stainless steel tubing as shown in Figure 4 with a = 1", b = 2".
       Connect a small funnel to open end of B with a flexible tubing sleeve.
       Fill the funnel halfway with barium chloranilate and thump the tube
       several times or use a vibrator (i.e., electric pencil engraver) to
       pack the solid in the column.  Continue the operation until B is com-
       pletely filledo  Remove the  funnel and cap the open end of B with a
       1/4" to 1/16" reducer fitted with 5 micron stainless steel frit.  The
       5 micron stainless steel frit in column A may be replaced with a stain-
       less steel wire screen with  nominal porosity of 10 microns.  Connect B
       to the HPLC pump, connect a  flexible tubing at A and direct the tubing
       to waste reservoir.   Fill HPLC pump reservoir with 60% IPA.  Activate
       HPLC pump according  to manufacturer's instructions, set flow at 3 - 4
       m&/min and let solvent flow  for at least 20 minutes.  This procedure
       removes the fines which can  cause background drift and at the same
       time compresses the  barium chloranilate crystals to fill the dead vol-
       ume in column A.  Deactivate the HPLC pump, disconnect the composite
       column from the pump, then column A from the composite column.  Con-
       nect a 1/4" to 1/16" reducer fitted with 5 micron frit to open end of A.

8.1.2  Cation exchange resin columns

8.1.2.1  Cation exchange resin hydrogen form.  Add strongly acidic cation ex-
         change resin, 100  - 200 mesh, to 160 m£ of 4 N HC1 in a 250 m&
         Erlenmeyer flask until a wet volume equivalent to 40 m£ has settled
         at the bottom.  Soak for at least 3 hours with occasional stirring
         with a glass rod.   Decant  the acid, add 100 m£ deionized water, stir
         and decant the liquid as soon as most of the solids have settled at
         the bottom.  This  procedure removes most of the fines.  Repeat
         rinsing procedure  several  times until the rinse liquid gives neutral
         reaction to pH paper.  Transfer half of the resin to a 150 m&
         Erlenmeyer flask for conversion to the silver form.

         Connect two sections of  1/4" O.D., 4 mm I.D. stainless steel tubing
         as in 8.1.1 with a = 9 and b = 5.  The reducer on the outlet end of
         A should have a 5  micron stainless steel filter frit.  These frits
         are available commercially.  The fritt must be able to withstand
         high pressure (1200 psi).   (If the cation exchange resin breaks
         through the frit and comes to contact with the barium chloranilate
         column, the column gets plugged.)  Connect a small funnel to open end
         of B with a flexible tubing sleeve.  Clamp composite column vertically
         and connect open end of A  to vacuum line equipped with liquid trap.


                                     450

-------
         Fill funnel with deionized water and turn vacuum slowly so that the
         column IB completely  filled with water.   Add enough water so that
         water level is in the funnel  cone;  stop  the vacuum and add the
         slurry of freshly washed  resin (H+  form).  Let the resin settle by
         gravity until the resin top is halfway in the funnel stem.  Open the
         vacuum slowly and keep adding the resin  slurry until the composite
         column is completely  filled.   Proceed as  in 8.1.1 beginning with the
         sentence "Remove funnel and cap open end of B...".

8.1.2.2  Cation exchange resin, silver form..   Add 60 m£ of 1 N AgNO3 solution
         to the other half of  the  washed cation exchange resin, hydrogen form,
         in a 150 m£ Erlenmeyer flask.   Stir with a glass rod, cover the
         the flask with aluminum foil  and soak the resin overnight.

         Decant the AgNO3 solution into a waste reservoir.  Add 100 m£ de-
         ionized water, stir and decant the  liquid as soon as most of the
         resins have settled at the bottom.   Repeat the rinsing procedure
         until the rinse liquid remains clear when treated with a few drops
         of 4 N HC1.

         Connect two sections  of 1/4"  O.D.,  4 mm  I.D. stainless steel tubing
         as in 8.1.1 with a =  6" and b = 5".   Load the column following the
         procedure described in section 8.1.2.1.

8.2   Priming System for Analytical Run

      Connect outlet end of cation exchange  resin  (Ag+ form) column to inlet
end of cation exchange resin  (H+ form)  column with a low dead volume 1/16"
to 1/16" stainless steel tubing connector.  Similarily connect the outlet end
of the second column to the barium chloranilate column.   Plumb the composite
column to the automated set up as  shown in Figure  3.   Fill solvent reservoir
LR with 60% IPA, activate the  HPLC pump, detector, recorder, sample injection
valve, sampler and peristaltic pump.   During this  initial operation dip the
samplling probe in 100 m£ of 60% IPA.   Set the liquid flow rate at 3 m£/min.
(Flow rate is conveniently measured by directing  the effluent from the UV
detector to a microburette and measuring the time  in seconds needed to fill
a volume of 3 m£.)  Let run for at least 30  minutes.   Deactivate sample in-
jection valve, sampler and peristaltic pump.   Leave other components in opera-
ting mode.  When absorbance background is stable  at the appropriate sensitivity
setting of the detector, the system is ready to analyze sulfate samples.

8.3   Preparation of Standards

      Sulfuric acid, sodium sulfate or ammonium sulfate may be used as stan-
dards.  Ammonium sulfate is preferred.

8.3.1  S04=  (100 ygs/m£) standard, alcoholic stock solution   DissoJve 275000
       ± 100 pgs of primary standard ammonium sulfate in 200 m£ of ^onized
       water in a 2000 m£ volumetric flask.   Add  300 m£ pure IPA  shake vig
       orously until thoroughly mixed,  and make up to volume wxth 60% IPA^
       Store in clean polypropylene bottles.   (Note:   a.  There is a volume
       decrease of about 2.7%  when two parts of water xs naxed wxth 3 parts


                                     451

-------
       of IPA.   b.   Do not use detergents nor dichromate-sulfuric acid solu-
       tion for cleaning glasswares.  Sulfate background from these sources
       are difficult to remove.  50% (v/v) nitric acid/water is preferable.)
       Prepare  from this stock solution S04=  calibration standards (0.5 to
       20 ygs/m&) by dilution of appropriate aliquots with 60% IPA.  Store
       standards in capped polypropylene bottles.

8.3.2  S04= (100 ygs/m&) standard, aqueous stock solution.  Dissolve 275000
       ± 100 ygs of primary standard ammonium sulfate in 200 m£ of deionized
       water in a 2000 m£ volumetric flask and make up to volume with de-
       ionized water.  Store in polypropylene bottle.

8.3.3  Alternative  method of preparing calibration standards.  Use a  repeti-
       tive dispenser, burette or automatic syringe pump.  Using either the
       alcoholic or aqueous S04= (100 ygs/m£) stock solution dispense appro-
       priate volumes containing 10, 20, 40 ..., 180, 200 ygs of SO4=  into 30
       of 60 mS, polypropylene bottles.   Prepare 10 for each sulfate level.
       Evaporate the liquids completely by placing the bottles uncapped in an
       oven maintained at 80 - 100°C.  Cool, cap the bottles and store until
       ready to use.  These solid standards can then be extracted in the same
       manner as the filter samples.

8.4   Ammonia Treatment of Filter Samples

      This treatment converts sulfuric acid particulate into ammonium sulfate.
Conversion of ammonium salt was observed to improve precision of sulfate mea-
surement. Sample losses from accidental contact of the filter surface with
another sufrace is  minimized.  An additional advantage is that the sulfate is
converted to the same form as the calibration standard.

      Figure 5 shows a simple schematic of an ammoniation set-up.  Filter
samples on open Petri dishes are placed face up on perforated shelves of the
ammoniation box.  The box is evacuated, valve V^ is closed, and valve V2 is
opened.  Ammonia from concentrated ammonium hydroxide fills the box and con-
verts sulfuric acid to ammonium sulfate. One hour exposure to ammonia vapor
is adequate.  V2 is closed, most of the ammonia is pumped out and passed
through a KI^PO^ scrubber column.  Vacuum is released by switching V^ to vent.

8.5   Extraction of Soluble Sulfates

      It is important that the water/IPA ratio in the sample and in the mobile
phase be the same.   A sample richer in water content than the mobile phase
will monentarily increase the solubility of barium chloranilate and will pro-
duce a positive peak above a flat background; that richer in IPA will produce
a negative peak. Variations will occur if the solvent used in the preparation
of the standards and in the extraction of the filter samples were taken from
a different stock as that used in the mobile phase.  Therefore, it is strongly
recommended that the extraction solvent be taken from the same stock solution
as the mobile phase or, if possible, directly from the liquid reservoir of the
HPLC pump.  The use of solid standards as prepared in 8.3.3 will eliminate
variability due to  IPA/water mismatch.
                                     452

-------
8.5.1  From Fluorocarbon Membrane Filters

       Place filter  in  appropriate size polypropylene bottle,  if approximate
       sulfate level is known from previous analysis of similar samples, mea-
       sure adequate volume of 60% IPA to give sulfate concentration of about
       10 ygs/mfi,.  Otherwise add 10 m£ 60% IPA and cap the bottle tightly.
       As a rule of  thumb, the sulfates constitute about 40 - 50% of the total
       particulate mass if most of the particulates on the filter is sulfuric
       acid.  Such cases are generally encountered with filter samples  from
       catalyst equipped cars run on non-leaded fuels.  Particulate mass
       loading of a  filter is an important piece of information in deciding
       appropriate volumes of extracting solution to use.

       Shake the bottle vigorously until the filter callapses and is complete-
       ly immersed in the solvent.  A vortex test tube mixer is recommended.
       Twenty seconds shaking with a vortex test tube mixer is adequate to
       leach the soluble sulfates from the filter.  If there are  no visible
       suspended particulates, the clear solution can be used directly for
       analysis.  If suspended carbon or other particles are apparent, filter
       about 5 m£ of the extract by using a syringe with a y pore size fluo-
       rocarbon in-line filter.  These filter syringes are available  commer-
       cially.  If the  particles are sufficiently large, they can be  removed
       from the bulk solution by centrifugation.

 8.5.2  From glass fiber filters

       This procedure is for the extraction of 47 mm diameter filters.  Ad-
       just the volume  of the extracting solvent accordingly for  different
       size filters.

       Place the  filter in appropriate size polypropylene  bottle.   If approxi-
       mate SO4=  level  is known, add adequate volume of 60% IPA to  give sul-
       fate concentration of about 10 ygs/mJl.  Otherwise,  add 30  m£ 60% IPA
       and  cap bottle tightly.  Shake with a vortex mixer  until the  glass
       fibers disintegrate.  Where large solvent volume and inordinately long
       shaking time  are required, stirring with a Teflon clad magnetic stir-
       ring bar is preferable.  Glass fibers are easily separated from the
       bulk liquid by using a centrifuge.  If finely suspended particles are
       present, as  in the case of diesel and some non-catalyst exhaust samples,
       syringe filtration as mentioned in 8.5.1 must be used.  In some cases
       a 0.2 y pore  size filter is necessary.

 8.6   Water extracts

      Water extracts of filter samples may also be analyzed b^he automated
 BCA method.  One  important requirement, however, is that the •*^a*™*£
 made up to 60% IPA   (e.g., 4 parts of the extract must be a^ ^^fSis
 pure IPA, v/v) before the sample can be analyzed,  it tn
 mixed directly with pure IPA, volume shrinkage  (about 2.7%) must be tax
 into account in the calculation of concentration.
                                      453

-------
      Another approach is to evaporate completely the solvent in a known vol-
ume of the extract in polypropylene bottle similar to the preparation of so-
lid calibration standard in 8.3.3.  The residue may be ammoniated in the
plastic bottle as in 8.4 and extracted as in 8.5-1 or 8.5.2.

8.7   Analysis

      Set instrument in operating mode, remove sample probe from the holder
and dip in 100 m£ 60% IPA.  Adjust the flow rate in 3 ml/win, allow the in-
strument to cycle several times until a stable background is obtained, then
remount sample probe to holder.  Adjust loading time or peristaltic pump rate
so that at least twice the volume of the sample as the volume of the sample
injection loop has exited the waste port of the sample injection valve.  Ad-
just sample injection time so that peaks from successive sample injections do
not overlap.  Fill sample cuvetted with samples and rinse solution (60% IPA)
and place according to a sampling pattern, blank, sample, blank, sample,
blank—  The rinse solution (blank)  is necessary to ascertain that there is
no memory effect from previous sample injection.

      A series of at least 6 calibration standards spanning the concentration
range of interest is run before samples are run.  A control standard may be
placed every 10 samples as a quality check on the stability of the system.
Dilution may be necessary if the sample peak height is beyond the range of
the calibration standard.  If a large number of samples needs dilution, it
may be more convenient to merely change the size of the sample loop.

      Plot peak height (or area if an integrator is also used) vs concentra-
tion in ygs/m£ of the sulfate standard.  The curve is not linear.  Alterna-
tively, the peak height or area concentration data may be fitted into a poly-
nomial of the form:
                            234
           y = a  + a  + ax  + ax  + ax  + ...


           where:  y = sulfate concentration in ygs/mJl

                   x = peak height or area

8.8   Calculations

      Calculate the concentration of sulfate as ygs SO^/mH using the calibra-
tion curve or the polynomial regression equation.  Total soluble sulfate
(SO4=)   in the filter is then given by:

      (S04=)F = (S04=)d x VQ x d

     where:   V  = total volume in m£ of original sample extract

                 = sulfate concentration of the diluted sample in Ugs/mil
                                     454

-------
                d = dilution factor




                  = 1, if there is no dilution of the extract before analysis
               V  = aliquot volume in mi of sample diluted with 60% IPA
                3.



               V, = final volume in m£ of the aliquot sample after dilution

                    with 60% IPA
Example :
      Suppose 10 m£ of 60% IPA was used to extract the soluble sulfated in

 the filter and that 2 m£ of this was further diluted with 4 mi 60% IPA to

 bring detector response within calibration range. Suppose that the concen

 tration of the diluted sample was found to be 5 ugs/m£ .  Then,



                (S04=)d = 5 ygs/m£



                     V  = 10 m£
                      o


                     V  = 2 m£
                      a


                     V, = 2 + 4
                      d


                        = 6 m£



                 (S04=)p = 5 x 10 x  (6/2)



                        = 150 ygs
                                      455

-------
    Figure 1.  Chromatogram at trace sulfate levels,  0.01,  0.02,  0.05,
0.01 ygs SC>4=/in£.  Flow rate at 3.2 m£/min.  Detector sensitivity at 0.01
      absorbance \anits full scale.  Sample volume  injected  = 0.5  m£
                                    456

-------
-f=>
01
      Figure 2.  Reproducibility of repetitive sample injections.  Flow at 3.2 m£/min.  Detector sensitivity

      at 0.5 AUFS.  Sample volume injected = 0.5 m£.  Numbers above peaks are sulfate concentrations in  |jgs/m£.

-------
43.
CJ1
00
         B  - buret
         LR - liquid reservoir
         LP - PHLC pump
         P  - Pressure Monitor
         SV - Sample Injection valve
         L  - Sample loop
                                                                                            Integrator
Position A
  Inject
                      CX-Ag+ - Cation exchange resin, silver form
                      CX-H+  - Ca,tion exchange resin, hydrogen  form
                      BCA    - Barium chloranilate
                      D      - UV detector
                      FM     - Flow monitor
                      PP     - Peristaltic pump
                      AS     - Automated sampler
                           Figure  3.   Schematic of  an  automated BCA sulfate set-up.

-------
                            •1/4" union-
                          5 micron Stainless
                            Steel Frit
                         1/4" to 1/16" Reducer
Figure 4.   Configuration for packing column.
                     459

-------
                                                             V-J
VACUUM «
                  *KH2P04'
                                      Perforated
                                        Shelf
                                       Vent
V,
                                       Filter on
                                    X* Petri dish
                                                                           Concentrated
                                                                              NH4OH
                          Figure 5.  Schematic pf an aitimoniation set-up.

-------
     DISCREPANCIES BETWEEN EPA AND DEPARTMENT OF EMISSIONS RESEARCH
                           BCA SULFATE PROCEDURE
4.1.1.1 - Bottled N2 gas is  used to drive 60% IPA from a reservior through
          the HPLC system  at average rate of 4.5 m£/min0

4.1.1.2 - The Dupont 837 sample  cell volume is 6.3 yJl.

4.1.1.3 - A dedicated 1 m£ sample loop is used on a manually-operated sample
          injection valve.

4.1.1.4 - Recorder is dual channel, 10 mv full scale, with variable chart
          speed set at 0.75  in/min.  Only one recorder  channel us used since
          chromatogram sensitivity is controlled at range switch on spectro-
          photometer.

4.1.1.6 - A 6-inch long cation column is used.

4.1.1.7 - A 2-1/2 inch long  BCA  column is used.

4.1.1.10 - One-quarter inch unions are used.

4.1.2   - Non-automated system is used.

6.5     - Polypropylene bottles  (30 mJl)  are  used for extraction of filters
          in 60% IPA solution.

6.8     - We don't use centrifuge.

7.2     - We use deionized water checked with AgNO3 for dissolved salts.

7.3     - We store in a 5-gallon stoppered glass bottle with dispenser.

7.7     - Ours is ACS reagent, oven dried, and stored over silica gel  in  a
          dessicator.

8.3.1   - Stock and calibration  standards are stored in glass bottles.

8.3.2   - standard solution  stored in glass  bottles.

8.3.3   - We don't use alternative  method of preparing calibration standards.
                                    461

-------
8.4     - The chamber containing sample  filters  in open Petri dishes is
          purged with ammonia from concentrated  ammonium hydroxide which
          is then vented out via hood.   The chamber is  purged for 3 minutes
          and then the filters are isolated in the ammonia-filled chamber
          for one hour before their extraction in 60% IPA.

8.5.2   - We do not use glass filters  for sulfate analysis.

8.6     - We do not do water extracts.

8.7     - Used in automated system only.
                                    462

-------
                            VALIDATION WORK AT SwRI
1.   Diesel fuel gives a positive interference

     a.   1 yg diesel fuel  gives response equal to 0.02  yg  SO 2~.
     b.   Response to diesel fuel occurs with or without BCA column in
          system indicating a nonsulfate interference in the fuel.
     c.   Diesel fuel response can be removed by washing diesel fuel doped
          filter with 25 m£ cyclohexane.
     d.   Sulfate response  unchanged by washing sulfate  doped filter with
          25 m£ cyclohexane.

2.   5-20 percent of the sulfate response on some filters collected from
     diesel fuel engines was not due to sulfate.

     a.   Filters washed with 25 m£ cyclohexane were  13-20 percent lower
          in apparent sulfate response than unwashed  filters.  These filters
          were duplicates collected from diesel powered  vehicles.  Cyclo-
          hexane does not remove sulfate from sulfate doped filters.  Cyclo-
          hexane does remove diesel interference from diesel doped filters.
     b.   Diesel filter samples run without BCA column in analysis system
          were 5-18 percent lower in apparent sulfate response than iden-
          tical samples run with the BCA column in place.*
* System designed to run with and without BCA column  in system was similar
  to the system described by N.  J. Khatri, J.  H.  Johnson and D. G. I^ddy
  in "The Characterization  of the Hydrocarbon and Sulfate Fractions of
  Diesel Particulate Matter,"  SAE. Paper No.  780111,  February-March, 1978.
  In this work they found a large portion of  the  sulfate due to what they
  attributed to hydrocarbon interferences.
                                     463

-------
      APPENDIX L





DMNA SAMPLING PROCEDURE
          464

-------
                DESIGN  AND CALIBRATION OF SECONDARY DILUTER


     A secondary diluter,  to be used in conjuction with a constant volume
sampler, has been constructed.   The secondary diluter was constructed:

     1.   to determine  the concentration of N-nitrosodimethylamine in auto-
          mobile exhaust

     2.   to determine  the composition of condensable organic vapors in auto-
          mobile exhaust.   Condensable organic vapors are those having mole-
          cular weights from about 140 to 300 atomic mass units.

     Factors which had  .to  be taken into consideration for construction of the
secondary diluter were:

     1.   a proportional sample of automobile exhaust is  needed

     2.   high levels of nitric oxide must be avoided because:

          a.   NO + amine  •* nitrosamine artifact,  which would interfere with
               the detection of N-nitrosodimethylamine.

          b.   the oxidation of NO to NO2 with materials  in the system and
               on the traps would cause interference.

     3.   the sample must  be presented to the Tenax GC trap at an appropri-
          ate flow rate and temperature so that collection is efficient and
          materials are not driven off the trap.   (We  sample at 1.6  £/min
          and 60°C) .

     The design criterion  was to avoid oxidation of NOX and have a sample of
5 ppm or less of NOX, which was accomplished by using  dry nitrogen to dilute
the automobile exhaust  at  25:1.  (Our dilution is  16.7:1.)

     The secondary diluter is shown schematically  in Figure 1 and its oper-
ating principels given  below:

     1.   diluted automobile exhaust is pulled into the secondary diluter
          through a 1/8" OD stainless steel tube at a  flow rate of approxi-
          mately 2.5 Vmin;  this flow is dependent upon the downstream pres-
          sure drop.
          Dry nitrogen  for  secondary  dilution  is add ed through a^oot j .meter
          and rotometer, in series, at  approximately 68 £/tain.  This "~ »
          independent of the downstream pressure drop.  (We run at 55 l/«an.)

                                    465

-------
                                                              Vacuum
                                                               Gage
CT>
                            ^
          mm
               Roots
              Blower

denotes line heated to 60°C
                                                Dilution
                                               Flowmeter
                                                                                                   Regulating
                                                                                                    Valve
d
                                                                                   Excess
                                                                                   Vacuum
                                                                                                  Tenax Trap
                                                                                                 Sample
                                                                                                 Flowmeter
Regulating
Valve
                                                                          Vacuum
                                         Figure 1.  Secondary diluter

-------
      3.   Diluted automobile exhaust and
           and pulled through a 3/8" 01
      4.   A magnehelic gauge is used to monitor the pressure drop across  the
           diluter   The reference side of the magnehelic is connected to  Se
           primary dilution tube; this allows dilution ratios to be determined
           without making corrections for differences in atmosoheric and™
           mary  dilution tube pressure.                                 P

      5.   Three  simultaneous samples are pulled through; (a)  glass fiber  fil-
           ters,  (b)  Tenax GC traps, (c)  rotometers, and (d)  needle valves
           The needle valves are set to maintain a flow of 2.5 Jl/min through
           each  sample trap.  (We use Fluoropore filters and sample at 1 6
           £/min . )

      6.   Excess  flow is discharged from the diluter through  a needle valve,
           which  is used to maintain the proper pressure drop  across  the
           diluter.

      During the  operation, it is suggested that the inlet rotometer  (2)  be
maintained at 80  and  the magnehelic gauge,  (4)  be maintained at 10" water,
which will give dilution ratio of 25:1.   These  gauges should be checked every
five minutes to be sure that these flows  do not change.   (Our rotometer is
set at 65 and the  magnehelic at 25" water for a dilution ratio of 16.7:1.)

CALIBRATION

      The rotometers  for the secondary  diluter  were  calibrated by using a
roots meter and dry test meter.

      The secondary diluter was  calibrated  by several different, but similar
experiments.  The  purpose of these experiments was to calibrate the dilution
ratio against diluter pressure drop.  Each  experiment involved measuring  di-
lution ratios obtained at various constant  diluter, pressure drops as meas-
ured by the magnehelic gauge.

      It was found early in the  experiments that a constant source of vacuum
is needed in order to maintain a steady pressure drop across the diluter.

      The experiments were conducted  as follows:

      1.   A sample of propane span gas was placed in a large empty Tedlar
           bag; the concentration of  propane  in the bag was determined by
           analyzing  the contents of  the bag  on a hydrocarbon analyzer.

           nn* c,amDie rotometer   (5)  was disconnected from its normal vacuum
           SrStS r™aV a'aiapfcra.  pu*p of sufficient size to p»U
           2 5 i/min    The reference  side of  the magnehelic was left open to
                      opelatio    V bag of propane span g,s »as connect


                                     467

-------
           to the sample inlet (3) .  The diluted sample was then collected in
           an emply Tedlar bag downstream of the diaphragm pump.  This bag
           was then analyzed on a hydrocarbon analyzer and its concentration
           determined.  By dividing the concentration of hydrocarbon in the
           span gas by the concentration found in the bag of diluted span gas,
           the dilution ratios were found for different magnehelic gauge
           readings .

      2.   The diluter was then set up as it normally will be used, except
           that one rotometer (5) was still connected to the diaphragm pump.
           A cylinder of span gas containing propane, carbon monoxide and
           carbon dioxide was then metered into the primary dilution tube.  A
           sample of this diluter span gas was collected in bags on the pri-
           mary diluter and at the same time bags were collected on the se-
           condary diluter.  The bags collected on both primary and secondary
           diluters were then analyzed on appropriate analyzers.  By dividing
           the concentration found in the primary bag by the concentration
           found in the bag of diluted span gas, the dilution ratios were
           found for different magnehelic gauge readings.

      3.   The diluter was set up as it would be for making a normal test,
           with the exception that a rotometer was attached to the sample
           inlet (1) .  Ambient air was pulled through this rotometer while
           the rest of the diluter was under normal operation.  By making
           corrections for pressure drop in the diluter, the flow of dry
           nitrogen into the diluter can be divided by the flow of ambient
           air into the diluter, giving dilution ratios for different magneh-
           lie gauge readings.  The expression used for correcting the pres-
           sure drop is PiVi = V2 where :
                P, = initial pressure
                V1 = initial volume
                PP = final pressure
                V  = final volume

      A correction for temperature was not made because the temperature was
stable at approximately 20°C and thought this difference would be of little
consequence .

      The dilution ratios determined by these three different experiments were
plotted against differential pressures measured by the magnehlic gauge using
the general expression for flow in pipe:


                    (ref ' 1J

where:  AP  = differential pressure
        f '   = friction
        L/D = length/diameter ratio
        g   = gravitational constant
        v   = specific volume of gas, I/density
        V   = linear velocity
                                    468

-------
      Since the  sample  inlet is of constant cross section,  the  linear velocity
is proportional  to  the  volumetric flow rate,  m any given  set  of experiments*
all variables in this equation other than velocitv anH ^
Because of this,  the equation reduced to: Vei°Clty and Pressure are constant.
             2      2
      AP = CV  = CQ
where O
          sample  flow rate
          - * — ; - - — : — -
             volumetric
      The dilution  ratio,  R, is given by:
      R •
                 °
 N2
-Q-
                                   N2
                                         Constant
 Therefore:  I = c..
            R

 and the plot  of the reciprocal of the dilution ratio against "YAP should be a
 straight  line.

      Figure  2 presents all the raw data gathered in these calibrations and
 plotted on  this basis.   A linear least squares fit for these data is also
 presented in  Figure 2.   A single equation was found to adequately represent
 the shole date base :

      1/R = 0.01858 1^ 0.02133

      r2  = 0.9819

      standard error of estimate = 3 percent of mean value

 Reference 1.   Mark's Standard Handbook for Mechanical Engineers, 7th Edition,
               McGraw-Hill, N.Y. (1967), p. 4-66.
                                      469

-------
O
                                                          Tunnel Propane Injection
                                                       n
                                                          DP " °-02030   /"SF0.0263
                                                                       O  Carbon Monoxj.de  Injection
                        0.030
                        0.020
                        0.016
                           2.0
                                                                                                       4.0 4.1
                          Figure  2,  Calibration of N2 secondary dilute*, February 10,  1976
                                                   (least square lines).

-------
TABLE 1.  N2  SECONDARY DILUTER CALIBRATION VALUES
                 FEBRUARY 10, 1976

TAP
2.0
7 O^fi
2 5495
9 TZRf,
2090/1
0 1 ftOT
j . J.O^ -J
•3 -31 CC
a • jXOO
3c; •3^'?
.D ODD
3*7 yl 1 "7
.7417
3Q"79Q Q
.0 /^yo
4-1 «-» *5 1 -i
.12311
A 0
B -0
R2 0
(2) 0
(4) 0

Carbon Monoxide
0.01712
0 02500


On^ARn


On AC. c: i
.U4D3-L

l/DF - A (fAP~)
.01709 0.02030
.01765 -0.02630
.99749 0.99414
.01653 0.01430
.05071 0.05491
l/DF
Propane
OA 1 QAO
.QlyQo

On "3 1 n^.
• UoJ. /D
On OQ 0*7
. U Jy J /
OARI i n
.UDJ.XU
Onc*7m
• UD /UX
+ B
0.01684
-0.01622
0.97560
0.01747
0.05115

Rotometer
Of\ f\ 1 ^ f
.02116
On oAm
.0300J.
OAOQQO

Ondnfi1?

n nsm ^

Overall Data
0.01858
-0.02133
0.0818
0.01584
0.05301
                         471

-------
               PROCEDURE USED BY SwRI FOR OBTAINING DMNA TRAPS
PREPARATION OF RUNNING

      1.   Turn on 4 rheostats in back of DMNA cart at least 30 minutes prior
           to test.

      2.   Remove trap holder with the filter case from the cart.  Insert fil-
           ter (FA) into filter case so that flow goes through the dull side
           of the filter (note direction of arrow).  Replace o-ring. The fil-
           ter and the o-ring are not to be handled with bare hands.  Fasten
           trap holder back onto cart and tighten nut.

      3.   Check nitrogen (N2) bottle to make sure it has at leaat 1000 psi.
           If necessary, N2 flow can be set with a dummy filter with the pump
           on. Check dryer on N2 bottle and replace molecular sieve and silica
           gel dryer if necessary.

      4.   Open the green valve on the sample line.

READY TO RUN

      1.   Obtain trap from refrigerator.  Leave it in the glass tube un*til
           ready to use.  With disposable gloves, place trap in holder.  Hand
           tighten both end fittings.  Replace spun glass ball in glass con-
           tainer and recap.

      2.   Just prior to start of test, turn on N2 flow.  At honk of horn,
           turn on sample pump and set the large N2 flowmeter with the top of
           the float at 65 and the magnehelic at 25" H20.  Both knobs need to
           be adjusted simultaneously since they are dependent on each other.
           These two will have to be monitored throughout the test.  Flow-
           meters 1, 2, and 3 should already be adjusted to their proper
           settings (10, 5.2, 10 respectively).  If not, use the large black
           knobs to readjust.  Temperature of the system is monitored at four
           points: sample inlet from CVS, N2 inlet, mixing point and at trap.
           It should be set at 60°C (140°F) by adjusting rheostats in back of
           cart.  These temperatures are recorded on the data sheet about
           once a week.

END OF TEST

      The traps are run the full cold FTP and then the hot 505.  During the
soak between FTP's, the pump and the N2 bottle are shut off.


                                    472

-------
1.   When the test is over,  the trap is again carefully handled with
     gloves  and placed back  in the glass container.  The container is
     labeled on the cap and on the side of the trap number  (from log
     book) ,  test number and run date.  The container is then placed in
     the  1-gallon can in the refrigerator.  Test number and run date are
     recorded  in the log book.  Test number is also recorded on the data
     sheet (same one temperatures were recorded on) .

 2.   Close green flow valve on sample line.

 3.   Turn off N2 regulator.

 4.   Turn rheostats  off.
                                   473

-------
      APPENDIX M




DMNA ANALYSIS PROCEDURE
          474

-------
THE MEASUREMENT OF DMNA  IN EXHAUST
            As  used by

    Research  Triangle  Institute
   Research Triangle Park,  N.C.
            Developed by

     Research Triangle Institute
    Research Triangle Park,  N.C.
            November 1978
                 475

-------
                         RESEARCH  TRIANGLE  INSTITUTE
     ANALYSIS FOR N-NITROSODIMETHYLAMINE IN EXHAUST GASES USING A TENAX
        GS CARTRIDGE AND GAS CHROMATOGRAPHY/MASS SPECTROMETRY/COMPUTER
EPA Contract No. 68-02-2767

RTI/1514/00-01S
                     Special Interim Technical Report

                                    by


                    E. D.  Pellizzari, Project Director
                         Date:   November 27,  1978
                              Project Officer
                                Ron Bradow
                  Mobile Source Emissions Reserach Branch
                        Mail  Drop  46, ERG Annex
                   U.S. Environmental Protection Agency
                   Research Triangle Park, N,  C.  27709
   Prepared for the Environmental Protection Agency,  Reserach Triangle
   Park, N. C.   27711
          RESEARCH  TRIANGLE  PARK,  NORTH CAROLINA  27709

                                   476

-------
      The information presented in this document is subject to the following
qualifications:

      (a)  The mention of  a specific company does not imply the  intent to
           regulate that company or its activities nor that, unless specifi-
           cally stated, the company is the source of a given compound;

      (b)  The  identification of compounds were determined by mass spectro-
           metric and retention index techniques and their identity are sub-
           ject to the limits of this methodology.

      (c)  The  mention of compounds in this report does not imply that they
           are  necessarily carcinogenic or mutagenic;

       (d)  The  possible mutagenic or carcinogenic activity attributed to a
           compound is based upon cited literature;

  and  (e)  The experimental findings and conclusion presented in this report
           should not be  cited, reproduced, or included in other publications
           without the expressed approval of the Project Director or Officer.
                                       477

-------
1.0   Introduction

      Because of the previous reports of the presence of N-Nitrosodimethyla-
mine in ambient air, interest in the determination of DMN in auto exhaust
heightened.  N-Nitrosodimethylamine (DMN)  levels in ambient air were deter-
mined for an area surrounding an industrial site in Baltimore, Maryland.
Using Tenax GC cartridge for concentrating DMN and glass capillary/glass-
liquid chromatography/mass spectrometery with specific ion (m/z 74)  monitoring,
DMN was detected and quantified.  On an industrial site DMN levels reached
32,000 ng/cm (10.67 parts per billion)  in the ambient air.

      On the basis of these  and other observations, research was conducted on
determineing whether N-nitrosodimethylamine was present in auto exhaust.  In
conduction with Southwest Research Institute, a study was conducted on the
detection of DMN in auto exhaust of automobiles which have been operated
under various test conditions.  SwRI was responsible for generating and par-
ticipating in the collection of the auto exhaust samples during the course of
this program.  The results of this research effort is described here.
                                     478

-------
2.0   Experimental  Procedures

      The collection and analysis techniques given in Appendix A were modi-
fied and used  for detecting DMN in auto exhaust gases.

      Tenax  GC sampling cartridges were prepared at the Research Triangle
Institute  (RTI)  and shipped by Federal Express to Southwest Research Institute
(SwRI) for the collection of auto exhaust samples and subsequent returned to
RTI for analysis.  In all cases the sampling cartridges were expended or
returned within four to five weeks and replaced with a fresh batch in order
to insure  a  low background level.  All samples were analyzed within two to
three weeks  after sample collection was completed.

      The  sampling procedure employed by SwRI consisted of a primary and
secondary  diluter and the secondary diluter and transfer lines were maintained
at 60°C.   The  sampling rate was 1.6 H/min and the test length was approxi-
mately 31  minutes giving a total sample volume of 50 £.  The average primary
cvs dilution was 11.1 to 1 and the secondary dilution was 16.7 to 1 giving
an actual  sample dilution of 185 to 1.  The secondary diluter was necessary
in order to  insure that the NOX and hydrocarbon values were at levels which
minimized  the  potential artifact formation of the N-nitrosodimethylamine on
the Tenax  GC sorbent.
                                      479

-------
                                 (APPENDIX A)

N-NITROSODIMETHYLAMINE IN AMBIENT AIR
ANALYTICAL METHOD

Analyte:      DMN                        Method No:

Matrix:       Air                        Range:           P.5 ppt - 10 ppb

Procedure:    Adsorption on Tenax GC,    Precision:       ±10%
               thermal desorption
               with He purge, measure-
               ment by capillary gas-
               liquid chromatography/
               mass spectrometry

Date Issued:                             Classification:  E  (Proposed)

Date Revised:


 1.     Principle  of Method

       N-nitrosodimetnylamine  (DMN)  is concentrated from ambient air on Tenax
 GC  in  a short  glass tube  (1,2) .  It is desorbed by heating and  purging with
 helium into  a  liquid nitrogen  cooled nickel  capillary trap and  then intro-
 duced  onto a high  resolution gas chromatographic  column where is is separated
 from interferences.  The  concentration of DMN is  measured from  the mass
 spectrometric  signal at m/e 74 (3).

 2.     Range  and  Sensitivity

  2.1  The range  of the mass spectrometric signal for the conditions listed
 corresponds  to 0.5 ppt to 10 ppb.

  2.2  A concentration of  0.5 ppt of DMN can  be determined in a  150-liter air
 sample.

 3.     interferences

       Interferences may result from materials having background ions  of  m/e_
 74  (C2H8N3,  C2H4N02, C2H6N20,  C3H3C1, C-jHgS, C3H6O2, or C3H1C)N2) , if  at  the"
same retention time of DMN.

4.     Precision  and Accuracy

  4.1  The precision of this method  has been  determined  to be ±10% or  relative
standard  deviation when replicate sampling cartridges were spiked with 50 ng
 (corresponding to  10 ppb  in 150 £ of air).   These data  were  obtained  using
10.0 cm long glass tubes  (1.5  cm i.d.)packed with 35/60 mesh of Tenax GC
 (bed dimensions: 1.5 cm x 6 cm in depth).
                                      480

-------
  4 2 The accuracy of the analysis is approximately ±10% of the amount re-
ported as determined from repeated analysis of several standards.

5.    Advantages and Disadvantages of the Method

  5.1 The gas  chromatography-mass spectrometry technique interfaced with a
Finnigan glass jet separator  (Model 01512-42158 Finnigan Corp., Sunnyvale
CA) is extremely sensitive and specific  for the analysis of DMN.   The high
resolution  gas chromatographic separation yields a retention time that  is
characteristic for DMN, and relatively specific for positive assignment of
the signal  as  DMN.  The mass spectrometer in combination with high resolution
gas chromatography yields a very high degree of specificity.  The base peak
of DMN is at m/e_ 74 which is also the parent ion.  In order to assign the
signal at m/e_ 74 to DMN it is absolutely necessary that the retention time
matches with the signal .

  5.2 Collected samples can be stored up to 1 month with less than 10% losses.

  5.3 Because  DMN is a suspected carcinogen in man it is extremely important
to exercise safety precautions in the preparation and disposal of liquid and
gas standards, cleaning of used glassware, etc., and the analysis of air
samples .

  5.4 Since the mass spectrometer can not be conveniently mobilized sampling
must be carried out away from the instrument.

  5.5 High  resolution gas chromatography /low resolution mass spectrometry is
not a convenient technique for handling  a large number of samples (>100/wk) .

  5.6 Efficiency of air sampling increases as the ambient air temperature
decreases  (i_-e_. sensitivity increases) .

  5.7 Ambient  air sampling is limited to cases where the NOX levels are less
than 3 ppm when dimethylamine is also present.

6.    Apparatus

  6.1 Sampling Tubes
                                                                        -' -
                                                                caps, and
 x 150 cm)  culture tubes, immediately sealed using
 cooled.
                                •4-v,  lonaer beds of sorbent may be prepared using
     6.1.2  Cartridge samplers with  longer oeas uj.
 a proportional amount of Tenax GC.
                                      481

-------
  6.2 Gas Chromatographic Column

    6.2.1 A 0.35 mm i.d. x 50 m glass SCOT capillary coated with DECS  station-
ary phase and 0.1% benzyl triphenylphosphonium chloride is used.  The  capil-
lary column is conditioned (detector end disconnected) for 48 hr at 210°C  @
1.5-2.0 m£/min helium flow.

  6.3 A Finnigan type glass jet separator on a magnetic or quadropole  instru-
ment is used at 200°C.

  6.4 Inlet-Manifold

    6.4.1 An inlet-manifold is fabricated and employed (Fiugre 1, ref, 1,2,4).

  6.5 Gas Chromatograph

    6.5.1 A Varian 1700 gas chromatograph or equivalent.  A gas chromatograph
employing a single column oven and a temperature programmer is adequate.

  6.6 Mass Spectrometer

    6.6.1 A mass spectrometer with a resolution of 500-2000 equipped with
single ion monitoring capabilities must be used in conjunction with a  gas
chromatograph.  A Varian-MAT CH-7 has been found to be satisfactory for this
purpose  (2,3).

  6.7 Syringes

    6.7.1 Syringes, l-m£ gas tight (Precision Sampling, Inc.) and 10 jl£
(The Hamilton Co., Inc.).

7.    Reagents and Materials

      All reagents must be analytical reagent grade.

  7.1 N-nitrosodimethylamine

  7.2 Acetone

  7.3 Isoclean®

  7.4 Tenax GC (35/60 mesh, Applied Science)

  7.5 Two 2-liter round bottom flasks fitted with injection ports.

  7.6 Soxhlet apparatus

8.    Procedure

  8.1 Cleaning of glassware


                                     482

-------
        GAS
       METER
                         FLOW
                         METER
           H3-
          NEEDLE
           VALVE

PUMP





CARTRIDGF 1

                                                                  GLASS
                                                                  FIBER
                                                                  FILTER
                         VAPOPs COLLECTION SYSTEM
                                                          PURGE
                                                           GAS
       ION
     CURRENT
    RECORDER
     MASS
    SPECTRO-
     METER
                  GLASS
                   JET
                SEPARATOR
                                 TWO
                                POSITION
                                 VALVE
                                                   THERMAL
                                                   DESORPTION
                                                   CHAMBER
                                                                \
                                                      CAPILLARY
                                                        TRAP
                                                      HEATED
                                                      BLOCKS
                                                              EXHAUST
                          ANALYTICAL SYSTEM
Figure 1.
Vapor collection and analytical  systems for analysis of
       hazardous  vapors in ambient  air.
                                  483

-------
      All glassware, glass sampling tubes, cartridge holders, etc. should be
washed in Isoclean ©/water, rinsed with doubly distilled water and acetone
and air dried.  Glassware is heated to 450°F for 2 hrs.

  8.2 Preparation of Tenax GC

    8.2.1 Virgin Tenax GC is extracted in a Soxhlet apparatus overnight with
acetone prior to its use.

  8.3 Collection of DMN in Ambient Air ,,

    8.3.1 Continuous sampling of ambient air may be accomplished using a
Nutech Model 221-A portable sampler (Nutec Corp., Durham, NC) or its equi-
valent (2).  Flow rates are adjusted with a metering valve through a cali-
brated rotameter.  Total flow is registered by a dry gas meter.

    8.3.2 For larger sample sizes it is important to realize that a larger
total volume of air may cause elution of DMN through the sampling tube.  It
has been demonstrated that exceeding a total of 385, 332, 280, 242, 224, 204,
163, 156, 148, 127, 107, 93, or 79-liters of air at temperatures of 50, 55,
60, 65, 70, 75, 80, 85, 90, 95, 100, 105 or 110°F, respectively will result
in elution of DMN from the cartridge sampler.  A flow of 10 cc/min to 30 £/
min may be used with the sampler described in 6.1.

    8.3.3 DMN has been found to be stable and quantitatively recoverable from
cartridge samplers after 4 weeks when tightly closed in cartridge holders,
protected from light and stored at 0°C.

  8.4 Analysis of Sample

    8.4.1 Instrument Conditions and Set-up.  The thermal desorption chamber
and six-port valve are set to 200°C.  fhe glass jet separator is maintained
at 200°C.  The mass spectrometer is set to monitor m/e_ 74 (Figure 2).

    8.4.2 Adjust the He purge through the desorption chamber to 50 mJl/min.
Cool  the Ni capillary trap at the inlet manifold with liquid nitrogen.

    8.4.3 Place the cartridge sampler in the desorption chamber and desorb
for 5 min.

    8.4.4 Rotate the six-port valve on the inlet-manifold to position "B",
heat the Ni capillary trap to 180°C with a wax bath.

    8.4.5 Temperature program the glass capillary column from 75 to 205°C at
4°C/min and hold at upper limit for 10 min.  The retention time of DMN is
approximately 26 min (Figure 3).

    8.4.6 The analytical column is cooled to ambient temperature and the
next sample is processed.
                                     484

-------
uu
90
80

70
60
50
40
30
20
10
0
.
.

•
H








m/







1 ,1 J


e 42








*"/ c 1 *±









, ,
  •FIT111  'I'l'I'P^I'I'I'l1  'I11
  10 20 30 40   60 70  80 90    110 120 130 140  160 170

0           bO          100          150
             m/e-
Figure 2.  Mass spectrum of N-nitrosodimethylamine.
                  485

-------
00
CTl
                 50 -


                 40 -



                 30 -


                 20 -


                 10 .
 0

50 .



40 •



30 "


20 .



10 -
                  0
                 50 .



                 40 -



                 30 '



                 20



                 10
                  0
                      x 3
                      x 3
                      x 3
                                                                         DMN (300 ng)
                                                                                          26 min.
                                                                                     xl
                                                                                          26 min.
                                                               xl
                                                                                          26 min.
                                                        TIME (MIN.)
              Figure  3.  Mass (m/e_ 74)  Chromatograms.   A = standard DMN, B,C =  replicate air  samples.

-------
9.    Calibration and Standards

  9.1 Preparation of Gas Standard
cno              tW°"2 llter rOUnd b°tt0m flask with helium, warm flasks to
50 °C with heating mantels and use magnetic bar to stir vapors.

    9.1.2 Inject 0.1-1 yJl of DMN into flask and let stir for  30 min   Make
further dilutions into second flask by transferring milliliter gas volumes
as needed.

    9.1.3 Purge air/vapor mixtures from second flask onto cartridge samplers.

  9.2 Calibration

    9.1.2 Prepare standard curve (with ten concentration points) by thermally
desorbing cartridge samplers loaded with 3 ng to 30 yJl of DMN. Plot m/e_ 74
response vs ng of DMN.  A linear response is observed.

10.   Calculations

  10.1  Hie  total quantity of DMN in ambient air is determined by comparting
m/e_ 74  response for samples of DMN with standard curve.

                                 ng DMN   24.45
                           ppb = — -    = —^~

      where :

      ng  DMN  = total ng concentration is determined in 9.2.1

      V       = volume of air in liters sampled at 25 °C and 760 torr

      24.45   = molar volume of an ideal gas at 25°C and 760 torr

      MW       = moleculat weight of DMN, 74.

11.   References
            148 pp.
                                           Analytical Techniques for Measuring
                                             Va
            1975, 187 pp.

       3.   PellaZ2ari, ... a... B««*. ..»• BerXiey ana ,.  McCrae. Bio^.ica,
            Mass Spec, submitted.
                                           v  T E  Bunch, and E.  Sawicki,
            Environ. Sci.
                                      487

-------
       APPENDIX N




BaP SAMPLING AND ANALYSIS
          488

-------
     SAMPLING AND ANALYSIS OF BaP
         Sample Collection at
     Southwest Research Institute
Analysis Method Developed and used at

   Environmental Protection Agency
     Research Triangle Park, N.C.
               June  1979
                  489

-------
                  SAMPLING AND ANALYSIS OF BENZO-CX-PYRENE
     The analysis for benzo-a-pyrene (BaP)  will be carried out by collecting
particulate samples on 8" x 10" glass fiber filters at Southwest Research
Institute and sending the filters to EPA-Research Triangle Park for analysis
by fluorscence spectroscopy.*

     Sampling for BaP has been successfully conducted in the past at South-
west Research by collecting diesel particulate on 8" x 10" glass fiber
filters.  Other filtering media and/or sizes are under consideration for
sampling and may be used in the future.  Sample flow rates will depend on
filter size and loading capacity.  The temperature of the dilute exhaust at
the sampling point will not exceed 125°F.

     After the filters have been loaded with sample they will be weighed,
folded in half so that the particulates are inside and folded again.  All
samples will be handled under yellow light as BaP is degraded in the pre-
sence of white light.  The filter will then be placed in a glassive envelope
and then in a manila envelope.  Several envelopes will be placed in a zip-
lock plastic bag purged with zero nitrogen and heat sealed.  The samples
will be stored at -30°C until shipped in an insulated container with dry
ice via air freight to EPA-RTP for analysis.

     At EPA-RTP particles on the glass fiber filters are extracted,with
methylene chloride.  A portion of the methylene chloride solution is analyzed
for BaP.  The methylene chloride solution is diluted with cyclohexane and
spotted on thin layer acetylated cellulose plates.  The plates are developed
and the fluorescence due to BaP is measured in a spectrophotometer.   The
concentration of BaP is determined by comparing the response of the sample
to the response of a series of standards.
* Private communication with EPA-RTP.
                                    490

-------
4. TITLE AND SUBTITLE
 ANALYTICAL PROCEDURES  FOR CHARACTERIZING UNREGULATED
 EMISSIONS FROM VEHICLES  USING MIDDLE-DISTILLATE FUELS
 Interim Report
 . AUTHOR(S)
 Lawrence R. Smith, Mary  E. Parness, E. Robert Fanick
 and  Harry E. Dietzmann
9. PERFORMING ORGANIZATION NAME AND ADDRESS"
 Southwest Research Institute
 8500 Culebra Road
 San  Antonio, Texas 78284
 REPORT NO.

 EPA-60Q/2-80-068
                                                           3. RECIPIENT'S ACCESSION NO.
                                                           5. REPORT DATE
                                                                 April IQftn
                                                             PERFORMING ORGANIZATION CODE
                                                           8. PERFORMING ORGANIZATION REPORT \o
                                                           10. PROGRAM ELEMENT NO.

                                                             07A1D  14-0457 (FY-80)
                                                           11. CONTRACT/GRANT NO.

                                                             68-02-2703
 2. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Sciences  Research Laboratory - RTF, NC
  Office of Research  and  Development
  U.S.  Environmental  Protection Agency
  Research Triangle Park, N.C. 27711
                                                            13. TYPE OF REPORT AND PERIOD COVERED
                                                              Interim.
                                                            14. SPONSORING AGENCY CODE

                                                             EPA/600/09
15. SUPPLEI*
           ITARY NOTES
 i e. ABS
       This research  program was initiated with the objective of developing, codifying
  and testing a group of chemical analytical methods for measuring toxic compounds in
  the exhaust of distillate-fueled engines (i.e. diesel, gas turbine, Stirling, or
  Rankin cycle powerplants).  It is a part of a larger effort to characterize these
  components from a number of prototype powerplants and, thus, represents a logical
  first step in the process.
       Methods of collection and analysis for aldehydes and ketones, for hydrogen
  cyanide and cyanogen,  for hydrogen sulfide, carbonyl sulfide and organic sulfides, for
  ammonia and amines, for nitrous oxide, sulfur dioxide, individual hydrocarbons, for
  soluble sulfate and N-nitrosodimethylamine, benzo-a-pyrene, and phenols were studied
  in detail.  Ten analytical procedures were developed and codified.  Interference
  studies and proof-tests in diesel engine exhaust were conducted with every procedure
  and the results of  these experiments are reported in detail.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
  *

  *
  * Air pollution
  * Vehicles
  * Distillates
    Fuels
    Emission
  * Collecting methods
  * Chemical analysis
18. DISTRIBUTION STATEMENT
  RELEASE TO PUBLIC
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                                              1Q SECURITY CLASS tThisReportl
                                               "UNCLASSIFIED
                                              20 SECURiT^ CLASS i This page
                                                UNCLASSIFIED
                                                                           COSATI I ield/Group
                                                                               138
                                                                               116
                                                                               148
                                                                               21D
21 NO. OF PAGES
    505
                                                                           PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION .s OBSOLETE
                                            491

-------