EPA-650/2-74-121

JULY 1974
Environmental Protection Technology Series
                   DEVELOPMENT  OF  METHOD
                          FOR  CARCINOGENIC
                            VAPOR  ANALYSIS
                  IN  AMBIENT ATMOSPHERES
                               Office of Research and Development

                              U S. Environmental Protection Agency

                                    Washington, DC 20460

-------
                              EPA-650/2-74-121
DEVELOPMENT  OF METHOD
     FOR  CARCINOGENIC
      VAPOR ANALYSIS
IN  AMBIENT  ATMOSPHERES
                by

            EdoE. Pellizzari

         Research Triangle Institute
            P.O. Box 12194
   Research Triangle Park, North Carolina 27709
          Contract No. 68-02-1228
          ROAP 21 BEG, Task 05
         Program Element No. 1AA010
      EPA Project Officer: Eugene Sawicki

       Chemistry and Physics Laboratory
     National Environmental Research Center
   Research Triangle Park, North Carolina 27711
             Prepared for

    OFFICE OF RESEARCH AND DEVELOPMENT
   U.S. ENVIRONMENTAL PROTECTION AGENCY
         WASHINGTON, D.C.  20460

               July 1974

-------
                                  ABSTRACT



      Analytical techniques and instrumentation were developed and evaluated


for the collection and analysis of carcinogenic and mutagenic vapors occurring


in ambient air.  The areas of investigation included (a) the design and

                                                                       3
testing of a cartridge sampler for concentrating trace quantities (ng/m )


of hazardous substances from air, (b) the design, fabrication and evaluation


of a thermal desorption inlet-manifold for recovering vapors trapped on an


analyte and sample transfer into an analytical system, (c) the evaluation


of thermal desorption as a technique for recovering hazardous vapors from


sorbents, (d) the development and performance of a field sampling system for


collecting trace quantities of vapors, and (e) the application of techniques


and instrumentation developed under this program to the analysis of hazardous


vapors in ambient air.
                                     ii

-------
                                CONTENTS






                                                                  Page




Abstract                                                           ii




List of Figures                                                    iv




List of Tables                                                     ix




Acknowledgments                                                    xi




Sections




I        Conclusions                                                1




II       Recommendations                                            3




III      Introduction and Background                                5




IV       Program Objectives and Experimental Rationale             13




V        Design of a Cartridge Sampler for Carcinogenic Vapors     19




VI       Gas Chromatographic Inlet-Manifold for Sample Analysis    70




VII      Thermal Desorption of Hazardous Vapors from Solid Sor-




         bents                                                     87




VIII     Design and Performance of a Field Sampler                 99




IX       Application of Developed Instrumentation and Methodo-




         logy                                                     109




X        References                                               123




XI       List of Papers Submitted for Publication                 129





XII      Appendix                                                 130
                                    iii

-------
                                 FIGURES




No.


1     Monitoring System for Hazardous Vapors in Cartridge


                                                                   22
      Effluents.


2     Elution Profile of Cartridge Effluent.  Sampling Rate

                                                                   24
      was 6 1/min.  Test Mixture II was used.


3     Elution Profile for Synthetic Air/Vapor Mixture I

                                                                   28
      from Spherical Chamber (Empty Cartridge).


4     Collection Efficiency Profile for BPL Carbon Using Test


      Mixture I.  Sampling Rate was 0.25 1/min; Sensitivity        29


      was 6.4 x 10~U AFS.


5     Collection Efficiency Profile for SAL9190 Carbon Using


                                                                   30
      Test Mixture I.


6     Calculated Pressure Differentials for 12/30 Mesh Par-


      ticles.  Cartridge Dimensions were 1.056 cm i.d. x           38


      3.0 cm in length.


7     Calculated Pressure Differentials for 35/60 Mesh Par-


      ticles.  Cartridge Dimensions were 1.056 cm i.d. x           39


      3.0 cm in length.


8     Calculated Pressure Differential for 100/120 Mesh Par-


      ticles.  Cartridge Dimensions were 1.056 cm i.d. x           40


      3.0 cm in length.


9     Comparison of Calculated and Experimental Pressure

                                                                   / O
      Differential for a Tenax GC (60/80) Cartridge.


10    Pressure Differentials (AP) for Cartridges Containing

                                                                   / O
      BPL Carbon (12/30) with a Cartridge i.d. of 0.5 cm.
                                    iv

-------
                             FIGURES CONT'D

No.                                                               Page
11    Pressure Differentials (AP) for Cartridges Containing
                                                                   44
      BPL Carbon (12/30) with a Cartridge i.d. of 1.056 cm.
12    Pressure Differentials (AP) for Cartridges Containing
      Tenax GC (35/60) with a Cartridge i.d. of 0.5 cm.
13    Pressure Differentials (AP) for Cartridges Containing
                                                                   46
      Tenax GC (35/60) with a Cartridge i.d. of 1.056 cm.
14    Pressure Differentials (AP) for Cartridges Containing
                                                                   47
      Tenax GC (60/80) with a Cartridge i.d. of 0.5 cm.
15    Pressure Differentials (AP) for Cartridges Containing
      Tenax GC (60/80) with a Cartridge i.d. of 1.056 cm.
16    Pressure Differentials (AP) for Cartridges Containing
      Chromosorb 101  (100/120) with a Cartridge i.d. of 1.056      49
      cm.
17    Effect of Particle Mesh Range on Pressure Differential
      and Relationship  to Cartridge Diameter.  BPL Carbon
      (12/30), Tenax  GC  (35/60) and Tenax GC  (60/80) are given     50
      by A, B, and C, respectively.  Packing Bed Depths were
      5 cm  (A), 3 cm  (B) , and 3 cm  (C) .
18    Effect of Particle Size on Pressure Differentials.           52
19    Pressure Differential Developed  for BPL  Carbon (12/30)
                                                                   53
      Cartridge with  an i.d. of 0.5 cm During  Vacuum Sampling.
20    Pressure Differentials for BPL Carbon (12/30) for Air
                                                                   54
      Drawn Through a 1.056 cm i.d. Cartridge.

-------
                              FIGURES CONT'D





No.                                                               Page



21    Pressure Differential Developed for Tenax GC (35/60)



      Cartridge with an i.d. of 0.5 cm During Vacuum Sampling.



22    Pressure Differential Developed for Chromosorb 101



      (60/80) Cartridge with an i.d. of 1.056 cm During Vacuum     56



      Sampling.



23    Pressure Differential Developed for Tenax GC (35/60)



      Cartridge with an i.d. of 1.056 cm During Vacuum Sampling.



24    Pressure Differential for Chromosorb 101 (60/80) for Air


                                                                   58
      Drawn Through a 1.056 cm i.d. Cartridge.



25    Thermal Desorption Inlet-Manifold.                           72



26    Thermal Desorption Chamber with Annular Space.   Sampling


                                                                   73
      Tube shown in Lower Figure.



27    Thermal Desorption Chamber.                                  74



28    Electronics Circuit Designed for Temperature Control on



      Inlet-Manifold System.



29    Differential Heating Rate in a Glass Cartridge Con-



      taining Tenax GC (60/80).                                    79



30    Temperature Rise Times in Sorbent Bed Using Annular


                                                                   81
      Spaced Chamber.



31    Comparison of Temperature Rise Times for Chambers with


                                                                   82
      and without an Annular Space.  Chamber was at 175°C.



32    Comparison of Heating Rates for Some Sorbents.   Curves



      A, B, and C correspond to PCB Carbon (12/30), Oxopro-



      pionitrile on Poracil C (80/100), respectively.  Thermal



      Desorption Chamber was Isothermal at 210°C.
                                    vi

-------
                              FIGURES CONT'D





No.



33    Heating Rate for Tenax GC at Different Isothermal


                                                                   85
      Desorption Chamber Temperatures.



34    Gas-liquid Chromatogram of Blank Porapak-Q Cartridge.        91



35    Background During Thermal Desorption of Tenax GC Car-


                                                                   93
      tridge Blank.                                                *



36    Background During Thermal Desorption of PCB Carbon


                                                                   94
      Cartridge Blank.



37    Gas-liquid Chromatogram of Synthetic Air/Vapor Mixture



      of Hazardous Substances.  Peaks A, B, C, D, and E are



      300 mg of Glycidaldehyde, Butadiene Diepoxide, N-nitro-



      sodiethylamine, 1,2-dichloroethyl ethyl ether, and ethyl



      methanesulfonate, respectively.  See test for glc



      parameters.



38    Gas-liquid Chromatogram of Vapors Desorbed from Tenax GC.



      Desorption Chamber at 225°C; see prior figure for Peak       _7



      Identity.  Background from Tenax GC is represented by



      Dashed Profile.



39    Relationship Between Flow Rate and Theoretical Power


                                                                  102
      Requirements at Various Tube Diameters and Particle Size.



40    Schematic of Universal Sampler 5-1068.                      105



41    Multiport Sampling Head.                                    107



42    Gas-liquid Chromatograph-Mass Spectrometer Computer



      (GLC-MS-COM) Outlay.                                        1U
                                  vii

-------
                             FIGURES CONT'D





No.




43    Total Ion Current Plot During Gas-Liquid Chromatography




      Mass Spectrometry of Air Sample from West Covina, CA.      114




44    Single Ion Plots of Ions Common to Aliphatic Cracking




      Series.                                                    120




45    Single Ion Plots for Ions Representative of Aromatic




      Cracking Series.                                           121




46    GLC-MS of Bis-(chloromethyl)ether.                         131




47    GLC-MS of Bis-(2-chloroethyl)ether.                        132




48    GLC-MS of g-Propiolactone                                  133




49    GLC-MS of Vinylene Carbonate.                              134




50    GLC-MS of N-diethylnitrosainine.                            135




51    GLC-MS of Nitromethane.                                    136




52    GLC-MS of Ethyl methanesulfonate.                          137




53    GLC-MS of Glycidaldehyde.                                  138




54    GLC-MS of Propylene Oxide.                                 139




55    GLC-MS of Styrene Oxide.                                   140




56    GLC-MS of Butadiene diepoxide.                             141




57    GLC-MS of Acrolein.                                        142




58    GLC-MS of Methylethylketone.                               143




59    Mass Spectrum of Maleic Anhydride.                         144




60    Mass Spectrum of Succinic Anhydride.                       145




61    Mass Spectrum of 1,3 Propanesultone.                       146




62    GLC-MS of Tetramethylene Sulfolane.                        147




63    GLC-MS of Aniline.                                         148






                                  viii

-------
                                 TABLES


No.                                                                Page

1     Commercially Available Materials with Chemically
                                                                    10
      Bonded Liquid Phases.

2     Sampling Parameters for Collecting 30 ng of Vapor
                                                                    15
      From Air.

3     Reagent Vapors for Evaluating Sampling Medium.                18

4     Synthetic Air-Vapor Mixtures for Cartridge Sampler
                                                                    25
      Evaluation.

5     Expulsion Rates from Chamber at Specified Flows.              26

6     Collection Efficiencies of Candidate Sorbents.                27

7     Approximate Collection Efficiency for Tenax-GC at
                                                                    31
      Various Sampling Rates.

8     Approximate Collection Efficiency for Chromosorb 101
                                                                    32
      at Various Sampling Rates.

9     Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    60
      1 cm Bed Depth.

10    Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    61
      1 and 3 cm Bed Depths.

11    Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    62
      2 cm Bed Depth.

12    Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    63
      3 cm Bed Depth.

13    Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    64
      2 cm Bed Depth and 12  1/min.

14    Elution Volume Characteristics for Tenax GC (35/60) -
                                                                    65
      2 cm Bed Depth and 24  1/min.
                                     ix

-------
                              TABLES CONT'D


No.                                                                Page

15    Elution Volume Characteristics for Tenax GC (35/60)  -
                                                                    66
      4 cm Bed Depth.

16    Elution Volume Characteristics for Tenax GC (35/60)  -
                                                                    67
      6 cm Bed Depth.

17    Pollutant Profile Breakthrough During Ambient Air
                                                                    69
      Sampling.

18    Percent Recovery of Vapors Adsorbed on Tenax GC
                                                                    98
      Cartridges using Thermal Desorption

19    Power Requirements to Deliver Various Sampling Rates.        103

20    Sampling Rate Characteristics for Universal Sampler  with
                                                                   108
      Multiport Head.

21    Protocol for Sampling Ambient Air in Los Angeles, CA.        113

22    Operating Parameters for GLC-MS-COMP System.                 115

23    Pollutants in Ambient Air from West Covina, CA.              116

24    Pollutants in Ambient Air from Santa Monica, CA.              118

-------
                              ACKNOWLEDGEMENTS






     The engineering support of Mr. B. Carpenter is gratefully acknow-




ledged for the design of a field sampling unit, pressure differential




calculations for cartridge samplers and a thermal desorption chamber




with an annular flow pattern.  The valuable assistance of Mr. J.  Bunch




for executing laboratory and field experimentation is appreciated.   Mr.




L. Retzlaff provided expert machining and construction of experimental




devices used in this research program; a sincere thanks for his support.




The design and fabrication of the temperature controller was performed




by Mr. R. L. Marguard and Mr. C. Cleary; their help is also gratefully




appreciated.




     The personnel at the CHAMP stations in Santa Monica and West Covina




are thanked for their help while field samples were acquired at these




sites and Drs. G. Lauer, J. Hribar and R. Myers at Rockwell International




Science Center in Thousand Oaks, CA for making available these facilities




and their extended courtesies during the author's stay.  Approval for use




of CHAMP sites was given by Mr. Ferris Benson of the Health Effects Research




at EPA, Research Triangle Park, N. C.




     The helpful suggestions of Dr. M. E. Wall throughout the program




and the computer program for single ion plotting made available by Dr. D.




Rosenthal are appreciated.

-------
                                SECTION I




                               CONCLUSIONS






      An analytical technique was developed for evaluating collection




efficiencies of candidate sorbents during the concentration of hazar-




dous vapors from a flowing stream.  The polymeric beads - Tenax GC,




Porapak Q, Chromosorb 101 and 104 - were >_ 90% efficient in trapping




hazardous vapors such as epoxides, $-lactones, sulfonates, sultones,




N-nitrosamines, chloroalkyl ethers, aldehydes and nitro compounds from




synthetic air/vapor mixtures at 0.25 1/min.  Tenax GC and Chromosorb




101 were also tested at sampling rates up to 9 1/min and efficiencies




of >_ 90% were maintained.  Carbowax 400 and 600 and oxypropionitrile




coated or chemically bonded to supports and activated carbons were also




highly efficient.




      A thermal desorption inlet-manifold for recovering and transferring




hazardous substances from sorbents to a gas-liquid chromatograph or a gas-




liquid chromatograph-mass spectrometer was developed.  The interface con-




sisted of a desorption chamber, a six-port two position high temperature




low volume valve, a Ni capillary trap and a temperature controller.  This




unit was utilized to determine the temperature rise times in the center




of cartridge samplers for a variety of sorbents under isothermal chamber




temperatures.  The heating rates were:  PCB and BPL activated carbons > oxy-




propionitrile and carbowax 400 chemically bonded to Poracil C > Chromosorb




104 > Tenax GC > Chromosorb 101.  The heating rates were linear for all




sorbents up to 65% of the set desorption chamber (60-90 sec) but required




several minutes thereafter to reach the final temperature.  Since  the per-




cent recovery of several hazardous vapors adsorbed on Tenax GC using this

-------
inlet-manifold was >_ 90% at the 50 ng level, it was concluded that the




designed system was satisfactory for analysis of cartridge samplers.




     Design criteria were developed for a field sampling system for col-




lecting pollutants using a cartridge containing a solid sorbent.  The




relationship between the ambient vapor concentration, the total volume




of air required and the time required for sampling at various flow rates




were important considerations in the design specifications.  It was con-




cluded that the power required to pump air through sorbent packed tubes




was a function of pressure differential (AP) across the tube under flow




conditions.  The AP was found to be related to (1) sampling rate, (2)




diameter and length of cartridge, (3) particle size distribution, and (4)




particle shape.  Based upon AP the power requirements for a pumping system




was calculated and applied to the fabrication of a field sampler.




     The methodology and instrumentation developed under this program was




also applied to the analysis of air samples from the Los Angeles Basin area.




Using glc-ms-comp techniques, many aliphatic and aromatic compounds were




identified in these preliminary studies.  The relative intensities of




single ion plots for ions representing aliphatic and aromatic cracking pat-




terns revealed that Tenax GC did not efficiently trap background aliphatic




constituents, a desirable feature since most hazardous vapor of interest




are semipolar/polar compounds.  It was concluded that improved techniques




for resolving background pollutants occurring at high concentrations from




trace hazardous vapors need to be developed.




     One oxygenated compound of significant interest tentatively identified




as styrene oxide was discovered in air from West Covina and Santa Monica, CA.

-------
                                SECTION II




                              RECOMMENDATIONS






     Four major phases of research in the current program are recommended




to be expanded and pursued.  These are listed as follows:




     (1)  Sampling -




          A concerted effort should be undertaken to collect field samples




          at geographical sites postulated to contain hazardous (mutagens,




          carcinogens and other alkylating agents) compounds under study.




          The effects of transportation and storage on adsorbents and




          reliability and accuracy of analysis for collected pollutants,




          and potential sources of contamination should be determined.




          Experimental criteria required for optimum performance of adsor-




          bents should be established.  It is recommended that an alternate




          backup sampling device also be examined.  The sampling devices




          should be improved and the system miniaturized for portability.




     (2)  Inlet-manifold unit -




          The strengths and weaknesses undercovered in this report should




          be considered for perfection of the design of the thermal desorp-




          tion unit; it should interface efficiently with a glc and glc-ms.




     (3)  Resolution of pollutant mixture -




          Techniques for the separation of hazardous substances under study




          from the many hundreds of organic pollutants which are of secondary




          interest at this stage should be developed and perfected.  It is




          recommended that the identity and relative quantities of background




          organic pollutants which interfere with the major goals be established,

-------
(4)   Characterization of hazardous compounds -




     The unequivocal characterization of atmospheric pollutants  of




     interest and those hazardous to living organisms is recommended




     using methodology and instrumentation developed under this  re-




     search program.  Finally,  the specifications for routine assay




     of hazardous substances based on the discovered ones should be




     delineated.

-------
                               SECTION III



                       INTRODUCTION AND BACKGROUND






CARCINOGENIC VAPORS IN AMBIENT ATMOSPHERE



                                                                         1-3
     Carcinogenic vapors have been postulated to occur in the atmosphere,



however until the present program was initiated no serious and thorough



endeavor had been made to collect and determine these substances.   Epoxides,



peroxides, aldehydes, ketones, lactones, sulfonates, sultones, and nitroso



and nitro compounds have been isolated in laboratory experiments during


                                                                     4

olefin oxidation, ozonization, sulfonation, nitrosation and nitration .



Some of these compounds and reaction mixtures have demonstrated carcinogenic



activity in animals    .  For example, the exposure of mice (strain A or



C57B) to an atmosphere containing ozonized gasoline increased the incidence




of tumors .



     Because unsaturated hydrocarbons constitute a large fraction of organic




air pollutants, it is reasonable to anticipate that their oxidation products



and their production of reaction with NO  and SO , whether spontaneously or
                                        X       X


photochemically induced, may also be present in ambient atmosphere .  Many



types of alkylating and arylating agents are being introduced directly at a



continually increasing rate into our environment, e.g. as industrial inter-



mediates in organic synthesis, organic solvents for various chemical pro-



cesses, as cross-linking agents in manufacturing processes, as medicines, and



as antibacterial and fungistatic agents.



     There is a strong indication that nitrosamines are present in the atmos-



phere; their identification in cigarette smoke has been confirmed  .  Alky-



lating compounds resulting from limited-aeration smouldering of plastics,



paper, and cellulose as well as auto-exhaust gases have been demonstrated to


                                                                     2
occur and have been postulated as a health hazard in industrial areas  .

-------
     On the basis of the current knowledge of air pollution, most of the




potentially deleterious vapors which could be formed in or expelled into




the atmosphere could be therefore classified as epoxides, $-lactones,




peroxides, hydroperoxides, sulfonates, sultones, nitrosamines and a-




chloroalkyl ethers.  Some of these pollutants may have a short lifetime




in the atmosphere.



     The National Academy of Sciences panel in a study of the biological




effects of atmospheric pollutants has concluded and recommended in their




report on Particulate Polycylic Organic Matter  "Research is needed on the




chemistry and biological activity of air pollutant cocarcinogens and tumor-



promoting agents, such as polyphenols and paraffin hydrocarbons, and on the




oxidation products of airborne olefins and aromatic hydrocarbons, including




the nature of the epoxides, hydroperoxides, peroxides, and lactones formed




and their biological properties".  Recently, Van Duuren     summarized a




review on the biological properties of carcinogenic vapors with the statement




"in view of the obvious importance of these aliphatic compounds (epoxides,




hydroperoxides and peroxides), it is imperative that studies be undertaken



on the analysis of volatile organic air pollutants".  Once the identity of




the physiologically active vapors present in polluted atmospheres are known,



then investigators can ascertain which substances need to be routinely ana-



lyzed, studied epidemiologically and eventually controlled.



     The primary mission of this research program has been to develop methodo-



logy for the reliable and accurate collection and analysis of mutagenic and



carcinogenic vapors present in the atmosphere down to nanogram per cubic



meter amounts.

-------
METHODS FOR COLLECTION AND ANALYZING POLLUTANTS FROM AMBIENT AIR



Collection Techniques



    The characterization and measurement of extremely minute amounts (ppb)



of these hazardous compounds in ambient air has been seriously hampered by



the lack of a reliable sampling system and sensitive instrumentation for



direct analysis.  Special systems have been developed for concentrating trace



organic vapors from large volumes of atmosphere and transferring the collec-


                                  12—28
ted vapors to an analytical system



    Many collection devices and analytes have been employed by investigators



in air pollution.  In general, the concentrating techniques have utilized


        . 22-27   ,     t.  17,28,31      .    „.  13-16,18-27 ..    .      „,,
cryogenic     , absorptive         or adsorptive            trapping methods.



Cryogenic (freeze-out) methods are particularly suitable for analysis of



highly volatile substances; however, if liquid nitrogen, oxygen, or solid



carbon dioxide/acetone is used as a coolant large quantities of water may



accumulate, which is a major problem during chromatographic analysis.  Aerosol



formation may be also experienced with this technique reducing the trapping



efficiency.  Drying the gas by passing the air over desiccants prior to cryo-


                                                                      29-31
genie trapping is not feasible since some solutes may also be scrubbed



Nevertheless, the major advantage of this approach is that it is the only



technique which permits the collection of low molecular weight air pollutants.



Also, oxidation or polymerization of constituents is minimized during their



concentration.


                                                                  21
    Activated carbons (including activated carbon molecular sieves  ) have



been shown to adsorb all of the classes of carcinogenic-type compounds to


                             13 14 32-34
be studied under this program  '  '      .  Where the adsorption is purely



physical, the compounds may be retained and subsequently released without



being decomposed.  Brooman used activated carbon to adsorb acrolein

-------
quantitatively up to the point of breakthrough, then used solvent extraction


                                14
to recover the trapped substance  .  Saunders used adsorbent carbon to iden-


                                                              35
tify 23 volatile compounds in closed environmental atmospheres   .  Ethylene



oxide has been successively recovered from activated carbon at 30°C; acet-


                      35
aldehyde required 175°



     The high surface activity of activated carbons also can produce arti-



facts during recovery.  Formaldehyde decomposes, while methyl ethyl ketone


                                                                     35
tends to form diacetyl compounds and acetic acid, on activated carbon



Epoxides and peroxides are usually destroyed, hence, activated carbon alone



has very limited potential as a sorbent for collecting highly polar and



reactive compounds.



     Compounds that are sensitive to polymerization or decomposition on


                                                          20
activated carbon are generally more so on molecular sieves  .  For this reason



the potential role of molecular sieves is probably limited to removing and



recovering simple molecules (H»0, NH«, CH,) from air.



     One unique approach which is based on gas and liquid chromatographic



principles employs liquid phases uniformly coated on solid supports (e.g.



silica, diatomaceous earth, polystyrenes) and these polymers exhibit solution



formation with trace organic vapors at ambient temperatures.  Williams has



reported a collection device for organic compounds such as hexane, benzene,



toluene, aldehydes, ketones, and chlorinated hydrocarbons at the part per



hundred million level  .  The device consisted of a tube packed with Chromo-



sorb P coated with the stationary phase, di-ii-butylphthalate.  During air



sampling the collection tube was cooled in dry ice and water was removed by



using a drying agent prior to drawing the air through the sampler.


                                                       17 9ft *}fi "^7
     Application of chemically bonded stationary phases  >  »  »   to



sampling of air pollutants appears to be potentially promising because

-------
selectivity can be incorporated by matching the physico-chemical properties



of the phase with the pollutants of interest.  Because the column packings



(glc and Ic) are essentially nonextractable, thermally and hydrolytically



stable, they exhibit lower backgrounds during thermal desorption than the



conventional support coated liquid phases.  Permaphase^ (E. I. duPont de



Nemours & Co.) is formed by reacting silane reagents with the surface of
the porous shell of Zipax^and then polymerizing the reagents to yield the


                        37
desired silicone coating  .   Ether bonded polymeric coatings are prepared



with a variety of functional groups, ranging from very polar to nonpolar,



providing a wide range of selectivities (Table 1).  The physical charac-



teristics of the stationary phase, such as concentration, film thickness,



and structure (i.e. linear or crosslinked), is controlled to obtain the



desired properties for affinity to specific classes of trace organics.



    By bonding the functional moiety chemically to the core material the



vapor pressure is reduced to near zero, hence a low bleed rate is observed.



This feature makes these solvents attractive during thermal desorption of



trapped organics since a minimal background interference from the polymer-



solid support occurs.



    Thermal degradation of the bonded functional moiety can result if



excessive temperatures are employed.  Table I depicts some representative



commercially available materials and their physical properties.



    Aue has employed a bonded silicone polymer (octadecyltrichlorosilane)


                                                                  12
to collect several organic compounds from fast-flowing gas streams



Silicone coatings have been shown to concentrate and release many chemical



classes of interest to this program; however, lactones, epoxides, and per-



oxides had not been examined for quantitative analysis.

-------
             Table 1.  COMMERCIALLY AVAILABLE MATERIALS WITH


                     CHEMICALLY BONDED LIQUID PHASES
p
Type
OPN/Porasil C
Carbowax 400/Porasil C
n-Octane/Porasil C
Carbowax 400/Porasil S
Phenyl isocyanate/Porasil C
Carbowax 4000/Porasil S
article Size
(mesh)
80/100
100/120
120/150
80/100
80/100
80/100
Temperature limit
Polarity3 °C
M
N
P
N
P
P
135
150
160
200
60
200
 M = medium, N = nonpolar, P - polar



     Other types of bonded functional groups are the "brush"-like stationary


phases.  These materials are prepared by esterification of the alcoholic


groups on silica surfaces.  One certain disadvantage of esters is their


suspectibility to hydrolysis.  Conceivably this might be a problem during


collection of trace organics from large volumes of polluted air, if water


vapor is also trapped during the solution process.  An example of a "Brush"


type is polyethylene glycol 400 on silica.

                                 /5\
     Similarly to the Permaphases^ (or Durapak, Waters Assoc.), Poragel-P


packings are marketed.  These are polystyrene resins with permanently


attached functional groups.  The surface affinity of the polymer and bonded


functional group also allows interaction and thus trapping of organic solutes.


     Several polystyrene type polymers are commercially available (e.g.


Chromosorb 101, 102, 104) which are used in gas-solid chromatography.  Chromo-


sorb 101, a styrene-divinylbenzene polymer adsorbs hydrocarbons, alcohols,


acids, esters, aldehydes, ketones, ethers, and glycols.  A polar surface is
                                    10

-------
provided by Chromosorb 104 which is an acrylonitrile-divinylbenzene resin.



The potential utility of these polymers has not been completely exploited



for trace organic vapor analysis.


                    38
     Leggett, et_ _al.   developed a collection cartridge containing 2 g of



Porapak Q-S sorbent for concentrating and determining trace organic vapors



at ppb levels.  Samples were successfully collected at 100 to 1,000 ml/min.


                   19
     In 1966 Hollis   reported the use of porous polyaromatic beads concen-



trating air samples on a chromatographic column at room temperature.  The



components were subsequently eluted by temperature programming.  With these



beads, water is not a major interferent since it is eluted as an early peak



without tailing.  Hollis successfully adsorbed and desorbed epoxides from



these polymer beads.  The use of polyalkyl styrene polymers to remove organic


                                                      39
vapors from a gas stream was patented in 1972 by Haigh


              40            41           3
     Dravnicks   , Crittenden  , and Jones  have described collection systems



using Chromosorb 102.  All employ thermal desorption for recovering trapped


              3

vapors.  Jones  used this solid sorbent to demonstrate the presence of trace



quantities of ethylene sulfite a suspected carcinogen in secular atmospheres.


                                 20
     More recently Bertsch et al.   described the collection of trace quan-



tities of organic volatiles from air using Tenax GC a porous polymer of 2,4-



diphenyl paraphenylene oxide with a high temperature stability.  The trapped



substances were  subsequently heat desorbed and the mixture resolved by high



resolution Ni capillary columns.  Several hundred hydrocarbons were recognized



and the identity of  about 100 was established.



Recovery and Analysis



     Because of  the  limited sensitivity of currently available detectors



hazardous substances need to be  concentrated from highly dilute samples.  A



step toward  the  solution of this problem can be achieved when  cartridges
                                     11

-------
containing an appropriate sorbent is used and large volumes of air are




forced or drawn through the sampling device whereupon the pollutants are




trapped.



     Even when the cartridge technique is used, only trace quantities of




hazardous pollutants can be expected to accumulate; thus it is imperative




that the entire sample be submitted for analysis.  Recovery of trapped vapors




has been accomplished using thermal     '   and vacuum      desorption which




allows for direct introduction of the total sample into an analytical system




(glc, glc-ms) in the absence of a. solvent as a carrier.  These methods are



subject to artifactual processes such as pyrolysis, polymerization or incom-




plete recovery.  Steam desorption  '   and solvent extraction      of the



sorbent alleviate the above mentioned problems; however, volatile pollutants



cannot be quantitatively concentrated from dilute solutions and since gas



chromatographic (gc) analysis is limited to small aliquots of liquid samples



only a fraction of the sample can be examined.  As a result the sensitivity




of the overall method is greatly reduced.
                                    12

-------
                               SECTION IV


                    PROGRAM OBJECTIVES AND RATIONALE



PURPOSE OF PROGRAM


     The general scope of this research program was to develop methodology


for reliable and accurate collection and analysis of mutagenic and carcino-


genic vapors (collectively referred to hereafter as hazardous compounds)

                                                          3
present in trace quantities in the atmosphere down to ng/m  amounts.  Once


the physiologically active vapors present in polluted atmospheres have been


determined, investigators can then ascertain what substances need to be


routinely analyzed, studied epidemiologically and eventually controlled.


The major objectives were:


     (1)  to determine, develop and fabricate a sampling device for the


          collection of atmospheric vapors at ambient temperatures, i.e.


          substances which are volatile solids or liquids in bulk at room


          temperature and pressure.


     (2)  to interface the sampling device to an analytical system(s) such


          as a gas chromatograph or gas chromatograph - mass spectrometer -


          computer for the characterization and assay of vapors such as


          N-nitrosamines, sulfones, sultones, sulfonates, epoxides, lactones,


          anhydrides, aromatic amines, peroxides, hydroperoxides or any


          other alkylating agents present in polluted atmospheres.


     (3)  to acquire performance data for the collection and analysis


          methodology embodying the sampling and interfacing devices.  This


          was to include  collection efficiency, reliability of collection


          and analysis as they relate to environmental factors such as


          sampling locations, temperature, possible  interferences from
                                     13

-------
          other gaseous pollutants, dust load and winds.  Artifacts and




          interferences were to be determined and an effort made to cor-




          rect or minimize them.



     (4)  to provide experimental support on -



          a.  the efficiency of pollutant collection and bleeding, and




              their relationship to atmospheric variables,



          b.  the effect, if any, of other gaseous pollutants on the




              reliability of the overall analysis methodology,




          c.  the qualitative composition of the ambient vapor sample in-



              cluding those constituents which are of secondary importance




              or interferences,




          d.  the characterization of those pollutants containing functional




              groups common to hazardous compounds (including alkylating



              agents)




     and  e.  the estimation of the amounts of some of the atmospheric  "•




              vapors and especially those containing the functional groups



              of interest.



GENERAL APPROACH TO PROBLEM



     Although the classes of compounds which are to be studied in ambient




air under this program cover a wide range of chemical functionalities and



physical properties they can be classified as semipolar and polar.  For




this reason, the reported techniques for collecting and analyzing atmospheric



pollutants need to be examined carefully prior to their acceptance as valid



methods for use in this program since the literature methods were originally




designed for aliphatic and aromatic pollutants occurring at the parts-per-



million level.
                                    14

-------
     The experimental approach adoped here was to thoroughly examine and


evaluate previously reported methods and provide appropriate modifications


which would permit the analysis of hazardous vapors for this program.  In


doing so, an in depth study beyond the superficial treatment provided in


the literature was to be conducted.


     Because the concentrations of hazardous vapors sought were anticipated

             3
to be in ng/ra  amounts or less, it was recognized that the major problems


would arise with the development of an adequate collection system for con-


centrating sufficient quantities of material.  The relationship between


the sampling parameters for collection, say 30 ng of a pollutant, from


ambient air is shown in Table 2.  It was readily apparent that the sampling


time required was directly proportional to the volume of air which must be


concentrated for analysis and inversely to the sampling rate at a given


pollutant concentration.  High sampling rates would be required in order to



           Table 2.  SAMPLING PARAMETERS FOR COLLECTING 30 NG


                            OF VAPOR FROM AIR
Quantity of Volume
Vapor in Air Required
u*/m
1
5
10
50
100
1000
30
6
3
0.6
0.3
0.03
0.25
2000
400
200
40
20
2
1
500
100
50
10
5
0.5
Sampling Time (hr) at
Various Flows (1/min)
4
125
25
12.5
2.50
1.25
0.125
9
56
11.1
5.6
1.12
0.56
0.056
20
25
5
2.5
0.5
0.25
0.025
30
16.7
3.33
1.6
0.33
0.17
0.017
62.5
8
4
0.8
0.16
0.08
0.008
                                    15

-------
collect enough of a compound within a feasible period of time.  For example,


a pollutant at a concentration of 1 ng/m  would require a sampling volume

                    3
of a minimum of 30 m ; at a 1 1/min ( a rate most commonly reported by


investigators) a sampling time of 500 hr would be needed.  Obviously, this


would be too long.  Sampling rates of 20 1/min or greater would be more


satisfactory.


     Many factors come into play when high flow rates are employed, parti-


cularly the pressure differential developed across the sampling device and


subsequently the power requirements needed to achieve the desired rates.


Such factors were considered in the development of a collection and sampling


system.


     During the consideration of a collecting system, specific criteria


emerged as important in its design.  These were:  (1) simplicity, (2) dura-


bility, (3) reliability, (4) ability to store samples for two or more weeks,


(5) convenience during its transportation to and from sampling site, (6)


convenience of its operation and maintaining the collection system during


field trails, and (7) readily interfaced with an analytical system.  The


general concept chosen was a cartridge sampler where the analyte consisted


of a solid material capable of trapping vapors of interest.


     Since one of the major objectives of this program was to characterize


the collected pollutants and because they were suspected to occur in trace


amounts, the analytical instrumentation of choice which offered the most


promise for achieving these goals were gas chromatography and combined gas


chromatography-mass spectrometry.  Interfacing the collection cartridge to


a gas phase analytical system was believed to best be served by using an


interface which would thermally desorb the vapors from the analyte and in-


troduce the entire sample for analysis.
                                    16

-------
      Reagent hazardous vapors listed in Table 3 were chosen to assist




in the evaluation of the developed methodology and instrumentation.  A




compilation of mass spectra has been assembled in the Appendix of this




report (Fig. 46-63) for many of these compounds.  Additional spectra will




be added to establish a library of hazardous vapors for future references




purposes.



      This report presents the results obtained during the execution of




the described experimental plan.
                                     17

-------
         Table 3.  REAGENT VAPORS FOR EVALUATING SAMPLING MEDIUM
  Chemical Class
        Substance
Chloroalkyl ethers
Lactones
Nitro and Nitroso


Sulfones


Sulfonates

Epoxides
Peroxides
Aldehydes and Ketones
Hydroperoxides

Acid Anhydrides


Sultones

Sulfolane

Aromatic Amines

Imino Heterocyclic
Bis-(chloromethy1)ether
alpha-dichloroethyl ether

3-propiolactone
3-butyrolactone
vinylene carbonate
parasorbic acid

diethy1 nitrosamine
nitromethane

trional
diethy1 sulfone

ethyl methanesulfonate

glycidaldehyde
propylene oxide
styrene oxide
1,2,3,4-diepoxybutane

methyl ethyl ketone peroxide
lauroyl peroxide

glycidaldehyde
acrolein
methyl ethyl ketone
phenylvinyl ketone

cyclohexene hydroperoxide

maleic anhydride
succinic anhydride

propane sultone

tetramethylene sulfone

Aniline

Aziridine
                                    18

-------
                                SECTION V




          DESIGN OF A CARTRIDGE SAMPLER FOR CARCINOGENIC VAPORS






     The development of a cartridge containing solid material for the




collection and analysis of carcinogenic vapors from ambient atmosphere




required a study of the physico-chemical properties of several sorbents.




These investigations included (1) the examination of collection effi-




ciencies of several sorbents, (2) the relationship between cartridge




dimensions (length and diameter), sorbent particle size, sampling rate




and pressure differential, and (3) an estimation of the breakthrough




volume for hazardous vapors.  During the evaluation of candidate sorbent




media, their collection efficiencies and recovery for analysis were in-




dependently determined.  This was important because a sorbent may have




good trapping characteristics but exhibit undesirable effects when a




certain type of recovery method was used.  The converse was also the case.




We therefore designed our experiments to allow an independent assessment




of each step.




DETERMINATION OF COLLECTION EFFICIENCIES FOR SEVERAL SORBENTS




     The performance of many sorbents as to their ability to extract and




retain hazardous vapors from a moving air stream has not been adequately




studied.  The parameters which are involved in determining the performance




of sorbents (collection efficiencies) can be divided into two categories.




There are those on the one hand which are related to sampling environment




such as flow rate, air temperature, and humidity, and those which are re-




lated to the physico-chemical properties of the sorbent such as surface




area, particle size and porosity, solute capacity, sorption mechanism,




degree of solute affinity, etc.  Furthermore, some of these factors which




influence sorbent performance are not independent of each other.

-------
     Because the collection and analysis of volatile hazardous substances

       3
in ng/m  amounts from ambient atmosphere requires the selection of a sor-


bent which is efficient under a variety of sampling conditions, an instru-


mental technique was designed for evaluating the collection efficiencies


of sorbents.  This section presents the performance of a number of candidate


solid materials for the concentration of substances such as epoxides, $-


lactones, sulfonates, sultones, nitrosamines, chloroalkyl ethers, aldehydes,


and nitro compounds from an air stream comparable to field sampling conditions.


Experimental


     Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 10A


(100/120) and Chromosorb W-HP (100/120) were purchased from Applied Science,


State College, Pa.  A series of stationary phases chemically bonded to


supports including carbowax 400/Poracil C (100/120), oxopropionitrile/Pora-


cil C (80/100), and phenylisocyanate/Poracil C (80/100) were also obtained


from Applied Science.  Stationary phases consisting of carbowax 600, didecyl


phthalate, and tricresyl phosphate and the sorbent Porapak Q were from


Supelco, Inc., Bellefonte, Pa.


     Carbon derived from coke (PCB and BPL, 12/30) was acquired from Pitts-


burgh Activated Carbon Division of Calgon Corp., Pittsburgh, Pa.  Cocoanut


derived carbons (SAL19190 and 580-26) were purchased from Barneby Cheney,


Columbus, Ohio.


     Ethyl methanesulfonate, 3-propiolactone, N-nitrosodiethylamine, 1,2-


dichloroethyl ethyl ether, nitromethane, methyl ethyl ketone, and aniline


were from Fisher Chemicals, Pittsburgh, Pa.  Glycidaldehyde and sulfolane


were obtained from Aldrich Chemicals, Milwaukee, Wis.  From Eastman Organic


Chemicals, Rochester, N. Y.  1,3 propane sultone, maleic anhydride, butadiene
                                    20

-------
diepoxide and propylene oxide were purchased.  Styrene epoxide, bis-(chloro-



methyl)ether and bis-(2-chloroethyl)ether were from K&K Labs., Plainview,



N. Y.



    The monitoring system shown in Figure 1 was designed and assembled



for measuring collection efficiencies.  Air (breathing quality, Linde Div.



Union Carbide, East Brunswick, N. J.) from a pressurized reservoir was



passed through a scrubbing tower (5 cm i.d. x 30 cm) which contained layers



of CaCl2 dessicant and BPL activated carbon (12 x 30 mesh, Calgon Corp.,



Pittsburgh, Pa.) to remove trace contaminants.  Purified air entered the



monitoring system at point A as shown in the schematic where the flow rate



was controlled with a Sho-Rate 250 flow meter (Model 1357-12F1BAA Brooks



Instruments Div., Emerson Electric Co., Hatfield, Pa.) equipped with a



teflon diaphragm regulator for compensating downstream pressure changes.



The air stream passed through a 2 1 cylindrical chamber (point B) fitted



with an injection port where known quantities of organic vapors were intro-



duced.  The chamber delivered synthetic air/vapor mixtures to the cartridge



sampler which contained the sorbent under study (point C) according to the



following relationship:



                              C - Coe-Ft/V                 (1)




where  C  = initial concentration in chamber
        o


        C = concentration in.chamber after time, t, has elapsed



        F = purging rate (ml/min or 1/hr)



        V = volume of chamber



The effluent stream from the sampler was split and a flow of 50-100 ml/min



was directed to a flame ionization detector  (Model 1200, Varian Instruments,



Corp., Walnut Creek, Ca.).  Because hazardous organic vapors were used in



this study, the apparatus between points A and D was contained in a glove box
                                    21

-------
Ni
K>
                     FLOW
                     METER/REGULATOR
                                                                               EXHAUST  TO
                                                                         _     CKY08ENIC  SAFETY
                                                                                  TRAPS
                                                                                                   RECORDER
7
1
" '"AIR
AMP.



MM
•MMMM^ •

m
-
                         Figure  1.   Monitoring system for hazardous vapors in cartridge effluents

-------
(Kewaunee Scientific Equipment, Adrian, Mich.) which was evacuated by vacuum




through cryogenic safety traps.  Hydrogen and air flow to the detector were




35 and 250 ml/min, respectively.  The detector output signal was amplified




(Varian Model 520) and recorded with an Omniscribe^ strip chart recorder




(Houston Instruments, Houston, Tx.).  This apparatus monitored total organic




vapor in the cartridge sampler effluent.




    Sorbents were packed in glass tubes (1.056 cm i.d. x 10 cm in length)




using 1 cm of silanized glass wool plugs for support.  The cartridge




samplers were inserted in canisters which were constructed from tube L




3/4 in copper fitted with 3/8 in Swagelock^unions.  The entrance and




exit lines at point C in the monitoring system were 3/8 in o.d. Teflon^




(Comco Plastics Corp., Raleigh, N. C.).




    To synthesize known concentrations of air/solute vapors mixtures




(Table 4), microliter quantities of each organic compound were added to a




2 1 cylinderical flask.  The flask was heated to 50° and the air/vapor mix-




ture was continuously stirred.  An aliquot from this stock reservoir was




transferred to the chamber (point B) in the monitoring system.  By con-




tinuously monitoring the cartridge sampler effluent with a flame ionization




detector the collection efficiency of each sorbent was determined.  A decay




curve (Fig. 2A) which represented the concentration of the air/vapor mixture




leaving the chamber per unit time was established for each mixture and at




each purging rate by using empty cartridge samplers.  The per cent collection




efficiency was estimated by comparing the areas under curves obtained for




samplers with and without sorbent (Fig. 2A and 2B).




Results and Discussion




    At a sampling rate of 0.25  1/min, over 90% of the synthetic air/vapor




mixture is purged through the  cartridge sampler in  less than 20 min  (Table 5).
                                     23

-------
to
                                                         (no  sorbent  in  cartridge)
                                                                          B   (Tenax GC  -  1.0  cm i.d. x 3.0 cm)
                                     Figure 2.
                                                            4     0
                                                            TIME  ( WIN.)
Elution profile of cartridge  effluent.   Sampling

rate was 6 1/min.   Test mixture III was  used.

-------
               Table 4.  SYNTHETIC AIR-VAPOR MIXTURES FOR

                      CARTRIDGE SAMPLER EVALUATION
Mixture                                        Componentsa


   I                                    ethyl methanesulfonate
                                        3-propiolactone
                                        N-nitrosodiethylamine
                                        1,2-dichloroethyl ethyl ether
                                        nitromethane
                                        methyl ethyl ketonec

  II                                    styrene epoxide
                                       %-nitrosodiethylamine
                                        butadiene diepoxide
                                        glycidaldehyde
                                        sulfolane
                                        propylene oxide

 III                                    aniline
                                        bis-(2-chloroethyl)ether
                                       bN-nitrosodiethylamine
                                        Bis-(chloromethyl)ether
                                        maleic anhydride
                                        1,3 propane sultone
Q
 Approximately 300 ng of each component was used for synthesizing the
 air/vapor mixture.
^Internal standard.
cAs decomposition product from MEK peroxide.
At 9 1/min it takes approximately one minute.  Since the decay process

for moderate concentrations of vapor (~0.05 ng/s) can be monitored with

a flame ionization detector, the ability of a sorbent to extract vapors

from a flowing gas stream was determined.  This phenomenon was defined

as collection efficiency, i.e. the fraction of solute vapor in the polluted

gas which was retained by the sorbent bed.

     The collection efficiencies for several sorbent media are given in

Table 6.  All of the polymeric bead sorbents were relatively effective

in extracting vapors from a flowing stream of 0.25 1/min.  The activated
                                    25

-------
                 Table 5.  EXPULSION RATES FROM CHAMBER




                           AT SPECIFIED FLOWS
Concentration (ng/1) remaining in chamber
Time elapsed, min
0
1
2
3
4
5
10
20
30
0.25 Urn
1800
1588
1402
1238
1092
964
516
148
42
4 l/m
3000
406
54.9
7.4
1.0
0.14
0.056


6 A/m 8 £/m 9 fc/m
3000 3000 3000
149 54.9 33.3
7.4 1.0 0.4
0.37 0.02 0.004
0.02




carbons were highly efficient in trapping the constituents in mixture I




(see also figures 3-5); however, they were relatively ineffective when




tested with mixtures II and III.  In separate experiments it was dis-




covered that propylene oxide and butadiene diepoxide were not effectively




trapped by two of the cocoanut type carbons and thus accounted for the




low collection efficiencies for synthetic air/vapor mixture II.  A com-




parison of several liquid phases coated on solid supports revealed a




considerable decrease in the trapping efficiency when the polarity of the




phase was increased (carbowax 600 vs tricresyl phosphate).  These results




suggested that the more non-polar vapors are not forming a "solution" with




the polar liquid phases (absorption via "like dissolves like") as readily




as might be expected if polar vapors were tested.  A discrimination in the
                                    26

-------
                      Table 6.   COLLECTION  EFFICIENCIES  OF CANDIDATE SORBENTS



S-4
(!) CO
S T3
rH 0)
0 pq
P-i

co
pi
o
f>
H
n)
CJ

0]
0)
CO
cd
P-4
13
•H
3
0"
•H


Sorbent

Tenax GC
Porapak Q
Chromosorb 101
Chromosorb 104
Activated carbons






Chemical Type

2,6-diphenyl-p-phenylene oxide
polyalkyl styrene
Styrene-divinyl benzene
acrylonitrile-divinyl benzene
PCB cocoanut (Pittsburgh Act.)
BPL coal (Pittsburgh Act.)
SAL9190 cocoanut (Barneby Cheney)
580-26 Cocoanut/pecan (Barneby-
Cheney)


20% Carbowax 600 on Chromosorb W(HP) 100/120 mesh
bCarbowax 400/Poracil C 100/120
bOxypropionitrile/Poracil C 80/100
25% Didecyl phthalate on Chrom P 100/120
20% Tricresyl phosphate on Chrom W(HP) 100/120
A
Percent Efficiency
Mixture
I

95
90
95
98
90
90
90
95



90
90
98
50
20
Mixture
II

90
95
95
90
95
90
30
30



90
90
96
80
20
Mixture
III

80
90
95
80
90
—
—
—



90
95
90
50
20
 Sampling rate was  0.25  1/min, packing  bed  dimensions were 1.056 cm I.D.  x 3.0 cm in length.
 exponential dilution flask was maintained  at  50°C.
^Chemically bonded  phases.
The

-------
                                  Sensitivity?
                                  No Sorbent
                                  Mixture I
                                Parameters

                                  Sampling Rate:   0.25  1/min
                                  Split Ratio:   5/1
       6.A x 10"11 AFS
                  20         40
                   TIME  (MIN)
60
Figure 3.  Elution Profile for Synthetic Air/Vapor
           Mixture I  from Spherical Chamber
           (Empty Cartridge)
                      28

-------
a
<
o
_
u.
u.
o
UJ
o
oc.
   50 r
   40
   30
    20
    10
                  20         40
                       TIME  (  MIN )
60
 Figure 4.   Collection efficiency profile for BPL carbon usjng
            test mixture  1.   Sampling  rate was 0.25 1/ir.in;
            sensitivity was  6.A >: 10    AFS .
                           29

-------
so r
                                                                 Parameters
                                 40
                                    Sampling Rate:   0.25 1/min
                                    Split ratio:   5/1
                                    Sensitivity:   6.4 x 10"11 AFS
                                    Sorbent:  SAL9190 carbon
                                    Mixture I
                              O

                              _J
                              _l
                              Z>
u>
o
30
                                                     BASELINE
                                              _ .XL	
                                               20
                         40         60
                          TIME  (MIN )
80
100
                                              Figure 5.   Collection Efficiency Profile for SAL9190
                                                         Carbon Using Test Mixture I

-------
In the affinity and therefore trapping of chemical classes may be pos-

sible with these media; a desirable feature since the final resolution

would be somewhat simplified.

    Upon reviewing the collection efficiencies for candidate sorbent

media, and the results from thermal desorption experiments (Section VII)

two polymeric beads, Tenax GC and Chromosorb 101, were selected and

further tested at higher flow rates.  As shown by the data in Tables 7

and 8 high collection efficiencies were also obtained at rates up to

9 1/min.

    The described monitoring system can be also used to determine the

effects of continuous sampling when no additional solute vapors are pre-

sent in the air stream.  This represents the extreme case encountered

during field sampling.  When polluted gas enters a sorbent bed an equili-

brium zone is established near the point of entry.  As more pollutant is

introduced this zone may expand through the packing length until the


             Table 7.  APPROXIMATE COLLECTION EFFICIENCY FOR

                   TENAX-GC AT VARIOUS SAMPLING RATES
Sampling Rate (1/min)
Test Mixture3
I
II
III
4
95b
95
>95
6
90
95
95
8
90
90
95
9
90
90
95
 Approximately 1500 ng/component/mixture was used to determine efficien-
 cies.  Packing dimensions were:  10.5 mm x 60 mm, Mesh 60/80.  The
 exponential dilution  flask was maintained at 50°C.
     values are average of duplicate runs.
                                    31

-------
              Table 8.  APPROXIMATE COLLECTION EFFICIENCY FOR

                CHROMOSORB 101 AT VARIOUS SAMPLING RATES
Sampling Rate (1/min)
Test Mixture8
I
II
III
A
95b
>95
>95
6
95
95
95
8
90
95
90
9
95
90
80
 Approximately 1500 ng/component/mixture was used to determine efficien-
 cies.  Packing dimensions were:  10.5 mm x 30 mm, 100/120.  Exponential
 flask was at 50°C.
 All values are average of duplicate runs.
capacity of the sorbent is exceeded.  However, if after an initial period

of time no additional polluted vapors are introduced and purging of the

packing bed continues, the zone of vapors may move through the packing

bed.  When the mass zone moves to the end of the available packing bed

and the vapors begin to leave, breakthrough has occurred.  Furthermore,

the elution volume (EV) can be calculated if the time required for the

zone to traverse and elute from the sorbent bed and sampling rate are

known.  In an ideal system, EV has an infinite value.  An investigation of

EV is presented later in this section.

     The amount of vapor adsorbed to a given quantity of adsorbent depends

                                                                  54
on the pressure, temperature and concentration of the solute vapor  .  The

higher the pressure or concentration, the greater the amount adsorbed.  When

an adsorbent and solute vapor are placed in contact with each other, an equili-

brium is reached between adsorption and desorption.  If the concentration of

the vapor increases or decreases, the mass of adsorbed vapor also increases
                                     32

-------
or decreases to establish a new equilibrium value.  This is an important




phenomenona during the collection of vapors from a flowing stream; the




binding affinity of the adsorbent for the solute must be very high in




order that the sampling rate to be relatively independent of collection




efficiency.  The data presented in this section supports this viewpoint.




    The attractive forces of atoms or molecules responsible for adsorp-




tion on the surface of a solid are attributed generally to two phenomena:




physical and chemisorption.  Physical adsorption is primarily due to




van der Waal's forces which is similar to the condensation of a gas.  The




magnitude of the heat evolved during adsorption and gas condensation are




similar.  The quantity adsorbed may be several monolayers.  When the pres-




sure of the vapor or its concentration is lowered, a subsequent decrease




in physical adsorption is observed.  In contrast, chemisorption is not




readily decreased and only a monolayer of solute vapor is adsorbed on the




surface.  Also, the heat evolved during chemisorption is much larger than




in physical; a surface compound is formed.




    Langmuir  '   advanced a model theory of adsorption which regards the




surface of a solid to be homogeneous, i.e., that the sites on the surface




all exhibited equivalent affinity for the solute vapor, and each site was




independent of another.  However, surfaces are known to be heterogeneous.




The Langmuir equation is represented as follows:





                          c/q - I/Kb + c/b                    (2)





where  c = concentration of solute (m/1)




       q = moles of solute adsorbed per gram of adsorbent




       b = moles of solute vapor adsorbed per gram of adsorbent




       K = equilibrium constant for V + S —*• VS
                                     33

-------
The complex VS represents the fraction of adsorbent sites occupied by vapor



molecules (V), and V is the fraction of free sites.  At a given value of



q, V and VS are constant and equation (2) reduces to




                                K' = 1/c




Thus, the enthalpy of adsorption, AH°, may be calculated by plotting



1/c vs 1/T at a given q value:





                  AH° - 2-303RTlT2  (iog K'  - log K' )         (3)

                          12~11



The Clausius-Clapeyron equation (3) can be used to calculate the heat of



desorption.  A larger amount of heat is required to desorb a mole of gas



when only a small fraction of the surface is covered than when a large



fraction of the surface is covered; indicating inhomogeneity of the surface.



     Since adsorption affinity constants are strongly temperature depen-



dent, the effects of temperature should be considered in any collection



efficiency and breakthrough study.  We are currently examining the influence



of temperature and humidity on the collection efficiency of Tenax GC and



Chromosorb 101 using the monitoring system described here.  The monitoring



system is also amenable to studies of humidity and temperature and their



effects on sorbent trapping performance.  Calibrated humidity levels



can be introduced into the chamber or can be heated with a heating mantel



for simulation of high temperature and humidity as observed in field condi-



tions .



     With respect to bed dimensions (length and i.d.) and particle size,


                                                                  20
it has been suggested that they play no major role on breakthrough  ; how-



ever our preliminary studies have shown that an increase in both collection
                                    34

-------
efficiency and EV can be expected.  These relationships need to be defined




more closely.




    Displacement chromatography may occur during sampling of polluted




air leading to early breakthrough.  Because each substance has a specific




affinity for the sorbent, the quantity adsorbed is characteristic for each




substance; furthermore one compound can be displaced by another, if the




latter has a higher adsorption affinity.  Thus, breakthrough studies should




be performed in the field under the full complexity of polluted air rather




than in the laboratory.  These results are discussed below.




    The capacity of Tenax GC for various compounds such as alkanes, alco-




hols, and amines have been reported to be higher than for aldehydes, ketones,




and phenols  .  Bertsch et al.   also reported that all sulfur compounds




examined were trapped in a narrow zone at the cartridge entrance.  Vola-




tile hydrocarbon compounds containing less than five carbon atoms are




not efficiently trapped by Tenax GC   while aromatics were  .




RELATIONSHIP BETWEEN SAMPLING RATE AND CARTRIDGE PRESSURE DIFFERENTIAL




    Most of the design principles that have been developed for the effi-




cient use of sorbents have been prescribed for the removal of vapors from




air to improve its quality.  For example, Jones   discusses the most




efficient physical arrangement of a charcoal bed to remove benzene vapor




from air.  His design presumes a steady flow of air of uniform concentration,




with the gradual accumulation of vapor until the sorbent becomes saturated




and breakthrough occurs  (frontal elution).




    However, additional  criteria were considered in this problem during




the design of a sampling system which is based upon a cartridge concept




for the collection and analysis of trace organic vapor pollutants.  The




most important factors are revealed in the aerodynamic features of the
                                    35

-------
cartridge, specifically  the pressure differential, AP, produced across a


specified sorbent bed.   The pressure differential is related to (1) the


air flow-rate,  (2) the cartridge shape  (diameter and depth of packing),


(3) the particle size distribution, (4) the particle shape, and (5) to a


lesser extent upon the air temperature  and humidity.  Since AP is an


important ingredient for relating the power requirement of a field sampler


(Section VIII)  to sampling rate and duration, it was carefully investigated


in this program.


Experimental


     An equation was derived from the mechanical energy balance and Leva's

           53
correlation   to predict AP which was valid at isothermal flow:
           2    2 _ 2SRG2T
           1    2 =  g M
V
In ~ +
 2   2fm
Vl   
      s
                               (4)
                                              2
where  p. = pressure at inlet to sampler, N/cm

                                               2
       p« = pressure at outlet of sampler, N/cm


        2 = gas compressibility factor (air = 1)


        R = gas constant, 831467 N x cm/kg moles x °K

                                            2
        G = fluid mass velocity, kg/sec x cm


        T = absolute temperature, °K


       g  = conversion factor, force to mass, kg x cm/N x sec


        M = molecular weight of gas flowing, kg/mole

                                                3
       V1 = specific volumes of gas at inlet, cm /kg

                                                3
       V,- = specific volume of gas at outlet, cm /kg


       fm = modified friction factor for flow through packed solids


        e = bed void fraction

                                     2/3
         = particle shape factor = V   /0.205 x surface area
                                     36

-------
      D  = particle diameter of packing, cm



       L = depth of sampler packing, cm






The particle diameter (Dp) can be calculated as I/(Ex/dp), where x is the



mass fraction with size d  and
                         P



                               .    ,,    , xO.5
                               dp = (dl x d2)






The minimum and maximum openings of adjacent sieves are designated as



d.. and d».  In the above equation, the mass-force conversion factor, gas



compressibility factor, and upstream and downstream specific gas volumes



are represented by g , 2, V- and V2, respectively.



    For the computations, Z and § were set equal to unity, n and fm were



evaluated as functions of the Reynolds number (particle diameter x mass air



rate/air viscosity), and E was evaluated as a function of the ratio particle



diameter/tube diameter.  Because of the relatively small magnitude of



ln(V-/V-), it was omitted.  All calculations were programmed and executed



on an IBM 360/70 computer at the Triangle Universities Computation Center,



Research Triangle, N. C.



Results and Discussion



    Figures 6-8 depict computer generated pressure differentials for



three different sorbent particle ranges (12/30, 35/60, and 100/120) and one



set of cartridge dimensions.  It was evident from these data that AP in-



creased exponentially as the flow rate was increased.  Recalling that 10.03


          2
Newtons/cm  is equivalent to 760 mm Hg, then a flow rate of 5 1/min produced



a AP of 11.4 mm Hg (Fig. 6) for a mesh range of 12/30.  On the other hand,



to achieve the same flow rate with mesh ranges of 35/60 and 100/120



(Fig. 2 and 3) a AP of 68.4 and 228 mm Hg, respectively would be produced.
                                    37

-------
Ol
   3.92

   2.99

   2*6

   2.33

    1.99

   1.66
z
a.  133

  Q998

  Q666

  Q334

 0.0029
      0.167  2.65   5.13   762   10.1   12.6   15.1   17.6   20.0  22.5  25.0
                           FLOW  RATE  IN  L/MIN
   Figure 6.   Calculated pressure  differential for 12/30 mesh particles.
              Cartridge  dimensions were 1.056 cm i.d. x 3.0 cm in length.
                                      38

-------
   8.33

   7.50

   6.67

   5.84

cT  &01
 E
 z  4.17
 o.
 < 3.34

    2.51

    1.68

  0.850

  0.0194
      0.167   1.82   3.47  5.12   6.77    842   10.1    11.7   13.4   15.0   16.7
                             FLOW  RATE  IN   L / MIN
   Figure 7.  Calculated pressure differentials for  35/60 mesh particles.
              Cartridge dimensions were  1.056  cm i.d. x 3.0 cm in length.
                                       39

-------
U
Q.
<
  3.48


   3.14


   2.81


  2.47


   2.13


   1.79


   1.45


   I.II


  0.771


 0.432


0.0929
                            j	I
i	  i
         0.167  0.650  1.13    1.62  2.10   2.58   3.07   3.55   4.03  452   5.00

                          FLOW  RATE IN  L/MIN

    Figure 8.  Calculated pressure  differential  for  100/120  mesh  particles,

               Cartridge dimensions were  1.056 cm  i.d. x  3.0 cm in  length.
                                        40

-------
     The calculated pressure drop curves for each mesh size allows some


judgement to be made with respect to the practical attainable flow rates


for each of these mesh sizes, and the most suitable bed packing dimensions


which will allow the attainment of desired field sampling rates.  Further-


more, large AP values were undesirable since vacuum desorption ("stripping")


of vapors initially trapped on a sorbent would occur if a sampling system


was used whereby the air sample was drawn through the cartridge.  This situa-


tion would thus seriously affect the collection efficiency and no doubt


breakthrough (elution volume decreased) would occur prematurely.  A pumping


system which forces the air through the cartridge would be preferrable for


these reasons.


     In addition to deriving a function which would allow AP to be pre-


dicted under a variety of conditions, AP was experimentally determined to


establish the validity of equation (1).  The pressure differential was


measured for air either drawn or forced through the cartridge containing


sorbent.  Figure 9 compares AP values which were calculated assuming an


average particle diameter, D , of 0.0211 cm to the experimentally determined
                            P

for a sorbent with a 60/80 U.S. mesh range.  This data demonstrates a good


correlation between AP values calculated using equation (1) and those ex-


perimentally measured; however, for exact calculation of pressure drop the


particle size and shape must be known.


     The effect of (1) bed packing diameter, (2) bed length, (3) sampling


rate, and (4) particle mesh range on the pressure differential developed


across a sampling'cartridge was examined.  These data are depicted in


Figure 10-17 for air forced through the sampler.  High AP values were


experienced when cartridge diameters of 0.5 cm were tested  (Figures 10, 12,


and  14) at high sampling rates, even when coarse sorbent particles
                                      41

-------
   700r
   600-
   900-
E
E
a.

g
a
ui
ui
   4OO-
   300
200-
    100-
                                                 A = experimental

                                                 O = calculated

                                                 D  = 0.0211 cm
                                                  P
                                                 Cartridge = 1.056 cm i.d.  x
                                                             3.0 cm
 Figure 9.
                                10

                     FLOW  RATE ,  liters / min


         Comparison of calculated  and experimental pressure

         differential for  a  Tenax  GC (60/80) cartridge.
                                                                 20
                                  42

-------
   1200
  1000
   800
I  600
   400
   200
                                                                   B
          2  4   6   8  10  12  14   16  18
              FLOW  RATE   ( l/min )
1200,-
1000-
 800-
600-
400-
200
                                 9 l/min
                                                                                    l/min
                                                                                 5 i/min
                                                                                 4  l/min
           2468
         PACKING  DEPTH (cm )
Figure 10.  Pressure  differentials (AP)  for cartridges containing BPL carbon (12/30) with a
            cartridge i.d. of 0.5 cm.

-------
  200
  160
E
E
   120
   80
   40
                6      10      14

               FLOW  ( l/min)
200
160
 120
 80
 40
                                                               B
  18 l/min
15 l/min
                                                                                9 l/min
           2468

          RUCKING BED DEPTH (cm)
 Figure 11.  Pressure differentials (AP)  for  cartridges containing BPL carbon (12/30)

            with a cartridge i.d. of 1.056 cm.

-------
                    6cm
  1400
  1200
  1000
— 800
o»
X
E
E
  600
Q.
<
  400
   200
  ft/      f
• AA   Ji       '
                      I   I   I    I   I
            4  6  8  K)  12  14  16  18
             FLOW   l/min
                              1400
                              1200
1000
                               800
                              600
                              400
                              200
          B
                                               2             4
                                        PACKING BED DEPTH (cm)
                                           6
  Figure  12.  Pressure  differentials (AP) for cartridges containing Tenax GC (35/60) with a cartridge
             i.d. of 0.5 cm.

-------
 1400 r
  1200
  1000
E
E
  800
  600
  400
  200
                                       8 cm
                                •
                                       5 cm
                                        3cm
                                         cm
                                     .D-
                              j	L
         2  4  6   8  10  12  14  16   18
                 FLOW   l/min
KOO

1200

1000

800

600

400

200

   O
                                                             B
           2468
           PACKING  BED  DEPTH (cm)
  Figure 13.  Pressure differentials (AP) for cartridges containing  Tenax GC (35/60)  with a
             cartridge i.d.  of 1.056 cm.

-------
                   4cm
E
E
  1400
  1200-
  1000 -
   800 -
                                                           B
   600-
   400 -
   200 -
                 6      10      14
                 FLOW  l/min
1400
1200
1000
 800
 600
400
 200
5 l/min
                                   l/min
                  2             4
        PACKING  BED DEPTH (cm)
   Figure  14.  Pressure differentials (AP)  for cartridges containing Tenax GC (60/80)
              with a cartridge i.d. of 0.5 cm.

-------
                  1400 r
00
1400
                                 6       10      14
                               FLOW  (  l/min )
                                                              1000
                                                               600
                                                               200
B
           2468
          RACKING  DEPTH  (cm)
                  Figure 15.   Pressure differentials (AP)  for  cartridges containing Tenax GC  (60/80)
                              with a cartridge i.d. of 1.056 cm.

-------
                                 8 cm
   l4CX)r
   1000-
E
E
Q.
<
   600-
   200-
              I
                                         5cm
                                        3cm
                                        Icm
1
             4       8      12      16
                  FLOW  ( l/min )
       20
               1400,-
               1000-
               600-
               200-
                                                                B
                                                                      18 l/min

                                                                         15 l/min    9 |/min
                                                                        D          o
                                                 5 l/min
                      j
2468
PACKING  BED DEPTH (cm)
    Figure 16.  Pressure differentials (AP)  for  cartridges containing Chromosorb  101  (100/120)
               with a cartridge i.d.  of 1.056 cm.

-------
  1400
  1000
E
E
a.
<
   600
   200
             18 l/min
1000
 600
 200
           0.5     1.0   1.5
       CARTRIDGE  DIAMETER
             (I.D.), cm
              B
           1400
           1000
           600
            200
                                                                     A 9  l/min
                                                                 5  l/min
          0.5
     CARTRIDGE
1.0    1.5
DIAMETER
          ( I.D.) . cm
     0.5    1.0    1.5
CARTRIDGE DIAMETER
      (ID.), cm
  Figure 17.  Effect of particle mesh range  on pressure differential and relationship to

              cartridge diameter.  BPL carbon  (12/30), Tenax GC (35/60) said Tenax GC (60/80)

              are given by A,  B, and C,  respectively.  Packing bed depths were 5 cm (A),

              3 cm (B) , and 3  cm (C) .

-------
(12/30 mesh) were used.  Similarly, a mesh range of 100/120 in 1.056 cm




diameter cartridges yielded large AP values (Figure 16).   Figure 17 and




18 summarizes the effect of cartridge diameter and particle size on pres-




sure drop respectively.  At a flow of 10 1/min AP doubles when 35/60 and




60/80 mesh ranges are compared.  It increases almost an order of magnitude




from mesh 12/30 to 60/80.  These data taken collectively indicated that




cartridge diameters of 0.5 cm and containing mesh ranges of >^ 35/60 pre-




cluded their use in any studies which would require field sampling rates




> A 1/min; likewise, a mesh of 100/120 and bed diameter of £ 1.056 cm would




not be suitable.




     Pressure differentials developed across cartridges while drawing air




through the sampler were also measured (Figures 19-24).  The general trends




observed in previous AP experiments were also apparent in these studies.




However, the restrictions on sampling rates were even greater.




ESTIMATION OF BREAKTHROUGH DURING FIELD SAMPLING




     On the basis of results obtained for collection efficiencies, thermal




desorption (Section VII) and pressure drop measurements, Tenax GC and




Chromosorb 101 were selected for further study as possible candidate ma-




terials for collecting hazardous vapors.  Field sampling experiments were




designed to provide relative breakthrough data for volatile vapors, i.e.




to determine when the mass transfer zone has moved to  the end of the




available packing bed.  The estimation of elution volume required for break-




through to occur provided a further assessment of sorbent materials and




their utility for collecting carcinogenic vapors from  ambient atmosphere.




Furthermore, -the packing bed depth required to minimize breakthrough yielded




information useful for designing reliable cartridge samplers.

-------
   700r
   600
   500
E  400
E
a.
o
I
LU
o:
Q.
   300
   200
   100
           TUBE DIAMETER,  1.06 cm
           PACKING  DEPTH,  5 cm
                   TENAX  GC 60/80  U.S. MESH
TENAX  GC  35/60
   U.S.  MESH
                                  BPL CARBON   12/30
                                        U.S. MESH
                      FLOW  RATE , liters /min

    Figure 18.  Effect of particle size on pressure differential.
                               52

-------
Ul
OJ
                X
                £
                E
                a.
                <
                   500
                  400
                   300
                   200
                   100
                          2    4    6   8   10   12
                             FLOW  RATE  (l/min)
5OOr
400 -
300-
200-
                                                                                               6  l/min
             23436
            PACKING  DEPTH  (cm)
                  Figure 19.  Pressure differential developed  for BPL carbon (12/30)  cartridge with an i.d. of

                             0.5 cm  during vacuum sampling.

-------
   600
   400
o>
x

E
E
   200
                                    7 cm
                                     &
          /**  6 cm

              .0

       A    /
          .,0  5 cm



A'  o^   *S 3cm
         2      6      10     14     18

            FLOW   RATE  (  l/min )
600
400
200
y
                                                                              14  l/min
                                                                              10  l/min
           2468

       PACKING  BED DEPTH  (cm)
  Figure  20.  Pressure differentials for BPL carbon  (12/30) for air drawn through a

             1.056 cm i.d. cartridge

-------
  600 r
                                            600r-
  400
c*
X

E
E
o.
<
  200
                    3 cm
                         2 cm
                              cm
              4        8        12

              FLOW RATE  (  l/min )
                                            400-
                                            200-
                                                                           5  l/mln
                                               01234

                                                 PACKING  DEPTH  ( cm )
 Figure 21,
Pressure  differential developed  for Tenax GC (35/60)  cartridge with an i.d,

of 0.5 cm during vacuum sampling.

-------
                   600 r
Ul
                    400
                 o>
                 X

                 E
                 E
                    200
600,-
                                        2 cm

                                         D
                                            1.5 cm
                                             11
                                            4O I cm
                       0246

                           FLOW  RATE ( l/min )
400
200
4  l/min
O
    01234

       PACKING  DEPTH (cm)
                   Figure  22.  Pressure differential  developed for Chromosorb 101 (60/80)  cartridge with

                              an i.d. of 1.056 cm during vacuum sampling;..

-------
   800
   600
E  400
<  200
                    7 cm
                        6 cm
                      m
                  /  0   | 5cm

          26       10      14      18
             FLOW  RATE  (l/min)
                                              800
600

                                             20°
                                                             B
                                                                            7 l/min


                                                                            6 l/min
                                                                            4 l/min
                                                        2468
                                                     PACKING  BED DEPTH (cm)
 Figure 23.  Pressure differential developed  for Tenax GC (35/60)  cartridge with an

            i.d.  of 1.056 cm during vacuum sampling.

-------
                      600 r
                      400
oo
                    X
                    E
                    E
                    o.
                    < 200
                                      7 cm
           600
                                                     cm
                                4      8       12
                              FLOW  RATE  (  l/min )
16
          400
           200
                                                                                               6 l/min
                                         5  l/min
                          l/min
   2468
RACKING BED DEPTH  (cm)
                     Figure 24.  Pressure differential for  Chromosorb 101 (60/80) for air  drawn  through

                                a 1.056 cm i.d. cartridge.

-------
Experimental




     To determine ET and EV a dual tandem cartridge arrangement and the




sampling system described in Section VIII were used.  At the beginning




of each sampling period a known quantity of vapor (methyl ethyl ketone,




phenyl methyl ether and/or nitrobenzene) was introduced directly into the




entrance corridor of the front cartridge.  Periodically (15 and 30 min




intervals) the back cartridge was replaced with a virgin one and the back-




up cartridge was examined by thermal desorption-glc (Section VII).  At the




end of the experiment the front cartridge of the duo-series was also analyzed.




Peak areas for each standard were calculated to determine the amount of vapor




disappearing from the front cartridge and appearing in the second, backup




cartridge.  The ET and EV values of four unidentified constituents (to be




identified by glc-ms) present in ambient air were also determined.




     Breakthrough on total "pollutant profile" was also evaluated by using




the duo-tandem cartridge arrangement.  The area under the chromatographic




peaks developed from a glc program run (75-165°C, 10°/min, 12 ft 2% DECS




column) for each cartridge was estimated with a planimeter and expressed




as a percent of the total area for each set of tandem cartridges.




Results and Discussion




     Table 9 presents the results observed for packing bed dimensions of




1.056 cm i.d. x 1 cm in length (front and back).  No methyl ethyl ketone




(MEK) or nitrobenzene (NB) could be detected in either the front or backup




cartridge at any time during sampling.  On the basis of these observations




it was concluded that breakthrough (and possible low collection efficiencies)




had occurred in less than 15 min.  On the otherhand PI, P2, P3 and P4 was




detected and appeared to remain relatively constant in the backup cartridge




throughout the entire sampling period.  Since P1-P4 were endogenous interfering
                                     59

-------
           Table 9.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                         (35/60)a - 1 CM BED DEPTH
Sampling Volume
Time (Min)









Front
15
45
75
105
165
195
225
255
315
Cartridge
Air Sampled
(1)
180
540
900
1260
1980
2340
2700
3060
3780

Peak Area
MEK
NDC
ND
ND
ND
ND
ND
ND
2.8
ND
ND
PI
1.
1.
2.
1.
3.
1.
2.
2.
-
9.

7
8
2
9
0
9
0
7

2
P2
1.
1.
2.
4.
4.
3.
1.
1.
-
11.

9
9
5
0
1
4
9
6

5
(Cm2)b
P3
2.
1.
3.
6.
4.
3.
4.
3.
-
13.

7
9
0
5
8
7
0
7

5
P4
1.9
0.7
1.0
1.9
2.6
1.8
2.2
2.2
-
10.0
NB
ND
ND
ND
ND
ND
ND
4.3
ND
ND
ND
lacking bed Dimensions - Front Cartridge:   1 cm dia. x 1 cm in length,
                          Backup Cartridge:  identical to front sampler.
 Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4 are unidentified peaks appearing at 100, 115, 135, and
 165°C, respectively in chromatogram of air sample.
CND = not detected.
constituents in ambient air and were probably present during the entire

sampling period (frontal analysis),  it could not be concluded whether

these results reflected low collection efficiencies, early breakthrough or

both.

     Because early breakthrough was suspect in the previous experiment,

another study was conducted which utilized a 1.056 i.d. x 1 cm long and

a 1.056 x 3 cm long front and back cartridges, respectively.  Table 10

shows these results.   In this case early breakthrough for MEK and NB was
                                     60

-------
         Table 10.   ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                       (35/60)a - 1 AND 3 CM BED DEPTHS
Sampling Volume
Time (Min)
15
30
60
90
120
150
180
210
240
Front Cartridge
Air Sampled
(1)
180
360
720
1080
1440
1880
2160
2560
2880


MEK
NDC
6.6
ND
ND
d1.2
2.0
ND
ND
ND
ND
Peak Area
PI P2 P3
1.8 5.3 7.3
1.6 3.7 4.8
2.6 2.7 5.3
2.6 5.3 6.3
3.2 5.1 8.0
3.0 6.3 8.5
_
_
_
4.9 6.0 5.4
(Cm2)b
P4
2.5
1.5
2.0
1.4
2.7
3.1
-
-
-
26.0

NB
12
13
14
19
25
26
30
24
20
2.5
 Packing bed dimensions - Front Cartridge:   1 cm dia. x 1 cm in length
                          Backup Cartridge:  1 cm dia. x 3 cm in length
 Sampling rate - 12 1/minj sampling location - Res. Tri. Prk.
bSee Table 1 for description of PI, P2, etc.
CND = not detected.
     be a contaminant peak.
detected which is consistent with the previous conclusions.  Since the air

sample was drawn through the cartridge, a vacuum gradient was created

across the packing bed.  This situation probably leads to vacuum desorp-

tion or "stripping" of the analyte and therefore early breakthrough of

solute vapors.  In effect a smaller EV occurs for the backup cartridge

than for the front cartridge (the pressure differential across increasing

packing bed lengths are shown to be nonlinear) .

     Tables 11 and 12 present EV characteristics for packing bed lengths
                                                               3
of 2 and 3 cm.  Breakthrough occurred after approximately 1.5 m  of air


                                    61

-------
         Table 11.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                          (35/60)a - 2 CM BED DEPTH
Sampling Volume Air Sample
Time (Min) (1)
15
45
75
105
135
165
195
225
255
180
540
900
1260
1620
1980
2340
2700
3060
id
PI
2.0
2.9
2.2
2.6
4.8
3.9
3.7
2.4
4.7

P2
0.8
3.1
1.6
3.3
5.2
3.6
3.2
6.5
6.0
Peak
P3
3.0
5.1
6.6
4.0
8.4
6.6
8.2
7.9
9.6
Area (Cm2)b
P4
2.2
1.9
1.6
1.4
3.5
2.8
3.3
3.2
5.0
NB
NDC
ND
ND
ND
75.6
47.2
40.0
40.0
35.0
lacking bed dimensions - Front Cartridge:   1 cm dia. x 2 cm in length
                          Backup Cartridge:  same as above
 Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4 are described in Table 6.
CND = not detected


was sampled for a 2 cm bed length of sorbent.  Increasing the bed depth

to 3 cm prevented breakthrough; all of the NB was detected in the front
                                        3
cartridge.  The EV for MEK was about 2 M  for a 1 x 3 cm cartridge.

    Two sampling rates 12 and 24 1/min were compared to determine whether

the elution volume characteristics were flow dependent.  The results shown

in Tables 13 and 14 indicated that breakthrough is relatively independent

of sampling rate when phenyl methyl ether (PME) and nitrobenzene (NB) were

used as the test vapors.  Analysis of the front cartridge revealed the

presence of PME and NB and therefore breakthrough had not occurred.
                                    62

-------
          Table 12.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                         (35/60)a - 3 CM BED DEPTH
Sampling Volume Air Samp]
Time (Min) (1)
15
45
75
105
135
165
195
225
255
Front Cartridge
180
540
900
1260
1620
1980
2340
2700
3060

Led
MEK
NDC
ND
ND
ND
ND
18.0
25.0
15.0
ND
ND
2
Peak Area (Cm )
PI
1.8
1.2
1.6
1.5
1.7
1.7
1.6
0.8
1.0
9.6
P2
2.5
2.4
2.2
4.2
4.5
3.8
3.9
3.4
3.7
11.5
P3
2.6
1.4
3.7
7.6
4.3
5.0
5.4
4.0
4.7
20.4
,b
P4
1.2
0.8
1.3
1.5
1.9
1.9
1.8
1.5
2.2
5.7

NB
ND
ND
ND
ND
ND
ND
ND
ND
ND
8.4
 i'acking bed dimensions - Front Cartridge:   1 cm dia. x 3 cm in length
                          Backup Cartridge:  same as above
 Sampling rate - 12 1/min; sampling location - Res. Tri. Prk.
bSee Table 6 for description of PI, P2, etc.
CND = not detected

     Using a flow of 24 1/min, 4 cm and 6 cm beds of Tenax GC (35/60)  were

evaluated and breakthrough did not occur for NB even after 15 m  of air

had been drawn through the front cartridge and analysis of the front car-

tridge by glc confirmed the presence of NB (Tables 15 and 16).  On the
                                                       3
otherhand, PME appeared to breakthrough after about 9m  of air had been

sampled through a 6 cm bed of Tenax GC packing but not for a 4 cm bed.

This was attributed to the higher vacuum experienced with a 6 cm bed to

maintain a 24 1/min flow rate and thus "vacuum stripping" of PME probably

occurred.
                                     63

-------
         Table 13.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                (35/60)a - 2 CM BED DEPTH AND 12 L/MIN
Sampling Volume Air Sam]
Time (Min) (1)
3 led

Peak Area (Cm^

PI
15
30
60
90
120
150
180
210
240
270
Front Cartridge
170
350
710
1060
1420
1770
2120
2480
2830
3190
3190
1.
1.
3.
3.
2.
3.
3.
1.
1.
1.
3.
3
3
2
1
5
2
4
9
2
4
6

PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7.2

P2
1.9
1.9
2.3
4.9
4.1
3.0
2.9
0.8
1.1
2.6
6.4

P3
4
5
2
10
11
3
3
1
1
6
12


.1
.5
.1
.5
.0
.8
.0
.8
.6
.8
.6

P4
1.9
1.5
1.5
3.6
5.6
3.0
3.3
0.9
1.1
3.7
So.o
)b



P5
0
0
0
1
2
1
1
0
1
1
9
.9
.6
.6
.3
.8
.8
.8
.9
.0
.7
.6


NB
NDC
ND
ND
ND
ND
ND
ND
ND
ND
ND
10.8
 Packing bed - Front Cartridge:   1.5 cm i.d. x 2.0 cm in length.
               Backup Cartridge:  1.5 cm i.d. x 3.0 cm in length.
 Sampling rate - 12 1/min; sampling location - Res. Tri, Prk.
bPl, P2, P3, P4, P5 were peaks eluting at 81°, 95°, 110°, 135°, and
 165°, respectively; see text for glc conditions.
CND = not detected

     A comparison of the "pollutant profile" from Los Angeles, CA. for

the front and backup cartridges were also made to evaluate overall break-

through for Tenax GC and Chromosorb 101 sorbents during field sampling.

These results are given in Table 17.  Early breakthrough from the front

cartridge resulted when Chromosorb 101 was used as the collection media.

In contrast, more than 90% of the "profile" collected on Tenax GC was
                                        3
still in the front cartridge after 5.6 m  of air had been sampled.
                                    64

-------
         Table 14.   ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                (35/60)a - 2 CM BED DEPTH AND 24 L/MIN
Sampling Volume Air Samp
Time (Min) (1)
led
Peak

PI
15
30
60
90
150
180
Front Cartridge
350
710
1420
2120
3540
4250
4250
1
1
2
1
1
2
1
.1
.7
.4
.4
.2
.5
.5

PME
ND
ND
ND
ND
ND
ND
13.0

P2
2.0
2.8
3.8
2.4
2.1
3.4
4.6

P3
3
7
8
4
5
10
8
Area


(cm2)b

P4
.4
.2
.8
.4
.7
.0
.8
2
3
4
4
4
2
9
.1
.1
.9
.8
.5
.1
.6

P5
1.1
1.1
1.9
2.0
1.6
0.9
5.0

NB
NDC
ND
ND
ND
ND
ND
3.4
T'acking bed - Front Cartridge:   1.5 cm i.d. x 2.0 cm in length
               Backup Cartridge:  1.5 cm i.d. x 3.0 cm in length
 Sampling rate - 24 1/min; sampling location - Res. Tri. Prk.
bPl, P2, P3, P4, P5 were unidentified peaks eluting at 80°, 95°, 110°,
 135°, and 165°, respectively; see text for glc conditions.
CND = not detected

-------
         Table 15.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                        (35/60)a- 4 CM BED DEPTH
Sampling Volume Air Samp
Time (min) (1)
30
60
90
120
150
180-
210
240
270
300
330
360
390
420
450
480
Front Cartridge
710
1420
2120
2830
3540
4250
4960
5660
6370
7080
7790
8500
9200
9910
10620
11330
11330
led
Peak
PI
2
1
2
0
3
2
2
3
1
3
2
2
2
2
2
2
1
.0
.7
.6
.8
.8
.9
.8
.4
.2
.5
.1
.0
.9
.0
.1
.8
.6
PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
7.4
P2
2.4
1.3
3.7
1.4
4.5
3.4
4.5
3.3
1.0
4.5
2.4
2.6
3.1
2.8
3.2
3.3
2.0
P3
5
2
7
3
7
7
10
7
2
10
5
5
7
7
6
8
4
Area (cm )

.8
.1
.0
.1
.7
.5
.5
.0
.4
.5
.2
.2
.5
.6
.7
.5
.6
P4
4.5
1.3
4.3
1.8
4.5
8.5
14.0
2.8
1.6
8.1
2.6
5.5
9.1
10.5
9.0
5.4
5.8
P5
1
0
2
0
1
3
3
1
0
3
1
2
4
3
4
2
4
.3
.7
.8
.5
.5
.7
.8
.8
.7
.5
.3
.4
.0
.6
.3
.1
.8
NB
NDC
ND
ND .
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
3.4
 racking bed - Front Cartridge:  1.5 cm i.d. x 4 cm in length
               Back Cartridge:   1.5 cm i.d. x 3 cm in length
 Sampling rate - 24 1/min; location - Res. Tri. Prk.
 See previous Table for explanation of PI, P2, etc.
CND = not detected.
                                    66

-------
Table 16.  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
              (35/60)a- 6 CM BED DEPTH
Sampling Volume Air Samp]
Time (rain) (1)
30
60
90
150
210
270
300
330
360
390
420
480
540
570
600
660
720
750
810
870
930
960
350
710
1060
1770
2480
3190
3540
3890
4250
4600
4960
5660
6370
6730
7080
7790
8500
8850
9560
10270
10970
11330
.ed
Peak Areas
PI
0
0
0
1
1
2
0
1
1
2
3






2
3
3
1
3
.84
.68
.50
.40
.68
.64
.55
.35
.20
.40
.38
-
-
-
-
-
-
.53
.00
.18
.62
.00
PME
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
10.8
3.0
ND
ND
ND
P2
2.5
1.5
1.4
2.9
3.0
4.9
0.9
0.9
1.2
1.9
1.5
4.8
1.2
2.1
3.2
2.2
1.7
1.0
3.1
3.2
1.7
3.0
P3
4
2
1
4
5
5
2
2
2
5
3
10
3
6
9
4
4
2
8
8
4
6
(cm
V
P4
.1
.8
.9
.5
.4
.5
.0
.0
.3
.5
.8
.5
.1
.5
.3
.8
.9
.3
.4
.2
.5
.2
2
4
1
2
1
2
1
0
3
2
1
6
1
3
7
1
2
1
5
7
4
6
.3
.6
.5
.4
.9
.8
.8
.9
.6
.6
.7
.3
.1
.7
.2
.8
.2
.9
.2
.5
.7
.8
P5
0.4
0.7
0.4
0.6
0.6
0.6
0.6
0.6
1.3
1.0
0.8
2.2
1.0
0.8
2.4
1.5
1.5
1.0
1.7
2.6
1.7
3.5
NB
NDC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
                           67

-------
    Table 16 (continued).  ELUTION VOLUME CHARACTERISTICS FOR TENAX GC
                          (35/60)a - 6 CM BED DEPTH
Sampling Volume Air Sam]
Time (min) (1)

1020
1080
1140
1200
1260
Front Cartridge

12040
12740
13450
14160
15000
15000
)led

PI
2.40
1.75
1.75
0.84
1.00
1.20
fy u
Peak Areas (cm )

PME
ND
ND
ND
ND
ND
ND

P2 P3 P4 P5 NB
2.0 4.1 2.5 3.0 ND°
2.4 3.8 6.0 3.2 ND
2.2 4.9 8.2 2.8 ND
1.7 4.0 2.1 2.0 ND
- ND
- 4.8
lacking bed - Front Cartridge:  1.5 cm i.d. x 6 cm in length
               Back Cartridge:   1.5 cm i.d. x 3 cm in length
 Sampling rate - 24 1/min; location - Res. Tri. Prk.
bSee"Table 14 for explanation of PI, P2, etc.
CND = not detected
                                    68

-------
                            Table  17.  POLLUTANT PROFILE BREAKTHROUGH DURING AMBIENT AIR SAMPLING3
vo
Sorbent Cartridge
Chromosorb 101 (60/80)
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Front
Backup
Tenax GC (35/60)
Front
Backup
Front
Backup

1
1
1
1
1
1
1
1
1
1
1
1

1
1
1
1
Volume Air
Dimensions Sampling Time (Min) Sampled (1)

.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5

.5
.5
.5
.5

X
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X

3
3
3
3
3
3
1
1
1
1
1
1

3
3
3
3

.0
.0
.0
.0
.0
.0
.5
.5
.5
.5
.5
.5

.0
.0
.0
.0

cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm
cm

cm
cm
cm
cm

30
30
60
60
90
90
30
30
60
60
120
120

60
60
120
120

450
450
1340
1340
3970
3970
400
400
1400
1400
5400
5400

1260
1260
5600
5600
Percent of
Combined Areab

75
25
40
60
40
60
90
10
55
45
40
60

90
10
95
5
              Experiments were performed in Santa Monica, CA.
            bThe  total area of the chromatogram between 75-160°  (lp°/min)  on  a 12  ft  2% DECS column was
              determined using a planimeter.

-------
                               SECTION VI


         GAS CHROMATOGRAPHIC INLET-MANIFOLD FOR SAMPLE ANALYSIS



     Because of the limited sensitivity of currently available detectors


hazardous substances need to be concentrated from highly dilute samples.


A step toward the solution of this problem is achieved when cartridges


containing an appropriate sorbent is used and large volumes of air are


forced or drawn through the sampling device whereupon the pollutants


are trapped.  The performance (collection efficiency, pressure differen-


tial, elution volume) characteristics for a selected number of sorbents


was described in Section V.


     Even when employing a cartridge technique, only trace quantities of


hazardous pollutants are accumulated; thus, it is imperative that the


entire sample be submitted for analysis.  The levelo of carcinogenic and


mutagenic vapors (e.g. epoxides, nitrosamines, sulfonates, sulfites, sul-


tones, aldehydes, ketones, 6-lactones, chloroalkyl ethers and nitro com-


pounds) to be collected and identified from ambient air under this research

                                        3
program are anticipated to occur at ng/m  amounts or less.  These severe


restrictions on sampling and analysis have required the development of a


technique(s) such as thermal desorption as a means of transferring the


entire amount of trapped vapors from the cartridge to the analytical system.


     This section describes (1) the design and fabrication of an inlet-


manifold for effecting desorption of vapors and efficient transfer of pol-


lutants to a glc, and (2) the heating characteristics of selected sorbents


using the inlet-manifold.


DESIGN AND FABRICATION OF INLET-MANIFOLD


     Thermal desorption systems of varying designs have been des-


cribed     '  '     for recovering trapped vapors and their subsequent


                                    70

-------
                                   46
analysis by glc and glc-ms.  Duel's   solvent-free sample inlet system



was designed so that with the sample fraction "in-line" with the sample,



the entire mixture could be introduced directly onto the glc column.



This concept was retained for the development in this research program



of an interface system for analyzing carcinogenic vapor samples by glc



and glc-ms.



     The described manifold design evolved from a consideration of



several criteria.  These were:  (1) the heat transfer characteristics of



cartridge samplers containing sorbents of various physical dimensions,



(2) the flow rate requirements for purging desorbed vapors from the cart-



ridge sampler and for packed and glc capillary columns, (3) the tempera-



tures necessary for effecting thermal desorption of trapped carcinogenic



vapors from sorbents (from Section VII), and (4) the ability to con-



veniently interchange and assemble on any standard glc or glc-ms instrument,



     The fabricated inlet/manifold system (Fig. 25) consisted of four main



components:  a desorption chamber, a six-port two position high temperature



low volume valve (Valco Inc., Houston, Tx.), a Ni capillary trap, and a



temperature controller.  One of the three prototype brass thermal desorp-



tion chambers constructed for use in this program is shown in Figure 26.



Brass was chosen for its good conductive properties.  The configuration



of the chamber was designed to allow an inert gas to enter through a side-



arm near the bottom and sweep up an annular space between the chamber and



glass cartridge.  This permitted the purge gas to be preheated to the



chamber temperature prior to passing down through the sorbent bed.  Two



additional prototype brass desorption chambers were designed so that the



purge gas entered near the top and passed directly through the glass



cartridge to the valve proper (Fig. 27).  The overall chamber lengths were
                                     71

-------
                         COMPRESSION SPRING
                               HEATING CARTRIOE

                                    CARRIER GAS
              TO GLC CAPILLARY



HEATING AND COOLING BATH


Nl CAPILLARY TRAP
                                                      VALVE POSITION A
                                                      (SAMPLE oeaonrriOM)
                                                   CARRIER
                                                      GAS
                                                      1
                                                       WOVE POSITION 8
                                                        (
                                                   CARRIER-.
                                                       S >"
Figure  25.   Thermal desorption inlet-manifold

-------
                                                                 i
                                10.5 cm
                                                     10.56 mm. 1.0.
                                                         1	
                                                                 r
                                                           13.0  mm.O.D.
Figure 26.  Thermal desorption chamber with annular space.  Sampling
            tube  shown  in lower figure.
                               73

-------
   OESORPTION CHAMBER
T
15.0 cm

   13.30 cm
il
                     TEFLON  INSERT
                         COMPRESSION  SPRING
                        D<	PURGE GAS
GLASS  CARTRIDGE  SAMPLER
    QT	"	""_	Q I3.0mm  ] 10.56mm

                         i~T~
    <	  10.5 cm    	M T




Figure 27. Thermal desorption chamber
                 74

-------
13.3 cm which accommodated a pyrex sampling cartridge 10.5 cm in length.



Two of the desorption chambers (Fig. 20 and 21) were designed to accept



cartridges 10.56 mm i.d. x 13.0 mm o.d.  While the third accepted a



larger cartridge of 15.6 mm i.d. x 16.5 mm o.d.



      An aluminum sandwich served as a heat sink (Fig. 25) which accepted



any one of the three desorption chambers.  Two 150 w, 115V heating



cartridges (Varian Part No. 22-0000-18-00) were used to heat the aluminum



sandwich and the temperature was monitored and controlled with a platinum



sensor probe (100ft, Varian Part No. 64-000009).  The desorbed vapors passed



via a short insulated capillary line through a six-port two position valve



which was also encased in an aluminum heating bath.  Temperature control



was identical to the thermal desorption chamber.  The electronics circuit



which controlled the temperatures on each heat sink is shown in Figure 28.



The temperature was monitored directly on a pyrometer; control was + 1°C.



      A nickel capillary (0.020 i.d. x 0.032 x 0.5 m in length) constituted



one loop of the valve proper which was cooled with liquid N~ or solid



carbon dioxide/isopropanol and served as a trap for collecting and concen-



trating desorbed vapors for their introduction into high resolution glc



columns.  The vapors were released from the capillary trap by rapidly



heating to 175° using a wax bath.



      The multiport valve used on the described inlet/manifold was chosen



for its polyimide internal stem to minimize the contact of desorbed trace



vapors with reactive metal surfaces, therefore, mimimizing contamination



or decomposition of sample constituents.



      In a typical thermal desorption cycle a sampling cartridge was placed



in the preheated (ca. 225°C) chamber, and N0 gas was purged through the
                                           /.


cartridge (ca. 20 ml/min) to purge the vapors into the liquid N9 cooled

-------
                                                                                        I  COW
                                                                                        HEATER
                                                                                              "*

Figure 28.   Electronics circuit  designed  for  temperature control on inlet-manifold system,

-------
Ni capillary trap; this constituted valve position "A" (Fig. 25).  After




the thermal desorption step was complete, the six-port valve was rotated




to position "B" (Fig. 25) and the temperature on the capillary loop was




rapidly raised (> 10°/min) whereupon the carrier gas carried the vapors




onto a glc column.




     As designed, the prototype thermal desorption chambers which were




easily interchangeable, accommodated cartridges of two different diameters




with up to 8 cm of packing (sorbent) depth.  Thus, comparisons of dif-




ferent cartridge sizes with respect to sorbent background during thermal




desorption could be made.  This inlet-manifold configuration also allowed




the desorbed vapors from one or more cartridges to be accumulated in the




capillary trap prior to analysis by glc or glc-ms.




     The inlet/manifold system described here was employed in studies on




thermal desorption of vapors from cartridges.  Its performance characteris-




tics are further discussed in Section VII.




HEAT TRANSFER CHARACTERISTICS FOR SELECTED SORBENTS AND THERMAL DESORPTION




CHAMBER




     An investigation was made of the heating rates for candidate sorbent




media.  Using the measured heat transfer coefficients as a guideline, the




required heating period and temperature for effecting quantitative desorp-




tion was selected (Section VII).  Furthermore, two prototype desorption




chamber designs with and without an annular space  (Fig. 26 and 27) were




compared with regard to the rate of heat transfer to the sorbent bed.




Experimental




     Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 104




(100/120) and Chrouosorb W-HP (100/120) were purchased from Applied Science,




State College, Pa.  Stationary phases chemically bonded to supports which
                                      77

-------
included carbowax 400/Poracil C (100/120) and oxopropionitrile/Poracil C




(80/100) were also obtained from Applied Science.  Carbowax 600 and didecyl




phthalate stationary phases and the sorbent Porapak Q were from Supelco,




Inc., Beliefonte, Pa.




     Carbon derived from coke (PCB, 12/30) was acquired from Pittsburgh




Activated Carbon Division of Calgon Corp., Pittsburgh, Pa.  A cocoanut




activated carbon (580-26) was purchased from Barneby Cheney, Columbus,




Ohio.




     The temperature rise times in the sorbent bed of the cartridge was




measured with a calibrated thermocouple (mV ys °C); signal output was




directly recorded on a strip chart recorder (Varian Model A-25, Varian




Instruments, Walnut Creek, CA.).  The rate of temperature increase in




the sorbent was determined using preset and isothermal desorption chamber




temperatures.  The rate of temperature increase was also monitored on the




inner glass wall of the cartridge.




Results and Discussion




     Prior to defining the parameters (temperature, heating time) for ef-




fecting thermal desorption of vapors trapped on sorbents (Section VII), the




rate of heat transfer from the thermal sink to the sorbent bed was measured




using the inlet-manifold interface depicted in Figure 25.  Operating the




desorption unit under isothermal conditions, the temperature increase in




the sorbent bed was monitored immediately after inserting a cartridge.




     Figure 29 depicts the thermocouple response time, and the heating rate




in the center of a cartridge containing Tenax GC using one of the previously




described thermal desorption chambers (Fig. 27).  Because the response of




the thermocouple significantly contributed to the temperature rise profile




(8 sec required to reach 63% of the upper temperature limit), its contribution
                                     78

-------
o
o
UJ

-------
was subtracted from all measurements.  These temperature rise times




(Fig. 29) indicated a cross-sectional gradient was produced immediately




after inserting the cartridge into the chamber.  The periphery of the




sorbent bed reached the temperature maximum in approximately 4 min, while




the center of the sorbent bed required an additional 2 min.  Furthermore




increasing the cartridge diameter from 1.0 to 1.5 cm increased the tem-




perature differential by a factor of 1.3.




     The heat transfer coefficient for sorbents and similarly their observed




differential temperature gradient, varied considerably.  The relative tem-




perature rise times for several sorbents which previously were shown to




have good collection efficiencies (Section V) were compared.  The order of




their heating rates were observed to be:  PCB and BPL carbon (12 x 30)




> oxopropionitrile and carbowax 400 chemically bonded to Poracil C (100/120)




> Chromosorb 104 (100/120) > Tenax GC (60/80) and Chromosorb 101 (100/120).




These results are depicted in Figure 30 which were obtained with a thermal




desorption chamber designed to have an annular space between the chamber




wall and glass cartridge.  The largest differences in heating rates (in the




center of the sorbent bed) were exemplified by the activated carbons and




Tenax GC (Fig. 30).




     Because inspection of the heating rates for Tenax GC (60/80) with and




without an annular space indicated relative differences (Fig. 29 and 30),




a more detailed study was conducted comparing these two thermal desorption




chamber designs.  Figure 31 presents this comparison.  This data clearly




shows that a 1 mm annular space in the chamber reduced the heating rate of




the inner glass cartridge wall as well as the center of the packing by at




least a factor of two.  Furthermore, a larger heating gradient was observed




between the center and periphery of the sorbent bed with the annular spaced
                                     80

-------
                                     2461-
oo
                                     25
                                                                                O= Chromosorb 104
                                                                                A= Chromosorb 101
                                                                                D= Tenax GC
                                                                                 O= BPL carbon
                                                                                A= PCB carbon
                                                                             D-A= oxypropionitrile/Poracil C 80/100
                                                                                • « 20% carbowax 600 on Chrom W (100/120)
                                       01       234       56
                                                             TIME   ( MIN )
                                  Figure  30.   Temperature rise  times in sorbent bed using
                                              annular  spaced  chamber.

-------
                              197 r
00
NJ
                                                          ,-D-r£T.-A-
                                                               —. • A
                                                                           temperature  in  center of cartridge,
                                                                           with (O)  and without (A) annular space
                                                                       B « temperature  of  inner glass wall of
                                                                           cartridge, with (/\) and without
                                                                           annular space
                                                                       Thermocouple response given by0
                                 0        1234567
                                                            TIME  (  MIN )

                            Figure 31.  Comparison of  temperature rise times for chambers with and
                                        without an annular space.  Chamber was at 175°C.

-------
chamber than the one providing direct contact between the brass wall and




glass cartridge.




     In view of the results obtained with the prototype desorption chamber




with an annular space (Fig. 26), the heating rates for candidate sorbent




media were also investigated with a chamber providing direct contact be-




tween the glass cartridge and chamber wall (Fig. 27).  In general the




trend was the same as previously observed (compare Fig. 30 and 32) except




all heating rates were significantly greater in this case.




     Since the heating rate was generally 1.7 times faster when the cart-




ridge was in direct contact with the brass desorption chamber wall and a




smaller temperature gradient was observed through the packing bed, it was




concluded that a chamber incorporating a 1 mm annular space 90% of the




length of the cartridge was not a feasible design for application to desorp-




tion of vapors.  Thus, the prototype desorption chamber which introduced the




carrier gas near the top of the chamber (Fig. 27) was employed in thermal




desorption of vapors trapped on sorbents.




     The temperature rise time for Tenax GC was also examined using three




isothermal conditions on the desorption unit (Fig. 33).  The heating rate




(slope) was only slightly increased by increasing the desorption unit




temperature; however, it was evident that the rate was linear upto 65% of




the final temperature or during the first 75 sec.  Thereafter, an additional




several minutes was required to reach a plateau.  These data, therefore,




indicate that the desorption unit should be set at a temperature which allows




the attainment of the required desorption temperature in 60-90 sec after




insertion of the cartridge sampler.




     When a thermal desorption chamber design permits rapid heating of the




cartridge sampler with a minimum of temperature gradient across the packing
                                     83

-------
220 r
                              3        4
                              TIME  (MIN.)
  Figure 32.   Comparison  of  heating  rates for some sorbents.  Curves
              A,  B,  and C correspond to PCB carbon (12/30), oxopro-
              pionitrile  on  Poracil  C (80/100), respectively.  Thermal
              desorption  chamber was isothermal at 210°C.
                               84

-------
  230 r
  205 -
                                   A = 180°C
                                   B = 205°C
                                   C = 228°C
      0              24               6
                      TIME  (MIN)
Figure 33.   Heating  rate for Tenax GC at different  isothermal
            desorption chamber temperatures.
                           85

-------
bed then the desorbed vapors may be introduced directly onto a conven-




tionally packed glc column.  The carrier gas flow rate was satisfactory for




efficiently purging the desorption chamber and maintaining column resolu-




tion.  Thus, the desorption chamber can be "in-line" with the glc column




for a time sufficient to desorb substituents with the greatest adsorption




affinities and then returned to the "by-pass" mode during the remainder




of the chromatographic period.  Under these conditions the background intro-




duced during heating of the sorbent can be minimized as well as artifactual




processes resulting from decomposition, polymerization, etc. of solute vapors.




The use of cartridges of <_ 1.0 cm i.d. and the prototype desorption chamber




shown in Figure 27 allowed the inlet-manifold to be operated in the manner




described.




     With larger diameters, the temperature gradient in the sorbent bed was




too great.  In this case, a cartridge "in-line" with the glc column during




the desorption cycle produced excessive solute band spreading and sample




resolution was decreased.  It was concluded that desorption was not uniform




across the sorbent bed because of the large temperature gradient which pro-




bably accounted for the loss in glc resolution.




     When sampler cartridges of > 1.0 cm i.d. were subjected to thermal




desorption or high resolution capillary columns were employed, the vapors




were concentrated in a small carrier gas volume in order to prevent the




excessive band spreading and decreased column efficiency.  This was achieved




as previously described by retrappiiig desorped vapors in a Ni capillary




(0.020 in i.d. x 0.5 m length) using liquid N • as the coolant.  After the




desorption period, the carrier gas was routed through the capillary trap




(Fig. 25) and the trap rapidly heated.
                                     86

-------
                               SECTION VII



       THERMAL DESORPTION OF HAZARDOUS VAPORS FROM SOLID SORBENTS






     The recovery of vapors adsorbed or absorbed on various sorbents by


       "I ft 01 OQ           / *% 7                                    /A

thermal     '   and vacuum     desorption has been reported.  Duel



employed a combination of vacuum-thermal stripping to remove pollutants



adsorbed to cocoanut charcoal.  The sample was heated from ambient to



170°C at 10°/min; partial fractionation was accomplished prior to glc


                 32
analysis.  Damico   also used cocoanut charcoal to trap glc fractions



for ms analysis.  Desorption of propionaldehyde and 2-nonanone occurred



at room temperature and 70°C, respectively, when loaded capillaries were



introduced into a high vacuum of the mass spectrometer.  Other investiga-



tors also reported the thermal desorption of vapors from charcoal;  *  '  '



however, analysis of pollutants had been restricted primarily to aliphatic



and relatively nonpolar aromatic compounds.



     The high surface activity of activated carbons has been reported to



produce artifacts during recovery.  Formaldehyde decomposes, as does methyl



ethyl ketone which forms diacetyl compounds and acetic acid.  Furthermore,



compounds sensitive to polymerization or decomposition on carbon are also


                                 20 21
generally sensitive to carbosieve  '



     Desorption of semi-polar and polar compounds by thermal means has been


       .,  i         j r      i    *  ^  j 15,17,20,21,38,40   „.,,,   17
successively achieved from polymeric beads                  .  Williams



used temperature programming up to 210°C to elute trace contaminants from


                                 38
Porapak columns.  Leggett, e_t^ al_.   reported significant amounts of con-



taminants from Porapak was produced if the temperature exceeded 110°C



during analysis.  Similar results were obtained by Krumperman   when



Porapak Q cartridges were heated above 170°C.
                                    87

-------
                                 20 21
     In contrast, Zlatkis, e£ jal.  '   were able to desorb many volatile



polar urine metabolites as well as atmospheric pollutants from Tenax GC



at 300°C.  The background from this polymer was extremely low.



     Thermal desorption of trace vapors absorbed on liquid phase coated


                                          31
beads was extensively employed by Williams   for analyzing atmospheric



pollutants.



     Although there are many reports on the use of thermal desorption



as a means for recovering and introducing the vapors into a glc, a thorough



study has not been made on the quantitative aspects of this method for



semi-polar and polar chemical classes of compounds such as carcinogens.



     This section discusses (1) the background contribution from sorbents,



(2) the desorption parameters (temperature, time) for effecting recovery



for sorbents, (3) the quantification of the thermal desorption step, and



(4) the percent recovery of hazardous substances of interest to this research



program from Tenax GC.



EXPERIMENTAL



     Tenax-GC (60/80 mesh), Chromosorb 101 (100/120), Chromosorb 104 (100/120)



and Chromosorb W-HP (100/120) were purchased from Applied Science, State



College, Pa.  Stationary phases chemically bonded to supports which included



carbowax 400/Poracil C (100/120) and oxopropionitrile/Poracil C (80/100)



were also obtained from Applied Science.  Carbowax 600 and didecyl phthalate



stationary phases and the sorbent Porapak Q were from Supelco, Inc., Belle-



fonte, Pa.



     Carbon derived from coke (PCB, 12/30) was acquired from Pittsburgh



Activated Carbon Division of Calgon Corp., Pittsburgh, Pa.  A cocoanut



activated carbon (580-26) was purchased from Barneby Cheney, Columbus,



Ohio.
                                    88

-------
     All sorbents were thermally conditioned 10°C below the maximum recom-




mended temperature limit for at least 12 hr under approximately 20 ml/min




of He flow.  After sorbents were packed into glass cartridges they were




conditioned again for 15 min in the thermal desorption unit prior to use.




     The standards-ethyl methanesulfonate, 3-propiolactone, N-nitrosodiethyl-




amine, 1,2-dichloroethyl ethyl ether, nitromethane, methyl ethyl ketone,




and aniline - were from Fisher Chemicals, Pittsburgh, Pa.  The source of




glycidaldehyde and sulfolane was Aldrich Chemicals, Milwaukee, Wise.  The




supply of 1,3 propanesultone, maleic anhydride, butadiene diepoxide and




propylene oxide was from Eastman Organic Chemicals, Rochester, N. Y.




Styrene epoxide, bis-(chloromethyl)ether and bis-(2-chloroethy1)ether were




acquired from K&K labs., Plainview, N. Y.




     In order to demonstrate the efficiency of collection plus thermal




desorption of trapped vapors, synthetic air/vapor mixtures were prepared




using N-nitrosodiethylamine (100 ng) as an internal standard; the quantity




of the other vapors tested was varied from 50 to 300 ng.  An aliquot of




this mixture was introduced directly through the thermal desorption cham-




ber which contained a glass cartridge packed with only glass wool and




the mixture resolved by glc.  The peak areas for each solute in this




calibration mixture (cm) was measured by triangulation and the relative




response ratios were calculated:






                                 ^cm = Ap/Ail x n* i?           (5)




where A  and A   were the areas of the solute peak and of N-nitrosamine,




respectively.  An identical aliquot of the synthetic air/vapor mixture was




purged (4.0 1/min) through a cartridge containing a sorbent using the moni-




toring system described earlier (Figure 25).  The trapped vapors were
                                    89

-------
desorbed in the thermal desorption chamber followed by glc analysis.  The



RR  for each constituent was calculated and the percent recovery was deter-
  s


mined as a ratio of the RR  values to those for RR  .
                          s                       cm
                     RR
   ,    /RR   x 100% = Percent recovery      (6)
sorbent   cm
     Gas-liquid chromatography (glc) was conducted on a Perkin-Elmer 900



series chromatograph (Perkin Elmer Corp., Norwich Conn.) equipped with dual



flame ionization detectors.  A 2.5 mm i.d. x 3.6 m silanized glass column



containing 2% DECS on Chromosorb W(HP) 80/100 mesh was used for resolving



synthetic air/vapor mixtures.  The column was programmed from 55 to 200°C



at 10°/min with an initial and final isothermal period of 2 and 10 min,



respectively.  Carrier gas (N«), hydrogen and air flow rates were 45, 30



and 250 ml/min, respectively.  The injection port, manifold and detector



temperatures were maintained at 250°C.



RESULTS AND DISCUSSION



     Prior to evaluating the desorption step of trace organics from sorbents,



the candidate sorbents were subjected to thermal desorption temperatures;



the background contribution from each was examined.  The polymeric beads-



Porapak Q, Chromosorb 101 and 104-all exhibited significant background even



when conditioned using the manufactured recommended procedures.  Of these



three, Porapak Q was the worst offender (Fig. 3A).  These results confirm


                                                1 f\ *^ft
those previously reported by other investigators  '  .  Polymeric beads were



extracted in a Soxhlet for 18 hr with acetone, methanol, or benzene and then



thermally conditioned; all attempts to reduce the background of Porapak Q



were unsuccessful.  Background was reduced significantly for Chromosorb 101



when extracted with methanol but this procedure was less successful for
                                     90

-------
              Parameters
VO
          20
                   Col - 55° - 185°
                   initial T° - 2 min delay
                   program - 13°/min
                   Flow rate - 40 ml/min
                   2% DECS Chrom W (HP) -  100/120
                   Thermal Desorption -  6' @
                   By-pass Loop - open 2'
                   Sorbent-Porapak Q
                                                            I
                                                I
18
16
14
12        10       8

   TIME (MIN)
                                                                                                              90
                                                                                        80
                                                                                                              70
                                                                                                              60
                                                                                                              50
                                                                                                                 <
                                                                                                                 o
                                                                                                              40
                                                                                                             30
                                                                                                              20
                                                                                                              10
                                                                                                                 u.
                                                                                                                 o
                                                                                                                 CJ

                                                  -nrrT-c'- of  blank Porarjak-Q cartridge.

-------
Chromosorb 104.  Although Porapak Q exhibited high collection efficiencies,




its background precluded any studies requiring trace analysis.




     On the otherhand background contamination from Tenax GC beads was



very low (Fig. 35).  Pre-extraction with methanol for 18 hr followed by




thermal conditioning at 325°C produced cartridges which allowed nanogram




quantities of hazardous vapors to be detected and quantitated.



     Sampling cartridges packed with any one of the activated carbons




also exhibited low background (Fig. 36).



     In contrast, inert glc supports coated or chemically bonded with liquid




phases gave relatively high contaminant peaks; the latter was somewhat better.




Procedures for treating these candidate sorbents to yield cartridges with low




background was not exhaustively investigated.  Further studies in this area




are warranted, especially with phases chemically bonded to supports.




     Because the desorption experiments indicated that the background con-




tribution from Tenax GC was least of the sorbents tested, and it exhibited



excellent collection efficiencies for selected hazardous substances (Section




V), the thermal recovery of vapors adsorbed to this polymer was examined.




Ten compounds which represent a broad spectrum of chemical properties and




are of particular interest in air pollution studies were chosen.  Each sub-



stance was introduced through a cartridge of Tenax GC (1.0 cm i.d. x 3.0 cm



in length)  at levels of 50, 100, 200, and 300 ng as a synthetic air/vapor



mixture.



     Initial experiments were designed to determine optimal recovery of



solutes using temperatures.  The principle reason was to use sufficient




temperature for vaporizing the trapped constituents while minimizing sor-



bent background and decomposition of labile compounds or inter-compound
                                    92

-------
  GLC Parameters



     FID-GLC

     Col. - 12 ft 2% DECS Chrom W-HP

     Program - initial hold 2 min, 70-180° @ 10°/min

     Detector - 200°

     Attenuation - 6.4 x 10~   AFS



  Thermal Desorption Parameters



     Chamber - 175°

     6-port Valve - 70°

     Cartridge Preheat Period - 5 min

     By-pass Open - 2 min
                                                                 90
                                                                 80
                                                                          70
                                                                          60
                                                                              ui
                                                                              fe

                                                                              I-
                                                                              LJ

                                                                          50 flc
                                                                          40g

                                                                              0.

                                                                              UJ
                                                                              tr

                                                                          30 oc
                                                                              UJ
                                                                              o
                                                                              tr
                                                                              o
                                                                              o

                                                                          20 £
                                                                          10
16
14
12
10       8

 TIME (min.)
          Figure  35.   Background during  thermal  desorption  of  Tenax GC


                      Cartridge  Blank
                                      93

-------
GLC Parameters

   FID-GLC
   Col - 12 ft 2% DECS Chrom W-HP
   Program - initial hold 2 min, 70-180° @ I0°/min
   Detector - 200°        ...
   Attenuation - 6.4 x 10    AFS

Thermal Desorption Parameters

   Chamber - 175°
   6-port Valve - 70°
   Cartridge Preheat - 5 min
   By-pass Open - 2 min
        14
12
10       8
  TIME (min.)
 Figure 36.   Background during thermal desorption of  PCB  carbon
             cartridge blank.
                                   94

-------
reactions.  At a temperature of 125° relatively little amounts of vapors




were desorbed from Tenax GC; 30-40% recovery was obtained at 175°C.  When




the desorption chamber was raised to 200°C, approximately 80-90% recovery




was observed for Mixtures I, II and III (Section V).




     Quantitative thermal desorption was achieved at 225°C in 90 sec.  A




comparison of the glc analysis for an aliquot of the synthetic air/vapor




mixture used for loading a cartridge and vapors desorbed from Tenax GC is




shown in Figures 37 and 38.  It was concluded that none of the vapors




studied had decomposed during the thermal desorption step since the chroma-




tograms were essentially identical to those obtained for mixtures directly



injected into the glc.




     The percent recoveries are given in Table 18.  Except for nitromethane,




all of the substances examined were quantitatively recovered.  Accuracy for




duplicate analysis was + 2%.



     In contrast to the results obtained for Tenax GC, attempts to desorb




these vapors from the activated carbons was not achieved even when tempera-




tures up to 330°C were used; higher temperatures began to exhibit chromato-



graphic peaks with retention indices different from the parent compounds




suggesting that decomposition was occurring.

-------
                               B
                                                                 ao
                                                                 6.0
                                                                    s
                                                                    X
                                                                    u
                                                                 4.0
                                                                    a:
                                                                    cc
                                                                    P
                                                                    a
                                                                 2.0
 8
        6
                             TIME ( MIN)
145
135
125
115
105
 r
95
85
75
                        TEMPERATURE (°C )
65
Figure 37
    Gas-liquid chromatogram of synthetic air/vapor mixture of
    hazardous substances.  Peaks A,  B,  C, D,  and E are  300 ng
    of glycidaldehyde, butadiene diepoxide,  N-nitrosodiethyi-
    amine, 1,2-dichloroethyl ethyl ether, and ethyl methane
    sulfonate, respectively.  See text  for glc parameters.
                                   96

-------
                                  4

                               TIME (WIN)
 145
135
125
115
     105      95


TEMPERATURE (°C)
85
75
                                                                    9.0
                                                                    8 JO
                                                                    7.0
                                                                    6.0
                                                                     ,0 2
                                                                    4.0
                                                                    3.0
                                                                    2.0
                                                                     1.0
                                                                        in
                                                                        CO
                                                                        a:
                                                                        ui
                                                                        o
65
Figure 38.   Gas-liquid chromatogran of vapors desorbed  from Tenax GC.



             Desorption chamber was  225°C; see prior  figure  for peak


             identity.  Background from Tenax GC is represented by


             dashed profile.

-------
            Table 18.  PERCENT RECOVERY OF VAPORS ADSORBED ON

              TENAX GC CARTRIDGES USING THERMAL DESORPTION3
Quantity Adsorbed (ng)
Compound
£
N-nitrosodiethylamine
B-propiolactone
ethyl methanesulfonate
nitromethane
glycidaldehyde
butadiene diepoxide
styrene epoxide
aniline
Bis (chloromethyl) ether
Bis- (2-chloroethyl) ether
50
100
105
105
-
100
100
100
95
100
95
100
100
100
100
-
100
100
100
95
100
90
200
-
100
95
70
95
100
105
95
100
90
300
-
100
100
70
80
100
90
60
90
-
 Tenax GC cartridge - 10.5 mm i.d. x 30 mm in length.  Synthetic air/
 vapor mixtures were introduced onto a Tenax GC bed @ A 1/min.  Desorp-
 tion Unit was at 225°C.
^Represents theoretical amount in synthetic air/vapor.
cValues, an average of duplicate runs, were calculated on basis of a ratio
 of peak areas for calibration mixture and from thermal desorption.
                                     98

-------
                             SECTION VIII                                    .


               DESIGN AND PERFORMANCE OF A FIELD SAMPLER



     Collection of pollutants by other investigators has been performed


at modest flow rates (20-2000 ml/min) because the concentrations of the


substances sought were relatively high.  In contrast, the sampling rates


and time required in this research program were much greater since the

                                                 3
hazardous vapors were anticipated at the low ng/m  levels.  This relation-


ship was previously shown in Table 2 (Section IV).


     In order to collect sufficient quantities of each atmospheric carcino-


gen for instrumental analysis, the field sampling unit had to meet several

                                                                     3
requirements.  These were:  (1) a sampling rate adjustable from 0-3 M /hr


at the pressure drop encountered with a sampler cartridge in-line, (2) a


capability of multiple cartridges on-line during a sampling period, (3)


uninterrupted 24 hr operation, and (4) the opportunity to "push" or "pull"


air samples through the cartridge sampler.  All of these factors ultimately


determine the power (milliamps/1/min) required for sampler operation.


     Also, these and additional factors were not independent of each other.


Other limitations were imposed by:   (1) the collection efficiencies of the


packing, (2) sorbent breakthrough characteristics, (3) the lowest detectable


concentrations of the carcinogenic compounds and (4) the size and shape of


the cartridge sampler and how pressure differential increases with increased


flow through it.  Because contamination of a sample was to be avoided, the


pump design was also important.  Of  these criteria considered in designing


a field sampling unit, the pressure  differential developed across a cartridge


at a specified flow rate was the most important.
                                    99

-------
     Questions regarding multiple sampling, continuous or pulsed flow were




also considered.  Flexibility, portability and durability were features




sought in the unit.




     The specifications for the design of a field sampling unit consisting




of a pump, multiport manifold, cartridge samplers, valves and flow indica-




tors were guided by all of the stated criteria.




     This section presents a discussion of these factors individually and




combined since many are interdependent.




SAMPLE VOLUME




     For the purpose of developing a sampling system it was assumed that




the amount of a compound required for its identification by a technique




such as gas chromatography-mass spectrometry is 30-50 ng and that the sor-




bent was quantitative in collecting the vapors at low concentrations.  The




relationships between the ambient vapor concentration and the total volume




of air that must be pumped through the cartridge, and thus, the time re-




quired for sampling, at various flow rates were given in Table 1.  The total




power required for sampling was directly related to sampling rate and dura-




tion.




POWER REQUIREMENT




     The power required to pump air through sorbent-packed tubes depends




upon the pressure differential across the tube under flow conditions, which




in turn depends mostly upon the air flow-rate, the shape of cartridge




(diameter and depth of packing), the particle size distribution, and the




particle shape of the sorbent, and to a lesser extent upon the air tempera-




ture and humidity.  A method for calculating power requirements from pres-




sure drop values was developed for this study.
                                     100

-------
     From the values of p  and ?„ in equation (4) , the theoretical power      •••?$.


(watts) required for compression or expansion of the air (compression if


the sampler were downstream from the pump) and its delivery through the


sampler was estimated using a formula derived from the Moss and Smith


equation for adiabatic horsepower.  The formula employed was:


                                                       k-1
                 Power (Watts) = SL/m x 5.968 x [ (pp   k   -1]    (7)
where H/m = liters per tnin, p. and p_ are the high and low pressures across


the sampler, and k is the ratio of specific heats for air, c  and c .  This


formula is based upon air at 14.7 psi, 23°C and 36 percent relative humidity;

                                    3
its density was taken at 0.075 Ib/ft ; k was set at 1.3947.


     Figure 39 shows the theoretical power required to pump air through


5-cm depths of sorbent packing.  For a flow-rate of 20 liters/min, the


change from 60/80 mesh particles to 18/20 in the 1.06 cm tube reduces power


requirements by almost a factor of 5 (30.4 watts to 6.4).  Increasing the


tube diameter from 1.06 cm to 1.82 cm decreases power for the 60/80 mesh


packing from 30.4 to 21 watts.  From these data it was concluded that tube


diameters of approximately 1.5 cm, particles of about 35/60 mesh, and bed


depths of about 5 cm will allow sampling rates of 20 liters /minute for a


theoretical power consumption of 15 watts (0.02 horse power) .  Practical


power requirements are expected to be at least twice the theoretical values


to allow for power losses in the pump itself.  For example, a cartridge of


Tenax GC, 35/60 mesh, in a 1.82-cm tube would require twice the theoretical


7,4 watts, or 14.8 watts to sample air at 20 £/tn.


     While the power requirements can be supplied by almost any 115 V,


50-60 Hz source, the power ratings of portable battery powered samplers


would have to be increased substantially to permit their use at low
                                     101

-------
                   lOOr
o
NJ
                     O.I
                                                                               D  (Cm) * U. S. Mesh
                                                                                P	
                                                               Tube  i.d.  (cm)
                                                                             A  0.0211
                                                                             O  0.0211
                                                                             A  0.035
                                                                                 80.093
                                                                                 0.035
                                                      60/80
                                                      60/80
                                                      35/60
                                                      18/20
                                                      35/60
 1.06
 1.82
 1.82
 1.06
 1.49
                                      10
100
                                                      POWER  (  WATTS )
                           Figure 39.
Relationship between flow rate  and  theoretical power requirements

at various tube diameters and particle  size.

-------
concentrations and sampling rates.  The required 6-volt battery ratings for

a sampler with a 5-cm depth of Tenax GC 30/65 in a 1.8 cm tube are given

in Table 19.

     With the exception of nickel-cadmium batteries, which provide up

to 45 mA for 10 hours, most of the sampling would therefore involve the

use of several batteries, possibly, in parallel, with recharging at regular

intervals, determined by the specific case.  The use of battery operated
                          3
samplers in detecting ng/m  concentrations of organic vapors depends upon

the effectiveness and capacity of sorbents with particle sizes in the 12/30

mesh range.  Pressure-drops at high flows (20 &/m) must be kept low to

avoid vacuum desorption if the air sample is drawn through the cartridge.

FLOW VECTOR REQUIREMENTS

     Whether the pumping unit would best serve in the capacity of pulling

or pushing the atmosphere through the cartridge also depended upon the

pressure differential developed.  For instance, if a large pressure drop


     Table 19.  POWER REQUIREMENTS TO DELIVER VARIOUS SAMPLING RATES
Sampling Rate
    A/min
Power Needed
    Watts
Six-Volt Battery Capacity
  Required Per Sample
        millamp-hr
ng/M3

1
4
9
20

0.022
0.34
2.06
14.8
1
1800
7500
19150
61500
10
180
750
1915
6150
100
18
75
192
615
1000
1.8
7.5
19.2
61.5
                                    103

-------
(> 100 mm Hg) was required to attain the desired sampling flew rates


then a "pull" system may significantly decrease the collection efficiency


profiles for each sorbent (vacuum stripping).  Since collection efficien-


cies in our studies have been acquired under positive pressure ("push")


then a "push" sampling system would deliver the best correlation between


field and laboratory trials.


     A continuous flow pumping unit was preferred to a pulse system since


the latter may have disturbed the packing material in the cartridge and/or


differ in trapping characteristics from those observed by our continuous


flow system in laboratory experiments.  A separate examination of collec-


tion efficiency and breakthrough volumes would be necessary in order to


determine the merits of pulse sampling.  Such a study was not conducted.


REPLICATE SAMPLING


     The uncertainties inherent in sampling for the hazardous organic vapors


imposed a need for a versatile system for multiple sampling, i.e., taking


several simultaneous samples.  For example, acquisition of duplicate samples


simultaneously circumvented the problem associated with diurnal fluctuations


of vapor concentrations occurring when duplicates were obtained sequentially.


Furthermore, different collection media could be compared by sampling the


same atmosphere concurrently.  The ability to collect with several car-


tridges at one time was conducive to overall shorter field sampling periods


as well as comparison of sample duration ys breakthrough (Section V).


     To draw 25 Jt/m of air through two sets of two samplers in parallel,


each containing a 3-cm depth of 0.035-cm diameter adsorbent in a 1.06-cm

                                                                        2
diameter tube, the pump must draw 50 £/m at a pressure drop of 8.34 N/cm


(626 nan Hg).  On the other hand it was also desirable to use cartridges


in tandem to determine whether breakthrough has occurred.  The Universal
                                    104

-------
Sampler 5 1068 (Research Appliance Co., Allison, Pa.) met these requirements

and has been used in our field sampling studies.  Its operation required

approximately 1/2 h.p. at 110 V, 50-60 Hz.  This unit is shown in concept

in Figure 40.

                     VACUUM GAUGE

                                            CRITICAL ORIFICE SET
                                         C3-
                      GAS       VACUUM
                       METER     RECORDER
                              RUNNING TIME
                                  METER

            Figure 40.  Schematic of Universal Sampler 5-1068


     The "pull" sampler would go at "A"; the push sampler, if employed,

at "B".  This system consists of those elements that will provide control

of the sampling at desired rates.  The system was quantified for a range

of flow rates up to 25 liters per minute, using the sampler employed in

testing adsorbents (a 3 cm deep packing of 0.035 cm diameter adsorbent

in a 1.06 cm diameter tube).  Calculations showed that, to draw  (or push)

25 1/m of air through two sets of two samplers in series, the pump must
                                                      2
draw 50 1/m (1.76 cfm) at a pressure drop of 8.34 N/cm  (12.1 psi).
                                     105

-------
     In order to accommodate the use of multiple cartridges during a sam-




pling period, a multiport chamber was designed and fabricated (Fig. 41).




The entire chamber was constructed of Teflon^.  Six ports were located




on the chamber equidistant from one another (60°) so that any multiple of




1,2,4, and 6 sampling cartridges could be employed simultaneously without




experiencing different drawing rates since a symmetrical geometry was main-




tained.  The air was drawn through a glass fiber filter (to remove "particu-




lates") and the cartridge, into the multiport head and then through the




pumping system.  The sampling rate through the cartridges was regulated




by a bleed valve located on the multiport chamber.  A vacuum/pressure gauge




on the chamber was used to monitor the pressure differential and by con-




sulting the pressure drop curves, a AP was selected and imposed in the cham-




ber which produced the desired sampling rate through each cartridge.




     The multiport chamber was utilized with the Universal sampling pump.




Table 20 depicts the maximum sampling rates which this system is capable




of achieving under various conditions (bleed valve closed).  When four




cartridges of Tenax GC (60/80) are compared to the same number of Chromo-




sorb 101 (100/120), the vacuum required to draw equivalent rates per cart-




ridge (and therefore total flow) was increased by 75 mm Hg.  This observation




corroborates the experimentally determined AP for various particle diameters.




The magnitude of the vacuum required to achieve a prescribed sampling rate




may ultimately influence the performance of the sorbent.  Decreased collec-




tion efficiencies and/or elution volumes may be experienced.




     The design parameters and systems described in this and previous sec-




tions were applied to the collection of trace quantities of hazardous




pollutants in ambient atmospheres.
                                     106

-------
                     CROSS- SECTION
 VACUUM PRESSURE

      GAUGE
     TEFLON -
     CHAMBER
GLASS FIBER FILTER

                      'S/SS/S//W//////A
                    '////S//////, I 'MSS///
PRESSURE  CONTROL
      VALVE
                                                GLASS COLLECTION CARTRIDGE
                                                 CARTRIDGE HOLDER
                           TO PUMP
                         TOP VIEW
           Figure 41.  Multiport sampling head,
                                  107

-------
         Table 20.   SAMPLING RATE CHARACTERISTICS FOR UNIVERSAL SAMPLER  WITH MULTIPORT HEAD
Sorbent
Tenax GC
(35/60)


Tenax GC
(60/80)


Chromosorb 101
(100/120)


b
No. of Cartridges
1
2
4
6
1
2
4
6
1
2
4
6
Sampling Rate
1/min/cartridge m^/hr/ cartridge
77.0
41.5
21.5
14.7
76.5
41.2
21.6
14.6
74.7
40.5
21.4
14.5
4.62
2.49
1.29
0.88
4.59
2.47
1.29
0.88
4.49
2.29
1.28
0.87
Total Volume
m3/hr
4.62
4.98
5.15
5.28
4.59
4.94
5.18
5.27
4.49
4.48
5.13
5.23
.The maximum pumping rate (no flow restriction)  for this commercial system is 89 1/min (5.28 m^/hr)
 Cartridge dimensions were 1.56 cm i.d.  x 6.0 cm in length.

-------
                              SECTION IX



             APPLICATION OF DEVELOPED INSTRUMENTATION AND



              METHODOLOGY TO THE ANALYSIS OF AMBIENT AIR





     In previous sections of this report detailed experimental design



and techniques were described for collecting and analysis of hazardous



vapors.  Preliminary results on the application of this methodology and



instrumentation to field sampling and the analysis of air samples by



combined gas-liquid chromatography/mass spectrometry/computer are dis-



cussed in this section.



EXPERIMENTAL



Site of sampling



     An urban site which presumably was concentrated with automobile ex-



haust in the presence of strong sunlight was chosen.  An environment con-



ducive to photochemical smog is known to exhibit high levels of ozone,



NO. and NO .  Two sites which had a history of high levels of ozone were
  £       A


at the CHAMP stations in Santa Monica and West Covina, CA  .  Furthermore,



West Covina during the month of April, 1974 also had appreciable amounts



of N0_ and NO .  On the basis of the presence of these pollutants and all-
     £»       X


phatic and aromatic hydrocarbons in air, the possibility of the formation of



epoxides and nitrosamines forming in ambient air was suspected to occur.



Sample Collection



     Ambient air samples were collected with a Universal Sampler 5-1068



equipped with a multiport head as described in Section VIII.  The sor-



bents selected for the collection of pollutants were Tenax GC and Chromo-



sorb 101 which was based upon their performance described in Section V.



The samplers  (cartridge holders plus glass cartridges) were prepared in
                                     109

-------
Inlet-
Hanilold
mem
GLC

-



1

                       Separator
Figure 42.  Gas-liquid chromatograph-mass  spectro-
            meter computer  (GLC-MS-CGti)  outlay.
                       Ill

-------
pattern total maximum and minimum m/e peak intensity, and standard deviation



from calibrated m/e, and (2) an electrostatic plot of total ion current plots



and/or normalized mass spectra.



    The operating parameters for the glc-ms-comp system for the analysis



of cartridges containing trapped pollutants are given in Table 22.  These



conditions were used throughout this entire study.



RESULTS AND DISCUSSION



    Figure 43 represents a total ion current plot from the mass spectro-



meter for an air sample from West Covina, CA.  This profile was typical



for samples also collected in Santa Monica.  The identity of several peaks



are given in Table 23 for this sample and Table 24 for a Santa Monica sam-



ple.  The major constituents were aromatics and aliphatics.  One oxygenated



compound, tenatatively,identified as styrene epoxide in this sample, was not



detected in samples from Santa Monica.  Styrene, itself, was likewise not



detected; however, methylvinyl benzene was present.  Styrene was reported


                                20
in Houston air by Bertsch et^ all.  ; styrene epoxide was not.  The discovery



of styrene epoxide was the most significant finding in this entire study.



    Mass spectra were also obtain of the background from Tenax GC (35/60);



it was concluded that ethylene oxide was an artifact generated from thermal



desorption of this sorbent.



    Because it was evident that the majority of the constituents were ali-



phatics and aromatics, single ion plots of 71 and 85 (Fig. 44) and 91, 105,



120, and 134 (Fig. 45) were obtained to demonstrate their distribution pat-



tern throughout a chromatogram.  The low intensities of the aliphatic ions



(Fig. 44) indicated that some selectivity was exhibited by the Tenax GC



cartridge, i.e. it was less effective in trapping aliphatic compounds than



aromatics or polar pollutants.  This selectivity was desirable because the
                                   112

-------
Table 21.  PROTOCOL FOR SAMPLING AMBIENT AIR IN LOS  ANGELES,  CA
Experiment Cartridge Type

1



2





3





Tenax GC (35/60)
1.5 cm i.d. x 6 cm
A
B
Tenax GC (35/60)
1.0 cm i.d. x 3 cm
A
B
C
D
Chromosorb 101 (60/80)
A
B
C
D
Sampling parameters
time (hr)


5
5


3
3
3
3

3
3
3
3
rate (1/min)


24
24


12
12
12
12

12
12
12
12
volume (m^)


6.6
6.6


2.07
2.07
2.07
2.07

2.07
2.07
2.07
2.07
Location Remarks
c1—.— *.-* it/ A_ .j .« .« /"* A






West Covina + internal
standards-
MEK & NB


West Covina + internal
standards-
MEK & NB


-------
UJ
UJ
>
UJ
UJ
tr
tr
o
                                            16 17
   O
   i
  mi |i|i|i M|i|i|rp I'M I MM
8700       8750       8800
                                885O       89OO       8950
                                MASS  SPECTRUM NO.
                                                'MI'I'I'IMM1!1!1!1!1!1!1!
                                                                    90OO
 it  i   i
50°
              |  i  i  i  i  i  i  i   i  iii  i  i  i  i  i   i    iii    i  i  i  i  i
             80         110        140        170        200        23O
                              COLUMN  TEMPERATURE  ( °C )
        Figure 43.  Total ion current plot during gas- liquid  chromatography mass
                   spectrometry of air sample  from West Covina, CA.   See text
                   for conditions.

-------
         Table 22.  OPERATING PARAMETERS FOR GLC-MS-COMP SYSTEM
  Parameter                                        Setting


Inlet-manifold
   desorption chamber                              225°C
   valve                                           125°C
   Capillary trap - minimum                       -195°C
                    maximum                       +175°C
   thermal desorption time                         4 min

GLC
   OV-17 Ni capillary, 450 ft
   column                                          50-230°C, 4°C/min
   carrier (He) flow                               ~3 ml/min
   transfer line to ms                             210°C

MS
   scan range                                      m/e 20 -»• 300
   scan rate, automatic-cyclic                     1 sec/decade
   filament current                                300 yA
   multiplier                                      6.0     ,
   ion source vacuum                               ~4 x 10   torr
                                   115

-------
    Table 23 (continued).  . POLLUTANTS IN AMBIENT AIR FROM WEST COVINA, CA
Peak No.                   RRT                       Name

   32                     2.030                  dimethylethylbenzene
   33                   •  2.035                  dimethylethylbenzene
   34                     2.059                  methyldiethylbenzene
   35                     2.070                  n-tridecane (tent.)
   36                     2.150                  unknown
   37                     2.162                  CiAH30
   38                     2.178                  tetramethylbenzene
   39                     2.233                  p_-tolualdehyde
   40                     2.261                  acetophenone
   41                     2.310                  unknown
   42                     2.330                  nitrobenzene
   43                     3.380                  unknown
   44                     3.530                  naphthalene
   45                     3.561                  unknown
                                      117

-------
   Table 24 (continued).  POLLUTANTS IN AMBIENT AIR FROM SANTA MONICA, CA
Peak No.                 RRT                       Name

   33                   2.161                acetophenone
   34                   2.330                nitrobenzene
                                     119

-------
to
o
                           m/e  71
                           m/e  85
     r i ' r r i ' r r r i1 rr r r PT  r rTTTrirm
            50          100          150         200
                                 MASS  SPECTRUM  NO.
                                                                         250
           • r j1 1 '
              300
 I  i  r  I  I  i  i   i  i   i  I   i  i   I  i   I  i   r  i   i  I   r  i   I

SO          80           110          140          170

                                  COLUMN TEMPERATURE  (°C)
i  |  i  i

 2OO
                                                                                     r  i  1  i

                                                                                      230
              Figure A4.  Single  ion plots of ions common to aliphatic cracking series.

-------
             m/e  134
             m/e  120
             m/e  105
             m/e  91
 I  '
50
 I
80
            I
I
110
         '  '   '  I  '   '  '
   140         170         200
COLUMN  TEMPERATURE (*C )
 I
230
 [rri 'I 'I'l'I'l'l' I 'I'M I1 IT1 [' I MM1 I1!1 I1 M PI 'I1 riTPIT(M MM
 0           50          100         150          200          25O         300      350
                               MASS  SPECTRUM   NO.
 Figure 45.  Single ion plots for ions representative of aromatic cracking series,

-------
hazardous vapors which are to be collected and identified under this pro-




gram are considered to be semi-polar to polar.




     In contrast the intensities of aromatic ions (Fig. 45) were relatively




large and constituted a high background between 140° to 230° which impeded




the detection and identification of possible hazardous vapors occurring at




trace levels.  It was concluded that the chromatographic conditions should




be improved whereby the aliphatic and aromatic compounds are resolved as a




group from hazardous vapors of interest.  The use of more polar stationary




phases and longer high resolution columns are currently under investigation.
                                     122

-------
                               SECTION X




                               REFERENCES






1.    Matz, J.  The Importance of Nitrosamines for Communal Hygiene.




      Z. Ges. Hyg. Ihre Grenzgebiete (Berlin).  18(12):903-8,  1972.




2.    Norpoth, K., Manegold, G., Brlicker, R.  and H. P.  Amann.   Investi-




      gations on the Problem of the Release of Alkylating Compounds  in the




      Course of the Smouldering Process.  Zbl. Bakt. Hyg.  156(B):341-52,




      1972.




3.    Jones, P. W.  Analysis of Nonparticulate Organic  Compounds in  Ambient




      Atmospheres.  67th Air Poll. Cont. Assoc. Mtg., Denver,  Colorado.




      Paper No. 74-265.  June 1973.




4.    Calvert, J. G. and J. N. Pitts, Jr.  Photochemistry.  John Wiley & Sons,




      Inc., New York, 1966.  pp. 366-557.




5.    Dunham, C. L. (ed.).   Biological Effects of Atmospheric Pollutants-




      Particulate Polycyclic Organic Matter.   Nat. Acad. Sci., Washington,




      D. C., 1972.  pp. 95-117.




6.    Ibid.  pp. 248-51.




7.    Van Duuren, B. L.  Carcinogenicity of Halo-ethers II Structure-Activity




      Relations of Analogs of Bis-(chloromethyl)ether.   J. Nat. Cancer Institute,




      48(5):1431-39, 1972.




8.    Van Duuren, B. L.  Dimethylcarbamyl Chloride, A Multipotential Carcinogen.




      J. Nat. Cancer Institute.  48(5):1539-40, 1972.




9.    Van Duuren, B. L.  Epoxides, Hydroperoxides, and  Peroxides in Air




      Pollution.  Int. J. Environ. Anal. Chem.  1(3):233-41, 1972.




10.   Van Duuren, B. L.  Interaction of Some Mutagens and Carcinogenic Agents




      with Nucleic Acids.  Proc. Int. Symp.,  1968.  p.  149.
                                     123

-------
11.   Van Duuren, B. L. and G. Witz.  "Phosphorescence Spectroscopy".   In



      Methods Pharmacol.  C. F. Chignell, Ed.  Appelton-Century-Crofts.,



      Kew York, 1972.  pp. 63-110.



12.   Aue, W. A.  Collection and Analysis of Organic Air Pollutant Trace



      Substances.  Environ. Health.  5:1-7, 1971.



13.   Altschuller, A. P.  Continuous Monitoring of Methane and Other



      Hydrocarbons in Urban Atmospheres.  J. Air Pollution Cont.   16:87-91,



      1966.



14.   Brooman, D. L. and E. Edgeley.  Concentration and Recovery  of Atmos-



      pheric Odor Pollutants Using Activated Carbon.  J. Air Pollution



      Cont. Assoc.  16:25-9, 1966.



15.   Jones, W. M.  The Absorption of Benzene Vapor from an Air Stream by



     ..Beds of Charcoal.  J. Appl. Chem.   16:345-9, 1966.



16.   Krumperman, P. H.  Erroneous Peaks from Porapak-Q Traps. J. Agr.



      Food Chem.  20:909, 1972.



17.   Williams, F. W. and M. E. Umstead.  The Determination of Trace Con-



      taminants in Air by Concentrating  on Porous  Polymer Beads.   Anal. Chem.



      40:2234-4, 1968.



18.   Raymond, A. and G. Guiochon.  Gas  Chromatographic Analysis  of C0-C-0
                                                                    0 lo


      Hydrocarbons in Paris Air.  Environ. Sci. Techn.  8:143-8,  1974.



19.   Hollis, D. L.  Separation of Gaseous Mixtures Using Porous  Polyaromatic



      Polymer Beads,  Anal. Chem.  38:309-16, 1966.



20.   Bertsch, W., Chang, R. C. and Z. Zlatkis.  The Determination of  Organic



      Volatiles in Air Pollution Studies:  Characterization of Profiles.



      J. Chromatog. Sci.  12:175-182, 1974.
                                    124

-------
21.   Zlatkis, A., Lichenstein,  H.  A.  and A.  Tishbee.   Concentration  and




      Analysis of Trace Volatile Organics in  Gases and Biological  Fluids




      with a New Solid Adsorbent.  Chromatographia.  6:67-70,  1973.




22.   Rasmussen, R. A.  A Quantitative Cryogenic Sampler-Design and Operation.




      American Lab.  4:19-27, 1972.




23.   Rohrschneider, L., Jaeschke,  A.  and W.  Kubik.  Air Pollution at




      Heights of up to 500 m above  Industrial Sites.   Chem.  Ing. Tech.




      43:1010-17, 1971.




24.   Kaiser, R. E.  Enriching Volatile Compounds by a Temperature Gradient




      Tube.  Anal. Chem.  45:965-7, 1973.




25.   Bellar, T. A., Brown, M. F. and J. E.  Sigsby, Jr.  Determination  of




      Atmospheric Pollutants in the Parts-per-Billion  Range  by Gas Chroma-




      tography.  Anal. Chem.  35:1924-27, 1963.




26.   Lonneman, W. A., Bellar, T. A. and A.  P. Altschuller.   Aromatic




      Hydrocarbons in the Atmosphere of the  Los Angeles Basin.  Environ. Sci.




      Technol.  2:1017-20, 1968.




27.   Altschuller, A. P., Lonneman, W. A., Sutlerfield, F. D.  and  S.  L.




      Kopczynski.  Hydrocarbon Composition of the Atmosphere of the Los




      Angeles Basin-1967.  Environ. Sci. Technol.  5:1009-16,  1971.




28.   Aue, W. A. and .P. M. Teli.  Sampling of Air Pollutants with  Support




      Bonded Chromatographic Phases.  J. Chromatog.   62:15-27, 1971.




29.   McEwan, D. J.  Automobile Exhaust Hydrocarbon Analysis by Gas Chroma-




      tography.  Anal. Chem.  38:1047-53, 1966.




30.   Farrington, P. S., Pecsok, R. L., Meeker, R. L.  and T. J. Olson.




      Detection of Trace Constituents by Gas Chromatography-Analysis  of




      Polluted Air.  Anal. Chem.  31:1512-16, 1959.
                                     125

-------
41.    Scheutzl,  D., Crittenden,  A.  L.  and R.  L.  Charleston.   Application




      of Computer Controlled High Resolution  Mass  Spectrometry  to  the




      Analysis of Air Pollutants.  J.  Air Poll.  Cont.  Assoc.  23:704-9,




      1973.




42.    Jeltes, R.  Sampling of Nonpolar Air Contaminants  on Porapak Porous




      Polymer Beads.  Atm. Envion.   3:587-8,  1969.




43.    Mann, J. R. and S. T. Preston.  Selection  of  Preferred  Liquid Phases.




      J. Chromatog. Sci.  11(4):216-220,  1973.




44.    Tiggelbeck, D.  Increasing Selective Efficiency  in Cigarette Filter




      Charcoals.  Proc. 4th Int. Tobacco  Sci. Cong., Athen,  1966.   pp.  923-




      44.




45.    Saunders,  R. A., Umstead,  M.  E., Smith, W. D., and R. H.  Gammon.




      The Atmospheric Trace Contaminant Pattern  of  Sea Lab II.   Proc.




      3rd Ann. Conf. Atmos. Contamination Confined Spaces, AMRL-TR-67-200,




      1967.




46.    Duel, C. L.  Collection and Measurement of Atmospheric  Trace Con-




      taminants.  Aerojet Electrosystems Co., Azusa, CA.  Final Report  Cont.




      NAS 1-8714, NASA Doc. No.  71-19636.




47.    Saunders,  R. A.  Analysis of Spacecraft Atmospheres.  NRL Rept.  5316,




      1962.




48.    Chiantella, A. J., Smith,  W. D., Umstead,  M.  E., and J. E. Johnson.




      Aromatic Hydrocarbons in Nuclear Submarine Atmospheres.  Am. Ind.




      Hyg. Assoc. J.  27:186-92, 1966.




49.    Saunders,  R. A.  Atmospheric Contamination in SEA-LAB  I.   Proc.  Conf.




      Atmos.  Contamination Confined Sapces.  AMRL-TR-65-230,  1965.
                                      127

-------
50.   Grob, K. and G. Grob.  Gas-liquid Chromatographic-Mass Spectrometric




      Investigations of C^-CL.. Organic Compounds in an Urban Atmosphere




      An Application of Ultra Trace Analysis on Capillary Columns.   J.




      Chroiaatog.  62:1-13, 1971.




51.   Jennings, G. and H. E. Nursten.  Gas Chromatographic Analysis of




      Dilute Aqueous Systems.  Anal. Chem.  39:521-3, 1967.




52.   Herbolsheimer, R., Funk, L. and H. Drasch.  Viability of Activated




      Carbon as Adsorbant for the Determination of Trichloroethylene in




      the Air.  Staub-Reinhalt. Luft.  32:31-3, 1972.




53.   Perry, R. H. and C. H. Chilton.  Chemical Engineers Handbook.  McGraw




      Hill, New York, N. Y., 1973.  pp. 5-53.




54.   Daniels, F. and R. A. Albertz.  Physical Chemistry.  John Wiley and




      Sons, Inc., New York, 1962.  pp. 607rll.




55.   Langmuir, I.  The Constitution and Fundamental Properties of  Solids




      and Liquids.  Part I.  Solids.  J. Amer. Chem. Soc.  38:2221-95,  1916.




56.   Langmuir, I.  The Absorption of Gases on Plane Surfaces of Glass, Mica




      and Platinum.  J. Amer. Chem. Soc.  40:1361-1403, 1918.




57.   Sune, J.  Private Communication.  Environmental Protection Agency,




      RTP, NC.
                                     128

-------
                               SECTION XI




          LIST OF PAPERS ACCEPTED OR SUBMITTED FOR PUBLICATION






1.   COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN




     AMBIENT ATMOSPHERES




     I.  A Technique for Evaluating the Concentration of Vapors by




         Sorbent Media.




     E. D. Pellizzari, J. Bunch, B. Carpenter and E. Sawicki




     J. Environmental Science and Technology.  Accepted for Publication.




2.   COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN




     AMBIENT ATMOSPHERES.




     II.  Studies on Thermal Desorption of Organic Vapors from Sorbent




          Media




     E. D. Pellizzari, B. Carpenter, J. Bunch and E. Sawicki




     J. Environmental Science and Technology.  Accepted for Publication.




3.   COLLECTION AND ANALYSIS OF TRACE ORGANIC VAPOR POLLUTANTS IN AMBIENT




     ATMOSPHERES




     III.  The Design of a Sampler System for Trace Quantities of Organic




           Vapors




     B. Carpenter, E. D. Pellizzari, J. Bunch and E. Sawicki




     J. Environmental Science and Technology.  Submitted for Publication.
                                   129

-------
APPENDIX
      130

-------

M
fc"
1 "
s
UJ •
p
5-





>




I
11 29 38












i







M,i









i
i » i • i • i







1


i in 1 in tl






1

11 	 i • i • i • i • i • i • i • t • i • i • i • i • i - i
•

•

•


ll,. 1.1
« ' ^o ' «!• ' Vo ' a'o ' 4i ' ilo ' lio ' lio ' ilo ' Uo ' lio ' iio ' do ' lio ' lio ' zlio ' 210 zio ' wo ' 1^0 ' zio



•100
• 10

10

• 40

• 20
• 0
M/e
Figure 46.  GLC-MS of bis-(chloromethyl)ether

-------

* •

s.
s
ssr
p
5,

• i • i • i • i •







L
11 28 38 «

i • i







,1 il
• i • i i







i
l|l|l|l|l|l|l|t|>|!I1|llV|»|lll1
a

•

*


1 II
h t'i Vi «'B 4i ib iio 120 lio ilo tio i«o 1^0 lio do 210 2*10 220 2&o 2^0 2^0
19V
10

(0

40

• 20
• 0
M/e
u>
KJ
                                        Figure 47.   GLC-MS of bis-(2-chloroethyl)ether.

-------
M

£
CO

5
UJ
p
5
£
         IB  ' 20  '  30  '  « '  da  '  «!o
                                                                                                                                          •100
                                                                                       ••  10
                                                                                       •• «o
                                                                                       ••  40
                                                                                       ••  20
do '  IQO ' iio'  iio'  i&o'  i4o'  lio' ito'
                  M/e
lilO  ' 190 '  2)10 '  210   220 ' 2^0 '  MO '  250
  10
  U)
                                                   Figure A8.   GLC-MS  of 3-propiolactone.

-------
2i

5


P
54-
•   i  •   i  '   i  '  i   •  i   •  i  '   i  '   i  'i  i   •  i  •   I  '   i  •  i   •  i   •   i  •   i   •  i   '  i  •   i  '   i  •  i   •  i   •  i  •   i  •   i  1100
                                                                                                                                    10
                                                                                                                                 ••  to
                                                                                                                                 •• 20
IB    28   39
+9   59   *9    70
                                            OB
                                            03   110   110  120   190   HO   190  1(0   110   190   ISO  200  210   220   290   210  250
                                                             M/e
                                         Figure 49.  GLC-MS of vinylene carbonate.

-------
K


M


5
UJ

p
oc
            »  «  i""|	.	dll_
—ii—j   i  |  i   p—i   (  P—f
 II    20    30    40   50   61
7B ' 8*0   Jo"
                                                                                                                                        100
                                                                                                                               •   10
                                                                                                                               •   60
                                                                                                                               . .  4Q
                                                                                                                                       •• 20
nO   110  120   190
            M/e
                                                                                      lid   1^0
200  210  220   230   240   250
 CO
 Ul
                                             Figure 50.  GLC-MS of N-diethylnitrosamine.

-------
JH
5-.
                              _L
        10   20   90   40   50   60   70   80   00
110   120  190  HO  150   160   170   180   ISO   200   210   220   290   210  250
       M/e
   u>
                                                 Figure 51.   GLC-MS of nitroir.ethane

-------
M
1 '
5
UJ •
p
(T
_J •

1 1 ' 1 ' 1 ' 1





11 2B 3D +0
i i i i i







1 i ii
do ' e!o











1



a



,1 ,1
Vo ' 90 ' Jo iko ' lio ' tio ' lio ' ito ' lio ' ito ' 1^0 ' lio ' lio ' zlio ' 210 ' 220 ' wo ' zio ' 2^0
M/e
'109
• 10
• (0

• 40
• 20
• o
Figure 52.  GLC-MS of ethylmethanesulfonate.

-------

r*
. .
M
a; .
5
UJ •
p
5-

• i * i • i • i








i in 1
18 28 ' ft ' 48
. i . i .,. i . i . i .,.,.,.,.,.,.,.,.,,,.,,,.,.,,,

.






| ,

•

ft ' ft ' Vo ' B'Q ' ft ' ibo ' no ' 120 ' lio ' ito ' lio ' ito ' 1^0 ' lio ' lio ' 250 ' 210 ' 220 ' 2^0 ' 2^0 ' sb
JVfl

10

• (0

• 10

• 20
• o
M/e
10
00
                                            Figure  53.  GLC-MS of glycidaldehyde,

-------

I*
. .
£
i.
5
UJ •
p
(E


• 1 • l ' I ' 1



i





I III
11 21 3D 4











|j
> 1 > 1 1 1 1 1 1 1 1 1 1 1 • 1 I 1 > 1 > 1 > 1 • 1 • 1 • 1 • 1 1 1 1 1 1 1 1 1 • 1









L „,!








1 1 i ll 1 Ilil 1 1 1 i
D 50 60 JO 80 4l lb 110 180 lIlO Uo l&O lU l!lO lIlO 190 290 210 220 2&0 2^0 2&0
•100

10

io

• 40


• 20
• o
M/e
VO
                                           Figure 54.  GLC-MS of propylene oxide.

-------

*
fc"
i.

z
1-4
UJ •
p
5 ;

.,.,.,.,.,.,.,.,.,








1
11 ' 2B 30 ' 4




,1
0 ' S


1 Ji
g ' 60






' >!' 1
• i < i i i i i • i • i • i • i • i • i • i > i • i > i • i • i
.

•





1
1/0 »'o ds do no i






k ' do ' do ' do ' do ' rio ' do ' do ' 2^0 ' 210 ' 220 ' »o ' 2^0 ' 250
19V
•0

(0


• «0

. 20
• 0
M/e
Figure 55.  GLC-MS of styrene oxide.

-------
5

P
uj
                           ill . 4 I
                                                                                                                              rrw*
                                                                                                                                -• 10
                                                                                                                                •  10
                                                   •   «0
                                                                                                                                •• 20
        18   21   31
8   60 '  Vo  '  8*8  '  ^8  '  180 '  lio  120  1^0
                                   M/e
lio   lio   1^0   lio   lio   200   210   220  zio  MO  »o
                                            Figure 56.  GLC-MS  of butadiene diepoxide.

-------
t
t-t
p
tc
UJ
                                                              I  '   I   '  I   •   I  «  I   '  I
                                                                                                              _J,I,I,_J,J_
                                                                                                                                       i  I  I100
                                                                                                                                              80
                                                                                                                                           •• «0
         ID    20    30   1-0    5D
7*8 ' 88  '  h '  ibo' iio'  iio  ' lio'  iio'  iio '
                             M/e
210  2J>0 '  230   M0250
                                                      Figure 57.   GLC-MS  of acrolein.

-------
LU


P
cn

LU
oe
                      ill
                                                                                                                                    100
                                                                                                                                  •• BO
                                                                                                                                   •• so
                                                                                                                                   •  40
                                                                                                                           •• 20
                               ILL
10    20    3D    +D    50
                                 &'o
70*8   flO   ibO   110   120  190  HO  150   160   110   150   ISO   200   210   220   290   240   250

                           M/e
                                             Figure 58.   CLC-MS of methyl ethyl ketorve.

-------

*
M
Z .
5
UJ •
P
tr

^

• i • i < i • i • i • i • i








J













i 1
1 i ' i "








1 I
l'o ' 20 ' 38 ' « ' 50 ' & ^0 ' 90 ' Jfl ' 1


i i i i i i i • 1 • i . i • i < i • i . i . i • i • , . , . ,
.

•






100
BO

• to

• 10


• 20

ho ' lio ' 120 ' iio ' ilo ' iio ' ilo ' do ' iio ' iio ' 290 ' 210 ' 220 ' afo ' 2^0 ' 2^0
M/e
Figure 59.  Mass spectrum of malelc anhydride.

-------
to
P
5
                        i
                                                                                                                             rrioo
                                                                                                                              .. 10
                                                                                                                              •• «o
                                                                                                                              •• 10
                                                                           •• 20
do'  do'  do'  do'  do'  do' do' do' do'  do'  210'  210'  220' s*io' 210' 2^0
            M/e
        10   20   3D   to   50    60   Vg   9*0
                                         Figure 60.  Mass  spectrum of succinic anhydride.

-------
M

2.
5
UJ •
P
5-



»

•
•

11 21 ' 3*8






1
i § i i i i • i i • i • i • i • i • i • i





ll Ll ,i ll

•



1 , ii
+'g h u Vi «'o SB ibo 110 i


o ' iio ' ito ' do ' lio ' do ' do ' lio ' 210 ' 210 ' 2^0 ' 2^0 ' 2^0 ' 2$o
M/e
109
• 10
10

• 40
20
• 0
Figure 62.  GLC-MS of tetramethylene sulfolane.

-------

» *

g.
s
y '
p
5-
^

,,,,.,,,.,. i . i . i • i







J



111 ill 1
11 28 38 ' *0 ' !?iJ 68







1 ,.,!
• i > i • i < i • i < i • i • i • i • i • i • i i i • i • i > i








100
10

(0

40

• 20

jo ' so ' dg ' ilio ' 110 ' lio ' lio ' lU ' ib ' lU ' i^ro ' lio ' lio ' ako ' 210 ' zio ' zio ' 2^0 ' zio
M/e
co
                                                Figure 63.  GLC-MS of aniline.

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
 EPA-650/2-74-121
                             2.
                                                           3. RECIPIENT'S >CCESSIOf*NO.
4. TITLE AND SUBTITLE
 Development of Method for Carcinogenic Vapor Analysis
 in Ambient Atmospheres
             5. REPORT DATE
               July  1974
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR
-------