EPA/600/R-92/055                                              EPA/600/R-92/055
                                                              April  1992
                           Theoretical Evaluation of Stability of
                             Volatile Organic Chemicals and
                        Polar Volatile Organic Chemicals in Canisters
                                         by
                                    R. W. Coutant
                                       Battelle
                              Columbus, Ohio 43201-2693
                                Contract No. 68-DO-0007
                                    Project Officer

                                 William A. McClenny
                  Atmospheric Research and Exposure Assessment Laboratory
                          U.S. Environmental Protection Agency
                            Research Triangle Park, NC  27711
                                   February 18, 1992
         ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
                     U.S. ENVIRONMENTAL PROTECTION AGENCY
                        RESEARCH TRIANGLE PARK, NC 27711

-------

-------
                                TECHNICAL REPORT DATA
 1. REPORT NO.

 EPA/600/R-92/055
                            2.
            3.r
             !      PB92-166941
 i>.  TITLE AND SUBTITLE

 THEORETICAL EVALUATION OF STABILITY OF VOLATILE
 ORGANIC CHEMICALS  AND POLAR VOLATILE ORGANIC
 CHEMICALS IN CANISTERS
            5.REPORT DATE

             01 APRIL 1992
            6.PERFORMING ORGANIZATION CODE

             EPA/ORD
7. AUTKOR(S)

R.  W.  COUTANT
                                                           8.PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION HAKE AND ADDRESS

Battelle
505  King Avenue
Columbus,  Ohio 43201-2693
            10.PROGRAM ELEMENT NO.

            DU Y105, A-04-01
            11. CONTRACT/GRANT NO.

            68-DO-0007
12. SPONSORING AGENCY NAME AND ADDRESS

Atmospheric Research  & Exposure Assessment Lab
Office  of Research  and Development
U.S.  Environmental  Protection Agency
Research Triangle Park,  N.C.  27711
                                                           13.TYPE OF REPORT AND PERIOD COVERED
            14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
v
A mathematical model  was developed  for  describing  loss by physical  adsorption  of
volatile  organic chemicals (VOCs) and polar volatile  organic chemicals (PVOCs)  in
stainless  steel canisters.  The model incorporates  compound specific properties
such  as polarizability,  vapor concentration, temperature, and equilibrium vapor
pressure.   Experimental  results show  that the model correctly predicts the loss  of
VOCs  in canisters from very dry samples.   A listing which documents the software
program that implements  the model is  included in the  report.  With  the program,  a
user  can  predict the  stability of VOCs  and PVOCs in multicomponent  mixtures
including water vapor under user-specified conditions of temperature and pressure.
Physicochemical data  needed for the model are provided for more  than sixty
compounds.  f-
17.
                                 KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                               b.IDENTIFIERS/ OPEN ENDED TERMS   c.COSATI
IB. DISTRIBUTION STATEMENT

RELEASE TO  PUBLIC
19. SECURITY CLASS (This Report)

UNCLASSIFIED
21.NO. OF PAGES

  48
                                               20. SECURITY CLASS CThis Page)

                                               UNCLASSIFIED
                                                                          ^^. PRICE

-------

-------
                                   DISCLAIMER
     The information in this document has been funded wholly or in part by the United
States Environmental Protection Agency under Contract No. 68-DO-0007 to Battelle.  It has
been subjected to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document.  In no event shall either the United States Environmental
Protection Agency or Battelle have any responsibility or  liability for any consequences of any
use, misuse, inability to use, or reliance on the information contained herein, nor does either
warrant or otherwise represent  in any way the accuracy,  adequacy,  efficacy, or applicability
of the contents hereof.  Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
                                         ii

-------

-------
                                     FOREWORD
      The Atmospheric Research and Exposure Assessment Laboratory, Research Triangle
Park, North Carolina, conducts intramural and extramural research in the chemical, physical,
and biological sciences.  This research is intended to characterize and quantify ambient air
pollutant levels and the resulting exposures of humans and ecosystems; to develop and
validate models to predict changes in air pollutant levels; to determine source-receptor
relationships affecting ambient air quality  and pollutant exposures; and to solve scientific
problems relating to EPA's mission through long-term investigation in the areas of
atmospheric methods, quality assurance, biomarkers, spatial statistics, exposure assessment,
and modeling.  The Laboratory provides support to Program and Regional Offices and state
and local groups in  the form of technical advice, methods research and development, quality
assurance, field monitoring, instrument development, and modeling for quantitative risk
assessment and regulation.  The Laboratory also collects, organizes, manages,  and distributes
data on air quality, human and ecosystem  exposures and trends for the Program and Regional
offices, the Office of Research and Development, the scientific community, and the public.
       Whole air collection using passivated stainless steel canisters is rapidly becoming the
method of choice for sampling of ambient volatile organic compounds. This methodology
has been validated through previous laboratory and field experience for a limited subset of
possible analytes. The current work provides a fundamental basis for development of
guidelines for the application of canister sampling to the ever expanding list of volatile polar
and non-polar organic compounds of importance to assessment of ambient air quality.
                                 Gary J. Foley
                                 Director
                                 Atmospheric Research and Exposure
                                  Assessment Laboratory
                                 Research Triangle Park, NC  27711
                                          m

-------
                                    ABSTRACT

      The potential for physical adsorption as a mechanism for loss of volatile organic
chemicals (VOC) and polar volatile organic chemicals (PVOC) from the vapor phase in
canister samples was assessed using the principles embodied in the Dubinin-Radushkevich
isotherm.  This isotherm provides a specific relationship between the tendency for adsorption
and compound/sample specific properties such as polarizability, vapor concentration,
temperature, and equilibrium vapor pressure.  In addition, the isotherm provides means for
distinguishing between surfaces having different physical and chemical properties.  A
computer based model was developed for predicting adsorption behavior and vapor phase
losses in multicomponent systems.  At present, the data base for the model contains relevant
physicochemical data for more than 60 compounds (42 VOC, 19 PVOC, and water), and
provisions for inclusion of additional compounds are incorporated  in the software.
      This report was submitted in partial fulfillment of Contract No. 68-DO-0007 by
Battelle under the sponsorship of the U.S. Environmental Protection Agency.  This report
covers a period from September, 1990 through February, 1991, and work was completed as
of February 28, 1991.

-------
                              CONTENTS


FOREWORD	iii

ABSTRACT  	iv

FIGURES	vi

TABLES  	vii

ACKNOWLEDGEMENT	viii

     1. INTRODUCTION	1

     2. CONCLUSIONS	3

     3. RECOMMENDATIONS	5

     4. MODEL DESCRIPTION AND RESULTS	6

          Background	6
          Canister Stability Software	8
              General Description of Software	8
              Program Operation	9
          Results	10
              Qualitative Results 	10
              Comparison with Experimental Results	22

REFERENCES   	30

APPENDIX A.  SOURCE CODE FOR CANISTER STABILITY PROGRAM	31

-------
                                    FIGURES


                                                                              Page

1 Adsorption Potential for Water Vapor	11

2 Adsorption Potentials for VOC and PVOC at 5 ppbv	17

3 Calculated Recoveries of Toluene, o-Xylene, 1,2,4-Trichlorobenzene from
  a Canister as a Function of Relative Humidity at a Sample Pressure of
  1 Atmosphere	18

4 Calculated Recoveries of Toluene, o-Xylene, 1,2,4-Trichlorobenzene from
  a Canister as a Function of Relative Humidity at a Sample Pressure of
  5 Atmospheres  	19

5 Effect of Temperature on Recovery of o-Xylene	21

6 Effect of Temperature on Recovery of 1,2,4-Trichlorobenzene	21
                                        VI

-------
                                    TABLES






Number                                                                    Page




Table 1. Physicochemical Properties of Target VOC and PVOC	12




Table 2. Experimental Results for RTI Canister Experiment	25




Table 3. Calculated Results for RTI Canister Experiment   	26




Table 4. Experimental and Calculated Recoveries for VOC	27
                                      Vll

-------
                              ACKNOWLEDGEMENT

     This work could not have been completed without the cooperation of representatives of
many laboratories who supplied information on their experiences with canister stability.
Special thanks are given to Dr. R.K.M. Jayanty (Research Triangle Institute), Mr. Michael
Holdren (Battelle), Ms. Deborah Smith (Battelle), Dr. Thomas J. Kelly (Battelle),
Dr. Sydney M. Gordon (Battelle), Dr. Rei Rasmussen (Oregon Graduate Center), and
Ms. Karen Oliver (ManTech Environmental Technology, Inc.) for their continued interest,
thoughtful discussion, and contribution of useful data.  We also wish to acknowledge the
inspiration and encouragement of Dr. William A.  McClenny who served as the EPA Work
Assignment Manager for this  task.
                                        viu

-------
                                     SECTION 1
                                  INTRODUCTION

      Physical adsorption of trace atmospheric constituents on the surfaces of sampling
apparatus has long been recognized as a contributing factor to apparent losses of some
organic compounds during sampling and sample storage. The use of passivated stainless
steel canisters for sample collection, shipping, and storage prior to analysis has diminished
this problem sufficiently that these devices are widely used for whole air sampling of non-
polar volatile organic compounds (VOC).  Nonetheless, it is recognized that not all VOC are
equally stable in canisters under all possible sampling temperatures and relative humidities.
The number of compounds that need to be accurately  sampled and analyzed is expanding as
the provisions of the Clean Air Act are implemented.  These new compounds include polar
volatile organic compounds (PVOC) and additional VOC,  some of which have lower vapor
pressures than VOC that are currently being sampled and analyzed using canisters.
      Experimental investigation of the stability of all possible combinations of important
VOC and PVOC at all concentration levels of interest and under all realistic sampling
conditions is not practical particularly given the dynamic nature of the provisions of the
Clean Air Act.  A more sensible approach is to develop guidelines for future evaluation and
application of canister sampling technology based on the implications of fundamental
processes that govern the stability  of whole air samples. Excluding compounds that are
inherently unstable (i.e. reactive) in the atmosphere, such as ozone,  these processes can be
limited to (1) heterogeneous chemical reaction (including chemisorption) with  surfaces within
the sampling system, and (2) physical adsorption on surfaces within the sampling system.
Passivation of a surface is generally performed to minimize the chemical reactivity of the
surface by either altering the chemical nature of the surface or by masking the surface by

-------
deposition of a less reactive coating.  In the case of Summa™ polished stainless steel
canisters, the surface area is reduced by the polishing process,  and the surface is
predominantly the relatively inert CrC^.  However, examination of the Summa polished
surface by ESCA, after exposure to air, shows the top 20 A to be covered by oxygen and
carbon species.  This is a common observation for most metallic surfaces that have been
exposed to air, and indicates that the main advantage of the Summa process may be in the
reduction of the surface area.
     The objective of this program was to evaluate the potential for physical adsorption for
a broad range of VOC and PVOC with the goal of developing a fundamentally consistent
model for assessing the stability of such compounds in canisters.  A subsidiary goal was to
collect and consolidate relevant experimental information for comparison with model
predictions.

-------
                                     SECTION 2
                                   CONCLUSIONS

      The potential for physical adsorption as a mechanism for loss of VOC and PVOC from
the vapor phase in canister samples was assessed using the principles embodied in the
Dubinin-Radushkevich isotherm.  This isotherm provides a specific relationship between the
tendency for adsorption and compound/sample specific properties such as polarizability,
vapor concentration, temperature, and equilibrium vapor pressure. In addition, the isotherm
provides means for distinguishing between surfaces having different physical and chemical
properties.  A computer based model was developed for predicting adsorption behavior and
vapor phase losses in multicomponent systems.  At present, the data base for the model
contains relevant physicochemical data for more than 60 compounds (42 VOC, 19 PVOC,
and water), and provisions for inclusion of additional compounds are incorporated in the
software.
      Based solely on the physicochemical properties of the compounds (i.e. independent of
surface considerations), the model predicts displacement of the more volatile VOC and
PVOC from a canister surface by water vapor at relative humidities in the range of 1 to 20
percent.  This is generally consistent with experimental observations, but in most cases, the
experimental conditions are not sufficiently characterized to permit detailed quantitative
comparison with the model. For example, relative humidities less than about 5 to 10 percent
are generally  not measured but rather are calculated based on the addition of a known
amount of water to a "dry" system.  A different kind of uncertainty arises when attempting
to compare the model results with field samples. In this case, the analysis is usually
conducted for a restricted set of analytes, whereas the model considers all components  to be
in competition for the surface.

-------
      Some conclusions inferred from this theoretical evaluation of canister sampling for

VOC and PVOC result from this work:

      (1)   Measurements of relative humidity and temperature should be made
           during the sampling process.  Under conditions where the relative
           humidity is so low that the model predicts target compound loss,
           provision should be made to add water vapor to the canister prior to
           analysis.

      (2)   The sample pressure should be as high as possible without causing
           precipitation of liquid water within the canister.  However,  pressure
           restrictions during shipment may apply. The pressure, P, for 100%
           RH occurs at: P = (100/RHi)EXP(-(AHv/R)(T(l-1 - T,'1)), where RH.
           is the relative humidity during sampling, AH,, is the heat of
           vaporization, R is the universal gas constant, and Ta and T, are  the
           analysis and sampling temperatures respectively.

      (3)   When considering the suitability of the canister sampling method for
           new compounds, the first parameters to be evaluated should be
           chemical reactivity and the vapor pressure of that compound.
           Compounds with equilibrium vapor pressures less than about 1 ton
           at ambient temperatures may require heating of the canister to effect
           good recovery, but heating of canisters should be done only with
           recognition of all of the effects this process may have on the
           recovery of the analytes.

      (4)   Inasmuch as all species present participate in the competitive
           adsorption process, retrospective considerations of the quality of data
           obtained from multiple  canisters at the same site should include at
           least semi-quantitative specification (e.g. total FID response) of non-
           target species contained in  the samples.

      Pending experimental confirmation of these conclusions under known controlled

conditions,  a  set of specific guidelines can be formulated for use in practical situations

encountered by  Regional, State and Local agencies as they implement canister-based
monitoring  programs.

-------
                                     SECTION 3
                               RECOMMENDATIONS

      The model developed in this program shows considerable promise for qualitative and
semi-quantitative explanation of physical adsorption phenomena associated with mixtures of
trace VOC and PVOC in canisters.  At this point, however, quantitative data are lacking with
respect to characterization of canister surface properties.  In particular, the surface roughness
factors for both electropolished and unpolished canisters are unknown.  At present, the model
calibration depends on an assumed value for this parameter, and uncertainty in the calibration
could be reduced by measurements of surface roughness. The model calibration also is
currently dependent on incompletely characterized experimental measurements at low relative
humidities. The accuracy of the calibration could be improved by a simple set of
experiments conducted under very controlled  conditions.  It would be desirable,  for example,
to prepare a standard cylinder of VOC in very dry air that could be used to charge a well-
cleaned and dried canister to various pressures. Analysis of the residual gas concentration at
each pressure would yield a more reliable calibration of the model than is currently in place.
      The parameters included in the model suggest a complex dependence of analyte
recovery on sample temperature. This aspect of canister analysis should be evaluated
systematically.

-------
                                    SECTION 4
                      MODEL DESCRIPTION AND RESULTS

BACKGROUND

     Physical adsorption is a widely studied phenomenon, and several excellent texts are
available on this topic (e.g. Flood, 1967). In brief, physical adsorption is characterized by
isosteric heats of adsorption that are no greater than a few kcal/mole (i.e. less than typical
chemical bond energies). A wide variety of theoretical and empirical isotherms have been
used to describe physical adsorption, but the most successful isotherm that is also
thermodynamically sound is the Dubinin-Radushkevich (DR) isotherm (Flood, op. tit.).  This
isotherm has received wide use in the description and modeling of adsorption of
multicomponent mixtures of VOC on activated carbons and in modeling the effect of relative
humidity on such systems (Coutant, 1987; Wemer and Winters, 1986).  Unlike many
empirical and other less well founded theoretical isotherms, the DR isotherm has been shown
to correctly describe physical adsorption  behavior over a very broad range of relative
pressures, extending from near saturation to values of P/P0 of the order of 10"'2 (Hobson,
1961).
     The general  form of the DR isotherm is given by
                              nr         i»       n  2
                              Wo         P2      Po

where W is the number of moles of adsorbed gas; W0 is  the total number of moles of gas
that could be adsorbed at saturation; B is a characteristic  of the surface; fl is the electronic

-------
polarizability of the sorbate, which is equal to the molar refractivity; R is the gas constant; T
is the absolute temperature; P is the experimental vapor pressure; and P0 is the equilibrium
vapor pressure at T.
      By invoking ideal solution theory, Equation 1 can readily be transformed for
application to mixtures of sorbates, including water vapor.  Then each component of the
system is represented by an equation of the type


                                  = exp-(—(K71n(—i-))2)                        (2)
                                          Pf
where Xj is the mole fraction of the ith component in the adsorbed phase.  The set of

resulting equations can then be solved by requiring

     (1)   Mass balance between starting vapor phase and final vapor phase plus adsorbed
           phase amounts for each component.  Inasmuch as formation of the adsorbed
          'layer takes place at the expense of the gas phase concentration, the decrease in
           gas phase pressure (AP) of each component is given by
           where V is the volume of the container.

     (2)   The sum of the mole fractions in the adsorbed phase is unity, i.e. ,
                                                                                  (4)
     (3)   The adsorption potentials for each component are equal at equilibrium (Grant and
           Manes, 1966), i.e.,

-------
                                                                                   (5)
The latter requirement provides a mechanism for directly determining the relative amounts of
each compound in the adsorbed and vapor phases.  Note that the adsorption potential
combines the effects of polarizability and vapor pressure and that these properties are the
drivers for determining the composition of the adsorbed phase.

CANISTER STABILITY SOFTWARE

General Description of Software
      A computer based model was developed to facilitate solution of the set of simultaneous
equations represented by Equations 2-5.  Code for this model was written in Microsoft
Quick-Basic. The software package includes the main program that is used for all
calculations, a Lotus- 123 file that contains the listing of compounds and their
physicochemical properties, and a .prn file that is generated from the Lotus- 123  file.  The
.prn file serves as a data base for the main program,  and provides a mechanism  for
expansion of the list of available compounds without requiring alteration of the program.  A
listing of the source code for the main program is given in Appendix A, and copies of the
compiled program and associated Lotus- 123 files have been supplied to the Project Officer
and Work Assignment Manager under separate cover.
      The program is menu driven,  with all inputs necessary for a computation being
specified through use of the menu.   Certain default values are specified at startup, but these
can be changed to suit the user's needs.  The program should be run on a system capable of
EGA  graphics and it requires a math co-processor for completion of a calculation in a
reasonable time.

-------
       The program assumes a spherical geometry for the canister, and the volume of the
canister is used to determine the geometrical surface area.  The surface roughness also is
used to determine the available area, which is related to W0. The surface roughness is
dependent on the extent of polishing — a perfect mirror surface would have a roughness
factor of 1, i.e., a true surface area equal to the geometrical surface area.  Electropolishing
typically yields roughness factors of the order of 1.5-2, and we have calibrated the model
using a default value of 2 for the surface roughness and experimental data derived using
Summa polished canisters.
      When trying to simulate experimental conditions, it is necessary to know the
temperatures for sampling and analysis.  The absolute pressure of the canister sample at the
completion of sampling also is needed. The program adjusts the  measured sampling
humidity to the analysis temperature and canister pressure, and provides a warning when the
sample properties are such as to  lead to water condensation.  The condensation of liquid
water within the canister could have a significant effect on the recovery of compounds that
are essentially miscible with water, e.g., low molecular weight alcohols and acids.
      Three output choices are available - screen, printer, and disk.  The screen and printer
outputs yield tables of compound names, adsorbed phase mole fractions (Xj), final gas phase
pressure (Pfn), initial gas phase pressure (PjJ, and the ratio of Pfo/Pj,,. The disk output
yields a file that can be imported into Lotus-123 for tabulation and comparison with
experimental data.

Program Operation
      Default settings are provided for all variables except sample composition which must
be supplied by the user.   Listings of compounds currently included in the data base can be
viewed either on screen by accessing the help feature or on hard copy by request from the
main menu. Once a particular composition is selected, it is stored by the program so that

-------
multiple computations involving changes in relative humidity, temperatures, etc. can be
conducted without reentry of the composition data.
      The program first uses the Newton-Raphson method to determine the values of X; at
the known initial values of PJ.  Then Equations 2 and 3 are used to estimate the amounts of
each component adsorbed and the corresponding residual gas phase pressures.  This process
of estimation of Xj and P4 is reiterated until the sum of the Xj converges to within I part-per-
million of unity. Approximately the same precision is achieved with the Pj except for the
case of very strongly adsorbed species.  To minimize computation time, any species for
which the estimated residual gas phase concentration is less than 1 percent of the starting
concentration, is arbitrarily set at the 1 percent level.

RESULTS
           Results
      The approach taken using the DR isotherm explicitly assumes that all species present in
the gas phase are in competition with each other for adsorption sites on the surface of a
canister, and that the ability for each species to compete is a function of (1) its concentration
in the gas phase, (2) its equilibrium vapor pressure, and (3) its polarizability. Quantification
of the combined effects of these variables on the total extent of adsorption depends on
experimental determination of B in Equation 1. Inasmuch as B is a function solely of the
surface, the usual approach is to measure W as a function of P using a single component
system.  For the current work, W0 was estimated by assuming a saturation density of 1014
molecules/cm2 and a surface area equal to the geometrical area times the roughness factor.
The value of B was then obtained as 7.05 x 10~5 by trial and error comparison of model
predictions with experimental data of Smith and Holdren (1989) on the recoveries for a 41
component VOC mixture.
                                          10

-------
      In principle, it is possible to explore the qualitative and even semi-quantitative

relationships embodied in the DR isotherm independent of consideration of the specific

surface involved.  For example, Figure 1 illustrates the dependence of the adsorption

potential for water vapor at different temperatures and gas phase concentrations.  This

information is combined in Figure 2 with similar data for the target VOC and PVOC

(calculated using physicochemical data shown in Table 1),
      -500
    0)
   -w
   s.
     i-iooo
    -1500
    e-
     -2000
     -2500
     -3000
100
1000
10000
20000
 0.33
 3.3
32.9
65.8
                                                           +^20000 ppmv
                                              *  10000 ppmv
         10      15      20      25      30      35
                                 Temperature, C


Figure 1. Adsorption Potential for Water Vapor.
                                            40
                                        45
50
                                        11

-------
Table 1. Physicochemical Properties of Target VOC and PVOC
Molar
Compound Molecular Refractive Refrac-
CA Index Name Wcinht,M_ Density j» Index, n tivity2
dichlorodifluoromethinc
methyl chloride
1 ,2-dichloro-l , 1 ,2,2-tetrafluoroethane
vinyl chloride
methyl bromide
ethyl chloride
trichlorofluoromethane
1 , 1-dichloroethcne
dichloromcthane
3-chloropropcne
1 , 1 ,2-trichloro- 1 ,2 ,2-trifluoroethane
1,1-dichloroethane
cis-1 ,2-dichlorocthenc
trichloromethane
1,2-dichlorocthanc
120.91
50.49
170.93
62.50
94.94
64.52
137.38
96.94
84.93
76.53
187.38
98.96
96.94
119.38
98.96
1.1834
0.9159
1.5312
0.9106
1.6755
0.8978
1.4940
1.2180
1.3266
0.9376
1.5635
1.1757
1.4459
1.4832
1.2351
NF3
1.3389
1.3092
1.3700
1.4218
1.3676
NF3
1.4249
1.4242
1.4157
1.3557
1.4164
1.4490
1.4459
1.4448
17.0
11.5
21.5
15.5
14.4
16.2
21.7
20.3
16.3
20.5
26.2
21.1
18.0
21.5
21.3
Vapor Pressure
Prtn«t*nt«l
A B
17.3
17.6
17.8
17.9
17.7
17.7
17.5
17.8
18.1
17.7
17.8
18.0
18.2
17.9
18.2
2598
2719
3079
2933
3048
3148
3221
3414
3610
3522
3578
3740
3843
3775
4113
Vapor Adsorption
Pressure Log(P) Potential
®25C,torr ©25 C ©5ppbv
5421.3
4578.0
1697.2
3315.8
1676.6
1216.3
796.9
590.8
414.5
366.4
326.0
222.0
202.9
192.1
81.8
3.73
3.66
3.23
3.52
3.22
3.09
2.90
2.77
2.62
2.56
2.51
2.35
2.31
2.28
1.91
-733
-1075
-550-
-785
-819
-718
-522
-549
-671
-532
-413
-501
-586
-489
-469
12

-------
Table 1. Physicochemical Properties of Target VOC and PVOC
Molar
Compound Molecular Refractive Refrac-
CA Index Name Weight.M., Density,/* Index, n tivity2
1,1 ,1-trichloroethanc
benzene
carbon tetrachloride
1 ,2-dichloropropane
trichloroethene
cis-1 ,3-dichloropropene
trans- 1 ,3-dichloropropene
1 , 1 ,2-trichlo methane
toluene
1 ,2-dibromoethanc
tetrachloroethene
chlorobenzene
ethylbenzene
m-xylene
p-xylene
133.41
78.12
153.82
112.99
131.39
110.97
110.97
133.41
92.15
187.87
165.83
112.56
106.17
106.17
106.17
1.3390
0.8787
1.5940
1.1560
1.4642
1.2170
1.2240
1.4397
0.8669
2.1792
1.6227
1.1058
0.8670
0.8611
0.8642
1.4379
1.5011
1.4601
1.4394
1.4773
1.4730
1.4682
1.4714
1.4961
1.5387
1.5053
1.5241
1.4959
1.4958
1.4972
26.1
26.2
26.4
25.7
25.4
25.6
25.2
25.9
31.1
27.0
30.3
31.1
35.8
36.0
36.0
Vapor Pressure
fniKstonKi1
A B
17.6
, 17.9
17.8
17.7
17.8
17.0
17.0
18.1
17.9
18.0
18.0
18.0
18.0
18.2
17.8
3813
3967
3909
4082
4016
4356
4356
4432
4306
4611
4487
4593
4639
4745
4614
Vapor Adsorption
Pressure Log(P) Potential
@25C,torr 025 C @5ppbv
124.4
95.2
109.3
53.2
75.3
10.9
10.9
24.9
30.2
12.9
19.5
12.9
11.1
9.5
10.4
2.09
1.98
2.04
1.73
1.88
1.04
1.04
1.40
1.48
1.11
1.29
1.11
1.04
0.98
1.02
-392
-385
-385
-379
-392
-344
-349
-359
-303
-330
-302
-286
-246
-242
-244
13

-------
Table 1. Physicochemical Properties of Target VOC and PVOC
Molar
Compound MoIecuUr Refractive Rcfrmc-
CA Index Name Weight.M., Density j> Index, n tivily2
ethenylbenzene
1 , 1 ,2.2-tetnchloiocthane
o-xyfcne
4-ethyl toluene
1 .3 ,5-trimethyIbenzcne
1 ,2,4-trimethylbenzene
benzyl chloride
m-dichlorobcnzcnc
p-dichlorobenzcnc
o-dichlorobcnzene
1 ,2,4-trichlorobenzene
hciuchlorobuudiene
oxirane
2-propencnitrilc
methyloxirane
104.16
167.85
106.17
120.19
120.20
120.20
126.59
147.01
147.01
147.01
181.45
260.76
44.05
53.06
58.08
0.9060
1.5953
0.8802
0.8620
0.8652
0.8758
1.1002
1.2884
1.2473
1.3048
1.4542
1.6820
0.8824
0.8060
0.8590
1.5468
1.4940
1.5055
1.4930
1.4994
1.5048
1.5391
1.5459
1.5285
1.5515
1.5717
1.5542
1.3597
1.3911
1.3670
36.4
30.6
35.8
40.6
40.8
40.7
36.1
36.1
36.3
36.0
41.0
49.7
11.0
15.6
15.2
Vapor Preuuic
PniKttnfol
A B
18.0
18.2
18.0
18.0
18.1
18.1
18.7
17.9
18.4
18.5
18.1
20.4
18.0
17.4
18.0
4764
4827
4753
4959
5017
5076
5449
5028
5278
5352
5577
6625
3239
3786
3501
Vapor Adsorption
Pressure Log(P) Potential
®25C,torr 025 C 05ppbv
7.7
7.1
7.9
4.0
3.5
Z9
1.5
2.8
2.1
1.7
0.5
0.2
1314.4
109.9
528.7
0.89
0.85
0.90
0.60
0.55
0.47
0.17
0.45
0.32
0.22
•0.26
-0.78
3.12
2.04
2.72
-236
-279
-241
-202
-199
-197
-211
-222
-215
-214
-171
-127
-1057
-650
-731
14

-------
Table 1. Physicochemical Properties of Target VOC and PVOC
* Molar
Compound Molecular Refractive Rcfrac-
CA Index Name Weight.M^ Density,/) Index, n tivity2
2-propenoic acid, ethyl ester
1,3-butadiene
acetonitrtie
2-propanone
methanol
ct Hanoi
2-butanone
2-melhoxy-2-methyl-propane
2-ethoxy-2-methyl-propane
2-propanol
n-butanol
acetic acid, ethenyl ester
acetic acid, ethyl ester
2-mcthyl- 1 ,3-butadiene
2,6,6-trimethylbicyclo[3. 1 . l]hept-2-ene
100.11
54.09
41. OS
58.08
32.04
46.07
72.12
88.15
102.18
60.11
74.12
86.09
88.12
68.13
136.24
0.9240
0.6211
0.7857
0.7899
0.7914
0.7893
0.8054
0.7405
0.7519
0.7855
0.8098
0.9317
0.9003
0.6810
0.8582
1.4050
1.4292
1.3442
1.3588
1.3288
1.3611
1.3788
1.3690
1.3794
1.3776
1.3993
1.3959
1.3723
1.4219
1.4658
26.6
22.5
11.1
16.2
8.2
12.9
20.7
26.9
31.4
17.6
22.2
22.2
22.3
25.4
44.0
Vapor Pressure
Pnnatflnfa'
A B
18.3
17.5
17.7
18.3
20.4
21.0
17.7
17.6
17.7
21.4
21.3
18.8
18.8
17.4
18.0
4361
2912
3939
3844
4632
5045
3903
3526
3897
5240
5733
4220
4252
3278
4857
Vapor Adsorption
Pressure Log(P) Potential
@25C,torr ©25 C ®5ppbv
40.7
2223.8
91.6
221.4
122.2
58.0
100.0
311.4
104.5
44.2
8.0
108.5
90.9
578.1
5.4
1.61
3.35
1.96
2.35
2.09
1.76
2.00
2.49
2.02
1.65
0.90
2.04
1.96
2.76
0.73
-361
-532
-909
-654
-1243
-758
-489
-402
-323
-547
-389
-458
-452
-439
-191
15

-------
Table 1. Physicochemical Properties of Target VOC and PVOC
Molar
Compound Molecular Refractive Refrac-
CA Index Name WciKhl,M_ Density j> Index, n livily2
water
ethanoic acid
18.02
60.05
1.0000
1.0490
1.3330
1.3718
3.7
13.0
Vapor Prctture
CVinrt.nt.l
A B
20.5
33.1
5178
10345
Vapor Adsorption
Preiiure Log(P) Potential
®25C,torr 025 C ®5ppbv
23.1
0.2
1.36
-0.71
-2498
-494
1. In (P.torr) = A - B/T

                       M     2  i
2. MolartefroctMty =  (—2:)(^-l
                       P   n2+
(estimated by group contributions when refractive index not given)
3. Refractive index not found
                                                   16

-------
       -200
e>  -400
"5
•^  -600
i
o  -800

S -1000
    £•-1200
    O

   
                                                                          0.1
                                                                          0.01

                                1234
                            Log Vapor  Pressure <8 Z5C
Figure 2. Adsorption Potentials for VOC and PVOC at 5 ppbv.


     Figure 2 shows the adsorption potentials of VOC and PVOC (at 5 ppbv) plotted as a

function of the logarithm of the pure compound vapor pressures at 25 C.  The right-hand

scale of Figure 2 shows the position of the adsorption potential for water at various relative

humidities (at 25 C).  Several observations and conclusions can be made from this plot:

     (1)   All but one of the VOC (methyl chloride) and most of the PVOC
           (excepting ethanol, methanol, acetonitrile, and ethylene oxide) fall within a
           fairly narrow band that is linear with log(Po). Therefore, for compounds
           that lie within this band, the  equilibrium vapor pressure is the most
           significant factor in determining their relative adsorption potential.  This
          plot also serves as a means for distinguishing those PVOC that can be
           expected to have physical adsorption behaviors that differ significantly
           from the majority of the compounds considered.
                                         17

-------
     (2)   Superposition of the information from Figure 1 on the VOC/PVOC data
           shows the relative position of water vapor at various relative humidities.
           In the sense of the adsorption potential as used here, the more positive the
           adsorption potential, the greater the tendency for adsorption.  Therefore,
           water vapor at  1-20 percent RH is expected to inhibit adsorption of many
           of the VCX: and PVOC considered.  It should be noted, however, that the
           adsorption potentials of the organics also are functions of their gas phase
           concentration and that increased concentrations will enhance their ability to
           compete with water vapor.


One further aspect of the information contained in Figure 2 is made more obvious in Figure

3, which shows the results  of application of the model to a three-component mixture of

aromatic compounds at various relative humidities.
   cc.

   I
   -+S
   -°
   &
                                                  PraBsura  - 1  atn.
                                                 2.4-Tri cl-il orobenzene
                10
20
30    40     50    60    70
  R*lativ«  Humidity. 2
80
100
Figure 3. Calculated Recoveries of Toluene, o-Xylene, 1,2,4-Trichlorobenzene from a
Canister as a Function of Relative Humidity at a Sample Pressure of 1 atm.
                                         18

-------
Figure 3 clearly illustrates the fact that compounds with lower vapor pressures require higher
relative humidities to effect good recoveries.  It also illustrates the effect of having more than
two components in the system. In this regard, the recoveries of toluene and o-xylene are
enhanced at low relative humidities by the presence of the trichlorobenzene.
      The effect of yet another variable, total sample pressure, is implicit in the model.
Sample concentrations are customarily expressed as ppbv  with the tacit assumption that the
sample is at a total pressure of 1 atmosphere.  In the model, the important variable is P^ the
partial pressure of each component.  In the sampling process, canister samples are usually
pumped to pressures greater than 1 atmosphere to facilitate recovery of the gas during
analysis. Under these conditions, the partial pressures of all components including water
vapor are greater than they were at 1 atmosphere.  The effect of this increase in pressure can
be seen by comparing Figure 4 with  Figure 3.
   .1
    a
120
110
100
 90
 80
 70
 eo
 so
 4O
 30
 20
 10
              To I
                                      SoupI a Praaaure —  5 otm.
                 o-/Xy I arts
                    I . H , 4 —Tr i el-i I orobenzene
                1O
                      2O
                      3Q    40
                        Rs I o t i
                                          SO
                                          Hum i
 BO    70
li ty. X
                                                              BO
                                                                    QO
                                                                           i on
Figure 4. Calculated Recoveries of Toluene, o-Xylene, and 1,2,4-Trichlorobenzene from a
Canister as a Function of Relative Humidity at a Sample Pressure of 5 Atmospheres.
                                          19

-------
      The curves shown in Figure 4 were derived using the model and the same overall
composition as that used for Figure 3, but with the sample pressure being increased to 5
atmospheres. Although canister pressures are usually considerably less than 5 atmospheres,
Figure 4 shows that there is a significant effect of total pressure on sample recovery.  To a
first approximation, the data in Figure 3 are compressed to the left by a factor of 5 because
of the increase in total pressure.
      Another effect that is sometimes ignored is the effect of temperature on the competitive
adsorption of water vapor. Increasing the temperature of the canister decreases the
adsorption potential of water in just the same way as it decreases the adsorption potentials of
the organic components. However, because of the different heats of vaporization for
different compounds, the adsorption potentials will not change at the same rate with
increasing temperature.  For  some sample situations, where the water adsorption potential
decreases more rapidly than does that of a particular VOC, the result of increased
temperature can be decreased recovery of the VOC.  This is illustrated in Figures 5 and 6.
      Figure 5 shows the calculated recovery of o-xylene with increasing temperature at two
different relative humidities.  At 10 percent RH, o-xylene is relatively unaffected by the
water vapor and the recovery increases with temperature as expected. However,  at 30
percent RH good recovery of o-xylene is predicted at room temperature,  and the effect of
increasing the temperature is dominated initially by the effect of lowering of the water
adsorption potential, resulting in poorer recovery of the xylene. As the temperature is raised
more, the effect of raising the vapor pressure of the o-xylene begins to dominate, and the
recovery curve becomes coincident with that of the low humidity case. At higher humidities,
the minimum in the response curve becomes progressively less pronounced.  Figure 6 shows
the results of a similar calculation for 1,2,4-trichlorobenzene.  The lower vapor pressure of
this compound causes the temperature/humidity effect to be even more pronounced than with
o-xylene, and even slight increases in temperature cause dramatic initial drops in the
                                         20

-------
       120
       1 IO
       IOO
        90
        8O
        7O
        SO
        50
        4O
        3O
        2O
        IO
          o~Xylana •  E ppbv
RH  - I O3T
         20     30
                             60    80
                                          70
                                         ra t.i_i r
                                                SO     90     IQO    110    12G
Figure 5. Effect of Temperature on Recovery of o-Xylene.
                                 1 .Z, 4-Trl cf> I orobanzana  • Sppbv
                                                                    14O
Figure 6. Effect of Temperature on Recovery of 1,2,4-Trichlorobenzene.
                                          21

-------
predicted recovery efficiency. Note that in both Figures 5 and 6, the minimum recovery
occurs at temperatures in the vicinity of SO C, which is comparable to temperatures used in
some current laboratory analyses.  Although heating of canisters to improve recovery of low
vapor pressure analytes is a common practice, to the best of our knowledge, no systematic
investigation of the effect of temperature on recovery has been conducted.

Comparison with Experimental Results
      Experiments to evaluate the stability of simulated or real ambient air samples in
canisters have been conducted by several laboratories. However, many of these experiments
have been conducted without recognition of the potential importance of several variables and,
therefore, experimental conditions have not always been adequately defined.  As examples,
neither the sample pressure nor the relative humidity (especially at low RH) are specified
accurately in many cases.  These experiments may provide valuable information for specific
sampling applications, but  they do not serve the broader goal of development of
understanding and generalization of conclusions. Two sets of data that come close to having
complete identification of conditions were obtained for use in this report.  The general
experimental procedures were similar for these works and consisted of the following:

      (1)   A pressurized  source cylinder of target compounds was prepared in "dry"
           gas and this cylinder was analyzed for the target compounds.
      (2)   Canister samples were prepared from the source cylinder with
           addition of measured amounts of water to the canisters.
      (3)   The canister samples were stored for various  time periods prior to
           analysis.

      The first set of data was received from Research Triangle Institute (RH) (Jayanty,
1990). Jayanty's  canister samples were prepared at a total pressure of 3 atmospheres
                                          22

-------
(absolute) and the samples were analyzed periodically over a 30-day period. Jayanty
included both polished and unpolished canisters in his experiments, but inasmuch as no
quantitative measures of the total surface areas (or surface roughness) of the two types of
canisters were given, only the results obtained with polished cylinders are cited here.
Furthermore, only the final analyses for the wet (RH ~  100% with excess water present)
and "dry" canisters are considered.  Table 2 shows the experimental results obtained by
Jayanty, and Table 3 shows the calculated results. Because of the uncertainty concerning the
actual moisture content in Jayanty's "dry"  experiment, the computations included several
relative humidities.  In general, the calculated recoveries are somewhat higher than the
experimental  results, with the calculated humidity effect stronger than the experimentally
measured effect. Quite possibly, the low experimental recoveries of some compounds from
the "wet" canisters could be due to dissolution in the excess liquid water that was present in
these canisters. Problems with the analysis of methanol are acknowledged, and the  most
notable discrepancies between the experimental and calculated results occur with the three
esters and n-butanol.
     The second data set was developed at Battelle (Smith and Holdren, 1989). In this
case, the canisters were pressurized to 19.1 psia and analyses were conducted after a holding
period  of 1-2 days.  As with Jayanty, Smith and Holdren did not measure  the relative
humidity of the pure air that was used for preparation of the "dry" source  cylinder.  Recent
work conducted with the  same air supply by this author suggests that the relative humidity of
this air is currently at least 2 percent and possibly as high as 7 percent.  It is possible,
therefore, that the relative humidity values specified by Smith and Holdren are lower than the
true values. Nonetheless, the model predictions are in at least qualitative agreement with the
experimental recoveries for most of the compounds shown in Table 4.  If the assumption is
made that the reported relative humidities are low by 7 percent, the agreement with  the
model is better except for the lowest vapor pressure compounds such as hexachlorobenzene
                                          23

-------
and 1,2,4-trichlorobenzene.  As shown in Figure 3, the latter compound is predicted to
require relative humidities in the neighborhood of 50 percent to achieve the recoveries shown
in the experimental data, and even higher relative humidities are predicted to be required for
hexachlorobenzene recovery.  It should be emphasized that this aspect of the model does not
depend on the calibration and is a function only of the properties of the compounds.  In this
regard, the vapor pressure constants for hexachlorobenzene (see Table 1) appear to be
inconsistent with those of similar compounds in Table 1, and this could be the cause of the
lack of agreement for this compound.
      Jayanty (private communication) reported that after completion of his 30-day
experiment, he added water to his dry canisters and reanalyzed with the result that good
recoveries were obtained.  With both the RTI and Battelle work, it would have been
desirable to investigate the effect of elevated temperatures  on the recovery of analytes from
the canisters.
                                          24

-------
Table 2. Experimental Results for RTI Canister Experiment
Compound C0,ppbv
1,1, 1-trichloroethane
toluene
2-propenenitrile
2-propenoic acid, ethyl ester
2-propanone
methanol
n-butanol
acetic acid, ethenyl ester
acetic acid, ethyl ester
5.3
5.0
13.9
4.9
7.2
13.1
5.8
5.7
5.4
Percentage Recovery
Dry
Day 0 Day 31
96.2%
60.0%
42.4%
0.0%
20.8%
188.5%
0.0%
0.0%
0.0%
95.9%
53.7%
10.9%
0.0%
17.2%
99.2%
0.0%
0.0%
0.0%
Wet1
Day 0 Day 31
103.8%
100.0%
89.9%
55.1%
84.7%
99.2%
29.3%
129.8%
66.7%
104.7%
100.4%
87.9%
56.0%
86.2%
141.4%
26.1%
129.8%
68.0%
1.  Estimated relative humidity =100% (excess liquid water present).
                                                      25

-------
Table 3. Calculated Results for RTI Canister Experiment*
Compound C0,ppbv
1 , 1 , 1 -trichloroethane
toluene
2-propenenitrile
2-propenoic acid, ethyl ester
2-propanone
methanol
n-butanol
acetic acid, ethenyl ester
acetic acid, ethyl ester
5.3
5.0
13.9
4.9
7.2
13.1
5.8
5.7
5.4
Recovery @ RH =
0.01% 1.5% 5% 7.5% 10% 100%
98.8%
66.8%
100.0%
95.7%
100.0%
100.0%
97.4%
99.8%
99.8%
98.8%
68.5%
100.0%
95.8%
100.0%
100.0%
97.3%
99.8%
99.8%
99.0%
75.1%
100.0%
96.4%
100.0%
100.0%
97.3%
99.8%
99.8%
99.3%
83.9%
100.0%
97.4%
100.0%
100.0%
97.6%
99.8%
99.8%
99.9%
98.3%
100.0%
99.7%
100.0%
100.0%
99.6%
100.0%
100.0%
100%
100%
100%
100%
100%
100%
100%
100%
100%
* Polished canister; sample pressure = 45 psia.
                                                       26

-------
Table 4. Experimental1 and Calculated Recoveries for VOC
Compound Co2,
Name ppbv
dichlorodifluoromcthane
methyl chloride
1 ,2-dichloro- 1 , 1 ,2,2-tetrafluoroethanc
vinyl chloride
mcthylbromide
ethyl chloride
trichlorofluoromethane
1,1-dichlorocthene
dichloromethanc
3-chloropropenc
1 ,1 ,2-trichloro-l ,2,2-lrifluoro ethane
1,1-dichloroethane
cis-1 ,2-dichloroethene
trichloromethanc
1 ,2-dichk>roethane
1,1,1 -trichloroethane
2.60
2.64
2.68
4.61
3.06
2.58
2.74
3.22
4.11
3.22
2.72
3.07
3.48
345
3.34
1.86
RH = 0% RH = 1.2* RH = 2.9% RH - 5.8* RH = 8.7* RH =11.6*
exp cal exp cal exp cal exp cal exp cal exp cal
ND3
101
106
96
96
95
NO
96
103
96
99
95
95
93
89
101
100
100
100
100
100
100
100
100
100
100
99.8
100
100
99.9
99.9
99.5
ND
102
106
100
97
98
ND
97
106
98
98
96
97
94
96
97
100
100
100
100
100
100
100
100
100
100
99.8
100
100
99.9
99.9
99.5
ND
108
110
100
101
99
ND
98
115
100
99
99
99
96
97
100
100
100
100
100
100
100
100
100
100
100
99.8
100
100
99.9
99.9
99.5
ND
100
108
94
97
96
ND
96
97
96
97
96
96
94
96
97
100
100
100
100
100
100
100
100
100
100
99.8
100
100
99.9
99.9
99.6
ND
102
109
94
96
97
ND
97
97
97

97
97
91
96
98
100
100
100
100
100
100
100
100
100
100
99.9
too
100
99.9
99.9
99.6
ND
98
100
106
97
98
ND
97
99
97
98
96
96
91
96
96
100
100
100
too
100
too
100
100
100
100
99.8
100
100
99.9
99.9
99.6
27

-------
Table 4. Experimental1 and Calculated Recoveries for VOC
Compound Cj,
Name ppbv
benzene
cubon tctnchloride
1 ,2-dichloroproptnc
trichloroethene
cii-1 ,3-dichloropropcne
trans- 1 ,3-dichloropropcnc
1,1,2-inchferoethuic
toluene
1 ,2-dibromocthanc
letrachloroethene
chlorobcnzcne
cthylbenzene
m-xylene
p-xykne
stymie
1 .1 ,2,2-tctnchloroethane
2.83
2.96
2.69
2.93
2.93
2.93
2.83
2.45
3.06
2.51
2.54
2.16
1.07
1.07
2.25
1.18
RH = 0% RH = 1.2% RH = 2.9ft RH = 5.8% RH = 8.7% RH =11 6%
exp cal exp cal exp cal exp ctl exp cal exp cal
95
96
90
92
82
63
79
88
69
91
81
75
72

33

99.4
99.4
99.1
99.4
95.8
96.3
97.9
91.8
94.3
89.6
82.1
46.9
41.1
42.4
31.7
75.2
97
97
96
94
98
94
95
95
94
95
95
95
93

82

99.4
99.4
99.1
99.4
95.8
963
98.0
92.0
94.4
89.7
82.4
47.4
41.7
430
32.2
75.5
100
109
98
98
109
98
100
96
96
%
95
96
94

84

99.4
99.4
99.1
99.4
95.9
96.3
98.0
92.1
94.5
89.9
82.8
48.3
42.5
43.8
33.0
76.0
97
97
96
95
97
93
97
95
96
94
95
98
97

83

99.4
99.5
99.1
99.4
96.0
96.5
98.1
92.5
94.7
90.4
83.5
49.9
44.1
45.5
34.5
76.8
98
99
96
92
98
93
96
97
95
95
94
97
97

89

99.5
99.5
99.2
99.5
96.2
96.6
98.1
92.9
94.9
90.8
84.3
51.8
46.0
47.4
36.3
77.8
98
96
98
93
98
93
100
97
95
95
98
96
97

94

99.5
99.5
99.2
99.5
96.3
96.7
98.2
93.3
95.1
91.4
85.2
54.0
48.3
49.6
38.5
79.0
28

-------
Table 4. Experimental1 and Calculated Recoveries for VOC
Compound C^2,
Name ppbv
o-xylcne
4-ethyl toluene
1 ,3 ,5-trimethyIbeiuenc
1 .2,4-trimethyIbenzene
benzylchloride
m-dichlorobcnzcne
p-dichk>robenzenc
o-dichlorobenzcnc
1 ,2,4-trichlorobenzcne
hexachlorobutadiene
1.1ft
1.83
1.83
1.88
1.10
1.10
1.78
2.16
1.64
1.37
RH - 0 % RH = 1.2% RH - 2.9% RH - 5.8% RH - 87% RH -11.6%
exp cal exp cal cxp col exp cal exp cal exp cal
62
37
37
28

40
37
35
5
19
37.2
5.9
5.1
4.3
9.7
16.1
11.4
11.0
1.0
1.0
90
87
87
84

80
78
78
45
63
37.8
6.0
5.2
4.5
9.9
16.5
11.6
11.3
1.0
1.0
93
90
88
87

85
87
83
55
70
38.6
6.3
5.4
4.7
10.2
16.9
12.0
11.6
1.0
1.0
97
99
91
94

91
94
92
66
81
40.2
6.8
5.8
5.0
10.8
17.9
12.7
12.3
1.0
1.0
94
94
93
92

88
85
88
64
79
42.0
7.4
6.4
5.5
11.6
19.1
13.6
13.2
1.0
1.0
96
95
95
94

89
87
90
69
85
44.2
8.2
7.1
6.1
12.5
20.5
14.7
14.2
1.1
1.0
1.  Temperature = 25 C; sample pressure = 19.1 psia; concentrations measured by GC/FID; (Smith and Holdren, 1989).
2.  Concentrations measured in master cylinder by GC/MSD.
3.  ND = not detected by FID.
                                                  29

-------
                                 REFERENCES

The Solid-Gas Interface, ed. E. A. Flood, Marcel Dekker, Inc., New York, 1967.

Coutant, R. W., Removal of Volatile Organics ftom Humidified Air Streams by Adsorption.
Final Report to Environics Division, Engineering and Services Laboratory, Tyndall Air
Force Base, September 15, 1987 (NTIS order No. AD-A192 43516/6/XAD).

Werner, M. D. and Winters N. L., "Predicting Gaseous Phase Adsorption of Organic
Vapors by Microporous Adsorbents," CRC Critical Rev. Envir. Control, 16, 327-356
(1986).

Hobson, J. P., "Physical Adsorption of Nitrogen on Pyrex at Very Low Pressures", J.
Chem. Phys.,  34, 1850 (1961).

Jayanty, R. K. M., Research Triangle Institute, Research Triangle Park, NC., personal
communication, 1990.

Smith, D. L. and Holdren, M. W., Development of Procedures for Performance Evaluation
of Ambient Air Samplers for Volatile Organic Compounds, final report for Work
Assignments 78 and 79 on EPA Contract No. 68-02-4127, September 1989.

Grant, R. J. and Manes, M., "Adsorption Behavior of Binary Hydrocarbon Gas Mixtures on
Activated Carbon," I&EC Fund., 5, 490-498 (1966).
                                       30

-------
     APPENDIX A. SOURCE CODE FOR CANISTER STABILITY PROGRAM
DECLARE SUB comp 0
DECLARE SUB menu 0
DECLARE SUB cmpdlist Q
DECLARE SUB help 0
DECLARE SUB molefrac Q
DECLARE SUB warn Q
DECLARE SUB done 0
DECLARE SUB putout Q
DIM SHARED fl(l TO 3), name$(l TO 100), i, mt, RHo, RH, P, Ps, R, B, Ts, Ta, V,
Wo, beta(l TO 100)
DIM SHARED Pin(l TO 100) AS DOUBLE, Pfh(l TO 100) AS DOUBLE, Po(l TO 100)
DIM SHARED X(l TO 100) AS DOUBLE, n, AP(1 TO 100), BP(1 TO 100), ofileS, y AS
DOUBLE
REM ' Default values'
     mt = 0: B =  .0000705: Ta = 298: Ts = 298: V = 6
     Wo = 2.66E-06: R = 2: Ps = 1: RHo = .01
FOR i = 1 TO 3: fl(i) - 2: NEXT
me: menu
molefrac
putout
GOTO me

SUB cmpdlist
CLS
       9: COLOR 11, 11
     OPEN "R", #1, "cancal.pmH, 69
     FIELD #1, 40 AS namf$, 9 AS fbet$, 9 AS fAP$, 9 AS fBP$, 2 AS ex$
     WHILE NOT EOF(l): j = j + 2
         GET #1, j: a$ = namfS: GET #1, j + 1: B$ = namf$
         PRINT a$; TAB(40); B$
         IF (j / 20 - INTO / 20)) = 0 THEN
         COLOR 12
         PRINT "More - press any key to continue"
         SLEEP
         COLOR 11
         END IF
     WEND

                                 31

-------
     CLOSE #1
END SUB

SUB comp
     CLS :  i - 1: j = 0
     IFRH > OTHEN
          name$(l) = "water vapor
          beta(l)  = 3.7025
          AP(1) - 20.514
          BP(1) = 5178
          i = 2
     END IF

     COLOR 12: INPUT "Number of Compounds including water = "; nc: COLOR 11
     OPEN "R", #1, "cancal.prn", 69: j = 0
     FIELD #1,  40 AS namfS, 9 AS fbet$, 9 AS fAP$, 9 AS fBP$, 2 AS ex$
     WHILE NOT  EOF(l): j = j + 1
          GET#l,j
          PRINT  namfS; : INPUT "  Concentration, ppbv = "; Pin(i): Pin(i) = 1E-09 *
Pin(i)
          IF Pin(i) > 0 THEN
               name$(i) = namfS: beta(i) = VAL(fbet$)
               AP(i) = VAL(fAP$): BP(i) = VAL(fBP$)
               i  = i +  1
          END IF
          n = i-  1
          IF n  =  nc THEN GOTO fin
     WEND
fin:  CLOSE #1
     FOR i = 1  TO n:  Pfn(i) = Pin(i): mt = mt + Pin(i): NEXT
     X(l) = Pin(l) / mt: X(2) = Pin(2) / mt
     FOR i = 1  TO n: Po(i) = (1 / 760) * EXP(AP(i) - BP(i) / Ta): NEXT

END SUB

SUB done
STATIC xpo
STATIC xpn
     IF (ABS(y)  <  1) THEN

                                    32

-------
           xpn = 10 + ABS(550 * (LOG(ABS(y)) / LOG(n * .000001)))
           IF xpn < xpo THEN LINE (10, 260)-(559, 270), 0, BF
           IF xpn > = 549 THEN xpn - 559
           LINE (10, 260)-(xpn, 270), 11, BF
          IF xpn > 285 THEN
                COLOR 14
                CIRCLE (20, 20), 20
                CIRCLE (20, 20), 10, ,  3.93, 5.5
                CIRCLE (15, 15), 2
                CIRCLE (25, 15), 2
                COLOR 11
           ENDIF
      xpo = xpn
      ENDIF
END SUB

SUB help
CLS
SCREEN 9: COLOR 15, 11
LINE  (0, 0)-(639, 349), 12, B
LINE  (2, 2)-(637, 347), 12, B
LOCATE 2, 2: PRINT "This program uses the Dubinin-Radushkevich isotherm and
assumes"
LOCATE 3, 2: PRINT "ideal solution behavior  in the adsorbed phase to calculate loses by "
LOCATE 4, 2: PRINT "adsorption on the canister wall. Canister and sample parameters
are"
LOCATE 5, 2: PRINT "changed from the default values by selection from the menu.  The "
LOCATE 6, 2: PRINT "initial composition of the sample is specified by selecting 'Cor 'c'."
LOCATE 7, 2: PRINT "The latter action will cause compound names to be displayed
sequentially,"
LOCATE 8, 2: PRINT "and requires that you enter the concentrations in ppbv. Any
compound for"
LOCATE 9, 2: PRINT "which you enter zero or no concentration will be excluded from the
computation."
LOCATE 10, 2: PRINT "The data file 'CANCAL.pm'  must be under the same directory as
this program"
LOCATE 11, 2: PRINT "before you can run this program. To change that file, use
'CANCAL.wkl'and"
                                     33

-------
LOCATE 12, 2: PRINT "Lotus-123. Data output choices can be toggled on/off by repeating
the choice."
LOCATE 13, 2: PRINT "The disk output will generate a file that can be imported into
Lotus-123."
LOCATE 14, 2: PRINT "After each calculation the program will return to the menu so that
the effects"
LOCATE 15, 2: PRINT "of changes in either sample or canister parameters can be explored
with the"
LOCATE 16, 2: PRINT "same starting composition.  To review list of available compounds,
press 'c'."
LOCATE 17, 2: PRINT "Note that the RH at analysis time is calculated based on the
sampling RH,"
LOCATE 18, 2: PRINT "the sample pressure, and the temperatures of sampling and
analysis, "
LOCATE 19, 2: PRINT "but only the analysis  time conditions show."
LOCATE 20, 2: COLOR 4: PRINT "Any comments or suggestions should be directed to
Bob Coutant (614)424-5247."
LOCATE 23, 2: COLOR 9: PRINT "To return to the menu, press any key."
ret: B$  * LCASE$(INKEY$): IF B$ - "" THEN  GOTO ret
   IF B$ = "c" THEN cmpdlist

END SUB

SUB menu
menu: CLS  : a$ =» INKEY$
SCREEN 9:  COLOR 7, 0
COLOR 9: LOCATE 1, 28: PRINT "Canister Sample Integrity"
COLOR 3: LOCATE 1, 70: PRINT "He(l)p"
COLOR 14:  LOCATE 3, 31: PRINT "Canister Properties"
COLOR 2: LOCATE 4, 15: PRINT "(V)olume -"; V; "LH
LOCATE 4, 50: PRINT "(R)oughness -"; R
COLOR 14:  LOCATE 6, 32: PRINT "Sample Properties"
COLOR 2: LOCATE 7, 15: PRINT "(Temperature - "; Ta - 273; "C"
P = Ps * Ta / Ts: RH = P * RHo * EXP(5178 * (Ta A -1  - Ts A -1))
LOCATE 7, 50: PRINT "Relative (H)umidity - "; : IF RH  > 100 THEN COLOR 12
LOCATE 7, 72: PRINT USING "###.##"; RH; : PRINT "%": COLOR 2
LOCATE 8, 15
     IF X(2) = 0 THEN
     COLOR  12

                                    34

-------
     ENDIF
PRINT "(C)oraposition"
COLOR 2: LOCATE 8, 50: PRINT "(P)ressure - "; : PRINT USING "##.##"; P; : PRINT
"atm"
COLOR 14: LOCATE 10, 33: PRINT "Output Choices"
COLOR H(l): LOCATE 11, 15: PRINT "(1) Screen"
COLOR fl(2): LOCATE 11, 33: PRINT "(2) Printer"
COLOR fl(3): LOCATE 11, 50: PRINT "(3) Disk"
   IF (X(2) = 0) OR fl(l) + fl(2) + fl(3) < 7 THEN
     LOCATE 13, 35
     COLOR 12
     PRINT "Not Ready"

   ELSE
     LOCATE 13, 37
     COLOR 10
     PRINT "Ready"
   ENDIF
LOCATE 15, 37: COLOR 9: PRINT "(S)tart"
COLOR 3: LOCATE 17, 1: PRINT "To change any of the above, press key corresponding
to letter or number"
PRINT  "in parentheses - - or Q to exit program. For a printed list of available"
PRINT  "compounds, press 4."
PRINT
COLOR 11
user:  a$ = LCASE$(INKEY$): IF a$ = "" THEN GOTO user
     IF a$ = "t" THEN
     INPUT "Analysis Temperature, C = "; Ta
     Ta = Ta + 273
     INPUT "Sampling Temperature, C = "; Ts
     Ts = Ts + 273
     ENDIF
     IF a$ = "1" THEN fl(l) =  12 - fl(l)
     IF a$ = "2" THEN fi(2) =  12 - fl(2)
     IF a$ = "3" THEN
     fl(3) =  12 - fl(3)
     INPUT "Name of output file = "; ofile$
     ofile$ = ofile$  + ".prn"
     ENDIF

                                   35

-------
     IF a$ = "4" THEN
          OPEN "R", #1, •cancal.pm", 69:j = 0
          FIELD #1, 40 AS namft, 9 AS fbet$, 9 AS fAP$, 9 AS fBP$, 2 AS ex$
          WHILE NOT EOF(l): j = j + 1
          GET #1, j
          LPRINTj;"  "; namf$
          WEND
          CLOSE #1: LPRINT CHR$(12)
     END IF
     IF a$ = "q" THEN END
     IF a$ - "v" THEN
          INPUT "Canister volume, L = "; V

     END IF
     IF a$ = "h" THEN
          INPUT "Relative humidity (at sampling temperature), % = "; RHo
          IF RHo = 0 THEN
          COLOR 12
          PRINT "Warning -- Zero RH is unrealistic, try again using some small number"
          INPUT "Relative humidity (at sampling temperature), % = "; RHo
          END IF
     END IF
     IF a$ = "r" THEN
          INPUT "Roughness factor = "; R

     END IF
     IF a$ = "p" THEN INPUT "Pressure at sampling temperature, atm. = "; Ps
     IF a$ = "c" THEN comp
     IF a$ = "1" THEN help
     IF a$ = "s" THEN
               IF (X(2) = 0) OR fl(l) + H(2) + H(3) < 7 THEN
               LOCATE 23, 5:  COLOR 12: PRINT  "Please select output mode or
composition — press any key to continue"
               SLEEP
               ELSE
               CLS
               LOCATE 15, 35: COLOR 12: PRINT "Working"
               LOCATE 18, 70: COLOR 11: PRINT "Done"
               LINE (8, 257H561, 272), 12, B

                                   36

-------
                LOCATE 18, 1: PRINT "0%"
                LOCATE 18, 35: PRINT "50%"
                GOTOes
           END IF
     END IF

     GOTO menu

es: END SUB

SUB molefrac
STATIC dy AS DOUBLE
STATIC dx AS DOUBLE
STATIC dP AS DOUBLE
STATIC w AS DOUBLE
     j - 0: Wo - R * 1.67E-08 * (12.57 * (V / 4.19) A .666)
     Po(l)  = (1 / 760) * EXP(AP(1) - BP(1) / Ta): Pin(l) =  .01 * RH * Po(l)
     mt = Pin(l): Pfh(l) = Pin(l)
     FOR i = 2 TO n: Pin(i) = P * Pin(i): Pfh(i) = Pin(i): mt = mt + Pin(i): NEXT
     X(l) = Pin(l) / mt: X(2) = Pin(2) / mt
     FOR i = 1 TO n: Po(i) = (1 / 760) * EXP(AP(i) - BP(i) / Ta): NEXT

repeat: FOR k -  1 TO 5
     y = l-X(l):dy = l:j =j + 1
     FOR i = 2 TO n
          X(i)  = (Pfh(i) / Po(i)) * (X(l) * Po(l) / Pfh(l)) *  (beta® / beta(l))
          y -  y - x(i)
          dx = (Pfh(i) / Po(i)) * (beta(i) / beta(l)) *  (Po(l) / Pfh(l))  * (X(l) * Po(l) /
Pfh(l)) * ((beta(i) / beta(l)) - 1)
          dy = dy + dx
     NEXT!
     WHILE (X(l) + y / (dy)) < 0: dy -  dy * 2: WEND
     IF ABS(y)  > 1 THEN X(l) = X(l) + y / dy ELSE X(l) = X(l) +  y / (n * dy)
     NEXTk
     done
     IF ABS(y)  > n  * .000001 THEN
           w = Wo * EXP(-(B / beta(l) x 2) * (2 * Ta * LOG(X(1) * Po(l) / Pfh(l)))
2)
          FOR i «  1 TO n

                                    37

-------
               dP = 0
               IF X(i) < 1 THEN
               dP = (.082055 * Ta / V) * w * X(i)
               IF dP  >  Pin(i) THEN dP » .99 * Pin(i)
               Pfh(i)  =  ((49) * Pfh(i) + (Pin(i) - dP)) / 50
               END IF
          NEXTi

          GOTO repeat
     END IF
     IF j < 50 THEN GOTO repeat
     CLS
     FOR i = 1 TO n: Pin(i) = Pin(i) / P: Pfh(i) = Pfh(i) / P: NEXT

END SUB

SUB putout
IF fl(l) = 10 THEN
     PRINT STRING$(79, 205)
     PRINT "Compound";  TAB(40); "mole frac"; TAB(55); "Cfin"; TAB(62); "Cinit";
TAB(70); "%Recovery"
     PRINT STRING$(79, 205)
     FOR i = 2 TO n
     PRINT RTRIM$(name$(i)); TAB(40); : PRINT USING "W.W~~   "; X(i); :
PRINT USING "##.##  "; Pfh(i) * 1E+09,  Pin(i)  * 1E+09; : PRINT USING "####.#"; 100
* Pfh(i) / Pin(i)
     NEXT
     PRINT STRING$(79, 205)
     LOCATE 23, 2:  COLOR 12: PRINT "Press any key to continue -"
     SLEEP
END IF
IF fl(2) = 10 THEN
     WIDTH LPRINT 80
     LPRINT STRING$(79, 61)
     LPRINT "Compound"; TAB(40); "mole frac"; TAB(55); "Cfin"; TAB(62); "Cinit";
TAB(70); "%Recovery"
     LPRINT STRING$(79, 61)
     FOR i = 2 TO n
                                   38

-------
      LPRINT RTRIM$(name$(i)); TAB(40); : LPRINT USING M##.#rA~   "; X(i); :
LPRINT USING "##.## "; Pfh(i) * 1E+09, Pin(i) * 1E+09; : LPRINT USING "####.#";
100 * Pfo(i) / Pin(i)
      NEXT
      LPRINT STRING$(79, 61)
      LPRINT "Relative humidity  = "; RHo; "% at sampling temperature = "; Ts - 273;
"C"
      LPRINT "Analysis temperature = "; Ta - 273; "C"; " and canister pressure = "; P *
14.7; "psia"
      LPRINT CHR$(12)
END IF
IF fl(3) = 10 THEN
      OPEN "O", #2, ofile$
      FOR i =  1 TO n
      mt = Pin(i)
      PRINT #2, name$(i); : PRINT #2, USING " ##.#r~A "; mt; : PRINT #2, USING "
##.###"; Pfh(i) / Pin(i)
      NEXT
      CLOSE #2
      fl(3) = 2
END IF

END SUB

SUB warn
CLS
SCREEN 9: COLOR 12, 11
LINE (0, 0)-(639, 349), 12, B
LINE (2, 2)-(637, 347), 12, B
LOCATE 10,  10
PRINT "WARNING ~ You have selected a combination of pressure and relative"
LOCATE 12,  10: PRINT "humidity that will cause liquid water condensation! Try again."
LOCATE 20,  10: COLOR 9: PRINT "Press any key to continue -"
SLEEP
Ps = 1: RHo  = .01
END SUB
                                    39

-------

-------