United States
              Environmental Protection
              Agency
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park, NC 27711
              Research and Development
EPA/600/SR-92/055  May 1992
EPA      Project  Summary
              Theoretical  Evaluation  of
              Stability of Volatile  Organic
              Chemicals  and Polar  Volatile
              Organic Chemicals  in Canisters
              Robert W. Coutant
                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 em-
              bodied in the Dubinin-Radushkevich
              isotherm. This isotherm provides a spe-
              cific relationship between the tendency
              for adsorption and compound/sample
              specific properties such as polarizabil-
              ity, vapor concentration, temperature,
              and equilibrium vapor pressure. In aid-
              dition, the isotherm provides the means
              for distinguishing between  surfaces
              having different physical and chemical
              properties. A computer-based model
              was developed for predicting adsorp-
              tion 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 PVOlC,
              and water),  and  provisions for inclu-
              sion of additional compounds are in-
              corporated 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 Febru-
              ary, 1991, and was completed as of
              February 28,1991.
                 This Project Summary was developed
              by EPA's Atmospheric Research and
              Exposure Assessment Laboratory, Re-
              search Triangle Park, NC, to announce
              key findings of  the research project
              that is fully documented in a separate
report of the same  title (see Project
Report ordering information at back).

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 passi-
vated stainless steel canisters for sample
collection, shipping, and storage prior to
analysis has diminished this problem suf-
ficiently that these devices are widely used
for whole air sampling of non-polar vola-
tile organic compounds (VOC).  Nonethe-
less, it is recognized that not all VOC are
equally stable in  canisters under all pos-
sible sampling temperatures and relative
humidities. The number of compounds that
need to be accurately sampled and ana-
lyzed is expanding as the  provisions of
the Clean Air Act are implemented. These
new compounds  include (PVOC) and ad-
ditional VOC, some of which have lower
vapor pressures  than  VOC that are cur-
rently being sampled and analyzed using
canisters.
  Experimental investigation of the stabil-
ity of all possible combinations of impor-
tant VOC and PVOC at all concentration
levels  of interest and under all realistic
sampling conditions is not practical  par-
ticularly given the dynamic nature of the
provisions of the Clean Air Act. A more
sensible  approach  is  to develop guide-
lines for future evaluation and application
of canister sampling technology based on
the implications of fundamental processes
that govern the stability of whole  air
                                                              Printed on Recycled Paper

-------
samples. Excluding compounds that are
inherently unstable (i.e.,  reactive) in the
atmosphere, such as ozone, these pro-
cesses  can be  limited to (1) heteroge-
neous chemical reaction (including chemi-
sorption) with  surfaces within the sam-
pling system, and (2) physical adsorption
on surfaces within the sampling  system.
Passivation of a surface is generally per-
formed to minimize the chemical reactivity
of the surface by either altering the chemi-
cal nature of the surface or by masking
the surface by deposition of a less reac-
tive coating.  In  the case of Summa™*
polished stainless steel canisters, the sur-
face area is reduced by the polishing pro-
cess,  and the  surface is predominantly
the relatively inert CrO2. However, exami-
nation of the Summa polished surface by
ESCA, after exposure  to air, shows the
top 20A to be covered by  oxygen and
carbon species. This is a common obser-
vation for most metallic surfaces that have
been  exposed to air, and indicates that
the main advantage of the Summa pro-
cess may be in the reduction of the sur-
face area.

Objective
  The objective of this program was to
evaluate the potential for physical adsorp-
tion for a broad range of VOC and PVOC
with the goal of developing a fundamen-
tally consistent model for assessing the
stability of such  compounds  in canisters.
A subsidiary goal was to collect and con-
solidate relevant experimental information
for comparison with model predictions.

Results
  A computer-based  model was devel-
oped  to facilitate solution of the set of
simultaneous equations representing mul-
ticomponent adsorption equilibria. Code
for  this model was written  in Microsoft
Quick-Basic. The  software  package in-
cludes the main program that is used for
all calculations, a Lotus-123 file that con-
tains  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 com-
pounds without requiring  alteration of the
program.
  The program  is menu driven, with all
inputs necessary for a computation being
specified through use of  the menu. Cer-
tain 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
  Mention ol trade names or commercial products
  doas not constitue endorsement or recommenda-
  tion (or USB.
requires a math co-processor for comple-
tion of a calculation in a reasonable time.
  The program assumes a spherical ge-
ometry for the canister,  and the volume of
the canister is used to determine the avail-
able surface area. The surface roughness
also is used to  determine the  available
area. This factor is dependent on the ex-
tent  of polishing — a perfect mirror sur-
face would have a roughness factor of 1,
(i.e., a true surface area equal to the geo-
metrical  surface  area). Electropolishing
typically yields roughness factors of 1.5-2,
and we have calibrated  the model using a
default value of 2 for the surface  rough-
ness 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 analysis condition is also
needed.  The program  adjusts the mea-
sured sampling humidity to the analysis
temperature and  canister pressure, and
provides a warning when the sample prop-
erties are such as to lead to water con-
densation.
  Three  output  choices are available-
screen, printer, and disk. The screen and
printer outputs yield fables of compound
names,  adsorbed phase  mole  fractions
(X|),  final gas phase pressure (PJ, initial
gas  phase pressure (Pln), and the ratio of
Ph/PIn. The  disk output yields a file that
can be imported into Lotus-123 for tabula-
tion  and comparison with experimental
data.
  Default settings are provided for all vari-
ables 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 "remembered" by the program so that
multiple computations involving changes
in relative humidity, temperatures, etc. can
be conducted without re-entry of the com-
position data.
  The program  first uses the Newton-
Raphson method to determine the values
of X, at  the known  initial values  of P,.
Then the amounts of  each component
adsorbed and the corresponding residual
gas  phase pressures are estimated. This
process of estimation of X, and P, is reiter-
ated until convergence is achieved to less
than 1 ppm.
  The model was calibrated by compari-
son  of model predictions with a single set
of experimental results for recovery of VOC
from "dry" canisters. Application  of the
model  to  other  experimental conditions
yields predictions that are qualitatively in
agreement with the experimental data. In
some cases, the  agreement between pre-
dictions and experimental results is nearly
quantitative,  suggesting that the use of
very well qualified data for model calibra-
tion  might improve the overall quality of
the predictions.

Conclusions  and
Recommendations
  Based  solely  on the physicochemical
properties of the compounds (i.e.  inde-
pendent of surface considerations),  the
model predicts displacement of the  more
volatile VOC  and PVOC from a  canister
surface by water vapor at relative humidi-
ties in the range of 1 to 20 percent. This is
generally consistent with experimental ob-
servations, but in most cases, the experi-
mental conditions are not sufficiently char-
acterized 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 addi-
tion 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 implications for application of can-
ister sampling for VOC and PVOC  result
from this work:

 (1)  Measurements of relative humidity
      and temperature should be made
      during the sampling process.  Un-
      der conditions where the relative
      humidity is low enough that the
      model  predicts loss of target  com-
      pounds, 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, but restrictions im-
      posed during shipment may gov-
      ern the maximum pressure allow
      able.
 (3)  When considering the suitability of
      the canister sampling method for
      new compounds, the first param-
      eters to be evaluated should  be
      chemical reactivity and the vapor
      pressure of that compound. Com-
      pounds with equilibrium vapor pres
      sures less than about 1 torr at am-

-------
     bient temperatures may require
     heating of the canister to effect good
     recovery, but heating of canisters
     should be done only with recogni-
     tion 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 ad-
  .   sorption process,  retrospective con-
     siderations 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.
  The model developed in this program
shows considerable promise for qualita-
tive and semi-quantitative explanation of
physical adsorption phenomena with mix-
tures  of trace VOC and PVOC in canis-
ters. At this point,  however,  quantitative
data are lacking with respect to character-
ization 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 depen-
dent on incompletely characterized experi-
mental measurements at low relative hu-
midities.  The accuracy  of the calibration
could be improved  by a  simple  set  of
experiments conducted under very  con-
trolled conditions.  It would be desirable,
for example, to prepare a standard cylin-
der 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 cali-
bration 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 as-
pect of canister analysis should be evalu-
ated systematically.
                                                                                      •U.S. Government Printing Office: 1992— 648-080/60022

-------
 Robert W. Coutantis with Battelle, Columbus, OH 43201.
 William A. McClenny is the EPA Project Officer (see  below).
 The complete report, entitled "Theoretical Evaluation of Stability of Volatile Organic
   Chemicals and Polar Volatile Organic Chemicals in Canisters," (Order No,
   PB92-166941/AS;  Cost: $17.00; subject to change) will be available only from:
         National Technical Information Service
         5285 Port Royal Road
         Springfield, VA 22161
         Telephone: 703-487-4650
 The EPA Project Off her can be contacted at:
         Atmospheric Research and Exposure Assessment Laboratory
         U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental
Research Information
Cincinnati, OH 45268
      BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35

Official Business
Penalty for Private Use $300
EPA/600/SR-92/055

-------