United States
 Environmental Protection
 Agency
 Environmental Monitoring
 Systems Laboratory
 Research Triangle Park NC  27711
 Research and Development
 EPA/600/S4-85/067  Jan. 1986
 Project  Summary
 Comparison  of Ambient Air
 Sampling  Techniques  for
 Volatile  Organic  Compounds

 M. W. Holdren, D. L Smith, and R. N. Smith
  The objective of this study was to
 carry out a comparison of ambient air
 sampling techniques. A series of four-
 teen experimental sampling runs were
 carried out at a field site adjacent to
 Battelle's chemistry laboratory. Ambi-
 ent air was drawn through a sampling
 manifold and was continuously spiked
 with a mixture of fifteen volatile or-
 ganic compounds (VOCs) to give  con-
 centrations 1 to 3 ng/l above back-
 ground  air. These compounds were
 chloroethene,  1,1-dichloroethene,
 dichloromethane, 3-chloropropene,
 1,1,2-trichloro-1,2,2-trifluoroethane,
 trichloromethane, 1,2-dichloroethane,
 1,1,1-trichloroethane, benzene, tetra-
 chloromethane, trichloroethene,
 toluene, tetrachloroethene, chloroben-
 zene, and 1,2-dimethylbenzene.
  During the sampling period passi-
 vated stainless steel canisters were uti-
 lized to collect  whole air integrated
 samples upstream and downstream of
 the spiking region. An automated gas
 chromatographic system was  em-
 ployed to analyze the contents of the
 sample canisters using capillary
 column separation and multiple detec-
 tors for sample analysis (electron  cap-
 ture, flame ionization and mass selec-
 tive detectors). Tenax GC adsorbent
 samples were also collected down-
 stream of the spiking region in parallel
 with the integrated canister samples.
 These samples were analyzed by stand-
 ard gas chromatographic/mass spectra-
 metric techniques.
  In comparing  analytical results,
whole air collection via canisters gave
better precision for the compounds of
interest in this study. An estimate of
 precision (% standard deviation) for the
 canister sampling method ranged from
 4 to 10 percent (using the mass selec-
 tive detector). For Tenax, precision val-
 ues ranged from 8 to 16 percent. When
 comparing measured  recoveries with
 expected concentrations, the canister
 sampling approach yielded values from
 89  to 120 percent. Recoveries, using
 Tenax adsorbent, ranged from 8 to 104
 percent.
  This Project Summary was devel-
 oped by EPA's Environmental Monitor-
 ing Systems Laboratory, Research Tri-
 angle 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
  The Methods Development Branch of
• the Environmental Monitoring Systems
 Laboratory (EMSL) is  responsible for
 the development  and evaluation of
 state-of-the-art and emerging analytical
 techniques for the determination of or-
 ganic compounds in ambient air. Re-
 cently, a priority listing of volatile or-
 ganics has been established and the
 EMSL is focusing on further develop-
 ment of analytical methodology associ-
 ated with the detection of these com-
 pounds.  Primary emphasis has been
 placed on developing field-compatible
 analytical systems.
  During the past two years a joint ef-
 fort by Battelle's Columbus Laborato-
 ries and EPA has resulted in the devel-
 opment and evaluation of a prototype
 system for the analyses of sixteen
volatile organic compounds (VOCs).

-------
The  system consists of a reduced-
temperature trap for condensing organ-
ics from  ambient air, and a capillary
column gas chromatograph (GC) with
flame ionization (FID) and electron cap-
ture  detectors  (ECD)  to quantitatively
monitor the VOCs. Software develop-
ment using the basic programming ca-
pability of the GC system permits auto-
matic sampling and  analysis to be
achieved  with minimal operator  inter-
facing. The prototype  system has been
tested with respect to sample drying
procedures (selective removal of water
vapor), co-collection of reactive ambi-
ent air species, and collection and re-
lease efficiency. The prototype design
has also been evaluated by comparing
two nominally identical systems using
calibration  mixtures and ambient  air.
Excellent  results were obtained during
these laboratory tests and the following
recommendations were made:
  • The automated GC system should
    be field tested. During the field
   tests the reduced temperature trap-
    ping  system  should be compared
   with  other preconcentration  tech-
    niques such as solid adsorbents
    and passive dosimeters.
  • Many of the target compounds
   tested  in  the laboratory also  co-
    elute with  other ambient air spe-
    cies.  Although the combination of
    capillary column and flame ioniza-
    tion and electron capture detectors
    alleviate some of the identification
    and quantitation concerns, other
    more  selective  detectors are
    needed. Integrating a mass selec-
    tive detector (MSD) into the  auto-
    mated  gas chromatographic sys-
    tem is  recommended. The greater
    specificity of this detector over
    other detection systems will allow
    better differentiation of co-eluting
    GC peaks and thus improve quanti-
    tative capability.
  The full report focuses on the com-
parison of two ambient air sampling
techniques. Specifically, whole air col-
lection into specially  treated canisters
was  compared with Tenax GC adsor-
bent sampling techniques.  Analysis of
the canisters' contents was accom-
plished with the automated GC system,
while Tenax adsorbed samples were
analyzed by standard gas chromato-
graphic mass spectrometric techniques.
The comparison study was carried  out
at a site adjacent to Battelle's chemistry
laboratory. To assure detectable con-
centrations, the sampling manifold was
spiked with a mixture of fifteen VOCs
and the comparison focused on these
compounds.

Procedure
  Tenax and whole air sampling (using
passivated stainless steel canisters)
were operated in parallel employing the
sampling manifold and apparatus
shown in Figure  1. Fourteen test runs
were carried out at the field site. Eleven
sampling runs were made with spiked
ambient air while the remaining three
runs involved the injection of calibra-
tion mixtures. Each test run lasted two
hours.
  Canister samples were analyzed with
an automated capillary column gas
chromatograph equipped with electron
capture, flame ionization and mass se-
lective detectors. This instrument was
also programmed to collect and analyze
"real time" integrated samples directly
from the manifold.
  Five, ten (duplicate), and twenty liter
samples were collected using Tenax ad-
sorbent sampling  techniques. These
samples were  analyzed by gas chro-
matography/mass spectrometry.

Results and Discussion

Overview
  The comparison study  was carried
out at a field site adjacent to Battelle's
chemistry laboratory. The evaluation
study consisted of fourteen sampling
runs as listed in Table 1. Runs 4 and 11
were  clean air experiments using dilu-
tion  mixtures from an  audit cylinder
supplied by EPA, while run 5 was a sim-
ilar run using Battelle's calibration cylin-
der. The remaining sampling runs were
made with spiked ambient air. The ana-
lytical precision of  the three detectors
from the automated GC system was ob-
tained by replicate analysis of canisters
collected from the sampling  manifold
during the tests. A comparison was also
made of concentrations of the six target
compounds that were detected by both
the electron capture and mass selective
detector. Whole air canister concentra-
tions  determined by the cryogenic GC
system (using the mass selective detec-
tor response) were also  compared with
those values obtained from Tenax GC
adsorbent devices.

Analytical Precision of MSD,
ECD, and FID Detectors
  An  estimate of the analytical preci-
sion of the cryogenic GC system was
determined from triplicate analyses  of
spiked and  background canisters col-
lected during each sample period. For
these analyses, twelve  compounds
were detected with the FID;  six com-
pounds were detected with  the ECD.
Due to the limitation in software only six
compounds  were monitored with the
MSD.
  A summary of the analytical precision
of all three detectors is shown in
Table 2. In viewing the spiked canister
data set, an average analytical precision
of ±7.3 percent was found for the six
compounds analyzed with the MSD. Te-
trachloroethene had the lowest value
(±4.0%) while (Freon-113) displayed the
highest relative standard deviation
(RSD) value (±10.1%). A significantly
lower average RSD value of  ±3.7 per-
cent was obtained for the same six com-
pounds analyzed with the ECD. The RSD
values ranged from ±1.3 percent (tetra-
chloromethane)  to  ±10.1  percent
(Freon-113). With the FID detector an av-
erage RSD value (±7.7%) similar to the
MSD was found. However when the six
additional compounds that were also
detected by the FID were included in the
calculations, the  average RSD  value
was lowered to ±5.8 percent. Toluene
exhibited the lowest RSD value (±1.8%)
while tetrachloromethane produced the
highest RSD value (±14.4%).
  The background canister data, also
shown in Table 2, follow the same trend
as the spiked canister data. An average
RSD  value of ±7.2 percent was found
for the six compounds analyzed with
the MSD, while an average RSD value of
±3.6 percent was obtained  with the
ECD. With the FID detector, the average
RSD value was ±8.5 percent. When the
five additional FID detected compounds
were included in the FID calculations,
the average RSD value was lowered to
±7.0 percent (chloroethene was not
found in background air i.e. <0.1 ng/l).

Comparison of ECD and MSD
Responses
  The  six compounds that were
detected by both the electron  cap-
ture and  mass  selective detectors
were   1,1,2-trich I o r o -1 ,2,2-
trifluoroethane, trichloromethane,
1,1,1-trich lo roetha ne, tetra-
chloromethane,  trichloroethene and
tetrachloroethene. A  comparison of
concentrations was carried out by de-
termining average  percent relative dif-
ference values between the two detec-
tors  i.e.,  [100 (ECD  response-MSD
response)/average response]. Three
data  sets were examined and include
the spiked canisters, the background

-------
                                                                                                         Exhaust
                                                                                                           \
                                                                                                       Variable Speed
                                                                                                         Pump and
                                                                                                       Dry Test Meter
                                                                              Solid
                                                                           Adsorbent
                                                                            Sampling
                                                                             System
Calibration Cyl.


    Audit Cyl.
Figure 1.    Diagram of sampling manifold and instrumentation employed during field program.
Table 1.    List of Sampling Runs
Run No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Sampling date
7/12/84
7/17/84
7/19/84
7/20/84
7/24/84
7/26/84
7/27/84
7/30/84
7/31/84
8/02/84
8/03/84
8/06/84
8/07/84
8/09/84
Description
Ambient air (trial run)
Ambient air
Ambient air
Audit cylinder
Calibration cylinder
Ambient air
Ambient air
Ambient air
Ambient air
Ambient air
Audit cylinder
Ambient air
Ambient air
Ambient air
                                                           canisters, and samples collected in real
                                                           time.
                                                             A summary of the  average percent
                                                           relative difference values from the three
                                                           data sets is shown in Table 3. For 1,1,2-
                                                           trichloro-1,2,2-trifluoroethane, negative
                                                           average percent relative difference val-
                                                           ues were observed for all three data sets
                                                           and indicated  that the MSD recorded
                                                           concentrations were in general higher
                                                           than the corresponding ECD values. On
                                                           the other hand, trichloromethane, tetra-
                                                           chloromethane  and trichloroethene
                                                           gave positive average percent relative
                                                           difference values for all three data sets.
                                                           The positive values denoted higher rela-

-------
Table 2.   Analytical Precision Data From Replicate Analyses of Spiked and Background Canisters

                                               Spiked canisters
                                            detector precision <±%)
                                                   Background canisters
                                                  detector precision (±%>
Compound
Chloroethene
1, 1,2-Trichloro- 1,2,2-trifluoroethane
Trichloromethane
1,2-Dichloroethane
1, 1, 1-Trichloroethane
Benzene
Tetrachloromethane
Trichloroethene
Toluene
Tetrachloroethene
Chlorobenzene
1,2-Dimethylbenzene
MSD
NM(t»
10.1
7.3
NM
8.7
NM
6.4
7.3
NM
4.0
NM
NM
ECD
NM
10.1
4.3
NM
2.0
NM
1.3
2.8
NM
1.7
NM
NM
FID
7.4
10.0
3.3
3.6
3.6
1.6
14.4
11.9
1.8
2.7
2.9
6.7
MSD
NM
10.4
12.4
NM
4.5
NM
4.9
8.6
NM
2.3
NM
NM
ECD
NM
3.9
9.2
NM
2.0
NM
1.7
2.7
NM
1.8
NM
NM
FID
NM
18.7
1.6
1.1
2.9
3.6
15.1
7.2
1.7
5.2
15.2
4.9
NM - not detected.

Table 3.   Average Percent Relative Difference Values Obtained By Comparing Electron Cap-
         ture and Mass Selective Detector Responses During the Field Study1"1

                                        Average relative difference values (%)
Compound
1, 1,2-Trichloro- 1,2,2-trifluoroethane
Trichloromethane
1, 1, 1-Trichloroethane
Tetrachloromethane
Trichloroethene
Tetrachloroethene

spiked
canisters
-11.8
9.2
3.2
23.2
19.6
-3.6
(ECD response -
background
canisters
-8.8
15.0
-6.0
0.5
36.9
5.7
MSDresponse\ ,„
real time
samples
-12.4
2.2
-1.9
23.4
24.1
-5.7

live ECD responses for these three com-
pounds. Tetrachloroethene and 1,1,1-
trichloroethane showed no consistent
bias. The spiked canister and real time
sampling data showed very similar per-
cent relative difference values  as one
would expect since both types of sam-
pling were carried out downstream of
the sample spiking port. Although the
same general trend  is also observed
with the background  canister data, the
somewhat larger variation resulted be-
cause these samples were collected up-
stream of the spiking port and therefore
generally  contained much  lower
concentrations of the  six  target
compounds.

Tenax GC Adsorbent Versus
Whole Air Collection/Cryogenic
Preconcentration  Comparison
Study for Ambient Air
  In order to compare the data from the
two collection methods, the appropriate
Tenax GS concentration values were ra-
tioed to the corresponding concentra-
tions found by cryogenic trapping/GC
analysis of canister samples. The ratios
provided a  direct measure of the rela-
                                  4
average response     /

  tive recovery of the Tenax GC adsorbent
  for each compound. The comparison
  was done on this basis because the can-
  ister/cryogenic trapping concentrations
  generally agreed better with the diluted
  concentration generated in the mani-
  fold (see full report). Likewise literature
  data has shown that several of the fif-
  teen target compounds have low break-
  through volumes on Tenax GC and are
  therefore not efficiently collected.
    In Table 4 a summary of the perform-
  ance data for Tenax GC relative to canis-
  ter/cryogenic trapping  is given. Each
  compound is listed along with its break-
  through volume and the average Tenax
  GC recovery relative to canister/cryo-
  genic trapping  obtained over the ten
  sampling runs. As indicated in the table
  for those compounds in which break-
  through  volumes were very low, the
  nominally 5 liter Tenax GC adsorbent
  sample was used in the comparisons;
  the remaining  concentration  values
  were taken from the 10 liter samples.
  The values used for the canister sam-
  ples were obtained using the mass se-
  lective detector, because this detection
  system was less subject to potential in-
  terferences. However, the toluene and
1,2-dimethylbenzene values were ob-
tained using FID due to limitations on
the number of ions which could be
monitored using the mass selective de-
tector.
  For the compounds, chloroethene
and 1,1-dichloroethene, no meaningful
data was obtained with the Tenax GC
adsorbent.  With the exception of the
anomalous behavior of 1,1,2-trichloro-
1,2,2-trifluoroethane, the compounds
dichloromethane through trichloro-
ethene gave very reasonable relative
recoveries ranging from 83 to  130 per-
cent. However a dramatic fall  off in
recovery is observed for the less vola-
tile  compounds, toluene, tetra-
chloroethene, chlorobenzene and 1,2-
dimethylbenzene. The reason for the
relatively  low recoveries for  these
higher boiling compounds is unclear.

Conclusions
  The significant findings from  this
study are presented below:
  (1) Replicate analysis  of canisters
     collected  upstream and down-
     stream of the spiking region with
     the automated GC  system pro-
     vided precision data for each de-
     tector. For concentrations varying
     from 1 to 70 ng/l an average BSD
     of ±3.7 percent was found for the
     six compounds detected by the
     electron capture detector (1,1,2-
     trichloro-1,2,2-trifluoroethane,
     trichlo rometha ne,  1,1,1-
     trichloroethane,  tetra-
     chloromethane, trichloroethene,
     and tetrachloroethene). An RSD
     of ±7.3 percent was obtained for
     the same six compounds ana-
     lyzed  with the mass selective de-
     tector. A similar value of ±7.7 per-

-------
Table 4.    Performance Data for Tenax Relative to Cryogenic Trappin'g/GC Analysis of Canis-
          ter Samples

                                                          AverageTenax GC
                                                          recovery relative to
     Compound                volume'31, liters/cartridge       cryogenic trapping, %
 Tenax GC breakthrough
volume1"1, liters/cartridge
Chloroethene
1, 1-Dichloroethene
Dichloromethane
3-Chloropropenelcl
1,1,2-Trichloro-1,2,
2-trifluoroethane
Trichloromethane<0>
1,2-Dichloroethanelc>
1, 1, 1-Trichloroethane
Benzene
Tetrachloromethanelc>
Trichloroethene
Toluene
Tetrachloroethene
Chlorobenzene
1,2-Dimethylbenzene
0.8
Not Given
4
6
Not Given

13
18
9
27
13
28
122
106
249
334
(b)
Ibl
83 (21)
87 (35)
39 (25)

WO (36)
100 (15)
130 (42)
100 (18)
110(37)
1 12 (26)
70 (19)
88 (27)
78 (35)
55 (21)
Data from reference 5 in the full report at 90°F.
Low volume (nominally 5 liters) Tenax GC value used for these compounds. Medium volume
  (nominally 10 liters) Tenax GC value used for all other compounds.
 Value in parentheses is standard deviation for all sampling runs (excluding audit and calibra-
  tion cylinder sampling runs).
     cent was found with the flame
     ionization detector.
  (2) The six compounds that were de-
     tected by both the  electron cap-
     ture detector and mass selective
     detector were compared using
     percent relative difference values
     [100  (ECD  response-MSD re-
     sponsel/average response]. Rela-
     tive difference values ranged
     from  -11.8 percent (1,1,2-
     trichloro-1,2,2-trifluoroethane) to
     36.9 percent (trichloroethene).
  (3) Tenax GC adsorbent samples also
     were collected in parallel with the
     integrated canister  samples and
     then were analyzed by standard
     gas chromatographic mass spec-
     trometric techniques. Analysis of
     duplicate Tenax samples (10 L) re-
     sulted in RSD values that ranged
     from 8 to 16 percent. Recovery rel-
     ative to canister sampling was as
     follows: chloroethene  and 1,1-
     dichloroethene  (no meaningful
     data), dichloromethane  (83%), 3-
     chloropropene (87%),  1,1,2-
     trichloro-1,2,2-trifluoroethane
     (39%), trichloromethane (100%),
     1,2-dichloroethane (100%), 1,1,1-
     trichloroethane (130%), benzene
     (110%), tetrachloromethane
     (110%), trichloroethene (112%),
     toluene (70%), tetrachloroethene
     (88%), chlorobenzene (78%) and
     1,2-dimethylbenzene (55%).
         Recommendations
           Additional studies should be under-
         taken under controlled conditions with
         analytical uncertainties minimized  as
         much as possible  by employing the
         same analysis procedure  for all col-
         lected samples. Techniques to be com-
         pared should  include distributive air
         volume sampling with Tenax  adsor-
         bent, passive sampling using personal
         exposure devices, and whole air collec-
         tion in canisters.
            M. W. Holdren, D. L. Smith, and R. N. Smith are with Battelle's Columbus
              Laboratories. Columbus, OH 43201.
            W. M. McClenny is the EPA Project Officer (see below).
            The complete report, entitled "Comparison of Ambient Air Sampling Techniques
              for Volatile Organic Compounds," (Order No. PB 86-120 953/AS; Cost: $11.95,
              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 Officer can be contacted at:
                   Environmental Monitoring Systems Laboratory
                   U.S. Environmental Protection Agency
                   Research Triangle Park, NC 27711
                                                                             U. S. GOVERNMENT PRINTING OFFICE: 1986/646-116/20759

-------
United States
Environmental Protection
Agency
Official Business
Penalty for Private Use $300


EPA/600/S4-85/067
                        Center for Environmental Research
                        Information
                        Cincinnati OH 45268
                                                                                                            0 i.
0000329

U  S
                       PS
                         PROTECTION  AGENCY

-------