United States
                Environmental Protection
                Agency
 Atmospheric Research and
 Exposure Assessment Laboratory
 Research Triangte Park NC 27711
                Research and Development
 EPA/600/S3-90/097  Feb. 1991
EPA       Project Summary
                Gas  Chromatography/
                Matrix Isolation-Infrared
                Spectrometry  for Air
                Sample  Analysis
               Jeffrey W. Childers
                  This report describes the applica-
               tion of gas chromatography/matrix iso-
               lation-infrared (GC/MI-IR) spectrometry
               to the analysis of several environmen-
               tal air sample extracts. Samples that
               were analyzed include  extracts from
               woodsmoke-impacted air, XAD-2 blanks,
               indoor air, and carpet  samples. The
               emphasis of this report is on the use of
               GC/MI-IR to identify semivolatile organic
               compounds  in these extracts. The
               complementarity of GC/MI-IR spectrom-
               etry and conventional electron-impact
               ionization gas chromatography/mass
               spectrometry (El-GC/JWS) is illustrated.
               The capability of GC/MI-IR to discrimi-
               nate between isomeric compounds that
               are difficult  to  separate chromato-
               graphically and to  distinguish  by
               EI-GC/MS is demonstrated. Preliminary
               results regarding the potential of
               GC/MI-IR spectrometry for the quanti-
               tative  analysis of polycyclic aromatic
               hydrocarbons in air sample extracts are
               presented. Problem areas and modifi-
               cations of a commercial GC/MI-IR sys-
               tem are discussed.
                  This Project Summary was devel-
               oped by EPA's Atmospheric Research
               and Exposure Assessment Laboratory,
               Research  Triangle Park, NC, to an-
               nounce 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 capabilities of gas chromatogra-
               phy/matrix  isolation-infrared (GC/MI-IR)
 spectrometry for characterizing semivola-
 tile organic compounds (SVOCs) in envi-
 ronmental air sample  extracts  were
 evaluated in this study. Since the devel-
 opment of GC/MI-IR spectrometry, several
 applications of the technique to the analy-
 sis of difficult samples have been reported
 in the literature. Many of these applications
 have been to environmental and health-
 related problems. Although several early
 reports of data obtained by GC/MI-IR dealt
 with the examination of the MI-IR spectra
 of environmentally important compounds,
 these studies were not direct applications
 of the technique to actual environmental
 samples. Thus, the capabilities, uses, and
 limits of GC/MI-IR for the analysis of real-
 world environmental samples must be de-
 termined.
   In GC/MI-IR spectrometry, the effluent
 from the GC is trapped in an inert matrix as
 it is deposited on the surface of a rotating,
 gold-plated, cryogenic disk. The  eiuates
 remain frozen indefinitely on the disk and
 are analyzed by MI-IR spectrometry after
 the GC run is completed.  Cryogenic trap-
 ping techniques, such as  GC/MI-IR, offer
 several advantages over  the more con-
 ventional, light-pipe-based  GC/iR systems.
 The principal advantage is an increase in
 sensitivity. The increase  in sensitivity is
 realized because the eluate  is concen-
 trated in a small cross-sectional area on
the cryogenic disk, and most of each GC
 peak is interrogated by the  IR beam during
 each  scan. Also, because the eluate is
frozen on the disk indefinitely, signal-av-
eraging is allowed  on any GC  peak to
 increase the signal-to-noise ratio to the
level necessary to obtain  an  identifiable

          S/Q Printed on Recycled Paper

-------
spectrum. Band broadening due to  mo-
lecular rotation and intermolecular interac-
tions is also minimized because the eluate
is trapped in an inert matrix at a very low
temperature. This results  in sharper IR
bands, which  affords greater molecular
specificity to aid in identifying components
in  complex mixtures and differentiating
between isomers and other closely related
compounds. In cryogenic-trapping GC/IR
systems, the separation and detection of
eluates are independent steps. Therefore,
the chromatography and the spectrometry
can be  optimized without compromising
one or the other.
    In this report, we demonstrate the ap-
plication of GC/MI-IR spectrometry to the
analysis  of extracts of woodsmoke-im-
pacted air, indoor air, and carpet samples.
Samples that were  analyzed include the
following:
 •   extracts of air samples collected on
     XAD-2 cartridges during afield study
     in  a woodsmoke-impacted residen-
     tial area
     extracts of XAD-2 blank cartridges
     that elicited an abnormally  high bio-
     assay response
 •   composite samples collected  on
     quartz-fiber filter from woodsmoke-
     and mobile source-impacted areas
 *   a nhro-substHuted polycyclic aromatic
     hydrocarbon  (PAH) fraction of a
     pooled mobile source air sample and
     selected  nitro-PAH standards
 •   selected  polyurethane foam (PUF)
     and filter extracts from a pilot resi-
     dential indoor air study
     extracts from air samples  collected
     on  XAD-2 cartridges in a complaint
     building
 •   extracts from selected carpet samples
    The  primary focus of the report is on
the evaluation of the capabilities of GC/MI-
IR for identifying specific SVOCs in com-
plex environmental air sample  extracts.
The complementarity of GC/MI-IR and GC/
mass spectrometry (MS) for the analysis of
environmental samples is also illustrated.
Most of the results described in this report
pertain to the qualitative identification  of
unknowns in various sample extracts. We
also present preliminary results regarding
the potential of GC/MI-IR for quantitative
analyses  and  discuss problems  encoun-
tered with the GC/MI-IR system.

Procedure
    All GC/MI-IR data were collected on a
Mattson  Instruments  (Madison,  Wl)
Cryolect system. This system consists of a
Mattson  Instruments Sirius  100 Fourier
transform IR spectrometer and Starlab data
system,  a Hewlett-Packard (HP) 5890A
capillary GC, and a Mattson Instruments
matrix-isolation cryogenic module.  In the
Cryolect system, the effluent of the analyti-
cal GC column is split, with 20% directed
to a flame ionization detector (FID); the
remaining 80% is directed through an open-
spirt cross and then through a heated,
fused-silica transfer line to the cryogenic
disk. The cryogenic disk is housed in an
evacuated chamber and  is maintained at
14 K during deposition and spectral analy-
sis of the sample extract. The MI-IR spec-
tra were obtained by coadding 128 scans
at a nominal resolution of 4 cm'1.
   The GC/MS results were obtained on
an HP  5970B mass  selective  detector
(MSD) interfaced to an HP 5980A capillary
GC. Similar chromatographic conditions,
including on-column  injection,  a (95%)-
dimethyl-(5%)-diphenyl pplysiloxane capil-
lary column, and  identical temperature
programs, were  used for both  GC/MI-IR
and GC/MS analyses.

Results and Discussion

Extracts of XAD-2 Cartridges from
a Field Study in a Woodsmoke-
impacted Residential Area

   Representative extracts of air samples
collected onto XAD-2 cartridges during a
field study designed to determine the im-
pact of residential wood combustion on
relatively simple  airsheds were analyzed
by GC/MI-IR and GC/MS. We analyzed
extracts of air samples that were collected
outdoors at night, outdoors during the day,
indoors at residences with a wood-fired
stove, and indoors at'residences without a
wood-fired stove.  The chromatographic
profiles of  the  different samples were
qualitatively very similar, with each  extract
containing the same major components.
Preliminary GC/MS analyses indicated that
the major components of each extract were
alkylbenzenes, including toluene, ethyl-
benzene, xylenes, ethyltoluenes,  and
trimethylbenzenes. Specific positional iso-
mers could not be  distinguished by
GC/MS but were identified by GC/MI-IR.
The MI-IR spectra of alkylbenzenes exhibit
characteristic strong absorption bands be-
tween 1000 and 650 cnr1, which are due to
the vibrational mode of the C-H  out-of-
plane deformation. The specific frequen-
cies of these bands are determined by the
position of the substituents on the ring and
the number of adjacent hydrogen atoms
remaining on the ring. Thus, the GC/MI-IR
analysis of the  XAD-2 extracts  of the
woodsmoke-impacted air enabled specific
isomers, such as 3- and 4-ethyltoluene, to
be identified even if they were not sepa-
rated chromatographically (see Figure 1).
In addition to the alkylbenzenes, guaiacol
(a potential woodsmoke tracer) was identi-
fied in an outdoor-nighttime residential air
sample but was not detected in the indoor
or daytime samples. Compounds identified
in the indoor air samples, but not the out-
door samples, included d-limonene,  1,4-
dichlorobenzene, normal aldehydes,  and
several alkanes.

Extracts of XAD-2 Blank Cartridges
that Elicited an Abnormally High
Bioassay Response

   Two concentrated  extracts of XAD-2
blank cartridges that yielded abnormally
high bioassay responses were analyzed
by GC/MI-IR and GC/MS. Both extracts
contained 2-ethyl-1-hexanol,  C4-ben-
zenes, acetophenone, substituted  aceto-
phenones, an ethylstyrene, a silane, an
unidentified  aldehyde, naphthalene,  and
several  esters,  phthalates, and  alkanes.
One extract also contained two major com-
ponents that were not detected in the other.
One of  these components was  identified
as 4-chlorophenylsulfone, and  the other
was tentatively identified as triphenylphos-
phine. No compounds were  identified in
the blank extracts that would  be expected
to contribute significantly to the mutage-
nic'rty of these extracts.

Quartz-Fiber Filter Composites
from Woodsmoke- and Mobile
Source-Impacted Areas

   Two nonpolar fractions of ambient filter
composite extracts, a woodsmoke com-
posite and a mobile  source composite,
were  analyzed by GC/MI-IR and GC/MS.
The chromatographic profiles of each ex-
tract were similar;  both contained com-
pounds associated'with woodsmoke, such
as 1-methyl-7-isopropylphenanthrene (re-
tene) and the methyl ester of dehydroabietic
acid, in  addition to a series of alkanes and
siloxanes and a large amount  of back-
ground material. In spite of the large amount
of background material  present  in  the
chromatogram, several PAH  compounds,
including  fluoranthene,  pyrene,  1-
methylpyrene, and 5-methylchryse.ne, were
identified in  both  extracts.  By using
GC/MI-IR, the presence of coeluting iso-
meric and closely  related PAHs  in  the
composite extracts was confirmed. For ex-
ample, an unambiguous distinction between
cyclopenta[c,d]pyrene and benz[a]anthra-
cene was  made,  and  the. chrysene-
triphenylene isomer  pair was distinguished
(see Figure 2).

-------
   1600
1400
   1200               1000

Wavenumber
800
Figure 1.    MI-1R spectra of (A) a component from an extract of an XAD-2 cartridge collected during a field study in a woodsmoke-impacted
          residential area, (B) 3-ethyltoluene reference standard, and (C) 4-ethyltoluene reference standard.

-------
        1600
                      r       r
1400              1200
            Wavenumber
1000
800
Figure 2.    Ml-IR spectra of (A) a component of a filter composite extract from woodsmoke- and mobile-source-impacted areas, (B) chrysene
          reference standard, and (C) triphenylene reference standard.              , .        ,

-------
 A Nitro-PAH Fraction and
 Selected Nitro-PAH Standards

    A nitro-PAH fraction of a pooled mo-
 bile source air sample extract was ana-
 lyzed by  GC/MI-IR.  Reference  MI-IR
 spectra of 1-nitropyrene, 2-nitrofluor-
 anthene,  2-nitropyrene, and  6-nitro-
 benzo[ajpyrene were also acquired.
 Although high-quality MI-IR reference
 spectra of the standard compounds could
 be obtained, no target nitro-PAH were
 detected  in the fractionated sample  ex-
 tract. The levels of nitro-PAHs in this frac-
 tion were estimated to be less than 100
 pg/uL, which is below the detection limits
 of the GC/MI-IR system  for these com-
 pounds. The reference spectra  obtained
 on the GC/MI-IR system were  compa-
 rable to MI-IR spectra of nitro-PAHs in the
 literature that were obtained under con-
 ventional slow-spray-on MI-IR conditions.

 Selected PUF and Filter Extracts
 from a Pilot Residential Indoor Air
 Study

    Selected  PUF and quartz-fiber filter
 extracts collected during a pilot residen-
 tial indoor air study were analyzed by GC/
 MI-IR and GC/MS. The results from these
 analyses illustrated the difference in sen-
 sitivity and selectivity of GC/MI-IR and
 GC/MS for different chemical classes. For
 instance, the GC/MI-IR analysis showed
 the presence of a number of carbonyl-
 containing compounds, such as  esters
 and carboxylic acids, whereas the scan-
 ning GC/MSD results were dominated by
 a  series of  lorig-chain, alkyl-substituted
 benzenes. The  three-ring PAH isomers,
 phenanthrene and anthracene, were dis-
 tinguished by GC/MI-IR in the presence
 of background material in the chromato-
 gram of a PUF extract.  Nicotine and
 tetraethylene glycol dimethylether  were
 identified as the  major components in  an
 extract of SVOCs collected on a quartz-
 fiber filter in a smoker's home.

 Extracts from XAD-2 Cartridges
 Collected in a Complaint Building

   The GC/MI-IR and  GC/MS systems
 were used to analyze an  extract of an
 XAD-2 cartridge collected from a room in
 an office complex during  an  episode  of
 health-related complaints by employees.
The extracts contained  several n-al-
kanes, branched alkanes, and alkylated
benzenes. Other major components  of
the extract that were identified included 2-
butoxyethanol, benzaldehyde, 1,4-dichlo-
robenzene, an  n-aldehyde, a-terpineol,
4-phenylcyclohexene,  and two  2,2,4-
 trimethylpentane(1,3)diol isomers. The
 presence  of 4-phenylcyclohexene, a by-
 product of the carpet manufacturing pro-
 -cess and  a suspected irritant  associated
 with the characteristic odor of new carpet,
 prompted  an investigation into the SVOCs
 and nonvolatile organic compounds associ-
 ated with new carpet.

 Extracts from Selected Carpet
 Samples

    Several carpet samples were Soxhlet-
 extracted  in methylene chloride, concen-
 trated, and then analyzed by GC/MI-IR and
 GC/MS. The GC/MI-IR analyses were used
 primarily to confirm tentative identifications
 made by GC/MS analyses. In addition, the
 unique features of MI-IR spectra, such as
 split carbonyl peaks and sharp O-H stretch-
 ing bands, that are not found in condensed-
 phase or vapor-phase IR spectra aided in
 identifying  components in the  extracts.
    In the analysis of these carpet samples,
 the  GC/MI-IR results were used to confirm
 an error in  a commercial mass spectral data
 base and  to distinguish between SVOCs
 that have  very similar mass spectra  but
 signficantly different MI-IR spectra, such as
 6-aminohexanoic acid and e-caprolactam.
 The MI-IR spectrum of e-caprolactam ex-
 hibits a split in the carbonyl absorption band,
 which would normally indicate the presence
 of more than one carbonyl moiety (see Fig-
 ure 3). However, e-caprolactam has only one
 carbonyl group. A search of the literature
 revealed that the MI-IR spectra of lactams,
 as  well  as many other compounds that
 contain a carbonyl group, can exhibit multiple
 carbonyl absorption bands. In the case of
 lactams, this multiplicity is attributed to the
 isolation of discrete conformers in the matrix.
 The MI-IR  spectra of several phenols and
 diols identified in the carpet extracts exhib-
 ited sharp O-H stretching bands, which indi-
 cated the lack of intermolecular hydrogen
 bonding under the GC/MI-IR deposition
 conditions.  The GC/MI-IR results were also
 used to  identify background material, which .
 was originally assigned to column bleed by
 interpretation of the GC/MS results, as an
 isocyanate.

 Preliminary Investigation of the
 Quantitative Capabilities of GC/MI-
 IR

   We have performed a preliminary inves-
tigation  of  the quantitative capabilities of
GC/MI-IR spectrometry for the determina-
tion  of target PAHs in environmental air
sample extracts. Our initial goals were to
determine the  repeatability, the  dynamic
range, and  the detection limits of the GC/
MI-IR system for PAH  compounds. To de-
 termine the repeatability  of the GC/MI-IR
 measurement, we analyzed  replicate in-
 jections of a 10-ng/u.L standard PAH mix-
 ture containing 10-ng/u.L of  1-methyl-
 benz[a]anthracene as an internal standard.
 The MI-IR  absorbance maximum of the
 analyte was normalized to the MI-IR ab-
 sorbance maximum of the internal stan-
 dard. Likewise, the FID  peak area was
 normalized to the  FID peak area of the
 internal standard.  The average  relative
 standard deviation of six  replicate analy-
 ses was approximately 20% for the nor-
 malized MI-IR absorbance  and on the order
 of 1% for the normalized FID peak areas.
 Because of the poor repeatability of the
 GC/MI-IR measurements, the dynamic
 range and the detection limits of the sys-
 tem for PAHs could not  be  determined.
 Fundamental experiments are currently
 underway to determine the sources of er-
 ror in GC/MI-IR quantitative analyses.

 Conclusions and
 Recommendations
    In general, we  have found  GC/MI-IR
 spectrometry to be very useful for identify-
 ing SVOCs in  a variety of environmental
 air sample  extracts. In many cases the
 information  obtained from the  GC/MI-IR
 analyses supported 'tentative ,;.•«;,ideations
 made from preliminary GC/MS  analyses.
 In  cases where a definitive  identification
 could not be made, the GC/MI-IR results
 often supplied enough  information for  a
 general compound  classification  to be
 made. In addition, the  GC/MI-IR analysis
 often revealed unique information that could
 not easily by obtained by other existing
 analytical techniques. In particular, we have
 demonstrated the capabilities of GC/MI-IR
 to discriminate between compounds such
 as  alkylbenzene  positional isomers and
 PAH isomers, which are difficult to sepa-
 rate chromatographically and to distinguish
 by conventional GC/MS. Although absolute
 detection limits have not been determined,
 we have found that the minimum quanti-
 ties of analyte needed for identifications by
 GC/MI-IR are very similar to those required
 by GC/MS in the scan  mode.  In practical
 terms, if a compound can be detected by
 scanning GC/MS, it is likely that an identifi-
 able MI-IR spectrum of that component
 can be obtained on the GC/MI-IR system.
   We recommend investigating the po-
 tential of interfacing the GC/MI-IR system
 to an MSD system to fully exploit the
 complementarity of the two  techniques.
 We also  recommend further evaluation of
the quantitative capabilities of the GC/MI-
 IR technique and a thorough investigation
to elucidate the sources of error in GC/MI-
 IR quantitative measurements.

-------
    B
     i  i  i   i  I   I
4000       3500
                                 i  i   |  r  i  i   i  |  i
                        3000       2500        2000
                                  Wavenumber
1500       1000
Figure 3.    MI-IR spectra of (A) a component of a carpet extract and (B) e-caprotactam reference standard.

-------

-------
Jeffrey W. Childersis with NSI Technology Services Corporation, Environmental
  Sciences, Research Triangle Park, NC 27709,
Nancy K. Wilson is the EPA Project Officer (see below).
The complete report, entitled "Gas Chromatography/Matrix Isolation-Infrared Spectrom-
  etry For Air Sample Analysis," (Order No. PB91-136317/AS; Cost: $23.00, 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 Officer 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/S3-90/097

-------