Analytica Chimica Acta 400 (1999) 153–162

Field air analysis with SPME device

Jacek Koziel, Mingyu Jia, Abir Khaled, Japheth Noah, Janusz Pawliszyn

∗

Department of Chemistry, University of Waterloo, Waterloo, Ont., Canada N2L 3G1

Received 15 April 1999; accepted 25 June 1999

Abstract

Solid-phase microextraction (SPME) devices were used for a wide scope of air-monitoring including field sampling and

analysis of volatile organic compounds (VOCs), formaldehyde, and particulate matter (PM) in air. Grab (instantaneous) and
time-weighted average (TWA) sampling were accomplished using exposed and retracted SPME fibers, respectively. Sampling
time varied from 1 to 75 min, followed by analysis with a gas chromatograph (GC). A portable GC equipped with unique,
in-series detectors: photoionization (PID), flame ionization (FID), and dry electrolytic conductivity (DELCD), provided
almost real-time analysis and speciation for common VOCs during an indoor air quality surveys. Indoor air samples collected
with SPME devices were compared with those collected using conventional National Institute for Occupational Safety and
Health (NIOSH) methods. Air concentrations measured with the SPME device were as low as 700 parts-per-trillion (ppt) for
semi-volatile organic compounds. SPME methodology proved to be more sensitive than conventional methods, and provided
a simple approach for fast, cost-effective sampling and analysis of common VOCs in indoor air. SPME technology combined
with fast portable GC reduced the sampling and analysis time to less than 15 min. The configuration offered the conveniences
of immediate on-site monitoring and decision making, that are not possible with conventional methods. In addition, SPME
fibers were applied to sampling of particulate matter in diesel engine exhaust. Linear uptake and particulate build-up on the
fiber were observed. Preliminary research suggests that SPME fibers could also be applied to sampling of airborne particulate
matter. ©1999 Elsevier Science B.V. All rights reserved.

Keywords: Air analysis; Formaldehyde; Indoor air quality; Particulate matter; Solid-phase microextraction; Volatile organic compounds

1. Introduction

There has been an increasing demand for improved

air sampling and analysis methods due to the: (1)
increasing number of regulated sources and pollu-
tants; (2) recognition of toxicity for chemicals that
were thought to be relatively harmless (e.g. formalde-
hyde); (3) recognition of toxic effects of more chem-

∗

Corresponding author. Tel.: [phone]; fax: +1-519-

746-0435
E-mail address: [e-mail] (J. Pawliszyn)

icals at even lower concentrations, and (4) increas-
ing public awareness of both indoor and outdoor air
quality. Conventional air-sampling methods use sor-
bent tubes, impingers, vacuum canisters, gravimetric
filters, pumps and light-scattering devices [1]. Many
of these methods require considerable sampling ex-
pertise and costly equipment, lengthy sample collec-
tion and preparation periods, and complicated clean-
ing and extraction procedures. With these methods it
may be difficult, if not impossible, to make on-site de-
cisions required for ‘hot spot’ emissions testing. In ad-
dition, many of commercially available real-time field

0003-2670/99/$ – see front matter ©1999 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 3 - 2 6 7 0 ( 9 9 ) 0 0 6 1 4 - 5

154

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

samplers are calibrated only for one or a few target
analytes.

There is subsequently demand for simple and

cost-effective sampling and analytical methods capa-
ble of achieving very low detection limits in real or
almost real-time. These air-sampling methods should
be applicable to a wide range of air pollutants and
concentrations, be appropriate for both instantaneous
and time-weighted average sampling, and be re-usable
and environmentally friendly.

Solid-phase microextraction (SPME) is an attractive

alternative to conventional air-sampling methods and
its feasibility for air-sampling has been shown previ-
ously [2–11]. Solid-phase microextraction combines
sampling and sample pre-concentration of analytes in
a single step, and allows for direct transfer into a
standard gas chromatograph (GC). The SPME device
consists of a re-usable, coated fiber in a syringe-like
SPME holder. The fiber is exposed to the sample ma-
trix, e.g. air, where analytes partition between coat-
ing and the sample until equilibrium is reached. The
equilibrium time varies from seconds to hours. The
amounts of analytes partitioned to the coating are pro-
portional to their initial concentrations in the matrix.
As such, SPME is particularly suited for air-sampling,
where the effects of partitioning on initial analyte air
concentration are negligible. Analysis of SPME fibers
usually consists of direct transfer of the SPME device
to a GC injector, where the analytes are thermally des-
orbed from the fiber coating and introduced to the GC
column. In most cases, after desorption in a GC injec-
tor lasting from several seconds to minutes, the SPME
fiber is analyte-free, and the device may be immedi-
ately deployed for another air sample.

In the case of sampling with a fiber retracted inside

the SPME needle (TWA sampling), equilibrium time
may be as long as several days, making the SPME
device applicable for long-term and occupational ex-
posure sampling. SPME fibers are also applicable for
non-equilibrium sampling and very short sampling
times. This case will be described later in Section 2 of
this paper. The flexibility of selecting a wide range of
sampling times which is possible with a single SPME
device, e.g. from less than 1 min to days, is currently
not achievable with conventional methods.

In this research, two approaches were applied

to field indoor air-sampling with SPME devices,
i.e. sample collection followed by sample preserva-

tion and subsequent analysis in the lab, and sample
collection followed by immediate on-site analysis
with portable instrumentation. SPME fibers were
used for field air-sampling and analysis of a wide
range of volatile organic compounds (VOCs) and
semi-VOCs. An analyte-specific approach was used
for formaldehyde sampling, where the SPME fiber
was first loaded with derivatizing agent before ac-
tual formaldehyde sample collection. Sampling with
SPME fibers was combined with GC/FID and fast,
portable GC/PID/FID/DELCD. Simultaneously with
the SPME devices, NIOSH methods were used for
air-sampling validation. SPME fibers were also used
for sampling and analysis of diesel exhaust.

2. Experimental

2.1. Chemicals and supplies

Benzene, toluene, ethylbenzene, p-xylene, all

n-alkanes,

chloroform,

and

derivatizing

reagent

(PFBHA)

were

purchased

from

Sigma-Aldrich

(Mississauga, Ont.). Formaldehyde was purchased
from BDH (Toronto, Ont.). Trichloroethene and
1,1,1-trichloroethane were purchased from Caledon
(Georgetown, Ont.). All SPME fibers and holders,
ORBO

TM

-25 large charcoal tubes, ORBO

TM

-32

adsorbent tubes, hydrocarbon traps, gas purifiers,
Teflon

TM

tubing, syringes, Thermogreen

TM

septa,

and vials were purchased from Supelco (Mississauga,
Ont.). The magnetic stirrer, magnetic stir bars, timers,
Tygon

TM

tubing were from VWR (Mississauga, Ont.).

Ultra-high purity hydrogen and nitrogen were from
Praxair (Waterloo, Ont.). Ultra-pure air for standard
gas generators and flame ionization detectors was
supplied from Whatman zero air generator (model
76-803).

2.2. Standard gas generators

National Institute of Standards and Technology

(NIST) traceable certified permeation tubes (Kin-Tech
Laboratories, La Marque, TX) were used for ben-
zene, toluene, ethylbenzene, p-xylene (BTEX) and
formaldehyde generation. A Kin-Tech standard gas
generator (model 491M-B) was used to generate
standard gas concentrations of formaldehyde [7]. For
n-alkanes (C

5

H

12

to C

15

H

32

) and other VOCs, two

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

155

standard gas generators were constructed and tested
using standard NIOSH methods. Ultra-high purity ni-
trogen (for Kin-Tech standard gas generator) and air
(for VOCs and n-alkanes generators) at 50 psi were
supplied using thoroughly cleaned copper tubing and
Swagelok

TM

connectors. Supplied air was addition-

ally scrubbed using Supelpure HC hydrocarbon trap
prior to entering the standard gas generating devices.

Permeation tubes for chloroform, 1,1,1-trichloro-

ethane, and trichloroethene were made by encapsulat-
ing pure analyte inside 100 mm long (1/4 in.) Teflon

TM

tubing capped with 20 mm long solid Teflon

TM

plugs

and (1/4) in. Swagelok caps. Emission rates for each
permeation tube were verified by periodic monitor-
ing of weight loss of individual analyte tubes using

an Ohaus® Analytical Plus scale. All permeation
tubes were placed inside a glass permeation cylinder
(Kin-Tech Laboratories, La Marque, TX) and swept
with constant flow of dilution air. The permeation
cylinder was held inside a permeation oven (100 mm
o.d., 45 mm i.d., 200 mm high) machined from a
solid aluminum rod. Two 100 W ’finger’ heating el-
ements were placed approximately halfway between
the inside and outside diameter of aluminum cylinder.
The temperature of the permeation cylinder was con-
trolled by a K-type Omega

TM

thermocouple (Omega,

Stamford, CT) and an electronic heat control device
designed and constructed by the Electronic Science
Shop (University of Waterloo, Ont.). For n-alkane
standard gas generation, a Razel

TM

Scientific In-

struments syringe pump (model A-99) and 500

␮

l

Gastight® Hamilton syringe was used to deliver the
n-alkane mixture into a heated Swagelok

TM

mixing

tee through a Thermogreen

TM

LB-2 septa.

Air flow rates were controlled by Sidetrack

TM

mass

flow controllers (Sierra Instruments, Monterey, CA)
placed on both the primary and dilution loops for each
standard gas generator. Actual air flow rate was ver-
ified using a primary gas flow standard, mini-Buck
calibrator (A.P. Buck, Orlando, FL). A wide range of
concentrations for formaldehyde, n-alkanes and other
VOCs was obtained by adjusting both air (or nitrogen)
flow rates and permeation cylinder temperature.

2.3. Sampling chambers

New sampling chambers for the three standard gas

generators were constructed and installed downstream

from the standard gas generators. These sampling
chambers facilitated a steady-state mass flow of VOCs,
n-alkanes and formaldehyde at constant temperatures
typical to indoor air. Sampling chambers consisted
of a 1.5 l glass bulb with several sampling ports
plugged with half-hole-type Thermogreen

TM

LB-1

septa. Omega

TM

120 W heating tape was wrapped

around the glass bulb to provide a controlled temper-
ature environment. An Omega

TM

K-type thermocou-

ple was attached to the outside surface of the glass
bulb. Both heating tape and thermocouple were con-
nected to an electronic heat control device designed
and constructed by the Electronics Science Shop at
the University of Waterloo (UW). Air temperatures
in the vicinity of the SPME fibers were maintained
within

±1.2% (±0.3

◦

C at room temperature). Stan-

dard gas flowrates ranged from 50 to 3000 ml min

−1

,

resulting in mean air velocities inside the glass bulb
of less than 10 mm sec

−1

. These air velocities were

within a range of air velocities encountered in in-
door air environments. Standard gas generators and
sampling chambers were validated using both SPME
fibers, and ORBO

TM

adsorbent tubes combined with

A.P. Buck I.H. personal air pumps for conventional
NIOSH methods.

2.4. VOC analysis and speciation on portable
GC/PID/FID/DELCD

A portable SRI-8610C GC (SRI Instruments,

Torrance, CA) was equipped with PID, FID, and
DELCD detectors in series (Fig. 1), and a fast,
dedicated injector [4]. This prototype GC com-
bined

with

polydimethylsiloxane/divinylbenzene

(PDMS/DVB) 65

␮

m fiber for air-sampling allowed

almost real-time sampling and analysis of three sam-
ples per hour. The SPME fiber was exposed for 1 min
(non-equilibrium sampling) in standard BTEX and
halogenated VOCs in the sampling chamber followed
by immediate analysis in the portable GC equipped

with a HP-PONA® (50 m

× 0.2 mm, 0.5

␮

m film)

column (Hewlett-Packard, Mississauga, Ont.). The
column program consisted of the beginning temper-
ature of 30

◦

C, followed by ramping at 15

◦

C min

−1

to 250

◦

C. Desorption time in the GC injector was

30 s. Ultra-high purity hydrogen was used as carrier
gas at 45 psi and 4 ml min

−1

flow rate. The carrier

156

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

Fig. 1. Schematic of PID, FID, and dry electrolytic conductivity detector (DELCD) in series in portable SRI-8610C portable gas
chromatograph.

gas was additionally scrubbed using Supelpure HC
hydrocarbon trap and OMI

TM

-1 carrier gas purifier.

Separation and speciation of VOCs on the portable

GC was achieved in less than 5 min (Fig. 2). As ex-
pected, the PID was much more sensitive for BTEX
compounds than the FID. Similarly, the DELCD was
more sensitive to halogenated VOCs in the stan-
dard gas mixture. Linear calibration curves on all
three detectors for non-equilibrium sampling us-
ing PDMS/DVB-65

␮

m fiber were developed for all

standard gases and concentrations ranging from ap-
proximately 20 ppt to 500 ppb. Exposure time ranged
from a few seconds to few minutes. Method de-
tection limits (MDLs) for BTEX standard gas and
PDMS/DVB-65

␮

m fibers were estimated to be ap-

proximately 1 ppb and 10 ppt, for FID and PID,
respectively. Similarly, MDLs for DELCD were es-
timated to be equal to 100 ppt for trichloroethene.
These MDLs were estimated as a product of stan-
dard deviation of detector response to n = 10 replicate
samples, and the two-tailed t-statistics value for n = 9
degrees of freedom at 95% confidence level.

For the estimation of total volatile organic com-

pounds (TVOCs), PDMS-100

␮

m fiber can be used

for grab air-sampling and analysis on GC/FID. This
approach uses a linear temperature-programmed re-

tention time index (LTPRI) to estimate the fiber-gas
partition coefficient (K

fg

) for each chromatogram

peak [2,5,6,8,11]. The mass of VOCs extracted and
desorbed from a SPME fiber can be estimated using
the FID response factor for general hydrocarbons
for each individual GC/FID. Application of the LT-
PRI approach for estimation of TVOCs in airborne
gasoline has been described elsewhere [6].

In some cases, simultaneous responses to the same,

unknown analyte on GC detectors in-series described
above, may provide a clue as to what class of organic
compound was detected. Similarly, back-calculated
values of K

fg

using the LTPRI approach may be

compared with those published in the literature for
many common toxic air pollutants. This method
may provide an alternative approach for approximate
identification and quantification (at equilibrium) for
individual airborne VOCs using a SPME device with
PDMS-100

␮

m fiber. To date, such speciation and

quantification could only be achieved using GC-MS
or similar analysis.

2.5. Time-weighted average sampling

A SPME device can also be used for long-term

or time-weighted average (TWA) sampling [8,10]. In

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

157

Fig. 2. Comparison of chromatograms for analysis of a standard gas mixture (approximately 1 ppm of each analyte in air) sampled by
PDMS/DVB-65

␮

m SPME fiber for 1 min, and analyzed on SRI-8610C portable GC equipped with PID, FID and DELCD detectors in series.

this case, the SPME fiber remains retracted inside the
needle and serves as a zero sink for analytes. For
TWA sampling, an equilibrium between airborne an-
alytes and analytes partitioned to the SPME should
not be reached. The uptake of airborne analytes is
much slower compared to the exposed fiber (outside
of the needle), and is a function of the analyte-specific
first-order rate constant (K

0

), air concentration, expo-

sure time and path length between opening of the nee-
dle and the fiber. K

0

is a function of the gas-phase

molecular diffusion coefficient, the needle opening
area, and the diffusion path length. This uptake rate
should not be affected, when the maximum amount of
analytes extracted to the fiber is less than 5–10% of
the equilibrium amount [8,10].

A commercial SPME holder was modified for TWA

sampling [12]. Six additional notches were made in the
existing Z-slot. These notches were spaced 5 mm apart
to allow for precise retraction of SPME fibers and to

control of the diffusion path inside the SPME needle
from 0 to 30 mm. An additional notch was placed at
the end of the existing Z-slot to enable fiber exposure
during GC injection. In addition, the plunger-retaining
screw was moved and the tensioning spring on the
SPME fiber assembly was removed to accommodate
full retraction of SPME fiber to 30 mm.

In this study, uptake rate constants were esti-

mated for n-alkanes, and PDMS-100

␮

m fiber, using

a standard gas generator and SPME sampling cham-
ber. In each case, amount of analyte extracted on
the fiber remained less than 10% of the equilibrium
amount. Resulting uptake rates were inversely pro-
portional to GC/FID retention time. Estimation of
uptake rate constants consisted of TWA sampling
of standard n-alkanes in air from several minutes to
days. Following the exposure, fibers were analyzed
on a Varian 3400CX GC/FID equipped with a DB-5
(30 m

× 0.25 mm, 1

␮

m film) column (Supelco, Mis-

158

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

sissauga, Ont.). The column oven program started at
45

◦

C for 1 min, followed by ramping at 15

◦

C min

−1

to 250

◦

C, where the temperature was held for 3 min.

The GC injector and detector temperatures were set
to 250 and 260

◦

C, respectively.

2.6. Field air sampling for VOCs

Grab air-sampling was conducted in a newly

painted residence building at the University of Wa-
terloo (UW). Both SPME 65

␮

m PDMS/DVB fibers

and NIOSH-1501 method (A.P. Buck I.H.pumps and
ORBO

TM

-32 large charcoal tubes) were used in this

survey. The sampling time for PDMS/DVB-65

␮

m

fibers was 1 min. Personal A.P. Buck pumps were
equipped and adsorbent tubes were used to collect
samples during a 2 h period, simultaneously with
grab-sampling by SPME fibers in each room. Both
samplers were positioned in a reasonable breathing
zone in each room. The SPME samples were analyzed
immediately after sample collection in the portable
SRI-8610C GC equipped with PID/FID/DELCD de-
tectors. Charcoal tube samples were analyzed during
the 2 days following sample collection using car-
bon disulfide extraction, direct liquid injection and
analysis with a Varian 3400CX GC/FID.

A SPME PDMS-100

␮

m fiber was also used for

an indoor air survey in a polymer laboratory at UW.
Fibers were deployed approximately in a breathing
zone for an average laboratory worker standing next
to the lab bench. Although for most VOCs SPME
equilibrium conditions can be reached in minutes,
sampling times were as long as 75 min, to allow for
equilibrium sampling in the semi-VOC range. Sample
collection was followed by sealing the SPME nee-
dle with a 10 mm long narrow-bore Teflon

TM

plug

for sample preservation. Samples were subsequently
injected into a Varian 3400CX GC/FID in our labora-
tory. The conventional NIOSH-1550 method was also
used for VOC concentration validation. Charcoal tube
samples were prepared and analyzed during the 2
days following sample collection using carbon disul-
fide extraction, direct liquid injection and analysis
with a Varian 3400CX GC/FID.

A newly remodeled student computer laboratory at

UW was a site for time-weighted average (TWA) sam-
pling using retracted PDMS-100

␮

m fiber and simul-

taneous grab-sampling with exposed fiber. Fibers were
deployed approximately at a breathing zone for an av-
erage student sitting in front of a computer. Sampling
time for both fibers was 75 min, followed by sealing
of the SPME needle with a 10 mm long narrow-bore
Teflon

TM

plug for sample preservation. Samples were

subsequently injected into a Varian 3400CX GC/FID
in our laboratory.

2.7. Field air-sampling for formaldehyde

Sampling for airborne formaldehyde was completed

in the graduate student mailroom at the Department
of Chemistry, UW. This location was selected for sus-
pected formaldehyde presence, due to a large particle
board shelving unit and recent installation of synthetic
carpeting. A PDMS/DVB-65

␮

m fiber was loaded

with

o-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine

hydrochloride (PFBHA) derivatizing agent by 2 min
headspace extraction from PFBHA solution in water
(17 mg l

−1

). Approximately 2 ml of this solution was

prepared in 4 ml vial, capped with Teflon

TM

-coated

septa, and stirred at 1800 rpm using a VWR stirrer.
This method is applicable for relatively low concen-
trations of formaldehyde encountered in indoor envi-
ronments and was previously described in detail [7].
Caution: some batches of PDMS/DVB-65

␮

m fibers

may contain interferences.

Simultaneously with 10 min grab-sampling with

PFBHA-loaded PDMS/DVB-65

␮

m fiber, conven-

tional NIOSH-2541 sampling method was used.
For NIOSH-2541, ORBO

TM

-25 adsorbent tubes and

an A.P. Buck I.H.pump was deployed for 4 h at
50 ml min

−1

. Both samplers were positioned approxi-

mately in the breathing zone of an average adult stand-
ing in front of the shelving. After exposure to ambient
air, SPME fibers were sealed with narrow-bore solid
Teflon

TM

caps and brought to the lab for analysis on

a Varian 3400 GC/FID. Chromatography conditions
were identical to those described previously [7].

2.8. Field air-sampling for particulate matter

The

feasibility

of

detecting

polycyclic

aro-

matic hydrocarbons (PAHs) in diesel exhaust using
PDMS-100

␮

m fiber has been described elsewhere

[3]. In this study, we attempted to use SPME fibers to

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

159

collect and analyze airborne particulate matter. Four
PDMS-7

␮

m fibers were simultaneously exposed to

diesel exhaust from an idling John Deere 344E loader.
All SPME holders were spaced within 20 mm of each
other, attached to a small ruler, and positioned approx-
imately 150 mm downstream from the exhaust pipe
opening. Exhaust temperature near the SPME fibers
was measured using an Omega

TM

K-type thermo-

couple and Omega

TM

microprocessor thermometer

(model HH22). Sampling temperature varied between
35 to 40

◦

C. Exposure times were 0.5, 1, 3, and 5 min,

for a single PDMS-7

␮

m fiber. After sample collec-

tion, each fiber was withdrawn inside the needle and
stored on ice before analysis on a Varian Saturn-IV
GC-MS. In addition, two PDMS-7

␮

m fibers were

simultaneously exposed to diesel exhaust for 1 min
in conditions similar to those described above. One
of these fibers was wrapped inside a hygienic tissue
during exposure to explore the possibility of size
and soluble/solid particle separation. Before and after
sample desorption in the GC injector, each fiber was
visually inspected under a 30

× microscope, to evalu-

ate particulate matter deposition and the effectiveness
of particulate removal from SPME fiber during 3 min
desorption in the GC injector.

Analytes were separated on a SPB-5 (15 m

× 0.25

mm, 0.25

␮

m film) column (Supelco, Mississauga,

Ont). The column was held at 40

◦

C for 3 min, ramped

at 20

◦

C min

−1

to 260

◦

C, followed by holding at

260

◦

C for 10 min. Injector and transfer line tem-

peratures were set to 250 and 270

◦

C, respectively.

Ultra-high purity helium was used as carrier gas at
5 psi and the linear flow velocity of 46.9 cm s

−1

. The

mass range scanned was from 50 to 600 amu.

3. Results and discussion

3.1. VOC grab-sampling

The presence of toluene in the newly painted apart-

ment was detected using PDMS/DVD-65

␮

m fiber

and a photoionization detector on a portable GC.
The toluene concentrations in the bedroom, kitchen,
washroom, and living room detected using both 1 min
extraction by PDME/DVB-65

␮

m fiber and analysis

on SRI-8610C GC, and the NIOSH-1501 method are

Table 1
Toluene concentrations (ppb) measured during indoor air qual-
ity survey using PDMS/DVB-65

␮

m fiber. 1 min sampling time

followed by immediate analysis on a portable GC/PID, and con-
ventional NIOSH-1501 method (approximately 120 min sampling
time, next day analysis by GC/FID)

Location

SPME fiber

NIOSH-1501

Living room

35.8

38.1

Kitchen

29.1

20.1

Bathroom

33.1

18.8

Master bedroom

24.1

22.4

presented in Table 1. These toluene concentrations
ranged from approximately 19–38 ppb. Good corre-
lation between observed concentrations was achieved
at all locations with the exception of the bathroom.
This discrepancy may be explained by the large dif-
ference in sampling times used in both methods, i.e. 1
versus 120 min. It may be possible that during the 2 h
of charcoal tube sampling, toluene concentrations did
not remain constant due to the fact that the bathroom
was the only location within the apartment with a
slightly opened window.

The SPME device combined with a fast and portable

GC system provided almost real-time measurement
of target VOCs. The total sampling and analysis time
was reduced to less than 15 min. In contrast, the to-
tal sampling and analysis time associated with the
NIOSH-1501 method was at least 24 h. It should be
emphasized that the SPME device/portable GC system
offered the convenience of almost immediate on-site
decision making related to VOC sampling. This, in
fact, was the case during one of our indoor air surveys.
The aforementioned system helped us to locate a sig-
nificant source of airborne VOCs inside a residential
house, and to remediate the problem. Such flexibil-
ity combined with very low detection limits, and low
sampling and analysis cost is currently not available
with conventional methods.

Significant amounts of both pentane and hexane

were found in a polymer laboratory at the UW (Table
2). No other analytes were detected. Good correlation
was achieved using both SPME and the NIOSH meth-
ods, for this relatively high analyte concentration. Sim-
ilarly, total sampling and analysis time was reduced
approximately 10-fold compared to the conventional
method.

160

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

Table 2
VOC concentrations (ppb) measured during indoor air quality sur-
vey in a polymer laboratory using (exposed) 100

␮

m PDMS fiber.

Sampling time of 75 min was followed by sealing of the SPME
needle with a narrow-bore solid Teflon

TM

cap, and subsequent

analysis by GC/FID

Analyte

SPME fiber

NIOSH-1550

Pentane

878

743

Hexane

183

172

Table 3
Dodecane air concentration (ppb) measured during indoor air qual-
ity survey in a computer laboratory using both retracted and ex-
posed PDMS-100

␮

m fiber. Sampling time of 75 min was followed

by sealing of the SPME needle with a narrow-bore solid Teflon

TM

cap and subsequent analysis by GC/FID

Analyte

TWA sampling

Exposed

(SPME fiber retracted 3 mm)

SPME fiber

Dodecane

0.70

0.72

3.2. Time-weighted average sampling

The time-weighted average sampling (TWA) was

completed using a retracted SPME fiber. The TWA
concentration for dodecane was approximately 700 ppt
(Table 3). This analyte was identified based on a
column retention time identical to that of dodecane
obtained during validation studies with standard gas
generators. The dodecane concentration estimated us-
ing retracted PDMS-100

␮

m fiber was almost equal

to the dodecane concentration estimated using si-
multaneously exposed (outside of the needle) fiber
(720 ppt). The amount of dodecane extracted during
TWA sampling was approximately 6% of the equilib-
rium amount extracted with exposed PDMS-100

␮

m

fiber. Although only 75 min TWA sampling was com-
pleted, longer sampling could have been used by
increasing the distance between the SPME needle
opening and the SPME fiber. Thus, if that distance
increased 10-fold, i.e. from 3 to 30 mm, the available
TWA sampling time should increase in the same ratio,
from 75 min to approximately 750 min, respectively
[10]. Thus, the use of SPME devices for TWA sam-
pling presents the potential user with great flexibility
of choosing time appropriate for both short-term to
long-term sampling. Such flexibility may be particu-
larly advantageous in occupational exposure assess-

Table 4
Formaldehyde air concentrations (ppb) in graduate student mail-
room measured using 65

␮

m PDMS/DVB fiber. 2 min on-fiber

loading with PFBHA derivatizing agent, followed by immediate
10 min sampling, sealing of the SPME needle with a narrow-bore
solid Teflon

TM

cap and subsequent analysis by GC/FID

a

Analyte

SPME fiber

NIOSH-2541

Formaldehyde

8.33 (RSD = 3.7%)

n/d

b

a

Air samples collected within 2 h period; RSD for n = 3 samples

shown in parenthesis.

b

Below method detection limits.

ments. In comparison, conventional methods require
processes that cannot be re-used, and provide very
limited flexibility for grab-sampling.

3.3. Formaldehyde sampling

As expected, airborne formaldehyde was found in

the mail room with particle board shelving (Table
4). Resulting instantaneous concentrations were re-
producible and equal to 8.33 ppb. This concentration
was not detected using the conventional NIOSH-2541
method, which has a detection limit of at least an or-
der of magnitude higher. SPME PDMS/DVB-65

␮

m

fiber loaded with PFBHA-derivatizing agent proved
to be very sensitive to relatively low formaldehyde
concentrations, undetectable by the conventional
NIOSH-2541 method. This finding is particularly
important considering the recent reassessment of
formaldehyde toxicity, and subsequent dramatic re-
duction of formaldehyde occupational threshold to
16 ppb [13]. The trend of lowering occupational
thresholds will likely continue for other airborne
pollutants, as analytical methods improve and more
knowledge related to compound toxicities becomes
available. Extending that knowledge may in fact be
possible using SPME devices, which allows for fast
and cost-effective sampling and analysis of airborne
contaminants in the parts-per-trillion concentration
range.

The SPME device allowed for a significant reduc-

tion of the total sampling and analysis time to approx-
imately 1 h. This time could be further reduced to ap-
proximately 30 min per sample, if a portable GC sys-
tem with either PID or FID could be used. None of
the existing methods for airborne formaldehyde can

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

161

Fig. 3. Particulate matter (idling diesel engine) loading on
PDMS-7

␮

m fiber.

achieve such a short total analysis time combined with
method detection limits of less than 5 ppb.

3.4. Particulate matter sampling

Increasing deposition of diesel exhaust material

with time on the surface of PDMS-7

␮

m fiber was

observed. Visual inspection of the fibers under a
microscope revealed that most of the deposit was re-
moved after desorption in the GC-MS injector. The
total area counts increased with exposure time (Fig.
3). Fig. 3 and visual observations with a microscope
indicate that it may be possible to achieve linear up-
take of airborne particulate matter on SPME fibers.
The preliminary data shown in Fig. 3 indicated a
non-zero intercept. The reason for this phenomenon is
likely to be associated with the extraction of analytes
dissolved in air. Preliminary data also indicated that
the uptake of several polycyclic aromatic compounds
was also proportional with sampling time. The above
results are in agreement with previous diesel exhaust
sampling using PDMS-100

␮

m [3].

4. Conclusions

The following major conclusions stem from this

study:

1. Air sampling with SPME devices proved to be a

powerful alternative to NIOSH-based field sam-
pling of VOCs and formaldehyde, particularly
where sampling at very low concentrations and
fast, solventless analysis is needed. The time

required to complete the analytical process has
been reduced by at least an order of magnitude
compared to conventional NIOSH techniques.

2. Fast separation and speciation of VOCs common

in indoor air environments was possible using a
portable GC equipped with PID/FID/DELCD de-
tectors in a series.

3. An SPME device can be used as a time-weighted

average sampler for occupational exposure assess-
ments.

4. Air sampling with SPME devices proved to

be more sensitive than conventional sampling
(NIOSH-2541 method) for airborne formalde-
hyde.

5. In all cases, the use of SPME devices allowed

for significant, i.e. at least 10-fold, reduction of
the total sampling and analysis time. The total
sampling and analysis time was less than 15 min in
cases where SPME devices were combined with
the use of fast, portable GC.

6. SPME showed good correlation with NIOSH

methods in cases where VOC concentrations
were above NIOSH method detection limits.

7. Preliminary data indicate that SPME fibers may

be used for sampling and analysis of airborne par-
ticulate matter and aerosols.

Acknowledgements

The authors would like to thank the Center for In-

door Air Research, the Centre for Research in Earth
& Space Technology, Supelco, NSERC, SRI Instru-
ments and Varian for funding this study. Special thanks
should go to Heather Lord for her help with editing
the manuscript.

References

[1] U.S. Environmental Protection Agency, Compendium of

Methods for the Determination of Toxic Organic Compounds
in Air, Research Triangle Park, NC, 1988.

[2] J. Pawliszyn, Solid Phase Microextraction – Theory and

Practice, Wiley-VCH, New York, 1997.

[3] M. Chai, J. Pawliszyn, J. Environ. Sci. Technol. 29 (1995)

693.

[4] T. Gorecki, J. Pawliszyn, Anal. Chem. 34 (1995) 3265.
[5] P. Martos, J. Pawliszyn, Anal. Chem. 69 (1997) 206.

162

J. Koziel et al. / Analytica Chimica Acta 400 (1999) 153–162

[6] P. Martos, A. Saraullo, J. Pawliszyn, Anal. Chem. 69 (1997)

402.

[7] P. Martos, J. Pawliszyn, Anal. Chem. 70 (1998) 2311.
[8] P. Martos, Air sampling with solid phase microextraction

Ph.D. dissertation, University of Waterloo, 1998.

[9] R. Eisert, J. Pawliszyn, G. Barinshteyn, D. Chambers, Anal.

Commun. 35 (1998) 187.

[10] P. Martos, J. Pawliszyn, Anal. Chem. 71 (1999) 1513.

[11] J. Pawliszyn (Ed.), Applications of Solid Phase Microextr-

action The Royal Society of Chemistry, Hertfordshire, UK,
1999.

[12] J. Koziel, A. Khaled, J. Pawliszyn, manuscript in preparation.
[13] National Institute for Occupational Safety and Health,

Documentation for immediately dangerous to life or health
concentrations (IDLHs), National Technical Information
Service, Springfield, VA, 1995.