Journal of Chromatographic Science, Vol.

38, July 2000

Theory of Solid-Phase Microextraction

Janusz Pawliszyn*
Department of Chemistry, University of Waterloo, Waterloo, ON, N2L 3G1, Canada

Abstract
The main objective of this contribution is to describe the
fundamental concepts associated with solid-phase microextraction
(SPME). Theory provides insight when developing SPME methods
and identifies parameters for rigorous control and optimization.
A mathematical model has been developed to understand the
principal processes of SPME by applying basic fundamental
principles of thermodynamics and diffusion theory. The model
assumes idealized conditions and is limited to air, liquid, or
headspace above liquid sampling. Theory for ideal cases can be
quite accurate for trace concentrations in simple matrices such as
air or drinking water at ambient conditions when secondary factors
such as thermal expansion of polymers and changes in diffusion
coefficients because of solutes in polymers can be neglected. When
conditions are more complex, theory for ideal cases still efficiently
estimates general relationships between parameters.

Introduction
Solid-phase microextraction (SPME) was developed to address
the need to facilitate rapid sample preparation both in the laboratory and on-site where the investigated system is located (1). In
the technique, a small amount of extracting phase that is dispersed on a solid support is exposed to the sample for a welldefined period of time. In one approach, a partitioning
equilibrium between the sample matrix and the extraction phase
is reached. In this case, convection conditions do not affect the
amount extracted. In a second approach that uses short-time preequilibrium extraction, if convection or agitation or both are constant, then the amount of analyte extracted is related to time.
Quantitation can then be performed based on timed accumulation of analytes in the coating. Figure 1 illustrates several implementations of SPME that have been considered. These mainly
include open-bed extraction concepts such as coated fibers, vessels, and agitation mechanism disks, but in-tube approaches are
also considered. Some implementations better address issues
associated with agitation, and others better address the ease of
implementing sample introduction to the analytical instrument.

It should be noted that SPME was originally named after the first
experiment that used an SPME device, which involved extraction
on solid fused-silica fibers. Then, it was later renamed to be a reference to the appearance of the extracting phase in relation to a
liquid or gaseous donor phase, even though it is recognized that
the extraction phase is not always technically a solid.
The configurations and operation of SPME devices are very
simple. For example, in the coated fiber implementation of the
technology, one who knows how to use a syringe is able to operate
an SPME device. In the case of automated in-tube extraction for
high-performance liquid chromatography (HPLC), fitting a piece
of the gas chromatography (GC) capillary into the system and
then turning on the autosampler is all that is required to start its
operation. The technology is designed to greatly simplify sample
preparation. However, this feature creates a false impression that
the extraction is a simple, almost trivial process. This misunderstanding frequently results in disappointments. It should be
emphasized that the fundamental processes involved in SPME are
similar to more traditional techniques, and therefore, challenges
to develop successful methods are similar. The nature of target
analytes and the complexity of the sample matrix determine the
level of difficulties in accomplishing a successful extraction. The
simplicity, speed, and convenience of the extraction devices primarily impact the costs of the practical implementation and
automation of the developed methods (2). The objective of this

Particle

Figure 1. SPME configurations.

* E-mail: [email protected]

270

Reproduction (photocopying) of editorial content of this journal is prohibited without publishers permission.

Journal of Chromatographic Science, Vol. 38, July 2000

contribution is to emphasize the fundamental principles of the

technique that define the advantages and limitations of SPME
technology.

Discussion
Principles of SPME

In SPME, a small amount of the extracting phase associated

with a solid support is placed in contact with the sample matrix
for a predetermined amount of time. If the time is long enough, a
concentration equilibrium is established between the sample
matrix and the extraction phase. When equilibrium conditions
are reached, exposing the fiber for a longer amount of time does
not accumulate more analytes. There are 2 different implementations of the SPME technique extensively explored to date. One
implementation is associated with a tube design, and the other is
associated with fiber design. The tube design can use very similar
arrangements (such as SPE); however, the primary difference (in
addition to the volume of the extracting phase) is that the objective of SPME is never that of an exhaustive extraction. This difference substantially simplifies the design of systems. For example,
in the analysis of liquids, in-tube SPME uses 0.25-mm-i.d. tubes
and approximately 0.1 L of the extraction phase; concern for
breakthrough is not relevant because exhaustive extraction is not
an objective. In fact, the objective of the experiment is to produce
full breakthrough as soon as possible, because this indicates that
equilibrium extraction has been reached.
A more traditional approach to SPME involves coated fibers.
The transport of analytes from the matrix into the coating begins
as soon as the coated fiber has been placed in contact with the
sample. There is a substantial difference in performance between
the liquid and solid coatings (Figure 2). A comparison with
adsorptive versus absorptive equilibrium extraction is useful. In
both cases, the extraction process begins by the adsorption of analytes at the extraction phasematrix interface, and then diffusion
of analytes into the bulk of the extraction-phase follows. If the diffusion coefficients of the analytes in the extraction phase are high,
then the analytes partition fully between the 2 phases, and absorp-

Figure 2. Schematic representation of absorptive versus adsorptive extraction

and adsorption in small versus large pores.

tive extraction is accomplished. This process is aided by thin

extraction phase coatings or the convection of the sample matrix
(if flowing liquid). However, if the diffusion coefficient is low, the
analyte remains at the interface and adsorption results. The principle advantage of absorption extraction (partitioning) is a linear
isotherm over a wide range of analyte and interference concentrations, because the property of the extraction phase does not
change substantially until the extracted amount is approximately
1% of the extraction phase weight. However, in adsorption extraction, the isotherm is highly nonlinear for higher concentrations
when the surface coverage is substantial. This causes a particular
problem in the equilibrium methods because the response of the
fiber for the analyte at high sample concentrations depends on
the concentrations of both analytes and interferences. The advantages of the solid sorbents include higher selectivity and capacity
for polar and volatile analytes.
Multiphase equilibria

SPME is a multiphase equilibration process. Frequently, the

extraction system is complex, such as in a sample consisting of an
aqueous phase with suspended solid particles having various
adsorption interactions with analytes, plus a gaseous headspace.
In some cases, specific factors have to be considered, such as analyte losses by biodegradation or adsorption on the walls of the
sampling vessel or stirring mechanism. In this discussion, only 2
phases are considered: (a) the fiber coating and (b) a homogeneous matrix such as pure water or air.
Typically, SPME is considered to be complete when the analyte
concentration has reached distribution equilibrium between the
sample matrix and the fiber coating. In practice, this means that
once equilibrium has been reached, the extracted amount is constant within the limits of experimental error and it is independent
of further increases of extraction time. The equilibrium conditions can be described as:
n=

(KfsVfVsC0)
(KfsVf + Vs)

(1)

where n is the mass of analyte extracted by the coating, Kfs is a

fiber coatingsample matrix distribution constant, Vf is the fibercoating volume, Vs is the sample volume, and C0 is the initial concentration of a given analyte in the sample (3).
Strictly speaking, this discussion is limited to the partitioning
equilibrium that involves liquid polymeric phases such as polydimethylsiloxane (PDMS). The method of analysis for solid sorbent coatings is the same for the low analyte concentration
because the total surface area available for adsorption is proportional to the coating volume if constant porosity of the sorbent is
assumed. For high analyte concentrations, saturation of the surface may occur, resulting in nonlinear isotherms (as will be discussed). Similarly, the high concentration of a competitive
interference compound can displace the target analyte from the
surface of the sorbent.
Equation 1 (which assumes that the sample matrix can be represented as a single homogeneous phase and no headspace is present in the system) can be modified to account for the existence
of other components in the matrix by considering the volumes of

271

Journal of Chromatographic Science, Vol. 38, July 2000

the individual phases and the appropriate distribution constants.

The extraction can be interrupted and the fiber analyzed prior to
equilibrium. However, to obtain reproducible data, constant convection conditions and careful timing of the extraction are necessary.
Simplicity and convenience of operation make SPME a superior alternative to more established techniques in a number of
applications. In some cases, the technique facilitates unique
investigations. Equation 1 indicates that after equilibrium has
been reached, there is a direct proportional relationship between
the sample concentration and the amount of analyte extracted.
This is the basis for analyte quantification. The most visible
advantages of SPME exist at the extremes of sample volumes.
Because the setup is small and convenient, coated fibers can be
used to extract analytes from very small samples. For example,
SPME devices are used to probe for substances emitted by a single
flower bulb during its life spanthe use of submicrometer-diameter fibers permits the investigation of single cells. Because SPME
does not exhaustively extract target analytes, its presence in a
living system should not result in a significant disturbance. In
addition, the technique facilitates speciation in natural systems,
because the presence of a minute fiber that removes small
amounts of analyte is not likely to disturb chemical equilibria in
the system. However, it should be noted that the fraction of analyte extracted increases as the ratio of coating to sample volume
increases. Complete extraction can be achieved for thick coatings
and small sample volumes when distribution constants are reasonably high. This observation can be used as an advantage if
exhaustive extraction is required. It is very difficult to work with
small sample volumes using conventional sample-preparation
techniques. Also, SPME allows for the rapid extraction and
transfer of a sample volume to an analytical instrument. These
features result in an additional advantage when investigating
intermediates in the system. For example, SPME was used to
study the biodegradation pathways of industrial contaminants
(4). The other advantage is that this technique can be used for
studies of the distribution of analytes in a complex multiphase
system (5) and to speciate different forms of analytes in a sample
(6).
In addition, when the sample volume is very large, equation 1
can be simplified to:
n = KfsVf C0

In the case of solid sorbents, the coating has a well-defined crystalline glass structure, which (if dense) substantially reduces the
diffusion coefficients within the structure. Therefore, within the
experimental time, the extraction occurs only on the surface of
the coating. This can be demonstrated by considering the extraction of proteins (illustrated in Figure 3). The original mixture
contains 3 compounds: myoglobin, cytochrome, and lysozyme.
During fiber extraction with polyacrylic acid, compounds with
weaker affinity are only observed at short extraction times. When
the extraction time is longer, the displacement of analytes with
lower affinities occurs. In this case, lysozyme (having a stronger
affinity for the coating) replaces the other 2 compounds during
extraction. This effect is associated with the fact that there is only
limited surface area available for adsorption. If this area is substantially occupied, then the displacement effects occur (7,8) and
the equilibrium amount extracted may vary with concentrations
of both the target and other analytes. However, in the extraction
of analytes with liquid coatings, partitioning between the sample
matrix and extraction phase occurs. In this case, equilibrium
extraction amounts vary only if the coating property is modified
by the extracted components, which only occurs when the
amount extracted is a substantial portion (a few percent) of the
extraction phase. This is very rarely observed because SPME is
typically used to determine trace contamination samples.
The only way to overcome this fundamental limitation of the
porous coatings is to use an extraction time much less than the
equilibrium time so that the total amount of analytes accumulated onto the fiber is substantially below the saturation value
(suggested by Figure 3). When performing such experiments, not
only is it critical to precisely control extraction times, but it is also
important to monitor convection conditions in order to ensure
that they are constant or can be compensated. One way of eliminating the need for compensation of convection is to normalize
or use the same agitation conditions. For example, the stirring
means used at well-defined rotation rates in the laboratory or fans
used for field air monitoring ensure consistent convection.
The short-time-exposure SPME measurement described has an
advantage associated with the fact that the rate of extraction is
defined by the diffusivity of analytes through the boundary layer of
the sample matrix and the corresponding diffusion coefficients

(2)

This equation points to the usefulness of the technique for field

applications. In this equation, the amount of extracted analyte is
independent of the volume of the sample. In practice, there is no
need to collect a defined sample prior to analysis because the fiber
may be exposed directly to ambient air, water, or the production
stream. The amount of the extracted analyte corresponds directly
to its concentration in the matrix without being dependent on the
sample volume. When the sampling step is eliminated, the whole
analytical process can be accelerated, thus errors associated with
analyte losses through decomposition or adsorption on the sampling container walls can be prevented. This advantage of SPME
could be enhanced practically by developing portable field devices
on a commercial scale.

272

Rapid sampling

Figure 3. Demonstration of myoglobin (M) and cytochrome (C) displacement by

lysozyme (L) with time (A) and protein extraction using a polyacrylic acidcoated fiber (B).

Journal of Chromatographic Science, Vol. 38, July 2000

rather than the distribution constants (Figure 4). The mass of

extracted analyte can be estimated from the following equation (9):
n(t) =

2Dg L
ln b +
b

( )

Cg t

(3)

where n is the mass of extracted analyte over sampling time (t), Dg

is the gas-phase molecular diffusion coefficient (cm2/s), b is the
outside radius of the fiber coating (cm), L is the length of the
coated rod (cm), is the thickness of the boundary layer surrounding the fiber coating (cm), and Cg is the analyte concentration in the bulk air (ng/mL). The linear uptake of the analytes with
respect to time for short exposures is demonstrated in Figure 5.
A precise understanding of the definition and thickness of the
boundary layer in this sense is useful. The thickness of the
boundary layer is determined by both the rate of convection (agitation) in the sample and analyte diffusion (10). Thus, in a single
sample, the boundary layer thickness will be different for various
analytes. Strictly speaking, the boundary layer is a region in
which analyte flux is progressively more dependent on analyte diffusion and less on convection when the extraction phase is
approached. However, for convenience, analyte flux in the bulk of

the sample outside of the boundary layer is assumed to be controlled by convection, whereas analyte flux within the boundary
layer is assumed to be controlled by diffusion. The thickness of
the boundary layer is defined as the position in which this transition occurs, or the point at which convection into the boundary
layer is equal to diffusion away from the boundary layer. At this
point, diffusion-controlled analyte flux from the thickness of the
boundary layer towards the extraction phase is equal to the analyte flux from the bulk of the sample towards the thickness of the
boundary layer, which is controlled by convection. The differences in the diffusion coefficients between compounds are small
compared to the differences in the distribution constants. This
makes it easier to calibrate the system. Because of the large differences in the distribution constants between analytes, the
resulting chromatograms are characterized by small peak areas
for compounds with small distribution constants and large peak
areas for those with large constants. With the uptake dependent
on diffusion coefficients, all compounds with similar molecular
masses in a chromatogram will have similar peak areas (given
they have similar detector responses). Also, it is relatively simple
to calculate the diffusion coefficient for a particular analyte and
also to correct for the small differences in it. It must be understood that this system is only suitable for trace analysis. When
sample concentrations become too high, saturation of the active
sites occurs and uptake rates are no longer linear. Shorter exposure times in which smaller amounts are extracted can solve this
problem. Also, at these higher concentrations, samples are easily
extracted and analyzed with a PDMS fiber using conventional
SPME extraction methods. The results of extraction by the diffusion-type approach are shown in Figure 6. The accumulation of
volatile components on the solid coating in 10 s is much larger
compared with the 10-min equilibrium extraction on PDMS. This
approach to extraction is not limited to devices using the fiber
geometry, but is generally applicable.

Figure 4. Boundary layer model.

Figure 5. Correlation of uptake rate with diffusion coefficients for short-sampling-time nonequilibrium extraction of VOCs by PDMSDVB fiber. Diffusion
coefficients: benzene, 0.088; toluene, 0.084; and p -xylene, 0.071.

Figure 6. Comparison of VOC mass loading between PDMS and PDMSDVB.

273

Journal of Chromatographic Science, Vol. 38, July 2000

Extraction modes with coated fiber SPME

SPME sampling can be performed in 3 basic modes: (a) direct

extraction, (b) headspace extraction, and (c) extraction with
membrane protection. Figure 7 illustrates the differences
between these modes. In direct extraction mode (Figure 7A), the
coated fiber is inserted into the sample and the analytes are transported directly from the sample matrix to the extracting phase. To
facilitate rapid extraction, some level of agitation is required to
transport the analytes from the bulk of the sample to the vicinity
of the fiber. For gaseous samples, the natural flow of air (e.g., convection) is frequently sufficient to facilitate rapid equilibration for

Figure 7. Modes of SPME operation: direct extraction (A), headspace SPME (B),
and membrane-protected SPME (C).

Figure 8. Headspace SPME of compounds characterized by a high (A) and low

(B) Henrys constant.

Figure 9. Temperature dependence of the absorption time profiles obtained for

methamphetamine.

274

volatile analytes, but for aqueous matrices, more efficient agitation techniques such as fast sample flow, rapid fiber or vial movement, stirring, or sonication are required to reduce the effect of
the depletion zone produced close to the fiber as a result of slow
diffusional analyte transport through the stationary layer of liquid
surrounding the fiber.
In the headspace mode (Figure 7B), the analytes are extracted
from the gas phase equilibrated with the sample. The primary
reason for this modification is to protect the fiber from adverse
effects caused by nonvolatile, high-molecular-weight substances
present in the sample matrix (e.g., humic acids or proteins). The
headspace mode also allows matrix modifications (including pH
adjustment) without affecting the fiber. In a system consisting of
a liquid sample and its headspace, the amount of an analyte
extracted by the fiber coating does not depend on the location of
the fiber (in the liquid or gas phase); therefore, the sensitivity of
headspace sampling is the same as the sensitivity of direct sampling as long as the volumes of the 2 phases are the same in both
sampling modes. Even when no headspace is used in direct
extraction, a significant sensitivity difference between direct and
headspace sampling may occur only for very volatile analytes.
However, the choice of sampling mode has a very significant
impact on the extraction kinetics. When the fiber is in the
headspace, the analytes are removed from the headspace first, followed by indirect extraction from the matrix. If the Henrys
Constant of a given compound is high, then the concentration of
analytes in the headspace is high, resulting in very rapid extraction because the extracted analytes originate primarily from the
gaseous headspace (Figure 8A). However, if the Henrys Constants
are low, then the extraction is long because the analytes need to
diffuse from the condensed phase before they reach the fiber
(Figure 8B). Therefore, in the case of the extraction of aqueous
samples, volatile and nonpolar analytes are extracted much faster
than semivolatiles or polar volatiles. Temperature has a significant effect on the kinetics of the process, because it determines
the vapor pressure of analytes above the condensed phase. In general, the equilibration times for volatile compounds are shorter
for headspace SPME than for direct extraction under similar agitation conditions because of the following 3 reasons: (a) a substantial portion of the analytes is present in the headspace prior to
the beginning of the extraction process, (b) there is typically a
large interface between the sample matrix and headspace, and (c)
the diffusion coefficients in the gas phase are typically higher by 4
orders of magnitude than in liquids. The concentration of
semivolatile compounds in the gaseous phase at room temperature is small, and headspace extraction rates for those compounds
are substantially lower. These compounds can be improved by
using very efficient agitation or by increasing the extraction temperature. Figure 9 illustrates the equilibration time profiles
obtained for the extraction of methamphetamine from a urine
sample at various temperatures. At 22C and 40C, the equilibration is very longexceeding 100 min as indicated in this graph.
It drops to approximately 20 min when the extraction temperature is 60C and to only a few minutes when the temperature is
73C. The dramatic change with the equilibration time is associated with the fact that an increase in temperate results in an
increase of the analytes Henrys Constant, an increase in the diffusion coefficient, and a decrease of the amount extracted at equi-

Journal of Chromatographic Science, Vol. 38, July 2000

librium. This decrease is associated with the fact that the distribution constant decreases when the temperature increases.
Therefore, it is important to carefully optimize the extraction
temperature for the shortest equilibration times and for acceptable sensitivities. In most SPME applications, equilibrium extraction is performed. However, often when the equilibration times
are long, pre-equilibrium quantification can be considered. It is
important in such experiments to ensure constant agitation conditions and acceptable extraction times in order to obtain good
precision.
In the third mode (SPME with membrane protection, Figure
7C), the fiber is separated from the sample with a selective membrane, which lets the analytes through while blocking the interferences. The main purpose for the use of the membrane barrier
is to protect the fiber against adverse effects caused by highmolecular-weight compounds when very dirty samples are analyzed. Although, extraction from headspace serves the same
purpose, membrane protection enables the analysis of lessvolatile compounds. The extraction process is substantially
slower than direct extraction because the analytes need to diffuse
through the membrane before they can reach the coating. The
use of thin membranes and an increase in extraction temperature
result in shorter extraction times.
Extraction modes with in-tube SPME

There are 2 fundamental approaches to in-tube SPME: (a)

active or dynamic (when the analytes are passed through the
tube) and (b) passive or static (when the analytes are transferred
into the sorbent by diffusion). In either of these approaches, the
coating may be supported on a fused-silica rod or coated on the
inside of a tube or capillary. The theoretical aspects of the extraction processes that use these geometric arrangements will be
briefly discussed.
Dynamic in-tube SPME
In this system, we assume (a) a piece of fused-silica capillary
internally coated with a thin film of extracting phase (a piece of
open tubular capillary GC column) is used or (b) the capillary is
packed with an extracting phase dispersed on an inert supporting
material (a piece of micro-liquid-chromatography capillary
column). During the introduction of the sample, the front of the
analyte migrates through the capillary with a speed proportional
to the linear velocity of the sample and inversely related to the
partition ratio (11,12). For short capillaries with a small dispersion, the extraction time can be assumed to be similar to the time
required for the center of the band to reach the end of the capillary. The extraction time is proportional to the length of the capillary and inversely proportional to the linear flow rate of the fluid.
Extraction time also increases with an increase in the
coatingsample distribution constant and an increase in the
thickness of the extracting phase, but decreases with an increase
in the void volume of the capillary. An increase in the
coatingsample distribution constant produces an increase in the
absolute amount extracted. It has been observed that increases in
the amounts extracted can be achieved in many cases by preconditioning the capillary with methanol or some other appropriate
solvent prior to extraction. Enhancement has even been observed
when a plug of methanol was aspirated into the capillary before

the sample was drawn in and the sample followed the plug while
in the capillary during the extraction aspiratedispense steps.
This is similar to the solvent preconditioning used in SPE to
enhance extraction.
In practice, in-tube SPME is implemented by replacing a section of the tubing in a commercially available autosampler and
then programming the autosampler to pass the sample in and out
of the extraction capillary until equilibrium or a suitable extraction level has been reached.
It should be emphasized that this is valid only for direct extraction when the sample matrix passes through the capillary. This
approach is limited to particulate-free gas and clean water samples. The headspace SPME approach can broaden the application
of in-tube SPME. In such cases, careful consideration of the mass
transfer between the sample and headspace should be given in
order to describe the process properly. Also, if the flow rate is very
rapidly producing turbulent behavior and the coatingsample
distribution constant is not very high, then perfect agitation conditions are met and equation 4 can be used to estimate equilibration times. In this case, equilibration time (te) is assumed to be
achieved when 95% of the equilibrium amount of the analyte is
extracted from the sample:
te = t95% =

(b a)2
2Df

(4)

In this equation, b a refers to the thickness of the sorbent

material and Df refers to analyte diffusion in the sorbent.
The removal of analytes from a tube is an elution problem similar to frontal chromatography and has been discussed in detail
(12). In general, if the desorption temperature of a GC is high and
thin coatings are used, then all the analytes are in the gas phase
as soon as the coating is placed in the injector and the desorbtion
time corresponds to the elution of 2 void volumes of the capillary.
For liquid desorption, the desorbtion volume can be even smaller
because the analytes can be focused at the front of the desorption
solvent (13).
Static in-tube SPME time-weighted average sampling
In addition to the analyte concentration measurement found at
a well-defined place in space and time (obtained by using the
approaches discussed previously), an integrated sampling is possible with a simple SPME system. This is particularly important in
field measurements when changes of analyte concentration over
time and place variations must often be taken into account.
When the extracting phase is not exposed directly to the sample
but is contained in a protective tubing (needle) without any flow
of the sample through it (Figure 10A), the extraction occurs
through the static gas phase present in the needle. The integrated
system can consist of the extraction phase coating the interior of
the tubing, or it can be an externally coated fiber withdrawn into
the needle. These geometric arrangements represent a very powerful method able to generate a response proportional to the integral of the analyte concentration over time and space when the
needle is moved through the space (14). In these cases, the only
mechanism for analyte transport to the extracting phase is diffusion through the gaseous phase contained in the tubing. During
this process, a linear concentration profile (shown in Figure 10)

275

Journal of Chromatographic Science, Vol. 38, July 2000

is established in the tubing between the small needle opening

characterized by the surface area (A) and the distance (Z) between
the needle openingand the position of the extracting phase.
The amount of analyte extracted (dn) during the time interval (dt)
can be calculated by considering Ficks first law of diffusion (15):
dn = ADg

dc
C(t)
dt = ADg
dt
Z
dz

(5)

where C(t)/Z is a value of the gradient established in the needle

between the needle opening and the position of the extracting
phase (Z). The function C(t) is equivalent to C(t) Cz, where C(t)
is a time-dependent concentration of the analyte in the sample
near the needle opening and Cz is the concentration of the analyte in the gas phase near the coating. If Cz is close to zero for a
high coatinggas distribution constant capacity, then C(t) is
equivalent to C(t). The concentration of the analyte at the coating
position in the needle (Cz) will increase with the integration time
but will be kept low compared to the sample concentration
because of the presence of the sorbent coating. Therefore, the
accumulated amount over time can be calculated as:

n = Dg A C(t)d
Z

(6)

As expected, the extracted amount of analyte is proportional to

the integral of the sample concentration over time and the diffusion coefficient of analytes in gaseous phase (Dg) in the area of the
needle opening (A), and it is inversely proportional to the distance
of the coating position with respect to the needle opening (Z). It
should be emphasized that equation 5 is valid only in a situation
in which the amount of analyte extracted onto the sorbent is a
small fraction (below RSD of the measurement, typically 5%) of
the equilibrium amount with respect to the lowest concentration
in the sample. To extend integration times, the coating can be
placed further into the needle (resulting in a larger Z), the
opening of the needle can be reduced by the introduction of an

Figure 10. Use of SPME for in-tube TWA sampling: schematic (A) and adaptation of a commercial SPME manual-extraction holder (B).

276

additional orifice (a smaller A), or a higher capacity sorbent can be

used. The first 2 solutions will result in a low measurement sensitivity. An increase of sorbent capacity presents a more attractive
opportunity. It can be achieved by either increasing the volume of
the coating or its affinity towards the analyte. An increase of the
coating volume will require an increase of the device size. The
optimum approach to an increased integration time is to use sorbents characterized by large coatinggas distribution constants.
The exploitation of restricting access to the absorbing medium
allows for the implementation of SPME for time-weighted
average (TWA) sampling. Whenever diffusion to the sorbent surface is limited, the sorbent can act as a sort of zero sink in which
extraction is very far from equilibrium under normal sampling
conditions. Therefore, in practice, any analytes that reach the sorbent surface are essentially exhaustively absorbed. However, the
rate of diffusion is still dependent on the sample concentration;
therefore, the total amount absorbed by the coating is proportional to the average of analyte concentrations over time, thus
TWA sampling is achieved. This has been implemented to date
with the conventional fiber assembly by retracting the fiber to a
known distance inside the needle (Figure 10B). The small size of
the needle orifice limits diffusion to the sorbent surface, and the
ultimate diffusion rate is a function of the distance from the fiber
tip to the end of the needle. Depending on the volatility and concentration of the analyte of interest, the fiber may be positioned
either closer to the end of the needle or further away from it in
order to achieve the desired degree of nonequilibrium extraction
and sensitivity. It would also be possible to implement this type of
sampling with the sorbent coated onto the interior wall of a capillary. However, to date, the retractable needle implementation
has gained the most attention because of its ease of use and
adjustability for the analyte and the given sample.
Prediction of distribution constants

In many cases, the distribution constants presented in equations 1 and 2 (that determine the sensitivity of SPME) can be estimated from physicochemical data and chromatographic
parameters. This approach eliminates the need for calibration.
For example, distribution constants between a fiber-coating and
gaseous matrix (e.g., air) can be estimated using isothermal GC
retention times on a column with a stationary phase identical to
the fiber-coating material (16). This is possible because the partitioning process in GC is similar to the partitioning process in
SPME and there is a well-defined relationship between the distribution constant and the retention time. The nature of the gaseous
phase does not affect the distribution constant unless the components of the gas (such as moisture) swell the polymer, thus
changing its properties. A most useful method for determining
coating-to-gas distribution constants uses the linear temperature-programmed retention index (LTPRI) system, which indexes
a compounds retention time in relation to the retention times of
n-alkanes. This system is applicable for the retention times of
temperature-programmed gasliquid chromatography. The logarithm of the coating-to-air distribution constants of n-alkanes
can be expressed as a linear function of their LTPRI values. For
PDMS, this relationship is logKfg= 0.00415 LTPRI 0.188 (17).
Thus, the LTPRI system permits interpolation of the Kfg values

Journal of Chromatographic Science, Vol. 38, July 2000

from the plot of logKfg versus the retention index. The LTPRI
values for many compounds are available in the literature; therefore, this method allows estimation of Kfg values without experimentation. If the LTPRI value for a compound is not available
from published sources, it can be determined from a GC run. It
should be noted that the GC column used to determine the LTPRI
value should be coated with the same material as the fiber
coating.
Estimation of the coatingwater distribution constant can be
performed using equation 3. The appropriate coatinggas distribution constant can be found by applying the techniques discussed previously, and the gaswater distribution constant
(Henrys constant) can be obtained from physicochemical tables
or estimated by the structural unit contribution method (18).
Some correlations can be used to anticipate trends in SPME
coatingwater distribution constants for analytes. For example, a
number of investigators have reported the correlation between
the octanolwater distribution constants, Kow and Kfw. This is
expected because Kow is a very general measure of the affinity of
the compounds to the organic phase. However, it should be
remembered that the trends are valid only for compounds within
a homologous series, such as aliphatic hydrocarbons, aromatic
hydrocarbons, or phenols. They should not be used to make comparisons between different classes of compounds because of different analyte activity coefficients in the polymer.
Effect of extraction parameters

Thermodynamic theory predicts the effects of modifying certain extraction conditions on partitioning and indicates the
parameters to control for reproducibility. This theory can be used
to optimize the extraction conditions with a minimum number of
experiments and correct for variations in the extraction conditions without the need to repeat calibration tests under the new
conditions. For example, SPME analysis of outdoor air may be
done at ambient temperatures that can vary significantly. The
relationship that predicts the effect of temperature on the
amount of analyte extracted allows for calibration without the
need for extensive experimentation (19). Extraction conditions
that affect Kfsinclude temperature, salting, pH, and the amount
of organic solvent content in water.
An extraction temperature increase causes an increase in the
extraction rate and simultaneously a decrease in the distribution
constant. In general, if the extraction rate is of major concern, the
highest temperature that still provides satisfactory sensitivity
should be used.
Adjustment of the pH of the sample can improve the sensitivity
of the method for basic and acidic analytes. This is related to the
fact that unless ion-exchange coatings are used, SPME can only
extract neutral nonionic species from water. By properly
adjusting the pH, weak acids and bases can be converted to their
neutral forms, in which case they can be extracted by the SPME
fiber. To make sure that at least 99% of the acidic compound is in
the neutral form, the pH should be at least 2 units lower than the
pKa of the analyte. For the basic analytes, the pH must be larger
than pKa by 2 units.
The volume of the sample should be selected based on the estimated distribution constant Kfs (8). The distribution constant
can be estimated by using literature values for the target analyte

or a related compound with the coating selected. The distribution

constant can also be calculated or determined experimentally by
equilibrating the sample with the fiber and determining the
amount of analyte extracted by the coating. Care must be taken to
avoid analyte losses from adsorption, evaporation, or microbial
degradation when very long extraction times are required to
reach the equilibrium.
The sensitivity of the SPME method is proportional to the
number of moles of the analyte (n) extracted from the sample and
(for direct extraction) is given by equation 1. When the sample
volume (Vs) increases, the amount of analyte extracted also
increases until the volume of the sample becomes significantly
larger than the product of the distribution constant and volume
of the coating (fiber capacity Kfs<<Vs).

Conclusion
A number of parallels can be drawn between the developments
and applications of SPME with electrochemical methods. The
coulometric technique corresponds to the total extraction
method. Although this technique is the most precise, it is not frequently used because of the time required to complete it. SPME
is capable of producing exhaustive extraction when the volume of
the extraction phase is large enough to combine with high-distribution constants. In fiber geometry, the larger volume translates
into thicker coatings, which results in long extraction times. The
alternative approach is to disperse the whole volume of the
extraction phase onto a larger surface area, resulting in a thinner
coating and faster equilibration times. For example, solid support
material may include particulate matter, a stirring mechanism, or
a vessels walls (Figure 1). However, in this case, there would be
more handling required in order to conveniently introduce the
extraction phase into the sample introduction system (GC or
HPLC). It might necessitate the use of an organic solvent to
desorb the analytes from the extraction phase. Equilibrium
potent-ion-metric techniques are more frequently used (pH electrode), particularly when the sample is a simple mixture or selectivity of the membrane is sufficient to quantify the target analyte
in complex matrices. The equilibrium SPME method has some
advantages in this regard because the technique is typically coupled with separation or MS detection methods or both, which
simultaneously allows for the identification and quantitation of
many components. The advantage of using electrochemical
methods is the response time resulting from the low capacities of
electrodes.
Several electrochemical methods, (e.g., amparometry) are
based on mass transport through the boundary layers, such as
pre-equilibrium SPME. Similarly, in SPME, calibration based on
diffusion coefficients can be accomplished when the agitation
conditions are constant, the extraction times are short, and the
coating has a high affinity towards the analytes. Figure 21 illustrates the results related to the 10-s extraction times using solid
coating. In some implementations of the technology, the rate of
the mass transfer to the extraction phase can be purposely
restricted by placing it in the needle, thus achieving the TWA
measurements of concentration in a specific time period.

277

Journal of Chromatographic Science, Vol. 38, July 2000

The potential savings in analysis time, reduced solvent use, and

the apparent simplicity of SPME techniques will continue to
attract interest among analytical chemists searching for
improved analysis methods. As long as analysts have a sound
understanding of the theory and principles behind this technique,
good accuracy and precision will follow.

8.
9.
10.
11.
12.

References
1. J. Pawliszyn. Solid Phase Microextraction. Theory and Practice.
Wiley, New York, NY, 1997.
2. Applications of Solid Phase Microextraction. J. Pawliszyn, Ed., RSC,
Cambridge, U.K., 1999.
3. D. Louch, S. Motlagh, and J. Pawliszyn. Anal. Chem. 64: 1187
(1992).
4. Applications of Solid Phase Microextraction. J. Pawliszyn, Ed., RSC,
Cambridge, U.K., 1999, pp 60921.
5. J. Poerschmann, F.-D. Kopinke, and J. Pawliszyn. Solid phase
microextraction to study the sorption of organation compounds
onto particulate and dissolved humic organic matter. Environ. Sci.
Technol. 31: 362936 (1997).
6. Z. Mester and J. Pawliszyn. Rapid Commun. Mass Spectrom. 13:
19992003 (1999).
7. J.-L. Liao, C.-M. Zeng, S. Hjerten, and J. Pawliszyn. Solid phase
micro extraction of biopolymers, exemplified with adsorption

278

13.

14.
15.
16.
17.

18.

19.

of basic proteins onto a fiber coated with polyacrylic acid.

J. Microcolumn Sep. 8: 14 (1996).
T. Gorecki, X. Yu, and J. Pawliszyn. Analyst 124: 64349 (1999).
J. Koziel, M. Jia, and J. Pawliszyn, submitted for publication in Anal.
Chem.
C.V. King. J. Am. Chem. Soc. 57: 82831 (1935).
J. Crank. Mathematics of Diffusion. Clarendon Press, Oxford, 1989,
p 14.
J. Pawliszyn. Kinetic model of Supercritical fluid extraction.
J. Chromatogr. Sci. 31: 3137 (1993).
R. Eisert and J. Pawliszyn. Automated in-tube solid-phase microextraction coupled to high-performance liquid chromatography.
Anal. Chem. 69: 314047 (1997).
M. Chai and J. Pawliszyn. Environ. Sci. Technol. 29: 693 (1995).
J. Crank. Mathematics of Diffusion. Clarendon Press, Oxford, 1989,
p 2.
Z. Zhang and J. Pawliszyn. J. Phys. Chem. 100: 17648 (1996).
P. Marlos, A. Saraullo, and J. Pawliszyn. Water analysis by solid
phase microextraction coupled to high-performance liquid chromatography. Anal. Chem. 69: 199296 (1997).
R. Schwarzenbach, P. Gschwend, and D. Imboden. Environmental
Organic Chemistry. John Wiley & Sons Inc., New York, NY, 1993, pp
10923.
P. Martos and J. Pawliszyn. Calibration of solid phase microextraction for an analyses based on physical chemical properties of the
coating. Anal. Chem. 69: 206215 (1997).
Manuscript accepted May 17, 2000.