Solid phase microextraction as a tool for trace element spec

Spectrochimica Acta Part B 56 2001 233

᎐260

Review

Solid phase microextraction as a tool for trace

element speciation

Zoltan Mester

, Ralph Sturgeon

, Janusz Pawliszyn

Institute for National Measurement Standards, National Research Council Canada, Ottawa, Ontario K1A 0R6, Canada

Uni

ersity of Waterloo, Department of Chemistry, Waterloo, Ontario N2L 3G1, Canada

Received 28 June 2000; accepted 30 November 2000

Abstract

Applications of solid phase microextraction SPME for trace element speciation are reviewed. Because of the

relative novelty of the technique in the inorganic analytical field, the first part of this review provides a short
overview of the principles of SPME operation; the second part describes typical SPME applications to elemental
speciation. Volatile organometallic compounds can be collected by SPME from the sample headspace or liquid
phase, directly or after derivatization. The usual separation method for the collected volatile species is gas
chromatography. Non-volatile analyte species can be collected from the sample liquid phase and separated by liquid
chromatography or capillary electrophoresis. Currently, most SPME applications in the inorganic field comprise
analyte ethylation and headspace extraction followed by gas chromatographic separation of tin, lead and mercury
species. The use of SPME for the study of equilibria in complex systems is also discussed and future roles of solid
phase microextraction in the inorganic analytical field are raised. Crown copyright

䊚 2001 Published by Elsevier

Keywords:

Solid phase microextraction; Inorganic speciation; Headspace sampling; Derivatization; Tin; Lead;

Mercury; Arsenic; Selenium

1. Introduction

An analytical procedure comprises several

steps: sampling, sample preparation, qualifica-

Corresponding author. Fax:

q1-613-9932451.

E-mail address: zoltan.mester@nrc.ca Z. Mester .

tion

᎐quantification, data evaluation, and decision

making. Each step is of equal importance for
ensuring correct results and, consequently, arriv-
ing at the right decision. Each step embodies a
reliability range which propagates through the
analytical procedure and contributes to the con-
fidence interval of the final results. Today’s ana-
lytical chemical research efforts are mainly fo-

0584-8547

PII: S 0 5 8 4 - 8 5 4 7 0 0 0 0 3 0 4 - 9

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

234

cused on the measurement qualitative or quanti-

tative step of the analytical procedure. Although
we are enveloped by the fascinating world of
high-tech instrumentation, many analytical prob-
lems waiting to be resolved by instrument devel-
opment efforts are actually chemical or physico-
chemical problems which could be, and must be,
resolved by other means. For example, stability of
a target compound during sampling and sample
preparation can be examined by isotopically label-
ing the analyte and following it over the course of
the analytical procedure using sophisticated mass
spectrometric instrumentation, or focusing on
the chemistry of the problem by means of sam-
pling

rsample preparation. In organic analytical

chemistry, sample preparation mainly comprises
extractions which serve to isolate components of
interest from a sample matrix, because most ana-
lytical instrumentation cannot handle the matrix.
Sample preparation in inorganic analytical chem-
istry is not generally concerned with extraction in
the classical context, but rather with digestions
which principally serve to liberate target com-
pounds from the sample matrix and convert the
various chemical forms of the analyte to a uni-

w x

form species 1 . Because of the high selectivity of
atomic spectroscopic detection, a separation step
is usually not required. This classical difference
between inorganic

relemental and organic analyt-

ical methodologies has largely disappeared over
the last decade because of the increasing interest
in speciation information in elemental analysis.
This has resulted in a strong cross-talk and cross-
fertilization of ideas between the two disparate
fields. Speciation of trace elements provides more
specific information about the real status and
impact of a given element in an environmental or
biological system. The term ‘chemical species’ was

originally used to refer to a specific form mono-

atomic or molecular or configuration in which an
element can occur, or to a distinct group of atoms
consistently present in different compounds or

w x

matrices 2 . More recently, the term speciation
has received broader acceptance as ‘the occur-
rence of an element in separate identifiable forms
Ž

i.e. chemical, physical or morphological state ’

w x

3 . To understand the impact

ravailability of tar-

get elements in any system requires different

approaches, which comprise the extended direc-
tions of speciation analysis:

Determination of the exact chemical form of
a given element includes not only its oxida-
tion state, but also the structure of the
molecule if the metal is covalently bonded. It
is well known that the different chemical
forms of an element may exhibit significant
differences in toxicity, mobility or reaction
kinetics.

Determination of the distribution of metallic
compounds in a real or pseudo-multiphase
system. For example, the toxic trace metals in
aqueous environmental systems containing
natural complexing agents, such humic and
fulvic acids, are partially associated with these
complexing agents, with the consequence that
the free metal content, which determines the
biological effect, can be seriously altered.
There is little doubt that studies devoted to
this will become one of the fastest growing
areas of speciation analysis because this type
of equilibrium

rdistribution problem arises

not only in the fields of the environment, but

w x

also the realm of bio-inorganic chemistry 4 .

Determination of the bioavailability of metals

arising from solid materials

such as sedi-

ments, soils, and fly and bottom ash through
solubilization with various ‘soft’ aqueous
leaching agents in an effort to model metal
release in different environmental situations.
For example, extracting fly ash samples with
weakly acidic solutions may simulate the ef-
fect of acidic rain; extraction of soils with
different complexing agents, such as EDTA,
may simulate the effect of the presence of
naturally occurring complexing substances in
soil systems.

Ž .

Approaches to speciation described in points i

and iii are well-studied and understood and, in
many cases, already routinely applied. Certified

reference materials are available

even if in

limited number

and virtually standard proce-

dures for analysis are already established, which
reflect the long history of work supporting them.
The situation in the field of equilibrium

rmulti-

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

235

phase speciation is less advanced. No standard,
recommended methods or reference materials are
available to support this area of activity.

The need for speciation particularly with re-

Ž .

spect to point i

has had a significant impact in

two areas of classical elemental analysis. The first
is sample preparation, wherein preservation of
the sample’s integrity requires use of ‘soft’ extrac-
tion techniques and, when GC analysis is to be
carried out, transfer of the organometallic com-
pounds from the aqueous to a non-polar organic
phase. The second aspect is the necessity to apply
separation techniques, such as liquid chromatog-

w x

raphy LC

5 , gas chromatography GC

6 and

w x

capillary electrophoresis

7 prior to ele-

ment-selective detection. As a consequence of the
need to interface chromatographic methods with
atomic

spectroscopic

detectors,

many

new

concerns arise relating to interface designs, prob-
lems relevant to lengthy data collection periods,
techniques for the evaluation of chromato-
graphic

rtransient data, etc. This review focuses

primarily

sampling

and

extraction

organometallic species using techniques based on
solid phase microextraction.

2. Extraction

The ideal analytical instrument would, without

human intervention, perform all analytical steps
from sampling to data analysis, and even under-
take decision making based on the results ob-
tained. Although today’s sophisticated instru-
ments can analyze complex samples and evaluate
results, most sampling and sample preparation
practices are based on 19

century technologies,

such as the common Soxhlet extraction method

w x

8 . Traditional sample preparation methods are

typically time consuming, employ multi-step
procedures having high risk for loss of analytes
and use extensive amounts of organic solvents.
These characteristics make such methods very
difficult to automate and integrate into modern
sampling

rseparation systems. As a consequence,

most of the analysis time is consumed by sam-
pling and sample preparation.

Extensive use of organic solvents in analytical

laboratories is no longer tolerated because of the
associated health risks and disposal concerns. As
a result, many solvent-free extraction methods
have been described or rediscovered in the last

Fig. 1. Classification of solvent-free sample preparation methods supercritical fluid extraction SFE , solid phase extraction SPE ,

and solid phase microextraction SPME .

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

236

decade. These can be classified according to the
nature of the extraction phase, as shown in Fig. 1,

w x

i.e. gas, membrane or sorbent 9 . Other low sol-
vent-consumption methods, such as single droplet

w x

extraction, are becoming more popular 10 . One
solvent-free extraction approach is solid phase

microextraction SPME which, as is evident from
Fig. 1, is a sorbent extraction technique similar to

solid phase extraction SPE . With SPME, the
sorbent material is attached to the surface of a

fiber rather than packaged into a cartridge tube
or used on the surface of a flat disk. Whereas the
SPE technique has been primarily designed for
use with liquid matrices and exhaustive extrac-

tion, SPME can be used in liquid aqueous or
gaseous matrices and primarily aims for partial or
equilibrium extraction of the analyte.

The principal approach of SPME is the use of a

small volume of extracting phase, usually less
than 1

␮l. The extracting phase can be a high

molecular-weight polymeric ‘liquid’ or a solid sor-
bent, typically a high surface-area porous mate-
rial. Fig. 2 illustrates the structure of a commer-
cially available SPME unit. A small-diameter
fused-silica fiber, coated with the extraction phase,

is mounted in a syringe-like device for protection
and ease of handling. The needle serves to conve-
niently pierce septa during sample extraction and
desorption operations. Using the syringe-like
mechanism of the holder unit, the fiber can be
extruded from the needle to expose the extraction

phase to the sample headspace or liquid . After
the sampling period, the same mechanism can be
used to withdraw the fiber inside the needle.
During the extraction and desorption periods, the
fiber is thus exposed by sitting outside of the
needle; during transfer of the SPME unit to a
desorption apparatus, the polymeric end of the
fiber is inside the protective needle.

3. General description of solid phase
microextraction

The transport of analytes from the matrix into

the extraction medium begins as soon as the
coated fiber has been placed in contact with the
sample. In most cases, SPME extraction is con-
sidered to terminate when the analyte concentra-
tion has reached distribution equilibrium between

w x Ž .

Ž .

Fig. 2. Commercial SPME device 58 . a SPME fiber holder; and b SPME holder and fiber assembly

ᎏ section view reprinted

with permission of Supelco Bellefonte, PA 16823, USA .

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

237

the sample matrix and the fiber coating. In prac-
tice, this means that once equilibrium is reached,
the amount extracted is constant, within the limits
of experimental error, and is independent of fur-
ther increases in extraction time.

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 investiga-
tions. The most dramatic advantages of SPME
exist at the extremes of sample volumes. Because
the set-up 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 span; the use of sub-

␮m-diam-

eter fibers permits the investigation of single cells.
Since SPME is an equilibrium technique, and
therefore does not extract target analytes exhaus-
tively, its presence in a system should not result in
significant disturbance. In addition, the technique
facilitates speciation in natural systems, since the
presence of a minute fiber, which removes small
amounts of analyte, is not likely to disturb chemi-
cal equilibria. It should be noted, however, that
the fraction of analyte extracted increases as the
ratio of coating volume to sample volume in-
creases. Complete extraction can be achieved for
small sample volumes when distribution constants
are large. This observation can be used to advan-
tage if exhaustive extraction is required. SPME
also allows rapid sample extraction and transfer
of analyte to the analytical instrument. These
features result in an additional advantage when
investigating intermediates in the system.

3.1. Extraction modes

Two basic types of extractions can be per-

formed using SPME: direct and headspace extrac-
tion. Fig. 3 illustrates the differences between

these modes. In the direct extraction mode Fig.

3b , the coated fiber is inserted into the sample
medium and the analytes are transported directly
to the extraction phase. To facilitate rapid extrac-
tion, some level of agitation is required to en-
hance transport of the analytes from the bulk of
the solution to the vicinity of the fiber. For gaseous

Ž .

Fig. 3. SPME operation modes: a headspace; and b direct

liquid phase immersion sampling.

samples, natural convection and diffusion in the
medium is sufficient to facilitate rapid equilibra-
tion. For aqueous matrices, more efficient agita-
tion techniques, such as fast sample flow, rapid
fiber or vial movement, stirring, or sonication are
required. These actions are undertaken to reduce
the effect caused by the ‘depletion zone’, which
occurs close to the fiber as a result of fluid
shielding and slow diffusion of analytes in liquid
media.

In the headspace mode Fig. 3a , the analytes

need to be transported through a layer of air
before they can reach the coating. This approach
serves primarily to protect the fiber coating from
damage by high molecular-weight species and
other non-volatile concomitants present in the
liquid sample matrix, such as humic materials or
proteins. This headspace mode also allows modi-
fication of the matrix, such as a change of the pH,
without damaging the fiber. Amounts of analyte
extracted into the coating from the same vial at
equilibrium using direct and headspace sampling
are identical, as long as sample and gaseous
headspace volumes are the same. This is a result
of the fact that the equilibrium concentration
is

independent

fiber

location

the

sample

rheadspace system. If the above condition

is not satisfied, a significant sensitivity difference
between the direct and headspace approaches
exists only for very volatile analytes. The choice
of sampling mode has a significant impact on
extraction kinetics. When the fiber coating is in
the headspace, the analytes are removed from the
headspace first, followed by indirect extraction

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

238

from the matrix. Therefore, volatile analytes are
extracted faster than semi-volatile components
since they are at a higher concentration in the
headspace, which contributes to faster mass
transport rates through the headspace. Tempera-
ture has a significant effect on the kinetics of the
process, by determining the vapor pressure of
analytes. In fact, the equilibration times for
volatile components are shorter in the headspace
SPME mode than for direct extraction under
similar agitation conditions. This outcome occurs
as a result of two factors: a substantial portion of
analyte is in the headspace prior to extraction,
and diffusion coefficients in the gas phase are
typically four orders of magnitude larger than in
liquid media. Since the concentration of semi-
volatile components in the gas phase at room
temperature is small, however, mass transfer rates
are substantially lower and result in longer ex-
traction times for such species. The situation can
be improved by use of even more efficient agita-
tion techniques, such as sonication, further reduc-
ing the extraction time. The other option is to
increase the temperature; this decreases the
amount extracted at equilibrium, but it may be

acceptable if target limits of detection can still be
attained.

3.2. Coatings

The efficiency of the extraction process is de-

pendent on the analyte distribution constant
between the coating and the sample matrix. This
characteristic parameter describes properties of a
coating and its selectivity toward the analyte ver-
sus other matrix components. Specific coatings
can be developed for a range of applications.
Coating volume also determines the method sen-
sitivity, but thicker coatings result in longer ex-
traction times because diffusion is slow within the
polymer extraction phase. Therefore, it is impor-
tant to use the appropriate coating for a given
application. Coating selection and design can be
based on chromatographic experience. For exam-
ple, a very pronounced difference in selectivity
toward target analytes and interferences can be
achieved by using surfaces common to affinity
chromatography or

molecular-imprinted po-

lymers.

To date, several experimental coatings have

Table 1

Commercially available SPME coatings Supelco, Bellefonte, PA

Stationary phase

Recommended use

(

)

Polydimethylsiloxane PDMS

100

␮mrnon-bonded

Volatiles

␮mrnon-bonded

Non-polar semivolatiles

␮mrbonded

Moderately polar to non-polar semivolatiles

(

)

Polydimethylsiloxane

rdi

inylbenzene PDMS

rDVB

␮mrpartially crosslinked

Polar volatiles

␮mrpartially crosslinked

General purpose for HPLC only

(

)

Polydimethylsiloxane

rCarboxen PDMSrCarboxen

␮mrpartially crosslinked

Trace-level volatiles

(

)

Carbowax

rdi

inylbenzene CW

rDVB

␮mrpartially crosslinked

Polar analytes

(

)

Carbowax

rtemplated resin CWrTPR

␮mrpartially

Cross-linked surfactants for HPLC

Polyacrylate

␮mrpartially crosslinked

Polar semivolatiles

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

239

been prepared and investigated for a range of
applications. In addition to liquid polymeric coat-

ings, such as polydimethylsiloxane PDMS for
general applications, other more specialized ma-
terials have been developed. Table 1 presents a
summary of commercially available fiber coatings
Ž

Supelco, Bellefonte, PA and their suggested ap-

plications.

4. Theoretical aspects of solid phase
microextraction

Theoretical aspects of SPME have been dis-

w x

cussed in detail by Pawliszyn 11 ; features which
are particularly relevant to method development
are summarized below.

4.1. Thermodynamics

Solid phase microextraction is based on multi-

phase equilibration processes. In this discussion,
we consider only three phases: the fiber coating,
the gas phase or headspace, and a homogeneous
matrix, such as pure water or air. During the
sampling period, the analytes migrate between
the three phases until equilibrium is achieved
Ž

this is naturally an ideal system without consider-

ing the inhomogeneity of the matrix or chemical
or physical characteristics of the analyte, such as

instability or degradation .

The total mass of analyte present during the

extraction is therefore represented by the fol-
lowing mass balance relationship:

⬁

Ž .

C V

sC V qC V qC V

0 s

where C is the initial concentration of analyte in

matrix, C

⬁

, C

⬁

, C

⬁

are the equilibrium or final

concentrations

analyte

the

coating,

headspace and sample, and V , V , and V are the

volumes of the coating, headspace and sample,
respectively.

The coating

r headspace distribution coeffi-

cient can be defined as:

⬁

Ž .

sC rC

and the headspace

r sample distribution coeffi-

cient can be defined as:

⬁

Ž .

sC rC

The mass of the analyte absorbed on or in the

coating is given by:

⬁

Ž .

sC V

Ž . Ž .

and can be further expressed using Eqs. 2

᎐ 4

as:

K K V C V

hs c

0 s

Ž .

K K V

qK V qV

In addition,

Ž .

sK K

Ž .

and Eq. 5 , therefore, can be simplified as:

K V C V

0 s

Ž .

K V

qK V qV

Ž .

It is significant that Eq. 7 states that the

amount of analyte extracted is independent of the
location of the fiber in the system. It may be
placed directly in the sample matrix or the
headspace.

Ž .

The three terms in the denominator of Eq. 7

represent the analyte capacity of each of the

three phases: fiber coating

K V , headspace

K V

and the sample matrix V . If there is no

headspace in the system such as liquid-phase

sampling from a completely filled vial , the term

Ž .

K V can be eliminated from Eq. 7 :

K V C V

0 s

Ž .

K V

In many cases, the fiber coating

rsample matrix

distribution constant K

is relatively small with

respect to the phase ratio of sample matrix to

coating volume V

<V . In this case, the capac-

ity of the sample matrix is significantly larger than

Ž .

the capacity of the fiber coating and Eq.

becomes:

Ž .

sK V C

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

240

which means that it is not necessary to sample a
well-defined volume of sample because the
amount of analyte extracted by the fiber coating

is independent of the sample volume V

pro-

vided:

⬁

K V

⬁

cs c

This implies that the analyte concentration ra-

tio between the sample and fiber coating at equi-
librium must compensate for the several orders of
magnitude volume difference between the two
phases. Therefore, SPME sampling can be easily
adapted to field applications and can be used for
direct sampling of unknown sample volumes. In
practice, there is no need to collect a defined
sample prior to analysis, as the fiber can be
exposed directly to ambient air, water, production
stream, etc. The amount of analyte extracted will
correspond directly to its concentration in the
matrix, without being dependent on the sample
volume.

The above equations are limited to liquid-type

polymer coatings

polydimethylsiloxane

where

extraction is based on absorption phenomena and
strongly related to the extraction phase volume.
However, considerations for solid sorbent coat-

ings adsorption based are, with some restric-
tions, analogous for low analyte concentrations.

4.2. Kinetics

A short overview of the most important con-

siderations related to the kinetics of the extrac-
tion procedure can be useful for enhanced under-
standing of the effects of optimization parame-
ters.

4.2.1. Ideal

ersus practical sample agitation

SPME is an equilibrium-based extraction

method but, naturally, in any matrix an exhaus-
tion effect occurs in close vicinity to the fiber,
which means that the analyte concentration in
the matrix close to the fiber is always lower than

that in the bulk solution. Perfect agitation of the
sample serves to eliminate this concentration gra-
dient. The time required to equilibrate the ana-
lyte between the extraction polymer phase and
sample matrix phase is dependent only on the
thickness of the coating and diffusion coefficient

w x

of the analyte in the coating 9 , i.e.:

2 D

where t is the equilibration time, a is the thick-

ness of the extraction polymer phase and D is the
diffusion coefficient of the target compound in
the extraction phase.

In practice, a concentration gradient occurs

around the SPME fiber. The sample medium is
always stationary in the immediate vicinity of the
fiber and can be considered as a boundary layer
around the fiber, where convection-based mass
transfer arising from external agitation is not
effective. As diffusion is the primary means of
mass transfer in this zone, the analyte equilibra-
tion time can be estimated with use of the fol-

w x

lowing equation 9 :

K a

where K

is the distribution coefficient between

the fiber coating and sample matrix, a is the
polymer coating thickness on the fiber and D is

the diffusion coefficient of the target compound
in the sample matrix.

Eq. 12 is particularly relevant to the applica-

tion of SPME technology for field sampling e.g.

indoor air, waste streams . Because the thickness
of the boundary layer around the fiber is strongly
dependent on the presence of convective mass

transfer which is the result of the agitated sam-

ple medium , precise control of this factor is
essential for achieving reproducible results. Ac-
complishing precise control of flow in air sam-
pling is not always a simple task, due to the
natural air-circulation patterns present in such an
environment.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

241

5. Solid phase micorextraction: step-by-step
method development

5.1. Extraction mode selection

Selection of the extraction mode is based on a

consideration of the sample matrix composition
and volatility of the analyte. Generally, for inho-
mogeneous matrices or pH- or polarity-incompat-
ible matrices, the only choice is headspace sam-
pling. For example, with inorganic analytical
methods typically used, acid leaching is not com-
patible with the fibers because their working range
is limited in pH and direct extraction cannot
therefore be performed. If the aqueous matrix
contains large non-polar species, such as lipid
components, the non-polar fibers cannot be ex-
posed to them because they rapidly saturate the
extraction phase. Obviously, for clean and com-
patible matrices, both headspace and direct sam-
pling can be used.

5.2. Fiber coating selection

The chemical nature of the analyte determines

the type of polymer used for the extraction. Selec-
tion of the coating is based primarily on the
polarity and volatility of the target compound. As
a first consideration, the simple rule ‘like dis-
solves like’ applies well for liquid polymeric coat-

ings. Polydimethylsiloxane PDMS is the most
useful liquid-type coating. It is very rugged and
the extraction characteristics can be easily esti-
mated from the wide gas-chromatographic experi-
ence gained with this coating. However, for spe-
cial cases, other solid adsorptive coatings can be
considered. The distribution constant of the ana-
lyte and the coating thickness influence the sensi-
tivity of the method. Thicker coatings require
longer equilibration times, and thus the coating
of choice is the thinnest one which will provide
sufficient sensitivity for the determination.

5.3. Deri

atization method selection

Analyte derivatization may prove advantageous

or necessary for several reasons. For example, if
the target analyte is not suitable for gas chro-

matographic analysis, or for the polarity-based

extraction offered by the SPME coating such as

ionic compounds , derivatization before, during,
or after the SPME procedure can be used to
enhance the detection power by introducing se-

lect functional groups into the analyte such as
halogen groups for electron capture detection
Ž

ECD or negative chemical ionization mass spec-

trometry . Derivatization is especially important
for speciation analysis, wherein most of the target
compounds are in an ionic form.

5.4. Optimization of desorption conditions

Optimal advantage of SPME methodology is

achieved when the smallest possible desorption
volume can be realized for use with either liquid
or gas chromatographic sample introduction. The
usual spilt

rsplitless types of GC injector were

designed for large-volume sample introduction
and are therefore normally equipped with a
large-volume glass insert sufficient to accommo-
date the vapour volume arising from a few

␮l of

organic solvent e.g. 1

␮l of hexane under 138

kPa pressure and 200

⬚C creates approx. 120 ␮l of

vapour and the injector inlet must be at least 150

␮l in volume . By introducing the analyte absor-
bed on or in a sorbent bed, there is no need for
such a large desorption volume, because there is
no solvent expansion effect. Narrow-bore injec-
tion liners can thus be used to enhance sensitivity
by minimizing dispersion. HPLC systems exhibit a
close analogy with GC systems, and to achieve
narrow injection bands, the volume of the desorp-
tion chamber must be minimized in order to
increase the linear velocity of the desorption solu-

tion usually the HPLC mobile phase .

5.5. Sample

olume optimization

The sensitivity achieved with SPME method-

ology is dependent on the number of mol of
analyte extracted from the sample, as evident

Ž .

from Eqs.

and

10 . Provided the sample

volume

4V , the amount of analyte extracted is

independent of the volume of sample if distribu-
tion equilibrium is achieved. If the available sam-
ple volume is not significantly greater than the

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

242

coating volume, conventional care is needed for
the sample volume measurement. Care must also
be taken to avoid analyte losses via evaporation,
adsorption, or microbiological activity.

5.6. Optimization of the extraction time

An optimal approach to SPME analysis is to

allow the analyte to reach equilibrium between
the sample and the fiber coating. The equilibra-
tion time is defined as the time after which the
amount of analyte extracted remains constant and
corresponds, within the limits of experimental
error, to the amount extracted after infinite time.
Care should be taken when determining the equi-
libration time, since, in some cases, a substantial
reduction of the slope of the curve might be
wrongly taken as the point at which equilibrium is
reached. Such

phenomena

often

occur

headspace SPME analysis of aqueous samples,
where a rapid rise of the equilibration curve,
corresponding to extraction from the gas phase
only, is followed by a very slow increase related to
analyte transfer from water through the headspace
to the fiber. Determination of the amount ex-
tracted at equilibrium allows calculation of the
distribution constants.

When equilibration times are excessively long,

shorter extraction times can be used. However, in
such cases, the extraction time and mass transfer
conditions have to be strictly controlled to assure
good precision. At equilibrium, small variations in
the extraction time do not affect the amount of
analyte extracted by the fiber. On the other hand,
over the steep part of the curve, even small varia-
tions in the extraction time may result in signifi-
cant variations in the amount extracted. The rela-
tive error becomes larger the shorter the extrac-
tion time. Autosamplers can reproduce the time
very precisely, and the precision of analyte de-
termination can then be very good, even when
equilibrium is not reached in the system. How-
ever, this requires that the mass transfer condi-
tions and the system temperature remain con-
stant for all experiments.

5.7. Calculation of the distribution constant

The sensitivity of the method is defined by the

distribution constant of the target analyte. It is
not necessary to calculate the fiber coating

rsam-

ple-matrix distribution constant, K , when the

calibration is based on isotopically labeled stan-
dards or standard addition, or when identical
matrix and headspace volumes are used for the
standard and sample with external calibration.
However, it is always advisable to determine K ,

since this value gives more information about the
experiment and aids optimization. K

can be

used for calculation of the headspace volume,
sample volume, and coating thickness required to
reach the desired sensitivity.

The distribution constant for the direct extrac-

tion mode can be calculated from the following

Ž .

dependence obtained from Eq. 8 :

V C V

0 s

5.8. Optimization of extraction conditions

An increase in extraction temperature causes

an increase in the extraction rate, but a simulta-
neous decrease in the distribution constant. In
general, if the extraction rate is of major concern,
the highest temperature which still provides satis-
factory sensitivity should be used.

Adjustment of the pH of the sample can im-

prove 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

extract only neutral non-ionic species from wa-
ter. By properly adjusting the pH, weak acids and
bases can be converted to their neutral forms for
extraction by the SPME fiber. To ensure that at
least 99% of the acidic compound is in the neu-
tral form, the pH should be at least two units
lower than the p K

of the analyte. For basic

analytes, the pH must be larger than p K

by two

units.

5.9. Determination of the linear dynamic range

Modification of the extraction conditions

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

243

affects both the sensitivity and the equilibration
time. It is advisable, therefore, to check the previ-
ously determined extraction time before proceed-
ing to the determination of the linear dynamic
range. This step is required if substantial changes
of the sensitivity occur during the optimization
process. SPME coatings include polymeric liquids,
such as PDMS, which by definition exhibit
capability for a very broad, linear range. For
solid sorbents, such as Carbowax

rDVB or

PDMS

rDVB, the linear range is narrower be-

cause of a limited number of sorption sites on the
surface, but it can still span several orders of
magnitude for typical analytes in pure matrices.
In some rare cases, when the analyte has ex-
tremely high affinity towards the surface, satura-
tion can occur at low analyte concentrations. In
such cases, the linear range can be expanded by
shortening the extraction time.

5.10. Selection of the calibration method

Standard calibration procedures can be used

with SPME. A fiber blank should be checked first
to ensure that neither the fiber nor the instru-
ment causes interferences with the determina-
tion. The fiber should be conditioned prior to first
use by desorption in a GC injector, or other
similarly designed conditioning device. This
process ensures that the fiber coating itself does
not introduce contaminants or interfering species.
Fiber conditioning may have to be repeated after
analysis of samples containing significant amounts
of high-molecular weight compounds, since such
compounds may require longer desorption times
than the analytes of interest.

When the matrix is simple e.g., air or ground-

water , the distribution constants are very similar
to those for the unadulterated phase. It has been
shown, for example, that typical moisture levels in
ambient air, as well as the presence of salt and

ror

alcohol in water in concentrations lower than 1%,
do not usually change the K

values beyond the

w x

5% RSD typical for SPME determinations 11 . In
many such instances, calibration might not be
necessary, since the appropriate distribution con-
stants, which define the external calibration curve,
are available in the literature or can be calculated

from chromatographic retention parameters. Ex-
ternal calibration can also be used successfully
when the matrix is more complex, but well-

defined e.g., process streams of relatively con-

stant composition . Of course, calibration stan-
dards have to be prepared in a matrix of the same
composition rather than in the pure medium in
such cases.

A special calibration procedure, such as iso-

topic dilution or standard addition, should be
used for more complex samples. With these meth-
ods, it is assumed that the target analytes behave
similarly to spikes during the extraction. This is
usually a valid assumption when analyzing homo-
geneous samples. However, it might not be true
when heterogeneous samples are analyzed, unless
the native analytes are completely released from
the matrix under the conditions used, or care is
taken to ensure the added spikes are fully equili-
brated with all components of the sample matrix.
Moreover, whenever any of these methods are
used, an inherent assumption is made that the
response is linear in the concentration range
between the original analyte concentration and
the spiked concentration. While this is usually
true for fibers extracting the analytes by absorp-

tion PDMS, PA , and detectors with a wide lin-
ear range are available, problems may arise

when porous polymer fibers PDMS

rDVB, Car-

bowax

rDVB are used, or when the detector

applied has a narrow linear range. It is important,
therefore, to verify the linearity of the response
using standard solutions before applying standard
additions or isotopic dilution for calibration. To
improve the accuracy and precision, multi-point
standard additions should be used whenever prac-
tical.

5.11. Precision of the method

The most important factors affecting precision

in SPME are presented below:

䢇

agitation conditions

䢇

sampling time if non-equilibrium conditions

are used

䢇

temperature

䢇

sample volume

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

244

䢇

headspace volume

䢇

vial shape

䢇

condition of the fiber coating cracks, adsorp-

tion of high MW species

䢇

geometry of the fiber thickness and length of

the coating

䢇

sample matrix components salt, organic mate-

rial, humidity, etc.

䢇

time between extraction and analysis

䢇

analyte losses

adsorption on the walls,

permeation through Teflon, absorption by

septum

䢇

geometry of the injector

䢇

fiber positioning during injection

䢇

condition of the injector integrity of septum

䢇

stability of the detector response

䢇

moisture in the needle

To ensure good reproducibility of the SPME

measurement, all experimental parameters listed
above should be kept constant.

5.12. Automation of the method

While SPME is a very powerful investigative

tool, it can also be the technique of choice in
many applications for processing large numbers
of samples. To accomplish this task requires auto-
mation of the methodology and, as automated
SPME devices with more advanced features and
capabilities become available, automation be-
comes easier. The currently available SPME
autosampler from Varian enables direct sampling
with agitation of the sample by fiber vibration and
static headspace sampling. In some cases, custom
modifications to the commercially available sys-
tems can facilitate operation of the method closer
to optimum conditions.

6. Solid phase microextraction for speciation

(

)

analysis structural speciation

Speciation analysis of solid samples requires a

soft digestion

rleaching technique to liberate the

Ž .

target organometallic compound s into the liquid
phase. For the determination of methylmercury in
biological samples, most methods employ HCl

w x

leaching 12 , although NaOH has also been used

w x

13 , as has acetone 14 . For sediment samples,

w x

steam distillation 15 , supercritical fluid extrac-

w x

tion SFE with CO

16 and solvent extraction

w x

with toluene 14 have been studied. For extrac-
tion of butyltin species from biological samples,

w x

combined HCl

᎐acetone leaching 17 , micro-

wave-assisted digestion with tetramethylammoni-

w x

umhydroxide

TMAH

18 , sonication with

w x

NaOH

qMeOH 17 , and leaching with HBr 19

have been described. For sediment samples,
butyltin determinations typically rely on acetic

w x

acid leaching 20 , which can be assisted by mi-

w x

crowave 18 or sonication 21 techniques. Lobin-

w x

ski and Adams 22 reviewed applications of gas
chromatography coupled with plasma source de-
tectors, wherein typical sample preparation

᎐

extraction methods for organometallic com-

w x

pounds were summarized. Camel 23 recently
reviewed microwave-based extraction techniques
fo r

e n viro n m e n ta l

sa m p le s,

in c lu d in g

organometallic compounds.

6.1. Volatile metal species

ᎏ gas chromatographic

determination

w x

Bayona 24 recently reviewed the determina-

tion of organometallic species using SPME sam-
pling with gas chromatographic separation

rdetec-

tion. Rapid advances in this field have served to
expand the scope of applications, interface devel-
opment, instrument technique and fiber prepara-
tion.

6.1.1. Ethylation

Most speciation applications of SPME are

based on classical gas chromatographic determi-
nation. Table 2 summarizes the SPME-GC meth-
ods used for speciation. Methods have been de-
scribed for mercury, tin, lead, arsenic and sele-
nium. Both headspace and direct extraction
methods have been used for the sampling of
organometallic compounds and, generally, some
type of derivatization process is necessary for
their GC separation. The typical derivatization
method used for tin, lead and mercury species is
ethylation using sodium tetraethylborate reagent
Ž

NaBEt . As is evident from Table 2, the work-

()

Z.
Mester

et
al.

Spectrochimica

Acta

Part

B:
Atomic

Spectroscopy

2001

233
᎐

260

245

Table 2

SPME methods for mercury, tin, lead, arsenic and selenium speciation with GC separation

Species

Sample type

Derivatization

Fiber

rextraction timer

Chromatographic column

Detector

Detection

Reference

extraction mode

temperature program

limit

w x

Met-Hg

Water, fish

NaBEt

racetate

100

␮m PDMSr5 minr

15 m

=0.53 mm=1.5 ␮m, DB1,

AFS

3.0 ng

tissue

buffer pH 4.5

y40⬚C 30 s y30⬚Crmin to 85⬚C
Ž

1 min 20

⬚Crmin to 200⬚C 1 min

w x

Met -Hg

Gas

None

100

␮m PDMSr30 sr

15 m

=0.53 mm=1.5 ␮m, DB1

MIP-AES

␮grl

condensate

direct sampling

isothermal 30

⬚C

w x

Met -Hg

Soil

None

100

␮m PDMSr20 minr

25 m

=0.32 mm=0.25 ␮m, HP-1,

MIP-AES

3.5

␮grl

Et -Hg

direct sampling

⬚C 5 min y40⬚Crmin to 200⬚C

1 min

w x

Met-Hg

Biological

Hydride

Fused-silica fiber

10 m

=0.25 mm, CPL-SIL 5CB

AAS

Not reported

samples,

generation

pretreated with conc.

coating, isothermal 40

⬚C,

quartz

sediments

KBH

racetate

HF acid for 3.5

᎐4 h r

tube

buffer pH 3

1.5

᎐2 hrHS

w x

Met-Hg

Soil

Hydride

Fused-silica fiber

30 m

=0.32 mm=0.25 ␮m, SPB-1,

AAS

Not reported

Et-Hg

generation

pretreated with conc.

⬚C 1 min y40⬚Crmin to 65⬚C

quartz

Phen-Hg

KBH

racetate

HF acid for 3.5

᎐4 h r

1 min , 150

⬚C 1 min , 200⬚C

tube

buffer pH 4

1.5

᎐2 hrHS

1 min

w x

Urine

NaBEt

rbuffer

100

␮m PDMSr15 minr

30 m

=0.25 mm=0.25 ␮m HP-5,

EI-MS

93 ng

Met-Hg

pH 4

⬚C 3 min y12⬚Crmin to 280⬚C

303 ng

w x

Water, fish

NaBEt

racetate

100

␮m PDMSr15 minr

30 m

=0.32 mm=1.8 ␮m, DB-624,

EI-MS

3.5

᎐8.7 ngrl

Met-Hg

tissue

buffer pH 4.5

HS and LPh

⬚C, y10⬚Crmin to 80⬚Cr

7.5

᎐6.7 ngrl

⬚Crmin to 260⬚C 2 min

w x

TeMT

Water,

NaBEt

racetic

100

␮m PDMSr20 minr

30 m

=0.32 mm=1.8 ␮m, DB-624,

FPD

41 ng

TMT

seawater

acid buffer pH 4

⬚C 2 min y10⬚Crmin to 150⬚C

15 ng

DMT

2 min

8.4 ng

MMT

8.6 ng

w x

MBT

River water

NaBEt

rethanoic

100

␮m PDMSr60 minr

30 m

=0.25 mm=0.25 ␮m

FPD

2 ng

DBT

acid buffer

LPh

methylsiloxane , 70

⬚C 1 min

2 ng

TBT

pH 4.8

y15⬚Crmin to 270⬚C 6 min

4 ng

MPhT

1 ng

DPhT

2 ng

TPhT

3 ng

()

Z.
Mester

et
al.

Spectrochimica

Acta

Part

B:
Atomic

Spectroscopy

2001

233
᎐

260

246

Table 2

Continued

Species

Sample type

Derivatization

Fiber

rextraction timer

Chromatographic column

Detector

Detection

Reference

extraction mode

temperature program

limit

w x

MBT

Environmental,

NaBEt

racetate

100

␮m PDMSr60 minr

30 m

=0.25 mm=1.0 ␮m, SPB-1,

FID

1.0

␮grl

DBT

sediment

buffer pH 4.0

⬚C 1 min y20⬚C-min to 140⬚C

1.2

␮grl

TBT

1 min

y20⬚Crmin to 220⬚C 1 min

0.9

␮grl

w x

MBT

Sediment,

NaBEt

rethanoic

100

␮m PDMSr60 minr

30 m

=0.25 mm=0.25 ␮m

FPD

0.031 ng

DBT

sewage sludge

acid buffer pH 4.8

LPh

methylsiloxane , 70

⬚C 1 min

0.007 ng

TBT

y15⬚C min to 270⬚C 6 min

0.006 ng

MPhT

0.114 ng

DPhT

0.167 ng

TPhT

0.583 ng

w x

MBT

Slurry of

NaBEt

racidified

␮m PDMSr45 minr

25 m

=0.32 mm=0.17 ␮m, HP-1,

MIP-AES

␮grl range

DBT

sediment

with HCl

LPh

⬚C, y20⬚Crmin to 230⬚C 0.75 min

TBT
TeBT

w x

Spiked water

NaBEt

racetate

100

␮m PDMSr15᎐20

30 m

=0.25 mm=0.25 ␮m,

EI-MS

0.2

␮grl

buffer pH 4.0

min

rHS

Omegawax

r40⬚C 1 min y20⬚Crmin

to 120

⬚C 1 min

w x

Blood, urine

NaBEt

racetate

100

␮m PDMSr15 minr

30 m

=0.25 mm=0.25 ␮m,

EI- MS

2-3

␮grl

buffer pH 4.0

SPB 5

r40⬚C 1 min -20⬚Crmin to

120

⬚C 4 min

w x

᎐

Deuterated

100

␮m PDMSr10 minr

30 m

=0.25 mm=0.25 ␮m, SPB 5,

EI- MS

95 ng

TML

NaBEt

racetate

⬚C isothermal

130 ng

TEL

buffer pH 4.0

83 ng

TeEL

90 ng

w x

River and

Piaselenol

Not reported

r30 minr

30 m

=0.25 mm=0.25 ␮m, 80⬚C,

EI-MS

6 ng

Total Se

tap water

formation

2 min

y10⬚Crmin to 280⬚C

w x

DMA

Urine

Thioglycol

100

␮m PDMSr40 minr

15 m

=0.25 mm=0.25 ␮m, SPB 5,

EI-MS

0.29

␮grl

MMA

methylate

LPh

110

⬚C 1 min y20⬚Crmin to 230⬚C

0.12

␮grl

derivatization

()

Z.
Mester

et
al.

Spectrochimica

Acta

Part

B:
Atomic

Spectroscopy

2001

233
᎐

260

247

Table 2

Continued

Species

Sample type

Derivatization

Fiber

rextraction timer

Chromatographic column

Detector

Detection

Reference

extraction mode

temperature program

limit

Multielement

w x

MBT

Sediment

NaBEt

racetate

100

␮m PDMSr10 minr

30 m

=0.25 mm=0.50 ␮m

ICP-MS

0.34 ng

w x

DBT

buffer pH 5.3

polydimethylsiloxane , 60

⬚C 1 min

2.1 ng

TBT

y20⬚Crmin to 200⬚C 0.5 min

1.1 ng

Met-Hg

4.3 ng

TML

0.19 ng

w x

MBT

Body fluids

NaBEt

racetate

100

␮m PDMSr10 minr

30 m

=0.25 mm=0.50 ␮m,

EI-MS-MS

9 ng

DBT

buffer pH 5.3

VA-5,

y40⬚C 3 min

13 ng

-10

⬚Crmin to 250⬚C

TBT

9 ng

Met-Hg

22 ng

18 ng

TML

7 ng

w x

Alkylmercury

Surface water,

NaBEt

racetate

100

␮m PDMSr10 minr

30 m

=0.25 mm=0.50 ␮m

ICP-MS

3.7 ng

Alkyltin

sediment

buffer pH 5.0

polydimethylsiloxane , 60

⬚C

0.38

᎐1.2 ngrl

Alkyllead

1 min

y20⬚Crmin to 200⬚C

0.13

᎐0.15 ngrl

0.5 min

TML, trimethyllead; TEL, triethyllead; TeEL, tetraethyllead; DMA, dimethylarsinic acid; MMA, monomethylarsonic acid; MBT, monobutyltin; DBT, dibutyltin;

TBT, tributyltin; MPhT, monophenyltin; DPhT, diphenyltin; TPhT, triphenyltin; NaBEt , sodium tetraethylborate; LPh, liquid phase; HS, headspace; AAS, atomic

absorption spectrometry; AFS, atomic fluorescence spectrometry; EI-MS, electron impact ionization-mass spectrometry; FID, flame ionization detection; FPD, flame
photometry detection; ICP-MS, inductively coupled plasma-mass spectrometry; and MIP-AES, microwave-induced plasma-atomic emission spectrometry;

Detection limit normally based on the concentration of solution sampled by SPME.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

248

ing pH range for the derivatization is 4.0

᎐5.3,

depending on the target compounds. An advan-
tage of NaBEt

is that derivatization can be ac-

complished in an aqueous environment, the natu-
ral medium for most environmental and biologi-
cal samples. There is no need to change phases,
as in the case of Grignard reagents. Derivatiza-
tion with NaBEt offers the unique possibility for

multi-elemental speciation of tin, mercury and
lead species. Fig. 4 shows such an example based
on NaBEt

derivatization of the analytes and

headspace sampling by SPME with detection by
gas chromatography coupled with inductively cou-

w x

pled plasma mass spectrometry. Moens et al. 25
reported a comparison of sensitivity between
‘conventional’ liquid

rliquid and headspace SPME

extraction of butyltin and organolead compounds.
They found that headspace SPME provided ap-
proximately 300-fold better sensitivity for butyltin
compounds and approximately 35-fold enhance-
ment for trimethyllead. Analysis of butyltin

species based on SPME extraction can thus be
accomplished using a conventional flame ioniza-

tion detector FID . One of the main limitations
of this technique is that it cannot be applied to
the speciation of ethyl ligand-containing species.
For example, reaction of triethyllead and inor-
ganic lead with NaBEt produces the same com-

w x

pound: tetraethyllead. Yu and Pawliszyn

overcame this limitation by utilizing deuterated
NaBEt for the derivatization of organolead com-

pounds.

A typical sampling procedure involves the fol-

lowing: for solid samples and for most biological
tissues a ‘soft’ digestion is used to drive the target
compounds into the liquid phase. For homoge-

neous liquid samples such as seawater , no diges-
tion step is necessary. The aqueous sample is
typically transferred to a 40-ml septum-sealed
glass vial, pH is adjusted to a predetermined
value, the derivatization agent is added and the
vial is tightly sealed. To facilitate mass transfer in

Fig. 4. Chromatograms derived from an organometal standard extracted with headspace SPME after ethylation with sodium

w x

120

126

202

208

Ž .

tetraethylborate 25 .

Sn,

Xe,

Hg and

Pb were simultaneously monitored. 1 Methylmercury; 2 Trimethyllead; 3

Ž .

. Ž .

Ž .

Diethylmercury; 4 Inorganic Sn tetraethyltin ; 5 Inorganic Pb tetraethyllead ; 6 Monobutyltin; 7 Dibutyltin; 8 Tributyltin;
X, unknown components.

䊚 American Chemical Society ᎏ reprinted with permission.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

249

the system, the solution is often continuously
stirred by use of a magnetic stirrer. Aguerre et al.

w x

27 examined the use of mechanical shaking for

liquid phase extraction of organotin species. The
SPME fiber is subsequently exposed to the
headspace or the liquid phase of the sample for a
predetermined time. After extraction, the fiber is
transferred to a suitable desorption unit.

6.1.2. Hydride generation

Because of the high volatility of metal hydrides,

few researchers have reported results on the
headspace sampling

rdetermination of metal

hydrides using solid phase microextraction tech-

niques. Jiang and co-workers 28,29 developed a
sampling method for organomercury species based
on hydride generation using KBH

reagent.

Commercially available fibers were reported as
not suitable for the extraction of mercury
hydrides and, consequently, new fibers were de-
veloped. Basically, a fused silica fiber was
immersed into a hydrofluoric acid solution for
typically 3

᎐4.5 h. Further increases in this treat-

ment time rapidly decreased the extraction capac-
ity of the fiber. The optimum extraction time was
reported to be 1.5

᎐2 h, which is relatively long. In

this case, extraction is based on surface interac-
tions between the organomercury hydrides and
the treated glass fiber. As this is predominantly
an adsorption phenomenon, the long equilibra-
tion time between the extraction surface and
sample solution can be understood. The hydro-
fluoric acid treatment likely increases the surface

w x

porosity of the silica fiber. Mester et al. 30
tested the compatibility of two different fibers
Ž

and two different extraction phenomena

for

sampling volatile metal hydrides coupled with

Fig. 5. Transient signal obtained by desorption of arsine from

Ž .

Carboxen- 1 or PDMS- 2 coated fibers with ICP-MS detec-

tion based on response from As

ICP-MS detection. An adsorption-based Car-
boxen coating provided better sensitivity than an
absorption-based extraction with a liquid-type po-

lymeric coating PDMS . The success of absorp-
tive sampling confirms the relatively high stability

of the metal hydrides studied As, Se, Sn and Sb
because they survive diffusion into the polymeric
liquid. However, as is evident from Fig. 5, the
desorption characteristics of the two fibers are
different. The transient signal from arsine ob-
tained with the PDMS fiber shows significant
tailing compared to the signal from a Carboxen
fiber, suggesting that partial decomposition of the
arsenic hydrides occurs during the extraction

rde-

sorption cycle. Table 3 summarizes detection
limits obtained by headspace sampling of metal
hydrides from a non-pressurized sample derivati-
zation system.

6.1.3. Other deri

atization methods

w x

Guidotti et al. 31 described the determination

of Se IV by headspace SPME-GC

rMS following

derivatization with piaselenol formation. Piase-

Table 3

w x

Figures of merit for metal hydride determination by SPME

᎐ICP-MS 30

ng ml

PDMS

0.32

11.0

3.2

1.8

0.96

35.0

9.7

5.3

Carboxen

0.007

5.3

0.008

0.31

0.02

15.0

0.024

0.92

Limit of detection 3

␴ .

Limit of quantification 10

␴ .

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

250

lenol formation is a well-known method for Se IV
derivatization prior to GC analysis. As in the case
of ethylation with sodium tetraethylborate, an
SPME method can serve as a substitute for liq-
uid

rliquid extraction. In this case, the piaselenol

complex is extracted from the liquid phase by
immersing a 100-

␮m PDMS-coated fiber into the

sample solution. The method can also be applied

to Se IV and Se VI speciation by determining

the Se IV content first, followed by the total
selenium concentration after converting all sele-

nium species into the Se IV oxidation state.

w x

Mester and Pawliszyn 32 reported a method

for the speciation of dimethylarsinic acid DMA

and monomethylarsonic acid MMA by SPME-

rMS. Thioglycolmethylate TGM was used

for derivatization. The method is based on the
known ‘affinity’ of the arsenic compound for the

w x

thiol group 33 .

(

)

6.1.4. Direct determination no deri

atization

Organometallic species which are normally sat-

urated non-ionic and sufficiently volatile can be
sampled by SPME and determined by GC without

w x

derivatization. Gorecki and Pawliszyn 34 de-
scribed a simple sampling procedure for teraethyl-
lead, wherein a 100-

␮m PDMS fiber was exposed

to the headspace of an aqueous sample. GC-MS

was used for quantitation. Similar studies were

w x

performed by Snell et al. 35 for the determina-
tion of dimethylmercury in the headspace of nat-
ural gas condensates. This method was character-

ized by very short sampling times 30 s . The
extremely complex matrix, comprising volatile or-

ganic material which is also extracted with the

dimethylmercury , presented a serious problem,
even for the extremely selective microwave-
induced plasma atomic emission detection system
used.

w x

Mester et al. 36 recently described a SPME

method

for

methylmercury

determination.

Headspace SPME sampling was performed above
a methylmercury solution which was previously
saturated with sodium chloride. This method was
based on the relatively high vapour pressure of
methylmercury chloride. A slightly polar, solid

coating PDMS

rDVB was used for extraction.

Sample introduction into an ICP-MS was achieved
with a unique thermal desorption interface, con-
sisting of a heated glass-lined splitless-type GC
injector placed directly at the base of the torch to
minimize the length of transfer line. This ar-
rangement provided for fast desorption and high
sample introduction efficiency. Direct liquid
immersion and headspace extraction of methyl-
mercury were studied. For clean solutions,

Fig. 6. Schematic illustration of the thermal desorption interface for SPME analyte introduction into ICP-MS.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

251

immersion sampling SPME provided good sensi-
tivity that was linear over two orders of magni-
tude, whereas headspace sampling showed 15%
lower sensitivity, but a linear range of more than
three orders of magnitude. Calibration by the
method of addition using direct extraction re-
vealed a severe matrix effect with biological tissue
samples, diminishing the methylmercury response
70-fold. That obtained by headspace extraction
was statistically indistinguishable from signals
generated using matrix-free standards. The real
novelty of the method was that speciation could
be achieved without employing any separation

method GC or HPLC , and was based entirely
on the selectivity of the extraction procedure and
the detection technique. As a secondary result of
this work, a new sample introduction approach
was developed, which can be used with any type
of atomic spectroscopic instrumentation. In addi-

tion to SPME, small volumes 1

᎐3 ␮l of organic

solvents may be introduced into the ICP with this
device. A schematic diagram of the thermal des-
orption ICP-MS interface is shown in Fig. 6.

6.2. Non-

olatile metal species

ᎏ liquid

chromatographic determination

SPME can also be used for sampling of non-

volatile

rionic metal species from the aqueous

phase. Following extraction, desorption can be
achieved off-line in a small vial, or on-line using a
‘T’ joint or commercially available liquid desorp-

tion chamber Supelco, PA as a substitute for the
sample introduction loop in an HPLC system.

One of the early applications of SPME in the

inorganic field was described by Otu and Pawliszyn

w x

37 . A PDMS fiber modified with a liquid ion-

w Ž

exchanger di- 2-ethylhexyl phosphoric acid was

used for the extraction of bismuth III from aque-

ous solution. The extracted Bi III was desorbed
into an acidic potassium iodide solution and the
yellow-colored BiI

complex was quantitated

spectrophotometrically.

w x

A similar study was undertaken by Jia et al. 38

for the determination of mercury. The SPME
fiber was modified for sampling, in that different
crown ethers, adsorbed onto the non-polar SPME
coating by simply dipping the fiber into a solution

of the crown ether, were used. An HPLC system
was subsequently employed for the separation of
free complexing agent from the metal complex
and, with UV-based detection, a detection limit
limit of approximately 500 ng

rml was achieved,

limited by the low sensitivity of the UV detector

w x

used in the study. Boyd-Boland 39 described a

method for speciation of chromium III and VI
based on the simultaneous extraction of an EDTA

complex of Cr III and direct extraction of Cr VI .
Different fibers were tested, with the best result
being obtained with Carbowax

rtemplated resin

CWAX

rTR . Following extraction, the analytes

were transferred to an HPLC for separation,

where the Cr III and Cr VI states were detected

with UV absorption. No interferences from Co II ,

Cu II or Fe III were observed, based on recov-

ery of Cr III .

Because there are no commercial ion-exchange

or specific metal-selective coatings available, the
study of new coating materials and extraction

w x

principles is significant. Caruso et al. 40 recently
reported

SPME

sol

᎐gel coatings for

organometal determination. Sol

᎐gel coatings are

very stable and durable in organic solvents, as
well as acidic and basic solutions, where commer-
cial SPME coatings cannot be used. Preparation
of fibers with hydrophobic coatings for SPME-
HPLC applications using the sol

᎐gel process was

described. The fibers were evaluated for SPME-
HPLC

applications

using

diphenylmercury,

trimethylphenyltin and triphenylarsine. The effect
of organic solvents and acids on the fibers was
also studied.

Electrochemical control can also be applied to

the extraction and

ror desorption of analyte

into

rfrom the extraction phase. In order to

achieve electrochemical control in the aqueous
sample system using SPME, it is necessary to
have a conductive extraction phase on the SPME
fiber. Commercial SPME fibers are not suitable
for this kind of work, because the extraction
phases and silica support are not conductive. Re-

w x

cently, Guo et al. 41 described a system in which
a gold-coated carbon-steel wire was used for col-
lecting mercury from aqueous samples. In this
case, both the extraction phase and the support
phase are conductive and suitable for elec-

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

252

trochemical control. The gold-coated wire served
as a working electrode in an electrochemical cell
and, by applying the necessary potential to it, the

Hg II present in the solution was reduced to
elemental mercury and collected as an amalgam
in the gold film. After the ‘extraction’ period, the
fiber was transferred to a capacitively heated GC
injector, wherein the mercury was rapidly liber-
ated from the amalgam and transported to an ion
trap mass spectrometer with electron impact ion-
ization. In this study, the electrochemical driving
force was used exclusively for the sampling
process. However, the potential difference can
also be applied to facilitate the release

rdesorp-

tion of the analyte from the polymeric extraction

w x

phase, as reported Gbatu et al. 42 . In this case,
the extraction phase was a conductive polymer,

namely poly 3-methylthiophene , coated on a
platinum wire. This special electrode, or fiber,
was used for sampling arsenate from aqueous

solutions. By applying

q1.2 V potential vs.

rAgCl , the arsenate anion AsO

migrated

into the polymeric layer to maintain electro-neu-
trality. Upon reversal of the potential to

y0.6 V

vs. Ag

rAgCl , the polymer was converted back

to its neutral hydrophobic form and the arsenate
ions were expelled into a smaller volume solution

and subsequently introduced into an ICP-MS via
flow injection using de-ionized water as the car-
rier.

w x

In a recent paper, Wu et al. 43 described the

application of a polypyrrole polymer coating for
SPME extraction of small anionic analytes, such
as Cl

, F

, Br

, NO

, SO

, SeO

SeO

and AsO

. Extraction was based on the

ion exchange characteristics of the polypyrrole
coating.

A different approach to the use of liquid phase

microextraction is to employ an open tubular
capillary substitute for the sample introduction
loop in an HPLC system. The interior wall of the
capillary tube is coated with a suitable extraction
polymer. One of the main advantages of such an
in-tube extraction system is simple achievement
of matrix separation. The polarity-based extrac-
tion should efficiently separate target compounds
from other small ionic constituents, such as chlo-
ride and phosphate, or from large macro-
molecules, such as proteins. Using an auto-injec-
tor, with a six-port valve in the ‘load’ position, the
capillary is washed several times with the sample
solution by applying ‘draw from sample

᎐eject into

sample’ extraction cycles. The cycles are repeated
a number of times until equilibrium is reached

Fig. 7. Schematic of the in-tube solid phase microextraction-liquid chromatography-electrospray ionization mass spectrometry
system.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

253

between the analyte concentration in the sample
and its concentration in the extraction phase.
After the extraction step, the valve is switched to
the ‘inject’ position and mobile phase passes
through the extraction capillary, transporting ana-
lyte into the HPLC column or direct to a detec-
tor. Fig. 7 schematically illustrates the in-tube
solid phase microextraction system.

w x

Mester et al. 44 described an in-tube SPME

system for the extraction of trimethyllead and
triethyllead from aqueous samples and evaluated
different coatings for extraction efficiency. Fig. 8
shows typical chromatograms for triethyl- and
trimethyllead obtained using various extraction
capillaries.

In-tube SPME has also been applied to the

determination of selenomethionine SeMet and

selenoethionine SeEt in several biological tis-

w x

sues 45 .

7. Distribution

requilibrium speciation using

solid phase microextraction

Solid phase microextraction can be readily ap-

plied to the study of the distribution of target
compounds amongst real- or pseudo-phases. The
bioavailability of a target compound depends on
which phase the analyte is located in. For exam-
ple, certain drugs entering the blood stream be-
come bound to blood proteins, diminishing their
therapeutic effect. Therefore, it is crucial to mon-
itor the free drug level in the blood during inten-
sive drug therapy. Similarly, the distribution and
fate of organic and organometallic compounds in
the environment depends on their sorption onto
inorganic particulates, such as clay minerals, oxy-
hydroxides, and hydrous iron and manganese
oxides, as well as humic material in sediments,
soils, and aquifers. Consequently, information

Fig. 8. Comparison of three polar GC capillaries for TML and TEL extraction. Solution concentrations are 1 ppm Pb by weight of

TML and TEL. Coatings: Supel-Q-Plot, a porous divinylbenzene polymer; OmegaWax 250, a bonded poly ethylene glycol ; and

Nukol, a poly ethylene glycol modified with nitroterephthalic acid.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

254

characterizing such distributions is invaluable for
modeling

rprediction.

The significant advantage of such an equilib-

rium-based extraction is that it can be applied to

the measurement of concentrations of free dis-

solved

portions of target compounds, as de-

w x

scribed by Vaes et al. 46 and by Poerschmann et

w x

al 71 . If the target compound X is in equilib-

rium with a matrix M in aqueous solution, the
equilibrium can be described as:

qM|XM

If an exhaustive extraction procedure is applied

to the target compound

X , the equilibrium

between the free and bound compound will shift
in the direction of dissociation and the results

obtained will not accurately reflect the free dis-

solved concentration. Using a non-exhaustive mi-
croextraction technique, the equilibrium between
X and M will be virtually undisturbed because of
the insignificant amount of analyte extracted, and
only the free, dissolved fraction will partition to
the SPME fiber according to:

qFiber|XyFiber

The concentration of the free

runbound com-

pound X in a matrix can be determined using a
calibration curve established in pure water for a
fixed time of exposure to the SPME fiber.

When measuring the free analyte concentra-

tion in samples containing a matrix that binds the
compound of interest, several criteria need to be
met to achieve accurate results:

The SPME extraction itself may not influence
the equilibrium that exists between the aque-
ous phase and the matrix. To ensure this, a
short

extraction

time,

low

coating

᎐r

sample

᎐volume ratio or low coating volume

should be used.

The matrix should not interfere with the mea-
surement by binding to the SPME fiber. The
selected coating must therefore be highly se-
lective between the target compound and the
pseudo-phase present in the solution.

The matrix

rwater equilibrium should not in-

fluence the amount of analyte absorbed by
the fiber; i.e. the absorption characteristics of
the target compound should be the same in a
calibration sample as in a sample containing
the matrix. In an aqueous solution containing
a matrix to which the substance is bound, any
decrease in the concentration due to uptake
on the SPME fiber, however small, will be
compensated

for

shift

the

matrix

rwater equilibrium. Therefore, the ex-

traction kinetics characterizing the SPME
fiber may be different between pure aqueous
solutions and solutions containing the matrix.

w x

Poerschmann et al. 47 used SPME for moni-

toring the sorption kinetics of butyltin species
onto dissolved and particulate humic organic mat-
ter. To study the sorption kinetics of the tin
species onto the dissolved humic substances, a

very short sampling time 15 s and 7-

␮m thick

PDMS fiber were used, offering the possibility of
quasi real-time monitoring. The effect of dis-
solved organic matter on the extraction kinetics
of the SPME fiber was also examined. The reason
for investigating this phenomenon was to ulti-
mately be able to save time with the SPME
procedure, i.e. to extract under non-equilibrium
conditions. Calibrating SPME under non-equi-
librium conditions is possible only if the sample

to be analyzed e.g., dissolved organic matter
Ž

DOM

and the calibration sample

without

DOM

display identical extraction kinetics. In

humic and fulvic acid-containing model solutions,
the extraction kinetics characterizing the SPME
fiber were different from those in a clean water
solution, likely due to the effects of the surfac-
tant-like properties of the humic compounds.

w x

Poerschmann et al. 47 concluded that the results
provided strong evidence that SPME must be
applied under equilibrium conditions when exter-
nal calibration is used for such studies.

8. Solid phase microextraction as an investigative
tool

Solid phase microextraction can, and has been,

used as a research tool. Because of the non-inva-

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

255

sive nature of the sampling process, it is suitable
for sampling in sensitive or difficult environments
and is also fast and easy to use with volatile and
semi-volatile compounds. For example, Barshick

and co-workers 48

᎐50 published several papers

focusing on the determination of mercury species
in soils using various detection methods. The
sampling and sample introduction methods ap-
plied were SPME-based.

w x

Gorecki et al. 51 applied SPME for sample

preparation and analyte introduction into a radio
frequency glow discharge mass spectrometer. The
distinctive fragmentation patterns obtained for
tetramethyltin and tetraethyllead species offer the
possibility of direct speciation without separation.

w x

Rosenkranz et al . 52 used SPME sampling

and sample introduction for the study of
transalkylation processes in different mercury
species.

w x

Mester and Pawliszyn 53 employed in-tube

solid phase microextraction for sample introduc-
tion into an electrospray mass spectrometer. The
main purpose of the study was the elucidation of
the collision-induced fragmentation patterns of
different organolead species.

9. Limitations of solid phase microextraction

The principal limitations of SPME technology

are strongly linked to the main advantages of the
technology. The volume of the polymer extraction
phase is very small and, although responsible for
the non-exhaustive nature of the extraction, re-
quires extreme precision during manufacturing of
the coating so as to reproducibly produce the
same coating quality. The reuse character of the
SPME fiber is also a significant cost-related ad-
vantage but, at the same time, some level of
degradation of the fiber occurs during repeated
usage, resulting in compromised accuracy and

w x

precision. Haberhauer-Troyer et al. 54 recently
presented a study of the surface characteristics of
SPME coatings. Using electron microscopy, the
surface

new

and

used

PDM S

and

PDMS

rCarboxen fibers were examined. Damage

to the coatings was detected on the tip of the

fiber and at the bond between the polymer-coated
fused silica fiber and the fiber attachment tubing.
The pore size ranges up to 40

␮m on the surface

of porous PDMS

rCarboxen fibers and the pore

density is highly variable between fibers. On the
surface of PDMS fibers, physical contamination
was detected; metal particles were present, which

likely came from the septum-piercing needle the
composition of the particles was determined and

found to closely match the needle composition .
Metallic contamination on the fibers can pose a
serious threat to the field of trace metal analysis.

w x

Martos and Pawliszyn 72 studied intra- and

inter-fiber reproducibility for 10 different fibers
employing 30-

␮m PDMS fibers with a standard

gas concentration of 34

␮grl for 11 VOCs at 298

K. They obtained excellent day-to-day consistency

for intra-fiber reproducibility

f2.5% RSD . An

inter-fiber estimation of method precision for the
10 different fibers yielded approximately 3

᎐9%

RSD for each of the 11 analytes studied. How-
ever, discrepancies were found in the accuracy of
the data obtained using different fibers. The
length of the individual fibers ranged from 0.98 to
1.06 cm, with an average length of 1.02

"0.30 cm.

When a correction for fiber length was applied to
the data, the differences between the individual
fibers were found to be statistically insignificant.

w x

Mester et al. 36 also reported data relating to

the intra- and inter-fiber reproducibility achieved
during methylmercury analysis. Table 4 shows re-
sults for intra- and inter-fiber performance ob-
tained with five replicate headspace samplings of
a 10-ng

rml sample solution of methylmercury

using a PDMS

rDVB fiber. The three fibers had

not been used prior to this process and were used
as provided by the supplier. For any given fiber,

the

intra-

reproducibility of measurement is

quite good, averaging approximately 2.3% RSD.
However, the RSD of the means for the three
fibers is greater than 20%. More alarming is the
difference between the two extremes, which is
more than 30%. These data are likely a reflection
of the quality of manufacture of the fibers. Sur-
prisingly, over the past several years of commer-
cial history of SPME and the hundreds of re-

search papers

organic analysis

utilizing this

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

256

technology, only a few discussions have arisen
concerning the quality of the results with respect
to fiber-to-fiber performance. The unique aspect
of manufacture of SPME fibers is the extremely
low volume of the extraction phase. Any irregu-
larity

inhomogeneity

the

polymer

phase

rsurface may result in a significant effect

on its extraction characteristics. This effect could
be more pronounced in the case of non-equi-
librium extraction with solid coatings, where the
extraction is based on adsorption phenomena
rather than absorption.

Although solid phase extraction SPE and cap-

illary GC column technology are similar, the re-
sultant effects of inhomogeneities in these coat-
ings are not as serious because, in both cases, the
extraction or separation is based on bulk charac-
teristics of the polymer phase. If the particle size
distribution in the case of SPE falls within speci-
fied limits, the variation between individual parti-
cles has no significant impact. If the capillary GC
column coating exhibits fluctuation in thickness
or, eventually, surface cracks appear along the
length of the column, these effects are buffered
by the 1

᎐100 m length of the column. If these, or

any other type of manufacturing irregularities,
occur on the surface of a 1-cm long SPME fiber,
the extraction characteristics could be seriously
altered.

The imprecision of the results obtained with

different fibers often makes it necessary to em-
ploy an internal standard or standard addition

method, which, unfortunately, can significantly in-
crease the analysis time.

10. Future potential of solid phase
microextraction

10.1. Equilibrium speciation

As stated in Section 1, it is the authors’ opinion

that the fastest-growing field of speciation analy-
sis will be the equilibrium

᎐distribution-type of

speciation. This will encompass not only the al-
ready described bound vs. free metal content
determination, but also molecular vs. dissociated,
active vs. inactive form of metal-containing
macromolecules and optically active vs. optically
inactive

conformations

metal-containing

molecules

such as seleno-amino acids . These

are only a few examples where non-exhaustive
microextraction methods such as SPME will play
key roles in the near future. Because this sam-
pling method permits different chemical and
physical processes to be followed without perturb-
ing them, studies now taking place in vitro will,
with some modification, eventually occur in vivo.
It is not difficult to imagine monitoring specific
drug levels during intensive drug therapy using a
syringe or an open tubular capillary designed for
continuous or semi-continuous sampling of body
fluids. With the miniaturization of sampling
probes, it may even be possible to sample single

Table 4

Ž .

Intra- and inter-fiber s precision

Replicate

Intra-fiber precision

Intensity counts

Average

RSD

Fiber 1

23 528

24 362

22 845

23 456

23 345

23 507

547

2.33

Fiber 2

18 635

18 523

18 856

17 854

18 886

18 551

418

2.25

Fiber 3

15 525

16 124

15 256

15 856

16 032

15 759

362

2.30

19 272

3924

w x

PDMS

rDVB fiber with non-equilibrium sampling of a 10 ngrml solution of methylmercury chloride 36 .

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

257

cells and follow events occurring at the cellular
level.

10.2. Industrial hygiene

Ž .

As is evident from Eq.

9 , the amount of

analyte collected by the SPME fiber is indepen-

dent of the sample size subject to specific condi-

tions being met . This means that sampling of
unknown volumes of air, or any other compatible
matrix, can be achieved without the necessity of
precisely measuring the sample volume. Most an-

alytical methods concerned with air quality e.g.,

volatile organic compounds VOCs or mercury

vapour determination in the workplace are based
on an active sampling process. A known volume

of air is pumped through a trap liquid or solid
sorbent or physical device, such as a cryogenic

trap and optimally 100% of the analyte is re-
moved from the air stream and retained on the
extraction cartridge. The process is one of ex-
haustive extraction. Quantification is based on
the known volume of sample and the instrumen-
tal response, which is easily externally calibrated
by simply injecting a known amount of analyte
into the detector and comparing the response
obtained for the standard with the response ob-
tained for the analyte.

Passive sampling by SPME does not require

calibrated pumping systems; the SPME fiber can
be directly exposed to the matrix for a prede-
termined time. After the sampling period, the
analyte can be measured in the field, or the fiber
can be sent to a central laboratory for analysis.
SPME sampling methodology thus offers the pos-
sibility for multiple sampling of a specified area
or mapping target compound distributions using a
large number of inexpensive SPME sampling de-
vices. This type of sampling cannot be easily
completed using an active sampling process, be-
cause every sampling point requires a pumping
system. Calibration of passive sampling devices
Ž

SPME fibers can be performed by exposure to a

standard atmosphere of target compound for the
same period of time as the sampling period. To
obtain sufficient precision with passive sampling,
it is necessary to work in the equilibrium-sam-

pling regime, wherein slight changes in the ex-
traction kinetics cannot perturb the results.

For collection of mercury or other volatile

metallic compounds from air, cryogenic trapping

is typically used. Donard and co-workers 55,56
published several papers devoted to this subject.
However, as was shown for sampling organomer-

cury or metal hydrides or basically any other

volatile metallic species , this can be conveniently
accomplished using solid phase microextraction.
The principal challenge is to create well-defined
standard atmospheres for inorganic compounds
necessary for accurate calibration of the passive
sampling devices. For some species, such as ar-
sine, this is already available, while for others,
conventional gas generation methods can be ap-
plied, based on evaporation of a known amount
of analyte into a reservoir of known volume. For
compounds which are stable in solution, such as
methylmercury halides, this evaporation approach
can easily be utilized, but for other species which

are not available in solution such as most of the

metal hydrides , other means must be devised.

Passive sampling with SPME also offers the

possibility of utilizing this approach

as a

personal-exposure monitoring device, by simply
attaching it to the clothes of the worker. A time-

weighted average TWA result can be obtained,
which may reflect the average concentration of
target pollutant to which exposure was made dur-
ing the sampling period. Martos and Pawliszyn

w x

57 have discussed the details of TWA sampling

with SPME.

References

w x

H.M.S. Kingston, P.J. Walter, The art and science of
microwave sample preparations for trace and ultratrace

elemental analysis, in: A. Montaser Ed. , Inductively
Coupled Plasma Mass Spectrometry, Wiley, New York
Ž

NY , 1998, pp. 33

᎐81.

w x

E. Nieboer, Speciation of elements in environmental
and biological sciences, XXVII-CSI post-symposium,

Analyst 117 1992 550.

w x

E. Nieboer, Y. Thomassen, Analyst 120

1995

30N

Second International Symposium on Speciation of Ele-

ments in Toxicology and in Environmental and Biologi-

cal Sciences .

w x

J. Szpunar, Bio-inorganic speciation analysis by hyphen-

ated techniques, Analyst 125 2000 963

᎐988.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

258

w x

C. Sarzanini, E. Mentasti, Determination and speciation
of metals by liquid chromatography, J. Chromatogr. A

789 1997 301

᎐321.

w x

R. Lobinski, F.C. Adams, Speciation analysis by gas
chromatography with plasma source spectrometric de-

tection, Spectrochim. Acta Part B 52 1997 1865

᎐1903.

w x

J.W. Olesik, J.A. Kinzer, E.J. Grunwald, K.K. Thaxton,
S.V. Olesik, The potential and challenges of elemental
speciation by capillary electrophoresis inductively cou-
pled plasma mass spectrometry and electrospray or ion
spray mass spectrometry, Spectrochim. Acta Part B 53
Ž

1998 239

᎐251.

w x

F. Soxhlet, Dingers Polytech. J. 232 1879 461.

w x

J. Pawliszyn, New directions in sample preparation for
analysis of organic compounds, Trends Anal. Chem. 14
Ž

1995 113

᎐122.

w x

H.H. Liu, P.K. Dasgupta, Analytical chemistry in a drop,

Trends. Anal. Chem. 15 1996 468

᎐475.

w x

J. Pawliszyn, Solid Phase Microextraction, Wiley, New

York NY , 1997.

w x

E. Bulska, D.C. Baxter, W. Frech, Capillary column gas
chromatography for methylmercury speciation, Anal.

Chim. Acta 249 1991 545

᎐554.

w x

H. Emteborg, N. Hadgu, D.C. Baxter, Quality control of
a recently developed analytical method for the simulta-
neous determination of methylmercury and inorganic
mercury in environmental and biological samples, J.

Anal. At. Spectrom. 9 1994 297

᎐302.

w x

A.M. Carro-Diaz, R.A. Lorenzo-Ferreira, R.J. Cela-Tor-
rijos, Speciaiton of oganomercurials in biological and
environmental samples by gas chromatography with mi-
crowave-induced plasma atomic emission detection, J.

Chromatogr. A 683 1994 245

᎐252.

w x

H. Hintelmann, R.D. Evans, J.Y. Villeneuve, Measure-
ment of mercury methylation in sediments by using
enriched stable mercury isotopes combined with meth-
ylmercury determination by gas chromatography-induc-
tively coupled plasma atomic emission spectrometry, J.

Anal. Atom. Spectrom. 10 1995 619

᎐624.

w x

H. Emteborg, E. Bjorklund, F. Odman, L. Karlsson, L.
Mathiasson, W. Frech, D.C. Baxter, Determination of
methylmercury in sediments using supercritical fluid ex-
traction and gas chromatography coupled with mi-
crowave-induced plasma atomic emission spectrometry,

Analyst 121 1996 19

᎐29.

w x

J.A. Stab, U.A.T. Brinkman, W.P. Cofino, Validation of
the analysis of organotin compounds in biological tissues
using alkylation

and gas

chromatography, Appl.

Organomet. Chem. 8 1994 577

᎐585.

w x

J. Szpunar, V.O. Schmitt, R. Lobinski, J.L. Monod,
Rapid speciation of butyltin compounds in sediments
and biomaterials by capillary gas chromatography

ᎏ

microwave-induced plasma atomic emission spectrome-
try after microwave-assisted leaching

rdigestion, J. Anal.

Atom. Spectrom. 11 1996 193

᎐199.

w x

J.M. Ombaba, E.F. Barry, Determination of organotin
species by capillary gas chromatography with alternating

current plasma emission detection, J. Chromatogr. 598
Ž

1992 97

᎐103.

w x

Y.K. Chau, F. Yang, R.J. Magurie, Improvement of
extraction recovery for monobutyltin species from sedi-

ment, Anal. Chim. Acta 320 1996 165

᎐169.

w x

M. Ceulemans, F.C. Adams, Evaluation of sample
preparation methods for organotin speciation analysis in
sediments

ᎏ focus on monobutyltin extraction, Anal.

Chim. Acta 317 1995 161

᎐170.

w x

R. Lobinski, F.C. Adams, Speciation analysis by gas
chromatography with plasma source spectrometric de-

tection, Spectrochim. Acta B 52 1997 1865

᎐1903.

w x

V. Camel, Microwave-assisted solvent extraction of envi-

ronmental samples, Trends Anal. Chem. 19

2000

229

᎐248.

w x

J.M. Bayona, Organometallic speciation by combining
aqueous phase derivatization with SPME-GC-FPD-MS,

in: J. Pawliszyn Ed. , Applications of Solid-Phase Mi-
croextraction, Royal Society of Chemistry, Cambridge,
1999, pp. 284

᎐295.

w x

L. Moens, T. De Smaele, R. Dams, P.V. Den Broeck, P.
Sandra, Sensitive, simultaneous

determination

organomercury, -lead, and

-tin

compounds

with

headspace solid phase microextraction capillary gas
chromatography combined with inductively coupled

plasma mass spectrometry, Anal. Chem. 69

1997

1604

᎐1611.

w x

X. Yu, J. Pawliszyn, Speciation of alkyllead and inor-
ganic lead by derivatization with deuterium-labeled
sodium tetraethylborate and SPME-GC

rMS, Anal.

Chem. 72 2000 1788

᎐1792.

w x

S. Aguerre, C. Bancon-Montigny, G. Lespes, M. Potin-

Gautier, Solid-phase microextraction

SPME : a new

procedure for the control of butyl- and phenyltin pollu-

tion in the environment by GC-FPD, Analyst 125 2000
263

᎐268.

w x

B. He, G.B. Jiang, Z.M. Ni, Determination of meth-
ylmercury in biological samples and sediments by capil-
lary gas chromatography coupled with atomic absorption
spectrometry after hydride derivatization and solid phase

microextraction, J. Anal. At. Spectrom. 13

1998

1141

᎐1144.

w x

B. He, G. Jiang, Analysis of organomercuric species in
soils from orchards and wheat fields by capillary gas
chromatography on-line coupled with atomic absorption
spectrometry after in situ hydride generation and
headspace solid phase microextraction, Fresenius J.

Anal. Chem. 365 1999 615

᎐618.

w x

Z. Mester, J. Lam, R. Sturgeon, Sampling and determi-
nation of metal hydrides by solid phase microextraction
thermal desorption inductively coupled plasma mass

spectrometry, J. Anal. At. Spectrom. 15

2000

1461

᎐1465.

w x

M. Guidotti, G Ravaioli, M. Vitali, Selective determina-
tion of Se

and Se

using SPME and GC

rMS, J.

High Resol. Chromatogr. 22 1999 414

᎐416.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

259

w x

Z. Mester, J. Pawliszyn, Speciation of dimethylarsinic
acid and monomethylarsonic acid by solid-phase mi-
croextraction-gas chromatography-ion trap mass spec-

trometry, J. Cromatogr. A 873 2000 129

᎐135.

w x

Z. Mester, G. Vitanyi, R. Morabito, P. Fodor, Speciation

of dimethylarsinic acid and monomethylarsonic acid by
gas chromatography-mass spectrometry, J. Chromatogr.

832 1999 183

᎐190.

w x

T. Gorecki, J. Pawliszyn, Determination of tetraethyl-
lead and inorganic lead in water by solid phase microex-

traction gas chromatography, Anal. Chem. 68 1996
3008

᎐3014.

w x

J.P. Snell, W. Frech, Y. Thomassen, Performance im-
provements in the determination of mercury species in
natural gas condensate using an on-line amalgamation
trap or solid-phase micro-extraction with capillary gas
chromatography-microwave-induced

plasma

atomic

emission spectrometry, Analyst 121 1996 1055

᎐1060.

w x

Z. Mester, J. Lam, R. Sturgeon, J. Pawliszyn, Determi-
nation of methylmercury by solid phase microextraction
ᎏ inductively coupled plasma mass spectrometry; a new
sample introduction method for volatile metal species, J.

Anal. At. Spectrom 15 2000 837

᎐842.

w x

E.O. Out, J. Pawliszyn, Solid phase micro-extraction of

metal ions, Mikrochim. Acta 112 1993 41

᎐46.

w x

C. Jia, Y. Luo, J. Pawliszyn, Solid phase microextraction
combined with HPLC for determination of metal ions
using crown ether as selective extracting reagent, J.

Microcolumn Sep. 10 1998 167

᎐173.

w x

A.A. Boyd-Boland, SPME-HPLC of environmental pol-

lutants, in: J. Pawliszyn Ed. , Applications of Solid-
Phase Microextraction, Royal Society of Chemistry,
Cambridge, 1999, pp. 327

᎐332.

w x

T.P. Gbatu, K.L Sutton, J.A. Caruso, Development of
new SPME fibers by sol

᎐gel technology for SPME-HPLC

determination of organometals, Anal. Chim. Acta 402
Ž

1999 67

᎐79.

w x

F. Guo, T. Gorecki, D. Irish, J. Pawliszyn, Solid-phase
microextraction combined with electrochemistry, Anal.

Commun. 33 1996 361

᎐364.

w x

T.P. Gbatu, O. Ceylan, K.L Sutton, J.F. Rubinson, A.
Galal, J.A. Caruso, H.B. Mark, Electrochemical control
of solid phase microextraction using unique conducting

polymer-coated fibers, Anal. Commun. 36 1999 205.

w x

J. Wu, X. Yu, H. Lord, J. Pawliszyn, Solid phase mi-
croextraction of inorganic anions based on polypyrrole

film, Analyst 125 2000 391

᎐394.

w x

Z. Mester, H. Lord, J. Pawliszyn, Speciation of trimeth-
yllead and triethyllead by in-tube solid phase microex-
traction high-performance liquid chromatography elec-
trospray ionization mass spectrometry, J. Anal. At. Spec-

trom 15 2000 595

᎐600.

w x

Z. Mester, H. Lord, J. Pawliszyn, Speciation of se-
lenomethionine and selenoethionine by in-tube solid
phase microextraction high-performance liquid chro-
matography electrospray ionization mass spectrometry,

J. Cromatogr. A submitted .

w x

W.H.J. Vaes, E.U. Ramos, H.J.M. Verhaar, W. Seinen,
J.L.M. Hermens, Measurement of the free concentra-
tion using solid-phase microextraction: binding to pro-

tein, Anal. Chem. 68 1996 4463

᎐4467.

w x

J. Poerschmann, F.D. Kopinke, J. Pawliszyn, Solid phase
microextraction to study the sorption of organotin com-
pounds onto particulate and dissolved humic organic

matter, Environ. Sci. Technol. 31 1997 3629

᎐3636.

w x

C.M. Barshick, S.A. Barshick, P.F. Britt, D.A. Lake,
M.A. Vance, E.B. Walsh, Development of a technique
for the analysis of inorganic mercury salts in soils by gas
chromatography mass spectrometry, Int. J. Mass Spec-

trom. 178 1998 31

᎐41.

w x

C.M. Barshick, S.A. Barshick, E.B. Walsh, M.A. Vance,
P.F. Britt, Application of isotope dilution to ion trap gas
chromatography

rmass spectrometry, Anal. Chem. 71

1999 483

᎐488.

w x

C.M. Barshick, S.A. Barshick, M.L. Mohill, P.F. Britt,
D.H. Smith, Elemental and organometallic analyses of
soil using glow discharge mass spectrometry and gas
chromatography

rmass spectrometry, Rapid Commun.

Mass Spectrom. 10 1996 341

᎐346.

w x

T. Gorecki, M. Belkin, J. Caruso, J. Pawliszyn, Solid
phase microextraction as sample introduction technique
for radio frequency glow discharge mass spectrometry,

Anal. Commun. 34 1997 275

᎐277.

w x

B. Rosenkranz, J. Bettmer, W. Buscher, C. Breer, K.
Cammann, The behaviour of different organometallic

compounds in the presence of inorganic mercury II :
transalkylation of mercury species and their analysis by
the GC-MIP-PED system, Appl. Organometal. Chem.

11 1997 721

᎐725.

w x

Z. Mester, J. Pawliszyn, Electrospray mass spectrometry
of trimethyllead and triethyllead with in-tube solid phase
microextraction sample introduction, Rapid Commun.

Mass Spectrom. 13 1999 199

᎐203.

w x

C. Haberhauer-Troyer, M. Crnoja, E. Rosenberg, M.
Grasserbauer, Surface characterization of commercial
fibers for solid-phase microextraction and related prob-
lems in their application, Fresenius J. Anal. Chem. 366
Ž

2000 329

᎐331.

w x

D. Amouroux, E. Tessier, C. Pecheyran, O.F.X. Donard,

Sampling and probing volatile metal loid

species in

natural waters by in-situ purge and cryogenic trapping
followed by gas chromatography and inductively coupled

plasma mass spectrometry

P-CT-GC-ICP-MS , Anal.

Chim. Acta 337 1998 241

᎐254.

w x

C.M. Tseng, A. de Diego, H. Pinaly, D. Amouroux,
O.F.X. Donard, Cryofocusing coupled to atomic absorp-
tion spectrometry for rapid and simple mercury specia-
tion in environmental matrices, J. Anal. At. Spectrom.

13 1998 755

᎐764.

w x

P.A. Martos, J. Pawliszyn, Time-weighted average sam-
pling with solid-phase microextraction device: implica-
tions for enhanced personal exposure monitoring to

airborne pollutants, Anal. Chem. 71 1999 1513

᎐1520.

(

)

Z. Mester et al.

rSpectrochimica Acta Part B: Atomic Spectroscopy 56 2001 233᎐260

260

w x

Supelco Data Sheet No. T713019A, Instructions for the
Supelco solid phase microextraction fiber holder for
manual use, 1998.

w x

X. Yu, H. Yuan, T. Gorecki, J. Pawliszyn, Determina-
tion of lead in blood and urine by SPME

rGC, Anal.

Chem. 71 1999 2998

᎐3002.

w x

Y. Cai, S. Monsalud, K.G. Furton, R. Jaffe, R.D. Jones,
Determination of methylmercury in fish and aqueous
samples using solid-phase microextraction followed by
gas chromatography atomic fluorescence spectrometry,

Appl. Organometal. Chem. 12 1998 565

᎐569.

w x

S. Mothes, R. Wenrich, Solid-phase microextraction and
GC-MIP-AED for the speciation analysis of organomer-

cury compounds, J. High Resol. Chromatogr. 22 1999
181

᎐182.

w x

Y. Morcillo, Y. Cai, J.M. Bayona, Rapid determination
of methyltin compounds in aqueous samples using solid
phase microextraction and capillary gas chromatography
following in-situ derivatization with sodium tetraethyl-

borate, J. High Resol. Chromatogr. 18 1995 767

᎐770.

w x

G. Lespes, V. Desauziers, C. Montigny, M. Potin-Gau-
tier, Optimization of solid-phase microextraction for the
speciation of butyl- and phenyltins using experimental

designs, J. Cromatogr. A 826 1998 67

᎐76.

w x

E. Milan, J. Pawliszyn, Determination of butyltin species
in water and sediments by solid-phase microextraction-
gas chromatography

ᎏ flame ionization detection, J.

Chromatogr. A 873 2000 63

᎐71.

w x

T. De Smaele, L. Moens, P. Sandra, R. Dams, Determi-
nation of organometallic compounds in surface water
and

sediment

samples

with

SPME-CGC-ICPMS,

Mikrochim. Acta 130 1999 241

᎐251.

w x

L. Dunemann, H. Hajimiragha, J. Begerow, Simultane-

ous determination of Hg II and alkylated Hg, Pb, and
Sn species in human body fluids using SPME-GC

rMS-

MS, Fresenius J. Anal. Chem. 363 1999 466

᎐468.

w x

S. Tutschku, S. Mothes, R. Wennrich, Preconcentration
and determination of Sn- and Pb-organic species in
environmental samples by SPME and GC-AED, Frese-

nius J. Anal.Chem. 354 1996 587

᎐591.

w x

Y. Cai, J.M. Bayona, Determination of methylmercury
in fish and river water samples using in situ sodium
tetraethylborate derivatization following by solid-phase
microextraction and gas chromatography-mass spec-

trometry, J. Chromatogr. A 696 1995 113

᎐122.

w x

T. De Smaele, L. Moens, R. Dams, P. Sandra, Metal

speciation by SPME-CGC-ICPMS, in: J. Pawliszyn Ed. ,
Applications of Solid-Phase Microextraction, Royal So-
ciety of Chemistry, Cambridge, 1999, pp. 296

᎐310.

w x

M. Guidotti, M. Vitali, Determination of urinary mer-
cury and methylmercury by solid phase microextraction

and GC

rMS, J. High Resol. Chromatogr. 21 1998

665

᎐666.

w x

J. Poerschmann, Z. Zhang, F.D. Kopinke, J. Pawliszyn,
Solid phase microextraction for determining the dis-
tribution of chemicals in aqueous matrices, Anal. Chem.

69 1997 597

᎐600.

w x

P.A. Martos, J. Pawliszyn, Calibration of solid phase
microextraction for air analyses based on physical
chemical properties of the coating, Anal. Chem. 69
Ž

1997 206

᎐215.

Wyszukiwarka

Podobne podstrony:
Solid phase microextraction as a clean up and preconcentrati
Solid phase microextraction a promising technique for sample
Solid phase microextraction for the detection of termite cut
Solid phase microextraction for herbicide determination in
Headspace solid phase microextraction for the determination
Trends in solid phase microextraction for
Solid phase microextraction for the analysis of biological s
Application of Solid Phase Microextraction Gas Chromatograp
Solid Phase Microextraction Analyses of Flavor Compounds in
Headspace solid phase microextraction profiling of volatile
A Practical Guide to Quantitation with Solid Phase Microextr
Vinyl chloride analysis with Solid Phase Microextraction
Solid phase microextration in biomedical analysis
Solid phase microextraction to concentrate volatile products
Kinetics of solid phase extraction and solid phase microextr
Solid phase microextraction in pesticide residue analysis
Optimisation of solid phase microextraction of volatiles
Applications of solid phase microextraction to
Comparison of Different Fibers in the Solid Phase Microextra

więcej podobnych podstron