For 25 mL of 0.144 g/L aqueous solution in a
initial rapid rise and equilibrium is reduced by
40-mL vial, the predicted mass of white phospho-
agitation. Based on this model and the physical
rus sorbed by the SPME phase is 97 pg. We found
properties of white phosphorus (Kow = 1200, KH =
2100 atm-cm3/mole), equilibration time for white
approximately 35 pg, or about one-third of the pre-
dicted mass. Assuming that K2 is accurately pre-
phosphorus should be only a few minutes for well-
dicted from KH, the amount of mass sorbed corre-
agitated samples. Even though equilibrium may
sponds to a K of 410. The ratio between Kow and K
not be reached within a few minutes when the
equals 2.9, which is slightly outside the range re-
sample is not agitated, the mass absorbed should
ported for BTEX. Therefore the model predictions
approach the mass at equilibrium because of the
and experimental results are in fair agreement.
shallow slope of the extraction profile.
The model also predicts the effect of variable
To see if the elimination of the sonication sig-
volumes of headspace. Predicted mass sorbed in-
nificantly decreased the mass of white phospho-
creases as headspace volume shrinks, but the ef-
rus sorbed to the SPME phase, four replicate 0.144-
g/L aqueous standards were prepared in reagent-
fect is minimal for the size vials we used when the
vials are at least half full. This characteristic is very
grade water and extracted and analyzed in ran-
useful in a field laboratory since sample volumes
dom order by headspace SPME by three methods:
need to be measured with a precision of approxi-
five minutes of sonication, ten minutes of sonica-
mately 1 mL. Sample aliquots could be obtained
tion, and five minutes static. One sample was ex-
by gentle pouring into graduated vials, eliminat-
tracted statically for 20 minutes, then 80 minutes.
ing the need for volumetric pipets and associated
The results (Table 7) were compared by blocked
glassware washing and rinsing.
ANOVA (Table 8), and a significant difference was
found indicating that equilibrium was not reached
Agitation
during five minutes of static headspace SPME.
However, the difference was small for practical
ple hastens equilibrium (Arthur et al. 1992c). We
purposes, 30.6 pg vs. 34.0 pg (90% recovery). For
used a sonic bath (Motlagh and Pawliszyn 1993),
the one sample extracted statically for 20 minutes
but found that temperature had to be carefully con-
and 80 minutes, the masses detected were 31.5 and
trolled. Fluctuating temperatures decrease preci-
35.2 pg, respectively. Since sonication or lengthy
sion. Higher temperatures increase the analyte va-
static extractions resulted in only modest increas-
por phase concentration but decrease analyte sorp-
tion into the fiber (MacGillivray et al. 1994, Zhang
and Pawliszyn 1995) since sorption is an exother-
Table 7. Mass of white phosphorus (pg)
found by headspace SPME for samples
mic process (Arthur et al. 1992c). Stirring was not
extracted with sonicated and static aque-
tested, but has been used by previous investiga-
ous phases.
tors (MacGillivray et al. 1994, Zhang and Pawliszyn
1995). The most convenient and least labor-inten-
White phosphorus mass (pg)
sive setup is static extraction for five minutes
Rep
5 min sonic 10 min sonic 5 min static
(length of the GC run) at room temperature.
1
30.3
32.7
29.1
Zhang and Pawliszyn (1993a) predicted equili-
2
36.8
34.5
31.5
bration time for headspace SPME using a model
3
36.8
35.1
30.8
based on one-dimensional diffusion described by
4
31.2
33.6
31.0
Mean
33.8
34.0
30.6
Fick's second law. Equilibration time was con-
Std dev
3.52
1.06
1.02
trolled by the gas/water and SPME phase/water
% RSD
10.4%
3.1%
3.3%
partition coefficients (K2 and K1 described above).
Extraction time profiles of analytes by headspace
Table 8. Blocked ANOVA comparing means
SPME are characterized by a three-part curve: first,
of sonicated and static samples.
by a rapid increase in mass sorbed within the first
minute of exposure of the SPME phase to the head-
ss
df
MS
F
space, followed by a much slower increase, then
Total
12971.74
12
no change at equilibrium. The initial increase cor-
Correction factor
12899.92
1
responds to the sorption of analyte initially present
Between block
28.02
3
9.34
3.58
in the headspace, and the slower increase is due to
Between treatment
28.14
2
14.07
5.39
Error
15.66
6
2.61
the mass transfer of analytes from the aqueous
phase (Pawliszyn 1995). The transition between the
F0.95(2,6) = 5.14
8
Previous Page
