GC and GC-MS Determination of Fluoroacetic Acid - Journal of ...

Journal of Analytical Toxicology, Vol. 22, May/June 1998
GC and GC-MS Determination of Fluoroacetic
Acid and Phenoxy Acid Herbicides via Triphasal
Extractive Pentafluorobenzylation Using a
Polymer-Bound Phase-Transfer Catalyst
A. Miki 1,*, H. Tsuchihashi 1, and M. Yamashita 2
1Forensic Science Laboratory, Osaka Prefectural Police Headquarters, 3-18, 1-Chome, Hommachi, Chuo-ward, Osaka 541-0053,
Japan and 2Department of Molecular Science and Technology, Doshisha University, Kyotanabe, Kyoto 610-0321, Japan
I Abstrad I
A simple and sensitive gas chromatography and gas
chromatography-mass spectrometry (GC-MS) procedure has been
developed for fluoroacetic acid (FA) and phenoxy acid herbicides
(PAHs) via triphasal extractive pentafluorobenzylation. The
triphasal system consisted of an aqueous sample, the extraction
solvent toluene containing pentafluorobenzyl bromide as the
derivatization reagent, and polymer-bound tri-n-butyl-
methylphosphonium bromide as a phase-transfer catalyst.
FA spiked in beverages, such as orange juice and milk, was
extracted as its pentafluorobenzyl (PFB) derivative under moderate
conditions (i.e., at a pH value of 6.5 at 60oc). The detection limits
were 0.10-0.20 pg/mL by GC with eledron-capture detection
(GC-ECD), and 0.42-0.50 pg/mL by full-scan GC-MS. PAHs were
also detectable in the same manner within the detection limits of
0.05-0.10 pg/mL by GC-ECD and 0.13-0.25 pg/mL by full-scan
GC-MS. Urine and serum which both contained 2,4-
dichlorophenoxyacetic acid could also be analyzed by GC-MS
after the triphasal pentafluorobenzylation. The detection limit was
0.20 pg/mL in the full-scan mode and 10 ng/mL in the seleded
ion monitoring mode both for the urine and serum.
Introdudion
Sodium fluoroacetate (compound 1080) is well known for its
high toxicity and has been widely used for the control of vertebrate
pests. The LDso values of 0.06 mg/kg for dogs and 0.22 mg/kg for
sewer rats have reportedly been effective (1). Although its use is
restricted in most countries because of its extreme toxicity in
vertebrates, a certain number of cases caused by careless use or
malicious use as a poison have been reported in recent years (2).
Phenoxy acid herbicides (PAHs), including 2,4-dichlorophen-
oxyacetic acid (2,4-D), are among the most frequently used agro-
chemicals throughout the world (3--5). Although PAHs are not
often used for homicides, several suicide cases involving these
herbicides are reported each year in Japan (1,2). Therefore, the
*Author to whom correspondence should be addressed.
development of a convenient and accurate analytical method for
fluoroacetic acid (FA) and PAHs is of great importance in envi-
ronmental health and in forensic toxicology.
Various analytical methods have been published for FA,
which include ion-selective electrode (6), high-performance
liquid chromatography (HPLC) (7,8), and headspace gas chro-
matography (GC) (9). PAHs in aqueous samples can be also
analyzed by HPLC with simple sample pretreatment (10,11).
In addition to these techniques, GC analysis after extraction
and an appropriate derivatization is very practical method for
such carboxylate-type substances (12,13). Particularly in
forensic toxicology, gas chromatography-mass spectrometry
(GC-MS) is one of the preferred confirmatory methods for
detecting these substances (14,15). Such GC and GC-MS pro-
cedures have been well documented for FA (16,17) and also for
PAHs with various derivatization methods, including (cya-
noethyl) dimethylsilyl (18), chloroethyl (19), ethyl ester
(20,21), and pentafluorobenzyl (22-24)derivatives. There are,
however, some inconveniences associated with these proce-
dures, which r.equire somewhat laborious and time-con-
suming sample treatments. In addition, unfavorable
partitioning of the highly polar compounds into the extraction
solvent leads to losses in yield.
In our previous papers, a convenient sample preparation for
GC and GC-MS for aqueous substances of forensic interest
via liquid-liquid-solid-phase-transfer-catalyzed derivatization
was established. The analytes included dialkyl phosphates
(25,26), phenolates (27), carboxylates (27), and toxicologically
important inorganic anions (28). In these procedures, a
polymer-bound quaternary phosphonium salt was employed as
the phase-transfer catalyst (PTC), and pentafluorobenzyi bro-
mide (PFB-Br) was the derivatization reagent. A scheme of
this triphasal procedure is shown in Figure 1. The polymer-
bound PTC captures anionic analytes in the aqueous sample
because the catalyst has an ion-exchange resin structure. The
derivatization proceeds inside or on the surface of the resin,
which is impregnated with organic solvent containing PFB-Br.
The anions are highly reactive in the polymer because of the
small degree of hydration. The most remarkable advantage of
this triphasal procedure is that the extraction, derivatization,
Reproduction (photocopyin 8) of editorial content of this journal is prohibited without publisher's permission. 237

and concentration of the resultant derivatives into small-
volume organic layer are simultaneously achieved. Because
the catalyst is in the form of bonded solid particles, this pro-
cedure leads to easy layer separation and allows the injection
of the organic layer into a GC instrument after only a simple
cleanup.
The present work has aimed to apply this triphasal procedure
and facilitate the GC and GC-MS determination of FA and PAHs
in various samples. The PAHs tested here are listed in Table I. In
forensics, it is of great importance to detect poisons in trace
amounts found at the scene (e.g., drink residues) and in biolog-
ical samples of the victims. In order to efficiently determine
these chemicals in various beverage samples, appropriate sample
pretreatments were studied for each type of drink. The GC-MS
analysis of 2,4-D in urine and serum was also optimized to
ensure the applicability of the triphasal extractive pentafluo-
robenzylation to biofluid samples.
Experimental
Materials
Sodium salts of 2,4-D, 2-methyl-4-chlorophenoxyacetic
acid (MCPA), and 2-methyl-4-chlorophenoxybutylic acid
(MCPB) were supplied by Nissan Kagaku (Tokyo, Japan).
Potassium salt of 2-(2-methyl-4-chlorophenoxy)propionic
acid (MCPP) was provided by Kei-ai Kasei (Tokyo, Japan).
Sodium salt of FA was obtained from Otsuka Yakuhin (Tokyo,
Japan). PFB-Br (99%+), 1,3,5-tribromobenzene (TBB) (98%),
and 2,4,6-trimethylbenzoic acid (TMBA) (99%+) were pur-
chased from Aldrich (Milwaukee, WI). Tri-n-butylmethyl-
phosphonium bromide polymer bound (0.75 meq. Br-/g
[TB-0.751) was prepared from tri-n-butylmethylphosphonium
chloride polymer bound (0.78 meq. Cl-/g 200-400 mesh,
polystyrene-l% divinylbenzene copolymer, Fluka, Buchs,
Switzerland) according to a previously described manner
(26). Sodium linear alkylbenzene sulfonate (LAS) and fatty
238
F F F F
F~CH2--Br F~>-- CH2--O-- C-R
II
F F F F O
PFB-Br ~x,~.~ /Org.
~ @ _ _ ~ CH2_ p+ ~-~eluti~
adsorption ~~/'~ Br, B~,~.-..~
~ Aq.
R-C-O- M + Br-
II
O
Figure 1. Scheme of the triphasal extractive pentafluorobenzylation of carboxylates.
)ournal of Analytical Toxicology, Vol. 22, May/June 1998
acid ethanolamide (FAE) (1:1 type) were obtained from Kao
(Tokyo, Japan). Florisil (60-100 mesh, P.R. grade) was pur-
chased from Wako (Osaka, Japan). Phosphate buffers (0.1M,
pH 4.5-9.2) were prepared by mixing 0.1M potassium dihy-
drogen phosphate and 0.1M disodium hydrogen phosphate.
All organic solvents and inorganic reagents were of analytical-
reagent grade. Deionized and distilled water was used
throughout.
The river water used for experiments was sampled from
Higashi-yokobori river at Chuo-ward (Osaka City, Japan). The
pH and chemical oxygen demand (COD) values were 6.8 and 11
rag/L, respectively.
Preparation of standards
Stock standard solutions of each analyte (1.0 mg/mL aqueous
solutions, adjusted to neutral pH with 0.1M sodium hydroxide
and/or phosphoric acid) were prepared weekly and stored at
0-5~ An aqueous standard solution containing FA, 2,4-D,
MCPA, MCPB, and MCPP at 8.0 IJg/mL each was prepared, and
2.5-mL aliquots were used for the characterization of each
derivative and optimization of the procedure. Spiked samples
were prepared by the addition of known amounts of appropriately
diluted standard solutions to blank samples.
Triphasal extractive pentafluorobenzylation procedure
River water and beverage samples. Before analysis, the
sample pH was adjusted to approximately 7 with diluted
sodium hydroxide and phosphoric acid. A 15-rag amount of the
polymer-bound catalyst TB-0.75 was placed in a 20-mL, Teflon-
sealed, stoppered glass test-tube. An aqueous sample (2.5 mL
for beverages and 8.0 mL for river water samples), 0.1M phos-
phate buffer (pH 6.5; 20% volume of the sample; i.e., 0.5 or 2.0
mL), and toluene (0.4 mL) containing 3.0 pL (6.0 IJL, if nec-
essary) of PFB-Br were successively added into the test-tube.
The test-tubes were stoppered, sent to the test-tube evapo-
rator "Vapor-mix" (Yamato rikagaku, Tokyo, Japan), and vig-
orously shaken in a 60~ water bath (without evaporation).
After an appropriate reaction time, the reaction was inhibited
by the addition of 1.0M phosphoric acid (1.0 mL), and I mL of
n-hexane containing TBB (I.S.) at 2-pg/mL
concentration was added. The organic layer
was then withdrawn and filtered and dried by
passage through the disposable Pasteur glass
pipette plugged with a piece of cotton and
filled with 0.5 g of anhydrous sodium sulfide.
A 1.0-pL aliquot was injected into the
GC-ECD. After concentrating the extract to
0.2-0.3 mL, a 1.0-1JL aliquot was subjected to
GC-MS.
Biofluid samples. Serum and urine sam-
ples (0.5 mL, adjusted to pH 6-7) were each
diluted with 2.0 mL water and mixed with 0.5
mL of 1.0-1Jg/mL TMBA (I.S.) aqueous solu-
tion. Derivatization was performed in the
manner described here; however, no buffer
solution was added. After a 45-rain reaction at
60~ n-hexane (1 mL for urine, 2 mL for
serum) was added. The organic layers were

Journal of Analytical Toxicology, Vol. 22, May/June 1998
Table I. Phenoxy Acid Herbicides (PAHs) Examined
Abbreviation Herbicide Structure
2,4-D
MCPA
MCPB
MCPP
Q
s
.m
n
o
n-
1
0.5 84
0.5
9
33
2,4-Dichlorophenoxyacetic acid
2-Methyl-4-chlorophenoxyacetic acid
2-Methyl-4-chlorophenoxybutylic acid
2-(2-Methyl-4-chlorophenoxy)propionic acid
a@o(~-coo~
- c~-ia
L 258 M +
61 161
then separated (the mixtures were cen-
trifuged if necessary) and applied onto florisil
mini-columns. The columns were prepared
in disposable glass pipettes filled with 0.3 g of
florisil and overlaid with 0.3 g anhydrous
sodium sulfide. After washing the columns
with 1 mL of n-hexane, they were eluted with
1.5 mL of 30% ethyl acetate in toluene. The
eluates were then concentrated to approxi-
mately 200 pL with a gentle stream of
nitrogen, and 1.0-pL aliquots were injected
into the GC-MS.
Instrumentation
GC-ECD. A GC-14A GC (Shimadzu, Kyoto,
Japan) fitted with an electron-capture detector
(Shimadzu clean ECD-C9, purge-gas type) was
used. The detector temperature was set at
l [ 1"
, 9 I ,I ,I J _1 h J ~, ' J _ h.
100 200 300
m/z
111 145
175
[]
F175 161
181 J 1 11 " ,,---'
F F lgT_J o el
J,l.,. ,. . . . .
L195
9 .I,
[
400 M*
100 2OO m/z 300 400
Figure 2. Electron impact mass spectra of the PFB derivatives of fluoroacetic acid (FA) (A) and 2,4-dichlorophenoxyacetic acid (2,4-D) (B).
A
B
239

300~ and the purge gas was nitrogen at a flow rate of 35
m~min. A DB-17 capillary column (30 m x 0.32-ram i.d., 0.25-
pm film thickness, J&W Scientific, Folsom, CA) was used at a
split ratio of 30:1. The column temperature was programmed
from 90 to 270~ at 10~ The injector temperature was set
at 300~ Helium was used as the carrier gas at a flow-rate of 2.0
mL/min.
GC-MS. A GCMS-QP5050 MS (Shimadzu) was used in
combination with a DB-1 megabore column (30 m x 0.53-ram
i.d., 1.5-pm film thickness, J&W Scientific). Helium was used
as carrier gas at a flow rate of 6.5 mL/min in the splitless
mode. The column temperature was programmed from 70
to 270~ at 10~ All measurements were performed
with electron impact ionization. The ionization energy and
interface temperature were set at 70 eV and 250~ respec-
tively.
Results and Discussion
Confirmation of the derivatives
The corresponding PFB esters were identified as the deriva-
tives of each analyte by GC-MS. All derivatives distinctively gave
both M § and m/z 181 (i.e., [C6FsCH2 ]§ ions as well as their char-
acteristic fragments. The mass spectra of the PFB derivatives of
FA and 2,4-D are shown in Figure 2. Particularly for the confir-
mation of FA, PFB derivatization was found to be more advan-
tageous than its methyl ester, which gave a very small molecular
ion peak by electron impact ionization (9).
Optimization of the derivatization
Our fundamental studies (27) allowed us to set several param-
eters for the triphasal extractive pentafluorobenzylation. These
include the use of toluene as the extraction solvent, the use of
TB-0.75 as the catalyst, the appropriate amounts of PFB-Br and
TB-0.75, and the derivatization temperature as described in the
experimental section.
Sample pH. Our previous study (27) also indicated that the
optimum sample pH was 4.5 for the derivatization of acetate
anion. However, pH optimization was done for FA and 2,4-D
because the pH of sample solution is one of the most sensitive
factors in extractive derivatization. Figure 3 shows the results.
Although the individual optimum pH values were 5.9 and 7.0
for FA and 2,4-D, respectively, the pH for their simultaneous
derivatization was set at 6.5 in the experiments described
here.
Derivatization time. The time course of the triphasal PFB
derivatization of FA and PAHs is illustrated in Figure 4. The
most appropriate reaction time was found to be 45 min at
60~ for the simultaneous process. According to our basic
experiments about triphasal pentafluorobenzylation of var-
ious carboxylic acids (27), the derivatization yields of a variety
of carboxylic acids ranged from 93 to 102%. (The exceptions
were hippuric acid [HA] and L-tryptophan [Trp], which both
have an amino group. The conversions to mono-substituted
HA and di-substituted Trp were 81 and 75%, respectively).
240
Journal of Analytical Toxicology, Vol. 22, May/June 1998
Although we did not measure the derivatization yields of FA
and PAHs, the previous results (27) suggested that all of the
PAHs and FA were semiquantitatively derivatized after a 45-
rain reaction. A gas chromatogram obtained by the simulta-
neous triphasal pentafluorobenzylation of FA and PAHs is
shown in Figure 5A.
Quantitative GC analysis. The calibration graphs for FA
and 2,4-D by GC-ECD were constructed using stepwise diluted
aqueous standard solutions and TBB as the internal standard.
The graphs were linear over the ranges of 5,0-100 pg/mL (y =
-0.02 + 1.19x, r 2 = 0.997) for FAand 2.5-50 IJg/mL (y = -0.01
+ 0.67x, r ~ = 0.998) for 2,4-D. Incomplete conversions of FA
I)
!
it
0.5
, i ~ J ;
0 ~/ 5
pH
Figure 3. Effect of pH in the triphasal pentafluorobenzylation of FA and
2,4-D. The derivatization yields were determined after a reaction time of
20 min.
0.5 ~ 0--~_
o MOPP
A MCPA
o MCPB
9 2,4-D
9 FA
13-
e.-
I I I ,
0 20 40 60 80
Reaction time (rain)
Figure 4. Time course of the triphasal pentafluorobenzylation of FA and
PAHs.

Journal of Analytical Toxicology, Vol. 22, May/June 1998
were observed at concentrations lower than 5.0 pg/mL. This
was presumably due to the adsorption of FA to the glassware
or vaporization of this polar but volatile compound. The rel-
ative standard deviations (RSD) of the five consecutive mea-
surements for FA and 2,4-D at 8.0 IJg/mL
were 3.6 and 2.7%, respectively. The absolute
detection limits of underivatized FA and 2,4-
D by GC-flame-ionization detection were
300 and 200 ng, respectively, based on the
injection amount. PFB derivatization led to
10 to 20 times more sensitive detection
limits because of less polar and fairly volatile
properties. PFB derivatives are preferably
analyzed by GC-ECD, which takes advan-
tage of the electron-capturing moiety. The
absolute detection limits of around 0.05 ng
were given for the PFB derivatives of FA and
PAHs by GC-ECD. _.....u
Analysis of beverage samples
In order to inspect the applicability of this
triphasal procedure to practical forensics,
recoveries were measured for FA and PAl-Is in
spiked orange juice, milk, and the nutrient-
refreshing drink called Oronamine C | The gas
chromatograms obtained from the spiked juice
and milk are shown in Figures 5B and 6A,
respectively. Table II lists the recoveries of FA
and PAHs from the spiked beverages at 8.0-
pg/mL concentration. Each recovery was cal-
culated by comparison of the relative peak
intensity of an analyte in a spiked sample with
that of one in the standard aqueous solution.
The low recovery of FA from the orange juice
can be explained as follows: the sample con-
tains certain concentrations of various car-
boxylic acids such as tertalic acid. During the
derivatization, PFB-Br was presumably con-
sumed for the derivatization of those acids.
Therefore, FA, which is relatively polar and not
quickly derivatized, was not completely defiva-
tized. The conversion of FA improved when the
quantity of PFB-Br was doubled (i.e., 6 pL).
The lower recoveries from the milk sample
were probably due to protein binding. The use
of a doubled quantity of PFB-Br did not
improve those recoveries.
Effect of deproteinization. In order to
achieve better recoveries of FA and PAHs from
the milk, some preparative methods such as
deproteinization were compared. The results
are listed in Table III. Deproteinization with
acetone was effective for PAHs; however, it
decreased the recovery of FA, presumably
because of the loss of the FA bound to the pro-
tein. The gas chromatograms obtained from
the spiked milk sample without any pretreat-
ment and the one obtained after the depm-
l
i
i
teinization are shown in Figure 6.
Sample dilution (five times weaker) before the derivatization
increased the recoveries of FA and PAHs. This simple dilution
was particularly effective in unbinding a polar compound such
o~
' 8 ' 1'2
Retention time (min)
~u4~
Q
3E
A
B
JL......I
i i !
Figure 5. Gas chromatograms obtained from the triphasal pentafluorobenzylation of FA and PAHs
in diluted aqueous standard solution (A) and spiked orange juice (B) at 8.0 pglmL each.
Table II. Recoveries of FA* and PAHs from Spiked-Beverages Samples
Analyle*
Amount
PFB-Br
Recovery (%)
(pL) Orange juice Oronamine C ~ Milk
FA 3 25 59 48
16
6 56 62 47
2,4-D 3 70 90 37
6 81 89 40
MCPA 3 75 98 39
6 85 96 42
MCPB 3 94 100 32
6 97 98 37
MCPP 3 92 96 55
6 95 96 58
" Abbreviations: FA, fluoroacetic acid; PAHs, phenoxy acid herbicides; PFB, pentafluorobenzyl;
2,4-D, 2,4-dichlorophenoxyacetic acid; MCPA, 2-methyl-4-chlorophenoxyacetic acid; MCPB,
2-methyl-4-chlorophenoxybutylic acid; MCPP, 2-(2-methyl-4-chlorophenoxy) propionic acid.
The concentrations were 8.0 pg/rnL
241

as FA. As we described in previous papers (26,27), the most
remarkable advantage of the polymer-bound PTC is the function
of capturing anions in a large-volume aqueous sample. FA and
0
<
IL
i i i i
8 12
Retention time (rain)
e4 9
2L
Journal of Analytical Toxicology, Vol. 22, May/June 1998
PAils could be efficiently concentrated into the organic layer in
the form of their PFB derivatives. This seemed most helpful for
the microanalysis of anionic analytes in peptide-rich samples like
milk and serum.
The detection limits of FA and PAHs in the
spiked samples by both GC-ECD and GC-MS
were measured by a combination of the most
A appropriate pretreatment and derivatization
t
i i i
16
Figure 6, Gas chromatograms obtained from the spiked milk sample with 8.0 IJg/mL of FA and
PAHs. Without any pretreatment (A) and after deproteinization with acetone (B).
Table III. Comparison of Pretreatment for the Triphasa[
Pentafluorobenzylation of FA* and PAHs in the Spiked-Milk Sample
Analyte t No pretreatment
Recovery (%)
Deproteinization* Dilutiont
FA 48 28
2, 4-D 37 63
MCPA 39 65 77
MCPB 32 78 86
MCPP 55 88 92
* Abbreviations: FA, fluoroacetic acid; PAHs, phenoxy acid herbicides; 2,4-D, 2,4-dichlorophenoxyacetic
acid; MCPA, 2-methyl-4-chlorophenoxyacetic acid; MCPB, 2-methyl-4-chlorophenoxybutylic acid; MCPP,
2 (2-methyl-4-chlorophenoxy) propionic acid.
* The concentrations were 8.0 p~mL.
* Deproteinization was done prior to the triphasal procedure as follows: acetone (2.0mL) was added to a
sample (2.5 mL), followed by vortex mixing and storage for 1 h at 4~ to ensure complete precipitation. After
centrifugation, a clear supernatant was separated. Acetone in the supernatant was then evaporated with a
gentle stream of nitrogen at 50~
A watered-down dilution (five times weaker) was made before the triphasal procedure.
242
B
conditions for each drink. The results are listed
in Table IV.
Effect of surface-active agents. The
extractive derivatization employed in this
study is a triphasal process, and polymer-
bound quaternary salt PTCs act as cationic
surface-active reagents. It is, therefore, of
great importance to examine the effects of
coexisting surface-active agents in the pro-
cedure. The authors tested the effects of
sodium linear alkylbenzene sulfonate (LAS)
as an anionic surface-active agent and fatty
acide ethanolamide (FAE) as a neutral sur-
face-active agent. Both are among the most
widely used surface-active agents, not only
in various industries, but also as major
ingredients in domestic detergents. In
domestic detergents, these two types of
agents are usually compounded at a ratio of
approximately 1:1. The samples used for the
experiment were the river water samples
spiked with each of the agrochemicals at a
concentration of 1.25 pg/mL and with LAS
or FAE or both at concentrations of 0-1000
IJg/mL. The results are listed in Table V.
Although the presence of LAS gave no
noticeable effects, the coexistence of FAE
at 1000 pg/mL obviously increased the
recovery of FA. This was possibly because
the presence of neutral surface-active agent
63 FAE led to a smoother phase transfer and
to a quicker derivatization of the polar
anion. The gas chromatograms obtained
76 from the spiked river water samples with
and without these surface-active agents are
compared in Figure 7.
Analysis of serum and urine samples
Generally, ingested FA is rapidly converted
to fluorocitrate, a potent inhibitor of aconi-
tase; therefore, the detection of FA in biological
tissues is impractical (1). According to
Allender (17), the concentration of FA in
stomach/gizzard contents of a magpie that had
ingested bait containing FA was quite low
(approximately 1 pg/g). Kancir et al. (5), how-
ever, reported that 82.4% of ingested 2,4-D
was excreted in urine, and only 12.8% was ex-
creted as conjugates.
In terms of biological samples, the sample

Journal of Analytical Toxicology, Vol. 22, May/June 1998
matrix often prevents promising methods moving from the out-
line stage to working procedures. Therefore, 2,4-D in serum
and urine was analyzed to ensure the applicability of the
triphasal procedure to biofluid samples. The analysis was carried
out by GC-MS in order to enhance the selec-
tivity for detection at low levels.
As the 2,4-D level in authentic biofluid sam-
pies may range very widely, the GC-MS was
performed both in the full-scan and selected
ion monitoring modes. Although PAH ana-
logues, such as 2,4,5-T (24), are commonly
used as the internal standard, we adopted the
aromatic carboxylic acid TMBA to avoid dupli-
cation of an authentic ingredient of forensic
interest. TMBA, which behaved almost identi-
cally to 2,4-D during the derivatization and
cleanup, was added before the analysis. Figure
8 shows the mass chromatograms obtained
from the fortified serum and urine samples
with 2,4-D at 0.25 pg/mL. The analysis was
performed in the selected ion monitoring
mode. The additional cleanup of the resultant
organic layer by passage through a florisil
mini-column was effective in reducing matrix
interference. The calibration graphs were
drawn by measuring the peak areas at rn/z 400
for 2,4-D against that at m/e 163 for TMBA
(internal standard). The graphs were both
linear over the range of 2.5 Bg/mL to 50 ng/mL
(.r = -0.061 + 5.61x, r 2 = 0.996 for serum and
y = -0.101 + 5.39x, r ~ = 0.997 for urine). The
detection limits were 10 ng/mL in the selected
ion monitoring mode and 0.20 pg/mL in the
full-scan mode for both the serum and urine.
The recoveries and RSDs of the whole proce-
dure evaluated at a concentration of 1.0 pg/mL
were 85 and 6.2%, respectively, for the serum
and 92 and 4.9%, respectively, for the urine.
Conclusion
In recent years, liquid chromatographic-
mass spectrometric technology has been
increasingly used in the confirmation anal-
ysis of aqueous substances. However, GC-MS
still has advantages, such as easier operation
and lower instrument costs. GC-MS has
made a large contribution to practical routine
analyses. The informative mass spectra pro-
vided by GC-MS with an appropriate deriva-
tization are indispensable for confirmation
analysis, particularly in forensics. In order
to establish a convenient sample pretreat-
ment for GC and GC-MS analysis of car-
boxylic acid-type agrochemicals, triphasal
extractive pentafluombenzylation was applied
6
I
as the key step in the sample pretreatment. The polymer-bound
phosphonium phase-transfer catalyst TB-0.75 allowed us to
simultaneously perform the extraction, concentration, and
derivatization of carboxylate analytes in an aqueous sample.
e~ Q
~E
Retention time (min)
Figure 7. Gas chromatograms obtained from the spiked river water with 1.25 pg/mL of FA and
PAHs. No surface-adive agents were spiked (A). The surface-active agents LAS and FAE were added
at 1000 pg/mL each (B).
Table IV. Detection Limits of FA and PAHs in Spiked Samples by GC-ECD
and GC-MS
Detection limit (pg/mL)
Orange juice* Milk* River Water ~
Anal~e * GC-ECDi GC-/VlS iN GC-ECDS GC-/VlS a GC-ECDS GC-IVlS it
FA 0.10 0.42 0.20 0.50 0.013 0.033
2,4-D 0.07 0.20 0.10 0.25 0.007 0.033
MCPA 0.07 0.20 0.10 0.25 0.005 0.020
MCPB 0.05 0.15 0.07 0.15 0.005 0.020
MCPP 0.05 0.13 0.07 0.20 0.005 0.025
* Six microliters PFB-Br was used.
A watered-down dilution (five times weaker) was made prior to the triphasal procedure.
The sample volume was 8.0 mL.
S/N =3. Average for three repeated measurements.
II Detection limit of five major peaks including M + in the full scan mode; this is the average for
three repeated measurements.
A
B
243

This concentration effect was potent for analytes in a rela-
tively large volume of aqueous sample. Using these advan-
tages, the detection of polar chemicals was successfully
performed even in protein-rich samples, such as milk and
Table V. Influences of Surface-Active Agents in the Triphasal
Pentafluorobenzylation of FA* and PAHs in Spiked River Water
Recovery (%)
Concentrations of surface-active agents (mg/t)
Anal~e t LAS* 0 250 1000 0 1000
FAEt 0 250 0 1000 1000
FA 44 44 42 59 62
2,4-D 58 58 57 56 54
MCPA 59 59 55 60 55
MCPB 60 64 63 64 65
MCPP 62 60 58 65 65
* Abbreviations: FA, fluoroacetic acid; PAHs, phenoxy acid herbicides; LAS, sodium linear alkybenzene; FAE,
fatty acid ethanolamide; 2,4-D, 2,4-dichlorophenoxyacetic acid; MCPA, 2-methyl-4-chlorophenoxyacetic
acid; MCPB, 2-methyl-4-chlorophenoxybutylic acid; MCPP, 2-(2-methyl-4-chlorophenoxy) propionic acid.
+ The concentrations were 1.25 pg/mL.
* Sodium linear alkylbenzene sulfonate.
w Fatty acid ethanolamide.
a
A
,- 402x2
. . . . 400 x 2
"----'--'---~ 177 x 2
~ "-----'--'--'--- 175 x 2
~---------~ 145 x 2
~ 1 1 1 x 2
- 344Xl
~-----"---~ 163 x 1
12.2 ?2 Is 12.2
Retention time (rain)
244
Journal of Analytical Toxicology, Vol. 22, May/June 1998
serum, after simple dilution with water. The authors will con-
tinue to pursue other extensive applications of this technique
in the forensic chemistry and toxicology fields.
r
References
Figure 8. Mass chromatograms obtained from the serum (A) and urine (B) samples containing 2,4-D at 0.25 pg/mL.
1. Pharmaceutical Society of Japan. 5tandard
Methods of Chemical Analysis in Poisoning.
Nanzan-do, Tokyo, Japan, 1992, pp 383-388.
2. Japanese National Research Institute of Police
Science. Annual Case Reports of Drug and Toxic
Poisoning in Japan. Vols. 20-39, Japanese
National Research Institute of Police Science,
Tokyo, 1977-1996.
3. World Health Organization. 2,4-Dichlorophen-
oxyacetic acid (2,4-D). Environmental Health
Criteria, Geneva, Switzerland. W.H.O. Monogr.
Ser. 29:1-44 (1984).
4. J. Osterloh, M. Lotti, and S. Pond. Toxicological
studies in a fatal overdose of 2,4-D, MCPP and
chlorpyrifos. J. Anal ToxicoL 7:125-129 (1983).
5. C.B. Kancir, C. Andersen, and A.S. Olesen.
Marked hypocalcemia in a fatal poisoning with
chlorinated phenoxy acid derivatives. Olin. Tox-
icoL 26(3-4): 257-264 (1988).
6. G. Livanos and P.J. Milham. Fluoride ion-selec-
tive electrode determination of sodium monoflu-
oroacetate in meat baits and formulations.
J. Assoc. Off. Anal Chem. 67(1): 10-12 (1984).
7. D.M. Collins, J.P. Fawcett and C.G. Rammell.
Determination of sodium fluoroacetate (com-
pound 1080) in poison baits by HPLC. Bull.
Environ. Contam. Toxicol. 26:669-673 (1981 ).
L 9
i
.L
B
181 x 1
402 x 5
400 x 5
-177x5
175x5
145x5
111x5
344x2
163x2
15 1~.5
Retention time (min)

Journal of Analytical Toxicology, Vol. 22, May/June 1998
8. A.C. Ray, L.O. Post, and J.C. Reager. High pressure liquid chro-
matographic determination of sodium fluoroacetate (compound
1080) in canine gastric content. J. Assoc. ~ Anal. Chem. 64(1):
19-24 (1981).
9. M. Mori, H. Nakajima, and Y. Seto. Determination of fluoroacetate
in aqueous samples by headspace gas chromatography. J. Chro-
matogr. 736:229-234 (1996).
10. Th. Keller, G. Skopp, M. Wu, and R. Aderjan. Fatal overdose of
2,4-dichlorophenoxyacetic acid (2,4-D). Forensic 5cL Int. 65:13-18
(1994).
11. A. Balinova. Ion-pairing mechanism in the solid-phase extraction
and reverse-phase high-performance liquid chromatographic deter-
mination of acidic herbicides in water. J. Chromatogr. 728:319-324
(1996).
12. J.D. Kohli, R.N. Kuanna, B.N. Gupta, M.M. Dahr, J.S. Tandon, and
K.P. Sircar. Absorption and excretion of 2,4-dichlorophenoxyacetic
acid in man. Xenobiotica 4(2): 97-100 (1974).
13. G.E. Schulze, J.W. Blake, and J.A. Dougherty. The metabolic fate of
2,4-dichlorophenoxyacetic acid-n-butyl ester in the wister rat. Arch.
Toxicol. 57:231-236 (1985).
14. K. Pfleger, H.H. Mourer, and A. Weber. Mass Spectral and GC Data
of Drugs, Poisons, Pesticides, Pollutants and Their Metabolites, 2nd
ed. VCH Verlag Sgesellschaft, Weinheim, Germany, 1992.
15. A.C. Moffat, J.V. Jackson, M.S. Moss, and B. Widdop, Eds. Clark's Iso-
lation and Identification of Drugs, 2nd ed. The Pharmaceutical
Press, London, U.K., 1986.
16. I. Okuno and D.L. Meeker. Gas-liquid chromatographic determi-
nation of sodium fluoroacetate (compound 1080). J. Assoc. Off.
Anal Chem. 63:49-55 (1980).
17. W.J. Allender. Determination of sodium fluoroacetate (compound
1080) in biological tissues. J. Anal Toxicol. 14:45-49 (1990).
18. M.J. Bertrand, A.W. Ahmed, B. Sarrasin, and V.N. Mallet. Gas chro-
matographic and mass spectrometric determination of chlorophe-
noxy acids and related herbicides as their (cyanoethyl) dimethylsilyl
derivatives. AnaL Chem. 59:1302-1306 (1987).
19. H. Agemian and A.S.Y. Chau. Analysis of pesticide residues by
chemical derivatization. J. Assoc. Off. Anal Chem. 60:1070-1078
(1977).
20. C. S,~nchez-Brunete, A.I. Garda-Valc,~rcel, and J.L. Tadeo. Deter-
mination of residues of phenoxy acid herbicides in soil and cereals
by gas chromatography-ion trap detection. J. Chromatogr. 675:
213-218 (I 994).
21. M. Tatsuno, M. Katagi, and H. Tsuchihashi. Fused silica capillary
gas chromatography of phenoxy acid herbicides with ethyl esteri-
fication. Rep. Natl. Res. Inst. Police Sci. 43(4): 154-160 (1990).
22. H.D. Meiring, G. den Engelsman, and A.P.J.M. de Jong. Determi-
nation of polar pesticides by phase-transfer catalysed derivatization
and negative-ion chemical ionization gas chromatography-
mass spectrometry. J. Chromatogr. 644:357-365 (1993).
23. M. Vink and J.M. van der Poll. Gas chromatographic determination
of acid herbicides in surface water samples with electron-capture
detection and mass spectrometric confirmation. ]. Chromatogr.
733:361-366 (I 996).
24. G.H. Oliveira and J. Palermo-Neto. Toxicology of 2,4-dichlorophe-
noxyacetic acid (2,4-D) and its determination in serum and brain
tissue using gas chromatography-electron-capture detection.
J. Anal. To 19= 251-55 (1995).
25. A. Miki, H. Tsuchihashi, K. Ueda, and M. Yamashita. Extractive
derivatization of aqueous drugs using polymeric phase transfer
catalysts I. Jpn. J. ToxicoL Environ. Health 38:168-I 75 (I 992).
26. A. Miki, H. Tsuchihashi, K. Ueda, and M. Yamashita. Gas chro-
matographic determination and gas chromatographic-mass spec-
trometric determination of dialkyl phosphates via extractive
pentafluorobenzylation using a polymeric phase transfer catalysts.
J. Chromatogr. 718:383-389 (1995).
27. A. Miki, H. Tsuchihashi, H. Yamano, and M. Yamashita. Extractive
pentafluorobenzylation using a polymeric phase-transfer cata-
lyst--A convenient one-step pretreatment for gas chromatographic
analysis of anionic compounds. AnaL Chim. Acta 356:165-175
(1997).
28. A. Miki, H. Tsuchihashi, K. Ueda, H. Yamano, and M. Yamashita.
Gas chromatographic and gas chromatography/mass spectro-
metric determination of aqueous poisons via extractive pentaflu-
orobenzylation using polymeric phase-transfer catalyst.
Proceedings of the 30th International Meeting of the Interna-
tional Association of Forensic Toxicologists, Fukuoka, Japan,
1992, pp 419-424.
Manuscript received June 9, 1997;
revision received October 30, 1997.
245