Anal. Chem.

1994,66, 1495-1499

Determination of Alachlor and I t s Sulfonic Acid Metabolite

in Water by Solid-Phase Extraction and Enzyme-Linked
Immunosorbent Assay
D. S. Aga, E. M. Thurman, and M. L. Pomes
U.S. Geological Survey, 482 I Quail Crest Place, Lawrence, Kansas 66049
Solid-phase extraction (SPE) and enzyme-linked immunosorbent assay (ELISA) were combined for the trace analysis of
the herbicide alachlor and its major soil metabolite, ethanesulfonic acid (ESA). The anti-alachlor antibodycross-reacted
with ESA, which produced false-positive detections of alachlor
in water samples by immunoassayscreens. Alachlor and ESA
were isolated from water by SPE on a CISresin and eluted
sequentially with ethyl acetate and methanol. Alachlor is
soluble in ethyl acetate while the anionic ESA is not. Thus
ESA remained adsorbed on the CISresin and was eluted later
with methanol. The combination of SPE withELISAeffectively
separatedand quantifiedboth alachlor and ESA using the same
antibody for two ELISA methods. The general method may
have applicability for the separation of other herbicides and
their ionic metabolites. The SPE-ELISA method has a
detection limit of 0.01 pg/L for alachlor and 0.05 pg/L for
ESA, with a precisionof *lo%. Analyses of surface and ground
water samples were confirmed by gas chromatography/mass
spectrometry and high-performance Liquid chromatography
with photodiode-arraydetection. Results showed widespread
occurrence of ESA in surface and ground water of the
midwestern United States, with concentrations ranging from
<0.10 to >10 pg/L.
Solid-phase extraction (SPE) and .enzyme-linked immunosorbent assay (ELISA) are two analytical techniques that
recently have found wide application in environmental
chemistry. For example, ELISA has been applied to the
analysis of herbicide residues in soil and water.l-I3 It is a
(1) Van Emon, J. M.; Lopez-Avila, V. Anal. Chem. 1992, 64, 79A-88A.
(2) Vanderlaan, M.; Stanker, L.; Watkins, B. In Immuncussaysfor Trace Chemical
Analysis; Vanderlaan, M., Stanker, L. H., Watkins, B. E., Roberts, D. W.,
Eds.; ACS Symposium Series 45 1;American Chemical Society: Washington,
DC, 1990; pp 2-13.
(3) Hammock, B. D.; st al. In Immunochemlcal Methods for Enuimnmental
Analysis; Van Emon, J. M., Mumma, R. O., Eds.; ACS Symposium Series
442; American Chemical Society: Washington, DC, 1990; pp 112-139.
(4) Harrison, R. 0.; Gee, S. J.; Hammock, B. D. In Biotechnology in Crop
Protection; Hcdin, P. A., Menn, J. J., Hollingworth, R. M., Eds.; ACS
Symposium Series 379; American Chemical Society: Washington, DC, 1988;
pp 316-330.
( 5 ) Vanderlaan, M.; Watkins, B. E.; Stanker, L. Enuiron. Sei. Technol. 1988,22,
247-254.
(6) Lawruk, T. S.;Hottenstein, C. S.;Herzog, D. P.; Rubio, F. M. Bull. Enuiron.
Conlam. Toxicol. 1992, 48, 643-650.
(7) Thurman, E. M.; Meyer, M.; Pomes, M.; Perry, C. A.; Schwab, A. P. Anal.
Chem. 1990,62, 2043-2048.
(8) Leavitt, R. A.; Kells, J. J.; Bunkellmann, J. R.; Hollingworth, R. M. Bull.
Enuiron. Conram. Toxicol. 1991, 46, 22-29.
(9) Goh, K. S.;Hernandez, J.; Powell, S.J.; Garretion, C.; Troiano, J.; Ray, M.;
Grccne, C. D. Bull. Enuiron. Contam. Toxicol. 1991, 46, 30-36.
(10) Bushway. R. J.; Perkins, E.; Savage, S.A.; Lekousi, S.J.; Ferguson, B. S.Bull.
Enuiron. Conram. Toxicol. 1988, 40, 647-654.
(1 1) Schlaeppi, J. M.; Fory, W.; Ramsteiner, K. J . Agric. Food Chem. 1989, 37,
1532-1 538.
This article not subJect to U S . Copyright.
Pubilshed 1994 by the American Chemical Society

sensitive, fast, and cost-effective technique that can be

conducted both in the laboratory and in the field. However,
many of the applications of ELISA to complex environmental
matrices are limited to screening of samples due to the potential
bias of analysis toward false-positive results.12J3 Less commonly, false-negative results are observed when the detection
limit of the assay is reached. Some immunoassay performance
may be affected significantly, not only by cross-reacting
compounds but also by matrix components that interfere with
the assay detection system and the antibodylantigen interactions.
SPE is used extensively as a cleanup procedure for clinical
and environmental sample^.^"" While ELISA is used to
detect known substances by using the specificity of the
antibodies, SPE is used to separate unknown substances by
mechanisms similar to that of high-performance liquid
~hromatography.~
The complimentary features of these two
techniques may be combined to providea selective and sensitive
analytical method analogous to a chromatographic separation
where the SPE cartridge is the column and ELISA is the
detector. With properly designed SPE procedures, closely
related compoundscan be separated, and proper interpretation
of immunoassay results can be made. Moreover, because SPE
is also a preconcentration technique, the overall detection limit
of the SPE-ELISA method may be improved by several orders
of magnitude. SPE can be automated easily for water
analysis;18J9hence the reproducibility of immunoassay is not
compromised. In addition, cross-reactivity of a structurally
similar compound could become a positive aspect of ELISA
because the same antibody may be used to quantify the crossreacting compound after it has been concentrated and
separated from the specific analyte by SPE.
Alachlor (2-chloro-2,6-diethyl-N-(methoxymethyl)acetanilide), a herbicide widely used in the United States, and its
metabolite 2-[(2,6-diethylphenyl)(methoxymethyl)amino]S.J.; Horton, S.R.; Sharp, C. R.; Logusch, E. W. J.
Agric. Food Chem. 1990, 38, 159-163.
(13) Feng, P. C.; Wratten, S.J.; Logussch, E. W.; Horton, S.R.; Sharp, C . R. In
Immunochemical Methods for Environmental Analysis; Van Emon, J. M.,
Mumma, R. O., Eds; ACS Symposium Series 442; American Chemical
Society: Washington, DC, 1990; pp 18&192.
(14) Appllcations Bibliography Sample Preparation Producrs; Varian Sample
Preparation Products, Harbor City, CA, 1991.
(IS) Tippins, E. Nature 1988, 334, 273-274.
(16) McDonald, P. D. Waters SepPak Cartridge Applications Bibliography;
Millipore Corp., Milford, MA, 1991.
(17) Morris, 2.;Ruthann, K. Solid Phase Extraction for Sample Preparation; J.
T.Baker, Inc.: Phillipsburg, NJ, 1988.
(18) Castellani, W. J.;Lente, F.V.;Chou,D.J. Autom. Chem. 1990,12, 141-144.
(19) Meyer, M.T.;Mills, M.S.;Thurman, E. M.J. Chromatogr. 1993,629,55-59.
(12) Feng, P. C.; Wratten,

Analytical Chemistry, Vol. 66, No. 9, May 1, 1994

1495

2-oxoethanesulfonic acid (ESA)FO are important target

compounds for the development of a sensitive method that
combines SPE and ELISA. Alachlor has been classified as
a possible human carcinogen and its maximum contaminant
level (MCL) is established at 2.0 pg/L by the U.S.EPA.21
EPA's recent toxicological studies on FSA suggest that the
metabolite is not mutagenic and does not bioaccumulate or
undergo further metabolism at the levels commonly found in
surface and ground water.22 However, ESA needs to be
monitored constantly because it is relatively persistent and
highly mobile23and may occur widely in ground water. The
frequency of false positives observed in the ELISA screening
kits for alachlor has been attributed to the significant crossreactivity of the ESA metabolite toward the anti-alachlor
a n t i b ~ d y .The
~ ~specific
~ ~ ~ objectives of the research described
herein were to (1) measure cross-reactivity of two alachlor
metabolites toward anti-alachlor antibodies, (2)understand
the principles involved in separating parent herbicides and
ionic metabolites by SPE and couple SPE and ELISA for
analysis of alachlor and ESA, and (3) apply the method to
samples of surface and ground water and compare the results
with gas chromatography/mass spectrometry (GC/MS) and
high-performance liquid chromatography with photodiodearray detection (HPLC/PDA).

EXPERIMENTAL SECTION
Reagents. Methanol (Burdick and Jackson, Muskegon,
MI) and ethyl acetate (Fisher, Springfield, NJ) were HPLCgrade solvents. Alachlor and its metabolites were obtained
from Monsanto Agricultural Co. (St. Louis, MO). The SPE
cartridges that were used (SepPak from Waters-Millipore,
Milford, MA) contained 360 mg of 40-pm C18-bonded silica.
Phenanthrene-& (EPA, Cincinnati, OH) was used as the
external standard for GC/MS quantification. Metribuzin
(EPA PesticideChemical Repository, ResearchTrianglePark,
NC) was used as the extemal standard for HPLC quantification. Standard stock solutions were prepared in methanol.
Solid-Phase Extraction Procedure. The SPE procedure
was automated with a Millipore Workstation (Waters,
Milford, MA) as described previ~usly.~J~
(Brand names in
this paper are for identification purposes only and do not
constitute endorsement by the U.S.Geological Survey.) The
c
1
8 cartridges were washed sequentially with 2 mL of
methanol, 6 mL of ethyl acetate, 2mL of methanol, and 2 mL
of distilled water. A 100-mL aliquot of sample was passed
through the cartridge (Figure 1) at a flow rate of 10mL/min.
The cartridge was eluted with 3 mL of ethyl acetate, followed
by a transfer step to remove the ethyl acetate (top layer)
containing the alachlor from the residual water (bottom layer)
in the eluate. Then, the cartridge was eluted with methanol
to remove ESA, which was collected in a separate test tube.
Both ethyl acetate and methanol extracts were evaporated to
(20) Baker, D. B.; Bushway, R. J.; Adam, S. A.; Macomber, C. S. Emiron. Sci.
Technol. 1993,27,562-564.
(21) U.S.EnvironmentalProtection Agency, Drinkingwater regulationsand health
advisories. Office of Water, Washington, DC,1992.
(22) Klein, A. J. Monsanto Co., personal communication, August 18, 1993.
(23) Goolsby, D. A.; Battaglin, W. A.; Fallon, J. D.; Aga, D. S.; Kolpin, D. W.;
Thurman, E. M. Absrracrs ofrhe Technical Meering, U.S.Geological Survey
Toxic Substances Hydrology Program, Sept 1993; p 83.
(24) Macomber, C.; Bushway, R. J.; Perkins, L. B.; Baker, D.; Fan,T. S.;Ferguson,
B. S. J. Agric. Food Chem. 1992.40, 1450-1452.

1496 Ana!vticaIChemklry, Vd. 66, No. 9, May 1, 1994

AlPchlor and ESA metabolite

( 1 0 0 " sample)

#I
El'fluent to waste

SPE Cia cartridge

Elute with3 mL of ethyl

acetate (containsalachlor)

J.

a) Analysis by E L I S A evaporate
to dryness, then reconstitutcwih
I m ~of20
. s methanohater or

b) Analysis by GUMS: spike with

phenanhnmcdio and evaporate to
100 @

Elute w i h 3 mL of methanol
(contains ESA and OXA metabolilts)

a) Analysis by E L I S A evaporate to
dryness. men econstitute with 5 mL

o f d i a i l ' d water or

b) Analysis by HPLC spike with

meuibuzin and evaporateto dryness.
then reconstitute with 100 pL of
20/80 huffer/methanol mixture

Flguro 1. Flow chart for SPE-ELISA method.

dryness under nitrogen at 45 OC using a Turbovap (Zymark,

Palo Alto, CA). The ethyl acetate extracts (containing
alachlor) were reconstituted with 1 mL of 20% methanol/
water and the methanol extracts (containing ionic metabolites)
with 5 mL of distilled water for analysis by,ELISA.
The preparation of samples for GC/MS and HPLC
analyses was performed in a similar SPE procedure. The
ethyl acetate extracts were spiked with phenanthrene-dlo and
evaporated to about 100 pL for the analysis of alachlor by
GC/MS. The methanol extract for HPLC analysis was spiked
with metribuzin, evaporated to dryness, and then redissolved
in 100 pL of 10 mM phosphate buffer/methanol (20:80)
mixture.
ELISA Procedure. The cross-reactivitiesof the two ELISA
kits from different manufacturers toward acetanilide herbicides and metabolites were determined. The alachlor EnviroGard assay (Millipore, Bedford, MA) had antibodiescoated
on the wellsof the microtiter plate, whereas the alachlor RaPID
assay (Ohmicron Corp., Newtown, PA) had antibodies
covalently attached to magnetic particles. Solutions of each
compound were prepared at concentrations of 0 , l .O, 5.0,10,
100,and 1000pg/L in water, and each solution was analyzed
in duplicate using both ELISA kits. For EnviroGard ELISA,
80 pL of sample and 80 pL of hapten-enzyme conjugate were
mixed in the well, and the mixture was incubated at 30 OC
in an orbital shaker (200rpm). After 1 h, the plate was rinsed
five times with deionized water, and excess water was removed.
Color reagent (160 pL, of 1:l mixture of 0.02% hydrogen
peroxide and 3,3',5,5'-tetramethylbenzidine) was added. The
color was allowed to develop for 30min, and then stop solution
(40 pL, 2 M sulfuric acid) was added. Optical densities were
read on a Vmax microplate reader using a Softmax software
(Molecular Devices, Menlo Park, CA).
For the alachlor RaPID assay, 250pL of samplewas mixed
with 250 pL of hapten-nzyme conjugate and 500 pL of
paramagnetic particles. This mixture was incubated for 30
min at r c " temperature and then a strong magnetic field
was applied to separate unbound analytes from the antibodybound analytes. The e x a s reagent was removed. A 500-pL
aliquot of color reagent was added next, and the mixture was

incubated for 20min at room temperature,followed by addition

of the stop solution (500 pL). The optical densities were read
using the RPA-I RaPID photometric analyzer (Ohmicron).
For the analysis of alachlor and ESA in actual water
samples, the alachlor RaPID assay kit was used. The
concentrations of alachlor were calculated from Ln/LogitB
transformed data using alachlor standards of 0, 0.1, 1.0, and
5.0 pg/L. ESA analysis was performed using the alachlor
RaPID assay, with the same procedure described for alachlor
except that the kit was calibrated with ESA standards prepared
in distilled water at concentrationsof 0 , l .O,5.0, and 20 pg/L.
The accuracy and precision of the SPE-ELISA method using
the RaPID assay kit was evaluated for several alachlor and
ESA mixtures. Analyses of alachlor and ESA were done in
duplicate.
GUMS Analysis. GC/MS analysis of alachlor was
performed on a Hewlett-Packard Model 5890A GC (Palo
Alto, CA) and 5970A mass-selective detector (MSD). Operating conditions were identical to those described by
Thurman et ala7 The detector was operated in a selected-ion
monitoring (SIM) mode, and confirmation was based on the
presence of the molecular ion peak, two confirmingions (with
area counts 4=20%), and a retention/time match of f0.2%
relative to phenanthrene-&,. The ratio of the areas of the
base-peak ion of alachlor to the 188 amu ion of phenanthrenedlo was used to construct the calibration curve. A 12-m,HP- 1
capillary column made of cross-linked methylsilicone with a
film thickness of 0.33 pm and 0.2 mm i.d. (Hewlett-Packard)
was used for separation. Helium was used as the carrier gas
with a flow rate of 1 mL/min and a head pressure of 35 kPa.
The samples were injected in the splitless mode by an
autoinjector. The column temperature was held at 60 OC for
1 min, then increased at 6 OC/min to 250 OC, and held at this
temperature for 10 min. Injector temperature was 280 OC.
HPLC/PDA Analysis. The HPLC analysisof the methanol
extracts for the confirmation of ESA was performed in a HP
Model 1090 series I1 liquid chromatograph with photodiodearray detector (Hewlett-Packard). The HPLC was equipped
with a 2.1 mm X 100.0 mm narrow-bore, reversed-phase
column packed with 5-pm Hypersil ODS (Hewlett-Packard).
The mobile phase consisted of 38.5% HPLC-grade methanol
and 61.5% 10 mM Na2HP04 (pH 7.0 buffer) prepared in
Nanopure water. The flow rate of the mobile phase was 1.200
mL/min. The sample injection volume was 90 pL. ESA was
monitored at a wavelength of 200 nm with a 4-nm bandwidth.
The reference wavelength was set at 450 nm with a 80-nm
bandwidth. To confirm the identity of the ESA, the ultraviolet
(UV) spectra were scanned from 190to 400 nm and the spectra
matched to standard spectra in an automated library search.
RESULTS AND DISCUSSION
Acetanilide Cross-Reactivity and E L S A Development. The
cross-reactivities of the antibodies in the RaPID (magnetic
particle-based) and EnviroGard (microtiter plate-based)
alachlor-ELISA with other acetanilides and alachlor metabolites were measured (Figure 2). Cross-reactivities were
expressed as the IC50, which is the concentration of the
compound that causes 50% inhibition of absorbance, and as
the least detectable dose (LDD), which is the concentration
at 90% inhibition of absorbance. Both ELISAs were most

cHiocH%

CHiOC\.

cocHIcl

-$J-

""$JCh"

ElhmesuIfonlc add (ESA)wt.bdite

N d o r

IC.%=

COCHISOIH

0.89 ,0.75

LDLk0.19,0.12
15.4, 1.7

LDD=0.07,0.06

W"

m c "
No aou-dvily 4.000 Wl.

Figure 2. Cross-reactivities of antibodies with chloroacetaniiide

herbicides and alechior metabolites. Cross-reactivity Is expressed as
the concentratlon required for 50% inhibition(ICm) in micrograms per
titer and in least detectable dose (LDD) at 90% inhibbn. The first
value Is for the alachlor RaPID assay kit (magnetic particle-based
ELISA),andthesecondvakreIsfortheEnvlroOerdala~
kit(mlcrotlter
plate-based ELISA).

sensitive for alachlor, with LDD of 0.07 pg/L for RaPID

assay and 0.06 pg/L for EnviroGard assay. The RaPID assay
had an IC50for alachlor of 0.89 pg/L whereas the EnviroGard
assay had an ICs0 of 0.75 pg/L. Of the compounds tested for
cross-reactivity with the alachlor-ELISA,ESA had the highest
cross-reactivity, with an IC50 of 5.4 pg/L and a LDD of 0.19
pg/L for the RaPID assay and an ICs0 of 1.7 pg/L and LDD
of 0.12 pg/L for the EnviroGard kit. 2-[2,6-Diethylphenyl)(methoxymethyl)amino]-2-oxoacetic acid (OXA) has an ICs0
of 335 and LDD of 21 pg/L (RaPID assay) and IC50 of 66
pg/L and a LDD of 2.7 pg/L (EnviroGard).
The cross-reactivity of the anti-alachlor antibodies with
ESA has caused problems in previous water-quality surveys
for alachlor by ELISA because of false-positive detections of
alachlor.20 Furthermore, it has been reported that the ELISA
for alachlor does not correlate well with GC/MS and gives
more than 10%false positives near the limit of detection, with
concentration range of 0.10-0.20 pg/L.l29l3 This poor
correlation has been attributed to the presence of high ESA
levels as confirmed by liquid chromatography with tandem
mass spectrometry (LC/MS/MS).24 ESA and OXA are two
major soil metabolites of a l a c h l ~ r ;however,
~~
only ESA
appears to have sufficient cross-reactivity (Figure 2) to cause
interference with ELISA.
Other soil metabolites and precursors of alachlor, such as
hydroxydiethylacetanilideand chlorodiethylacetanilide, re(25) Sharp,D. B.Herbicides: Chentisry,DegradationandModeofAction.Kearney,
P. C., Kaufman, D.D..Eds.: Dekker: N e w York, 1988; pp 301-333.

Analytical Chemlstry, Voi. 00, No. 9, May 1, 1994

1497

Table 1. Breakthrough Capacity (mL) of Sep-Pak C1, Cartridges

for Alachlor and Other Ionlc Compounds

compound
alachlor
ESA
OXA
2,4-D

10% breakthrough

100%breakthrough

>7000
175
150
10

750
750
20

Table 2. Recovery and Precldon of Alachlor and ESA from a

CI8 RmIn by SrqumflaI E M h wtlh Ethyl Acetato and Methanol
by SPE-ELISA Udng the RaPID A m y Klt ( n = 7)
% recovered
concn of fortified

distilled water (fig/L)

alachlofl

ESAb

A. 0.20ESA
B. 0.050 alachlor + 0.10 ESA
C. 0.10 alachlor + 0.050 ESA
D. 0.025 alachlor 0.025 ESA
E. 0.20 alachlor

0
100f 10
97 f 10
105 f 10
98 f 10

92 f 10
97 f 10
99 f 10
91 f 10
0

spectively, did not show any cross-reactivity up to a concentration of 1000pg/L. Metolachlor, another chloroacetanilide
herbicide, showed low cross-reactivity, with an ICs0 of 109
and 27 pg/L for the RaPID and EnviroGard assay, respectively. This cross-reactivity pattern suggests that the binding
of alachlor toward the antibody is affected by the presence
of the methoxymethyl side chain and the alkyl group in the
ring. These results are not surprising because the antibodies
against alachlor are usually generated by the use of an
alachlor-protein conjugate, formed through the chlorinebearing carbon of alachlor via a thioether bond.12J3 The
antibodies develped for alachlor recognize ESA to a significant
degree relative to the other metabolites because the sulfur
atom in ESA is similar in size to the chlorine atom.
Separation by SPE. The separation of alachlor and ESA
was achieved by SPE using a CISresin. Both compounds
adsorbed quantitatively from water onto the CIS cartridge
(Table 1). Alachlor had a large capacity on the CISresin
with more than 7 L of sample passing through the cartridge
before a 10% breakthrough was ~bserved.~
In contrast, ESA
had considerably less capacity with 10% and 100% breakthroughs observed after 175 and 750 mL of sample, respectively, had passed through the cartridge. This difference in
capacity is caused by the greater solubility of ESA in water.
Likewise, OXA had a small capacity on the c18 resin, with
a 10% breakthrough beginning at 150 mL, and a 100%
breakthrough observed after 750 mL of water had passed
through the cartridge. A more soluble herbicide, 2,4dichlorophenoxyacetic acid (2,4-D), was also examined. It
was found that 2,4-D had a very small capacity for sorption,
with a 10% breakthrough observed after only 10 mL, and a
100%breakthrough observed after 20 mL of sample had passed
through the cartridge. These results suggest that the isolation
of ionic compounds on CISresin is possible, but this isolation
technique is quite dependent on the aqueous solubility of the
analyte. Thurman and others26reported that ionic character
was critical in the sorption of organic acids onto a macroporous
acrylic resin and was a function of chain length. The same
concept appears to apply to the isolation of ionic metabolites
on c18 resin because both isolations involve hydrophobic
interactions.
The separation of alachlor and ESA occurs in the elution
step. Alachlor is eluted first with ethyl acetate without removal
of ESA from the resin, and then ESA is eluted with methanol.
The recovery of spiked ESA from distilled water when analyzed
by SPE-ELISA ranged from 91% to 99% in the methanol
extract and none in ethyl acetate (Table 2). Alachlor was
recovered from 97% to 105% in the ethyl acetate extract and
0% in the methanol extract. Hence, there is a complete
(26) Thurman, E.M.;Malcolm, R.L.;Aiken, G.R. A n d . Chem. 1978,50,775779.

1490

Analytical Chemism, Vol. 66,No. 9, M y 1, 1994

In ethyl acetate extract. In methanol extract.

separation of ESA and alachlor by SPE. After separation

and concentration by SPE, it was possible to use the alachlorELISA for quantification of ESA due to the sufficient crossreactivity of the latter toward the anti-alachlor antibody. The
OXA coeluted with ESA in methanol as observed from HPLC
data, but because OXA has almost no cross-reactivity with
the alachlor-ELISA, it does not interfere with the analysis of
ESA. Themechanism of differential solubility in ethyl acetate
for ESA and OXA is probably a result of the fact that the
ionic organic molecules require a cation to be eluted with
them, either a sodium or calcium ion for most natural waters.
Apparently, the solubility of ESA-Ca or OXA-Ca in ethyl
acetate is low, probably because of its ionic character and
because of the necessity that its inorganic cation and water
also must be present in solution. The result is a separation
of parent compound and metabolites for ELISA. The
importance of this finding is that many ionic compounds,
particularly metabolites of herbicides, may be separated from
parent compounds by this procedure.
SPE-ELISA Method. The detection limit for alachlor of
the RaPID assay using the SPE-ELISA method described in
Figure 1 isO.01 pg/L. Thisis 10-foldlowerthan thedetection
limit of the conventional ELISA (detection limit is 0.1 pg/L).
ESA was eluted with methanol and was analyzed by ELISA
using the alachlor-RaPID assay but calibrated with ESA
standards. The calibration curve for ESA was linear for
concentrationsof 1.O-20 pg/L. Because SPE preconcentrates
the sample, the detection limit can go as low as 0.05 pg/L if
a 100-mL sample is used and the dried extract is diluted to
5 mL with water for ELISA. Samples containing ESA were
verified by HPLC. It should be noted that solvent exchange
of the organic solvent to an aqueous medium was performed
prior to ELISA because methanol may affect the antibody/
antigen interaction and produce an erroneous result. Although
some antibodies can tolerate more than 50%methanol, others
may be disrupted by minor amounts of methanol. The RaPID
assay kit tolerates 20% methanol, but the reproducibility of
the Enviroguard alachlor kit was affected, with increased 7%
CV observed using 20% methanol in the sample matrix.
The automated workstation for SPE acts as a liquid
chromatograph with a fraction collectorby pumping the sample
through the cartridge, separating the compounds by sequential
elution, and collecting the fractions for analysis by immunoassay, GC/MS, and HPLC. The automated SPE procedure
is capable of reproducing the separation within *lo%.
The ability of SPE to separate alachlor and ESA was
demonstrated with ground water samples from 12 locations.
The presence of both alachlor and ESA in the ground water

Table 3. Analyrk of Ground Water Samples for Alachlor and

ESA by ELISA, SPE-ELISA, and GC/MS

alachlor concn oCg/L)

ESA concn 01 /L)
direct anal.
SPE-ELIS!
sample by ELISA GUMS SPE-ELISA
1
2
3
4
5
6
7
8
9
10
11
12

0.6
0.16
0.92
0.33
1.83
1.10
0.22
8.60
0.22
2.85
1.79
4.63

0.45
C0.05
K0.05
<0.05
0.07
0.21
C0.05
1.50
0.06
0.52
0.36
2.95

0.50
0.04
0.05
0.03
0.10
0.39
0.04
1.26
0.09
0.61
0.33
2.78

0.06
0.83
2.18
1.65
4.80
1.88
1.10
3.50
0.91
9.32
4.29
5.41

samples resulted in an overestimation of alachlor when they

were analyzed directly by ELISA. Four out of these 12
samples produced "positive detection" of alachlor by ELISA
but negative detection by GC/MS (detection limit of 0.05
pg/L), and 8 samples had significantly higher alachlor
concentrations by ELISA compared to that by GC/MS (Table
3). The discrepancy between the amount of alachlor measured
by GC/MS and that measured by ELISA was proportional
to the amount of ESA in the sample. However, when the
water samples were passed through SPE and the ethyl acetate
eluate was analyzed for alachlor by ELISA as described above,
the concentration of alachlor agreed with that obtained from
GC/MS. In addition, analysis of the methanol eluates by
ELISA showed high levels of ESA whenever there was a
discrepancy between ELISA and GC/MS for alachlor
concentrations. This indicatesfurther that ESA, which caused
the false alachlor concentration in ELISA, was recovered in
the methanol and not in the ethyl acetate extracts. These
results suggest that the SPE-ELISA method is quite effective
for the separation and analysis of alachlor and ESA.
Figure 3A shows the correlation of ESA analysis by HPLC
and SPE-ELISA for 22 surface water samples from reservoirs
in the midwestern United States. The correlation coefficient
( r )of 0.93 demonstrates that the two methods correlate quite
well, and the slope of 0.69 indicates that HPLC concentrations
generally were lower. The smaller ESA concentrations
indicated by HPLC analysis may be attributed to high levels
of humic and fulvic acids that were coextracted in themethanol
fraction. It was difficult to measure ESA at concentrations
less than 0.10 pg/L using HPLC because natural organic
substances were recovered in the methanol fraction and gave
a large response in HPLC. Ground-water samples from several
wells in the midwestern United States were analyzed by HPLC
and results showed better correlation (r = 0.97) with ELISA
(Figure 3B), which again may be the result of the smaller

'

8.0

0.0

y=O.16+1.2x

r =0.97

2.0

4.0

8.0

8.0

10.0

CONCEN'RATlONOF ESA DEERMNED

BY ELISA. IN MICROGRAMS PER LITER

Flgure 3. Concentrationsof ESA in (A) 22 surfacbwater samples and

(6) 20 ground water samples determined by HPLC and SPE-ELISA
using the Alachbr RaPID assay.

amounts of humic and fulvic acids in the samples. Also, a

higher slope (1.2) compared with that observed from the
surface water samples was obtained. The samples from the
various wells and reservoirsvaried in ESA concentration from
less than 0.1 pg/L to greater than 10 pg/L. Further studies
are underway to understand the fate and transport of the
ESA in the soil and in aquatic environment.

CONCLUSIONS
The coupling of SPE and ELISA resulted in a reproducible
immunoassay method for the trace analysis of alachlor and
its major soil metabolite, ESA. Alachlor and ESA were
separated by sequential elution with ethyl acetate and
methanol. The SPE method is a simple and elegant procedure
because cleanup, concentration, and separation occur in a
single step. Because ESA cross-reacts strongly with the antialachlor antibody, it was possible to analyze ESA by
immunoassay using the ELISA kits designed for alachlor
analysis. The method was viable for the analysis of both
surface and ground water samples and comparable to GC/
MS and HPLC analyses, for alachlor and ESA, respectively. The method is sensitive, with a detection limit of 0.01
pg/L for alachlor and 0.05 pg/L for ESA, with a precision
of *lo%.
Received for review October 18, 1993. Accepted February 4,
1994.@
Abstract published in Advance ACS Absrracrs, March 15, 1994.

Analytical Chemistty, Vol. 88, No. 9,May 1, 1994

1499