Journal of Environmental Science and Health Part B (2008) 43, 134–140Copyright C© Taylor & Francis Group, LLCISSN: 0360-1234 (Print); 1532-4109 (Online)DOI: 10.1080/03601230701795072

Determination of the herbicide fluroxypyr in oil matrices

HALIMAH B. MUHAMAD1, TAN YEW AI1 and ISMAIL B. SAHID2

1Malaysian Palm Oil Board, Kuala Lumpur, Malaysia2School of Environmental and Natural Resource Sciences, University Kebangsaan Malaysia, Bangi, Selangor, Malaysia

The purpose of this study was to develop a method for the determination of fluroxypyr (4-amino-3,5-dichloro-6-fluro-2-pyridyloxyacetic acid) residue in palm oil namely crude palm oil (CPO) and crude palm kernel oil (CPKO). The method involvesthe extraction of the herbicide from the oil matrix followed by low temperature precipitation and finally quantification of the residuesusing the high performance liquid chromatography (HPLC). The extraction efficiency of the method was evaluated by conductingrecovery studies. The recovery of fluroxypyr from the fortified CPO samples ranged from 78%–111% with the relative values for thecoefficient of variation ranging from 1.4 to 8.6%. Furthermore, the recovery of fluroxypyr from the spiked CPKO samples ranged from91–107% with the relative values for the coefficient of variation ranging from 0.6 to 4.5%. The minimum detection limit of fluroxypyrin CPO and CPKO was 0.05 µg/g. The method was used to determine fluroxypyr residues from the field-treated samples of CPO andCPKO. When fluroxypyr was used for weed control in oil palm plantations no residue was detected in CPO and CPKO irrespectiveof the sampling interval and the dosage applied at the recommended or double the manufacturer’s recommended dosage.

Keywords: Herbicide; fluroxypyr; crude palm oil; crude palm kernel oil.

Introduction

Weeds are the most common problem in oil palm planta-tions especially at the immature stage of the palm. Her-bicides are regularly applied to immature oil palm standswhen the palms are less than eight years old. Flurox-ypyr is commonly used to control broadleaved weeds in-cluding leguminous cover crops such as Calopogoniumcaeruleum, Centrosema pubescens, Pueraria phaseoloides,and other weeds such as Asystasia gangetica, and Mika-nia micrantha.[1]

Fluroxypyr is a herbicide recently introduced for post-emergence control of broad-leaved weeds in cereal andperennial crops. Fluroxypyr is the common name for4-amino-3,5-dichloro-6 fluro-2-pyridinyloxy acetic acid,(C7H5Cl2FN2O3) and is sold under the trade name ofStarane©R . It is grouped under the pyridyloxyacetic acidweedicides and is a colorless crystal with a melting point of232–233◦C. It has low solubility in water, about 91 mg/Lat 27.7◦C.

Application of herbicides in oil palm plantations maycause contamination to the palm oil. Since palm oil is a

Address correspondence to Ismail B. Sahid; School of Envi-ronmental and Natural Resource Sciences, Faculty of Scienceand Technology, Universiti Kebangsaan Malaysia, 43600 UKM,Bangi, Selangor, Malaysia; E-mail: ismail@ukm.myReceived June 11, 2007.

traded commodity, monitoring pesticide residue levels isvery important to maintain quality. The palm oil shouldbe free from chemical residues as it used for human con-sumption. Therefore, it is necessary to develop a suitablemethod for analyzing pesticide residues in the oil to en-sure that the palm oil produced in Malaysia can meet thepesticide residue regulatory requirements of the importingcountry.

Reports on the method development to analyze pesti-cide residues in palm oil are limited.[2−5] A method for de-termination of chlorpyrifos in crude palm oil (CPO) andcrude palm kernel oil (CPKO) using liquid-liquid extrac-tion in combination with solid liquid clean-up has beenpublished.[4,5] Gillespie and Walters[6] reported a methodfor extraction of organophosphorus pesticide (OP) residuein fats using C18 solid phase extraction cartridges. Reportson the method of extraction of fluroxypyr from water andsoil have also been published.[7−10] However, to date there isno method or data available for the determination of flurox-ypyr in either CPO or CPKO. Since fluroxypyr is used inoil palm plantations, there is a probability of it contami-nating the palm oil. Therefore, it is necessary to develop anaccurate and efficient method for detecting the compoundin palm oil. The present study was designed to develop pre-cise, accurate and reproducible methods for determinationof fluroxypyr in both CPO and CPKO. The procedure ob-tained from this study was used to analyze the fluroxypyrresidue in CPO and CPKO from fields where fluroxypyr

Fluroxypyr in oil matrices 135

was used for weed control. The data obtained can be usedto establish MRLs for this herbicide. For this purpose, anaccurate method for determination of fluroxypyr is very im-portant to ensure that the data obtained are reliable. Cur-rently, CODEX MRLs of fluroxypyr have been based onthe information from other commodities such as vegetablesand fruits. In general, data on the residue trials on MRLsof CPO and CPKO in Malaysia are very limited, although areport has been published on the residue trial of trichlorfonin CPO and CPKO.[11]

Materials and methods

Experimental design

The study was conducted in the oil palm plantation of theMalaysian Agricultural Horticultural Company. The plan-tation is situated in central Selangor, Peninsular Malaysia,at an altitude of 70–100 m above sea level. The study plotwas on hilly land of slope 45◦ with predominantly clayeysoil. The soil had high clay content (52.2%), cation exchangecapacity (CEC) 6.9%, coarse sand (15%), fine sand (27%)and silt (5.8%).

The field trials on the fate of fluroxypyr in the oilpalm plantation were conducted during the wet season ofNovember 2001 to February 2002. The study plot covered6 hectares, and consisted of 9 subplots of 0.66 ha each.The nine subplots received three treatments namely: control(no herbicide treatment), treatment at the manufacturer’srecommended dosage (72.05 g a.i./ha) and treatment atdouble the recommended dosage (144.10 g a.i./ha). Eachtreatment was done in triplicate, and each subplot was sep-arated by a 2 m buffer zone. Fluroxypyr (Starane©R 200 EC;Dow Elanco Ltd) was applied as an aqueous spray usinga knapsack sprayer (nozzle 5/64) at a spraying volume of250 L/ha. Five of the samples were collected in November2001, one in December 2001, and the remaining two in Jan-uary 2002 and February 2002. The daily average rainfall,evaporation and atmospheric temperature at the oil palmplantation during the study period were recorded.

Method of sampling from field trials

Fresh fruit bunches (FFBs) were harvested the day beforetreatment, and on days 0 (immediately after spraying), 1,3, 5, 7, 14, 21, 30, 60 and 90 after spraying. Samplingsof the control plot were done concurrently with those oftreated plots. A total of nine palm trees were selected foreach sampling. Nine samples of FFB were harvested fromeach treatment, including the control plot. Random sam-pling was done on the FFBs from each plot to obtain abulked 5 kg sample of loose fruits. The fruits from eachreplicate that had received the same treatment were bulkedtogether to get a homogeneous sample. The samples wereplaced in plastic bags and sent to the Pesticide Laboratory

of the Malaysian Palm Oil Board (MPOB) in Bangi imme-diately after harvesting. Samples were kept in a cold roomat 5◦C before being processed for oil extraction.

Method of sampling from blank sample plot

For the blank oil samples, five FFBs were taken fromthe Malaysian Palm Oil Board (MPOB)/Universiti Ke-bangsaan Malaysia (UKM) Research Station. These FFBswere free of fluroxypyr residues as no spraying had beendone. Random sampling was done on the FFBs to obtaina bulked 3 kg sample of loose fruits. The blank oil sampleswere to be used for determining the percentage of recoveryby spiking with known amounts of fluroxypyr.

Sample processing and extraction of CPO and CPKO

The FFBs of both treated and blank sample plots were cutinto spikelets and put into separate autoclave-safe plasticbags to avoid contamination. The spikelets were then ster-ilized in an autoclave (Sakura Neoclave ASV-302) at 120◦Cand 1.2 kg/cm2 for 40 minutes. In order to avoid crosscontamination, sterilization was done in sequence, startingfirst with the samples from the blank plots, control plots,followed by the samples where the palms were sprayed atthe recommended dosage and at double the recommendeddosage. The sterilized fruits were than manually strippedfrom the spikelets and placed into a mini hydraulic handpress for extraction of the oil. A centrifuge set at 3,000 rpmfor 20 min was used to separate the extracted crude palm oil(CPO) from fiber and other solid material. The CPO wasthen decanted, filtered through Whatman No. 4 filter pa-per containing anhydrous sodium sulfate to remove water.The CPO obtained was kept in a freezer at 0◦C prior to theextraction of the fluroxypyr residue. After the extraction ofCPO, the fruits were kept at 0◦C in the freezer before theextraction of CPKO.

Before extracting CPKO, the mesocarp fibers from thepressed sterilized fruits was peeled off the nut with a knife.The nut was then cracked using a hammer to obtain thekernel. The kernel was then broken into smaller pieces usinga mortar and pestle. The broken kernel was crushed in ablender to obtain a homogenous mixture. Ten gm of theground kernel were weighed into an extraction thimble andthe thimble was plugged with a wad of cotton wool. Tenthimbles with 10 gm each of ground kernel were solventextracted with 150 mL of n-hexane for 6 hours to obtainthe CPKO for analysis. The solvent was evaporated using arotavapor (N-Evap Model 1111; Organomation Assoc. Inc.USA). After removal of the solvent, nitrogen was passedthrough the oil mixture to remove the remaining solvent.The CPKO was then stored in brown bottles and kept ina freezer at −20◦C prior to the extraction of fluroxypyrresidue.

136 Muhamad et al.

Recovery of fluroxypyr from the spiked blankCPO and CPKO

Samples of CPO and CPKO obtained from fruits in theblank sample plot were melted at 60◦C and then homog-enized. The oil samples were weighed into containers andthe appropriate amount of standard solution was individu-ally added into them to obtain concentrations ranging from0.05 to 1.0 µg/g of fluroxypyr. The extraction of fluroxypyrresidue was carried out as described below prior to HPLCanalysis.

Extraction of fluroxypyr from CPO

Ten g of melted CPO from fruits in the blank sample plotwere weighed individually into 250 mL conical flasks andspiked with 0.5 to 1.0 µg/mL of fluroxypyr standard solu-tion in acetone. Each flask was shaken for 15 seconds usinga vortex mixer. Each of the spiked samples was dissolved in40 mL hexane and mixed thoroughly on a vortex mixer fora few seconds. Then 20 mL acetonitrile were added, and thesamples were then mixed using a vortex mixer and placedinto an ultrasonic bath for 30 min. The contents were thentransferred into a 250 mL separating funnel. The acetoni-trile phase (lower phase) was drained into a 100 mL roundbottomed flask, stoppered and kept in a freezer for 2 hoursat –18◦C, after which the contents were filtered, leaving thesolid oil on the filter paper, and the filtrate was collected ina 100 mL round bottomed flask. Ten mL of the filtrate werethen taken and dried in a stream of nitrogen gas. The residuewas then redissolved in 1 mL acetonitrile before being in-jected into the high performance liquid chromatography-diode array detector (HPLC-DAD).

Extraction of fluroxypyr from CPKO

Twenty ± 0.001g samples of CPKO blank were weighed intoindividual 20 mL screw cap centrifuge tubes and spikedwith standard fluroxypyr in acetone (0.2–1 µg/mL). The

Fig. 1. Calibration curve of standard fluroxypyr against the high performance liquid chromatography (HPLC) peak area using a diodearray detector (DAD).

analysis of fluroxypyr was replicated five times. Fluroxypyrwas then extracted by adopting the following procedure: 10mL of methanol was added and the contents were mixedfor 3 min on a vortex mixer. The tube was then allowedto stand in the freezer at −20◦C for 45 min. The contentswere then filtered using Whatman No. 1 filter paper. FivemL of the filtrate were then taken and dried in a stream ofnitrogen gas. Finally, the residue was redissolved in 1 mLacetonitrile before being injected into the HPLC-DAD.

Calibration of HPLC using standard solution

Analytical grade acetonitrile, acetone, hexane andpetroleum ether (boiling point 60◦C-80◦C), all supplied byMerck, were used. Standard fluroxypyr of 99.7% purity waspurchased from the Laboratories of Dr. Ehrenstonfer, Ger-many. A sample of 0.01 g analytical grade fluroxypyr wasdissolved in 50 mL acetone and made up to a stock solu-tion of 200 µg/mL. Working standard solutions containing0.05 µg/mL to 3 µg/mL were prepared by appropriate di-lution of the standard stock solution with acetonitrile. Allthe standard solutions were stored at –20◦C in glass bot-tles with teflon-lined screw caps prior to injection into theHPLC.

HPLC-DAD: conditions for the analysis of fluroxypyr

An Agilent HPLC 1100 Series fitted with a diode arraydetector (DAD) was used. The column used was a LUNARP18, 5 µm (250 mm × 4.6 mm I.D.). A guard cartridgeRP 18, 5 µm (Phenomenex) with the RP-18 column wasused. The mobile phase was acetonitrile and acetate buffer(1% acetic acid in water) in the ratio of 60:40. The flowrate and injection volume were 0.8 mL/min and 100 µL,respectively. The wavelength used for the recovery study andquantification of fluroxypyr in the CPO and CPKO sampleswas 254 nm.

Fluroxypyr in oil matrices 137

Calculation of fluroxypyr residue in CPO and CPKO

The peak response (area) of standard fluroxypyr in thechromatogram of the working standards was measured.The linear least square analysis on the relationship be-tween the standard amount injected (µg/mL) versus thepeak area was calculated to obtain the calibration line.Five working standard solutions of fluroxypyr at lev-els from 0.05 to 3 µg/mL were used for the calibrationcurve.

Recovery was calculated using the following Equation 1:

%recovery = A1/A2 × 100 (1)

Fig. 2. High performance liquid chromatography (HPLC) chromatograms for a) fluroxypyr standard solution of 1.0 µg/mL, b)untreated sample of crude palm oil (CPO) and c) sample of CPO, fortified at level of 0.1 µg/g fluroxypyr.

where A1 is the peak area of the spiked sample and A2 is thepeak area of actual standard fluroxypyr. For each sampleinjection, the fluroxypyr concentration (in µg/g) in the oilsample was calculated from the sample peak response andthe standard peak response.

The fluroxypyr content (µg/g) was calculated by the fol-lowing formula (Equation 2):

Conc. Std × V1/V2 × A1/A2 × FV/W (2)

where Conc. Std is the concentration of standard injection,V1 is the volume of standard injection, V2 is the volume ofsample injection, A1 is the area of sample response, A2 isthe area of standard response, FV is the final volume, and

138 Muhamad et al.

W is the weight (g) of the oil sample. The average of twoinjections was taken for each sample (to one decimal place).

Results and discussion

Figure 1 shows the calibration data obtained from trip-licate analysis, with each solution being injected thrice.

Fig. 3. High performance liquid chromatography (HPLC) chromatograms for a) fluroxypyr standard solution of 1.0 µg/mL, b) un-treated sample of crude palm kernel oil (CPKO) and c) sample of CPKO, fortified at level of 0.1 µg/g fluroxypyr.

The linear regression (r2) was found to be 0.9992 andthe equation derived from the calibration area data wasy = 163.06x – 4.4158 where y was the area of flurox-ypyr obtained from the HPLC analysis and x the con-centration of fluroxypyr in µg/mL. The reproducibilityand linearity of the injection technique were found tobe acceptable and the performance of HPLC-DAD forthe determination of fluroxypyr was at 99.9% confidence.

Fluroxypyr in oil matrices 139

The detection limit of fluroxypyr for HPLC-DAD was0.05 µg/mL.

The recovery of fluroxypyr residue in CPO and CPKOsamples spiked with standard fluroxypyr solution at five dif-ferent concentrations ranging from 0.05 to 1 µg/g is shownin Table 1. The recovery rates of fluroxypyr ranged from78 to 111% with the coefficient of variation ranging from1.4 to 9.9%. Figure 2 shows the chromatograms obtainedusing HPLC-DAD for standard fluroxypyr solution of 1µg/mL, an untreated sample of CPO and a spiked sam-ple of CPO containing 0.1 µg/g. The recovery results were

Fig. 4. High performance liquid chromatography (HPLC) chromatograms for field treated samples a) control CPO, b) control crudepalm kernel oil (CPKO, c) fluroxypyr treated sample of crude palm oil (CPO), and d) fluroxypyr treated sample of CPKO.

satisfactory for residue analysis as shown in Table 1. Thelimit of detection of fluroxypyr in CPO was 0.05 µg/g. Theretention time for fluroxypyr was 10.5 min.

The recovery rates of fluroxypyr from spiked CPKO sam-ples ranged from 91–109% with standard deviation less than5% (Table 1). This indicated that this analytical methodfor fluroxypyr analysis in CPKO was precise, accurate andacceptable for application in determining residual levelof fluroxypyr in CPKO. Figure 3 shows the HPLC chro-matograms corresponding to fluroxypyr standard solution,untreated samples of CPKO and a spiked sample of CPKO,

140 Muhamad et al.

Table 1. Percentage recovery of fluroxypyr from crude palm oil (CPO) and crude palm kernel oil (CPKO)

CPO CPKO

Concentration Relative Relativeof fluroxypyr Recovery (%) standard Recovery (%) standard(µg/g) N = 5 deviation (%) N = 5 deviation (%)

0.05 111 7.0 99 4.50.1 109 2.8 107 1.00.2 97 3.0 98 1.40.4 109 9.9 109 4.00.5 96 8.6 98 2.51.0 78 1.4 91 0.6

respectively. The minimum detection limit of fluroxypyr inCPKO was 0.005 µg/g and the retention time of fluroxypyrwas 10.8 min.

The extraction method used in this experiment for thedetermination of fluroxypyr in palm oil and its productswas a modification of the multiresidues method outlinedby Gillespie and Walters[6] and Halimah et al.[4]. Petroleumether-saturated acetonitrile was used to extract OP pesti-cides from oil and it was found that 10 mL of the solventmixture was needed for optimum extraction of chlorpyri-fos from the oil matrix. However, in this study althoughthe same solvent mixtures were used for the extraction offluroxypyr, it was found that 20 mL of the solvent mixturewas needed for complete extraction of fluroxypyr in CPKO.Commercial C18 (reverse phase) cartridges were used forcleaning up step by Gillespie and Walters[6] but low tem-perature precipitation were used in the present study.

In this study, no residue was found in oil extracted fromfruits where oil palm plantation plots were treated in thefield with fluroxypyr, even as long as 60 days after treat-ment (Figure 4). Therefore, it appears that fluroxypyr issafe for use in oil palm plantations, as the residue was notfound in the CPO and CPKO extracted from palm fruitson 1, 5, 7 or 14, 21, 30 and 60 days after application ofthe herbicide. Since the amount of fluroxypyr deposited inthe soil was found to be very low,[12] there was low prob-ability of the uptake of fluroxypyr residue into the plantsystem. Furthermore, since fluroxypyr has low solubility inwater (91 µg/L), it is unlikely for the plant roots absorbthe residue. Hence, it is highly improbable that the residuecould be present in either the fruits or the oil.

Conclusion

From the results obtained it can be concluded that the pro-posed method is suitable for quantification of fluroxypyrresidue in an oil matrix. This method appears to be suitablefor routine monitoring of fluroxypyr residue in palm oil andits products as the results show good recovery with low co-

efficients of variation. The data also showed that there wasno extractable residue from fruits where plantations weretreated with fluroxypyr. Therefore, based on the informa-tions obtained in this study it is suggested that fluroxypyris safe for use as a weedicide in oil palm plantations as noresidue was found in the CPO and CPKO from fruits col-lected from palms after herbicide spraying, irrespective ofthe time of spraying.

References

[1] Chung, G.F.; Balasubramaniam, R.; Cheah, S.S. Recent develop-ment in spray equipment for effective control of pests and weeds.The Planter 2000, 76, 65–84.

[2] Cheah, U.B. Ozonation-microbial degradation of 2,4-D. Proceedingsof the International Conference on Pesticides in Tropical Agricul-ture, Kuala Lumpur, Malaysia, 1987, 56–68.

[3] Cheah, U. B.; Kirkwood, R.C.; Lum, K.Y. Adsorption-desorptionand mobility of four commonly used pesticides in Malaysian agri-cultural soils. Pestic. Sci. 1997, 50, 53–63.

[4] Halimah, M.; Osman, H.; Ainie, K.; Tan, Y.A.; Md.-Fauzi, A. De-termination of chlorpyrifos in refined palm olein by GC-FPD andGC-ECD. J. Palm Oil Res. 1999, 11 (2), 89–97.

[5] Halimah, M.; Md.-Pauzi, A.; Chian, S.S. Optimization of sweepco- distillation cleanup method for determination of organochlorinepesticide residues in oil matrix. J. Palm Oil Res. 2004, 16 (2), 30–36.

[6] Gillespie, A.M.; Walters, S.M. Rapid clean-up of fat extracts fororganophosphorus pesticide residue determination using C18 solidphase extraction cartridges. Anal. Chimica Acta 1991, 245, 259–265.

[7] Lehmann, R.G.; Miller, J.R.; Soil catalysed hydrolysis of fluroxypyr-methylheptyl ester. Weed Res. 1989, 29, 385–389.

[8] Lehmann, R.G.; Miller, J.R.; Cleveland, C.B. Fate of fluroxypyr inwater. Weed Res. 1993, 33, 197–204.

[9] Halimah, M.; Tan, Y.A.; Ainie, K.; Ismail, B.S. Method develop-ment for determination of fluroxypyr in water. J. Environ. Sci. Health2003, B 38, 429–440.

[10] Halimah, M.; Tan, Y.A.; Ismail, B.S. Method development for deter-mination of fluroxypyr in soil. J. Environ. Sci. Health 2004, B39(5),765–777.

[11] Halimah, M.; Tan, Y.A.; Ainie, K. Supervised trial of trichlofon incrude palm oil and palm kernel oil. Bangi: Malaysian Palm Oil Board.2002, pp. 1–12.

[12] Halimah, M.; Tan, Y.A.; Ismail, B.S. Fate of fluroxypyr in soil underoil palm agroecosystem. Weed Biol. Manag. 2005, 5, 184–189.