Research Article Received: 3 July 2013
Revised: 8 November 2013
Accepted article published: 19 November 2013
Published online in Wiley Online Library: 26 December 2013
(wileyonlinelibrary.com) DOI 10.1002/ps.3684
Simplified analysis of glyphosate and aminomethylphosphonic acid in water, vegetation and soil by liquid chromatography–tandem mass spectrometry LeEtta J Mareka∗ and William C Koskinenb Abstract BACKGROUND: There is a need for a simple, fast, efficient and sensitive method for analysis of glyphosate and its degradate aminomethylphosphonic acid (AMPA) in diverse matrices such as water, vegetation and soil. RESULTS: Aqueous extracts from water, vegetation and soil were passed through reverse-phase and cation-exchange columns and directly injected into a tandem mass spectrometer using only a guard column for separation. Extraction efficiencies from the three matrices were >80% for both glyphosate and AMPA. The method reporting levels (MRLs) for glyphosate in water, vegetation and soil were 3.04 µg L−1 , 0.05 mg kg−1 and 0.37 mg kg−1 respectively. AMPA MRLs were 5.06 µg L−1 for water, 0.08 mg kg−1 for vegetation and 0.61 mg kg−1 for soil. CONCLUSIONS: A validated, simple and efficient liquid chromatography–tandem mass spectrometry (LC-MS/MS) method for routine analysis of glyphosate and AMPA in water, vegetation and soil that uses minimal sample handling and clean-up will facilitate the additional environmental research needed to address the continuing concerns related to increasing glyphosate use. Published 2013. This article is a U.S. Government work and is in the public domain in the USA. Keywords: glyphosate; AMPA; LC-MS/MS analysis; soil; water; vegetation
1
INTRODUCTION
1158
Glyphosate [N-(phosphonomethyl)glycine], a non-selective, broad-spectrum herbicide, has become the most widely used herbicide in the world. Since its introduction and subsequent commercialization in 1974, glyphosate has been considered to be a toxicologically and environmentally safe and effective herbicide. However, recently there have been increasing reports of incidences of glyphosate-resistant (GR) weeds, as well as weed shifts to naturally tolerant weed species. This has resulted in an increased number of glyphosate applications and increased rates of application for weed control, which in turn has raised concerns about adverse impacts of glyphosate use. For instance, several recently published papers have concluded that glyphosate adversely affects the mineral nutrition of GR crops, which has led to other problems; however, a recent review of the literature indicates that the majority of the literature shows no effect of glyphosate application on mineral nutrition of GR crops.1 The increasing applications and rates have also raised concerns about its environmental effects after long-term use. To address the continuing concerns related to increasing glyphosate use, additional research will continue to be conducted. To facilitate glyphosate research, simple, sensitive methods of analysis are needed for glyphosate in water, plant and soil samples. In addition, it has been suggested that the simple, sensitive methods developed for glyphosate can be used for analysis of chemical weapons such as phosphonic-acid-containing Pest Manag Sci 2014; 70: 1158–1164
nerve agents; glyphosate has been used as a surrogate for these chemicals.2 Numerous analytical techniques used to quantify glyphosate have proven to be reliable and sufficiently sensitive for most purposes. However, most of the methods are rather complex and time consuming, which leads to numerous ‘new’ or ‘revised’ methods being reported every year in the literature for water, plants or soil. The method modifications usually involve varying the details in the extraction and/or clean-up procedures, separation techniques or detection methods of previously published methods for each matrix, or the use of a previously published method with a new matrix. For instance, in the past 5 years alone, numerous ‘new’ methods have been published on the analysis of glyphosate and its principal degradate aminomethylphosphonic acid (AMPA) in water. They have included capillary electrophoresis (CE) with electrochemiluminescence detection,3 CE with contactless conductivity detection,4 CE with online isotachophoresis (ITP) pretreatment of waters on a column-coupling chip,5 capillary ion chromatography with a suppressed conductive detector,6
∗
Correspondence to: LeEtta J. Marek, Minnesota Department of Agriculture, St Paul, MN, USA. E-mail: [email protected]
a Minnesota Department of Agriculture, St Paul, MN, USA b USDA-ARS, St Paul, MN, USA
www.soci.org
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
Glyphosate and AMPA analysis in water, vegetation, and soil by LC–MS/MS anion-exchange chromatography with coulometric detection,7 high-performance ion chromatography coupled to inductively coupled plasma dynamic reaction cell mass spectrometry,8 high-performance liquid chromatography (HPLC) coupled to electrochemical9 and ultraviolet10 detectors and tandem mass spectrometry.11 Different online water preconcentration techniques have been reported,4,11 along with offline enrichment techniques such as using ion-exchange resins8 and solid-phase extraction discs.9 Clean-up procedures using cation-exchange resins have been reported.12 Liquid chromatography (LC) columns used for analyte separation include anion-exchange,9 reverse-phase10,11 and mixed-mode columns.13 To lower levels of detection in water, new pre- and post-column methods to derivatize glyphosate and AMPA have been reported.10,14 – 16 Fewer new methods have been reported in the past 5 years for more complex matrices such as vegetative material, soil and sewage sludge. These have involved minor variations of previously published methods or a new matrix: use of a hydrophilic interaction chromatography (HILIC) column with UV-visible or electrospray ionization (ESI)-MS/MS detection for analysis of glyphosate in selected fruits and vegetables,17 sample clean-up by protein precipitation for LC-MS/MS analysis of glyphosate and AMPA in soybean,18 optimization of derivatization methods using 9fluorenylmethylchloroformate (FMOC) for analysis of glyphosate and AMPA in maize by HPLC coupled to fluorescence and ESIMS,19 an optimized extraction procedure and chromatographic separation for analysis of glyphosate and AMPA in soil,20 online FMOC derivatization on a strong anion-exchange resin as a solid support for analysis of glyphosate and AMPA in sewage sludge21 and LC-ESI-MS or MS/MS analysis of wheat and rye22 and soybean.18 In spite of these recently published methods, which have primarily been developed for a single matrix, there is a need for a simplified method of analysis for glyphosate and AMPA that will work for multiple, diverse matrices. The objective of the present project was to develop and validate a simple and efficient LC-MS/MS method for routine analysis of glyphosate and its degradation product AMPA in water, vegetation and soil. The method uses minimal sample handling and clean-up and requires no derivatization. Samples of river water, aqueous plant extracts and acidic aqueous soil extracts were passed through IC-RP SPE cartridges into LC vials and injected into an LCMS/MS system with only a cation guard column for separation in the HPLC.
2
MATERIALS AND METHODS
Pest Manag Sci 2014; 70: 1158–1164
2.2 Standards Separate stock standard solutions (1.00 g L−1 ) of glyphosate and AMPA were prepared in plastic volumetric flasks using certified standards of known purity. An amount (∼0.0100 g) of standard was weighed, and an appropriate amount of freshly opened/poured HPLC-grade water was added to obtain a solution concentration of 1.00 g L−1 . The stock standards prepared in HPLC-grade water are reported to be stable for at least 1.3 years and up to 23 months for glyphosate and AMPA, respectively, if stored in a refrigerator (Monsanto, unpublished report ‘Determination of glyphosate in ground, surface and finished drinking waters by LC-MS/MS’; DuPont, 2008). Freshly opened/poured HPLC-grade water was used to prepare intermediate standards with 0.5, 5.0 and 50 mg L−1 concentrations of both glyphosate and AMPA for analysis of water, vegetation and soil. Soil and vegetation analyses additionally required two internal standard mixes which were prepared with [13C]3 , [15 N]1 glyphosate and (d2) 13C, 15 N AMPA obtained from Monsanto. HPLC-grade water was used to prepare 25 mL of a 50 mg L−1 mixed standard solution from which a 100 µL aliquot was diluted to 10 mL to obtain a 0.5 mg L−1 solution. This 0.5 mg L−1 internal standard mix was used to spike the samples as well as in the calibration standard. A series of calibration standards were used, ranging from 2 to 200 µg L−1 , which were prepared through dilution of intermediate standards with HPLC-grade water and 30% phosphoric acid to obtain a concentration of 1% phosphoric acid. Calibration solutions must be stored in a refrigerator in plastic containers until use. Plastic containers are employed to avoid possible chelation of the glyphosate with metal ions on the glassware. The calibration standards prepared in aqueous acid should be stable for 99 days. 2.3 Water sample preparation Surface water samples were taken from seven rivers in south central Minnesota. Visual evaluation of the river water samples showed that three of them had high concentrations of sediments, two had medium concentrations of sediments and two had high concentrations of algae. For each sample, including quality control (QC) samples, 15 mL plastic centrifuge tubes were used. To each spike control centrifuge tube, a known amount of intermediate standard (i.e. 200 µL of the 0.5 mg L−1 standard) was added, along with a known amount of HPLC-grade water (i.e. 4.8 mL). An aliquot of each water sample equal to the volume of the standard (5 mL) was transferred into a centrifuge tube. An equal volume of HPLC-grade water was added to a centrifuge tube as a blank. Each centrifuge tube was acidified with concentrated phosphoric acid (85%) (final volume 5 mL per 1.0% phosphoric acid), capped and mixed with a vortex mixer. Samples were then passed through IC-RP SPE cartridges (Grace Maxi-CleanTM , 0.5 mL; Deerfield, IL) attached to 3 mL disposable syringes with nylon syringe filters (Acrodisc, 13 mm × 0.2 µm; Pall Life Sciences, Ann Arbor, MI,) connected to the outlet of the cartridge and collected in polypropylene autosampler vials with preslit PTFE/silicone septa. Samples were stored in the refrigerator or chilled autosampler until LC-MS/MS analysis. 2.4 Vegetation sample preparation For method development, vegetative samples from diverse sources were mixed and then ground and homogenized using a Robot CoupeTM (Jackson, MS) sample processor and returned to the freezer if not immediately analyzed. Additionally, samples of individual plants were ground and homogenized for recovery
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
wileyonlinelibrary.com/journal/ps
1159
2.1 Chemicals/reagents Glyphosate (purity 99.8%) and AMPA (purity 98.5%) were provided by the EPA Pesticide Standard Repository (Fort Meade, MD). The solvents used were pesticide-grade (nanograde, distilled in glass) acetonitrile (EMD OmniSolv; Merck, Darmstadt, Germany) and HPLC-grade water (EMD OmniSolv). Formic acid (88%) (J.T. Baker, Phillipsburg, NJ) diluted to 0.1% with HPLC-grade water was also used as a component of the LC mobile phase. Reagent-grade phosphoric acid (85%) (Ricca Chemical, Arlington, TX) was diluted with HPLC-grade water to create 30, 10 and 1% phosphoric acid solutions. For soil and vegetation analyses, [13C3 ], [15 N1 ] glyphosate and (d2) 13C, 15 N AMPA (Monsanto, St Louis, MO) were internal standards.
www.soci.org
www.soci.org
LJ Marek, WC Koskinen
Table 1. Soil physical and chemical properties Soil Composite Verndale Drummer
Texture
Type
Sandy loam Sandy loam Silty clay loam
Sand (%)
Mixed soils Mixed frigid Udic Argiboroll Typic Endoaquoll
75 80 19
Clay (%)
Organic carbon (%)
12 7 32
pH
0.75 1.4 4.1
7.1 6.1 5.7
Table 2. Mass spectrometry operating conditions for analysis of glyphosate and AMPA
Analyte
Data acquisition range (min)
Retention time (min)
Glyphosate
0.0–7.0
1.7
Internal standardb AMPA
3.5–15.0
12.0
Internal standardc a b c
Parent ion (Da) 167.8 167.8a 171.8 109.9 109.9 113.9
Daughter ion (Da) 62.8 149.9 62.8 62.8 80.9 80.8
Cone voltage (V) 20 20 20 30 30 30
Collision energy (eV) 22 11 22 20 13 13
Quantitation ions are in bold. [13C]3 , [15 N]1 glyphosate. (d2) 13C, 15 N AMPA.
tests. A quantity of 5 g of each sample was weighed into a 50 mL plastic centrifuge tube. Matrix vegetative spikes (similar samples previously shown to contain no glyphosate or AMPA) were prepared by adding 500 µL of the 5 mg L−1 standard, which produced a sample concentration of 0.5 mg kg−1 and an on-column injection amount of 1.5 ng. Reagent water (20 mL) was added to each centrifuge tube, and the tubes were briefly vortex mixed to homogenize the samples. If the entire amount of vegetation did not thoroughly mix with water, a known amount of additional water was added. Two 9.5 mm diameter stainless steel ball bearings were placed in each centrifuge tube, and samples were shaken for 3 min on a GenoGrinder (SPEX Sample Prep, Metuchen, NJ) at a speed of approximately 800 strokes min−1 . Samples were then centrifuged for 10 min at ∼5000 rpm or higher. Acidified HPLC-grade water (1.5 mL) and the centrifuged sample extract (0.50 mL) were added into 15 mL plastic centrifuge tubes (final volume 2 mL per 1.0% phosphoric acid). The tubes containing samples were capped and mixed with a vortex mixer. The acidified aqueous extracts were processed for analysis by being passed through IC-RP SPE cartridges/nylon syringe filters, as in the water analysis.
1160
2.5 Soil sample preparation Soils used in the method development were a Drummer silty clay loam surface soil from Oxford, Indiana, a Verndale sandy loam surface soil from Verndale, Minnesota, and a composite soil made from mixing together various Minnesota surface soils (Table 1). Soil samples were sieved and homogenized prior to processing, and then returned to the freezer if not immediately analyzed. Each soil sample (5 g) was weighed into a 50 mL plastic centrifuge tube. Matrix spikes (similar soil known to contain no glyphosate or AMPA) were prepared by adding the 50 mg L−1 standard (100 or 200 µL) to the soil (5 g), which produced a sample concentration of 1.0 or 2.0 mg kg−1 and an on-column concentration of 8.1 or 16.3 µg L−1 . To each soil sample, 10% phosphoric acid (20 mL) was added. After briefly mixing the soil, samples were placed on
wileyonlinelibrary.com/journal/ps
a shaker for ∼1 h. The samples were then centrifuged for 25 min at 2500 rpm or higher. The extraction solvent from each sample was poured into a second 50 mL labelled centrifuge tube. The soils were extracted with additional 10% phosphoric acid (10 mL) by shaking for ∼30 min, after which the tubes were centrifuged for 20 min at 2500 rpm or higher. The extraction solvents from the second wash were combined with the first extraction solvent, capped and vortex mixed. A 10.8 µg L−1 internal standard mix (3.8 mL) was added to the centrifuged extracts (0.2 mL) in plastic centrifuge tubes, and the solution was vortex mixed. Preconditioned IC-Chelate cartridges (Grace Maxi-CleanTM , 0.5 mL) were attached to 10 mL disposable syringes, and samples were passed through the cartridges and filters and dispensed back into the original 15 mL centrifuge tubes. An aliquot of the extracted solution was acidified with 30% phosphoric acid solution (final volume 2 mL per 1.0% phosphoric acid), and the acidified aqueous extracts were processed for analysis by being passed through IC-RP SPE cartridges/nylon syringe filters, as in the water analysis. 2.6 LC-MS/MS procedure A Waters AllianceTM liquid chromatograph (Milford, MA) coupled to a Waters Quattro LCTM mass spectrometer/mass spectrometer was employed for sample analysis. Method development was performed using a Bio-Rad Cation H microguard (Hercules, CA) in a Bio-Rad cartridge holder. An Opti-SolvTM minifilter (0.2 µm) was used to remove debris before the samples entered the guard cartridge. No additional column was used to achieve separation. For soil and vegetation, external/internal standard calibration was done by injecting each calibration standard by the same technique that was employed to introduce the samples into the LC-MS/MS. The instrument was recalibrated each day of use or, if the instrument was running the method the previous day/batch, a calibration check was used; if the response of the analyte of interest varied by more than 25%, a new calibration curve was run. Calibration checks or complete second curves were incorporated at the end of each run.
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
Pest Manag Sci 2014; 70: 1158–1164
Glyphosate and AMPA analysis in water, vegetation, and soil by LC–MS/MS
www.soci.org
(a) Water spike
100
%
0 0.00
Time 2.00
4.00
6.00
100
8.00
10.00
12.00
14.00
10.00
12.00
14.00
(b) Vegetation spike
%
0 0.00
2.00
4.00
6.00
100
Time
(c) Soil spike
%
0 0.00
8.00
AMPA
Glyphosate
2.00
4.00
6.00
100
8.00
10.00
12.00
14.00
10.00
12.00
14.00
Time
(d) Soil blank
%
0 0.00
2.00
4.00
6.00
8.00
Time
Figure 1. Chromatograms of glyphosate and AMPA in samples of water (15 µg L−1 ), vegetation (0.5 mg kg−1 ) and soil (2.0 mg kg−1 ) and a soil blank.
Analysis was performed according to the following parameters: acquisition mode MRM; photomultiplier voltage 750 V; quad temperature 120 ◦ C; oven temperature 50 ◦ C; autosampler temperature 5 ◦ C. The injection volume was 50 µL. The LC mobile phase was isocratic with 20:80 acetonitrile:0.1% formic acid with a flow rate of 0.5 mL min−1 .
3
RESULTS AND DISCUSSION
Pest Manag Sci 2014; 70: 1158–1164
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
wileyonlinelibrary.com/journal/ps
1161
3.1 LC-MS/MS For a simple and efficient method for routine analysis of glyphosate and its degradation product AMPA, LC-MS/MS is the instrument of choice as it provides confirmation and it is sufficiently sensitive for quantitation of glyphosate and AMPA at trace levels without derivatization. Derivatization methods, such as using FMOC, may increase the sensitivity of analyses, but they add cost and time to the analyses. Also, the derivatization technique developed for one matrix may not work with a different matrix. For instance, a method based on liquid chromatography coupled to fluorescence and electrospray ionization mass spectrometry (LC-FLD-ESI-MS) to determine glyphosate and AMPA in one matrix had to be adapted and modified (derivatization step changed) for use with a new matrix.19
For confirmation and quantitation, multiple reaction monitoring (MRM) was used for both glyphosate and AMPA, with two transitions for each: m/z 167.8 → 62.8 and m/z 167.8 → 149.9 for glyphosate, and m/z 109.9 → 62.8 and m/z 109.9 → 80.9 for AMPA (Table 2). For quantitation, m/z 167.8 → 149.9 was used for glyphosate and m/z 109.9 → 80.9 was used for AMPA. For soil analyses, [13C]3 , [15 N]1 glyphosate and (d2) 13C, 15 N AMPA internal standards overcame any possible matrix effects. For calibration standards in the concentration range 2.0–200 µg L−1 , the response was linear, R2 ≥ 0.997. Sample chromatograms (actually, overlays of the two different functions, so that both compounds can be seen in a simulated chromatogram) are shown in Fig. 1. Adequate separation of glyphosate and AMPA proved possible using only a guard column (Monsanto, unpublished report ‘Determination of glyphosate in ground, surface and finished drinking waters by LC-MS/MS’). Previous articles have shown that a variety of LC columns have been used for analyte separation, including anion-exchange,9 HILIC,17 multimode (reverse-phase and weak-anion-exchange packing)13 and reverse-phase columns.10,11 Although each column has worked for a specific matrix, other matrices may require different conditions or column packings for adequate separation of analytes. For instance, the length of a C18
www.soci.org
LJ Marek, WC Koskinen
Table 3. Recoveries of glyphosate and AMPA river waters
River water 1
Visual characterization
Chemical concentration (µg L−1 )
Number of samples
Glyphosate recovery ± standard deviation (%)
AMPA recovery ± standard deviation (%)
High sediment
5 15
1 1
89 109
120 103
High sediment, filtered
5 15 5 15 5 15 5 6 10 15 25 30 50 5 15 5 15 5 15 3
1 1 1 1 1 1 1 7 7 1 3 3 3 1 1 1 1 1 1 5
98 107 136 101 106 117 115 93 ± 7 106 ± 6 112 90 ± 8 105 ± 8 96 ± 2 107 108 117 112 108 116 98 ± 12
111 104 117 114 135 121 112 93 ± 3 107 ± 8 108 97 ± 3 100 ± 3 98 ± 3 114 97 105 101 92 111 105 ± 23
2
High sediment
3
High sediment
4
Medium sediment
5
Medium sediment
6
High algae
7
High algae
Lab DI water
column had to be increased for separation of glyphosate and AMPA in soil extracts as compared with other matrices.20 The present authors tried various columns, including a porous graphitic carbon column for polar compounds and an HILIC column, with limited success for all matrices; the peak shapes were not as sharp for glyphosate as with the present method, which used only a cation guard, which in turn increased the sensitivity of the method. In a simplified method, sample handling must be minimized: the more steps in the method, the more potential opportunities for losses or interferences. All objects that come into contact with the sample must be washed and solvent rinsed prior to being used for another sample. Very little sample processing was done in this procedure, so disposable lab ware could be used. Glyphosate adsorbs to glass surfaces, especially in the absence of a matrix, so plasticware used in this procedure minimized this problem. Losses of ∼20% of glyphosate and AMPA were reported when extraction glassware was not silanized.20
1162
3.2 Water The developed method for analysis of glyphosate and AMPA in diverse water samples when spiked at levels from 5 to 50 µg L−1 was quantitative: ≥88% for glyphosate and ≥92% for AMPA (Table 3). Analysis of glyphosate and AMPA was not affected by differences in river waters. Prefiltering of a river water sample containing a high level of sediment did not affect the recovery of glyphosate or AMPA. The average recovery of glyphosate from river water spiked at 5 and 15 µg L−1 was 99% with no prefiltering, as opposed to 103% with prefiltering (Table 3). For AMPA, recovery was 112% with no prefiltering and 108% with prefiltering. Therefore, to save time and money, aqueous samples were not prefiltered. The average recovery of glyphosate from
wileyonlinelibrary.com/journal/ps
samples of seven different rivers (with medium to high amounts of sediment and with low to high contents of algae) spiked at 5 µg L−1 was 110%, and recovery was 110% for samples spiked at 15 µg L−1 . For AMPA, recoveries were 113 and 107% for samples spiked at 5 and 15 µg L−1 respectively. Recoveries from laboratory deionized (DI) water spiked at 3 µg L−1 was 98% for glyphosate and 105% for AMPA. The method reporting level (MRL) for glyphosate in water was 3.04 µg L−1 , and 5.06 µg L−1 for AMPA. There are very few recently published methods for these two chemicals in water that do not require preconcentation of the analytes. The MRLs range from 1 to 320 µg L−1 for glyphosate and from 1 to 50 µg L−1 for AMPA.9,13 Analysis using an anion-exchange column with a coulometric detector has been reported to have lower levels of detection.7 Lower detection levels using methods that required preconcentration have also been reported for both glyphosate (
Revised: 8 November 2013
Accepted article published: 19 November 2013
Published online in Wiley Online Library: 26 December 2013
(wileyonlinelibrary.com) DOI 10.1002/ps.3684
Simplified analysis of glyphosate and aminomethylphosphonic acid in water, vegetation and soil by liquid chromatography–tandem mass spectrometry LeEtta J Mareka∗ and William C Koskinenb Abstract BACKGROUND: There is a need for a simple, fast, efficient and sensitive method for analysis of glyphosate and its degradate aminomethylphosphonic acid (AMPA) in diverse matrices such as water, vegetation and soil. RESULTS: Aqueous extracts from water, vegetation and soil were passed through reverse-phase and cation-exchange columns and directly injected into a tandem mass spectrometer using only a guard column for separation. Extraction efficiencies from the three matrices were >80% for both glyphosate and AMPA. The method reporting levels (MRLs) for glyphosate in water, vegetation and soil were 3.04 µg L−1 , 0.05 mg kg−1 and 0.37 mg kg−1 respectively. AMPA MRLs were 5.06 µg L−1 for water, 0.08 mg kg−1 for vegetation and 0.61 mg kg−1 for soil. CONCLUSIONS: A validated, simple and efficient liquid chromatography–tandem mass spectrometry (LC-MS/MS) method for routine analysis of glyphosate and AMPA in water, vegetation and soil that uses minimal sample handling and clean-up will facilitate the additional environmental research needed to address the continuing concerns related to increasing glyphosate use. Published 2013. This article is a U.S. Government work and is in the public domain in the USA. Keywords: glyphosate; AMPA; LC-MS/MS analysis; soil; water; vegetation
1
INTRODUCTION
1158
Glyphosate [N-(phosphonomethyl)glycine], a non-selective, broad-spectrum herbicide, has become the most widely used herbicide in the world. Since its introduction and subsequent commercialization in 1974, glyphosate has been considered to be a toxicologically and environmentally safe and effective herbicide. However, recently there have been increasing reports of incidences of glyphosate-resistant (GR) weeds, as well as weed shifts to naturally tolerant weed species. This has resulted in an increased number of glyphosate applications and increased rates of application for weed control, which in turn has raised concerns about adverse impacts of glyphosate use. For instance, several recently published papers have concluded that glyphosate adversely affects the mineral nutrition of GR crops, which has led to other problems; however, a recent review of the literature indicates that the majority of the literature shows no effect of glyphosate application on mineral nutrition of GR crops.1 The increasing applications and rates have also raised concerns about its environmental effects after long-term use. To address the continuing concerns related to increasing glyphosate use, additional research will continue to be conducted. To facilitate glyphosate research, simple, sensitive methods of analysis are needed for glyphosate in water, plant and soil samples. In addition, it has been suggested that the simple, sensitive methods developed for glyphosate can be used for analysis of chemical weapons such as phosphonic-acid-containing Pest Manag Sci 2014; 70: 1158–1164
nerve agents; glyphosate has been used as a surrogate for these chemicals.2 Numerous analytical techniques used to quantify glyphosate have proven to be reliable and sufficiently sensitive for most purposes. However, most of the methods are rather complex and time consuming, which leads to numerous ‘new’ or ‘revised’ methods being reported every year in the literature for water, plants or soil. The method modifications usually involve varying the details in the extraction and/or clean-up procedures, separation techniques or detection methods of previously published methods for each matrix, or the use of a previously published method with a new matrix. For instance, in the past 5 years alone, numerous ‘new’ methods have been published on the analysis of glyphosate and its principal degradate aminomethylphosphonic acid (AMPA) in water. They have included capillary electrophoresis (CE) with electrochemiluminescence detection,3 CE with contactless conductivity detection,4 CE with online isotachophoresis (ITP) pretreatment of waters on a column-coupling chip,5 capillary ion chromatography with a suppressed conductive detector,6
∗
Correspondence to: LeEtta J. Marek, Minnesota Department of Agriculture, St Paul, MN, USA. E-mail: [email protected]
a Minnesota Department of Agriculture, St Paul, MN, USA b USDA-ARS, St Paul, MN, USA
www.soci.org
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
Glyphosate and AMPA analysis in water, vegetation, and soil by LC–MS/MS anion-exchange chromatography with coulometric detection,7 high-performance ion chromatography coupled to inductively coupled plasma dynamic reaction cell mass spectrometry,8 high-performance liquid chromatography (HPLC) coupled to electrochemical9 and ultraviolet10 detectors and tandem mass spectrometry.11 Different online water preconcentration techniques have been reported,4,11 along with offline enrichment techniques such as using ion-exchange resins8 and solid-phase extraction discs.9 Clean-up procedures using cation-exchange resins have been reported.12 Liquid chromatography (LC) columns used for analyte separation include anion-exchange,9 reverse-phase10,11 and mixed-mode columns.13 To lower levels of detection in water, new pre- and post-column methods to derivatize glyphosate and AMPA have been reported.10,14 – 16 Fewer new methods have been reported in the past 5 years for more complex matrices such as vegetative material, soil and sewage sludge. These have involved minor variations of previously published methods or a new matrix: use of a hydrophilic interaction chromatography (HILIC) column with UV-visible or electrospray ionization (ESI)-MS/MS detection for analysis of glyphosate in selected fruits and vegetables,17 sample clean-up by protein precipitation for LC-MS/MS analysis of glyphosate and AMPA in soybean,18 optimization of derivatization methods using 9fluorenylmethylchloroformate (FMOC) for analysis of glyphosate and AMPA in maize by HPLC coupled to fluorescence and ESIMS,19 an optimized extraction procedure and chromatographic separation for analysis of glyphosate and AMPA in soil,20 online FMOC derivatization on a strong anion-exchange resin as a solid support for analysis of glyphosate and AMPA in sewage sludge21 and LC-ESI-MS or MS/MS analysis of wheat and rye22 and soybean.18 In spite of these recently published methods, which have primarily been developed for a single matrix, there is a need for a simplified method of analysis for glyphosate and AMPA that will work for multiple, diverse matrices. The objective of the present project was to develop and validate a simple and efficient LC-MS/MS method for routine analysis of glyphosate and its degradation product AMPA in water, vegetation and soil. The method uses minimal sample handling and clean-up and requires no derivatization. Samples of river water, aqueous plant extracts and acidic aqueous soil extracts were passed through IC-RP SPE cartridges into LC vials and injected into an LCMS/MS system with only a cation guard column for separation in the HPLC.
2
MATERIALS AND METHODS
Pest Manag Sci 2014; 70: 1158–1164
2.2 Standards Separate stock standard solutions (1.00 g L−1 ) of glyphosate and AMPA were prepared in plastic volumetric flasks using certified standards of known purity. An amount (∼0.0100 g) of standard was weighed, and an appropriate amount of freshly opened/poured HPLC-grade water was added to obtain a solution concentration of 1.00 g L−1 . The stock standards prepared in HPLC-grade water are reported to be stable for at least 1.3 years and up to 23 months for glyphosate and AMPA, respectively, if stored in a refrigerator (Monsanto, unpublished report ‘Determination of glyphosate in ground, surface and finished drinking waters by LC-MS/MS’; DuPont, 2008). Freshly opened/poured HPLC-grade water was used to prepare intermediate standards with 0.5, 5.0 and 50 mg L−1 concentrations of both glyphosate and AMPA for analysis of water, vegetation and soil. Soil and vegetation analyses additionally required two internal standard mixes which were prepared with [13C]3 , [15 N]1 glyphosate and (d2) 13C, 15 N AMPA obtained from Monsanto. HPLC-grade water was used to prepare 25 mL of a 50 mg L−1 mixed standard solution from which a 100 µL aliquot was diluted to 10 mL to obtain a 0.5 mg L−1 solution. This 0.5 mg L−1 internal standard mix was used to spike the samples as well as in the calibration standard. A series of calibration standards were used, ranging from 2 to 200 µg L−1 , which were prepared through dilution of intermediate standards with HPLC-grade water and 30% phosphoric acid to obtain a concentration of 1% phosphoric acid. Calibration solutions must be stored in a refrigerator in plastic containers until use. Plastic containers are employed to avoid possible chelation of the glyphosate with metal ions on the glassware. The calibration standards prepared in aqueous acid should be stable for 99 days. 2.3 Water sample preparation Surface water samples were taken from seven rivers in south central Minnesota. Visual evaluation of the river water samples showed that three of them had high concentrations of sediments, two had medium concentrations of sediments and two had high concentrations of algae. For each sample, including quality control (QC) samples, 15 mL plastic centrifuge tubes were used. To each spike control centrifuge tube, a known amount of intermediate standard (i.e. 200 µL of the 0.5 mg L−1 standard) was added, along with a known amount of HPLC-grade water (i.e. 4.8 mL). An aliquot of each water sample equal to the volume of the standard (5 mL) was transferred into a centrifuge tube. An equal volume of HPLC-grade water was added to a centrifuge tube as a blank. Each centrifuge tube was acidified with concentrated phosphoric acid (85%) (final volume 5 mL per 1.0% phosphoric acid), capped and mixed with a vortex mixer. Samples were then passed through IC-RP SPE cartridges (Grace Maxi-CleanTM , 0.5 mL; Deerfield, IL) attached to 3 mL disposable syringes with nylon syringe filters (Acrodisc, 13 mm × 0.2 µm; Pall Life Sciences, Ann Arbor, MI,) connected to the outlet of the cartridge and collected in polypropylene autosampler vials with preslit PTFE/silicone septa. Samples were stored in the refrigerator or chilled autosampler until LC-MS/MS analysis. 2.4 Vegetation sample preparation For method development, vegetative samples from diverse sources were mixed and then ground and homogenized using a Robot CoupeTM (Jackson, MS) sample processor and returned to the freezer if not immediately analyzed. Additionally, samples of individual plants were ground and homogenized for recovery
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
wileyonlinelibrary.com/journal/ps
1159
2.1 Chemicals/reagents Glyphosate (purity 99.8%) and AMPA (purity 98.5%) were provided by the EPA Pesticide Standard Repository (Fort Meade, MD). The solvents used were pesticide-grade (nanograde, distilled in glass) acetonitrile (EMD OmniSolv; Merck, Darmstadt, Germany) and HPLC-grade water (EMD OmniSolv). Formic acid (88%) (J.T. Baker, Phillipsburg, NJ) diluted to 0.1% with HPLC-grade water was also used as a component of the LC mobile phase. Reagent-grade phosphoric acid (85%) (Ricca Chemical, Arlington, TX) was diluted with HPLC-grade water to create 30, 10 and 1% phosphoric acid solutions. For soil and vegetation analyses, [13C3 ], [15 N1 ] glyphosate and (d2) 13C, 15 N AMPA (Monsanto, St Louis, MO) were internal standards.
www.soci.org
www.soci.org
LJ Marek, WC Koskinen
Table 1. Soil physical and chemical properties Soil Composite Verndale Drummer
Texture
Type
Sandy loam Sandy loam Silty clay loam
Sand (%)
Mixed soils Mixed frigid Udic Argiboroll Typic Endoaquoll
75 80 19
Clay (%)
Organic carbon (%)
12 7 32
pH
0.75 1.4 4.1
7.1 6.1 5.7
Table 2. Mass spectrometry operating conditions for analysis of glyphosate and AMPA
Analyte
Data acquisition range (min)
Retention time (min)
Glyphosate
0.0–7.0
1.7
Internal standardb AMPA
3.5–15.0
12.0
Internal standardc a b c
Parent ion (Da) 167.8 167.8a 171.8 109.9 109.9 113.9
Daughter ion (Da) 62.8 149.9 62.8 62.8 80.9 80.8
Cone voltage (V) 20 20 20 30 30 30
Collision energy (eV) 22 11 22 20 13 13
Quantitation ions are in bold. [13C]3 , [15 N]1 glyphosate. (d2) 13C, 15 N AMPA.
tests. A quantity of 5 g of each sample was weighed into a 50 mL plastic centrifuge tube. Matrix vegetative spikes (similar samples previously shown to contain no glyphosate or AMPA) were prepared by adding 500 µL of the 5 mg L−1 standard, which produced a sample concentration of 0.5 mg kg−1 and an on-column injection amount of 1.5 ng. Reagent water (20 mL) was added to each centrifuge tube, and the tubes were briefly vortex mixed to homogenize the samples. If the entire amount of vegetation did not thoroughly mix with water, a known amount of additional water was added. Two 9.5 mm diameter stainless steel ball bearings were placed in each centrifuge tube, and samples were shaken for 3 min on a GenoGrinder (SPEX Sample Prep, Metuchen, NJ) at a speed of approximately 800 strokes min−1 . Samples were then centrifuged for 10 min at ∼5000 rpm or higher. Acidified HPLC-grade water (1.5 mL) and the centrifuged sample extract (0.50 mL) were added into 15 mL plastic centrifuge tubes (final volume 2 mL per 1.0% phosphoric acid). The tubes containing samples were capped and mixed with a vortex mixer. The acidified aqueous extracts were processed for analysis by being passed through IC-RP SPE cartridges/nylon syringe filters, as in the water analysis.
1160
2.5 Soil sample preparation Soils used in the method development were a Drummer silty clay loam surface soil from Oxford, Indiana, a Verndale sandy loam surface soil from Verndale, Minnesota, and a composite soil made from mixing together various Minnesota surface soils (Table 1). Soil samples were sieved and homogenized prior to processing, and then returned to the freezer if not immediately analyzed. Each soil sample (5 g) was weighed into a 50 mL plastic centrifuge tube. Matrix spikes (similar soil known to contain no glyphosate or AMPA) were prepared by adding the 50 mg L−1 standard (100 or 200 µL) to the soil (5 g), which produced a sample concentration of 1.0 or 2.0 mg kg−1 and an on-column concentration of 8.1 or 16.3 µg L−1 . To each soil sample, 10% phosphoric acid (20 mL) was added. After briefly mixing the soil, samples were placed on
wileyonlinelibrary.com/journal/ps
a shaker for ∼1 h. The samples were then centrifuged for 25 min at 2500 rpm or higher. The extraction solvent from each sample was poured into a second 50 mL labelled centrifuge tube. The soils were extracted with additional 10% phosphoric acid (10 mL) by shaking for ∼30 min, after which the tubes were centrifuged for 20 min at 2500 rpm or higher. The extraction solvents from the second wash were combined with the first extraction solvent, capped and vortex mixed. A 10.8 µg L−1 internal standard mix (3.8 mL) was added to the centrifuged extracts (0.2 mL) in plastic centrifuge tubes, and the solution was vortex mixed. Preconditioned IC-Chelate cartridges (Grace Maxi-CleanTM , 0.5 mL) were attached to 10 mL disposable syringes, and samples were passed through the cartridges and filters and dispensed back into the original 15 mL centrifuge tubes. An aliquot of the extracted solution was acidified with 30% phosphoric acid solution (final volume 2 mL per 1.0% phosphoric acid), and the acidified aqueous extracts were processed for analysis by being passed through IC-RP SPE cartridges/nylon syringe filters, as in the water analysis. 2.6 LC-MS/MS procedure A Waters AllianceTM liquid chromatograph (Milford, MA) coupled to a Waters Quattro LCTM mass spectrometer/mass spectrometer was employed for sample analysis. Method development was performed using a Bio-Rad Cation H microguard (Hercules, CA) in a Bio-Rad cartridge holder. An Opti-SolvTM minifilter (0.2 µm) was used to remove debris before the samples entered the guard cartridge. No additional column was used to achieve separation. For soil and vegetation, external/internal standard calibration was done by injecting each calibration standard by the same technique that was employed to introduce the samples into the LC-MS/MS. The instrument was recalibrated each day of use or, if the instrument was running the method the previous day/batch, a calibration check was used; if the response of the analyte of interest varied by more than 25%, a new calibration curve was run. Calibration checks or complete second curves were incorporated at the end of each run.
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
Pest Manag Sci 2014; 70: 1158–1164
Glyphosate and AMPA analysis in water, vegetation, and soil by LC–MS/MS
www.soci.org
(a) Water spike
100
%
0 0.00
Time 2.00
4.00
6.00
100
8.00
10.00
12.00
14.00
10.00
12.00
14.00
(b) Vegetation spike
%
0 0.00
2.00
4.00
6.00
100
Time
(c) Soil spike
%
0 0.00
8.00
AMPA
Glyphosate
2.00
4.00
6.00
100
8.00
10.00
12.00
14.00
10.00
12.00
14.00
Time
(d) Soil blank
%
0 0.00
2.00
4.00
6.00
8.00
Time
Figure 1. Chromatograms of glyphosate and AMPA in samples of water (15 µg L−1 ), vegetation (0.5 mg kg−1 ) and soil (2.0 mg kg−1 ) and a soil blank.
Analysis was performed according to the following parameters: acquisition mode MRM; photomultiplier voltage 750 V; quad temperature 120 ◦ C; oven temperature 50 ◦ C; autosampler temperature 5 ◦ C. The injection volume was 50 µL. The LC mobile phase was isocratic with 20:80 acetonitrile:0.1% formic acid with a flow rate of 0.5 mL min−1 .
3
RESULTS AND DISCUSSION
Pest Manag Sci 2014; 70: 1158–1164
Published 2013. This article is a U.S. Government work and is in the public domain in the USA.
wileyonlinelibrary.com/journal/ps
1161
3.1 LC-MS/MS For a simple and efficient method for routine analysis of glyphosate and its degradation product AMPA, LC-MS/MS is the instrument of choice as it provides confirmation and it is sufficiently sensitive for quantitation of glyphosate and AMPA at trace levels without derivatization. Derivatization methods, such as using FMOC, may increase the sensitivity of analyses, but they add cost and time to the analyses. Also, the derivatization technique developed for one matrix may not work with a different matrix. For instance, a method based on liquid chromatography coupled to fluorescence and electrospray ionization mass spectrometry (LC-FLD-ESI-MS) to determine glyphosate and AMPA in one matrix had to be adapted and modified (derivatization step changed) for use with a new matrix.19
For confirmation and quantitation, multiple reaction monitoring (MRM) was used for both glyphosate and AMPA, with two transitions for each: m/z 167.8 → 62.8 and m/z 167.8 → 149.9 for glyphosate, and m/z 109.9 → 62.8 and m/z 109.9 → 80.9 for AMPA (Table 2). For quantitation, m/z 167.8 → 149.9 was used for glyphosate and m/z 109.9 → 80.9 was used for AMPA. For soil analyses, [13C]3 , [15 N]1 glyphosate and (d2) 13C, 15 N AMPA internal standards overcame any possible matrix effects. For calibration standards in the concentration range 2.0–200 µg L−1 , the response was linear, R2 ≥ 0.997. Sample chromatograms (actually, overlays of the two different functions, so that both compounds can be seen in a simulated chromatogram) are shown in Fig. 1. Adequate separation of glyphosate and AMPA proved possible using only a guard column (Monsanto, unpublished report ‘Determination of glyphosate in ground, surface and finished drinking waters by LC-MS/MS’). Previous articles have shown that a variety of LC columns have been used for analyte separation, including anion-exchange,9 HILIC,17 multimode (reverse-phase and weak-anion-exchange packing)13 and reverse-phase columns.10,11 Although each column has worked for a specific matrix, other matrices may require different conditions or column packings for adequate separation of analytes. For instance, the length of a C18
www.soci.org
LJ Marek, WC Koskinen
Table 3. Recoveries of glyphosate and AMPA river waters
River water 1
Visual characterization
Chemical concentration (µg L−1 )
Number of samples
Glyphosate recovery ± standard deviation (%)
AMPA recovery ± standard deviation (%)
High sediment
5 15
1 1
89 109
120 103
High sediment, filtered
5 15 5 15 5 15 5 6 10 15 25 30 50 5 15 5 15 5 15 3
1 1 1 1 1 1 1 7 7 1 3 3 3 1 1 1 1 1 1 5
98 107 136 101 106 117 115 93 ± 7 106 ± 6 112 90 ± 8 105 ± 8 96 ± 2 107 108 117 112 108 116 98 ± 12
111 104 117 114 135 121 112 93 ± 3 107 ± 8 108 97 ± 3 100 ± 3 98 ± 3 114 97 105 101 92 111 105 ± 23
2
High sediment
3
High sediment
4
Medium sediment
5
Medium sediment
6
High algae
7
High algae
Lab DI water
column had to be increased for separation of glyphosate and AMPA in soil extracts as compared with other matrices.20 The present authors tried various columns, including a porous graphitic carbon column for polar compounds and an HILIC column, with limited success for all matrices; the peak shapes were not as sharp for glyphosate as with the present method, which used only a cation guard, which in turn increased the sensitivity of the method. In a simplified method, sample handling must be minimized: the more steps in the method, the more potential opportunities for losses or interferences. All objects that come into contact with the sample must be washed and solvent rinsed prior to being used for another sample. Very little sample processing was done in this procedure, so disposable lab ware could be used. Glyphosate adsorbs to glass surfaces, especially in the absence of a matrix, so plasticware used in this procedure minimized this problem. Losses of ∼20% of glyphosate and AMPA were reported when extraction glassware was not silanized.20
1162
3.2 Water The developed method for analysis of glyphosate and AMPA in diverse water samples when spiked at levels from 5 to 50 µg L−1 was quantitative: ≥88% for glyphosate and ≥92% for AMPA (Table 3). Analysis of glyphosate and AMPA was not affected by differences in river waters. Prefiltering of a river water sample containing a high level of sediment did not affect the recovery of glyphosate or AMPA. The average recovery of glyphosate from river water spiked at 5 and 15 µg L−1 was 99% with no prefiltering, as opposed to 103% with prefiltering (Table 3). For AMPA, recovery was 112% with no prefiltering and 108% with prefiltering. Therefore, to save time and money, aqueous samples were not prefiltered. The average recovery of glyphosate from
wileyonlinelibrary.com/journal/ps
samples of seven different rivers (with medium to high amounts of sediment and with low to high contents of algae) spiked at 5 µg L−1 was 110%, and recovery was 110% for samples spiked at 15 µg L−1 . For AMPA, recoveries were 113 and 107% for samples spiked at 5 and 15 µg L−1 respectively. Recoveries from laboratory deionized (DI) water spiked at 3 µg L−1 was 98% for glyphosate and 105% for AMPA. The method reporting level (MRL) for glyphosate in water was 3.04 µg L−1 , and 5.06 µg L−1 for AMPA. There are very few recently published methods for these two chemicals in water that do not require preconcentation of the analytes. The MRLs range from 1 to 320 µg L−1 for glyphosate and from 1 to 50 µg L−1 for AMPA.9,13 Analysis using an anion-exchange column with a coulometric detector has been reported to have lower levels of detection.7 Lower detection levels using methods that required preconcentration have also been reported for both glyphosate (
