U.S. Geological Survey
|
Multiply | By | To obtain |
---|---|---|
cubic centimeter (cm³) | 0.06102 | cubic inch |
gram (g) | 2.205 x 10-3 | pound |
liter (L) | 2.642 x 10-1 | gallon |
meter (m) | 3.281 | foot |
microliter (µL) | 2.642 x 10-7 | gallon |
micrometer (µm) | 3.937 x 10-5 | inch |
milligram (mg) | 3.53 x 10-5 | ounce |
millimeter (mm) | 3.937 x 10-2 | inch |
ounce (oz) | 0.02957 | liter |
pound per square inch (lb/in²) | 6.895 | kilopascal |
Miscellaneous Abbreviations |
---|
mass to charge (m/z) minute (min) second (sec) volt (V) |
Abbreviated Water-Quality Units |
liter per minute (L/min) microgram per liter (µg/L) microgram per milliliter (µg/L) milligram per milliliter (mg/mL) mililiter (mL) milliliter per minute (mL/min) millimoles (mM) |
An analytical method for the determination of glyphosate, its principal degradation compound, aminomethylphosphonic acid (AMPA), and glufosinate in water with varying matrices has been developed. Four different sample matrices fortified at 0.2 and 2.0 µg/L (micrograms per liter) were analyzed using precolumn derivatization with 9-fluorenylmethylchloroformate (FMOC). After derivatization, cleanup and concentration were accomplished using automated online solid-phase extraction followed by elution with the mobile phase allowing for direct injection into a liquid chromatograph/mass spectrometer (LC/MS). Analytical conditions for MS detection were optimized, and quantitation was carried out using the following representative ions: 390 and 168 for glyphosate; 332, 110, and 136 for AMPA; and 402, 180, and 206 for glufosinate. Matrix effects were minimized by utilizing standard addition for quantification and an isotope-labeled glyphosate (2-13C,15N) as the internal standard. Method detection limits (MDLs) were 0.084 µg/L for glyphosate, 0.078 µg/L for AMPA, and 0.057 µg/L for glufosinate. The method reporting limits (MRLs) were set at 0.1 µg/L for all three compounds. The mean recovery values ranged from 88.0 to 128.7 percent, and relative standard deviation values ranged from 5.6 to 32.6 percent.
Glyphosate [N-(phosphonomethyl)glycine] is a broad-spectrum, nonselective, postemergence herbicide that is used extensively in the United States in various applications for weed and vegetation control. Aminomethylphosphonic acid (AMPA) is a degradation product of glyphosate. Glufosinate [ammonium DL-homoalanin-4-(methyl)phosphinate] is similar to glyphosate in chemical structure and use.
The three compounds are very polar and highly soluble in water. The detection of these compounds requires the use of a derivatization step. Published methods outline the use of precolumn derivatization using 9-fluorenylmethylchloroformate (FMOC) coupled with high-performance liquid chromatography (HPLC) with fluorescence detection (Spark-Holland, 1996) or tandem mass spectrometry (MS/MS) detection (Vreeken and others, 1998). Fluorescence detection has sensitivity but lacks specificity, and the MS/MS method can be subject to matrix variation in derivatization and fragmentation.
Utilization of isotope-labeled (2-13C,15N) glyphosate as an internal standard carried through all steps of a method developed by the U.S. Geological Survey (USGS) Organic Geochemistry Research Group in Lawrence, Kansas, addresses the variations in derivatization and fragmentation for the analysis of glyphosate. Analysis of samples using the standard-addition method, adding a known amount of standard(s) to a matching replicate of each unknown sample further addresses matrix variations for glyphosate, AMPA, and glufosinate.
The method of analysis described in this report has been assigned the USGS method code "0-2136-01." This unique code represents the automated method of analysis as it is described in the report and can be used to identify the method.
This report provides a detailed description of the method, including the apparatus, reagents, instrument calibration, and the solid-phase extraction (SPE) procedure required for sample analysis. Method detection limits (MDLs), mean extraction recoveries, and relative standard deviations for the method also are presented.
The method using HPLC/MS and online SPE is suitable for the determination of low concentrations (in micrograms per liter) of glyphosate, AMPA, and glufosinate in water samples. Associated molecular weights and USGS parameter codes for these compounds and their derivatized compounds are listed in table 1. Because suspended particulate matter is removed by filtration, the method is suitable only for dissolved-phase compounds. The calibration range for the method is equivalent to concentrations from 0.1 to 2.0 µg/L without dilution.
[FMOC, 9-fluorenylmethylchloroformate; --, not applicable]
Compound | Molecular weight (atomic mass units) |
Parameter code |
---|---|---|
Glyphosate | 169.1 | 62722T |
Glyphosate-FMOC | 391.3 | -- |
Isotope-labeled glyphosate | 171.1 | -- |
Isotope-labeled glyphosate-FMOC | 393.3 | -- |
Aminomethylphosphonic acid | 111.0 | 62649T |
Aminomethylphosphonic acid-FMOC | 333.3 | -- |
Glufosinate | 181.1 | 62721T |
Glufosinate-FMOC | 403.4 | -- |
Cysteic acid | 169.2 | -- |
Cysteic acid-FMOC | 391.4 | -- |
Water samples are filtered at the collection site using glass-fiber filters with nominal 0.70-µm pore diameter to remove suspended particulate matter. In the laboratory, 10 mL of sample(s) are dispensed into two labeled, 19-mL, screw-capped plastic tubes. The sample in the tube labeled "standard addition" is fortified with 1 µg/L of each compound to be analyzed. Internal standard solutions are added to both tubes, the sample is buffered to pH 9.0 by adding borate buffer, and after mixing, a solution of FMOC is added to all tubes. Derivatization is carried out in the dark in a water bath at 40 °C. After 24 hours, the reaction is stopped and stabilized by adding 2-percent phosphoric acid. All tubes are stored in the dark until analyzed.
A 5.5-mL aliquot of each sample and the matching standard-addition sample are diluted 1:1 with reagent water in autosampler vials, capped, and placed in the tray of the autosampler. The autosampler provides the sample loading for the automated, online SPE system. The SPE cartridge is conditioned with methanol and reagent water. Ten milliliters (10 mL) of the diluted sample are loaded onto the cartridge. After a 500-µL reagent-water rinse, the cartridge is placed into the flow path of the liquid chromatograph (LC) preceding the column. The conditions and gradient of the mobile phase are set to elute the compounds of interest and leave the excess derivatization reagent on the cartridge.
The sample's compounds are separated by the LC column and detected by the mass spectrometer (MS). Compounds are identified by comparing retention times with the retention times of the standard-addition sample and further by comparision of the selected fragment ions. The concentration of each compound is calculated by determining the ratio of the compound to the internal standard to the ratio of the same compound in the standard-addition sample minus the ratio of the sample. The sample and standard-addition sample are analyzed sequentially, using the same method and instruments.
Sampling methods capable of collecting water samples that accurately represent the water-quality characteristics of the ground water or surface water at a specific time or location are used. Detailed descriptions of sampling methods for obtaining ground-water samples are given in Hardy and others (1989). Similar descriptions of sampling methods used by the USGS for obtaining depth- and width-integrated surface-water samples are given in Edwards and Glysson (1988) and Ward and Harr (1990).
Sample-collection equipment must be free of tubing, gaskets, and other components made of nonfluorinated plastic material that might leach interfering compounds into water samples or absorb the compounds from the water. The water samples from each site are composited in a single container and filtered through a nominal 0.7-µm glass-fiber filter using a peristaltic pump. Filters are preconditioned with about 200 mL of sample prior to filtration of the sample. The filtrate for analysis is collected in baked, 125-mL amber glass bottles with Teflon-lined lids. Samples are chilled immediately and shipped to the laboratory within 3 days of collection. At the laboratory, samples are logged in, assigned identification numbers, and refrigerated at 4 °C until derivatized and analyzed.
Background absorbance readings, peak shape, and system pressure are used to evaluate LC performance. Background absorbance readings should remain stable and low and indicate that the LC column has equilibrated with the mobile-phase flow. If peak shape deteriorates, the columns may need to be replaced. If the pressure reading is high, there may be a clog in the mobile-phase flow path, or the column-compartment thermostat may not have reached the required temperature.
Mass spectrometer performance is evaluated by assessing isotopic ratios and abundance. The MS is tuned in electrospray, negative-ion mode before each HPLC/MS analytical run using the solutions, procedure, and software supplied by the manufacturer.
A calibration table is prepared from an analyzed standard using the LC/MSD Chemstation software (Hewlett Packard, Wilmington, Delaware). Manufactures' instructions are followed for using the internal standards as time references and for quantitation.
The relative retention time (RRTc) is calculated for each selected compound in the calibration solution or in a sample as follows:
(1),
RRTc = RTc/RTi,
where RTc = uncorrected retention time of the selected compound, and
where RTi = uncorrected retention time of the internal standard.
See table 2 for retention times, relative retention times, molecular, and fragment ions.
[m/z, mass-to-charge ratio; FMOC, 9-fluorenylmethylchloroformate; --, not applicable]
Compound | Retention time (minutes) |
Relative retention time (minutes) |
Molecular ion (m/z) |
Fragment ion 1 (m/z) |
Fragment ion 2 (m/z) |
|||||
---|---|---|---|---|---|---|---|---|---|---|
Compounds (in order of increasing retention time) | ||||||||||
Glyphosate-FMOC | 11.22 | 1.00 | 390 | 168 | -- | |||||
Glufosinate-FMOC | 14.35 | 1.28 | 402 | 180 | 206 | |||||
Aminomethylphosphonic acid-FMOC | 16.31 | 1.45 | 332 | 110 | 136 | |||||
Internal standards | ||||||||||
Isotope-labeled glyphosate-FMOC | 11.22 | 1.00 | 392 | 170 | -- | |||||
Cysteic acid-FMOC | 14.16 | 1.00 | 390 | 168 | -- |
The expected retention time (RT) of the peak of the selected compound needs to be within ±2 percent of the expected retention time on the basis of the RRTc obtained from the internal-standard analysis. The expected retention time is calculated as follows:
The samples are derivatized upon arrival in the laboratory, then stored in a refrigerator in the dark until analyzed on the instruments.
The LC/MSD Chemstation software (Hewlett Packard, Wilmington, Delaware) is used with the previously prepared calibration table for identification of compounds.
If a selected compound has passed the qualitative identification criteria, the concentration in the sample is calculated as follows:
(3),
C = [ Ac / Ai / [ Acsa / Aisa - Ac/Ai ]] x SAcx DF,
where C = concentration of the selected compound in the sample, in micrograms per liter;
where Ac = area of peak of the (molecular or fragment) ion for the selected compound;
where Ai = area of peak of the molecular ion for the internal standard;
where Acsa = area of peak of the (molecular or fragment) for the selected compound in the standard-addition sample;
where Aisa = area of peak of the molecular ion for the internal standard for the standard-addition sample;
where SAc = concentration of standard addition; and
where DF = dilution factor, for samples that have exceeded upper range of method.
Glyphosate, AMPA, and glufosinate are reported in concentrations ranging from 0.1 to 2.0 µg/L. If the concentration is greater than 2.0 µg/L, a portion of the original sample is diluted appropriately with reagent water and reanalyzed through the entire procedure.
A reagent-water sample, a ground-water sample collected from a well in Sedgwick County, Kansas, a surface-water sample from the Kisco River below Mt. Kisco, New York, and a surface-water sample from the spillway below Clinton Lake in Kansas were used to test the performance of method 0-2136-01. All samples were filtered through a nominal 0.7-µm glass-fiber filter and stored at 4 °C.
Samples of each matrix were spiked with glyphosate, AMPA, and glufosinate to concentrations of 0.2 and 2.0 µg/L and analyzed on different days during August 2001. In addition, unspiked samples of each matrix were analyzed. Comparisions of the different matrices and concentrations included bias from day-to-day variations. Method recoveries from the analyses are included in tables 3, 4, 5, and 6.
[µg/L, microgram per liter]
Eight samples spiked at 0.2 µg/L | Eight samples spiked at 2.0 µg/L | |||||||
---|---|---|---|---|---|---|---|---|
Mean recovery | Mean recovery | |||||||
Compound | (µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
(µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
Glyphosate | 0.199 | 100 | 0.028 | 14 | 2.14 | 107 | 0.19 | 9 |
Aminomethylphosphonic acid | .224 | 112 | .026 | 12 | 2.41 | 120 | .38 | 16 |
Glufosinate | .220 | 110 | .019 | 9 | 2.57 | 129 | .18 | 7 |
Average | .214 | 107 | .024 | 11 | 2.37 | 119 | .25 | 11 |
[µg/L, microgram per liter]
Eight samples spiked at 0.2 µg/L | Eight samples spiked at 2.0 µg/L | |||||||
---|---|---|---|---|---|---|---|---|
Mean recovery | Mean recovery | |||||||
Compound | (µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
(µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
Glyphosate | 0.194 | 97 | 0.040 | 21 | 2.04 | 102 | 0.46 | 23 |
Aminomethylphosphonic acid | .228 | 114 | .054 | 24 | 2.24 | 112 | .62 | 28 |
Glufosinate | .220 | 110 | .037 | 17 | 2.53 | 126 | .45 | 18 |
Average | .214 | 107 | .044 | 20 | 2.27 | 113 | .51 | 23 |
[µg/L, microgram per liter]
Eight samples spiked at 0.2 µg/L | Eight samples spiked at 2.0 µg/L | |||||||
---|---|---|---|---|---|---|---|---|
Mean recovery | Mean recovery | |||||||
Compound | (µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
(µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
Glyphosate | 0.188 | 94 | 0.040 | 21 | 1.93 | 96 | 0.49 | 26 |
Aminomethylphosphonic acid | .223 | 117 | .043 | 18 | 2.23 | 111 | .56 | 25 |
Glufosinate | .235 | 118 | .044 | 19 | 2.52 | 126 | .82 | 33 |
Average | .219 | 109 | .042 | 19 | 2.22 | 111 | .62 | 28 |
[µg/L, microgram per liter]
Eight samples spiked at 0.2 µg/L | Eight samples spiked at 2.0 µg/L | |||||||
---|---|---|---|---|---|---|---|---|
Mean recovery | Mean recovery | |||||||
Compound | (µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
(µg/L) | (percent) | Standard deviation (µg/L) |
Relative standard deviation (percent) |
Glyphosate¹ | 0.176 | 88 | 0.044 | 25 | 2.32 | 116 | 0.43 | 19 |
Aminomethylphosphonic acid | .204 | 102 | .038 | 19 | 2.02 | 101 | .29 | 14 |
Glufosinate | .208 | 104 | .012 | 6 | 2.32 | 116 | .37 | 16 |
Average | .196 | 98 | .031 | 16 | 2.22 | 111 | .36 | 16 |
The reagent-water sample, ground-water sample, and surface-water sample from the Kisco River did not require correction for background concentrations of glyphosate, AMPA, and glufosinate. The surface-water sample from Clinton Lake contained glyphosate at 0.31 µg/L (table 6) but did not contain AMPA or glufosinate. The data from this water sample were corrected for the background concentration of glyphosate.
A method detection limit (MDL) is defined as the minimum concentration of a substance that can be identified, measured, and reported with a 99-percent confidence that the compound concentration is greater than zero. MDLs were determined according to procedures outlined by the U.S. Environmental Protection Agency (1992). Eight replicate samples of buffered reagent water spiked with 0.20 µg/L of each of the compounds were analyzed to determine MDLs (table 7). Each sample was analyzed on different days during August 2001 so that day-to-day variation is included in the results.
[µg/L, microgram per liter]
Eight samples spiked at 0.2 µg/L | |||
---|---|---|---|
Compound | Mean concentrations (µg/L) |
Standard deviation (µg/L) |
Method detection limit (µ) |
Glyphosate | 0.199 | 0.028 | 0.084 |
Aminomethylphosphonic acid | .224 | .026 | .078 |
Glufosinate | .220 | .019 | .057 |
Minimum | .057 | ||
Maximum | .084 |
The MDL was calculated using the following equation:
(4),
MDL = (S)t(n-1,1-a)=0.99) ,
where S = standard deviation of replicate analysis, in micrograms per liter, at the spiked concentration;
where t(n-1, 1-a= 0.99) = Student's t-value for the 99-percent confidence level with n-1 degrees of freedom (U.S. Environmental Protection Agency, 1992); and
where n = number of replicate analyses.
The estimated mean MDL was 0.084 µg/L for glyphosate, 0.078 µg/L for AMPA, and 0.057 µg/L for glufosinate (table 7). According to the U.S. Environmental Protection Agency (1992) procedure, the spiked concentrations should be no more than five times the estimated MDL. The spiked concentrations were within five times the MDL.
Mean recoveries in reagent-water, ground-water, and surface-water samples were determined by comparing the mean analyzed concentration (see "Quantitation" section) from the eight replicate samples to the spiked concentration. Mean recoveries were highest overall in reagent-water samples at the 2.0-µg/L level and lowest overall in the Clinton Lake water at the 0.2-µg/L level. Relative standard deviations of the recoveries ranged from 5.8 to 32.6 percent.
An HPLC method utilizing online SPE and fluorescence detection was reported by Spark-Holland (1996) as being quite sensitive but lacking the confirmation available using mass spectrometry. Also, the mobile phase used was not readily compatible with most mass spectrometers currently (2001) in use. The use of disposable SPE cartridges, as outlined by Spark-Holland (1996), improves LC column life and reduces the possiblity of carryover.
An HPLC/MS/MS method utilizing online SPE was reported by Vreeken and others (1998). This HPLC/MS/MS method used the same column repeatedly for cleanup and lacked the accuracy afforded by the use of internal standards. Both of the HPLC and HPLC/MS/MS methods used precolumn derivatization with FMOC.
An GC/MS/MS method utilizng ion-exchange chromatography followed by derivatization was reported by Royer and others (2000). The GC/MS/MS method required elaborate preparation of columns and an elaborate derivatization procedure.
The incorporation of an isotope-labeled (2-13C,15N) glyphosate as an internal standard carried through the derivatization, extraction, separation, and detection steps enhanced the reproducibility and accuracy of the online SPE and HPLC/MS method described in this report. The use of standard addition for quantitation overcomes the variation in derivatization and fragmentation observed from analyzing compounds with different matrices. The use of the more readily available single quadrapole mass spectrometer allowed for simpler and less-expense operation.
Figure 1 shows a total ion chromatogram of a FMOC-derivatized, 2.0-µg/L spiked surface-water sample from Clinton Lake. Figure 2 shows the chromatograms of the FMOC-derivatized ions of each compound. Baselines are relatively clean, and separations are adequate for quantitation. The cysteic acid internal standard, which was used before the labeled glyphosate became available, has been left in the method as a retention-time reference.
The method described in this report provides for routine analysis of glyphosate, AMPA, and glufosinate in environmental water samples. Derivatization with FMOC, online SPE, and HPLC/MS are shown to be a reliable and sensitive method for low concentrations.
Good precision and accuracy for the analysis of glyphosate, AMPA, and glufosinate were demonstrated for reagent water, ground water, and surface water. Method detection limits were 0.084 µg/L for glyphosate, 0.078 µg/L for AMPA, and 0.057 µg/L for glufosinate. The mean recoveries of water samples spiked at 0.2 and 2.0 µg/L ranged from 88.0 to 128.7 percent with relative standard deviations ranging from 5.8 to 32.6 percent.
Information about the fate and transport of glyphosate, its degradation compound AMPA, and glufosinate in water can be acquired from the analysis of ground water and surface water. This method also can be used for water-quality determinations.
Appendix 1. Programmed steps for autosampler
Glyphosate method 0-2136-01
Basic system parameters:
Step | Action | Function | Notes |
---|---|---|---|
1 | Wait | 5 sec | |
2 | Aux port 3 | OFF | Unfreezes online SPE instrument. |
3 | Compressor | ON | |
4 | Syringe valve | Needle | |
5 | Injection valve | Load | |
6 | Aspirate | 10,000-µL sample-speed 5-height 4 mm | Draws sample into sample loop. |
7 | Wait for input 2 | Low | Waits for signal from online SPE instrument that cartridge preparation is done. |
8 | Injection valve | Inject | Puts sample loop into cartridge path. |
9 | Dispense | 2,000-µL waste-speed 4 | Empties excess sample from buffer loop to waste. |
10 | Syringe valve | Wash | |
11 | Aspirate | 2,000-µL wash-speed 5 | Refills syringe with wash solution. |
12 | Syringe valve | Needle | |
13 | Wait | 10 sec | |
14 | Wait for input 2 | Low | Waits for signal from online SPE instrument that cartridge preparation is done. |
15 | Injection valve | Load | Removes sample from loop from cartridge path. |
16 | Wait | 10 sec | |
17 | Dispense | 10,000-µL waste-speed 5 | Dispenses syringe contents through buffer loop and sample loop for cleaning. |
18 | Syringe valve | Wash | |
19 | Load syringe | Volume 10,000-µL-speed 9 | Refills syringe with wash solution. |
20 | Syringe valve | Needle | |
21 | Dispense | 10,000-µL waste-speed 5 | Dispenses syringe contents through buffer loop and sample loop for cleaning. |
22 | Needle wash | 300 µL | Washes needle. |
23 | Compressor | OFF | |
24 | Wait for input 2 | Low | Waits for signal from online SPE instrument that cartridge preparation is done. |
25 | Wait | 3 sec | |
26 | Aux port 3 | ON | Freezes online SPE instrument. |
27 | Wait for input 1 | Low | Waits for signal from LC/MS Chemstation generated by the injection. |
28 | Wait | 5 sec | |
29 | END | Returns to step 1 until all vials in series are done. |
Appendix 2. Programmed steps for automated online solid-phase extractor
Glyphosate method 0-2136-01
Sample preparation program
Solvent 1 Methanol
Solvent 2 Acetonitrile
Solvent 3 Distilled, deionized water
SP program # 12:
Step | Time (minutes) | Action |
---|---|---|
1 | 0.00 | Valve 1 purge |
2 | .05 | Valve 1 elute |
3 | 9.05 | Valve 1 purge |
4 | 9.15 | Change cartridge |
5 | 9.15 | Solvent 1 |
6 | 9.15 | 2.0 mL/min |
7 | 11.16 | Solvent 3 |
8 | 13.16 | Aux 2 ON |
9 | 13.16 | Aux 2 OFF |
10 | 13.17 | 2.0 mL/min |
11 | 18.30 | 0.0 mL/min |
12 | 18.35 | Aux 2 ON |
13 | 18.36 | Aux 2 OFF |
14 | 18.40 | Valve 2 * * * |
15 | 18.41 | Solvent 2 |
16 | 18.42 | 2.0 mL/min |
17 | 19.41 | Solvent 1 |
18 | 20.11 | 0.0 mL/min |
19 | 20.12 | Valve 2 - - - |
20 | 23.43 | Aux 2 ON |
21 | 23.44 | Aux 2 OFF |
22 | 23.49 | End of program |
For additional information contact:
Betty Scribner
U.S. Geological Survey
4821 Quail Crest Place
Lawrence, KS 66049-3839
Telephone: (785) 832-3564
Fax: (785) 832-3500
Email: scribner@usgs.gov