AGRICULTURAL MATERIALS

Liquid Chromatographic Determination of Glyphosate in

Water-Soluble Granular Formulations: Collaborative Study
MORLIER & TOMKINS: JOURNAL OF AOAC INTERNATIONAL VOL. 80, NO. 3, 1997
LYNN W. MORLIER
Monsanto Company, PO Box 174, Luling, LA 70070
DAVID F. TOMKINS
Monsanto Company, PO Box 473, Muscatine, IA 52761

Collaborators: B. Koppen; B. Declercq; O. Weller; A. Gregoric; C. Focht; G. Roberts; D. Harbin; R.H. Schreuder; H. Hicks; F.
Verbist; T. Wieland; T. Bjorholm; L. Manso; S. Cohen; E. Lee; T. Zurcher; M.D. Muller; D.F. Tomkins; A. Noble; H.D.
Schlesinger; B. Smallage; K.C. Tam; Y.-W. Kim; T. Sato; D.I. Schultz; A. Parsons; I.M. Galoux; M. Hussain; F.M.E. Masoller

A liquid chromatographic (LC) method for the deter-
mination of glyphosate in various water-soluble
granular formulations was tested by 28 laboratories.
Samples were dissolved in mobile phase and in-
jected directly into an LC system with a 25 cm SAX
column and a mobile phase of 96% % aqueous buffer
solution, 0.0062M KH2PO4, and 4% % methanol. Detec-
tion was by UV absorption at 195 nm. Manual or auto-
mated injections were made via fixed-volume loops.
Calculations were based on peak area comparisons
with external standards. The collaborative study ana-
lyzed 5 matched pairs of 4 water-soluble granular for-
mulations, with one determination per sample. The
study generated 138 matched pairs, which were ana-
lyzed by using the AOAC spreadsheet and Youdens
matched-pair calculations. Coefficients of variation
%. The
for the 4 formulations ranged from 0.65 to 1.37%
LC method for the determination of glyphosate in
water-soluble granular formulations has been
adopted first action by AOAC INTERNATIONAL.
lyphosate, N-(phosphonomethyl)glycine, is the active
Methods for glyphosate analysis based on liquid chroma-
tography (LC) have been used by Monsanto Company for ap-
proximately 20 years (2). This collaborative study evaluated
the method that has been adapted for glyphosate in WSG for-
mulations (3). A key goal was to determine the methods preci-
sion for and applicability to the new formulations.
Collaborative Study
Five matched pairs of 4 glyphosate WSG formulations, to-
gether with standards, column, practice samples, and detailed
instructions, were sent to 28 laboratories in Europe, North and
South America, Asia, Australia, and the Middle East. Collabo-
rators were instructed to prepare mobile phase, inject standards
to test linearity of response, establish good resolution, and ad-
just integration parameters with a practice sample. Sample
chromatograms were provided. Participants were presented
with 10 samples (5 blind pairs) and asked to make single deter-
minations according to a work sheet with sample calculations.
Work sheets were reported with raw data such as standard
weights and area counts of standards and samples. Collabora-
tors were asked to report any deviations from the method as
written. In general, instructions were followed closely. The de-

G ingredient in Roundup herbicide as well as in more re-

cent formulations of dry water-soluble granules. The
dry formulations are ammonium salts of glyphosate. Some of
these formulations have surfactants. The water-soluble granu-
lar (WSG) products have been marketed in various countries
under such names as Trounce, Pacer, Rival, and Roun-
dupWSD. All formulations are water-soluble nonselective her-
bicides that are applied to foliage (1).
viations mentioned were practical adjustments such as manual
rather than automated injection, use of smaller injection loops,
detection at 200 nm, solution preparations based on 100 mL
instead of 50 mL, etc. These adjustments had little effect on the
quality of results.
996.12 Glyphosate in Water-Soluble Granular
Formulations, Liquid Chromatographic Method

Submitted for publication July 6, 1996. First Action 1996

The recommendation was approved by the Methods Committee on
Pesticide and Disinfectant Formulations, and was adopted by the Official
Methods Board of the Association. See Official Methods Board Actions Caution: See Appendix B: safety notes on safe handling of
(1996) J. AOAC Int. 79, 111A, and Official Methods Board Actions acidsphosphoric acid, and safe handling of special chemical
(1996) The Referee, October issue.
hazardspesticides and methanol. Dispose of waste solvents
in an appropriate manner compatible with applicable environ-
mental rules and regulations.
Method Performance:
See Table 996.12A for method performance data.
A. Principle
Samples are dissolved in phosphate buffer mobile phase and
injected into ion-exchange chromatographic system. Peak area
response, as measured by UV detection, is quantitated by exter-
nal standard technique.
B. Apparatus
(a) Liquid chromatographic (LC) system.Capable of
generating >3000 psi; equipped with isocratic LC pump capa-
ble of accurately metering 2.3 mL/min and with UV detector
set at 195 nm.
(b) LC column.Anion-exchange column. See E, System
Suitability.
(c) Injector.Fixed loop or autoinjector. Capable of accu-
rately delivering 50 L; capable of producing peak areas within
1% on the same glyphosate working standard or test sample
solution. (Note: Some anion-exchange columns require lower
loading, 1020 L, with the same reproducibility criteria. See
Table 996.12B for ASTM column suitability parameters.)
(d) Electronic integrator for chromatographic data.
Data-handling system may be used.
(e) Analytical balance.Accuracy of 0.1 mg.
(f) Volumetric flasks.100 mL.
C. Reagents
(a) Potassium dihydrogen phosphate.Primary standard
grade.
(b) Water.LC grade.
(c) Phosphoric acid (H3PO4).85%, reagent grade.
(d) Methanol.LC grade.
(e) Glyphosate reference standard.99.9% Purity (avail-
able from Monsanto Company, 800 N. Lindbergh Blvd, St.
Louis, MO 63167). Reference standard is stable at least 2 years
when stored in closed, dry container at 030C. Material of
lesser purity is available commercially as N-(phos-
phonomethyl)glycine, FW 169.1.
(f) LC mobile phase.Dissolve 0.8437 g KH2PO4 in
960 mL H2O. Add 40 mL methanol and mix well. Using pH
meter (calibrated at pH 2.0), adjust pH to 2.1 with 85% H3PO4.
Filter LC mobile phase and degas before use. LC mobile phase
is stable at least 2 weeks when stored in closed containers under
ambient conditions (2030C). When in use, lightly vent con-
tainer into atmosphere.
D. Preparation of Standard Solutions
Prepare 0.2, 0.4, and 0.6% glyphosate working standard so-
lutions as follows: Accurately weigh 0.2, 0.4, and 0.6 g
glyphosate reference standard to the nearest 0.1 mg into sepa-
rate flasks containing 100 mL LC mobile phase. Calculate ex-
act concentrations of glyphosate, taking into account purity of
glyphosate reference standard.
E. Calibration
Equilibrate LC column in LC mobile phase ca 1 h at
2.3 mL/min.
Inject glyphosate working standard solutions from D onto
LC column. Ensure reproducibility of injections to within 1%.
To obtain glyphosate retention time of 3.95.9 min, adjust pH
of LC mobile phase and adjust flow rate (2.3 mL/min and
pH 2.1 are typical). Ensure linearity of standard injections.
Using practice samples ensure baseline separation of
glyphosate peak from both surfactant peaks as well as small
impurity immediately after glyphosate (see Figure 996.12). If
necessary, adjust integration parameters to achieve known
analyses of practice samples.
F. System Suitability
The quality of LC column is critical to accurate and precise
glyphosate analysis. The criterion for acceptable LC column is
reliable integration of glyphosate peak in glyphosate formula-
tions. Interference from surfactant (peak A; see Figure 996.12)
with front end of glyphosate peak or especially interference
from impurity (peak C) with back end of glyphosate peak must
be avoided.
LC columns must be tested for suitability by using perform-
ance criteria that include peak efficiency, resolution between
glyphosate and peak C, and asymmetry (see Table 996.12B for
ASTM column suitability parameters.)
The apparent mode of separation of glyphosate is based on
ion suppression using ion-exchange resin as partition media.
This indicates that weaker exchangers or less efficient ion-ex-
change resins may perform better in resolving glyphosate from
its major impurities (see Table 996.12B).
The ideal criterion for adequate LC columns is performance
on control charts with a typical sample. Relative standard de-
viation should not be more than 0.25%.
G. Determination
Into 100 mL volumetric flask, accurately weigh 0.4 g test
sample to the nearest 0.1 mg and dilute to volume with LC mo-
bile phase. Sample response should be bracketed by 0.2 and
0.4% glyphosate working standard solutions.
Establish regression line of concentration vs response using
glyphosate working standard solutions. Make duplicate deter-
minations (i.e., 2 weighings, 1 injection/weighing) of each test
sample and sequence between 0.2 and 0.4% glyphosate work-
ing standard solutions.
Most samples contain a late-eluting peak that approximately
doubles the run time required for full elution. Injecting LC mo-
bile phase after each sample injection enables use of autoinjec-
tors to produce minimal analysis time for the required injec-
tions. Glyphosate working standard solutions bracketing
samples should produce slopes and intercepts within control
limits of established regression curve. With fixed-volume loop
injection, variance depends on fluctuations in laboratory tem-
perature. Sample concentrations determined should agree
within 1% relative basis.
Report average of 2 determinations.
 H. Calculations
Calculate regression line as follows:
Pg = mx + b
where Pg = peak area of glyphosate in glyphosate working
standard solution; m = slope of regression line; x = concentra-
tion of glyphosate in glyphosate working standard solution, %;
and b = intercept of regression line.
Calculate glyphosate concentration, %, in pesticide formu-
lation as follows:
(Ps b) V
Glyphosate, % =
mW
where Ps = peak area of glyphosate in test sample, V = volume
of diluted test sample, mL; and W = weight of test sample, g.
Ref. : J. AOAC Int. 80, 464468(1997)
CAS-1071-83-6 (glyphosate)
Results and Discussion
The collaborative data in Table 1 includes 138 matched
pairs from 28 laboratories. A statistical summary is presented
in Table 996.12A. The AOAC spreadsheet and Youdens
matched-pair calculations were used.
Table 996.12A gives performance parameters for each sam-
ple. The table contains n, the number of_ laboratories used in
estimating the performance parameters; x, the mean; sr, the re-
peatability standard deviation; sR, the reproducibility standard
deviation; RSDr, the repeatability coefficient of variation or
relative standard deviation; and RSDR, the reproducibility coeffi-
cient of variation or relative standard deviation.
Outliers are indicated in Table 1. These data were eliminated
from the statistical summary by either Cochrans or Grubbs
test for determining outliers. Only 4% of the data were rejected
as outliers. The method showed good performance by a large
number of collaborators, and thus, it is recommended for adop-
tion. The performance statistics in Table 996.12A show the co-
efficients of variation ranged from 0.65 to 1.37. The coeffi-
cients of variation were lower for MON 8750 matched pairs.
These samples contained no surfactants and had higher
glyphosate assay. The current statistics also compare favorably
with similar statistics obtained from the original 1983 study (4)
of glyphosate and (liquid) formulations presented in Table 2.
By comparison, the coefficients of variation from the original
study ranged between 0.88 and 1.70%. The basic method exten-
sion for WSG formulations is a valid modification to the original
glyphosate method, which already has moved to final action.
Recommendation
On the basis of the results of this study, it is recommended
that the LC method for determination of glyphosate in WSG
formulations be adopted first action, replacing AOAC Official
Method 983.10, Glyphosate (Technical) in Pesticide Formula-
tions, LC Method.
Acknowledgments
We thank Margaret Nemeth, Committee Statistician, and the
following collaborators and their associates for their cooperation:
Benny Koppen, Miljoministeriet National Environmental
Research Institute, Roskilde, Denmark
Bernard Declercq, Laboratoire Interrgional de la Rpres-
sion des Fraudes, Massy (Paris), France
O. Weller, AgrEvo GmbH, Frankfurt am Main, Germany
Ana Gregoric, Kmetijski Institut Slovenije, Hacquetova,
Slovenia
Charles Focht, Nebraska Department of Agriculture Labo-
ratory, Lincoln, NE
Graham Roberts, Department of Agriculture State Chemis-
try Lab, Melbourne, Australia
Don Harbin, Bayer Corporation, Agriculture Division, Kansas
City, MO
R.H. Schreuder, Ministerie van Landbouw, Natuubeheer en
Visserji, The Netherlands
Heidi Hicks, Montana Department of Agriculture, Bozeman, MT
Frans Verbist, Monsanto Europe NV, Antwerp, Belgium
Thomas Wieland, GTZ Pesticide Formulation, Schopfheim,
Germany
Thomas Bjorholm, Cheminova Agro A/S, Lemvig, Denmark
Luis Manso, Ministerio de Agricultura, Pesca Y Alimen-
tacion, Madrid, Spain
Sharryn Cohen, Nevada State Department of Agriculture,
Reno, NV
Elizabeth Lee, Zeneca Western Research Center, Richmond, CA
T. Zurcher and M.D. Muller, Swiss Federal Research Station,
Wadenswil, Switzerland
Alan Noble, Queensland Department of Primary Industries,
Indooroopilly, Australia
H.M. Schlesinger, Analyst Research Laboratories, Rehovot,
Israel
Bob Smallage, Office of the Indiana State Chemist, West
Lafayette, IN
K.C. Tam, Laboratory Services Division, West Agriculture
and Agri-Feed Canada, Calgary, Canada
Youn-Whan Kim, Oriental Chemical Industries, Inchon,
Korea
Tatsuo Sato, Monsanto Agricultural Research Center Asia-
Pacific, Ibaraki Prefecture, Japan
D.I. Schultz, Bayer AG, Monheim, Germany
Allan Parsons, G.C. Laboratories Ltd., Bedford, United
Kingdom
I.M. Galoux, Ministre de lAgriculture, Centre de Recher-
ches Agronomiques, Gembloux, Belgium
M. Hussain, Agrochemicals Unit, Joint United Nations Food
and Agricultural Organization/IAEA Programme, Vienna, Austria
F.M. Elena Masoller, Ministerrio de Ganaderia, Agricultura
y Pesca, Montevideo, Uruguay
References
(1) Technical Data Sheets on MON 8750, 60696, 14420-G,
14445, Monsanto Company, St. Louis, MO
(2) Tomkins, D.F. (1979) J. Chromatogr. Sci. 17, 333
(3) Buchanan, R.W. (1993) Monsanto Company, Internal Publi-
cation, Monsanto AQC Method No. 765
(4) Burns, A.J. (1983) J. Assoc. Off. Anal. Chem. 66, 12141219
(5) Youden, W.J., & Steiner, E.H. (1975) Statistical Manual of
the AOAC, AOAC, Arlington, VA
Table 996.12A. Method performance for determination of glyphosate in water-soluble granular formulations by
liquid chromatography
Glyphosate _
formulationa No. of labs x, % sr sR RSDr, % RSDR, % rb Rc

60696 27 63.40 0.69 0.69 1.09 1.09 1.93 1.93

14445 28 73.07 0.84 0.92 1.14 1.26 2.35 2.58
14420G 28 68.90 0.65 0.94 0.94 1.37 1.82 2.63
8750 26 87.38 0.41 0.57 0.47 0.65 1.15 1.60
8750 26 87.35 0.67 0.67 0.77 0.77 1.88 1.88
a
Youden pairs.
b
r = 2.8 sr.
c
R = 2.8 sR.
Table 996.12B. ASTM column suitability parameters for glyphosate separations using anion-exchange columns
(flow rate, 2.3 mL/min)
Resolution ratio
Injection volume, Glyphosate Column efficiency between glyphosate Asymmetry
LC column, type and length L retention time, min (glyphosate)a and peak C ratio

b
Whatman 4226 SAX, 250 mm 50 4.05 1341 2.17 0.68
Hamilton PRPX 100 SAX, 100 mm 10 4.41 1776 1.33 0.90
c
Mac-Mod Zorbax SAX, 250 mm 50 2.99 2842 1.74 0.69
d
Whatman 4120 WAX, 250 mm 10 3.70 6228 4.71 1.02
e
Whatman MAX I WAX, 300 mm 50 4.47 2784 4.78 0.38
e
Whatman MAX I WAX , 300 mm 10 4.27 4387 5.21 0.58

2
RT
a
Column efficiency (glyphosate), N, calculated as theoretical plates as follows: N = 5.54 , where RT = retention time, min; and Pw =
P
w
peak width, min.
b
Used in collaborative study.
c
Limited applicability because of front-end interference with peak A (surfactant, see Figure 996.12).
d
Some front-end interference observed at higher loading.
e
Columns are based on weak-anion exchange (WAX) and updated technology including 5 m silica rather than 10 m and spherical rather
than irregular packing. Weak-anion exchangers show promising improvements in efficiency and resolution, although column loading may
need to be decreased for optimal peak shape.
Table 1. Collaborative results of determination of glyphosate in water-soluble granules by LC
Matched pair 1, Matched pair 2, Matched pair 3, Matched pair 4, Matched pair 5,
Collaborator MON 60696 MON 14445 MON 14420-G MON 8750 MON 8750

1 64.05 63.92 74.33 73.68 69.31 69.51 87.78 87.58 87.73 87.80
2 a 63.19 75.22 73.22 69.08 70.12 87.13 87.66 87.79 86.95
3 63.44 63.97 73.90 73.23 69.61 69.79 87.54 87.86 87.91 88.01
b
4 63.92 62.37 73.13 72.04 67.95 67.64 88.08 86.84 82.16b 86.64
5 63.21 63.36 71.85 73.43 68.58 66.48 87.06 87.68 87.85 86.91
6 63.72 62.67 73.73 72.81 68.38 68.79 86.74 86.52 87.04 86.96
7 63.34 62.93 73.68 73.58 69.51 69.39 87.66 87.43 87.72 87.81
8 63.16 63.32 73.22 72.66 68.88 69.17 86.76 87.14 86.46 86.92
b
9 62.76 63.82 72.99 69.80 68.77 66.71 86.29 86.67 83.25b 87.45
10 63.77 63.50 74.15 73.10 68.69 68.66 87.67 88.33 87.43 87.69
11 62.94 64.05 72.43 73.34 68.58 69.41 87.11 86.58 86.04 86.98
12 63.74 63.48 72.54 73.56 69.30 68.99 87.16 86.93 87.96 87.23
13 61.77 62.09 71.24 72.11 68.20 67.90 86.78 86.88 86.53 86.80
14 63.72 a 72.16 72.60 70.72 70.27 88.03 88.16 87.17 88.74
15 64.20 63.66 73.11 73.50 70.14 70.47 87.96 87.37 87.18 88.79
16 62.16 65.09 72.30 73.04 68.60 68.02 87.17 87.40 86.94 86.71
17 62.90 62.66 72.88 72.62 68.24 68.42 86.95 87.08 87.04 86.70
18 63.78 63.25 73.56 73.18 68.83 68.80 87.25 87.01 87.31 87.34
19 62.95 65.38 73.08 74.21 69.38 68.96 88.49 88.07 88.31 87.32
20 63.70 63.72 74.07 73.88 70.14 69.98 87.62 86.68 87.24 86.96
21 64.19 64.02 74.30 73.46 69.72 70.02 88.09 89.06 87.69 87.64
22 63.75 62.79 73.83 72.49 68.97 69.75 86.70 88.15 87.84 87.63
23 63.86 63.17 72.92 73.09 68.44 66.82 87.44 87.73 85.71 87.14
24 63.37 63.29 73.44 73.63 69.36 69.28 87.23 87.15 87.48 87.66
c b
25 63.27 60.61b 70.48 73.60 69.71 67.31 85.25 b
83.93b 86.40 88.98
26 63.57 63.13 73.24 73.54 69.12 69.00 87.41 87.24 87.36 87.20
27 63.31 63.50 73.15 73.29 68.55 68.30 87.27 87.64 87.96 87.38
b
28 62.89 62.26 72.70 71.65 68.59 66.97 86.68 90.27b 85.89 88.04
a
Duplicate injections too far apart to report.
b
Results identified as outliers.
c
Alternative calibration procedure used.
Table 2. Comparative summary of 1983 collaborative
results for determination of glyphosate in Roundup and
technical-grade formulations (matched pairs)
Weight of isopropylamine salt, %

Technical
Statistic Roundup Amine salt glyphosate

_n 10 10 10
x 41.86 62.72 97.73
sr 0.57 0.55 0.55
sR 0.71 0.55 0.88
RSDR 1.70 0.88 0.90
Figure 996.12. Liquid chromatographic separation of
surfactant (A); glyphosate (B); impurity 1 (C); and
2 isomers of late-eluting peaks, impurity 2 (D); in
glyphosate herbicide sample.