2302269

Liquid Chromatography/Mass Spectrometry of the
Thermally Labile Herbicides, Chlorsulfuron and
Sulfometuron Methyl
Lamaat M. Shalaby
Agricultural Chemicals Department, E. I. du Pont de Nemours & Company, Inc., Research Division, Experimental
Station, Wilmington, Delaware 19898, USA
An on-line liquid chromatographic/mass spectrometric technique with direct liquid introduction (DLI) was investigated for the analysis of chlorsulfuron, the active ingredient of [email protected] herbicide and sulfometuron methyl, the
active ingredient of [email protected] herbicide. These compounds are thermally labile and cannot be analysed by gas
chromatography. Our work reported here demonstrates that with DLI-liquid chromatographyhnass spectrometry
(LC/MS) we can eliminate thermal decomposition and obtain a mass spectrum with complete fragmentation pattern
and abundant protonated molecular ion. The amount of fragmentation and abundance of the protonated molecular
ion can be changed by varying the source pressure or temperature. This effect is valuable for identification and
quantification purposes. In addition, the study includes a comparison of chlorsulfuron spectra obtained by both
DLI-LC/MS and commercial thermospray LC/MS interface.
INTRODUCTION
Both Du Pont [email protected] [email protected] have a wide
spectrum of activity. [email protected] herbicide is particularly
useful for controlling broadleaf weeds and some grasses
in cereal crops. The active ingredient of [email protected] herbicide is chlorsulfuron, 2-chloro-N-[[(4-methoxy6-methyl- 1,3,5-triazin-2-yl)amino]carbonyl]benzenesulfonamide. [email protected] is used for controlling many
grasses and broadleaf weeds growing in non-cropland
areas. The active ingredient of [email protected] herbicide, sulfometuron methyl, is methyl 2-[[[[(4,6-dimethyl-2pyrimidinyl) amino]carbonyl]amino]sulfonyl]benzoate.
Both structures are shown below:
N - 4\
om,
chlorsulfuron
[email protected] herbicide
sulfometuron methyl
Oust" herbicide
Chlorsulfuron and sulfometuron methyl are thermally
labile compounds and cannot be analysed by gas
chromatography. Consequently, liquid chromatography
@
Registered trademark of E. I. du Pont de Nemours and Company,
Inc.
is the primary analytical technique used for these compounds so as to avoid decomposition during analysis.'**
Normally, mass spectrometric identification of these
sulfonylureas in a sample mixture would require an
off-line liquid chromatographic separation with fraction
collection and possible subsequent derivatization. Both
techniques are tedious, time consuming and, in most
cases, inefficient.
Generally, electron ionization (EI) spectra of sulfonylureas contain only low mass fragment ions and
no molecular ions. On the other hand, with soft ionization techniques such as chemical ionization (CI), the
protonated molecular ion is detected and a complete
fragmentation pattern is generated. For this reason, we
normally use chemical ionization for identification and
structure elucidation of sulfonylureas and their derivatives.
In this investigation, on-line LC/MS was selected for
the analysis of chlorsulfuron and sulfometuron methyl
as it offers both the liquid chromatographic separation
and chemical ionization conditions recommended for
those compounds. Among the most commonly used
LC/MS interfaces are direct liquid introduction3 (DLI)
and the moving belt.4 In our case, the direct liquid
introduction approach was preferred because it involves
lower temperatures at the interface, thus reducing potential thermal degradation.
In the present investigation, we describe the
instrumentation and report the analytical results for
chlorsulfuron and sulfometuron methyl by on-line DLILC/MS. A study of the effect of the source temperature
and pressure on the fragmentation pattern in addition
to a comparison of the spectra obtained by DLI and a
commercial thermospray interface is discussed.
EXPERIMENTAL
A Finnigan Model 4500 mass spectrometer was adapted
to a Hewlett-Packard direct liquid introduction (DLI)
CCC-0306-042X/S5/060261-08 $04.00
0 Wiley Heyden
Ltd, 1985
BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985 261
L. M. SHALABY
VESPEL" DESOLVATION
CHAMBER
TO THE
QUADRUPOLE
f
THE DLI PROBE
THE ION
VOLUME
Figure 1. DLI probe and the modified ion volume assembly
interfacesy6 (Hewlett-Packard, Palo Alto, California,
USA) to carry out this investigation. This required the
addition of a liquid nitrogen cryogenic pump to improve
the pumping capacity of the mass spectrometer. A
desolvation chamber was also needed to desolvate the
effluent and connect the DLI probe to the ionizer. This
was accomplished by replacing the 0.7 cm long Finnigan
stainless steel ion volume adapter (which is part of the
Finnigan ion volume assembly) with a 2.8 cm long [email protected] adapter as shown in Fig. 1. This device served as
an interchangeable desolvation chamber, which is inserted and removed with the regular Finnigan insertion
removal probe. Such an interchangeable desolvation
chamber eliminates the need to drill in the source block
or vent the instrument to disconnect the chamber. Using
[email protected] to fabricate the desolvation chamber reduced
the amount of heat conducted to the DLI probe from
the source and eliminated the need for cooling the probe.
The liquid chromatograph employed was a MACSO
100 pump with a Rheodyne valve loop injector and
either a 15 cm ~ 4 . mm
6 i.d. [email protected] ODS or silica prepacked column (Du Pont Company, Analytical Instruments Division, Wilmington, Delaware). Chlorosulfuron and sulfometuron methyl samples were standard
reference materials (Du Pont Company, Wilmington)
and all solvents used were high-performance liquid
chromatography (HPLC) grade (Burdick and Jackson
Lab., Muskegon, Michigan) filtered with a [email protected]
filtering apparatus (Millipore Corporation, Bedford,
Massachusetts) before use. Data were acquired and processed using an INCOS data system.
RESULTS AND DISCUSSION
A start-up procedure for DLI-LC/ MS includes checking
the liquid chromatographic separation and the quality
of the jet before introducing the probe to the mass
spectrometer. In some instances we have modified existing liquid chromatographic conditions to make them
more compatible with the mass spectrometer. These
modifications included using volatile organic acids or
buffers and modifying the mobile phase to contain less
water. In some cases, we adjusted the flow rate to
increase the amount of sample injected to the mass
spectrometer after splitting by the DLI interface.
Checking the jet quality and performance is crucial
to ensure a successful LC/MS run. The diameter of the
diaphragm pinhole, the liquid chromatographic flow rate
and DLI split ratio all affect the length of the jet. On
the other hand, the jet direction is strictly dependent on
the uniformity of the orifice. Occasionally the pinhole
262 BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985
may crack or suffer partial blockage by tiny particles
which will affect the jet direction and render it
~ n c o a x i a l .Backflushing
~
and rubbing the diaphragm
surface around the orifice can sometimes help remove
the imbedded particles from the orifice. For flow rates
of 15 p1 min-' (*3) at the diaphragm orifice, we found
that the optimum pinhole diameter is 5 pm (*1). With
larger orifices (>5 pm) higher flow rates (> 15 ~ 1 are
)
required to produce a jet, and in this case, the mass
spectrometer may not accommodate such amounts of
liquid. For our studies here, the optimum jet length was
found to be 2.5-3 cm, depending on the type of mobile
phase we used. Once the DLI jet was optimized, we
generally continued monitoring it for a few minutes to
verify its stability and then inserted the probe into the
mass spectrometer.
Initially, the probe was positioned a few microns away
from the desolvation chamber while both the ionizer
pressure and the analyser high vacuum were monitored.
Once the pressure stabilized, the probe position was
then adjusted to yield the maximum tolerable ionizer
pressure. Typical ionizer pressure was 0.25 Torr and
4.0 x
Torr for the high vacuum.
The performance of the jet under vacuum is easily
judged by the amount of fluctuation observed in the
ionizer pressure. A typical fluctuation should not exceed
k0.02 Torr. A continuing pressure buildup or sharp
fluctuation could indicate a defective jet.
The solvent cluster ions were utilized to tune the mass
spectrometer' and to select the mass scan range. Normally the scan range was set to eliminate the solvent
ions' and to improve sensitivity in the mass range of
interest.
Figure 2(a) shows the reconstructed total ion current
(RIC) trace obtained by DLI-LC/MS for chlorsulfuron
using the ZorbaxO ODS column with 6Ooh (v/v)
acetonitrile in pH 2.9 water as a mobile phase at a flow
rate of 1.25 ml min-' and a DLI split ratio of 1/99. The
chromatogram contains a single sharp peak which indicates that the thermally labile molecule elutes intact
without any decomposition. The stable baseline is due
to the combination of adequate pumping capacity of the
mass spectrometer, optimized jet and a pulse-free liquid
chromatographic pump.
The type of mass spectrum obtained with DLI-LC/ M S
is shown in Fig. 2(b). It is a CI type of spectrum with
a protonated molecular ion m / z 358 [M-tH]' as the
base ion. Furthermore, the spectrum contains a complete
fragmentation pattern which is more informative than
the reference EI spectrum shown in Fig. 2(c). Among
the new high mass fragment ions formed under DLILC/MS conditions were m / z 324 [MH -Cl]+, m / z 294
[ MH - SO,]' and m/ z 184 [M - C,H,SO,Cl]+.
These ions are valuable for confirmation and structure
elucidation. Although the intact molecule did not form
adducts with the solvent ions, some of the key fragment
ions did form strong adduct ions as shown. The adduct
ion of m / z 182 is of special interest because it confirms
the
m / z 140
(2-amino-4-methoxy-6-methyl-1,3,5triazine) fragment ion which was originally eliminated
by the selected scan range (150-500u).
The mass spectrometer ionizer pressure and temperature were found to influence the quality of the
obtained spectrum. Figure 3 shows the DLI-LC/MS
spectra of chlorsulfuron with 50% (v/v) acetonitrile in
LCMS OF HERBICIDES
:HLORSULFURON
136
182
I
SCAN
TIME
100
3.20
2
6 40
Figure 2. (a) Total ion current DLI-LC/MS chromatogram of chlorsulfuron using acetonitrile: pH water 2.9 (60:40); (b) the
corresponding LC/MS spectrum of chlorsulfuron; (c) the probe El reference spectrum of chlorsulfuron.
m/z233
[ac1 I
SO,NH,
LcH~O
tonated molecular ion then the reverse ratio [ M S H 140]+/[ M H]+ indicates the percentage fragmentation.
Table 1 contains those defined ratios as a function of
the source pressure and illustrates that the degree of
fragmentation is inversely proportional to the abundance of the protonated molecular ion. When the source
pressure was increased the percentage fragmentation
was reduced from 67% at 0.24Torr source pressure to
3% at 0.35 Torr. On the other hand, the abundance of
the protonated molecular ion is almost 17 times greater
at 0.35 Torr source pressure than at 0.24 Torr.
The signal level at higher source pressure is larger
than at lower source pressure due to the high sample
transmission efficiency. As we retract the probe further
away from the desolvation chamber, the mobile phase
reagent gas and sample molecule are pumped out before
they reach the ionizer. The loss in the mobile phase
reagent gas and the pressure drop will inhibit the
+
+ CH3CN + H
J
pH 2.3 water (50: 50) as the mobile phase at three different source pressures, 0.24, 0.26 and 0.35 Torr. The
spectra show that higher source pressure increases the
relative abundance of the protonated molecular ion [M +
HI+ at m/z358 and decreases the degree of fragmentation.
If the intensity ratio of m / z 358 [M+H]+/rn/z 218
[M + H - 140]+ gives the relative abundance of the pro-
Table 1. The effect of source pressure (Torr) on the relative
abundance of [M +HI+ and fragmentation
Source
pressure (Torr)
Relative abundance of [M +HIC
0.24
0.26
0.35
2
[M+H]+/[M +H-140]+
Fragmentation (%)
[M+H-l40]+/[M
4
67
29
33
3
+HI’
BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985 263
L. M. SHALABY
CHLORSULFURON
OCH 3
MOL. WT 357
SOURCE PRESSURE IS 0.24 TORR
50-
CHLORSULFURON
MOL WT 357
SOURCE TEMP. 12OoC
302
218
208
252
2T8
I I
324
150
200
250
m/z
300
350
100
'
O
0
y
CHLORSULFURON
SOURCE TEMP. 14OoC
-
[MfH]'
358
7
184
I
1
218 233
177
194
225
208
252
324
I,
m/z
100
167
CHLORSULFURON
SOURCE TEMP. 17OoC
CHLORSULFURON
SOURCE PRESSURE
[M+H]'
358
--
NO
m /z
Figure 3. Liquid chromatography/chemical ionization spectra of
chlorsulfuron with acetonitrile: pH 2.3 water (50:50) as reagent at
0.24, 0.26 and 0.35 Torr source pressure.
interaction process between reagent ions and sample
molecules. These conditions favor sample fragmentation
and decrease the abundance of the protonated molecular
ion. Our study showed that the absolute signal level and
the relative abundance of the protonated molecular ion
264 BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985
1
1
7
7
350
Figure 4. Liquid chromatography/chemical ionization mass spectra
of chlorsulfuron with acetonitrile: pH 2.3 water (50:50)as reagent
at 120 "C, 140°C and 170°C source temperature.
is maximized when the probe tip is butted against the
desolvation chamber. These conditions are useful for
the molecular weight determination and when optimum
sensitivity is required. When the probe tip is retracted
by a few microns (sometimes just rotating outward
<30°), we allow a breathing distance between the probe
tip and the desolvation chamber. Under this condition,
the ionizer pressure drops and both the absolute signal
LCMS OF HERBICIDES
100
RI(
SULFOMt
0
W
U
SULFOMETURON METHYL
199
50-
[email protected] HERBICIDE
I
I
I
560
9 20
I
I
I
580
9 40
600
620
10 00
10 20
1 1 I,
1
241
, ,, , ,
I
250
300
350
, , '14
150
200
I
640 SCAN
10 40 TIME
t
Mol. w t 364
1210
, ,
m/z
Figure5. (a) Total ion current DLI-LC/MS chromatogram of sulfometuron methyl using methylene chloride +acetic acid +methanol (97 :2: 1);
(b) the corresponding L C / M S spectrum of sulfometuron methyl; (c) the probe/El reference spectrum of sulfometuron methyl.
level and relative abundance of the protonated
molecular ion decrease. The loss in the absolute signal
level is proportional to the breathing distance. Those
conditions are useful to induce fragmentation for structure elucidation. For sample applications we limited the
overall pressure changes by positioning the probe
(<0.2 Torr total pressure drop), so that the loss in
absolute signal level was not more than 50%.
A similar effect could be induced by changing the
ionizer temperature. Figure 4 shows the DLI-LC/MS
spectra of chlorsulfuron with 50% (v/v) acetonitrile in
pH 2.3 water as the mobile phase at three different source
temperatures, 120"C, 140 "C and 170°C. The data indicate that a lower source temperature will increase the
abundance of the protonated molecular ion and decrease
fragmentation.
As shown in Table 2, the ratio of [ M + H]+/[M+ H 140]+is 34 times higher at a source temperature of 120 "C
than at 170 "C. Similarly, the percentage fragmentation
is 3% at a 120°C (source temperature) compared with
91% at 170°C.
For practical reasons, controlling the amount of fragmentation and abundance of the protonated molecular
ion is best accomplished by adjusting the source pressure
rather than the source temperature. Normally it is done
by retracting the DLI probe a few microns away from
the desolvation chamber so that some of the mobile
phase reagent gas and sample molecules can escape to
the manifold. The ionizer source pressure could also be
changed by adjusting the liquid chromatographic flow
rate or the DLI split ratio, bat this could affect the
performance of the jet and may case major pressure
fluctuation. Varying the position of the probe is a more
efficient way to adjust the ionizer pressure.
This type of behavior is useful particularly for identification of thermally labile compounds and their related
compounds (metabolites and other by-products). An
initial analysis at higher source pressure would maximize
sample transmission efficiency and yield a spectrum with
a defined molecular ion. At the same time, a subsequent
analysis at lower source pressure would induce fragmentation, which could be valuable for structure elucidation. Furthermore, analysis at relatively high source
pressure would improve the detection level with selected
ion monitoring within the molecular ion region.
The experimental setup described thus far could
efficiently handle reversed phase mobile phases containing up to 50% water. Higher water content mobile phases
Table 2. The effect of source temperature on the relative
abundance of [M +HI' and fragmentation
Source
Relative abundance of [M +HI+
temperature ( " C )
[M + H ] + / [ M + H - 140]+
Fragmentation (%)
[M +ti - 140]+/[Mt HIC
120
140
34
3
3
29
170
1
91
BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985 265
L. M. SHALABY
0
100.0
232
97
:
4.3.
20
40
60
0:40
1:20
2:oo
80
2:OO
I
I00
100
320
I
120
4:OO
SCAN
TIME
RIC
2-chlorobenzene sulfonomide
~6
14
24
33
4
6, ,
Figure6. The DLI-LC/MS ion chromatogram of a mixture of 1.6 +g chlorsulfuron
and 25 ng 2-chlorobenzene sulfonamide, eluted with 60% acetonitrile in 0.1 M
ammonium acetate and the mass chromatogram of m / z 233 [M +CH,CN +H I+
ion of the 2-chlorobenzene sulfonamide.
would probably require higher pumping capacity to
maintain the ionizer pressure and the high vacuum in
the operating ranges. In addition, we found that volatile
organic acids such as trifluoroacetic acid and acetic acid
could be added to the mobile phase to adjust the pH
with no apparent problems.
As part of these studies, we also used normal phase
chromatography with a methylene chloride +
methanol + acetic acid (97% :2% : 1 % ) mobile phase.
This was used to separate sulfometuron methyl on a
25 cm x4.6 mm i.d. [email protected] Sil prepacked column at a
flow rate of 1.4 ml min-' and a 1/99 DLI split ratio. The
system was extremely stable with the ionizer forepressure
at 0.1 Torr and the high vacuum at 2.4 x
Torr.
Figure 5(a) shows the reconstructed
ion
chromatogram of sulfometuron methyl which shows a
single peak with no apparent decomposition. The
LC/MS spectrum obtained under these conditions (Fig.
266 BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985
5(b)) yielded a molecular ion abundance of 80% and a
fragmentation pattern similar to the conventional EI
spectrum (Fig. 5(c)) with a new high mass fragment ion
at rn/z258 [M+H-107]+.
The DLI-LC/MS system utilized in this investigation
is applicable for separation and identification purposes
of chlorsulfuron and sulfometuron methyl by-products.
Figure 6 is the reconstructed ion chromatogram for
separation of a mixture of 1.6 kg chlorsulfuron and 25 ng
2-chlorobenzene sulfonamide eluted with 60%
acetonitrile in 0.1 M ammonium acetate and from a
25 cm ~ 4 . mm
6 i.d. [email protected] ODS column at 1 mi min-*
flow rate. This chromatogram illustrates the applicability
of the technique to the separation and identification of
trace components in a major product. In this case it is
the 2-chlorobenzene sulfonamide in chlorsulfuron.
Similar application can include sulfonylurea metabolites
in biological sample extracts or residue samples.
LCMS OF HERBICIDES
CH3
Chlorsulfuron
rnol w t 357
I
46
20
40
0 44
1 28
I I
68
2 12
80
2 56
100
3 40
120
424
140
508
160 SCAN
5 5 2 TIME
Figure 7. (a) The thermospray ionization mass spectrum and (b) the
thermospray total ion chromatogram of 200 ng chlorsulfuron separated by 25%
(v/v) acetonitrile in 0.1 M ammonium acetate.
Figure 6 also illustrates the detection limit for the
LC/MS system used in this investigation of approximately 20 ng of sample injected to the mass spectrometer
(2 Fg injected on column) in the full scan range. With
selected ion monitoring at the molecular ion region, a
detection limit of 200 pg may be achievable.
With the DLI technique and conventional HPLC flow
rates, ( 1- 1.5 ml min-') the sample effluent is split 1 :99
before the mass spectrometer to maintain the high
vacuum. Only 1-2% of the material injected onto the
column is actually introduced into the mass spectrometer
(10-15 pl min-').
Microbore
columns
(25 cm x
1 mm i.d.) with flow rate of 50 p1 min-' and a split ratio
of 1 :2 were utilized in our laboratory and found to
improve the sample transmission efficiency to 30%. Preliminary evaluation of microbore HPLC shows that it
can produce separations comparable to the conventional
HPLC. In addition, it can offer a better detection limit
with the DLI interface due to the enrichment of the
HPLC effluent in sample relative to solvent. However,
microbore columns are more difficult to use and
maintain.
As we have discussed so far, on-line LC/MS can be
a very efficient technique for handling thermally labile
sulfonylurea samples. In an effort to expand the applicability of this technique to handle highly aqueous mobile
phase (>5O% water content) and maximize sample
transmission efficiency, we have recently looked into the
new thermospray" technique. Initially we were concerned about the stability of those fragile molecules in
the hot vaporizer region of the thermospray interface' I
(>200 "C) and the complexity of the spectrum obtained.
Figure 7(b) shows the thermospray" reconstructed
ion chromatogram for 200 ng chlorsulfuron utilizing a
commercially available thermospray unit manufactured
by Finnigan MAT (San Jose, California). The separation
BIOMEDICAL MASS SPECTROMETRY, VOL. 12, NO. 6, 1985 267
L. M. SHALABY
was carried out on 15 cm x4.6 mm i.d. ZorbaxO ODS
prepacked column with 25% (v/v) acetonitrile in 0.1 M
ammonium acetate at 1.3 ml min-I. The thermospray’’
spectrum (Fig. 7(a)) showed a less abundant protonated
molecular ion (<50%) relative to the DLI generated
spectrum (Fig. 2) and contained a molecular ion adduct
at m /z 375 [ M + NH4]+. In addition, the fragmentation
pattern contained fewer fragment ions, but was similar
in structure to the one obtained by DLI. The data also
show a detection limit below 50ng of sample injected
on the column. With thermospray LC/MS the total
sample effluent is introduced to the mass spectrometer.
Comparable sample transmission efficiency could be
achieved with the DLI interface only with microbore
columns. The detection limit in terms of amount of
sample introduced to the mass spectrometer by either
DLI or thermospray is equivalent. Although the ther-
mospray technique is not yet full characterized, our
initial results show that it can efficiently handle our
thermally labile herbicides in a highly aqueous mobile
phase (>75% aqueous). Although the spectrum
obtained by DLI is more valuable for molecular weight
determination and structure elucidation, the thermospray technique offers high sample transmission
efficiency and is more adaptable to reversed phase liquid
chromatography using conventional columns and flow
rates.
Thus far, DLI and thermospray techniques complement each other to render on-line LC/MS applicable to
most of the liquid chromatographic conditions currently
used in chromatography laboratories. The specific needs
of each laboratory coupled with the performance of the
specific LC/ MS interface will dictate which technique
is best suited for a given application.
REFERENCES
1. E. W. Zahnow, J. Agric. Food Chem. 30,854 (1982).
2. R. V. Slates, J. Agric. Food Chem. 31, 113 (1983).
3. M. A. Baldwin and F. W. McLafferty, Org. Mass Spectrom. 7 ,
111 (1973).
4. W. H. McFadden, H. L. Schwartz and S. J. Evans, J. Chromafogr.
122, 389 (1976).
5. P. Arpino, M . A. Baldwin and F. W. McLafferty. Biomed. Mass
Spectrom. 1, 80 (1974).
6. A. Melera, Adv. Mass Specfrom. 86, 1597 (1980).
7. B. Mauchamp and P. Krien, J. Chromatogr. 236, 17 (1982).
268 BIOMEDICAL MASS SPECTROMETRY, VOL. 12. NO. 6, 1985
8. R. D. Voyksner, C. Parker and R. Hass, Anal. Chem. 54, 2583
(1982).
9. R. D. Voyksner, R. Hass and M. M. Bursey. Anal. Chem. 54,2465
(1982).
10. C. R. Blakley and M. L. Vestal, Anal. Chem. 55, 750 (1983).
11. M. L. Vestal, Int. J. Mass Specfrom. /on Phys. 46, 193 (1983).
12. W. H. McFadden, Finnigan MAT (San Jose, California).
Received 9 July 1984; accepted (revised) 19 December 1984