GAS CHROMATOGRAPHIC EVALUATION OF SOME N-DERIVATIVES OF INTACT CARBARYL E. D. MAGALLONA and F. A. GUNTHER Department of Entomology Universityof the Philippines College, Lagana Philippines and
Departmentof Entomology Universityof California Riverside, California 92502
Gas chromatographic (GLC) detection of aryl N-methylcarbamates as the N-acetyl-Clx or N-nitroso derivatives of the intact molecule was explored using carbaryl as the model compound. These derivatives were designed for use with the microcoulometric detector for chlorine and the solution conductivity detector for nitrogen but response to the electron capture detector was also studied. GLC characteristics of the derivatives were evaluated in terms of thermal stability and detector response. N-Trichloracetyl and N-nitroso carbaryl underwent extensive decomposition in the GLC system. Although the N-monochloroacetyl derivatives were also subject to thermal decomposition, they gave sufficient linearity and sensitivity with the microcoulometric detector for analysis at the residue level. N-Acetyl carbaryl was thermally stable and the response of electrolytic conductivity and the electron capture detectors were adequate for residue analysis. Because o f the thermal instability o f most carbamates when analyzed by gas chromatography (GLC), indirect detection methods are becoming popular. Two o f the more promising approaches are detection as derivatives o f the phenol hydrolysis product and that o f the intact carbamate. Although the latter has been used only to a limited extent (see review by Williams 1971) it appears very promising and has distinct advantages over the former, especially in metabolite detection. T h e gas c h r o m a t o g r a p h i c e v a l u a t i o n o f two N - d e r i v a t i v e s , the N - a c e t y l - C l x ( x = 0 , 1 , 2 , 3 ) andN-nitroso derivatives is reported here using carbaryl as model compound a n d p , p ' - D D T as the comparison standard. Carbaryl was used because it is one o f the most popular carbamate insecticides and apparently one of the most unstable to GLC conditions. These derivatives were developed primarily for G L C with detectors specific for chlorine or nitrogen but the electron capture detector may also be used. Archives of Environmental Contamination and Toxicology Vol. 5, No. 2 185-190 (1977) 1977 by Springer-VerlagNew York Inc.
185
186
E. D. Magallona and F. A. Gunther
Experimental Synthesis and/or purification of experimental compounds. Experimental compounds were synthesized, purified, and characterized as described in Magallona et al. (1976). Microcoulometric G L C . The gas chromatograph, operated in the reductive mode, consisted of a Dohrmann Instruments Co. Model C-200 microcoulometer, S-200 pyrolysis furnace and a T-300 S titration cell attached to a Loenco, Inc. Model 15-C-E column oven by an independently heated quartz transfer line. Nitrogen was the carrier gas and hydrogen was the reaction gas. Thermal stability of the chlorinated acetyl derivatives was determined by injecting known quantities of the derivatives a n d p , p ' - D D T and plotting detector response percent of theoretical versus ng of chlorine injected. Calibration curves were also constructed for the N-chloroacetyl and N-dichloroacetyl derivatives of carbaryl (C-CIt and C-C12, respectively) with the detector in the sensitive mode to determine detector linearity and the minimum detectable ranges. Solution conductivity G L C . Detection was in the reductive mode using a Coulson Instruments Co. Model 1 gas chromatograph described by Coulson (1965) and incorporating the suggestions of Patchett (1970). Hydrogen was the carrier and reaction gas. Thermal stability of experimental compounds was determined by constructing calibration curves and comparing detector response using varying column temperatures to obtain essentially the same retention times. Electron capture G L C . A Varian Aerograph Model 1700 gas chromatograph, operated isothermally and equipped with a tritium foil, was used, with nitrogen as the carrier gas. Calibration curves were constructed to determine linearity ranges, minimum detection limits, and detector response to experimental compounds relative t o p , p ' - D D T . Comparative response was expressed as relative response factor (RRF) defined as: nanomoles of derivative to give response x RRF = nanomoles ofp,p' -DDT to give response x
Results and discussion Microcoulometric G L C . p,p"DDT was used as the standard material for evaluating the thermal stability and detector linearity for experimental materials because of its known GLC characteristics. A fortuitous consequence of this choice was the comparable retention times of the compound with C-C12 and N-trichloroacetyl carbaryl (C-CIa) in several columns. This minimized variations in recovery due to differences in column temperatures or delays due to the need to adjust column temperature to obtain comparable retention times. In an attempt to find the optimum reaction gas flow rate for sample combustion, detector response to constant sample size was plotted as a function of total gas flow with the carrier gas flow kept constant. As can be seen in Figure 1, detector response was essentially constant for the 130 to 220 ml per rain range of total gas flow. A total gas flow of
GLC Evaluation of N-Derivatives of Carbaryl
187
100 9
9
=
DDT
-=
C-CI=
,n
C-Ch
80.
6O
4O
Q
r
o
20.
1~o
1~o
1~o
2~o
2~,o
F l o w rate (ml/min)
Fig. I. Microcoulometric detector response top,p'-DDT and twoN-acetyl-Clx derivatives of carbaryl as a function of reaction gas flow rate using a 6-foot, 3% DC-200 column at 205~ 200 ml per min was therefore arbitrarily adapted. Although response was unaffected by gas flow rates, it appears that both carbaryl derivatives have been subject to thermal decomposition, especially CI-C13. For C-C12 and C-C13, detector response also increased with the number of injections. The preliminary injection of large slugs of sample (about 3 x 30/zl of 100/zg/ml solution) adequately "conditions" the column for C-C12 so that successive injections gave acceptable reproducibility. On the other hand, response with C-Ci3 was still not reproducible and continued to increase with the number of injections. The data in Figure 1 for this compound were therefore taken from the fourth injection after "conditioning" the column. Detector response for the derivatives as a function of quantity of material injected (Figure 2) also showed lower recovery compared to p,p'-DDT and increased response with increase in quantity of chlorine injected. The latter was indicative of thermal decomposition. With C-C13 in particular therefore, the theoretical advantages in its use seem to be negated by its thermal instability and further work on this derivative was discontinued. Although it appears that the N-acetyl-Clx derivatives of carbaryl underwent significant thermal decomposition, it was still impo~ant to find the detectability limit and linearity range of the microcoulometric GLC (MC-GLC) especially for C-CIt and C-C12. This is because differences in "recovery" may be due, at least in part, to structural differences which influence completeness of combustion or formation of titrable chlorine. The calibratipn curve obtained by operating the MC-GLC in the sensitive mode (furnace effluent impinging on the sensor electrode) shows only slight upward curvatures and 25 and 15 ng of C-CI1 and C-C12, respectively, give about 8% full-scale recorder deflection (Figure 3). Therefore, from the standpoint of residue detection, these derivatives could be used. Solution conductivity GLC. This detector, with its high specificity for nitrogen, is well suited for the evaluation of the thermal stability of carbamates and/or carbamate derivatives. Since the theoretical basis for the quantitative relationship between electrolytes introduced and detector response has not been established, and empirical approach based
E. D. Magallona and F. A. Gunther
188
1(111. 80.
!. Q
=
,
DOT
C-CI=
,7-- 60-
~ ~
~
, . ~ . . . _ ~
~
C-CIt
/ r
40-
x
C-CI=
20-
1O0
200
300
Chlorine
(ng)
Fig. 2. Microcoulometric detector response to p, p'-DDT and the N-acetyi-C I x derivatives of carbaryl as a function of quantity of material injected using a 4-foot, 3% SE-30 column.
on the construction of calibration curves using stock solutions of pure compounds, has to be resorted to. Differences in slopes of these curves, plotted as detector response v e r s u s ng of nitrogen, would reasonably reflect differences in thermal stabilities.
From Figure 4, it was evident that with the possible exception of N-acetyl carbaryl (C-CIo), the derivatives of carbaryl underwent thermal decomposition with the GLC parameters used. For these derivatives, thermal stability appears to be related to degree of chlorination. The N-nitroso derivative (C-NO) has to be ruled out for use in carbaryl detection because of its thermal instability. The data obtained here are also confirmed by mass spectroscopy-GLC.
/•0-CII
1000
ol r
--~
500
. . . .
i
9
-
9
50 Derivative
t
100
.
.
.
.
i
150
injected
(ng)
Fig. 3. Response of the microcoulometric detector operating at sensitive mode to N-monochloroand N-dichloro-acetyl carbaryl using a 2-foot, 3% DC-200 column.
GLC Evaluation of N-Derivatives of Carbaryl
1200-
. c-c~,0920c)
1000~
189
/
/
/ /
8oo.
~
//
C-CI,(205oc)
/ C-C,(210"C)
4OO
200
~
~
C-NO(135~
Nitrogen (ng)
Fig. 4. Comparativeelectrolytic conductivity detectorresponseto carbaryl and its N-derivatives using a 3.5-foot, 2% OV-17 column.
t
/
DOT(183oc)
'~176 t /
. C'CI,(174~)
//~
10 Nanomoles
c'c' ('a~176
20 x 10~
Fig. 5. Comparative electron capture response to p, p'-DDT and the N-acetyl-C Ix derivatives of carbaryl using a 4-foot 3% DC-200 column.
Electron capture G L C . Since the electron capture detector has found use in the detection of pesticides down to the picrogram range, it was decided to find the detectability limits of the derivatives with this detector. This detectability limit is expressed relative to p,p'-DDT or any other standard compound. Temperatures were varied for each compound to obtain a retention time in the 3.2 to 3.5 min range because preliminary work had shown that detector response increased with retention time. Figure 5 shows that detector response was linear for p , p ' - D D T , C-CI0, and C-Cll, whereas response to C-CIz was non-linear. This curvature could be due to column decomposition or may be influenced by other factors discussed by Gaston (1964). Detector response to C-CIa was no longer determined because it was found to decompose by MC-GLC and detector response increased with increasing number of injections. Under our conditions of use, C-CII gave an RRF of 4.0 while with C-C10, this was 5.6. On the weight basis, the lower limits of detection were about 0.1 ng for both derivatives.
190
E. D. Magallonaand F. A. Gunther References
Coulson, D. M.: Electrolytic conductivity detector for gas chromatography. J. Gas Chromatog. 3, 134 (1965). Gaston, L. K.: Gas chromatography using an electron absorption detector. Residue Reviews 5, 21 (1964). Magallona, E. D., F. A. Gunther, and Y. Iwata: Synthesis and spectroscopic characterization of acetyl-Cl~ = 0,1,2 or 3) and nitroso derivatives of carbaryl and three possible metabolites. Arch. Environ. Contamin. Toxicol. 5, 177 (1976). Patchett, G. G." Evaluation of the electrolytic conductivity detector for residue analysis of nitrogen-containing pesticides. J. Chromatog. Sci. 8, 155 (1970). Williams, I. H.: Carbamate insecticide residues in plant materials determination by gas chromatography. Residue Reviews 38, 1 (1971). Manuscript received April 16, 1975; accepted August 24, 1975
Departmentof Entomology Universityof California Riverside, California 92502
Gas chromatographic (GLC) detection of aryl N-methylcarbamates as the N-acetyl-Clx or N-nitroso derivatives of the intact molecule was explored using carbaryl as the model compound. These derivatives were designed for use with the microcoulometric detector for chlorine and the solution conductivity detector for nitrogen but response to the electron capture detector was also studied. GLC characteristics of the derivatives were evaluated in terms of thermal stability and detector response. N-Trichloracetyl and N-nitroso carbaryl underwent extensive decomposition in the GLC system. Although the N-monochloroacetyl derivatives were also subject to thermal decomposition, they gave sufficient linearity and sensitivity with the microcoulometric detector for analysis at the residue level. N-Acetyl carbaryl was thermally stable and the response of electrolytic conductivity and the electron capture detectors were adequate for residue analysis. Because o f the thermal instability o f most carbamates when analyzed by gas chromatography (GLC), indirect detection methods are becoming popular. Two o f the more promising approaches are detection as derivatives o f the phenol hydrolysis product and that o f the intact carbamate. Although the latter has been used only to a limited extent (see review by Williams 1971) it appears very promising and has distinct advantages over the former, especially in metabolite detection. T h e gas c h r o m a t o g r a p h i c e v a l u a t i o n o f two N - d e r i v a t i v e s , the N - a c e t y l - C l x ( x = 0 , 1 , 2 , 3 ) andN-nitroso derivatives is reported here using carbaryl as model compound a n d p , p ' - D D T as the comparison standard. Carbaryl was used because it is one o f the most popular carbamate insecticides and apparently one of the most unstable to GLC conditions. These derivatives were developed primarily for G L C with detectors specific for chlorine or nitrogen but the electron capture detector may also be used. Archives of Environmental Contamination and Toxicology Vol. 5, No. 2 185-190 (1977) 1977 by Springer-VerlagNew York Inc.
185
186
E. D. Magallona and F. A. Gunther
Experimental Synthesis and/or purification of experimental compounds. Experimental compounds were synthesized, purified, and characterized as described in Magallona et al. (1976). Microcoulometric G L C . The gas chromatograph, operated in the reductive mode, consisted of a Dohrmann Instruments Co. Model C-200 microcoulometer, S-200 pyrolysis furnace and a T-300 S titration cell attached to a Loenco, Inc. Model 15-C-E column oven by an independently heated quartz transfer line. Nitrogen was the carrier gas and hydrogen was the reaction gas. Thermal stability of the chlorinated acetyl derivatives was determined by injecting known quantities of the derivatives a n d p , p ' - D D T and plotting detector response percent of theoretical versus ng of chlorine injected. Calibration curves were also constructed for the N-chloroacetyl and N-dichloroacetyl derivatives of carbaryl (C-CIt and C-C12, respectively) with the detector in the sensitive mode to determine detector linearity and the minimum detectable ranges. Solution conductivity G L C . Detection was in the reductive mode using a Coulson Instruments Co. Model 1 gas chromatograph described by Coulson (1965) and incorporating the suggestions of Patchett (1970). Hydrogen was the carrier and reaction gas. Thermal stability of experimental compounds was determined by constructing calibration curves and comparing detector response using varying column temperatures to obtain essentially the same retention times. Electron capture G L C . A Varian Aerograph Model 1700 gas chromatograph, operated isothermally and equipped with a tritium foil, was used, with nitrogen as the carrier gas. Calibration curves were constructed to determine linearity ranges, minimum detection limits, and detector response to experimental compounds relative t o p , p ' - D D T . Comparative response was expressed as relative response factor (RRF) defined as: nanomoles of derivative to give response x RRF = nanomoles ofp,p' -DDT to give response x
Results and discussion Microcoulometric G L C . p,p"DDT was used as the standard material for evaluating the thermal stability and detector linearity for experimental materials because of its known GLC characteristics. A fortuitous consequence of this choice was the comparable retention times of the compound with C-C12 and N-trichloroacetyl carbaryl (C-CIa) in several columns. This minimized variations in recovery due to differences in column temperatures or delays due to the need to adjust column temperature to obtain comparable retention times. In an attempt to find the optimum reaction gas flow rate for sample combustion, detector response to constant sample size was plotted as a function of total gas flow with the carrier gas flow kept constant. As can be seen in Figure 1, detector response was essentially constant for the 130 to 220 ml per rain range of total gas flow. A total gas flow of
GLC Evaluation of N-Derivatives of Carbaryl
187
100 9
9
=
DDT
-=
C-CI=
,n
C-Ch
80.
6O
4O
Q
r
o
20.
1~o
1~o
1~o
2~o
2~,o
F l o w rate (ml/min)
Fig. I. Microcoulometric detector response top,p'-DDT and twoN-acetyl-Clx derivatives of carbaryl as a function of reaction gas flow rate using a 6-foot, 3% DC-200 column at 205~ 200 ml per min was therefore arbitrarily adapted. Although response was unaffected by gas flow rates, it appears that both carbaryl derivatives have been subject to thermal decomposition, especially CI-C13. For C-C12 and C-C13, detector response also increased with the number of injections. The preliminary injection of large slugs of sample (about 3 x 30/zl of 100/zg/ml solution) adequately "conditions" the column for C-C12 so that successive injections gave acceptable reproducibility. On the other hand, response with C-Ci3 was still not reproducible and continued to increase with the number of injections. The data in Figure 1 for this compound were therefore taken from the fourth injection after "conditioning" the column. Detector response for the derivatives as a function of quantity of material injected (Figure 2) also showed lower recovery compared to p,p'-DDT and increased response with increase in quantity of chlorine injected. The latter was indicative of thermal decomposition. With C-C13 in particular therefore, the theoretical advantages in its use seem to be negated by its thermal instability and further work on this derivative was discontinued. Although it appears that the N-acetyl-Clx derivatives of carbaryl underwent significant thermal decomposition, it was still impo~ant to find the detectability limit and linearity range of the microcoulometric GLC (MC-GLC) especially for C-CIt and C-C12. This is because differences in "recovery" may be due, at least in part, to structural differences which influence completeness of combustion or formation of titrable chlorine. The calibratipn curve obtained by operating the MC-GLC in the sensitive mode (furnace effluent impinging on the sensor electrode) shows only slight upward curvatures and 25 and 15 ng of C-CI1 and C-C12, respectively, give about 8% full-scale recorder deflection (Figure 3). Therefore, from the standpoint of residue detection, these derivatives could be used. Solution conductivity GLC. This detector, with its high specificity for nitrogen, is well suited for the evaluation of the thermal stability of carbamates and/or carbamate derivatives. Since the theoretical basis for the quantitative relationship between electrolytes introduced and detector response has not been established, and empirical approach based
E. D. Magallona and F. A. Gunther
188
1(111. 80.
!. Q
=
,
DOT
C-CI=
,7-- 60-
~ ~
~
, . ~ . . . _ ~
~
C-CIt
/ r
40-
x
C-CI=
20-
1O0
200
300
Chlorine
(ng)
Fig. 2. Microcoulometric detector response to p, p'-DDT and the N-acetyi-C I x derivatives of carbaryl as a function of quantity of material injected using a 4-foot, 3% SE-30 column.
on the construction of calibration curves using stock solutions of pure compounds, has to be resorted to. Differences in slopes of these curves, plotted as detector response v e r s u s ng of nitrogen, would reasonably reflect differences in thermal stabilities.
From Figure 4, it was evident that with the possible exception of N-acetyl carbaryl (C-CIo), the derivatives of carbaryl underwent thermal decomposition with the GLC parameters used. For these derivatives, thermal stability appears to be related to degree of chlorination. The N-nitroso derivative (C-NO) has to be ruled out for use in carbaryl detection because of its thermal instability. The data obtained here are also confirmed by mass spectroscopy-GLC.
/•0-CII
1000
ol r
--~
500
. . . .
i
9
-
9
50 Derivative
t
100
.
.
.
.
i
150
injected
(ng)
Fig. 3. Response of the microcoulometric detector operating at sensitive mode to N-monochloroand N-dichloro-acetyl carbaryl using a 2-foot, 3% DC-200 column.
GLC Evaluation of N-Derivatives of Carbaryl
1200-
. c-c~,0920c)
1000~
189
/
/
/ /
8oo.
~
//
C-CI,(205oc)
/ C-C,(210"C)
4OO
200
~
~
C-NO(135~
Nitrogen (ng)
Fig. 4. Comparativeelectrolytic conductivity detectorresponseto carbaryl and its N-derivatives using a 3.5-foot, 2% OV-17 column.
t
/
DOT(183oc)
'~176 t /
. C'CI,(174~)
//~
10 Nanomoles
c'c' ('a~176
20 x 10~
Fig. 5. Comparative electron capture response to p, p'-DDT and the N-acetyl-C Ix derivatives of carbaryl using a 4-foot 3% DC-200 column.
Electron capture G L C . Since the electron capture detector has found use in the detection of pesticides down to the picrogram range, it was decided to find the detectability limits of the derivatives with this detector. This detectability limit is expressed relative to p,p'-DDT or any other standard compound. Temperatures were varied for each compound to obtain a retention time in the 3.2 to 3.5 min range because preliminary work had shown that detector response increased with retention time. Figure 5 shows that detector response was linear for p , p ' - D D T , C-CI0, and C-Cll, whereas response to C-CIz was non-linear. This curvature could be due to column decomposition or may be influenced by other factors discussed by Gaston (1964). Detector response to C-CIa was no longer determined because it was found to decompose by MC-GLC and detector response increased with increasing number of injections. Under our conditions of use, C-CII gave an RRF of 4.0 while with C-C10, this was 5.6. On the weight basis, the lower limits of detection were about 0.1 ng for both derivatives.
190
E. D. Magallonaand F. A. Gunther References
Coulson, D. M.: Electrolytic conductivity detector for gas chromatography. J. Gas Chromatog. 3, 134 (1965). Gaston, L. K.: Gas chromatography using an electron absorption detector. Residue Reviews 5, 21 (1964). Magallona, E. D., F. A. Gunther, and Y. Iwata: Synthesis and spectroscopic characterization of acetyl-Cl~ = 0,1,2 or 3) and nitroso derivatives of carbaryl and three possible metabolites. Arch. Environ. Contamin. Toxicol. 5, 177 (1976). Patchett, G. G." Evaluation of the electrolytic conductivity detector for residue analysis of nitrogen-containing pesticides. J. Chromatog. Sci. 8, 155 (1970). Williams, I. H.: Carbamate insecticide residues in plant materials determination by gas chromatography. Residue Reviews 38, 1 (1971). Manuscript received April 16, 1975; accepted August 24, 1975
