Screening and optimization of the derivatization of polar herbicides with trimethylanilinium hydroxide for GC-MS analysis

J. Biochem. Biophys. Methods 69 (2006) 3 – 14 www.elsevier.com/locate/jbbm

Screening and optimization of the derivatization of polar herbicides with trimethylanilinium hydroxide for GC-MS analysis A. Ranz, E. Lankmayr * Graz University of Technology, Institute of Analytical Chemistry and Radiochemistry, Technikerstr. 4, A-8010 Graz, Austria Received 1 December 2005; received in revised form 19 January 2006; accepted 9 February 2006

Abstract In the present study, a derivatization method for the determination of acidic herbicides has been investigated. This procedure involves the methylation with the quaternary ammonium salt trimethylanilinium hydroxide (TMAH) directly in the gas chromatographic auto-sampler vial for analysis by gas chromatography combined with mass spectrometry. The derivatization reaction has been screened for influential factors and statistically significant parameters. The identified factors, reaction time, temperature and hold-up time were optimized by a complete factorial response surface design and optimal reaction conditions were generated. Finally, the optimized methylation procedure was compared to different alkylation methods and obtained results demonstrated the applicability of derivatization with trimethylanilinium hydroxide. Acidic herbicides used in the study consist of several families of compounds like derivatives of acetic acid (2,4-D and 2,4,5-T), butanoic acid (MCPB), benzoic acid (chloramben, dicamba), phenol (dinoseb and dinoterb), propanoic acid (mecoprop) and other miscellaneous acids such as pyridinecarboxlyic acid (picloram). A reliably working, rapid method for the preparation of methyl compounds is generated with respect to automation for routine analysis. D 2006 Elsevier B.V. All rights reserved. Keywords: Acidic herbicides; Derivatization; Gas chromatography; Optimization; Screening; Trimethylanilinium hydroxide

* Corresponding author. Tel.: +43 316 873 8305; fax: +43 316 873 8304. E-mail address: [email protected] (E. Lankmayr). 0165-022X/$ - see front matter D 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jbbm.2006.02.007

4

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

1. Introduction Parallel to the increase of agricultural productivity, the environmental insertion of pesticides has led to a more and more important challenge to environmental analysis. Among the pesticides, polar herbicides play the most important role in agriculture, forestry and industrial weed application. Owing to their physical properties, they are frequently detected in ground and surface waters [1]. Due to the fact that groundwater is a major source for drinking water, monitoring of acidic and polar herbicides is necessary, also because of their persistence in the environment and toxicity. Acidic herbicides used in the present study consist of several families of compounds to investigate the derivatization procedure for a large range of different chemical classes. This compounds are derivatives of acidic functional groups including chlorophenoxy acid (2,4-D, MCPB, mecoprop and 2,4,5-T), benzoic acid (chloramben, dicamba and picloram) and dinitrophenol (dinoseb and dinoterb). It is well known that herbicides with the carboxylic acid directly at the aromatic system like chloramben, dicamba and picloram cause more problems regarding esterification than derivatives of the chlorophenoxy acid. Steric hindering effects at the active hydrogen are also given in the case of dinoseb and dinoterb. These herbicides are characterized by relatively high water solubility and high physiological and biological activity. So they can affect toxicological effects, even in low concentrations [2]. For analysis of herbicides, gas chromatography (GC) is in common use owing to the unsurpassed limits of detection. Among the detection methods available for GC, mass spectrometry (MS) plays a leading role due to its sensitivity and selectivity. Owing to their polarity, acidic herbicides are usually less volatile and thermally instable. For gas chromatographic analysis they must be derivatized therefore [3]. Derivatization is usually performed after the sample preparation steps pre-concentration and extraction [4]. One of the most often used derivatizing reagents to prepare methyl ester derivatives and also implemented into a standard method of the Environmental Protection Agency (EPA) is diazomethane in spite of its disadvantages [5–8]. Diazomethane is explosive (N90 8C), toxic, carcinogenic and an irritant. And because of its instability it has to be generated freshly each time. A less hazardous substitute for diazomethane, which has been shown to react quickly at room temperature, is trimethylsilyldiazomethane [9,10]. These two methods are pointed out, because they are chosen for a comparison to methylation with trimethylanilinium hydroxide. Quaternary ammonium salts like tetramethylammonium hydroxide, tetramethylammonium acetate, tetramethylsulfonium hydroxide or trimethylanilinium hydroxide have to been shown to be alternatives to traditionally classical derivatization procedures [11,12]. Depending on the class of the analytical compounds, different methods and reactions have been developed and published. Methylation with quaternary ammonium salts is described as pyrolytic methylation, thermally assisted methylation [13] and injection port methylation or intra-injector derivatization [14–17]. Studies on the application of tetramethylammonium hydroxide for methylation were conducted by McKinney et al. [18] reporting a batch thermochemolyses. Different tetraalkyland phenyltetraalkyl-ammonium salts are quite readily available, an optimum reagent can be chosen for a given analytical task and compound [19,20]. In this study, trimethylanilinium hydroxide is the reagent of choice, because dimethylanilinium is a better leaving group in this mechanism [21]. The quaternary ammonium compound replaces the active hydrogen of an acid herbicide by a methyl group as shown in Fig. 1. Methylation is realized directly in a gas chromatographic auto-sampler vial with respect to a

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14 CH3 OH-

O R

C

+

+

H3C N CH3

CH3

O R

O H

+

H3C N CH3

C

5

+

H2O

O

H3C

O R

+

C

CH3 N

OCH3

Fig. 1. Derivatization of a carboxylic acid with trimethylanilinium hydroxide.

reliably working, rapid method for the preparation of methyl compounds. First experiments have resulted to better reaction yields and recoveries for this procedure than for pure injection port reactions. This method has been generated with regard to automation for routine analysis and can accomplish this analysis without a working step between derivatization and injection into the gas chromatographic system. In the present study, this method is investigated concerning influential parameters like temperature, reaction time, hold-up time, which means the time between end of reaction and injection, stirring and reagents. An optimization using a response surface design is followed and a validation of this procedure was subject of this study. 2. Materials and methods 2.1. Reagents and chemicals Pesticide standards were supplied by Labor Dr. Ehrenstorfer (Augsburg, Germany): chloramben (3-amino-2,5-dichlorobenzoic acid; purity: 98.5%), dicamba (3,6-dichloro-2methoxybenzoic acid; purity: 98.3%), dinoseb (2-(1-methylpropyl)-4,6-dinitrophenol; purity: 99.0%), dinoterb (2-(1,1-dimethylethyl)-4,6-dinitrophenol; purity: 97.0%), MCPB (4-(4-chloro2-methylphenoxy)butanoic acid; purity: 99.9%), mecoprop (2-(4-chloro-2-methylphenoxy)propanoic acid; purity: 99.8%), picloram (4-amino-3,5,6-trichloro-2-pyridinecarboxylic acid; purity: 97.5%). 2,4-D ((2,4-dichlorophenoxy)acetic acid; purity: 99.0%) and 2,4,5-T ((2,4,5trichlorophenoxy)acetic acid; purity: 99.0%) were purchased from Riedel-de Hae¨n (Seelze, Germany). All ester–respectively ether–reference compounds (-ME) were supplied by Labor Dr. Ehrenstorfer (Augsburg, Germany): chloramben-ME in cyclohexane (100 ng AL 1; purity: 99.0%), 2,4-D-ME (purity: 96.0%), dicamba-ME in n-hexane (100 ng AL 1; purity: 99.0%), dinoseb-ME in methanol (100 ng AL 1; purity: 99.5%), dinoterb-ME in cyclohexane (10 ng AL 1; purity: 99.5%), MCPB-ME (purity: 97.5%), mecoprop-ME (purity: 98.5%), picloram-ME in cyclohexane (100 ng AL 1; purity: 98.4%), 2,4,5-T-ME (purity: 98.0%). Stock standard solutions were prepared and stored at 4 8C protected from light. Trimethylanilinium hydroxide in methanol (0.1 M; puriss.) was acquired from Fluka (Buchs, Switzerland). Cyclohexane and methanol were supplied by Baker (Deventer, The Netherlands) (purity: ultra resi-analyzed) and ethyl ether was acquired from Merck (Darmstadt, Germany). NMethyl-NV-nitro-N-nitrosoguanidine (purity: z 97%) and trimethylsilyldiazomethane (2 M;

6

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

techn.) were supplied by Fluka (Buchs, Switzerland), KOH was purchased from Merck (Darmstadt, Germany). Nitrogen (purity: 99.9990%) and helium (purity: 99.9990%) were obtained from Air Liquide (Graz, Austria). 2.2. Derivatization 2.2.1. Derivatization with trimethylanilinium hydroxide The derivatization reaction with trimethylanilinium hydroxide was carried out under the following conditions: all measurements for screening design, optimization and validation were accomplished with an herbicide concentration of 200 pg AL 1 in cyclohexane. Derivatization was performed directly in the gas chromatographic auto sampler vial by adding 1 mL of solvent, herbicide stock solution and quaternary ammonium salt. For screening design, a 10-fold (0.9 AL trimethylanilinium hydroxide solution) and a 50-fold (4.5 AL trimethylanilinium hydroxide solution) excess were chosen. To realize stirring, reaction was prepared under magnetic stirring. Sodium sulfate was added. For the response surface design derivatization was done with a 20-fold excess of trimethylanilinium hydroxide for different times and temperatures. Higher temperatures were realized by heating the reaction solution directly in the auto-sampler vial in a heated aluminium block. 2.2.2. Derivatization with diazomethane Generation of diazomethane: Preparation is performed in a diazomethane generator, which was maintained at 0 8C. A quantity of 1 mL ethyl ether was added to the collection vial and 80.0 mg N-methyl-NV-nitro-N-nitrosoguanidine to the second impinger. After connecting the generator 220 AL deionized water and 300 AL KOH (5 M) were added with a syringe and within 30 min a 0.1 molar solution of diazomethane was generated. This solution may be used over a period of 48 h. Derivatization: Using a Pasteur pipette, diazomethane solution was added to the reaction solution until the solution remains slightly yellow in colour. After 30 min, unreacted diazomethane was removed under a nitrogen flow. 2.2.3. Derivatization with trimethylsilyldiazomethane 1 mL cyclohexane and stock solution were added to an auto-sampler vial and the herbicides were esterified with 20 AL trimethylsilyldiazomethane solution after shaking vigorously at a vortex and sonicated for 30 min. 2.3. Analysis and quantification Analysis was carried out by gas chromatography combined with mass spectrometry (GCMS). GC-MS-determination was performed by using an Hewlett-Packard (Waldbronn, Germany) HP 6890 gas chromatograph equipped with an HP7683 auto-sampler and a split/ splitless injector operated in splitless mode (purge delay: 0.60 min; purge flow: 35.0 mL min 1; pressure: 68.7 kPa). The injector was equipped with a single taper glass insert and kept at 250 8C. An injection volume of 1 AL was selected for all analyses. The capillary column used was an HP-5MS, 30 m  0.25 mm i.d. and 0.25 Am film thickness. Helium at a constant flow rate of 1.1 mL min 1 was used as carrier gas. The oven temperature was: 60 8C for 1.00 min ramped to 160 8C at 100 8C min 1, to 210 8C at 7 8C min 1 and held for 1 min, to 220 8C at

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

7

Table 1 Retention times and selected ions for MS-detection in SIM mode Derivative

Retention time (min)

Target-ion (m/z)

Qualifier (m/z)

Chloramben-ME 2,4-D-ME Dicamba-ME Dinoseb-ME Dinoterb-ME MCPB-ME Mecoprop-ME Picloram-ME 2,4,5-T-ME

8.76 7.22 6.06 9.73 9.57 8.94 6.27 10.98 8.94

188 199 203 225 239 101 169 196 233

190 201 205 254 254 107 171 198 235

219 234 234 195 209 142 228 254 270

221 236 188

230

10 8C min 1 and held for 1 min, to 300 8C at 100 8C min 1 and held for 8 min. The gas chromatograph was coupled to an HP5973 mass selective detector with electron impact ionization, operated in single ion monitoring (SIM) mode using the m/z values listed in Table 1. The interface temperature was maintained at 280 8C. Analytes were quantified by linear calibration at seven concentration levels in the range from 10 to 500 pg AL 1 using methyl esters as standards in solutions. 2.4. Experimental design for screening and optimization A fractional factorial design was performed for the derivatization. The screening was carried out in one series of 24 experiments. The parameter setting of the six factors is given in Table 2. A 33 response surface design with a 3-fold repetition of the centre was generated for optimization. The design matrix of the response surface design is shown in Table 3. For statistical calculations, the software package STATGRAPHICS PLUS Version 3 for Windows (Manguistics, Rockville, USA) was used for creating experimental designs and analyzing experimental data. 3. Results 3.1. Screening Derivatization was investigated concerning influential parameters followed by an optimization of the complete methylation procedure. Therefore, a screening design and a response surface design were used to show the most significant factors of this reaction. The first purpose was to select the most important main effects on the derivatization procedure. For that reason, a fractional factorial design for six factors was generated in a first approach. The factors of interest Table 2 Factors and their settings for the screening design Factor

Minimum

Maximum

Temperature (8C) Hold-up time (min) Excess reagent Reaction time (min) Na2SO4 Stirring

60 0 10 30 No (0) Off (0)

100 120 50 90 Yes (1) On (1)

8

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

Table 3 Response surface design matrix for the optimization with three-fold repetition of the centre Run

Reaction time (min)

Temperature (8C)

Hold-up time (min)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

30 90 30 60 60 30 60 30 30 60 90 60 60 30 90 60 30 90 90 30 60 60 90 90 30 60 90 60 90

120 70 70 95 70 95 95 120 70 120 95 95 95 95 120 120 70 70 70 95 70 120 95 95 120 70 120 95 120

240 10 10 125 125 240 125 10 125 10 125 125 240 125 10 125 240 125 240 10 240 240 10 240 125 10 240 10 125

were reaction time, reaction temperature, hold-up time, excess of trimethylanilinium hydroxide, as well as stirring and drying over Na2SO4. Maximum reaction yield in percent of each compound was chosen as target function for both, screening and optimization. A computationally obtained Analysis of Variance (ANOVA) table reports statistically significant effects. The variability for each compound is parted into separate pieces for each of the effects. The statistical significance of each effect is tested and the result is given in Table 4. The P-value in Table 4 serves as a measure of significance as a component of this ANOVA table. Bold numbers show significant factors as identified by P-values less than 0.05 indicating that they are different from 0 at 95.0% confidence level. From the calculated results, it can be clearly seen that temperature exhibits the biggest influence on the derivatization—strongly positive and for all investigated herbicides. Different results are obtained for the other factors. For some compounds also reaction time is a statistically significant parameter, particularly for those which are not so easily derivatized like the derivatives of benzoic acid and the dinitrophenols. The other four investigated factors were not significant for nearly all herbicides. For visualization of the result of the screening design, Pareto charts and main effect plots were drawn. In Fig. 2, one example for each chemical compound class is given. Main effect plots show the most influential parameters by drawing a line between the low and high levels of the factor setting. The slope and the length of this line are proportional to the effect; the higher the

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

9

Table 4 Values and R 2 obtained from ANOVA

Chloramben 2,4-D Dicamba Dinoseb Dinoterb MCPB Mecoprop Picloram 2,4,5-T

Reaction time

Hold-up time

Excess

Temperature

Na2SO4

Stirring

R 2 (%)

0.0988 0.1358 0.4362 0.0288 0.0243 0.5407 0.4453 0.0366 0.0588

0.3072 0.9283 0.6166 0.4539 0.4344 0.4690 0.4275 0.4465 0.8615

0.8601 0.7269 0.2041 0.1323 0.1174 0.2933 0.3592 0.9077 0.3421

0.0000 0.0000 0.0000 0.0002 0.0004 0.0000 0.0000 0.0001 0.0000

0.4204 0.1057 0.3266 0.3233 0.2834 0.7393 0.5562 0.9997 0.1631

0.2304 0.9987 0.9259 0.4871 0.4267 0.7411 0.9608 0.3395 0.5073

92.47 90.15 87.54 66.47 64.34 82.23 85.62 65.56 95.03

Bold numbers indicate statistically significant factors.

slope, the higher is the influence of the corresponding parameter. The direction of the line specifies a positive or negative influence of the parameter. Pareto charts give a useful supplementation to the main effect plots. The length of each bar is proportional to the standardized effect and bars of the most influential factors are grouped at the top of the list. Standardized effects show each effect divided by its standard error. The chart includes a vertical line for testing the significance, so factors overpassing this line exert a statistically significant influence on the result within the 95.0% confidence level. 3.2. Optimization Optimization by means of a response surface design is a process for locating optimum conditions in a more dimensional parameter space. Therefore, three experimental factors, identified as significantly influential by the screening design, have been investigated, namely reaction temperature, reaction time and hold-up time. Hold-up time was chosen in a bigger time range than for screening design to exclude another time factor. Similar to the fractional factorial design, maximum reaction yield was chosen as target function for optimization and a 33 response surface design with a 3-fold repetition of the centre was generated to obtain the optimum of the experimental region. Hypersurfaces, as they are expressed in Eq. (1), allow defining optimum conditions and showing the reaction yield as a function of the investigated factors. y ¼ b0 þ

k X

b i xi þ

i¼1

k X 1ViVj

bij xi xj þ

k X

bii x2i

ð1Þ

i¼1

where y is the reaction yield, x i and x j are the variables considered for the optimization and b i and b j are the parameters to be calculated. For a 33 response surface design this leads to the function given in Eq. (2): y ¼ b0 þ b1 x1 þ b2 x2 þ b3 x3 þ b12 x1 x2 þ b13 x1 x3 þ b23 x2 x3 þ b11 x21 þ b22 x22 þ b33 x23 ð2Þ x1 x2 x3

Reaction time Temperature Hold-up time.

10

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

Pareto Chart for Picloram

Main Effects Plot for Picloram 50 Reaction Yield [%]

Temperature Reaction time Stirring Hold-up time Excess Sodium Sulfate

40 30 20 10 0

0

1

2 3 4 Standardized effect

5

6

Sodium sulfate Excess Reaction time Stirring Hold-up time Temperature

Pareto Chart for Dinoseb

Main Effects Plot for Dinoseb 18 Reaction Yield [%]

Temperature Reaction time Excess Sodium sulfate Hold-up time Stirring

15 12 9 6 3 0

0

1

2 3 4 Standardized effect

Sodium sulfate Reaction time Excess Stirring Hold-up time Temperature

5

Pareto Chart for Mecoprop

Main Effects Plot for Mecoprop 119 Reaction yield [%]

Temperature Excess Hold-up time Reaction time Sodium sulfate Stirring

109 99 89 79 69 59

0

2

4 6 8 Standardized effect

10

Sodium sulfate Reaction time Excess Stirring Hold-up time Temperature

Fig. 2. Main effect plots and Pareto charts of the screening design for picloram, dinoseb and mecoprop.

One result of the constructed hypersurfaces is the R 2-value, which measures the proportion of variability in the model for the dependent variables. An approximation of R 2 to 100% is indicative for a correct consideration of all experimental parameters. Both values, R 2 and P, are shown in Table 4, whereas bold numbers are indicative of statistically significant factors identified by the ANOVA at the 95% confidence level. Similar to the screening design, Pareto charts and main effect plots let visually identify the most important effects. One example is given in Fig. 3 for the analyte mecoprop. Interactions between parameters like AB, AC and BC are also listed; interactions between one and the same factor like AA, BB and CC are just result of the mathematic model without any physical relevance. The result of the analysis of variance identified the temperature as the most significant factors for all target compounds. Significant for nearly all herbicides is the time of reaction with the sole exception of the chlorophenoxy acid mecoprop having a factor bar nearly at the vertical

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

Pareto Chart for Mecoprop

11

Main Effects Plot for Mecoprop 135 Reactin yield [%]

B:Temperature BB AB A:Reaction time CC C:Hold-up time AC BC AA

125 115 105 95 85

0

2

4 6 Standardized effect

30.0

8

90.0 70.0

Reaction time

120.0 10.0

Temperature

240.0

Hold-up time

Fig. 3. Main effect plots and Pareto charts of the response surface design.

line. Without influence on the reaction yield is the hold-up time, a fact, which makes the procedure much easier because of the drop out of another time dimension. It can be clearly seen that two of the investigated variables exhibit a distinguished influence on the derivatization efficiency, but not alone the factors, also the interaction between those parameters, reaction time and temperature. The calculated R 2-values are around 79% to 96%, even higher than 90% for most of the investigated compounds. This indicates that the results from the computation are sufficiently correct to describe the effect of the variables in the parameter space. The result of the optimization allows determining and defining optimum conditions for each compound. Since the aim of the response surface design was to provide the overall best instrumental conditions for all of the target compounds, as a result of compromise, the combination of 75 min reaction time, 105 8C reaction temperature and 10 min hold-up time were considered to provide the best conditions (Table 5). 4. Discussion In order to verify the performance of derivatization with trimethylanilinium hydroxide, this procedure was compared to two different established procedures using diazomethane and trimethylsilyldiazomethane. For comparison purposes, measurements were done four times for each methylation method at a concentration of 200 pg AL 1. For all measurements, a Grubb’s Test for outliers was performed (T(0.95; 4)). Derivatization with trimethylanilinium hydroxide was accomplished with a parameter setting which resulted from optimization. Conditions for derivatization with diazomethane and Table 5 P-Values obtained from ANOVA: (A) reaction time, (B) temperature, (C) hold-up time

Chloramben 2,4-D Dicamba Dinoseb Dinoterb Mecoprop MCPB Picloram 2,4,5-T

A

B

C

AA

AB

AC

BB

BC

CC

0.011 0.043 0.024 0.001 0.002 0.071 0.146 0.047 0.010

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

0.304 0.186 0.901 0.868 0.608 0.569 0.039 0.114 0.231

0.406 0.368 0.616 0.117 0.053 0.981 0.802 0.579 0.203

0.029 0.042 0.011 0.002 0.001 0.008 0.005 0.018 0.088

0.853 0.771 0.375 0.706 0.769 0.625 0.639 0.209 0.589

0.337 0.000 0.000 0.000 0.000 0.000 0.005 0.000 0.000

0.883 0.713 0.912 0.575 0.295 0.652 0.136 0.143 0.878

0.701 0.311 0.577 0.572 0.497 0.306 0.426 0.684 0.214

Bold numbers indicate statistically significant factors.

12

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

Table 6 Reaction yields for the optimized derivatization with trimethylanilinium hydroxide (TMAH) with trimethylsilyldiazomethane (TMS-D) and diazomethane in percent reaction yield and standard deviation (n = 4)

Dicamba-ME Mecoprop-ME 2,4-D-ME Chloramben-ME MCPB-ME 2,4,5-T-ME Dinoterb-ME Dinoseb-ME Picloram-ME

TMAH

TMS-D

Diazomethane

91.1 F1.2 118.0 F 2.0 90.9 F 1.4 71.7 F 4.7 108.9 F 2.0 88.0 F 2.0 41.2 F 3.0 45.0 F 2.1 36.7 F 2.9

79.6 F 0.9 106.5 F 0.6 73.7 F 1.2 63.1 F1.1 84.3 F 2.9 66.2 F 0.8 27.1 F1.1 35.2 F 0.2 50.2 F 6.0

52.2 F 8.1 77.6 F 5.9 57.2 F 9.4 63.5 F 5.3 66.5 F 4.1 56.8 F 7.7 32.2 F 6.5 43.3 F 4.6 56.6 F 10.3

trimethylsilyldiazomethane are described in Section 2. The results for all three methods are listed as means and standard deviations in Table 6. The data obtained by derivatization with trimethylanilinium hydroxide are in good agreement with those obtained by methylation with diazomethane and trimethylsilyldiazomethane. Overall, the quaternary ammonium salt yields a better result for most of the target compounds. Finally, the limits of detection (LOD) and limits of quantification (LOQ) of the TMAH derivatization procedure were determined. Calculation was performed with the Excel Macro Validata Version 3.02.54ger (Wegscheider-Rohrer-Neubo¨ck, Leoben, Austria) at the 95% confidence interval, following the Eurachem/CITAC Guide [22]. Therefore, a 3-fold repetition of six concentration levels was performed. These results are listed in Table 7. With the purpose of monitoring the progress of the derivatization reaction after defined times, methylation with trimethylanilinium hydroxide was investigated by high performance liquid chromatography (HPLC). Because of the use of a hot injection port for gas chromatography and the possibility of a partially injection port derivatization, monitoring of the reaction just by heating in the aluminium block is not possible with gas chromatography. Therefore, the methylation of the chlorophenoxy acid mecoprop representing the investigated herbicides was screened with HPLC. As can be seen from Fig. 4, derivatization has already started after a few minutes and was nearly completed after 75 min. No special emphasis was given to a reduction of adsorption phenomena causing peak tailing. On the left-hand side of the chromatogram the peak of the acid is shown, vice versa on the right side the ester of mecoprop. For visualization of the course of derivatization, chromatograms of different time levels are overlaid. Recapitulating,

Table 7 Limit of detection (LOD) and limit of quantification (LOQ)

Mecoprop 2,4-D Dicamba Dinoseb Dinoterb Chloramben MCPB Picloram 2,4,5-T

LOD (pg AL 1)

LOQ (pg AL 1)

14.50 25.73 20.74 36.85 38.86 62.87 34.14 45.03 29.65

52.27 92.23 74.51 128.97 130.51 222.21 122.07 159.42 106.01

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14 0 min 30 min 75 min 90 min

mAU 12

13

acid

10 8

ester

6 4 2 0 -2 2.5

3

4

4.5

5

5.5

6

6.5

Fig. 4. HPLC chromatogram of mecoprop. HPLC: Agilent 1100 system with UV-detection at 230 nm. Column: RP-C18 (ODS2; 4,6  150 mm; 5 Am). Solvent: (A) MeOH, (B) 1 % H3PO4. Linear gradient from 46% A to 57% A in 9 min, held for 4 min, to 62% A in 7 min, held for 8 min, to 46% A in 3 min.

methylation is completed after heating for 75 min in the aluminium block directly in the autosampler vial at temperatures of 105 8C. 5. Conclusion Derivatization with trimethylanilinium hydroxide for the determination of acidic herbicides is a powerful, fast and high throughput methylation procedure. In combination with gas chromatography coupled with mass spectrometry, it is a reliably working method for the preparation of esters and ethers of acidic compounds with respect to automation for routine analysis. It could be shown that it is possible to methylate these substrates directly in the gas chromatographic auto-sampler vial without any working steps between methylation and injection. As could be shown, temperature and time of reaction were the most important and significant factors. After a careful screening design and optimization of the derivatization reaction, best conditions were found as a result of compromising for all target compounds in a combination of 75 min time of reaction, 105 8C reaction temperature and 10 min hold-up time. A comparison to different alkylation procedures demonstrated the applicability of the optimized derivatization method. References [1] Stetter J, Lieb F. Innovation in crop protection: trends in research. Angew Chem Int Ed 2000;39(10):1725 – 44. [2] Cuong NV, Bachmann TT, Schmid RD. Development of a dipstick immunoassay for quantitative determination of 2,4-dichlorophenoxyacetic acid in water, fruit, and urine samples. Fresenius’ J Anal Chem 1999;364(6): 584 – 9. [3] Baugh PJ. Gaschromatographie, Vieweg, Braunschweig/Wiesbaden, Germany; 1997. [4] Rompa M, Kremer E, Zygmunt B. Derivatisation in gas chromatographic determination of acidic herbicides in aqueous environmental samples. Anal Bioanal Chem 2003;377(4):590 – 9. [5] Ngan F, Ikesaki T. Determination of nine acidic herbicides in water and soil by gas chromatography using an electron-capture detector. J Chromatogr 1991;537(1–2):385 – 95. [6] Ngan F, Toofan M. Modification of preparation of diazomethane for methyl esterification of environmental samples analysis by gas chromatography. J Chromatogr Sci 1991;29(1):8 – 10. [7] U.S. Environmental Protection Agency. Method 515.3 1996. [8] U.S. Environmental Protection Agency. Method 8041 1996.

14

A. Ranz, E. Lankmayr / J. Biochem. Biophys. Methods 69 (2006) 3–14

[9] Rimmer DA, Johnson PD, Brown RH. Determination of phenoxy acid herbicides in vegetation, utilising highresolution gel permeation chromatographic clean-up and methylation with trimethylsilyldiazomethane prior to gas chromatographic analysis with mass-selective detection. J Chromatogr A 1996;755(2):245 – 50. [10] Hayamizu T, Kudoh S, Nakamura H, Methylated N. Vepsiln.-dansyl-l-lysine as a fluorogenic reagent for the chiral separation of carboxylic acids. J Chromatogr B Biomed Sci Appl 1998;710(1–2):211 – 8. [11] Brondz I. Development of fatty acid analysis by high-performance liquid chromatography, gas chromatography, and related techniques. Anal Chim Acta 2002;465(1–2):1 – 37. [12] Amijee M, Cheung J, Wells RJ. Direct on-column derivatization in gas chromatography: III. On-column benzylation reagents and the development of 3,5-bis(trifluoromethyl)benzyldimethylphenylammonium fluoride, an efficient new on-column derivatization reagent. J Chromatogr A 1996;738(1):57 – 72. [13] Pastorova I, van der Berg KJ, Boon JJ, Verhoeven JW. Analysis of oxidized diterpenoid acids using thermally assisted methylation with TMAH. J Anal Appl Pyrolysis 1997;43(1):41 – 57. [14] Brondz I, Olsen I. Intra-injector methylation of free fatty acids from aerobically and anaerobically cultured Actinobacillus actinomycetemcomitans and Haemophilus aphrophilus. J Chromatogr 1992;576(2):328 – 33. [15] Brondz I, Olsen I. Intra-injector formation of methyl esters from phenoxy acid pesticides. J Chromatogr 1992;598(2):309 – 12. [16] Ogierman L. Gas–liquid chromatography of selected benzimidazole fungicides by flash-heater methylation with trimethylanilinium hydroxide. J Chromatogr 1981;210(1):83 – 92. [17] Ogierman L. On-column esterification and N-methylation of carbetamide with N,N,N-trimethylanilinium hydroxide. Pestic Sci 1989;25(2):155 – 9. [18] McKinney DE, Carson DM, Clifford DJ, Minard RD, Hatcher PG. Off-line thermochemolysis versus flash pyrolysis for the in situ methylation of lignin: is pyrolysis necessary? J Anal Appl Pyrolysis 1995;34(1):41 – 6. [19] Rompa M, Kremer E, Zygmunt B. Derivatization of acidic herbicides with selected tetraalkylammonium and trimethylsulfonium hydroxides for their GC analysis. Anal Lett 2004;37(15):3299 – 312. [20] Engewald W. 24th International symposium on chromatography held in Leipzig, Germany September 2002. Chromatographia, vol. 57. 2003. Suppl. [21] Meierhenrich U, Thiemann WH-P, Rosenbauer H. Pyrolytic methylation assisted enantioseparation of chiral hydroxycarboxylic acids. J Anal Appl Pyrolysis 2001;60(1):13 – 26. [22] Eurachem/CITAC guide: quantifying uncertainty in analytical measurement. second edition; 2004.