Investigation and application of liquid chromatography mass spectrometry in the analysis of polar, less volatile and thermal unstable organic pollutants in environmental and biological samples 5

5.2.3 Sample Extraction 5.2.3.1 Optimization strategy for MAE In our previous study, very good results above 95% of recovery had been obtained for propoxur extraction from soil with 30

As already emphasized previously, LC-MS is one of the most powerful techniques for the analysis of polar pesticide residues In the past few years, much work has been done by LC-MS to analyze carbamate pesticides contained in various types of matrices, such as aquatic samples [4-6], and fruits and vegetables [7-11] To our knowledge, however, little effort has been made to analyze carbamates residues in biological samples with LC-API-

MS, except in one study done by Kawasaki et al wherein they used LC-APCI-MS to

analyze eight carbamate pesticides (isoprocarb, metocarb, fenobucarb, xylylcarb, XMC, ethiofencarb, propoxur, and carbaryl) in blood from patients suffering from acute poisoning [12] In addition, liquid-liquid extraction (LLE) has been commonly used to extract carbamates (for example, carbofuran, furathiocarb, benfuracarb, carbaryl, and propoxur) from various kinds of biological samples [13-15] However, LLE has many disadvantages - namely, it requires large amounts of toxic organic solvents and is time-consuming As described above, MAE has become a very attractive sample preparation technique as it overcomes the aforementioned disadvantages Yet, extraction from biological samples by MAE has been limited to the determination of methylmercury [16-19], arsenic speciation [20-22], and fatty acids [23] Currently, there are very few MAE methods available for the analysis of pesticide residues in biological samples, except chlorinated pesticides, which can be detected in a freeze-dried mussel tissues using focused MAE [24]

Based on the above considerations, in this study, LC-APCI-MS technique coupled to MAE, was selected to monitor trace residues of propoxur and its TPs found in biological materials in the aquatic environment Our work is firstly concentrated on the MAE optimization by using orthogonal array design (OAD) procedure Secondly, the application of optimized techniques for the extraction of target pesticides from spiked

biological materials [goldfish, tortoises (Trachemys scripta elegans) and sea lettuce (Ulva lactuca)] is described Finally, the stability studies of propoxur under microwave

irradiation and in tested biological extracts at their respective optimum conditions are discussed

5.2 EXPERIMENTAL

5.2.1 Reagents and Standards

Propoxur was purchased from ChemService (See chapter 2)

Both 2-isopropoxyphenol and N-methylformamide were obtained from Aldrich (Steinheim, Germany) Stock standard solutions were prepared in methanol at concentrations of 1000 µg/ml for each compound and stored at - 4°C Working solutions were prepared by diluting the stock solutions with methanol

5.2.2 Sample Preparations

The preparation of these biological samples is described in detail in Chapter 2

5.2.3 Sample Extraction

5.2.3.1 Optimization strategy for MAE

In our previous study, very good results (above 95% of recovery) had been obtained for propoxur extraction from soil with 30 ml of methanol, an 80 °C extraction temperature, and 6-min microwave heating The differences between the distribution of propoxur and its transformation products in biota and soils, and additionally the difference in the

sample preparation processes (soils were spiked at ambient conditions and air-dried, whereas biota tissues in the present study were freeze-dried), required that there be optimization of extraction step For this, a two-level orthogonal array design (OAD) was applied

In this study, two different experiments were designed to optimize MAE conditions for the extraction of the target According to our mode designed earlier, the four important variables selected for optimizing MAE are: (1) extraction solvents (factor A), (2) extraction temperature (factor B), (3) extraction time (factor C), and (4) solvent volume (factor D) One factor not considered was the power setting for the microwave heating This setting was proportional to the number of samples undergoing extraction during one run Furthermore, according to previous experience with MAE and intuition, one two-variable interactions to be considered was B x C (interaction between different extraction temperature and extraction time) Because four two-level variables and one two-variable interaction were to be considered, a total of five degrees of freedom was required, and the

OA 8 (2 7) matrix was, therefore, chosen so as to have sufficient degrees of freedom for the assignment of the variables under consideration The assignment of the main variables (A, B, C and D), two-variable interaction, and the level setting values of the main variables applied to aquatic animals (goldfish and tortoises) and sea lettuce are displayed in Table 5-1

Table 5-1 Assignment table of variables and the arrangement of the experiment runs

*A: Extraction solvent; B: Extraction temperature; C: Extraction time; D: Volume of

extraction solvent; B x C: Interaction between extraction temperature and time

5.2.3.2 MAE procedures and the treatment of extracts

For every sample in each species extracted by MAE, 1 g of a spiked sample was

accurately weighed out and quantitatively transferred to the PIFE-lined extraction vessel

After adding the needed volume of the extracting solvent (Table 5-1 and Table 5-2),

10-min equilibration was allowed before extraction During optimization of the MAE

conditions, the extraction conditions of each pre-designed experimental trial (a total of 8)

were set according to the two-level OAD procedure given in Table 5-1 and Table 5-2

Table 5-2 The OA8 (27) matrix with the experimental results

* Average response of each level

Column No Average recovery (%)

5.2.4 LC-MS Measurements

The LC-APCI-MS instrument was initially tuned with a tuning solution (a mixture of caffeine, MRFA, and ultromark 1621) in both positive and negative ionization modes In order to ensure optimal tuning conditions, the propoxur standard was delivered into the APCI source through an electronically controlled syringe pump Typical tuning parameters were: vaporizer temp: 450.00 °C, sheath gas flow rate: 80 arb, aux gas flow rate: 20 arb, discharge current: 5.00 µA, capillary temp: 150.00 °C, capillary voltage: 35.00 v, tube lens offset: 5.00 v, and corona voltage: 4.50 kv

Selective monitoring (SIM) was performed at m/z 60 ([M+H]+), 151([M-H]-) and 210 ([M+H]+) The dwell time for each channel was 0.1 s, with an interchannel delay of 0.02

s and a mass span of 1 mass unit Quantitation was based on the area under the peak from the mass chromatogram of the above molecular ions at m/z 60, 151, and 210

For the LC separation of propoxur and its two transformation products (N-methyl formamide and 2-isopropoxyphenol), a mixture of ultrapure water-methanol (50:50) was used as a mobile phase at a constant flow-rate of 0.6 ml/min A Phenomenex ODS 150 x 3.2 mm column was utilized for separation The HPLC system was interfaced with the ion trap through the APCI source Mass spectra collected in full-scan mode were obtained

by scanning over a range of m/z 50 to 250 Maximum injection time was set at 150 ms Time scheduled mass conditions were shown as following scheme:

30 LC time (min) 22

5.3 RESULTS AND DISCUSSION

5.3.1 LC-MS Analysis

The structures of propoxur and its two metabolites (2-isopropoxyphenol and N-methyl formamide) are given in Fig 5-1, together with their typical APCI spectra in full scan mode (m/z 50-250) acquired from direct injections of standard solution of each compound MS2 spectra were recorded by isolating the quasi-molecular ions, followed by

a 30% collision-induced dissociation (CID) energy, and are also presented in Fig 5-1 From Fig 5-1(a), it can be seen in the mass spectrum that the quasi-molecular ion of propoxur at m/z 210 is the most intense peak The MS2 spectrum reveals that two fragment ions were detected One is [M+H-42] + at m/z 168, probably because of the loss

of a 1-propene (C3H6) molecule, the other is the base peak at m/z 153 [M+H-57]+, which probably resulted from the neutral loss of a CH3NCO group MS studies of other carbamate pesticides have reported an identical product ion [M-CONCH3+H]+ [15] that is believed to be a key characteristic of N-methylcarbamate pesticides Fig 5-1 (b) shows that the most abundant ion in the MS spectrum of 2-isopropoxyphenol is [M-H]- at m/z

151 The MS2 analysis of that ion, at m/z 151 with 30% CID energy, shows fragment ions corresponding to cleavage of C-C and C-O bonds A fragment ion at m/z 121 [M-H-30]-was detected as two CH3 groups were lost, whereas the m/z 109 [M-H-42]- probably resulted from the neutral loss of 1-propene, similar to the formation of a fragment ion of propoxur at m/z 168 For N-methylformamide, one peak at m/z 60 [M+H]+ was detected

in the MS spectrum, whereas no MS2 spectrum was obtained, as m/z values of fragment ions are too small to be recorded (Fig 5-1 (c)) In all the cases, small peaks can be ignored when their relative intensities are less than 20% of the corresponding base peak

In this study, SIM of quasi-molecular ions in each tested compound at m/z 210, m/z 151 and m/z 60, were selected for quantitative analysis Fig 5-2 shows the individual SIM chromatogram for all three compounds in a mixture of 0.10 ng/µl of a solvent-based standard

(a)

+C Full MS 2 210.00 @ 30 [50.00-250]

+C APCI Full MS (m/z 50-250)

%

(c) (b)

Fig 5-2 SIM chromatogram for (a) N-methylformamide, (b) propoxur and (c)

2-isopropoxyphenol in a standard mixture (0.10 ng/µl of each analyte in methanol)

Retention Time (min)

5.3.2 Data Analysis Strategy for MAE Optimization

After implementing the eight experimental trials, which were pre-designed according to the OA8 (27) matrix, the corresponding average recovery (AR) for each experimental trial was calculated and then tabulated (Table 5-2) The average of the responses (r1 and r2) for each factor at different levels was also calculated and is shown in Table 5-2 Table 5-2 reflects the values for goldfish, tortoises, and sea lettuce

Based on methods described in previous work [22, 23], the results of the sums of squares for different variables and so-called interaction columns for each extraction species were calculated (see Table 5-3, 5-4 and 5-6) For each tested species, two-variable interactions were assigned to column 6 Furthermore, columns 3 and 5 were treated as dummies and used for estimating the error variance in this optimization Subsequently, the ANOVA tables were constructed (Table 5-3 to 5-5)

Table 5-3 Variance analysis table in the OA 8 (27) matrix for the optimization of MAE

from goldfish

Items

Extraction solvent

Extraction temperature

Extraction time

Solvent volume

SS= sum of squares; d.f.= degrees of freedom;

aThe critical F value is 18.51 (**P < 0.05) and 8.53 (*P< 0.1)

From Table 5-3, it can be seen that only factor A (extracting solvent) and factor B

(extraction temperature) are statistically significant (p < 0.05), whereas no statistical

differences are observed for any other variables and interactions considered (p < 0.1) It is

clear from the average recovery (AR) results that the combination of A1 and B2 would

result in a maximum response Since effects of duration time and volume of the solvents

(factors C and D) are minor, less solvent and time are preferred Thus, the optimum MAE

conditions for the extraction of propoxur and its two TPs from goldfish tissue can be

summarized as follows: 20-ml of methanol, an optimum microwave heating temperature

of 80 °C, and a 6-min of microwave heating

Table 5-4 Variance analysis table in the OA 8 (27) matrix for the optimization of MAE

from tortoises

Items

Extraction solvent Extraction temperature Extraction time Solvent volume

SS= sum of squares; d.f.= degrees of freedom;

aThe critical F value is 18.51 (**P < 0.05) and 8.53 (*P< 0.1)

Table 5-4 shows the results, which are used to optimize the extraction from tortoises

From the table, it is clear that factor A and factor B are statistically significant (p < 0.05),

exactly as in the extraction from fish tissue However, the small difference between the

extractions from goldfish and tortoises tissues is that the two-variable interaction B х C (interaction between extraction, temperature, and time) is statistically significant (p < 0.1) for the extraction from the latter, whereas it has a minor effect in extraction from goldfish Comparing the r1 and r2 for factors B and C (see Table 5-2), it is clear that the optimum levels are B2 and C1 On the other hand, because the interaction B х C is statically significant (p < 0.1), the choice of the optimum experimental conditions for factors B and C must depend on their interaction, which can be evaluated using a 2 х 2 table (Table 5-5) The method of construction of a 2 х 2 table is described in detail elsewhere [22]

Table 5-5 Two-by-Two table for the analysis of the B х C interaction during the

the following variables: Factor B is significant at p < 0.05, while both factors A and D are

statistically significant at p < 0.1 Thus, the optimization conditions selected for algal

extraction are 30 ml of methanol and 6 min microwave heating at 80 °C Basically

according to experience, the magnetron power of the microwave heating was at 80%

outputs (960W) The solvent temperature was able to reach the pre-set microwave

temperature within 1-2 minutes of microwave heating during the extraction An important

point to note when employing the OAD procedure is that although some parameters

(duration of extraction and volume of solvent) may not be statistically significant during

this optimization exercise, it does not mean that they are not important for the recoveries

of target pesticides during MASE The selection of levels for each parameter may be too

close or too far apart in magnitude to be statistically significant at all Thus, it should be

emphasized that adequate judgment and preexisting experience are necessary for properly

identifying appropriate factors and their levels; this is normally a case-by-case situation

Table 5-6 Variance analysis table in the OA 8 (27) matrix for the optimization of MAE

from sea lettuce

Items

Extraction solvent Extraction temperature Extraction time Solvent volume

SS= sum of squares; d.f.= degrees of freedom;

aThe critical F value is 18.51 (**P < 0.05) and 8.53 (*P< 0.1)

Table 5-7 The Recoveries of different analytes in spiked biological samples under

respective optimized MAE

Extraction Samples

Recoveries (%) ± RSD (%) n=6 Propoxur 2-isopropoxyphenol N-methyl-formamide

the 2-isopropoxyphenol was introduced by direct spiking and not by oral ingestion Furthermore, another interesting result was that although methanol was demonstrated to

be a very good extractant for MAE, a slightly better recovery of 2-isopropoxyphenol was obtained in algal samples when dichloromethane was used as the extractant

5.3.3 Stability of Propoxur under Microwave Irradiation

One of the main objectives in speciation analysis is the assurance that the targets are stable during the whole process The effects of the microwave irradiation during extraction were studied by analyzing a standard propoxur solution (1 µg/ml) An aliquot

of this solution was subjected to the microwave extraction under the optimum conditions established The experiments were performed in two replicates and each replicate was measured twice The results showed that neither evaporation losses nor propoxur decomposition were produced in a closed microwave oven Thus, the stability of the analytes tested during testing is clearly assured Evidence is presented in Fig 5-3, which shows the typical chromatograms obtained before and after microwave extraction (80°C,

8 min extraction time, and 20 ml of methanol)

5.3.4 Stability of Propoxur and Its TPs in Tested Extracts with Time

In order to determine the stability of biota extracts prior to measurement, freeze-dried biological samples considered in this work containing propoxur, N-methylformamide, and 2-isopropoxyphenol, were extracted under MAE at their respective optimal conditions Aliquots of the extracts were analyzed after 12, 24, 48, and 72h The results