Primary and secondary amine material based on crosslinked polystyrene: synthesis and initial application for multiresidue pesticides analysis

Weak anion exchange sorbent based on cross-linked polystyrene with primary secondary amine group was prepared by substitution nucleophilic reaction (SN2) between methylene chloride group and 1,2-ethylene diamine. The effect of factors, namely the weight ratio of amine over methylene chloride, reaction time and temperature on nitrogen percentage were studied using experimental design approach. The amination yield rose as all of factors increased but was reduced while both temperature and time increased simultaneously. Nitrogen percentage of the products were varied from 4.0% to 6.3%. Sorbents with predicted capacity of 4.5%, 5.0%, 6.3%, and 6.5% were synthesized. The results showed that the actual capacities of the products were close to the predictions, especially for those in the experimental domain, indicating a good model that can be used to prepare sorbents of any desired capacity. The sorbent application ability of multiresidue pesticides analysis in food were initially investigated through both aspects: interference elimination and analyte content conservation. Keywords—Anion exchange, crosslinked polystyrene, experimental design, multiresidue pesticides analysis, primary and secondary amine, QuEChERS


INTRODUCTION
rosslinked polystyrene and its modified materials are popular materials which were applied as sorbent of various analytes [1][2][3][4] due to their advantageous properties, namely, high surface area, chemical resistance, rigid structure. Crosslinked polystyrene with full of phenylene rings in its structure offers p-p interaction to aromatic analytes [5]. Nevertheless, surface modification by polar or ionic functional groups, e.g. sulfonated -SO 3 -, would support polar-polar and electrostatic interaction [6].
QuEChERS (stand for Quick, Easy, Cheap, Effective, Rugged, and Safe) -introduced by Anastassiades et al. [7] was developed as a sample preparation method for multiresiduepesticide determination in fruits and vegetables. The method includes three main steps (i) the extraction of pesticides with acetonitrile (ii) partition the analytes into acetonitrile phase by adding salts and (iii) a dispersive solid phase extraction for clean-up. This method and several modified versions have been applied for different types of matrices and pesticides [8][9][10][11]. In the third step of QuEChERS, adsorbents, such as C18, primary secondary amine (PSA), graphitized carbon black (GCB), play an important role in interference elimination process. Silica is generally used as support for C18 and PSA sorbent thanks to its availability and hydrophilic surface. The aim of this work was to synthesize and test whether the PSA sorbent based on cross-linked polystyrene can be used in QuEChERS method.

Chemicals, apparatus and software
1-Dodecanol, toluene, 1,2-ethylene diamine, sodium hydroxide, sodium carbonate, formic acid, hydrochloric acid, nitric acid, sulfuric acid, boric acid, ammonium acetate, styrene, divinylbenzene, and vinylbenzyl chloride were of synthesized grade and purchased from Merck (Germany). Crosslinked polystyrene materials which contain various levels of vinylbenzyl chloride (VBC) were synthesized based on the procedure of our previous work [1]. 2MA is the name assigned for the cross-linked polymer prepared from monomers (divinylbenzene and vinylbenzyl chloride) and then modified in order to have primary secondary amine groups on the surface. 2MA was synthesized via two steps, (1) preparation of ethylene chloride cross-linked polystyrene, and (2) amination of the polymer. The synthesis procedure of crosslinked polystyrene with methylene chloride group was conducted as in a previous work. Briefly, the monomers (14g VBC, 26g DVB), porogen (19g toluene, 41g dodecanol) and benzoyl peroxide (3g) were mixed. The polymerization was performed at 80 o C for 24h. The un-polymerized components were removed by Shoxlet extraction with methanol for 20h and dried at 60 o C for 6h.
The polymer was then wetted with toluene, to which 1,2-ethylene diamine was added for the amination. The products were washed three times with 30mL of 2% hydrochloric acid in acetone and drying at 60 o C overnight. To study effect of reaction conditions capacity of the sorbents, the mole ratio of amine to methylene chloride was varied from 10 to 70 times, temperature from 30 to 80 o C, reaction time from 8 to 24h while mass of polymer (1g) and toluene volume ((20-V amine ) mL) were fixed.

Chloride and amine content analysis
The chloride contents of pre-and postamination materials were determined by the procedure described in our previous publication [12]. Briefly, samples were treated by alkaline fusion method with mixture of Na2CO3 and NaOH. Then, their aqueous solutions were analyzed by indirect spectrophotometry of the chloride based on the adsorption at 460 nm of Fe(SCN) 2+ , a product of the reaction between chloride ion and a mixture of mercury (II) thiocyanate and ferric ion.
Additionally, %N was determined by Kjeldahl method. Sample (0.200 g) was digested with a mixture of 0.5 g CuSO 4 , 5.0 g Na 2 SO 4 and 10 mL H 2 SO 4 (conc.). The solution was then alkalized with 60mL 7M NaOH. The ammonia gas was absorbed into a solution containing an excess of H3BO3. The nitrogen content is then determined by titration of the NH 4 HBO 3 formed with standardized HCl solution using Tashiro as indicator.

Design of experiment (DOE) for amination reaction
Two-level full factorial design (denoted as 2 3 ) was chosen for the design of experiment (DOE) in this study. Reaction time (Time), temperature (Temp), and the mole ratio of amine to methylene chloride (Ratio) were factors; and %N was the response. The reaction conditions of 11 experiments were tabulated in Table 1. Experiments (N1-N8) were at high and low levels of each factor. Three replicated experiments at the center values (N9-N11) were used to evaluate the reproducibility of synthesis and model. The run order of the experiments was randomized by software.

Investigation of interference elimination
Food, namely cucumber, lemon, cabbage, green bean, garlic, onion, strawberry, green tea, tomato, and apple, consisting of chlorophyll, organic acid, sugar, dye, and essential oil as interference was extracted by QuEChERS [13]. 10g of grinded sample was extracted with 10mL of ACN, 4g MgSO4, and 1g NaCl for 1 min. 1mL of the decant was mixed with 25 mg 2MA and 150mg MgSO4. Resulted solution was analyzed by HPLC UV at 210nm, gradient eluent (mixture of acetonitrile: ammonium formate) from 50:50 to 95:5 (v/v) for 5 min, then keep in 5 min before returned to the initial condition.

Investigation of pesticide content conservation
10 µg of each pesticide was added to 100 g of grinded samples and kept at room temperature overnight. QuEChERS sample preparation was carried out same as procedure of interference elimination investigation. However, resulted solution was analyzed by HPLC MS with instrumental parameters shown in Table 2.

Regression model for the amination of crosslinked polystyrene
Based on the experimental data the regression model was built for the amination procedure (Eq. 1). It should be noted that the regression coefficients are scaled and centered. This means that they are not expressed in original measurement scales of the factors, but in the coded -1/+1 unit corresponding to the lowest and the highest values. Y = 5.932 + 0.394x1 + 0.287x2 + 0.542x3 -0.178x1x2 (Eq. 1) Where Y, x1, x2 and x3 denoted Capacity, Ratio, Temp and Time, respectively. After refining the model i.e. removing coefficients, which had uncertainty covering zero value, the resulting model having large goodness of fit factor (R2 = 0.974) and prediction power (Q2 = 0.775). Regression coefficients and factors evaluating the quality of the model are presented in Fig. 1. It was found that all of factors, including temperature, time, and amine to methylene chloride ratio give an increase in capacity. The results also revealed that the three main factors were not independent. In fact, there were significant interaction coefficients, which only can be investigated using the DOE approach. It is obvious that increases in amine level (Ratio) led to high yields of the reaction because of the higher chance of amine reagent and methylene chloride site get into contact. The same effect of temperature could be explained by the reduced viscosity of the reaction medium which promoted the contact between the reagent and surface of the porous material. Moreover, the longer reaction time, the more effective reaction sites between methylene chloride and amine reagent. However, the coefficient "Temp*Time" had the most profound negative effects which showed in response surface plot illustrated a quadratic regression between Temp and Time factors (Fig. 2A). The reason of their negative effect can be the destruction of peripheral reacted layers of 2MA sorbent to submicron scale particle which eliminated in post-synthesis treatment process while reaction was carried out at high temperature for a long time. The model can be visualized by means of response surface plots. The curvature in plots involving the factor "Temp*Time" confirmed its negative effect on the capacity when the reaction time and temperature were further increased simultaneously (Fig. 2B, C). At the bottom, higher ratio of amine to methylene chloride and longer reaction time gave high nitrogen content resulted sorbent.

Preparation of primary secondary amine sorbents with desired nitrogen percentage
To examine a model applicability, the model of amination process was used to design suitable conditions to prepare 2MA with nitrogen percentage from 4.5% to 6.5%. It should be noted that there were several reaction conditions for a desired nitrogen percentage. Considering the fact that nitrogen content would be fallen down while both temperature and time increased simultaneously, the conditions were selected with low temperature to facilitate the procedure (Table  3). It was found that, the predicted and actual values were well agreed, indicating a very good model.

Initial application in simultaneous multiresidue pesticide analysis
The extracts of ten kinds of food which were treated by 2MA were analyzed by HPLC-UV and HPLC-MS/MS to examine the interference elimination of the home-made sorbents.
The results illustrated that extracts without sorbent treated would content many UVresponsive compounds which are interferences in pesticide analysis in food. After sample preparation procedure with adsorbents (2MA or commercial PSA, there are the losses of peaks of chromatograms (Fig. 3A). However, the interference elimination ability depended on the sample nature. In case of simple matrices, like apple, tomato, green bean, onion, and cabbage, both of 2MA and commercial PSA offered effective elimination. With complex sample matrices, such as lemon, garlic, strawberry, and green tea, both of adsorbents could not remove their interferences. Moreover, the total ion chromatograms (Fig. 3B) showed that most of polar compounds which eluted before 40 mins had been removed by 2MA in simple matrix samples, the later peaks were washed out of the reversed phase column by neat acetonitrile, while in case of other complex matrix ones, early 40 mins peaks still appeared. The reason could be the high content of essential oils, polyphenols, organic acids and other polar compounds in garlic, green tea or lemon which cannot be eliminated completely by 2MA. It could be overcome by the combination of 2MA and other sorbents (C18, GCB) in QuEChERS. Besides interference elimination, analyte conservation is one of the most important requirement of adsorbent. Recoveries of 13 pesticides (retention time from 10 mins to 37 mins) in 10 matrices which were prepared by 2MA or commercial PSA as sorbent in QuEChERS were shown in Table 4. Data compatibility was checked by Student's t-test which provided p value equals 0.216, higher than 0.05 (with 95% significance level). It means the null hypothesis is accepted, there is no significant difference between recoveries of 13 pesticides in 10 kinds of food samples which were prepared by both home-made 2MA and commercial PSA.