Development of solid phase extractants for pesticide analysis

Abstract

Solid phase extraction (SPE) is a well-known sample preparation technique which is used widely by many ISO 17025 accredited laboratories around the world and is also the focus of many articles in the academic literature. Furthermore, with regards to the use of SPE sorbents, molecularly imprinted polymers (MIPs) have received significant research interest due to the need for selective and precise extraction sorbents. The four pesticide templates used in this study were selected specifically for their relevance in the South African context, namely two triazine pesticides: atrazine and terbuthylazine, and two chloroacetanilide class pesticides: acetochlor and alachlor. All four of the selected pesticides are highly ranked locally in terms of their weighted hazard potential based on the quantity used, their toxicity and potential environmental impacts. Another factor that played an important role in the pesticide selection process was compatibility with gas chromatography-mass spectrometry (GC-MS) analysis. Even though liquid chromatography tandem mass spectrometry (LC-MS/MS) is a powerful, more sensitive alternative, it is not readily available in many commercial laboratories due to its high cost. More analyte concentration and sample clean-up steps were thus necessary prior to sample analysis by GC-MS. In this study, emphasis was additionally placed on pesticides that have the potential to be used in the South African medicinal cannabis field. As of 2017, there has been a heightened interest in cannabis as South Africa started moving towards the legalisation thereof for personal and medicinal use. In this study, MIPs and non-imprinted polymers (NIPs) were synthesised utilising methacrylic acid (MAA) as functional monomer, ethylene glycol dimethacrylate (EGDMA) as cross-linker and 2,2′-azobis (2-methylpropionitrile) (AIBN) as a radical polymerization activator. Several MIPs were synthesised, including a multi-template MIP, which included all four target pesticide templates in one MIP. The template molecules were removed by repeated washing of the polymer with a mixture of methanol: acetic acid (9:1 v/v). Size fractionation was performed utilising wet sieving, with 25 and 53 μm stainless steel sieves and deionized water. The washed and size-fractionated polymers were subsequently packed into cartridges suitable for SPE. The materials were characterized by scanning electron microscopy (SEM) and Fourier-transform infrared spectroscopy (FTIR). The adsorption capacities were determined for each synthesised MIP, as well as the NIP. The NIP adsorption capacity ranged from 0.52 to 0.69 mg/g for atrazine and alachlor, respectively. Comparing the average adsorption capacity of the two pesticide classes on the NIP indicates that the chloroacetanilide pesticides (namely acetochlor and alachlor) have higher adsorption capacities. A correlation between log Kow values of the pesticides and adsorption capacities was observed. Furthermore, the average MIP was found to have an adsorption capacity of 0.99 mg/g, with no significant statistical differences observed between the pesticide classes. The correlation between adsorption capacity and log Kow was not observed as was the case in the NIP, indicating that more than just hydrophobic interactions were responsible for the increased adsorption capacity in the MIP. The MIP has adsorption sites or cavities left in the shape of the template molecule allowing for greater adsorption capacity when compared to the NIP. Several variations of MIPs and their corresponding NIPs were synthesised and characterized in terms of adsorption capacity and selectivity. The variations included increasing the amount of template added during synthesis, as it was hypothesised that an increase in the number of template molecules should increase the number of cavities in the subsequent MIP, thereby enhancing the adsorption capacity, and these MIPs were referred to as enhanced adsorption capacity (EAC) MIPs. A novel multi template MIP was also synthesised where both triazine and chloroacetamide pesticide class template molecules were added during synthesis. Packed molecularly imprinted SPE (MISPE) cartridges were compared to commercially available C18 SPE cartridges in terms of extraction and elution efficiency. Under ideal conditions and a relatively high concentration of all four selected pesticides in the loading fractions (0.2 μg/mL), recoveries ranged from 90 to 97% for both the MISPE and C18 sorbents. Furthermore, it was found that the pesticides eluted more easily from the C18 sorbent than from their respective MIPs, as more methanol elution solvent had to be passed through the MIP to fully elute the analytes. Thus the cavities in the MIP provide high affinity adsorption sites for the template pesticide, making elution thereof more difficult. Cannabis flower samples were spiked with the four selected pesticides, at the relevant concentration of 0.05 mg/kg, which is the South African maximum residue limit (MRL) for the selected pesticides on crops. The spiked flowers were then extracted utilising the MISPEs to good effect. With the spiked samples, the MIP outcompeted the C18 sorbent in terms of selectivity at the South African limit, as many non-polar molecules (oils and waxes found in cannabis plant material) were trapped on the C18 sorbent but can more easily pass through the selective synthesised MISPEs during the washing steps, resulting in less background interference. Pesticide recoveries from the MIP ranged between 58.5% for atrazine on an atrazine MIP with water extraction, to 85% recovery of acetochlor on the acetochlor MISPE with water extraction. It is theorised that the triazine pesticides have more sites for hydrogen bonding with MAA and EDGMA in the molecular cavities, as the NIP did not perform well for pesticide extraction from plant samples spiked at 0.05 mg/kg as no pesticides were detected in the NISPE extract. In conclusion, the synthesised MIPs were effective at extracting pesticides from spiked cannabis material. The triazine MIPs proved to be slightly more efficient than the chloroacetamide MIPs as more background can be removed during the MISPE procedure, as observed by the number of matrix interference peaks present in the resulting chromatograms. This is attributed to the wash solvent, as atrazine and terbuthylazine loaded on the triazine MIPs are amenable to higher concentration methanol solvent fractions before the pesticide analytes are washed out of the cavities. Atrazine had the best recovery at 78.6% on the atrazine MISPE and acetochlor had the best recovery on the acetochlor MISPE at 79.1%, these recoveries were not found to be statistically different from one another with a t-test. For the chloroacetamide MIPs, it was found that the analytes were removed from the cavities far more easily, requiring the use of more polar loading and washing fractions. The enhanced adsorption capacity (EAC) MIP proved to have higher adsorption capacities for both the triazine and chloroacetamide pesticide classes. For triazine EAC MIPs the average adsorption capacity was increased from 0.93 to 1.32 mg/g for the triazine pesticides. Similarly, the adsorption capacity for the chloroacetamide pesticides increased from 1.02 to 1.24 mg/g on the chloroacetamide EAC MIPs. Thus proving the hypothesis that more cavities in the MIP increases the mass of analytes that can be adsorbed. It was, however, concluded that at a relevant spiking concentration of 0.05 mg/kg the increased adsorption capacity becomes irrelevant as only a few μg of pesticide analyte would be available for extraction prior to analysis. The novel multi template (MT) MIP proved to be effective at extracting and adsorbing all the targeted analytes from an aqueous solution and spiked cannabis flowers respectively, with recoveries ranging from 76.5% for atrazine and 83.2% for acetochlor with optimized extraction methods for each analyte. In comparison with commercial C18 cartridges, the MISPEs performed better in terms of selectivity when spiked cannabis samples were analysed resulting in less background noise. However, the MIPs were found to be much more prone to channelling when the sorbent bed dried, which made it necessary to omit the drying step before the final elution of the analytes, which contributed to the increase of the elution fraction volume. The larger elution volumes do not, however, have a significant impact on the extraction recoveries, as they are dried and reconstituted with methanol containing internal standard prior to GC-MS analysis.