Development of analytical methods for N-nitrosamine disinfection byproducts in drinking water

Abstract

N-nitrosamines (NAs) are widely recognized as cancer-causing and genotoxic substances that emerge in water primarily as a result of chlorination or chloramination disinfection processes when nitrogen-containing compounds are present. N-nitrosodimethylamine (NDMA) is an NA that has garnered considerable attention due to its highly carcinogenic properties. Solid-phase extraction (SPE) is the preferred method for isolating NAs from water samples, with analysis primarily relying on gas chromatography-mass spectrometry (GC−MS) or liquid chromatography-mass spectrometry (LC−MS). Despite this, comprehensive research into the distribution and potential consequences of NAs in water remains largely unexplored on the African continent. There is a pressing need for extensive research geared towards the development of sensitive yet user-friendly analytical techniques for detecting NAs and monitoring their presence in African water sources, especially in drinking water, to better assess potential health risks for the human population.

Molecularly imprinted polymers (MIPs) are widely acknowledged for their durability, adaptability, and ability to effectively emulate antibody-antigen systems. These polymers were thus synthesized with microcavities designed with the intent to exclusively recognize and bind to NDMA for utilization in the extraction of NDMA from water and methanol samples using SPE and dispersive solid-phase extraction (dSPE) techniques. Computational simulations were undertaken to identify an appropriate template for the synthesis of MIPs with selectivity towards NDMA. The difference in binding energies between the dimethylformamide and methacrylic acid complex (DMF−MAA) and N-nitrosodimethylamine and methacrylic acid complex (NDMA−MAA) in both acetonitrile and water was found to be -1.081 kcal.mol-1 and -1.012 kcal.mol-1, respectively. This observation suggested that DMF could serve as a suitable template for the synthesis of MIPs targeting NDMA, as the binding energies of these complexes exhibited relatively close values. Additionally, it is of paramount importance that the template is complementary in shape and size to the target molecule to allow the creation of microcavities that would be able to bind to it. In this study, DMF was the template of choice for NDMA due to similarities in their structures. The choice of water and acetonitrile for the simulations was deliberate, as these solvents mirror those employed in the actual MIP synthesis process. The DMF template was efficiently removed from the synthesized polymer by heat (158 oC for one hour) after an initial washing step to remove unreacted reagents.

This approach is more environmentally friendly than typical methods that employ large solvent volumes for removing templates from MIPs. The optimal synthetic method for MIPs and non-imprinted polymers (NIPs) employed 1,1’-azobis(cyclohexanecarbonitrile) as an initiator and temperatures around 75-76 oC for 24 hr. The characterization of the resulting polymers was done using scanning electron microscopy, thermogravimetric analysis, Fourier-transform infrared spectroscopy, and the Brunauer-Emmett-Teller (BET) method. Thermal analysis showed that the backbone of all the polymers collapsed around 290 oC and that there was close to 100% thermal decomposition for all polymers around 470 oC which was comparable to temperatures that were found in the literature for similar MIPs. The FTIR spectrum of DMF had a completely different fingerprint region when compared to that of MIP−DMF. In addition to the results of the analysis of the polymer washings by GC−MS, this proved that DMF was efficiently removed from the MIP−DMF polymer. Despite the large particles that were observed for the non-imprinted polymers (NIPs) which may indicate a smaller specific surface area, the BET method showed that the NIP contained more pores and thus the additional spaces and interconnected pores in it rendered a larger overall surface area, which resulted in a higher specific surface area than that for the MIPs which had smaller particles. LC−MS and GC−MS methods were developed for NAs analysis. For the LC−MS, atmospheric pressure chemical ionization (APCI) outperformed electrospray ionization (ESI) and hence was chosen as the ionization mode of choice for further analyses. The LC−MS method was used for the analysis of eluates from the SPE and dSPE experiments. The GC−MS method was used in the stability analysis of NAs, for the analysis of anhydrous extracts from the SPE and dSPE experiments, and to provide further evidence that the DMF removal from the MIPs was successful. Despite the extensive precautions taken to preserve the nitrosamine mix standard in securely sealed amber bottles and vials stored under refrigeration, along with conducting experiments under low-light conditions, a stability assessment revealed that NAs exhibit significant instability, particularly those with lower molecular masses, such as NDMA and N-nitrosomethylethylamine (NMEA). These compounds were observed to degrade progressively over time, with eight of the nine NAs in a nitrosamine mix standard displaying a decline in response (in terms of peak areas) of 60% or more by the fifth day of the stability analysis. Consequently, the detection of NAs at trace levels becomes even more challenging due to their inherent instability.

The optimal conditions for the SPE of NAs by LC−MS were found to be: one hour sorbent drying time, 10 mL elution volume, 1 M ionic strength, and a basic pH of 11.60 using HC-C18 cartridges. 500 mg MIP and NIP cartridges were found to block during the sample loading step when SPE experiments were conducted, but this challenge was overcome by reducing the sorbent mass to 90 mg. Upon loading the MIP−DMF cartridges with 500 mL water samples spiked with 1 mL of 10 µg/mL of NDMA to provide a final concentration of 0.02 µg/mL, it was found that the sorbents were either unable to retain NDMA or interacted very strongly with it, making it difficult to elute NDMA by dichloromethane (DCM), as NDMA was not detected in the eluates by GC−MS. To provide further evidence of whether the synthesized MIPs could retain NDMA and whether they are selective, dSPE experiments were attempted as they are relatively easy and can be completed in a short time as opposed to the cartridge based SPE method, which makes dSPE potentially attractive for commercial laboratory applications. 5 mg of MIPs and NIP were determined to be a suitable sorbent mass for the dSPE procedure, but 20 mg was used for the HC C18 as it was difficult to weigh 5 mg of it accurately due to its high density. The amount of NDMA that was detected by the GC−ToFMS in the supernatant of the MIP−NDMA was 1.4 greater than that detected in the HC-C18 supernatant as determined by the ratio of the NDMA peak areas. The amount of NDMA detected in the supernatants of the MIP−DMF and NIP was 9.2 and 5.7 greater, respectively, than that detected in the HC-C18 supernatant. The MIP−NDMA performed more similarly to HC-C18, whilst the MIP−DMF and the NIP did not perform as well. There was no NDMA detected in the extracts from either the loaded synthesized polymers or the HC-C18. It is possible that the extraction solvents employed (DCM and methanol) may not have been sufficiently strong to release any extracted NDMA, making it present at concentrations that were not detectable by the GC−ToFMS. Alternatively, NDMA might have been strongly bound to the materials, hindering its release, or the bound NDMA underwent degradation during the extraction procedure. Further dSPE experiments were conducted using a nitrosamine mix standard as a spiking solution to determine the selectivity of the MIPs. Percentages of how much was retained and eluted, degraded or irreversibly retained, or not retained by the sorbents for each N-nitrosamine were determined after analysis by UPLC−QToFMS. Of the three NAs that were retained and eluted from the sorbents, MIP−DMF extracted 14% of the NDBA whilst MIP−NDMA and HC C18 extracted 10% and 20% of the NDBA, respectively. The MIP−DMF extracted more of the N-nitrosopiperidine (NPip) (8%), followed by the MIP−NDMA (7%) and the HC-C18 (3%). The MIP−DMF also extracted more of the N-nitrosodiphenylamine (NDPhA) (15%), followed by the MIP−NDMA (14%) and the HC-C18 (12%). The remaining analytes or portions of analytes that were not retained and eluted were either irreversibly retained by the sorbent material or degraded, whilst some portions were not retained at all. Of the two NAs extracted by the NIP, 7% and 6% were extracted for N-nitrosopyrrolidine (NPyr) and NPip, respectively. Most of the NAs were either irreversibly retained by the sorbent material or were degraded. The dSPE method utilized did not prove effective in fully assessing the potential sorption of NDMA by the MIPs synthesized in this study. In comparison to the cartridge SPE experiments, the dSPE results for the HC-C18 were notably poorer (although it is acknowledged that a lower mass of sorbent was used). This suggests that the dSPE method employed may not be well-suited for extracting the nine target NAs in this study. Nonetheless, it is worth noting that the dSPE method offers convenience and rapidity in contrast to the cartridge based SPE method. Overall, this study contributes to the understanding of N-nitrosamine extraction and detection, specifically NDMA, from water matrices. It highlights the development of MIPs as a potential solution for improved extraction and the challenges associated with the stability and selective sorption of N-nitrosamines, providing valuable insights into the field of water quality assessment and the need for more effective analytical techniques.