Analytical approaches for haloacetic acid disinfection byproducts in treated water

Abstract

Water quality is vitally important for basic human health and key in growing economies and in overcoming poverty. Disinfection of source water is therefore indispensable, however, it may introduce new adverse health risks to consumers due to the formation of disinfection byproducts (DBPs). Many of the over 700 DBPs discovered to date are cytotoxic, genotoxic, potentially carcinogenic, or pose other health risks. Minimizing exposure of the public to these compounds is therefore crucial. The focus of this study was the investigation and development of analytical approaches for haloacetic acid (HAA) DBPs, which over the past two decades have been increasingly included in drinking water guidelines globally. The United States Environmental Protection Agency (US EPA) and the European Union (EU), for example, limit the maximum concentration to 60 μg/L in drinking water for a combination of five HAAs (HAA5: monochloroacetic acid (MCAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA), monobromoacetic acid (MBAA), and dibromoacetic acid (DBAA)), whilst the World Health Organization (WHO) stipulates maximum concentrations of 20, 50, and 200 μg/L for MCAA, DCAA and TCAA, respectively. Although the current South African National Standard for drinking water quality (SANS 241), does not include HAAs, a draft which proposes their inclusion was released in 2022 for public comment. Implementation of these regulations calls for accurate and especially sensitive methods for HAAs, as they are expected to be present at trace (low μg/L) levels in water. Despite two standard methods for HAAs published by the US EPA, extensive research is being done to develop improved methods, specifically addressing the issues of time and labour costs, whilst considering the environmental impact which has grown in importance since these methods emerged. However, much of this research disregards the needs of laboratories in developing countries. This study thus focused on improving HAA analysis methods in a developing country context, with consideration of the environmental footprint thereof. Since gas chromatography (GC)-based analysis, and therefore the appropriate instruments, are more commonly available than high resolution and comprehensive chromatography-based methods, the US EPA Method 552.3 (from hereon called EPA Method), which utilizes GC-electron capture detection (ECD) analysis, was considered as a basis for this research. Due to the acidic nature of the analytes and the water matrix, an extraction and derivatization step are required prior to analysis, making the method time and labour intensive. Each of these steps was considered and alternative approaches to simplify them were explored, starting with the analysis, followed by the derivatization, and finally the extraction. In terms of analysis, GC-mass spectrometry (MS) was contrasted with GC-ECD, and the feasibility of liquid chromatography (LC)-MS analysis was investigated, as this would eliminate the need for extraction and derivatization. Although both GC-based instruments had their respective advantages, the sensitivity achieved with the GC-ECD remained superior (limit of detection: 0.34 – 5.45 μg/L for GC-ECD versus 5.00 – 30.0 μg/L for GC-MS). Analyte concentrations in Pretoria tap water samples were below these detection limits using the EPA Method with GC-ECD analysis. Sensitivities achieved with LC-MS analysis (Waters® Synapt ultra performance LC – high definition MS (UPLC-HDMS) quadrupole/time of flight (QToF) mass analyser and Waters® Xevo-G2 LC-QToF mass analyser), were even poorer than with the GC-based instruments. Nonetheless, seven analytes could be tentatively identified based on mass spectra in spiked real water samples. Furthermore, experiments revealed that filtration of real water samples did not negatively influence analyte peak areas, whilst addition of formic acid (FA) had no benefit. Furthermore, two analytes (DCAA and TCAA) could be detected with the Waters® Xevo-G2 LC-QToFMS in swimming pool water obtained from LC de Villiers sports campus. Although two different solvent elution methods were explored, further method development could potentially improve sensitivity. The Fischer esterification of HAAs to the methyl esters required by the EPA Method is the most time-consuming step, thus an alternative derivatization to the octyl esters was explored and optimized. Reaction time and temperature could be drastically reduced from 2 hours to 30 min, and 50 °C to 30 °C, which consequently saved energy, making the octyl derivatization a greener option than the Fischer esterification. However, application of this derivatization in an adaptation of the EPA Method with GC-ECD analysis, demonstrated that the liquid-liquid extraction (LLE) into the hexane solvent of the octyl derivatives was unsatisfactory. Therefore, despite successful improvement of the derivatization step, integration thereof with the other steps in the complete HAA analysis protocol and application thereof to spiked deionized water and treated swimming pool water proved difficult. Further exploration utilising other solvents which are appropriate for the octyl derivatization, and which may improve the extraction efficiency of HAAs from water is necessary. As an alternative to the labour- and time-intensive LLE, thin film extraction was considered. This extraction is fast, simple, and environmentally conscious as sample handling is reduced and solvent volumes minimized. Polydimethylsiloxane (PDMS) thin films were prepared in-house with a novel spin-coating technique which allowed for effortless addition of sorbents to the PDMS in different variations, as well as geometric alterations. Two sorbents, hydrophilic-lipophilic balance (HLB) and Carboxen® particles, were either added to the PDMS prior to film synthesis or during the curing process, followed by characterization using microscopy. PDMS, HLB and Carboxen® were thus sorbents of interest, however, due to the proprietary nature of Carboxen® only HLB and PDMS could be interfaced with different analyte variations in a computational model. This model was verified by comparing the binding energies to experimentally obtained extraction efficiencies of the various film types. These were determined by quantifying the residual analyte concentrations in water after extraction with the films by immersion in HAA spiked deionized water for 24 hours. Further information on the interactions occurring at a molecular level could be obtained from the model, as well as insights into which ester derivatives of the analytes would be best extracted by each respective sorbent. Both, the experimental and computational data predicted good extraction potentials with HLB particles. From the experimental data it was determined that Carboxen® showed great extraction efficiencies for the analytes of a lower molecular mass (95 – 218 g/mol), whereas HLB extracted those of a larger molecular mass (163 – 297 g/mol) better. Moreover, films with the sorbents added to the surface during the curing process achieved the highest extraction efficiencies, with HLB reaching efficiencies up to 93% and Carboxen® up to 58%. The application of thin film extractions in HAA methods needs to be explored further but could be beneficial in a variety of applications: firstly, in a pre-concentration step for LC-based analysis to increase the sensitivity, secondly, in a GC-based method with thermal desorption and in-line gaseous derivatization, or thirdly, simply as the analyte extraction step, followed by solvent back-extraction, derivatization and analysis. The research presented in this dissertation thus critically evaluates existing methods for the analysis of HAAs in treated waters and explores a number of alternatives to make these methods faster, more environmentally friendly, and cost effective, whilst not compromising on sensitivity and accuracy. In this manner, this research contributes to the evolution of new analytical approaches for HAA water disinfection byproducts which are of global importance.