Determination of Oxyanions in Drinking Water by Direct Injection Using Liquid Chromatography-Tandem Quadrupole Mass Spectrometry

The purpose of this work is to demonstrate a direct injection method for the determination of chlorate, perchlorate, bromate and chlorite in drinking water that exceeds the requirements for the new EU Drinking Water Directive 2020. The method performance study was completed on an ACQUITY UPLC I-Class PLUS System with a Xevo TQ-XS using the Anionic Polar Pesticide Column. Method performance was assessed using 3 spike levels (20, 50, and 100 µg/L for chlorate and perchlorate, 2, 5, and 10 µg/L for bromate and 100, 250, and 500 µg/L for chlorite) with 22 replicates at each level split over 2 batches for each water type studied. The water types studied were mineral water, soft and hard drinking water. Residues of chlorate were detected in soft (54.3 µg/L) and hard (12.0 µg/L) water and quantified by standard addition, which are significantly lower than the parametric values listed by the EU Drinking Water Directive. Method performance for average trueness values across all water types and spike levels for perchlorate 96.2–100.3%, chlorate 98.1–100.5% (mineral water only), bromate 92.2–102.1% and chlorite 84.1–114.1%. RSDs were significantly below 10% except for chlorite where higher RSDs were observed for some water types. Retention time repeatability was recorded through all method performance batches and an additional repeatability study, where RSDs for all analytes across over 550 injections were less than 1%.

Benefits

• Robust and reproducible analytical method allowing for a simpler laboratory workflow with method performance that exceeds regulatory requirements
• Consistent method performance across all 6 method evaluation batches in the 3 different water types (mineral, soft, and hard drinking water)
• Simple sample handling before determination stage

Bromate, chlorite, and chlorate are disinfection byproducts formed during the water treatment process. Chlorate and chlorite are formed during the disinfection process when using hyperchlorite, whereas bromate is formed as a byproduct from water ozonation. Perchlorate can be found in water either through natural or man-made sources such as run off from land, where natural fertilizers have been used. Perchlorate can also be formed during the degradation of sodium hypochlorite, during water disinfection, as well as being a byproduct of various industrial processes.¹ Perchlorate, chlorate, and chlorite are known to be hazardous to human health while bromate is classed as a possible human carcinogen.¹ Monitoring of these compounds in water is essential for public health.

A fast, efficient method for the determination of oxyanions in drinking water is important as we recognize and manage risks associated with chemical contamination in drinking water supplies. The World Health Organization (WHO) updated drinking water guidelines in 2017² and the EU are to set new standards in chemical contamination in an updated EU Drinking Water Directive.³ The new Directive adds chlorite (with a new parametric value of 250 µg/L), whilst maintaining a requirement to monitor for chlorate (with an unchanged parametric value of 250 µg/L) and bromate (with an unchanged parametric value of 10 µg/L). The WHO has an additional guideline for perchlorate (with a level of 70 µg/L) and has a higher concentration of 700 µg/L for chlorate and chlorite. The US EPA includes bromate (MCL 0.01 mg/L) and chlorite (MCL 1 mg/L) in their National Primary Drinking Water Regulations.⁴

Common methods for the analysis of oxyanions require separation with ion chromatography using suppressed conductivity detection. Often to achieve the required detection limits, in some cases for bromate, post column derivatization (PCD)⁵ is required, using UV detection. We have developed a fast, direct injection method with low injection volume utilizing the Anionic Polar Pesticide Column. This approach allows for rapid analysis of the 4 oxyanions within 5 minutes on a tandem quadrupole mass spectrometer which is becoming a popular tool for the analysis of these compounds. The benefit of this technology is that it gives reliable quantification of these analytes using multiple reaction monitoring (MRM), detecting all 4 oxyanions in drinking water without the need for PCD or specialized chromatography equipment.

Sample Description

Several 1 liter samples of tap water from a soft water (Wilmslow, UK) and hard water (York, UK) area were collected and transported to the Applications laboratory (Wilmslow, UK). Several 1 liter samples of a generic mineral water were purchased in the UK for use as matrix blanks.

Method Conditions

Once the drinking water samples were received in the laboratory, they were split into aliquots of 50mL in plastic falcon tubes. The samples were then stored at 4 °C in a fridge until analysis.

Just before analysis, the sample was transferred to an LC vial, isotopically labelled internal standards were added (Cl¹⁸O₃- and Cl¹⁸O₂- at 100 µg/L in sample and Br¹⁸O₃- at 10 µg/L in sample) for the determination of the oxyanions. Only chlorite was not internally calibrated as no suitable internal standard was identified for this analyte.

After the method development phase, a method performance study was conducted for mineral water, soft water, and hard water which consisted of 2 batches for each water type containing 33 spiked water samples: 11 spikes at a designated low, mid and high concentration; 2, 5, and 10 µg/L for bromate, 20, 50, and 100 µg/L for chlorate and perchlorate and 100, 250, and 500 µg/L for chlorite. The MRMs listed in Table 1 highlight the optimized transitions used for quantification and confirmation of the oxyanions in this application.

The injection diluent was investigated during the method development phase of this project. We investigated the requirement to add acetonitrile to ensure acceptable peak shape for a HILIC Column. The results shown in Figure 1 confirm that there is no significant loss of peak integrity when injecting undiluted water compared to the standard that was composed of 50% acetonitrile. Injection volume was investigated to allow an increase in sensitivity if required. The results shown in Figure 2 confirm that there is no significant loss of peak integrity when injecting 1 µL compared to 5 µL of an undiluted water sample. For these initial investigations it was concluded that 5 µL of an undiluted water could be injected onto the Anionic Polar Pesticide Column without significant loss of peak integrity and this format was used for the method performance study to improve method limits of quantification.

The drinking and mineral water samples collected were screened for the presence of oxyanions before the method performance study was started. No oxyanions were detected in mineral water; however, chlorate was detected in both the hard and soft water at levels significantly lower than the legal limits, but at concentrations that would interfere with the method performance study. The residue concentrations of chlorate were calculated by standard addition (n=6) and the results are included in the method performance data displayed in Table 2 and were significantly lower than the parametric limit.

Method performance was assessed with 2 analytical batches per water type, where each batch was comprised of 33 spikes, with 11 at each designated low, mid, and high levels of each of the oxyanions. The results for chlorate were excluded for both the soft and hard water types due to the presence of chlorate residues in the samples. Mineral water was used to assess method performance for chlorate. Matrix matched calibration standards were used for all method performance study batches. Calibration solutions were prepared over the range 10–200 µg/L for perchlorate and chlorate, 1–20 µg/L for bromate and 100–1000 µg/L for chlorite. Figure 3 demonstrates expected calibration graphs from this approach. Coefficients of determination were 0.995 or greater, for chlorate, perchlorate, and bromate and 0.99 or better for chlorite. Residuals for perchlorate and chlorate were all under 10%, bromate were all under 12%, with chlorite 10% or under.  

Chromatographic performance of the method was consistent throughout the method performance study. Figure 4 gives an example of the chromatography that was achieved for the lowest calibration standard from soft, hard, and mineral water calibration series. Across the three different types of water investigated there was no significant change in peak shape. Results from quantification of the spikes of the 6 method performance batches are displayed in Table 2. A consistent method performance for perchlorate and bromate was noted across all 6 batches and for chlorate in the mineral water batch. More variation for chlorite results was observed in hard water.

Retention time stability is an important factor to investigate during a method performance study. To this end a repeatability study of 300 injections of an alternating soft, hard, and mineral water standards was run. All retention times across the 6 method performance batches and the retention time stability study were plotted and displayed in Figure 5. There is no significant change in retention time noted across the whole method performance study and repeatability study with all 4 oxyanions having RSDs under 1% across the whole study.

The method performance study data demonstrates a robust analytical method for the determination of oxyanions by direct injection of water samples onto an Anionic Polar Pesticide Column fitted in an ACQUITY UPLC I-Class PLUS System coupled to a Xevo TQ-XS. This application exceeds the new requirements in the EU Drinking Water Directive (EU DWD) coming into effect in late 2020³, current WHO drinking water guidelines², and US EPA National Primary Drinking Water Regulations.⁴ Limits of quantification for chlorate and perchlorate are 20 µg/L, for bromate 2 µg/L, and for chlorite 100 µg/L. For chlorate, this is 12.5 times lower than the proposed new EU DWD limit, for bromate 5 times lower and, for chlorite 2.5 times lower. Perchlorate is 2.5 times lower than the WHO recommended level. Retention time stability was proven over the course of the study with RSDs for all 4 oxyanions under 1% across all 6 method performance batches and retention time stability study. This short method time of 5 minutes with direct injection of samples allows for a high sample number throughput.

Scientists must validate the method in their own laboratories and demonstrate that the performance is fit for purpose and meets the needs of the relevant analytical control assurance system.

1. Constantinou. P, Louca-Christodoulou D, Agapiou A. LC-ESI-MS/MS Determination of Oxyhalides (Chlorate, Perchlorate, and Bromate) in Food and Water Samples, and Chlorate on Household Water Treatment Devices along with Perchlorate in Plants. Chemosphere 2019, 235: 757–766.
2. WHO Guidelines for Drinking-Water Quality Incorporating the 1st Addendum, 2017 (accessed September 2020) https://www.who.int/water_sanitation_health/water-quality/guidelines/en/.
3. Directive of the European Parliament and of the Council on the Quality of Water Intended for Human Consumption, Inter-institutional File 2017/03329(COD), 5 Feb 2020, 5813/20.
4. US EPA National Drinking Primary Drinking Water Regulations, (accessed October 2020) https://www.epa.gov/ground-water-and-drinking-water/national-primary-drinking-water-regulations#six.
5. Cordeior F, Schmitz F, Verbist I, Emteborg H, Charoud-Got J, Lpoex MC, Robouch P, Taylor P, de la Calle MB, JRC Scientific and Technical Reports, IMEP-25a: Determination of Bromate in Drinking Water.