Introduction

Per- and polyfluoroalkyl substances (PFAS) are emerging contaminants linked to the manufacturing of many products, such as semiconductors, coatings for food packaging and textiles, and aqueous firefighting foams (Prevedouros et al. 2006; Renner 2001; Domingo et al. 2017; Martin et al. 2019; Langberg et al. 2021; Glüge et al. 2020; Wu et al. 2021). Owing to the strong carbon and fluorine bonds in their structures as well as their hydrophilic and lipophilic characteristics, these substances are extremely persistent, thermally and chemically durable, and favour bioaccumulation (Buck et al. 2011; Niu et al. 2019; Foguth et al. 2020; Stoiber et al. 2020). Among PFAS, perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) have been extensively investigated worldwide in the environment, animal tissues, and human blood (Giesy and Kannan 2001; Kannan et al. 2004; Jantzen et al. 2016; Olsen et al. 2017).

Consequently, public health and environmental concerns regarding PFAS have increased around the world, and thus, the large-scale production and utilisation of PFAS have been curtailed. In Thailand, although the amounts of PFAS used and imported are currently unknown, investigations on their occurrence in different media have been ongoing since 2007. PFAS have been reported in Thailand in numerous environments, consumer products and other materials, such as river water, wastewater, tap water, bottled water, air, cosmetics, food packaging, and textiles (Boontanon et al. 2012; Keawmanee et al. 2015; Kunacheva 2009a; Kunacheva et al. 2009b; Kunacheva et al. 2010; Pattanasuttichonlakul et al. 2014; Poothong et al. 2012; Shivakoti et al. 2010; Supreeyasunthorn et al. 2016; Vo et al. 2020; Shigei et al. 2020).

Groundwater is a precious freshwater resource that is increasingly being exploited in Thailand, especially in rural areas, where surface water is limited and often polluted. Although groundwater is purified as it flows through soil and deep-rock layers, its quality can be severely affected by pollution from numerous sources. Furthermore, in Thailand, substantial amounts of municipal and industrial refuse are improperly disposed of due to ineffective management and low monitoring budgets. Leachate derived from landfills has been reported in many studies as a potential source of groundwater contamination. Leachates commonly contain compounds that negatively impact the environment and human health (Eggen et al. 2010; Brusseau et al. 2020; Menger et al. 2020). PFAS levels in landfill leachates reported in studies from different countries vary from the range of nanograms to micrograms per L of leachate (Benskin et al. 2012; Busch et al. 2010; Lang et al. 2017; Wei et al. 2019; Solo-Gabriele et al. 2020; Wang et al. 2020; Cui et al. 2020; Yong et al. 2021). However, there have been no studies involving PFAS levels in groundwater in Thailand. Therefore, the present study aimed to determine PFAS levels in groundwater in Thailand, identify potential sources of these compounds, and highlight their spatial distribution. The findings of the present study improve our understanding of groundwater contamination in Thailand associated with PFAS, which can be exploited for further studies and for the implementation of environmental standards and regulations.

The groundwater sampling sites were selected from three cities in Thailand; these are shown in Fig. 1. Groundwater samples were collected in February 2017 from domestic wells near the Bang Sai and Sena municipal waste disposal sites (MWDS) in Ayutthaya (n = 12), near the Nong Nae industrial waste disposal sites (IWDS) in Chachoengsao (n = 4), which are associated with continuous operation and near the Map Pai IWDS in Chonburi (n = 15), which is ceased operations. Analyses were conducted in the Faculty of Engineering of Mahidol University in Thailand.

Fig. 1
figure 1

Map showing points from which groundwater samples were collected, including near the 1a Bang Sai MWDS, 2a Sena MWDS, b Nong Nae IWDS, and c Map Pai IWDS

Full size image

Materials and methods

Standards and reagents

Seven PFAS standards were investigated in the present study, including perfluoroheptanoic acid (PFHpA), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnA), perfluorohexane sulfonate (PFHxS), and perfluorooctane sulfonate (PFOS). High purity solvents, including methanol HPLC grade (> 99.99%), methanol ACS grade, acetonitrile HPLC grade (> 99.8%), and ammonium acetate (98%), were purchased from Merck KGaA (Millipore, Germany). Standard solutions were prepared using the HPLC grade methanol, while Milli-Q water produced using a RiOs-DI® water purification system (Millipore, Germany) was employed in all analyses.

Sample collection

The 1.5 L polyethylene terephthalate bottles used were rinsed with methanol and dried before collecting samples. Faucets connected to pumps in the groundwater wells were used to fill the bottles after rinsing them thrice using the well water. Subsequently, the samples were stored in a cooler and transported to the Water Quality Engineering Laboratory at Mahidol University. In the laboratory, the samples were filtered using a glass fibre filter (GF/B). Teflon and glass equipment were avoided during the analyses because these could add to or cause adherence of some target compounds (Hansen et al. 2002; Yamashita et al. 2004).

Solid-phase extraction

The PrecepC-Agri (C18) cartridges were pre-conditioned with 10 mL of methanol, after which they were rinsed twice with 10 mL of Milli-Q water before the samples were loaded. The 1.5 L water samples were filtered through 1 μm GF/B glass fibre filters to remove the suspended solids. The filtrates were then loaded onto cartridges using a concentrator at a flow rate of 10 mL min−1. After loading, the target analytes were eluted using 4 mL of methanol. The extracts were purged with nitrogen gas and reconstituted using 0.5 mL of 30% HPLC grade acetonitrile (Kunacheva 2009a).

Instrumental analysis and quality control

Data for parameters associated with the high-performance liquid chromatography-tandem mass spectrometry (HPLC–MS/MS) analysis are presented in Table 1. The target PFAS in the samples were analysed using an Agilent 1200SL HPLC coupled with an Agilent 6400 MS/MS. Approximately 10 μL of each extract was injected into a 2.1 × 100 mm (5 μm) Agilent Eclipse XDB-C18 column. The mass spectrometer was operated in the negative electrospray ionisation (ESI) mode. The mobile phases involved the following: (A) 10 mM ammonium acetate in ultrapure water and (B) 100% acetonitrile (HPLC grade). The initial mobile phase contained 30% (B), and this was ramped up to 60% (B) at 16.5 min and maintained for 3.5 min. At 23 min, (B) was increased to 70% and then raised to 90% at 26 min. Subsequently, the mobile phase gradient was decreased to 30% (B) for 4 min, yielding a running time of 30 min. The calibration curves prepared for quantification, which comprised five concentrations between 0.1 and 10 μg L–1, exhibited determination coefficients (R2) > 0.999. The limit of detection (LOD) and limit of quantification (LOQ) were defined by signal-to-noise ratios (S/N) of 3:1 and 10:1, respectively (Yamashita et al. 2004). Recoveries of the seven target PFAS in the groundwater matrix were evaluated by spiking 10 μg L–1 of each PFAS standard into 1 L of the sample. Ultrapure water was used to prepare a blank sample following the procedure employed for the spiked samples. The recovery rates of the target compounds are listed in Table 1.

Table 1 Summarised data for analytical parameters involved in the HPLC–MS/MS analysis and recovery rates (%) of PFAS in spiked water samples
Full size table

Statistical analysis

Cluster analysis was used as a multivariate approach to determine the source apportionment of organic pollutants (Xiao et al. 2012). Cluster analysis is generally used to identify groups of similar individuals or objects. In this study, the source identification was evaluated using hierarchical cluster analysis with Ward’s method in the SPSS® Statistics 20 (IBM®) software package. The square of the Euclidean distance was used as an agglomeration technique. Preceding the analysis, any concentrations greater than or equal to the LOD, but less than the LOQ, were assigned a value twice that of the LOD, while any concentrations at or less than the LOD were assigned a value of zero (Yao et al. 2014). Each of the seven target PFAS concentrations was normalised using the total PFAS concentration to avoid misclassification resulting from their differing orders of magnitude. Thus, the PFAS exhibited similarities that were agglomerated in clusters.

Results and discussion

Concentrations of PFAS in groundwater around the MWDS and the IWDS and their distribution patterns

The concentrations of PFAS in the groundwater samples from around the MWDS and IWDS are presented in Table 2. Six target compounds, including PFOS, PFOA, PFHpA, PFNA, PFUnA, and PFHxS, were detected in samples collected near the Bang Sai MWDS. Five of the seven PFAS (PFOA, PFOS, PFNA, PFUnA, and PFHpA) were detected in samples collected near the Sena MWDS. The total PFAS concentrations in the groundwater samples collected near the MWDS vary from 1.68 to 7.75 ng L–1. The concentrations of PFAS in samples collected near the Bang Sai MWDS are in the following order: PFOS > PFOA > PFHpA > PFNA > PFUnA > and PFHxS, whereas PFOA is dominant in samples collected near the Sena MWDS. Evidently, the PFAS distribution display variations among the areas, although these areas are all near the MWDS. These variations are attributed to rain input and waste arrangement differences associated with the waste disposal sites. These differences likely impacted the initial leachate composition from which the PFAS in the groundwater samples originated (Eschauzier et al. 2013; Hepburn et al. 2019; Yong et al. 2021). The total PFAS concentrations in groundwater samples collected near the Nong Nae and Map Pai IWDS, where illegal industrial waste dumping occurred, varied as 4.43–10.80 ng L–1 and 2.64–42.01 ng L–1, respectively. Obviously, these concentrations are higher than those for samples collected near the MWDS. In samples collected near the Nong Nae IWDS, the target compounds PFOA, PFOS, PFHxS, and PFHpA were detected, whereas all target PFAS were detected in samples collected near the Map Pai IWDS. Among the target compounds, PFOA and PFOS were dominant, which is consistent with results obtained in studies from other countries. These results indicate that PFOS and PFOA are still utilised in industrial processes or are among the chemicals used in manufacturing consumer products. PFHxS was observed in most of the groundwater samples collected near the two IWDS. Although some studies have reported that PFHxS is being used as a substitute for PFOS-based compounds and has been reported in Thailand’s POPs Inventory Assessment Report (MTEC & NSTDA 2021), PFHxS substitution might not be directly implied in this case. PFHxS is not an appropriate alternative, owing to its long average half-life (Li et al. 2018) and higher liver toxicity than PFOS (Lloyd-Smith and Senjen 2015). Therefore, it poses a greater human health concern if such contaminated water is used for drinking. Notably, the total PFAS concentrations for samples collected near the Map Pai IWDS are significantly higher than that for samples collected near the Nong Nae IWDS. This difference is likely because of other factors, such as the soil components.

Table 2 PFAS concentrations (ng L–1) in groundwater near municipal waste disposal sites and industrial waste disposal sites in Thailand
Full size table

Potential source identification

The possible sources of the PFAS in the groundwater samples were classified based on hierarchical cluster analysis performed using the IBM® SPSS® Statistics 20 for Windows. The results produced three clusters of PFAS distribution patterns, and a dendrogram of these results is displayed in Fig. 2.

Fig. 2
figure 2

Dendrogram presenting hierarchical cluster analysis results

Full size image

Most groundwater samples collected near the Bang Sai MWDS, Sena MWDS and Nong Nae IWDS belong to clusters 1 and 2; these areas are surrounded by rural areas. Although Nong Nae is considered an IWDS, the PFAS concentrations in groundwater samples from this area fall within the group for samples collected near MWDS. Therefore, in addition to waste disposal, other factors likely account for the PFAS observed in groundwater samples from these areas.

Samples from the IW_CB07, IW_CB11 and IW_CB14 wells belong to clusters 3 and 4. This cluster displays the following unique PFAS distribution: PFOS > PFOA > PFHxS > PFNA > PFHpA > PFDA. Although wells IW_CB07, IW_CB11 and IW_CB14 are 2.15 km, 1.5 km, and 3.4 km away, respectively, from the Map Pai IWDS, the associated groundwater samples exhibit high PFAS concentrations. This might be explained by contributions from other sources because the IW_CB07 and IW_CB14 wells are near large, abandoned ponds, while the IW_CB11 well is adjacent to a pig farm, as shown in Fig. 2. However, it is unlikely that the pig farm is a potential PFAS source, as pig farms do not use PFAS extensively in their practices. Although the intake and elimination of PFAS by animals, particularly pigs, has rarely been reported, the presence of PFAS in wastewater from livestock farming cannot be ignored. Lai et al. (2016) indicated that wastewater from the livestock industry is a potential PFAS source in Kinmen Lake, Taiwan. However, it is unclear whether the pig farm near well IW_CB11 in this study is a potential source of the PFAS observed in the groundwater samples. Further analyses should be conducted to confirm or refute the role of this pig farm. Regarding wells IW_CB07 and IW_CB14, the PFAS could not be assigned to other sources because the function of the abandoned ponds was undetermined.

Clusters 5 and 6 comprise most groundwater samples collected near the Map Phai IWDS. These clusters exhibit similar PFAS concentration patterns, with PFOA dominating, followed by PFOS. However, the total PFAS concentration in cluster 6 is significantly higher than that for cluster 5. The highest total PFAS concentration was measured in a groundwater sample collected near the Map Phai IWDS, which is in an industrial area. Therefore, similar to other studies, PFAS in samples from industrial or urban areas largely surpassed those in samples from rural areas (Wang et al. 2012; Chen et al. 2016). Notably, groundwater samples collected near the Nong Nae, Chachoengsao and Map Phai and Chonburi IWDS are in different clusters, despite their collection near IWDS. However, aside from the potential sources, PFAS contamination and transport in the associated groundwater were likely influenced by location and flow (i.e., whether the sampling points were located upstream or downstream of the potential sources). Therefore, further research with a larger sampling size should be conducted, including investigations of the groundwater flow and its direction, to further explore the associations between the potential sources and contaminant concentrations.

Spatial distribution of PFAS in groundwater around the MWDS and IWDS

According to the results presented in Table 2, high PFAS concentrations are associated with groundwater samples collected near IWDS, indicating that IWDS significantly contributed to groundwater contamination. However, PFAS levels in groundwater samples collected near the Nong Nae and Map Pai IWDS differ remarkably, even though both sets of samples were collected near similar contamination sources. Therefore, other factors likely influenced the concentrations of PFAS in groundwater around both sites, and these account for the observed differences. Studies on the lateral distribution of PFAS may illustrate such factors and their characteristics. The horizontal distribution of PFAS was evaluated using ArcGIS 10.1 in the present study based on a soil map obtained from the Land Development Department (LDD) of Thailand.

Figure 3a and b reveal that soils in the Ayutthaya area, where the sampling points are located, comprise soil series Ayutthaya (Ay) and Sena (Se). The Ay and Se soils contain mostly clay, and thus, their main physical property is very low permeability, while their major chemical properties are high acidity values (pH 5.5–6 and 4–5.5, respectively) and high cation exchange capacities (CEC) (LDD 2010). Therefore, the low PFAS concentrations in groundwater samples collected near MWDS are likely linked to interactions between the PFAS and cations in the soils, which is consistent with observations by Xiao et al. (2015). In addition, Wang and Shih (2011) indicated that adsorption increases with decreasing pH. They reported that Ca2+ and Mg2+ can form bridges with PFOA anions and that PFOS can be bridged by Ca2+. Therefore, adsorption probably represents the principal mechanism through which PFAS contamination occurs in these areas.

Fig. 3
figure 3

Map displaying PFAS concentrations (ng L–1) in groundwater, their distribution, and soil characteristics near the a Bang Sai MWDS and b Sena MWDS

Full size image

The concentrations and distribution of PFAS from samples collected near IWDS are displayed together with soil types in the areas in Fig. 4a and b. Evidently, the PFAS concentrations in groundwater samples collected near the Nong Nae IWDS are significantly lower than those for samples collected near the Map Pai IWDS. These differences are probably linked to the properties of the soils in the two areas. In the area hosting the Nong Nae IWDS, soils around the sampling points include the Klaeng (Kl) and Don Rai (Dr), whereas, in the Map Pai IWDS area, samples were collected in areas with the Ban Bueng (Bbg) and Chonburi (Cb) soils. The Kl and Dr soils are characterised by moderate CEC, low permeability and pH values ranging between 4.5 and 6.4, and these properties are consistent with those of other soils in the Ayutthaya area. In contrast, low CEC, high permeability and pH values varying between 5.5 and 8.5 have been reported for the Bbg and Cb soils (LDD 2010). Therefore, permeability and the interaction of negatively charged PFAS and cations in the soils in association with the CEC level likely significantly influenced the distribution of PFAS in the groundwater of the areas studied.

Fig. 4
figure 4

Map of showing PFAS concentrations (ng L–1) in groundwater, their distribution, and associated soil characteristics near the a Nong Nae IWDS, and b Map Pai IWDS

Full size image

Conclusion

The conclusions drawn from the findings of this study and opportunities for future considerations are as follows:

  • PFOA and PFOS were the most abundant of the target PFAS in all samples, and their concentrations in samples collected near IWDS surpassed those of samples collected near MWDS, which indicates that PFAS were released in higher quantities from IWDS.

  • The data suggest that hierarchical cluster analysis can be used to identify potential PFAS sources. Groundwater sample classification was based on the total PFAS concentrations and patterns. Excluding direct contamination sources, the PFAS concentrations in the groundwater samples suggest the involvement of additional factors. In addition to the impact of direct sources, interactions between the associated soils and PFAS influenced the groundwater PFAS levels in various areas. Therefore, deep clay layers, which represent major components of soils near the two MWDS in Ayutthaya and the Nong Nae IWDS in Chachoengsao, can protect the aquifers, reduce PFAS transport, and better adsorb these contaminants than the sandy soil near the Map Pai IWDS in Chonburi.

  • This study demonstrates that waste disposal site leachates represent a likely major PFAS source in groundwater in Thailand. Further research related to PFAS transport in the soil column is required to elucidate their transport mechanisms.