Occurrence and risk assessment of perfluoroalkyl substances in the river basin around fluorine industry parks, China

ABSTRACT

Fluorine industry parks (FIPs) play a fundamental role in the advancement of China’s fluorine industry. In this study, Target analysis was combined with traceability analysis to determine the type of PFASs present in the surface water and sediment around FIPs in China. Meanwhile, the risk they posed to the surrounding river was examined through risk characterisation ratio. Among the areas surveyed, 10 and 15 PFASs were detected in the surface water and sediment, with concentrations of up to 32,674.0 ng/L and 17.84 ng/g, respectively. There were a larger proportion of short-chain PFASs (C4-C7), and perfluoropentanoic acid was particularly prevalent. Source analysis indicates that the PFASs in the surface water may have originated from precursor degradation and atmospheric deposition. The fluorine rubber industry was the main source for the sediment PFASs, particularly for short-chain compounds. A risk characterisation ratio analysis identified perfluorododecanoic acid, perfluorobutane sulfonate, perfluoroheptanoic acid, perfluoropentanoic acid, and perfluorohexanoic acid as compounds with relatively high risks.

KEYWORDS:

• fluorine industry parks
• surface water
• sediment
• source analysis

1. Introduction

Perfluoroalkyl substances (PFASs) have unique physical properties, and therefore have wide applications both in industry and daily life [1–5], however there are several disadvantages to these compounds. The increase in production and application of PFASs inevitably leads to their release into the environment, where they tend to persist due to limited photolysis, hydrolysis, and biodegradation [6], causing them to accumulate in organisms [7]. They thus pose a potential threat to the ecological environment and human health.

The long-chain PFASs (C8), especially, Perfluoro-octanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS), have been researched in the most detail [8–10]. Owing to their adverse effects, PFOS, PFOA, and related chemicals and precursors appear on various control and restriction lists, such as the Stockholm Convention on Persistent Organic Pollutants [11]. Despite this, some of these toxic long-chain derivatives are not restricted, and short-chain PFASs (C4-C7), which can be used as substitutes for long-chain compounds can increase atmospheric emissions [12]. It is therefore necessary to determine the distribution and environmental risks of PFASs.

Direct emissions from the fluorine chemical enterprises, processing facilities, and effluents from sewage treatment plants are the primary sources of PFASs in the environment [13]. As one of the largest PFASs producers and consumers globally, China’s fluorine enterprises are clustered in over ten fluorine industrial parks (FIPs). This study focuses on the Chinese FIPs to determine the distribution, sources, and risks of 17 PFASs in surface water and sediment. The study’s main goals are (1) investigate the type and concentration level of PFASs in the surface water and sediment around FIPs; (2) qualitatively analyses the source and contribution rates of PFASs through the ratio method and principal component analysis-linear regression analysis (PCA-MLR) model; (3) evaluate the ecological risk in the aquatic environment by the risk characterisation ratio (RCR).These results clarify the types of PFASs in FIPs and provide technical support for the formulation of emission limits.

2. Materials and methods

2.1. Chemicals

A total of 17 PFASs standards and 9 internal standards were purchased from Wellington Laboratory (Guelph, Ontario, Canada); the purities of all chemicals were ≥ 98% (Table S1). Methanol and acetonitrile (HPLC grade) were purchased from Fisher Scientific Ltd (Loughborough, UK). Ammonium acetate and ammonia (HPLC grade) were purchased from Anpel Laboratory (Shanghai, China). Ultrapure water was obtained from Watson (Guangdong, China).

2.2. Sample collection

In June 2021, a total of 33 samples comprising 23 surface water and ten sediments were collected from the rivers around FIPs in China. A detailed map of the sampling site is shown in Figure S1. The FIPs was in the lower reaches of the Yangtze River, with surrounding rivers including Zouma Pond, Fushan Pond, Shacao River, Cuipu Pond. Affected by runoff and tide, the hydrographic features uneven, complex, and diverse characteristics. Twenty-three surface water samples were collected from Zouma Pond (C-1, Z-1 to Z-6), Fushan Pond (F-1 to F-5), Cuipu Pond (P-1 to P-5), and the Shacao River (Y-1 to Y-6), and ten sediment samples were collected from Cuipu Pond (W-0 to W-2), Fushan Pond (W-3 to W-5), and the Shacao River (W-6 to W-9). 1 L amber polyethylene terephthalate (PET) bottle was used to collect water and sediment samples. All samples were transferred to the laboratory and analysed within 24 hours. The sample during the transfer process was stored at 4°C.

2.3. Sample pretreatment and instrumental analysis

2.3.1. Sample pretreatment

Suspended solids of surface water samples were removed by a 0.45 μm glass fiber filter (Whatman, Maidstone, UK). All filters had been previously combusted at 450°C for 5 h prior to use. A mix of internal standards (5 ng, Table S1) was added to 100 mL of filtrate to obtain a final concentration of 50 ng/L. Solid-phase extraction (SPE) by a WAX cartridge (6 cc/500 mg, Agela & Phenomenex, CA, U.S.A) was used to extract PFASs from the surface water samples. Before extraction, the WAX cartridge was preconditioned by sequentially adding 4 mL of 0.1% ammonia/methanol solution, methanol, and ultrapure water. Water samples were then passed through a cartridge at a rate of 1 drop/s. Following extraction, the analytes were eluted with 4 ml 25 mM ammonium acetate buffer solution, which was then dehydrated under negative pressure. After elution with 4 mL of methanol and 4 mL of ammonia/methanol solution, the analysis samples were collected in 15 mL centrifuge tubes. The extract was concentrated to 1 mL under a gentle nitrogen stream. Finally, the extracts were collected in amber autovials and stored at −20°C.

For sediment pretreatment, 50 ng of the internal standard and 10 g of sediment were added to methanol-rinsed polypropylene (PP) tubes. Then, 10 mL of a 50% acetonitrile (ACN) water solution was added to the samples. The mixture was placed in a sonication bath for 30 min. After sonication, the mixture was centrifuged at 4000 rpm for 10 min This process was repeated twice, and the collected supernatant was concentrated to 1 mL by nitrogen blowing. The concentrated solution was mixed with 8 mL of 2% formic acid and 42 mL of ultrapure water and treated according to the water extraction procedure.

2.3.2. Instrumental analysis

Quantitative analysis of the 17 PFASs was performed using LC-MS/MS operated in the electrospray negative ionisation mode. A total of 10 μL analyte was injected onto a 2.1 mm × 100 mm ACQUITY UPLC BEH C18 column (1.7 µm), with a mobile phase of ACN (A) and 2 mM ammonium formate solution (B), at a flow rate of 0.4 mL/min. Multiple reaction monitoring (MRM) was performed to identify the analytes (Table S2). The mobile phase was initially 5% A, which then increased from 5% to 95% in 10 min, stayed constant for 2 min, decreased to 5% at 0.1 min, and then maintained constant for 3 min. The MS operating conditions were set as follows: ion source temperature, 150°C; Capillary voltage, 2.5 kV; Dissolvent gas temperature, 400 ºC; Dissolvent gas flow, 1000 L/h; The air flow of the taper hole was 50 L/h; the collision gas was argon and dissolvent gas was nitrogen.

2.4. Quality control and quality assurance

Parallel samples were collected (10% of the sample sites, including three water and two sediment sites), and following this, field and transportation blanks prepared in ultrapure water were used to monitor cross-contamination. The 17 PFASs were quantified using an internal standard calibration curve diluted with methanol (0.1, 0.5, 1, 2, 5, 10, 20, 50 and 100 g/L). The instrumental limits of detection (LOD) and quantification (LOQ) had signal-to-noise (S/N) ratios of 3 and 10, respectively [14]. The LODs varied from 0.037–0.169 ng/L (ng/g) and the corresponding R² ranged from 0.990–0.997 (Table S3). Sampling and extraction procedures were considered uncontaminated when all pollutants in the procedural blank were below the LOQ. The recoveries of the spiked samples ranged from 61.0% to 117.0% in surface water and from 61.3% to 115.0% in sediment samples, respectively.

2.5. Data analysis

2.5.1. Source identification

Whether the PFASs were sourced from atmospheric deposition, rainfall, or precursor degradation is determined by the ratio method (PFHpA/PFOA, PFOS/PFOA, and PFOA/PFNA) [15,16]. If the PFHpA/PFOA ratio is greater than 1, it is likely due to atmospheric deposition. When the PFOS/PFOA ratio is lower than 1, rainfall input is the main source. When the PFOA/PFNA ratio was greater than 15, the degradation of the precursor has had an impact.

Through PCA, the principal components (PCs) was obtain by using Kaiser normalisation and varimax rotation of the concentration data matrix of PFASs. MLR was used to determine the contribution weight of sources by establishing a relationship between PCA factor scores and the average concentration of ∑ PFASs, which the contribution amount (ng/L) to ∑PFASs at each site can be calculated using Equation (1)(1) $\begin{array}{rl}& \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{b}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{s}=\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\sum \mathrm{P}\mathrm{F}\mathrm{A}\mathrm{S}\mathrm{s}\\ & \times \left(\mathrm{B}\mathrm{i}/\sum B\mathrm{i}\right)+\mathrm{B}\mathrm{i}\times \delta \times \mathrm{F}\mathrm{S}\mathrm{i}\end{array}$(1) :

(1) $\begin{array}{rl}& \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{b}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{s}=\mathrm{a}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\sum \mathrm{P}\mathrm{F}\mathrm{A}\mathrm{S}\mathrm{s}\\ & \times \left(\mathrm{B}\mathrm{i}/\sum B\mathrm{i}\right)+\mathrm{B}\mathrm{i}\times \delta \times \mathrm{F}\mathrm{S}\mathrm{i}\end{array}$(1)

B_i represents the contribution of each factor to the MLR, where B_i/∑B_i is the percentage contribution of each factor. FSi is PCA factor score for factor I; average ∑PFASs (ng/L) is the average value of ∑PFASs concentration; δ is the standard deviation of the total concentrations.

2.5.2. Environmental risk assessment

Based on the technical guidelines for environmental and health risk assessment of chemical substances [17], the environmental risk assessment of PFASs in surface water around FIPs was computed using the environmental RCR. The RCR is measured as the ratio of the environmental concentration of pollutants (MEC) to the predicted no-effect concentration (PNEC), which is classified as high (RCR > 1), medium (0.1 < RCR < 1), or low risk (RCR < 0.1). There are two ways to obtain PNEC: 1) species sensitivity distribution (SSD) and 2) the evaluation coefficient method. Previous studies have shown that the PNEC of PFOA is 386.09 μg/L, using SSD curves [10]. The PNEC of the remaining detected PFASs were estimated using the evaluation coefficient method, and the value was calculated as the ratio between NOEC, LC₅₀ or EC₅₀, and the assessment factor (AF). The toxicities of PFASs in fish, invertebrates, and algae were obtained from the USEPA ECOTOX database and EPI Suite (Table S4). The AF was determined based on the acute and chronic toxicity data of the pollutants.

3. Results and discussion

3.1. Concentrations and composition of PFASs

As shown in Table 1, ten PFASs were detected in 23 surface water samples around the FIPs, and perfluorobutanoic acid (PFBA), perfluorohexanoic acid (PFHxA), perfluoroheptanoic acid (PFHpA), PFOA, and perfluorobutane sulfonate (PFBS) were all detected. The detection rates of perfluoropentanoic acid (PFPeA) and perfluorononanoic acid (PFNA) were 91.3% and 83.0%, respectively. The remaining PFASs were detected sporadically with a rate of < 15.0%. The total concentrations of PFASs ranged from to 14,055.1–32674.0 ng/L, which were higher than those in the surface waters of eastern China, the Yangtze River and the Pearl River [16]. This result is consistent with previous studies [18]. Previous studies also shown that the total concentrations level of PFASs in fluorine industry impacted region was higher than that in another regions. Among the detected PFASs, PFPeA had the highest concentration (11623.4 ng/L), followed by PFHxA (5523.2 ng/L), PFBS (2462.6 ng/L) and PFHpA (2436.0 ng/L). PFPeA was dominant in the surface water of the FIPs, contributing 47.9% of the total PFASs, followed by PFHxA (23.2%), PFHpA (10.5%), PFBS (10.6%), PFBA (6.9%), and others (0.9%). The results showed that short-chain PFASs (C4-C7) were dominant in the surface water around the FIPs, which is consistent with previous studies [18,19]. This implies that, owing to their widespread production and use, short-chain PFASs (C4-C7) have gradually become the main pollutants in FIPs. In addition, the PFOA concentrations in surface water ranged from 33.7–533.0 ng/L, which is higher than that in the Chongqing and Yichang section of the Yangtze River [20,21]. PFOA is primarily derived from polytetrafluoroethylene (PTFE) [22]. High levels of PFOA in the surface water around the FIPs may be related to PTFE production. PFOS was not detected in the surface water, reflecting the local government’s efforts to implement the Stockholm Convention.

Table 1. The concentrations PFASs in surface water (ng/L) and sediment (ng/g) samples.

Matrix	Compound	Range	Mean	Median	Detection rate (%)
surface water	PFBA	815.3–4791.7	1623.7	1295.2	100(23/23)
	PFPeA	ND-15019.4	11623.4	12580.3	91.3(21/23)
	PFHxA	2333.6–15039.2	5523.2	4658.1	100(23/23)
	PFHpA	2304.3–2579.2	2436.0	2414.9	100(23/23)
	PFOA	33.7–533.0	192.5	137.5	100(23/23)
	PFNA	ND-7.1	2.6	2.6	83.0(19/23)
	PFDA	ND-1.7	0.2	ND	13.0(3/23)
	PFDoDA	ND-3.7	0.3	ND	8.6(2/23)
	PFBS	2271.8–2829.3	2462.6	2465.4	100(23/23)
	PFHxS	ND-1.1	ND	ND	4.3(1/23)
sediment	PFBA	ND-7.0	0.7	ND	10.0(1/10)
	PFPeA	ND-7.8	1.9	1.4	80.0(8/10)
	PFHxA	ND-1.5	0.6	0.6	90.0(9/10)
	PFHpA	ND-0.8	0.2	0.1	70.0(7/10)
	PFOA	ND-4.1	0.7	0.3	80.0(8/10)
	PFNA	ND-0.3	0.02	ND	10.0(1/10)
	PFDA	ND-2.3	0.4	ND	30.0(3/10)
	PFUnDA	ND-2.0	0.4	ND	30.0(3/10)
	PFDoDA	ND-1.4	0.5	0.4	90.0(9/10)
	PFTrDA	ND-0.6	0.2	0.09	50.0(5/10)
	PFTeDA	ND-0.3	0.08	ND	30.0(3/10)
	PFHxDA	ND-0.2	0.02	ND	10.0(1/10)
	PFODA	ND-0.16	0.02	ND	10.0(1/10)
	PFBS	ND-0.5	0.2	0.2	90.0(9/10)
	PFOS	ND-1.57	0.2	ND	20.0(2/10)

The concentrations of PFASs in the sediment samples collected from the surroundings of the FIPs are shown in Figure 1. With the exception of PFHxS and PFDS, the remaining 15 PFASs were detected at total concentrations ranging from 0.25–17.84 ng/g, and mean concentration of 6.2 ng/g. Previous studies have shown that the mean concentrations in the sediments of the Pearl River Delta, Laizhou Bay, Bohai Sea, Zhujiang River and Huangpu River were 3.54 ng/g, 5.25 ng/g, 0.42 ng/g, 1.8 ng/g, and 0.7 ng/g, respectively, which were lower than demonstrated by the results of this study [22–25]. Six of the 17 target compounds were the most frequently detected PFASs, with detection rates higher than 70.0%. PFPeA had the highest concentration (1.9 ng/g) among the detected PFASs, followed by PFOA (0.7 ng/g) and PFHxA (0.6 ng/g). PFPeA and PFHxA were the dominant PFASs in the sediment samples from the surrounding FIPs, accounting for 33.2% and 11.1% of the concentrations, respectively. Previous studies have found that short-chain PFASs have greater transport potential than long-chain, in soils and sediment [16]. PFPeA and PFHxA, which are typical short-chain compounds, may have a greater potential for transport in sediment.

Figure 1. Concentrations of the PFASs detected in the river basin around the FIPs. (a) surface water; (b) sediment.

3.2. Source analysis

The degradation of precursor compounds is an indirect source of environmental PFASs, which are primarily derived from production, transportation, use, and disposal [3]. In the surface water, PFHpA and PFOA were detected at all sites, with PFHpA/PFOA ∈ [4.5, 68.7] > 1, indicating that atmospheric deposition has a relatively important contribution [15]. In addition, PFOA/PFNA in surface water ranged from 11.9 to 283.8, which is greater than PFOA/PFNA ∈ [2,15], indicating that the degradation of precursors is a source of PFASs [26]. In the surface water of the FIPs, the PFOS/PFOA ratio was zero at all sites, indicating that rainfall had a notable effect.

PCA was applied to distinguish the sources of PFASs in the surface water (Figure 2a). After varimax rotation, three factors accounting for 65.4% of the total loading were identified. Long-chain PFASs, including PFOA, PFNA, PFDA, and PFHxS, were included in the first component (32.6%). PTFE is considered to be the source of PFOA [22], and PFNA is a necessary processing aid in the production of fluoropolymers [19]. The second component (18.31%) comprised PFBS, PFHxA, and PFPeA. PFBS is used primarily in food packaging. The second component is an indicator of food packaging and metal plating sources. The third component contained PFHpA, PFBA, and perfluorododecanoic acid (PFDoDA), which are involved in the synthesis of precious metals. As shown in Figure 2b, site Y-3 was located far from the other sampling sites in the score plot, implying an alternative source. The principal components of Y-3 included PFPeA and PFHxA, indicating that the PFASs in Y-3 may have originated from metal plating.

Figure 2. Three-dimensional principal component in surface water. (a) loading plot and (b) score plot.

Three factors were identified in the sediment samples, accounting for 82.13% of the total load (Figure 3). The loadings of factor 1 were 39.86% for PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFBS, and PFOS, which originated from processing aids, precursor conversion, and food packaging. Principal component 2 (26.31%) was closely associated with PFBA, PFPeA, PFHxA, and PFHpA, which are indicators of precious metal and coating industries. Component 3 (15.96%) represented PFTeDA, PFHxDA, and PFODA, which are long-chain perfluorinated compounds. These PFASs are considered to be sources of fluorine rubber and PTFE powder. In terms of regional distribution, the source of PFASs at point W-1 was mainly affected by the conversion of processing aids and precursors, point W-5 by the fluorourubber industry, and point W-8 by the precious metals and coating industries.

Figure 3. Three-dimensional principal components in sediment. (a) Loading plot and (b) Score plot.

(2) $Z=0.24\ast FS{ }_1+0.22\ast FS{ }_2+0.54\ast FS{ }_3$(2)

The contribution rates of the three factors in the PCA were determined using MLR. The PCA-MLR model is expressed in Equation (2)(2) $Z=0.24\ast FS{ }_1+0.22\ast FS{ }_2+0.54\ast FS{ }_3$(2) , yielding an excellent coefficient of determination (R² = 0.982). Component 3 contributed to the concentration of PFASs in the sediment with a mean contribution of 54%. The results show that the PFASs in the sediments around the FIPs mainly originate from the fluorine rubber and PTFE industries, followed by the precious metal and coating industries. The conversion of processing aids and precursors can introduce perfluorinated compounds into sediments.

3.3. Risk assessment

To investigate the risk that the detected PFASs pose to aquatic organisms in the surface waters of the FIPs, RCR values for aquatic organisms were calculated (Figure 4). Among the ten PFASs, PFDoDA posed a high risk (RCR > 1) at two sampling sites (Y-3 and P-3), indicating that in some surface water basins it may have adverse effects on the aquatic ecosystem. Short-chain PFASs (C4-C7) posed a median risk (0.1< RCR < 1), with the point ratios of PFBS, PFHpA, PFHxA, and PFPeA being 100%, 100%, 100%, and 91%, respectively. For the C4-C7 PFASs, the mean RCR of short-chain PFASs followed the trend: PFHpA > PFHxA > PFPeA > PFBS. Among long-chain PFASs, PFNA showed only slight impact on aquatic organisms, obtaining an RCR value between 0.01 and 0.1, indicating a potential ecological risk. The RCR of short-chain PFASs was higher than that of long-chain compounds, differing from the results for other regions [18,27,28]. Previous studies have shown that long-chain PFASs pose a higher environmental risk to aquatic organisms in lakes impacted by the fluorine industry in China. These results indicate that it is necessary to strengthen the management and control of short-chain PFASs in FIPs, particularly PFBS, PFHpA, PFHxA, and PFPeA, to reduce ecological disturbance and protect the biodiversity in this region.

Figure 4. Risk characterisation ratio (RCR) of the 10 PFASs to aquatic organisms.

4. Conclusion

This study showed that the distribution, sources, and risks of PFASs in river basins around the FIPs through experimental analysis. Short-chain PFASs (C4-C7) gradually accumulated to become dominant in the surface water and sediments around the FIPs, which indicated the short-chain substitution effect has gradually appeared in the FIPs of China, due to China was striving to implement the Stockholm Convention. PFASs in surface water may originate from the degradation of precursors and atmospheric deposition, with the fluorine rubber industry identified as the main source in the sediment, particularly for short-chain compounds. Risk assessment indicated that PFDoDA, PFBS, PFHpA, PFHxA, and PFPeA were found to pose relatively higher risks to aquatic organisms in the surrounding river around the FIPs. This indicates the necessity of strengthening regulations surrounding the emissions of short-chain PFASs in FIPs. However, the present study still has limitations. Although 33 samples covered the internal and surrounding rivers of the FIPs, the frequency of the samples is still limited to accurately reflect the actual condition of PFASs. Therefore, the following research should increase the sampling frequency to reveal the spatiotemporal distribution of PFASs in the river basin around FIPs.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

All data reported here can be made available on request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/26395940.2023.2218569

Additional information

Funding

This research was funded by Research Project of Ecological Environment in Jiangsu Province (2022014, 2022003) and Research and Development Project of Jiangsu Environmental Engineering Technology Co., Ltd (No. JSEP-GJ20220017-RE-ZL).