Iron and Molybdenum Isotope Application for Tracing Industrial Contamination in a Highly Polluted River

Abstract

Rivers adjacent to industrial zones usually suffer from severe pollution issues. Industrial wastewater that has undergone sewage treatment processes may be legally discharged into rivers under water quality permits. Previous studies have frequently employed isotopic tracers to identify potential contaminants for pollution control. Conventional radiogenic isotopes utilized in tracing studies cannot discern whether the source is untreated (primary) industrial wastewater, which can have serious impact to the environment. By analyzing the iron (Fe) and molybdenum (Mo) isotopic compositions in original industrial wastewater and treated effluent, this study aims to investigate whether the heavily polluted Agongdian River is contaminated by the untreated wastewater. Based on the results from this study, the original industrial wastewater exhibits higher concentrations of metallic elements and heavier Fe and lighter Mo isotopic compositions, compared to the treated effluent. Consequently, it appears that Agongdian River water indeed exhibits evidence of untreated industrial wastewater. Furthermore, the volume of original industrial wastewater entering the river can be estimated from these results. This research offers a more precise and accurate approach to monitor potential industrial wastewater pollution in natural water bodies, contributing to the goal of environmental protection and sustainable development.

1. Introduction

Pollution control over water resources is crucial in all human activities. Contaminated water may raise human health risks [1,2] and cause ecological hazards [3], and the pollutant monitoring protocols are thus necessary for water resources management [4,5]. While water quality monitoring alone can only provide safety information about water consumption, relying solely on water quality monitoring is far from sufficient to protect water resources. Consequently, efficient and accurate means to identify the sources of pollutants and tracking the pathways of their dispersion have become critical in water resources management. It is essential to intercept and stop the pollutants from entering the natural environment.

The application of isotopic composition in source tracing has been employed on the lithosphere, hydrosphere, atmosphere, and even biosphere [6,7,8,9]. The utilization of traditional stable isotopes (i.e., H, N, and O) to trace the sources of pollutants in water bodies is a well-established research technique [10,11]. The isotopic signature of metal elements has become increasingly important because of their significant inputs from urban and industrial sewages [12]. Traditionally, radiogenic isotope (i.e., Sr, Nd, and Pd) systems were the tools used in source tracing studies; this is because their isotopic signals will not be fractionated during transport [13,14], and the isotopic signals also remained unmodified after physical and chemical processes. Consequently, radiogenic isotopes seem to be promising tools for pollutant source tracking [15,16,17,18,19]. While the principal strategy for wastewater management is the removal of pollutants from sewage [20], and it requires additional technologies capable of identifying whether industrial wastewater has undergone sewage treatment before discharge. Traditional radiogenic isotope systems were incapable of discerning whether discharged wastewater has undergone treatment, because their isotopic signals will not be fractionated after these processes. Therefore, non-traditional stable isotopes, such as Mo and Fe isotopes, are used to assist in the process.

The sewage treatment methods for industrial wastewater could be classified into three major categories, including physical, chemical, and biological processes [21,22], and the main purpose of these methods is to remove pollutants from aqueous phases [20,22]. Many of the material removal procedures in liquid phase happening in nature, i.e., precipitation, hydrothermal mineral deposits, and biological absorption, have been observed and explained via mass-dependent metal stable isotopic fractionation. It has been shown that a significant Mo isotopic offset was found between seawater and ferromanganese crusts, and the observed isotopic fractionation occurred when Mo in the seawater was incorporated into Fe-Mn oxides [23,24,25]. While Mo was removed from the water column into the underlying sediments, its isotopic composition could be very sensitive to the redox condition [23,26,27,28]. Therefore, the Mo isotopic signature in sediments varies significantly from oxic to euxinic, and has been utilized to study the paleo-redox conditions [25,26,27,28,29]. Molybdenum isotopic variations in molybdenites were also presented, and the fractionation processes most likely were associated with the formation of molybdenites in different ore deposits [30,31]. Living organism preferred lighter Mo isotopes was also observed in bacteria [32,33], which indicates that both abiotic and biotic processes could lead to mass-dependent isotopic fractionation of Mo. The variations in Mo isotopes in different natural reservoirs due to anthropogenic pollution have also been observed in the literature. The Mo isotopic fractionation can be utilized to discern whether the natural water reservoirs, such as rivers and oceans, have been affected by anthropogenic pollution [34,35,36]. Lastly, Mo isotopic differentiations could also offer the further investigation of anthropogenic contamination in various nature environments such as sediments, soils, and aerosols [37,38,39].

Molybdenum is one of the transition metals that with observed isotopic fractionation during material transferring/removing. Besides Mo, scientists also have a wide interest in Fe isotopes, since it is one of the most abundant elements on Earth. Previous studies have shown that during chemical and physical weathering/reaction, the leached Fe would be isotopically lighter comparing to the host or the source rocks [40,41]. Iron isotopes could also be fractionated during the formation of ore deposits, e.g., sulfide deposits, nickel deposits, and iron oxide-apatite deposits [42,43,44]. Lastly, Fe has also shown isotopic fractionation during high temperature processes, including partial melting, fractional crystallization, and even core formation [45,46,47,48]. Consequently, both Fe and Mo isotopes are highly effective tools in studies related to material transport. The fractionations of these isotopic systems can be used to study the removal of pollutants from sewage.

This study aims to provide constraints for better understanding of the source tracing and transport pathways of heavy metal pollutants and to minimize the impact of anthropogenic pollution on the natural environment. In order to investigate the correlation between pollutants in rivers and industrial wastewater, the concentrations and isotopic characteristics of metal elements in the river water and the industrial wastewater were analyzed. Based on the isotopic fractionation between treated and untreated sewage from different individual industries, it is possible to identify the source of pollutants, and to develop potential solutions to remedy the pollution issues affecting the rivers near industrial areas.

2. Materials and Methods

2.1. Study Area

Kangshan Benjhou Industrial Park is about 2.1 km² and is dominated by metal processing and manufacturing industries. It locates in the southwestern Taiwan (Figure 1) and resides at the Agongdian River catchment. The Agongdian River is 38 km long and originates from Wushanding, Kaohsiung City. The catchment area is 137 km² in size and comprises mostly Quaternary deposits. Since the major tributary of the Agongdian River runs through Kangshan Benjhou Industrial Park, the industrial wastewater was designed to be released into the Agongdian River system after treatment. Consequently, the Agongdian River has long been one of the most severely polluted rivers in Taiwan. In 2002, only 16% of its length was considered relatively clean, while over 50% of the length was severely polluted [49]. Despite years of remediation efforts on sewage restriction policy, still only less than 30% was considered relatively clean, with more than 15% of its length still remaining under severe pollution in 2022 [50].

2.2. Sample Collection and Preparation

Industrial wastewater samples, including untreated (before wastewater treatment) and treated (after wastewater treatment) wastewater, were collected from 12 different industries and the sewer system of the industrial park on 5 May and 7 July 2020. These sewage samples were collected as bulk water samples (unfiltered) to study all metal phases. For a better comparison with industrial wastewater, a time-series water sampling was scheduled during June to August of 2020 on a tributary of the Agongdian River following the suggested protocol (NIEA W104) of Environmental Protection Administration (EPA) of Taiwan as briefed below. Pre-cleaned polypropylene (PP) bottles were utilized in the process of sampling, and the sampling point was located approximately 60 cm below the water surface. Following the sampling procedure, the collected samples were acidified with double-distilled concentrated nitric acid (the volumetric ratio of sample to nitric acid is 1000:1). River water samples were collected twice daily at 9 a.m. and 9 p.m. during the period from 29 June to 10 August 2020, and the monitoring site (22.805161° N, 120.287799° E) was marked in Figure 1. The summer season represents the primary precipitation period in Taiwan, predominantly influenced by typhoons and tropical depressions [51]; therefore, samples collected under varying rainfall intensities were selected for analysis to comprehend the impact of precipitation on the study.

All of the samples were digested following the standard protocol of Taiwan EPA (NIEA W313, same as United States Environmental Protection Agency Method 200.8), which aims to release the total recoverable analytes from the samples. Two milliliters of 8 N HNO₃ and 1 mL of 6 N HCl were added to a 100 mL aliquot of acidified sample and heated on a hotplate at 85 °C. After the mixture was reduced to 20 mL by gentle heating at 85 °C, the lid was closed, and the sample was refluxed for 30 min. We diluted the remaining sample to 50 mL with Milli-Q^® water and filtered out the survived particles with 0.45 μm PTFE membrane.

2.3. Ion Exchange Chromatography

Ion chromatography is commonly used to separate the element of interest from the matrix, and purified single element solution is necessary for MC-ICPMS isotopic measurement. Two aliquots were taken from each digested sample, and they were spiked with Fe (⁵⁷Fe, ⁵⁸Fe) and Mo (¹⁰⁰Mo, ⁹⁷Mo) double isotope tracers, respectively, prior to column chemistry for correction of instrumental and laboratory mass bias. A spike-sample ratio of ∼1 was applied to all samples for both Fe and Mo double spike techniques.

2.3.1. Fe Purification

The chemical purification methods for Fe were largely followed that of Liang et al. (2020) [52]. Spiked samples were taken up in 1 mL 8 N HCl for chemical separation. Samples were loaded on 0.1 mL of Biorad AG1-X8 anion exchange resin, and rinsed with 1 mL 8 N HCl to remove most of matrix elements and possible isobaric interference (i.e., Cr and Ni), and then eluted with 1.5 mL 0.7 N HCl to collect Fe. The purification procedure was carried out twice for better purification of Fe. The total procedural blanks were <2 ng of Fe, and more than 95% of Fe was recovered after the column chemistry.

2.3.2. Mo Purification

The chemical purification procedure for Mo was followed that of Liang et al. (2017) [53]. Spiked samples were taken up in 5 N HCl, and subsequently loaded on 2 mL of Biorad AG1-X8 anion exchange resin. After being rinsed with 40 mL of 5 N HCl to remove most of the major and trace elements, samples were eluted with 30 mL of 1 N HCl. The eluents were dried and loaded on 2 mL of Biorad AG50W-X8 cation resin in 0.5 N HCl to remove Fe. The cation procedure was performed twice for complete separation of Mo from Fe, which may form polyatomic isobaric interference of Mo (i.e., FeAr) during mass spectrometry analysis. The total procedural blanks were <0.3 ng of Mo, and more than 95% of Mo was recovered after the purification procedure.

2.4. Mass Spectrometry

Iron and molybdenum isotopic measurements were performed via a Thermo Scientific Neptune Plus^TM High Resolution MC-ICP-MS using CETAC AridusII Desolvating Nebulizer System at the Institute of Earth Sciences, Academia Sinica, Taiwan. The Fe isotopic reference material IRMM-014 and the Mo isotopic reference material NIST SRM 3134 were used as laboratory and measurement standards.

The 10¹¹Ω resistors were used for both Fe and Mo isotopic measurement, and each measurement was the average of 63 cycles with an integration time of 4 s for each cycle. The solutions for Mo isotopic measurement contain between 10 and 15 ppb of Mo, and the ion beam intensities were usually kept between 150 and 250 V per ppm. The Mo isotopes that have isobaric interferences with Zr isotopes were not used in the calculation, and ⁹⁹Ru was monitored to ensure that there was no Ru interference above the instrument background level. For the Fe isotope analysis, in order to resolve analyte and polyatomic interference beams (i.e., ArO and ArN), the measurement was operated at the high-resolution mode, which reduced the transmission by 80%. Therefore, the concentrations of Fe in the measurement solutions were between 100 and 150 ppb, while the measurement ion beam intensities were usually between 30 and 50 V per ppm. All four of the Fe stable isotopes (⁵⁴Fe, ⁵⁶Fe, ⁵⁷Fe, and ⁵⁸Fe) were measured, and both Cr and Ni (⁵³Cr and ⁶⁰Ni) were monitored simultaneously to ensure that there was no Cr and Ni interference above the instrument background level. All the isotopic compositions were presented by the delta notation as the deviation in parts per thousand (‰):

$${\mathsf{\delta }}^{56/54}\mathrm{F}\mathrm{e}=\left(\frac{{\left({}^{56}\mathrm{F}\mathrm{e}/{}^{54}\mathrm{F}\mathrm{e}\right)}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}}{{\left({}^{56}\mathrm{F}\mathrm{e}/{}^{54}\mathrm{F}\mathrm{e}\right)}_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}}-1\right)\times 1000$$

$${\mathsf{\delta }}^{98/95}\mathrm{M}\mathrm{o}=\left(\frac{{\left({}^{98}\mathrm{M}\mathrm{o}/{}^{95}\mathrm{M}\mathrm{o}\right)}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}}{{\left({}^{98}\mathrm{M}\mathrm{o}/{}^{95}\mathrm{M}\mathrm{o}\right)}_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}}-1\right)\times 1000$$

The external reproducibility over the time period of sample analysis is ±0.047‰ (2σ) for δ⁵⁶/⁵⁴Fe values and ±0.08‰ (2σ) for δ⁹⁸/⁹⁵Mo values. To verify the accuracy of our isotopic measurement, a well-characterized USGS international rock standard (BHVO-2) was also included in each batch of sample processing and isotopic measurement, and the results (δ⁵⁶/⁵⁴Fe = 0.118 ± 0.046‰, 2σ, and δ⁹⁸/⁹⁵Mo = −0.03 ± 0.07, 2σ) showed a good agreement with values in the literature [54,55,56,57].

3. Results and Discussion

3.1. Industrial Wastewater

The Fe and Mo isotopic compositions for industrial sewage are shown in

Table 1. Fe and Mo isotope compositions and metal concentrations of Industrial wastewater.

Classification	Industry	Sampling Date	Type of Sewage	Cr (ppb)	Cu (ppb)	Fe (ppb)	Sr (ppb)	Mo (ppb)	Pb (ppb)	d56/54Fe	2 s.d.	d98/95Mo	2 s.d.
Manufacture of Basic Metals
	BM1	7 May 2020	Untreated	2.42 × 104	1.16 × 104	1.38 × 107	3768	2387	244	0.360	0.069	0.06	0.09
		7 May 2020	Treated	194	18	979	429	24.2	0.663	0.376	0.051	0.67	0.09
Manufacture of fabricated metal
	FM1	7 July 2020	Untreated	637	2896	7.60 × 105	414	286	139	0.370	0.048	−0.19	0.09
		7 July 2020	Treated	38.2	10.0	216	179	124	0.880	0.246	0.048	0.31	0.09
Manufacture of Food Products
	FP1	7 July 2020	Untreated	18.2	96.9	1360	819	4.98	8.72	0.216	0.051	0.37	0.09
Treatment of metal surface
	MS1	7 May 2020	Untreated	8.10 × 104	1678	4.05 × 105	62	1618	49.8	0.384	0.048	0.01	0.09
		7 May 2020	Treated	33.9	7.65	234	1377	340	1.05	−0.020	0.051	0.18	0.09
	MS2	7 May 2020	Treated	15.5	16.8	153	558	1.90	1.36	0.450	0.041
	MS3	7 May 2020	Treated	28.0	3.63	133	563	9.76	0.613	0.098	0.095	0.66	0.06
	MS4	7 May 2020	Untreated	8.58 × 105	1903	1.50 × 105	650	1088	62.8	0.462	0.069	0.22	0.09
		7 May 2020	Treated	5.96	2.40	56.2	511	88.5	0.613	−0.610	0.051	0.54	0.09
		7 July 2020	Treated	11.5	1.25	61.6	555	411	0.180	0.414	0.136	0.47	0.09
	MS5	7 July 2020	Treated	1.86	4.34	449	186	59.1	0.320	−0.946	0.013	0.50	0.09
	MS6	7 July 2020	Untreated	1.25 × 106	268	4.25 × 104	618	4.48	18.5	0.314	0.130	0.37	0.07
		7 July 2020	Treated	3.34	7.69	247	1391	1.69	0.580	−0.463	0.048	0.78	0.08
Industrial sewer system
	SS1	7 May 2020	Untreated	19.9	22.04	5705	471	2666	1.000	0.323	0.095	0.10	0.07
		7 May 2020	Treated	7.45	3.89	548	497	3334	0.825	−0.161	0.079	0.23	0.09
		7 July 2020	Treated	2.89	2.19	472	388	569	N/A	0.040	0.051	0.22	0.09
	SS2	7 May 2020	Treated	31.2	3.95	176	682	29.2	1.188	0.057	0.051	0.16	0.09
	SS3	7 May 2020	Treated	4.36	55.1	111	501	42.4	0.300	0.153	0.041	0.67	0.03

(expressed relative to the composition of the IRMM 014 and NIST SRM 3134 standards), together with Cr, Cu, Fe, Sr, Mo, and Pb concentration data published by Wu (2022) [58]. Seven out of these twelve industries have been analyzed for untreated wastewater. Metal concentrations showed a huge variation between different samples. Comparing to other types of manufacture, the industries related to metal processing and manufacturing have remarkably high metal content in the untreated sewage (Table 1). In the untreated sewage, the Cr concentration could be up to more than 0.1%, and the Fe concentration could be even more than 1%. The results also show that the sewage treatment is efficient in removing metal components (Table 1). After the treatment, there are no more than 200 and 1000 ppb of Cr and Fe, respectively, in the treated effluent (Table 1).

Both the Fe and Mo isotopic variations in sewage between different sorts of industries are not as significant as the concentration data. The δ⁵⁶/⁵⁴Fe values range from −0.946 ± 0.013‰ to 0.462 ± 0.069‰, while the δ⁹⁸/⁹⁵Mo values range from −0.19 ± 0.09‰ to 0.78 ± 0.08‰. There is no significant difference between metal industries and other industries (p values are 0.99 for Fe isotopes, and 0.65 for Mo isotopes). In contrast, the isotopic fractionations before and after sewage treatment were quite significant (p values are 0.03 and 0.007 for Fe and Mo isotopic signals, respectively). Studies in the literature have shown that the removal of materials from a reservoir may lead to Mo isotopic fractionation [23,25,26]. Reducing pollutants in wastewater is the primary objective of wastewater treatment, and metal elements present in sewage should also be simultaneously removed during the wastewater treatment process. According to this study, the metallic content within the wastewater of Kangshan Benjhou Industrial Park has indeed been successfully reduced following the implementation of sewage treatment (Table 1). The observed isotopic differences between untreated and treated wastewater (Table 1) might be due to the material removal process.

For further investigation, pairs of treated and untreated sewage samples were collected from six individual factories (Table 1) on the same day. Their Fe isotopic compositions generally become lighter after sewage treatment (the δ⁵⁶/⁵⁴Fe decreasing from −0.124 to −1.072) apart from industry BM1, which is indistinguishable within analytical errors (Table 1). In contrast, Mo isotopic signatures of treated effluent are notably heavier than untreated wastewater (the δ⁹⁸/⁹⁵Mo increasing from 0.13 to 0.61, Table 1 and Figure 2). In Figure 2, the isotopic differences (Δ^56/54Fe and Δ^98/95Mo) between treated and untreated sewages are plotted against the remaining concentration level of wastewater ([E]/[E]*). While no apparent correlation can be deduced for the Fe isotopes, there seem to be some linear correlations between the Mo isotopic signatures and metal contents (Figure 2). The correlations in Figure 2b indicate that with the successful removal of metal (Cu, Fe, and Pb) contents, the remaining Mo in the residual (treated effluent) would become heavier. The better correlation in Figure 2b than Figure 2a also suggests that the variation in Mo isotopes between treated and untreated sewage samples (Δ^98/95Mo) is more indicative over the reducing level of metal content in industrial wastewater than Fe isotopes (Δ^56/54Fe). Furthermore, our results suggest that metal stable isotopic fractionation could also be used to decipher between treated and untreated industrial water pollution, since the untreated sewage samples appear to be heavier in Fe isotopes and lighter in Mo isotopes (Table 2 and Figure 2).

3.2. Monitoring of River Water

During the monitoring period, the metal concentrations of the Agongdian River exhibited significant fluctuations. Within the period of our sampling, the variation in Fe concentration is the most pronounced by more than two orders (36-7814 ppb, Table 2). In addition to Fe, the concentrations of several transition metals in the Agongdian River also exhibit significant variability. Chromium and lead, albeit at relatively low concentrations relative to Fe, display nearly 40-fold variations (Table 2). The content of Cr in the monitoring site varies between 0.6 and 25.3 ppb, while Pb varies between 0.344 and 13.05 ppb (Table 2). Copper, on the other hand, demonstrates lower variability (slightly less than an order) in river water, and its concentration in the monitoring site ranges between 2.67 and 23.85 ppb (Table 2). The variation in Mo concentration is also minor, and it fluctuates within the range of 3.5 to 49.3 ppb (Table 2). The concentration of Sr in the river water at the monitoring site is relatively stable. Although the Sr abundances change between 176 and 853 ppb, this fluctuation is less than 5-fold (Table 2).

Table 2. Fe and Mo isotope compositions and metal concentrations of Agongdian River water.

Sample ID	Sampling		Precipitation (mm)	Cr (ppb)	Cu (ppb)	Fe (ppb)	Sr (ppb)	Mo (ppb)	Pb (ppb)	d56/54Fe	2 s.d.	d98/95Mo	2 s.d.
	Date	Time
B10930197	29 June 2020	21:00	0	2.72	4.44	323	675	49.3	0.887	−0.080	0.081	0.64	0.08
B10930198	30 June 2020	9:00	31	12.5	5.05	509	853	45.1	1.301	−0.092	0.085	0.40	0.08
B10930202	2 July 2020	9:00	4	3.81	9.76	767	547	17.0	2.717	−0.060	0.051	1.07	0.08
B10930210	6 July 2020	21:00	0	3.35	4.78	492	641	26.9	1.034	−0.105	0.061	0.85	0.07
B10930281	7 July 2020	21:00	2.5	2.59	4.05	389	707	26.1	1.155	−0.212	0.085	0.68	0.03
B10930294	12 July 2020	21:00	0	0.94	3.49	114	535	22.8	0.618	−0.356	0.111	0.49	0.07
B10930325	14 July 2020	9:00	0	5.44	4.41	454	584	24.0	0.803	−0.089	0.048	0.49	0.08
B10930329	16 July 2020	9:00	20	20.2	17.39	5796	322	6.8	10.03	0.075	0.085	0.67	0.03
B10930330	16 July 2020	21:00	20	25.3	22.93	7815	331	6.2	13.05	0.035	0.085	0.90	0.08
B10930362	25 July 2020	21:00	1	2.30	2.77	223	562	12.3	0.344	−0.024	0.048	0.71	0.08
B10930364	26 July 2020	21:00	0	0.64	2.67	36	559	13.1	0.407	−0.149	0.085	0.53	0.03
B10930369	28 July 2020	9:00	5.5	1.51	3.91	360	577	18.4	0.900	−0.230	0.051	0.59	0.08
B10930371	29 July 2020	9:00	16	2.68	23.85	637	604	14.9	3.297	0.015	0.048	0.71	0.08
B10930375	31 July 2020	9:00	3.5	5.27	6.10	707	501	24.4	1.281	0.022	0.048	0.58	0.08
B10930378	1 August 2020	21:00	0	2.11	4.39	244	466	16.0	1.243	−0.091	0.048	0.54	0.08
B10930389	4 August 2020	9:00	103	12.0	20.40	5753	176	3.7	6.905	0.082	0.081	0.59	0.03
B10930390	4 August 2020	21:00	103	11.3	19.46	5680	176	3.5	6.745	0.093	0.081	0.54	0.08
B10930393	6 August 2020	9:00	4.5	4.16	5.39	1009	414	14.9	2.909	−0.008	0.085	0.49	0.08
B10930394	6 August 2020	21:00	4.5	4.70	5.45	992	421	14.9	2.673	−0.036	0.081	0.53	0.08
B10930395	7 August 2020	9:00	4.5	16.1	11.64	2189	473	25.0	4.752	0.145	0.048	0.61	0.08
B10930396	7 August 2020	21:00	4.5	15.3	11.69	2025	489	24.6	4.337	0.120	0.048	0.57	0.09
B10930402	10 August 2020	21:00	2	1.38	4.48	241	695	45.3	0.591	−0.030	0.085	0.46	0.08

Table 2. Fe and Mo isotope compositions and metal concentrations of Agongdian River water.

Sample ID	Sampling		Precipitation (mm)	Cr (ppb)	Cu (ppb)	Fe (ppb)	Sr (ppb)	Mo (ppb)	Pb (ppb)	d56/54Fe	2 s.d.	d98/95Mo	2 s.d.
	Date	Time
B10930197	29 June 2020	21:00	0	2.72	4.44	323	675	49.3	0.887	−0.080	0.081	0.64	0.08
B10930198	30 June 2020	9:00	31	12.5	5.05	509	853	45.1	1.301	−0.092	0.085	0.40	0.08
B10930202	2 July 2020	9:00	4	3.81	9.76	767	547	17.0	2.717	−0.060	0.051	1.07	0.08
B10930210	6 July 2020	21:00	0	3.35	4.78	492	641	26.9	1.034	−0.105	0.061	0.85	0.07
B10930281	7 July 2020	21:00	2.5	2.59	4.05	389	707	26.1	1.155	−0.212	0.085	0.68	0.03
B10930294	12 July 2020	21:00	0	0.94	3.49	114	535	22.8	0.618	−0.356	0.111	0.49	0.07
B10930325	14 July 2020	9:00	0	5.44	4.41	454	584	24.0	0.803	−0.089	0.048	0.49	0.08
B10930329	16 July 2020	9:00	20	20.2	17.39	5796	322	6.8	10.03	0.075	0.085	0.67	0.03
B10930330	16 July 2020	21:00	20	25.3	22.93	7815	331	6.2	13.05	0.035	0.085	0.90	0.08
B10930362	25 July 2020	21:00	1	2.30	2.77	223	562	12.3	0.344	−0.024	0.048	0.71	0.08
B10930364	26 July 2020	21:00	0	0.64	2.67	36	559	13.1	0.407	−0.149	0.085	0.53	0.03
B10930369	28 July 2020	9:00	5.5	1.51	3.91	360	577	18.4	0.900	−0.230	0.051	0.59	0.08
B10930371	29 July 2020	9:00	16	2.68	23.85	637	604	14.9	3.297	0.015	0.048	0.71	0.08
B10930375	31 July 2020	9:00	3.5	5.27	6.10	707	501	24.4	1.281	0.022	0.048	0.58	0.08
B10930378	1 August 2020	21:00	0	2.11	4.39	244	466	16.0	1.243	−0.091	0.048	0.54	0.08
B10930389	4 August 2020	9:00	103	12.0	20.40	5753	176	3.7	6.905	0.082	0.081	0.59	0.03
B10930390	4 August 2020	21:00	103	11.3	19.46	5680	176	3.5	6.745	0.093	0.081	0.54	0.08
B10930393	6 August 2020	9:00	4.5	4.16	5.39	1009	414	14.9	2.909	−0.008	0.085	0.49	0.08
B10930394	6 August 2020	21:00	4.5	4.70	5.45	992	421	14.9	2.673	−0.036	0.081	0.53	0.08
B10930395	7 August 2020	9:00	4.5	16.1	11.64	2189	473	25.0	4.752	0.145	0.048	0.61	0.08
B10930396	7 August 2020	21:00	4.5	15.3	11.69	2025	489	24.6	4.337	0.120	0.048	0.57	0.09
B10930402	10 August 2020	21:00	2	1.38	4.48	241	695	45.3	0.591	−0.030	0.085	0.46	0.08

The river water monitoring points of this study are set along the tributary of the Agongdian River, which flows through urban areas and flat plains for agriculture. Prior to entering the industrial zones, the upstream sources of this tributary consist mainly of agricultural drainage. Considering its purpose for drainage, the river channel has been subject to specialized management, resulting in relatively low sedimentation. The overall variations in the trace element contents in the river water derived from the sediments should also be insignificant. This is consistent with the negative correlation observed between Fe and Sr in the river water (Figure 3a). If the variations in these metal contents attributed directly to the sediments, the abundance of these two elements enriched in the continental crust should corelate positively in the river water.

The concentration variations caused by dilution effect resulting from precipitation also needs to be considered. During the monitoring period, the day with the highest precipitation was at 4 August 2020, and the concentrations of Cr, Cu, and Fe in the river water were elevated by 1–2 orders of magnitude on that day compared to non-rainy days. In addition, the highest abundance of these three elements was observed on the day with a 20 mm rainfall event (16 July 2020). These suggest that the variations in metal contents are not controlled sorely by the dilution effect.

In contrast to Figure 3a, Cr, Cu, and Pb concentration in Agongdian River water exhibits positive correlations with Fe concentration (Figure 3b–d). While Cr, Cu, Pb, and Fe are all at considerably high levels in the untreated industrial wastewater (Table 1), these correlations suggest a likely connection between industrial wastewater discharge and the observed fluctuations of these metal contents in Agongdian River water. This also provides an explanation for the minor variation in Sr in the river water. In this study, the concentration differences of Sr in industrial wastewater (approximately 60-fold, Table 1) were significantly lower than other metallic elements (3–5 orders of magnitude, Table 1). The Sr concentrations in industrial wastewater were comparable to or even lower than those in the river water, with some untreated wastewater (i.e., MS1, Table 1) exhibiting Sr concentrations much lower than the river water. Therefore, the discharge of industrial wastewater may have a negligible impact on Sr concentrations. Based on the earlier discussion, the metal concentrations in river water significantly increased on days with precipitation. While industrial wastewater emission was artificially controlled, the variations in metal abundances can be considered as the industrial signals overcomes the natural signals.

In order to compare with the industrial wastewater, the Fe and Mo isotopic compositions of the river water were also analyzed. The Fe isotope in Agongdian River water at the monitoring site varies from δ⁵⁶/⁵⁴Fe = −0.356 ± 0.111‰ to 0.145 ± 0.048‰, while the Mo isotope changes from δ⁹⁸/⁹⁵Mo = 0.40 ± 0.08‰ to 1.07 ± 0.08‰. Previous studies have indicated that most of the rivers in Taiwan exhibit relatively subtle changes in Mo and Fe isotopic compositions during wet and dry seasons in the absence of anthropogenic interference [39,59,60]. The observed Fe and Mo isotopic variations in Agongdian River water appear to be more than simple natural fluctuations over such a short period, and more likely due to human activities, particularly the discharge of industrial wastewater in the vicinity. The Fe isotopic compositions become heavier, while Cr, Fe, Cu, and Pb concentrations increase (Figure 4). As discussed in Section 3.1, the untreated wastewater contains more abundant transition metals and heavier Fe and lighter Mo isotopic compositions, while treated effluent contains the opposite signals (lower metal contents, and lighter Fe and heavier Mo isotopic ratios). The variations observed in Figure 4 are likely influenced by the mixing between upstream river water, untreated industrial wastewater, and treated industrial effluent.

3.3. Fe and Mo Stable Isotopes as Tracers

From Section 3.1, it is evident that the untreated wastewater exhibits heavier Fe and lighter Mo isotopic signatures. In addition, a positive correlation between Fe isotopes and transition metal concentrations was also observed in Agongdian River water. These suggest that the Fe and Mo isotopic ratios, along with metal concentrations, can be employed to assess whether the river is affected by untreated wastewater, but the dilution effect on element concentrations in river water must also be taken into consideration. In Section 3.2, it was observed that Sr concentrations in the river water of the monitoring site remained relatively stable (Table 2) and showed no correlation with the elements associated with industrial pollution (i.e., Fe and Cr). Therefore, normalizing Fe and Cr abundances to Sr contents may mitigate the dilution effect.

Diagrams of δ⁵⁶/⁵⁴Fe versus Fe/Sr ratios and δ⁹⁸/⁹⁵Mo versus Cr/Sr ratios are plotted in Figure 5 to investigate whether the signals of untreated wastewater were in Agongdian River water. Based on previous report, the Fe/Sr and Cr/Sr ratios in the upstream fall within the ranges of 0.21–1.42 and 0.0008–0.0084, respectively [61]. Therefore, the river water samples with Fe/Sr and Cr/Sr ratios in these ranges would be considered comparable to upstream signal (Figure 5). By analyzing the variations in isotopic compositions and metal concentrations of the river water in conjunction with known industrial wastewater signatures (Section 3.1), patterns of pollution affecting the Agongdian River at the monitoring site could be discerned. When the Fe/Sr ratio decreases alongside a decrease in δ⁵⁶/⁵⁴Fe, or when the Cr/Sr ratio does not change while δ⁹⁸/⁹⁵Mo become heavier, it indicates that the pollution source is dominated by treated industrial effluent (Figure 5). Conversely, if the Fe/Sr ratio increases alongside a shift towards heavier δ⁵⁶/⁵⁴Fe, or if the Cr/Sr ratio increases without significant changes in δ⁹⁸/⁹⁵Mo, it suggests the potential influx of untreated industrial wastewater into the river (Figure 5). Following these principles, the dates significantly affected by the untreated wastewater were highlighted in Figure 5. It is noteworthy that not only the dates correspond to both in Fe and Mo isotope plots, but the level of contamination is also approximately consistent. Water samples collected on 16th of July and 4th of August emerge as the two most pronounced periods of untreated wastewater signals, with 7th of August following closely. Although samples collected on 31st of July and 6th of August exhibit less intense untreated wastewater signals, they remain distinguishable. The monitoring results on 30th of July demonstrate a more obvious untreated wastewater signal in Figure 5b than in 5a, possibly due to a particular source with higher Cr content compared to Fe. Using the isotopic composition and concentration data analyzed in this study, it is possible to further estimate the quantity of untreated wastewater that was potentially discharged into the Agongdian River (

Table 3. Estimation of discharged untreated wastewater.

Date	Estimated Discharged Untreated Wastewater
	Minimum (m3/day)	Maximum (m3/day)
30 June 2020	129	285
16 July 2020	1117	1561
31 July 2020	44	200
4 August 2020	1154	1391
6 August 2020	35	214
7 August 2020	471	683

). Based on the estimations, during the monitoring period of this study, there may have been between 2950 and 4334 m³ of untreated industrial wastewater discharged into Agongdian River.

4. Conclusions

Iron and molybdenum isotopes were employed to study the potential pollutions of rivers from the discharging of industrial wastewater. This is the first study demonstrating that Fe and Mo isotope systems could be applied to differentiate between treated and untreated sewage and the extent of industrial wastewater contamination in natural river bodies.

I. In Kangshan Benjhou Industrial Park, treated industrial effluent exhibited significantly reduced metal contents, and also displayed a notable shift towards a lighter δ⁵⁶/⁵⁴Fe. Additionally, the δ⁹⁸/⁹⁵Mo in treated effluent showed a substantial increase in heavy Mo isotopes, and the degree of increase positively correlated with the proportion of metallic element removal.

II. Agongdian River water exhibited a more than two orders of fluctuation in Fe concentration during the monitoring period, and the variation was independent of local precipitation. Significant variations in Cr, Cu, and Pb concentrations in river water were also observed, and these variations were positively correlated with the Fe concentration. Furthermore, the Cr, Fe, Cu, and Pb abundances in the river water were also positively correlated with δ⁵⁶/⁵⁴Fe. Comparing these results with that of industrial wastewater clearly demonstrated the varying degrees of pollution in the Agongdian River.

III. The Fe/Sr vs. δ⁵⁶/⁵⁴Fe and Cr/Sr vs. δ⁹⁸/⁹⁵Mo plots of monitored river water clearly demonstrated that untreated wastewater and treated industrial effluent represent two distinct end-members of contamination sources for the Agongdian River. This is useful for identifying the degree and timing of contamination from untreated wastewater. It appeared that between 2950 and 4334 m³ of untreated industrial wastewater entered the natural water bodies during the monitoring duration (43 days).

These results show the effectiveness of non-traditional stable isotopes in providing valuable insights into water resource management. Employing this technology to other types of wastewaters, i.e., urban, technology industry, agriculture, or livestock industry wastewater, would be the future works of the authors of this study.