Transport of bisphenol A, bisphenol S, and three bisphenol F isomers in saturated soils

With the limitation of the use of bisphenol A (BPA), the production of its substitutes, bisphenol S (BPS), and bisphenol F (4,4’-BPF) is increasing. Understanding the fate and transport of BPA and its substitutes in porous media can help reduce their risk of contaminating soil and groundwater systems. In this study, column and batch adsorption experiments were performed with 14C-labeled bisphenol analogs and combined with mathematical models to investigate the interaction of BPA, BPS, 4,4’-BPF, 2,2’-BPF, and 2,4’-BPF with four standard soils with different soil organic matter (SOM) contents. The results show that the transport capacity of BPS and 4,4’-BPF in the saturated soils is significantly stronger than that of BPA. Meanwhile, the mobility of the three isomers of bisphenol F exhibits variability in saturated soils with high SOM content. The two-site nonequilibrium sorption model was applied to simulate and interpret column experimental data, and model simulations described the interactions between the bisphenol analogs and soil very well. The fitting results underscore SOM’s role in providing dynamic adsorption sites for bisphenol analogs. Hydrophobicity primarily accounts for the disparity in adsorption affinity between BPA, BPS, 4,4’-BPF, and soil, whereas hydrogen bonding forces may predominantly influence the differential adsorption affinity between 4,4’-BPF and its isomers and soil. The results of this study indicate that BPS and three isomers of BPF, as alternatives to BPA, have higher mobility in saturated soils and may pose a substantial risk to groundwater quality. This study enhances our understanding of bisphenol analogs’ behavior in natural soils, facilitating an assessment of their environmental implications, particularly regarding groundwater contamination.


Introduction
2,2-Bis(4-hydroxyphenyl) propane (BPA) is a derivative of phenol and acetone (Rochester 2013;Staples et al. 1998).The structural rigidity inherent to BPA enables it to serve as a foundational scaffold within polymer systems, profoundly influencing their thermo-mechanical properties (Baluka and Rumbeiha 2016;Manzoor et al. 2022).BPA is widely used to make polycarbonate plastics and epoxy liners for food and beverage containers.In addition to this, BPA is used as a developer for thermal paper, such as paper receipts and airport baggage tags (Kataria et al. 2022).With annual BPA production reaching 5 million tons in 2015, BPA stands as one of the most produced chemicals globally within industrial sectors (Corrales et al. 2015).However, studies have found that BPA is an endocrine disruptor (EDC), which disrupts the endocrine system equilibrium in animals and humans (Akash et al. 2020;Saal et al. 2012).Many countries have started restrictions on BPA usage, and emphasizing the urgency of identifying and adopting BPA alternatives Responsible Editor: Kitae Baek Highlights • BPS and 4,4'-BPF have higher mobilities than BPA in four standard soils.
• The adsorption affinity of bisphenol analogs to soil is significantly correlated with soil organic matter.(Catenza et al. 2021;Chen et al. 2016;Rochester 2013).Among the BPA alternatives, bisphenol S (BPS) and bisphenol F (BPF) are the prominent alternatives, with annual production or imports reported 1000-10,000 tons and 10,000-100,000 tons, respectively, in Europe, as reported by the ECHA (Chen et al. 2016).They have a similar structure to BPA and both have a wide range of commercial applications.BPS, with its close resemblance to BPA (Table 2), contains a sulfone group with strong electron absorbing ability and two hydroxyl groups, rendering it more stable than BPA (Hu et al. 2019;Liu et al. 2021).It predominantly substitutes for BPA in thermal paper production and serves as a component of phenolic resins.Similarly, BPF, resembling BPA in having two phenol rings, is connected via a methylene bridge and primarily contributes to the production of polycarbonate plastics and resins (Hu et al. 2019).BPF used in industry usually refers to 4,4'-BPF; however, the production of two isomers, 2,2'-BPF and 2,4'-BPF, as side products, is inevitable.Nevertheless, limited attention has been dedicated to 2,2'-BPF and 2,4'-BPF (Guo et al. 2019;Sun et al. 2017).
The extensive utilization of bisphenol analogs has led to their widespread occurrence across various environmental media (Gao et al. 2023;Maturi et al. 2023;Qian et al. 2021;Xue et al. 2016), with soils representing a notable reservoir for these compounds (Xu et al. 2021).Bisphenol analogs in the soil can enter the aqueous environment through a variety of pathways, such as surface runoff, drainage flow, and groundwater flow (Dueñas-Moreno et al. 2022).According to Ying and Kookana (2003), the degradation of BPA in aerobic conditions within seawater exhibits limited progress within 35 days, followed by rapid degradation in the subsequent week.Conversely, due to unfavorable redox and degradation conditions in the groundwater, contaminants degrade very slowly (Yamazaki et al. 2015;Zhi et al. 2019).BPS and BPF are currently not regulated and can be used without restriction (Soria et al. 2015).However, due to their similar structures, studies have found comparable anti-androgenic effects between BPS, 4,4'-BPF, and BPA (Gogola-Mruk et al. 2023;Fol et al. 2017;Wang et al. 2021).Consequently, once the contaminants reach the groundwater, the risk to human health is high.Therefore, understanding the mechanisms that control the transport of BPA and its substitutes through soil is essential for developing management practices to protect soil and groundwater.
Numerous studies have highlighted the high mobility of BPA, a polar contaminant, with the potential to traverse soil profiles and infiltrate groundwater.For example, Dai (2020a) reported BPA's high mobility in porous media, both under saturated and unsaturated conditions.Notably, the presence of a gas phase in soil, as occurs in unsaturated conditions, diminishes BPA mobility due to pronounced adsorption, driven by hydrophobic interactions, at the air-water interface (AWI).Additionally, BPA's transport in soil, under both saturated and unsaturated conditions, exhibits a negative correlation with soil organic matter (SOM) content.BPA is adsorbed onto SOM through a combination of hydrophobic interactions and hydrogen bonding, which attenuates its mobility in soil (Guo et al. 2022;Zakari et al. 2016).Likewise, BPS, as an alternative to BPA, appears to possess greater mobility, as reported in previous studies (Shi et al. 2018(Shi et al. , 2019)).Adsorption plays a key role in governing the transport and fate of organic pollutants across various environments.BPA appears to have a stronger affinity for soil adsorption when in a nondissociated state (pH < pK a ) due to its greater hydrophobicity than BPS (Choi and Lee 2017).Consequently, BPA tends to accumulate in the soil, while BPS exhibits a proclivity to migrate into groundwater.Current studies on the fate and transport of bisphenol analogs in porous media have focused mainly on BPA and BPS, and little attention seems to have been paid to BPF.However, surveys have revealed that the concentration of BPF in Japanese, Korean, and Chinese waters seems to be one to two orders of magnitude higher than those of BPA (Qian et al. 2021;Yamazaki et al. 2015).BPA was detected in water collected near a manufacturing plant in Jiaxing City, Zhejiang Province (China), at a concentration of 0.28 ug/L, while BPF was detected at concentrations up to 15.3 ug/L (Chen et al. 2016;Wang et al. 2022).Current research in BPF seems to be limited to the development of new acid catalysts and optimization of BPF isomer product distribution (Sun et al. 2017;Wang et al. 2016).Only a handful of studies have delved into BPF retention and distribution, predominantly concentrating on 4,4'-BPF (Liu et al. 2018;Wu et al. 2019).For example, due to the stronger hydrophobicity of BPA, 4'-BPF exhibited lower adsorption onto polyvinyl chloride (Wu et al. 2019).Furthermore, 4,4'-BPF, as a polar organic pollutant, has high mobility in saturated soils, and there is no resolution hysteresis in the co-transport of microplastics (MPS) with 4,4'-BPF.Importantly, MPS exerts no discernible influence on the transport of 4,4'-BPF (Liu et al. 2018).There is a gap in studies on the adsorption and transport of 2,2'-BPF and 2,4'-BPF in soils.Moreover, the transport of bisphenol analogs in different types of soils thus might differ; this has not been tested.
The overarching goal of this study is to advance the current understanding of the fate and transport of bisphenol analogs in soil.The objectives are as follows: (1) to compare the transport capacity of BPA and its substitutes, BPS and 4,4'-BPF in different soils, and to reveal the mechanism of interaction with soil.(2) The transport capacity of 4,4'-BPF and its isomers 2,2'-BPF and 2,4'-BPF in different soils was compared, and the interaction mechanism with soil was revealed.(3) To establish and test mathematical models for fate and transport of bisphenol analogizes in porous media.

Materials
Lufa soils were purchased from Lufa Speyer, Germany (www.lufa-speyer.de).Lufa soils are harvested from natural soils of common soil types in selected areas of Germany.They are not applied with pesticides, fungicidal fertilizers, or organic fertilizers for at least 5 years under agricultural use.Mineral fertilizers were not used more than 3 months before sampling.The soils were usually sampled from a depth of 0-20 cm, prepared and sieved with a 2-mm sieve.Standard soils have been widely used for hazard assessment of soil contaminants, combining representativeness with ecological relevance for accurate risk assessment (Bastos et al. 2014;Garcia-Velasco et al. 2017;Mocova et al. 2022;Turner et al. 2020).The specific properties of the soil are shown in Table 1.BPA, BPS, 4,4'-BPF, 2,4'-BPF, and 2,2'-BPF (with purity > 99%) were purchased from Sigma-Aldrich (St. Louis, MO).The physicochemical characteristics of the bisphenol analogs are shown in Table 2.

Determination of pollutant concentration
14 C-labeled BPA (0.74 GBq/mmol), BPS (2.40 GBq/mmol), 2,2'-BPF (2.82 GBq/mmol), 2,4'-BPF (2.82 GBq/mmol), and 4,4'-BPF (2.82 GBq/mmol) were synthesized using 14C-labeled phenol as a precursor.Nonlabeled BPA, BPS, 2,2'-BPF, 2,4'-BPF, and BPF (all with purity > 99%) were purchased from Sigma-Aldrich (St. Louis, MO).By mixing the labeled and unlabeled contaminants, we obtained contaminant stock solutions with high response values that can be diluted for specific experimental applications.The amount of contaminants was calculated by quantitative determination of radioactivity using a liquid scintillation counter (LS6500; Beckman Coulter, USA).The instrument settings were the following: The energy scale was 0 ~ 6700 keV, the 2 σ value was 1%, and the counting time was 5 min.The control and standard radioactivity samples were provided by the company.The calibration curve was obtained using the radioactivity standards.About 0.5 ~ 5 mL of experiment samples were mixed with 10 mL of scintillation cocktail (Gold Star multipurpose; Meridian Biotechnologies  Ltd., UK).The samples had a background of < 0.5 Bq and no chemiluminescence due to the cocktails was observed.By controlling the radioactivity of the stock solution, liquid scintillation counter was able to detect the contaminants in this experiment within the range of 0.05 to 100 ug/L.The error range is within 0.05 ug/L.

Column transport experiments
Filling of the Lufa soil columns followed the same procedure described in our previous study (Liu et al. 2019).Specifically, column experiments were performed in a 10-cm long borosilicate glass column apparatus with a core diameter of 0.66 cm.Air-dried, screened Lufa soil was uniformly filled in the borosilicate glass column.Carbon dioxide gas (purity > 99%) was introduced into the soil column from the lower part of the end soil column for about 1 h (Liu et al. 2018;Yang et al. 2015).The residual air in the column was exhausted.Then, deionized (DI) water was pumped from the bottom of the soil column to the top to flush the filling medium.Subsequently, approximately 200 mL of background electrolyte (0.5 mM NaCl solution) was injected from bottom to top by an injection pump to flush the soil column to equilibrium (Liu et al. 2018).When the soil column is equilibrated by long-term saturation with the background solution, experimental solutions containing NaCl (0.5 mM) and the target bisphenol analogs (10 µg/L each of BPA, BPS, 4,4'-BPF, 2,2'-BPF, and 2,4'-BPF) were injected into the columns at a flow rate of 10 m/day with a constant dasher.An effluent sample is collected every 2-3 well volumes (PV) (Ren et al. 2021;Yang et al. 2015).After collecting 60 PV (contaminants reached migration equilibrium in the soil column), the column was flushed with background solution (free of contaminants) and an effluent sample was collected every 2-3 PV until no contaminants were detected in the effluent.The effluent samples were measured with a liquid scintillation counter (Liu et al. 2019).The basic parameters of the soil column in the test are shown in Table 3.

Two-site nonequilibrium sorption model
To further better understand the transport mechanism of bisphenol analogs in the soil column, we used a two-site nonequilibrium sorption model to fit the breakthrough curves (BTCs) of bisphenol analogs (Nkedi-Kizza et al. 1984).The model divides the adsorption of bisphenol analogs that occurs during transport in the soil column into two parts, assuming that adsorption on the type 1 site is transient and follows the equilibrium isotherm, while adsorption on the type 2 site is a first-order kinetic process (Avila and Breiter 2009).The two-site nonequilibrium adsorption model has gained wide where S e (µg/g) is the adsorbed amount at the equilibrium adsorption site.S k (µg/g) is the adsorbed amount at the kinetic adsorption site.We combined the results of batch adsorption experiments that the adsorption of bisphenol analogs on soil obeyed the linear isothermal adsorption characteristics (Table 4).It is assumed that the adsorption of bisphenol analogs on soil during migration obeys the linear isothermal adsorption characteristics (Eq.3): where f is the proportion of instantaneous equilibrium adsorption sites; ω (min −1 ) is the first-order adsorption rate coefficient; and K d (cm 3 /g) is the solid-liquid partition coefficient. ( where R d is the delay factor; β represents the fraction of the instantaneous equilibrium adsorption sites in all the adsorption sites.The tracer breakthrough curves were fitted using the Hydrus-1D software (version 4.16.0110,PC-progress, Czech Republic/USA) to fit the tracer breakthrough curves, inversely solving for the dispersion (λ), and the dispersion coefficient (D = λv) was calculated, and fit the breakthrough curves of bisphenol analogs, inversely solving for the model parameters K d , f, and ω, from which R d , β. (5)

Batch adsorption experiments
The adsorption experiment of organic pollutants on Lufa soil was carried out by the adsorption experiment method used in previous literature (Choi and Lee 2017;Dai et al. 2020b).First, add approximately 0.5 g of soil to a series 20-mL brown glass vial; the use of brown glass vials is effective in preventing photodegradation of bisphenol analogs (Liu et al. 2018).Each target chemical was dissolved in sterile 0.5 mM NaCl solution to achieve concentration of 3, 5, 10, 20, 30, 40, and 50 μg/L.After that, different amounts of 14 C of labeled organic pollutant reserve liquid were carefully injected into the glass bottle with a micro syringe.The bottle was then sealed with a polytetrafluoroethylene pad and cap and placed on a rotating disk in a darkroom that ran continuously for 7 days at a rate of three cycles per minute to equalize the adsorption of organic pollutants on the soil.The concentration of bisphenol analogs in solution was determined when the Lufa soils reached adsorption equilibrium for the bisphenol analogs (Choi and Lee 2017;Guo et al. 2022).
Freundlich model and linear model were used to fit the adsorption isotherm of bisphenol analogs in Lufa soil.The calculation formulas of the two models are as follows: Freundlich model (FM): Linear model (LM): where q e (mg/kg) is the adsorption state concentration of organic pollutants after adsorption equilibrium, C e (mg/L) is the aqueous phase concentration of organic pollutants after adsorption equilibrium, K F ((mg/kg)/(mg/L)) is the Freundlich adsorption coefficient, n (unitless) is the Freundlich index, and K d (L/kg) is the Lufa water partition coefficient.

Transport of BPA and its substitutes in saturated Lufa soils
In all four typical saturated soil media, the maximum penetration rates (C/C 0 ) of the BPA alternatives exceeded that of BPA, indicating that BPS and BPF exhibit higher mobility than BPA. Figure 1 shows the measurement and fitting breakthrough curves (BTCs) of BPA, BPS, and 4,4'-BPF columns filled with four different types of standard soils under saturated flow conditions.The results of the column experiments revealed that the breakthrough concentrations of BPS and BPF increased rapidly in Lufa a, Lufa b, and ( 11) Lufa d, reaching their peak earlier than that of BPA.And BPA exhibited a significant trailing phenomenon in all four Lufa soils.In addition, the maximum penetration rates (C/C 0 ) of BPS and 4,4'-BPF in four typical soils ranged from 89.3 to 101.3%, while for BPA, the C/C 0 in these four soils were 51.6-77.5%.Furthermore, the differences in mobility of three bisphenol analogs in these four typical soils appeared to become more pronounced with increasing SOM content.The SOM content of four Lufa soils follows the order of Lufa b (2.94%) > Lufa c (1.84%) > Lufa a (1.23%) > Lufa d (1.01%).Notably, the higher SOM content in Lufa b tends to amplify the distinctions in mobility among BPA and its substitutes.In Lufa b, the soil with the most organic matter, BPA had the strongest affinity for binding to the soil SOM, which gave it the lowest maximum penetration rate (51.6%) among three bisphenol analogs.Following closely was BPF, which had a maximum penetration rate of 76% in Lufa b.The maximum penetration of BPA in the four Lufa b soils was 51.6%.In contrast, BPS displayed high mobility in all four Lufa soils, with a maximum penetration rate of 90.8% in Lufa b.The effect of SOM content on the transport of bisphenol analogs in porous media has been reported in other studies (Chen et al. 2022;Dai et al. 2020a;Shi et al. 2019).For example, in a study by Shi et al. (2019), BPA and BPS had low mobility in soils with high SOM content, but their transport in the sand column was significantly enhanced after the removal of SOM from the soil.
To further understand the transport of bisphenol analogs, batch adsorption experiments were performed under the same solution conditions as the column experiments.The results of batch adsorption experiment effectively elucidate the observed phenomena in the aforementioned column experiments.The adsorption affinity of five bisphenol analogs with four typical soils was quantified by batch isothermal experiments (Fig. 2).The fitting of the adsorption models revealed that the Freundlich model, with correlation coefficients (R 2 ) exceeding 0.98, provided an excellent fit for the adsorption isotherm data of bisphenol analogs on Lufa soils.The Freundlich model fitting yielded values of 1/n ranging from 0.75 to 1.25, indicating a dominance of linear distribution (Dai et al. 2020a).This result suggests that the concentrations of bisphenol analogs used in the experiments were not high enough to capture nonlinear adsorption (i.e., multilayer adsorption).According to relevant studies, it was found that the SOM content in soil affects the adsorption of bisphenol analogs in soil, and the SOM content in soil is positively correlated with the adsorption of bisphenol analogs in soil (Ahmed et al. 2014;Choi and Lee 2017;Li et al. 2016).The K f value reflects the adsorption affinity of bisphenol analogs to Lufa soil.However, as we employed four different soils in the batch adsorption experiments, it is important to note that different soils typically yield different n values, making direct comparisons of K f values between different soils impractical.Therefore, we used the approximate linear distribution coefficient (K d ) of the adsorption isotherm to compare the adsorption affinity of bisphenol analogs to soil.In the case of BPA, BPS, and 4,4'-BPF within the four Lufa soils, the K d values followed the order BPA > 4,4'-BPF > BPS.This suggests that the difference in sorption affinity of BPA, BPS, and 4,4'-BPF in the four Lufa soils represents the primary factor contributing to differences in their mobility.
The adsorption of bisphenol analogs onto soils is affected by lots of factors.The properties of bisphenol (e.g., molecular structure, hydrophobicity, polarity, and spatial configuration) analogs are the main factor influencing their adsorption behavior in soil.Log K ow is considered as a parameter to assess the hydrophobicity of organic chemicals (Christensen et al. 2022).Notably, the Log K ow values of the three bisphenol analogs were ranked in the order of BPA > 4,4'-BPF > BPS.This was in accordance with their K d values.This seems to indicate that the adsorption capacity of bisphenol analogs is positively correlated with their hydrophobicity.The importance of hydrophobic interactions for the adsorption of bisphenol analogs has been reported in prior studies (Choi and Lee 2017;Guo et al. 2022;Wu et al. 2019).For example, in a study by Wu et al. (2019), the adsorption of BPA, BPS, BPB, and BPAF on PVC was found to have a good linear correlation with their hydrophobicity.Moreover, Shi et al. (2018) demonstrated that BPA possessed greater hydrophobicity compared to BPS, rendering it more prone to adsorption onto limestone porous media.In addition to hydrophobic interactions, electrostatic and hydrogen bonding forces also affect the adsorption of bisphenol analogs by soil (Choi and Lee 2017).Previous studies have shown that bisphenol analogs become negatively charged through dissociation when pH > pK a and as soils typically bear a negative charge; electrostatic repulsion diminishes the adsorption affinity of bisphenol analogs to soil (Choi and Lee 2017).In our column experiments, the pH levels ranged from 6.7 to 7 (Table 3), which is lower than the pK a of bisphenol analogs, suggesting that they existed in a non-dissociated state.Consequently, electrostatic forces did not significantly influence the transport of bisphenol analogs.Interestingly, hydrogen bonding seems to affect the adsorption of bisphenol A by soil, attributed to BPA's organic nature featuring hydrophilic hydroxyl groups (Pan et al. 2017), rendering it more prone to binding with SOM through hydrogen bonding (Dai et al. 2020a, b).As alternatives to BPA, BPS and 4,4'-BPF also feature hydroxyl groups, making hydrogen bonding forces an important determinant in the transport of bisphenol analogs.Notably, DFT calculations have revealed that the reactivity of the BPS molecule surpasses that of the BPA molecule, both through electrophilic reactions via the sulfonyl oxygen on BPS and nucleophilic reactions via the hydroxyl hydrogen on BPS (Guo et al. 2022).However, among the four soils, the adsorption affinity of BPA and its isomers was consistently higher than that of BPA substitutes.Therefore, we infer that differences in hydrophobicity might be the primary driver of variations in the transport capacities of bisphenol analogs in the four standard soils.The two-site nonequilibrium sorption model has proven to be a valuable tool for estimating the transport properties of contaminants in saturated porous media (Cao et al. 2023;Wikiniyadhanee et al. 2015;Zakari et al. 2016).In this study, the model well simulated the BTCs of bisphenol A and its substitutes (R 2 > 0.95).The corresponding fitting parameters are detailed in Table 5.It is noteworthy that the K d values calculated from the column experimental fit were marginally smaller than those obtained from batch adsorption experiments, although the overall trend remained consistent (Table 4).This disparity is reasonable as the K d calculated during the intermittent experiment represents equilibrium adsorption, while the K d calculated during transport signifies nonequilibrium adsorption (Ge et al. 2018;Wikiniyadhanee et al. 2015).The fitted delay factor (R d ) is an applicable tool for assessing the transport of BPA and its analogs in terrestrial soil-filled columns, where a higher R d value indicates a more delayed transport of the contaminant in the soil column (Nkedi-Kizza et al. 1984).The fitting results also showed that the order of R d values in the four soils was BPA > 4,4'-BPF > BPS, indicating that BPA is less mobile than its alternatives in all four soils (Fig. 3).The parameter f denotes the ratio of the adsorption at the instantaneous adsorption site to the total adsorption (Chen et al. 2022;Dai et al. 2020b).The fitting results showed that the f values of BPA (0.11-0.23) were smaller than those of BPS (0.47-0.99) and 4,4'-BPF (0.35-0.68) in all four soils.This suggests that more dynamic adsorption of BPA occurs during transport.Previous studies have  demonstrated that SOM would provide more dynamic adsorption sites for pharmaceuticals and personal care products (PPCPs) (Chen et al. 2016;Dai et al. 2020a).This aligns with our findings where the lowest f-values of bisphenol analogs were reached in Lufa b, characterized by the highest organic matter content, indicating more dynamic adsorption.Furthermore, the equilibrium pointsite adsorption rate constant β of BPA was smaller than that of BPS and 4,4'-BPF, implying that the transport of all three bisphenol analogs was influenced by soil SOM content, with greater SOM content resulting in increased adsorption of bisphenol analogs at dynamic adsorption sites (Cao et al. 2023).Moreover, BPA had the strongest binding capacity to SOM, which led to its weaker transport than BPS and 4,4'-BPF.
Interestingly, the results of the batch adsorption experiments do not seem to align well with the results of the column experiments.Examination of the adsorption batch isotherms (Fig. 5) and the adsorption data (Table 4) reveals that 2,4'-BPF and 2,2'-BPF exhibit greater adsorption affinity for soil Lufa a, Lufa c, and Lufa d compared to 4,4'-BPF.In Lufa b, this trend was also maintained in the low concentration range (3-20 μg/L), where the bisphenol analogs used in the column experiments (10 μg/L) were located (Fig. 5b).This starkly contrasts with the results of the column experiments where differences in the transport of the three BPF isomers in soil Lufa a, Lufa c, and Lufa d were not significant.Pan et al. (2017) demonstrated that the strength of hydrogen bonds between the three isomers of BPF follows the order 2,2'-BPF > 2,4'-BPF > 4,4'-BPF.The SOM content of all three Lufa soils was low, and the batch sorption experimental procedure lasted longer compared to the column experimental procedure, thus allowing the three isomers of BPF more time to engage with these small amounts of SOM.Due to their ability to form stronger Fig. 5 Adsorption isotherms of 2,2'-BPF, 2,4'-BPF, and 4,4'-BPF to Lufa soils.The insets are drawn at a smaller scale to show differences in the adsorption affinity of pollutants to soil at low concentrations hydrogen bonds with SOM and bind more tightly to the soil, 2,2'-BPF and 2,4'-BPF exhibited a stronger affinity for soil adsorption than 4,4'-BPF.Based on the results of bulk sorption experiments and column experiments, it was shown that the three isomers of BPF differ in their sorption affinity for soil, and this difference is likely to become more significant as the SOM content increases.This may lead to their different fates and transport in the soil.
The fitting results of the two-site nonequilibrium sorption model showed that among the four soils, for BPF and its isomers, 2,2'-BPF and 2,4'-BPF, had smaller R d values than 4,4'-BPF (Fig. 6).This indicated that 4,4'-BPF undergoes more adsorption during transport (Xing et al. 2021;Zou and Zheng 2013).In the four Lufa soils, the transient adsorption coefficient β of 4,4'-BPF was smaller than that of 2,2'-BPF and 2,4'-BPF, and it occurred more dynamic adsorption (Cao et al. 2023).The most significant difference in β values of the three BPF isomers was found in Lufa b, which has the highest organic matter content.This can be due to the fact that SOM provides more dynamic adsorption sites (Dai et al. 2020a).In the fitted model, the dynamic adsorption sites were consistent with first-order kinetic adsorption, affirming that 4,4'-BPF hydrogen bonds were more easily broken, and 4,4'-BPF binds to SOM through hydrogen bonding interactions during migration (Guo et al. 2019;Pan et al. 2017).This leads to increased adsorption of 4,4'-BPF at the dynamically adsorbed sites, potentially accounting for its lower mobility than 2,2'-BPF and 2,4'-BPF.

Conclusion
Our study, encompassing batch adsorption tests, column migration experiments, and mathematical modeling, has deepened our understanding of bisphenol analog behavior in natural soils.Utilizing 14 C-labeled bisphenol analogs in conjunction with standard soils for column and batch experiments has greatly enhanced the precision and reproducibility of experimental data.Collectively, all experimental and model fitting data consistently indicate that the transport capacity of BPA was higher than that of its substitutes BPS and 4,4'-BPF in the four different standard soils.Furthermore, our findings suggest that the mobility of the three isomers of BPF may exhibit variations with increasing soil SOM content, underscoring the significance of SOM as a critical factor influencing the fate and transport of these compounds.The effective application of the two-site kinetic model in describing the experimental breakthrough curves (BTCs) of bisphenol analogs signifies its utility as a predictive tool for assessing the fate and migration of bisphenol analogs in soil.Our results emphasize the environmental risk posed by BPA alternatives, especially BPS and 4,4'-BPF, which exhibit a greater potential for groundwater contamination.Furthermore, the heightened mobility of 2,2'-BPF and 2,4'-BPF, by-products of BPF production, warrants attention in environmental assessments.This study provides important insights into the fate and transport of bisphenol analogs in natural soils and helps to assess their potential environmental risks.

Fig. 1
Fig.1BTCs of BPA, BPS, and 4,4'-BPF in saturated Lufa soils.Symbols are experimental data, and the lines are the fitted breakthrough curves using HYDRUS-1D

Table 1
Properties of Lufa soils