The simultaneous removal of bisphenol A, manganese and ammonium from groundwater by MeO_x: the role of chemical catalytic oxidation for bisphenol A Open Access

At present, the pollution problem of organic micro-pollutants (OMPs) in drinking water sources is becoming more and more serious, which has attracted widespread attention. In the production of cosmetics and plastic products, bisphenol A (BPA) as a typical OMP is severely overused. Due to the poor degradability, stable chemical properties and low concentration of BPA in water, it has become one of the most common and harmful OMPs in the water environment (Chen et al. 2016; Wang et al. 2019; Wang et al. 2020). BPA may damage the endocrine, reproductive and nervous systems in the human body, and long-term exposure to BPA may cause cancer (Moreira et al. 2019). BPA has been found in some rivers, lakes, surface water, and even tap water in China (Nie et al. 2015; Bai & Acharya 2019; Yang et al. 2020). Unfortunately, BPA cannot be effectively removed by the traditional water treatment processes in water treatment plants. The maximum contaminant level for BPA of 0.01 mg/L has been strictly required for tap water in China.

BPA can be efficiently removed by chemical catalytic oxidation, chemical adsorption, and photocatalytic and biological methods. Under certain pH conditions (3.0–9.0), the removal efficiency of BPA (20 mg/L in influent) reached almost 100% with a new type of catalyst Co₃O₄-Bi₂O₃ (0.30 g/L) in chemical catalytic oxidation, and the BPA was removed mainly by the surface-bound SO4^•− radicals (Hu et al. 2017). Mg-Al-layer double hydroxide had a strong adsorption capacity (about 216.50 mg/g) for BPA in chemical adsorption (Zhong et al. 2020). Cu_0.5Mn_0.5Fe₂O₄ was used to remove BPA under UV light irradiation, and it exhibited high photocatalytic activity for BPA (the removal efficiency exceeding 95.2%) (Yang et al. 2020). Bacillus thuringiensis was used to remove BPA by metabolic and proteomic approaches, and degradation efficiency was up to 85.0% (Li et al. 2018). In general, the chemical catalytic oxidation method is simple and low cost, but it is not easy to be generally used in traditional water treatment due to the complex preparation condition of the catalyst. The chemical adsorption method is efficient, but it is easy to cause secondary pollution after adsorption saturation. Although the biological method is widely used to remove BPA, it is affected by low-temperature conditions and long start-up periods. Even though the photocatalytic method has high removal efficiency for BPA, the high treatment cost makes it difficult to apply to actual water treatment plants. The removal of BPA is limited by many influencing factors, such as initial concentration of BPA, temperature and pH (Balgooyen et al. 2017; Huang et al. 2017).

Due to the high oxidation–reduction potential, manganese oxides were used to remove some metal ions (Fe(II), Mn(II) and As(V)) and OMPs (phenolic compounds and BPA) (Jiang et al. 2017; Wang et al. 2018; Jung et al. 2019). In our previous studies, potassium permanganate was used to continuously oxidize Mn(II) and Fe(II) in the influent, and the quartz sand surface quickly formed an iron and manganese oxide filter film (MeO_x) in a filter column (Guo et al. 2017). Moreover, the chemical catalytic oxidation process has some obvious advantages, such as higher efficiency, lower cost of water production, shorter start-up period (about 17 days), and stronger adaptability (Cheng et al. 2017; Guo et al. 2019).

The purpose of this study was to clarify the oxidation of BPA on conventional sand filters (MeO_x), the influence of NH₄⁺ on the removal of BPA and the simultaneous removal of Mn²⁺, NH₄⁺ and BPA from groundwater by the MeO_x. Two filter columns were accomplished and used for our experimental research, and the quartz sand with iron–manganese oxide film (MeO_x) as the filter material was packed in the filter column. During the experiment, the changes in the surface properties of the oxide film were explored, and the catalytic oxidation process of BPA was proposed.

The raw water was sampled from a groundwater well in northwest China, and the water quality is shown in Table 1. From Table 1, the concentrations of Mn²⁺ and NH₄⁺ were significantly lower than the Groundwater Quality Standards, and BPA was not found in the raw water.

The schematic of the experimental filter column is shown in Figure 1. From Figure 1, two filter columns were used in parallel. This system mainly included two filter columns which consist of plexiglass tube (internal diameter = 100 mm, height = 3,000 mm), dosing pumps and water distribution system. The raw water was added into the two columns by a booster pump. After the raw water fell into the filter column, the concentration of DO was increased to about 8.5 mg/L by drop aeration. The quartz sand with iron–manganese co-oxide filter film (MeO_x) on the surface was used as the filter material in this study, and its diameter was about 1.2–1.5 mm. The filter bed fixed was 1,500 mm high in the tube. At the bottom of the filter bed, the column had a 300-mm-high cobblestone in the support layer. The different initial concentrations of Mn²⁺, NH₄⁺ and BPA were added into the influent using several dosing pumps. The DO uptake rate (OUR) was calculated according to Garcia-Ochoa et al. (2010). The filtration rate was about 8.0 m/h throughout the operation in all experiments, and the overflow rate was 62.8 L/h. When the water level reached about 1.5 m above the bed layer or the effluent water quality deteriorated, the pilot-scale column was backwashed. The operation method of backwashing the filter column was as in a previous study (Guo et al. 2017).

Some scholars have studied the removal of BPA in aqueous solutions and drinking water, and the spiking concentrations were 0.50 mg/L (Xie et al. 2012) and 5.00 mg/L (Zhao et al. 2015) in their studies. Thus, concentrations of BPA of about 0.30–0.70 mg/L were selected in this study.

To determine the removal of BPA, BPA in different initial concentrations (about 0.30, 0.50 and 0.70 mg/L) was added into the influent using a dosing pump. Each experiment was conducted in triplicate and for at least 240 hours, and the concentrations of BPA and DO were measured every four hours. Depletion of BPA concentration in relation to the bed depth and a linear regression analysis of BPA depletion in relation to the empty bed contact time (EBCT) were discussed.

The influence of NH₄⁺ on the removal of BPA was studied. The BPA concentration was about 0.40 mg/L, and NH₄⁺ in different initial concentrations (about 0.60, 0.90 and 1.20 mg/L) was added by two dosing pumps. Each experiment was performed for at least 240 hours.

When NH₄⁺ and Mn²⁺ have been used as pollutants in a water environment, the typical concentration range of 0.50–2.00 mg/L has been used in some studies (Cheng et al. 2017; Guo et al. 2017). So concentrations of NH₄⁺ and Mn²⁺ of about 0.60–2.00 mg/L were selected in this study.

BPA (about 0.40 mg/L), NH₄⁺ (about 1.21 mg/L) and Mn²⁺ (0–2.00 mg/L) were added into the influent using three dosing pumps, and the concentration changes of the above pollutants in the influent and effluent were determined. Each experiment was performed for at least 240 hours.

The concentration of BPA was determined by a UV spectrophotometer (HITACHI, U-3310, Japan), and the concentrations of Mn²⁺ and NH₄⁺ were determined by the method of potassium periodate oxidation spectrophotometry and Nessler reagent spectrophotometry, respectively (SEPA of China 2002) (UV-1800PC, Shanghai Mepda Instrument Co., Ltd, China). DO, pH and T were detected using a multi-parameter water quality analyzer (HACH, HQ30d, USA).

The filter sands were frozen, vacuum-dried and kept in some sealed vacuum tubes after taking the sample. The surface characteristic of the oxide film was characterized using scanning electron microscopy (SEM, 600F, USA EDAX Company), and the elemental composition of the film was analyzed by an energy dispersive spectrometer (EDS, GENESIS XM, USA EDAX Company). The main compound structures of the film were analyzed by X-ray diffraction (XRD, MiniFlex600, Rigaku, Japan). The binding energies of the Mn, C and O on the surface of the film were characterized by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, USA), and the XPS spectra of the Mn(2p3/2), C(1s) and O(1s) were peak-fitted and analyzed by bundled software (Avantage) (Cheng et al. 2018).

The BPA removal and the change of the backwashing interval were investigated in the filter column. The role of the column is only to provide a place for the quartz sand filter material, which does not enhance the removal of BPA. As shown in Figure 2(a), BPA was not detected in the effluent when the BPA concentration was increased from 0.30 to 0.50 mg/L in the influent. When the influent concentration of BPA was up to 0.70 mg/L, the effluent concentration of BPA was about 0.23 mg/L, and the backwashing interval was shortened from 29 hours to ten hours. The backwashing interval was too short (only ten hours), which would not be conducive to the filtration process, so the concentration of BPA in the influent water should not be too high (less than 0.50 mg/L). As shown in Figure 2(b), most of the BPA had been removed before the 45 cm depth of the filter layer, and the BPA concentration did not change obviously in the rest of the filter layers. The OUR was about 29.03 g/(m³·h) when the influent concentration of BPA was about 0.50 mg/L (about 0.75 g/d). The concentration of BPA was increased to 0.70 mg/L (about 1.06 g/d), the consumption of DO was about 6.59 mg/L (about 9.93 g/d), and the residual DO concentration (2.04 mg/L) was not enough to remove excess BPA in the entire filter column. By maintaining the pH value and DO concentration in the influent basically constant, suppose the BPA depletion rate is pseudo-first order: −d[BPA]/dt = k [BPA] (Guo et al. 2017). From Figure 2(c), the plot of log{[BPA]_t/[BPA]_o} versus EBCT is linear (k = 0.197), proving that the BPA oxidation process was pseudo-first-order kinetics.

The effect of NH₄⁺ on the removal of BPA is shown in Figure 3. Before adding NH₄⁺ into the influent, the BPA in the water could be continuously and effectively removed, and the effluent concentration of BPA was always close to 0. The influent concentration of NH₄⁺ was gradually increased from 0 to 1.21 mg/L, and the effluent concentration of BPA began to increase and finally remained at 0.14 mg/L, and the was reduced from 0.47 to 0.26 mg/L in the filter column (Figure 3(a)). From Figure 3(b), the higher the ammonium concentration in the influent, the slower the BPA concentration decreased along with the bed depth, and it can be presumed that NH₄⁺ was oxidized prior to BPA. From Figure 3(c), when the influent concentration of NH₄⁺ was increased by more than 0.90 mg/L, the effluent concentration of DO was less than 1.96 mg/L, so the removal efficiency of BPA was significantly reduced. DO was a limiting factor for the removal of BPA, and the lower concentration of DO was not enough to oxidize the remaining BPA. The oxidation process of ammonium also needed to consume DO in water (Guo et al. 2017). Accordingly, the presence of NH₄⁺ would hinder the removal of BPA.

The influent concentrations of BPA and NH₄⁺ were approximately 0.40 mg/L and 1.20 mg/L, respectively. With the increase of Mn²⁺ concentration (0–2.00 mg/L) in the influent, the removal efficiency of BPA and NH₄⁺ was slightly increased, and the concentration of BPA remained at about 0.10 mg/L in the effluent (Figure 4(a)). From Figure 4(b), when the Mn²⁺ concentration was about 2.00 mg/L in the influent, more than 90% of BPA, Mn²⁺ and NH₄⁺ were removed before the 65 cm depth of the filter layer, and their concentration change curves were consistent with the change of DO (Figure 4(c)). The OUR was about 8.11 g/(m³·h), which was basically the same as the concentration change of DO in Figure 3(c). Therefore, it is speculated that Mn²⁺ was mainly oxidized to manganese oxide on the surface of the film by MeO_x, and there was no need to consume the DO in the water during the removal process of Mn²⁺, which is consistent with our previous research (Guo et al. 2019). To sum up, the presence of Mn²⁺ would assist in the removal of BPA and NH₄⁺ by MeO_x.

The filter materials were characterized by SEM before and after adding BPA. As shown in Figure S1(a) and S1(b) (Supplementary Information), the oxide film surface was relatively complete and covered the quartz sand surface, and the structure was compact and the pores were developed, which were more conducive to the adsorption of pollutants in water. After adding BPA in the influent (Figure S1(c) and S1(d)), the oxide film showed a slight cracking, and some had fallen off. The originally well-developed pore structure was blocked, which was mainly caused by the excess substance from the oxidation process of BPA. The active sites on the oxide film surface would be gradually decreased, thereby affecting the effective removal of BPA in water. As shown in Figure S1(e) and S1(f), the film structure was restored after continuous dosing of Mn²⁺ into the influent. Because the Mn²⁺ could be oxidized to manganese oxide by MeO_x, these oxides were continuously adsorbed on the surface of the oxide film to ensure that the film was gradually renewed.

The EDS spectrum is used to characterize the change of the elemental composition on the film surface in Figure 5. The major elements are O, Mn and Fe in the original filter materials, so the iron and manganese oxides are the main components of the oxide film. When the BPA was added into the influent, the composition of C and O were both increased by about 9.00%, resulting in a significant decrease from 63.48% to 44.55% in the proportion of Mn elements. Therefore, the reduced Mn oxides could lead to decrease in the removal efficiency of BPA. After adding Mn²⁺ into the influent, the content of C, O and Si on the film surface decreased, and the content of Mn and Fe increased significantly. The structure of the filter film was gradually repaired.

The oxide film with different experimental stages was characterized by XRD, and the characteristic peaks and compounds are shown in Figure 6. The first characteristic peak was high and narrow, and this structure was a typical manganese oxide Mn₆O₁₂·3H₂O from the PDF#84-1714 card, which was conducive to the catalytic oxidation of BPA. The weak peak (at 30.22°) represents SiO₂, which was known from the PDF#32-1128 card. In addition, the third characteristic peak was wide and of low intensity, and its structure was typical of Bermanite minerals (Mn(H₂O₄)Mn₂(OH)₂(PO₄)₂) from the PDF#71-1793 card. Both the manganese oxide Mn₆O₁₂·3H₂O and Bermanite minerals might be used as effective substances of MeO_x to oxidize BPA in water.

The oxide film with different experimental stages was analyzed by XPS, and the spectra of C(1s), O(1s) and Mn(2p3/2) are shown in Figure 7. As shown in Figure 7(a), the main elements of the film were C, O, Mn and Fe, and the Mn and Fe had obviously changed after adding Mn²⁺. After adding BPA into the influent, the manganese oxide had not changed significantly, and the forms of manganese oxide mainly included Mn₃O₄ (Audi & Sherwood 2002) and MnFe₂O₄ (Junta & Hochella 1994) (Figure 7(b)). From Figure 7(c), the C = O (Shchukarev & Korolkov 2004) bond and the possible formation of C-containing intermediates [−CH₂C − H(OH)]_n (Stoyanov et al. 1990) were increased significantly on the surface of the oxide film, and these increased C-compounds covered the oxide film surface and blocked the pores, which led to the decrease in the removal of BPA. From Figure 7(d), the O element on the oxide film surface was mainly from MnO_x (Strohmeier & Hercules 1984). In addition, Si − O (Nohira et al. 2002) and C = O bonds were increased after adding BPA, and the organic matter covered the dense loose pores of the filter film, resulting in the activity decrease of the filter film. These generated manganese oxides could be used to restore the activity of the oxide film after adding Mn²⁺. Therefore, the addition of Mn²⁺ can synergize the removal of BPA.

BPA could be efficiently removed by an iron and manganese oxide filter film in a pilot-scale filter system, and the maximum removal concentration of BPA was 0.52 mg/L (about 0.78 g/d), which consumed 5.44 mg/L (about 8.20 g/d) of DO. Continuous dosing of BPA into the influent would cause surface blockage and reduced activity of the oxide film, and it could also cause a significant decrease in the backwash interval of the filter column. The BPA oxidation process was pseudo-first-order kinetics. NH₄⁺ can hinder the removal of BPA, because NH₄⁺ can consume DO in water before BPA, but the presence of Mn²⁺ in water has a synergistic effect on the removal of BPA. When the DO in the water was lower than 2.00 mg/L, BPA was not effectively removed. The EDS energy spectrum showed that the content of C and O elements on the film surface was increased, while the content of Mn was decreased after adding BPA. The XPS spectrum showed that the main components of the oxide film were MnFe₂O₄ and Mn₃O₄, but the increase of carbonaceous substance in the oxide film after adding BPA resulted in a decrease in the BPA removal. By continuously adding Mn²⁺, the activity of the oxide film could be gradually restored to ensure the effective removal of BPA in water.

This work was financially supported by the Scientific Research Project of Shaanxi Education Department, China (19JC017), the Xi'an Municipal Science and Technology Project, China (2020KJRC0025), Natural Science Basic Research Plan in Shaanxi Province of China (2021JQ-688), the Doctoral Scientific Research Foundation (107020335), and Scientific Research Program Funded by Shaanxi Provincial Education Department (21JK0650). Thanks for the anonymous reviewers' invaluable advice.

All relevant data are included in the paper or its Supplementary Information.