Surface chemical environment regulation of carbon-based sorbent for the purification of organic impurities in H2O2 Authors: Jianguang Li, Tong Wang, Ziqin Chen, Zhikang Bao, Kang Shao, Jian Zhang, Feng Lin, Liang Lv Published: 2026 ABSTRACT AbstractPurification of organic impurities is critical for electronic-grade hydrogen peroxide (E-H2O2), since residual organics severely deteriorate the performance of semiconductor devices. Herein, we report a carbon-based sorbent (BAC-x) constructed via sodium metaborate (NaBO2), which shows outstanding organic impurities removal efficiency in 27.25 wt% and 38.84 wt% H2O2 solutions. Structure characterizations reveal that NaBO2 modification effectively tailors the surface chemical environment of activated carbon by introducing abundant surface functional groups, which act as active adsorption sites and promote chemisorption between the sorbent and organic impurities. Temkin adsorption isotherm and pseudo-second-order kinetic models confirm that surface chemical environment governs the rate-limiting adsorption step. As a result, BAC-800 exhibits an exceptional organic impurities removal efficiency of 96.28% with a residual total organic carbon (TOC) of 6.38 mg·L-1 that meets the standard of E-H2O2.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13065-026-01865-3.Keywords: Chemisorption, Electronic grade H2O2, Organic impurities, Purification, Wet electronic chemicals Introduction Wet electronic chemicals play a vital role in the microelectronics manufacturing chain, including wet etching, surface cleaning, photoresist developing and tripping process. [1–2] Among these, electronic-grade H2O2 receives extensive attention with its unique role in photoresist stripping and microcircuit patterns etching [3–5]. E-H2O2 is mainly produced from industrial-grade H2O2 (I-H2O2) through electrolytic approach, continuous resin adsorption or Anthony reaction [6]. The purity of I-H2O2 directly determines the quality and stability of the final E-H2O2 product. In particular, soluble organic impurities tend to adsorb on device surface that frequently cause incomplete cleaning and persistent residue [7]. Moreover, the organic contaminants can even provide a nutrient source for microbial growth, thus reducing the safety and quality of E-H2O2. Therefore, precise control the organic impurity in E-H2O2 is particularly important. The concentration of organic impurities in E-H2O2 is recommended to be less than 20 ppm that according to the most influential international SEMI (Semiconductor Equipment and Materials International) standard [8]. Organic impurities in I-H2O2 mainly derive from heavy aromatic hydrocarbons, trioctyl phosphate, anthraquinone and the corresponding derivatives [9]. Previous studies demonstrate that carbon-based absorbents exhibit excellent performance in the field of organic contaminants purification [10]. The highly developed porous structure and large surface area enable efficient adsorption of organic impurities at accessible active sites to facilitate the effective separation and purification [11]. Meanwhile, the versatile surface structure of carbon material allows the introduction of functional groups that can establish tailored π–π interactions with aromatic organic molecules. [12–13] In addition, the acid–base sites on the carbon surface regulate the interfacial chemical environment, which is essential for suppressing undesired H₂O₂ decomposition. However, high-performance carbon-based adsorbents capable of deeply removing organic impurities from I-H₂O₂ to meet electronic-grade specifications remain scarce. Activated carbon (AC) has been widely applied in aqueous pollution remediation. However, systematic studies targeting the purification of organic impurities from H₂O₂ solutions to satisfy the total organic carbon (TOC) criteria of the C8 electronic-grade H₂O₂ standard remain limited. Lin et al. reported that microporous structure of AC greatly influenced the adsorption capacity for organic impurities in H2O2 [14]. By adjusting the exposed graphite phase, Wang et al. displayed that steam-etched AC could remove nearly 100% organic impurities from 27.25 wt% H2O2 [15]. Besides, by constructing surface π-rich electron groups, KAC-800 achieved an organic impurities removal efficiency of 98.2% in 50 wt% H2O2 [16]. Despite these advances, there are still few studies focusing on the relationship between the polar chemical environment on AC surface and adsorption process, and the corresponding adsorption mechanism remains unclear. In particular, inappropriate surface modification may inadvertently generate sites that promote H₂O₂ decomposition rather than selective adsorption of organic impurities. Herein, we revealed the adsorption mechanism of organic pollutants in I-H₂O₂ by the polar chemical environment on the surface of AC. Surface functional groups were found to optimize the surface properties of carbon adsorbents, thereby substantially improving organic impurity removal performance. Meanwhile, these functional groups did not induce the decomposition of H₂O₂. Among them, BAC-800 exhibited excellent organic impurities purification ability from 27.25 wt% H2O2 and 38.84 wt% H2O2 with removal rate of 96.28% and 58.89%, respectively, which met the requirement of organic impurities in E-H2O2. Kinetic and thermodynamic investigations confirmed that surface base functional groups accelerated the chemisorption between adsorbates and sorbent, thereby resulting in promoted mass transfer of organic impurities to accelerate the removal efficiency. This study provides a new pathway for the purification of wet electronic chemicals via sorbent surface modification engineering, which is expected to facilitate the development of microelectronic industry. Results and discussion Scanning Electron Microscope (SEM) and Transmission Electron Microscope (TEM) were used to characterize the morphologies of samples, respectively. The well-preserved surface microtopography‌ in SEM images (Fig. 1a and Fig. S1) and unaltered morphologies in TEM images (Fig. 1b and Fig. S2) displayed that NaBO2 high temperature modification did not severely damage the framework of activated carbon. X-ray diffraction (XRD) patterns in Fig. 1c exhibited two obvious diffraction peaks. The peak at 2ϴ=23° corresponded to the (002) crystal plane, which represented the spatial regularity of graphite microcrystalline layer. While the peak at 2ϴ=43° was assigned to the (100) crystal plane, which was related to the size of the crystal plane of graphite crystallites. With the treated temperature increased, the intensity of (002) crystal plane peak slightly decreased, while the (100) peak increased. [17] This variation in crystallite-related peaks indicated a weakened regular arrangement of graphite layer and a slight increased in the graphitization degree of BAC-x. The peak at 26.5° is generally attributed to the (002) crystal plane diffraction of graphitic layered microcrystalline structures in activated carbon. [18] The low content and lack of surface functional groups mean it does not play a decisive role in the adsorption process‌. As displayed in Raman spectra (Fig. 1d), the band at ~ 1350 cm−1 corresponded to the amorphous carbon (D band), while the broad band at ~ 1580 cm−1 was the G band corresponded to the graphitized carbon. [19, 20] The degree of graphitization in BAC-x was correlated with the ID/IG ratio (Fitting peak area). The ID/IG values of AC, BAC-600, BAC-700, BAC-800, BAC-900 and BAC-1000 were 1.80, 1.76, 1.70, 1.69, 1.72 and 1.78, respectively. The decrease in ID/IG ratios attributed to the increased content of graphitic carbon (BAC-800 reaches the maximum value), which was consistent with the XRD results. Furthermore, N2 adsorption-desorption isotherms provided valuable insights into the pore structure of BAC-x (Fig. S3). The presence of a obvious H4-type hysteresis loop confirmed the retention of the porous structure after modification. The BET surface area and pore structure parameters of the samples (Table S1) showed that the surface area and pore volume of the AC sample slightly increased after treatment at 600 °C, followed by a slight decrease with further temperature elevation‌, which may be attributed to the combined effects of pore development and structural collapse at high temperatures. At 600 °C, the activation process promotes the formation of new pores and the expansion of existing ones, leading to increased surface area and pore volume. However, as the temperature rises beyond this point, excessive carbon burn-off and pore wall collapse may occur, resulting in a reduction in both parameters. The average pore diameter did not show significant change‌, remaining relatively stable throughout the temperature variation process. a SEM images of AC and BAC-800; (b) TEM image of BAC-800; (c) XRD patterns; (d) Raman spectra; (e) FT-IR spectra The surface chemical environment of the sorbents was further investigated via Fourier-transform infrared (FT-IR) spectroscopy (Fig. 1e). Characteristic adsorption bands corresponded to the stretching vibration peak of C-H bond (2832 cm-1), C-C vibration (1367 cm-1) and the C = C stretching vibration peak of conjugated graphitic domains (1610 cm-1) were obvious observed. [21] The reduced intensities of the C-C and C = C were attributed to the bond cleavage at high temperature. These cleaved C-C and C = C bonds generated additional unsaturated sites, which were conducive to the adsorption of organic impurities. To further invest the surface chemical structure, X-ray photoelectron spectroscopy (XPS) was conducted. The survey XPS spectra (Fig. S4) and high-resolution elemental spectra (Fig. S5) confirmed the absence of Na and B elements signals, indicating the complete removal of residual NaBO2 after modification. High-resolution XPS C 1s spectra could be deconvoluted into four subpeaks at 284.8 eV, 286.2 eV, 288.6 eV and 290.9 eV (Fig. 2a), which corresponded to the C-C, C-O, C = O and O-C = O, respectively. [22] After the modification process, the relative content of C-O、C = O、O-C = O increased (Table S2), which could be the results of the chemical reaction between NaBO2 and surface carbon atoms of activated carbon under high temperature. The increased C = O content represented an increased contents of surface functional groups on the sorbent surface, which can also be obtained from the acid-base titration results (Table S3), thereby enhancing the adsorption capacity for aromatic organic impurities. This result confirmed that the precise regulation of the surface chemical environment was critical for optimizing the performance of carbon-based sorbents in organic impurity purification. a high resolution C 1s XPS spectra and (b) O 1s XPS spectra Then, the purification efficiency of organic impurities was evaluated by monitoring the total organic carbon (TOC) content. The initial TOC concentrations of 27.25 wt% and 38.84 wt% were 195.7 mg·L-1 and 226.4 mg·L-1, respectively. As shown in Fig. 3a, the organic impurity removal efficiency of BAC-x significantly superior to the unmodified AC sample, which confirmed the critical role of surface functional groups in adsorption. Among them, BAC-800 exhibited the excellent purification ability with a TOC removal rate of 86.7% in 27.25 wt% H2O2 with a residual organic impurity concentration of only 6.38 mg·L⁻¹, which met the strict organic purity requirements for electronic-grade H₂O₂. The removal efficiency of BAC-800 was comparable to that of state-of-the-art carbon-based sorbents (Table S4). However, the TOC removal rate of BAC-900 decreased to 58.89%, which was attributed to the strong oxidizing property of high-concentration H₂O₂ that compromised sorbent stability. The concentration of hydrogen peroxide after the reaction was measured. The results showed that the 38.84 wt% concentration hydrogen peroxide solution decreased to 35.08 wt% after the reaction, while the 27.25 wt% concentration hydrogen peroxide reduced to 24.80 wt% after the adsorption reaction. The slight decrease ruled out the possibility that the decrease in TOC was caused by the decomposition of hydrogen peroxide. Figure 3b showed the adsorption time effect on removal rate of BAC-800. When the adsorption time was 45 min, BAC-800 reached the adsorption equilibrium and exhibited a TOC removal rate of 86.51% (residual TOC: 23.14 mg·L-1) and 58.89% (residual TOC: 91.87 mg·L-1) in 27.25 wt% and 38.84 wt% H2O2, respectively, indicating rapid adsorption kinetics that was favorable for practical applications. a organic impurities removal performances of BAC-x; (b) adsorption time effect of BAC-800; (c) temperature effect of BAC-800; (d) dose effect of BAC-800 Figure 3c showed the influence of adsorption temperature on the purification performance of BAC-800. For 27.25 wt% H₂O₂, the optimal adsorption performance was achieved at 20 ℃, where the TOC was reduced to 23.14 mg·L⁻¹ with a removal rate of 97.19%. In contrast, the removal efficiency of 38.84 wt% H₂O₂ is not sensitive to changes in temperature. For 38.84 wt% H₂O₂, the maximum TOC removal rate of 59.86% (residual TOC: 90.87 mg·L⁻¹) was obtained at 10 ℃. Moreover, dose effect was also investigated and the results were presented in Fig. 3d. It was found that the dose of adsorbents related to the removal rate. The TOC value of 27.25 wt% H2O2 was decreased to 16.10 mg·L-1 by using 400 mg, while 8.49 mg·L-1 with 600 mg BAC-800. When the sorbent was 800 mg, the TOC decreased to 6.38 mg·L-1 accompanied with 96.28% removal rate. For 38.84 wt% H₂O₂, a maximum removal rate of 77.2% (residual TOC: 51.62 mg·L⁻¹) was achieved with 800 mg of BAC-800. The relatively lower removal efficiency for 38.84 wt% H2O2 attributed to the ultra-strong oxidizing capacity of high-concentration H₂O₂, which partially deactivated the sorbent’s active sites. Cyclic adsorption experiments and post-adsorption characterizations were conducted to evaluate the stability of the material. As shown in Fig. S6a, the crystal structure of BAC-800 did not show any significant changes before and after the adsorption reaction. In the FT-IR spectra shown in Fig. S6b, the decrease in the intensity of the absorption peak at around 1050 cm-1, which is assigned to C-O stretching vibration, indicates that these groups act as active sites for chemisorption of organic pollutants. After the four-round adsorption experiments (Fig. S6c), the adsorption performance of BAC-800 decreased by 25–30%, which is caused by the reduction of surface active functional groups after interaction with organic pollutants. Besides, the blank adsorption experiment - activated carbon and water under the same conditions was performed (Fig. S6d). After the adsorption process, the TOC content in pure water increased from 0.90 mg/L to 6.8 mg/L. This was caused by the release of organic pollutants contained in the AC material. After using BAC-800 for adsorption, the TOC content in 27.25 wt% H2O2 also dropped to this level, indicating that the organic pollutants inside have been basically absorbed. Moreover, after adsorption, the pH value of 38.84 wt% H2O2 changed from 1.64 to 1.78, while the pH value of 27.25 wt% H2O2 changed from 2.08 to 2.25. To elucidate the underlying purification mechanism, adsorption kinetics and thermodynamics were systematically investigated. Adsorption isotherms provided critical insights into the interaction between the sorbents and organic impurities. [23] The experimental data were fitted with the Langmuir, Freundlich, and Temkin adsorption isotherm models (Fig. 4a and Fig. S7), respectively, while the corresponding fitting parameters were summarized in Table 1. The Temkin adsorption isotherm model yielded the highest correlation coefficients (R2). It was 0.993 for AC and 0.957 for BAC-800 in 27.25 wt% H2O2, while 0.991 and 0.998 in 38.84 wt% H2O2, respectively, which were better than the correlation coefficients of Freundlich adsorption isotherm and Langmuir adsorption isotherm models. This indicated that Temkin adsorption isotherm model was most suitable for describing the adsorption process of organic impurities in H2O2. As revealed by Temkin model, the binding energy on the surface of adsorbent was uniformly distributed and the binding energy deceased linearly with the adsorption. This characteristic confirmed that the interaction between the functional groups on BAC-800 and organic impurities were dominated by chemisorption. Meanwhile, the correlation coefficients in Freundlich adsorption were 0.985 (AC, 27.25 wt% H2O2), 0.914 (BAC-800, 27.25 wt% H2O2), 0.996 (AC, 38.84 wt% H2O2) and 0.984 (BAC-800, 38.84 wt% H2O2), respectively. The well fitting of the Freundlich model suggested the occurrence of multilayer adsorption, reflecting that the overall adsorption process involved both physisorption and chemisorption. This dual-mode adsorption behavior was the result of the modified surface chemical environment. The porous structure facilitated physisorption via physical confinement, while the enriched surface functional groups promoted chemisorption through strong specific interactions with organic impurities. a Temkin adsorption isotherm plot; (b) pseudo-second-order kinetic plot Table 1Fitting parameters of Langmuir, Freundlich and Temkin modelsModelParameter27.25 wt% H2O238.84 wt% H2O2ACBAC-800ACBAC-800Langmuirqm/mg·g− 1-53.11-149.7-19.3628.30KL/L·mg− 1-0.005-0.004-0.005-0.917 R 2 0.7720.1950.9370.852Freundlich 1/n 1.4731.2542.1711.986KF/L·mg− 10.0600.2600.0010.003 R 2 0.9850.9140.9960.984Temkin B1 36.4536.4050.5656.22KT/L·mg− 10.0350.0560.0140.017 R 2 0.9930.9570.9910.998 Fitting parameters of Langmuir, Freundlich and Temkin models To further clarify the adsorption mechanism, adsorption kinetic analysis was studied using four kinds of kinetic modes including pseudo-first-order kinetic model, pseudo-second-order kinetic model, intra-particle diffusion model and Elovich model (Fig. 4b and Fig. S8). The fitting results were summarized in Table S5. Compared with other models, pseudo-second-order kinetic model exhibited the highest linearly dependent coefficient (R2 > 0.999) for both 27.75 wt% and 38.83 wt% H2O2 (Table S5). These results confirmed that chemisorption was the rate-limiting step in the organic impurity purification process, and the adsorption rate was directly determined by the number of active sites provided by the surface functional groups of BAC-x. This further highlighted the significance of precise surface chemical environment for optimizing the adsorption performance of carbon-based sorbents. Moreover, thermodynamic analysis was also performed to evaluate the spontaneity and nature of the adsorption process for 27.25 wt% and 38.84 wt% H₂O₂ (Table 2). For 27.25 wt% H₂O₂, the enthalpy change (ΔH) values were 0.600 kJ·mol⁻¹ for BAC-800 and 0.427 kJ·mol⁻¹ for AC. The positive ΔH values indicated that the adsorption of organic impurities was an endothermic process. With the increase of surface functional groups after NaBO₂ modification, the entropy change (ΔS) increased, confirming that the adsorption process was spontaneous adsorption. This enhanced entropy was attributed to the increased disorder at the sorbent-solution interface during the specific interaction between surface functional groups and organic impurities, resulting in a strong affinity of BAC-800 for organic contaminants. However, the opposite thermodynamic parameters were observed in 38.84 wt% H2O2. The reduced entropy might be the results of the increasing entropy effect caused by organic impurities desorption exceeded the decreasing entropy effect caused by adsorption. [15] Therefore, BAC-800 exhibited the relatively lower adsorption efficiency at high-concentration H₂O₂ as the oxidative environment disrupted the spontaneous chemisorption process. Table 2Thermodynamic parameters of AC and BAC-800 under 27.25 wt% H2O2 and 38.83 wt% H2O2SamplesT / KlnKΔG KJ·mol− 1ΔHKJ·mol− 1ΔS KJ·mol− 127.25 wt%H2O2AC293-0.45-34.970.427120.79298-0.94-35.57303-0.97-36.1727.25 wt%H2O2BAC-8002930.25-51.340.600177.26298-0.36-52.22303-0.47-53.1138.83 wt%H2O2AC293-1.62-6.410.12222.29298-1.70-6.52303-1.77-6.6338.83 wt%H2O2BAC-800293-1.23-4.270.08614.87298-1.30-4.34303-1.33-4.42 Thermodynamic parameters of AC and BAC-800 under 27.25 wt% H2O2 and 38.83 wt% H2O2 27.25 wt% H2O2 27.25 wt% H2O2 38.83 wt% H2O2 38.83 wt% H2O2 Conclusion Electronic grade H2O2 has strict requirements on organic impurities concentration. Herein, a high-performance activated carbon adsorbent (BAC-x) was prepared by high-temperature modification with NaBO2 to purity industry grade H2O2. The increased surface functional groups and improved graphitization of BAC-x mediated the acid/base environments and accelerated the adsorption of aromatic compounds. Therefore, BAC-800 exhibited excellent organic removal efficiency of 96.28% removal rate accompanied with 6.38 mg·L-1 residual total organic carbon content, which achieved the requirement of E-H2O2. The calculated adsorption isotherm and kinetic showed that the surface sites directly determined the chemical adsorption, which played an important role in industrial H2O2 purification. This work provides a new insight to design carbon-based sorbent toward organic impurities for electronic grade H2O2 purification. Experimental section Materials Activated carbon, NaBH4, NaOH and HCl were purchased from Sinopharm Chemical Reagent Co., Ltd. Industrial grade H2O2 was obtained from Jiangshan Hydrogen Peroxide Co., Ltd. (Jiangshan, China). From the enterprise, after receiving the freshly produced hydrogen peroxide, we used it to plot a standard curve via spectrophotometry, and then stored it in a refrigerator at 10 °C under dark conditions using a dedicated polyethylene (PE) plastic container. The container was fitted with a hole to release oxygen generated by decomposition and maintained internal pressure balance. Synthesis of BAC-x The carbon-based sorbents (BAC-x) were synthesized by a high temperature solid phase reaction. Firstly, activated carbon (20 g) was ground and sieved to a particle size of ~ 150 μm to obtain the fine powder. Then, the powder was sequentially washed with 10 wt% HCl, 10 wt% NaOH and deionized water successively. After drying, the activated carbon powders were mixed and ground with NaBO2 solution (1 g/L). The pretreated samples were heated to different temperature (600 °C, 700 °C, 800 °C, 900 °C, 1000 °C) at a heating rate of 10 °C/min and held at the target temperature for 2 h. The obtained samples were named as BAC-x (x = 600 °C, 700 °C, 800 °C, 900 °C, 1000 °C). After cooling to room temperature, the samples were rinsed with deionized water until the ionic conductivity approached zero by using ionometer (PXSJ 216 F). Finally, the BAC-x sorbents were stored in a vacuum drying oven and used without further purification. Characterization X-ray diffraction (XRD, Burker D8 Advance), FT-IR (Nicolet iS50), Raman (Thermo Scientific DXR 3Xi) and XPS (Thermo Scientific K-Alpha, USA) were used to analyze the crystalline, chemical functional groups and surface chemical structures of samples. Scanning electron microscope (ZEISS Sigma 300) and transmission electron microscopy (FEI Talos F200S) used to identify the morphologies. Brunauere-Emmette-Teller (BET, Micromeritics ASAP 2460, USA) used to analyze the pore volume and specific surface area. Acid-base titration was performed with an automatic potentiometric titrator. Briefly, an appropriate amount of sample was soaked in hydrochloric acid or sodium hydroxide solution for 48 h at a constant temperature of 25 °C. After filtration and washing, the filtrate was collected for titration. The titration endpoint was set to pH 7, and the content of surface acid-base functional groups was calculated based on the consumption volume of standard hydrochloric acid or sodium hydroxide solution. Purification process Before the adsorption experiments, an appropriate amount of hydrogen peroxide was poured from the PE container and placed in a water bath to equilibrate to the experimental temperature. Then we measured its concentration three times using spectrophotometry and take the average value, by monitoring the absorbance intensity of the TiIV–H2O2 complex at 410 nm. The organic impurity purification process was conducted using a thermostatic shaker. Firstly, 300 mg carbon-based sorbent was added to an Erlenmeyer flask containing 40 mL industrial H2O2. Then, the flask was oscillated with a speed of 140 rpm for one hour. After adsorption, the adsorbent was filtered out using a 0.22 μm microporous filter membrane, and the filtrate was analyzed with total organic carbon / total nitrogen analyzer (Multi C/N 2100s, Germany). The organic impurities removal efficiency (K) was calculated as followed: Where C0 represented the initial concentration of organic impurities; Ct was related to the residual concentration of organic impurities. Moreover, adsorption isotherm experiments were conducted following a similar operational procedure, with only the concentration of organic impurities adjusted. The Langmuir and Freundlish models were employed to fit the isotherm data and determine the isotherm constants. The adsorption kinetic experiments were performed by mediating the adsorbed time. And the pseudo-first-order and pseudo-second-order models were applied for fitting the kinetic data. Supplementary Information Supplementary Material 1 Acknowledgements Not applicable. Abbreviations Electronic-grade hydrogen peroxide Carbon-based sorbent Total organic carbon Industrial-grade H2O2 Activated carbon X-ray diffraction Fourier-transform infrared X-ray photoelectron spectroscopy Brunauere-Emmette-Teller Scanning Electron Microscope Transmission Electron Microscope Authors’ contributions JL designed the experimental direction and route, TW and ZC synthesized the samples and carried out the characterization experiments, ZB tested the performance of the samples, KS analyzed the data of the characterization results, JZ wrote the first draft of the paper, FL and LL modified and polished the first draft. All authors read and approved the final manuscript. Funding This work was supported by the the Science and Technology Research Competitive Projects of Quzhou No. 2024K142, Scientific Research Startup Foundation of Quzhou University No. KYQD006223006, Joint Fund of Zhejiang Provincial Natural Science Foundation of China under Grant No. LQZQN25E020003, Natural Science Foundation of Zhejiang Province No. LGG22B060002 and National Natural Science Foundation of China under Grant No. 22508225. Data availability All data generated or analysed during this study are included in this published article and its supplementary information files. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Jian Zhang, Email: jianzhang6666@163.com. Feng Lin, Email: linfeng@qzc.edu.cn. Liang Lv, Email: lianglv@qzc.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All data generated or analysed during this study are included in this published article and its supplementary information files.