Coal gangue (CG) is a complex solid waste generated from coal mining and coal washing, whose major chemical composition is SiO₂ and Al₂O₃ with carbon content of 15%–30% (Li et al. 2010; Wu et al. 2021; Zhang et al. 2022). The accumulation of coal gangue not only occupies land resources but also results in serious environmental pollution (Sun et al. 2009). In particular, the residual element of C and S in the waste pile can spontaneously ignite, releasing many harmful gases like CO, CO₂, SO₂, H₂S, and NO_x (Zhao et al. 2021). However, coal gangue could also be a useful source like heat energy, soil improvement, and (Al/Si/carbon)-based chemical products because of its high content of carbonaceous and clay minerals (Zhou et al. 2012; Li et al. 2023).

Moreover, coal mining and coal chemical can result in pollution of surface water and groundwater like ammonia nitrogen compounds, phenols and polycyclic aromatic hydrocarbons (PAHs) that are health-threatening and poor removal performance (Li and Zhang 2022; Mougnol et al. 2022; Wang et al. 2022). In particular, phenols (C₆H₅OH) and PAHs (two or more fused benzene ring molecules) are classified as priority pollutants. They are capable of undergoing bioaccumulating through food webs, and subsequently into humans (Zango et al. 2020). Therefore, it is necessary to comprehensive utilization of coal gangue and treat the (phenols and PAHs) contaminated wastewater flowing through mining areas.

A sub-category of advanced oxidation processes (AOPs), sulfate radical-based AOPs (SR-AOPs) have become a hotspot for the degradation of recalcitrant pollutants due to the long half-life (30–40 µs) and their selectivity to electron-rich moieties (Giannakis et al. 2021). Among the activation without energy input, SO₄²⁻ can be easily generated by the transition metals activation of peroxymonosulfate (PMS) (Ahn et al. 2019; Zhang et al. 2021a, 2021b) since its unsymmetrical structure and large O–O bond length (Kohantorabi et al. 2021). However, transition metals are practically limited by their relatively high cost and potential secondary leaching pollution. Thus, the development of efficient catalyst without secondary pollutants is imminent (Song and Liu 2018; Li et al. 2020).

In previous studies, carbon-based materials including (reduced) graphene oxides (Sun et al. 2012), carbon nanotubes (Duan et al. 2015), and activated carbons (Huang et al. 2016) are promising metal-free candidates for PMS activation, which would solve the secondary pollutants of transition metals. However, the above-mentioned efficient carbonaceous materials suffer from high preparation costs and complicated preparation procedures. Thus, it is highly desirable to develop some novel carbon-based materials with distinct activation mechanisms that are more suitable for practical applications (Xu et al. 2020). To date, coal gangue based material-ZnCl₂ (CGM-ZnCl₂) (Li et al. 2022) and algae-derived boron-doped biochar (Hung et al. 2022), have been applied in activating PMS for phenol and PAHs removal, respectively.

In the present contribution, we efficiently utilized the carbon and silicon aluminum skeleton within coal gangue to prepare a novel and more efficient PMS catalyst for naphthyl, phenanthrene, and phenol removal than CGM-ZnCl₂ (82.6% degradation rate of phenol in 25 min) with negligible heavy metal leaching. Not only taking advantage of coal gangue but also attaining the clean treatment of wastewater. We further explored the phase transition and gas product of CG + FeCl₂·4H₂O mixture during the heating process by XRD and TG-FTIR-MS, respectively, which has not been researched to our knowledge. And we also identify the reactive oxygen species (ROS) by integrating several strategies. This study is dedicated to the mechanistic study of PMS activation over coal gangue-based catalysts for the removal of organic contaminants in wastewater.

2.1 Materials

Ferrous chloride tetrahydrate (FeCl₂·4H₂O), sodium dihydrogen phosphate (NaH₂PO₄), tertiary-butyl alcohol (TBA, C₄H₁₀O) were purchased from Xilong Chemical Co., Ltd (Guangzhou, China). Potassium peroxymonosulfate (PMS, 2KHSO₅·KHSO₄·K₂SO₄) were purchased from Maya Reagent Co. (Zhejiang, China). L-histidine (C₆H₉N₃O₂), furfuryl alcohol (FFA, C₅H₆O₂), sodium chloride (NaCl), and phenol (C₆H₆O) were purchased from Aladdin Industrial Corporation. Sodium bicarbonate (NaHCO₃) were purchased from Beijing Reagent Co. (Beijing, China). Methanol (MeOH, CH₃OH) were purchased from Macklin Biochemical Co., Ltd (Shanghai, China). The coal gangue with the fixed carbon content of 16.7% was obtained from Tunlan City, Shanxi Province. All chemicals were used without further purification.

2.2 Composites preparation

The coal gangue (CG) was first broken and ground to 100–200 mesh. Then 4.0 g of coal gangue and a certain amount of ferrous chloride tetrahydrate were mixed and dispersed in 15 mL deionized water, and the magnetic stirrer was used to impregnate the reaction at room temperature for 3 h. The mixture was dried overnight at 80 °C and ground. Finally, the mixture was placed in a tubular furnace and calcined in a nitrogen (N₂) atmosphere at 750 °C for 90 min with a heating rate of 10 °C/min. The influence of ferrous chloride tetrahydrate amount on sample preparation was investigated, including the mass ratio of CG to FeCl₂·4H₂O was 4:1, 4:1.5, 4:2, 4:3, and 4:4. The prepared materials were denoted as CG-FeCl₂(4:1), CG-FeCl₂(4:1.5), CG-FeCl₂(4:2), CG-FeCl₂(4:3) and CG-FeCl₂(4:4), respectively.

2.3 Characterization

X-ray powder diffraction (XRD) was recorded with the range of 2θ from 5° to 80°. X-ray fluorescence (XRF) was applied to analyze the chemical composition of the CG and CG-FeCl₂(4:2). Scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), and selected area electron diffraction (SAED) were employed to observe the morphologies and microstructure of the CG and CG-FeCl₂ materials. Thermogravimetric (TG) under N₂ atmosphere and thermogravimetric analyzer coupled with fourier transform infrared and mass spectrometry (TG-FTIR-MS) under an Ar atmosphere were applied to the analysis of weight changing and gas products during the heating process. The pore structure parameters of the as-prepared samples were obtained by using multipoint N₂ adsorption/desorption isotherm with the methods of BET and BJH. Fourier transformed infrared (FTIR) spectroscopy was recorded with the scanning range from 400 to 4000 cm⁻¹. Raman spectroscopy was employed to detect surface structure changes in CG and CG-FeCl₂(4:2). X-ray photoelectron spectroscopy (XPS) was performed to the analysis of surface chemical states of the CG-FeCl₂(4:2) material. Inductively coupled plasma-optical emission spectrometry (ICP-OES) was used to analyze the Fe leaching of CG-FeCl₂(4:2). Total Organic Carbon (TOC) was applied to the analysis of TOC removal efficiency in the degradation process. Electron paramagnetic resonance (EPR) spectroscopy was performed to identify the reactive oxygen species (ROS) involved in the CG-FeCl₂/PMS system. The reaction products and potential reaction pathways of naphthyl/phenanthrene/phenol were revealed by a high-performance liquid chromatography-mass spectrometer (LC-MS).

2.4 Evaluation of catalytic activity

In a typical experiment, 60 mg of CG-FeCl₂(4:X) were dispersed into 30 mL of phenol solution (20.0 mg/L). The suspension was stirred for 30 min to achieve the adsorption-desorption equilibrium. Subsequently, 2.0 mL of PMS solution (30 mmol/L) was added to the above suspension. At desired time intervals, 0.5 mL of suspension was sampled and mixed with 0.5 mL of methanol to end the reaction, and then filtrated by a 0.22 μm filter.

The residual concentration of phenol was evaluated on a HPLC (LC-2030 C Shimadzu, Japan) with a C18 reverse-phase column (4.6 mm × 250 mm, 5 μm). The mobile phase was composed of acetonitrile (50%) and ultrapure water (50%) with a flow rate of 1.0 mL/min. The injection volume was 20 µL and the detection wavelength was 254 nm.

3.1 Characterization of catalysts

The XRD patterns of the CG and prepared CG-FeCl₂ composites are shown in Fig. 1a. The XRD pattern of CG showed the obvious characteristic peak at 26.64°, which corresponds to the (011) lattice of quartz (JCPDS No.85-0798). Meanwhile, the characteristic peaks at 12.38° and 33.05° in the XRD pattern of CG can be assigned to the (001) of kaolinite (JCPDS No.80-0886) and (200) lattice of pyrite (JCPDS No.71-1680), respectively. As for the prepared CG-FeCl₂ composites, the characteristic peaks of kaolinite and pyrite disappeared, while the characteristic peaks of iron sulfide (JCPDS No.65-0408), sekaninaite (JCPDS No.54-0679), mullite (JCPDS No.15-0776) and aluminum iron silicon (JCPDS No.45-1205) appeared, meaning that phase transition has appeared during the heating process.

Furthermore, the phase transition of main component of CG including pyrite, kaolinite, coal (for studying the carbon in CG), and their phase transformation products after adding FeCl₂·4H₂O are discovered through changes in the XRD patterns as shown in Fig. 1b. The XRD pattern of pyrite-FeCl₂(4:2) showed the presence of pyrrhotite (JCPDS No.29-0725), which was owing to the partial decomposition of the pyrite (Oliveira et al. 2018). However, for the prepared CG-FeCl₂ composites as displayed in Fig. 1a, only CG-FeCl₂ (4:1) appeared with the characteristic peak at 43.50°, which may be owing to the (2012) reflection of crystalline pyrrhotite. While other CG-FeCl₂ composites appear a characteristic peak at around 43.22°, which is due to the (102) reflection of crystalline iron sulfide.

Meanwhile, the XRD profiles of coal showed the peak at 20° < 2θ < 30°, which corresponds to the (002) reflection of carbon due to the stacking structure of aromatic layers (Takagi et al. 2004). While the XRD pattern of coal-FeCl₂ (4:2) showed the peaks at 43.20°, 43.60° and 44.70°, which corresponded to the (102) lattice of iron sulfide, (2012) lattice of pyrrhotite (maybe owing to the S element in coal combined with the Fe from FeCl₂·4H₂O), and (220) lattice of aluminum iron silicon, respectively (Fig. 1a). Moreover, X-ray diffractogram of the prepared kaolinite-FeCl₂ material doesn’t show any kaolinite peak but the presence of a halo peak band at 18° < 2θ < 25°indicated the formation of amorphous alumino-silicate phase (Elimbi et al. 2011). While the prepared CG-FeCl₂ (from 4:1.5 to 4:4) and calcined kaolinite-FeCl₂ material all showed the presence of mullite. Overall, the above XRD pattern difference between kaolinite-FeCl₂, CG-FeCl₂ composites, and calcined kaolinite-FeCl₂ may be due to the structural and impurity variations of kaolinite (Pask and Tomsia 1991). Besides, the results of HRTEM images and SAED pattern further proved the presence of mullite in the prepared CG-FeCl₂ materials as presented in Figs. 2j-m. However, the formation temperature of mullite from kaolinite is normally 1000 °C, which was obviously reducd after the introduction of FeCl₂·4H₂O. This is due to the iron could enter the lattice of mullite, affecting the morphology and reducing the temperature of mullite crystallization (Ilić et al. 2014; Ye et al. 2019).

XRD patterns of a CG and the prepared CG-FeCl₂ composites b Pyrite, Pyrite-FeCl₂(4:2), Kaolinite, Kaolinite-FeCl₂(4:2), Calcined Kaolinite-FeCl₂(4:2), Coal and Coal-FeCl₂(4:2) composites

Table 1 presents the chemical composition of CG and CG-FeCl₂ (4:2). The content of iron trioxide increases from 6.87 wt% for CG to 29.22 wt% for CG-FeCl₂ (4:2), which originated from pyrite and the FeCl₂·4H₂O activation. By combining the results of XRD (Fig. 1), FTIR (Fig. 4a), and XPS (Fig. 5c), it can be further confirmed that iron intercalated into the lattice during the activation process.

SEM images of a and b Coal gangue, c and d CG-FeCl₂ composites, e – l HRTEM images of CG-FeCl₂ composites, m SAED pattern of CG-FeCl₂ composites

The SEM and HRTEM images illustrate the morphological characteristics of CG and CG-FeCl₂ samples. As presented in Figs. 2a–d, the CG displayed a typically smooth 2D lamellar structure (Sun et al. 2020). While for CG-FeCl₂ samples, their structure became porous with many irregular cylindrical particles, which may be caused by the dehydroxylation of kaolinite and phase change (Cao et al. 2016) during the heating process. Furthermore, the HRTEM images of CG-FeCl₂ material revealed the interplanar spacing of 0.25, 0.54, and 0.26 nm, which could be indexed to the (111), (110) lattice planes of mullite (Kong et al. 2003) and (512) lattice plane of sekaninaite, respectively. The crystalline structure of mullite could also be demonstrated by the diffraction rings observed from the corresponding SAED pattern of CG-FeCl₂ samples.

a Adsorption/desorption isotherms and Pore size distribution plots of b CG and the prepared CG-FeCl₂ composites

To further analyze the pore structure of CG and the CG-FeCl₂ composites derived from different activator (FeCl₂·4H₂O) amounts, the obtained N₂ adsorption-desorption isotherms and pore diameter distribution curves were shown in Fig. 3. The shapes of the isotherms indicate changes in pore size distribution (Lopez et al. 2013). As shown in Fig. 3a, the curve shape of CG showed a typical II isotherm (IUPAC classification) with probable pore diameter of 12.756 nm, which indicates the presence of nonporous and macro-porous structures. On the other hand, the CG-FeCl₂ composites all exhibit Type IV isotherm with the adsorption-desorption hysteresis loop, which indicates a combination of micropores and mesopores with the pore diameter ranging from 3.879 to 10.216 nm (Zhang et al. 2014).

The calculated structure parameters of a series of coal gangue-based materials, including specific surface area (SSA), total pore volume, and average pore diameter are summarized in Table 2. It can be deduced that the addition of activation agent FeCl₂·4H₂O can drastically increase the SSA (4.45 to 66.50 m²/g) and total pore volume (0.014 to 0.084 cm³/g) of CG by almost 15 and 6 times, respectively. The change in SSA and total pore volume corresponds to a decrease in average pore diameter from 12.8 to 4.4 – 5.4 nm.

Moreover, the results in Fig. 3b and Table 2 suggested that the introduction of FeCl₂·4H₂O developed new mesopores during the mass ratio of CG/FeCl₂·4H₂O from 4:1 to 4:1.5. When the ratio ranged from 4:1.5 to 4:3, FeCl₂·4H₂O not only prompted the generation of mesopores but also enlarged existing pores, resulting in significant increases in SSA, total pore volume, and average pore diameter. However, when the ratio went up to 4:4, the SSA of CG-FeCl₂ decreased with the increased total pore volume and average pore diameter, which may be owing to its micropores being enlarged into mesopores (Qian et al. 2007). The above variation suggests that the amount of activation agent has a strong effect on the pore structure of CG-FeCl₂ materials, especially from 4:1 to 4:2. This is due to the generation of mesopores and the enlargement of existing pores, thus significantly altering the adsorption, migration, and catalytic degradation of materials against organic contaminant.

Figure 4a shows the FT-IR spectra of CG and the prepared CG-FeCl₂ materials. All the samples exhibited adsorption bands at 3442 − 3434 cm⁻¹, 1632 − 1616 cm⁻¹, 799 − 795 cm⁻¹, and 471 − 457 cm⁻¹, which corresponds to adsorbed H₂O, the C–C stretching vibrations of polyaromatic C = C (Zhou et al. 2007), symmetric Si–O stretching vibration, and in-plane Si–O–Si bending vibration within SiO₄ tetrahedra, respectively (Zhang et al. 2011). For CG, the adsorption bands at 935/913 cm⁻¹ and 539 cm⁻¹ represent the Al-OH and Al-O-Si bending vibrations, respectively (Zhu et al. 2016). The changes between the FTIR spectrum of CG and the prepared CG-FeCl₂ materials can be attributed to the activation of FeCl_2·4H₂O and the heating process. Especially, the characteristics peak of Si-O-Si shifted from 1033 cm⁻¹ for CG to 1107 − 1081 cm⁻¹ for CG-FeCl₂ materials. These are mainly ascribed to the formation of Si-O-Fe bonds since the coupling of silicon with iron (Zhang et al. 2021a, 2021b), which is consistent with the results of XRD and XPS. Besides, the adsorption bands at around 1180, 900 and 570 cm⁻¹ represent the Si-O and Al-O bending vibrations, respectively, and they were corresponding to the spectrum of mullite (Yoshida et al. 2010), which prove the presence of it in the prepared CG-FeCl₂ material again. Finally, bending vibrations of C = O and C-O may be covered by the adsorption bands at 1690 − 1550 cm⁻¹ (C–C) and 1107–1081 cm⁻¹ (Si–O–Fe).

a FT-IR spectra of CG and the prepared CG-FeCl₂ composites,  b Raman spectra of CG and the CG-FeCl₂ (4:2) material

XPS spectra of a survey, b C 1s, c O 1s, d Fe 2p for fresh and used CG-FeCl₂(4:2) material

Figure 4b shows the Raman spectra of CG and the prepared CG-FeCl₂(4:2) material. Among the spectra, well-known D and G bands were observed near 1350 and 1585 cm⁻¹, respectively. The ratio of the intensities of the disorder-induced D band to the symmetry-allowed G band ratio (I_D/I_G) has been used to characterize the defect density in carbon materials (Escribano et al. 2001; Saito et al. 2011). The value of I_D/I_G was higher for CG-FeCl₂(4:2) (1.87) than that of CG (1.78), indicating that the CG-FeCl₂(4:2) material possess more structural disorder (amorphousness) compared with CG (Sreńscek-Nazzal et al. 2016), enabling CG-FeCl₂(4:2) to exhibit a higher level of reaction possibility, thereby enhancing activities of CG-FeCl₂ for PMS activation (Wang et al. 2020).

Furthermore, the surface chemical states of the CG-FeCl₂(4:2) composite were identified by XPS analysis as shown in Fig. 5. The survey scan of both fresh and used CG-FeCl₂(4:2) (Fig. 5a) appeared the characteristic peaks of C 1s, O 1s, Si 2p, Al 2p, and Fe 2p. In the C 1s spectra (Fig. 5b), the deconvoluted peaks at 284.5, 286.0, 287.3, 288.9, and 291.0 eV in the fresh spectrum as well as 284.4, 286.0, 287.3, 288.9 and 291.2 eV in the used spectrum can be assigned to C–C, C–O, C = O, O = C–O bonds and adsorbed (CO and CO₂), respectively (Fen-rong et al. 2011; Chen et al. 2020). In the O 1s spectra (Fig. 5c), the deconvoluted peaks at 531.5, 532.0, 533.0, and 533.8 eV can be assigned to Al–O/C = O, Si–O–Fe, Si–O, and O = C–O bonds, respectively (Figueiredo et al. 2011; Wang et al. 2018; Hu et al. 2021). In the Fe 2p spectra (Fig. 5d), the fitted peaks at 710.9 and 724.3 eV in the fresh spectrum as well as 711.1 and 724.3 eV in the used spectrum assigned to Fe³⁺ (Yamashita and Hayes 2008), indicating that Fe³⁺ is the main valence state in CG-FeCl₂(4:2). Owing to the slow transformation of Fe³⁺ back to Fe²⁺, it is implied that Fe in CG-FeCl₂(4:2) has little effect on PMS activation.

The thermal behavior (50 to 750 °C) of CG, CG + 50% FeCl₂·4H₂O, and FeCl₂·4H₂O at the heating rate of 10 °C/min under an inert atmosphere was performed through thermogravimetry and differential scanning calorimetry (TG-DSC). As displayed in Fig. 6, CG + 50% FeCl₂·4H₂O has three stages of mass loss. The first two stages were around 50 to 135 °C and 135 to 190 °C, respectively, which both can be attributed to the evolution of H₂O and H₂ (Fig. 7a), corresponding to the heat absorption peaks at 130 and 180 °C on its DSC curve. The difference between these two stages is the type of water, including the physically absorbed moisture and the decomposition of FeCl₂·4H₂O (Bi et al. 2021). The third stage (around 335 to 750 °C) coincides with the dissociation of molecules (kaolinite and organic matter etc.) (Zhou et al. 2012) and the release of volatiles including H₂, HCl, and Cl (Fig. 7a and c), corresponding to the heat absorption peaks at 530 and 630 °C on its DSC curve. Thus, the dehydroxylation of kaolinite (Al₂O₃·2SiO₂·2H₂O) in coal gangue transformed to the crystalized mullite (Al₂O₃·2SiO₂) (Cao et al. 2016) occurrence in 400 to 750 °C can be drawn. Besides, considering the difference between TG-DSC results of CG + 50% FeCl₂·4H₂O and FeCl₂·4H₂O, especially around 600 to 750 °C. The conclusion can be proved again that the element of Fe is mostly retained after the heating process, which is consistent with the results of XRF (Fe₂O₃ from 6.87 wt% to 29.22 wt%), FTIR and XPS (the formation of Si-O-Fe bonds).

TG-DSC curves of the CG, CG + 50% FeCl₂·4H₂O (the mass ratio of CG to FeCl₂·4H₂O was 4:2), FeCl₂·4H₂O

Furthermore, the MS and FTIR investigation of the volatile products evolved during TG measurements (TG-MS-FTIR) offered overall thermal behavior of CG + 50% FeCl₂·4H₂O mixture under an inert atmosphere, as displayed in Figs. 7a–d. The standard FTIR spectra of H₂O, HCl, and CO from the National Institute of Standards and Technology (NIST) (Chu et al. 1999) were displayed in Fig. 7e as well. The MS signal for the M/Z = 18 of H₂O⁺ and M/Z = 17 of HO⁺ have similar evolution patterns in the 50–750 °C (Fig. 7a). Therefore, we can safely consider that HO⁺ was mainly resulted from H₂O⁺ splitting. Besides, the evolution of H₂ can be divided into two main stages. In the first stage, at 50 –300 °C, H₂ was resulted from water splitting. The second stage was above 400 °C, H₂ comes from the condensation of (hydro) aromatic or the decomposition of substrate abundant in hydrogen and heterocyclic compounds (Wang et al. 2013). As shown in Fig. 7b, even though the M/Z = 28 of CO⁺ has one peak at 67 °C, the typical bands at 2178 and 2114 cm⁻¹ were not detected in Fig. 7d, thus the evolution of CO didn’t take place in the heating process. It is noticeable that there is no peak for the M/Z = 32 of O₂⁺ and M/Z = 44 of CO₂⁺ (Fig. 7b), suggesting an almost insulated oxygen heating environment, which induced the results of FTIR (the rupture of C-C bonds) and Raman (I_D/I_G) was higher for CG-FeCl₂(4:2) (1.87) than that of CG (1.78). As shown in Fig. 7c, HCl is formed in the temperature ranges (above 485 °C rather than begin at 83 °C), which is similar to the peak of CO⁺ at 67 °C. Because HCl treatment can increase the pore volume and openness of coal samples (Yi et al. 2021), the evolution of HCl played a positive role in changing the pore structure of CG-FeCl₂. In addition, HCl is hazardous, and placing the outlet of the gas in water during the synthesis process can avoid harm to the human body to some extent.

Evolution of a–c MS curves d FTIR spectra during TG analysis of CG + 50% FeCl₂·4H₂O mixture under Ar-pyrolysis at 50 to 750 °C e Standard FTIR spectra of HCl, H₂O, and CO from NIST

FTIR analysis was helpful to distinguish the evolution stages of the volatile products of CG + 50% FeCl₂·4H₂O mixture. The characteristic peaks of H₂O appeared at 100 °C and gradually strengthened at 100 to 550 °C and kept stable at 550 to 750 °C (Fig. 7c). Meanwhile, the bands at around 2944 and 2798 cm⁻¹ for HCl had peaks above 500 °C. The characteristic peaks of HCl gradually strengthened at 550 ~ 600 °C and gradually weakened at 650 to 750 °C, which is agreement with the third mass loss stage of CG + 50% FeCl₂·4H₂O mixture (Fig. 6). Besides, there are three kinds of oxygen-containing functional groups at 500 to 750 °C, i.e. OH-ether O hydrogen bond, cyclic OH and C–O (from phenols, alcohols, and ethers), which is located near 3265, 3180 and 1188 cm⁻¹, respectively (Scaccia 2013; Yue et al. 2021).

Overall, the synthesis process from CG + 50% FeCl₂·4H₂O mixture to CG-FeCl₂(4:2) has occurred the following changes: (1) Phase transition and chemical composition: The principal component of coal gangue includes kaolinite, pyrite, and carbon. After the activation process, the crystalline phase of kaolinite and pyrite disappear, while the characteristic peaks of iron sulfide, sekaninaite, mullite, and aluminum iron silicon appear. Moreover, the content of Fe₂O₃ from 6.87 wt% for coal gangue increased to 29.22 wt% for CG-FeCl₂(4:2); (2) Evolved gases and pore structure: The volatilization of H₂O, H₂, and HCl leads to the weight loss of CG + 50% FeCl₂·4H₂O mixture, which has played an important role in the changing of the pore structure of the synthesized material as well (Yi et al. 2021; Gao et al. 2022); (3) Chemical bond features: The rupture of the C–C bond, more defects, electron-rich C = O groups, and the formation of Si–O–Fe bonds lead to an efficient and environmentally friendly catalyst.

3.2 Catalytic activity of catalysts

The catalytic performance of CG and the prepared CG-FeCl₂ materials were evaluated through the degradation of phenol under the same conditions. As can be seen from Fig. 8a, CG, CG-FeCl₂(4:1), and CG-FeCl₂(4:1.5) did not have a satisfactory performance due to a lack of adsorption capacity and low PMS activation ability for oxidizing phenol. In contrast, CG-FeCl₂(4:2), CG-FeCl₂(4:3), and CG-FeCl₂(4:4) could oxidize phenol with a degradation percentage of 95%, 96%, and 100% within 25 min, indicating that they could efficiently form ROS through PMS activation. Because the phenol degradation efficiency was consistent with the mass ratio of CG to FeCl₂·4H₂O. Hereupon, the mass ratio of 4:2 was chosen as the optimum dosage for the following experiments considering the cost and economy.

The removal of phenol (a) and linear transform ln(C₀/C) of the kinetic curves of phenol (b) with CG and the prepared CG-FeCl₂ composites, c phenol removal with CG-FeCl₂(4:2), Coal-FeCl₂(4:2), Pyrite-FeCl₂(4:2), Kaolinite-FeCl₂(4:2) and Coal, d The effect of phenol concentration, e effect of catalyst dosage, f effect of PMS concentration, g effect of anions-50 mM, h effect of pH on the removal of phenol; i Fe leaching concentration of CG-FeCl₂(4:2) during the degradation process. General conditions: C₀ (phenol) = 20 mg/L, C₀ (catalyst) = 2.0 g/L, C₀ (PMS) = 2.00 mM, V = 30 mL, initial pH = 6.48, temperature = 293 K

The pseudo-first-order kinetic pattern can be expressed as -Ln(C/C₀) = kt, where C is the concentration of phenol at instant t, C₀ is the initial phenol concentration, k presents the kinetic rate constant, and t is the time of reaction. The k results of phenol degradation on CG and some selected CG-FeCl₂ composites were displayed in Fig. 8b. The k (0.06469 min⁻¹) value of CG-FeCl₂(4:2) was nearly 45 times higher than that of CG (0.00143 min⁻¹), giving the superior generation rate of ROS in CG-FeCl₂(4:2)/PMS system. Moreover, the influence of coal (for studying the carbon), pyrite (Rahimi et al. 2021) and kaolinite (Li et al. 2019) (both have been applied to PMS activation) in coal gangue was considered for CG-FeCl₂/PMS system. Figure 8c depicts phenol degradation in various materials including CG-FeCl₂(4:2), coal-FeCl₂(4:2), pyrite-FeCl₂(4:2), kaolinite-FeCl₂(4:2) and coal. Among them, CG-FeCl₂(4:2) and coal-FeCl₂(4:2) could oxidize phenol completely after 25 min. However, pyrite-FeCl₂(4:2), kaolinite-FeCl₂(4:2) and coal only attain 29%, 14% and 7% removal percentage of phenol, respectively. Therefore, the FeCl₂·4H₂O activated carbon component within coal gangue as well as the introduced variable-valence iron play an important role for phenol removal in the CG-FeCl₂/PMS system.

The effect of phenol concentration, catalyst dosage, and PMS concentration was considered in the CG-FeCl₂(4:2)/PMS system. As can be seen from Fig. 8d, the degradation efficiency of phenol by CG-FeCl₂(4:2)/PMS depends on the initial concentration of phenol. Complete phenol removal can be achieved at phenol concentrations of 10 and 20 mg/L, while phenol removal efficiency decreased to 84% and 66% at phenol concentrations of 30 and 40 mg/L, respectively. The phenomenon may be owing to the mutual competition of phenol for ROS and a higher organic load under the same amount of PMS.

It is believed that phenol degradation depended on catalyst dosage in the CG-FeCl₂(4:2)/PMS system as well. Figure 8e shows phenol degradation at various catalyst dosages ranging from 1.75 to 2.5 g/L. The complete degradation is observed at 2.25 and 2.5 g/L of CG-FeCl₂(4:2) in 10 min, while in the same duration at catalyst dosages of 1.75 and 2.0 g/L, the phenol removal percentage reached 89% and 94%, respectively. The enhanced degradation efficiency of phenol was due to the more active sites under same organic load (Zeng et al. 2020; Madihi-Bidgoli et al. 2021).

PMS concentration is also a variable influencing phenol degradation, which is representative of the amounts of free radicals generated in SR-AOPs (Saputra et al. 2013). As displayed in Fig. 8f, phenol degradation efficiencies were > 90% at PMS concentrations between 1.33 and 3.33 mM, suggesting the great catalytic efficiency of CG-FeCl₂(4:2)/PMS system. Besides, the removal of phenol was faster with the PMS concentration of 3.33 mM than other lower PMS concentrations, which can be ascribed to that more radicals could be formed within unit time (Zhang et al. 2022).

The effect of anions (H₂PO₄⁻, Cl⁻, and HCO₃⁻) on the degradation of phenol in the CG-FeCl₂(4:2)/PMS system was also investigated (Fig. 8g). The removal efficiency increased with the introduction of HCO₃⁻, which might be attributed to that HCO₃⁻ could activate PMS to generate active species for degrading phenol (Yang et al. 2010; Wang and Wang 2020). However, the removal of phenol was suppressed in the background of Cl⁻/H₂PO₄⁻. Because both ^·OH and SO₄^·− can react with Cl⁻/H₂PO₄⁻ to generate the lower reaction efficiency radical of Cl^·, PO₄^·2−, HPO₄^·−, and H₂PO₄^·. Compared with H₂PO₄⁻, the lower suppression effect of Cl⁻ on the degradation of phenol may be owing to Cl^· only formed under acid conditions in the presence of ^·OH (Wang and Wang 2021).

Since the value of pH can directly influence catalytic activity, the degradation of phenol in the CG-FeCl₂(4:2)/PMS system in terms of pH (from 3 to 9) was studied as well. As shown in Fig. 8h, CG-FeCl₂(4:2) displayed efficient degradation of phenol over a pH range from 5 to 7 (removal percentage > 84%), whereas the degradation of phenol decreased at pH = 3 and 9. The phenomenon is due to the H-bond between H⁺ and the O–O group of PMS would be more significant in a strong acid environment, while at the pH of 9, the self-decomposition of PMS initiated, resulting in the decreased degradation efficiency (Wang et al. 2017). Overall, CG-FeCl₂(4:2)/PMS system exhibits good performance (removal percentage > 74%) in the pH range from 3 to 9, suggests it is favorable for treating wastewater in weak acid and weak base conditions.

Furthermore, the Fe leaching of CG-FeCl₂(4:2) in the phenol degradation process was studied by ICP-OES analysis (Fig. 8i). The amount of Fe concentrations for CG-FeCl₂(4:2) catalyst in the water after 3 cycles was 0.6512 µg/L, 0.0094 µg/L, and 0.0011 µg/L, respectively, which confirms negligible Fe leaching compared with Chinese Surface Water Environmental Quality Standard (GB 3838 − 2002, the standard value of Fe is 0.3 × 10³ µg/L). Moreover, since Fe is the only one major heavy metal element in CG-FeCl₂(4:2) (Table 1), the security of it for practical application was proved by results of ICP-OES.

a The TOC removal efficiency of phenol and b The removal efficiency of naphthyl and phenanthrene in the CG-FeCl₂(4:2)/PMS system. General conditions: C₀ (phenol) = 20 mg/L, C₀ (naphthyl) = 5 and 10 mg/L, C₀ (phenanthrene) = 5 and 10 mg/L, C₀ (catalyst) = 2.0 g/L, C₀ (PMS) = 2.00 mM, V = 30 mL, temperature = 293 K

To quantify the mineralization efficiency of phenol, the removal of TOC for this degradation process was determined (Fig. 9a). The TOC removal value of 66% was achieved within 1 min, suggesting fast mineralization of phenol in the CG-FeCl₂(4:2)/PMS system. Moreover, the removal value of TOC is 78% after 25 min. Thus, we can consider that intermediate compounds are produced and phenol can be efficiently degraded and mineralized in the CG-FeCl₂/PMS system. To investigate the removal efficiency of PAHs in the CG-FeCl₂(4:2)/PMS system, the degradation experiments of naphthyl and phenanthrene were conducted (Fig. 9b). High naphthyl (97%) and phenanthrene (96%) removal percentages can be achieved at their concentrations of 5 and 10 mg/L after 60 min of reaction. Therefore, we can safely consider that CG-FeCl₂(4:2)/PMS system can also be applied in the degradation of PAHs.

3.3 Reaction mechanism and degradation pathways

To confirm that CG-FeCl₂(4:2) or PMS cannot directly remove phenol alone, related experiments were conducted. As shown in Fig. 10a, the removal percentages of phenol only reach 19% in the CG-FeCl₂ system and 7% in the PMS system. The result suggesting that ROS are mainly generated from the catalytic effect of CG-FeCl₂(4:2) on PMS.

To identify involved reactive oxygen species (ROS) during the degradation process, related experiments were conducted in the presence of L-histidine and furfuryl alcohol (FFA) (scavenger of ¹O₂), tertiary butanol (TBA, scavenger of ^·OH), and methanol (MeOH, scavenger of ^·OH and SO₄^·−) (Yang et al. 2020). The removal efficiency decreased following the trend of L-histidine > FFA > TBA > MeOH (Fig. 10b), suggesting that ¹O₂ are the main species that contribute to the degradation of phenol in the presence of PMS. In addition, TBA and MeOH showed inhibitory effect with a phenol degradation percentage of 57% and 82%, respectively. This confirms that both ^·OH and SO₄^·− radicals show some responsibility for the degradation of phenol in the CG-FeCl₂/PMS system.

To further confirm the generation of ROS, electron paramagnetic resonance (EPR) with 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) and 2,2,6,6-tetramethyl-4-piperidinyloxyl (TEMP) as spin trapping agents was employed to capture radicals produced in the CG-FeCl₂/PMS system. Characteristic signals of ^·OH, SO₄^·−, and ¹O₂ were observed in the CG-FeCl₂/PMS system (Fig. 10(c-d). In addition, the signal intensity of ^·OH, SO₄^·− and ¹O₂ continually increased along with time, indicating that CG-FeCl₂ can continuously activate PMS to form ROS (Zhang et al. 2021a, 2021b). While in the CG-FeCl₂ alone system, no signals for ^·OH, SO₄^·−, and ¹O₂ were observed, indicating that ROS was generated from CG-FeCl₂ activation of PMS. These results agree well with the above experiments, together confirming that all of ¹O₂, ^·OH, and SO₄^·− radicals show responsibility for the degradation of phenol in the CG-FeCl₂/PMS system.

The Removal of phenol in a Different systems b The presence of L-histidine (5 mM), furfuryl alcohol (FFA, 5 mM), tertiary butanol (TBA, 500 mM), and methanol (MeOH, 500 mM ), electron paramagnetic resonance spectra of c ^·OH and SO₄^·−, d ¹O₂ over CG-FeCl₂ material

We further investigated the intermediates of naphthyl, phenanthrene, and phenol degradation in the GC-FeCl₂/PMS system by LC-MS. Two possible pathways of naphthyl, phenanthrene, and phenol were proposed based on recognized intermediates in the GC-FeCl₂/PMS system, as shown in Figs. 11a–c.

a The potential degradation pathways of naphthyl b Phenanthrene, and c phenol in CG-FeCl₂/PMS system. General conditions: C₀ (phenol) = 20 mg/L, C₀ (naphthyl) = 10 mg/L, C₀ (phenanthrene) = 10 mg/L, C₀ (catalyst) = 2.0 g/L, C₀ (PMS) = 2.00 mM, V = 30 mL, temperature = 293 K

Based on all the above results, the efficient degradation of phenol in the CG-FeCl₂/PMS system was proposed and summarized as followed: CG-FeCl₂ materials have larger specific surface area, pore volume, and smaller pore size compared to that of CG, which leads to better adsorption ability for PMS and phenol, eventually speeding up the process of PMS activation to form highly reactive complexes (part 1); The defects in CG-FeCl₂ can promote the transfer of electrons from phenol to activate PMS and elongate the O-O bond of PMS (HSO₅⁻) to generate ^·OH and SO₄^·−, which also have strong reactivity with aromatic compounds like phenol (part 2) (Duan et al. 2016). In addtion, the low oxygen content of CG-FeCl₂ materials are beneficial to the oxidation of organic contanminants in PMS system, and electron-rich C = O groups in them are the critical reactive sites via direct electron transfer and the generation of ¹O₂ or cleave PMS into ^·OH and SO₄^·− (part 3) (Wang et al. 2020; Huang et al. 2021) (Fig. 12).

The diagram of the possible degradation mechanism of phenol in the CG-FeCl₂/PMS system

To synthesize novel carbon-based materials and efficient utilization of coal gangue towards catalytic oxidation of organic contaminants in wastewater, the catalyst CG-FeCl₂ is developed in this study. The introduction of activator FeCl₂·4H₂O participated in the phase transformation and formed the Si–O–Fe bonds. And the main gaseous products including H₂O, H₂, and HCl during the heating process was proved by various characterizations. The resulting CG-FeCl₂ catalyst shows excellent peroxymonosulfate catalytic activity in the removal of naphthyl, phenanthrene and phenol (≥ 95%) under optimized conditions. This unique catalytic oxidation activity is mainly due to the synergetic effect of pore structure, defects, and electron-rich C = O groups. From the scavenging experiments, multiple active oxygen species including ^·OH, SO₄^·− and ¹O₂ are generated in the CG-FeCl₂/PMS system, and the significance order was ¹O₂>^·OH > SO₄^·−. Through the analysis of CG-FeCl₂/PMS/(naphthyl/phenanthrene/phenol) system by LC-MS, the potential degradation pathways of naphthyl, phenanthrene, and phenol were obtained as well. This work provides a feasible solution for the preparation of pollution-free coal gangue-based catalysts for the removal of PAHs and phenol.