1 Introduction

Recently, biochar/clay mineral composites have attracted much attention in the environmental remediation due to the good adsorption property toward environmental pollutants (Gao and Goldfarb 2021; Sewu et al. 2019; Viglasova et al. 2018; Yang et al. 2020). In fact, the choice of the carbon sources is crucial to the microstructure and adsorption properties of the biochar/clay mineral composites, which directly decides the pore structure, specific surface area, surface functional groups of composites (Gao and Goldfarb 2021; Viglasova et al. 2018). Therefore, different carbon sources were employed to prepare biochar/clay mineral composites including cellulose, macroalgae (Sewu et al. 2019), bamboo (Viglasova et al. 2018), camellia oleifera shells (Gao and Goldfarb 2021; Ssa et al. 2021; Viglasova et al. 2018), etc. Generally, the common preparation strategy of biochar/clay mineral composites involved pyrolysis of the mixture of biomass and clay minerals, but it was difficult to adjust the structure and performance of biochar/clay mineral composites at the molecular level due to the poor surface/interfacial interaction and the large gaps in the size and microstructure between biomass and clay minerals (Sewu et al. 2019). Therefore, there is an urgent need to find new carbon sources for the preparation of the well-defined biochar/clay mineral composites at the molecular level.

Oil shale semi-coke (OSSC) is a typical industrial waste derived from oil shale after distillation of oil shale, and it is composed of multi-minerals (kaolinite, illite, quartz, feldspar, etc.) and organic matters (Huang et al. 2019; Raukas and Punning 2009; Yörük et al. 2018), in which organic matters are tightly bound to minerals each other during the formation of oil shale. Due to the high mineralization degree and low oil content of oil shale, large numbers of OSSC were generated after extracting shale oil. In the past decades, the relevant researches mainly focused on the combustion characteristics of OSSC, OSSC with organic matters or its extracts as the auxiliary fuel based on the calorific value of OSSC (Hda et al. 2021; Qin et al. 2017; Wang et al. 2011; Yang et al. 2017). However, these ways did not realize the full-component utilization of OSSC accompanied with the discharging of numerous wastes. With the increasingly stringent of the environmental legislation and environmental management standards, the generated OSSC was mainly treated by land stockpiling at present, which took up much land and polluted the environment to different degrees as well (Liang et al. 2021), and thus OSSC presented dual attributes of pollution and resources. Taking the compositions and microstructure of OSSC into account, it may be a promising candidate for the fabrication of biochar/clay mineral nanocomposites at the molecular level based on its total compositions of clay minerals and organic matters.

From the perspective of reducing environmental pollution and full-component utilization of OSSC, the biochar/clay mineral nanocomposites were prepared by H3PO4 hydrothermal treatment followed by KOH activation using OSSC as a precursor for the removal of organic pollutants of methylene blue (MB), methyl violet (MV), malachite green (MG) and tetracycline (TC). The obtained biochar/clay mineral nanocomposites were characterized to investigate the evolution rules of the structure, composition, and morphology by using X-ray diffraction, X-ray fluorescence, Fourier transform infrared spectroscopy, scanning electron microscopy, transmission electron microscopy analyses and N2 adsorption–desorption isotherm. In addition, the effect of the contact time and initial concentration of pollutants on the adsorption properties of nanocomposites was evaluated, and the cyclic regeneration of the spent nanocomposites after the adsorption of pollutants was also studied. Based on the above results of structure and properties, the structure–activity relationship and adsorption mechanism were proposed.

2 Materials and methods

2.1 Materials

OSSC was provided by Yaojie Coal and Electricity Group Co. Ltd., Lanzhou, Gansu Province, China. It was smashed and ground into powder and passed through a 200-mesh sieve before use. H3PO4 (85 wt.%) and KOH were supplied by Sichuan Xilong Chemical Co. Ltd. All solutions were directly prepared using the deionized water.

2.2 Samples preparation

Firstly, OSSC and H3PO4 were mixed uniformly at different liquid–solid ratios of 1:4, 1:2, 1:1, 2:1 and 4:1, in which the mass of OSSC was fixed to be 10 g, and then the mixture was transferred into a Teflon autoclave to be maintained at 200 °C for 12 h. After being cooled to room temperature, the solid products were washed with water to neutral and dried at 70 °C for 12 h. The obtained products were labeled as OSSC-X, and X denoted the liquid–solid ratios of H3PO4 to OSSC of 0.25, 0.5, 1, 2 and 4, respectively. As a control, OSSC was hydrothermally treated without addition of H3PO4 using the same procedures and the sample was labeled as OSSC-0. The optimal liquid–solid ratio of H3PO4 to OSSC was optimized according to the maximum capacity of OSSC-X toward MB and TC, and then OSSC-2 was selected for the KOH activation. Typically, OSSC-2 was mixed with KOH with a mass ratio of 1:1 and calcined at 500 °C, 600 °C, 700 °C, 800 °C and 900 °C for 2 h in a tube furnace at a heating rate of 10 °C min−1 under nitrogen atmosphere, respectively. The obtained samples were washed with water to neutral and named as OSSC-2-Y, in which Y represented the calcining temperature.

2.3 Characterization

The surface morphology of the products was observed using a scanning electron microscope (SEM) with a JSM-6701F field emission. Surface element content and elemental mapping were measured with an Energy Dispersive X-Ray Spectrometer (EDS) equipped on the scanning electron microscope. The microstructure of the sample was measured with a JEM 2100 transmission electron microscope (TEM, Field Electron and Ion Co, USA). The X-ray diffraction (XRD) patterns were collected using an X′pert PRO X-ray power diffractometer equipped with a Cu-Kα radiation source (40 kV, 40 mA) (PAN analytical Co., Netherlands) from 3° to 80° with a step interval of about 0.167° using Cu-Kα radiation of 0.1542 nm. Thermo Nicolet 6700 FTIR spectrometer (Thermo Fisher Scientific, USA) was used to measure the infrared spectrum of the sample in the wavenumber range of 400–4000 cm−1. The pore structural parameters were measured with an ASAP 2010 analyzer (Micromeritics, USA) at 77 K, and the relevant parameters were calculated by a Brunauer–Emmett–Teller (BET) method. The specific surface area (SBET), total pore volume (Vtotal) and the micropore volume (Vmicro) were obtained from the volume of liquid N2 held at a relative pressure p/p0 = 0.99 and estimated by at plot method ultimately. Zeta potentials of the samples were determined using a ZEN3600 Zeta voltmeter (Malvern, Britain). Before testing, the powder sample was dispersed uniformly into deionized water by high-speed stirring (10,000 rpm) to form a 0.5 wt.% of aqueous suspension.

2.4 Evaluation of adsorption performance

The batch adsorption experiments were conducted to evaluate the adsorption properties of biochar/clay mineral nanocomposites toward MB, MV, MG and TC. Typically, 10 mg of the samples were added into 20 mL of aqueous solutions containing MB, MV, MG or TC, and then placed in a constant temperature shaker (THZ-98A) at 140 rpm and 25 °C. The concentration of organic pollutants was measured by UV–visible spectrophotometer (Shimadzu Instruments (Suzhou) Co., China) after adsorption, and the maximum absorption wavelengths were 664 nm, 581 nm, 618 nm and 392 nm for MB, MV, MG and TC, respectively. The equilibrium adsorption capacity (Qe, mg g−1) and removal ratio (R, %) of nanocomposites were calculated according to Eqs. 1 and 2, respectively.

$${Q}_{e}=\frac{\left({C}_{0}-{C}_{e}\right)\times V}{m}$$
(1)
$${{R}}=\frac{{C}_{0}-{C}_{e}}{{C}_{0}}\times100\%$$
(2)

where V (mL) is the volume of the adsorbate solution and m (mg) is the mass of adsorbent, C0 and Ce are the initial and equilibrium concentrations of organic pollutants, respectively.

3 Results and discussion

3.1 Optimization of preparation conditions

The preparation conditions of adsorption materials derived from OSSC were optimized based on the effects of modification method, added amount of H3PO4 and calcination temperature on the adsorption performance to the cationic organic pollutants of MB and TC. In this process, the control variable method was used, and the details of the relevant parameters are shown in Additional file 1: Table S1, in which the initial concentration and contact time of MB and TC were fixed to be 200 mg L−1 and 8 h. Figure 1a presents the effect of the different modification methods on the adsorption performance of the samples. It was found that the adsorption capacity of OSSC toward MB and TC was very small (8.92 mg g−1 for MB and 25.04 mg g−1 for TC). After two-step modification, the adsorption properties of samples increased significantly, and OSSC-2-800 exhibited the optimal adsorption capacity of 123.23 mg g−1 (MB) and 102.58 mg g−1 (TC). This indicated that the two-step modification method combing with H3PO4 and KOH was more favorable for enhancing the adsorption properties of OSSC. Therefore, the optimal added amount of phosphoric acid and calcination temperature continued to be investigated based on the above results, respectively. As shown in Fig. 1b and c, when the liquid–solid ratio of H3PO4 to OSSC was 1:2 and the calcination temperature was 700 °C, the corresponding sample (OSSC-2-700) presented the best adsorption performance toward MB and TC (153.39 mg g−1 for MB and 122.65 mg g−1 for TC). Therefore, OSSC-2-700 was selected to be characterized and analyzed.

Fig. 1
figure 1

Effect of preparation conditions on the adsorption capacity toward MB and TC including a different modification methods, b liquid–solid ratios of H3PO4 to OSSC and c calcination temperatures

Full size image

3.2 Effect of the added amount of H3PO4 during hydrothermal treatment

XRD patterns of OSSC before and after the H3PO4-assisted hydrothermal treatment are provided in Figs. 2a and 2b. The phase compositions of OSSC mainly consisted of quartz (2θ = 20.74° and 26.60°, JCPDS No. 46–1045) (Yang et al. 2020), kaolinite (2θ = 12.36° and 19.84°, JCPDS No. 78–1996) (Viglasova et al. 2018) and feldspar (2θ = 27.85°) (Zhao et al. 2022a, b). After the hydrothermal treatment without H3PO4, the characteristic diffraction peaks of OSSC-0 did not change significantly compared with raw OSSC, indicating that the crystal structure of OSSC was retained during the hydrothermal process. It also suggested that the minerals and organics of OSSC were stable and chemically inert combining with the adsorption properties toward MB and TC (Fig. 1a). After the addition of small amounts of H3PO4 (OSSC-0.25), the diffraction peaks of kaolinite and feldspar obviously weakened, which was mainly due to the dissolution of aluminum ions in the crystal skeletons of kaolinite and feldspar by H3PO4, and thus their crystal structures were destroyed by H3PO4 (Xu et al. 2022). As the liquid-to-solid ratio increased from 1:4 to 4:1, the diffraction peaks of kaolinite and feldspar continued to weaken until completely disappeared from OSSC-0.25 to OSSC-4 (Fig. 2a), especially the characteristic reflection of the (001) plane of kaolinite (2θ = 12.36°), indicating that high added amounts of H3PO4 accelerated the dissolution of aluminum ions in kaolinite and feldspar. This result was also evidenced by the fact that the content of Al2O3 gradually decreased with increasing added amount of H3PO4 (Table 1).

Fig. 2
figure 2

a XRD patterns of OSSC before and after modification at 5° ~ 35°, b local magnification at 20°–28°, c FT-IR spectra of OSSC before and after modification at 4000 cm−1–400 cm−1, d local magnification at 1300 cm−1–400 cm−1 (Q, K, F and B denoted quartz, kaolinite, feldspar and aluminum phosphate, respectively)

Full size image
Table 1 XRF chemical compositions (wt.%) of raw OSSC and OSSC-X
Full size table

In addition, it was also found from the XRF chemical compositions that the content of all metallic elements of OSSC decreased to varying degrees with the increase in the added amount of H3PO4. It indicated that high concentration of H3PO4 could etch most of the metal ions from OSSC compared with the chemical compositions of OSSC and OSSC-0 (Li et al. 2020). Notably, the diffraction peaks of aluminum phosphate (AlPO4) appeared at 2θ = 20.71° and 26.37° after the addition of H3PO4 (Fig. 2b), which implied that some of PO43− reacted with the free aluminum ions derived from kaolinite or feldspar. With the increase in the added amount of H3PO4, the peak intensity of AlPO4 enhanced first and then weakened. This phenomenon suggested that the moderate amounts of H3PO4 promoted the formation of AlPO4, while the excess H3PO4 prevented from generating AlPO4 or dissolved the formed AlPO4 (Dai et al. 2018). In other words, the etched aluminum ions could react with phosphate anion to produce the insoluble AlPO4 due to the low solubility product constant (6.3 × 10–19). With the increase in the added amount of H3PO4, more aluminum ions were etched and more AlPO4 was generated while the formation rate of AlPO4 was greater than its dissolution rate under the pH condition of the system. Finally, the generated AlPO4 might be dissolved again by the excess of H+ with the continuous increase in the liquid–solid ratio of H3PO4 to OSSC from 1:1 to 4:1, and thus the Al2O3 content of the samples continuously decreased (Table 1).

Figure 2c shows the FTIR spectra of OSSC, OSSC-0, OSSC-0.25, OSSC-1, OSSC-2, and OSSC-4. The absorption peaks of kaolinite were observed at 1114 cm−1, 1035 cm−1, 912 cm−1, 543 cm−1 and 467 cm−1, which corresponded to the asymmetric stretching of Si–O–Al, Si–O–Si, and Al–OH vibrations, as well as the bending vibrations of Si–O–Al and Si–O–Si (Liu et al. 2022). The peaks at 789 cm−1 and 694 cm−1 belonged to the characteristic absorption peaks of quartz, which were attributed to the stretching vibrations of Si–O on or between the silica-oxygen tetrahedral structures (Xu et al. 2022). The broad absorption band (1016 cm−1–1304 cm−1) was a superposition of the Si–O–Al bond and phosphate group (PO43−) absorption peaks (Xu et al. 2022). The bands at 3445 cm−1 and 1627 cm−1 were assigned to the H–O–H of water molecules (Ssa et al. 2021; Viglasova et al. 2018; Huang et al. 2019). In addition, the stretching vibrations of –CH2 (2944 cm−1), –CH3 (2847 cm−1) and bending vibrations of –C=O (1438 cm−1) were also observed, indicating that OSSC contained some organic matters (Raukas and Punning 2009; Yörük et al. 2018). After hydrothermal treatment without addition of H3PO4, all absorption peaks of OSSC-0 remained almost unchanged compared with those of raw OSSC, indicating that the hydrothermal environment did not affect the structure of organic and inorganic compositions of OSSC in the absence of H3PO4 (Hda et al. 2021).

By contrast, the FT-IR spectra of OSSC-0.25, OSSC-1, OSSC-2 and OSSC-4 presented some different changes compared with OSSC-0. With increasing added amount of H3PO4, the absorption peaks of the aluminum-containing functional groups (912 cm−1 and 543 cm−1) gradually they disappeared until disappeared. This indicated that the aluminum-oxygen octahedral skeleton and the Si–O–Al bonds connecting the tetrahedra and octahedra of kaolinite were destroyed (Wang et al. 2019a, b), and the damage became more and more serious with increasing H3PO4 concentration, which was also consistent with the results of XRD and XRF. The increase in the relative intensity of Si–O–Si at 467 cm−1 implied that more Si–O–Si was produced at a higher concentration of H3PO4 from OSSC-0 to OSSC-4 (Fig. 2d). This phenomenon might be due to the breaking of Al–O bonds of Si–O–Al functional groups by H3PO4, which resulted in increasing the amount of free Si–O− followed by bonding with each other to form Si–O–Si bonds (Lu et al. 2020). These findings were also in agreement with the results of XRF elemental analysis (Table 1). In addition, the characteristic peaks of Si-based compounds (789 cm−1 and 694 cm−1) and organics (2944 cm−1, 2847 cm−1, and 1438 cm−1) were still observed, suggesting that the excess amount of H3PO4 had no significant effect on the quartz and organics of OSSC (Lu et al. 2020). However, the relative intensity of the superimposed absorption band (1016 cm−1 ~ 1304 cm−1) exhibited a clear trend of first decreasing, increasing and finally decreasing again with increasing concentration of H3PO4. This was mainly due to the fact that H+ gradually accelerated the damage of Al–O bonds on Si–O–Al functional groups with the increase of H3PO4 content from OSSC-0 to OSSC-1. Meanwhile, the possible binding of PO43− with the dissolved free Al3+ increased (Dai et al. 2018). As the liquid–solid ratio was less than 1, the dissolution rate of Si–O–Al bonds was greater than the formation rate of AlPO4, and the relative intensity of the corresponding absorption bands weakened. With the continuous increase in the liquid–solid ratio to 2, the dissolution rate of Si–O–Al bonds was smaller than the formation rate of AlPO4, and the relative intensity of this absorption band increased. However, AlPO4 was dissolved by the excess H+ as the concentration of H3PO4 increased to a certain level (OSSC-4), and thus the relative intensity of the superposition band weakened (Quesada-Plata et al. 2016). This phenomenon was also consistent with the XRD results.

Figure 3 illustrates the microscopic morphology and the corresponding elemental compositions of OSSC before and after modification. Figure 3a, d, g and j indicates that the raw OSSC exhibited a large and disordered accumulation of lamellar morphology with slight agglomeration at the microscopic level. The EDS spectra and elemental mappings suggested that the elemental compositions of the lamellar structure were mainly composed of carbon, oxygen, aluminum and silicon elements (Fig. 3a inset and Additional file 1: Fig. S1), which were uniformly distributed on the surface of the nanosheets. Combined with the above XRD and FTIR results, the lamellar structure consisted of carbon-containing organic matter and inorganic silicates (kaolinite, quartz, etc.). After being hydrothermally treated without addition of H3PO4 (OSSC-0), the lamellar structure was preserved, and the elemental components had no obvious changes compared with OSSC (Fig. 3b inset). However, the lamellar morphology of OSSC-0 became thinner and smaller than that of OSSC accompanied with partial curling and agglomeration (Fig. 3e, h, k). Combining the results of FTIR spectra and XRF chemical compositions, this phenomenon might be ascribed to the following reasons: (i) The hydrothermal reaction induced the hydrothermal carbonization of part of the organic matter in OSSC to soluble carbonates and gases, which was indirectly demonstrated by the decrease in the total amount of carbon and oxygen in the EDS analysis (Additional file 1: Fig. S1); (ii) The consumption of some components of OSSC during hydrothermal process led to the disruption of the force field symmetry of some ions retained on the lamellae, which created a residual bonding force and thus increased the surface energy of the lamellae. Based on the principle of energy minimization, the ions containing residual bonding forces interacted spontaneously to reduce the overall energy by surface relaxation and surface reconstruction, which led to the curling of the lamellae at the microscopic level. After the addition of H3PO4, it was observed that the curling of the lamellae obviously disappeared (OSSC-2), but a large number of pits were formed on the retained lamellar morphology (Fig. 3c, i, l). This was mainly due to the dissolution of metal ions by high concentrations of H3PO4, which resulted in the collapse of the octahedral and tetrahedral skeletons of kaolinite and feldspar (Dai et al. 2018). In addition, a large number of nanoparticles could be found on the surface and edges of the nanosheets (Fig. 3c, f, i, l). Taken into account the distribution of aluminum elements (Fig. 3c), these nanoparticles might be the generated AlPO4 during the H3PO4-assisted hydrothermal treatment, which was also consistent with the results of XRD, FTIR and XRF analyses.

Fig. 3
figure 3

SEM images, elemental mappings (the insets) and TEM images of OSSC (a, d, g, j), OSSC-0 (b, e, h, k) and OSSC-2 (c, f, i, l)

Full size image

3.3 Effect of calcination temperature during KOH activation

Figure 4 shows the structural and morphological changes of OSSC-2 before and after KOH activation. As shown in Fig. 4a, no diffraction peaks of aluminum phosphate were observed for OSSC-2-700 compared to OSSC-2, indicating that the activation process decomposed aluminum phosphate. In addition, the diffraction peaks of kaliophilite were observed (2θ = 28.73° and 34.21°), indicating that introduction of KOH during the activation process participated in the reaction and promoted the formation of the new phase of kaliophilite (Dai et al. 2018). This conclusion also could be demonstrated by comparing the changes in the aluminum and potassium content before and after KOH activation (Table 1). It was noteworthy that the aluminum content of OSSC-2-700 reduced compared with that of OSSC-2 (Table 1). This suggested that not all aluminum from aluminum phosphate was converted into kaliophilite. As a result, the remaining aluminum elements were removed due to the formation of water-soluble compounds (Raymundo-Pinero et al. 2005). Moreover, the intensity of the characteristic diffraction peaks of quartz weakened with a slight shift (from 26.58° to 26.64°), indicating a decrease in the crystallinity and content of quartz accompanied by a small amount of lattice distortion (Wang et al. 2019a, b). In fact, this phenomenon was mainly due to the following reasons: (i) Addition of KOH disrupted the spatial arrangement of the silica-oxygen tetrahedral, resulting in the formation of more unsaturated silica-oxygen groups, such as (SiO4)4−, (Si2O6)n4n−, etc., which subsequently reacted with potassium ions to form water-soluble potassium silicate salts. As a result, the intensity of the characteristic diffraction peaks of quartz reduced; (ii) The high-temperature activation excited the production of more highly reactive free potassium ions, which diffused into the structural interstices of the rest of silicon-oxygen tetrahedra and then replaced some of the lattice nodal positions of the silicon elements, causing lattice distortion in the tetrahedra (Ishida et al. 1987).

Fig. 4
figure 4

a XRD patterns, b FTIR spectra, c Raman spectra, d zeta potential, e N2 adsorption–desorption isotherms and f pore size distribution curves of different samples prepared by two steps. B, Q and Ka denote aluminum phosphate, quartz and kaliophilite, respectively

Full size image

Figure 4b shows the FTIR spectrum variation of OSSC-2 at different activation temperatures. It was found that the relative intensity of the absorption band of 1016 cm−1–1304 cm−1 significantly weakened after calcination at 500 °C (OSSC-2-500) compared with the main absorption peak of OSSC-2 at 1114 cm−1. It indicated that AlPO4 was decomposed and no absorption peaks of new Al-containing compounds were observed, and thus AlPO4 might be converted to water-soluble substances such as potassium metabaluminate and phosphorus-containing compounds (Dai et al. 2018). In addition, the characteristic absorption bands of kaliophilite were observed from 404 cm−1 to 694 cm−1 (Fig. 4b and Additional file 1: Fig. S3), which also confirmed the XRD and XRF conclusions (Zhao et al. 2022a, b). Meanwhile, the absorption peaks of organic functional groups (2944 cm−1, 2847 cm−1, 1438 cm−1, and 1396 cm−1) and silicon-oxygen groups (789 cm−1, 694 cm−1, and 467 cm−1) also diminished, suggesting that some organic and silicon-oxygen compounds in OSSC-2 were consumed during the calcination process with the help of KOH. When the calcination temperature increased from 500 to 700 °C (from OSSC-2-500 to OSSC-2-700), the absorption peaks of AlPO4 and organic matter disappeared completely, while the absorption bands of kaliophilite became more obvious, so AlPO4 and organic matter might be maximally consumed and converted at this temperature. When the temperature rose to 900 °C, the relative intensity of the absorption peaks of kaliophilite diminished, while the relative intensity of the Si–O bond increased slightly (OSSC-2-900). It could be inferred that the higher temperature might destroyed the functional groups of the formed kaliophilite and promoted the combination of free oxygen ions with silicon ions to re-form Si–O bonds (Xu et al. 2020). Compared with the photographs of the samples before and after modification (Additional file 1: Fig. S2), it was found that the color of OSSC-2-700 did not change significantly and remained as a black powder, which implied that the activation process did not completely consume the organic matter of OSSC-2.

In order to analyze the structure change of organic matter of the samples, the Raman spectra of different samples were conducted (Yuan et al. 2019). As shown in Fig. 4c, the bands in OSSC at 1366 cm−1 and 1599 cm−1 represented the D-band (defect band) and G-band (graphite band) of carbonaceous matter, respectively (Xu et al. 2020). It suggested that OSSC contained more carbonaceous organic matter, and the ID/IG value of OSSC was calculated to be 0.903, indicating the low graphitization of carbon in organic matter of OSSC (Xu et al. 2020). After being hydrothermally treated with the assistance of H3PO4 (OSSC-2), the position of D-band shifted to the right (from 1366 cm−1 to 1372 cm−1) accompanied with a left shift for G-band (from 1599 cm−1 to 1595 cm−1), implying a decrease in the degree of aromatization and the overall structural order of organic matter (Álvarez-Docio et al. 2019). It might be due to the etching of metal ions from OSSC by H3PO4, which destroyed the stable organic–inorganic binding form. In addition, the ID/IG value of OSSC-2 decreased to 0.881, suggesting an improvement in the degree of graphitization after hydrothermal process (Álvarez-Docio et al. 2019). In the case of OSSC-2-700, the D-band continued to shift to the right, while the G-band shifted to the left, and it implied the decomposition of the aromatic ring in the organic matter in the activation process (Xu et al. 2020), which was also consistent with the results of FTIR analysis. Furthermore, the ID/IG value increased to 0.979, demonstrating a decrease in the graphitization due to the depletion of the ordered aromatic compounds (Xu et al. 2020). By contrast, the intensity of the D and G peaks of OSSC-2-700 sharply reduced with respect to OSSC and OSSC-2. It indicated that the activation process consumed a large amount of carbon from organic matter, and only a small amount of carbonaceous matter was retained and co-existed in an ordered and disordered state (Álvarez-Docio et al. 2019; Xu et al. 2020).

The values of zeta potential of the samples before and after KOH activation at different temperatures are depicted in Fig. 4d. It was found that OSSC-2 presented a negative charge (− 31.11 mV). With the increase in the calcination temperature from 500 °C to 900 °C, the zeta potential values of the calcined samples increased compared with that of OSSC-2 (Additional file 1: Table S2). It indicated that the calcination process contributed to exposing or producing more negatively charged groups (Zhao et al. 2022a, b). Combined with the results of XRD, FTIR and Raman spectra, the negatively charged groups might be composed of the following types: (i) the (SiO4)4− groups in OSSC-2, and the unsaturated Si–O− bonded with organic matter before calcination (Wang et al. 2021a, b); (ii) the negatively charged carbonaceous organic groups retained after activation (Zhao et al. 2022a, b); (iii) the free oxygen and silicon ions to combine with each other to form new Si–O−during the high-temperature calcination (Yuan et al. 2019) and (iv) the unsaturated K/Si/Al-O− bonds at the edge of the kaliophilite lamellar structure (Additional file 1: Fig. S4). In addition, it could be seen from Fig. 4d that the value of zeta potential first increased and then decreased from 500–900 °C, and the maximum was obtained at 700 °C. This was mainly due to the generation and disappearance of the negatively charged groups (Yang et al. 2022). Below 700 °C, the increase in the temperature promoted the consumption and transformation of organic matter and aluminum phosphate after incorporation of KOH, and thus more Si–O− and negatively charged organics and (K/Si/Al)-O− bonds occurred. With the continuous increase in the calcination temperature from 700 °C to 900 °C, the excessive temperature resulted in destroying some of the negatively charged groups again, and thus the zeta potential value of OSSC-2800 and OSSC-2-900 decreased (Wang and Kaskel 2012), which also could be supported by the FTIR results (Fig. 4b).

To analyze the change of pore structural parameters of samples before and after two-step modification, the samples were characterized using the automated surface area and pore size analyzer. As shown in Fig. 4e and Table 2, OSSC had an SBET value of 15.22 m2 g−1 exhibiting a little adsorption of N2 between a relative pressure of 0 ~ 1 (Fig. 4e), which might be related to the dense structure of OSSC and the existence of organic matters (Zhao et al. 2022a, b). In addition, the pore size analysis also showed that OSSC contained a small number of pores, which were predominantly mesoporous (PD = 3.55 nm). After H3PO4-assisted hydrothermal treatment, the pore structural parameters of OSSC-2 enhanced (SBET = 26.06 m2 g−1). At 0 < p/p0 < 0.8, the adsorption amount of N2 gradually increased, indicating that more micropores and mesopores were produced, while the pore size distribution presented a predominance of mesopores (Fig. 4f and the inset, PD = 3.94 nm), which was mainly due to the damage of the aluminum-oxygen octahedral structure of kaolinite and feldspar in OSSC (Wang et al. 2021a, b). With the increase in the value of p/p0 from 0.8 to 1, the adsorption amount of N2 also continuously increased due to the formation of macropore structure, which might be derived from the disordered accumulation of the generated aluminum phosphate on the nanosheets combining with the analysis of Fig. 2 and Fig. 3 (Lu et al. 2020). Subsequently, the overall adsorption amount of N2 increased significantly at a relative pressure of 0–1 after hydrothermal treatment followed by KOH activation at 700 °C, and the specific surface area sharply increased to 199.29 m2 g−1, suggesting that the number and size of pores of OSSC-2-700 were obviously improved. Based on the results of Fig. 4a–d, the formation of micro- and mesopores (p/p0 < 0.8) was due to the consumption of quartz and residual organic matter during calcination, as well as the transformation of kaliophilite. In contrast, the large pores (0.8 < p/p0 < 1) were related to the collapse of the silica-oxygen tetrahedral structural units and the dissociation of the laminated tetrahedral structural groups due to the breakage of the Si–O bonds (Sun et al. 2008). Furthermore, the pores were still mesoporous (inset of Fig. 4f, PD = 3.47 nm), and the number of macropores in OSSC-2-700 seemed to be less than that of OSSC-2 (Fig. 4f) due to the decomposition of AlPO4 during KOH activation (Dai et al. 2018). This result also could be confirmed by comparing with the morphologies of OSSC-2 and OSSC-2-700 (Fig. 5).

Table 2 Pore structural parameters of OSSC, OSSC-2 and OSSC-2-700
Full size table
Fig. 5
figure 5

Morphology changes before and after KOH activation at 700 °C: a OSSC-2; b OSSC-2-700 and c partial enlarged image of OSSC-2-700

Full size image

As shown in Fig. 5a, TEM images of OSSC-2 presented a thin lamellar structure loaded with large numbers of AlPO4 nanoparticles. After being calcined at 700 °C, the AlPO4 nanoparticles completely disappeared due to the synergistic effect between the calcination and incorporation of KOH (OSSC-2-700, Fig. 5b). It also provided the evidence for the statement that AlPO4 nanoparticles affect the number of macropores. In addition, the etched pits on the nanosheets of OSSC-2700 were more obvious than those of OSSC-2 (Fig. 5c), which contributed to the increase in the number of pores and specific surface area (Dai et al. 2018). Thus, the incorporation of H3PO4 and KOH during two-step modification improved the aggregation phenomenon of OSSC and promoted the conversion of organic and inorganic matters to produce more abundant negatively charged groups. This might also be the reason for the higher adsorption of OSSC-2-700 toward MB and TC (Fig. 1).

3.4 Adsorption performance of OSSC-2-700

In order to test the adsorption performance of the as-prepared OSSC-2-700, MV and MG were also selected to be the model organic contaminants besides MB and TC. The adsorption capacities of OSSC-2-700 toward four pollutants were investigated and analyzed at different contact times and initial concentrations, respectively. Figure 6a indicates the effect of contact time on adsorption property of OSSC-2-700, and the concentration of organic contaminants is 200 mg L−1. It was obvious that the adsorption capacities of four organic contaminants increased rapidly with the contact time when the contact time was below 90 min, indicating that a large number of adsorption sites on OSSC-2-700 were gradually occupied by organic pollutant molecules (Dong et al. 2020). However, when the contact time was below 180 min, the adsorption amount continued to gently increase, which suggested that the adsorption of OSSC-2-700 to organic contaminants started to approach the saturation state (Dong et al. 2020). After adsorption for 240 min, the adsorption of OSSC-2-700 to four organic contaminants completely saturated, and the maximum adsorption capacities were 145.05 mg g−1, 107.39 mg g−1, 138.83 mg g−1 and 384.40 mg g−1 for MB, TC, MV, and MG, respectively. In order to further analyze the adsorption types, the kinetics models of pseudo-first-order (Eq. 3), pseudo-second-order (Eq. 4) and Avrami fractional order (Eq. 5) were introduced to fit the adsorption data (Fig. 6b, c and Additional file 1: Fig. S6), and the corresponding parameters are listed in Additional file 1: Table S3.

Fig. 6
figure 6

a Effect of contact time on the adsorption properties of OSSC-2-700 toward four organic contaminants, b pseudo-first-order model and c pseudo-second-order model for fitting the adsorption data. d Effect of initial concentration of four organic contaminants on the adsorption properties of OSSC-2-700, e Langmuir model and (f) Freundlich model for fitting the adsorption data

Full size image
$${{Q}}_{{t}}={{Q}}_{{c}}\left({1} - {{e}}^{-{{k}}_{1}{{t}}}\right)$$
(3)
$${{Q}}_{{t}}=\frac{{\text{K}}_{2}{{{Q}}}_{{c}}^{2}{{t}}}{{1} + {\text{K}}_{2}{{{Q}}}_{{c}}{{t}}}$$
(4)
$$\it {\text{Q}}_{\text{t}} = {\text{Q}}_{\text{c}}\left\{\text{1-exp}{\left[-\left({\text{K}}_{3}{\text{t}}\right)\right]}^{\text{n}}\right\}$$
(5)

where, Qc and Qt (mg g−1) are the adsorbed amounts of MB, MV, MG and TC (mg g−1) at equilibrium and time t, respectively. K1, K2 and K3 are the rate constants obtained by fitting with pseudo-first-order, pseudo-second-order models and Avrami fractional order models, respectively. By contrast, the pseudo-second-order was more fitting to describe the adsorption process of MB, TC, and MV on OSSC-2-700 than the pseudo-first-order with high R2 values of 0.9627 (MB), 0.9909 (TC), and 0.9924 (MV). It suggested that the adsorption rate of OSSC-2-700 for MB, TC, and MV was mainly controlled by chemisorption processes rather than physical adsorption (Yang et al. 2020). In contrast, it presented a good fit using pseudo-first-order (R2 = 0.9912) and pseudo-second-order models (R2 = 0.9772) for the adsorption process of MG. This phenomenon revealed that the adsorption of OSSC-2-700 to MG was the result of a combination of physical and chemical adsorption (Wang et al. 2021a, b). Furthermore, the fitting of the Avrami fractional order model of OSSC-2-700 toward MG was also excellent (R2 = 0.9756), which indicated that MG undergone multiple kinetic processes during the adsorption on OSSC-2-700. It also explained why both the pseudo-first-order model and the pseudo-second-order model exhibited the excellent fitting results for the adsorption process of MG (de Azevedo et al. 2023).

Figure 6d depicts the effect of different initial organic pollutant concentrations on the adsorption properties of OSSC-2-700 as the adsorption time is 240 min. Obviously, the adsorption amounts of organic pollutants rapidly increased with the increase in the initial concentration of the pollutions at the initial stage. It was due to the increased concentration gradient of adsorbates between bulk solution and adsorbent with increasing initial concentration, which provided a relatively stronger driving force to accelerate the diffusion of adsorbates from solution to adsorbent (Yu et al. 2021a, b). When the initial concentration of them increased to 600 mg L−1, the adsorption capacities of OSSC-2-700 for MB, TC and MV reached the maximum value (165.30 mg g−1 for MB, 145.89 mg g−1 for TC and 159.02 mg g−1 for MV). However, the adsorption capacity of MG continued to increase until the initial concentration of the solution reached 2000 mg L−1, and the maximum adsorption capacity was 2046.35 mg g−1. As the concentration increased to 3000 mg L−1, the adsorption capacity of OSSC-2-700 for MG almost achieved the absorption equilibrium, which suggested that the available adsorption sites were completely occupied (Pokharel et al. 2020; Yu et al. 2021a, b). This result also could be proved by the decrease in the removal ratios of organic pollutant molecules with increasing initial concentrations (Additional file 1: Fig. S5). It was obvious that the adsorption capacity of OSSC-2-700 to MG was much greater than other three organic pollutants. To analyze the reasons for this phenomenon, different isothermal models were introduced to fit the adsorption data including Langmuir (Eq. 6) and Freundlich (Eq. 7):

$${{Q}}_{{e}} = {{ Q}}_{{c}}\times{\text{K}}_{{L}}\times\frac{{{C}}_{{e}}}{{1} + {\text{K}}_{{L}}\times{{C}}_{{e}}}$$
(6)
$${{Q}}_{{e}} = {\text{ K}}_{{F}}\times{{{C}}_{{e}}}^{\frac{1}{{{n}}}}$$
(7)

where the Qe is the experimental maximum adsorption capacity, KL and KF are the constants associated with the affinity of the adsorption sites using Langmuir model and Freundlich model, respectively. n represents the supportive force of the adsorption process, Qc (mg g−1) refers to the theoretical adsorption capacity at equilibrium. As shown in Fig. 6e, the Langmuir model was closer to describe the adsorption behavior of MB, TC, and MV on OSSC-2-700 compared with the Freundlich model according to the R2 values of 0.9941, 0.9797, and 0.9948 (Additional file 1: Table S4), respectively (Fig. 6f). This result revealed that the adsorption of MB, TC, and MV onto the surface of OSSC-2-700 belonged to the monolayer coverage mechanism, and the adsorption happened at a specific uniform location in the adsorbent (Wang et al. 2015). On the contrary, the fitting results of MG adsorption indicated that the adsorption behavior of OSSC-2-700 to MG could be well described by the Freundlich model with a higher R2 value of 0.9872 than that of Langmuir model (R2 = 0.9157). It revealed that the adsorption of OSSC-2-700 to MG might be attributed to a multilayer adsorption on a local non-uniform surface. Meanwhile, the mutual adsorption might occur between MG molecules on the surface of MG (Yu et al. 2021a, b; Yu et al. 2021a, b). Therefore, the adsorption behavior of MG on the adsorption sites of OSSC-2-700 depended on the physical and chemical adsorption, in which the physical adsorption was predominant (Wang et al. 2021a, b). Compared with the similar biochar/clay mineral composites, the as-prepared OSSC-2-700 exhibited the excellent adsorption capacity for organic pollutants from aqueous solution (Table 3), suggesting its great potential for removal of organic pollutants.

Table 3 Comparison the adsorption capacity of OSSC-2-700 to MB, MV, MG and TC with the similar biaochar/clay mineral composites
Full size table

3.5 Possible adsorption mechanisms of OSSC-2–700 to organic pollutants

In order to clearly understand the possible adsorption mechanisms of OSSC-2-700 for the involved organic contaminants, this section analyzed and discussed them in terms of energy and structure. The site energy distribution theory (SED) not only demonstrates the distribution of adsorption sites on the adsorbent surface, but also allows counting the number and energy of these sites (Zhou et al. 2019). Therefore, the SED model was introduced to analyze the formation of the interaction between the adsorbent and the adsorbate. Since the Langmuir model could well describe the adsorption behavior of OSSC-2-700 toward MB, TC and MV (Fig. 6e), while the Freundlich model was more suitable for describe the adsorption of MG on OSSC-2-700 (Fig. 6f), the Langmuir-based SED function (Eq. 11) was obtained after associating with Eq. 6 and Eqs. 8–10, and the Freundlich-based SED function (Eq. 12) was deduced combining Eqs. 7–10. The relevant parameters are summarized in Additional file 1: Table S5, and the fitting results are depicted in Fig. 7a and b.

Fig. 7
figure 7

The analysis of SED model (a, b) and statistical physical model (c, d). Zeta potentials (e) and FT-IR spectra (f) of OSSC-2-700 before and after adsorption of organic contaminants

Full size image
$${{C}}_{{e}}={{C}}_{{s}}{{exp}}\left[\left({{E}}_{{s}}{-E}{)}/{\text{RT}}\right)\right]={{C}}_{{s}}{{exp}}\left[-\left({\text{E}}^{*}/{\text{RT}}\right)\right]$$
(8)
$${\text{F}}\left({\text{E}}^{*}\right)={\text{-dQ}}_{\text{e}}\left({\text{E}}^{*}\right)/{\text{dE}}^{*}\text{(}{\text{C}}_{\text{e}}/\text{RT)}*{\text{dQ}}_{\text{e}}\left({\text{E}}^{*}\right)/{\text{dC}}_{\text{e}}$$
(9)
$$\int \limits_{0}^{+\infty}{\text{F}}\left({\text{E}}^{*}\right){\text{d}}{\text{E}}^{*}={\text{Q}}_{\text{m}}$$
(10)
$${\text{F}}\left({\text{E}}^{*}\right)={\text{Q}}_{\text{m}}{{\text{K}}}_{\text{L}}{{\text{C}}}_{\text{s}}/{\text{RT}}{\text{exp}}\left({\text{-E}}^{*}/{\text{RT}}\right){\left[\text{1} + {\text{K}}_{\text{L}}{{\text{C}}}_{\text{s}}{\text{exp}}\left({\text{-E}}^{*}/{\text{RT}}\right)\right]}^{-2}$$
(11)
$${\text{F}}\left({\text{E}}^{*}\right)={\text{(K}}_{\text{F}}{{\text{C}}}_{\text{S}}^{\text{n}}\text{n/RT)exp}\left(\text{-n}{\text{E}}^{*}/{\text{RT}}\right)$$
(12)

As shown in Fig. 7a, the value of E* was negatively correlated with Ce, indicating that the organic pollutant molecules were mainly adsorbed at high-energy sites at low concentrations, and then gradually diffused to low-energy sites (Lu et al. 2021). The curves of F(E*) to (E*) demonstrates the distribution of site energies, and it was clear that the energy sites of four pollutants were normally distributed on OSSC-2-700 (Fig. 7b). It indicates that the energy and number of adsorption sites were not uniformly distributed (Zhao et al. 2022a, b). Moreover, the area covered by the OSSC-2-700-MG curve is much larger than that of other three pollutants, and it implies that the theoretical adsorption of MG molecules by the adsorbent was much higher than that of MB, TC and MV (Rama Krishna and Philip 2008), which also was experimentally proven (Fig. 6d). Moreover, the Em* is the extreme point of F(E*), so (Em*, F(Em*)) can demonstrate the number and energy of active sites on the adsorbent that are more favorable for capturing the pollutants. The variation of the curve of F(E*) indicates that the values of E* for MB, MV, TC and MG were 37.79 kJ mol−1, 20.51 kJ mol−1, 28.69 kJ mol−1, and 35.68 kJ mol−1, respectively, with small energy differences, while the values of F(Em*) for MB, TC, MV and MG were 18.08, 14.96, 25.80 and 45.44 (mg-mol) (g kJ) −1, respectively, and the number of optimal adsorption sites for MG was much higher than other three pollutants. It suggested that the large number of adsorption sites was responsible for the adsorption of MG provided by the adsorbent (Su et al. 2020).

Furthermore, the analysis of statistical physical models was introduced to further explore the formation of the interaction between the adsorption sites and organic pollutants. The isothermal thermodynamic analysis indicated that the adsorption of OSSC-2-700 to MB, TC and MV was mono-layers, while the adsorption of MG was multi-layers. Therefore, a statistical physical model based on monolayer and multilayer adsorption was introduced, as shown in Eqs. 13 and 14, in which Dm, n, C1, C2 denote the number of adsorption sites on the adsorbent, the number of pollutant molecules captured by a single adsorption site, the concentration at half-saturation associated with the first layer and more than one layer, respectively (Sellaoui et al. 2017). In addition, the value of the n parameter provides the adsorption direction of the pollutant molecules on the adsorbent surface and the aggregation of pollutant molecules in the aqueous solution. Generally, it indicates that the adsorption is a multi-interaction process (n < 1) in the monolayer adsorption process, i.e., at least two adsorption sites were required to capture a pollutant molecule; For n > 1, the adsorption was a multimolecular process, i.e., multiple pollutant molecules can be captured at one adsorption site (Sellaoui et al. 2021). In the multilayer adsorption process, there is also a relationship that n < 0.5 corresponds to parallel anchoring, 0.5 < n < 1 implies the mixed anchoring, and n > 1 represents non-parallel anchoring in addition to holding for the above conclusion (Thomas 2021). Figure 8c exhibits that the Dm values of OSSC-2-700-MG (9.957) were much larger than other organic pollutants of 0.842 (MB), 0.513 (TC), 0.451 and (MV), suggesting that a large number of sites on OSSC-2-700 were available for the immobilization of MG molecules (Pang et al. 2019), which was consistent with the results of SED theory. As shown in Fig. 7d, the n values for OSSC-2-700-MB, OSSC-2-700-TC and OSSC-2-700-MV were 0.574, 0.676 and 0.815, respectively. It could be inferred that the adsorption of MB, TC and MV was a multi-interaction process in monolayer adsorption (Xue et al. 2022). In contrast, the n value of OSSC-2-700-MG was 0.541, and thus the adsorption of OSSC-2-700 toward MG molecules was a multi-interaction process during multilayer adsorption, and the adsorbed MG molecules were mixed and immobilized on the adsorbent as well. In other words, the adsorbed MG molecules were parallelly and non-parallelly anchored onto the adsorbent surface (Zhao et al. 2022a, b).

Fig. 8
figure 8

Schematic diagram of the possible adsorption mechanisms of OSSC-2-700 toward organic contaminants using MG as an example

Full size image
$${\text{Q}}_{\text{e}}={\text{n}}{\text{D}}_{\text{m}}/\text{[1+}{\left({\text{C}}_{1}/{\text{C}}_{\text{e}}\right)}^{\text{n}}\text{]}$$
(13)
$${\text{Q}}_{\text{e}}={\text{n}}{\text{D}}_{\text{m}}\left[{\left({\text{C}}_{\text{e}}/{\text{C}}_{1}\right)}^{\text{n}}+2{\left({\text{C}}_{\text{e}}/{\text{C}}_{2}\right)}^{\text{n}}\right]/\text{[1+}{\left({\text{C}}_{\text{e}}/{\text{C}}_{1}\right)}^{\text{n}}\text{+}{\left({\text{C}}_{\text{e}}/{\text{C}}_{2}\right)}^{{2}{\text{n}}}\text{]}$$
(14)

As depicted in Fig. 7e, the zeta potential of OSSC-2-700 also changed obviously after the adsorption of the four organic pollutants. The zeta potential of OSSC-2-700 was − 46.93 mV before adsorption, and then it transformed to + 1.71 mV and + 7.79 mV after adsorption of cationic MB and MV, respectively, suggesting that MB and MV were adsorbed on OSSC-2-700 (Dong et al. 2020). Combining the results of the pseudo-second-order fittings, it could be safely inferred that the strong electrostatic attraction between positively charged pollutants and the negatively charged adsorbent induced the adsorption of OSSC-2-700 to dyes, which might result in the formation of a large number of chemical bonds (Zhao et al. 2022a, b). Subsequently, the adsorption continued due to the presence of excess MB and MV molecules in the solution, and this process also involved some physical interactions (Wang et al. 2021a, b), such as Van der Waals’ force and hydrogen bonds between the adsorbent and the adsorbate. As a result, some of the cationic groups on the last captured MB and MV were retained when the adsorption was complete, resulting in an overall positive zeta potential value. In contrast to MB and MV, the zeta potential value of OSSC-2–700 remained negative after adsorption of TC, and no obvious difference was even observed in their zeta potential values. This result suggested that the adsorption process of TC did not involve the direct combination between positive and negative charges to neutralize the excess negative charge on the adsorbent surface. According to the kinetic fitting results, the adsorption of TC was mainly due to the formation of new coordination bonds after the lone pair of electrons carried by the ionic or molecular groups on the adsorbent occupied the empty orbitals of the TC groups (Dong et al. 2019). Similarly, the zeta potential of OSSC-2-700-MG also presented a negative value, and the coordination bonds were also formed during the adsorption of MG. In addition, there were physical adsorption and multilayer adsorption in this process. Therefore, it could be concluded that the first layer adsorption was mainly chemisorption driven by the coordination interaction in the adsorption process of MG, and then the physical adsorption was dominant for the following multi-layer adsorption (Khoury and Schoop 2021). It might be due to the negative charge field of OSSC-2-700 causing the polarization within the MG molecules to change its dipole moment, thus making it much easier to form hydrogen bonds, H–π interactions, π–π stacking among of MG molecules (Perez 2012).

To corroborate the conclusions of the zeta potential analysis and further explore the adsorption mechanisms, FTIR spectra of OSSC-2–700 before and after adsorption were collected and compared (Fig. 7f). The absorption peaks of OSSC-2–700 at 1624 cm−1, 1026 cm−1, 690 cm−1, 470 cm−1 and 444 cm−1 were due to the vibration of Si–OH and organics-OH, stretching vibration of Si–O bond, Si–O–Si bending vibrations and the characteristic absorption peaks of kaliophilite vibrations (see the details of Fig. 4b). After adsorption of organic contaminants, the characteristic absorption peaks of MB, TC, MV, and MG were observed in their FTIR spectra (Wang et al. 2019a, b), including the stretching vibration of the C–N bond at 1492 cm−1 (MB), the vibration of the aromatic ring backbone at 1454 cm−1 (TC), the stretching vibration of the C–H bond at 1395 cm−1 (MB), the stretching vibration of the C–H bond at 1377 cm−1 (MV), and the stretching vibration of the C–C bond at 1169 cm−1 (MV). It indicated that these pollutants were successfully captured by OSSC-2-700 (Yu et al. 2021b). In addition, there were significant shifts for the vibrational peaks of the Si–O and kaliophilite bonds after adsorption from 1026 cm−1 to 1038 cm−1 (or 1036 cm−1 and 1029 cm−1), 690 cm−1 to 688 cm−1 (or 685 cm−1 and 692 cm−1), 470 cm−1 to 454 cm−1, and 444 cm−1 to 439 cm−1 (or 436 cm−1), respectively. It suggested that the mineral adsorption sites were mainly derived from unsaturated Si/Al/K–O bonds on the quartz and kaliophilite skeletons, which carried negative charge to promote more physical and chemical adsorption (Dong et al. 2019). The newly emerging absorption peaks indicated the chemisorption of organic matters to pollutants, such as the newly formed peaks at 2918 cm−1 and 2851 cm−1 for C–S bonds (superimposed with –CH2, MB) and S–O bonds (superimposed with –CH3, MB), C–N bonds (1377 cm−1 and 1168 cm−1, MV and MG), etc. (Zhang et al. 2019). These new chemical bonds suggested that the carbon and oxygen elements of organic matters interacted with the nitrogen and sulfur elements of pollutants, which reaffirmed the conclusions of the relevant kinetic fittings. Meanwhile, the shifts of O–H groups were observed for OSSC-2-7000-MG from 3433 cm−1 and 1624 cm−1 to 3428 cm−1 and 1628 cm−1, respectively. That was to say, the adsorption behavior of OSSC-2-700 to organic contaminants involved various physical interactions consisting of H-bonds, H–π interactions, π–π stacking for adsorption of MG (Leodopoulos et al. 2014). These findings further explained the results of zeta potential analysis.

In summary, the adsorption behavior of the four organic contaminants on OSSC-2-700 was a complex process, including both physical and chemical adsorption. The SED theoretical analysis confirmed that there were a large number of active adsorption sites on OSSC-2-700, especially MG, which provided a strong driving force for the adsorption process. The results of statistical physical modeling further indicated that the active adsorption sites of OSSC-2-700 captured all of contaminants in the form of multi-interaction process, in which MG molecules covered the adsorbent surface in a different form from other organic contaminants. In addition, the results of FTIR spectra and zeta potential analysis also predicted various possible binding modes such as H-bonding, Van der Waals force and chemical bonding. Finally, the process of MG adsorption by OSSC-2-700 was used as an example, and a schematic diagram of the approximate adsorption mechanism was shown based on the above analysis (Fig. 8).

3.6 Cycling adsorption stability

Reusability is an important parameter to assess the properties of an adsorbent. In this section, MG solution with a concentration of 500 mg L−1 was chosen to study the cycling stability of OSSC-2-700. The spent OSSC-2-700 after adsorption of MG was first washed with deionized water followed by the alternating washing using anhydrous ethanol until colorless, and then the sample was dried at 70 °C for subsequent cyclic adsorption experiments. As shown in Fig. 9, the sample exhibited a good adsorption performance (912.82 mg g−1) in the first cycle, and then the adsorption capacity decreased slightly with the increase in the number of adsorption cycles. Nevertheless, the adsorption capacity of 676.45 mg g−1 was still available after the fifth cycle, indicating that the adsorbent presented a good cycling adsorption stability to remove organic pollutants.

Fig. 9
figure 9

Adsorption capacity of OSSC-2-700 toward MG at different cycle times

Full size image

4 Conclusions

In this work, novel biochar/clay mineral nanocomposites were fabricated using oil shale semi-coke solid waste as precursor by H3PO4-assisted hydrothermal treatment followed by KOH activation. The results indicated that the dispersion, pore structural parameters, and zeta potential of OSSC-2-700 were significantly improved compared with the raw OSSC. The adsorption experiments confirmed that OSSC-2-700 exhibited good adsorption performance toward cationic MB, TC, MV and MG with the maximum adsorption capacity of 165.30 mg g−1, 159.02 mg g−1, 145.89 mg g−1 and 2137.36 mg g−1, respectively. The kinetic fittings suggested that the adsorption behavior of OSSC-2-700 to MB, TC, and MV was mainly attributed to the chemi-sorption, while the adsorption on MG was the result of both physical and chemical adsorption. The isothermal fitting results showed that the adsorption of MB, TC and MV occurred as monolayer adsorption on a locally homogeneous surface, while the adsorption of MG was multilayer adsorption. Furthermore, the theoretical analysis suggested that a large number of active adsorption sites existed on the surface of OSSC-2-700, which contributed to enhancing the adsorption properties to organic contaminant molecules. Combining the theoretical simulations and characterization results, the adsorption process for MB, TC and MV was mainly due to the strong electrostatic attraction to prompt the formation of new chemical bonds between the adsorbent and the adsorbate. In the case of adsorption of MG, it was controlled by various strong–weak interactions including electrostatic attraction, hydrogen bonds, H–π interactions and π–π stacking. Beyond all doubt, this study provided a feasible strategy for the preparation of clay mineral/biochar nanocomposites with excellent adsorption performance for removal of organic pollutants based on full-components resource utilization of OSSC, which also proposed new insights into the interaction mechanisms between clay mineral/biocarbon nanocomposites and organic pollutants.