Persistent and non-biodegradable organic substances, such as organic dyes, are dangerous to the ecosystem and human health [1]. Due to their complex structure, removing dyes from industrial wastewater by traditional methods including adsorption, coagulating sedimentation, chemical oxidation, and biological digestion is not effective [2, 3]. Photocatalysis is an oxidative process that has been established as one of the most useful and environmental friendly methods for the degradation of dye pollutants [4]. Among the materials used for the photocatalytic degradation of dye pollutants, semiconducting layered double hydroxides (LDH) with easily tunable chemical composition have attracted increasing attention as a class of anionic clay materials, because of their high adsorption and photocatalytic activity. A diverse range of transition metal ions can be stabilized in the octahedral sites of the LDH lattices [5, 6, 32]. Especially, ZnCr-LDH exhibited a fairly high photocatalytic activity for the UV light-induced generation of orange Ⅱ and 4-chlorophenol [7], and for the photodegradation of the dyes (rhodamine B, rhodamine 6G) under visible light [8]. However, the application of ZnCr-LDH materials is limited by their aggregation tendency, weak adsorption, and difficult separation from wastewater [9].

Sepiolite, a hydrated magnesium silicate clay mineral, is fibrous, has a theoretically large surface area, and high chemo-mechanical stability [10], which are related to the powerful adsorbent properties for organic or inorganic molecules and ions [1]. Sepiolite has been studied as an effective material for the photocatalytic treatment of pollutants due to its unique properties and structures as well as its abundance and low cost [11]. For example, a TiO₂-supported sepiolite composite showed high photocatalytic activity for gaseous formaldehyde degradation under UV light, and a positive synergistic effect on TiO₂ photocatalysis was observed [12]. Akkari et al. [13] prepared ZnO/Fe₃O₄-sepiolite composites for degrading methylene blue dye under UV light and the composites exhibited good photoactivities and superparamagnetic properties. Thus, sepiolites are ideal carrier materials for the ZnCr-LDH catalyst, which can improve catalyst adsorption capacity and separation from wastewater.

In this study, acid-activated sepiolite (Sep) served as a carrier for ZnCr-LDH, which was assembled on the surface of the Sep by in situ co-precipitation to prepare the composites (Sep@LDH). To determine the photocatalytic performances of the prepared Sep@LDH, methyl orange (MO) and methylene blue (MB) were chosen as model pollutants present in the same solution. The structure, morphology, texture, and optical features of the composites were determined by X-ray diffraction (XRD), Scanning electron microscopy (SEM), N₂ adsorption-desorption, Thermogravimetry and differential thermal gravity (TG/DTG), Ultraviolet-visible diffuse reflectance spectra (DRS) and Photoluminescence (PL) analyses. The photocatalytic performances of the composites were evaluated and compared for degradation of the two co-existing dyes under visible-light irradiation. The photodegradation mechanism of Sep@LDH was proposed based on the contribution of specific scavengers and kinetics of the photocatalytic reaction of co-existing MO and MB. Moreover, to the best of our knowledge, few reports have been published regarding the photocatalytic performances of Sep@LDH photocatalysts under visible light.

MO, MB, and other reagents were of analytical grade (AR) and used without further purification, as-received from Nan-Jin Chemical Reagent Co., Ltd. (China). All solutions were made with deionized water, and 0.1 mol∙L^–1 NaOH and HCl solutions were used for pH adjustments. A pH electrode (Mettler Toledo S40K) was used for pH measurements.

Raw sepiolite was purchased from Xiangtan YuanYuan Sepiolite New Materials Co. Ltd. (Hunan, China), and the physicochemical properties of the sepiolite are listed in Table S1. The sepiolite was treated before the experiments as follows: the suspension containing 10 g·100 mL^-1 sepiolite was mechanically stirred for 5 h in a HCl solution (1.0 mol·L^-1). After waiting for approximately 3 min, the supernatant suspension was filtered, washed until a neutral pH was reached with deionized water to remove the excess Cl^-, and then dried at 80 ℃ for 24 h. The solid sample was ground and sieved using a 200-mesh sieve for further experiments, and the pre-acidified sepiolite was denoted as Sep.

The sepiolite@LDH composites were prepared using in situ co-precipitation (Scheme 1). First, solution A (200 mL) was prepared by mixing solutions of Zn and Cr nitrates with a Zn/Cr molar ratio of 1.0 (Zn²⁺ + Cr³⁺ = 1.2 mol∙L^-1). Solution B was prepared by dissolving NaOH (1.75 mol∙L^-1) and Na₂CO₃ (0.75 mol∙L^-1) in 200 mL of deionized water, and 100 mL suspension solution containing different amounts of Sep was prepared. Subsequently, the A and B solutions were added simultaneously to the Sep suspension solution at 60 ℃ under vigorous stirring for 3 h. The precipitate was aged at 80 ℃ for 12 h, filtered and washed until the pH was neutral, and subsequently dried at 90 ℃ for 24 h to obtain the resulting sample that was denoted Sep@LDH. For convenience, the samples produced with different Sep masses (1.0, 2.0, 4.0, 6.0, and 10.0 g) in the 100 mL suspension solution were designated as Sep₁@LDH, Sep₂@LDH, Sep₄@LDH, Sep₆@LDH, and Sep₁₀@LDH, respectively.

For comparison, the ZnCr-LDH samples with Zn/Cr molar ratios of 0.5, 1.0, and 2.0 were also prepared by co-precipitation. The preparation procedure was similar to that of the Sep@LDH without the Sep suspension, where solutions A and B (Zn²⁺ + Cr³⁺ = 1.2 mol∙L^-1) were added simultaneously to a beaker under vigorous stirring. In preliminary work, the ZnCr-LDH with a Zn/Cr molar ratio of 1.0, designated as LDH, was determined to be the best candidate due to its low band-gap energy in the visible-light region, large surface area, and weak electron-hole recombination (Figs. S1 and S2). It also exhibited the highest photocatalytic activity in the solution of co-existing MO and MB dyes (Fig. S3). Thus, solution A was prepared by mixing an equal molar ratio of Zn²⁺ and Cr³⁺ for the preparation of Sep@LDH.

The photocatalytic performance of the prepared samples was evaluated in synthetic wastewater containing equal concentrations of MO and MB under visible light as a model reaction. Various amounts of catalyst in synthetic wastewater were magnetically stirred in the dark for 30 min to reach adsorption-desorption equilibrium, and were then exposed to visible light from a 300W Xe lamp equipped with a UV cutoff filter (λ ≥ 420 nm) at a specific time. A 3mL aliquot of the reaction solution was withdrawn by a syringe at given time intervals and centrifuged to remove the catalyst for dye concentration measurement. A control reaction was performed following the same procedure without adding any catalyst. All results were obtained and reported with reference to the corresponding controls. The concentrations of the two dyes were determined by measuring the absorbance at characteristic wavelengths by UV-vis spectrophotometry (Hitachi U-2910), where the γ_max values were 465 nm for MO and 665 nm for MB (Fig. S4). The MO (or MB) removal ratio (degradation and adsorption ratio) was determined using the following equation:

where η is the removal ratio (%), C_o is the initial concentration of MO (or MB), and C_t is the concentration (mg·L^–1) at time t (min).

Powder XRD measurements were performed using a Japan Rigaku D/max 2550PC (λ = 1.5405 Å ) instrument with Cu K_α irradiation. The scan step was 0.02° (2θ) with a filament intensity of 30 mA and a voltage of 40 kV. SEM (JEOL JSM-6700F) was used to characterize the surface morphology of the samples. The specific surface area, pore volume, and diameter of the samples were determined by the Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Hallender (BJH) methods using a NOVA-e1000 system (Quantachrome, USA). TG/DTG were performed under a nitrogen atmosphere using a Seiko 6300 TG/DTG instrument with a heating rate of 10 ℃∙min^-1 under a He stream (60 mL∙min^-1). UV-visible DRS were measured using a UV-visible spectrophotometer (Shimadzu UV-2550, Kyoto, Japan). PL spectra were recorded at room temperature using a F-4600 spectrophotometer at an excitation wavelength of 290 nm.

The XRD spectra measured for the pure Sep, virgin LDH, and Sep@LDH composites are shown in Fig. 1. The virgin LDH displayed a typical of ZnCr-LDH structure with sharp and intense (003), (006), (009), (110), and (113) reflections, and no other crystalline phase was observed [5]. The XRD patterns of the pure Sep and Sep@LDH composites clearly showed a sepiolite phase, indicating that the sepiolite structure was maintained despite the less intense reflections with an unstable baseline. The XRD patterns of the Sep@LDH composites indicated the presence of ZnCr-LDH and sepiolite. For the Sep₁₀@LDH sample, reflections distinctive of sepiolite were observed, albeit with weak reflections from the crystalline phase of the ZnCr-LDH, implying that the sepiolite structure was the main phase. Compared to the Sep₁@LDH, Sep₂@LDH, and Sep₆@LDH composites, the Sep₄@LDH sample exhibited more intense reflections with a more stable baseline, especially for the (110) and (113) reflections which were more clearly distinguished. This indicated that the Sep₄@LDH sample had a more regular layered structure and higher crystallinity [14].

Fig. 2 depicts the morphology of the pure Sep and Sep@LDH composites, where the ZnCr-LDH lamellar structure and sepiolite filler are clearly distinguished in the SEM images. The pure Sep exhibited a fibrous morphology and numerous fibers were congregated into bundles. From the SEM images of the Sep@LDH composites, ZnCr-LDH was observed to assemble on the surface of the Sep. The Sep₁₀@LDH exhibited a morphology featuring a ZnCr-LDH lamellar structure, and an articulated fiber structure of sepiolite was observed due to the large amount of Sep. With decreasing Sep content, significant amounts of LDH particles were observed in the Sep₁@LDH, Sep₂@LDH, Sep₄@LDH, and Sep₆@LDH samples, where the Sep fibers appeared to be fully coated by the LDH lamella. Especially for the Sep₄@LDH sample, the SEM image showed a typical "sand-rose" morphology characteristic of hydrotalcite-like compounds with interconnected hexagonal lamellar plates almost uniform in size, which was consistent with the XRD results that showed Sep₄@LDH had the highest crystallinity of all samples. This indicated that a balanced amount of ZnCr-LDH coating on the surface of the Sep fibers caused a decrease in agglomeration of the composites.

To examine the textural characteristic of the prepared photocatalysts, BET gas sorptometry measurements were used to characterize their pore volume (V_P), pore diameter (D_P), and specific surface area (S_BET). The adsorption isotherms of the pure Sep, virgin LDH, and Sep@LDH composites are shown in Fig. 3. Except for Sep₄@LDH, all samples exhibited a typical H3 hysteresis loop in the high relative pressure (P/P₀) range of 0.60 to 0.97. The pure Sep and Sep₁₀@LDH exhibited similar adsorption isotherm curves with very small type H3 hysteresis loops from 0.80 to 0.97 (P/P₀) corresponding to a type Ⅱ isotherm according to IUPAC classification, indicating the presence of tubular pores [15, 31]. For the virgin LDH, Sep₁@LDH, and Sep₂@LDH, isotherms with typical H3 hysteresis loop in a relatively high pressure range of 0.70 to 0.97 (P/P₀) was observed due to aggregates of plate-like particles causing slit-like pores with nonuniform size and largely without micropores, belonging to typical type Ⅳ isotherms of mesoporous materials [2]. Only the Sep₆@LDH sample showed a hysteresis loop in the range of 0.60 to 0.97 (P/P₀) corresponding to a type Ⅳ isotherm, indicating the presence of pores formed by the parallel plates. Especially for the Sep₄@LDH sample, the isotherm with a hysteresis loop between type H3 and H4 was observed at a moderate relative pressure from 0.4 to 0.97 (P/P₀) corresponding to a type Ⅳ isotherm, indicating a broad pore size distribution and mesoporous structure. The pore formation was more likely caused by coverage of the LDH particles on the surface of the Sep fibers. As seen in Table 1, the corresponding surface areas (S_BET) of the samples suggested that the LDH coating with lamellar structure exerted a small impact on the S_BET of the Sep@LDH. The Sep₄@LDH composite showed the highest surface area (148 m²∙g^-1), likely due to the presence of LDH with well-dispersed particles on the surface of the Sep fibers. These results indicated the mesoporous nature of Sep@LDH, and that appropriate coating of ZnCr-LDH on the surface of the Sep fibers caused an increase in the S_BET of the composites.

The TG/DTG curves of the pure Sep, virgin LDH, and Sep@LDH composites are shown in Fig. 4. The pure Sep was heated in a multistep dehydration process from 25 to 700 ℃. The weight loss was attributed to the removal of surface-adsorbed water, zeolite water, coordinated water, and hydroxyl groups. Another degradation step was observed due to the phase transformation of sepiolite to enstatite (MgSiO₃) at >700 ℃, which was in good agreement with the literature [16, 17]. For the virgin LDH, the thermal decomposition was evidenced by two well-differentiated peaks, which corresponded to the loss of surface and interlayer water (the first endothermic peak at 50–180 ℃) and hydroxyl groups from the brucite sheets and the interlayer CO₃^2– anions for the second peak (180–400 ℃) [18, 19]. The TG and DTG curves of the Sep@LDH composites demonstrated similar pyrolysis behavior, where the endothermic step of the hydroxyl groups in the brucite sheets and interlayer CO₃^2– anions became gradually blurred with increasing coating amount of Sep. On the other hand, the total weight loss of the pure Sep was approximately 14.0 wt% at 700 ℃, while the total weight loss of the virgin LDH was approximately 25.6 wt%. For the Sep@LDH composites, the total weight loss increased with increasing LDH, and the Sep₁@LDH exhibited the highest total weight loss of 25.2 wt%. Thus, it can be concluded that ZnCr-LDH was present in the Sep@LDH composites.

UV-Vis DRS measurement is very simple, and can be used as a sensitive measure of changes in the molecular structure of materials. The UV-Vis DRS spectra of the pure Sep, virgin LDH, and Sep@LDH composites are shown in Fig. 5(A), where their band gaps were also estimated. The pure Sep exhibited a weak peak at 200–800 nm, while the virgin LDH showed three strong absorption bands at ˂320, 380–500, and 500–800 nm. This indicated that the ZnCr-LDH had strong absorption in the visible-light region due to the photoexcitation of the CrO₆ octahedron at approximately 430 nm and the d-d transition of Cr³⁺ at approximately 600 nm [8, 18]. Compared to the pure Sep, a new peak appeared at 600 nm for the Sep₁₀@LDH, which indicated the presence of Cr³⁺. The Sep@LDH composites, except for Sep₁₀@LDH, exhibited a new peak at approximately 430 nm, indicating the presence of polymeric species with CrO₆ octahedral coordination. Moreover, the band gap transition of the composites was determined from the plots of (αhν)² as a function of (hν) based on the expression (hν)² = A(hν - E_g) [20], and the related curves were plotted in Fig. 5(A) (inset). The band gap energies of the virgin LDH, Sep₁@LDH, Sep₂@LDH, Sep₄@LDH, and Sep₆@LDH were 2.30, 2.29, 2.27, 2.22, and 2.25 eV, respectively, suggesting that the virgin LDH and Sep@LDH composites showed potential for photocatalytic decomposition of MO and MB under visible light. It should be noted that the band-gap energies of the Sep@LDH composites were lower than that of virgin LDH, which can be attributed to the effect of the sepiolite carrier on the LDH crystal structure. Meanwhile, Sep₄@LDH exhibited the lowest band-gap energy among the prepared composites, suggesting that a balanced amount of the LDH increased the photoexcitation of the CrO₆ octahedron because of its high crystallinity and large surface area. The results were in good agreement with those of the XRD and SEM analyses.

PL spectra were measured to investigate the luminescence properties of the virgin LDH and Sep@LDH composites and are shown in Fig. 5(B). The main emission peaks for all samples were centered at approximately 350 nm corresponding to the band gap of ZnCr-LDH [21], and the PL spectra agreed well with the UV-Vis spectra. The emission can be attributed to band-band PL phenomena where the electron transitions from the CB bottom to VB top occurred. The emission of the band gap transition was equal to the band gap energy, consistent with previous research [22]. The highest PL emission intensity was observed for the Sep₁₀@LDH, due to its low amount of LDH coating. Stronger peak intensities of the Sep@LDH composites, except for that of Sep₄@LDH, were observed compared with that of the pure LDH. It should be noted that the Sep₄@LDH sample exhibited the largest extent of quenching. This phenomenon was attributed to its high crystallinity and dispersity, leading to many active sites on the surface of the Sep₄@LDH composite, enhancing electron-hole separation/transport [23]. The results suggested that a suitable LDH coating significantly enhanced the electron-hole separation/transport, where the abundant electron-hole pairs (active sites) promoted carrier transport efficiency, leading to high photocatalytic activity towards the visible light-driven decomposition of MO and MB. This is further discussed in the following section.

The photocatalytic activities of the composites were examined by monitoring the time-dependent degradation of MB and MO under visible-light irradiation (λ > 420 nm). As shown in Fig. 6, the adsorption of MB and MO on the samples contributed to their degradation. An MO adsorption ratio of 21.2% and MB adsorption ratio of 45.7% was observed for the pure Sep, while the MO and MB adsorption ratios were 40.4% and 20.7%, respectively, for virgin LDH after 30 min in the dark. When the LDH coating amount increased stepwise, the adsorption of MO gradually dropped, but the adsorption of MB increased on the Sep@LDH composites. On the other hand, it was clear that pure Sep showed little photocatalytic activity towards MO, while it exhibited weak catalytic activity towards MB (15.2% degradation ratio) under visible light after 120 min. However, after coating with LDH, the Sep@LDH composites generally showed enhanced photocatalytic activity. Specifically, the Sep₄@LDH sample showed superior photocatalytic activities (removal ratios of 86.9% for MO and 81.5% for MB) under visible light after 120 min. In contrast, the virgin LDH only exhibited a removal ratio of 73.7% for MO and 61.8% for MB under the same conditions. The Sep₄@LDH composite exhibited the highest photocatalytic activity due to its large surface area, high crystallinity of the ZnCr-LDH formed on the surface of Sep, and low band-gap energy (2.22 eV). Thus, the appropriate LDH coating amount on the surface of Sep enhanced the photocatalytic activity, verifying the analytical UV-Vis DRS and PL results. To further optimize Sep₄@LDH, the experiments were performed by varying the catalyst loading from 0.25 to 2.0 g·L^-1 (Fig. S5). The highest photocatalytic efficiency was observed at a loading of 1.25 g·L^-1 Sep₄@LDH, which was selected to further evaluate the photocatalytic performance of the Sep₄@LDH composite. For practical application, the effect of the initial concentration of the dye substrates on the photocatalysis was also investigated, and the results are shown in Fig. 7. It was clearly demonstrated that the initial concentrations of MO and MB did not have a significant effect on the time required to reach equilibrium, and their removal ratios increased with decreasing initial concentrations. When the initial concentrations of MO and MB were 10 mg·L^-1, the corresponding removal ratios were 99.0% and 93.3%, respectively, after 120 min.

To quantitatively understand the reaction mechanisms, the photocatalytic degradation of co-existing MB and MO was studied as a function of time over the Sep₄@LDH catalyst, and was found to follow the pseudo first-order kinetic model. The kinetics were well-described by the Langmuir-Hinshelwood (L-H model) which could accurately represent the experimental data of the photocatalytic reaction systems [24-26]. From the data obtained in this study (Fig. 7), the kinetic parameters were calculated according to the L-H model and are listed in Table 2. The calculated correlation coefficient (R²) obtained from the model was higher than 0.96 (R² > 0.96), indicating that the L-H model was suitable for describing the photocatalytic reaction. As seen in Table 2, higher k values indicate better removal efficiency of the co-existing MO and MB under visible light, showing a similar trend where the catalytic activity increased with decreasing initial concentrations of MO and MB. To further investigate the reaction mechanism, the kinetic parameters of the photodegradation reaction over the virgin LDH and all Sep@LDH composites were calculated using the L-H model based on the data shown in Fig. 6 and Table S2. As seen in Table S2, the kinetic investigation confirmed that the photodegradation of MO and MB followed the L-H model for all studied catalysts (R² > 0.95). The results suggested that the active sites significantly influenced the rate of photodegradation, which verified the UV-Vis DRS and PL analyses. The classical L-H model was used under the hypothesis that the surface-active sites of the photocatalysts were energetically homogeneous [27] and confirmed that the moderate LDH coating on the surface of the Sep formed copious and homogeneous active sites leading to better photocatalytic activity.

Catalyst recycling has often been problematic because of instability and leaching; therefore, a recycling experiment was performed. Each run was performed in the dark for 30 min, then exposed under visible light for 120 min, following which the Sep₄@LDH catalyst was reused for the next run under the same conditions. A total of five runs using the filtered catalyst were performed with washing to remove the reaction solution attached to the catalyst, and then drying at 80 ℃ for reuse (Fig. 8). Slight leaching of the active Cr species was detected, with Cr concentrations of 5×10^-6 mol∙L^-1 in the first cycle, 1×10^-6 mol∙L^-1 in second cycle, and no Cr detected after the third cycle. The catalytic activity (94.4% for MO and 90.2% for MB) declined mainly due to Cr leaching during the second cycle. In subsequent cycles, the catalyst was recycled without an obvious decrease in activity and was maintained at over 86.2% for MO and 84.9% for MB in the fifth cycle. These results suggested that the Sep₄@LDH catalyst has strong potential for good photocatalytic activity and stability.

To obtain further insight into the photocatalytic mechanism of Sep₄@LDH, the reactive oxygen species in the photocatalytic process were investigated by adding scavengers into the reaction system (Fig. 9). The formation of intermediate active species, such as superoxide anion radicals (•O₂^-), hydroxyl radicals (•OH), and photogenerated holes (h⁺) during the photoreaction process, and their role, were indirectly investigated using three quenchers, i.e., benzoquinone (BQ, an •O₂^- scavenger), disodium ethylenediaminetetraacetate (Na₂-EDTA, a h⁺ scavenger), and tert-butanol (t-BuOH, an •OH scavenger) [28, 29]. Upon introducing t-BuOH, the photocatalytic activity was almost completely quenched, implying that hydroxyl radicals (•OH) were the dominant reactive species contributing to the photodegradation of co-existing MO and MB. When Na₂-EDTA was added, the h⁺ holes were scavenged, but •OH could still participate in the photo-degradation process, so the photocatalysis was only partially suppressed. With the addition of BQ, most of the MO and MB were degraded, indicating that •O₂^- radicals played a very minor role in the photocatalytic degradation. Based on the above results, a photodegradation mechanism of co-existing MO and MB by the Sep₄@LDH composite under visible-light irradiation was proposed as follows,

(1)

(2)

(3)

(4)

(5)

(6)

(7)

Upon visible-light excitation, photo-induced electrons and holes were generated, and h⁺ reacted with adsorbed H₂O to form •OH radicals which degraded the dyes (MO and MB). The e^--absorbed dioxygen on the surface of the catalyst produced •O₂^- radicals, which were then reduced to •OH via disproportionation of •O₂^- [30]. Lastly, the MO and MB molecules were oxidized and decomposed by the •OH radicals. Therefore, according to these findings, it can be concluded that •OH radicals played a critical role in the photocatalytic degradation process of co-existing MO and MB by the Sep@LDH composites (Scheme 2).

Sep@LDH composites were successfully synthesized by a facile in situ co-precipitation method and showed photocatalytic activities toward the degradation of co-existing MO and MB under visible light, where the acid-activated Sep was as an excellent carrier. The composites were characterized by XRD, SEM, BET/BJH, TG/DTG, UV-Vis DRS, and PL methods, indicating that ZnCr-LDH was coated on the Sep surface successfully. A moderate amount of coating of the ZnCr-LDH generated a good distribution of the Sep@LDH particles with high crystallinity and large surface area, which increased the number of active sites on the surface of Sep@LDH resulting in high photocatalytic activity. Among the Sep@LDH composites, Sep₄@LDH exhibited the best photocatalytic performance. In addition, the degradation kinetics were fitted to a pseudo first-order kinetic model, which confirmed that the active sites were a crucial factor in the photodegradation. The good degradation of the dye pollutants was attributed to the generation of highly oxidative •OH radicals at the active sites and electron/hole pairs during the photodegradation reaction.