1 Introduction

Due to the extensive demand and utilization of organic dyes in textile, cosmetics, papermaking, and other industries, there are high concentrations of toxic organic dyes in wastewater, which have serious impacts on the environment and human health. As one of the common anionic dyes (Mu'azu et al. 2018), RR120 contains azo (–N=N–) groups, resulting in low biodegradability, and therefore poses many risks to the ecosystem and human health (Irem et al. 2013; Rahman et al. 2013). The high concentration of RR120 leads to skin irritation, dermatitis, carcinogenicity, mutagenicity, and toxicity to aquatic life. Hence, how to remove organic dyes before they are discharged into the water environment is extremely important for water treatment.

Numerous approaches include adsorption (Yagub et al. 2014), nanofiltration (Gerçel 2016), ion exchange (Liu et al. 2007), electrochemical catalysis (Zhang et al. 2019), and Fenton degradation (Hou et al. 2016) have been widely used to remove organic pollutants in wastewater. Compared to other technologies, adsorption is considered as one of the most efficient methods for removing organic dyes because of its easy operation, low cost, and high efficiency (Jang et al. 2018). Recently, biochar (BC) based materials as cost-efficient adsorbents have attracted much attention (Laird 2008; Zhang et al. 2017). The surfaces of most of the original biochars are usually negatively charged, which is related to their rich oxygen-containing functional groups, thus showing highly efficient adsorption of cation pollutants (Wang et al. 2017). However, most biochars have limited adsorption capacity for anionic dyes such as RR120, partly due to the lack of positive active sites on their surface (Yang et al. 2019; Zheng et al. 2020). This has greatly hindered the use of biochar to remove RR120 from aqueous solutions. Thus, improving the anionic dye adsorption capacity of biochar has become an important direction to expand the environmental applications of biochar technology.

Recently, metal oxide/biochar composites have been increasingly applied in adsorption for various water pollutants (Li et al. 2018; Yao et al. 2011). Chemical modification by loading metal oxides on the surface of biochar can enhance its surface properties by adding more sorption sites and adjusting its surface charges, thereby improving its adsorption capacity (Zheng et al. 2020). Ball milling has been applied to fabricate carbon/metal oxide composites due to its low cost, environmental friendliness, and high efficiency (Kumar et al. 2020). On the one hand, ball milling can grind biochar into powder, which can decrease particle size, increase specific surface areas, and introduce surface functional groups on biochar, thus enhancing its adsorption ability (Cai et al. 2018; Lyu et al. 2020). On the other hand, ball milling of metal oxide can introduce lattice defects (such as oxygen vacancies) on its surface (Aggelopoulos et al. 2017). The presence of oxygen vacancies can change the state of the adsorbed molecules (e.g. bond length, bond angle, coordination mode, or intermediate), thereby increasing the affinity to the molecules (Zhao et al. 2020). Thus, it is very meaningful to adjust the adsorption properties of materials with the fabrication of oxygen vacancies by ball milling.

In this work, a series of oxygen vacancy-rich BiOCl/biochar nanocomposites were prepared by solvent-free ball-milling of BiOCl particles and biochar. A series of adsorption experiments were conducted to evaluate the RR120 adsorption capacity of composite materials. The ball-milling parameters such as time, speed, and ratio of balls-to-BiOCl were investigated. The physicochemical properties of BiOCl/biochar nanocomposites were characterized. Moreover, the adsorption behaviors and mechanisms for RR120 on BiOCl/biochar nanocomposites were studied in particular. The presented investigation provided new ideas in the field of biochar-based adsorbents.

2 Experimental section

2.1 Catalyst preparation

The bamboo was cut into small pieces and washed with deionized water for further use. Fast pyrolysis of bamboo at 800 °C was conducted to produce biochar (Hien et al. 2020). The obtained biochar was washed thoroughly with deionized water and dried at 60 °C overnight for further use.

For the preparation of BiOCl, Bi(NO3)3·5H2O (0.02 mol) and KCl (0.02 mol) were added into ethylene glycol (90 mL, AR 98%) with stirring to obtain a homogeneous solution. After magnetic stirring for 30 min, the mixture was transferred to a hydrothermal container (100 mL) and was heated in an oven at 160 °C for 12 h. After heating, the resulting precipitate was centrifuged and alternately washed three times each by ethanol and distilled water. The obtained sample was named as BiOCl.

The nanocomposites of BiOCl/BC were made by the facile ball milling method. The procedures of ball-milling were as follows: mixtures (4 g) of BiOCl and biochar with different mass percentages of BiOCl (1%, 10%, 30%, 50%, 60%, and 80%) were put in 250 mL ball milling jars and then 200 g balls were added into each jar. The ball milling was operated at a rotational speed of 400 rpm for 9 h and the rotation direction was altered every 1.5 h. A series of BiOCl/BC nanocomposites with different mass percentages of BiOCl (1%, 10%, 30%, 50%, 60%, and 80%) were obtained and marked as 1%-BiOCl/BC, 10%-BiOCl/BC, 30%-BiOCl/BC, 50%-BiOCl/BC, 60%-BiOCl/BC, and 80%-BiOCl/BC. The diameter of 50%-BiOCl/BC nanoparticles is about 0.2–1 μm. The BM-BiOCl and BM-BC were also prepared with the same procedures in the absence of biochar or BiOCl, respectively.

2.2 Catalyst characterization

Powder X-ray diffraction measurements were characterized by X-ray diffractometer (Bruker D8 Advance-A25, Germany) in a scanning range of 5°–80° (2 theta) with Cu Ka (λ = 0.15418 nm) radiation. Fourier-transform infrared spectroscopy (FT-IR) spectra of samples were measured from 400 to 4000 cm−1 on an FT-IR spectrometer (Bruker VERTEX 70, Germany). The surface microstructure was obtained using a Nova Nano SEM-450 (FEI, USA) field emission scanning electron microscopy (FE-SEM). Transmission electron microscopy (TEM, JEOL 2100F, Japan) was employed to observe the internal structural morphology. X-ray photoelectron spectroscopy (XPS) measurements were acquired on an X-ray photoelectron spectrometer (Thermo Scientific ESCALAB 250, USA) with an Al Kα radiator. All the binding energy values were calibrated by the C 1s peak at 284.8 eV. Brunauer–Emmett–Teller (BET) surface area and pore size measurements were tested by nitrogen adsorption via an adsorption–desorption instrument (Quantachrome Autosorb-iQ, USA). Electron paramagnetic resonance (EPR) spectra were recorded at room temperature using a Bruker A300 (Germany, Bruker). Zeta potential data were measured with a 90Plus PALS high sensitivity zeta potential and particle size analyzer (Brookhaven, USA).

2.3 Batch adsorption experiment

The optimum ball milling parameters were investigated by batch adsorption experiments. The preparation details of BM-BiOCl under different ball-milling conditions can be seen in the Additional file 1 (S1). For each experiment, adsorbent (dosage of 0.4 g L−1) was added into 15 mL of RR120 solution (10 mg L−1) under mechanical shaking at 25 °C for 24 h. After shaking, the suspension was filtered with 0.22 μm filter membrane (MCE, Troody). RR120 concentration was measured at 533 nm by UV mini-1285 (Shimadzu, Japan). To make the experimental results more accurate, all experiments were repeated three times.

The sample with the highest RR120 removal rate (50%-BiOCl/BC) was selected to further investigate the adsorption behavior of RR120 on BiOCl/BC nanocomposites. The adsorption capacity (qe, mg g−1) and removal efficiency (R%) can be calculated by Eqs. (1) and (2):

$${q}_{e}=\frac{({C}_{0}- {C}_{e})V}{m}$$
(1)
$$\mathrm{R}= \frac{{C}_{0}- {C}_{e}}{{C}_{0}} \times 100\%$$
(2)

where C0 and Ce are the initial and ultimate detected concentrations (mol L−1) of the RR120 solution, respectively, and V (L) and m (g) represent the volume of RR120 solution and the mass of adsorbent.

3 Results and discussion

3.1 Characterization of BiOCl/BC nanocomposites

The crystal structures of the as-prepared samples were studied by XRD. As shown in Fig. 1a, the XRD pattern of the BM-BC showed a low-intensity diffraction peak around 26.6°, corresponding to hexagonal carbon (JCPDS NO. 26-1076). The XRD pattern of BM-BiOCl exhibited the characteristic diffraction peaks at 11.9°, 26.0°, 32.5°, 33.6°, 40.9°, 46.8°, and 58.6°, corresponding to (001), (101), (110), (102), (112), (200), and (212) crystal planes of tetragonal phase of BiOCl (JCPDS No. 85-0861), respectively. The diffraction peaks of BiOCl also existed in the XRD patterns of BiOCl/BC nanocomposites, indicating that the formation of the nanocomposites did not significantly change the crystallinity of BiOCl. The peak of biochar almost disappeared in the composites, due to the good dispersion of biochar in composites.

Fig. 1
figure 1

XRD patterns (a) of BM-BiOCl, 60%-BiOCl, 50%-BiOCl, 30%-BiOCl and BM-BC; FTIR spectra (b), N2 adsorption–desorption isotherms (c), and pore volume distribution curves (d) of BM-BiOCl, 50%-BiOCl/BC, and BM-BC

Full size image

Figure 1b shows the FTIR spectra of the as-prepared BM-BiOCl, 50%-BiOCl/BC, and BM-BC to better understand the chemical structure of the prepared BiOCl/BC nanocomposites. The observed peaks in BM-BC were allocated as: at 3459 cm−1 (O–H vibrational stretching and N–H stretching vibration) (He et al. 2019); 2812 cm−1 (asymmetric CH2 stretching) (Amjed et al. 2020); 1600 cm−1 (O–H bending) (Micheal et al. 2019); 1400, 1362 cm−1 (O–H stretching) (Tang et al. 2021); 814, 755 cm−1 (aromatic C–H bending) (Mingke et al. 2019). The absorption peaks at 2830 (Boukaoud et al. 2021), 1600, 1362, and 772 cm−1 (Wang et al. 2015) in BM-BiOCl can be attributed to the O–H stretching, O–H bending, O–H stretching, and the asymmetrical stretching vibration of the Bi–O bond, respectively. The similar FTIR spectra of BM-BiOCl and 50%-BiOCl/BC indicate that the structures of BiOCl are not changed after the introduction of BC.

The nitrogen adsorption–desorption isotherms (Fig. 1c) show that the adsorption isotherm of BM-BiOCl was type IV isotherm with H3 hysteresis loop according to the IUPAC classification (Kruk and Jaroniec 2001; Sing 1985), implying that it comprised of aggregates of platelike particles forming slitlike pores. The isotherms of BiOCl/BC nanocomposites and BM-BC did not enclose, which may attribute to the existence of a large amount of very narrow slit pores or bottle-shaped pores in the biochar (Wang et al. 2018). Pore diameter analysis (Fig. 1d) showed that BM-BiOCl had many pores ranging from 2 to 4 nm, while BM-BC and 50%-BiOCl/BC exhibited many micropores which is due to the pore structure of biochar. The BET surface area, pore size, and pore volume of BM-BC, BM-BiOCl, and 50%-BiOCl/BC were estimated by N2 adsorption–desorption isotherms. As shown in Table 1, the specific surface area of 50%-BiOCl/BC (62.15 m2 g−1) was lower than that of BM-BC (163.7 m2 g−1) but higher than that of BM-BiOCl (43.96 m2 g−1), indicating that the introduction of BC increased the specific surface area of the BiOCl/BC nanocomposites compared to BM-BiOCl, which may enhance the adsorption capacity of RR120.

Table 1 Specific surface areas and pore characteristics of the BM-BiOCl, 50%-BiOCl/BC, and BM-BC samples
Full size table

The SEM images (Fig. 2a, b) of 50%-BiOCl/BC showed that the diameters of nanocomposite nanoparticles were about 0.2–1 μm. EDS analysis (Fig. 2e–i) proved that the C, O, Bi, and Cl elements were uniformly distributed in 50%-BiOCl/BC. To further examine the microstructure of 50%-BiOCl/BC, TEM and HRTEM images were obtained and shown in Fig. 2c and d. Small BiOCl nanoparticles with diameters around 10–25 nm were distributed in biochar. Clear lattice fringes of 0.275 nm can be observed corresponding to the (110) crystal facets of BiOCl. The TEM and HRTEM analysis further proved that BiOCl nanoparticles were evenly distributed on BC.

Fig. 2
figure 2

SEM images (a, b) of 50% BiOCl/BC; TEM image (c) and HRTEM image (d) of 50%-BiOCl/BC; the EDS mapping images (f–i) of 50%-BiOCl/BC image (e)

Full size image

The XPS measurement was performed to reveal the surface composition and chemical state of as-prepared samples. Figure 3a shows that the 50%-BiOCl/BC was composed of the elements of Bi, O, Cl, and C. The O 1s XPS spectrum (Fig. 3b) of BM-BiOCl can be resolved into three peaks. The binding energies at 529.9 eV, 530.8 eV, and 531.9 eV can be assigned to Bi–O bond, oxygen vacancies, and adsorbed oxygen such as hydroxyl groups or O2 molecules, respectively (Lu et al. 2019; Yang et al. 2020). In the O 1s of BM-BC, the binding energies at 532.3 eV and 533.8 eV can be assigned to adsorbed oxygen and C–O bond (Quan et al. 2020). The O 1s spectrum of 50%-BiOCl/BC can be convoluted into four peaks: Bi–O bond (530.3 eV), oxygen vacancies (531.1 eV), adsorbed oxygen (532.1 eV), and C–O bond (532.8 eV). The peaks of O 1s of 50%-BiOCl/BC shifted to the higher binding energy compared to BM-BiOCl and to lower binging energy compared to BM-BC, indicating the interactions between BiOCl and BC in the composite. Figure 3c shows two peaks in Cl 2p of BM-BiOCl spectrum at 198.0 eV and 199.6 eV can be assigned to Cl 2p3/2 and Cl 2p1/2, respectively, which were characteristic peaks of Cl atoms in the BM-BiOCl (Cai et al. 2018). Two peaks in Bi 4f spectrum of BM-BiOCl at 159.3 eV and 164.6 eV (Fig. 3d) belong to Bi 4f7/2 and Bi 4f5/2, respectively (Wu et al. 2017). Compared with BM-BiOCl, the characteristic peaks of Cl and Bi in 50%-BiOCl/BC moved to higher binding energy, indicating the electron density on the surface of Bi and Cl decreased in the composites, which is also due to the strong interactions between BiOCl and biochar in composites.

Fig. 3
figure 3

Survey spectra (a) and O 1s spectra (b) of BM-BiOCl, 50%-BiOCl/BC, and BM-BC; Cl 2p (c) and Bi 4f (d) spectra for BM-BiOCl and 50%-BiOCl/BC

Full size image

In addition, electron paramagnetic resonance (EPR) measurements were carried out to detect surface oxygen vacancies, because the presence of oxygen vacancy leads to obvious electron paramagnetic resonance (EPR) signals when g = 2.003 (Zhang et al. 2020). As shown in Fig. 4, the 50%-BiOCl/BC showed a strong characteristic EPR signal, indicating the existence of large amounts of oxygen vacancies. The concentration of oxygen vacancies in 50%-BiOCl/BC was higher that in BiOCl, BC (Liu et al. 2017; Sun et al. 2018), and BM-BC samples, which may provide more active sites for adsorption.

Fig. 4
figure 4

EPR spectra of BC, BM-BC, BiOCl, and 50%-BiOCl/BC

Full size image

3.2 Comparison of the best ball milling parameters and the adsorption capacity of different ratios in composites

Because BiOCl was the modification component in nanocomposites, ball milling parameters were optimized based on the RR120 adsorption capacities of BiOCl. BM-BiOCl powders with different ball-milling times, ball-milling speeds, and mass ratio of ball to BiOCl were prepared, and then subjected to batch adsorption experiments to obtain the best ball milling parameters. The RR120 adsorption capacity of BM-BiOCl samples under different ball milling conditions are shown in Fig. 5a–c. The mass ratio of ball to BiOCl had little effect on the adsorption capacity of RR120 on BiOCl, while the ball milling speed and ball milling time showed greater influences. The optimum milling parameters were as follows: milling time was 9 h, the mass ratio of balls to BiOCl was 50:1, and milling speed was 400 rpm. Subsequently, ball milling samples were obtained under the optimum milling parameters.

Fig. 5
figure 5

RR120 sorption capacity on BiOCl with different milling time (a), mass percentage (b), and speed (c); effect of BiOCl content (d) on RR120 adsorption on BiOCl/BC nanocomposites

Full size image

Figure 5d shows the effect of different mass percentages of BiOCl to BC on the adsorption capacity of RR120. Compared with BM-BC, all the BiOCl/BC nanocomposites showed higher removal rate of RR120, which reflected that the adsorption capacity of biochar to RR120 was significantly enhanced after ball milling with BiOCl. The removal rate of RR120 by BiOCl/BC nanocomposites with the increase of BiOCl content, and the removal rate reached the highest at 50%-BiOCl/BC. However, further increase of the content of BiOCl in BiOCl/BC nanocomposites reduced the RR120 adsorption capacity of nanocomposites. This might be due to the aggregations of BiOCl nanoparticles in the composites when BiOCl contents were further increased.

3.3 Adsorption behaviors

The parameter of pH is a key factor that affects the dye adsorption process, because it controls the degree of ionization of adsorbed molecules and the surface charge of the adsorbent. Figure 6a describes the effect of initial solution pH on the adsorption of RR120 on 50%-BiOCl/BC. It was observed that with the rising of pH from 2 to 10, the removal rate gradually decreased from 62.5 to 1.6%. The highest adsorption capacity of RR120 dye was observed at solution pH 2. The pHpzc (point of zero charge) of 50%-BiOCl/BC was tested to be 2.1 (Additional file 1: Fig. S1) (Azalok et al. 2021). Errais et al. measured aqueous solution of RR120 dye by potentiometric titration, the results showed that RR120 dye still had anionic morphology at acidic pH (pH < 4) (Errais et al. 2012). Therefore, when the solution pH was below 2.1, the surface of 50%-BiOCl/BC can be converted to positively charged, thus promoting the uptake of negatively charged species (RR120) via electrostatic attraction (Jawad et al. 2020a, b). With the pH increase, the electrostatic attraction gradually weakened, while hydrogen bond and π–π interactions were the main functional mechanisms. In summary, 50%-BiOCl/BC exhibited better adsorption ability at lower pH for RR120 dye.

Fig. 6
figure 6

The effect of solution pH on the adsorption of RR120 (a); kinetics of RR120 adsorption on 50%-BiOCl/BC (vs time) (b); kinetics of RR120 adsorption on 50%-BiOCl/BC (vs square root of time) (c); isotherms of RR120 on 50%-BiOCl/BC (d). The lines are simulations of various models

Full size image

The efficiency of adsorbents was characterized by adsorption kinetics. Figure 6b shows the experimental data and model fitting of adsorption kinetics. It can be observed that the adsorption kinetic rate of RR120 on 50%-BiOCl/BC increased sharply in the first 1 h, then slowed down with time, reaching adsorption equilibrium within 4 h. Table 2 summarizes the two different adsorption kinetics models’ best model parameters and corresponding correlation coefficients. Compared with the pseudo-first-order model (R2 = 0.794), the pseudo-second-order model (R2 = 0.920) fitted better. This suggests that the adsorption of RR120 dye on the 50%-BiOCl/BC was more likely to be controlled by the chemical adsorption process (Zubair et al. 2020). The curve of adsorption kinetics versus t0.5 showed three straight lines (Fig. 6c), indicating that the adsorption kinetics of RR120 on nanocomposites might be divided into three steps (Lyu et al. 2018). The initial rapid adsorption might be due to the electrostatic attraction between the positively charged 50%-BiOCl/BC surface and the negatively charged RR120, while the later slow adsorption suggested that intra-particle diffusion may be involved (Lalley et al. 2016). Finally, the adsorption of RR120 by the composites hardly changed.

Table 2 Best-fit parameters of the kinetics and isotherm models for RR120 adsorption on 50%-BiOCl/BC
Full size table

Figure 6d shows the adsorption isotherms of RR120 on 50%-BiOCl/BC. It can be seen that 50%-BiOCl/BC showed high adsorption ability of RR120 in a wide range of equilibrium concentrations, reflecting that the 50%-BiOCl/BC had a large adsorption capacity for RR120. The experimental data were simulated with Freundlich and Langmuir models, and the best model parameters were summarized in Table 2. Freundlich (R2 = 0.998) model and Langmuir (R2 = 0.999) both described the isotherm data well with the R2 values close to each other. The Langmuir model fitted slightly better than the Freundlich model, indicating that the single-layer adsorption behavior was dominant (Fan et al. 2019; Jang et al. 2018). The adsorption kinetics and isotherms results showed that 50%-BiOCl/BC could be used as an effective adsorbent for RR120 dye.

To evaluate the stability of 50%-BiOCl/BC in RR120 solution, the XRD patterns before and after the adsorption of RR120 by 50%-BiOCl/BC were measured (Additional file 1: Fig. S2). It can be seen that the XRD patterns of the sample before and after adsorption did not change significantly, which can prove that it was stable in aqueous solution. The corresponding pictures of 50%-BiOCl/BC sample before and after adsorption were barely changed (Inserts in Additional file 1: Fig. S2), which further proved the stability of the adsorbent.

3.4 Adsorption mechanisms

Based on the above discussion of BET surface area, oxygen vacancies, pH effects, adsorption kinetics, and adsorption isotherms, the possible adsorption mechanisms of 50%-BiOCl/BC were analyzed and illustrated in Scheme 1. High specific surface area and porous structures contributed to the adsorption performance of the prepared BiOCl/BC composites. Abundant oxygen vacancies on the surface of the 50%-BiOCl/BC provided a large number of adsorption sites and also improved its adsorption for RR120 dye (Zhao et al. 2020). When pH < pHpzc (pHpzc 2.1), electrostatic attractions existed between negatively charged sulfonate (–SO3−) acid groups of RR120 and the positively charged surface of 50%-BiOCl/BC. Hydroxyl bond can occur between –OH on the surface of 50%-BiOCl/BC and aromatic ring or O and N atoms in RR120 dye (Jawad et al. 2019, 2020c). The benzene rings of RR120 molecules could interact with the benzene ring of biochar in 50%-BiOCl/BC through π–π interactions (Gai et al. 2020). Thus, it can be inferred that high specific surface area, abundant oxygen vacancies, electrostatic effect, hydrogen bond, and π–π interaction determined the adsorption capacity of 50%-BiOCl/BC.

Scheme 1
scheme 1

Schematic diagram of adsorption mechanism for RR120 on 50%-BiOCl/BC

Full size image

4 Conclusion

In this work, the 50%-BiOCl/BC showed high adsorption capacity for RR120 dye (62%), which was much higher than that of BM-BC (20%). Through batch adsorption experiments, the effects of key adsorption parameters such as RR120 concentration, solution pH, and contact time were studied. Langmuir model, Freundlich model, and pseudo-second-order kinetic model described the adsorption process well. The electrostatic effect, π–π interaction, and hydrogen bond might be the adsorption mechanism on the heterogeneous surface, and intra-particle diffusion will be the main limitation of RR120 adsorption on 50%-BiOCl/BC. As a result, the ball-milled BiOCl/BC nanocomposites enhanced physicochemical properties and efficient removal ability for RR120 dye from wastewater.