International Journal of Environmental Research and Public Health

Article Synergy of Photocatalysis and Adsorption for Simultaneous Removal of Hexavalent Chromium and Methylene Blue by g-C3N4/BiFeO3/Carbon Nanotubes Ternary Composites

Huiwen Huo 1, Xinjiang Hu 1,*, Hui Wang 1,2, Jiang Li 3, Guangyu Xie 1, Xiaofei Tan 4,5, Qi Jin 1, Daixi Zhou 1, Chuang Li 1, Guoqiang Qiu 2 and Yunguo Liu 4,5

1 College of Environmental Science and Engineering, Central South University of Forestry and Technology, Changsha 410004, China 2 Faculty of Life Science and Technology, Central South University of Forestry and Technology, Changsha 410004, China 3 School of Architecture and Art, Central South University, Changsha 410082, China 4 College of Environmental Science and Engineering, Hunan University, Changsha 410082, China 5 Key Laboratory of Environmental Biology and Pollution Control (Hunan University), Ministry of Education, Changsha 410082, China * Correspondence: [email protected] or [email protected]; Tel.: +86-0731-85623096



Received: 6 August 2019; Accepted: 30 August 2019; Published: 3 September 2019


Abstract: A novel graphite-phase carbon nitride (g-C3N4)/bismuth ferrite (BiFeO3)/carbon nanotubes (CNTs) ternary magnetic composite (CNBT) was prepared by a hydrothermal synthesis. Using this material, Cr(VI) and methylene blue (MB) were removed from wastewater through synergistic adsorption and photocatalysis. The effects of pH, time, and pollutant concentration on the photocatalytic performance of CNBT, as well as possible interactions between Cr(VI) and MB species were analyzed. The obtained results showed that CNTs could effectively reduce the recombination rate of electron-hole pairs during the photocatalytic reaction of the g-C3N4/BiFeO3 composite, thereby improving its photocatalytic performance, while the presence of MB increased the reduction rate of Cr(VI). After 5 h of the simultaneous adsorption and photocatalysis by CNBT, the removal rates of Cr(VI) and MB were 93% and 98%, respectively. This study provides a new theoretical basis and technical guidance for the combined application of photocatalysis and adsorption in the treatment of wastewaters containing mixed pollutants.

Keywords: hexavalent chromium; methylene blue; adsorption; photocatalysis; synergistic treatment

1. Introduction With the development of many industrial technologies, pollutants discharged into the environment have become increasingly toxic and chemically complex [1]. Generally, industrial wastewaters tend to contain both organic pollutants and heavy metals, and their proper treatment is a very challenging task because it requires the use of special techniques and reaction conditions [2]. Chromium has two valence states in water, Cr(VI) and Cr(III), among which Cr(VI) is highly toxic, water-soluble, mobile, and easy to accumulate. It is allergenic, carcinogenic, teratogenic, and mutagenic as it could interfere with the transcription process of DNA [3], therefore Cr(VI) is classified as a surface water pollutant of the first type. Comparatively, Cr(III) is less toxic and can be easily removed by forming the Cr(OH)3 precipitate under alkaline conditions. Methylene blue (MB) is a commonly used cationic dye that is chemically stable and can affect the growth of aquatic plants and microorganisms, thereby causing

Int. J. Environ. Res. Public Health 2019, 16, 3219; doi:10.3390/ijerph16173219 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2019, 16, 3219 2 of 18 serious ecological problems [4,5]. Cr(VI) and MB simultaneously exist in wastewaters produced during electroplating, printing, and dyeing as well in the leather, metallurgical, cosmetics, and pigments industries [6]. Therefore, the effective treatment of wastewater aimed at the removal of heavy metals and organic pollutants has become the focus of many studies in the field of environmental chemistry [7]. In recent years, various methods for treating such mixed wastewaters have been developed including photocatalysis, adsorption, ion exchange, membrane filtration, chemical precipitation, flocculation, oxidation, and aerobic/anaerobic biological treatment (for low and high concentration organic wastewater, respectively) [8,9]. The advantages of the first two techniques over other methods include their simplicity, lower costs, and higher treatment and Cr(VI) decontamination efficiencies [10]. These technologies are widely used in the industry, and proper selections of the catalyst and adsorbent are the main factor affecting their performance [11,12]. Graphite-phase carbon nitride (g-C3N4) is a typical nontoxic metal-free semiconductor [10]. It consists of only two elements, C and N, which are abundant in the earth core and can be chemically modified through simple inexpensive routes [13]. g-C3N4 is an allotrope of carbon nitride with a stacked two-dimensional structure and band gap of ca. 2.7 eV, which suggests its possible application in sunlight harvesting as well as high chemical and thermal stabilities [14]. The C and N atoms of g-C3N4 2 are sp -hybridized to form highly delocalized conjugated π-bonds [15]. g-C3N4 has been widely used in the photocatalytic reduction of heavy metals [16], the photocatalysis hydrogen production [17], and photodegradation of organic pollutants [18]. However, bulk g-C3N4 has a small specific surface area, very few active sites, and a high electron-hole recombination rate, which do not allow significant enhancement of its photocatalytic performance [19]. Hence, various researches have tried to improve the photocatalytic activity of this material by coupling [20] and surface modification [21]. Bismuth ferrite (BiFeO3) is a photocatalytic material with a wide range of potential applications, which possesses a narrow band gap and is responsive to visible light [22]. Furthermore, it can simultaneously exhibit ferroelectric and magnetic properties at room temperature [21]. Combing g-C3N4 with BiFeO3 to produce the g-C3N4/BiFeO3 (CNB) composites may reduce the recombination rate of photogenerated electron-hole pairs and expand the photo-response range, while the strong magnetic properties of BiFeO3 can be used in the solid-liquid separation by a magnet [23]. In addition, under visible light radiation, BiFeO generates photoexcited electrons that react with H O to form OH, 3 2 · highly reactive free radicals, that can effectively degrade organic pollutants such as MB [24]. However, we also found that the photoreduction rate of Cr(VI) by CNB was not very high, indicating that although the coupling of g-C3N4 and BiFeO3 could improve the photocatalytic efficiencies of these compounds to some extent, the efficiency of the electron transfer between them remained limited. Carbon nanotubes (CNTs) are nanomaterials with a large specific surface area, good electrical conductivity, and low solubility that are widely utilized in adsorption treatment processes [25]. The p-electrons of their C atoms form a large range of delocalized π-bonds with a strong conjugation effect, which promotes the storage and transfer of photoexcited electrons [26]. Therefore, combining g-C3N4, BiFeO3, and CNTs to form a ternary magnetic composite may not only increase the specific surface area and number of active sites [27], but also reduce the recombination rate of photogenerated carriers, thus improving the photocatalytic properties of the resulting material. Although many research groups have investigated various technologies for the treatment of solutions containing the Cr(VI) or MB species, very few of them discussed their simultaneous removal through synergistic photocatalysis and adsorption. Therefore, in this study, g-C3N4, BiFeO3, and CNTs were combined to produce a ternary magnetic composite CNBT, and its synergistic adsorption and photocatalytic properties towards the removal of Cr(VI) and MB from wastewaters were explored. Additionally, the influences of the solution pH, pollutant concentration, and reaction time on the adsorption and photocatalytic processes were determined, and possible mechanisms of the adsorption and photocatalytic reactions were suggested. Int. J. Environ. Res. Public Health 2019, 16, 3219 3 of 18

2. Materials and Methods

2.1. Preparation of Functionalized Ternary Magnetic Composite CNBT

Fe(NO3)3, Bi(NO3)3,K2Cr2O7, melamine, methyl alcohol, ethyl alcohol, ethylene glycol monomethyl ether, and methyl ether were purchased from China National Pharmaceutical Group Chemical Reagent Co. Ltd. (Shanghai, China). Citric acid was supplied by Shanghai Mclean Biochemical Technology Co. Ltd. (Shanghai, China). The chemicals were of the analytical grade. CNTs (Muti Wall Carbon nanotubes, Main Range of Diameter: 10–20 nm, purity > 97%) was purchased from Shenzhen Nanotech Port Co. Ltd. (Shenzhen, China). First, 20.0 g of melamine was washed by distilled water and absolute ethanol in a crucible. The supernatant was removed, and the obtained precipitate was placed into a muffle furnace for dehydration at 70 ◦C in air for 20 min without capping. Subsequently, the crucible was covered and calcined at 600 ◦C for 3 h in air. After cooling at ca. 25 ◦C, the precipitate was ground and sieved (#100 mesh size) to obtain g-C3N4. BiFeO3 was prepared by a sol-gel method [26]. First, 0.08 mol of Fe(NO3)3 and 0.08 mol of Bi(NO3)3 were dissolved in 200 mL of ethylene glycol methyl ether followed by the addition of 0.2 mL of nitric acid (0.1 mol/L) obtained a mixed solution. Next, 0.08 mol of citric acid was dissolved in 100 mL of ethylene glycol and added to the above mixed solution. The obtained mixture was stirred at 60 ◦C for 1 h and then heated to 100 ◦C for 17 h in air. The resulting yellow gel was transferred to a crucible and heated first to 200 ◦C for 1 h and then to 500 ◦C for 2 h in a muffle furnace in air. After cooling at ca. 25 ◦C, the produced solid was ground to obtain the BiFeO3 powder. CNB and CNBT were synthesized by ultrasonically dispersing their components in methanol at specified ratios for 3 h followed by dehydration. The mass ratio of g-C3N4 to BiFeO3 for CNB was 1:1, and that of g-C3N4, BiFeO3, and CNTs for CNBT was 3:3:1.

2.2. Characterization The material microstructure was characterized by a QUANTA250 field emission scanning electron microscope (FE-SEM; FEI, Hillsboro, OR, USA), the samples were directly attached to the conductive adhesive, then were tested with the ETD morphology mode under a 20 KV working voltage and 0◦ angle after gold spraying. Transmission electron microscope (TEM) analyses were applied a Tecnai G2 F20 (TEM; FEI, Hillsboro, OR, USA). The samples were analyzed by a NICOLET 5700 Fourier transform infrared (FTIR) spectrometer (Thermo Nicolet Corporation, Madison, WI, USA) under the 1 1 acquisition range of 400–4000 cm− (number of scans 64, resolution 4 cm− ). Thermogravimetric (TG) and differential scanning calorimetry (DSC) curves were recorded in air from room temperature to 1000 ◦C using an SDT Q600 synchronous thermal analyzer (New Castle, TA, USA) at a flow rate of 100 mL/min and heating rate of 10 ◦C/min. The Brunner–Emmet–Teller (BET) specific surface areas of the studied materials were determined using an ASAP2460 4MP instrument (Norcross, GE, USA). The magnetic properties of the samples were evaluated by a MPMS-XL-7 vibrating sample magnetometer (VSM; Quantum Design Instruments, O’Fallon, MO, USA). The electron spin resonance (ESR) signals produced by spin-trapped radicals were examined on a JES FA200 spectrometer (JEOL, Tokyo, Japan) using the spin-trapping reagent 5,5-dimethyl-1-pyrroline N-oxide (DMPO) under visible light irradiation. Phase compositions of the materials were analyzed by the X-ray diffraction (XRD) on a D/max-2500 (Rigaku, Tokyo, Japan) diffractometer with a Cu Kα radiation, while their elemental compositions were determined by the X-ray photoelectron spectroscopy (XPS, ESCALAB 250XI, Thermo Fisher Scientific, Waltham, MA, USA). The XPS data acquisition was under the standard lens mode, Al Ka source type (hv = 1486.6 eV), 30.0 eV pass energy, 500 µm spot size, C1s 284.8 eV as the reference point, and the energy step size of the full XPS spectra and the characteristic spectra were 1.0 eV and 0.5 eV, respectively. The photoluminescence (PL) spectra were recorded on a FLS 980 fluorescence spectrophotometer (Edinburgh Instruments, Livingston, UK). Ultraviolet (UV)-visible diffuse reflectance spectroscopy (DRS) analyses of the samples were conducted by a diffuse reflectance Int. J. Environ. Res. Public Health 2019, 16, 3219 4 of 18

spectrophotometer (UH4150 UV-vis, Hitachi, Tokyo, Japan) using BaSO4 as a reference under the condition of start wavelength 800 nm, end wavelength 200 nm, scan speed 300 nm/min, lamp change wavelength 350 nm, path length 10 mm, step size 0.50 nm.

2.3. Photoelectrochemical Measurements The photocurrent responses of the studied materials were measured using a photochemical workstation (CHI 660E) with a conventional three-electrode model [28]. g-C3N4, BiFeO3, CNTs, CNB, and CNBT samples were examined by a current-time method, in which the Pt and Ag/AgCl electrodes served as the counter and reference electrodes, respectively. The working electrode was fabricated as follows. First, 20 mg of the sample was suspended in a 2 mL nafion (0.25 vol%) to form a slurry, which was fully dispersed by ultrasonication for more than 2 h. Next, 0.2 mL of the dispersed slurry was uniformly dropped onto the conductive side of the clean FTO glass sheet at 80 ◦C, then used as a working electrode [5]. To obtain a working area of 1 1 cm2, the extra parts of the slurry were scraped × off. The electrolyte of the three-electrode system consisted of a 0.5 mol/L Na2SO4 solution. The light source was a xenon lamp with a power of 300 W (CEL-HXF300, Beijing China Education Au-light Co. Ltd., Beijing, China) containing a UV filter (λ > 400 nm). The distance between the light source and the surface of the working electrode was set to 15 cm. In addition, the light and dark states were alternated in 20-s intervals during the test.

2.4. Photocatalytic Activity Testing To prepare the samples for photocatalytic activity testing, 2.829 g of potassium dichromate was heated in an oven to 110 ◦C for 2 h, after which the annealed powder and 1 g of MB were dissolved separately in ultrapure water inside 1000-mL volumetric flasks to obtain 1 g/L Cr(VI) and MB stock solutions, respectively. Different concentrations of the Cr(VI) and MB solutions used in the experiments were achieved by diluting their corresponding stock solutions. The schematic diagram of the photocatalytic activity test reaction scheme is shown in Figure S3. A 300-W xenon lamp (CEL-HXF300, Beijing China Education Au-light Co. Ltd.) with an ultraviolet filter (λ > 400 nm) was used as a visible light source. Multiple 200-mL mixed solutions containing 5 mg/L of Cr(VI) and 20 mg/L of MB were prepared, and their pH values were adjusted using 0.1 or 1 mol/L HCl and NaOH solutions followed by the addition of 0.5 g of the catalyst. Each solution was first stirred magnetically at a speed of 1000 rpm under dark conditions for 1 h to reach the adsorption-desorption equilibrium between the contaminant and the photocatalyst, and then radiated with visible light for 4 h. The xenon lamp was equipped with an ultraviolet filter, and the beam was directed to the lens to shine vertically on the liquid surface, with a vertical distance of about 15 cm. The solution was multiplied in the 500 mL beaker and placed on a magnetic stirrer at a speed of 1000 rmp for 4 h. In the process, ca. 5 mL of the treated solution was passed through a 0.45-µm filter (Hydrophilic, MCE) to filtration every 30 min. The concentrations of the Cr(VI) and MB species left in the reacted solutions were measured by a UV-visible spectrophotometer (UV2600, UNICO, Shanghai, China) at wavelengths of 540 nm and 665 nm, respectively. To ensure the accuracy of the experimental results, each group of experiments was conducted in triplicate.

2.5. Adsorption Studies After adjusting the pH of the mixed solutions of Cr(VI) and MB with HCl and NaOH, 0.05 g of the composite was added to the conical flasks containing 20 mL of these solutions. The flasks were placed in a rotary shaker and agitated at a constant temperature of 25 ◦C and speed of 150 rpm for 5 h. After the reaction, 5 mL of each solution was passed through a 0.45-µm filter, and the remaining concentrations of Cr(VI) and MB were measured by a colorimetric method at wavelengths of 540 nm and 665 nm, respectively, using the UV-visible spectrophotometer. Each group of experiments was done in triplicate. In addition, the effects of different materials, pH value (2.0–11.0), initial concentrations of the Cr(VI) Int. J. Environ. Res. Public Health 2019, 16, 3219 5 of 18

(1–50 mg/L)Int. J. and Environ. MB Res. (5–200 Public Health mg 201/L)9, 16 species,, x and reaction time (5–300 min) on the adsorption5 of 18 capacity were studied. The adsorbed amount was calculated by the following equation: (C0 - Ce)V qe = (1) W (C 0 Ce)V where qe represents the amount of theq adsorbed= material− at equilibrium (mg/g); C0 and Ce are the (1) e W initial and equilibrium concentrations of Cr(VI) or MB in the solution (mg/L), respectively, V is the solution volume (L), and W indicates the amount of the added material (g) [29]. The adsorption where qe represents the amount of the adsorbed material at equilibrium (mg/g); C0 and Ce are the models applicable under different control conditions are described in the Supplementary Material initial and equilibrium concentrations of Cr(VI) or MB in the solution (mg/L), respectively, V is the section. The different model parameters are listed in Table S1 and Table S2, the fitted models are solution volumeshown in (L), Figure and S1 W and indicates Figure S2, the respectively. amount of the added material (g) [29]. The adsorption models applicable under different control conditions are described in the Supplementary Material section. The different3. Results model and parameters Discussion are listed in Tables S1 and S2, the fitted models are shown in Figures S1 and S2, respectively.3.1. Sample Characterization

3. Results3.1.1. and Electron Discussion Microscopy

The morphologies and structures of g-C3N4, BiFeO3, CNTs, CNB, and CNBT were analyzed by 3.1. Samplethe Characterization scanning electron microscopy (see Figure 1). The g-C3N4 sample depicted in Figure 1a exhibits many irregular small crystals, which are closely packed in layers. As shown in Figure 1b, BiFeO3 3.1.1. Electronrepresents Microscopy a fine-grained crystal with a size of about 100 nm, and CNTs are a dense tubular material as shown in Figure 1c. In the CNB and CNBT composites depicted in Figure 1d,e, respectively, the The morphologiessurface of the original and structures granular BiFeO of g-C3 was3N 4covered, BiFeO with3, CNTs, a number CNB, of CNTs, and CNBT which werereduced analyzed the by the scanning electronsharpness microscopyof BiFeO3 and increased (see Figure the 1specific). The surface g-C 3 Narea4 sample due to the depicted combination in Figure of CNTs1a and exhibits g- many C3N4; g-C3N4 has a lamellar shape with a flat surface and randomly dispersed BiFeO3 particles; and irregular small crystals, which are closely packed in layers. As shown in Figure1b, BiFeO 3 represents a fine-grainedCNTs crystalhave large with aspect a sizeratios ofand about are loosely 100 nm,coupled and to CNTsthe other are two a components. dense tubular material as shown The CNB and CNBT samples were also analyzed by the transmission electron microscopy (the in Figure1obtainedc. In the results CNB are and shown CNBT in Figure composites 2). Figure 2a,c depicted characterize in Figure the microscopic1d,e, respectively, morphology of the the surface of the originalCNB granular material. BiFeOBoth its3 wasg-C3N covered4 and BiFeO with3 components a number are of irregularly CNTs, which block-shaped reduced andthe tightly sharpness of bound, while the color of BiFeO3 is relatively dark. The resolution of the crystal plane is ca. 0.395 nm, BiFeO3 and increased the specific surface area due to the combination of CNTs and g-C3N4; g-C3N4 which is similar to the results of the literature studies [30,31]. Figure 2b,d show that the coupling of has a lamellar shape with a flat surface and randomly dispersed BiFeO3 particles; and CNTs have large linear CNTs with g-C3N4 produces the composite with a large specific surface area, while the aspect ratiosmorphology and are of loosely CNTs does coupled not significantly to the other change. two components.

Figure 1. FigureSEM 1. images SEM images of (a of) graphite-phase(a) graphite-phase carbon carbon nitride nitride (g-C3 (g-CN4), (b3N) bismuth4), (b) ferrite bismuth (BiFeO ferrite3), (c) (BiFeO3), carbon nanotubes (CNTs), (d) g-C3N4/BiFeO3 (CNB), and (e) carbon nanotubes ternary magnetic (c) carbon nanotubes (CNTs), (d) g-C3N4/BiFeO3 (CNB), and (e) carbon nanotubes ternary magnetic composite (CNBT). composite (CNBT).

The CNB and CNBT samples were also analyzed by the transmission electron microscopy (the obtained results are shown in Figure2). Figure2a,c characterize the microscopic morphology of the CNB material. Both its g-C3N4 and BiFeO3 components are irregularly block-shaped and tightly bound, while the color of BiFeO3 is relatively dark. The resolution of the crystal plane is ca. 0.395 nm, which is similar to the results of the literature studies [30,31]. Figure2b,d show that the coupling of linear CNTs Int. J. Environ. Res. Public Health 2019, 16, 3219 6 of 18

with g-C3N4 produces the composite with a large specific surface area, while the morphology of CNTs does not significantlyInt. J. Environ. Res. Public change. Health 2019, 16, x 6 of 18

Int. J. Environ. Res. Public Health 2019, 16, x 6 of 18

FigureFigure 2. TEM 2. TEM images images of (a (a,c,)c CNB,) CNB, (b,d ()b CNBT.,d) CNBT.

3.1.2. FTIR3.1.2. FTIR Figure 2. TEM images of (a,c) CNB, (b,d) CNBT. The FTIR spectral characteristics of g-C3N4 and BiFeO3 were determined in previous studies [26]. The FTIR spectral characteristics of g-C3N4 and BiFeO3 were determined in previous studies [26]. The absorption band of g-C3N4, was located close to 1639 cm−1 due to the C=N stretching vibration The absorption3.1.2. FTIR band of g-C N , was located close to 1639 cm 1 due to the C=N stretching vibration [32], [32], and its absorption3 peaks4 were centered at around 808 cm−−1 and 1200–1600 cm−1, representing the The FTIR spectral characteristics of g-C3N4 and BiFeO3 were1 determined in previous studies1 [26]. and its absorptionout-of-plane peaks bending were modes centered of C-N heterocycles at around and 808 the typical cm− aromaticand 1200–1600 C-N stretching cm− vibrations,, representing the The absorption band of g-C3N4, was located close to 1639 cm−1 due to the C=N stretching vibration out-of-planerespectively bending [26]. modes The absorption of C-N peak heterocycles of BiFeO3 was and centered the typical near 550 aromatic cm−1 [26]. The C-N FTIR stretching spectra ofvibrations, [32], and its absorption peaks were centered at around 808 cm−1 and 1200–1600 cm−1, representing the the CNB and CNBT composites and CNTs are displayed in Figure 3. For CNTs, the1 absorption peak respectivelyout-of-plane [26]. The bending absorption modes peakof C-N of heterocycles BiFeO3 was and centeredthe typical aromatic near 550 C-N cm stretching− [26]. Thevibrations, FTIR spectra of was weaker than those of the other materials and centered at a frequency of 3435 cm−1 characteristic respectively [26]. The absorption peak of BiFeO3 was centered near 550 cm−1 [26]. The FTIR spectra of the CNB andof the CNBT O-H bonds, composites which might and be CNTs due to are the displayed adsorption inof Figurewater molecules3. For CNTs, on the thenanotube absorption surface peak was the CNB and CNBT composites and CNTs are displayed in Figure 3. For CNTs, the absorption1 peak weaker than[9,33]. those The ofbroadband the other spanning materials from and 1000centered to 1400 cm at−1 acorresponded frequency to of the 3435 C=O cm bond− characteristic stretching of the was weaker than those of the other materials and centered at a frequency of 3435 cm−1 characteristic O-H bonds,vibrations which [34]. might The beFTIR due peak to of the the adsorption CNBT composite of water centered molecules at 1650–1709 on cm the−1 was nanotube weaker than surface [9,33]. of the O-H bonds, which might be due to the adsorption of water molecules on the nanotube surface that of CNB, owing to the presence of CNTs.1 In addition, the CNBT peaks were compared with the The broadband[9,33]. spanningThe broadband from spanning 1000 to from 1400 1000 cm− tocorresponded 1400 cm−1 corresponded to the C =toO the bond C=O stretchingbond stretching vibrations [34]. characteristic peaks of the g-C3N4, BiFeO3, and CNTs individual components; the obtained results vibrations [34]. The FTIR peak of the CNBT composite centered at1 1650–1709 cm−1 was weaker than The FTIR peakrevealed ofthe thatCNBT the basic composite structure of centered each material at 1650–1709 was not destroyed cm− wasduring weaker the compounding than that process. of CNB, owing to the presencethat ofof CNTs.CNB, owing In addition, to the presence the CNBT of CNTs. peaks In addition, were comparedthe CNBT peaks with were the compared characteristic with the peaks of the g-C N , BiFeOcharacteristic, and CNTspeaks of individual the g-C3N4, components; BiFeO3, and CNTs the obtainedindividual resultscomponents; revealed the obtained that the results basic structure 3 4 revealed3 that the basic structure of each material was not destroyed during the compounding process. of each material was not destroyed during the compounding1600 process. 1200 550 CNBT

1600 1200 550 CNBT CNB

CNB

Transmittance (%) Transmittance CNTs 3435

Transmittance (%) Transmittance CNTs 4000 35003435 3000 2500 2000 1500 1000 500 -1 Wavenumber (cm )

4000Figure 3500 3. FTIR 3000 spectra 2500 of CNTs, 2000 CNB 1500 and 1000CNBT. 500 -1 Wavenumber (cm )

FigureFigure 3. FTIR 3. FTIR spectra spectra of of CNTs, CNTs, CNB CNB and andCNBT. CNBT.

3.1.3. TG-DSC

Figure S4 shows that both g-C3N4 and BiFeO3 exhibit a weak endothermic DSC peak at 60 ◦C because of the volatilization of the adsorbed water. However, no obvious weight loss was detected Int. J. Environ. Res. Public Health 2019, 16, x 7 of 18

3.1.3. TG-DSC

Figure S4 shows that both g-C3N4 and BiFeO3 exhibit a weak endothermic DSC peak at 60 °C Int. J. Environ. Res. Public Health 2019, 16, 3219 7 of 18 because of the volatilization of the adsorbed water. However, no obvious weight loss was detected on the corresponding TG curve, indicating that the analyzed sample contained a very little amount ofon water. the corresponding The mass loss TG of curve, g-C3N indicating4 was about that 90% the at analyzed 550–705 °C, sample and containedthe corresponding a very little DSC amount curve hadof water. an exothermic The mass peak loss of at g-C 7033 N°C4 wascaused about by 90%structural at 550–705 damage.◦C, andIn addition, the corresponding BiFeO3 produced DSC curve no otherhad an apparent exothermic endothermic peak at 703 or ◦exothermicC caused by peaks structural throughout damage. the Inheating addition, process BiFeO and3 produceddid not lose no anyother weight. apparent CNTs endothermic lost about or92% exothermic of their original peaks throughoutweight in the the temperature heating process region and of did560–660 not lose°C. Strongany weight. and weak CNTs exothermic lost about peaks 92% of were their observed original at weight 607 °C in and the 635 temperature °C, which region might ofbe 560–660 due to the◦C. oxidativeStrong and decomposition weak exothermic of the peaks surface were groups observed and oxidative at 607 ◦C combustion and 635 ◦C, of which the materials might be [9]. due The to CNBTthe oxidative composite decomposition exhibited a strong of the exothermic surface groups peak andat 527 oxidative °C, and combustionthe mass loss of at the 480–580 materials °C was [9]. aboutThe CNBT 52% (the composite further exhibited increase in a strongtemperature exothermic did not peak reduce at 527 its◦ massC, and significantly). the mass loss at 480–580 ◦C was about 52% (the further increase in temperature did not reduce its mass significantly). 3.1.4. VSM Measurements 3.1.4. VSM Measurements The magnetization hysteresis loops of BiFeO3, CNB, and CNBT were recorded by vibrating a sampleThe magnetometry magnetization to hysteresis determine loops their ofmagnetic BiFeO3 pr, CNB,operties. and Figure CNBT S5 were illustrates recorded the by data vibrating listed in a Tablesample S3, magnetometry shows the magnetic to determine properties their of magnetic BiFeO3, CNB, properties. and CNBT, Figure the S5 magnetic illustrates retention the data listedof all threein Table materials S3, shows was theweak magnetic (0.29, 0.14, properties and 0.12 of emu/g BiFeO for3, CNB,BiFeO and3, CNB, CNBT, and theCNBT, magnetic respectively), retention and of theall threemagnetization materials wasphenomenon weak (0.29, almost 0.14, anddisappeared 0.12 emu in/g forthe BiFeOabsence3, CNB, of an and applied CNBT, magnetic respectively), field, suggestingand the magnetization that the samples phenomenon exhibit a almost superparamagnetic disappeared in behavior the absence at room of an temperature, applied magnetic and their field, hysteresissuggesting loops that thewere samples sigmoidal exhibit closed a superparamagnetic curves [35]. In addition, behavior the atanalysis room temperature,shows that BiFeO and3 their is a strongerhysteresis superparamagnetic loops were sigmoidal than closedCNB and curves CNBT, [35 ].wh Inile addition, CNBT has the the analysis weakest shows magnetic that BiFeOproperties.3 is a Therefore,stronger superparamagnetic after the wastewater than treatment, CNB and the CNBT, separa whiletion and CNBT recycling has the of weakest the material magnetic can be properties. realized byTherefore, applying after a magnetic the wastewater field, thereb treatment,y simplifying the separation the treatment and recycling process. of the material can be realized by applying a magnetic field, thereby simplifying the treatment process. 3.1.5. ESR 3.1.5. ESR Since the ·OH, e−, and ·O2− species are important intermediates formed during the photocatalytic Since the OH, e , and O species are important intermediates formed during the photocatalytic process [36], the· synthesized− · 2material− was analyzed by the ESR, and the amounts of both the free radicalsprocess and [36], e the− were synthesized determined material from the was detected analyzed signal by st therengths. ESR, andFigure the 4a amounts displays of the both ESR the curves free radicals and e− were determined from the detected signal strengths. Figure4a displays the ESR of ·OH and ·O2− produced by the CNBT composite. It shows that the number of the ·OH and ·O2− curves of OH and O produced by the CNBT composite. It shows that the number of the OH species generated· in· the2− dark is very small, and their corresponding signals are also very weak. After· and O species generated in the dark is very small, and their corresponding signals are also very the light· 2− was applied for 5 min, the number of the two free radicals increased significantly. When the lightweak. was After applied the light for was10 min, applied the number for 5 min, of these the number free radicals of the twoalmost free doubled radicals as increased compared significantly. with that obtainedWhen the after light 5 was min applied of illumination. for 10 min, These the number result ofs revealed these free that radicals the almostCNBT doubledcomposites as comparedcould be with that obtained after 5 min of illumination. These results revealed that the CNBT composites excited to generate a large amount of the ·OH and ·O2− species under more permanent light could be excited to generate a large amount of the OH and− O2− species under more permanent conditions. Figure 4b,c show the ESR spectra of ·OH ·and e recorded· for g-C3N4, BiFeO3, CNB, and light conditions. Figure4b,c show the ESR spectra of OH and e recorded for g-C N , BiFeO , CNB, CNBT after 10 min of illumination. The higher was the· test signal− intensity, the greater3 4 was the3 ability ofand a material CNBT after to produce 10 min the of illumination. ·OH radicals. TheHowever, higher the was opposi the testte trend signal was intensity, observed the for greater the e− wasspecies the ability of a material to produce the OH radicals. However, the opposite trend was observed for the (the higher the intensity, the smaller· the number of e−). From the signal intensity curves depicted in Figuree− species 4b,c, (the it can higher be concluded the intensity, that the abilities smaller theof different number materials of e−). From to generate the signal ·OH intensity and e− may curves be depicted in Figure4b,c, it can be concluded that the abilities of di fferent materials to generate OH and ranked as follows: CNBT > CNB > BiFeO3 > g-C3N4, indicating that the CNBT composite ·has the strongeste− may be ·OH ranked and aselectron follows: generation CNBT > CNBcapabilities.> BiFeO 3 > g-C3N4, indicating that the CNBT composite has the strongest OH and electron generation capabilities. ·

(a) - ·O2 - CNBT Light 10 min

Light 5 min Dark

·OH - CNBT Light 10 min Intensity (a.u.)

Light 5 min Dark

320 322 324 326 328 Magnetid Field (mT) Figure 4. Cont. Int.Int. J.J. Environ.Environ. Res. Public Public Health Health 20120199, 1166, x 3219 8 8 of 1818

Int. J. Environ. Res. Public Health 2019, 16, x 8 of 18

(b) ·OH CNBT (b) ·OH CNBT

CNB

CNB Intensity (a.u.) Intensity BiFeO3 Intensity (a.u.) Intensity BiFeO3 g-C3N4

g-C3N4 317.9 318.0 318.1 318.2 318.3 Magnetid Field (mT) 317.9 318.0 318.1 318.2 318.3 Magnetid Field (mT) (c) - CNBT e (c) - CNBT e CNB CNB

BiFeO3

BiFeO3 Intensity (a.u.) Intensity g-C3N4 Intensity (a.u.) Intensity g-C3N4

318.4 318.5 318.6 318.7 318.8 318.4 318.5Magnetic 318.6 Field 318.7(mT) 318.8 Magnetic Field (mT) FigureFigure 4.4. (a) ESR spectraspectra ofof radicalradical adductsadducts trappedtrapped byby O·O22−− andand ·OHOH for CNBT in the dark and under Figure 4. (a) ESR spectra of radical adducts trapped by ··O2− and ·OH· for CNBT in the dark and under − visiblevisible lightlight irradiation.irradiation. ((bb,,cc)) ESRESR spectraspectra ofof g-Cg-C33NN44, BiFeOBiFeO3,, CNB, and CNBT byby ·OHOH ((bb)) andand− ee− (c) visible light irradiation. (b,c) ESR spectra of g-C3N4, BiFeO3, CNB, and CNBT by ·OH· (b) and e (c) exposed to visible light for 10 min. exposedexposed to tovisible visible light light for for 10 10 min. min. 3.1.6. XRD 3.1.6.3.1.6. XRD XRD The XRD patterns recorded for the CNTs, CNB, and CNBT are shown in Figure5. In our TheThe XRD XRD patterns patterns recorded recorded forfor thethe CNTs,CNTs, CNB, andand CNBTCNBT areare shown shown in in Figure Figure 5. 5.In Inour our previous study, the XRD patterns of the g-C3N4 and BiFeO3 individual components were analyzed [26]. previousprevious study, study, the the XRD XRD patterns patterns ofof thethe g-Cg-C33N4 and BiFeOBiFeO33 individualindividual components components were were analyzed analyzed The g-C3N4 peaks were centered at 2θ = 12.9◦ and 27.6◦, while the characteristic peaks of BiFeO3 [26].[26]. The The g-C g-C3N3N4 peaks4 peaks were were centered centered at at 22θθ == 12.9°12.9° and 27.6°,27.6°, while while the the characteristic characteristic peaks peaks of ofBiFeO BiFeO3 3 formed a diffraction pattern containing the (012), (104), (110), (006), (202), (024), (116), (122), (018), formedformed a diffractiona diffraction pattern pattern containing containing thethe (012),(012), (104), (110),(110), (006),(006), (202), (202), (024), (024), (116), (116), (122), (122), (018), (018), (202), (214), (300), (208), and (220) reflections of the planar BiFeO3 single-phase perovskite structure. (202),(202), (214), (214), (300), (300), (208), (208), and and (220) (220) reflections reflections ofof the planar BiFeOBiFeO3 3single-phase single-phase pero perovskitevskite structure. structure. The results presented in Figure5 show that the CNTs produced two di fferent diffraction peaks centered TheThe results results presented presented in in Figure Figure 55 showshow thatthat the CNTs producedproduced two two different different diffraction diffraction peaks peaks at 2θ = 25.9 and 42.8 , which corresponded to the (002) and (101) crystal planes, respectively (JCPDS centeredcentered at ◦at2θ 2 θ= 25.9°= 25.9° ◦and and 42.8°, 42.8°, which which corresponded corresponded toto thethe (002) (002) and and (101) (101) crystal crystal planes, planes, respectively respectively 01-0646; note that the intensity of the latter peak was relatively small) [37]. The XRD patterns of CNB (JCPDS(JCPDS 01-0646; 01-0646; note note that that the the intensity intensity ofof thethe latterlatter peak waswas relatively relatively small) small) [37]. [37]. The The XRD XRD patterns patterns andof CNBTCNB and contain CNBT the contain characteristic the characteristic peaks of g-C peaks3N4 ,of BiFeO g-C3N3,4 and, BiFeO CNTs.3, and However, CNTs. However, the diffraction the of CNB and CNBT contain the characteristic peaks of g-C3N4, BiFeO3, and CNTs. However, the peakdiffraction of the CNTs peak is of very the weak, CNTs which is very also weak, confirms which that also the confirms content ofthat CNTs the incontent the composite of CNTs is in smaller the diffraction peak of the CNTs is very weak, which also confirms that the content of CNTs in the thancomposite those of is the smaller g-C3N than4 and those BiFeO of 3thecomponents. g-C3N4 and BiFeO At the3 samecomponents. time, no At new the peakssame time, were no detected new composite is smaller than those of the g-C3N4 and BiFeO3 components. At the same time, no new in thepeaks di ffwereraction detected patterns in the of alldiffraction materials, patterns indicating of all that materials, no impurities indicating were that introduced no impurities during were the peaks were detected in the diffraction patterns of all materials, indicating that no impurities were materialintroduced preparation during the and material that the preparation purity of the and composite that the purity was relatively of the composite high. was relatively high. introduced during the material preparation and that the purity of the composite was relatively high.

CNBT CNBT

CNB

Intensity (a.u.) Intensity CNB Intensity (a.u.) Intensity

CNTs

CNTs 10 20 30 40 50 60 70 80 Two-Theta (deg) 10 20 30 40 50 60 70 80 Figure 5. Two-Theta (deg) Figure 5.XRD XRD patternspatterns of CNTs, CNB, and and CNBT. CNBT.

Figure 5. XRD patterns of CNTs, CNB, and CNBT.

Int. J. Environ. Res. Public Health 2019, 16, 3219 9 of 18 Int. J. Environ. Res. Public Health 2019, 16, x 9 of 18

3.1.7.3.1.7. XPS XPS InIn order order to to determine determine the the chemical chemical compositions compositions of of the the CNBT CNBT composite composite and and possible possible interaction interaction betweenbetween its its various various components, components, the the XPS XPS analysis analysis was was performed performed on on the the CNTs, CNTs, CNB, CNB, and and CNBT CNBT materialsmaterials (the (the full full XPS spectra of them are shown in FigureFigure6 6a).a). TheThe characteristiccharacteristic peakspeaks ofof thethe individualindividual components components had had different different degrees of corre correspondencespondence in the composite. Figure Figure 66bb contains thethe C C 1s XPS spectrum ofof thethe CNBT,CNBT, whichwhich was was fitted fitted with with three three di differentfferent Gaussian Gaussian peaks peaks centered centered at at284.80, 284.80, 285.55, 285.55, and and 288.71 288.71 eV. eV. The The first firs twot two peaks peaks corresponded corresponded to to either either the the sp sp22hybrid hybrid NN−CC=N=N or or − CC−CC structure structure and and sp sp33 hybridhybrid structure structure of C −NN3 inin g-C g-C3NN4, ,respectively respectively [38]. [38]. The The peak peak at at 288.71 288.71 eV eV − − 3 3 4 waswas attributed to the O −C=OC=O structure structure in in the the constituent constituent CNTs CNTs [39, [39,40].40]. In In Figure Figure 66c,c, the characteristic − peakspeaks ofof OO 1s 1s centered centered at 530.06,at 530.06, 531.56, 531.56, and and 533.41 533 eV.41 were eV attributedwere attributed to the presenceto the presence of Bi O (BiFeOof Bi−O), − 3 (BiFeOC=O, or3), OC=O,C= orO, O and−C=O, the hydroxyland the hydroxyl O H (CNTs) O−H species, (CNTs) respectively species, respective [26,41].ly In [26,41]. Figure 6Ind, Figure the N 6d,C − − − 3 thetertiary N−C3 nitrogen tertiary nitrogen structure structure and sp2 CandN sp=C2 C hybrid−N=C hybrid structure structure formed formed by the by N atomthe N asatom well as as well the − asC theN CH2−N structure−H structure formed formed by the by attachment the attachment of the amideof the N-containingamide N-containing group to group the edge to the of theedge layer of 2− − thewith layer binding withenergies binding ofenergies 399.31, of 399.06, 399.31, and 399.06, 401.17 and eV, 401.17 respectively eV, respectively [36,38]. As [36,38]. shown As in shown Figure6 ine, Figure 6e, the XPS spectrum of Bi can be distinguished by the presence of the 4 f7/2 and 4 f5/2 Gaussian the XPS spectrum of Bi can be distinguished by the presence of the 4 f 7/2 and 4 f 5/2 Gaussian peaks peakswith binding with binding energies energies of 159.47 of 159.47 eV and eV 164.77 and 164. eV,77 respectively, eV, respectively, which correspondwhich correspond to the spin-orbitto the spin- of orbitBi [31 of]. Bi From [31]. these From results, these itresults, can be it concluded can be conc thatluded the Bithat element the Bi existedelement in existed the form in the of Bi form3+ oxide of Bi in3+ oxidethe produced in the produced composite composite [42]. [42].

(a) (b) C 1s CNBT 284.80 Bi 4f CNB O 1s Fe 2p N 1s

C 1s 288.71 285.55 Intensity(a.u.) Intensity (a.u.) Intensity

CNTs

1200 1000 800 600 400 200 0 292 291 290 289 288 287 286 285 284 283 282 Binding Energy (eV) Binding Energy (eV) (d) (c) O 1s N 1s

530.06 399.31

531.56 401.17 399.06 Intensity (a.u.) Intensity (a.u.) Intensity

533.41

536 535 534 533 532 531 530 529 528 527 404 403 402 401 400 399 398 397 396 Binding Energy (eV) Binding Energy (eV)

(e) Bi 4f 159.47

164.77 Intensity(a.u.)

168 166 164 162 160 158 156 Binding Energy (eV) FigureFigure 6. 6. ((aa)) Full Full XPS spectraspectra ofof CNTs, CNTs, CNB, CNB, and and CNBT. CNBT. Characteristic Characteristic spectra spectra of CNBT, of CNBT, (b) C(b 1s,) C ( c1s,) O ( 1s,c) O(d 1s,) N ( 1s,d) N (e )1s, Bi ( 4f.e) Bi 4f.

3.2. Investigation on Charge Separation and Optical Properties

Int. J. Environ. Res. Public Health 2019, 16, 3219 10 of 18

3.2. Investigation on Charge Separation and Optical Properties The results of the photocurrent test are depicted in Figure7a. They show that g-C N , BiFeO , Int. J. Environ. Res. Public Health 2019, 16, x 310 4of 18 3 CNB, and CNBT can produce electrons at the moment of illumination and generate an electric current due to theThe photocatalytic results of the ephotocurrentffect [26]. The test photocurrent are depicted in in Figure CNBT 7a. was They much show stronger that g-C than3N4, BiFeO those3, in the otherCNB, materials, and CNBT confirming can produce that electrons the former at the possessed moment aof higher illumination separation and generate efficiency an electric of photogenerated current hole-electrondue to the pairsphotocatalytic and larger effect electron [26]. The conductivity. photocurrent in CNBT was much stronger than those in the otherThe PLmaterials, properties confirming of the synthesizedthat the former materials possessed were a examinedhigher separation to further efficiency determine of the recombinationphotogenerated rates hole-electron of photoexcited pairs and electron-hole larger electron pairs conductivity. [43]. Generally, the higher the fluorescence The PL properties of the synthesized materials were examined to further determine the signal, the greater the recombination rate of such pairs [28]. In Figure7b, the pure g-C N exhibits the recombination rates of photoexcited electron-hole pairs [43]. Generally, the higher the fluorescence3 4 strongest fluorescent peak, owing to its highest recombination rates of photogenerated electron-hole signal, the greater the recombination rate of such pairs [28]. In Figure 7b, the pure g-C3N4 exhibits the pairs.strongest After combining fluorescent g-C peak,3N 4owingwith to BiFeO its highest3 and CNTs,recombination the PL peakrates intensityof photogenerated was significantly electron-hole reduced, whichpairs. indicated After combining that the additiong-C3N4 with of BiFeOBiFeO3 and CNTs, CNTs the could PL inhibitpeak intensity the photoelectron-hole was significantly pair recombination,reduced, which thereby indicated improving that the theaddition photocatalytic of BiFeO3 and effi ciencyCNTs could of the inhibit composite the photoelectron-hole [38]. pairThe recombination, optical properties thereby of aimproving material directlythe photocatalytic affect its efficiency light absorption of the composite capacity [38]. [44 ]. In this work, The optical properties of a material directly affect its light absorption capacity [44]. In this work, UV-visible diffuse reflection studies of g-C3N4, BiFeO3, CNB, and CNBT were performed to examine theirUV-visible light absorption diffuse reflection characteristics. studies of Figure g-C3N74,c BiFeO clearly3, CNB, shows and that CNBT the were edge performed of the light to examine absorption their light absorption characteristics. Figure 7c clearly shows that the edge of the light absorption region of g-C3N4 is ca. 480 nm, which is close to the absorption regions of g-C3N4 prepared in previous region of g-C3N4 is ca. 480 nm, which is close to the absorption regions of g-C3N4 prepared in previous studies [45–47]. BiFeO3 had a relatively high visible light absorption capacity, and the edge of its studies [45–47]. BiFeO3 had a relatively high visible light absorption capacity, and the edge of its absorptionabsorption region region was was ca. ca. 775 775 nm nm [ 48[48].]. FigureFigure 77cc shows that that CNB CNB has has a wider a wider photo-response photo-response range range thanthan that that of the of the pure pure g-C g-C3N34Ndue4 due to to the the additionaddition of of BiFeO BiFeO3 3andand that that the the photo-response photo-response range range of the of the CNBTCNBT composite composite is further is further broadened broadened after after adding adding the theCNTs. CNTs. The band gap gap energy energy was was calculated calculated by the followingby the following formula: formula: E = 1240/λ (2) Egg = 1240/λ (2) λ wherewhereis λ the is the wavelength wavelength (nm) (nm) of ofthe the absorptionabsorption edge, edge, and and Eg E isg isthe the band band gap gapenergy energy of the of the semiconductorsemiconductor (eV) (eV) [43 [43,49].,49]. This This formula formula shows shows that the the band band gaps gaps of of g-C g-C3N34 Nand4 and BiFeO BiFeO3 are 32.58are eV 2.58 eV and 1.6and eV, 1.6 respectively.eV, respectively. According According to to Figure Figure7 7c,c, the the bandband gaps gaps of of the the studied studied materials materials can canbe ranked be ranked in thein orderthe order of g-C of g-C3N43N>4 BiFeO> BiFeO3 3> > CNBCNB > CNBT.CNBT. The The lowest lowest band band gap gap of CNBT of CNBT can canfacilitate facilitate the the excitationexcitation and and transition transition of of the the electrons, electrons, therebythereby improving improving the the photocatalytic photocatalytic performance performance of the of the materialmaterial [38]. [38].

Figure 7. (a) Photocurrent response density. (b) Photoluminescence spectra. (c) UV-Vis spectra. Figure 7. (a) Photocurrent response density. (b) Photoluminescence spectra. (c) UV-Vis spectra.

Int. J. Environ. Res. Public Health 2019, 16, 3219 11 of 18

3.3. Adsorption Studies Before conducting the photocatalytic performance tests, a series of adsorption experiments were performed to compare the adsorption properties of different materials. By taking into account the characteristics of wastewater, the influences of pH, and the concentrations of Cr(VI) and MB on their adsorption properties were analyzed.

3.3.1. Adsorption Properties of Different Materials

Figure S6 shows the adsorption efficiencies of g-C3N4, BiFeO3, CNB, and CNBT in the mixed solution of MB and Cr(VI). After 5 h of reaction, the adsorption capacity of g-C3N4 towards MB was about 0.63 mg/g; however, no apparent adsorption of the Cr(VI) ions was observed. The adsorption capacities of BiFeO3 towards MB and Cr(VI) reached 4.85 mg/g and 0.78 mg/g, respectively. While the adsorption efficiency of CNB was lower than that of BiFeO3 due to the combination of g-C3N4 and BiFeO3 that likely blocked some adsorption sites on the BiFeO3 surface. Compared with CNB, CNBT has a higher specific surface area and more adsorption sites, which promoted the adsorptions of MB and Cr(VI) reaching the adsorption capacities of 5.33 mg/g and 1.04 mg/g, respectively. According to the results of the BET measurements, CNBT had the highest pore specific surface area of 38.3291 m2/g and a pore volume of 0.1576 cm3/g, which also contributed to its superior adsorption properties.

3.3.2. Adsorption Efficiencies at Different pH The pH value of the solution is an important parameter of the adsorption process, which affects not only the surface charge of the adsorbent, but also the physicochemical properties of the adsorbate [7,34]. In this work, the pH values were adjusted to 2.0–11.0 using 0.1 and 1 mol/L of the NaOH and HCl solutions, and the adsorption reaction was performed for 5 h at 25 ◦C. According to the results presented in Figure8, the adsorption e fficiency strongly depends on the pH value. When the pH was above three, the adsorption of Cr(VI) was suppressed with an increase in pH, whereas the opposite trend was observed for MB because the higher pH values increased the number of negative charges on the CNBT surface due to protonation [50]. Cr(VI) exists mainly in the form of HCrO4− under acidic conditions and in the form of Cr2O7− in the alkaline solutions. Such Cr(VI) species with negative charges are electrostatically repelled from the negatively charged CNBT species. In contrast, the cationic dye MB Int.is J. positively Environ. Res. charged Public Health because 2019, 1 it6, x can ionize chloride ions and color groups in the solution and12 thus of 18 is attracted by the negative charges of CNBT, which increases the MB removal rate. 4.8 2.0

4.6 1.6 ) g 4.4 / ) g

g 1.2 / m g m ) ( 4.2 M B I (

0.8 V B ( C r r M C

4.0 (VI) f f o o

0.4 e q 3.8 q e 0.0 3.6

2 4 6 8 10 12 pH

Figure 8. Effect of pH values on the adsorption of Cr(VI) and MB with CNBT: C0Cr(VI) = 5 mg/L, FigureC0MB =8. 20Effect mg/ L,of mpH/V values= 2.5 gon/L, the T = adsorption25 ◦C. of Cr(VI) and MB with CNBT: C0Cr(VI) = 5 mg/L, C0MB = 20 mg/L, m/V = 2.5 g/L, T = 25 °C.

3.4. Photocatalytic Experiments

3.4.1. Photocatalytic Properties of Different Materials Figure 9 describes the removal of the Cr(VI) and MB species from the mixed wastewater by the four materials under the same reaction conditions. After 1 h of the treatment, the removal rates of Cr(VI) and MB by the CNBT composite were 50% and 75% within the first half an hour and 55% and 79% by the end of the first hour, respectively, indicating that the adsorption-desorption equilibrium was almost reached within the study period. After the application of light, the removal rates of both pollutants significantly increased. During the 4 h photocatalytic treatment, g-C3N4 and BiFeO3 exhibited different adsorption capacities toward Cr(VI) (see Figure 9a), but no significant differences in their photocatalytic abilities were observed. After the photocatalytic treatment, the removal rate of Cr(VI) exceeded 98%; however, g-C3N4 demonstrated a poor photocatalytic degradation of MB, as shown in Figure 9b. After the synergistic treatment involving both the adsorption and photocatalysis reactions, about 50% of MB was removed from the mixed solution, which might be caused by the relatively low number of the ·OH radicals produced by the g-C3N4-containing photocatalytic systems, resulting in the undesirable photodegradation of MB. In addition, CNBT exhibited a stronger photocatalytic effect than that of CNB because the added CNTs played the role of channels that facilitated the electron transfer process. As a result, the number of active sites increased due to the large surface area, which increased the reaction efficiency.

(a) (b)

0 1.0 1.0 MB - g-C N Light on 0 3 4 Cr(VI) - g-C3N4 Photocatalysis MB - BiFeO Cr(VI) - BiFeO 3 0.8 3 Cr(VI) - CNB 0.8 MB - CNB Cr(VI) - CNBT MB - CNBT 0.6 0.6

0.4 0.4 Ligth on 0.2 0.2 0.0

Photocatalysis C/C MB of ratio Remaining

Remaining ratio of Cr(VI) C/C Cr(VI) of ratio Remaining 0.0 -0.2 -60 0 60 120 180 240 -60 0 60 120 180 240 Time (min) Time (min)

Figure 9. Effects of different kinds of materials on Cr(VI) (a) and MB (b) reduction by g-C3N4, BiFeO3,

CNB, and CNBT under visible light irradiation when Cr(VI) mixed with MB: C0Cr(VI) = 5 mg/L, C0MB = 20 mg/L, m/V = 2.5 g/L, pH = 2.0. (−60–0 min: Adsorption process, 0–240 min: Photocatalysis process).

Int. J. Environ. Res. Public Health 2019, 16, x 12 of 18

4.8 2.0

4.6 1.6 ) g 4.4 / ) g

g 1.2 / m g m ) ( 4.2 M B I (

0.8 V B ( C r r M C

4.0 (VI) f f o o

0.4 e q 3.8 q e Int. J. Environ. Res. Public Health 2019, 16, 3219 12 of 18 0.0 3.6

When the pH was between two2 and three,4 the6 actual adsorption8 10 effect of12 the composite on the two pollutants was opposite to the overall trend. This phenomenonpH can be explained by the fact that under strong acidic conditions, MB was protonated to become positively charged, whereas Cr(VI) possessed

a negative charge in the form of HCrO4−. These two compounds could react to form a flocculent chelate complex, which reduced the amount of Cr(VI) adsorbed by the material. Therefore, in the Figure 8. Effect of pH values on the adsorption of Cr(VI) and MB with CNBT: C0Cr(VI) = 5 mg/L, C0MB = case of a large MB concentration, the reduction in the number of the Cr(VI) adsorption sites provided more20 mg/L, adsorption m/V = 2.5 sites g/L, for T = MB. 25 °C. As a result, the material exhibited stronger MB adsorption properties. When the pH was between six and eight, the form of Cr(VI) changed dramatically, mainly reflected in 3.4. Photocatalytic Experiments 2 the transformation of HCrO4− to CrO4 −, which made the unstable trend for Cr(VI) a possibility [26].

3.4.1.3.4. Photocatalytic Photocatalytic Properties Experiments of Different Materials

3.4.1.Figure Photocatalytic 9 describes Propertiesthe removal of Diofff theerent Cr(VI) Materials and MB species from the mixed wastewater by the four materials under the same reaction conditions. After 1 h of the treatment, the removal rates of Figure9 describes the removal of the Cr(VI) and MB species from the mixed wastewater by the Cr(VI)four and materials MB by under the CNBT the same composite reaction conditions. were 50% Afterand 75% 1 h of within the treatment, the first the half removal an hour rates and of Cr(VI) 55% and 79% andby the MB end by the of CNBTthe first composite hour, respectively, were 50% and indicating 75% within that the the first adsorption-desorption half an hour and 55% and equilibrium 79% by was almostthe end ofreached the first within hour, respectively, the study period. indicating After that the adsorption-desorptionapplication of light, equilibriumthe removal was rates almost of both pollutantsreached significantly within the study increased. period. AfterDuring the applicationthe 4 h photocatalytic of light, the removal treatment, rates ofg-C both3N pollutants4 and BiFeO3 exhibitedsignificantly different increased. adsorption During capaci the 4ties h photocatalytic toward Cr(VI) treatment, (see Figure g-C3N 9a4 and), but BiFeO no 3significantexhibited di differencesfferent in theiradsorption photocatalytic capacities abilities toward Cr(VI)were observed. (see Figure 9Aftea), butr the no photocatalytic significant di fferences treatment, in their the photocatalytic removal rate of abilities were observed. After the photocatalytic treatment, the removal rate of Cr(VI) exceeded 98%; Cr(VI) exceeded 98%; however, g-C3N4 demonstrated a poor photocatalytic degradation of MB, as shownhowever, in Figure g-C 9b.3N 4Afterdemonstrated the synergistic a poor treatment photocatalytic involving degradation both the of adsorption MB, as shown and in photocatalysis Figure9b. After the synergistic treatment involving both the adsorption and photocatalysis reactions, about 50% of reactions, about 50% of MB was removed from the mixed solution, which might be caused by the MB was removed from the mixed solution, which might be caused by the relatively low number of the relatively low number of the ·OH radicals produced by the g-C3N4-containing photocatalytic systems, OH radicals produced by the g-C N -containing photocatalytic systems, resulting in the undesirable · 3 4 resultingphotodegradation in the undesirable of MB. In photodegradation addition, CNBT exhibited of MB. a strongerIn addition, photocatalytic CNBT exhibited effect than a that stronger of photocatalyticCNB because effect the addedthan CNTsthat of played CNB the because role of channelsthe added that CNTs facilitated played the electron the role transfer of channels process. that facilitatedAs a result, the electron the number transfer of active process. sites increasedAs a resu duelt, the to thenumber large surfaceof active area, sites which increased increased due the to the largereaction surface e area,fficiency. which increased the reaction efficiency.

(a) (b)

0 1.0 1.0 MB - g-C N Light on 0 3 4 Cr(VI) - g-C3N4 Photocatalysis MB - BiFeO Cr(VI) - BiFeO 3 0.8 3 Cr(VI) - CNB 0.8 MB - CNB Cr(VI) - CNBT MB - CNBT 0.6 0.6

0.4 0.4 Ligth on 0.2 0.2 0.0

Photocatalysis C/C MB of ratio Remaining

Remaining ratio of Cr(VI) C/C Cr(VI) of ratio Remaining 0.0 -0.2 -60 0 60 120 180 240 -60 0 60 120 180 240 Time (min) Time (min)

Figure 9. Effects of different kinds of materials on Cr(VI) (a) and MB (b) reduction by g-C3N4, BiFeO3, CNB, Figure 9. Effects of different kinds of materials on Cr(VI) (a) and MB (b) reduction by g-C3N4, BiFeO3, and CNBT under visible light irradiation when Cr(VI) mixed with MB: C0Cr(VI) = 5 mg/L, C0MB = 20 mg/L, CNB, and CNBT under visible light irradiation when Cr(VI) mixed with MB: C0Cr(VI) = 5 mg/L, C0MB = m/V = 2.5 g/L, pH = 2.0. ( 60–0 min: Adsorption process, 0–240 min: Photocatalysis process). 20 mg/L, m/V = 2.5 g/L, pH =− 2.0. (−60–0 min: Adsorption process, 0–240 min: Photocatalysis process). 3.4.2. The Relationship between the Two Pollutants In order to examine possible reactions between Cr(VI) and MB in the studied photocatalytic systems, the removal efficiency of a single pollutant by CNBT was compared with that of the mixed pollutants (the obtained results are shown in Figure 10). It was found that the presence of MB significantly promoted the removal of Cr(VI). When CNBT was used for the photocatalytic treatment of Cr(VI) alone, Int. J. Environ. Res. Public Health 2019, 16, x 13 of 18

3.4.2. The Relationship between the Two Pollutants In order to examine possible reactions between Cr(VI) and MB in the studied photocatalytic systems, the removal efficiency of a single pollutant by CNBT was compared with that of the mixed pollutants (the obtained results are shown in Figure 10). It was found that the presence of MB significantly promoted the removal of Cr(VI). When CNBT was used for the photocatalytic treatment of Cr(VI)Int. J. Environ.alone, Res.the Public removal Health 2019rate, 16 was, 3219 only about 50%; however, when both pollutants were 13 ofpresent 18 in the solution, the Cr(VI) removal rate was 93%. Which means that Cr(VI) in the mixed substrates is higherthe than removal that rateof the was single only aboutsystem, 50%; the however, phenomen whenon bothis mainly pollutants due wereto the present synergistic in the relationship solution, that takesthe Cr(VI) place removal between rate the was oxidation 93%. Which of MB means and thatreduction Cr(VI) inof theCr(VI) mixed onto substrates the surface is higher of the than materials, that the synergisticof the single effects system, are the benefical phenomenon to minimize is mainly the due recombination to the synergistic rate relationship of the charge that carriers takes place [51]. As for thebetween influence the oxidation of Cr(VI) of MBon the and reductionMB removal, of Cr(VI) the ontoadsorption the surface of ofCr(VI) the materials, was suppressed the synergistic but the photocatalyticeffects are beneficaldegradation to minimize was hardly the recombination affected beca rateuse of theCr(VI) charge competed carriers [ 51with]. As MB for for the the influence adsorption of sites. Cr(VI) on the MB removal, the adsorption of Cr(VI) was suppressed but the photocatalytic degradation was hardly affected because Cr(VI) competed with MB for the adsorption sites.

1.0 Ligth on Cr(VI) - Alone Cr(VI) - Mixture MB 0.8 MB - Alone MB - Mixture Cr(VI) 0.6

0.4

0.2

0.0 Photocatalysis Remaining ratioRemaining of Cr(VI) and MB -0.2 -60 0 60 120 180 240 Time (min) Figure 10. Photocatalysis of Cr(VI) and MB alone or mixed by CNBT under visible light irradiation: Figure 10. Photocatalysis of Cr(VI) and MB alone or mixed by CNBT under visible light irradiation: C = 5 mg/L, C = 20 mg/L, m/V = 2.5 g/L, pH = 2.0. ( 60–0 min: Adsorption process, 0–240 min: 0Cr(VI) 0MB − C0Cr(VI)Photocatalysis = 5 mg/L, C0MB process). = 20 mg/L, m/V = 2.5 g/L, pH = 2.0. (−60–0 min: Adsorption process, 0–240 min: Photocatalysis process). 3.4.3. Possible Photocatalytic Mechanism in Reaction System 3.4.3. PossibleThe proposedPhotocatalytic photocatalytic Mechanism mechanism in Reaction of CNBT System is shown in Figure 11. When the CNBT photocatalyst is irradiated with visible light that possesses a higher photon energy than its band gap, The proposed photocatalytic mechanism of CNBT is shown in Figure 11. When the CNBT the catalyst becomes excited to generate a large number of electron-hole pairs [52]. The energy levels photocatalyst is irradiated with visible light that possesses a higher photon energy than its band gap, of g-C3N4 was calculated by the Mulliken electronegativity theory [53]: the catalyst becomes excited to generate a large number of electron-hole pairs [52]. The energy levels of g-C3N4 was calculated by the MullikenE electronegativity= X Ee 0.5 E theory[53]:g (3) CB − − ECB = X − Ee − 0.5 Eg (3) EVB = ECB + Eg (4)

where X is the absolute electronegativity ofEVB a semiconductor = ECB + Eg (X-g-C3N4 = 4.72 eV); Ee is the energy(4) of free electrons compared to hydrogen (4.5 eV) [43]; Eg is the band-gap of the g-C3N4 and BiFeO3 where X is the absolute electronegativity of a semiconductor (X-g-C3N4 = 4.72 eV); Ee is the energy of are about 2.58 eV and 1.6 eV, respectively. The semiconductor coupling effect between BiFeO3 and free electronsg-C3N4 triggers compared both theto hydrogen electron transfer (4.5 eV) from [43]; the Eg conduction is the band-gap band of g-C of3 theN4 tog-C that3N of4 and BiFeO BiFeO3 and3 are abouthole 2.58 migration eV and from1.6 eV, the respectively. valence band ofThe BiFeO semiconductor3 to that of g-C coupling3N4 [38], effect while CNTsbetween serve BiFeO as highly3 and g- C3N4 conductivetriggers both transfer the electron paths that transfer facilitate from the the migration conduction of these band photoexcited of g-C3N4 charges. to that of In BiFeO addition,3 and the hole migrationsimultaneous from the treatment valence of theband organic of BiFeO pollutant3 to MBthat and of heavyg-C3N metal4 [38], Cr(VI) while can CNTs further serve promote as thehighly conductiveseparation transfer of photoexcited paths that electron-hole facilitate the pairs. migratio Specifically,n of these photoexcited photoexcited electrons charges. can react In withaddition, O2 in the the photocatalytic system to form the O species that are able to reduce Cr(VI) [29]. Furthermore, simultaneous treatment of the organic pollutant· 2− MB and heavy metal Cr(VI) can further promote the these electrons can also migrate to the material surface and react directly with Cr(VI) [54]. Meanwhile, separation of photoexcited electron-hole pairs. Specifically, photoexcited electrons can react with O2 the photoexcited holes interact with the water adsorbed on the CNBT surface to form the OH radicals, in the photocatalytic system to form the ·O2− species that are able to reduce Cr(VI) [29].· Furthermore, which can subsequently oxidize MB into CO2 and H2O. these electrons can also migrate to the material surface and react directly with Cr(VI) [54]. Meanwhile, + CNBT + hv CNBT (e− + h ) (5) → CB VB

e− + O O − (6) 2 → · 2 Int. J. Environ. Res. Public Health 2019, 16, x 14 of 18 the photoexcited holes interact with the water adsorbed on the CNBT surface to form the ·OH radicals, which can subsequently oxidize MB into CO2 and H2O.

CNBT + hv → CNBT (e−CB + h+VB) (5)

e− + O2 → ·O2− (6) Int. J. Environ. Res. Public Health 2019, 16, 3219 14 of 18 h+ + H2O → ·OH + H+ (7)

+ + −h + H+ O − OH +3+H (7) HCrO4 + 7H 2 + 3e→ ·→Cr + 4H2O (8) + 3+ HCrO4− + 7H + 3e− Cr + 4H2O (8) ·OH + MB → CO→2 + H2O (9) OH + MB CO + H O (9) · → 2 2

Figure 11. Mechanism diagram of the CNBT composite photocatalytic Cr and MB. Figure 11. Mechanism diagram of the CNBT composite photocatalytic Cr and MB. 4. Conclusions

4. ConclusionsIn this study, the CNBT ternary magnetic composite photocatalytic material with a high photocatalytic activity was prepared by the hydrothermal synthesis method and then used for Inthe this adsorption study, and the photocatalytic CNBT ternary treatment magnetic of the wastewatercomposite containingphotocatalytic the heavy material metal Cr(VI)with anda high photocatalyticorganic dye activity MB. During was theprepared adsorption by process,the hydrothermal the Cr(VI) and synthesis MB species method competed and for then the adsorption used for the adsorptionsites, but and the photocatalytic simultaneous photocatalysis treatment of of the these wa twostewater pollutants containing actually increased the heavy their metal removal Cr(VI) rates. and organicIn thedye lower MB. pH During range, athe stronger adsorption Cr(VI) adsorptionprocess, the eff ectCr(VI) was observed.and MB Thespecies obtained competed equilibrium for the adsorptionisotherms sites, were but fittedthe simultaneous with different photocatalys two-parameteris of models, these andtwo thepollutants Langmuir actually model increased produced atheir removalbetter rates. fit than In the the lower Freundlich pH range, model, a suggestingstronger Cr(VI) that the adsorption distribution effect of the was active observed. sites on The the CNBTobtained equilibriumsurface wasisotherms homogeneous. were fitt Theed resultswith different of the kinetic two-parameter studies showed models, that the and PS-order the Langmuir kinetic model model better described the adsorption process and that the adsorption process reached saturation after 0.5 h. produced a better fit than the Freundlich model, suggesting that the distribution of the active sites on The photocatalytic performances of different materials can be ranked in the order of CNBT > BiFeO > the CNBT surface was homogeneous. The results of the kinetic studies showed that the PS-order3 CNB > g-C3N4. The spectra obtained by UV-Vis, ESR, VSM, PL, and other characterization techniques kineticrevealed model that better the CNBTdescribed composite the adsorption had a reasonable process band and gap, that could the be adsorption easily separated process from reached the saturationsolution, after and 0.5 possessed h. The photocatalytic good visible light performa response,nces high of chemical different stability, materials and can low recombinationbe ranked in the orderrates of CNBT of photoexcited > BiFeO3 > charges,CNB > g-C which3N4. contributed The spectra to obtained its high photocatalyticby UV-Vis, ESR, efficiency. VSM, PL, Moreover, and other characterizationphotocatalysis techniques likely broke revealed the adsorption-desorption that the CNBT composite equilibrium had achieveda reasonable over band a short gap, time could and be easilypromoted separated the from deep the degradation solution, ofand pollutants possessed by thegood studied visible materials. light response, Therefore, hi thegh resultschemical obtained stability, and lowin this recombination work can provide rates a practicalof photoexcited and theoretical charges, basis which for the contributed treatment of to wastewaters its high photocatalytic containing efficiency.organic Moreover, pollutants andphotocatalysis heavy metals. likely broke the adsorption-desorption equilibrium achieved over Supplementarya short time Materials:and promotedThe following the aredeep available degradati onlineon at http: of //pollutantswww.mdpi.com by/ 1660-4601the studied/16/17 /3219materials./s1; Therefore,S1. Adsorption the results models; obtained S2. Effect in of this contact work time andcan adsorptionprovide kinetics;a practical S3. Adsorption and theoretical isotherms. basis Table for S1: the treatmentKinetic of parameters wastewaters of the containing Cr(VI) and MB organic adsorption pollutants by CNBT: andC0Cr(VI) heavy= 5 mg metals./L, C0MB = 20 mg/L, m/V = 2.5 g/L, pH = 3.0, T = 25 ◦C; Table S2: Comparison of the different isotherm parameters of the Cr(VI) and MB adsorption by CNBT: C0Cr(VI) = 5 mg/L, C0MB = 20 mg/L, m/V = 2.5 g/L, pH = 3.0, T = 25 ◦C; Table S3: Magnetic properties of SupplementaryBiFeO3, CNB, Materials: and CNBT The measured. following Figure are available S1: Sorption online kinetics at www. data andmdpi.com/link; fitted models S1. of Cr(VI)Adsorption and MB models; by CNBT; Figure S2: Different adsorption isotherm models of Cr(VI) (a) and MB (b) by CNBT; Figure S3: Schematic S2. Effect of contact time and adsorption kinetics; S3. Adsorption isotherms. Table S1: Kinetic parameters of the diagram of the photocatalytic activity test reaction scheme. Figure S4: TG-DSC curves of g-C3N4, BiFeO3, CNTs,

Int. J. Environ. Res. Public Health 2019, 16, 3219 15 of 18

and CNBT; Figure S5: Hysteresis loops of BiFeO3, CNB, and CNBT measured; Figure S6: Comparison of the adsorption effects of g-C3N4, BiFeO3, CNB, and CNBT: C0Cr(VI) = 5 mg/L, C0MB = 20 mg/L, m/V = 2.5 g/L, pH = 2.0, T = 25 ◦C. Author Contributions: The authors H.H., C.L., and G.Q. did the experimental work and wrote the manuscript; X.H., H.W., and Y.L. conceived this study and revised the manuscript; J.L., X.T., G.X., Q.J., and D.Z. contributed to the data analysis and prepared the figures. All authors reviewed the manuscript and contributed to the scientific discussion. Acknowledgments: This research was funded by the National Natural Science Foundation of China (grant number 51608208), the Natural Science Foundation of Hunan Province (grant numbers 2018JJ3887, 2019JJ51005, and 2018JJ3096), the Research Foundation of Education Department of Hunan Province, China (grant number 17K105). Conflicts of Interest: The authors declare that they have no competing interest.

References

1. Han, D.; Currell, M.J.; Cao, G. Deep challenges for China’s war on water pollution. Environ. Pollut. 2016, 218, 1222–1233. [CrossRef][PubMed] 2. Feng, Y.; Yang, S.; Xia, L.; Wang, Z.; Suo, N.; Chen, H.; Long, Y.; Zhou, B.; Yu, Y. In-situ ion exchange electrocatalysis biological coupling (i-IEEBC) for simultaneously enhanced degradation of organic pollutants and heavy metals in electroplating wastewater. J. Hazard. Mater. 2019, 364, 562–570. [CrossRef][PubMed] 3. Xiao, K.; He, J.; Zhang, J.; Yang, B.; Zhao, X. Simultaneous Cr(VI) removal and bisphenol A degradation in a solar-driven photocatalytic fuel cell with dopamine modified carbon felt cathode. Appl. Surf. Sci. 2019, 471, 912–920. [CrossRef] 4. Katheresan, V.; Kansedo, J.; Lau, S.Y. Efficiency of various recent wastewater dye removal methods: A review. J. Environ. Chem. Eng. 2018, 6, 4676–4697. [CrossRef] 5. Mitoraj, D.; Lamdab, U.; Kangwansupamonkon, W.; Pacia, M.; Macyk, W.; Wetchakun, N.; Beranek, R. Revisiting the problem of using methylene blue as a model pollutant in photocatalysis: The case of

InVO4/BiVO4 composites. J. Photochem. Photobiol. A Chem. 2018, 366, 103–110. [CrossRef] 6. Zhang, Y.; Xu, M.; Li, H.; Ge, H.; Bian, Z. The enhanced photoreduction of Cr(VI) to Cr(III) using carbon dots

coupled TiO2 mesocrystals. Appl. Catal. B Environ. 2018, 226, 213–219. [CrossRef] 7. Ning, J.; Wang, M.; Luo, X.; Hu, Q.; Hou, R.; Chen, W.; Chen, D.; Wang, J.; Liu, J. SiO2 Stabilized Magnetic Nanoparticles as a Highly Effective Catalyst for the Degradation of Basic Fuchsin in Industrial Dye Wastewaters. Molecules 2018, 23, 2573. [CrossRef][PubMed] 8. Gan, L.; Zhou, F.; Owens, G.; Chen, Z. Burkholderia cepacia immobilized on eucalyptus leaves used to simultaneously remove malachite green (MG) and Cr(VI). Colloids Surf. B Biointerfaces 2018, 172, 526–531. [CrossRef][PubMed] 9. Deng, Y.; Ok, Y.S.; Mohan, D.; Pittman, C.U.; Dou, X. Carbamazepine removal from water by carbon dot-modified magnetic carbon nanotubes. Environ. Res. 2019, 169, 434–444. [CrossRef][PubMed] 10. Sheng, Y.; Wei, Z.; Miao, H.; Yao, W.; Li, H.; Zhu, Y. Enhanced organic pollutant photodegradation via

adsorption/photocatalysis synergy using a 3D g-C3N4/TiO2 free-separation photocatalyst. Chem. Eng. J. 2019, 370, 287–294. [CrossRef] 11. Chen, F.; An, W.; Liu, L.; Liang, Y.; Cui, W. Highly efficient removal of bisphenol A by a three-dimensional graphene hydrogel-AgBr@rGO exhibiting adsorption/photocatalysis synergy. Appl. Catal. B Environ. 2017, 217, 65–80. [CrossRef]

12. Mu, C.; Zhang, Y.; Cui, W.; Liang, Y.; Zhu, Y. Removal of bisphenol A over a separation free 3D Ag3PO4 -graphene hydrogel via an adsorption-photocatalysis synergy. Appl. Catal. B Environ. 2017, 212, 41–49. [CrossRef] 13. Cao, S.; Low, J.; Yu, J.; Jaroniec, M. Polymeric Photocatalysts Based on Graphitic Carbon Nitride. Adv. Mater. 2015, 27, 2150–2176. [CrossRef][PubMed]

14. Kong, L.; Mu, X.; Fan, X.; Li, R.; Zhang, Y.; Song, P.; Ma, F.; Sun, M. Site-selected N vacancy of g-C3N4 for photocatalysis and physical mechanism. Appl. Mater. Today 2018, 13, 329–338. [CrossRef]

15. Cai, X.; He, J.; Chen, L.; Chen, K.; Li, Y.; Zhang, K.; Jin, Z.; Liu, J.; Wang, C.; Wang, X.; et al. A 2D-g-C3N4 nanosheet as an eco-friendly adsorbent for various environmental pollutants in water. Chemosphere 2017, 171, 192–201. [CrossRef] Int. J. Environ. Res. Public Health 2019, 16, 3219 16 of 18

16. Tan, J.Z.Y.; Nursam, N.M.; Xia, F.; Sani, M.-A.; Li, W.; Wang, X.; Caruso, R.A. High-Performance Coral Reef-like Carbon Nitrides: Synthesis and Application in Photocatalysis and Heavy Metal Ion Adsorption. ACS Appl. Mater. Interfaces 2017, 9, 4540–4547. [CrossRef]

17. Wu, X.; Wang, X.; Wang, F.; Yu, H. Soluble g-C3N4 nanosheets: Facile synthesis and application in photocatalytic hydrogen evolution. Appl. Catal. B Environ. 2019, 247, 70–77. [CrossRef]

18. Carvalho, K.T.G.; Nogueira, A.E.; Lopes, O.F.; Byzynski, G.; Ribeiro, C. Synthesis of g-C3N4/Nb2O5 heterostructures and their application in the removal of organic pollutants under visible and ultraviolet irradiation. Ceram. Int. 2017, 43, 3521–3530. [CrossRef] 19. Zhao, H.; Yu, H.; Quan, X.; Chen, S.; Zhang, Y.; Zhao, H.; Wang, H. Fabrication of atomic single layer

graphitic-C3N4 and its high performance of photocatalytic disinfection under visible light irradiation. Appl. Catal. B Environ. 2014, 152, 46–50. [CrossRef] 20. Wang, H.-H.; Zhang, B.; Li, X.-H.; Antonietti, M.; Chen, J.-S. Activating Pd nanoparticles on sol-gel

prepared porous g-C3N4/SiO2 via enlarging the Schottky barrier for efficient dehydrogenation of formic acid. Inorg. Chem. Front. 2016, 3, 1124–1129. [CrossRef] 21. Wang, X.; Yang, C.; Zhou, D.; Wang, Z.; Jin, M. Chemical co-precipitation synthesis and properties of

pure-phase BiFeO3. Chem. Phys. Lett. 2018, 713, 185–188. [CrossRef] 22. Gao, F.; Chen, X.Y.; Yin, K.B.; Dong, S.; Ren, Z.F.; Yuan, F.; Yu, T.; Zou, Z.G.; Liu, J.M. Visible-Light

Photocatalytic Properties of Weak Magnetic BiFeO3 Nanoparticles. Adv. Mater. 2007, 19, 2889–2892. [CrossRef] 23. Wang, X.; Mao, W.; Zhang, J.; Han, Y.; Quan, C.; Zhang, Q.; Yang, T.; Yang, J.; Li, X.; Huang, W. Facile fabrication

of highly efficient g-C3N4/BiFeO3 nanocomposites with enhanced visible light photocatalytic activities. J. Colloid Interface Sci. 2015, 448, 17–23. [CrossRef][PubMed] 24. Luo, W.; Zhu, L.; Wang, N.; Tang, H.; Cao, M.; She, Y. Efficient Removal of Organic Pollutants with Magnetic

Nanoscaled BiFeO3 as a Reusable Heterogeneous Fenton-Like Catalyst. Environ. Sci. Technol. 2010, 44, 1786–1791. [CrossRef][PubMed] 25. Chen, L.; Chen, N.; Wu, H.; Li, W.; Fang, Z.; Xu, Z.; Qian, X. Flexible design of carbon nanotubes grown on carbon nanofibers by PECVD for enhanced Cr(VI) adsorption capacity. Sep. Purif. Technol. 2018, 207, 406–415. [CrossRef]

26. Hu, X.; Wang, W.; Xie, G.; Wang, H.; Tan, X.; Jin, Q.; Zhou, D.; Zhao, Y. Ternary assembly of g-C3N4/graphene oxide sheets /BiFeO3 heterojunction with enhanced photoreduction of Cr(VI) under visible-light irradiation. Chemosphere 2019, 216, 733–741. [CrossRef][PubMed] 27. Noimark, S.; Weiner, J.; Noor, N.; Allan, E.; Williams, C.K.; Shaffer, M.S.P.; Parkin, I.P. Dual-Mechanism Antimicrobial Polymer-ZnO Nanoparticle and Crystal Violet-Encapsulated Silicone. Adv. Funct. Mater. 2015, 25, 1367–1373. [CrossRef] 28. Deng, Y.; Tang, L.; Zeng, G.; Feng, C.; Dong, H.; Wang, J.; Feng, H.; Liu, Y.; Zhou, Y.; Pang, Y. Plasmonic

resonance excited dual Z-scheme BiVO4/Ag/Cu2O nanocomposite: Synthesis and mechanism for enhanced photocatalytic performance in recalcitrant antibiotic degradation. Environ. Sci. Nano 2017, 4, 1494–1511. [CrossRef] 29. Anirudhan, T.S.; Shainy, F.; Christa, J. Synthesis and characterization of polyacrylic acid- grafted-carboxylic graphene/titanium nanotube composite for the effective removal of enrofloxacin from aqueous solutions: Adsorption and photocatalytic degradation studies. J. Hazard. Mater. 2017, 324, 117–130. [CrossRef]

30. Tseng, W.J.; Lin, R.D. BiFeO3/alpha-Fe2O3 core/shell composite particles for fast and selective removal of methyl orange dye in water. J. Colloid Interface Sci. 2014, 428, 95–100. [CrossRef] 31. Sankar Ganesh, R.; Sharma, S.K.; Sankar, S.; Divyapriya, B.; Durgadevi, E.; Raji, P.; Ponnusamy, S.; Muthamizhchelvan, C.; Hayakawa, Y.; Kim, D.Y. Microstructure, structural, optical and piezoelectric

properties of BiFeO3 nanopowder synthesized from sol-gel. Curr. Appl. Phys. 2017, 17, 409–416. [CrossRef] 32. Kang, H.W.; Lim, S.N.; Song, D.; Park, S.B. Organic-inorganic composite of g-C3N4-SrTiO3:Rh photocatalyst for improved H2 evolution under visible light irradiation. Int. J. Hydrog. Energy 2012, 37, 11602–11610. [CrossRef] 33. Cheng, N.; Tian, J.; Liu, Q.; Ge, C.; Qusti, A.H.; Asiri, A.M.; Al-Youbi, A.O.; Sun, X. Au-Nanoparticle-Loaded Graphitic Carbon Nitride Nanosheets: Green Photocatalytic Synthesis and Application toward the Degradation of Organic Pollutants. ACS Appl. Mater. Interfaces 2013, 5, 6815–6819. [CrossRef][PubMed] Int. J. Environ. Res. Public Health 2019, 16, 3219 17 of 18

34. Chang, Y.P.; Ren, C.L.; Qu, J.C.; Chen, X.G. Preparation and characterization of Fe3O4/graphene nanocomposite and investigation of its adsorption performance for aniline and p-chloroaniline. Appl. Surf. Sci. 2012, 261, 504–509. [CrossRef] 35. Hu, X.J.; Liu, Y.G.; Zeng, G.M.; Wang, H.; Hu, X.; Chen, A.W.; Wang, Y.Q.; Guo, Y.M.; Li, T.T.; Zhou, L.; et al. Effect of aniline on cadmium adsorption by sulfanilic acid-grafted magnetic graphene oxide sheets. J. Colloid Interface Sci. 2014, 426, 213–220. [CrossRef][PubMed] 36. Zhou, C.; Lai, C.; Huang, D.; Zeng, G.; Zhang, C.; Cheng, M.; Hu, L.; Wan, J.; Xiong, W.; Wen, M.; et al. Highly porous carbon nitride by supramolecular preassembly of monomers for photocatalytic removal of sulfamethazine under visible light driven. Appl. Catal. B Environ. 2018, 220, 202–210. [CrossRef]

37. Yang, M.Q.; Zhang, N.; Xu, Y.J. Synthesis of fullerene, carbon nanotube, and graphene-TiO2 nanocomposite photocatalysts for selective oxidation: A comparative study. ACS Appl. Mater. Interfaces 2013, 5, 1156–1164. [CrossRef][PubMed]

38. An, J.; Zhang, G.; Zheng, R.; Wang, P. Removing lignin model pollutants with BiFeO3/g-C3N4 compound as an efficient visible-light-heterogeneous Fenton-like catalyst. J. Environ. Sci. 2016, 48, 218–229. [CrossRef] [PubMed]

39. Wang, X.; Lu, M.; Ma, J.; Ning, P.; Che, L. Synthesis of K-doped g-C3N4/carbon microsphere@graphene composite with high surface area for enhanced adsorption and visible photocatalytic degradation of tetracycline. J. Taiwan Inst. Chem. Eng. 2018, 91, 609–622. [CrossRef]

40. Di, J.; Li, S.; Zhao, Z.; Huang, Y.; Jia, Y.; Zheng, H. Biomimetic CNT@TiO2 composite with enhanced photocatalytic properties. Chem. Eng. J. 2015, 281, 60–68. [CrossRef] 41. Li, S.; Li, Z.; Ke, B.; He, Z.; Cui, Y.; Pan, Z.; Li, D.; Huang, S.; Lai, C.; Su, J. Magnetic multi-walled carbon nanotubes modified with polyaluminium chloride for removal of humic acid from aqueous solution. J. Mol. Liq. 2019, 279, 241–250. [CrossRef] 42. Booshehri, A.Y.; Chun-Kiat Goh, S.; Hong, J.; Jiang, R.; Xu, R. Effect of depositing silver nanoparticles on

BiVO4 in enhancing visible light photocatalytic inactivation of bacteria in water. J. Mater. Chem. A 2014, 2, 6209–6217. [CrossRef] 43. Bai, X.; Du, Y.; Hu, X.; He, Y.; He, C.; Liu, E.; Fan, J. Synergy removal of Cr (VI) and organic pollutants over

RP-MoS2/rGO photocatalyst. Appl. Catal. B Environ. 2018, 239, 204–213. [CrossRef] 44. Ning, J.; He, Q.; Luo, X.; Wang, M.; Liu, D.; Wang, J.; Li, G.; Liu, J. Determination of Uric Acid in Co-Presence of Dopamine and Ascorbic Acid Using Cuprous Oxide Nanoparticle-Functionalized Graphene Decorated Glassy Carbon Electrode. Catalysts 2018, 8, 407. [CrossRef]

45. Zhang, Y.; Zhang, Q.; Shi, Q.; Cai, Z.; Yang, Z. Acid-treated g-C3N4 with improved photocatalytic performance in the reduction of aqueous Cr(VI) under visible-light. Sep. Purif. Technol. 2015, 142, 251–257. [CrossRef] 46. Liu, W.; Cao, L.; Cheng, W.; Cao, Y.; Liu, X.; Zhang, W.; Mou, X.; Jin, L.; Zheng, X.; Che, W.; et al. Single-site active cobalt-based photocatalyst with long carriers lifetime for spontaneous overall water splitting. Angew. Chem. Int. Ed. 2017, 56, 9312–9317. (In English) [CrossRef][PubMed]

47. Liu, W.; Wang, M.; Xu, C.; Chen, S. Facile synthesis of g-C3N4/ZnO composite with enhanced visible light photooxidation and photoreduction properties. Chem. Eng. J. 2012, 209, 386–393. [CrossRef] 48. Qin, F.; Wang, R.; Li, G.; Tian, F.; Zhao, H.; Chen, R. Highly efficient photocatalytic reduction of Cr(VI) by bismuth hollow nanospheres. Catal. Commun. 2013, 42, 14–19. [CrossRef]

49. Wang, H.; Yuan, X.; Wang, H.; Chen, X.; Wu, Z.; Jiang, L.; Xiong, W.; Zeng, G. Facile synthesis of Sb2S3/ultrathin g-C3N4 sheets heterostructures embedded with g-C3N4 quantum dots with enhanced NIR-light photocatalytic performance. Appl. Catal. B Environ. 2016, 193, 36–46. [CrossRef] 50. Ibrahim, S.; Shuy, W.Z.; Ang, H.-M.; Wang, S. Preparation of bioadsorbents for effective adsorption of a reactive dye in aqueous solution. Asia Pac. J. Chem. Eng. 2010, 5, 563–569. [CrossRef]

51. Alam, U.; Khan, A.; Bahnemann, D.; Muneer, M. Synthesis of Co doped ZnWO4 for simultaneous oxidation of RhB and reduction of Cr(VI) under UV-light irradiation. J. Environ. Chem. Eng. 2018, 6, 4885–4898. [CrossRef] 52. Tang, L.; Feng, C.; Deng, Y.; Zeng, G.; Wang, J.; Liu, Y.; Feng, H.; Wang, J. Enhanced photocatalytic activity of

ternary Ag/g-C3N4/NaTaO3 photocatalysts under wide spectrum light radiation: The high potential band protection mechanism. Appl. Catal. B Environ. 2018, 230, 102–114. [CrossRef] Int. J. Environ. Res. Public Health 2019, 16, 3219 18 of 18

53. Zhang, D.; Cui, S.; Yang, J. Preparation of Ag2O/g-C3N4/Fe3O4 composites and the application in the photocatalytic degradation of Rhodamine B under visible light. J. Alloy. Compd. 2017, 708, 1141–1149. [CrossRef] 54. Zhang, Y.; Ma, H.-L.; Peng, J.; Zhai, M.; Yu, Z.Z. Cr(VI) removal from aqueous solution using chemically reduced and functionalized graphene oxide. J. Mater. Sci. 2012, 48, 1883–1889. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).