Magnetically separable CeO₂/CoFe₂O₄ heterojunction photocatalysts for dye degradation: characterization and mechanism

ABSTRACT

This study developed magnetically separable CeO₂/CoFe₂O₄ heterojunction photocatalysts for dye degradation to reduce water pollution. The CeO₂/CoFe₂O₄ nanocomposites with different CeO₂ ratios (10, 20, and 30 wt%) were prepared by a precipitation method. The nanocrystalline 20%CeO₂/CoFe₂O₄ photocatalyst with a surface area of 117 m².g⁻¹, E_g 2.37 eV, and uniform distribution of quasi-spherical CeO₂ minor phase (cubic fluorite structure with particle size 3 ± 1 nm) throughout the distorted spherical CoFe₂O₄ major phase (cubic spinel structure with particle size 9 ± 3 nm) showed the best photocatalytic performance for methylene blue (MB) degradation under UV light irradiation at the efficiency of 88.7% and the apparent rate constant (k) of 0.0138 min⁻¹. The 20%CeO₂/CoFe₂O₄ photocatalyst had a good magnetic response and could be easily separated from the dye solution by applying an external magnetic field. The reusability test showed lower photodegradation efficiency due to the adsorption of residual MB dye on the photocatalyst surface. The formation of heterojunction at the interfaces of CeO₂/CoFe₂O₄ facilitated the photogenerated charge separation and increased availability of active species of holes (h⁺) and hydroxy radicals (^∙OH), which were the key factors in improving the photocatalytic activity.

GRAPHICAL ABSTRACT

HIGHLIGHTS

• Improved photocatalysis in CeO₂-coupled CoFe₂O₄ composites.

• Heterojunctions in CeO₂/CoFe₂O₄ increase the availability of active species.

• External magnetic field separates the photocatalyst from the dye solution.

KEYWORDS:

• Photocatalyst
• dye degradation
• heterojunction
• composite
• cobalt ferrite

1. Introduction

Water pollution associated with the growth of industrialization and urbanization can cause several adverse effects on ecosystems and human health [1,2]. The use of synthetic dyes in various industries, including the textile, printing, tannery, paint, paper, pulp, cosmetic, food, pharmaceutical, and photoelectrochemical industries, is one of the main causes of water pollution as the dyeing process consumes and discharges a large amount of water and dye effluent [2–4]. Many treatment processes and technologies have been used to remove synthetic dyes from industrial wastewater to reduce their environmental impact [2]. Semiconductor photocatalysis is a green and cost-effective technology [5] that can be used to eliminate organic synthetic dyes from wastewater by photo-inducing charge generation of electron and hole pairs (e⁻/h⁺) at semiconductor surfaces to produce reactive oxygen species (ROS) through oxidation and reduction reactions. Subsequently, these reactive species react with the dye molecules and decompose them into less toxic products [1,4,6,7].

Heterogeneous photocatalysts, e.g. TiO₂, ZnO, CdSe, WO₃, Bi₂WO₆, BiVO₄, and Fe₂O₃, have been used to treat organic dye wastewater [4,8–14]. Magnetic nanosized spinel ferrite photocatalysts, e.g. CoFe₂O₄, NiFe₂O₄, CuFe₂O₄, ZnFe₂O₄ and SnFe₂O₄ [1,15–18], have recently received great attention as they can overcome a technical problem in separating and recovering the photocatalyst from a dye solution after the photodegradation reaction that limits the practical application and may cause secondary pollution to the environment [1,15,16]. The suspended magnetic photocatalyst particles can simply be separated by applying an external magnetic field [1,19]. CoFe₂O₄, an n-type semiconductor with a cubic inverse spinel structure, is a ferrimagnetic material with high coercivity, moderate saturation magnetization, high absorption capability, high physical and chemical stability, and non-toxicity [1,20–22]. Even though CoFe₂O₄ photocatalysts possess a high potential to be used as magnetically separable photocatalysts, their photocatalytic activity is relatively low due to a narrow band gap energy. The low E_g of CoFe₂O₄ at 1.1–2.3 eV [1,19,22,23] causes a fast electron-hole pair recombination rate and poor charge separation performance [1,24]. Coupling CoFe₂O₄ with other dissimilar band structure semiconductors to form heterojunction photocatalysts is an effective approach to improve the photocatalytic efficiency since the formation of heterojunction interface can prolong the photo-induced electron–hole pair recombination, which interns increasing the charge participation in the photocatalytic reactions [1,9,25–27].

Heterojunction systems of TiO₂/CoFe₂O₄ [28–30], ZnO/CoFe₂O₄ [31], AgBr/CoFe₂O₄ [32], BiVO₄/CoFe₂O₄ [33], Ni/CoFe₂O₄ [34], graphene/CoFe₂O₄ [35], rGO/CoFe₂O₄ [36], Ag₃PO₄/CoFe₂O₄ [19,37], g-C₃N₄/CoFe₂O₄ [38], and polyaniline/CoFe₂O₄ [39] composite photocatalysts were reported to increase the degradation efficiency of organic dyes, e.g. methylene blue (MB), methyl orange (MO), rhodamine B (RhB), reactive red 120 (RR-120) crystal violet (CV), acridine orange (AO), rhodamine 6 G (R-6 G), active black (BL-G), and active red (RGB), as well as showing a good magnetic response. CeO₂, an n-type semiconductor with an E_g of 2.3–3.2 eV [40–43], is an interesting choice for coupling with CoFe₂O₄ [44,45] since it has desirable photocatalytic properties, including high UV absorption, physiochemical stability, and large oxygen storage and release capacity [46,47]. Moreover, CeO₂ has suitable band edge positions to form heterojunction at the interface of CoFe₂O₄, which can increase the charge separation, so coupling CeO₂ with CoFe₂O₄ would benefit to optimize the photocatalytic activity of CoFe₂O₄. The synthesis of CoFe₂O₄/CeO₂ nanocomposites was studied for possible application in photocatalysis [44], and core-shell CoFe₂O₄/CeO₂ photocatalyst was reported to effectively activate sodium persulfate (Na₂S₂O₈) for orange II dye degradation [45]. However, the nanocomposite system of minor phase CeO₂ with CoFe₂O₄ heterojunction photocatalyst for direct dye degradation has not been studied in detail yet.

In this study, magnetically separable CeO₂/CoFe₂O₄ heterojunction photocatalysts were prepared and characterized. The photocatalytic efficiency and rate constant for dye degradation were examined, and the possible mechanism for the photocatalytic degradation over the CeO₂/CoFe₂O₄ system was also proposed.

2. Material and methods

2.1. Photocatalyst preparation

CoFe₂O₄ nanoparticles were prepared by the sono-assisted-precipitation method [21]. Firstly, 0.1 mol of cobalt(II) nitrate hexahydrate (Co(NO₃).6 H₂O; 99% purity, Carlo Erba) and 0.2 mol of ferric(III) nitrate nonahydrate (Fe(NO₃).9 H₂O; 99% purity, Carlo Erba) were dissolved in 100 mL of deionized water and stirred with a magnetic bar at 80°C for 2 h. The mixed solution was added dropwise into 100 mL of 10 M sodium hydroxide solution (NaOH; 98% purity, Carlo Erba) that was simultaneously irradiated with the high-intensity ultrasonic wave at 20 kHz 150 W/cm² (Ti-horn, Sonics model VCX 750, Vibracell). Notably, the highly basic concentration of precipitation medium at 10 NaOH was used as it could increase the nucleation rate, pH, and ionic strength, which resulted in CoFe₂O₄ with a small crystallite size [48]. The reaction was conducted for 30 min. After that, the black precipitated were separated in a centrifuge, washed with distilled water until neutral (pH ~ 7), and dried at 90°C for 24 h to obtain CoFe₂O₄ nanoparticles. CeO₂ nanoparticles were prepared by a precipitation method [42,43]. 100 mL of 0.1 M ammonium cerium(IV) nitrate solution ((NH₄)₂Ce(NO₃)₆; 99% purity, Acros Organics) was adjusted to pH 12 with 0.5 M NaOH solution, followed by magnetic stirring at room temperature for 2 h. Next, the pale-yellow precipitate was separated in a centrifuge, washed with distilled water until neutral (pH ~ 7), and dried at 90°C for 24 h. To prepare CeO₂/CoFe₂O₄ heterojunction photocatalysts, the required amounts of CeO₂ (10, 20, and 30 wt%) and 2 g of CeFe₂O₄ powders were dispersed in 100 mL deionized water under a high-intensity ultrasonic irradiation at 20 kHz 150 W.cm⁻² for 15 min. The products were separated in a centrifuge, dried at 90°C for 24 h, and finally ground in an agate mortar to obtain 10%CeO₂/CoFe₂O₄ and 20%CeO₂/CoFe₂O₄ and 30%CeO₂/CoFe₂O₄ nanocomposite photocatalysts.

2.2. Characterization

Crystal structures of photocatalysts were examined by an X-ray diffractometer (XRD; Rigaku Smartlab, CuKα, 40 kV 30 mA), using indices from the JCPDS (ICDD) database. The lattice parameter (a) was calculated from the Bragg equation, and the crystal spacing was determined by assuming a cubic symmetry. The crystallite size (D) was estimated using the Debye-Scherrer’s equation as shown in Equation 1(1) $\mathrm{D}=\frac{\mathrm{k}\mathit{\lambda }}{{\mathit{\beta }}_{\mathrm{hkl}}cos\mathit{\theta }}$(1) [49]:

(1) $\mathrm{D}=\frac{\mathrm{k}\mathit{\lambda }}{{\mathit{\beta }}_{\mathrm{hkl}}cos\mathit{\theta }}$(1)

where k is a constant taken as 0.94, λ is X-ray wavelength (CuKα radiation = 0.15418 nm), β_hkl is the full width at half maximum (FWHM) of the intensity of the observed diffraction peak, hkl are the Miller indices of the planes being analyzed, which are (311) plane for CoFe₂O₄, (111) plane for CeO₂ and (311) plane of CoFe₂O₄ phase for the CeO₂/CoFe₂O₄ composites, and θ is the diffraction angle. Microstructures of the photocatalysts were studied by a scanning electron microscope (SEM; Hitachi SU3500) and transmission electron microscope (TEM; FEI Tecnai G2 20 Twin). Elemental compositions were examined by an energy-dispersive spectrometer (EDS; Oxford instrument X-MAX 20) equipped with SEM and energy-dispersive X-ray fluorescence spectrometer (EDXRF; Rigaku NEX CG). The average particle size and distribution were obtained from the ImageJ program by measuring at least 100 sample particles observed in TEM images. The Brunauer–Emmett–Teller (BET) specific surface area was measured by adsorption-desorption of nitrogen gas (N₂) with Autosorb (Autosorb-1). The optical property was studied using a diffuse reflectance ultraviolet-visible spectrophotometer (UV-Vis DRS; Thermo Scientific evolution 201). The optical band gap energy (E_g) of semiconducting photocatalysts was calculated using the Tuac’s equation (Equation 2(2) $\mathit{\alpha }\mathrm{h\nu }={\left(\mathit{\beta }\mathrm{h\nu }-\mathrm{Eg}\right)}^n$(2) ) [34,50]:

(2) $\mathit{\alpha }\mathrm{h\nu }={\left(\mathit{\beta }\mathrm{h\nu }-\mathrm{Eg}\right)}^n$(2)

When α is the absorption coefficient, h is the Plank’s constant (6.626 × 10⁻³⁴ J.s), ν is the frequency of light (s⁻¹), and n is a constant depending on the nature of band transitions, where n = 2 for the indirect transition in CoFe₂O_4, CeO₂, and CeO₂/CoFe₂O₄ composites [51,52], β is the absorption constant for an indirect transition metal, which is equal to 1. By plotting between the (αhν)² in the y-axis versus hν in the x-axis, E_g in a unit of eV was evaluated from the intercept of the linear region with the x-axis ((αhν)² = 0) in the plot. The magnetic property to evaluate the magnetic separability of the photocatalysts from the dye solution was tested at room temperature using a vibrating sample magnetometer (VSM, in-house Kasetsart University; Thailand) at a maximum applied field of 10,000 Oe. The molecular structure of photocatalysts before and after the photodegradation reaction was studied using a Fourier transform infrared spectrometer (FT-IR; Perkin Elmer, Spectrum Gx, potassium bromide (KBr) mixing in transmission mode).

2.3. Photocatalytic degradation of dye

The photocatalytic activity of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ composites for degradation of methylene blue dye (MB; C₁₆H₁₈N₃SCl.3 H₂O, cationic dye, 93% purity, Carlo Erba) and Congo red dye (CR; C₃₂H₂₂N₆Na₂O₆S₂, anionic azo dye, 35% purity, Carlo Erba) in aqueous solution under UV light radiation was studied. 14 mg of photocatalyst powders were dispersed in 50 mL of 10 mg.L⁻¹ (ppm) MB solution or 30 mg.L⁻¹ CR solution by stirring with a magnetic bar, and the pH of the solution at which the reaction carried out was pH 7.3 for MB degradation and pH 7.1 for CR degradation. At the first 30 min, the photocatalyst suspension was stirred in the dark to reach adsorption-desorption equilibrium. After that, the suspension was irradiated with a UV-C mercury lamp (Tokiva, 15 W, luminous intensity 7800 cd, UV-C output 4.9 W with peak light emission intensity at a wavelength of 254 nm and other emissions at 365, 408, and 436 nm) for 240 min. The suspension was collected every 30 min and centrifugally separated to remove the suspended photocatalysts. The reusability was tested by magnetically separating the solid photocatalysts from the dye solution, washing them with distilled water, drying at 90°C for 24 h, and reusing in the same experimental procedures. The adsorption and photodegradation of dye were examined by measuring the absorbance of MB solution at λ_max of 664 nm and CR solution at λ_max of 493 nm using a UV-visible spectrophotometer (Thermo Scientific evolution 201). To optimize the photodegradation reaction, the effects of photocatalyst dosage and initial dye concentration were studied by varying the photocatalyst dosages at 7, 14, and 21 mg and the MB concentrations at 5, 10, 15, and 20 mg.L⁻¹. The dye adsorption and photodegradation efficiency was calculated from Equation 3(3) $\mathrm{Degradation}\mathrm{efficiency}\left(\mathrm{\%}\right)=\frac{{\mathit{C}}_{0-}\mathit{C}}{C_0}\times 100\mathrm{\%}$(3) [34,50]:

(3) $\mathrm{Degradation}\mathrm{efficiency}\left(\mathrm{\%}\right)=\frac{{\mathit{C}}_{0-}\mathit{C}}{C_0}\times 100\mathrm{\%}$(3)

When C₀ is the initial concentration of the dye solution, and C is the concentration of the dye solution after the adsorption and photodegradation at a certain time. The kinetics of dye photodegradation were obtained by fitting the pseudo-first-order equation (Equation 4(4) $-\mathit{ln}\frac{C}{C_0}=\mathit{kt}$(4) ) [34,50], where k is the apparent rate constant (min⁻¹), and t is irradiation time (min). The apparent rate constant (k) was estimated from the slope of the plot between -ln (C/C₀) on the y-axis versus t on the x-axis.

(4) $-\mathit{ln}\frac{C}{C_0}=\mathit{kt}$(4)

3. Results and discussion

3.1. Materials characterization

XRD patterns of CoFe₂O₄, CeO₂ and CeO₂/CoFe₂O₄ composites with different wt% of CeO₂ prepared by the precipitation method without a calcination process are shown in Figure 1. The XRD pattern of CoFe₂O₄ showed diffraction peaks at 2θ = 30.4°, 35.9°, 43.4°, 57.1° and 63.2° corresponded to the (220), (311), (400), (511), and (440) planes of CoFe₂O₄ with a cubic spinel structure (JCPDS No. 22–1086, space group Fd-3 m). The XRD pattern of CeO₂ showed broad diffraction peaks at 2θ = 28.5°, 33.0°, 47.4°, and 56.3° corresponded to the (111), (200), (220), and (311) planes of CeO₂ with a cubic fluorite structure (JCPDS No. 34–0394, space group Fm-3 m). For 10%CeO₂/CoFe₂O₄, 20%CeO₂/CoFe₂O₄, and 30%CeO₂/CoFe₂O₄ composites, XRD patterns showed the diffraction peaks indexed to both major and minor phases of CoFe₂O₄ and CeO₂ in which the intensity of CeO₂ diffraction peaks increased with increasing %wt of CeO₂. Table 1 shows the calculated cubic lattice parameter (a) and crystallite size of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ composites. The lattice parameters of pure CoFe₂O₄ (8.3235 Å) and CeO₂ (5.4261 Å) were close to those reported in the JCPDS (ICDD) database (CoFe₂O₄, JCPDS No. 22–1086 a = 8.391Å and CeO₂, JCPDS No. 34–0394 a = 5.4113Å). No significant change in the lattice parameters of the CeO₂/CoFe₂O₄ composites (8.3213–8.3274 Å) compared to CoFe₂O₄, suggesting the formation of composites without cation substitution. The crystallite size of CoFe₂O₄ and the CeO₂/CoFe₂O₄ composites, which were estimated from (311) plane of CoFe₂O₄, varied from 12.1–13.5 nm, while CeO₂ ((111) plane) was 3.2 nm, which was in the range of nanocrystalline materials.

Figure 1. XRD patterns of CoFe₂O₄, CeO₂ and CeO₂/CoFe₂O₄ composites.

Table 1. Lattice parameter, crystallite size, and surface area of CoFe₂O₄, CeO₂ and CeO₂/CoFe₂O₄ composites * estimated from CoFe₂O₄ phase.

Photocatalyst	Lattice parameter [Å]	Crystallite size (nm)	Surface area (m^2.g^−1)
CoFe2O4	8.3235	12.1	163
CeO2	5.4261	3.2	228
10%CeO2/CoFe2O4	8.3274	13.2*	116
20%CeO2/CoFe2O4	8.3274	13.5*	117
30%CeO2/CoFe2O4	8.3213	12.1*	142

SEM and TEM images of pure CoFe₂O₄ and CeO₂, with particle size distributions inset, are shown in Figure 2. CoFe₂O₄ had distorted spherical particles, while CeO₂ was non-uniform quasi-spherical particles. Particle agglomeration was observed in both CoFe₂O₄ and CeO₂. The average particle size of CoFe₂O₄ (9 ± 3 nm) and CeO₂ (3 ± 1 nm) agreed well with the crystallite sizes estimated by the XRD study (Table 1), confirming the formation of nanocrystalline CoFe₂O₄ and CeO₂ [53]. Figure 3 shows SEM images of 10%CeO₂/CoFe₂O₄, 20%CeO₂/CoFe₂O₄, and 30%CeO₂/CoFe₂O₄ composites plus EDS elemental mappings and fractions (wt%). The EDS elemental mappings showed uniform distribution of the CeO₂ minor phase throughout the CoFe₂O₄ major phase, and it is interesting to note that a higher degree of CeO₂ agglomeration was observed in 30%CeO₂/CoFe₂O₄ composite. When considering the elemental fractions, the Ce fraction in the composites increased with increasing the CeO₂ content, whereas the Co and Fe fraction decreased. The EDXRF was used to determine the oxide composition of composites and the actual wt% of CeO₂ in 10%CeO₂/CoFe₂O₄, 20%CeO₂/CoFe₂O₄ and 30%CeO₂/CoFe₂O₄ was 7.73 wt%, 11.17 wt%, and 23.90 wt%, respectively. Figure 4 shows TEM and high-resolution TEM (HRTEM) images of 20%CeO₂/CoFe₂O₄. The composite consisted of a mixed morphology of distorted spherical-shaped particles of CoFe₂O₄ and quasi-spherical CeO₂. The lattice fringes could be seen in the high-resolution image (Figure 4(b)). The interplanar spacings (d-spacings) of 0.21 nm and 0.25 nm corresponded to the reflection of the (400) and (311) planes of CoFe₂O₄, while the interplanar spacing of 0.31 nm was attributed to (111) plane of CeO₂. The interfacial contact between CoFe₂O₄ and CeO₂ phases observed in the HRTEM image confirmed the formation of heterojunctions of CeO₂/CoFe₂O₄.

Figure 2. SEM, TEM images and particle size distributions of (a and b) CoFe₂O₄ and (c and d) CeO₂.

Figure 3. SEM images plus elemental mappings and elemental fractions of CeO₂/CoFe₂O₄ composites.

Figure 4. (a) TEM and (b) HRTEM and lattice fringes of 20%CeO₂/CoFe₂O₄.

The BET surface area of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ composites are shown in Table 1. The specific surface area of CoFe₂O₄ (163 m².g⁻¹) was lower than that of CeO₂ (228 m².g⁻¹), which corresponded well with the larger particle size of CoFe₂O₄ (9 ± 3 nm) than CeO₂ (3 ± 1 nm). The surface area of CoFe₂O₄ and CeO₂ obtained in this work are comparable to those previously reported for nanosized CoFe₂O₄ (160–285 m².g⁻¹ [53–55]) and CeO₂ (200–457 m².g⁻¹ [56–58]). The composites, 10%CeO₂/CoFe₂O₄ (116 m².g⁻¹), 20%CeO₂/CoFe₂O₄ (117 m².g⁻¹) and 30%CeO₂/CoFe₂O₄ (142 m².g⁻¹), had lower surface area than those of the pure CoFe₂O₄ and CeO₂, which could be attributed to particle agglomeration due to the physical bonding between CeO₂ and CoFe₂O₄ phases [44,45]. The surface area of the composites tended to increase with increasing the amount of higher surface area phase of CeO₂. Notably, the surface area of 10%CeO₂/CoFe₂O₄ and 20%CeO₂/CoFe₂O₄ was close at 116 and 117 m².g⁻¹, and this was probably because, at these ratios, CeO₂ was uniformly dispersed throughout the CoFe₂O₄ phase to create the interfacial contact between both phases.

Figure 5 shows the diffuse reflectance UV-visible spectra inset with Tuac’s plot for determining optical band gap energy (E_g) of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ composites. The absorption spectra showed that pure CoFe₂O₄ with a black surface could absorb a broad absorption range of UV light (200–400 nm) and visible light (400–800 nm) regions, while CeO₂ mainly absorbed UV light. The absorption intensity of CeO₂/CoFe₂O₄ composites was lower than CoFe₂O₄ but tended to increase compared to CeO₂, especially in the visible light region. The E_g values in Table 2 show that the E_g of pure CoFe₂O₄ at 2.15 eV and CeO₂ at 3.07 eV

Figure 5. Diffuse reflectance UV-visible spectra inset with Tuac’s plot for E_g determination of CoFe₂O₄, CeO₂ and CeO₂/CoFe₂O₄ composites.

Table 2. E_g and degradation efficiency of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ photocatalysts.

Photocatalyst	Eg(eV)	% Adsorption (30 min)	% Photodegradation (30–180 min)	% Total (180 min)	k (min^−1) (30–150 min)
CoFe2O4	2.15	14.4	19.6	34.0	0.0014
CeO2	3.07	6.8	45.2	52.0	0.0024
10%CeO2/CoFe2O4	2.19	15.5	55.6	71.1	0.0039
20%CeO2/CoFe2O4	2.37	19.6	54.2	73.8	0.0041
30%CeO2/CoFe2O4	2.58	10.6	57.1	67.7	0.0034

are comparable to the values reported previously (CoFe₂O₄ 1.1–2.3 eV [1,19,22,23] and CeO₂ 2.3–3.2 eV [40–43]). For the CeO₂/CoFe₂O₄ composites, coupling CoFe₂O₄ with a higher E_g phase, CeO₂, resulted in an increase of E_g values, in which the E_g values of 10%CeO₂/CoFe₂O₄, 20%CeO₂/CoFe₂O₄, and 30%CeO₂/CoFe₂O₄ were 2.19 eV, 2.37 eV, and 2.58 eV, respectively.

3.2. Photocatalytic degradation of dye

The photocatalytic activity of CoFe₂O₄, CeO₂, and CeO₂/CoFe₂O₄ photocatalysts was determined by methylene blue dye (MB) degradation under UV irradiation. The degradation efficiency is shown in Figure 6 and Table 2. The blank control of 10 mg.L⁻¹ MB in the absence of the photocatalysts showed a stable concentration of MB under UV irradiation over the studied period at 240 min. In this study, the suspension of the 14 mg photocatalysts in 50 mL of 10 mg.L⁻¹ MB solution at pH 7.3 was stirred with a magnetic bar in the dark for 30 min to reach the adsorption-desorption equilibrium. At 30 min before the UV irradiation, CoFe₂O₄ had higher adsorption efficiency than CeO₂, although it possessed a lower surface area and this could be because the pH of 10 mg.L⁻¹ MB solution at 7.3 was higher than the point of zero charge (pH_pzc) of CoFe₂O₄ = 7.2 [59] but lower than the pH_pzc of CeO₂ = 8.1 [60], which resulted in a greater electrostatic attraction between the negatively charged CoFe₂O₄ surface and the cationic MB dye molecules [59,61]. In the case of the composites, the adsorption efficiency of 20%CeO₂/CoFe₂O₄ was the highest due to the good dispersion of the CeO₂ throughout the CoFe₂O₄ matrix. Increasing the CeO₂ content excess 20 wt% to form 30%CeO₂/CoFe₂O₄ led to the agglomeration of the CeO₂ phase as observed in the SEM-EDS study and thus lowered the adsorption efficiency. After the UV irradiation, the MB photodegradation efficiency of all photocatalysts increased sharply with increasing the reaction times from 30 to 180 min and became steady from 180 to 240 min. According to the photodegradation efficiency shown in Table 2, CeO₂ is more active for MB degradation than CoFe₂O₄. The composite photocatalysts, which possessed lower surface area (Table 1), showed better photocatalytic performance than pure CoFe₂O₄ and CeO₂, and this suggested that the formation of CeO₂/CoFe₂O₄ heterojunctions played more role in the photocatalytic activity than the surface area. The total MB degradation efficiency at the reaction time of 180 min was shown to follow the order: CoFe₂O₄ (34.0%) < CeO₂ (52.0%) < 30%CeO₂/CoFe₂O₄ (67.7%) < 10%CeO₂/CoFe₂O₄ (71.1%) < 20%CeO₂/CoFe₂O₄ (73.8%). The kinetics of MB photodegradation estimated by the plot between -ln (C/C₀) versus UV irradiation time (t) for 120 min time intervals from the time at 30 to 150 min are shown in Figure 7. The plots of all photocatalysts followed the pseudo-first-order reaction and the apparent rate constant (k) obtained from the slope was summarized in Table 2. Notably, the k values were estimated from the longest period data (30–150 min) that produced the slope with R² values close to 1. The composite photocatalyst possessed a higher photodegradation rate than the pure phases in which 20%CeO₂/CoFe₂O₄ showed the best performance with the k value of 0.0041 min⁻¹. This suggested that the optimum ratio for improving the photocatalytic activity was 20 wt% CeO₂ loading to CoFe₂O₄. Further increasing CeO₂ loading to 30 wt% lowered the photocatalytic activity due to agglomeration of the CeO₂ phase, which led to less interfacial contact between CeO₂ and CoFe₂O₄ phases.

Figure 6. The photocatalytic degradation of MB under UV irradiation.

Figure 7. The kinetics plots for the pseudo-first-order reaction of photodegradation of MB.

The photocatalytic degradation of anionic dye, Congo red (CR), under UV irradiation was also studied. The photodegradation of 50 mL of 30 mg.L⁻¹ CR solution (pH 7.1) by using 14 mg of 20%CeO₂/CoFe₂O₄ photocatalysts is shown in Figure 8(a,b). The initial concentration of CR solution was used at 30 mg.L⁻¹ due to the high adsorption capacity of anionic CR dye molecules on the positively charged surface of 20%CeO₂/CoFe₂O₄ (30 mg.L⁻¹ CR solution had pH = 7.1, which was lower than pH_pzc of CoFe₂O₄ = 7.2 [59] and CeO₂ = 8.1 [60]). The CR degradation efficiency at the reaction time of 240 min was 51.3%. The apparent rate constant (k) for photodegradation estimated from 210 min time intervals (30 to 240 min) was 0.0025 min⁻¹, which was lower than the photodegradation rate of cationic MB dye at 0.0041 min⁻¹. The reduction of the rate constant was probably related to the high adsorption capacity of anionic CR dye that prevented the photons from reaching the photocatalyst surface to activate the photocatalytic reaction [32].

Figure 8. The photocatalytic degradation and kinetics plots for the pseudo-first-order reaction of 20%CeO₂/CoFe₂O₄ in (a and b) CR, (c and d) MB at different photocatalyst dosage, and (e and f) MB at different initial concentration under UV irradiation.

The effects of photocatalyst dosage (7, 14, and 21 mg of 20%CeO₂/CoFe₂O₄) and initial MB dye concentration (5, 10, 15, and 20 mg.L⁻¹) at pH ~ 7.3 on the photodegradation reaction were examined, and the results are shown in Figure 8(c)-(f). The degradation efficiency increased with increasing the photocatalyst dosage, in which the photodegradation rate constant obtained from using 7, 14, and 21 mg of 20%CeO₂/CoFe₂O₄ photocatalysts in 50 mL of 10 mg.L⁻¹ MB solution were 0.0016, 0.0041, and 0.0065 min⁻¹, respectively. The increase of the rate constant with increasing the photocatalyst dosage was due to the higher number of active sites available for the photocatalytic reaction [32]. However, it should be noted that the excess amount of photocatalyst dosage may decrease the photocatalytic activity due to the light scattering factor [28,34]. Figure 8 (ae,bf) show the photocatalytic degradation of 14 mg 20%CeO₂/CoFe₂O₄ in 50 mL of MB solution with different initial concentrations at 5, 10, 15, and 20 mg.L⁻¹. The degradation efficiency of 5, 10, 15, and 20 mg.L⁻¹ MB solution at 180 min was 88.7%, 73.8%, 52.7%, and 42.7%, respectively. The photodegradation rate constant (k) decreased with the increase of initial MB concentration from 5 mg.L⁻¹ (0.0138 min⁻¹), 10 mg.L⁻¹ (0.0041 min⁻¹) to 15 mg.L⁻¹ (0.0036 min⁻¹) and became stable at 20 mg.L⁻¹ (0.0036 min⁻¹). The decrease in photodegradation rate in higher MB concentrations was because the darker color of the dye solution decreased the path length of photons reaching the photocatalyst surface [28,32,37]. Based on these results, the optimum photoreaction condition with 88.7% photodegradation efficiency and k = 0.0138 min⁻¹ was to use 14 mg of 20%CeO₂/CoFe₂O₄ photocatalyst in 50 mL of 5 mg.L⁻¹ MB solution under UV irradiation for 180 min. To further enhance the photocatalytic activity of 20%CeO₂/CoFe₂O₄ in the degradation MB may be done by varying the photoreaction condition by adjusting the pH of the dye solution to pH = 10 [32] to increase the surface adsorption/desorption of cationic dye molecules or adding electron acceptor such as peroxomonosulphate species (SO₅²⁻) [28] since these methods were reported to optimize the activity of CoFe₂O₄-based photocatalysts. Table 3 summarizes the activity of CoFe₂O₄-based photocatalysts for dye degradation.

Table 3. Summary of CoFe₂O₄-based photocatalysts for dye degradation.

Photocatalyst	Fabrication method	Dye	Light source	Degradation efficiency (%)	Rate constant, k (min^−1)	Irradiation time (min)	Ref.
TiO2/CoFe2O4	Precipitation	RR-120	Vis (150 W W-halogen lamp)	N/A	2.99 × 10^−7	60	[28]
5%CoFe2O4/TiO2	Hydrothermal method	MB, MO, CV, AO, RhB, and R-6G	UV (UVA, 15 W lamp) and Vis (15 W lamp)	75.3 (MB, UV) 50.4 (MB, Vis)	N/A	18	[29]
Core-shell CoFe2O4/ZnO	Sol-gel method	MB	UV (300 W Hg lamp)	N/A	0.0447	60	[31]
20%CoFe2O4/AgBr	Precipitation–deposition	MO	Vis (10 W LED lamp)	95	0.0710	60	[32]
0.2CoFe2O4/0.8BiVO4	Hydrothermal method	MB	Vis (50 W halogen lamp)	72.5	N/A	240	[33]
10%Ni/CoFe2O4	Mechanical alloying method	MB	Vis (150 W lamp)	93.6	0.0340	120	[34]
7.5%CoFe2O4/Ag3PO4	Precipitation	MB and RhB	Vis (500 W W-halogen lamp)	100 (MB)	N/A	30	[37]
41%CoFe2O4/ g-C3N4	Sol-gel method and calcination	MB	Vis (300 W Xe lamp)	97.3 with H2O2	N/A	180	[38]
40%graphene/CoFe2O4	Hydrothermal method	MB, RhB, MO, BL-G, and RGB	Vis (500 W Xe lamp)	100 (MB) 94 (RhB) 71 (MO) 66 (BL-G) 61 (RGB)	N/A	240	[35]
CoFe2O4/rGO	Precipitation-deposition	MB, MO, and RhB	Vis (150 W lamp)	89 (MB) 58 (MO) 64 (RhB)	0.0079 (MB) 0.0024 (MO) 0.0033 (RhB)	180	[36]
Core-shell CeO2/CoFe2O4	Hydrothermal method	Na2S2O8 activated O II	Vis	98.5	N/A	60	[45]
20%CeO2/CoFe2O4	Precipitation	MB and CR	UV (15 W Hg lamp)	88.7 (MB) 51.3 (CR)	0.0138 (MB) 0.0025 (CR)	180	This work

UV: Ultraviolet light; Vis: Visible light; RR-120: Reactive Red 120; MB: Methylene blue; MO: Methyl orange; CV: Crystal violet; AO: Acridine orange; RhB: Rhodamine B; R-6G: Rhodamine 6G; BL-G: Active black; RGB: Active red; O II: orange II; CR: Congo red.

The mechanism for photocatalytic degradation of MB over heterojunction CeO₂/CoFe₂O₄ photocatalyst under UV irradiation is shown in Figure 9. The energy bands were used for deriving the diagram by considering the relative positions of the conduction band (CB), valance band (VB) and oxidation potentials. The band edges of the VB (E_VB) and CB (E_CB) of CoFe₂O₄ and CeO₂ semiconductors were calculated from Equation 5(5) ${\mathrm{E}}_{\mathrm{VB}}=\mathit{\chi }-{\mathrm{E}}_{\mathrm{e}}+0.5{\mathrm{E}}_{\mathrm{g}}$(5) and 6(6) ${\mathrm{E}}_{\mathrm{CB}}={\mathrm{E}}_{\mathrm{VB}}-{\mathrm{E}}_{\mathrm{g}}$(6) [62]:

Figure 9. Photocatalytic degradation of MB dye over CeO₂/CoFe₂O₄ heterojunction photocatalysts.

(5) ${\mathrm{E}}_{\mathrm{VB}}=\mathit{\chi }-{\mathrm{E}}_{\mathrm{e}}+0.5{\mathrm{E}}_{\mathrm{g}}$(5)

(6) ${\mathrm{E}}_{\mathrm{CB}}={\mathrm{E}}_{\mathrm{VB}}-{\mathrm{E}}_{\mathrm{g}}$(6)

where χ is the absolute electronegativity of the semiconductors, 5.81 eV for CoFe₂O₄ [24] and 5.56 eV for CeO₂ [62], E_e is the energy of free electron on the normal hydrogen scale (NHE), ~4.5 eV [62]. E_g is the band gap energy of the semiconductors, 2.15 eV for CoFe₂O₄ and 3.07 eV for CeO₂. Thus, the calculated E_VB and E_CB relative to the NHE of CoFe₂O₄ were 2.39 eV and 0.24 eV, while those of CeO₂ were 2.60 eV and −0.48 eV, respectively. According to an alignment of E_VB and E_CB of CoFe₂O₄ and CeO₂, the band edges of CB and VB of CeO₂ are higher and lower, respectively, than those of CoFe₂O₄, so CeO₂/CoFe₂O₄ system is classified into type I heterojunction photocatalyst [1,9]. The photodegradation of MB over the CeO₂/CoFe₂O₄ photocatalyst started when the surface of CoFe₂O₄ and CeO₂ semiconductors were irradiated with UV light and absorbed energy equal to or higher than their E_g to produce photogenerated electron (e⁻)- holes (h⁺) pairs from the electrons excited to the CB and holes left in the VB. In the proposed mechanisms, h⁺ from the CeO₂ VB is transferred to the CoFe₂O₄ VB through the interface, as the CeO₂ E_VB (2.60 eV) is more positive than the CoFe₂O₄ E_VB (2.39 eV). Meanwhile, the photoexcited e⁻ in the CeO₂ CB (−0.48 eV) migrated to the more positive CoFe₂O₄ CB (0.24 eV). Then, these photogenerated charges reacted with the absorbed species on the semiconductor surface, e.g. O₂, H₂O, and hydroxide ion (OH⁻), to generate reactive oxygen species (ROS) for MB degradation. When considering the standard oxidation-reduction potentials, the CoFe₂O₄ E_CB (0.24 eV) is not negative enough to reduce O₂ to superoxide radicals (^∙O₂⁻) via e⁻ acceptation (O₂/^∙O₂⁻ = −0.046 eV vs. NHE [24]), while the CoFe₂O₄ E_VB (2.40 eV) is more positive to oxidize OH⁻ into hydroxyl radical (^∙OH) via h⁺ acceptation (OH⁻/^∙OH = 1.99 eV vs. NHE [24]). This suggested that in the CeO₂/CoFe₂O₄ photocatalyst system, the photogenerated h⁺ and ^∙OH were the main active species responsible for reacting with MB dye. As the h⁺ and ^∙OH species were required for ring-opening and complete degradation of MB molecules into less harmful products of CO₂, H₂O, or inorganic ions [63], the formation of heterojunction interfaces between CeO₂/CoFe₂O₄ could improve the photocatalytic activity by facilitating the interfacial charge separation and increasing the availability of reactive species of h⁺ and ^∙OH. It is interesting to note that some of the photogenerated e⁻ in the CeO₂ CB might also react with the adsorbed O₂ to form reactive ^∙O₂⁻ as the CeO₂ E_CB (−0.31 eV) is more negative than the single electron reduction potential of O₂ (-0.046 eV vs. NHE [24]). The quenching experiment by using radical scavengers may require confirming the main active species during the photocatalytic reaction [64,65].

The magnetic properties of CoFe₂O₄ and 20%CeO₂/CoFe₂O₄ photocatalysts at room temperature were studied by VSM technique. The magnetic hysteresis loops (M-H loops) of CoFe₂O₄ and 20%CeO₂/CoFe₂O₄ (Figure 10) showed a superparamagnetic behavior by having nearly no hysteresis loops and low coercivity (H_c) and remanent magnetization (M_r) [7,14]. CoFe₂O₄ had H_c = 106.4 Oe, M_r = 4.7 emu.g⁻¹, and saturated magnetization (M_s) = 44.1 emu.g⁻¹. In the case of 20%CeO₂/CoFe₂O₄, the Ms value decreased to 33.4 emu.g⁻¹ while H_c and M_r increased to 209.1 Oe and 5.3 emu.g⁻¹ due to the presence of the non-magnetic phase of CeO₂ [44]. The inset in Figure 10 shows that the suspended magnetic 20%CeO₂/CoFe₂O₄ photocatalysts could easily be separated from the treated solution by applying an external magnetic field. The 20%CeO₂/CoFe₂O₄ photocatalysts still preserved a good magnetic response and facilitated the separation and recovery of photocatalysts from a dye solution after the photodegradation reaction, which benefits the practical use of photocatalysis technology in wastewater treatment.

Figure 10. The M-H loops of CoFe₂O₄ and 20%CeO₂/CoFe₂O₄ under a magnetic field. Inset: a photograph showing the magnetic 20%CeO₂/CoFe₂O₄ photocatalysts separated by an external magnetic field.

The reusability of the magnetic 20%CeO₂/CoFe₂O₄ heterojunction photocatalyst for MB dye degradation was examined by magnetically separating the solid photocatalysts from the solution, washing, drying, and reusing in the same experimental procedures by using 14 mg photocatalyst dosage in 10 mg.L⁻¹ MB concentration. The photodegradation efficiency of the reused 20%CeO₂/CoFe₂O₄ in the second cycle was 25.2%, which was 65.9% lower than the efficiency of the fresh photocatalyst (73.8%). The reduction in activity through the recycling test probably resulted from strong adsorption of residual MB on the photocatalyst surface that decreased the surface-active sites. Figure 11 shows FT-IR spectra of the fresh and reused magnetic 20%CeO₂/CoFe₂O₄ photocatalyst and MB. The reused 20%CeO₂/CoFe₂O₄ showed the characteristic absorption bands of CoFe₂O₄ and CeO₂ at 595 cm⁻¹ for spinel-type oxide Fe-O and 409 cm⁻¹ for Co-O and Ce-O [19,66], together with the absorption bands due to the vibrational modes of MB (C₁₆H₁₈N₃SCl·3 H₂O) at 3390 cm⁻¹ for O-H stretching, 3100, 3025, and 2804 cm⁻¹ for C-H stretching of the aromatic group, 2770 cm⁻¹ for C-H stretching of an alkyl group, 1590–1330 cm⁻¹ for C-C stretching in aromatic rings, 1140 cm⁻¹ for C-N stretching and 1021 cm⁻¹ for C-S-C stretching [67,68]. These results confirmed that adsorbed MB on the photocatalyst surface was responsible for reducing photocatalytic activity in the reusability test. Therefore, removing the adsorbed MB molecules by treating the photocatalysts with a mind acid solution of 0.0001 M hydrochloric acid (HCl) for 20 min prior to reuse is required to regenerate the photocatalysts and maintain the performance [65].

Figure 11. FT-IR spectra of fresh and reused 20%CeO₂/CoFe₂O₄ photocatalysts and MB dye.

4. Conclusion

This study showed the improvement of photocatalytic activity CoFe₂O₄ by coupling with CeO₂. The magnetically separable CeO₂/CoFe₂O₄ nanocomposites were prepared by a precipitation method with different CeO₂ loadings at 10, 20, and 30 wt%. The CeO₂/CoFe₂O₄ nanocomposites had a higher MB dye degradation under UV irradiation than pure CoFe₂O₄ and CeO₂. 20%CeO₂/CoFe₂O₄ photocatalyst showed the highest photocatalytic efficiency and rate constant due to the uniform dispersion of the CeO₂ minor phase throughout the CoFe₂O₄ major phase. The formation of CeO₂/CoFe₂O₄ heterojunctions improved the photocatalytic activity by facilitating the effective charge separation and increasing the availability of reactive species of h⁺ and ^∙OH for MB degradation. The magnetically separable CeO₂/CoFe₂O₄ photocatalysts are suitable for practical application in wastewater treatment as they can easily be recovered after the photodegradation reaction by applying an external magnetic field.

Acknowledgments

This work was supported by the [School of Science, King Mongkut’s Institute of Technology Ladkrabang (KMITL)] under Grant [number 2565-02-05-019]. We thank the Scientific Instruments Center, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work was supported by the School of Science, King Mongkut’s Institute of Technology Ladkrabang (KMITL) [2565-02-05-019].