Characterization of Co 2+ - and Fe 3+ -Codoped TiO 2 Nanomaterials for Photocatalytic Degradation of Organic Pollutants under Visible Light Irradiation

Abstract

In this study, TiO₂ nanomaterials were prepared using a solvothermal method and codoped with Co²⁺ and Fe³⁺ ions for the photocatalytic degradation of organic pollutants under visible light. The physicochemical properties of the obtained materials were studied by powder X-ray diffraction, field emission electron scanning microscopy, energy-dispersive X-ray spectroscopy, and nitrogen adsorption isotherms. Optical absorption was characterized by UV-vis absorption spectroscopy. The photocatalytic activities of the prepared materials were evaluated through methylene blue (MB) degradation under visible light irradiation. Results showed the excellent performance of MB degradation investigated on TiO₂ samples codoped with Co²⁺ and Fe³⁺ in comparison with undoped and Co²⁺-doped TiO₂ samples. The codoped TiO₂ samples degraded 85%–90% of MB after 120 min, whereas all the prepared TiO₂ samples were composed of pure anatase phase and had a spherical-like shape and mean crystalline size ranging from 6.2 nm to 7.8 determined by Scherrer’s equation. The optical absorption of the TiO₂ codoped with Co²⁺ and Fe³⁺ was significantly enhanced toward the visible light region. The pseudo-second-order kinetic model fits well for the degradation of MB on as-prepared TiO₂ codoped with Co²⁺ and Fe³⁺.

1. Introduction

The development of metal oxide nanoparticles has been extensively attracted as adsorbents and photocatalysts for the treatment of dye-containing wastewater [1–3]. Titanium dioxide (TiO₂) nanoparticles have considerable attention because of their long-term stability, low cost, and nontoxicity [4]. However, the major drawback of TiO₂ is a large bandgap energy (~3.2 eV and 3.0 eV for anatase and rutile phases, respectively), and it requires ultraviolet (UV) irradiation (wavelength, $\lambda <387\text{nm}$ ) for activation. Many attempts have been conducted to extend the photocatalytic activity of TiO₂ from a UV to visible light region (wavelength, $\lambda$ : 400-700 nm). Doping of TiO₂ with transition metals or nonmetals is the most promising approach used to activate it due to the bandgap energy reduction [5]. Transition metals can provide additional energy levels within the bandgap of a TiO₂ photocatalyst. Electron transfer from one of these levels to the conduction band requires lower photon energy than those of undoped TiO₂ [6]. Various transition metal ions have been studied to dope with TiO₂, including Co [7], Fe [8], Cr [9], Ni [10], and V [11], etc. Furthermore, many researchers have also focused on codoping of TiO₂ with two or more metal ions for enhancement of photocatalytic degradation of organic pollutants under solar light irradiation such as Mn and Co [12]; Ni and Cr [13]; Fe, Co, and S [14]; and Fe and Pr [15]. It indicated that codoped TiO₂ could expand its light absorption range toward visible light and thus enhance its photocatalytic efficiency, which is higher than that of single-doped TiO₂. It was reported that Mn²⁺- and Co²⁺-codoped TiO₂ could degrade about 97% of enoxacin, which was higher than that of Co²⁺ single-doped TiO₂ (degradation percentage ~48%) after 80 min of exposure to solar light [12]. A recent work also demonstrated that Ni/Cr-codoped TiO₂ exhibited a higher efficiency of 95.6% for MB degradation after 90 min under sunlight, compared to Ni-doped TiO₂ (Ni-doped TiO₂ degraded ~28%-59% of MB depending on the amount of Ni doping) [13]. The photocatalytic degradation of Acid Orange 7 azo dye under visible light using Fe³⁺- and Pr³⁺-codoped TiO₂ was much more enhanced compared to undoped TiO₂ and Fe³⁺single-doped TiO₂ [15]. Those studies indicated that the photocatalytic activity of doped TiO₂ material is greatly dependent on the nature of the dopant ions and their doping concentration and preparation, in which optimal conditions for the synthesis decide the ability of nanomaterials applied for the photodegradation of organic pollutants in wastewater. Fe³⁺ has been considered the best dopant among the transition metals used for increasing the photocatalytic activity of TiO₂ thanks to its similar radius to Ti⁴⁺ (0.63 Å and 0.60 Å, respectively). TiO₂ doping with Co²⁺ showed excellent degradation activity of dye molecules; however, the optimal concentration of Co²⁺ is needed to enhance activity, because the high percentage of Co²⁺ leads to a reduction in the photocatalytic activity [14].

In this study, we aimed to synthesize and characterize TiO₂ nanoparticles codoped with Co²⁺ and Fe³⁺ ions using the solvothermal method. The effect of the doping concentration of Co²⁺ was investigated to determine the optimal condition for preparing a photocatalyst with high activity in the visible light region. The obtained nanomaterials were investigated by various physicochemical methods, including XRD, FE-SEM, EDXS, nitrogen adsorption-desorption isotherms, and UV-vis absorption. Photocatalytic activities of prepared nanomaterials were evaluated for degradation of methylene blue (MB). Our results indicated that the prepared TiO₂ materials codoped with Co²⁺ and Fe³⁺ showed a highly photocatalytic efficiency compared to single-doped TiO₂. This nanomaterial could be an effective alternative for the treatment of wastewater in the textile industry.

2. Materials and Methods

2.1. Chemicals and Preparation

All chemicals were of analytical grade and used as received without further purification. Tetraisopropyl orthotitanate ( $\text{Ti}{\left(i‐\text{O}{\text{C}}_3{\text{H}}_7\right)}_4$ ), acetylacetone (C₅H₈O₂), ethanol (C₂H₅OH), and methylene blue C₁₆H₁₈ClN₃S (MB) were obtained from Merck. Cobalt (II) acetate tetrahydrate ((CH₃COO)₂Co·4H₂O) and iron (III) nitrate nonahydrate (Fe(NO₃)₃·9H₂O) were purchased from China. Double-distilled water was used for preparing all solutions.

Single doping of TiO₂ with Co²⁺ and codoping with Co²⁺ and Fe³⁺ were conducted by the solvothermal synthesis. The obtained samples were referred to as $\text{Ti}{\text{O}}_2‐x\%\text{Co}‐y\%\text{Fe}$ , with $x$ and $y$ showing the molar Co/Ti and Fe/Ti ratio in the starting materials, respectively. In a typical experiment, a mixture consisting of 14.5 mL of C₂H₅OH and 0.5 mL of C₅H₈O₂ was stirred for 15 min. The mixture was slowly added with 2.96 mL of $\text{Ti}{\left(i‐\text{O}{\text{C}}_3{\text{H}}_7\right)}_4$ as a precursor and stirred for another 30 min (mixture 1). Desired amounts of (CH₃COO)₂Co·4H₂O and Fe(NO₃)₃·9H₂O were dissolved in another mixture consisting of 5 mL of C₂H₅OH, 1 mL of C₅H₈O₂, and 0.18 mL of H₂O for 30 min. The mixed solution was then dropped into mixture 1 with continuous stirring for 30 min and then transferred to a 100 mL Teflon-lined stainless-steel autoclave maintained at 180°C for 12 h. The product was washed several times with double-distilled water until pH 7, dried at 100°C for 24 h, and ground on agate mortar. Single-doped and codoped TiO₂ samples were designated as TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe. TiO₂ nanoparticles were also prepared as a control and denoted as undoped TiO₂.

2.2. Material Characterization

X-ray diffraction (XRD) patterns of the prepared TiO₂ samples were recorded with a D8 ADVANCE Bruker Diffractometer within $2\theta$ range of 20°–80° with a scan step of 0.03°·s^-1 under Cu-Kα radiation ( $\lambda =\mathrm{0.1.54056}\text{nm}$ ). The surface morphology of the TiO₂ samples and representative doped TiO₂ samples was observed by field emission scanning electron microscopy (FE-SEM, Hitachi S-4800, Japan). The average nanoparticle sizes were calculated from the scanning electron microscopic (SEM) images by using ImageJ software. The elemental composition of the samples was analyzed through energy-dispersive X-ray spectroscopy (EDXS, Nova NanoSEM 450, FEI). The nitrogen adsorption-desorption isotherms of the representative TiO₂ samples were measured at -196°C with a Micromeritics ASAP 2020 apparatus. The total surface areas of the samples were determined by the Brunauer–Emmett–Teller (BET) method, and the pore size distribution was determined through the Barrett–Joyner–Halenda (BJH) method using desorption curves. UV-vis absorption spectra were recorded on the Jasco V-670 spectrometer.

The photocatalytic activity of the prepared TiO₂ samples was evaluated by degradation of MB under visible light irradiation. A 250 W Osram mercury lamp equipped with a UV cut-off filter was used as a visible light source. In a typical experiment, 40 mg of the prepared TiO₂ sample was added into a 100 mL quartz photoreactor containing 50 mL of 14 mg·L^-1 MB solution. The reaction mixture was then stirred in the dark at a constant speed of 150 rpm for 30 min to reach the adsorption-desorption equilibrium. The resulting mixture was irradiated under the visible light source for up to 120 min. After predefined intervals (30, 60, 90, and 120 min), the samples were removed from the photoreactor and centrifuged. The residual concentration in the supernatant was measured with a UV-vis spectrometer (Agilent 8453, USA) at a wavelength of 665 nm. MB concentration was determined using a linear regression equation obtained by plotting a calibration curve of MB within a range of known concentrations. The photocatalytic ability of the TiO₂ samples was evaluated through the percentage of MB degradation as follows: $\begin{array}{}\text{(1)}& \text{Degradation}\text{percent}=\frac{C_{\text{o}}-C}{C_{\text{o}}}·100\%,\end{array}$ where $C_{\text{o}}$ is the initial concentration of MB and $C$ is the remaining concentration of MB at a specific time of measurement.

3. Results and Discussion

3.1. Characterization of the Synthesized Materials

The XRD patterns of the synthesized samples (undoped, single-doped, and codoped TiO₂) are shown in Figure 1. The results demonstrated the formation of monocrystalline TiO₂ in the anatase phase in all samples. The XRD pattern of undoped TiO₂ samples (Figure 1(a)) shows typically crystalline peaks at $2\theta ={25.1}^{°}$ , 37.67°, 47.71°, and 54.11° corresponding to crystal planes (101), (111), (200), and (211), respectively, which can be indexed to the anatase phase of TiO₂ (JCPDS card no. 21-1272). For single-doped and codoped TiO₂ samples, those diffraction peaks are observed at $2\theta ={25.22}^{°}$ , 25.25°, 25.30°, 25.34°, and 25.37°; 37.85°, 37.90°, 38°, 38.09°, and 38.12°; 47.82°, 47.88°, 47.92°, 47.99°, and 48.11°; and 54.35°, 54.41°, 54.44°, 54.46°, and 54.51° (see Figures 1(b)–1(f)) for TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe, respectively. The data indicated that the typical peaks of the anatase TiO₂ in single-doped and codoped TiO₂ samples were slightly shifted to higher values than those of undoped TiO₂ samples. The codoped TiO₂ samples (Figures 1(e) and 1(f)) exhibited a decrease in peak intensity and slight broadening of the peaks compared with those of undoped TiO₂ and single-doped TiO₂ samples (Figures 1(a)–1(d)). Moreover, no characteristic peaks attributed to cobalt and iron compounds were detected in the XRD patterns, which can be attributed to the very low dopant amount in the samples. The observations suggested that Co²⁺ and Fe³⁺ cations were successfully introduced into the lattice of TiO₂. The average particle sizes of the samples were calculated from the full width at half maximum (FWHM) of the (101) diffraction peak according to the Debye–Scherrer equation [16]: $\begin{array}{}\text{(2)}& D=\frac{0.9\lambda }{\beta \cos \theta }\end{array}$ where $\lambda$ is the X-ray wavelength (0.15406 nm), $\beta$ is the line width at half maxima of peaks, and $\theta$ represents the Bragg angle of X-ray diffraction. The obtained results (Table 1) showed that the average particle sizes of all samples were at a nanosized scale ranging from 6.2 nm to 7.8 nm. The single-doped and codoped TiO₂ samples had a smaller size than the undoped TiO₂ samples. The undoped TiO₂ samples had the largest mean particle size of 7.8 nm. Among the three Co-doped TiO₂ samples, TiO₂-1%Co samples had the smallest particle size with a mean particle size of 6.8 nm. The average particle size of Co-doped TiO₂ samples increased from 6.8 nm to 7.5 nm with an increasing Co/Ti molar ratio from 1% to 5% in the samples. The high doping amount of Co resulted in an increase in the particle size, which was attributed to the larger radius of Co²⁺ (0.74 Å) than that of Ti⁴⁺ (0.60 Å). Therefore, the Co/Ti molar ratio of 1% was considered to be optimal. This ratio was further used to prepare TiO₂ codoped with Co and Fe with varying Co/Ti molar ratios from 1% to 2.5% and a fixed Fe/Ti molar ratio of 2.5%. Previous studies [14, 15] reported that the Fe-doped TiO₂ can extend more excellent absorption in the longer wavelength range than TiO₂. As a result, we chose iron as a codopant element with cobalt for extending the light absorption of TiO₂ nanoparticles toward the visible region. The Fe/Ti molar ratio of 2.5% was used in the synthesis of codoped TiO₂ samples, because the higher Fe content may lead to the formation of iron oxide during thermal treatment at 180°C, decrease the synthesis yield, and/or distort the lattice of TiO₂. As shown in Table 1, TiO₂-1%Co-2.5%Fe samples had smaller sizes (their mean particle sizes are 6.2 nm) than TiO₂-2.5%Co-2.5%Fe samples (their mean particle sizes are 7.6 nm). These results demonstrated that TiO₂ codoped with Co and Fe inhibited the crystal growth of the particles and effectively decreased the particle size; the high doping amount of Co also resulted in an increase in the particle size of the codoped TiO₂ samples.

Figure 1

XRD patterns of (a) undoped TiO₂, (b) TiO₂-1%Co, (c) TiO₂-2.5%Co, (d) TiO₂-5%Co, and co-doped TiO₂: (e) TiO₂-1%Co-2.5%Fe and (f) TiO₂-2.5%Co-2.5%Fe.

Table 1

Mean crystalline size and phase of the prepared TiO₂ samples.

Samples	Undoped TiO2	TiO2-1%Co	TiO2-2.5%Co	TiO2-5%Co	TiO2-1%Co-2.5%Fe	TiO2-2.5%Co-2.5%Fe
Crystalline phase	Anatase	Anatase	Anatase	Anatase	Anatase	Anatase
Mean crystalline size	7.8 nm	6.8 nm	7.4 nm	7.5 nm	6.2 nm	7.6 nm

The morphology of the synthesized TiO₂ samples was further observed through FE-SEM imaging. Figure 2 presents the FE-SEM images of the representative samples: undoped TiO₂ (Figures 2(a) and 2(b)), TiO₂-1%Co (Figures 2(c) and 2(d)), TiO₂-5%Co (Figures 2(e) and 2(f)), and TiO₂-1%Co-2.5%Fe (Figures 2(g) and 2(h)) at magnifications of 100 k and 200 k. The results indicated that the synthesized TiO₂ samples consisted of numerous crystalline particles that agglomerated to form tiny clusters. The FE-SEM images demonstrated that the particles were spherical and had uniform size distribution. The calculations from the FE-SEM images confirmed that the average sizes were 26.5, 23.8, 27, and 20.8 nm for undoped TiO₂, TiO₂-1%Co, TiO₂-5%Co, and TiO₂-1%Co-2.5%Fe samples, respectively. These results showed deviations from the data obtained from XRD due to the agglomeration of the nanoparticles in the FE-SEM images. However, the results were in agreement with the XRD results, confirming that Co and Fe doping can suppress TiO₂ particle growth and the TiO₂-1%Co-2.5%Fe sample had the smallest particle size with a narrow distribution of the particle size (inset in Figure 2).

Figure 2

FE-SEM images of undoped TiO₂ (a, b), TiO₂-1%Co (c, d), TiO₂-5%Co (e, f), and TiO₂-1%Co-2.5%Fe (g, h) samples with magnifications of 100 k and 200 k. Inset shows a particle size distribution of TiO₂-1%Co-2.5%Fe samples.

The presence of doping metals on the TiO₂ samples was determined through EDXS analyses. The EDXS spectra of the representative TiO₂ samples (undoped TiO₂, TiO₂-1%Co, TiO₂-5%Co, and TiO₂-1%Co-2.5%Fe) are presented in Figure 3. As shown in Figures 3(a), O and Ti were found to be the major components of undoped TiO₂ with 65.35 at% and 33.65 at%, respectively. Co was detected as the doping component of TiO₂-1%Co and TiO₂-5%Co samples because tiny peaks were observed in the EDXS spectra (Figures 3(b) and 3(c)), which were attributed to the presence of Co in the samples (0.21 at% and 0.89 at%, respectively). The contents of Co and Fe in TiO₂-1%Co-2.5%Fe samples were 0.28 at% and 0.74 at%, respectively (Figure 3(d)). The results confirmed that Co and Fe were successfully doped in the TiO₂ samples.

Figure 3

EDXS spectra of (a) undoped TiO₂, (b) TiO₂-1%Co, (c) TiO₂-5%Co, and (d) TiO₂-1%Co-2.5%Fe samples.

(b) (c) (d)

Nitrogen adsorption-desorption isotherms were determined for the representative samples to analyze and quantify the pore structure and surface area of doped and codoped TiO₂ samples. Figure 4 presents the nitrogen adsorption-desorption isotherms of three selected samples: TiO₂-1%Co (Figure 4(a)), TiO₂-5%Co (Figure 4(b)), and TiO₂-1%Co-2.5%Fe (Figure 4(c)). Results in Figure 4 indicate that the three samples had a similar hysteresis loop, which could be associated with capillary condensation in mesopores. The hysteresis loops of these samples, which were observed in the $P/P_{\text{o}}$ range of 0.42–0.82, were characterized by a type IV isotherm with an H₂ hysteresis loop [17]. The results suggested that the TiO₂ samples contained ink-bottle pores, leading to pore blocking/percolation effects during desorption in mesopore networks. Pore size distribution was calculated by the BJH method using data on the desorption curves (Figure 5). The surface areas of the synthesized TiO₂ samples were determined by using the BET method. The surface characteristics of these samples are summarized in Table 2 and compared with those of undoped TiO₂. TiO₂-1%Co and TiO₂-1%Co-2.5%Fe samples showed a narrow pore size distribution (Figures 5(a) and 5(c)) with average mesopore diameters of 3.97 and 4.1 nm, respectively (Table 2). TiO₂-5%Co samples exhibited a wider pore size distribution that scattered from 2.5 nm to 6.5 nm and mainly concentrated at 5.1 nm (Figure 5(b) and Table 2). Undoped TiO₂ showed a larger mean pore size of 6.7 nm than the doped and codoped TiO₂. As shown in Table 2, the prepared TiO₂ samples did not have significant discrepancies in their BET surface area. Relatively high BET surface areas were determined as follows: 170 m²/g for TiO₂-1%Co, 164 m²/g for TiO₂-5%Co, and 174 m²/g for TiO₂-1%Co-2.5%Fe, which are slightly higher than that of undoped TiO₂ (160 m²/g). The large BET surface areas of the doped and codoped TiO₂ samples confirmed the better adsorption ability of the frameworks of TiO₂. This finding may be due to the linkage between the dopant ions (Co²⁺ and Fe³⁺) and titanium by an oxygen bridge, which effectively increased their specific surface areas.

Figure 4

Nitrogen adsorption-desorption isotherms of representative samples: (a) TiO₂-1%Co, (b) TiO₂-5%Co, and (c) TiO₂-1%Co-2.5%Fe samples.

(b) (c)

Figure 5

Pore size distributions of representative samples: (a) TiO₂-1%Co, (b) TiO₂-5%Co, and (c) TiO₂-1%Co-2.5%Fe samples.

(b) (c)

Table 2

Surface characteristics of the representative TiO₂ samples.

Samples	$S_{\text{BET}}$ a (m2/g)	$V_{\text{BJH}}$ b (cm3/g)	$D_{\text{p}}$ c (nm)
TiO2-1%Co	170	0.18	3.97
TiO2-5%Co	164	0.23	5.1
TiO2-1%Co-2.5%Fe	174	0.16	4.1
Undoped TiO2	160	0.29	6.7

^aBET surface area. ^bTotal pore volume determined using desorption curves of the isotherms. ^cPeak pore sizes from the pore size distributions.

Table 3 compares the BET surface areas of doped and codoped TiO₂ materials of the present work with those of other doped and codoped TiO₂ materials reported previously. The BET surface areas of doped and codoped TiO₂ materials of our work were within 164–174 m²/g, which were higher than those of Fe-doped TiO₂ materials [18, 19], Co-doped TiO₂ materials [20], Cr- and Fe-doped TiO₂ materials [18], and Fe- and La-doped TiO₂ materials [19] (Table 3). The larger BET surface areas of the doped and codoped TiO₂ materials of the present work were possibly due to their smaller particle sizes ranging from 6.2 nm to 7.5 nm (Table 3), depending on the doping amount. The doped TiO₂ materials of our work were prepared under more favorable conditions (solvothermal treatment at 180°C without calcination). The other TiO₂ materials were prepared by hydrothermal treatment or sol-gel method, followed by calcination, which resulted in the formation of larger particle sizes and significant loss of the BET surface area (Table 3).

Table 3

Comparison of the BET surface areas of the TiO₂ samples prepared in this work with those of previous works.

Materials	Synthetic conditions	$S_{\text{BET}}$ (m2/g)	Particle size (nm)	References
TiO2-1%Co	Solvothermal treatment at 180°C without calcination	170	6.8	This work
TiO2-5%Co	Solvothermal treatment at 180°C without calcination	164	7.5	This work
TiO2-1%Co-2.5%Fe	Solvothermal treatment at 180°C without calcination	174	6.2	This work
TiO2-Fe	Hydrothermal treatment at 180°C, calcined at 400°C	89	18	[18]
TiO2-0.2%Fe	Sol-gel, calcined at 500°C	78	12	[19]
TiO2-0.2%Fe-2%La	Sol-gel, calcined at 500°C	84	10	[19]
TiO2-0.085%Co	Sol-gel, with following heat treatment from 200°C to 900°C for 30 min	39.7	25	[20]
Fe-Cr-codoped TiO2	Hydrothermal treatment at 180°C, calcined 500°C	93		[18]

The UV-vis absorption spectra of undoped TiO₂, doped TiO₂, and codoped TiO₂ samples were recorded to study their optical properties (Figure 6). As shown in Figure 6 (curve a), the undoped TiO₂ sample was characterized by a narrow absorption spectrum, ranging from 250 nm to 370 nm in the UV region. No visible light absorption was observed for the undoped TiO₂ sample. Co-doped TiO₂ (Figure 6, curves b–d) and Co- and Fe-doped TiO₂ samples (Figure 6, curves e and f) had broader absorption spectra, which shifted toward longer wavelengths in the visible light region. The absorption spectra of Co-doped TiO₂ samples had maximum peaks at about 330–340 nm, but the shoulder of the peaks shifted to the visible range (400-700 nm). TiO₂ samples codoped with Co and Fe were characterized by a broader maximum peak at 400–500nm and stronger intensity of the absorption shoulder in the visible region compared with those of single-doped TiO₂ samples. Figure 6 shows that the absorption edges of TiO₂ codoped with Co and Fe moved remarkably, with a redshift to the visible region in comparison with undoped TiO₂. These optical absorption characteristics of TiO₂ codoped with Co and Fe were possibly due to absorption induced by dopants. The absorption edges can be extrapolated by making a tangent line of the UV-vis absorption spectra [21]. Therefore, the bandgap energies of the prepared TiO₂ samples were estimated from the UV-vis spectra via the following equation: $\begin{array}{}\text{(3)}& E_g=\frac{1240}{\lambda }\end{array}$

Figure 6

UV-vis spectra of undoped TiO₂, TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe samples.

The bandgap energy levels of Co-doped TiO₂ samples were estimated to be 2.03 eV, 2.00 eV, and 1.99 eV (for TiO₂-1%Co, TiO₂-2.5%Co, and TiO₂-5%Co, respectively), which were higher than those of Co- and Fe-codoped TiO₂ samples (their bandgap energies are 1.65 eV and 1.59 eV for TiO₂-1%Co-2.5%Fe and TiO₂-2.5%Co-2.5%Fe, respectively). The calculated bandgap energies of single-doped and codoped TiO₂ samples were lower than that of undoped TiO₂ samples (~3.25 eV). The results revealed that the dopant elements were successfully incorporated into the TiO₂ crystal lattice and extended the optical absorption toward the visible light region. TiO₂ samples codoped with Co and Fe had stronger optical absorption of the visible light than single-doped and undoped TiO₂ samples. This means that the TiO₂ codoped with Co and Fe can absorb visible light in a much wider range of wavelengths than undoped TiO₂ and single-doped TiO₂ samples, which is beneficial for increasing the photocatalytic efficiency under visible light.

3.2. Photocatalytic Activities of the Prepared TiO₂ Materials for Degradation of MB

Before the photocatalytic reaction, the adsorption for MB removal on the synthesized TiO₂ materials was conducted under dark conditions. Results obtained showed that the percentages of MB removal were only 5%-6% after 60 min for all the synthesized materials, but desorption of MB was observed by prolonging adsorption time for those materials. Therefore, the photocatalytic activities of the prepared TiO₂ samples were evaluated by degrading MB solution under visible light irradiation. The reaction mixture was first stirred in the dark for 30 min to reach the adsorption-desorption equilibrium and to ensure that the degradation of MB obtained is due to the photocatalytic reaction with the presence of the synthesized TiO₂ materials, but not due to the adsorption. Figure 7 shows the percentage of MB degradation over all of the TiO₂ samples (undoped TiO₂, TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe samples) as a function of irradiation time. For comparison, control experiments were performed under the same conditions but without the presence of a photocatalyst (Figure 7, curve MB). As shown in Figure 7 (curve MB), almost no degradation of MB was observed without the photocatalyst (only 7.95% of MB was degraded after 120 min of exposure to visible light irradiation), indicating that MB was unable to self-degrade. Undoped TiO₂ and Co-doped TiO₂ with low doping concentration (TiO₂-1%Co and TiO₂-2.5%Co) showed pure photocatalytic activity. The degradation efficiencies of MB on undoped TiO₂ samples were comparable with those of Co-doped TiO₂ samples with low Co doping concentration for all the tested time points (Figure 7, curves undoped TiO₂, TiO₂-1%Co, and TiO₂-2.5%Co). The degradation percentages of MB were 10.95%, 17.85%, and 17.5% after 120 min of irradiation for undoped TiO₂, TiO₂-1%Co, and TiO₂-2.5%Co, respectively. However, the degradation of MB is enhanced with higher Co doping concentration. The degradation percentages of MB increased three times on TiO₂-5%Co samples and were higher than those on TiO₂-1%Co and TiO₂-2.5%Co samples for all the tested time points. About 39.8% of MB was degraded within 30 min on TiO₂-5%Co and then gradually increased to ~50% by increasing irradiation time up to 120 min. Moreover, the significantly enhanced degradation of MB was observed on codoped TiO₂ samples (Figure 7, curves TiO₂-1%Co-2.5%Fe and TiO₂-2.5%Co-2.5%Fe). The highest degradation of MB was obtained by TiO₂ codoped with 1%Co and 2.5%Fe for all the time points tested. The degradation of MB was almost complete on codoped TiO₂ samples for the tested time; 90% and 85% of MB were degraded on TiO₂-1%Co-2.5%Fe and TiO₂-2.5%Co-2.5%Fe after 120 min, respectively.

Figure 7

The degradation of MB using different catalysts: undoped TiO₂, TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co,TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe samples under visible light irradiation.

The results showed that TiO₂ samples codoped with Co and Fe exhibited higher photocatalytic degradation of MB under visible light than the undoped and single-doped TiO₂ samples, which could be attributed to their lower bandgap values compared to those of the undoped and single-doped TiO₂ samples. These observations indicated that codopant ions had a favorable effect on the photocatalytic performance of the prepared TiO₂ materials. These ions can provide additional energy levels within the bandgap of TiO₂. The bandgap of TiO₂ consists of a contribution from the 2p orbitals of O for the valence band and the 3d orbitals of Ti toward the conduction band, which have a large difference in energy, leading to the activation of TiO₂ in the UV light region with extremely high energy. The band structures of codoped TiO₂ materials are mainly affected by the 3d energy states of the transitional metal ions (Co²⁺ and Fe³⁺). In fact, the UV-vis absorption studies (Figure 6) revealed that the respective absorption bands of codoped TiO₂ samples effectively shifted toward wavelengths longer than 400 nm. Thus, the partially filled Co/Fe 3d bands were located below the conduction band of TiO₂. Hence, when photons with wavelengths longer than 400 nm are used for illumination, the electrons in the Co 3d and Fe 3d bands, instead of electrons in the valence band of TiO₂, are excited to the conduction band, while Co²⁺ and Fe³⁺ loses one electron and becomes Co³⁺ and Fe⁴⁺ (see Scheme 1). This phenomenon can induce more photogenerated electrons and holes to participate in the photocatalytic reaction.

Scheme 1

Band structure of TiO₂ and TiO₂ codoped with Co²⁺ and Fe³⁺. VB: valence band; CB: conduction band.

As calculated above, the bandgap energy values of codoped TiO₂ samples (1.59 eV-1.65 eV) were lower than those of single-doped TiO₂ samples (1.99-2.03 eV) and undoped TiO₂ (3.25 eV), indicating the lower energy of photons necessary to generate electron transition and holes ( $h^{+}$ ) for codoped TiO₂ samples compared to single-doped and undoped TiO₂. These holes can react with water to produce the highly reactive ^·OH, and both holes and ^·OH have a strong oxidation potential for the degradation of MB.

The plausible reaction mechanism of the photodegradation is given below: $\begin{array}{}\text{(4)}& {\text{TiO}}_2+hv⟶e^{-}+h^{+}\text{(5)}& {\text{Co}}^{2+}+hv⟶{\text{Co}}^{3+}+e^{-}\text{(6)}& {\text{Fe}}^{3+}+hv⟶{\text{Fe}}^{4+}+e^{-}\text{(7)}& h^{+}+{\text{H}}_2{\text{O}}_{\text{ads}}⟶·\text{OH}+{\text{H}}^{+}\text{(8)}& e^{-}+{\text{O}}_{2\text{ads}}⟶{\text{O}}_2{·}^{-}\text{(9)}& \text{MB}+·\text{OH}/{\text{O}}_2{·}^{-}⟶\text{degradation}\text{products}\end{array}$

It is suggested that the photodegradation of MB can be divided into three main steps: (1) the initial step is the formation of electrons ( $e^{-}$ ) in the CB and holes ( $h^{+}$ ) in the VB upon a light incident on the photocatalyst surface (see equations (4), (5), (6)); (2) the intermediate step includes the formation of ^·OH and H⁺ through the interaction between the holes and H₂O adsorbed on the photocatalyst surface (equation (7)) and of O₂ ^·- due to the reduction of the adsorbed O₂ molecule (equation (8)) that has strong oxidation power, which in turn promotes the decomposition of MB; and (3) the final step is the degradation of MB by ^·OH/O₂ ^·- radicals.

Thus, this work proposed a study to understand the role of Co and Fe in codoped TiO₂ nanoparticles in yielding better performance as a visible light-driven photocatalyst.

The kinetics of photodegradation of MB on as-prepared TiO₂ materials was further investigated using the pseudo-first-order kinetic model [22] as follows: $\begin{array}{}\text{(10)}& \ln \frac{C_{\text{o}}}{C}=k_{1}t,\end{array}$ where $C_{\text{o}}$ is the initial concentration of MB, $C$ is the concentration of MB at time $t$ , and $k_1$ is the apparent reaction rate constant. According to equation (10), a linear regression plot of $\ln \left(C_{\text{o}}/C\right)$ vs. time $t$ gives the value of $k_1$ . The experimental data were analyzed using the pseudo-first-order equation; the resulting plots of $\ln \left(C_{\text{o}}/C\right)$ vs. time $t$ are shown in Figure 8(a). From the obtained plots in Figure 8(a), the rate constants were found to be 0.0009 min^-1, 0.0014 min^-1, 0.0013 min^-1, 0.005 min^-1, 0.0186 min^-1, and 0.0162 min^-1 with correlation coefficients ( $R^2$ ) of 0.96, 0.91, 0.90, 0.78, 0.85, and 0.98 for the undoped TiO₂, TiO₂-1%Co, TiO₂-2.5%Co, TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe, respectively. The experimental degradation data of all the TiO₂ samples were also analyzed using the pseudo-second-order kinetic equation proposed by Blanchard et al. [23]: $\begin{array}{}\text{(11)}& \frac{1}{q_t}=\frac{1}{t}\frac{1}{k_2{q}_e^2}+\frac{1}{q_e},\end{array}$

Figure 8

Plots of (a) first- and (b) second-order reaction rates for the degradation of MB over as-prepared TiO₂ materials.

(b)

where $q_t$ and $q_e$ are the amounts of dye adsorbed at time $t$ and at equilibrium (mg·g^-1) and $k_2$ is the pseudo-second-order rate constant (g·mg^-1·min^-1) for the adsorption process. The linear plots of $1/q_t$ vs. time $1/t$ shown in Figure 8(b) give the values of $k_2$ and $q_e$ . It could be seen that three TiO₂ samples: TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe showed the best fit in the linear plots of $1/q_t$ vs. $1/t$ with higher values of $R^2$ (0.83, 0.95, and 0.99, respectively) than those of the pseudo-first-order equation (0.78, 0.85, and 0.98). The values of $q_e$ obtained from the linear plots are 9.03 mg·g^-1, 18.28 mg·g^-1, and 21.78 mg·g^-1 for TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe, which were higher than those of the experimental values (8.75 mg·g^-1, 14.86 mg·g^-1, and 15.75 mg·g^-1 for TiO₂-5%Co, TiO₂-1%Co-2.5%Fe, and TiO₂-2.5%Co-2.5%Fe, respectively), whereas the undoped TiO₂, TiO₂-1%Co, and TiO₂-2.5%Co samples did not follow the pseudo-second-order equation because of smaller values of $R^2$ (0.82, 0.55, and 0.54) of the plots $1/q_t$ vs. $1/t$ (not shown), compared to those of the pseudo-first-order plots.

The observations revealed that the degradation of MB on the undoped TiO₂ and Co single-doped TiO₂ samples with 1% Co and 2.5% Co doping can be described by the pseudo-first-order equation. Meanwhile, the pseudo-second order fits better compared to the pseudo-first order for describing the degradation of MB on the TiO₂ doped with 5%Co and codoped with Co and Fe, considering that $R^2$ of the pseudo-second-order equation was greater than that of the first order for those samples. This clearly explains that the enhanced degradation efficiency of MB is attributed to the high surface area of the TiO₂ codoped with Co and Fe and their reduced bandgap energies.

Table 4 compares the degradation of MB on the TiO₂ materials as prepared in this work with those published in previous works [24–28]. It was found that the TiO₂ codoped with 1%Co and 2.5%Fe in this work showed the highest photocatalytic efficiency with 90% of MB degraded after visible light irradiation of 2 h. The metal oxide (Bi₂O₃) [24] and mixed metal oxides (e.g., CeO₂/V₂O₅ [24], CeO₂/CuO [25], and ZnO/CdO [26]) exhibited much lower degradation efficiency of MB compared with the TiO₂ materials of our work, which degraded about 64.2%-75% of MB for a longer irradiation time of 4 h. The complex metal oxide CaBi₆O₁₀ [24] and nanocomposite g-C₃N₄-CdS [27] demonstrated relatively high degradation percentages of MB (~97% and 90.45% for CaBi₆O₁₀ and g-C₃N₄-CdS, respectively), which could be comparable with that of the TiO₂ codoped with 1%Co and 2.5%Fe of our work, but these materials require a quite long irradiation time of 4 h (for CaBi₆O₁₀) and 3 h (g-C₃N₄-CdS) to obtain the complete degradation of MB under visible light. Moreover, a similar study reported that TiO₂ doped with 10% F [28] also revealed a high degradation percentage of MB (~90%), but after a much longer irradiation time of 4 h, compared to that of the TiO₂ codoped with Co and Fe of our work. The comparison revealed that the photocatalytic efficiency of the materials depends not only on their band structure but also on their surface characteristics. The excellent photocatalytic performance of the TiO₂ codoped with Co and Fe of our work could be attributed to their larger BET surface area compared to those of the other materials. As a result, the TiO₂ codoped with Co and Fe has more active sites and simultaneously can absorb much more visible light and generate more electron-hole pairs, which in turn enhanced the photocatalytic activity and reduced the reaction time as compared to the other materials.

Table 4

Comparison of MB degradation of the TiO₂-codoped samples of this work with other materials.

Materials	Synthetic conditions	$S_{\text{BET}}$ (m2·g-1)	Degradation percentage of MB	References
Bi2O3	Calcination method at 500°C for 6 h	2.9	70% after 4 h under visible light irradiation	[24]
CaBi6O10	Impregnation-calcination method at 650°C for 12 h	2.5	97% after 4 h under visible light irradiation	[24]
CeO2/V2O5	Thermal decomposition at 400°C for 30 min	—	64.2% after 4 h under visible light	[25]
CeO2/CuO	Thermal decomposition at 400°C for 30 min	—	70.1% after 4 h under visible light	[25]
ZnO/CdO	Thermal decomposition at 350°C for 3 h	—	75% after 4 h under visible light	[26]
g-C3N4-CdS	Precipitation method	166.5	90.45% after 3 h under visible light	[27]
F-doped TiO2	Sol-gel method, calcined at 500°C for 1 h	39.6	90% after 4 h under visible light	[28]
TiO2-1%Co-2.5%Fe	Solvothermal method at 180°C	174	90% after 2 h under visible light	This work

4. Conclusions

In this work, we presented the characterization of TiO₂ nanoparticles codoped with Co²⁺ and Fe³⁺ synthesized through solvothermal treatment at 180°C without further calcination at high temperatures. Undoped TiO₂ and Co single-doped TiO₂ materials were also prepared using the same procedure to compare with codoped TiO₂ in terms of optical absorption and photocatalytic efficiency for MB degradation. The doped and codoped TiO₂ materials of the present work exhibited better adsorption characteristics (e.g., their BET surface areas of 164-174 m²/g) than doped and codoped TiO₂ materials of previous works; as-prepared codoped TiO₂ materials had significantly enhanced optical adsorption toward the visible light region. Among the TiO₂ samples studied, the codoped TiO₂ materials showed the highest photocatalytic activity for the degradation of MB; they degraded about 90% of MB within 120 min under visible light irradiation. The TiO₂ samples codoped with Co and Fe in this work showed higher photocatalytic efficiency for the degradation of MB than those of metal oxides, mixed metal oxides, and other materials in the previous works. The pseudo-second-order kinetic model fits well for describing the degradation of MB on the TiO₂ codoped with Co and Fe. Hence, this study provides a simple route to synthesize an effective photocatalyst for the degradation of dye compounds under visible light for wastewater treatment.