Titanium dioxide - activated carbon composite for photoelectrochemical degradation of phenol

In this study, titanium dioxide (TiO 2 ) and titanium dioxide – activated carbon composite (TiO 2 – AC) were prepared by sol-gel method for photoelectrochemical (PEC) applications. Characterization of the materials was performed by scanning electron microscope, energy dispersive X-ray analysis, Fourier transform infrared spectroscopy, X-ray diffraction, and diffuse reflectance spectroscopy. The results show that TiO 2 was successfully loaded on activated carbon (AC), producing TiO 2 – AC with 2.61 eV of bandgap energy, lower than that of TiO 2 (3.15 eV). Photoanodes based on TiO 2 and TiO 2 – AC were fabricated and applied to PEC experiments for phenol degradation. In comparison with the TiO 2 photoanode, the TiO 2 – AC one exhibited superior photocatalytic activity, which was indicated by a high current density and effective phenol removal. A mechanism of phenol PEC degradation on the TiO 2 – AC photoanode was proposed, which includes interaction between protonated phenol and active sites bearing oxygen on the photoanode surface. A kinetic model according to this mechanism was also established and fitted to experimental findings, resulting in rate constants of elementary reactions.


Introduction
Phenol and phenolic compounds are commonly used in pharmaceuticals, insecticides, cosmetics, and other industrial substances [1]. Phenol is known as a hazardous pollutant due to its toxicity and high stability for a long period of time in the environment [2,3]. Phenol exposure may cause acute and/or chronic diseases on the skin, eye, respiratory and nervous systems [4,5]. There are some techniques for phenol removal from wastewater, as summarized in [6]. Among advanced oxidation processes, photocatalysis is described as an effective choice for phenol degradation. Based on TiO2 photocatalyst synthesized by the solgel method, phenol degradation was attained at 58.8% after 240 minutes of UV illumination [7]. In the presence of carbon, photogenerated charge recombination of TiO2 was delayed [8], suggesting an improvement in phenol degradation. Synergy effect between TiO2 and carbon-based materials was explored in the previous reports on multi-walled carbon nanotubes -TiO2 [9], activated carbon -TiO2 [10], graphene -TiO2 [11], and carbon fiber -TiO2 [12].
Difficult recovery and fast photoexcited electron-hole pairs recombination of catalysts are two considerable drawbacks of photocatalysis, which can be minimized with the photoelectrochemical (PEC) method [13,14]. According to the PEC principle, a photocatalyst is coated onto a photoelectrode, to which a bias voltage is applied to improve photogenerated charge carrier separation, thus enhancing the activity of the photocatalyst [15,16]. Therefore, the PEC method has attracted considerable interest in water splitting [17] as well as organic pollutants degradation [13]. Us-Chimica Techno Acta 2022, vol. 9(4), No. 20229416 ARTICLE 2 of 11 ing a TiO2-based photoanode for the PEC experiment, phenol degradation achieved 73.76% after 120 minutes of UV illumination under 0.8 V of applied voltage [18]. By adding peroxymonosulfate into the PEC system, a photoelectrode based on Co3O4-loaded carbon fiber demonstrated 100% phenol degradation within 90 minutes under UV radiation at 1.5 V of applied voltage [19]. The contribution of carbon to PEC behavior of the C/TiO2 composite was explored by Haro et al. [20]. It was found out that carbon promoted charge transfer reactions on C/TiO2 photoelectrode surface through enhancement of charge carrier generation and separation. However, the application of C/TiO2 in PEC degradation of organic compounds has not been reported in the literature.
In this study, TiO2 and TiO2-activated carbon (TiO2-AC) composites were synthesized by a sol-gel method for PEC degradation of phenol. Effects of AC on PEC properties of the TiO2/AC photoanode for phenol were determined. Moreover, a mechanism of phenol PEC degradation was proposed, revealing a kinetic model describing the surface responses of the TiO2/AC photocatalyst.
Characterization of materials was conducted with analytic methods such as scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) using Prisma E SEM, Xray diffraction (XRD) recording by D2 PHASER, diffuse reflectance spectra carried out by FL-1039 (HORIBA), Fourier-transform infrared spectroscopy (FT-IR) with Nicolet iS5, current − voltage (J-V) curve and electrochemical impedance spectroscopy (EIS) using MPG2 Biologic system.

Preparation of TiO 2 , TiO 2 -AC composite, and photoanodes
10 ml of TiCl4 and 10 g of AC were added into an Erlenmeyer flask containing 200 ml of hexane under N2 atmosphere in a glovebox. The Erlenmeyer flask was then covered and moved to an ultrasonic tank. The mixture was dispersed under 40 kHz of sonication for 15 minutes before adding 500 ml of distilled water. After 30 minutes of sonication, the solid phase was separated and washed with distilled water in a vacuum filtration system until the filtrate reached neutral pH. The obtained solid was thermally treated at 550 °C for 30 minutes to produce the TiO2-AC photocatalytic composite.
To prepare the TiO2 photocatalyst, the above procedure was applied without AC.
A coating mixture (consisting of 80% w/w of polyester, 19% w/w of ethanol, and 1% w/w of BMIM PF6) was deposited onto a SUS 304 stainless sheet (0.8 mm of thickness, 100100 mm of dimension) as a photoanode substrate by the dip-coating method. Then, an abundant amount of photocatalyst was spread on the photoanode surface and pressed under 3 N/cm 2 of pressure to create contact between the photocatalyst and the photoanode substrate. The obtained photoanodes based on TiO2 and TiO2-AC photocatalysts were stored at room temperature for one week before being utilized in the PEC system.

Photoelectrochemical measurements
A three-electrode cell was used for PEC measurements with a Pt grid as a counter electrode, an Ag/AgCl as a reference electrode, the photoanode as a working electrode, and a UV-C mercury lamp (9 W) vertically soaked in a solution. All PEC measurements were conducted under fluorescent light (400−600 lux of illumination) in a 10 mg/L of phenol solution, pH = 5 with UV and non-UV radiation. The J-V curves were recorded according to the linear sweep voltammetry technique in a potential range of 0 V to +1.5 V at a scan rate of 50 mV/s. The EIS characteristics were obtained over a frequency range of 10 mHz to 10 kHz with 10 mV amplitude.

Photoelectrochemical degradation of phenol
Bath experiments for PEC degradation of phenol under UV and non-UV illumination were performed in a stirred PEC reactor containing 1 liter of phenol solution. The Pt grid and photoanode were vertically dipped into phenol solution and connected to a controllable DC voltage source. Between the Pt grid and the photoanode, the UV-C lamp was also immersed in the solution. Phenol concentration was determined by the colorimetric method using HI 3864 phenol test kit with the instrumental error up to 0.1 mg·L −1 . The efficiency of phenol removal was calculated by Equation (1).
where Ct, C0 are the concentrations of phenol after contact time t and at the start (initial concentration) in mg·L −1 , respectively.

Characterization of materials
The morphology of TiO2, AC and TiO2-AC can be observed in SEM images. As shown in Figure 1a, TiO2 particles are irregular polygonal shapes that are less than 25 m in size. Porous structure with cavities and pits of AC is disclosed in Figure 1b. For TiO2-AC, a large number of TiO2 particles was loaded on AC structure by filling into the pits instead of the surface of cavities, as shown in Figure 1c. Similar observation was also reported in a previous publication [21]. Located in pits, TiO2 particles could avoid being washed out from AC in photoanode preparation and utilization. As expected, C, O, and Ti elements in TiO2-AC were indicated in the EDX results, as shown in Figure 1d.
Crystal phase characteristics of TiO2, AC, and TiO2-AC were explored with X-ray diffraction analysis as shown in Figure 2. The reflections from the (002) and (100) planes of aromatic rings of amorphous carbon structure caused the diffraction peaks at 26.2° and 43.5° [22]. For the XRD pattern of TiO2, peaks appearing at 27.7°, 36.1°, 41.2°, 54.2°, and 56.7° were indexed as (110), (101), (111), (211), and (220) planes of rutile phase, whereas the presence of anatase phase was identified by the peaks at 25.3° and 48.3°, according to (101) and (200) planes [23]. However, only characteristic peaks of rutile TiO2 exist in TiO2-AC material, as shown in the XRD pattern of TiO2-AC. This may be due to the effect of carbon on anatase to rutile transformation of TiO2 through the formation of oxygen vacancies, as mentioned in [24][25][26]. This is a disadvantage of our TiO2-AC because rutile TiO2 exhibits lower photocatalytic activity than anatase TiO2 [27].
Chemical bonds in TiO2, AC, and TiO2-AC were identified by FTIR analysis ( Figure 3). As shown in the FTIR spectrum of TiO2, a strong absorption band at about 500 cm −1 reflects the vibration of Ti−O bond, and a band at 1653 cm −1 corresponds to O−H bending in absorbed water molecules [28]. Stretching vibrations of C−O and C=C bonds in AC structure were identified by bands at 1097 cm −1 and 1554 cm −1 , respectively [29]. A band appearing at 3469 cm −1 is associated with the stretching vibration of O−H bond and free water [28]. It can be observed in the FTIR spectrum of TiO2-AC that characteristic bands of the bonds in TiO2 and AC appeared again without a new band, indicating TiO2 did not conjugate to AC in TiO2-AC by a chemical bond. A similar result could be found in [21,30]. As presented in Figure 4a, UV-V is diffuse reflectance spectra of TiO2 and TiO2-AC shows the absorption edge of the TiO2 and TiO2-AC at about 370 nm and 420 nm, respectively. It indicates the red-shift of the TiO2-AC towards the visible region in comparison with TiO2. Tauc plots, relationships of h vs (h) 1/2 of materials, are illustrated in Figure  4b, revealing band gap energies of 3.15 eV and 2.61 eV for TiO2 and TiO2-AC, respectively. Carbon not only promotes the 4p → 4s electronic transition at defect points in the titanium atoms [31], but also improves the electron transfer due to the high electronic conductivity of carbon, thereby resulting in the lower band gap energy of TiO2-AC compared to TiO2. This result indicates that TiO2-AC can effectively separate electron-hole pairs in the visible region, which promotes photocatalytic activity under the solar illumination.

Characterization of photoanodes
A TiO2-AC photoanode was selected for morphological analysis by the SEM technique. It can be observed in Figure 5a that TiO2-AC particles in random shapes were widely spread on the photoanode surface generating a rough structure. The original morphology of TiO2-AC (Figure 1c) exposed on the photoanode surface (as shown in Figure 5b) indicated that TiO2-AC particles were not immersed completely in the coating mixture. It proves that photoactive sites in TiO2-AC were disclosed on the photoanode surface, enabling photoexcitation of the TiO2-AC photoanode in the photoelectrochemical system. The mean thickness of a coated layer on the photoanode was determined to be 136 m.
Photoelectrochemical properties of the TiO2 and TiO2-AC photoanodes were examined using a three-electrode cell in phenol solution at pH = 5. Figure 6 demonstrates the current density of the TiO2 and TiO2-AC photoanodes under non-UV and UV illumination applying linear sweep voltammetry. In the absence of UV (non-UV), photoanodes present low current density. Under UV radiation, photocurrent densities of photoanodes significantly increase. The TiO2-AC photoanodes generated a current density of 283 A/cm 2 (at 1.45 V vs. Ag/AgCl), which is approximately 2.2 times higher than that of the TiO2 photoanode. This result may be due to the decrease of TiO2 band gap in the presence of AC. Photocatalytic activity of a photoanode driving a reaction can be evaluated through an onset potential, which is a potential at the intersection point between J-V curve in non-UV radiation and the tangent line with a maximum slope of J-V curve in UV radiation [32,33]. The onset potential of the TiO2 photoanode driving phenol oxidation (~1.0 V vs Ag/AgCl) is about 350 mV higher than that of the TiO2-AC photoanode (~0.66 V vs Ag/AgCl), indicating heterojunction formation of TiO2/AC in the TiO2-AC material [34]. The low onset potential of the TiO2-AC photoanode demonstrated an effective charge separation and transfer, manifesting a favorable application of the TiO2-AC photoanode for PEC degradation of phenol.     Figure 7 presents Nyquist plots of TiO2 and TiO2-AC photoanodes under UV and non-UV illumination. There is only one semicircle for each Nyquist plot, indicating that PEC process on the photoanodes for phenol was controlled by trap electron transfer [35]. Smaller diameter semicircle demonstrates lower charge transfer resistance [36]. It can be observed clearly from Figure 7 that the semicircle in the Nyquist plot of the TiO2 and TiO2-AC photoanodes under UV radiation is smaller than that under the non-UV excitation. This finding proves that the charge transfer of photoanodes was improved under UV illumination. The TiO2-AC photoanode exhibited a better charge transfer in comparison to TiO2 one because Nyquist plots of the TiO2-AC photoanode show smaller semicircles in both UV and non-UV excitation. This result indicates that the presence of AC not only did not reduce the photocatalytic properties of TiO2, but also improved the electrical conductivity of TiO2-AC, advancing the applicability of the TiO2-AC photoanode in PEC process.

Photoelectrochemical degradation of phenol
Applied external voltage (Vapp) can improve PEC performance of TiO2 by enhancing the generation and separation of charge carriers [37]. After 30 minutes of PEC treatment under different Vapp in the range of 0−0.7 V, phenol and COD removal at the TiO2-AC photoanode were recorded and shown in Figure 8a. Phenol and COD removal exhibit similar trends under Vapp variation. Phenol can be converted to organic intermediates during the photocatalysis process before it is completely oxidized to CO2 [38]. If COD removal equates to phenol removal, the total removed phenol is oxidized to CO2 without the formation of intermediates. As illustrated in Figure 8a (Figure 8a), which is close to the onset potential of the TiO2-AC photoanode, suggesting a decisive contribution of the applied external voltage to the PEC degradation of phenol.
Significant improvement of phenol degradation with PEC process as compared to photocatalysis is shown in Figure 8b. Under UV illumination, electrons and holes were photogenerated on TiO2, enhancing charge transfer between phenyl ring and photoanodes [39] and promoting phenol degradation. However, fast recombination of these electron-hole pairs causes a low limit of phenol removal, as seen in the TiO2/UV and TiO2-AC/UV curves. The recombination time of electron-hole pairs can be longer than charge transfer time in the redox reaction under appropriate Vapp. As expected, at Vapp = 0.7 V, TiO2/PEC and TiO2-AC/PEC curves in Figure 8b exhibit an increase in phenol removal. Moreover, the TiO2-AC photoanode presents a higher efficiency of PEC degradation of phenol than the TiO2 photoanode owing to the contribution of electric conductivity of AC. After 60 min of contact time, phenol PEC degradation achieved 75.9% on TiO2-AC, which is lower than that on TiO2 nanotubes [40], but significantly higher than on TiO2 [18].
Chimica Techno Acta 2022, vol. 9(4), No. 20229416  As discussed above, the following experiments for the kinetic study of phenol PEC degradation were conducted at the TiO2-AC photoanode under Vapp = 0.7 V and UV illumination.

Kinetic of photoelectrochemical degradation of phenol in acidic solution
Photocatalytic degradation of phenol can be described by the following Langmuir -Hinshelwood (L-H) kinetic mechanism [41][42], in which phenol is first adsorbed onto the photocatalyst surface and then decomposed. Therefore, this L-H model (2) was used to fit our experimental data, revealing negative values of KB (as shown in Table 1). However, this is unreasonable because the adsorption equilibrium constant KB must be positive.
• L-H model: • First-order model: In the case of chemical reaction control, a first-order kinetic model (3) well described the photodegradation of phenol on TiO2/AC [43] as well as ZnO, TiO2 and ZnO-TiO2 photocatalysts [44]. Photoelectrochemical degradation rate following the first-order model was identified for acidic red 17 dye on ammonium persulphate [45], and phenol on PbO2 anode [46]. However, the determination coefficients (R 2 ) obtained from fitting our experimental data with this model were not close to 1 (as shown in Table 1). It signifies that the PEC degradation of phenol occurred in a complex mechanism and was not controlled exclusively by a chemical reaction process.
We proposed a mechanism of PEC degradation of phenol through reactions (4)-(6) based on the previous reports. Spallart et al. [47] proved that water competed with aromatic compounds in PEC oxidation. Moreover, oxygen atoms from water molecules can form O 2 •− radicals at photoexcited points [39]. Hence, the active site on the photoanode surface ( * + ) was suggested to interact with a water molecule under the applied external potential and form an active site bearing oxygen ( * + O 2− ) following reaction (4). In other consideration, phenol (Ph) can be protonated into H + Ph form [48] in acid solution, promoting the transfer of electron pair of oxygen in −OH group into the aromatic ring [49] and resulting in polar structure H + Ph − according to reaction (5  We assume that reactions (4)-(6) occurred in the presence of large amounts of H + ions and H2O. Therefore, the rates of these reactions depend on the concentration of phenol (x1),  (7), (8), (9).
where x1, x2, x3 are in mmol·L −1 , and [ * + ] is an apparent concentration of active sites in mmol·L −1 . Under UV illumination, TiO2-AC particles on the photoanode surface were excited to create photoexcited sites ( * ). Then, under the applied external potential, the photoexcited site lost n electrons and became an active site ( * + ). Supposing that the total number of * + does not change with an apparent concentration [ * + ] , equation (10) is obtained: The value of [ * + ] can be determined from (10) and substituted in to (9) to result in (11): The parameters in the proposed model ( + , w − , p + , p − , kd, and [ * + ] ex ) were determined by fitting experimental data with the model using the least-square method for x1 objective: where, 1 is the phenol concentration predicted by the model, i = 1, 2, …, n denotes i th value, and n = 12 is the number of experimental data points. The determination coefficient (R 2 ) is used to evaluate the goodness of fit of the model as presented in formula (13).
where SSS = ∑( 1 − 1 ) 2 is the residual sum of squares, is the mean value of x1.
Numerical solutions of ordinary differential equations (7), (8), and (11) were carried out using Runge -Kutta 4 th order method with initial conditions: 1 (0) = [Ph] 0 = 0 94.11 , 2 (0) = 3 (0) = 0 (94.11 is molecule weight of phenol). The minimization problem (12) was solved with the help of the Excel Solver tool (ver. 2016) with a GRG non-linear option. Figure 9 shows the effects of initial phenol concentration (C0) and initial pH (pH0) on phenol removal under both observations of the experiment and simulation. The proposed model exhibited a good description of the experimental data due to the closeness of R 2 to 1, obtaining kinetic parameters as summarized in Table 1.
Experiments of PEC degradation of phenol were performed at different C0 from 5 to 20 mg·L −1 at pH0 = 5, revealing kinetic behavior as presented in Figure 9a. Phenol exhibits a property of UV light interception [39,50], causing a decrease in the active site quantity (* n+ ) and oxidation rate constant (kd) while increasing C0 as shown in Table 1. Consequently, the rate of phenol removal is lower with higher C0. Moreover, contact time for 99.95% phenol removal was predicted to be 1110, 725, 490, and 450 minutes at C0 = 5, 10, 15, and 20 mg·L −1 , respectively, according to the proposed model. It proves that the PEC degradation of phenol on TiO2-AC photoanode did not reach equilibrium and tended to complete phenol removal.
pH0 of the solution is an important parameter in the photodegradation of phenol because of the variation of charge properties of phenol at different pH values [51]. In this study, PEC degradation of phenol was studied at various pH0 values of solution (3, 4, 5, and 6) with C0 = 20 mg·L −1 . Phenol is protonated to H + Ph − in the presence of ion H + according to reaction (4). Therefore, the lower pH0 value was, the more H + Ph − was produced, resulting in an improvement in phenol degradation (as shown in Figure 9b).  (Table  1) at 5, 10, 15, and 20 mg·L −1 of C0, respectively, which proves that high C0 is a disadvantage to protonation of phenol.
The curves in Figure 10b present the relationships between the amount of * n+ O 2− and the contact time. In all studied C0, [* n+ O 2− ] quickly increases to reach the equilibrium concentration ([* n+ O 2− ]eq). The ratios of [* n+ O 2− ]eq and [* n+ ]ex are larger than 0.9, as shown in the inset in Figure 10b. It means that most photoexcited sites changed to form active sites bearing oxygen, which acts as a reactant in PEC degradation of phenol. Although [* n+ O 2 − ]eq decreased with the increase of C0, the high value (>350 mmol·L −1 ) was enough to interact entirely with protonated phenol molecules.  Figure 11b. The inset in Figure 11b presents the effect of pH0 on the ratio of [ * + O 2− ]/[ * + ] ex . It can be seen that this ratio at pH0 = 5 is the highest (close to 1). It means that more separation between the point of zero charge and pH0 caused lower [ * + O 2− ]/[ * + ] ex .

Conclusions
In this work, TiO2 and TiO2-AC photocatalysts were successfully synthesized by the sol-gel method to apply to the PEC degradation of phenol. The TiO2-AC photoanode exhibits higher photoactivity for degrading phenol under UV-C illumination than TiO2. Effects of applied external voltage, pH, and initial concentration of phenol on the kinetics of phenol PEC degradation were also experimentally investigated in this study. The interaction mechanism between protonated phenol and the active site bearing oxygen was well-demonstrated for the phenol PEC degradation. The kinetic constants and the concentration variations of protonated phenol and the active site bearing oxygen were determined by fitting the established kinetic model to the experimental data.

Supplementary materials
No supplementary materials are available.

Funding
This research had no external funding.