Anodic Titanium Oxide Layers Modified with Gold, Silver, and Copper Nanoparticles

Titanium(IV) oxide is commonly used in photocatalysis. However, it has some drawbacks, e.g., a high rate of electron-hole recombination and a wide bandgap. In here, the surface of anodic titanium(IV) oxide (ATO) was modi ﬁ ed with metal nanoparticles (gold, silver, and copper) in order to enhance its photoelectrochemical (PEC) and photocatalytic (PC) properties. SEM analysis revealed that Au, Ag, and Cu nanoparticles obtained on an ATO surface by chemical methods had an average diameter of 50 ± 10 nm, 30 ± 6nm, and 25 ± 3 nm, respectively. Enhancement of photoelectrochemical water-splitting current e ﬃ ciency in the wavelength range of 300 – 400 nm was observed due to the occurrence of the Schottky barriers. However, the nanoparticles had no e ﬀ ect on the current e ﬃ ciency in the range of 400 – 600nm which meant that the surface plasmon resonance (SPR) e ﬀ ect was not observed. A rate of methyl red photodecomposition was improved after the modi ﬁ cation of the ATO surface. The best results were obtained for ATO decorated with gold nanoparticles.


Introduction
Titanium(IV) oxide is an n-type semiconductor that is known from its good stability, high resistance to corrosion, nontoxicity, biocompatibility, and low cost of production. Due to these properties, TiO 2 is widely used in medicine, industry, and the military. Recently, titanium(IV) oxide nanostructures, like nanoparticles [1], nanowires [2], nanorods [3], nanotubes [4], or nanopores [5] are being intensively studied because of their unique properties. They can be obtained by methods such as the sol-gel process [6], solvothermal synthesis [7], electrodeposition [8], and anodization [4,5,9]. Among those techniques, anodic oxidation of titanium substrate is favorable due to the possibility of obtaining highly ordered nanostructures in a simple, non-time-consuming, and low-cost production way. What is more, dimensions (tube length and diameter) of those structures can be easily controlled by several parameters, like external voltage, time, temperature, and type and composition of the electrolyte [10][11][12][13].
As is well known, TiO 2 is commonly used in photocatalysis due to its ability to absorb the UV irradiation. It is possible because of the value of its bandgap, which is 3.2 eV and 3.0 eV for anatase and rutile, respectively [14,15]. When a semiconductor absorbs UV light, an electron from the valence band is excited to the conduction band. Simultaneously, an electron hole is created in the valence band. Then, electron-hole pairs can recombine, but this process is undesirable. What is beneficial from the point of view of photocatalysis is that electrons and holes, responsible respectively for reduction and oxidation reactions, can migrate to the surface of a material and react with adsorbed compounds, e.g., water molecules [16]. What is worth mentioning is the fact that the band edge positions of conduction and valence bands of the efficient photocatalyst have to be close to the redox potential of the adsorbed species. For titanium(IV) oxide, those band edges are located near the redox potential of water molecules [15]. As a result of the reaction occurring at the surface of the catalyst, highly reactive oxygen species, e.g., radicals are produced which can react in the next step with other compounds from the electrolyte.
Despite the advantages mentioned above, TiO 2 has some crucial drawbacks that limit its utilization in photocatalytic processes. Titanium(IV) oxide has a wide bandgap which allows absorbing light only from the UV range. On the other hand, the solar spectrum consists of approximately 5% of UV and 40% of visible light. That is why TiO 2 cannot be used as a photocatalyst under sunlight irradiation which would have been preferable in terms of lowering processing costs. Additionally, the rate of electron-hole recombination is high and significantly decreases photocatalytic efficiency. In order to extend potential applications of TiO 2 , numerous modifications have been performed that expand the spectrum of the light absorption and enhance the efficiency of photocatalysis and photoelectrocatalysis [17]. Strategies are based on, e.g., nonmetal [18,19] and metal [20,21] doping, coupling with other semiconductors [22,23], and surface decoration with nanoparticles [24,25].
The decoration of the TiO 2 surface with metal nanoparticles can be made by using photo-or electrodeposition, precipitation, and ion-exchange techniques. The most commonly used metals are noble metals (Au, Ag, and Pt [24][25][26]) and some transition metals (Ni, Fe, and Cu [27][28][29]). It is well known that coupling some metals with the semiconductor results in Schottky junctions at their interface [30]. This phenomenon is described as a flow of electrons from the semiconductor to the metal, that leads to bending of the energy bands of the semiconductor. The presence of the Schottky barrier prevents a return of electrons to the semiconductor. In other words, electrons can be trapped into the metal, and thus the rate of recombination of the electron-hole pair is reduced. That is why this type of junction can enhance photocatalytic and photoelectrocatalytic efficiencies. Paramasivam et al. [24] compared the photocatalytic efficiency of the degradation of acid orange 7 on an unmodified flat TiO 2 film and nanotubular anodic TiO 2 with anodic TiO 2 nanotubes decorated with gold and silver nanoparticles. It was found, that in the presence of the catalyst loaded with noble metal nanoparticles the rate of dye degradation significantly increased.
On the other hand, the second phenomenon that can occur at the metal-semiconductor interface is the surface plasmon resonance (SPR). Surface plasmons are defined as the collective oscillation of conduction electrons (surface plasmons) at the interface between a metal and other materials. Under light irradiation, charge-density oscillations might occur and the formation of the electric field can be observed [31]. Metals such as gold [32], silver [33], and copper [34] are known for their ability to undergo the SPR effect. According to the literature, there are three possible mechanisms of the SPR that may have an influence on photocatalytic reactions, namely scattering of resonant photons by metal, plasmon resonance energy transfer, and hot electron transfer [31,[35][36][37].
Silver, gold, and copper nanoparticles, which are characterized by the presence of absorption bands at 400 nm, 520 nm, and 580 nm, respectively, are of particular interest [38]. Zhao et al. [39] studied the effect of size, distribution, and morphology of gold-covered TiO 2 nanotubes on their photoelectrochemical properties. The theoretical calculations and experimental results confirmed that the SPR effect led to a significant increase in the magnitude of generated photocurrents. Other examples of the effect of nanoparticle deposition on photoelectrochemical and/or photocatalytic properties are presented in Table 1.
In this work, we present the photoelectrochemical and photocatalytic properties of anodic titanium(IV) oxide decorated with chemically deposited metal nanoparticles. The anodic materials were compared in terms of the morphology, amount of nanoparticles on the oxide surface, generated photocurrent at different wavelengths and applied potentials, and rate of methyl red decolorization. Additionally, we showed whether the SPR effect or Schottky barrier occurs during surface modification of anodic TiO 2 under the proposed experimental conditions.

Materials and Methods
2.1. Synthesis of Anodic Titanium Oxide (ATO). Titanium foil (99.5% purity, 0.25 mm thick) was polished electrochemically, 2 Journal of Nanomaterials and then chemically [46]. The anodization process was carried out in a two-electrode cell, where the polished Ti sample was used as an anode, and a titanium plate was used as a cathode. Electrodes were kept at the constant distance of 2 cm. Nanoporous anodic titanium oxide layers were synthesized in an ethylene glycol-based solution containing NH 4 F (0.38 wt.%) and H 2 O (1.79 wt.%). A three-step anodization was carried out at the constant voltage of 40 V at 20°C. The first and second anodizing steps lasted 3 h, and after each step, an adhesive tape was used to remove the resulting oxide layer. The third anodizing step was carried out for 10 min in a freshly prepared electrolyte. Each anodizing step was performed at the constant stirring rate of 200 rpm [47]. Afterwards, TiO 2 samples were annealed in air at 500°C for 2 h with a heating rate of 2°C min −1 using a muffle furnace (FCF 5SHM Z, Czylok) [48]. The morphology of ATO layers was characterized using a field emission scanning electron microscope (SEM, Hitachi S-4700).

Deposition of Nanoparticles.
Copper nanoparticles (CuNPs) were deposited using a SILAR method [41]. In the first step, anodic TiO 2 was immersed in a 0.01 M copper (II) acetate solution for 10 min. Then, the sample was rinsed with distilled water and placed in a 0.01 M NaBH 4 solution for 10 min. This cycle was performed five times. Silver nanoparticles (AgNPs) were synthesized using the citrate method [41]. The same volumes (5 mL) of 1 mM silver nitrate solution and 1 mM sodium citrate were mixed together in a beaker and placed in an ultrasonic bath. Anodic TiO 2 was immersed in the solution and 2.5 mL of a 0.01 M sodium borohydride solution was added dropwise. Finally, ATO was rinsed with water and dried in air. Gold nanoparticles (AuNPs) were obtained in a mixture of 1 mM sodium citrate and 1 mM chloroauric acid(III) (3 : 1 in vol.) heated to 60°C. Anodic TiO 2 was immersed in the mixture, and 1 mL of 0.01 M sodium borohydride was added. The samples were rinsed with water and dried in air.
In order to confirm the presence of gold and silver nanoparticles, the postreaction solutions were analyzed by using a UV-Vis spectrophotometer (Evolution 220, Thermo Fisher Scientific). UV-Vis spectra of Au and Ag nanoparticles were recorded against distilled water in the range of 200-780 nm.

Photoelectrochemical Measurements.
Photoelectrochemical measurements were performed using a three-electrode cell with a quartz window, where the nanostructured ATO layer was used as a working electrode (WE), a platinum foil was used as a counter electrode (CE), and a Luggin capillary with a saturated calomel electrode (SCE) was used as a reference electrode. The generated photocurrents were measured using a photoelectric spectrometer (Instytut Fotonowy, Poland) equipped with the 150 W xenon arc lamp and combined with a potentiostat [47,48]. The photoelectrochemical characterization was carried out in 0.1 M KNO 3 at the potential range of 0-1 V vs. SCE and wavelengths ranging from 200 to 800 nm. Incident photon to charge carrier efficiency (IPCE) was calculated from the following equation [49]: where I p λ and P λ are the photocurrent density (A m −2 ) and incident power density of light (W m −2 ), respectively, at wavelength λ (nm). The constant of 1240 has the unit W nm A −1 .
where C 0 is the initial MR concentration and C i is the MR concentration after illumination.

Results and Discussion
Typical morphology of anodic TiO 2 layers obtained by three-step anodization at the potential of 40 V is shown in Figure 1(a). Such nanostructures are characterized by a top porous layer and nanotube structure in cross-sectional view (Figure 1(a) insert). A thickness of the oxide layers was 2.2 ± 0.2 μm. The surface of anodic TiO 2 layers with deposited copper, gold, and silver nanoparticles is presented in Figures 1(b)-1(d), respectively. Prior to the deposition of noble metal nanoparticles, titanium oxide layers were heat treated at 500°C in order to transform amorphous oxide into the anatase phase [48]. As shown in Figure 1, nonaggregated nanoparticles are randomly distributed over the ATO surface regardless of the type of deposited metal. The presence and distribution of nanoparticles on the surface of ATO layers was confirmed by EDS analyses (Figures 2 and 3), and the content of metals was different depending on the applied method of synthesis. The size of metal nanoparticles was estimated on the basis of SEM micrographs. In the case of copper nanoparticles obtained by the SILAR method, the average diameter equals to 50 ± 10 nm, and the copper content in the ATO layer was found to be 0.4 at.%. The deposited gold nanoparticles have the average diameter of 30 ± 6 nm, and the Au total content of 2.2 at.% was estimated in the sample. The average size of silver nanoparticles was 25 ± 3 nm, and the Ag content determined by EDS was about 0.1 at.%. According to XPS studies performed for similar anodic materials [42,50,51], there are no chemical bonds between a noble metal and Ti. After deposition of Au and Ag nanoparticles on the surface of ATO layers, the solutions with metal nanoparticles were subjected to 3 Journal of Nanomaterials spectrophotometric analyses (Figure 3). The obtained spectrum characterized by maximum absorbance at approx. 520 nm (Figure 3(a)) confirms the presence of gold nanoparticles. A similar analysis performed for a solution containing silver nanoparticles revealed that AgNPs absorb at approx. 400 nm, which is in good agreement with the literature reports [52]. For both nanoparticle solutions, the observed wide band in the UV-Vis spectrum (instead of a narrow well-defined peak) can be attributed to an agglomeration of nanoparticles, which can occur in aqueous solutions.
The ATO layers decorated with noble metal nanoparticles were used as photoanodes for photoelectrochemical water splitting. It was expected that additional maxima in the photocurrent density vs. wavelength spectra     Journal of Nanomaterials (AuNPs~520 nm, AgNPs~400 nm, and CuNPs~570 nm) might be observed as a result of a modification of the anodic TiO 2 surface with gold, silver, and copper nanoparticles. However, for all tested anodic materials, there was no significant increase observed in the generated photocurrent in the wavelength range of 400-600 nm. In the case of a photoanode modified with copper nanoparticles, the metal surface could have oxidized and, therefore, the SPR effect was not observed [53]. In other cases, it is possible that the amount of metal nanoparticles on the ATO surface was too small to observe the SPR effect or/and the size of particles was too big to observe quantum effects. On the other hand, an improvement in photoelectrochemical properties was seen in the wavelength range of 300-400 nm ( Figure 4). These results suggest that for the studied photoanodes the Schottky barrier is present. The radiation is absorbed by the semiconductor, and the photogenerated electrons are trapped in metal nanoparticles, which promotes the charge separation and contributes to the efficiency of the process. The complex photoelectrochemical behavior of the tested photoanodes was studied at the potential range of 0-1 V vs. SCE and wavelengths ranging from 300 to 400 nm. The photocurrent densities as a function of incident light wavelength and applied potential were recorded for anodic TiO 2 samples annealed at 500°C and decorated with CuNPs, AuNPs, and AgNPs ( Figure 5). For the nonmodified anodic TiO 2 layer, the maximum photocurrent density of about 70 μA cm −2 is observed at 350 nm ( Figure 5(a)). In general, the deposition of nanoparticles results in an improvement of the photoelectrochemical response of nanostructured anodic TiO 2 films, which is in good agreement with the literature (see Table 1). For instance, the maximum photocurrent densities of about 90 μA cm −2 and 80 μA cm −2 were recorded at 350 nm and 1 V vs. SCE for samples modified with copper ( Figure 5(b)) and silver ( Figure 5(d)) nanoparticles, respectively. For the ATO sample decorated with AuNPs, much higher photocurrents (about 100 μA cm −2 ) at 350 nm were observed. However, it should be mentioned that the estimated atomic content of deposited noble metals was   5 Journal of Nanomaterials different for each electrode. What is more, IPCE values obtained for the samples modified with metal nanoparticles are mostly higher than those observed for bare anodic titania ( Figure 5(e)). For the ATO surface decorated with Au nanoparticles, the higher conversion efficiencies close to 45% are visible at the higher applied potentials.    Journal of Nanomaterials In order to verify the photocatalytic effectiveness of ATO layers decorated with metal nanoparticles in decomposition of MR, photodegradation tests were performed. Prior to photocatalytic tests, a series of standard solutions to generate a calibration curve was prepared at the MR concentration range of 0-5 mg L −1 . The typical UV-Vis spectra obtained in 0.01 M hydrochloric acid are presented in Figure 6(a). Based on the band at 520 nm, the calibration curve was created and used for further determination of dye concentration (Figure 6(b)). During the photocatalytic tests, the dye solution was irradiated with 350 nm UV light in the presence of the synthesized materials, and a decrease in the intensity of the 520 nm band was monitored. Using the calibration curve, the concentration of MR was determined at given time intervals. The dye concentration as a function of exposure time to the UV radiation is presented in Figure 6(c). As can be seen, the dye decomposition profiles are very   Figure 6(c)). In order to determine the kinetic rate constant, the ln(C) values were plotted with respect to time for all studied ATO samples (Figure 6(d)). The plotted data were best fitted by straight lines of different slopes which indicate that the reaction follows pseudofirst order kinetics. From the slopes of those plots, the reaction rate constants were determined. The best photocatalytic properties, indicated by the highest rate constant of 0.30 h −1 when compared with nonmodified ATO (k = 0 24 h −1 ), were observed for the ATO layer decorated with AuNPs. A slight improvement in photocatalytic properties was found also for the ATO layer modified with silver nanoparticles (k = 0 25 h −1 ). For ATO decorated with CuNPs, no enhancement in the photodegradation rate was observed. It can be attributed to a partial oxidation of copper nanoparticles during the photocatalytic process. What is more, the percentages of dye loss (DEG%) obtained after 6 hours of irradiation was calculated and gathered in Table 2. As it can be seen, the most effective photocatalytic behavior was observed for ATO decorated with gold nanoparticles (about 84% of MR decomposition), and the second promising modification of the ATO surface was deposition of silver nanoparticles (DEG% = 79%).
The synthesized Au-ATO photocatalysts have been used multiple times in order to determine the stability of such materials. As can be seen in Figure 7(a), the decolorization rate of methyl red is similar when using the same sample four times. The average MR degradation percent was determined to be 84 ± 2% as shown in Figure 7(b). Both gold and silver nanoparticles are known for their antibacterial properties [54]. Covering TiO 2 nanotubes with noble metal particles can be a great advantage in the context of the use of such materials for water purification. The presence of Au and Ag may prevent the growth of microorganisms near the photocatalyst, thus allowing its permanent exposure to solar radiation.

Conclusions
In this study, the effect of the ATO surface decorated with Cu, Au, and Ag nanoparticles on photoelectrochemical and photocatalytic properties were investigated. Nanoparticles were synthesized by a chemical method, and the presence of metals on TiO 2 was confirmed by EDS analysis. All obtained materials were used as photoanodes in photoelectrochemical water splitting and as photocatalysts in a methyl red decomposition process. It was found that in studied conditions the Schottky barrier is formed between the metal/ATO interface, which leads to enhancing photocurrent generation in the range of 300-400 nm. However, in the visible light spectrum (400-600 nm) no additional effect from SPR was observed. Nevertheless, deposition of noble metal nanoparticles on the ATO surface resulted in improved efficiencies in both studied processes. It was shown that deposition of gold nanoparticles resulted in high photocurrent densities (100 μA cm −2 for 350 nm and 1 V) with a high IPCE value of~45%. The best photocatalytic properties was observed for ATO samples with AuNPs (DEG% = 84%, k = 0 30 h −1 ) and AgNPs (DEG% = 79%, k = 0 25 h −1 ).

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Additional Points
Highlights. (i) Nanostructured TiO 2 (ATO) layers were obtained by electrochemical oxidation. (ii) Copper, gold, and silver nanoparticles were deposited on ATO surface. (iii) The effect of ATO surface modification was studied. (iv) Photoelectrochemical performance of obtained materials was investigated. (v) Photodegradation of methyl red in the presence of modified ATO was performed.