A solution (hereinafter named solution S₂) containing 0.075 M of molybdenum trioxide (MoO₃, 99.97% purity, Sigma Aldrich), 0.075 M of nickel (NiSO₄·6H₂O, 99.00% purity, Merck), 0.50 M of sodium citrate (Na₃C₆H₅O₇·2H₂O, 99.00% purity, Merck) and 0.70 M of ammonium hydroxide (NH₄OH, 25% NH₃ in H₂O, Merck) was used as an electrolyte for the electrodeposition of thin (nanometric) films of molybdenum oxide (s) doped with nickel on FTO-coated glass.

Electrodeposition experiments done with a similar electrolyte solution but without nickel (hereinafter named solution S₁) are also discussed in this research to compare the microstructural and compositional characteristics and the photoelectrochemical properties of MoO_x and MoO_x-Ni films. All solutions were prepared using deionized (DI) water. The initial pH of the solutions was adjusted to 9.0 using sulfuric acid (97% purity, Merck), to favor the presence of molybdenum in the form of molybdate ( MoO4(aq.)2-) and nickel in the form of the dimeric nickel-citrate complex [Ni2(OH)2H2cit2](aq.)4- during electrodeposition.

FTO-coated glass slides (TEC 8, 3.0 mm thickness, Sigma-Aldrich) were cut into pieces of 1.0 cm width and 2.5 cm length to be used as working electrodes. Copper foil and silver paste were used to fabricate the electrical connections of the FTO-coated glass pieces. The electrodeposition active area of the FTO-coated glass electrodes was a rectangle of 1 cm in width and 1.5 cm in length. A platinum foil (1.25 cm²) was used as a counter electrode and an Ag/AgCl (3 M KCl) electrode as reference. The three electrodes were mounted in a glass-jacketed cell and the temperature was controlled at 25°C. A cathodic potential of −1.377 V vs Ag/AgCl (3M) was applied for 3 hours to the working electrode (using a Gamry Reference 3000 potentiostat/galvanostat/ZRA) to promote electrodeposition of molybdenum oxide (s) and nickel metal/hydride, considering that the reduction of molybdate ions occurs solely at high overpotentials.

The microstructure, chemical composition, and UV visible light absorbance of the deposits obtained were characterized by SEM, EDS, XPS, AFM, and UV-visible spectroscopy.

A JEOL JSM-IT300 SEM coupled to an Oxford Instruments x-act Penta EDS device was used to perform the microstructural and elemental analysis, with a penetration depth of 2 m at different magnifications.

A Physical Electronics 1257 XPS system was used to corroborate the elemental composition and oxidation states of the elements present in the films deposited on the photoanode surfaces. Its surface analysis chamber is equipped with radiation (non-monochromatic) at 1.4866 kV from an aluminum K_α source. The electrons are captured at 35 to the normal’s sample. Spectra were acquired in the region of interest using the 0 to 1200 eV experimental range of binding energy, areas where it is expected to find signals of the elements: Mo, Ni, O, Sn, F, and C.

The surface morphology and thickness of the films were examined using an atomic force microscope model NaioAFM (Nanosurf GmbH, Langen, Germany). The measurements were performed in the contact mode and with a maximum scan field area of 50 × 50 microns.

Finally, the UV-visible light absorbances (in the range between 300 and 800 nm) of the films produced were recorded using an Agilent 8553 UV-visible spectrometer with a tungsten lamp. The absorption coefficient and band-gap of the films produced were determined by fitting the data of the UV-visible spectra obtained using Tauc plots.

The performance of the photoanodes S₁ and S₂ was tested in a photoelectrochemical cell of polytetrafluoroethylene (PTFE), which had a quartz crystal window to let light into the cell. Photoelectrochemical measurements were performed using white (Chanson, 30 W, 6000-6500 K color temperature) and UV (Chanson, 20 W, λ_max = 385 nm) LED lights coupled to a power source with adjustable voltage/current output, and a Gamry Reference 3000 potentiostat/galvanostat/ZRA.

The photoanodes were used as working electrodes, after cleaning with acetone and distilled water in an ultrasonic bath for 5 minutes and drying in air. A platinum foil (1.25 cm²) was used as a counter electrode and an Ag/AgCl (3 M KCl) electrode was used as a reference. A solution of 0.1 M sodium sulfate (Na₂SO₄, 99.00% purity, Sigma Aldrich) and pH equal to 3.5 adjusted using sulfuric acid (H₂SO₄, 98% purity, Merck) was used as electrolyte.

These three electrodes were mounted in the PTFE cell, one of them, with the photoanode placed in front of the quartz crystal window of the cell. In this setup, the light strikes perpendicularly to both the quartz glass window and the photoanode surface placed at 4 cm distance from the light source, with the MoO_x-Ni (MoO_x) film facing the light. The white light intensity was regulated to reproduce the illuminance level of full daylight (10 Klux), assuming neglectable reaction losses. A digital light lux-meter (Outest, GM1010) was used for this purpose. Likewise, the UV light intensity was regulated using a digital radiometer (Tenmars, TM-213) to provide an irradiance of 2 mW cm⁻² at a distance of 4 cm from the light source.

Linear voltammetries under dark, UV, and white lights at a scan rate of 25 mV s⁻¹ and between 0 and 1.6 V vs Ag/AgCl (3M KCl) were carried out to study the photoactivity of the photoanodes.

Fig. 2 depicts the curves of current densities as a function of time obtained in the electrodeposition experiments at a potential of −1.377 V vs. Ag/AgCl1) (3M) applied for 3 hours for the synthesis of photoanodes S₁ and S₂.

Fig. 2a shows that the cathodic current density curve obtained in the synthesis of photoanode S₂ is larger than the cathodic current density curve obtained in the synthesis of photoanode S₁ under the same experimental conditions. A zoom at the beginning of the current density curve of the photoanode S₁ (Fig. 2b) reveals a small decrease in cathodic current density (< 2 min). This initial shape of the current density would be related to the formation of molybdenum oxides of the MoO₂ type, which are probably partially blocking the electrode surface. The above would be caused because the high electronegativity of oxygen compared to hydrogen at room temperature (even in environments where there is a greater amount of hydrogen) would give greater stability to binary d-metal oxides, such as MoO₂ instead of metal forms with lower states of oxidation as reported by Borgschulte et al. [25]. Studies carried out by Shembel et al. [26] confirm that at potentials lower than −0.68 V vs SHE the most stable species due to the reduction of molybdate ions, MoO₄²⁻ is molybdenum dioxide, MoO₂. This justifies the formation of molybdenum oxides instead of metallic molybdenum, even in the presence of hydrogen generated during HER.

Despite the current density curve obtained throughout the experiment would be a product of both the hydrogen evolution reaction and the reduction of molybdate ions to molybdenum dioxide, the steady-state value of the current density illustrated in Fig. 2a would correspond mainly to the HER as reported by Kuznetsov et al. [27-29]. Probably the high electrical conductivity2) of the MoO₂ deposited on FTO-coated glass would have a positive impact on the HER since it would allow the passage of electrons easily from the substrate to the electrode | electrolyte interface promoting HER at the cathode surface.

Considering the above, the possible reactions that describe the reduction of molybdate ions to molybdenum oxide and the evolution of hydrogen during electrodeposition are the following:

On the other hand, Fig. 2c shows that for the photoanode S₂ the current density has a concave shape during the first 70 minutes; this behavior could be attributed to two main factors:

The initial current density curve shape could be related to the partial blockage of the photoanode surface by an intermediate complex of the type [Nicit-MoO₂]_ads which would be formed under alkaline conditions at the working potential, as reported by Kuznetsov et al. [29]. The possible formation of the intermediate complex mentioned above would be given by the following [29]:

Afterward, the complex [NicitMoO₂]_ads gives rise to the formation of [Ni2(OH)2H2cit2](aq.)4- and MoO_2(s) as shown in Eq. (6). It is possible to assume that due to the complexing ability of the citrate, the adsorbed species would slowly give way to the MoO_2(s)-Ni₂H_(s) alloy formation within the first 70 minutes as illustrated in Fig. 2c (Eqs. (6,7)).

Complex [Ni2(OH)2H2cit2](aq.)4- would form Ni₂H_(s) under the experimental conditions of electrodeposition, as shown in Eq. (7).

In the next 110 minutes, the current density value continued to increase reaching approximately −29 mA cm⁻² at 180 minutes as illustrated in Fig. 2a. This current density had a more cathodic value than that presented by photoanode S₁ during the same period. This behavior would be linked to the cathodic surface type. Chassaing [31] has reported that the nickel-molybdenum alloys are a good catalyst for the HER under similar experimental conditions (pH 9.5 and polarization range: from −1.3 to −1.6 V vs SSE) to those used in chronoamperometries made in this research. The above would explain the cathode current density increase in the last 110 minutes. This suggests that the MoO_2(s)-Ni₂H_(s) alloy on the photoanode S₂ surface would exhibit better electrical conductivity than the undoped molybdenum dioxide deposited on the photoanode S₁ surface, promoting the HER on the cathode surface with ease, as shown in Fig. 2a.

In order to characterize the morphology of the undoped and nickel-doped molybdenum oxide films deposited on FTO-coated glass, the photoanodes S₁ and S₂ were analyzed by SEM and AFM.

The surface morphology of the photoanodes was observed through SEM and the micrographs are presented in Fig. 3. According to this figure, SEM analysis revealed the presence of several cracks on the surface of the photoanodes S₁ and S₂. The loss of water from the deposited films on FTO-coated glass and the HER during the electrodeposition process is probably the main cause of the formation of cracks in these electrodes.

An SEM image of the photoanode S₁ surface is shown in Fig. 3a. Unlike the photoanode S₂ (Fig. 3b), it does not show the presence of irregularly shaped conglomerates. Indeed, it forms an orderly and homogeneous pattern over the entire surface of the photoanode S₁. This comparison reveals the advantage of the MoO_2(s)-Ni₂H_(s) alloy in the photoanode S₂, since the conglomerates of irregular shape in this type of film would favor the PEC oxygen evolution reaction, due to the greater surface area and availability of active sites on it.

An AFM image of a nickel-doped molybdenum oxide film of the photoanode S₂ is shown in Fig. 4b. According to this figure, MoO_2(s)-Ni₂H_(s) films on the photoanode S₂ surface exhibit an irregular morphology. Conglomerates and randomly located clusters are clearly observed on the film surface, which agrees with Fig. 3b obtained by SEM analysis of the photoanode S₂. A surface examination revealed that these hill-like structures have an average width of 182 nm. The height of these conglomerates varied from 250 to 1,159 nm. Moreover, Fig. 4b shows that the film surface does not follow the pattern of the FTO-coated glass. In contrast to this, Fig. 4a corroborates that the photoanode S₁ presents a film deposited without the presence of irregularly shaped conglomerates. The AFM image of the undoped molybdenum oxide film on the surface of the photoanode S₁ presented in Fig. 4a, indicates the presence of a homogeneous film on the FTO-coated glass surface. Previously, thicknesses ranging between ca. 219 nm [15] and 1150 nm [13] have been reported for molybdenum oxide films obtained by electrodeposition. In this research, the thicknesses determined by AFM showed that the photoanode S₂ presents a film3) on the substrate with an average thickness of 763 nm, while the photoanode S₁ has a deposited film with an average thickness of 438 nm on FTO-coated glass, which is thinner than the film obtained on the substrate of the electrode S₂. Both values are within the ranges reported in the literature [13,15]. For the photoanode S_2, it is possible to assume that the nickel presence could modify the morphology and the thickness of the generated films. Indeed, Popczyk et al.[32] has reported that nickel-based coatings are among the most often used to obtain electrodes with a very developed, rough, thick, or porous electrode surface. This is due to the ease with which Ni²⁺ penetrates within interstitial spaces and besides the reaction of Ni with the basal (molybdenum) or lateral (oxygen) atoms of the metal oxide structure, modifying the morphology and properties of the semiconductor material [33].

It is important to mention that to determine the thickness of the deposits in the AFM analysis, the step formed between the clean FTO-coated glass surface and the deposit was considered. The above is illustrated in Fig. 5.

The composition and chemical state of the elements forming the films deposited on the surface of the photoanodes were characterized by EDS and XPS.

Figs. 6a and 6b show the EDS analyses for selected points (same micro-area used in the SEM analysis presented in Fig. 3) of the photoanodes S₁ and S₂.

The EDS spectra of the photoanodes S₁ and S₂ corroborate a strong presence of molybdenum (over 40 wt.% in both cases) in all the deposited films, which would be in the photoanodes mainly as MoO₂ (as suggested by the XPS analysis) and MoO₃ (formed from the partial oxidation of MoO₂ in contact with the atmosphere).

The presence of oxygen at the photoanodes is also shown in the EDS spectra. The high amount of oxygen (over 55 wt.% in both cases) corroborates the presence of MoO₂/MoO₃.

For the case of the photoanode S₂, as can be seen in Fig. 6b, it is possible to find nickel on the film surface of the photoanode S₂ (1.1 wt.%), which would be mainly as Ni₂H obtained from the electrodeposition and NiOOH obtained from the partial oxidation of nickel hydride in contact with the electrolyte during the electrodeposition.

Figs. 7a and 7b show the Mo 3d XPS spectra for the photoanodes S₁ and S₂. As can be seen, the deposited films on the photoanodes contain the characteristic peaks of Mo⁴⁺ situated approximately at binding energies of 230 and 232 eV, attributed to the presence of MoO₂ [34], which agrees with the reaction in Eq. (4) for both types of electrodes in section 3.1. However, at the same time, it is possible to see in all the films two characteristic peaks of Mo⁶⁺ situated approximately at binding energies of 232 and 235 eV, which would indicate the presence of MoO₃ as well [34]. The presence of this type of oxide in the photoanodes would be explained by the exposure of the photoanodes to the atmosphere before the XPS analysis, which would lead to the surface oxidation of deposited films. Consistent with this research, Scanlon et al.[33] reported that for the peak Mo 3d, samples of MoO₂ seemed to have an MoO₃ surface layer when these samples were exposed to the atmosphere before XPS analysis and, therefore, it is likely that the data included components associated with changes in the sample induced by exposure to oxygen.

The O1s spectra for the films of the photoanodes S₁ and S₂ are shown in Figs. 7c and 7d. In all the films, it is possible to appreciate two peaks situated approximately at binding energies at 531 and 532 eV. In both photoanodes, the peaks at lower binding energies mainly correspond to metal oxides such as MoO₂ and/or MoO₃, whereas the peaks at higher binding energies correspond to hydroxyl groups or adsorbed water on the surface i.e. Mo-OH [35,36].

Fig. 8 illustrates the XPS spectra in the region of Ni 2p for the photoanode S₂. For this sample, the Ni 2p exhibits a doublet Ni 2p_1/2 and Ni 2p_3/2. Analysis of the Ni 2p_3/2 region shows the main component centered at 858 eV for the peak Ni³⁺. This peak is assignable to nickel oxyhydroxide (NiOOH) [37,38], which could be formed from the partial oxidation of the nickel hydride formed in the film (see section 3.1) due to the presence of water in the electrolyte. The above can be described by the following equation:

An additional peak at higher binding energy (centered at 861 eV) can be seen in Fig. 8, which could correspond to a satellite structure4) [37,38]. It is important to mention that the presence of satellite structures for nickel compounds with 0, +2, and +3 oxidation states is common at higher binding energies [37,38]. In fact, in the case of the photoanode S₂, the correct assignment can be compromised by the complex mainline splitting due to multip contributions in oxyhydroxides and by satellite structures at higher binding energies, particularly in the most intense Ni 2p spectra where often the satellite structures are wide.

To have a complete optical characterization, analysis by UV-Visible spectroscopy and Tauc plots for the photoanodes S₁ and S₂ were performed.

UV-Visible spectroscopy analysis was used to determine the absorbance of the photoanodes S₁ and S₂. Fig. 9 shows the absorption spectra of these photoanodes.

According to Fig. 9, the photoanodes S₁ and S₂ formed a similar pattern in the w range used in UV-Visible analysis. As can be seen, in these photoanodes, the absorption peak is reached at wavelengths below 400 nm. It is evident that the photoanode S₂ had a greater absorption of light energy in the whole range of analyzed wavelengths, reaching its maximum in the UV region (the maximum absorbance of this electrode was ca. 3.1 at λ = 330 nm). As mentioned in section 3.2, this behavior can be attributed to the film morphology on the photoanode S₂ surface. The presence of irregularly shaped conglomerates on the photoanode surface promotes large surface areas favoring the absorption of light energy [14].

In solar energy conversion systems, the optical band-gap energy (Eg) of a semiconductor electrode is a key parameter. Only photons of equal or higher energy than the band-gap energy will be absorbed by the semiconductor film. The optical band-gap value of the deposited films was determined by the analysis of Tauc plots.

Tauc plots of photoanodes S₁ and S₂ are depicted in Figs. 10a and 10b. According to Fig. 10a, the photoanode S₁ has a higher band-gap (ca. 3.15 eV) than the values obtained indirect transitions in other research [13,15] for electrodeposited molybdenum oxides in similar conditions to this research, which implies that this type of photoanode could present a greater resistance to the passage of electrons from the valence band to the conduction band of the molybdenum oxide deposited on the FTO-coated glass. Semiconductor materials’ large band-gap as in the case of the photoanode S₁ results in a reduced use of the solar spectrum [39]. On the other hand, the photoanode S₂ (Fig. 10b) exhibits a band energy value of 2.58 eV, which could be attributed to the incorporation of Ni³⁺ in the MoO₂/MoO₃ structures, contributing to the formation of a shallow donor energy level below of the conduction band of the molybdenum oxide.

Representative linear sweep voltammograms of the photoanodes S₁ and S₂ obtained in an electrolyte of Na₂SO₄ 0.1 M at pH 3.5 are presented in Figs. 11a and 11b. Additionally, in order to obtain a better understanding of the phenomena in the system in which the photoelectrochemical experiments were carried out, Eh-pH for the nickel and molybdenum species have been elaborated. Figs. 12a and 12b show the obtained which will be used to interpret and describe the processes that occur during the LSVs.

The analysis of the LSV (Fig. 11b) for the photoanode S₂ shows two peaks centered at 0.2 V (ca. 0.4 V vs SHE) and 0.75 V (ca. 0.95 V vs SHE) when the photoanode is illuminated (white light and UV light). These peaks could be associated to the redox transition of Mo⁴⁺ to Mo⁶⁺. Likely, in the acidic electrolyte used in this voltammetry, the formation of H₂MoO_4(aq.) from the MoO_2(s) present in the photoanode surface begins at this potential range according to the following reaction mechanism:

The above agrees with the information provided by the Eh-Ph diagram in Fig. 12b, where it is possible to appreciate that at the indicated pH and potential range, the formation of H₂MoO_4(aq.) from the MoO_2(s) is favored. On the other hand, the second peak also could be caused by the formation of NiO₂(H₂O)_(s) from NiOOH_(s) present in the film surface, according to the Eh-pH diagram in Fig. 12a. The formation of NiO₂(H₂O)_(s) is given by the following reaction:

As can be seen in the voltammogram in Fig. 11b, when the photoanode is illuminated under UV and white light there is an increase in the current density curves from ca. 1.3 V (ca. 1.5 V vs SHE) that would correspond mainly to the PEC oxygen evolution reaction present at the anode surface, which is to be expected in that potential zone. On the other hand, in the LSV under dark polarization conditions of the photoanode S₂, the presence of a small peak around 0.2 V can be observed. This peak shows that the MoO_2(s) present in the deposited film could be giving way to H₂MoO_4(aq.) formation under the experimental conditions, which would indicate a possible instability of the deposited film. The current increase at 1.4 V under darkness is related to the electrochemical oxidation of water molecules.

About the photoanode S₁, it is possible to see that under ultraviolet illumination the current density curve (Fig. 11a) presents similar peaks to the voltammograms obtained for the photoanode S₂ under light conditions. This suggests that the formation of H₂MoO_4(aq.) is present during the study performed by voltammetry in this electrode. When the voltammetries are performed with white light or in dark conditions, the presence of a small peak around 0.5 V (ca. 0.8 V vs SHE) can be observed. This peak indicates that the MoO_2(s) present in the deposited film may not be fully stable. In the case of this photoanode, an increase in the current density curve under UV illumination can also be observed from ca. 1.4 V (ca. 1.6 V vs SHE) which correspond to the PEC oxygen evolution reaction present at the anode surface.