Deposition of MoO3 with (011) and (102) planes exposed
First, the MoO3 material was deposited at 450°C in O2 atmosphere (0.5 mbar) for 120 minutes. The resulting MoO3 film uniformly covered the entire surface of the FTO substrate. The SEM micrograph of the deposited film is presented in Fig. 2a. The polycrystalline film consisted of thin plates with sizes up to 500 nm. The crystallites appeared to be predominantly oriented perpendicular to the substrate, although a mixture of different orientations was observed. The XRD pattern showed that the most intense reflexes originated from the (102) and (011) planes, while the (002), (004), and (006) planes were also clearly visible (see Fig. 2b, PLD_450°C(O2)). Additionally, deposition was performed at different temperatures, and deposition at 400°C also resulted in a random orientation of crystallites, while deposition at 500°C did not lead to the formation of a distinct layer (see Fig. S1a and 1b, respectively). In order to achieve a more ordered arrangement of α-MoO3 crystallites, slight modifications were made to the deposition parameters. The deposition process commenced at room temperature in an argon atmosphere. After 10 minutes, oxygen was introduced into the chamber, and the temperature was raised to 450°C (sample labeled as PLD_450°C(Ar,O2)). The SEM image of resulting layer is presented in Fig. 2c. The morphology of the layer was relatively similar, consisting of crystallites with a similar appearance. However, upon closer examination, it was observed that the majority of crystallites were perpendicular to the FTO substrate. Moreover, the intensities of the reflections from the (00k) planes were significantly diminished, and the peaks appear broader in comparison with the conventionally deposited film pattern (Fig. 2b). Based on this results, it can be concluded that the exposed facets observed in the SEM image were (011) and (102), while the (00k) planes were oriented perpendicular to the substrate. Considering this findings, it can be inferred that vdW gaps are also perpendicular to FTO and the direction of crystallites growth is related to the initial stage of the process when metallic Mo is deposited in argon atmosphere. To validate this hypothesis, a 20 nm layer of metallic Mo was deposited on FTO using a magnetron sputtering machine, and then the FTO/Mo was used as a substrate for MoO3 deposition using PLD (O2 atmosphere, 120 min, 450°C). The resulting layer exhibited a similar morphology and exposed (102) and (011) facets (SEM micrograph is shown in Fig. S1c). Thus, the presence of a thin Mo film during the initial stage of MoO3 deposition is crucial for obtaining a film with exposed (011) and (102) facets.
Deposition of MoO3 with (00k) planes exposed
The same PLD system, substrates, and target were used to obtain a MoO3 film with differently oriented crystallites and exposed (00k) planes. The deposition process was performed at room temperature for the same duration (120 minutes) and under the same oxygen pressure (0.5 mbar). The SEM image of the as-deposited film is presented in Fig. 3a, where no distinguishable crystallites can be observed, indicating an amorphous nature of as-deposited film. To crystallize the layer, the sample was annealed in an air atmosphere. However, conventional annealing at 450°C for 4 hours with a slow heating rate resulted in the formation of relatively large crystallites that poorly covered the substrate and exhibited completely random arrangement, see Fig. S2a. Nevertheless, through optimization of the crystallization procedure, it was possible to obtain a film with exposed (00k) facets. The amorphous films were directly placed in a heated oven and annealed at 575°C for varying periods of time (0–20 minutes). The selected temperature was limited by thermal stability of the FTO substrates (lower temperatures did not yield the expected effect). The SEM images of the post-annealed films are presented in Fig. 3b-f. The morphology of the films underwent significant changes with increasing annealing time. After 0.5 and 1 minute of heating, there can be found epitaxial-like areas. Prolonging the annealing time resulted in fully polycrystalline layers, with larger crystallites formed as the heating duration increased. Notably, after 20 minutes, the crystal size expanded to several micrometers, causing partial exposure of the FTO substrate surface, as it is shown in Fig. S2b. In the studied case, where the aim is to obtain a film of deposited material, the observed effect of patchy layer formation is unfavorable. Extending the annealing time to 60 minutes at a temperature of 575°C revealed an intriguing phenomenon concerning the behavior of MoO3. It appeared that the material underwent sublimation from the FTO substrate, as evidenced by the absence of MoO3 on part of the FTO surface (Fig. S3a). This sublimation phenomenon was unexpected since it is generally claimed that the sublimation temperature of MoO3 is higher than 780°C 26. However, it has been reported that in the case of the MoO3 nanoplates, the sublimation occurred even at prolonged heating at about 400°C 27. This sublimation phenomenon highlights the sensitivity of MoO3 in a form of thin films to elevated temperatures and emphasizes the importance of carefully controlling the annealing conditions to achieve the desired layer deposition and maintain film integrity. Based on these findings, a post-annealing duration of 5 minutes was determined to be optimal for the sample, labeled as PLD_RT(575°C,5). A digital photo of the sample is shown in Fig. S3b for comparison with the sample annealed for 60 minutes. The morphology of the films presented in Fig. 3 differs significantly from the morphology of MoO3 deposited at 450°C. In contrast to the crystallites deposited at high temperatures, the crystallites in post-annealed samples after room temperature deposition appeared to be oriented parallel to the FTO substrate. The XRD patterns of the films after annealing at 575°C are shown in Fig. 3g. As expected, the as-deposited layer was amorphous, with only reflexes originating from the FTO substrate. After 30 seconds of heating, low-intensity reflections characteristic of α-MoO3 appeared and as the heating time increased, the intensity of these signals grew, indicating a more crystalline material. Moreover, after 5 minutes of annealing, the intensity of the MoO3 peaks exceeded that of the peaks originating from the FTO substrate. Notably, the intensities of the peaks did not increase uniformly as the reflexes originating from the (002), (004), and (006) planes exhibited much higher intensity compared to the (011) plane, while the peak from (102) plane nearly disappeared. On the basis of XRD results, it can be inferred that the majority of the crystallites is arranged in such a way that the (00k) planes are exposed. Considering the crystal structure of α-MoO3 (see Fig. 1a), the vdW gaps for this type of samples were parallel to the substrate. The post-annealing of amorphous MoO3 was also performed at lower temperatures (400 and 450°C) using the same method of directly placing the samples in the hot oven for 60 minutes. The obtained films were polycrystalline and uniformly covered the substrate (no sublimation of MoO3 was observed), see Fig. S4a and S4b. However, the XRD results demonstrated that exposure of the (00k) plane was not achieved at this temperature (Fig. S4c). Therefore, to obtain a MoO3 film with the (00k) planes exposed, the “rapid” annealing at 575°C for a short period of time was required. Two samples, labeled as PLD_450°C(Ar,O2) and PLD_RT(575°C,5), were selected for further comparison of their properties using electrochemical methods
Electrochemical properties
Both types of samples were subjected to investigation using electrochemical methods in a 1 M AlCl3 aqueous electrolyte. Previous studies have already reported the satisfactory electrochemical properties of MoO3 in such an electrolyte 28. First, cyclic voltammetry curves were recorded to compare the behavior of the PLD_450°C(Ar,O2) and PLD_RT(575°C,5) electrode materials, as shown in Fig. 4a and 4b. Generally, during the first scan, both samples exhibited cathodic peaks occurring at the same potentials. Notably, the current density for the PLD_450°C(Ar,O2) sample was higher than that of the PLD_RT(575°C,5) one. These cathodic peaks correspond to the reduction of Mo(VI) centers and simultaneous cation insertion. The complexity of the AlCl3 aqueous solution chemistry, including hydrolysis, makes it unclear which form of ion is being intercalated. However, the ex-situ EDX measurements confirms the presence of Al in the sample after cathodic polarization (-0.1 V vs Ag/AgCl (3 M KCl) as it is shown in Fig.S5, indicating that the observed electroactivity is related to the insertion of Al-containing cations into the MoO3 structure. The presence of chlorides was also detected, suggesting that the electrolyte was simply adsorbed on the electrode surface, however, the excess of Al (considering the stoichiometry of AlCl3) clearly indicates the incorporation of Al-containing ions into the structure. Regarding the shape of the cyclic voltammetry curves, both samples exhibited clear irreversibility of the electrochemical processes. MoO3 is sensitive to the potential range during polarization 29, indicating that cations are likely irreversibly intercalated into the electrode material structure. The cathodic peak observed during the 1st scan at around E = 0.05 V vs. Ag/AgCl (3 M KCl) disappeared and was not seen during the 2nd scan. Similar behavior has already been reported for MoO3-based electrodes tested in Mg2+-containing electrolytes 30. In the case of the sample with (011) and (102) facets exposed, the electrochemical activity related to cation intercalation/deintercalation was clearly observed also during 2nd scan, as well. The peaks registered for the electrode material with (00k) planes exposed almost disappeared, but a similar shape of cyclic voltammogram has been reported before 31. This suggests that the PLD_450°C(Ar,O2) sample, with vdW gaps perpendicular to the substrate, is a more suitable electrode material for energy storage due to the occurrence of intercalation/deintercalation processes during subsequent scans. However, as shown in the CV curves, the PLD_RT(575°C,5) sample exhibited a significantly lower overpotential for the hydrogen evolution reaction (HER). The Tafel parameters determined based on polarization curves are presented in Fig. 4c. The results prove that the MoO3 sample with exposed (00k) planes was characterized by higher electrocatalytic activity and faster kinetics in promoting the HER. The electrocatalytic properties of MoO3 towards HER have already been reported 32. Moreover, it has been presented that MoO3 in the form of nanobelts, with (00k) planes exposed, exhibits superior properties for H2 evolution compared to commercially available MoO333.
Additionally, electrochemical impedance spectroscopy measurements were performed and the comparison of spectra recorded during cathodic polarization (E = -0.1 V vs Ag/AgCl (3 M KCl)) is shown in Fig. 5a. The measurements were performed at the potential that Al-containing ions are intercalated to the electrode material. As expected, due to the mechanism of charge storage in MoO334, there is a range of frequencies where the spectra exhibit diffusional behavior. Thus, the Warburg coefficients were estimated as the slope of Z = f(ω−0.5) function, as shown in Fig. 5b. This analysis suggests that apparent diffusion coefficient of cations (it is assumed that it is Al3+ diffusion) is higher for the PLD_450°C(Ar,O2) electrode material compared to the PLD_RT(575°C,5) one. In order to compare the electrical properties of the electrodes, an electric equivalent circuit was proposed, as shown in the inset of Fig. 5a. The simple model consist of 4 elements: R1 – electrolyte resistance, R2 – charge transfer resistance on the electrode/electrolyte interface, CPE1 – capacitive properties of the electrode material, and W1 – Warburg element (diffusional processes). The goodness of fitting of about 3–5·10− 5 was achieved and the results are shown in Fig. 5c. The values of R1 are comparable due to the electrochemical setup, which ensures constant distances between the electrodes. However, a significant difference was observed for the R2 values, with 261 and 3 Ωcm2 for the PLD_RT(575°C,5) and the PLD_450°C(Ar,O2) electrode, respectively. Assuming that this value is related to cation insertion from the electrolyte to the electrode material, results indicate that exposing the (001) and (102) planes to the electrolyte facilitates the intercalation phenomenon. The values of the constant phase element parameters are related to the capacitive properties of the electrode materials. The “n” values close to 0.7 suggest that it is not “pure” double-layer capacitance, but rather a more complex modeled process 35, thus the electric equivalent circuit used is simplified. Nevertheless, the higher “P” value indicated a higher capacitance of the PLD_450°C(Ar,O2) material. The values of the Warburg coefficients obtained from modeling are not exactly the same as those from the analysis shown in Fig. 5b, however, the trend remains consistent. The apparent diffusion coefficient of Al3+ ions was estimated using a formula derived by rearranging the definition of the Warburg coefficient, assuming that the diffusion coefficients of intercalation and deintercalation are the same:
$${D}_{{Al}^{3+}}={\left(\frac{RT}{\sqrt{2}{n}^{2}\sigma c{F}^{2}}\right)}^{2}$$
1
where R – gas constant [J·K− 1·mol− 1], T – temperature [K], n – charge of ion, σ – Warburg coefficient [Ωcm2s− 0.5], c – concentration of ions [mol·cm− 3] (it is assumed that c = 1 M), F - Faraday constant [C·mol− 1].
The apparent diffusion coefficient of Al3+ ions estimated using the Warburg coefficient from modeling was found to be 1.9·10− 14 and 3.9·10− 15 cm2s− 1 for PLD_450°C(Ar,O2) and PLD_RT(575°C,5), respectively. Notably, different diffusion coefficients were determined for electrodeposited MoO3, depending on the size and charge of the cation; for instance, DNa+ was higher than DAl3+36. Moreover, the values of DAl3+ are in good agreement with values presented here. In this work, the diffusion coefficients of the same cation were compared for polycrystalline films with differently oriented crystallites. Moreover, it can be expected that intercalated ions are mobile in vdW gaps. According to the results, in the case of the polycrystalline film with vdW gaps perpendicular to the substrate, diffusion is facilitated, resulting in a higher diffusion coefficient compared to the material with vdW gaps parallel to the substrate. Furthermore, the values of charge transfer resistance suggest that the intercalation phenomenon occurs on the (011) and (102) facets, while the mobility of the cation through the (00k) plane is hindered. On the other hand, the electrode material with exposed (00k) planes exhibits enhanced electrocatalytic properties towards the hydrogen evolution reaction. The presented conclusions are schematically shown in Fig. 6.