Synthesis of VO_x powders and solution

Precursor solutions for VO_x powders were prepared by using 0.1 M solution of vanadium (IV) oxide sulfate hydrate (VOSO₄·xH₂O) and 1 M solution of sodium hydroxide (NaOH), as previously suggested [25]. These two precursor solutions were mixed, resulting in a blue color solution with a pH ~2 tested with a MColorHast^TM. After that, the blue solution was heated at 80 °C for 3 h to obtain VO_x precipitates. These precipitates were dispersed in methanol to dissolve possible impurities and then centrifuged. After centrifugation, VO_x precipitates were dried at 80 °C for 2 h to obtain VO_x powders.

To prepare a VO_x sol-gel solution, VO_x powders were dispersed in isopropyl alcohol under ultrasonic vibration at room temperature for 10 min. Then, the mixture solution was filtered with a cotton filter to remove the undispersed precipitates and obtain a transparent solution for thin film deposition by spin-coating. Pure VOx films were called samples S1.

Preparation of VO_x:Co solutions

To obtain cobalt doped VO_x precursor solutions, dried VO_x powders and CoCl₂·6H₂O salt were dispersed/dissolved in isopropyl alcohol. Then, the solutions were subjected to ultrasonic vibration at room temperature for 10 min. Finally, the solutions were filtered with a cotton filter to remove the undispersed precipitates and obtain transparent VO_x:Co solutions for thin film deposition by spin-coating. Four VO_x:Co solutions were prepared with VO_x/CoCl₂⋅6H₂O weight ratios of 0.7/0.3, 0.5/0.5, 0.3/0.7 and 0.1/0.9, called samples S ₂ , S ₃ , S ₄ and S ₅ , respectively.

VO_x and VO_x:Co thin film deposition

The five precursor solutions, VO_x and VO_x/Co, were used to form thin films by spin-coating. Microscope glass slides from VE-P10 Velab™, to be used as substrates for vanadium oxide solutions, were washed with Alconox® detergent, rinsed with distilled water and air-dried. The clean substrates were placed in the Chemat Thecnology Spin-Coated KW-4A, 50 µL of the VO_x or VO_x:Co solutions were deposited at 2500 rpm for 2 min on the glass substrates. The obtained coatings were thermally annealed in air at 200 °C for 1 h and became solid thin films.

Thin films characterizations

X-ray photoelectron spectroscopy (XPS) was used to assess the chemical composition of the films. The tool employed was PHI 5000 Versa Probe II, with a monochromatic Al Kα (1486.6 eV) X-ray source with a 0.1 eV step size and a constant pass energy of 23.50 eV. The scale of binding energies was corrected by calibrating the main C 1s peak at a binding energy of 284.8 eV. Atomic compositions of VO_x/Co films were estimated from the XPS results. Crystalline structures of VO_x/Co films were probed with X-ray diffraction (XRD) employing a Rigaku DMAX-2200 diffractometer equipped with a Cu Kα (λ = 1.54 Å) X-ray source. A Jeol JSM-7800F field emission scanning electron microscope (FE-SEM) was used to evaluate the surface morphology of the films. The surface morphology of each film sample was analyzed by atomic force microscopy (AFM) with a Veeco Dimension Icon, Bruker instrument. Optical transmittance spectra of VO_x/Co films were measured with a Perkin Elmer Lambda 20 UV-visible spectrometer. Thicknesses of VO_x and VO_x:Co thin films were measured with an Ambios Technology XP 200 profilometer.

Profilometry and main characteristics

The first assessment of the possible formation of continuous VO_x and VO_x:Co thin films from the above mentioned precursor solutions is film thickness measurement by profilometry. Table 1 lists the five types of VO_x:Co thin films with their respective thickness. In the Alpha-Step profiles of all film samples, large grains of 100 to 400 nm of height were observed and were embedded in continuous solid coatings with thickness from 133 to 246 nm.

Optical properties

Optical transmittance spectra of VO_x:Co thin films can be appreciated in Figure 1a with transmittance percentage of 65 - 75 % in the 800 -1350 nm wavelength (λ) range. The main influence of Co²⁺ inclusion is observed in UV-A to blue region, at 325 - 500 nm, as reported elsewhere (^{Hajzeri et al., 2012}; ^{Martínez-Gil et al., 2020}). Furthermore, samples with larger Co²⁺ concentration, such as S ₄ and S ₅ , show lower transmittance at the visible region (400 - 700 nm). Optical absorption coefficient spectrum of each film sample, α(λ) (Figure 1b), is calculated from the simplified Beer-Lambert equation: a=-1dln⁡T , where d is film thickness and T(λ) the transmittance spectrum. The optical absorption onset (or band gap E _g ) value is estimated by intersection of two straight fitting lines. The E _g values of pure vanadium oxide and vanadium with different amounts of Co²⁺, are around 2.3 eV, which is very close to that previously reported for V₂O₅ [27]. It is also observed that the absorption coefficients of all the cobalt doped vanadium oxide thin films (from S ₂ to S ₅ ) are higher compared to the pure vanadium oxide sample (S ₁ ), keeping the same forms of the spectra. Higher absorption region (> 3.8 eV) in these samples should be related to the glass substrate absorption.

Figure 1 (a) Transmittance and (b) optical absorption coefficient spectra of vanadium oxide (S ₁ ) and cobalt doped vanadium oxide (S ₂ to S ₅ ) thin films. 

Figura 1: Espectros de transmitancia (a) y coeficiente de absorción óptica (b) para películas delgadas de óxido de vanadio (S ₁ ) y óxido de vanadio dopado (S ₂ to S ₅ ). 

X-ray diffraction

X-ray diffraction pattern of VO_x film sample (S ₁ ) is presented in Figure 2, which appears to correspond to an amorphous material due to the broad signal around 22°, plus a monoclinic phase of HNaV₆O₁₆•4H₂O with the plane orientation (100) at around 2θ ∼ 8° (JCPDS# 49-0996 (^{Channu et al., 2011})). This crystalline compound could be a residual from VOSO₄·xH₂O + NaOH, reaction and was not be eliminated during the 200 °C annealing or the thin film processing. Interestingly, after Co²⁺ addition, the reflection peak around 2θ ∼ 8° gradually reduces. The higher the concentration of Co²⁺, the smaller the intensity of that peak, until its complete elimination in the sample with the highest concentration of Co²⁺ (sample, S ₅ ). S ₅ shows the diffraction pattern typical of an amorphous film, hence it is a predominantly amorphous material. Since no diffraction peaks of CoO_x or VCoO_x compounds appear in the five VO_x:Co²⁺ samples, it seems that low temperature annealing leads to amorphous materials, as indicated in (^{Hu et al., 2017}; ^{Martínez-Gil et al., 2020}).

Figure 2 X-ray diffractions patterns of vanadium oxide with different Co²⁺ concentrations. 

Figura 2:  Patrones de difracción de rayos X de óxido de vanadio con diferentes concentraciones de Co⁺². 

Thin films surface morphology

Figure 3 shows AFM three dimensional surface topography images in scales of 2 m x 2 µm of (a) S₁, (b) S₂, (c) S₃, (d) S₄ and (e) S₅ VO_x:Co²⁺ film samples. The root-mean-square roughness of the same samples is plotted in Figure 3f. The surface topography of pure VO_x film (S ₁ ) presents lumps and the lowest roughness. As the cobalt ions are incorporated into VO_x host, variations in shape and roughness are observed in AFM images. Notable changes are observed in samples S ₃ and S ₄ that exhibit a ring like morphology and increased roughness (8 - 9 nm) in comparison with S₁ and S₂ (4 - 5 nm) samples. sample S ₅ that contains the highest cobalt concentration loses the ring pattern in its surface morphology and shows the highest roughness (∼23 nm) among the five samples. Similar results are reported elsewhere, suggesting that the inclusion of Co²⁺ in VO_x host by solution methods modifies the superficial morphology of cobalt doped vanadium oxide thin films (^{Fuentes-Ríos et al., 2021}). Such porous morphology encourages potential use of VO_x:Co²⁺ thin films as a cathode material for lithium battery applications (^{Wang et al., 2011}).

Figure 3 Three-dimensional AFM surface morphology images of (a) S₁, (b) S₂, (c) S₃, (d) S₄, (e) S₅ samples. (f) Roughness values of the same samples. 

Figura 3: Imágenes tridimensionales por AFM de la morfología superficial de las muestras (a) S₁, (b) S₂, (c) S₃, (d) S₄ y (e) S₅. (f) Valores de aspereza de las mismas muestras. 

To complement the AFM results, SEM micrographs of S ₁ , S ₃ and S ₅ samples were analyzed and are shown in Figure 4. Without cobalt addition, vanadium oxide sample S ₁ shows a granular ground with shallow leaves. The S₃ sample (VO_x:Cobalt salt = 50:50), presents an interconnected rings or porous arrangement, very similar to the AFM image of the same sample (Fig. 3c). The pore size distribution histogram of this film sample is in Figure 4d. It shows that the pore size varies from 100 to 300 nm, and those with ~150 nm predominates. By chemical element mapping in the SEM image of sample S ₃ , vanadium and cobalt elements have been detected to have a homogeneous distribution inside the pores. Finally, the sample with the highest concentration of cobalt, S ₅ , shows an irregular porous granular pattern at the surface, and under such pattern a continuous film is observed.

Figure 4 SEM micrographs with different magnifications of a) S₁, b) S₃, c) S₅ and d) histogram of S₃ pores. 

Figura 4: micrografías de microscopia electrónica de barrido con diferentes magnificaciones de las muestras a) S1, b) S3, c) S5 y d) histograma de los poros de la muestra S3. 

Chemical composition

To assess the chemical state of the VO_x:Co²⁺ film samples, an accurate peak-fitting analysis of their X-ray photoelectron spectra (XPS) (A. Herera-Gomez, no date) was carried on. Figure 5 shows the photoelectron spectra of the five samples corresponding to the (a) Co 2p and (b) O 1s and V 2p core-levels. From the Co 2p spectra (Figure 5a) we observe that the experimental data are congruent to that of the Co²⁺ chemical state, showing the main peak centered at 780.80 eV, but at the same time, three satellite peaks at higher binding energies that are characteristics of the Co²⁺ state (^{Yang et al., 2010}; ^{Biesinger et al., 2011}; ^{Martínez-Gil et al., 2020}). It is important to mention that the relative intensities of those satellite peaks suggest that there are neighboring atoms arranged in a short-range order, possibly having distortions that deviate from the octahedral Co²⁺ symmetry. That is, the intensities of the satellite peaks are proportional to the molecular defects or amorphous state of the material, which correlates to the amorphous XRD results for all samples.

Figure 5 X-ray photoelectron spectra of VO_x:Co²⁺ films with different concentrations of Co²⁺. 

Figura 5: Espectro de rayos X de películas de VO_x:Co²⁺ con diferentes concentraciones de Co²⁺. 

The V 2p and the O 1s spectra share the same binding energy range (Figure 5b), where we can observe the sole presence of a vanadium oxide. Due to the large spread of the reported binding energy values of this compound, to assess the present films we have employed the method proposed by ^{Hryha et al. (Hryha, Rutqvist and Nyborg, 2012}), where instead of using only the energy position of the V 2p _3/2 for chemical state determination, we have taken into account the difference in the binding energy between the O 1s and the V 2p _3/2 core levels (Δ= BE(O1s) - BE(V2p _3/2 )). The results show that for all samples the energy separation is around 13.50 eV, which suggests that the oxide obtained with our chemical process is VO₂ (^{Hryha, Rutqvist and Nyborg, 2012}). Here it is interesting to note that the energy position of the V 2p _3/2 in our samples is at 516.50 eV, which is almost the average value of those reported for VO₂ (515.95 eV) and V₂O₅ (517.20 eV), however, due to also the large spread in binding energy calibrations in the literature, we discard the possibility of the V₂O₅ compound.

From peak intensities directly derived from the photoelectron spectra, chemical compositions of our VO₂:Co²⁺ film samples have been determined. We have corrected the peak intensities with the appropriate physical parameters accounting for the photoemission signal attenuation due to scattering (^{Cabrera-German, Gomez-Sosa and Herrera-Gomez, 2016}). The results expressed in terms of atomic percentage for each of the observed chemical species are presented in Figure 6a. Here, we observe that as the Co²⁺concentration in the precursor solution increases, the amount of Co²⁺ in the deposited films also increases, from 0 to 16.55 at %. On the other hand, the V⁴⁺ atomic concentration in all five samples vary slightly, from 31 at % (sample S ₁ ) to 27 at % (sample S ₅ ), which suggests that the VO _x yield is high in all the probed samples. We can also observe that the percentage of oxygen atoms bonded to metals is coherent to the chemical species: VO₂ and CoO, reaching 69 at % for sample S₁ and decreasing up to 56.4 at % for sample S₅ as the Co²⁺ in the films is the largest. The atomic ratio of vanadium versus cobalt (V:Co) in each sample is listed in Table 1, changing from 6.6:1 (S ₂ ), 4.9:1 (S ₃ ), 2.9:1 (S ₄ ) to 1.6:1 (S ₅ ).

Figure 6 XPS chemical composition assessment: (a) atomic percentage (at%) of elements O, V and Co, and (b) O/V and O/(V+Co) atomic ratios of VO_x:Co²⁺ thin films. 

Figura 6: Evaluación de la composición química por XPS: (a) porcentaje atómico (at%) de los elementos O, V y Co, y (b) proporciones atómicas O/V y O/(V+Co) de las películas delgadas de VO_x:Co²⁺. 

For a better assessment of metal oxide compositions in the film samples, the ratio between oxygen and metal concentrations is plotted in Figure 6b. Here, the O/V atomic ratio lies around the expected stoichiometry of VO₂ for all the samples, which further confirms our previous analysis. Yet, if we determine the atomic ratio of oxygen to the sum of the two metallic species, O/(V+Co), we observe that such ratio decreases with increasing concentration of Co²⁺ in the reaction solution, a trend that follows the expected stoichiometry of the CoO compound.