Temperature dependence of the interband transition in a V₂O₅ film Open Access

Of all the vanadium oxides, including VO₂, V₂O₃, and V₆O₁₁, vanadium pentoxide (V₂O₅) is the most stable compound, and it exhibits highly anisotropic electrical and optical properties due to its orthorhombic structure (α-V₂O₅).¹ Thin-film V₂O₅ has attracted much attention owing to its unique electronic, chemical, and optical properties.^2–5 The outstanding properties of V₂O₅ films are of interest for use in various applications such as high-capacity lithium batteries, chemical sensors, solar cell windows, and electrochromic devices.^6–8 

One of the most important parameters for characterizing the properties of materials is the bandgap energy (E_g). Generally, the E_g of a semiconductor or an insulator has been found to decrease with increasing temperature. The variation of the fundamental E_g with temperature is very important for both basic science and technological applications. Previous investigations of the optical properties of amorphous and polycrystalline V₂O₅ films and single crystals have been conducted,^9–11 and it has been shown that V₂O₅ is an insulator with an indirect E_g of 2.2–2.4 eV.¹² However, understanding the optical properties of V₂O₅ films requires detailed studies on the interband transition in the fundamental absorption edge.

In the present study, we report the temperature dependence of the interband transition in a crystalline V₂O₅ thin film based on absorption and photoluminescence (PL) spectra taken at temperatures ranging from 10 to 300 K. The behavior of the visible emission in the film was determined by analyzing the indirect and direct E_gs.

An α-V₂O₅ thin film with orthorhombic structure was prepared on a (0001) Al₂O₃ substrate by RF magnetron sputtering using a V₂O₅ (99.99%) disk target with a diameter of 10 cm. Sputtering was performed for 120 minutes at an RF power of 200 W and a substrate temperature of 500°C. Ar and O₂ with 99.999% purity were used as sputtering and reactive gases, respectively, with an O₂ partial pressure of 10% and a total flow rate of 30 sccm. The base pressure of the chamber was less than 5.0 × 10⁻⁶ Torr, and the working pressure of the chamber during sputtering was approximately 1.0 × 10⁻³ Torr. The thickness of the grown V₂O₅ film was determined to be approximately 115 nm using a spectroscopic ellipsometer (Horiba Jobin-Yvon, Uvisel UV/NIR).

The microstructure of the V₂O₅ film was investigated using scanning electron microscopy (SEM; JEOL, JSM6335F) and using X-ray diffraction (XRD; Rigaku, D/MAX-Rc) with Cu Kα radiation at a wavelength of 1.54 Å. Raman spectra were obtained using a Raman/PL spectrometer (Horiba Jobin-Yvon, LabRAM HR) equipped with a He-Ar laser at a wavelength of 514.5 nm and a maximum power of 200 mW.

Transmission measurements were obtained to investigate the interband transition in V₂O₅ films at various temperatures using a cryogenic system (JANIS C210-4, USA) with a 100 W tungsten-halogen lamp. The absorption data was detected by a photomultiplier tube using a grating monochromator (SPEX 1702, f = 3/4 m) and analyzed with a computer-aided data acquisition system. PL measurements were also performed at temperatures from 10 to 300 K using a Raman/PL spectrometer equipped with a He-Cd laser at a wavelength of 325 nm and a maximum power of 200 mW.

The XRD pattern of the V₂O₅ film, measured at room temperature, shows two peaks belonging to the α-V₂O₅ (i.e., the orthorhombic structure; Fig. 1(a)). A sharp peak corresponding to α-V₂O₅ (001) appeared at 2θ = 20.09°.¹³ The film had a rod-like surface morphology (Fig. 1(b)).

The Raman spectrum of the film is consistent with the Raman-active mode for α-V₂O₅ (Fig. 2), indicating an orthorhombic structure in agreement with the XRD results. The intense peak at 143 cm⁻¹ is related to vibrations of the V–O–V chains, and its presence indicates the layered orthorhombic structure of the V₂O₅ phase. The peak at 995 cm⁻¹ corresponds to the terminal oxygen (V = O) stretching mode, which is related to the structural quality and stoichiometry of the film.¹⁴ 

In order to study the temperature dependence of the interband transition in the α-V₂O₅ film, we measured its optical transmission at a normally incident angle as a function of temperature. In optical measurements, light intensity in material is characterized by applying Beer's law (I = I₀exp ^−αd), where α is the absorption coefficient of the material and d is the distance from the surface (or film thickness). In the absorption edge of a light-absorbing film, since the decay of the light intensity is mostly governed by the absorption, the reflectance losses in deducing α are sometimes ignored. In our study, the transmission was measured as light intensity and α was obtained by the light intensity ratio (I/I₀) and the film thickness. The α-V₂O₅ film has a high refractive index of approximately 2.8 near the absorption edge (see Fig. 8(a)). Thus, we included reflectance losses of 12% in calculating α.

Transmission spectra of the V₂O₅ film were collected at various temperatures from 300 to 20 K; as the temperature decreased, the transmission was enhanced for wavelengths in the region above 450 nm (Fig. 3(a)). To analyze the interband transition in the V₂O₅ film, we calculated absorption spectra using the measured transmission spectra; as the temperature decreased from 300 to 20 K, the spectra shifted toward higher energies over the entire photon energy range (Fig. 3(b)). The spectra revealed two distinct interband transitions that are attributable to indirect and direct transitions, which are denoted by two dotted lines in Fig. 3(b).¹⁵ This result indicates that V₂O₅ is a material with indirect bandgap and direct transition energies.

In order to clarify the interband transition characteristics in the V₂O₅ film, we calculated the optical bandgap energy (E_g) of the film using a Tauc plot. The theoretical interband absorption at a fundamental absorption edge is given by (αhv) = A(hv − E_g)ⁿ, where A is a constant and the exponent n determines the transition type, being 1/2 for a direct transition and 2 for an indirect transition.¹⁵ (αhv)^1/n is plotted against hνand then E_g is determined by extrapolating the plot to the photon energy intercept. We obtained Tauc plots (n = 1/2 and 2) at 20 K and 300 K (Fig. 4); in the extrapolations of the indirect gap, we used the region with (αE)^1/2 of 0.8 − 1.4 × 10⁴ m^−1/2 eV^1/2. The extrapolation in the region below 0.9 × 10⁴ m^−1/2 eV^1/2 corresponds to an E_g of about 1.9 eV (652 nm); this is too low to be consistent with the results from the absorption spectra (Fig. 3(b)), which estimate the indirect E_g to be above 2.1 eV by extrapolation in the region with α below 2.5 × 10⁷ m⁻¹. Thus, we selected the (αE)^1/2 region of 0.8 − 1.4 × 10⁴ m^−1/2 eV^1/2 for the extrapolation, which also estimates the indirect E_g to be about 2.1 eV. The E_g values at 20 K and 300 K were respectively found to be 2.26 eV and 2.16 eV for the indirect transition and 2.67 eV and 2.64 eV for the direct transition. This suggests that α-V₂O₅ is a material with two interband transitions.

A series of PL spectra of the V₂O₅ film were collected over a range of temperatures from 10 to 300 K (Fig. 5). These spectra included an intense peak centered near 710 nm and a weak peak centered near 530 nm, corresponding respectively to photon energies of 2.34 eV and 1.74 eV. As the temperature decreased from 300 K to 10 K, the PL peak centered at 710 nm became approximately 25 times more intense, without any considerable shift in the peak. The peak at 710 nm may be due to oxygen vacancies introduced during the growth of the film.^16,17 Reducing the temperature from 300 K to 10 K also enhanced the PL peak centered near 530 nm approximately threefold. This 530 nm (2.34 eV) peak is likely due to the band edge transition since the bandgap of V₂O₅ is approximately 2.2–2.4 eV.^12,18,19

To confirm whether the 530 nm peak was caused by an interband transition in the V₂O₅ film, we analyzed this peak in detail, cumulating enlarged spectra over the range of 2.0–3.4 eV to confirm variations near 530 nm (Fig. 6). This revealed a remarkable peak shift from 2.45 to 2.32 eV with decreasing temperature (from 300 K to 10 K), which may be due to a reduction in emissions caused by crystalline defects such as vacancies and disorders.^20–22 

Both E_gs of the α-V₂O₅ film were found to increase with decreasing temperature (Fig. 7). The direct E_g increased during the 300–150 K temperature change to a greater extent than was caused by further changes below 150 K; the indirect E_g varied more stably below 150 K than above. This may be because the lattice dilatation effect decreases with decreasing temperature. The temperature dependence of E_g in a semiconductor or insulator can be explained by the sum of two distinct mechanisms: thermal expansion of the lattice and the electron–phonon interaction.^23,24 In the case of CdTe, it is known that the lattice dilatation effect contributes more to the bandgap shift than the electron–phonon interaction above 100 K, while below 100 K the electron–phonon interaction is responsible for the energy shift.^23,25

The PL peak centered near 530 nm, which is thought to be caused by an interband transition, varied within a range of energies between the indirect and direct E_g values (Fig. 7). Moreover, the PL peak neared the indirect bandgap energy with decreasing temperature, with a difference at 20 K of only 0.06 eV. In contrast, the PL peak differed by a much larger 0.34 eV from the 2.67 eV direct transition energy at 20 K, making it clear that the PL near 530 nm is not induced by the direct transition, and rather is due to the indirect transition in α-V₂O₅.

The PL peak position exhibits a different temperature dependence from that of the indirect and direct E_g (Figs. 6 and 7). The PL peak position increases in energy with rising temperature (Fig. 7). This may be explained by the thermal expansion of the lattice and increasing electron–phonon interaction; as mentioned above, the lattice-dilatation effect and the electron–phonon interaction increase with temperature. These may act as the factors that interrupt the indirect transition in the V₂O₅ film. Recall that the PL in the V₂O₅ film near 530 nm is very weak (Fig. 5); moreover, the PL weakens further with rising temperature. This may indicate that the PL is caused by the indirect transition, and weakens due to the enhancement of the lattice-dilatation effect and electron–phonon interactions. It may also be the case that the PL is induced by some defects, having higher energy than the indirect bandgap energy, such as V vacancies and donor–acceptor levels introduced during the growth of the film. From these results, we presume that the lattice-dilatation effect and the electron–phonon interaction are responsible for the blue shift of the PL position.

The spectra of the complex refractive index (n and k) of the α-V₂O₅ film were obtained to confirm the transmission measurements. Ellipsometric data (Ψ and Δ) on the film were obtained using a spectroscopic ellipsometer (Jobin-Yvon, Uvisel UV/NIR) with a photon energy range of 0.75–4.0 eV at an incident angle of 70°. SE data for the V₂O₅ film were analyzed by means of an optical model based on the Bruggeman effective medium approximation (BEMA).²⁶ The double new amorphous (DNA) formula, combined with two oscillators, is utilized to describe the dispersion in the optical constants of the V₂O₅ film over the entire spectral range.²⁷ Details of the modeling and analysis are reported in Ref. 28.

Both the n and k spectra of the V₂O₅ film show the typical dispersion shape of an α-V₂O₅ film (Fig. 8(a)).^28,29 Above the fundamental absorption edge (at approx. 2.4 eV) indicating an indirect transition, the k spectra abruptly increased, and another direct transition was observed near 3.0 eV. For comparison with transmission measurements, the transmittance at room temperature was simulated from the n and k obtained via the SE (see Fig. 8(b)), and then the absorption spectra (α) and the E_g of the film were calculated using Beer's law and a Tauc plot, respectively. Essential Macleod (Ver. 8.16) was used for simulating the transmittance. The E_g values were 2.25 eV for the indirect transition and 2.62 eV for the direct transition at 300 K (Fig. 8(c)). The indirect E_g shows a discrepancy of 0.09 eV, due to the difference in sensitivity between SE and transmission measurements, this is in good agreement, and makes it clear that the α-V₂O₅ is a material with two interband transitions.

Absorption coefficients at room temperature were obtained from the transmission data, the simulated transmission, and the k (Fig. 9); despite considering the reflectance losses of 12%, the α determined from the transmission data using Beer's law showed somewhat high values over the entire photon energy range. This result may be due to the scattering losses from the surface of the film and the interface between the film and the substrate, as well as the sensitivity of the photomultiplier tube used as the detector. Although there is some disagreement between the transmission measurements and the ellipsometric measurements, both techniques clearly confirm two interband transitions in the α-V₂O₅.

The temperature dependence of the interband transition in an α-V₂O₅ film was investigated by analysis of absorption and PL spectra at temperatures ranging from 10 to 300 K, and the visible emission behavior due to the interband transition was determined by analyzing the indirect bandgap and direct transition energies. The variations in transmission with temperature indicate that the α-V₂O₅ film has two distinct interband transitions, implying indirect and direct transitions. This result was confirmed by SE measurements. The blue shift of both the transitions in the α-V₂O₅ film with decreasing temperature was attributed to diminishing lattice dilatation effect and electron–phonon interaction. Based on variations in PL with temperature, we attributed the emission near 530 nm in the α-V₂O₅ film to the indirect transition.

This work was supported by the 2012 Research Fund of the University of Ulsan.