SrNbO3 as a transparent conductor in the visible and ultraviolet spectra

The study addresses the need for transparent conductor materials that combine high electrical conductivity with high optical transparency, especially in the ultraviolet (UV) spectral range required for applications such as solar-blind detection, UV curing, biomolecule sensing, germicidal irradiation, lithography, phototherapy, and UV-enhanced plant growth. While visible and UVA LEDs can exhibit high external quantum efficiencies (EQEs), deep-UV LEDs suffer from poor EQEs due to challenges including hole injection, defects in wide bandgap active regions, and the absence of suitable transparent electrodes with both high transparency and conductivity in the deep UV. Indium tin oxide (ITO), the prevalent transparent electrode, absorbs strongly near 360 nm and is unsuitable for deeper UV. Ultrawide bandgap transparent conducting oxides (TCOs) like β-Ga2O3 and ZnGa2O4 provide UV transparency but have electrical conductivities far below those of degenerately doped ITO. To overcome these limitations, the authors propose SrNbO3, a correlated metal, as a transparent conductor with excellent performance from the visible into the 260–320 nm UV range. The design concept leverages an energetically isolated conduction band derived from Nb 4d orbitals and moderate electron correlations that shift the plasma frequency and interband absorption to enable transparency while maintaining high conductivity.

Conventional transparent electrodes focus on visible transparency with high conductivity, including ITO and multilayer oxide-metal-oxide stacks (ITO/Cu/ITO, IZO/Ag/IZO), multicomponent oxides (Zn–In–Sn–O), and indium-free doped ZnO (Al, Ga). Ultrathin metals (Ag, Ni, Cr) have also been explored. However, these approaches lack sufficient transmittance in the deep UV. Ultrawide bandgap semiconductors such as β-Ga2O3 (Eg ~4.5 eV) and ZnGa2O4 (Eg ~5.0 eV) offer UV transparency but exhibit conductivities orders of magnitude lower than ITO, limiting their usefulness as low-resistance UV electrodes. Improvements in UV LED EQE have often relied on flip-chip designs with metal reflectors to enhance light extraction, but these add cost and complexity. Correlated oxides like SrVO3 have been identified as visible transparent conductors, yet their interband absorption edge near ~2.9 eV (427 nm) compromises visible transparency. Theory suggests that increasing the electronegativity difference between the transition metal cation and oxygen (e.g., replacing V4+ with less electronegative Nb4+) can enlarge the O 2p–t2g separation, blue-shifting the interband absorption edge into the UV while reducing correlation strength (larger 4d orbital overlap), potentially improving both transparency and conductivity.

Computational: Density functional theory (DFT) calculations were performed using the linearized augmented plane wave method (Wien2k). A 25×25×25 Monkhorst–Pack k-point grid was used to converge the density of states; 12×12×12 grids sufficed for other calculations. Fully charge self-consistent DFT+DMFT calculations used the EDMFT package with the Luttinger–Ward functional. On-site Coulomb interaction U = 6 eV and Hund’s coupling J = 0.7 eV were employed with nominal double counting, appropriate for early d1 oxides. The DMFT self-energy near the Fermi level was Fermi-liquid-like (small imaginary part, linear real part), validating extraction of renormalization constants. DFT and DMFT were used to compute band structures, spectral functions, band widths, and to estimate plasma frequencies and correlation renormalization constants. Sample growth and structural characterization: SrNbO3 films were grown on (100) KTaO3 substrates by pulsed laser deposition (PLD) using a sintered Sr2Nb2O7 target and a KrF excimer laser (λ = 248 nm). Substrates were heated to 700 °C with an infrared lamp heater. Films were grown under oxygen-deficient conditions; comparable properties have been reported by other groups under similar conditions. Film crystal structure and thickness were characterized by X-ray diffraction and X-ray reflectivity using a Bruker AXS D8 Discover diffractometer with 2D/1D detectors. Optical characterization: Spectroscopic ellipsometry was performed at room temperature. Substrate (KTaO3) spectra (Ψ, Δ) were collected prior to film measurements. SrNbO3 films with thicknesses 10, 23, and 29 nm were measured at incidence angles 50°, 60°, and 70° using an M-2000 Ellipsometer (0.734–5.043 eV) and IR-VASE (0.044–0.814 eV), and at 65.23° using an M-2000F Focused Beam Ellipsometer (1.242–6.458 eV). Data were modeled in CompleteEase, extracting the complex dielectric function ε(ω) = ε1 + iε2 across IR–UV ranges. The screened plasma frequency was determined from ε1(ω)=0. An extended Drude model was applied to extract the renormalization constant Z from optical data. Transmittance spectra of freestanding films were computed from the dielectric function, including reflections and interference; thickness-dependent sheet resistance incorporated surface/grain boundary scattering (Fuchs–Sondheimer) in figure-of-merit calculations. Electrical characterization: Temperature-dependent resistivity and Hall effect were measured using van der Pauw geometry in a Quantum Design PPMS with source current of 500 μA and magnetic fields up to 18 T perpendicular to the film plane. Carrier concentration and mobility were extracted from Hall measurements. Residual resistivity ratios (RRR) were determined from room temperature to 2.2 K. Electron mean free path was estimated via the Sommerfeld model using measured n and μ. Analysis: Figures of merit (Haacke Φ_TC = T^10/R_s) were computed for visible (400–700 nm) and UV (260–320 nm) spectral ranges by averaging transmittance over the respective wavelength windows and using thickness-dependent R_s. Comparisons were made to SrVO3 and ITO using literature optical data.

• DFT indicates SrNbO3 is metallic with three Nb 4d t2g bands crossing the Fermi level; O 2p valence bands lie well below the Nb 4d states. The O 2p–Nb t2g interband transition threshold is >4 eV; t2g→eg/Sr 5s transitions lie ~2.5–3 eV. Compared to SrVO3, SrNbO3 exhibits larger separations, blue-shifting interband absorption by >1.5 eV. • DMFT shows weaker correlations in SrNbO3 than in SrVO3 with a larger renormalization constant Z ≈ 0.72 (SrNbO3) vs ~0.55 (SrVO3), leading to a smaller reduction of conduction band width and a somewhat less red-shifted plasma edge. • Electrical transport (10–60 nm films on KTaO3): sheet resistance ranges from 67.5 Ω/sq (10 nm) to 7.3 Ω/sq (60 nm); resistivity 6.9×10^−5 to 3.8×10^−5 Ω·cm. RRR values are modest (1.6 for 23 nm, 3.2 for 37 nm), indicating significant temperature-independent defect scattering. • Hall measurements: room-temperature mobility ≈ 8 cm^2 V^−1 s^−1; carrier concentration ≈ 1×10^22 cm^−3 (about half of SrVO3 and over an order of magnitude higher than typical TCOs and UV TCOs). The high n compensates the lower μ, yielding superior conductivity to UV TCOs. • Estimated electron mean free path Λ ≈ (3.5 ± 0.7) nm, much shorter than noble metals (Ag ~52 nm) and comparable to SrVO3 (5.6 nm) and CaVO3 (3.9 nm), enabling aggressive thickness scaling with limited surface scattering penalties. • Optical properties from ellipsometry: IR response dominated by a Drude peak; measured screened plasma energy ħω_p = (1.98 ± 0.03) eV. DFT predicts reduced plasma energy 2.15 eV; correcting for DFT–DMFT mass renormalization gives 1.82 eV. Extended Drude analysis yields experimental Z = 0.89 ± 0.02, leading to a corrected reduced plasma frequency of (1.91 ± 0.04) eV, in good agreement with measurement. • Interband absorption edge is at ~4.8 eV (~260 nm), consistent with DFT and blue-shifted relative to SrVO3 by >1.5 eV. ε1 and ε2 are small throughout the visible and into the UV up to the ~4.8 eV edge, implying low absorption. • Figure of merit Φ_TC (Haacke): visible (400–700 nm) maximum ≈ 5×10^−3 Ω^−1 at 10 nm thickness, exceeding SrVO3 (≈ factor 2 lower) and ITO (≈1.6×10^−3 Ω^−1 at ~150 nm). In the UV (260–320 nm), 10-nm SrNbO3 maintains Φ_TC > 10^−3 Ω^−1, comparable to ITO in the visible and an order of magnitude higher than ITO in the UV. SrVO3 is unsuitable in UV (Φ_TC lower by ~2 orders). Ultrawide bandgap UV TCOs such as Ga2O3 exhibit Φ_TC in the mid 10^−7 Ω^−1 due to low conductivity. • A weak DFT-predicted t2g→eg absorption near ~2.7 eV was not observed experimentally, potentially due to defects also responsible for low RRR.

The findings confirm that SrNbO3 meets the dual requirements of UV-transparent electrodes: high conductivity and high transmittance extending into the 260–320 nm range. The energetically isolated Nb 4d t2g conduction band increases the O 2p–t2g interband transition energy, pushing the absorption edge to ~4.8 eV, while moderate electron correlations reduce the plasma frequency sufficiently to limit free-carrier reflection in the visible without sacrificing metallic carrier concentrations (n ~10^22 cm^−3). Compared to SrVO3, SrNbO3’s larger O 2p–t2g separation and weaker correlations yield significantly improved blue/UV transparency while maintaining competitive electrical performance. As a result, SrNbO3 achieves Φ_TC values in the visible that exceed ITO and SrVO3 at thin film thicknesses, and uniquely sustains high Φ_TC in the 260–320 nm UV band where conventional options (ITO, ultrawide bandgap TCOs) fail due to either absorption or inadequate conductivity. These properties directly address the bottleneck of UV LED EQEs by enabling low-resistance, UV-transparent top electrodes, with implications for improved efficiency, lifetime, and device architectures across UV sanitation, photochemistry, medical therapies, lithography, sensing, and solar-blind detection.

SrNbO3 thin films are demonstrated, both theoretically and experimentally, as high-performance transparent conductors spanning the visible and deep into the UV (260–320 nm). DFT/DMFT reveal an energetically isolated conduction band and moderate correlations that yield a blue-shifted interband edge (~4.8 eV) and a reduced plasma frequency consistent with optical measurements. Transport shows metallic conductivity with high carrier concentration, and ellipsometry confirms low absorption up to the UV edge. Figures of merit surpass ITO in the visible at thin thicknesses and exceed ITO by an order of magnitude in the UV, far outperforming ultrawide bandgap TCOs. These results position SrNbO3 as a superior UV-transparent electrode, enabling advancements in UV LED technologies and related applications. Future work could focus on reducing defect densities to increase mobility and RRR, integrating SrNbO3 into complete UV optoelectronic device stacks, optimizing growth conditions for scalability and stability, and exploring substrate/strain engineering to further tune optical and electronic properties.

The films exhibit relatively low residual resistivity ratios (RRR ~1.6–3.2), indicating significant temperature-independent defect scattering, which likely also affects optical features (e.g., the predicted weak ~2.7 eV interband absorption was not observed). The correlation-induced red-shift in SrNbO3 does not fully push the carrier reflection edge into the infrared, leading to somewhat lower transmission at the longest visible wavelengths compared to SrVO3. Thin-film properties are thickness dependent due to surface scattering, and the performance evaluations (e.g., Φ_TC) are based on modeled freestanding films derived from ellipsometric dielectric functions, which may differ from device-integrated films on specific substrates or with encapsulation.