Theoretical Discovery and Experimental Synthesis of Ultra-wide-band-gap Semiconductors for Power Electronics

Abstract

Semiconductors have unique electrical properties such as variable electrical conductivity through doping, making them an essential component of modern electronics. Silicon is the traditional semiconductor material that governs modern microelectronic technology. However, with the advent of artificial intelligence (AI), big data, autonomous vehicles, and Internet of Things (IoT), there is a need for developing advanced semiconductor materials that can operate more energy-efficiently at high power and high frequency. Power electronics is the application of electronics to the control and conversion of electrical power and it seeks to enhance energy conversion efficiency by utilizing ultra-wide-band-gap (> 3.4 eV, UWBG) semiconductors with high carrier mobility and high thermal conductivity. However, materials with wide band gaps generally have heavy effective masses which lead to inefficient charge transport and doping. The state-of-the-art materials suffer from doping asymmetry and/or poor thermal conductivity, which motivates alternative UWBG semiconductors with enhanced material properties.

This thesis investigates theoretical discovery and experimental synthesis of novel UWBG semiconductors that can overcome the challenges faced by the state-of-the-art materials. To discover the extreme limits to semiconductor band gap, wide-band-gap materials are surveyed and the key material factors are identified which enables semiconductivity of materials. It is found that materials composed of light elements and crystallized in densely packed structures give rise to a combination of wide band gap and light effective mass that enables shallow dopant, high mobility, and weakly bound polarons. For the candidate semiconductors, atomistic calculations are performed to explicitly calculate dopant ionization energies, formation of DX centers, and carrier mobility. Calculation results revealed promising semiconductor materials with band gaps up to 11.6 eV (even wider than insulators), which challenges the conventional gap-based criterion to distinguish semiconductors from insulators.

Among the materials, rutile GeO2 (r-GeO2) is identified to be a promising, yet unexplored UWBG (4.68 eV) semiconductor for rapid transformative impact of power electronic applications. Hybrid density functional theory predicted shallow ionization energies for donors such as SbGe, AsGe, and FO, and the ionization energy of 0.45 eV for Al acceptor that can be lowered by heavy Al doping to enable ambipolar doping. The electron and hole mobilities are also calculated (289 cm2 V−1 s−1 and 28 cm2 V−1 s−1, respectively), which are close to the state-of-the-art semiconductors such as GaN. Thermal conductivity is also measured for hot-pressed r-GeO2 polycrystals using laser-flash. The measured value is 51 W m−1 K−1, 3 times higher than β-Ga2O3. Though thin film growth of r-GeO2 is challenging due to the presence of kinetically stable glass phase and high vapor pressure of GeO, the first synthesis of single crystal r-GeO2 thin films is demonstrated using molecular beam epitaxy. Due to the competitive phase space, growth conditions that utilize a novel preoxidized molecular precursor as well as buffer layers with reduced misfit strain are key to realizing the rutile phase. Though the available substrates all have large lattice difference with r-GeO2 (> 4%), flux synthesis technique and mechanical polishing allow the fabrication of 4 × 2 mm2 size r-GeO2 single crystal substrates with highly crystalline surfaces that can be utilized for epitaxial film growth. This work provides opportunities to realize new UWBG semiconductors with enhanced material properties that can drive energy-efficient power electronics.