Nitride-based semiconductors are a centerpiece of light-emitting diodes (LEDs) for solid-state lighting, select high-frequency/power electronics, and future multi-junction photovoltaics. Progress towards commercialization of these technologies demands high current densities that compel improved thermal management to lower operating temperatures. While packaging strategies improve heat dissipation, preliminary data show that the device itself has a large thermal resistance. Thermal conductivity (k) measurements on device-grade nitride films show very low values of k in thin 100nm aluminum nitride (AlN) and 100nm gallium nitride (GaN) layers. Electron microscopy images suggest that these reductions are caused by structural imperfections inherent to industrial growth processes. As yet, a clear scientific explanation is lacking. Intellectual Merit: The objective of this GOALI proposal is to study thermal transport in nitride semiconductor heterostructures composed of GaN, AlN, and indium gallium nitride (InGaN) thin films (~100nm), fabricated by scalable growth processes. An interdisciplinary team of academic investigators at Carnegie Mellon University (CMU), in partnership with Kyma Technologies, will study the nature of thermal transport by phonons in nitride nanostructures. Defective films and interfaces separated by distances commensurate to the bulk phonon mean free paths will be considered. Open scientific questions include: What is the thermal boundary resistance (R) at the device-substrate interface and between nitride layers? Can highly defective crystals transport heat in a manner similar to disordered materials? How do growth techniques impact thin film thermal properties?
To answer these questions, the investigation will focus on a common base structure that includes a sapphire, silicon carbide, or GaN substrate with an AlN nucleation layer followed by an InGaN buffer layer. Specific inquiries include: (i) the effect of the substrate on R and k of the AlN nucleation layer, (ii) the effect of growth technique on k and R of the AlN nucleation layer, and (iii) the effect of indium concentration on k and R of the InGaN buffer layer. Controlled growth of nitrides and imaging of defect density and structure (Davis, Paskova) will be used to inform the atomic structure for molecular dynamics simulations (McGaughey). Simulation results will be compared with direct measurements of k and R on these samples, made using a pump-probe optical method called Frequency Domain Thermoreflectance (Malen).
Broader Impact: Partnership with Kyma Technologies makes this research directly transferable to industry, where nitride devices have the potential to revolutionize lighting and to outperform silicon-based electronics. Kyma recognizes the critical need for incorporating thermal-management into the nitride device structure. Access to Kyma's growth technologies, which represent the future of nitrides, will make any discoveries far-reaching. Continuing collaborations on nitride science and technology with the Cree Corporation and the Naval Research Laboratory further support the need for this research. The educational activities will expose students to industry-driven academic research through curriculum development, guest lectures and seminars, and summer internships at Kyma. Kyma, in turn, will receive exposure at CMU through integration of the research topics and results within courses taught by the academic PIs. An LED-based outreach activity "Lighting: The Next Generation" will be developed for the Society of Women Engineers High School Day, Pittsburgh public schools, and the YWCA TechGyrls program.
Nitride-based semiconductors are a centerpiece of light-emitting diodes (LEDs) for solid state-lighting, select high-frequency/power electronics, and future multi-junction photovoltaics. Progress towards commercialization of these technologies demands high current densities that compel improved thermal management to lower operating temperatures. While packaging strategies improve heat dissipation, our data generated through this NSF project show that the device itself has a large thermal resistance. The work specifically contributes the following conclusions to the nitride community: Thermal conductivities of the individual layers of gallium nitride (GaN) based LEDs were directly measured using the 3-omega method. Base layers of aluminum nitride (AlN), GaN, and indium gallium nitride (InGaN), grown by organometallic vapor phase epitaxy (OMVPE) on silicon carbide (SiC), have effective thermal conductivities that are much lower than bulk values. The AlN layer, used to nucleate device layers on the SiC substrate, had the highest overall thermal resistance caused by a severely defective AlN-SiC interface that caused additional phonon (vibrations that carry heat in crystalline solids) scattering [1]. Thickness dependent thermal conductivity measurements were made on AlN thin films grown by two methods (OMVPE and plasma vapor deposition) on SiC and sapphire substrates with differing surface roughness. We find that the AlN itself makes a small contribution to the overall thermal resistance. Instead, the thermal boundary resistance between the AlN and substrate is equivalent to 240 nm of highly dislocated AlN or 1450 nm of single crystal AlN [2]. Thermal conductivity in non-metallic crystalline materials results from cumulative contributions of phonons that have a broad range of mean free paths. We probed the phonon mean free path spectra of semiconductors relevant to nitride devices, including GaN, AlN, and 4H-SiC. We found that phonons with MFPs greater than 1000 nm, 2500 nm, and 4200 nm contribute 50% of the bulk thermal conductivity of GaN, AlN, and 4H-SiC near room temperature. This finding explains significant suppression of thermal conductivity in thin films and suggests that similar suppression will be realized for devices with feature sizes of order 1000 nm [3]. Thermal conductivities of model compound semiconductors, where the two species differ only in mass, were predicted using lattice dynamics calculations and the Boltzmann transport equation. The thermal conductivity varies non-monotonically with mass ratio, with a maximum value that is four times higher than that of a monatomic semiconductor of the same density. The very high thermal conductivities are attributed to a reduction in the scattering of optical phonons when the acoustic-optical frequency gap in the phonon dispersion approaches the maximum acoustic phonon frequency. These results shed light on trends of high thermal conductivity in nitride semiconductors [4]. Select publications from this grant that are herein referenced: 1. Su, Z.H., L. Huang, F. Liu, J.P. Freedman, L.M. Porter, R.F. Davis, and J.A. Malen, Layer-by-layer thermal conductivities of the Group III nitride films in blue/green light emitting diodes. Applied Physics Letters, 2012. 100(20). 2. Su, Z.H., J.P. Freedman, J.H. Leach, E.A. Preble, R.F. Davis, and J.A. Malen, The impact of film thickness and substrate surface roughness on the thermal resistance of aluminum nitride nucleation layers. Journal of Applied Physics, 2013. 113(21). 3. Freedman, J.P., J.H. Leach, E.A. Preble, Z. Sitar, R.F. Davis, and J.A. Malen, Universal phonon mean free path spectra in crystalline semiconductors at high temperature. Scientific Reports (Nature Publishing Group), 2013. 3: p. 2963. 4. Jain, A. and A.J.H. McGaughey, Thermal conductivity of compound semiconductors: Interplay of mass density and acoustic-optical phonon frequency gap. Journal of Applied Physics, 2014. 116(7).
