Jacob Melby — 2009-10 Fellow
Twenty-two percent of the electricity and an associated 250 million metric tons of CO2 are generated/year in the USA for lighting. Forty percent is derived from incandescent bulbs, which are less than 5% efficient . In contrast, light emitting diodes (LEDs), have very high efficiencies (approaching 100% in some cases). Conclusions of government-sponsored studies  show that significant use of LEDs for solid-state white lighting (SSL) would eliminate 130 new coal-fired power stations and the associated CO2 and save roughly 275 TWh/year in energy.
White light is produced commercially via LEDs by the interaction of blue light with and transmission through a yellow phosphor; however, this approach is inefficient and results in poor color rendering. The combination of green, blue and red light produces significantly better color matching; however, the green LEDs experience a significant drop in efficiency and a shift in wavelength with increasing power that must be eliminated for use in SSL.
In a green LED, electrical current is converted directly into light via radiative recombination of electrons (-) and holes (+) in a multi-quantum well (MQW) InGaN/GaN "active region". Several theories regarding the sources of energy loss in these regions have been promulgated; however, fundamental studies combining theory and experiment to understand and reduce these inefficiencies are lacking. Our research focuses on surmounting the marked and very important energy loss at the p-type contact resulting from poor hole injection efficiency due to high contact resistance. The inability to inject high concentrations of holes into the device via the p-type contact also causes heat generation and degradation of the device and has a serious impact on electron/hole recombination and light emission in the MQW. Enhanced doping and annealing to achieve low resistance contacts to p-type GaN have not proven sufficient.
As such, we are investigating a novel science-based approach to increase the p-type hole concentration and decrease the aforementioned contact resistance via (1) fabrication of ohmic contacts to p-type GaN using a polarization-induced doping method and (2) investigation of the energy loss occurring at the contacts. This research integrates theoretical modeling with optical and electronic characterization of the device contacts and LEDs.
Group III-nitrides (AlN, GaN, InN and their alloys) possess 'spontaneous' polarization due to their ionic nature and their non-centrosymmetric crystal structures. The different lattice parameters in these materials also introduce piezoelectric polarization when pseudomorphic layers of different compositions are grown atop each other. Nitride-based heterostructures possess strong polarizations that create a large electric field and high sheet charge, or two-dimensional carrier gases, without the disadvantages of impurity doping . In this project we exploit these properties in InGaN/GaN heterostructures to produce tunneling (ohmic) contacts to the p-type layer of GaN-based LEDs.
i "DOE Solid State Lighting Status and Future" SPIE Annual Meeting (2004).
ii "The Promise of Solid State Lighting" OIDA Report (2001).
iii Ambacher et. al., J. Appl. Phys. 85, 3222, (1999).