Electronic/Optoelectronic Nitride Devices

Ultraviolet Emitters

Further development of group-III nitride based optoelectronics requires solution of a crucially important problem - obtaining high-density hole currents. It is well established that the difficulties in achieving high hole concentrations mainly originate from high values of activation energy of known acceptors (about 250 meV for Mg in GaN). Use of Al-containing compounds desired for the devices operating in the deep UV range leads to an increase in acceptor activation energy and, consequently, to a lower efficiency of activation. To overcome the low acceptor activation problem, it has been suggested that the average hole concentration can be enhanced in a p-doped ternary compound material with a spatially modulated chemical composition - a superlattice. Calculations and Hall measurements support the idea of improved acceptor activation in Mg-doped ternary superlattices: the average hole concentration can be increased up to one order of magnitude. However the main drawback of this approach is that the most of the holes ionized from the acceptors are localized inside the quantum wells with the potential barriers as high as 100 to 400 meV. These barriers hinder participation of the holes in vertical transport required in typical light-emitting devices. Thus, the high local hole concentration cannot be efficiently used in the traditional LEDs. Moreover, the lack of high-density hole current makes development of a nitride-based UV laser with vertical carrier transport virtually impossible. To resolve these problems, we are pursuing two novel concepts for the design of light-emitting devices, which utilize the planar nature of potential profile induced local enhancement of two-dimensional hole density.

ECE Nano Image - Ultraviolet Emitters

To increase the over-barrier hole concentration and the vertical hole current, we are modifying the traditional design of LEDs by introducing a two-terminal hole injector schematically as shown in (a) at right. The injector consists of a p-doped SL base and two contacts S and D. The injector is separated from the rest of the device by an i region. A bias voltage applied between the S and D contacts provides lateral hole acceleration and increases the effective temperature of the holes Th. This is known as the real-space transfer effect. An increase in Th will result in significant enhancement of over-barrier hot-hole concentration. This device can be thought of as a three terminal device, or Light-Emitting Triode (LET) schematically shown in (b) at right, with a hole-injector region, an intrinsic barrier i layer, and an n-doped region (contact C). If a p-doped SL is used as a hot-hole injector, the device can operate as a charge injection transistor. With the D contact as the ground and a positive voltage VS applied to the S contact, the injector would yield a lateral hole current and heating of the holes. Assuming that a negative bias VC is applied to the cathode C, both the hot holes from the SL and the electrons from the n region would be injected into the i layer, as illustrated in (b). Emission of light would occur in the i layer as the result of electron-hole recombination. The proposed AlGaInN-based LET will significantly increase the intensity of emission compared to the LED structure made of the same material, especially in the deep UV range. Preliminary estimates show that even in AlGaN structures, a modest lateral electric field of about 3-5 kV/cm may lead to more than an order of magnitude increase in over-barrier hole concentration at room temperatures (and more than two orders of magnitude at higher barriers or lower temperatures). In the LET structure, the tunneling current is also expected to be higher than in LEDs.

The planar character of the proposed hole injector will enable integration of a number of such injectors onto the same wafer. A large area hot-hole injector can be achieved by fabricating multiple contacts to the SL in a sequence of S-D units, as shown in Fig. 1(b) by the dashed lines, for the case of an identical lateral electric bias for all injectors. In general, the lateral bias applied to each of the integrated injector as well as the particular design (composition, etc.) can be different. This allows additional control and manipulation of hole injection into the active layer of the emitting device and introduces the ability for such a chip to simultaneously emit/absorb a broad range of wavelengths.

Runaway effects in nanoscale group-III nitride semiconductor structures

ECE Nano Image - Chart of effects in nanoscale group-III nitride semiconductor structures

We have revisited the problem of electron runaway in strong electric fields in polar semiconductors focusing on nanoscale group-III nitride structures. By developing a transport model that accounts for the main features of electrons injected in short devices under high electric fields, we have investigated the electron distribution as a function of electron momenta and coordinates. Runaway transport has been analyzed in detail. The critical field of this regime is determined for InN, GaN, and AlN. We found that the transport in the nitrides is always dissipative ~i.e., no ballistic transport!. For the runaway regime, however, the electrons increase their velocities with distance, which results in average velocities higher than the peak velocity in bulk-like samples. We have demonstrated that the runaway electrons are characterized by a distribution function exhibiting a population inversion.

We have calculated the distribution functions for total populations of different energy stairs, and average velocities at different electric fields ε. For small ε (<0.2) we found that for any z, the electrons occupy approximately one or two lowest-energy stairs, and the distributions Fs reproduce themselves at each stair resulting in distance-independent steady dissipative transport. The figure to the right displays the average velocity versus distance for different fields. For comparison, the velocity in the ballistic regime is shown by the dashed line.