Technical: This project addresses phase transitions and defect formation in antimonide III-V materials, including direct gap quaternary materials GaInAsSb and InAs/Ga(In)Sb superlattices having bandgaps covering the 2-5 micron spectral range. These materials are expected to be of growing importance to infrared optoelectronic devices, including photodiode detectors, light emitting diodes, laser diodes, and thermophotovoltaics. However, GaInAsSb alloys grown under thermodynamic equilibrium conditions unstrained on GaSb substrates have immiscibility regions in the alloy phase space according to thermodynamic calculations, potentially limiting stability of desired compositions. This project will study and map out the influence of strain and growth kinetics on quantum well GaInAsSb alloys across the thermodynamic immiscibility region. Imaging techniques will be used to study characteristic defect formation, phase separation, and spinodal decomposition. Photoluminescence, absorption and optical lifetime measurements will be used to correlate morphology and optical properties, including radiative and nonradiative decay processes. Alternative materials to GaInAsSb alloys with direct gaps covering the 2-5 micron wavelength region, short-period, type-II InAs/Ga(In)Sb superlattices (T2SL) will be addressed as well. They have several characteristics in common with GaInAsSb alloys, including common elemental composition, as well as GaInAsSb regions at interfaces and in layers due to intermixing. The project aims to advance state-of-the-art understanding of crystal growth and material properties of GaInAsSb alloys and InAs/Ga(In)Sb superlattices, and allows testing at both research grade MBE reactors at U. IA as well as industrial grade MBE reactors at IQE. New, comprehensive data will be obtained on comparisons between measured and calculated properties of these materials, with theoretical calculations provided by ASL Analytical.
This GOALI (Grant Opportunities for Liaison with Industry) project addresses basic research issues in a topical area of materials science with high technological relevance. Collaborative links between the University of Iowa academic group and industrial partners, IQE and ASL Analytical are synergistic, and provide advantageous opportunities for scientific, technological, and educational impacts. The PI will recruit and hire a new graduate and undergraduate student to work on the project and advise the graduate student through the completion of his or her PhD. The graduate student will get research experience working both in academic (U. IA) and industrial (IQE, ASL) environments. The PI will reinvigorate a U. IA Physics and Astronomy outreach program called "Family Adventures in Science," featuring a series of lectures on a theme by different faculty members to the public; one of the themes will be semiconductors. Additionally, the PI will propose a one credit seminar for first year students on approaches to science through study of several famous scientists.
In this NSF grant, we examined the growth of GaInAsSb by molecular beam epitaxy. GaInAsSb is a quaternary semiconductor alloy important for infrared optoelectronics, with bandgap adjustable from 1.7 to 4.9 um while keeping the lattice constant matched to GaSb. However, this alloy is only thermodynamically metastable or unstable over a significant portion of this wavelength range at typical growth temperatures employed by epitaxial techniques. Molecular beam epitaxy is a so-called nonequilibrium growth technique due to the large driving force for growth, but also because surface kinetics allow materials to be grown far from thermodynamic equilibrium. By simply lowering growth temperature, and limiting adatom diffusion length, we were able to demonstrate growth of high quality GaInAsSb without phase separation across the compositional range. For example, Fig. 1 shows bright, room temperature photoluminescence peaked from 1.7 to 4.9 um. Figure 2 shows how material quality of Ga(50)In(50)As(45)Sb(55) is high quality when grown from 400-440C. At about 460 C, as adatom diffusion length increases, the alloy abruptly phase separates, as evidenced by a dramatic increase in surface roughness and width of the X-ray diffraction peak. Even though phase separation can be largely suppressed with growth temperature, there is still some broadening of the alloy X-ray diffraction peak, indicating a low degree of alloy fluctuations. These fluctuations can be further suppressed by straining the epilayer. Figure 3 shows a Ga(50)In(50)As(45)Sb(55) epilayer grown at 450 C, so that the alloy did not phase separate at the edges, but did phase separate in the center of the sample where temperature is slightly hotter during growth. As the alloy is progressively strained, the phase separation is suppressed across the wafer, and the low degree alloy fluctuations at the edge of the wafer are also suppressed, as shown by the sharply decreasing width of the alloy X-ray diffraction peak, and the increase in photoluminescence. While strain can limit layer thickness, we found no evidence of relaxation 3-4x the critical thickness of the layer, and no reduction in optical quality up to 10x the critical thickness. In growing these epilayers, we were plagued early on by pyramidal defects, which are shown in Fig. 4 top and top inset, and which have also been reported by other groups. We found these pyramids are terraced from bottom to top, and form due to an Ehrlich-Schwoebel step edge potential which leads to an upward adatom current. By putting down a thin layer of AlAsSb, the very short diffusion length of Al atoms alters the surface from step flow to islands, thus disrupting the pyramidal growth. With subsequent GaSb growth, the pyramids disappear, and atomically smooth, step flow surface results, as shown in Fig. 4 (bottom). We looked at the dependence of the Shockley-Read-Hall defect-assisted recombination rate in mid-wave InAs/GaSb superlattices on interface growth method. While growth methods such as migration enhanced epitaxy can significantly smooth the interfaces, because they slow down the average growth rate, they lead to lower quality epilayers. A comparison of low density minority carrier lifetime in superlattices with InSb-like interfaces (for strain balancing the superlattice) grown by migration enhanced epitaxy, by InSb codeposition, and by "free" growth (i.e. let them form freely) showed InSb codeposition gave the best results, as summarized in Table 1. In our InAs/GaSb superlattice time-resolved carrier studies, we were able to obtain low density lifetimes of 90 ns, whereas we have obtained low density lifetime in similar bandgap InAsSb layers of over 1 us. We are currently studying whether the reduction in lifetime is caused by adding Ga to material, and to what degree the lifetime is limited by background carriers from native defects versus defect energy levels in the bandgap.
