Papers

AVS2004 Session SS+EM+SC-ThA: Compound Semiconductor Growth and Surface Structure

Thursday, November 18, 2004 2:00 PM in Room 210C

Thursday Afternoon

Start	Invited?	Item
2:00 PM		SS+EM+SC-ThA-1 III/V Semiconductor Surfaces during Metalorganic Vapor-phase Epitaxy R.F. Hicks (University of California, Los Angeles) {Thin films made from compound semiconductors, such as indium phosphide, gallium arsenide and their alloys, have key applications in electronic and photonic devices. These single-crystal materials are fabricated by metalorganic vapor-phase epitaxy (MOVPE). In our laboratory, an ultra-high vacuum system has been connected to a state-of-the-art MOVPE reactor so that the semiconductor surfaces may be characterized in the growth environment. The atomic composition and structure of these materials has been determined by scanning tunneling microscopy, infrared spectroscopy, reflectance difference spectroscopy, x-ray photoelectron spectroscopy, and ab initio molecular cluster calculations. It has been found that indium phosphide and gallium arsenide exhibit very different surface structures under MOVPE process conditions. A disordered double-layer of As atoms with a few alkyl radicals are adsorbed on GaAs (001), while on InP (001), the surface is terminated with H atoms adsorbed on P dimers. This latter structure exhibits a single-domain (2x1) reconstruction. Another interesting example of III/V semiconductor surface chemistry is the formation of InGaAs/InP interfaces. Localized strain produces atomic arrangements that are distinct combinations of InP and InAs reconstructions. The application of this knowledge to the growth of epitaxial device structures will be discussed. invited by David Castner.}
2:40 PM		SS+EM+SC-ThA-3 Quantitative Analysis of Indium Concentration in InGaAs Quantum Dots and Wetting Layers Using Cross-sectional Scanning Tunneling Microscopy N. Liu, S. Govindaraju, A.L. Holmes Jr., C.-K. Shih (University of Texas at Austin) Scanning tunneling microscopy has been employed to explore self-assembled InGaAs quantum dots (QDs) grown by migration enhanced epitaxy (MEE). With atomic resolution, compositional analysis has been done for both the QDs and wetting layers quantitatively. We found that both vertical and lateral segregation play important roles during the formation of the islands and thereafter capping procedure. Depletion of the wetting layer, due to the formation of the QDs, is demonstrated. More importantly, it is found that after capping the amount of existing indium in the QDs and WL is less than that of deposited indium, indicating a portion of deposited indium atoms was evaporated to the vacuum during overgrowth of GaAs. This observation is different from previous results, which proposed indium re-distribution within GaAs matrix after capping growth. Based on the observed data, a growth model is also proposed.
3:00 PM		SS+EM+SC-ThA-4 Surface Science of Gallium Nitride and Related Alloys R.M. Feenstra (Carnegie Mellon University) The formation and structure of various faces of GaN, including adsorbed layers of Al, In, or H, are discussed. The primary preparation method is plasma-assisted molecular beam epitaxy although a comparison of results from metal-organic vapor phase epitaxy will also be presented. Experimental results from scanning tunneling microscopy allow us to broadly determine the characteristics of the surface structures, and first principles theory is then used to determine the precise atomic arrangements. For the case of surfaces grown by vapor phase epitaxy spectroscopic ellipsometry is used to identify the relevant surface phases. In contrast to other semiconductor surfaces, a central feature of many GaN reconstructions is their tendency to form metallic overlayers of metal (Ga, In, or Al) atoms. The terminating layers of metal atoms also leads to novel aspects of the surface kinetics - N atoms are predicted to diffuse easily between the metal layers, thus yielding enhanced surface diffusivity for those surfaces which are terminated by more than one layer of metal atoms. Work performed with Y. Dong, C. D. Lee, H. Chen, A. R. Smith (CMU); J. E. Northrup (PARC); J. Neugebauer (FHI, Berlin); C. Cobet, T. Schmidtling, M. Drago, N. Wollschlaeger, N. Esser, W. Richter (TU, Berlin); and supported by NSF and ONR.
3:40 PM		SS+EM+SC-ThA-6 First Scanning Tunneling Microscopy and Spectroscopy Study of c-GaN(001)-4x1 Tetramer Structure and SIESTA Surface Simulation H.A. Al-Brithen, M.B. Haider, N. Sandler, A.R. Smith (Ohio University); P. Ordejon (Institudo de Cienias de Materiales, Spain) Although early papers of the surface structure of c-GaN(001) reported 2x2 and c(2x2) reconstructions,[a] it was later shown both experimentally[b] and theoretically[c] that the intrinsic reconstruction is 4x1. However, until now, this 4x1 reconstruction has never been reportedly observed in real space. We have grown c-GaN on MgO(001) using radio frequency nitrogen plasma molecular beam epitaxy under Ga-rich conditions. RHEED patterns show that GaN(001) clearly exhibits 1x1 reconstruction during and after the growth; in fact, after cooling to ~ 200 °C a reversible disorder-order transition from 1x1 to 2x occurs, which may have been neglected or confused in earlier experiments with 2x2. This 2x is imaged using STM, finding that it is actually c(4x16) for Ga-rich growth and c(4x20) for more Ga-rich growth; STS spectra suggest that GaN(001)-c(4x16) is metallic. Annealing the film at T_s ~ 700-800 °C leads to the 4x1 reconstruction, as indicated by RHEED. STM performed on c-GaN(001)-4x1 shows that the surface consists of rows aligned along [110] with row spacing of 12.8 Å. Dual-bias STM images show a 180° phase shift of the filled and empty states profiles, as the sample bias changes from -1.2 V to +1.2 V, consistent with our recent simulated STM images, calculated using SIESTA code based on the tetramer model, showing that the filled state peak centered on the tetramer corresponds to the empty state minimum. STS acquired on the tetramer surface agrees with the semiconducting nature of 4x1, having a surface gap of 1.3 eV. In fact, the 4x1 tetramer structure was also predicted for c-AlN(001)[d], which widens the importance of understanding this reconstruction. Work is supported by NSF. [a] Brandt et al., Phys. Rev. B R2253 (1995). [b] Feuillet et al., Appl. Phys. Lett. 70(8) (1997). [c] Neugebauer et al., Phys. Rev. Lett. 80(14) (1998). [d] Felice et al., Appl. Phys. Lett. 74(15) (1999).
4:00 PM		SS+EM+SC-ThA-7 Ion Induced Step Debunching of GaN B. Cui, P.I. Cohen (University of Minnesota); A.M. Dabira (SVT Associates, Inc.) The development of surface morphology during ion bombardment has been described in terms of the curvature dependence of the sputtering yield [1] and asymmetric kinetics for the attachment of surface adatoms and vacancies at step edges [2]. We have used a Kaufman ion source to study the low energy ion effects during the MBE growth of GaN on sapphire substrates and GaN templates, comparing the results to these models. From a macroscopic point of view our measurements on GaN show quantitative agreement with the curvature driven theories. In particular we use the cross-over between Ga-limited growth and N-limited growth to estimate the N adatom concentration, a key ingredient of the theory. From a microscopic view, however, our RHEED and AFM studies have observed step debunching of multilayer steps and the elimination of hillock spirals. In these measurements, the starting GaN(0001) templates had 20 layer high mesas. After growth, round, nanoscale dimple structures, ranging from 90 nm to 850 nm, with bilayer steps were produced. This was seen with both Ar and nitrogen ions at energies ranging from 100-1200 eV. The size of the dimples and the terrace length of the debunched steps decrease with increasing sample temperature. After ion assisted growth, islands are found at the edges of the debunched steps. By tuning the ion energy and growth rate, uniform distributions of GaN nanoparticles, with means ranging from 50 nm to 200 nm, can be prepared. By combining ion induced step debunching and growth, step flow growth at the debunched steps is obtainable. Partially supported by the NSF and the AFOSR. 1. R. M. Bradley and J. M. E. Harper, J. Vac. Sci. Technol. A 6, 2390 (1988). 2. J. Kim, D. G. Cahill, and R. S. Averback, Phys. Rev. B 67, 045404 (2003).
4:20 PM		SS+EM+SC-ThA-8 Thermal Desorption of Deuterium from GaN(0001): A Sensitive Probe of Surface Preparation C.M. Byrd, J.N. Russell, Jr. (Naval Research Laboratory) Gallium nitride (GaN) is a wide band gap semiconductor with applications in high temperature, power and frequency optoelectronic devices. The surface chemistry of hydrogen on GaN affects growth rates and electronic passivation, while annealing temperatures impact both ohmic contacts and thermal stability. In this work, the preparation of a GaN(0001) thin film surface was investigated as a function of anneal temperature (300-1100K) using Auger electron spectroscopy (AES), electron energy loss spectroscopy (EELS), low energy electron diffraction (LEED), and temperature programmed desorption (TPD). After the GaN(0001) surface was sputter cleaned with nitrogen (N₂⁺) ions, N₂ desorption was observed at 950 K from embedded nitrogen and then above 1200 K from GaN decomposition. EELS and AES showed subtle changes as the anneal temperature increased, and the LEED pattern sharpened. TPD spectra were collected for a series of anneal temperatures. After annealing the surface and cooling to room temperature, the surface was dosed with D atoms. When heated at 1 K/s, D₂ thermal desorption was observed, but not ammonia or gallane. There were four D₂ thermal desorption peaks at 430, 600, 730 and 810 K, the appearance and relative intensities of which were related to whether the anneal occurred at, above, or below the embedded nitrogen desorption temperature. Correlation of the anneal temperature dependence of the D₂ thermal desorption with the EELS, AES, and LEED data aided in identifying the origins of the D₂ desorption states. This work demonstrates D₂ thermal desorption is very sensitive to the quality of the GaN(0001), and explains differences in hydrogen on GaN(0001) TPD results in the literature.
4:40 PM		SS+EM+SC-ThA-9 Metal/Semiconductor Phase Transition in CrN Grown by Molecular Beam Epitaxy and Scanning Tunneling Microscopy C. Constantin, M.B. Haider, A.R. Smith (Ohio University) Considerable interest has been of late in transition metal nitrides thin films/surfaces, which have both magnetic and electronic properties with potential applications in spintronics. CrN is a particularly interesting case, having a known correlation of structural and magnetic transition from a B₁ NaCl-paramagnetic to an orthorhombic-antiferromagnetic at T_Neel=273-286K¹. However,the reported electronic properties of CrN are controversial, and there has been no consensus whether the material is a metal or a semiconductor^2,3. In this study, CrN is grown on MgO(001) at a substrate temperature of 450°C by a novel molecular beam epitaxy method for obtaining smooth surfaces. Bulk measurements reveal that the films are single crystal, and stoichiometric. The 1x1 face-centered cubic (fcc) surface structure is clearly distinguishable as obtained (for the first time) in room temperature atomic resolution scanning tunneling microscopy. In addition to the atomic resolution, long-range topographic distortions [LTD] are also seen on the surface, as also observed for the semiconductor ScN (001)³ and other nonpolar III-V surfaces. LTDs are characteristic of semiconductor surfaces, and are related to localized charge accumulation from impurities. Resistivity was measured from 77 to 450K; metallic behavior is found up to 260K (in contrast with some earlier reports) and semiconductor behavior above 285K. The bandgap obtained from resistivity data,71±0.315meV, agrees with the tunneling spectroscopy of the surface which show a very small gap. Consistent results have now emerged in which CrN has a semiconductor-metal phase transition corresponding to its magnetic transition. ¹ A. Filippetti et.al, Phys. Rev. B 59, 7043 (1999) ²J. D. Browne et.al, Phys. Status Solidi 1, 715 (1970) ³P. S. Herle et.al, J. Solid State Chem. 134, 120 (1997) ⁴H. A. Al-Brithen et.al, submitted to Phys. Rev. B.