Mechanism of Arsine Adsorption on the Gallium-Rich GaAs(001)-(4 × 2) Surface Qiang Fu, ² Lian Li, ‡ Connie H. Li, ² Michael J. Begarney, ² Daniel C. Law, ² and Robert F. Hicks* ,² Chemical Engineering Department, UniVersity of California, Los Angeles, California 90095-1592, and Department of Physics and Laboratory for Surface Study, UniVersity of Wisconsin, Milwaukee, Wisconsin 53201 ReceiVed: February 14, 2000; In Final Form: April 6, 2000 The kinetics and mechanism of arsine adsorption on the (4 × 2) surface of gallium arsenide (001) has been studied by scanning tunneling microscopy, infrared spectroscopy, and ab initio quantum chemistry calculations. Arsine forms a dative bond to a gallium dimer. Then, this species either desorbs from the surface or decomposes to an AsH 2 or AsH fragment with hydrogen transfer to an arsenic site. Finally, desorption of hydrogen leaves arsenic dimers on the surface. The energy barriers for arsine desorption and dissociation into AsH 2 are estimated to be 9.3 and 16.5 kcal/mol, respectively. Gallium hydride is not produced upon dissociation of AsH 3 because this process is not energetically favorable. 1. Introduction Compound semiconductor devices have found many applica- tions in the fast-growing telecommunications industry, including uses as fiber-optic transceivers, broadband cable modems, and satellite photovoltaics. The active layers in these devices are grown on the (001) plane of gallium arsenide by metalorganic vapor-phase epitaxy (MOVPE). 1-10 In this process, volatile precursors of the group III and group V elements in a hydrogen carrier gas are fed into a reactor containing a substrate heated to between 773 and 923 K. 1-3 The precursors adsorb and decompose on the wafer surface, depositing a single-crystal film that is lattice-matched to the GaAs substrate. The film surface plays a crucial role in this process, because it mediates the decomposition of the group III and group V reagents. 4-13 It is essential to understand the heterogeneous decomposition reac- tions because they affect the growth rate, film composition, film morphology, and dopant profiles, all of which impact device performance. Arsine is widely used as the group V source during MOVPE, and its adsorption on GaAs(001) surfaces has been studied by several groups. 14-17 Arsine undergoes dissociative adsorption at temperatures as low as 140 K. Upon an increase in the substrate temperature from 140 to 450 K, most of the arsine desorbs from the surface, while a small fraction decomposes to form arsenic dimers. White and co-workers 15,16 found that the hydrogen from arsine transfers to As and Ga sites, generating both As-H and Ga-H stretching vibrations, as shown in their high-resolution electron-energy-loss spectra. However, this point has been disputed by Qi et al., 17 who observed only As-H vibrational bands by infrared spectroscopy following arsine dosing of GaAs(001). In addition to this discrepancy, there are several unresolved issues regarding the arsine adsorption mechanism. In particular, the adsorbed intermediates have not been identified, and except for estimates of the sticking coefficient, little information is available on the reaction kinetics. In this paper, we report on a study of arsine adsorption and decomposition on the GaAs(001)-(4 × 2) reconstructed surface using scanning tunneling microscopy, internal-reflection infrared spectroscopy, and ab initio calculations with density functional theory. We have identified each of the reaction intermediates and, with the aid of theory, determined their relative stabilities on the semiconductor surface. In addition, a kinetic model for the dissociative adsorption of arsine is presented. 2. Experimental and Theoretical Methods The samples were prepared by growing GaAs films, 0.5 µm thick, on gallium arsenide (001) substrates in an MOVPE reactor. 9 After growth, samples were transferred directly into an ultrahigh vacuum (UHV) chamber with a base pressure of 2.0 × 10 -10 torr. The GaAs(001) crystals were annealed at 793 K for 30 min to obtain a clean and well-ordered gallium-rich (4 × 2)/c(8 × 2) reconstruction, as verified by low-energy electron diffraction. The surface composition was measured with an X-ray photoelectron spectrometer (PHI 5000), equipped with a hemispherical analyzer. Scanning tunneling micrographs were obtained at a sample bias of -2.0 to -4.0 V and a tunneling current of 0.5 nA. 3 Arsine was introduced into the chamber at 5 × 10 -7 torr through a precision leak valve. Dosing was continued for up to 65 min to ensure that a constant coverage was obtained. A series of spectra was collected before and during arsine adsorption at 8 cm -1 resolution and with co-addition of 1024 scans. These spectra were collected by multiple internal reflection through GaAs(001) crystals that were cut into trapezoids 10-mm-wide by 40-mm-long by 0.64-mm-thick with 45° bevels at each end. Thirty-one reflections occurred off the front face of the crystals. The long axis of the crystals used in the experiments was parallel to the [110] direction. The reflectance spectra presented here were obtained by taking the ratio of the spectra recorded at saturation coverage to that recorded before dosing. During the arsine adsorption experiments, all of the filaments in the chamber were turned off. It was important to do this, because otherwise, the arsine molecule dissociated on the filament, causing AsH x and H fragments to adsorb onto the surface and yield anomalous results. * Author to whom correspondence should be addressed. E-mail: rhicks@ucla.edu. Fax: 310-206-4107. ² University of California. ‡ University of Wisconsin. 5595 J. Phys. Chem. B 2000, 104, 5595-5602 10.1021/jp0005827 CCC: $19.00 © 2000 American Chemical Society Published on Web 05/18/2000