Papers by Keyword: III-V

Abstract: Atomic layer deposition (ALD) of high dielectric constant (high-k) materials for ULSI technologies is now widely adopted in Si-based CMOS production. Extending the scaling of integrated circuit technology has now resulted in the investigation of transistors incorporating alternative channel materials, such as III-V compounds. The control of the interfacial chemistry between a high-k dielectric and III-V materials presents a formidable challenge compared to that surmounted by Si-based technologies. The bonding configuration is obviously more complicated for a compound semiconductor, and thus an enhanced propensity to form interfacial defects is anticipated, as well as the need for surface passivation methods to mitigate such defects. In this work, we outline our recent results using in-situ methods to study the ALD high-k/III-V interface. We begin by briefly summarizing our results for III-As compounds, and then further discuss recent work on III-P and III-Sb compounds. While arsenides are under consideration for nMOS devices, antimonides are of interest for pMOS. InP is under consideration for quantum well channel MOS structures in order to serve as a better nMOS channel interface. In all cases, a high-k dielectric interface is employed to limit off-state tunneling current leakage.

Abstract: III-V compound semiconductors have been recognized among the potential options for continuing transistor power-performance scaling owing to their ultra high charge carrier mobility. In order to realize their potential in high performance and lower-power digital logic applications, there must be strong gate control and a high Ion-Ioff ratio, achieved by integrating a stable, ultra thin high-K dielectric between the semiconductor and the gate [1, 2]. Unlike Si, which has long benefited from its very stable native oxide, III-V materials suffer from their poor native oxides that cause charge traps and Fermi level pinning at the semiconductor-oxide interface. Attempts to deposit high-K directly on III-V often produce MIS structures with fast surface state and CV instability [3].

Abstract: We have investigated the hydrogenation of the zinc acceptor in GaP and InP, and of the phosphorus acceptor in ZnTe, by computer modeling. We used a density-functional supercell code and pseudopotentials to deal with the core electrons. However zinc 3d electrons were explicitly taken to be valence electrons. We have determined the relaxed atomic geometry for seven hydrogen sites. We have found that, in the lowest total energy configuration, hydrogen sits in a bond centered position between zinc and arsenic atoms in all GaP, InP and ZnTe semiconductors and is bonded to the phosphorus atom. We found metastable states, by 0.4, 0.4 and 0.5 eV, for structures where H is antibonding to the phosphorus atom for GaP, InP and ZnTe, respectively. The calculated local vibrational modes (LVM) for the bond-centered configuration agree, within 1%, with the experimental values of 2379.0 cm-1 for GaP:Zn-H, 2287.7 cm-1 for InP:Zn-H and 2193 cm-1 for ZnTe:P-H. The isotopic shift due to the replacement of deuterium by hydrogen is reproduced by less than 2.5% using experimental data. The decrease in the LVM when going from GaP to ZnTe, as the perfect bond length increases, is also well-reproduced. A wag mode at 496 cm-1 and lower LVM, a doublet at 329 cm-1 and a singlet at 242 cm-1, are predicted for P-H in ZnTe.