| AVS 55th International Symposium & Exhibition | |
| Surface Science | Friday Sessions |
| Session SS+EM+NC-FrM |
| Session: | Semiconductor Surfaces |
| Presenter: | J.B. Clemens, University of California, San Diego |
| Authors: | J.B. Clemens, University of California, San Diego T. Song, University of California, San Diego A.C. Kummel, University of California, San Diego |
| Correspondent: | Click to Email |
Atomic Layer Deposition requires the substrate to have a chemical passivation/termination layer consisting of reactive groups that initiate the ALD reaction. A suitable passivation/termination layer would have ligands that mimic the surface during growth, such as hydroxyl (OH-).1,2 Scanning tunneling microscopy was used to study the initial bonding configuration of hydroxyl onto the group III-rich InAs(001)-(4x2)/c(8x2) surface, which is almost identical to the InGaAs(001)-(4x2)/c(8x2) surface. These surfaces are more resistant to oxidation than group V-rich surfaces, which is true of many III-arsenide semiconductors, and therefore is a better starting template for ALD.3 Aqueous (30%) hydrogen peroxide vapour is used as the OH source. After annealing, the surface reaction forms well-ordered interfaces that terminate at about one ML indicating that this process is self-limiting. Substrate lattice disruption is minimal following OH desorption after annealing at 350° C. If pure water vapour is used as a control dose, less surface reaction occurs and it centers at surface defect sites. Density functional theory was used to model the interaction of InGaAs(001)-(4x2) with OH, H, and H2O. Energies and kinetics of adsorption and desorption of OH, H, and H2O are presented, which compares the stability of the HOOH/H2O vs the H2O only termination/passivation methods. DFT results show that the OH passivation method using HOOH is stable at high temperatures that are typically found under ALD growth conditions, while the passivation method using only water is not. The electronic structure was probed using scanning tunneling spectroscopy. On the clean as-prepared substrates, both n- and p-type InAs(001)-(4x2)/c(8x2) show n-type behavior, consistent with literature.4,5 Upon OH termination, both surfaces exhibit n-type behavior, with the Fermi level about 0.1 eV below the CB minimum. This shows no evidence for midgap Fermi level pinning, suggesting that this method has potential for high-κ gate oxide ALD on III-V semiconductor surfaces.
1 K. Kukli, et.al., J. Appl. Phys., 92, 1833 (2002).
2 J. Aarik, et.al., Appl. Surf. Sci., 161, 385 (2000).
3 D. Winn, et.al., J. Chem. Phys., 127, 134705 (2007).
4 L. Olsson, et.al., Phys. Rev. Lett., 76, 3626 (1996).
5 P. De Padova, et.al., Surf. Sci., 482-485, 587 (2001).