| AVS 54th International Symposium | |
| Surface Science | Friday Sessions |
| Session SS1-FrM |
| Session: | Surface Dynamics |
| Presenter: | K. Mitchell, University of British Columbia, Canada |
| Authors: | K. Mitchell, University of British Columbia, Canada A. Yazdi, University of British Columbia, Canada N. Ingle, University of British Columbia, Canada T. Tiedje, University of British Columbia, Canada |
| Correspondent: | Click to Email |
Under bombardment by a rastered 30keV Ga+ ion beam, a flat epitaxially grown Au surface is found to exhibit the well known sputter ripple instability as first observed by Bradley and Harper.1 These ripples exhibit a characteristic lateral length scale on the order of 100nm and a short-range RMS saturation height on the order of 10nm after receiving a fluence of approximately 1500 ions/nm2. The starting surface is grown under UHV at 350°C on a freshly cleaved mica substrate resulting in a 0.5µm film with flat areas up to 8µm2 having an RMS roughness of 0.5nm. Accurate topography data are gathered using ex situ AFM on areas exposed to increasing ion fluence to track the increase in roughness associated with the pattern formation, while in situ SEM imaging is used to observe the dynamics before and after the ripples have formed. These experimental data are compared to 2D numerical solutions of a non-linear partial differential equation which captures the essential features of height saturation, characteristic length scale and parabolic ripple shape. A state of the art fourth-order "exponential time differencing" method as perfected by Kassam and Trefethen2 is used to advance the solution in time while high accuracy spectral methods are used to compute spatial derivatives. The equation is similar to the chaotic Kuramoto-Sivashinsky equation, but with an additional higher order non-linear term as derived by Castro and Cuerno.3 Like the Kuramoto-Sivashinsky equation, linear terms of competing stability set the initial characteristic length and growth rate, while the non-linear terms are responsible for the height saturation and the surface shape. The additional non-linear term causes an increase in the characteristic time, length and height scales after saturation has occurred. By adjusting the strength of the new non-linear term, the solutions can be tuned to match the morphology of the experimentally observed surface. The resulting equation parameters are then able to give information about the physical constants involved in the experiment.
1 Bradley, Harper, J. Vac. Sci. Technol. A 6,2390 (1988)
2 Kassam, Trefethen, SIAM J. Sci. Comp. 26, 1214 (2006)
3 Castro, Cuerno Phys. Rev. Lett. 94, 016102 (2005)