Papers

AVS2004 Session SS3-ThA: Surface Diffusion and Transport

Thursday, November 18, 2004 2:00 PM in Room 213B

Thursday Afternoon

Start	Invited?	Item
2:00 PM		SS3-ThA-1 Molecular Transport on the Nano-Second Timescale; Co/cu(001). G. Alexandrowicz, A.P. Jardine, H. Hedgeland (University of Cambridge, UK); P. Fouquet (Institut Laue-Langevin, France); S. Dworski, W. Allison, J. Ellis (University of Cambridge, UK) The measurement of surface diffusion on microscopic length scales presents a considerable experimental challenge. Measurements are typically restricted to specific systems which either diffuse sufficiently slowly for real time techniques, such as STM, to be applied or fast enough to be seen using Quasielastic Helium Atom Scattering (QHAS). Here, we present the first surface diffusion measurements from the unique new Cambridge ³He Spin Echo Spectrometer¹². The apparatus makes use of the ³He Spin Echo technique³ to improve upon existing QHAS resolution by three orders of magnitude, showing new insights into previously inaccessible experimental regimes. We have studied the prototypical molecular system; CO/Cu(100). Existing QHAS measurements on this system were very limited⁴ and the intense debate over the difference between potential energy surfaces extracted from the measurements and from first principle calculations⁵ was left unresolved. Our measurements show the system is diffusing well within the previously unmeasurable regime. Quasielastic broadenings were measured as function of momentum transfer and temperature and interpreted using MD simulations. The analysis reveals a Potential Energy Surface, illustrating the success and limitations of recent first principle calculations⁶, and serves as a bench mark for further theoretical improvements. footnote1 AP Jardine, P Fouquet , S Dworski, G Alexandrowicz, J Ellis, W Allison, To be published footnote2 S Dworski, G Alexandrowicz, P Fouquet, AP Jardine, W Allison, J Ellis. Rev. Sci. Inst. In press. 2004. footnote3 C Shmidt, D Scholz, U Spinola, M Dekivit, D Dubbers. Phys. Rev. Lett. 75(10):1919-1922 footnote4 AP Graham, JP Toennies, Surf. Sci. 427-428 (1999) footnote5 Q G, DA King, J. Chem. Phys. 114(2001) 1053 footnote6 P Fouquet, RA Olsen, EJ Baerends, J. Chem. Phys. 119(2003) 509-514.
2:20 PM		SS3-ThA-2 CO Diffusion on Cu(111): Effects of CO-CO Interactions B.V. Rao, R. Perry, A. Liu, L. Bartels (University of California, Riverside) Diffusion is the surfaces process that occurs most frequently and, thus, may arguably be regarded the most important one. With the advent of low-temperature scanning tunneling microscopy one is able to image and track the motion of individual atoms and molecules on surfaces for long periods of time¹. We probed the diffusion of CO molecules on Cu(111) at temperatures between 17 and 50 K. At 17K, molecular hops are infrequent and occur on the time-scale of hours, whereas at 50K the hopping frequency has reached >1 Hz. Measurements at various temperatures indicate that the diffusion of isolated CO molecules follows an Arrhenius behavior. Our investigation puts special attention on the effect of neighboring CO molecules on the hopping direction and frequency. Preliminary results suggest, that the CO-CO interaction enhances or decreases CO diffusion depending on the separation of the molecules. The distance dependence resembles the oscillation patter of the Cu(111) surface state, similar to Cu adatom interactions previously reported by Repp et al.². ¹Briner, B.G., Doering, M., Rust, H.-P., Bradshaw, A. M., Science 278 (1997) 257. ²Repp J., Moresco F., Meyer G., Rieder K.H., Hyldgaard P., Persson M., Phys. Rev. Lett., 85 (2000) 2981 .
2:40 PM		SS3-ThA-3 How Trace Amounts of S Enhances Self-Diffusion on Cu(111) W.L. Ling, K.F. McCarty, J. De la Figuera, K. Pohl, N.C. Bartelt (Sandia National Laboratories) We find that <0.01 monolayer (ML) of S enhances surface diffusion on Cu(111) by several orders of magnitude. Using low-energy electron microscopy (LEEM) and scanning tunneling microscopy (STM), we probe the mechanisms of this enhancement by monitoring the decay rate of island stacks and the ripening of 2D island arrays as a function of S coverage and temperature. The decay rate of islands in stacks increases roughly as the S coverage cubed, consistent with the proposal that the enhanced surface diffusion is due to the formation of Cu₃S₃ clusters¹. However, we also find a dramatic change in the mechanism of island-stack decay and the ripening of 2D island arrays as a function of S coverage. While island decay on clean Cu is diffusion limited², this is not the case in the presence of a small amount of S. However, at higher S coverages, the kinetics revert to diffusion limited. Similarly, how island arrays evolve is strongly influenced by the S coverage. For clean Cu and at high S coverages, large islands grow at the expense of nearby small islands. In contrast, local ripening does not occur with smaller amounts of S -- mass flows from islands to the steps bounding large terraces. We show that the strong dependence of the surface kinetics on S coverage can be quantitatively understood if the Cu diffusion at low S coverages is limited by the rate of reaction of S and Cu to form Cu₃S₃ clusters. This work was supported by the Office of Basic Energy Sciences, Division of Materials Sciences of the U.S. DOE under Contract No. DE-AC04-94AL85000. ¹P.J. Feibelman, Phys. Rev. Lett. 85, p. 606 (2000). ²M. Giesen & H. Ibach, Surf. Science 431, p. 109, (1999) .
3:00 PM		SS3-ThA-4 Diffusion of Ir and W Atoms on W(110) ^* G. Antczak, G. Ehrlich (University of Illinois at Urbana-Champaign) Using field ion microscopy we have carried out extensive measurements of the migration of Ir and W atoms on the W(110) plane at temperatures from 300 to 370 K. In addition to nearest-neighbor jumps in the <111> direction, we recognized three types of long jumps, which may contribute to diffusion: double jumps along <111>, vertical jumps in the <110> direction and the horizontal jumps in the <100> direction. From analysis of the distribution of displacements, constructed from at least 1200 observations, we have found that long jumps play a significant role in diffusion at temperature higher than 340 K. The temperature dependence of all types of jumps was measured to derive activation energies and prefactors for diffusivities. Long jumps proceed with a higher activation energy and a higher prefactor than do single jumps for tungsten as well as for iridium atoms. The rate of single jumps is diminished by contributions from long jumps. However, contrary to expectation, the presence of long jumps does not raise the prefactor for diffusivity above the usual value. Comparisons are made of Ir with W and simple models accounting for our finding are proposed. ^*Supported by the Department of Energy under Grant No. DEFG02-96ER45439 to the Materials Research Lab., and by the Petroleum Research Fund, under Grant ACS PRF No. 36919-AC5.
3:20 PM		SS3-ThA-5 Temperature-Dependent Thresholds for Ion-Stimulated Surface Diffusion: A Comparison of Si and Ge Substrates Z. Wang, E.G. Seebauer (University of Illinois at Urbana-Illinois) Ion-surface interactions at low energies (<100 eV) characterize an increasingly diverse array of material processing steps in ion beam assisted deposition (IBAD), plasma enhanced deposition, reactive ion etching (RIE), and other applications. The governing kinetic phenomena are often tacitly considered to lie at one of two poles: physical effects where momentum matching dominates, and chemical effects involving thermal activation of atomic bonds according to Arrhenius expressions. Here we report surface diffusion measurements demonstrating behavior that lies at neither pole. Optical second harmonic microscopy is used to image diffusion of indium adsorbed on Si(111) and Ge(111) under low-energy noble-gas ion bombardment. Both systems exhibit a surprising tradeoff between substrate temperature and the energy threshold at which ion influences become manifest. Molecular dynamics simulations suggest that the effect originates from changes in surface point defect concentrations. Instantaneous nonuniformities in net surface potential induced by thermal vibrations provide a mechanism by which ions can affect these concentrations nonlinearly. The effects generally increase the rate of mass transport across the surface. The simulations and experiments agree in finding that an important factor determining the magnitude of the tradeoff is the strength of the adsorbate-substrate bond, which is lower for In/Ge than for In/Si.
3:40 PM		SS3-ThA-6 LEEM Observation of Island Decay on Si(110) F. Watanabe, S. Kodambaka, W. Swiech, G.D. Cahill (University of Illinois at Urbana-Champaign) Laser texturing enables the observation of island decays on (1x1) high temperature phase of Si(110) surface in low energy electron microscopy by artificially creating large flat terraces. At temperatures above the phase transition (T = 790 ~ 980 Â°C), the decay of the island areas exhibits a nonlinear dependence on time, indicating the rate limiting process is surface diffusion. The aspect ratios of the islands during the decay show a weak temperature dependence. Following a methodology developed previously¹, we have obtained kinetic parameters involved in the island decay. The activation energy of mass transport and the free energy of kink formation have been determined to be 1.67 Â± 0.15 eV and 0.22 Â± 0.03 eV respectively. Based on this value for the kink energy and the equilibrium island shapes, we derive the angular dependence of the step energy and the step stiffness. ¹ S. Kodambaka, V. Petrova, S. V. Khare, D. D. Johnson, I. Petrov, and J. E. Greene, Surf. Sci. 513, 468 (2002).
4:00 PM		SS3-ThA-7 Effects of Si Deposition on Electromigration Induced Step Bunching on Si(111) B.J. Gibbons (The Ohio State University); J. Noffsinger (The University of Kansas); J.P. Pelz, C. Ebner (The Ohio State University) We have studied the effects of Si deposition on direct current (DC) heating induced step bunching on Si(111) using Si samples with spherical dimples ground into the surface to create a range of surface miscut. With no Si deposition, we observe the well-known behavior that only "step-down" current produces bunching in temperature "Regime I": (<950°C), while bunching in "Regime II": (1050°C-1250°C) only occurs for "step-up" current. But in contrast to a report by Métois et al. [Surf. Sci. 440 (1999) 407] we very clearly do not observe that net growth conditions reverses the current direction required for bunching in Regime II. This is not consistent with the proposal [S. Stoyanov, Surf. Sci. 416 (1998) 200] that the primary bunching mechanism in Regime II is due to increased step permeability. However we do observe that there is a strong reduction in the density of "crossing steps" close to zero net deposition/sublimation conditions, qualitatively consistent with the simultaneous bunching/debunching model of Kandel and Weeks [Phys. Rev. Lett. 74 (1995) 3632]. We will discuss on-going work to quantify this reduction and compare it with 1D analytic and 2D Monte Carlo models. By measuring areas of the dimpled samples with different miscut, we have also found that the average step bunch height increases roughly linearly with sample miscut, but does not depend significantly on Si deposition conditions. We are also comparing this observed bunch-height dependence on miscut with 1D and 2D models to evaluate which existing model can best explain a range of step-bunching behavior. Work supported by NSF Grant DMR-0074416.
4:20 PM		SS3-ThA-8 Pb/Si(111) Domain Boundary Fluctuations D.B. Dougherty (University of Maryland at College Park); W.G. Cullen, E.D. Williams (MRSEC) The techniques of fluctuation statistics are applied to a common 1D interface on solid surfaces: a domain boundary between two surface phases. We describe variable temperature STM observations of fluctuating boundaries between Si(111)-(1x1)-Pb and Si(111)-√3x√3R30-Pb¹. We establish the dominant mass transport mechanism for the fluctuations from the power-law growth in time of the temporal correlation function with an exponent of about 0.5. This behavior implies that fluctuations of the (1x1) boundaries occur via exchange of mass with a 2D adatom gas on the √3x√3 phase. This conclusion is corroborated by the extraction of the persistence probability from the same data, giving a persistence exponent of 0.75. This 2D exchange mechanism allows comparison of the nanoscale fluctuation kinetics with previous macroscopic surface science measurements. The 2D adatom gas has long been established for this system from the observation of zeroth-order desorption kinetics. Previous studies indicate a barrier for this rate-limiting exchange step of 2.3 eV². Our experimental correlation functions yield an effective activation barrier of about 1.6 eV. This provides a direct physical confirmation of the indirect deduction of 2D exchange of mass at the domain boundary. Finally, the autocorrelation function of the boundary fluctuations can also be determined and used to extract correlation lengths that compare well with the size of (1x1) domains. In addition, these lengths are compared with correlation lengths for fluctuating steps on clean metal surfaces. *Supported by UMD-NSF-MRSEC under DMR-00-80008. ¹ J. Slezak et al., Surf. Sci. 454 (2000) 584; O. Custance et al., Surf. Sci. 482 (2001) 878. ² M. Saitoh et al., Surf. Sci. 154 (1985) 394.
4:40 PM		SS3-ThA-9 Temperature-Dependence of Submonolayer Island Structure and Multilayer Growth Morphologies in Ag/Ag(111) Homoepitaxy M. Li, E. Cox, C. Chung, C. Ghosh, J. Evans, P.A. Thiel (Iowa State University) We have performed STM studies to analyze the irregular structure of submonolayer islands and the rough wedding-cake-like multilayer growth morphologies formed during Ag/Ag(111) homoepitaxy between 120-180K. Previous STM and surface scattering studies have probed only selected aspects of these features of film structure, which are expected due to restricted periphery diffusion (PD) and limited interlayer transport. However, no comprehensive real-space characterization of film morphology across this temperature regime has been available. KMC simulation of atomistic lattice-gas models is also viable in this regime where island separations are below ~100nm, and film morphology is not controlled by dislocations (in contrast to behavior at 300K). We have thus also developed a realistic atomistic model to describe the submonolayer growth of individual islands, and thereby quantified the observed transition with increasing temperature from dendritic shapes with triangular envelopes (due to anisotropy in incorporation at corners) to more isotropic fractal shapes (reflecting high edge and kink rounding barriers), and then to compact shapes (reflecting efficient PD). More complete modeling of the overall island nucleation and growth process revealed a transition to reversible island nucleation in this temperature range. Finally, by extending of our realistic model for submonolayer island formation to the multilayer regime incorporating negligible interlayer transport, we described quantitatively the observed kinetic roughening and wedding-cake-like growth morphologies.