Event
MSE Seminar - Prof. Raymond Phaneuf, University of Maryland
Friday, September 26, 2008
1:00 p.m.
Rm. 2108, Chemical and Nuclear Engineering Bldg.
Annette Mateus
301 405 5207
amateus@umd.edu
"Patterning for Directed Self-Organization and as a Probe of Interface Stability"*
The challenge posed by the drive toward nanotechnology is not simply fabrication of structures of nm-scale dimensions, but also to generate extremely high densities of these structures, with a controlled relative arrangement, and at a practical time-scale. Self-assembly/self-organization might seem the obvious answers, but a problem is how to direct these processes so that the structures which form spontaneously do so at controllable positions. In this seminar Ill discuss our recent work investigating directed self-organization, in which we use a topographical template, produced using a combination of lithography and etching, to direct fast self-organization processes; these include epitaxial growth, and diffusion/sublimation during annealing. At the same time these studies allow us to address a crucial problem in Materials science: Under what conditions and over what length scales is a moving interface stable?
As a first illustrative example Ill discuss the behavior of artificially-patterned corrugations during molecular beam epitaxial growth of GaAs(001) and GaAs/AlAs multilayer structures. Mounds form on patterned and unpatterned substrates during the early stages of growth on both patterned and unpatterned substrates, but disappear as growth continues. Atomic force microscopy (AFM) images subsequent to growth under standard conditions for GaAs, T = 585ºC, 1 ML/s, PAs2/PGa = 10:1 reveal a transient instability, in which evolution of patterned structures in this system is nonmonotonic with respect to length scale and time: beneath a critical length scale growth causes decay in the amplitude of the structures, but above this value instead an amplification. This critical pattern period increases monotonically with the overall thickness grown increases. Growth at lower temperatures (~500ºC) produces a mode of topographical evolution which is qualitatively different: ridges build up outside of lithographically defined pits. Numerical simulations based upon continuum height equations rule out the KPZ model for the instability, but are consistent with the conserved versions of this equation proposed by Sun, Guo and Grant at ~600ºC, and by Lai and Das Sarma at ~500ºC. I discuss possible interpretations of the nonlinear term in these equations, and the apparent switch in sign near 530ºC.
As a second example, Ill discuss the evolution of patterned structures on stepped Si(111), in which sublimation and diffusion compete in determining how the interference of those steps resulting from lithographic patterning (loop steps) and those resulting from a slight misorientation of the average surface from (111) (vicinal steps). The evolution of the topography during high temperature annealing also in this case shows an interesting dependence on the lateral period of our cylindrical pit arrays. At the lower end of the range we have explored, the vicinal steps interfere with the pits to form nearly straight step bunches quickly; the height of bunches relaxes much more slowly. At the upper end of the range of pattern periods the relaxation of the lateral corrugation, or waviness of the step bunches, and the bunch heights themselves converges to a common relaxation time which scales as the inverse fourth power of the spatial period. We find that numerical simulations based upon step creation free energies and step-step interactions, along with sublimation allow us to reproduce this intriguing set of observations.
Each of these systems demonstrate how the competition between kinetic and thermodynamic effects can be used in the fast generation of novel templates which in turn can be exploited in directing self-assembly of technologically important structures such as quantum dots, for applications in photonics, optoelectronics and quantum information systems.
*Work supported by the National Science Foundation, the University of Maryland MRSEC and The Laboratory for Physical Sciences.
