Speaker
Description
Oxide films are customarily manufactured using physical vapor deposition (PVD) at or near room temperature. In such conditions, the covalent bonds formed by oxide atoms (or order a few eV's) cannot be broken by thermal agitation; atoms only rearrange during the impacts of deposited particles. This process hampers the formation of crystalline domains and thus facilitates the growth of homogeneous and amorphous, transparent films. These films, in sharp contrast with bulk glasses, have never experienced or even approached any thermal equilibrium state. They present unusual, protocol-dependent, physical properties, which implies that their microstructures also differ from those of usual glasses and are sensitive to the details of the growth process. How can the film microstructure be characterized? How does it depend on the details of the deposition process? How does it determine macroscopic physical properties? These questions are open, yet hold the key to advancing our ability to control the film physical properties, and in particular, the mechanical losses limiting the resolution of gravitational wave detectors.
In this talk, I will summarize a recent study [1] in which we constructed a numerical model of low-T oxide growth, tested it against experimental data, and analyzed the resulting film microstructure. Our aim was to identify the physical origin of large, anisotropic, in-situ stresses (a few hundreds of MPa's) commonly found in the films produced by e-beam vaporization of silica. Realizing that deposition may involve compound particles enabled us to access steady film growth for the first time using numerical simulations. Microstructural analysis then revealed that the observed stress anisotropy is caused by the slight (a few percents, in line with ion-beam analysis data) oxygen-deficiency of the films. I will emphasize that: (i) low-T film growth is amenable to numerical modelling because thermally activated processes are irrelevant, so that only the series of impact events, during which atoms rearrange, need to be simulated; (ii) however, this effort depends on assumptions concerning the nature (stoichiometry, charge) and speed of the particles impacting the films, which remain largely unknown.
[1] S. Gelin, D. Poinot, S. Chatel, P. J. Calba, and A. Lemaitre. Microstructural origin of compressive in situ stresses in electron-gun-evaporated silica thin films. Phys. Rev. Materials, 3(5):055608, (2019).
