AN INVESTIGATION OF THE OXIDATION OF LASER AND FURNACE-ANNEALED SPUTTER-DEPOSITED NITI THIN FILMS USING REFLECTIVITY MEASUREMENTS

F. Khelfaoui,Y. Bellouard,T. Gessmann,Xi Wang,J. Vlassak,M. Hafez
2004-01-01
Abstract:Shape Memory Alloy (SMA) thin films are of particular interest as micro-actuators. Thus far, their use in microsystems is still limited due to numerous technological issues that have yet to be overcome. In particular, film oxidation during annealing has to be carefully controlled to preserve the material properties. In Ni-Ti thin films, the oxide layer can make up a significant fraction of the total film thickness. Native oxides like TiO2 rapidly form on Ni-Ti alloy surfaces in the presence of oxygen. The oxide layer thickness depends on annealing conditions (i.e. temperature, time and atmosphere). During the oxidation process, compositional changes may occur in the Ni-Ti film. Hence, the properties of the film can be significantly altered and the martensitic transformation reduced or even inhibited. In this paper, we present a study of oxide growth on Ni-Ti thin films. Oxide thickness measurements were performed using optical reflectometry. Thin films were deposited on fused quartz substrates using magnetron sputtering in a UHV chamber with confocal sputter guns. Films of three different compositions near equiatomic and Ti-rich were considered. The asdeposited amorphous films were heat treated using either furnace or laser annealing. We compare the two methods in terms of oxide growth kinetics and discuss the influence of film composition. INTRODUCTION Thanks to their shape memory and superelastic properties, Ni-Ti thin films are of particular interest for application in microsystems [1]. Shape memory and superelastic properties are only observed in crystalline films that undergo a thermoelastic martensitic phase transformation. Sputter-deposited films are usually deposited at room temperature and are amorphous as-deposited. Heat treatments to crystallize the films can be performed in-situ in the deposition chamber or in a furnace under high vacuum conditions to prevent oxidation. To fully exploit thin-film SMA properties and, in particular, to integrate a two-way shape memory effect, it would be desirable to have shape memory in certain areas of the film and not in others. Bellouard et al. [2] have demonstrated that this can be achieved by selectively annealing certain areas of the film. Among the various methods that have been proposed, laser annealing [2, 3] is the most promising: it is a flexible process, it is easy to implement and unlike annealing techniques based on Joule heating it does not require any particular geometry. Heating is done locally, at the point where the laser beam is focused. Depending on how sophisticated the focusing optic is, this zone can be almost as small as the dimension of the laser wavelength. Although the heating time is several orders of magnitude shorter than for furnace annealing, the laser annealing process still results in the formation of an oxide
What problem does this paper attempt to address?