TEM observations of the mechanism of delamination of chromium films from silicon substrates
- PDF / 2,388,190 Bytes
- 8 Pages / 576 x 792 pts Page_size
- 53 Downloads / 170 Views
We have observed the complete delamination of polycrystalline chromium films from single crystal silicon substrates during deposition due to the formation of high internal stresses. These intrinsic stresses can give rise to interfacial defects which assist in the separation of the film from the substrate. Stresses in the film are balanced by stresses in the substrate, which cause mechanical failure in the substrate near the interface. Extensive arrays of dislocations and cracking of the substrate have been observed. We find that the delamination of the films from the substrate is initiated by the formation of damage in the substrate, rather than to the film or the interface.
I. INTRODUCTION Large internal stresses can be induced when thin films are formed by deposition from the vapor phase onto a substrate. 12 The stresses typically do not cause plastic deformation of the film because the small dimensions of the specimens prevent dislocation multiplication from occurring.1'3 The stresses, however, may lead to delamination of the film from the substrate and thereby cause catastrophic failure of the film-substrate system. Various studies have shown that the stress in a tensile loaded film decreases rapidly with the increasing film thickness and yet the failure of the system occurs when a welldefined thickness is exceeded. The results of Ng4 and Namaroff5 show that the stresses in the film decrease rapidly with increasing film thickness and beyond a 'critical thickness' the stress becomes constant. The stresses in the film become constant as a consequence of the failure of the film substrate composite after the critical thickness has been reached, and the film failure occurs while the stress in the film is reducing, rather than increasing. The occurrence of this type of failure is therefore something of a puzzle. Although it is well established that pre-existing cracks traversing the film may propagate under these conditions,6 the mechanism by which a through-film crack can be nucleated remains an open question, given that it occurs while the film thickness is increasing and the stress is decreasing. We have demonstrated in other papers7'8 that the stress in a thin film is not homogeneous and this may have some effect upon the mechanism of the failures. In this paper we report on experiments that were undertaken to elucidate the mechanism of failure initiation. II. EXPERIMENTAL In the present studies chromium films with different thicknesses were thermally deposited onto silicon single J. Mater. Res., Vol. 7, No. 2, Feb 1992
http://journals.cambridge.org
Downloaded: 25 Mar 2015
crystal wafers of (111) and (100) surface orientations. The substrates were nominally at room temperature for all depositions. A deposition rate of 0.15 nm per second was used for all the coatings and a range of coating thicknesses from 5 to 500 nm was prepared. Transmission electron microscopy specimens were prepared using a modification of the apparatus suggested by Booker and Stickler.9 Disk specimens were trepanned from the whole wafers
Data Loading...
