Effect of composition and environment on crack self-healing in oxide glasses

Success in designing self-healing materials has important consequences for material safety, product performance and enhanced fatigue lifetime. For example, solid oxide electrolyser and fuel cells development is hindered by a lack of long-term durability of seals caused by cracking induced by thermal expansion mismatch. In these applications, self healing glasses need to be developed to unsure a reasonable durability of the seal. It is consequently interesting to better understand and analyse the self healing behaviour of cracks in glass sealant, and more generally in oxide glasses. Although the process of crack opening has been widely studied, the driving forces controlling crack healing are not clearly identified [1,2]. Viscous flow is obviously an important parameter but environment has also been observed to play a crucial role. And it is also evident that residual stresses cannot be ignored. In this study we analyse by micro-indentation the healing of cracks generated in different silicate and borosilicate glasses. Healing thermal treatment is performed at temperatures both above and below Tg in humid and dry environment. Basically, we analyse the shape and the length of crack as a function of treatment's time. Our results confirm that temperature and thus viscosity plays a major role. But although thermal treatments were realized at the same viscosities for the different glasses and under the same atmosphere, our results clearly evidence that the healing rate depends on the glass composition, being faster in the borosilicate glass. Our results also evidence that water vapour enhanced cracks healing both in silicate and borosilicate glasses. In the case of borosilicate glass in humid environment, almost complete healing is observed even 20°C below Tg. For thermal treatment realized above Tg, we can observe a change in the shape of the crack, with a spheroidization. This leads us to the conclusion, that surface tension can not be neglected. In this perspective, we develop a specific device allowing us to measure surface tension. [1] B.A. Wilson, E.D. Case, J. of Materials Science 32 (1997) 3163-3175 [2] P. Hrma, W.T. Han, A.R. Cooper, J. of Non-Crystalline Solids 102 (1988) 88-94 [3] K. Kese, M. Tehler, B. Bergman, J. of the Eur. Cer. Soc. 26 (2006) 1003-101