Baixe zanotto 1993 e outras Notas de estudo em PDF para Engenharia de Produção, somente na Docsity! EXPERIMENTAL STUDIES OF SURFACE NUCLEATION AND . CRYSTALLIZATION OF GLASSES Edgar D. Zanotto DEMa - Federal University at Sao Carlos 13560 - Sao Carlos - SP, Brazil 1. INTRODUCTION Surface crystallization is much more frequent than internal crystallization but, paradoxically, has been much less studied. One possible reason for the lack of detailed studies on the subject is the inherent difficulty of controlling and characterizing the surface properties of glass. From a technological point of view, the establishment of a firm understanding and the ability to control the kinetics of surface nucleation and growth in glasses may lead to the development of a wide variety of information devices for optical memories and related applications [1]. Additionally, controlled surface crystallization has proved to be quite an effective method of enhancing the strength of glass [2). The mechanisms of surface crystallization are still a matter of controversy . Several factors which promote free surface crystallization have been alluded in the past and include: cracks or irregularities in: the surface, alteration due to chemical reactions with the atmosphere, differences between bulk and surface composition, contamination with solid particles, etc. Most studies carried out so far, however, are qualitative and the relative importance of these potential origins of surface crystallization in various glasses have not been determined. The relevance of surface crystallization was recognized by the devitrification committee (TC7) of the ICG which started a multinational research effort on the subject in 1989. The objective of this paper is to review and analyze most of the relevant papers on surface crystallization To the extent authorized under the laws of the United States of America, all copyright interests in this publication are the pro of The American Ceramic Society. Any duplication, reproduction, or republication of this publication or any part thereof, vidhou the express written consent of The American Ceramic Society or fee paid to the Copyright Clearance Center, is prohibited. Nucleation and Crystallization in Glasses and Liquids 65 and also to present a few results generated by the TC7 Committee on the subject. The goal is to provide a firmer understanding of the causes for the predilection of glasses to crystallize from the surface rather than in the bulk. In addition, the question of whether or not there is an intrinsic preference for crystallization at the free surfaces will be addressed. 2 - SURFACE NUCLEATION 2.l. Effects of Surface Quality, Chemical Composition and Temperature Only a few comprehensive studies were devoted to heterogeneous nucleation in glass, on crucible walls and on metallic particles ,see for instance [3-6], and the subject certainly deserves further attention. In this article ,however, I deal only with isothermal studies om free surface crystallization. Starting in the early thirties, a number of papers of qualitative nature were published concerning the surface devitrification of several glasses. Observations of crystallization around internal bubbles were specially important because their surfaces are in principle much cleaner than the external glass surfaces. These papers were summarized recently [7] and are listed in Table 1. Table 1. Surface crystallization studies. | | Author Year Glass crystals on bubbles ? Morey 30 Soda-lime-silica no Scott 61 Na20.2Si0Oz no Klingsberg 67 BaO-Al203-TiO2z-Si0Qz no Ernsberger 66 Soda-lime-silica only on a few bubbles Mattox 67 CaO-A1203-B203 in those with solid particles Hishinuma 86 PbO.SiOz, Na20.2SiOoz | no DN Despite some controversial results, the majority of papers indicate that only those bubbles contaminated with solid particles are preferential nucleation sites, Most studies provided no evidences for nucleation on (clean) bubble surfaces. It should also be emphasized that it is generally accepted by glass technologists, that glasses with dirty surfaces crystallize much easier than those with clean surfaces. 66 Nucleation and Crystallization in Glasses and Liquids measured. The identity of the x-crystals and why the surface nucleation rates are sufficiently slow to allow for their measurement are not known at present. Table 2 summarizes the crystallite densities reported by several authors for different glasses. Table 2. Surface crystallite densities: 2 Author Year Treatment Glass Ns (mm *) Ground surface McMillan 82 840ºc / 10min ZAS 3:10* Polished ; McMillan 82 840 / 10min ZAS 1:10 zanotto 86 820 / 0-4h CMS2 8.10" zanotto 90 700 / 0-24h NCS 4105 Zanotto 90 720 / 0-26h NCS 3:10 Kalinina 90 140-860/450h M2A2S5 1105 Muller 90 900 /10 min M2A2S5 5:10 As-received Zanotto 90 750-800 / 0-30h Float 0-3-10"* Fractured 2 Muller 90 900-980 /0-80min M2A258S5 2:10 Fire-polished zganotto 90 730 / 0-7.5h NCS Zero Yamane 90 | GM 30870 "lowest" * range of values for different crystal phases. Summarizing , the results of Table 2 and Figure 2 show that the surface crystallite densities strongly depend on the surface condition. The smoother (and presumably the cleaner) the surface the smaller is Ns. Some fire polished specimens and some (as-received) float glasses, with pristine surfaces, . did, not crystallize at all, while Ns was almost 10 / mm for some ground and for mechanically polished specimens. The results of [7-12] also indicate that, in most cases, surface nucleation finishes in the early stages of the transformation, and saturation occurs before any measurement of the heterogeneous nucleation rate can be made, in a wide temperature range, from Ty to Tm (the x-phase was an exception). A possible explanation for the apparent lack of temperature dependence of the surface nucleation rates is provided by the classical theory, which is capable of describing reasonably well the temperature dependence of homogeneous nucleation rates in glasses Nucleation and Crystallization in Glasses and Liquids 69 [6]. If one uses the classical equation, and allows the reduced interfacial energy, à = & Na “vu! ?/AHm (these parameters were defined elsewhere [6]) to vary from typical values for homogeneous nucleation (0.4< a <0.6) to the expected small values for heterogeneous nucleation (a < 0.3), it is seen that the smaller the value of a, the flatter, and consequently the less dependent on temperature, is the nucleation curve. Finally, the absence of crystals on some clean, pristine, surfaces and also on most bubble surfaces provide clear evidences that the free surfaces are not preferential nucleation sites, in agreement with the thermodynamic arguments of Unlmann [17): 2.2. Effects of the Atmosphere The effects of the surrounding atmosphere on the crystal growth rates are reasonably well documented [8]. However, much less is known about the effects of atmosphere on surface nucleation densities. A detailed study was performed by Takahashi and Sakaino [13] who measured the crystal growth rates and also the crystallite densities on the surface of a Na20.2SiO2 glass (T9g-=470'c) for a series of heat treatments, from 200 to 500“C in N2, COz and water vapor. The number of crystals increased in the following order; in dry N2, in Hz0 vapor, being much larger in dry CO2, with a maximum at 250-300ºCc. They demonstrated that NazCOs crystals formed at the glass surface provided active sites for crystallization of sodium disilicate crystals at the development temperature (600'C for 20min). Partridge and McMillan [14] have shown that atmospheres containing reduced oxygen and water contents inhibit surface nucleation of a Zn0-Al203-SiOz glass. 3. CRYSTAL GROWTH In the vast majority of experimental studies of crystal growth in glasses, nucleation spontaneously initiates or is induced at the glass surface and the time evolution of the crystallized layer growing towards the specimen center is determined by microscopy techniques. Thus, the growth Kinetics refer to molecular rearrangements in the bulk of the glass. Recently, a different experiment was carried out with an almost stoichiometric cordierite glass supplied by 70 Nucleation and Crystallization in Glasses and Liquids shott Class for the TC 7 Committee members. The full set of results will be presented in a forthcoming congress [15], but a relevant summary is given bellow. Three types of growth kinetics were determined, i.e, of: i) crystals which nucleated at the surface and grew in the bulk, ii) surface nucleated crystals which grew two- dimensionally on the surfaces, in the early stages of the transformation, and àãii) internal crystals, which nucleated on foreign particles, and grew in the bulk. The three types of growth rates vwere equal, within experimental error, implying that the surface and bulk diffusion processes for growth, and inferentially for nucleation, are similar. In a previous paper [7] 1 suggested that the bulk and surface diffusional mechanisms might be different. That suggestion was based on molecular dynamics calculations and also on the observed differences between the activation enthalpies for crystal growth, Hc, and viscous flow, Hn, in a float glass which crystallized from the surface. This matter is clarified in [18] were we demonstrate that the differences in Hc and Hm are genuine for most glasses and not only for those which only nucleate at the surfaces. 4. SUMMARY surface nucleation saturates in the early stages of crystallization for most glasses. The nucleation rates are too fast to be measured in wide temperature ranges, from Tg to Tm. This insensitivity to temperature is due to the small values of surface energy. | Nucleation is not observed on pristine, clean, surfaces, as predicted by simple thermodynamic arguments. Surface nucleation is mainly due to impurity particles whose number is inversely proportional to the degree of surface perfection and cleanliness. “There is much scope for further work on surface crystallization. Acknowledgements TI thank R. Basso, A.V. Cardoso, E.B. Ferreira, E. Wittman, N. Mora and E.C. Ziemath for performing most experiments and for useful discussions. I also thank M.Weinberg and the TC7 Committee colleagues, R.Muller and W.Pannhorst, for useful exchange of information. Financial help by FAPESP, contract no. 85/0725-3 and by PADCT/CNPq contract no. 620058/91-9 are appreciated. Nucleation and Crystallization in Glasses and Liquids nf