Baixe zanotto 1994 e outras Notas de estudo em PDF para Engenharia de Produção, somente na Docsity!
Chimica Chronica, New Series, 23, 3-17 (1994) CRYSTALLIZATION OF GLASS: A TEN YEAR PERSPECTIVE 1993 "Vittorio Gottardi Prize" Lecture Edgar Dutra Zanotto DEPARTMENT OF MATERIALS ENGINEERING FEDERAL UNIVERSITY OF SAO CARLOS k 13565-905, São Carlos, SP- Brazil This paper reviews four aspects of glass crystallization: i) the effects of liquid-liquid phase separation on crystal nucleation and growth; ii) the validity of the classical nucleation theory; iii) some trends observed on homogeneous and heterogeneous nucleation in oxide glasses and; iv) the state-of-art on surface crystallization kinetics. INTRODUCTION First of all I should say that it is a distinct pleasure to deliver this review lecture to such knowledgeable audience, looking at the magnificent scenery of the Greek sea (it can be seen from the lecture theater!). In this occasion, I decided to review my own work, carried out jointly with several colleges and graduate students in the past ten years, because this is a unique opportunity. However, I would like to point out that most of the research described here has been inspired in the earlier work of Russian scientists (V. Fokin, A. Kalinina and V. Filipovich), the Bulgarian group (I. Gutsow and co-workers), the Sheffield group (P. James and students), the Arizona team (M. Weinberg and D. Uhlmann) and my Brazilian colleagues (A. Craievich and E. Meyer). I dedicate this talk to these distinguished scientists. 1 learned much from them. Throughout the years I discovered how one can be mentally "in phase", at least in our restricted research field, just by carefully reading a few of their papers in chronological order, listening to some of their talks and occasionally talking to them. Sometimes I try to predict what these persons are thinking or which experimental or theoretical problems they are planning to attack next. Surprisingly, I succeed several times! I quess that is how science develops. I see it (science) as a long standing game, played internationally and being slowly solved, piece by plece, year after year. In the next sessions I will show some of the pieces I have put together in the past decade, in chronological order... I will also describe a little the history and motivation behind each research topic. Thus, this review does not cover many other important developments on crystal nucleation and growth in glass. The reader is urged to refer to the above mentioned authors if he intends to be "in phase" in the field. This paper discusses the effects of liquid-liquid phase separation on crystal nucleation, the applicability of the Classical nucleation theory to glass crystallization, scme remarkable trends observed on homogeneous and heterogeneous 4 E. D. Zanotto nucleation in oxide glasses and, finally, the state-of-art on surface crystallization. It does not review the relevant theories nor . the experimental procedures due to space limitations. Here instead, I present solely a minimum of relevant equations, necessary to follow the article. 1. THE EFFECTS OF LIQUID PHASE SEPARATION ON CRYSTAL NUCLEATION In the seventies there was a tremendous scientific excitement and research activity on liquid-liquid phase separation (LLPS) as well as on crystal nucleation and growth. (CNG) in glasses, mainly due to the potential development of glass-ceramics having unusual properties and applications. Of special interest were the possible relationships between the two phenomena. Several authors speculated that LLPS occurred before CNG and was a necessary step for the production of fine grained glass-ceramics. A strong controversy on how LLPS could affect NCG existed, as shown in a meeting on "The Vitreous State" held at The University of Bristol-UK [1]. Thus, some authors defended that compositional changes induced by liquid phase separation could affect CNG, while others believed that tne interfaces between the glassy phases should play an important role on crystallization, by providing favorable sites for heterogeneous nucleation. I started working on this specific subject, as a MSc student of physics, with Aldo Craievich, in Sao Carlos, Brazil, in 1977. At the same time, Peter James had a PhD student, Anthony Ramsdem, working on the same topic in Sheffield. We exchanged some information, and after finishing my MSc dissertation in 1979, I received a grant from the Brazilian government and decided to apply to the 'Glass Mecca' at the time, Sheffield University, and there 1 stayed for three years as a PhD candidate. I continued working on the same topic initiated in Sao Carlos, using the excellent library and glass laboratories of Sheffield, under Peter's quidance. We did a systematic, detailed, work with Liz0-SiO2 and BaO-Si0Oz glasses, having compositions inside and outside the spinodal and binodal areas of the respective phase diagrams. We used controlled thermal treatments to induce the simultaneous development of LLPS and CNG. The kinetics of these processes could be decoupled with SAXS, TEM and optical microscopy. The experimental details are described in [2-4]. Please note that the actual problem was raised in 1970, our research work started in 1977 and the first publication with conclusive results only came out in 1983! 1.1 Main Results l.1l.a. BaO-SiQO2 Glasses: | Figure 1 shows the partial phase diagram of the baria- silica system, with the miscibility gap, glass compositions end thermal treatment temperatures used (black dots). Figure 2 6 | E. D. Zanotto that for the establishment of steady-state crystal nucleation rates (Figure 2). Thus, the compositional effects are clear: as LLPS develops, the amorphous matrix phase gets progressively richer in Ba, and the crystal nucleation rates increase till a constant matrix composition is reached ( the nucleation rate of a glass with 33.3 mole$% Ba0O, outside the miscibility gap, is much higher). Additionally, there is no correlation between the number and surface area of the amorphous droplets and the number of crystals nucleated for all glass compositions and treatments tested. This finding eliminates the possibility that droplet interfaces play an important role on crystallization. 0270 743º€ o 28-38 743ºC o 28:38 760% Lo ) 5 15 t (hours) es Figure 3. Total integrated SAXS intensity, Q, for the same glasses of Figure 2. Several other experiments confirmed the overall picture shown above [4]. A limited, but conclusive, number of experiments were carried out to test the effects of LLPS on the crystal growth rates. The results were quite similar to those of the crystal nucleation experiments, being explained by the compositional effects of LLPS. l.1l.b. Li2z0-SiO Glasses Very similar findings emerged from the study carried out with Liz0-SiOz glasses. The crystal nucleation rates for compositions having widely different modifier content, a glass with 31.0 mole% LizO (in the nucleation and growth region of the miscibility gap) and other with 17.7 mole$ Liz0 (well inside the spinodal region) initially increased and then became identical when the steady-state regime was reached, after about three hours at 481 C (Figure 4). As the glass transition range is about 450 -460 C for these glasses, we concluded that the initial increase in nucleation was due to the compositional change of the amorphous matrix caused by phase separation, as in the case of Ba-glasses. When LLPS was completed, the matrix composition of the two glasses were iIdentical (given by the binodal line) and so were the crystal nucleation rates. A different experiment was devisad to test the effects of the advanced stages of LLPS on crystal nucleation. Specimens Crystallization of Glass ] of the two glasses were first heated to 497 C for 5 hours to bring LLPS to the coarsening stage. After that they vwere nucleated at 481 C and "developed" at 570 C (the standard way to allow crystal growth to optical microscopy sizes). Figure 5 shows that the crystal nucleation rates are constant and equal from the beginning. For comparison we also plot the steady- state nucleation rates of as-quenched glasses (dashed lines). The previous treatment at 497 C eliminates the curvatures observed in the initial stages (Figure 4) and decreases the nucleation rates. The intercepts on the Nv axis are due to nucleation in the initial treatment at 497 C. The smaller nucleation rates are due to the different matrix composition of glasses which had been previously phase separated at 497 € (less Liz0) compared to those phase separated at 481 C. In this case, secondary phase separation has not been observed. 12 me Or E e E E: DE VU Ls 5 6 Ss < 2z ç o 4 53 > >
