Transparent Hafnia-Containing b-Quartz Glass Ceramics: Nucleation and Crystallization Behavior Lothar Wondraczek w and Philippe Pradeau Corning European Technology Center, Corning S.A.S., Avon, France Heterogeneous nucleation and crystallization of lithium alumo- silicate glasses with hafnia-containing nucleation agents was ex- plored. Kinetic, thermodynamic, and structural data were considered to assess nucleation efficiency and to characterize the crystallization process. It is shown how lattice parameters and, particularly, anisotropy of the nuclei phase depend on the amount of ZrO 2 –HfO 2 substitution in a specific base glass. For a given crystallization treatment, the size of the derived crys- tallites of b-quartz structure is used as a measure of nucleation rates. A nonlinear dependence of nucleation efficiency on HfO 2 content was established, with the supposedly most efficient nu- cleation corresponding to the lowest degree of anisotropy of the nuclei crystallites, i.e., when about 20%–30% of ZrO 2 was substituted by HfO 2 . Apparent activation energies and estimates of the Avrami coefficient were determined from nonisothermal crystallization experiments for selected compositions to high- light the differences between hafnia and zirconia. Hafnia can be used alone or in combination with other agents to nucleate nano- crystalline glass ceramics with a low coefficient of thermal expansion. I. Introduction L OW thermal expansion glass ceramics of the system Li 2 O– Al 2 O 3 –SiO 2 (LAS) are today, commercially, the most suc- cessful example of this class of materials. 1–3 Moreover, they count among the very first examples of internally nucleated glass ceramics. Their successful development is based on Stookey’s original work on the use of TiO 2 as nucleation agent, 4 and sub- sequent work by Kondratev, 5 Alekseev and Zasolotskaya, 6 Beall et al., 7 Tashiro and Wada, 8 Doherty et al., 9 Sack and Scheidler, 10 and Petzoldt 11 in the 1960s. To achieve a low or negative coefficient of thermal expansion, the major crystalline phase is usually either a solid solution of b-quartz (often also referred to as high quartz, b-eucryptite) or of b-spodumene (keatite) structure, containing different amounts of MgO and/or ZnO. 11 Although nucleation proved to be the key step in ob- taining fine-grained glass ceramics, only a small number of pos- sible agents are known today that, in LAS glasses, enable precipitation of b-quartz phases with crystallite sizes well below 100 nm. Among these, historically TiO 2 , 4 ZrO 2 , 8 and combina- tions of the two 7,10,11 are still by far the most prominent repre- sentatives. ZrO 2 - and/or TiO 2 -induced nucleation in precursor glasses for LAS glass ceramics has been studied comprehensively by Schiffner and Pannhorst. 12,13 They showed that nucleation in those systems is highly nonstationary, 12 and they quantitatively described the impact of changes in the TiO 2 /ZrO 2 ratio as well as total molar fraction on nucleation rates. Today, these two agents, in different quantities, are the essential constituents of practically all commercially produced, internally nucleated LAS glass ceramics. Heterogeneous nucleation in LAS systems is often believed to be associated with initial amorphous phase separation, 14 which is followed by the formation of crystalline phases that contain nucleating oxides. 7–10,15 The formation of the first crystalline phases may occur either homogeneously or at the internal interfaces that are developed during phase separation. It is now generally accepted that these crystallites then act as nucleation sites for the formation of crystalline silicate phases. Using transmission electron microscopy, Maier and Mu¨ ller 15 provided direct evidence that epitaxy of b-quartz phases on crystalline ZrTiO 4 is the critical step in the crystalli zation of LAS glasses that contain ZrO 2 and TiO 2 . They also showed that while each b-quartz crystallite contains a ZrTiO 4 crystallite at its center, b-quartz crystallites are also surrounded by a much larger number of peripheral ZrTiO 4 crystallites of different shapes. They suggested that these peripheral crystallites are the result of secondary phase separation in the residual glass phase because of increasing residual concentrations of TiO 2 and ZrO 2 as silicate precipita- tion progresses. The search for alternative nucleation agents constantly con- tinues, and compounds such as Ta 2 O 5 16 and, more recently, SnO 2 have been reported to be, at least in combination with either or both TiO 2 and ZrO 2 , effective alternatives. On the other hand, efficient nucleation has also been reported for a se- ries of metallic clusters or particles, e.g., silicon, 17 platinum, or rhodium. However, the latter are usually not an alternative be- cause they generally result in dark colors, either directly or by disadvantageously affecting redox states of other batch compo- nents. In the present study, we elucidated how glass ceramics with a low coefficient of thermal expansion and high optical transparency can be nucleated with hafnia or solid solutions of hafnia, zirconia, and/or titania. II. Experimental Procedure (1) Sample Preparation A series of glass samples, containing different amounts of HfO 2 and ZrO 2 , was produced by melting 1 kg batches of the approximate (70.5 mol%) composition (mol%) 74SiO 2 – 13Al 2 O 3 –8Li 2 O–3(MgO1ZnO1BaO)–0.5TiO 2 –(xHfO 2 1yZrO 2 ) (x and y given in Table I) in Pt crucibles. Filled crucibles were inserted into a preheated electrical furnace at 14001C and melted for 3 h at 16801C. Melts were then poured onto a preheated steel plate and rolled into a 4-mm-thick sheet. For further analysis, part of the sample was transformed into a glass ceramic by re- heating to 7801C, annealing for 2 h at this temperature, further heating to 9201C, and annealing for 1 h. This procedure yielded highly transparent glass ceramics. Densities were measured be- fore and after crystallization using a He pycnometer (Micro- meritics, Norcross, GA) (10 measurements per sample), and thermal expansion was measured with a horizontal dilatometer (DIL 402C, Netzsch, Selb, Germany; heating rate 10 K/min, air). D. Johnson—contributing editor w Author to whom correspondence should be addressed. e-mail: lothar.wondraczek @ww.uni-erlangen.de Manuscript No. 24100. Received December 14, 2007; approved January 24, 2008. J ournal J. Am. Ceram. Soc., 91 [6] 1945–1951 (2008) DOI: 10.1111/j.1551-2916.2008.02373.x r 2008 The American Ceramic Society 1945