berthier 2007

berthier 2007

(Parte 1 de 3)

New large grain, highly crystalline, transparent glass–ceramics

Thiana Berthier a, Vladimir M. Fokin b, Edgar D. Zanotto a,* a Vitreous Materials Laboratory, Department of Materials Engineering, Federal University of Sao Carlos, 13.565-905 Sao Carlos, SP, Brazil b S.I. Vavilov State Optical Institute, ul. Babushkina 36/1, St. Petersburg 192131, Russia

Received 28 November 2006; received in revised form 13 August 2007 Available online 1 October 2007


Two main reasons assure the transparency in the visible of some glass–ceramics (TGC): their crystal sizes are much smaller than the wavelength of light or the difference between the refractive index of glass matrix and crystal phase is small. The majority of traditional TGC have nano-size crystals and small to moderate crystallized volume fraction, usually between 3% and 70%. In this article we present a new type of transparent glass–ceramics having large (micrometric) grain size and very high crystallized volume fraction, which reaches up to 97%. Their high transparency mainly results from simultaneous variations of the glass matrix and crystal compositions during crystallization, which considerably decreases the difference between the respective refractive indexes, and this factor prevails, regardless of crystal size. Preliminary tests of their optical properties indicate that this new family of TGC can be further developed by doping with transition metals and rare-earths. 2007 Elsevier B.V. All rights reserved.

PACS: 61.43.Fs;; 64.60.Q Keywords: Crystallization; Crystal growth; Glass ceramics; Nucleation; Oxide glasses; Alkali silicates; Silicates; Soda-lime–silica

1. Introduction

Glass–ceramic technology is based on controlled crystal nucleation and growth in certain glasses, and has several advantages over conventional powder-processed ceramics, such as very low or null porosity, as well as uniformity and reproducibility of microstructure. The crystalline phases precipitated in some glass matrices typically enhance and sometimes lead to entirely new combination of properties. Some of the technologically most important glass–ceramics have, for instance, low thermal expansion coefficients coupled with high thermal and chemical stability, high mechanical strength and optical transparency [1].

Recently, much attention has been devoted to optically transparent glass–ceramics that have improved optical properties, greater thermal stability and strength than their parent glasses [2,3]. Dozens of papers and patents on transparent glass–ceramics have been registered worldwide in the last thirty years (see, e.g. [4]). Applications of these transparent crystalline materials, which are capable of being shaped by fast and flexible glass-forming processes, are numerous and include cooking ware, fire resistant plates, security windows, hosting medium for transition metals and rare-earths among others [5].

This paper focuses on the development and characterization of a new type of transparent glass–ceramic. The originality of this material is based on a tight control of homogeneous nucleation and growth of solid solution crystals with a continuous variation of the compositions of both crystalline and glassy phases during phase transformation. This unique evolution reduces the difference between the refractive indexes of crystals and residual glass. The crystallinity of this new TGC can reach up to 97% (spontaneous cracking takes place for higher crystallinities) with grain sizes of about 5–7 lm, in contrast to conventional TGC that have nano-metric microstructures

02-3093/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2007.08.052

E-mail address: (E.D. Zanotto). URL: (E.D. Zanotto).

Available online at Journal of Non-Crystalline Solids 354 (2008) 1721–1730 and low to moderate crystallized volume fraction. Some of the characteristic optical properties of these new, large grain transparent glass–ceramics are presented below together with an analysis of their crystallization process.

2. Brief literature review

Some of the most important proposed applications of

TGC refer to optical components and devices including: laser media; saturable absorber media; illumination devices using infrared light; heat-resistant materials that absorb ultraviolet light, TGCs that reflect infrared light and are transparent to visible light; substrates for arrayed waveguide grating (AWG) [6,7]; radiation sources of lamps; laser pumps; substrate materials; wafer stages; mirror support for telescopes; second harmonic generation (optical energy converting elements) [8,9]; front windows in heating furnaces or slow-cooling reactors; precise photolithography; blue or green state lasers, LCD substrates, optical amplifiers for wavelength up-conversion [10]; ring laser gyroscopes; materials that absorb UV and fluoresce in the red/IR (possibly used as tunable lasers) [1]; solar collectors; printed optical circuits; gradient refraction optical lenses; optical waveguides [12–14], etc.

There are several patents on transparent glass–ceramics that contain a major crystalline phase of low thermal expansion coefficient (0.1 · 10 6–1.0 · 10 6/K) and thus present high thermal shock resistance [4]. Transparency is usually observed in glass–ceramics containing relatively small to moderate volume fractions of crystals (3–70%) with nanoscale size (1–15 nm) [15]. The majority of these low expansion materials belong to the lithia–alumina–silica family with titanium and/or zirconia as nucleating agents, and a number of minor elements to impart color or to improve processability, and have beta-quartz solid solution as the major phase [16].

3. Theoretical background

There are two distinct ways in which light intensity can be attenuated by a medium: atomic absorption and scattering [17,18]. A combination between these two mechanisms plus light reflection from the external surfaces can be expressed by the following equation:

and I0 and I are the initial and transmitted light intensity, respectively, b the linear absorption coefficient (absorption index), S the scattering coefficient, x the optical path length, and n is the refractive index of the TGC. The scattering coefficient S can be written as or assuming spherical particles, as where N and V are the number density and volume fraction of the scattering particles, respectively, and K is a scattering factor.

K increases with particle-size, reaches a maximum when the particle-size is close to k, and decreases for larger particle-sizes, approaching a constant value for r k. Thus, according to Eq. (3), for a fixed concentration (V) of scattering phase the measured scattering coefficient decreases with increasing particle-size (see, e.g. [19]).

If a scattering medium has particles with sizes less than 0.1–0.2 of the typical wavelength of visible light, Rayleigh scattering takes place and the angular distribution of scattered light can be written as [3] where h is the scattering angle, L is the distance from the scattering centers of radius r, and M is the ratio of the refractive index of the particle to that of the surrounding medium. According to Eq. (4) the scattered intensity depends mainly on the ratio of the particle-size to the wavelength of light, r/k, and the ratio of refractive indexes, M. Thus, for any given value of h, L, and k, the smaller the first ratio and the closer to unity the second, the most transmitting is the medium.

If the sizes of the scattering centers are comparable or larger than the wavelength of visible light (as in the case of the micrometric crystals of our TGC), the simple Rayleigh theory breaks down. For such relatively large crystal sizes, the dependence of the scattered light intensity on k becomes weaker than that for the Rayleigh scattering, and is given by where p < 4 and decreases with increasing particle-size. Due to the interference of light scattered by different parts of each scattering particle, the I(h) dependence becomes more complicated. With increasing particles sizes forward scattering (acute angles) prevails back scattering (obtuse angles), the so-called Mie effect. This effect may lead to a decrease of the measured scattering coefficient (contrary to the case of small particles) with increasing particle-sizes.

4. Materials and experimental procedures

Sodiumandcalciumcarbonates,andcrushedquartzwere used as starting materials. After melting the proper mixtures in a platinum crucible at 1300–1500 C for about 4 h, the melts were cast on massive metallic plates to give about 5 m thick specimens. The compositions of the glasses were close to the pseudo-binary join CaO Æ SiO2–Na2O Æ SiO2 in the interval (Na2O Æ 2CaO Æ 3SiO2(N1C2S3)–Na2O Æ CaO Æ

2SiO2(N1C1S2)) of solid solution formation. Table 1 shows typical compositions of the TGCs here investigated.

1722 T. Berthier et al. / Journal of Non-Crystalline Solids 354 (2008) 1721–1730

The degree of crystallinity and crystal sizes were measured by optical microscopy with a pattern analyzer. The crystal and glassy matrix compositions were estimated via energy dispersive spectroscopy (EDS) using Stereoscan 440 Link/Oxford. The excitation volume of the electron beam was about 2–3 lm in diameter. The parent glass and a fully crystallized glass were employed as standards. About six measurements were performed for both glassy and crystalline phases to evaluate the average values.

Spectra of the partly and fully crystallized glasses, shaped as parallel-sided polished plates, were recorded in the range of 200–10 nm to characterize their transmittance as a function of crystallinity. To estimate the param-

absorption and scattering, respectively, the transmittance

I/I0 was measured as a function of sample thickness, x. Fig. 1 shows a ln(I/I0) versus x plot approximated by a lin-

ear function Y = A Bx where A =l n(P1)a nd B = P2. The refractive indexes of the parent glasses and TGCs were measured with an Abbe refractometer using the

NaD line (k = 589 nm). The Abbe numbers (td) were calculated by the following equation where nd, nf, and nc are the refractive index at k = 589, 486, and 643 nm, respectively [20].

A Netzsch DSC 404 equipment was employed for the determination of the characteristic temperatures in the heating and cooling curves using bulk samples of about 20–30 mg.

The crystals grown in the present glasses are solid solu- tions (s/s) having the chemical formula Na4+2x-

At a temperature Tpm, well below the glass transition temperature, Tg, the solid solutions with 0 < x < 0.5 reveal a reversible polymorphic transition, which is accompanied by a clear thermal effect in the DSC curves (Fig. 2) also reported in Ref. [24]. The value of Tpm strongly decreases with decreasing sodium content (see Fig. 3) and thus can be employed to characterize the crystal composition, see Ref. [25].

Depending on the parent glass composition, the crystals are approximately spheric or cubic shaped, such as those shown in Fig. 4. Crystals with mixed shapes were occasionally observed as has been already noted in Ref. [26]. Excess of SiO2 relative to the metasilicate composition favors spherical crystals, while a deficit in SiO2 favors the formation of cubic shaped crystals.

Fig. 5 shows the XRD patterns of almost fully crystallized samples TGC1 and TGC2 with crystals of cubic and spherical morphologies, respectively. Both morphologies refer to solid solutions based on the hexagonal structure

Table 1 Examples of some our TGC compositions in mol%

TGC SiO2 Na2O CaO a By analysis. b This TGC was doped with 0.2 and 0.5 wt% Er.

ln(I/I x, m

Fig. 1. ln(I/I0) versus x typical plot for TGC-0 with about 3% of crystallinity fitted by a linear function.

TLTcr-c Tcr-h

Tg Tpm exo

T,oC endo

Fig. 2. Typical DSC heating and cooling curves for glass TGC-0. The dotted and solid lines refer to heating and cooling with rates ±10 K/min, respectively. Arrows denote the characteristic temperatures: glass transi- tion, Tg, crystallization during heating, Tcr-h, and cooling, Tcr-c, liquidus,

TL, and the reversible phase transition, Tpm. The ‘pre-melting’ peak is likely caused by the peculiarities of the s/s melting between solidus and liquidus. It should be noted that this peak is typical for compositions that are shifted relative to stoichiometric composition N1C2S3. The inset shows DSC heating and cooling curves for a fully crystallized TGC-2.

T. Berthier et al. / Journal of Non-Crystalline Solids 354 (2008) 1721–1730 1723

Crystal morphology is one of the factors which affect the transparency of crystallized glasses. Fig. 6 shows examples of a typical transmission spectrum and micrographs of some TGCs with spherical and cubic shaped crystals. The volume fraction crystallized is close to 1.0 in both cases. It is clear that the cubic form favors the transparency of our TGC.

(Parte 1 de 3)