Kerr studies of several tellurite glasses

Kerr studies of several tellurite glasses

LETTER TO THE EDITOR Journal of Non-Crystalline Solids 355 (2009) 2195–2198 Contents lists available at ScienceDirect Journal of Non-Crystalline So...

150KB Sizes 1 Downloads 35 Views

LETTER TO THE EDITOR

Journal of Non-Crystalline Solids 355 (2009) 2195–2198

Contents lists available at ScienceDirect

Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/locate/jnoncrysol

Letter to the Editor

Kerr studies of several tellurite glasses J.R. Duclère a, A.A. Lipovskii b,*, A.P. Mirgorodsky a, Ph. Thomas a, D.K. Tagantsev c, V.V. Zhurikhina b a

Sciences des Procédés, Céramiques et de Traitements de Surfaces, UMR 6638 CNRS, Faculté des Sciences et Techniques, 123 av. A. Thomas, 87060 Limoges, France St. Petersburg State Polytechnical University, Polytechnicheskaja 29, St. Petersburg 195251, Russia c Research and Technological Institute of Optical Materials Science, Babushkina 36-1, St. Petersburg 193171, Russia b

a r t i c l e

i n f o

Article history: Received 3 April 2009 Received in revised form 13 July 2009 Available online 17 August 2009 PACS: 61.43.Fs 78.20.Jq 81.05.Kf Keyword: Tellurites

a b s t r a c t Estimates of Kerr electrooptical sensitivity of several tellurite glasses are presented. The highest value of Kerr coefficient B  190  10 16 m V 2 is registered for 0.6TeO2–0.3TlO0.5–0.1ZnO glass. This evidences the prospects of thallium–tellurite glass system for electrooptical applications. A gradual decrease of B from 41  10 16 to 26  10 16 m V 2 in (1 x) TeO2 – xNbO2.5 system is revealed for x increasing from 0.1 to 0.15. No crystalline phase was found in that system, thus allowing attributing its Kerr sensitivity to the intrinsic properties of the glass matrix. The Kerr coefficient variation from 66 to 81  10 16 m V 2 was observed for 0.85TeO2–0.15WO3 glasses co-doped with small amounts of silver and cerium. The analysis of optical absorption spectra of several silver-containing tellurium–tungsten oxide glasses makes it possible to think that introducing cerium provokes formation of new mid-range orderings. Ó 2009 Elsevier B.V. All rights reserved.

1. Introduction The design of glasses with high electrooptical sensitivity is important for modern optics, however, the lack of clear understanding of the origin of Kerr sensitivity of glassy materials, except well studied family of alkaline-silica-niobate glasses [1,2], hinders anticipating prospective glass compositions. Nevertheless, the relation of quadratic optical nonlinearity and Kerr electrooptical sensitivity [3,4] allows treating tellurite glasses which demonstrate remarkable magnitude of non-linear coefficients [5,6] as good candidates for electrooptical applications. This brief communication presents results of the characterization of Kerr electrooptical sensitivity of several tellurite glasses. 2. Experimental The samples of glasses were prepared as described elsewhere [7,8] by melting TeO2 prepared in the laboratory (via the decomposition of commercial H6TeO6 at 550 °C) together with intimate mixtures of commercial chemicals in platinum crucibles. The melting temperature was equal to 700 °C for thallium-containing glass and 800 °C for all other glasses. After 20–30 min, the melts were quickly quenched in a brass frame placed on a brass block to obtain 2–3 mm thick rectangular samples 15  20 mm2. Then the

* Corresponding author. Tel./fax: +7 812 5345821. E-mail addresses: [email protected], [email protected] (A.A. Lipovskii). 0022-3093/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2009.07.022

glasses were annealed for 10 h at glass transition temperatures determined by the data of differential thermal analysis performed for the most of these glasses earlier [7,8]. The compositions of several melted glasses are listed in Table 1. The choice of 0.6TeO2– 0.3TlO0.5–0.1ZnO glass composition (sample #1) was because previous studies within the TeO2–Tl2O system have shown that nonlinear index of thallium–tellurite glasses could be from 50 up to 100 times larger than that of SiO2 [5]. Unfortunately such glasses are very brittle (thermal stability of about 50 °C (Tc Tg)) and so no suitable for the elaboration of ‘large glassy objects’ [9]. That is why we used the addition to such glasses of an oxide improving their mechanical strength (ZnO). The choice of the 85TeO2– 15WO3 (mol%) glass (samples ## 4, 5, 10, 14 and 16) takes the following points into account: it presents a good thermal stability and a high mechanical strength which means that it is suitable for elaboration of ‘large samples’ and it has previously demonstrated relatively high third-order non linear susceptibility [10,11]. Moreover this composition is of interest since it allows obtaining by heating the crystallization of the non-centrosymmetric metastable c-TeO2 phase [10]. This c-TeO2 compound presents a second harmonic signal 70 times higher than that of a-quartz which corresponds to 6% of the signal of the LiNbO3 frequency doubling crystal [12]. We can expect in such glass c-TeO2-like crystal motifs, that is why the addition of species which could improve formation of such motifs has been performed. As concerns niobium–tellurite glasses (samples #18–20), their choice has the same motivation: good thermal stability in the range of 10–15 mol% NbO2.5 (Tc Tg of 80–100 °C) suitable for large glassy pieces elaboration [7] and interesting

LETTER TO THE EDITOR

2196

J.R. Duclère et al. / Journal of Non-Crystalline Solids 355 (2009) 2195–2198

Table 1 Composition (in moles where not indicated) and Kerr coefficient B of glasses. Sample 1 4 5 10 14 16 18 19 20

Composition 0.6TeO2–0.3TlO0.5–0.1ZnO 0.85TeO2– 0.15WO3 + 0.1 wt%Ag2O + 0.076 wt%CeO2 0.85TeO2–0.15WO3 + 0.1 wt% Ag2O + 0.056 wt%CeO2 0.85TeO2–0.15WO3 0.85TeO2–0.15WO3 + 0.1 wt% Ag2O 0.69TeO2–0.21WO3–0.1Ag2O 0.85TeO2–0.15NbO2.5 0.875TeO2–0.125NbO2.5 0.9TeO2–0.1NbO2.5

B (10

16

mV

4. Discussion

2

)

190 66 81 – – – 26 32 41

non-linear optical properties [11]. Also crystallization of the cTeO2 phase has been shown [7] from the glass for 12.5% and 15%). For the composition 10 mol% NbO2.5 we also have noticed the crystallization of the d-TeO2 compound [7]. All surfaces of the rectangular samples were polished, and, after measuring optical absorption spectra in the visible range, their bigger surfaces were coated with 200-nm copper electrodes using thermal evaporation. These electrodes were used to apply voltage in the measurements of Kerr electrooptical sensitivity of the glasses at He–Ne laser wavelength k = 0.63 lm using technique reported by the authors earlier [1,13]. Optical radiation going through the sample was modulated by ac voltage applied to the electrodes while samples were placed between two crossed polarizers, and the response of this modulator was measured at the second harmonic of the input modulation frequency. The comparison of the voltages providing the same output signals for a calibrated electrooptical modulator placed between the input polarizer and for the sample under study gave the information about Kerr coefficient B with 5% accuracy. This technique allowed electrical compensating of the initial birefringence observed in all experimental samples of the glasses. Too high birefringence which could not be compensated and, in some cases, high level of light scattering hindered Kerr measurements for about a half of the synthesized glass samples. 3. Results The results of Kerr coefficient measurements are presented in Table 1 in which a high value of Kerr coefficient for thalliumcontaining glass (sample #1) can be readily pointed out. Such a high electrooptical sensitivity is close to that observed for Nb-rich alkaline-silicate glasses [1,14], and is in excellent correlation with high optical nonlinearity registered for thallium–tellurite glasses [5], which makes prospective further studies of this glass system. The quality of synthesized silver-containing tellurite glasses has admitted of a Kerr coefficient measurement for samples #4 and #5 solely. In spite of practically the same composition, Kerr coefficients of these two samples differ by 20%. Consequently, a structural difference of these two glasses should be supposed to explain the Kerr coefficient variation observed. The appearance of such a difference could be deduced from the optical spectra of samples ##5, 10, 14, and 16 presented in Fig. 1. Samples #10, #14, and #16 differ in amount of silver, and the increase of silver content induces the shift of short-wavelength optical absorption edge towards red, the spectral shape of optical absorption being similar. Introducing cerium into the glass results in essential (even in comparison with the sample #16, containing 10 mol% of silver oxide) red shift of the absorption edge – see Fig. 1.

Most probable reason of the difference of the spectra in Fig. 1 is the formation of the pre-nuclei of crystalline phase in the glass (compare the spectra of samples #14 and #5 differing only in cerium). Our previous measurements performed with silica-based glasses with about the same cerium doping showed the position of Ce3+ absorption peak at 310 nm, and the width of this peak was relatively narrow – the influence of Ce-doping on optical density vanished at wavelength of 360 nm. This corresponds to zerolevel width of the absorption peak of about 100 nm. Thus, taking into account wide spectral shape of the increase of optical absorption in sample #5, it can hardly be attributed to Ce+3 absorption. Another possible reason of the observed red shift in sample #5 optical absorption relatively to sample #14 which has no cerium, could be the formation of silver nanoparticles and related plasmonic absorption. Indeed, the high index surrounding of tellurite glass results in the red shift of surface plasmon resonance. Nevertheless, the absence of pronounced peak corresponding to this resonance in sample #5 allows concluding that the peak, if exists, either falls in the region of fundamental absorption of the host glass, or a very small peak lays in the region of slope of the absorption curve. In these cases one could hardly await for increase of optical absorption in whole visible range which is demonstrated by sample #5 in Fig. 1. It is also worth to note that the increase of glass refractivity and related reflection due to the addition of 0.1 wt% Ag2O can not provide observed spectral change comparatively to initial sample #10. Thus one should attribute the difference of spectra #5 and #14 to scattering due to formation of prenuclei or, possibly, to spectral features of these pre-nuclei which can be formed around metallic centers. Such pre-nuclei (called crystal motifs (CMs) in Ref. [1]) are usually too small to be observed by X-ray diffractometry. According to [1], they determine the Kerr coefficient value. Because of a high thermal expansion coefficient, as of a fast viscosity change with temperature just above the glass transition point of these glasses (very ‘short’ glasses), their structure is very sensitive to the annealing temperature, thus indicating a high sensitivity of the CM atomic arrangement to this factor. In other words, to provide a certain structure of CMs in these glasses and, therefore, the value of Kerr coefficient, one needs to choose and keep annealing temperature (or annealing viscosity) with a high precision, and it is because both CM structure and Kerr coefficient are rather sensitive to changes in the fictive (or so-called structural) temperature of the

Fig. 1. Optical absorption spectra of synthesized silver-containing tellurite glasses.

LETTER TO THE EDITOR

J.R. Duclère et al. / Journal of Non-Crystalline Solids 355 (2009) 2195–2198

glass1. This can explain the difference in Kerr coefficients measured for samples #4 and #5. It is essential that the formation of nuclei in silver-doped glasses and subsequent glass ceramisation can be facilitated by cerium addition [15]. In the beginning of this process, the reaction Ce3+ + Ag+ ? Ce4+ + Ag0 takes place, and neutral silver atoms behave as centers of nucleation [16] thus favoring the CM formation to be more steady. UV-irradiation is usually applied to stimulate this process, but to some extent this reaction can be promoted by chemical surrounding, especially, if one takes into account a high mobility of silver ions in tellurite glasses [17]. It is necessary to note that crystallization of glass without silver and cerium doping (#10) [8] as well as the crystallization of glass containing 10 mol% of silver oxide (sample #16)2 starts above 380 °C, and the phenomena observed are caused by the addition of cerium only. It should be also mentioned that pre-nucleation process can hardly be governed by UV-irradiation, as in the case of silicate glass–ceramics, due to low optical transmission of these glass system in UV-blue optical range. Studied tellurite–niobate glasses (samples ##18–20) demonstrated decrease of electrooptic Kerr coefficient with the growth of niobium oxide concentration. In our experiments, even a small (5 mol% of NbO2.5) increase of Nb content in glass led to 50% drop in quadratic electrooptical sensitivity. Such a behavior differs from that of niobium-alkaline-silicate glasses, in which the Kerr coefficient value markedly increases with increasing niobium concentration [1,14]. That effect was explained by augmentation of the content of electrooptical CMs, i.e., those having composition and structure of known electrooptical crystals of alkaline niobates. In the present case, studied glasses did not contain alkaline ions at all. So, niobium atoms did not have partners to form electrooptical CMs, while niobium oxide itself is not a good electrooptical medium. At the same time, in TeO2-based glasses one may expect the formation of CMs like (TeO2)n chains corresponding to a structural form possessing high polarazibility and, respectively, electrooptical/non-linear sensitivity [18]. Introduction of niobium is supposed to lead to diluting the glass with non-electrooptical CMs of niobium oxide that results in decreasing Kerr coefficient. It should be noted that in accordance with the simplest effective medium approximation [1] the dependence of Kerr coefficient on the concentration of CMs (responsible for electrooptical sensitivity of glassy medium) is a linear function of variable C/(1 C)2, where C is the molar concentration of CMs. This approximation gives a good description of the electrooptical behavior of glasses with low concentrations of electrooptical CMs, but the general character of this dependence appears to remain the same for high C as well, that is, Kerr coefficient increases with C with acceleration, and at C close to 1 small variation in the concentration of electrooptical CMs should influence Kerr coefficient stronger as compared to the same variation at C close to 0. In the present case, the concentration of electrooptical CMs is supposedly determined by the concentration of TeO2, which for the studied glasses is close to 1. This can explain why even small additions of niobium oxide, having low electrooptical sensitivity, to tellurite glasses led to 50% drop in electrooptical Kerr coefficient. 1 Fictive temperature of a glass at room temperature can be determined as the one in the equilibrium temperature dependence of a glass melt property (usually, it is equilibrium viscosity measured above Tg) to which one can extrapolate nonequilibrium branch of this dependence measured below Tg. Fictive temperature characterizes non-equilibrium glass structure as corresponding to the structure of one of its equilibrium sates frozen just at this temperature. In case of complete structure stabilization in annealing, fictive temperature coincides with the annealing temperature. 2 Differential thermal analysis of this glass was performed within the frames of the present study.

2197

In parallel, the observed behavior of Kerr sensitivity in tellurite– niobate glasses may evidence the changes of the structure of CMs responsible for Kerr phenomenon, since the change of total polarizability of constituent glass ions is sufficiently small when one goes from sample #20 (0.9TeO2–0.1NbO2.5) to sample #18 (0.85TeO2–0.15NbO2.5). An indirect prove of CM transformation is the formation of different crystalline phases in tellurite–niobate glasses varying in composition as reported in [7]. Spectral measurement showed that optical absorption of samples ##18–20 was very similar, and the position of optical absorption edge remained the same for these glasses3. Thus it could be thought that either the spectral behavior of the structural entities responsible for the electrooptical sensitivity is the same (this seems to be reasonable if they are just CMs of different TeO2 forms), or their spectral features are in the range of wavelength shorter than optical absorption edge of the glass matrix. 5. Conclusion The present study shows that a thallium–tellurite system is likely to be the most prospective glass system for electrooptical applications, which is in line with published previously data on non-linear optical susceptibility [5]. Actually, its Kerr electrooptical sensitivity is several times higher comparatively to the one of tellurium–tungsten and tellurium–niobium glasses, and it is comparable with the highest Kerr coefficients registered in uniform glasses. Tellurium–niobium glasses demonstrate decreasing of Kerr coefficient with an increase of niobium content, most probably this is due to the dilution of high-sensitive electrooptical structure of formed by TeO2 groups and rearrangement of the structure of CMs (mid-range ordering). Tellurium–tungsten glass shows clear dependence of its properties on the presence of small amount of cerium accompanying doping of this glass with silver. Potentially, this could allow controlling electrooptical properties of this glass via formation of neutral silver in glass matrix, however high optical absorption of this tellurite glass in UV-blue–green region prevents governing this process by optical irradiation. Acknowledgements We acknowledge the support of Ministry of Education and Science of Russian Federation (project #988). References [1] A.A. Lipovskii, D.K. Tagantsev, B.V. Tatarintsev, A.A. Vetrov, J. Non-Cryst. Solids 318 (2003) 268. [2] V.N. Sigaev, S.Yu. Stefanovich, B. Champagnon, I. Gregora, P. Pernice, A. Aronne, R. LeParc, P.D. Sarkisov, C. Dewhurst, J. Non-Cryst. Solids 306 (2002) 238. [3] A. Yariv, P. Yeh, Optical Waves in Crystals: Propagation and Control of Laser Radiation, Wiley, New York, 1984. [4] I.A. Denisov, A.G. Selivanov, K.V. Yumashev, A.V. Anan’ev, L.V. Maksimov, N.V. Ovcharenko, V.N. Bogdanov, A.A. Lipovskii, A.N. Vlasova, J. Appl. Spectrosc. 74 (2007) 866. [5] B. Jeansannetas, S. Blanchandin, P. Thomas, P. Marchet, J.C. ChamparnaudMesjard, T. Merle-Méjean, B. Frit, V. Nazabal, E. Fargin, G. Le Flem, M.O. Martin, B. Bousquet, L. Canioni, S. Le Boiteux, P. Segonds, L. Sarger, J. Sol. State Chem. 146 (1999) 329. [6] R.F. Souza, M.A.R.C. Alencar, J.M. Hickmann, R. Kobayashi, Appl. Phys. Lett. 89 (2006) 171917. [7] S. Blanchandin, P. Thomas, P. Marchet, J.C. Champarnaud-Mesjard, B. Frit, J. Mater. Chem. 9 (1999) 1785. [8] S. Blanchandin, P. Marchet, P. Thomas, J.C. Champarnaud-Mesjard, B. Frit, A. Chagraoui, J. Mater. Sci. 34 (1999) 4285. [9] B. Jeansannetas, P. Marchet, P. Thomas, J.C. Champarnaud, B. Frit, J. Mater. Chem. 8 (1998) 1039.

3 The spectra are not presented as they do not add any useful information. Moreover they do not differ.

LETTER TO THE EDITOR

2198

J.R. Duclère et al. / Journal of Non-Crystalline Solids 355 (2009) 2195–2198

[10] S. Blanchandin, P. Thomas, P. Marchet, J.C. Champarnaud-Mesjard, B. Frit, J. Mater. Sci. 34 (1999) 1. [11] R.A.F. El-Mallawany, Tellurite Glasses Handbook: Properties and Data, CRC, Boca Raton, FL, 2002. [12] G. Vrillet, C. Lasbrugnas, P. Thomas, O. Masson, V. Couderc, A. Barthelemy, J.C. Champarnaud-Mesjard, J. Mater. Sci. Lett. 40 (2005) 4975. [13] A.A. Vetrov, A.A. Lipovskii, D.K. Tagantsev, Instrum. Exp. Technol. 45 (2002) 550.

[14] A.A. Lipovskii, V.G. Melekhin, D.K. Tagantsev, Y. Kaganovskii, O.V. Yanush, J. Non-Cryst. Solids 354 (2008) 1245. [15] S.D. Stookey, Ind. Eng. Chem. 41 (1949) 856. [16] F.E. Livingston, P.M. Adams, H. Helvajian, Appl. Surf. Sci. 247 (2005) 526. [17] B.V.R. Chowdari, P.P. Kumari, J. Mater. Sci. 33 (1998) 3591. [18] A.P. Mirgorodsky, M. Soulis, P. Thomas, T. Merle-Mejean, M. Smirnov, Phys. Rev. B 73 (2006) 134206.