New tungstate fluorophosphate glasses

New tungstate fluorophosphate glasses

Journal of Non-Crystalline Solids 351 (2005) 293–298 www.elsevier.com/locate/jnoncrysol Review New tungstate fluorophosphate glasses Gae¨l Poirier a,...

212KB Sizes 0 Downloads 41 Views

Journal of Non-Crystalline Solids 351 (2005) 293–298 www.elsevier.com/locate/jnoncrysol

Review

New tungstate fluorophosphate glasses Gae¨l Poirier a, Marcel Poulain

b,* ,

Younes Messaddeq a, Sidney J. L Ribeiro

a

a

b

Instituto de Quı´mica, UNESP, CP 355, CEP 14801-970, Araraquara, SP, Brazil Laboratoire des Mate´riaux Photoniques, Baˆt 10B, Campus de Beaulieu, Universite´ de Rennes I, 350432 Rennes, France Received 13 October 2004 Available online 5 January 2005

Abstract A new class of tungstate fluorophosphate glasses was identified in the NaPO3–BaF2–WO3 ternary system. The variation of several physical properties was determined with respect to chemical composition. Characteristic temperatures, density and refractive index increase as tungsten oxide content increases. The optical transmission range and specially the energy bandgap depend of the WO3 amount. No crystallization could be observed for the most WO3 concentrated vitreous samples (P20% molar). Color and optical properties of the glasses depend of the melting time because of the presence of reduced tungsten species like W5+ and W4+. In addition, photodarkening is observed in tungsten rich glass samples under UV laser illumination and this phenomenon can be reversible by heat treatment near the glass transition temperature. Ó 2004 Published by Elsevier B.V. IDT: G180; G183; P147; T135; T355; U110 PACS: 61.43.Fs; 65.60.+A; 78.20.e; 78.40.Pg

1. Introduction Phosphate and fluorophosphate glasses make a special group of optical glasses of technological interest. Their specific properties include larger thermal expansion coefficient, smaller liquidus viscosity and softening temperatures than silicate glasses. They have been developed for UV transmission and also as laser hosts with reduced non-linear refractive index. Other non-optical fields of interest include ionic conductivity and sealing [1]. Another interesting feature of these glasses is their ability to incorporate large amounts of transition metal, alkali and rare earth oxides without reduction of glass

*

Corresponding author. Tel.: +33 2 23 23 62 63; fax: +33 2 23 23 69

72. E-mail address: [email protected] (M. Poulain). 0022-3093/$ - see front matter Ó 2004 Published by Elsevier B.V. doi:10.1016/j.jnoncrysol.2004.11.017

forming ability. Spectroscopic studies of doped glasses have been implemented in the scope of active applications while the ability of tungsten to have several oxidation states opens possibilities for electro-optical applications [2–5]. Numerous fluorophosphate glasses based on the association of sodium polyphosphate and divalent fluorides have already been reported [6,7]. This work intents to define a new group of fluorophosphate glasses, tungstate fluorophosphate glasses based on the ternary NaPO3–BaF2–WO3 association. It was expected that tungsten oxide content would influence significantly thermal and optical properties [8–10]. Especially, large WO3 concentrated vitreous samples would be very stable against devitrification [11–15] and would present special properties observed in crystalline WO3 such as thermochromism, photochromism [16,17] or non-linear optical properties [18–23]. The aim of this study is to determine the effect of WO3 incorporation on the

294

G. Poirier et al. / Journal of Non-Crystalline Solids 351 (2005) 293–298

thermal, physical and optical properties of a fluorophosphate-based glass and determine the influence of synthesis conditions on the oxidation state of tungsten atoms in the glass.

2. Experimental 2.1. Glass preparation These glasses were synthesized by conventional method. Powdered starting materials were: – Tungsten oxide WO3 (99.8% pure). – Sodium polyphosphate NaPO3 (99+% pure). – Barium fluoride BaF2 (99.8% pure). In a first step, powders were mixed and heated at 400 °C for 1 h to remove water and adsorbed gases. Then, the batch was melted at a temperature ranging from 800 °C to 1100 °C, depending on WO3 content. Liquid was kept at this temperature for 15–75 min to ensure homogenization and fining. In practice, the melting time must be longer for higher WO3 concentrations to remove the blue coloration of the vitreous samples expected to be due to tungsten reduction processes. The influence of the melting time on physical and optical properties is reported in this work. Finally the melt was cooled in a brass mold preheated below the glass transition temperature Tg (Tg – 20 °C). Annealing was implemented at this temperature for several hours in order to minimize mechanical stress resulting from thermal gradients upon cooling. Bulk samples were cut and polished before performing optical measurements. 2.2. Physical measurements Characteristic temperatures (Tg for glass transition, Tx for onset of crystallization and Tp for maximum of crystallization peak) were determined by differential scanning calorimeter (DSC). The estimated error on the temperature is 2 K for glass transition and onset of crystallization witch are obtained from tangents intersection and 1 K for the position of the crystallization peak. Powdered samples were set in aluminum pans under N2 atmosphere at 10 K/min heating rate. Infrared transmission spectra were recorded with a BOMEM Michelson Spectrophotometer in the 5000–2000 cm1 range. Ultraviolet transmission between 300 and 500 nm was studied using a Varian spectrophotometer Cary 5. The densities were determined by pycnometry under He pressure (Micromeritics Accupyc), which allows a precision of about 104. The refractive index was measured at 633 nm (He–Ne laser) using a Metricon prism coupler and the estimated error on these measurements is ±0.001.

3. Results 3.1. Vitreous domain Glass forming area in the NaPO3–BaF2–WO3 ternary system is drawn in Fig. 1. Limiting compositions correspond to quenched glasses. Sodium polyphosphate (NaPO3)n which exists as a glass forms binary glasses with BaF2 and WO3. Up to 80 mol% of WO3 could results in vitreous phases, even in the NaPO3–WO3 binary system. Only ceramics were obtained in the WO3– BaF2 system. However vitrification could occur with the addition of 10 mol% NaPO3 or less. This suggests that rare earth alkali fluorotungstate glasses could be synthesized. A set of samples with increasing WO3 content were prepared according to the following composition rule: (80–0.8x)NaPO3–(20–0.2x)BaF2–xWO3, x varying between 0 and 60 mol%. These compositions contain a constant [NaPO3]/[BaF2] ratio in order to study the effect of WO3 incorporation. As tungsten concentration increases, samples appear yellowish. 3.2. Thermal analysis Characteristic temperatures were recorded using the DSC technique up to 600 °C. They are reported in Table 1 which also gives the value of the thermal stability range Tx–Tg as an estimate of glass stability. This value is not reported when crystallization peak is not observed at the applied heating rate (10 K min1). As shown in Fig. 2, glass transition temperature increases from 240 °C to 524 °C as the amount of WO3 rises from 0% to 60%. This increase of Tg with respect to the WO3 concentration was already described for several glass matrix [9–11,24,25].

NaPO3 Glass Ceramic

BaF2

WO3

Fig. 1. Phase diagram in the NAPO3–BaF2–WO3 ternary system.

G. Poirier et al. / Journal of Non-Crystalline Solids 351 (2005) 293–298 Table 1 Characteristic temperatures and thermal stability range for glasses from this study Sample

Tg (°C)

Tx (°C)

Tp (°C)

Tf (°C)

Tx–Tg (°C)

NBW0 NBW10 NBW20 NBW30 NBW40 NBW50 NBW60

240 280 320 370 418 465 524

303 430 – – – – –

320 470 – – – – –

534 589 – – – – –

63 150 – – – – –

Tg is the temperature for glass transition, Tx for onset of crystallization, Tp for maximum of crystallization peak and Tf for melting temperature.

500 450

3.4. Optical properties As one could expect, refractive index increases when tungsten oxide is incorporated. This is exemplified by Fig. 4 which reports the evolution of the refractive index versus WO3 content, from 1.5 to 1.85 [18]. Optical transmission of these glasses was recorded in the ultraviolet and in the infrared spectra. Transmission is limited by the band gap at short wavelength and by the multiphonon absorption in the mid infrared. The evolution of the transmission spectra with respect to glass composition is displayed in Fig. 5 in the UV–visible range and Fig. 6 in the infrared range. The observed redshift of the optical bandgap, from 3.6 to 3 eV, when WO3 concentration increases, is consistent with sample color [26,27]. A large absorption band is observed near 2700 cm1 and corresponds to the multiphonon absorption of the O–P–O chains in polyphosphate structure [28].

400 350

1.9 300 250

1.8

200 0

10

20

30

40

50

60

Molar concentration of WO3

Fig. 2. Evolution of vitreous transition temperature versus WO3 concentration.

Refractive index

Vitreous transition temperature ( C)

550

295

3.3. Density

1.7

1.6

1.5

Measured density increases roughly linearly versus WO3 molar concentration, as shown in Fig. 3. Limits of variation are 3.2 g cm3 for tungsten free sample and 5.2 g cm3 at 60 mol% WO3.

0

10

20 30 40 Molar concentration of WO3

50

60

Fig. 4. Evolution of refractive index versus WO3 concentration.

100 80

40 20

3.5

NBW 10 NBW 20 NBW 3 NBW 0 40 NBW 50 NBW 60

4.0

60

0

4.5

NBW

Transmittance (%)

3

Density (g/cm )

5.0

0 300

3.0 0

10

20 30 40 Molar concentration of WO3

50

Fig. 3. Evolution of density versus WO3 concentration.

60

400

500

Wavelength (nm)

Fig. 5. UV–visible transmittance of the samples versus WO3 concentration. Sample thickness: 3 mm.

296

G. Poirier et al. / Journal of Non-Crystalline Solids 351 (2005) 293–298

Transmittance (arbitrary units)

100

Transmittance (%)

80 60

NBW60

40 20 0 5000

NBW0 4000 3000 Wavenumber (cm-1)

1.4

NBW50 - 60 min

1.2

NBW50 - 30 min NBW50 - 10 min

1.0

NBW50 - 5 min

0.8 0.6 0.4 0.2 0.0 800

2000

Fig. 6. Infrared transmittance of the samples versus WO3 concentration. Sample thickness: 3 mm.

1200

1600

2000

Wavenumber (cm-1 )

Fig. 7. Infrared transmittance of the NBW50 sample versus melting time.

1.6

3.5. Influence of the melting time

NBW50 - 5 min

Color variations were observed in the large WO3 concentrated vitreous samples when we changed the melting time. In fact, a very strong blue coloration appeared when we applied a short melting time whereas this coloration disappeared progressively when we increased the melting time. Four samples with the same composition (NBW50) were synthesized using the same melting temperature (1050 °C) and four different melting times (5 min, 10 min, 30 min and 60 min). Therefore, several physical and optical characterizations were performed to determine this color change origin. Thermal analysis were performed by DSC in the 200–600 °C temperature range and all the samples show the same thermal characteristics with a vitreous transition temperature of 465 °C and no crystallization phenomenon. A structural investigation was realized by infrared spectroscopy in the 1500–400 cm1 range using glass powders dispersed in KBr pellets. These absorption spectra are presented in Fig. 7. A large absorption band constituted of several elementary bands can be observed between 1500 and 400 cm1. The central band at 910 cm1 is attributed to the symmetric stretching vibrations of WO6 octahedra [29] whereas the shoulders observed at 700 cm1 and 1180 cm1 are due to the symmetric stretching vibrations of P–O–W bonds and Q1 phosphate tetrahedra respectively [30,31]. In addition, the absorption spectra are similar for all the samples and seems to be independent of the melting time. The absorption spectra in the UV–visible-near infrared were obtained between 300 nm and 1000 nm range as shown in Fig. 8. A large absorption band centered in the visible range (600–1000 nm) disappears progressively by increasing the melting time and is attributed to reduced tungsten species W5+ and W4+ [32,33].

Absorbance (a.u)

1.4 1.2 1.0 0.8 NBW50 - 10 min

0.6 0.4

NBW50 - 30 min

0.2

NBW50 - 60 min

0.0 400

600

800 1000 Wavelength (nm)

1200

Fig. 8. UV–visible-near infrared transmittance of the NBW50 sample versus melting time.

As shown in Table 2, refractive index depends on the time for which the glass is maintained in the molten state at room atmosphere. 3.6. Photosensitivity Samples containing 40 and 50 mol% of tungsten oxide have been submitted to the 350 nm exposure of a UV excimer laser. Photodarkening was observed and this phenomenon appeared to be reversible when an Table 2 Color and refractive index of the different NBW50 studied samples Sample NBW50 NBW50 NBW50 NBW50

– – – –

5 min 10 min 30 min 60 min

Melting time (min)

Color

Refractive index

5 10 30 60

Dark blue Blue Light blue Transparent

1.769 1.773 1.785 1.785

G. Poirier et al. / Journal of Non-Crystalline Solids 351 (2005) 293–298

annealing stage was implemented for 2 h below glass transition temperature.

4. Discussion This study has shown that large amounts of tungsten oxide could be incorporated in a fluorophosphate glass. As a result the vitreous transition temperature increases with increasing WO3 content. The glass forming range is very large and stability against devitrification increases. This is exemplified by the value of the thermal stability range Tx–Tg that increases from 63 to 150 °C when WO3 concentration rises from 0 to 10 mol%. At large WO3 content compositions, neither crystallization nor melting peaks were observed at 10 K min1 heating rate. This behavior demonstrates that crystallization kinetics is largely reduced, which allows one to keep the glass melt above Tg for a long time before crystallization becomes noticeable. This makes a serious advantage for the fabrication of optical components and fiber drawing. The linear evolution of the physical properties, such as the vitreous transition temperature, the density or the refractive index, versus WO3 concentration suggests that the evolution of the glass structure is monotonic. These evolutions are consistent with the larger atomic weight, polarizability and bond strength of tungsten VI cations by comparison to the other cations of the glass. A general structural study of these glasses by vibrational spectroscopy (IR, Raman), 31P MAS-NMR and X-ray absorption spectroscopy (XANES, EXAFS) is now in progress. The redshift of the UV–visible optical bandgap is exemplified by the color change of these glasses from transparent for 0% of WO3 to yellow and green for large WO3 concentrations, as reported elsewhere [27]. Tungsten at the VI oxidation level shows no UV absorption; consequently this optical absorption arises from defects located nearly tungsten ions, such as anion vacancies in the coordination polyhedron or electrons trapped in defect sites. It is assumed that these defects account for the photodarkening effect reported in Section 3.6. In addition, while W6+ cation does not show absorption in the UV–visible range, the evolution of the glass color as a function of melting time and the associated absorption band described in Section 3.5 are assumed to be due to reduced tungsten atoms (W5+ and W4+) [32,33]. More particularly, the blue coloration is caused by polaron transitions between two non-equivalent sites as described below [34]: hm þ W5þ ðAÞ þ W6þ ðBÞ ! W5þ ðBÞ þ W6þ ðAÞ; hm þ W5þ ðAÞ þ W4þ ðBÞ ! W5þ ðBÞ þ W4þ ðAÞ: These reduced species can be found in the starting products or formed during the melting stage by the chemical

297

action of reducing impurities such as cellulose, organics or hydrocarbides. A long melting time helps to reoxidise these tungsten atoms and to obtain transparent samples containing W6+. On the other hand, these glasses are thermochromic because the colored samples can be partially bleached by thermal treatment under oxidizing atmosphere near the glass transition temperature. In the same way the transparent samples can be partially blue colored by thermal treatment under reduced atmosphere near the glass transition temperature. The vitreous samples with large amounts of WO3 proved to be photosensitive to UV laser illumination and this photodarkening can be reversible by heat treatment. In addition, these glasses were identified to be very good rare earth matrix that shows interesting emission properties [35] while the high WO3 content samples have a large non-linear absorption behavior [36].

5. Conclusion New tungstate fluorophosphate glasses have been synthesized and characterized in the NaPO3–BaF2– WO3 ternary system. Tungsten oxide increases the thermal stability of the glass versus devitrification. Physical and optical properties are very sensitive to WO3 concentration. The color of most WO3 concentrated samples show to be dependant on melting time during the synthesis because of oxidation–reduction processes involving tungsten atoms in the melt and these compositions are thermochromic. Photodarkening has been observed in some composition range. Further studies include photosensitivity and structural characterizations.

References [1] B.C. Sales, Mater. Res. Bull. 12 (1987) 32. [2] G. Chiodelli, A. Magistris, M. Villa, J.L. Bjorkstam, J. Non Cryst. Solids 51 (1982) 143. [3] U. Selvaraj, K.J. Rao, J. Non Cryst. Solids 72 (1985) 315. [4] S. Muthupari, S. Prabakar, K.J. Rao, J. Phys. Chem. 98 (1994) 2646. [5] A.S. Linsey, E. Owen, F.M. Hayatee, J. Non Cryst. Solids 4 (1970) 208. [6] J. Jiang, G. Zhang, M. Poulain, J. Non Cryst. Solids 213&214 (1997) 11. [7] J. Jiang, G. Zhang, M. Poulain, J. Alloy. Compd. 275–277 (1998) 733. [8] J.J. Rothermel, J. Am. Ceram. Soc. 32 (1949) 153. [9] A. Mansingh, A. Dhawan, R.P. Tandon, J. Non Cryst. Solids 27 (1978) 309. [10] A. Mansingh, R.P. Tandon, Phys. Rev. B 21 (1980) 4829. [11] T. Kosuge, Y. Benino, V. Dimitrov, R. Sato, T. Komatsu, J. Non Cryst. Solids 22 (1998) 154. [12] I. Shaltout, Y. Tang, R. Braunstein, E.E. Shaisha, J. Phys. Chem. Solids 57 (1996) 1223.

298

G. Poirier et al. / Journal of Non-Crystalline Solids 351 (2005) 293–298

[13] S.K.J. Al-Ani, C.A. Hogarth, R.A. El-Malawany, J. Mater. Sci. 20 (1985) 661. [14] V. Kozhukharov, M. Marinov, G. Grigorova, J. Non Cryst. Solids 28 (1978) 429. [15] S. Blanchandin, P. Marchet, J. Mater. Sci. 34 (1999) 4285. [16] B.W. Faughnan, Appl. Phys. Lett. 27 (1975) 275. [17] B.W. Faughnan, IEEE. T. Electron. Dev. 22 (1975) 1063. [18] S.H. Kim, T. Yoko, J. Am. Ceram. Soc. 78 (1995) 1061. [19] H. Nasu, T. Uchigaki, M. Nakamura, K. Kamiya, S. Saka, N. Soga, in: Proc. Int. Conf. of Science and Technology of New Glasses, Tokyo, 1991, p. 175. [20] A. Berthereau, Y. Le Luyer, R. Olazcuaga, Mater. Res. Bull. 29 (1994) 933. [21] M.E. Lines, Phys. Rev. B 43 (1991) 1197. [22] M.E. Lines, Phys. Rev. B 41 (1990) 3383. [23] M.E. Lines, J. Appl. Phys. 69 (1991) 6876. [24] I. Shaltout, Y.I. Tang, R. Braunstein, J. Phys. Chem. Solids 56 (1995) 141.

[25] F.P. Koffyberg, F.A. Benko, J. Non Cryst. Solids 40 (1980) 7. [26] T. Nishida, M. Suzuki, S. Kubuki, J. Non Cryst. Solids 194 (1996) 23. [27] M. Wollenhaupt, H. Ahrens, P. Frobel, K. Barner, J. Non Cryst. Solids 194 (1996) 191. [28] M. Nalin, Y. Messaddeq, S.J.L. Ribeiro, M. Poulain, V. Briois, J. Optoelectron. Adv. Mater. 3 (2001) 553. [29] P. Subbalakshmi, N. Veeraiah, J. Non Cryst. Solids 298 (2002) 89. [30] A.M. Efimov, J. Non Cryst. Solids 209 (1997) 209. [31] A.M. Efimov, J. Non Cryst. Solids 253 (1999) 95. [32] Y.F. Lu, H. Qiu, J. Appl. Phys. 88 (2) (2000). [33] G. Leftheriotis, S. Papaefthimiou, P. Yianoulis, A. Siokou, Thin Solid Films 384 (2001) 298. [34] O.F. Schirmer, J. Phys. C 6 (1980) 479. [35] G. Poirier, V.A. Jerez, C.B. de Araujo, Y. Messaddeq, S.J.L. Ribeiro, M. Poulain, J. Appl. Phys. 93 (3) (2003). [36] G. Poirier, C.B. de Araujo, Y. Messaddeq, S.J.L. Ribeiro, M. Poulain, J. Appl. Phys. 91 (12) (2002) 10221.