@
Paper3. Effect of Oxygen Diffusion from the Atmosphere on Sn Depth
Profile of Float Glass at Temperatures Above Tg
Satoshi Takeda*, Ryoji Akiyama* and Hideo Hosono**
@
@The effect of oxygen diffusion from the atmosphere on tin depth profile in the bottom face of a soda - lime - silica float glass at temperatures above Tg was investigated. The sheet glass samples were heat - treated under 18O2/N2 and argon (Ar) atmospheres. The significant diffusion of tin from the inside of the glass to the surface was observed for the glass heat - treated in 18O2/N2 atmosphere, resulting in the formation of a tin - enriched layer near the surface region. It was found that the tin was supplied from the region shallower than the "hump" which is commonly observed in the tin depth profile of a commercial soda - lime - silica float glass. No significant change in the tin depth profile was observed for the glass heat - treated in Ar atmosphere. These results indicate that 18O diffusion into the glass, which causes the change in valence of tin from Sn2+ to Sn4+ , induces the significant diffusion of tin. Furthermore, the precipitation of crystalline SnO2 particles with a diameter of `1 nm was clearly recognized in the tin - enriched layer. This fact indicates that a phase separation was induced by the oxygen diffusion into the glass. Consequently, Sn2+may be supplied to the surface in order to compensate for the marked decrease in Sn2+ concentration in the glass system. The significant diffusion of tin to the surface was suppressed by increasing the iron content in the glass. This suppression was ascribed to the increase in Sn4+ concentration as a result of the redox - reaction between tin and iron because the diffusion coefficient of Sn4+ is much smaller than that of Sn2+ .

*Research Center **Tokyo Institute of Technology
@

–ΪŽŸ
@ 1. Introduction
@ 2. Experimental
@ 3. Results
@ 4. Discussion
@ 5. Conclusions

@
‚P. Introduction
@Float glass is most widely used in industrial uses for architectures, automobiles and displays such as liquid crystalline displays (LCDs)or plasma display panels (PDPs) because of its high productivity and excellent flatness over a large area. In the float process, molten glass is floated on a molten tin bath, so that tin is penetrated into the glass. It is known that the tin is not uniformly diffused into the glass and the depth profile has an anomalous hump(1)|(5) .
@When the float glass is used for architectural or automotive applications, it is sometimes tempered or bent by thermal toughening process. Occasionally, the appearance of the glass often becomes milky or hazy by the process, which is called gbloom h (1) . This is a serious problem for the glass manufacturing industry because the transparency is lost. Therefore, it is important to clarify the formation mechanism of the bloom. This phenomenon is observed only for the bottom face of the glass, indicating that the diffused tin may be closely related to the formation of the bloom. Until now, many researchers have extensively studied the phenomenon, and reported that the tin - enrichment layer and wrinkling structure were detected near the surface region after the heat treatment (3)(6)(7) . These findings suggest that the bloom may be induced by the formation of the tin - enrichment layer. However, the mechanism of the significant tin enrichment is not clearly elucidated so far.
@The purpose of the present study is to clarify the mechanism of the significant tin diffusion to the surface. A part of this work has been reported in Ref (8)(15) . Here, we investigated the effect of oxygen diffusion from the atmosphere on the tin diffusion. The heat treatment was performed in 18O2/N2 and argon (Ar)atmospheres. The depth profiles of tin and 18O for the bottom face of the glass were measured using secondary ion mass spectrometry (SIMS). In this experiment, the behavior of oxygen from the atmosphere could be traced in detail even if much oxygen is present in the glass because the oxygen tracer ( 18O)gas was used. Namely, direct information can be obtained about the oxygen and tin depth profile. The tin - enriched layer was explored by transmission electron microscopic (TEM)observation and selected area diffraction (SAD)analyses.
@In addition, we also investigated the effect of iron on the tin diffusion. It is known that the impurity iron is commonly present in the glass, and that the iron affects the chemical states of tin (9)(10)(11) , causing the change in tin depth profile. From the results obtained, the behavior of tin in the bottom face of the float glass by the heat treatment was discussed.
@
£Page Top

@
2. Experimental
@Commercial soda - lime - silica float glasses were used in this study. The compositions of major elements in the glasses are listed in Table 1. These glasses were manufactured on the same float line although the iron content was different. The glass transition temperature (Tg)of the glasses is 562 Ž. A schematic illustration of the experimental apparatus for the heat treatment was described in Ref (8) . The bottom face of the glasses was set upward, then the glasses were heat - treated at 740 Ž(}2 Ž)for 15 min in gas mixture ( 18O2/N2=1/4)and in argon (Ar)atmosphere. Prior to the heat treatment, the sample chamber (Quartz glass vessel)was purged with Ar gas (99.999% purity)flow for 5 hr in order to diminish the residual oxygen. Then, the vessel was evacuated and then gas mixture ( 18O2/N2, 99%purity for 18O and 99.999%purity for N2 )or Ar gas was introduced in the vessel up to the atmospheric pressure.
@Observation of surface morphology and the quantitative analysis of the surface roughness of the glass were performed using an atomic force microscope (AFM). The optical transmission spectra of the glass were measured at room temperature in air using a dual beam spectrometer. The appearance change of the glass due to the heat treatment was evaluated as haze value change using a haze meter. The haze value is defined as Td / Tt ~100%(Td;scattered light, Tt; transmitted light).

@The depth profiles of 120Sn, 18O and 54Fe for the bottom face of the glass were measured using SIMS. Positive secondary ions were detected using an O2+ primary ion beam operated at 8 keV, 100nA. The angle of incidence was 60 ‹to the normal of the sample surface and an area of 100 ~100 ƒΚm 2 was sputtered. Negative secondary ions were detected using a Cs+ primary ion beam operated at 6 keV, 20 nA. The angle of incidence was 45 ‹to the normal of the sample surface and an area of 400 ~400 ƒΚm2 was sputtered. The charge neutralization was accomplished using an electron flood gun. The etching rate was determined by using a contact probe level difference meter after SIMS measurements. The accuracy of the apparatus was within }5nm. The tin - enriched layer near surface region of the glass was observed by TEM.The crystalline phases of the layer was identified by selected area electron diffraction (SAD)analyses. The chemical compositions of the layer were determined by X - ray photoelectron spectroscopy (XPS).
@
£Page Top

@
3. Results
3.1 Effect of heat treatment atmosphere on tin depth profile
@Table 2 shows the haze value for Sample - A heat - treated in the gas mixture ( 18O2/N2 )or Ar atmosphere. It is found that the haze value significantly increases after the heat treatment in the gas mixture. This phenomenon is not observed when heat - treating in Ar atmosphere. This result suggests that the increase in haze value may be due to the incorporation of oxygen into the glass surface from the atmosphere.
@Fig.1 shows the AFM images of Sample - A before and after heat treatment in 18O2/N2 or Ar atmosphere. The wrinkling structure is observed on the surface of the glass heat - treated in 18O2/N2 atmosphere. The surface roughness (Rms) estimated from the image is `50 nm. This result indicates that the haze increase, as shown in Table 2 , is due to the geometrical light scattering by the wrinkling structure. The wrinkling structure is not recognized for the sample heat - treated in Ar, suggesting that the formation of the wrinkling structure is concerned with oxygen in the atmosphere.
@Fig.2 shows the SIMS depth profiles for Sample - A before and after heat treatment in 18O2/N2 or Ar atmosphere. Before the heat treatment, a hump in tin profile is clearly observed around at the depth of `3 ƒΚm from the surface. After the heat treatment in 18O2/N2 atmosphere, the significant diffusion of tin to the surface is recognized, resulting in the formation of a tin enriched layer near the surface region. The tin profile shallower than the hump distinctly changes after the heat treatment. This means that the diffused tin to the surface is supplied from the region shallower than the hump. No significant change in tin depth profile is observed for the glass heat - treated in Ar atmosphere.
@Fig.3 shows the SIMS depth profile of 18O and 120Sn near surface region for the glass before and after the heat treatment in 18O2/N2 atmosphere. The diffusion of 18O into the glass is also observed and the amount of incorporated 18O is larger by two orders of the magnitude than that without the heat treatment. The diffused depth of 18O, which is defined as the depth that the secondary ion intensity reach to the same level as that without the heat treatment, is `300 nm of the surface. The depth at the maximum concentration of 18O is almost the same as that of tin.
@Fig.4 shows (a)cross - sectional TEM image and (b)SAD pattern of the tin - enriched layer. The precipitation of nanometer particles is clearly observed in the tin - enriched layer and the average size of the particles is `1 nm. The SAD pattern reveals that these particles are crystalline and identified as SnO2. The chemical compositions of the tin - enriched layer determined by XPS is:58.0 - SiO2, 16.9 - Na2O, 6.5 - CaO, 18.2 - SnO2 (wt %), which significantly differs from that of the glass before the heat treatment.


3.2 Effect of iron on tin depth profile
@Fig.5 shows the SIMS depth profile for Sample - A and B heat - treated in 18O/N2 atmosphere. These glasses were manufactured on the same float line, and the iron concentration of Sample - B is larger than that of Sample - A, as shown in Table 1.It is found that the depth at the maximum concentration of tin is slightly shifted to the deeper side for Sample - B compared with that of Sample - A. This indicates that the diffusion of tin to the surface is suppressed by increasing the iron content in the glass. Furthermore, the surface iron concentration of Sample - B is larger than that of Sample - A. No significant change was observed in
 
£Page Top

@
4. Discussion
4.1 Mechanism of significant tin diffusion to the surface
@As shown in Fig.2, the significant diffusion of tin to the surface is distinctly observed after the heat treatment in 18O2/N2 atmosphere. In addition, the tin depth profile shallower than the hump position markedly changes. These results suggest that the tin is supplied from a shallower region than the hump. As mentioned before, it is known that the chemical states of tin are changed with depth, and that the depth profile shows an anomalous hump (1)|(5) . According to the research of Williams et al. (3) using M oN ssbauer spectroscopy, the majority of the tin near the surface exists as Sn 2+ . Nomura (5) also investigated the Sn2+ /Sn4+ ratio with depth, and reported that almost all the species were Sn4+ at the deeper layers than the hump. Furthermore, it is known that the diffusion coefficient of Sn4+ is much smaller than that of Sn2+ . These suggest that the diffused tin to the surface is considered to be stannous tin (Sn2+ ).
@On the other hand, no marked change in the tin depth profile is observed for the glass heat - treated in Ar atmosphere, indicating that the significant diffusion of tin is induced by oxygen diffusion from the atmosphere. As is seen in Fig.3, the depth of diffused 18O is `300 nm of the surface. This fact suggests that the effect of oxygen diffusion on the oxidation states of tin is within `300 nm. It is known that the small amount of Sn2+ is oxidized to Sn4+ by the heat treatment in air(3) .

@The precipitation of crystalline SnO2 particles with a diameter of `1 nm is clearly observed in the tin - enriched layer, as shown in Fig.4.This observation indicates that a phase separation occurs in the tin - enriched layer. Here, it is known that the coordination structure of oxygen around tin is different between Sn2+ and Sn4+ in tin - soda - lime - silica glass system, and that Sn4+ is in an octahedral coordination (the bond is ionic)and Sn2+ is in tetragonally - pyramidal coordination (the bond is rather covalent) (12) . This suggests that the change in oxidation state of tin from Sn2+ to Sn4+ by oxygen diffusion (3) may subsequently cause the change in the coordination structure of tin from tetragonally - pyramidal to octahedral coordination. This valence change is rather drastic because the valence increases by 2 and the bonding nature changes from covalent to ionic. Thus, the formation of non - bridging oxygens is required so as to meet the local electroneutrality around Sn4+ . These suggest that Sn4+ ion is not compatible with silica network. As a consequence, phase separation is induced (13) . In fact, it is known that SnO2 - SiO2 glass system is phase - separated into SnO2 and SiO2:Sn phases when the SnO2 concentration increases (14) .
@Based on these analyses, we conclude that the significant diffusion of tin should result from the concentration decrease in Sn2+ ions as a result of elimination of Sn ions from the supercooled liquid state, i. e. , when O2 gases from air are diffused into the tin - enriched layer to oxidize Sn2+ into Sn4+ , resulting Sn4+ ions are precipitated as SnO2 crystals, and Sn2+ ions in the region shallower than the hump are diffused towards the surface layer to compensate the drop of Sn2+ ions concentration in the layers. Consequently, the wrinkling structure on the surface is formed, as shown in Fig.1 (b), as a result of mismatching of the thermal expansion coefficient between the tin - enriched layer and the glass.


4..2 Iron effect on the tin diffusion
@As is seen in Fig.5, it is found that the significant diffusion of tin to the surface is slightly suppressed by increasing the iron content in the glass. It is known that the iron in the glass can react with tin, as follows (9)(10)(11) ,

Sn2+ + 2 Fe3+ ¨ Sn4+ + 2 Fe2+

Namely, the increase in iron content causes the increase of Sn4+ concentration in the glass. It is considered that increasing the Sn4+ concentration decreases the mobility of tin because the diffusion coefficient of Sn4+ is much smaller than that of Sn2+ . Namely, the suppression of the tin diffusion, as shown in Fig.5 , is due to the increase in Sn4+ content as a result of redox reaction between Sn2+ and Fe3+ . This result suggest that adding the oxidizing species for Sn2+ such as Fe3+ is useful for controlling the tin diffusion by the heat treatment.
@
£Page Top

 
5. Conclusions
@In this paper, we have investigated the effect of oxygen diffusion from the atmosphere on tin concentration profile of a soda - lime - silica float glass at temperatures above Tg. The significant diffusion of tin to the surface was observed for the glass heat - treated in 18O2/N2 atmosphere, resulting in the formation of tin - enriched layer near the surface  region. It was found that the precipitation of nanometer - sized SnO2 crystals was clearly recognized in the tin - enriched layer, indicating that a phase separation was induced by the oxygen diffusion into the glass. Based on these analyses, we conclude that the tin (Sn2+ )was supplied from the region shallower than the hump in order to compensate for the marked decrease in Sn 2+ concentration of the glass system. Furthermore, it was also found that the significant diffusion of tin was suppressed by increasing the iron content in the glass. The suppression was due to the decrease in the mobility of tin, which was induced by the increase in Sn4+ concentration as a result of the redox - reaction between tin and iron, because the diffusion coefficient of Sn4+ is much smaller than that of Sn2+ . Therefore, to add the oxidizing species for Sn2+ such as Fe3+ is useful for controlling the tin diffusion by the heat treatment.
@We expect that the present findings offer a novel clue to control of the tin depth profile and to improvement of the thermal durability of float glass.

|References |
(1) L. A. B. Pilkington, The float glass process, Proc. Roy Soc. Lond, A314 , 1 (1969).
(2) J. S. Sieger, J. Non - Cryst. Solids, 19, 213 (1975).
(3) K. F. E. Williams, C. E. Johnson, J. Greengrass, B. P. Tilley, D. Gelder, J. A. Johnson, J. Non - Cryst. Solids, 211 , 164 (1997).
(4) K. F. E. Williams, C. E. Johnson, O. Nikolov, M. F. Thomas, J. A. Johnson, J. Greengrass, J. Non - Cryst. Solids, 242 , 183 (1998).
(5) K. Nomura, gM oN ssbauer spectroscopy in Material Science h, ed by M. Miglierini and D. Petridis (Kluwer Academic Publishers, the Netherlands, 1999), p. 63.
(6) J. Deubener, R. Bruckner, H. Hessenkemper, Glastech. Ber. , 65 , 256 (1992).
(7) C. G. Pantano, V. Bojan, Supplement to Rivista della Stazione Sperimentale del Vetro, Vol XX E, 1993, p. 285.
(8) S. Takeda, R. Akiyama, H. Hosono, J. Non - Cryst. Solids, 281 , 1 (2001).
(9) A. Kumar, S. P. Singh, R. Pyare, Glastech. Ber. , 64 , 106 (1991).
(10) K. F. E. Williams, M. F. Thomas, C. E. Johnson, B. P. Tilley, J. Greengrass and J. A. Johnson, gFundamentals of Glass Sci. and Technol h. 1997, p. 127.
(11) P. D. Townsend, N. Can, P. J. Chandler, B. W. Farmery, R. Lopez - Heredero, A. Peto, L. Salvin, D. Underdown, B. Yang, J. Non - Cryst. Solids, 223 , 73 (1998).
(12) H. Kawazoe, J. Nishii, H. Hosono, T. Kanazawa, H. Imagawa, J. Phys. , C9 , 156 (1982).
(13) H. Rawson, gInorganic Glass - Forming System h, Academic Press, London and New York, 1967.
(14) C. J. Brinker, G. W. Scherer, gSol - Gel Science h, Academic Press, 1990;V. V. Sidorchuk and V. M. Chertov, Sov. J. Inorg. Mater. , 22 , 1692 (1986).
(15) S. Takeda, R. Akiyama, H. Hosono, J. Non - Cryst. Solids, in press.
@
£Page Top

back next
 
COPYRIGHT(C) 1996-2004 ASAHIGLASS COMPANY. ALL RIGHT RESERVED.