(1)
where Sga, Sgt, and Sta are the surface tensions at the glass–atmosphere, glass–molten tin, and tin–atmosphere interfaces, respectively, g is the gravitational constant and ρt and ρg are the density of the molten tin and glass, respectively (Table 2). For soda‐lime silicate glass floating on clean molten tin under a nitrogen‐hydrogen atmosphere, Te is 6.9 mm (Figure 8), a thickness that is actually insensitive to small changes in the chemical compositions of the atmosphere, metal bath, or glass [1, 3–9].
5.2 Float Bath
The float bath contains several metric tons of molten glass. It is a large unit with a length of up to more than 50 m enclosed by a steel shell that is lined with thick insulating and nonreactive refractory materials and holds a pool of molten tin whose depth is 50–100 mm and total amount is up to more than 200 metric tons kept at temperatures decreasing from about 1000 to 600 °C from the hot to the cold end (Figure 9). A reducing gas mixture made up of 2–8% hydrogen and 98–92% nitrogen is supplied at a high rate of the order of 103 m3/h from above to the bath to prevent oxidation of the molten tin and to maintain a positive pressure difference with the atmosphere at the bath exit where leakages are highest. The heaters, coolers, and other devices are installed and inserted in the bath. The molten glass is continuously supplied from the furnace conditioner via a canal where its flow rate is precisely controlled by an adjustable gate called a tweel. It arrives to a ceramic spout lip, which is an inlet of the float bath, through which it falls freely onto the molten tin. After many years of struggle at Pilkington Brothers to achieve excellent quality, the design and engineering of the inlet area were an outstanding invention to force the contaminated molten glass in contact with the refractory lip to flow outwardly so as to be brought forward at the outer edges of the ribbon [9].
Table 2 List of symbols regarding equilibrium thickness mechanism.
| Symbol | Denotation |
|---|---|
| T e | Equilibrium thickness |
| ρ t | Density of molten tin |
| ρ g | Density of molten glass |
| S ga | Surface tension at glass–atmosphere interface |
| S gt | Surface tension at glass–molten tin interface |
| S ta | Surface tension at tin–atmosphere interface |
| g | Gravitational constant |
Figure 8 Equilibrium thickness of floating glass on the molten tin when the gravitational forces and surface tensions are balanced [3].
Once poured onto the tin bath with a thickness of about 50 mm, the glass spreads out and thins to its equilibrium thickness in the upstream area in the float bath. As formed to the required thickness and width in the forming area (see Section 5.3), the glass ribbon is taken out from the bath either to receive appropriate reflective, low‐emissivity, solar‐control, self‐cleaning, or other specific coatings (Chapters 6.7 and 6.8) or to enter directly the annealing lehr at the temperature at which the viscosity is about 1010 Pa·s (i.e. about 600 °C for soda‐lime silicate). At the end of the lehr, whose length can reach 120 m, the ribbon is finally cooled down to room temperature and brought into the cutting area. Whereas both edges are cut out (to be recycled as cullet) because of the imprint left by the top rolls, the ribbon itself is cut either according to customers' specifications or as standard sheets, for instance, 6.0 × 3.21 m in Europe where tools used in the flat‐glass transportation industry have been fitted to this size (which, by the way, is too large to allow flat glass to be shipped in containers).
5.3 Thinner (Top‐Roll Process) and Thicker (Fender Process) Glass Ribbons
For forming thin sheets, the molten glass with its initial equilibrium thickness in the upstream area in the bath is subjected at the same time to longitudinal and lateral forces. The former are exerted by conveyor rolls that stretch the ribbon from the annealing lehr and pull it at a typical speed of up to 25 m per minute. The latter are exerted outwardly on the ribbon edges by pairs of top rolls, which are water‐cooled rotating gears, to reduce the narrowing of the glass ribbon because the imposed longitudinal stretching reduces not only its thickness but also its width (Figure 10a). In parallel, the glass ribbon is cooled down to prevent it from returning to its equilibrium thickness until its width is constant at the end of the forming area. In view of the fundamental influence of viscosity within the glass ribbon upon stretching and thinning, the temperature distribution and the top‐roll operations must be controlled very tightly to ensure a good forming quality. Besides, keeping the glass ribbon as wide as possible is important to maximize productivity.
For producing float glass thicker than the equilibrium thickness, a pair of water‐cooled carbon fenders serves as slipping guides to the flowing glass in the bath (Figure 10b). The glass thus proceeds with a restricted width and a large thickness. As it passes down the fender area, the effects of gravitational forces and surface tensions make both its upper and lower surfaces flat and thickness uniform. The glass is then cooled to an appropriate temperature in the downstream area of the fender where its viscosity is high enough not to allow width changes. In contrast to what is taking place in the top‐roll process, stretching is not significant at all and there is no drive to return to the equilibrium thickness because there is no glass–tin–atmosphere interface in the fender area.
Figure 9 Sketch of the tin bath part of the float process: (a) on vertical plane along centerline; (b) on horizontal plane. A reducing nitrogen–hydrogen gas mixture is supplied from above. Heaters