Simulation of tin penetration in the float glass process 

The flat glass produced by the float glass process has a tin-rich surface due to the contact with molten tin. The penetration of tin into the glass surface is assumed to involve coupled diffusion of stannous (Sn2+) and stannic (Sn4+) ions. The diffusion coefficients of these ions were calculated using the modified Stocks–Einstein relation with the oxidation velocity of stannous ions depending on the oxygen activity in the glass. The ion diffusion was analyzed using a coupled diffusion simulation with a modified diffusion coefficient to compensate for the negative effect of the glass ribbon’s stretching or compressing in the glass forming process. Tin penetration simulations for both green glass and clear glass show an internal local tin concentration maximum in green glass which is quite different from that in clear glass. The local maximum in the profile is associated with the accumulation of stannic ions where the greatest oxygen activity gradient occurs. Since more float time is needed in the manufacture of thicker glass plate, the tin penetrates to a greater depth with the maximum deeper in the glass and the size of the maximum larger for thicker glass.

The float glass process, which was originally developed by Pilkington Brothers in 1959 (Haldimann et al., 2008), is the most common manufacturing process of flat glass sheets. More than 80–85% of the global production of float glass is used in the construction industry (Glass for Europe, 2015a). In the float glass process, the ingredients (silica, lime, soda, etc.) are first blended with cullet (recycled broken glass) and then heated in a furnace to around 1600°C to form molten glass. The molten glass is then fed onto the top of a molten tin bath. A flat glass ribbon of uniform thickness is produced by flowing molten glass on the tin bath under controlled heating. At the end of the tin bath, the glass is slowly cooled down, and is then fed into the annealing lehr for further controlled gradual cooling down. The thickness of the glass ribbon is controlled by changing the speed at which the glass ribbon moves into the annealing lehr. Typically, glass is cut to large sheets of 3 m × 6 m. Flat glass sheets of thickness 2–22 mm are commercially produced from this process. Usually, glass of thickness up to 12 mm is available in the market, and much thicker glass may be available on request. A schematic diagram of the production process of float glass is shown in Fig. 5.2.The float glass process was invented in the 1950s in response to a pressing need for an economical method to create flat glass for automotive as well as architectural applications. Existing flat glass production methods created glass with irregular surfaces; extensive grinding and polishing was needed for many applications. The float glass process involves floating a glass ribbon on a bath of molten tin and creates a smooth surface naturally. Floating is possible because the density of a typical soda-lime-silica glass (~2.3 g/cm3) is much less than that of tin (~6.5 g/cm3) at the process temperature. After cooling and annealing, glass sheets with uniform thicknesses in the ~1–25 mm range and flat surfaces are produced. The ultra clear float glass process is used to produce virtually all window glass as well as mirrors and other items that originate from flat glass. Since float glass is ordinarily soda-lime-silica, the reference temperatures and behavior of this glass are used in the discussion below.

Figure 3.48 shows the basic layout of the clear float glass line. The glass furnace is a horizontal type, as described above. For a float line, the glass furnace is typically on the order of ~150 ft long by 30 ft wide and holds around 1200 tons of glass. To achieve good chemical homogeneity, the glass is heated to ~1550–1600°C in the furnace, but is then brought to about 1100–1200°C in the forehearth. From there, the glass flows through a channel over a refractory lipstone or spout onto the tin bath. As it flows, the glass has a temperature of about 1050°C and viscosity of about 1000 Paradical dots. A device, called a tweel, meters the flow of the molten glass.Imperfections include bubbles (or ‘seeds’) that may have a number of possible sources, the most common being gas evolved during firing. Bubbles may contain crystalline materials formed during cooling of the glass that may provide clues to the origin of the bubbles. Cords are linear features within the glass that may result either from imperfectly homogenized raw materials, dissolved refractories or devitrified material. Figure 360 shows the appearance of soda–lime–silica glass that exhibits bubbles and cords. ‘Stones’ are solid crystalline substances occurring in glass that are regarded as defects. They are usually derived either from the batch material, refractories, or devitrification. Figure 361 shows the appearance of soda–lime–silica glass that contains a devitrification ‘stone’. These may develop as the result of incomplete mixing of the molten glass constituents and/or too low a firing temperature. The ‘stone’ shown in Figure 361 contains an aggregation of tridymite crystals (see 362).

As the floating glass ribbon traverses down the length of the tin bath, its properties change dramatically. The glass enters as a viscous liquid and exits virtually a solid at a temperature very close to its glass transition temperature. The details of how the temperature changes and the viscosity builds are complicated. On one side, the free surface of the glass is exposed the atmosphere; heat can leave this surface by radiation or convection. Cooling and heating apparatuses are stationed above the glass ribbon down the length of the bath to allow adjustment of the ribbon temperature. On the other side, the glass is in contact with the tin bath, which can absorb some of the heat and transport it away from the ribbon. The tin bath is in constant motion due to the moving glass above it as well as the thermal convection currents. Unfortunately, no simple approximations can be made to make the modeling of the heat transfer.

The thickness of the tinted float glass sheet is adjusted by controlling flow onto the tin bath as well as by tension exerted along the length of the bath by rollers in the annealing lehr and sometimes by rollers in the bath unit itself. In the Pilkington design, the melt enters the bath and spreads out laterally to a thickness near the equilibrium value. If a sheet thicker than the equilibrium is required, then this spreading is constrained with physical barriers. If a sheet thinner than equilibrium is needed. then the glass ribbon is pulled in tension by rollers. In the PPG design, thickness is regulated by the tweel position and by tension from rollers in the lehr. The thermal profile allows the thinning deformation to take place effectively. A short distance away from the entry point, the temperature of the ribbon drops and the viscosity rises. Overhead coolers help this process. The glass viscosity is high enough so that knurled rollers contact the glass ribbon and pull it forward (and in some operations, laterally as well). Heaters are placed shortly downstream of these edge rollers to raise the temperature of the ribbon and create a deformable zone. This zone is followed by coolers that again lower the temperature and raise the viscosity. At