The results of the considered raw‐material analysis is reported in the shaded gray area of the matrix.
Table 3 Redox factors R(i) of selected active raw materials i; these factors refer to batch compositions normalized to a sand amount of 2000 kg.
| Raw material i | Chemical formula | R(i) per 2000 kg sand |
|---|---|---|
| Carbon | C : 100, 85, 65% | −6.70 |
| Iron sulfide | FeS | −1.60 |
| Pyrite | FeS2 | −1.20 |
| Fluorspar | CaF2 | −0.10 |
| Calumite | Multicomponent slag | −0.073 |
| Iron red | Fe2O3 | +0,25 |
| Chili saltpeter | NaNO3 | +0.32 |
| Heavy spar | BaSO4 | +0.40 |
| Gypsum | CaSO4·2 H2O | +0.56 |
| Potassium dichromate | K2Cr2O7 | +0.65 |
| Salt cake; sulfate | Na2SO4 | +0.67 |
| Gypsum anhydrite | CaSO4 | +0.70 |
| Sodium dichromate | Na2Cr2O7 | +0.77 |
| Manganese oxide | MnO2 | +1.09 |
Final adjustment of the batch composition still requires allotments of the appropriate agents for controlling glass color, fining (as described in Section 5.2), redox conditions, and, thus, valence states and oxygen complex formation of polyvalent ions (cf. Chapter 5.6) at the industrial scale. For adjustment of the redox state, the so‐called redox number concept [2] is widely accepted and empirically applied in industry. This incremental system assigns a specific redox factor Ri to every member of a set of redox‐active ingredients i (Table 3). In the example of Table 4, the batch composition from Table 2 is complemented by 4 kg of sulphate (the amount of soda ash being reduced accordingly to maintain an identical amount of Na2O in the glass). Then the batch composition is normalized to amounts mIII(i) per 2000 kg of sand, and the total redox number of the batch is calculated from the weighted sum R = ∑ Ri·mIII(i). It is true that this number R does not have a straightforward scientific meaning but it allows one to set in a well‐defined way the redox state to a desired level. For redox numbers in the interval −25 < R < 25, a fair estimate of the Fe2+/Fetotal ratio for is given by 0.4 – 0.015·R. Because the chemical composition of the batch can no longer be corrected after charging, these rather simple calculations are mandatory for any successful melting process.
Finally, the raw materials are automatically weighed in proportions determined by batch calculations, conveyed to a mixer, and thoroughly mixed. During mixing, typically 2–3% of water is added to suppress dust formation and segregation induced either originally by transportation or subsequently by mixing. In batches containing soda ash, small amounts of this product dissolve in the water before reprecipitating on other batch grains. This process termed “impregnation” actually enhances the kinetics of batch melting.
4 The Conversion of Batch into Melt
4.1 The Basic Importance of Convection
In principle, the melting compartment (the tank) is a shallow basin whose typical dimensions (length L × width W × depth D) are 10 × 6 × 1 m3 for medium‐size container‐glass furnaces and 30 × 10 × 1.4 m3 for float‐glass furnaces. The required energy is delivered in a combustion space right above the melting compartment and transferred to the melt chiefly by black‐body radiation. Additional energy (5–20%) is delivered by direct electrical heating (boosting) of the melt.
The very melting process, i.e. steps M1–M3, takes place in one single box‐shaped compartment as sketched in Figure 2a, b. The sketches illustrate some essential features only (see Chapter 9.7 for details). Spatially, the individual process steps are separated – not in a rigorous but effective way – by two convective vortices. In furnaces with transversal flame direction (so‐called side port fired furnaces; flow pattern Figure 2a), these are mainly generated thermally by an appropriate distribution of fuel input to a series of burners arranged along the L axis above the melt. This results in two well‐developed vortices with a well‐localized hot spot at the position of maximum energy input. In furnaces with longitudinal flame direction (end port fired furnaces; flow pattern Figure 2b), the energy input via combustion along the L axis cannot be controlled at the same distinction as in a side port fired furnace. Here, the flow pattern is typically dominated by a single predominant vortex; the hot spot is shifted toward the end of the furnace. Local stabilization of the vortices is achieved by electrical heating from below, and – for the second type of furnaces – also mechanically by the action of air bubblers or by implementation of a solid barrier (a wall) at the bottom of the tank. As seen from Figure 2, the batch floats on the surface of the molten phase and melts continuously in the L direction, thereby typically covering an area whose length is about L/3, or more than 2L/3 for side and end port firing, respectively. Processes M2 (sand dissolution) and M3 (fining) then take place in the first and second vortices, respectively. Note