Silicate melts activities

The viscosity of liquid silicates such as drose containing barium oxide and silica show a rapid fall between pure silica and 20 mole per cent of metal oxide of nearly an order of magnitude at 2000 K, followed by a slower decrease as more metal oxide is added. The viscosity then decreases by a factor of two between 20 and 40 mole per cent. The activation energy for viscous flow decreases from 560 kJ in pure silica to 160-180kJmol as the network is broken up by metal oxide addition. The introduction of CaFa into a silicate melt reduces the viscosity markedly, typically by about a factor of drree. There is a rapid increase in the thermal expansivity coefficient as the network is dispersed, from practically zero in solid silica to around 40 cm moP in a typical soda-lime glass. 

The plot in figure 10.8 clearly shows that silicate melts may exert marked fractionation on REE acidic melts with relatively greater free oxygen activity display the highest mean activity coefficients in fight REE (TREE). LREE are thus less soluble in these melts than heavy REE (HREE). The opposite is observed in basic melts, which have the highest mean activity coefficients at the HREE level. In other words, REE relative concentrations in silicate melts simply reflect their relative solubilities, as we saw for crystal structures. 

Figure 10,8 Mean activity coefficients of various REE in silicate melts with different values of activity of dissolved free oxygen Uq -, plotted as a function of increasing atomic number. Reprinted from Ottonello (1983), with kind permission of Theophrastus Publishing and Proprietary Co.

Fraser D. G. (1975a). Activities of trace elements in silicate melts. Geochim. Cosmochim. Acta, 39 1525-1530. 

Richardson F. D. (1956). Activities in ternary silicate melts. Trans. Faraday Soc., 52 1312-1324. 

Toop G. W. and Samis C. S. (1962b). Activities of ions in silicate melts. Trans. Met. Soc. AIME, 224 878-887. 

In silicate melts, typical D values at 13(X)°C are of the order 10 m /s, and typical activation energy is 250 kj/mol. Highly charged cations (such as Si " and Zr +) or strongly bonded ions (such as bridging oxygen) diffuse more slowly and... 

The thermodynamics of corrosive alkali salt-oxide interaction is not well established. In an assessment of research needs for materials in coal conversion, the need for carbonate-silicate melt studies, including activity and phase equilibrium measurements, was stressed (31 ). The lack of thermodynamic data for fused salts, and their reactions with oxides and alloys leading to models of hot corrosion, was also indicated. 

Although there is no general model for activity-composition relationships in silicate melts, several studies have shown that, under certain circumstances, a semi-empirical approach works reasonably well. Blundy et al. (1995) investigated the partitioning behavior of sodium between clinopyroxene and silicate melts over wide ranges of pressure and temperature. They showed that the crystal-liquid partition coefficient Dfja bears a very simple relationship to the equilibrium constant Xj,ja for the melting reaction ... 

The behavior of some siderophile elements is controlled by the composition of the metal, and, in particular, the Fe/Ni ratio. Because the activity coefficient of a siderophile element, M, may be different in a nickel-rich metal than an iron-rich metal, an understanding of this effect is necessary before partition coefficients can be successfully applied to a natural system. A good example is tin, which has a low activity coefficient in iron-rich metal, and a high activity coefficient in nickel-rich metal (Figure 8 Capobianco et al., 1999). Metal composition can change as a function of/o, and this can then change the solubility in the silicate melt. In addition, the Fe/Ni ratio in metal is sometimes used to impose a specific/o on a system. These effects are linked and must be unraveled first in order to understand tin in planetary mantles (Righter and Drake, 2000). 

Ryerson F. J. (1985) Oxide solution mechanisms in silicate melts systematic variations in the activity coefficient of Si02. Geochim. Cosmochim. Acta 49, 637—649. 

Activities of individual components are calculated on the basis of the theory of conformal solutions (Reiss et al., 1962). This theory was derived for ionic systems without the formation of complex ions, in which both anions and cations have identical charges. This theory was later applied to systems containing ions of various valences (Saboungi and Blander, 1975). It should be emphasized that the application of this theory to silicate melts has only a formal character. 

Let us define the polymerization degree of silicate melts as the fraction of bridging oxygen atoms in a single formula unit. The assumptions, on the basis of which activities of components in the mixture are calculated, are as follows ... 

Equation (3.133) for the activity of a component derived in this way is completely universal and may be used in any system. The calculation of the activity of a component in an actual system using this equation thus respects the characteristic features of silicate melts, e.g. the different energetic state of individual atoms of the same kind. Such cases may... 

The energy barrier, U, that molecules have to overcome in order to become attached to a nucleus surface is of importance in the solidification of silicate melts and organic liquids, especially of polymeric substances. In this case U signifies the activation energy in the process of the diffusion of a molecule (or its segments) from the bulk of the liquid phase to the surface of a nucleus. A drastic decrease in the diffusion rate in such liquids, related to the increase in viscosity as the temperature is lowered, causes a maximum to appear in the curve representing nuclei formation frequency as a function of temperature. The position of this maximum corresponds to some supercooling, AT as shown in Fig. IV-9. ... 

---