The impact of using a ternary parameter L, defined by the excess Gibbs energy, Gexc= L·x1·x2·x3, is elaborated. It is emphasized that this impact is very much counter-intuitive.
For example, the activity of Hg in the Bi-Sn-Hg system at 350°C along the section xBi = xSn is shown as the solid line in Fig. 1, calculated without a ternary parameter (L=0). In this example, we have calculated aHg from the Redlich-Kister (Mugg ianu) extrapolation of the binary data, but that is completely unimportant for the following. We get the same impact of L for any other extrapolation scheme.
It is is not possible to push the activity curve aHg down (or up) for all alloys on that section by using any value of this simplest ternary L-parameter. An increase of aHg for xHg <0.5 always comes with a decrease of aHg for xHg >0.5 (or vice versa) and aHg (xHg=0.5) is constant for all values of L. This impact of L is very much against intuition, see Fig. 1.
This is a general feature. In any ternary
1-2-3 system the ternary excess term L·x1·x2·x3
leaves a1 unchanged along the section x1=0.5.
The reason is that
= L·x2·x3·(1-2x1).
Conclusion: This ternary L-parameter should not be used as a fitting tool for phase diagrams without carefully checking its impact on activities. It is better to not use it at all, but rather improve on the binary descriptions.