Reinjection of formation waters creates few chemical problems if no large change in temperature has occurred, no gas or vapor has separated, and access of air has been prevented. The fluids already have had an opportunity to react with the minerals to the point of compatibility. Injection of incompatible fluids, however, may cause chemical problems; thus, prediction of fluid-mineral reactions should be attempted.

Real geologic systems chosen where reactions are known to occur, and where the reactant minerals and fluids and product minerals and fluids can be identified and analyzed, rarely have stable equilibrium (minimum Gibbs free energy) conditions. In some situations, metastable equilibrium conditions have been found. The assumption of a stable equilibrium state is an unwise assumption for a model, especially at temperatures of 100°C and below.

The general lack of attainment of equilibrium in no way impairs prediction of reactions, however. Admittedly, there is no a priori thermodynamic method of predicting what phase will react, nor can the necessary departures from equilibrium (G_R) be predicted before a reaction will occur at a significant rate, because thermodynamic arguments are time independent. However, in geologic systems now accessible, enough reactions are occurring that, by observation, an empirical knowledge will provide a base for predicting reactions.

Reactions should be studied in both reaction directions. Most reactions are asymmetric in that the G_R required to dissolve most phases is of different values than the G_R necessary to form the mineral from solution. The reactions should all be treated as congruent. All the species in a solution generated by complete solution of the solid must be considered. Using incongruent reactions (reactions producing a new solid directly from the reactant solid) introduces the unwarranted assumption of equilibrium.

In general, hydroxide species such as amorphous silica, limonite, and brucite, as well as simple carbonates, have been found to dissolve or precipitate in natural systems where G_R values are 1 kcal or less. More complex silicates including serpentine, kenyaite, and magadiite may not precipitate from solutions where G_R values favor supersaturation by as much as 7 kcal. Sulfides such as covellite, chalcocite, and pyrite may not precipitate even where supersaturation exceeds 30 kcal, although some sulfide minerals will dissolve readily where oxidation of sulfur can take place.