#12 Total Gibbs increases for DebeyeHuckel object
The total Gibbs free energy is shown to increase during some iterations of the equilibrium solver. Some obvious issues have been eliminated as possibilities for why this is happening.
Note, the final state of the solver seems to be fine. delta G is equal to zero for all reactions, and the state may be achievable by various different paths. However, it's just the case that some paths exhibit a lower total G than the final solution, even though delta G_i isn't equal to zero at those spots.
It appears that this behavior is limited to cases where the DebyeHuckel object is used. This may indicate that the Gibbs-Duhem relation is not being satisfied for this object, and it is believed that lack of satisfaction of the Gibbs-Duhem relation would trigger the behavior that is seen.
i.e., the following doesn't hold:
G = sum_k (n_k mu_k) where dG/dn_k = mu_k
Of course other reasons for the bug could exist.
