On the computational approach to immobilized pH gradients.

The unified treatment for computing the pH of complex mixtures of mono- and polyprotic buffers, including ampholytes, as utilized in the gradient simulation program PGS, is presented. Its ability to compute pH, buffering power and ionic strength is shown by discussing a few simulations. The problems arising in the automatic formulation of optimal mixtures are presented, as well as the merits and limits of several target functions utilized in such optimizations. It is shown that no universal target function exists and that a proper optimization method should account for the fact that more than one formulation is possible for a given pH range.