Critical comparison of evaluation methods of potentiometric titration data using a microcomputer-controlled automatic potentiometric titration system

A microcomputer-controlled automatic potentiometric titrator withP controlled intelligent auto-sampler has been built for serial analysis. Controlling and evaluating programs have also been written. In order to use the most powerful evaluating method, a comparison was made between the methods. The data of the titration curves were obtained by the implicit relation characterizing the curve. The different evaluating methods were tested by using these data. The effects of random errors stemming from measurements were also investigated. The methods used were: two point differentiating formula, differentiation of interpolating and smoothing spline functions, the Gran method, and implicit regression with the Gauss-Newton-Marquardt method. Regression is outstanding among the methods, but it needs much more computing time compared with others. The second-best method is the differentiation with smoothing spline function with end point determination based on finding the sign reversal of the second derivative. This method is much faster than regression.