Using Recurrent Dependences to Control, Adjust, and Recover Values of the Physicochemical Properties of Organic Compounds

Abstract

A way of controlling, adjusting, and (if needed) recovering values of the physicochemical properties of organic compounds is considered using the example of normal boiling and melting points of perfluoroalkylcarboxylic acids ${{{\text{C}}}_{n}}{{{\text{F}}}_{{2n + {\text{ }}1}}}{\text{C}}{{{\text{O}}}_{2}}{\text{H}}$. Its relevance is due to the values of most properties from different sources of information often failing to correspond to one another, and the discrepancies can be substantial. At the first stage of data verification, the algorithm can include rejecting the greatest abnormalities using polynomial approximation of the dependence of physicochemical properties А on number n of carbon atoms in a molecule. The most reliable results are provided by a subsequent approximation of the data using recurrent relations A(n + Δn) = aA(n) + b. For boiling points, Δn = 1; for melting points, Δn = 2. This means verifying reference data to correspond to those for remaining homologs of the same series. For most homologs, we can select the data that corresponds best to others from several alternative values of different properties. Less often, it is necessary to adjust or exclude data for individual homologs. It should be noted that the considered procedure does not substantially distort the initial arrays of physicochemical characteristics, since the required adjustings do not generally exceed the uncertainties of many modern ways of calculating them.