Comparative refinement of correct and incorrect structural models of tetrabutylammonium tetrabutylborate – pitfalls arising from poor-quality data

Abstract

This paper demonstrates how numerical parameters usually used to assess the quality of a crystal structure solution (R, wR and S) may be misleading when studying a model refined against poor-quality data. Weakly diffracting crystals of tetrabutylammonium tetrabutylborate, a low-density organic salt comprising isoelectronic cations and anions, were measured using Cu and Mo Kalpha radiation. Along with the correct structural model, six erroneous structural models were constructed and refined against the same data. For both data sets it was found that models based on an incorrect unit-cell choice give lower values of R and wR than the correct one, thus apparently being in better agreement with measured data. Closer inspection of the measured data shows that this is in fact not the case.