The emphasis of the research that has been reviewed in this article has been to highlight the subtle effects that undetected crystallographic disorder may play in the determination of molecular structures by single crystal X-ray diffraction. Such effects of disorder can range from the incorrect determination of bond lengths, to the incorrect formulation of compounds. These examples serve to emphasize that the observation of low R values, low esd's, and well-behaved thermal parameters are not always sufficient indications of a true structure. Many of the topics described in this article have centered on disorder due to the presence of impurities in single crystals. It is very likely that one of the principal reasons for neglecting to consider compositional disorder when performing X-ray structure determinations is undoubtably due to the common belief that single crystals are inherently pure, even though it is recognized that such a method of doping paramagnetic complexes into crystals of isostructural diamagnetic complexes is used extensively in single crystal EPR studies.131 It is this general belief that single crystals are pure, compounded by the subtlety of the disorder problem, that results in many of the misinterpretations described in this review. The interpretation of X-ray diffraction data in providing apparent experimental verification for bondstretch isomerism in m-mer-MoOCl2(PMe2Ph)3 represents an extreme example of the subtlety of the disorder problem. Moreover, it is also a reflection of the influential power and the exactness that is often believed to be inherent in a X-ray structure determination. Indeed, as cited in the introduction: “There can be no doubt that when a compound has to be identified there is nothing to beat a crystal structure determination. After all, the final outcome—a drawing of the molecule, with all the atoms nicely shaded and labeled—is convincing and will silence all but the most stubborn critics.” While this statement may be correct for the majority of structures, it is important to recognize the limitations of the technique so that results are not overinterpreted.48 In summary, the following points should be highlighted: (i) Crystallographic disorder may be difficult to detect, and unsuspected crystallographic disorder can result in both (a) the incorrect determination of bond lengths and (b) the incorrect formulation of compounds. (ii) Cocrystallization of structurally related molecules, resulting in the formation of single-crystal solid solutions, may be more common than presently appreciated. Hence, the possibility of compositional disorder should be considered when performing X-ray structure determinations, especially if there are any unusual measured bond lengths. In this regard, spectroscopic analysis of the same crystal used in the diffraction study may prove informative, as would repeating the structure determination on a crystal obtained from a different batch. (iii) The observation of a long Mo=O bond length in the green “isomer” of cis-mer-MoOCl2(PMe2Ph)3 is an artifact due to compositional disorder with isostructural mer-MoCl3(PMe2Ph)3. There is no evidence for bondstretch isomerism in the first system that was proposed to exhibit the phenomenon. (iv) The 3 cr criterion for establishing whether or not two bond lengths are significantly different is only valid if there is no systematic error in the data. Systematic errors may be difficult to detect, but, as in ii above, repeating the structure determination on a crystal obtained from a different batch may provide a more accurate estimate of the error, especially if compositional disorder is the source of the systematic error. Finally, although the original example of bond-stretch isomerism has been shown to be the result of an artifact and that there is presently no definitive evidence of bond-stretch isomerism in other complexes, the concept itself still remains. Only a combination of time and a demanding series of analytical experiments will reveal whether or not it is actually possible for bond-stretch isomers to exist under normal conditions. © 1993, American Chemical Society. All rights reserved.
