Jack D. Dunitz

Stereochemistry of reaction paths at carbonyl centres

Abstract Results of recent experimental and theoretical studies of nucleophilic addition to carbonyl groups are described. The reaction paths found by different methods for different nucleophiles show some striking similarities that appear to be characteristic for the reaction type.

Electronic properties of transition-metal oxides-II: Cation distribution amongst octahedral and tetrahedral sites

Abstract The distribution of transition-metal ions amongst octahedral and tetrahedral sites in oxides, particularly of the spinel type, is discussed from the viewpoint of crystal field theory. It is shown that the available experimental information can be rationalized m terms of crystal field stabilizations, the magnitudes of which are determined from optical and magnetic data.

Electronic properties of transition-metal oxides—I: Distortions from cubic symmetry

Abstract The distortions from cubic symmetry which occur in certain transition-metal oxides, particularly those with the spinel structure, are discussed in terms of crystal (ligand) field theory. It is shown that many such distortions, including all the large ones, are related in a simple fashion to the electronic configuration of the metal ion and may be considered to arise as a consequence of a Jahn-Teller type of distortion.

Organic Fluorine: Odd Man Out

Linus Pauling had a superb intuitive understanding of chemistry, backed by deep intelligence and a prodigious memory. He seldom made mistakes. The best known is perhaps his ill-fated three-stranded DNA structure, but one of the few other examples concerns hydrogen bonds involving fluorine. This is evident from a comparison between the various editions of his classic TMThe Nature of the Chemical Bond∫. In the early editions he wrote: TMonly the most electronegative atoms should form hydrogen bonds, and the strength of the bond should increase with increase in the electronegativity of the two bonded atomso It is found empirically that fluorine forms very strong hydrogen bonds, oxygen weaker ones, and nitrogen still weaker ones.∫ [1] Pauling went on to discuss the strong hydrogen bond in hydrofluoric acid (HF)n and the very strong one in the hydrogen fluoride ion [HF2] and correctly concluded that the proton in the latter should lie in a single minimum potential well or in a double minimum potential with a very small barrier. It was only in the third edition, published in 1960, some twenty years after the first, that Pauling conceded: TMIt is interesting that in general fluorine atoms attached to carbon do not have significant power to act as proton acceptors in the formation of hydrogen bonds in the way that would be anticipated from the large difference in electronegativity of fluorine and carbon.∫ [2] Over the years, many chemists have followed Paulings first line of thought, and more or less taken it for granted that organic fluorine acts as a powerful acceptor in the formation of hydrogen bonds. Others have looked at the available structural evidence as collected in the Cambridge Crystallographic Structural Database (CSD) and concluded that organic fluorine is at best only a weak hydrogen-bond acceptor. 4] A further intensive search of the CSD, including detailed inspection of individual crystal structures and backed by ab initio calculations on model systems, confirmed that organic fluorine hardly ever accepts hydrogen bonds, that is, it does so only in the absence of a better acceptor. Even in such compounds like crystalline ammonium trifluoroacetate, in which there are four hydrogen donors per anion, there is no hint of N H¥¥¥F hydrogen bonding; the four N H bonds all point towards oxygen atoms of the trifluoroacetate anions. In the ammonium monofluoro structure there is just a hint of a bifurcated hydrogen bond involving carboxylate O and the syn-planar F atom, but the latter is 0.26 ä more distant from the H atom (Figure 1). Likewise, the evidence for hydrogen bonding to organic fluorine in protein±ligand complexes was examined and found to be unconvincing. Hydrogen bonding involving B F bonds should be stronger than that involving C F bonds. Yet even in crystalline ammonium tetrafluoroborate NH4 BF4 , surely the exemplar of such an expected interaction, no short N H¥¥¥F lengths are observed. The authors commented TMit is believed that hydrogen bonding contributes negligibly to the lattice energy of this crystal∫. A search of the CSD reveals only very few crystal structures showing short intermolecular X H¥¥¥F B lengths; one example is 4,4,8,8-tetrafluoropyrazabole, another is 2,2-difluoro-4,6-dimethyl-3-phenyl1,3,2-difluorodiazaborine (Figure 2). One can hardly deny that these should be classified as genuine hydrogen bonds, but there are few such specimens. It seems clear that with its low polarizability and tightly contracted lone pairs, fluorine is unable to compete with stronger hydrogen-bond acceptors such as oxygen or nitrogen. The few authentic examples of O H¥¥¥F or N H¥¥¥F hydrogen bonding involve systems in which approach of the hydrogen atom to other better acceptor atoms is sterically hindered. Indeed, nowadays the occurrence of a genuine hydrogen bond involving organic fluorine seems to be regarded as sufficiently noteworthy that it deserves special mention in the title of the publication, as, for example, ref. [10] .

The non-planar amide group☆

Crystal structures of medium-ring lactams and their crystalline hydrochlorides furnish detailed information on the nature of the out-of-plane distortions of the amide group. Caprylolactam has a non-planar transoid amide group in the crystal but exists in solution as an equilibrium mixture of at least two conformations, one with a nearly planar cis-amide group, one with a non-planar transoid-amide group. The corresponding hydrochloride salt is protonated on oxygen and contains a nearly planar cis-amide group. For non-planar amide groups it is found that out-of-plane bending at nitrogen is about as important as pure twisting, the contribution from out-of-plane bending at the carbonyl carbon being smaller. These results, together with rough data on the energy differences between the lactam conformations, can be used to test various potential functions that have been proposed. It is concluded that although the equilibrium conformation of the amide group may be planar or close to it, out-of-plane deformation can be made at a very modest energy cost. In the construction of models of polypeptide chains or in fitting such models to electron-density maps of protein crystals, the strictly planar peptide unit thus incorporates restrictions that may be rather easily relaxed in the actual molecule.

Crystal structure prediction of small organic molecules: a second blind test.

The first collaborative workshop on crystal structure prediction (CSP1999) has been followed by a second workshop (CSP2001) held at the Cambridge Crystallographic Data Centre. The 17 participants were given only the chemical diagram for three organic molecules and were invited to test their prediction programs within a range of named common space groups. Several different computer programs were used, using the methodology wherein a molecular model is used to construct theoretical crystal structures in given space groups, and prediction is usually based on the minimum calculated lattice energy. A maximum of three predictions were allowed per molecule. The results showed two correct predictions for the first molecule, four for the second molecule and none for the third molecule (which had torsional flexibility). The correct structure was often present in the sorted low-energy lists from the participants but at a ranking position greater than three. The use of non-indexed powder diffraction data was investigated in a secondary test, after completion of the ab initio submissions. Although no one method can be said to be completely reliable, this workshop gives an objective measure of the success and failure of current methodologies.

A New Interpretation of the Disordered Crystal Structure of Ferrocene

X-ray analyses of ferrocene [Fe(CsHs) 2, CmH~0Fe] at room temperature and at 173 K yield vibrationalellipsoid patterns that are incompatible with pure rotational disorder of the cyclopentadienyl rings. Below 164 K ferrocene becomes triclinic. Relationships between the diffraction patterns of the monoclinic hightemperature (HT) and triclinic low-temperature (LT) structures suggest that the formally centrosymmetric molecule of the disordered HT phase can be described in good approximation as an averaged superposition of the four molecules present in the primitive LT cell. The molecular centre of symmetry required by the HT space group is thus only statistical in nature, and the crystallographic evidence for the staggered arrangement of the cyclopentadienyl rings in ferrocene and other isomorphous metallocenes has to be revised.

Phase transitions in molecular crystals from a chemical viewpoint

Why do molecules pack in crystals as they do? The trivial answer: because they thereby achieve periodic arrangements corresponding to minimum potential energy. In general, for any given molecule, there must be many such arrangements of closely similar energy; experience accumulates to show that polymorphism is of widespread Occurrence in the world of molecular crystals. It is often accompanied by radical changes in conformation and hydrogen bonding patterns. Indeed, as far as molecular crystals are concerned, there seems to be no clear cut boundary between polymorphic transitions and solid state chemical reactions. Of particular interest are those cases where crystal orientation is preserved during the structural change, for there, in principle, the atomic positions can be mapped quite precisely on both sides of the transition and mechanisms can be proposed for the molecular motions that are involved. Not all such processes Occur under topochemical control, i.e., with a minimum amount of atomic or molecular movement from the known starting structure. Many seem to occur, or at least to be initiated, at crystal defects where the regular arrangement of molecules is interrupted. Examples of both types will be discussed, and the attempt will be made to describe these and related phenomena from a chemical point of view.

Temperature dependence of thermal motion in crystalline naphthalene

Single-crystal data for naphthalene have been measured at five temperatures between 90 and 240 K. Positional and thermal parameters for C and H atoms at each temperature were refined by conventional least-squares techniques. The effect of varying the weighting scheme was examined. Contributions of internal molecular modes to the motions of the atoms turn out to be important. They were estimated for the C atoms at each temperature from a standard force field and subtracted from the experimental U~/ values. The corrected UiTs were then analysed to determine rigid-body translational and librational tensors for the naphthalene molecule. The absolute magnitudes and temperature dependence of these quantities have been compared with values calculated from atom-atom potentials and from spectroscopic data.

Crystal structure analyses of 1,4,7,10,13,16-hexaoxacyclooctadecane and its complexes with alkali thiocyanates

The results of crystal structure analyses of 1,4,7,10,13,16-hexaoxacyclooctadecane (CH2CH20)6 and its complexes with NaNCS, KNCS, RbNCS, CsNCS [and Ca(NCS)2] are discussed. In the K*, Rb*, Cs* and Ca 2+ complexes the unsubstituted hexaether adopts a conformation with virtual Ds~ symmetry although the larger Rb + and Cs + ions are displaced by more than 1 ,~, from the mean plane of the ligand. In the Na* complex, the ring is strongly distorted from its symmetrical conformation to accommodate the smaller cation. The uncomplexed hexaether has a centrosymmetric conformation containing three different types of monomeric subunit. The shortening of the C-C bonds found in these and related complexes is discussed and judged to be mainly an artificial effect arising from inadequate treatment of curvilinear vibrations.