Chemical Crystallography Group Autumn Meeting 1997

The 1997 CCG Autumn Meeting was held in the School of Chemistry at the University of Bristol and was organised and hosted there by Professor Guy Orpen on Wednesday 12th November.

An audience of over 60 assembled to hear a series of lectures on the subject of "Disorder, Twinning and Incommensurate Structures" and following a welcome by Guy Orpen, these commenced with one given by Mike Hursthouse (Cardiff) entitled "The origin of disorder in crystal structures and its treatment in refinement". He provided a definition of 'real disorder': "Discontinuities in the lattice repetition of a crystal structure, resulting in partial or mixed occupancy of lattice sites in the crystallographic unit cell." The distinction was drawn between disorder in molecular and in framework structures. For the former the following types were identified:
(a) alternate displacement disorder, which may be accompanied by marked "thermal motion" and require restraints or constraints to molecular geometry and to occupancies;
(b) alternative orientational disorder often leads to severe correlation and unstable refinement, necessitating restraints or constraints;
(c) continuous orientational disorder is best tackled by acquiring a new set of data at a lower temperature.
In framework structures disorder can affect either the guest atoms or ions in cavities within the framework or the framework itself. In the latter case substitution (e.g. Al for Si in aluminosilicates) or partial occupancy can occur. A caution was sounded regarding "virtual disorder", where the "disorder" arises from an incorrect definition of the lattice or the choice of the wrong space group.
Polysynthetic twinning in Labradorite OUM17161

Photograph published by permission of the Oxford University Museum of Natural History.

Thanks also to David Watkin for this beautiful example of a twin used to illustrate his talk.

David Watkin (Oxford) had chosen the reassuring title "Twinning - don't give up - yet" for his examination of this phenomenon. He described how twinning involves a specific relationship between singletons and depending on the exact relationship can be divided into various types: in merohedral twins the lattices of the components exactly coincide, while each component of a non-merohedral twin generates its own diffraction pattern. Twinning can be detected by a variety of symptoms including geometric relationships between supposedly independent unit cell parameters; the assignment of space groups such as Pmmm which exhibit a suspicious lack of translational symmetry; strange patterns of systematic absences; poor convergence and high difference Fourier residuals with apparently good data. Historically, the move from photography to diffractometry led to a decrease in the number of twinned structures reported: the spread of area detectors is beginning to restore this balance.

After an excellent lunch, the sessions resumed with Howard Flack in the chair and our third speaker, Mike Glazer (Oxford), spoke on "Incommensurate diffraction as an end member of disorder", beginning with some illustrations of the phenomenon. Using examples such as the structures of nickel squarate and barium titanate, he showed how two different structural arrangements A and B could exist within the same crystal, and how a version of the structure could contain mistakes, for example arrangement A existing within a structural motif of B. Experimentally, incommensurate samples can give rise to additional, diffuse intensity maxima superimposed on the Bragg scattering, and Mike proceeded to demonstrate these effects by using optical transforms. Incommensurate structures represent an intermediate situation between short-range disorder involving atoms or molecules and the long-range phenomenon of twinning. 20 Acenaphthalene is ordered at 130 K, shows molecular overlap at 149 K and adopts an incommensurate structure in the intervening range.

Chick Wilson (ISIS) outlined some "New approaches to the study and modelling of disorder", including variable temperature techniques, supercell methods and diffuse scattering. Time-of-flight single crystal diffraction retains three- dimensional information because of the "time-binning" it uses. At 5 K, even lead shows some diffuse reflection. In the study of static disorder it is possible to map trends as a function of sample conditions and effectively scan thermodynamic variables. In a study of the temperature-dependent behaviour of disordered hydrogen atoms in the benzoic acid dimer, these were shown to be ordered at 20 K but already extensively disordered by 50 K. In another study, hydrogen atoms of torsionally disordered methyl groups show significant libration even at 100 K. Full three-dimensional datasets were acquired at eight temperatures between 20 K and 330 K, and subsequent analysis demonstrated that even at 20 K there was still some librational motion. From such studies it is possible to quantify the zero point motion.
Chick described the use of a model-free description of disorder in order to improve the description of the thermal motions. Although the method is computationally intense, it allows the study of systems such as plastic crystals where standard methods are not applicable. Information on local structure or transient effects can be obtained from the analysis of diffuse scattering.

We then moved on to the short contributions, which were chaired by Marjorie Harding. The first of was by Howard Flack (Geneva) on "Chirality, polarity, lattice symmetry and standard uncertainties". Howard noted that removing atoms from the lattice can increase the symmetry. An flipped atomic arrangement may not correspond with the original, and new data (collected on a new crystal) may not correspond to the previously established coordinates. The symmetry of multiple cells and subcells can play a role in twinning. In the absence of actual twinning, the inversion twinning parameter (the Flack parameter - definitely a relation!) may be used to determine the absolute structure. With chiral materials this amounts to establishing the enantiomorph. When the Flack parameter is 0.5, a conglomerate could be present.

In a talk entitled "Novel approaches to difficult data sets", a series of views of the nearby Clifton Suspension Bridge allowed Friedemann Hahn (Stoe Darmstadt) to demonstrate the importance of the viewpoint adopted when processing information, especially when dealing with X-ray effects which cannot be handled using a conventional orientation matrix. 2D Indexing and optimisation are important points, but automatic programs often fail with difficult cases. His conclusion was that graphics are very helpful, becoming more so with increasing numbers of data. A program called RECIPE shows peaks and separates out the lattices of intergrown crystals: it is possible to derive two orientation matrices and process a dataset for each component: these sets can be used separately or in combination. Several examples of the application of this technique were given.

Simon Parsons (Edinburgh) made the tempting offer "Multiple crystals - buy one, get one free", and then outlined how to get the best from it. The determination of the twin law is critical for the modelling of merohedral twinning, but non-merohedral twins are now being reported more commonly since the advent of area detectors and better indexing routines. Simon uses the DIRAX program to suggest the different orientation matrices for the components in twinned crystal, and from the relationship between these matrices it is possible to derive the twin law. He describes how supercells can arise, for example in an orthorhombic cell where a � 1/2 b. He included some impressive examples of rather complex but clear derivation of twin laws.

In the final contribution, "Twinned or disordered?" was the question posed by Aggi Steiner (Liverpool), who gave the example of the product of the reaction between Ti(OⁱPr)₄ and acetone which after two days produces colourless needles. The product is
[Ti₃O( m -OⁱPr)₃(OⁱPr)₄{Me₂C(O)CH = C(O)CH₂C(O)Me₂} ], in which three acetone molecules have trimerised via an aldol-condensation to give the tridentate ligand Me₂C(O)CH = C(O)CH₂C(O)Me₂. The resulting trititanate complex shows no symmetry at all, but the X-ray data imply the space group R-3m, in which the tridentate ligand is disordered about the three-fold axis as well as over the three associated mirror planes. The space group symmetry can be reduced either to R-3 (with the loss of the three mirror planes) or to C2/m (with the loss of the three-fold axis). In both cases the corresponding twin law has to be introduced. In case of C2/m this requires rather awkward and time-consuming matrix transformations both for cell constants and twin law (and worse: in addition all three mirror planes have to be checked). This can be avoided by adopting the space group R-1m, which is the rhombohedral setting of C2/m. Acceptable models could be refined both in R-3 and R-1m (C2/m), but these are still disordered in one way or the other and represent only a compromise between twinning and disorder. Aggi is continuing the work using a model involving a non-disordered, six-component twin in space group
P-1 (or R-1 as it appears in the rhombohedral world).

The Meeting concluded with a summary by Marjorie Harding which drew together the various topics discussed during the Meeting and stressing their growing significance in modern chemical crystallography. Thanks were expressed to Guy Orpen and his team for the smooth organisation and running of the Meeting.

A.J. Blake

University of Nottingham

Editor's Note: David Watkin's talk has been stored on the Internet at URL
http://darkstar.xtl.ox.ac.uk/refs_bris97.html

Click here to return to BCA homepage