SIG #4

Special Interest Group
on Electron Crystallography
of the European Crystallographic Association

Promotors:

-Prof. Sven Hovmoller, Stockholm University
-Prof. Jon Gjonnes, Oslo University

Proponents:
four ECA councillors:
-Prof. Karimat El-Sayed, Egypt
-Prof. Giovanni Ferraris, Italy
-Prof. Berit Fjaertoft, Norway
-Prof. Rolf Norrestam, Sweden

1. Introduction

The aim of the SIG is to raise the awareness, acceptance and general standard of Electron crystallography, i.e. to standards comparable to those of X-ray diffraction. This will need some concerted effort of groups and workers in this field – and external support. Development of electron crystallography is usually carried out in electron microscopy labs, with relatively small research groups as one of several materials research activities. The contrast between this kind of ‘in-house labs’ and the organization associated with large facilities, e.g in the synchrotron field, has been pointed out by A. Howie (in ‘Current Opinion in Solid State & Materials Science’ vol 4, no 3). It is our view as SIG organizers that we should find ways to offset this drawback in order that the full potential of electron optical methods in structure research can be realized.

2. Electron crystallography, the situation:

Recent crystallographic work in several groups have demonstrated that electron optical methods are viable options for structure solution over a wide field, including various kinds of inorganic structures, small organic molecules, biological macromolecules, organic polymers. Successful refinement has been obtained from intensity data collected by selected area electron diffraction (SAED) or micro-diffraction. Refinement of structure factor amplitude and phases to very high accuracy has been attained by the convergent-beam (CBED) technique for crystals with small unit cells. The main advantage of electron crystallography remains the facility to study minute crystals, and to relate these to their surroundings, e.g. in multiphase materials. Other benefits include sensitivity to ionic states, bond charges and charge fluctuations; combination with other electron microscopy techniques, such as spectroscopy with high spatial resolution. The problems and tasks one is facing in development of the field have two main aspects. One needs to:
i) establish reliable, well-documented procedures, that extend from data collection, through workable procedures for determination and verification to satisfactory refinement of crystal structure. The difficulties, (especially related to dynamical scattering effects) must be dealt with – and the specific advantages exploited.
ii) focus on classes of structure problems that depend upon electron crystallographic methods for satisfactory solutions. Flexibility and variations in the method is expected to remain a characteristic feature of electron crystallography. Combination with other crystallographic technique, such as X-ray powder diffraction, will increase. Other aspects besides precision in position parameters may be brought into focus, e.g. ionicity of the atom species, disorder and defects.

3. Development of the method

i) Structure solution: the phase problem Solving a crystal structure is equivalent to assigning structure factor phases to a sufficient number of reflections. A selection of methods for extracting structure factor phases from electron diffraction intensities and/or high resolution images have been established – or proposed: . Direct methods applied to electron diffraction intensities, assuming kinematical scattering, have been shown to work in a variety of cases, notably for organic crystals. Recent experience indicates that direct methods can be applied also to inorganic crystal, including three-dimensional electron diffraction data, even from quite thick crystals. Crystallographic structure factor phases can be extracted directly from the Fourier transform of HREM images from single images, or by more extensive procedures involving through-focus series. The increased information limit and high source coherence of FEG microscopes will increase the power of these methods. Multiple-beam dynamical diffraction effects containing phase information are readily observed in CBED, but have so far not been much used in actual structure determination. Such phase information may be very useful in conjunction with other methods. Various proposals for ‘inversion of dynamical scattering’ in order to obtain structure factors (amplitudes and phases) have been published, but have not been tested as a means for solving unknown structures. The availability of coherent CBED and holography in modern TEM may increase the interest in this problem. . The inclusion of approximate corrections for multiple beam dynamical diffraction at various stages should be addressed, including methods for independent measurement of thickness. Development in this field could benefit vastly from exchange of views, experience and structure problems.
ii) Intensity measurement Intensity data are crucial. There is a lot of work to do in establishing good procedures for intensity measurement. New recording media: imaging plates and slow-scan CCDs (with associated soft-ware) should encourage this work. It is also important to develop quantification procedures for the many labs that cannot afford to buy this expensive equipment. Define experimental conditions: electron diffraction theory is usually formulated for a parallel beam, whereas crystallographic procedures assume integrated intensities. The way integration is performed in reciprocal space should be analyzed in relation to experimental situations, specimen characteristics, recording procedures. Background subtraction procedures, correction factors, influence of specimen quality etcetera need to be considered. Spot patterns with stationary beam, precession patterns, CBED-technique must be treated. Theoretical expressions corresponding to real experimental situations should be developped and tested. Special attention to merging of data to three-dimensional sets.
iii) Refinement The situation is characterized by vast variations in the quality of intensity data, from CBED-profiles obtained from perfect crystal areas under well-defined condition, to spot intensities from bent crystals of poor quality. The quality of data within one structure determination may also contain large variations. Parallel with improvement of the data it is important to work on the refinement procedures. Refinement in Fourier space is one option. Another possibility is to introduce chemical constraints. A major challenge appears to be to combine limited intensity data of high quality, e.g. from CBED-measurement, with extensive data sets from spot patterns. First step may be a discussion of refinement strategies – and their relation to ways of intensity measurement.
iv) Application of coherent beams and holographic techniques

4. When do we need electron crystallography?

The motivation for these efforts must be connected with needs in structure and materials research:
a) Small crystals encountered in chemistry and materials research. The options are: grow larger crystals, use powder diffraction – or use electrons. Electron microscopy is often necessary anyway, in order to characterize the material, ascertain Bravias lattice and unit cell dimension – or because small crystal size, internal surfaces, defects or precipitates are inherent aspects of the material.
b) Combination with powder diffraction. In addition to supplying the correct Bravais lattice, electron diffraction can provide important additional information, such as modulations, anisotropic temperature factors, ionicities and bond charges.
c) Nanostructured materials. With materials that basically and deliberately is made up of very small crystals, with properties different from bulk material, it should be obvious that electron crystallographic techniques must have a major r=F4le. It is important to communicate this message in materials communities.
d) Materials characterized with CTEM/HREM as a main tool. At present the possibility to extract quantitative crystallographic information in addition to other TEM technique is often overlooked.
e) “Ill-defined” structures, i.e. materials with poor or varying crystallinity. With electron diffraction and high resolution imaging one can select the most ordered parts for structure studies.
f) Biological macromolecules. Membrane proteins are important examples. Comments and views on these and other aspect on electron crystallography are welcome!