dd Crystallography dd on this page: Rietveld Method Composite Structures Preferred Software		d

My interest in Crystallography

My main interest is applied crystallography and crystal chemistry. I.e. to take x-ray or neutron diffraction patterns, refine them employing a certain structural model and finally understand the properties by regarding the structure.
There are two major collections of sites with crystallographic interest on the web. You can use the following general entry points:

• The crystallographers guide to the internet - crystallography world wide
• The International Union for Crystallography - IUCr

For software you might directly choose:

• The The SINCRIS software database
• The Computing project 14 (Crystallography)

The Rietveld Method top of page

hard

Start with Hugo Rietveld's original summarizing paper in Journal of Applied Crystallography 2, 65-71 (1969).
Continue acheiving all following literature concerning e.g. magnetic refinements, preferred orientation, background modelling ...

easier

Buy the book "The Rietveld Method" edited by R.A. Young.
Allthough it may be hard to find and you might find it hard to pay for it (it's not quite cheap), this is better than searching several catalogues in libraries and copying all important stuff yourself.

easy

Try to find someone in your local crystallography, geology or mineralogy department who practices the method (and has some time left for you). A personal contact will help you through most problems. Avoid people who do not understand the maths behind.

Once you are quite familiar with the software and the theory, you might still have problems. A list of frequently asked questions is available on the Rietveld Mailinglist Homepage. You will find numerous hints in the FAQ-list and almost every important link on the web concerning the Rietveld method there.

I take the opportunity to express gratitude to my "personal contact", J–rg Ihringer at the Institute of Crystallogaphy, University of T¸bingen. He was allways very patient when I came with my "chemical" understanding of maths and statistics.

Composite Structures top of page

These are usually structures that are no longer to be described by only three-dimensional symmetries. 4D group theory is quite well developed, but 5D and 6D are not yet. Two routes are acccessible in order to describe such a structure with 3D-methods:

1. An experimental trick might be possible (we are working on this one): Using anomalous scattering should allow for the ‚at least numerically possible‚ extinction of one of the substructures. Usually, anomalous scattering occurs when the wavelength used in the experiment is close to the absorption edge of an element. Then, the scattering factor is affected due to absorption of the x-ray radiation. Recording diffraction patterns at different wavelengths in the neighbourhood of the absortion edge should allow for the numeric "subtraction" of the scattered intensity of the element in quest. Hence in the case of composite materials you can make the contribution of one substructure "vanish".
2. A numerical method is to refine two modulated 3D phases during refinement. Each of the substructures is attributed a 3D symmetry and unit cell, but as both substructures "see" each other, they both possess a modulation vector that reflects the other substructure. A software being able to calculate such composite structures including all constraints that are imposed due to the "composite" properties is JANA by XXX. It has long time been a software for single crystal structure refinement, but powder calculations should be possible, too.

My Favourite Software top of page

SIMREF, the software programmed by Maichle, Ihringer, Prandl & Ritter did not really simplify my first contact with the Rietveld method, but I do not know any software that did by that time! Two special features of SIMREF are noteworthy:

• The possibility to use diffraction patterns from different instruments (e.g. x-ray tube, synchrotron and neutron sources) and to include all this information into one single structure refinement.
• The ability to decend symmetry without the need to change space groups (by applying characteristic, symmetry-given displacement modes to the atomic positions; a way to understand phase transitions in the terms of Landau-theory).

For one-source spectra I prefer FullProf of Carvajal for the moment. It does a perfect job and it is a bit easier to handle. Furthermore, it allows for some more variance in fitting the profile (e.g. asymmetry funcions, background fitting, constraints etc.) than SIMREF.