5. Usage
This is a rough guide to how to proceed:
1) Generate your h, k, l and F and phase values. If you have HREM data available you may want to determine and fix some phases, otherwise all phases should be set to zero (unknown). (Be careful how you do this, for instance paying attention to the CTF, non-linear effects and where you place the origin of the cell.) The h, k, l, F, phase list should be placed in name.hkl in free format.
2) Find your symmetry, or at least what symmetries it might be. For surface data, STM or AFM can help.
3) Make an estimate of the number of atoms and the temperature factors. (This need not be too accurate, but a good estimate is much better than a really bad one.)
4) Prepare your files in the appropriate format and run div2d. Check the output in "Divergence.log". If there are reflections which are specified with a high probability ( > 0.9) by Sigma1 relationships you could include these (i.e. preset them - safest not to). Copy the output file name.Dsort to name.hkl or let div2D.sh do this for you.
5) Run a genetic analysis. Depending upon your computer and the size of the structure this may take from minutes to hours.
6) Look at the solutions, perhaps using xfom first to find unique solutions. This can be done both visually and using peaks2D.
7) If you have something close to a solution, you may want to look at the calc2.log file to see if there are some reflections which (very consistently) have the same value. If this is the case you can include them in your name.hkl file and go back to step 4) and rerun. Often this will give better results.
8) Repeat everything for different symmetries, different numbers of atoms (within reasonable limits) and so forth.
9) Check the solutions using chemical reasoning and (ideally) by a refinement.