Space Group from Absences
The determination as it is actually done: you have a data set, some classes of reflection are systematically missing, and the question is which of the 230 space groups you are allowed to be in. Answer the classes you have determined and the page names every group that fits — and the measurement that would shorten the list.
- You supply
- What you observe, class by class. Each menu offers only conditions some space group actually produces, so an answer can always be satisfied by something. Anything you have not determined is left alone and constrains nothing.
- Reading it
- A class left at not determined constrains nothing; answering it no condition asserts you looked and found nothing missing, which is what actually narrows the list. The strong answer is the one that can be wrong — and a wrong one removes the right group without saying so.
Worked examples: P21/c – the commonest determination there is · P21/c – the same data, empty classes too · Pbca – one pattern, two systems · P4cc – four groups, one reflection apart · Fd3m – diamond
Input
Answer only the classes you have actually determined. A class left at not determined constrains nothing, and that is the honest state for a zone you have not measured out far enough to be sure about — every answer you can give narrows the list, and a wrong one removes the right group without saying so.
11 space groups
These conditions are produced by 11 of the 230 space groups.
| No. | Symbol | System |
|---|---|---|
| 14 | P21/c | monoclinic |
| 30 | Pnc2 | orthorhombic |
| 38 | Amm2 | orthorhombic |
| 39 | Aem2 | orthorhombic |
| 56 | Pccn | orthorhombic |
| 57 | Pbcm | orthorhombic |
| 60 | Pbcn | orthorhombic |
| 61 | Pbca | orthorhombic |
| 130 | P4/ncc | tetragonal |
| 138 | P42/ncm | tetragonal |
| 205 | Pa3 | cubic |
Systematic absences see the lattice centring, the glide planes and the screw axes and nothing else. They cannot see whether the structure has a centre of inversion, so a centrosymmetric group and its non-centrosymmetric subgroups extinguish exactly the same reflections. Separating those is a job for intensity statistics, for the structure solving, or for chemistry — never for more careful indexing.
Measure this next
The 0kl class separates them best. Whichever of these 5 outcomes your data show, the list shortens to at most 4.
| If 0kl shows | Groups left |
|---|---|
| no condition | 1 |
| k + l = 2n | 2 |
| l = 2n; k = 2n | 1 |
| l = 2n | 3 |
| k = 2n | 4 |