Mr. Sharma,

The space group and system are determined experimentally by single crystal XRD.

Towards “prediction”, you can use the methods of quantum chemistry by Gaussian (solid-state computations, www.gaussian.com, Guide for solid-state computations by Gaussian: http://scuseria.rice.edu/gau/PBC-Guide.pdf), for example.

You should do, however, a significant set computations (it is really significant set!), taking into account for different: (i) number of molecules per unit cell; (ii) included crystallohydrate (crystallosolvate); (iii) a variation of typical unit cell dimensions to given substance according to a widely distributed space systems; (iv) different models of space groups to any spaces system; and (v) more. To each of this models you should compute the energy of the ensemble of interacting species. Thus, having, a statistically representative set of energies you could deduce, which space system/group is highly probable to be observed experimentally on the base on it energetics.

As I have pointed out, however, this is associated with a significant set models of systems to each molecule, with a significant statistical representativeness.

For example, you would like to model possible space groups/systems to pyridine, then:

(A) You should take a models of a triclinic unit cell with number of molecules Z = 1, 2, 4, 6, 10 and 12 (these are typical values);

(B) To each of these 6 models in point A, you should study both noncenrosymmetric space group P1 and a centrosymmetric one (P-1). Therefore you have 12 models of unit cells.

(C) Towards periodic boundary condition computations, you should involved unit cell dimensions selected set (N) from 5 - 35 A (those are typical values for small organics). Therefore in point C you have 12.N number of models.

(D) You should involved in each model in point A examples with included M = 1,2,4..10 molecules crystallohydrates (K solvates different type). So you have to compute 6.K models in pint A; 12K models according point B and 12N.K models according point C.

The models shown in A-D should be repeated to each 230 space groups to each space systems. Furthermore pyridine has a fixed geometry.

When your molecule is flexible, then:

(E) You should take into account for a set possible conformations of the molecular skeleton, because of the molecular conformation in the crystalline state very frequently is different from the most preferred geometry corresponding to the energy minimum at the equilibrium state.

Thank You Prof. Bojidarka B. Ivanova for long discussion !

I want to ask here, does every molecule has HOMO-LUMO gap ? When it will crystallize in 3D periodic system then it's band gap will be more or less compare to HOMO-LUMO gap.

Shilendra

Mr. Sharma,

Yes, they have!