# Could Gaussian or Spartan software determine stereoisomers from conformational landscape of a molecule alone?

I read that, on some papers, conformers were generated by stereoconfiguration of the molecules and from there, they evaluated stability of each stereoisomer in terms of free energy differences between two stereoisomers. Why people illustrated the energy profile of each stereoisomers from their fixed stereoconfiguration, not from global landscape of conformers (conformational distribution) of the molecule and sorting out the conformers by giving plane of symmetry or other methods defining and distinguishing stereoisomers of the molecule? A molecule is always allowed to rotate freely on 3D space and the results of its ability to rotate into a mirror image of its origin is the chirality we are interested in.

Is it possible to tell Gaussian or Spartan to assign stereoconfiguration of each stereoisomers by solely from its generated conformers (conformational distribution) alone without any preconfiguration of its stereoconfiguration? Or I missed a theory that I can’t follow up the general methods on deducting 3D structure of a stereoisomer?

• I really don't understand what you are asking for. A molecule is allowed to rotate in a 3D space, but it will never rotate into a mirror image—that's just not possible. Conversion between stereoisomers requires an inversion, which cannot happen via any bond rotation that you might conceive. Besides, sampling the full conformational space is very expensive, particularly with QM methods. Jun 14 at 8:22
• @ShoubhikRMaiti This figure said it’s possible to do so nature.com/articles/s41557-018-0043-6/figures/1 Jun 14 at 9:02
• Yes, but as you can see from the images, there are some which are easily interconvertible and some which aren't. Generating a global landscape of conformers is very costly and time consuming. It is far easier to generate possible conformers and then generate stereoisomers from them. (AFAIK Gaussian does not have any option to deal with stereoisomers, at least not automatically. I have never used Spartan so can't comment on that.) Jun 14 at 18:50

Because Jmol may assign stereogenic centers, it is likely other programs (perhaps including Gaussian) are equally capable to perform this task. For example, on a hypothetical bromofluoromethanol:

you select the carbon atom in question, and type on Jmol's console the commands

set labelalignment right;
set labeloffset -20 0;
label %[chirality];


to obtain the requested CIP-label (further examples and details here).

For the figure in question:

(credit to Canfield et al.)

the e.g., line about the interconversion of atropisomers has to be read altogether with the energy scale / time scale provided below the same figure, too. For 1,1'-biphenyls with little or none steric hindrance at 2,2',6,6'-position, it is indeed plausible that enough heating might suffice to interconvert the $$\Lambda$$ into the $$\Delta$$ atropisomer (or, P into M)

Assuming typical reaction conditions, steric hindrance for larger molecules like e.g., BINAP, or BINOL establishes an energetic barrier this high that you observe them in either the P, or M configuration, even if there is some conformational flexibility left. The barrier is this high that these molecules are reliably used as chiral ligands in chiral syntheses.

(credit)

If you have access to a molecule kit, I recommend you build these molecules by your own. Experience the steric hindrance the hydrogen atoms not explicitly drawn impose if one attempts to pass for the other enantiomer; this requires to break the $$\ce{C-C}$$ bond.

Reference: Canfield, Peter J. et al., A new fundamental type of conformational isomerism. Nat. Chem. 10 (2018), 615–624; doi: 10.1038/s41557-018-0043-6; the author's preprint.