Is there a way to tell which one is the racemic or meso compound with 1H-NMR if all you had was the structure and spectra?

Question

I know that diastereomers will have different chemical shifts but can you predict which one is which via up/down field shifts in the spectra?

For instance, is it like alkenes vicinal H’s when cis and trans are up and down field respectively?

I ran a 1H-NMR of hydrobenzoin and when comparing it to the literature I noticed that the racemic mixture was further upfield than the meso.

I made the Newman projections below and assumed the most energetically favorable positions to show the cis and trans H’s in the diastereomers.

Is this the right way of thinking about this?

Answer 1

Hydrobenzoin has 3 stereoisomers: meso, R,R and S,S. You have not included the S,S. Also, your Newman projections might make sense based on steric grounds, but largely ignore electronic factors such as pi-stacking and hydrogen bonding. A very quick run through the Avogadro Bable geometry optimisation has these molecules optimising with the phenyl rings somewhat gauche.

While there are many methods in NMR that can assist in determining relative and absolute stereochemistry, there is no method that offers universal application. Generally it is not possible to identify stereoisomers from first principles in non-rigid skeletons. This is because conformers are rarely locked into a single position. Properties such as spin coupling, and nOe are useful, and the method by Murata is considered groundbreaking in this area. It relies on measurable 3J(HH) and 2,3J(CH). Derivatisation, database comparison, measurement of residual dipolar couplings in aligned media (liquid crystals), and computational methods. Determination of absolute configuration usually involves complexation with a enantiomerically pure reagent.

And I'm not confident that your method for assigning cis/trans geometry based on chemical shifts is reliable.