Regarding possible structure predictions of BiInO3 apart from the ones mentioned in Materials Project

I have been trying to grow BiInO3 (one of the new polar materials predicted by Materials Project). For the past semester, I have been trying to grow BiInO3 on a couple of substrates. There are two equally likely structures predicted by MP Pnma and Pna21. Out of these Pna21 has been made in bulk in literature. Other than that there is no experimental information about any such oxide’s structure. Thereby, we are heavily relying on the prediction by MP. Recently we got some results which show 2 possibilities, either multidomain Pnma unit cells (which is very unlikely) or any other structure not mentioned in the MP structure predictor (maybe hexagonal). We have already checked for Pnma (mp-561102), Pna21 (mp-556892), In2Bi2O7 (mp-768468) which have the least Energy above Hull values. We just want to confirm in case any other structure is possible apart from the ones mentioned above.

Hi @Megha_Acharya,

The simple answer is that our database is by no means exhaustive, so there is always the potential that some other structure in this system does in fact exist and might outcompete the polymorph you’re trying to make. It looks like all the BiInO3 structures in MP have ICSD IDs. I would follow those back to the original publication and check that first.

I’m glad you’ve looked at structure prediction. While the structure prediction is very useful, it also has its limitations, mostly due to the fact that the substitution probabilities are trained for a wide range of structures and chemistries rather than specific systems. For instance, it’s not good at identifying all possible distortions of perovskite for the given chemistry. Thus, there might be BiInO3 structures in slightly different space groups that are close to or even more stable.

Hi Shyam,

Thank you for your reply. So, the ICSD IDs you mentioned are the ones that I refer to as they have been experimentally verified. Both of them (Pnma and Pna21) are related to the same paper by Belik et al where they have proven using Rietveld Refinement that the final structure of BiInO3 in bulk is actually Pna21 and not Pnma. This is the only piece of literature available about this material. The problem is my experimental data matches best with multidomain Pnma which is highly unlikely though not impossible. Also, our results suggest this material is non-polar even below room temperature. So, I just wanted to confirm that in case there are other structures possible which would fit our hypothesis about domains better.

Is it possible using the structure predictor tool or any other tool in MP that we can get a better idea of all the possible distortions of this specific perovskite?

Hi Megha_Acharya,

If you think the structure could be hexagonal, one possible candidate would be the YInO3 (mp-504607) structure with 5-coordinate In, although it would be a little surprising to see that for a cation as large as Bi3+ .

Hi Steven_Hartman,

Thanks for your suggestion. I will try to check out the peak positions for YInO3 for my film in case I can find anything. Yes, I agree it will be very surprising considering the difference in atomic radii.