Doubts on cif files of ceria DOI 10.17188/1195334

At the attention of the community,
I am working on ceria surfaces for ethanol oxidation. I have really appreciated the explanatory work at the reported DOI. Probably I am too naive, but I can not understand the CIF data reported. There are 4 CIF files. I will just focus on the first two: “computed” and “conventional standard”. The cells described by the two files are quite different and it seems to me that the “computed” structure is a cell that exhibits the (111) surface, while the “conventional standard” one is a cell that exposes the (100) face. Am I correct?

I am deeply grateful for your time.

Have a nice day.


Hello Jonathan,

Welcome! :slight_smile:
There is a great post explaining the differences between the cif files that you can download from MP:

Miller indices are defined with respect to the lattice vectors of the conventionally centered unit cell. So, for your purpose (to create surfaces) you should use the conventional standard cif.

Hope this helps!



Thank you for your kind reply, I deeply appreciate the time you have dedicated. As a new user, I can not upload data, but I will try to explain myself here. The “computed” file and the “standard one” have the same cell parameters: angles and distances are pretty the same. But when I open the two cif files with a visualization program ( in my case is VESTA) and I extend the cell on the xy plane, the structure appears to me pretty different. In particular they seems to refer to different surfaces: the 111 for the comnputed file and the 100 for the conventional one.

Sure, happy to help!
Mhm strange - the computed (based on primitive cell) and standard conventional cifs of mp-20194 have different unit cell parameters. Computed has a = b = c = 3.866 A and alpha = beta = gamma = 60 degrees, whereas conventional standard has a = b = c = 5.467 A and alpha = beta = gamma = 90 degrees. Could you double-check that you downloaded the correct cif files?
But regardless, you should use the conventional standard cif file for surface generation!

1 Like

Thank you again for your time. This is exactly my doubt: the two files refer to different unit cell parameters. I have downloaded the files at the url mp-20194: CeO2 (cubic, Fm-3m, 225) from the same drop down menu.
Anyway, I will use the conventional standard file.

Oh, I see. Your previous post got me confused because you wrote “they have the same cell parameters”.
So, yes, they refer to different unit cells of the same crystal. Unit cells are not unique and you can describe the same crystal with different unit cells. As an example, look at Figure 3 (a) vs. (b) in this explanation for the FCC structure: Close Packed Structures: fcc and hcp | Physics in a Nutshell (a) displays the conventional unit cell of FCC and (b) the primitive unit cell of FCC. Those two unit cells have different cell parameters (e.g. angles in (a) are 90 degrees and in (b) they are 60 degrees) but describe the same structure (FCC).
Miller indices are defined with respect to the conventional unit cell (not the primitive one) because the unit cell vectors of the conventional cell are aligned with high-symmetry directions of the crystal.

1 Like

I deeply appreciate your kindness. Sure, the two files refer to the same crystal, and thank you for the example. Yeah, the doubt is not about the different shapes of the cell. The doubt is about the different expose surfaces when the two cells are propagated along the XY plane. I am sorry if my question was not so clear.

No worries! Yes, exactly - in general, different unit cells may expose different surfaces along their basis vectors. That’s exactly why a clear definition of Miller indices is necessary to avoid confusion. Starting from the conventional unit cell you can create any surface orientation by using Pymatgen, for example.
Good luck and let me know if you have further questions regarding surfaces!