I am trying to model the surface relaxation on the TiO2(Anatase-101) surface, but it confuses me that when I try to set some atoms as fixed, unlike the crystal that only contains one element, there is no clear boundary between each layer of atoms.

My question is:
For metal oxide, there is no boundary line between each layer, how should I choose the atoms that I want to fix?

enter image description here

  • 1
    $\begingroup$ How is the "boundary between layers" factor into this problem? Also, there are layers in oxides, too. Eg this looks like a 3 layer slab of anatase (101). $\endgroup$
    – Greg
    Mar 18, 2022 at 17:41
  • $\begingroup$ Why do you need to fix atoms? What are your purposes? $\endgroup$
    – Binh Thien
    Mar 21, 2022 at 5:36
  • $\begingroup$ The bottom two or three layers of atoms are to be fixed generally when dealing with a slab. This is done to simulate the bulk of the material, and also prevents your structure from disintegrating. It is easier to choose the layers in the case of orthogonal structures, tetragonal or cubic, but bit tricky in other cases. $\endgroup$
    – ansonthms
    Mar 21, 2022 at 6:44

1 Answer 1


Worst option in terms of precision, but simplest one: make fixed atom every crystall cell. This way you will see familiar crystalline structure, but a lot of forces will remain between these fixed points. You will only see changes in extreme case, like change of a crystal cell. Example: research about high pressure change to a crystal structure, probably superconductors.

Popular method with some precision: make atoms that are closer than a certain distance to a surface fixed. Idea is that in practice oxide is often attached to a substrate, and that substrate doesnt allow the oxide to move. You can study surface effects, catalytic activity for small molecules. Example: surface catalysts used in cars exchausts, hand warmers. Substrate plays an important role if oxide is a few nanometers, as it defines forces that keep the catalyst in a certain extended or contracted state. You can account for this by spacing atoms of the oxide's first layer accordingly. You will likely be able to see only a few layers, and then atoms' positions will become more chaotic, especially if spacing is far from what is found in a bulk crystal.

Emerging field, nanotechnology: make only one or a few far away atoms fixed. This allows to study most complex topics like nucleation, nanoparticle growth and stabilization. You will likely not see any familiar patterns, small nucleation sites are full of defects, and structure looks chaotic. Example: some UV protection for plastics, window films.

As usual, 'it depends'. What effects, what industry, do you want this study to be useful for? Select an appropriate complexity of a model based on that.

  • $\begingroup$ This answer needs improvement while it defines when you might do some of these things, the definitions across the board are slightly incorrect. $\endgroup$ Mar 23, 2022 at 15:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .