You can make a layer of the thickness of even only one atom.

Yes, you have to somehow generate a surface/slab geometry to represent the thin film.

Ok thanks for your answer but can we consider slab/surface as thin film  even the thickness is just only couple of nanometer?

What general kind of DFT calculation do you have in mind with respect to the type / symmetry of your system: a calculation for molecules with a fixed volume and localized basis functions, or a calculation for crystals with periodic boundary conditions and, typically, a plane-wave basis? In case of the latter, be aware to include enough "layers" of vacuum, since a thickness of a single atom will give you a bulk calculation.

Ok thank you for your answer.By the way, i'm using wien2k to calculate crystal structure.

I think that it depends, also, on the nature of the material used. Could you tell us, please, what kind of material are you using ?

The kind of material is not necessary to be known I think .the code of simuation has to take into account the thickness. For ex Wien2k treat only the bulk. It depends mainly on the simulation code used

Afiq Radzwan: "... but can we consider slab/surface as thin film even the thickness is just only couple of nanometer?" What do you think is a possible reason that this might not be the case?

Jan-Martin Wagner, I just misunderstand between thin film and ultra thin. Because I read some papers, they calculate the surface/slab to represent thin film. but I think it is ultra thin film. My focus is to calculate thin film with thickness up to 100nm but I think it is  impossible to use WIEN2K because WIEN2K only allow calculation below 1000 atoms.

Then I don't understand your starting question at all -- shouldn't it have been "What is the MAXimum thickness of thin film structure for DFT calculation" instead?