Yes upto a great extent, but you need to cool down the sample at least up to liquid nitrogen level.

You could try dopants, which increase the crystallization temperature.

The answer is yes but needs cooling with liquid nitrogen

It depends on the materials. Some material is easy to get smooth amorphous film, such as Al2O3, SiO2. But some material is hard to get amprphous film, such as ZnO.

As mentioned before by Palermo and Chen it depends on material, and then maybe on process parameters for a given temperature. For a given temperature you could play with the energetics of the arriving species, lowering energy might help (which you can always do while not everybody has a cooled manipulator/substrate holder). I don' t know what material you use, but in case evaporated films would already show nanograins you might be out of luck any way...

Thank you all for contribution. It seems like I need to clarify some aspects more. I am sputtering a double component A-B alloy where A and B are both metals. The film that I get is no doubt already amorphous but the surface is not smooth. It has nanograined structure. As the machine that I use is very simple RF magnetron sputter without any features like liquid nitrogen cooler for substrate or I can not apply negative potential to it (and machine is not ours and we can not modify it in any way), I'm looking for more simple solution (if it exists) as playing with sputtering rate, or pressure inside the chamber or the temperature of the substrate (It can be heated up).

Single component metal films deposited with standard magnetron are usually polycrystaline.

- Standard way to make amorphous films is to cool the substrate. You may try to use a heat sink or a rotational substrate holder to keep the temperature low.

- Another way to obtain an amorphous material, is to use multi-component alloy. You have to play with composition and dopants - there is a number of publications and patents in this field.

- You have mentioned that "unwanted nanograins are amorphous and not crystalline. " This is a surprise for me. Usually the grains are just microcrystals. You may check if they are not some sort of technological debris, coming e.g. from the walls of a vacuum chamber or a composite target.

By the way, "amorphous" is imprecise term, its meaning depends on your requirements, applications and limitations of measurement apparatus.

To get an amorphous film, on has to cool faster depending on the material one uses. Depending on the sample holder distance and speed of sample rotation, one can get considerably reduce the surface roughness of the film.

Based on my experience in RF sputtering, deposition of metals under high pressures lead to smoother films with poor crystalinity, but for the case of oblique angle deposition. And depending on the material of your target you need to optimize and see at which high pressure you can get a smoother film with poor crystalinity. On the other hand, the manipulation of power can also help to smoothen the film. In general, large power results in developing films with small grain sizes. Again this one also needs to be optimized based on the target's material. This two papers may help:

"Effect of Sputtering Power on the Nucleation and Growth of Cu Films

Deposited by Magnetron Sputtering"

http://pubs.acs.org/doi/abs/10.1021/la104891u

Generally, thin films prepared under lower temperature tend to be amorphous. But it is strongly dependent on the material/

We have carefully studied ZrO2 and polymer (PC, PMMA) thin films deposited by pulsed laser deposition. They are (also for large film thicknesses of > 1µm) absolutely smooth with a RMS roughnesses of less than 0.5 nm.

........ The reason is the high kinetic energys of the deposited particles. So you should try to sputter deposit the materials at lowest argon pressure possible to increase the kinetic energy of the deposited particles.

I have seen some kind of similar behaviour for PLD deposition, where the high energy of incoming species lead to amorphization of the surface of an oxide thin film. But bear in mind that energies for PLD without any or very low moderator gas pressure are typically an order of magnitude or more higher than in the case of sputtering at low pressures (ballistic regime)...

yes, you need to cool down the substrate via cooling down the sample holder normally at liquid nitrogen temperature. if also possible along with cooling, spinning will also be helpful

If someone is still interested in this question. We investigated the sputtering process and one can find the results in our paper "Formation of nanostructured metallic glass thin films upon sputtering". Full text can be found in my profile