It is hard to be precise in answering your question as there are many parameters. Let us assume that the initial chemical formulation for the sol-gel system and the post drying/annealing process are the same. Then the only way to get a different film thicknes is by changing the spin speed. The higher the spin speed, the thinner the film. However, it is possible for these films to exhibit a Marangoni type instability ( as a result of the variation in surface tension due to temperature (evaporative cooling) and concentration effects- loss of solvent due to evaporation) that will generate "roughness" when the films are annealed. To avoid the roughness effects one requires optimizing the solvent and the acceleration ramp of the spin coater. You can download some of my papers from ResearchGate that deal with this issue.

It is hard to be precise in answering your question as there are many parameters. Let us assume that the initial chemical formulation for the sol-gel system and the post drying/annealing process are the same. Then the only way to get a different film thicknes is by changing the spin speed. The higher the spin speed, the thinner the film. However, it is possible for these films to exhibit a Marangoni type instability ( as a result of the variation in surface tension due to temperature (evaporative cooling) and concentration effects- loss of solvent due to evaporation) that will generate "roughness" when the films are annealed. To avoid the roughness effects one requires optimizing the solvent and the acceleration ramp of the spin coater. You can download some of my papers from ResearchGate that deal with this issue.

@Brian G Higgins

Dear Professor Higgins,

Thank you so much for the time to write and respond me. I have downloaded and read your paper entitled, “Marangoni flows during drying of colloidal films”. Although our material is more like a solution than a colloid, what I understood from your paper is that when Marangoni stresses are sufficiently strong based on annealing process (which is the only different item in our samples fabrication process), the agglomerated particles can be formed more easily on the surface of thin film. So, and according to your research, it can be convincing that annealing can be an important factor in roughness and disturbance growth rate.

Now, I have two clues, Marangoni type instability and the fact that in formation of a thin film, the crystallites join together from the lower parts and form integrated layers from the bottom. Therefore, by more bonded crystallites we can obtain a thin film with lower roughness or, in other words, with higher thickness of grown film.

Finally, can I conclude that, in our case, annealing has increased the disturbance growth rate and this has led to a reduction in the thickness of interconnected crystallites which results in higher values of roughness than thickness? Is this inference reasonable?

@Thomas Breuer

Dear Professor Breuer,

Thanks a lot for your response. Powder samples of the deposited material showed the nanoparticles size in the mentioned range. However, we are not facing with the problem of cutting the samples for SEM cross-sections test, because we cut the substrates (suitable for SEM surface and cross-section test) before deposition and spin coating process.

Therefore, from this point of view, your suggestion about the size of nanoparticles and the analysis of Professor Higgins are in the same direction and both of these were very helpful to me. Thank you so much.

@Thomas Breuer

Dear Professor Breuer,

It is my pleasure to write you again. Our material is Nb doped TiO2 (TNO) and according to our previous work and your suggestion, the deposited TNO on silica is sensitive to heat treatment. We are investigating the effects of annealing on TNO's electrical and optical properties and in our previous work we performed different annealing conditions result in different electro-optical properties. However, in our last work with a different annealing mechanism we saw larger roughness in deposited thin film than its thickness which was unconventional for us. According to the discussions of this page, it seems that it can be related to the new annealing treatment we used. Although, I found the article of Professor Higgins suitable as a reference of this issue, I would appreciate it if you could recommend me another reference about it.

@Thomas Breuer

Dear Professor Breuer,

Thanks for your attention. Although highly conductive and transparent TNO thin films can be obtained by sputtering method, industrial productions are limited by the expensive vacuum technique. Furthermore, the homogeneous and large-area films preparation is also an upfront challenge yet. A sol–gel method has several advantages including simplicity, low cost, easily controlled doping levels and feasible preparation of large area films. In fact, the most reason of using sol gel spin coating method is its low cost processing and we think it is an urgent for thin films to be inexpensive. Therefore, we are trying to improve the electro-optical properties of TNO by this method.

Something is probable. 1-your machine is not working well.

2- its round per minute is not selected carefully, try higher rmp

3- check your precursors

Milad Fallah

Check out the work by Shari et al: Effect of annealing time on memristive behavior of sol-gel spincoated ZnO-based memristive device (Conference Paper Effect of annealing time on memristive behavior of sol-gel s...

In this work the researchers address the effect of annealing on ZnO films. Bottom line: As the annealing time is increased the film roughness decreases , but if the annealing is done over a long period ( 120 mins) roughness starts to increases and the authors attributed this behavior to the growth of grains. So what you suggest is possible. Note: surface roughness due to Marangoni instability occurs during the evaporation of the solvent. Once you stop the spinner the instability is "frozen in" and the amplitude may decrease slightly with additional drying and film heating. The Marangoni instability generally results in complex surface striations near the periphery of the wafer. At the center you see the typical Bernard pattern- cellular structure. Hope this helps.

@Elham Easy

Dear Elham,

Thank you very much for your response and suggestions. I think our spin coat machine is working well because we are fabricating other thin films of ITO, ZNO:Al or FTO with good results, but your suggestions about rpm or precursors can be two topics for further exploration of this issue.

However, in this case, we believe that this phenomenon occurs during an annealing process based on the principle of reproducibility in our work and we seek a reasonable reason for that. Thank you very much.

@Brian G Higgins

Dear Professor Higgins,

Thank you again for your time. The reference you provided was very useful and I can use it along with your article. Also, thank you very much for introducing me with the Marangoni type instability and Bernard pattern- cellular structure concept. I will definitely use these concepts in analyzing my work. Thanks for helping me with this.