The non-uniformity in the produced films could stem from a number of factors.

Be sure to shake and/or stir your polymer solution for an appreciable amount of time before applying the spin coating process, so that the solutes can be as homogeneous as possible in the volume of the polymer solution dispensed on the substrate.

With reference to the solvent used for the spin coating process, toluene has a significantly lower vapor pressure compared to solvents such as chloroform, and in that light, toluene alone is not responsible for the non-uniformity. However, one thing you could do for further suppressing the aforementioned factor, is to produce a solvent saturated overhead gas phase in the remote vicinity of the substrate, so that the toluene will evaporate at a lower rate during the spin coating process. In the relevant literature, extremely volatile solvents or the absence of a solvent saturated overhead gas phase have been found to increase non-uniformity in thin polymer films.

As a third measure, I would advise you to make your solutions denser, i.e. 1.5 or 2% (w/v). Of course, you should check whether the produced solution acts like a Newtonian liquid at these concentrations, because otherwise striations on the film surface may be formed due to non-Newtonian behavior during the speed coating process.

I am a new comer but still I have a suggestion... try to use dispersed phase other than solution to spin coat if possible.

To get a homogeneous and uniform coating you can use a PDMS prepared in a plasmapolymerization process (PE-CVD). Thicknesses between 10 nm and some microns are possible. No degassing, no other elements only PDMS 3-dimensionally crosslinked.

I think the first step will be to try to increase the concentration.

Does the film peels off nicely?

Once you get a thin and uniform film, follow the suggestions by others to clean the substrate surface.

How you are measuing the film thickness and the uniformity?

Spin coating the good old tried and proven method for film formation. Practice with other polymers to get a feeling for the controls (e.g., cellophane). Nect try with the speed variation.

It would be helpful if you could describe in detail the type of non-uniformities you are seeing. For example, non-uniformities related to pin-holes, areas of de-wetting, striations, phase separation ( between PDMS and curing agent). Also are you using a commercial spin coater. What is the coating RPM? What is the ramp-up acceleration? Are you working in a clean environment ( e.g. using a Class 100 hood). How do you apply the solution? Do you completely flood the wafer while stationary, or do apply your coating solution to a slowly rotating wafer etc. What is your target dry thickness?

As mentioned by others, prepping the wafer surface before coating so that you have a clean surface that is completely wetted by the solution, is essential.

You mention that you add a curing agent. Is that curing agent completely soluble in the toluene, in the presence of PDMS? Or are you getting some phase separation effects as the toluene evaporates....

Spin coating sounds deceptively simple, but it is a dynamic process in which film flow, solvent evaporation, and coating solidification are all occurring though often with different time scales. Managing these (often competing) processes during the spinning is challenging......

As suggested by Athanasios Kakalis, you have to stir the solution for half an hour.

For spin coating solution should be viscous. Then you will get uniform coating.

Increase the spincoating speed should provide uniform spin i reckon.

You can try biasing the substrate or preheating the substrate.

For many (but not all) polymer/solvent systems the "average" film thickness decreases as the square root of the spin speed RPM^(1/2). But in general an increase in spin spin does not guarantee good quality films. The key is first to understand why there are imperfections/defects. For example, it is well known that the presence of striations in the spin coated film arise from Marangoni instabilities due to surface tension gradients, which can be suppressed by the addition of a suitable surfactant. Defects due to phase separation can be minimized/reduced by designing a suitable solvent mixture, where the "good" solvent has a lower vapor pressure. Wetting defects can be treated by coating the surface of the wafer with an appropriate wetting agent, and pin hole defects due to airborne contaminates an be minimized by working in a clean-room environment., and so on. Check my publications for article on spin coating of colloidal suspensions ...

@Raja Baddula

Check the following two papers:

1. Yimsiri, P., Mackley, M.R., «Spin and Dip Coating of Light-Emitting Polymer Solutions: Matching Experiment with Modelling», Chemical Engineering Science 61 (2006) 3496-3505.

2. Chang, C., Pai, C., Chen, W., Jenekhe, S.A., «Spin Coating of Conjugated Polymers for Electronic and Optoelectronic Applications», Thin Solid Films 479 (2005) 254-260.

The most uniform film has an optimal thickness, which depends on both structures of substrate and the film. If the film is too thick then a non-uniform self-organization occurs but if it is too thin then island structure appears. So, you should find this optimal thickness; usually it is around 20 nm and appear just at concentration 1% and 3000 ppm. Of course, the surface should be very clean!

Try using an alkanethiol or better a PDMS-thiol coating aid. The thiol functionality will bind to the gold surface and the rest will interact with the PDMS, making the PDMS feel more at home". I agree with the person who suggested increasing the concentration, as you may have too little PDMS to form a monolayer. I am not sure what to advise about coating on silicon, but if the surface is SiO2, carboxylate-ended amphiphiles rather than thiol-ended may work as described for gold.

I think you have to increase the concentration of PDMS and increase the r.p.m...

from my experience,....the solution must be homogen, for certain viscosity find the right rpm more viscous more rpm. check the spincoater whether flat or not during turning. chose the method to coat, drop the solution and then turn on the spin coater or turn on the spin coater and then drop the solution. The subtrat must be really clean, there some method to clean the substrat.

You have to find the exhaustive crosslink for turning the viscosity suitable a long with the extension of reaction. If you are using a photochemistry curing, you can find the best point and after assembled, let the sample crosslinking again to the final.

I hope it be usefull

Regarding Brian Higgins answer if you face some Marangoni effect one possible solution is to use cyclohexane (good solvent of pdms) instead of toluene. You will have less prononce surface gradient effects

Have you ever tried to introduce some surfactant to the system ?

I agree with the first answer. Substrate state is primordial. Otherwise, I think that you mus try another solvent more volatile;for instance carbon tetrachloride. You can look for our method in : Vibrational and Electronic Excitations in Poly(methylphenylsilane) and

Poly(diphenylsiloxane) Films: Surface Aspects

Ana Maria Botelho do Rego, Olivier Pellegrino and Manuel Rei Vilar

In addition to the above views, i would like to add that using a two step process of spinning i.e. start at a lower spinning speed and then after the solution uniformly spreads over the entire substrate, then spin at a high speed to control the thickness.

RCA method is adequate for silicon substrates pre-treatment.

If you take a rapid evaporation solvent as tetrachloride carbon, you should use one-single step spinning for instance at 2000 tpm with an acceleration of 800 tpm/s. The thickness of the resulting film mostly depends on the concentration of the polymer in solution.

Thank you all very much. So far a used Toluen and it worked fine if a stir the pdms with the curing agent dissolved in toluen over night. prior to spin-coating. The films are now uniform on both the silicon and gold crystals up to 1% concentration. I also tried hexane as solvent which works even better than Toluen, the films have less gradient effects or almost none. Mr. Malloggi i am also interested in trying cyclohexane. Thanks for the advice. And so far i am cleaning my surfaces by short exposure to piranha solution, but am interested also in UVO cleaning procedure, but first the UVO chamber has to be acquired.

I would suggest to use Hexane or C-Hexane instead of Toluene. I've got good results with them.

I do wish we could have some preciseness. It is not PDMS when cured into a silicone rubber. You probably have some local gelation of the silicone into rubbery particles. Toluene can be quite wet promoting re-equilibration of the PDMS into chains and rings and in the presence of catalyst one could expect a non-uniform film. Hexane will be dry and so this is the reason it works better. Silica surfaces can be sufficiently acidic to promote these reactions.

Well i did some basic light microscopy on pdms films spin-coated from toluen on a silica surface and at higher concentrations some local differences in the film thickness can be observed which look like some kind of aggregates or might even be rubbery particles. At very low concentrations this is not observed but a have prepared some films from hexane but they are yet to be characterized.

I have obtained the highest quality films of PDMS by spin coating from n-octane solution. Sylgard has filler particles which make very thin films less than optimal quality. If the films spin coated from octane are still not good enough, try one of the ambient cure solutions sold by Gelest in methyl-THF solution and dilute them with n-octane.

Hexane worked very well with me. but how can we prove that the resultant thin film is purely PDMS (Hexane-free) from a scientific point of view? I mean the boiling point of hexane is 65, I baked my sample at 100C, so the hexane should evaporate at the temperature. But some reviewers of my paper argue that maybe the PDMS thin-film may not be purely PDMS . how can I argue ?

Well Mutasem, it would be nice if you have a possibility to measure some XPS on your films. You could then compare the ratio of the Si, O, and C atoms in your material with theoretical values of the PDMS, which should be 25%, 25%, and 50% respectively. But might be difficult to see an increase in C if you have really small amounts of Hexane still left in the film. 

Thanks Matej, Good approach. I will perform EDS analysis instead to test the elements ratio.

I do not think this is a question you would be asked in review. To address you could take film thickness measurements as a function of bake time, doubling the bake time interval at each measurement. when you get the same thickness after 2-3 measurements your film is dry.