!!! A paper in JMEMS (vol. 7, n. 2., june 1998, pp. 252-260) describes the process !!!
This short page has been written because the information on this powerful process was not as freely available
as it should be (Hilton Head proceedings are quite rare among non-american
research staff...). I hope Internet could help in spreading this information
to a broader public...
For sure, no copyright or whatever similar is implied by myself on the process, but before using it in a commercial application, better check if a patent does exist or not...
My teacher was Andrew Yeh (Thank you Andrew!), from Norman C. Tien MEMS group at Cornell University. It is an application of the revised process described by M. Houston, R. Maboudian and R. Howe, from UC Berkeley, at the Solid-State Sensor and Actuator Workshop, 'Hilton-Head', in 1996 (pp.42-47) (see other references in this paper).
This coating seems very useful for decreasing the occurrence of release-related and in-use stiction phenomena with micromachined polysilicon devices. It reduces the work of adhesion by several orders of magnitude (3µJ/m2 instead of 20.000µJ/m2 for an oxidized hydrophilic surface) and increases the water contact angle above 90° suppressing the meniscus force (114° instead of 0-30° for hydrophilic surface). Practically, 2 µm-thick and 10 µm-wide beam over a 2 µm-thick sacrificial layer still had a 50% chance of release when they were 1mm long (110 µm long beam had the same probability of sticking with hydrophilic surface)! The release was 90% successful for beam length up to 600 µm (about 70 µm for hydrophilic surface).
It shows very good time stability at room temperature, with no change observed in air for period longer than 18 months.
Its thermal stability in vacuum or inert atmosphere exceeds 400°C, adequate for some packaging process. However its qualities decrease quickly in air (or oxygen) when the temperature increases beyond 100°C, which place an upper limit to the operation temperature with non-sealed packages.
It has shown superior wear ability as dry lubricant inside a wobble motor (80 million cycles over a nine months period).
The principle of this process is to keep the device in wet environment until it is fully coated with a hydrophobic self-assembled monolayer film. Thus care should be taken to avoid drying the sample before the very last step, and a layer of liquid has to be kept at every time on the surface, particularly when changing bath...
rem: you may use any process here, like diluted HF (5:1 H2O:HF) solution.
rem: this step may be suppressed if H2O2 is to damage metalization, but in that case SAM quality is thought to be poorer (cf. remove bulk water step)
rem: if bulk water is undesirable because it could cause OTS molecule to aggregate, a thin film of water on the surface is thought to be desirable to enhance the quality of the coating by increasing the mobility of the precursor on the surface. Hence the reason for the previous oxidation step (creating a hydrophilic surface).
rem: the original paper says 1 drop (~20µl) of OTS per 50ml of solution, I think it is per 5ml of solution, or, alternatively, the drops are smaller in my case (maybe ~2µl or 2mm3). From the original paper, a typical solution would be 1mM OTS solution. A proper way to prepare the solution is to add drop one by one and when the solution become 'viscous' you know you have been too far... Throw away the solution (it will certainly create polymer deposit on your structure), start a fresh one, count the drop and stop before the previous number... If you still observe many polymer aggregate on your sample you know you have been too far :-) According to the paper, if the concentration of OTS exceeds 5mM polymer formation on sample surface may happen.
rem: the temperature of the SAM solution should never exceed 28°C, the transition temperature of the solution below which the SAM films display best quality.
Created 03/1998, revised 09/2000