Understanding structure formation in thin polymer films

This month’s issue of Nature Materials has a paper from my group which provides new insight into the way structure can emerge in ultrathin polymer films by self-assembly. It’s very easy to make a very uniform polymer film with a thickness somewhere between 5 and 500 nanometers; in a process called “spin-casting” you just flood a smooth, flat substrate with a solution of the polymer in an organic solvent like toluene, and then you spin the substrate round at a couple of thousand RPM. The excess solution flies off, leaving a thin layer from which the solvent quickly evaporates. This process is used all the time in laboratories and in industry; in the semiconductor industry it’s the way in which photoresist layers are laid down. If you use, not a single polymer, but a mixture of two polymers, as the solvent is removed then the two polymers will phase separate, like oil and water. What’s interesting is that sometimes they will break up into little blobs in the plane of the film, but other times they will split into two distinct layers, each of which might only be a few tens of nanometers thick. The latter situation, sometimes called “self-stratification”, can be potentially very useful. It’s an advantage for solar cells made from semiconducting polymers to have two layers like this, and Henning Sirringhaus, from Cambridge, (whose company, Plastic Logic, is actively commercialising polymer electronics) has shown that you can make a polymer field effect transistor in which the gate dielectric layer spontaneously self-stratifies during spin-coating.

The paper (which can be downloaded as a PDF here) describes the experiments that Sasha Heriot, who is a postdoc in my group, did to try and disentangle what goes on in this complex situation. Our apparatus (which was built by my former graduate student, James Sharp, now a lecturer at Nottingham University) consists of a spin-coating machine in which a laser shines on the film as it spins; we detect both the light that is directly reflected and the pattern of light that is scattered out of the direct beam. The reflected light tells us how thick the film is at any point during the 5 seconds which the whole process takes, while the scattered light tells us about the lateral structure of the film. What we find is that after the spin-coating process starts, the film first stratifies vertically. As the solvent is removed, the interface separating the two layers becomes wavy, and this wave grows until these two layers break up, leaving the pattern of droplets that’s seen in the final film. We don’t exactly know why the interface between the two self-stratified films becomes unstable, but we suspect it’s connected to how volatile the solvent is. When we do understand this mechanism properly, we should be able to design conditions for the spin-coating to get the final structure we want.

The relevance of this is that this kind of solvent-based coating process is cheap and scalable to very large areas. The aim is to control the nanostructure of thin films of functional materials like semiconducting polymers simply by adjusting the processing conditions. We want to get the system to make itself as far as possible, rather than having to do lots of separate fabrication steps. If we can do this reliably, then this will get us closer to commercial processes for making, for example, very cheap solar cells using simple printing technology, or simple combinations of sensors and logic circuits by ink-jet printing.

2 thoughts on “Understanding structure formation in thin polymer films”

  1. Polymer solar cells will be a nanotech killer app. I’m curious about the laser that was used in this experiment. Was it customized for this procedure or is it a typical spin-cater “accessory”? Seems some newer polymer R+D is cheap enough to attempt at home; I like.

  2. The laser is a standard lab HeNe, and the detector is a relatively cheap CCD camera. What’s home-built is actually the spin-coater, which needs to be more mechanically stable than the bought models, and the laser and detector mount; all these were made in our workshop. The whole thing is driven by Labview on a PC. I’d guess it cost about $10,000, depending on how you cost the workshop time (it came in hugely under budget, thanks to James’s ingenuity and experimental acumen).

    I’m also hugely excited about polymer solar cells, and we’re planning to do much more work on them. I should, however, say that we do have some more expensive kit around which is going to be useful for this as well – in particular my colleague David Lidzey has a scanning near field optical microscope coupled to a fairly sophisticated spectroscopy set-up, which allows him to look at the optoelectronic properties of individual, phase separated domains of semiconducting polymer on a sub-100 nm length-scale. Structure on the nm length-scale is important too. We can’t of course see this with the light-based techniques, so at some point soon we’ll have to bite the bullet and put together an x-ray version of the laser instrumented spin coater.

Comments are closed.