There was an interesting presentation that happened towards the end of SPIE Litho – it seemed to catch the audience off guard, and I frankly went away with the sense that there was some confusion in the room.

The presentation discussed an experiment that was done at Osaka University as part of the overall effort to optimize EUV exposure. It all relates to this seemingly inviolate triumvirate of “RLS”: resolution, LWR (line-width roughness), and sensitivity. Improvements within these three have to come at the expense of something within these three – they form a zero-sum game.

Normally, you expose the photoresist through the mask for the entire length of the exposure. The photons create acid where they interact with the resist, and this acid provides for the selective removal of resist material during development.

This experiment changed that. The exposure was broken into two steps:

• A short exposure through the mask
• After 10-15 minutes, then, with no mask, just a flood of UV across the entire wafer.

The first exposure seemed to create some acid, but mostly “sensitized” the photoresist (and I frankly didn’t come away understanding what that “sensitizing” meant from a chemical standpoint). The strange thing then was that flooding with the second exposure created the normal amount of acid only in the sensitized area.

This provided about 9 times the prior sensitivity, with no apparent tradeoff in LWR or resolution.

Note that no special resists were used; these were the same resists as are currently being used.

I didn’t get the sense that they had a real handle on what the underlying mechanisms were, and it was surprising to the audience. Assuming the data are correct, it’s certainly an interesting result. We’ll have to see if anything further comes of it, or if it goes the way of cold fusion…