photoelectric lithography

That's basically how the original electron microscope worked (before the invention of the SEM), only backwards. There are a lot of difficulties with it, especially in maintaining decent image correction over a large field.

Cheers

Phil Hobbs

The best lithography is e-beam, and it's used to make masks. But the beam currents are limited by space-charge distortion, so throughput is too low for volume production.

The new-new thing is EUV, 13.4 nm light:

I saw a presentation on the Zeiss lens that's used on these things. It looks more like a big diesel engine than any sort of lens. It does something like 4:1 reduction from the mask to the wafer. There are no lenses at this wavelength, so it works entirely with grazing-incidence metal mirrors. I still can't understand how that's possible.

Hi,

I guess its like a transmission electron microscope in reverse, I think there are some modern aberration corrected electron optics that seem pretty amazing, not sure how they work exactly, but can do spatial resolution down to 0.05nm apparently.

cheers, Jamie

I saw a strange lens recently too, for terahertz light, it is the same shape as a glass convex lens, but it is made out of teflon, I guess it works by changing the permeability and permittivity of free space?

cheers, Jamie

IBM was going to do direct-write E-beam lithography. They even spent $1B on their own synchrotron (the only privately owned synchrotron) for the source. ...then some wise-ass invented the phase-shift mask.

The ASTC (Fishkill Bldg 600) synchrotron was for X-ray litho, not e-beam. I used to collaborate with some folks up there.

Cheers

Phil Hobbs

Of course not. You can't change free space. You can make it not so free, though. ;-)

You're right.

They were doing direct-write E-Beam before that (for a while, anyway).

It's still used in research and development, though it's beginning to hit its resolution limits. When I was at IBM, I made my nanoantennas using one of the Leica E-beam writers in Yorktown. They were good to about 10 nm if you used PMMA resist and developed it in IPA+water. IPA+water needs more exposure than MIBK/IPA but doesn't make the resist swell, so it's easier to hold the dimensional accuracy.

Cheers

Phil Hobbs

It has to want to change itself.

Space charge doesn't come into it with regular scanned-spot electron beam microfabricators. The restriction comes from the electron source, which defines the amount of current you can put on the specimen - the the smaller the spot on the specimen, the less electrons you can put into it.

"Shaped beam" electron beam microfabricators get around this - to some extent - by illuminating a defined area on the specimen rather than a scanned spot, but then space charge does restrict the area that you can illuminate. The machine I worked on a Cambridge Instruments could

- in principle - illuminate a 10 micron by 10 micron square, but anything over about 10 square microns would get messed up by space charge, so the patterns to be written were broken up into rectangles that could be up to 10 microns long or high, but only if they were narrower than one micron in the other direction.

There were a bunch of shaped beam machines at Aix-en-Provence in the

- and Cambrigde Instruments was using them in production instruments when I worked there (until November 1991).

-- Bill Sloman, Nijmegen

Hi,

The throughput for e-beam is limited because it is scanning instead of projecting a 2D image, but if you project a 2D electron beam image the throughput could be the same as photo lithography and the beam current density equal (or less) than scanning e-beam.

cheers, Jamie

People have been claiming that for a long time now, but they've yet to put anything together that works.

-- Bill Sloman, Nijmegen