Yet Another Escape Hatch for Moore’s Law
July 15, 2009 | Tom Tritton
Moore’s law states that the number of transistors in integrated circuits doubles about every two years. This is why computers keep on getting smaller, why memory chips keep increasing in storage capacity, and why digital cameras keep having more megapixels.
But can Moore’s law hold forever?
Of course not, even though it has held steady for nearly half a century. It’s hard to imagine a transistor being smaller than a single atom, and even Gordon Moore himself has pointed out that exponentials always crash at some point.
The limitation at the moment is the photolithography process that lays down circuit diagrams on silicon chips. Current technology can produce features as small as 45 nanometers. But suppose you could use carbon nanotubes (about tenfold smaller at 4nm in diameter) as the mask for the lithographic process?
You can, at least in principle, thanks to new work from Columbia University chemists. Liu et al. deposited carbon nanotubes on silicon to mask the wafer for etching by hydroxide ions (Journal of the American Chemical Society (2 June 2009), doi: 10.1021/ja903333s). They cajoled the reaction sufficiently to get trenches about 4–6 nanometers deep, but alas ten times as wide.
The authors speculate that the resolution could be improved by clamping down the wandering tendency of the carbon nanotubes through clever chemistry. This would then yield the hoped-for tenfold increase in masking potential over current photolithography.
So Moore’s law may continue to hold for some time to come. With luck, we’ll postpone the inevitable exponential crash just long enough to come up with Moore’s law 2.0.