the problem
Silicon nanowires are crystals of silicon which diameter is smaller than 200 nm and can be several tens of micons long.
Those seen on the picture are about 50 nm in diameter and one micon long. This is 1000 times thiner than a human hair!!!
Numerous applications have been indentified in the field of energy, electronics, nanoelectromechnical systems, biology and chemistry.
The usual way to fabricate silicon nanowires is to use golden nanoparticles to catalyse their growth. Unfortunately this approach is
incompatible with the semiconductor industry because gold poisons silicon electrical properties.
solution
It seems natural to try using another metal to catalyse the growth.
Unfortunately, growth can be achieved only at very high temperatures with other metals (>500°C).
This is unfortunate because high temperatures (>450°C) destroy devices (transistors) already fabricated on the chip.
One has therefore the difficult task to work at low T with non-gold metals. These conditions are usually refered to as
CMOS-compatible.At Leti, we found a way to synthesise silicon nanowire at 400°C with copper-based catalyst. We therefore achieved the long waited CMOS-compatibility. Interesting is the way we achieved it by oxidizing copper prior to growth.
This is at odd with usual practice because oxygen is though to be very bad for nanowire growth. This belief is inheritated for the
long tradition of work with golden catalysts. We showed that it is the exact oposite with copper and that it is even helpfull to decrease
synthesis temperatures. In addition, we performed a detailed analysis of the nanowires and in particular its evolution after exposure to room oxygen. The catalyst degrades as illustrated be the tomographic image (A 3D reconstruction of the nanowire using Transmission Electron Microscope).
More details available in the Nature Nanotechnology article or in the free version found on Cornell's archive.