Shrinking Electronics

Chris Phoenix is providing live blog coverage for us on all the presentations from an important conference on Productive Nanosystems: Launching the Technology Roadmap...

Next talk: "Atomic-Scale Device Fabrication in Silicon"
Michelle Simmons, School of Physics, University of New South Wales, Australia

Michelle will be talking remotely from Australia, about making silicon electronic devices at the atomic scale.

In 2020, Moore's Law says we'll be at the atomic scale. We'll need deterministic doping [putting atoms where they will affect the electronic properties of the silicon]. We'll also need atomic level control of the interfaces between different materials.

The plan is to use single phosphorus atoms as quantum dots.

Doping error [variance] is the square root of the number of atoms. So if you have 10,000 atoms, the error is 100, which is 1%. But with 100 atoms, the error is 10, or 10%. That means the threshold voltage is not reproducible between transistors.

Quantum effects dominate at this scale [for electrons, not atom position!] - can this be used?

Silicon atoms in a surface can't be moved around the way Eigler moved metal-on-metal atoms; the silicon atoms are strongly bonded. [But Oyabu did manage to remove and replace single atoms; but that's more cumbersome than being able to push them around the surface.]

The goal is to make atomic features. Remove hydrogen atoms from a hydrogen-terminated surface, deposit phosphorous-containing gas, heat it to incorporate the P in the surface, then deposit more silicon on top, then deposit electrodes above the buried dopant atoms.

To understand what the microscope was seeing, when looking at PH3 gas on the surface, they had to calculate the energy level of lots of different configurations, then use density functional theory to simulate what it would look like to the microscope... then they could go back and identify surface features from the microscope image.

By calculating what the phosphorous does as it loses hydrogens, and how it incorporates itself in the silicon surface, they can now place P atoms with atomic precision.

She's talking about amount of phosphorous vs. temperature. That doesn't sound atomically precise. I guess it's a different research direction.

It's possible to see buried 7-nm-wide wires reflected in the electronic properties at the surface.

It's possible to count the dopant atoms. And then, by building a Hall effect structure, count the charge carriers - and each atom creates a charge carrier. Similarly, they can demonstrate that the STM tip can completely remove the H protectant and let all possible P in.

OK, this next thing is really cool. They can build structures narrow enough to affect quantum effects. It goes like this: if electrons are able to make a coherent quantum loop between dopant atoms, they go around the loop in both directions at once, interfere, and are blocked; this increases the electrical resistance. Lower temperature and magnetic field allow bigger coherence length, bigger loops, and thus more resistance. But if they build a sufficiently narrow wire, then the biggest loops can't form, and the resistance doesn't spike as high. They can calculate the size of the loops that didn't form, and it matches the width of the wire they built.

Something cool that I didn't catch while writing the previous paragraph: ordered dopant vs. random dopant leads to an Arrhenius voltage curve... or something like that.

They can make single silicon dioxide layers by depositing a layer of oxygen, then a layer of silicon (at low temperature), then heating it.

They can build circuits using combinations of surface electrodes and buried gates.

They're working on nanoscale MOSFETs, 3D transistor architectures, atomically precise resonators, silicon-based quantum computers.

My observation: Although this is not moving/mechanical nanostructures, it is an example of atomic-precision fabrication. It's mainstream, it's semiconductors (which means that there'll be commercial attention), and it leaves no doubt that stable single- and multiple-atom, atomic-precision structures are being built by scanning probe microscope. This should go a long way to blunting claims that atomic precision fabrication is impossible on either practical or theoretical grounds.

Chris Phoenix

Tags: nanotechnology nanotech nano science technology ethics weblog blog