Toward computing with atoms and molecules

[I wrote this blog post on November 17, 2017, but somehow I failed to publish it until Sep 28, 2019..]

Last summer, within a short span two papers showed up on arxiv: 

1. Passivation of dangling bonds on hydrogenated Si(100)-2×1: a possible method for error correction in hydrogen lithography by Niko Pavliček, Zsolt Majzik, Gerhard Meyer, and Leo Gross of IBM Zurich (on arxiv and published in Appl. Phys. Lett.)

2. Binary Atomic Silicon Logic by Taleana Huff et al. in the Wolkow's group at the University of Alberta (on arxiv and published in Nature Electronics)

The group at IBM Zurich, the site of the invention of scanning tunneling microscopy, is undoubtedly a world leader in atomic and molecular manipulation. I had the pleasure of visiting their lab in September. Their AFM/STM work with planar, conjugated molecules on an insulator-covered metal surface is elegant and impressive. So it’s nice to see that they haven’t quite forgotten about silicon.

The Wolkow group has been working toward this outcome for many years. They have conducted pioneering work on passivated silicon surfaces (H:Si(100)), including the styrene chain reaction. Here, they presented a case of "binary atomic silicon logic (BASIL)" by constructing an OR gate using a ‘Y’ pattern of silicon dangling bonds.

One essential element to achieve atomic-scale computation at surfaces is the ability to instantly reuse the circuit, to write and erase the bits in the BASIL. And this is where the IBM Zurich group’s work comes in. Their preprint described a procedure for correcting errors in hydrogen lithography. In other words, they can erase exposed dangling bonds. (So, I later realized that the Wolkow group did something very similar in this paper, covering dangling bonds with a hydrogen atom carried by the tip.)

Before we achieve an actual atomic computer, I still see huge challenges with scaling, programming, speed of calculation (including I/O), and error correction. Nevertheless, these are two important papers along the road to the ultimate miniaturization of computers — it doesn’t get much smaller than atoms!