Tikalon Blog by Dev Gualtieri

Reaching the Bottom

May 3, 2011

Everyone is familiar with Richard Feynman's December 29, 1959, talk, "There's Plenty of Room at the Bottom ." As we have found, there's a lot of playing field at the nanoscale . Thirty years ago I was working with magnetic bubbles devices with three micrometer (3000 nm ) features. After a few years, we had pushed this scale to 800 nanometers . Now, integrated circuits are routinely made using 100 nm design rules .

Very small photograph of Richard Feynman

Feynman on the nanoscale.
This very small version of Richard Feynman's Los Alamos National Laboratory ID badge photograph is the largest clear public domain image available on the Internet.
(US Government photograph, via Wikimedia Commons ).

Aside from the practical difficulty in producing multiple instances of smaller and smaller features, is there any intrinsic limit as to how small a nanoscale device can be? After all, materials are made from atoms , and the properties of a small clump of atoms are not the same as those of a larger chunk. The question of how small we can actually go - where the "bottom" really exists - is one that Michael Kelly of the Cambridge University Centre for Advanced Photonics and Electronics asks in a recent paper in the journal, Nanotechnology .[1-2]

There are two methods available for production of a nanoscale device. The subtractive process, which Kelly calls the "top-down" manufacturing approach, involves the traditional semiconductor processing step of etching a substrate . The additive, or bottom-up approach, involves pick-and-place of useful elements onto a substrate. A tour de force example of the bottom-up approach is the use of an atomic force microscope to arrange atoms into a corporate logo. Kelly sees top-down as the only viable method for production of commercial nanoscale devices.

Kelly sees two limits in top-down manufacturing. The first is a 3 nm limit on creation of one-off structures. There are about fifty silicon atoms per cubic nanometer, so this 3 nm limit corresponds to a little more than a thousand atoms. Since most useful nanomaterials are binary compounds (e.g., ZnO ) this would be just a few hundred functional units, not all of which would be in the bulk crystal and exhibit the bulk crystal properties.

A silicon unit cell with silicon tetrahedra emphasized.

There are eight silicon atoms per unit cell.

(Image by Pieter Kuiper, via Wikimedia Commons ).

Kelly's second limit is 7 nm, or about 17,000 silicon atoms. This is the limit that he sees in manufacturing nano-sized arrays of acceptable yield to a particular device specification. It's not difficult to see why this would be true. The lattice constant of silicon is 543 pm, or about half a nanometer. When you're forming silicon patches of 7 nm size, one unit cell plane is about 8% of this dimension.

A good model system is an array of vertical pillars, and Kelly does a statistical analysis of such arrays that confirms our premonitions in the preceding paragraphs.

In summary, Kelly has a caveat for the nanotechnology community:

"If more work is devoted to the hard problem of understanding just what can be manufactured and how, at the expense of more studies of things that cannot be manufactured under the conditions of the present theorem, then that too is good for science and for technology."[2]

Obligatory "bottom-up" nanoscale photograph.
The IBM logo rendered with xenon atoms in 1990.
(Via IBM ).[3]

Reaching the Bottom

References: