Using peptides to assemble smaller computer circuits

Smaller computer chips have helped make computers faster and more efficient. Think about why an older computer can seem slow and outdated next to a newer one: improving the components changes what the whole machine can do. The opening establishes why continued miniaturization matters, then introduces a problem. The methods being used are approaching a size limit.

Read the date and qualification carefully. This is an article from 2000 predicting a wall around 2010. More importantly, the limit applies to properly functioning transistors made with current methods. The claim isn't that nothing in nature can be smaller than 25 nanometers. Living cells already produce complex structures below that size.

Could researchers borrow those biological construction methods? That's the possibility driving the account. Rather than keep shrinking a transistor made the usual way, perhaps they could use molecules that perform its function or help assemble a circuit from other materials. Other research focuses on DNA; Angela Belcher and Evelyn Hu are investigating peptides, which are shorter amino acid chains than proteins. Keep their particular project in view as the discussion narrows.

The clue comes from Belcher's work on abalone shells. A specific peptide causes calcium carbonate to crystallize into the structure of the tough shell. Why would a computer researcher care about that? Because the peptide helps control how another material forms. If a peptide could direct the growth of the semiconductor materials used in transistors, it might become a tiny construction tool.

That is the connection to chip size. The promise isn't merely a larger supply of semiconductor material. It is a way to organize material precisely at the scale where biological molecules operate. The shell finding supplies a model for what the researchers hope peptides can do with circuit materials.

There is a substantial missing step: no known peptide does the needed job with semiconductors. So the researchers grow a random assortment of a billion different peptides and screen them against three semiconductor materials. Instead of assuming the shell peptide will work, they look for candidates suited to the new materials.

Some bind, and their selectivity is striking. A peptide attaches exclusively to one of the tested crystals, even to a particular face of that crystal. That's more useful for controlled construction than a molecule that sticks indiscriminately. The researchers then use the promising candidates to develop related peptides through a process resembling accelerated evolution: start with useful characteristics and develop further candidates from them.

The final paragraph tells us why a few successful binders aren't yet a complete circuit-building method. A circuit requires a set of tools suited to different component materials. The team expands the search to 20 more semiconductors, and its collection of useful peptides grows into the hundreds.

Now consider the new two-crystal peptides. If a peptide can bind to two different crystals at once, it can act like a tiny daub of glue between them. That gives us a concrete picture of the required finesse: the right molecule has to connect the right materials in the right way. Simply finding a peptide that sticks to something isn't the entire goal.

The project remains a work in progress. The researchers have found promising molecular tools and are designing more, but the account doesn't announce finished, commercially available self-assembling circuits. Our main point should preserve both the promise and that qualification: peptide research may provide a way to build chips below the size limit imposed by existing methods.