It’s only got 178 transistors, but it’s an important proof-of-concept that’s poised to keep Moore’s Law right on track. The breakthrough, in which a basic computer was powered by microscopic chains of carbon atoms, means we may have finally found a viable alternative to silicon chips.
Back in 1965, Intel Corp. co-founder Gordon Moore famously predicted that the density of transistors would double about every two years, resulting in smaller, faster, and cheaper electronic devices. Trouble is, smaller and faster has resulted in whole lot of on/off transistor switching in increasingly smaller spaces, leading to intense heat dissipation.
No doubt. Today’s silicon-based laptops can get absolutely scorching at times, often making them impossible to use in the way they were literally intended. Eventually, these transistors could start to melt or burn electronic components. What’s more, it’s a tremendous waste of power. It’s a concern that’s plagued designers for years, leading them to worry that Moore’s Law may eventually come to an end.
Long Chains of Carbon Atoms
A ray of hope in all this, however, has been the potential for carbon nanotubes (CNT) — long chains of carbon atoms that are exceptionally efficient at conducting and controlling electricity. But they can also be fashioned into transistors within semiconductors.
And in fact, CNTs were first used as transistors 15 years ago — but engineers faced terrible issues when trying to make them work in the exact way needed. Specifically, the CNTs didn’t grow in accordance to the strict parallel lines required by engineers. In addition, depending on how they grew, some CNTs ended up behaving like metallic wires that perpetually conducted electricity instead of acting like proper semiconductors which can be switched off. This intermittent problem made the prospect of mass production a nightmare.
But scientists did not want to give up on CNTs. They’re amazing conductors. And because they’re so thin — thousands of chains can fit side-by-side in a human hair — they require a ridiculously small amount of energy to switch them off. Think of it as a garden hose; the thinner the hose, the less effort is needed to shut off the flow. Should CNTs be made to work, they could operate an order of magnitude in performance beyond silicon-based chips.
An Imperfection-immune Design
To address these problems, a Stanford team came up with a novel solution — a two-pronged approach they’re calling an “imperfection-immune design.”
To get rid of the wire-like nanotubes, the researchers switched off all the good CNTs. They then shot a burst of electricity into the semiconductor, which was collected in the metallic nanotubes. This caused them to grow so hot that they burned up and vaporized into tiny puffs of carbon dioxide. First part of the problem solved.
Then, to bypass the misaligned tubes, the researchers developed a sophisticated computer program capable of mapping out a circuit layout guaranteed to work no matter whether or where CNTs might be misaligned.
Faster, Energy Efficient, Cooler
With these problems solved, the research team, which was led by by Stanford professors Subhasish Mitra and H.S. Philip Wong, was able to fashion a basic computer with 178 transistors. The machine can perform tasks like counting and number sorting. It also supports a basic operating system that allows it to switch between these two tasks. It can also run MIPS, a commercial instruction set designed back in the 1980s, allowing it to run over 20 different instructions.
Though people have long considered CNTs a viable alternative to silicon transistors, this is the first proof that actually works — and it's the most complex carbon-based electronic system yet realized. It’s one of several recent and remarkable breakthroughs, including the ‘teleportation’ of information across an electronic circuit and the first quantum hub-and-spoke digital communications network.
Once perfected and scaled-up to industrial-scale levels, the Stanford approach could revolutionize the way electronics are designed and produced. It may even represent the next-generation of chip design. As a result, the ongoing miniaturization revolution will be allowed to continue, as will Moore’s Law. Future devices will continue to run at increasingly faster rates, require significantly less energy than silicon chips, and remain cool.