Most breakthroughs in miniaturisation are important but boring; this substance can be stretched thinner than before, that manufacturing process is now 8 per cent cheaper. This has always been in pursuit of a day when enough fundamental nano-breakthroughs have come together from materials and manufacturing that we can start inventing whole machines on that scale. Nobody’s ever written a Star Trek episode about the world’s smallest microchip, only about the world’s smallest computer.
Now, a team from the University of Michigan has built not just a very small microchip, but a whole functioning computer, and it’s less than a cubic millimetre in size. It’s called the Michigan Micro Mote, or M3, and this tiny computer features processing, data storage, and wireless communication. Researcher Pabral Dutta thinks it will be the “next revolution in computing.”
These M3 units are, of course, very limited. They are about as simple as they can be, and for now they’re being designed for very simple sensory tasks; load one up with a detector, say for the fluid pressure in a person’s skull, and it might be able to monitor that value until it reaches some threshold, then alert its closest neighbour. (This is all hypothetical, as the concept has only just started animal testing).
The technology works with a very low-powered and low-range wireless standard to broadcast its latest state every few minutes. Presumably, if the signal is more than the usual “all clear,” it gets ferried on by its nearest brothers until finally reaching a base station with real computing power.
This implies that the chips are designed to necessarily work as a swarm, and indeed the term “smart dust” seems to have been a rallying cry for the researchers. Yet few of their proposed uses seem to fit well with the technology itself – will we have a swarm of M3 computers in our eye, or just one? Even if we have hundreds of the little monitors in various parts of the body, these millimetre-scale computers might struggle to communicate.
Additionally, as is always the case with extreme miniaturisation, power generation (and storage) has been a big problem. Wireless power is an obvious choice, but that technology is having efficiency problems even on the human scale, let alone the dust one. To power the M3, researchers fitted it with a tiny solar cell – but how much sunlight will be available while monitoring intracranial conditions? Passive power generation techniques, like MicroGen’s ambient vibration harvester known as Bolt (see the image below), have already been scaled down quite nicely, and their meagre energy output is well-suited to the equally tiny energy requirements of the new smart dust. And, since these are mostly being put forward for medical or at least bio-technological uses, body heat is another obvious potential source of energy.
Most possible applications require some faith that the micro-motes will improve over time. The researchers present their “swarmputers” as the next stage in overall processor evolution, from desktop to laptop to cell phone to bloodstream. And yet these computers are limited by both processing power and communication, relaying information to one another over very small windows, like deep space satellites. This does not seem to lend itself well to distributed computing.
Samsung’s proposed graphene micro-antennae might offer some hope here, but even these function only on the centimetre scale. This would work for a tightly clustered network of chips, but how many applications will actually see these things bound closely together?
By far the most plausible short-term application is in diagnosis. If we can inject a patient with a small swarm of computers (made to break down slowly in the blood, so we don’t clog any arteries) we could send them on all kinds of missions. They could flush down larger blood vessels to detect a blockage or circulate freely until they bind to some programmed target. Autonomous endobots are one of the holy grails of diagnostic medicine.
There are also possible non-medical uses, though for those we must get quite speculative indeed. Being so small, they could easily get airborne with even tiny little wings; imagine the sensor ball idea from Twister, but on the millimetre scale. Now the solar panel begins to make more sense, and the possible applications much easier to imagine. What if you could paint a laser security system onto a door frame, a transparent medium laced with thousands of tiny computers that randomly fire bursts of light and wait for them to bounce back off a body moving through the door?
One other possible use for these computers is to make their size a temporary measure and have them reassemble into a larger whole. This would be a good way to get complex computers through tight bottlenecks, medical or otherwise. Civil engineers might find it useful to flood a city’s sewer system with little surveyor computers which periodically meet up to connect and to pool their power in order to send a signal back to home base.
All these possible applications hinge on detection or monitoring, though, never on taking modern processing power and breaking it up. This does not seem to be a further step in the miniaturisation of computer power, but rather a new and independent path that could have benefits of its own.