## Decreasing Size Increases Power

How about speed of response? A bee moves its wings up and down far faster than we can flap our arms. The speed of a moving appendage may be about the same in the bee and us, but the bee's wing travels a much shorter distance. So in small systems the speed of walking and flying movements increases. Think of a mosquito's whine. Or a factory might perform ten steps a second, but fast enzymes operate about a million times each second. The density of power or power density measures power, force or strength, times speed, but strength is proportional to area. If speed is constant, a machine ten times larger can produce one hundred times as much power. But power per unit volume or power density remains unidimensional!

Think of it this way. If a system ten centimeters cubed creates a thousand watts of power, then an engine one centimeter cubed may produce only ten watts of power or one hundredth the power of the ten times larger engine. But are you ready for this? Suppose a thousand of the one centimeter cubed engines occupies the same volume as the one ten centimeter cubed engine, so now together, the smaller engines create ten thousand watts. So by building a thousand times as many machines and by making each machine ten times smaller, the same mass and volume can be designed to deliver ten times as much power.

Suppose we now consider frequency of operation. Frequency of operation increases as the size of our system decreases, so miniature engines may run at ten times the rate of smaller ones. When a design shrinks by a factor of ten, its number of parts increases by a factor of one thousand. This relationship is the functional density, and functional density remains proportional to our system's volume. We can pack in a million, million, million, or 1018 more nanoscale parts that are a million times smaller into the same volume.

Suppose we pack these parts into a bee's thorax. Shrinking by a factor of one hundred, as might be the difference between today's transistors and today's molecular electronics, allows us to confine a million times more circuitry within the same volume. But here's the rub. Suppose each additional component costs extra money, or our parts or machines have short lives, then taking a thousand times more parts to just increase our current performance ten times becomes non-economical.

But what if we might coerce bees into producing our parts using their evolutionarily honed, massively parallel, reliable and fault tolerant, processing designs, so that our parts, now made by contented bees, might last as long as a bee itself? It might now be worth attempting.

And lastly, let's consider efficiency. A large-scale system that is ninety percent efficient may grow well over ninety-nine point nine percent efficient if we shrink it into the nanoscale and reduce its speed to keep its power and functional density constant.

But what about friction? After all, isn't friction the bane of all machines, even nature's? That friction is proportional to force or area implies that frictional power must be proportional to what power is consumed, regardless of scale. Let's say the thickness of a protective cuticle that is available to erode through rubbing and attrition decreases as we shrink our bee system. What happens? Even when very thin, the covalent bonds of cuticle remain strong enough to resist forces between sliding surfaces that are smooth and clean, so frictional wear alone should not break these bonds, as rubbing will never generate enough heat or force to break cova-lent bonds. Remember I have said nothing about chemicals. Most systems will not shrink all the way down to the nanoscale, however, as problems having to do with how the system's parts connect, intervene.

0 0