Field notes on things that run themselves
Nothing Is Spent to Keep It Moving
Start a current flowing around a closed loop of superconducting wire, take away whatever started it, and watch. No battery, no gradient, nothing pushing the charge forward — and it does not slow down, not over an hour, not over a year. As far as anyone has measured, it does not decay. Every system this publication has covered holds its shape by spending something, continuously, to stay standing. This one is the deliberate exception: a loop built to need nothing spent to keep it moving. Except that isn’t quite true.
Superconductivity was discovered by the Dutch physicist Heike Kamerlingh Onnes in 1911, when he cooled a thread of mercury below about 4 kelvin and watched its electrical resistance vanish completely. The persistent current followed almost at once, the discovery’s strangest demonstration: induce a current in a closed superconducting loop, remove whatever induced it, and the loop needs nothing further. No resistance means no energy lost to heat on each pass, so nothing degrades the flow, cycle after cycle, in principle forever.
In principle is not measured, so in 1963 physicists J. File and R. G. Mills, at Princeton’s Plasma Physics Laboratory, watched one for as long as they could. They started a current in a superconducting ring and used nuclear magnetic resonance, the physics behind an MRI scan, to track its field over stretches of 21 and then 37 days. If the current were decaying by ordinary resistance, however small, that exponential signature would have shown up. It didn’t. Their result, and the similar experiments it opened, is why physicists today can honestly say only that a persistent current’s decay time is bounded at longer than roughly 100,000 years — not infinite, just bounded from below by however long anyone has been patient enough to watch.
That is nearly the cleanest possible story, and the cleanest is rarely the whole one. File and Mills didn’t just report a null result; their paper tries to correlate the faint signal they saw against flux creep, a real process in practical wire, which carries its current alongside bundles of magnetic flux pinned by microscopic defects in the metal. Pin every flux line perfectly and the current is permanently loss-free. But thermal energy can occasionally jostle one loose, letting it hop to a neighboring site and release energy on the way. Enough of that, at random, and the trapped field relaxes — not exponentially, the signature of resistance, but logarithmically, the signature of thermal creep, worked out by Anderson and Kim in the early 1960s. The honest answer isn’t simply zero; it may be flux quietly resettling, while the resistance stays indistinguishable from nothing.
Here is the part worth paying for. The current may cost nothing to sustain, but the condition that allows it — being a superconductor at all — is not free. The wire must stay below its critical temperature; for the niobium-titanium in most hospital MRI magnets, that means below about 9 kelvin, run in practice at 4.2 kelvin, liquid helium’s boiling point. Once an MRI magnet is charged, technicians open a switch, disconnect the power supply entirely, and close the loop on itself. The current keeps circulating on its own for years, sometimes over a decade, without being fed again. Everything paid for afterward — cryocoolers running around the clock, periodic helium top-offs — has nothing to do with keeping the current moving. It is the cost of keeping the wire cold enough to let a current that needs no help keep helping itself.
That arrangement has a hard edge. If the cooling fails, or a small disturbance dumps energy into even a sliver of wire, that sliver briefly stops superconducting and becomes an ordinary resistor with an enormous current forced through it. It heats, and that heat pushes the resistive zone into its neighbors faster than plain conduction would, since current starts diverting into a resistive backup layer built into the wire for this moment. The loop can turn resistive in milliseconds, and everything it was storing comes out as heat at once — a single large accelerator dipole can hold some 7 megajoules, enough, dumped that fast, to melt roughly ten kilograms of copper. Engineers call this a quench: not a slow running-down, but losing, all at once, the cold that let the current not have to.
This publication’s chemiosmotic gradient (No. 36) makes an instructive foil. A mitochondrion spends real energy, every second, pumping protons uphill to hold a voltage across its membrane — stop pumping and the charge drains away within moments, because holding that flow open costs something continuously. A persistent current inverts the arrangement: the flow itself asks for nothing, ever, once started; everything around it — the temperature, not the motion — must be paid for continuously, or lost all at once. The throughput this publication keeps finding did not disappear here. It moved: off the current, onto the cold that lets it stop needing anything at all.
One loop I’m watching
Next: a cave pearl. In certain limestone caves, a dripping pool agitates a grain of sediment on the pool floor just enough to keep it tumbling instead of cementing in place, and mineral-saturated water deposits calcite around it in concentric layers — the same process that builds a stalagmite, turned sideways and shared out over a moving target instead of a fixed point. Stop the drip, or let the grain settle for too long, and it stops forming a sphere. It just cements to the floor and becomes an ordinary, irregular lump of stone.
Tip: the ← and → arrow keys move between issues.
Prefer feeds? Follow by RSS or JSON Feed.
Want something playable between issues? Thought Toys has its own email list and RSS feed.
New to The Standing Wave? Start here →