Slashdot Mirror
Liquid Metal Capsules Used To Make Self-Healing Electronics
MrSeb writes "A crack team of engineers at the University of Illinois has developed an electronic circuit that autonomously self-heals when its metal wires are broken. This self-healing system restores conductivity within 'mere microseconds,' which is apparently fast enough that operation can continue without interruption. The self-healing mechanism is delightfully simple: The engineers place a bunch of 10-micron (0.01mm) microcapsules along the length of a circuit. The microcapsules are full of liquid metal, a gallium-indium alloy, and if the circuit underneath cracks, so do the microcapsules (90% of the time, anyway — the tech isn't perfect yet!). The liquid metal oozes into the circuit board, restoring up to 99% conductivity, and everything continues as normal. This even works with multi-layer printed circuit boards (PCBs), such the motherboard in your computer, too. There's no word on whether this same technology could one day be used by Terminators to self-heal shotgun blasts to the face, but it certainly sounds quite similar. The immediate use-cases are in extreme environments (aerospace), and batteries (which can't be taken apart to fix), but long term we might one day buy motherboards with these self-healing microcapsules built in."
I don't know if I'd want to be on a crack team. I'm more of a coke team kind of guy.
Having to work for a living is the root of all evil.
A crack engineering unit was sent to prison by a military court for a crime they didn't commit. These men promptly escaped from a maximum security stockade to the Los Angeles underground. Today, still wanted by the government, they survive as soldiers of fortune. If your circuits have a problem, if no one else can help, and if you can find them, maybe you can hire... The A-Team.
I think they'd probably design the circuits so that there wasn't enough of the liquid metal to reach the next wire over. That being said, this probably only works under normal gravity, so what you're suggesting might be something of an issue in space.
Bio questions? Ask me to start a Q&A journal. Computer analogies available for most topics!
I know it would be an alloy... but Gallium isn't such a great thing to be shipping around in airplanes, etc..watch this youtube video of gallium eating an aluminum can for an idea why.
The liquid metal oozes
Sounds a lot like gravity is the main mechanism for deploying the liquid, in which case any circuit that is not facing "Up" cannot utilize this technology otherwise the liquid will just pour whichever direction is down, which is not always toward the circuit... Or am I just understanding this concept incorrectly?
I once worked for a company that tried to get something like this to work. Wetting was a major problem. PCB traces are prone to oxidation anyway, and if they are in buried layers then they are prone to surface contamination from the epoxy. Although in theory cracks should be clean surfaces, the GaIn has to get there in the first place, and in doing so its own surface may be contaminated. Even a very thin layer of oxide or an organic monolayer may well be enough to prevent wetting. I suspect that this will succeed up to the point they try to make it work successfully in real circuit boards.
From scarped cliff or quarried stone she cries "A thousand types are gone, I care for nothing, no not one."
What happens when it breaks a second time? Then it's just as broken.
This kind of thing may help resist a sudden, one-time shock, but it won't do a thing to protect electronics from ongoing wear. Perhaps if there were a way of notifying the device that it had been broken so that it could quickly inform the user and void its own warranty then it would be more useful.
EMPs are a greatly overstated risk, and science does not back Hollywood. There's a video of an actual upper atmospheric detonation of a nuclear weapon, that shows some LLNL physicists on a beach eating hot dogs and steaks. The nuke detonates and temporarily interrupts the transistor radio that's playing, and then it starts working again a few seconds later. No vacuum tubes required.
The only "EMP weapons" that have done anything require direct conductivity (think Tazer). It's a non-issue.
And that's why this is probably useless for consumer grade electronics.
I mean really - how often do you break TRACES in a motherboard or PCB in any home consumer product? I haven't ever seen a failure like that get out of QC. The things that kill consumer electronics are corrosion, solder point failure (usually from overpressured heatsinks or heat based warping, see RROD), bad/exploding capacitors, and the occasional power surge or ESD damage.
MAYBE in aeronautics? Maybe maybe MAYBE in automobiles, if you have a PCB somewhere controlling a multifuel system. But for consumer grade home electronics? Not remotely necessary.
Considering the incredible marketing effort and designed failure in consumer electronics to always buy new crap, I really have to wonder if the average consumer electronic would survive long enough to need this technology.
I have a Number Nine 128 video card still working on an old P4 server. That's damn near 20 years old I think. No cracks in the PCB on that.
If I have motherboards that are 3,5,10 years old still working just fine and I fail to see the point of this technology in consumer products. Military and harsh environments certainly.
The article states this technology is intended to automatically repair integrated circuits via "microcapsules, as small as 10 microns in diameter". Being charitable and going with 90 nm geometries (which we still used in our company last year - we are a bit slow) that's too large by a factor of 100. Interesting for PCBs, but not for integrated circuits.
The article also states that the technology would fix things "so fast that the user never knew there was a problem" and then explains that "a failure interrupts current for mere microseconds".
The summary corrupts that somewhat into the claim that "operation can continue without interruption". It's far too slow for that. Let's assume a rather slow 33 MHz bus - that gives us a clock period of 30 ns - so we'd miss at least 33 clock cycles in this scenario. This interruption might not be noticed by the user, if an error correcting protocol is used on the bus and the system retransmits. Otherwise you would get wrong data, and you have to assume that will be noticed sooner or later.
Interesting technology on PCBs or communication wires, I could see it being used in safety-critical applications. On integrated circuits it doesn't seem feasible. Basically you make the transistors and wires on ICs already as small as you can. To repair the wires on the IC you now need to insert capsules into the wires to do the automatic repair - so they would be way smaller than the wires. If you could manufacture these structures you'd make the wires smaller though and then you'd lose your ability to insert the microcapsules ... there is no way to win that race.
I recall a similar idea about a "self-healing" plastic that had microspheres with chemicals that would form some new plastic when they broke. These material advances are cool in that they make materials that are more durable and can last longer before ultimate failure. In normal consumer electronic usage, this material is not very useful like you say. Consumer electronic internals aren't subjected to a lot of physical stress like bending or shear forces and therefore are not a major fail point. In the testing of the material, they most likely happened to find that it heals fast enough to not affect electronic circuits, but I doubt that was their original intention.
I think this sort of material that can heal itself is best used where an object gets physically damaged or worn down over time, not in electronic circuits. Imagine an engine block head that can stop a crack before it gets anywhere. To me, that is a much better use for healing metal alloys.
Authorities who disagree with you include:
The Encyclopedia Britannica
The Cambridge Advanced Learner's Dictionary & Thesaurus
The US National Institute of Standards and Technology
and about 16.5 million other hits on Google.
For some reason, having the homonyms ton/tonne variously refer to a short ton (907.18474 kg), a tonne (1000 kg), or long ton (1,016.0469088 kg a.k.a. English ton) vexes some people. They prefer to specify a "metric ton" rather than so overemphasize "tonne" that they sound as if they have a speech impediment.
The unit of measure exists by virtue of its pervasive use. The fact that you prefer an alternate equivalent does nothing to change that fact.
traces don't break. they suffer from electromigration. I.e., where the constant collision of electrons with the metal lattice eventually creates voids in the metal. Becomes more of a problem with higher power processors and narrower conductors. some metals are more susceptible as well. (aluminum more than copper, i think).
And similarly, they would get hot (due to the high current density in the near break) before they break, and this heat could trigger the liquid metal release. There are applications for high-reliability electronics. I think the automotive sector is the one that most easily comes to mind for the consumer market. Long use equiment, like medical equipment maybe too.
Also, don't forget, the equipment you have is designed to operate as long as necessary without the types of failures this would solve. Given this tool, could they be designed differently? More efficiently? Smaller? Maybe.
NO, EMP only destroys semiconductors. Won't bother resistors, coils, capacitors, or vaccuum tubes. If you want an EMP-proof circut, use tubes rather than semiconductors and you're good to go.
Free Martian Whores!
When you have something like a telecommunications satellite that costs $250 million and has to last 15+ years without maintenance, you aren't looking at the cost of materials for making micro capsules.
You are paying upwards of $100 million / ton for the whole thing anyhow.
The SI unit that equals 1000 Kg is a megagram (Mg, or 10^6 grams). The tonne is not an SI unit, but, in a fit of nostalgia, has been metricized and accepted for use with the SI system.
Agreed 100%. Its highly unusual for a PCB to fail, 90% of the time it's been bad solder joints or bad caps which can then escalate into other problems. Solder joints go bad due to heat or vibration or just being poorly soldered in the first place.
This problem is going to get much worse before it gets any better. lead based solders help prevent joint cracking and they're now illegal in the EU. As a result all new electronics use lead-free formulations. This means more heat/vibration related failures than ever, all because more politicians demanded we 'think of the children!'(tm)
And that's why this is probably useless for consumer grade electronics.
I mean really - how often do you break TRACES in a motherboard or PCB in any home consumer product? I haven't ever seen a failure like that get out of QC. The things that kill consumer electronics are corrosion, solder point failure (usually from overpressured heatsinks or heat based warping, see RROD), bad/exploding capacitors, and the occasional power surge or ESD damage.
MAYBE in aeronautics? Maybe maybe MAYBE in automobiles, if you have a PCB somewhere controlling a multifuel system. But for consumer grade home electronics? Not remotely necessary.
I don't know.....if they could find a way to apply this to BGA chips......
Seriously.... the modern BGA package was the stupidest cost cutting measure in history that has caused the average laptop to last maybe 20% as long as laptops made 10 years ago. I doubt Taiwanese 6-yr-olds in the sweatshop X-Ray every board and make sure the solder balls are perfectly uniform.
I want a REAL computer again instead of a disposable consumer entertainment devices. But since the consumer market is so large, pro users and hobbyists with a clue are not at all their target market anymore despite the fact that without the demanding, picky, quality-conscious geeks the home computer would have NEVER taken off. PERIOD.