Slashdot Mirror
Ask Slashdot: How Would Room-Temp Superconductors Affect Us?
Bananatree3 writes "While we have sci-fi visions of room temperature superconductors like in the movie Avatar, the question still remains: How would the discovery of a such a material impact our everyday lives? How would the nature of warfare change? How would the global economy react? What are the cultural pros and cons of such a technological shift?" And just as important, in what contexts would you want to see it first employed?
By the standards of the physical universe, "room temperature" is pretty arbitrary. For a spacecraft, keeping superconductors cold is reasonably easy.
The most realistic answer, but not the one you want to hear, is: Nobody really knows.
If history teaches us one thing than it is that we are horrible at predicting the outcomes of anything major. In hindsight, we can "explain" things, but our predictions suck so badly, it's a surprise we haven't given up on the subject. And that's for both experts and non-experts.
Nobody came even close to predicting the impact of computers. Or electricity. People didn't think WW1 would become the slaughterhouse it did. There are refugees around the globe who are living in "temporary" shelters, waiting to return home because the conflict will surely be over any day now. Some of them have been waiting for a decade and more.
The real impact of this technology, as most, will most likely not be anything that anyone today predicts, but something that someone in the future comes up with that nobody thought of before. That includes the inventors. I don't think Graham Bell ever thought that "please turn off your mobile phones" would be a screen shown in these newfangled movie theatres that just came about in his time.
Assorted stuff I do sometimes: Lemuria.org
Warfare? Who'd go to war when they had a hoverboard at home?
The first use will be warfare as is always the case sadly. You'll probably first see rail- and coil-guns show up. Next you'll find its uses in radars and specifically in trying to make them useless. Then it will proceed into gimmicks for rich people. After that it'll go to civil scientists (space exploration, particle accelerators, ...) and maybe a few years later into people's houses. Somewhere in between all of that somebody might find a use for it in medicine (other than improving your standard NMRI).
OLED monitor floating in midair. pen floating in midair. FLUX PIN ALL THE THINGS
100% computational efficiency, 0% heat release
You can't do that. Any non-reversible computation causes an increase in entropy, and reversible computation is not particularly practical. Achieving practical reversible computation would be a leap at least as large as room temperature superconductors.
Finally! A year of moderation! Ready for 2019?
</pedantic>
http://skepticsplay.blogspot.com.au/2012/01/superconductors-picture-of-progress.html
For those wondering, the highest critical temperature as of 2012 is 135K. Room temperature is about 300K. So no, unobtanium hasn't been discovered yet.
If I have seen further it is by stealing the Intellectual Property of giants.
Maglevs comes to mind - you only once load the magnets along the track, and then they will keep the magnetic field forever.
Imagine roadrails along the interstates which keep the cars on track. Also the hover car will suddenly be feasible - as soon as the car moves forward, induction will load the magnets inside the car and let it hover along the supra conducting magnets in the road. You can see the effect already today at some science shows where they have supraconducting maglevs. Zero friction against the track, just air friction left. One can imagine subways with supracontucting tracks, which work with air pressure along the tubes.
Super strong magnets can be build, which you once load with electricity and which then keep the magnetism forever. Construction could get rid of glue and screws, just put the elements together, load the magnets once, and they will keep everything in shape. You could lock your house with magnetic bars, which once locked, keep tight until you unload the electricity from the bars and they open again.
You could store electricity in giant coils instead of chemical cells, making loading and unloading the electricity much faster, and enabling lots of non-constant electricity creators like windwheels and solar panels to work within a giant grid and finally overcome the problem of the electric base load.
Well, we currently only lose about 6-7% of the electric energy we generate to transmission losses.
6% of a trillion-dollar industry is "not all that much"?
Let's be realistic here. Like so many other technological advances, what's going to make it take off is SEX.
Levitating sex will sell, I have no doubts.
From there, the applications will (if you excuse the language) trickle down to more mundane uses. I'm sure there are lots of kitchen uses, for example. But sex first, cause that's where the money is.
Most people think of superconductors as merely a "perfectly efficient" conductor. While this is true, it just scratches the surface of what's possible with superconductors. Using superconductors just to improve efficiency wouldn't be that big a deal by itself. It would improve battery life a little bit, and maybe drop bulk electricity transmission overheads, but not by much, and certainly not immediately. Making most superconductors into high-tensile wire is a non-trivial exercise, even if cooling isn't a problem -- and it will be! Just because a material is discovered that can conduct at "room" temperature isn't helpful for wire outdoors in direct sunlight, or in a hot environment inside high-temperature machinery. Last but not least, superconductors have current and magnetic field limits that increase as they are cooled past the transition temperature. A superconductor with a transition temperature of 26C would probably have only a few limited applications above 20C.
The other uses are more interesting, and often more amenable to thermal control:
The Meissner_effect provides magnetic shielding, which is useful for all sorts of things, like amplifiers, or for protecting sensitive electronics. This is also what causes magnets to levitate above Type 2 superconductors. I assume that a room-temperature superconductor would be Type 2, so levitation would likely be possible.
The London moment could be used in gyroscopes and the like.
Josephson junctions provide all sorts of functions, like ultra-sensitive magnetic field sensors (think hard-drives and MRIs).
Still, all of that is a bit... meh. I mean sure, you get less noise in your now ultra-sensitive amplifier, and electricity will cost 10% less than it would have otherwise in 30 years. Is this life changing? Probably not really.
A much more interesting potential application than all of those combined is Rapid Single Flux Quantum digital circuitry. That stuff makes silicon look like vacuum tubes. Think 100GHz+, self-clocking, 1000x as efficient as CMOS, and manufacturable now, with only the cooling requirement the big down-side. If RSFQ could be made to work at room-temperature (or even near it), you could be looking at a sudden massive leap forward in computer power like never before. For example, with a power draw 1000x lower, it would be possible to stack every chip in a typical computer into a little "cube", with much shorter wire lengths, and hence, latencies. We can't do this now, because that cube would literally melt in seconds form the heat.
The reality-check of all this is that many MRI machines are still cooled by liquid helium, even though superconductors that work at liquid nitrogen temperatures have been available for a while. This tells you a lot about the limitations that might restrict the application of even a hypothetical room-temperature superconductor. For example, ultra-sensitive sensors and RSFQ may not work at all, because the tiny signal quanta may be swamped by the background thermal noise. Similarly, manufacturability of wire and maximum magnetic field strength is a key requirement for a lot of applications, like MRIs and electric motors.
Personally, I suspect that the first room-temperature superconductor will be initially manufacturable in bulk only as a thin-film, so expect the first decade or two to be mostly about improved circuitry and sensors more than anything else. This might be closer than people think. For example, there's a harmless quack who claims to have achieved superconductivity at 28C by manufacturing extremely complex copper-based crystals as a thin layer between two different traditional copper-based superconductors. Assuming for a second that he's onto something, it gives you an idea
It's rather tough to predict the impact of room temperature semi-conductors without knowing a lot more about the specifics of the technology.
For example, is the material suitable for long-haul power lines? Does it have the tensile strength to be deployed as multi-kilometer wiring? If it is, we can expect to see a dramatic improvement in the efficiency of power distribution, resulting in delays in the deployment of new power plants because the old ones would suddenly be delivering 10-20% more power to the home/business instead of losing it in the wiring.
Is the material suitable for fine wiring? If so, we may see some marginal improvements in the power drain of general electrical and electronic equipment.
No matter what happens in this field, we can expect that the military will be the first to apply the technology. They're really the only ones with the budget to become "early adopters" of such a shift in technology, other than research prototypes coming out of the likes of IBM.
All in all, though, I really wouldn't expect a very dramatic shift in power systems, though. Efficiency is great, but it rarely is an earth-shattering improvement.
Improving the efficiency of transmission doesn't change the speed of transmission, so it really wouldn't affect the raw computing horsepower of machines, just their power consumption. It's not like anyone has been talking about any superconductors that could replace the metal wiring layers on VLSI chips -- having a material and being able to vapour deposit or lithograph the material are two dramatically different technologies, and it could be decades after the discovery of the material before someone comes up with a practical way to use it on the microscopic scale of chips.
Personally I'm more interested in some of the "light switch" technologies that are being experimented with, because those technologies could change the fundamental physics of computing far more dramatically than reducing power consumption would.
I do not fail; I succeed at finding out what does not work.
And in order to achieve that, it's necessary to keep it quite rare for any large amount of electricity to be transmitted for long distances. With room-temperature superconductors, it's possible for electricity generated anywhere in the world to be used anywhere in the world. That's gonna make for a big change.
You really, really don't have any idea of superconductors are capable of, do you? It's lot more than just making electricity transmission more efficient.
First of all, mod+1 for the reference to the minimum amount of heat -- I knew that such a limit existed but it was good to see the estimate and have links to the formal argument and beyond. Second, while we may or may not be able to reduce the heat released from the bits themselves as they change state, room temperature superconductors will still make two very significant improvements in processor design. First, reducing the resistance of everything BUT the bits will reduce the heat released by a chip by a nontrivial amount, rather a nontrivial fraction -- presuming that one can lay down the superconductor in VLSI circuits and mass produce them, as opposed to build them a molecule at a time. Second, electrical superconductors are usually thermal superconductors as well.
It is this latter property that is probably by far the most important. Note e.g. this article: http://www.sciencedaily.com/releases/2003/11/031112072719.htm -- if one were able to make the base of a chip out of a superconductor in good thermal contact with the actual semiconductor matrix a thin film on top of it, and couple that base directly to a superconducting heat sink, one could e.g. produce 10x to 50x the heat in the actual CPU and still remove it fast enough to keep the chip itself sufficiently cool. If the traces within the chip itself were superconducting, if clever use of superconducting material let one reduce the heat associated with switching closer to the limit, so much the better. Ultimately, it would probably mean that one could run chips at higher voltage and higher clock to produce faster reliable switching and still deal with the heat.
I don't have time to do a formal estimate of the speedup possible, but I'm guestimating that a real thermal superconductor -- one with "zero" resistance to the flow of heat -- suitable for use as the base material for a chip would permit a very rapid scale-up of chip speed by up to an order of magnitude in clock or effective clock. It also might make it possible to build a three dimensional CPU -- one reason chips are 2D is so that one can get the heat out; if one had a thermal/electrical superconductor one could in principle stack up layers and scale performance by one or more orders of magnitude, at first multiple cores on steroids but all at much higher clocks, later true 3d design and layout.
In any event, the impact would very probably be profound, at least if the hypothetical RTS was cheap and suitable for nanoscale integration as a substrate and/or trace material (and functioned as a thermal superconductor as well as noted).
Still, I think that simply eliminating resistivity in power transmission would have the greatest societal impact. PV solar power, for example, "instantly" becomes feasible because one can generate in the Mojave and use the electricity in Maine without transmission loss. That isn't huge, that is game-changing enormous. The Sahara become the electrical source for Europe and Africa, India for Asia, etc. Depending on the hypothetical materials magnetic properties (big if, actually!) it may well revolutionize electrical motor design, maglev trains and roadways, and more, but just letting us move power for free to where we use it makes Edison have the last laugh over Tesla -- human civilization can convert to low voltage DC electrical service. A civilization run on 5 VDC would make electrocution a historical oddity from pre-RTS times -- one can manage to kill yourself with as little as 9 volts (see my favorite Darwin Award, "Resistance is Futile" -- http://www.darwinawards.com/darwin/darwin1999-50.html) but 50 mA should be below the fatal threshold even for somebody that tries very hard.
rgb
Even when the experts all agree, they may well be mistaken. --- Bertrand Russell.
Reversible computing is not that hard, you just have to use reversible operations. You will need an instruction to throw away data though to be Turing complete though, but at least it would make the non-reversible instruction very clear.
Almost all math operation can be written as a reversible operation by make the operation produce a result and remainder.
A + B : ADDSUB(A,B) => (A+B, A-B)
A * B: MULMOD(A,B) => (A*B, A # B)
etc.
Unfortunately, not, not for more recent definitions of modern. We lose a pretty fair amount of energy now through "insulators", or so I was told by a chip designer once. Ah, not quite insulators, but "off" MOSFETs: http://en.wikipedia.org/wiki/Subthreshold_leakage
So we don't need better conductors, we need better not-conductors.
The DC to DC conversions inside portable electronic devices would get a lot better. All the circuits inside the devices would be more efficient. It might enable some kind of better battery technology also. Room temp superconductor is not going to be able to be produced on a large scale at low price anytime soon after discovery, so transmission lines are the last thing on the list to get made.
to see some room-temperature semiconductors in the future. keeping the quantum computer in the living room has been a tough job lately. Ive found that while the couch and the area rug dont mind hovering below absolute-zero, the cats certainly dont appreciate it very much.
Good people go to bed earlier.
So how would you implement a hash function then? Or some iterative functions? Reversible does not work for anything remotely practical. Quite simply there are more inputs than outputs.
Of course we are so far away from the kT limit per bit that we don't really need it either. At least for a very long while.
If information wants to be free, why does my internet connection cost so much?
If you use reversible logic blocks you can still run non-reversible algorithms...as i understand it you could reverse it on the CPU that ran it forwards, but when you write the result to permanent storage or put it out on the network you throw out the extra information (wasting some energy) and it becomes no longer reversible.
length shrinks proportionally with the cross sectional area so nothing changes
All other things being equal, no it doesn't. The cross sectional area is two dimensional, whereas the length is one dimensional. This is really basic stuff. If you need an illustrative example, consider a bar with a square cross section. We will set the length of our example bar to ten times its other edges. We'll call the length of the cross-sectional square's edges n. So, the cross sectional area will be n^2 and the length will be 10n for any given n. ,90 so ratio is 9/10 ,80 so ratio is 4/5 ,70 so ratio is 7/10 ,60 so ratio is 3/5 ,50 so ratio is 1/2
For n=10, cross section is 100 and length is 100 so ratio is 1/1
For n=9, cross section is 81 and length is
For n=8, cross section is 64 and length is
For n=7, cross section is 49 and length is
For n=6, cross section is 36 and length is
For n=5, cross section is 25 and length is
For n=4, cross section is 16 and length is 40, so ratio is 2/5
For n=3, cross section is 9 and length is 30, so ratio is 3/10
For n=2, cross section is 4 and length is 20, so ratio is 1/5
For n=1, cross section is 1 and length is 10, so ratio is 1/10
So, by scaling down the length of the bar by a factor of 10, without altering the shape, the ratio of cross-section to length went from 1/1 to 1/10. Scaling down the length by another factor of ten would yield a ratio of 1/100 and so forth and so on.
The practical considerations for applications it ends up in depend tremendously on how much it costs. If this room temperature supercondictor costs more than the current cryogenic cooling of a conventional superconductor, because it's made from a super exotic material or requires a prohibitively expensive process to manufacture, it's not likely to displace it from most current applications, let alone get into many new ones. Of course that still depends on the price difference; If they're comparable you'll see some change over. Power companies would love it, but if the conductor costs significantly more than the percentage of power they are losing to resistive heating in a given section, it won't get changed. Chip applications may be a notable exception if it's not terribly expensive, but they have the additional consideration of manufacturing: if it can't be laid down on silicon in a process that is compatible with the current lithography, they are almost certainly going to stay in a niche market for a long time even if the bulk material is dirt cheap.
So folks can do the Glass half full thing and figure out places where it can be used, but without an answer to "How much does it cost" there is no way to predict the paramount information of where it *will* be used.
Even people that believe in pre-destiny look both ways before crossing the street.
This is an active area of research because at the quantum level everything is reversable. If the hardware implementation difficulties of quantum computers ever get solved, we need to have both theories and "practical" languages to handle it. (By "practical" I mean one that actually looks like programming versus "programming" in terms of Hilbert spaces.)
My understanding is that to implement a hash function you have one of two choices. The first is to pay the "kT" cost by calling the "erase" operator (i.e. pipe it to /dev/null). The second is to have it generate "garbage" bits. These bits provide enough information for the computation to be reversed and are not hard to define (even for something like hash functions). For example, with a hash function, the data from which the hash was computed can be used as the garbage. By carrying the garbage bits around instead of erasing them, you might be able to (1) use them in some other computation, (2) be able to localize where and when you pay the heat cost of the garbage, or (3) be able to cheaply backtract the computation.
If you are interested in this, James and Sabry ("Information Effects" POPL 2012, "The Two Dualities of Computation", etc.) are actively developing the foundational theories of a language for programming in a reversable language (disclaimer: the authors are both personal friends of mine). Their stuff might still be a bit heavy for Joe Programmer, but it should be accessable to anyone familiar with higher-order, typed languages.
The inefficiency of modern digital circuits comes from two things:
1. Leakage through gate oxides (insulators) and switched-off transistors (semiconductor action).
2. Charging and discharging transistor gates during turn-on and turn-off (capacitance).
Unfortunately, superconductors won't help with either of those. Even switching power supplies lose a lot of power through transistor switching and diode drops. For most electronic products, imperfect semiconductor devices are a bigger problem than imperfect conductors.
Visit the