Slashdot Mirror
Ultra Efficient Chip Cooling Passes Boeing Tests
joelgrimes writes "A company called Cool Chips plc is showing off a cooling device that claims unbelievable efficiencies using what they call 'quantum mechanical electron tunneling'. A choice quote from their press release: "A panel of Cool Chips one inch square will provide enough cooling for a refrigerator; a panel about two inches square will have the capacity to provide the air conditioning for a living room". They also mention using them to cool microprocessors. I used to think this company was nuts, but Boeing is making me think twice. Oh, and by the way, they work in reverse to make electricity from heat. Should I sell my baseball cards and buy their stock now, or can an army of slashdotters poke holes in their claims?" Fascinating stuff. Makes peltier coolers look pretty old school. In the press release they claim up to 80% efficiency, compared to 5-8% for peltier coolers and 50% for conventional refrigeration. I will say the cool chips corporate logo is baffling, though.
Here's the Google cache for those too lazy to find it themselves.
COOL CHIPS DISCLOSES APPLICATION OF QUANTUM MECHANICS IN HIGH-EFFICIENCY NANOTECH COOLING DEVICES
Refers to COLCF and BOREF
Cool Chips plc
Gibraltar
14 May 2002
Cool Chips plc (COLCF) said that its Cool Chips(tm), wafer-thin discs designed to produce cooling or refrigeration more efficiently than any competing technology, use quantum mechanical electron tunneling as the primary cooling mechanism. The Cool Chip is one of the first transformative technologies to emerge from the nanotechnology revolution.
The Cool Chip technology could eventually replace nearly every existing form of cooling, air conditioning, and thermal management. Prototype devices are being shown publicly for the first time at the Nanotech Planet Conference in San Jose, California, that begins today. The company has not previously disclosed the full scientific basis for its technology.
Because of the inherent advantages in cooling across a gap using electron tunneling, Cool Chips are projected to attain efficiencies much higher than those previously available in cooling systems, and they are much less than 10% of the size and weight of compressors. Cool Chips are modular, and can be packaged in arrays to cool virtually any size heat load.
The company expects its Cool Chip(tm) technology, which has been in development since 1994, to replace all thermoelectrics and compressors for cooling, in applications ranging from electronics and infrared sensors, to computer components, refrigeration, and air conditioning. Cool Chips are on target to have an overwhelming cost advantage.
Cool Chips will enable many new and improved consumer products. They will enable laptops to run cooler, for example, and make possible in-car soda and grocery coolers. A panel of Cool Chips one inch square will provide enough cooling for a refrigerator; a panel about two inches square will have the capacity to provide the air conditioning for a living room; and a panel about five inches square will supply enough cooling power to cool an entire house.
Most existing cooling systems use compressors and environment-damaging fluids and are 40-50% efficient. Smaller thermoelectric cooling devices, despite more than $1 billion spent on research, are only 8% efficient. Cool Chips are projected to operate at 70-80% of the maximum theoretical efficiency (Carnot) for cooling.
Cool Chips prototypes are small electronic devices similar in appearance to computer chips. When an electric current is applied, one side of the chip will become cold and the other side hot, as electrons "tunnel" across a 1-to-10 nanometre gap separating the two sides, carrying heat with them. Innate device advantages include high efficiency, solid-state design, silent operation, environmentally friendly materials and operation, and compact size for easy integration.
"We have demonstrated the capability to make multiple prototypes that show a tunneling current in excess of 10 amps, using a wafer area approximately 9 square cm in area," said Isaiah Cox, Cool Chips' president. "This is, by far, the largest tunneling current that has ever been reported across a gap, and we expect Cool Chips to make the first use of this quantum tunneling effect in a primary commercial application."
The tunneling current can be harnessed to provide cooling of very high density. The theoretical heat flux for flat electrodes suspended 50 Angstroms from each other is on the order of 5000 watts per square centimetre. Cool Chips(tm) will be more than adequate for cooling the next generation of microprocessors, which will produce upwards of 100 watts of heat per square centimetre.
Cool Chips are currently in development, and it is expected to take over a year to complete prototypes which demonstrate high output and efficiency. Current prototypes are being used to increase the quantum tunneling, and cooling has not been directly measured to date. Once the tunneling output has been increased to a certain level, our scientists intend to begin increasing cooling output.
An IV curve and other information is now available on the Cool Chips website at http://www.coolchips.gi.
The Cool Chips technology is protected by an extensive patent portfolio. This coverage extends to include a broad array of techniques related to this unique thermal management system, which offers solutions for nearly any thermal management application.
Cool Chips plc, based in Gibraltar, is a majority-owned subsidiary of Borealis Exploration Limited (BOREF) and has 7,281,785 shares outstanding. Borealis' business is reinventing the core technologies used by basic industries, including electric motors, steelmaking, electrical power generation, and cooling and thermal management.
For further information contact:
Chris Bourne
Director of Public Relations
Cool Chips plc
+44 20 8571 5216
pr@coolchips.gi
Forward Looking Statement at http://www.coolchips.gi/fwdlook.shtml
I do not read or respond to AC's. If you want a discussion, log in. Otherwise, don't waste your time.
From http://www.borealis.com/technology/patents.shtml:
Patent 5981866(StampPE)
PROCESS FOR STAMPABLE PHOTOELECTRIC GENERATOR
Abstract
Manufacture of a photoelectric converter by a photolithographic or stamping process prior to coating with a photoelectrically emissive material is described. This gives an economic and simple means of mass-producing photoelectric converter cells, and in one aspect is analogous to that used for pressing optical discs.
which is getting /.ed.
t ml
http://www.coolchips.com/technology/overview.sh
What is Cooling with Electrons?
"Hot" and "cold" are words we use to describe the presence (or absence) of heat. Heat is best described as energy contained within something else. So a cup of hot coffee has more energy than that same cup an hour later, after much of the heat has dissipated.
The energy which makes up "heat" is the kinetic energy of the atoms which carry the heat. So if the atoms in the cup of coffee are very active, the coffee is "hot". If the atoms become less active, the coffee is "cold". And if the atoms get cold enough so that the atoms are no longer in a fluid form, the coffee freezes into a solid.
While atoms in a solid themselves tend to be pretty immobile, the sub-atomic particles within them are always moving. At any temperature above absolute zero, electrons are constantly in motion, spinning around the atom, but also (especially in metals) swapping places with the electrons of surrounding atoms.
Of course, some electrons have high energy, while some electrons have low energy. The low energy electrons are cold, while the high energy electrons are hot.
Cooling with electrons involves encouraging the high energy electrons to escape, bringing in low energy electrons to replace them. It is analogous to removing the loudest people from a party: the party gets quieter.
What makes Cool Chips special?
There are other technologies which use electron migration to reduce heat. These fall under the rubric of "thermoelectrics". These technologies all use special materials and geometries to move the hottest electrons to one side, keeping the coldest electrons at the other.
The biggest problem with thermoelectrics is that while electrons are used to carry heat in one direction, the material itself returns most of that heat through conduction!
Cool Chips are special because the electrons move across a gap -- and that gap, since it is not a solid, is an excellent insulator. Once heat is trapped on one side, it cannot easily return.
How do we get the electrons to move across the gap?
The difficulty in getting lots of electrons to flow across a gap is that electrons do not naturally leave their atoms to go into space. Electrons do jump around a lot (it is called tunneling), but those jumps are pretty short, from one to ten nanometers, or just a few billionths of a meter long.
Researchers at Cool Chips plc have figured out how to get two materials very close to each other so that electrons can tunnel from one material to the next, carrying their heat with them. With the addition of a voltage bias, which encourages the electrons to move in a given direction, the heat is then transferred from one side to the other. And because there is a gap between the two materials, the heat cannot simply flow back!
Why hasn't this been done before?
Thermotunneling has not been done before because nobody imagined that it was possible to get large surfaces areas close to each other without making occasional contact. Cool Chips' scientists not only imagined a way to do it, but we have accomplished this goal and are currently refining our patented process.
Once these devices become commercially available, they will not only revolutionize the industries of refrigeration and cooling, but all of those industries that depend on them.
Cooling not yet measured? So, the device works in theory, but there might be an unanticipated roadblock ahead which significantly delays or hinders their ability to produce devices that actually cool something. :/
Justin
"Why would God give us a waist if we wasn't supposed to rest our pants on it?" - Rev. Roy McDaniels
No, it isn't that sinister a concept.
Right of first refusal means that you get a chance to buy something before anyone else does. It is the business equivalent to the concept of having "dibs" on something.
It also does not represent too great a risk on Boeing's part. They aren't obligated to buy this technology. They just have the chance to buy it before anyone else does. While they are certainly paying for this privilege in some manner (maybe the press release is the payment), they aren't jumping in with both feet.
It means that Boeing has secured the right that, if any other company tenders an offer for (exclusive?) rights to the technology, Boeing is guaranteed the right to match or beat the offer before the deal is made.
No, it is NOT restricted to the medium they are using. electrons can tunnel through anything, because they are borrowing some needed energy (to get through the barrier) from time, because delta t*E = h_bar...so since the uncertainty in time * energy cannot be less, electrons can borrow some energy and give it back later. The bigger the potential, clearly the less likely it is they will tunnel. also, when u speak of vibration at atomic scale, it is more the OTHER way around: particles at the atomic scale act like waves...that is really the ONLY reason there is quantum tunneling at all!
QED
BSD is for people who love UNIX. Linux is for those who hate Microsoft.
Actually, from the press release, Boeing isn't backing it-- yet. They just said thqt they were in testing, and notice the lack of results or data.
Without data from even a not-so-independent reviewer like Boeing (not so independent since they seem to have some financial interests in the company), I'm far from thinking this is close to reality.
I demand a million helicopters and a DOLLAR!
If the heatsink on the hot side of the coolchip isn't radiating as much heat as the CPU is producing then (assuming the coolchips heat pumping properties work) the hot side of the coolchip will keep getting hotter until the radiation of the heatsink matches the heat output by the CPU. You argument would work if the coolchip was just an excellent conductor of heat, but it's a heat pump - it can shift heat from a side that is cool to a side that is hotter than the side the heat came from.
This is what heatsinks are for, a 1 inch cube heatsink can have a huge surface area (which air is then blown through), and there's no reason to stick to one cubic inch, the heatsync can be much larger than the coolchip provided it can conduct the heat sufficently to all it's tiny fins. If two coolchips can actually do the heat pumping work of an air conditioner, then transferring that into the actual air should be no trickier than with conventional aircon units.
They claim to have invented a highly efficient (~80%) Peltier device ("CoolChip") using "quantum electronic tunneling" across a near-perfect temperature insulating "gap" of nanoscale width. They claim heat-transfer capabilities on the order of 500w/cm**2 (theoretically, but there aren't any _measurements_ yet, that they have, er... published).
It's difficult to attack these claims, simply because they haven't _explained_ the physics or materials or construction beyond trendy buzzwords and, by the way, they seem not to have actually _built_ any devices. This is typical of bunco artists hyping seemingly wonderful new technology. See all the "zero point energy" hucksters, for example.
However, a little common-sense physics is enough to demolish this scam. I'd like to hear their answers to the following questions and objections. But, I bet they won't do it.
There is no such thing as a near-perfect (or even really good) temperature insulating solid material - the only pretty good temperature insulation is... a vacuum. Any decent vacuum over a nano-scale gap is going to close the gap, real quick (especially if there is the strong electroforce attraction between negative and positive semiconductors helping); that's Strike One.
Such a Peltier-like device has to work by pumping electrons into the cold side and removing them from the hot side. But injecting electrons into the cold side _excites_ the existing n-doped semiconductor's electron-states, and it's only the rapid migration of those excited electrons away from that layer that removes heat (and the device has to pull away unbound electrons marginally faster than they are injected to provide cooling). It's impossible to extract more electrons than are added without entirely stripping the substrate eventually, and long before that happened you'd see _reverse_ tunneling of electrons into the very depleted cold substrate; here's Strike Two.
Then there's the claimed energy transfer. At the rate of 500w/cm**2, the hot substrate is going to start generating _photons_ (which have no charge, so they're not going to be bashful about moving _back_ across the "insulating" gap) and they will carry... heat; ergo, Strike Three.
Sure, "Any sufficiently advanced technology is indistinguishable from magic" [A.C. Clarke], and great technological leaps are desireable. But the only "magic" these people have in mind is moving significant amounts of money from scientifically naive, greedy, and gullible investors into their own pockets. But, it were ever thus: caveat emptor.
70-80% of the maximum theoretical efficiency (Carnot) for cooling.
Just so everyone is aware, in thermodynamics the Carnot engine is not a 100% efficient engine. Actually, depending on a few variables, the carnot engine can be incredibly inefficient.
Stating that the efficiency is 70-80% of the maximum theoretical efficient (Carnot engine) for cooling doesn't mean that much, since it doesn't fit the equation we all think about.
70-80% = Energy Out/Energy In
Instead we get
70-80% = Energy Out/(Energy In * Carnot Efficiency)
Since |Carnot Efficiency| 1, we end up with a artificial increase in the actual efficiency of the engine.
I would personally like to see the results of the actual efficiency, not this skewed statistic.
.....Marvin Mouse.....
(Math, CS, Physics, Psychology Undergrad)
~ kjrose
The person is right, you need to sink the heat away, but so do you with ANY cooling device, the heat dissipates indoors for a fridge, outdoors for AC.
The idea is you wind up with the same amount of heat, just rearranged. Peltiers have a bimetallic junction, which acts as a diode of sorts. A diode will cool on one side, and re-emit the heat at the other. Small effect, but get enough surface area and you have something. The battery op coolers have these things. How they actually work is any dissimilar junction electron needs to overcome a barrier, and the energy it uses normally comes from the voltage applied. Instead, peltiers use the raw thermal energy at the junction gap(the - side) to go to the + side, picking up energy. When they recombine at + side, energy is released as heat.
Big drawback is most junctions don't take kindly to being heated on either side, so you need to sink them quick.
This seems to be way more efficient, either not allowing an electron to 'get lucky' and jump over an impurity, or has to continually pick up more thermal energy on the way over, or has a ridiculously efficient manufacture process at nanoscale, and can afford to get just the right material thickness(1-10nM)
my 2c adjusted for inflation...
This mind intentionally left blank.
The KKK a bunch of sheetheads? You decide!
require extraordinary wads of cash money.
As experience has shown - suckering a major company with X does not mean X is true.
That said, actually, I believe this could work. The "efficiency" claim, however, is somewhat bogus. Quoth their webpage:
to a projected 70-80% of the maximum (Carnot) theoretical efficiency for heat pumps. Conventional refrigerators operate at up to 50% efficiency and current thermoelectric systems (Peltier Effect) operate at 5-8% efficiency.
The Carnot efficiency is not 100%; it is (Th-Tc)/Th x 100%. Th is the temp of "hot" half of the engine cycle and Tc is the cold. Both are in kelvin. So, if your car engine runs at 400K (boiling water) on the compression stroke and 300K (freezing water) on the expansion stroke the maximum efficiency you can theoretically get is 25%.
Now, they seem to be comparing the percentage-of-theoretical efficiency that their device gets with the actual efficiency of other devices. The upshot is that I believe refrigerators also run at about 80-90% of the Carnot efficiency, which is 50% actual efficiency, but I might be making a mistake.
I suppose this maps somehow to a total kinetic energy operator for the individual electrons they are moving (1 minute chemistry - heat is "thermal motion", the degree to which particles are bouncing around. Every "observable" feature of a particle - position, kinetic energy, momentum, and so on - is actually "random", and is related to the "wave function" of the particle, which is a function that tells you the probability of finding the particle at any given position, by an operator, the position operator is the number 1, which is itself a function that maps from a set of algebraic functions to a set of algebraic functions. The math for these operators is hoary as all hell, not analytically soluble, and they can generally only be dealt with pproximately/computationally.)
Clearly - and I'm talking about the second law of thermodynamics, here - they can't actually convert environmental heat into an electrical potential. A heat differential, on the other hand, could very well be done, so they might be usable (in the long run) as a way to generate electricity while venting waste heat from nuclear reactors and the like.
The good and new comes from no quarter where it is looked for, and is always something different from what is expected.
Oh I really hope your post is a deeply ironic humourous one, sessamoid. I notice all your examples are from the USA.
Having travelled to India, Spain, several other hot countries, I've seen a lot of architecture designed to work with the temperatures: narrow alleyways which are always in shade, houses with thick insulating walls and small windows, air ways through the houses to channel slight breezes into cooling air flows. These things work. These things have been working for hundreds if not thousands of years.
I believe you are refering to architecture and town planning dependant on artificial cooling techniques: big, open pavements exposed to the sun, large glass and steel buildings with huge windows, big doors facing into the sun.... Also check out people's life styles. Remember Rudyard Kipling's famous line 'only mad dogs and Englishmen go out in the noon day sun'. If you're in India or the southern Mediterranean, people get off the streets by 11 in the morning. It's just the crazy tourists wandering around out there. Everybody else is working or relaxing out of the sun in the nice cool shaded buildings. If you want to see real genius, check out the Alhambra in Granada, Spain (and I am sure there are many other fine examples).
Please tell me your email was one of those 'wind up the dumb rednecks' style postings...
"India isn't blessed with AC all over the place and yet the hot streets aren't littered with burnt corpses."
t h_asia/ newsid_1991000/1991215.stm
How about some right here:
http://news.bbc.co.uk/hi/english/world/sou
That was in today's news. 450 dead in Indian heat wave.
I too heard this some time ago, but I believe it was dropped because the harmless gas in question was a either a strong greenhouse gas or a destroyer of ozone.
btw. it's Corus steel (I know, it's a very silly name)
You can't win Darth. If you mod me down, I shall become more powerful than you could possibly imagine
according to the CIA world factbook luxembourg has the highest per capita GDP in the world. US is second by a couple of hundred dollars ($36.4k Vs $36.2k gdp/per capita)
I use Friend/Foe + mod-point modifiers as a karma/reputation system.
- Rob
Unforntnately retarding your cpu to the worst possible conditions each time is silly.
No, it's sound engineering.
Who would buy a P4 if in its worst possible stage, data and enviroment vareables it run at 1G?
What are you talking about? This is true right now, but the P4 is still selling well. This is because if you have proper cooling and are running the cpu in spec, you never see the clock halving occur. The observation is that while the maximum power is X, you rarely if ever see power usage above Y in real workloads. Without this feature, you would still have to design for power X, since it -might- happen. This way, you can design for power Y.
When you start having clock gating it will only cool down the processor on average executions (pending loops etc.).And thus has no point as the chip would still work at exactly thesame speed as the worst possible scenario with none of the gating being turned on.
Once again, what are you talking about? Clock-gating is a power saving technique. You turn off the parts of the chip you aren't using, thus saving switching power on those parts. It does -not- reduce maximum power, except to the degree to which it is impossible to use all parts of the chip at once. The purpose is to reduce -average- or -typical- power, not to increase the max frequency by reducing the worst case. So it does indeed have a point.
The next set of high power precessors will have finer temperature detection and will retard only parts of the chip that are going mad.
Now this, however, would have no point. If a part of your machine is "going mad", that means it is being used heavily, and thus it a safe bet that the code you are running depends on that unit, and thus retarding that part of the chip would decrease your performance, quite possibly exactly to the degree to which you reduce the frequency. At that point, you might as well slow down the entire chip, since you've already screwed yourself by slowing down the critical path.
Eather that or async logic will kick in.
The only thing async logic will do for this is reduce the need for the margin built in to the frequencies of clocked chips. Hey, I'm all for that, but the fact remains that the part of your chip that is hot and thus running slower is almost certainly the critical path for performance, and thus you'll still see performance degrade. Getting rid of that margin is a noble goal, but there are ways to do it without having to resort to the complexity of an asynchronous design.
The enemies of Democracy are