Slashdot Mirror
Sapphire: A Liquid That Won't Get Things Wet
eaglebtc writes "Tuesday on Good Morning America, a representative from Tyco Fire & Security demonstrated an amazing new substance called Sapphire: a water-like fluid that does not get things wet. He filled a small fish tank with Sapphire and submerged a book, a laptop, and a flat panel TV. Both electronics were turned on when submerged; all three items came out completely unharmed. Click here for a slideshow of the demonstration. The official name for Sapphire is actually Novec 1230. Read about it here (PDF). Tyco sees practical applications of Sapphire in fire extinguisher systems for museums and libraries. By the same token of practicality, regular readers of Slashdot probably have something else in mind: total-immersion watercooling. Just think of the possibilities!"
Offtopic, but the submitter opened the door: according to their specs sheet (PDF warning), this stuff has a boiling point of 49.2C (120.6F). Processors burn hotter than that, how useful would it still be for cooling purposes if it were a gas? I also have to wonder what the long-term effects of exposure would be... it's one thing to dunk a laptop for a few seconds, it's something else entirely to have it swimming all day long. At least your machine would never catch on fire.
;)
They might have some information there about how well the stuff will conduct heat, but I got a lousy grade in Chemistry, so I'll leave it to the experts.
Auto-reply to ACs: "Truly, you have a dizzying intellect."
A very interesting note is that Saphire/Novec 1230 has a freezing point at -162.4*F according to 3M's white paper
Remembering that you are going to die is the best way I know to avoid the trap of thinking you have something to lose.
So is this concept of non-conductive-water-like substances all that new?
Wouldn't distilled water work just fine for total-submergion water cooling? After all, it's the ions in water that make it a conductor, correct?
of the chemical in action.
This was on Fark already. Read the PDF. It's stored as a liquid in the tank, but its a cyclic fluorocarbon that vaporizes on release. This is simply the latest version of Halon (TM) fire extinguishers, not an "non-wetting water" or an "non-flammable organic with a water-like viscosity" At work we have carbon dioxide jets in the server room in case of fire. We've never had the building burn down so I don't know the merits of either method.
This is of course not the first liquid that does not cause harm to electronics, and can be used for total immersion water cooling. Fluorinert (3m) has been around for a while. One version of it is(was) also used for liquid breething deep diving (same as used on "The Abyss").
main(i){putchar(177663314>>6*(i-1)&63|!!(i<5)<<6)&&main(++i);}
Auto-reply to ACs: "Truly, you have a dizzying intellect."
Fluorinert does the same thing, and it's been around for many years. That's what was used in some Cray machines.
In addition the stuff is photolytic by UV light. The PDF states the stuff would be expected to last about five days when exposed to the atmosphere. Fluorescent lights put out a fair amount of UV, so if it were used for cooling, it would have to be a well-sealed opaque tank.
If you can't beat them, embrace and extend them.
Tell that to the NCO's in Viet Nam who had to drink glasses of Agent Orange to show the men it was safe. Many had children with severe birth defects, assuming the stuff didn't make them completely infertile.
You have to open the slideshow by clicking on the link between the short opener in bold and the beginning of the article. Here's the link Look for the link "IMAGES: Liquid in action on Good Morning America" in between the two.
The PDF says it breaks down quickly under UV light...
> how to you clean it up?
From the article:
The fire protection fluid will quickly evaporate, safeguarding
these items - and leaving them unaffected.
George II -- Spreading Freedom and American values, one bomb at a time.
Not for long anyways, heres the rundown of tryed and failed experments:
-immersion in tap water: its conductive, one person was stupid enough to try this on his shiny new system, lets just say the power supply did somehting intersting.....
-immersion in distilled/de-ionised water: it gets contaminated by the computer and becoms slightily conductive, all the traces corrode.
-immersion in mineral oil: works for a few days but then stopped working with no obvious damage. Probily the capacitors soaked up the oil and that changed their electrical properites.
So theonly this stuff will work is if you use some kind os sealent on the board around the capicators and that might not even work...
Boiling Point @ 1 atm 49.2 C
Heat of Vaporization @ boiling point 88.0 kJ/kg
Vapor Pressure 0.404 bar
This is a liquid that will readily evaporate (a little slower than ether would). If a limited quanitity is used (such as in a hand-held extinguisher), it will probably evaporate before you get the chance to clean it up.
The article also states that the LC50 is over 10% by volume, which tells this substance is probably not very dangerous, unless specific medical problems arise.
As it seems to be safe to the atmosphere as well, i guess the 'plan' is to just let it sit there and evaporate.
This may sound dangerous, but we do the same with CO2 - which is more lethal to anyone entering the room and possibly to the environment (global warming) as well.
Their PDF blurbs talk about the ozone question, and say that the stuff breaks down in about 5 days in sunlight and doesn't bother the ozone. It doesn't say exactly what happens to the Fluorine in the process, but it does say that it's much different from the hydrogen-fluorine-carbon compounds like Freons. From what I remember, the freons catalyze the breakdown of ozone; perhaps the breakdown products from this compound don't do that (e.g. maybe the fluorine atoms all end up as F2.)
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
Okay, chemistry geeks, it's:
3 -PENTANONE
1,1,1,2,2,4,5,5,5-NONAFLUORO-4-(TRIFLUOROMETHYL)-
Anyone know if that's conductive?
$8.95/mo web hosting
Asinine is defined as "Utterly stupid or silly." Doesn't that disqualify true statements? Caffeine is more toxic than plutonium.
Mod down posts with a "Free Mac Mini/iPod" sig, they're spam!
But when will we have "Liquid you can drink and not be accused of modding on crack"?
Even those who arrange and design shrubberies are under considerable economic stress at this period in history.
It's only useful in total immersion cooling if and only if it's cheaper than fluorinert (by 3M).
Of course, due to the low boiling point of Sapphire, it would be necessary to isolate the cpu and probably the gpu and cool them with normal watercooling and the rest can be submerged in sapphire, which will also help reduce the build up of dust, enemy of efficient cooling.
BTW, TIC is nothing new. There are these crazy New Zealanders who bought 2 gallons of fluorinert, priced at 500 bucks a gallon (and you thought gas prices in california was bad), submerged their entire setup in it and cooled the liquid.
Unfortunately, LN2's temp was well below the freezing point of fluorinert. Here's a linkie
It specifically states that it's NOT like HALON - it puts out fires by cooling vs. interference with fire chemistry (as HALON does). And it can also be used in "streaming" applications (like water).
Sean
The MSDS sheet for ansul is here: http://www.ansul.com/docs/msds/F-2003263.pdf
Specific Physical Form: Liquid
Odor, Color, Grade: clear colorless, low odor
General physical form: Liquid
Autoignition temperature Not Applicable
Flash Point Not Applicable
Flammable Limits - LEL Not Applicable
Flammable Limits - UEL Not Applicable
Boiling point 46C
Vapor Density 11.6 [RefStd: Air=1]
Vapor Pressure 244mmHg [Details: @20C]
Specific Gravity 1.6 [RefStd: Water =1]
pH Not Applicable
Melting Point -108C
Solubility in Water None
Evaporation Rate >1 [RefStd: BUOAC=1]
Volatile Organic Compunds No Data Available
Percent volatile 100%
VOC Less H20 No Data Available
Viscosity 0.5 centiposise
One problem. Try heating a frying pan with nothing on it. After you can sense that it's hot enough, sprinkle a little water on it. The water will float over the pan.
When the water makes contact with the hot pan, it turns to steam, which then insulates the remaining water above the pocket. The temperature of that pocket of steam gets quite high since it has little opportunity to escape and doesn't really get cooled. More importantly, the pan gets very little cooling effect from the water evaporation.
Therefore, you should never rely on coolant when any part of it is at or very close to its boiling point. The coolant properties of the fluid break down.
The above doesn't really match your example, since it's not immersed within a coolant environment. For a better example, use a boiling pot of water. Examination of the locations of steam nucleation reveal that those areas (however small) do not get wet, and gets insulated as illustrated above.
Go back and read the datasheet. This saphire stuff has very low heat of evaporation. Basically, it's designed to boil off right away when you pour it on fire. It has high vapor pressure, which allows a large amount of it to stay in gaseous form mixed into the air, extinguishing the fire by oxigen displacement. A good coolant, on the other hand, would have a high heat of evaporation and low vapor pressure, like water. Water puts out the fire by using up the thermal enerrgy on evaporation and cooling down the materials, not by oxigen displacement. If you want submersion cooling, look for something that won't evaporate easily.
I've seen submerged mineral oil cpu boards before. Heck, here's one right here at the very top of the Google.
If you put it deep enough in the fluid, the bubbles will condense on the way up.
However the shockwaves from their formation and collapse (cavitation) could cause physical damage to the chip packaging, especially where conductors penetrate it. If the chip package isn't designed for it, total immersion is proably out. Back to liquid-cooled clampons. (In which case, why not use water, which has extremely high specific and vaporization heats?)
If the heat of vaporization is anything reasonable, this should work quite well to remove heat from your chip--the fluid changing to a gas absorbs a bunch of heat,
But the heat of vaporization is extremely low compared to water - by a factor of 25! (That's why it can be "stored as a liquid and used as a gas" - the small amount of heat in the air causes a spary to immediately evaporate).
Specific heat wasn't stated - but with such a low heat of vaporization it is also probably low and/or doesn't matter. You're going to have to circulate this stuff REALLY FAST to get usable cooling.
Note that its use as fire suppression is not relevant to its use as cooling. Though this stuff DOES suppress fires by cooling (unlike halon, which interferes with the chemical reactions), fire suppression is a one-pass rather than multi-pass function. So the cooling can be accomplished by breaking up the molecule - using the heat of formation, in addition ot the the specific or vaproization heats, to cool the fuel. I doubt that you want to be continuously consuming your coolant and disposing of the resulting fluorinated alkyl radicals in your home system.
Also, I'm concerned about the toxicity.
This is being sold as a fire suppressant. Fires, and their combustion products, are SO toxic that a suppression system chemical can be quite hellish and still be a drastic improvement. But long-term exposure as an alternative to non-exposure is a far different can of worms.
One document touts that the LD50 (concentration that kills 50% of those exposed) and cardiac sensitization NOAEL (no observable effects level) - both ACCUTE (immediate) poisoning measures - are both "over 10% v/v". But another document, touting its rapid vaporization, point out that the equilibrium vapor pressur in air is four times that: 40% (nearly half the air replaced by vapor). And given how easily this stuff vaporizes, it can approach that damned quickly. So dumping warm coolant might quickly displace nearly half the air with this stuff's vapor and put you in jepoardy - of suffocation if nothing else. Not a problem if it's putting out a fire - BIG problem if it's not.
With that high vapor pressure and low heat of vaporization, exposure would tend to be very high during handling or in the presense of even a tiny leak. So if there are even small long-term toxic effects you'd want to avoid having this where it could result in repeated and prolonged contact.
Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
In the chemistry of ozone depletion, an O-X bond is formed, where X is a halogen. Stratospheric UV breaks this bond for X = Cl or X = Br, allowing one atom halogen to destroy many ozone molecules (catalytic destruction). But, stratospheric UV can't break the O-F bond. So one F atom destroys one ozone molecule only.
The clearance system sounds logical. It is not. It is completely arbitrary. -- John Bolton
"I seem to remember hearing..."
Maybe this glossary is where the person who told you read it from. Under PFIB it says:
"Perfluoroisobutylene. A toxic, colorless, odorless gas that can be produced when Fluorinert liquid thermally decomposes when exposed to open flames, glowing electric heating elements, electric arcs, or temperatures above 200 degrees Celsius."
Esteem isn't a zero sum game
Being a Mechanical Engineering by training I used this technology back in early 1990s while doing my undergraduate degree at Washington State University.
It is expensive as hell (at the time it was expensive).
It is by no means a new break through, unless they are considering the barrier of entry being no longer cost prohibitive as a break through.
Some reactors (namely, Boiling Water reactors) run right at the boiling point. They use nucleate boiling at the surface of the fuel rod to break up a laminate layer that tends to insulate the rod from the rest of the water. In other words, the turbulance caused by the boiling increases the heat transfer rate.
"Fire suppression systems such as those that use Halon (which was outlawed in the '90s due to its ozone-destroying side-effects) put out fires by displacing oxygen with some other gas."
Nope. Halon systems work by absorbing free radicals in the fire. It literally interferes with the chemical processes required to sustain burning. The same reason they are so dangerous to the ozone layer. It's also the reason the newer gas based fire suppressants aren't nearly as effective.
Free radicals are also the reason pre-burned wood can be restarted burning again so easily, and why burned toast is more likely to give you bowel cancer than lightly tanned toast.
Government of the people, by corporate executives, for corporate profits.
Nuclear reactors have a lot of design time to make sure they work. They're also made to exacting tolerances - thus things like the surface roughness are precisly known and controlled.
More importantly, a boiling water reactor uses the water as a moderator. When as a gas, it's much less effective as a moderator than as a liquid. This operates as a feedback system (too much heat generated - water boils - reaction rate slows - system cools), which is critical to the design here. The water would be more efficent at cooling, if the system was run at a lower temperature. However, the system of reactor - turbine - generator is more efficent as a whole when the water is run near it's boiling point (because the heat exchanging systems work more efficently with a greater temperature differenctial).
So, yes, it is used in those cases - but that's not the most efficent method of using the water _as a coolant_. Within a microprocessor, you have no feedback loop to reduce heat production when the temperature peaks the boiling point [0], and no desire to maximise the running temperature.
Personally, I'll stick with water for electronics cooling.
I was a small time firefighter for awhile, and we trained to deal with this stuff all the time.
:)
First off, you can put a fire out by using water to cover all the surfaces and deprive it of oxygen. One one fire the gire started in a basement room that was closed on all sides. We didn't get there fast enough to snuff it out at the point of origin. But the fire damned sure went out when we pumped the basement full. That's not the goal, but it works, guaranteed.
The ideal way of dealing with ordinary house fires is to get there fast enough so it's confined to one room. You advance a hoste team a bit into the room, set the nozzle to a 30 degree fog pattern and move thestream in a clockwise pattern thru the seat of the fire, up to the ceiling and back down. A few sweeps like that and a small fire will be out and you can get to work. At no time is the fire actually "drowned."
If the room is already mostly engulfed, you just crack the door, pump a bunch of fog in, and let your old friend vapor phase cooling suck the heat out of the fire. The only downer is that anybody in there is gonna cook for sure.
In our training burns we would start a fire in the tower, let it build up, and then run the hose team in. The boundary between smoke and clear air is actually quite distinct, i'd say no more than a foot high. When you cut loose with the water everything gets dak fast, the layer drops down, and life starts to suck. It does give ou confidence in your gear though. You have to experience it if you get the chance.
Theoretically you can put out a room fire without causing any water damage by putting in just the right amount of a fog stream into the room and letting all the water be vaporized. But fire fighting isn't ballet, and I've never seen it happen. Usually water damage is pretty bad.
One effect that can't be ignored is the mechanical force of a stream of water coming out of a hose. It can and will tear stuff up. If you spread out the stuff that is burning, stuff gets cooler and is easier to deal with. A nice straight bore nozzle hooked up to a good size pump can and will tear thru roofs without a problem. You absolutely have to get a hold of one of these things at least once in your life
This is so not my area, but Metal fires, Class D in the U.S., are only fought with powder extinguishers. Any extinguisher that isn't marked as a Class D is useless and a health hazard. One of the main reason you don't spray water onto a Class D fire is that you tend to get violent and explosive spalling. Life will start to suck in a big way if a moron tries that. It's a specialty area and any shop that workes with these metals, zirconium, and a bunch of others should put bucks into their training and safety gear. It's not to be messed with by amateurs.
Why do I have this? I don't smoke.
They used the Flourinert from 3M (link in another post). Drives are NOT submersed, only the Mobo, CPU and cards. Since it's the CPU, GPU, chipset and memory that really need the extra cooling when overclocking anyway that's fine.
The liquid Tech TV used cost over $900 US per gallon, so Saphire is just a cheaper version of the same basic thing. It's chemically inert (won't cause corrosion), non conductive (won't cause short circuits), and non toxic (fish can live in it if you add a standard aquarium air pump). Overclockers.com also has articles on submersion cooling.
I'm glad to see this, the only thing that stopped me from building an E-quarium (complete with fake swimming fish (no fish poop on my Mobo, thank you!)) was the price, since I calculated it would take three gallons to fill a recirculating system.
Tommy
Open Source for Open Minds
how useful would it still be for cooling purposes if it were a gas? Potentially very useful depending on the properties of that gas.
Of course!
More importantly, though, if the "water" is boiling because of the heat of your motherboard, it's undergoing a phase change - while it does that, it will consume all available heat to continue the phase change rather than elevate the temperature.
A pot of boiling water will never get over 100C until after all the water has boiled off (or if you increase the pressure, ie. a pressure cooker or a steam engine - PV = nRT!). Likewise, this will never let the processor get above 50C until all the coolant has boiled off. But if you capture the vapor, condense it, and drip it back into the computer's enclosure, you've got a closed system which is good indefinitely. I would worry, however, that if this stuff doesn't "wet", it probably has a lot of surface tension - so the "water" to processor interface won't be as tight as it would be with water, and therefore there might be a little more localized heating effects, similar to water droplets flying across a hot skillet. On the other hand, I'm sure the liquid, being capable of convective flow and in direct contact with the processor, will probably couple heat at least as well as a conventional heat sink. We should also look up the specific heat of this stuff. (Too lazy, didn't check to see whether it was in any of the cited links.)
Personally, I doubt you'd actually maintain the entire vessel at the temperature of the hottest component (the processor) - convective flow within the enclosure will move the hot "water" to the outsides of the container, where the surface area (thousands of times greater than the surface area of the processor) will couple away the heat to the atmosphere.
But why can't you do this with regular water? Submerge the motherboard and cards only - not the drives or the power supply. The thermal transfer grease isn't water soluble, so I don't think you'll make it into a conductive ionic solution. Voltages from pin to pin are pretty low, and pure water is a good insulator - the only problem is getting pure enough water, and keeping it pure enough. Corrosion will also not be a problem, again if the water is pure, the motherboard is continually submerged (preferably with a small surface area to the air so that less oxygen dissolves in it), and connectors are tin/nickel/gold plated - as most of them are. Wash the motherboard/cards/cables in distilled water before putting them into the bath, to make sure that you don't take any ionic contaminants (salts, etc. in dust, fuzzies around old CMOS batteries, manufacturing chemical remnants, little bits of leakage from electrolytic capacitors) in with you.
Components? Modern components are usually sealed anyway - the last step of manufacturing is removing the soldering flux, and that's usually done in what is, essentially, a dishwasher. They're not actually rated for immersion, but most of them do take a good spraying. I'd take out PC-board mounted piezo speakers, but offhand, I think that's the only part that would really have a problem with it.
I think I'd try this with an old computer before doing it with my real one. I seem to have an old 486DX-33 with 30-pin SIMMs and 16-bit ISA slots only... it's been begging for a job. I'll check out retail distilled water on my megaohm meter first.
Fire and Meat. Yummy.
From the article: "The chemical has all the firefighting properties of water..."
Except one. It doesn't wet. It will slide right off anything it touches, allowing the fire back onto it.
We also don't know what its evaporative cooling properties are. Someone might, but we don't. The misting of water in a burning room cools gases and reduces flashing probabilities.
The smothering property is nice, but it doesn't have all the firefighting properties of water.
But then, it has firefighting properties water doesn't have. Water, for instance, will cause electrical fires, and electrical explosions if the voltage of the electrical system is particularly high (first it electrolyzes to hydrogen and oxygen, then it recombines explosively to become water again).
Things 'wet' due to very small forces between molecules. For instance, water molecules are more attracted to paper than to itself, so paper gets wet. Water molecules are not attracted to grease/oil, so they dont mix.
Same thing here - basically the liquid is more attracted to itself than other objects, so it basically stays in it's own droplets.
That's about as simple as it gets.