Sign in with
Sign up | Sign in
Your question

What liquid shoud we put into water coolers?

Last response: in Overclocking
Share
May 1, 2006 1:59:25 AM

I undersand to raise the temperature of 1L of water(pure H20) it takes 1 kilojoule (lets say its energy expressed interms of heat).

However if we add impurities to water eg Salt (Sodium Chloride) etc it will now take more energy to rase the waters temperature!!!!

Wouldnt this be a good thing? Imagine 1L of cooolant in your water cooler now needs 100 kilojoules to raise its temp by 1 degree. This would allow the water in your water cooler to stay cool longer (correct me if im wrong)??

I work in a lab with pulse field gel electrophoresis. The buffers we use take 240wats over a 6 hr period and the buffer doesent raise in temperature by more than 5 degrees!

So who uses what liquids in thier water coolers?
May 1, 2006 2:03:23 AM

Ugh, i never heard about liquid coolers but normal water should work?
May 1, 2006 2:46:28 AM

Yea. Wow thats a great idea. Yea lets add salt to a liquid cooling system. While your at it chuck some salt in your car radiator aswell and see how that goes.
Related resources
May 1, 2006 3:00:56 AM

i've heard some of the best things to use are thing like anti-freeze or car coolant for the most extreme users, also have seen references to other products used for cars but i can't remember them right now...
May 1, 2006 3:02:45 AM

What it gonna do if we put vinegar?
May 1, 2006 5:01:10 AM

vinegar is used to clean up build up in yoru loops when you flush it every 6 months

the ideal fluid is indeed water. anything you add will hinder its thermal transfer properties.

if you have all metals with around the same electronegativity, then water and some algecide will work. if you have different ones (ie. copper and alu) use anti-corrosion factors.(they'll slow the corrosion down, but not prevent it)
a b K Overclocking
May 1, 2006 6:15:23 AM

I use water with a little bit of that chemical used to prevent algea growth in humidifiers. It all works nice!
May 1, 2006 7:11:07 AM

Quote:
I undersand to raise the temperature of 1L of water(pure H20) it takes 1 kilojoule (lets say its energy expressed interms of heat).


actually it takes closer to 4000 kilojoules to rase 1 litre of water by 1c.
May 2, 2006 2:18:05 PM

By far the best liquid to use for 99% of closed loop applications is water, with some corrosion and microbe/algae inhibitors.
May 2, 2006 3:06:04 PM

Quote:
I undersand to raise the temperature of 1L of water(pure H20) it takes 1 kilojoule (lets say its energy expressed interms of heat).

However if we add impurities to water eg Salt (Sodium Chloride) etc it will now take more energy to rase the waters temperature!!!!

Wouldnt this be a good thing? Imagine 1L of cooolant in your water cooler now needs 100 kilojoules to raise its temp by 1 degree. This would allow the water in your water cooler to stay cool longer (correct me if im wrong)??

I work in a lab with pulse field gel electrophoresis. The buffers we use take 240wats over a 6 hr period and the buffer doesent raise in temperature by more than 5 degrees!

So who uses what liquids in thier water coolers?


Firstly, a lesson in physics/chemistry. The heat capacity for pure water is 4.2J per gram of water. And since the density of water is 1 gram per cubic centimetre, add your numbers and it would take 4.2 kilojoules of energy to raise the temperature of a litre of water by 1 degree celsius. Adding common salt (Sodium Chloride) will not only increase the heat capacity significantly, but also lower the freezing point of water. However, saltwater is very corrosive for use as a coolant in a cooling system, particularly when exposed to copper. Strange that a person who works in a lab forgot such a simple fact! :lol: 

Anyhoo, the best coolant you can get is distilled water, purely for the fact it still has a high heat capacity and isn't as corrosive.
May 2, 2006 3:07:52 PM

There's nothing wrong about using anti-freeze and distilled water for liquid cooling. But personally I used a non-electrical conductive, non-corrosive and lubricant pc cooling fluid like Fluid XP+ and PrimoChill ICE. I'm new to water cooling and I used this somewhat costly cooling fluid to protect my hardwares as well. :D 
May 2, 2006 4:24:16 PM

Quote:
However if we add impurities to water eg Salt (Sodium Chloride) etc it will now take more energy to rase the waters temperature!!!!

Wouldnt this be a good thing? Imagine 1L of cooolant in your water cooler now needs 100 kilojoules to raise its temp by 1 degree. This would allow the water in your water cooler to stay cool longer (correct me if im wrong)??


You're thinking is backwards. No, it would not be a good thing, if it requires more heat to raise the temp of the liquid, that is more heat the cpu has to produce to raise the temp of the liquid. The idea of liquid cooling is to remove the heat from the item to be cooled. Having any one component of the system that will retains heat or has a high heating index is counter-productive.

You do not want any impurities (or at least as little as possible) in the cooling liquid itself. This is to reduce corrosion, algae, and keep particles from clogging the lines, pumps, resevoir, etc...

And you say you work in a lab...

Quote:
So who uses what liquids in thier water coolers?


You can use distilled water...a mix of distilled water and regular car-anti-freeze...you can purchase coolant and additives from places like Danger Den...and, actually if you peruse the Danger Den site, they offer many resources to answer your liquid cooling questions...that and a Google search yields scads of water cooling info.
May 2, 2006 4:56:05 PM

wonder what would happen if you used LSD as a coolent aditive, other then you computer being a bad trip when playing Doom 3.
May 2, 2006 4:57:27 PM

Quote:
So who uses what liquids in thier water coolers?


If your goal is to remove more heat from your PC, or if you want to lower the operating temperature of a component, then one possible solution is to lower the temperature of the coolant, be it air, water, whatever. Realize that such an approach can bring with it additional problems, such as condensation, noise, a large quantity of exhaust heat, etc. I have not yet witnessed a perfect solution - but I'm workin' on it!
May 2, 2006 5:28:44 PM

i wanna try icy hot?? think about it its cool then hot.....oh wait that hot part isn't gunna work is it? nevermind :?
May 2, 2006 5:59:48 PM

Lol, isnt the whole idea to get heat away from the CPU/GPU?

So therefor the FASTER the liquide picks up the energy from the cpu/gpu the better...

----

Or if u just want to have a as "cool" liquide as possible then dont even mount it on the cpu just let it sitt there pumping liquide around (should be room temp...)

=P
May 2, 2006 6:24:54 PM

Quote:
i've heard some of the best things to use are thing like anti-freeze or car coolant for the most extreme users, also have seen references to other products used for cars but i can't remember them right now...


actually despite popular belief, antifreeze is not a better coolant than water. Antifreeze does just that, it prevents the water in your engine block from freezing and (without freeze plugs) cracking your block. With freeze plugs, theoretically the water begins freezing and it pops the freeze plugs out preventing a cracked block.

Also, antifreeze prevents corrosion to the cooling system in the car (i.e. radiator).

However; if you look at any race car...they don't run antifreeze...straight water.

I honestly don't know the mechanics behind this but water is superior to water + antifreeze or antifreeze.

edit-->however there is another additive called "30 Below" or "40 below" that is used in some hotrods. I'm not sure if this would be a better choice to water...
May 2, 2006 7:03:35 PM

What about Acetone???
May 3, 2006 1:16:11 AM

no, you have it all backwards, would you rather blow HOT air at a CPU or COLD air at a CPU, samething applies here with watercooling the cooler the water the better it do at removing the heat from the CPU, that being said thats why adding salt would be a good idea if not for its corosive factors...
May 3, 2006 2:00:15 AM

Quote:
What about Acetone???


Acetone or a some type of form of it, is what's used in heatpipes in HSF's. It's dissipates heat by condensation then back to liquid form again and repeats the cycle. It is a good thermal conductor however it is corrosive and it can dry out silicon tubes, seals, pumps and other plastic components that is why it is only used with metals like copper of that of HSF. Salt well I don't know much about it thermal conduction properties, but one thing, it can develope mineral deposits and deteriorates the liquid cooling components just like calcuim. That's why for most water cooling applications manufacturers recommend distilled water in addition of the anti-freeze in a ration of 3:1, 3 parts water and 1 part anti-freeze can help by protecting parts from corrosion and as well as inhibits the growth of algea. But if you don't want to bother with that, there's cooling fluid that are made specifically for pc cooling system and although costly, still with it's benifits can outweigh the cost. It's non-electrical conductive, non-toxic, lubricates and protects components, inhibits algea and it has about as close of thermal conduction properties that of water.
May 3, 2006 2:10:54 AM

Okay. If it takes more heat to raise the temp of the water with salt, then where the he!! is all the heat from the cpu going? If the water doesn't take the heat away, your cpu will fry. Therefore, you want a liquid that both heats up quickly (at the cpu) AND equally importantly cools down quickly (in the radiator/fan assembly) That is the basis of the entire watercooling concept. Therefore salt would be a bad idea.

The idea is to remove the heat from the cpu, not to prevent heat absorption.
May 3, 2006 4:59:34 AM

Quote:
What about Acetone???


Acetone or a some type of form of it, is what's used in heatpipes in HSF's. It's dissipates heat by condensation then back to liquid form again and repeats the cycle. It is a good thermal conductor however it is corrosive and it can dry out silicon tubes, seals, pumps and other plastic components that is why it is only used with metals like copper of that of HSF. Salt well I don't know much about it thermal conduction properties, but one thing, it can develope mineral deposits and deteriorates the liquid cooling components just like calcuim. That's why for most water cooling applications manufacturers recommend distilled water in addition of the anti-freeze in a ration of 3:1, 3 parts water and 1 part anti-freeze can help by protecting parts from corrosion and as well as inhibits the growth of algea. But if you don't want to bother with that, there's cooling fluid that are made specifically for pc cooling system and although costly, still with it's benifits can outweigh the cost. It's non-electrical conductive, non-toxic, lubricates and protects components, inhibits algea and it has about as close of thermal conduction properties that of water.

thanks good to know...and i have a heatpipe on my sapphire 9800pro ultimate never came to my mind that there is something inside...doh :D 
May 3, 2006 5:47:19 AM

Quote:
What about Acetone???


It is a good thermal conductor however it is corrosive and it can dry out silicon tubes, seals, pumps and other plastic components that is why it is only used with metals like copper of that of HSF.

I'd apreciate it if you would clarify your comment about acetone being corrosive. It is a very strong solvent for polar organics. It can swell some polymers and even dissolve some polymers. Still, it's possible to put together a recirculating system that could work with acetone. But it's very volatile and flammable and for those reasons, I'd never consider it for a PC liquid cooling system. Combine it with a x1900XTX and you're looking at a bomb in the hood.
May 3, 2006 5:49:29 AM

I recommend using honey dijon mustard.
May 3, 2006 6:04:00 AM

Quote:
I recommend using honey dijon mustard.


So if I fry my cpu, it will taste good?
May 3, 2006 6:05:29 AM

Quote:
I recommend using honey dijon mustard.


So if I fry my cpu, it will taste good?

As long as you use Grey poopon 8O
May 3, 2006 6:47:22 AM

What I mean by corrosive is that acetone can deteriorates plastics and rubber products. Also acetone has a low boiling point of 57c*/134f* at normal atmospheric pressure, making it volitile and it is also highly flammable. But if you're talking about why not make a recirculation cooling system with acetone, well it's possible to do so, but then that's what a refrigerant is for and it is better.

Like this one:

http://www.extremeprometeia.com/acatalog/Prometeia_Mach...
May 3, 2006 10:18:56 AM

Quote:
no, you have it all backwards, would you rather blow HOT air at a CPU or COLD air at a CPU, samething applies here with watercooling the cooler the water the better it do at removing the heat from the CPU, that being said thats why adding salt would be a good idea if not for its corosive factors...


What?! :?: I can't determine if this reply is sarcasm or ignorance...read some of the other replies in this thread...do a little research... and then post another statement acknowledging you're a dumass.
May 3, 2006 2:08:39 PM

My bro is using our dads radiator fluid and it cools great, the problem with this, is that the liquid is dense henc the pump works harder to circulate throuhout the system, but cools great. Another option is distilled water, clean water with no contamination. My neighbor is using distilled water with silver shards, the silver absorbs the heat more efficiently than any other liquid, temps wer 8-14*c lower than just plain h20!!
Tthats crazy :twisted:
May 3, 2006 2:50:21 PM

Deep fried silicon with some dijon, Hell yea it'd taste good :twisted:
May 3, 2006 3:51:07 PM

well if you really wanna have a boost...I mean to get something cooler...put some KNO3 in the water in the reservoir...it will cool the water down by about 5-15°C...but just for a few minutes...and if you want something really exotic put gold shards...they would work a little better than silver...
May 3, 2006 4:43:57 PM

I see your dijon mustard and raise you HAM JELLY.
May 3, 2006 6:35:01 PM

Quote:
Okay. If it takes more heat to raise the temp of the water with salt, then where the he!! is all the heat from the cpu going? If the water doesn't take the heat away, your cpu will fry. Therefore, you want a liquid that both heats up quickly (at the cpu) AND equally importantly cools down quickly (in the radiator/fan assembly) That is the basis of the entire watercooling concept. Therefore salt would be a bad idea.

The idea is to remove the heat from the cpu, not to prevent heat absorption.


Maybe I’m missing something. But basic Thermodynamics would suggest you DON'T want the coolant to heat up quickly. Because the rate of cooling is dependent on the temperature difference between the two conductive surfaces. The higher the temperature difference the faster the heat transfer.

Also, the thermal conductivity of water is not very good when compared to lets say copper.

Copper 401 W/(m*K)
Water 0.613 W/(m*K)
Air 0.026 W/(m*K)

However… when you compare water to air you can see water is an order of magnitude better then air.

I guess what I’m trying to say is heat capacity is only a small part of the equation. The flow rate of fluid through the heat sink is another very important factor as it affects the rate you will transfer heat from the heat sink to the water and remove the warm water from the heatsink.

Given two liquids that have absorbed the same amount of energy the liquid with the higher heat capacity will have the least temperature increase. The other side of this is whether you have adequate cooling to remove this energy from the water before you send it back to the CPU heatsink.

However, the thermal conductivity is equally important to compare.

Ok, I better stop now, its been too long since I had to think about this stuff and I don’t want to have to get out my text books.
May 3, 2006 10:01:45 PM

Quote:
Maybe I’m missing something. But basic Thermodynamics would suggest you DON'T want the coolant to heat up quickly. Because the rate of cooling is dependent on the temperature difference between the two conductive surfaces. The higher the temperature difference the faster the heat transfer.

And...
Quote:
Given two liquids that have absorbed the same amount of energy the liquid with the higher heat capacity will have the least temperature increase. The other side of this is whether you have adequate cooling to remove this energy from the water before you send it back to the CPU heatsink.


I gotta go with the concept that you DO want the coolant to heat up quickly. If the rate of cooling is dependent on the temperature difference between the two conductive surfaces (in this case, the copper heatsink and the coolant) then, to in order to get get the best cooling effect, the coolant would need to remove as much heat as possible as quickly as possible from the heatsink. The more heat transfered from the cpu to the heat sink to the coolant, the more heat is going to be moved away from the heatsink and cpu.

Quote:
I guess what I’m trying to say is heat capacity is only a small part of the equation. The flow rate of fluid through the heat sink is another very important factor as it affects the rate you will transfer heat from the heat sink to the water and remove the warm water from the heatsink.


Absolutely! The ability of the liquid to absorb heat is just one part of the system. You need to have an adequate flow rate to ensure a supply of "cold" coolant to remove the heat from the heatsink. And, ensure that the heat exchanger/radiator can remove the heat from the coolant quickly enough to return it to a nominal temperature.

With all that said, I still go with the concept that you want the coolant to absorb as much heat as possible as quickly as possible.
May 3, 2006 11:13:45 PM

so if you use anything but water...you want something colder not something that is gunna resist heating up
May 4, 2006 12:15:51 AM

Quote:
so if you use anything but water...you want something colder not something that is gunna resist heating up


"I gotta go with the concept that you DO want the coolant to heat up quickly. "


/sigh.

This is basic thermodymanics. The material with a higher heat capacity will heat up slower, but remove more "heat" (actually energy) then a fluid with a lower heat capacity assuming they both have the same thermal conductivity. This is because the fluid that remains cooler will have a larger temperature differece and drive the heat transfer.

The calculation to determine the rate of heat tranfer is something like q=-(k/L) (T2-T1).

L = thickness of the material
K =thermal conductivity constant

So the total transfer is greatly influenced by the temperature difference.

Regardless of what people "feel" and "assume".
May 4, 2006 12:57:03 AM

Quote:
Mountain Dew


BEER! Vodka (at least it's clear like water). Alcohol!
May 4, 2006 1:11:41 AM

Don't forget the moonshine there buddy! :D 
May 4, 2006 1:43:17 AM

Quote:
Real-world performance is what matters here. (extensive testing required)
Having loads and loads of theory will get you nowhere unless they're tested and proven.


I guess an example would provide a better explanation.

Lets take two liquids, acetone and water.

Heat capacity Thermal conductivity
Acetone 0.5 cal/(g c) 0.514 W/(m K)
Water 1 cal/(g c) 0.613 W/(m K)

Both liquids have similar thermal conductivities with water having approximately 20% advantage. However, the heat capacity of water is twice that of acetone.

Now, if you ran a system with acetone, the rate of energy absorption would be slower because of the lower thermal conductivity, but the temperature of the acetone would end up higher then the water would in the same system because acetones heat capacity is so much lower.

I just chose these two liquids because I could readily find the properties. It just shows that the rate a liquid “heats up” is not good indicator as to its effectiveness.

In order to understand this, you must understand the difference between heat capacity and thermal conductivity. It has been a long time since I have studied this so it is possible I’ve made a mistake or two. But I’m confident in the general principle. These are “Laws” of thermodynamics. Not “theory”. They quit testing this a long time ago.
May 4, 2006 2:23:38 AM

Long story short:

The companies that spend millions developing cooling solutions have engineers who have a greater knowledge base on this subject than we do. If some other liquid is more capable as a cooling solution, don't you think someone, somewhere would have come up with a way to utilize it effectively? Saying that "X" is corrosive or "Y" is a better conductor may, in and of themselves, be true statements. However, Zalman, Thermaltake, Coolermaster and every other company KNOWS THIS TOO. With the need for better cooling reaching such great heights, don't you think that one of these companies could come up with a way to implement these ideas if they could be done? Just follow the directions that come w/ your stuff and be content that you are one of the few lucky ones who have such a cool setup that you need such a specialized cooling device in the first place!
May 4, 2006 2:31:37 AM

since it is called WATERcooling and everyone knows that water is a sufficient enough cooler unless more extreme cooling is required and other substitutes are actually quite expensive and the people who actually use this stuff aren't your typical "whats the difference between memory and a harddrive" the general watercooling population would not accept a substitute that costs money, let alone an expensive one when all they need is FREE water

and thank you -feller- for saying what i couldn't to backup what i said earlier
Quote:

What?! Question I can't determine if this reply is sarcasm or ignorance...read some of the other replies in this thread...do a little research... and then post another statement acknowledging you're a dumass.


ok well how about you read some earlier posts dumbass as my post is still credible and correct in most situations

Quote:
High/race performance cars uses WaterWetter which is what I'm using with my water system.


thanks for clearing up the antifreeze note that i posted, your WaterWetter stuff is what i meant to say when i said there are other things that cars use that can be used in watercooling but i forgot the name, thx for the clearup
May 4, 2006 2:47:01 AM

Were you talking to me or did you just press the nearest reply button?
May 4, 2006 3:33:59 AM

Quote:
Real-world performance is what matters here. (extensive testing required)
Having loads and loads of theory will get you nowhere unless they're tested and proven.


I guess an example would provide a better explanation.

Lets take two liquids, acetone and water.

Heat capacity Thermal conductivity
Acetone 0.5 cal/(g c) 0.514 W/(m K)
Water 1 cal/(g c) 0.613 W/(m K)

Both liquids have similar thermal conductivities with water having approximately 20% advantage. However, the heat capacity of water is twice that of acetone.

Now, if you ran a system with acetone, the rate of energy absorption would be slower because of the lower thermal conductivity, but the temperature of the acetone would end up higher then the water would in the same system because acetones heat capacity is so much lower.

I just chose these two liquids because I could readily find the properties. It just shows that the rate a liquid “heats up” is not good indicator as to its effectiveness.

In order to understand this, you must understand the difference between heat capacity and thermal conductivity. It has been a long time since I have studied this so it is possible I’ve made a mistake or two. But I’m confident in the general principle. These are “Laws” of thermodynamics. Not “theory”. They quit testing this a long time ago.

This thread has officially jumped the shark...
May 4, 2006 3:38:32 AM

Agreed. Let's just let it die.
May 6, 2006 2:26:54 AM

Quote:
Were you talking to me or did you just press the nearest reply button?


nearest reply
June 17, 2006 7:27:20 PM

Quote:
There's nothing wrong about using anti-freeze and distilled water for liquid cooling. But personally I used a non-electrical conductive, non-corrosive and lubricant pc cooling fluid like Fluid XP+ and PrimoChill ICE. I'm new to water cooling and I used this somewhat costly cooling fluid to protect my hardwares as well. :D 



I'm going this route to. How much fluid did you buy? The reservoir I'm thinking on getting is the swiftech MCRES ( 4.5 fl oz ), + tubing + CPU block + GPU (crossfire) blocks + pump + radiator ( thermochill PA120.3). A bottle size is 32 oz.
June 17, 2006 8:10:38 PM

I hear gasoline works well as a coolant, however don't get it to vapor form.
June 18, 2006 1:37:37 AM

I think you guys missed to see the hole picture here, with both the cpu and the radiator.
Cold liquid goes into the cpu block and is heated by x degrees depending on processor load and heat capacity of the liquid. The liquid then goes to the radiator where it is coled by x degrees.
A large heat capacity means that this temperature difference x will be small, while still trancfering the same heat energy.

All we can see by looking at the cpu block is how much the water increases in temperature, not the temperature itself.

To see that we must look at the radiator. The radiators energy transfer to the air is proportional to the temperature difference between the radiator and the air. This means that it does not matter whether there is a great or small difference in the in and the out temperature; the cooling only depends on the mean temperature.
This means that the mean temperature of the water and thus the processor is independent of the heat capacity.


Conclusions

The heat capacity does not effect the average temperature, but whith a low heat capacity, you will have a cold and a hot side of the processor.

Good heat conductivety is favorable at both the processor and the radiator.
To get high heat conduction you need high water flow and a liquid with low viscosity.

Ps. I hope I cleared out more confusion than I created.
June 18, 2006 2:24:23 AM

Quote:
The companies that spend millions developing cooling solutions have engineers who have a greater knowledge base on this subject than we do. If some other liquid is more capable as a cooling solution, don't you think someone, somewhere would have come up with a way to utilize it effectively?

Please don't take this as a personal attack, because I do not intend it to be offensive in any way, but I must say that this statement inspired me to respond to the thread. The idea that "someone else will figure it out" is one that has always bugged me. Granted that it is important to know your limitations in any given universe of discourse, but it is also important to knock down those limitations and strive to both understand and innovate. It is only through discussion and collaboration of ideas that innovation is possible. With that said, I will now do my best to lay out the key concepts of this topic as I perceive them.

It seems that some people are confused about the difference between heat and temperature. Heat is merely radiant energy, which is nothing more than electromagnetic radiation (photons) in the infrared range of the spectrum. Temperature is the observed effect on a material as it is exposed to heat.

Feller was very precise in his explanation of specific heat capacity, but it can be summed up quite simply: the heat capacity of a material indicates the amount of energy that a material must displace in order to change its temperature by some given amount. In general, a material with a high specific heat capacity has a low thermal conductivity. Likewise, a material with a high thermal conductivity tends to have a low specific heat capacity.
The charts on this page illustrate the idea well:
http://www.overclockers.com/articles609/
For reference purposes, here are tables of the heat capacities and thermal conductivities of some common materials:
http://hypertextbook.com/physics/thermal/heat-sensible/
http://hyperphysics.phy-astr.gsu.edu/hbase/tables/thrcn...
It is important to note that water has one of the highest specific heat capacities for liquids, second only to ammonia, which is highly corrosive. Therefore, adding salt or ethylene glycol (antifreeze) to water will actually result in a mixture with a lower heat capacity than pure water. This is why you add salt to water in order to make it boil more quickly:
http://www.swri.edu/10light/water.htm

To comment on the original suggestion of adding salt to the circulated water in a cooling system, it seems like that would defeat the point of using water in the first place. While this is in agreement with what some people said, it is for the completely opposite reason. Adding salt will make the water change temperature more rapidly, which is not the desired effect in this case. As for increasing corrosion, as long as you use aluminum water blocks and assuming that your pump has no internally exposed metal parts, you *shouldn’t* have to worry. A little known fact about aluminum is that it is extremely resistant to corrosion between pHs of 4 and 8.5 because of a natural oxidation layer covering the entire surface. Since the addition of NaCl alone does not change the pH of water (7.0), and no abrasion is taking place inside the water block, there is no reason why corrosion should occur. You can read more about the topic here:
http://www.corrosion-doctors.org/MatSelect/corralumin.h...

So now that the chemistry part is over with, let’s talk more about how it relates to our cooling discussion. There are several popular ways to remove heat from a CPU. For the purpose of simplifying this list, let’s simply define a heat sink as a material with a high thermal conductivity, usually aluminum or copper.
1. Air cooling - a heat sink directly dissipates heat into the air
2. Liquid cooling - a heat sink dissipates heat into a liquid, which later deposits the heat at a second heat sink that dissipates the heat into the air
3. Peltier cooling - a semiconductor device uses electrical energy to displace heat into a heat sink, which then dissipates the heat into the air
4. Vapor-phase cooling - a heat pipe or compression refrigerator displaces heat into a heat sink, which then dissipates the heat into the air

What all four systems have in common is the fact that they all end with a heat sink dissipating the heat into the air. The net effect is always the same, but we have different ways of getting there. If you notice, I have listed these cooling methods in order of increasing effectiveness (not necessarily reliability), which happens to line up directly with the order of increasing complexity. If removing heat were as simple as wanting a cooling device to change temperature very quickly (which is what some people explicitly stated here), then it would be logical to infer that air cooling is the most ideal system. We all know this to be false however, and the reason lies in the concept of specific heat capacity as it has been explained thus far.

Liquid cooling relies on a liquid with a high heat capacity, ideally pure H2O, to absorb a large amount of energy. The higher the heat capacity of the liquid, the lower the sustained temperature of this liquid will be as it absorbs heat (energy), and the lower the resulting temperature of the CPU will be. The problem with liquid cooling is that it relies on a liquid with a natural and nearly constant heat capacity. This means that the limit of the temperature buffer is fairly low.

Peltier and vapor-phase cooling are ways of getting around this barrier. By creating a heat vacuum (a large deficit of heat) on the input side of the system, the effect of a material with a very high heat capacity can be emulated. The details of how these two methods work are readily available to those who wish to read:
http://en.wikipedia.org/wiki/Peltier-Seebeck_effect
http://en.wikipedia.org/wiki/Vapor-compression_refriger...


FINAL THOUGHTS:
In order to have an effective liquid cooling system, you need to have heat sinks with very high thermal conductivities and a liquid with a very high specific heat capacity. In order to maximize the efficiency of the system you want to have a high liquid flow rate, an effective heat sink design, and a high air flow rate at the final output of the system.

The main thing that I feel is lacking in current water cooling systems is the effective heat sink design. A little over a year ago I was looking at water cooling systems and decided against trying one just because I was dissatisfied with everything on the market. I finally decided to spend some time designing my own water block to mill out in the university machine shop. Numerous hours and several revisions later, I came up with this:
http://people.clarkson.edu/~lanahamc/media/pc/cpu_block...
I designed this block to fit a Socket A system since I was running an Athlon XP system at the time (and still am for now). The block meets the exact dimensions of a stock AMD cooler and was designed with a high flow-rate in mind. This picture illustrates how the water would be channeled away as it enters the block:
http://people.clarkson.edu/~lanahamc/media/pc/cpu_block...
I have yet to machine this block, as I plan to make revisions so that it fits the LGA775 form factor. I also need to devise a decent mounting system. I must admit that newer designs such as Swiftech’s Storm and Apogee water blocks are an improvement, but I still think they can do better.

If anyone has suggestions or comments about my design, I am definitely open to criticism. Please send me a PM with your thoughts so as not to hijack the thread.
!