Endothermic cpu cooling?

[Gam3r01]

So my school sends a few students who do reasearch each year to the junior science and humanities symposium to do research and hopefully present it. I went a year ago and decided to work on something. I concluded that I would try to figure out a way to get a continuous endothermic reaction in a closed loop to cool computer hardware. Before I even try to find a way to reverse the reaction such as HVAC, I would like to know if it would damage the system somehow first. Im not asking for a reaction or a way to reverse it, just if it is simply safe.
I have decided (as much as I can) that amonium chloride or amonium nitrate reacting with water might be the best to use, as they yield at leat 15c cooling over plain water.

Would this corrode the blocks or tubing/radiators being used? Would there be any toxic by products? (note this would be a closed system as I would need all the products to be recombined into reactants which would create heat, hence the radiator.)

I could find a way to reverse it most likely, I just need to know if it would run in a loop.

 

[Calculatron]

I honestly don't have a clue, but, I just wanted to say that this sounds like a bloody awesome idea.

 

[Gam3r01]

Yeah, Ive been putting thought into it for a while now. If that could be done that would drop cpu temps 15-25C over custom water cooling.

 

[rogue909]

I love weird cooling experiments... and this is a weird one
Bad news;
I was thinking from the perspective of the waterblock; you will need to have 2 intakes and a single output. Else the reaction would occur in the tubes as opposed to at the block. They would lose a chunk of the cooling en-route.
A reaction will cause temps to drop 15-25C; room temp is 20C, you will definitely have condensation issues. I'm not trying to discourage, merely educate.
A by-product of the reaction is chlorine... this is bad for all water-cooled systems. It degrades the acrylic and can accelerate issues with metals.
Both ammonia nitrate and chloride are solids at STP. I am assuming you are going into chemistry but I'm not sure about a deliver system for that.
Good news;
Loss of temps in pipes and condensation at the block have all been dealth with before. Research "Phase change cooling" (imagine using a compressor & radiator from an AC unit to cool your PC) and liquid nitrogen (you are probably familiar with that one) systems. Both of those are rather exotic but have been done before. Super OCers use liquid N for competitions, its been done.
As far as the deliver system & chlorine....
I'm not sure

 

[Gam3r01]

Ive taken CHEM last year so I know the basics and whatnot. As for the condensation I would most likely isolate the block/insulate it so that the air cant get cold enough to condense anywhere that would harm anything.
I might have to use the ammonium nitrate instead, as chlorine would be troublesome. AS for the delivery it would intake as an aqueous solution and once mixed with water or another reactant it would then exchange heat and turn to gas(in the block), which would then escape though the output tubing.
The way around this I was thinking of actually came from my welding class. Oxy acetylene welding torches use one way gaseous (and can be liquid) valves, so if I get one of those (actually 3) then the reaction would have a strict area to react and could not unless it got there, and could not leave in anything but a gaseous state.
Once left a system like HVAC would then compress the gas which would release all the heat it took in (at a rad with fans to vent) then it would be condensed back into its aqueous form.
All of this is just a theory I have and the whole purpose it to find work-arounds. I appreciate telling me why it wont work so I can think about it 😀

 

[schmuckley]

In to see what happens

 

[Gam3r01]

In to see what happens

If all goes well I might make a tutorial on how to do the setup haha

 

[rogue909]

Seems to me you would end up with a figure 8 loop where the center would be the CPU block & separation chamber (most likely a fractional distillation system, check out petroleum distillers).
Water loop
From separation chamber the water would go into a standard water cooling loop; block -> separation chamber -> resevoir -> pump -> radiator -> block
Ammonium Nitrate
the ammonium nitrate would end up going into a phase change loop to take the gas and turn it into a liquid; block -> separation -> compressor -> radiator -> expansion valve -> block

Academically? this all seems awesome and incredibly fun.
Practically? you're building a double loop that could be replaced with a single phase change loop with freon instead of trying to do a chemical reaction and probably achieve better temperatures

If your school will fund it? awesome! If its for yourself? I would be weary.

 

[Gam3r01]

This is all for school, I know there are normal composite coolants that would perform better this is all just a proof of concept.
My school should cover most of it, and if I can get a working concept I have a chance to earn alot of money in scholarships and cash through the Symposium.

 

[rogue909]

Then go for it!
Let us know how it goes, I know I would be interested to see the results

 

[Gam3r01]

So I have had another idea. Rather than use use a custom loop type cooler which would be large, hard to maintain, and expesive, I had another plan.
Take a cpu heatsink such as the stock AMD cooler (which I have a spare one) or an EVO for example. What would happen if I used an endothermic reaction chamber INSIDE the cooling fins of the heatsink? Rather than cool the plate directly with a condencer and block system, shouldnt supercooling the fins of the heatsink produce better temps as well?
Having the compresor mounted on the heatsink with a fan blowing across the fins into the compressor would then cool the cpu, then vent that heat with the heat from reversing the reaction out the case. This would be compact and depending on the heatsink used, smaller than an EVO. Using a shroud around the fins would keep the cool air from filling the case causing condensation, and would isolate the fins in case they condenced the water on them. This would also keep all the cold air/fins where heat is expended.
Opinions on probability of that working?
A good way to think of it would be a 212 EVO with an H60 slapped on the back of it, but instead of a water block, it was a reaction chamber INSIDE the heatsink. This way it would provide airflow across the heatpipes which would be 25C cooler than air, and keep airflow through the condesor then out the back.