Cooling Chips With Nanotube “Forests”

[HardOCP News]

Researchers say that growing forests of carbon nanotubes onto the surfaces of computer chips will not only work better than conventional cooling but it will be cheaper too.

The two engineers recently have shown how to grow forests of tiny cylinders called carbon nanotubes onto the surfaces of computer chips to enhance the flow of heat at a critical point where the chips connect to cooling devices called heat sinks. The carpetlike growth of nanotubes has been shown to outperform conventional "thermal interface materials."

 

[cyclone3d]

"In a personal computer, laptop and portable electronics, the better your thermal interface material, the smaller the heat sink and overall chip-cooling systems have to be," Cola said.

HAHAHAHAHAHA.

The smaller the heatsink has to be???? What???? So dumping MORE heat into a SMALLER heatsink is going to provide the same level of cooling?

I don't think so.

It MIGHT help keep the chips cooler / reduce hot spots, but it in now way is going to be able to make any significant change in the size heatsinks have to be.

It MIGHT also help reduce sudden temp changes when a chip goes from idle to load.

 

[Dr.Khron]

I can understand your skepticism, but don't be so negative, this is very promising technology. Remember, heat transfer is all about transfer rates, so if a thermal interface can move heat faster then you can reduce the size of the heat sinks. Depending on your application, probably not by much, but its a theoretical possibility.

Thermal transfer is like a chain, and right now the physical interface between the heat sink and chip can be a very weak link... why else would extensive lapping produce such repeatable improvements in heat transfer?

The point of this technology is to be able to produce a really good thermal interface AT A REASONABLE COST. 2 hours of lapping does not come cheap.

 

[cyclone3d]

Having a better "TIM" would be good... but you really can't claim that with a better TIM, you can reduce the size of the cooler by any noticeable amount.

The only real feasable way to reduce the actual "size" of the required cooler (air cooling) is to increase the surface area (thinner/more fins - and they are getting to be pretty thin now days), or by increasing airflow.

If they try to pump more heat into the same or smaller size heatsink, yeah.. it may be cooler for a little while.. until the heatsink gets heatsoaked.

It will probably help with OEM systems.. but that is only because they use total junk to begin with.... In most laptops, they use a pad... just by replacing that pad with cheapo thermal compound, you can drop the CPU temps by 5-10c.

Also.. are you only going to be able to apply the heatsink once with this stuff? Or will it return to the original state when the heatsink is removed?

 

[jstnomega]

Color me overly paranoid, but considering the poor record of PC parts disposal, the idea of *ANYTHING* nano making its way into the environment gives me extreme pause.

 

[deeznuts]

cyclone3d, you provide rebuttals to your own arguments in your very own posts.

you use the example of a laptop, if you reduce the pad with a good TIM, you can reduce temps by 5-10c. Well if you have a certain threshold of heat the laptop can tolerate without ill-effects, and you effectively reduce the temps by 5-10c, well all of a sudden you don't need such a robust cooling system. this may include reduction in size of the HSF.

Not sure what you find so ridiculous or funny.

 

[Jakalwarrior]

Heat travels from hot to cold right? Better thermal interface would keep the heatsink closer to the temperature of the core so it would dissipate heat into the air faster. Essentially speeding up the rate it dissipates heat. I guess the idea of a big heatsink is that it can overcome part of the bad transfer rate by having a higher temperature differential.

Think of touching cold things through a rag vs paper vs bare hand. The Thicker the thing between you and whatever is cold, the colder the thing has to be before it really saps your heat quickly.

 

[Grimlaking]

My worry on this is a carbon fiber nanotube's strength and structual rigidity would be severly impeded by the weight of a heatsink on top of it.

Perhaps they mean more in die than otherwise. Bascially channeling the heat out much like the heat channel heatsinks we have today.

That would be neat carbon fiber nanotube heat channels. Woot!

 

[silent-circuit]

My worry on this is a carbon fiber nanotube's strength and structual rigidity would be severly impeded by the weight of a heatsink on top of it.

Perhaps they mean more in die than otherwise. Bascially channeling the heat out much like the heat channel heatsinks we have today.

That would be neat carbon fiber nanotube heat channels. Woot!

I see no reason why the nanotubes would be weak/fragile. Carbon is tough stuff when tightly packed, and cylinders are very strong structures, though only in one direction. So long as they were boxed in with steel, copper, or even aluminum to prevent being crushed from the sides the nanotubes could probably support a great deal of weight as long as it's a flat surface placed directly atop them... just like a heatsink would be.

 

[Alias2]

Think of it this way. More surface area equals better cooling, roughly anyways. Right? So Trillions of nano tubes have a HUGE surface area compared to a heatsink with fins.

Or am I missing something?

 

[cyclone3d]

cyclone3d, you provide rebuttals to your own arguments in your very own posts.

you use the example of a laptop, if you reduce the pad with a good TIM, you can reduce temps by 5-10c. Well if you have a certain threshold of heat the laptop can tolerate without ill-effects, and you effectively reduce the temps by 5-10c, well all of a sudden you don't need such a robust cooling system. this may include reduction in size of the HSF.

Not sure what you find so ridiculous or funny.

Hehe.. the problem with the thermal pads in laptops is that they have issues with frying stuff.. RAM, MB, CPU, etc. because the thermal pads basically act as an insulator.

When replacing the thermal pad with transfer compound, the heatsink (all of them use heatpipes for the most part now) gets a lot hotter... way hotter than it should ever be. and the CPU still runs a lot hotter than I would like.

The fact of the matter is, most laptops NEED bigger heatsinks, even when replacing the thermal pad with thermal paste. The heatsinks as they are, cannot effectively dissipate the amount of heat that is currently being pumped into them.

If they keep on with the current design trend and incorporate these nanotubes as the TIM, it would only drop temps maybe 1c over cheap thermal compound.

However, if they actually used proper sized/designed heatsinks, they could easily drop 15-20c.

Oh.. and using thermal compound in the first place instead of the crappy thermal pads would be cheaper than using the thermal pads.

And.. it would significantly reduce failure rates.

Have you been working on/building computers for over 16 years?

 

[Nemesis999]

The smaller the heatsink has to be???? What???? So dumping MORE heat into a SMALLER heatsink is going to provide the same level of cooling?

Yes, that's EXACTLY what it will do. Heat transfer rates are related to temperature difference and thermal interface. High effeciency heat transfer from chip to heatsink means the heatsink stays closer to the temperature of the chip. The higher the heatsink temperature the faster it loses heat to the surrounding air.

The net effect is either that ,using the same size heatsink/fan, your chip temps drop and your HSF temps rise, or, using a smaller heatsink/fan, the chip temp stays the same but the heatsink/fan is hotter making transfer to the air more efficient.

 

[Nemesis999]

Yes, that's EXACTLY what it will do. Heat transfer rates are related to temperature difference and thermal interface. High effeciency heat transfer from chip to heatsink means the heatsink stays closer to the temperature of the chip. The higher the heatsink temperature the faster it loses heat to the surrounding air.

The net effect is either that ,using the same size heatsink/fan, your chip temps drop and your HSF temps rise, or, using a smaller heatsink/fan, the chip temp stays the same but the heatsink/fan is hotter making transfer to the air more efficient.

"edit" Think of the heatsink as a bucket with a hole. The more water in the bucket the faster it leaks. Better transfer from the chip to the bucket keeps it fuller making it leak faster.

 

[HAF72]

wow that'll be interesting.