A Semiconductor is for Ever...

[Althornin]

I was thinking along the lines of the diamond actually cutting into the metal, creating a larger total contact surface than is possible with a "softer" paste, and would obviously work best if the diamond crystals themselves were in contact with both surfaces.

I suppose you wouldn't want to also sacrifice the thermal conductivity of the paste itself, so you wouldn't want them to be too big, and you definitely wouldn't want the heatsink to move once it is placed (even a fraction of a micron could drastically reduce the conductivity, though it *might* be recovered if the heatsink/paste settles...).

And this would definitely require exacting specifications on actually placing the heatsink. Might be better for factory-applied heatsinks than user-applied ones...

do you know the pressure required to deform copper?
the yield strength of rather pure (soft) copper is nominally 9000psi (62MPa) at 0.2% offset, or 0.5% extension under load
a 120mm² die (~0.186in²) with a 24lbf load results in an applied load of ~130psi (900kPa), which is approx. spec for AMD procs - and is still far far below the required force. In addition, copper used in HS units is not that pure. Sure, the force will be greater on the points of the diamond (smaller surface area) but that will balance out before full "penetration" occurs, and then you've just created a longer thermal gradient and reduced efficiency - and the above data indicates that "penetration" will be miniscule before the pressure is too weak to deform the copper.
besides, how are you going to align the diamond particles so that the flat sides all sit on the silicon die and the sharp points dig into the copper? And even if you do, then you've just got another thermal joint to deal with - the diamond to silicon one. You've gained nothering, and added a thermal joint, and created a longer thermal path!

Research using small diamond partcles under 100psi (roughly at rated HSF force) still shows that they lose to carbon black - did you even bother to read the research i presented?

Its an unworkable idea.

 

[Simon F]

Read that, then go and download her research here:
http://zzz.com.ru/chung.zip
then read the paper, and understand.

Hmmm. Handle with care?

 

[KimB]

do you know the pressure required to deform copper?

The pressure required to deform copper would merely set the amount of the paste that would contain diamonds.

And as for alignment, I would expect that the crystals would cut into both metals. After all, it would be most likely that a sharp side would (at first) contact both materials, such that when pressed together, either the crystal cuts into one of the materials, or the crystal will be rotated by the pressure so that a flat side is in contact with both. Either way is good for contact.

Edit:
Oh, and don't forget that that study only examined one specific size of diamonds in the paste.

Btw, I don't doubt that the carbon black is an excellent paste due to its properties. It just operates on a different idea than including diamond crystals.

 

[Althornin]

do you know the pressure required to deform copper?

The pressure required to deform copper would merely set the amount of the paste that would contain diamonds.

And as for alignment, I would expect that the crystals would cut into both metals. After all, it would be most likely that a sharp side would (at first) contact both materials, such that when pressed together, either the crystal cuts into one of the materials, or the crystal will be rotated by the pressure so that a flat side is in contact with both. Either way is good for contact.

Edit:
Oh, and don't forget that that study only examined one specific size of diamonds in the paste.

Btw, I don't doubt that the carbon black is an excellent paste due to its properties. It just operates on a different idea than including diamond crystals.

I'm sorry, but you missed the point.
your diamond idea is unworkable. You dont seem to understand that anything that seperates a direct core to HSF contact is BAD. Your diamond idea does that, and has a bunch of other problems, most of which you seem to ignore.

As per your edit - yes, and it examined the smallest sized diamond particles that are feasible. Smaller is better. I'd suggest you do a little more research. The entire basis of your idea is simply flawed.

Simon - yep, but hey, if it shaves off a few degrees, great.

 

[Simon F]

Simon - yep, but hey, if it shaves off a few degrees, great.

Speaking of risky chemicals, the paper said that this new paste is superior to solder which was the leading thermal material.

I remember reading that, in mercury tilt switches (i.e. where you have a small glass vessel with two electrical contacts and a blob of mercury), when the mercury covers the contacts, the electrical resistance is lower than if the contacts had been soldered together.

IIRC, given that heat is generally transmitted in metals by electron movement, I wonder how well using mercury between the metal layers would work. (I could just imagine a new phrase "Mad as an Overclocker" coming into fashion if it did function well.)

 

[nelg]

IIRC, given that heat is generally transmitted in metals by electron movement, I wonder how well using mercury between the metal layers would work. (I could just imagine a new phrase "Mad as an Overclocker" coming into fashion if it did function well.)

Great, then we could get bigger overclocks and bigger testicles all at the same time.

 

[Althornin]

Simon - yep, but hey, if it shaves off a few degrees, great.

Speaking of risky chemicals, the paper said that this new paste is superior to solder which was the leading thermal material.

I remember reading that, in mercury tilt switches (i.e. where you have a small glass vessel with two electrical contacts and a blob of mercury), when the mercury covers the contacts, the electrical resistance is lower than if the contacts had been soldered together.

IIRC, given that heat is generally transmitted in metals by electron movement, I wonder how well using mercury between the metal layers would work. (I could just imagine a new phrase "Mad as an Overclocker" coming into fashion if it did function well.)

electrical conductivity is a decent indicator of thermal conductivity, normally.
let me google mercury.
Thermal conductivity [/W m-1 K-1]: 8.3
thats pretty good.

this study suggests that it would perform quite well, as the main criterion is spreadability and small particle size. People have found in many tests that water is a superior TIM to every paste currently on the market - it just evaporates quickly. This is due to both the decent thermal conductivity of water and its incredible "spreadability".

 

[nelg]

People have found in many tests that water is a superior TIM to every paste currently on the market - it just evaporates quickly. This is due to both the decent thermal conductivity of water and its incredible "spreadability".

I wonder how glycerin would do. I remember a product photographer telling me how they used it in place of water to spray on foods in the studio because it does not evaporate. IIRC he used Kodak Photo Flow.

 

[Dio]

IIRC, given that heat is generally transmitted in metals by electron movement, I wonder how well using mercury between the metal layers would work.

Thermal conductivity [/W m-1 K-1]: 8.3
thats pretty good.

It's a low-temperature equivalent of liquid sodium...

 

[Gubbi]

People have found in many tests that water is a superior TIM to every paste currently on the market - it just evaporates quickly. This is due to both the decent thermal conductivity of water and its incredible "spreadability".

I wonder how glycerin would do. I remember a product photographer telling me how they used it in place of water to spray on foods in the studio because it does not evaporate. IIRC he used Kodak Photo Flow.

I once used olive oil instead of thermal grease on an Athlon XP (1466Mhz). To my surprise real thermal grease only reduced the temperature 1.5 degree (centigrade). The real bonus is that synthetic grease doesn't decompose.

Cheers
Gubbi

 

[KimB]

When using a liquid like olive-oil, I'd be more worried that it would run out from between the CPU and heatsink, leaving you with no "contact help" after some amount of time.

 

[Simon F]

I once used olive oil instead of thermal grease on an Athlon XP (1466Mhz).

We've fried chips with olive oil - they taste quite nice.

 

[Mariner]

I once used olive oil instead of thermal grease

Extra-virgin or standard? 8)

 

[Bambers]

Hmmm. Handle with care?

bah, the harmful label gets stuck on pretty much everything and it means pretty much nothing. Well actually it means don't eat/sniff it or stick it in your eyes/ears but then you don't want to go around doing the same with oil, ink, hand cream, glass or power transformers for that matter. such a lable's function is soley to protect the legal arses of schools/companies from the terminally stupid.

Carbon black could become dangerous if mixed with powdered potassium nitrate (/remembers A level chemistry practicals where activated charcol, potassium nitrate and powered magnesium were all out in a single lesson )

 

[Gubbi]

When using a liquid like olive-oil, I'd be more worried that it would run out from between the CPU and heatsink, leaving you with no "contact help" after some amount of time.

I justed dipped the tip of my finger in some oil, and then touched the top of the chip package, way less than a drop, but it lowered temperature around 12 degree centigrade compared to dry contact.

But then again it was some of the finest olive oil you can get (Frantoio Portofino extra virgin).

Cheers
Gubbi

 

[The Lamb]

Mercury has a high surface tension, so it wouln't make very good contact between metallic surfaces without a fluxing or wetting agent to counteract it along with any grease or dust that the chip and mercury pick up.
Mercury is also quite volatile and has a low viscosity, so it readily evaporates - leaving you with a burnt out chip, and a strange sense of humour!