The Official Thermal Interface Material thread

[SpittinFax]

Let's get into the semantics of what "pump out" means with TIM right here
I agree with the OP of that thread, its a nonsense term.

Pump out is definitely a nonsense term. We can debate until the end of time about pump out because it involves too many variables that are impossible to quantify. Application life, dryout (chemical stability), heat cycle duty, the type of cooler used and application technique all have some kind of impact on the concept of "pump out". It's just too unclear and it muddies the water.

I've seen many conversations on pump out and they always lead nowhere, with the conclusion always being something like "well it can't be tested, but it exists, so it's an important consideration and I'll argue with anyone who disagrees". Waste of time to even talk about it.

 

[delshay]

I'm not an expert, but this is how I look at thermal paste.

When it dry's out, I look at what's left behind filling in the tiny gaps. Those tiny material/particles must now do it's work when the moisture is all gone. This is where I think the material that's filling in the gaps is now important. So here you want the best material that conducts heat.

After Diamond, "Silver" is the next best conductor of heat, this is why I have stuck with Arctic Silver 5, It contains real particles of silver tested in a Lab many years ago. It's the performance when moisture is no longer available. This part is important to me. My laptop has Arctic Silver 5 & i applied it about 8 or 9 years ago with no real issues.

TLDR see link below

Top 10 Thermally Conductive Materials (thermtest.com)

 

[Shrek]

But silver can oxidize, so I prefer something like zinc oxide or aluminum nitride which is done reacting.

Artic silver 5 is slightly capacitive
Arctic Silver Incorporated - Arctic Silver 5

 

[freeagent]

So it’s slightly capacitive.. it has been that way since the early 2000s.. I don’t recall hearing anyone who lost parts because of sloppy use of AS5.

 

[lexluthermiester]

But silver can oxidize, so I prefer something like zinc oxide or aluminum nitride which is done reacting.

Not in the formulation used in AS5. Zinc Oxide or Aluminium Nitride do not conduct heat as well as Silver.

Artic silver 5 is slightly capacitive
Arctic Silver Incorporated - Arctic Silver 5

As long as it is not applied like a monkey flinging it's poo, that is not a problem.

 

[Blaeza]

Arctic MX2, good or bad? Repasted my 1660 super yesterday and it seems alright so far.

 

[GerKNG]

Arctic MX2, good or bad? Repasted my 1660 super yesterday and it seems alright so far.

not the best performing paste in the world but there is like 2°C between low and high end paste. (MX4 and Kryonaut Extreme is basically identical on my 12700k)
but it's just a 1660 Super and not a 3090.

 

[P4-630]

Using Thermal Grizzly Kryonaut Extreme , the pink bubblegum.
When I was building my new build I accidentally installed the CPU cooler the wrong way around, so had to dissasemble again.
This time had a chance though to see if my CPU IHS was still straight, and it was, no issues.

The pink bubblegum works well for me and the CPU temp doesn't reach 60c during the day, even during gaming mostly just under 50c.
Highest temp I have seen was 71c on one core during 10 minutes multi core Cinebench R23, not bad at all for air cooling a i7 12700K.

BTW I used the "spread" method, and seems to work well.

 

[Blaeza]

That Kryonaut has got me tempted, but I'll be getting NTH1 free with the Noctua NH U12A I've found at a BARGAIN price. Just got to ignore that it's brown.

 

[P4-630]

Noctua NH U12A

Great little 7-heatpipe cooler!! One of the best performing coolers out there.

 

[lexluthermiester]

Arctic MX2, good or bad? Repasted my 1660 super yesterday and it seems alright so far.

Good. Carry on!

but there is like 2°C between low and high end paste.

Exactly. To say something isn't the best in the world when we're talking about differences of 0.8% to 1.3% is a bit disingenuous..

 

[Blaeza]

Great little 7-heatpipe cooler!! One of the best performing coolers out there.

I know and will go well with the 5800X bargain I've found too both brand new too

 

[Shrek]

But silver can oxidize, so I prefer something like zinc oxide or aluminum nitride which is done reacting.

Actually, I'd like to question this!

In the thermite reaction iron oxide gives up its oxygen to aluminum powder, so one must be careful to use the right oxide or things are not 'done reacting';
fortunately the people who make thermal compound know this, unlike stupid me.

 

[P4-630]

I know and will go well with the 5800X bargain I've found too both brand new too

Nice upgrades incoming man!

 

[Blaeza]

Nice upgrades incoming man!

Later this year, need 1440p monitor and RTX 3070/6750XT first. I'm unfortunately not printing money out my buttocks.

 

[Muck Muster]

Artic MX-4. I've used it for years. Never had issues. always stayed cool.
Even used it on a Vega64. In the same die block with the GPU, HBM memory sits next to it, at an uneven height. Gave it a little bit more MX4 to even it out and it ran it OC with temps under 50c. Used a Raijintek Morpheus 2 cooler and the Vega64 ran under 40c.. . using Artic MX-4.

 

[cvaldes]

BTW I used the "spread" method, and seems to work well.

I'm curious. What is the "spread" method?

 

[lexluthermiester]

I'm curious. What is the "spread" method?

Spreading the TIM flat & even across the entire surface of the CPU/IHS. See below.

Noctua NT-H1


Copper Grease


Arctic MX-5

 

[delshay]

Spreading the TIM flat & even across the entire surface of the CPU/IHS. See below.
View attachment 251648View attachment 251649
Noctua NT-H1

View attachment 251650View attachment 251651
Copper Grease

View attachment 251652View attachment 251653
Arctic MX-5

When I apply thermal paste, I want it just cover the shape of the core. Excessive paste where the IHS is not making contact with the cold plate will act as a blanket. The same rule is applied here when it come's to thermal pads. What i'm trying to say here is, the thermal paste on the coldplate should be just over the complete shape of the IHS. Excessive thermal paste on the coldplate where there is zero contact with the IHS effectively is trapping a little heat. Note that everyone has they preferred preference/method which works.

 

[lexluthermiester]

When I apply thermal paste, I want it just cover the shape of the core.

And while that can work, there is something to be said about covering the entire surface of the IHS for maximum surface contact and heat transfer.

Excessive paste where the IHS is not making contact with the cold plate will act as a blanket.

I think there is a misunderstanding. Those photos are showing a very small amount of TIM spread across the surface. In a couple of those photos you can see through the TIM to the surface. This is the way it should be applied. What little excess there is ends up being squished out the sides as the heatsink is secured.

Excessive thermal paste on the coldplate where there is zero contact with the IHS effectively is trapping a little heat.

I've seen no empirical evidence that such takes place...

Note that everyone has they preferred preference/method which works.

Agreed, to each there own. At the end of the day, as long as there is good contact between the CPU & heatsink, thermal performance will be good.

 

[ThrashZone]

Hi,
Yeah spread or large X both work pretty well
.

 

[lexluthermiester]

Hi,
Yeah spread or large X both work pretty well
.

Exactly this. I'd like to see him do this again but with a thicker glass square.

 

[Toss]

I hate paste, all troubles are because of that.
My paste in GPU dried out somehow in 1 year, repasted today, temps from 100C to 70C LUL

 

[ThrashZone]

Hi,
I should of taken a picture of my evga 1080ti ftw3 stock paste after less than 1 year
It looked like a cracked and crushed cookie was there

 

[Pictus]

I like to spread and add a very small bit in the center.
I prefer to apply the paste over the cooler and keep it warm to be able
to proper apply thick pastes, a hair dryer/heat gun is a good option.

Wanted :

1 - Last longer
2 - Lower thermal impedance
3 - Higher Thermal conductivity
4 - No break-in
5 - No curing time

The Thermalright TFX/Zezzio ZT-GX/ZT-GXS(nano diamond) fits the bill.
The Shin-Etsu 7921/Thermalright Tf3 and Honeywell 7950 film
are also very interesting, specially the Honeywell 7950 film
for endurance with direct die(GPU/notebook).

I just bought the Zezzio ZT-GX and will mix with the
Alseye T9+ Platinum Edition to make it a bit less dry.
When the Frankenpaste degrade/lost performance, I will get
the dry paste and try to revive by mixing with a bit of Krytox.

BTW, the new Zezzio ZT-GXS is nano diamond based.
You know, "Diamonds Are a Girl's Best Friend", but if it does not scratch
the CPU/DIE surface, maybe it can be a friend to your PC too.