Spreading it Thin TIM Roundup 2007

Cooling/Thermal Compounds by KeithSuppe @ 2007-12-13

Arctic Silver dominated the TIM (Thermal transfer Material) industry for many years. They may have invested more in R&D then all their competition combined. Many ask is there anything else out there? Today we hope to answer this question.

  • next



In the PC industry HEAT is not just the by-product of ever increasing transistor counts and shrinking micro-processes, it’s also our arch nemesis. As the micro-processor industry seems to bear out Moore’s Law a multi-million dollar industry based on what are effectively ad hoc cooling solutions has evolved. One area where the term micro in micro-processing industry haunts us are those micro-pores and striations which exist in all heatsink base and IHS (Integrated Heat Spreader) surfaces. The illustrations below borrowed from the AMD white-paper Thermal Interface Material Comparison (pdf) and Koolance give an example of where these imperfections live.

Madshrimps (c)

The human eye cannot detect imperfections this minute and if left "untreated" you’re either going to experience very high processor temps affecting performance or CPU failure all together. Shortly after micro-processors began to produce more heat than their surfaces could dissipate the first heatsinks were born. At that point the need for a thermal interface material, to literally fill the gap in the new technology, was needed. As this new industry grew, choices and formulas for thermal interface materials were first on the R&D list. Silicon was the minimum standard and then metal based formulas hit the market. Silver in particular not only sounded great to the consumer, its high thermal conductivity made it an ideal ingredient (in small quantities). Of course it's not just conductivity, a TIM (Thermal Interface Material) must also fill surface gaps and hopefully displace the air that would remain without it. Air compared to Silver has a low conductivity as expressed in this formula W/m K = 0.026, while silver = 429.0. Between air and paste containing 10% silver, the latter seems a much better option.

Madshrimps (c)

The TIM industry has been working to develop the ultimate formula by attempting to incorporate those materials with the highest thermal conductivity and an ability to fill gaps. Of course those may be very important, but there are other attributes. AMD's white paper Thermal Interface Material Comparison lists five "Guidelines" or criteria for thermal interface materials. Listed verbatim they are:

• Thermal conductivity of the material
• Electrical conductivity of the material
• Spreading characteristics of the material
• Long-term stability and reliability of the material
• Ease of application

Electrical conductivity is of course high on the list and as we will discover this conundrum for manufacturers has prompted experimentation with many different materials. Albeit silver, silicon, or diamond most manufacturers have to consider all the guidelines if they want to produce a desirable product. Perhaps other criteria might be cost, although I really haven't heard too many PC-Enthusiasts complain once they've discovered how critical a robust thermal interface material is to the life of their processor.

Thermal Pastes Tested:

Madshrimps (c)

1.) Arctic Cooling - MX-2
2.) Arctic Silver - AS5
3.) Arctic Silver - Ceramique
4.) Innovation Cooling - 7-Carat Thermal Compound
5.) Sunbeamtech - Tuniq TX-2
6.) ESG Nanotherm - PCM+

Onto thermal paste formulas ->
  • next
Comment from Rutar @ 2007/12/13
that TX-2 seems to be good stuff, but I think you made an error in the ranking of AS5 and MX2 because AS5 definatly looses for long term stability and electrical safety

Have you tested the difference between the line and the blob method?
Comment from zerotol @ 2007/12/14
is it difference between applying tim with a blob that BIG ??

how many times did u guys test that , thats a huge difference in temps
Comment from jmke @ 2007/12/14
been using blob method for all my CPU HSF reviews last 2 years and it allows for more consistent results compared to spreading it out over the IHS
Comment from Rutar @ 2007/12/14
But there is the method of using a line on intel CPUs or a blob, both without spreading and it has to be solved by scientific testing.
Comment from thorgal @ 2007/12/15
Great little write-up Keith, I learned quite a bit

You've certainly got some attention from other sites as well, even an xtremesystems forum thread about this article
Comment from Arctucas @ 2007/12/15
I see a lot of these TIM reviews, but never see my favorite; Shin-Etsu X23.
Comment from Rutar @ 2007/12/16
Originally Posted by Arctucas View Post
I see a lot of these TIM reviews, but never see my favorite; Shin-Etsu X23.
MX-2 is said to be the same