Does thermal paste expire? 20-year-old TIM tested - no degradation seen, but poor compared to modern products

old thermal paste test
(Image credit: Igor's Lab)

A syringe of ancient thermal paste has been tested by Igor’s Lab. With some surprise, the hardware testing stalwart reports that Arctic Cooling’s Silicone Paste worked just as well as a factory fresh industrial alternative with a similar formulation. However, Wallossek’s testing clarified that modern PC thermal pastes aimed at PC enthusiasts perform much better.

A PC hardware dealer in Germany discovered some 20-year-old thermal paste at the back of a drawer and decided to send it to Igor’s Lab. Wallossek looked at the Intel Pentium 4-era paste from Arctic Cooling and thought it was like finding an old can of soup – but one that “can’t spoil.” Without much hesitation, testing began.

The initial finding was that, since the syringe had a good seal, the paste squeezed out as it should, with an acceptable consistency.

The ingredients also piqued Wallossek’s interest. According to the packaging, Arctic Cooling's old Silicone Paste featured 50% silicone, 20% carbon, and 30% metal oxide. This required verification, decided the veteran hardware tester and ex-Tom’s Hardware writer. So, he employed a laser-induced breakdown spectroscopy device to analyze the paste’s ingredients. Wallossek reckons that the ingredients on the label were bunk, as the analysis tool reported the paste was made from 50% silicone and 50% zinc oxide. He concluded the mention of carbon on the packaging was solely for marketing purposes…

Moving onto the comparison testing, Igor’s Lab noted that a TCTG-4.0 from MCT, an industrial “durability-focused reference paste,” performs very similarly to the antique Arctic Cooling paste. This indicates that little, if any, degradation occurred during the old paste’s slumber at the back of the PC hardware dealer’s drawer.

Alphacool Apex

(Image credit: Alphacool)

The Pentium 4 era paste significantly underperformed compared to a modern thermal paste, though. Wallossek tested an Intel Core i9-13900K with both the old Arctic paste and the “modern, good” Alphacool Apex thermal paste. The graph of the comparative test run (below) shows a delta of six degrees Celsius, representing 20 years of progress in thermal paste formulation. It's not astonishing, but a very welcome improvement.

(Image credit: Igor's Lab)

From the above tale, we might conclude that thermal paste packaging that features ‘use by’ or ‘best before’ advice from manufacturers might be wise to ignore.  This may only apply if the paste is kept well, like in a sealed tube in a dark, cool drawer.  I just checked my drawer and found three partly-used thermal paste syringes, and none of them seemed to have any ‘use by’ date indicated on the packaging.

For more thermal paste comparisons, look at our colossal roundup in our Best Thermal Paste for CPUs 2023: 90 Pastes Tested and Ranked feature.

Mark Tyson
News Editor

Mark Tyson is a news editor at Tom's Hardware. He enjoys covering the full breadth of PC tech; from business and semiconductor design to products approaching the edge of reason.

  • artk2219
    It makes sense, there's not really anything in it that "goes bad" so long as it has a good seal and in a decent temperature range.
    Reply
  • bit_user
    I expected this to be an article about whether TIM degrades after being applied. I've never had to re-apply TIM, but I've been using Arctic MX-4 the longest - it's said to have good durability.
    Reply
  • COLGeek
    This reminds me, I need to clean out my parts pins. No telling what relics I have in them...
    Reply
  • helper800
    With no active ingredients, radioactive decay, or dehydration of the paste it should last for many decades if not centuries... I still have a bulk tube of arctic silver 5 from 2010 or so I have used on many people's computers with zero issues.
    Reply
  • HaninTH
    Without a good seal, I imagine the solvent would evaporate over time, but as noted, the constituent parts of the TIM should not separate nor expire.
    Reply
  • Alvar "Miles" Udell
    I can say from experience that TIM -CAN- and -DOES- expire, I had a tube of IC Diamond 7K (the small tube) from 2019 properly stored and after about two years it went bad, essentially providing zero thermal conductivity and causing my 5950X to overheat under the lightest stress, even after multiple proper cleanings and reapplications. A new tube of Arctic MX-5 solved the issue.

    The more "traditional" formulations, like AS5, likely will have no issues for years as long as the tube is closed and is stored properly.
    Reply
  • helper800
    Alvar Miles Udell said:
    I can say from experience that TIM -CAN- and -DOES- expire, I had a tube of IC Diamond 7K (the small tube) from 2019 properly stored and after about two years it went bad, essentially providing zero thermal conductivity and causing my 5950X to overheat under the lightest stress, even after multiple proper cleanings and reapplications. A new tube of Arctic MX-5 solved the issue.

    The more "traditional" formulations, like AS5, likely will have no issues for years as long as the tube is closed and is stored properly.
    Do you know what the general formulation of the IC diamond 7k is, or at least know anything more about it as to the reason it "went bad?"
    Reply
  • Alvar "Miles" Udell
    I don't, but in
    helper800 said:
    Do you know what the general formulation of the IC diamond 7k is, or at least know anything more about it as to the reason it "went bad?"
    I don't, but in an older version of TomsHardware's thermal paste roundup articles they tested both the IC Diamond 24K and 7K (same thing, different sizes), and they noted a higher temperature with the 24K vs the 7K which they attributed to age due to lower stock turnover. Also I have noticed that Amazon and Newegg both no longer carry IC Diamond, and that in the latest thermal paste roundup article from TH it has IC Diamond performing basically the same as AS-5, when I know I remember it being far better (but TH seems to have removed that test as it was not on their 2013 version).

    Anyway, with MX-5 being a thing, there's no reason to even consider IC Diamond anymore.

    Reply