Can thermal paste degrade with use?
Can thermal paste degrade with use?
I've owned my computer for six years and it's always been running nonstop. I never changed the thermal paste. The CPU typically stays at about 65% during regular use, but when I play games—especially older titles—it hits its limit without ever overheating. It only has one fan, which is located on the power supply unit. I don't believe it degrades over time. The room temperature is usually around 75°F.
Based on my observations, thermal paste can degrade over time. I recently replaced a processor in a self-built 900MHz AMD Athlon T-Bird system. After about four years of use in a hot environment with no air conditioning, it failed. When I removed the processor, I found only a small amount of grease residue on the core sides. I believe temperature fluctuations played a significant role as well, since the device ran very hot—often exceeding 130°F at idle and slightly more under load. I’m not sure about the cost differences between cheaper and more expensive alternatives.
I question if this can be linked to any issue with the paste.
I would suggest reapplying every 2 or 3 years.
Thermal paste is liquid, just like any other liquid, it will gradually evaporate. This evaporation leads to the paste shrinking and cracking over time, which reduces the amount in contact with the heating element and the CPU. Even worse, as the paste dries out it becomes brittle and can form small air bubbles in what remains of the paste. According to the latest information, air is actually a poor conductor of heat.
On a related but completely different subject, I came across some information years ago about testing heat conduction with a single paste. After applying it, removing it, and then using another paste, the results showed that you can't fully remove all the old paste. It seems the best approach is to choose your paste from the start and continue using it.
The issue lies in figuring out if the improvement came from aging paste or poor initial use, unless thorough benchmarking was done at first. It isn't just about temperatures, but also fan speeds and room conditions. A detailed setup and patience would be needed to fully evaluate this.
Well AS5 contains 99.9% pure silver, which clearly handles heat transfer effectively. Even if the remaining 0.1% evaporates, it shouldn’t affect performance. That small amount of liquid is intended to keep the silver dust attached to the CPU until the HSF is applied, preventing it from dispersing before installation.
On another point regarding cooling paste, I’ve noticed a liquid metal like mercury used as a conductor between a CPU and an HSF. The substance I observed wasn’t mercury because that’s toxic, but it was still a liquid metal—something that could work well despite the risks.
I generally reapply my thermal paste each time I perform a thorough cleaning of my computer, usually every six months or so... because removing the HSF forces me to clean and reapply it again.
Based on my personal observations, high-quality thermal paste tends to keep devices running efficiently over time. That’s what I’ve noticed from my own experience. At least so far. When I first got a P4 (a 1.8 Northwood), I saw a slight rise in average temperature after about 18 months of use. I remember upgrading to a P4 3.0-800fsb model in January 2004 with the thermal pad, which likely helped. For my current setup, I use Artic Silver paste. It kept me within 38-40°C during idle and reached up to 50°C during intense gaming sessions like Half-Life2: Episode 1 or video encoding tasks I handle frequently. I often switch between different video formats using multithreaded applications, which increases CPU heat. These readings are definitely accurate because I’m very concerned about overheating. When I stop a game after hours or perform heavy encoding, I always check the CPU temperature first. I also recall being surprised by the jump from 47-48°C to 50°C compared to my older Northwood 1.8 overclocking to 2.4-533—something I now fully understand. My two cents. 😀
the 99.9% silver refers to the purity of the silver, not the amount in the paste. this is often misunderstood. if it were 99.9% silver content, it would appear like silver rather than a grayish paste. i think the paste will remain effective until the computer stops working, after which its performance won't change. mcg