Thermal Paste
Also known as: thermal compound, thermal grease, TIM, thermal interface material, CPU paste, heatsink compound
Thermal paste (also called thermal compound, thermal grease, or thermal interface material/TIM) is a substance applied between a processor (CPU or GPU) and its heatsink or cooling solution.
Both the processor lid and heatsink surface have microscopic imperfections: air pockets in these gaps are poor heat conductors. Thermal paste fills these gaps, dramatically improving heat transfer from the chip to the cooler. Without thermal paste, air gaps cause the processor to overheat. Over time (typically 3-7 years), thermal paste dries out or degrades, reducing its effectiveness: a common cause of gradually increasing temperatures and thermal throttling in aging computers.
Thermal Paste, explained at three levels
At home
Thermal paste works like a filler for tiny gaps. If you press two solid metal surfaces together (like a processor and a heatsink) they look flat but are actually covered in microscopic peaks and valleys. Air gets trapped in those gaps, and air is a terrible conductor of heat. Thermal paste fills these gaps with a material that conducts heat much better than air, so heat flows efficiently from the hot processor to the cooler heatsink. Without it, the processor would overheat and either slow down (throttle) to protect itself or potentially be damaged. Most processors come with thermal paste pre-applied to the included cooler.
At work
Thermal paste degradation is a common cause of computer performance problems in systems 3-5 years old. As thermal paste dries out, heat transfer efficiency decreases, causing the processor to run hotter: triggering thermal throttling (automatic reduction of processing speed to reduce heat). Symptoms: a computer that used to be fast now slows down under load; fan speeds are much higher than when the computer was new; temperature monitoring software shows CPU/GPU temperatures near or above thermal limits. Replacing dried thermal paste (reapplication is a straightforward repair) often restores full performance in affected systems.
At home
Thermal conductivity rating (W/mK): thermal pastes are rated by their thermal conductivity in watts per metre-kelvin (W/mK). Higher W/mK means better heat transfer. Standard included thermal pastes achieve 4-8 W/mK; high-performance pastes (Thermal Grizzly Kryonaut) achieve 12-14 W/mK; liquid metal compounds (gallium-indium) achieve 38-73 W/mK. The real-world temperature difference between a standard included paste and a premium compound is typically 3-8°C: meaningful for overclockers and enthusiasts but not critical for most users. Application technique (even coverage, correct quantity) has more impact on temperatures than compound selection for average users.
At work
Thermal paste is a suspension of conductive filler particles (zinc oxide, aluminum oxide, silver, or carbon) in a silicone or synthetic oil base. The oil lets it flow into microscopic valleys under mounting pressure, while the filler carries heat across the gap. Performance depends on bond line thickness: a layer tens of micrometres thick conducts best, so excess paste adds resistance and too little leaves air pockets. Over years, heat cycling pumps paste out from the centre and dries the oil, so temperatures creep up.
At home
Die-direct vs. integrated heat spreader (IHS) thermal architecture: most desktop processors have an Integrated Heat Spreader (IHS): a metal lid soldered or affixed above the processor die, providing a flat, durable surface for heatsink contact. Thermal paste is applied between the IHS and heatsink. Some enthusiasts 'delid' processors (removing the IHS) to apply thermal paste or liquid metal directly to the processor die, bypassing the original thermal paste between die and IHS: achieving temperature reductions of 10-30°C in some cases. AMD's desktop Ryzen processors have mostly used solder between die and IHS since the first generation (2017), making delidding unnecessary; many mainstream Intel chips from 2012 to 2018 used paste under the IHS, which made delidding popular with extreme overclockers, though Intel's high-end desktop chips have been soldered again since 2018.
At work
Thermal interface materials in server and embedded hardware: enterprise servers use a variety of TIMs beyond traditional paste: phase-change materials (solid at room temperature, liquefying at processor temperatures to conform perfectly to surface irregularities), thermal pads (pre-cut, pre-applied solid TIM for volume manufacturing), and indium solder (directly bonding die to heat spreader in high-performance applications). Server TIMs must maintain performance over 5-7 year operating lifetimes and 24/7 operation: a different requirement from consumer TIMs designed for shorter operating cycles. Thermal pad vs. paste choice in embedded designs affects manufacturing process (pads are more consistent and automatable) and performance (paste typically slightly outperforms pads at the same material specification).
Real-world examples
At home
Sophie's 4-year-old gaming laptop started running slowly during games that previously ran smoothly. Temperature monitoring showed the CPU hitting 100°C and throttling. She took it to a repair shop, which cleaned the cooling system (dust in the fan and heatsink) and replaced the dried thermal paste. Post-repair temperatures under full load stayed below 85°C, and game performance returned to its original level. The $40 service fee restored the laptop's performance without any hardware replacement.
At work
Marcus managed a fleet of 200 workstations deployed in a manufacturing facility (dusty environment, warm ambient temperatures). After 3 years, field reports showed increasing numbers of workstations with fan speed alarms and performance complaints. He implemented a proactive annual maintenance cycle: cleaning dust filters and heatsinks, and replacing thermal paste on the 20% of units showing the highest temperatures. The maintenance program reduced thermal-related failures (unplanned downtime) by 65% year-over-year and extended the planned refresh cycle by 18 months.
Common confusions
- Many people believe that applying more thermal paste improves heat transfer — Thermal paste fills microscopic gaps between surfaces; if there are no gaps, additional paste adds nothing and can cause problems. The ideal thermal paste layer is as thin as possible while completely covering the contact area: excess paste squeezes out from under the heatsink and can flow onto circuit components around the CPU socket (potentially causing short circuits with electrically conductive compounds). Thermal paste should be used in the minimum quantity needed for full coverage: a pea-sized dot (approximately 3-4mm diameter) in the centre of the CPU is typically sufficient for most desktop processors. More paste does not mean better cooling: it means more mess and potential contamination.
- Some assume that any grease or compound can substitute for thermal paste in a pinch — Thermal paste is specifically formulated for high thermal conductivity and stability at CPU/GPU operating temperatures (typically 50-100°C continuous). Common substitutes that do not work: regular white petroleum grease (petroleum decomposes at CPU temperatures, thermal conductivity is essentially zero); toothpaste (abrasive, dries rapidly, poor thermal conductivity); butter or food greases (decompose, attract pests). Even temporary substitutes can cause sustained overheating damage. The only practical substitute for proper thermal paste is a thermal pad designed for heatsink-to-die contact. Purpose-made thermal paste is inexpensive (a few dollars provides multiple applications) and the correct solution.
For analysts
Thermal paste maintenance is a relevant consideration in hardware lifecycle management for any client with desktop computers, workstations, or servers past 3-4 years of age. Including thermal paste reapplication in preventive maintenance schedules (along with fan cleaning and dust removal) prevents thermal throttling that reduces productivity without any hardware fault being obvious to the user. For client environments with servers or workstations operating in high-temperature or dusty environments, annual thermal maintenance cycles provide measurable protection against thermal-related failures. The cost is minimal; the downside of neglect is gradual performance degradation that may be misdiagnosed as software or hardware obsolescence.
History & origin
The concept of using conductive compounds to fill surface gaps in thermal assemblies predates modern computers: it has roots in industrial heat transfer engineering. Early computer cooling used dry metal-to-metal contact or simple greases. Indium foil (a soft, highly thermally conductive metal) was used in some early IBM mainframes. Consumer-grade thermal compounds became standardized as Intel and AMD's processors exceeded 50W TDP in the late 1990s, requiring effective heatsink mounting. Arctic Silver 5 (released in the early 2000s) became a long-running standard enthusiast thermal compound known for high conductivity. Modern thermal pastes range from simple white silicone-based compounds (included with most coolers and processors) to high-performance metal-based and carbon-based compounds (Thermal Grizzly Kryonaut, Conductonaut) used by enthusiasts. Liquid metal thermal compounds (primarily gallium-indium alloys) achieve the highest conductivity but require careful application due to electrical conductivity.