Figure 1. Dried thermal paste loses its contact with the surfaces - and the component slowly runs hotter.
If fans, airflow and load are unchanged, a common cause is degradation of the thermal paste between the chip and the heat sink. At sustained high temperature, the oil carrier in a paste can separate from the thermally conductive filler, flow out of the interface or evaporate. What remains is a drier, harder layer that no longer wets the surfaces well, so thermal resistance rises and the chip runs hotter. This is called dry-out. It is tested with long high-temperature storage (for example 125–150°C for 1,000 hours or more) while tracking thermal resistance and weight loss.
A server, inverter or LED driver that ran comfortably at launch slowly gets hotter. Nothing in the cooling system has changed. Eventually the component throttles, a margin alarm trips, or field failures start to appear.
One of the most common causes is the thinnest part of the thermal path: the paste layer. This article explains how dry-out develops, how to confirm it, and how to separate it from other causes that look the same from the outside.
How Dry-Out Happens
Figure 2. Fresh paste wets both surfaces. After long exposure to heat, the oil carrier separates and evaporates, leaving a dry layer with micro-gaps - and higher thermal resistance.
A thermal paste is a thermally conductive filler dispersed in an oil carrier. The carrier lets the paste flow into surface roughness and keeps it in close contact with both surfaces. Over time at high temperature:
The carrier separates. Oil migrates away from the filler and can flow out of the interface first, leaving a filler-rich layer behind.
Volatile content evaporates. Lighter fractions leave as vapor, which shows up as weight loss.
The layer stiffens. With less carrier, the paste becomes dry, crumbly or hard and stops conforming to the surfaces.
Contact resistance rises. Micro-gaps open between the dried layer and the surfaces, and the component temperature climbs.
Humidity can also change the thermal resistance of some pastes, depending on the filler and carrier system.
Symptoms
A slow, steady temperature rise over months at the same load and ambient conditions.
On disassembly: dry, crumbly, cracked or hard residue across the interface, sometimes with an oily ring at the edges where carrier has migrated.
The heat sink lifts off easily, with little paste sticking to either surface.
Temperature recovers after cleaning the surfaces and applying fresh paste.
Rule Out the Look-Alikes First
| Cause | How to check | What you find |
|---|---|---|
| Cooling-side problem | Fan speed, filters, fin fouling, coolant flow, ambient | Heat sink itself runs hotter; interface temperature drop unchanged |
| Loss of clamping force (parts loosening over time) | Fastener torque, spring clips, warped brackets | Loose fasteners; paste may look normal |
| Pump-out | Disassembly; cycling history | Paste still soft but moved - thin center, excess at edges |
| Dry-out | Disassembly; time at temperature | Paste hard or crumbly across the interface |
If you can measure both the case (or junction) temperature and the heat sink base temperature, the difference between them isolates the interface. A growing difference with a stable heat sink temperature points to the interface; a rising heat sink temperature points to the cooling side.
How to Test for Dry-Out Before Field Use
High-temperature storage (bake). Age assembled samples or test fixtures at an elevated temperature above normal operation - published reliability reports commonly use 125–150°C for 1,000–2,000 hours.
Track thermal resistance at time zero and at regular intervals, using the same measurement method and pressure throughout.
Track weight loss of the material - an indicator of evaporation and dry-out.
Inspect at the end. Check the paste for hardening, cracking and separation, and note any oil migration at the edges.
Add damp heat (for example 85°C / 85% RH) if the product sees humidity.
Compare candidate materials under identical conditions. The trend over time - not the starting value - is what predicts field behavior.
How to Reduce Dry-Out
Know the real interface temperature. Dry-out accelerates with temperature. Measure or simulate the interface temperature at worst-case load and ambient, and select a material with long-term data at or above it.
Ask for aging data, not just initial values. High-temperature storage results with thermal resistance over time tell you more than a conductivity number.
Look at volatility and oil bleed. Lower values mean less carrier lost from the interface.
Keep the layer thin and well clamped. A controlled bond line under stable pressure keeps contact even as the material ages.
Plan maintenance where possible. In serviceable equipment, define when the interface is inspected and reworked.
Next Step: Qualify Materials on Long-Term Data
If dry-out is showing up in your product, the fix is usually a material qualified on aging performance - whether from your current supplier or a second source. See How to Qualify a Second-Source Thermal Interface Material.
Temperatures Rising in the Field?
Share your interface temperature, operating hours and what the paste looked like on disassembly. Our engineers can help you narrow down whether dry-out is the cause.
References
Electronics Cooling - Reliability Testing of Thermal Greases
NREL - Degradation Characterization of Thermal Interface Greases (preprint)
Laird - Tflex CR350S Thermal Reliability Report (example of isothermal bake testing)
Published by: Fong Yong Chemical Co., Ltd. - specialty chemical and thermal management materials manufacturer, Taiwan. Established 1983.
