Thermal Grease Selection: Conductivity vs Pump-Out
Short answer: Pick thermal grease by balancing bulk conductivity against mechanical stability, not by the headline W/m·K number alone. For most extruded or CNC-machined heat sinks running under 60 °C at the interface, a 3–6 W/m·K silicone or silicone-free paste at 0.05–0.10 mm bond line thickness (BLT) is enough. Above roughly 80 °C junction temperature with repeated power cycling, high-loading pastes above 8 W/m·K often pump out faster because their oil phase migrates; a lower-conductivity but more cohesive paste can outlast them. BQUQ machines heat sink bases to ±0.005 mm flatness targets, which reduces the BLT you need to fill — and that changes which grease wins.
Why the W/m·K Number Is Only Half the Story
Every thermal interface material (TIM) sits in a stack: die or package case, TIM, heat sink base, then the bulk metal. The TIM's job is to displace air, which conducts at roughly 0.026 W/m·K, from the microscopic gaps between two imperfectly flat surfaces. Because the TIM layer is thin — typically 0.03 to 0.15 mm — its thermal resistance matters more than its conductivity.
Thermal resistance is conductivity divided by thickness. A 12 W/m·K paste applied at 0.20 mm BLT performs worse than a 5 W/m·K paste at 0.05 mm. That is why surface finish, flatness, and mounting pressure on the heat sink are not secondary concerns — they determine whether a premium paste ever gets to do its job.
Three variables govern real-world performance:
- Bulk thermal conductivity (W/m·K) — the marketing number.
- Bond line thickness — set by paste viscosity, mounting pressure, and heat sink flatness.
- Mechanical stability over time — whether the paste stays where you put it.
The third variable is where pump-out lives.
What Is Pump-Out and Why Does It Kill High-Conductivity Greases?
Pump-out is the progressive migration of TIM out of the interface, driven by thermal cycling. As the assembly heats, the die and heat sink expand at different rates (CTE mismatch), and the gap between them opens and closes. Each cycle acts like a tiny pump: paste squeezes outward from the center of the interface toward the edges, and it does not fully return on the cold half of the cycle.
Highly loaded pastes — those with 80–90 % metal or ceramic filler by weight — tend to be the worst offenders. They have less silicone or synthetic oil phase to hold the structure together, and the oil that does remain migrates preferentially. Once the filler network separates from the carrier, dry-out follows, and thermal resistance climbs.
The failure signature
Pump-out rarely causes an immediate failure. It shows up as a slow drift: junction temperature rises 3–8 °C over 500 to 2,000 cycles, then accelerates. By the time a field unit trips a thermal limit, the interface may be 40 % depleted.
What makes it worse
- Large CTE mismatch (copper heat sink against a plastic-packaged device).
- High power cycling frequency with large ΔT swings.
- Low mounting pressure, which lets the gap breathe.
- Thick initial BLT, which gives paste more room to travel.
What makes it better
- Higher mounting pressure and a flatter heat sink base.
- A paste with higher oil retention and lower filler loading.
- Phase-change materials or pre-cured thermal pads, which are dimensionally stable but generally higher resistance.
How Do You Match Grease Viscosity to Your Heat Sink Base?
Viscosity determines how thin the paste can be squeezed under a given mounting force. It also determines whether the paste stays put during shipping and vibration.
| Grease class | Typical viscosity | Typical conductivity | Best-fit heat sink | Pump-out risk |
|---|---|---|---|---|
| Low-viscosity silicone | 50–150 Pa·s | 1–3 W/m·K | Stamped or thin extruded, low pressure | Low |
| Mid-viscosity silicone | 150–400 Pa·s | 3–6 W/m·K | Extruded aluminum, spring-clip mount | Low–moderate |
| High-loading silicone | 400–900 Pa·s | 6–10 W/m·K | CNC-machined copper base, high pressure | Moderate–high |
| Silicone-free / hydrocarbon | 100–500 Pa·s | 2–8 W/m·K | Sensitive optics, silicone-averse assemblies | Moderate |
| Liquid metal / metallic | Very low | 20–70 W/m·K | Sealed, non-aluminum, expert assembly only | Very high (alloying risk) |
Note that liquid metal is not a general-purpose grease. It alloys with aluminum and will destroy an extruded aluminum heat sink over time. If you need that conductivity tier, the base must be copper or nickel-plated.
For most extruded heat sinks in LED lighting, power supplies, and industrial electronics, the mid-viscosity band is the practical sweet spot. It wets a machined or milled surface well, tolerates 20–50 psi mounting pressure, and does not pump out under normal duty cycles.
Conductivity vs Pump-Out: A Practical Decision Table
Use this as a first-pass screen. The resistance figures are indicative for a 0.06 mm BLT at moderate pressure.
| Application profile | Interface temp | Cycling severity | Recommended direction | Indicative TIM resistance |
|---|---|---|---|---|
| LED street light, extruded sink | 45–65 °C | Low, slow diurnal | 2–4 W/m·K, mid viscosity | 0.15–0.30 °C·cm²/W |
| Industrial PSU, forced air | 60–85 °C | Moderate | 4–6 W/m·K, mid-high viscosity | 0.10–0.20 °C·cm²/W |
| IGBT module, black anodized sink | 80–110 °C | High, PWM | 6–8 W/m·K, high oil retention | 0.08–0.15 °C·cm²/W |
| Dense server CPU, copper base | 70–95 °C | High, frequent | 8–12 W/m·K or phase change | 0.05–0.12 °C·cm²/W |
| Outdoor telecom, wide ΔT | −40 to +85 °C | Severe | Phase-change or pad, not grease | 0.15–0.35 °C·cm²/W |
The pattern: as cycling severity rises, the optimal conductivity stops rising. Past a certain point you are trading long-term stability for a number that only holds on day one.
How Does Heat Sink Flatness Change the Grease You Need?
This is the part most TIM selection guides skip, and it is the part a manufacturer can actually control. A heat sink base that is flat to 0.02 mm across the contact area needs less paste to fill the gap than one that is flat to 0.10 mm. Less paste means thinner BLT, which means lower resistance — and less material available to pump out.
BQUQ machines heat sink bases on CNC equipment holding ±0.005 mm on critical features, and we can specify flatness and surface roughness (Ra) on the drawing. If your interface is 0.05 mm BLT instead of 0.12 mm, you can often drop one conductivity tier and still hit the same junction temperature, while gaining pump-out margin.
Three manufacturing levers matter:
1. Base flatness — controls BLT directly.
2. Surface roughness — a slightly textured surface (Ra 0.4–0.8 µm) holds paste better than a mirror finish, which can starve the interface.
3. Mounting hole position and spring pressure — even pressure distribution prevents the gap from breathing at one corner.
If you are still deciding how thick the base should be, our guide on heat sink base thickness covers the trade-off between spreading resistance and thermal mass.
Application Method: Where Most Pump-Out Is Actually Born
Field failures attributed to "bad grease" are frequently application failures. The most common mistakes:
- Too much paste. A line or small dot that spreads under pressure is correct. A thick hand-spread layer guarantees a 0.15 mm+ BLT.
- Inconsistent pressure. Hand-tightened screws produce uneven BLT across the interface.
- Contaminated surfaces. Oils and oxide layers raise contact resistance before the paste is even applied.
- No burn-in. Most pastes need one or two thermal cycles to reach steady-state BLT. Test after burn-in, not before.
For assemblies where the TIM choice interacts with the mechanical design, our article on thermal interface selection walks through the broader decision between grease, pads, and phase-change materials.
When Should You Abandon Grease Entirely?
Grease is not always the answer. Move away from it when:
- The assembly must survive 5,000+ power cycles with wide ΔT.
- The gap between surfaces exceeds 0.3 mm — a gap filler or pad is better.
- Field service is impossible and rework cost is high.
- The product is exposed to vibration that would migrate any viscous fluid.
In those cases, pre-cured thermal pads, phase-change materials, or a properly specified gap filler will beat even a premium grease on lifetime cost, even though their day-one resistance is higher.
What Should Be on Your TIM Spec Sheet?
A useful spec sheet for grease does not stop at W/m·K. Ask for:
- Bulk conductivity and test method.
- Viscosity and oil-bleed data.
- Volatile content after bake.
- Recommended BLT range.
- Thermal cycling data (resistance change after N cycles).
Our thermal spec sheet guide shows how to structure this so procurement and engineering read the same document.
Frequently Asked Questions
Q: Is higher thermal conductivity always better for thermal grease?
A: No. Higher conductivity usually means higher filler loading, which reduces the oil phase that holds the paste together. Under thermal cycling, these pastes pump out and dry faster. For interfaces above roughly 80 °C with frequent cycling, a 4–6 W/m·K paste with good oil retention often outperforms an 8–12 W/m·K paste over the product's life.
Q: What bond line thickness should I target for thermal grease?
A: Target 0.05–0.10 mm for most electronics. Below 0.03 mm you risk an incomplete fill and dry contact spots. Above 0.15 mm the paste's own resistance dominates. Achieving a thin BLT depends on heat sink flatness and mounting pressure as much as on the paste itself.
Q: Can I use thermal grease with an aluminum heat sink?
A: Yes, with silicone or hydrocarbon-based greases. Avoid liquid metal and any gallium-based TIM on aluminum — it alloys with the aluminum and degrades the interface. If you need liquid-metal-tier conductivity, the heat sink base should be copper or nickel-plated.
Q: How do I test for pump-out before shipping a product?
A: Run a power-cycling test that reproduces your field ΔT and cycle count, measuring junction-to-sink resistance at intervals. A rise of more than 10 % from the post-burn-in baseline indicates pump-out. Accelerated tests at higher ΔT can compress months of field life into weeks.
Q: Does BQUQ supply heat sinks with a specified flatness for TIM use?
A: Yes. BQUQ machines heat sink bases to ±0.005 mm on critical features and can hold flatness and surface roughness callouts on the drawing. Tighter flatness lets you use a thinner bond line, which often means a lower-cost grease still meets the thermal target. Quotes in 12 working hours.
Related Resources
- About BQUQ and our Dongguan factory: /about/
- Heat sink product range: /heat-sinks/
- Extruded heat sink profiles: /extruded-heat-sinks/
- CNC-machined heat sinks with tight base flatness: /cnc-machined-heat-sinks/
- Industry trends in thermal management: /industry-dynamics/
- Technical articles and engineering guides: /bquq-blog/
- Contact our engineering team: /contact/
Authored by the BQUQ Engineering Team. BQUQ (Dongguan) runs CNC machining (±0.005 mm), metal stamping, custom springs, and heat sink production in one ISO9001 factory. Source-direct from Dongguan, China — quote in 12 hours: sc@bquq.com | WhatsApp +86 13713157787 | www.bquq.com
