{"id":1824,"date":"2026-08-05T09:49:46","date_gmt":"2026-08-05T09:49:46","guid":{"rendered":"https:\/\/haktak.com\/?p=1824"},"modified":"2026-08-05T09:49:50","modified_gmt":"2026-08-05T09:49:50","slug":"thermal-grease-pump-out","status":"publish","type":"post","link":"https:\/\/haktak.com\/ja\/thermal-grease-pump-out\/","title":{"rendered":"\u30b5\u30fc\u30de\u30eb\u30b0\u30ea\u30fc\u30b9\u306e\u30dd\u30f3\u30d7\u30a2\u30a6\u30c8\uff1a\u539f\u56e0\u3001\u30ea\u30b9\u30af\u3001\u304a\u3088\u3073\u89e3\u6c7a\u7b56"},"content":{"rendered":"
\n

\u30b5\u30fc\u30de\u30eb\u30b0\u30ea\u30fc\u30b9\u306e\u30dd\u30f3\u30d7\u30a2\u30a6\u30c8\u3068\u306f\u3001\u767a\u71b1\u90e8\u54c1\u3068\u51b7\u5374\u90e8\u54c1\u306e\u5b9f\u63a5\u89e6\u9818\u57df\u304b\u3089\u30b0\u30ea\u30fc\u30b9\u304c\u5f90\u3005\u306b\u79fb\u52d5\u3059\u308b\u73fe\u8c61\u3067\u3059\u3002\u901a\u5e38\u306f\u3001\u81a8\u5f35\u3068\u53ce\u7e2e\u306e\u7e70\u308a\u8fd4\u3057\u3001\u53cd\u308a\u3001\u632f\u52d5\u3001\u307e\u305f\u306f\u63a5\u89e6\u5727\u529b\u306e\u5909\u5316\u306b\u3088\u3063\u3066\u5f15\u304d\u8d77\u3053\u3055\u308c\u307e\u3059\u3002\u3053\u3046\u3057\u305f\u52d5\u304d\u306b\u3088\u3063\u3066\u3001\u8584\u3044\u30b5\u30fc\u30de\u30eb\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30fc\u30b9\u30de\u30c6\u30ea\u30a2\u30eb\u3001\u3059\u306a\u308f\u3061 TIM \u306b\u305b\u3093\u65ad\u304c\u751f\u3058\u307e\u3059\u3002\u30b0\u30ea\u30fc\u30b9\u306e\u914d\u5408\u3068\u30dc\u30f3\u30c9\u30e9\u30a4\u30f3\u539a\u3055\u306b\u3088\u3063\u3066\u3001\u6750\u6599\u304c\u305d\u306e\u5834\u306b\u3068\u3069\u307e\u308b\u304b\u3001\u7aef\u90e8\u306b\u5411\u304b\u3063\u3066\u30af\u30ea\u30fc\u30d7\u3059\u308b\u304b\u3001\u30dc\u30a4\u30c9\u304c\u767a\u751f\u3059\u308b\u304b\u304c\u6c7a\u307e\u308a\u307e\u3059\u3002.<\/p>\n\n\n\n

\"\u30b5\u30fc\u30de\u30eb-\u30b0\u30ea\u30fc\u30b9-\u30dd\u30f3\u30d7-\u30a2\u30a6\u30c8\"<\/figure>\n\n\n\n

\u305d\u306e\u7d50\u679c\u3001\u71b1\u30a4\u30f3\u30d4\u30fc\u30c0\u30f3\u30b9\u306e\u4e0a\u6607\u3001\u30db\u30c3\u30c8\u30b9\u30dd\u30c3\u30c8\u306e\u62e1\u5927\u3001\u30d5\u30a1\u30f3\u30ce\u30a4\u30ba\u3001\u30b9\u30ed\u30c3\u30c8\u30ea\u30f3\u30b0\u3001\u6c5a\u67d3\u3001\u90e8\u54c1\u5bff\u547d\u306e\u77ed\u7e2e\u306b\u3064\u306a\u304c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308a\u307e\u3059\u3002\u4e00\u3064\u306e\u5bfe\u7b56\u3060\u3051\u3067\u89e3\u6c7a\u3067\u304d\u308b\u3082\u306e\u3067\u306f\u3042\u308a\u307e\u305b\u3093\u3002\u8010\u4e45\u6027\u306e\u3042\u308b\u89e3\u6c7a\u7b56\u306b\u306f\u3001\u30b0\u30ea\u30fc\u30b9\u3001\u5857\u5e03\u91cf\u3001\u30af\u30e9\u30f3\u30d7\u30b7\u30b9\u30c6\u30e0\u3001\u63a5\u5408\u9762\u3001\u71b1\u30d7\u30ed\u30d5\u30a1\u30a4\u30eb\u3001\u307e\u305f\u306f\u5225\u306e TIM \u5f62\u614b\u306e\u898b\u76f4\u3057\u304c\u542b\u307e\u308c\u308b\u5834\u5408\u304c\u3042\u308a\u307e\u3059\u3002\u5727\u529b\u3092\u9ad8\u3081\u305f\u308a\u3001\u30da\u30fc\u30b9\u30c8\u3092\u539a\u304f\u3057\u305f\u308a\u3059\u308b\u3053\u3068\u3067\u6539\u5584\u3067\u304d\u308b\u8a2d\u8a08\u3082\u3042\u308a\u307e\u3059\u304c\u3001\u3044\u305a\u308c\u3082\u72b6\u6cc1\u3092\u60aa\u5316\u3055\u305b\u308b\u53ef\u80fd\u6027\u304c\u3042\u308a\u307e\u3059\u3002\u30a2\u30bb\u30f3\u30d6\u30ea\u5168\u4f53\u3092\u8003\u616e\u3059\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002.<\/p>\n\n\n\n

What Is Thermal Grease Pump-Out?<\/h2>\n\n\n\n

A grease interface looks simple: heat source -> thin grease bond line -> heat sink or cold plate<\/strong>. In practice, those parts do not stay perfectly still. Packages, baseplates, cold plates, heat sinks, and PCBs expand by different amounts. Even tiny movement matters in a very thin grease layer.<\/p>\n\n\n\n

During each cycle, the interface is compressed or sheared. Grease can creep from the hottest region, collect near the perimeter, and leave a thin or bare center. Small voids may join, so cooling can be excellent on day one and worse after many cycles.<\/p>\n\n\n\n

Think of a sandwich being flexed over and over. The filling slowly works toward the crust, even though no one deliberately removed it. That is close to the basic pump-out effect. It is progressive movement, not simply too much paste spilling out during installation.<\/p>\n\n\n\n

Pump-Out vs Dry-Out, Squeeze-Out, and Poor Initial Coverage<\/h3>\n\n\n\n

These failure modes are often lumped together because they can leave a similar teardown pattern. They are not the same problem.<\/p>\n\n\n\n

Failure Mode<\/strong><\/td>When It Happens<\/strong><\/td>Typical Pattern<\/strong><\/td>Primary Driver<\/strong><\/td><\/tr>
\u6c72\u307f\u51fa\u3057<\/td>Over repeated operating cycles<\/td>Thin or bare center with edge buildup<\/td>CTE mismatch, warpage, cyclic shear<\/td><\/tr>
\u4e7e\u71e5<\/td>Over time under heat<\/td>Hard, cracked, or filler-rich residue<\/td>Carrier loss, oil bleed, oxidation, separation<\/td><\/tr>
Squeeze-out<\/td>During initial mounting<\/td>Immediate perimeter overflow<\/td>Excess deposit, clamp force, or very low viscosity<\/td><\/tr>
Poor initial coverage<\/td>From first operation<\/td>Local dry area without a migration history<\/td>Deposit pattern, flatness, or mounting error<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Several modes can occur together. Grease may first migrate and later become dry at the exposed edge. A poor mount may also make pump-out easier. This is why one photograph cannot prove the root cause. The temperature history, assembly geometry, and material condition all matter. For general maintenance expectations, the guide to how long thermal paste can be used<\/a> helps separate routine aging from a movement-driven failure.<\/p>\n\n\n\n

How CTE Mismatch and Warpage Pump Grease Out<\/h2>\n\n\n\n
\"How<\/figure>\n\n\n\n

Different Materials Expand by Different Amounts<\/h3>\n\n\n\n

The coefficient of thermal expansion, or CTE, describes dimensional change with temperature. Silicon, copper, aluminum, ceramics, organic packages, and FR-4 boards have different CTE values and constraints.<\/p>\n\n\n\n

As an assembly heats, its plate and package try to grow by different amounts. Fasteners prevent free motion, so stress, bowing, or microscopic sliding appears at the TIM. Cooling reverses the direction. The grease therefore sees a mechanical workload, not just heat.<\/p>\n\n\n\n

Temperature alone is therefore an incomplete explanation. A stable hot interface may fare better than one cycling through a wide range every few minutes. Swing size, heating rate, geometry, and restraint all influence movement.<\/p>\n\n\n\n

Baseplate and Cold-Plate Warpage<\/h3>\n\n\n\n

Large IGBT or SiC modules make the effect easy to picture. Uneven heat can bow a baseplate, while the cold plate has its own gradient and stiffness. Their gap may open in one area while closing in another.<\/p>\n\n\n\n

That changing gap acts like a shallow pump. Compression pushes grease outward, while relaxation can leave weak regions underfilled. Large contact areas amplify small flatness errors, so a module that looks flat on a bench may behave differently at full load.<\/p>\n\n\n\n

Cyclic Shear in a Thin Bond Line<\/h3>\n\n\n\n

The grease must wet surface asperities while staying continuous across the active area. Relative lateral movement drags that thin layer back and forth. If the material yields easily and does not recover its structure, it creeps. If it resists movement too strongly, stress may move into the package, bond, or cured TIM instead.<\/p>\n\n\n\n

The material needs enough mobility to fill microscopic roughness during assembly, but enough structure to resist repeated shear. A practical analysis of paste pump-out on CPUs and GPUs<\/a> shows this interaction well, though its estimation approach is not a universal qualification standard.<\/p>\n\n\n\n

Vibration and Board Flex<\/h3>\n\n\n\n

Thermal movement is not the only driver. Vehicle vibration, server fan vibration, handling shock, laptop chassis flex, and PCB bending can disturb the contact. Shared cooling assemblies are especially sensitive. Tightening one spring screw may slightly lift another corner. A nearby thermal pad may be too thick and tilt the cooler. Tiny changes can become a repeating shear cycle.<\/p>\n\n\n\n

The real vibration and movement profile should guide TIM selection, especially in vehicles, mobile products, and industrial equipment.<\/p>\n\n\n\n

Material Properties That Control Pump-Out Resistance<\/h2>\n\n\n\n
\"Material<\/figure>\n\n\n\n

Why Viscosity Alone Is Not Enough<\/h3>\n\n\n\n

Viscosity is easy to quote, so it often becomes the headline. Yet a room-temperature viscosity number says little about how grease behaves at 100\u00b0C after 2,000 cycles. Test method, shear rate, temperature, and material history can all change the result.<\/p>\n\n\n\n

A thick grease can still migrate when hot. Two products with similar viscosity can have very different yield stress, oil separation, adhesion, and recovery. Viscosity matters, but it is only one piece.<\/p>\n\n\n\n

Yield Stress, Thixotropy, and Viscoelasticity<\/h3>\n\n\n\n

Yield stress<\/strong> is the force needed before a material starts to flow. Higher yield stress can help grease resist small mechanical disturbances, but too much can hurt spreading and increase bond line thickness.<\/p>\n\n\n\n

Thixotropy<\/strong> means the material flows more easily under shear and rebuilds structure at rest. That helps dispensing and may improve stability after mounting. Recovery at operating temperature matters most.<\/p>\n\n\n\n

Viscoelasticity<\/strong> is the mix of liquid-like flow and solid-like recovery. It changes with frequency and temperature, so test data should reflect the intended duty cycle.<\/p>\n\n\n\n

Filler Loading, Carrier Stability, and Oil Bleed<\/h3>\n\n\n\n

Thermal grease is normally conductive filler dispersed in a carrier. Particle type, size, shape, treatment, and loading affect conductivity and flow. A dense filler network may resist movement but hurt dispensing or wetting. If carrier oil bleeds away, the residue becomes stiff and filler-rich, so migration and dry-out can overlap. Check storage, dispensing, and time-at-temperature behavior.<\/p>\n\n\n\n

Curing and Non-Curing Compounds Trade Different Risks<\/h3>\n\n\n\n

A curing material limits free flow after it sets, which sounds ideal. The tradeoff is reduced ability to accommodate strain. A published thermal-cycling visualization study using ultrasonic imaging<\/a> observed void formation in a non-curing grease and cracking in a curing grease under its specific repeated-pressure setup. That does not rank all curing and non-curing formulations. It shows why changing the failure mode is not the same as removing it.<\/p>\n\n\n\n

High W\/mK Does Not Predict Pump-Out Life<\/h3>\n\n\n\n

Bulk thermal conductivity describes heat flow through material under defined conditions. It does not tell an engineer whether the grease will maintain coverage under warpage and shear. A high W\/mK product can perform poorly if its bond line becomes thick, develops voids, or leaves the hot region.<\/p>\n\n\n\n

Selection should consider initial thermal impedance and its change after aging. HAKTAK’s guide to \u30b5\u30fc\u30de\u30eb\u30b0\u30ea\u30b9\u306e\u71b1\u4f1d\u5c0e\u7387\u306e\u9078\u629e\u3068\u30c6\u30b9\u30c8<\/a> explains why conductivity alone is not the finished interface result.<\/p>\n\n\n\n

Bond Line Thickness, Mounting Pressure, and Surface Flatness<\/h2>\n\n\n\n
\"Bond<\/figure>\n\n\n\n

Why Bond Line Thickness Changes Heat Flow and Movement<\/h3>\n\n\n\n

For an ideal, uniform layer, bulk thermal resistance can be approximated as:<\/p>\n\n\n\n

Rbulk = BLT \/ (k x A)<\/p>\n\n\n\n

\u3053\u3053\u3067\u3059\u3001, BLT<\/strong> \u7d50\u5408\u7dda\u306e\u592a\u3055\u306f\u3001, k<\/strong> \u71b1\u4f1d\u5c0e\u7387\u306f\u3001 A<\/strong> is contact area. Real interfaces also include contact resistance, roughness, voids, and uneven pressure. A thick line stores more mobile material; an extremely thin one can lose continuity over warped surfaces. Target the thinnest complete bond line the tolerance stack can hold reliably.<\/p>\n\n\n\n

Does More Mounting Pressure Prevent Pump-Out?<\/h3>\n\n\n\n

Enough pressure improves wetting, collapses excess thickness, and establishes contact. Stable spring load can also reduce gap variation. Beyond that point, more force is not free performance. Excess pressure may squeeze out fresh paste, bow a board, deform a baseplate, crack a bare die, or disturb pads elsewhere in a shared cooler.<\/p>\n\n\n\n

Uneven force is often worse than modest force. Follow device and cooler specifications. Use a controlled torque sequence, the correct fasteners or springs, and measured compression where possible. A proper \u30b5\u30fc\u30de\u30eb\u30b0\u30ea\u30b9\u5857\u5e03\u306e\u30ef\u30fc\u30af\u30d5\u30ed\u30fc<\/a> helps make deposit and mounting repeatable, but it cannot compensate for poor flatness or unstable hardware.<\/p>\n\n\n\n

Flatness and Parallelism Come First<\/h3>\n\n\n\n

Check the heat spreader, die, module baseplate, heat-sink base, and cold plate across the relevant temperature range. A room-temperature flatness check can miss powered warpage. Verify that brackets do not tilt the cooler and that nearby gap pads are not holding it away from the primary device.<\/p>\n\n\n\n

For shared assemblies, inspect the imprint at every contact. Grease cannot bridge a real millimeter-scale gap while preserving a thin interface. That is a geometry problem.<\/p>\n\n\n\n

Thermal Grease Pump-Out Risks<\/h2>\n\n\n\n

Pump-out does not guarantee sudden failure. It often causes gradual, uneven degradation. That can be harder to catch because average temperatures may look acceptable while one local region runs much hotter.<\/p>\n\n\n\n

Risk<\/strong><\/td>What Changes<\/strong><\/td>Possible System Effect<\/strong><\/td><\/tr>
Void growth<\/td>Real contact area falls<\/td>Higher thermal impedance<\/td><\/tr>
Center starvation<\/td>Hot region loses TIM coverage<\/td>Local hot spot and larger temperature delta<\/td><\/tr>
Thermal drift<\/td>Interface resistance rises over cycles<\/td>Fan noise, throttling, or reduced output<\/td><\/tr>
Edge migration<\/td>Grease leaves the intended area<\/td>Contamination or electrical concern<\/td><\/tr>
Repeated overheating<\/td>Junction temperature rises<\/td>Faster aging elsewhere in the system<\/td><\/tr>
Unit variation<\/td>Mounts pump out at different rates<\/td>Unpredictable yield and field life<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Electrical risk depends on the compound. Use its electrical data and keep-out requirements rather than assuming every gray paste is harmless.<\/p>\n\n\n\n

How to Identify Thermal Paste Pump-Out<\/h2>\n\n\n\n

Look for a Trend, Not One Hot Reading<\/h3>\n\n\n\n

A classic pattern is good performance immediately after a correct repaste, followed by gradual temperature rise under the same workload. A GPU may show a widening core-to-hot-spot delta. A CPU may need higher fan speed or begin throttling at a power level it previously sustained. A power module may show rising case-to-sink thermal impedance.<\/p>\n\n\n\n

Control ambient, power, fan or pump speed, and coolant temperature before comparing results. Dust, a worn fan, a failing pump, a loose fastener, or a changed power limit can imitate a TIM problem.<\/p>\n\n\n\n

Inspect the Teardown Pattern Carefully<\/h3>\n\n\n\n

Useful clues include a bare or unusually thin center, a thick perimeter ridge, directional streaks, channels, voids, or cracks. Compare the pattern with the heat source location. Also look for one-sided contact, untouched areas, hardware witness marks, and evidence that a pad or bracket tilted the cooler.<\/p>\n\n\n\n

Some excess at the edge is normal. The key question is whether material left the active region during service.<\/p>\n\n\n\n

\u75c7\u72b6<\/strong><\/td>Possible Pump-Out Clue<\/strong><\/td>Competing Cause<\/strong><\/td>Next Check<\/strong><\/td><\/tr>
Temperature rises over weeks<\/td>Edge accumulation after a good initial mount<\/td>Dust or fan wear<\/td>Clean and repeat a controlled test<\/td><\/tr>
Sudden temperature jump<\/td>Cooler movement disturbed the interface<\/td>Pump failure or loose fastener<\/td>Inspect cooling hardware first<\/td><\/tr>
Large GPU hot-spot delta<\/td>Center or corner starvation<\/td>Cold-plate flatness or sensor variation<\/td>Compare imprint, power, and mount<\/td><\/tr>
Failure returns after repaste<\/td>TIM and assembly are poorly matched<\/td>Wrong pad thickness or bracket load<\/td>Review the whole stack-up<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Where Pump-Out Causes Real Trouble<\/h2>\n\n\n\n

Desktop CPUs<\/h3>\n\n\n\n

An integrated heat spreader offers a robust mounting surface, but uneven parts and repeated power cycles can still drive migration. Stable mounting and controlled application matter more than adding a giant blob of premium paste.<\/p>\n\n\n\n

GPUs and Laptop Direct Dies<\/h3>\n\n\n\n

Direct-die cooling, spring screws, PCB flex, and surrounding pads make contact sensitive. Over-thick memory or VRM pads may reduce die pressure. Repeated cycling then works paste outward. \u201cGreat after repaste, bad again in a month\u201d is a clue, not proof.<\/p>\n\n\n\n

AI Servers and Accelerators<\/h3>\n\n\n\n

AI accelerators run at high sustained power, often under cold plates. Start-stop events, load transitions, maintenance, and coolant changes still create cycles. At fleet scale, small interface drift means extra cooling energy or service work. Consistent torque, traceable materials, telemetry, and lot-level qualification matter.<\/p>\n\n\n\n

Power Electronics, EVs, and Industrial Equipment<\/h3>\n\n\n\n

IGBT, SiC, and GaN assemblies combine high heat flux with complex material stacks. EV inverters add wide ambient conditions, vibration, and coolant variation. Industrial drives, renewable-energy inverters, telecom gear, and LEDs may be expensive to service. Here, pump-out resistance is a reliability requirement.<\/p>\n\n\n\n

How to Prevent Thermal Grease Pump-Out<\/h2>\n\n\n\n

Start at the Material Level<\/h3>\n\n\n\n

Ask suppliers for pump-out, bleed, rheology, and aging data at relevant temperatures. Review operating range, yield behavior, oil separation, electrical properties, dispensing, and surface compatibility. A pump-out-resistant grease is often the easiest option for a thin, serviceable interface, but it still needs assembly-level cycling under the intended gap, pressure, movement, and temperature.<\/p>\n\n\n\n

Control the Interface<\/h3>\n\n\n\n

Use clean, compatible surfaces and a controlled deposit. Define minimum, nominal, and maximum bond line thickness. Record material lot, application method, and fastening results. Confirm pressure distribution and post-mount coverage. Rework should follow the production process because a casual hand application can hide a marginal design.<\/p>\n\n\n\n

Fix Mechanical Movement<\/h3>\n\n\n\n

Improve flatness and parallelism. Reduce warpage, maintain spring load through expansion, support flexible boards, control torque order, and check whether hoses or chassis parts pull on the cooler. The visible grease may be blamed when the real cause is a warped plate or over-thick neighboring pad.<\/p>\n\n\n\n

Reduce System Stress<\/h3>\n\n\n\n

Lower unnecessary temperature swing and power overshoot. Stabilize coolant flow and avoid controls that repeatedly drive the device from very hot to very cool. Monitor hot-spot delta and thermal drift. Condition-based maintenance is more useful than a universal replacement calendar.<\/p>\n\n\n\n

When Another TIM Is Better Than Grease<\/h2>\n\n\n\n

Recurring pump-out may mean the interface needs a different format. The goal is not to replace grease at any cost. It is to choose a material whose mechanical behavior matches the gap and movement.<\/p>\n\n\n\n

TIM Option<\/strong><\/td>Pump-Out Advantage<\/strong><\/td>Main Tradeoff<\/strong><\/td><\/tr>
Pump-out-resistant grease<\/td>Thin bond line and easy rework<\/td>Still requires cycling validation<\/td><\/tr>
\u76f8\u5909\u5316\u6750\u6599<\/td>Solid handling with heat-activated wetting<\/td>Needs suitable activation temperature and pressure<\/td><\/tr>
Curing gel or compound<\/td>Less free flow after cure<\/td>Cure stress, cracking risk, harder rework<\/td><\/tr>
Graphite interface<\/td>No liquid carrier to migrate<\/td>Electrical conductivity and conformity limits<\/td><\/tr>
\u71b1\u4f1d\u5c0e\u30d1\u30c3\u30c9<\/td>Stable preformed geometry<\/td>Thicker interface and compression force<\/td><\/tr>
Putty or gap filler<\/td>Accommodates uneven or variable gaps<\/td>Higher thickness and tighter process control<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Phase-Change Materials<\/h3>\n\n\n\n

A PCM is firm at room temperature and softens near its activation temperature. It can wet surfaces during burn-in and may improve stability in a thin joint. Activation temperature, pressure, thickness, and cycling still need review. The overview of phase-change thermal pads<\/a> explains where this format fits.<\/p>\n\n\n\n

Curing Gel, Graphite, Pads, and Gap-Filling Materials<\/h3>\n\n\n\n

A lightly curing compound reduces free flow but may crack. Graphite removes carrier migration but can be electrically conductive and less conformable. Pads offer stable geometry but add thickness and compression force.<\/p>\n\n\n\n

For a true gap, grease is usually the wrong tool. The comparison of thermal putty versus thermal paste<\/a> explains why putty can suit irregular component heights. Larger or complex cavities may call for thermal conductive liquid gap fillers<\/a> that cure in place. A broader review of alternatives to thermal paste<\/a> can help shortlist pads, PCM, graphite, putty, or liquid materials before testing.<\/p>\n\n\n\n

How to Test Thermal Grease Pump-Out<\/h2>\n\n\n\n

No universal pump-out test predicts every application. A useful plan combines thermal measurement, realistic movement, and physical inspection.<\/p>\n\n\n\n

Establish an Initial Baseline<\/h3>\n\n\n\n

Record material lot, deposit, bond line, surface condition, clamp load, power, ambient or coolant condition, and the thermal result. Use enough samples to expose assembly variation. Measure initial thermal impedance or a repeatable device-temperature metric after any burn-in, and retain controls.<\/p>\n\n\n\n

Use Thermal Cycling and Power Cycling for Different Questions<\/h3>\n\n\n\n

IEC 60068-2-14 and JEDEC JESD22-A104 provide useful frameworks for temperature-change or temperature-cycling exposure. They do not prescribe one universal grease pump-out profile. Select temperatures, ramp rates, dwell times, and cycle counts based on the product environment.<\/p>\n\n\n\n

Chamber cycling changes the whole assembly temperature. Power cycling heats the device from within and can reproduce local gradients and warpage. Many products need both. Add vibration or flex only when it represents service conditions.<\/p>\n\n\n\n

Accelerate Movement Without Losing Field Relevance<\/h3>\n\n\n\n

A mechanical fixture can impose controlled displacement or pressure change. This compares formulations quickly and separates movement from chemical aging. Correlation is the hard part. Confirm that teardown patterns, impedance drift, and migration resemble field or power-cycle results.<\/p>\n\n\n\n

Measure Before and After Aging<\/h3>\n\n\n\n

ASTM D5470-17(2024)<\/a> describes measurement of thermal transmission properties for thermally conductive electrical insulation materials. It can support controlled initial and post-aging thermal impedance comparisons. It does not define a pump-out cycling profile or guarantee performance in a specific CPU, GPU, or power module.<\/p>\n\n\n\n

Combine thermal results with imaging or teardown. Ultrasound can reveal internal voids, while thermal imaging can locate hot regions. Microscopy, coverage analysis, and photographs document migration after disassembly.<\/p>\n\n\n\n

Define Pass\/Fail Criteria Before Testing<\/h3>\n\n\n\n

Useful criteria may include:<\/p>\n\n\n\n

    \n
  • maximum change in thermal impedance;<\/li>\n\n\n\n
  • maximum device-temperature or hot-spot rise at controlled power;<\/li>\n\n\n\n
  • minimum covered area in the critical region;<\/li>\n\n\n\n
  • no unacceptable void growth;<\/li>\n\n\n\n
  • no grease migration into an electrical keep-out zone;<\/li>\n\n\n\n
  • no cracks, delamination, or hardware damage;<\/li>\n\n\n\n
  • repeatable results across lots and assembly operators.<\/li>\n<\/ul>\n\n\n\n

    Set limits from system margin and reliability needs. \u201cLooks okay\u201d is not an engineering acceptance criterion.<\/p>\n\n\n\n

    Practical Pump-Out Root-Cause Checklist<\/h2>\n\n\n\n
      \n
    1. Did temperatures start low and rise after repeated operating cycles?<\/li>\n\n\n\n
    2. Is the center thin or bare while material has collected near the edge?<\/li>\n\n\n\n
    3. Are the package, heat sink, baseplate, and cold plate flat and parallel in operation?<\/li>\n\n\n\n
    4. Is clamp load stable, even, and within the component specification?<\/li>\n\n\n\n
    5. Is bond line thickness controlled at minimum, nominal, and maximum gap?<\/li>\n\n\n\n
    6. Does the grease have relevant pump-out, bleed, and cycling data?<\/li>\n\n\n\n
    7. Are vibration, board flex, hose load, or baseplate warpage present?<\/li>\n\n\n\n
    8. Can peak temperature or temperature swing be reduced?<\/li>\n\n\n\n
    9. Would PCM, gel, graphite, a pad, putty, or gap filler better match the interface?<\/li>\n\n\n\n
    10. Has the complete assembly passed realistic cycling with defined limits?<\/li>\n<\/ol>\n\n\n\n

      \u7d50\u8ad6<\/h2>\n\n\n\n

      Thermal grease pump-out is a system-level mechanical reliability problem. CTE mismatch, warpage, cyclic shear, bond line thickness, clamp load, surface geometry, and grease rheology work together. That is why repeated repasting may restore temperatures for a while without fixing the cause.<\/p>\n\n\n\n

      Start with evidence. Confirm the performance trend, control the test conditions, inspect the contact pattern, and rule out the rest of the cooling system. Then work outward from the interface: select a suitable material, control the deposit and bond line, stabilize the mount, reduce movement, and manage thermal swings where practical.<\/p>\n\n\n\n

      The best solution is not always thicker grease, more mounting force, or a fashionable PCM. It is the combination that maintains contact in the real assembly. Validate that combination under representative thermal and mechanical cycles, and judge it by thermal drift plus physical condition. That is how a quick thermal fix becomes a durable design.<\/p>\n\n\n\n

      \u3088\u304f\u3042\u308b\u3054\u8cea\u554f<\/h2>\n\n\n\n

      What Is Thermal Grease Pump-Out?<\/h3>\n\n\n\n

      Thermal grease pump-out is gradual displacement from the active contact area. Repeated warpage, vibration, and pressure change move the bond line, often leaving edge buildup and weak center coverage.<\/p>\n\n\n\n

      What Causes the Pump-Out Effect in Thermal Paste?<\/h3>\n\n\n\n

      Relative movement is the main driver. CTE mismatch, warpage, vibration, board flex, changing clamp load, bond line geometry, and grease rheology control the migration rate.<\/p>\n\n\n\n

      How Can You Tell If Thermal Paste Has Pumped Out?<\/h3>\n\n\n\n

      Look for good initial performance that degrades over cycles, plus a thin hot region and edge buildup on teardown. Control power and cooling, then rule out dust, hardware, mounting, and flatness faults.<\/p>\n\n\n\n

      Is Pump-Out the Same as Thermal Paste Dry-Out?<\/h3>\n\n\n\n

      No. Pump-out is mechanical displacement; dry-out involves carrier loss, separation, oxidation, or hardening. They may occur together, but their root causes differ.<\/p>\n\n\n\n

      Why Is Pump-Out Common on GPUs and Laptops?<\/h3>\n\n\n\n

      Bare dies, concentrated heat, flexible structures, spring screws, and surrounding pads make contact sensitive. Frequent cycling can then move a marginal paste away from the center.<\/p>\n\n\n\n

      Does Mounting Pressure Prevent Thermal Grease Pump-Out?<\/h3>\n\n\n\n

      Correct, even pressure improves wetting and controls the bond line. Excess force can squeeze out material or deform parts, so follow the hardware specification instead of simply tightening more.<\/p>\n\n\n\n

      Does Thicker or More Viscous Paste Resist Pump-Out?<\/h3>\n\n\n\n

      Not necessarily. Room-temperature viscosity misses yield stress, structural recovery, bleed, and hot behavior. Validate resistance under the intended gap, pressure, temperature, and movement.<\/p>\n\n\n\n

      Can Phase-Change Material Prevent Pump-Out?<\/h3>\n\n\n\n

      A phase-change TIM may reduce grease migration because it is firm during handling and wets when heated. It still needs suitable activation temperature, pressure, thickness, and cycling data.<\/p>\n\n\n\n

      How Is Thermal Grease Pump-Out Tested?<\/h3>\n\n\n\n

      Engineers establish a baseline, apply representative thermal, power, mechanical, or vibration cycles, then compare thermal performance. Ultrasound and teardown can reveal voids, cracks, and migration.<\/p>\n\n\n\n

      What Is the Best Long-Term Solution for Recurring Pump-Out?<\/h3>\n\n\n\n

      Reduce the movement first by checking flatness, warpage, pad stack-up, clamp stability, and temperature swing. Then qualify a suitable grease or alternative TIM in the complete assembly.<\/p>\n<\/div><\/div>","protected":false},"excerpt":{"rendered":"

      Thermal grease pump-out is the gradual movement of grease away from the active contact area between a hot component and […]<\/p>\n","protected":false},"author":4,"featured_media":1355,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"disabled","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1824","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-haktak-blog"],"_links":{"self":[{"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/posts\/1824","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/comments?post=1824"}],"version-history":[{"count":1,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/posts\/1824\/revisions"}],"predecessor-version":[{"id":1825,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/posts\/1824\/revisions\/1825"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/media\/1355"}],"wp:attachment":[{"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/media?parent=1824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/categories?post=1824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/haktak.com\/ja\/wp-json\/wp\/v2\/tags?post=1824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}