You apply fresh thermal grease, mount the cooler, start the system and expect a happier temperature reading. Instead, the fan gets loud and the temperature climbs. Annoying, yes. But the paste itself may not be the villain.
Most thermal grease application mistakes come from one of five things: dirty surfaces, the wrong amount, poor coverage, uneven mounting or an unsuitable material. The visible pattern is only part of the story. A reliable interface also needs good wetting, a thin bond line, stable pressure and a process that can be repeated.

This guide explains what commonly goes wrong, how to judge the risk and how to correct it. CPU and GPU examples make the ideas easy to picture. The same fundamentals also apply to IGBT power modules, LED assemblies, power supplies, telecom hardware and industrial electronics.
What Thermal Grease Is Supposed to Do
Thermal grease fills microscopic valleys between two clamped surfaces and replaces trapped air with a more thermally conductive material. It is meant to form a thin, continuous interface. It is not a structural adhesive, and it is usually not the right material for filling a large mechanical gap.
Metal surfaces can look smooth while behaving more like tiny mountain ranges under magnification. When a processor lid, module baseplate or LED board touches a heat sink, only the highest points meet directly. Air remains in the low spots. Since air transfers heat poorly, those small voids raise contact resistance.
A suitable thermal grease material wets both surfaces and fills those low spots. Mounting pressure then pushes the grease into a thin bond line. The metal still carries much of the heat across its direct contact points; the grease improves the places where metal does not meet metal.
That last bit matters. Grease usually conducts heat better than air but worse than aluminum or copper. A thick blanket of paste is not the goal. You want enough material to remove air and maintain contact, but not so much that the grease becomes an unnecessarily long heat path.
Why the same deposit can behave differently
A pea-sized dot might work on one desktop CPU and miss part of a larger rectangular heat spreader. A stencil print that works on one power module may leave dry zones when the baseplate flatness changes. Results depend on the whole interface:
- Contact area and heat-source location
- Surface flatness and roughness
- Grease viscosity and thixotropy
- Mounting pressure and fastener pattern
- Final bond-line thickness
- Operating temperature and orientation
- Movement during thermal cycling
So, yes, application method matters. It just does not act alone.
Mistake 1: Applying Thermal Grease Over Old Paste or Contamination
Fresh grease cannot repair a dirty interface. Old compound, dust, fingerprints, lint and cleaner residue become extra layers in the heat path. They may also stop the new material from wetting the surface evenly, which creates local resistance and unpredictable spread.

The temptation is understandable. The old paste still looks soft, so adding another dab feels harmless. In practice, the old layer has already been compressed, heated and exposed to its environment. Mixing it with fresh material creates a bond line with unknown thickness and rheology.
What should be removed before reapplication
- Previous thermal paste from both mating surfaces
- Dust, lint and loose particles
- Skin oil and fingerprints
- Machining or handling oil
- Removable protective film on a new cold plate
- Cleaner residue and moisture
Use the cleaning method approved for the component and cooler. For many metal CPU lids and cold plates, high-purity isopropyl alcohol and a lint-free wipe are common choices. Sensitive plastics, coatings, labels and optical parts may require a different cleaner, so check the manufacturer instructions first. Intel’s basic thermal-paste application guidance also recommends removing old material and letting the surfaces dry before installation.
A simple preparation routine
- Power down and make the assembly safe to handle.
- Remove bulk residue without scratching the surface.
- Wipe both mating faces with an approved cleaner.
- Let the surfaces dry fully.
- Inspect them under good, angled light.
- Keep fingers away from the cleaned contact area.
Do not use a fuzzy tissue because it happens to be nearby. A single fiber across a thin bond line is surprisingly large at interface scale. Haktak’s broader guide to preparing and applying thermal grease covers the practical sequence in more detail.
Mistake 2: Using Too Much Thermal Grease
Too much grease is not automatically a disaster, especially when a nonconductive paste can escape under strong, even clamping pressure. Still, excess material can produce a thicker bond line, create a mess, contaminate nearby parts and hide an inconsistent application process. The actual risk depends on the assembly.
Online discussions often reduce this to a shouting match: “too much is fine” versus “too much insulates the CPU.” Both statements lose the conditions that matter.
If two flat surfaces are clamped firmly and the grease can flow outward, much of the excess may squeeze out. Thermal performance can remain acceptable, although material is wasted and cleanup becomes unpleasant. If pressure is low, the surfaces are warped or the paste is highly viscous, the excess may remain trapped as a thicker layer. That adds resistance.
The relationship is easy to picture. Heat crossing a thin smear takes a short route. Heat crossing a thick mound travels farther through a material that is less conductive than the surrounding metal. Haktak’s explanation of bond-line thickness and thermal performance explores this effect in depth.
Signs that the application volume is excessive
- Heavy squeeze-out around most of the interface
- Grease reaching a socket, connector or PCB keep-out area
- A cooler that slides easily after first contact
- A visibly thick transfer layer after controlled disassembly
- Large unit-to-unit variation on a production line
- Difficulty reaching the intended clamp position
Overflow also changes the electrical risk. Many common silicone thermal greases are electrically insulating, but that should never be assumed from appearance. Metal-filled compounds and liquid metal can be electrically conductive. Check the exact technical data and safety instructions before allowing any spill near traces, pins or exposed components.
Mistake 3: Using Too Little or Missing the Active Heat Area
Too little thermal grease leaves dry zones and air-filled voids. The full outer edge does not always need perfect cosmetic coverage, but the active heat source and intended contact area must be coupled continuously to the cooler. A bare patch over the hot region is far more serious than a small dry corner away from it.
A universal “pea size” is convenient advice, not a universal engineering unit. Package geometry varies. A square desktop CPU, a long server processor, a bare GPU die and an IGBT baseplate do not have the same area or heat-flux map.
Clues that coverage may be insufficient
- One region, core or device runs much hotter than comparable areas
- Temperature rises immediately under load
- A controlled transfer imprint shows a dry patch over the heat source
- Paste reaches only the center of a large rectangular interface
- Thermal results vary when the cooler is remounted with the same hardware
Do not keep adding tiny amounts around the edge of an assembled interface. If coverage is genuinely uncertain, remove the cooler safely, document the imprint, clean both surfaces and start again with a controlled application.
For high-value assemblies, a transparent plate or witness coupon can help develop the process before working on the real component. It is a development tool, not a substitute for final thermal verification.
Mistake 4: Treating the Dot, Line, X or Spread Pattern as the Whole Answer
There is no single pattern that wins for every interface. The useful method is the one that provides repeatable wetting and a controlled final bond line for the actual geometry, paste and mounting system. Volume, placement, pressure and verification matter more than the shape drawn on top.
The center dot is simple and often effective on moderate, clamped surfaces. A line may fit an elongated heat spreader. An X can help distribute material over a larger area, but intersecting beads can become too thick if the volume is not controlled. Manual spreading offers visible coverage, yet a dirty tool, ridges or repeated scraping can introduce new problems.
Industrial assemblies often use stencil printing, screen printing or automated dispensing because “about this much” does not scale well to thousands of units.
| Method | Useful Situation | Common Error |
| Center dot | Small or moderate clamped surfaces | Assuming one dot size fits every package |
| Line or X | Elongated or larger heat spreaders | Depositing thick, intersecting beads |
| Manual spread | High-viscosity paste or visible coverage requirement | Contamination, ridges and variable film thickness |
| Stencil or screen print | Repeatable power-module production | Wrong aperture, worn tooling or unverified print thickness |
| Automated dispense | Controlled production volume and location | Calibration drift, trapped air or a poor dispense path |
The right choice should be developed around the part, not internet tradition. See Haktak’s comparison of thermal paste application methods for a closer look at each technique. For power semiconductors, Littelfuse also provides an official application note on stencil generation for thermal grease.
Mistake 5: Lowering or Tightening the Cooler Unevenly
A correct deposit can still fail when the cooler is tilted, dragged sideways or tightened fully from one corner. Uneven mounting moves grease away from some regions, changes local bond-line thickness and creates an uneven pressure map across the interface.

Lower the cooling surface squarely once the grease is in place. Avoid sliding it around to “help the paste spread.” Start each fastener lightly, then use the sequence and torque specified by the hardware or module manufacturer. A diagonal cross pattern is common, but the documented assembly procedure should win.
Why more torque is not always better
Higher clamp load can reduce the bond line and improve wetting up to a point. Past that point, more force may bow a PCB, stress solder joints, distort a module baseplate or damage a bare die. It can also squeeze too much grease out of the active area.
Spring hardware, washers and brackets matter because they help maintain pressure as the assembly expands and contracts. A loose or bottomed-out fastener cannot be fixed by choosing a more expensive paste.
Mistake 6: Reusing Grease After Lifting the Cooler
Once the interface has been separated, the old grease film may not return to the same continuous state. The safest long-term practice is to clean both surfaces and apply fresh material, even when the first application is only a few minutes old.
Lifting a cooler breaks the compressed film. Paste stretches, tears and moves into ridges. Dust may enter before the cooler is replaced. Reseating can trap gaps between those disturbed regions. Sometimes temperatures still look fine, which is why this shortcut survives. But “it booted” is not the same as a controlled process.
For a temporary diagnostic check, an engineer may accept a different level of risk. A final product, customer sample or production assembly needs the documented process. If the cooler comes off, reset the interface.
Do not add fresh paste on top of the disturbed layer. That creates a mixture with uneven thickness and unknown distribution. Clean, inspect and reapply.
Mistake 7: Selecting Thermal Grease by W/mK Alone
A high thermal-conductivity number does not guarantee the lowest device temperature. A grease with strong wetting and a thin stable bond line can outperform a higher-W/mK material that is difficult to spread, too thick in use or unstable during cycling.
Thermal conductivity describes the material. Thermal impedance describes opposition to heat flow through the tested interface under stated conditions. Device temperature also includes contact resistance, spreading resistance, cooler performance and the rest of the heat path.
That is why the label on the syringe is only the start. Review the thermal grease parameters that affect real use, including viscosity, oil bleed, volatility, dielectric behavior and temperature range. Then interpret conductivity alongside thermal impedance and interface conditions.
Properties that should be considered together
- Thermal impedance at a known bond line and pressure
- Apparent or bulk thermal conductivity and test method
- Viscosity, thixotropy and spread behavior
- Operating and storage temperature
- Oil bleed, volatility and filler separation
- Pump-out resistance under cycling
- Electrical conductivity or insulation
- Compatibility with metals, plastics, coatings and seals
A technically modest grease used correctly can beat an impressive datasheet number used badly. Not glamorous, perhaps, but common.
Mistake 8: Ignoring Storage, Separation and Shelf Life
An application can look neat and still fail because the material was expired, separated or stored outside its specified conditions. The syringe, cartridge or jar is part of the process. Grease condition affects dispensing, filler distribution, wetting and long-term stability.

Before use, confirm the product identity, lot number, expiry date and storage history. Look for unusual oil separation, dry crust, lumps or inconsistent color. Some products have an approved conditioning or mixing procedure; others should not be remixed in the field. Follow the technical data sheet.
Never restore a thickened grease by adding solvent or oil. That changes the formulation and invalidates the supplier data. Do not mix two pastes because their colors look similar. Filler chemistry, carrier fluid, additives and electrical behavior may be different.
If old material has hardened at the interface, the detailed explanation of why thermal paste dries out can help separate storage problems from in-service aging.
Mistake 9: Assuming Good Initial Temperatures Prove Long-Term Reliability
A new grease interface can perform well and then degrade through pump-out, carrier loss, oil bleed or changing contact pressure. Qualification should represent the real thermal cycles, vibration, orientation and service life. A short benchmark proves initial assembly performance, not lifetime stability.
Pump-out happens when repeated expansion and contraction move grease away from part of the active interface. It is especially important around large dies, power modules and assemblies with significant coefficient-of-thermal-expansion mismatch. A vertical orientation may add migration risk for a material with poor slump control.
Haktak’s guide to thermal grease pump-out during cycling explains the mechanism and validation approach.
A realistic power-module example
Imagine an inverter module that passes its end-of-line temperature test. After repeated heat-up and cool-down cycles, one corner begins running hotter. The catalog W/mK value has not changed, but local contact has. Baseplate movement, uneven fastener load and grease displacement may have created a growing dry region.
The useful test is not simply “does it turn on after aging?” Measure the same thermal response before and after the environmental profile. Inspect movement, bleed, cracking and contact where the failure analysis allows it.
Mistake 10: Blaming the Grease Before Checking the Whole Cooling System
High temperature after repasting does not prove the thermal paste is defective. Poor cooler seating, a stopped pump, reversed airflow, the wrong bracket, changed power limits or a misleading sensor can create the same symptom. Troubleshooting should follow the complete heat path.
| Symptom | Possible Grease-Related Cause | Other Cause to Check |
| Immediate high temperature | Poor coverage, forgotten film, uneven seating | Pump disconnected, fan error, wrong bracket |
| One region or core is hotter | Local dry zone or pressure variation | Die layout, sensor behavior, warped surface |
| Temperature rises over weeks | Pump-out, dry-out or migration | Dust buildup, fan wear, coolant problem |
| Large unit-to-unit variation | Uncontrolled dispensing or torque | Part tolerance, cooler flatness, test variation |
| Paste reaches nearby components | Excess volume or poor placement | Missing keep-out control or handling error |
Start with what changed. If the cooler was removed, inspect mounting and paste application. If only software changed, check workload and power behavior. If every unit in one batch runs hot, look at material handling, dispensing calibration and assembly settings. Random guesses burn time.
How to Correct a Bad Thermal Grease Application
Do not keep adding paste to an uncertain interface. Record the symptom, rule out obvious cooling-system faults, remove the cooler safely, clean both mating surfaces and rebuild the interface with a controlled material volume and mounting process.
Corrective workflow
- Record a baseline. Note ambient temperature, workload, power, fan or pump state and the sensor being read.
- Check the cooling hardware. Confirm airflow, coolant flow, brackets, springs, screws and electrical connections.
- Remove the cooler safely. Follow the component or module procedure and avoid pulling at an angle.
- Document the transfer pattern. A clear photograph can help reveal dry areas, heavy squeeze-out or pressure imbalance.
- Clean both surfaces. Remove old material and contamination completely.
- Inspect flatness and damage. Look for scratches, bowing, dents and foreign particles.
- Confirm the grease condition. Verify the product, lot, storage and expiry.
- Apply a controlled amount. Use the developed pattern, stencil or dispense program.
- Mount evenly. Follow the approved fastener sequence and torque.
- Repeat the same test. Compare results under the original load and ambient conditions.
If the second application is still poor, stop repasting blindly. Investigate the cooler, mechanical stack, power level and measurement method.
Preventing Thermal Grease Application Errors in Production
Repeatability comes from controlling material condition, dispense volume, location, tooling and clamp load. Operator experience is useful, but a production process that depends entirely on feel will drift between people, shifts and lots.
Manual syringe application may be suitable for prototypes or low volume. As production increases, define what operators can inspect and what equipment must control. A calibrated dispenser can manage shot volume and location. A stencil can create a repeatable printed area. Torque-controlled fastening can reduce pressure variation.
Process controls worth documenting
- Approved material, supplier and lot traceability
- Storage, conditioning and open-time limits
- Dispense mass, volume or printed thickness
- Needle, nozzle, screen or stencil condition
- Placement path and PCB keep-out zones
- Fastener sequence and torque
- Rework and cleanup method
- First-article and periodic thermal checks
- Aging requirements for high-risk products
A compact inspection plan
| Control | What It Catches | Possible Record |
| Material and lot check | Wrong or expired grease | Traveler, barcode or batch record |
| Dispense volume or weight | Too much, too little and equipment drift | Shot log or sample weight |
| Visual placement check | Offset deposit, void, overflow and contamination | Approved limit sample or vision result |
| Fastener control | Uneven or insufficient pressure | Torque-tool record |
| Thermal baseline | Combined interface and assembly problems | Temperature rise or thermal-impedance result |
| Post-aging comparison | Pump-out, dry-out and contact loss | Before-and-after thermal data |
The goal is not to inspect quality into the product after assembly. It is to make a bad application difficult to produce in the first place.
What ASTM D5470 Can and Cannot Tell You
ASTM D5470 supports thermal-impedance measurement and apparent-conductivity calculation under defined fixture conditions. It is valuable for comparing thermal interface materials, but it does not certify a finished application volume, mounting process or aged electronic assembly.
The official ASTM D5470 standard page describes a method covering materials from liquid compounds to hard solids. Results depend on specimen thickness, pressure, temperature, contact surfaces and the method used to separate bulk and contact effects.
When reviewing data, ask:
- What pressure was applied?
- What bond-line thickness was measured?
- At what temperature was the test run?
- Were the contact surfaces comparable?
- Is the reported number conductivity, resistance or impedance?
- Was the material tested before or after aging?
Use a standard to improve measurement consistency, not to avoid assembly testing. Haktak’s overview of common TIM testing standards explains where D5470 and related methods fit.
When Thermal Grease Is the Wrong Interface Material
Grease works best between thin, reasonably flat and mechanically clamped surfaces. If the design has a large variable gap, needs clean preformed placement, cannot maintain pressure or requires structural attachment, another thermal interface format may be more reliable.
Consider alternatives when the assembly needs:
- Gap filling across large height tolerance
- A controlled preformed thickness
- Clean placement with no squeeze-out
- Better resistance to migration in a vertical orientation
- Adhesive attachment between parts
- Simplified field service or automation
A thermal pad can cover a designed gap and provide electrical insulation. A liquid gap filler can conform around mixed component heights. A phase-change material offers solid handling before it softens in operation. A thermally conductive adhesive creates a bonded heat path.
For thin, flat and strongly clamped interfaces, low thermal resistance grease remains a strong option. The point is to use it where its behavior fits the mechanics, not because paste is familiar.
Information Haktak Needs to Review an Application
A useful thermal-grease recommendation starts with the actual heat source, mating surfaces, expected bond line, clamping method, application process and reliability profile. A close-up photo helps, but it cannot show pressure, power or service conditions by itself.
Send as much of the following as the project allows:
- Component, module or assembly type
- Normal and peak power dissipation
- Heat sink or cold-plate material and finish
- Contact area, flatness, gap and tolerance
- Target component temperature or interface resistance
- Clamp pressure, screw layout and service orientation
- Manual dispense, automation, stencil or screen-print process
- Operating temperature, thermal cycling, vibration and life target
- Electrical insulation, outgassing or silicone restrictions
- Sample quantity and expected production volume
Haktak can use this information to support thermal material selection and application testing before a process moves into production.
Conclusion
The most common thermal grease mistakes are not mysterious: contaminated surfaces, uncontrolled quantity, incomplete coverage, uneven mounting, reused paste and weak verification. What makes them tricky is that several can create the same symptom.
Do not judge the interface by the drawing pattern alone. Clean both surfaces, use a suitable grease, control the amount, mount the cooler evenly and test the complete cooling system under a repeatable load. For production equipment, add traceability, calibrated dispensing, torque control and aging checks.
A good grease application is a little boring. It looks the same from unit to unit, behaves predictably and gives engineers very little to argue about later. That is exactly what you want.
Frequently Asked Questions
What is the most common mistake when applying thermal grease?
Using an uncontrolled amount on an unprepared surface is the broadest recurring mistake. Too much, too little and contamination all change the final interface. Clean both mating surfaces, confirm the correct material, control the deposit and mount the cooler evenly. The pattern matters less than the final wetting, bond line and pressure.
What happens if I apply too much thermal paste?
Excess paste may squeeze out harmlessly under strong, even pressure, but it can also leave a thicker bond line, make the cooler slide and contaminate nearby parts. Electrically conductive compounds create additional risk. If the paste has reached a socket or PCB area, follow the material and hardware cleaning instructions before powering the assembly.
Can too little thermal grease cause overheating?
Yes. Too little material can leave dry contact zones and air-filled voids, especially over the active heat source. That raises local interface resistance and may create hot spots. The correct quantity depends on the contact area, package shape, grease rheology and mounting pressure rather than a universal dot size.
Is it better to spread thermal paste or let mounting pressure spread it?
Either method can work. A center deposit is simple when pressure and geometry spread it reliably. Manual spreading can help with viscous material or a coverage requirement, but the tool and surface must stay clean and the film must remain thin. Production parts often benefit from a developed stencil, screen or automated dispense process.
Do I need to remove old thermal paste before applying new paste?
Yes for a dependable final assembly. Old compound may contain dried regions, contamination and an already-compressed filler structure. Adding new paste on top creates an uneven, unknown bond line. Remove the old material from both surfaces with an approved method, let them dry and then apply fresh grease.
Must I reapply thermal grease after removing the cooler?
Reapplication is the safest long-term practice. Removing the cooler breaks the compressed film and can create ridges, dry zones or trapped voids when it is reseated. Even if the paste is fresh, clean both surfaces and rebuild the interface before final use or qualification.
Can air bubbles in thermal paste cause high temperatures?
Large or poorly located voids can reduce contact and create a hot spot because air conducts heat badly. Still, not every tiny visible bubble causes a measurable problem. Focus on complete wetting over the active region, controlled volume, straight mounting and even pressure rather than trying to eliminate every microscopic void by guesswork.
Can thermal grease damage or short a circuit board?
It depends on the product. Many thermal greases are electrically insulating, while metal-filled compounds and liquid metal may conduct electricity. Even insulating grease can contaminate connectors or make later service difficult. Check the exact datasheet and keep the material inside the approved application area.
Why are temperatures still high after replacing thermal paste?
The cooler may be seated unevenly, a protective film may remain, the fan or pump may not be operating, airflow may be wrong or the test workload may have changed. The paste may also have poor coverage or an excessive bond line. Compare the complete cooling system against a controlled baseline before blaming one material.
How can a manufacturer keep thermal grease application consistent?
Define the approved grease, storage rules, deposit volume, placement area, application equipment, fastener sequence and torque. Use calibrated dispensing or printing where volume justifies it. Record material lots, inspect tooling, establish visual limits and compare thermal performance before and after relevant aging. Repeatability should come from the process, not operator memory.

