Laptop Thermal Paste Ageing: An Advanced TIM Diagnosis — Dry-Out, Pump-Out, Liquid Metal, Phase-Change and Graphite

A laptop that keeps getting hotter, with fans spinning harder, does not always have a fan problem or a heat pipe problem. Often the real fault lies in a layer so thin it is almost invisible: the thermal interface material (TIM). This article covers only the TIM layer itself: thermal paste, liquid metal, phase-change materials, thermal pads, graphite sheets, application methods, bondline thickness, and contact pressure. Fan motors, fan control, temperature sensors, and the heatsink body itself are used only as comparisons; motherboard power delivery is out of scope. Do not open the machine yourself to work on the TIM layer; internal inspection should be left to qualified technicians. What readers can do is observe from outside, record temperatures, and describe symptoms accurately.

What the TIM layer does, and where it fails

Neither a chip surface nor a heatsink base is perfectly flat at microscopic scale. When the two faces are pressed together, only the high points truly touch, and the remaining gaps are filled with air, which conducts heat very poorly. The TIM fills those microscopic gaps, replacing air with a better-conducting material and lowering interface resistance. Intel's official guidance describes TIM as filling the natural microscopic surface defects and gaps between heatsink and chip; AMD's guidance likewise lists material conductivity, electrical conductivity, spreading behaviour, long-term stability and reliability, ease of use, and the pressure the heatsink applies through the TIM onto the package.

TIM failure is therefore not a vague statement that "the paste went bad". It happens at one or more of three places:

Failure location	What happens	Field clue
Chip-to-TIM interface	TIM fails to wet the chip surface, or an oxide film, contamination, or old residue prevents bonding, leaving voids	Improvement after a re-paste fades quickly
Inside the TIM, the bondline	Carrier loss, filler separation, cracking, or voids raise resistance inside the layer	The temperature curve under the same load shifts upward as a whole
TIM-to-heatsink interface	TIM is pumped out, migrates, or fails first at the edges because pressure is uneven, leaving the central hot spot uncovered	Temperature swings unpredictably and the hot spot drifts

Many users ask: What are the signs that thermal paste has dried out? The answer is not one symptom but a gradual pattern: temperatures first rise slightly at the end of heavy loads, throttling then appears earlier, and finally temperatures jump irregularly even under ordinary loads. TIM failure is rarely a sudden on-off fault.

Dry-out: the carrier leaves, the filler loses support

Thermal paste is generally a filler suspended in a carrier: the filler may be alumina, zinc oxide, boron nitride, or carbon-based particles, and the carrier is silicone oil or a synthetic oil. Dry-out is not the whole paste evaporating; it is the breakdown of the carrier-to-filler ratio. A reliability study in Electronics Cooling defines dry-out precisely: filler separates from the polymer matrix at high temperature, the matrix tends to flow out of the interface first, and the paste inside the interface dries out, raising field thermal resistance.

Dry-out stage	Paste condition	Temperature behaviour
Early	Carrier begins to leave or migrate, paste viscosity rises, the filler skeleton barely holds contact	Slight rise only at the end of heavy loads
Middle	Filler loses carrier support, micro-voids appear inside the layer, resistance rises	The whole temperature curve under the same load shifts upward and throttling appears earlier
Late	Paste becomes powdery, hard lumps, or crumbs, and can no longer flow to refill gaps	Contact at the central hot spot deteriorates first, with irregular temperature jumps

A common misjudgement is to see wet-looking paste at the edges and conclude the TIM is still healthy. Mechanically this is unreliable: carrier can migrate to the edges, leaving the edges looking wet while the filler skeleton under the central hot spot has already lost its carrier support. Dry-out cannot be judged by appearance alone; it needs temperature curves and before-and-after load comparisons.

Dry-out versus oil separation	Appearance	Mechanism	Consequence
Dry-out	The whole layer becomes dry and hard, powdery or lumpy	Carrier loss leaves the filler unsupported	Voids increase inside the layer, resistance rises
Oil separation	Oily shine at the edges, hard residue at the centre	Low-molecular-weight oil bleeds out of the thickener structure	One part of the same interface looks oily, another looks powdery, so resistance is uneven
Both together	Oil film and powdery lumps coexist	Separation and loss proceed together	Local high-resistance zones, gradual temperature drift rather than a jump

Pump-out: every thermal cycle pushes a little more

Many users ask: What is the pump-out effect? Pump-out is the apparent seeping or migration of a low-viscosity material, such as thermal paste, out from between two faces over time. The cause is repeated mechanical stress from temperature change: the two faces expand and contract at different rates, stressing the material and causing separation. AMD's guidance uses the same language: low-quality compounds harden or leak out over time, causing overheating or early device failure, while higher-viscosity pastes on bare-die devices generally resist pump-out better.

Many users ask: What is the difference between pump-out and dry-out? Pump-out is a change of material position, while dry-out is a change of material composition, and the two often accelerate each other: pump-out thins the centre, so the remaining filler loses carrier more easily; dry-out removes the paste's ability to flow back and refill the gaps pump-out created. A diagnosis that mentions only one will miss half the mechanism.

Four stages of pump-out	What happens	Result
Heating stage	Chip, heatsink base, and retainer expand at different coefficients, creating microscopic shear and a pressure gradient; paste softens and viscosity falls	Paste flows sideways more easily under pressure
Cooling stage	Materials contract out of step, paste is pulled and squeezed, and its viscoelastic nature tries to recover its shape	Layer thickness changes further
Repeated accumulation	Each heating and cooling cycle pushes a small amount of paste from the central high-pressure, high-temperature zone toward the low-pressure edges	The centre thins gradually, while paste piled at the edges contributes almost nothing to cooling
Critical point	The centre loses coverage, air or voids re-enter the interface	Local resistance jumps suddenly, and the temperature curve becomes unstable with a drifting hot spot

Conditions that accelerate pump-out	Mechanism	Why laptops are especially exposed
Low-viscosity formula	Low flow resistance, so shear pushes it out more easily	Thin machines have light heatsinks and short retainer travel
Large, frequent temperature cycles	Every cycle adds shear displacement	Long heavy loads alternating with shutdown or idle is the everyday laptop pattern
Excessive bondline thickness	A thicker layer has a larger absolute shear displacement	Retainer pressure is limited by the need to avoid cracking a bare die
Uneven contact pressure	The centre has the highest pressure and the largest temperature swings	Base or chip unevenness amplifies local differences
Uneven or bowed surfaces	Local pressure and layer thickness vary	Pump-out starts first at the weakest edge

Long-term testing by Igor's Lab adds an important geometric point: pump-out always starts at the edges and works inward; if the application leaves no material reserve at the edges, shear is no longer spread through volume but converts directly into directed flow, with binder transported outward along the interface first. Conversely, a somewhat thicker application that keeps an edge reserve acted as a mechanical buffer in testing. This overturns the intuition that thinner is always better: a layer too thin to keep a continuous skeleton is structurally worse under cycling. That is not encouragement to apply thick paste; it is a reminder that the layer must be thick enough to keep a continuous skeleton.

Cracking: the structural end-stage of dry-out

Cracking is the structural result of dry-out and phase separation reaching a late stage, not a separate chemical reaction. When carrier loss passes a threshold, the filler skeleton loses the binder's continuous support, and the layer can no longer follow the two faces as a continuous film under repeated expansion and contraction, so cracks and fragments appear.

Crack source	How it forms	Diagnostic meaning
Shrinkage cracks	Carrier loss shrinks volume, internal tensile stress exceeds bond strength	Net-like or radial cracks; each crack is itself an air channel
Shear cracks	Different expansion coefficients shear an already hardened layer repeatedly	Cracks extend along the shear direction; temperatures swing with temperature and retainer micro-movement
Edge peel cracks	Pump-out thins or piles the edges unevenly, concentrating stress at the thickness step	Cracks run from the edge toward the centre; the hot spot shifts with temperature

Phase separation and oil bleed: oil and powder part ways

Phase separation means filler and carrier go their separate ways: oily carrier gathers as a bright film or bleeds outward, while filler gathers as powder or dry lumps. ARCTIC lists non-bleeding alongside non-drying as a long-term performance claim.

Phase separation triggers	Mechanism	Note
Prolonged high temperature	Accelerates carrier evaporation and migration, weakening filler coating	Effect depends on the formula
High humidity	Research shows high humidity can change paste-layer resistance	Direction and size depend on filler and resin system; no blanket rule applies
Mismatched formula	Filler particle size, surface treatment, and carrier viscosity out of balance	Separation is easier under long cycling
Pressure and shear	Pump-out preferentially transports binder outward	The centre's filler fraction rises abnormally, accelerating separation

Bleeding oil does more than weaken the TIM itself; it can contaminate neighbouring components and interfaces, further harming contact. Judgement should look for the uneven state of oil film and powder coexisting across the layer, not only whether the centre looks dry.

Oxidation and surface degradation: a new paste can fail in the same place

Oxidation in TIM failure has two levels that must be kept apart.

Oxidation level	Where it happens	Effect
Inside the TIM and its filler	Some metal fillers and carriers change chemically under prolonged heat and moisture	Paste structure changes, volume changes, or resistance rises, not always visible
On the contact surfaces	Oxide films form on copper, aluminium, and nickel-plated copper bases	Surface energy changes, wetting falls, paste cannot fully bond, and voids remain in the interface

Fingerprint oil, mould release agents, cutting fluids, dust, and packaging residues add further barriers. The diagnostic point matters: even with a new paste, if the oxide film or old residue on the contact face is untreated, the new TIM layer will bond poorly in the same place, looking briefly better and then returning to the old behaviour. This is why "old paste not fully removed" and oxidation are often confused: the results look alike (poor bonding), but one cause is residue raising the bondline and the other is surface chemistry preventing wetting.

Seven TIM material classes compared: numbers cannot be ranked across brands

All values below are manufacturer nominal values, not measurements made for this article; different manufacturers test differently, so direct cross-brand ranking is not valid. Measurements also show that a conventional paste's effective conductivity in a real thin layer, at real pressure and temperature, is often lower than the nominal bulk value on the package, because contact resistance takes a large share.

Class	Form and filler	Manufacturer nominal conductivity	Electrical conductivity	Durability tendency	Suitable use	Risks and limits
Conventional silicone paste	Silicone oil carrier with alumina or zinc oxide fillers	Generally low to mid-range, commonly about 1 to 6 W/m·K	Generally insulating, per individual datasheet	Moderate; oil separation and pump-out under long thermal cycling	Everyday and office laptops, a safe general choice	An over-thick layer raises resistance instead
Carbon-based paste	Silicone gel or oil carrier with carbon particles	Nominal values vary widely by manufacturer; check the individual datasheet	For ARCTIC MX-6, officially non-conductive	Relatively good; high viscosity and low-bleed design target long-term stability	Gaming and office machines seeking a balance of performance and low short-circuit risk	High-viscosity paste is hard to spread evenly with a blade
Metal-oxide performance paste	Silicone oil carrier with fine nano alumina or zinc oxide fillers	Thermal Grizzly Kryonaut older datasheet 12.5 W/m·K; Kryonaut Extreme older datasheet 14.2 W/m·K	Manufacturer states non-conductive	Better early performance, but some reviews note performance can fall back after thermal cycling	High-power gaming machines, machines sensitive to early temperatures	High performance does not equal high durability; judge the two separately
Liquid metal	Gallium-indium-tin alloy, liquid at room temperature	Thermal Grizzly's current official page no longer publishes a general W/m·K value since late 2020, so this article does not use an old number as a current nominal value	Conductive; the manufacturer warns in bold about conductivity and never contacting aluminium	The metal itself does not dry out, but alloying, migration, and pump-out remain	Only confirmed copper or nickel-plated copper bases with full barriers and insulation protection	Absolute limits are listed in the next section
Phase-change material	Polymer phase-change system with conductive filler, a solid sheet at room temperature	Honeywell PTM7950 manufacturer or distributor data: 8.5 W/m·K	Generally non-conductive, per individual datasheet	More resistant to pump-out and dry-out	Gaming machines, ultra-thin machines, machines where fewer re-pastes are wanted	In low-power use that never reaches phase-change temperature, forming is incomplete
Silicone thermal pad	Silicone base with ceramic or boron nitride fillers, pre-formed sheet	By grade, commonly about 1 to 12 W/m·K in manufacturer nominal values	Generally insulating, per individual datasheet	More stable, no dry-out, no pump-out, but can take a permanent compression set	Memory, power stages, and chipsets with larger, uneven gaps	The absolute limit is using the wrong thickness in place of the original pad
Graphite sheet	Pyrolytic or flexible graphite sheet, layered carbon structure	In-plane values can reach hundreds to over a thousand W/m·K, but through-plane values are far lower, often below 20 W/m·K	Conductive	No dry-out, no pump-out, better long-term stability	Office machines or specific ultra-thin designs wanting fewer re-pastes	Damaged, frayed sheets must not be reused; alignment matters

Three common misreadings of conductivity	Misreading	Correct understanding
A bigger number always means lower temperature	Choosing by nominal conductivity alone	Total interface resistance combines the paste's own resistance with two contact resistances; in thin layers, contact resistance often dominates
All manufacturers' numbers share one ruler	Ranking brands by one number	Test method, pressure, thickness, temperature, and surface roughness all change results; Noctua publicly explains why it does not publish a single conductivity figure, and ARCTIC’s datasheets likewise give no single figure
Laboratory effective conductivity equals the package number	Treating a bulk-material number as installed-interface performance	Bulk material and installed interface are two different measurements; real thin-layer effective values are often lower

Many users ask: How often should thermal paste be replaced? There is no single schedule that fits every model, because replacement timing depends on formula, load pattern, cycle count, retainer pressure, and environment. Judge by temperature-curve drift, earlier throttling, and widening load-to-idle temperature gaps, not by a fixed number of years. For inspection arrangements, see the diagnostic service; for fee structure, see pricing.

Liquid metal: chemical risk behind the performance

Many users ask: Is liquid metal suitable for every laptop? The answer is no. Liquid metal suits only confirmed copper or nickel-plated copper bases with full barriers and insulation protection in an original or professionally designed application. Commuting vibration and repeated thermal cycling increase migration risk, so mobile laptop use carries higher risk than a machine that stays on a desk.

Many users ask: Why must liquid metal never touch aluminium? Because gallium reacts very strongly with aluminium: Indium Corporation notes gallium can corrode through thin aluminium foil within hours at room temperature, faster at high temperature, and Thermal Grizzly likewise forbids aluminium contact. On copper, CuGa2 forms; on nickel-plated copper, NiGa5 forms a thinner reaction layer, and the manufacturer recommends nickel-plated copper as the more stable long-term contact face.

Liquid metal risk	Mechanism	Consequence
Aluminium corrosion	Gallium reacts strongly with aluminium	Thin aluminium foil can be corroded through within hours at room temperature, faster when hot
Alloying	CuGa2 on copper, NiGa5 on nickel-plated copper	Contact-face properties change, so long-term behaviour differs from early behaviour
Migration	Liquid at room temperature, high flow and migration risk	A single spilled drop can cause a short circuit
Conductive short circuit	Liquid metal conducts electricity	Use near exposed circuitry without protection can short the board

Liquid metal suitability and limits	Situation	Judgement
Confirmed copper or nickel-plated copper base, full barrier and insulation protection	A professionally designed application	Decided by the design, never assumed
Aluminium base, or base material unknown	Absolute limit	Must not be considered
Inadequate protection around the die, exposed circuitry nearby	Absolute limit	Spill risk is unacceptable
Frequent commuting vibration and mobile use	Risk increased	Migration risk is higher than for a desk-bound machine

Phase-change pads: a solid-soft-solid cycle aimed at pump-out

Many users ask: What is the difference between a phase-change pad and conventional thermal paste? A phase-change material is a solid sheet at room temperature that softens or melts when heated, refills micro-gaps, and resolidifies on cooling, so its position is relatively fixed at the end of each cycle and it resists pump-out and dry-out better. Conventional paste stays paste-like across the whole working temperature range, so its position moves more easily under cycling.

Many users ask: Does a phase-change pad need a burn-in? A phase-change pad needs to reach its phase-change temperature to complete forming. For Honeywell PTM7950, sources record a phase-change range of about 45 to about 52 degrees, with a recommendation to hold at least 60 degrees for a period to reach minimum bondline; minimum bondline is about 17 to 20 micrometres. In low-power use that never reaches phase-change temperature, forming stays incomplete and early resistance stays high.

Phase-change pad points	Content	Caution
Form	Solid sheet at room temperature, commonly 0.25 mm, with a 0.2 mm version also listed	Check genuine specifications; do not substitute a non-original thickness
Phase-change temperature	About 45 to about 52 degrees; full change needs the whole sheet near 50 degrees	Low-power use may never reach it
Recommended clamping pressure	About 30 to 40 psi	Insufficient pressure affects forming and bondline
Minimum bondline	About 17 to 20 micrometres, typically below 50 micrometres	Wrong thickness changes the bondline
Reliability test conditions	Manufacturer reliability data mentions 1,000 hours at 150 degrees and 1,000 thermal cycles with low resistance maintained	These are manufacturer test conditions, not an everyday lifespan guarantee

Thermal pads: the wrong thickness lifts the heatsink

Many users ask: What happens if a thermal pad of the wrong thickness is used? Thermal pads suit memory, power stages, and chipsets with larger, uneven gaps; they are not a direct replacement for the thin paste layer on a processor or graphics chip. Too thick, and the pad lifts the heatsink so the processor interface loses contact; too thin, and gaps stay partly empty, leaving sections suspended.

Thermal pad parameters	Content	Selection note
Hardness	Measured in Shore OO, commonly about 15 to 80 Shore OO	Too hard a pad bonds poorly at low pressure; too soft a pad is over-compressed or squeezed out at high pressure
Thickness	Commonly 0.5 to 5.0 mm, with compression allowance needed	Must match the original thickness; hardness and compression must be judged together with gap and pressure
Durability	More stable, no dry-out, no pump-out	Can take a permanent compression set, harden at the edges, and lose surface bonding
Electrical conductivity	Generally insulating	Check the individual datasheet

Graphite sheets: the in-plane number is not the through-plane number

Many users ask: Why cannot a graphite sheet's conductivity be judged by its in-plane value alone? Because graphite's layered carbon structure makes in-plane and through-plane conduction very different: in-plane values can reach hundreds to over a thousand W/m·K, but through-plane values are far lower, often below 20 W/m·K. A processor interface needs through-plane transfer, not in-plane spreading, so using an in-plane number to predict interface performance is a misreading.

Many users ask: What risk does graphite's conductivity create? Technical data states clearly that graphite conducts electricity, so contact with exposed conductors, battery tabs, or antenna traces can cause short circuits or change grounding and radio-frequency behaviour. Graphite sheets are flexible but brittle; edges wear, delaminate, and shed debris; no-flow, no-dry-out, and reusability apply only while the sheet is undamaged.

Graphite misreadings and facts	Misreading	Fact
The in-plane number equals interface performance	Expecting a large temperature drop from a hundreds-scale number	The interface needs through-plane transfer, whose value is far lower
Graphite sheets can be cut and placed freely	Continuing to use a damaged edge	A sheet with damaged edges or debris must not be reused; misalignment creates hot spots
Graphite sheets insulate	Allowing contact with exposed circuitry	Graphite conducts; without insulation separation, contact with exposed conductors can short

Application methods and bondline: coverage matters more than the pattern name

Many users ask: What is the difference between dot, cross, and spread application? The table below describes how paste spreads under pressure in principle, not an instruction manual. Every method must answer the same question: under retainer pressure, can the paste form a complete, continuous, thin film with no large trapped bubbles?

Method	Principle	Coverage tendency	Bubble risk	Main drawback
Centre dot	One central dot, spread outward radially by retainer pressure	Thickest at the centre, thinning outward; good for small dies, corners may fall short on large dies	Lower, as paste pushes air outward from a single point	Central thickness raises local resistance; large-die corner starvation risk
Cross and X shape	Two diagonals crossing at the centre, starting points already closer to the corners	Better large-die coverage, higher corner reach than a dot	Lower, lines still push outward	Line ends near the edge can overflow; the crossing point runs thick
Lines	One or more parallel lines spreading sideways	More even thickness distribution, suits rectangular dies	Low to moderate; lines must be continuous and even, broken lines trap air	Line-end overflow risk; uneven spacing gives uneven thickness
Multiple dots	A centre dot plus surrounding dots, pre-distributing paste	More complete coverage, higher corner reach	Moderate to high; air can be trapped where dots meet	Harder dose control; excess overflows; bubble risk at dot junctions is higher than a single dot
Blade spread and pre-applied film	A blade or pre-applied film forms an even thin layer over the whole contact face first	Theoretically most even, thickness pre-controlled	Depends on technique; repeated strokes, an uneven blade, or dry paste can trap air; some high-viscosity products must not be blade-spread	Spreading most easily introduces bubbles and uneven thickness; pre-applied films are limited by film thickness and alignment accuracy

Three judgement principles follow: coverage matters more than the pattern name, since a film with missing corners, missing edges, or an over-thick centre fails regardless of the pattern used; a bubble is resistance, not a minor flaw, because air conducts extremely poorly and one large bubble's local resistance is far higher than the surrounding paste layer's, creating a hot spot; and the dose logic is "just enough to fill the micro-gaps", not "more is safer", because excess paste is simply squeezed out at the edges, which for non-conductive paste means mainly mess and waste, but for liquid metal and conductive graphite means short-circuit and contamination risk.

Many users ask: Does applying more thermal paste give better cooling? The answer is no. The thicker the bondline, the higher the resistance, and bondline thickness is set jointly by face flatness, TIM thickness, and retainer clamping force. Too much paste thickens resistance; too little paste, uneven coverage, bubbles, uneven pressure, old residue, or an untreated oxide film can all make a re-paste worse than before.

Many users ask: What is bondline thickness? Bondline thickness is the final thickness of the TIM layer clamped between chip and heatsink. Thermal resistance is set jointly by face flatness, TIM thickness, and retainer clamping force, so bondline is not the paste's own labelled thickness but the actual installed layer thickness. A phase-change pad's minimum bondline is about 17 to 20 micrometres, typically below 50 micrometres; a conventional paste's bondline varies with dose, viscosity, and pressure, and there is no single number for all models.

Bondline and resistance	Situation	Effect
Bondline too thick	The paste's own resistance rises with thickness	Overall interface resistance rises, temperatures shift upward
Bondline too thin to keep a skeleton	Filler sits almost two-dimensionally, losing three-dimensional skeleton support	Under load and thermal cycling the structure destabilises quickly and cracks more easily
Uneven bondline	Some areas thick, some thin, pressure uneven	Thin areas fail first, thick areas run high resistance, hot spots drift
Old residue raising the bondline	Old paste or oxide film left untreated	The new layer bonds poorly in the same place

Many users ask: What problems does uneven contact pressure cause? Uneven contact pressure makes local layer thickness vary, and the position with the highest central pressure and the largest hot-spot temperature swings is where the pump-out driving force is greatest, while the loss first shows at the edges: paste is pushed from the central high-pressure zone toward the low-pressure edges, the edges fail first, and the central hot spot then loses coverage. Insufficient or uneven pressure lets the layer slide more easily under cycling; pressure distribution depends jointly on retainer design, base flatness, and pad thickness, and simply tightening harder does not fix unevenness.

Pressure problems	Mechanism	Field behaviour
Insufficient pressure	The layer slides under cycling, bonding is inadequate	Acceptable bonding at first, gradual temperature drift as cycles accumulate
Uneven pressure	Local thickness varies, thin areas fail first	Temperatures swing unpredictably, hot-spot position is unstable
Over-thick pad lifting the heatsink	The processor interface loses contact	Temperatures stay high even with brand-new paste
Uneven surfaces	Local pressure and thickness vary	Pump-out starts first at the weakest edge

Layered exclusion: prove the problem is really in the TIM layer

What the TIM layer cannot do matters as much as what it can: it cannot fix a warped cold plate, a loose retainer, or an over-thick thermal pad lifting the heatsink; it cannot replace airflow and exhaust, nor unblock dust-clogged fins; it cannot correct wrong power settings or fan curves; and it cannot raise the heatsink module's own cooling ceiling. So when temperatures rise, the first question should not be which paste is best, but whether the resistance increase is really in the TIM layer.

Seven layered-exclusion steps	What to do	Judgement focus
Step 1, fixed-condition recording	Fix room temperature, power mode, and load program; record idle and heavy-load temperatures	Without fixed conditions, temperature numbers cannot be compared
Step 2, curve shape	Check whether temperature rises smoothly, jumps suddenly, or jumps irregularly	Smooth rise points to overall resistance increase, a sudden jump to dry-out or a contact break, irregular jumps to cracking or unstable contact
Step 3, fan comparison	Note fan speed, noise, and exhaust temperature changes	When the fan is at fault, exhaust and speed change first; see the diagnostic service for inspection steps and pricing for fee structure
Step 4, heat pipe and fin comparison	Note fin temperature distribution and exhaust temperature	Fins cool while the chip is very hot points to heat not being moved; fins very hot with weak exhaust points to heat not being removed
Step 5, sensor cross-check	Compare readings from different sensor points for contradictions	When one sensor disagrees while others look normal, suspect the sensing or reading chain first
Step 6, time pattern	Record whether the problem appears at boot, after a period of use, or intermittently	At boot points to contact or installation, gradual onset to ageing, intermittent behaviour to unstable contact or cracking
Step 7, combined judgement before acting	Put all six steps together before deciding on TIM-layer work	One symptom alone cannot settle it, and a re-paste cannot fix fan, heat pipe, or sensor faults

Many users ask: How can you tell whether a temperature problem comes from the TIM layer or the fan? If fan speed, noise, and exhaust temperature change together, the fan or airflow path is more likely at fault; if the fan behaves normally but the temperature curve under the same load shifts upward as a whole and throttling appears earlier, interface resistance is more likely involved. Both can coexist, and layered exclusion exists precisely to avoid fixing only one side.

Many users ask: How can you tell whether a temperature problem comes from the TIM layer or the heat pipe? A heat pipe fault typically shows heat not being moved: the chip is very hot while the fins stay relatively cool, or the pipe tail is unusually cold; TIM failure is a rise in interface resistance, where heat enters the heatsink less efficiently but the pipe's own transport ability may be intact. Curve shapes differ too: heat pipe dry-out often jumps suddenly, while TIM ageing usually drifts gradually.

Comparison: TIM versus other parts	TIM layer failure	Fan path	Heat pipe path	Sensor path
Typical temperature curve	Gradual upward shift, earlier throttling	Speed or exhaust changes first, temperatures follow	Sudden jump, or unusually cold pipe tail	One reading abnormal, others not in step
Key clue	Load-to-idle temperature gap widens, curve unstable	Noise, speed, and exhaust temperature change together	Heat not moved, abnormal fin temperature distribution	Readings contradict each other
Does a re-paste help	Only if the fault is really in the TIM layer	No	No	No

Hong Kong context: humidity, commuting, and thermal cycling combined

Many users ask: How does Hong Kong's humid weather affect the TIM layer? The Hong Kong Observatory's 1991 to 2020 average relative humidity is 78%, and high humidity has been shown to change paste-layer resistance, with the exact effect depending on the filler and resin system's response to moisture. It cannot be said flatly that humidity always worsens paste, nor that it never matters. Humidity also accelerates oxidation and contamination of contact surfaces, worsening wetting and bonding of new paste.

Many users ask: Do commuting vibration and temperature cycling accelerate TIM failure? Yes, and the three factors combine. Daily commuting vibration adds mechanical disturbance to the interface, raising migration risk for flowable materials such as liquid metal and alignment risk for graphite sheets; moving between air-conditioned rooms and outdoors adds repeated expansion and contraction, amplifying pump-out shear; high humidity affects the chemical stability of paste and contact surfaces. When all three are present, material choice should favour pump-out-resistant and separation-resistant structures, such as high-viscosity paste or phase-change material, rather than chasing the highest nominal number.

Hong Kong factors	Mechanism	Selection and judgement implication
High humidity, average relative humidity 78%	Can change paste-layer resistance, accelerates surface oxidation and contamination	Effect varies by formula; judge by temperature curves, not by weather alone
Commuting vibration	Extra mechanical disturbance at the interface	Higher migration risk for flowable materials, higher alignment drift for graphite sheets
Air-conditioned room to outdoor cycling	Repeated expansion and contraction, amplifying shear displacement	Higher pump-out and cracking risk; temperature records must fix room temperature and load conditions
Long heavy loads alternating with idle	Many cycles, large amplitude	Gradual drift reflects TIM ageing better than a single hot reading

Common misdiagnoses: six reasons a re-paste can make things worse

Many users ask: Why might temperatures be higher after reapplying thermal paste? Six TIM-layer reasons below often occur together.

Reason a re-paste gets worse	Mechanism	How to avoid misjudgement
Too much paste	Bondline thickens, resistance rises	Dose logic is just enough to fill micro-gaps, not more is safer
Too little paste or uneven coverage	Corner starvation, central hot spot uncovered	Coverage matters more than the pattern name
Bubbles trapped in the layer	Air conducts extremely poorly, local resistance jumps	A bubble is resistance, not a minor flaw
Uneven pressure	Local thickness varies, thin areas fail first	Tightening harder alone does not fix unevenness
Old paste not fully removed	Residue raises the bondline	The new layer bonds poorly in the same place
Oxide film on the contact face untreated	Surface energy changes, new material cannot wet	Brief improvement, then a return to the old behaviour

Other common misdiagnoses include treating fan noise as TIM failure, treating heat pipe dry-out as dried paste, treating an abnormal sensor reading as real overheating, and concluding from wet-looking edge paste that the TIM is still healthy. Layered exclusion exists for exactly this reason: prove the problem is really in the TIM layer before deciding on treatment. If the faulty layer is unclear, check the diagnostic service for inspection steps and pricing for fee structure.

Repair, replace, or upgrade decision framework

Situation	Decision direction	Reason
Temperature curve drifting upward gradually, layered exclusion points to the TIM layer, paste dried out or pumped out	Replace with the same type of paste, or switch to a phase-change material	Treats the confirmed failure mechanism
Paste repeatedly pumps out, model cycles thermally often, retainer pressure is adequate	Consider a phase-change material	Resolidifying on cooling keeps its position relatively fixed, resisting pump-out better
Fewer re-pastes wanted, office machine or specific ultra-thin design	Consider a graphite sheet, but only as the design decides	No dry-out, no pump-out, but conductive and demanding on flatness
Larger, uneven gaps on memory or power-stage components	Use a thermal pad of the correct thickness	Paste cannot fill large gaps; wrong thickness lifts the heatsink or leaves sections suspended
Aluminium base, unknown base material, or inadequate surrounding protection	Never consider liquid metal	Aluminium corrosion and conductive short-circuit risks are unacceptable
Layered exclusion points to fan, heat pipe, or sensor	Fix that part first, do not change TIM	Changing TIM cannot fix other parts' faults
Old model with retainer or base already deformed	Judge by inspection results, do not assume an upgrade helps	Deformation makes pressure and thickness uneven, so a material change alone may not improve things

Fee note: battery from HK$580, screen from HK$1,280, motherboard chip-level from HK$880, and keyboard from HK$480 are confirmed reference fees, and they do not represent the fee for TIM-layer work in this article; all other fees, including this topic, are subject to a written quotation after inspection. For inspection steps, see the diagnostic service; for fee structure, see pricing.

What to record before sending a machine for inspection

Readers should not handle the internal TIM layer themselves, but they can record useful information for layered judgement: idle and heavy-load temperature readings, changes in fan noise and exhaust feel, whether the problem appears at boot or after a period of use, whether temperature rises smoothly or jumps suddenly, and any major recent change in usage environment, such as moving from long air-conditioned use to outdoor use. Records should fix power mode and load program, or numbers from different conditions cannot be compared. For inspection arrangements, see the diagnostic service; for fee structure, see pricing.

Viscosity, filler, and durability: a three-way trade-off

Viscosity, filler, and durability pull against each other and cannot be optimised one at a time. High viscosity resists pump-out and leakage better, but spreads with difficulty and demands more from retainer pressure and surface flatness; under a low-pressure structure it may leave coverage gaps. Low viscosity spreads easily and bonds well at first, but is more easily displaced and separated under thermal cycling. Larger filler particles set a higher floor for minimum bondline; finer particles allow thinner layers, but filling ratio and viscosity must stay balanced, since over-filling makes paste dry and brittle. Carrier stability matters equally: low-bleed formulas and thixotropic designs exist to keep filler and carrier uniform under prolonged heat and pressure, rather than separating. A genuinely high-performance paste must therefore answer three questions at once: what is its early resistance, what is its resistance after a thousand cycles, and can it form a complete thin film under this model's retainer pressure? Any comparison that answers only the first question is incomplete.

Selection decision tree: by model and usage pattern

For office machines, stability and low risk come first: conventional silicone paste or carbon-based paste is enough, and the priorities are complete coverage and even pressure, not chasing the highest nominal number. For gaming machines, cycling durability comes first: if paste repeatedly pumps out, consider a phase-change material; if power is high but the environment is fixed and retainer pressure is adequate, a performance paste is acceptable, provided its durability is not assumed equal to a conventional paste's. For ultra-thin machines, bondline and the pressure ceiling come first: bare-die direct contact and short retainer travel mean selection must follow the model's original thickness and pressure conditions, and any substitution with the wrong pad or phase-change thickness changes the bondline. For older machines, retainer and base condition come first: if the base is deformed or retainer spring force has faded, a material change alone may not help, and inspection results must decide. Every model shares one hard line: aluminium or unknown base material, or inadequate surrounding protection, means liquid metal is never considered; and any design where a graphite sheet would touch exposed conductors is likewise never considered.

Conclusion

TIM-layer failure is rarely a single cause: dry-out changes composition, pump-out changes position, cracking opens air channels, phase separation splits oil from powder, oxidation stops new material from wetting, and pressure and thickness set the whole layer's resistance starting point. Liquid metal, phase-change pads, thermal pads, and graphite sheets each have suitable uses and absolute limits, and no single material suits every model. The safest judgement order is always the same: record temperature curves under fixed conditions first, exclude fan, heat pipe, and sensor layers next, and only then decide on TIM-layer work. If inspection arrangements or fees are unclear, see the diagnostic service and pricing, and rely on a written quotation after inspection.

Shop information: Room 1209, Daly Plaza, Shek Mun, Shatin; telephone +852 2186 8010; WhatsApp +852 6558 6806; business hours 12:00–19:00.

If the symptoms above are already present, record the temperature curve first, then check the diagnostic service and pricing.

Disclaimer: This article is for reference only; actual conditions must be judged by professional inspection results. Different models have different structures and original specifications, and handling should follow the written quotation and professional advice after inspection.