Thirty minutes into your assembly, the epoxy still feels tacky. Two hours later, it's barely set. By morning, it never fully hardened, leaving a rubbery, weak bond that fails under minimal stress. Poor epoxy cure is rarely a defect in the adhesive itself. Instead, most cure failures trace to controllable factors: mixing errors, environmental conditions, or substrate interactions that stop the resin-hardener reaction from completing. This matters just as much when selecting a high-strength epoxy for metal assemblies, since even the best-specified system fails if it never fully cures. Mixing Errors: The Leading Cause of Failed Cures Two-part epoxies depend on precise stoichiometry, the chemical ratio of resin to hardener, and getting it wrong is the single largest cause of incomplete cure. Visual estimation, spoon-measures, or counting drops introduce ratio errors of 5–15%, each dramatically weakening the cured epoxy. A 10% hardener deficiency reduces final strength by 15–30% and delays gelation by hours; excess hardener cures quickly but embrittles the adhesive, making it prone to cracking. Weigh each component on a scale accurate to 0.1 grams and confirm the ratio against the data sheet before combining. Getting the ratio right isn't enough if mixing falls short. Stir vigorously for 60–90 seconds with a stiff stick, scraping the container's walls and bottom repeatedly—circular stirring misses corners that a figure-eight pattern reaches. Incomplete mixing leaves pockets of unmixed resin or hardener that cure unevenly, creating weak zones in the bond line. The container matters too: plastic bags, wax-coated paper, and certain plastics can react with epoxy components or absorb hardener, so mix only in rigid glass, metal, or supplier-validated containers. Temperature and Humidity: The Invisible Variables Epoxy cure is a thermally driven chemical reaction, and temperature controls both its speed and completeness. If the epoxy specifies a 70–77°F minimum cure temperature and your assembly sits at 50°F—an unheated garage in winter—the reaction slows dramatically: a 20°F drop can stretch gel time from 30 minutes to 4+ hours and push full cure from 24 hours to 5+ days. Below a critical minimum (often 50°F for general-purpose epoxies), the reaction can stall entirely. Measure the actual bond-line temperature, not ambient air—metal conducts heat away, running 5–15°F cooler than the room. Humidity compounds the problem in both directions. Cold, humid conditions let moisture condense on the epoxy surface and inhibit hardener activation, so wait for temperatures above 60°F and humidity below 85% before bonding outdoors. During the first 4–8 hours of cure, humidity above 85% RH lets moisture migrate into the partially cured epoxy and disrupt cross-linking, leaving a soft, undercured interface—cure in 40–60% RH where possible and shield outdoor work from dew and rain. At the opposite extreme, large exothermic masses (1+ quart mixed at once) can spike past 200°F, cracking the cured epoxy; mix only what you need, and apply thick sections in layers. Surface Preparation: The Foundation for Cure Poor substrate preparation prevents the epoxy from chemically bonding to metal surfaces, leaving a mechanically adhesive interface that never fully cures. Mill scale,…

A marketing sheet claims 5,000 psi shear strength. A competitor advertises 6,500 psi. A third lists tensile strength at 70 MPa. These numbers blur together, leaving engineers guessing whether the specified strength actually predicts performance in their application. The gap between lab-tested epoxy strength and real-world performance often exceeds 30%—revealing that material data sheets alone cannot answer the question: how strong is high-strength structural epoxy in my assembly? Understanding the Three Strength Metrics That Matter Structural epoxies resist failure through three distinct mechanical properties, each dominating different loading scenarios. Tensile Strength measures the epoxy's resistance to pulling forces along the adhesive axis. Most structural epoxies develop 4,000–8,000 psi tensile strength when tested as neat adhesive (bulk material without substrates). This property dominates only in specific geometries—thin adhesive layers under direct tension—and is often misleading for typical bond-line applications. Shear Strength indicates resistance to sliding forces across the adhesive layer. This is the dominant stress in lap-shear joints, the most common assembly geometry, and it's what ASTM D1002 was written to measure: the apparent shear strength of a single-lap adhesive joint under tension loading. Structural epoxies typically develop 3,000–6,000 psi shear strength under this test, with high-performance formulations exceeding 7,000 psi. Shear strength is your primary selection criterion for structural bonding, and it's a natural companion metric to review alongside our guide on selecting a metal-bonding epoxy system in the first place. Peel Strength quantifies the epoxy's resistance to peeling forces that concentrate stress at the bond-line edge. Peel is the most damaging stress mode—it concentrates loads into a narrow front rather than distributing them across the bond area. Most structural epoxies develop 2–5 pli (pounds per linear inch) peel strength, meaning a 1-inch-wide epoxy layer withstands 2–5 pounds of perpendicular pulling force before failure initiates at the interface edge. Real assemblies experience all three stress modes simultaneously. Selecting epoxy based on tensile or shear strength alone ignores peel risk and application geometry. How Strength Specifications Are Generated (and Why They're Incomplete) Standardized lap-shear testing produces the numbers on data sheets—but it measures epoxy in isolation, under a defined 1x1-inch overlap area, ideal surface preparation, controlled cure temperature, and minimal gap. Real assemblies introduce variables that reduce measured strength: Cure Temperature Variance. Standard test protocols assume 77°F ambient during cure. Assembly in winter (40°F) or summer heat (95°F) alters cure kinetics, potentially reducing final strength by 10–25%. An epoxy curing slowly in cold temperatures doesn't develop full cross-link density. Surface Preparation Variation. Test surfaces are chemically cleaned and grit-blasted to uniform roughness—the kind of controlled prep described in surface-preparation guides for metal adherends. Production surfaces, contaminated with oil or handled with bare hands, develop weaker interfaces, and poor surface prep alone can cut strength by 30–50%. Bond-Line Thickness. Standard lap-shear testing typically uses 0.05-inch adhesive thickness. Thinner bond lines (0.010–0.020 inch) experience higher stress concentrations and may fail at 80–90% of the rated strength. Thicker layers (0.15+ inch) trap voids and cure slower, reducing strength. Gap-Filled Joints. Many assemblies tolerate 0.5–2mm…

Metal-to-metal bonds demand more than surface-level adhesion. When equipment fails, the consequences ripple through production schedules, maintenance budgets, and safety protocols—making the choice of epoxy adhesive far more consequential than many engineers realize. Selecting the right structural epoxy for metal applications requires understanding load requirements, environmental exposure, and cure chemistry. This guide cuts through the noise to help you identify the epoxy systems that deliver reliable metal bonding without compromise. What Makes an Epoxy "Structural" for Metal Applications Structural epoxies for metal bonding differ from general-purpose formulations through their cure mechanisms and mechanical properties. A true structural epoxy develops high shear strength, tensile strength, and impact resistance while maintaining these properties across thermal cycles and environmental stress. Metal substrates present unique challenges. Unlike composites or plastics, metals conduct heat rapidly, affecting cure rates and exotherm intensity. Surface oxidation and mill scale require specific pre-treatment protocols—outlined in guides such as ASTM D2651—to ensure chemical bonding rather than mechanical adhesion alone. Structural formulations are engineered to penetrate microscopic surface irregularities and create chemical cross-links that resist peel and shear forces. Key Properties for Metal-to-Metal Bonds When evaluating structural epoxies for metal applications, four properties separate adequate from exceptional systems: Shear Strength. Metals experience predominant shear loading in lap joints and overlapping bond lines. Structural epoxies designed for metal bonding typically achieve 3,000–5,000 psi shear strength at room temperature under ASTM D1002 test conditions, with high-performance systems exceeding 6,000 psi. This property matters more than tensile strength for most metal assemblies. Temperature Resistance. Industrial applications demand epoxies that maintain strength across temperature swings. Glass transition temperature (Tg) indicates when an epoxy transitions from rigid to rubbery. For metal assemblies, Tg above 150°C ensures strength retention in elevated-temperature environments like automotive engine compartments or industrial machinery. Thermal Expansion Coefficient (CTE). Metals expand and contract with temperature change. Epoxy CTE typically ranges from 40–80 ppm/°C, while metals sit around 10–20 ppm/°C. This mismatch creates internal stress at the interface. High-performance structural epoxies minimize CTE through fillers or specialty resin chemistry, reducing thermal fatigue. Gap-Fill Capability. Real-world metal surfaces are never perfectly smooth. Structural epoxies that fill gaps of 0.5–2mm without voids maintain load paths and resist micro-crack initiation. Thixotropic (paste) formulations outperform thin liquids in this regard. How Metal Bonds Actually Fail in Service Understanding failure mechanisms clarifies what "high-strength" needs to mean for your application. Three failure modes dominate metal-to-metal epoxy joints: adhesive failure (the epoxy separates cleanly from the metal, indicating poor surface prep or contamination), cohesive failure (the epoxy itself splits internally, usually the sign of a correctly bonded, well-matched system reaching its material limit), and interfacial fatigue failure (repeated thermal or vibration cycling slowly propagates a crack along the bond line until sudden separation). A field example illustrates the stakes: a stamped-steel bracket assembly bonded with a general-purpose epoxy rated at 3,200 psi shear passed initial quality checks, but began delaminating within four months of outdoor service. Post-failure inspection showed adhesive failure at the interface—the bond had never chemically keyed to…

An electronic component rated for 10,000 operating hours fails after 3,000 hours in an unencapsulated assembly. The same component, potted with a high-temperature compound, operates for 9,000 hours — approaching its rated specification. Potting doesn't change component ratings. It eliminates the failure mechanisms that keep components from reaching rated life in real-world environments. Rated Life vs. Real-World Failure Component datasheets specify lifetime at rated conditions — for example, an electrolytic capacitor rated "105°C, 5,000 hours at rated voltage" means the component fails within 5,000 operating hours if held continuously at 105°C and rated voltage (median life; some units fail earlier, some later). In an unencapsulated assembly sitting in a 100°C environment, that capacitor reaches its 105°C rating almost immediately, and thermal cycling between −20°C and +100°C adds stress on top. Combined thermal and electrical stress can cut actual life from a rated 5,000 hours to 1,000–2,000. In a potted assembly, thermal management drops peak capacitor temperature to 95°C — now operating below rating rather than at it, with margin extending life beyond the datasheet number. Why Thermal Margin Matters So Much Component reliability broadly follows the Arrhenius model: doubling temperature margin above rated operation can extend life by 2–10x depending on the failure mechanism. An electrolytic capacitor rated 105°C/5,000 hours running at 105°C gets roughly its rated 5,000 hours; the same part at 95°C (10°C margin) reaches roughly 10,000 hours; at 85°C (20°C margin) it reaches 20,000–25,000 hours — a 4–5x improvement from margin alone. The relationship holds broadly across semiconductors, capacitors, and solder joints alike. How Potting Delivers That Margin Thermally-conductive potting flattens temperature gradients, typically cutting peak component temperature 10–20°C versus an unencapsulated design. Elastomer-toughened potting damps mechanical vibration, reducing solder joint oscillation and extending fatigue life 5–10x — the same mechanism covered in our piece on vibration damage prevention. Moisture exclusion prevents the electrolytic corrosion on traces, leads, and solder joints that is the dominant failure mode for unencapsulated electronics in humid environments, and full encapsulation shields components from salt spray, oil mist, and industrial chemicals that would otherwise degrade insulation. Put together, these mechanisms typically deliver: 3–5x life extension on electrolytic capacitors (1,500–3,000 hours unpotted vs. 5,000–8,000 hours potted), 3–5x on solder joints under cycling (500–1,500 cycles vs. 2,000–5,000), 3–5x on copper traces against corrosion (2–4 years vs. 7–10+ years), 2–3x on power semiconductor junctions (3,000–5,000 hours vs. 8,000–12,000), and 3–5x on connector pin reliability (1–2 years vs. 5–8 years). Vibration-driven solder fatigue sees the largest swing, often 5–20x, since unencapsulated leads oscillate essentially unconstrained. Cost-Benefit of Extended Component Life For a 20-year industrial control system, an unpotted component rated for 10,000 hours but actually lasting 3,000 hours under thermal and vibration stress needs replacement every 3–4 years — 5–6 replacement cycles over the system's life, each with parts and labor cost. The same component potted reaches 9,000–10,000 hours of actual life, meaning 0–1 replacements over 20 years. Potting material cost of $5–20 per assembly is easily justified against eliminating 4–5 replacement cycles at $100–500 each…

Machinery vibration stresses solder joints and component leads. An unencapsulated PCB in a 50Hz/5G vibration environment fails within 6–12 months. The same PCB, potted with the right compound, survives 5+ years in identical conditions. Potting doesn't eliminate vibration stress — it constrains component movement and absorbs energy through material damping, turning destructive vibration into manageable stress. How Vibration Damages Unencapsulated PCBs Machinery vibration excites PCB resonances, and component leads oscillate at the board's natural frequency with amplified motion. In worst-case resonance, a 1G environmental vibration can produce 5–20G acceleration at the lead tips. Cyclic bending under that acceleration drives three failure modes. Solder fatigue accumulates as cyclic bending plastically deforms the solder's lead-tin microstructure; after 10,000–100,000 stress cycles a crack initiates and propagates to open circuit. Component leads with brittle plating or stress-concentrated geometry crack after just 1,000–10,000 cycles. Copper traces thinner than 0.25mm or longer than 50mm without support can fracture after 100,000+ cycles. In continuous automotive or industrial vibration, this progression — micro-cracks at 1,000 cycles, degrading continuity at 10,000, open circuit by 50,000–100,000 — corresponds to just 6–18 months of service life. Potting's Protection Mechanism Potting reduces stress on solder joints by 70–95% through two combined effects. Mechanical constraint solidifies the assembly into a rigid mass that resists flex; component leads can no longer oscillate freely, and PCB natural frequency shifts upward (often into the ultrasonic range) or disappears from the excitation spectrum entirely, eliminating resonance amplification. Vibration damping, present in elastomer-toughened formulations, dissipates energy as heat through micro-deformation of the elastomer particles — proportional to elastomer loading (5–15% by weight). Un-toughened rigid potting delivers constraint but minimal damping; toughened potting delivers both, cutting vibration stress by 80–95%. The gap between the two is dramatic. At 5G environmental acceleration with no potting, PCB resonance amplifies lead acceleration to 20G and solder fails within 50,000 cycles. Rigid potting (no damping) eliminates resonance and holds lead acceleration near the environmental 5–6G, pushing fatigue failure out to 500,000 cycles — a 10x improvement. Elastomer-toughened potting adds damping on top of constraint, cutting peak lead acceleration to 2–3G and pushing fatigue failure beyond 2 million cycles — practically indefinite service, and another 2–4x improvement over rigid potting alone. Specifications for Vibration Duty Target 8–15% elastomer toughening by weight (10–12% is optimal for most applications — higher loading improves damping but can reduce thermal stability and ultimate stiffness), Shore D hardness 70–80 for a balance of damping and mechanical support, damping capability that doesn't degrade significantly across the operating temperature range, and elongation at break above 10% so the compound absorbs localized strain without cracking. Property Rigid Potting Elastomer-Toughened Silicone Damping Poor Excellent Excellent Mechanical support Excellent Good Fair Thermal stability Excellent Good Good Vibration protection Moderate Excellent Excellent Thermal cycling Fair (brittle) Excellent Excellent For a broader look at how these same chemistries stack up outside vibration duty, see our comparison of silicone, epoxy, and polyurethane potting compounds. Design Techniques That Complement Potting Potting alone is insufficient for extreme vibration environments.…

An outdoor LED street-light controller endures UV exposure, temperature swings from −30°C winter nights to +70°C day heat, rain and morning dew, and salt spray from roadway de-icing. After 18 months, its potting yellows, turns brittle, and cracks — exposing the electronics inside to the moisture it was supposed to keep out. Outdoor potting faces a combination of thermal, UV, and environmental stress that indoor potting never encounters, and standard formulations degrade quickly once UV enters the picture. What Makes Outdoor Exposure Different Solar UV breaks polymer bonds directly: unprotected potting yellows within months and turns brittle within one to two years, at which point cracking opens a path for moisture. Outdoor components also see more severe thermal cycling than typical industrial duty — commonly −40°C to +80°C multiple times daily between night and day, cloud and sun — and that cycling arrives alongside direct rain, standing water, and sustained high humidity that potting has to resist for years, not months. Coastal or roadside deployments add salt spray or airborne chemical contamination on top, and the worst combination is thermal cycling plus moisture together: a component wet from rain, then heated by sun, drives internal moisture migration and stress accumulation that neither factor produces alone. How the Options Compare Standard epoxy is the wrong default outdoors — UV resistance is poor, it yellows within six months and degrades within one to two years, and moisture absorption (1–3%) accelerates in the wet/dry cycling outdoor equipment sees. Expect field failure within 12–18 months. Silicone is the strongest performer: UV resistance holds for 3+ years, moisture absorption stays low (0.5–1.5%, hydrophobic), and its inherent flexibility tolerates the full −40°C to +100°C swing outdoor deployments demand, at $80–150/lb. UV-stabilized epoxy splits the difference — formulated with UV absorbers to prevent yellowing, it holds up reasonably well for 2–3 years at $60–100/lb, a useful compromise where silicone's cost isn't justified. Polyurethane is the one family to avoid outdoors regardless of its other merits: UV resistance is poor enough that visible degradation shows up in 3–6 months. The Properties That Actually Matter UV absorbers, antioxidants, and light stabilizers cost $10–30/lb to formulate in but are non-negotiable for outdoor service — validate any "UV-stabilized" claim against ASTM G154, the standard practice for operating fluorescent UV lamp apparatus for exposure testing of nonmetallic materials, rather than taking the label at face value. Moisture resistance should target under 0.5% absorption, achieved through hydrophobic resin selection and minimal hygroscopic filler, and validated through 1,000+ hours at 85°C/85% RH. Low CTE (35–45 ppm/°C) limits thermal-cycling stress on solder joints given the wide swings outdoor equipment experiences, and 8–10% elastomer toughening absorbs the vibration that wind load or traffic adds on top of thermal stress. Salt-fog resistance — validated per ASTM B117, the standard for operating salt spray (fog) apparatus — depends on low moisture absorption, corrosion inhibitors in the formulation, and potting thickness above 3mm to provide multiple protective layers. What the Numbers Look Like in the Field Across 500 units…

A potting decision splits into two competing requirements. One assembly dissipates 100W and needs potting that conducts heat away from hot components. Another houses a temperature sensor that needs potting that insulates it from external heat. No single compound satisfies both. Choosing between thermally conductive and thermally insulating potting starts with an honest read of your application's power dissipation and sensing requirements, not the compound's marketing name. Thermally Conductive Potting (2–4 W/m·K) Conductive potting is filled with aluminum oxide, boron nitride, or silicon carbide particles that build a heat-conduction pathway through the matrix. Thermal conductivity reaches 2–4 W/m·K versus 0.2–0.5 W/m·K for unfilled resin, at 1.5–2x the material cost. Filler loading raises viscosity, so vacuum de-gassing is usually required to avoid trapped voids, and ultimate strength drops slightly unless the formulation includes elastomer toughening. The payoff is real: conductive potting reduces component hot spots by 15–25°C through heat spreading, which extends service life proportionally. It suits power supplies above 50W, LED drivers, motor drives, and automotive under-hood electronics with continuous power draw. Its one drawback for mixed assemblies — it shunts heat toward any nearby temperature sensor, corrupting the reading. Thermally Insulating Potting (0.2–0.8 W/m·K) Unfilled or lightly filled epoxy, polyurethane, or silicone conducts heat slowly by design. Lower viscosity means easier pouring and, for non-critical assemblies, no vacuum step at all. Without filler constraints, formulators can optimize for strength, elastomer toughening, and low CTE simultaneously — properties that are harder to balance once thermal filler enters the mix. Insulating potting is the default choice for control circuits, precision measurement electronics, and — critically — any assembly built around a temperature sensor, since it won't shunt ambient heat away from the sensing element. Its weakness shows up only when power dissipation climbs: trapped heat can push components 10–20°C above what an equivalent conductive-potted design would reach, and thermal cycling stress concentrates in those unmanaged hot spots. Selective Potting for Mixed Assemblies When one assembly carries both a high-power section and a sensor, pot them separately. Encapsulate the power components with conductive compound, pot the sensor region with insulating compound, and connect the two regions with thin wires or a thermal isolator to limit coupling. A battery management system with high-current MOSFETs and onboard temperature sensors illustrates the approach: the MOSFETs get 3 W/m·K conductive potting for cooling, the sensor gets 0.3 W/m·K insulating potting to preserve calibration, and only signal wires cross between the two regions. The two-pour approach costs more than a single monolithic pour but avoids compromising either function — a trade-off examined in more depth in Incure's guide on whether high-temperature potting compound affects heat dissipation. Quantifying the Difference On a 100W power supply, an unencapsulated MOSFET runs near 150–160°C. Potted with insulating compound (0.3 W/m·K), the same MOSFET climbs to roughly 165°C because trapped heat has nowhere to go — 5°C worse than no potting at all. Potted with conductive compound (3 W/m·K), the MOSFET drops to about 135°C, a 30°C swing versus the insulating option.…

A power transformer rated 150°C operates in a 120°C ambient environment; its copper windings self-heat to 160°C. The potting is standard "high-temperature" epoxy with 180°C Tg. After 3 years, insulation resistance drops sharply and the transformer fails — not from a component defect, but from potting degradation and moisture ingress the epoxy was never formulated to resist. Transformers, coils, and motors have unique potting demands: continuous high temperature (not just cycling), high electrical stress, moisture exposure, and a service life measured in decades. Standard potting is inadequate here; specialized transformer-grade potting is required. Thermal and Electrical Stresses Unique to Transformers Copper and core losses generate continuous heat, pushing hotspot temperature to 150–200°C for years at a time — a far more severe duty cycle than the intermittent stress of thermal cycling. Insulation stress from 400V–10,000V windings demands dielectric strength above 15 kV/mm even at 150°C, not just at room temperature. Outdoor or wet-environment transformers absorb moisture during cool-down cycles, and that moisture reduces insulation resistance while enabling electrochemical corrosion of the windings. Compounding all of this, transformers are typically expected to run 20–30+ years, so potting must hold its properties over a duration most other electronics never approach. How Transformers Punish Potting Differently Moisture at the copper-potting interface initiates electrochemical corrosion, and under voltage stress, ionic migration accelerates it — corroded copper raises winding resistance, which generates more heat, which accelerates corrosion further in a self-reinforcing loop. Unlike thermal cycling, sustained high temperature causes oxidative degradation of the resin itself: over years the potting becomes brittle, loses mechanical support, and develops micro-cracks that let moisture in faster. Dielectric strength erodes with both heat and moisture — a new potting rated 18 kV/mm can drop to 10–12 kV/mm at 150°C and 3% moisture absorption, marginal for a 400V transformer facing 10 kV peak switching transients. And because copper expands at roughly 17 ppm/°C against steel core's 12 ppm/°C, potting CTE has to be chosen carefully to accommodate that mismatch without overstressing either the windings or the potting itself. Performance Requirements for Transformer Duty Dielectric strength should exceed 15 kV/mm at 23°C per ASTM D149, stay above 12 kV/mm at 150°C, hold above 10 kV/mm after moisture conditioning (85°C/85% RH, 500 hours), and remain above 10 kV/mm after 1,000 hours of aging at 150°C — most standard potting drops to 8–10 kV/mm under any one of these conditions, an inadequate margin for transformer duty. Moisture resistance should stay under 0.3% absorption per ASTM D570 with insulation resistance above 10 MΩ retained after conditioning. Copper corrosion resistance should show no visible corrosion after 500 hours at 85°C/85% RH and no electro-migration under 150V stress at elevated temperature. Long-term aging should retain 80%+ tensile strength and 90%+ dielectric strength after 1,000 hours at 150°C. Mechanically, target CTE 30–45 ppm/°C, elongation at break above 3%, and modulus 2–5 GPa — stiff enough to support windings, compliant enough to limit stress. Matching Material to Transformer Class Standard epoxy is unacceptable for this duty: dielectric strength drops…

A pressure sensor deployed in a 250°C industrial furnace fails within weeks — not because the sensor itself can't handle the heat, it's rated for 250°C, but because the potting encapsulating it was never built for that duty. Sensor potting fails in ways power-electronics potting never does. Sensor potting has to preserve calibration accuracy and signal integrity, not just survive thermal exposure — a materially different bar than the one power-supply potting has to clear. What Makes Sensors Different Potting that expands significantly under thermal stress mechanically loads the sensor element itself, shifting its response curve. A pressure sensor potted with an incompatible compound can drift 1–5% per 50°C swing purely from mechanical stress, with no change in the sensor's underlying physics. Moisture compounds the problem for humidity and pressure sensors specifically, and any potting with meaningful electrical conductivity or temperature-variable conductivity can add noise to high-impedance outputs like thermistors, strain gauges, or capacitive elements. Layer daily thermal cycling on top, and delicate sensor elements fatigue faster than the rest of the assembly. Matching Potting to Sensor Type RTDs and thermistors need low thermal conductivity (0.2–0.5 W/m·K) so potting doesn't shunt heat away from the sensing element, plus low CTE to limit mechanical loading on the resistance element. Pressure sensors and load cells are more demanding still — a diaphragm or strain gauge translates CTE mismatch directly into output error, so very low-CTE potting (20–30 ppm/°C) with strong moisture resistance is the baseline. Capacitive sensors (humidity, displacement, pressure) are sensitive to the potting's own dielectric constant, which needs to stay within ±5% across the operating range and demands moisture absorption under 0.2%. Integrated sensor-plus-electronics modules split the difference, needing moderate thermal conductivity (1–2 W/m·K) to manage IC heat alongside low CTE for the sensor itself. Standard epoxy — CTE 50–70 ppm/°C, often 1–3% moisture absorption — is generally inadequate for anything but the least precise sensors. Low-CTE potting (25–35 ppm/°C, <0.2% absorption, roughly 2x the cost) handles pressure, displacement, and load-cell duty well. Silicone, at 3–5x standard epoxy cost, adds superior thermal-cycling tolerance and hydrophobic moisture resistance for the most demanding precision applications. Thermally conductive low-CTE blends (2–3 W/m·K, 30–40 ppm/°C) suit intelligent sensor modules combining sensing and signal conditioning on one board. Isolating the Sensor Element Mechanically Where possible, don't pot directly over the sensing element. Pot the signal-conditioning electronics separately, connect to the sensor through thin wires that act as a stress-relief conduit, and encapsulate only the junction area — leaving the element itself unpotted or minimally covered. Multi-element sensors (thermocouples, differential pressure) benefit from a layered approach: compliant potting on the base wiring, a stiffer structural layer above it, and minimal potting thickness directly over the elements themselves. For temperature sensors specifically, avoid thermally conductive potting near the element — heat shunted in from external sources through a conductive path is exactly what you're trying to prevent. Always verify sensor output after potting: cure for the full 48 hours, then calibration-check at room temperature and full operating…

An aircraft engine starts cold at −40°C. Within seconds, nearby BMS and power electronics see a jump to 100°C in under 30 seconds as combustor heat radiates through the nacelle. That extreme thermal transient stresses potting compound and embedded components simultaneously — and it's a different failure mode than the thermal cycling most datasheets report. Standard potting compounds, designed for gradual cycling, can crack or delaminate under rapid temperature swings. Thermal shock tolerance is a distinct property from thermal cycling endurance, and confusing the two is a common specification mistake. Thermal Shock vs. Thermal Cycling Gradual thermal cycling changes temperature at under 1°C per minute; stress distributes through the potting and components over minutes, strain accumulates gradually, and failure comes from cumulative fatigue over many cycles. Thermal shock changes temperature at more than 10°C per minute; stress concentrates at component surfaces and interfaces, strain is nearly instantaneous, and failure comes from crack initiation at a localized stress concentration. A potting formulation that survives 1,000 gradual thermal cycles may crack after just 50 thermal shocks — the two stresses are not interchangeable, and passing one test says nothing about the other. Where the Stress Concentrates When temperature changes rapidly, the potting's surface cools or heats before its interior, creating a temporary internal-to-surface temperature gradient. A 30°C rise applied instantly can create enough stress difference to exceed the potting's local tensile strength and initiate a crack. Severity depends on thermal diffusivity (lower diffusivity means a steeper gradient and higher stress), CTE (higher CTE means more volume change per degree), and mechanical toughness (brittle materials crack at lower stress than tough, high-elongation materials). Thermal shock testing is typically done by cycling between temperature extremes rapidly — IEC 60068-2-14 covers standard temperature cycling at 3–5°C/minute, while ASTM D2305 and liquid-to-liquid immersion methods push samples between hot and cold baths in seconds. The qualification framework in IPC-CC-830 references similar thermal-shock conditioning for insulating compounds used on printed wiring assemblies. Most suppliers provide gradual thermal cycling data but not true thermal shock data — worth asking about explicitly, since the absence of the data is itself informative. Failures show up as surface micro-cracks radiating from stress concentration points, delamination from the PCB due to expansion mismatch, component lead cracking from rigid potting stressing leads during the transient, or internal cracks around pre-existing voids that aren't visible until the assembly fails in the field. What Makes a Formulation Shock-Tolerant Higher thermal conductivity flattens the interior-surface gradient and reduces stress — thermally-conductive potting handles shock better than insulating potting. Elastomer toughening absorbs stress through particle deformation, letting the compound tolerate higher localized stress before cracking. Low CTE reduces the magnitude of stress from any given gradient, and elongation at break above 5% allows localized strain without cracking. A potting with Tg 200°C also tolerates shocks up to 160–180°C better than a rigid compound with Tg 150°C, since operating well below Tg keeps the material in its stiffer, more predictable regime. For rapid transients above 10°C/minute, target: thermal conductivity…