๐ก In This Article
- Why an epoxy adhesive with excellent initial adhesion can lose adhesion after reliability testing
- How differences in coefficient of thermal expansion (CTE) affect the adhesive interface
- How cracks initiate and propagate under repeated thermal shock conditions
In the previous article (Does More Toughener Always Mean Better Adhesion?), I covered the optimum toughener content and toughening mechanisms used to improve the impact resistance of epoxy adhesives. Flexibility and toughness matter as much as initial adhesion, but there’s a harder problem that keeps coming up in real-world development and evaluation work.
I hear some version of this question constantly:
“The initial adhesion was excellent, but the adhesive failed after the reliability test.”
The instinctive assumption is that the adhesive simply wasn’t strong enough. Failure analysis usually tells a different story: the real culprit is stress accumulated during repeated thermal cycling โ not a lack of bonding strength. This matters most in automotive electronics and other electronic materials, where long-term reliability outweighs initial bond strength. A product can look completely fine right after bonding and still develop failures after repeated exposure to changing temperatures.
What Is Thermal Shock?
Thermal shock is a reliability test that repeatedly exposes a product to low- and high-temperature extremes. Typical conditions include:
- โ40ยฐC โ 85ยฐC
- โ40ยฐC โ 125ยฐC
- โ55ยฐC โ 150ยฐC
Depending on the application, automotive electronics, battery systems, and semiconductor packages can undergo hundreds or even thousands of cycles.
The point of thermal shock testing isn’t just whether a material survives a temperature extreme โ it’s whether the bonded structure can withstand the mechanical stress that repeated temperature swings generate. That distinction is central to evaluating adhesive reliability.
Why Does Thermal Shock Generate Stress?
The main driver is the mismatch in coefficient of thermal expansion (CTE) between the bonded materials. Take a simple case: an adhesive bonding aluminum to glass. Both expand as temperature rises, but not by the same amount โ aluminum has a CTE of roughly 23 ppm/ยฐC, while glass runs much lower, depending on composition.
The adhesive layer sits between two materials trying to move by different amounts. As temperature rises, that mismatch builds stress in the adhesive layer and at the interfaces; as it falls, the materials contract and the stress reverses direction. Every thermal cycle puts the adhesive through this expansion-contraction cycle again.
Typical CTEs of Common Substrates
| Substrate | Typical CTE (ppm/ยฐC) | Key Consideration in Thermal Shock Testing |
|---|---|---|
| Silicon wafer / Glass | 2.5โ4.0 | Very low expansion; limited ability to accommodate deformation |
| FR-4 PCB substrate | 14โ18 (X/Y) | Anisotropic thermal deformation must be considered |
| Aluminum (Al) | ~23.0 | High expansion can generate significant stress when bonded to low-CTE materials |
| Neat Epoxy | ~50โ80 | CTE can increase sharply above Tg; filler loading is often needed to bring it down |
Actual values vary by material grade, formulation, temperature range, and measurement direction. What matters isn’t the absolute number โ it’s the CTE mismatch between the bonded materials.
Why Does the Adhesive Look Fine Initially but Fail Later?
This is a hallmark of thermal-shock-related failure. Right after bonding, an adhesive can pass peel testing, shear testing, and visual inspection with no apparent problems, and the picture only changes after repeated thermal cycling.
As thermal stress accumulates, microscopic cracks can begin forming at the interface or within the adhesive layer. At first these have almost no measurable effect on adhesion, but as cycling continues, they gradually propagate โ a defect too small to detect even under a microscope can grow into a significant crack. Eventually the remaining bonded area can no longer carry the applied stress, and adhesion drops sharply or the interface delaminates outright.
That’s the mechanism behind adhesives that look excellent initially but fail after reliability testing. The adhesive didn’t weaken overnight โ a small defect grew under repeated mechanical stress until it hit a critical threshold.
Which Adhesives Are Particularly Vulnerable?
Highly rigid adhesive systems tend to be more vulnerable to thermal-shock-induced failure, and a high initial bond strength doesn’t guarantee high reliability. A highly crosslinked, rigid epoxy can show excellent initial adhesion and strength, but if the cured network can’t accommodate repeated thermal strain, stress builds up at the interface or within the layer, and cracks are more likely to initiate and spread.
This is why stress relaxation capability matters so much in electronic-material adhesives. Tougheners and similar modifications aren’t added just to soften the adhesive โ they help it absorb and redistribute mechanical stress. That said, if a formulation is already flexible enough, softening it further isn’t necessarily the answer; tuning Tg, crosslink density, and network structure is often more effective at that point.
The goal was never simply “soft” or “hard.” It’s finding a formulation that manages the stresses the adhesive will actually see in service.
How Is Thermal-Shock Failure Analyzed?
Checking the adhesion value alone rarely tells the full story. Several methods are typically used together:
| Analysis Method | What It Can Reveal |
|---|---|
| Cross-sectional analysis | Cracks, voids, and interfacial conditions |
| Peel / shear test | Changes in mechanical adhesion strength |
| Microscopic observation | Location and morphology of failure |
| SEM analysis | Microcracks and detailed fracture morphology |
| TMA | CTE and dimensional changes with temperature |
| DMA | Temperature-dependent mechanical behavior |
In practice, the location and mode of failure can be more informative than the adhesion number itself. A drop in shear strength tells you something changed; knowing whether the failure was adhesive (at the interface) or cohesive (within the adhesive) tells you where the actual problem lies.
The Adhesive May Not Have Failed Because It Was Too Weak
Thermal-shock failure analysis often lands on an unexpected conclusion: the adhesive had plenty of initial bonding strength โ it just couldn’t withstand repeated stress over time.
That distinction matters. A formulation selected purely on initial adhesion can perform beautifully in the lab and still cause problems once it’s in mass production. Adhesive development can’t be reduced to “how strong is the bond?” The better question is how well the bonded system holds up under the stresses it will actually experience โ and that’s a fundamentally different lens for thinking about reliability.
Thermal Shock Is Not the Only Reliability Test
This article has focused on thermal shock, but it’s only one piece of reliability evaluation for automotive and electronic materials. Other key tests include:
- 85ยฐC / 85% RH
- PCT (Pressure Cooker Test)
- HAST (Highly Accelerated Stress Test)
Thermal shock mainly evaluates repeated temperature swings and the mechanical stress they cause; high-temperature/high-humidity testing focuses more on moisture penetration, interfacial degradation, and long-term environmental durability. It’s entirely possible for an adhesive to show excellent initial adhesion and still degrade faster than expected under heat and humidity. I’ll cover moisture-related reliability failure in the next article (Why Does Epoxy Look Fine When Dry but Fail After Absorbing Moisture? โ โ Why 85/85, PCT, and HAST Testing Exist).
A Bond Is More Than a Single Strength Number
Every time I dig into a thermal-shock failure, the same lesson comes back: a single adhesion value can’t fully describe how reliable an adhesive system is. A product can look perfectly healthy right after bonding and still develop a failure mechanism over time, because an adhesive’s job isn’t just to hold two substrates together โ it has to keep absorbing and redistributing mechanical stress for its entire service life.
That’s why adhesive development takes more than maximizing initial adhesion. A formulation has to be built around the real combination of adhesion, cohesion, toughness, CTE, Tg, stress relaxation, interface stability, and environmental resistance. The best adhesive isn’t the one with the highest initial adhesion value โ it’s the one that holds up after thousands of thermal cycles and years in the field. That’s where the real work of adhesive development begins.
