# Underfill, warpage, thermal paths and qualification

Canonical: https://semiaifoundry.com/learn/atlas/package-thermomechanical-reliability/
Revision: 2026-10-07
Rights: https://semiaifoundry.com/ai-atlas/licenses.html

Packaging / Assembly and reliability · Atlas v1.3.0

# Underfill, warpage, thermal paths and qualification

Connects package materials and stress histories to measurable failure mechanisms.

[Explore in Atlas](https://semiaifoundry.com/ai-atlas/#package-thermomechanical-reliability)Joint surface, thin die and complete package; deformation exaggerated

Learning objectiveSeparate assembly damage, thermomechanical mechanisms, screening and qualification.

## The mechanism

A package combines materials that expand differently and conduct heat differently. Underfill and molding support joints and protect structures, but also change stress, moisture response and thermal behavior. Lids and thermal interface materials help transfer heat toward cooling hardware; voids or poor contact can impair that path. Warpage changes connection geometry during assembly and later use. Qualification evaluates selected product/process populations under defined stresses, while production screening checks manufactured units. Failure analysis must connect electrical symptoms to physical evidence on identified samples. A thermal-cycle failure is evidence about a tested configuration; it is not by itself a calibrated prediction of field lifetime.

1.  Record incoming components, assembly materials and the baseline electrical/thermal condition.
2.  Evaluate deformation and contact during relevant assembly and operating temperatures.
3.  Apply qualification stresses chosen for plausible mechanisms and intended use conditions.
4.  Localize failures with correlated electrical, acoustic, X-ray, cross-section or thermal evidence.
5.  Use a validated acceleration model only when relating a specific mechanism to field conditions.

## Science and assumptions

### Thermal expansion mismatch

$$
\Delta\varepsilon\approx(\alpha_1-\alpha_2)\Delta T
$$

Free mismatch strain is approximately the difference between expansion coefficients times temperature change.

Symbols and units

Δε

Free thermal-expansion mismatch strain

dimensionless

α₁,α₂

Linear expansion coefficients of the two materials

K⁻¹

ΔT

Temperature change

K

Expansion coefficients α use K⁻¹ and temperature change ΔT uses K; mismatch strain is dimensionless.

Assumptions Free-expansion mismatch estimate, not actual joint stress. Geometry, moduli, constraint, viscoelasticity and plasticity govern the stress field.

## Engineering example

Synthetic materials differing by 10×10⁻⁶ K⁻¹ across a 100 K change have 0.001 free mismatch strain. That 0.1% estimate does not predict solder fatigue life without geometry and a validated mechanism model.

## Distinctions and tradeoffs

Passing a qualification sample is not proof that every unit has been stress-tested, and test hours do not directly equal field hours.

A stiffer or more conductive material can improve one requirement while worsening stress transfer, adhesion or process compatibility.

Accelerators require sustained high-power operation; reliable joints and thermal interfaces preserve useful performance after deployment.

## Specifications and interfaces

Stress variables

Temperature, humidity, voltage and cycling

Retain the actual waveform, dwell, bias and test population.

Thermal boundary

Junction → package → cooling interface

Thermal resistance depends on the named path and conditions.

### Interfaces

Assembled material stack Input

Joint, underfill, molding, TIM and lid histories

Qualification and failure evidence Output

Electrical, structural and thermal responses under defined stress

## Check your understanding

### What should precede a destructive cross-section of every failed sample?

1.  Convert cycle count directly into years of life
2.  Preserve identified samples and collect discriminating nondestructive evidence
3.  Assume every open is a solder crack

Answer and reasoning

Preserve identified samples and collect discriminating nondestructive evidence

Preserving evidence allows competing incoming, assembly and stress-driven explanations to be investigated.

## Evidence and scope

No HBM lifetime, product qualification standard or test acceptance criterion is invented. Illustrative deformation is exaggerated for teaching.

Use the manufacturing investigation to trace compatible inputs, assembly and operating constraints; no current standalone calculator models this lesson’s full mechanism.

## Sources

-   [Texas Instruments — Reliability testing](https://www.ti.com/quality-reliability/reliability/testing.html)
-   [imec — Wafer-to-wafer hybrid bonding: 400 nm interconnect pitch](https://www.imec-int.com/en/articles/wafer-wafer-hybrid-bonding-pushing-boundaries-400nm-interconnect-pitch)
-   [SK hynix — Small Size, Big Impact](https://news.skhynix.com/en/small-size-big-impact/)

[Published lesson data](https://semiaifoundry.com/ai-atlas/lesson-data/package-thermomechanical-reliability-723dba57eff2f204.json) · [Atlas releases and licensing](https://semiaifoundry.com/ai-atlas/licenses.html)

## Continue exploring

### Foundations

-   [Flip-chip assembly and organic substrates](https://semiaifoundry.com/learn/atlas/flip-chip-abf-substrate/)
-   [Microbumps, reflow and thermocompression](https://semiaifoundry.com/learn/atlas/bump-thermocompression-assembly/)

### Related lessons

-   [Thermal interface material](https://semiaifoundry.com/learn/atlas/thermal-interface/)
-   [Lid and heat spreader](https://semiaifoundry.com/learn/atlas/heat-spreader/)
-   [Hybrid bonding: surface state and alignment](https://semiaifoundry.com/learn/atlas/hybrid-bond-surface-alignment/)