Keywords: solder fatigue, Coffin-Manson, solder joint reliability, thermal cycling, fatigue life
Keywords: solder fatigue, Coffin-Manson, solder joint reliability, thermal cycling, fatigue life
Introduction
Solder joints are the structural connections that hold electronic assemblies together—both electrically and mechanically. Every time a powered electronic device heats up and cools down, the mismatch in coefficient of thermal expansion (CTE) between silicon, copper, solder, and laminate materials forces solder joints into cyclic plastic strain. Over thousands of thermal cycles, this repeated deformation accumulates fatigue damage until cracks initiate, propagate, and eventually cause electrical open circuits. Solder joint fatigue is the dominant failure mechanism in thermal cycling of PCBAs, accounting for an estimated 70% of field failures in thermal environments. The Coffin-Manson model, originally developed for metal fatigue in nuclear reactor applications, has become the cornerstone analytical tool for predicting solder joint fatigue life. This article provides a rigorous technical guide to solder joint fatigue mechanisms, the Coffin-Manson equation and its derivatives, application to different package types, and the impact of lead-free solder alloys.
Solder Joint Failure Mechanisms
Understanding solder fatigue requires understanding the physical mechanisms that cause it. Solder joint failure under thermal cycling is not a single process but a sequence of interacting degradation mechanisms.
CTE Mismatch: The Root Cause
The fundamental driver of solder fatigue is the difference in coefficient of thermal expansion between the materials joined by solder:
| Material | CTE (ppm/°C) | Role in PCBA |
|---|---|---|
| Silicon | 2.6–3.0 | IC die |
| Alumina (ceramic) | 6.0–7.0 | Ceramic package substrate |
| Copper | 16.5–17.0 | Pads, traces, PTH |
| FR-4 (in-plane) | 14.0–17.0 | Laminate substrate |
| FR-4 (Z-axis) | 50–70 | Laminate thickness |
| SAC305 solder | 21.0–23.0 | Solder joint |
| SnPb solder | 23.0–25.0 | Solder joint |
| Aluminum | 23.0 | Lead frame, heat sink |
When temperature changes, these materials expand and contract at different rates, imposing shear strain on solder joints. A BGA package with a silicon die on a BT substrate mounted on an FR-4 board has at least three CTE mismatches working simultaneously.
Fatigue Damage Accumulation
Solder joint fatigue progresses through three stages:
Stage 1: Crack Initiation - Micro-cracks form at stress concentration points: solder-component interface, solder-pad interface, or voids within the solder - Typical initiation period: 10–30% of total fatigue life - Driven by cyclic plastic strain accumulation and microstructural degradation
Stage 2: Crack Propagation - Micro-cracks coalesce and propagate through the solder bulk along the path of maximum shear strain - Propagation rate depends on strain amplitude, temperature, and solder microstructure - Typical propagation period: 60–80% of total fatigue life - Crack path: typically through the bulk solder near the component side for SAC alloys
Stage 3: Fracture - Crack propagates completely through the joint, causing electrical open circuit - Final fracture area typically shows classic fatigue striations on one portion and ductile overload on the remainder
Role of Intermetallic Compounds (IMC)
The interfacial intermetallic compound layer between solder and pad is both a necessary bond and a potential weakness:
| IMC Type | Location | Growth Rate | Impact on Fatigue |
|---|---|---|---|
| Cu₆Sn₅ | Solder/Cu pad interface | Moderate (temperature-dependent) | Brittle; crack initiation site |
| Cu₃Sn | Cu₆Sn₅/Cu interface | Slow | Generally beneficial for adhesion |
| Ag₃Sn | Solder bulk (SAC alloys) | Fast at high temperature | Can form plate-like brittle phases |
| Ni₃Sn₄ | Solder/Ni (ENIG) interface | Moderate | Alternative to Cu₆Sn₅; similar brittleness |
Excessive IMC growth (especially at elevated temperatures during service or testing) embrittles the interface, shifting the failure mode from bulk solder fatigue to interfacial fracture, which is more sudden and less predictable.
Low-Cycle Fatigue and the Coffin-Manson Equation
Distinguishing Low-Cycle from High-Cycle Fatigue
Solder joint fatigue under thermal cycling is classified as low-cycle fatigue because each thermal cycle produces plastic (permanent) deformation in the solder. This contrasts with high-cycle fatigue (e.g., vibration), where stresses are below the yield point and damage accumulates through elastic mechanisms.
| Fatigue Type | Strain Regime | Cycles to Failure | Dominant Model |
|---|---|---|---|
| Low-cycle fatigue | Plastic strain | 10²–10⁴ | Coffin-Manson |
| High-cycle fatigue | Elastic strain | 10⁵–10⁸ | Basquin (S-N curve) |
| Mixed regime | Elastic + plastic | 10⁴–10⁶ | Combined Coffin-Manson + Basquin |
The Coffin-Manson Equation
The original Coffin-Manson relationship, developed independently by L.F. Coffin (1954) and S.S. Manson (1953), relates plastic strain amplitude to fatigue life:
Δε_p = ε'_f · (2N_f)^c
Where: - Δε_p = plastic strain range per cycle - ε'_f = fatigue ductility coefficient (material constant, typically 0.2–0.6 for solder) - 2N_f = number of reversals to failure (1 cycle = 2 reversals) - c = fatigue ductility exponent (typically −0.4 to −0.7 for solder)
Rearranging to solve for fatigue life:
N_f = 0.5 · (Δε_p / ε'_f)^(1/c)
Application to Solder Joints
To apply Coffin-Manson to solder joints, we need to determine Δε_p—the plastic strain range in the solder joint per thermal cycle. This can be obtained through:
- Analytical calculation: Using CTE mismatch, temperature range, and joint geometry to estimate shear strain
- Finite Element Analysis (FEA): Nonlinear viscoplastic FEA models (e.g., Anand model) provide detailed strain distributions
- Experimental measurement: Moiré interferometry or digital image correlation can measure strain directly
For a simplified analytical estimate, the plastic shear strain range is:
Δγ_p ≈ (Δα · ΔT · L) / h
Where: - Δα = CTE mismatch (ppm/°C) - ΔT = temperature cycle range (°C) - L = distance from neutral point (DNP) to the joint of interest (mm) - h = solder joint height (mm)
This simplified form ignores elastic constraints and creep relaxation but provides an order-of-magnitude estimate useful for preliminary design.
The Engelmaier Model: Refining Coffin-Manson for Solder
While the basic Coffin-Manson equation captures the fundamental relationship between plastic strain and fatigue life, it does not account for several critical factors specific to solder joint reliability. Werner Engelmaier extended the model in the 1980s to address these limitations.
Engelmaier's Modifications
Engelmaier's model adds temperature and dwell time dependence through a modified fatigue exponent:
N_f = 0.5 · (Δγ_p / ε'_f)^(1/c)
Where the fatigue exponent c is not constant but depends on mean temperature and cycling frequency:
c = −0.442 − 6×10⁻⁴ · T_sj + 1.74×10⁻² · ln(1 + 360/t_dwell)
Where: - T_sj = mean cyclic solder joint temperature (°C) - t_dwell = dwell time at temperature extremes (minutes)
Engelmaier Model Parameters
| Parameter | SnPb (63/37) | SAC305 | SAC-Bi |
|---|---|---|---|
| ε'_f (fatigue ductility coefficient) | 0.325 | 0.231 | 0.183 |
| c (at 20°C mean, 10 min dwell) | −0.512 | −0.582 | −0.621 |
| c (at 50°C mean, 10 min dwell) | −0.530 | −0.600 | −0.639 |
| c (at 80°C mean, 10 min dwell) | −0.548 | −0.618 | −0.657 |
The temperature and dwell dependence of c reflects the role of creep in solder fatigue. At higher temperatures and longer dwells, creep relaxation reduces the effective plastic strain per cycle, but it also causes microstructural coarsening that accelerates crack propagation. The net effect depends on the specific alloy and conditions.
Engelmaier Acceleration Factor
The Engelmaier model produces an acceleration factor for converting between test and field conditions:
AF = N_field / N_test = (ΔT_test / ΔT_field)^(1/c) · (f_field / f_test)^(1/3) · [F(field) / F(test)]
Where: - ΔT = temperature range - f = cycle frequency - F = a factor accounting for the temperature dependence of c
This AF equation is widely used in IPC-9701 and industry reliability calculations.
Temperature Cycle Parameters and Their Effects
The parameters of a thermal cycle profile significantly influence the fatigue life measured in testing. Understanding these effects is essential for both test design and field life prediction.
Temperature Range (ΔT)
The temperature range is the single most influential parameter. Wider ΔT produces more thermal expansion mismatch and more plastic strain per cycle. In the Coffin-Manson framework, N_f is proportional to ΔT^(1/c), so even small increases in ΔT dramatically reduce fatigue life.
| ΔT (°C) | Relative Fatigue Life (c = −0.6) | Typical Application |
|---|---|---|
| 40 | 1.0 (reference) | Consumer, indoor |
| 80 | 0.132 | Industrial, general |
| 120 | 0.037 | Automotive, harsh |
| 165 | 0.011 | Aerospace, military |
| 200 | 0.006 | Extreme environment |
A product that lasts 5,000 cycles at ΔT=40°C may last only 55 cycles at ΔT=165°C.
Mean Temperature
Higher mean temperature increases creep contribution, which has two competing effects:
- Beneficial: Creep relaxation during dwell reduces peak stress, lowering the effective strain amplitude
- Detrimental: Creep causes microstructural coarsening, IMC growth, and grain boundary sliding, accelerating crack initiation and propagation
For SAC alloys, the net effect is typically detrimental above 75°C mean temperature, as creep becomes significant relative to the alloy's melting point (homologous temperature > 0.6 Tm).
Dwell Time
Longer dwell times allow more creep relaxation, which can reduce plastic strain amplitude in subsequent cycles. However, longer dwells at high temperatures also accelerate IMC growth and microstructural coarsening. The optimal dwell time for testing balances these effects:
| Dwell Time | Effect on Measured Life | Recommendation |
|---|---|---|
| < 5 min | Insufficient creep relaxation; unrealistically high strain | Too short for accurate life prediction |
| 5–10 min | Moderate creep; acceptable for comparative testing | Standard for IPC-9701 |
| 10–15 min | Near-complete creep relaxation; most accurate life prediction | Preferred for quantitative ALT |
| > 30 min | Excessive IMC growth; shifts failure mode | Too long; may not represent field cycling |
Ramp Rate
Ramp rate affects thermal gradients across the assembly, which can introduce additional stress:
- Fast ramp (> 20°C/min): Large thermal gradients between components and board; may cause warpage-induced stress in addition to CTE-driven shear
- Moderate ramp (10–15°C/min): Standard for IPC-9701; balances test time with realistic stress
- Slow ramp (< 5°C/min): Minimizes thermal gradients but extends test time significantly; rarely used except for very large assemblies
Package Type Comparison: BGA vs QFP vs Chip Components
Different package types impose different stress states on solder joints, resulting in significantly different fatigue lives under the same thermal cycling conditions.
BGA (Ball Grid Array)
BGA solder joints are spheres of solder that connect the package substrate to the board. Their fatigue behavior is characterized by:
- High DNP effect: Outermost joints (largest DNP) experience the most strain and typically fail first
- Under-die stress: Joints under the silicon die experience additional stress due to the CTE mismatch between silicon (3 ppm/°C) and substrate (15–17 ppm/°C)
- Failure location: Cracks typically initiate at the package side of the joint, near the pad-solder interface, and propagate through the bulk solder
- Void sensitivity: Voids in BGA joints significantly reduce fatigue life because they reduce the effective load-bearing cross-section
| BGA Type | Typical Nf (TC3, 500 cycles) | Primary Failure Location |
|---|---|---|
| Plastic BGA (PBGA) | 1,000–3,000 | Package-side interface |
| Ceramic BGA (CBGA) | 300–800 | Board-side interface (high CTE mismatch) |
| Chip Scale Package (CSP) | 1,500–4,000 | Bulk solder, mid-joint |
| Wafer-Level CSP (WLCSP) | 500–2,000 | Package-side interface |
QFP (Quad Flat Package)
QFP components have gull-wing leads that provide mechanical compliance, absorbing much of the CTE mismatch strain:
- Lead compliance reduces solder strain: The gull-wing lead bends, reducing the strain imposed on the solder joint
- Longer fatigue life: QFP joints typically last 2–5× longer than BGA joints under the same conditions
- Failure location: Cracks initiate at the heel of the solder fillet where stress concentration is highest
- Lead-free transition impact: Stiffer SAC solder reduces the lead compliance benefit, narrowing the life advantage over BGA
Chip Components (Resistors, Capacitors)
Chip components present a unique fatigue challenge because they are rigid ceramic bodies mounted directly on the board:
- Maximum CTE mismatch: Ceramic CTE (6–7 ppm/°C) vs. FR-4 CTE (14–17 ppm/°C) creates large shear strain
- Board flexure coupling: The rigid component constrains board expansion, creating flexure that adds to CTE-driven strain
- Failure mode: Cracks typically initiate at the component termination-solder interface and propagate horizontally through the solder
- Size effect: Larger components (e.g., 2512, 2010) experience more strain than smaller ones (0402, 0603) due to greater DNP
| Component Size | Relative Strain | Typical Nf (TC3) | Common Failure Mode |
|---|---|---|---|
| 0201 | 0.2× | > 5,000 | Rarely fails in thermal cycling |
| 0402 | 0.4× | > 4,000 | Termination crack |
| 0603 | 0.6× | 2,000–4,000 | Solder fillet crack |
| 0805 | 0.8× | 1,500–3,000 | Solder fillet crack |
| 1206 | 1.0× (reference) | 1,000–2,500 | Heel crack → bulk |
| 2512 | 2.5× | 300–800 | Termination crack |
Impact of Lead-Free Solder on Fatigue Life
The transition from tin-lead (SnPb) to lead-free (SAC) solder, driven by the EU RoHS directive, fundamentally changed solder joint fatigue behavior. SAC alloys have different mechanical properties, microstructures, and fatigue characteristics than SnPb.
Mechanical Property Comparison
| Property | SnPb (63/37) | SAC305 | SAC-Bi |
|---|---|---|---|
| Melting point (°C) | 183 | 217–220 | 207–213 |
| Young's modulus (GPa) | 30–35 | 45–50 | 40–45 |
| Yield strength (MPa) | 30–40 | 25–35 | 35–45 |
| UTS (MPa) | 40–50 | 40–50 | 50–60 |
| Elongation (%) | 35–50 | 30–45 | 20–35 |
| CTE (ppm/°C) | 23–25 | 21–23 | 20–22 |
| Creep rate at 25°C | High | Low | Moderate |
| Creep rate at 80°C | Very high | Moderate | High |
Fatigue Life Comparison
The fatigue life difference between SnPb and SAC depends on the test conditions:
| Condition | SnPb Nf | SAC305 Nf | Ratio (SAC/SnPb) |
|---|---|---|---|
| TC1 (0–100°C, 10 min dwell) | 3,000–6,000 | 5,000–10,000 | 1.5–2.0× |
| TC2 (−25–100°C, 10 min dwell) | 1,500–3,000 | 2,500–5,000 | 1.3–1.8× |
| TC3 (−40–125°C, 10 min dwell) | 500–1,500 | 800–2,000 | 1.2–1.5× |
| TC3 (−40–125°C, 1 hr dwell) | 300–800 | 300–600 | 0.8–1.0× |
Key observations: - SAC outperforms SnPb at moderate conditions (low ΔT, short dwell) because its higher stiffness and lower creep rate result in less plastic strain per cycle - SAC advantage diminishes at harsh conditions (high ΔT, long dwell) because creep and microstructural coarsening become dominant - At very long dwells, SAC may underperform SnPb because SAC's microstructure (Ag₃Sn plates, Sn grain colonies) coarsens more rapidly than SnPb's equiaxed structure
Microstructural Effects
SAC solder has a complex microstructure that evolves during thermal cycling:
- Sn-rich dendrites: The matrix is large β-Sn grains that form colonies. Cycling causes grain boundary sliding and recrystallization at colony boundaries.
- Ag₃Sn intermetallics: Fine Ag₃Sn particles strengthen the as-solidified alloy but coarsen during thermal cycling, reducing strengthening effect.
- Cu₆Sn₅ particles: Distributed in the bulk; also coarsen over time. Their distribution affects crack path.
- Recrystallization: During cycling, new fine equiaxed grains form along crack paths, providing an easier fracture path. This phenomenon is more pronounced in SAC than in SnPb and is a key mechanism in SAC fatigue.
Coffin-Manson Exponent for Lead-Free Solder
The fatigue ductility exponent c for SAC alloys is typically −0.5 to −0.7, compared to −0.4 to −0.6 for SnPb. The more negative exponent means that SAC fatigue life is more sensitive to strain amplitude—it performs relatively better at low strain (moderate conditions) but degrades faster at high strain (harsh conditions).
| Solder Alloy | c (typical) | c range | Recommended for |
|---|---|---|---|
| Sn63/Pb37 | −0.51 | −0.42 to −0.60 | Legacy, repair |
| SAC305 | −0.58 | −0.50 to −0.70 | General lead-free |
| SAC405 | −0.59 | −0.51 to −0.70 | High-reliability lead-free |
| SAC-Bi (SAC105+Bi) | −0.62 | −0.55 to −0.72 | Low-Ag lead-free |
| SnAg (no Cu) | −0.55 | −0.48 to −0.65 | High-temperature |
Life Prediction Workflow
Putting it all together, a practical solder joint fatigue life prediction follows this workflow:
Step 1: Gather Input Data
- Component package type and dimensions
- Board material, thickness, and CTE
- Solder alloy and joint geometry
- Field temperature profile (min, max, mean, cycle frequency, dwell)
- DNP for the critical joint (usually the corner joint with largest DNP)
Step 2: Calculate Plastic Strain Range
- Use simplified analytical formula or FEA
- For BGA: Δγ_p ≈ (Δα · ΔT · DNP) / h
- For chip components: account for board flexure in addition to CTE mismatch
Step 3: Select Model Parameters
- Choose fatigue ductility coefficient ε'_f and exponent c based on solder alloy
- Adjust c for mean temperature and dwell time using Engelmaier's correction
Step 4: Calculate Fatigue Life
- N_f = 0.5 · (Δγ_p / ε'_f)^(1/c)
- Apply statistical correction: typical designs target 1% failure, so divide N_f by a safety factor (typically 2× for Nf at 50% failure → 1% failure)
Step 5: Calculate Acceleration Factor
- If using test data: AF = (ΔT_test / ΔT_field)^(1/c) · (f_field / f_test)^(1/3)
- Field life = test cycles × AF
Step 6: Validate
- Compare prediction with available test data
- Adjust model parameters if prediction deviates significantly (>2×) from test data
- Conduct failure analysis to confirm the assumed failure mechanism matches observations
Conclusion
Solder joint fatigue is the primary life-limiting failure mechanism for PCBAs in thermal cycling environments. The Coffin-Manson model, refined by Engelmaier for solder-specific behavior, provides a physics-based framework for predicting fatigue life from strain amplitude, temperature, and dwell time. The model's accuracy depends on correct identification of the dominant failure mechanism, accurate strain determination (analytical or FEA), and appropriate material parameters for the specific solder alloy.
The transition to lead-free solder has added complexity—SAC alloys generally outperform SnPb under moderate conditions but may underperform under harsh conditions due to microstructural coarsening. Engineers must use alloy-specific parameters and validate predictions with physical testing. When applied carefully, the Coffin-Manson framework enables data-driven design decisions that balance reliability, cost, and time-to-market.
Frequently Asked Questions
1. What is the Coffin-Manson equation and what does it predict?
The Coffin-Manson equation relates plastic strain amplitude to the number of cycles to failure in low-cycle fatigue. It states that N_f = 0.5 × (Δε_p / ε'_f)^(1/c), where Δε_p is the plastic strain range, ε'_f is the fatigue ductility coefficient, and c is the fatigue ductility exponent. In electronics, it predicts how many thermal cycles a solder joint can withstand before failure, based on the strain imposed by CTE mismatch.2. How does the Engelmaier model differ from the basic Coffin-Manson equation?
The Engelmaier model extends Coffin-Manson by making the fatigue exponent c temperature- and dwell-time-dependent, reflecting the role of creep in solder fatigue. It also incorporates cycle frequency effects and provides a standardized acceleration factor equation. The Engelmaier model is the basis for IPC-9701 solder joint reliability calculations and is more accurate for solder joints than the basic Coffin-Manson equation.3. Why do BGA solder joints fail faster than QFP joints in thermal cycling?
BGA solder joints are rigid spheres that directly transmit CTE mismatch strain to the solder, while QFP gull-wing leads provide mechanical compliance that absorbs much of the strain. The lead bending reduces the plastic strain in the solder joint, giving QFP joints 2–5× longer fatigue life. However, lead-free SAC solder is stiffer than SnPb, reducing this compliance benefit.4. How does lead-free SAC solder compare to SnPb in thermal cycling fatigue?
SAC solder generally lasts 1.2–2.0× longer than SnPb under moderate thermal cycling (ΔT < 100°C, short dwells) because its higher stiffness and lower creep rate result in less plastic strain per cycle. However, under harsh conditions (large ΔT, long dwells at high temperature), SAC's advantage diminishes or reverses because its microstructure coarsens more rapidly, and recrystallization along crack paths accelerates failure.5. What is the distance from neutral point (DNP) and why does it matter?
DNP is the distance from the center of the component to the solder joint of interest. In thermal cycling, the shear strain in a solder joint is proportional to DNP—joints farther from the center experience more displacement due to CTE mismatch. This is why corner joints on BGA packages typically fail first. Designers can reduce fatigue by placing high-reliability components near the board center or using smaller packages.6. How accurate are Coffin-Manson fatigue life predictions?
Coffin-Manson predictions are typically accurate within a factor of 2× when correct material parameters are used and the dominant failure mechanism matches the model's assumptions. Accuracy improves when the model is calibrated with test data at multiple conditions. Predictions are less accurate when: (1) multiple failure mechanisms compete, (2) the temperature range spans a material transition (e.g., Tg), or (3) the strain distribution is complex (e.g., under-chip stress concentration). Always validate predictions with physical testing.References
- IPC-9701A-2006, "Performance Test Methods and Qualification Requirements for Surface Mount Solder Attachments," IPC International. https://www.ipc.org/TOC/IPC-9701A.pdf
- Engelmaier, W., "Fatigue Life of Leadless Chip Carrier Solder Joints During Power Cycling," IEEE Transactions on CHMT. https://ieeexplore.ieee.org/document/1665631
- Lee, W.W., Nguyen, L.T., Selvaduray, G.S., "Solder Joint Fatigue Models: Review and Applicability to Chip Scale Packages," IEEE Transactions on Electronics Packaging Manufacturing. https://ieeexplore.ieee.org/document/820042
- Pang, J.H.L., "Lead-Free Solder: Mechanics and Reliability," Springer. https://link.springer.com/book/10.1007/978-1-4614-0462-2
- Syed, A., "Accumulated Creep Strain and Energy Density Based Thermal Fatigue Life Prediction Models for SnAgCu Solder Joints," ECTC Proceedings. https://ieeexplore.ieee.org/document/1469315
Meta Description: Technical guide to solder joint fatigue analysis using the Coffin-Manson model and Engelmaier extension, covering failure mechanisms, plastic strain calculation, package type comparison (BGA/QFP/chip), and lead-free solder impact on thermal cycling reliability.