Microvia Failure Analysis
Quality & Inspection Standards
PRODUCTS CENTER
Get Custom Quote
PRODUCTS CENTER
Microvia Failure Analysis – IPC Class 2/3 Acceptance & HDI Reliability
Microvia failures are the number one reliability risk in HDI boards – and they are often invisible until after thermal cycling. We have seen 0.5mm pitch BGAs pass electrical test, pass as‑polished microsection, then fail catastrophically after 500 thermal cycles because of microvia corner cracks or target pad separation. The physics is unforgiving: CTE mismatch between copper (17 ppm/°C) and the dielectric (50‑70 ppm/°C) concentrates strain at the microvia bottom and the via‑to‑via interface in stacked structures. This guide walks through IPC‑TM‑650 microsection techniques to identify the four primary microvia failure modes, and the acceptance limits per IPC‑6018 for Class 2 and Class 3.
Table of Contents
1. The Four Primary Microvia Failure Modes
Per IPC‑6018 §4.4 and §5.2, microvia failures fall into four distinct categories. Each has different root causes and acceptance rules.
| Failure Mode | Microsection Appearance | Primary Root Cause |
|---|---|---|
| Target pad separation | Gap between microvia copper and target pad; often at one edge | Laser smear, inadequate desmear, poor electroless adhesion |
| Corner crack at via base | Crack initiating at the 90° corner where microvia meets target pad | Stress concentration from CTE mismatch, brittle copper |
| Stacked interface separation | Gap between first‑level and second‑level microvia copper | Incomplete recrystallization, oxide contamination at interface |
| Microvia void (bottom or barrel) | Missing copper plating in the via, often at the target pad center | Plating bath contamination, insufficient agitation, laser debris |
Table 1: Microvia failure modes defined in IPC‑6018 and IPC‑A‑600J.
2. Target Pad Separation – The HDI Class 3 Showstopper
Target pad separation is the most common microvia failure in our lab. Per IPC‑6018 §5.2, for Class 3 products, any visible separation between the microvia and its target pad after thermal stress is rejectable. For Class 2, separation up to 10% of the via circumference is acceptable provided the remaining contact area has at least 75% of the target pad width.
Root causes we have identified:
- Laser smear (charred resin) left on the target pad after CO₂ or UV drilling. Desmear process (plasma or permanganate) must remove all residue.
- Insufficient electroless copper deposition: minimum thickness per IPC‑6018 is 0.5µm in the microvia bottom. We have seen labs with only 0.2µm – separation is guaranteed.
- Glass fiber protrusion: glass bundles that are not fully vaporized by the laser, creating a non‑uniform surface for plating.
In one AI server program, we rejected three consecutive HDI prototype lots for target pad separation >15%. The supplier had reduced plasma desmear time by 30% to increase throughput. The cost of requalification was $60,000 – but catching it before production saved over $2M in field failures.
3. Corner Cracks at the Microvia Base
Corner cracks initiate at the 90° junction between the microvia barrel and the target pad. Per IPC‑A‑600J §3.3.9, these are treated as Type B or Type E copper cracks but specific to microvias. The mechanism is stress concentration: during reflow (260°C peak), the target pad expands radially, pulling the microvia wall outward. The corner experiences the highest strain.
Acceptance per IPC‑6018 §5.2:
- Class 3: Zero corner cracks allowed after thermal stress.
- Class 2: Cracks ≤10% of the via circumference are acceptable, provided they do not propagate into the target pad.
We have seen corner cracks in microvias with aspect ratios >1:1 (depth/diameter). IPC‑6018 §4.4.1 recommends a maximum aspect ratio of 1:1 for reliable microvias. Designs exceeding 1.2:1 have a 3x higher crack rate in our data.
4. Stacked Microvia Interface Fatigue
Stacked microvias (2‑n‑2 or 3‑n‑3 builds) have an interface between the first‑level and second‑level microvia copper. This interface is a plane of weakness because the electroless seed layer and the subsequent electrolytic copper do not fully recrystallize. Per IPC‑6018 §5.2, after thermal stress, any separation at the stacked interface is rejectable for Class 3.
In a 2‑n‑2 build, we have observed that the failure rate after 1000 thermal cycles (-40°C to +125°C) is 4x higher than 1‑n‑1 builds. The third interface in 3‑n‑3 builds amplifies the risk non‑linearly. Our design rule: Avoid stacked microvias beyond 2 levels unless the application requires >1000 cycles and you have validated the material stack with IPC‑TM‑650 thermal shock testing.
Engineering Insight
Stacked microvias are not inherently unreliable – but they require perfect process control. We have seen suppliers achieve <0.1% failure with 2‑n‑2 when using plasma desmear and pulse plating. The difference is process maturity. Always audit your supplier's HDI capability before committing to stacked microvias.
5. Acceptance Criteria: Class 2 vs Class 3 per IPC‑6018
| Failure Mode | Class 2 | Class 3 | IPC Reference |
|---|---|---|---|
| Target pad separation | Accept ≤10% of via circumference, pad width ≥75% of min | Zero separation | IPC‑6018 §5.2 |
| Corner crack | Accept ≤10% of circumference, no propagation into pad | Zero cracks | IPC‑6018 §5.2 / IPC‑A‑600J |
| Stacked interface separation | Accept if ≤5% of interface and isolated | Zero separation | IPC‑6018 §5.2 |
| Microvia void (bottom) | Accept ≤20% of via diameter, ≥80% contact remaining | Accept ≤10% of via diameter | IPC‑6018 §4.4.3 |
Table 2: Microvia acceptance criteria per IPC‑6018. For Class 3, target pad separation of any size is a lot rejection.
Microvia Failure Inspection Checklist (IPC‑Anchored)
- Examine as‑polished at 200‑500x. Record any separation or cracks before micro‑etch.
- Apply reduced etch (IPC‑TM‑650 Note 6.2). Standard etch can smear over microvia defects.
- Measure target pad contact percentage. For Class 3, must be 100%.
- Check corner radius. Cracks initiate at sharp 90° corners; a small fillet reduces risk.
- For stacked microvias, examine each interface separately. Separation at any interface is a failure for Class 3.
- Verify microvia aspect ratio ≤1:1 per IPC‑6018 §4.4.1. Higher ratios require supplier qualification.
- Document with 500x micrograph. Include scale bar and failure mode classification.
6. FAQ
Q: Can a microvia with target pad separation be repaired?
A: No. Target pad separation indicates a permanent bonding failure. The affected microvia cannot be reworked. For Class 3, any such panel must be scrapped. Prevention is the only cure: specify plasma desmear and electroless copper thickness verification.
Q: What is the maximum reliable aspect ratio for a microvia?
A: IPC‑6018 §4.4.1 recommends 1:1 (depth equal to diameter). With advanced plating (pulse reverse current), some suppliers achieve 1.2:1, but reliability drops sharply above 1:1. We have seen 1.5:1 microvias fail after only 200 thermal cycles.
Q: Why do stacked microvias fail more often than staggered?
A: Stacked microvias have a copper‑copper interface that is a plane of weakness. Staggered microvias land on solid copper pads, eliminating the interface. Stacked designs are sometimes necessary for routing density, but they require much tighter process control. Always validate with thermal shock testing.
Q: How many microsections are needed for HDI qualification?
A: Per IPC‑6018 Table 4‑1, at least 5 coupons per panel, with 3 microsections per coupon (one as‑polished, one after thermal stress, one after thermal shock for Class 3). For critical designs, we require 10 coupons to achieve 95% confidence.
Continue Reading – HDI & Microvia Series
Found microvia failures in your HDI boards? Our engineers will review your microsections against IPC‑6018, identify root cause (laser smear, plating, or material), and recommend corrective actions.
Request Microvia Failure Review →References: IPC‑TM‑650 2.1.1 Note 6.2, 2.4.13.1; IPC‑6018 §4.4, §4.4.1, §4.4.3, §5.2; IPC‑A‑600J §3.3.9; IPC‑2221B §9.1.2. IPC‑TM‑650 Official Method.
Get Custom Quote
PRODUCTS CENTER


