GLOBAL

Delamination in Laminated Busbars – Causes and Prevention

11/08

2026

By A&JLINK Engineering Team · Last updated November 2026 · 9 min read
The short answer
Delamination is usually decided before the part reaches the press — the press only freezes a result.
  • ✓Read the failure mode first. Adhesive = pre-treatment · Cohesive = parameters · Edge peel = geometry or moisture.
  • ✓Thermal stress is an accelerator, not an origin.
  • ✓PD testing is the one non-destructive screen that works in production.
Most laminated busbar delamination investigations start in the wrong place. The part is cut open, the bond line looks weak, and the conclusion is "the adhesive failed" or "the press pressure was too low".
Both conclusions are usually wrong. The bond line looks weak because it never formed properly, and the reason for that sits upstream — on the surface of the copper, or in how the film was handled before it went into the stack.
Scope: this article is about what causes delamination and how to prevent it. It is not a defect catalogue (that is a separate article) and it does not cover which insulation film to choose — only why a chosen film can fail to bond.

1. Delamination Is Rarely the Adhesive — It Is the Interface

On laminated busbars, three things have to be true for a bond to form — and two of them are decided before the press closes.
The copper surface has to be active. Clean copper oxidises within hours in a normal workshop atmosphere. A surface that has been sitting, handled, or stored without protection is no longer the surface the process was qualified on.
The film has to be free of residue. Handling marks, dust, release-agent transfer from storage — all of it sits between the two materials that are supposed to bond.
The parameters have to be held across the whole part. Not at the centre thermocouple — across the platen, including the edges, which is where the heat loss is.
The press does not create a bond. It completes one that the surface condition has already made possible or impossible. That is why two factories with the same press and the same film can produce different failure rates: the difference is upstream.

2. Three Failure Modes

Reading a laminated busbar delamination case starts with one question: where did it separate? When a delaminated part is cross-sectioned, the failure mode itself tells you where to look — and it is the same set of modes you will find in the wider busbar failure modes picture.
Failure modeWhere it separatesWhat it tells you
AdhesiveAt the copper surfacePre-treatment problem — oxide or residue
CohesiveInside the bond lineParameter or material problem
Edge peelFrom a cut edge or windowStress concentration, or a moisture path
Adhesive failure is the most common and the most misread. The copper comes out clean — no residue of bond line on the metal. That is not a weak adhesive; it is evidence that the adhesive never adhered. Look at pre-treatment first, not at the press.
Cohesive failure means the bond formed but the bond line itself is weaker than it should be. Both surfaces carry material. This points at cure, dwell, or a material lot.
Edge peel starts at a cut edge or a window. It is a geometry and handling issue more than a bonding one — and it is the mode most likely to be triggered by moisture.

3. The Causes, in Order of Frequency

1. Interface contamination. The single most common cause, and the cheapest to fix. Oxidised copper, incomplete degreasing, residue on the film. It is invisible in the finished part and it is decided hours before lamination — and it is why the choice of laminated busbar insulation film matters less than how that film was stored and handled.
2. Parameter window too narrow. Temperature, pressure and dwell have to be held across the whole platen, not just at the centre. A window that works at the thermocouple but not at the edge produces parts that fail at the edge.
3. Mechanical stress in assembly. This one appears after the part leaves the factory: a bar forced into position, a bolt torqued beyond spec, a stack compressed during handling. The bond line survives the press and then gets peeled apart at installation.
4. Moisture ingress. Whether this matters depends entirely on the edge construction — see below.
Where thermal stress actually fits. CTE mismatch between copper and film is real, but it does not create delamination on its own. A sound bond line absorbs the differential through the adhesive's compliance; that is what the adhesive layer is for. Thermal cycling promotes an existing weakness — a marginal bond, a small void, an edge peel that had not propagated yet. Treat thermal stress as an accelerator, not an origin. It is why a part can pass at goods-in and fail after a year in service.
Moisture depends on the edge. With a closed edge, the bond line is sealed and moisture has almost no path in. With an open edge, the cut edge is an open door: moisture enters the interface, and at reflow or operating temperature it vaporises and lifts the bond. Writing about moisture without distinguishing the two is a sign that the author has not looked at the actual construction.

4. Why It Shows Up Late

Three reasons a laminated busbar delamination case is rarely caught at goods-in.
Non-destructive inspection is difficult on a multilayer stack. Ultrasonic inspection works, but attenuation through many layers limits resolution. You can see a trend; you cannot see a 50 µm void.
Initial peel strength can be adequate and still marginal. A bond that measures acceptably today can be sitting at the edge of its life. The number passes; the margin does not.
The trigger comes later. Thermal cycling, humidity, and mechanical stress in service are what convert "marginal" into "failed". A part that is fine on the bench and fails after eighteen months was not good and then broke — it was marginal the whole time.
That is the uncomfortable conclusion: goods-in pass is not evidence of a good bond, only of an adequate one today.

5. Prevention: Design Side and Process Side

Process side — where most of the gain is
  • Pre-treatment is the single most important action. Define the window between cleaning and lamination, and hold it. A copper surface left overnight is a different surface.
  • Qualify the parameter window at the edge, not the centre. Put thermocouples where the heat loss is.
  • Control cooling rate. Fast cooling builds in residual stress at the interface before the part ever sees a thermal cycle.
Design side — where geometry sets the risk
  • Avoid sharp internal corners. They concentrate peel stress exactly where delamination likes to start.
  • Keep windows and cut-outs sensible. Large openings remove the bonded area that holds the edge.
  • Avoid abrupt thickness changes. A step in the stack is a stress riser, and it is also where the parameter window is hardest to hold.
If you want to see how these are handled on our side, our laminated busbar factory page covers the pre-treatment and lamination controls behind the parts.

6. How It Is Detected

Four methods, in order of usefulness in production.
Partial discharge as an indirect indicator. This is the one worth knowing. Delamination lowers the PD inception voltage noticeably, and PD testing is non-destructive and can be run on every part. It does not tell you where the void is, but it tells you the insulation system is not what it should be — which is enough to pull a part. Our article on laminated busbar partial discharge covers what the measurement can and cannot prove.
Ultrasonic C-scan. Good for locating voids and mapping the bond line. Read the trend across a batch rather than the absolute value on one part, because attenuation limits what a single scan can claim.
Cross-sectioning. Destructive, slow, and still the only way to identify the failure mode — which is what tells you where the problem came from.
Humidity ageing followed by re-test. 85/85 for a defined period, then repeat PD or peel. This is the closest bench approximation to what an open-edge part experiences in service.

7. What We'd Recommend

  • Look at pre-treatment before the press. If the copper came out clean, the press was never the problem.
  • Cross-section one failed part and read the failure mode. Adhesive, cohesive or edge peel each point at a different process step.
  • Qualify the window at the edge. The centre thermocouple has never failed a part.
  • Treat thermal cycling as an accelerator, not a cause. It explains late failures; it does not explain the defect's origin.
  • If you remember one thing: delamination is decided before the press, and only revealed after it.

8. Frequently Asked Questions

Can a delaminated busbar be repaired?
No. Once the bond line has separated, there is no rework that restores the original interface. The part is scrap — which is why prevention is the only economic answer.
Why did it pass goods-in and fail later?
Because goods-in proves the bond is adequate today, not that it has margin. Thermal cycling, humidity and assembly stress convert a marginal bond into a failure.
How do I tell whether it is the supplier's process or my assembly?
Read the failure mode and the location. Failure at a cut edge or window points at geometry or handling; failure in the middle of a bonded area points at the lamination process.
Does a closed edge eliminate the moisture risk?
It removes the main path, not every path. The interface can still be reached at terminals and transitions, but the risk is a different order of magnitude.
Can PD testing replace cross-sectioning?
No. PD tells you a part is bad; cross-sectioning tells you why. Use PD to screen, cross-section to diagnose.

9. Summary

Laminated busbar delamination is usually decided before the press closes — on an oxidised copper surface, on a film with residue, in a parameter window that does not reach the edge. The failure mode tells you which of those it was: adhesive failure at the copper means pre-treatment, cohesive failure means parameters or material, edge peel means geometry or moisture. Thermal stress and humidity do not create the defect; they convert a marginal bond into a failed one, which is why the part can pass at goods-in and fail a year later. Prevent it upstream, screen it with PD, and diagnose it with a cross-section.
Send Us One Failed Part
A cross-section and a PD reading tell us more than a description. We will tell you which failure mode you have and which process step to look at.
Talk to an Engineer

References & Standards

  1. IEC 60216-1 — Electrical insulating materials — Thermal endurance properties — Part 1: Ageing procedures and evaluation of test results. International Electrotechnical Commission. webstore.iec.ch

Prev: none

Next: What to Send for a Laminated Busbar Quote: A Data Checklist

Recommended Reading

Enter your information for download

If you are interested in our products, please contact us as soon as possible!

Submit
%{tishi_zhanwei}%