Destructive Failure Analysis Methods for PCB Assemblies

Some failures cannot be seen from the outside. A cracked joint under a ball grid array, a delaminated interface, a void at the die attach and a crater in the laminate all require the assembly to be opened before the cause can be established. Destructive failure analysis is how that is done, and the choice of method determines what can be learned before the evidence is consumed.

Why Destructive Analysis Is Necessary

Non destructive methods show where a problem is, not why it happened. X-ray shows a void or an open, acoustic microscopy shows a delamination, and both are valuable screening tools. The mechanism, the interface that failed and the sequence of events require the joint to be examined directly under a microscope.

The analysis is also needed to distinguish between causes that look identical from the outside. A joint that failed by fatigue, one that failed by a brittle fracture and one that was never properly bonded can all appear as an open circuit, and the corrective action is different in each case.

Dye and Pry

Dye and pry is a method for examining the joints of an area array package. A dye penetrant is applied to the assembly and allowed to wick into any gap or crack, and the package is then pried from the board. The fracture surfaces are stained, so the extent and the location of the separation are visible when the package is examined.

The method is fast, needs no sectioning and preserves the fracture surface for examination, which makes it useful for a first look at an area array failure. It requires the package to be removed, so it is destructive, and the pry force has to be controlled so that good joints are not separated in a way that confuses the result. The method is often paired with the <a href="https://www.gopcba.com/bga-inspection-x-ray-vs-aoi-guide/” title=”X-ray inspection”>X-ray inspection that identified the suspect component.

Cross Sectioning

Cross sectioning cuts through the assembly and polishes the surface so that the internal structure can be examined at high magnification. It shows the joint shape, the intermetallic layers, the void distribution, the plating in a barrel and the condition of the laminate beneath a pad, all in one image.

The preparation determines the value of the result. The sample has to be cut in the right place, mounted so that the joint is supported and ground and polished without smearing the soft alloy. A poor preparation destroys the evidence, and the position of the cut has to be chosen from the failure location rather than selected for convenience. The technique is the same one described in PCB quality assessment for routine verification.

Delamination and Interface Analysis

Delamination between layers or between a coating and a board is examined by a combination of sectioning and selective removal of the layers. Where the interface is between a coating and the laminate, a peel or a tape test can show the location and the mode of the failure without destroying the whole assembly.

The examination should identify whether the failure is adhesive or cohesive. A failure at the interface between two materials indicates a preparation or a compatibility problem, while a failure within one of the materials indicates a bulk property problem. The distinction decides whether the corrective action is a surface treatment or a material substitution.

Planning the Analysis

The analysis has to be planned before the first cut. The questions should be defined, the method chosen to answer them and the sample selected accordingly. Where several assemblies are available, some can be sectioned and others used for a different method, which is more informative than repeating one method on several samples.

Photographs and measurements should be recorded at each stage, because the sample is consumed and the evidence cannot be revisited later. A record that includes the magnification, the measurement and the location lets the result be reviewed by someone else and compared with other samples from the same failure population.

Limitations and Sequence

Every destructive method answers a specific question and destroys its own evidence. Dye and pry shows the extent of separation but not the metallurgy. A cross section shows the structure at one plane but not the extent across the joint. Neither shows the loading history, which has to be inferred from the failure mode.

The sequence should therefore start with the least destructive and the most informative, and move to the more destructive only when the earlier methods have narrowed the question. Starting with a cross section on the first sample often wastes the only evidence available.

Preserving the Evidence Before Analysis

The evidence begins to degrade as soon as the assembly fails. Handling, storage, cleaning and even the act of removing a suspect part can destroy the fracture surface or introduce contamination that misleads the examination. The sample should be handled as little as possible, kept in a clean container and photographed in place before anything is removed.

The electrical history should be recorded with the sample. The conditions at failure, the number of cycles or hours of operation, and any prior rework all inform the interpretation of the fracture mode. Where a failure happened in the field, the sequence of operations before it failed is often the most valuable information available and the easiest to lose.

Examining Voids and Their Distribution

A cross section used to measure voids has to be interpreted carefully. A single plane through a joint can pass through the centre of a void and show a large area, or it can miss the void entirely. The void content is a volume property, so an imaging method that scans the joint in three dimensions is more appropriate for measuring it, and the section is used to confirm the interface condition.

The void distribution matters as much as the total. Voids concentrated at the interface reduce the bonded area, while voids in the bulk of the joint mainly affect the thermal path. The two have different consequences and different causes, and the examination should state which is present. The joint criteria that apply are described in solder joint acceptance criteria.

When to Use Each Method

Dye and pry is the first choice for an area array joint failure. A cross section is the choice for a joint structure, plating or laminate question. A peel or tape test is the choice for a coating or an interface question. Where the failure is in a component rather than on the board, the analysis moves to the component supplier, and the board level work stops at establishing that the board was not the cause.

The gopcb failure analysis group plans the sequence with the customer before any sample is cut, because the choice of method determines what can be concluded. Where the evidence is limited to one assembly, the plan is written down and agreed, and the photographs and measurements are recorded so that the conclusion can be reviewed rather than taken on trust.

FAQ

Can dye and pry be used on a leaded package? It is designed for area array joints. A leaded joint can be examined by sectioning, where the fracture surface is less confined.

Does a cross section show the whole joint? No. It shows the plane that was cut, which may miss a defect that is not at that position. Several cuts or a dye method may be needed.

Why not start with a cross section? Because it destroys the sample and may not be the method that answers the question. The sequence should start with the method that gives the most information for the least destruction.