Reason for the capacitor failure?

[ar__systems]

I'm investigating capacitor failure in the production board. The cap is 1210 35V 4.7uF. I'm waiting for the part number from the assembler. Obviously I researched the issue already; and I've dealt with stressed cracked caps in the past, but this is something else. First of all, this cap is in the center of the board and the nearest mounting hole is 2 inches away. Also, the failure does not look similar to typical stress cracked caps. In fact they don't look like any other failed cap.

The cap was sitting between 24V line and ground, so pretty much it sees only DC. There were 5 more on the same bus they all are fine. The board was tested at production and I believe they run them for 1 hr during QC. It worked for some time in the field, and then failed.

Upon investigating the board the failed cap was identified. It was presenting itself as a 15.0 Ohm resistor. The failure originated in the cap: after it was removed from the circuit it worked fine.

So... I'm looking for some advice -- whether it is random manufacturing defect, or something else. Attached is a picture of the cap from a side. As you can see it is split perfectly.

The assembler was scrambling to get any caps, maybe they got some bad stuff...

Appreciate any help.

 

[chris_leyson]

That's one of the big problems with MLCCs higher capacitance means thinner layers and a lot more manufacturing steps. The more layers you put in then the probability of an axial fracture increases. In a high volume commericial product you would see leaded electrolytic caps used in that application reliability being the simple reason. Hopefully it's a one off defect in your case.

[wraper]

That line running along the cap looks like severe layer delamination. I'd assume it was defective from the beginning.

[Niklas]

Could it be related to the temperature profile during the soldering process? A larger package would most likely need a slightly slower temperature slope for both heating up and cooling down.