Underfill Inspection Criteria

On "Flip Chip/uBGA Under-fill Visual Standards" �

The third "Satisfactory" [The under-fill if by design should be visible � ], here the notation "if by design" is unclear. Does "according to product documentation" get to the point?

The first "Acceptable", second sentence [Flow out on to test points �], move to "Unacceptable" portion of these criteria.

The second "Acceptable" [Small bubbles in the surface of � ], define "small" [eg, less than one forth the gap between the board and the interposer?].

The third "Acceptable [Evidence of a needle dispense pattern � ], be more specific about the "minimum" size of the pattern. We don�t get concerned about the pattern. We get concerned about gaps, thin areas, and voids when too large a pattern or too high speed needle movement are used.

The forth "Acceptable" [The under-fill if by design should be � ], here the notation "if by design" is unclear. Does "according to product documentation" get to the point? [As in the third "Satisfactory" above]. Continuing, change "reach a height of no less than 20% of the thickness of the package in the center of any side" to "reach a height of no less than 20% of the thickness of the package in the center of each side and uniformly taper to cover at least the edge of the package at the corner".

The first "Unacceptable" [There is no evidence of � ], delete "there is no".

The second "Unacceptable", second sentence [No evidence of colour change is � ], change "No evidence of color change is unacceptable." to "Unacceptable if the color change indicates anything less than a complete cure."

The third "Unacceptable" [The under-fill is not visible � ], change "along the complete side of any device" to "along the entire side of the package".

The fifth "Unacceptable" [There should be no � ], delete "there should be no".

The sixth "Unacceptable" [After curing there should be no � ], delete "there should be no".

The seventh "Unacceptable [There should be no � ], delete "there should be no".

On "CSAM to asses the under-filling/curing process" criteria �

The "Acceptable", second sentence [Under-fill is complete and there is � ], be more specific about "no evidence of voids around solder joints" by specifying how close [eg, diameter of a solder ball, touching, diameter of a void, etc.] voids be to solder connections.

The "Unacceptable" [After curing there should be no � ], delete "there should be no".

reply »

Flip Chip/uBGA Underfill Visual Standards

The aim of this document is to produce some draft text and then select photographs to make up a draft visual standard for inspection. There is also a need to produce some reference material for C Scan images.

The following images provide a guide to the inspection requirement after the underfilling operation. They may also be applied after curing of the underfill.

CLASSIFICATION OF QUALITY STANDARDS

During the assessment the following classifications, which are pictorially illustrated, should apply:

SATISFACTORY

This is a satisfactory condition which should be achieved and used as the standard for manufacture.

ACCEPTABLE

This condition represents the maximum acceptable departure from the "Satisfactory" condition. Examples within this limit of deviation will not require reworking. Individual clarification accompanies each example illustration. Consideration should be given to modification to the process materials or conditions.

UNACCEPTABLE

This applies to an unacceptable condition which should not be reworked without the cause of the fault being established. Rework may be possible after assessment of the fault and corrective action taken on the process. Flip Chip/uBGA Underfill Visual Standards

Satisfactory The underfill should be visible at the edge of the device and extend completely around the perimeter. Agreed

Satisfactory The underfill should be visible on all four sides, there may be evidence of more material on one or two sides to aid complete filling of device base. Agreed, you generally have more material at the feed in side(s), this can be minimised by multiple passes and cotrol of the under board heating or use of a seal pass around the other side of the component only after flow out has been completed

Satisfactory The underfill if by design should be visible on the side of the device and reach a minimum height of 25% of the thickness of the package. Disagreed, the underfill, depending on the material selection and it's attraction to the device , silicon / whatever will flow up the side of the device, and occasionally over it. The danger of it flowing over is the posibility of it creating a downward force on the orner / edge of the die and breaking it while under a cycle load. Typically you can see good fill's reachine upto 75% of the edge of the die.

Satisfactory The base of the device should be completely filled with no evidence of voids in the underfill or solder joints being visible around the perimeter of the device. Agreed

Satisfactory Curing of the underfill material should be compared with the reference colour standards. Evidence of colour change illustrates a correctly cured sample. Agreed (to a point), however this will only ensure the outer material has cured the centre of the die, depending on the mass can take longer, you would need to stipulate that it is the same on all 4 sides with evidence of flow out.

Acceptable Flow out of underfill from the device is acceptable provided that it does not completely cover any solder joints. Flow out on to test points, through holes is unacceptable and needs to be reworked and the process modified. Disagreed you require flow out from the device, failure to get this mean's that you have not underfilled the device. In term's of flow's onto components, this can only be achied if the design of the PCB is correct and you don't have thing's like chip's acting like a dam focusing the material. Via's should , I would have thought, require to be tented arounnd such devices.

Acceptable Small bubbles in the surface of the underfill fillet is acceptable provided that no solder joint is visible.

Acceptable Evidence of a needle dispense pattern around the device is acceptable, the width of the pattern should be kept to a minimum. Disagreed, you will be able to see the leading edge, i.e. the dispence edge but if you have the underside hating too high you'll cure the material before flow so you would see evidence of the needle.

Acceptable The underfill if by design should be visible on the side of the device and reach a height of no less than 20% of the thickness of the package in the centre of any side. Not sure about this, you will get flow up the device and it's as much a function of the underfill material as anything else.

Acceptable Underfill fillets which are not visible on the side of the device are not rejectable, however modifications to the dispense process should be made Not sure, you require to see flow out and this will form a fillit, if only with the underside / edge of the component.

Unacceptable There is no evidence of underfill on the top side of the device Disagreed, you don't want underfill on top of a component.

Unacceptable Curing of the underfill should be compared with the reference colour standards. No evidence of colour change is unacceptable. Agreed

Unacceptable The underfill is not visible along the complete side of any device Agreed

Unacceptable Solder joints are visible without underfill on the side of the device Agreed

Unacceptable There should be no evidence of underfill flowing out through holes under the device. Agreed, but this is a design related falut and unless you have tented via's or small enough you may get the fault. It would mean that potentially your getting void's around these area's or inducing stressing on the via.

Unacceptable After curing there should be no visual evidence of cracks between the device surface, underfill or board. Agreed

Unacceptable There should be no evidence of damage to the board or device caused by contact with the dispensing needle or height sensor. Agreed, if your machine is doing this then it's not set right anyway. Some system's use plastic EFD needles which will minimised this anyway.

When using CSAM to asses the underfilling/curing process the following criteria may apply:

Satisfactory Underfill is complete and there is no evidence of voiding between the device and the board. Agreed

Acceptable Underfill is complete and there is evidence of small voids of less than 50% of the solder bump diameter. There is no evidence of voids around solder joints. Agreed

Unacceptable After curing there should be no evidence of cracks between the device surface, underfill or board. Agreed

Unacceptable Underfill is incomplete and there is evidence of voids around or between the solder joints Agreed

Unacceptable Underfill is complete but there is evidence of voids which are larger than 50% of the solder bumpAgreed

Unacceptable Underfill filler shows evidence of separation or uneven flow and will expansion matching and affect the long-term reliability. Agreed

John

reply »