|
X-Ray Inspection of IC Packages and PWBs
|
Seeing Voids
To detect air bubbles in epoxy-type materials, the system must be able to image x-rays down to around 10 kV or less, requiring a low-energy imaging system, such as a beryllium image intensifier or a flat scintillator panel. These devices are more expensive, but are necessary if soft materials are to be inspected.
In general, images are produced by different levels of x-ray absorption in different parts of a sample. To detect these levels, the x-rays have to pass through the sample and into the x-ray detector.
If a sample contains some very subtle differences within its structure, but the whole structure is shielded in a thick copper casing, then it is most likely that any small signal differences will become lost in the copper.
A typical example of this may be found in the structure of a conventional copper leadframe package. In this case, the silicon die will be invisible in the package, since the absorption of x-rays in the silicon is negligible compared to that in the copper.
The basic laws of physics are immutable, and x-ray system design is no exception.
Online or Inline Systems?
Although the video image from an offline system runs at 25-30 frames per second, to capture a good quality image for analysis takes one or two seconds as the series of images is taken and averaged. Inline systems, however, must capture images for analysis in about 40ms-impossible without a loss of resolution.
Why does resolution suffer with inline systems? The reason is simple, only about 0.6 percent of the electron beam power is converted to x-ray. The remaining 99.4 percent is dissipated as heat in the focal spot. To avoid target burn, the x-ray output dose from a microfocus tube must be very low.
Therefore, obtaining a good low-noise image in a single video frame requires a high-output dose and large focal spot. Thus, as the speed increases, the resolution decreases.
Applications
One of the most useful applications in electronics for x-ray is the inspection of PWB assemblies.
These assemblies are composed of components and materials of different types and atomic numbers ranging from carbon to molybdenum. Because of this diversity, an x-ray image, particularly one of solder joints on a glass fiber substrate, will contain lots of contrast and resolvable detail.
Where x-rays pass through areas of bare PWBs, their attenuation is low, and the pixels in the imaging system that they strike become very bright. On the other hand, x-rays passing through the solder joints are heavily attenuated, and these areas appear dark.
The outline of the solder joints is, in general, an obvious transition from dark to light and so short circuits and irregular shaped joints are very easy to detect.
The detection of a missing ball under a µBGA package or flip chip will only be possible if the operator knows what should be there. Once this information is at hand, however, solder joint omissions are as easy to detect as shorts.
By increasing the kVs to the level where about 10% of the x-rays are passing through a solder joint, the joint will become semi-transparent, revealing any voids inside as light circles. Performing a simple area calculation will yield the percentage void for the joint.
Software functions are included on most systems to detect and analyze shorts, voids, shape factor and missing balls for BGAs. The most challenging defect to find, however, is a non-wetted or open joint.
|
|
|
Figure 3. X-ray image of copper balls on a substrate
|
Figure 4. A motherboard undergoes inspection for solder-joint defects in the X-Tek CPX160.
|
BGA Wetting
Standard collapsible BGA solder joints of up to 1 mm will form a barrel-shaped joint with the ends of the barrel matching the size of the solderable area of the pad on the PWB or device substrate (Figure 3). By obtaining a reasonably angled view of the balls, any non-wetted balls will show up as having a spherical end sitting on a flat pad.
If the pad diameter is much smaller than the ball diameter, then the wetted ball and the non-wetted ball become similar in shape and are very difficult to differentiate.
Pad Design
Some circuit designers add wetting indicators to the pads to aid x-ray inspection. These are solderable areas of pad stretching outside the ball diameter so that a face-down x-ray view of the joint will show a dark extension of the joint when reflowed correctly. If the wetting indicator is not visible after reflow, then the solder has not flowed onto it, suggesting a poor joint.
Misplaced devices are not usually a problem; during reflow the surface tension of the joints pulls the device into alignment.
Collapse may occur if the temperature profile across the device is uneven, or if the device has some mechanical pressure applied before the joints are solid, causing the device to sit high on one side or corner. A simple check for this condition analyzes the aspect ratio of the corner balls in a fairly steep-angled x-ray view.
A correctly soldered device will have the same aspect ratio on all of the joints. A warped device or PWB will show the differences between center and edge.
Flip-Chip Underfill
The integrity of flip-chip underfill is difficult to check with x-ray, because the silica-filled epoxy is fairly x-ray transparent. Some work has been done on specific element imaging that does produce results, but the procedure requires an understanding of the materials in the underfill and substrate.
Lead-Free Solder
Thus far, the use of lead-free solder has not changed x-ray's capabilities, since the absorption in tin, like lead, is high.
System Types
X-ray systems are produced in a variety of form factors for use in electronics, and in manual, semi-automated or fully automated versions (Figure 4). In the past few years, x-ray inspection has moved from strictly a laboratory tool to a valuable production assistant.
|
Mr. Wright has been sales and marketing manager for X-Tek since March 1993. Earlier, he was an applications engineer in Europe for Nicolet X-Ray. He earned a BSc degree in electrical and electronic engineering from the University of Hertfordshire. [stuartw@xtek.co.uk]
|
|