Issue



X-ray Inspection Technology: Finding Hidden Defects


09/01/2006







BY DON NAUGLER, VJ Electronix, Inc.

Not all packaging and interconnect defects can be detected through in-circuit and functional test. X-ray inspection provides a non-destructive means to identify and track process trends.

Ever since Wilhelm Roentgen discovered X-ray radiation in 1896, X-ray technology has evolved into one of the most powerful inspection tools available. It has been said that information from X-ray imaging, even in its early forms, is so intuitive that its viability has never been questioned.

X-rays are a form of electromagnetic energy similar to visible light, but with a shorter wavelength and higher energy. When visible light strikes most objects, the longer wavelengths are either reflected, or absorbed by the material. X-ray radiation properties allow the waves to travel through intermolecular space within solid materials.


Figure 1. X-ray penetration.
Click here to enlarge image

Materials of relatively low density will allow more of the X-rays to pass than will relatively high density materials (Figure 1). This phenomenon is used by doctors to find broken bones, because bones are significantly denser than soft tissue and create more salient shadows on X-ray images. The same principle applies to electrical assembly and package inspection. Dense materials such as metal interconnect traces, wires, etc. are far more dense than organic package materials.

kV and X-ray Wavelength

X-rays are made up of a continuum of wavelengths in the range of 0.1 to 100 Å. The shorter the wavelength of the X-ray radiation, the more readily it will pass through solid materials. X-ray system operators have the ability to control output wavelength by setting the potential, or kV, of the X-ray source; the higher the kV, the shorter the wavelength and the greater its penetrating capability.

To inspect assemblies with dense materials, a system with a high kV source would be required. Most systems designed for inspection of electronic packages and PCBs have X-ray sources capable of operating at 75 to 160 kV.

In addition to controlling the wavelength via the voltage at the source, the total quantity of X-rays is controlled by adjusting the current. The greater the current applied, the more photons of X-ray energy emitted.

Obtaining a good X-ray image requires a balance of optimizing wavelength (or kV) to provide desired contrast between dense and less-dense materials, and current (mA) to control the image brightness.

Often times an operator is tempted to use the kV setting as a brightness knob. Doing so may result in a low contrast image. The kV setting should be adjusted to the minimum required to achieve penetration and the power adjusted to set desired brightness.

Spot Size

Another important parameter is the X-ray source’s spot size. The spot is the area on a part of the source called a target, where the X-rays are generated. The smaller the spot, the better the resolution capability will be, due to a phenomenon called penumbral blur (Figure 2).


Figure 2. Spot size vs. image blur.
Click here to enlarge image

If all the X-rays were generated from a point source there would be no blurring, but because they are generated from a finite spot there will always be some blurring of the image.

The size of the spot is limited by the power required because heat is also generated at the X-ray target. If too much energy is focused on a small spot, the heat will damage the target. Most systems will vary the spot size with X-ray power (the product of kV × mA) to protect the target and maximize the useful life of the source.

Magnification

X-ray systems can provide a high magnification of the sample being inspected. Geometric magnification is defined as the ratio of the distance between the source and detector to the distance between the source and the sample (Figure 3). In addition to geometric magnification, many systems offer optical and/or digital magnification. Both may be useful but will result in loss of image quality.


Figure 3. Geometric magnification.
Click here to enlarge image

Obtaining the best resolution under high magnification requires a high level of geometric magnification. This can be achieved using a large distance between the source and detector. However, a large source-to-detector distance requires a large cabinet and is often not practical. Therefore, the most effective way to achieve high geometric magnification is to place the sample as close to the source as possible.

The trade-off for high magnification is field-of-view; the greater the magnification, the smaller the field-of-view. The magnification required is dependent on the nature of the sought-after defects. Inspection of eutectic bonds in a semiconductor package may require 40× magnification. Wire-bond inspection may exceed 200 to 400×. When using high magnification it is likely that multiple images (or multiple regions of interest) will be required - an important factor when calculating inspection throughput.

Available Inspection Tools

There are a number of tools available to enhance X-ray images and provide additional data. Frame averaging reduces signal noise and provides a crisp video image. Various filters, such as edge filters, are available to enhance or highlight subtle changes in a complex image. X-ray images in raw form are black and white, but may be colorized or rendered into 3-D plots (Figure 4) to emphasize differences.


Figure 4. Standard view vs. pseudo 3-D rendering.
Click here to enlarge image

Systems with advanced image processing capabilities can automatically make measurements. One of the more useful tools is void analysis. Voids are volumetric defects caused by encapsulated air pockets or remains of resins of solder paste. Automated void analysis applies a grayscale discrimination of areas that comprise a bond between two surfaces such as between a substrate and component. The analysis will highlight voids (areas of relatively high grayscale value) and present data about the total bond and void areas, and the percentage of void to bond area.

Ball grid array (BGA) solder joint X-ray inspection is often performed to monitor the production process. Automated BGA analysis provides data on solder joint diameter, area, shape, percent void area, and size of largest void, which may be used to process limits and automatically flag pass/fail conditions.


Figure 5. Void analysis.
Click here to enlarge image

Void analysis (Figure 5) and bond area analysis provide valuable feedback for process control. Excess voiding may indicate surface preparation, thermal, mechanical, or metallurgical issues. It is important to identify bond issues as part of process controls because this type of problem may not be revealed in other testing, but may result in premature failure once the joint is stressed under high electrical current, mechanical, or thermal load.


Figure 6. Wire sweep image.
Click here to enlarge image

Wire-bond inspection may be used for failure analysis or as part of process controls. High magnification may be applied to view the integrity of the wire bond. Wire sweep analysis (Figure 6) provides feedback on the wire bond process as well as the over-mold or encapsulation process.

New Technology

Improvements to the source, primarily related to spot size, have allowed greater resolution. Improvements to X-ray detectors have provided greater sensitivity and sharper images. Most recently, new digital detectors have been developed, providing huge improvements in image contrast.

The most common X-ray detector in use is called an image intensifier, which converts X-ray energy into visible light. The output of the image intensifier is transmitted through a lens into a CCD camera. The camera signal is commonly digitized into an 8-bit grayscale image with 256 shades of gray - not a problem for computers, but human eyes can only discriminate between 30 - 60 gray scales.

Flat panel digital detectors are capable of providing 12-bit or 16-bit images. The image is transferred directly to the system computer - eliminating loss from lens and electrical coupling problems. The resulting image contains 4096 grayscale levels or more.

The benefit of increased data available through use of digital detectors can be leveraged through software image analysis tools, which can manipulate this data to highlight extremely subtle changes. This can be especially important when inspecting samples, such as ceramic packages that have very dense materials in some areas and thin and/or less dense materials in others, by compressing large contrast differences.

DON NAUGLER, general manager, may be contacted at VJ Electronix Inc., 234 Taylor St., Littleton, MA 01460; 978-486-4777; Fax: 978-486-4550; E-mail: dnaugler@vjt.com.