Inspecting the future

Probably the most imminent development is the drive to remove lead from solder and replace it with other metals whose toxicity often exceeds that of lead. (The fact that the electronics industry contributes only about half of one percent of the environmental lead doesn't seem to matter. The worst culprit is lead-acid batteries in cars.) Manufacturers of x-ray inspection equipment have assured their customers that x-ray systems, with but a few system adjustments, can "see" lead-free solder as readily as the leaded variety.

Unfortunately, solder-joint inspection includes much more than just seeing the joints. Lead-free solder does not flow as readily as the leaded kind, and it generally presents a considerably higher melting point. As a result, the joint shape may not conform to the conventional definition of a "good joint" and may prove less consistent from process to process and from product to product. If an inspection system makes pass/fail determinations using the same criteria and tolerances as with leaded solder, it may generate a lot of false failures.

On the other hand, relaxing pass/fail criteria to accommodate circuit variations will likely miss many failures. Note that because the circuit must function much as it did before, manufacturers might succumb to the temptation to migrate away from inspection and back toward conventional test. Yet, lack of access, increasing speed, logic complexity, and other issues will continue to make any movement in that direction exceedingly difficult.

One challenge to the electronics industry has always been to create materials that can withstand hostile environments—in aircraft, spacecraft, and automobiles, for example. A leading candidate has always been silicon carbide (SiC) because it is much more resistant to temperature and other extremes than conventional silicon. A system built with SiC would not need elaborate cooling systems and other infrastructure to prevent circuit failure in the field, so the system could be smaller, lighter, and less expensive to produce.

Unfortunately, the robust nature of SiC has created its own set of problems. Since SiC doesn't melt at high temperatures, the normal process that manufacturers use to produce high-quality wafers won't work. The traditional technique for producing SiC—condensing the material from supersaturated vapor—cannot produce circuits of adequate quality for complex devices like microprocessors and memory chips.

Researchers led by Daisuke Nakamura at Toyota Central R&D Laboratories in Aichi, Japan, have discovered that if you build SiC wafers by growing the crystals in several stages, you can minimize defects (Ref 1). The researchers have built perfect wafers up to 3 in. in diameter—no competition for today's 12-in. silicon wafers, but a promising beginning. They expect practical applications for SiC in this form in about six years.

The impact on inspection of such a development is not clear. Again, electrically these devices should work the way today's equivalent ones do. Whether their physical appearance resembles current technology—and whether much stricter (or much looser) tolerances will challenge the capabilities of current inspection technology—remains to be seen.

Scientists at Ohio State University recently discovered a special type of the germanium/selenium semiconductor glass used in DVDs and information storage systems that softens under a low-power laser and then re-hardens to its original condition when the laser is removed (Ref 2). Ratnasingham Sooryakumar and former doctoral student Jared Gump discovered the odd behavior by accident, as they attempted to study the properties of the glass to determine why a combination of 20% germanium (hard) and 80% selenium (soft) represents a "magic formula"—the optimum mix for producing the glass.

To determine the material's hardness, they measured the speed of sound waves through it using a low-power red laser. When they achieved a different result every day, they thought they had a defective batch of glass.

It took them a while to realize that the glass was fine—just very sensitive to light. A laser power of a mere 6 mW softened the material by 50%. Yet, the glass always returned to its original condition. Even the latticework of atoms appeared unchanged.

Sooryakumar suggested that these special glasses might be used in rewriteable computer memory and other electronics applications. If so, inspection systems that use lasers to determine a board's position or to perform post-reflow inspection on solder joints might get inconsistent or erroneous results. It may become necessary to be much more careful about where the laser touches the board during inspection, such as by implementing an algorithm that shines the laser only at the joints and turns it off between joints. Such a technique would be clumsier and would consume more power than current systems do. This discovery bears watching.

Printer manufacturer Seiko Epson has come up with a new technique for imprinting circuits on boards that will dramatically reduce their size, weight, and (most importantly) their cost (Ref 3). Researchers at the company propose drawing circuit lines using tiny droplets of metallic ink as small as 10 nm in diameter using a technique similar to that used in today's inkjet printers. Multilayer circuits would include both conductive and insulating inks. Conductive inks can be made of silver, aluminum, nickel, or magnesium. The insulating layers are made from an organic compound that the company will not disclose.

Seiko Epson researchers produced a 20-layer board using a variant of ink-jet technology. The tiny board was only 20 mm2 and 200 microns thick. Courtesy of Seiko Epson.

Inspecting such smaller circuit elements would require higher-resolution imaging and ever-faster image processing. The cost advantages of the process would put pressure on back-end costs from test and inspection, squeezing the industry's ability to comply.

Each of these innovations may still be on the horizon. Their impact on inspection and on the mix between inspection and test may range from critical to insignificant. The influences may turn out to be technical or purely economic. Nevertheless, it is better to be aware of these technologies now and follow their progress than to wait until products begin arriving, when it is too late to react.