Getting the Lead Out

Dr. Dobb's Journal April 2000

Sometimes you don't need the Internet to be reminded how small and interconnected the world really is. Nor do you need the likes of James Burke, host of the long-running PBS series "Connections," to explain how the headbone is connected to the footbone. For instance, few places have as little in common as southern Missouri's Ozark Mountains and northern California's Silicon Valley. On one hand, you have pristine streams, miles and miles of oak forests, and quaint villages. On the other, you have clogged freeways, outrageous real-estate prices, and cell-phone hell. Nevertheless, the two dissimilar environs are as connected as, well, "Micro" is to "soft." What Silicon Valley needs, the Ozark Mountains have -- and therein lies the rub.

Fundamental to the electronic systems designed and built in Silicon Valley is lead-based solder, which connects components to printed circuit boards and each other. Because it is cheap and easy to work with, lead-based solder is just dandy for connecting everything from water pipes to ICs. However, all lead derived from ore bodies has natural contamination of alpha emitters -- U-238, TH-232, and the like. This isn't a problem for water pipes, but having high-energy alpha particles close to devices utilizing high-density silicon-based memory leads to "soft errors" -- transient, nonreproducible errors in memory, most notably in DRAM and SRAM memory chips.

As Dr. Glenn Lykken, professor of Physics at the University of North Dakota, graciously points out, there's nothing new about this alpha particle lead thing -- engineers have known about it for at least 20 years (see IEEE Transactions on Electron Devices, January 1979). And, according to Don Weeks of Spectrum Sciences (http://www.alphacounting.com/), this soft-error phenomenon has also been associated with ceramic and glass IC packages. Alpha emitter contaminants have also been discovered in other IC materials -- sputtering targets, molding compounds, gold-bonding wire, and the like. However, thanks to materials purification research, most of these materials are now alpha free. But newer IC packages, C-4, Flip Chip, and BGA (all of which use lead/tin solder as interconnects), have led to a resurgence of the alpha particle problem, again creating reliability problems for IC manufacturers.

So the question isn't whether the alpha-lead problem exists, but what can we do about it. In general, there seem to be three solutions:

Refine the lead to separate offending isotopes -- an expensive process that requires laser separation (Pure Technologies, http://www.puretechnologies.com/, does this). Dr. David Robertson and researchers at the University of Missouri-Rolla, in collaboration with The Doe Run Company, have also developed a smelting method that produces "clean" Low-Alpha Lead (LAL).
Salvage "antiquity lead" (in which alpha emitters have decayed) from antique stained-glass windows, old buildings, pre-WW II ships, or similar sources. Don Weeks reports one European company has recovered more than 200 tons of lead originally smelted more than 2000 years ago. Antiquity lead recovery projects have reportedly been undertaken by companies such as Intel and IBM, too, but there's simply not enough old lead; it's often too hard to get; and, at $175 a pound, it's expensive. Again, Dr. Robertson and crew have developed a method whereby antiquity lead can be cleaned of impurities rendering it suitable for use in high-density semiconductor memory devices.
Finally, locate and mine naturally occurring LAL or No-Alpha Lead (NAL), which, as their names suggest, emit lower amounts of radiation than that commonly used in solder. Alas, this approach isn't without problems, as even LAL can be contaminated with Pb-210 during the smelting process. Cleaning up this process is the focus of Dr. Lykken's research.

As it turns out, the Mark Twain National Forest, located in the heart of Missouri's Ozark Mountains, is one place that has naturally occurring LAL (Alaska is another). Even though demand for it is low, LAL still sells for $100.00 per pound, compared to 24 cents per pound for normal lead. Clearly, lead miners see gold, not lead, in meeting Silicon Valley's needs. But there is -- or rather was -- a catch (isn't there always?).

Last year, U.S. Interior Secretary Bruce Babbitt was on the verge of declaring parts of the Mark Twain National Forest off limits to lead mining. This didn't sit well with lead-mining companies that stood to profit -- and that had ponied up heavily to congressional campaign coffers. Riding to the miners' rescue was U.S. Senator Christopher Bond (R-Mo.), a recipient of the lead-mining industry's generosity, who quietly attached an amendment to last year's Interior Department appropriations bill that prevented the Department from halting lead prospecting and mining in the National Forest.

In exchange for pieces of silver, Bond offered up parts of a National Forest for sacrifice. It's ironic that, unless more payola from interested parties starts pouring into Congress, lead solder will likely be banned from electronic manufacturing within the next few years anyway. Instead of bending to personal greed and special interests, Congress should be promoting research into environmentally safe and economical alternatives to lead solder, an already known hazard. But in Washington, where votes are connected to money rather than integrity, common sense, and the common good, logic seldom makes a difference.

Jonathan Erickson
editor-in-chief
jerickson@ddj.com