Leibson's Law

I just picked up the June, 2008 issue of Asimov’s Science Fiction magazine (a recently acquired guilty pleasure) and read a Robert Silverberg article titled “The Death of Gallium.” This isn’t a science fiction story. It’s a fact article describing some predictions of Dr. Armin Reller, a materials chemist at the University of Augsburg in Germany. Although Silverberg’s article isn’t really targeted at the electronics industry, the outcomes—should they come to pass—are pretty dire for our little piece of the world.

Reller’s thesis is that several rare elements used in the manufacture of many useful products are about to go extinct, as in we don’t have any more of the stuff. A little Googling shows I’m late to the party for this disaster scenario. It seems the story got a lot of coverage by the “big” publications, including the Wall Street Journal and the UK’s New Scientist, back in the middle of 2007. Nevertheless, it’s news to me and worthy of a blog entry.

Here are the New Scientist’s predictions for the remaining time before we use up known reserves of these key elements:

I don’t know about you, but to me this looks like a doomsday chart for the electronics industry as we know it. Here we are worrying about the end of oil in 50 years, and this list makes it look as though the electronics industry won’t last as long as the oil industry. Although all sorts of bad things happen when we run out of any of the items on the list (after all, we need platinum for making those memristors I just wrote about), let’s just deal with hafnium.

The electronics industry has merrily jaunted along for more than 100 years without needing much hafnium, until now. Suddenly, at the 45nm lithography node, we need it to tame static leakage in CMOS ICs. That’s going to be good for another decade or so and then no more hafnium, according to Dr. Reller. Gallium and indium, used in LEDs, are also in short supply. The metals-centric electronics industry is in pretty serious trouble here.

Now elemental extinction isn’t like biological extinction. Once the apatosaurus and the wooly mammoth become extinct, they can’t return (barring the kind of technology in Michael Crichton’s Jurassic Park). However, more of these elemental metals exist on the Earth and even more can be had elsewhere in the Solar System, we merely need to find it and extract or refine it. That doesn’t mean we will. It may not make economic sense.

Who knows what electronics will look like in a decade? Perhaps all circuitry will be carbon based as we transition from silicon semiconductors to etched graphene and carbon nanotubes. After all, we’ve got so much carbon we’re figuring out ways to sequester it. We might as well do something useful with all that extra carbon.

Couldn’t happen, you say? Too far fetched? We’ve made such radical transitions before. Electronics was born in 1906 with Lee de Forest’s invention of the Audion vacuum-tube triode. Tubes held sway for about 50 years. The germanium transistor was invented at Bell Labs in 1947 but didn’t take over most of the jobs done by tubes for another 10 years (that’s Leibson’s Law in action). Silicon transistors took about another 10 years (that’s Leibson’s Law in action too). ICs also took about 10 years, from their invention in 1959, to capture the bulk of the electronics market (Leibson’s Law again). Ditto microprocessors, which showed up in 1971 (another example of Leibson’s Law). So we’ve made many big shifts in electronics over the years. We will no doubt make more.

Elemental extinction may drive the next big shift in electronics. Something certainly will.