Safety Last

By Robert X. Cringely, the host of PBS’s NerdTV (THE NEW YORK TIMES, 01/09/06):

OVER the past two weeks Apple Computer and Dell Computer have recalled approximately six million lithium-ion batteries powering their notebook computers. The batteries, all made by Sony, had an annoying problem: they were prone to explode. While we as consumers might wonder why vendors are selling batteries intended to go in pockets and briefcases that have greater energy storage density than dynamite, recent history actually shows that we’ll accept almost any risk for more power.

Lithium-ion is just the latest in a succession of relatively toxic rechargeable battery technologies, each intended to pack more available electrons in smaller and smaller packages, with li-ion besting the field by a resounding five times, explosions be damned. Earlier rechargeable batteries were mainly nickel-cadmium (these can explode, too, if short-circuited) and nickel-metal-hydride (the most benign rechargeable, these are in your Toyota Prius, natch).

In most li-ion batteries, either overcharging or a manufacturing defect in the plastic membrane separating the battery’s anode and cathode can lead to an explosion that pops the battery’s metal can, releasing steam at up to 600 degrees. If this happened to the computer resting on your knees, it would be a very bad day; fortunately, computers get hot to the touch before they blow. Even deep into writing your conclusion to the Great American Novel, you’d know something was about to happen, believe me.

In a worst-case situation, metallic lithium itself can explode in the battery, causing flame as well as heat. Luckily for Sony, these flamers mainly happen to another battery variety called lithium-polymer, not li-ion.

No one builds batteries expecting them to explode in normal use. But just as Ford knew about the exploding Pinto gas tanks, li-ion battery manufacturers know it’s statistically probable that a small percentage of their cells are going to blow. They still sell these batteries because they’ve calculated that the failures will be few. This is done with something called the M.T.B.F., or mean time between failure.

The way M.T.B.F. is measured has little to do with the way the batteries are used in the real world. It is a sham. A random sample of batteries, say 1,000, are put to use in a target device and operated until one or more fail in service for whatever reason. If it takes 100 hours for the first battery to fail, the M.T.B.F. is set at 100,000 hours — that is, 1,000 times 100, an essentially meaningless formula, but a satisfying number for all concerned, as it implies that if your cellphone were to explode it would most likely do so after thousands of hours of use.

It’s unlikely, however, that your battery will last that long. Li-ion batteries start to age practically from the moment they are made, and hardly ever live past their third birthdays. (When they are dying of old age, the batteries don’t explode, they just stop working.) Three years is only 26,280 hours, suggesting that if we really believe in the M.T.B.F. fairy our li-ion batteries will never explode. But of course they do — often enough that Sony is now eating a roughly $200 million loss to replace six million of them.

The limited life expectancy of li-ion batteries was a rude shock to buyers of early Apple iPods, which had non-removable li-ion batteries. Silly people — they thought they were buying a device with longevity. Mobile phones have a life expectancy of about 18 months, making them appear to be immune to catastrophic li-ion battery failure, which they aren’t.

One might think that we’d be working on safer technologies, and we are, up to a point. Safer lithium-ion batteries are available, but computer and mobile phone manufacturers, now duking it out in a market based on talk time and battery life, have decided that we don’t really need them. And judging from the reckless way we use these devices while driving cars, the manufacturers are probably correct about our risk tolerance.

It is possible to replace the lithium cobalt oxide cathode material in li-ion batteries with lithiated metal phosphate cathodes that don’t explode and even have a longer shelf life. But for the moment these safer li-ion batteries seem mainly destined for electric cars and other large-capacity applications, where the safety issues are more critical. (Explosive power is in part a function of battery volume. An errant cellphone might burn your hip, but a li-ion hybrid car compromised in a collision might actually kill someone). The fact is that lithiated metal phosphate batteries hold only about 75 percent as much power and we’ve decided (or the li-ion companies have decided for us) that when it comes to talk time we’d rather risk it.

Of course, there are other emerging energy-generating and storage technologies. Ultra-capacitors built with carbon nanotubes look good, if still a few years from the market. We could use fuel cells, but that could require storing compressed and highly explosive hydrogen in your notebook computer. Direct methanol fuel cells could run your mobile phone on vodka but they’d likely leak in your pocket.

It would be glib to claim that we’ll never be happy with any amount of bandwidth or battery life. But each new advance in energy storage seems to offer us a Faustian choice, and we always take the devil’s side: knowledge (talk- or surfing-time) over safety. Thirty percent more talk time with a one-in-10-million risk of burning to death? No problem.

Still, maybe there are limits to our need for speed. If we wait long enough, we’ll have batteries that store energy directly in the protons and neutrons of nuclear isomers like hafnium-178, delivering to our cellphones the energy density of a plutonium bomb.

That ought to be enough.