Common sense says you can't keep an atom's nucleus from decaying simply by looking at it. Quantum mechanics says you can. Now two Israeli physicists have come up with a way in which watching a nucleus might make it decay faster.
The decay-preventing process, known as the quantum Zeno effect, has fascinated physicists for 25 years. Imagine an alpha particle, two protons and two neutrons, lodged inside a much larger, radioactive nucleus. It can escape via a strange quantum tunneling, in which the particle exists in a "superposition" of states that puts it, in effect, both inside and outside the nucleus at once. But if someone measures the particle, by, say, bouncing a photon off it, the superposition collapses and the particle must instantly commit itself to being either inside or outside the nucleus. By repeatedly prodding the particle with photons, the theory goes, scientists can destroy superposition before it starts and dash a particle's chances of escaping the nucleus. In short, the watched pot never boils.
Now physicists Gershon Kurizki and Abraham Kofman of the Weizmann Institute of Science in Rehovot, Israel, argue in the 31 May issue of Nature that the same procedure can sometimes have the opposite effect: Under certain conditions, the watched pot always boils. Repeated measurements bump a particle and give it a broader range of energies than it would otherwise occupy, they say--including energy states that correspond to the particle's existing outside the nucleus. "If you [repeat measurements] sufficiently fast, you would see an increase of the decay rate," Kurizki says.
Although nobody has yet seen the anti-Zeno effect in action, Kurizki believes experiments will verify it within a few years. In fact, he thinks the anti-Zeno effect ought to be much more common than the Zeno effect. If so, that could be bad news for scientists trying to develop quantum computers, says Los Alamos physicist Peter Milonni, who points out that some physicists have proposed using the Zeno effect to keep quantum bits from losing information.