You are presented with a sealed crate of bombs, some of which are coupled to detonators and some of which are not. The detonators are so sensitive that they will be triggered if struck by the smallest quantum packet of light conceivable – a single photon. You must figure out, without explosions, which bombs have detonators and which do not.
To put it another way: even simply peeking at an object requires that it be illuminated with light, which will then be absorbed or reflected. So can something ever be 'seen' without its absorbing photons of light – indeed, without photons ever striking it? According to the physicists in the USA and Austria who have come up with a new, improved solution to this dastardly challenge of probing without interacting, the answer is 'yes'.
The 'superbomb' conundrum above was posed in 1993, by physicists A. Elitzur and L. Vaidman, to illustrate the problem of 'interaction-free' measurements. It is a restatement of the old quantum-theory chestnut: the observer and the observed are inextricably linked.
Elitzur and Vaidman proposed that, far from foiling attempts at interaction-free measurement, quantum theory makes them possible. Conventional thinking says that, if you want to inspect the bombs, there's no option but to bounce at least one photon off them. Yet according to quantum theory, because light is both wave and particle, a photon can have a ghostly influence on itself. A beam of light that is split in two and then recombined by a system of mirrors will generate interference effects – even if only one photon passes through the apparatus at a time.
Elitzur and Vaidman showed that, because of this 'self-interference', a primed superbomb could be detected with a one in two chance even without the photon actually striking it. This is all very well as far as it goes – but the option of being blown up in half of the measurements is none too attractive.
Now Paul Kwiat from the Los Alamos National Laboratory in New Mexico, and colleagues, have improved the odds from one in two to almost one in four. As they explain in
This trick is named after Zeno, the Greek philosopher of the fourth century BC renowned for his relish of paradoxes. His take on the 'measurement problem' has entered folklore as the adage that 'a watched pot never boils'. In the quantum world, this can acquire some truth: in some situations, repeated measurements made on a quantum system can prevent it from changing its state. This is the quantum Zeno effect. It operates even if the interaction between the probe and the system is extremely weak. The interaction can then be inferred from the invariance of the system.
The effect holds even in the case of vanishingly small interaction strengths, if the measurements are made often enough. Now, Kwiat and colleagues report that, with just six such measurements, a single photon of laser light could detect a particular quantum system two times out of three without being absorbed.
By combining Elitzur and Vaidman's beam-splitting arrangement with an arrangement of polarizing filters to bring about the quantum Zeno effect, the researchers figured that they could, in principle, make interaction-free measurements as reliably as they liked.
In practice, the limited number of repeat measurements in their set-up gave them a hypothetical probability of interaction-free detection of 85 per cent. Experimentally, the efficiency was a little less – about 74 per cent, or three out of four – because of losses in the light signal within the apparatus. But the researchers claim that increasing the number of sequential measurements to 100 would increase the practical efficiency to more than 93 per cent. So there is hope yet for the quantum detective.