Interaction-Free Measurement

Abstract

We show that one can ascertain the presence of an object in some sense without interacting with it. One repeatedly, but weakly, tests for the presence of the object, which would inhibit an otherwise coherent evolution of the interrogating photon. The fraction of "interaction-free" measurements can be arbitrarily close to 1. Using single photons in a Michelson interferometer, we have performed a preliminary demonstration of some of these ideas. PACS numbers: 03.65.Bz, 42.50.Dv One of the commonly cited differences between classical and quantum physics is that in the former the disturbance of the system under observation can be made arbitrarily small, while in the latter the measurement process in general disturbs the system. Yet, Renninger used the notion of a "negative-result measurement" to describe the nonobservance of a particular result as a measurement of a quantum system, seemingly without disturbing it [1]. The concept of an "interaction-free" quantum measurement was then considered by Dicke, who analyzed the change of an atom's wave function effected by the nonscattering of a photon from it [2]. Elitzur and Vaidman (EV) extended these ideas, so that the presence of an object modified the interference of a photon, even though the photon and the object need not have interacted [3]. The maximum attainable efficiency in the EV scheme is 50%. We have discovered an improved method where the fraction of interaction-free measurements can be arbitrarily close to 1 [4]. In the new version, which may be viewed as an application of a discrete form of the quantum Zeno effect, one coherently repeats the interrogation of the region that might contain the object. As an intermediate step we have used the single-photon states available from spontaneous parametric down-conversion to experimentally demonstrate the principle of an interaction-free measurement. The initial proposal of EV employs an interferometer aligned so that an incident photon (or any other interfering particle) will with certainty exit via a given output port, the "bright" port. Thus, in the absence of any object within the interferometer, the photon will never be detected in the "dark" output port. The presence of an absorbing (or, more generally, nontransmitting) object in one of the arms changes completely the possible outcomes by destroying the interference. For a beam splitter of reflectivity R (and transmissivity T = 1-R), any incident photon will encounter the object with probability R and be absorbed. There is a probability R that the photon will still exit to the bright port; since this yields no information, the experiment should be repeated (either with the same photon or with a new one). However, there is also a probability RT that the photon will exit to the dark output port. Detecting this photon, one can conclude that an object was certainly in one arm of the interferometer, even though the photon could not have interacted with it. To make the argument more dramatic, EV proposed that the object could be an ultrasensitive bomb, triggered by the absorption of a single photon. The complementarity of a single quantum is essential for the above method: In the absence of the object, it is the wavelike nature of the incident light which allows us to establish, through destructive interference, a condition in which the photon never uses the dark output;