Focus: Reflection of a Ghost
Under special optical arrangements, physicists can create an image of an object using light that has never interacted with the object. This “ghost imaging” has been around for more than a decade, but it has yet to find much practical use. That could change now that researchers have generated a ghost image of an opaque object, as described in the April Physical Review A. The technique could lead to improved satellite imaging through clouds. But the debate over how to theoretically describe the effect continues.
Ghost imaging involves two light detectors–one pointed at the object to be imaged and the other at the light source illuminating the object. In one version of the experiment, a weak beam emits only a few photons of light at a time. The beam is split, with part going to the object before hitting a photon-counting detector and part going an equal distance to a digital camera. When the source emits two photons nearly simultaneously, they are likely to be headed for the same spot on the object, but one may go to the camera instead. Any simultaneous detection of photons signals a pixel that the camera should count in the final image. Although researchers have been making the pictures for years, they still don’t agree on whether the explanation is purely quantum-mechanical in every case, or whether classical physics can explain some experiments.
Until now, the images have all come from light passing through stenciled patterns in a mask. But Yanhua Shih of the University of Maryland, Baltimore County, working with Ron Meyers and Keith Deacon at the US Army Research Laboratory in Adelphi, Maryland, wanted to make a ghost image with the small amount of light that bounces or scatters off of an opaque object. Meyers says that with reflected ghost imaging, one might someday put a detector on a satellite and have a second camera take images directly of the Sun. This combination could generate ghost images of the Earth’s surface, even if clouds or smoke partially obscured the satellite’s view.
The team placed their object, a toy soldier, 45 centimeters from the light source, with a photon detector positioned to capture light scattered from the toy. The other beam traveled an equal distance to the camera. The ghostly soldier began appearing after about a thousand coincident pairs of photons were recorded.
The argument among researchers is whether the proximity in space of a pair of photons–or its wavelike analog in terms of electric fields–is purely a quantum mechanical effect. Some researchers believe a classical interpretation is sufficient, in which fluctuations in the source cause photons to be emitted in bunches that all point in roughly the same direction.
But in this experiment, the source is so close that it is not just a point. Classical theory is inadequate here because nothing would prevent differently-directed bunches from separate points within the source from all striking the detectors at nearly the same time, the team argues. These uncorrelated bunches–or background noise–would destroy the ghost image. The ghost’s “reality,” they say, is thanks to a quantum mechanical effect involving photon interference, which forces the photon pair to follow related trajectories along the separate optical paths.
The team contends that this experiment settles the argument over the theoretical basis of ghost imaging, but others disagree. “Shih and co-workers do not show in this new [work] that classical theory fails,” says Morten Bache of the Technical University of Denmark in Lyngby, although he believes the paper is a good contribution to the debate. Suhail Zubairy of Texas A&M University’s campus in Qatar thinks the authors make a good argument, but adds, “I don’t think the last word has been written on this subject.”
Michael Schirber is a freelance science writer in Lyon, France.
A theory paper published one day earlier attempts to clarify the classical and quantum issues in ghost imaging: B. I. Erkmen and J. H. Shapiro, “Unified Theory of Ghost Imaging with Gaussian-State Light,” Phys. Rev. A 77, 043809 (2008)