The Dark Side of Ghost Imaging

Displays of candy corn and costumes may soon be replacing sunscreen and beach towels, but this post isn’t meant to detract from what’s left of the summer. Ghost imaging is a technique for imaging something that you can’t see directly. It does seem a bit spooky—imagine getting detailed images of the ground from a satellite-based optical system even when clouds or smoke obscure the line-of-sight. However, ghost imaging isn’t a supernatural feat. It’s just another strange and mind-bending application of quantum mechanics.

First demonstrated in 1995, ghost imaging has sparked a lot of research, lively discussions and disagreements, and intriguing possible applications in encryption and military intelligence gathering. Research published earlier this month in the American Physical Society journal Physical Review A adds to the discussion by considering the differences between ghost imaging systems based on two different kinds of “light” sources. The work provides insight on how to optimize the image quality for each source.

A cartoon highlighting the “dark ghost” image produced by fermions and the “bright ghost” image produced by bosons.
Image Credit: Hongchao Liu

When you take a photograph, your camera detects light reflecting off of the objects in view. The intensity and color are stored digitally as pixels, so that the image can be recreated. You can also create an image by illuminating an object from one side and putting a detector on the other side. This results in a shadow-like silhouette. In both of these cases, the quality of the image depends on the number of pixels in the detector. The more you know about the intensity of light at each particular point on the detector, the better the image. This isn’t the case with ghost imaging.

In ghost imaging you need one light source and two (or more) detectors. You split the light beam in two and only send half of it toward your object. There is a detector on the far side of the object, but it only has one pixel. That means it can tell you something about intensity, but nothing about location so it can’t construct an image. The second half of the beam is a reference beam, sent directly to a detector with lots of pixels. This gives you information on the structure of the beam, but tells you nothing about the object.

The key to creating a ghost image is that the two beams are correlated via quantum mechanics (remember, they start off as the same beam). It turns out that if you know the intensity of the first beam that passes through the object, the distribution of the second beam in space, and the mathematical relationship that describes how the beams are correlated, you can combine that information and get a pretty good image.

The new research in Physical Review A was carried out by Hongchao Liu from the University of Birmingham in the UK. The project was inspired by a talk he attended on electron quantum optics given by Jean-Marc Berroir in 2014. The field of quantum optics explores how photons interact with matter and gave birth to the laser, among other technologies. The new field of electron quantum optics is like quantum optics, but the photons are replaced by electrons. Electrons have mass, electric charge, and other properties that are different than photons, but it turns out that they can be used to image and study matter in some similar ways.

The talk inspired Liu to consider the correlation differences between photons and electrons, or more generally, between bosons (e.g. photons) and fermions (e.g. electrons and protons), in the context of ghost imaging. No one has been able to create a ghost image with fermions yet, but it should be possible. Although the “light” source and optical elements would be different, the overall structure would be the same. The resulting image would also be different. Photons produce a bright image on a dark background. Electrons would produce a dark image on a bright background.

Using mathematical analysis and computer simulations, Liu explored the differences between ghost imaging with bosons and fermions. In particular, he compared their correlation functions, the physics that describes how two beams are linked. As a result of this process he discovered some important behavioral differences between particles in the two types of beams. Based on these differences, he was able to provide some recommendations on optimizing the data processing to get the best quality images. It turns out that this is much more straightforward for bosons than it is for fermions.

Liu is eagerly awaiting the first dark ghost image. About six weeks ago a group of researchers from Australian National University released a preprint paper on ghost imaging experiments with helium-4. Although not yet published in a peer-reviewed journal, this paper describes using helium-4 to create a ghost image, the first one ever formed using particles with mass. Like photons, helium-4 is a boson. However, Liu points out, helium-3 is a fermion. It probably won’t be long before ghost imaging is done with fermions, he says. When the time comes, his work should help experimentalists make the most of the information they capture.

If you haven’t been inspired by the Halloween costumes already in stores, why not grab a friend and make your own this year? I guarantee you’ll be the only ghosts inspired by electron quantum optics at the party!

Kendra Redmond

You may also read these articles