This experiment deals with light propagation through a disordered material. Typically, light going through such a material will get scattered in many different directions. Depending on the thickness of the material, very little, if any, of the light will make it through. But ah, there can be another way to squeeze that light through under the right condition.
A typical light wave is a regular rippling of electromagnetic fields traveling through space, much like evenly spaced ripples lapping at a straight lakeshore. When such waves of light hit a disordered material, different parts of the wave reflect off the myriad surfaces in material in a random way; most of the light scatters from the surface and only a little of the light "diffuses" through. Less and less light gets through as the material is made thicker.
However, it should be possible to wedge light through the disordered solid. The idea is to shape the waves of light, pulling some parts of the wave forward and pushing others back so that when the different parts of the wave scatter in the material they reinforce each other in a process called constructive interference and bounce through an "open channel" (see figure). In the 1980s, theorists used a mathematical scheme called random matrix theory to show that, in principle, there should always be an open channel through a disordered material: Increasing the thickness reduces the number of these open channels, but some always remain. "However thick a material is, it should be possible to create a wave that can be transmitted," says Allard Mosk of the University of Twente in the Netherlands.
Mosk and Ivo Vellekoop, both at the University of Twente, did just that!
The paper is to appear in an "upcoming" PRL, but you can read the preprint here:
I'll include the proper citation here when it appears in PRL.
Edit. There's a very nice Viewpoint article written by John Pendry review this paper. In fact, at this link, you can also get the paper for free. That's nice!