Turbid materials, such as opal glass, white paint, or skin, strongly scatter light, making it impossible to see through them.
However, it is well known that light can still be focused through even the most turbid objects, if only you know exactly how that particular object scatters light. In a typical experiment, a megapixel camera is used to record how the light gets scattered. But how much light do you need to accurately record how the light was scattered? Intuitively one would expect one photon per pixel to be the lower limit. Surprisingly, this answer is far from correct. Researchers from the University of Twente and the California Institute of Technology demonstrate that there is no such lower limit, and experimentally demonstrated light focusing with as little as 0.004 harvested scattered photons per camera pixel.
Light has many applications in biology and medicine, not only in microscopy, but also in measuring perfusion, detecting tumors, or activating targeted drugs. Unfortunately, scattering prevents light from penetrating deep into the tissue. Therefore, physicists have set on a bold quest to mitigate the effects of scattering, in order to see deep inside scattering tissue. One of the key technologies to make this possible is optical phase conjugation. In optical phase conjugation, one first records how the light gets scattered by the material, and then generates a time-reversed copy of the scattered wave. This time-reversed copy will trace its way back through the material exactly to focus back at the original source.
Focusing light back at its original source is not very useful by itself, unless this source is already embedded deep inside the scattering medium. Such an embedded source is called a ‘guide star’, and many approaches for generating guide stars inside tissue have been demonstrated in proof-of-concept experiments. Unfortunately, these guide stars are currently extremely weak, and the general concern is that they do not produce enough light for practical use. Motivated by this concern, researchers of the University of Twente and the California Institute of Technology set out to establish the fundamental lower limit of the amount of light required for optical phase conjugation.
Imagine that the guide star generates as little as 1000 photons, while the phase conjugation system uses a camera with 200.000 pixels. One may think that in this case the camera can only measure useful information with 1000 of its pixels, at best. Surprisingly, when the weak signal is interfered with a reference beam, the information from each single scattered photon is recorded by all camera pixels simultaneously. The intuitive lower limit of one scattered photon per camera pixel is not only a very pessimistic estimate: it turns out that there fundamentally is no such a lower limit at all. In an experiment, the researchers could go as low as 0.004 photons per pixel, using 1000 detected photons altogether to focus light through a slab of opal glass.
This result is particularly good news for practical applications of optical phase conjugation such as deep tissue imaging and high-speed imaging, where the low number of photons is the ultimate limiting factor.
The paper ‘Optical phase conjugation with less than a photon per degree of freedom’, by Mooseok Jang, Changhuei Yang, and Ivo M. Vellekoop has been published in Physical Review Letters https://doi.org/10.1103/PhysRevLett.118.093902
Particle or wave. A crucial step in digital optical phase conjugation is to accurately measure the wavefront of a scattered or distorted beam. When the light intensity is high, the wavefront can be detected with full fidelity (center). At extremely low light intensities, one may think that the few arriving signal photons only permit measuring the wavefront at a set of scattered points (left), a situation that would not allow for effective phase conjugation. In this paper, however, we demonstrate phase conjugation with as little as 1000 signal photons, scattered over 220000 camera pixels. Our results can only be explained by a surprising feature of interferometric detection: each signal photon contributes some information to all individual camera pixels simultaneously, adding a small statistical bias towards the correct measurement (right). Our message is good news for applications where the photon budget is low, such as ultra-fast phase conjugation and deep-tissue imaging with time-reversed ultrasound-encoded light (TRUE).