PIONEERING EXPERIMENT
FOCUSING COHERENT Light
10 incident wavefronts per second, so that we needed several minutes to make a high-intensity focus. That speed was unacceptable for applications, especially in dynamic biological systems. Second, we needed to place a detector in the desired focal point. That almost completely defeated the purpose. While we could focus through, we could not yet focus inside an opaque material. Third, because we used feedback from a single point only, we could not easily use our method to make images. Even though our method served well to demonstrate enhanced transmission due to the theoretically predicted open channels [ 4 ], clearly more work was needed to enable further applications. Fortunately, at this point our work was picked up by an energetic, open, and incredibly creative community, with many more amazing results than we can review here. Many beautiful examples have been collected on the website maintained by Sebastian Popoff, www. wavefrontshaping. net as well as in several recent in-depth reviews [ 2,3,5,6,7 ].
A profound development with tremendous impact was the demonstration of optical transmission matrix measurements by the team of Sylvain Gigan [ 8 ]. On a superficial level, it provides a generalization of wavefront shaping to an entire image instead of a focus. More importantly, it links the theoretical concept of the transmission matrix in mesoscopic transport theory directly to a measurable object in optics [ 3 ]. Measuring the transmission matrix enables the use of the powerful methods of linear algebra to analyze, optimize and control the propagation of light in a wide variety of materials and for a wide variety of purposes, reaching from imaging and light delivery to computing, machine learning and quantum optics [ 6 ]. Advances in mesoscopic theory quantified the effects of limited control, so that even partial transmission matrices can be meaningfully compared to theory [ 9 ]. One of the most prominent application-oriented developments is the use of wavefront shaping for imaging through multimode optical fibers [ 10 ]. Because multimode fibers are very thin, they enable minimally invasive imaging inside sensitive organs such as the brain. However, they cannot be used as an endoscope directly since the differential phase delay between the fiber modes effectively scrambles the incident light into a speckle pattern. Wavefront shaping the input field leads to a focused beam at the distal end of the fiber, which may then be scanned to construct an image of a fluorescent or reflective object [ 10 ]. Similarly, scanning microscopy techniques were developed to image through opaque layers, or even inside scattering biological tissue. Most of these techniques rely on first creating a focus inside the material, and then raster scanning the focus to obtain a( fluorescence) image. A wide range of approaches for creating the initial focus were demonstrated: using local fluorescent or nonlinear particles, acousto-optics, as well as statistical matrix methods or even digital twins. A form of correlations( the optical‘ memory effect’) then facilitates scanning the focus [ 2 ].
Figure 4. Everything is a lens! An impression of some of the materials we were able to focus through:( a) Daisy petals,( b) white paint,( c) human teeth,( d) chicken tissue( cooked),( e) eggshells,( f) porous gallium phosphide.
32 www. photoniques. com I Photoniques 140