Coherent perfect absorbers (CPAs) are novel linear optical elements that can absorb arbitrary bodies or aggregates perfectly at discrete frequencies when a precise amount of dissipation is added under specific conditions of coherent monochromatic illumination. This phenomenon arises from the interaction of optical absorption and wave interference, and corresponds to moving a zero of the elastic S-matrix onto the real wavevector axis. It is the time-reversed process of lasing at threshold. The effect is demonstrated in a simple Si slab geometry illuminated in the 500–900nm range. CPAs are absorptive interferometers that may be useful for controlled optical energy transfer. A CPA is an optical system that can absorb radiation perfectly when illuminated by the time-reverse of a lasing mode. This occurs when the incident radiation corresponds to the specific eigenvector of the S-matrix with eigenvalue zero. The CPA zeros are distinct from absorption resonances of the atomic or molecular medium, which do not require specific illumination conditions. The CPA process arises from the interplay of interference and absorption: in the presence of specific amounts of dissipation, there exist interference patterns that trap the incident radiation for an infinite time. Even small rates of single-pass absorption can lead to perfect absorption, and hence media that normally do not absorb radiation well at certain frequencies can be made to do so, albeit within narrow frequency bands. The simplest possible CPA is a single port reflector, but important properties of the CPA are only revealed when there are multiple input ports and hence non-trivial eigenvectors. A two-port case is studied to illustrate the concept fully. The CPA zeros are determined by the analytic properties of the S-matrix. For a uniform dielectric slab, the CPA zeros are found by solving the equation $ e^{i n k a} = \pm \frac{n-1}{n+1} $. The solutions for ka = 664.7 are shown in Fig. 2. The CPA zeros are found by varying the real and imaginary parts of n, leading to zeros at different k-points. In an indirect bandgap semiconductor such as Si, the CPA zeros can be found by varying k so as to pass very close to several CPA zeros. The CPA zeros are a-dependent and hence tunable within a given material. The S-matrix eigenvalue intensities are bounded by a tight lower bound, which goes to zero at the n values given in (7) and locates the interesting operating regions. For Si, the optimal wavelength occurs at around 750 nm for a = 10 μm, and around 1000 nm for a = 150 μm. The exact value depends on n(k), which in turn depends on the doping. The CPA functions as a compact absorbing interferometer. Unlike an ordinary interferometer, it does not deflect the input beams into another output channel, but causes it to be absorbed entirely within the materialCoherent perfect absorbers (CPAs) are novel linear optical elements that can absorb arbitrary bodies or aggregates perfectly at discrete frequencies when a precise amount of dissipation is added under specific conditions of coherent monochromatic illumination. This phenomenon arises from the interaction of optical absorption and wave interference, and corresponds to moving a zero of the elastic S-matrix onto the real wavevector axis. It is the time-reversed process of lasing at threshold. The effect is demonstrated in a simple Si slab geometry illuminated in the 500–900nm range. CPAs are absorptive interferometers that may be useful for controlled optical energy transfer. A CPA is an optical system that can absorb radiation perfectly when illuminated by the time-reverse of a lasing mode. This occurs when the incident radiation corresponds to the specific eigenvector of the S-matrix with eigenvalue zero. The CPA zeros are distinct from absorption resonances of the atomic or molecular medium, which do not require specific illumination conditions. The CPA process arises from the interplay of interference and absorption: in the presence of specific amounts of dissipation, there exist interference patterns that trap the incident radiation for an infinite time. Even small rates of single-pass absorption can lead to perfect absorption, and hence media that normally do not absorb radiation well at certain frequencies can be made to do so, albeit within narrow frequency bands. The simplest possible CPA is a single port reflector, but important properties of the CPA are only revealed when there are multiple input ports and hence non-trivial eigenvectors. A two-port case is studied to illustrate the concept fully. The CPA zeros are determined by the analytic properties of the S-matrix. For a uniform dielectric slab, the CPA zeros are found by solving the equation $ e^{i n k a} = \pm \frac{n-1}{n+1} $. The solutions for ka = 664.7 are shown in Fig. 2. The CPA zeros are found by varying the real and imaginary parts of n, leading to zeros at different k-points. In an indirect bandgap semiconductor such as Si, the CPA zeros can be found by varying k so as to pass very close to several CPA zeros. The CPA zeros are a-dependent and hence tunable within a given material. The S-matrix eigenvalue intensities are bounded by a tight lower bound, which goes to zero at the n values given in (7) and locates the interesting operating regions. For Si, the optimal wavelength occurs at around 750 nm for a = 10 μm, and around 1000 nm for a = 150 μm. The exact value depends on n(k), which in turn depends on the doping. The CPA functions as a compact absorbing interferometer. Unlike an ordinary interferometer, it does not deflect the input beams into another output channel, but causes it to be absorbed entirely within the material