Materials that can absorb all incident light are essential to fabricate efficient light detectors. Very good light absorbers exist, but absorption occurs in a layer that is much thicker than the wavelength of light. To construct an efficient and sensitive detector, the absorption must occur in a volume that is as small as possible in order to detect the energy deposited into the material. We have investigated the potential of niobium nitride (NbN), which can absorb all incident light in a very thin, 4.5nm-thick film (about 200 times thinner than the incident-light wavelength used in our experiment).
The absorption process in NbN converts light energy to heat. Thinner detectors are more sensitive because the same amount of heat in a small volume leads to a larger change in temperature. In fact, the layer in our study is so thin that it can be used for detectors that are sufficiently sensitive to detect absorption of a single photon.1 Typical detectors (see Figure 1) currently have limited efficiencies (10–20%), largely because of their limited optical absorption. Improving efficiency requires a deeper understanding of the optical properties of thin NbN films.
Figure 1. Scanning-electron-microscope image of a typical niobium nitride (NbN) detector. (Credit: Sander Dorenbos, Delft University of Technology, the Netherlands.)
We considered the simplest geometry, consisting of a thin film with dielectrics on either side. Absorption is a function of film thickness, and its absorption properties can be calculated using Fresnel's equations for reflection and transmission. If the film is too thin, most of the light is transmitted, while if it is too thick, it becomes a good reflector. At normal incidence, an optimal film thickness exists where the absorption (A) is maximal,2
Here, n1 and n2 are the refractive indices of the materials on either side of the film, which is illuminated through material 1.
For a symmetric system (n1=n2) the absorption cannot exceed 50%. On the other hand, for an asymmetric system it can be increased by illumination from the side of the material with the highest refractive index. The absorption is still limited, however, because it is hard to find a material with a refractive index at visible wavelengths greater than 3.5. To move beyond this limit, one needs to change the angle of incidence.3 Figure 2 shows calculations (curves) and measurements of the absorption by a 4.5nm-thin NbN film (illuminated from the sapphire substrate) for different angles of incidence and polarizations. At the critical angle for total internal reflection (dashed line), the absorption of light that is polarized parallel to the film (s polarization: blue curve) reaches a measured maximum of 94%. Destructive interference occurs between the multiple reflections from the film. Since the light cannot be transmitted at this angle, all light is absorbed. For the same angle, the orthogonal p polarization (red curve) reaches a minimum. All incident light is now reflected. The same principle applies to a detector geometry of parallel lines (see Figure 1), so that the absorption by a realistic detector could be as high as 94% for a given polarization.
Figure 2. Measured (solid and open circles) and calculated (curves) absorption for a 4.5nm-thin NbN film as a function of angle of incidence. The dash-dotted line indicates the angle for total internal reflection.
Figure 3. Possible setup for a polarization-independent detector. s and p: Parallel and orthogonally polarized beams, respectively. 4: Quarter-wave plate (λ: Wavelength).
In summary, a realm in optics that had not been explored to date has been opened up by studying very thin films with extremely efficient absorption. Special materials are needed that have a very high resistivity (orders of magnitude higher than common metals such as gold or silver). In addition, the films must be very thin and uniform. NbN combines both properties and offers the possibility to make detectors with significantly improved efficiency. For a single polarization, 94% absorption can be achieved. Polarization dependence can be overcome by taking advantage of the fact that nonabsorbed light is reflected. A second, well-placed detector (or the same detector using a quarter-wave plate and a mirror) can detect the reflected beam (see Figure 3). Our future directions include demonstration of a working detector that uses enhanced optical absorption. We also aim to gain a deeper understanding of the origin of the unusual dielectric constant of NbN and related materials.
Eduard Driessen, Michiel de Dood
Huygens Laboratory, Leiden University
Leiden, The Netherlands
1. G. N. Gol'tsman, O. Okunev, G. Chulkova, A. Lipatov, A. D. Semenov, K. V. Smirnov, B. M. Voronov, A. Dzardanov, C. Williams, R. R. Sobolewski, Picosecond superconducting single-photon optical detector, Appl. Phys. Lett. 79, pp. 705, 2001. doi:10.1063/1.1388868