Nanophotonic devices have enabled microscopic control of light with an unprecedented spatial resolution by employing subwavelength optical elements that can strongly interact with incident waves. However, to date, most nanophotonic devices have been designed based on fixed-shape optical elements, and a large portion of their design potential has remained unexplored. It is only recently that free-form design schemes have been spotlighted in nanophotonics, offering routes to make a break from conventional design constraints and utilize the full design potential. In this review, we systematically overview the nascent yet rapidly growing field of free-form nanophotonic device design. We attempt to define the term “free-form” in the context of photonic device design, and survey different strategies for free-form optimization of nanophotonic devices spanning from classical methods, adjoint-based methods, to contemporary machine-learning-based approaches.

The optical properties of thin-film light emitting diodes (LEDs) are strongly dependent on their structures due to light interference inside the devices. However, the complexity of the design space grows exponentially with the number of design parameters, making it challenging to optimize the optical properties of multilayer LEDs with rigorous electromagnetic simulations. In this work, we demonstrate an artificial neural network that can predict the light extraction efficiency of an organic LED structure in 30 ms, which is ∼103 times faster than the rigorous simulation in a single-treaded execution with root-mean-squared error of 1.86 × 10−3. The effective inference time per structure is brought down to ∼0.6 μs with unaltered error rate with parallelization. We also show that our neural networks can efficiently solve the inverse problem – finding a device design that exhibits the desired light extraction spectrum – within the similar time scale. We investigate the one-to-many mapping issue of the inverse problem and find that the degeneracy can be lifted by incorporating additional emission spectra at different observing angles. Furthermore, the forward neural network is combined with a conventional genetic algorithm to address additional large-scale optimization problems including maximization of light extraction efficiency and minimization of angle dependent color shift. Our approach establishes a platform for tackling computation-heavy optimization tasks with one-time computational cost.

Increasing the light extraction efficiency has been widely studied for highly efficient organic light-emitting diodes (OLEDs). Among many light-extraction approaches proposed so far, adding a corrugation layer has been considered a promising solution for its simplicity and high effectiveness. While the working principle of periodically corrugated OLEDs can be qualitatively explained by the diffraction theory, dipolar emission inside the OLED structure makes its quantitative analysis challenging, making one rely on finite-element electromagnetic simulations that could require huge computing resources. Here, we demonstrate a new simulation method, named the diffraction matrix method (DMM), that can accurately predict the optical characteristics of periodically corrugated OLEDs while achieving calculation speed that is a few orders of magnitude faster. Our method decomposes the light emitted by a dipolar emitter into plane waves with different wavevectors and tracks the diffraction behavior of waves using diffraction matrices. Calculated optical parameters show a quantitative agreement with those predicted by finite-difference time-domain (FDTD) method. Furthermore, the developed method possesses a unique advantage over the conventional approaches that it naturally evaluates the wavevector-dependent power dissipation of a dipole and is thus capable of identifying the loss channels inside OLEDs in a quantitative manner.

It is well known fact that human localize the direction of sound source in transverse plane according to interaural time difference in low frequency and interaural level difference on high frequency. In this research, we tested possibility for interaural time difference to be interpreted as interaural level difference when it is perceived by brain. To test the hypothesis, we used sinusoidal wave and commercial music to study the effect of interaural time difference on the perception of loudness. Each sound sample was regulated to have interaural time difference and given to 9 subjects. Subjects were asked to answer which side was louder. For sinusoidal waves, more than 60% of subjects answered side with earlier sound arrival was louder. The effect of interaural time difference on interaural level difference was larger when it interaural time difference was given to lower frequencies. Music samples were regulated with 100 and 500 interaural time difference and 100% of subjects answered side with earlier sound arrival was louder. Results from the experiments imply the possibility for interaural time difference interpretation by interaural level difference on localization of sound sources.