Precise control over the properties of laser light is becoming increasingly important with rapid technological advancement. Even in isotropic materials, we can achieve diverse intensity profiles of laser beams, and even more complex ones can be created using optically anisotropic materials. Liquid crystal molecules have interesting optical properties, one of the basic ones being optical birefringence. By orienting the molecules of the liquid crystal, the optical axis of the material changes, which affects
the propagation of light. If we use an optically birefringent liquid crystal in a laser resonator, the resulting light modes can be controlled by rotating the liquid crystal molecules. In the master’s thesis, I explore the design of laser modes in optically birefringent resonators based on liquid crystals. The generation of a training set used for training a neural network, which accelerates the calculation of the resonator’s light eigenstates, is presented. The advantages and disadvantages of using machine learning for this purpose are investigated. Finally, an algorithm is presented with which we can efficiently design a birefringent laser resonator using a neural network and numerical optimization, so that the desired light eigenmode is established in it.
|
