Thin-film silicon technology is a major candidate to comply with the ever-increasing global energy demand. The small thickness of silicon allows high industrial throughput and low material usage and subsequently opens new avenues to mass-production of low-cost solar cells. This small thickness, together with the indirect bandgap of silicon and thus its relatively weak absorption of long-wavelength light, requires methods to improve absorption. This is often called light trapping. Light absorption can be enhanced in a thin film solar cell by introducing light scattering functionalities, for example textured interfaces. Of common interest is to know: 1) how far the photocurrent and efficiency of a cell can be improved by applying these geometrical changes, 2) which geometries can satisfy the criterion of high photocurrent generation, and 3) how much the incident angle affects the two previous points. This thesis addresses the mentioned questions. Specifically, amorphous silicon (a-Si) cells are the main subject of focus in the thesis, however, the results and conclusions are applicable to other types of solar cells, and similar structures such as light-emitting diodes. The first part of the thesis is devoted to the electromagnetic theory for thin film multilayers and the numerical methods which were used for optical simulations during the PhD work. These methods are described and compared, and some common sources of numerical error in them are identified. To address the first and the third questions, the limits of light absorption enhancement, photocurrent generation and efficiency in thin-film solar cells are studied. As a result, we obtain the limits of absorption enhancement in thin films with periodic texture, over a wide angular and wavelength range. More specifically, first we extend the statistical temporal coupled-mode theory to the case of thin films with wavelength-scale grating couplers. Then, we use this theory to study the effect of the incident angle and the grating period on the absorption enhancement in an idealized thin film with a thickness of 200 nm and refractive index n=4. We show that absorption in a thin-film solar cell depends strongly on the grating period and angle of incidence; therefore, consideration of oblique incident of light in these cells is a necessity. We provide guidelines for the design of thin-film solar cells with periodic texture. Afterwards, we obtain the limit of absorption enhancement for different structures including a full thin-film a-Si solar cell stack for different grating geometries. We show that for thin-films, hexagonal gratings enhance absorption more significantly compared to square gratings. We identify parasitic absorption as a major bottleneck for photocurrent generation. To deal with realistic cases, we investigate the guided modes of thin-film a-Si solar cells by rigorous simulations (the second and the third question). First, we extract the guided modes of the cells and study them in an equivalent planar model. We show that a plasmonic mode exists for very thin buffer layers. Then, we focus on the effect of texture geometry over a broad angular range by comparing the short-circuit current density (Jsc) of the cells for different grating patterns. We find that based on the cell configuration, the optimal texture may be symmetric or asymmetric. We show that TM polarized light produces higher photocurrent at large incident angles regardless of the texture geometry. In the final part of the thesis, we study two novel configurations for thin-film solar cells. First a plasmonic a-Si cell is considered which does not include a buffer layer. We demonstrate that the Jsc of the plasmonic cell is sensitive to the thickness of the n-doped silicon layer and we find that for an n-Si thickness of less than 10 nm, the plasmonic cell outperforms a conventional a-Si cell. Then we simulate an a-Si cell with a periodic array of ZnO nanowires inside the active layer. Our simulations indicate that assuming a periodicity of around 500 nm, the Jsc is highest for a nanowire diameter of about 300 nm.