Experimental and numerical investigation of two-dimensional photonic crystals for application in integrated optics

Two-dimensional (2D) photonic crystals (PhCs) at near-infrared wavelengths are promising candidates for novel integrated optics applications. The main focus being on 'all PhC' monolithically integrated optical circuits. Because of well established fabrication technologies particular emphasis is on 'quasi' 2D PhCs where a 2D lattice of air holes is etched through a planar step-index waveguide providing the optical confinement in the third dimension. This approach has been successfully demonstrated both in GaAs and InP-based systems. We have studied vertical low-index-contrast structures both in GaAs and InP. These structures suffer inherently from radiation losses that are strongly affected by the air hole depth and shape. A semi-analytical 'ε"-model' enables out-of-plane scattering to be described by an effective imaginary dielectric constant in the air-holes which can be used in 2D finite difference time domain (FDTD) calculations. The model, once validated by 3D FDTD, allows us to deduce from optical simple slab transmission measurements the structural parameters (e.g. size, depth and form of the hole). The experimental transmission spectra are obtained with the so-called 'internal light source technique' (ILS), which allows for quantitative normalised transmission measurements, thus eliminating the uncertainties due to the in-out coupling arising in alternative characterisation techniques, i.e. end-fire. The ILS technique proved to be a versatile method without stringent limits for assessing the fabrication parameters (air-filling factor and loss) but also proved to be suitable for testing building blocks. In this thesis different building blocks were fabricated in the GaAs system and analysed using diverse modeling techniques (e.g. plane wave expansion, FDTD and coupled mode theory). The first building block was used to study contra-directional coupling of the fundamental mode with higher-order modes in channel waveguides. It is an extension of the work on waveguides with symmetric periodic corrugation to the case of multi periodicity. We showed that by breaking the translational symmetry in these structures referred to as hybrid waveguides, the number of mini-stopbands is increased and not, as one could have expected, decreased. A second building block analysed to guide light around the corner, are cavity resonant bends (CRBs). As our analysis showd that CRBs work best with non-degenerate cavity modes, different low-symmetry order cavities were studied with differ coupling coefficients between cavity and waveguides. Unfortunately, the resulting peak transmission and overall performance was below our expectations. An important building block in integrated optics are directional couplers. They have a potential for applications such as 3dB-intensity splitting or channel add-drop filtering. To reduce the coupling length, transmission measurements of a series of coupler structures of different lengths were taken. In the case of two W3 waveguides separated by a single row, the coupling length proved to be very dependent on the hole diameter in the single row. For this case the experimentally determined coupling length, which has been validated by both PWE and FDTD simulations, was of the order of 350 periods which corresponds to an absolute length of about 80 μm, a value that is small for conventional optics, but rather large for the domain of PhC optics where due to the unusually high propagation losses the integrated circuit has to be kept small. This value could, however, be reduced by a factor of at least three when reducing the ratio between the barrier hole and the nominal hole radius to 0.5. Finally we studied the waveguides needed to connect the different building blocks. The standard solution are channel waveguides, line-defects obtained by omitting a number of rows of holes. An alternative approach are coupled cavity waveguides (CCWs). They not only offer the potential for waveguiding but may also be used for non-linear optics by exploiting the special dispersion properties of CCWs (e.g. low group-velocity). A tight-binding model has been set-up that describes the interaction between the neighbouring cavities and that allows one to deduce the dispersion properties of the chain from the single cavity field distribution.