Red, yellow and green LEDs based on GaAs and related materials have been available since the 1960s. However, wide band gap semiconductors suitable for blue LEDs were difficult to develop and the first efficient, long-lifetime blue GaN-based LED was not fabricated until 1993. This revolutionised the field of solid-state lighting and kick-started a considerable research effort focusing on GaN and related materials. However, it soon became clear that high densities of threading dislocations in the GaN films were severely limiting the performance of GaN-based devices that emit in the violet and UV region. Much research effort has therefore been focused on dislocation density reduction, in an effort to develop efficient violet and UV-emitting LEDs and lasers. These emitters are useful in their own right for water sterilisation, data storage and communications, but are also needed for white LEDs, which are commonly based on the combination of a violet or UV LED and phosphors which convert the short-wavelength light into an appropriate mixture of red, green and blue wavelengths. If the performance of white LEDs can be improved via defect reduction, then solid-state lighting with a luminous efficacy and lifetime considerably better than conventional technology will result. Such an initiative is particularly important in light of a recent EU announcement regarding the banning of sale of conventional light bulbs. We believe that bright, white LEDs offer the best solution for lighting the UK; fluorescent energy-saving bulbs contain harmful metals such as mercury (creating a waste disposal problem), are large in size and often provide poor colour rendering. The aim of our short proposed project is therefore to reduce or eliminate (a) threading dislocations in hexagonal GaN (a form of GaN particularly suitable for LEDs) and (b) stacking faults in cubic GaN (a form of GaN particularly suitable for lasers). This ambitious aim is prompted by our initial research, which indicated that novel transition metal nitride interlayers have the ability to block such dislocations from propagating into the active light-emitting region of a device. We intend to develop this concept by growing such layers in-situ in a single growth process using the same technique that is used to grow the GaN, namely MOCVD for hexagonal GaN or MBE for cubic GaN. This is expected to provide both practical benefits (no interruption of the growth process, as was previously required) and to provide significant improvements in dislocation density reduction via improvement in the roughness of the layers, among other factors. However, we also intend this work to pave the way for integration of transition metal nitrides with GaN devices, which could improve LED contact technology, improve buffer layer technology and allow superconducting TM-nitride/insulating GaN heterostructures (Josephson junctions) of high structural quality to be fabricated, leading to application such as ultra-sensitive magnetometers for use in healthcare and geophysics.