"The aim of this project is to advance the performances of a spinelectronics based microwave oscillator. These spintronics oscillators (STO) make use of the recently discovered spin transfer torque effect by which the magnetization of a thin magnetic layer can be set into perpetual steady state oscillations. These magnetization oscillations are translated into an ac output voltage by virtue of the giant or tunnelling magneto-resistance while its Gigahertz frequency can be tuned by the spin polarized DC current. Due to their extremely small size (100 nm) it is expected that a single conventional oscillator (of the type VCO voltage controlled oscillator) can be replaced by 50 spin torque oscillators thus largely increasing the covered frequency range at the same cost of CMOS space. This is of great potential for future telecommunications applications, where incorporating as many standards as possible on a single device, at reduced cost and space, poses a major challenge. However, one of the major bottlenecks for the implementation of STOs is their relatively large microwave emission linewidth (10 – 100 MHz) that is little understood. The aim of this project is therefore to investigate the physical mechanisms that limit the temporal coherence using improved magnetic stacks of the oscillator device that lead to stable microwave emission lines compatible with standard microwave applications. In particular, for the given oscillator configurations, the frequency and amplitude fluctuations will be identified and characterized with respect to their corresponding timescales and correlation. This will be done using recently developed time domain experiments. The ultimate goal is to extract phase noise, which is the parameter relevant for applications. This will be done by using numerical analysis techniques of signal traces from time domain measurements and by phase locked loop techniques."