Reliable Pulse-Retrieval for Frequency-Resolved-Optical-Gating

Frequency-resolved-optical-gating (FROG) measures the full temporal information of arbitrary ultrashort laser pulses' fields using two-dimensional data in the time-frequency domain. By benefiting from the properties of the acquired 2D data, FROG not only measures stable pulses accurately, but also it indicates the presence of pulse-shape instability in the pulse train by a discrepancy between the measured and retrieved FROG traces using its pulse-retrieval algorithm. However, this discrepancy could also arise when the pulse-retrieval algorithm stagnates. When the current best algorithm, generalized projections (which is known to stagnate as much as half the time) is used to retrieve the field from a trace of unstable pulse trains, the retrieved fields may vary significantly from one retrieval to the other, while never returning a pulse that matches the measured trace. So, it is difficult to distinguish the presence of instability from the stagnation of the algorithm. In this dissertation, the Retrieved-Amplitude N-grid Algorithmic (RANA) approach is introduced for the 100% reliable pulse-retrieval from FROG traces of stable pulses with even high complexities and significant noise for the commonly used FROG geometries. Further, the performance of the RANA approach for traces contaminated with artifacts due to averaging over a pulse train with unstable pulse shapes is investigated. The results indicate that the RANA approach returns pulses that yield the lowest achievable discrepancy between the measured and retrieved traces for the traces of unstable trains. Thus RANA approach convincingly indicates the stability or instability of the pulse train, solving the problem. It also yields the correct approximate pulse length, spectral width, and time-bandwidth product, including some structure when present.Ph.D