Speakers
Description
Advances in laser technology have enabled the generation of optical pulses approaching a single optical cycle in duration and achieving peak powers of up to 10 PW. Such ultrashort pulses are essential for a wide range of applications, including high-harmonic generation (HHG), fusion science, and attosecond science, where pulse-shape stability is of critical importance. Since the early days of ultrashort-pulse characterization, shot-to-shot variations in pulse intensity and phase have posed challenges for accurate pulse measurements. Intensity autocorrelation, one of the earliest techniques used to characterize ultrashort laser pulses, often shows a narrow spike atop a broad background when measuring unstable pulse trains. The spike corresponds to the coherent component of the pulse train. Because this spike is always narrower than the average pulse in the train, it tends to represent the shortest temporal features present rather than the true average pulse duration. This issue is not limited to intensity-based measurements and can also affect interferometric techniques that measure spectral phase. The underlying reason is that long, temporally complex pulses generally possess a more structured spectral phase, whereas shorter and simpler pulses are associated with a flatter spectral phase. For a given optical spectrum, the transform-limited pulse corresponds to a flat spectral phase and therefore represents the shortest achievable pulse duration. Consequently, averaging the spectral phase over an ensemble of different complex pulses inevitably produces a measured spectral phase that is artificially flatter than that of the individual pulses [1]. Fortunately, multi-shot variants of FROG have shown qualitatively pulse-shape instabilities through discrepancy between measured and retrieved traces [2]. However, similar discrepancies can also arise from stagnation in the retrieval algorithm. The recently developed RANA approach addresses this issue. It provides highly reliable reconstruction even for complex even up to TPB of 100 and in presence of noise [3].
We introduce a quantitative method to measure pulse-shape instability by analyzing systematic errors between measured and retrieved FROG traces. The approach extends the classical one-dimensional runs test to two-dimensional FROG data by counting connected regions of identical sign in the difference trace. Random noise produces many small alternating regions, whereas systematic errors form larger connected areas. To reduce noise-dominated contributions at the trace edges, each 2D run is weighed by the measured trace
