Abstract. We demonstrate for the first time active dispersion and amplitude correction in a fiber laser producing sub-45 fs pulses. The approach is based on single-shot second and third order dispersion measurement based on multiphoton intrapulse interference. The same principle is applied to obtain time-resolved measurements of the transient dispersion induced by an intense laser pulse on a silica window.
PrincipleUltrafast laser systems have proven to be particularly adequate for applications such as material ablation and micromachining, nonlinear spectroscopy and sensing, because of their ability to easily deliver peak power densities of >10 12 W/cm 2 . While these laser sources have experienced great progress in terms of output pulse characteristics, reliability, and size, they are still notoriously complex and sensitive to changes in ambient environment, leading to alignment drifts, deteriorated pulse characteristics, and depreciated performance over time. Applications of femtosecond lasers outside of laser laboratories thus require automated dispersion and amplitude drift compensation in order to maintain optimum performance without human assistance.It is reasonable to assume that such phase drifts would manifest themselves primarily in the second-and sometimes third-order dispersion (SOD and TOD, respectively) of the original pulse waveform. Here we take advantage of the inherent sensitivity of nonlinear optical processes to phase distortions. We apply the concept of multiphoton intrapulse interference phase scan (MIIPS) [1], in order to determine SOD and TOD from changes in the second harmonic generation (SHG) spectrum of the a reference pulse. Intensity drifts are simply measured and corrected by adjusting the pulseshaper transmission (amplitude) mask. This real-time version of MIIPS (RT-MIIPS) scheme can be used for monitoring and correction of pulse energy and peak power drifts, allowing for optimum unattended ultrafast laser performance over an indefinite period of time.The principle behind RT-MIIPS is outlined in Figure 1, where a single local SHG maximum is formed by adding a cubic reference phase mask on an otherwise transform limited pulse. Changes in the magnitude and sign of SOD affect the maximum position within the SHG spectrum. By establishing a one-to-one correspondence between the SHG peak position and the amount of SOD through calibration, one can later accurately measure and correct the change in SOD after acquiring a single SHG spectrum.