Pitfall in autocorrelation measurements of laser radiation

Spectrally broad laser radiation from continuous wave (cw) lasers can exhibit second-order autocorrelation traces virtually indistinguishable from those of mode-locked lasers. Consequently, based only on autocorrelations, one might erroneously conclude that a cw laser is mode-locked. This pitfall in interpretation can be avoided by carefully characterizing radio frequency transients and spectra. However, optoelectronics are often too slow for lasers with an axial mode spacing in the multi-GHz range. Carefully evaluated autocorrelations then remain the last resort for validating mode locking. We demonstrate in detail what needs to be observed. We compare autocorrelation measurements and calculations of a mode-locked titanium-sapphire (Ti:Sa) laser with 76 MHz repetition rate and a spectrally broad monolithic cw Ti:Sa laser and devise a new, additional measurement to safeguard against misinterpretation of their autocorrelations.