Noise in fundamental and harmonic modelocked semiconductor lasers: Experiments and simulations

Measurement of record-low residual jitter in 40-GHz monolithic mode-locked lasers

“…The noise spectra have a characteristic white noise plateau, which is as low as -123dBc/Hz for the l-QW device, showing good qualitative CThN2 agreement with the theoretical predictions in [10], and quantitative agreement with the absolute phase noise measurements on the same devices in [2]. The "white" spontaneous emission noise is filtered by the cavity with a characteristic corner frequency and a 30-35 dB roll-off, which again qualitatively agree well with the theoretical results of [10] and with the experimental results in [4] and [6]. The lower graph of figure 2 shows the intrinsic laser jitter integrated from a variable lower limit to the Nyquist frequency of 20 GHz.…”

“…Our measurement technique is based on the von der Linde method [3], but overcomes the drawbacks of oscillator noise and spectrum analyzer thermal noise, which reduces the important high frequency dynamic-range. Similar residual phase-noise measurements have been performed at lower repetition rates [4][5][6][7] such as 10 GHz or lower. However, when increasing the repetition rate the upper limit of integration scales with this rate and requires a very high bandwidth of the electronics.…”

Noise of mode-locked lasers (Part II): timing jitter and other fluctuations

“…It has been pointed out [34] that for a given pulse repetition rate one can achieve a lower timing jitter by harmonic mode locking, i.e., by using a longer laser cavity with multiple equally spaced circulating pulses. This result becomes obvious through the discussion above.…”

“…This result becomes obvious through the discussion above. However, harmonic mode locking is not only affected by the greater difficulty of making a longer cavity mechanically stable, but is also plagued by supermode noise [34], which can strongly increase the noise powers near harmonics of the pulse repetition frequency. Therefore, harmonically mode-locked lasers need special efforts [28] to realize the theoretical potential for very low timing noise over a wide frequency span.…”

Wide-band residual phase-noise measurements on 40-GHz monolithic mode-locked lasers

“…We have recently reported design principles for all-active mode-locked lasers with ultralow amplitude and timing jitter [3], but were unable to measure anything but a rough upper estimate of the jitter of the lasers: 400 fs in the integration range 10 kHz-20 GHz. Important figures of merit such as phase-noise plateau level and knee position [4] were not measurable. No distinction between different designs was possible.…”