OpenWorldLab

Loading interactive view…

Lasers

Passive mode-locking

A laser running freely oscillates in many longitudinal modes c/2L apart, with random phases. Lock the phases and the modes add up to one short pulse bouncing back and forth; each time it reaches the output coupler a little leaves. A fast saturable absorber does the locking by itself, because it lets an intense spike through more easily than the weaker light around it.

What you are looking at

A Nd:YVO₄ crystal beside a flat high reflector, a lens that focuses the mode near the far end, a semiconductor saturable absorber and a 95% output coupler. A second lens collimates the output, a pick-off sends a little of the train through an ND filter to a fast photodiode, and an autocorrelator measures the pulses.

How the bench is set up

853 mm between the mirrors, a 3 mm Nd:YVO₄ crystal pumped with 15 W, a 50 mm lens 53 mm from the output coupler, and an absorber passing 99% unbleached that recovers in a picosecond. A second 50 mm lens collimates the output. Photodiode bandwidth 1 GHz.

The theory

The pulse returns to the coupler once per round trip, so the train repeats at the free spectral range,

frep=c2L,f_{\text{rep}} = \frac{c}{2L},

with LL the optical length (the crystal counts nn times its length). Haus’s master equation holds the pulse in shape: each round trip the gain line narrows the pulse’s spectrum and so stretches it, while the absorber, saturating, trims its weak wings. The two balance in a sech2\mathrm{sech}^2 pulse. With a weakly saturated absorber of depth q0q_0, a gain gg on a line of half-width Ωg\Omega_g, intracavity pulse energy WW and absorber saturation power PAP_A,

τ≈4gPAΩg2 q0 W.\tau \approx \frac{4 g P_A}{\Omega_g^2\, q_0\, W}.

The lab solves the saturated case exactly. An intensity autocorrelator cannot see the pulse directly; it reports the width of its autocorrelation, 1.543 times the pulse for sech2\mathrm{sech}^2.

What to measure

Repetition rate c/2L and a sech² pulse

  1. Read the photodiode: a train of spikes about 5.7 ns apart, 175 MHz. The detector is too slow to resolve each pulse, so each spike has the photodiode’s own shape.
  2. Work out c/2L from the mirror spacing plus (n − 1) × 3 mm for the crystal, n = 2.17.
  3. Read the autocorrelator: about 16 ps, so the pulses last about 10 ps.
  4. Raise the absorber’s transmission to 99.5%, halving its depth: the pulses lengthen to about 14 ps.

What the model leaves out

  • The absorber responds instantly and the gain line is taken as a parabola near its peak. Dispersion and the Kerr effect are left out of the bench, so there is no soliton shaping.
  • Q-switched mode-locking is flagged, not simulated: if the pulse is too weak to keep the absorber bleached the inspector warns that the train would break into bursts.

Things to try

  • Move the high reflector and the crystal 40 mm to the left: the rate drops to c/2L, about 167 MHz, and the pulses barely change.
  • Lower the pump to 4 W: the inspector warns that the pulse is too weak to keep the train steady (Q-switched mode-locking).
  • Swap the absorber for Cr:YAG, which recovers in microseconds: the laser Q-switches instead.

References

  • A. E. Siegman (1986), Lasers, University Science Books.
  • H. A. Haus (2000), Mode-locking of lasers, IEEE Journal of Selected Topics in Quantum Electronics 6, 1173.
  • U. Keller et al. (1996), Semiconductor saturable absorber mirrors (SESAMs) for femtosecond to nanosecond pulse generation in solid-state lasers, IEEE Journal of Selected Topics in Quantum Electronics 2, 435.
  • C. Hönninger et al. (1999), Q-switching stability limits of continuous-wave passive mode locking, Journal of the Optical Society of America B 16, 46.

More experiments, and how the lab works, are in the laser lab’s overview.