Getting a Ti:sapphire laser to work
Thu Oct 26 10:56:43 EST 2006
1) Get the laser to lase. Use a bandpass filter at the laser wavelength in front of a powermeter on the output to detect when you're getting close. Without this filter there is too much scattered light in the non-laser modes to tell whether you are going in the right direction. Ti:sapphire has a gain region covering hundreds of nanometres.
2) Align the crystal to be at Brewster's angle to the laser cavity
mode to minimize losses. This is quite fiddly since there is no
rotation stage to adjust the crystal angle, so it must be done
freehand. Diagnostic is to operate the laser with low input power
and rotate crystal, realign cavity end mirrors to close the cavity
and translate the crystal until maximum power is reached for that
angle. Repeat this for different crystal angles, find the angle
that gives the maximum power.
The laser is operated at low power in the above step to avoid
jumping between different nonlinear lasing regimes with different
power outputs eg CW lasing versus modelocking. Pump laser power is
controlled by a waveplate and beamsplitter before the lens that
matches pump light to the cavity mode.
3) Align the mirror after the crystal to minimize the total astigmatism of the cavity. The crystal itself introduces astigmatism between the horizontal and vertical beam sizes, which can be cancelled by rotating the curved mirror after the crystal. Again, operate at low power, tweak horizontal adjust of mirror, reclose cavity with the end mirrors, then translate crystal and pump beam to overlap the new position of the cavity mode waist. This requires walking the translation of these two elements. Repeat this for a range of horizontal angles of the mirror after the crystal, chose the angle that gives maximum output power.
4) Longitudinal translation of the mirror after the crystal to map out cavity stability region. Translate mirror, close cavity with end mirrors, walk pump and crystal to give maximum power. Do this for a range of translations, plot out power versus mirror position. There will be two stability regions as the mirror is translated away from the crystal, with a sharply defined region in the middle where the laser stops lasing.
5) Choose a position of the mirror after the crystal just at the start of the second stability region ie further from the crystal. This seems to be better for modelocking. Then translate the saturable absorber back and forth, looking at the laser output on the optical spectrum analyser (nb: terminology confusion here, in my old lab we called scanning reference cavities optical spectrum analysers, here the device actually gives you the spectrum of a light source over a range of wavelengths simultaneously using a grating). Modelocking is indicated by the spectrum changing from a single narrow peak (CW behaviour) to a broad (several nm) hump.
[physics]
[permlink]