Optics Group -

Physics and Mathematical Physics ,   Adelaide University

The High Power Laser Project

One of the most important endeavours of 21st century physics is to detect gravitational radiation, and in so doing open a new window on the universe for astronomy. At Adelaide the optics group is part of a worldwide collaboration to build large scale interferometric detectors of gravitational waves. This collaboration comprises a collection of consortia in various countries throughout the world. In Australia there is the ACIGA consortium of which we are a part.
Our contribution is the development of high power lasers for the phase-II LIGO interferometers. LIGO is the consortium of US institutions which is building two interferometers. These lasers are not particularly powerful from the point of view of materials processing or users of pulsed lasers, but their combination of CW power, mode quality and stability is unprecedented. The LIGO II laser specification calls for a laser in excess of 100W.

The picture to the right is the LIGO I interferometer. (It's also a link!)

This photo shows the slab and the totally internally reflected path of the beam inside it.

Our approach is a Nd:YAG laser (1064nm wavelength) with the laser medium in the form of a slab, pumped (i.e. excited) from the side by diode laser light that is brought up to the slab with fibres. The innovative features of our design are management of the waste heat so that temperature gradients in the crystal don't cause excessive thermal lensing, and a stable/unstable cavity design. We use a two stage injection locking scheme to transfer the stability of a low power laser (the master oscillator) to our high power device (the slave).

The neodymium laser is pumped by diode laser light. This thumbnail shows the diode pump lasers in an earlier incarnation (only 80W pump power). Click on it to see the 520W rack of diode lasers that we are currently working with, (and Martin Ostermeyer doing up a nut). (60k)

Power scalable TEM00 cw Nd:YAG laser with thermal lens control

D.Mudge, M.O.Ostermeyer, P.J.Veitch, J. Munch, B.Middlemiss, D.J.Ottaway and M.W.Hamilton

IEEE Journal of Selected Topics in Quantum Electronics 6, 643-649 (2000)

(pdf) copyright IEEE

Not all of the pump light is actually converted into laser light and this means that there is waste heat generated in the slab. As this flows out, a non-uniform temperature distribution is created. The refractive index of the crystal is temperature dependent, so the temperature distribution causes unwanted prismatic and lensing effects. To control the temperature distribution we cool the slab from the sides (the same sides as the pump light enters) and inject extra heat from the top and bottom. This makes the temperature distribution in the vertical axis uniform, at least over the volume occupied by the laser light. The zig-zag path of the light in the horizontal direction, a function of the slab geometry, averages out the effect of temperature gradients in that direction. We have demonstrated that this works!

Because we control the temperature gradients in the vertical axis, but not the horizontal, we separate the design of the horizontal and vertical sections of the laser cavity (i.e. the mirror arrangement that provides feedback of the laser light to the crystal). In the horizontal we have a fairly standard stable cavity design, with natural apertures in the cavity giving mode control. The cavity in the vertical direction is unstable, with magnification of the beam within the cavity and a graded reflectivity output mirror. We have demonstrated that this works too!

Click on thumbnail to see slide showing temperature gradient control of stable/unstable operation.

Stable cavity: the focussing of the light by the cavity mirrors ensures that there are ray paths which do not wander out of the cavity.
Unstable cavity: ray paths are allowed to wander out of the cavity, and this loss from the cavity is the output light.
The advantage of an unstable cavity is that the laser beam inside the cavity can be broader and make better use of the pumped laser medium.

The injection locking scheme is a two stage arrangement because our stable low power laser (200mW) doesn't have enough output power to reliably enslave the frequency of the high power laser on its own. At an earlier stage of this project we demonstrated that we can use this low power laser to stabilise a medium power laser (5W). This process added negligible extra noise, so the stability of the low power laser really does govern the overall stability. (Stabilising low power lasers directly is much easier than with high power ones.) The medium power laser will then be used as the master oscillator for the high power system.