|The Web Newsletter||Front Page||Hanford||Livingston||MIT|
LIGO Caltech NewsRecycling Experiment Entering Final Phase at 40 Meter Laser Interferometer
Since our last newsletter, the major effort of the 40 meter facility has been to modify the optical configuration of the interferometer through the addition of a "recycling mirror" (see Figure 1 below). Laser light is modulated using a Pockels cell to create a "carrier" or unmodulated frequency and "sidebands" at frequencies above and below the carrier. In the previous "recombined" configuration, the recycling mirror was not present. For noise reasons the photodetector was operated on a dark fringe for the carrier so that carrier light returning from the arms interfered constructively in the direction of the laser and was consequently lost from the system. The newly installed recycling mirror now reflects this light back into the interferometer, coherently adding it to the input laser beam. This increases the circulating power and therefore increases the overall interferometer sensitivity.
Many modifications were made to the interferometer in order to achieve recycling. The recycling mirror and its associated suspension and control electronics were added, a new optical "pickoff" control servo was added, and the RF modulation source and an associated resonant transformer were changed to operate at 32.7 MHz. (This is the frequency that allows both the carrier light and the sidebands to be simultaneously resonant in the recycling cavity formed between the recycling mirror and the input mirrors in each arm.) Also the configuration of optical levers used to sense and control angular alignment was adjusted to accommodate the installation of the additional mirror.
Demonstration of operation in a power recycled mode proceeded through a step-by-step process of bringing subsystems of increasing complexity within the interferometer into their correct operating configurations, or "lock." At each step, comparisons of the observed interferometer behavior were made to the lock acquisition model ("SMAC") developed earlier. (SMAC, the Single Mode Acquisition Code, was used to design the interferometer control topology and we utilized its predictions here to lock the subsystems in the correct order. For more about SMAC, see the articles in newsletter Volume 1, Number 6 and Volume 2, Number 6.) Some differences with the model were found. For example, intuitively we expected a sign flip in the beamsplitter servo when the second arm came into lock; this is also what the software model predicts. When we removed the sign flip the interferometer leapt into lock and we observed solid lock sections of 10 minutes or more. There are two most likely reasons why the sign flip is unnecessary. One is that the carrier is undercoupled in the interferometer, i.e., the light coming back into the recycling cavity from the front mirror of the Fabry-Perot cavity is mostly from reflection off the back of the mirror and only a little bit of the light is due to leakage of stored light out of the Fabry-Perot cavity. Alternatively, it may be due to poor mode matching into the interferometer, i.e., the intensity profile of the beam stored in the Fabry-Perot cavity does not smoothly match the intensity profile of the light within the recycling cavity, allowing us to effectively lock the beamsplitter on non-mode matched light. Both possibilities are now under investigation.
Figure 1. The configuration of the 40 Meter Laser Interferometer on the Caltech campus. The recycling mirror, seen at left in the diagram, was installed last fall.
Figure 2 belows shows the power recycled interferometer acquiring and then losing lock. There are four traces shown; the lower two show the south and east arm transmitted light levels (SAT and EAT voltages). These are measures of the amount of optical power circulating in each arm: a higher level indicates greater circulating power. The upper trace (labeled APD) shows the light level incident on the photodetector of Figure 1. The trace below (labeled SPD) shows the amount of light returning to the laser from the recycling cavity. In the diagram, the ground level for each trace is indicated by the arrow beside the trace number. Also, the APD and SPD have reversed polarities so that increasing light levels are indicated by decreasing voltages. At the beginning of the plot the power recycled Michelson is resonant for the sidebands and anti-resonant for the carrier. The east arm comes into resonance for the carrier after approximately 100 msec. Since the power recycled Michelson cavity is essentially anti-resonant for the carrier light, very little light leaks into the east arm and hence the light build up in this cavity is small. Note that the light which does return from this cavity satisfies the conditions for resonance in the recycling cavity and thus a small fraction of the carrier light starts to resonate within the power recycled Michelson. After approximately 600 msec the south arm comes into resonance. Now the carrier is resonant in the recycling cavity and thus the light incident on the arm cavities is enhanced. This is shown by the dramatic increase in circulating power in the arms. As this happens, the amount of light reflected back towards the laser decreases as shown by the SPD trace. Note that the APD, EAT, and SPD traces show correlated fluctuations. By observing the changing shape of the laser spot due to the light incident on the photodetector of Figure 1, it is evident that the fluctuation is mainly due to orientation changes in the mirrors. The implementation of wave front sensing (see below) is expected to improve the angular alignment stability.
Our attention is now focused primarily on the implementation of wave-front sensing in order to achieve more precise angular alignment of the suspended optics. Wave front sensing makes use of the fact that higher order modes of the resonant cavity can be excited by a TEM-00 gaussian beam entering a resonant cavity if the end mirrors of the cavity are misaligned. The amplitudes of the higher order modes therefore contain information about the angular misalignments of the mirrors that can be used to feed back into a servo system in order to maintain alignment stability. The hardware and software needed to implement wave front sensing has been installed and is currently being tested.