Caltech 40 Meter Team locks Advanced-LIGO-like Prototype Interferometer
- Contributed by Alan Weinstein
A small but dedicated team of physicists and engineers, working at LIGO's 40 Meter Prototype Interferometer on the Caltech campus, has been striving towards the control of a suspended-mass interferometer with the planned Advanced LIGO optical configuration. Advanced LIGO optics include an additional suspended mirror in the core interferometer, increasing the number of length degrees of freedom to be controlled from four (in Initial LIGO) to five. In November 2004, the team achieved a significant milestone: the control of all five degrees of freedom. More work is required to bring the interferometer into the correct operating point for gravitational wave detection, but that ultimate goal is now on the horizon.
This milestone was established back in 2001, when Initial LIGO was coming on line at the Hanford and Livingston sites. Many of the key technologies for Initial LIGO were first tested at the 40 Meter lab, and it's where many of LIGO's interferometer scientists learned their trade. The lab continues to play these same important roles for Advanced LIGO.
When a LIGO interferometer is first turned on, the mirrors are swinging freely from pendulum suspensions; sensing and controlling their microscopic positions is virtually impossible. The interferometer must "acquire lock" in stages, through a procedure that requires cleverness, patience, a thorough understanding of the interferometer dynamics, and fast embedded computers.
The Advanced LIGO optical configuration may seem at first glance to be only a small tweak on Initial LIGO: the addition of a "signal mirror" at the asymmetric port of the Michelson interferometer to perform "signal recycling" or "resonant signal extraction." But this extra mirror plays a very different role from all the others. The beam splitter, test masses, and power recycling mirrors in the Initial LIGO configuration serve to store carrier laser light in a resonant "reservoir." When a gravitational wave passes, some of that light is converted to "signal sidebands" which exit the asymmetric port of the interferometer towards a photodetector, unimpeded. For Advanced LIGO the addition of a signal mirror now stands in the way of the signal sidebands, recycling them back into the interferometer at some frequencies and resonantly extracting them at others, in order to shape the frequency response of the detector for optimal sensitivity to gravitational waves. The signal mirror does not "see" the carrier laser light until a gravitational wave induces the signal sidebands, but when that happens it must be precisely pre-positioned (microscopically) in order to manipulate the signal as intended. This means that the sensing and control of the signal mirror must be accomplished without the help of carrier laser light, in contrast to the way all mirror positions are sensed in Initial LIGO.
As a result, the "power- and signal-recycled Fabry-Perot Michelson" optical configuration of Advanced LIGO requires a new and much more complex scheme for sensing and controlling the microscopic positions of all the mirrors. The LIGO Scientific Collaboration (LSC) Advanced Interferometer Configurations working group established the need for this more complex configuration five years ago; it is one consequence of the need for higher laser light power. But it was clear that new approaches to the sensing of the mirrors were required, and that thorough tests of these schemes must be performed at a "full-up" LIGO-like interferometer; the 40 Meter lab was the obvious place to do this. Over the last five years, the 40 Meter lab has been completely rebuilt (keeping only the vacuum enclosure and seismic isolation stacks) in order to test lock acquisition and control schemes for the Advanced LIGO interferometers.
For most of the last five years, the incredibly talented team of physicists and electrical, mechanical, and optical engineers that helped build Initial LIGO have labored to assemble the infrastructure for yet another LIGO interferometer, with a pre-stabilized laser, ten precision optics suspended as pendula, optical beamlines laid out precisely in a claustrophobic laboratory and vacuum enclosure, and a full complement of servos, slow controls, and data acquisition systems. A steady stream of Caltech undergrads, NSF Research Experiences for Undergraduate summer students, and scientific visitors from LSC collaborating institutions contributed to the emerging interferometer. Then, in the last six months, efforts to lock and control the full interferometer were in full swing, greatly aided by visitors from Japan, Germany, Scotland, and Australia. By Thanksgiving eve 2004, all that hard work finally paid off: lock was acquired in all five length degrees of freedom of the dual-recycled Fabry-Perot Michelson interferometer, which was kept under control for on the order of one minute (despite rush hour traffic on Del Mar Avenue a block away). The configuration was not the correct one for resonant signal extraction (the arm cavities were locked somewhat away from resonance, for technical reasons), and moving to the correct configuration is the next step. Meanwhile, small parties celebrating crucial steps towards this milestone culminated in a particularly festive celebration that Friday.
Of course, this means that the real work at the 40 Meter lab is only just beginning. Acquiring lock is "easy" when all the servo controls are tuned precisely, but that tuning can't be accomplished unless the interferometer is already in lock--it's a chicken-and-egg problem. In order to acquire lock for the first time, we relied on "crutches"--clumsy intermediate steps that won't be necessary once the interferometer is fully tuned up. We "dither-locked" the Michelson by shaking the beam-splitter at 1230 Hz to help us sense its position relative to the "dark fringe." We "offset-locked" the high-finesse Fabry-Perot arms to help us find its narrow resonant point. We acquired lock with one set of servo sensors, then "handed-off" the control to the set of low-noise servo sensors that will, ultimately, keep the interferometer in the correct locked configuration for detecting gravitational waves (which the 40 Meter is perfectly capable of doing, albeit with reduced sensitivity due to its puny arm lengths).
Now we have to tune things up so that we can lock all five degrees of freedom in the correct configuration for resonant signal extraction. We have to teach our computers to do all this in a fast, smooth, optimized and fully automated sequence. We have to verify, through detailed simulations, that all this will work seamlessly with four-kilometer arms and multiple-pendulum suspensions. Also, in the next six months, we plan to design, build and install an output mode cleaner and a "homodyne" gravitational-wave detection system. Further on, we hope to use an MIT-designed system to inject squeezed light into the asymmetric port, to reduce the quantum noise and thereby further improve the sensitivity to gravitational waves. All this will be occasion for more celebration on the challenging road towards gravitational-wave detection and study.
Members of the 40 Meter team, including visiting scientists, expressing pleasure at acquiring lock. From left: Bob Taylor, Steve Vass, Osamu Miyakawa, Rob Ward, Alan Weinstein, Virginio Sannibale, Seiji Kawamura (visiting from TAMA), Hartmut Grote (visiting from GEO).