|The Web Newsletter||Front Page||Livingston||MIT|
LIGO Hanford Observatory NewsMode Cleaner Lights Up!
Smoking Out Signals in the Control Room
The Mode Cleaner, the heart of the input optics for the 2-km interferometer at LIGO Hanford Observatory, resonated for the first time this month with light delivered by the prestabilized laser.
In Figure 1 at left, MIT graduate student Rana Adhikari (standing) and Hanford scientist Haisheng Rong (seated) share a victory smile as the first pictures of success appear on the closed-circuit TV monitor. The mode cleaner is an optical cavity, approximately twelve meters long, composed of three suspended mirrors that sit on vibration-isolated optical tables in the vacuum of the HAM (Horizontal Access Module) chambers on the input side of the interferometer. The mode cleaner and its associated input optics were built by an experimental team from the University of Florida that was tightly interfaced to the lasers and optics group at Caltech for design and fabrication. Haisheng Rong was technical liaison at the Hanford end. MIT scientists and the Caltech controls group provided design and fabrication of the locking electronics.
The job of the mode cleaner is to receive the powerful (~ 8 W), but imperfect, output of the prestabilized laser system; isolate and purify a single set of spatially perfect, sinusoidal light waves; and then pass them onto the main interferometer. The mode cleaner achieves this perfection by resonating the light between the nearly perfect surfaces of its mirrors, which float in space with minimal vibrations. A single pattern of light waves is capable of reproducing itself as it reflects again and again among the mirrors. Energy is continuously extracted from the laser beam at the input mirror and cloned into this perfect pattern of light within the mode cleaner, much the way that the energy from a bow "skritching" over the string of a violin is extracted into the beautiful notes of that instrument. The oscillation of the violin string builds up over many skritches as the bow alternately sticks and slips over the string. On each oscillation, a small fraction of that violin string's energy, contained in the vibrating modes of the string, leaks into the surrounding air to make music. Our case here is similar. The light builds up to high power levels in the mode cleaner over many oscillations while a small transmission in the mode cleaner's output mirror allows the perfect optical wave to slowly leak out for use in the main interferometer. The TV monitor displays the single optical mode isolated by the mode cleaner as our photographer catches the moment.
Although the mode shown above does look beautiful, unfortunately it's the wrong one! Upon closer inspection, we see that the light spot captured on the TV monitor consists of two lobes (blotches displaced right to left) separated by a dark line in the center. This is one of the solutions of Maxwell's Equations for the light trapped inside the cavity, but it is not the solution that is capable of extracting the highest power from the input laser beam. This odd mode (called TEM10) gets excited due to a slight horizontal misalignment of the mirrors of the mode cleaner relative to the direction of the laser beam. In fact the dark line corresponds to a node of the light, a place where interference of the multiple passes is destructive, whereas the bright lobes correspond to the electric fields of the light interfering constructively but with opposite phases. A little bit of tweaking results in a better alignment and, mirabile dictu, the fundamental, disk-like, TEM00 mode appears, shown in Figure 2 at right. Note the beautiful uniformity of the disk (minus a few striations on the lower right and near the top, which are artifacts of how the beam is extracted onto the camera monitor). This fundamental mode of the mode cleaner is the one that can be coupled most efficiently from the laser beam and the one that can best be matched into the main interferometer.
Even as the mode cleaner was being resonated with light from the laser, other crews were readying optics in one 2-km arm of the interferometer. When this arm achieves resonance, it will be the largest optical cavity ever lit. Stay tuned...
For several months now, our newsletter has focused on the huge effort of getting interferometer hardware delivered from our many manufacturers across the United States and around the world. Testing this hardware to ensure quality, assembling components into structures, then vacuum-prepping and installing the equipment, requires a major effort by a great many people, not to mention of coordination to keep the operation running smoothly. And as with any complex system, glitches can occur come integration time. Bonding techniques clash with cleaning techniques. A powerful tornado wipes out a contractor's factory and suddenly your parts are scattered...who knows where. Supercleaned metallic surfaces inadvertently form cold welds and freeze up. Plus a host of other unexpected obstacles. It's all in the job description of building a first-ever engineering design.
Somewhat less picturesque, but certainly at least as complex, are the electronics, computing and software systems that knit all this hardware into a precision scientific apparatus. Pictured in Figure 1 at left is the LIGO Hanford Observatory control room as it looked in early May. The control room will be the nerve center of the Observatory where scientists and engineers can monitor the heartbeats of the two interferometers that will run at Hanford. (These are commonly called the "4K" or the "2K," after the separations of the mirrors in their long arms.) In addition, there are "housekeeping" functions, such as monitoring and controlling dust levels, as well as the lighting, heating, ventilation and air conditioning systems for the five experimental-hall buildings and the building housing the support laboratories and offices. Another big housekeeping task is controlling the vacuum system, consisting of 22 room-sized vacuum chambers, five miles of vacuum tubing, about 90,000 gallons of liquid nitrogen, and more than 500 signal and control points on gauges, pumps and valves.
A Physics Environment Monitoring (PEM) system draws in about 100 channels of data from seismometers, tiltmeters, magnetometers, electrical-line monitoring equipment, weather stations, radio-frequency monitors and a cosmic-ray-shower detector. PEM data will be used to learn how the interferometers are affected by changes in the physical environment and to check whether environmental parameters were "nominal" when candidate gravitational-wave events are recorded. The interferometers themselves generate large rates of data in addition to the gravitational-wave channel. Consider that a single suspended optical component has five shadow sensors and five voice-coil actuators that are used to position and align the optic, and damp vibrations at the resonances of the optics suspension system. Now consider that there are about a dozen suspended optics in an interferometer! There are special photosensors that study the phase of light selected at various points throughout the interferometer from which we deduce the lengths and orientations of the resonant optical cavities formed by these mirrors. Then there is the laser system, with multiple servocontrol mechanisms that lock the laser frequency to a reference cavity, stabilize the laser intensity, and prefilter the mode shape. All told, the flood of incoming data flowing through the optical fibers at Hanford is about six Mbytes/s (enough to fill four Zip disks a minute if left uncompressed).
Our Hanford engineers--Dave Barker, Richard McCarthy, Josh Meyers and Chris Patton--supported by a large control and data system group run by Rolf Bork at Caltech, have been rapidly building up and troubleshooting the data acquisition and network capabilities to handle these blistering data speeds. Some of the hardware involved is pictured above. Starting with Figure 2 at left, we see a rack of mirror suspension controllers, followed in Figure 3 by a data collection unit. Next is seen the ATM fiber network controller in Figure 4, followed by a SUN 450 frame builder in Figure 5. Frame builders pack the collected data into a standardized format that will be readable by all gravitational-wave observatories around the world. Network data servers feed data out to data-monitoring number crunchers that can process significant amounts of data "on the fly" to diagnose the health of the machine and ensure that good data is being recorded. Three separate computing networks have been established to handle different functions. CDSnet (control and data system) handles the massive real-time acquisition, monitoring and mass storage functions, and supports control room operating stations like the one pictured in Figure 6 at right. LDASnet supports the on-line LIGO Data Analysis System, a parallel computing facility that reduces the data and does a "quick-look" scan of the data for gravitational-wave signals. GCnet supports the general computing functions for the scientists and engineers at the observatory, including e-mail, web services, on-line journals, document delivery and library services, as well as scientific computing and modeling, CAD and drawing services, business software, and more.
We are just beginning the process of testing and analyzing the first data channels with our software tools and we expect to start installing LDAS computing capacity early this summer. Around this time also the data system will begin supporting research teams composed of high-school teachers and students who from their classrooms will eventually monitor and analyze some of LIGO's PEM data as part of research-based course offerings at their schools. But that is a story for a future time...