Laser Interferomter Gravitational-Wave Observatory

Go to LIGO Home
LIGO 1998 Summer Undergraduate Research Projects

Caltech | Hanford

  Some pictures from the 1998 field trip to the LIGO Hanford Observatory!
  Students inspecting Hanford site Movable clean-room assembly inside LVEA Raab, Savage, and students Students framed by vacuum chambers

Projects at Caltech

  Laser Intensity Stabilization
Kaiwen Xu
Mentor: Eric Black

The principal objective of this project is to get a well stabilized laser intensity by the use of a feedback control circuit. The laser I have been working on is a standard commercial YAG Diode-pump laser, which has a power adjustment input. By applying a positive or negative voltage to the input, the laser output intensity could be modulated around a relatively stable value (~500mW). The idea of a feedback servo is to get the fluctuation signal of the intensity through a photo-detector (so we get a voltage signal), properly amplify the signal and negatively feed it into the power adjustment input of the laser controller. Since the aim of the stabilization is shot noise level of the laser beam, the components of the circuit should be carefully chosen and tested through measurements. So far, the circuit has been put together with the open-loop transfer functions (i.e. frequency response) of the components measured. Theoretically, when the loop is closed, the intensity of the laser should be stabilized to the level limited mostly by the Johnson noise of the circuit components. Practically, however, there is a problem with the shielding of the laser controller. The circuit is picking up the 60Hz environmental noise from the connection to the power adjustment input. In the coming weeks, I will try to solve that problem and hopefully make the circuit work.

  Laser Frequency Noise in LIGO's Thermal Noise Interferometer
Sinead E. Quinn
Mentor: Eric Black

The objective of this project is to measure residual frequency noise in a stabilized laser as part of the Thermal Noise Interferometer (TNI).

The laser is locked to a fixed length reference cavity using the Pound-Drever-Hall technique. A measurement of the in-loop error signal gives a lower limit on the frequency noise in the laser. An analyzer cavity is then locked to the laser by actuating its temperature. The error signal from this cavity will contain both frequency noise and cavity noise and so will give an upper bound for the laser frequency noise.

Thermal noise is the major limiting factor in a gravitational wave detector and the TNI intends to study and minimize it. Laser frequency noise competes with thermal noise in the TNI and must be understood before an effective measurement can be made.

  Theoretical Properties and Fundamental Limits of Electrostatic Actuators for Use with LIGO
Nathan Becker
Mentor: Kenneth G. Libbrecht

Electrostatic actuators, which are arrays of capacitors, may be used to increase the sensitivity of the LIGO gravity-wave detectors by reducing the thermal noise. Electrostatic actuators exert a force on a dielectric mirror without touching it. This allows one to move the LIGO test mass mirrors or possibly even levitate them without spoiling the mechanical quality factor.

We investigated the theoretical properties of various electrostatic actuator configurations. Both numerical and analytic techniques were employed to study the performance, the most efficient design, and the drawbacks of electrostatic actuators for use in LIGO. In particular, the thermal noise generated by intrinsic electrical loss in the dielectric mirrors was calculated.

  Simulation of Thermal Noise in Non-Uniform Fused-Silica Fibers and Experimental Measurement of Suspended Mass Oscillation Modes
Mukund Thattai
Mentor: Phil Willems

Suspensions for high-quality pendula using uniform steel fibers are well understood. However, the fused-silica fibers that will be used to suspend test-masses for the second phase of LIGO are non-uniform by design, with thick ends tapered to a thin mid-section. Most of the loss in the suspension occurs at fiber endpoints, where bending is largest; but this is exactly where the non-uniformity of the fiber will tend to decrease Q. We want to determine the magnitude of this effect, and to suggest optimal taper designs.

We modeled the fiber numerically, using a finite-element analysis. Three experimentally relevant taper geometries were investigated: linear, exponential, and inverse-square-root. The model predicted substantial (15%-50%) decreases in Q for typical tapers relative to the uniform fiber, and showed interesting variations between different fiber geometries. These results are relevant, since the technology already exists to construct fibers with specified taper shapes.

Further losses in the LIGO suspensions occur due to the attachment of actuation magnets to the fused-silica test mass using an indium alloy. We set up a locked Michelson interferometer to measure the effect of indium solder on the oscillation modes of the suspended test-mass. This research is ongoing.

  Data Analysis and Visualization Software for LIGO Data
Steven C. Drasco
Mentor: Roy D. Williams

Einstein's General Theory of Relativity predicts gravitational radiation. This gravitational analog to electromagnetic radiation has never been directly detected. Detectors must be sensitive to changes in length on the order of one-thousandth the radius of a proton. Caltech and MIT are building the Laser Interferometer Gravitational-wave Observatory (LIGO) in a joint effort. LIGO is designed to detect gravitational radiation with enough sensitivity such that astrophysical research can be done based solely on gravitational radiation.

Existing data analysis programs written specifically for the purpose of analyzing gravitational radiation data have been merged with a professionally developed problem-solving environment. The result of this merger is a MATLAB based package that provides users with Graphical User Interface (GUI) and command line access to GRASP (Gravitational Radiation Analysis and Simulation Package). The GUI can read data from URLs or local disks. Visualization is provided via graphics and audio playback. The push of a button produces time and frequency domain plots while more advanced applications include simulations of compact binary inspirals or chirps and optimal filter based searches for such systems. Data can also be viewed in a three dimensional template space that displays the output of a grid of matched filters.

  Validation of the End-to-End Simulation Model for LIGO
Juan C. Rodriguez Domenech
Mentor: Biplab Bhawal

The End-to-End simulation program (written in C++) for the LIGO interferometer is being developed at Caltech. It provides a very user-friendly virtual laboratory and allows the users to perform various computer experiments to test interferometer performance.

The principle objective of this project was to validate some optical parts of this program. For this purpose various optical configurations were setup and their transfer functions (i.e. demodulated signals with respects to dither in some mirror) at different output ports were measured by a computer-equivalent of a spectrum-analyzer. These results were compared with those of another program called "Twiddle" which was previously developed at Caltech using Mathematica for the purpose of measuring only transfer functions of such optical setups.

Specifically, we studied a Fabry-Perot cavity, the recycling cavity of the LIGO interferometer (basically a recycled Michelson interferometer) and the full LIGO interferometer. Some discrepancies between the results of Twiddle and End-to-End model were observed. This helped the End-to-End modeling team to focus on those problems (related to both physics and programming), and to remove the discrepancies.

  Making an Improved Optical Spectrometer
Micah Boyd
Mentor: William P. Kells

I spent my summer in the Laser Interferometer Gravitational Wave Observatory 40 Meter Lab studying ways to more accurately measure the power of the laser sidebands. This measurement is performed in the 40 Meter Lab with a scanning optical spectrometer. This instrument is a Fabry-Perot optical cavity with one moveable mirror on a piezoelectric transducer. The mirror spacing is varied rapidly, producing an output signal which may be displayed on an oscilloscope as a spectrum of the laser power. These signals are quite small, and external disturbances of the spectrometer through effects such as heating or acoustic vibration cause undesirable degradation of the signal. These effects were thoroughly studied, and several methods were developed to stabilize the output signal. I constructed a thermally and acoustically insulated enclosure with an electronic temperature control system. I also created a more refined experimental setup to help ensure the critically sensitive alignment of the resonant cavity.

  Non-Gaussian Noise in the LIGO Interferometer
Francisco Acevedo
Mentor: Dennis Coyne

The LIGO interferometer will measure test mass displacements on the order of 1018 meters. Any sort of mechanical noise produced by its components, such as due to spontaneous strain relaxation events, will affect its sensitivity and make gravitational wave detection very difficult. The objective of this project is to test some of the components of the interferometer for the production of non-Gaussian noise. The suspects we investigated are the coil springs used in the isolation stack, the cable used to connect the electronics, and the bolts used throughout the interferometer.

The experimental setup consists of a vacuum chamber with a two stage isolation stack placed on a floatable optical table. The top of the isolation stack is configured according to the component being tested. The sensor is a piezo electric transducer connected to a low noise amplifier.

Null tests were repeated several times until the proper settings were obtained. The results of the cable test showed that it met the requirements for LIGO. The data for the spring test shows a source of interference, but further analysis is required to show if the spring is the actual source. The design for the bolt test was developed, and the piece that will be used to test the bolts was built in the shop.

  Searching for Gravitational Radiation from Pulsars
Réjean Dupuis
Mentor: Stuart Anderson

Several gravitational-wave detectors are currently being built around the world. The probability that those detectors are successful in their search will depend on our ability to extract signals that are buried in noise. This is especially true for pulsars, rapidly rotating neutron stars, where the signals are expected to be weak. In order for a pulsar to emit gravitational radiation it must be non-axisymmetric. Some mechanisms which lead to this breaking of symmetry will be discussed. Code has been developed to search for periodic signals using the parallel computers available at Caltech. This algorithm was applied to data from the 40-meter prototype interferometer. The search strategy used involve coherent accumulation of data from the interferometer over a long period of time. Frequency shifting due to the motion of the earth and spindown of the pulsars are effects that must be taken into consideration. Results from a search for known pulsars will be presented. Further work is required in order to expand the search to unknown pulsars.

  LIGO Data Analysis in the Root Environment
Efrem A. Tekle
Mentor: Walid A. Majid

ROOT, an interactive object-oriented data analysis framework developed at CERN, has been selected as a candidate LIGO data analysis tool. The main features of ROOT are its ability to analyze large volumes of data, a C/C++ interpreter, an object-oriented framework, and interface to FRAME-a common data format adopted by the Gravitational-Wave community-and its visualization tools. Since ROOT, however, was not initially intended for gravitational data analysis, additional signal processing tools were needed.

Our goal was to develop these signal-processing tools in the ROOT environment. Currently we are developing Line Removal techniques, a process that removes a narrow frequency band and its harmonics from a given data set. When its development is completed, it will be used as part of the LIGO data analysis package.

  The LIGO 40 Meter Interferometer
Jameson G. Rollins
Mentor: Richard Gustafson

The LIGO 40-meter interferometer on the Caltech campus was built in the late '80's as a 1/100-scale model and testing ground for the full-scale 4-kilometer LIGO interferometers now under construction in Hanford, Washington and Livingston, Louisiana. My work here at the 40-meter, which will extend ultimately until the end of the year, is to assist in the development and implementation of the power recycling Michelson subsystem of the 40-meter interferometer. Phase One of the full-scale LIGO interferometers will utilize what is know as a recycling mirror placed on the input side of the beam splitter to recycle the signal leaving the interferometer out of the input back into interferometer. The 40-meter has recently been fitted with a power-recycling mirror in order to model the full-scale design. Along with this primary goal, I have also been helping out in all other aspects of the upkeep of the interferometer. This has involved the development of various other interferometer improvements, assistance in reaching the overall goal of the lab (namely getting the interferometer up to its full working potential and maintaining it at a stable working condition for the taking of data), and, of course, to gain experience and knowledge in the operation of an interferometer as a gravity wave detector.

Up to this point, my biggest project has been the development of a beam stop system to assist in the alignment of the interferometer subsystems. There are three main parts to the interferometer. Each arm of the overall Michelson interferometer is itself a Fabry-Perot cavity consisting of two parallel mirrors, one at the end of each arm and another near the beam splitter. These vertex mirrors, as they are called, along with the beam splitter and the recycling mirror constitute the power recycling Michelson interferometer subsystems. All of these interferometer subsystems are controlled by servos to keep them in a stable configuration. Each of these subsystems needs to be properly aligned. The beam stops that I have been developing will go in the vacuum, on either side of the vertex masses, so that each subsystem can be isolated to expedite the alignment process.

This task has proven quite challenging. Everything that will be placed in the vacuum system must undergo an extremely rigorous evaluation and cleaning process to assure that the integrity of the vacuum is maintained. The beam stops had to meet certain strict mechanical criteria as well, such as be able to move reliably in an out of the path of the beam, so as not interfere with the normal operation of the interferometer.

This project is nearing completion and I have begun work on other projects needed for the upkeep of the system and the development of the power recycling subsystem.


Projects at the LIGO Hanford Observatory


Simulations of the Alignment Sensitivity in LIGO Lock Acquisition
Martin R. Zwikel
Mentor: Daniel Sigg

The laser interferometric gravitational wave observatories (LIGOs) presently under construction in Hanford, WA and Livingston Parish, LA aim to detect a gravitational wave strain on the order of 10-21. To obtain this sensitivity, the interferometer must first be length locked on resonance. A scheme has been devised which sequentially locks the various LIGO cavities, but misalignments in the optical elements of these cavities reduce the signals used in lock acquisition. Of particular interest is the nominal "critical alignment," where length signals are degraded to half their aligned strengths. By simulating the LIGO lock signals for seismically driven mirror motions, we determine the critical alignment angles of the system. The final, and most sensitive, stages of lock acquisition are examined, with all possible misalignments of the optics expressed in a basis which is diagonal to signal sensitivity. We find a mirror speed independent critical alignment at 2x10-7 rad in the more sensitive angular degrees of freedom, while the other degrees have a lesser, but speed and stage dependent effect. The alignment signals and power readouts which will aid in lock acquisition are also examined, and are seen to remain useful at this critical alignment.

  RGA Analysis and Hydrocarbon Contamination
Betsy Weaver
Mentor: Frederick J. Raab

The intent of this project was to study the effects of contamination inside the LIGO Hanford Observatory bake oven, particularly hydrocarbon contamination with atomic mass units 41, 43, 53, 55, and 57. The scope of this project included the construction of the bake oven, heating system assembly, and control panel assembly. The completed apparatus was then evaluated to ensure that it would meet demands on the system, such as acceptable pump down pressures, and temperature stability. Once these systems were in place and tested, a residual gas analyzer (RGA) was used to analyze the gaseous content of the bake oven. The analysis consisted of identifying species in the bake oven, observing how well they are pumped out via turbopumps, as well as by heating to elevated temperatures over time. Pump speeds, outgassing rates, and dwell times for the five hydrocarbons were calculated. Further analysis will determine how the contaminants move around in the beam tube and what surfaces they stick to, making elimination of them more difficult. Combined with the previous measurements and calculations, this analysis work will aid in the modeling of potential contamination of the LIGO mirrors.

  Performance Evaluation of the LIGO Observatory Technical Slab Foundations
Tanim Islam
Mentor: Richard Savage

The Parsons Infrastructure Company designed and constructed the Hanford observatory in order to meet the stringent vibration isolation criteria that LIGO requires. Phase and acceleration measurements at Hanford were performed on the y-arm end station to assess vibration levels resulting from equipment noise, specifically from the air handling fans housed in the machine room. Measured fan-induced vibrations levels are below the LIGO requirements. The fan isolation was shorted to provide measurable peaks above background at the fan rotation frequency and harmonics (29.75 Hz, 59.5 Hz, and 89.25 Hz). As expected, the concrete foundation and air gaps attenuate vibrations better at higher frequencies. The vibration level for the fan is better than expected; the Parsons model, it appears, is conservative by approximately a factor of ten.

  Missing abstract for Janet Casperson.